Advertisement
JBC

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M401610200 on July 1, 2004

J. Biol. Chem., Vol. 279, Issue 37, 38313-38324, September 10, 2004
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
279/37/38313    most recent
M401610200v1
Right arrow Submit a Letter to Editor
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowRequest Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Pinte, S.
Right arrow Articles by Leprince, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Pinte, S.
Right arrow Articles by Leprince, D.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

The Tumor Suppressor Gene HIC1 (Hypermethylated in Cancer 1) Is a Sequence-specific Transcriptional Repressor

DEFINITION OF ITS CONSENSUS BINDING SEQUENCE AND ANALYSIS OF ITS DNA BINDING AND REPRESSIVE PROPERTIES*

Sébastien Pinte{ddagger}, Nicolas Stankovic-Valentin{ddagger}, Sophie Deltour{ddagger}§, Brian R. Rood¶, Cateline Guérardel{ddagger}, and Dominique Leprince{ddagger}||

From the {ddagger}CNRS UMR 8526, Institut de Biologie de Lille, Institut Pasteur de Lille, 1 Rue Calmette, Lille Cedex 59017, France, and the Division of Pediatric Hematology/Oncology, Children's National Medical Center, Washington, D. C. 20010

Received for publication, February 13, 2004 , and in revised form, June 8, 2004.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
HIC1 (hypermethylated in cancer 1) is a tumor suppressor gene located at chromosome 17p13.3, a region frequently hypermethylated or deleted in human tumors and in a contiguous-gene syndrome, the Miller-Dieker syndrome. HIC1 is a transcriptional repressor containing five Krüppel-like C2H2 zinc fingers and an N-terminal dimerization and autonomous repression domain called BTB/POZ. Although some of the HIC1 transcriptional repression mechanisms have been recently deciphered, target genes are still to be discovered. In this study, we determined the consensus binding sequence for HIC1 and investigated its DNA binding properties. Using a selection and amplification of binding sites technique, we identified the sequence 5'-C/GNGC/GGGGCAC/A CC-3' as an optimal binding site. In silico and functional analyses fully validated this consensus and highlighted a GGCA core motif bound by zinc fingers 3 and 4. The BTB/POZ domain inhibits the binding of HIC1 to a single site but mediates cooperative binding to a probe containing five concatemerized binding sites, a property shared by other BTB/POZ proteins. Finally, full-length HIC1 proteins transiently expressed in RK13 cells and more importantly, endogenous HIC1 proteins from the DAOY medulloblastoma cell line, repress the transcription of a reporter gene through their direct binding to these sites, as confirmed by chromatin immunoprecipitation experiments. The definition of the HIC1-specific DNA binding sequence as well as the requirement for multiple sites for optimal binding of the full-length protein are mandatory prerequisites for the identification and analyses of bona fide HIC1 target genes.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
Hypermethylated in cancer 1 (HIC1)1 is located in 17p13.3 in a region frequently hypermethylated or deleted in human tumors (1–6). HIC1 has been considered a candidate tumor suppressor because its enforced expression by stable transfection in various cancer cell lines results in a significant decrease in their clonogenic survival and its expression is up-regulated by p53 (1, 7). Clues to the tumor suppressor function of HIC1 have recently come from the study of HIC1-deficient mice (8, 9) because heterozygous HIC1+/– mice develop, after 70 weeks, many different spontaneous malignant tumors (8).

HIC1 could also be implicated in a contiguous-gene syndrome, the Miller-Dieker syndrome, a severe form of lissencephaly accompanied by developmental anomalies (10). Although the isolated lissencephaly sequence is caused by point mutations or intragenic deletions in the LIS1 gene (11), the more severe cortical phenotype as well as the craniofacial and limb defects seen in the Miller-Dieker syndrome are the result of heterozygous deletions in another critical region in 17p13.3 including HIC1 (12). Together with perinatal death and a reduction in overall size, HIC1–/– mouse embryos have other developmental anomalies resembling those found in Miller-Dieker syndrome patients, including craniofacial dysmorphology, defects of the limbs and digits, and omphalocele (9). Finally, a high HIC1 expression is detected in the precursors of certain tissues that exhibit abnormalities in Miller-Dieker syndrome patients (13, 14).

HIC1 encodes a protein with five Krüppel-like C2H2 zinc fingers in the C terminus and a protein-protein interaction domain called the BTB/POZ domain at the N terminus (1, 15–17). The BTB/POZ domain is a conserved structural motif found mainly in transcription factors and actin-binding proteins. Crystal structures of the PLZF and BCL6 BTB/POZ domains have demonstrated that this domain is a tightly intertwined obligate dimer with a conserved dimerization interface (18, 19). Focusing on BTB/POZ transcription factors, they can function as transcriptional repressors, activators, or both, but in most cases, the homo/heterodimerization or even oligomerization properties of the BTB/POZ domain appear essential for their biological function. Indeed, the BTB/POZ domain can negatively regulate the DNA binding of the full-length protein to a single site (16). Likewise, the oligomerization of the protein via its BTB/POZ domain can mediate cooperative DNA binding to multiple sites as shown for GAGA (20–22), PLZF (23, 24), or Bab (25). Finally, the BTB/POZ domain is essential for the function of transcriptional repressors by directly recruiting nuclear corepressors (SMRT, N-CoR, or B-CoR)-histone deacetylase complexes, as shown for the human PLZF and BCL6 proteins (19, 26–28). Previously, we have shown that HIC1 is a transcriptional repressor (17) as is its avian ortholog, called {gamma}FBP-B ({gamma}F1-binding protein B) because it has been isolated as a sequence-specific transcriptional repressor of the {gamma}F-crystallin (Crygf) gene (29, 30). In addition, HIC1 and {gamma}FBP-B contain a short sequence, GLDLSKK, which is involved in the recruitment of the C-terminal binding protein (CtBP) corepressor (see Fig. 1) (31). Thus, HIC1 is clearly a transcriptional repressor, but so far no bona fide target genes for mammalian HIC1 proteins have been described. Although {gamma}FBP-B is expressed during chicken lens differentiation (29, 30), murine HIC1 gene expression is not detected in lenses at embryonic, postnatal, or adult stages by in situ hybridization studies, suggesting that it is not implicated in the regulation of Crygf gene (13).



View larger version (21K):
[in this window]
[in a new window]
  		FIG. 1.
Schematic structure of the proteins and oligonucleotides used in the HIC1 binding site selection. A, schematic drawing of the full-length FLAG-HIC1 1–714, of the FLAG-HIC1 5ZF-Cterm, and of the GST-HIC1 5ZF-Cterm fusion proteins. The BTB/POZ domain, the binding site for the CtBP corepressor (31), and the conserved C-terminal end (14) are represented, respectively, as a shaded, a black, and a dotted box. The five C2H2 Krüppel-like zinc fingers (ZF) are shown as black ovals. A FLAG epitope tagged at the N-terminal part of the HIC1 proteins is shown as a black triangle. B, sequences of the oligonucleotides used in this study. In the first set of experiments, the oligonucleotides used were 64 bp in length with constant region of 20 and 18 bp, flanking a random sequence region of 26 bp, N26 (34). In the second and third set, two modifications were introduced: a central TGC motif is flanked by 12 bp of random DNA sequence and one of the two C in the 3'-nonrandom sequence, which has been selected by several clones (Table I) has been replaced by a T (lowercase letters), yielding N12-TGC-N12.

 
As a first step toward the identification of HIC1 target genes, we have determined the HIC1-specific DNA binding sequence using multiple rounds of SAAB (selection and amplification of binding sites) (32). In silico analyses using the zinc finger recognition code (33) and electrophoretic mobility shift assays (EMSAs) with various mutated sequences fully validated this consensus and highlighted the GGCA core motif bound by zinc fingers 3 and 4. Notably, this consensus sequence is highly divergent from the {gamma}F1 site except for the presence of this GGCA core. A {Delta}POZ mutant protein can bind efficiently to a probe containing a single optimized sequence, whereas the full-length HIC1 protein cannot. By contrast, the full-length HIC1 protein binds cooperatively and in a BTB/POZ-dependent manner to a probe containing five concatemerized binding sites and represses the transcription of a reporter gene through its direct binding to these sites, as confirmed by ChIP experiments. These results have been obtained either with exogenous proteins expressed by transient transfection in RK13 cells or, more importantly, with endogenous HIC1 proteins found in the DAOY medulloblastoma cell line. The definition of the HIC1-specific DNA binding sequence as well as the requirement for multiple sites for optimal binding of the full-length protein are mandatory prerequisites for the identification and analyses of bona fide HIC1 target genes.


   EXPERIMENTAL PROCEDURES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
RESULTS
 DISCUSSION
 REFERENCES
 
Production of a GST-HIC1 Fusion Protein—A 1.2-kbp SmaI fragment of a human HIC1 clone encoding amino acids 348–714 was fused in-frame downstream of GST into the SmaI-digested pGEX-3X vector (Amersham Biosciences). This construct was introduced in Escherichia coli BL21 cells, and the GST-HIC1 5ZF-Cterm fusion protein was purified on glutathione-Sepharose beads (Amersham Biosciences) and eluted from the beads, as described previously (17, 28).

SAAB—To identify the HIC1 DNA binding consensus sequences, the SAAB method (32) was carried out with minor modifications. For the first set of binding site selections, a library of double-stranded random oligomers was obtained by PCR using the 64-base random oligonucleotide N26 (5'-GGCTGAGTCTGAACGGATCC-N26-CCTCGAGACTGAGCGTCG-3') as a template (10 ng) and 1 µg of primer 1 (5'-CGACGCTCAGTCTCGAGG-3') and primer 2 (5'-GGCTGAGTCTGAACGGATCC-3') for amplification (34). The double-stranded DNA (6 µl) was incubated on ice for 20 min with 0.3 µg of GST-HIC1 5ZF-Cterm fusion protein in a final 20-µl volume of binding buffer (20 mM Tris, pH 7.5, 80 mM NaCl, 0.1% Triton X-100, 2 mM dithiothreitol, 10 µM ZnCl2, 5% glycerol) containing 5 µg/ml poly(dI·dC). The resulting DNA complexes were separated by electrophoresis on an 8% nondenaturing polyacrylamide gel, excised from the gel, and eluted at 4 °C in 500 µl of water. The recovered 10 µl of DNA was amplified by 35 cycles of PCR with primers 1 and 2. After four rounds of selection, a fraction of the last PCR product was cloned directly using the TOPO TA Cloning Kit (Invitrogen) and sequenced.

A second set of experiments was performed with double-stranded random oligomers obtained by PCR using 10 ng of the 65-base random oligonucleotide N12-TGC-N12 (5'-GGCTGAGTCTGAACGGATCC-N12-TGC-N12-TCTCGAGACTGAGCGTCG-3') as a template and 1 µg of primer 1' (5'-CGACGCTCAGTCTCGAGA-3') and primer 2, as described above with the GST-HIC1 5ZF-Cterm protein. Finally, another set of selection was performed with this pool of N12-TGC-N12 oligomers and a FLAG-HIC1 5ZF-Cterm protein produced in rabbit reticulocyte lysate (2 µl).

Cell Culture and Transient Transfections—COS-7, RK13 (rabbit kidney), and the DAOY medulloblastoma cell line were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. Cells were transfected in OptiMEM by the polyethyleneimine method as described previously (31) either in 90-mm dishes for COS-7 cells (nuclear extracts) with 2.5 µg of DNA or in 12-well plates with 500 ng of DNA for RK13 cells (repression assays), for 6 h and then incubated in fresh complete medium. The cells were rinsed in phosphate-buffered saline 48 h after transfection and processed for nuclear extracts (COS-7) or luciferase assays (RK13). DAOY cells were transfected with LipofectAMINE 2000 (Invitrogen).

Repression Assays—The repression assays were carried out as described previously (17, 31). Luciferase and {beta}-galactosidase activities were measured by using, respectively, beetle luciferin (Promega) and the Galacto-light Kit (Tropix, Bedford, MA) with a Berthold (Nashua, NH) chemoluminometer. After normalization to {beta}-galactosidase activity, the data were expressed as Luc activity relative to the activity of pSV40-Luc with empty control vector (pcDNA3), which was given an arbitrary value of 1. Results presented are the mean values ± S.D. from one experiment representative of three independent transfections in triplicate.

Plasmid Construction—The full-length FLAG-HIC1 1–714 and {Delta}POZ mutant containing an N-terminal FLAG epitope cloned in the pcDNA3 expression vector have been described previously (31). The C521S point mutant (35) was generated by a two-round PCR strategy and verified by sequencing. To construct the 5xHiRE-SV40-Luc reporter plasmid, two complementary 5xHiRE (HIC1-responsive element) oligonucleotides containing overhanging GATC sequences (see below) were annealed and cloned into the BglII site of the pGL3-promoter vector (Promega) upstream of the luciferase gene driven by an SV40 promoter.

Nuclear Extracts, Western Blot, and EMSAs—Nuclear extracts were prepared from transiently transfected COS-7 or from DAOY cells according to the mini-extracts protocol (36). Protein expression was monitored by SDS-PAGE followed by Western blot analyses with the M2 monoclonal or with anti-HIC1 polyclonal (31) antibodies. These nuclear extracts or the various constructions, expressed in 2-µl reticulocyte lysates according to the manufacturer's instructions (Promega), were incubated on ice for 20 min in a final 20-µl volume of binding buffer as described above with the relevant {gamma}-32P-labeled probe. For competition assays, the cold competitor was included in the binding reaction. For antibody supershift experiments, the mixtures were incubated for 20 min at room temperature with the relevant antibody prior to the addition of the labeled probe and incubation on ice for a further 20 min. The sequences of the wild-type clone 5 probe and of the various mutated oligonucleotides used in competitive EMSAs are detailed in Table II. The oligonucleotides used for the cooperative binding assays and for the construction of the 5xHiRE reporter are as follows with the 5 HIC1 binding sites in bold: 5xHiRE probe wt (sense), 5'-GATCTGGGTTGCCCCCAGGGGGCAACCCAAGGGGGCAACCCTTCTAGAAGGGTTGCCCCCGGGGCAACCCA-3'; 5xHiRE probe wt (antisense), 5'-GATCTGGGTTGCCCCGGGGGCAACCCTTCTAGAAGGGTTGCCCCCTTGGGTTGCCCCCTGGGGGCAACCCA-3'.


View this table:
[in this window]
[in a new window]
  		TABLE II
Sequences of the oligonucleotides used in the competitive EMSAs

 
The sequence of the mutated 5xHiRE oligonucleotide used in competitive EMSAs is shown (sense strand) with the mutation in the TGC or GCA core sequences indicated in lowercase letters: 5xHiRE probe mutated (sense), 5'-GATCTGGGTcatCCCCAGGGGatgACCCAAGGGGatgACCCTTCTAGAAGGGTcatCCCCGGGatgACCCA-3'.

ChIP Assay—Briefly, 2 x 107 RK13 or DAOY cells transfected for 48 h with the indicated vectors were fixed in 1% formaldehyde for 10 min at room temperature. Nuclei were isolated, resuspended in 1.7 ml of sonication buffer (10 mM Tris, pH 8, 1 mM EDTA, 0.5 mM EGTA, and protease inhibitors), and sonicated. After centrifugation, the supernatant was removed (at this point, 160 µl was saved as input) and precleared for 15 min at 4 °C with 100 µl of preblocked protein A-agarose beads. After centrifugation the supernatants were removed (800 µl) and incubated with 10 µl of the precipitating antibody and 200 µl of immunoprecipitation dilution buffer (0.04% SDS, 4.4% Triton X-100, 4.8 mM EDTA, 66.8 mM Tris, pH 8, 167 mM NaCl, and protease inhibitors) overnight at 4 °C. 200 µl of preblocked protein A-agarose beads was added and incubated for 30 min at room temperature. Immune complexes were washed twice with dialysis buffer (2 mM EDTA, 50 mM Tris, pH 8, 0.2% sarkosyl) and four times with immunoprecipitation wash buffer (100 mM Tris, pH 9, 500 mM LiCl, 1% Nonidet P-40, 1% sodium deoxycholate). DNAs were eluted twice for 30 min in 500 µl of elution buffer (100 mM NaHCO3, 1% SDS). Then, NaCl (0.3 M final concentration) and 2 µl of RNase A (5 mg/ml) were added. The two eluates were combined and incubated at 67 °C for 5 h to reverse the cross-links. After ethanol precipitation, DNAs were extracted and used for PCR. Briefly, the PCRs were performed with 5 µl of DNAs, 200 ng of each primer (reverse primer, 5'-GCTGTCCCCAGTGCAAGTGC-3'; and forward primer, 5'-CTAATTGAGATGCAGATCGC-3') for 40 cycles.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
DISCUSSION
 REFERENCES
 
HIC1 Binding Site Selection—To define an optimal HIC1 DNA binding sequence, a GST-HIC1 5ZF-Cterm fusion protein was incubated with a mixture of double-stranded 64-mers containing 26 nucleotides of random core sequence, N26 (Fig. 1). Bound oligonucleotides were recovered and amplified by PCR (32). After four sequential rounds of EMSA/PCR selection, the bound oligonucleotides were subcloned, and 45 clones were sequenced. They were different, but strikingly, two-thirds (33 of 45) contained a TGC triplet just upstream from the CCTCGAG flanking sequence of the 3'-primer (Table I). In fact, all selected clones contain at least one copy of the TGCC (or of its complementary GGCA) sequence found in the {gamma}F1 binding motif (5'-TTCCTGCCAACACAGCAGACCTC-3'), which allowed the cloning of {gamma}FBP, the chicken ortholog of HIC1, by Southwestern screening of a lens cDNA library (29, 30). These experiments thus highlighted the TGCC sequence as a core motif, but the consensus binding site sequence could not be fully deduced because most of the selected sequences overlap with the common 3'-oligonucleotide (Table I).


View this table:
[in this window]
[in a new window]
  		TABLE I
Overview of the sequences (34 of 45) selected by the GST-5ZF-Cterm HIC1 fusion protein

Each sequence contains at least one TGCC (or GGCA) sequence (underlined and bold). Parts of the 5' and 3' nonrandom sequences are indicated in lowercase letters.

 
A second set of selection was performed with oligomers containing the central TGC sequence found in the first selection flanked by 12 nucleotides of random sequences, N12-TGC-N12 (Fig. 1B). After six rounds of selection with the GST-HIC1 5ZF-Cterm protein, the sequence of 74 clones revealed a 13-nucleotide GC-rich consensus centered on a TGCCC core, 5'-CCGGGTGCCC(G/C)(G/C)(G/C)-3' (Fig. 2A). Notably, the {gamma}F1 motif is highly divergent from this consensus, except for the presence of a TGCC motif (Fig. 2A).



View larger version (64K):
[in this window]
[in a new window]
  		FIG. 2.
Consensus sequence of the HIC1 DNA recognition sequence. A, the HIC1 binding sequences selected by the GST-HIC1 5ZF-Cterm fusion protein using the pool of N12-TGC-N12 oligonucleotides. For each position, the frequency of each nucleotide is indicated as a percentage with the most abundant underlined and in bold type. The following abbreviation are used: K for T or G; S for G or C; and B for C, G, or T. The consensus obtained is 5'-CCGGGTGCCC(G/C)(G/C)(G/C)-3'. B, the HIC1 binding sequences selected from the same pool of oligonucleotides by the FLAG-HIC1 5ZF-Cterm protein expressed in reticulocyte lysates. A slightly divergent and longer consensus was obtained, 5'-GGG(T/G)TGCCC(G/C)CN(G/C)GCC-3'. A sequence selected in both experiments, clone 5, is indicated. Except for the TGCC sequence, both consensus are highly divergent from the {gamma}F1 oligonucleotide shown below. C, comparison between the in silico and the experimentally defined HIC1 binding sequences. A potential model of the residue-base contacts between the HIC1 zinc fingers 2–4 and DNA has been deduced using an in silico code (33). Only the unambiguous contacts are shown. The arrows indicate the residue-base contacts. The DNA primary strand is shown in the anti-parallel direction (3' -> 5') compared with the zinc fingers (ZF2 -> ZF5). The consensus defined experimentally by the SAAB method is aligned below.

 
Finally, we performed a third set of selection with the N12-TGC-N12 oligomers and a FLAG-HIC1 5ZF-Cterm HIC1 protein expressed in rabbit reticulocyte lysate. After five rounds of selection, 32 clones were sequenced, and a similar consensus was obtained, albeit slightly extended in its 3'-end. In addition, a T or a Gis observed at similar frequency upstream of the TGC sequence (Fig. 2B) instead of the G observed preferentially with the bacterial fusion protein (Fig. 2A). One of the sequences, clone 5 (Cl5) found in the second set of selection with the GST-HIC1 5ZF-Cterm protein, was also selected three times by the FLAG-HIC1 5ZF-Cterm protein. Because its sequence very closely matched the consensus (Fig. 2B), it was chosen for the confirmatory EMSAs.

The crystal structure of the three zinc fingers of Zif268 bound to DNA, and numerous functional studies have established that each zinc finger contains a two-stranded antiparallel {beta}-sheet and an {alpha}-helix and binds 3 bp in the DNA major groove through specific contacts with the N-terminal part of the {alpha}-helix (37–39). Thanks to the constant accumulation of new data, a "recognition code" that relates the amino acids of a given zinc finger, notably residues at positions –1, +3, and +6 of the {alpha}-helix, to its DNA target has been proposed (33) and validated for natural (38) or engineered (40) zinc fingers. We have applied this code to zinc fingers 2–5 of HIC1, which are separated by a 7–8-amino acid conserved H/C link typical of Krüppel-like zinc fingers making them likely to be involved in sequence-specific DNA binding (Fig. 1A) (37). Indeed, a construct containing only the zinc fingers 2–4 binds specific DNA as efficiently as a 5ZF or 5ZF-Cterm construct (data not shown). According to the matrix proposed by Choo and Klug (33), an in silico defined binding sequence can be assigned at some positions (Fig. 2C). Interestingly, the TGC triplet that emerged from our first selection (see Table I) and was thus imposed in the second and third sets of selection is unambiguously predicted for the zinc finger 3 of HIC1 (Fig. 2C). In addition, an arginine at –1 and at +6, when present, is always associated with binding to a G at the third and the first positions of two consecutive subsites, respectively (39), a fact that validates our experimentally defined TGCC core (Table I; Fig. 2, A and B). Finally, because these analyses also defined the primary DNA strand with which most of the amino acid contacts are made, the HIC1 core consensus motif should thus be viewed in the following orientation, 5'-GGCA-3', which is shown in the antiparallel direction to zinc fingers (C-terminal finger 5 to N-terminal finger 2 versus DNA in the 5' -> 3' orientation) (Fig. 2C) (37–39).

In conclusion, our experiments defined a HIC1 consensus binding sequence centered on a GGCA core motif. Notably, the {gamma}F1 motif is highly divergent except for the presence of this GGCA motif.

Validation of the HIC1 Consensus Binding Site—To confirm the consensus, we used the binding site selected in clone 5 as a probe and performed competitive EMSAs using either the wild-type clone 5 or the {gamma}F1 sequences as cold competitor. As shown in Fig. 3, the FLAG-HIC1 5ZF-Cterm protein expressed in reticulocyte lysates bound to the labeled site 5 used as a probe (lane 2), and a 100-fold excess of cold site 5 almost completely abolished this binding (lane 3). By contrast, a 100-fold excess of the cold {gamma}F1 sequence inhibited only weakly this binding (lane 4). Thus, although the chicken {gamma}FBP-B/HIC1 (30) and the human HIC1 (data not shown) zinc fingers domain can bind the {gamma}F1 sequence, this sequence is not an optimal consensus binding site.



View larger version (75K):
[in this window]
[in a new window]
  		FIG. 3.
Validation of the HIC1 consensus binding sequence by competitive EMSA. Binding of the FLAG-HIC1 5ZF-Cterm protein expressed in rabbit reticulocyte lysates was examined by EMSA using the radiolabeled site 5 as probe (lanes 2–12). In lane 1, an unprogrammed lysate was incubated with the probe as control. Specificity of the binding obtained in the absence of competitor (–, lanes 2 and 5) was analyzed by a competition assay using a 100-fold excess of the indicated cold competitor. wt refers to the sequence selected in clone 5 and {gamma}F1 to a sequence derived from the {gamma}-F crystallin gene (see Table II). The various mutant oligonucleotides containing triple base substitutions in the clone 5 sequences are shown schematically above, and their complete sequence is listed in Table II.

 
The majority of base contacts between zinc fingers and DNA occurs in 3-bp segments along one strand of the DNA, called the primary strand (38). To confirm the specificity of the interaction of HIC1 with clone 5 and to define the contribution of each zinc finger to this binding, we designed oligonucleotides containing triple base substitutions in the segment predicted for each zinc finger and performed competitive EMSAs as described above. First, we tested two mutants of the 3 bp flanking either in 5' or in 3' the 12-bp consensus predicted for four zinc fingers. Both the mut b (5') and the mut a (3') (Table II) oligonucleotides inhibited the binding of the FLAG-HIC1 5ZF-Cterm protein to the labeled wild-type probe 5 (Fig. 3, lanes 5, 7, and 12), as efficiently as the wild-type clone 5 used as its own cold competitor (Fig. 3, lane 6). Even though a weak consensus sequence has been obtained in the binding site selection for these 5'- and 3'-sequences (Fig. 2B), these results demonstrate that they are not crucial for the HIC1-specific DNA binding activity. Thus, the HIC1 consensus binding sequence can be restricted to 12 bp. The mutants containing triple base substitutions in the segment predicted for each zinc finger yielded different results. Indeed, mut II, mut III, and mut IV mutated in the sequence recognized, respectively, by zinc fingers 2, 3, and 4 competed poorly (mut II) or very poorly (mut III and IV) with the wild-type probe 5 (Fig. 3, lanes 5 and 8–10). Notably, these mutants contained base substitutions at positions that were the most invariant in our binding site selection and included the GGCA core. Finally, mut V corresponding to the sequences bound by zinc finger 5 competed very efficiently (Fig. 3, lane 11). Accordingly, this portion of the defined consensus is the less specific (5-G/CNG-3') (Fig. 2, B and C). Taken together these results suggest that the residues bound by zinc fingers 3 and 4 are the most important in the consensus in contrast to residues bound by zinc finger 2 and especially zinc finger 5.

Having established the contribution of each zinc finger, we tried to define the critical contact residues within the consensus binding site. To that end, another series of oligonucleotides containing single base transversions within the sequences bound by zinc fingers 2–4 were tested in competitive EMSAs. All of these mutants competed very efficiently for the binding of the FLAG-HIC1 5ZF-Cterm protein to the wild-type consensus site 5 used as a probe, except for mutants 3, 4, and 6 (Fig. 4, lanes 3, 5, and 7). In fact, these mutated sites contain base transversions in the GGCA core derived from the binding site selection (Fig. 2 and Table II). Strikingly, mutant 5, where a T replaces the C in the GGCA core, behaves as a very efficient competitor (Fig. 4, lane 6). Because this nucleotide corresponds to the G of the TGC triplet imposed in the selection procedure and predicted by the zinc finger code, it thus seemed important to analyze the effect of each possible nucleotide substitution at this position. Although the wild-type (GGCA) and mut 5 (GGTA) sequences are efficient competitors (Fig. 4, lanes 16 and 18), the mut 5' (GGGA) and mut 5'' (GGAA) sequences are unable to compete for the binding of HIC1 on its consensus sequence (Fig. 4, lanes 17 and 19).



View larger version (75K):
[in this window]
[in a new window]
  		FIG. 4.
Mutational analyses of the HIC1 consensus binding sequence by competitive EMSAs. Left panel, binding of the FLAG-HIC1 5ZF-Cterm protein expressed in rabbit reticulocyte lysates was examined by EMSA using the radiolabeled site 5 as probe (lanes 1–14). Specificity of the binding obtained in the absence of competitor (–, lane 14) was analyzed by a competition assay using a 400-fold excess of the indicated cold competitor. Mut 1 to mut 9 (lanes 1–3, 5–7, and 9–11) contain a single point mutation in the sequence selected in clone 5 (wt) (see Table II). Mut IV, III, and II containing triple substitutions (Table II) were also used in this assay (lanes 4, 8, and 12). The various mutant oligonucleotides containing single base substitutions in the clone 5 sequences are shown schematically above, and their complete sequence is listed in Table II. Right panel, similar analyses were performed with mut 5, 5', and 5'' (lanes 17–19), which represent each possible point mutation of the C nucleotide in the GGCA core (Table II). EMSAs were also performed in the absence (–, lane 15) or in the presence of the wild-type (lane 16) or mut III (lane 20) cold competitors.

 
Collectively, these results demonstrate that the optimized consensus binding site for HIC1 is centered on a core GG(C/t)A motif and that these nucleotides are essentially those bound by zinc fingers 3 and 4.

The BTB/POZ Domain Inhibits the Binding of HIC1 to a Single Consensus Binding Site in Transfected Cells—Because our binding site selection and mutational analyses have been performed with BTB/POZ-deleted proteins, as in most previously published cases (16, 25, 34, 41, 42), and because the BTB/POZ domain can influence the binding properties of full-length proteins (16, 20, 21), it was thus important to verify that the full-length HIC1 protein was readily able to bind our defined consensus site. To that end, we transfected into COS-7 cells expression vectors for the full-length FLAG-HIC1 1–714 protein and for a protein lacking the BTB/POZ domain (FLAG-HIC1 {Delta}POZ), both tagged in the N terminus with a FLAG epitope (Fig. 5A). Western blotting analyses with the anti-FLAG monoclonal antibody (M2) of nuclear extracts prepared from transfected or mock-transfected cells demonstrated that both proteins are expressed at similar levels (Fig. 5A, lanes 2 and 3). Using the 32P-labeled probe 5 carrying a single HIC1 consensus binding site in the EMSA, several retarded complexes were observed with equal amounts of nuclear extracts from the FLAG-HIC1 {Delta}POZ-transfected cells (Fig. 5B, lanes 4 and 7). The lowest mobility complex (arrowhead in Fig. 5B) could correspond to the FLAG-HIC1 {Delta}POZ protein, and the other (asterisk in Fig. 5B) to N-terminal truncated products. Indeed, all of these complexes are specific because they are not observed in the mock-transfected cell extracts (Fig. 5B, lanes 1–3) and because they are competed by an excess of cold wild-type probe 5 (Fig. 5B, lanes 5 and 8) but not by the cold mut III oligonucleotide mutated in the GGCA core (Table II) (Fig. 5B, lanes 6 and 9). Notably, equal amounts of nuclear extracts from the FLAG-HIC1 1–714 expression vector generated a significantly lower amount of protein-DNA complexes that comigrated with those observed in the FLAG-HIC1 {Delta}POZ-transfected cells (Fig. 5B), suggesting that they could represent complexes containing truncated proteins generated during the nuclear extracts preparation instead of the full-length HIC1 protein. This hypothesis was confirmed by supershift experiments using two antibodies directed against the N-terminal and C-terminal ends of these proteins. Indeed, the complexes observed in the FLAG-HIC1 {Delta}POZ-transfected cells are supershifted both by the M2 monoclonal antibody directed against the N-terminal FLAG epitope and by a rabbit polyclonal antibody directed against the C-terminal residues of HIC1 (701–714) (31) (Fig. 5C, lanes 7 and 8). By contrast, the lowest mobility complexes observed with the expression vector for the full-length HIC1 protein are supershifted only by the anti-HIC1 Cterm antibody (Fig. 5C, lane 3) but not by the M2 monoclonal antibody (Fig. 5C, lane 4) directed against the N-terminal epitope (Fig. 5A). Thus, the proteins engaged in these complexes are not full-length proteins but rather proteins truncated in their N-terminal end, notably in their BTB/POZ domain.



View larger version (68K):
[in this window]
[in a new window]
  		FIG. 5.
{Delta}POZ HIC1, but not full-length HIC1, can bind to a single optimal consensus binding sequence. A, the full-length FLAG-HIC1 1–714 and the FLAG-HIC1 {Delta}POZ proteins are depicted schematically on the left. The localization of the antigenic determinants recognized by the M2 monoclonal antibody raised against the N-terminal FLAG epitope (black triangle) and the polyclonal anti HIC1 Cterm (Ct) antibodies raised against the last 14 residues (31) are shown as arrows. Nuclear extracts prepared from COS-7 cells transfected by the empty pcDNA3-FLAG (lane 1) or by the indicated expression vector (lanes 2 and 3) were analyzed by Western blot (WB) using the M2 monoclonal antibody. B, DNA binding activity of HIC1 transfected cells. Nuclear extracts, prepared from COS-7 cells transfected with the empty pcDNA3-FLAG (lanes 1–3) or the FLAG-HIC1 1–714 (lanes 4–6) and FLAG-HIC1 {Delta}POZ (lanes 7–9) vectors, were analyzed for DNA binding activity with the 32P-labeled probe 5. Specificity of the binding obtained in the absence of competitor (–, lanes 1, 4, and 7) was analyzed by a competition assay using a 400-fold excess of the above indicated wild-type (probe 5, lanes 2, 5, and 8) or mut (mut III, lanes 3, 6, and 9) cold competitor. The filled triangle indicates the position of the retarded complexes with the lowest mobility. The asterisk corresponds to specific complexes with truncated proteins. C, DNA·HIC1 complexes observed with the full-length protein contain N-terminally truncated proteins. Nuclear extracts of the FLAG-HIC1 1–714 (lanes 1–4) and FLAG-HIC1 {Delta}POZ (lanes 5–8) transfected cells were analyzed by EMSA as described in B. Antibody supershift experiments were performed using 4 µl of preimmune rabbit serum (PI, lanes 2 and 6), 4 µl of anti-HIC1 Cterm immune serum from the same rabbit (I, lanes 3 and 7), or 1 µg of the M2 monoclonal antibody (M2, lanes 4 and 8). The filled triangle indicates the same retarded complexes as in B) (lowest mobility), and the arrows indicate the supershifted bands.

 
These results thus demonstrate that the BTB/POZ domain of HIC1 impedes the binding of the full-length protein to a single site.

HIC1 Binds Cooperatively to a Probe Containing Multiple Consensus Binding Sites—The fact that the full-length HIC1 protein is unable to bind with high affinity to a probe containing a single site in contrast to {Delta}POZ proteins is highly reminiscent of the situation described previously for the GAGA factor (20–22). Indeed, the BTB/POZ domain through its dimerization (oligomerization) properties mediates cooperative binding of GAGA to natural promoters or to a probe containing multiple binding sites. To address this issue, we synthesized a 71-bp oligonucleotide containing five copies of the optimal binding site 5 (5xHiRE) in the orientation used by Katsani et al. (21). This labeled probe was tested in EMSAs with a range of concentrations of unprogrammed and FLAG-HIC1 1–714- or {Delta}POZ-programmed reticulocyte lysates. The FLAG-HIC1 {Delta}POZ protein generated a ladder of complexes which gradually increased in size and intensity when the concentration of expressed protein increased (Fig. 6, lanes 3, 6, and 9), presumably reflecting an increase in the occupancy of independent sites by individual FLAG-HIC1 {Delta}POZ proteins. By contrast, with the full-length HIC1 protein, a single low mobility complex was observed, which indicates that all of the binding sites become occupied at once (Fig. 6, lanes 2, 5, and 8). In addition, we also observed a high mobility complex that presumably corresponds to the binding of a single N-terminally truncated protein as observed with nuclear extracts (Fig. 5). Notably, at the highest protein concentration, the DNA/FLAG-HIC1 full-length complexes are not very abundant, whereas the probe is almost completely shifted by a similar amount of FLAG-HIC1 {Delta}POZ protein (Fig. 6, lanes 8 and 9).



View larger version (49K):
[in this window]
[in a new window]
  		FIG. 6.
Full-length HIC1 can bind cooperatively to multiple optimal consensus binding sequences. An oligonucleotide containing five copies of the optimal HIC1 binding site selected in clone 5 (5xHiRE) closely packed in various orientations as described by Katsani et al. (21) for the GAGA factor was synthesized. This labeled probe was used in EMSAs with increasing amounts of calibrated (see left inset) unprogrammed reticulocyte lysates (Retic., lanes 1, 4, and 7) and lysates programmed with FLAG-HIC1 1–714 (lanes 2, 5, and 8) or FLAG-HIC1 {Delta}POZ (lanes 3, 6, and 9) vectors. The filled triangles indicate the ladder-like retarded complexes presumably corresponding to the increasing occupancy of the probe by one, two, or three HIC1 {Delta}POZ proteins (lanes 3, 6, and 9). The white triangle corresponds to a single low mobility complex observed with the full-length HIC1 protein (lanes 2, 5, and 8). A high mobility complex probably corresponding to truncated proteins as observed in the nuclear extracts (Fig. 5) is also visible in these lanes.

 
To demonstrate further that these low mobility complexes are specific for the full-length HIC1 proteins, we performed two additional control experiments. First, we engineered by PCR mutagenesis a HIC1 full-length protein where the second cysteine in zinc finger 3, which binds the GCA triplet in the core motif, has been mutated to serine, HIC1 C521S (Fig. 7A). Such a point mutation has been shown to abolish the binding of the Drosophila Krüppel protein to its site (35). Indeed, in the context of the FLAG-HIC1 5ZF-Cterm protein, this mutation abrogates the binding to the clone 5 probe (data not shown). Similarly, no complexes were observed between the full-length HIC1 C521S protein and the 5xHiRE probe in EMSAs (Fig. 7B, lanes 4–6) in contrast to the wild-type protein (Fig. 7B, lanes 1–3). Conversely, we synthesized a 71-bp 5xHiRE mutant probe where each GCA or TGC motif has been mutated in ATG or CAT, as in mut III (see Table II). When this 5xHiRE mutated probe was used in EMSAs, no complexes were observed with the full-length FLAG-HIC1 1–714 or C521S proteins nor even with the FLAG-HIC1 {Delta}POZ protein (Fig. 7B, lanes 7–10).



View larger version (38K):
[in this window]
[in a new window]
  		FIG. 7.
Mutations either in zinc finger 3 of full-length HIC1 protein or in each GCA core motif of the 5xHiRE probe abrogate the formation of DNA·HIC1 complexes. A, schematic drawing of the FLAG-HIC1 C521S mutant protein in which the second cysteine in zinc finger 3 is replaced by a serine. This mutation has no impact on the production in reticulocyte lysate compared with the wild-type protein (lanes 2 and 3). Lane 1 is an unprogrammed lysate shown as control. B, left panel, binding activity of the wild-type or C521S FLAG-HIC1 full-length proteins expressed in rabbit reticulocyte lysates was examined by EMSA using the 32P-labeled 5xHiRE probe (lanes 1–6). Specificity of the binding obtained in the absence of competitor (–, lanes 1 and 4) was analyzed by a competition assay using a 400-fold excess of the indicated cold competitor. wt is the 5xHiRE probe used as its own competitor (lanes 2 and 5); mut corresponds to a 5xHiRE probe in which each GCA triplet of the core GGCA has been mutated in ATG, thus mimicking mut III (see Table II and "Experimental Procedures") (lanes 3 and 6). The asterisk (*) represents a nonspecific band also observed in unprogrammed reticulocyte lysates (data not shown). The white triangle corresponds to a single low mobility complex observed with the full-length HIC1 protein; the black triangle corresponds to a high mobility complex containing truncated proteins. Right panel, conversely, the 32P-labeled 5xHiRE mutated oligonucleotide was used in EMSAs with calibrated amounts of unprogrammed reticulocyte lysates (Retic., lane 7) and lysates programmed with FLAG-HIC1 {Delta}POZ (lane 8), FLAG-HIC1 1–714 (lane 9), or FLAG-HIC1 C521S (lane 10) vectors.

 
These results demonstrate that full-length HIC1 proteins expressed in reticulocyte lysates can bind cooperatively and in a BTB/POZ-dependent manner to a probe containing multiple copies of its optimal binding sequence albeit with a much lower affinity than a {Delta}POZ protein.

Full-length HIC1 and HIC1 {Delta}POZ Bind the 5xHiRE Probe with Similar Affinity in Transfected Cells—We then investigated the binding to the 5xHiRE probe of full-length FLAG-HIC1 proteins expressed in vivo, using nuclear extracts from transfected cells. As shown in Fig. 8, at least three types of high molecular weight complexes, called a, b and c, were observed with the wild-type full-length FLAG-HIC1 1–714 proteins expressed in transfected COS-7 cells (Fig. 8B, lanes 8 to 12). These complexes are specific because they are not observed in the mock (Fig. 8B, lanes 2 to 6) or in the FLAG-HIC1 C521S transfected cells extracts (Fig. 8B, lanes 20 to 24) and because they are competed by an excess of cold 5xHiRE probe but not by an excess of mutated probe (Fig. 8C, lanes 2 and 3, respectively). These complexes contain the full-length FLAG-HIC1 1–714 protein because their mobility is very low as compared with the DNA/FLAG-HIC1 {Delta}POZ complexes (Fig. 8B, compare lanes 8–12 with lanes 14–18). One of them, complex a, is even retained at the top of the gel. More importantly, the complexes b and c are supershifted both by the N-terminal anti-FLAG and C-term anti-HIC1 antibodies (Fig. 8C, compare lanes 1, 5, and 6). Finally and in striking contrast with the EMSA performed with the proteins expressed in reticulocyte lysates (Fig. 6), at each protein concentration the amount of complexes obtained between the 5xHiRE probe and the FLAG-HIC1 full-length 1–714 or {Delta}POZ proteins are very similar (Fig. 8B, compare lanes 11 and 12 with lanes 17 and 18, respectively).



View larger version (82K):
[in this window]
[in a new window]
  		FIG. 8.
Full-length HIC1 binds cooperatively to multiple optimal consensus binding sequences. A, nuclear extracts prepared from COS-7 cells transfected by the empty pcDNA3-FLAG (lane 1) or by the indicated expression vector (lanes 2–4) were analyzed by Western blot (WB) using the M2 monoclonal antibody. B, increasing amounts (from 0.125 to 2 µl) of nuclear extracts from COS-7 cells transfected by the empty pcDNA3 (lanes 2–6), by the FLAG-HIC1 1–714 (lanes 8–12), by the FLAG-HIC1 {Delta}POZ (lanes 14–18), and by the FLAG-HIC1 C521S (lanes 20–24) vectors were tested by EMSA. The white triangles correspond to three low mobility complexes (a, b, and c) observed only with the full-length HIC1 1–714 protein (lanes 8–12). Complex a is retained at the top of the gel. Lanes 1, 7, 13, and 19 correspond to the migration of the probe alone. C, the specificity of the complexes obtained between the FLAG-HIC1 1–714 protein and the 5xHiRE probe (lane 1) was validated by competitive EMSA with a 400-fold excess of the wild-type (wt, lane 2) or mutated (mut, lane 3) cold 5xHiRE sequence as well as antibody supershift experiments using the preimmune (PI, lane 4) and immune (I, lane 5) anti-HIC1 C-term rabbit polyclonal antibodies or the anti-FLAG M2 (M2) monoclonal antibody (lane 6). Complexes a, b, and c are indicated as in B).

 
Thus, when expressed in cells by transient transfection, full-length HIC1 and {Delta}POZ proteins bind to a probe containing multiple copies of their optimal binding site with similar affinity, albeit with different mechanisms (individual versus cooperative binding).

HIC1 Represses the Transcriptional Activity of a Reporter Gene Containing the 5xHiRE Sequence—We previously identified several autonomous repression domains in HIC1, namely in its BTB/POZ domain and in its central region, using chimeras with the GAL4 DNA binding domain (17, 31). Characterization of a high affinity binding site for HIC1 thus allowed us to test directly whether HIC1 could function as a transcriptional repressor when bound to this site in vivo. To that end, the 5xHiRE element was cloned upstream of the luciferase gene driven by the SV40 promoter in a pGL3-promoter reporter vector. Transfection of this pGL3–5xHiRE-SV40-Luc reporter with the FLAG-HIC1 1–714 expression vector in RK13 cells resulted in a clear dose-dependent repression of the reporter activity (Fig. 9A). Conversely, and in close agreement with the EMSAs (Fig. 8), no repression was observed with a similar dose of FLAG-HIC1 C521S mutated in the DNA binding domain (Fig. 9A).



View larger version (21K):
[in this window]
[in a new window]
  		FIG. 9.
HIC1 is a transcriptional repressor when bound to its multiple optimal consensus-binding sequence. A, the structure of the 5xHiRE-SV40-Luc reporter gene is shown schematically. RK13 cells were transiently cotransfected with 225 ng of the SV40-Luc (white columns) or 5xHiRE-SV40-Luc (black columns), with the indicated amount of pcDNA3 FLAG-HIC1 1–714 or C521S expression vectors. The luciferase activity was normalized to the {beta}-galactosidase activity of a cotransfected {beta}-OS-LacZ construct (50 ng). The total amount of DNA was kept constant (500 ng) in each transfection by the addition of pcDNA3. The results are the mean values ± S.D. from one independent transfection in triplicate representative of three independent experiments. B, ChIP assay. RK13 lysates were prepared 48 h after cotransfection of the 5xHiRE-SV40-Luc reporter gene and the indicated expression vector. Chromatin was immunoprecipitated with the anti HIC1 Cterm immune serum (I) or preimmune serum (PI) from the same rabbit as control, purified, and used as template for PCR with oligonucleotides flanking the concatemerized HIC1 binding sites. Input corresponds to 10% of the chromatin. A control PCR in the absence of template was also performed.

 
To confirm that this transcriptional repression readily relies on the binding of HIC1 to its target sites, we performed a ChIP assay. RK13 cells were transfected with the pGL3–5xHiRE-SV40-Luc reporter gene together with the FLAG-HIC1 1–714 or FLAG-HIC1 C521S expression vectors. Chromatin was then immunoprecipitated using the preimmune or immune anti-HIC1 Cterm serum, and the presence of immunoprecipitated DNA was checked by PCR using oligonucleotides flanking the HIC1 binding sites in the reporter gene. A PCR fragment was obtained only with the immune serum, in the cells transfected by the full-length HIC1 protein, whereas no DNA was coimmunoprecipitated by the C521S mutant (Fig. 9B). As control, the input DNA for each sample gave a similar signal, indicating that equivalent amounts of DNA were present.

These results thus demonstrate that the full-length HIC1 protein can repress transcription when bound to its specific DNA binding sequence.

Endogenous HIC1 Proteins Repress Transcription via the 5xHiRE Sequence—All of the above described results thus indicate that the 5xHiRE sequence could be a suitable target to study HIC1-mediated transcriptional regulation in vivo. Recently, we detected by reverse transcription-PCR analyses HIC1 expression in a human medulloblastoma cell line, DAOY.2 In line with these results, endogenous HIC1 proteins can be detected in DAOY nuclear extracts by Western blot analyses using our anti-HIC1 antibodies (Fig. 10A). In addition, upon incubation of these nuclear extracts with the 5xHiRE probe, we observed two specific complexes as demonstrated by the use of wild-type and mutated competitors (Fig. 10B). Notably, one of them is retained at the top of the gel as described previously for the full-length HIC1 protein transiently expressed in COS-7 cells (Fig. 8). Transfection of the pGL3–5xHiRE-SV40-Luc reporter gene in the DAOY cells resulted in a 5-fold decrease of the transcriptional activity of the SV40 promoter compared with the native pGL3-SV40-Luc vector (Fig. 10C). This suggests that endogenous protein(s) recognizing the 5xHiRE sequences are able to repress transcription. Because endogenous HIC1 proteins are present in DAOY nuclear extracts, it is likely that one of these endogenous proteins is the HIC1 gene product itself. This hypothesis was confirmed by a ChIP experiment using the anti-HIC1 antibodies (Fig. 10D).



View larger version (28K):
[in this window]
[in a new window]
  		FIG. 10.
Endogenous HIC1 proteins repress transcription of the pGL3 5xHiRE-SV40-Luc reporter gene. A, detection of endogenous HIC1 proteins in the DAOY medulloblastoma cell line. Nuclear extracts were resolved by SDS-PAGE and analyzed by Western blotting using polyclonal antibodies raised against the C-terminal end of HIC1 (590–714) fused to GST (31) (lane 1); by the same immune serum adsorbed with an excess of the bacterial GST peptide (Ads GST, lane 2) or adsorbed with an excess or the purified GST-HIC1 polypeptide used to immunize the rabbit (Ads GST-HIC1, lane 3). B, nuclear extracts from DAOY cells were tested by EMSA. The specificity of the complexes obtained with the 5xHiRE probe (lane 1) was validated by competitive EMSA with a 400-fold excess of the wild-type (wt, lane 2) or mutated (mut, lane 3) cold 5xHiRE sequence. The arrows correspond to two specific complexes, one of them being retained at the top of the gel. C, DAOY cells were transiently transfected with 900 ng of the SV40-Luc (white column) or 5xHiRE-SV40-Luc (black column) reporter plasmid. The luciferase activity was normalized to the {beta}-galactosidase activity of a cotransfected {beta}-OS-LacZ construct (100 ng). The results are the mean values ± S.D. from one transfection in triplicate representative of two independent experiments. D, ChIP assay. DAOY lysates were prepared 48 h after transfection of the 5xHiRE-SV40-Luc reporter gene, and a ChIP assay was performed with the anti HIC1 Cterm immune serum (I) or preimmune serum (PI) exactly as described in Fig. 9.

 
Thus, endogenous HIC1 proteins can bind and repress transcription through their binding to the 5xHiRE sequence that we have defined.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
REFERENCES
 
We have shown previously that HIC1 is a transcriptional repressor that contains several autonomous repression domains and is capable of recruiting the CtBP corepressor (17, 31). However, the HIC1-specific DNA binding sequence and hence its target genes, involved in tumorigenesis or developmental regulation, remain ill defined. To date, the only putative target gene described is the Crygf gene (29, 30), although the lack of expression of murine HIC1 in the eye at any developmental stage has cast some doubt on these results (13). As a first step toward the identification of bona fide HIC1 target genes, we performed a site selection using the SAAB method based on the amplification by PCR of oligonucleotides selected by the HIC1 protein from a pool of oligomers containing random sequences. The first set of selection highlighted a TGC triplet located just upstream of the common sequence used for PCR amplification (Table I). Two other series of in vitro binding experiments with oligonucleotides containing a central TGC motif defined a 12-bp sequence as our empirically determined HIC1 binding site. This consensus, and notably the imposed TGC, was consistent with the in silico zinc finger code established by Choo and Klug (33). Interestingly enough, this code also defines the primary DNA strand with which most of the amino acid contacts are made, hence giving the 5' -> 3' orientation of the target sequence. These in vitro and in silico analyses thus defined the 12-bp sequence 5'-C/GNGC/GGGGCAC/ACC-3' as an optimal binding site for HIC1. Mutational analyses of the site highlighted the functional importance of a core GGCA motif (Figs. 3 and 4), which notably, is also found in the {gamma}F1 motif, albeit embedded in a significantly divergent context. These results could thus explain why the zinc finger domain of chicken {gamma}FBP-B has been isolated during the screening of a chicken lens cDNA library using the concatemerized {gamma}F1 motif as a probe (30). Indeed, the HIC1 5ZF-Cterm protein is also able to bind this {gamma}F1 motif, albeit with a lower affinity than the consensus clone 5 probe (data not shown). However, these results, together with the lack of expression of murine HIC1 in various differentiation stages of the eye (13), strongly suggest that the Crygf gene is not a bona fide target gene, at least for the mammalian HIC1 proteins.

The full-length HIC1 protein expressed either in reticulocyte lysates or in transfected COS-7 cells is unable to bind to a single optimized site, and this DNA binding inhibition is clearly BTB/POZ-dependent. Indeed, the only retarded bands obtained with extracts from cells transfected with the full-length HIC1 expression vector contain N-truncated proteins, as demonstrated by supershift experiments with the anti-FLAG antibody (Fig. 5C). Conversely, the {Delta}POZ mutant binds this probe very efficiently. Such a situation is not unprecedented because this BTB/POZ inhibitory effect, even on heterologous DNA binding domains, was described when the founder of this protein family was identified a decade ago (16). A similar effect has also been shown for the Drosophila GAGA factor (20–22) and the human BCL6 protein (41, 43). However, this is not a general property of the BTB/POZ transcription factors because the PLZF protein can bind single sites in the HoxD (23) or Hoxb2 (24) homeotic genes as well as in the c-myc proto-oncogene (44) at least in EMSA, although the transcriptional repression observed in vivo relies on the cooperative binding to multiple sites (23, 24).

Although the HIC1 BTB/POZ domain impedes binding to a single site, it mediates strong cooperative binding to a probe containing multiple optimized sites (Figs. 6 and 8). In fact, the crystal structure of the PLZF BTB/POZ domain demonstrated that it forms an obligate tightly interlaced dimer that can even oligomerize (18). This property is also shared by the BCL6 BTB/POZ domain, indicating that this architectural feature could be a common core fold for this protein family (19). Consequently, BCL6 can bind two adjacent sites located in the first noncoding exon of its gene. These sites constitute a major negative autoregulation loop controlling BCL6 expression in normal cells and are a major target for the deregulation of BCL6 associated with some human leukemias. Indeed, they are eliminated either by promoter substitution because of the various recurrent 3q27 chromosomal translocations affecting BCL6 in diffuse large B cell lymphomas (DLBCLs) or by point mutations identified in DLBCLs without any obvious 3q27 translocations (45, 46). Similarly, the BTB/POZ protein FBI (factor binding to IST) binds, in the HIV promoter, two half-sites, each containing an imperfect palindrome but whose orientation and spacing can greatly vary (47).

Besides these two examples, several studies provide compelling evidence that the BTB/POZ proteins can bind several, not necessarily optimal, binding sites to exert their biological function. This model stems from the role of the Drosophila in the regulation of the Ultrabithorax (ubx) homeotic gene (20–22, 48). As shown by DNase footprinting, transient transfection assays, or electron microscopy, GAGA factor could first, in conjunction with chromatin remodeling activities such as NURF (49, 50) or FACT (51), create a nucleosome-free region. Then, through its BTB/POZ-mediated oligomerization GAGA factor could simultaneously bind several distant binding sites, thereby linking cis or even trans enhancers or silencers and their promoters into a single regulatory unit (for a recent review, see Ref. 48). This local open chromatin environment could ultimately facilitate different regulatory responses, through transcriptional regulators recruiting genetically antagonistic polycomb (PcG) or trithorax (trxG) complexes. Such an architectural role has also been proposed for the Drosophila Bab proteins, which bind several sites in the large bab locus (25).

In line with the above described model, the vertebrate PLZF protein can form DNA loops and bridge together multiple distant PLZF binding sites in the HoxD gene (23) or cooperate through its binding to a single PLZF site with an A/T-rich binding factor bound to a distant site in the regulation of the Hoxb2 enhancer (24). Notably, PLZF interacts with the mammalian PcG protein Bmi-1, colocalizes with PcG nuclear bodies (23) and PLZF–/– mice show striking patterning defects in the limbs and axial skeleton (52).

In this paper, we have shown that both transiently expressed and endogenous HIC1 proteins can bind to a probe containing multiple binding sites designed according to the 5xGAGA probe template used by Katsani et al. (21). In reticulocyte lysates, although the {Delta}POZ proteins bind this probe additively, as single molecules, the full-length HIC1 proteins generate a single low mobility complex. This "all or nothing" pattern suggests that a consistent number of binding sites become occupied all at once by a single HIC1 oligomer whose protein stoichiometry is not known. The formation of these low mobility DNA-full-length HIC1 complexes appeared to be highly favored in nuclear extracts compared with reticulocyte lysates (Fig. 8). These results raise the possibility that post-translational modification(s) and/or interaction(s) with protein partners would be required for the optimal HIC1 DNA binding activity.

Finally, such specific low mobility complexes are also observed upon incubation of nuclear extracts prepared from the DAOY cell line expressing endogenous HIC1 proteins with the 5xHiRE probe (Fig. 10), a finding that fully validates our in vitro defined consensus.

As a whole, our results would suggest that HIC1 target promoters could contain several copies of HIC1 binding sites and that the HIC1 transcriptional repression could be brought about by a cooperative binding to multiple, even suboptimal, binding sites or in conjunction with other transcription factors bound to distant sites. To begin to address these issues, we have searched in promoter data bases for potential target genes containing this site. For example, these analyses identified several candidate HIC1 response elements in a 1.6-kbp promoter fragment of the human cyclin D1 gene, one of them perfectly matching the GGGCAACC consensus established in this study. Preliminary data demonstrate that a {Delta}POZ HIC1 protein binds this site in EMSAs, further suggesting that HIC1 may play a role in the regulation of cyclin D1 transcription (data not shown).

Undoubtedly, the HIC1 consensus binding sequence as well as the binding mechanisms characterized here will be helpful for the identification of direct target genes implicated in the tumorigenesis and/or developmental defects associated with a loss or decrease in HIC1 expression and will provide insight for the functional analyses of their promoters.


   FOOTNOTES
 
* This work was supported by funds from CNRS, the Pasteur Institute, and the Association pour la Recherche contre le Cancer. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Back

§ Present address: Welcome Trust, University of Cambridge, Cambridge, United Kingdom. Back

|| To whom correspondence should be addressed. Tel.: 33-3-2087-1119; Fax: 33-3-2087-1111; E-mail: dominique.leprince{at}ibl.fr.

1 The abbreviations used are: HIC1, hypermethylated in cancer 1; BTB/POZ, broad complex, Tramtrack and Bric à brac/poxviruses and zinc finger; ChIP, chromatin immunoprecipitation; EMSA, electrophoretic mobility shift assay; {gamma}FBP-B, {gamma}F1-binding protein B; GST, glutathione S-transferase; HiRE, HIC1-responsive element; Luc, luciferase; SAAB, selection and amplification of binding sites; ZF, zinc finger; CtBP, C-terminal binding protein. Back

2 B. Rood, unpublished results. Back


   ACKNOWLEDGMENTS
 
We thank Jean Coll, Jean-Louis Couderc, and Alexis Verger for critical reading of the manuscript.



   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Makos-Wales, M., Biel, M., El Deiry, W., Nelkin, B. D., Issa, J. P., Cavenee, W. K., Kuerbitz, S. J., and Baylin, S. B. (1995) Nat. Med. 1, 570–577[CrossRef][Medline] [Order article via Infotrieve]
  2. Baylin, S. B., Esteller, M., Rountree, M. R., Bachman, K. E., Schuebel, K., and Hermann, J. G. (2001) Hum. Mol. Genet. 10, 687–692[Abstract/Free Full Text]
  3. Eguchi, K., Kanai, Y., Kobayashi, K., and Hirohashi, S. (1997) Cancer Res. 57, 4913–4915[Abstract/Free Full Text]
  4. Fujii, H., Biel, M. A., Zhou, W., Weitzman, S. A., Baylin, S. B., and Gabrielson, E. (1998) Oncogene 16, 2159–2164[CrossRef][Medline] [Order article via Infotrieve]
  5. Melki, J. R., Vincent, P. C., and Clark, S. J. (1999) Leukemia 13, 877–883[CrossRef][Medline] [Order article via Infotrieve]
  6. Rood, B. R., Zhang, H., Weitman, D. M., and Cogen, P. H. (2002) Cancer Res. 62, 3794–3797[Abstract/Free Full Text]
  7. Guérardel, C., Deltour, S., Pinte, S., Monté, D., Bègue, A., Godwin, A. K., and Leprince, D. (2001) J. Biol. Chem. 276, 3078–3089[Abstract/Free Full Text]
  8. Chen, W. Y., Zeng, X., Carter, M., Morell, C. N., Chiu Yen, R.-W., Esteller, M., Watkins, D. N., Herman, J. G., Mankowski, J. L., and Baylin, S. B. (2003) Nat. Genet. 33, 197–202[CrossRef][Medline] [Order article via Infotrieve]
  9. Carter, M. G., Johns, M. A., Zeng, X., Zhou, L., Zink, M. C., Mankowski, J. L., Donovan, D. M., and Baylin, S. B. (2000) Hum. Mol. Genet. 9, 413–419[Abstract/Free Full Text]
  10. Yingling J., Toyo-oka, K., and Wynshaw-Boris, A. (2003) Am. J. Hum. Genet. 73, 475–488[CrossRef][Medline] [Order article via Infotrieve]
  11. Wynshaw-Boris, A., and Gambello, M. J. (2001) Genes Dev. 15, 639–651[Free Full Text]
  12. Cardoso, C., Leventer, R. J., Ward, H. L., Toyo-oka, K., Chung, J., Martin, C. L., Allanson, J., Pilz, D. T., Olney, A. H., Mutchinick, O. M., Hirotsune, S., Wynshaw-Boris, A., Dobyns, W. B., and Ledbetter, D. H. (2003) Am. J. Hum. Genet. 72, 918–930[CrossRef][Medline] [Order article via Infotrieve]
  13. Grimm, C., Spörle, R., Schmid, T. E., Adler, I.-D., Adamski, A., Schughart, K., and Graw, J. (1999) Hum. Mol. Genet. 8, 697–710[Abstract/Free Full Text]
  14. Bertrand, S., Pinte, S., Stankovic-Valentin, N., Deltour-Balerdi, S., Guérardel, C., Bègue, A., Laudet, V., and Leprince, D. (2004) Biochim. Biophys. Acta 1678, 57–66[Medline] [Order article via Infotrieve]
  15. Albagli, O., Dhordain, P., Deweindt, C., Lecocq, G., and Leprince, D. (1995) Cell Growth Differ. 6, 1193–1198[Abstract]
  16. Bardwell, V. J., and Treisman, R. (1994) Genes Dev. 8, 1664–1677[Abstract/Free Full Text]
  17. Deltour, S., Guérardel, C., and Leprince, D. (1999) Proc. Natl. Acad. Sci. U. S. A. 96, 14831–14836[Abstract/Free Full Text]
  18. Ahmad, K. F., Engel, C. K., and Privé, G. G. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 12123–12128[Abstract/Free Full Text]
  19. Ahmad, K. F., Melnick, A., Lax, S., Bouchard, D., Liu, J., Kiang, C.-L., Mayer, S., Takahashi, S., Licht, J. D., and Privé, G. G. (2003) Mol. Cell 12, 1551–1564[CrossRef][Medline] [Order article via Infotrieve]
  20. Espinas, M. L., Jimenez-Garcia, E., Vaquero, A., Canudas, S., Bernues, J., and Azorin, F. (1999) J. Biol. Chem. 274, 16461–16469[Abstract/Free Full Text]
  21. Katsani, K. R., Hajibagheri, M. A., and Verrijzer, C. P. (1999) EMBO J. 18, 698–708[CrossRef][Medline] [Order article via Infotrieve]
  22. Mahmoudi, T., Katsani, K. R., and Verrijzer, C. P. (2002) EMBO J. 21, 1775–1781[CrossRef][Medline] [Order article via Infotrieve]
  23. Barna, M., Merghoub, T., Costoya, J. A., Ruggero, D., Branford, M., Bergia, A., Samori, B., and Pandolfi, P. P. (2002) Dev. Cell 3, 499–510[CrossRef][Medline] [Order article via Infotrieve]
  24. Ivins, S., Pemberton, K., Guidez, F., Howell, L., Krumlauf, R., and Zelent, A. (2003) Oncogene 22, 3685–3697[CrossRef][Medline] [Order article via Infotrieve]
  25. Lours, C., Bardot, O., Godt, D., Laski, F. A., and Couderc, J. L. (2003) Nucleic Acids Res. 31, 5389–5398[Abstract/Free Full Text]
  26. Melnick, A., and Licht, J. D. (1999) Blood 93, 3167–3215[Free Full Text]
  27. Melnick, A. M., Carlile, G., Ahmad, K. F., Kiang, C. L., Corcoran, C., Bardwell, V., Prive, G. G., and Licht, J. D. (2002) Mol. Cell. Biol. 22, 1804–1818[Abstract/Free Full Text]
  28. Dhordain, P., Albagli, O., Lin, R. J., Ansieau, S., Quief, S., Leutz, A., Kerckaert, J. P., Evans, R. M., and Leprince, D. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 10762–10769[Abstract/Free Full Text]
  29. Liu, Q., Tini, M., Tsui, L.-C., and Breitman, M. L. (1991) Mol. Cell. Biol. 11, 1531–1537[Abstract/Free Full Text]
  30. Liu, Q., Shalaby, F., Puri, M. C., Tang, S., and Breitman, M. L. (1994) Dev. Biol. 165, 165–177[CrossRef][Medline] [Order article via Infotrieve]
  31. Deltour, S., Pinte, S., Guérardel, C., Wasylyk, B., and Leprince, D. (2002) Mol. Cell. Biol. 22, 4890–4901[Abstract/Free Full Text]
  32. Blackwell, T. K., and Weintraub, H. (1990) Science 250, 1104–1110[Abstract/Free Full Text]
  33. Choo, Y., and Klug, A. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 11168–11172[Abstract/Free Full Text]
  34. Kawamata, N., Miki, T., Ohashi, K., Suzuki, K., Fukuda, T., Hirosawa, S., and Aoki, N. (1994) Biochem. Biophys. Res. Commun. 204, 366–374[CrossRef][Medline] [Order article via Infotrieve]
  35. Reemann, N., Gaul, U., and Jackle, H. (1988) Nature 322, 90–92
  36. Schreiber, E., Matthias, P., Müller, M. M., and Schaffner, W. (1989) Nucleic Acids Res. 17, 6419[Free Full Text]
  37. Pavletich, N. P., and Pabo, C. O. (1991) Science 252, 809–817[Abstract/Free Full Text]
  38. Wolfe, S. A., Nekludova, L., and Pabo, C. O. (1999) Annu. Rev. Biomol. Struct. 3, 183–212
  39. Iuchi, S. (2001) Cell Mol. Life Sci. 58, 625–635[CrossRef][Medline] [Order article via Infotrieve]
  40. Corbi, N., Perez, M., Maione, R., and Passananti, C. (1997) FEBS Lett. 417, 71–74[CrossRef][Medline] [Order article via Infotrieve]
  41. Seyfert, V. L., Allman, D., He, Y., and Staudt, L. M. (1996) Oncogene 122, 2331–2342
  42. Li, J.-Y., English, M. A., Ball, H. J., Yeyati, P. L., Waxman, S., and Licht, J. D. (1997) J. Biol. Chem. 272, 22447–22455[Abstract/Free Full Text]
  43. Deweindt, C., Albagli, O., Bernardin, F., Dhordain, P., Quief, S., Lantoine, D., Kerckaert, J. P., and Leprince, D. (1995) Cell Growth Differ. 6, 1495–1503[Abstract]
  44. McConnell, M. J., Chevallier, N., Berkofsky-Fessler, W., Giltname, J. M., Malani, R. B., Staudt, L. M., and Licht, J. D. (2003) Mol. Cell. Biol. 23, 9375–9388[Abstract/Free Full Text]
  45. Wang, X., Li, Z., Naganuma, A., and Ye, B. H. (2002) Proc. Natl. Acad. Sci. U. S. A. 99, 15018–15023[Abstract/Free Full Text]
  46. Pasqualucci, L., Migliazza, A., Basso, K., Houldsworth, J., Chaganti, R. S. K., and Dalla-Favera, R. (2003) Blood 101, 2914–2923[Abstract/Free Full Text]
  47. Pessler, F., and Hernandez, N. (2003) J. Biol. Chem. 278, 29327–29335[Abstract/Free Full Text]
  48. Lehmann, M. (2004) Trends Genet. 20, 15–22[CrossRef][Medline] [Order article via Infotrieve]
  49. Tsukiyama, T., and Wu, C. (1995) Cell 83, 1011–1020[CrossRef][Medline] [Order article via Infotrieve]
  50. Tsukiyama, T., Daniel, C., Tamkun, J., and Wu, C. (1995) Cell 83, 1021–1026[CrossRef][Medline] [Order article via Infotrieve]
  51. Shimojima, T., Okada, M., Nakayama, T., Ueda, H., Okawa, K., Iwamatsu, A., Handa, H., and Hirose, S. (2003) Genes Dev. 17, 1605–1616[Abstract/Free Full Text]
  52. Barna, M., Hawe, N., Niswander, L., and Pandolfi, P. P. (2000) Nat. Genet. 25, 166–172[CrossRef][Medline] [Order article via Infotrieve]

Add to CiteULike CiteULike   Add to Complore Complore   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us   Add to Digg Digg   Add to Reddit Reddit   Add to Technorati Technorati    What's this?


This article has been cited by other articles:


Home page
 Genes Dev.Home page
K. J. Briggs, I. M. Corcoran-Schwartz, W. Zhang, T. Harcke, W. L. Devereux, S. B. Baylin, C. G. Eberhart, and D. N. Watkins
Cooperation between the Hic1 and Ptch1 tumor suppressors in medulloblastoma
Genes & Dev., March 15, 2008; 22(6): 770 - 785.
[Abstract] [Full Text] [PDF]

Home page
 Mol. Cell. Biol.Home page
N. Stankovic-Valentin, S. Deltour, J. Seeler, S. Pinte, G. Vergoten, C. Guerardel, A. Dejean, and D. Leprince
An Acetylation/Deacetylation-SUMOylation Switch through a Phylogenetically Conserved {psi}KXEP Motif in the Tumor Suppressor HIC1 Regulates Transcriptional Repression Activity
Mol. Cell. Biol., April 1, 2007; 27(7): 2661 - 2675.
[Abstract] [Full Text] [PDF]

Home page
 CarcinogenesisHome page
C. Hellerbrand, E. Bumes, F. Bataille, W. Dietmaier, R. Massoumi, and A. K. Bosserhoff
Reduced expression of CYLD in human colon and hepatocellular carcinomas
Carcinogenesis, January 1, 2007; 28(1): 21 - 27.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Res.Home page
D. Arlt, W. Huber, U. Liebel, C. Schmidt, M. Majety, M. Sauermann, H. Rosenfelder, S. Bechtel, A. Mehrle, D. Bannasch, et al.
Functional Profiling: From Microarrays via Cell-Based Assays to Novel Tumor Relevant Modulators of the Cell Cycle
Cancer Res., September 1, 2005; 65(17): 7733 - 7742.
[Abstract] [Full Text] [PDF]

Home page
 Nucleic Acids ResHome page
T. Morinaga, A. Enomoto, Y. Shimono, F. Hirose, N. Fukuda, A. Dambara, M. Jijiwa, K. Kawai, K. Hashimoto, M. Ichihara, et al.
GDNF-inducible zinc finger protein 1 is a sequence-specific transcriptional repressor that binds to the HOXA10 gene regulatory region
Nucleic Acids Res., July 26, 2005; 33(13): 4191 - 4201.
[Abstract] [Full Text] [PDF]

Home page
 J. Cell Sci.Home page
K. F. Kelly, A. A. Otchere, M. Graham, and J. M. Daniel
Nuclear import of the BTB/POZ transcriptional regulator Kaiso
J. Cell Sci., December 1, 2004; 117(25): 6143 - 6152.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
279/37/38313    most recent
M401610200v1
Right arrow Submit a Letter to Editor
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowRequest Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Pinte, S.
Right arrow Articles by Leprince, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Pinte, S.
Right arrow Articles by Leprince, D.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
 All ASBMB Journals   Molecular and Cellular Proteomics 
 Journal of Lipid Research   ASBMB Today 
Copyright © 2004 by the American Society for Biochemistry and Molecular Biology.
Advertisement
spacer
Advertisement
Advertisement