Influence of local duplex stability and N6-methyladenine on uracil recognition by mismatch-specific uracil-DNA glycosylase (Mug)

Article abstract of DOI:10.1021/tx020062y

To maintain genomic integrity, DNA repair enzymes continually remove damaged bases and lesions resulting from endogenous and exogenous processes. These repair enzymes must distinguish damaged bases from normal bases to prevent the inadvertent removal of n

Full text of DOI:10.1021/tx020062y

Chem. Res. Toxicol. 2002, 15, 1595-1601
1595
In flu en ce of Loca l Du p lex Sta bility a n d
N6-Meth yla d en in e on Ur a cil Recogn ition by
Mism a tch -Sp ecific Ur a cil-DNA Glycosyla se (Mu g)
Victoria Valinluck,† Pingfang Liu,†,‡ Artur Burdzy,† J unichi Ryu,† and
Lawrence C. Sowers*,†,‡
Department of Biochemistry and Microbiology, School of Medicine, Loma Linda University,
Loma Linda, California 92350, and Graduate School of Biological Sciences, City of Hope National
Medical Center, 1500 East Duarte Road, Duarte, California 91010
Received J uly 2, 2002
To maintain genomic integrity, DNA repair enzymes continually remove damaged bases and
lesions resulting from endogenous and exogenous processes. These repair enzymes must
distinguish damaged bases from normal bases to prevent the inadvertent removal of normal
bases, which would promote genomic instability. The mechanisms by which this high level of
specificity is accomplished are as yet unresolved. One member of the uracil-DNA glycosylase
family of repair enzymes, Escherichia coli mismatch-specific uracil-DNA glycosylase (Mug),
is reported to distinguish U:G mispairs from U:A base pairs based upon specific contacts with
the mispaired guanine after flipping the target uracil out of the duplex. However, recent studies
suggest other mechanisms for base selection, including local duplex stability. In this study,
we used the modified base N6-methyladenine to probe the effect of local helix perturbation on
Mug recognition of uracil. N6-Methyladenine is found in E. coli as part of both the mismatch
repair and restriction-modification systems. In its cis isomer, N6-methyladenine destabilizes
hydrogen bonding by interfering with pseudo-Watson-Crick base pairing. It is observed that
the selection of uracil by Mug is sequence dependent and that uracil residues in sequences of
reduced thermostability are preferentially removed. The replacement of adenine by N6-
methyladenine increases the frequency of removal of the uracil residue paired opposite the
modified adenine. These results are in accord with suggestions that local helix stability is an
important determinant of base recognition by some DNA repair enzymes and provide a potential
strategy for identifying the sequence location of modified bases in DNA.
Escherichia coli mismatch-specific uracil-DNA glyco-
sylase (Mug)1 is a member of the uracil glycosylase family
and only cleaves its substrate in the context of a DNA
duplex, explaining its initial name, double-strand specific
uracil-DNA glycosylase (10, 11). Later, it was reported
that Mug recognizes U:G mispairs, but not U:A base pairs
(10-12), and that Mug specifically recognizes the “wid-
owed” guanine opposite the target uracil via three specific
hydrogen bonds (12, 13), thus the designation mismatch-
specific uracil-DNA glycosylase (Mug) (12). However,
other investigators refer to this base excision repair
enzyme as ethenoC-DNA glycosylase based on its high
excision activity against ethenocytosine in duplex DNA
(14). Recent studies suggest that thermal stability rather
than recognition of the “widowed” guanine is the primary
determinant of Mug substrate selectivity (15). Mug has
been reported, by several groups, to excise uracil from
pseudo-Watson-Crick U:A base pairs as well as wobble
U:G mispairs, though at a lower frequency than the U:G
mispairs (14-16). In this study, we investigate whether
thermal destabilization (∆Tm) may be a key factor in the
mechanism for Mug substrate selectivity. We do so by
In tr od u ction
The human genome is under constant insult from both
exogenous and endogenous DNA damaging agents. The
number of lesions resulting from endogenous DNA dam-
age sources such as water, reactive oxygen species, and
alkylating agents is estimated to be on the order of
10 000/cell/day (1-3). To maintain genomic integrity and
stability, DNA repair pathways are implemented. The
major mechanisms that exist for the repair of these
lesions include base excision repair, mismatch repair,
nucleotide excision repair, and recombination repair,
although base excision repair is most commonly employed
in the repair of endogenous lesions (4-9). Base excision
repair glycosylases must distinguish the estimated 104
damaged bases in need of repair from the 109 normal
bases also present. The selectivity of these glycosylases
must be greater than 1 in 105 to prevent the excision of
undamaged bases, thereby preventing inadvertent dam-
age to the genome that would contribute to genomic
instability. The mechanisms by which glycosylases find
a damaged base and distinguish that base from normal
bases are under investigation.
1
Abbreviations: Mug, mismatch-specific uracil-DNA glycosylase;
* To whom correspondence should be addressed. Telephone: (909)
558-4480. Fax: (909) 558-4035. E-mail: lsowers@som.llu.edu.
† Loma Linda University.
APE, human apurinic/apyrimidinic endonuclease; m6A, N6-methylad-
enine; tms, trimethylsilyl; Tm, oligonucleotide melting temperature;
dU, 2′-deoxyuridine; GC/MS, gas chromatography-mass spectrometry;
DMT, dimethoxytrityl.
‡ City of Hope National Medical Center.
10.1021/tx020062y CCC: $22.00 © 2002 American Chemical Society
Published on Web 11/20/2002

1596 Chem. Res. Toxicol., Vol. 15, No. 12, 2002
Valinluck et al.
F igu r e 2. Sequence of 30-mer duplexes used in Mug-mediated
uracil excision assays. The duplex without m6A is oligo A.
Introduction of m6A into the duplex at position 1 gives oligo
A1, at position 2 gives oligo A2, and at position 3 gives
oligo A3.
F igu r e 1. (A) Base pairing between uracil and N1-cis rotational
isomer of N6-methyladenine, and (B) pseudo-Watson-Crick base
pairing between uracil and N1-trans rotational isomer of N6-
methyladenine. Dashed lines represent hydrogen-bonding in-
teractions.
Ma ter ia ls a n d Meth od s
m 6A P h osp h or a m id ite P r ep a r a tion (23-25). N6-Methyl-
2′-deoxyadenosine (Aldrich) was coevaporated with anhydrous
pyridine (Aldrich). The 1.13 mmol of dry product was tritylated
with 1.22 mmol DMT-Cl (Aldrich) in 7 mL of anhydrous pyridine
overnight at room temperature. Completion of the reaction was
determined via thin-layer chromatography on general purpose
silica gel glass plates with UV indicator (Aldrich) using 5% (v/
v) MeOH in dichloromethane (Aldrich) as the solvent. Extraction
of the tritylation reaction was completed using dichloromethane
and saturated sodium bicarbonate (Aldrich). The organic phase
was retained and evaporated. The dried product was dissolved
in dichloromethane with 0.5% (v/v) triethylamine (Aldrich). A
Silica H gel column (Aldrich) with dichloromethane with 0.5%
(v/v) triethylamine as the solvent was used to purify the N6-
methyl-2′-deoxyadenosine-DMT, with a total yield of 75%. N6-
methyl-2′-deoxyadenosine-DMT (0.85 mol or 0.483 g) was
phosphitylated with 300 µL of 2-cyanoethyl tetraisopropyl-
phosphoramidite 97% (Aldrich), in the presence of 75 mg of N,N-
diisopropylammonium tetrazolide (Aldrich) and 2 mL of aceto-
nitrile overnight, under argon at room temperature. Completion
of the reaction was determined via thin-layer chromatography
using 5% (v/v) MeOH in dichloromethane as the solvent.
Extraction of the phosphitylation reaction was completed with
20% (v/v) sodium chloride (Aldrich) in dH2O with equal volume
of triethylamine and ethyl acetate. The organic phase was
retained and evaporated. The dried product was dissolved in
ethyl acetate. The phosphoramidite purification was completed
using a Silica H gel column with ethyl acetate and hexane at
varying concentrations with 0.5% (v/v) triethylamine as the
solvent. The final product was dried under reduced pressure.
d U P h osp h or a m id ite P r ep a r a tion . 2′-Deoxyuridine was
converted to its corresponding phosphoramidite using the same
protocol as described previously for the N6-methyl-2′-deoxyad-
enosine phosphoramidite (24, 25), except on the gram scale
rather than the milligram scale.
studying the effect of modest base pair destabilization,
induced by m6A base paired with uracil in duplex
oligonucleotide DNA, on Mug substrate recognition.
N6-Methyladenine (m6A) is involved in mismatch
repair in E. coli. After DNA replication, the newly
synthesized daughter strand is transiently unmethylated
at adenine in GATC sequences while the parent strand
retains the adenine methylation pattern (17). This ad-
enine hemimethylation allows for strand discrimination
and directs the mismatch repair in E. coli (17-19). For
the free nucleoside, there is a 20:1 preference for the m6A
cis rotational isomer over the trans rotational isomer (20).
However, in DNA, the methyl group of m6A exists as
either cis or trans to N1 with a preference for the trans
isomer (20) (Figure 1). In both isomers, the methyl group
is coplanar with the heterocyclic rings of adenine. When
in the trans configuration, the methyl group of the m6A
protrudes into the major groove of the helix. This
positioning of the methyl group in the trans orientation
does not disturb hydrogen bonding between the m6A and
the opposite pyrimidine, allowing for pseudo-Watson-
Crick base pairing. However, when in the cis configura-
tion, the methyl group of m6A is located on the N1 side
of the base, interfering with pseudo-Watson-Crick base
pairing (20, 21). It has been shown that the presence of
m6A in DNA destabilizes base pairing and helix stability
(22). In this study, we use this base-pairing destabiliza-
tion property of m6A to create selected sites of instability
within a duplex to study the substrate recognition
mechanism of the base excision repair enzyme, mismatch-
specific uracil-DNA glycosylase (Mug).
Oligon u cleotid e Syn th esis a n d P u r ifica tion . Oligonucle-
otide 30-mers (Figure 2) were prepared by standard solid-phase
synthesis using either the Gene Assembler Plus (Pharmacia)
or Expedite Nucleic Acid Synthesis System (Applied Biosystems)
automated DNA synthesizers (24, 25). Phosphoramidites other
than those for m6A and dU, were obtained from Glen Research.
The oligonucleotides were removed from the solid support and
deprotected in aqueous ammonia (Aldrich) at 60 °C overnight,
after which they were purified with Poly-Pak II cartridges (Glen
Research). The composition of the oligonucleotides was verified
by gas chromatography-mass spectometry (GC/MS) after hy-
drolysis in 200 µL of 88% (v/v) formic acid (Aldrich) at 140 °C
for 30 min, and silylation in 100 µL of dry acetonitrile (Aldrich)
and 100 µL of bis-(trimethylsilyl) trifluoroacetamide containing
1% (v/v) trimethylchlorosilane (Aldrich) (26, 27). The GC/MS
We prepared an oligonucleotide containing a run of
three adenines base-paired with three uracils, as well as
a series of oligonucleotide duplexes containing uracil
paired with m6A at selected positions. These oligonucle-
otides were then subjected to uracil excision by Mug.
Examination of Mug-induced uracil excision opposite a
run of three adenine residues revealed that the uracil
excision efficiency is consistent with the local stability
of the DNA duplex. We also find that, within the same
sequence context, uracil opposite m6A is rendered more
susceptible to Mug-mediated uracil excision as compared
to uracil opposite adenine. These results, in conjunction
with Tm data obtained from corresponding 11-mer du-
plexes, confirm the importance of slight differences in
thermal stability for Mug substrate recognition. Further-
more, the recognition of modest base pair destabilization
by Mug can potentially be exploited as a strategy for
determining the location of m6A in a DNA duplex,
applicable in the characterization of restriction-modifica-
tion systems.
analysis was conducted with
a Hewlett-Packard 5970 GC
interfaced with a 5890 mass selective detector. The m/z of
predominant ions from the mass spectrum of N6-methyladenine
is shown in Figure 3. Oligonucleotides for the melting temper-
ature studies and for the oligonucleotide size markers used in
the Mug-mediated uracil excision assays were synthesized in
the same manner as described for the 30-mer oligonucleotides.

Effects of Local Duplex Stability and N6mA on Mug
Chem. Res. Toxicol., Vol. 15, No. 12, 2002 1597
Ta ble 1. Com p a r ison of Tm (°C) for 11-m er Du p lexes Con ta in in g Ad en in e or N6-Meth yla d en in e a t Differ en t P osition s
P a ir ed w ith eith er Ur a cil or Th ym in ea
GCAGA1A2A3TCGA
(A)
GCAGMA2A3TCGA
(m 6A1)
GCAGA1MA3TCGA
(m 6A2)
GCAGA1A2MTCGA
(m 6A3)
TCGAUUUCTGC (U)
TCGATTTCTGC (T)
44.0 ( 0.6
46.6 ( 0.6
41.5 ( 1.1
42.6 ( 0.3
41.4 ( 0.4
43.1 ( 0.4
42.0 ( 0.5
44.4 ( 0.9
a
The measurements were carried out in 0.1 M NaCl, 0.01 M sodium phosphate, and 0.1 mM EDTA (pH 7). The Tm values shown were
determined at a total strand concentration of 25 µM. The values shown are the averages of five to six separate determinations. M denotes
N6-methyladenine. The names of the single-stranded 11-mers are in parentheses, italicized and bolded.
F igu r e 3. Mass spectrum of doubly silylated N6-methylad-
enine.
F igu r e 4. Twenty minute Mug-mediated uracil excision of 30-
mer duplexes with uracil paired with either adenine or N6-
methyladenine at different positions. Lane 1: 19-, 20-, and 21-
mer size markers. Lanes 2-5: 20 min Mug-mediated uracil
excision of oligo A, oligo A1, oligo A2, and oligo A3,
respectively. Lane 6: negative control, oligo A incubated in 20
mM Tris-HCl, pH 8.0, 0.1 mg/mL BSA, 1 mM EDTA, 1 mM
EGTA, and 1 mM DTT for 20 min without the addition of Mug.
This reaction was conducted in a 10 µL reaction volume, and
was treated with human APE as described for all other Mug-
mediated uracil excision reactions.
The sequences of the 11-mer oligonucleotides used in the melting
temperature measurements are seen in Table 1. The sequences
of the oligonucleotides used as the oligonucleotide size markers
in the Mug-mediated uracil excision assays are as follows:
19-mer: 5′ CCGGATCCGTCGACCTCGA 3′
20-mer: 5′ CCGGATCCGTCGACCTCGAT 3′
21-mer: 5′ CCGGATCCGTCGACCTCGATT 3′
dinic sites were cleaved by the addition of 1 µL of human APE
(1 unit/µL) under 10 mM HEPES-KOH, pH 6.5, 100 mM KCl,
10 mM MgCl2 buffer conditions in a volume of 15 µL and then
incubated for 1 h at 37 °C. Trevigen defines 1 unit of human
APE as the amount that cleaves 1 pmol of a 32P-labeled apurinic/
apyrimidinic site oligonucleotide in 1 h at 37 °C. The reactions
were stopped by the addition of equal volumes of Maxam-
Gilbert loading buffer (98% formamide, 0.01 M EDTA, 1 mg/
mL xylene cyanol, and 1 mg/mL bromophenol blue). The
quenched reactions were electrophoresed on 20% (v/v) denatur-
ing polyacrylamide gels, and the bands corresponding to the
Mug-mediated uracil excision products were visualized and
quantified using a phosphorimager (Molecular Dynamics) and
the ImageQuant 5.0 software (Molecular Dynamics).
Oligon u cleotid e La belin g a n d An n ea lin g. The 30-mer
oligonucleotide with the run of three uracil residues (Figure 2,
bottom strand) was 5′-32P-end labeled by T4 polynucleotide
kinase (New England Biolabs) with [γ-32P]ATP (ICN Life
Sciences) under conditions recommended by the enzyme sup-
plier. The labeled oligonucleotides were purified using G50
Sephadex columns (Boehringer Mannheim). The labeled oligo-
nucleotides were annealed with a complementary strand con-
taining m6A at one of three different positions, resulting in three
different duplexes with m6A replacing adenine at positions 1,
2, or 3 (Figure 2). The annealing was carried out in 20 mM Tris-
HCl, pH 8.0, 0.1 mg/mL BSA, 1 mM EDTA, 1 mM EGTA, and
1 mM DTT, with
a 2-fold molar excess of the unlabeled
complementary strand. The annealing mixture was heated to
95 °C for 5 min and then allowed to slowly cool to room
temperature. To confirm duplex formation, the duplexes were
digested with SalI (New England Biolabs) under conditions
recommended by the enzyme supplier, and the products were
sized on denaturing 20% (v/v) polyacrylamide gels (not shown).
The 19-, 20-, and 21-mers used for the oligonucleotide size
markers (Figure 4, lane 1) were also 5′-32P-end labeled and
purified as described above.
Mu g-m ed ia ted Ur a cil Excision Assa ys. E. coli Mug,
Human AP Endonuclease (APE), and their respective reaction
buffers were obtained from Trevigen Inc. Each Mug-mediated
uracil excision reaction was conducted with approximately 2
pmol of double-stranded oligonucleotide substrate (Figure 2) and
1 µL of Mug (1 unit/µL) in 20 mM Tris-HCl, pH 8.0, 0.1 mg/mL
BSA, 1 mM EDTA, 1 mM EGTA, and 1 mM DTT in a 10 µL
reaction volume for 5, 10, 15, 20 or 25 min at 37 °C. Trevigen
defines 1 unit of Mug as the amount that cleaves 1 pmol of a
32P-labeled oligonucleotide probe containing 3,N4-ethenocytosine
within an oligonucleotide duplex in 1 h at 37 °C. The apyrimi-
Deter m in a tion of Meltin g Tem p er a tu r es of Oligon u cle-
otid es. Samples of the single-stranded 11-mer oligonucleotides,
previously described, were prepared by combining aliquots of
the appropriate complementary single strands (Table 1) into
cuvettes for a 12.5 µM final concentration of each strand with
100 or 300 µL (depending upon cuvette size) of the melting buffer
(0.1 M NaCl, 0.01 M sodium phosphate, and 0.1 mM EDTA,
pH 7.0) (28). Oligonucleotide melting temperatures (Tm) were
determined using a Varian Cary 100 Bio UV-Visible spectro-
photometer. Four temperature ramps were performed on each
sample per run: (1) 15 to 98 °C at a rate of 1 °C/min, (2) 98 to
15 °C at a rate of 0.5 °C/min, (3) 15 to 98 °C at a rate of 0.5
°C/min, and (4) 98 to 15 °C at a rate of 0.5 °C/min. Data were
collected at 0.5 °C intervals while monitoring the temperature
with a probe inserted into a cuvette containing only buffer. The
oligonucleotides were allowed to anneal in ramp 2. Only ramps
2-4 were used for data analysis. Using the Cary WinUV
Thermal software (Varian), the Tm of each duplex was deter-
mined using a total of five to six independent Tm measurements
(Table 1) (29).

1598 Chem. Res. Toxicol., Vol. 15, No. 12, 2002
Valinluck et al.
Resu lts a n d Discu ssion
Oligon u cleotid e Syn th esis a n d Ch a r a cter iza tion .
The sequence of the oligonucleotide duplexes used in this
study is shown in Figure 2. The 30-mer was chosen based
upon studies done by Taylor et al. (30) on the substrate
recognition and selectivity of EcoR124I methyltrans-
ferase. The sequence originally used by Taylor et al. (30)
was slightly modified. Adenine instead of thymine was
included as the eleventh base from the 5′ end (Figure 2,
upper strand) as to allow for a run of three adenine
residues. This sequence includes the recognition sequence
for the type I restriction-modification system EcoR124I
(31). The three tandem adenine residues are paired with
three uracil residues in the duplexes used for the Mug-
mediated uracil excision assays. Introduction of m6A at
position A2 in the duplex (Figure 2) corresponds to the
natural methylation site of EcoR124I methyltransferase
in the 5′ portion of its recognition sequence (31-32). Of
the four 30-mer oligonucleotide duplexes synthesized, one
duplex, oligo A, does not contain modifications in the
run of three adenines in the upper strand (Figure 2),
whereas the other three duplexes have m6A at either
position A1, A2, or A3, and will be referred to as oligo
A1, oligo A2, and oligo A3, respectively (Figure 2).
F igu r e 5. Quantification of Mug-mediated excision of uracil
opposite adenine and N6-methyladenine in 30-mer duplexes. The
data shown is an average obtained from four sets of 20 min Mug-
mediated uracil excision reactions on the four duplexes tested.
The sequence shown includes only the central five base pairs
in the 30-mer duplexes. Oligo A does not contain m6A, and the
excision pattern seen, 2 > 3 > 1, is reproducible and significant
with a p-value of 0.018, determined using a two-tailed Friedman
test. Systematic replacement of adenine with m6A in the duplex
at each of the three positions significantly increases the Mug
excision of uracil opposite the m6A. In oligo A1, oligo A2, and
oligo A3, adenine was replaced with m6A at positions 1, 2, and
3, respectively. The p-value for the increase in uracil excision
at positions 1, 2, and 3 after introduction of m6A were deter-
mined to be 0.0105, 0.0105, and 0.0415, respectively, using a
one-tailed Mann-Whitney test.
The composition of the oligonucleotides was confirmed
via GC/MS. The mass spectrum for the doubly silylated
m6A is shown in Figure 3. The doubly silylated m6A,
with a retention time of 11.71 min, is the major silylated
form observed. The singly silylated m6A has a retention
time of 10.48 min and is the minor form. We observe the
parent ion (m) for the doubly silylated m6A at 293 amu.
Also observed are a major peak at 278 amu (m-15), and
minor peaks at 73 amu (tms), 206 amu (m-N-tms), and
220 amu (m-tms). The observation of these peaks indi-
cates the correct incorporation of the m6A into the
oligonucleotides.
(Table 1). Though we can detect the thermal instability
induced by the introduction of m6A into the duplex, the
decrease in Tm of the duplex is the same with m6A at
each of the three positions examined. Therefore, we
cannot distinguish the position of the m6A within the
duplex based upon the Tm data alone.
Meltin g Tem p er a tu r e Stu d ies. Table 1 shows the
average melting temperatures for the 11-mer duplexes.
The relative melting temperatures obtained for the four
11-mer duplexes are representative of the relative melt-
ing temperatures of the corresponding 30-mer duplexes
used in the Mug-induced uracil excision assays. The
shorter 11-mer duplexes were used in the melting tem-
perature studies because the effect of destabilization of
one base pair on the melting temperature is more readily
apparent in a shorter duplex. Melting temperatures
determined for the U/m 6A1, U/m 6A2, and U/m 6A3
duplexes are the same within experimental error (Table
1), but the average of the three is 2.4 °C lower than the
Tm determined for the U/A duplex, which does not contain
m6A. The same trend is true for the T/m 6A1, T/m 6A2,
and T/m 6A3 duplexes. The melting temperatures of the
three duplexes containing m6A are experimentally in-
distinguishable (Table 1), but they are on average 3.2 °C
lower than the Tm determined for the T/A duplex without
m6A. As expected, the melting temperatures obtained for
the duplexes containing three uracil residues are on
average 2.6 °C lower than the corresponding T containing
duplexes. The decreased base-stacking interactions of
uracil in the duplex, as compared with thymine, de-
creases the duplex melting temperature (33). These
melting studies show that the addition of a single m6A
base to the 11-mer duplex lowers the Tm of that duplex,
whether the opposite strand is the U or the T oligo
Effect of N6-Meth yla d en in e-In d u ced Th er m a l In -
sta bility on Ur a cil Excision by Mu g. Mug-mediated
uracil excision reactions were completed on the set of four
30-mer duplexes (Figure 2). Mug-mediated uracil excision
was allowed to proceed for varying lengths of time: 5,
10, 15, 20, and 25 min. Multiple data sets revealing
essentially identical excision patterns for all time points
were reproducibly obtained. Though the excision pattern
is the same at all time points, the percent of product
formation is proportional to the length of time the
substrates were incubated with Mug. These data indicate
that Mug will only excise one of the three possible target
uracils from each 30-mer duplex. Figure 4 shows a
denaturing polyacrylamide gel of 20 min Mug-mediated
uracil excision reactions completed with the four du-
plexes. The Mug uracil excision pattern seen in this gel
is representative of that seen at all experimental time
points tested. Cleavage of uracil at positions 1, 2, and 3
(Figure 2) generates 21-, 20-, or 19-mer bands, respec-
tively (Figure 4). Figure 5 shows the quantitation of four
sets of data obtained from 20 min Mug-mediated uracil
excision of the four duplexes.
The pattern seen with Mug-mediated uracil excision
of oligo A (no m6A) provides the baseline to which
excision patterns of oligo A1, oligo A2, and oligo A3,
with m6A at positions 1, 2, and 3, respectively, are

Effects of Local Duplex Stability and N6mA on Mug
Chem. Res. Toxicol., Vol. 15, No. 12, 2002 1599
compared. Mug-mediated uracil excision of oligo A
produces the reproducible pattern where the observed
excision frequency is uracil at position 2 > 3 > 1. The
addition of one m6A residue opposite uracil at any of the
three positions alters this pattern in a reproducible
manner. The new pattern resulting from substitution of
one m6A residue is indicative of the position of the
substitution. Without variation, the uracil opposite the
m6A at any of the three positions tested is rendered more
susceptible to Mug excision as compared to uracil op-
posite adenine within the same sequence context (Figures
4 and 5). The methyl group of m6A in its N1-cis isomer
is known to disrupt pseudo-Watson-Crick base pairing
(Figure 1) (20-22). Disruption of the interstrand hydro-
gen bonding between uracil and m6A due to the methyl
group destabilizes the base pair, thereby decreasing the
thermal stability of that base pair (Table 1) and increas-
ing its susceptibility to Mug-mediated uracil excision.
Although the position of the m6A within the substituted
oligonucleotides cannot be distinguished from the melting
studies, the Mug-mediated uracil excision pattern of
oligo A is specifically altered by the m6A substitution
at each of the three target sites. Mug excision of the uracil
residue opposite m6A is selectively enhanced relative to
the corresponding U:A base pair in oligo A (Figure 5).
Exa m in a tion of Mod els to Exp la in th e Sequ en ce-
Dep en d en ce of Ur a cil Excision by Mu g. The mis-
match-specific uracil-DNA glycosylase, Mug, was named
for its reported ability to remove uracil mispaired with
guanine, but not uracil paired with adenine (10, 12). The
apparent mechanism for this selectivity was suggested
to result from the formation of specific hydrogen-bonding
interactions between the “widowed” guanine opposite the
target uracil and the glycosylase (12, 13). However, the
results reported here, in accord with those of Liu et al.
(15) and Sung et al. (16), demonstrate that Mug cleaves
uracil opposite adenine and opposite m6A. A recent study
suggests that selection of the target base by Mug is more
likely influenced by local helix stability rather than the
specific recognition of the base opposite the target uracil
residue, reflecting the energy cost of flipping the target
base from the helix (15).
In the experimental system examined here, Mug can
only excise one of the three uracil residues within a target
duplex. The data presented in Figure 5 show that the
Mug substrate selection in the duplex containing three
uracil residues paired with adenine is specific, resulting
in a pattern of uracil excision that is reproducible, with
substantial and highly significant differences in excision
frequency of the three uracils. The uracil in the central
position, U2, is cleaved more frequently than U3, while
U1 is the least frequently excised. It is known that Mug,
like several other DNA glycosylases and DNA methy-
lases, flips the target base from the helix (12, 34-36). It
is plausible that the observed Mug-mediated uracil
excision pattern (Figure 5) is related to the relative
energy required to extrude a given uracil residue from
the helix prior to enzymatic cleavage of the glycosidic
linkage.
F igu r e 6. Models for observed Mug sequence-dependent uracil
excision preferences. Lines with double-headed arrows represent
base-stacking interactions, and dotted lines represent hydrogen-
bonding interactions. (A) Model of Mug-mediated uracil excision
based upon base stacking and base pairing of the target uracil.
In this model, disruption of the base stacking and base pairing
of only the target uracil is considered in Mug uracil excision
preference. According to this model, the predicted pattern of
uracil excision of oligo A would be position 3 > 2 > 1, which is
inconsistent with the observed relative excision frequency 2 >
3 > 1. (B) Model of Mug-mediated uracil excision based upon
base-stacking interactions of the target base pair. This model
takes into consideration the energy required to disrupt the base-
stacking interactions of the target base-pair rather than simply
the target uracil. According to this model, the predicted pattern
of uracil excision of oligo A would be position 2 > 1 > 3, which
is not the pattern experimentally observed. (C) Model of Mug-
mediated uracil excision based upon base-stacking and base-
pairing interactions of the target base and adjacent base pairs.
In this model, the energy required to disrupt the hydrogen
bonding as well as the energy that is required to disrupt base-
stacking interactions of the target uracil and the adjacent base
pairs are considered. This model predicts the relative uracil
excision frequency of oligo A to be 2 > 3 > 1 which coincides
with the observed frequency.
In this model, the magnitude of the base-stacking inter-
actions would depend on the identity of the adjacent
bases, while the disruption of the base-pairing hydrogen-
bonding interaction with adenine would be sequence
independent. Previously, Kollman et al. (37) calculated
inter- and intrastrand base-stacking interactions for all
possible nearest neighbors in DNA. Their calculations
were performed with thymine rather than uracil. Al-
though the replacement of thymine with uracil in a
duplex decreases the oligonucleotide melting temperature
due to comparatively decreased stacking interactions of
the uracil (33), we assume that the differences in stacking
interactions between uracil and thymine would be similar
for the three sequence positions examined. Upon the
basis of the data of Kollman et al. (37), the inter- and
intrastrand base-stacking interactions for a thymine
residue at positions 1, 2, and 3 (Figure 5), corresponding
to the triplets 3′ CT1T, TT2T, and TT3A in the sequence
3′ CT1T2T3A, would be -17.0, -16.2, and -15.1 kcal/mol,
respectively. The T residue demonstrating the least
stacking interactions, and therefore the most easily
excluded from the helix according to the model shown in
Figure 6A, would be T3, followed by T2 and T1.
We considered several models to explain the observed
Mug sequence-dependent uracil excision preferences as
diagramed in Figure 6. The simplest model involves
disrupting the inter- and intrastrand base-stacking in-
teractions between the target uracil residue and the
adjacent base pairs, and the hydrogen-bonding interac-
tion of the target U:A base pair as shown in Figure 6A.
The observed Mug excision preference for oligo A is
position 2 > 3 > 1 (Figure 4, lane 2), whereas the relative
excision frequency predicted based upon disruption of
only base stacking with adjacent bases is position 3 > 2

1600 Chem. Res. Toxicol., Vol. 15, No. 12, 2002
Valinluck et al.
> 1. We therefore considered a model in which the
stacking interactions of the target base pair, rather than
simply the target base, is disrupted by the glycosylase
prior to base flipping (Figure 6B). The corresponding base
pair stacking interaction energies based upon the data
reported by Kollman et al. (37) would be -39.2, -38.2,
and -40.0 kcal/mol for positions 1, 2, and 3, respectively.
The predicted relative excision frequency based upon this
model, position 2 > 1 > 3, does not correspond with the
observed Mug excision frequency for oligo A.
We then considered a third model (Figure 6C) in which
both base-stacking and base-pairing interactions are
disrupted for the target and adjacent base pairs. The
corresponding base-stacking and hydrogen-bonding in-
teraction energies would be -66.9, -58.9, and -60.9 kcal/
mol for positions 1, 2, and 3, respectively, predicting a
relative excision frequency of 2 > 3 > 1 (37). This trend
coincides with the observed relative Mug-mediated uracil
excision frequency of oligo A, suggesting that the prefer-
ence of Mug for excision of a target uracil residue is
inversely related to the energy required to disrupt both
hydrogen-bonding and base-stacking interactions of the
target and adjacent base pairs.
The variation of oligonucleotide melting temperature
with sequence has been attributed to both sequence-
dependent base-stacking and base-pair dependent hy-
drogen-bonding interactions (38-40). Several experimen-
tal data sets have been reported which have been used
to interpret the observed sequence-dependent differences
in thermostability. Doktycz et al. (41) measured the
melting temperatures of a series of octanucleotides in
which the terminal and central three base pairs were
systematically altered. It was demonstrated that this
data set could be used to predict the melting tempera-
tures of octanucleotides of any sequence (41). The melting
temperatures for a series of octanucleotides containing
the central trimers 3′CTT, TTT, and TTA, whose central
residues correspond to positions 1, 2, and 3, were reported
to be 36.5, 33.7, and 35.2 °C, respectively. Upon the basis
of the disruption of the trimer surrounding the target
base as diagramed in Figure 6C, this relative thermo-
stability predicts the relative Mug-mediated uracil exci-
sion frequency of oligo A to be position 2 > 3 > 1. This
trend is consistent with that obtained from the theoretical
studies discussed above for the third model.
Whereas the melting of short oligonucleotides is coop-
erative and fits well to a two-step denaturation model,
the melting behavior of longer oligonucleotides is more
complex. Local regions of lower thermostability melt with
greater ease, creating denatured bubbles that then
propagate throughout the sequence. We used the data
of Doktycz et al. to predict the relative thermostability
along the sequence of the oligonucleotide sequence uti-
lized in this study and shown in Figure 2. The melting
temperature of each overlapping duplex trimer (Table 1
in ref 41) was compared with the average duplex trimer
melting temperature for the 30-base oligonucleotide
duplex. Relative thermostability at each nucleotide posi-
tion was calculated by taking the difference between the
average and “local” thermostability. Differences in ther-
mostability across the 30-base oligonucleotide are plotted
in Figure 7. Positive values correspond to regions of lower
thermostability.
F igu r e 7. Thermal stability of each overlapping triplet duplex
relative to the average thermal stability of the triplets within
the 30-mer sequence corresponding to oligo A. The run of uracil
residues in oligo A is replaced with thymine residues in the
sequence seen above. The region in which the three target
uracils would be located is the least thermally stable region of
the oligonucleotide. Among the three target bases, indicated as
solid black bars, the relative order of decreased thermal stability
corresponds with the observed relative Mug excision frequency
of oligo A, position 2 > 3 > 1.
bases in the target region, the order of relative local
decreased thermostability is position 2 > 3 > 1 (Figure
7), which corresponds to the observed relative excision
efficiency demonstrated by Mug at each of the three
positions (Figure 5). Similar results were obtained with
the experimental oligonucleotide melting temperature set
of Gotoh et al. (42). While we do not wish to overinterpret
our data, the observed relative cleavage of the uracil
residues in the target oligonucleotide by Mug is best
explained by differences in local thermostability. These
data are in accord with our studies, discussed above, in
which the introduction of an N6-methyladenine residue
decreases the stability of the substitutent base pair,
rendering that uracil residue more susceptible to Mug
excision.
The model presented here is consistent with the pinch-
pull-push mechanism described by Mosbaugh and co-
workers for uracil-DNA glycosylase (43). Our model is
similar to thumbing through a book in search of the page
that will fall open due to the presence of a bookmarker.
In this analogy, the thickness of the bookmarker would
represent the degree of local instability of the target base.
Mug would thumb through the helix, inducing stress to
disrupt areas of the duplex with weaker base-stacking
and base-pairing interactions, causing that region of the
duplex to open up (Figure 6C, bottom). This mechanism
of recognition would provide a conceivably simple and
quick way for Mug to scan a duplex in search of its target
lesion.
The significance of thermal stability as a parameter
in substrate recognition for repair enzymes has been
discussed previously (15, 28, 44-46). Our results confirm
the importance of thermal stability as a recognition
mechanism for Mug. We provide a model where Mug
recognition is inversely correlated with local duplex
thermostability due to the base stacking of adjacent base
pairs as well as interstrand hydrogen bonding. Further-
more, the ability of Mug to target and preferentially
excise uracil paired with m6A over uracil paired with
adenine can be exploited as a method for locating m6A
in duplexes.
The region of the oligonucleotide in which the three
target uracils are located has the lowest relative ther-
mostability within the oligonucleotide. Among the three
Ack n ow led gm en t. This study was supported by
NIH Grants GM50351 and CA84487.

Effects of Local Duplex Stability and N6mA on Mug
Chem. Res. Toxicol., Vol. 15, No. 12, 2002 1601
propylphosphorodiamidite for in situ preparation of deoxyribo-
nucleoside phosphoramidites and their use in polymer-supported
synthesis of oligodeoxyribonucleotides. Nucleic Acids Res. 14,
7391-7403.
Note Ad d ed a fter ASAP . This article posted to the
Web as an ASAP article November 20, 2002. Subse-
quently, two corrections in the text were made. The
article reposted to the Web December 16, 2002.
(25) Gait, M. J . (1984) Oligonucleotide Synthesis, A Practical Approach,
IRL Press, Oxford, U.K.
(26) Dizdaroglu, M. (1985) Application of capillary gas chromatogra-
phy-mass spectrometry to chemical characterization of radiation-
induced base damage of DNA: Implications for assessing DNA
repair processes. Anal. Biochem. 144, 593-603.
(27) Fujimoto, J ., Tran, L., and Sowers, L. C. (1997) Synthesis and
cleavage of oligodeoxynucleotides containing a 5-hydroxyuracil
residue at a defined site. Chem. Res. Toxicol. 10, 1254-1258.
(28) Hang, B., Sagi, J ., and Singer, B. (1998) Correlation between
sequence-dependent glycosylase repair and the thermal stability
of oligonucleotide duplexes containing 1,N6-ethenoadenine. J .
Biol. Chem. 273, 33406-33413.
(29) Marky, L. A., and Breslauer, K. J . (1987) Calculating thermody-
namic data for transitions of any molecularity from equilibrium
melting curves. Biopolymers 26, 1601-1620.
(30) Taylor, I., Watts, D., and Kneale, G. (1993) Substrate recognition
and selectivity in the type IC DNA modification methylase
M.EcoR124I. Nucleic Acids Res. 21, 4929-4935.
(31) Kan, N. C., Lautenberger, J . A., Edgell, M. H., and Hutchison,
C. A., III (1979) The nucleotide sequence recognized by the
Escherichia coli K12 restriction and modification enzymes. J . Mol.
Biol. 130, 191-209.
(32) Bickle, T. A. (1982) The ATP-dependent restriction endonucleases.
Nucleases. Cold Spring Harbor Laboratory Press, Plainview, NY.
(33) Sowers, L. C., Shaw, B. R., and Sedwick, W. D. (1987) Base
stacking and molecular polarizability: Effect of a methyl group
in the 5-position of pyrimidines. Biochem. Biophys. Res. Commun.
148, 790-794.
(34) Roberts, R. J ., and Cheng, X. (1998) Base flipping. Annu. Rev.
Biochem. 67, 181-198.
(35) Slupphaug, G., Mol, C. D., Kavli, B., Arvai, A. S., Krokan, H. E.,
and Tainer, J . A. (1996) A nucleotide flipping mechanism from
the structure of human uracil-DNA glycosylase bound to DNA.
Nature 384, 87-92.
(36) Stivers, J . T., Pankiewicz, K. W., and Watanabe, K. A. (1999)
Kinetic mechanism of damage site recognition and uracil flipping
by Escherichia coli uracil DNA glycosylase. Biochemistry 38, 952-
963.
(37) Kollman, P. A., Weiner, P. K. and Dearing, A. (1981) Studies of
nucleotide conformations and interactions. The relative stabilities
of double-helical B-DNA sequence isomers. Biopolymers 20, 2583-
2621.
(38) Devoe, H., and Tinoco, I., J r. (1962) The stability of helical
polynucleotides: Base contributions. J . Mol. Biol. 4, 500-517.
(39) Crothers, D. M., and Zimm, B. H. (1964) Theory of the melting
transition of synthetic polynucleotides: Evaluation of stacking
free energy. J . Mol. Biol. 9, 1-9.
(40) Petruska, J ., Sowers, L. C., and Goodman, M. F. (1986) Compari-
son of nucleotide interactions in water, proteins, and vacuum:
Model of DNA polymerase fidelity. Proc. Natl. Acad. Sci. U.S.A.
83, 1559-1562.
(41) Doktycz, M. J ., Morris, M. D., Dormady, S. J ., Beattie, K. L., and
J acobson, K. B. (1995) Optical melting of 128 octamer DNA
duplexes. Effects of base pair location and nearest neighbors on
thermal stability. J . Biol. Chem. 270, 8439-8445.
(42) Gotoh, O., and Tagashira, Y. (1981) Stabilities of nearest-neighbor
doublets in double-helical DNA determined by fitting calculated
melting profiles to observed profiles. Biopolymers 20, 1033-
1042.
(43) Wong, I., Lundquist, A. J ., Bernards, A. S., and Mosbaugh, D.
W. (2002) Presteady-state analysis of a single catalytic turnover
by Escherichia coli uracil-DNA glycosylases reveals a “Pinch-Pull-
Push” mechanism. J . Biol. Chem. 277, 19424-19432.
(44) Sagi, J ., Hang, B., and Singer, B. (1999) Sequence-dependent
repair of synthetic AP sites in 15-mer and 35-mer oligonucle-
otides: role of thermodynamic stability imposed by neighbor
bases. Chem. Res. Toxicol. 10, 917-932.
Refer en ces
(1) Ames, B. N., Shigenaga, M. K., and Hagen, T. M. (1993) Oxidants,
antioxidants, and the degenerative diseases of aging. Proc. Natl.
Acad. Sci. U.S.A. 90, 7915-7922.
(2) Mullaart, E., Lohman, P. H. M., Berends, F., and Vijg, J . (1990)
DNA damage metabolism and aging. Mutat. Res. 237, 189-210.
(3) Loeb, L. A. (1989) Endogenous carcinogenesis: Molecular oncology
into the twenty-first century - Presidential address. Cancer Res.
49, 5489-5496.
(4) Lindahl, T. (1993) Instability and decay of the primary structure
of DNA. Nature 362, 709-715.
(5) Lindahl, T., and Wood, R. D. (1999) Quality control by DNA
repair. Science 286, 1897-1905.
(6) Cleaver, J . E., Karplus, K., Kashani-Sabet, M., and Limoli, C. L.
(2001) Nucleotide excision repair “a legacy of creativity”. Mutat.
Res. 485, 23-36.
(7) Hoeijmakers, J . H. J . (2001) Genome maintenance mechanisms
for preventing cancer. Nature 411, 366-374.
(8) Thompson, L. H., and Schild, D. (2001) Homologous recombina-
tional repair of DNA ensures mammalian chromosome stability.
Mutat. Res. 477, 131-153.
(9) Seeberg, E., Eide, L., and Bjoras, M. (1995) The base excision
repair pathway. Trends Biochem. Sci. 20, 391-397.
(10) Gallinari, P., and J iricny, J . (1996) A new class of uracil-DNA
glycosylases related to human thymine-DNA glycosylase. Nature
383, 735-738.
(11) Pearl, L. H. (2000) Structure and function in the uracil-DNA
glycosylase superfamily. Mutat. Res. 460, 165-181.
(12) Barrett, T. E., Savva, R., Panayotou, G., Barlow, T., Brown, T.,
J iricny, J ., and Pearl, L. H. (1998) Crystal structure of a G:T/U
mismatch-specific DNA glycosylase: mismatch recognition by
complementary-strand interactions. Cell 92, 117-129.
(13) Barrett, T. E., Scharer, O. D., Savva, R., Brown, T., J iricny, J .,
Verdine, G. L., and Pearl, L. H. (1999) Crystal structure of a
thwarted mismatch glycosylase DNA repair complex. EMBO J .
18, 6599-6609.
(14) Saparbaev, M., and Laval, J . (1998) 3,N4-ethenocytosine, a highly
mutagenic adduct, is a primary substrate for Escherichia coli
double-stranded uracil-DNA glycosylase and human mismatch-
specific thymine DNA-glycosylase. Proc. Natl. Acad. Sci. U.S.A.
95, 8508-8513.
(15) Liu, P., Burdzy, A., and Sowers, L. C. (2002) Substrate recognition
by a family of uracil-DNA glycosylases: UNG, MUG and TDG.
Chem. Res. Toxicol. 15, 1001-1009.
(16) Sung, J . S., and Mosbaugh, D. W. (2000) Escherichia coli double-
strand uracil DNA glycosylase: involvement in uracil-mediated
DNA base excision repair and stimulation of activity by endonu-
clease IV. Biochemistry 39, 10224-10235.
(17) Marinus, M. G., and Morris, N. R. (1973) Isolation of deoxyribo-
nucleic acid methylase mutants of Escherichia coli K-12. J .
Bacteriol. 114, 1143-1150.
(18) Pukkila, P. J ., Peterson, J ., Herman, G., Modrich, P., and
Meselson, M. (1983) Effects of high levels of DNA adenine
methylation on methyl-directed mismatch repair in Escherichia
coli. Genetics 104, 571-582.
(19) Radman, M., and Wagner, R. (1984) Effects of DNA methylation
on mismatch repair, mutagenesis, and recombination in Escheri-
chia coli. Curr. Top. Microbiol. Immunol. 108, 23-28.
(20) Engel, J . D., and von Hippel, P. H. (1978) Effects of methylation
on the stability of nucleic acid conformations. Studies at the
polymer level. J . Biol. Chem. 253, 927-934.
(21) Sternglanz, H., and Bugg, C. E. (1973) Conformations of N6-
monosubstituted adenine derivatives. Crystal structure of N6-
methyladenine. Biochim. Biophys. Acta 308, 1-8.
(22) Quignard, E., Fazakerley, G. V., Teoule, R., Guy, A., and Guschl-
bauer, W. (1985) Consequences of methylation on the amino group
of adenine. A proton two-dimensional NMR study of d(GGATATCC)
and d(GGm6ATATCC). Eur. J . Biochem. 152, 99-105.
(23) Ono, A., and Ueda, T. (1987) Synthesis of decadeoxyribonucle-
otides containing N6-methyladenine, N4-methylcytosine, and
5-methylcytosine: recognition and cleavage by restriction endo-
nucleases (nucleosides and nucleotides part 74). Nucleic Acids Res.
15, 219-232.
(45) Sagi, J ., Perry, A., Hang, B., and Singer, B. (2000) Differential
destabilization of the DNA oligonucleotide double helix by a T:G
mismatch, 3,N4-ethenocytosine, 3,N4-ethanocytosine, or an 8-(hy-
droxymethyl)-3,N4-ethenocytosine adduct incorporated into the
same sequence contexts. Chem. Res. Toxicol. 13, 839-845.
(46) Biswas, T., Clos, L. J ., II, SantaLucia, J ., J r., Mitra, S., and Roy,
R. (2002) Binding of specific DNA base-pair mismatches by
N-methylpurine-DNA glycosylase and its implication in initial
damage recognition. J . Mol. Biol. 320, 503-513.
(24) Nielsen, J ., Taagaard, M., Marugg, J . E., van Boom, J . H., and
Dahl, O. (1986) Application of 2-cyanoethyl-N,N,N′,N′-tetraiso-
TX020062Y