JOURNAL OF BACTERIOLOGY, Aug. 2011, p. 3815–3821
Copyright © 2011, American Society for Microbiology. All Rights Reserved.
Vol. 193, No. 15
Role of High-Fidelity Escherichia coli DNA Polymerase I in Replication
Bypass of a Deoxyadenosine DNA-Peptide Cross-Link?
Kinrin Yamanaka,1,2Irina G. Minko,1Steven E. Finkel,4
Myron F. Goodman,4,5and R. Stephen Lloyd1,3*
Center for Research on Occupational and Environmental Toxicology,1Department of Physiology and Pharmacology,2and
Department of Molecular and Medical Genetics,3Oregon Health & Science University, Portland, Oregon 97239, and
Department of Biological Sciences4and Department of Chemistry,5University of Southern California,
Los Angeles, California 90089
Received 23 December 2010/Accepted 14 May 2011
Reaction of bifunctional electrophiles with DNA in the presence of peptides can result in DNA-peptide
cross-links. In particular, the linkage can be formed in the major groove of DNA via the exocyclic amino group
of adenine (N6-dA). We previously demonstrated that an A family human polymerase, Pol ?, can efficiently and
accurately synthesize DNA past N6-dA-linked peptides. Based on these results, we hypothesized that another
member of that family, Escherichia coli polymerase I (Pol I), may also be able to bypass these large major
groove DNA lesions. To test this, oligodeoxynucleotides containing a site-specific N6-dA dodecylpeptide cross-
link were created and utilized for in vitro DNA replication assays using E. coli DNA polymerases. The results
showed that Pol I and Pol II could efficiently and accurately bypass this adduct, while Pol III replicase, Pol IV,
and Pol V were strongly inhibited. In addition, cellular studies were conducted using E. coli strains that were
either wild type or deficient in all three DNA damage-inducible polymerases, i.e., Pol II, Pol IV, and Pol V.
When single-stranded DNA vectors containing a site-specific N6-dA dodecylpeptide cross-link were replicated
in these strains, the efficiencies of replication were comparable, and in both strains, intracellular bypass of the
lesion occurred in an error-free manner. Collectively, these findings demonstrate that despite its constrained
active site, Pol I can catalyze DNA synthesis past N6-dA-linked peptide cross-links and is likely to play an
essential role in cellular bypass of large major groove DNA lesions.
DNA-protein cross-links represent a class of DNA lesions
that are formed as a consequence of endogenous metabolic
processes and exposure to various chemical toxicants, such as
acrolein (1, 18). Although the prokaryotic nucleotide excision
repair machinery can function directly on a subset of DNA-
protein cross-links, it has been demonstrated to act more effi-
ciently on DNA-peptide cross-links, the proteolytic degrada-
tion products of DNA-protein cross-links (20, 22). However,
little is known about the replication past DNA-peptide cross-
link lesions. This lack of knowledge is an obstacle to under-
standing how cells tolerate these lesions.
Unrepaired DNA-peptide cross-links are hypothesized to be
substrates for replication bypass by DNA polymerases. Previ-
ously, we demonstrated that when a single-stranded shuttle
vector containing a major groove-linked DNA-peptide cross-
link was replicated in African green monkey kidney cells
(COS7 cells), this lesion was only marginally miscoding (19).
Such data show that in a mammalian system, a subset of DNA-
peptide cross-links can be bypassed in a relatively error-free
manner. Germane to these observations, we recently demon-
strated that polymerase ? (Pol ?), an A family human polymerase
possessing in vitro translesion DNA synthesis (TLS) activity, can
efficiently and accurately bypass an N6-dA dodecylpeptide cross-
link lesion (Fig. 1) (31).
In Escherichia coli, Pol III carries out DNA synthesis at
the replication fork, while specialized DNA damage-induc-
ible polymerases, i.e., Pol II, Pol IV, and Pol V, mainly
conduct replication bypass of DNA lesions (10). Pol I, a
member of the A family of DNA polymerases, is the most
abundant DNA polymerase in E. coli in the absence of
exogenous stressors and has many important cellular func-
tions, including the processing of Okazaki fragments and
DNA repair (25). In addition, Pol I has been shown to be
capable of catalyzing replication bypass of various N6-dA-
linked DNA lesions, ranging from a small styrene oxide-
induced DNA adduct to more bulky benzo[a]pyrene-in-
duced DNA adducts (5, 17).
To determine if high-fidelity bypass of an N6-dA dodecyl-
peptide cross-link by human Pol ? could be extrapolated to
other A family polymerases, we evaluated the ability of E. coli
Pol I and its exonuclease-deficient Klenow fragment (KFexo?)
to replicate past this lesion in vitro, and for comparative pur-
poses, we conducted similar analyses with all other E. coli
DNA polymerases. In addition, the efficiency and fidelity of
intracellular TLS past the cross-link were assessed using either
wild-type E. coli or an isogenic triple mutant that lacks DNA
MATERIALS AND METHODS
Materials. [?-32P]ATP was purchased from PerkinElmer Life Sciences (Wal-
tham, MA). P-6 Bio-Spin columns were obtained from Bio-Rad (Hercules, CA).
T4 polynucleotide kinase, Pol I (10,000 U/ml), and KFexo?(5,000 U/ml) were
* Corresponding author. Mailing address: Center for Research on
Occupational and Environmental Toxicology, Oregon Health & Sci-
ence University, Mail Code L606, 3181 SW Sam Jackson Park Rd.,
Portland, OR 97239. Phone: (503) 494-9957. Fax: (503) 494-6831.
?Published ahead of print on 27 May 2011.
purchased from New England BioLabs (Beverly, MA). Sodium cyanoborohy-
dride was obtained from Sigma (St. Louis, MO). Slide-A-Lyzer dialysis cassettes
with a molecular weight cutoff of 10,000 were purchased from Thermo Scientific
(Rockford, IL). The peptide Lys-Phe-His-Glu-Lys-His-His-Ser-His-Arg-Gly-Tyr
(KFHEKHHSHRGY) was obtained from Sigma-Genosys (St. Louis, MO). E.
coli Pol II (4), Pol IV (14, 29), UmuD?2C (3), and RecA (6) were purified as
previously described. The Pol III replicase (Pol III*), a holoenzyme that contains
two molecules of Pol III core connected by the ? complex clamp loader but lacks
the ? subunit sliding clamp, was a generous gift from Michael O’Donnell (Rocke-
feller University, New York, NY) and was purified and reconstituted as previ-
ously described (21, 24, 26, 28). Pol ? was purchased from Enzymax, LLC
Bacterial strains. The Pol II-, IV-, and V-deficient strain SF2018 (ZK126
polB::SpcrdinB::KanrumuDC::Camr) is derived from the E. coli K-12 strain
ZK126 (W3110 ?lacU169 tna-2) (33). This triple mutant strain was constructed
by bacteriophage P1 transduction into ZK126, using the following donor strains:
SF2003 (polB::Spcr), SF2006 (dinB::Kanr), and SF2009 (umuDC::Camr) (32).
The original donor alleles and strains for the polymerase mutants mentioned
above were as follows: for SF2003, SH2101 (polB?1::? Sm-Sp) (2); for SF2006,
RW626 (dinB::Kan); and for SF2009, RW82 (umuDC::Cam) (both RW626 and
RW82 were generous gifts from Roger Woodgate [National Institutes of Health,
Bethesda, MD]). The triple mutant was selected by growing cells in the presence
of spectinomycin (100 ?g/ml), chloramphenicol (30 ?g/ml), and kanamycin (50
Oligodeoxynucleotide synthesis. Nondamaged oligodeoxynucleotides were
synthesized by the Molecular Microbiology and Immunology Research Core
Facility at Oregon Health & Science University (Portland, OR). An oligodeoxy-
nucleotide containing ?-HO-PdA (?-hydroxypropanodeoxyadenosine) was a
generous gift from Carmelo J. Rizzo (Vanderbilt University, Nashville, TN). An
oligodeoxynucleotide containing an N6-dA dodecylpeptide cross-link (Fig. 1)
(5?-GCTAGTACTCGTCGACAATTCCGTATCCAT-3?) at the underlined nu-
cleotide was prepared according to a previously published procedure (31).
DNA replication assays. For replication assays, primers were designed for
conditions with either a running or standing start, where the primer 3?-OH was
positioned three nucleotides upstream of the lesion site (?3 primer) or imme-
diately prior to the lesion site (?1 primer), respectively. The sequences of the ?3
and ?1 primers were 5?-AAAATGGATACGGAAT-3? and 5?-AAAATGGAT
ACGGAATTG-3?, respectively. DNA replication assays with Pol III*, Pol I,
KFexo?, Pol II, Pol IV, and human Pol ? were carried out in reaction mixtures
containing 25 mM Tris-HCl (pH 7.5), 10% (vol/vol) glycerol, 100 ?g/ml bovine
serum albumin, and 5 mM dithiothreitol. The Pol III*-, Pol I-, KFexo?-, and Pol
?-catalyzed reactions were carried out in the presence of 8 mM MgCl2; the Pol
II- and Pol IV-catalyzed reactions were carried out in the presence of 5 mM
MgCl2. DNA replication assays with UmuD?2C were carried out in reaction
mixtures containing 20 mM Tris-HCl (pH 7.5), 8 mM MgCl2, 4% (vol/vol)
glycerol, 10 mM sodium glutamate, 5 mM dithiothreitol, 100 ?M EDTA, 1 ?M
single-stranded 36-mer oligodeoxynucleotide, 6 ?M RecA, and 0.5 mM adeno-
sine 5?-[?-thio]triphosphate. The single-stranded 36-mer oligodeoxynucleotide
and RecA were preincubated in reaction buffer in the presence of adenosine
5?-[?-thio]triphosphate at 37°C for 3 min. Reactions were initiated by the addi-
tion of deoxynucleoside triphosphates (dNTPs) and UmuD?2C. All reactions
were carried out at 37°C. DNA substrate, DNA polymerase, and dNTP concen-
trations, as well as reaction times, are given in the figure legends. Reaction
products were resolved through 15% acrylamide denaturing gels containing 8 M
urea and were visualized using a PhosphorImager screen. The percentage of
primer extension was determined by calculating the ratio of the amount of
extended primers beyond the adducted or corresponding unadducted base to the
total amount of primers (unextended primers plus extended primers up to and
opposite the site of the adduct plus extended primers beyond the site of the
adduct). Quantifications were done using ImageQuant 5.2 software.
Replication of DNAs containing site-specific N6-dA dodecylpeptide cross-link
in E. coli. The pMS2 shuttle vector was the generous gift of Masaaki Moriya
(State University of New York, Stony Brook, NY). A single-stranded pMS2
vector containing a site-specific N6-dA dodecylpeptide cross-link (pMS2-XL12)
was constructed following a previously published protocol (21). Initially, individ-
ual transformations with either unadducted reference pBR322 plasmid or pMS2-
XL12 plasmid were carried out using a wild-type E. coli strain to determine the
ratio of plasmid concentrations that would result in approximately equal num-
bers of transformants. The exact ratio of plasmids used for the transformations
was not possible to calculate because following ligation of the lesion-containing
insert into the pMS2 vector, the percentage of closed circular plasmids with the
insert could not be determined accurately.
Following determination of a desirable ratio, the unadducted pBR322 and
lesion-containing pMS2-XL12 plasmids were mixed and used to transform wild-
type E. coli and an isogenic strain (SF2018 [Pol II?Pol IV?Pol V?]). The
transformants were selected overnight on LB agar plates containing ampicillin,
and individual colonies were grown first in LB broth containing ampicillin (100
?g/ml) in 96-well plates at 37°C. After 6 h of incubation, a 20-?l aliquot from
each well was transferred to other plates containing LB broth with tetracycline
(12.5 ?g/ml) and incubated overnight at 37°C. This procedure allowed the dis-
tinction of the tetracycline-sensitive pMS2 transformants from the tetracycline-
resistant pBR322 transformants. In order to identify the population of transfor-
mants that originated from insert-containing pMS2-XL12 vectors, tetracycline-
sensitive colonies were subjected to differential hybridization as previously
described (21). As a probe, we used the oligodeoxynucleotide 5?-TACCAGCG
ATGCTAGTACT-3?, which hybridizes to the 5? junction of the pMS2 backbone
and the lesion-containing 30-mer oligodeoxynucleotide insert. To ensure the
accuracy of this hybridization screen, plasmids from approximately 10% of col-
onies that did not hybridize with this probe were isolated and sequenced. The
numbers of pBR322 and pMS2-XL12 transformants in the wild-type and Pol II?
Pol IV?Pol V?strains were determined, and the relative transformation effi-
ciency of pMS2-XL12 in each strain was calculated by normalizing the transfor-
mation efficiency of pMS2-XL12 to that of pBR322. Mutational analyses were
carried out by transforming the wild-type and Pol II?Pol IV?Pol V?strains
with a pMS2 vector containing either an unmodified or N6-dA dodecylpeptide
cross-link-containing 30-mer oligonucleotide. The probe that was complemen-
tary to the DNA sequences encompassing the lesion, assuming no mutations
were introduced, was used for differential hybridization. Vectors that did not
hybridize to the probe were isolated by use of Qiagen plasmid minipreparation
kits and digested with ScaI restriction endonuclease. Since both the original
pMS2 vector and the insert had an ScaI recognition site, two DNA fragments
were generated (?1.9 and ?3.2 kb) from vectors that contained inserted se-
quences. Subsequently, such vectors were analyzed by DNA sequencing.
Pol III*-catalyzed replication is blocked by N6-dA dodecyl-
peptide cross-link in vitro. It has been shown previously that
Saccharomyces cerevisiae replicative Pol ? cannot efficiently
bypass an N6-dA dodecylpeptide cross-link (31). However,
the ability of prokaryotic replicative polymerases to bypass
this lesion has not been investigated. Therefore, primer
extension reactions were conducted using nondamaged and
N6-dA dodecylpeptide cross-link-containing templates with
Pol III* (holoenzyme that lacks the ? subunit sliding clamp).
The results revealed that Pol III* could not bypass the
cross-link because it was severely blocked at a site one
nucleotide prior to the lesion, with only 2% of primers
extended beyond the adducted site, compared to 65% with
an unadducted template (Fig. 2).
Replication bypass of N6-dA dodecylpeptide cross-link by
Pol I and KFexo?in vitro. In order to test the hypothesis that
Pol I could be involved in replication bypass of the N6-dA
FIG. 1. Structure of a ?-HO-PdA-derived peptide cross-link. The
DNA-peptide cross-link consists of a dodecyl-(Lys-Phe-His-Glu-Lys-
His-His-Ser-His-Arg-Gly-Tyr) peptide attached via an acrolein moiety
at the N-6 position of dA.
3816 YAMANAKA ET AL.J. BACTERIOL.
dodecylpeptide cross-link, primer extension reactions were
carried out using Pol I. As shown in Fig. 3A, Pol I fully ex-
tended the primers annealed to the adducted template; under
the conditions used, minimal blockage of DNA synthesis was
observed one nucleotide prior to the lesion. As evident from
the bar graph, the percentages of primers that were extended
beyond the adducted site and the corresponding unadducted
site were comparable; specifically, 63% and 51% primer exten-
sion was observed with a low concentration of Pol I on unad-
ducted and adducted templates, respectively. Thus, Pol I can
catalyze replication bypass of a large N6-dA peptide cross-link.
In order to assess the identity of the nucleotide incorporated
opposite the lesion, single-nucleotide incorporation reactions
were conducted with KFexo?. KFexo?was used instead of Pol
I in this experiment because the proofreading exonuclease
activity of Pol I could potentially remove the misincorporated
nucleotide, and thus the fidelity of nucleotide incorporation
would be overestimated. In the presence of a 1 ?M concen-
tration of each dNTP individually, only the correct nucleotide,
dT, was incorporated opposite the lesion (Fig. 3B), demon-
strating that this cross-link can be bypassed accurately by
KFexo?and suggesting that Pol I-catalyzed replication past
the lesion is error-free.
Replication bypass of N6-dA dodecylpeptide cross-link by
DNA damage-inducible polymerases Pol II, Pol IV, and Pol V
in vitro. Since DNA damage-inducible polymerases are known
to be involved in bypass of many lesions, the ability of these
polymerases to bypass DNAs containing an N6-dA dodecyl-
peptide cross-link was examined. Primer extension reactions
were carried out with Pol II, Pol IV, and Pol V (UmuD?2C and
RecA complex). Pol IV could not efficiently bypass this lesion.
It was strongly blocked at a site one nucleotide prior to the
lesion, with only 11% of primers being extended beyond the
adducted site, even at the highest concentration of Pol IV used.
This is in contrast to 90% primer extension beyond the corre-
sponding unadducted site (Fig. 4).
Pol V also catalyzed inefficient bypass of the N6-dA dodecyl-
peptide cross-link. It could extend primers annealed to the ad-
FIG. 2. Replication blockage of Pol III* by N6-dA dodecylpeptide
cross-link. Primer extension reactions were catalyzed by 2 nM Pol III*
under running-start conditions (?3 primer), using nondamaged (ND)
and N6-dA dodecylpeptide cross-linked (XL12) substrates. Reactions
were carried out in the presence of 2 nM primer-template and a 200
?M concentration of each dNTP for 20 min at 37°C. A*, adducted site.
The bar graph indicates the percentage of primers extended beyond
the unadducted or adducted site. ND1, primer extension reactions with
nondamaged substrate in the absence of Pol III*; ND2, primer exten-
sion reactions with nondamaged substrate in the presence of Pol III*;
XL1, primer extension reactions with N6-dA dodecylpeptide cross-
linked substrate in the absence of Pol III*; XL2, primer extension
reactions with N6-dA dodecylpeptide cross-linked substrate in the
presence of Pol III*.
FIG. 3. Replication bypass of N6-dA dodecylpeptide cross-link by
Pol I and KFexo?. (A) Primer extension reactions were catalyzed by
0.32 mU or 1.6 mU Pol I under running-start conditions (?3 primer),
using nondamaged (ND) and N6-dA dodecylpeptide cross-linked
(XL12) substrates in the presence of a 200 ?M concentration of each
dNTP. The bar graph indicates the percentage of primers extended
beyond the unadducted or adducted site. ND1, primer extension re-
actions with nondamaged substrate in the absence of Pol I; ND2,
primer extension reactions with nondamaged substrate in the presence
of 0.32 mU Pol I; ND3, primer extension reactions with nondamaged
substrate in the presence of 1.6 mU Pol I; XL1, primer extension
reactions with N6-dA dodecylpeptide cross-linked substrate in the ab-
sence of Pol I; XL2, primer extension reactions with N6-dA dodecyl-
peptide cross-linked substrate in the presence of 0.32 mU Pol I; XL3,
primer extension reactions with N6-dA dodecylpeptide cross-linked
substrate in the presence of 1.6 mU Pol I. (B) Single-nucleotide in-
corporation reactions and primer extension reactions were catalyzed
by 0.1 mU KFexo?under standing-start conditions (?1 primer), using
the same substrates as those described for panel A in the presence of
a 1 ?M concentration of individual or all dNTPs. All reactions were
carried out in the presence of 2 nM primer-template for 30 min at
37°C. A*, adducted site.
VOL. 193, 2011 BYPASS OF DNA-PEPTIDE CROSS-LINK BY DNA POLYMERASE I3817
ducted template up to the lesion, but the percentage of primers
extended beyond the adducted site was only 4%, compared to
44% beyond the unadducted site (Fig. 5A). However, this limited
replication bypass appeared to be accurate, because only the
correct nucleotide (dT) was incorporated opposite the lesion
Relative to Pol IV and Pol V, Pol II was significantly more
competent in bypass of the N6-dA dodecylpeptide cross-link.
In particular, using a low concentration of Pol II, the percent-
ages of primers extended beyond the corresponding unad-
ducted and adducted sites were 48% and 13%, respectively
(Fig. 6A). As shown in Fig. 6B, single-nucleotide incorporation
reactions revealed that only the correct nucleotide (dT) was
incorporated opposite the lesion. Collectively, these data sug-
gest that Pol II could be another polymerase involved in rep-
lication bypass of the N6-dA dodecylpeptide cross-link.
Efficiency and accuracy of replication of plasmid DNAs con-
taining N6-dA dodecylpeptide cross-link in vivo. Although our
in vitro results showed that Pol II was able to bypass the N6-dA
dodecylpeptide cross-link lesion, Pol I manifested a greater
ability. In contrast to Pol II, Pol I is known to be expressed
constitutively. Therefore, DNA damage-inducible polymerases
may not be required for the replication bypass of this lesion,
and Pol I may play a major intracellular role in catalyzing
replication bypass of such large peptide cross-links.
In order to test whether DNA damage-inducible poly-
merases are required for replication bypass of the N6-dA do-
decylpeptide cross-link, 30-mer oligodeoxynucleotides contain-
ing this lesion were ligated into the single-stranded pMS2
vector. A mixture of lesion-containing pMS2-XL12 vector and
pBR322 nondamaged reference vector was prepared and
transformed into either wild-type or Pol II?PolIV?PolV?E.
coli. The transformants were then selected for ampicillin resis-
tance and tested for tetracycline resistance to identify tetracy-
cline-sensitive pMS2-containing clones. Furthermore, colonies
were identified by differential hybridization using a probe that
hybridizes to the 5? junction of pMS2 DNA and the insert
sequence. The colonies containing the progeny of either
pBR322 or pMS2-XL12 were counted, and the transformation
efficiency of pMS2-XL12 vector relative to that of pBR322
vector was calculated for each strain. No differences in the
transformation efficiency of pMS2-XL12 vector were observed
between the wild-type and Pol II?PolIV?PolV?strains (Fig.
7). These data demonstrate that DNA damage-inducible poly-
FIG. 4. Replication blockage of Pol IV by N6-dA dodecylpeptide
cross-link. Primer extension reactions were catalyzed by 1.25 nM and
5 nM Pol IV under running-start conditions (?3 primer), using non-
damaged (ND) and N6-dA dodecylpeptide cross-linked (XL12) sub-
strates. Reactions were carried out in the presence of 5 nM primer-
template and a 100 ?M concentration of each dNTP for 20 min at
37°C. A*, adducted site. The bar graph indicates the percentage of
primers extended beyond the unadducted or adducted site. ND1,
primer extension reactions with nondamaged substrate in the absence
of Pol IV; ND2, primer extension reactions with nondamaged sub-
strate in the presence of 1.25 nM Pol IV; ND3, primer extension
reactions with nondamaged substrate in the presence of 5 nM Pol IV;
XL1, primer extension reactions with N6-dA dodecylpeptide cross-
linked substrate in the absence of Pol IV; XL2, primer extension
reactions with N6-dA dodecylpeptide cross-linked substrate in the
presence of 1.25 nM Pol IV; XL3, primer extension reactions with
N6-dA dodecylpeptide cross-linked substrate in the presence of 5 nM
FIG. 5. Replication bypass of N6-dA dodecylpeptide cross-link by
Pol V. (A) Primer extension reactions were catalyzed by 200 nM
UmuD?2C under running-start conditions (?3 primer), using nondam-
aged (ND) and N6-dA dodecylpeptide cross-linked (XL12) substrates
in the presence of a 500 ?M concentration of each dNTP, activated
RecA, and 5 nM primer-template. The bar graph indicates the per-
centage of primers extended beyond the unadducted or adducted site.
ND1, primer extension reactions with nondamaged substrate in the
absence of UmuD?2C; ND2, primer extension reactions with nondam-
aged substrate in the presence of UmuD?2C; XL1, primer extension
reactions with N6-dA dodecylpeptide cross-linked substrate in the ab-
sence of UmuD?2C; XL2, primer extension reactions with N6-dA do-
decylpeptide cross-linked substrate in the presence of UmuD?2C.
(B) Single-nucleotide incorporation reactions and primer extension
reactions were catalyzed by 200 nM UmuD?2C under standing-start
conditions (?1 primer), using the same substrates as those described
for panel A in the presence of a 100 ?M concentration of individual or
all dNTPs, activated RecA, and 2 nM primer-template. All reactions
were carried out for 20 min at 37°C. A*, adducted site.
3818YAMANAKA ET AL.J. BACTERIOL.
merases, including Pol II, are dispensable for bypass of an
N6-dA dodecylpeptide cross-link in vivo.
In vitro, Pol I bypassed the N6-dA dodecylpeptide cross-link
in an error-free manner. In order to test whether accurate
bypass of this lesion also takes place within cells, mutational
analyses were conducted by transforming pMS2-XL12 vector
into the wild type or the Pol II?PolIV?PolV?strain. After
overnight selection in the presence of ampicillin, 95 and 96
colonies of wild-type and Pol II?PolIV?PolV?cells, respec-
tively, were analyzed for mutations as described in Materials
and Methods. In parallel experiments, 89 and 96 individual
clones of wild-type and Pol II?PolIV?PolV?cells, respec-
tively, were tested following replication of the corresponding
nondamaged control vectors. The results revealed that no mu-
tations were introduced opposite either nondamaged or dam-
aged dA, demonstrating the accuracy of TLS past the N6-dA
dodecylpeptide cross-link in vivo (data not shown).
Pol I is known to be essential for ColE1-type plasmid rep-
lication (8, 15). Since pMS2 possesses a ColE1 origin (9), a Pol
I-defective strain could not be used for the plasmid-based
approach described above to directly assess the role of Pol I in
bypass of DNA lesions. However, given the strong inhibitory
effect of the N6-dA dodecylpeptide cross-link on Pol III*-
catalyzed DNA synthesis in vitro (Fig. 2) and the efficient
replication of plasmids containing this lesion in the Pol II?
PolIV?PolV?strain (Fig. 7), it can be inferred that Pol I likely
plays a key role in bypass of such large major groove peptide
In the current investigation, we demonstrated that Pol I and
Pol II could replicate DNAs containing an N6-dA dodecylpep-
tide cross-link in an error-free manner in vitro, while Pol III*
and Pol IV were strongly inhibited by this type of lesion.
Although Pol V could fully and accurately extend a primer on
a damaged substrate, its efficiency was low. Since Pol I is
expressed constitutively, we proposed that Pol I in cells can
carry out TLS past an N6-dA dodecylpeptide cross-link and
that the DNA damage-inducible polymerases are not required.
Consistent with this idea, in vivo studies using wild-type and
Pol II?Pol IV?Pol V?E. coli strains revealed that DNA
damage-inducible polymerases were dispensable.
Although these data suggested the essential role of Pol I in
replication bypass of the N6-dA dodecylpeptide cross-link, an
involvement of alternative DNA polymerases in this process
cannot be ruled out completely. In particular, Pol III and
FIG. 6. Replication bypass of N6-dA dodecylpeptide cross-link by
Pol ??. (A) Primer extension reactions were catalyzed by 0.5 nM and 5
nM Pol II under running-start conditions (?3 primer), using nondam-
aged (ND) and N6-dA dodecylpeptide cross-linked (XL12) substrates
in the presence of a 250 ?M concentration of each dNTP and 5 nM
primer-template. The bar graph indicates the percentage of primers
extended beyond the unadducted or adducted site. ND1, primer ex-
tension reactions with nondamaged substrate in the absence of Pol II;
ND2, primer extension reactions with nondamaged substrate in the
presence of 0.5 nM Pol II; ND3, primer extension reactions with
nondamaged substrate in the presence of 5 nM Pol II; XL1, primer
extension reactions with N6-dA dodecylpeptide cross-linked substrate
in the absence of Pol II; XL2, primer extension reactions with N6-dA
dodecylpeptide cross-linked substrate in the presence of 0.5 nM Pol II;
XL3, primer extension reactions with N6-dA dodecylpeptide cross-
linked substrate in the presence of 5 nM Pol II. (B) Single-nucleotide
incorporation reactions and primer extension reactions were catalyzed
by 0.5 nM Pol II under standing-start conditions (?1 primer), using the
same substrates as those described for panel A in the presence of a 250
?M concentration of individual or all dNTPs. All reactions were car-
ried out and 2 nM primer template for 20 min at 37°C. A*, adducted
FIG. 7. Ability of E. coli strains to replicate DNA containing N6-dA
dodecylpeptide cross-link. A mixture of nondamaged pBR322 vector
and site-specifically modified single-stranded pMS2 vector (pMS2-
XL12) was transformed into an E. coli wild-type strain and a mutant
strain lacking all DNA damage-inducible polymerases (Pol II?Pol
IV?Pol V?). The transformation efficiency of pMS2-XL12 vector (a
function of the replication bypass efficiency) was measured for each
strain. The data were normalized using the transformation efficiency of
pBR322 vector as a reference. The data were obtained from three
independent experiments. Error bars represent standard errors.
VOL. 193, 2011 BYPASS OF DNA-PEPTIDE CROSS-LINK BY DNA POLYMERASE I3819
possibly other DNA polymerases may manifest higher bypass
efficiencies in cells than those observed in vitro due to stimu-
lation of bypass by the ? sliding clamp and additional accessory
factors. It is also important to recognize that the division of
labor between DNA polymerases in replication bypass is ex-
pected to be more complex in the context of chromosomal
DNA, as opposed to a site-specific single lesion on plasmid
DNA. Following chemical exposure, multiple structurally dif-
ferent DNA lesions are likely to be formed that may trigger the
SOS response. The contribution of Pol II to TLS past N6-dA
polypeptide cross-links could be quite prominent under such
conditions, since its intracellular levels are significantly in-
creased (10). Pol IV, the most abundant E. coli DNA polymer-
ase in stressed cells (10), may also be involved in replication
bypass of lesions, displacing more efficient but less abundant
DNA polymerases from the primer termini. Such a phenome-
non was recently observed during double-strand-break repair
(11). In addition, the location of the lesions on the chromo-
some may also influence the choice of polymerase, similar to a
site-specific polymerase switch during spontaneous mutagene-
The active site of Pol I is constrained, and Pol I has a
stringent requirement for accommodating nucleobase pairs
with the correct Watson-Crick geometry (13, 30). Therefore, it
is intriguing that Pol I was capable of bypassing a very large
N6-dA dodecylpeptide cross-link that not only blocked repli-
cative polymerases (E. coli Pol III* and yeast Pol ?) but also
blocked the low-fidelity polymerases specialized in lesion by-
pass (E. coli Pol IV and Pol V and human Pol ?) (Fig. 8). One
possible explanation may be that this lesion has conformational
flexibility, with the bulky dodecylpeptide pointed away from
the active site of Pol I, and has no effect on the Watson-Crick
geometry. The DNA-peptide structure might resemble that of
Lys-Trp-Lys-Lys linked at N2-dG via the N-terminal Lys
through an acrolein moiety. In the latter structure, the Trp
indolyl group does not intercalate into the DNA, while the
C-terminal Lys is exposed to solvent, interacting minimally
with the DNA (12). However, a detailed analysis of the mech-
anism of action of Pol I in the bypass of the N6-dA dodecyl-
peptide cross-link awaits the co-crystal structure determination
of Pol I in complex with this lesion.
With regard to the biological significance of N6-dA peptide
cross-links, a previous report focused exclusively on the iden-
tification of E. coli DNA polymerases that can catalyze repli-
cation bypass of N2-dG peptide cross-links (21). The poly-
merases responsible for the replication bypass of N6-dA
peptide cross-links, however, were not studied. The N6-dA
peptide lesion used in this study was derived from the ring-
opened aldehydic form of the acrolein-induced ?-HO-PdA
adduct. Nuclear magnetic resonance spectroscopy analyses
showed that ?-HO-PdA in the ring-opened form can be de-
tected readily (23). This suggests that ?-HO-PdA could poten-
tially interact with peptides, resulting in the formation of
DNA-peptide cross-links. Such a mechanism would be concep-
tually similar to that for the aldehydic group of ring-opened
?-HO-PdG, which reacts with peptides and yields chemically
identical lesions, except that the lesions are located in the
minor groove of DNA (16). Consistent with this prediction, it
was demonstrated that the peptide Lys-Trp-Lys-Lys is trapped
efficiently and stably with oligodeoxynucleotides containing
?-HO-PdA in the presence of a reducing agent (19). Thus,
although N6-dA peptide cross-links have not been identified in
vivo to date, their biological relevance cannot be disregarded,
and therefore the present investigation examining the identity
of polymerases involved in the processing of these lesions has
It is worth noting that in E. coli, TLS past N2-dG peptide
cross-links requires the DNA damage-inducible polymerase
Pol IV (21). In contrast, such specialized polymerases were not
essential for replication bypass of the N6-dA dodecylpeptide
cross-link. This is interesting given the fact that the previously
investigated N2-dG dodecylpeptide cross-link is chemically
identical to the N6-dA dodecylpeptide cross-link used here,
except that the former lesion is positioned in the minor, not the
major, groove of DNA. Germane to this observation, DNA
polymerases of the A family interact with DNA at the minor
groove (27). Thus, Pol I may have evolved to be particularly
proficient in the bypass of major groove lesions, including
relatively small adducts (5, 17) and large degradation products
of DNA-protein cross-links.
In summary, we have shown the ability of E. coli Pol ? to
accurately replicate past an N6-dA dodecylpeptide cross-link in
vitro and have generated data suggesting its role in bypass of
this lesion in vivo. Since no data about the identity of E. coli
DNA polymerases responsible for processing the major groove
cross-links are available to date, the findings reported here
promote our understanding of how cells maintain genome in-
tegrity upon induction of DNA-peptide and DNA-protein
We thank Carmelo J. Rizzo (Vanderbilt University, Nashville, TN)
for the gift of ?-HO-PdA-containing oligodeoxynucleotides, Masaaki
Moriya (State University of New York, Stony Brook, NY) for the gift
of the pMS2 vector, Michael O’Donnell (Rockefeller University, New
York, NY) for the gift of Pol III*, and Roger Woodgate (National
Institutes of Health, Bethesda, MD) for the gifts of E. coli strains
RW626 and RW82.
This work was supported by Public Health Service grants ES05355
(R.S.L.) and ES012259 (M.F.G.) from the National Institute of Envi-
ronmental Health Sciences, grant CA106858 (R.S.L.) from the Na-
tional Cancer Institute, grant GM21422 (M.F.G.) from the National
FIG. 8. Replication blockage of human Pol ? by N6-dA dodecyl-
peptide cross-link. Primer extension reactions were catalyzed by 1 nM
Pol ? under running-start conditions (?3 primer), using nondamaged
(ND) and N6-dA dodecylpeptide cross-linked (XL12) substrates. Re-
actions were carried out in the presence of a 100 ?M concentration of
each dNTP and 5 nM primer-template for 30 min at 37°C. A*, ad-
3820 YAMANAKA ET AL.J. BACTERIOL.
Institute of General Medicine, and an NSF Career Award
1. Barker, S., M. Weinfeld, and D. Murray. 2005. DNA-protein crosslinks: their
induction, repair, and biological consequences. Mutat. Res. 589:111–135.
2. Bonner, C. A., et al. 1992. Processive DNA synthesis by DNA polymerase II
mediated by DNA polymerase III accessory proteins. J. Biol. Chem. 267:
3. Bruck, I., R. Woodgate, K. McEntee, and M. F. Goodman. 1996. Purification
of a soluble UmuD?C complex from Escherichia coli. Cooperative binding of
UmuD?C to single-stranded DNA. J. Biol. Chem. 271:10767–10774.
4. Cai, H., H. Yu, K. McEntee, and M. F. Goodman. 1995. Purification and
properties of DNA polymerase II from Escherichia coli. Methods Enzymol.
5. Chary, P., and R. S. Lloyd. 1995. In vitro replication by prokaryotic and
eukaryotic polymerases on DNA templates containing site-specific and ste-
reospecific benzo[a]pyrene-7,8-dihydrodiol-9,10-epoxide adducts. Nucleic
Acids Res. 23:1398–1405.
6. Cox, M. M., K. McEntee, and I. R. Lehman. 1981. A simple and rapid
procedure for the large scale purification of the recA protein of Escherichia
coli. J. Biol. Chem. 256:4676–4678.
7. Curti, E., J. P. McDonald, S. Mead, and R. Woodgate. 2009. DNA polymer-
ase switching: effects on spontaneous mutagenesis in Escherichia coli. Mi-
crobiol. Mol. Biol. Rev. 71:315–331.
8. del Solar, G., R. Giraldo, M. J. Ruiz-Echevarria, M. Espinosa, and R.
Diaz-Orejas. 1998. Replication and control of circular bacterial plasmids.
Microbiol. Mol. Biol. Rev. 62:434–464.
9. Fernandes, A., et al. 1998. Mutagenic potential of stereoisomeric bay region
(?)- and (?)-cis-anti-benzo[a]pyrene diol epoxide-N2-2?-deoxyguanosine
adducts in Escherichia coli and simian kidney cells. Biochemistry 37:10164–
10. Goodman, M. F. 2002. Error-prone repair DNA polymerases in prokaryotes
and eukaryotes. Annu. Rev. Biochem. 71:17–50.
11. Hastings, P. J., et al. 2010. Competition of Escherichia coli DNA poly-
merases I, II and III with DNA Pol IV in stressed cells. PLoS One 5:e10862.
12. Huang, H., et al. 2010. Minor groove orientation of the KWKK peptide
tethered via the N-terminal amine to the acrolein-derived 1, N2-?-hydroxy-
propanodeoxyguanosine lesion with a trimethylene linkage. Biochemistry
13. Hwang, H., and J. S. Taylor. 2005. Evidence for Watson-Crick and not
Hoogsteen or wobble base pairing in the selection of nucleotides for inser-
tion opposite pyrimidines and a thymine dimer by yeast DNA pol ?. Bio-
14. Kobayashi, S., M. R. Valentine, P. Pham, M. O’Donnell, and M. F. Good-
man. 2002. Fidelity of Escherichia coli DNA polymerase IV. Preferential
generation of small deletion mutations by dNTP-stabilized misalignment.
J. Biol. Chem. 277:34198–34207.
15. Kogoma, T., and R. R. Maldonado. 1997. DNA polymerase I in constitutive
stable DNA replication in Escherichia coli. J. Bacteriol. 179:2109–2115.
16. Kurtz, A. J., and R. S. Lloyd. 2003. 1,N2-Deoxyguanosine adducts of acro-
lein, crotonaldehyde, and trans-4-hydroxynonenal cross-link to peptides via
Schiff base linkage. J. Biol. Chem. 278:5970–5976.
17. Latham, G. J., C. M. Harris, T. M. Harris, and R. S. Lloyd. 1995. The
efficiency of translesion synthesis past single styrene oxide DNA adducts in
vitro is polymerase-specific. Chem. Res. Toxicol. 8:422–430.
18. Minko, I. G., et al. 2009. Chemistry and biology of DNA containing 1,N2-
deoxyguanosine adducts of the alpha, beta-unsaturated aldehydes acrolein,
crotonaldehyde, and 4-hydroxynonenal. Chem. Res. Toxicol. 22:759–778.
19. Minko, I. G., et al. 2008. Mutagenic potential of DNA-peptide crosslinks
mediated by acrolein-derived DNA adducts. Mutat. Res. 637:161–172.
20. Minko, I. G., et al. 2005. Initiation of repair of DNA-polypeptide cross-links
by the UvrABC nuclease. Biochemistry 44:3000–3009.
21. Minko, I. G., et al. 2008. Replication bypass of the acrolein-mediated
deoxyguanine DNA-peptide cross-links by DNA polymerases of the DinB
family. Chem. Res. Toxicol. 21:1983–1990.
22. Minko, I. G., Y. Zou, and R. S. Lloyd. 2002. Incision of DNA-protein
crosslinks by UvrABC nuclease suggests a potential repair pathway involving
nucleotide excision repair. Proc. Natl. Acad. Sci. U. S. A. 99:1905–1909.
23. Nechev, L. V., C. M. Harris, and T. M. Harris. 2000. Synthesis of nucleosides
and oligonucleotides containing adducts of acrolein and vinyl chloride.
Chem. Res. Toxicol. 13:421–429.
24. Onrust, R., et al. 1995. Assembly of a chromosomal replication machine: two
DNA polymerases, a clamp loader, and sliding clamps in one holoenzyme
particle. I. Organization of the clamp loader. J. Biol. Chem. 270:13348–
25. Patel, P. H., M. Suzuki, E. Adman, A. Shinkai, and L. A. Loeb. 2001.
Prokaryotic DNA polymerase I: evolution, structure, and “base flipping”
mechanism for nucleotide selection. J. Mol. Biol. 308:823–837.
26. Pritchard, A. E., H. G. Dallmann, B. P. Glover, and C. S. McHenry. 2000. A
novel assembly mechanism for the DNA polymerase III holoenzyme DnaX
complex: association of ??? with DnaX4forms DnaX3???. EMBO J. 19:
27. Rothwell, P. J., and G. Waksman. 2005. Structure and mechanism of DNA
polymerases. Adv. Protein Chem. 71:401–440.
28. Stukenberg, P. T., P. S. Studwell-Vaughan, and M. O’Donnell. 1991. Mech-
anism of the sliding ?-clamp of DNA polymerase III holoenzyme. J. Biol.
29. Tang, M., et al. 2000. Roles of E. coli DNA polymerases IV and V in
lesion-targeted and untargeted SOS mutagenesis. Nature 404:1014–1018.
30. Yadav, P. N., J. S. Yadav, and M. J. Modak. 1992. A molecular model of the
complete three-dimensional structure of the Klenow fragment of Escherichia
coli DNA polymerase I: binding of the dNTP substrate and template-primer.
31. Yamanaka, K., et al. 2010. Novel enzymatic function of DNA polymerase ?
in translesion DNA synthesis past major groove DNA-peptide and DNA-
DNA cross-links. Chem. Res. Toxicol. 23:689–695.
32. Yeiser, B., E. D. Pepper, M. F. Goodman, and S. E. Finkel. 2002. SOS-
induced DNA polymerases enhance long-term survival and evolutionary
fitness. Proc. Natl. Acad. Sci. U. S. A. 99:8737–8741.
33. Zambrano, M. M., D. A. Siegele, M. Almiron, A. Tormo, and R. Kolter. 1993.
Microbial competition: Escherichia coli mutants that take over stationary
phase cultures. Science 259:1757–1760.
VOL. 193, 2011 BYPASS OF DNA-PEPTIDE CROSS-LINK BY DNA POLYMERASE I3821