Oncogene
SEARCH     advanced search my account e-alerts subscribe register
Journal home
Advance online publication
Current issue
Archive
Press releases
For authors
For referees
Contact editorial office
About the journal
For librarians
Subscribe
Advertising
naturereprints
Contact NPG
Customer services
Site features
NPG Subject areas
Access material from all our publications in your subject area:
Biotechnology Biotechnology
Cancer Cancer
Chemistry Chemistry
Dentistry Dentistry
Development Development
Drug Discovery Drug Discovery
Earth Sciences Earth Sciences
Evolution & Ecology Evolution & Ecology
Genetics Genetics
Immunology Immunology
Materials Materials Science
Medical Research Medical Research
Microbiology Microbiology
Molecular Cell Biology Molecular Cell Biology
Neuroscience Neuroscience
Pharmacology Pharmacology
Physics Physics
Browse all publications
 
6 September 2001, Volume 20, Number 39, Pages 5409-5419
Table of contents    Previous  Article  Next   [PDF]
Original Paper
FLRG, an activin-binding protein, is a new target of TGFbold beta transcription activation through Smad proteins
Laurent Bartholin1, Véronique Maguer-Satta1, Sandrine Hayette1,2, Sylvie Martel1, Mylène Gadoux2, Suzanne Bertrand2, Laura Corbo1, Christine Lamadon1, Anne-Marie Morera3, Jean-Pierre Magaud1,2 and Ruth Rimokh1

1Unité INSERM U453, Centre Léon Bérard, 69373 Lyon, France

2Laboratoire de Cytogénétique Moléculaire, Hôpital Edouard Herriot, 69437 Lyon, France

3Unité INSERM U329, Hôpital Debrousse, 69322 Lyon, France

Correspondence to: R Rimokh, Unité INSERM U453, Centre Léon Bérard, 69373 Lyon, France. E-mail: rimokh@lyon.fnclcc.fr

Abstract

The FLRG gene encodes a secreted glycoprotein that binds to activin and is highly homologous to follistatin, an activin ligand. We cloned the promoter region of the human FLRG gene, and defined the minimal region necessary for transcription activation in a reporter-system assay. We showed that the fragment between positions -130 and +6, which consists of multiple consensus Sp1-binding sites, is required for the constitutive expression of the FLRG gene. We demonstrate here that FLRG mRNA expression is rapidly induced by TGFbeta or by transfection with Smad protein expression vectors in human HepG2 cells. We investigated the transcription-regulation mechanism of FLRG expression in HepG2 cells following treatment with TGFbeta. By deletion and point-mutation analysis of the FLRG promoter, we identified a Smad-binding element involved in the TGFbeta-inducible expression of the FLRG gene. Moreover, transactivation of the FLRG promoter by TGFbeta was compromised by dominant-negative mutants of Smad3 and Smad4 proteins. In addition, gel electrophoresis mobility-shift assays demonstrated the specific interaction of Smad3 and Smad4 proteins with the Smad-binding element consensus motif found in the FLRG promoter. Taken together, our data imply that Smad proteins participate in the regulation of expression of FLRG, a new target of TGFbeta transcription activation. Oncogene (2001) 20, 5409-5419.

Keywords

FLRG; TGFbeta; Smad; activin; follistatin

Introduction

We have already reported the molecular characterization of a t(11;19)(q13;p13) translocation in B-cell chronic lymphocytic leukemia, which led to the identification of a new evolutionarily-conserved gene, the Follistatin-related gene (FLRG), located on chromosome 19 at band 19p13 (Hayette et al., 1998). In normal human tissues the highest steady-state levels of FLRG transcripts have been observed in the placenta, amniocytes, chorionic villosities and bone marrow stromal cells. FLRG encodes a secreted glycoprotein which is highly homologous to follistatin, most of whose effects are mediated by its binding with activin, a member of the transforming growth factor-beta (TGFbeta) superfamily of growth/differentiation factors. The biological activity of activin, in a broad range of cells, is antagonized by follistatin (Phillips and de Kretser, 1998; Ying et al., 1997). These cytokines have multifunctional properties, and are important local regulators of growth and differentiation in a variety of tissues at various stages in the life cycle (DePaolo, 1997). Activin, produced by bone marrow stromal cells (Yamashita et al., 1992) such as FLRG, is also involved in the regulation and development of hematopoietic progenitor differentiation (Johansson and Wiles, 1995; Yu and Dolter, 1997), particularly in the erythroid lineage (Kitamura et al., 2000; Shao et al., 1992; Shiozaki et al., 1992). It was the strong homology between follistatin and FLRG that led to our demonstration that FLRG, like follistatin, interacts physically with activin (Maguer-Satta et al., 2001). The interaction of FLRG with activin suggests that FLRG could be involved in the regulation of the biological activity of activin, in particular in hematopoiesis. We have previously analysed FLRG expression in the hematopoietic system and shown that FLRG and activin were expressed in the same cells, and that they were up regulated during hematopoiesis. Moreover, in human bone marrow stromal cells, FLRG mRNA expression was rapidly induced by TGFbeta (Maguer-Satta, et al., 2001), a cytokine known to be involved in the regulation of cell proliferation and differentiation during hematopoiesis (Dybedal and Jacobsen, 1995; Krystal et al., 1994; Pierelli et al., 2000).

TGFbeta exerts a wide range of biological effects on a large variety of cell types, e.g. growth control, matrix production, cell adhesion and migration (Kingsley, 1994). The transduction of the TGFbeta signal is initiated by its binding to the heteromeric complex of type I and type II receptors, both of which exhibit serine/threonine kinase activity. Following ligand binding, the type II receptor phosphorylates and thereby activates the type I receptor cytoplasmic domains. Signaling from the receptor to the nucleus is mediated by the phosphorylation of cytoplasmic mediators, the Smad proteins (Heldin et al., 1997; Massague and Wotton, 2000; Wrana, 2000). The identification of Smad proteins as critical intracellular mediators of TGFbeta superfamily-transcriptional responses has contributed greatly to an understanding of how this family of cytokines produces their biological effect (Derynck et al., 1998). When phosphorylated by the activated receptors, Smad proteins transduce signals to the nucleus via homo- and heterooligomeric interactions. They activate the transcription of target genes in a cell-type-specific manner through interaction with DNA and with other nuclear factors (Massague and Wotton, 2000; ten Dijke et al., 2000). Regulation of the transcription of specific sets of genes by TGFbeta mediates many of these physiological roles. To date, few of these transcriptional targets have been identified. Molecular characterization of the mechanisms involved in the regulation of TGFbeta-responsive genes should provide an insight into TGFbeta physiological role.

The aim of this study was to characterize the FLRG minimal promoter and to analyse its transcriptional regulation by TGFbeta. FLRG is expressed as two major transcripts of 2.5 and 1.2 kb (Hayette et al., 1998). Here we show that the two FLRG transcripts were initiated at the same transcription start site, and that their transcription was directed by the same promoter. We cloned the promoter region of the human FLRG gene and defined the minimal region necessary to transcription activation in a reporter-system assay. The human hepatoma HepG2 cells respond to a variety of growth factors such as TGFbeta, and in this study we found that TGFbeta dramatically increased FLRG mRNA expression in these cells. We characterized the regulation of the expression of FLRG by TGFbeta at the transcriptional level, and showed that TGFbeta up-regulates FLRG expression through the action of Smad proteins, as demonstrated by FLRG promoter-reporter constructs, Northern blot, and electrophoretic mobility-shift assays.

Results and discussion

Differential polyadenylation of FLRG

We previously showed that FLRG is expressed as two major transcripts of 2.5 and 1.2 kb (Hayette et al., 1998). The 2.5 kb transcript (accession No. U76702) is the most abundant. To find out whether the shorter 1.2 kb transcript results from alternative transcription initiation or post-transcriptional modification such as polyadenylation and splicing, we screened a human cDNA library, using a fragment of the 5' end region of the 2.5 kb cDNA as a probe. Of 24 positive clones, eight did not hybridize to a probe specific of the 3' untranslated region of the 2.5 kb cDNA. Restriction mapping and partial sequences showed that all of them contained sequences coding from exon 1. The two longest clones were entirely sequenced; they were identical to the region encompassing nucleotides 1-1139 on the 2.5 kb transcript. Strangely, there was no recognizable polyadenylation sequence (AATAAA) in the sequence of the short transcript, but a close variant, GATAAA, was found at position 1112. A poly(A) tail of 48A was found at the 3' end of the cDNA, indicating that polyadenylation had occurred. The lack of the canonical AATAAA motif in a short transcript, coding for the same protein as the long transcript, has already been reported for other genes such as the human cyclin D1 (Xiong et al., 1991).

In summary, nucleotide sequence analysis revealed that the two FLRG transcripts had the same 5' end, and coded for the same protein. The smaller transcript corresponded to a shortened form of the 2.5 kb transcript resulting from the use of a different polyadenylation signal. These results indicate that both FLRG transcripts were initiated at the same transcription start site, and that their transcription was directed by the same promoter.

The 5' flanking region of the human FLRG gene exhibits promoter activity in HepG2 cells

In order to study the transcription regulation determining FLRG gene expression, we cloned and analysed the promoter activity of this gene. The 5400-bp human genomic DNA fragment containing sequences upstream from the FLRG cDNA was cloned from a human genomic library. The complete sequence of this region was found in GenBank under the description of the cosmid R33586 localized on chromosome 19 and containing the FLRG gene (accession No. AC004156). The promoter activity of this sequence was tested in transient transfection experiments in HepG2 cells using two reporter plasmids (pGL3-basic and pGL3-MLP) which contain different regions of the FLRG putative promoter in front of a luciferase reporter gene. The FLRG sequence, extending from positions -5386 to +6 relative to the major transcription start site (cloned into the pGL3-basic plasmid) mediated transcriptional activity in HepG2 cells (Figure 1a). In addition, transfection of HepG2 cells with a pGL3-basic plasmid containing the fragment extending from positions -398 to +6 resulted in a high level of luciferase activity (Figure 1a). The level of transcriptional activity observed with the -386 to +6-bp region is higher than the one obtained with the -5386 to +6-bp region suggesting the existence of putative inhibitory elements in the -5386 to -398-bp region.

These assays indicate that the promoter sequence from positions -398 to +6 is sufficient to mediate the transcriptional activation of the FLRG gene. Computer-assisted analysis revealed that this fragment (Figure 1b) did not contain the canonical CAAT box, but that it did exhibit several features of a promoter: a putative TATA homology element (in the noncoding strand at nucleotide -49) and a GC-rich flanking area (83% G/C between -130 and +6). There are four sequence motifs, at positions -114, -95, -83 and -78, which exactly match the recognition sequence of the transcription factor Sp1 in this sequence. Sp1 is a ubiquitously-expressed transcription factor with a zinc-finger DNA-binding domain that recognizes G/C-rich DNA sequences. Sp1 produces its transcriptional properties by interacting directly with factors of the basal transcription machinery and cooperating with several transcriptional activators (Cook et al., 1999; Suske, 1999).

To test the possibility that the Sp1 consensus sites may be involved in the mediation of FLRG promoter induction, various FLRG promoter fragments were cloned into the pGL3-MLP vector. The transcriptional activation induced by the sequence extending from positions -5386 to +6 was severely impaired by the deletion of the sequence between -398 and +6 (Figure 1c). The highest level of transcriptional activity in HepG2 cells was obtained with the reporter construct containing the -130/+6 region, which contains all the Sp1 consensus sites. We deleted these Sp1 sites of the FLRG promoter in the pGL3-MLP(-130/+6). Deletion of these sites in the pGL3-MLP(-130/+6;DeltaSp1) construct reduced the activity of the reporter to almost the background level (Figure 1c). Our initial functional characterization of the FLRG promoter suggests, that the proximal -130 to +6-bp region, which contains multiple Sp1-binding consensus sites, is required for constitutive FLRG promoter activity in hepatocytes, and that it contains an endogenous transcription initiation site controlled by the basic transcriptional machinery.

TGFbeta up-regulates FLRG expression at a transcriptional level in hepatocarcinoma cell lines

We have previously reported that both FLRG mRNA and protein were dramatically increased by TGFbeta in human bone marrow stromal cells (Maguer-Satta et al., 2001). TGFbeta increased the expression of activin and its ligand follistatin (Maguer-Satta et al., 2001; Zhang et al., 1997). Here we investigated whether TGFbeta modulated FLRG expression in hepatoma cell lines known to respond to a variety of growth factors such as TGFbeta. HepG2 and Hep3B cell lines, known to contain functional TGFbeta receptors, were treated with TGFbeta for different periods of time, and total cellular RNA was analysed by Northern blot using an FLRG cDNA probe. As shown in Figure 2a, TGFbeta treatment led to an accumulation of both the 2.5 and 1.2 kb FLRG transcripts in these cell lines. We observed a large increase in FLRG mRNA levels 3 h after the addition of TGFbeta to HepG2 and Hep3B cells. The time-course analysis showed a constant increase in the FLRG mRNA levels over time, the maximal effect being reached after 15 h, declining after longer incubation times.

To make sure that the increased FLRG mRNA levels correlated with an increase in protein synthesis, we performed immunoblotting experiments using a polyclonal antibody directed against FLRG, a secreted glycoprotein which has been shown by Western blot analysis to exist as a 33 kDa form under reducing conditions in the conditioned medium (Hayette et al., 1998). Figure 2b shows an accumulation of FLRG protein in cell lysate and the conditioned medium from HepG2 cells treated for 48 h with TGFbeta.

To find out whether the increase in FLRG expression was due to increased FLRG gene expression or to changes in mRNA stability, we analysed FLRG mRNA levels in HepG2 cells treated with the transcription inhibitor actinomycin D in the absence or presence of TGFbeta. Up-regulation of FLRG by TGFbeta was abolished by the addition of actinomycin D, suggesting that the increase in mRNA levels was at least partly due to a transcriptional event (Figure 2c). To make sure that the increase in FLRG expression in the presence of TGFbeta resulted from increased transcription of the FLRG promoter, we transfected HepG2 cells with the pGL3-MLP(-5386/+6) plasmid containing FLRG promoter elements between nucleotides -5386 and +6 driving the expression of the luciferase reporter gene. As shown in Figure 2d, TGFbeta produced a 3.5-fold induction of luciferase activity, indicating that the increase in luciferase activity reflected an increase in de novo transcription of the FLRG promoter in response to the addition of TGFbeta. All these results indicate that TGFbeta up-regulates FLRG expression at a transcriptional level in hepatocarcinoma cell lines.

Identification of TGF beta-responsive elements in the FLRG promoter

The process of identifying the TGFbeta-responsive elements in the FLRG promoter began with the cloning of various FLRG promoter 5' deletion fragments in the pGL3-MLP vector. Each construct was then fused with MLP-luciferase at position +6, and deletion constructs of the FLRG gene promoter were transfected into HepG2 cells in the presence or absence of TGFbeta. Deletions up to position -2306 relative to the initiation site did not significantly change the induction of the FLRG promoter by TGFbeta (Figure 3a). By contrast, the deletion of nucleotides -2306 to -1920 significantly reduced the TGFbeta activation of the FLRG promoter. These results are consistent with the existence of a TGFbeta-responsive element somewhere between the -2306 and -1920 nucleotides of the FLRG promoter.

Smad proteins are critical components of the TGFbeta signaling pathway (Massague and Wotton, 2000; ten Dijke et al., 2000). The MH1 domains of Smad3 and Smad4 is able to bind to specific DNA sequences that contain CAGA- or GTCT-like sequences, termed Smad-binding element (SBE). Optimal binding is achieved with the GTCTAGAC sequence (Zawel et al., 1998), identified by PCR-based oligonucleotide screening using the MH1 domain of Smad3 and Smad4. One or more SBE-like sequences have been identified in the promoters of several TGFbeta-responsive genes (Dennler et al., 1998; Hua et al., 1999; Jonk et al., 1998; Nagarajan et al., 1999; Vindevoghel et al., 1998). Transcriptional regulation by TGFbeta can be achieved by Smad proteins binding directly to this regulatory element within the promoter. Alternatively, Smad proteins can serve as co-activators of transcription through physical interactions and functional cooperation with other DNA-bound transcription factors (Chen et al., 1996a). Sequence analysis of the -2306 to -1920 region of the FLRG promoter has shown that there is an SBE, AGCCAGACA, at -2028, which perfectly matches the SBE consensus site identified in the human PAI-1 promoter (Dennler et al., 1998). To explore the biological significance of the SBE in the TGFbeta-mediated induction of the FLRG promoter, we mutated the AGCCAGACA sequence to AACGCGT in the pGL3-MLP(-2306/+6;DeltaSBE) construct. This mutation significantly reduced the luciferase activity of the FLRG promoter in the presence of TGFbeta by comparison with the pGL3-MLP(-2306/+6) construct (Figure 3b). This result indicates that this consensus Smad-binding motif in the FLRG promoter is involved in TGFbeta-mediated transcription activation.

Previous studies have shown that overexpression of Smad proteins activates transcription of several genes, in the absence of TGFbeta (Denissova et al., 2000; Feng et al., 2000; Jonk et al., 1998). Expression vectors for Smad2, Smad3 and Smad4 activate the luciferase activity of the pGL3-MLP(-2306/+6) FLRG promoter construct in the absence of TGFbeta (Figure 3c). When the pGL3-MLP(-2306/+6;DeltaSBE) mutated construct is cotransfected with Smad2, Smad3 and Smad4 expression vectors, the luciferase activity is significantly reduced. Nevertheless, with the mutated pGL3-MLP(-2306/+6;DeltaSBE) construct, a residual induction is observed in the presence of either TGFbeta or Smad proteins, suggesting the existence of another TGFbeta-responsive element. Altogether, our results indicate that this consensus Smad-binding motif in the FLRG promoter is involved in TGFbeta-mediated transcription activation and that Smad proteins are able to activate FLRG transcription via the SBE site located at position -2028. This result is in accordance with previous reports showing that the ability of several genes to respond to TGFbeta requires the presence of one or more SBE (Dennler et al., 1998; Jonk et al., 1998; Nagarajan et al., 1999; Zhang et al., 2000).

Several reports on different promoters have identified Sp1 sites as major TGFbeta-responsive promoter elements (Brodin et al., 2000; Feng et al., 2000; Pardali et al., 2000). To analyse whether the Sp1 DNA binding sequences found on the FLRG promoter are involved in response to TGFbeta, we generated a reporter construct where the Sp1 sites were deleted. Within the pGL3-MLP(-2306/+6) vector containing the SBE site, the region containing the Sp1 consensus sites (nt -104 to -66) in the FLRG promoter was deleted giving the pGL3-MLP(-2306/+6;DeltaSp1) construct. As we observed using the pGL3-MLP(-130/+6;DeltaSp1) construct (Figure 1c), deletion of the Sp1 sites in the pGL3-MLP(-2306/+6;DeltaSp1) vector reduced the basal promoter activity. This deletion significantly reduced the response to TGFbeta or Smad proteins of the FLRG promoter reporter construct (Figure 3b,c). As previously reported for Smad7 gene (Brodin et al., 2000), our results suggest that Sp1 transcription factor is essential for basal transcription and for the optimal inducibility by TGFbeta of the FLRG promoter.

Regulation of the FLRG promoter by Smad proteins

Having determined that the increase in FLRG transcription by TGFbeta required an SBE motif, we further analysed the involvement of Smad family proteins in FLRG transcriptional activation. The effects of Smad proteins on the FLRG promoter were examined by cotransfection of HepG2 cells with the pGL3-MLP(-3293/+6) promoter-luciferase construct along with expression vectors for Smad2, Smad3 and Smad4, either independently or in combination. As previously shown for other target genes transcriptionally regulated by TGFbeta (Chen et al., 1996b; Dennler et al., 1998; Nagarajan et al., 1999), transfection of Smad3 alone can stimulate FLRG promoter activity (Figure 4a). The highest level of transactivation was observed when all three Smad proteins were cotransfected. Our results are in agreement with previous reports (Feng et al., 2000; Moustakas and Kardassis, 1998; Nakao et al., 1997) that show a cooperation of Smad2, Smad3 and Smad4 in the TGFbeta-inducible transcriptional response. The fact that transactivation of the FLRG promoter by Smad proteins was higher than that achieved by TGFbeta suggests that endogenous Smad proteins are not expressed at saturating amounts in HepG2 cells.

We also examined the possibility that Smad protein expression correlates with induction of FLRG mRNA in HepG2 cells. For this purpose, we transfected expression vectors for Smad2, Smad3 and Smad4 proteins, independently or in combination, into HepG2 cells, and total cellular RNA were analysed by Northern blot analysis using FLRG cDNA as a probe. The expression pattern of endogenous FLRG mRNA (Figure 4b) shows a strong correlation with the results obtained using the FLRG promoter fused to luciferase as a cooperative effect of Smad2, Smad3 and Smad4 was also observed in FLRG mRNA expression.

Further evidence for the involvement of Smad proteins in the TGFbeta-mediated activation of FLRG promoter was obtained by the use of dominant-negative mutants of Smad3 and Smad4 from which the C-terminus, the MH2 domain, had been deleted. It has already been shown that these carboxyl-terminal truncated mutants can block TGFbeta signaling through Smad proteins (Lagna et al., 1996; Zhang et al., 1996). Transfection of Smad3 and Smad4 dominant-negative mutants significantly reduced the TGFbeta-mediated activation of the FLRG promoter (Figure 4c). Taken together, our results indicate that Smad proteins are involved in FLRG transcriptional activation by TGFbeta in HepG2 cells.

Binding of Smad3 and Smad4 to the SBE site

It is well documented that only Smad3 and Smad4 proteins can bind DNA. In order to activate the DNA binding of Smad3, the Smad3 MH1-MH2 interaction must be released. This occurs naturally by the phosphorylation of Smad3 by TGFbeta-type I receptor or can be achieved artificially when the C-terminus, the MH2 domain, is removed (Kawabata et al., 1998; Song et al., 1998, Zawel et al., 1998). To find out whether Smad3 and Smad4 proteins could interact directly with the SBE site at position -2028 of the FLRG promoter, we performed an electrophoretic mobility-shift assay (EMSA) using Smad3, Smad3DeltaC and Smad4 GST fusion proteins. We used as probe a DNA fragment of the FLRG promoter (nt -2093 to -1955) containing the wild-type (139-bp) or the mutated (137-bp) SBE site. As expected, the full-length GST-Smad3 protein showed no detectable binding to the probe (data not shown), while the wild-type probe was significantly shifted by the MH2-deletion mutants of Smad3 (Smad3DeltaC) and the full-length Smad4 fusion protein (Figure 5a). The binding of Smad3 and Smad4 was found to be specific, in that no shift was observed with the 137-bp probe containing the mutated SBE site. In addition, no shift was observed when the wild-type or the mutated probes were incubated with GST alone.

Having demonstrated that Smad proteins are able to bind the SBE motif of the FLRG promoter, we investigate whether TGFbeta induces binding of nuclear proteins to the SBE motif. We performed an EMSA using Mv1Lu cells nuclear extracts, treated or not with TGFbeta, and using a probe, containing the wild-type or the SBE mutated site, and covering the region between -2034 and -2014 of the FLRG promoter in three copy number. A specific TGFbeta-induced band with retarded mobility was observed using the wild-type probe and nuclear extracts from cells that had been treated with TGFbeta for 30 min, while no DNA-protein complex was observed with nuclear extracts from untreated cells (Figure 5b). In addition, no TGFbeta-induced bands were observed with the mutated probe. To analyse whether Smad proteins are present in nuclear complexes induced by TGFbeta, antibodies directed against Smad3 were added to the nuclear extracts before the addition of the radiolabeled probe. Smad3 antibody supershifted the TGFbeta-induced complex, while the negative control Smad7 antibody was unable to do so (Figure 5b). These results indicate that TGFbeta induces the formation of a DNA-protein complex containing Smad proteins and involving the Smad-binding motif, which is present in the FLRG promoter.

In conclusion, we are presenting here the first functional characterization of the FLRG promoter. We have shown that the fragment between positions -130 and +6, which consists of multiple consensus Sp1-binding sites, is required for the constitutive expression of the FLRG gene. Furthermore, our results indicate that TGFbeta induces the expression of the FLRG gene in hepatoma cells through a transcriptional mechanism, and the Smad proteins are involved in the activation of the FLRG promoter by TGFbeta. The SBE motif identified at position -2028 of the FLRG promoter matched the consensus site previously described, and can directly associate with Smad3 and Smad4 proteins. Our study provides strong evidence that this SBE is involved in the response of FLRG to TGFbeta. Thus the FLRG gene can be considered as a new target of TGFbeta transcriptional activation through the Smad proteins. A functional cooperation and a physical interaction between Smad and Sp1 proteins has been recently documented for TGFbeta-induced genes such as the cell cycle inhibitors p15(Ink4B) and p21/Waf1/Cip1 (Feng et al., 2000; Moustakas and Kardassis, 1998; Pardali et al., 2000). In the case of FLRG, a cooperation between Smad and Sp1 proteins might be involved in the transcription regulation of this gene. Further studies will be required to confirm this hypothesis.

FLRG protein, known to be highly homologous to follistatin, has also recently been shown to bind to activin, member of the TGFbeta superfamilly (Maguer-Satta et al., 2001). Follistatin, through its binding to activin, modulates its biological activity in a broad range of cells. This suggests that FLRG may constitute another partner in the regulation of the biological activity of activin (Tsuchida et al., 2000). TGFbeta increases the expression of both activin ligands, FLRG and follistatin (Maguer-Satta et al., 2001). Altogether, these data suggest that FLRG and follistatin might be involved in the biological effects of these TGFbeta superfamily cytokines. The involvement of FLRG in the cellular response to TGFbeta and activin, such as cell proliferation and differentiation, is currently under investigation.

Materials and methods

Cell cultures

The human hepatoma cell lines HepG2 and Hep3B, and the mink lung epithelial cell line Mv1Lu were obtained by American Type Culture Collection (ATCC), and cultured in Dulbecco's modified Eagle's medium containing 10% fetal calf serum, 0.03% L-glutamine, 100 mug/ml penicillin and 100 mug/ml streptomycin sulfate.

cDNA and DNA libraries

A human placental cDNA library in lambdaDR2 phage was obtained from Clontech Laboratories. The human placental genomic library cloned in EMBL3 phage (Stratagene) has already been described (Hayette et al., 1998).

Plasmid constructs

By screening the human placental genomic library with a probe specific to the 5' end of human FLRG cDNA, we isolated about 10 kb of the 5' sequence of the FLRG gene and subcloned insert from positive phages into the pBluescript KS(-) vector (Stratagene).

To study the transcriptional regulation of human FLRG gene expression, we subcloned the FLRG promoter extending from positions -5386 to +6 into the pGL3-basic luciferase reporter gene vector (lacking promoter and enhancer sequences; Promega) and into the pGL3-MLP vector containing a minimal promoter consisting of a TATA box and the initiator sequence of the adenovirus major late promoter (MLP), kindly provided by JM Gauthier (Glaxo Wellcome laboratory, Les Ullis, France) (Dennler et al., 1998). Various FLRG promoter fragments and 5' deletion fragments were cloned into pGL3-basic and pGL3-MLP vectors. Deletions of DNA were obtained by the unidirectional exonuclease III digestion method (Erase-a-base system, Promega). Within the pGL3-MLP(-2306/+6) vector, the SBE consensus site in the FLRG promoter was moved by site-directed mutagenesis (PCR based mutagenesis) from 5'-AGCCAGACA-3' to 5'-AACGCGT-3', giving the pGL3-MLP(-2306/+6;DeltaSBE) construct. Within the pGL3-basic(-130/+6) and the pGL3-MLP(-2306/+6) vectors, the region containing the Sp1 consensus sites (nt -104 to -66) in the FLRG promoter was deleted using the USE Mutagenesis kit (Amersham Pharmacia Biotech), giving respectively the pGL3-basic(-130/+6;DeltaSp1) and the pGL3-MLP(-2306/+6;DeltaSp1) constructs. All the constructs were sequence-checked.

Mammalian expression vectors encoding flag-tagged human Smad2, Smad3 and Smad4 were kindly provided by P ten Dijke (Ludwig Institute for Cancer Research, Uppsala, Sweden). Mammalian expression vectors encoding the dominant-negative human Smad3 (DN-Smad3) and Smad4 (DN-Smad4) proteins were kindly provided by R Derynck (University of California, San Francisco, USA).

Bacterial vectors encoding the full length Smad3 and Smad4 proteins and the MH2-deletion mutant of Smad3 (Smad3DeltaC), fused to GST, were kindly provided by A Mauviel (Hôpital Saint Louis, INSERM U532, Paris).

Transient tranfections and reporter assays

For transient transfections, HepG2 cells were seeded (1.4 ´ 105 cells per well) in 12-well plates and transfected 48 h later by Exgen 500 (Euromedex, Souffelweyersheim, France) with the indicated reporter constructs (1 mug) and the pRL-CMV internal control vector (25 ng) under the control of the cytomegalovirus promoter (Promega). When increasing amounts of Smad-expression vectors were transfected, total DNA was kept constant (1.5 mug) by the addition of an empty vector (pcDNA3.1 wild-type vector, Invitrogen). Ten ng/ml of human recombinant TGFbeta1 (R&D) were added 24 h after transfection, and after a further 24 h the luciferase activity was quantified in the cell lysates using the Dual Luciferase Assay (Promega). In all the experiments, performed in triplicate, luciferase activity was normalized by reference to the renilla luciferase activity expressed by the pRL-CMV vector.

For Northern blot analysis, HepG2 cells were seeded (4 ´ 105 cells per well) in 6-well plates and transfected 24 h later by Exgen 500 with expression vectors for Smad2, Smad3 and Smad4, either independently or in combination. The same amount of Smad expression vector (650 ng) was transfected in each case, total DNA was kept constant (2 mug) by the addition of an empty vector. Total RNA from transfected cells were extracted 48 h later.

Northern blot analysis

Total cellular RNA was purified by the acid guanidium thiocyanate-phenol-chloroform method. For Northern blot analysis, 10 mug of total RNA were size-fractionated in formaldehyde-1.2% agarose gels and transferred onto nylon filters. The probes were labeled with alpha-32P-dCTP using the Rediprime labeling kit (Amersham Pharmacia Biotech) and hybridized according to standard procedures. RNA integrity and equal loading was assessed by Ethidium Bromide staining.

Western blot analysis

The conditioned medium from a low-concentration serum culture concentrated five times and cell lysates (40 mug) from HepG2 cells either treated or not treated with TGFbeta1 (10 ng/ml) were subjected to electrophoresis on a 12% SDS-polyacrylamide gel. The separated proteins were transferred onto membrane (Millipore) by electroblotting. The filters were incubated with the rabbit anti-GST-FLRG antibody (Hayette et al., 1998) and with an anti-actin monoclonal antibody (Roche Molecular Biochemicals). Bound primary antibodies were detected with horseradish peroxydase-labeled anti-rabbit or anti-mouse secondary antibodies, and developed with enhanced chemiluminescence reagents purchased from Amersham Pharmacia Biotech. A pre-immune serum was used as a negative control (data not shown).

Electrophoresis mobility-shift assay

To prepare nuclear extracts, Mv1Lu cells were plated in 10-cm plates and grown to 75% confluency. Cells were then incubated in the presence or absence of TGFbeta (10 ng/ml) for 30 min. Cells, washed in cold phosphate-buffered saline, were lysed in hypotonic lysis buffer (20 mM HEPES pH 7.9, 20% glycerol, 10 mM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 0.1% Triton X-100, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium orthovanidate, and protease inhibitors). After 10 min, nuclei were spun down and incubated on ice during 45 min in the same lysis buffer containing 500 mM NaCl. The nuclei were lysed by 10 strokes of a Dounce all-glass homogenizer. Supernatants were then clarified and protein concentration was determined using a commercial assay kit (Bio-Rad). Full-length Smad3 and Smad4 proteins and the MH2-deletion mutant of Smad3 (Smad3DeltaC) fused to GST were expressed in Escherichia coli. The resulting recombinant proteins were purified over a GST-Sepharose column according to the manufacturer's instructions (Amersham Pharmacia Biotech).

Two probes spanning the region of the FLRG promoter between -2093 and -1955 were generated by polymerase chain reaction (PCR), using the pGL3-MLP(-2306/+6) or the pGL3-MLP(-2306/+6;DeltaSBE) plasmid as a template, and cloned into the pBluescript KS(-) vector (Stratagene). Wild-type 139-bp and mutated 137-bp DNA fragments containing, respectively, the Smad-binding element (SBE) at position -2028 (AGCCAGACA) and the mutated SBE at position -2028 (AACGCGT), excised from Bluescript SK(-) vector by BamHI.

Double stranded synthetic oligonucleotides spanning the region between -2034 and -2014 of the FLRG promoter in three copy number, and containing the wild-type or the mutated SBE site, were used as probes in the electrophoresis mobility-shift assay with Mv1Lu nuclear extracts: SBE-wt, 5'-GGCACGCCAGCCAGACAATGACCCACGCCAGCCAGACAAT
GACCCACGCCAGCCAGACAATGA-3', SBE-mut, 5'-GGCACGCCAGCTACATAATGACCCACGCCAGCTACATAATG
ACCCACGCCAGCTACATAATGA-3' (sense strands shown; the core binding site is underlined).

The probes were end-labeled by filling in using the Kleenow fragment of DNA polymerase (Biolabs) and alpha-32P-dCTP. The labeled probes purified by push-column chromatography were incubated with equal amounts of nuclear extracts (10 mug) or 300 ng of GST-Smad fusion proteins and 1 mug of polyd1-dC in a total volume of 30 mul of binding buffer (hypotonic lysis buffer containing 100 mM NaCl). Protein-DNA complexes were resolved on a 5% non-denaturing polyacrylamide gel in 0.5 ´ TBE buffer and visualized by autoradiography. For supershifts analysis, nuclear extracts were preincubated for 30 min at 37°C with 2 mug of antibodies against Smad3 or Smad7 (Santa Cruz Biotechnology), followed by addition of the radiolabeled probes.

Acknowledgements

This study was supported by grants from INSERM, the Association pour la Recherche contre le Cancer, the Ligue contre le Cancer (Comités du Rhône et de la Saône et Loire). We would like to thank JM Gauthier for providing the pGL3-MLP plasmid, P ten Dijke for the mammalian expression vectors encoding the flag-tagged human Smad2, Smad3 and Smad4, R Derynck for the mammalian expression vectors encoding human dominant-negative Smad3 (Smad3DeltaC) and Smad4 (Smad4DeltaC), and A Mauviel for bacterial vectors encoding GST-Smad proteins. We would like to thank MJ N'Guyen for her technical assistance. L Bartholin held a doctoral fellowship from the Ligue contre le Cancer, comité de la Haute Savoie.

References

Brodin G, Ahgren A, ten Dijke P, Heldin CH, Heuchel R. (2000). J. Biol. Chem. 275, 29023-29030. MEDLINE

Chen X, Rubock MJ, Whitman M. (1996a). Nature 383, 691-696. MEDLINE

Chen Y, Lebrun JJ, Vale W. (1996b). Proc. Natl. Acad. Sci. USA 93, 12992-12997. Article MEDLINE

Cook T, Gebelein B, Urrutia R. (1999). Ann. NY Acad. Sci. 880, 94-102. MEDLINE

Denissova NG, Pouponnot C, Long J, He D, Liu F. (2000). Proc. Natl. Acad. Sci. USA 97, 6397-6402. Article MEDLINE

Dennler S, Itoh S, Vivien D, ten Dijke P, Huet S, Gauthier JM. (1998). EMBO J. 17, 3091-3100. Article MEDLINE

DePaolo LV. (1997). Proc. Soc. Exp. Biol. Med. 214, 328-339. MEDLINE

Derynck R, Zhang Y, Feng XH. (1998). Cell 95, 737-740. MEDLINE

Dybedal I, Jacobsen SE. (1995). Blood 86, 949-957. MEDLINE

Feng XH, Lin X, Derynck R. (2000). EMBO J. 19, 5178-5193. Article MEDLINE

Hayette S, Gadoux M, Martel S, Bertrand S, Tigaud I, Magaud JP, Rimokh R. (1998). Oncogene. 16, 2949-2954. MEDLINE

Heldin CH, Miyazono K, ten Dijke P. (1997). Nature 390, 465-471. Article MEDLINE

Hua X, Miller ZA, Wu G, Shi Y, Lodish HF. (1999). Proc. Natl. Acad. Sci. USA 96, 13130-13135. Article MEDLINE

Johansson BM, Wiles MV. (1995). Mol. Cell. Biol. 15, 141-151. MEDLINE

Jonk LJ, Itoh S, Heldin CH, ten Dijke P, Kruijer W. (1998). J. Biol. Chem. 273, 21145-21152. Article MEDLINE

Kawabata M, Inoue H, Hanyu A, Imamura T, Miyazono K. (1998). EMBO J. 17, 4056-4065. Article MEDLINE

Kingsley DM. (1994). Genes Dev. 8, 133-146. MEDLINE

Kitamura K, Aota S, Sakamoto R, Yoshikawa SI, Okazaki K. (2000). Blood 95, 3371-3379. MEDLINE

Krystal G, Lam V, Dragowska W, Takahashi C, Appel J, Gontier A, Jenkins A, Lam H, Quon L, Lansdorp P. (1994). J. Exp. Med. 180, 851-860. MEDLINE

Lagna G, Hata A, Hemmati-Brivanlou A, Massague J. (1996). Nature 383, 832-836. MEDLINE

Maguer-Satta V, Bartholin L, Jeanpierre S, Gadoux M, Bertrand S, Martel S, Magaud J, Rimokh R. (2001). Exp. Hematol. 29, 301-308. MEDLINE

Massague J, Wotton D. (2000). EMBO J. 19, 1745-1754. Article MEDLINE

Moustakas A, Kardassis D. (1998). Proc. Natl. Acad. Sci. USA 95, 6733-6738. MEDLINE

Nagarajan RP, Zhang J, Li W, Chen Y. (1999). J. Biol. Chem. 274, 33412-33418. Article MEDLINE

Nakao A, Imamura T, Souchelnytskyi S, Kawabata M, Ishisaki A, Oeda E, Tamaki K, Hanai J, Heldin CH, Miyazono K, ten Dijke P. (1997). EMBO J. 16, 5353-5362. Article MEDLINE

Pardali K, Kurisaki A, Moren A, ten Dijke P, Kardassis D, Moustakas A. (2000). J. Biol. Chem. 275, 29244-29256. Article MEDLINE

Phillips DJ, de Kretser DM. (1998). Front Neuroendocrinol 19, 287-322. MEDLINE

Pierelli L, Marone M, Bonanno G, Mozzetti S, Rutella S, Morosetti R, Rumi C, Mancuso S, Leone G, Scambia G. (2000). Blood 95, 3001-3009. MEDLINE

Shao L, Frigon Jr NL, Young AL, Yu AL, Mathews LS, Vaughan J, Vale W, Yu J. (1992). Blood 79, 773-781. MEDLINE

Shiozaki M, Sakai R, Tabuchi M, Nakamura T, Sugino K, Sugino H, Eto Y. (1992). Proc. Natl. Acad. Sci. USA 89, 1553-1556. MEDLINE

Song CZ, Siok TE, Gelehrter TD. (1998). J. Biol. Chem. 273, 29287-29290. MEDLINE

Suske G. (1999). Gene 238, 291-300. Article MEDLINE

ten Dijke P, Miyazono K, Heldin CH. (2000). Trends Biochem. Sci. 25, 64-70. Article MEDLINE

Tsuchida K, Arai KY, Kuramoto Y, Yamakawa N, Hasegawa Y, Sugino H. (2000). J. Biol. Chem. 275, 40788-40796. MEDLINE

Vindevoghel L, Kon A, Lechleider RJ, Uitto J, Roberts AB, Mauviel A. (1998). J. Biol. Chem. 273, 13053-13057. MEDLINE

Wrana JL. (2000). Cell 100, 189-192. MEDLINE

Xiong Y, Connolly T, Futcher B, Beach D. (1991). Cell 65, 691-699. MEDLINE

Yamashita T, Takahashi S, Ogata E. (1992). Blood 79, 304-307. MEDLINE

Ying SY, Zhang Z, Furst B, Batres Y, Huang G, Li G. (1997). Proc. Soc. Exp. Biol. Med. 214, 114-122. MEDLINE

Yu J, Dolter KE. (1997). Cytokines Cell Mol. Ther. 3, 169-177. MEDLINE

Zawel L, Dai JL, Buckhaults P, Zhou S, Kinzler KW, Vogelstein B, Kern SE. (1998). Mol. Cell 1, 611-617. MEDLINE

Zhang W, Ou J, Inagaki Y, Greenwel P, Ramirez F. (2000). J. Biol. Chem. 275, 39237-39245. MEDLINE

Zhang Y, Feng X, We R, Derynck R. (1996). Nature 383, 168-172. MEDLINE

Zhang YQ, Kanzaki M, Shibata H, Kojima I. (1997). Biochim. Biophys. Acta. 1354, 204-210. MEDLINE

Figures

Figure 1 FLRG promoter activity in HepG2 cells. (a) HepG2 cells were transiently cotransfected with the indicated FLRG promoter-luciferase constructs in the pGL3-basic vector together with the pRL-CMV internal control vector. Luciferase activity was determined in the whole-cell lysate 48 h later. Relative luciferase activity (mean ± s.d.) was measured in three separate experiments, each performed in triplicate. One representative set of results is given here. (b) DNA sequence of the human FLRG promoter region -398/+10. The transcription initiation site (+1) and relevant endonuclease restriction sites are indicated. The ATG initiation codon is given in bold type. Consensus binding sites for the TATA box and the transcription factor Sp1 are underlined. (c) HepG2 cells were transiently cotransfected with the indicated FLRG promoter-luciferase constructs in the pGL3-MLP vector together with the pRL-CMV internal control vector. The four Sp1 consensus sites present at positions -114, -95, -83 and -78 in the FLRG promoter (see b) of the pGL3-MLP(-130/+6) vector were deleted, giving the pGL3-MLP(-130/+6;DeltaSp1) construct. Relative luciferase activity is given as the mean (±s.d.) of three independent experiments, each performed in triplicate. In all the experiments, luciferase activity was normalized by reference to the renilla luciferase activity expressed by the cotransfected pRL-CMV control plasmid

Figure 2 FLRG induction by TGFbeta in hepatocarcinoma cells. (a) Northern blot analysis with a specific FLRG cDNA alpha-32P-labeled probe using 10 mug of total RNA from Hep3B and HepG2 cells after treatment with TGFbeta for the indicated times. Equal amounts of RNA were checked by ethidium bromide (EtBr) staining. (b) Western blot analysis of supernatants (S) and lysates (L) from HepG2 cells, either treated or not with TGFbeta for 48 h. Proteins separated by SDS-12% polyacrylamide gel electrophoresis were immunodetected using the rabbit anti-GST-FLRG antibody. The same blot was probed for actin to control for equal protein loading. (c) Northern blot analysis with a specific FLRG cDNA alpha-32P-labeled probe using 10 mug of total RNA from HepG2 cells after treatment with TGFbeta (10 ng/ml) and actinomycin D (10 mug/ml) for the indicated times. Equal amounts of RNA were checked by ethidium bromide (EtBr) staining. (d) Transcriptional activation of FLRG by TGFbeta. HepG2 cells were cotransfected with the pGL3-MLP(-5386/+6) FLRG promoter-luciferase construct together with the pRL-CMV internal control vector. Twenty-four hours before harvesting, the transfected cells were either stimulated by TGFbeta or left unstimulated, and luciferase activity was determined. The relative luciferase activity is given as the mean ± s.d. of an experiment performed in triplicate, representative of three experiments

Figure 3 Identification of a TGFbeta-responsive element in the human FLRG promoter. (a) HepG2 cells were transiently cotransfected with the indicated 5' deletion constructs of the FLRG gene promoter in pGL3-MLP vector, along with the pRL-CMV internal control vector in the presence or absence of TGFbeta. Relative luciferase activity is expressed as the ratio (fold induction, mean ± s.d. of three experiments) of TGFbeta-activation to basal activity detected in HepG2 cells transfected with the reporter constructs in the absence of TGFbeta. (b) In the pGL3-MLP(-2306/+6;DeltaSBE) construct, the AGCCAGACA SBE motif at position -2028 in the pGL3-MLP(-2306/+6) was mutated by site-directed mutagenesis to AACGCGT. In the pGL3-MLP(-2306/+6;DeltaSp1) construct, the region containing the Sp1 consensus sites (nt -104 to -66) is deleted. HepG2 cells were transiently cotransfected with the indicated constructs of the FLRG gene promoter along with the pRL-CMV internal control vector, in the presence or absence of TGFbeta. Relative luciferase activity is given as the mean ± s.d. of an experiment performed in triplicate, representative of three experiments. (c) HepG2 cells were transiently cotransfected with the indicated constructs of the FLRG gene promoter in the pGL3-MLP vector together with the pRL-CMV internal control vector, in the presence or absence of the Smad2, Smad3 and Smad4 expression vectors. Relative luciferase activity is given as the mean ± s.d. of an experiment performed in triplicate, representative of three experiments

Figure 4 Activation of the FLRG promoter by Smad proteins. (a) HepG2 cells were cotransfected with the pGL3-MLP(-3293/+6) vector (or the wild-type pGL3-MLP vector) together with the pRL-CMV internal control vector and expression vectors for Smad2, Smad3 and Smad4 proteins, either separately or together, in the presence or absence of TGFbeta. The same amount of Smad expression vector (150 ng) was transfected in each case. The amount of DNA transfected was kept constant (1.5 mug) by the addition of an empty vector. Relative luciferase activity is given as the mean ± s.d. of an experiment performed in triplicate, and representative of three experiments, is expressed as the ratio of activation (mean ± s.d.) to basal activity detected in HepG2 cells transfected with the constructs in the absence of TGFbeta. (b) Northern blot analysis with a specific FLRG cDNA alpha-32P-labeled probe using 10 mug of total RNA from HepG2 cells cotransfected with the indicated Smad expression vectors in the presence or absence of TGFbeta. The same amount of Smad expression vector (650 ng) was transfected in each case. Total transfected DNA was kept constant (2 mug) by the addition of an empty vector. Equal amounts of RNA were checked by ethidium bromide (EtBr) staining. (c) HepG2 cells were cotransfected with the pRL-CMV internal control vector, along with the pGL3-MLP(-3293/+6) vector (1 mug) and expression vectors (500 ng) for dominant-negative forms of Smad3 (DN-Smad3) and Smad4 (DN-Smad4). Total transfected DNA was kept constant (1.5 mug) by the addition of an empty vector. Relative luciferase activity is expressed as the ratio of activation (mean ± s.d. of two experiments performed in triplicate) to basal activity detected in HepG2 cells transfected with the constructs in the absence of TGFbeta

Figure 5 Association of Smad3 and Smad4 with FLRG promoter. (a) Binding of Smad3 and Smad4 to the SBE of the FLRG promoter by gel mobility-shift assay. GST, GST-Smad3DeltaC and GST-Smad4 (300 ng/reaction) were incubated with the wild-type 139-bp (SBE-wt) and the mutated 137-bp (SBE-mut) DNA fragments used as probes. (b) An EMSA was performed using labeled probes containing three repeats of the wild-type or the mutated SBE site, as described in Materials and methods. Nuclear proteins (10 mug) were prepared from Mv1Lu cells either treated or not with TGFbeta for 30 min. The Smad3 antibody was used in the supershift assay to detect the presence of Smad3 protein in the TGFbeta-induced bands. The anti-Smad7 was used as control. The arrows denote the position of the TGFbeta-induced complex as well as the position of supershifted DNA-protein complex

Received 13 December 2000; revised 31 May 2001; accepted 14 June 2001
6 September 2001, Volume 20, Number 39, Pages 5409-5419
Table of contents    Previous  Article  Next    [PDF]