Base pairing and replicative processing of the formamidopyrimidine-dG DNA lesion

Article abstract of DOI:10.1021/ja0549188

The 2,6-diamino-4-hydroxy-5-formamidopyrimidine of 2′-deoxyguanosine (FaPydG) is one of the major DNA lesions found after oxidative stress in cells. To clarify the base pairing and coding potential of this major DNA lesion with the aim to estimate its mut

Full text of DOI:10.1021/ja0549188

Published on Web 12/01/2005
Base Pairing and Replicative Processing of the
Formamidopyrimidine-dG DNA Lesion
Matthias Ober, Heiko Mu¨ller, Carsten Pieck, Johannes Gierlich, and Thomas Carell*
Contribution from the Department of Chemistry and Biochemistry
Ludwig-Maximilians-UniVersity Munich, D-81377 Munich, Germany
Received July 22, 2005; E-mail: Thomas.Carell@cup.uni-muenchen.de
Abstract: The 2,6-diamino-4-hydroxy-5-formamidopyrimidine of 2′-deoxyguanosine (FaPydG) is one of the
major DNA lesions found after oxidative stress in cells. To clarify the base pairing and coding potential of
this major DNA lesion with the aim to estimate its mutagenic effect, we prepared oligonucleotides containing
a cyclopentane based analogue of the DNA lesion (cFaPydG). In addition, oligonucleotides containing the
cyclopentane analogue of 2′-deoxyguanosine (cdG), and oligonucleotides containing 8-oxo-7,8-dihydro-
2′-deoxyguanosine (8-oxodG) were synthesized. The thermodynamic stability of duplexes containing these
building blocks and all canonical counterbases were determined by concentration dependent melting-point
measurements (van’t Hoff plots). The data reveal that cFaPydG greatly destabilizes a DNA duplex (∆∆G°298K
≈ 2-4 kcal mol-1). The optimal base pairing partner for the cFaPydG lesion is dC. Investigation of duplexes
containing dG and cdG shows that the effect of substituting the deoxyribose by a cyclopentane moiety is
marginal. The data also provide strong evidence that the FaPydG lesion is unable to form a base pair with
dA. Our computational studies indicate that the syn-conformation required for base pairing with dA is
energetically unfavorable. This is in contrast to 8-oxodG for which the syn-conformation represents the
energetic minimum. Kinetic primer extension studies using S. cerevisiae Pol η reveal that cFaPydG is
replicated in an error-free fashion. dC is inserted 2-3 orders of magnitude more efficiently than dT or dA,
showing that FaPydG is a lesion which retains the coding potential of dG. This is also in contrast to 8-oxodG,
for which base pairing with dC and dA was established.
Introduction
For both guanine derived lesions, it is known that the
Watson-Crick hydrogen bonding mode with cytosine is largely
DNA is constantly damaged by reactive oxygen species
(ROS), which are either formed directly in the cell during
aerobic respiration or generated after the exposure of cells to
certain chemical agents. The resulting base lesions interfere with
the replication fidelity and therefore endanger the integrity of
the genetic information. Oxidatively generated lesions are one
main reason for mutagenesis and are in addition believed to be
involved in the aging process.1-3
One of the most prevalent DNA lesions formed after the
attack of DNA by ROS is the 2′-deoxyguanosine-based ring-
opened 2,6-diamino-4-hydroxy-5-formamidopyrimidine DNA
lesion (FaPydG).4 This hydrolysis product of the nucleobase
dG is generated in addition to the well studied 8-oxo-7,8-
dihydro-2′-deoxyguanosine lesion (8-oxodG) from a common
radical precursor5,6 (Scheme 1a). Both lesions were isolated from
DNA of cells exposed to oxidative stress or γ-irradiation, which
underpins their biological relevance.7,8
retained. Both lesions, however, possess in principle the ability
to participate in alternative base pairing modes. For 8-oxodG,
it is well documented that this lesion forms a stable base pair
with dC in the anti-conformation. 8-oxodG, however, can rotate
around the glycosidic bond to adopt a syn-conformation.9 This
syn-conformation allows the lesion to form a Hoogsteen base
pair with dA as depicted in Scheme 1b. A similar dual base
pairing behavior was recently postulated for the FaPydG lesion.
Investigations by our group with a FaPydG analogue (cFaPydG,
see below) and by Ide et al.10 with a methylated derivative of
the FaPydG lesion, however, question the ability of the FaPydG
lesion to base pair with dA and claim instead a base pairing
potential with dT.
The problem of any study with the FaPydG lesion is the
â-FaPydG lesion readily anomerizes to give the R-FaPydG
lesion under conditions generally required for DNA synthesis.
Whether this anomerization is of biological relevance is unclear.
Due to rapid anomerization of the lesion under these chemical
synthesis conditions, it has only been possible to incorporate
(1) Lindahl, T. Nature 1993, 362, 709-715.
(2) Friedberg, E. C.; Walker, G. C.; Siede, W. DNA repair and mutagenesis;
ASM Press: Washington, DC, 1995.
(3) Scha¨rer, O. D. Angew. Chem., Int. Ed. 2003, 42, 2946-2974.
(4) Douki, T.; Martini, R.; Ravanat, J. L.; Turesky, R. J.; Cadet, J. Carcino-
genesis 1997, 18, 2385-2391.
(8) Collins, A. R.; Cadet, J.; Mo¨ller, L.; Poulsen, H. E.; Vin˜a, J. Arch. Biochem.
Biophys. 2004, 423, 57-65.
(5) Burrows, C. J.; Muller, J. G. Chem. ReV. 1998, 98, 1109-1151.
(6) Steenken, S. Chem. ReV. 1989, 89, 503-520.
(9) Plum, G. E.; Grollman, A. P.; Johnson, F.; Breslauer, K. J. Biochemistry
1995, 34, 16148-16160.
(7) Pouget, J.-P.; Frelon, S.; Ravanat, J.-L.; Testard, I.; Odin, F.; Cadet, J.
Radiat. Res. 2002, 157, 589-595.
(10) Asagoshi, K.; Terato, H.; Ohyama, Y.; Ide, H. J. Biol. Chem. 2002, 277,
14589-14597.
9
10.1021/ja0549188 CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 18143-18149
18143

A R T I C L E S
Ober et al.
Scheme 1. (a) Formation of the Lesions 8-oxodG and FaPydG, (b) Pairing of anti-8-oxodG with dC (Watson-Crick) and syn-8-oxodG with
dA (Hoogsteen), (c) Proposed Pairing of FaPydG with dC (Watson-Crick) and dT (“Wobble” Base Pair)
an R/â-mixture of the FaPydG lesion into DNA.11,12 The exact
R to â ratio in these synthetic DNA strands is unknown, although
a recent detailed enzymatic investigation provided evidence that
the â-form may be the prevalent anomer in duplex DNA after
annealing.13 Hydrolysis of the glycosidic bond at elevated
temperatures is another problem associated with the FaPydG
lesion.
Both other research groups14 und ourselves15 have recently
reported about the development of nonhydrolyzable and epimer-
izable â-analogues of the FaPydG lesion. We reported synthesis
of a bioisosteric cyclopentane analogue of the FaPydG lesion.
This analogue (cFaPydG) features a cyclopentane unit instead
of the naturally occurring 2-deoxyribose moiety (O to CH2
modification). This analogue is kinetically stable, which allowed
us to obtain cocrystal structures of DNA containing this lesion
in complex with the formamidopyrimidine DNA glycosylase
repair enzyme16 (Fpg) and with a high fidelity DNA poly-
merase.17
the O to CH2 “chemical mutation”. The conclusions of this study
were complemented by computational investigations.
Results and Discussion
Synthesis: For the synthesis of the FaPydG analogue
cFaPydG (Scheme 2), we started with the enantiomerically pure
cyclopentylamine 1, which was prepared in six steps from
(1R,4S)-2-azabicyclo[2.2.1]hept-5-ene-3-one using the synthetic
route recently communicated by Cullis and co-workers.18 The
pyrimidinone 2 was prepared from 4,6-dichloro-2-pyrimidiny-
lamine as published.19,20 The critical coupling step between these
two compounds was accomplished using dimethylformamide
as solvent at 70 °C under strictly anhydrous conditions in order
to avoid hydrolysis of the base labile acetyl protecting group.
The configuration of compound 3 was proven by X-ray
crystallography. Crystallization was achieved by vapor diffusion
of water into a saturated solution of 3 in water/DMF 1:1 over
several weeks (Figure 1).
In this study, we present a detailed analysis of the base pairing
and coding potential of â-FaPydG employing the cyclopentane
cFaPydG analogue using melting point and primer extension
experiments. A comparative study of dG and its cyclopentane
analogue cdG was undertaken in order to estimate the effect of
Protection of the hydroxyl groups with TBDMS-Cl yielded
compound 4. Reduction of the nitro group furnished the highly
air-sensitive amine 5, which was subsequently handled under
inert gas. Formylation was possible with EDC and formic acid
with fair yields. The subsequent deprotection of the TBDMS
groups proved to be difficult due to parallel cleavage of the
acetyl protecting group under basic fluoride deprotection
conditions such as TBAF in THF. The formamide group on
the other hand was labile under most acidic deprotection
conditions. The cleavage reaction was finally possible with
pyridine buffered pyridine-HF complex in ethyl acetate.
(11) Haraguchi, K.; Greenberg, M. M. J. Am. Chem. Soc. 2001, 123, 8636-
8637.
(12) Haraguchi, K.; Delaney, M. O.; Wiederholt, C. J.; Sambandam, A.; Hantosi,
Z.; Greenberg, M. M. J. Am. Chem. Soc. 2002, 124, 3263-3269.
(13) Patro, J. N.; Haraguchi, K.; Delaney, M. O.; Greenberg, M. M. Biochemistry
2004, 43, 13397-13403.
(14) Delaney, M. O.; Greenberg, M. M. Chem. Res. Toxicol. 2002, 15, 1460-
1465.
(15) Ober, M.; Linne, U.; Gierlich, J.; Carell, T. Angew. Chem., Int. Ed. 2003,
42, 4947-4951.
(16) Coste, F.; Ober, M.; Carell, T.; Boiteux, S.; Zelwer, C.; Castaing, B. J.
Biol. Chem. 2004, 279, 44074-44083.
(18) Dom´ınguez, B. M.; Cullis, P. M. Tetrahedron Lett. 1999, 40, 5783-5786.
(19) Pfleiderer, W.; Bu¨hler, E. Chem. Ber. 1966, 99, 3022-3039.
(20) Burgdorf, L. T.; Carell, T. Chem.sEur. J. 2002, 8, 293-301.
(17) Hsu, G. W.; Ober, M.; Carell, T.; Beese, L. S. Chem. Biol., submitted.
9
18144 J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005

Base Pairing and Replication of the FaPydG-DNA Lesion
A R T I C L E S
Scheme 2. Synthesis of the Carbocyclic Analogues of â-FaPydG (â-cFaPydG) and 2′-deoxyguanosine (cdG)a
a (j) DIPEA, DMF, 70 °C, 90 min, 86%, (k) TBDMSCl, imidazole, DMF, 25 °C, 3 h, 79%, (l) Pd/C, H2, EtOH, 25 °C, 3 h (m) HCOOH, DIPEA,
EDC, DMF, 25 °C, 48 h, 60% for two steps, (n) HF‚pyridine, pyridine, EtOAc, 25 °C, 20 h, 94%, (o) DMTCl, pyridine, DMAP, 0-20 °C, 3 h, 56%,
(p) P[N(iPr)2]2OCH2CH2CN, diisopropylaminotetrazolide, CH2Cl2, 25 °C, 20 h, 53%, (q) diethoxymethyl acetate, 25 °C, 20 h, then 100 °C, 75 min, 37%,
(r) HF‚pyridine, EtOAc, 25 °C, 20 h, 85%, (s) DMTCl, pyridine, 20 °C, 2.5 h, 85%, (t) ClPN(iPr)2OCH2CH2CN, DIPEA, THF, 25 °C, 2 h, 52%.
into the phosphoramidite 13 was possible using standard
procedures. Full experimental procedures for the synthesis of 9
and 13 are given in the Supporting Information. Incorporation
of the cFaPydG building block into oligonucleotides was
possible using standard conditions for coupling and deprotection.
For capping, however, we had to replace phenoxyacetyl
anhydride with the sterically more demanding pivaloylic
anhydride in order to avoid transamidation of the formami-
dopyrimidine.12 Cleavage of the fully assembled DNA from the
solid support and of all base labile protecting groups was
performed using a saturated solution of ammonia in water/
ethanol 1:3 at 15 °C. Oligonucleotides containing 8-oxodG were
synthesized with a commercially available 8-oxodG building
block and deprotected using a saturated solution of ammonia
in water/ethanol 1:3 and mercaptoethanol at 60 °C, as suggested
by the manufacturer. All oligonucleotides were purified by
reversed-phase HPLC. DMT-on oligonucleotides were depro-
Figure 1. X-ray crystal structure of compound 3.
tected after HPLC purification on a Sep-Pak cartridge (Waters)
The DMT protecting group was introduced using standard
as described in the Supporting Information. All oligonucleotides
conditions. The phosphoramidite was finally prepared with
prepared for this study are listed in Table 1.
3-(bis(diisopropylamino)phosphanooxy)propanenitrile and di-
isopropylaminotetrazolide.21
To prove the correct incorporation of the cFaPydG into DNA,
an HPLC-MS/MS experiment of the total enzymatic digest of
oligonucleotide d1 was performed. The HPL chromatogram
showed five peaks, of which four could be assigned to the
canonical nucleotides dA, dC, dG, and dT by their mass and
fragmentation pattern. The fifth peak had the same retention
time, mass, and fragmentation pattern as the fully deprotected
To determine whether the replacement of the deoxyribose
by a cyclopentane ring system significantly affects the biophysi-
cal characteristics of nucleobases in DNA, we prepared the
carbocyclic 2′-deoxyguanosine (cdG) derivative 13 to compare
with dG.
Compound 13 was synthesized from the air-sensitive diamino
precursor 5. Purine formation proceeded in moderate yields
using diethoxymethyl acetate.22 The silyl protecting groups were
removed with pyridine-HF in ethyl acetate. Conversion of 11
(21) Kurz, M.; Gobel, K.; Hartel, C.; Gobel, M. W. HelV. Chim. Acta 1998, 81,
1156-1180.
(22) Vince, R.; Daluge, S. J. Org. Chem. 1980, 45, 531-533.
9
J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005 18145

A R T I C L E S
Ober et al.
Table 1. Synthesized Modified Oligonucleotides
interaction, is observed with dC in a cFaPydG:dC base pair.
The absolute value of ∆G°298K for the melting process of
duplexes containing a cFaPydG:dC is, however, not much higher
than for a dG:dT mismatch. The second best pairing interaction
of cFaPydG is observed with dT as the counterbase. This base
pair is only 1.9 (d1) to 2.6 (d2) kcal/mol less stable than the
cFaPydG:dC base pair. cFaPydG opposite any purine base gives
extremely strong destabilized duplexes (about 3.0-3.5 kcal/
mol weaker paired than cFaPydG:dC). This destabilization is
even higher than that measured for a normal mismatch. If we
correct the data for the effect of the O f CH2 chemical
alteration, we obtain even stronger destabilizing effects. From
these data it is evident, that cFaPydG induces a large destabi-
lization of any DNA duplex regardless of the counterbase. The
best interaction is obtained with dC. Some constructive interac-
tions may also be possible with dT (Scheme 1c). The measured
destabilization of â-cFaPydG facing dA is, in our experiments,
so high that the formation of a cFaPydG:dA base pair has to be
completely ruled out.
This conclusion contrasts results reported by Haraguchi and
Greenberg12 who observed that FaPydG is able to base pair with
dA. The reason for the discrepancy may be that their seminal
results were obtained with the already mentioned lesion R/â-
mixture. Synthesis of a carbocyclic R-FaPydG derivative is
underway to investigate this issue. The result of our study,
however, is in agreement with earlier data obtained using a
OHC-NCH3-methylated analogue of FaPydG.14
a
ꢀ260, calcd
Mcalcd
Mfound
(amu)
1
name
sequence
(L
µ
mol-1 cm-
)
(amu)
d1
d2
d3
5′GCGATcFaPydGTAGCG
5′TGCAGTcFaPydGACAGC
5′TAGcFaPydGCCTGGTCATT
0.1152
0.1255
0.1355
3413.0 3413.4
3686.2 3686.3
4285.6 4285.5
d4
d5
d6
5′GCGAT8-oxodGTAGCG
5′TGCAGT8-oxodGACAGC
5′TAG8-oxodGCCTGGTCATT
0.1152
0.1255
0.1355
3413.1 3413.6
3686.3 3687.0
4285.7 4285.1
d7
d8
d9
5′GCGATcdGTAGCG
5′TGCAGTcdGACAGC
5′TAGcdGCCTGGTCATT
0.1214
0.1316
0.1416
3395.2 3395.8
3668.3 3669.2
4267.7 4268.2
a MALDI-TOF, hydroxypicolinic acid matrix, negative mode.
cFaPydG proving that cFaPydG was not modified during DNA
synthesis and purification (See Supporting Information for
details).
Thermodynamic Measurements: The thermodynamic pa-
rameters ∆H°, ∆S°, and ∆G°298K of the dissociation process of
oligonucleotide duplexes containing cFaPydG and 8-oxodG
were derived from concentration dependent melting-point
investigations using van’t Hoff plots.23 In total we determined
van’t Hoff plots for two different DNA double strands with dG,
cdG, 8-oxodG, or cFaPydG at a defined position. Each strand
was hybridized with four different counterstrands to form all
possible base pairs. UV melting points of the duplexes were
determined for concentrations ranging from 0.3 to 27 µM (Table
2).
Density Functional Calculations: The melting point data
are at a first glance difficult to explain in light of the rather
similar structures of the FaPydG and 8-oxodG lesions. Why
does 8-oxodG base pair with dA and the structurally similar
FaPydG not? To investigate this problem one has to ask whether
FaPydG is able to rotate around the glycosidic bond into a stable
syn-like conformation in order to form the required Hoogsteen
base pair with dA. To analyze this question we performed
theoretical studies. We calculated first the torsion potential of
FaPydG and cFaPydG around the C1′-N bonds in order to
clarify how much the data from the analogue deviate from the
natural lesion. The results of the calculations are presented in
Figure 2a. Each data point represents a full geometry optimiza-
tion using B3LYP/6-31G* density functional calculation with
a constraint on the investigated torsion angle.24
The potential functions of cFaPydG and FaPydG are as
expected very similar, which supports again the validity of our
model compound. Both compounds exhibit two local minima,
which correspond to the syn- and the anti-conformation.25 The
anti-conformation, which allows Watson-Crick base pairing
with dC, however, is in both cases about 6 kcal/mol more stable
than the syn-conformation needed to establish the Hoogsteen
binding mode with dA. Because of the calculated low barrier
We first observed the effect of replacing the 2-deoxyribose
moiety with cyclopentane on the obtained ∆G°298K, ∆H°, and
∆S° values. To this end oligonucleotides containing carbocyclic
2′-deoxyguanosine cdG were studied in comparison to strands
containing dG.
As expected, the absolute value of ∆G°298K at 25 °C for the
melting process of duplexes containing dG at the site of interest
paired with dC was highest. Duplexes containing a mismatch
exhibited a strongly reduced ∆G°298K (∆∆G°298K ≈ 4 kcal/mol).
The carbocyclic cdG nucleotide exhibits a very similar behavior.
The observed melting point differences between duplexes
containing dG and cdG are very small (between 0.2 and 2.0
°C). The cyclopentane replacement generates a slightly more
stable duplex due to a small increase of the absolute value of
the dissociation enthalpy ∆H°, counteracted by a small increase
of ∆S°. However, these effects are small leading to the
conclusion that ∆G°298K for duplexes containing dG and cdG
are the same within the error margin of the experiment (see
Table 2).
We next investigated the base pairing preferences of 8-oxodG
and determined in accord with the literature that the 8-oxodG
lesion features the highest absolute value of ∆G°298K when base
paired with dC and dA. The data therefore reflect the well-
known phenomenon that 8-oxodG pairs with dC in the anti-
and with dA in the syn-conformation as depicted in Scheme
1c; 8-oxodG:dG or 8-oxodG:dT-pairs behaved like dG:dG or
dG:dT mismatches, respectively.
(24) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A.; Stratmann,
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,
K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,
J. V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.;
Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham,
M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-
Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.9;
Gaussian, Inc.: Wallingford, CT, Pittsburgh, PA, 1998.
We finally studied the base pairing characteristics of the
FaPydG lesion using the cFaPydG analogue. The presence of
the cFaPydG lesion in DNA conferred rather large duplex
destabilizations. The highest melting point, and hence the best
(25) Blackburn, C. M.; Gait, M. J. Nucleic acids in chemistry and biology;
Oxford University Press: New York, USA, 1996.
(23) Marky, L. A.; Breslauer, K. J. Biopolymers 1987, 26, 1601-1620.
9
18146 J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005

Base Pairing and Replication of the FaPydG-DNA Lesion
A R T I C L E S
Table 2. Thermodynamic Melting Data of Duplexes Containing dG, cdG, 8-oxodG, and cFaPydG
5
′
′
GCGATXTAGCG (d1)
CGCTAYATCGC
5′
3′
TGCAGTXACAGC (d2)
ACGTCAYTGTCG
3
b
b
b
b
base paira
X, Y
∆
(
G
°
∆
H
°
∆
S
°
∆
(
G
°
∆
H
°
∆S°
298
298
1
1
1
1
°
C)
(kcal mol-
)
(cal mol-1 K-
)
°
C)
(kcal mol-
)
(cal mol-1 K-
)
dG, dA
-9.0
-14.2
-9.9
-46 ( 1.1
-74 ( 1.5
-51 ( 0.7
-54 ( 1.8
-122 ( 3.8
-199 ( 4.7
-137 ( 2.2
-146 ( 5.7
-11.5
-16.0
-11.5
-12.4
-68 ( 1.3
-85 ( 2.3
-68 ( 1.5
-73 ( 0.9
-189 ( 4.1
-231 ( 6.9
-190 ( 4.9
-204 ( 2.7
dG, dC
dG, dG
dG, dT
-10.3
cdG, dA
cdG, dC
cdG, dG
cdG, dT
-9.8
-13.5
-11.7
-10.5
-49 ( 2.2
-70 ( 2.6
-62 ( 2.5
-57 ( 2.6
-130 ( 7.0
-191 ( 8.1
-170 ( 8.7
-157 ( 9.6
-11.5
-15.6
-13.0
-12.9
-70 ( 1.3
-87 ( 1.7
-86 ( 1.7
-82 ( 2.4
-196 ( 12.6
-238 ( 11.5
-245 ( 10.8
-232 ( 14.6
8-oxodG, dA
8-oxodG, dC
8-oxodG, dG
8-oxodG, dT
-12.6
-13.1
-9.3
-66 ( 1.4
-70 ( 1.1
-49 ( 0.5
-56 ( 1.4
-178 ( 4.3
-190 ( 3.5
-133 ( 1.5
-154 ( 4.6
-13.3
-15.3
-11.9
-12.2
-74 ( 1.6
-85 ( 1.4
-75 ( 1.3
-78 ( 1.4
-204 ( 4.9
-234 ( 4.2
-211 ( 4.0
-221 ( 4.4
-9.9
cFaPydG, dA
cFaPydG, dC
cFaPydG, dG
cFaPydG, dT
-7.0
-10.2
-6.9
-42 ( 2.0
-51 ( 1.0
-43 ( 2.0
-47 ( 1.0
-118 ( 7.0
-137 ( 3.2
-121 ( 6.8
-129 ( 3.4
-10.4
-13.4
-9.9
-64 ( 1.6
-80 ( 1.8
-61 ( 1.1
-68 ( 1.1
-181 ( 5.0
-222 ( 5.7
-173 ( 3.7
-193 ( 3.7
-8.3
-10.8
b
a Condititions: 150 mm NaCl, 10 mm Tris/HCl, pH 7.4, coligo) 0.3-27 µm. The error is calculated by propagation of the standard error of the linear
regression.
Figure 2. Comparison of the rotation potential around the glycosylic bond. Each data point represents the relative energy of a complete geometry optimization
employing B3LYP/6-31G* density functional calculations. The data point of each curve with the lowest energy was set to 0. The gray bars highlight the
torsion angles ø corresponding to the conformations syn-, anti-, and high anti- as defined by Blackburn et al.25 (a) Comparison of natural FaPydG and the
carbocyclic analogue; (b) comparison of 8-oxodG with 2′-deoxyguanosine. 8-oxodG has a second minimum at the syn-region.
of only 12-16 kcal/mol, we can assume free rotation at room
temperature, which enables cFaPydG and FaPydG to exhibit a
strong preference for the anti-conformation. The molecular
geometries for the global minima of the monomers are almost
superimposable (Supporting Information Figure s1). These
results are furthermore supported by a two-dimensional torsional
screen around both the C1′-NH and the NH-C4 bond of both
monomers using the force field mmff94s.26,27 The rotation
potentials of both compounds are first of all again almost
identical. In addition, the computational data support a global
minimum with a torsional angle around the glycosidic bond in
the anti-conformation (Supporting Information Figure s2).
Calculation of the energy functions of 8-oxodG shows,
however, the expected additional minimum with the syn
conformation of the glycosidic bond. The syn conformation is
in fact the calculated global minimum, which explains why
8-oxodG base pairs preferentially with dA and dC. The
computed data are in full agreement with experimental observa-
tions with 8-oxodG and other 8-substituted purines.28,29
The structural reason for 8-oxodG preferring the syn-
conformation is a simple steric clash of the C8-carbonyl group
and the sugar C5′ methylene group in the anti-conformation as
observed also in crystal structures.30 In the FaPydG lesion,
however, this C8dO is part of the formamide, which possesses
additional degrees of freedom. This freedom allows the forma-
mide group to rotate out of the plane with the pyrimidine ring.
By rotation of the formamide function the steric clash can be
avoided, which allows this lesion to adopt the anti-conformation
with great preference.
Primer Extension Kinetics: The obtained thermodynamic
data clarify the base pairing properties of the FaPydG lesion
but not the coding potential during replication. Replication
requires base pairing within the active site of a DNA poly-
merase. This is next to thermodynamic parameters determined
by the steric constraints within the active site of the DNA
polymerase. We have recently determined the efficiency of base
incorporation opposite a cFaPydG present in a template strand
(28) Tavale, S. S.; Sobell, H. M. J. Mol. Biol. 1970, 48, 109-123.
(29) Uesugi, S.; Ikehara, M. J. Am. Chem. Soc. 1977, 99, 3250-3253.
(30) Hsu, G. W.; Ober, M.; Carell, T.; Beese, L. S. Nature 2004, 431, 217-
221.
(26) Halgren, T. A. J. Comput. Chem. 1999, 20, 720-729.
(27) SYBYL, V6.8; Tripos Inc., 1699 South Hanley Rd., St. Louis, Missouri,
63144, USA, 2003.
9
J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005 18147

A R T I C L E S
Ober et al.
Table 3. Kinetic Parameters of dNTP (Y) Insertion by G. stearothermophilus Pol I and Yeast Pol η Opposite dG, cdG, or cFaPydG (X)
G. stearothermophilus Pol I:
S.cerevisiae Pol
-TAGXCCTGGTCATT
GGACCAGTAA(Fl)-3
(X ) dG, cdG, cFaPydG)
η:
5′
-CATXCGAGTCAGGCT
5′
GCTCAGTCCG (31P)-3
′
′
(X
Km
mol)
)
dG, cFaPydG)
a
a
a
incorporation
kcat
(min-
kcat/Km
mol-1min-
kcat
Km
mol)
kcat/Km
mol-1 min-
1
1
1
1
b
Y
f
X
)
(
µ
(µ
)
fins
(min-
)
(
µ
(µ
)
fins
dCTP f dG
dTTP f dG
dATP f dG
dGTP f dG
750
0.25
1.5
58
111
86
500
1
31.3 ( 2.8
2.57 ( 0.10
1.71 ( 0.10
3.27 ( 0.67
190 ( 25
103 ( 26
9.58 ( 2.8
1
4.3 × 10-3
6.3 × 10-5
9.3 × 10-6
8.6 × 10-6
1.3 × 10-7
1.9 × 10-8
(1.35 ( 0.2) × 10-2
(1.4 ( 0.7) × 10-3
7 × 10-4
(1.65 ( 0.5) × 10-2
(1.7 ( 1.1) × 10-3
8 × 10-4
dCTP f cdG
dTTP f cdG
dATP f cdG
42.8 ( 2.3
8.54 ( 0.65
0.84 ( 0.02
4.10 ( 0.46
359 ( 78
84.0 ( 6.9
10.4 ( 1.7
1
(2.38 ( 0.7) × 10-2
(2.3 ( 1.0) × 10-3
(10.0 ( 0.1) × 10-2
(1.0 ( 0.3) × 10-3
dCTP f cFaPydG
dTTP f cFaPydG
dATP f cFaPydG
dGTP f cFaPydG
520
1.3
97
84
400
0.36
1
35.0 ( 3.2
1.85 ( 0.08
1.72 ( 0.13
12.1 ( 2.5
83.0 ( 13
154 ( 34
2.90 ( 0.9
1
35.2
12.1
0.06
9.0 × 10-4
3.5 × 10-4
1.2 × 10-6
(2.24 ( 0.5) × 10-2
(7.7 ( 3.8) × 10-3
0.14
(1.11 ( 0.3) × 10-2
(3.8 ( 2.3) × 10-3
130
4.6 × 10-4
b
a The error is the standard error of the nonlinear regression or calculated by error propargation. (kcat/Km)dNTP/(kcat/Km)dCTP
.
by a high fidelity polymerase (G. stearothermophilus DNA
Pol I). The obtained data are for comparative reasons listed in
Table 3 together with new data obtained with the low-fidelity
polymerase Pol η. Pol η tolerates base modifications and is
capable in rescueing damaged DNA. The polymerase possesses
a less constrained active site allowing us to better estimate the
coding flexibility of the cFaPydG lesion.
To investigate, how this enzyme processes cFaPydG lesions
in the template strand, steady-state kinetics of dNTP incorpora-
tion by S. cereVisiae Pol η opposite dG, cdG, and cFaPydG
were performed. To this end a 54 amino acid C-terminally
truncated Pol η31 protein was prepared from the RAD30 gene
of the yeast strain YPH499 (ATCC 76625) as described in the
Supporting Information. This truncation has no effect on the
polymerase activity but simplifies polymerase purification
procedures. For the kinetic studies, we used the template-primer
system depicted in Table 3 and standard Pol η primer extension
conditions.32-37 For a better quantification of the fluorescein
labeled primer and its extension products, we employed capillary
electrophoresis coupled to a laser-induced fluorescence detector
instead of 31P phosphor imaging techniques. The sensitivity of
this method is comparable to radioactive quantification. Peak
integration, however, is more accurate. Figure 3 shows a typical
plot of a primer extension experiment. The primer strand was
detected with a retention time of 22.4 min. The one nucleobase
elongated primer strand (n + 1) appeared at a slightly larger
retention time of 22.7 min.
Figure 3. Incorporation of dNTPs opposite cFaPydG by yeast Pol η. (a)
Stacked plot of normalized fluorescence signals determined by capillary
electrophoresis. The tracks refer to raising dCTP concentrations.
directly to the Michaelis-Menton equation to calculate the
paramters Vmax and Km. The value kcat was derived as an enzyme
concentration independent constant via the equation kcat)Vmax
cenz
/
.
Comparison of the efficiency constants kcat/Km (Table 3) for
the incorporation of dCTP, dATP, and dTTP indicates that
correct insertion of dCTP opposite dG and cdG is in the range
of (4.4 ( 3.5) × 102 and (1.0 ( 0.3) × 103 more efficient than
misincorporation of dATP or dTTP. The values for dNTP
incorporation opposite dG and cdG are again as expected the
same within the error margin, showing that the effect of the
chemical modification O to CH2 is small even inside the active
site of the polymerase.
To identify the preferred dNTP inserted by the polymerase
opposite cFaPydG, primer extension reactions were performed
in the presence of a single type of triphosphate at a time using
increasing dNTP concentrations.38 The relationship between the
rate and dNTP concentration was fitted by nonlinear regression
dCTP incorporation opposite cFaPydG is performed by the
polymerase with an efficiency of about 30 (Pol η) - 80%
(Pol I) compared to the insertion of dCTP opposite dG or cdG.
The studies show that cFaPydG is not a block but a rather good
templating base, again in contrast to earlier observations by the
Greenberg group39 and also in contrast to results from the
OHC-NCH3-FaPydG derivative of Ide et al.10 Incorporation
of dATP and dTTP opposite cFaPydG by Pol η is (2.6 ( 1.5)
× 102 and (1.3 ( 0.7) × 102 times less efficient compared to
incorporation of dCTP. Compared to the data for G. stearo-
thermophilus Pol I, the selectivity of the low fidelity Pol η is
(31) Kondratick, C. M.; Washington, M. T.; Prakash, S.; Prakash, L. Mol. Cell.
Biol. 2001, 21, 2018-2025.
(32) Kusumoto, R.; Masutani, C.; Iwai, S.; Hanaoka, F. Biochemistry 2002, 41,
6090-6099.
(33) Haracska, L.; Yu, S. L.; Johnson, R. E.; Prakash, L.; Prakash, S. Nature
Genet. 2000, 25, 458-461.
(34) Johnson, R. E.; Washington, M. T.; Prakash, S.; Prakash, L. J. Biol. Chem.
2000, 275, 7447-7450.
(35) Washington, M. T.; Johnson, R. E.; Prakash, S.; Prakash, L. J. Biol. Chem.
1999, 274, 36835-36838.
(36) Hwang, H. S.; Taylor, J. S. Biochemistry 2004, 43, 14612-14623.
(37) Masutani, C.; Kusumoto, R.; Iwai, S.; Hanaoka, F. EMBO J. 2000, 19,
3100-3109.
(38) Creighton, S.; Bloom, L. B.; Goodman, M. F. Method Enzymol. 1995, 262,
232-257.
(39) Wiederholt, C. J.; Greenberg, M. M. J. Am. Chem. Soc. 2002, 124, 7278-
7279.
9
18148 J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005

Base Pairing and Replication of the FaPydG-DNA Lesion
A R T I C L E S
reduced by 1 order of magnitude, but the specificity sequence
dCTP , dTTP < dATP is retained. In summary, polymerases
process cFaPydG efficiently in an error-free fashion. Some
reduction of selectivity is observed, which may be responsible
for the mutational effect. The thermodynamic melting point
studies and the primer extension studies with high and low
fidelity polymerases argue that FaPydG may be responsible for
the incorporation of dTTP, which would result in dG to dA
transition mutations. The incorporation of dATP opposite the
lesion is less likely based on our studies with Pol I or Pol η.
This is in agreement with crystallographic studies17 and with
an earlier report by Ide et al.10 employing the OHC-NCH3-
FaPydG analogue and E. coli Pol I Klenow fragment.
the largely destabilizing effect of DNA duplexes by the FaPydG
lesion, the lesion fully retains its coding potential and also does
not constitute a polymerase blocking unit. dCTP was efficiently
inserted when the lesion was processed by DNA Pol I or Pol η.
However, the lesion gives rise to a reduced polymerase fidelity,
which may allow formation of a FaPydG:dT base pair to some
extent. This could be the basis for dG to dA transition mutations.
Acknowledgment. The work was supported by the Volk-
swagenstiftung, the Fonds der chemischen Industrie, the Deut-
sche Forschungsgemeinschaft (DFG), funds from the State of
Bavaria and support from the Gottfried Wilhelm Leibniz
program of the DFG. Calculations were performed on the CuP
computer cluster of the LMU Munich. We thank Dr. Glenn
Burley for carefully reading the manuscript and the Boehringer
Ingelheim Fonds for a predoctoral fellowship to M.O.
Concluding Remarks
The cyclopentane analogue cFaPydG of the FaPydG lesion
was synthesized, and the base pairing capabilities inside and
outside a polymerase active site were investigated and compared
to data obtained with dG and cdG in the same sequence context.
dG and cdG show similar properties in our studies, indicating
that data derived with cyclopentane analogues of 2′-deoxygua-
nosine-derived DNA lesions are solid. cFaPydG lowers DNA
duplex stability significantly. The favored cFaPydG:dC base pair
has, in DNA, a free dissociation enthalpy similar to a typical
dG:dT mismatch. According to our computational and thermo-
dynamic data, cFaPydG does not adopt a syn-conformation,
which would be a prerequisite for base pairing with dA. Despite
Supporting Information Available: An Experimental Section
including information concerning the synthesis and chemical
data for 1-13 (including full information of the synthesis of
compound 1), oligonucleotide synthesis and analysis via HPLC-
MS/MS, van’t Hoff plots of the melting point experiments,
cloning and overexpression of S. cereVisiae Pol η, further details
on the kinetic experiments. CIF coordinates of compound 3.
This material is available free of charge via the Internet at
http://pubs.acs.org.
JA0549188
9
J. AM. CHEM. SOC. VOL. 127, NO. 51, 2005 18149