Báo cáo khoa học: 5-Bromodeoxyuridine induces transcription of repressed genes with disruption of nucleosome positioning pptx

5-Bromodeoxyuridine induces transcription of repressed
genes with disruption of nucleosome positioning
Kensuke Miki
1
, Mitsuhiro Shimizu
2
, Michihiko Fujii
1
, Shinichi Takayama
1
,
Mohammad Nazir Hossain
1
and Dai Ayusawa
1
1 Department of Genome System Science, Yokohama City University, Yokohama, Kanagawa, Japan
2 Department of Chemistry, Meisei University, Hino, Tokyo, Japan
Introduction
5-Bromodeoxyuridine (BrdU), which is frequently used
to measure DNA synthesis immunochemically in living
cells, is also well known to modulate various biological
functions when incorporated into DNA as 5-bromoura-
cil instead of thymine. Previously we have found that
BrdU very clearly induces a senescent-like phenomenon
in every type of mammalian cell and also in yeast cells
[1,2]. Historically, BrdU has been used as a modulator
of cellular differentiation with cAMP and butyrate
[3,4]. The latter two are found to target protein kinase
A and histone deacetylase, respectively, leading to func-
tional understanding of cell signaling and gene expres-
sion. BrdU is thought to influence the expression of key

genes involved in cellular differentiation. However, the
molecular mechanism underlying the actions of BrdU
still remains a mystery in spite of many efforts [5].
We have extensively characterized genes up- or
down-regulated by the addition of BrdU in various cell
Keywords
AT-tract; 5-bromodeoxyuridine; BAR1;
nucleosome positioning; transcriptional
derepression
Correspondence
D. Ayusawa, Department of Genome
System Science, Yokohama City University,
Seto 22-2, Kanazawa-Ku, Yokohama,
Kanagawa 236-0027, Japan
Fax: +81 45 787 2193
Tel: +81 45 787 2193
E-mail:
(Received 5 January 2010, revised 2 August
2010, accepted 2 September 2010)
doi:10.1111/j.1742-4658.2010.07868.x
5-Bromodeoxyuridine (BrdU) modulates the expression of particular genes
associated with cellular differentiation and senescence when incorporated
into DNA instead of thymidine (dThd). To date, a molecular mechanism
for this phenomenon remains a mystery in spite of a large number of stud-
ies. Recently, we have demonstrated that BrdU disrupts nucleosome posi-
tioning on model plasmids mediated by specific AT-tracts in yeast cells.
Here we constructed a cognate plasmid that can form an ordered array of
nucleosomes determined by an a2 operator and contains the BAR1 gene as
an expression marker gene to examine BAR1 expression in dThd-auxotro-
phic MATa cells under various conditions. In medium containing dThd,

BAR1 expression was completely repressed, associated with the formation
of the stable array of nucleosomes. Insertion of AT-tracts into a site of the
promoter region slightly increased BAR1 expression and slightly destabi-
lized nucleosome positioning dependent on their sequence specificity. In
medium containing BrdU, BAR1 expression was further enhanced, associ-
ated with more marked disruption of nucleosome positioning on the pro-
moter region. Disruption of nucleosome positioning seems to be sufficient
for full expression of the marker gene if necessary transcription factors are
supplied. Incorporation of 5-bromouracil into the plasmid did not weaken
the binding of the a2 ⁄ Mcm1 repressor complex to its legitimate binding
site, as revealed by an in vivo UV photofootprinting assay. These results
suggest that BrdU increases transcription of repressed genes by disruption
of nucleosome positioning around their promoters.
Abbreviations
BrdU, 5-bromodeoxyuridine; dThd, thymidine; MNase, micrococcal nuclease; TK, thymidine kinase.
FEBS Journal 277 (2010) 4539–4548 ª 2010 The Authors Journal compilation ª 2010 FEBS 4539
lines using PCR-based cDNA subtractive hybridization
and DNA microarray analysis [6,7]. Such BrdU-
responsive genes behave similarly in normal human
fibroblasts undergoing replicative senescence. BrdU
decondenses particular regions of chromosomes after
incorporation into DNA [8], suppresses position effect
variegation [9] and restores expression of silenced
genes [10]. Consistent with this, BrdU-responsive genes
are located on particular regions of human chromo-
somes, forming clusters on or nearby Giemsa-dark
bands of human chromosomes [11,12]. AT-tract minor
groove binders, such as distamycin A, netropsin, Hoe-
chst 33258 and the AT-hook protein HMG-I, have all
been shown to markedly potentiate the effects of BrdU

[11,13]. On the basis of the above observations, we
suggest that BrdU targets certain types of AT-rich
sequence and alters the chromatin structure to induce
particular genes.
In eukaryotes, DNA fibers exist as regularly arrayed
beads of nucleosomes. Nucleosomes restrict the acces-
sibility of transcription factors to promoters and regu-
latory sequences of genes. Thus, an alteration of
nucleosome positioning is an essential step for the
transition from a repressed state to an active state by
aid of chromatin remodeling complexes [14–16]. In our
previous studies, we addressed the effect of BrdU on
nucleosome positioning in vivo using TALS plasmids
[17], which have been successfully utilized to study
nucleosome positioning in Saccharomyces cerevisiae.
We have clearly shown that 5-bromouracil incorpo-
rated into the plasmids disrupts nucleosome position-
ing by inducing A-form-like DNA conformation in
yeast cells [18]. In mammalian cells, histones were
shown to bind more tightly to 5-bromouracil-substi-
tuted DNA in vitro than normal DNA containing thy-
mine [19,20].
In this study, we examined whether BrdU induces
the expression of genes associated with the destabiliza-
tion of nucleosome positioning with model plasmids
containing the BAR1 gene as a marker gene. We
showed that BrdU increases the transcription of
repressed genes by disruption of nucleosome position-
ing around their promoters. These data will facilitate
the understanding of the role of BrdU and AT-tracts

in the induction of particular genes.
Results
Construction of model plasmids
In yeast MATa cells, the a2 ⁄ Mcm1 repressor binds to
the a2 operator and acts to repress MATa cell-spe-
cific genes such as BAR1, STE2 and STE6 [21–25].
Previously, our high-resolution mapping of micrococ-
cal nuclease (MNase) cleavage sites has indicated that
nucleosomes were well positioned around the pro-
moter region of the genomic BAR1 gene [26] and
such nucleosome positioning was required for its full
repression [27]. Here, we constructed a plasmid pRS-
BAR1 that contains the BAR1 gene as a model sys-
tem (Fig. 1) to easily examine the role of BrdU in
gene expression and nucleosome positioning in vivo.
To further examine the effects of AT-tracts on nucle-
osome positioning, we inserted them into a KpnI site
adjacent to the TATA box. These plasmids were
introduced into a thymidine (dThd)-auxotrophic
Inserts at the
Kpn
Isite
pRS001-BAR1
pRS801-BAR1
:T
3
CCT
6
CT
5

GCT
5
CT
7
:A
34
BAR1
-UTR -UTR
pRS-BAR1
TATA
2op
mRNA
Kpn
I
+1130
––219 66 +95 +255
I
II
III IV
V
60
–
500
Xho
I
–––235
+420
–158
Eco
RV

+1667
Probe A Probe B
5′ 3′
BAR1
Fig. 1. Schematic representation of pRS-BAR1 and its derivatives. Plasmid pRS-BAR1 contains a genomic fragment spanning the BAR1
gene regulated by a2 ⁄ Mcm1 repressor. The a2 operator (a2 op), the TATA box (TATA) and the BAR1 coding sequence are denoted by
hatched, filled and dotted boxes, respectively. The transcriptional initiation site is indicated by a bent arrow. The site of KpnI, to which
AT-tracts are inserted, is indicated by a vertical line. The probes A and B are indicated by gray boxes. Positions of stable nucleosomes are
shown by shadowed ellipses with numbers.
Molecular mechanism of 5-bromodeoxyuridine K. Miki et al.
4540 FEBS Journal 277 (2010) 4539–4548 ª 2010 The Authors Journal compilation ª 2010 FEBS
mutant of yeast cells to ensure quantitative incorpora-
tion of BrdU into DNA [2].
Nucleosome positioning on pRS-BAR1
We first examined whether pRS-BAR1 forms a stable
and precise array of nucleosomes in dThd medium
using the indirect end-labeling method. DNA samples
were completely digested with XhoI after partially
digestion with MNase, and were subjected to Southern
blot analysis with probe A. Five MNase cleavage sites
indicated by *a, *b, *c, *d and *e were observed in the
naked DNA sample, and these sites were protected in
a chromatin sample of pRS-BAR1 (Fig. 2A, lanes 1
and 2). Also, the cleavage sites showed equal intervals
(140–150 bp) in the chromatin sample (Fig. 2A, lane
1). These results showed that at least five stable nucle-
osomes were precisely and stably formed from the a2
operator to the coding region of the BAR1 gene on
pRS-BAR1 in vivo (Fig. 1, bottom panel).
We then examined how BrdU affects nucleosome

positioning on the pRS-BAR1 plasmid. When the
MATa cells were cultured in BrdU medium, the two
bands (*b and *c) that were protected in the chromatin
sample prepared in dThd medium were evident in the
chromatin sample prepared in BrdU medium (Fig. 2A,
lanes 1 versus 3). Densitometrical measurement clearly
showed these differences (Fig. 2B). Consistently, the
band corresponding to the linker region between nucle-
osomes II and III became broader in the same sample
(Fig. 2A, lane 3). Because there was no significant dif-
ference in MNase cleavage patterns between naked
DNA samples prepared in dThd and BrdU medium,
BrdU did not affect the specificity and sensitivity to
MNase (Fig. 2A, lanes 2 versus 4). These results indi-
cate that BrdU destabilized nucleosome positioning.
We then examined the sequence in nucleosome II, and
found the following: the AT-content of the sequence
was 69%, whereas in other nucleosomes it was approx-
imately 60%, and A
n
,T
n
or (AT)
n
tracts (n ‡ 6) were
found in six sites in nucleosome II, but not in nucleo-
some III. Therefore, BrdU seemed to destabilize the
positioning of nucleosome II through AT-rich
sequences in nucleosome II, which then led to the
destabilization of nucleosome III. The same results

were obtained when nucleosome positioning was deter-
mined from the EcoRV site with probe B (Fig. 1, S1,
lanes 1–4).
Effects of AT-tracts on nucleosome positioning
We examined two derivatives of pRS-BAR1, pRS001-
BAR1 and pRS801-BAR1, containing T
3
CCT
6
CT
5
GCT
5
CT
7
and A
34
, respectively, in a promoter region
of the BAR1 gene (Fig. 1A). Insertion of these AT-
tracts did not affect MNase cleavage patterns in naked
DNA samples (Fig. 2A, even-numbered lanes). The
MNase cleavage pattern in a chromatin sample of
pRS001-BAR1 was similar to that in the chromatin
sample of pRS-BAR1 in dThd medium, although the
bands *c and *d were more evident in the former sam-
ple (Fig. 2A, lanes 1 versus 5). This suggests that the
positioning of nucleosomes I and II on pRS001-BAR1
was less stable than on pRS-BAR1 in dThd medium.
The bands *b, *c and *e in the chromatin sample of
pRS001-BAR1 were more marked in BrdU medium

than in dThd medium (Fig. 2A, B, lanes 5 versus 7),
although the MNase cleavage patterns in the naked
DNA sample were similar between dThd and BrdU
medium. Similar results were obtained when nucleo-
some positioning was mapped in the opposite direction
from the EcoRV site with probe B (Fig. S1, lanes 5–8).
We next analyzed the nucleosome positioning on
pRS801-BAR1. Interestingly, the MNase mapping of
nucleosomes with probe A showed that the three
bands *b to *d (especially band *c) were not protected
in a chromatin sample of pRS801-BAR1 (Fig. 2A, lane
9), indicating that insertion of A
34
disrupts nucleosome
positioning more remarkably than that of
T
3
CCT
6
CT
5
GCT
5
CT
7
. In BrdU medium, all of the
bands that corresponded to the linker regions were
eliminated and the bands *a, *b and *e were more
clearly detected than in dThd medium (Fig. 2B, lanes 9
versus 11). These results were also confirmed by indi-

rect end-labeling with probe B (Fig. S1, lanes 9–12).
Taken together, these observations indicate that
nucleosome positioning on pRS001- and pRS801-
BAR1 were modestly and almost completely disrupted,
respectively, by BrdU.
Effect of BrdU on BAR1 expression
We measured the expression of the BAR1 gene on the
model plasmids by Northern blot analysis. In dThd
medium, the BAR1 mRNA level for pRS-BAR1 was
30 times less in the MATa cells than in the MATa cells
(Fig. 3A). These results indicate that the episomal
BAR1 gene is regulated similarly to the genomic BAR1
gene. In the MATa cells, the mRNA level was 2.4
times higher in BrdU medium than in dThd medium
(Fig. 3B). Insertion of T
3
CCT
6
CT
5
GCT
5
CT
7
(pRS001-
BAR1) slightly increased the mRNA level in dThd
medium, but additionally increased it in BrdU medium
(Fig. 3B). On the other hand, insertion of A
34
(pRS801-BAR1) significantly increased the mRNA

level in dThd medium and markedly increased it in
K. Miki et al. Molecular mechanism of 5-bromodeoxyuridine
FEBS Journal 277 (2010) 4539–4548 ª 2010 The Authors Journal compilation ª 2010 FEBS 4541
a
b
c
d
a
b
c
d
2op
c
d
a
b
c
d
a
b
c
e
a
b
c
e
a
b
pRS-BAR1
CDCD

dThd BrdU
pRS001-BAR1
CD CD
dThd BrdU
pRS801-BAR1
CD C D
dThd BrdU
Lane 3 412 56 78 9 10 1112
M
2.0
1.5
1.0
0.5
(kb)
e
e
e
e
d
d
I
II
III
IV
V
A
Lane
1
3
5

7
9
11
B
2 op
IIIII IV VI
Naked DNA
e d c b a
Fig. 2. Nucleosome positioning on pRS-
BAR1 and its derivatives. Chromatin (indi-
cated by C) and naked DNA (indicated by D)
samples were prepared from cells transfect-
ed with the plasmid indicated and cultured
in dThd or BrdU medium as indicated. The
samples were partially digested with
5UÆmL
)1
(odd-numbered lanes) or
0.5 UÆmL
)1
(even-numbered lanes) MNase,
completely digested with XhoI, and sub-
jected to the indirect end-labeling analysis
with probe A as described in Materials and
Methods. At least three independent analy-
ses for each plasmid gave similar results.
(A) Autoradiography of MNase cleavage pat-
terns. DNA size markers (M), positions of
nucleosomes I–V and a2 operator (a2 op)
are shown to the left. Specific cleavage

sites on naked DNA samples are marked
with *a, *b, *c, *d and *e. Ellipses with
dotted lines indicate nucleosomes whose
positioning is unstable. Open stars on some
lanes denote bands that changed in BrdU
medium. The transcriptional start site and
the KpnI site are indicated by a bent arrow
and arrowheads, respectively. (B) Densito-
metric profiles of autoradiography. The odd-
numbered lanes in (A) were densitometrical-
ly scanned. The vertical dotted lines denote
the bands (*a, *b, *c, *d and *e) specific to
naked DNA. The positions of stable nucleo-
somes are shown at the top.
Molecular mechanism of 5-bromodeoxyuridine K. Miki et al.
4542 FEBS Journal 277 (2010) 4539–4548 ª 2010 The Authors Journal compilation ª 2010 FEBS
BrdU medium (Fig. 3B). We also examined the effect
of BrdU on pYBT1 containing the thymidine kinase
(TK) gene driven by the yeast constitutive ADH1 pro-
moter as a control plasmid. BrdU did not significantly
affect the expression of the gene (Fig. 3B). These
results show that the levels of expression of the BAR1
gene are parallel to those of the disruption of nucleo-
some positioning in dThd and BrdU medium.
To confirm that our experimental conditions can
induce specific genomic genes, we examined the expres-
sion of some genomic genes having an AT-tract on
their promoter regions. The DED1 gene, having
T
3

CCT
6
CT
5
GCT
5
CT
7
, and the MAK16 gene, having
T
24
, were significantly up-regulated by the addition of
BrdU (Fig. 4), suggesting that the mechanism found in
the episomal genes also operates in the genomic genes.
Effect of BrdU on a2/Mcm1 repressor–operator
complex
We examined whether the a2 ⁄ Mcm1 repressor changes
its binding to the a2 operator on pRS-BAR1 upon
incorporation of 5-bromouracil by an in vivo UV
photofootprinting assay. The a2 ⁄ Mcm1 operator has
numerous thymine bases necessary for recognition by
the repressor [28] and thus its binding to the repressor
may be disturbed by substitution of 5-bromouracil. In
the noncoding strand of a naked DNA sample, several
thymine dimers were found around the a2 operator
(Fig. 5A, lane 2), but three sites (marked with + in
Fig. 5A) were protected in chromatin samples
(Fig. 5A, lanes 1 and 3). Although slight differences in
the thymine dimers formed were observed in the naked
DNA samples containing thymine or 5-bromouracil

(Fig. 5A, lanes 2 versus 5), the three thymine dimers
were equally protected in the chromatin samples con-
taining thymine or 5-bromouracil (Fig. 5A, lanes 4 and
6). Similar results were obtained with the coding
strand (Fig. 5B). These results show that BrdU does
not weaken the formation of the a2 ⁄ Mcm1 repressor–
operator complex, excluding the possibility that the
DED1
MAK16
ACT1
dThd BrdU
Fig. 4. Northern blot analysis of genomic genes. Total RNA sam-
ples prepared from the MATa cells cultured in dThd or BrdU med-
ium were subject to Northern blot analysis with probes derived
from the genes indicated.
TB
pRS-BAR1
0
2
4
6
TB
pRS001-BAR1
##
TB
pRS801-BAR1
###
Relative mRNA level
TB
BAR1 TK

pYBT1
Relative BAR1 mRNA level
20
0
MAT
MAT
a
pRS-BAR1
10
30
AB
Fig. 3. Gene expression profiles of pRS-BAR1, its derivatives and a reference plasmid. (A) BAR1 mRNA levels in MATa and MATa cells.
Total RNA and DNA samples were prepared from the cell type indicated in dThd medium and subjected to Northern and Southern blot analy-
ses as described in Materials and Methods. BAR1 mRNA levels were expressed relative to actin mRNA levels after normalization by copy
numbers of plasmids. (B) Effects of BrdU on BAR1 and TK mRNA levels. Total RNA and DNA samples were prepared from the MATa cells
transfected with the plasmid indicated and cultured in dThd (T) or BrdU medium (B), and processed as in (A). ***P < 0.001 compared with
the values in dThd medium.
##
P < 0.01 and
###
P < 0.001 compared with the values of pRS-BAR1 in dThd medium. Histograms represent
means ± standard error. At least four independent analyses carried out for each plasmid gave similar results.
K. Miki et al. Molecular mechanism of 5-bromodeoxyuridine
FEBS Journal 277 (2010) 4539–4548 ª 2010 The Authors Journal compilation ª 2010 FEBS 4543
disruption of nucleosome positioning by BrdU is due
to a decrease in the formation of the a2 ⁄ Mcm1 com-
plex [18].
Effect of BrdU on pRS-BAR1 lacking the promoter
activity
We addressed whether the above changes in nucleo-

some positioning are affected by the promoter activity
of the BAR1 gene, because transcription factors can
affect nucleosome positioning. We constructed a plas-
mid, pRS-DTA ⁄ BAR1, in which the TATA box of
BAR1 was disrupted (Fig. 6A) [23]. With pRS-DTA ⁄
BAR1 we were able to determine the change in nucleo-
some positioning without the effect of the expression
of BAR1 (Fig. 6B).
Disruption of the TATA box caused the disappear-
ance of band *d in naked samples of pRS-DTA ⁄ BAR1
prepared in dThd and BrdU medium. Band *d corre-
sponds to the nuclease hypersensitive site located at
the TATA box on pRS-BAR1 (Fig. 2A, lanes 2 and 4
versus Fig. 6C, lanes 2 and 4) as described in the pre-
vious reports by Shimizu et al. [26] and Cooper et al.
[23]. In dThd medium, the MNase cleavage pattern in
a chromatin sample of pRS-DTA ⁄ BAR1 showed well-
ordered nucleosome positioning (Fig. 6C, lane 1),
which was identical to that on the chromatin sample
of pRS-BAR1 (Fig. 2A, lane 1), except for the pres-
ence of band *d.
In BrdU medium, the bands corresponding to the
linker regions between nucleosomes I–III were
broader, and the two bands (*b and *c) that were pro-
tected in the chromatin sample prepared in dThd med-
ium became more evident (Fig. 6C, lane 3). When
dThd
CDC
BrdU
CDC

+
+
+
+
+
+
+
+
+
A
dThd
CDC
BrdU
CDC
+
+
+
+
+
+
B
123 456 789 101112
3′-CGTACATTAATGGCATTTTCCTTTAATGTAC-5′
3′-GTACATTAAAGGAAAATGCCATTAATGTACG-5′
Fig. 5. In vivo UV photofootprinting of a2 operator on pRS-BAR1.
Intact cells (indicated by C) and naked DNA (indicated by D) were
irradiated with UV at dosages of 250 mJ (lanes 1, 4, 7 and 10),
500 mJ (lanes 3, 6, 9 and 12) and 60 mJ (lanes 2, 5, 8 and 11). UV
photoproducts were analyzed using primer extension mapping on
the noncoding (A) and the coding (B) strands as described in Mate-

rials and Methods. The a2 operator sequence is shown to the left
of each panel. The thymine bases protected from UV damage are
indicated by +.
a
b
c
e
I
II
III
IV
V
a
b
c
e
BrdU
–137 –125
AC
CAGTATAAAAGTG
TATA
pRS-BAR1
CAGTGGATCCGTG
Bam
H I
pRS-ΔTA/BAR1
pRS-ΔTA/BAR1
dThd
2 op
Lane 3 41 2

B
ACT1
BAR1
BAR1
TA/
BAR1
pRS-
C DC D
Fig. 6. Nucleosome positioning on p0RS-
DTA ⁄ BAR1. (A) Sequences of the TATA box
and disrupted TATA box of the BAR1 pro-
moter. (B) Northern blot analysis of the
BAR1 gene. (C) Autoradiography of MNase
cleavage patterns. Nucleosome positioning
was analyzed as in Fig. 2.
Molecular mechanism of 5-bromodeoxyuridine K. Miki et al.
4544 FEBS Journal 277 (2010) 4539–4548 ª 2010 The Authors Journal compilation ª 2010 FEBS
cultured in dThd or BrdU medium, the overall band
patterns were almost identical in the chromatin sam-
ples between pRS-BAR1 and pRS-DTA ⁄ BAR1, except
for the presence of band *d. These results indicate that
BrdU changes nucleosome positioning in the absence
of the transcription factors involved. In line with this
observation, we have shown that incorporation of
BrdU into DNA converts the DNA structure into an
unusual conformation [18]. It is thus reasonable to
suggest that such a structural change in DNA induced
by BrdU affects nucleosome positioning and results in
altered gene expression at particular regions.
Discussion

We were able to show a positive correlation between
gene expression and the disruption of nucleosome posi-
tioning in yeast cells harboring minichromosomes and
cultured with BrdU as the only source of thymine. To
validate this observation, BrdU must directly affect
DNA structure, but not interactions between DNA
and DNA-binding proteins, to induce a change in
nucleosome positioning. In our model plasmids used,
the binding of a2 ⁄ Mcm1 repressor to its legitimate
binding site has a critical role in the formation of sta-
bly ordered nucleosomes. As expected, BrdU did not
significantly affect the formation of the a2 ⁄ Mcm1
repressor–operator complex [18]. In support of this,
expression of the genomic STE2 gene regulated by the
a2 ⁄ Mcm1 complex was not affected by the addition of
BrdU (data not shown). In addition, some DNA-bind-
ing proteins and enzymes examined to date cannot
functionally distinguish between 5-bromouracil and
thymine on DNA [29–32]. These results suggest that
an unusual DNA conformation induced by BrdU is
the primary cause of altered nucleosome positioning.
In fact, we have demonstrated that the incorporation
of 5-bromouracil into DNA reduces the bending of
DNA [11] and converts to A-form-like DNA or a rigid
DNA structure [18]. However, a possibility cannot be
ruled out that a change in interactions between 5-bro-
mouracil-substituted DNA and specific proteins may
additionally affect nucleosome positioning.
We showed here that levels of BAR1 expression are
parallel to those of the destabilization of nucleosome

positioning with the use of model plasmids. In our pre-
vious study employing different minichromosomes [18],
disruption of nucleosome positioning by BrdU was
shown to depend on the length and sequence specificity
of AT-tracts located at particular sites of minichromo-
somes. In this study, destabilization of nucleosome
positioning by BrdU did not require the presence of
the promoter or expression of the BAR1 marker gene.
As shown in the Results, AT-tracts alone can destabi-
lize nucleosome positioning in dThd medium. How-
ever, not all AT-tracts have the ability to induce the
destabilization of nucleosome positioning or expression
of the marker genes.
The nucleosome disruption did not lead to full
expression of the BAR1 gene. This can be explained by
the absence of an activator function of Mcm1 in
MATa cells. Mcm1 acts as an activator of a-cell-spe-
cific genes in MATa cells, whereas it acts as a repressor
in MATa cells. Taken together, nucleosome position-
ing seems to be sufficient for full repression of genes
[27]. Also, disruption of it seems to be sufficient for
full derepression of genes if necessary transcription fac-
tors are supplied.
Can our findings obtained with the episomal genes
apply to genomic loci? The episomal BAR1 gene was
shown to behave similarly to the genomic BAR1 gene
[27] when A
34
was inserted into their promoters. Simi-
larly, the genomic DED1 gene was significantly up-reg-

ulated by BrdU, as the episomal BAR1 gene has the
AT-tract derived from the DED1 promoter. Further-
more, the genomic MAK16 gene having T
24
on its pro-
moter region was also up-regulated by BrdU. These
results suggest that episomal and genomic genes
behave similarly, and prove to be useful in studying
gene regulation controlled by the higher-order struc-
ture of chromatin.
However, the presence of AT-tracts does not always
affect the expression of their adjacent genes if the
genes have a strong promoter. Their promoter regions
seem to be reluctant to form a stable nucleosome
structure. In these genes, BrdU does not seem to cause
an additional change in the nucleosome structure sur-
rounding the genes and increase their expression. For
example, the expression of the TK gene driven by the
yeast constitutive ADH1 promoter on pYBT1 was not
affected by BrdU (Fig. 3B). In the case of TALS-GFP,
EGFP expression is driven by the ADH1 promoter
(Fig. S2). pOM801-GFP, having A
34
inserted in the
ADH1 promoter region (Fig. S2), showed significantly
increased promoter activity even in dThd medium
(Fig. S4). However, BrdU did not further increase the
expression of GFP (Fig. S4) or the state of the already
opened nucleosome structure in pOM801-GFP
(Fig. S3). These results support our hypothesis that

BrdU induces the expression of genes through the dis-
ruption of nucleosome positioning on their promoter
regions.
In mammalian cells, BrdU is thought to induce the
expression of genes when AT-tracts are located adja-
cent to their promoters, similar to yeast systems [11].
In contrast to yeast genes, most of the mammalian
K. Miki et al. Molecular mechanism of 5-bromodeoxyuridine
FEBS Journal 277 (2010) 4539–4548 ª 2010 The Authors Journal compilation ª 2010 FEBS 4545
genes are silenced during lifetime, embedded in con-
densed chromatin. As described previously, BrdU can
restore the expression of silenced genes [9,10], and
BrdU-responsive genes are frequently located on inac-
tive chromatin regions, such as AT-rich Giemsa-dark
bands of human chromosomes [11,12]. In this context,
BrdU seems to disrupt nucleosome positioning around
AT-rich condensed chromatin and results in the induc-
tion of the expression of silenced or repressed genes.
Finally, the data of this study may lead to a new
understanding of the molecular mechanism of BrdU.
They may answer the new and old question of why
BrdU modulates the expression of particular genes
associated with cellular differentiation and senescence.
Materials and methods
Plasmids and yeast strains
To construct BAR1 expression plasmid pRS-BAR1, the
)500 to +2064 sequence containing the promoter with a
KpnI site ()158 to )154) [27], a coding sequence and
300 bp of 3¢-UTR of the BAR1 gene, were cloned into the
XhoI-SacI site of pRS424DKpnI in which one KpnI site was

filled in. Plasmid pRS-BAR1 derivatives were constructed
by inserting oligonucleotides into the KpnI site of pRS-
BAR1 to yield pRS001-BAR1 (T
3
CCT
6
CT
5
GCT
5
CT
7
) and
pRS801-BAR1 (A
34
).
To disrupt a TATA box in pRS-BAR1, the sequence of
the TATA box at )134 (
TATAAAA) was changed to a
sequence (T
GGATCC) that contained a BamHI site by
amplifying a KpnI-BglII sequence of pRS-BAR1 with the
following two primers: 5¢ -TATTGGTACCGTGTGTTTT
TTGATAACAGT
GGATCCGTG-3¢ and 5¢-GTGGAAGA
TCTATGCTCATTATAAGTACTC-3¢. The amplified
sequence was digested with KpnI and BglII, and cloned into
the KpnI–BglII site of pRS-BAR1 to yield pRS-DTA ⁄ -
BAR1, which lacks a functional TATA box.
These plasmids were introduced into the yeast dThd-

autotrophic strains YKH2 (MAT a ura2-52 trp1 his3 leu2
cdc21::LEU2 pYBT1) or YKH4 (MATa ura2-52 trp1 his3
leu2 cdc21::LEU2 pYBT1) established as described previ-
ously [2]. Plasmid pYBT1 contains the herpes simplex virus
TK gene driven by the yeast ADH1 promoter.
Chromatin preparation and nuclease digestion
Yeast cells harboring plasmids were selected in SC medium
(2% glucose, 0.67% yeast nitrogen base without amino
acids) supplemented with appropriate amino acids (except
for tryptophan) and 1 mm dThd. Cells were grown in
30 mL of medium containing 1 mm dThd or BrdU at 30 °C
for 15 h to an optimal density of 0.6–1.0 at 600 nm. Chro-
matin and naked DNA samples were prepared according to
the method of Balasubramanian & Morse [33]. Each sample
was digested with MNase (Takara, Kyoto, Japan) at 37 °C
for 10 min. The reactions were initiated by the addition of
0.15% Nonidet P-40, and halted by the addition of SDS
and proteinase K. Samples were purified with a phe-
nol ⁄ chloroform extraction and ethanol precipitation.
Indirect end-labeling
DNA samples were completely digested with XhoIor
EcoRV together with RNase A, run on a 1.4% agarose gel
and transferred on to a Nylon membrane (Biodyne B, Pall,
Port Washington, NY, USA) followed by cross-linking with
UV light (Stratalimker 2400, Stratagene, La Jolla, CA,
USA). The membrane was incubated at 65 °C for 16 h in
hybridization solution (0.5 m Na-Pi, 1 mm ETDA and 7%
SDS] containing a probe labeled with [a-
32
P] dCTP using a

random-primed DNA labeling kit (Mega-prime, Amersham,
Piscataway, NJ, USA). The XhoI–MspI fragment of pRS-
BAR1 was used as a probe to detect nucleosomes in sam-
ples digested with XhoI. Likewise, a sequence amplified with
the primers 5¢- ATCTTATAATTATCGAGATCG- 3¢ and 5¢-
AAGTGTTCCACTG TCTAGTTTG- 3¢ from pRS-BAR1 was
used as a probe to detect nucleosomes in samples digested with
EcoRV. After washing, the membrane was subjected to autora-
diography and densitometric analysis using an image analyzer
FLA-5000 (FUJIFILM, T okyo, Japan).
Gene expression analysis
Total RNA samples were prepared, subjected to electropho-
resis on a 1% formaldehyde agarose gel and blotted on to
a Nylon membrane as described previously [34]. To deter-
mine copy numbers of plasmids in yeast cells, DNA sam-
ples were prepared as described above. The DNA samples
were digested with XhoI to linearize and run on a 1% aga-
rose gel and blotted on to a Nylon membrane. The mem-
brane was hybridized with appropriate
32
P-labeled BAR1,
TK, DED1, MAK16 or ACT1 coding sequences as probes.
After washing, the membrane was subjected to autoradiog-
raphy followed by imaging analysis as described above.
In vivo UV photofootprinting
An in vivo UV photofootprinting assay was performed as
described previously [35,36]. Yeast cells harboring a plasmid
were grown in SC (Trp
)
) medium containing 1 mm dThd or

BrdU at 30 °C for 15 h, irradiated with UV at 254 nm using
a Stratalinker 2400, and DNA samples purified. As controls,
DNA samples were purified from unirradiated cells and irra-
diated similarly. Sites and levels of the formation of UV
photoproducts were determined by primer extension map-
ping, using the IR-Dye-800-labeled MS-9 primer ()387 to
)358 of the BAR1 coding strand) and MS-10 primer ()87 to
Molecular mechanism of 5-bromodeoxyuridine K. Miki et al.
4546 FEBS Journal 277 (2010) 4539–4548 ª 2010 The Authors Journal compilation ª 2010 FEBS
)116 of the BAR1 noncoding strand) with a LI-COR 4000L
DNA sequencer (LI-COR, Lincoln, NE, USA) [26,37].
Acknowledgements
This work was supported in part by Grants-in-Aid for
Scientific Research from the Ministry of Education,
Science and Culture of Japan.
References
1 Michishita E, Nakabayashi K, Suzuki T, Kaul SC,
Ogino H, Fujii M, Mitsui Y & Ayusawa D (1999)
5-Bromodeoxyuridine induces senescence-like phenomena
in mammalian cells regardless of cell type or species.
J Biochem (Tokyo) 126, 1052–1059.
2 Fujii M, Ito H, Hasegawa T, Suzuki T, Adachi N &
Ayusawa D (2002) 5-Bromo-2¢-deoxyuridine efficiently
suppresses division potential of the yeast Saccharomyces
cerevisiae. Biosci Biotechnol Biochem 66, 906–909.
3 Wilt FH & Anderson M (1972) The action of 5-bromo-
deoxyuridine on differentiation. Dev Biol 28 , 443–447.
4 Weintraub H (1973) Size of the BUdR sensitive targets
for differentiation. Nat New Biol 244, 142–143.
5 Goz B (1977) The effects of incorporation of 5-haloge-

nated deoxyuridines into the DNA of eukaryotic cells.
Pharmacol Rev 29, 249–272.
6 Suzuki T, Minagawa S, Michishita E, Ogino H,
Fujii M, Mitsui Y & Ayusawa D (2001) Induction of
senescence-associated genes by 5-bromodeoxyuridine in
HeLa cells. Exp Gerontol 36, 465–474.
7 Minagawa S, Nakabayashi K, Fujii M, Scherer SW and
Ayusawa D (2005) Early BrdU-responsive genes consti-
tute a novel class of senescence-associated genes in
human cells. Exp Cell Res 304, 552–558.
8 Zakharov AF, Baranovskaya LI, Ibraimov AI,
Benjusch VA, Demintseva VS & Oblapenko NG (1974)
Differential spiralization along mammalian mitotic
chromosomes. II. 5-bromodeoxyuridine and 5-bro-
modeoxycytidine-revealed differentiation in human
chromosomes. Chromosoma 44, 343–359.
9 Suzuki T, Yaginuma M, Oishi T, Michishita E,
Ogino H, Fujii M & Ayusawa D (2001) 5-Bromodeoxy-
uridine suppresses position effect variegation of
transgenes in HeLa cells. Exp Cell Res 266, 53–63.
10 Fan J, Kodama E, Koh Y, Nakao M & Matsuoka M
(2005) Halogenated thymidine analogues restore the
expression of silenced genes without demethylation.
Cancer Res 65, 6927–6933.
11 Suzuki T, Michishita E, Ogino H, Fujii M &
Ayusawa D (2002) Synergistic induction of the
senescence-associated genes by 5-bromodeoxyuridine
and AT-binding ligands in HeLa cells. Exp Cell Res
276, 174–184.
12 Minagawa S, Nakabayashi K, Fujii M, Scherer SW &

Ayusawa D (2004) Functional and chromosomal clus-
tering of genes responsive to 5-bromodeoxyuridine in
human cells. Exp Gerontol 39, 1069–1078.
13 Satou W, Suzuki T, Noguchi T, Ogino H, Fujii M &
Ayusawa D (2004) AT-hook proteins stimulate induc-
tion of senescence markers triggered by 5-bromodeoxy-
uridine in mammalian cells. Exp Gerontol 39, 173–179.
14 Kingston RE, Bunker CA & Imbalzano AN (1996)
Repression and activation by multiprotein complexes
that alter chromatin structure. Genes Dev 10, 905–920.
15 Wolffe AP & Hayes JJ (1999) Chromatin disruption
and modification. Nucleic Acids Res 27, 711–720.
16 Morse RH (2007) Transcription factor access to pro-
moter elements. J Cell Biochem 102, 560–570.
17 Shimizu M, Mori T, Sakurai T & Shindo H (2000)
Destabilization of nucleosomes by an unusual DNA
conformation adopted by poly(dA) ⁄ poly(dT) tracts in
vivo. EMBO J 19, 3358–3365.
18 Miki K, Shimizu M, Fujii M, Hossain MN &
Ayusawa D (2008) 5-Bromouracil disrupts nucleosome
positioning by inducing A-form-like DNA conforma-
tion in yeast cells. Biochem Biophys Res Commun 368,
662–669.
19 Lin S, Lin D & Riggs AD (1976) Histones bind more
tightly to bromodeoxyuridine-substituted DNA than to
normal DNA. Nucleic Acids Res 3, 2183–2191.
20 Fasy TM, Cullen BR, Luk D & Bick MD (1980) Stud-
ies on the enhanced interaction of halodeoxyuridine-
substituted DNAs with H1 histones and other polypep-
tides. J Biol Chem 255, 1380–1387.

21 Roth SY, Shimizu M, Johnson L, Grunstein M &
Simpson RT (1992) Stable nucleosome positioning and
complete repression by the yeast alpha 2 repressor are
disrupted by amino-terminal mutations in histone H4.
Genes Dev 6, 411–425.
22 Ganter B, Tan S & Richmond TJ (1993) Genomic foot-
printing of the promoter regions of STE2 and STE3
genes in the yeast Saccharomyces cerevisiae. J Mol Biol
234, 975–987.
23 Cooper JP, Roth SY & Simpson RT (1994) The global
transcriptional regulators, SSN6 and TUP1, play dis-
tinct roles in the establishment of a repressive chroma-
tin structure. Genes Dev 8, 1400–1410.
24 Davie JK, Trumbly RJ & Dent SY (2002) Histone-
dependent association of Tup1-Ssn6 with repressed
genes in vivo. Mol Cell Biol 22, 693–703.
25 Ruiz C, Escribano V, Morgado E, Molina M &
Mazon MJ (2003) Cell-type-dependent repression of
yeast a-specific genes requires Itc1p, a subunit of the
Isw2p-Itc1p chromatin remodelling complex.
Microbiology 149, 341–351.
26 Shimizu M, Roth SY, Szent-Gyorgyi C & Simpson RT
(1991) Nucleosomes are positioned with base pair
K. Miki et al. Molecular mechanism of 5-bromodeoxyuridine
FEBS Journal 277 (2010) 4539–4548 ª 2010 The Authors Journal compilation ª 2010 FEBS 4547
precision adjacent to the alpha 2 operator in
Saccharomyces cerevisiae. EMBO J 10, 3033–3041.
27 Morohashi N, Yamamoto Y, Kuwana S, Morita W,
Shindo H, Mitchell AP & Shimizu M (2006) Effect of
sequence-directed nucleosome disruption on cell-type-

specific repression by alpha2 ⁄ Mcm1 in the yeast gen-
ome. Eukaryot Cell 5, 1925–1933.
28 Zhong H, McCord R & Vershon AK (1999) Identifica-
tion of target sites of the alpha2-Mcm1 repressor
complex in the yeast genome. Genome Res 9, 1040–
1047.
29 Goeddel DV, Yansura DG, Winston C &
Caruthers MH (1978) Studies on gene control regions.
VII. Effect of 5-bromouracil-substituted lac operators
on the lac operator–lac repressor interaction. J Mol Biol
123, 661–687.
30 Brennan CA, Van Cleve MD & Gumport RI (1986)
The effects of base analogue substitutions on the cleav-
age by the EcoRI restriction endonuclease of octade-
oxyribonucleotides containing modified EcoRI
recognition sequences. J Biol Chem 261, 7270–7278.
31 Hayakawa T, Ono A & Ueda T (1988) Synthesis of
decadeoxyribonucleotides containing 5-modified uracils
and their interactions with restriction endonucleases Bgl
II, Sau 3AI and Mbo I (nucleosides and nucleotides
82). Nucleic Acids Res 16, 4761–4776.
32 Risse G, Jooss K, Neuberg M, Bruller HJ & Muller R
(1989) Asymmetrical recognition of the palindromic
AP1 binding site (TRE) by Fos protein complexes.
EMBO J 8, 3825–3832.
33 Balasubramanian B & Morse RH (1999) Binding of
Gal4p and bicoid to nucleosomal sites in yeast in
the absence of replication. Mol Cell Biol 19, 2977–
2985.
34 Takayama S, Fujii M, Kurosawa A, Adachi N & Ayus-

awa D (2007) Overexpression of HAM1 gene detoxifies
5-bromodeoxyuridine in the yeast Saccharomyces cerevi-
siae. Curr Genet 52, 203–211.
35 Murphy MR, Shimizu M, Roth SY, Dranginis AM &
Simpson RT (1993) DNA-protein interactions at the
S. cerevisiae alpha 2 operator in vivo. Nucleic Acids Res
21, 3295–3300.
36 Shimizu M & Mitchell AP (2003) Hap1p photofoot-
printing as an in vivo assay of repression mechanism in
Saccharomyces cerevisiae. Methods Enzymol 370, 479–
487.
37 Morohashi N, Nakajima K, Kuwana S, Tachiwana H,
Kurumizaka H & Shimizu M (2008) In vivo and in vi-
tro footprinting of nucleosomes and transcriptional acti-
vators using an infrared-fluorescence DNA sequencer.
Biol Pharm Bull 31, 187–192.
Supporting information
The following supplementary material is available:
Fig. S1. Nucleosome positioning on pRS-BAR1 and
its derivatives.
Fig. S2. Schematic representation of TALS-GFP and
its derivative.
Fig. S3. Nucleosome positioning on TALS-GFP and
pOM801-GFP.
Fig. S4. Effects of BrdU on EGFP mRNA levels.
This supplementary material can be found in the
online version of this article.
Please note: As a service to our authors and readers,
this journal provides supporting information supplied
by the authors. Such materials are peer-reviewed and

may be re-organized for online delivery, but are not
copy-edited or typeset. Technical support issues arising
from supporting information (other than missing files)
should be addressed to the authors.
Molecular mechanism of 5-bromodeoxyuridine K. Miki et al.
4548 FEBS Journal 277 (2010) 4539–4548 ª 2010 The Authors Journal compilation ª 2010 FEBS