NF1 (nuclear factor 1) binds to two upstream elements of the human ANT2 (adenine nucleotide translocator-2) promoter and actively represses expression of the gene in growth-arrested diploid skin fibroblasts [Luciakova, Barath, Poliakova, Persson and Nelson (2003) J. Biol. Chem. 278, 30624–30633]. ChIP (chromatin immunoprecipitation) and co-immunoprecipitation analyses of nuclear extracts from growth-arrested and growth-activated diploid cells demonstrate that NF1, when acting as a repressor, is part of a multimeric complex that also includes Smad and Sp-family proteins. This complex appears to be anchored to both the upstream NF1-repressor elements and the proximal promoter, Sp1-dependent activation elements in growth-arrested cells. In growth-activated cells, the repressor complex dissociates and NF1 leaves the promoter. As revealed by co-immunoprecipitation experiments, NF1–Smad4–Sp3 complexes are present in nuclear extracts only from growth-inhibited cells, suggesting that the growth-state-dependent formation of these complexes is not an ANT2 promoter-specific event. Consistent with the role of Smad proteins in the repression complex, TGF-β (transforming growth factor-β) can fully repress ANT2 transcription in normally growing fibroblasts. Finally, pull-down experiments of in vitro transcribed/translated NF1 isoforms by GST (glutathione transferase)–Smad and GST–Smad MH fusion proteins indicate direct physical interactions between members of the two families. These findings suggest a possible functional relationship between the NF1 and Smad proteins that has not been previously observed.

INTRODUCTION

The NF1 (nuclear factor 1) transcription factors are encoded in four genes (Nf1-a, Nf1-b, Nf1-c and Nf1-x) [1] that are expressed as a large number of splice variants [24]. The expressed forms of NF1 bind to DNA as homo- or hetero-dimers [5,6], thus creating the potential for a large and varied transcriptional network in which NF1 family members activate or repress transcription depending on promoter and cellular context [1]. Many genes are dependent on NF1 for proper and optimal expression, and NF1 targets the promoters of genes with widely varying functions (reviewed in [1]). NF1 has also been implicated in several growth [7,8] and developmental [911] processes, but never in a central regulatory capacity.

We recently described a unique role for NF1 as an active repressor of gene expression in serum-starved cells [12] using the human ANT2 (adenine nucleotide translocator-2) gene as a model [13]. ANT proteins, which play an important role in maintaining cellular energy homoeostasis by catalysing the exchange of cytosolic ADP for mitochondrial ATP, are encoded in four genes [14]. The ANTs have also been implicated as both positive and negative regulators of apoptosis [15].

ANT2 differs from the other ANT isoforms in that its expression is regulated by the growth and differentiation states of cells. In general, ANT2 expression is repressed in growth-inhibited [12,16,17] or differentiating [1719] cells. Furthermore, re-activation of growth-inhibited cells activates ANT2 expression on the same time scale as the immediate-early genes [17], suggesting a functional role in the early events of cell growth. In spite of this, the importance of growth-modulated ANT2 expression remains unclear.

Growth-arrest repression of ANT2 is mediated through NF1 bound to two upstream repressor elements [Go-1 and Go-2 (Go NF1-binding repressor elements 1 and 2)] [12,16]. The Go-1 and Go-2 repressor elements are occupied by NF1 only in the growth-arrested state, and are required and sufficient for NF1-mediated transcriptional repression [12]. Since ANT2 expression is maintained by Sp1 bound to two juxtaposed Sp1 activation elements (AB boxes) in the proximal promoter [20], NF1 either directly or indirectly regulates the activity of Sp1.

In the present study, we demonstrate that, when acting as a growth-dependent repressor, NF1 is a component of a protein complex that includes members of the Smad- and Sp-family proteins. This repressor complex contacts both the upstream NF1-binding repressor elements and the proximal promoter Sp1-dependent activation element. In growth-activated cells, NF1 leaves the promoter and the multimeric repressor complex dissociates. The presence of Smad proteins in the NF1 repressor complex, together with our observations that NF1 and Smad proteins interact physically and that TGF-β (transforming growth factor-β) represses ANT2 transcription in normally growing cells, suggest a functional collaboration between NF1 and Smad that has not been observed previously.

EXPERIMENTAL

Cell culture

Human primary diploid foreskin fibroblasts in passages 7–20 were grown as described in [12]. To arrest cell growth, cells were washed twice with PBS and incubated for 48 h in serum-free medium. For growth activation, serum-starved cells were grown for an additional 24 h in the presence of 10% (v/v) serum.

Expression plasmids

The GST (glutathione transferase)–Sp1, GST–Smad2, GST–Smad3, GST–Smad4 and the GST–Smad4 MH1 plasmids were generously provided by Dr A. Moustakas (Ludwig Institute for Cancer Research, Uppsala, Sweden). The Smad2 MH1 domain was prepared by amplifying the domain from amino acids 1 to 251, and the MH2 domain was prepared by amplifying the domain from amino acids 251 to 467. Similarly, the Smad3 MH1 (amino acids 1–199) and MH2 (amino acids 199–425) domains were prepared; and the Smad4 MH2 domain (amino acids 321–552) was prepared. As a template, the appropriate full-length Smad in pGEX was used. Amplified MH domains were cloned into pGEX-6P-1 (Amersham Biosciences) and verified by sequencing. cDNAs coding for NF1-A1, -B2, -C2 and -X1 were provided by Dr M. Imagawa (Graduate School of Pharmaceutical Sciences, Nagoya City University Nagoya, Japan). For in vitro transcription/translation, inserts containing the whole coding region of individual NF1s were recloned into the expression vector pETM11 (EMBL, Heidelberg, Germany).

Treatment of cells with TGF-β

Cells in complete medium were treated with 1 ng/ml of TGF-β for 48 h [21].

Preparation of nuclear extract

Cells from starved and serum-induced human diploid fibroblasts were harvested, the nuclei were prepared [22] and nuclear proteins were extracted as described in [23]. The protein was measured using the Bio-Rad Protein Assay (Bio-Rad). Nuclear extracts were stored at −70 °C.

Immunoprecipitation

Antibodies were covalently linked to Protein A–Sepharose using a Seize X Protein Immunoprecipitation kit (Pierce). Binding, antigen immunoprecipitation and elution of the antigen were performed according to the manufacturer's instructions.

SDS/PAGE and Western-blot analysis

Eluted immunoprecipitated samples were precipitated for 20 min on ice in 10% (v/v) trichloroacetic acid, followed by a 15 min centrifugation at 10000 g and two washes with ice-cold acetone. Samples were air-dried, dissolved in sample buffer and separated by SDS/10% PAGE [24]. After electrophoresis, samples were electroblotted on to a Hybond ECL® membrane (Amersham Biosciences). Membranes were incubated with appropriate antibodies and developed with SuperSignal West Pico chemiluminescent substrate (Pierce).

ChIP (chromatin immunoprecipitation)

ChIP on growth-arrested and growth-induced human diploid cells was performed as in [25], except that 100 μl of Protein A–Sepharose was used instead of Staphylococcus aureus cells. Immunoprecipitation was performed with either 2 μl of antiserum against NF1 (No. 8199, kindly provided by Dr N. Tanese, New York University Medical Center, New York, NY, U.S.A.) or 2 μg of antiserum against Smad4 (sc-7154; Santa Cruz Biotechnology), Smad 1/2/3 (sc-7960; Santa Cruz Biotechnology), Sp1 (sc-59; Santa Cruz Biotechnology), Sp3 (sc-644; Santa Cruz Biotechnology), Smad 2 (sc-6200; Santa Cruz Biotechnology) and p300 (sc-585; Santa Cruz Biotechnology). Amplification of immunoprecipitated DNA fragments (2 μl) was performed by using the primers −931 (5′-AGTTCTTAACCTTCCTAAGCCTC-3′) and −666 (5′-CAAGTGAATGCCTCACTCTTCC-3′) covering the GoR (Go repressor region); and −151 (5′-TCCACGACTCCTCCTCCTGCGAGC-3′) and +39 (5′-CTGCAGGACGGGACTGGGCCGGGAAC-3′) covering the AB boxes. PCR was performed for 26–30 cycles, with 45 s of denaturation at 94 °C, followed by 45 s of annealing at 60 °C and 45 s of extension at 72 °C. The last step included extension for 10 min at 72 °C.

In vitro and bacterial expression of proteins

In vitro expression and radiolabelling of proteins were performed using the Promega Coupled Transcription/Translation System (TNT). Plasmids were transcribed using the T7 RNA polymerase according to the manufacturer's instructions. The quality of translated proteins was verified by SDS/PAGE. GST fusion proteins were expressed in Escherichia coli strain BL21 and purified as described in [26]. Solubilization and purity of the expressed proteins were monitored by SDS/PAGE.

Pull-down experiments

A 1 μl portion of the 35S-labelled NF1s, obtained by in vitro transcription/translation, was incubated with 1 μg of GST or GST–Smad2, GST–Smad3 or GST–Smad4 bound to glutathione–Sepharose beads (Amersham Biosciences). Binding of NF1 to GST fusion proteins was carried out in 50 mM Tris/HCl (pH 8), 50 mM KCl, 0.02% Tween 20, 0.02% BSA and 0.5 mM DTT (dithiothreitol), and washing of the bound proteins was performed in 50 mM Tris/HCl (pH 8), 50 mM KCl, 1% Triton X-100, 0.02% BSA and 0.5 mM DTT. Associated proteins were separated by SDS/PAGE and visualized on a Typhoon Amersham phosphoimager.

RNA purification and RT–PCR (reverse transcription–PCR)

RNA was isolated using the acidic guanidinium thiocyanate/phenol/chloroform extraction [27]. RNA (1 μg) was reverse-transcribed to generate the first-strand cDNA by using a RevertAid First Strand cDNA Synthesis kit (Fermentas) following the manufacturer's instructions. One-fifteenth of the cDNA was used as a template in PCR (25 μl) using primers for the ANT2 and ANT3 coding regions respectively. The 5′-primer was common to all three ANT isoforms and covered nucleotides +160/+181 in ANT2 and nucleotides +184/+195 in ANT3 (5′-GGGTCAAGCTGCTGCTGCAGG-3′) respectively. The 3′-primer of ANT2 covered nucleotides +509/+532 (5′-CGGAATTCCCTTTCAGCTCCAGC-3′), and the 3′-primer for ANT3 covered nucleotides +443/+464 (5′-CCGGCCGCACCGCCGGAGGC-3′). PCR reaction was performed for 28 cycles with 1 min of denaturation at 94 °C, 30 s annealing at 63 °C and extension for 1 min at 72 °C. PCR products were separated on 2% (w/v) agarose and stained with ethidium bromide.

RESULTS

The ANT2 promoter contains three repressor elements that modulate the activity of Sp1 on the AB box (Figure 1). However, repression of ANT2 transcription in growth-arrested cells (either by serum starvation or by growth to confluence) is mediated solely through NF1-binding elements (Go-1 and Go-2) in the GoR [12,16]. The ANT2 promoter also contains a number of putative Smad-binding elements (5′-GTCT-3′ or 5′-AGAC-3′) that are distributed around and between the Go-1 and Go-2 elements, between the two Silencer elements, and immediately upstream of the Sp1-activating elements (Figure 1). This distribution, together with data suggesting a link between NF1 and the TGF-β signalling system [28], raised the question whether Smad proteins could play a role in NF1-dependent repression of ANT2 in serum-starved cells.

The human ANT2 promoter

Figure 1
The human ANT2 promoter

Positions of known regulatory regions are placed relative to the transcription start (arrow). Empty boxes represent the Sp1 AB-activation elements. Three repressor regions, the GoR, Silencer and the C box are marked as grey boxes. Black ovals represent putative Smad-binding elements. The Go repressor and the Silencer bind NF1, whereas the C box binds Sp1.

Figure 1
The human ANT2 promoter

Positions of known regulatory regions are placed relative to the transcription start (arrow). Empty boxes represent the Sp1 AB-activation elements. Three repressor regions, the GoR, Silencer and the C box are marked as grey boxes. Black ovals represent putative Smad-binding elements. The Go repressor and the Silencer bind NF1, whereas the C box binds Sp1.

To address this issue, we first carried out a series of co-immunoprecipitation experiments to determine whether NF1 and Smad proteins interact, either directly or as components of a common multimeric complex. Experiments were performed on nuclear extracts from serum-starved (growth-arrested) and serum-refed (growth-activated) diploid fibroblasts (Figure 2). The existence of Smad/NF1 nuclear complexes could be demonstrated by co-immunoprecipitation with antibodies to both Smad4 (Figure 2, left panel) and NF1 (Figure 2, right panel). Unexpectedly, NF1/Smad4 complexes were detected in nuclear extracts predominantly, if not exclusively, in serum-starved cells (Figure 2, lane S compared with lane I). Thus formation of nuclear complexes containing NF1 and Smad proteins appears to be growth-state-dependent.

NF1 and Smad4 co-immunoprecipitate in growth-arrested cells

Figure 2
NF1 and Smad4 co-immunoprecipitate in growth-arrested cells

To arrest cell growth, cells were washed twice with PBS and incubated in serum-free medium for 48 h. To re-activate growth, serum-starved cells were grown for an additional 24 h in the presence of 10% serum. Nuclear extracts (500 μg) from serum-starved (S) and serum-induced (I) cells were isolated and immunoprecipitated (IP) with antibodies against Smad4 or NF1. Proteins eluted from Protein A–Sepharose were resolved by SDS/PAGE, transferred on to nitrocellulose membranes and probed with specific antibodies. WB, Western blots. Arrows indicate the molecular mass standards. A representative image of four independent experiments is shown.

Figure 2
NF1 and Smad4 co-immunoprecipitate in growth-arrested cells

To arrest cell growth, cells were washed twice with PBS and incubated in serum-free medium for 48 h. To re-activate growth, serum-starved cells were grown for an additional 24 h in the presence of 10% serum. Nuclear extracts (500 μg) from serum-starved (S) and serum-induced (I) cells were isolated and immunoprecipitated (IP) with antibodies against Smad4 or NF1. Proteins eluted from Protein A–Sepharose were resolved by SDS/PAGE, transferred on to nitrocellulose membranes and probed with specific antibodies. WB, Western blots. Arrows indicate the molecular mass standards. A representative image of four independent experiments is shown.

To determine whether Smad/NF1 nuclear complexes are associated with the ANT2 promoter, ChIP analysis was performed on chromatin isolated from serum-starved and serum-refed cells. The results show that Smad4 (Figure 3A, upper panel, compare lanes 4 and 5) and NF1 (Figure 3B, lanes 4) are both associated with the Go-2/Go-1 elements in serum-starved cells. By contrast, neither protein is associated with the ANT2 promoter in growth-activated cells, suggesting that release of NF1 is accompanied by release of Smad4. However, Smad4 is detected on the AB box in refed cells (Figure 3A, lower panel, lane 5). The growth-activated loss of NF1 and Smad4 from the Go repressor and the appearance of Smad4 associated with the AB activation boxes suggest a reorganization of the repressor complex in growth-activated cells.

The NF1–Smad4 repressor complex binds to the GoR only in growth-arrested human diploid cells

Figure 3
The NF1–Smad4 repressor complex binds to the GoR only in growth-arrested human diploid cells

ChIP was carried out on formaldehyde-cross-linked cells either serum-deprived for 48 h (lane 4) or serum-deprived and subsequently grown in the presence of serum (lane 5). Chromatin was isolated and immunoprecipitated with: (A) anti-Smad4 or (B) anti-NF1 antibodies. DNA was amplified with primers covering the Go-2/Go-1 repressor elements (nucleotides −931/−666) and the AB box Sp1-activation elements (nucleotides −151/+39). Primer amplification was also carried out on: a negative control containing no template (lane 1), a template isolated from chromatin that was not immunoprecipitated (lane 2), chromatin immunoprecipitated with an unrelated antibody, anti-F1-ATPase (lane 3) or total chromatin (lane 6). A 100-bp gene ruler (MBI Fermentas) marker is also shown (M). A representative image of four independent experiments is shown.

Figure 3
The NF1–Smad4 repressor complex binds to the GoR only in growth-arrested human diploid cells

ChIP was carried out on formaldehyde-cross-linked cells either serum-deprived for 48 h (lane 4) or serum-deprived and subsequently grown in the presence of serum (lane 5). Chromatin was isolated and immunoprecipitated with: (A) anti-Smad4 or (B) anti-NF1 antibodies. DNA was amplified with primers covering the Go-2/Go-1 repressor elements (nucleotides −931/−666) and the AB box Sp1-activation elements (nucleotides −151/+39). Primer amplification was also carried out on: a negative control containing no template (lane 1), a template isolated from chromatin that was not immunoprecipitated (lane 2), chromatin immunoprecipitated with an unrelated antibody, anti-F1-ATPase (lane 3) or total chromatin (lane 6). A 100-bp gene ruler (MBI Fermentas) marker is also shown (M). A representative image of four independent experiments is shown.

We also tested for the presence of R-Smads (Smad2/3) [29] on the GoR and AB box elements. ChIP analysis using antibodies to Smad2/3 or Smad2 show that Smad2, and perhaps also Smad3, are associated with both the GoR and the AB box regulatory elements in serum-starved cells (Figure 4A, lane 4), but are associated with neither of these elements in growth-activated cells (Figure 4A, lane 5). Thus Smad2/3 is associated with both the Go and AB elements, but only in the repressed state. However, co-immunoprecipitation of nuclear extracts (Figure 4B) shows that NF1–Smad2–Smad3 complexes are present in the nucleus under all conditions of growth, suggesting that they may not dissociate after leaving the Go repressor element. By contrast, nuclear complexes containing both Smad4 and Smad2/3 appear to decrease in serum-activated cells (Figure 4B, lane I), suggesting that the Smads might dissociate upon growth activation. Since co-precipitation experiments probe the total population of nuclear complexes, not just the subset of those bound to the ANT2 promoter, identification of Smad–NF1 complexes by this method provides reason to believe that such complexes participate in the regulation of additional genes.

Smad2 and Smad3 form a complex on the GoR

Figure 4
Smad2 and Smad3 form a complex on the GoR

(A) ChIP was carried out on formaldehyde-cross-linked serum-starved (lane 4) or serum-activated (lane 5) cells. The primers used and the definition of the remaining lanes are described in Figure 3. (B) Immunoprecipitation of Smad2/3 with antibodies to NF1 or Smad4. Nuclear extracts (500 μg) from serum-starved (S) and serum-induced (I) cells were isolated and immunoprecipitated (IP) with antibodies against the indicated proteins. Smad2/3 detected by Western blotting is indicated by an arrow on the right side. The arrow on the left side shows the molecular mass standards. A representative image of two independent experiments is shown.

Figure 4
Smad2 and Smad3 form a complex on the GoR

(A) ChIP was carried out on formaldehyde-cross-linked serum-starved (lane 4) or serum-activated (lane 5) cells. The primers used and the definition of the remaining lanes are described in Figure 3. (B) Immunoprecipitation of Smad2/3 with antibodies to NF1 or Smad4. Nuclear extracts (500 μg) from serum-starved (S) and serum-induced (I) cells were isolated and immunoprecipitated (IP) with antibodies against the indicated proteins. Smad2/3 detected by Western blotting is indicated by an arrow on the right side. The arrow on the left side shows the molecular mass standards. A representative image of two independent experiments is shown.

The association of NF1–Smad complexes with the AB box is most probably mediated through the Sp proteins. We therefore conducted ChIP analysis of Sp1 and Sp3 binding to the ANT2 promoter in serum-starved and serum-activated cells (Figure 5). Sp3 was included in the analysis since it binds the same element as Sp1, but is the only member of the Sp family that acts as a repressor [30]. Sp1 (Figure 5A, lane 4) and Sp3 (Figure 5B, lane 4) are associated with both the Go element and the AB box elements in serum-starved cells, suggesting that these elements might be linked through a larger repressor complex. Upon serum re-activation, Sp3 is no longer detected (Figure 5B, lane 5) and appears to leave the promoter, while Sp1 remains bound only to the AB boxes (Figure 5A, lower panel, lane 5). Thus it would appear as though Sp1 and Sp3 are anchored to the Go repressor and the AB activation boxes together with NF1 and Smad2/3 in growth-arrested cells. However, upon refeeding with serum, Sp3 leaves the promoter, while Sp1 remains anchored to the AB activation element (Figure 5, lane 5, lower panels).

Sp1 and Sp3 are associated with the Go repressor

Figure 5
Sp1 and Sp3 are associated with the Go repressor

(A) ChIP was carried out on serum-starved (lanes 4 and 6) and serum-induced (lanes 5 and 7) diploid fibroblasts using antibodies against Sp1 or p300. (B) ChIP was carried out on starved (lane 4) and induced (lane 5) cells using the antibodies against Sp3. DNA was amplified with primers covering the GoR (Go-2/Go-1) and the AB box (AB box). Lanes 1–3 are as described in Figure 3; lane 8 (A) and lane 6 (B) represent total chromatin. A representative image of three independent experiments is shown.

Figure 5
Sp1 and Sp3 are associated with the Go repressor

(A) ChIP was carried out on serum-starved (lanes 4 and 6) and serum-induced (lanes 5 and 7) diploid fibroblasts using antibodies against Sp1 or p300. (B) ChIP was carried out on starved (lane 4) and induced (lane 5) cells using the antibodies against Sp3. DNA was amplified with primers covering the GoR (Go-2/Go-1) and the AB box (AB box). Lanes 1–3 are as described in Figure 3; lane 8 (A) and lane 6 (B) represent total chromatin. A representative image of three independent experiments is shown.

To gain further insights into the repressor complex, we investigated the association between the Sp and Smad proteins under different growth states. Sp1 antibodies co-precipitate Smad4 from nuclear extracts prepared from both serum-starved and refed cells (Figure 6A), indicating the presence of Sp1–Smad4 nuclear complexes under all conditions of growth. However, a second Smad protein is also co-precipitated by Sp1 antibodies, but only from serum-starved cells (Figure 6A, asterisk). The growth-state dependency of this latter Smad–Sp1 complex suggests a possible role in Go repressor assembly/disassembly. However, the identity and functional significance of this Smad protein remain unknown.

Interactions of Sp proteins and p300 with the NF1–Smad complexes

Figure 6
Interactions of Sp proteins and p300 with the NF1–Smad complexes

(A, B) Nuclear extracts (500 μg) from serum-starved (S) and serum-induced (I) cells were immunoprecipitated (IP) with antibodies against Sp1 or p300. Proteins resolved by SDS/10% PAGE were probed with Smad4 or NF1 antibodies (Santa Cruz Biotechnology). Smad4 or NF1 detected by Western blotting are indicated by an arrow on the right side. Arrows on the left side show the molecular mass standards. (C) Nuclear extracts from serum-starved (S) and serum-induced (I) cells were immunoprecipitated with antibodies against NF1 and Smad4, and then probed with antibodies against Sp3. A representative image of two independent experiments is shown.

Figure 6
Interactions of Sp proteins and p300 with the NF1–Smad complexes

(A, B) Nuclear extracts (500 μg) from serum-starved (S) and serum-induced (I) cells were immunoprecipitated (IP) with antibodies against Sp1 or p300. Proteins resolved by SDS/10% PAGE were probed with Smad4 or NF1 antibodies (Santa Cruz Biotechnology). Smad4 or NF1 detected by Western blotting are indicated by an arrow on the right side. Arrows on the left side show the molecular mass standards. (C) Nuclear extracts from serum-starved (S) and serum-induced (I) cells were immunoprecipitated with antibodies against NF1 and Smad4, and then probed with antibodies against Sp3. A representative image of two independent experiments is shown.

In contrast with the presence of nuclear Sp1–Smad4 complexes in both serum-starved and serum-refed cells, Sp3–Smad4 complexes are detected by co-immunoprecipitation only in nuclear extracts of serum-starved cells (Figure 6C). Thus the dissociation of nuclear Sp3–Smad4 complexes appears to be related to the growth repression/activation transition in diploid fibroblasts. This conclusion is strengthened by ChIP analysis showing that Sp3 leaves the AB boxes in growth-activated cells (Figure 5B), resulting in the Sp1-dependent activation of ANT2 expression [20].

Finally, Sp1 antibodies co-precipitate NF1 from nuclear extracts prepared from growth-arrested cells, but not from growth-activated cells (Figure 6B). Thus, as with the Sp3–Smad4 complexes, formation of nuclear Sp1–NF1 complexes in human diploid fibroblasts is growth-state-dependent. This conclusion is also consistent with ChIP experiments showing that Sp1–NF1 complexes bind to the ANT2 promoter only in serum-starved, growth-inhibited cells (Figure 3).

p300 interacts directly with Smad proteins [31,32] and functions as a co-activator. We therefore tested whether p300 acts as a co-activator on the ANT2 promoter. Antibodies to p300 co-precipitate Smad4 from nuclear extracts of serum-activated cells, but not from serum-starved cells (Figure 6A). Thus the formation of p300–Smad4 complexes is also growth-state-dependent in diploid fibroblasts, favouring the growth-activated state. However, even though nuclear p300–Smad4 complexes are formed in growth-activated fibroblasts (Figure 6A), a significant increase in p300 binding to the AB activation boxes could not be detected by ChIP analysis (Figure 5A, lower panel). Thus p300 may not act as a co-activator on the ANT2 promoter. Antibodies to p300 co-precipitated little or no NF1 from nuclear extracts of either growth-arrested or growth-activated cells (Figure 6B), suggesting that the population of nuclear complexes containing both p300 and NF1 is relatively small, if existent, in human diploid fibroblasts. This result supports our conclusions from ChIP analysis that p300 is not associated with the Go repressor complex (Figure 5A).

The question remains whether nuclear NF1–Smad complexes are formed through direct or indirect physical contacts. Smad [31,33] and NF1 [34] interact directly with Sp1, providing a possible assembly pathway for NF1/Smad-containing complexes. To test for physical interactions, GST fusion proteins (full-length Smads or their MH1 or MH2 domains) were used to pull down in vitro translated, radiolabelled full-length NF1s. Smad3 and Smad4 interact relatively strongly with NF1-A and -C (Figure 7, left panel) through their MH1 domain (Figure 7, right panel). Smad3 and Smad4 appear to interact more weakly with full-length NF1-B and NF1-X (left panel). However, pull down with the MH domains (right panel) suggests that Smad3 and Smad4 might interact with NF1-X through the MH1 domain. Interactions of Smad2 are less clear. Although it appears to interact as strongly as Smad3 and 4 with full-length NF1-A (Figure 7, left panel), this was not confirmed using the MH-domain fusion proteins (Figure 7, right panel). Taken together, the results shown in Figure 7 show for the first time a direct physical interaction between the transcription factor NF1 and Smad2, Smad3 and Smad4.

Smad proteins physically interact with the transcription factor NF1

Figure 7
Smad proteins physically interact with the transcription factor NF1

GST–Smad2, GST–Smad3 and GST–Smad4 and their individual MH-domain fusion proteins were expressed as described in the Experimental section. GST–Smad protein (1 μg) was bound to glutathione beads and incubated with radiolabelled NF1 isoforms produced by in vitro transcription/translation in rabbit reticulocyte lysates. 5% input, input of the 35S-labelled in vitro-synthesized NF1 proteins. GST, empty vector carrying GST alone. Left panel: NF1 isoform synthesized in vitro. A representative picture of four independent experiments is shown.

Figure 7
Smad proteins physically interact with the transcription factor NF1

GST–Smad2, GST–Smad3 and GST–Smad4 and their individual MH-domain fusion proteins were expressed as described in the Experimental section. GST–Smad protein (1 μg) was bound to glutathione beads and incubated with radiolabelled NF1 isoforms produced by in vitro transcription/translation in rabbit reticulocyte lysates. 5% input, input of the 35S-labelled in vitro-synthesized NF1 proteins. GST, empty vector carrying GST alone. Left panel: NF1 isoform synthesized in vitro. A representative picture of four independent experiments is shown.

The inclusion of Smad proteins in the Go-repression complex suggests that TGF-β might be a signalling ligand in ANT2 repression. To test this, we measured ANT2 transcripts in normally growing diploid fibroblasts treated with TGF-β. The results (Figure 8) show that TGF-β prevents ANT2 expression in growing cells and that the magnitude of the TGF-β inhibitory effect is similar to that obtained by serum removal (Figure 8; [12]). As expected [35], serum starvation does not effect the expression of ANT3 (Figure 8).

TGF-β inhibits expression of ANT2

Figure 8
TGF-β inhibits expression of ANT2

Total RNA (1 μg) from human diploid foreskin fibroblasts was reverse-transcribed and amplified with ANT2- and ANT3-specific primers. RNA was isolated from cells grown in the absence of serum for 48 h (S), from serum-starved cells subsequently induced for 24 h with 10% serum (I) and from cells grown for 48 h in complete medium and in the presence of TGF-β (1 ng/ml). A representative image of three independent experiments is shown.

Figure 8
TGF-β inhibits expression of ANT2

Total RNA (1 μg) from human diploid foreskin fibroblasts was reverse-transcribed and amplified with ANT2- and ANT3-specific primers. RNA was isolated from cells grown in the absence of serum for 48 h (S), from serum-starved cells subsequently induced for 24 h with 10% serum (I) and from cells grown for 48 h in complete medium and in the presence of TGF-β (1 ng/ml). A representative image of three independent experiments is shown.

DISCUSSION

Expression of the human ANT2 gene is repressed in serum-starved cells and in cells grown to confluence [12,16]. We earlier mapped and identified two upstream promoter elements (Go-1 and Go-2) required for repression [12]. These elements bind NF1, which acts as an active repressor. To gain insights into the mechanism by which NF1 represses Sp1-dependent activation, we studied the proteins associated with the Go and AB box elements in growth-repressed and growth-activated human skin fibroblasts. We show that NF1 bound to the Go elements in growth-repressed cells is part of a larger repressor complex that includes both Smad and Sp-family proteins. Figure 9 summarizes the proteins found by ChIP analysis to be associated with the Go and AB box elements in growth-activated and growth-repressed cells. Complexes containing NF1, Smads2/3, Smad4, Sp1 and Sp3 are associated with both the Go and AB Box elements in serum-starved cells. These results could be explained by the binding of separate and distinct complexes to each set of elements, as depicted in Figure 9. However, it is equally possible that NF1 and Sp proteins anchored to the Go and AB box elements respectively are linked in a repressor complex that includes Smad proteins.

The role of NF1, Smad and Sp proteins in growth repression/activation of the ANT2 gene

Figure 9
The role of NF1, Smad and Sp proteins in growth repression/activation of the ANT2 gene

A complex containing NF1, Smad and Sp proteins binds to the GoR and Sp1 promoter elements in serum-starved cells. Upon activation of cell growth, NF1–Smad2/3 and Sp3 are released from the Go element, leaving a complex of Sp1 and Smad4 bound to the proximal Sp1 transcription-activation elements (AB box). The repressor elements of the GoR and the C box (see the main text) are depicted by grey boxes. The filled ovals are putative Smad-binding elements. An arrow indicates the transcription initiation site and a cross indicates transcription inhibition.

Figure 9
The role of NF1, Smad and Sp proteins in growth repression/activation of the ANT2 gene

A complex containing NF1, Smad and Sp proteins binds to the GoR and Sp1 promoter elements in serum-starved cells. Upon activation of cell growth, NF1–Smad2/3 and Sp3 are released from the Go element, leaving a complex of Sp1 and Smad4 bound to the proximal Sp1 transcription-activation elements (AB box). The repressor elements of the GoR and the C box (see the main text) are depicted by grey boxes. The filled ovals are putative Smad-binding elements. An arrow indicates the transcription initiation site and a cross indicates transcription inhibition.

Whatever the nature of the complexes, growth activation results in an extensive reorganization on the promoter in which NF1 and the associated Smad and Sp proteins leave the Go element. These findings are in accord with our earlier failure to detect protein binding to the Go elements in growth-activated cells by using in vivo footprinting [12]. Release of the repressor complex from the Go elements in growth-activated cells is also accompanied by reorganization around the AB elements (Figure 9). Sp1 remains on the AB boxes, but is no longer associated with NF1, Smad2/3 or Sp3. Sp1 is, however, associated with Smad4. Reorganization of Sp1 bound to the AB elements in growth-activated cells is also suggested by a shift in the specific nucleotides contacted by Sp1 [36].

NF1 binding to the Go element in the growth-repressed state, and its release in the growth-activated state [12], suggest a central role for this protein in ANT2 repression. However, the mechanism of NF1 recruitment to the promoter and the order in which the components of the repressor complex are assembled are difficult to assess since Sp1 interacts directly with the Smads [31,33] and NF1 [34] and, as demonstrated in the present study, NF1 also interacts with Smads. The dynamic redistribution of Smads is only one of many possible mechanisms by which assembly of a repressor complex might be regulated, especially as the cytoplasmic/nuclear distribution of Smad proteins is a dynamic process, and some of these proteins are present in the nucleus even in non-stimulated cells [37,38]. Furthermore, all three proteins can be modified by various combinations of phosphorylation, acetylation or SUMOylation and are thus potential targets for signalling pathways that initiate repressor assembly/disassembly. Despite these complications, our observation that TGF-β induces repression of the ANT2 promoter in proliferating human diploid fibroblasts implies involvement of a Smad-dependent pathway. Furthermore, since TGF-β is released from serum-starved human diploid fibroblasts [39] and is responsible for initiating a senescence response [39], it is a prime candidate as the ligand that initiates ANT2 repression in the absence of serum.

A mechanism of NF1-dependent repression similar to that described here and in our previous studies [12] on the ANT2 promoter was recently reported for the p21Waf1/Cip1 promoter in human skin fibroblasts [40]. With both promoters, NF1 binding represses transcription, whereas release of NF1 is associated with transcriptional activation. Furthermore, Sp3 is associated with NF1 on both promoters in the repressed state, but not in the activated state (discussed below). However, regulation of the two genes by NF1 differs in that repression of ANT2 takes place in growth-arrested human skin fibroblasts, whereas p21 repression occurs in proliferating cells. Thus NF1, acting as a direct repressor, is able to regulate genes with vastly different growth-related functions in skin fibroblasts.

NF1 has been implicated in a variety of growth-related processes, even though a central role for NF1 in growth regulation has not been established. NF1 represses gadd153 [7] and activates p53 [8] expression, both of which have key roles in growth arrest of damaged or stressed cells. Furthermore, NF1-X, together with c-myc and mdm2 (murine double minute 2), were the only genes identified in a screen performed to identify complementary DNAs that prevent TGF-β-induced growth arrest of mink lung epithelial cells [28]. Finally, expression of NF1 in chick embryo fibroblasts has also been reported to prevent transformation induced by nuclear, but not cytosolic, oncogene products [41]. These studies, together with the present results demonstrating the existence of NF1–Smad nuclear complexes (Figures 2 and 4B), suggest that NF1 may have a more general role in regulating growth-related processes than previously appreciated.

NF1-mediated repression of ANT2 described here is reminiscent of the TGF-dependent repression of the myc and Id1 genes, which are mediated through Smads co-operating with a transcription factor acting as a co-repressor. myc repression is mediated through the p107 repressor protein, which is recruited to a Smad–E2F–DP1 complex [42], and Id1 repression is mediated via ATF3 (activating transcription factor 3) in co-operation with Smad3 [43]. In the case of ANT2, assembly of the NF1–Smad complexes on the promoter is also associated with recruitment of a known repressor protein, Sp3. Sp3 is associated with an NF1-repressor complex on both the Go and the AB box elements in growth-arrested cells, and is released from the promoter in growth-activated cells. Similarly, Sp3, but not Sp1, is bound to the NF1-repressed p21Waf1/Cip1 promoter in proliferating skin fibroblasts [40]. These results suggest that Sp3 could be a specific component of the NF1-dependent repression machinery. Sp3, unlike Sp1, is a strong transcriptional repressor [30]. Silencing is dependent on the lysine residue in an IKEE repressor motif, which, when mutated, converts Sp3 into a transcription activator [44]. In vivo acetylation of Sp3 activates expression of the TGF-β type II receptor [45] and the SOCS (suppressor of cytokine signalling) [46] genes, suggesting a possible signalling mechanism for the Sp3 repressor/activator switch. However, the IKEE motif is also SUMOylated, which, as with acetylation, eliminates Sp3 silencing and promotes Sp3 activation [4749]. Thus the specific role of Sp3 in ANT2 repression remains to be sorted out.

To our knowledge, no structural or functional associations between members of Smad and NF1 families have been reported previously, even though, as discussed above, several reports suggest interfacing between the NF1 proteins and TGF-mediated processes. However, the present study demonstrates clearly: (i) that Smad and NF1 proteins interact physically, as shown by in vitro pull-down assays and (ii) that nuclear complexes of Smad/NF1 are formed and disassembled in human skin fibroblasts in a growth-state-dependent manner. These findings, coupled with our observation from ChIP analysis that Smad proteins are components of an NF1 complex that acts to repress ANT2 expression, raise the possibility that NF1/Smad may have a broader function in the regulation of gene expression than heretofore appreciated.

This study was supported by the Slovak Science and Technology Assistance Agency (APVT) No. 26-002102 (to K. L.), the Slovak Grant Agency VEGA No. 2/6060/26 (to P. B.) and the Swedish Research Council (to B. D. N.). We thank Dr M. Imagawa, Dr A. Moustakas and Dr N. Tanese for gifts of NF1 isoform cDNA vectors, GST–Smad expression vectors and NF1-C antibodies respectively. We declare no conflict of interest.

Abbreviations

     
  • ANT

    adenine nucleotide translocator

  •  
  • ChIP

    chromatin immunoprecipitation

  •  
  • DTT

    dithiothreitol

  •  
  • NF1

    nuclear factor 1

  •  
  • Go-1 and Go-2

    Go NF1-binding repressor elements 1 and 2

  •  
  • GoR

    Go repressor region

  •  
  • GST

    glutathione transferase

  •  
  • TGF-β

    transforming growth factor-β

References

References
1
Gronostajski
R. M.
Roles of the NFI/CTF gene family in transcription and development
Gene
2000
, vol. 
249
 (pg. 
31
-
45
)
2
Santoro
C.
Mermod
N.
Andrews
P.
Tjian
R.
A family of human CCAAT/box/binding proteins active in transcription and DNA replication: cloning and expression of multiple cDNAs
Nature
1988
, vol. 
334
 (pg. 
218
-
224
)
3
Rupp
R.
Kruse
U.
Multhaup
G.
Gobel
U.
Beyreuther
K.
Sippel
A.
Chicken NF1/TGGCA proteins are encoded by at least three independent genes: NFI-A, NFI-B, and NFI-C with homologues in mammalian genomes
Nucleic Acids Res.
1990
, vol. 
18
 (pg. 
2607
-
2616
)
4
Kruse
U.
Sippel
A. E.
The genes for transcription factor nuclear factor I give rise to corresponding splice variants between vertebrate species
J. Mol. Biol.
1994
, vol. 
238
 (pg. 
860
-
865
)
5
Kruse
U.
Sippel
A. E.
Transcription factor nuclear factor I proteins form stable homo- and heterodimers
FEBS Lett.
1994
, vol. 
348
 (pg. 
46
-
50
)
6
Gounari
F.
De Francesco
R.
Schmitt
J.
van der Vliet
P.
Cortese
R.
Stunnenberg
H.
Amino-terminal domain of NF1 binds to DNA as a dimer and activates adenovirus DNA replication
EMBO J.
1990
, vol. 
9
 (pg. 
559
-
566
)
7
Nakamura
M.
Okura
T.
Kitami
Y.
Hiwada
K.
Nuclear factor 1 is a negative regulator of gadd153 gene expression in vascular smooth muscle cells
Hypertension
2001
, vol. 
37
 (pg. 
419
-
424
)
8
Furlong
E. E.
Rein
T.
Martin
F.
YY1 and NF1 both activate the human p53 promoter by alternatively binding to a composite element, and YY1 and E1A cooperate to amplify p53 promoter activity
Mol. Cell. Biol.
1996
, vol. 
16
 (pg. 
5933
-
5945
)
9
das Neves
L.
Duchala
C. S.
Tolentino-Silva
F.
Haxhiu
M. A.
Colmenares
C.
Macklin
W. B.
Campbell
C. E.
Butz
K. G.
Gronostajski
R. M.
Godinho
F.
Disruption of the murine nuclear factor I-A gene (Nfia) results in perinatal lethality, hydrocephalus, and agenesis of the corpus callosum
Proc. Natl. Acad. Sci. U.S.A.
1999
, vol. 
96
 (pg. 
11946
-
11951
)
10
Steele-Perkins
G.
Plachez
C.
Butz
K. G.
Yang
G.
Bachurski
C. J.
Kinsman
S. L.
Litwack
E. D.
Richards
L. J.
Gronostajski
R. M.
The transcription factor gene Nfib is essential for both lung maturation and brain development
Mol. Cell. Biol.
2005
, vol. 
25
 (pg. 
685
-
698
)
11
Steele-Perkins
G.
Butz
K. G.
Lyons
G. E.
Zeichner-David
M.
Kim
H. J.
Cho
M. I.
Gronostajski
R. M.
Essential role for NFI-C/CTF transcription-replication factor in tooth root development
Mol. Cell. Biol.
2003
, vol. 
23
 (pg. 
1075
-
1084
)
12
Luciakova
K.
Barath
P.
Poliakova
D.
Persson
A.
Nelson
B. D.
Repression of the human adenine nucleotide translocase-2 gene in growth-arrested human diploid cells: the role of nuclear factor-1
J. Biol. Chem.
2003
, vol. 
278
 (pg. 
30624
-
30633
)
13
Luciakova
K.
Li
R.
Nelson
B. D.
Differential regulation of the transcript levels of some nuclear- and mitochondrial-encoded respiratory chain components in response to growth activation
Eur. J. Biochem.
1992
, vol. 
207
 (pg. 
253
-
257
)
14
Dolce
V.
Scarcia
P.
Iacopetta
D.
Palmieri
F.
A fourth ADP/ATP carrier isoform in man: identification, bacterial expression, functional characterization and tissue distribution
FEBS Lett.
2005
, vol. 
579
 (pg. 
633
-
637
)
15
Le Bras
M.
Borgne-Sanchez
A.
Touat
Z.
El Dein
O. S.
Deniaud
A.
Maillier
E.
Lecellier
G.
Rebouillat
D.
Lemaire
C.
Kroemer
G.
, et al. 
Chemosensitization by knockdown of adenine nucleotide translocase-2
Cancer Res.
2006
, vol. 
66
 (pg. 
9143
-
9152
)
16
Barath
P.
Luciakova
K.
Hodny
Z.
Li
R.
Nelson
B. D.
The growth-dependent expression of the adenine nucleotide translocase-2 (ANT2) gene is regulated at the level of transcription and is a marker of cell proliferation
Exp. Cell Res.
1999
, vol. 
248
 (pg. 
583
-
588
)
17
Battini
R.
Ferrari
S.
Kaczmarek
L.
Calabretta
B.
Chen
S.-T.
Baserga
R.
Molecular cloning of a cDNA for a human ADP/ATP carrier which is growth-regulated
J. Biol. Chem.
1987
, vol. 
262
 (pg. 
4355
-
4359
)
18
Lunardi
J.
Hurko
O.
Engel
W. K.
Attardi
G.
The multiple ADP/ATP translocase genes are differentially expressed during human muscle development
J. Biol. Chem.
1992
, vol. 
267
 (pg. 
15267
-
15270
)
19
Stepien
G.
Torroni
A.
Chung
A. B.
Hodge
J. A.
Wallace
D. C.
Differential expression of adenine nucleotide translocator isoforms in mammalian tissues and during muscle cell differentiation
J. Biol. Chem.
1992
, vol. 
267
 (pg. 
14592
-
14597
)
20
Li
R.
Hodny
Z.
Luciakova
K.
Barath
P.
Nelson
B. D.
Sp1 activates and inhibits transcription from separate elements in the proximal promoter of the human adenine nucleotide translocase 2 (ANT2) gene
J. Biol. Chem.
1996
, vol. 
271
 (pg. 
18925
-
18930
)
21
Poncelet
A. C.
Schnaper
H. W.
Sp1 and Smad proteins cooperate to mediate transforming growth factor-β1-induced α2(I) collagen expression in human glomerular mesangial cells
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
6983
-
6992
)
22
Dignam
J. D.
Lebovitz
R. M.
Roeder
R. G.
Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei
Nucleic Acids Res.
1983
, vol. 
11
 (pg. 
1475
-
1489
)
23
Barath
P.
Poliakova
D.
Luciakova
K.
Nelson
B. D.
Identification of NF1 as a silencer protein of the human adenine nucleotide translocase-2 gene
Eur. J. Biochem.
2004
, vol. 
271
 (pg. 
1781
-
1788
)
24
Laemmli
U. K.
Cleavage of structural proteins during the assembly of the head of bacteriophage T4
Nature
1970
, vol. 
227
 (pg. 
680
-
685
)
25
Weinmann
A. S.
Farnham
P. J.
Identification of unknown target genes of human transcription factors using chromatin immunoprecipitation
Methods
2002
, vol. 
26
 (pg. 
37
-
47
)
26
Kardassis
D.
Papakosta
P.
Pardali
K.
Moustakas
A.
c-Jun transactivates the promoter of the human p21WAF1/Cip1 gene by acting as a superactivator of the ubiquitous transcription factor Sp1
J. Biol. Chem.
1999
, vol. 
274
 (pg. 
29572
-
29581
)
27
Chomczynski
P.
Sacchi
N.
Single-step method of RNA isolation by acid guanidinium thiocyanate–phenol–chloroform extraction
Anal. Biochem.
1987
, vol. 
162
 (pg. 
156
-
159
)
28
Sun
P.
Dong
P.
Dai
K.
Hannon
G. J.
Beach
D.
p53-independent role of MDM2 in TGF-β1 resistance
Science
1998
, vol. 
282
 (pg. 
2270
-
2272
)
29
Massague
J.
Gomis
R. R.
The logic of TGFβ signaling
FEBS Lett.
2006
, vol. 
580
 (pg. 
2811
-
2820
)
30
Hagen
G.
Müller
S.
Beato
M.
Suske
G.
Sp1-mediated transcriptional activation is repressed by Sp3
EMBO J.
1994
, vol. 
13
 (pg. 
3843
-
3851
)
31
Pardali
K.
Kurisaki
A.
Moren
A.
ten Dijke
P.
Kardassis
D.
Moustakas
A.
Role of Smad proteins and transcription factor Sp1 in p21Waf1/Cip1 regulation by transforming growth factor-β
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
29244
-
29256
)
32
Park
S. R.
Lee
E. K.
Kim
B. C.
Kim
P. H.
p300 cooperates with Smad3/4 and Runx3 in TGFβ1-induced IgA isotype expression
Eur. J. Immunol.
2003
, vol. 
33
 (pg. 
3386
-
3392
)
33
Feng
X. H.
Lin
X.
Derynck
R.
Smad2, Smad3 and Smad4 cooperate with Sp1 to induce p15Ink4B transcription in response to TGF-β
EMBO J.
2000
, vol. 
19
 (pg. 
5178
-
5193
)
34
Rafty
L. A.
Santiago
F. S.
Khachigian
L. M.
NF1/X represses PDGF A-chain transcription by interacting with Sp1 and antagonizing Sp1 occupancy of the promoter
EMBO J.
2002
, vol. 
21
 (pg. 
334
-
343
)
35
Dörner
A.
Pauschinger
M.
Badorff
A.
Noutsias
M.
Giessen
S.
Schulze
K.
Bilger
J.
Rauch
U.
Schultheiss
H.-P.
Tissue-specific transcription pattern of the adenine nucleotide translocase isoforms in humans
FEBS Lett.
1997
, vol. 
414
 (pg. 
258
-
262
)
36
Luciakova
K.
Hodny
Z.
Barath
P.
Nelson
B. D.
In vivo mapping of the human adenine nucleotide translocator-2 (ANT2) promoter provides support for regulation by a pair of proximal Sp1-activating sites and an upstream silencer element
Biochem. J.
2000
, vol. 
352
 (pg. 
519
-
523
)
37
Schmierer
B.
Hill
C. S.
Kinetic analysis of Smad nucleocytoplasmic shuttling reveals a mechanism for transforming growth factor β-dependent nuclear accumulation of Smads
Mol. Cell. Biol.
2005
, vol. 
25
 (pg. 
9845
-
9858
)
38
Nicolas
F. J.
De Bosscher
K.
Schmierer
B.
Hill
C. S.
Analysis of Smad nucleocytoplasmic shuttling in living cells
J. Cell Sci.
2004
, vol. 
117
 (pg. 
4113
-
4125
)
39
Zeng
G.
McCue
H. M.
Mastrangelo
L.
Millis
A. J.
Endogenous TGF-β activity is modified during cellular aging: effects on metalloproteinase and TIMP-1 expression
Exp. Cell Res.
1996
, vol. 
228
 (pg. 
271
-
276
)
40
Ouellet
S.
Vigneault
F.
Lessard
M.
Leclerc
S.
Drouin
R.
Guerin
S. L.
Transcriptional regulation of the cyclin-dependent kinase inhibitor 1A (p21) gene by NFI in proliferating human cells
Nucleic Acids Res.
2006
, vol. 
34
 (pg. 
6472
-
6487
)
41
Schuur
E. R.
Kruse
U.
Iacovoni
J. S.
Vogt
P. K.
Nuclear factor I interferes with transformation induced by nuclear oncogenes
Cell Growth Differ.
1995
, vol. 
6
 (pg. 
219
-
227
)
42
Chen
C. R.
Kang
Y.
Siegel
P. M.
Massague
J.
E2F4/5 and p107 as Smad cofactors linking the TGFβ receptor to c-myc repression
Cell
2002
, vol. 
110
 (pg. 
19
-
32
)
43
Kang
Y.
Chen
C. R.
Massague
J.
A self-enabling TGFβ response coupled to stress signaling: Smad engages stress response factor ATF3 for Id1 repression in epithelial cells
Mol. Cell
2003
, vol. 
11
 (pg. 
915
-
926
)
44
Braun
H.
Koop
R.
Ertmer
A.
Nacht
S.
Suske
G.
Transcription factor Sp3 is regulated by acetylation
Nucleic Acids Res.
2001
, vol. 
29
 (pg. 
4994
-
5000
)
45
Ammanamanchi
S.
Brattain
M. G.
Sp3 is a transcriptional repressor of transforming growth factor-β receptors
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
3348
-
3352
)
46
Ehlting
C.
Haussinger
D.
Bode
J. G.
Sp3 is involved in the regulation of SOCS3 gene expression
Biochem. J.
2005
, vol. 
387
 (pg. 
737
-
745
)
47
Ross
S.
Best
J. L.
Zon
L. I.
Gill
G.
SUMO-1 modification represses Sp3 transcriptional activation and modulates its subnuclear localization
Mol. Cell
2002
, vol. 
10
 (pg. 
831
-
842
)
48
Sapetschnig
A.
Rischitor
G.
Braun
H.
Doll
A.
Schergaut
M.
Melchior
F.
Suske
G.
Transcription factor Sp3 is silenced through SUMO modification by PIAS1
EMBO J.
2002
, vol. 
21
 (pg. 
5206
-
5215
)
49
Sapetschnig
A.
Koch
F.
Rischitor
G.
Mennenga
T.
Suske
G.
Complexity of translationally controlled transcription factor Sp3 isoform expression
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
42095
-
42105
)