| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Laboratoire de Génétique, UMR 5641 CNRS, UCBL1, and
* Laboratoire de Biologie de la Reproduction et du Développement, UCBL1, 69373 Lyon cedex 08, France
1Correspondence: Laboratoire de Génétique, UMR 5641 CNRS, Domaine Rockefeller, UCBL1, 8 avenue Rockefeller, 69373 Lyon cedex 08, France. E-mail: dante{at}cismsun.univ-lyon1.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: DNA methylation • oocytes • spermatozoa • breast cancer
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. It is likely that BRCA1
acts as a tumor suppressor gene (2)
. Indeed, in breast
cancers linked to BRCA1, as expected for a tumor suppressor
gene, allelic deletions at this locus invariably involve the wild-type
allele (3
, 4)
. BRCA1 involvement in breast
cancers does not seem to be restricted to familial cancers. Despite the
absence of somatic mutation in the breast tissues, a down-regulation of
BRCA1 expression is associated with malignancy in human
sporadic breast cancers (4
5
6)
.
Although its biological function is still unknown, BRCA1 may
have an important role in cellular differentiation and proliferation.
In transgenic mice, homozygous disruption of the Brca1 gene
results in embryonic lethality (7)
. The progressive
changes in Brca1 expression during mouse embryogenesis
(8)
also imply a role for Brca1 in the
differentiation process. In addition, variations of Brca1
expression are observed during postnatal mammary gland development
(9
, 10)
. More recently, it had been shown that
Cre-mediated invalidation of this gene affects final differentiation of
the gland during gestation (11)
. In humans, the
up-regulation of BRCA1 gene expression observed during the
first stages of prenatal development of the mammary gland also suggests
a role for BRCA1 in the differentiation of the mammary gland
(12)
. Taken together, these data indicate that variations
of BRCA1 expression may have some physiological consequences
in human breast tissue.
Analysis of the genomic region containing this gene indicates that an
another gene, NBR2, lies head to head with the
BRCA1 gene (13)
. Site-directed deletion
mutagenesis experiments led to the identification of a bidirectional
promoter region, position -258 to +43 (14)
, and a minimal
positive regulatory region has been mapped at position -198 to -162
(15)
. Structural studies have shown that a minor part of
breast cancers (~10%) exhibits hypermethylated sites in the region
containing the putative BRCA1 promoter
(16
17
18)
. However, a down-regulation of BRCA1
expression has been observed in most sporadic breast cancers
(4
5
6)
, indicating that abnormal methylation of this 5'
end region of BRCA1 does not account for the decrease in
BRCA1 expression in most cases of sporadic breast cancers.
In addition, in a small series (37 cases) of sporadic breast cancers,
the down-regulation of BRCA1 was not correlated with DNA
hypermethylation in the vicinity of the promoter region
(5)
.
Although transient transfection assays indicate that essential
regulatory sites are not present in the upstream region (position
-1528 to -202; ref 15
), it had been shown for other
promoters that methylation of the surrounding sequences may repress
gene expression (19)
. The efficiency of the inhibition
seems to be dependent on CpG density and promoter strength
(19)
.
We have therefore determined the methylation status of the 5' end of BRCA1 in a variety of cell lines and tissues, including fetal and cancer breast tissues. This analysis led to an unexpected finding, since the 5' end of BRCA1, which is embedded in a large CpG island, appears to be regionally methylated in all somatic tissues analyzed, suggesting that this region may participate in the regulation of BRCA1 gene expression.
Differences in DNA methylation are associated with differentiation and
carcinogenesis, CpG methylation correlating with the silencing of many
genes (20
, 21)
. Two types of mechanism could be involved
in silencing genes by DNA methylation. CpG methylation can
down-regulate gene expression by preventing the binding of
transcription factors to their recognition sequences or through
repressor molecules that bind to methylated DNA (22
, 23)
.
The methyl-CpG binding proteins (MBD), a family of vertebrate proteins,
bind to methylated DNA in any sequence context (22)
; for
some members of this family it has been shown that the binding of such
proteins represses gene expression at a distance (24)
.
Among the DNA binding proteins potentially involved in the negative
regulation of gene expression, the methyl-CpG binding protein MeCP2
seems to play an important role (25
, 26)
. MeCP2 is a
chromosomal protein that binds specifically to methylated CpG and
represses densely methylated genes in association with a histone
deacetylase complex (27
, 28)
. Although inhibition of
histone deacetylase by trichostatin A can relieve the transcriptional
repression mediated by MBD (27
, 28)
, for some genes it has
been shown that dense CpG island methylation might be a dominant factor
in gene silencing (29)
.
We investigated, therefore, whether methyl-CpG binding proteins such as MeCP2 might contribute to the control of BRCA1 expression.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Human oocytes that had failed to fertilize 3 days after in vitro insemination were collected from the In Vitro Fertilization Laboratory (Hospital E. Herriot, Lyon, France). To remove the follicular cells linked to the zonae pellucida, oocytes were treated by enzymatic digestion with hyaluronidase (150 units, type VIII hyaluronidase, Sigma, France) to discard contaminating somatic cells. Then a digestion by trypsin was performed to remove the zonae pellucida and the remaining somatic cells. Oocytes were rinsed several times in PBS 1x and stored in liquid nitrogen until use. The cumulus cells were collected after hyaluronidase digestion for further studies.
Cell culture
Human breast cell lines (MCF7, BT20, and HBL 100), cervix cell
line (HeLa), or kidney cell line (Bosc 23) were obtained from ATCC
(Rockville, Md.) and grown in Dulbecco’s modified Eagle’s medium
(Sigma, L’isle d’Abeau, France) supplemented with FCS 5% and 0.5
µg/ml insulin for MCF7 and BT 20 cell lines. All cells were grown at
37°C in a humidified 5% CO2 atmosphere.
PCR-based methylation assay
DNA extracted from tissue samples, cell lines, and cells treated
with 5azadC were digested with a fivefold excess of RsaI or
RsaI plus HpaII or CfoI and incubated
overnight at 37°C in the appropriate buffer (Roche Diagnostics,
Meylan, France). Control experiments were performed using the
methyl-insensitive enzyme MspI. Enzymes were inactivated by
heating at 65°C for 1 h and an aliquot of the reaction was used
for polymerase chain reaction (PCR) amplification. The PCR
amplification was performed in the following conditions: 10 mM Tris-HCl
(pH 8.3), 3 mM MgCl2, 50 mM KCl, 0.1 mg/ml
gelatin, 4% DMSO, 100 µM of each of the four deoxyribonucleoside
triphosphates, 0.25 µM of the primers (forward: 5' TTG GGA GGG GGC
TCG GGC AT 3'; reverse: 5' CAG AGC TGG CAG CGG ACG GT), and 0.6 units
of Taq DNA polymerase (Roche Diagnostics) after 35 cycles in
an Eppendorf thermocycler (1 min denaturation at 94°C, 2 min
annealing at 55°C, and 3 min extension at 72°C). In each
experiment, the sample digested with RsaI, which does not
cleave the sequence between the two primers, was amplified to verify
the efficiency of the amplification. The sequence analyzed contains 9
HpaII sites and 10 CfoI sites and PCR
amplification occurs only when the sites are methylated and uncut.
5-aza-2'-deoxycytidine treatments
For 5-aza-2'-deoxycytidine (5azadC, Sigma, France) treatments,
cell lines were seeded at low density (3–4x105
cells/100 mm dish) 16 h before treatment with a final
concentration of 1 µM 5azadC. The medium was changed 24 h after
drug addition and every subsequent day. RNA and DNA were isolated after
72 h. DNAs were extracted and quantitated (Hoechst staining
method) from treated cells and control cells. Different aliquots were
digested by RsaI (in order to normalize the size of the DNA
molecules) plus: HpaII, CfoI
(methylation-sensitive enzymes), MspI
(methylation-insensitive), and no enzyme. The 5' end of the
BRCA1 gene was then amplified by PCR from these samples and
the signals obtained were quantitated by densitometry. No signal was
observed after MspI digestion. The extent of demethylation
was evaluated by the ratio between the signal obtained after digestion
by HpaII, CfoI, or both and the signal was
obtained from DNAs digested with RsaI alone. The data
indicated that > 90% of the DNA molecules contained unmethylated
CfoI and HpaII sites after 5azadC treatment. Exon
1, which was found unmethylated in normal and tumoral tissues, was
amplified as a control experiment of complete digestion with
methylation-sensitive enzymes.
Transfections
pCMV-MeCP2-HA (kindly provided by Dr. A. Bird), pCMV-ßgal
plasmids, or pGL3-control plasmid (Promega, Lyon, France) were
transfected using the calcium phosphate precipitation technique
(30)
. Cells were collected 48 h after transfection.
For the immunofluorescence assay, Bosc 23 cells were grown on Lab-tek
Permanox (Nunc) 2-well chambered coverslips at 1 x
105 cells per well and transfected as described
previously. After 48 h, the cells were washed twice in PBS 1x,
fixed in paraformaldehyde 4% for 15 min at room temperature, and
rinsed several times with PBS 1x. Then cells were permeabilized in 0.1
M glycine-PBS 1x buffer, followed by incubation in 0.5% Triton
X-100-PBS 1x buffer, and blocked in 0.2% gelatin-PBS 1x buffer for
15 min at room temperature. The anti-tag HA monoclonal antibody (12
CA5, Roche Diagnostics) diluted 1:40 in 0.5% Triton X-100-PBS 1x
buffer was incubated with the cells for 1 h. The cells were rinsed
several times with PBS 1x and incubated for 1 h with a 1:200
dilution of the FITC-conjugated secondary antibody (goat anti-mouse
IgG, Dako, France). Fluorescence was visualized with a 25x or 40x
immersion lens on a Leica microscope. In these experiments, MeCP2
expression was consistently observed in more than 50% of the cells and
ß-galactosidase activity (30)
was observed in more than
60% of the cells; five adjacent fields, at magnification 40x, were
counted.
pGL3 constructs, methylation, and luciferase assay
A 1757 bp fragment of the BRCA1 gene containing a
part of the 5' CpG island and 43 bp of exon 1a was isolated from
genomic DNA by PCR using primers containing KpnI site at the
5' end and BglII site at the 3' end (primer
forward-KpnI: 5' CTG GTA CCT TGG GAG GGG GCT CGG
GCA A3'; primer reverse-BglII: 5' GAA GAT CTT CCA
GGA AGT CTC AGC GAG C 3'). The PCR products were ligated between
KpnI and BglII sites into the pGL3 basic vector
(Promega) and transformed into competent JM-109 Escherichia
coli cells (Promega).
In vitro methylation was performed by incubating the p5'-BRCA1-Luc vector with one unit of either HpaII (CCGG sites), HhaI (GCGC sites), or SssI (CG sites) methylases per microgram of plasmid DNA in the conditions recommended by the manufacturer (Biolabs, Beverly, Mass.); completeness of the modification was checked with the corresponding restriction enzyme. p5'-BRCA1-Luc constructs were transfected into Bosc 23 cells as described previously. Transfections were optimized for 12-well plates. The cells were lysed and assayed for luciferase expression (Luciferase assay, Promega) after 48 h and the light emissions were measured in a scintillation counter (Packard, Downers Grove, Ill.). A plasmid (pGL3-control; Promega) containing a promoter and an enhancer derived from SV40 was also transfected in the same plates; values obtained for this control vector were comparable between each experiments.
DNA extraction
High molecular weight DNA was extracted from frozen pulverized
tissue samples and cells by standard procedures (30)
.
Briefly, samples were resuspended in 10 mM Tris-0.1M EDTA buffer and
digested with proteinase K (300 µg/ml final concentration) in the
presence of SDS. A similar method with the addition of 0.001% (V/V)
ß-2 mercaptoethanol was used to prepare decondensed DNA from
spermatozoa. When DNA was extracted from a small number of cells (6 to
10 oocytes), 2 µg of pGEM-T plasmid (Promega) was added as carrier.
The mixture was incubated at 37°C, phenol/chloroform extracted, and
DNA was ethanol-precipitated.
RNA extraction
RNA was isolated in a single step procedure by
acid-guanidium-thyocyanate-phenol-chloroform extraction as described
previously (31)
. After extraction, total RNA was
precipitated in isopropanol and resuspended in an appropriate volume of
sterile water. The integrity and quantity of RNA were examined by gel
electrophoresis. Total RNA was quantified by densitometry on a 1.2%
agarose gel containing 0.1 µg/ml ethidium bromide in comparison with
serial dilutions of known amount of standard RNA (Roche Diagnostics).
Quantitative reverse transcription PCR (RT-PCR)
The RT-PCR assay was performed as described previously
(32)
by coamplification of 0.3 µg of total RNA and a
known amount of competitor RNA. Primers used were designed to amplify
the cDNA fragment from exon 6 to 8 of the BRCA1 gene (C3F:
5' TGT GCT TTT CAG CTT GAC ACA GG 3' and C3R: 5' CGT CTT TTG AGG TTG
TAT CCG CTG 3'). Reactions were performed in 100 µl containing 10 mM
Tris-HCl (pH 8.3), 3 mM MgCl2, 50 mM KCl, 0.1
mg/ml gelatin, 200 µM of each of the 4 deoxyribonucleoside
triphosphates, and 0.25 µM of the C3 primers.
After initial denaturation at 92°C for 2 min, six units of Expand Reverse Transcriptase (Roche Diagnostics) were added to the reaction mixture and incubated for 35 min at 42[de]C. Reverse transcriptase was then inactivated by heating and after cooling to 0°C, PCR amplification was accomplished by adding 0.6 units of Taq DNA polymerase (Roche Diagnostics) after 35 cycles in an Eppendorf thermocycler (1 min denaturation at 94°C, 2 min annealing at 55°C, and 3 min extension at 72°C). Aliquots were analyzed on a 2% agarose gel containing 0.1 µg/ml ethidium bromide and the intensity of the bands corresponding to the wild-type BRCA1 PCR products and to the competitor PCR products was determined using Image Analyzer Software (Wayne Rasband, National Institutes of Health).
Southern blot analysis
Methylation patterns were determined by Southern blot
experiments using methylation-sensitive restriction endonuclease
CfoI (GCGC site) and HpaII (CCGG site) or
methylation-insensitive isoschizomer MspI (CCGG site). In a
typical experiment, 30 µg of total DNA was cleaved with a 10-fold
excess of TaqI or PstI for 10 to 12 h in the
appropriate conditions (Roche Diagnostics). Two-thirds of the sample
was then digested overnight by CfoI or HpaII or
MspI. Restriction endonuclease products were separated by
electrophoresis on a 1.2% agarose gel and transferred to Hybond
N+ Nylon membrane (Amersham, France). After
drying, the membranes were hybridized overnight at 65°C to a randomly
primed 32P-labeled probe (Random Primed DNA
labeling kit, Roche Diagnostics) in the hybridization solution (0.25 M
sodium phosphate, 7% SDS, 1% BSA, 50 µg/ml yeast tRNA). Membranes
were washed with increasing stringency (from 2 to 0.1 x SSC;
0.1% SDS) at 65°C and exposed to Hyperfilm (Amersham, France) for 1
to 7 days at -70°C. Membranes could be used for rehybridization
after two successive treatments with SDS 1% at 80°C for 30 min.
Gene probes
The probe used for Southern blot analysis of the CpG island of
the BRCA1 gene was synthesized by PCR amplification of a 239
bp region (-1244 to -1005; ref 33
). The PCR fragments
were cloned into a pGEM-T vector (Promega). After digestion with the
appropriate enzyme, the insert was purified by agarose gel
electrophoresis, followed by electroelution. Each filter was
rehybridized with the exon 1 probe encompassing exon 1a of the
BRCA1 gene and the bidirectional promoter (5)
in order to verify that DNA cleavage with restriction endonucleases was
complete.
Bisulfite modification
The sodium bisulfite modification method, followed by the
sequencing of PCR products, is used to determine the CpG methylation
pattern. Sodium bisulfite converts unmethylated cytosines to uracils
whereas the methylated cytosines remain unmodified. In the resultant
modified DNA, uracils are replicated as thymines during PCR
amplification. The sodium bisulfite reaction was carried out on 4 µg
of DNA (3 µg of carrier DNA and 1 µg of human genomic DNA).
Alkali-denatured DNA was incubated in 3 M NaHSO3
and 5 mM hydroquinone for 16 h at 50°C. Modified DNA was
purified using the Wizard DNA Clean-up System (Promega) and eluted into
50 µl of sterile water. Modification was completed by 0.3 M NaOH; DNA
was precipitated by 0.5 M ammonium acetate (pH 4.6) and resuspended in
water.
DNA was amplified using strand-specific primers designed to amplify a 258 bp region in the CpG island of the BRCA1 gene in two separate reaction mixtures. The first round of PCR amplification was accomplished in 100 µl in a buffer containing 10 mM Tris-HCl (pH 8.3), 3 mM MgCl2, 50 mM KCl, 0.1 mg/ml gelatin, 100 µM of each of the four deoxyribonucleoside triphosphates, 0.25 µM of the primers (forward: 5' TTT TGT TTT GTG TAG GGC GGT T 3'; Reverse: 5' CCT TAA CGT CCA TTC TAA CCG T 3'), and 0.6 units of Taq DNA polymerase (Roche Diagnostics) after 35 cycles in an Eppendorf thermocycler (1 min denaturation at 94°C, 2 min annealing at 55°C, and 3 min extension at 72°C). An aliquot of the first amplification was reamplified with internal primers (forward: 5' TGA GAA TTT AAG TGG GGT GT 3'; reverse: 5' AAC CCT TCA ACC CAC CAC TAC 3') in the same conditions.
PCR products were first analyzed by digestion with restriction enzymes DdeI, EcoRI (Roche Diagnostics), and HphI (Biolabs) in the buffers recommended by the manufacturers.
Then PCR products were cloned in a pGEM-T vector (Promega) and 10 random clones were analyzed by automatic sequencing (Eurogentec, Belgium) to determine the proportion of methylated (CpG) or unmethylated (TpG) sites.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and 30-fold in sporadic breast cancers
(5)
, in the amount of BRCA1 mRNA are not
related to the methylation status of the putative promoter sequence
located at the 5' end of the exon 1a (5)
. However, this
unmethylated region (nt -167 to nt +484) is included in a large CpG
island, spanning nucleotides -2200 to +500 (33 and Fig. 1E
), which suggests that hypermethylation of this CpG island
could be associated with gene silencing. Analysis of the methylation
status of the -1714 to -1023 region using a PCR-based methylation
assay indicated that this CpG island was methylated at CfoI
and HpaII sites in human somatic tissues (including normal
and tumoral somatic tissues and fetal tissues, 65 samples analyzed) and
cell lines (data not shown). Since this assay is qualitative rather
than quantitative, methylation patterns were further determined by
Southern blot experiments (representative samples are shown in Fig. 1A
).
|
DNAs were first digested with the methylation-insensitive restriction
endonuclease PstI. Using probe 1 (position -1244 to -1005,
Fig. 1C
), PstI generates two bands of 3.8 kb for
the BRCA1 gene and 5 kb for the BRCA1 pseudogene
(Fig. 1A
, lanes 1 and 9). When DNA extracted from the HBL
100 cell line is digested with the methylation-insensitive enzyme
MspI, this probe reveals a 1008 bp band corresponding to the
BRCA1 gene (Fig. 1A
, lane 2). A faint 202 bp band
corresponding to the BRCA1 pseudogene was detected at the
bottom of the gel. DNAs extracted from normal tissues (Fig. 1A
) were digested with the methyl-sensitive enzyme
HpaII that cleaves CCGG sites when the internal cytosine at
CpG dinucleotide is unmethylated. Probe 1 maps a 1449 bp band (Fig. 1A
, lanes 4–7) corresponding to the methylation of 9 CCGG
sites located within the 5' region of the BRCA1 gene (Fig. 1B
and 1C). In spermatozoa, the CCGG sites analyzed using
probe 1 are unmethylated (Fig; 1A, lane 8). DNA from the HBL 100 cell
line and fetal breast tissue (Fig. 1A
, lanes 10 and 11,
respectively) were cleaved with the methylation-sensitive enzyme
CfoI in order to map several other CpG sites. Probe 1
reveals a 924 bp band in the HBL 100 cell line and reveals a 736 bp
band in tissues, indicating that the proximal CfoI sites are
unmethylated in normal tissues. In spermatozoa, CpG dinucleotides
located within GCGC sites are unmethylated, giving rise to a 455 bp
band after hybridization with probe 1.
Taken together, these data indicate that DNAs extracted from somatic
tissues and cell lines were methylated at CCGG and GCGC sites within
the region spanning nucleotides -2000 to -1000 or -720 for the HBL
100 cell line, with the exception of human spermatozoa DNA, which was
unmethylated at these sites (Fig. 1A
).
Analysis of the methylation pattern of the body of BRCA1
gene was carried out, as described previously (5)
, on
several fetal and adult human tissues, cell lines, and male germ cells
by Southern blot after digestion by the methylation-sensitive enzymes
HpaII and CfoI (Fig. 1D
). The filters
were successively hybridized with several probes corresponding to the
different exons present in the 3' region of the BRCA1 gene.
The body of the gene (from the end of exon 1b to the end of exon 24) is
methylated in somatic tissues (5)
and mature male germ
cells, as expected for non-CpG island regions. In somatic cells, the
promoter region of the gene is unmethylated, as described previously
(5)
. In cell lines and tissues, the 5' region is
methylated at CCGG and GCGC sites in the various samples. On the other
hand, this region is unmethylated in spermatozoa as expected for a CpG
island.
The 5' CpG island of BRCA1 is unmethylated in human
gametes
The gamete-specific pattern was further confirmed by scaling-down
the bisulfite sequencing method in order to investigate CpG methylation
from very small amounts of DNA, and experimental conditions were chosen
for obtaining a quantitative assay of the ratio methylated CpG vs.
unmethylated CpG.
DNA extracted from purified human spermatozoa was modified by sodium
bisulfite (34)
. This chemical treatment converts
unmethylated cytosines to uracils while methylated cytosines remain
unmodified. After modification, DNA was amplified by a two step PCR
method, as described in Materials and Methods. The PCR product was
digested by specific restriction endonucleases to determine the global
methylation status of the sample. Completeness of the modification was
monitored by digestion with DdeI, which cleaves only
unconverted DNA. PCR products obtained from methylated molecules
exhibit a new EcoRI site at position 138, whereas
unmethylated molecules exhibit a new HphI site at position
165.
The sensitivity of PCR amplification after bisulfite modification was
monitored by mixing different proportions of unmethylated DNA from
spermatozoa (from 25 to 100%) and methylated DNA from HBL 100 (from 0
to 75%). For each assay, an aliquot of the PCR product was cleaved
with DdeI (unmodified DNA), EcoRI (methylated
DNA), or HphI (unmethylated DNA), loaded on a 2% agarose
gel, and visualized by ethidium bromide staining. The results indicate
that the amount of PCR product cleaved by enzymatic digestion is
directly related to the percentage of methylated or unmethylated DNA
used in the coamplification assay (Fig. 2A
). DNAs from somatic tissues and gametes were therefore
modified using this method and PCR products were cloned and sequenced.
|
Within the region analyzed, -1643 to -1358, all the 24 CpG sites were
unmethylated in DNA from human oocytes and spermatozoa (Fig. 2B
). As expected from the Southern blot experiments, all
these CpG sites were methylated in all somatic tissues and cell lines,
including the somatic cells of the corona radiata surrounding the
oocytes (Fig. 2B
). The absence of DNA methylation within the
CpG island in human gametes did not extend to the body of the
BRCA1 gene, since control experiments indicated that two
regions of the exon 11 are methylated both in somatic tissues and
gametes (data not shown), suggesting that the methylation of the CpG
island might play regulatory role in BRCA1 expression.
Chemically induced hypomethylation elevates BRCA1
expression
To test the potential role of DNA methylation in the control of
BRCA1 expression, demethylation was induced by
5-aza-2'-deoxycytidine (5aza-dC) treatments. Human cell lines (three
breast cell lines: BT20, MCF7, and HBL 100; one cervix cell line, HeLa;
and one kidney cell line, Bosc 23) were grown for 72 h in the
presence of 1 µM of 5aza-dC. The PCR-based methylation assay
indicated that the 5' end of BRCA1 was efficiently
demethylated at the end of the treatment. Then the amounts of
BRCA1 mRNA were determined using a competitive RT-PCR method
(32)
.
5aza-dC treatment invariably elevated BRCA1 expression in
the cell lines analyzed by up to fivefold in BT20 cells and from 1.5 to
2.5 in the other cell lines (Table 1
), indicating that CpG methylation is associated with a low level of
BRCA1 expression.
|
In vitro methylation induces a MeCP2-dependent
repression of the BRCA1 promoter activity
The methylated domain of the BRCA1 CpG island is
relatively far (~0.7 to 1 kb, Fig. 1C, E
) from the start
site of the transcription. MeCP2 is able to drive a long-range
repression and might down-regulated BRCA1 expression. This
hypothesis was investigated using an expression vector
(p5'-BRCA1-Luc) containing the -1714 to +43 region of
BRCA1 fused to the luciferase enzyme as a reporter gene.
Analysis of the 5' end BRCA1 sequence indicates that the
CCGG sequences (HpaII sites) are located outside the minimal
promoter region required for full promoter activity and only one GCGC
sequence (HhaI sites) is located at the 3' end of this
region, deletion of this latter sequence inducing only a weak
inhibition of the promoter activity (15)
. Therefore, we
take this opportunity to investigate whether the methylation of
surrounding sequences might affect BRCA1 promoter activity.
p5'-BRCA1-Luc vector was in vitro methylated
using HpaII methylase or HhaI methylase and
transfected in Bosc 23 cells in the presence or absence of a vector
(MeCP2-Tag-HA) encoding for MeCP2. In the absence of the MeCP2-Tag-HA
vector, in vitro methylation induced only a two- to
threefold drop in luciferase activity; meanwhile, in the presence of
MeCP2, the in vitro methylated vectors exhibited an almost
total loss of transcriptional activity (Fig. 3
). Control experiments using expression vectors containing the
ß-galactosidase gene driven by a CMV-promoter (pCMV-ß-gal plasmid)
and a pGL3-control plasmid containing a SV40 promoter fused to the
luciferase gene indicated that the expression of these reporters genes
was not or was only minimally affected by the overexpression of the
pCMV-MeCP2-HA-vector.
|
Therefore, cotransfection experiments in the Bosc 23 cell line
indicated that partial methylation of the vector containing a part of
the 5' end of BRCA1, which mimics the methylation pattern
observed in somatic cells, induced a MeCP2-dependent repression of the
transcriptional activity of this vector (Fig. 3)
.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
).
Although these sequences are usually unmethylated, hypermethylation of
CpG islands associated with gene silencing has been described for
crucial regulator of growth during cancer progression
(20)
. In addition, abnormal methylation at multiple
CpG-rich sequences and gene silencing are also characteristic of human
and murine cell lines (35)
.
In normal somatic tissues, methylated CpG islands have been found for X
chromosome-inactivated genes and for imprinted genes, suggesting that
the methylation of a subset of CpG islands also occurs during normal
physiological process. In line with this observation, it has been shown
in colon tissue that DNA methylation at CpG islands takes place during
the aging process (36)
.
In this study we have found that the CpG island of BRCA1 is regionally methylated in all somatic human cells and tissues analyzed, suggesting that this epigenetic modification might have some consequences on BRCA1 gene expression. Since some promoters are not or are only minimally inhibited by in vitro methylation, we have investigated the methylation sensitivity of the p-5'-BRCA1-Luc expression vector (containing the 5' end of BRCA1). In vitro methylation using M-HpaII generates a methylated site at position -167 that is not methylated in vivo, whereas the other nine sites are methylated in the cell lines and tissues analyzed.
Several papers have described a bidirectional promoter in this region.
For example, Thakur and Croce (15)
have localized the
promoter region at position -195 to -162 and a minimal positive
regulatory region at position -195 to -177. This short region seems
to be involved in the binding of the nuclear proteins. More recently,
Suen and Gross (37)
also mapped the bidirectional
BRCA1 promoter in this region at position -204 to -149,
and band-shift assays indicated that within this promoter the -167 to
-149 region was the target of nuclear proteins. These data might
suggest that the nucleotide ‘C’ at position -167 is not involved in
the binding of potential transcription factors.
In addition, the methylation of this site (-167) does not seem to be crucial for the expression of p-5'-BRCA1-Luc vector since the methylation of the HhaI sites (the more proximal sites are at positions -564 and -20) has a more pronounced inhibitory effect. Furthermore, the methylation of all CpGs induces (using SssI methylase) a very strong inhibition of expression of the p-5'-BRCA1-luc vector, even in the absence of pCMV-MeCP2-HA vector, suggesting that the density of methylated sites is the main point of the inhibitory effect.
Although MeCP2 can bind DNA segments containing only one methylated
CpG, it had been shown that the inhibitory effect of MeCP2 is strongly
dependent on the density of methyl-CpGs (38)
. Using
p-5'-BRCA1-Luc vector cotransfected with a MeCP2 expression
vector, in our experiments the inhibition seems dependent on the number
of methylated CpGs, since the inhibition observed with HhaI
methylase (16 sites) is stronger than the inhibition obtained after
methylation with HpaII methylase (10 sites). These data also
suggest that the inhibition of the expression of the p-5'-BRCA1-Luc
vector by DNA methylation is not the result of the methylation of a
specific site, but seems to be dependent on the methylation density.
The regional methylation of the 5' end BRCA1 CpG island is
clearly not the consequence of a pathological process, since this
pattern was observed in all normal somatic tissue samples analyzed. In
contrast to the imprinted genes, this region is unmethylated in human
spermatozoa and in oocytes. Furthermore, in fetal and adult tissues
differential methylation between alleles was not detected. The
presence, at the 5' end of BRCA1, of the complete long
terminal repeat element pTR5, spanning nucleotides -3123 to -1273
(33)
, which is partially included in the
BRCA1-CpG island (nt -2200 to +500), might account for the
partial methylation of this region. In addition, the 5' end of
BRCA1 exhibits a complex organization since the
transcription start site of BRCA1 is separated from that of
the NBR2 gene by 218 bp and, therefore, its 5' end region in
included in the intron 1 of NBR2 gene (13)
. In
line with this hypothesis, the promoter region of the mouse
Brca1 gene, which lacks this repetitive element and complex
organization (39)
, appears to be unmethylated in mouse
somatic tissues and in the NIH 3T3 cell line (unpublished data),
suggesting that insertion of the LTR element in somatic cells might
have some consequences on the methylation patterns of the human
BRCA1 gene.
This methylation pattern of the BRCA1 CpG island confers potential regulatory features. Our results show that BRCA1 expression is regulated at least in part by methyl-CpG binding proteins. Among this family, MeCP2 might be a good candidate since transient expression of the corresponding gene led to a methylation-dependent inhibition of BRCA1 expression. Despite the low level of MeCP2 expression (unpublished data) in the cell lines analyzed, inhibition of DNA methylation resulted in an elevated level of BRCA1 mRNA. This chemically induced hypomethylation by 5aza-dC might suggest that other members of the methyl-DNA binding proteins family are also involved in this regulation.
Transient transfection of fusion proteins have identified a MeCP2
domain that is capable of long-range repression of vectors (19
, 24)
. This protein, which binds to methylated CpG, forms a
complex with histone deacetylase and the transcription repressor Sin3A,
leading to formation of transcriptionally repressive chromatin
architecture (27
, 28)
. MBD2 and MBD3, two other methyl-DNA
binding proteins, have recently been characterized and shown to
participate in other histone deacetylase complexes and in gene
silencing mechanisms (40
, 41)
. The expression patterns of
these methyl-DNA binding protein genes are not fully determined, but it
has been shown in the rat that the level of MeCP2 expression
is dependent on the tissues (42)
, MBD3 is associated with
metastatic-associated protein 1, a protein that is overexpressed in
several human cancers (43)
, and MBD2 belongs to the MeCP1
deacetylase complex, which is also present at various levels depending
on the cell type and the differentiation state (40)
.
Although, genes possessing a 5' end methylated CpG island should represent only a very small part of the mammalian genomes, differences in the amount of methyl-DNA binding proteins between tissues or cells could represent an additional mechanism for tissue-specific gene expression.
|
ACKNOWLEDGMENTS |
|---|
Received for publication September 7, 1999.
Revision received March 2, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
hänen, S., Herman, J. G., Baylin, S. B. (1999) Synergy of demethylation and histone deacetylase inhibition in the re-expression of genes silenced in cancer. Nat. Genet. 21,103-107[Medline]
This article has been cited by other articles:
|
|
 |
|
  C. Jordan and U. Francke Ube3a expression is not altered in Mecp2 mutant mice Hum. Mol. Genet., July 15, 2006; 15(14): 2210 - 2215. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Wei, T. A. Grushko, J. Dignam, F. Hagos, R. Nanda, L. Sveen, J. Xu, J. Fackenthal, M. Tretiakova, S. Das, et al. BRCA1 Promoter Methylation in Sporadic Breast Cancer Is Associated with Reduced BRCA1 Copy Number and Chromosome 17 Aneusomy Cancer Res., December 1, 2005; 65(23): 10692 - 10699. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Auriol, L.-M. Billard, F. Magdinier, and R. Dante Specific binding of the methyl binding domain protein 2 at the BRCA1-NBR2 locus Nucleic Acids Res., July 28, 2005; 33(13): 4243 - 4254. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Thakur, T. Nakamura, G. Calin, A. Russo, J. F. Tamburrino, M. Shimizu, G. Baldassarre, S. Battista, A. Fusco, R. P. Wassell, et al. Regulation of BRCA1 Transcription by Specific Single-Stranded DNA Binding Factors Mol. Cell. Biol., June 1, 2003; 23(11): 3774 - 3787. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. H. WEI, R. BROWN, and T. H.-M. HUANG Aberrant DNA Methylation in Ovarian Cancer: Is There an Epigenetic Predisposition to Drug Response? Ann. N.Y. Acad. Sci., March 1, 2003; 983(1): 243 - 250. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. El-Osta, M. Lubbert, P. W. Wijermans, T. Licht, and P. A. Jones On the use of DNA methylation inhibitors and the reversal of transcriptional silencing Blood, February 15, 2003; 101(4): 1656 - 1657. [Full Text] [PDF]
|
|
|
|
 |
|
 J. L. Hilton, J. P. Geisler, J. A. Rathe, M. A. Hattermann-Zogg, B. DeYoung, and R. E. Buller Inactivation of BRCA1 and BRCA2 in Ovarian Cancer J Natl Cancer Inst, September 18, 2002; 94(18): 1396 - 1406. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Roder, M.-J. Latasa, and H. S. Sul Silencing of the Mouse H-rev107 Gene Encoding a Class II Tumor Suppressor by CpG Methylation J. Biol. Chem., August 16, 2002; 277(34): 30543 - 30550. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Magdinier, S. G. d'Estaing, C. Peinado, B. Demirci, C. Berthet, J. F. Guerin, and R. Dante Epigenetic marks at BRCA1 and p53 coding sequences in early human embryogenesis Mol. Hum. Reprod., July 1, 2002; 8(7): 630 - 635. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Ballestar and M. Esteller The impact of chromatin in human cancer: linking DNA methylation to gene silencing Carcinogenesis, July 1, 2002; 23(7): 1103 - 1109. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. D. Braastad, M. Leguia, and E. A. Hendrickson Ku86 autoantigen related protein-1 transcription initiates from a CpG island and is induced by p53 through a nearby p53 response element Nucleic Acids Res., April 15, 2002; 30(8): 1713 - 1724. [Abstract] [Full Text] [PDF]
|
|
|
|
