|Home | About | Journals | Submit | Contact Us | Français|
Overexpression of the inhibitory Smad, Smad7, is used frequently to implicate the Smad pathway in cellular responses to transforming growth factor β (TGF-β) signaling; however, Smad7 regulates several other proteins, including Cdc42, p38MAPK, and β-catenin. We report an alternative approach for more specifically disrupting Smad-dependent signaling using a peptide aptamer, Trx-SARA, which comprises a rigid scaffold, the Escherichia coli thioredoxin A protein (Trx), displaying a constrained 56-amino acid Smad-binding motif from the Smad anchor for receptor activation (SARA) protein. Trx-SARA bound specifically to Smad2 and Smad3 and inhibited both TGF-β–induced reporter gene expression and epithelial-to-mesenchymal transition in NMuMG murine mammary epithelial cells. In contrast to Smad7, Trx-SARA had no effect on the Smad2 or 3 phosphorylation levels induced by TGF-β1. Trx-SARA was primarily localized to the nucleus and perturbed the normal cytoplasmic localization of Smad2 and 3 to a nuclear localization in the absence of TGF-β1, consistent with reduced Smad nuclear export. The key mode of action of Trx-SARA was to reduce the level of Smad2 and Smad3 in complex with Smad4 after TGF-β1 stimulation, a mechanism of action consistent with the preferential binding of SARA to monomeric Smad protein and Trx-SARA-mediated disruption of active Smad complexes.
Transforming growth factor-β (TGF-β) is a pluripotent cytokine that regulates diverse biological processes such as cell proliferation, differentiation, apoptosis, and epithelial-to-mesenchymal transition (EMT; ten Dijke et al., 2002 ; Siegel and Massague, 2003 ). TGF-β signaling has also been associated with fibrosis and cancer (Derynck et al., 2001 ; Leask and Abraham, 2004 ; Flanders, 2004 ). TGF-β signaling initiates when the TGF-β ligand binds to and actives the type II and type I receptor kinases, resulting in Smad2 and Smad3 phosphorylation. Phosphorylated Smad2 and Smad3 form heterodimeric and heterotrimeric complexes with Smad4 that accumulate in the nucleus and bind over two dozen different transcription factors, coactivators, and corepressors to activate or inhibit the expression of specific target genes (Derynck and Zhang, 2003 ; Shi and Massague, 2003 ; ten Dijke and Hill, 2004 ; Feng and Derynck, 2005 ; Massague et al., 2005 ). TGF-β also activates several Smad-independent signaling pathways that involve proteins such as Tak1, ERK/MAPK, and RhoA (Derynck and Zhang, 2003 ).
Inhibition of TGF-β signaling is an important tool for elucidating the multiple biological functions of TGF-β and has become of significant interest as a potential therapeutic strategy in fibrotic diseases and metastatic cancer (Dumont and Arteaga, 2003 ; Yingling et al., 2004 ). The ligand itself has been the target for several methods of inhibition include blocking synthesis of the ligand with antisense (Fakhrai et al., 1996 ), binding peptide inhibitors to the ligand (Santiago et al., 2005 ), or sequestering the ligand using neutralizing antibodies (Ziyadeh et al., 2000 ; Mead et al., 2003 ) or soluble TGF-β–binding proteins (Komesli et al., 1998 ; Kolb et al., 2001 ; Bandyopadhyay et al., 2002 ). The TGF-β receptors have also been targeted to disrupt the pathway by using neutralizing antibodies to the receptor (Kasuga et al., 2001 ), dominant-negative forms of the receptors (Gorelik and Flavell, 2001 ), dominant-negative forms of Smad2 or Smad3 that bind to receptor but cannot be phosphorylated (Macias-Silva et al., 1996 ) or by inhibiting the kinase activity of the receptor with small molecule compounds (Laping et al., 2002 ; DaCosta Byfield et al., 2004 ; Singh et al., 2004 ; Grygielko et al., 2005 ). Chemical inhibitors of the kinase activity, such as SB431542, block the phosphorylation of Smad2 and Smad3 (Inman et al., 2002a ) and inhibit TGF-β–induced cell proliferation and motility of glioma cells (Hjelmeland et al., 2004 ) or fibrosis in skin fibroblasts (Mori et al., 2004 ). Small molecule inhibitors have also been used to inhibit Smad-independent TGF-β activation of, for example, c-Abl (with imatinib mesylate; Daniels et al., 2004 ), phosphatidylinositol 3-OH kinase (with LY294002; Bakin et al., 2000 ), p38 mitogen–activated protein kinase (with SB202190; Bakin et al., 2002 ), or mitogen activated protein kinase (with U0126; Xie et al., 2004 ).
A natural inhibitor of TGF-β signaling, the inhibitory Smad, Smad7, is frequently constitutively expressed, or overexpressed, to demonstrate that Smad-dependent signaling is involved in a specific gene expression or biological response to TGF-β. Smad7 is induced by TGF-β and binds to both TGF-β type I receptor and bone morphogenetic protein (BMP) type I receptors to inhibit phosphorylation of the regulated Smads (R-Smads; Nakao et al., 1997 ; Souchelnytskyi et al., 1998 ; Mochizuki et al., 2004 ) and to target the type I receptors for proteolytic degradation (Kavsak et al., 2000 ; Ebisawa et al., 2001 ). Smad7 targets the type I receptors for degradation in the lipid raft-caveolar internalization pathway by recruiting Smurf ubiquitin ligase to the receptors (Kavsak et al., 2000 ; Di Guglielmo et al., 2003 ; Murakami et al., 2003 ). Smad7 also functions as a scaffold or adaptor protein to assemble GADD34 and protein phosphatase 1 (PP1) at the activated type I receptor, where PP1 de-phosphorylates and inactivates the type 1 receptor (Shi et al., 2004 ). However, Smad7 appears to have a variety of other functions. In some cells, e.g., PC-3U and HaCaTs, Smad7 mediates TGF-β–induced apoptosis (Landstrom et al., 2000 ). Overexpression of Smad7 interferes with the prosurvival functions of NFκB and activates apoptosis through stress-activated MAPKs (Lallemand et al., 2001 ; Yu et al., 2002 ; Undevia et al., 2004 ). Smad7 serves as a scaffold for assembling a complex of TAK1-MKK3 and p38 MAPK that causes activation of p38MAPK in PC3U cells (Edlund et al., 2003 ). One function of activated p38 MAPK in these cells is to inactivate GSK-3β, leading to stabilization and activation of β-catenin–mediated apoptosis (Edlund et al., 2005 ). Smad7 also forms a protein complex with β-catenin and LEF1/TCF in response to TGF-β to facilitate transcriptional activation of β-catenin responsive genes such as Myc (Edlund et al., 2005 ). This may be cell type specific in that Smad7 activates a different MAPK pathway, the SAPK/JNK pathway, in MvLu1 cells and MDCK cells to trigger apoptosis (Lallemand et al., 2001 ). Smad7 is localized in the nucleus and can also interact with histone deacetylase 1 (Bai and Cao, 2002 ). Smad7 is also reported to mediate TGF-β activation of Cdc42, reorganization of actin filaments, and membrane ruffling (Edlund et al., 2004 ). This multitude of Smad7 functions indicates that the consequences of overexpressing Smad7 perhaps should not be interpreted as simple inhibition of Smad-dependent signal transduction.
We have been interested in targeting the transcriptionally active Smad complexes to inhibit signaling and provide selective inhibition of Smad-dependent TGF-β responses. To develop such reagents, we are identifying constrained peptide inhibitors that are targeted to Smad proteins (Cui et al., 2005 ). Constrained peptide motifs expressed on a rigid scaffold protein, also called peptide aptamers, have been used to disrupt signal transduction events in mammalian cells by binding to a target protein and disrupting normal protein–protein interactions, e.g., cyclin-dependent kinase 2 (Colas et al., 1996 ), E2F-DP1 (Fabbrizio et al., 1999 ), p53-mdm2 (Bottger et al., 1997 ), epidermal growth factor (EGF) receptor-Stat3 interaction (Buerger et al., 2003 ), and Stat3-DNA binding or Stat3 dimerization (Nagel-Wolfrum et al., 2004 ). The development of protein scaffolds began with the recognition that immunoglobulins function through the use of a conserved framework region and a spatially defined hypervariable region. Subsequently several proteins of small size, high stability and ease of production, such as E. coli thioredoxin A (Trx), have been tested as protein scaffolds (Skerra, 2000 ). The benefits of expressing a constrained peptide on a scaffold can include improved binding affinity, e.g., a peptide aptamer that inhibits cyclin-dependent kinase 2 has an IC50 1000-fold lower than the free 20-amino acid peptide (Colas et al., 1996 ). The aptamers to cdk2 also showed selective inhibition in blocking the interaction of cdk2 with histone H1 but not affecting the interaction with Rb (Cohen et al., 1998 ). Importantly, peptide aptamers on the E. coli thioredoxin A scaffold are efficiently expressed in mammalian cells, e.g., expression of a Trx peptide aptamer from a CMV promoter in Saos-2 cells inhibited cell cycle progression (Cohen et al., 1998 ), and introduction of a 12 amino acid Mdm2-binding Trx peptide aptamer into cells caused an accumulation of p53 protein and activation of p53-responsive reporter genes (Bottger et al., 1997 ). We have reported that Trx aptamers containing peptides from xFoxH1b and Lef1 bind to Smads and selectively inhibit TGF-β responsive reporter genes (Cui et al., 2005 ). The Trx-Lef1 aptamer inhibited TGF-β activation of the TwnTop-luciferase reporter and Trx-XFoxH1b aptamer inhibited TGF-β activation of the A3-luciferase reporter. The activities of several other reporter genes were not altered by either aptamer (Cui et al., 2005 ).
We have now generated a peptide aptamer that displays the Smad-binding domain from the protein “Smad anchor for receptor activation” (SARA) on the Trx scaffold. SARA was originally identified as a Smad2-binding protein and is localized to the early endosomal membrane through a double zinc finger FYVE domain (Tsukazaki et al., 1998 ; Itoh et al., 2002 ). SARA also binds to the type I TGF-β receptor and recruits Smad2 or Smad3 to the type I receptor for phosphorylation (Tsukazaki et al., 1998 ). Phosphorylation of Smad by the type I receptor leads to release of the Smad from SARA and assembly of multimeric Smad2/4 and Smad3/4 complexes (Tsukazaki et al., 1998 ; Qin et al., 2002 ). The Smad-binding domain (SBD) of SARA, amino acids 665–750, is sufficient to bind Smad2 or Smad3 in vitro and was cocrystallized with the MH2 domain of Smad2 (Wu et al., 2000 ; Qin et al., 2002 ). Three structural motifs of the SARA SBD bind along the hydrophobic groove of the Smad MH2 domain. The MH2 monomer bound to SARA is in a different conformation than the MH2 incorporated into a heterotrimeric complex with Smad4, consistent with biochemical data showing that SARA preferentially binds to the monomeric form of Smad and can disrupt the Smad3–4 complex in vitro (Wu et al., 2000 ).
Mouse mammary gland (NMuMG) cells were obtained from Dr. Caroline Alexander (University of Wisconsin-Madison, WI). NMuMG cells were maintained in DMEM and supplemented with 10% fetal bovine serum (FBS), 10 μg/ml insulin, 100 μg/ml penicillin, and 100U/ml streptomycin. 293 cells were obtained from ATCC (Manassas, VA) and grown in 10% DMEM and supplemented with 10% FBS, 100 μg/ml penicillin, and 100 U/ml streptomycin. All cells were grown at 37°C in 5% CO2. TGF-β1 was purchased from R&D Systems (Minneapolis, MN). TNF-α was purchased from Sigma-Aldrich (St. Louis, MO). Human recombinant EGF was purchased from Upstate Signalling (Charlottesville, VA).
For mammalian expression of the aptamers, fragments of human SARA (accession AF104304) were obtained by PCR using primers with terminal RsrII sites and cloned into the unique RsrII site of pCI-NLS-HA-Trx (Cui et al., 2005 ). The specific amino acid sequences included in each aptamer are shown in Figure 1.
The hemagglutinin (HA)-tagged Trx-SARA and the Flag-tagged Smad7 were cloned into the pCMMV-IRES-GFP vector (Ory et al., 1996 ; Kennedy and Sugden, 2003 ) using the NotI restriction site. Plasmids were transfected in 293 cells to generate virus as described previously (Cui et al., 2005 ). NMuMG cells were infected with virus multiple times to generate NMuMGTrx-GA, NMuMGTrx-SARA, and NMuMGSmad7cells. Cells were sorted by flow cytometry (UWCCC Flow Cytometry Facility, Madison, WI) for GFP intensity, and the brightest 10% of the cells were collected in 2 ml of FBS. These cell populations were expanded and used for all experiments.
Cells were washed with ice-cold phosphate-buffered saline (PBS) and lysed with TNE buffer (10 mM Tris-HCl, pH 8.0, 100 mM NaCl, 1% Nonidet P-40, 1 mM EDTA) supplemented with proteinase inhibitors (Roche Applied Science, Indianapolis, IN) and phosphatase inhibitors (1 mM EDTA, 20 mM NaF, and 1 mM sodium orthovanadate). Total cell lysates were centrifuged at 14,000 rpm for 10 min at 4°C. Protein concentrations were determined by BCA assay kit (Pierce, Rockford, IL). Protein A/G plus agarose beads (Santa Cruz Biotechnology, Santa Cruz, CA) were used in coimmunoprecipitation assays. Total cell lysates or immunoprecipitates were resolved by 4–20% SDS-PAGE and transferred to PVDF membrane. The following antibodies were used: rabbit phospho-Smad2(Ser465/467), rabbit p38, phospho-p38, rabbit p42/44, phospho-p42/44, Akt and phospho-Akt (Cell Signalling Technology, Beverly, MA), rabbit anti-phospho-Smad3 (gift from Ed Leof, Mayo Clinic. Rochester, MN), N-cadherin, SP1, and Smad4 (B-8; Santa Cruz Biotechnology), anti-E-cadherin (BD Biosciences, San Jose, CA), β-actin, vimentin, Flag-M2, and β-tubulin (Sigma-Aldrich), Smad2 and Smad3 (Zymed Laboratories, South San Francisco, CA), and HA-peroxidase (Roche Applied Science, Indianapolis, IN).
About 10,000 NMuMG cells were seeded in 12-well plates. After 24 h, NMuMG cells were transfected with reporter constructs and CMV-β-galactosidase plasmid using LipofectAMINE 2000 (Invitrogen, Carlsbad, CA) or Fugene 6 (Roche Applied Science). Six hours after the transfection, cells were treated with or without TGF-β (100 pM) in 0.2% FBS DMEM. After 20 h, cells were lysed and luciferase (Promega, Madison, WI) and β-galactosidase assays (Galacto-Star, Tropix, Bedford, MA) were performed according to manufacturer’s instructions. For the 3xκb-luc reporter assay, NMuMG cells were treated with 10 ng/ml TNF-α for 12 h. Luciferase values were normalized to the β-gal activity.
Cells were grown on coverslips (VWR) or eight-well chamber slides (VWR). After the TGF-β (100pM) treatment, cells were fixed with 4% (wt/vol) paraformaldehyde in PBS and permeabilized with 0.5% Triton X-100 in PBS for 1 min. Cells were incubated with blocking solution (3% BSA in PBS). Primary antibodies were added at 1:200 dilutions in blocking solution, incubated for at least 1 h at room temperature, and then washed three times with 0.05% Triton X-100 in PBS. Subsequently, secondary antibodies diluted in blocking solution were added and incubated for 30 min. Finally, cells were washed with PBS and mounted with DAPI (Santa Cruz Biotechnology) and visualized by fluorescence microscopy (Carl Zeiss, Thornwood, NY). The following antibodies were used: Texas red–conjugated phalloidin (Sigma-Aldrich), anti-Smad2 and anti-Smad3 antibody (Zymed Laboratories), and anti-E-cadherin antibody (BD Biosciences). Alexa Fluor 594 goat anti-rabbit IgG, Alexa Fluor 594 anti-mouse, and Alexa Fluor 594 anti-HA mouse antibody (16B12) were purchased from Invitrogen. Images were processed using Adobe Photoshop software (Adobe Systems, Mountain View, CA).
Cells grown to 80% confluency were washed with ice-cold PBS. Cytoplasmic and nuclear fractions were extracted using Nuclear and Cytoplasmic Extraction Reagent (Pierce, Rockford, IL) following the manufacturer’s instructions. Nuclear and cytoplasmic protein concentrations were determined by BCA assay (Pierce).
Cells grown in P100 plates were lysed with TNE buffer (described above). Equal amounts of total cell lysate were incubated overnight at 4°C with streptavidin-agarose beads (Pierce) and 1 μg of the double-stranded SBE4 oligonucleotide (biotin-5′-GTACCCAGACAGTCAGACAGTCAGACAGTCAGACAGTC-3′; Warner et al., 2003 ). Streptavidin-agarose beads were washed four times with TNE buffer. Bound proteins were resolved by SDS-PAGE on a 4–20% gel.
In a previous study, the minimum Smad-binding domain (SBD) of SARA was identified as amino acids 665–721 (Wu et al., 2000 ). We introduced this 56 amino acid SARA SBD into the active site loop of E. coli Trx to generate the Trx-SARA aptamer. An HA epitope tag and nuclear localization sequence were added to the N-terminus of the Trx scaffold (Figure 1A). The SARA SBD contains three motifs: the rigid coil, the amphipathic α-helix and the β-strand, which are all implicated in binding to the hydrophobic groove of the Smad MH2 domain (Wu et al., 2000 ; Figure 1A). To determine the importance of these secondary structure elements in mediating the aptamer–Smad interaction, we constructed a series of deletion mutants of the Trx-SARA, the N-terminal deletion (removes rigid coil), the C-terminal deletion (removes the β-strand) and the N and C terminal deletion (removes both the rigid coil and the β-strand; Figure 1A). We tested whether the Trx-SARA SBD aptamer bound to Smad2 in mammalian cells by coimmunoprecipitation. Expression plasmids for Flag-tagged Smad2 and HA-tagged Trx-SARA SBD were cotransfected into 293 cells. Cell extracts were immunoprecipitated with anti-Flag antibody, followed by immunoblotting with anti-HA antibody. As shown in Figure 1B, Trx-SARA was detected in the Flag-Smad2 immunoprecipitate. None of the Trx-SARA deletion mutant proteins were detected in the Flag-Smad2 immunoprecipitate, suggesting that the Trx-SARA aptamer required all three structural elements to bind Smad2.
Eight mammalian Smads have been identified, including Smad1, 5, and 8, the R-Smads in the BMP signaling pathway, and Smad2 and Smad3, the R-Smads in the TGF-β signaling pathway. Smad6 and 7 are inhibitory Smads in the TGF-β– and BMP-signaling pathways. HA-tagged Trx-SARA aptamer was expressed in 293 cells together with different Flag-tagged Smad proteins. Examination of anti-Flag immunoprecipitates by anti-HA westerns indicated that the Trx-SARA protein associated with Smad2 and Smad3 but not Smads 1, 4, or 7 (Figure 1C). A control Trx aptamer, Trx-GA, containing an 11-amino acid repeat of Gly-Ala was not detectable in any of the immunoprecipitates. These data indicate that Trx-SARA maintains the same Smad-binding specificity as full-length SARA (Tsukazaki et al., 1998 ).
To determine the transcriptional effects of Trx-SARA, we tested several TGF-β–responsive reporter genes in NMuMG cells. Expression of Trx-SARA greatly reduced the TGF-β–induced luciferase expression from three reporter genes: SBE(CAGA)12-luc, 3TP-luc, or PAI-1-luc (Figure 2, A–C). To test the specificity of Trx-SARA inhibition, we also used a non-TGF-β–responsive luciferase reporter, the 3xκB-luc reporter, which is activated by TNF-α treatment (Huang et al., 2000 ). As shown in Figure 2D, Trx-SARA caused no inhibition of the 3xκb reporter.
To study the effects of Trx-SARA on TGF-β–induced responses in NMuMG cells, we generated retroviral expression vectors that constitutively expressed the aptamers and GFP, or Smad7 and GFP. NMuMG cells were infected with retroviruses encoding Trx-SARA, Trx-GA, or the inhibitory Smad protein, Smad7. High-GFP–expressing cells were recovered by FACS and expanded in culture to provide populations of cells (NMuMGTrx-SARA, NMuMGTrx-GA, and NMuMGSmad7) uniformly expressing the respective aptamer or Smad7. We transiently transfected the 3TP luciferase reporter into these three cell populations and demonstrated that Trx-SARA inhibited TGF-β–induced luciferase activity by 75%, whereas Smad7 inhibited the activity by 85% (Figure 2E). We next examined whether overexpression of Smad2, Smad3, and Smad4 could rescue the inhibitory effect of Trx-SARA and Smad7 on 3TP-luc reporter activity. As shown in Figure 2E, transfection of Smad2/3/4 increased basal 3TP reporter activity in all three cell populations and increased 3TP-luciferase activity in TGF-β–treated NMuMGTrx-SARA and NMuMGSmad7cells. The inability of Trx-SARA to inhibit activation of the reporter by the combination of Smad2/Smad3/Smad4 in the absence of ligand-induced phosphorylation of the R-Smads suggests that Trx-SARA may be acting on the phosphorylated R-Smads or the trimeric complex formed by phosphorylated R-Smads.
TGF-β induces EMT, which comprises loss of epithelial cell morphology, dissociation of cell–cell contacts, reduction in proteins mediating cell–cell contacts, remodeling of the actin cytoskeleton, and acquisition of an elongated, mesenchymal cell shape (Thiery, 2002 , 2003 ). In normal development, EMT has been associated with processes in gastrulation, heart formation, somitogenesis, palate formation, and Mullerian tract regression (Savagner, 2001 ; Shook and Keller, 2003 ). EMT has been causally linked to tumor invasion and metastasis (Gotzmann et al., 2004 ). In renal fibrosis, tubular epithelial cells undergo EMT in response to primary insults such as hypertension or diabetes, which results in the production and deposition of extracellular matrix proteins (Zeisberg and Kalluri, 2004 ). Multiple signaling proteins have been implicated in the induction of EMT by TGF-β, including Ras/MAPK (Oft et al., 1996 , 1998 , 2002 ), integrin β-1 (Bhowmick et al., 2001a ), integrin-linked kinase (Li et al., 2003b ), p38MAPK (Bakin et al., 2002 ), Rho kinase (Bhowmick et al., 2001b ), phosphatidylinositol 3 (PI3)-kinase (Bakin et al., 2000 ), Jagged1/Notch (Zavadil et al., 2004 ), and NF-κB (Huber et al., 2004 ). The Smad proteins are also important for the EMT response as demonstrated by studies on Smad2 (Oft et al., 2002 ), overexpression of Smad2 or Smad3 with Smad4 (Piek et al., 1999 ), use of a type I receptor mutant that is unable to phosphorylate R-Smads (Itoh et al., 2003 ), use of conditional Smad4 knockout mice (Li et al., 2003a ), use of Smad3 knockout mice (Sato et al., 2003 ; Saika et al., 2004 ; Zavadil et al., 2004 ), and use of dominant negative Smads (Valcourt et al., 2005 ).
TGF-β–induced EMT in NMuMG cells was assayed by the morphological change of the cells and by a number of well-characterized molecular changes. We used the cell populations generated by retroviral infection and cell sorting for GFP expression to test the effect of the aptamers on the EMT response in NMuMG cells. In the control NMuMGTrx-GA cells, treatment with 100 pM TGF-β for 24 h induced cells to change shape from a cobblestone-like appearance of epithelial cell morphology to a more elongated mesenchymal cell morphology; this morphological change was more apparent at 48 h (Figure 3A). Neither NMuMGTrx-SARA nor NMuMGSmad7 cells changed morphology in response to TGF-β (Figure 3A). To verify that Trx-SARA and Smad7 expression in NMuMG cells blocked TGF-β–induced EMT, we examined several epithelial and mesenchymal molecular markers in the cells. As shown in Figure 3B, NMuMGTrx-GA cells treated with 100 pM TGF-β for 48 h lost E-cadherin staining. In contrast, the epithelial cell marker E-cadherin was detected at the junctions between neighboring NMuMGTrx-SARA cells even in the presence of 100 pM TGF-β (Figure 3B). NMuMGSmad7 cells showed discontinuous and diffuse E-cadherin staining at the cell–cell junction. We also examined stress fiber formation by phalloidin staining and the expression pattern of vimentin. As shown in Figure 3B, F-actin distribution was mainly at the cell–cell borders and no stress fibers were observed in non-TGF-β–treated cells. However, when cells were treated with TGF-β for 48 h, the F-actin distribution changed, and stress fibers were formed throughout NMuMGTrx-GA cells. In contrast, NMuMGTrx-SARA and NMuMGSmad7 cells did not exhibit mesenchymal patterns of phalloidin or vimentin straining (Figure 3B). Western blot analysis was performed to analyze the expression level of E- and N-cadherin. As shown in Figure 3C, E-cadherin levels decreased, whereas N-cadherin increased in NMuMGTrx-GA cells in response to TGF-β treatment. In contrast, there was no detectable change in E- or N-cadherin levels in NMuMG cells expressing Trx-SARA or Smad7. Western analysis confirmed that the NMuMGTrx-SARA and NMuMGTrx-GA cell populations expressed similar levels of HA-tagged Trx protein (Figure 3C).
To determine if expression of Trx-SARA blocked TGF-β–induced phosphorylation of Smad2 and Smad3, we analyzed the phosphorylated Smad2 and Smad3 level in the three cell populations: NMuMGTrx-GA, NMuMGTrx-SARA and NMuMGSmad7 (Figure 4). Western blot analysis of total cell lysates from NMuMGTrx-GA and NMuMGTrx-SARA cells indicated a reduced level of phospho-Smad2 protein in NMuMGTrx-SARA cells 15 min after addition of 100 pM TGF-β (Figure 4A). By 30 min after TGF-β addition, however, the levels of phospho-Smad were similar in both cell populations, and this similarity was maintained through the 24-h time course (Figures 4, A and B). By contrast, expression of Smad7 in NMuMG cells decreased levels of phosphorylated Smad2 and Smad3 (Figure 4B).
p38 mitogen–activated protein kinase (Bhowmick et al., 2001a ), the phosphoinositide-3′kinase, Akt (Bakin et al., 2000 ), and the ERK/MAPK (Xie et al., 2004 ) also contribute to TGF-β–induced EMT in NMuMG cells. Treatment with TGF-β for 1 or 2 h moderately increased phosphorylation of Erk/MAPK, p38MAPK, and Akt in NMuMGTrx-GA cells (Figure 4C), and these increases were not altered by expression of Trx-SARA. Reduced TGF-β–induced phosphorylation of p42/44 ERK/MAPK and higher phosphorylation of p38 and Akt were observed in NMuMGSmad7 cells. The increased phosphorylation of p38 is in agreement with reports that Smad7 can serve as a scaffold for assembly of kinases that activate p38 kinase (Edlund et al., 2003 ). Phosphorylation of ERK/MAPK p42/44 or p38MAPK by EGF or osmotic shock, respectively, were not altered in the NMuMGTrx-SARA and NMuMGSmad7 cells (Figure 4, D and E).
Smad2 and Smad3 are continually shuttling between the nucleus and cytoplasm even in their nonphosphorylated state (Inman et al., 2002b ; Xu et al., 2002 ). A previous study demonstrated that overexpression of a fragment of SARA prevented Smad2 nuclear import into isolated nuclei by masking the nuclear import signal of Smad2 (Xu et al., 2000 ). Trx-SARA and the control aptamer, Trx-GA, were both predominantly localized in the nucleus of the NMuMG cell populations by immunostaining for the HA-epitope tag (Figure 5A). In the nonstimulated NMuMGTrx-GA cells, Smad2 (Figure 5B) and Smad3 (Figure 5C) were distributed in both the cytoplasm and the nucleus. After TGF-β stimulation, Smad2 and Smad3 were localized predominantly in the nucleus, consistent with normal Smad protein translocation in the Trx-GA cell population (Figure 5, B and C). In contrast, Smad2 and Smad3 were localized in the nucleus in NMuMGTrx-SARA cells even without TGF-β stimulation (Figure 5, B and C). Treatment of NMuMGTrx-SARA cells with TGF-β only slightly increased Smad2 and Smad3 nuclear levels. In the NMuMGSmad7 cells, the Smad2 and Smad3 cellular localization was cytoplasmic, similar to GA cells in the absence of TGF-β (Figure 5, B and C). When NMuMGSmad7 cells were treated with TGF-β, there was much less nuclear Smad2 or Smad3 staining detected, indicating that nuclear accumulation of Smad2 or Smad3 is reduced; this is consistent with the reduced Smad2 and Smad3 phosphorylation in NMuMGSmad7cells (Figure 4A). Smad 4 shuttles into and out of the nucleus independently from the R-Smads (Inman et al., 2002b ; Xu et al., 2002 ). Because Trx-SARA does not bind to Smad4, the cellular distribution of Smad4 should not be altered in Trx-SARA expressing cells. Immunohistochemical detection of Smad4 revealed no differences in the distribution of Smad4 between NMuMGTrx-GA and NMuMGTrx-SARA cells (Figure 5D).
To validate the immunohistochemical observations, subcellular fractionation experiments were also performed using cell populations of NMuMGTrx-GA, NMuMGTrx-SARA, and NMuMGSmad7(Figure 5E). The accumulation of Smad 3 in the nuclear fraction after TGF-β treatment was detected in the NMuMGTrx-GA cell fractions. In agreement with the immunohistochemical data, Smad3 was found predominantly in the nuclear fraction of NMuMGTrx-SARA cells even without TGF-β treatment, and treatment of TGF-β did not increase Smad3 in the nuclear fraction (Figure 5E). Although Trx-SARA apparently impedes nuclear export, leading to high levels of nuclear R-Smads in the absence of TGF-β, the normal phosphorylation levels of Smad2 and Smad3 in response to TGF-β stimulation of Trx-SARA cells (Figure 4 A and B) indicate that nuclear-cytoplasmic shuttling still occurs in the Trx-SARA cells because the Smads must have access to the cell surface receptors to maintain the phosphorylated state (Inman et al., 2002b ; Xu et al., 2002 ). When compared with the control cells after TGF-β stimulation, the TGF-β–stimulated Trx-SARA cells contain comparable levels of phosphorylated R-Smad and comparable levels of total nuclear R-Smad. Therefore, the ability of Trx-SARA to inhibit TGF-β–induced reporter genes and EMT is not explained by the altered basal subcellular localization of the R-Smads.
The assembly of phosphorylated Smad2 into a trimeric complex results in a 12 Å conformational change in the Smad MH2 domain that reduces binding to SARA through disruption of the β-sheet formed between the beta strand in SARA and the N-terminal β1′-strand of the MH2 domain (Wu et al., 2000 ). This conformational change facilitates the Smad– Smad interface binding in the heterotrimeric complexes comprising Smad4 and phosphorylated Smad2 or Smad3. Previous studies have shown that Smad2 could bind either SARA or Smad4 but not both proteins together, suggesting mutually exclusive binding of SARA and Smad4 by the two different conformations of Smad2 (Tsukazaki et al., 1998 ). We tested whether expression of Trx-SARA in NMuMG cells disrupts the Smad2–Smad4 or Smad3–Smad4 complexes. NMuMGTrx-GA, NMuMGTrx-SARA, and NMuMGSmad7 cells were treated with or without TGF-β for 2 h, and cell lysates were immunoprecipitated with anti-Smad2 or anti-Smad3 antibodies. Immunoprecipitated Smad2 (Figure 6A) or Smad3 (Figure 6B) protein complexes were analyzed by Western blot analysis using anti-Smad4 antibody. TGF-β treatment increased Smad2/4 and Smad3/4 complex formation in NMuMGTrx-GA cells. We observed lower levels of Smad2/4 (Figure 6A) and Smad3/4 (Figure 6B) in NMuMGTrx-SARA cells, suggesting that the interactions between endogenous Smad4 with Smad2 or Smad3 were disrupted by the binding of Trx-SARA to Smad2 or Smad3. Consistent with this ability to disrupt or reduce the formation of R-Smad-Smad4 trimers, the Trx-SARA aptamer immunoprecipitated either total Smad2 and 3 or phospho-Smad2 and 3 at similar efficiencies from cell lysates (Figure 6C). Because most of the detectable R-Smad is in the nucleus of NMuMGTrx-SARA cells (Figure 5), the immunoprecipitates from NMuMGTrx-SARA cells examined primarily the association of nuclear R-Smad with Smad4. Although the Smad2 and Smad3 immunoprecipitates from NMuMGTrx-SARA cells contained less Smad4 protein, they did contain detectable Trx-SARA, indicating association of the aptamer with endogenous Smad2 and Smad3. The amounts of Trx-SARA in the immunoprecipitates from nonstimulated or TGF-β–stimulated NMuMGTrx-SARA cells were similar, even though the TGF-β–treated NMuMGTrx-SARA cells contain normal amounts of phosphorylated R-Smads (Figure 4). NMuMGSmad7cells also exhibited a significant decrease in the amount of Smad2 and Smad3 in Smad4 complexes, but this was expected because of the reduced Smad2 and Smad3 phosphorylation in NMuMGSmad7cells (Figure 4).
The Smad2 and Smad3 complexes with Smad4 bind to DNA to regulate gene expression. We tested whether the binding of Smad3–Smad4 complex to the Smad binding element was altered in the cell populations using an oligonucleotide pulldown assay (Figure 6D). TGF-β caused a significant increase in Smad3 and Smad4 from NMuMGTrx-GA cell lysates bound to the Smad-binding element DNA. The level of complex that bound to Smad-binding element DNA in cell lysates from NMuMGTrx-SARA or NMuMGSmad7cells was reduced. The level of Smad complex in cell lysates from TGF-β–treated NMuMGTrx-SARA cells that bound to DNA was similar to the basal level of DNA-bound complex from untreated NMuMGTrx-GA cells.
The ability of the Trx-SARA peptide aptamer to inhibit TGF-β activation of gene expression and induction of EMT demonstrates that peptide aptamers can be expressed at sufficient levels and with sufficient protein binding affinity to perturb R-Smad function in living cells. Generation of peptide aptamers that bind other sites on R-Smads should be a useful strategy for determining which protein-binding sites on R-Smads are required for specific responses to TGF-β. Previous studies using overexpression of Smad7 or dominant negative Smads have produced conflicting conclusions on the importance of the endogenous Smads for EMT in NMuMg cells (Bhowmick et al., 2001b ; Valcourt et al., 2005 ); the results reported here with the Trx-SARA peptide aptamer provide further support that Smad2 and Smad3 are required for TGF-β–induced EMT in NMuMg cells.
Smad2 and Smad3 nuclear import and export are mediated by phenylalanine-glycine-repeat–containing nucleoporins such as CAN/Nup214 and Nup153 and do not require Smad phosphorylation, importin-α, or importin-β (Xu et al., 2000 , 2002 , 2003 ). The nucleoporins bind the “hydrophobic corridor” in the Smad MH2 domain (Xu et al., 2002 , 2003 ). Nuclear import of Smad2 or Smad3 in an in vitro assay was inhibited by binding of the SARA SBD to hydrophobic corridor and nuclear export was inhibited by overexpression of the FoxH1 transcription factor, which also binds hydrophobic corridor (Hoodless et al., 1999 ; Xu et al., 2000 , 2002 , 2003 ). Missense mutations in the Smad hydrophobic corridor have no effect on the levels of TGF-β–induced Smad phosphorylation or Smad2 association with Smad4 but block nuclear import and thereby reduce the activation of TGF-β–responsive reporter gene expression (Xu et al., 2002 , 2003 ). In the absence of TGF-β, most of the cellular pool of Smad2 and Smad3 is located in the cytoplasm. This distribution may be the result of Smad association with cytoplasmic retention proteins such as SARA (Tsukazaki et al., 1998 ), microtubules (Dong et al., 2000 ), or filamin (Sasaki et al., 2001 ) or simply may be due to a more rapid rate of nuclear export than import (Schmierer and Hill, 2005 ). TGF-β–induced phosphorylation of Smad proteins results in their accumulation in the nucleus in approximately 30 min. When TGF-β receptor activity is terminated, the Smads return to a predominantly cytoplasmic localization in about 2 h (Inman et al., 2002b ). The Trx-SARA aptamer caused the subcellular distribution of Smad2 and Smad3, but not Smad4, to change from primarily cytoplasmic to primarily nuclear. We believe the pools of R-Smads accumulate in the nucleus because binding of the Trx-SARA aptamer to the two R-Smads impedes nuclear export. Trx-SARA, and Trx-GA contain NLS sequences and, by immunohistochemical staining, are preferentially enriched in the nucleus of the NMuMG cells. The higher concentrations of Trx-SARA in the nucleus should be more effective in inhibiting Smad export than lower amounts of cytoplasmic Trx-SARA are in impeding Smad import.
Although the Trx-SARA aptamer alters the subcellular location of the Smad2 and 3 pools, this does not alter the levels of phosphorylated Smads generated by receptor activation, except at the earliest time point 15 min after TGF-β addition, indicating there are sufficient cytoplasmic pools of Smad2 and Smad3 to interact with receptor. This is analogous to the finding that during TGF-β signaling most of the detectable Smad2 is located in the nucleus within 30 min after addition of TGF-β, but there is sufficient nucleocytoplasmic shuttling of Smad to sustain on-going rephosphorylation by activated receptor (Inman et al., 2002b ). If nuclear export of Smad was completely blocked by Trx-SARA, there should be reduced levels of cytoplasmic Smad available to the receptor and reduced levels of phosphorylated Smad when cells are treated with TGF-β. The temporary lower level of phospho-Smad protein observed 15 min after TGF-β addition may be caused by the initially lower levels of cytoplasmic R-Smads in the Trx-SARA–expressing cells and/or the nuclear export rate providing a rate-limiting step. Although we cannot rule out that lower level of phospho-Smad at 15 min may contribute to some of the biological effect in Trx-SARA cells, this seem unlikely in that sustained R-Smad activation of several hours is needed to generate biological responses (Inman et al., 2002b ; Nicolas and Hill, 2003 ). The observation that Smad phosphorylation levels are maintained in Trx-SARA cells over the same time course as in control cells also argues that there must be sufficient nuclear export of Smads for rephosphorylation by the active receptors. The activity state of the receptor is monitored by a continual process of Smad de-phosphorylation, export to the cytoplasm, rephosphorylation by the active receptors, nuclear import, and reassembly of active Smad heterotrimers with Smad4 (Inman et al., 2002b ). Therefore alteration in the basal subcellular location of the R-Smad pools does not appear to be the cause of inhibition of TGF-β–activated reporter gene expression and EMT by Trx-SARA.
The level of Smad2–Smad4 and Smad3–Smad4 complexes formed after TGF-β stimulation was significantly reduced in cells expressing Trx-SARA. This was also true in cells expressing Smad7, but in those cells the levels of phosphorylated Smad proteins were greatly reduced, and nonphosphorylated Smad proteins do not form a complex with Smad4. Reduction in the level of the Smad2–Smad4 complex or the Smad3–Smad4 complex would be expected to impede gene expression responses to TGF-β because the heterocomplexes are the form of Smad that participate in transcriptional regulatory complexes. Previous work has established that binding of SARA or Smad4 to Smad2 is mutually exclusive and the structural basis for this has been correlated with the two different conformations of R-Smad bound to SARA or bound to Smad4. The SARA binding site is available on the surface of the trimer and is not physically occluded by the trimer interactions (Qin et al., 2002 ). Both phosphorylated and nonphosphorylated Smad proteins bind to Trx-SARA in coimmunoprecipitation experiments, and binding of the SARA SBD to a trimeric phospho-Smad3 protein caused the trimeric protein to dissociate into monomers (Qin et al., 2002 ). The prior work on Smad-SARA interactions are consistent with a mechanism in which high levels of Trx-SARA in the nucleus bind to phospho-Smad2 or phospho-Smad3 in the heterotrimeric complex and drive the conformation of the Smad to a SARA-binding state, thereby dissociating the phospho-R-Smad from the Smad4 complex. This would lead to the observed reduced levels of Smad2–Smad4 and Smad3–Smad4 complexes in Trx-SARA cells and reduce Smad-dependent transcriptional responses to TGF-β. This distinct mechanism of Smad inhibition by Trx-SARA should avoid interference with Smad-independent functions of the TGF-β receptors and should have a more selective inhibitory effect on TGF-β signaling than overexpression of Smad7. We have not ruled out that Trx-SARA may also compete for the binding with another critical nuclear factor required for transcriptional activity by R-Smads. The ability of Trx-SARA to block the induction of EMT in mammary epithelial cells in culture demonstrates that this peptide aptamer can be used to explore the roles of TGF-β–induced Smad-dependent signaling in normal physiology and disease.
We are grateful to Drs. Joan Massagué, Peter ten Dijke, Shigeki Miyamoto, and Rik Derynck for expression constructs and luciferase reporter constructs and Dr. Ed Leof for antibody to phospho-Smad3. This work was supported by National Cancer Institute Grant R01 CA090875 to F.M.H. and Cancer Center Support Grant P30 CA014520. B.Z. was supported by the predoctoral training grant in Cancer Biology T32-CA09135.
This article was published online ahead of print in MBC in Press (http://www.molbiolcell.org/cgi/doi/10.1091/mbc.E05-10-0990) on June 14, 2006.