Tài liệu Báo cáo khoa học: The stereochemistry of benzo[a]pyrene-2¢-deoxyguanosine adducts affects DNA methylation by SssI and HhaI DNA methyltransferases pptx

The stereochemistry of benzo[a]pyrene-2¢-deoxyguanosine
adducts affects DNA methylation by SssI and HhaI DNA
methyltransferases
Oksana M. Subach
1
, Diana V. Maltseva
1
, Anant Shastry
2
, Alexander Kolbanovskiy
2
,
Saulius Klimas
ˇ
auskas
3
, Nicholas E. Geacintov
2
and Elizaveta S. Gromova
1
1 Chemistry Department, Moscow State University, Russia
2 Department of Chemistry, New York University, NY, USA
3 Laboratory of Biological DNA Modification, Institute of Biotechnology, Vilnius, Lithuania
The polycyclic aromatic hydrocarbons are a well-
known class of ubiquitous environmental pollutants
which are generated by incomplete combustion of
organic matter. These compounds require metabolic
activation to highly reactive diol epoxides to elicit their
detrimental genotoxic effects [1]. Benzo[a]pyrene
(B[a]P), one of the most widely studied polycyclic
aromatic hydrocarbons, is metabolically activated

in vivo to the highly mutagenic [2] and tumorigenic [3]
(+)-7R,8S-diol 9S,10R-epoxide of benzo[a]pyrene
Keywords
benzo[a]pyrene-2¢-deoxyguanosine adducts;
DNA methyltansferases; environmental
pollutants; stereochemistry
Correspondence
E. S. Gromova, Chemistry Department,
Moscow State University, Moscow,
119992, Russia
Fax: +7495 939 31 81
Tel: +7495 939 31 44
E-mail:
(Received 19 July 2006, revised 19 January
2007, accepted 21 February 2007)
doi:10.1111/j.1742-4658.2007.05754.x
The biologically most significant genotoxic metabolite of the environmental
pollutant benzo[a]pyrene (B[a]P), (+)-7R,8S-diol 9S,10R-epoxide, reacts
chemically with guanine in DNA, resulting in the predominant formation
of (+)-trans-B[a]P-N
2
-dG and, to a lesser extent, (+)-cis-B[a]P-N
2
-dG
adducts. Here, we compare the effects of the adduct stereochemistry and
conformation on the methylation of cytosine catalyzed by two purified
prokaryotic DNA methyltransferases (MTases), SssI and HhaI, with the
lesions positioned within or adjacent to their CG and GCGC recognition
sites, respectively. The fluorescence properties of the pyrenyl residues of the
(+)-cis-B[a]P-N

2
-dG and (+)-trans-B[a]P-N
2
-dG adducts in complexes with
MTases are enhanced, but to different extents, indicating that aromatic
B[a]P residues are positioned in different microenvironments in the DNA–
protein complexes. We have previously shown that the (+)-trans-isomeric
adduct inhibits both the binding and methylating efficiencies (k
cat
) of both
MTases [Subach OM, Baskunov VB, Darii MV, Maltseva DV, Alexandrov
DA, Kirsanova OV, Kolbanovskiy A, Kolbanovskiy M, Johnson F,
Bonala R, et al. (2006) Biochemistry 45, 6142–6159]. Here we show that the
stereoisomeric (+)-cis-B[a]P-N
2
-dG lesion has only a minimal effect on the
binding of these MTases and on k
cat
. The minor-groove (+)-trans adduct
interferes with the formation of the normal DNA minor-groove contacts
with the catalytic loop of the MTases. However, the intercalated base-
displaced (+)-cis adduct does not interfere with the minor-groove DNA–
catalytic loop contacts, allowing near-normal binding of the MTases and
undiminished k
cat
values.
Abbreviations
AdoHcy, S-adenosyl-
L-homocysteine; AdoMet, S-adenosyl-L-methionine; B[a]P, benzo[a]pyrene; B[a]PDE, r7,t8-dihydroxy-t9,10-epoxy-
7,8,9,10-tetrahydrobenzo[a]pyrene; B[a]P-DNA, DNA containing benzo[a]pyrene; C5 MTase, C5-cytosine DNA methyltransferase; EMSA,

electrophoretic mobility shift assay; k
cat
, multiple turnover rate constant; K
d
, dissociation constant; M.SssI, SssI DNA methyltransferase;
M.HhaI, HhaI DNA methyltransferase; MTase, DNA methyltransferase; V
0
, initial rate of methylation; V
max
, maximal rate of methylation.
FEBS Journal 274 (2007) 2121–2134 ª 2007 The Authors Journal compilation ª 2007 FEBS 2121
{(+)-B[a]PDE} [1]. This metabolite reacts with DNA
predominantly at the N
2
-exocyclic amino groups of
guanine [4,5] via trans or cis opening of the epoxide
ring to form the (+)-trans-B[a]P-N
2
-dG and (+)-cis-
B[a]P-N
2
-dG adducts (Fig. 1). The relative yields of
the two stereoisomeric adducts are generally higher in
the case of the (+)-trans-B[a]P-N
2
-dG adduct (in some
cases more than 90%) than in the case of the (+)-cis-
B[a]P-N
2
-dG adduct (up to 13%) [4,6,7]. Although the

enantiomer (–)-7S,8R-diol 9R,10S diol epoxide of
B[a]P is not formed in eukaryotic cells [8], it is often
used in structure–function studies of B[a]P-N
2
-dG
adducts because of the different conformational char-
acteristics of the (–)-trans-B[a]P-N
2
-dG and (+)-trans-
B[a]P-N
2
-dG adducts [6].
The structures of (+)-trans-B[a]P-N
2
-dG and (+)-cis-
B[a]P-N
2
-dG adducts in dsDNA are very different
from one another, the former being characterized by
an external minor-groove conformation and the latter
by a base-displaced intercalative conformation [6,9,10].
The different structural characteristics have a pro-
nounced effect on the cellular processing of these stereo-
isomeric DNA adducts. First, both prokaryotic and
eukaryotic nucleotide excision repair systems eliminate
the (+)-cis-B[a]P-N
2
-dG adducts more efficiently than
the (+)-trans-B[a]P-N
2

-dG adducts [11,12]. The lesions
that escape repair can influence DNA replication, tran-
scription, and the interaction of different proteins with
DNA. All four B[a]P-N
2
-dG adducts inhibit DNA repli-
cation [13,14]. The successful, although error-prone,
translesional synthesis past both stereoisomeric adducts
has been reported [13,15,16]. The fidelity of trans-
lesional synthesis depends on adduct stereochemistry,
nucleotide sequence context, and the DNA polymerase
[15,16]. In the case of DNA transcription, T7 RNA
polymerase is blocked more efficiently by the (+)-trans-
B[a]P-N
2
-dG adduct than by the (+)-cis-B[a]P-N
2
-dG
adduct [17]. The binding of the transcription factor Sp 1
to B[a]PDE-modified DNA is highly dependent on the
B[a]P-N
2
-dG conformation [18], whereas no apparent
differences in the binding affinities of the Ap 1 tran-
scription factor to DNA containing different stereoiso-
meric B[a]P-N
2
-dG adducts was observed [19]. The
B[a]P-DNA adducts also affect the function of human
topoisomerase I by alteration of DNA cleavage patterns

[20]. The greatest disturbance of DNA cleavage is
caused by the (+)-trans-B[a]P-N
2
-dG and (+)-cis-
B[a]P-N
2
-dG adducts [20,21]. In the present work, we
explored the hypothesis that DNA methylation is
dependent on the absolute configurations and confor-
mations of (+)-trans-B[a]P-N
2
-dG and (+)-cis-B[a]P-
N
2
-dG lesions.
DNA methylation plays an important role in dif-
ferent cellular processes such as regulation of tran-
scription, cell development, and chromatin structure
[22,23]. Mammalian genomes are methylated at cer-
tain CpG sites, resulting in different patterns of
DNA methylation [22,23]. Disruption of methylation
patterns can lead to cancer [24–28]. In eukaryotes,
methylation of CpG sites is carried out by several
C5-cytosine DNA methyltransferases (C5 MTases;
EC 2.1.1.37). Prokaryotic C5 MTases are good mod-
els of biological methylation because they share with
mammalian C5 MTases a number of conserved
amino-acid motifs that have structural roles and are
involved in catalysis [29]. The prokaryotic C5 MTases
SssI and HhaI transfer a methyl group to the C5

position of the target cytosine (
C) in their CG and
G
CGC recognition sites, respectively. The M.SssI has
substrate specificity identical with that of the mam-
malian MTases [30].
HO
HO
OH
N
NH
O
N
N
R
NH
HO
HO
OH
N
NH
O
N
N
R
NH
X
+
- (+)-cis-B[a]P-N
2

-dG
Y
+
- (+)-trans-B[a]P-N
2
-dG
Fig. 1. Chemical structures of the (+)-trans-B[a]P-N
2
-dG and (+)-cis-
B[a]P-N
2
-dG adducts.
Stereochemistry of B[a]P-N
2
-dG affects methylation O. M. Subach et al.
2122 FEBS Journal 274 (2007) 2121–2134 ª 2007 The Authors Journal compilation ª 2007 FEBS
B[a]P-N
2
-dG lesions are formed efficiently at the
guanine residue in CpG sequence contexts [31] that are
recognition sites of mammalian MTases. The efficiency
of such damage is enhanced in the presence of m
5
dC
instead of dC in 5¢-CpG targets [31–33]. Such damage
in the promoter region of a gene may disturb the nor-
mal functioning of MTases and change the genomic
methylation pattern. It has previously been found that
the concentrations of methylated cytosines in the
DNA of mammalian cells treated with racemic

(+ ⁄ –)-B[a]PDE are lower than normal [34,35]. In these
earlier investigations, only the overall concentrations
of B[a]P-N
2
-dG adducts were compared with the levels
of DNA methylation, and thus the effect of the abso-
lute configurations and conformations of the adducts
on DNA methylation was not evaluated. More
recently, the effect of stereochemically distinct (+)-
trans-B[a]P-N
2
-dG and (–)-trans-B[a]P-N
2
-dG adducts
on DNA methylation by the prokaryotic MTases Eco-
RII [36], SssI, and HhaI [37] was examined. In most
cases, the methylation efficiency of oligodeoxynucleo-
tide duplexes containing trans adducts in MTase recog-
nition sites by these MTases was diminished. These
effects were attributed to the conformation of the
trans-B[a]P-N
2
-dG adducts in the minor groove of
B-DNA [10], which interfere with the formation of the
normal and critical minor-groove MTase-DNA con-
tacts [37].
Because of the markedly different (+)-trans-B[a]P-
N
2
-dG and (+)-cis-B[a]P-N

2
-dG adduct conforma-
tions, it is of structural interest to compare the effects
of these conformations on DNA methylation. In this
work, the effect of the intercalated [9] (+)-cis-anti-
B[a]P-N
2
-dG adduct on the DNA binding and catalytic
activity of SssI and HhaI was examined and compared
with the effects of the minor-groove (+)-trans-B[a]
P-N
2
-dG adduct [37]. The hypothesis was tested that
the (+)-cis-B[a]P-N
2
-dG adducts, because of their
intercalative conformations, inhibit methylation to a
lesser extent because the DNA minor groove remains
available for interaction with the critical amino-acid
groups of the MTases. Using biochemical and spectro-
scopic methods, we show here that the (+)-cis-anti-
B[a]P-N
2
-dG adducts indeed do not significantly
inhibit methylation, demonstrating that the stereo-
chemistry of B[a]P metabolite-derived DNA adducts
can affect this potentially important epigenetic mech-
anism of cancer initiation [1,38].
Results
The (+)-cis-B[a]P-N

2
-dG lesions (X
+
) were site-specifi-
cally incorporated into the single-stranded oligonucleo-
tides shown in Table 1. The corresponding duplexes
are shown in Table 2. The X
+
residues were intro-
duced into the overlapping recognition sites of both
M.SssI (
CpG) and M.HhaI (GCGC) on either the
Table 1. Oligodeoxynucleotide sequences synthesized. M, m
5
dC;
X
+
, (+)-cis-B[a]P-N
2
-dG; Y
+
, (+)-trans-B[a]P-N
2
-dG.
GCG
ref
5¢-GAGCCAAGCGCACTCTGA
CGM
ref
5¢-TCAGAGTGMGCTTGGCTC

GCG 5¢-CACCCTTGCGCTCTCTCA
CGC 5¢-TGAGAGAGCGCAAGGGTG
CGM 5¢-TGAGAGAGMGCAAGGGTG
X
+
CG 5¢-CACCCTTX
+
CGCTCTCTCA
GCX
+
5¢-CACCCTTGCX
+
CTCTCTCA
Y
+
CG 5¢-GAGCCAAY
+
CGCACTCTGA
GCY
+
5¢-GAGCCAAGCY
+
CACTCTGA
Table 2. Properties of the oligodeoxynucleotide duplexes containing (+)-cis-B[a]P-N
2
-dG adduct as substrates of M.HhaI and M.SssI. The tar-
get dC are underlined. M.SssI ⁄ M.HhaI sites are in bold. The other designations are as in Table 1.
Designation DNA duplex
M.HhaI M.SssI
K

d
(pM) k
cat
(min
)1
) K
d
(nM) k
cat
(min
)1
)
G
CG ⁄ CGC5¢-CACCCTTGCGCTCTCTCA
3¢-GTGGGAA
CGCGAGAGAGT
52.7 ± 1.3 3.4 ± 0.4 6.8 ± 1.2 0.9 ± 0.4
X
+
CG ⁄ CGC5¢-CACCCTTX
+
CGCTCTCTCA
3¢-GTGGGAA
CGCGAGAGAGT
41.6 ± 1.2 2.7 ± 0.3 3.0 ± 0.3 0.7 ± 0.3
G
CX
+
⁄ CGC5¢-CACCCTTGCX
+

CTCTCTCA
3¢-GTGGGAA
CGCGAGAGAGT
42.8 ± 13.9 2.0 ± 0.2 3.8 ± 0.6 0.7 ± 0.2
G
CG ⁄ CGM 5¢-CACCCTTGCGCTCTCTCA
3¢-GTGGGAA
CGMGAGAGAGT
13.0 ± 3.9 2.1 ± 0.3 4.1 ± 0.8 0.4 ± 0.2
X
+
CG ⁄ CGM 5¢-CACCCTTX
+
CGCTCTCTCA
3¢-GTGGGAA
C GMGAGAGAGT
61 ± 14 1.5 ± 0.3 4.2 ± 0.5 0.13 ± 0.05
G
CX
+
⁄ CGM 5¢-CACCCTTGCX
+
CTCTCTCA
3¢-GTGGGAA
CGM GAGAGAGT
13.6 ± 2.5 1.7 ± 0.2 1.9 ± 0.2 0.5 ± 0.1
O. M. Subach et al. Stereochemistry of B[a]P-N
2
-dG affects methylation
FEBS Journal 274 (2007) 2121–2134 ª 2007 The Authors Journal compilation ª 2007 FEBS 2123

5¢- or the 3¢-side of the target dC residue. In hemi-
methylated oligonucleotide duplexes (X
+
CG ⁄ CGM
and G
CX
+
⁄ CGM), 5-methylcytosine was introduced
into one of the strands of the recognition site instead
of the target cytosine. The melting curves of the hemi-
methylated X
+
CG ⁄ CGM and GCX
+
⁄ CGM duplexes
containing (+)-cis-B[a]P-N
2
-dG were cooperative with
the melting temperature ranging from 57 to 61 °C,
being 4–8 °C lower than that of the unmodified
G
CG ⁄ CGM duplex (data not shown). Therefore, the
(+)-cis-B[a]P-N
2
-dG lesions destabilize the 18-mer
duplexes.
Effects of M.HhaI and M.SssI binding on the
fluorescence of the (+)-trans-B[a]P-N
2
-dG

and (+)-cis-B[a]P-N
2
-dG adducts
To examine how the stereochemical and conforma-
tional features of the (+)-cis-B[a]P-N
2
-dG (X
+
) and
(+)-trans-B[a]P-N
2
-dG (Y
+
) adducts are affected by
the binding of the MTases, the fluorescence properties
of the pyrenyl residues were examined when the oligode-
oxynucleotide duplexes were titrated with various
amounts of M.HhaI or M.SssI. The duplexes containing
Y
+
are defined in Table 3. The Y
+
residues were intro-
duced into the overlapping recognition sites of both
M.SssI (
CpG) and M.HhaI (GCGC) on either the 5¢-side
(Y
+
CG ⁄ CGM) or the 3¢-side (GCY
+

⁄ CGM) of the
target dC residue or distant from it [Y
+
(N)
4
CG ⁄
C(N)
4
GM]. The emission spectra of X
+
CG ⁄ CGM,
G
CX
+
⁄ CGM, Y
+
CG ⁄ CGM and GCY
+
⁄ CGM dup-
lexes alone or in complexes with MTases exhibit the
usual broad maxima at 384 and 404 nm (Fig. 2A), con-
sistent with those previously reported [39,40].
The binding of M.HhaI to the X
+
CG ⁄ CGM and
G
CX
+
⁄ CGM duplexes containing (+)-cis-B[a]P-N
2

-dG
adduct in the HhaI recognition site results in a similar
3.5-fold and fourfold increase in the fluorescence
Table 3. Oligodeoxynucleotide duplexes containing (+)-trans-B[a]
P-N
2
-dG adduct. N is any nucleotide residue. The other designa-
tions are as in Tables 1 and 2.
Designation DNA duplex
G
CG ⁄ CGM 5¢-GAGCCAAGCGCACTCTGA
3¢-CTCGGTT
CGMGTGAGACT
Y
+
CG ⁄ CGM 5¢-GAGCCAAY
+
CGCACTCTGA
3¢-CTCGGTT
C GMGTGAGACT
G
CY
+
⁄ CGM 5¢-GAGCCAAGCY
+
CACTCTGA
3¢-CTCGGTT
CGM GTGAGACT
Y
+

(N)
4
CG ⁄ C(N)
4
GM 5¢-GCTY
+
GTGGCGTAGGC
3¢-CGAC CACC
GMATCCG
Fig. 2. Fluorescence titration of DNA containing (+)-cis-B[a]P-N
2
-dG
(X
+
) or (+)-trans-B[a]P-N
2
-dG (Y
+
) adducts with M.HhaI or M.SssI.
(A) Typical fluorescence emission spectra of the M.HhaI•B[a]P-
DNA•AdoHcy complexes and the free B[a]P-DNA duplexes; 500 n
M
GCX
+
⁄ CGM or GCY
+
⁄ CGM duplexes were incubated with 875 nM
M.HhaI in the presence of 0.1 mM AdoHcy in buffer D. The fluores-
cence excitation wavelength was 350 nm. (B) 500 n
M GCX

+
⁄ CGM
(s), G
CY
+
⁄ CGM (r), and Y
+
CG ⁄ CGM (j), or 200 nM of X
+
CG ⁄
CGM (n) were titrated with M.HhaI in buffer D at 25 °C and then
the emission at 384 nm was measured with excitation at 350 nm.
(C) 100 n
M GCX
+
⁄ CGM (s), X
+
CG ⁄ CGM (n), Y
+
CG ⁄ CGM (j),
G
CY
+
⁄ CGM (r)orY
+
(N)
4
CG ⁄ C(N)
4
GM (·) were titrated with

M.SssI in buffer B at 25 °C. The excitation and emission wave-
lengths are the same as in (B).
Stereochemistry of B[a]P-N
2
-dG affects methylation O. M. Subach et al.
2124 FEBS Journal 274 (2007) 2121–2134 ª 2007 The Authors Journal compilation ª 2007 FEBS
emission intensity of the aromatic pyrenyl residue,
respectively (Fig. 2B). However, in the case of the
(+)-trans-B[a]P-N
2
-dG adducts, the increase in fluores-
cence intensity is much more pronounced upon the
binding of M.HhaI to the Y
+
CG ⁄ CGM and
G
CY
+
⁄ CGM duplexes (by factors of 30 and 20,
respectively). The fluorescence enhancement in practice
does not depend on the position of the (+)- trans or
(+)-cis adduct.
The binding of M.SssI to the X
+
CG ⁄ CGM and
GCX
+
⁄ CGM duplexes containing (+)-cis-B[a]P-N
2
-

dG adduct in the recognition site or in the flanking
sequence leads to a 1.6–1.7-fold increase in the fluores-
cence emission intensity of the B[a]P residue (Fig. 2C).
The fluorescence intensity of the B[a]P residue increa-
ses by factors of 2.8–16 upon the binding of M.SssI to
the Y
+
CG ⁄ CGM, GCY
+
⁄ CGM and Y
+
(N)
4
CG ⁄
C(N)
4
GM duplexes containing (+)-trans-B[a]P-N
2
-dG
adduct. Thus, the fluorescence enhancement depends
on the position of the (+)-trans adduct. When M.SssI
binds to G
CY
+
⁄ CGM duplexes containing (+)-trans-
B[a]P-N
2
-dG adducts in the recognition site, the fluor-
escence intensity increases by a factor of 16. Upon
M.SssI binding to Y

+
CG ⁄ CGM or Y
+
(N)
4
CG ⁄
C(N)
4
GM duplexes with the Y
+
adducts flanking the
CpG recognition site on the 5¢ side, or positioned four
nucleotide residues distant from the CpG site on the 5¢
side, respectively, the fluorescence emission yield
increases by a factor of only  3.
Overall, these results indicate that the fluorescence
properties of B[a]P-DNA in the complexes with MTases
strongly depend on the (+)-cis-B[a]P-N
2
-dG and
(+)-trans-B[a]P-N
2
-dG adduct stereochemistry and on
the location of the adduct in either the MTase recogni-
tion site or the flanking sequences.
Binding of M.SssI and M.HhaI to
oligodeoxynucleotide duplexes containing
the (+)-cis-B[a]P-N
2
-dG adduct

The binding of M.SssI and M.HhaI to the oligodeoxy-
nucleotide duplexes was performed in the presence of
the cofactor analog S -adenosyl- l-homocysteine (Ado-
Hcy). In the case of C5 MTases, AdoHcy facilitates
the formation of specific complexes with DNA [41,42].
To determine the K
d
values of the M.SssI or M.HhaI
complexes with the oligodeoxynucleotide duplexes con-
taining the (+)-cis-B[a]P-N
2
-dG adduct, we used a
competition equilibrium binding assay. In these com-
petition experiments, unlabeled B[a]PDE-modified and
32
P-labeled GCG ⁄ CGM
ref
duplexes were mixed before
the addition of MTase. The formation of the com-
plexes of M.SssI and M.HhaI with DNA was moni-
tored by electrophoretic mobility shift assay (EMSA)
(Fig. 3A,B). The competition curves (Fig. 3C,D) are
characteristic of equilibrium competition processes
[43].
In the case of M.HhaI, the K
d
values for the
B[a]PDE-modified X
+
CG ⁄ CGC and GCX

+
⁄ CGC
duplexes are 1.2–1.3 times smaller than for the
unmodified parent G
CG ⁄ CGC duplex, and the K
d
value for the GCX
+
⁄ CGM duplex is about the same
as for the parent G
CG ⁄ CGM duplex (Table 2). How-
ever, a 4.7-fold reduction in the binding affinity was
observed in the case of binding of the M.HhaI with
the X
+
CG ⁄ CGM duplex containing the (+)-cis-B[a]P-
N
2
-dG adduct on the 5¢-side of the target dC residue.
The binding of M.SssI to X
+
CG ⁄ CGC and GCX
+
⁄
CG
C is favored by a factor of  2 relative to the
unmodified G
CG ⁄ CGC duplex. In the case of
M.SssI•X
+

CG ⁄ CGM•AdoHcy and M.SssI•GCX
+
⁄
CGM•AdoHcy complexes, the K
d
values are about the
same as the K
d
of the M.SssI•GCG ⁄ CGM•AdoHcy
complex. Thus, for both enzymes, the K
d
values of
the ternary MTase•(unmethylated cis-B[a]P-DNA)•
AdoHcy and MTase•(hemimethylated cis-B[a]P-DNA)•
AdoHcy complexes are comparable to the K
d
values of
the ternary complexes of MTases with the correspond-
ing unmodified unmethylated (G
CG ⁄ CGC) or hemi-
methylated (G
CG ⁄ CGM) duplexes.
Steady-state kinetics of methylation of
oligodeoxynucleotide duplexes containing
(+)-cis-B[a]P-N
2
-dG adduct by M.SssI and M.HhaI
The rates of methylation of the X
+
CG ⁄ CGC, GCX

+
⁄
CG
C, X
+
CG ⁄ CGM and GCX
+
⁄ CGM duplexes by
SssI and HhaI MTases were determined under steady-
state conditions (Fig. 4), and the k
cat
values were cal-
culated (Table 2). The k
cat
values of methylation of
B[a]PDE-modified unmethylated X
+
CG ⁄ CGC and
G
CX
+
⁄ CGC duplexes or hemimethylated X
+
CG ⁄
CGM and G
CX
+
⁄ CGM duplexes by M.HhaI were
decreased by factors of 1.2–1.7 in comparison with the
k

cat
values of the corresponding unmodified duplexes.
In the case of M.SssI, the largest effect on DNA
methylation was a 3.1-fold decrease in k
cat
for the
hemimethylated X
+
CG ⁄ CGM duplex containing (+)-
cis-B[a]P-N
2
-dG on the 5¢-side of the target dC residue.
The k
cat
value for the hemimethylated duplex
G
CX
+
⁄ CGM was about the same as that for the
G
CG ⁄ CGM duplex. The k
cat
values determined for
the unmethylated X
+
CG ⁄ CGC and GCX
+
⁄ CGC
duplexes were only 1.3 times smaller than k
cat

for the
G
CG ⁄ CGC duplex. In summary, the presence of the
(+)-cis-B[a]P-N
2
-dG adduct practically does not affect
O. M. Subach et al. Stereochemistry of B[a]P-N
2
-dG affects methylation
FEBS Journal 274 (2007) 2121–2134 ª 2007 The Authors Journal compilation ª 2007 FEBS 2125
DNA methylation rates catalyzed by either M.SssI or
M.HhaI. The extent of the methylation is almost
independent of the position of the damaged guanine
residue X
+
.
Discussion
The goal of our study was to elucidate the effect of
stereochemistry and adduct conformation of (+)-cis-
B[a]P-N
2
-dG and (+)-trans-B[a]P-N
2
-dG adducts on
DNA methylation by prokaryotic MTases SssI and
HhaI. The conformations of the stereoisomeric B[a]
P-N
2
-dG adducts have been investigated in detail
(summarized in [6]). Briefly, the bulky pyrene-like aro-

matic ring system in the (+)-trans-B[a]P-N
2
-dG
adducts is positioned in the minor groove and is 5¢-
directed relative to the modified guanine residue with
all base pairs intact, including the modified G*•C base
pair [10]. In contrast, the (+)- cis-B[a]P-N
2
-dG adduct
assumes an intercalated base-displaced adduct confor-
mation with the modified dG residue and the partner
base dC in the opposite strand displaced into the
minor and major grooves, respectively [9]. Molecular
views of these (+)-cis and (+)-trans adduct conforma-
tions are shown in Fig. 5A.
The structure of the MTase–DNA complexes con-
taining B[a]P-N
2
-dG adducts in the MTase recognition
sites has not been studied. According to the available
crystal structures of complexes of M.HhaI with
unmodified DNA and AdoHcy(AdoMet) [44], M.HhaI
consists of two domains, the large domain containing
the S-adenosyl-l-methionine (AdoMet) binding site
and the catalytic center, and the small domain contain-
ing the target recognition domain (Fig. 5B). The DNA
molecule is located in the cleft formed between the two
domains with the major groove facing the small
domain and the minor groove facing the large domain.
Before methylation, the target dC residue flips out of

the DNA double helix into the M.HhaI active-site
pocket [45]. The flipped out cytosine forms contacts
with the catalytic loop of the enzyme from the DNA
minor-groove side. The contacts of the amino-acid
A
B
CD
Fig. 3. Equilibrium competitive binding of
duplexes containing (+)-cis-B[a]P-N
2
-dG
adduct and unmodified duplexes, to the
MTases SssI and HhaI. Autoradiographs of
EMSA of competitive binding of the unlabe-
led B[a]PDE-modified X
+
CG ⁄ CGM and
32
P-labeled GCG ⁄ CGM
ref
duplexes to
M.HhaI (A) and the unlabeled X
+
CG ⁄ CGC
and
32
P-labeled GCG ⁄ CGM
ref
duplexes to
M.SssI (B). The concentrations of compet-

itor X
+
CG ⁄ CGM were 0, 0.5, 1, 5, 10, 40,
80, 120, 200 n
M in lanes 1–9, respectively
(A), and 0, 10, 20, 50, 100, 200, 300, 400,
500 n
M in lanes 1–9, respectively (B). Equi-
librium competition curves for complexes of
M.HhaI (C) and M.SssI (D) with
32
P-labeled
duplex GCG ⁄ CGM
ref
in the presence of
increasing concentrations of the competitor
duplexes G
CG ⁄ CGC(j), X
+
CG ⁄ CGC(m),
G
CX
+
⁄ CGC(d), GCG ⁄ CGM (h),
X
+
CG ⁄ CGM (n)orGCX
+
⁄ CGM (s). The
relative fraction of bound

32
P-labeled DNA
(R) is the ratio of the fraction of bound
32
P-labeled DNA in the presence of the
competitor DNA (cpm
bound
⁄ cpm
total
) to the
fraction of bound
32
P-labeled DNA in the
absence of the competitor DNA
(cpm
bound
°⁄cpm
total
°).
Stereochemistry of B[a]P-N
2
-dG affects methylation O. M. Subach et al.
2126 FEBS Journal 274 (2007) 2121–2134 ª 2007 The Authors Journal compilation ª 2007 FEBS
residues within the catalytic domain of the M.HhaI
with the DNA minor groove play an important role in
the methylation reaction. The contacts of the M.HhaI
small domain with the DNA major groove are respon-
sible for the specific MTase-DNA recognition. Similar
structural features are likely to be present in the tern-
ary M.SssI•DNA•AdoHcy complex, as suggested by a

recent modeling study [46].
Fluorescence properties
The fluorescence of the B[a]P residues is quenched
by factors of 100–200 in (+)-trans-B[a]P-N
2
-dG and
(+)-cis-B[a]P-N
2
-dG mononucleoside adducts [47] by
a solvent-dependent proton-coupled electron-transfer
mechanism [48]. The fluorescence lifetimes are 1.4 ±
0.1 and 0.71 ± 0.2 ns, respectively, in aqueous solu-
tions [47], but are longer in oligonucleotide duplexes.
For example, the fluorescence decay profiles of the
(+)-cis-B[a]P-N
2
-dG and (+)-trans-B[a]P-N
2
-dG with-
in oligonucleotide duplexes are well described by the
sums of three exponential decay components with
mean lifetimes of 4.0 ± 0.2 and 2.4 ± 0.2 ns
(Y. Tang, A. Durandin, and N. E. Geacintov, unpub-
lished). Thus, in the absence of protein, the fluores-
cence characteristics of the (+)-trans-B[a]P-N
2
-dG and
(+)-cis-B[a]P-N
2
-dG adducts are not too different, a

conclusion that is supported by the similar fluorescence
yields of the two types of adduct in the absence of
protein (Fig. 2A).
The fluorescence properties of the pyrenyl residues
in the B[a]P-N
2
-dG adducts are known to be sensitive
to their microenvironments [49,50], particularly in
complexes with proteins [51,52]. To elucidate possible
differences in the microenvironments of the (+)-cis
and (+)-trans adducts within MTase•B[a]P-DNA com-
plexes, we examined changes in fluorescence intensities
of the two stereoisomeric B[a]P residues in the
duplexes depicted in Tables 2 and 3 when the two dif-
ferent MTases were added to aqueous solutions of
these duplexes. The enhancement in the fluorescence
yield is substantially greater when MTase binds to
oligonucleotide duplexes containing the (+)-trans
adduct than the (+)-cis adduct (Fig. 2). We observed
a 3.5–4-fold fluorescence increase upon M.HhaI bind-
ing to duplexes containing the (+)-cis-B[a]P-N
2
-dG
adduct and a 20–30-fold fluorescence enhancement
upon the binding of M.HhaI to duplexes containing
the stereoisomeric (+)-trans-B[a]P-N
2
-dG adduct. A
1.6–1.7-fold fluorescence increase occurred upon
M.SssI binding to the duplexes containing (+)-cis-

B[a]P-N
2
-dG adduct and a 2.8–16-fold upon binding
of M.SssI to the duplexes containing (+)-trans-B[a]P-
N
2
-dG adduct. Therefore, the larger enhancement of
the fluorescence yield of the (+)-trans adduct relative
to the (+)-cis adduct reflects the difference in the local
microenvironments of the two aromatic pyrenyl resi-
dues in the protein–DNA complexes.
It is known from previous studies that the fluores-
cence yields of (+)-trans-B[a]P-N
2
-dG mononucleoside
adducts are dramatically increased as the concentration
of organic solvents is increased in aqueous mixtures
[53]. The differences in the fluorescence yields upon
formation of the M.HhaI•G
CY
+
⁄ CGM•AdoHcy
and M.HhaI•Y
+
CG ⁄ CGM•AdoHcy complexes sug-
gest that the (+)-trans adducts are situated in a
more hydrophobic environment in the protein com-
plexes than in aqueous solution in the absence of
A
B

Fig. 4. Steady-state kinetic analysis of methylation of unmodified
and (+)-cis-B[a]P-N
2
-dG adduct-containing oligodeoxynucleotide
duplexes by M.SssI (A) and M.HhaI (B). The concentrations of
G
CG ⁄ CGC, X
+
CG ⁄ CGC, GCX
+
⁄ CGC and GCG ⁄ CGM duplexes were
250 n
M, and the concentrations of X
+
CG ⁄ CGM and GCX
+
⁄ CGM
duplexes were 350 n
M. Designations of unmodified and B[a]PDE-
modified duplexes on the curves are as shown in Fig. 3.
O. M. Subach et al. Stereochemistry of B[a]P-N
2
-dG affects methylation
FEBS Journal 274 (2007) 2121–2134 ª 2007 The Authors Journal compilation ª 2007 FEBS 2127
protein (Fig. 2). Such enhancements are consistent
with the effects observed in the case of the nucleoside
(+)-trans-B[a]P-N
2
-dG adducts when water is replaced
by more hydrophobic organic solvents [53]. Our hypo-

thesis is that the change in the hydrophobicity of the
local environment upon protein binding is less pro-
nounced in the case of the (+)-cis-B[a]P-N
2
-dG adduct
than in the case of the (+)-trans-B[a]P-N
2
-dG adduct.
Thus, the different microenvironment of the pyrenyl
residue in the (+)-cis-B[a]P-N
2
-dG adduct and (+)-
trans-B[a]P-N
2
-dG adduct in MTase•B[a]P-DNA com-
plexes is revealed by fluorescence studies.
It has been postulated that the flipping or extrusion
of the target base from the DNA duplex is an import-
ant intermediate step in DNA methylation catalyzed
by C5 MTases [54]. We postulated that the fluores-
cence of the pyrenyl residue in the B[a ]P-N
2
-dG
adducts would be particularly sensitive to changes in
the microenvironment when this adduct is flanked by a
target cytosine that undergoes flipping in the MTase–
DNA complexes. In accordance with this, the depend-
ence of the fluorescence of the (+)-trans adduct on its
position relative to the target dC was revealed in the
case of the formation of the complexes of M.SssI

with G
CY
+
⁄ CGM, Y
+
CG ⁄ CGM and Y
+
(N)
4
CG ⁄
C(N)
4
GM duplexes (Fig. 2C). It is well established
that neighboring bases in their normal positions in
DNA quench the fluorescence of (+)-trans-B[a]P-N
2
-
dG introduced into oligodeoxynucleotide duplexes
[39,55,56]. We suggest that the observed large increase
in fluorescence in the case of the complex of M.SssI
with the G
CY
+
⁄ CGM duplex containing the
(+)-trans adduct in the CpG site may be caused by
diminished quenching by the target dC residue that is
flipped in the MTase–DNA complex. When the B[a]P
residue is separated by four nucleotides from the
target dC residue in the Y
+

(N)
4
CG ⁄ C(N)
4
GM dup-
lex, the fluorescence enhancement upon formation
of the M.SssI–DNA complex is significantly smaller
(Fig. 2C). In the case of the Y
+
CG ⁄ CGM duplex,
when the B[a]P aromatic ring system is out of the
CpG site but near the target dC, the fluorescence
A
B
Fig. 5. (A) Conformations of the B[a]PDE-
modified duplexes containing the (+)-trans-
B[a]P-N
2
-dG and (+)-cis-B[a]P-N
2
-dG adducts
obtained by NMR methods and adapted
from [62] with permission of the American
Chemical Society. (B) Three-dimensional
structure of the ternary complex of M.HhaI
with the 12-mer duplex containing GCGC
and the cofactor analog AdoHcy derived
from the RCSB Protein Data Bank (3mht
[63]). The catalytic loop, the flipped out cyto-
sine, and AdoHcy are depicted in dark grey.

The enzyme is shown in the ribbon repre-
sentation. DNA and AdoHcy are shown in
the stick representation.
Stereochemistry of B[a]P-N
2
-dG affects methylation O. M. Subach et al.
2128 FEBS Journal 274 (2007) 2121–2134 ª 2007 The Authors Journal compilation ª 2007 FEBS
enhancement is also small. We showed previously that
methylation of this duplex was essentially inhibited,
even under single-turnover conditions for 2 h, and it
was assumed that the flipping of the target base was
impeded [37]. Therefore, in this case there was prob-
ably no effect of the flipping of the target cytosine on
the B[a]P fluorescence. Overall, the changes in fluores-
cence intensities are clearly due to changes in the local
microenvironment of the B[a]P residues in the DNA
duplexes, and are consistent with a base-flipping model
of the dC target residue.
Binding and methylation studies
The interactions of M.HhaI and M.SssI with DNA
containing site-specifically positioned (+)-trans-B[a]P-
N
2
-dG adduct have recently been investigated [37]. The
K
d
and k
cat
values for the (+)-cis-B[a]P-N
2

-dG and
(+)-trans-B[a]P-N
2
-dG adducts in different sequence
contexts are compared with one another in Fig. 6. The
minor-groove position of the (+)-trans-B[a]P-N
2
-dG
adduct did not significantly affect M.SssI binding to
DNA, but reduced M.HhaI binding by 1–2 orders of
magnitude (Fig. 6). Therefore, the bulky B[a]P residue
positioned in the DNA minor groove severely inhibits
DNA binding to M.HhaI by perturbing the minor-
groove DNA–M.HhaI contacts and does not signifi-
cantly influence DNA binding to M.SssI [37]. Our
observations indicate that the introduction of the (+)-
cis-B[a]P-N
2
-dG into DNA does not cause any signifi-
cant changes in K
d
for either M.SssI or M.HhaI
(Table 2, Fig. 6). This observation can be accounted
for by the intercalative conformation of the B[a]P resi-
dues in the (+)-cis adducts which interferes less signifi-
cantly with DNA–protein interactions on either side of
the modified base pair. Thus, the stereochemistry of
the B[a]P-N
2
-dG adducts in DNA does not influence

DNA binding in the case of M.SssI, but, in contrast,
does affect DNA binding in the case of M.HhaI.
The (+)-trans-B[a]P-N
2
-dG adduct greatly dimin-
ishes the methylating efficiency of hemimethylated (by
factors of 185–5000) and unmethylated (by factors of
1.3–9) DNA catalyzed by either M.SssI or M.HhaI [37]
when the (+)-trans-B[a]P-N
2
-dG adduct is positioned
5¢ to the target dC base (Fig. 6). On the other hand,
the (+)-cis-B[a]P-N
2
-dG adduct has practically no
effect on the methylation rate constant, k
cat
, in either
case (Table 2, Fig. 6). These differences are a direct
consequence of the strikingly different conformational
characteristics of the stereoisomeric (+)-cis-B[a]P-N
2
-
dG and (+)-trans-B[a]P-N
2
-dG adducts. It is likely
that, in the (+)-trans-B[a]P-N
2
-dG adduct, the bulky
B[a]P residue situated in the minor groove interferes

with the interactions between the catalytic loops of
SssI and HhaI MTases and the minor groove of the ol-
igodeoxynucleotide duplexes [37]. However, in the case
of the (+)-cis-B[a]P-N
2
-dG adduct in the unbound
duplex, the B[a]P residue is intercalated into the DNA
A
B
Fig. 6. Bar graphs representing relative K
d
(K
rel
d
) and k
cat
(k
rel
cat
) values for binding and
methylation of DNA containing (+)-cis-B[a]P-
N
2
-dG and (+)-trans-B[a]P-N
2
-dG adducts by
M.SssI (A) and M.HhaI (B). The K
rel
d
and k

rel
cat
values for duplexes containing the (+)-cis-
B[a]P-N
2
-dG adduct were calculated relative
to the canonical, unmodified duplex
G
CG ⁄ CGM from the data presented in
Table 2. The K
rel
d
and k
rel
cat
values for
duplexes containing the (+)-trans-B[a]P-N
2
-
dG adduct were calculated in a similar way
from the data presented in [37]. G* is (+)-
cis-B[a]P-N
2
-dG (X
+
) or (+)-trans-B[a]P-N
2
-dG
(Y
+

). The target dC residue is underlined.
O. M. Subach et al. Stereochemistry of B[a]P-N
2
-dG affects methylation
FEBS Journal 274 (2007) 2121–2134 ª 2007 The Authors Journal compilation ª 2007 FEBS 2129
helix, and the modified dG residue is displaced into
the minor groove. These findings suggest that the
B[a]P aromatic ring system remains stacked between
neighboring base pairs, thus exerting relatively minor
effects on K
d
and k
cat
. In this model, the bulky B[a]P
residue does not significantly disturb the contacts
between the M.HhaI (or M.SssI) catalytic loops
with the minor groove of the oligodeoxynucleotide
duplexes.
Relative to the unmodified G
CG ⁄ CGM duplex, a
small decrease in the efficiency of methylation by
M.SssI of the X
+
CG ⁄ CGM duplex is observed when
the (+)-cis adduct X
+
is positioned on the 5¢-side of
the target dC residue (Fig. 6, Table 2). On the other
hand, k
cat

remains unchanged when the (+)-cis adduct
in the G
CX
+
⁄ CGM duplex is positioned on the 3¢-side
of the target dC residue. In the case of M.HhaI, the
k
cat
values are practically unchanged in the presence of
the (+)- cis adduct in both hemimethylated duplexes
(Fig. 6).
Conclusions
In contrast with the (+)-trans-B[a]P-N
2
-dG adduct
[37], the introduction of the stereoisomeric (+)-cis-
B[a]P-N
2
-dG adduct into the DNA recognition sites of
the prokaryotic MTases M.HhaI and M.SssI does not
have any significant effect on DNA methylation rates.
This difference may be associated with the intercalative
conformation of the (+)-cis adduct and the minor-
groove conformation of the (+)-trans adduct, the lat-
ter interfering with interactions of the catalytic loops
of the MTases and the minor groove of DNA. In
accordance with this hypothesis, the fluorescence prop-
erties of the pyrenyl residues of the (+)-cis-B[a]P-N
2
-

dG or (+)-trans-B[a]P-N
2
-dG adduct in complexes
with MTases are enhanced, but to different extents,
indicating that aromatic B[a]P residues are positioned
in different microenvironments in these DNA–protein
complexes. Such effects of adduct stereochemistry on
hypomethylation may also exist in the case of mamma-
lian MTases, and these possibilities are being investi-
gated in our laboratory.
Experimental procedures
Chemicals and enzymes
AdoMet and AdoHcy were purchased from Sigma (St
Louis, MO, USA). [CH
3
-
3
H]AdoMet (77 CiÆmmol
)1
,
13 lm) was from Amersham Biosciences (Little Chalfont,
UK). [c-
32
P]ATP (1000 CiÆmmol
)1
) was bought from Izotop
(Obninsk, Russia). M.HhaI (4.4 mgÆmL
)1
) was prepared as
described previously [57]. Also we used His

6
-tagged M.SssI
(6.7 lm). To obtain His
6
-tagged M.SssI, an appropriate
hybrid plasmid was produced [58] using the vector pCAL7
provided by New England BioLabs (Beverly, MA, USA).
The kinetic parameters determined with wild type or His
6
-
tagged M.SssI were practically identical in value. The
MTases were found to be homogeneous on 12% polyacryl-
amide gels in the presence of 0.1% SDS. T4 polynucleotide
kinase was obtained from MBI Fermentas (Vilnius, Lithu-
ania). Buffers A–F were prepared using Milli-Q water:
A, 10 mm Tris•HCl (pH 7.9) ⁄ 50 mm NaCl; B, buffer A
containing 1 mm dithiothreitol; C, buffer B, containing
0.1 mgÆmL
)1
acetylated BSA; D, 50 mm Tris•HCl
(pH 7.5) ⁄ 50 mm NaCl ⁄ 10 mm EDTA ⁄ 5mm 2-mercaptoeth-
anol; E, 50 mm Tris•HCl (pH 7.5) ⁄ 50 mm NaCl ⁄ 10 mm
EDTA ⁄ 5mm 2-mercaptoethanol ⁄ 0.2 mgÆmL
)1
acetylated
BSA; F, 50 mm Tris•H
3
BO
3
(pH 8.3) ⁄ 2mm EDTA.

Oligodeoxynucleotides
The sequences of the oligodeoxynucleotides used are sum-
marized in Table 1. GCG
ref
, CGM
ref
, GCG, CGC and
CGM were purchased from IDT (Coralville, IA, USA) and
Syntol (Moscow, Russia).
Y
+
CG and GCY
+
oligodeoxynucleotides containing a
single (+)-trans-B[a]P-N
2
-dG adduct were obtained as des-
cribed [49]. The site-specifically modified X
+
CG and
GCX
+
oligodeoxynucleotides containing a single (+)-cis-
B[a]P-N
2
-dG lesion were obtained by treatment of GCG
with racemic B[a]PDE solution using previously described
methods [59]. The (+)-trans-B[a]P-N
2
-dG, (–)-trans-B[a]

P-N
2
-dG and (+)-cis-B[a]P-N
2
-dG adducts at the 18-mer
oligodeoxynucleotide level were separated and purified by
reverse-phase HPLC on an X Terra C18 column (Waters,
Milford, MA, USA) [59].
All oligodeoxynucleotides were further purified by elec-
trophoresis on denaturing 20% polyacrylamide gels and
desalted by passing the solutions through C18 September-
Pack cartridges (Waters). The sequences were labeled by
the standard
32
P-5¢-phosphorylation of oligodeoxynucleo-
tides using T4 polynucleotide kinase and [c-
32
P]ATP.
Oligodeoxynucleotide concentrations were estimated spec-
trophotometrically. The absorption coefficients of unmodi-
fied and B[a]PDE-modified oligonucleotides were calculated
as described [36].
Fluorescence measurements
The fluorescence of the X
+
CG ⁄ CGM, GCX
+
⁄ CGM,
Y
+

CG ⁄ CGM, GCY
+
⁄ CGM and Y
+
(N)
4
CG ⁄ C (N)
4
GM
duplexes was recorded on a Perkin–Elmer spectrofluorime-
ter with slit widths of 5–10 nm for excitation and 3–5 nm
for the emission monochromator. All titrations were per-
formed in a micro quartz cuvette (10 mm · 10 mm,
100 lL; Starna Cells, Atascadero, CA, USA). X
+
CG ⁄
Stereochemistry of B[a]P-N
2
-dG affects methylation O. M. Subach et al.
2130 FEBS Journal 274 (2007) 2121–2134 ª 2007 The Authors Journal compilation ª 2007 FEBS
CGM, GCX
+
⁄ CGM, Y
+
CG ⁄ CGM, G CY
+
⁄ CGM or
Y
+
(N)

4
CG ⁄ C (N)
4
GM duplexes were incubated with var-
ious concentrations of M.HhaI or M.SssI in buffer D or B
in the presence of 100 lm AdoHcy at 25 °C for 10 min.
The fluorescence emission spectra of the B[a]P-DNA
adducts were measured at 360–480 nm with an excitation
wavelength of 350 nm. In separate control experiments,
M.HhaI or M.SssI was incubated in buffer D or B, respect-
ively, in the presence of 100 lm AdoHcy at 25 °C for
10 min in the absence of B[a]P-DNA, and emission spectra
were measured at 360–480 nm with excitation at 350 nm.
These control spectra were subtracted from those of the
M.HhaI•B[a]P-DNA•AdoHcy or M.SssI•B[a]P-DNA•Ado-
Hcy complexes. For each oligodeoxynucleotide duplex
alone and in complexes with MTases, the fluorescence
intensities at 384 nm were calculated.
Determination of the amounts of the active form
of M.SssI and M.HhaI
The amount of the active form of M.SssI (338 ± 22 nm)
and M.HhaI (120 ± 10 lm) was determined as described
previously [37].
Equilibrium competition experiments
Determination of K
d
by the EMSA method
To obtain modified oligodeoxynucleotide duplexes X
+
CG ⁄

CG
C, GCX
+
⁄ CGC, X
+
CG ⁄ CGM and GCX
+
⁄ CGM, the
unmodified strands CGC or CGM were mixed with a two-
fold excess of the B[a]PDE-modified strands, X
+
CG or
GCX
+
, in buffer A, or in 50 mm Tris•HCl (pH 7.5) ⁄
50 mm NaCl. The mixtures of oligodeoxynucleotides were
heated to 80 °C and allowed to cool to room temperature.
The MTases HhaI and SssI do not bind X
+
CG or GCX
+
strands (data not shown). To obtain unmodified GCG ⁄
CG
C and GCG ⁄ CGM duplexes, the oligodeoxynucleotide
strands were mixed in the ratio 1 : 1. In the case of M.SssI,
the reference
32
P-labeled GCG ⁄ CGM
ref
duplex (100 nm)

was mixed with increasing concentrations of the B[a]PDE-
modified competitor duplex (0–250 nm) in buffer C contain-
ing 8% glycerol and AdoHcy (1 mm). In the case of
M.HhaI, the
32
P-labeled GCG ⁄ CGM
ref
duplex (2 nm) was
mixed with increasing concentrations of the B[a]PDE-modi-
fied competitor duplex (0–90 nm) in buffer E containing
8% glycerol and AdoHcy (0.1 mm). M.SssI or M.HhaI was
added to a final concentration of 50 nm or 0.6 nm, respect-
ively, and samples were incubated at room temperature for
5 min and then at 0 °C for 10 min, and were analyzed by
nondenaturing 8% PAGE in 0.5 · buffer F. When exposed,
radioactive gels were processed as described in [37], and
the values of k ¼ K
r
d
⁄ K
d
were determined as described in
[37], where K
r
d
and K
d
are the dissociation constants of the
MTase•G
CG ⁄ CGM

ref
•AdoHcy and MTase•B[a]P-DNA•
AdoHcy complexes. The k ratio specifies the relative bind-
ing efficiency of the reference unmodified DNA and the
competitor damaged DNA to the MTase. The K
d
values
were calculated by dividing K
r
d
(see below) by the experi-
mental values of k.
Determination of K
r
d
for the M.HhaI(M.SssI)•
G
CG ⁄ CGM
ref
•AdoHcy complexes
We were unable to obtain an accurate K
r
d
values for the
M.HhaI•G
CG ⁄ CGM
ref
•AdoHcy complex by direct titra-
tion of solutions of the G
CG ⁄ CGM

ref
duplex with increas-
ing enzyme concentrations. This was due to the
considerable experimental error associated with the dilution
of the enzyme and the low radioactivity of the DNA at low
(picomolar) concentrations. The K
r
d
value for the
M.HhaI•G
CG ⁄ CGM
ref
•AdoHcy complex was calculated
by multiplying the ratio k (obtained from the competitive
binding of G
CG ⁄ CGM
ref
and Y
+
CG ⁄ CGM to M.HhaI)
by the K
d
value of the M.HhaI•Y
+
CG ⁄ CGM•AdoHcy
complex. The K
d
value of the M.HhaI•Y
+
CG ⁄ CGM•Ado-

Hcy complex was obtained by direct titration using EMSA.
The
32
P-labeled oligodeoxynucleotide duplex, Y
+
CG ⁄ CGM
(0.2 nm), was incubated in the presence of 0.1 mm AdoHcy
with various M.HhaI concentrations (0.3–5 nm) in buffer E
containing 8% glycerol at 37 °C for 5 min and at 0 °C for
10 min. The further experimental procedures and data ana-
lysis were the same as described in [37]. The K
r
d
value for
the M.SssI•G
CG ⁄ CGM
ref
•AdoHcy complex was deter-
mined as described [37].
Methylation assay
Oligodeoxynucleotide duplexes were obtained as described
above. The B[a]PDE-modified X
+
CG or GCX
+
strands
alone are not methylated by MTases HhaI and SssI (data
not shown). The efficiency of methylation was monitored
by the radioactivity of tritium (CH
3

-
3
H) incorporated into
the oligodeoxynucleotide duplexes [60]. The reactions were
carried out in buffer B for M.SssI, or buffer D for M.HhaI.
The mixtures contained G
CG ⁄ CGC, X
+
CG ⁄ CGC, GCX
+
⁄
CG
C, GCG ⁄ CGM, X
+
CG ⁄ CGM or GCX
+
⁄ CGM, M.SssI
(17 nm) or M.HhaI (5 nm), and [CH
3
-
3
H]AdoMet (1.3 lm).
In the case of both enzymes, the saturating duplex concentra-
tions were found at which the V
0
values were not changed.
These DNA concentrations were 0.25–0.35 lm or 0.05–
0.5 lm in the M.SssI and M.HhaI reactions, respectively.
The reactions were started by the addition of the MTase.
After 0.5–15 min of incubation at 37 °C, aliquots of the reac-

tion mixtures were pipetted on to DE-81 paper disks (What-
man, Brentford, UK) and treated as described [61]. The
amounts of methylated DNA were computed as described
[60]. The V
0
values for all duplexes were determined from the
initial linear portions of the product versus time profiles. In
O. M. Subach et al. Stereochemistry of B[a]P-N
2
-dG affects methylation
FEBS Journal 274 (2007) 2121–2134 ª 2007 The Authors Journal compilation ª 2007 FEBS 2131
the used duplex concentration ranges, the measured V
0
values were constant for each unmodified or B[a]PDE-modi-
fied duplex, indicating that the V
max
limit was reached. Using
the V
max
values thus obtained, the k
cat
values were calculated
(Table 2).
Acknowledgements
This research was supported by a US Public Health
Service grant No. TW05689 from the Fogarty Interna-
tional Center (New York University and Moscow
State University), NIH Grant CA 099194 (New York
University), and RFBR Grants 04-04-49488 and 05-04-
49690 (Moscow State University). We thank Dr

B. Jack and Dr G. Wilson from New England Biolabs
for their gift of the M.SssI plasmid, Dr F. Johnson
and Dr R. Bonala for synthesis of trans-B[a]P-N
2
-dG
adducts, Dr N. N. Veiko and Dr. N. A. Cherepanova
for assistance with the fluorescence experiments.
References
1 Luch A (2005) Nature and nurture: lessons from chemi-
cal carcinogenesis. Nat Rev Cancer 5, 113–125.
2 Wei SJ, Chang RL, Hennig E, Cui XX, Merkler KA,
Wong CQ, Yagi H, Jerina DM & Conney AH (1994)
Mutagenic selectivity at the HPRT locus in V-79 cells:
comparison of mutations caused by bay-region
benzo[a]pyrene 7,8-diol-9,-10-epoxide enantiomers with
high and low carcinogenic activity. Carcinogenesis 15,
1729–1735.
3 Buening MK, Wislocki PG, Levin W, Yagi H, Thakker
DR, Akagi H, Koreeda M, Jerina DM & Conney AH
(1978) Tumorigenicity of the optical enantiomers of the
diastereomeric benzo[a]pyrene 7,8-diol-9,10-epoxides in
newborn mice: exceptional activity of (+)-7beta,8alpha-
dihydroxy-9alpha,10alpha-epoxy-7,8,9,10-tetra-
hydrobenzo[a]pyrene. Proc Natl Acad Sci USA 75,
5358–5361.
4 Cheng SC, Hilton BD, Roman JM & Dipple A (1989)
DNA adducts from carcinogenic and noncarcinogenic
enantiomers of benzo[a]pyrene dihydrodiol epoxide.
Chem Res Toxicol 2, 334–340.
5 Meehan T & Straub K (1979) Double-stranded DNA

steroselectively binds benzo(a)pyrene diol epoxides.
Nature 277, 410–412.
6 Geacintov NE, Cosman M, Hingerty BE, Amin S,
Broyde S & Patel DJ (1997) NMR solution structures
of stereoisometric covalent polycyclic aromatic carcino-
gen-DNA adduct: principles, patterns, and diversity.
Chem Res Toxicol 10, 111–146.
7 Jiang G, Jankowiak R, Grubor N, Banasiewicz M,
Small GJ, Skorvaga M, Van Houten B & States JC
(2004) Supercoiled DNA promotes formation of
intercalated cis-N2-deoxyguanine adducts and base-
stacked trans-N2-deoxyguanine adducts by (+) -7R,
8S-dihydrodiol-9S,10R-epoxy-7,8,9,10-tetra-hydro-
benzo[a]pyrene. Chem Res Toxicol 17, 330–339.
8 Conney AH (1982) Induction of microsomal enzymes
by foreign chemicals and carcinogenesis by polycyclic
aromatic hydrocarbons. Cancer Res 42, 4875–4917.
9 Cosman M, de los Santos C, Fiala R, Hingerty B,
Ibanez V, Luna E, Harvey RG, Geacintov NE, Broyde
S & Patel DJ (1993) Solution conformation of the
(+)-cis-anti-[BP]dG adduct in a DNA duplex: intercala-
tion of the covalently attached benzo[a]pyrenyl ring into
the helix and displacement of the modified deoxyguano-
sine. Biochemistry 32, 4145–4155.
10 Cosman M, de los Santos C, Fiala R, Hingerty BE,
Singh SB, Ibanez V, Margulis LA, Live D, Geacintov
NE, Broyde S & Patel DJ (1992) Solution conformation
of the major adduct between the carcinogen (+)-anti-
benzo[a]pyrene diol epoxide and DNA. Proc Natl Acad
Sci USA 89, 1914–1918.

11 Hess MT, Gunz D, Luneva N, Geacintov NE & Naegeli
H (1997) Base pair conformation-dependent excision of
benzo[a]pyrene diol epoxide-guanine adducts by human
nucleotide excision repair enzymes. Mol Cell Biol 17,
7069–7076.
12 Zou Y, Luo C & Geacintov NE (2001) Hierarchy of
DNA damage recognition in Escherichia coli nucleotide
excision repair. Biochemistry 40, 2923–2931.
13 Chiapperino D, Kroth H, Kramarczuk IH, Sayer JM,
Masutani C, Hanaoka F, Jerina DM & Cheh AM
(2002) Preferential misincorporation of purine nucleo-
tides by human DNA polymerase eta opposite
benzo[a]pyrene 7,8-diol 9,10-epoxide deoxyguanosine
adducts. J Biol Chem 277, 11765–11771.
14 Rechkoblit O, Zhang Y, Guo D, Wang Z, Amin S,
Krzeminsky J, Louneva N & Geacintov NE (2002)
trans-Lesion synthesis past bulky benzo[a]pyrene diol
epoxide N2-dG and N6-dA lesions catalyzed by DNA
bypass polymerases. J Biol Chem 277, 30488–30494.
15 Fernandes A, Liu T, Amin S, Geacintov NE, Grollman
AP & Moriya M (1998) Mutagenic potential of stereoiso-
meric bay region (+)- and (-)-cis-anti-benzo[a]pyrene diol
epoxide-N2-deoxyguanosine adducts in Escherichia coli
and simian kidney cells. Biochemistry 37, 10164–10172.
16 Ponten I, Kroth H, Sayer JM, Dipple A & Jerina DM
(2001) Differences between the mutational consequences
of replication of cis- and trans-opened benzo[a]pyrene
7,8-diol 9,10-epoxide-deoxyguanosine adducts in
M13mp7L2 constructs. Chem Res Toxicol 14, 720–726.
17 Choi D-J, Marino-Alessandri DJ, Geacintov NE &

Scicchitano DA (1994) Site-specific benzo[a]pyrene diol
epoxide–DNA adducts inhibit transcription elongation
by bacteriophage T7 RNA polymerase. Biochemistry 33,
780–787.
Stereochemistry of B[a]P-N
2
-dG affects methylation O. M. Subach et al.
2132 FEBS Journal 274 (2007) 2121–2134 ª 2007 The Authors Journal compilation ª 2007 FEBS
18 MacLeod MC, Powell KL, Kuzmin VA, Kolbanovskiy
A & Geacintov NE (1996) Interference of benzo[a]py-
rene diol epoxide-deoxyguanosine adducts in a GC box
with binding of the transcription factor Sp1. Mol Carci-
nog 16, 44–52.
19 Persson AE, Ponten I, Cotgreave I & Jernstrom B
(1996) Inhibitory effects on the DNA binding of AP-1
transcription factor to an AP-1 binding site modified by
benzo[a]pyrene 7,8-dihydrodiol 9,10-epoxide diastereo-
mers. Carcinogenesis 17, 1963–1969.
20 Pommier Y, Kohlhagen G, Pourquier P, Sayer JM,
Kroth H & Jerina DM (2000) Benzo[a]pyrene diol epox-
ide adducts in DNA are potent suppressors of a normal
topoisomerase I cleavage site and powerful inducers of
other topoisomerase I cleavages. Proc Natl Acad Sci
USA 97, 2040–2045.
21 Tian L, Sayer JM, Kroth H, Kalena G, Jerina DM &
Shuman S (2003) Benzo[a]pyrene-dG adduct interference
illuminates the interface of vaccinia topoisomerase with
the DNA minor groove. J Biol Chem 278, 9905–9911.
22 Jeltsch A (2002) Beyond Watson and Crick: DNA
methylation and molecular enzymology of DNA methyl-

transferases. Chembiochem 3, 274–293.
23 Hermann A, Gowher H & Jeltsch A (2004) Biochemis-
try and biology of mammalian DNA methyltransferases.
Cell Mol Life Sci 61, 2571–2587.
24 Momparler RL & Bovenzi V (2000) DNA methylation
and cancer. J Cell Physiol 183, 145–154.
25 Trinh BN, Long TI, Nickel AE, Shibata D & Laird PW
(2002) DNA methyltransferase deficiency modifies can-
cer susceptibility in mice lacking DNA mismatch repair.
Mol Cell Biol 22, 2906–2917.
26 Kane MF, Loda M, Gaida GM, Lipman J, Mishra R,
Goldman H, Jessup JM & Kolodner R (1997) Methyl-
ation of the hMLH1 promoter correlates with lack of
expression of hMLH1 in sporadic colon tumors and
mismatch repair-defective human tumor cell lines. Can-
cer Res 57, 808–811.
27 El-Osta A (2003) DNMT cooperativity-the developing
links between methylation, chromatin structure and can-
cer. Bioessays 25, 1071–1084.
28 Esteller M (2005) Aberrant DNA methylation as a can-
cer-inducing mechanism. Annu Rev Pharmacol Toxicol
45, 629–656.
29 Kumar S, Cheng X, Klimasauskas S, Mi S, Posfai J,
Roberts RJ & Wilson GG (1994) The DNA (cytosine-5)
methyltransferases. Nucleic Acids Res 22, 1–10.
30 Nur I, Szyf M, Razin A, Glaser G, Rottem S & Razin
S (1985) Procaryotic and eucaryotic traits of DNA
methylation in spiroplasmas (mycoplasmas). J Bacteriol
164, 19–24.
31 Denissenko MF, Chen JX, Tang MS & Pfeifer GP

(1997) Cytosine methylation determines hot spots of
DNA damage in the human P53 gene. Proc Natl Acad
Sci USA 94, 3893–3898.
32 Chen JX, Zheng Y, West M & Tang MS (1998) Carci-
nogens preferentially bind at methylated CpG in the
p53 mutational hot spots. Cancer Res 58, 2070–2075.
33 Weisenberger DJ & Romano LJ (1999) Cytosine
methylation in a CpG sequence leads to enhanced
reactivity with benzo[a]pyrene diol epoxide that corre-
lates with a conformational change. J Biol Chem 274,
23948–23955.
34 Wilson VL, Smith RA, Longoria J, Liotta MA, Harper
CM & Harris CC (1987) Chemical carcinogen-induced
decreases in genomic 5-methyldeoxycytidine content of
normal human bronchial epithelial cells. Proc Natl Acad
Sci USA 84, 3298–3301.
35 Wojciechowski MF & Meehan T (1984) Inhibition of
DNA methyltransferases in vitro by benzo[a]pyrene
diol epoxide-modified substrates. J Biol Chem 259,
9711–9716.
36 Baskunov VB, Subach FV, Kolbanovskiy A,
Kolbanovskiy M, Eremin SA, Johnson F, Bonala R,
Geacintov NE & Gromova ES (2005) Effects of
benzo[a]pyrene-deoxyguanosine lesions on DNA methy-
lation catalyzed by EcoRII DNA methyltransferase
and on DNA cleavage effected by EcoRII restriction
endonuclease. Biochemistry 44, 1054–1066.
37 Subach OM, Baskunov VB, Darii MV, Maltseva DV,
Alexandrov DA, Kirsanova OV, Kolbanovskiy A,
Kolbanovskiy M, Johnson F, Bonala R, et al. (2006)

Impact of benzo[a]pyrene-2¢-deoxyguanosine lesions on
methylation of DNA by SssI and HhaI DNA methyl-
transferases. Biochemistry 45, 6142–6159.
38 Bombail V, Moggs JG & Orphanides G (2004) Pertur-
bation of epigenetic status by toxicants. Toxicol Lett
149, 51–58.
39 Geacintov NE, Cosman M, Mao B, Alfano A, Ibanez V
& Harvey RG (1991) Spectroscopic characteristics and
site I ⁄ site II classification of cis and trans benzo[a]py-
rene diolepoxide enantiomer-guanosine adducts in oligo-
nucleotides and polynucleotides. Carcinogenesis 12,
2099–2108.
40 Zou Y, Liu TM, Geacintov NE & Van Houten B
(1995) Interaction of the UvrABC nuclease system with
a DNA duplex containing a single stereoisomer of
dG-(+)- or dG-(-)-anti-BPDE. Biochemistry 34, 13582–
13593.
41 Wyszynski MW, Gabbara S, Kubareva EA, Romanova
EA, Oretskaya TS, Gromova ES, Shabarova ZA &
Bhagwat AS (1993) The cysteine conserved among
DNA cytosine methylases is required for methyl trans-
fer, but not for specific DNA binding. Nucleic Acids
Res 21, 295–301.
42 Klimasauskas S & Roberts RJ (1995) M.HhaI binds
tightly to substrates containing mismatches at the target
base. Nucleic Acids Res 23, 1388–1395.
43 Lin SY & Riggs AD (1972) Lac repressor binding to
non-operator DNA: detailed studies and a comparison
O. M. Subach et al. Stereochemistry of B[a]P-N
2

-dG affects methylation
FEBS Journal 274 (2007) 2121–2134 ª 2007 The Authors Journal compilation ª 2007 FEBS 2133
of equilibrium and rate competition methods. J Mol
Biol 72, 671–690.
44 Sankpal UT & Rao DN (2002) Structure, function, and
mechanism of HhaI DNA methyltransferases. Crit Rev
Biochem Mol Biol 37, 167–197.
45 Klimasauskas S, Kumar S, Roberts RJ & Cheng X
(1994) HhaI methyltransferase flips its target base out
of the DNA helix. Cell 76, 357–369.
46 Koudan EV, Bujnicki JM & Gromova ES (2004)
Homology modeling of the CG-specific DNA methyl-
transferase SssI and its complexes with DNA and
AdoHcy. J Biomol Struct Dyn 22, 339–345.
47 Geacintov NE, Solntsev K, Johnson LW, Chen JX,
Kolbanovskiy AD, Liu TM & Shafirovich VY (1998)
Photoinduced electron transfer and strand cleavage in
pyrenyl-DNA complexes in adducts. J Phys Org Chem
11, 561–565.
48 Shafirovich VY, Courtney SH & Naiqi Ya & Geacintov
NE (1995) Proton-coupled photoinduced electron trans-
fer, deuterium isotope effects, and fluorescence quench-
ing in noncovalent benzo[a]pyrenetetraol–nucleoside
complexes in aqueous solutions. J Am Chem Soc 117,
4920–4929.
49 Geacintov NE, Zinger D, Ibanez V, Santella R,
Grunberger D & Harvey RG (1987) Properties of cova-
lent benzo[a]pyrene diol epoxide-DNA adducts investi-
gated by fluorescence techniques. Carcinogenesis 8, 925–
935.

50 Undeman O, Lycksell PO, Graslund A, Astlind T,
Ehrenberg A, Jernstrom B, Tjerneld F & Norden B
(1983) Covalent complexes of DNA and two stereoi-
somers of benzo (a) pyrene 7,8-dihydrodiol-9,10-epoxide
studied by fluorescence and linear dichroism. Cancer Res
43, 1851–1860.
51 Wittung P, Funk M, Jernstrom B, Norden B & Takaha-
shi M (1995) Fluorescence-detected interactions of
oligonucleotides in RecA complexes. FEBS Lett 368,
64–68.
52 Zou Y, Bassett H, Walker R, Bishop A, Amin S,
Geacintov NE & Van Houten B (1998) Hydrophobic
forces dominate the thermodynamic characteristics of
UvrA–DNA damage interactions. J Mol Biol 281, 107–
119.
53 Geacintov NE, Mao B, France LL, Zhao R, Chen J,
Liu T-M, Ya N-Q, Margulis LA & Sutherland JC
(1992) Photoinduced electron transfer and fluorescence
mechanisms in covalently linked polynuclear aromatic
nucleotide complexes. SPIE Time-Resolved Laser Spec-
troscopy Biochemistry III 1640, 774–783.
54 Roberts RJ & Cheng X (1998) Base flipping. Annu Rev
Biochem 67, 181–198.
55 Mao B, Xu J, Li B, Margulis LA, Smirnov S, Ya NQ,
Courtney SH & Geacintov NE (1995) Synthesis and char-
acterization of covalent adducts derived from the binding
of benzo[a]pyrene diol expoxide to a -GGG- sequence in
a deoxyoligonucleotide. Carcinogenesis 16, 357–365.
56 Ponten I, Kim SK, Graslund A, Norden B & Jernstrom
B (1994) Spectroscopic studies of the trans adducts

derived from (+)- and (-)-anti-benzo[a]pyrene-7,8-di-
hydrodiol-9,10-epoxide and the oligonucleotide 5¢-d
(CCTATAGATATCC). Carcinogenesis 15, 2207–2213.
57 Kumar S, Cheng X, Pflugrath JA & Roberts R (1992)
Purification, crystallization and preliminary X-ray
diffraction analysis of an M.HhaI–AdoMet complex.
Biochemistry 31
, 8648–8653.
58 Darii MV, Kirsanova OV, Drutsa VL, Kochetkov SN
& Gromova ES (2007) Isolation and site-directed muta-
genesis of DNA methyltransferase SssI. Mol Biol
(Mosk) 41, 110–117.
59 Cosman M, Ibanez V, Geacintov NE & Harvey RG
(1990) Preparation and isolation of adducts in high yield
derived from the binding of two benzo[a]pyrene-7,8-di-
hydroxy-9,10-oxide stereoisomers to the oligonucleotide
d (ATATGTATA). Carcinogenesis 11, 1667–1672.
60 Brennan CA, Van Cleve MD & Gumport RI (1986)
The effects of base analogue substitutions on the methy-
lation by the EcoRI modification methylase of octa-
deoxyribonucleotides containing modified EcoRI
recognition sequences. J Biol Chem 261, 7279–7286.
61 Subach OM, Khoroshaev AV, Gerasimov DN,
Baskunov VB, Shchyolkina AK & Gromova ES (2004)
2-Pyrimidinone as a probe for studying the EcoRII
DNA methyltransferase–substrate interaction. Eur J
Biochem 271, 2391–2399.
62 Huang W, Amin S & Geacintov NE (2002) Fluores-
cence characteristics of site-specific and stereochemically
distinct benzo[a]pyrene diol epoxide-DNA adducts as

probes of adduct conformation. Chem Res Toxicol 15,
118–126.
63 O’Gara M, Klimasauskas S, Roberts RJ & Cheng X
(1996) Enzymatic C5-cytosine methylation of DNA:
mechanistic implications of new crystal structures for
HhaI methyltransferase-DNA-AdoHcy complexes.
J Mol Biol 261, 634–645.
Stereochemistry of B[a]P-N
2
-dG affects methylation O. M. Subach et al.
2134 FEBS Journal 274 (2007) 2121–2134 ª 2007 The Authors Journal compilation ª 2007 FEBS