Synthesis and characterization of oligodeoxynucleotides containing formamidopyrimidine lesions and nonhydrolyzable analogues

Article abstract of DOI:10.1021/ja012135q

Oligodeoxynucleotides containing formamidopyrimidine lesions and C-nucleoside analogues at defined sites were prepared by solid-phase synthesis and in some cases enzymatic ligation. Formamidopyrimidine lesions were introduced as dinucleotides to prevent r

Full text of DOI:10.1021/ja012135q

Published on Web 03/12/2002
Synthesis and Characterization of Oligodeoxynucleotides
Containing Formamidopyrimidine Lesions and
Nonhydrolyzable Analogues
Kazuhiro Haraguchi, Michael O. Delaney, Carissa J. Wiederholt,
Aruna Sambandam, Zsolt Hantosi, and Marc M. Greenberg*
Contribution from the Department of Chemistry, Colorado State UniVersity,
Fort Collins, Colorado 80523
Received September 6, 2001
Abstract: Oligodeoxynucleotides containing formamidopyrimidine lesions and C-nucleoside analogues at
defined sites were prepared by solid-phase synthesis and in some cases enzymatic ligation. Formami-
dopyrimidine lesions were introduced as dinucleotides to prevent rearrangement to their pyranose isomers.
Oligodeoxynucleotides containing single diastereomers of C-nucleoside analogues of Fapy‚dA were
introduced by using the respective phosphoramidites. The formamidopyrimidine lesions reduce the TM of
dodecamers relative to their unmodified nucleotide counterparts when opposite the nucleotide proper base-
pairing partner. However, duplexes containing Fapy‚dG-dA mispairs melt significantly higher than those
comprised of dG-dA. All duplexes containing Fapy‚dA-dX or its C-nucleoside analogue melt lower than the
respective complexes containing dA-dX. Studies of the alkaline lability of oligodeoxynucleotides containing
formamidopyrimidine lesions indicate that Fapy‚dA is readily identified as an alkali-labile lesion with use of
piperidine (1.0 M, 90 °C, 20 min), but Fapy‚dG is less easily identified in this manner.
DNA undergoes a variety of chemically induced structural
changes when exposed to reactive oxygen species (e.g. hydroxyl
radical) and other chemical agents.1,2 Many of the lesions
produced in DNA result from attack by radicals and other
electrophiles on the nucleobases and are involved in aging and
diseases such as cancer.3 The effects of these lesions on
polymerase enzyme activity and their recognition by DNA repair
enzymes are critical in determining their biological role.4 These
interactions are a manifestation of the lesions’ structures and
their effect on the shape and stability of the DNA duplex as a
whole. Synthesis of a biopolymer containing a specific lesion
at a defined site provides a valuable tool for elucidating the
structural and functional properties of damaged DNA. Although
synthesis of unmodified nucleic acids is routine, preparation of
DNA containing damaged nucleotides can be challenging due
to the decreased stability of a lesion to synthesis and deprotection
conditions.5,6
recognized by a variety of DNA repair enzymes.7 γ-Radiolysis
produces Fapy‚dG in greater amounts in vivo than the respective
8-oxopurine lesion (OxodG), which is formed from a common
intermediate (Scheme 1).8 Fapy‚dA and 8-oxodeoxyadenosine
(OxodA) are formed via an analogous mechanism. The effects
of formamidopyrimidines on polymerase enzymes are uncertain,
but evidence suggests that both are premutagenic lesions.9
Cleavage of the purine imidazole ring in deoxyadenosine and
deoxyguanosine introduces additional functional groups and a
greater number of degrees of freedom in the formamidopyri-
midine lesions than are present in the native nucleotides or the
8-oxopurines. This expands the number of possible hydrogen
bonding motifs that can interact with enzymes, but also increases
the synthetic challenge. Synthesis of oligodeoxynucleotides
containing Fapy‚dA and Fapy‚dG are complicated by the fact
that the monomeric species readily isomerize to their pyranose
congeners.10 We developed an approach that overcomes this
and other obstacles and enables one to synthesize an oligode-
The formamidopyrimidines are examples of DNA lesions that
are of biological and chemical interest. These lesions are
(6) (a) For a recent review see: Butenandt, J.; Burgdorf, L. T.; Carrell, T.
Synthesis 1999, 1085. (b) Taylor, J.-S.; O’Day, C. L. J. Am. Chem. Soc.
1989, 111, 401. (c) Shibutani, S.; Takeshita, M.; Grollman, A. P. Nature
1991, 349, 431. (d) Iwai, S.; Shimizu, M.; Kamiya, H.; Ohtsuka, E. J. Am.
Chem. Soc. 1996, 118, 7642. (e) Kobertz, W. R.; Essigmann, J. M. J. Am.
Chem. Soc. 1997, 119, 5960. (f) Sambandam, A.; Greenberg, M. M. Nucleic
Acids Res. 1999, 27, 3597. (g) Iwai, S. Angew. Chem., Int. Ed. 2000, 39,
3874.
(7) (a) Boiteux, S.; Gajewski, E.; Laval, J.; Dizdaroglu, M. Biochemistry 1992,
31, 106. (b) Krakaya, A.; Jaruga, P.; Bohr, V. A.; Grollman, A. P.;
Dizdaroglu, M. Nucleic Acids Res. 1997, 25, 474. (c) Dizdaroglu, M.;
Bauche, C.; Rodriguez, H.; Laval, J. Biochemistry 2000, 39, 5586.
(8) Pouget, J.-P.; Douki, T.; Richard, M.-J.; Cadet, J. Chem. Res. Toxicol. 2000,
13, 541.
(1) (a) von Sonntag, C. The Chemical Basis of Radiation Biology; Taylor and
Francis: London, 1987. (b) Greenberg, M. M. Chem. Res. Toxicol. 1998,
11, 1235. (c) Greenberg, M. M. In ComprehensiVe Natural Products
Chemistry, Vol. 7, DNA and Aspects of Molecular Biology; Kool, E. T.,
Ed.; Pergamon: Oxford, 1999; pp 371-426.
(2) Gates, K. S. In ComprehensiVe Natural Products Chemistry, Vol. 7, DNA
and Aspects of Molecular Biology; Kool, E. T., Ed.; Pergamon: Oxford,
1999; pp 491-552.
(3) (a) Hoeijmakers, J. H. Nature 2001, 411, 366. (b) DePinho, R. A. Nature
2000, 408, 248.
(4) (a) Lindahl, T.; Wood, R. D. Science 1999, 286, 1897. (b) Laval, J.; Jurado,
J.; Saparbaev, M.; Sidokina, O. Mutation Res. 1998, 402, 93. (c) Delaney,
J. C.; Essigmann, J. M. Chem. Biol. 1999, 6, 743. (d) Wang, D.; Kreutzer,
D. A.; Essigmann, J. M. Mutat. Res. 1998, 400, 99.
(9) Graziewicz, M.-A.; Zastawny, T. H.; Oli’nski, R.; Tudek, B. Mutation Res.
1999, 434, 41.
(10) (a) Raoul, S.; Bardet, M.; Cadet, J. Chem. Res. Toxicol. 1995, 8, 924. (b)
Berger, M.; Cadet, J. Z. Naturforsch. 1985, 40b, 1519.
(5) Iyer, R. P.; Beaucage, S. L. In ComprehensiVe Natural Products Chemistry;
Kool, E. T., Ed.; Pergamon: Amsterdam, 1999; Vol. 7, pp 105-152.
9
10.1021/ja012135q CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 3263-3269
3263

A R T I C L E S
Haraguchi et al.
Scheme 1
Scheme 2
Scheme 3
oxynucleotide containing Fapy‚dA or Fapy‚dG at a defined
site.11 The ability to prepare oligodeoxynucleotides containing
DNA lesions at defined sites greatly facilitates examination of
biological, chemical, and structural issues. Nucleoside analogues
also play useful roles in such studies and we have prepared
configurationally stable C-nucleosides of Fapy‚dA (R,â-1) for
this purpose.12,13
indicated that formamidopyrimidine nucleosides exist predomi-
nantly as the R,â-pyranosyl isomers and minor amounts of the
respective R,â-furanosyl species.10 We reasoned this apparently
facile rearrangement would prohibit a traditional oligodeoxy-
nucleotide synthesis approach in which monomers are added
one nucleotide at a time in the 3′ f 5′ direction. In principle,
the formamidopyrimidine lesions could also be incorporated into
growing chains as the mononucleotide by using reverse phos-
phoramidites in which the dimethoxytrityl group (DMT) protects
the 3′-OH and the phosphoramidite is appended to the primary
alcohol. However, this strategy would require carrying the labile
phosphoramidite functionality through several steps of its
synthesis and was not deemed likely to be successful. We
avoided enabling rearrangement of the formamidopyrimidine
nucleosides to their pyranose forms by introducing the lesions
as the 3′-component of a dinucleotide. Our retrosynthetic
analysis (Scheme 2) relied upon the assumption that a 5-nitro
substituent would destabilize the acyclic imine intermediate
necessary for isomerization and that we could transform this
functional group into a formamide following formation of the
dinucleotide phosphate triester (Scheme 3). This strategy would
not prevent epimerization of later synthetic intermediates
containing formamidopyrimidine groups. However, stereocontrol
at the anomeric center was considered unnecessary because the
anomers of Fapy lesions readily epimerize in water. For instance,
we determined that 2, which cannot rearrange to a pyranose
Results and Discussion
Synthesis of Dinucleotide Phosphoramidites Containing
Formamidopyrimidine Lesions. We recently reported the
synthesis of oligodeoxynucleotides containing Fapy‚dG at
defined sites.11 The successful approach involved a strategy that
respected the incompatibility of the formamidopyrimidine and
primary hydroxyl group. Radiolysis experiments on nucleosides
(11) Haraguchi, K.; Greenberg, M. M. J. Am. Chem. Soc. 2001, 123, 8636.
(12) For some recent examples see: (a) Kool, E. T. Annu. ReV. Biophys. Biomol.
Struct. 2001, 30, 1. (b) Sun, L.; Wang, M.; Kool, E. T.; Taylor, J.-S.
Biochemistry 2000, 39, 14603. (c) Morales, J. C.; Kool, E. T. J. Am. Chem.
Soc. 2000, 122, 1001. (d) Tae, E. L.; Wu, Y.; Xia, G.; Schultz, P. G.;
Romesberg, F. E. J. Am. Chem. Soc. 2001, 123, 7439.
(13) (a) Zharkov, D. O.; Rosenquist, T. A.; Gerchman, S. E.; Grollman, A. P.
J. Biol. Chem. 2000, 275, 28607. (b) Chepanoske, C. L.; Porello, S. L.;
Fujiwara, T.; Sugiyama, H.; David, S. S. Nucleic Acids Res. 1999, 27, 3197.
9
3264 J. AM. CHEM. SOC. VOL. 124, NO. 13, 2002

Synthesis and Characterization of Oligodeoxynucleotides
A R T I C L E S
Scheme 4a
a Key: (a) TMSN3, TMSOTf, CH2Cl2. (b) NaOMe, MeOH. (c) DMTCl,
pyridine. (d) TBDMSCl, imidazole, pyridine. (e) H2, Pd/CaCO3, EtOH. (f)
6-Amino-4-chloro-5-nitropyrimidine, Et3N, dioxane. (g) AcOH/H2O (4:1),
CH2Cl2.
isomer, reaches equilibrium between its R- and â-anomers in
less than 7 h at room temperature (10 mM phosphate buffer,
pH 7.5).14
The desired isomers (8, 9) were also readily distinguished
by using NMR from the pyranose rearrangement products (10,
11). The pyranose isomers (10, 11) were independently syn-
thesized from the benzyl acetal (12) by a similar route and were
1
easily distinguished from 8 and 9 by examination of the H
NMR chemical shifts.18 The C1′ protons in the furanose species
were significantly more deshielded than the analogous protons
in 10 and 11.19 This trend is consistent with that observed for
deoxyglucopyranosyl containing analogues of DNA.20 The
pyranose isomers were also distinguishable from 8 and 9
qualitatively by reacting each with NaIO4. The compounds
identified by NMR to contain furanose rings failed to react under
conditions that rapidly consumed 10 and 11. Importantly, these
experiments enabled us to determine rearrangement had not
occurred during detritylation of 6. This is in contrast to the
analogous Fapy‚dG species, which are moderately reactive under
acidic conditions (see below).
All attempts to acylate the N4 and/or N6 amino groups in 6
were unsuccessful. Presumably, the combination of the 5-nitro
substituent and electron-poor pyrimidine ring accounted for the
poor nucleophilicity of these amines. Consequently, we elected
to proceed by using the major isomer (7R) with the amines
unprotected (Scheme 5). It was necessary to carry out dinucle-
otide formation with 7 using the O-methyl phosphoramidite of
thymidine because the â-cyanoethyl group was unstable to the
subsequent hydrogenation conditions. Excellent yields of the
phosphate triester (13) were obtained when tert-butylhydrop-
eroxide was used as oxidant. There was no evidence for
epimerization during this reaction, but isomerization did occur
during the next step in which the nitro group is reduced.
Reduction was carried out using large amounts Lindlar’s catalyst
to complete reduction overnight. The crude amine was then
formylated in the presence of excess acetic formic anhydride.
No attempts were made to separate the mixture of anomers of
14, because we established previously that the formamidopy-
rimidine anomers rapidly equilibrate in water.14 Formylation
The stability of nitropyrimidine 7 was crucial to the success
of this approach for the synthesis of an appropriate phosphora-
midite. This intermediate was prepared via nucleophilic aromatic
substitution of the crude amine obtained upon selective reduction
of the azide in 4. There was no evidence for detritylation during
the hydrogenation of 4, which was prepared in 7 steps from
deoxyribose via the known triacetate (3, Scheme 4).15 The azide
(4) was obtained in 38% yield overall and this method was
superior in our hands to a slightly shorter one proceeding
through 2-deoxyribonolactone that was not amenable to produc-
ing the necessary azide in suitable quantities.16 The tritylated
glycosylation product (6) was obtained as a mixture of anomers
(3:1) that were separable upon deprotection to the configuration-
ally stable primary alcohols (7). The R-product was found to
be the major isomer. Stereochemical assignments were based
upon chemical shifts of the carbohydrate protons in 7R,â, 8,
and 9, as well as NOE experiments on the diols (8, 9). 1H NMR
analysis of analogous compound 2 revealed that the cis
orientation of the exocyclic heteroatom substituents at C1′ and
C3′ resulted in a significant chemical shift difference of the
diastereotopic C2′ protons in the R-isomer.14 In contrast, the
effects of the heteroatom substituents at these positions opposed
one another in the â-anomer, giving rise to a pair of protons
whose chemical shifts were unresolved. These trends were
upheld in 7-9.17 Furthermore, NOE experiments on 8 and 9
affirmed the assigned stereochemistry. Irradiation of the C1′
proton in 9 gave rise to an enhancement at C4′. In contrast,
irradiation of the C1′ proton in 8 resulted in an enhancement of
the more deshielded C2′ proton but no effect on the respective
C4′ proton was observed.
(14) Greenberg, M. M.; Hantosi, Z.; Wiederholt, C. J.; Rithner, C. D.
Biochemistry 2001, 40, 15856.
(18) See Supporting Information.
(19) The following H1′ chemical shifts were recorded at 300 MHz in CD3OD:
8, 6.42; 9, 6.37; 10, 5.90; 11, 5.83.
(15) Gold, A.; Sangaiah, R. Nucleosides Nucleotides 1990, 9, 907.
(16) Wichai, U.; Woski, S. A. Org. Lett. 1999, 1, 1173.
(17) The following H2′ chemical shifts were recorded at 300 MHz: 7R
(CDCl3): pro-S, 2.30-2.46, pro-R, 2.03; 7â (CDCl3): pro-S and pro-R,
2.15-2.34; 8 (CD3OD): pro-S, 2.33-2.44, pro-R, 1.96-2.04; 9 (CD3-
OD): pro-S and pro-R, 2.20-2.32.
(20) (a) Otting, G.; Billeter, M.; Wu¨trich, K.; Roth, H.-J.; Leumann, C.;
Eschenmoser, A. HelV. Chim. Acta 1993, 76, 2701. (b) Bo¨hringe, M.; Roth,
H.-J.; Hunziker, J.; Go¨bel, M.; Krishnan, R.; Giger, A.; Schweizer, B.;
Schreiber, J.; Leumann, C.; Eschenmoser, A. HelV. Chim. Acta 1992, 75,
1416.
9
J. AM. CHEM. SOC. VOL. 124, NO. 13, 2002 3265

A R T I C L E S
Haraguchi et al.
Scheme 5a
Scheme 6a
a Key: (a) Bu4N+ F-, AcOH, THF. (b) 5′-Silyl thymidine phosphora-
midite, then t-BuOOH. (c) H2, 10% Pd/C, MeOH, THF. (d) HC(O)OAc,
pyridine, THF. (e) Bu4N+ F-, AcOH, THF. (f) 2-Cyanoethyl phosphora-
midic chloride, diisopropylethylamine, CH2Cl2.
comparable transformations were carried out in a similar manner
as described for the synthesis of 15 (Scheme 6). However, some
specific steps deserve mention. For instance, it was necessary
to carry out the desilylation of 19 and 21 with buffered TBAF
to prevent cleavage of the phenoxyacetyl group. The lability of
this group was also problematic during the reduction of the nitro
group in 20 and it was necessary to carry out the reduction with
the more reactive catalyst (10% Pd/C) to minimize exposure of
the phenoxyacetamide to the C5-amino group.
a Key: (a) R-7, tetrazole, then t-BuOOH. (b) H2, Pd/CaCO3, MeOH,
THF. (c) HC(O)OAc, pyridine, THF. (d) Bu4N+ F-, AcOH, THF. (e)
2-Cyanoethyl phosphoramidic chloride, diisopropylethylamine, CH2Cl2.
occurred exclusively at the N5-position, reflecting the effect of
the pyrimdine ring nitrogen atoms, which are ortho and para to
the N4 and N6 amines, but meta with respect to the position
where reaction occurred. Attempts to protect the free amines in
the presence of the formamide group were again unsuccessful.
Under forcing conditions deformylation and/or transacylation
at N5 were observed. We reasoned that if the free amines were
so unreactive with acylating agents then there was no need to
protect them during oligodeoxynucleotide synthesis. Conse-
quently, 14 was carried on to the requisite dinucleotide
phosphoramidite (15) by desilylating with buffered TBAF and
phosphitylating with standard conditions.5
A similar overall approach was reported in preliminary form
for the synthesis of a dinucleotide phosphoramidite suitable for
preparing oligodeoxynucleotides containing Fapy‚dG.11 How-
ever, the properties of the Fapy‚dG nucleoside were sufficiently
different from those of the adenine analogue so as to require
significant modifications in the synthesis described above.
Despite extensive investigation, cleavage of the 5′-dimethox-
ytrityl group in 16 under acidic conditions yielded the desired
product (17) accompanied by varying amounts of pyranosyl
isomers (eq 1). After considering several alternative protecting
Synthesis of a C-Nucleoside Analogue of Fapy‚dA. The
issue of furanose pyranose equilibration was replaced by other
concerns in the synthesis of a nonhydrolyzable analogue of
Fapy‚dA (R,â-1). The core structure was assembled as a mixture
of epimers via a Wittig reaction and subsequent recyclization
(Scheme 7).22 The lactol (22) was prepared from the lactone
via reduction with DIBAL-H.16 Following demethylation the
dicarbonyl compound (25) was converted into 26. The overall
yield of the dichlorination/ammonolysis procedure was low. This
was not due to the nucleophilic substitution step, which
proceeded with high selectivity for reaction at C4 versus C2
(∼6:1). A considerable amount of material was lost due to
desilylation under the chlorination conditions. Despite extensive
experimentation with alternative reagents and variation of
reaction conditions, the yield of the dichlorination step could
not be improved significantly and we elected to move forward.
The remaining chloride was removed and the nitro group
reduced in one pot. The crude diamine was formylated with
use of acetic formic anhydride. Selective formylation occurred
at the N5-position to yield 27, but care had to be taken to prevent
formyl migration to the N4-amino group. Regioisomeric for-
mamides were separable from one another and distinguishable
by 1H NMR. The desired N5-formamides exhibited a significant
NOE between the formyl and benzylic like methylene groups,
as well as a 0.2-0.3 ppm upfield shift of the C2 vinyl protons
relative to the N4-formylation products. The desired formamide
was obtained as the exclusive formylated product in 75% yield
from 26 by using a modest excess of pyridine at 0 °C. Acyl
migration indicated that the N4-amine needed to be protected
during oligodeoxynucleotide synthesis. This was accomplished
in the absence of a nucleophilic base by the in situ formation
of phenoxyacetic anhydride. Following desilylation and subse-
group ensembles that would enable us to unmask the primary
hydroxyl group under nonacidic conditions, we proceeded with
the 5′-silyloxy, 3′-dimethoxytrityl combination.11 Although this
approach committed us to synthesize oligodeoxynucleotides
containing Fapy‚dG in the 5′ f 3′ direction, it enabled us to
prepare the requisite phosphoramidite (18) with the minimum
number of protecting group manipulations.21 Many of the
(21) Rippe, K.; Jovin, T. M. Methods in Enzymology; Lilley, D. M. J., Dahlberg,
J. E., Eds.; Academic Press: San Diego, CA, 1992; Vol. 211, pp 199-
231.
(22) Cupps, T. L.; Wise, D. S.; Townsend, L. B. J. Org. Chem. 1986, 51, 1086.
9
3266 J. AM. CHEM. SOC. VOL. 124, NO. 13, 2002

Synthesis and Characterization of Oligodeoxynucleotides
A R T I C L E S
Scheme 7a
the substrates were tested for their stability to the appropriate
synthesis and deprotection (anhydrous K2CO3, MeOH) reagents.
All of the lesions were stable to K2CO3 (0.05 M) in anhydrous
MeOH for 3 h at 25 °C. This was particularly important for the
C-nucleoside analogues given the above-described acyl migra-
tion. The instability of the formamidopyrimidine nucleosides
to acid was accounted for by their incorporation as dinucleotides.
The substrates underwent transacylation to varying degrees in
the presence of the standard capping conditions (Ac2O, lutidine,
N-methylimidazole). Dodecameric (31, 32) oligodeoxynucle-
otides containing R,â-1 were prepared with isobutyryl anhydride
in place of acetic anhydride after the modified nucleotide is
incorporated. In addition, the coupling time of 30 was extended
to 5 min and a second batch of activated phosphoramidite was
reacted (“double-coupling”) prior to proceeding onto the capping
step of the synthesis cycle. In each instance the overall yield,
as measured by the trityl response, of the synthesis was ∼2-
fold greater when â-30 was coupled. Isolated yields of oligode-
oxynucleotides following deprotection (0.05 M K2CO3/MeOH)
and gel purification were consistent with the measured trityl
responses. Analysis by ESI-MS suggested a minor impurity that
corresponded to dehydration product, presumably resulting from
cyclization between the formyl group and N4. Further examina-
tion revealed that dehydration was occurring in the spectrometer.
The cone voltage of the probe affected the relative amounts of
dehydrated product and uncyclized material. Furthermore,
dehydrated material is evident in the ESI-MS of monomeric
material (35), for which 1H NMR shows no evidence for
cyclized product. There was no evidence for transacylated
product in any of the dodecamers.
a Key: (a) i) Toluene, ∆; (ii) NaOMe, MeOH. (b) NaI, TMSCl, CHCl3.
(c) POCl3, diethylamine, toluene. (d) NH3, THF. (e) H2, Na2CO3, Pd/C,
EtOH. (f) HC(O)OAc, pyridine, THF. (g) Phenoxyacetic anhydride, CH2Cl2.
(h) Et3N‚3HF, THF. (i) DMTCl, DMAP (cat.), pyridine. (j) 2-Cyanoethyl-
N,N,N′,N′-tetraisopropyl phosphane, diisopropylamino tetrazolide, CH2Cl2.
Longer oligodeoxynucleotides (33, 34) containing 1 (36mers)
were prepared by enzymatic ligation of the above dodecamers
(Scheme 8).23 After phosphorylating the dodecamer containing
the modification (e.g. 31, 32) and the 3′-terminal oligodeoxy-
nucleotide, these oligodeoxynucleotides and the 5′-HO-terminal
fragment were hybridized to a DNA template (42mer) and
reacted with T4 DNA Ligase. The desired products were
obtained following gel purification in yields as high as 98%,
but were typically between 53% and 66%. As previously
described, a comparable approach was taken for the synthesis
of oligodeoxynucleotides containing Fapy‚dG.11 Dodecamers
(36, 37) were synthesized by solid-phase synthesis and 37 was
Figure 1. Determination of stereochemistry in 29 by NOE experiments.
quent dimethoxytritylation, the epimeric mixture of products
was separated by column chromatography and the stereochem-
istry at the nominal anomeric center was determined by NOE
experiments (Figure 1). Irradiation of the C1′ proton results in
an enhancement at C4′ in â-29 but not the R-anomer. In contrast,
an interaction is observed between C1′ and C3′ in the R-isomer
but not in â-29. Phosphoramidite synthesis was completed by
using the tetrazolide that was generated in situ.
Oligodeoxynucleotide Synthesis and Characterization.
Prior to carrying out solid-phase oligodeoxynucleotide synthesis
(23) (a) Ordoukhanian, P.; Taylor, J.-S. Nucleic Acids Res. 1997, 25, 3783. (b)
Smith, C. A.; Taylor, J.-S. J. Biol. Chem. 1993, 268, 11143.
9
J. AM. CHEM. SOC. VOL. 124, NO. 13, 2002 3267

A R T I C L E S
Haraguchi et al.
Scheme 8
Table 1. Alkaline Lability of Fapy‚dA and Fapy‚dGa
% cleavage
temp
(°C)
treatmenta
piperidine (1.0 M)
piperidine (1.0 M) + BME (0.25 M)
piperidine (1.0 M)
Fapy‚dA
Fapy‚dG
90 99.0 ( 1.5 39.4 ( 3.4
90 96.5 ( 1.1 54.9 ( 9.9
37 34.9 ( 3.7
6.15 ( 0.1
7.8 ( 1.8
2.2 ( 0.03
N,N′-dimethylethylenediamine (0.1 M) 37 19.6 ( 3.3
NaOH (1.0 M)
NaOH (1.0 M)
NaOH (0.1 M)
NaOH (0.1 M)
55
8.9 ( 0.7
90 73.5 ( 2.5 52.5 ( 5.5
55 3.5 ( 0.7 1.3 ( 0.1
90 12.2 ( 1.4 16.5 ( 0.2
a All cleavage reactions were carried out for 20 min.
Table 2. UV-melting Temperature of Duplex DNA Containing
Fapy‚dG
used to prepare 38. However, oligodeoxynucleotides (39-41)
containing Fapy‚dA as long as 30 nucleotides (41) were
successfully prepared via solid-phase synthesis. The coupling
yields of 15 were greater than 90% with use of a 15 min double-
coupling procedure. We have no evidence to explain why 15
coupled so much more efficiently than 18, but it is tempting to
speculate that the use of a reverse dinucleotide phosphoramidite
during the preparation of Fapy‚dG containing oligodeoxynucle-
otides is a factor. To avoid transacylation capping was carried
out with pivalic anhydride and lutidine in the absence of
N-methyl imidazole activating agent after the dinucleotide was
coupled. To help make up for the less reactive acylation
conditions, capping reaction times were extended to 30 s (from
5 s). There was no evidence for transacylation in the ESI-MS
of any of the formamidopyrimidine containing oligodeoxy-
nucleotides.18
TM (°C)
X
dG‚X
Fapy‚dG‚X
dC
dA
dG
T
56.6 ( 0.4
43.5 ( 0.4
47.1 ( 0.5
46.8 ( 0.3
53.7 ( 0.4
52.1 ( 0.4
43.8 ( 0.1
44.0 ( 0.3
Scheme 9
Table 3. UV-Melting Temperature of Duplex DNA Containing
Fapy‚dA
TM (°C)
X
dA‚X
Fapy‚dA‚X
Alkaline Lability of Oligodeoxynucleotides Containing
Formamidopyrimidine Lesions. Alkali lability of DNA lesions
can serve as a useful initial probe for their presence. For
instance, the formation of alkali labile lesions at deoxyguanosine
is useful for the quantitative analysis of electron hole migration
in DNA.24 Similarly, the formation of adducts of a DNA
cleavage product serves as a fingerprint for the presence of the
2-deoxyribonolactone lesion.25 We used 5′-32P-41 and 5′-32P-
38 to determine the alkaline lability of Fapy‚dA and Fapy‚dG
(Table 1). None of the alkaline conditions completely cleaved
DNA at the Fapy‚dG site in 20 min. Also, unlike OxodG, the
presence of thiol reducing agent had no effect on cleavage of
oligodeoxynucleotides containing Fapy‚dA or Fapy‚dG.26 The
T
54.1 ( 0.2
41.8 ( 0.4
45.5 ( 0.3
40.7 ( 0.5
47.0 ( 0.3
40.7 ( 0.3
41.9 ( 0.6
38.6 ( 0.2
dA
dG
dC
relative amounts of cleavage of the two lesions under various
conditions parallel the greater thermal stability of Fapy‚dG and
are consistent with the intermediate formation of an abasic site.14
The formation of two products containing different 3′-end
termini in varying ratios depending upon the degree of harshness
of the cleavage conditions is also consistent with this interpreta-
tion. Treatment with 1 M piperidine at 90 °C produced a faster
moving product consistent with a 3′-phosphate terminus as the
major product. By analogy to previous studies on abasic sites,
the minor slower moving product, which becomes the major
one under milder cleavage conditions (e.g. N,N′-dimethyleth-
ylenediamine), is believed to be the intermediate â-elimination
product.27,28
(24) (a) Schuster, G. B. Acc. Chem. Res. 2000, 33, 253. (b) Giese, B. Acc. Chem.
Res. 2000, 33, 631. (c) Williams, T. T.; Odom, D. T.; Barton, J. K. J. Am.
Chem. Soc. 2000, 122, 9048.
(25) Hwang, J. T.; Tallman, K. T.; Greenberg, M. M. Nucleic Acids Res. 1999,
27, 3805.
(26) (a) Bodepudi, V.; Shibutani, S.; Johnson, F. Chem. Res. Toxicol. 1992, 5,
608. (b) Cullis, P. M.; Malone, M. E.; Merson-Davies, L. A. J. Am. Chem.
Soc. 1996, 118, 2775.
UV-Melting Studies on Duplex DNA Containing Forma-
midopyrimidines and C-Nucleoside Analogues. Melting tem-
9
3268 J. AM. CHEM. SOC. VOL. 124, NO. 13, 2002

Synthesis and Characterization of Oligodeoxynucleotides
A R T I C L E S
Table 4. UV-Melting Temperature of Duplex DNA Containing 1
possible base pairing schemes. Two hypothetical conformations
of Fapy‚dG that enable it to present a thymidine like hydrogen
bonding pattern are presented for illustrative purposes (Scheme
9). These simple pictures do not take into account possible
distortions of the DNA duplex required to accommodate the
hydrogen bonding patterns shown, and further studies are
needed.
TM (°C)
X
dA‚X
R-1‚X
â-1‚X
T
56.6 ( 0.5
46.7 ( 0.5
48.3 ( 0.6
44.7 ( 0.5
46.8 ( 0.1
46.6 ( 0.5
41.8 ( 0.1
46.4 ( 0.6
46.8 ( 0.1
46.4 ( 0.5
42.6 ( 0.5
45.6 ( 0.9
In contrast, melting temperatures of duplexes containing Fapy‚
dA (Table 3) or its C-nucleoside analogues (Table 4) do not
indicate as obvious a preference for base pairing to a base other
than thymine than does Fapy‚dG. All three compounds depress
the TM compared to the native nucleotide (dA). It may be
significant to note that the duplex containing a Fapy‚dA‚dA
mispair melts relatively higher compared to the appropriate
duplex containing dA than do duplexes containing other Fapy‚
dA mispairs. It is also worth noting that the qualitative effects
of R,â-1 on duplex TM are similar to those of Fapy‚dA,
suggesting that the C-nucleoside analogues are good models of
the lesion.
dA
dG
dC
peratures of duplexes containing the lesions were compared to
the respective DNA dodecamers containing native nucleotides.
All melting temperatures were measured at 2.2 micromolar
duplex. Substitution of the purine in a native base pair by any
of the corresponding lesions depressed the duplex TM. The
decrease in TM was smallest for the dodecamer containing Fapy‚
dG (Table 2). Moreover, placing dA opposite Fapy‚dG produced
a duplex that melted almost as high as that containing a Fapy‚
dG‚dC base pair and significantly higher than the dodecamer
containing dA opposite dG. Although it could be fortuitous, it
is interesting to note that a similar preference for base pairing
with dA was observed for OxodG.29 Given the ability of the
formamidopyrimidines to epimerize one can envision several
Acknowledgment. Financial support of this work from the
National Institutes of Health (CA-74954) is greatly appreciated.
Supporting Information Available: Detailed experimental
procedures, ESI-MS, and MALDI-TOF MS of 31-34 and 36-
41 (PDF). This material is available free of charge via the
Internet at http://pubs.acs.org.
(27) Bailly, V.; Verly, W. G. Biochem. J. 1988, 253, 553.
(28) McHugh, P. J.; Knowland, J. Nucleic Acids Res. 1995, 23, 1664.
(29) Plum, G. E.; Grollman, A. P.; Johnson, F.; Breslauer, K. J. Biochemistry
1995, 34, 16148.
JA012135Q
9
J. AM. CHEM. SOC. VOL. 124, NO. 13, 2002 3269