Tolerance of base pair size and shape in postlesion DNA synthesis

Article abstract of DOI:10.1021/ja311434s

The influence of base pair size and shape on the fidelity of DNA polymerase-mediated extension past lesion-containing mispairs was examined. Primer extension analysis was performed with synthetic nucleosides paired opposite the pro-mutagenic DNA lesion O⁶-benzylguanine (O ⁶-BnG). These data indicate that the error-prone DNA polymerase IV (Dpo4) inefficiently extended past the larger Peri:O⁶-BnG base pair, and in contrast, error-free extension was observed for the smaller BIM:O ⁶-BnG base pair. Steady-state kinetic analysis revealed that Dpo4 catalytic efficiency was strongly influenced by the primer:template base pair. Compared to the C:G pair, a 1.9- and 79 000-fold reduction in Dpo4 efficiency was observed for terminal C:O⁶-BnG and BIM:G base pairs respectively. These results demonstrate the impact of geometrical size and shape on polymerase-mediated mispair extension.

Full text of DOI:10.1021/ja311434s

Communication
pubs.acs.org/JACS
Tolerance of Base Pair Size and Shape in Postlesion DNA Synthesis
Hailey L. Gahlon,† W. Bernd Schweizer,‡ and Shana J. Sturla*,†
†Department of Health Sciences and Technology, Institute of Food Nutrition and Health, ETH Zurich, Schmelzbergstrasse 9, 8092
̈
Zurich, Switzerland
̈
‡Laboratory for Organic Chemistry, D-CHAB, ETH Zurich, Wolfgang-Pauli-Strasse 10, 8093 Zurich, Switzerland
̈
̈
S
* Supporting Information
ABSTRACT: The influence of base pair size and shape
on the fidelity of DNA polymerase-mediated extension
past lesion-containing mispairs was examined. Primer
extension analysis was performed with synthetic nucleo-
sides paired opposite the pro-mutagenic DNA lesion O6-
benzylguanine (O6-BnG). These data indicate that the
error-prone DNA polymerase IV (Dpo4) inefficiently
extended past the larger Peri:O6-BnG base pair, and in
contrast, error-free extension was observed for the smaller
BIM:O6-BnG base pair. Steady-state kinetic analysis
revealed that Dpo4 catalytic efficiency was strongly
influenced by the primer:template base pair. Compared
to the C:G pair, a 1.9- and 79 000-fold reduction in Dpo4
efficiency was observed for terminal C:O6-BnG and BIM:G
base pairs respectively. These results demonstrate the
impact of geometrical size and shape on polymerase-
mediated mispair extension.
Figure 1. Structures and DNA sequences used in this study. (a)
Wobble base pair between O6-BnG and C in the active site of Dpo4
(b) synthetic base-modified BIM (1) and Peri (2) and (c) biochemical
scheme of Dpo4-mediated primer extension assay.
NA polymerase enzymes are responsible for gene
D
replication; dysfunctions during DNA synthesis can
result in mutations that may lead to cancer. Bulky DNA
lesions, resulting from exposure to genotoxic agents (e.g.,
polycyclic aromatic chemicals, nitrosamines, epoxides) can
impede replication and transcription processes.1,2 Specialized
DNA polymerases (i.e., Y-family, e.g. DNA polymerase IV,
Dpo4) are devoted to processing these bulky lesions that may
otherwise impede DNA replication.3 Important aspects of
error-prone DNA replication include binding of a Y-family
polymerase to primer:template DNA, dNTP selection,
phosphodiester bond formation, pyrophosphate release, trans-
location to the subsequent templating base, and potentially
repeating the selection and bond-forming process.4 Y-family
polymerases are poorly processive compared to high-fidelity
polymerases (e.g., Thermus aquaticus DNA polymerase
processivity is 50−80 nucleotides per binding event at 55
°C).5 Dpo4 will dissociate from primer:template DNA after
∼16 dNTP incorporation events (at 37 °C),6 and these
extension steps (i.e., immediately following the lesion) are
potentially mutagenic.
analysis revealed that the G:T pair contained a reverse wobble
that misaligns the 3′-hydroxyl necessary for primer extension.
Much of our current understanding regarding chemical
details of polymerase-mediated DNA replication has been
gained by studies utilizing non-natural nucleosides as chemical
probes.8−10 A study investigating minor-groove interactions
with a deoxyadenosine isostere indicated that mispair extension
is dependent on formation of a hydrogen bond between the
enzyme (Klenow DNA polymerase) and N3 of the base on the
primer oligonucleotide.11 These studies suggested that base pair
geometry may be more important for primer extension vs
insertion. The above studies do not address mispair extension
with physiological DNA adducts, i.e., PLS. The goal of the
present study was to investigate chemical features, i.e., base pair
size and shape, which can impact PLS. The strategy involves
probing PLS with synthetic nucleosides paired opposite the
DNA adduct O6-BnG. To our knowledge this is the first direct
probe to test the influence of base pair size and shape of PLS on
O6-alkylguanine containing templates.
Bulky O6-alkylguanine DNA adducts, like O6-BnG, are
informative models for investigating DNA polymerase mech-
anisms and nitrosamine carcinogenesis.12,13 N-Methylbenzyl-
nitrosamine can form O6-alkylguanine adducts that, if not
The process of postlesion DNA synthesis (PLS) has received
little attention despite relevance to DNA damage tolerance.
PLS is the extension process immediately following dNTP
insertion opposite a DNA adduct (see Figure 1c) and may
account for relevant mutations. Studies investigating the ability
of Dpo4 to extend past canonical mispairs observed that a G:T
mismatch was inefficiently extended by Dpo4.7 Crystallographic
Received: November 21, 2012
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja311434s | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
repaired, can generate G to A transition mutations during DNA
replication.14 O6-BnG is bypassed by Dpo4 (dCTP insertion is
the major product, approximately 70%).15 And crystal
structures containing a C:O6-BnG (primer:template) terminal
base pair within the active site of Dpo4 reveal a wobble base
pair interaction (Figure 1a).15 Thus, insertion opposite O6-BnG
has been investigated, but PLS remains to be explored.
To our knowledge, there is no information regarding
chemical requirements for O6-alkylguanine PLS. A recent
report concerning Dpo4 conformation dynamics during
synthesis past 8-oxoG found translocation was inhibited during
PLS, suggesting that the inhibition in DNA translocation may
facilitate polymerase switching between low- and high-fidelity
polymerases.16 A similar result has been observed in a study
regarding pol eta bypass of a TT dimer. Here pol eta observed a
higher error rate and lower fidelity while bypassing the 3′
thymine of the dimer in comparison to the 5′ dimer, suggesting
that during lesion translocation pol eta can sense differences in
dimer location that may be responsible for inducing polymerase
switching during lesion bypass.17
Previously, we reported an adduct-specific synthetic nucleo-
side that forms stable DNA duplexes when paired opposite O6-
BnG vs the natural DNA bases.18 However, no information is
available concerning how this or analogous probes behave
during polymerase-catalyzed mispair extension. The exper-
imental approach of this study involves evaluating two site-
specifically modified oligonucleotides that position the adduct-
directed nucleoside probes at the 3′ terminal position of the
primer (Figure 1). BIM (1) and Peri (2) contain nucleobase
modifications that promote adduct:probe pairing in a DNA
duplex. Each structure contains a similar imidazole moiety
along the Watson−Crick hydrogen-bonding face, but BIM and
Peri differ in size as the result of having one (BIM) or two
(Peri) aromatic rings (Figure 1b). Phosphoramidites of BIM
and Peri were synthesized for chemical incorporation into
oligonucleotides. Crystal structures of BIM and Peri 2′-
deoxynucleosides were obtained to confirm relative stereo-
chemistry (Figures S3 and S4). Mutagenesis studies were
completed to elucidate the efficiency and fidelity on PLS using
the base-modified primers containing BIM and Peri (Figure
1c). Single nucleotide insertion and full length extension
studies were performed to assess the ability of Dpo4 to perform
PLS on damaged (O6-BnG) and primer extension on
nondamaged (G) templates (Figure 2). Two conditions were
investigated for primer extension assays with Dpo4, DNA, and
dNTPs: (1) 20 nM, 10 nM, and 100 μM, respectively (Figure
2) and (2) 5 nM, 10 nM, and 50 μM, respectively (Figure S9).
Dpo4-mediated primer extension studies involving the
natural 24mer primer revealed reduced fidelity for the G-
containing template vs the O6-BnG-containing template
(Figure 2a). Extension of a primer involving a canonical C:G
pair at the primer:template terminus (denoted throughout this
text as the first base, here C, originating from the primer strand
and the second base, here G, from the template strand)
revealed low fidelity, and in some cases (dGTP and dATP
insertion) multiple incorporations were observed. In the case of
dATP insertion, a prominent n + 2 extension product band was
present due to the template sequence; where T is the n + 2
templating base. Additionally, prominent dGTP n + 2 bands
were observed for the C:G and C:O6-BnG pairs (albeit lower
for C:O6-BnG, Figure 2). This result was not unexpected since
enzyme concentrations were in excess to DNA. However, to
test this further, lower enzyme and dNTP concentrations were
Figure 2. Analysis of Dpo4-mediated primer extension. (B, blank; 4, all
four dNTPs; G, dGTP; A, dATP; T, dTTP; C, dCTP). (a) Extension
of a natural 24mer primer (see Figure 1c for DNA sequences)
opposite G or O6-BnG. (b) Extension of 24mer 3′-modified BIM
primer opposite G or O6-BnG. (c) Extension of 24mer 3′-modified
Peri primer opposite G or O6-BnG.
investigated. This resulted in decreased n + 2 band intensity for
both C:G and C:O6-BnG pairs (Figure S9). Full-length primer
extension was highly efficient for C:G and C:O6-BnG pairs
(100% extension products). PLS on the C:O6-BnG pair
revealed high selectivity for products of correct extension
(dGTP insertion), with trace amounts of dATP and dTTP
extension. With lower enzyme (5 nM) and dNTP (50 μM)
concentrations, relative to DNA (10 nM), the correct insertion
of dGTP was observed for both G and O6-BnG templates, and
these results revealed a less mutagenic profile (Figure S9) in
comparison to higher enzyme and dNTP concentrations. It
seems that Dpo4, as a translesion DNA polymerase, is well
suited for lesion processing even when the lesion is in the n − 1
position (i.e., PLS).
Extension with BIM:G and BIM:O6-BnG pairs was selective
(i.e., only dGTP insertion). A higher percentage of full-length
extension was observed for the nondamaged (56%) vs damaged
O6-BnG template (37%), suggesting that the size of a terminal
BIM:G base pair is more suited for efficient catalysis to occur in
comparison to the BIM:O6-BnG pair. Lower mutagenicity was
observed with the BIM:G pair in comparison to both the C:G
and Peri:G pairs. Dpo4-mediated extension with the Peri:G
terminal pair contains low levels of dGTP or dTTP extension
products (approximately 10% each) and high levels of dCTP
extension products (50%). The high amount of dCTP
incorporation seems interesting since this would result in a
C:C mismatch, i.e. due to the presence of a C in the template
strand (Figure 1c). A possible explanation for the apparent C:C
mismatch product may invoke a slippage mechanism. Template
slippage has been observed during Dpo4-meditated DNA
replication on repetitive DNA sequences, and results in single-
base deletions.19 Full-length extension as well as a small
percentage of dGTP insertion (7%) was observed with the
B
dx.doi.org/10.1021/ja311434s | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Peri:O6-BnG pair. However, these levels were low compared
with those resulting from Dpo4-mediated extension on C:O6-
BnG or BIM:O6-BnG terminal pairs. It seems the large size and
shape distortion from the Peri:O6-BnG termini impedes PLS
more than C:O6-BnG or BIM:O6-BnG termini.
To quantitatively address the differences in the ability of
Dpo4 to extend from canonical vs mismatched (probe:
template) termini, steady-state kinetics were performed. Kinetic
parameters, Km and kcat, were determined under enzyme-
limiting conditions and time-course measurements of n + 1
product formation (dGTP incorporation) were performed
(Supplementary Methods). The Km for C:G or C:O6-BnG was
9.8 and 30 μM, respectively. Catalytic efficiency was decreased
1.9-fold when Dpo4 performed extension on C:O6-BnG
termini, in comparison to the canonical C:G terminal pair
(Table 1). Steady-state parameters for the 3′-base-modified Peri
Figure 3. Modeling images of base-modified probes opposite O6-BnG
(a) BIM:O6-BnG and (b) Peri:O6-BnG.
a closer contact for H-bonding. The modeling results and the
crystal structure lattices of BIM and Peri nucleosides reveal a
striking resemblance in the hydrogen-bonding pattern (Figure
S2). The crystal lattice of BIM shows an intermolecular H-bond
between an imidazole nitrogen from one BIM molecule and a
5′-OH group from a neighboring BIM. In contrast, Peri utilizes
two water molecules to bridge H-bonding between neighboring
Peri molecules, possibly to form H-bonding without close
contacts with the naphthalene moiety. These structural models
suggest a potential difficulty in probing the influence of size and
shape on PLS without distorting H-bonding relationships
between the probe-lesion base pair. Nonetheless, these
structures are useful for visualizing the size constraints within
the postinsertion site.
The impact of increasing the size of alkyl groups placed in
either the major or minor groove of DNA upon polymerization
by Y-family polymerases has been studied previously in great
detail.20,21 However, there are no studies reported to date that
investigate how bulky constituents at the primer terminus
modulate extension past DNA lesions by specialized enzymes.
In this study, we utilized 3′ base-modified nucleoside analogs
for probing size and shape tolerances on Dpo4-mediated O6-
BnG-PLS.
Biochemical characterization of the four DNA polymerases
present in Sulfolobus solfataricus (Dpo1−4) indicated that Dpo4
was most proficient at lesion bypass and subsequent
extension.22 In mammalian systems, DNA damage bypass can
require two polymerases, i.e., one polymerase to insert and the
second to extend.23 In this context, Dpo4 has limitations in
probing eukaryotic PLS. It seems DNA polymerase ζ, an
“extender” polymerase in eukaryotic systems, would be a good
candidate for further PLS investigation.
In the case of Dpo4-mediated PLS, the template strand has a
significant role in PLS-extension fidelity. The data from the
system tested in this study suggest that O6-BnG, at the n − 1
position, imparts higher fidelity in comparison to the
nondamaged template (at excess enzyme and dNTP concen-
trations relative to DNA) but that the efficiency with which the
primer strand is extended seems to be strongly influenced by
the base pair at the primer:template termini during PLS. The
smaller BIM probe resulted in error-free and efficient n + 1
primer extension, when paired opposite G or O6-BnG, in
comparison to the larger Peri probe. However, relative to either
BIM or Peri, the natural primer with a terminal 3′ cytosine was
the most efficiently extended (dGTP incorporation) in the case
of nondamaged or damaged templates.
Table 1. Steady-State Kinetic Parameters for dGTP
Incorporation by Dpo4
Δ relative efficiency
a
(NX)
Km (μM)
kcat (min−1
)
to C:G
C:G
9.8 0.4
30 5.0
1.82 0.03
2.94 0.29
1
C:O6-BnG
1.9-fold less
79 000-fold less
BIM:G
3390 700
0.008 0.0054
a
Describes the ratio of (kcat/Km,dGTP
)
C:G/(kcat/Km)N:X; NX represents
termini base pair; BIM:O6-BnG, Peri:G, and Peri:O6-BnG not
determined due to insufficient product formation.
primer paired opposite nondamaged and damaged templates
were investigated, however no product bands were observed
under all conditions tested (time points up to 72 h). The Km
for BIM:G was 3390 μM. For BIM:O6-BnG the Km was not
accurately determined (no product observed at high dGTP
concentrations (20 mM) nor time points up to 23 h). Taken
together, these data suggest that alterations in base pair size and
shape are important factors that influence Dpo4 catalytic
efficiency during O6-BnG-PLS and invoke the following trend
in PLS rates from natural to more perturbed structures: C >
BIM > Peri.
Thermal melting analysis was performed to evaluate
primer:template duplex stability for placing BIM or Peri
opposite G or O6-BnG. Duplexes containing a C:G or C:O6-
BnG terminal pairs had a Tm of 70 and 68 °C, respectively.
There was no difference in Tm for damaged or nondamaged
templates when paired with BIM or Peri in comparison to
duplexes with the natural primer (Supplementary Table 2). It
was anticipated that modifications at the 3′ end of the primer,
whether the terminal primer:template base pairs are matched or
mismatched, would impart very little change in primer:template
DNA melting behavior, and the results support this assertion.
To provide further rationale of base pairing interactions
between BIM:O6-BnG and Peri:O6-BnG, computational studies
were performed on the basis of the published crystal structure
of a ternary complex of Dpo4 containing a C:O6-BnG terminal
pair and an incoming dGTP (pdb code 2jef).15 The 3′ cytosine
on the primer strand was replaced with BIM or Peri, energy
minimized, and visualized. A hydrogen bond was observed
between the exocyclic amino group of O6-BnG and an
imidazole nitrogen of BIM (2.11 Å, Figure 3a). No hydrogen
bonds were observed between O6-BnG and Peri in the
calculated structure (Figure 3b). Although suitably arranged
for H-bonding, the bulky naphthalene on Peri seems to hinder
In summary, these results implicate that (1) the size and
shape of the base pair at the penultimate (n − 1) position
during Dpo4-mediated PLS influences extension fidelity and
efficiency; (2) the template strand has a strong impact on PLS
fidelity; and (3) the structure of the base at the terminal
position on the primer strand largely impacts extension
C
dx.doi.org/10.1021/ja311434s | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
efficiency during PLS. Further studies are warranted using new
nucleoside probes that would promote complementary
adduct:probe base pairing in order to evaluate additional
chemical factors (such as H-bonding) that could affect PLS.
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental procedures and characterization data. Full
crystallographic data has been deposited with the Cambridge
Crystallographic Data Centre with deposition numbers CCDC
910347 (1) and CCDC 910348 (2). This material is available
free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
sturlas@ethz.ch
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank funding from the European Research
Council 260341. We gratefully acknowledge the University of
Minnesota Supercomputing Institute as well as Prof. Elizabeth
Amin and Dr. Ting-Lan Chiu for helpful feedback and insight
regarding the computational modeling experiments.
REFERENCES
■
(1) Zhang, L.; Rechkoblit, O.; Wang, L.; Patel, D. J.; Shapiro, R.;
Broyde, S. Nucleic Acids. Res. 2006, 34, 3326.
(2) Gowda, P. A. S.; Krishnegowda, G.; Suo, Z.; Amin, S.; Spratt, T.
E. Chem. Res. Toxicol. 2012, 25, 1195.
(3) Sale, J. E.; Lehmann, A. R.; Woodgate, R. Nat. Rev. Mol. Cell Biol.
2012, 13, 141.
(4) Berdis, A. J. Chem. Rev. 2009, 109, 2862.
(5) Merkens, L. S.; Bryan, S. K.; Moses, R. E. Biochim. Biophys. Acta
1995, 1264, 243.
(6) Fiala, K. A.; Suo, Z. Biochemistry 2004, 43, 2116.
(7) Trincao, J.; Johnson, R. E.; Wolfle, W. T.; Escalante, C. R.;
Prakash, S.; Prakash, L.; Aggarwal, A. K. Nat. Struct. Mol. Biol. 2004,
11, 457.
(8) Jung, K. H.; Marx, A. Cell. Mol. Life Sci. 2005, 62, 2080.
(9) Lee, I.; Berdis, A. J. Biochim. Biophys. Acta 2010, 1804, 1064.
(10) Dahlmann, H. A.; Vaidyanathan, V. G.; Sturla, S. J. Biochemistry
2009, 48, 9347.
(11) Morales, J. C.; Kool, E. T. J. Am. Chem. Soc. 1999, 121, 2323.
(12) Yoo, J. S. H.; Guengerich, F. P.; Yang, C. S. Cancer Res. 1988,
48, 1499.
(13) Peterson, L. A. Chem. Res. Toxicol. 1997, 10, 19.
(14) Pauly, G. T.; Moschel, R. C. Chem. Res. Toxicol. 2001, 14, 894.
(15) Eoff, R. L.; Angel, K. C.; Egli, M.; Guengerich, F. P. J. Biol.
Chem. 2007, 282, 13573.
(16) Maxwell, B. A.; Xu, C.; Suo, Z. J. Biol. Chem. 2012, 287, 13040.
(17) McCulloch, S. D.; Kokoska, R. J.; Masutani, C.; Iwai, S.;
Hanaoka, F.; Kunkel, T. A. Nature 2004, 428, 97.
(18) Gong, J.; Sturla, S. J. J. Am. Chem. Soc. 2007, 129, 4882.
(19) Wu, Y.; Wilson, R. C.; Pata, J. D. J. Bacteriol. 2011, 193, 2630.
(20) Choi, J. Y.; Guengerich, F. P. J. Biol. Chem. 2008, 283, 23654.
(21) Maddukuri, L.; Eoff, R. L.; Choi, J. Y.; Rizzo, C. J.; Guengerich,
F. P.; Marnett, L. J. Biochemistry 2010, 49, 8415.
(22) Choi, J. Y.; Eoff, R. L.; Pence, M. G.; Wang, J.; Martin, M. V.;
Kim, E. J.; Folkmann, L. M.; Guengerich, F. P. J. Biol. Chem. 2011, 286,
31180.
(23) Livneh, Z.; Ziv, O.; Shachar, S. Cell Cycle 2010, 9, 729.
D
dx.doi.org/10.1021/ja311434s | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX