|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 277, Issue 38, 35434-35439, September 20, 2002
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the
Received for publication, April 10, 2002, and in revised form, July 15, 2002
Although mammalian MBD3 contains the mCpG-binding
domain (MBD) and is highly homologous with the authentic mCpG-binding
protein MBD2, it was reported that the protein does not bind to mCpG
specifically. Using recombinant human wild type and mutant MBD3
proteins, we demonstrated that atypical amino acids found in MBD3 MBD,
namely, His-30 and Phe-34, are responsible for the inability of
MBD3 to bind to mCpG. Interestingly, although H30K/F34Y MBD3
mutant protein binds to mCpG efficiently in vitro, it was
not localized at the mCpG-rich pericentromeric regions in mouse cells.
We also showed that Y34F MBD2b MBD, which possesses not the
mCpG-specific DNA-binding activity but the nonspecific DNA-binding
activity, was localized at the pericentromeric regions.
These results suggested that the mCpG-specific DNA-binding activity is
largely dispensable, and another factor(s) is required for the
localization of MBD proteins in vivo. MBD3 was identified
as a component of the NuRD/Mi2 complex that shows chromatin
remodeling and histone deacetylase activities. We demonstrated
that MBD3 MBD is necessary and sufficient for binding to HDAC1 and
MTA2, two components of the NuRD/Mi2 complex. It was therefore
suggested that mCpG-binding-defective MBD3 has evolutionarily conserved
its MBD because of the secondary role played by the MBD in
protein-protein interactions.
DNA methylation is the major modification in eukaryote genomes.
This modification occurs predominantly at position 5 of cytosine when
followed by guanosine (CpG site) in vertebrates. Approximately 60-90%
of the total CpG sites is methylated in vertebrates. One of the direct
consequences of mCpG is the transcriptional repression of nearby genes.
Several molecular mechanisms are thought to be responsible for this
methylcytosine-mediated gene repression. Among them, the repression
mediated by mCpG-binding proteins has been most extensively studied.
Five mCpG-binding proteins,
MBD1-41 and MeCP2, have been
identified in mammals and are collectively called MBD family proteins
because these proteins share the mCpG-binding domain (1, 2). MBD4
recognizes the T/G mismatched base pair, and is a component of the DNA
repair machinery (3, 4). MBD1, MBD2, and MeCP2 have been well
characterized for their gene-repressing activities. These proteins bind
to mCpG via MBD and recruit histone deacetylase (HDAC) complexes to the
binding sites via their separate domain called TRD (transcriptional
repression domain). The deacetylation by HDAC of the core histones H3
and H4 is partly responsible for the gene repression mediated by these
MBD proteins (1, 5-7). In contrast to these well characterized MBD
proteins, the function of MBD3 remains unclear. Although MBD2 and MBD3
are highly homologous in and outside MBD, experiments performed
in vitro and in vivo failed to demonstrate the
mCpG-binding activity of recombinant mouse MBD3 (2). A controversy
arose when Xenopus MBD3 was isolated and demonstrated to
possess mCpG-binding activity (8). However, recent studies have pointed
out that amino acids in MBDs critical for binding to mCpG are present
in Xenopus MBD3 but not in mammalian MBD3. These amino acid
substitutions are likely responsible for the inability of mammalian
MBD3 to bind to mCpG (9, 10).
Recently, it was reported that the NuRD/Mi2 complex contains MBD3 as a
component (8, 11). The complex consists of Rpd3-like histone
deacetylase, RbAp46/p48, MTA2, CHD4/Mi2beta, and MBD3 and possesses
DNA-dependent ATPase, chromatin remodeling activity, and
histone deacetylase activity (12). Thus, it has been proposed that the
NuRD/Mi2 complex plays a role in gene repression through histone
deacetylation and chromatin packaging in methyl-CpG-rich regions (13).
However, the role of MBD3 in the NuRD/Mi2 complex is again
controversial. In one study, it was shown that the human NuRD/Mi2
complex does not bind to mCpG, but is recruited to mCpG by interacting
with the authentic mCpG-binding protein MBD2 (11, 14). In another
study, the Xenopus NuRD/Mi2 complex and its component xMBD3
were shown to have specific mCpG-binding activity on their own (8).
In this study, we aimed at clarifying the mCpG-binding activity of
human MBD3. Specifically, we constructed mutant MBD3 proteins in which
atypical amino acids occupying the critical residues in other MBD
proteins were "corrected" to the conserved amino acids. We found
that such MBD3 mutants indeed showed specific binding activity to mCpG
in vitro as proposed previously. However, unexpectedly, such
mutants did not associate with murine pericentromeric regions that are
highly rich in mCpG. These results suggest that other factors are
necessary for the targeting of MBD proteins in vivo in
addition to the mCpG-binding activity.
Plasmids--
Mutant MBD3 cDNAs having point mutations were
generated by the site-directed mutagenesis kit (TaKaRa). Other
truncated cDNAs were generated by PCR. All cDNAs were confirmed
for nucleotide sequences. Recombinant proteins were expressed in
Escherichia coli using pGEX5x-1 (Amersham Biosciences) or
pET32c (Novagen) vectors. For the analysis of subcellular localization,
cDNAs encoding MBD2b and MBD3 tagged with the HA epitope at
their N termini were subcloned in pcDNA3 (Invitrogen), transfected,
and expressed in NIH 3T3 cells.
Recombinant Proteins--
GST-fused recombinant proteins were
expressed in E. coli XLIblue MRF' (Stratagene) and purified
using glutathione-Sepharose CL-4B beads (Amersham Biosciences).
His-tagged MTA2 and HDAC1 proteins were expressed in E. coli
BL21(DE3). The cells were lysed by sonication in buffer containing 500 mM NaCl, 10 mM Tris-Cl (pH 8.5), 0.5% Triton
X-100, and 10 mM 2-mercaptoethanol supplemented with the
complete protease inhibitor tablet (Roche). After centrifugation, the
supernatant was found not to contain soluble recombinant proteins. The
pellet was dissolved in 8 M urea, 50 mM Tris-Cl
(pH 8.5), and 0.5% Triton X-100. These solutions were dialyzed against
500 mM NaCl, 10 mM Tris-Cl (pH 8.5), 0.05%
Triton X-100, 10% glycerol, and 10 mM 2-mercaptoethanol.
After removing insoluble materials by ultracentrifugation, partially
purified MTA2 or HDAC1 was obtained.
GST Pull-Down Assay--
GST-fused MBD3 proteins were incubated
with glutathione-Sepharose CL-4B beads for 60 min in PBS. After washing
the beads twice with PBS to remove unbound GST-fused proteins, the
beads were incubated with target proteins suspended in the binding
buffer (PBS with 0.05% Nonidet P-40, 1 mM dithiothreitol,
and protease inhibitor mixture (Complete, Roche)). The binding buffer
for HDAC1 contained 0.3 M NaCl. After washing the beads
three times with the binding buffer, the protein bound to the beads was
detected by SDS-PAGE and immunoblotting.
Electrophoretic Mobility Shift Assay (EMSA)--
EMSA was
performed essentially as previously described (15), except that the
buffer contained 20 ng of sonicated E. coli genomic DNA in
20 µl of binding reaction mixture. The double-stranded DNA
oligonucleotides used as the probe was:
5'-GCCTTCAGACCCGGAGATCCAG-3' (upper strand) and
5'-ACTGGGATCTCCCGGTCTGAAGG-3' (lower strand). The
two underlined cytosines were unmethylated in the unmethylated probe
(U) and both methylated in the methylated probe (M). Approximately 100 ng of recombinant proteins was added to 20 µl of the reaction mixture
containing the probes. Protein-DNA complexes were fractionated in 4%
polyacrylamide gel supplemented with 3% glycerol in 0.5× TBE
at 4 °C. The gel was analyzed using a phosphoimager (Fuji, BAS-2000).
Indirect Immunofluorescence--
After washing with PBS, cells
were permeabilized with 0.5% Triton X-100 in PBS for 2 min at room
temperature. Cells were washed with PBS twice and fixed with 4%
paraformaldehyde in PBS for 10 min at room temperature. Cells were
treated twice with TBS containing 0.2% Tween 20 followed by incubation
with anti-HA antibodies for 60 min at room temperature. Cells were then
incubated with Alexa488-conjugated anti-mouse Ig antibodies (Molecular
Probes) for 60 min at room temperature and mounted on slides with the
mounting medium containing DAPI (Vectashield, Vector Laboratories).
The samples were examined under a fluorescence microscope (Zeiss).
DNA-binding Assay--
The purified GST-fused recombinant
proteins and 0.1 pmol of labeled oligonucleotides were incubated in 100 µl of PBS with glutathione-Sepharose CL-4B beads for 30 min. After a
brief centrifugation to sediment the beads, the supernatant was
collected for scintillation counting. After washing three times with
PBS to remove unbound oligonucleotides, the beads were suspended in
PBS. The slurry and the supernatant were transferred to Ready Cap
(Beckman) and subjected to scintillation counting. DNA-binding
efficiency was calculated as Efficiency = Ppt/(Sup+Ppt), where Sup
is the count of the first supernatant and Ppt is that of the slurry (beads).
His-30 and Phe-34 Are Responsible for the Inability of MBD3 to Bind
to mCpG--
An NMR study revealed that five amino acid residues in
the MBD domain of human MBD1 play important roles in recognizing the methyl groups of mCpG/mCpG (16). These amino acid residues, Val-20,
Arg-22, Tyr-34, Arg-44, and Ser-45, are highly conserved in MBD family
members. Interestingly, these residues are also conserved in xMBD3 but
not in human and mouse MBD3s where Tyr-34 is replaced with Phe-34 (the
numbering of equivalent amino acids in MBD1 and MBD3 is the same). In
addition, the same study showed that Arg-30 of MBD1 MBD interacts with
the DNA backbone. Basic amino acids are present at this position in
most MBD proteins including xMBD3. However, mammalian MBD3s contain an
atypical residue, His, at this position. Therefore, it was hypothesized that these two amino acid substitutions specific to mammalian MBD3
might be responsible for the reported inability of the protein to bind
to mCpG (9, 16). To test this hypothesis, we prepared three human MBD3
mutants (Fig. 1A). MBD3-H30K
and MBD3-F34Y encode mutant MBD3s in which His-30 and Phe-34 are
replaced with Lys and Tyr, respectively. MBD3-H30K/F34Y encodes mutant
MBD3 having both substitutions. These mutants were designed to correct
the MBD3-specific atypical residues to conserved ones.
The N-terminal 92 amino acids of wild type and mutant MBD3s were
expressed in E. coli as N-terminally GST-fused proteins. These fusion proteins contain the entire MBD3 MBD that comprises the
N-terminal 75 amino acids. The recombinant proteins were purified using
glutathione-Sepharose CL-4B beads (Fig. 1B), and subjected to EMSA (Fig. 1C). A 23-bp oligonucleotide containing a
single CpG site at the center was used as a probe for EMSA. Two
versions of oligonucleotides, U and M, were used. The single CpG/CpG is methylated not in U but in both strands in M. The GST-fused N-terminal wild type MBD3 (GST-wt(N)) produced a shifted band upon incubation with
the labeled M probe. However, this shifted band was competed at similar
efficiencies by a 100-fold excess of either unlabeled M or U. Thus, it
was concluded that GST-wt(N) binds to DNA in a methylation-independent
manner, as reported previously (2). The GST-fused N-terminal MBD3-H30K
(GST-H30K(N)) behaved in a manner similar to GST-wt(N). In contrast,
the GST-fused N-terminal MBD-F34Y (GST-F34Y(N)) produced a shifted band
upon incubation with the labeled M probe, and significantly, this band
was competed more efficiently by 100-fold excess of unlabeled M than by
unlabeled U. Therefore, it was suggested that substituting Tyr for Phe
at residue 34 of MBD3 significantly increased the mCpG-specific
DNA-binding activity. The GST-fused N-terminal MBD-H30K/F34Y
(GST-H30K/F34Y(N)) showed essentially similar DNA-binding activities to
those of GST-F34Y(N).
Because the above experiments did not give a quantitative idea of the
DNA-binding activities of mutant proteins, we quantitatively measured
the DNA-binding efficiency of those proteins. Different amounts of
purified GST-MBD fusion proteins were incubated with 1 nM
of labeled U or M oligonucleotides in the absence of any cold DNA. The
fusion proteins were precipitated with glutathione-Sepharose CL-4B
beads, and both pellet fractions and supernatant fractions were
scintillation-counted for the labeled probes. The bound fraction of the
DNA probe was determined from the following equation: Bound Fraction = Ppt/(Sup+ Ppt), where Ppt and Sup represent the
scintillation counts of Ppt and Sup, respectively. The values obtained
in this way are plotted against the protein concentration (Fig.
2). Both GST-wt(N) and GST-H30K(N) did
not show significant binding to either U or M in this assay (Fig. 2,
A and B). In contrast, GST-F34Y(N) showed weak
but significant binding to the M probe (Fig. 2C). It was not
possible to deduce the Kd value for this binding, however, because protein-DNA-binding was not saturated in the range of
protein concentrations examined. Interestingly, GST-H30K/F34Y(N) showed
a remarkable increase both in the binding affinity to DNA and in the
specificity to bind to M rather than to U. Kd between GST-H30K/F34Y(N) and M was estimated to be 9.32 × 10
It was reported that Tyr-34 in MBD1 plays important roles both directly
and indirectly in MBD1-mCpG binding (16). Given the positive roles of
this residue, it was expected that the substitution of Ala instead of
Tyr for Phe-34 would not improve the DNA-binding activity of MBD3.
Unexpectedly, however, GST-fused F34A(N) showed significantly higher
DNA-binding activity than GST-wt(N) (Fig. 3, A and B).
Interestingly, GST-F34A(N) did not show any specificity to U or M,
suggesting that Tyr-34 may be important for the mCpG-specific binding
activity. The equivalent mutant of MBD2b, GST-MBD2b Y34A(N), also
showed DNA-binding activity comparable with that of GST-F34A(N) (Fig.
3, A and C). Similar observations have been
previously reported for MeCP2. MeCP2 that had mutations at Tyr-123, the
equivalent residue of Tyr-34 in MBD1, showed a reduced but significant
binding activity to mCpG (18). The simplest interpretation of these results is that Phe-34 in MBD3 has an inhibitory effect on nonspecific DNA-binding activity, and substituting any amino acid for this residue
may lead to the recovery of the nonspecific DNA-binding activity.
Neither Wild Type nor Mutant MBD3 MBD Associates with mCpG-Rich
Heterochromatin in Vivo--
Questions regarding the ability of
MBD3 to bind with mCpG in vivo were raised in an
experiment in which MBD3-GFP proteins were expressed in mouse cells
(2). In mouse cells, ~50% of total mCpG is concentrated at
pericentromeric regions that consist of the major satellite DNA.
Pericentromeric regions are easily detected in interphase cells by
virtue of their being brightly stained by DAPI. In a previous study (2,
22), it was found that overexpressed mouse MBD1, MBD2, MBD4, and
rat MeCP2 were localized at pericentromeric regions, supporting the
proficient mCpG-binding activity of these proteins in vivo.
In contrast, the overexpressed MBD3 protein was not localized at
pericentromeric regions and instead was homogeneously distributed in
the nuclei when MBD3 was expressed at a low level or accumulated as
nuclear foci when it was expressed at a high level. Together with the data that human MBD3 does not bind to mCpG in vitro, it was
concluded that MBD3 does not function as an mCpG-binding protein
in vivo (2). Because we have found that the H30K/F34Y MBD3
mutant binds to mCpG efficiently in vitro, we were
interested in examining whether or not this mutant protein associated
with mCpG in vivo.
The MBD domains of wild type and mutant human MBD3s, as well as MBD2b,
were overexpressed in mouse NIH 3T3 cells as HA-tagged proteins by
transient DNA transfection. The cells were stained with DAPI for
pericentromeric regions and with anti-HA antibodies for MBD
localization (Fig. 4). MBD2b MBD was
localized at the pericentromeric regions (Fig. 4, c),
whereas wild type MBD3-derived MBD were distributed diffusely in
the nuclei, as expected (Fig. 4, a). Neither H30K, F34Y
(data not shown), nor H30K/F34Y mutant MBD3 MBD (Fig. 4, b)
co-localized at the pericentromeric regions. Because MBD3 H30K/F34Y
specifically binds to mCpG in vitro (Fig. 2E),
these results suggested that the ability of MBD proteins to associate
with mCpG in vivo is not a simple reflection of the ability
of these proteins to bind to mCpG in vitro.
Next, we performed similar analyses for the mutated versions of MBD2b.
MBD2b MBD carrying the two substitutions found in wild type MBD3 (MBD2b
K30H/Y34F(N)) was not co-localized at the pericentromeric regions, as
expected (Fig. 4, e). Unexpectedly, MBD2b Y34F, which binds
to DNA in an mCpG-independent manner (Fig. 3D), was
co-localized at the pericentromeric regions (Fig. 4, d).
Because MBD2b Y34F and MBD2b K30H/Y34F showed similar binding behaviors
(i.e. mCpG-independent DNA-binding activity) (Fig. 3,
D and E), the different localizations of these
two versions of MBD2b MBD in vivo again suggest the presence of other factors that determine the localization of MBD proteins in vivo in addition to the binding ability of the protein to
DNA measured in vitro.
MBD3 Does Not Bind to Either the mCpG/TpG Mismatch or
Methyl-cytosine in the CpA Context--
MBD4 specifically recognizes
the mCpG/TpG mismatch sequence (3, 4). Because we found that MBD3 MBD
did not bind to mCpG/mCpG specifically, we next examined the
possibility that MBD3 binds to the mCpG/TpG mismatch sequences using
the assay system described in Fig. 2. GST-fused recombinant protein of
MBD4 MBD binds to an oligonucleotide containing CpG/TpG and at a higher
affinity to one containing mCpG/TpG, as expected (Fig.
5A). Under the same conditions, GST-fused MBD3 MBD did not bind to either CpG/TpG or
mCpG/TpG.
Recently, it was reported that a minor population of methylcytosine in
the genome of mouse embryonic cells is present as CpA-dinucleotides (19). We also tested the possibility that MBD3 MBD might bind specifically to this sequence. The recombinant MBD3 MBD protein did not
show appreciable binding ability to either an mCpA/TpG- or a
CpA/TpG-containing oligonucleotide (Fig. 5B). These results indicate that MBD3 does not recognize these relatively minor
populations of methylcytosine in the genome.
MBD Is Necessary and Sufficient for Binding to MTA2 and
HDAC1--
MBD3 was identified as a component of the NuRD/Mi2 complex
(8, 11). In vitro, it was shown that human MBD3 directly
binds to many NuRD/Mi2 components, including MTA2, HDAC1, RbAp46, and RbAp48. We examined regions in MBD3 responsible for binding to MTA2 and
HDAC1. MBD3 (also called MBD3a) (11) and its deletion mutants were
expressed as GST-fusion proteins in E. coli (Fig. 6A). MBD3 proteins were
incubated with recombinant full-length MTA2 or HDAC1 and sedimented
with glutathione-Sepharose CL-4B beads. After washing the beads, MBD3
and associated proteins were examined by SDS-PAGE. All MBD3 proteins
were confirmed for their expression and absorption by
glutathione-Sepharose CL-4B beads (Fig. 6, B and
C). Both MTA2 and HDAC1 were associated with all versions of
MBD3 except one that lacks the N-terminal 92 amino acids that contain
MBD ( In this study, we demonstrated that human MBD3 has no appreciable
mCpG-specific DNA-binding activity. The lack of DNA-binding activity in
MBD3 appears to be due to two reasons. First, the amino acid
substitution of the well conserved Lys-30 and Tyr-34 in MBD3 is the
major reason why MBD3 lost its mCpG-specific DNA-binding activity. This
was proven by the fact that simultaneous corrections of these amino
acids to the conserved residues resulted in conferring significant
mCpG-binding activity to MBD3. Second, the presence of Phe at residue
34 positively inhibits the potential DNA-binding activity of MBD3
because the F34A mutant showed significant DNA-binding activity, albeit
nonspecific to methyl or non-methyl CpG. One potential explanation for
the lack of any DNA-binding activity in the presence of Phe at residue
34 is that this bulky amino acid may significantly change the
structural framework of the protein. However, given that MBD2b Y34F and
MBD2b K30H/Y34F bind to DNA nonspecifically (Fig. 3, D and
E), the effect of Phe at residue 34 appears to be
context-dependent. Collectively, the presence of Tyr at
residue 34 (and equivalent positions) in conventional MBDs is important
for conferring mCpG specificity, a biochemical result consistent with
that of the structural study (16). Another unexpected finding is that
even if the H30K/F34Y MBD3 mutant protein showed significant
mCpG-binding activity in vitro, when expressed in
vivo it did not associate with the mCpG-rich pericentromeric regions in NIH 3T3 cells. This observation led us to suggest that the
in vitro binding activity of the MBD proteins does not
directly lead to the association of the proteins with the mCpG-rich
regions in vivo.
Functions of MBD in MBD3--
MBD2b and MBD3 are highly homologous
(71% identity), and MBD2 and MBD3 genes have
very similar genome structures although mapped to different chromosomes
in mouse and human (20). Therefore, it was suggested that these genes
have diverged from a relatively recent common ancestor (20). Why then
did MBD3 lose its DNA-binding activity during evolution, and does it
function as a core component of the NuRD/Mi2 complex while maintaining
MBD? In this study, we demonstrated that MBD3 MBD is necessary and
sufficient for the physical interaction with MTA2 and HDAC1. We believe
that this secondary role played by MBD is the reason why MBD3 has
evolutionarily conserved the apparently "useless" MBD. The dual
roles of MBD, namely binding to mCpG and interacting with NuRD/Mi2
components, are not exclusive of each other because we have noted that
the MBD3 H30K/F34Y mutant protein associates with MTA2 or HDAC1 (Fig. 6D) and that MBD2b MBD also interacts with MTA2 and HDAC1
in vitro.2
However, we do not know whether one MBD molecule simultaneously interacts with mCpG and proteins or not. The substitutions at Lys-30
and Tyr-34 to His-30 and Phe-34, respectively, should have specified
the role of MBD3 as a protein-interacting protein rather than as a
DNA-binding protein.
Previously (15), we have shown that MBD2 and MBD3 form a heterodimer
via an interaction between the coiled-coil domains of the two proteins.
Therefore, it can be imagined that the ancestor of MBD2/MBD3 forms a
similar dimer that has two mCpG-interacting domains in one complex.
During the course of evolution, MBD3 MBD lost its mCpG-binding activity
and became specialized as an adaptor protein for MTA2 and HDAC1. The
current MBD2-MBD3 heterodimer has a single mCpG-interacting domain
provided by MBD2 and two potential MTA2/HDAC1-interacting domains
provided by MBD2 and MBD3. However, it is possible that the ability of
MBD2 to bind to MTA2 or HDAC1 is not functional in vivo
because the same region binds to mCpG.
There are other examples of domain structures that are evolutionarily
conserved while losing their original function. PAF400, a component of
the PCAF complex, has a typical domain characterizing the ATM
family (21). However, the PAF400 ATM kinase-like domain has mutations
at amino acid residues responsible for the kinase activity, and indeed,
no catalytic activity was demonstrated with PAF400. As proposed for
MBD3 in this study, a second important role may be achieved by the
catalytically inactive ATM domain of PAF400.
Independence of the in Vitro mCpG-Binding Activity from the in Vivo
mCpG-Rich Heterochromatin-associating Activity in MBD
Proteins--
One unexpected observation was that the H30K/F34Y MBD3
mutant did not associate with the mCpG-rich pericentromeric regions in
NIH 3T3 cells. Two possibilities can be envisioned. First, it is
possible that the MeCP2, MBD1, MBD2, and MBD4 proteins, which show
proficient pericentromeric association in vivo, may have a
second domain in addition to MBD that facilitates the localization of
proteins at specific regions in vivo. MBD3 may not have this second domain. There have been precedent studies suggesting this possibility. It was reported that when a MeCP2 mutant protein that had
a deletion in MBD was overexpressed, a small fraction of the protein
still associated with the pericentromeric regions. This result
suggested the presence of a second domain in MeCP2 that
MBD-independently facilitated the localization of MeCP2 at the
pericentromeric regions (22). In another study, it was found that
GFP-MBD1 localization at the pericentromeric regions did not change
markedly in DNMT1
The second possibility is that MBD3 may have an additional domain that
facilitates the recruitment of the protein to the NuRD·Mi2 complex,
and this domain is not present in other MBD proteins. Recently, the
zinc-finger protein called p66 was identified as a component of the
MeCP1 complex. p66 also binds to MBD3, but not to MBD2, and possibly
determines the intra-nuclear localization of MBD3 (23). p66 or its
related proteins may be a candidate for the factor that positively
influences the localization of MBD3 in the NuRD·Mi2 complex and
outside the pericentromeric regions.
Interestingly, MBD2b MBD Y34F accumulated at the pericentromeric
regions although it had no mCpG-specific DNA-binding activity. This
result suggests that the mCpG-specific DNA-binding activity is largely
dispensable for the pericentromeric localization of MBD proteins and
supports the notion that a second determinant is required for the
correct targeting of the protein in vivo. Because we have
not identified any mutant MBD that lacks nonspecific DNA-binding
activity but retains the ability to accumulate at the pericentromeric
regions, it remains unclear whether the nonspecific DNA-binding
activity of MBD proteins is required for the heterochromatin targeting
in vivo or not. The second mutation of K30H in MBD2b abolished the pericentromeric accumulation of the protein when it was
present simultaneously with Y34F. This observation suggests that a
region distinct from MBD but overlapping with Lys-30 may be important
to recruit wild type MBD2 to the mCpG-rich regions in vivo,
according to the first hypothesis. Although both MBD2b MBD and MBD3
H30K/F34Y associate with MTA2 and HDAC1 in vitro (discussed
above), MBD2b MBD is localized at the pericentromeric regions, whereas
MBD3 H30K/F34Y is not. This observation indicates that the interaction
of MBD with MTA2 and HDAC1 in vitro is not the determinant
of the protein localization in vivo. Therefore, there should
be a second domain in MBD3 outside of amino acid 30 or 34 that is
recognized by the NuRD/Mi2 complex, according to the second hypothesis.
In this study, we have expressed and analyzed recombinant proteins
containing the MBDs of MBD2b and MBD3 but lacking unique regions such
as the glutamate-rich C-terminal tail in MBD3. Therefore, the
additional domain unique to MBD2b or MBD3 should reside in the highly
conserved MBDs. No noticeable sequence was recognized in the MBD
sequences of MBD2b and MBD3, however. Further studies are necessary to
determine why the H30K/F34Y MBD3 mutant did not associate with the
mCpG-rich region in vivo.
We thank Dr. Yi Zhang for providing mouse
MTA2 cDNA. We thank F. Nishizaki, A. Orii, and A. Katayama for
excellent secretarial work and M. Tamura for technical assistance.
*
This work was supported in part by the COE Grant and
a grant-in-aid for Cancer Research from the Ministry of Education,
Culture, Sports, Science and Technology, Japan.The costs of publication of this
article were defrayed in part by the
payment of page charges. The article
must therefore be hereby marked
"advertisement" in accordance with 18 U.S.C. Section
1734 solely to indicate this fact.
§
Supported by the Japan Society for the Promotion of Science
Research Fellowship for Young Scientists.
Published, JBC Papers in Press, July 17, 2002, DOI 10.1074/jbc.M203455200
2
M. Saito and F. Ishikawa, unpublished data.
The abbreviations used are:
MBD, mCpG-binding domain;
mCpG, methylated CpG;
HDAC, histone deacetylase;
GST, glutathione S-transferase;
PBS, phosphate-buffered
saline;
EMSA, electrophoretic mobility shift assay;
U, unmethylated
probe;
M, methylated probe;
DAPI, 4',
6-diamidino-2-phenylindole;
wt, wild type;
N, N-terminal;
HA, haemagglutinin A.
The mCpG-binding Domain of Human MBD3 Does Not Bind to mCpG but
Interacts with NuRD/Mi2 Components HDAC1 and MTA2*
§ and¶
Laboratory of Molecular and Cellular
Assembly, Department of Biological Information, Graduate School of
Bioscience and Biotechnology, Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama 226-8501, Japan and the
¶ Laboratory of Cell Cycle Regulation, Department of Gene
Mechanisms, Graduate School of Biostudies, Kyoto University,
Kitashirakawa-Oiwake-cho, Kyoto 606-8502, Japan
ABSTRACT
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
MATERIALS AND METHODS
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
RESULTS
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
View larger version (38K):
[in a new window]
Fig. 1.
mCpG-binding activities of human MBD3 MBD.
A, amino acid sequences of wild type and mutant MBD3
proteins as well as the corresponding region of MBD2b. B,
SDS-PAGE analyses of recombinant proteins. Purified recombinant
proteins were fractionated by SDS-PAGE and visualized by silver
staining. Lane 1, GST-wt(N); lane 2, GST-H30K(N);
lane 3, GST-F34Y(N); and lane 4,
GST-H30K/F34Y(N). The asterisk indicates the position of the
full-length recombinant proteins. The faster migrating bands are
assumed to be the degradation products of the recombinant proteins.
C, EMSA of human MBD3. GST-fused wild type N-terminal MBD3
(amino acids 1-92) (GST-wt(N)) and mutant versions
(GST-H30K(N), GST-F34Y(N), and
GST-H30K/F34Y(N)) were analyzed for mCpG-binding activity.
U and M indicate unmethylated and methylated
oligonucleotide DNA, respectively. Cold competitors were included in
the reaction in 100-fold excess of the labeled probe. Due to the
different binding affinities of different proteins, the amounts of
protein used in the binding reaction were different: ~100 ng for
GST-wt(N) and GST-H30K(N), 20 ng for GST-F34Y(N), and 2 ng for
GST-H30K/F34Y(N). The binding reaction mixture contained 20 ng of
sonicated E. coli genomic DNA, which presumably lacks mCpG
modification, to minimize the nonspecific DNA binding of
proteins.
8 M (Fig. 2D). In a similar
experiment, Kd between M and the GST-fused MBD
domain of MBD2b (amino acids 1-88, the region equivalent to the MBD3
MBD used in this study), a well characterized methyl-DNA-binding
protein (17), was estimated to be 2.67 × 108
M (Fig. 2E). Therefore, simultaneous
substitutions of H30K and F34Y conferred to MBD3 a methyl-DNA-specific
binding activity that was comparable with that of MBD2b. These results
supported the previous hypothesis that the two atypical amino acids at
residues 30 and 34 are responsible for the inability of MBD3 to bind to mCpG specifically.
View larger version (24K):
[in a new window]
Fig. 2.
Kd analyses of recombinant
MBD3 proteins. 1 pmol of labeled probe M (open circles)
or U (open squares) in 100 µl of reaction mixture was
incubated with indicated concentrations of recombinant proteins, and
the protein-bound fraction of the probe was measured (see "Materials
and Methods" for more details). The DNA-binding efficiency of
GST-wt-MBD3(N) (A), GST-H30K(N) (B), GST-F34Y(N)
(C), GST-H30K/F34Y(N) (D), and GST-MBD2b(N)
(E) is plotted against the protein concentration. At least
two samples were analyzed. Kd values were calculated
and are shown in D and E, where more than three
samples were analyzed for each protein concentration.
View larger version (27K):
[in a new window]
Fig. 3.
mCpG-binding activities of recombinant MBD3
and MBD2 proteins. A, amino acid sequences of wild type and
mutant forms of MBD3 and MBD2b. DNA binding activities of GST-F34A(N)
(B), GST-MBD2b Y34A(N) (C), GST-MBD2b Y34F(N)
(D), and GST-MBD2b K30H/Y34F(N) (E) were analyzed
as in Fig. 2. The open circles and the open
squares indicate the DNA-binding efficiency against the methylated
probe and the non-methylated probe, respectively.
View larger version (65K):
[in a new window]
Fig. 4.
Localization of overexpressed MBD3 and MBD2
proteins in NIH 3T3 cells. HA-tagged indicated proteins were
transiently overexpressed in mouse NIH 3T3 cells. Locations of the
proteins and DNA were examined by the indirect immunofluorescence
technique using anti-HA antibodies and DAPI staining,
respectively.
View larger version (18K):
[in a new window]
Fig. 5.
Binding activities of MBD3 to unusual
methylcytosine sequences. A, GST-fused recombinant proteins
of MBD3-derived and MBD4-derived MBDs were examined for binding
activity to oligonucleotides containing the CpG/TpG mismatch or the
mCpG/TpG mismatch. The dinucleotide sequence of mCpG/TpG is shown on
the left. B, GST-fused recombinant proteins of
MBD3-derived MBD as well as GST only were examined for binding activity
to oligonucleotides containing CpA/TpG or mCpA/TpG. The dinucleotide
sequence of mCpA/TpG is shown on the left.
N92). Reciprocally, the MBD3 mutant consisting of MBD only
(C93) associated with MTA2 and HDAC1. Therefore, MBD is necessary
and sufficient for binding to MTA2 and HDAC1. However, a significant
reduction of the recovery of MTA2 and HDAC1 was observed in an
experiment using C174 compared with that in experiments using
mutants with shorter C-terminal deletions (such as C221). Therefore,
it was suggested that there might be a region facilitating MTA2- and
HDAC1-MBD3 association in amino acids 174-220. We also examined
whether MBD3 H30K/F34Y, a proficient type of the protein in methyl-CpG
binding, associates with MTA2 and HDAC1 or not (Fig. 6D).
The results showed that this mutant protein interacts with MTA2 and
HDAC1 as efficiently as the wild type MBD3.
View larger version (25K):
[in a new window]
Fig. 6.
Physical interactions between MBD3 regions
and NuRD/Mi2 components. A, schematic structures of
GST-fused recombinant MBD3 proteins used in the experiments.
B, purified GST-fused MBD3 proteins and His-tagged MTA2 were
incubated, and proteins associated with GST-MBD3 were sedimented using
glutathione-Sepharose CL-4B beads. The bound proteins were fractionated
by SDS-PAGE and analyzed for MTA2 (upper panel) and MBD3
(lower panel) using anti-MTA2 antibodies and Coomassie
brillant blue staining, respectively. One-tenth amounts of the input
proteins were loaded onto the input lane. C, His-tagged
HDAC1 was analyzed as in B. Upper panel,
immunoblot using anti-HDAC1 antibodies. Lower panel,
Coomassie brillant blue staining. D, binding
abilities of recombinant MBD3 MBDs to MTA2 and HDAC1. GST-wt(N) (the
same construct as C93 in A) and GST-H30K/F34Y(N) (C93
containing two amino acid substitutions) were analyzed as in
B. MTA2 and HDAC1 were detected by immunoblotting using the
specific antibodies.
DISCUSSION
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
/ ES cells that contained a
very low level of mCpG (2). This finding suggested that the
localization of MBD1 at the pericentromeric regions might be partly
mCpG-independent.
ACKNOWLEDGEMENTS
FOOTNOTES
To whom correspondence should be addressed. E-mail:
fishikaw@lif.kyoto-u.ac.jp
ABBREVIATIONS
REFERENCES
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
1.
Nan, X.,
Meehan, R. R.,
and Bird, A.
(1993)
Nucleic Acids Res.
21,
4886-4892 2.
Hendrich, B.,
and Bird, A.
(1998)
Mol. Cell. Biol.
18,
6538-6547 3.
Hendrich, B.,
Hardeland, U., Ng, H. H.,
Jiricny, J.,
and Bird, A.
(1999)
Nature
401,
301-304[CrossRef][Medline]
[Order article via Infotrieve]
4.
Riccio, A.,
Aaltonen, L. A.,
Godwin, A. K.,
Loukola, A.,
Percesepe, A.,
Salovaara, R.,
Masciullo, V.,
Genuardi, M.,
Paravatou-Petsotas, M.,
Bassi, D. E.,
Ruggeri, B. A.,
Klein-Szanto, A. J.,
Testa, J. R.,
Neri, G.,
and Bellacosa, A.
(1999)
Nat. Genet.
23,
266-268[CrossRef][Medline]
[Order article via Infotrieve]
5.
Fujita, N.,
Shimotake, N.,
Ohki, I.,
Chiba, T.,
Saya, H.,
Shirakawa, M.,
and Nakao, M.
(2000)
Mol. Cell. Biol.
20,
5107-5118 6.
Bird, A. P.,
and Wolffe, A. P.
(1999)
Cell
99,
451-454[CrossRef][Medline]
[Order article via Infotrieve]
7.
Boeke, J.,
Ammerpohl, O.,
Kegel, S.,
Moehren, U.,
and Renkawitz, R.
(2000)
J. Biol. Chem.
275,
34963-34967 8.
Wade, P. A.,
Gegonne, A.,
Jones, P. L.,
Ballestar, E.,
Aubry, F.,
and Wolffe, A. P.
(1999)
Nat. Genet.
23,
62-66[Medline]
[Order article via Infotrieve]
9.
Ohki, I.,
Shimotake, N.,
Fujita, N.,
Nakao, M.,
and Shirakawa, M.
(1999)
EMBO J.
18,
6653-6661[CrossRef][Medline]
[Order article via Infotrieve]
10.
Wakefield, R. I.,
Smith, B. O.,
Nan, X.,
Free, A.,
Soteriou, A.,
Uhrin, D.,
Bird, A. P.,
and Barlow, P. N.
(1999)
J. Mol. Biol.
291,
1055-1065[CrossRef][Medline]
[Order article via Infotrieve]
11.
Zhang, Y., Ng, H. H.,
Erdjument-Bromage, H.,
Tempst, P.,
Bird, A.,
and Reinberg, D.
(1999)
Genes Dev.
13,
1924-1935 12.
Guschin, D.,
Wade, P. A.,
Kikyo, N.,
and Wolffe, A. P.
(2000)
Biochemistry
39,
5238-5245[CrossRef][Medline]
[Order article via Infotrieve]
13.
Wade, P. A.
(2001)
Oncogene
20,
3166-3173[CrossRef][Medline]
[Order article via Infotrieve]
14.
Feng, Q.,
and Zhang, Y.
(2001)
Genes Dev.
15,
827-832 15.
Tatematsu, K. I.,
Yamazaki, T.,
and Ishikawa, F.
(2000)
Genes Cells
5,
677-688[Abstract]
16.
Ohki, I.,
Shimotake, N.,
Fujita, N.,
Jee, J.,
Ikegami, T.,
Nakao, M.,
and Shirakawa, M.
(2001)
Cell
105,
487-497[CrossRef][Medline]
[Order article via Infotrieve]
17.
Ng, H. H.,
Zhang, Y.,
Hendrich, B.,
Johnson, C. A.,
Turner, B. M.,
Erdjument-Bromage, H.,
Tempst, P.,
Reinberg, D.,
and Bird, A.
(1999)
Nat. Genet.
23,
58-61[Medline]
[Order article via Infotrieve]
18.
Free, A.,
Wakefield, R. I.,
Smith, B. O.,
Dryden, D. T.,
Barlow, P. N.,
and Bird, A. P.
(2001)
J. Biol. Chem.
276,
3353-3360 19.
Ramsahoye, B. H.,
Biniszkiewicz, D.,
Lyko, F.,
Clark, V.,
Bird, A. P.,
and Jaenisch, R.
(2000)
Proc. Natl. Acad. Sci. U. S. A.
97,
5237-5242 20.
Hendrich, B.,
Abbott, C.,
McQueen, H.,
Chambers, D.,
Cross, S.,
and Bird, A.
(1999)
Mamm. Genome
10,
906-912[CrossRef][Medline]
[Order article via Infotrieve]
21.
Vassilev, A.,
Yamauchi, J.,
Kotani, T.,
Prives, C.,
Avantaggiati, M. L.,
Qin, J.,
and Nakatani, Y.
(1998)
Mol. Cell
2,
869-875[CrossRef][Medline]
[Order article via Infotrieve]
22.
Nan, X.,
Tate, P., Li, E.,
and Bird, A.
(1996)
Mol. Cell. Biol.
16,
414-421[Abstract]
23.
Feng, Q.,
Cao, R.,
Xia, L.,
Erdjument-Bromage, H.,
Tempst, P.,
and Zhang, Y.
(2002)
Mol. Cell. Biol.
22,
536-546
Copyright © 2002 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  S. E. Brown and M. Szyf Epigenetic Programming of the rRNA Promoter by MBD3 Mol. Cell. Biol., July 1, 2007; 27(13): 4938 - 4952. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J J L Wong, N J Hawkins, and R L Ward Colorectal cancer: a model for epigenetic tumorigenesis Gut, January 1, 2007; 56(1): 140 - 148. [Full Text] [PDF]
|
|
|
|
|
|
P. Gu, D. Le Menuet, A. C.-K. Chung, and A. J. Cooney Differential Recruitment of Methylated CpG Binding Domains by the Orphan Receptor GCNF Initiates the Repression and Silencing of Oct4 Expression Mol. Cell. Biol., December 15, 2006; 26(24): 9471 - 9483. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. P. Kransdorf, S. Z. Wang, S. Z. Zhu, T. B. Langston, J. W. Rupon, and G. D. Ginder MBD2 is a critical component of a methyl cytosine-binding protein complex isolated from primary erythroid cells Blood, October 15, 2006; 108(8): 2836 - 2845. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Rosendorff, S. Sakakibara, S. Lu, E. Kieff, Y. Xuan, A. DiBacco, Y. Shi, Y. Shi, and G. Gill NXP-2 association with SUMO-2 depends on lysines required for transcriptional repression PNAS, April 4, 2006; 103(14): 5308 - 5313. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. J. Noh, E. R. Jang, G. Jeong, Y. M. Lee, C. K. Min, and J.-S. Lee Methyl CpG-Binding Domain Protein 3 Mediates Cancer-Selective Cytotoxicity by Histone Deacetylase Inhibitors via Differential Transcriptional Reprogramming in Lung Cancer Cells Cancer Res., December 15, 2005; 65(24): 11400 - 11410. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Datta, S. Majumder, S. Bai, K. Ghoshal, H. Kutay, D. S. Smith, J. W. Crabb, and S. T. Jacob Physical and Functional Interaction of DNA Methyltransferase 3A with Mbd3 and Brg1 in Mouse Lymphosarcoma Cells Cancer Res., December 1, 2005; 65(23): 10891 - 10900. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Sharma, J. Blum, X. Yang, N. Beaulieu, A. R. Macleod, and N. E. Davidson Release of Methyl CpG Binding Proteins and Histone Deacetylase 1 from the Estrogen Receptor {alpha} (ER) Promoter upon Reactivation in ER-Negative Human Breast Cancer Cells Mol. Endocrinol., July 1, 2005; 19(7): 1740 - 1751. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Kondo, Z. Gu, A. Horii, and S. Fukushige The Thymine DNA Glycosylase MBD4 Represses Transcription and Is Associated with Methylated p16INK4a and hMLH1 Genes Mol. Cell. Biol., June 1, 2005; 25(11): 4388 - 4396. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S.-G. Jin, C.-L. Jiang, T. Rauch, H. Li, and G. P. Pfeifer MBD3L2 Interacts with MBD3 and Components of the NuRD Complex and Can Oppose MBD2-MeCP1-mediated Methylation Silencing J. Biol. Chem., April 1, 2005; 280(13): 12700 - 12709. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Tong, T. Aune, and M. Boothby T-bet antagonizes mSin3a recruitment and transactivates a fully methylated IFN-{gamma} promoter via a conserved T-box half-site PNAS, February 8, 2005; 102(6): 2034 - 2039. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Ruzov, D. S. Dunican, A. Prokhortchouk, S. Pennings, I. Stancheva, E. Prokhortchouk, and R. R. Meehan Kaiso is a genome-wide repressor of transcription that is essential for amphibian development Development, December 15, 2004; 131(24): 6185 - 6194. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C.-L. Jiang, S.-G. Jin, and G. P. Pfeifer MBD3L1 Is a Transcriptional Repressor That Interacts with Methyl-CpG-binding Protein 2 (MBD2) and Components of the NuRD Complex J. Biol. Chem., December 10, 2004; 279(50): 52456 - 52464. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Andrin and M. J. Hendzel F-actin-dependent Insolubility of Chromatin-modifying Components J. Biol. Chem., June 11, 2004; 279(24): 25017 - 25023. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M.-S. Hung and C.-K. J. Shen Eukaryotic Methyl-CpG-Binding Domain Proteins and Chromatin Modification Eukaryot. Cell, October 1, 2003; 2(5): 841 - 846. [Full Text] [PDF]
|
|
|
|
 |
|
 A. Berg, T. J. Meza, M. Mahic, T. Thorstensen, K. Kristiansen, and R. B. Aalen Ten members of the Arabidopsis gene family encoding methyl-CpG-binding domain proteins are transcriptionally active and at least one, AtMBD11, is crucial for normal development Nucleic Acids Res., September 15, 2003; 31(18): 5291 - 5304. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. F. Fraga, E. Ballestar, G. Montoya, P. Taysavang, P. A. Wade, and M. Esteller The affinity of different MBD proteins for a specific methylated locus depends on their intrinsic binding properties Nucleic Acids Res., March 15, 2003; 31(6): 1765 - 1774. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Sakai, T. Urano, K. Ookata, M.-H. Kim, Y. Hirai, M. Saito, Y. Nojima, and F. Ishikawa MBD3 and HDAC1, Two Components of the NuRD Complex, Are Localized at Aurora-A-positive Centrosomes in M Phase J. Biol. Chem., December 6, 2002; 277(50): 48714 - 48723. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |