A highly effective nonpolar isostere of deoxyguanosine: synthesis, structure, stacking, and base pairing.

Article abstract of DOI:10.1021/jo025884e

We describe the preparation and structure of the deoxyribonucleoside of 4-fluoro-6-methylbenzimidazole, abbreviated dH (8), which acts as a close shape mimic of the nucleoside deoxyguanosine. The nucleoside is prepared from 2-fluoro-4-methylaniline in seven steps. The X-ray crystal structure reveals a (-sc) glycosidic orientation, an S conformation for the deoxyribose moiety, and quite close shape mimicry of guanine by the substituted benzimidazole. Conformational studies by (1)H NMR and (1)H-(1)H ROESY experiments reveal an S-type conformation and an anti glycosidic orientation in solution (D(2)O), essentially the same as that of deoxyguanosine. Base-stacking studies in a "dangling end" context reveal that the benzimidazole base mimic stacks more strongly than all four natural bases, and more strongly than its counterpart guanine by 1.1 kcal/mol. Base-pairing studies in a 12mer DNA duplex show that, like other nonpolar nucleoside isosteres, H is destabilizing and nonselective when paired opposite natural bases. However, when paired opposite another nonpolar isostere, difluorotoluene (F), a mimic of thymine, the pair exhibits stability approaching that of its natural analogue, a G-T (wobble) base pair. The nucleoside analogue dH will be useful in studies of protein-DNA interactions, and the H-F base pair will serve as a structurally and thermodynamically close mimic of G-T in studies of DNA mismatch repair enzymes.

Full text of DOI:10.1021/jo025884e

VOLUME 67, NUMBER 17
AUGUST 23, 2002
© Copyright 2002 by the American Chemical Society
A High ly Effective Non p ola r Isoster e of Deoxygu a n osin e:
Syn th esis, Str u ctu r e, Sta ck in g, a n d Ba se P a ir in g
Bryan M. O’Neill,† J essica E. Ratto,† Kristi L. Good,† Deborah C. Tahmassebi,*,†
Sandra A. Helquist,‡ J uan C. Morales,‡ and Eric T. Kool*,‡
Department of Chemistry, University of San Diego, 5998 Alcala Park, San Diego, California 92110, and
Department of Chemistry, Stanford University, Stanford, California 94305-5080
kool@stanford.edu
Received April 30, 2002
We describe the preparation and structure of the deoxyribonucleoside of 4-fluoro-6-methylbenz-
imidazole, abbreviated dH (8), which acts as a close shape mimic of the nucleoside deoxyguanosine.
The nucleoside is prepared from 2-fluoro-4-methylaniline in seven steps. The X-ray crystal structure
reveals a (-sc) glycosidic orientation, an S conformation for the deoxyribose moiety, and quite close
1
shape mimicry of guanine by the substituted benzimidazole. Conformational studies by H NMR
1
and H-1H ROESY experiments reveal an S-type conformation and an anti glycosidic orientation
in solution (D2O), essentially the same as that of deoxyguanosine. Base-stacking studies in a
“dangling end” context reveal that the benzimidazole base mimic stacks more strongly than all
four natural bases, and more strongly than its counterpart guanine by 1.1 kcal/mol. Base-pairing
studies in a 12mer DNA duplex show that, like other nonpolar nucleoside isosteres, H is destabilizing
and nonselective when paired opposite natural bases. However, when paired opposite another
nonpolar isostere, difluorotoluene (F), a mimic of thymine, the pair exhibits stability approaching
that of its natural analogue, a G-T (wobble) base pair. The nucleoside analogue dH will be useful
in studies of protein-DNA interactions, and the H-F base pair will serve as a structurally and
thermodynamically close mimic of G-T in studies of DNA mismatch repair enzymes.
In tr od u ction
methylbenzimidazole (dZ)4 or 4-methylindole5 replaces
the adenine base.
Nonpolar nucleoside isosteres1 have proven to be useful
tools in probing the active sites of DNA polymerase
enzymes and DNA repair enzymes.2 These shape-
mimicking molecules lack hydrogen bonding ability, and
so, by comparison with their natural counterparts, act
as probes of electrostatic effects in the absence of large
steric differences. To date we and others have reported
on properties of an isostere for thymidine (dF) in which
difluorotoluene replaces the thymine nucleobase.3 In
addition, we have described the structures and properties
of nonpolar isosteres for deoxyadenosine, in which 4-
To date there has been no report of a proposed shape
mimic for deoxyguanosine. Such a molecule has a shape
different from that of the previous isosteres, as a result
of distinct fluorine and methyl substitution. This differ-
ence is expected to be useful, since variations in shapes
of nucleobases are thought to have large effects on
polymerase efficiency and fidelity.6 In addition, a shape
mimic for dG would be very beneficial in probing mech-
anisms of DNA damage repair. The “wobble” mismatch
G-T appears in DNA relatively frequently as a result of
misinsertion errors by DNA polymerases, and from
† University of San Diego.
‡ Stanford University.
(3) (a) Moran, S.; Ren, R.; Rumney, S.; Kool, E. T. J . Am. Chem.
Soc. 1997, 119, 2056. (b) Moran, S.; Ren, R.; Kool, E. T. Proc. Natl.
Acad. Sci. U.S.A. 1997, 94, 10506. (c) Guckian, K. M.; Kool, E. T.
Angew. Chem., Int. Ed. 1998, 36, 2825. (d) Liu, D.; Moran, S.; Kool, E.
T. Chem. Biol. 1997, 4, 919.
(4) Morales, J . C.; Guckian, K. M.; Sheils, C. J .; Kool, E. T. J . Org.
Chem. 1998, 63, 9652.
(5) Moran, S.; Ren, R.; Sheils, C. J .; Rumney, S.; Kool, E. T. Nucleic
Acids Res. 1996, 24, 2044.
(6) (a) Kunkel, T. A.; Bebenek, K. Annu. Rev. Biochem. 2000, 69,
497. (b) Goodman M. F.; Fygenson K. D. Genetics 1998, 148, 1475. (c)
Kool, E. T. Biopolymers 1998, 48, 3-17.
(1) Schweitzer, B. A.; Kool, E. T. J . Org. Chem. 1994, 59, 7238.
(2) (a) Morales, J . C.; Kool, E. T. Biochemistry 2000, 39, 2626. (b)
Morales, J . C.; Kool, E. T. J . Am. Chem. Soc. 2000, 122, 1001. (c) Kool,
E. T. Annu. Rev. Biophys. Biomol. Struct. 2001, 30, 1. (d) Morales, J .
C.; Kool, E. T. Biochemistry 2000, 39, 12979. (e) Sun, L.; Wang, M.;
Kool, E. T.; Taylor, J . S. Biochemistry 2000, 39, 14603. (f) Schofield,
M. J .; Brownewell, F. E.; Nayak, S.; Du, C.; Kool, E. T.; Hsieh, P. J .
Biol. Chem. 2001, 276, 45505. (g) Drotschmann, K.; Yang, W.;
Brownewell, F. E.; Kool, E. T.; Kunkel, T. A. J . Biol. Chem. 2001, 276,
46625. (h) Sun L. P.; Zhang, K. J .; Zhou, L.; Yuan, F. H.; Kool, E. T.;
Wang, Z. G.; Taylor, J . S. A. Biochemistry 2001, 40, 177.
10.1021/jo025884e CCC: $22.00 © 2002 American Chemical Society
Published on Web 07/09/2002
J . Org. Chem. 2002, 67, 5869-5875
5869

O’Neill et al.
F IGURE 1. The solid-state structure of nucleoside mimic dH (8) compared to that of deoxyguanosine (dG). Bond lengths are
given in angstroms.
SCHEME 1a
deamination of 5-methylcytosine.7 For the majority of
DNA polymerases, the G-T mismatch is the most com-
mon error made. Because of the high frequency of G-T
mismatches in cells, there are several known repair
enzymes that recognize and correct these errors in
prokaryotic and eukaryotic systems.7
a
Conditions: (a) Ac2O, CHCl3 (94%); (b) HNO3 (28%); (c) KOH/
MeOH (86%); (d) H2, Pd/C (95%); (e) formic acid, heat (67%).
The “fluorine up” and “fluorine down” isomers were
initially distinguished by 2D 1H-1H ROESY experiments,
and the faster moving isomer (by column chromatogra-
phy) was assigned as the desired “fluorine up” isomer 8.
The assignment was later confirmed by X-ray crystal-
lography (see below). Full details of synthesis and
characterization are given in the Supporting Information.
Nucleoside dH (8) was converted to the standard 5′-
O- dimethoxytrityl-3′-O-cyanoethylphosphoramidite de-
rivative (10) in two additional steps in preparation for
automated DNA synthesis (Scheme 2). Incorporation into
oligodeoxynucleotides was carried out with standard
coupling cycles, and stepwise yields for coupling of this
compound were >98% by trityl monitoring. Its incorpora-
tion into DNA was confirmed both by NMR spectroscopy
of short oligonucleotides containing dH and by MALDI-
TOF mass spectrometry (see Supporting Information).
Solid -Sta te Str u ctu r e. The crystal structure of dH
and the known crystal structure of dG9 are shown in
Figure 1. The nucleobase structures are quite similar;
the only significant deviations occur where the nitrogen-
to-carbon replacements were made. This affects the bond
lengths of C6-C7 (a difference of 0.05 Å) and C7-C8
(0.06 Å). The exocyclic substituents also show reasonable
mimicry in bond lengths, with the C-methyl bond length
in dH 0.11 Å longer than that of C-NH2 in dG, and the
C-F bond in dH within 0.01 Å of the corresponding CdO
bond in dG.
Here we describe the preparation and properties of the
new isostere dH (8), a highly effective shape mimic for
dG. We find that it stacks surprisingly strongly in DNA,
and it selectively pairs with dF, resulting in a fully
hydrophobic, non-H-bonded pair (H-F) that mimics the
natural G-T mismatch quite closely.
Resu lts
Syn th esis. Starting with commercial 2-fluoro-4-meth-
ylaniline, the nucleobase mimic 4-fluoro-6-methyl-1H-
benzimidazole (5) was synthesized in five steps as
detailed in Scheme 1. Benzimidazole 5 was then coupled
with 1-chloro-2-deoxy-3,5-di-O-p-toluoyl-R-D-erythro-pento-
furanose in the presence of sodium hydride.8 The two
resulting isomers (“fluorine up” and “fluorine down”) were
then deprotected with sodium methoxide and separated
by column chromatography to yield 5-(4-fluoro-6-methyl-
benzimidazol-1-yl)-â-D-deoxyriboside 8 (abbreviated dH).
Although the bond lengths and base shapes in the two
structures are quite close, there are differences in the
(7) (a) Hsieh, P. Mutat. Res. 2001, 486, 71. (b) Hardeland, U.;
Bentele, M.; Lettieri, T.; Steinacher, R.; J iricny, J .; Schar, P. Prog.
Nucleic Acid Res. Mol. Biol. 2001, 68, 235.
(9) Institute for Biomedical Computing, Washington Univer-
sity, 1994 websites: http://www.ibc.wustl.edu/moirai/klotho/html/
deoxyguanosine.html, http://www.ibc.wustl.edu/moirai/klotho/html/
guanosine.html.
(8) Hoffer, M. Chem. Ber. 1960, 93, 2777.
5870 J . Org. Chem., Vol. 67, No. 17, 2002

A Highly Effective Nonpolar Isostere of Deoxyguanosine
SCHEME 2a
a
Conditions: (a) NaH, CH3CN (46%, “fluorine up”); (b) NaOMe, MeOH (65%); (c) DMT-Cl, pyridine, DMAP (98%); (d) cyanoethylphos-
phonamidic chloride, DIPEA (80%).
TABLE 1. 1H NMR Da ta (D2O) for Nu cleosid e 8 (d H)
a n d Deoxygu a n osin e
chemical shift
H1′
H2′
H2′′
H3′
H4′
H5′
nucleoside 8 (ppm) 6.36
2.75
2.73
2.50
2.46
4.56
4.56
4.05
4.06
3.70
3.73
deoxyguanosine
6.24
nucleosides are essentially identical, and are classified
as 70% S for dG and 66% S for dH.
1
Relative intensities of H-1H cross-peaks, as deter-
mined by 2D 1H-1H ROESY experiments, were examined
for both dH and dG in D2O. While there are small
differences in orientation, the data indicate that both
nucleosides adopt an anti conformation about the glyco-
sidic bond. For dH, a strong H2-H2′ nuclear Overhauser
effect, a weak H2-H2′′ effect, and a medium H7-H2′
effect were detected.
Sta ck in g a n d P a ir in g Stu d ies in DNA. The stack-
ing properties of dH were measured in the context of
duplex DNA by the dangling end method. For these
preliminary studies, one standard10 sequence context was
chosen, with the dangling dH residue adjacent to C in
the self-complementary duplex. Table 2 shows the data
in comparison to literature values for other related
compounds.11
F IGURE 2. Space-filling models of the natural nucleobase
guanine and the nonpolar mimic H (5). Electrostatic potentials
are mapped on the surfaces. Red denotes negative potential
and blue positive potential.
orientation of the base with respect to the furanose ring.
The natural product dGadopts a typically anti conforma-
tion with a torsional angle (ø) of -122°. The base of dH
is twisted by 58° relative to this, falling between typical
anti and syn ranges, and displaying a -sc (high anti)
glycosidic conformation with a measured torsional angle
(ø) of -65°. The two nucleosides adopt different sugar
conformations in the solid state. Analogue dH has a C2′-
endo (south) sugar pucker and dG has an O4′-endo
conformation.
The data show strong stabilization by the nucleobase
mimic H in this context, with stacking considerably
stronger than that of G, and close to that of the most
strongly stacking molecules (pyrene, 5-nitroindole) in a
recent study.11
Space-filling models of the nucleobase G and mimic H
show very close similarity in shape (Figure 2). Once
again, the single noticeable difference is at the proton at
C-7 in H (compared to N3 of G). Comparison of calculated
electrostatic potentials (AM1) on these molecules con-
firms the highly polar nature of G, and shows by contrast
that H is expected to be quite nonpolar, except at the
nitrogen in the imidazole ring (analogous to N7 of
guanine).
Pairing of base mimic H opposite natural bases and
opposite another nonpolar base mimic was evaluated in
a 12 base pair sequence context. The analogue H is quite
destabilizing when paired opposite natural nucleobases
(Table 3), and shows little if any selectivity between
them. However, when paired opposite the nonpolar
thymine mimic difluorotoluene (F) it regains some stabil-
ity, and the overall duplex with the H-F pair approaches
Con for m a tion in Solu tion . There are even greater
structural similarities between the two nucleosides in
D2O solution as indicated by the nuclear magnetic
resonance data (Table 1). Sugar conformations of both
(10) Petersheim, M.; Turner, D. H. Biochemistry 1983, 22, 256-
263.
(11) Guckian, K. M.; Ren, R. X.-F.; Chaudhuri, N. C.; Tahmassebi,
D. C.; Kool, E. T. J . Am. Chem. Soc. 2000, 122, 2213.
J . Org. Chem, Vol. 67, No. 17, 2002 5871

O’Neill et al.
TABLE 2. F r ee En er gy of Sta ck in g for Na tu r a l Nu cleosid es a n d Rela ted An a logs, As Mea su r ed by Da n glin g En d
Th er m a l Den a tu r a tion Stu d ies w ith Self-Com p lem en ta r y Str a n d s (d XCGCGCG)a
-∆H° (kcal)
(van’t Hoff)
∆S° (eu)
(van’t Hoff)
-∆G°37 (eu)
(van’t Hoff)
-∆G°37 (kcal)
∆∆G°
stacking
dangling residue
T
m (°C)b
(fits)c
none (core duplex)e
thyminee
41.7
48.1
51.6
46.2
51.5
55.7
54.6
60.6
46
48
55
50
43
70
67
54
120
130
140
130
110
190
180
140
8.1 ( 0.2
9.2 ( 0.2
10.1 ( 0.2
9.1 ( 0.2
9.4 ( 0.2
11.6 ( 0.2
11.2 ( .2
11.4 ( .2
8.1 ( 0.1
9.2 ( 0.2
10.0 ( 0.4
8.9 ( 0.1
9.9 ( 0.3
10.7 ( .05
10.5 ( .1
11.6 ( .3
1.1 ( 0.2
2.0 ( 0.2
1.0 ( 0.2
1.3 ( 0.2
3.5 ( 0.5
3.1 ( .3
adeninee
cytosinee
guaninee
4-F-6-Me-benzimidazole (H)
4-methylindolee
5-nitroindolee
3.4 ( .3
a
Free energy of stacking (∆∆G°) is obtained by substracting the free energies of the duplexes with two dangling residues from the
b
energy of the core hexamer duplex. Conditions: 1 M NaCl, 10 mM Na‚phosphate, pH 7.0; 5.0 µM DNA strand concentration for Tm
value shown. c Average free energies from fits to individual melting curves. Concentration 6 µM. e Data from ref 11.
d
with the deoxyguanosine in the syn conformation.14 We
surmise (i) that rotation about this bond is relatively
TABLE 3. Ba se P a ir in g Da ta for Gu a n in e An a log H
Op p osite Na tu r a l Ba ses a n d Op p osite Th ym in e An a log F
unhindered for dH (as with dG) and (ii) that crystal
packing effects have influenced the dH conformation. A
similar effect was seen for a nonpolar isostere of deoxy-
adenosine in the solid state,4 and later structural studies
in the context of DNA5 revealed it to adopt the anti
orientation, like its natural counterpart dA.
The crystal structure of dH is useful in pointing out
the extent of shape mimicry by the base mimic 4-fluoro-
6-methylbenzimidazole. Comparison with the guanine
heterocycle in the dG structure shows very similar
structures. The one most substantial structural difference
between dG and dH is the presence of the proton on dH
a
at the 7 position (see Figure 2); this is missing in dG,
where a ring nitrogen occurs. This difference of ca. 0.5 Å
in steric bulk should be considered in enzyme active sites
where close minor groove contacts might be made. In
addition, the lack of a hydrogen bond acceptor at this
position may have significant effects on DNA stability
and on activity in DNA replication.15
Conditions: 100 mM NaCl, 10 mM Na‚PIPES, pH 7.0; 10 mM
b
MgClL2, 5.0 µM DNA strand concentration for Tm. Average free
energies from fits to individual melting curves. c Data from ref 17.
that of the duplex containing the analogous G-T “wobble”
pair.
Str on g Ba se Sta ck in g. The base-stacking studies
show that the nucleobase mimic H is quite effective at
stacking and stabilizing the double helix, at least in the
one sequence context studied. Further studies in other
contexts will be needed to test the generality of this. We
find that H stacks considerably more strongly in this
context than all four DNA bases, and surpasses guanine
by a large 1.1 kcal/mol per residue. Since calculated
polarizability (data not shown) and the sizes of G and H
are essentially the same, we hypothesize that this dif-
ference is due to enhanced hydrophobic effects for H
rather than to differences in dispersion forces. Compari-
son of this molecule with other nonpolar nucleoside
replacements11,16 shows that H stacks somewhat more
strongly than 4-methylindole, and essentially the same
as 5-nitroindole and even the tetracyclic hydrocarbon
pyrene. Some of the enhanced stacking may be due to
increased surface area, increased hydrophobicity of fluo-
rine, and/or altered electrostatics. Work is underway to
test some of these effects more systematically.
Discu ssion
Str u ctu r e Mim icr y by d H. The NMR data show that
the new nucleoside dH is an effective shape mimic of
deoxyguanosine in solution. Both adopt the “S” class of
sugar conformation in water, and are virtually identical
in this conformation, as judged by coupling constants in
the sugar protons.12 In addition, both the natural nucleo-
base and the substituted benzimidazole base mimic
appear to adopt predominantly anti glycosidic orienta-
tions.
It is not surprising that in the crystalline state the new
nucleoside dH adopts a glycosidic orientation twisted
significantly from that in the published structure of dG.
Purines are known to adopt both syn and anti conforma-
tions under varying conditions, both in the crystalline
form and in solution. Since rotation about the glycosidic
bond is relatively unhindered in purines, a wide variety
of torsional angles have been reported in the X-ray
structures of deoxyguanine, deoxyguanosine, and gua-
nosine monophosphate.9,13 For example, a 1:1 complex of
deoxyguanosine and 5-bromodeoxycytidine crystallizes
Ba se P a ir in g. The results show that nonpolar isostere
dH pairs almost equally with all four natural bases. It is
(14) Haschemeyer, A. E. V.; Sobell, H. M. Acta Crystallogr. 1965,
19, 125.
(15) (a) Morales, J . C.; Kool, E. T. J . Am. Chem. Soc. 1999, 121,
2723. (b) Spratt, T. E. Biochemistry 2001, 40, 2647.
(16) Guckian, K. M.; Schweitzer, B. A.; Ren, R. X.; Sheils, C. J .;
Paris, P. L.; Tahmassebi, D. T.; Kool, E. T. J . Am. Chem. Soc. 1996,
118, 8182.
(12) Rinkel L. J .; Altona C. J . Biomol. Struct. Dyn. 1987, 4, 621.
(13) (a) Bloomfield, V. A.; Crothers, D. M.; Tinoco, I., J r. Nucleic
Acids Structure, Properties, and Function; University Science Books:
2000; Chapter 2. (b) Thewalt, U.; Bugg, C. E.; Marsh, R. E. Acta
Crystallogr., Sect. B 1970, 26, 1089. (c) Son, T. D.; Gueron, M. J . Am.
Chem. Soc. 1972, 94, 7903.
5872 J . Org. Chem., Vol. 67, No. 17, 2002

A Highly Effective Nonpolar Isostere of Deoxyguanosine
quite substantially destabilizing relative to a standard
A-T base pair. The H-X pairs are all approximately as
destabilizing as a mismatch in this context.17 We at-
tribute this in part to the lack of hydrogen bonds and,
perhaps more importantly, to the energetic cost of de-
solvation of the polar partners paired across from H.
Virtually the same results have been observed previously
for the other nonpolar isosteres dF and dZ.4,18 We note
that nonpolar, non-hydrogen bonding nucleosides are
among the most effective candidates for “universal” DNA
bases that pair equally with all four bases.19
Interestingly and notably, the data show that pairing
of dH with another nonpolar nucleoside (dF) results in
greater stability for the duplex, relative to the pairs of
H with polar bases. The stability of the H-F pair is ∼1
kcal/mol more stable than the H-T pair, for example.
Thus the H-F pair is selective for pairing in the context
of the four natural pairs. We attribute this preference of
nonpolar base for another one to removal of the desol-
vation cost that is incurred when one partner is polar,
and perhaps also to more favorable hydrophobic effects.17
The Watson-Crick pairing faces of dH and dF are
essentially identical in shape and size to those of the
natural nucleosides dG and dT. Since H-F is potentially
a close shape analogue of the G-T “wobble” mismatch,
it is interesting to compare H-F to G-T in pairing
stability. The results here show that G-T is more stable
than H-F by a relatively small amount (∼0.6 kcal). We
hypothesize that the reason for the lower stability for
H-F, despite the strong stacking of F and H, may by
due either to a lack of hydrogen bonds or to disruption
of minor groove solvation.20
Facility of the Department of Chemistry and Biochem-
istry at the University of California, San Diego.
X-r a y Cr ysta llogr a p h ic Meth od s. Single crystals of
nucleoside 8 were grown from a concentrated ethanol
solution. A clear crystal of dimensions 0.20 × 0.13 × 0.10
mm3 was mounted on a glass fiber with epoxy. The X-ray
intensity data were collected at 100(2) K, at a wavelength
of 0.71073, with a θ range of 1.58-27.50. Frames were
integrated to yield a total of 10252 reflections, of which
2722 were independent (Rint ) 2.98%). The crystal
belongs to the orthorhombic crystal system. The unit cell
parameters were a ) 5.8291(6) Å, b ) 7.9860(8) Å, c )
25.797(3) Å, and V ) 1200.9(2) Å3. The space group was
assigned as P2(1)2(1)2(1). The structure was solved with
direct methods included in the SHELXTL program pack-
age and refined by full-matrix least-squares on F2. For a
Z value of 4, there is one molecule in the asymmetric unit.
All non-hydrogen atoms were refined anisotropically with
hydrogens included in idealized positions giving a data:
parameter ratio of approximately 15:1. The largest peak
in the final difference map was 0.321 e/Å3. The structure
refined to a goodness of fit (GOF) on F2 of 0.915 and final
residuals of R1 ) 3.42% and wR2 ) 8.60%.
N-(2-F lu or o-4-m eth ylp h en yl)a ceta m id e (1). A so-
lution of 2-fluoro-4-methylaniline (4.90 g, 39.2 mmol)
diluted in 50 mL of chloroform was stirred and cooled to
0 °C. To that solution was added dropwise acetic anhy-
dride (4.0 g, 39.2 mmol) diluted in 100 mL of chloform
over 10 minutes. The mixture was stirred for an ad-
ditional 30 min. Water was added, the solution was
neutralized with a saturated solution of Na2CO3, and the
product was extracted into the organic layer. The com-
bined organic layers were dried with magnesium sulfate
and concentrated to yield 6.27 g (94%) of 1 as a white
solid: 1H NMR (CDCl3) δ 8.11 (1H, dd, J ) 8.1, 8.7 Hz),
7.31 (1H, NH), 6.91 (2H, m), 2.31 (3H, s), 2.20 (3H, s).
13C NMR (CDCl3) δ 168.50, 150.92, 125.13, 122.02,
115.60, 115.36, 24.72, 21.08. HRMS calcd for C9H10FNO
M + H m/e 168.0825, found 168.0822.
We anticipate that analogue dH will be useful in
studies of DNA replication and repair. For example, a
recently discovered error-prone human polymerase, Pol
iota, is reported to synthesize G-T mismatches with
exceptionally high frequency.21 The analogue H could be
useful in probing steric and hydrogen bonding effects in
the active sites of such error-prone enzymes. In addition,
G-T mismatches, for reasons described above, occur
commonly in cells. The analogue H, and the analogous
H-F pair, may well be useful in probing mechanisms of
repair enzymes that correct these mutagenic mis-
matches.7
2-F lu or o-4-m eth yl-6-n itr oa n ilin e (3). The protected
aniline derivative (1) (6.27 g, 37.5 mmol) was dissolved
in ice cold 70% HNO3 (70 mL). Maintaining the temper-
ature at 0 °C, 90% HNO3 (60 mL) was added dropwise
over 1 h. The solution was stirred for one additional hour
then poured over ice to form a yellow precipitate. The
solid was collected by vacuum filtration, washed with
water until the filtrate was neutral, and dried to yield
6.66 g (83.8%) of a mixture of nitrated product 2 (N-
acetamido-2-fluoro-4-methyl-6-nitroaniline) and the 5-ni-
tro isomer as a pale yellow solid. The mixture was
dissolved in 100 mL of methanol and stirred. Solid KOH
was added until the pH was 10. While the reaction was
stirred for several hours, solid KOH was added as needed
to maintain the pH at 10. The methanol was removed
by evaporation under reduced pressure and the depro-
tected product 3 was extracted with methylene chloride.
The combined organic layers were dried with magnesium
sulfate. The crude product was concentrated and purified
by silica column chromatography (CHCl3-EtOAc, 5:1) to
obtain 1.50 g (24%) of 3 as a yellow/orange oil: 1H NMR
(DMSO) δ 7.66 (1H, s), 7.34 (1H, d, J ) 11.7 Hz), 2.22
(3H, s). 13C NMR (CDCl3) δ 20.40, 124.79, 124.88, 132.95,
133.19, 150.24, 153.45. HRMS calcd for C7H7FN2O2 M +
Exp er im en ta l Section
Gen er a l Rem a r k s. 1H, 13C, and ROESY NMR spectra
were obtained on a 300 MHz instrument. Column chro-
matography was performed with Merck silica gel 60
(230-400 mesh). 2-Fluoro-4-methylaniline was pur-
chased from a commercial source. Anhydrous dimethyl-
formamide and acetonitrile were purchased from a
commercial source in Sure-seal bottles. X-ray crystal-
lography was performed at the Small Molecule X-ray
(17) Kool, E. T.; Morales, J . C.; Guckian, K. M. Angew. Chem., Int.
Ed. 2000, 39, 990.
(18) Morales, J . C.; Kool, E. T. Nature Struct. Biol. 1998, 5, 950.
(19) (a) Nichols, R.; Andrews, P. C.; Zhang, P.; Bergstrom, D. E.
Nature 1994, 369, 492. (b) Loakes, D.; Brown, D. M. Nucleic Acids Res.
1994, 22, 4039. (c) Matray, T. J .; Kool, E. T. J . Am. Chem. Soc. 1998,
120, 6191. (d) Berger, M.; Wu, Y. Q.; Ogawa A. K.; McMinn, D. L.;
Schultz, P. G.; Romesberg, F. E. Nucleic Acids Res. 2001, 28, 2911.
(20) Lan T.; McLaughlin, L. W. J . Am. Chem. Soc. 2000, 122, 6512.
(21) Zhang, Y.; Yuan, F.; Wu, X.; Wang, Z. Mol. Cell. Biol. 2000,
20, 7099.
+
NH4 m/e 188.0835, found 188.0840.
J . Org. Chem, Vol. 67, No. 17, 2002 5873

O’Neill et al.
1-F lu or o-5-m et h yl-2,3-d ia m in ob en zen e (4). Ap-
proximately 25 mg of Pd(C) (5 wt %) was added to 3 (2.24
g, 13.2 mmol) dissolved in 5 mL of anhydrous DMF. The
mixture was hydrogenated at 52 psi overnight. The
purple-colored mixture was filtered over Celite and
concentrated to yield 1.75 g (95%) of 6 as a blue/purple
oil: 1H NMR (CDCl3) δ 6.36 (1H, d, J ) 10.5 Hz), 6.31
(1H, s), 2.19 (3H, s). 13C NMR (CDCl3) δ 20.74, 112.16,
119.06, 129.55, 137.18, 151.46, 154.57. HRMS calcd for
C7H9FN2 M + H m/e 141.0828, found 141.0831.
4-F lu or o-6-m eth yl-1H-ben zim id a zole (5). Formic
acid (1.6 mL, 88%) was added to 1.75 g (10.7 mmol) of
diamine 4 dissolved in 20 mL of methanol. The mixture
was heated to reflux. During the reflux the pH was
maintained at approximately 7 by the addition of 88%
formic acid. After approximately 48 h of reflux, the
reaction mixture was cooled then concentrated to remove
the methanol. Water and ethyl acetate were added, and
the product was extracted into the organic layer. The
combined organic layers were dried, then concentrated.
The crude material (1.58 g, 10.53 mmol) was purified by
silica column chromatography (CHCl3:EtOAc, 3:1). Benz-
imidazole 5 was obtained as a pale yellow oil, 1.25 g
(67%): 1H NMR (CDCl3) δ 8.02 (1H, s), 7.19 (1H, s), 6.82
(1H, d, J ) 11.1 Hz), 2.47 (3H, s). 13C NMR (CDCl3) δ
155.19, 151.94, 142.71, 135.47, 110.21, 109.97, 21.80. GC/
MS calcd for C8H7N2F M m/e 150.15, found 150.
tracted. The combined organic layers were dried, concen-
trated, and purified with use of silica column chroma-
tography to yield 415 mg (65%) of nucleoside 8 as a clear
1
oil. H NMR (D2O) δ 8.26 (1H, s), 7.25 (1H, s), 6.92 (1H,
d, J ) 12.3 Hz), 6.36 (1H, dd, J ) 3.6, 6.6 Hz), 4.56 (1H,
m), 4.05 (1H, m), 3.70 (2H, m), 2.75 (1H, m), 2.50 (1H,
m), 2.41 (3H, s). 13C NMR (D2O) δ 152.97, 152.94, 121.09,
120.87, 118.31, 107.23, 97.84, 96.07, 81.92, 72.44, 49.66,
31.94, 26.68. HRMS (DCI) calcd for C13H15O3N2F (M +
H) 267.1145, found 267.1150.
1′,2′-Did eoxy-5′-(4,4′-d im eth oxitr ityl)-3′-h yd r oxy-
1′-(4-flu or o-6-m et h yl-1H -b en zim id a zole)-â-D-r ib o-
fu r a n ose (9). The deprotected nucleoside (8), 200 mg,
0.751 mmol) was coevaporated with dry pyridine (2 × 5
mL) and dissolved in pyridine (9 mL). To the above
mixture was added diisopropylethylamine (315 µL) in one
portion. A solution of 4,4′-dimethoxitrityl (DMT) chloride
(382 mg, 1.12 mmol) in pyridine (5 mL) was added slowly
over 40 min. The mixture was stirred for 22 h and
quenched by adding methanol (20 mL). The mixture was
concentrated and purified by flash column chromatog-
raphy (hexanes-ethyl acetate-triethylamine, 1:2:0.1) to
obtain 438 mg (98%) of the 5′-DMT ether derivative as a
1
yellow foam. H NMR (CDCl3) δ 7.95 (1H, s), 7.40-7.15
(10 H, m), 6.80 (1H, d, J ) 12 Hz), 6.77 (4H, d, J ) 9.0
Hz), 6.25 (1H, t, J ) 7.2 Hz), 4.80 (1H, m), 4.68 (1H, br
s), 4.24 (1H, m), 3.77 (6H, s), 3.35 (2H, m), 2.70-2.38
(2H, m), 2.30 (3H, s); 13C NMR (CDCl3) δ 158.28, 154.71,
151.37, 144.34, 140.13, 135.66, 135.54, 135.46, 134.53,
134.44, 130.19, 129.96, 129.77, 127.99, 127.54, 126.59,
112.91, 109.58, 109.35, 106.89, 86.34, 86.20, 85.21, 71.42,
63.68, 60.15, 54.84, 40.22, 21.33. HRMS (DCI) calcd for
C34H33O5N2F (M + 1) 569.2452, found 569.2452.
1′,2′-Did eoxy-3′,5′-d i-O-tolu oyl-1′-(4-flu or o-6-m eth -
yl-1H-ben zim id a zole)-â-D-r ibofu r a n ose (7). Benzim-
idazole 5 (1.25 g, 8.33 mmol) was dissolved in dry CH3CN
(100 mL). Sodium hydride (60% suspension in mineral
oil) (367 mg, 9.2 mmol) was added and the mixture was
stirred in a nitrogen atmosphere for 1 h at room tem-
perature while hydrogen gas evolved. The mixture was
cooled to 0 °C, then 3.28 g of 1′-R-chloro-3′,5′-di-O-toluoyl-
2′-deoxyribose9 (8.33 mmol) was added in one portion.
The mixture was stirred and slowly warmed back to room
temperature. The reaction was stirred overnight then
quenched by the addition of a saturated solution of
NaHCO3. The reaction mixture was filtered and concen-
trated, and the product was extracted with ethyl acetate.
The combined organic layers were dried and concentrated
to recover the mixture of isomers (6 and 7) as a brown
oil. The isomers were separated and purified by silica
column chromatography (hexane:EtOAc, 1:1). Compound
1′,2′-Did eoxy-5′-(4,4′-d im eth oxitr ityl)-1′-(4-flu or o-
6-m eth yl-1H-ben zim id a zole)-â-D-r ibofu r a n ose-3′-O-
cya n oeth yl-N,N-d iisop r op ylp h osp h or a m id a te (10).
The 5′-O-tritylated compound (9) (100 mg, 0.176 mmol)
was dissolved in dry dichloromethane (2.5 mL), and to
this were added diisopropylethylamine (0.170 mL, 0.66
mmol) and 2-cyanoethyl N,N-diisopropylchlorophos-
phoramidite (0.08 mL, 0.35 mmol). The reaction mixture
was stirred at room temperature for 1 h and 45 min.
Hexanes (5 mL) was added and the concentrated mixture
was purified by flash column chromatography (hexanes-
ethyl acetate-triethylamine, 3:2:0.1). The two isomers
of 10 were obtained as oils, DMT phosphoramidite A (52
mg, 38%), and DMT phosphoramidite B (56 mg, 42%).
Spectroscopic data for DMT phosphoramidite A: 1H NMR
(CDCl3) δ 7.98 (1H, s), 7.42-7.13 (10H, m), 6.85-6.72
(5H, m), 6.26 (1H, m), 4.81-4.70 (1H, m), 4.33 (1H, m),
3.90-3.59 (3H, m), 3.80 (6H, s), 3.43-3.25 (2H, m), 2.80-
2.25 (5H, m), 2.30 (3H, s), 1.26-1.12 (12H, m). 13C NMR
(CDCl3) δ 158.48, 144.27, 140.13, 135.72, 135.43, 134.46,
134.37, 129.96, 128.09, 127.71, 126.81, 117.16, 113.04,
109.57, 109.34, 106.63, 106.60, 86.40, 85.59, 85.55, 85.18,
73.55, 73.32, 63.16, 58.37, 58.11, 55.07, 43.31, 43.15,
39.44, 39.40, 24.49, 24.40, 21.45, 20.12, 20.03. HRMS
(FAB, 3-NBA matrix) calcd for C43H50O6N4FP (M + 1)
768.3530, found 768.3546. Spectroscopic data for DMT
phosphoramidite B: 1H NMR (CDCl3) δ 7.99 (1H, s),
7.42-7.15 (10H, m), 6.85-6.72 (5H, m), 6.28 (1H, m), 4.70
(1H, m), 4.31 (1H, m), 3.90-3.52 (3H, m), 3.80 (6H, s),
3.42-3.27 (2H, m), 2.80-2.59 (5H, m), 2.33 (3H, s), 1.22-
1.12 (12H, m). 13C NMR (CDCl3) δ 158.52, 144.30, 140.15,
1
7 was recovered as a clear oil (1.93 g, 46%). H NMR
(DMSO) δ 8.53(1H, s), 7.80 (2H, d, J ) 7.8 Hz), 7.83 (2H,
d, J ) 7.8 Hz), 7.35 (1H, s), 7.34 (4H, m), 7.01 (1H, d,
J ) 12.6 Hz), 6.61 (1H, dd, J ) 6.9, 7.2 Hz), 5.70 (1H,
m), 4.53 (3H, m), 3.09 (1H, m), 2.81 (1H, m), 2.41 (3H,
s), 2.38 (3H, s), 2.84 (3H, s). 13C NMR (CDCl3) δ 166.06,
165.87, 155.35, 152.00, 144.53, 144.18, 140.00, 135.65,
135.53, 134.79, 134.70, 129.93, 129.74, 129.55, 129.27,
126.60, 126.37, 109.87, 109.64, 106.41, 85.48, 82.50,
74.69, 63.90, 38.06, 21.65, 21.57. HRMS calcd for
C29H27O5N2F M + H m/e 503.1982, found 503.1990.
1′,2′-Did eoxy-3′,5′d ih yd r oxy-1′-(4-flu or o-6-m eth yl-
1H-ben zim id a zole)-â-D-r ibofu r a n ose (8). To a solu-
tion of 7 bis-toluoyl ester (1.18 g, 3.85 mmol) in absolute
ethanol (20 mL) was added NaOMe (in methanol, 25%,
0.5 mL). The reaction mixture was stirred for 3 h. Solid
ammonium chloride was added to quench the reaction.
The mixture was filtered then concentrated. Ethyl ac-
etate and water were added and the product was ex-
5874 J . Org. Chem., Vol. 67, No. 17, 2002

A Highly Effective Nonpolar Isostere of Deoxyguanosine
135.70, 135.45, 134.50, 129.97, 128.10, 127.72, 126.80,
117.18, 113.07, 109.60, 109.32, 106.65, 106.61, 87.01,
85.80, 85.72, 85.55, 74.69, 74.47, 63.84, 58.81, 58.55,
55.29, 44.06, 43.93, 39.84, 39.81, 24.63, 24.54, 21.52,
20.50, 20.41. HRMS (FAB, 3-NBA matrix) calcd for
C43H50O6N4FP(M + 1) 768.3530, found 768.3501.
edges the University of San Diego Faculty Research
Grants for support and Dr. Tammy Dwyer for helpful
discussions.
Su p p or tin g In for m a tion Ava ila ble: Proton NMR spectra
for all compounds and tables of supporting data and an ORTEP
diagram for the X-ray structure of compound 8. This material
is available free of charge via the Internet at http://pubs.acs.org.
Ack n ow led gm en t. We thank the National Insti-
tutes of Health (GM52956) for support. D.T. acknowl-
J O025884E
J . Org. Chem, Vol. 67, No. 17, 2002 5875