Toward a New Genetic System with Expanded Dimensions: Size-Expanded Analogues of Deoxyadenosine and Thymidine

Article abstract of DOI:10.1021/ja038384r

We describe the design, preparation, and properties of two key building blocks of a size-expanded genetic system. Nucleoside analogues of the natural nucleosides dA and dT are reported in which the fusion of a benzo ring increases their size by ca. 2.4 A.

Full text of DOI:10.1021/ja038384r

Published on Web 01/13/2004
Toward a New Genetic System with Expanded Dimensions:
Size-Expanded Analogues of Deoxyadenosine and Thymidine
Haibo Liu, Jianmin Gao, Lystranne Maynard, Y. David Saito, and Eric T. Kool*
Contribution from the Department of Chemistry, Stanford UniVersity,
Stanford, California 94305-5080
Received September 7, 2003; E-mail: kool@stanford.edu.
Abstract: We describe the design, preparation, and properties of two key building blocks of a size-expanded
genetic system. Nucleoside analogues of the natural nucleosides dA and dT are reported in which the
fusion of a benzo ring increases their size by ca. 2.4 Å. The expanded dA analogue (dxA), having a tricyclic
base, was first reported by Leonard nearly three decades ago. We describe a shortened and more efficient
approach to this compound. The expanded dT analogue (dxT), a methylquinazolinedione C-glycoside, was
previously unknown; we describe its preparation in eight steps from 5-methylanthranilic acid. The key
glycoside bond formation employed Pd-mediated coupling of an aryl iodide precursor with a dihydrofuran
derivative of deoxyribose. Both nucleosides are shown to be efficient fluorophores, emitting light in the
blue-violet range. The base-protected phosphoramidite derivatives were prepared, and short oligonucleotides
containing them were characterized. The two size-expanded nucleosides are key components of a new
four-base genetic system designed to form helical paired structures having a diameter greater than that of
natural DNA. Elements of the design of this expanded genetic molecule, termed xDNA, are discussed,
including the possibility of up to eight base pairs of information storage capability.
Introduction
natural nucleobases so that the molecules can recognize DNA
or RNA of the natural genetic system.
The design and synthesis of modified versions of the DNA
double helix has received much attention in the past two
decades, primarily boosted by the development of automated
solid-phase chemistry for the assembly of oligonucleotides.1 The
motivation behind such work has been multifold: first is the
basic scientific exploration of how the natural helix assembles
and functions biochemically; second is the possibility that
modified oligonucleotide analogues could have biomedical
applications in the detection and treatment of disease; third is
the possibility that chemists might expand the natural genetic
alphabet by providing new base pairs; and finally, there is the
exploration of possible primordial precursors of life on earth.
In addition to work with modified backbones, the nucleobases
themselves have recently been the targets of modification, often
with the goal of designing new base pairs. Early work by Benner
proposed a number of new hydrogen bonded pairs that might
provide a larger genetic alphabet for hybridization and for
encoding genetic information.7 Work in other laboratories has
explored not only hydrogen-bonded pairs,8 but also nonpolar
pairs9-11 and metal-mediated pairs as well.12,13 A main focus
of those approaches has been to match the geometry of the
natural double helical DNA framework, since that is believed
to be a chief factor in biochemical function of replication.14,15
Because much of the preceding work has focused on
maintaining the geometry of pairing within the natural DNA
framework, little attention has been devoted to preparation of
This work has involved chemical approaches to modification
both of the phosphodiester backbone of the DNA or of the bases
themselves. Modification of the backbone can lead to enhanced
properties in hybridization and biochemical stability. A few
prominent recent examples include the peptide nucleic acids,2
locked nucleic acids,3 phosphoramidate nucleic acids,4 and
threose nucleic acids.5,6 Those analogues have generally retained
(7) Piccirilli, J. A.; Krauch, T.; Moroney, S. E.; Benner, S. A. Nature 1990,
343, 33-37.
(8) Rappaport, H. P. Nucleic Acids Res. 1988, 16, 7253-7267.
(9) (a) Schweitzer, B. A.; Kool, E. T. J. Am. Chem. Soc. 1995, 117, 1863-
1872. (b) Matray, T. J.; Kool, E. T. J. Am. Chem. Soc. 1998, 120, 6191-
6192. (c) Matray, T. J.; Kool, E. T. Nature 1999, 399, 704-708.
(10) (a) McMinn, D. L.; Ogawa, A. K.; Wu, Y. Q.; Liu, J. Q.; Schultz, P. G.;
Romesberg, F. E. J. Am. Chem. Soc. 1999, 121, 11585-11586. (b) Tae, E.
L.; Wu, Y.; Xia, G.; Schultz, P. G.; Romesberg, F. E. J. Am. Chem. Soc.
2001, 123, 7439-7440.
(1) (a) Shimidzu, T.; Letsinger, R. L. J. Org. Chem. 1968, 33, 708-711. (b)
Ogilvie, K. K.; Kroeker, K. Can. J. Chem. 1972, 50, 1211-1215. (c)
Beaucage, S. L.; Caruthers, M. H. Tetrahedron Lett. 1981, 22, 1859-1862.
(2) Egholm, M.; Buchardt, O.; Christensen, L.; Behrens, C.; Freier, S. M.;
Driver, D. A.; Berg, R. H.; Kim, S. K.; Norden, B.; Nielsen, P. E. Nature
1993, 365, 566-568.
(3) Petersen, M.; Wengel, J. Trends Biotechnol. 2003, 21, 74-81.
(4) Gryaznov, S. M. Biochim. Biophys. Acta 1999, 1489, 131-140.
(5) Schoning, K.; Scholz, P.; Guntha, S.; Wu, X.; Krishnamurthy, R.;
Eschenmoser, A. Science 2000, 290, 1347-1351.
(6) Chaput, J. C.; Ichida, J. K.; Szostak, J. W. J. Am. Chem. Soc. 2003, 125,
856-857.
(11) Mitsui, T.; Kitamura, A.; Kimoto, M.; To, T.; Sato, A.; Hirao, I.; Yokoyama,
S. J. Am. Chem. Soc. 2003, 125, 5298-5307.
(12) Meggers, E.; Holland, P. L.; Tolman, W. B.; Romesberg, F. E.; Schultz,
P. G. J. Am. Chem. Soc. 2000, 122, 10714-10715.
(13) Tanaka, K.; Tengeiji, A.; Kato, T.; Toyama, N.; Shionoya, M. Science 2003,
299, 1212-1213.
(14) (a) Moran, S.; Ren, R. X.-F.; Rumney, S.; Kool, E. T. J. Am. Chem. Soc.
1997, 119, 2056-2057. (b) Kool, E. T. Annu. ReV. Biochem. 2002, 71,
191-219.
(15) Kunkel, T. A.; Bebenek, K. Annu. ReV. Biochem. 2000, 69, 497-529.
9
1102
J. AM. CHEM. SOC. 2004, 126, 1102-1109
10.1021/ja038384r CCC: $27.50 © 2004 American Chemical Society

A Genetic System with Expanded Size
A R T I C L E S
combined with modified enzymes that could replicate it, yielding
a new information-storing and amplifiable molecular system.
Here, we describe the preparation of the two new key
components of xDNA, namely, the size-expanded forms of
deoxyadenosine, dxA, and thymidine, dxT (Figure 1). When
paired with the natural pairing partners thymidine and deoxy-
adenosine, these compounds create the possibility of a large-
diameter four-base paired system, with information encoding
potential the same as the natural genetic system. Aspects of the
base pair designs are also described.
Results and Discussion
Design Aspects. In considering the structure of the expanded
adenine base, Leonard pointed out that, if paired with T, the
xA base would be too large for the natural DNA helix. Indeed,
he designed a possible new size-compressed pairing partner to
compensate for this extended base, although it was apparently
never successfully tested.20 We envisioned, by contrast, that
rather than adjust for the added size of xA, one might take
advantage of its geometric extension. A fully analogous
geometric extension of pyrimidines and purines (Figure 1),
paired with natural DNA bases, could potentially yield a regular
helix with expanded diameter. Modeling studies in our labora-
tory (data not shown) suggested that the natural DNA backbone
could accept this stretching of all four pairs, requiring only small
adjustments of bond angles and requiring no large change to
sugar conformations. Subsequent experimental studies have
borne this out.21 There is one significant adjustment that is likely
needed to adapt to the large cylindrical diameter of such a helix,
however. Because the outer circumference that the backbone
must traverse is larger, it is expected that there will be a greater
number of base pairs per turn. Our modeling suggested that
roughly 14 base pairs per turn might be optimum for xDNA;
this contrasts with natural B-DNA (at ca. 10.5 pairs per turn).
Our molecular design for benzo fusion of xT involves a
C-nucleoside framework (Figure 1). This design moves the point
of attachment to C-1 of deoxyribose by ca. 2.4 Å, virtually the
same length increase that occurs with insertion of benzene into
the xA base. Thus, the lengths of the two are fully analogous,
and the extensions push the hydrogen-bonding surfaces outward
toward the backbone of the opposite strand. As a result, the
major and minor groove widths in an xDNA duplex are likely
to be increased as well. The vector orientations of the two
extensions are expected to be slightly different, with the xT
base extension oriented about 12° toward the minor groove
relative to the xA extension (Figure 1B). This difference arises
from the fact that xA has a five-membered-ring glycosidic
attachment, whereas xT is attached via a six-membered ring.
This difference is relatively small, and (we expect) might be
accommodated with modest changes in bond torsional angles
in the sugars and backbones.
Figure 1. Structures of benzo-fused deoxynucleosides dxA and dxT. (A)
The free nucleosides. (B) Proposed structures of hydrogen-bonded base pairs
involving extended bases (right), with comparison to the natural analogous
pairs (left). The arrows illustrate the vectors of extension of the heterocyclic
bases upon insertion of the benzene ring into the framework. The C1′-C1′
glycosidic distances are expected to be ca. 13.0 Å for the extended pairs;
by comparison, the distance for B-form DNA is ca. 10.7 Å.
nucleotide analogues of larger size. However, in pioneering work
by Leonard nearly three decades ago, a “stretched-out” analogue
of adenosine was described in which a benzene ring was inserted
between the two rings of adenine.16 This ribonucleoside and its
triphosphate derivative were used to probe the active sites of
ATP-dependent enzymes.17,18 Later work described conversion
of the ribonucleoside to the deoxyribonucleoside derivative (here
termed dxA) as well.18 That work was far ahead of its time,
and the analogue was never incorporated into oligodeoxynucle-
otides (automated synthesizers were not yet invented). More
recently, a thienyl group has been used as a spacer in tricyclic
purine analogues screened for antiviral activity.19 Leonard
recognized that a “stretched out” adenine base would likely not
pair well within the natural DNA framework, unless it were
paired with a shortened version of thymine to compensate for
the size differential.20
We have undertaken a program to develop a new artificial
genetic system that is not constrained by the dimensions and
geometries of the natural DNA helix.21 By combining size-
expanded nucleobase analogues with the natural ones, we
envision size-expanded pairs (Figure 1). Assembly of strands
containing only size-expanded pairs might result in a fully
expanded helix (termed xDNA) with diameter larger than the
natural one, and this might avoid the geometric problem of
including an isolated larger pair within the natural framework.
Ultimately, such a nonnatural genetic system might one day be
The two nucleosides dxA and dxT enable the possibility of
a full four-base genetic pairing system. The four pairs of this
system are xA-T, T-xA, xT-A, and A-xT (Figure 1). This
is the same information-encoding potential as the natural genetic
system. If the natural bases A and T do not pair together well
within this expanded framework, and if xDNA bases do not
pair well within the natural framework, then the xDNA system
is orthogonal to the natural DNA system. Note also that
analogous extended cytosine and guanine base analogues are
(16) Leonard, N. J.; Sprecker, M. A.; Morrice, A. G. J. Am. Chem. Soc. 1976,
98, 3987-3994.
(17) Leonard, N. J.; Scopes, D. I. C.; Vanderlijn, P.; Barrio, J. R. Biochemistry
1978, 17, 3677-3685.
(18) Lessor, R. A.; Gibson, K. J.; Leonard, N. J. Biochemistry 1984, 23, 3868-
3873.
(19) Seley, K. L.; Januszczyk, P.; Hagos, A.; Zhang, L.; Dransfield, D. T. J.
Med. Chem. 2000, 43, 4877-4883.
(20) Czarnik, A. W.; Leonard, N. J. J. Am. Chem. Soc. 1982, 104, 2624-2631.
(21) Liu, H.; Gao, J.; Lynch, S. R.; Maynard, L.; Saito, D.; Kool, E. T. Science
2003, 302, 868-871.
9
J. AM. CHEM. SOC. VOL. 126, NO. 4, 2004 1103

A R T I C L E S
Liu et al.
Scheme 1 a
a
Reagents and conditions: (a) KMnO4, tBuOH, water, 75 °C, 1.5 h, 60%. (b) Ac2O, 155 °C, 3 h, 90%. (c) TMSN3, 90-95 °C, 3 h. (d) Formamidine
acetate, DMF, 155 °C, 3 h, 55% (two steps). (e) P2S5, pyridine, 140 °C, 36 h, 100%. (f) CH3I, KOH, 1 h, 64%. (g) NaH, CH3CN, chlorosugar, 6 h, 37%.
(h) NH3, EtOH, 150 °C, 4 × 12 h, 61%. (i) N,N-Dimethyl acetamide dimethylacetal, MeOH, 60 °C, 72 h, 73%. (j) 4,4′-Dimethoxytrityl chloride, DIPEA,
pyridine, 6 h, 72%. (k) N,N-Diisopropyl cyanoethyl chlorophosphoramidite, CH2Cl2, DIPEA, 5 h, 100%.
envisioned; combination of those with natural G and C (along
with the above four pairs) leads to a possible eight-base paired
genetic system.
that reported by Leonard.18 That previous approach involved
deoxygenation of a ribonucleoside derivative; we envisioned
instead a more direct approach to the deoxyribonucleoside. The
previously known extended base precursor 8-methythioimidazo-
[4,5-g]quinazoline (3, Scheme 1) was prepared repeating a six-
step procedure in 19% overall yield, starting from 5,6-
dimethylbenzimidazole. In a new approach to the deoxyriboside,
compound 3 was then subjected to coupling with Hoffer’s
R-chlorosugar24 in anhydrous acetonitrile, using NaH (95%, dry)
as a strong base. Due to a lack of stereocontrolling factors, this
coupling reaction afforded four products as two approximately
1:1 mixtures of deoxyribosides (Scheme 2), with methylthio
group oriented down (toward the sugar) (4b) or up (4a). Each
mixture was composed of two C1′-anomers, with the â-isomers
dominant. Regioisomers 4a and 4b were easily separated by
silica chromatography. However, no separation of C1′-isomers
was accomplished until at a later stage. The structure of each
regioisomer was assigned by 2D-ROESY NMR experiments.
Different ROE signal patterns were observed, as shown in
Scheme 2. In compound 4b, there was an aromatic proton that
had correlations with both H1′ and SMe, while there was no
such proton in 4a.
Interestingly, these expanded bases are expected to be either
orthogonal to or compatible with natural DNA (RNA) strands,
depending on the sequence design. If the extended bases are
placed in both strands of an expanded duplex, the pairing is
expected to be orthogonal by selective size requirements (three
rings paired with one or two paired with two but selective
against two paired with one, as occurs in DNA). However, if
all expanded bases are segregated into one strand, then this
xDNA strand can pair with natural DNA or RNA strands to
form an expanded helix.21
The reasons for our study of xDNA are multifold. First, we
expect that the new helix may be more stable thermodynamically
than that of natural DNA because the bases are larger and thus
might stack more efficiently.22 This could enable useful ap-
plications in hybridization with natural DNA and RNA. Second,
the greater electronic conjugation xDNA bases could yield useful
properties such as fluorescence. Third, we hope to use such size-
expanded nucleobases to probe steric effects in the active sites
of polymerase enzymes.14,15,23 Fourth, if xDNA can indeed
undergo stable and selective self-recognition, then it is a
candidate for the genetic component of possible extraterrestrial
forms of life. Finally, there is our aforementioned goal of
developing a new genetic system that is orthogonal to the natural
one.
The desired regioisomers 4a were then treated with ammonia
in ethanol at 150 °C in a sealed tube. This step simultaneously
converted the methylthio group to an amino group and removed
3′ and 5′ hydroxyl protections, thus furnishing the free dxA
nucleoside 1. The overall preparation from 3 was accomplished
in eight steps in 4.3% overall yield (assuming a single major
â-epimer); this compares favorably to Leonard’s approach20 (12
steps, <2.9% yield).
Expanded Deoxyadenosine Analogue. The expanded deoxy-
adenosine nucleoside 1 was prepared by a route different from
(22) Guckian, K. M.; Schweitzer, B. A.; Ren, R. X.-F.; Sheils, C. J.; Tahmassebi,
D. C.; Kool, E. T. J. Am. Chem. Soc. 2000, 122, 2213-2222.
(23) Strerath, M.; Cramer, J.; Restle, T.; Marx, A. J. Am. Chem. Soc. 2002,
124, 11230-11231.
(24) Hoffer, M. Chem. Ber. 1960, 93, 2777-2781.
9
1104 J. AM. CHEM. SOC. VOL. 126, NO. 4, 2004

A Genetic System with Expanded Size
A R T I C L E S
nm maximum had a molar absorptivity of 11 080 L/mol‚cm.
The compound fluoresced in the blue-violet (λmax,em ) 393 nm)
with a quantum yield of 0.44. This compares to other fluorescent
adenine analogues such as 2-aminopurine, which has a low
molar absorptivity and emits in the UV (λmax,em ) 365 nm),25
and ethenoadenine, which has a similar quantum yield of 0.56
(λmax,em ) 410 nm).26
Simple AM1 calculations were carried out for the free base
of dxA to explore tautomeric preferences. The results suggested,
as expected by analogy to adenine, that the benzo-fused adenine
base of dxA preferred the amino tautomer over imino variants
(Figure 3).
For eventual incorporation into oligonucleotides, this nucleo-
side required base protection and sugar functionalization steps.
The amino functional group in nucleoside 1 was protected with
N,N-dimethylacetamide dimethylacetal in 73% yield (compound
5). In the next step, the 5′-OH was selectively protected with
dimethoxytrityl. Interestingly, this tritylation reaction furnished
the 1′-â anomer (6) as the only isolated product in 72% yield.
The C1′-â configuration was assigned by means of COSY and
2D-ROESY NMR spectroscopy. ROEs between H1′ and H4′,
H2 and H3′, H5′ were observed (data not shown), which were
used as proof for the C1′-â configurational assignment. Finally,
the dxA phosphoramidite derivative (7) was prepared in
quantitative yield by treating the DMT-dxA ether with 2-cya-
noethyl-diisopropyl-chlorophosphoramidite. To demonstrate in-
tact incorporation into DNA, we prepared the oligonucleotide
dT-xA-T following the reported procedure21 and characterized
its proton NMR spectrum without purification. The spectrum
(see Supporting Information) confirmed the removal of the
protecting group as well as high yield of coupling.
Expanded Thymidine Analogue. We designed the meth-
ylquinazolinedione analogue dxT (2) as a benzo-fused variant
of thymidine having analogous geometric changes as was found
in the benzoadenine moiety of dxA (Figure 1). The methyl group
of dxT was retained in analogy to that of thymidine. A
previously reported quinazolinedione nucleoside27 did not have
expanded pairing dimensions due to a different position of
glycosidic attachment.
Figure 2. Fluorescence properties of extended nucleosides dxA and dxT.
(A) Excitation (dashed line, λem ) 393 nm) and emission (solid, λex ) 333
nm) spectra of dxA nucleoside 1 in methanol. (B) Excitation (λem ) 377
nm) and emission (λex ) 320 nm) spectra for quinazolinedione dxT
nucleoside 2 in methanol.
Scheme 2
This second building block, with 6-methyl-quinazoline-2,4-
(1H,3H)-dione as the nucleobase, was prepared by a route
somewhat shorter and more efficient than the dxA analogue
(Scheme 3). The 8-iodo derivative of 6-methyl-quinazoline-2,4-
(1H,3H)-dione 10 was made through two-step synthesis with
5-methyl-2-aminobenzoic acid 8 as starting material. First,
iodination with iodine monochloride of 8 gave 2-amino-3-iodo-
5-methylbenzoic acid 9 in high yield; subsequent cyclization
of 9 in urea melt afforded compound 10 in excellent yield.
Attempts were made to use compound 10 directly in a Heck
coupling reaction with 1,2-dehydro-3-O-(tert-butyl-diphenylsi-
lyl)-5-hydroxymethyl-furan 11, a ribofuranoid glycal designed
for stereospecific formation of â C-glycosyl bonds.28 However,
none of the expected product was obtained under several
reaction conditions, presumably due to nitrogen complexation
with palladium. To circumvent this problem, compound 12, a
(25) Ward, D. C.; Reich, E.; Stryer, L. J. Biol. Chem. 1969, 244, 1228-1237.
(26) Secrist, J. A.; Barrio, J. R.; Leonard, N. J.; Weber, G. Biochemistry 1972,
11, 3499-3506.
As was reported previously for the ribonucleoside derivative
of the xA base,16 our data showed that the deoxynucleoside dxA
was fluorescent (Figure 2). In methanol solution it absorbed
light with maxima at 231, 260, 319, 333, and 348 nm; the 333
(27) (a) Dunkel, M.; Pfleiderer, W. Nucleosides Nucleotides 1992, 11 787-
819. (b) Michel, J.; Toulme, J. J.; Vercauteren, J.; Moreau, S. Nucleic Acids
Res. 1996, 24, 1127-1135.
(28) Zhang, H. C.; Daves, G. D. J. J. Org. Chem. 1992, 57, 4690-4696.
9
J. AM. CHEM. SOC. VOL. 126, NO. 4, 2004 1105

A R T I C L E S
Liu et al.
Figure 3. Preliminary calculations of relative tautomeric preferences for the xT and xA bases in the gas phase. Spartan software (Wavefunction, Inc.) was
used to perform AM1 geometry optimizations and heats of formation calculations.
Scheme 3 a
a
Reagents and conditions: (a) ICl, HCl (aq), 96%. b) Urea, 150 °C, 92%. (c) POCl3, N,N-Diethylaniline, reflux, 90%. (d) NaOCH3/CH3OH, reflux, 87%.
(e) 1,2-Dehydro-3-O-(tert-butyl-diphenylsilyl)-5-hydroxymethyl-furan, Pd(OAc)2, AsPh3, N(Bu)3, 70 °C, 64%. (f) TBAF, THF, 0 °C, 84%. (g) NaB(OAc)3H,
THF, AcOH, 15 °C, 93%. (h) NaI, AcOH, 60 °C, 100%. (i) 4,4′-Dimethoxytrityl chloride, DIPEA, pyridine, 84%. (j) N,N-Diisopropylammonium tetrazolide,
2-cyanoethyl tetraisopropylphosphoramidite, CH2Cl2, 82%.
dimethyl-protected derivative of 10, was made through a two-
step synthesis in very good yield.
The free expanded T nucleoside dxT 2 was obtained by
removing the methyl-protecting groups on compound 15. The
overall approach afforded this new nucleoside in eight steps
and 34% yield.
Because of the added benzo fusion, the extended conjugation
of the xT base rendered it fluorescent, in contrast to the parent
thymine base. The absorption and emission properties of the
nucleoside 2 were measured in methanol. As with the xA base,
xT (as the deoxynucleoside) showed multiple absorption
maxima, and the long-wavelength maximum of xT (at 320 nm)
had a molar absorptivity of 3400 L/mol‚cm (Figure 2). The
Heck coupling between 11 and 12 was relatively efficient
with Pd(OAc)2/AsPh3 combination as catalyst. The coupled
product 13 was obtained in 64% yield. The tert-butyl-diphen-
ylsilyl protecting group was easily removed with TBAF, and
the resulting compound 14 was then stereoselectively reduced
with sodium triacetoxyborohydride to form the â-nucleoside of
dimethoxy-protected quinazoline 15 as the only product. The
stereochemistry was confirmed by 1-D NOE experiments:
irradiation of 2′H-â gave an NOE signal on 3′H (6.2%), whereas
irradiation at 2′H-R yielded significant enhancement on the 1′H
(4.8%).29
(29) Chaudhuri, N. C.; Ren, R. X.-F.; Kool, E. T. Synlett 1997, 341-347.
9
1106 J. AM. CHEM. SOC. VOL. 126, NO. 4, 2004

A Genetic System with Expanded Size
A R T I C L E S
was added 95% dry NaH (583 mg). After being stirred at room
temperature for 1 h, Hoffer’s chlorosugar24 (4.90 g) was added in one
portion. The reaction mixture was stirred for 6 h at r.t. under nitrogen
atmosphere and then was filtered, concentrated, and redissolved in ethyl
acetate (250 mL), which was washed with 2% HCl (aq), 5% NaHCO3
(aq), and brine (125 mL each). After drying over Na2SO4, the crude
product was purified by silica column chromatography (33-100%
EtOAc in hexanes). An ∼1:4 mixture of 1′-R,â epimers was obtained
as a yellow solid foam (1.86 g, 37%). 1H NMR (CDCl3, 500 MHz): δ
8.90 (s, 1H), 8.54 (s, 0.20H), 8.51 (s, 0.80H), 8.46 (s, 0.20H), 8.36 (s,
0.80H), 8.10 (s, 0.24H), 8.08 (s, 0.76H), 7.92-7.89 (m, 2H), 7.79-
7.77 (m, 1.50H), 7.64-7.63 (d, 0.5H), 7.24-7.19 (m, 2H), 7.14-7.12
(m, 2H), 6.53-6.52 (m, 0.23H), 6.47-6.44 (m, 0.80H), 5.70-5.68 (m,
1H), 4.82-4.81 (m, 0.21H), 4.67-4.59 (m, 2.90H), 3.10-3.07 (m,
fluorescence emission maximum appeared at 377 nm, in the
blue-violet range, and had a quantum yield of 0.30. These
properties were similar to that of a previously reported quinazo-
linedione derivative.27
For this new nucleobase analogue there were several possible
variations of keto-enol tautomers (Figure 3). Preliminary
semiempirical calculations of heats of formation suggested,
however, that the bis-keto tautomer was preferred (at least in
the gas phase) by a wide margin of >11 kcal/mol.
DMT protection of the 5′-OH of the dxT deoxynucleoside
was done following standard procedures (Scheme 3). In the
subsequent reaction between DMT-protected nucleoside 16 and
2-cyanoethyl N,N-diisopropylchlorophosphoramidite, a com-
monly used reagent to introduce phosphoramidite groups at the
3′-OH group of nucleosides, a double phosphorylation occurred.
Presumably, nucleophilicity of O4 on the quinazolinedione is
quite significant; thus, it can be phosphorylated as well. In
response to this, the milder reagent 2-cyanoethyl bis(diisopro-
pylamino)phosphoramidite30 was used instead, which gave the
desired xT-phosphoramidite 17 in good yield. This reactivity
at O4 foreshadowed the possibility of undesired reactions with
phosphoramidites during oligonucleotide synthesis; such reac-
tions could lead to branched side products. Thus, we also
retained the O-methyl-protected nucleoside and prepared the
5′-dimethoxytrityl-3′-phosphoramidite derivatives of it as well,
for possible future application to oligonucleotide synthesis.
Preliminary studies have suggested, however, that this protection
is not necessary and leads to complications in the subsequent
deprotection.21 To demonstrate intact incorporation, we prepared
the oligonucleotide dT-xT-T following the reported proce-
dure21 and characterized its proton NMR spectrum without
purification. The spectrum is given in the Supporting Informa-
tion and confirms removal of the protecting groups as well as
high yield of coupling.
0.22H), 2.84-2.79 (m, 1.96H), 2.71 (s, 3H), 2.38-0.230 (m, 6H). 13
C
NMR (CDCl3, 125 MHz): δ 173.14, 173.04, 171.36, 166.37, 166.16,
152.19, 146.27, 145.70, 144.91, 144.52, 144.47, 144.17, 143.79, 138.16,
138.03, 130.05, 129.92, 129. 87, 129.82, 129.60, 129.56, 126.90, 126.71,
126.50, 126.13, 120.62, 120.57, 115.05, 114.88, 107.83, 107.49, 86.89,
85.81, 84.74, 83.13, 75.06, 74.96, 64.30, 64.07, 60.61, 38.74, 38.41,
22.00, 21.91, 21.28, 14.43, 12.97. HRMS: (M + H) calcd for
C31H29N4O5S, 569.1859; found, 569.1877. A second product was eluted
with 100% ethyl acetate and was identified as the next regioisomer.
8-Methylthio-1-(2′-deoxy-3′,5′-di-O-toluoyl-D-ribofuranosyl)-imi-
dazo[4,5-g]quinazoline. 1H NMR (CDCl3, 500 MHz): δ 8.94 (s, 0.2H),
8.93 (s, 0.8H), 8.54 (s, 0.2H), 8.45 (s, 0.8H), 8.39 (s, 1H), 8.20 (s,
0.2H), 8.17 (s, 0.8H), 7.95-7.91 (m, 2H), 7.73-7.71 (m, 2H), 7.26-
7.22 (m, 2H), 7.15-7.10 (m, 2H), 6.49-6.47 (dd, 1H), 5.72-5.70 (m,
1H), 4.69-4.66 (m, 2H), 4.64-4.60 (m, 1H), 2.94-2.86 (m, 2H), 2.77
(s, 0.6H), 2.71 (s, 2.4H), 2.39-2.31 (m, 6H). 13C NMR (CDCl3, 125
MHz): δ 166.10, 165.93, 151.23, 145.62, 144.70, 144.26, 144.09,
129.81, 129.66, 129.58, 129.54, 129.35, 129.32, 126.39, 126.14, 120.20,
114.98, 114.79, 106.48, 86.66, 85.52, 84.59, 82.95, 74.77, 74.65, 64.00,
63.79, 38.59, 38.24, 21.76, 21.66, 12.89.
3-[2′-Deoxy-D-ribofuranosyl]-8-aminoimidazo[4,5-g]quinazo-
line (1). 4a (790 mg) was dissolved in EtOH (12 mL), ammonia gas
was bubbled through at 0 °C for 35 s, and then the solution was heated
at 150 °C for 12 h in an Ace pressure tube. To complete the reaction,
the same operation was repeated 4 times. The product was purified by
silica column chromatography (33-50% MeOH in CH2Cl2). The
product was obtained as a light yellow solid (252 mg, 61%), which
Studies of DNAs that contain dxA and dxT are underway
and will be reported in due course. Further work will be aimed
at synthesis and study of analogous size-expanded variants of
cytosine and guanine as well.
1
was a mixture of ∼2:5 1′-R and â epimers. H NMR (CD3OD, 500
Experimental Section
MHz): δ 8.83 (s, 0.28H), 8.79 (s, 0.71H), 8.57 (s, 1H), 8.38 (s, 1H),
8.00, 7.98 (s, 1H), 6.56-6.52 (m, 1H), 4.61-4.59 (m, 0.30H), 4.56-
4.55 (m, 0.70H), 4.28-4.27 (m, 0.32H), 4.10-4.08 (m, 0.72H), 3.86-
3.78 (m, 1.49H), 3.73-3.70 (m, 0.64H), 2.93-2.90 (m, 0.30H), 2.80-
2.74 (m, 0.66H), 2.59-2.53 (m, 1H). 13C NMR (DMSO-d6, 125
MHz): δ 163.07, 154.74, 154.64, 147.19, 147.09, 146.22, 143.45,
137.97, 114.14, 114.01, 107.66, 107.59, 88.19, 85.41, 85.28, 71.15,
71.00, 62.11. HRMS: (M + H) calcd for C14H16N5O3, 302.1253; found,
302.1243.
General Methods. Reagents were purchased from Aldrich and used
without further purification. All water-sensitive reactions were carried
out in oven-dried glassware with a stirring bar under a nitrogen or argon
atmosphere. Anhydrous solvents pyridine, CH3CN, CH2Cl2, and THF,
were distilled under a nitrogen atmosphere. THF was dried over Na0,
and the other three solvents were dried over CaH2. All other anhydrous
solvents were purchased from Aldrich and used directly. Thin-layer
chromatography was carried out using EM Science Silica Gel 60 F254
plates. Column chromatography was performed using Selecto Scientific
Silica Gel, sizes 32-63. All 1H, 13C, and 31P NMR spectra were
recorded on Varian Mercury-400 or Inova-500 instruments as noted.
31P NMR spectra were referenced to 85% H3PO4 in water as external
standard (0.0 ppm). 2D-COSY and 2D-ROESY spectra were acquired
on an Inova-500 instrument. High-resolution mass spectrometry was
performed at UC Riverside Mass Spectrometry Facilities.
3-[2′-Deoxy-D-ribofuranosyl]-8-[N-[1-(dimethyl-amino)ethylidine]-
amino]-imidazo[4,5-g]quinazoline (5). 1 (700 mg) was dissolved in
12 mL of dry MeOH, to which was added 90% N,N-dimethylacetamide
dimethyl acetal in MeOH (8 mL). After being heated at 60 °C for 72
h, the reaction mixture was concentrated and purified by silica column
chromatography (5-20% MeOH in CH2Cl2). A light yellow foamy
solid was obtained (626 mg, 73%), which was a mixture of ∼1:3 1′-R
and â epimers. 1H NMR (D2O, 500 MHz): δ 8.20 (s, 0.24H), 8.07 (s,
0.71H), 7.95 (s, 0.22H), 7.88 (0.61H), 7.55 (0.24H), 7.44 (s, 0.68H),
6.96 (s, 0.25H), 6.86 (s, 0.70H), 5.95-5.94 (m, 1H), 4.45-4.42 (m,
0.76H), 4.39-4.37 (m, 0.31H), 4.09-4.08 (m, 0.29H), 3.95-3.93 (m,
0.78H), 3.65-3.53 (m, 2H), 2.99 (s, 6H), 2.51-2.41 (m, 1H), 2.37-
2.27 (m, 1H), 1.93 (s, 2.3H), 1.96 (s, 0.82H). 13C NMR (D2O 125
MHz): δ 166.39, 166.16, 163.85, 152.24, 145.63, 143.83, 141.27,
141.07, 136.55, 116.88, 114.9, 104.68, 87.68, 86.99, 85.45, 84.61, 70.86,
8-Methylthio-3-(2′-deoxy-3′,5′-di-O-toluoyl-D-ribofuranosyl)-imi-
dazo[4,5-g]quinazoline (4). 8-Methythioimidazo[4,5-g]quinazoline31
(1.90 g) was suspended in 300 mL anhydrous acetonitrile, to which
(30) Barone, A. D.; Tang, J. Y.; Caruthers, M. H. Nucleic Acid Res. 1984, 12,
4051-4061.
(31) (a) Leonard, N. J.; Kazmierczak, F.; Rykowski, A. J. Org. Chem. 1987,
52, 2933-2935. (b) Leonard, N. J.; Morrice, A. G.; Sprecker, M. A. J.
Org. Chem. 1975, 40, 356-363.
9
J. AM. CHEM. SOC. VOL. 126, NO. 4, 2004 1107

A R T I C L E S
Liu et al.
61.58, 61.31, 39.09, 38.68, 16.86. HRMS: (M + H) calcd for
C18H23N6O3, 371.1832; found, 371.1840.
111.30, 86.97, 19.89. HRMS m/z: [M]+ (EI+) calcd for C8H8NO2I,
276.9600; found, 276.9606.
8-Iodo-6-methylquinazoline-2,4-dione (10). 2-Amino-3-iodo-5-
methylbenzoic acid 9 (7.8 g, 28.1 mmol) and urea (16.9 g, 281 mmol)
were heated to 150 °C in a round-bottom flask, and the melt was stirred
for 12 h. The mixture was then allowed to cool to 100 °C, and a volume
equivalent of water was added. The resulting mixture was then stirred
for a few minutes to allow urea to dissolve. The solid was filtered,
resuspended in 200 mL of 0.5 M NaOH (aq), and then heated to 100
°C to allow formation of the sodium salt of the product. The mixture
was retriturated with glacial acetic acid. The mixture was filtered, and
the precipitate was collected and allowed to dry to a fine brown powder
3-[2′-Deoxy-5′-O-(4,4′-dimethoxytrityl)-â-D-ribofuranosyl]-8-[N-
[1-(dimethyl-amino)ethylidine]amino]-imidazo[4,5-g]quinazoline (6).
5 (544 mg, 1.47 mmol) was coevaporated with anhydrous pyridine (3
× 8.0 mL) and then dissolved in 20 mL of anhydrous pyridine, to which
was added DIPEA (1.65 mL), DMAP (50 mg), and 4,4′-dimethoxytrityl
chloride (944 mg, 2.65 mmol). After being stirred at r.t. under nitrogen
atmosphere for 8 h, the reaction was quenched with addition of MeOH.
The reaction mixture was concentrated, redissolved in EtOAc (120 mL),
and washed with 5% NaOH (aq) (2 × 50 mL), followed by washing
with brine (50 mL). The organic phase was then dried over Na2SO4,
filtered, concentrated, and purified by silica column chromatography
(0-5% MeOH in CH2Cl2, with 1% TEA). The title compound was
obtained as a yellow foam solid (710 mg, 72%). The 1′-â configuration
was assigned based on 2D ROESY spectra. 1H NMR (CDCl3, 500
MHz): δ 8.68 (s, 1H), 8.58 (s, 1H), 8.23 (s, 1H), 7.88 (s, 1H), 7.30-
7.07 (m, 9H), 6.67-6.64 (m, 4H), 6.50 (t, 7 Hz), 4.63-4.62 (m, 1H),
4.29-4.28 (m, 1H), 3.64 (s, 6H), 3.30-3.29 (m, 2H), 3.18 (s, 3H),
3.09 (s, 3H), 2.60-2.57 (m, 2H), 2.15 (s, 3H). 13C NMR (CDCl3,125
MHz): δ 168.05, 161.68, 158.70, 154.46, 146.48, 145.09, 144.78,
143.71, 138.24, 135.93, 135.83, 130.25, 130.21, 128.35, 128.11, 127.11,
118.55, 117.19, 113.41, 105.59, 86.72, 86.64, 85.57, 72.12, 64.22, 55.42,
53.19, 46.15, 41.40, 38.76, 17.07, 9.48, 8.25. HRMS: (M + H) calcd
for C39H41N6O5, 673.3138; found, 673.3168.
1
(7.8 g, 92%). H NMR (DMSO-d6, 500 MHz): δ 11.51 (s, 1H), 9.48
(s, 1H), 8.00 (s, 1H), 7.74 (s, 1H), 2.30 (s, 3H). 13C NMR (DMSO-d6,
125 MHz): δ 126.85, 150.48, 146.48, 139.81, 134.53, 128.05, 116.06,
84.21, 20.20. HRMS m/z: [M]+ (EI+) calcd for C9H7N2O2I, 301.9552;
found, 301.9540.
8-Iodo-2,4-dichloro-6-methylquinazoline (11). To a flame-dried
round-bottom flask outfitted with a condenser, 8-iodo-6-methylquinazo-
line-2,4-dione 10 (1.20 g, 4 mmol) was dissolved in phosphorus
oxychloride (8.00 mL, 87 mmol). To this solution was added N,N-
diethylaniline (1.2 mL, 7.6 mmol). The mixture was stirred under reflux
for 3 h. Then the reaction mixture was allowed to cool to room
temperature, and volatiles were removed under vacuum. The remaining
residue was dissolved in 60 mL of chloroform and washed twice with
ice-cold water. The organic layer was collected, dried, and concentrated.
Silica column chromatography (hexanes/ethyl acetate 4:1) gave the
product 11 as yellowish powder (1.23 g, 90%). 1H NMR (CDCl5, 400
MHz): δ 8.42 (s, 1H), 8.02 (s, 1H), 2.59 (s, 3H). 13C NMR (CDCl3,
100 MHz): δ 163.88, 155.48, 150.86, 148.65, 141.43, 125.64, 122.91,
99.82, 21.67. HRMS m/z: [M]+ (EI+) calcd for C9H5N2Cl2I, 337.8874;
found, 337.8866.
3-[2′-Deoxy-3′-O-(2-cyanoethyl-N,N-diisopropylphosphino)-5′-O-
(4-4′-dimethoxy-trityl)-â-D-ribofuranosyl]-8-[N-[1-(dimethyl-amino)-
ethylidine]amino]-imidazo-[4,5-g]quinazoline (7). 6 (215 mg, 0.30
mmol) was dissolved in 12 mL of anhydrous methylene chloride, to
which was added 0.80 mL of DIPEA and 143 µL (0.60 mmol) of
2-cyanoethyl diisopropylchlorophosphoramidite. After being stirred at
room temperature under nitrogen atmosphere for 7 h, the reaction was
quenched with a few drops of methanol, then purified by silica column
chromatography (0-3% MeOH in CH2Cl2). A light yellow foamy solid
was obtained (100%). 1H NMR (CDCl3, 500 MHz): δ 8.78 (s, 0.5H),
8.77 (s, 0.5H), 8.69 (s, 0.5H), 8.68 (s, 0.5H), 8.37 (s, 0.5H), 8.34 (s,
0.5H), 7.98 (s, 0.5H), 7.95 (s, 0.5H), 7.40-7.17 (9H), 6.89-6.74 (4H),
6.44-6.41 (m, 1H), 4.72-4.69 (m, 1H), 4.36-4.33 (m, 1H), 3.88-
3.87 (m, 1H), 3.81-3.80 (d, 1H), 3.76-3.74 (m, 6H), 3.67-3.55 (m,
2H), 3.44-3.41 (m, 1H), 3.37-3.34 (m, 1H), 3.30 (s, 3H), 3.22 (s,
3H), 2.79-2.78 (m, 1H), 2.71-2.69 (m, 1H), 2.67-2.65 (t, 1H), 2.49-
2.47 (t, 1H), 2.30 (s, 3H), 1.28-1.12 (m, 12H). 13C NMR (CDCl3, 125
MHz): δ 158.32, 144.65, 144.39, 143.47, 143.35, 137.83, 137.74,
135.37, 135.31, 135.28, 135.00, 129.89, 129.84, 127.96, 127.89, 127.78,
127.72, 126.76, 126.71, 117.73, 117.49, 117.31, 117.00, 116.81, 112.99,
105.48, 86.78, 86.34, 86.31, 85.58, 85.35, 85.14, 73.94, 73.80, 73.41,
73.28, 63.26, 63.11, 58.20, 58.14, 58.05, 57.99, 57.95, 57.50, 55.04,
53.47, 52.74, 46.92, 46.75, 46.20, 45.63, 45.46, 45.14, 45.09, 34.20,
43.14, 43.10, 43.04, 39.85, 38.66, 24.45, 24.42, 23.21, 22.81, 22.73,
22.71, 22.47, 22.32, 22.17, 20.33, 20.28, 20.10, 20.04, 19.97, 19.91,
18.91, 17.20, 8.49, 8.00. 31P NMR (CDCl3, 202 MHz): δ 146.57,
146.52. HRMS: (M + H) calcd for C53H65N10O8P, 873.4217; found,
873.4198.
8-Iodo-2,4-dimethoxy-6-methylquinazoline (12). 8-Iodo-2,4-dichloro-
6-methylquinazoline 11 (1.20 g, 3.5 mmol) was suspended in 0.5 M
sodium methoxide solution in methanol. The reaction mixture was
heated to reflux and stirred under reflux for 14 h. Then the reaction
was allowed to cool to room temperature and was neutralized with 1
M HCl (aq). A precipitate was formed upon neutralization. After cooling
in ice water for 30 min, the precipitate was filtered to give the crude
product, which was further purified by silica column chromatography
to give 12 as a yellowish powder (1.02 g, 87%). 1H NMR (CDCl3, 400
MHz): δ 8.13 (s, 1H), 7.81 (s, 1H), 4.16 (s, 3H), 4.14 (s, 3H), 2.42 (s,
3H). 13C NMR (CDCl3, 100 MHz): δ 169.70, 162.31, 150.17, 145.73,
135.76, 123.71, 113.97, 98.35, 55.34, 55.19, 20.08. HRMS m/z: [M]+
calcd for C11H11N2O2I, 329.9865; found, 329.9855.
2,4-Dimethoxy-6-methyl-8- {(2′R)-cis-3-[2′,3′-dehydro-3′-(tert-
butyl-diphenylsilyloxy)]-5′-hydroxymethyl-2′-furanyl}quinazoline
(13). 1,2-Dehydro-3-O-(tert-butyl-diphenylsilyl)-5-hydroxymethyl-
furan28 (1.20 g, 3.38 mmol) and the dimethoxy-protected quinazoline
12 (1.12 g, 3.38 mmol) were charged into a 250 mL round-bottom
flask. Anhydrous DMF (40 mL) was added to form a suspension. Pd-
(OAc)2 (152 mg, 0.68 mmol) and AsPh3 (416 mg, 1.36 mmol) were
charged into a 100 mL conical flask, and 60 mL of dry DMF was
added. The catalyst-ligand mixture was stirred at room temperature
for 30 min before being transferred to the glycal-quinazoline suspension
via syringe. Tri(n-butyl) amine (1.29 mL, 5.41 mmol) was added to
the reaction mixture in one portion. Then the reaction mixture was
heated to 75 °C and was stirred for 36 h. Volatiles were removed under
vacuum, and the residue was purified with silica column chromatog-
raphy (hexanes/ethyl acetate 4:1). Product 13 was obtained as yellow
2-Amino-3-iodo-5-methylbenzoic Acid (9). In a round-bottom flask,
5-methyl-2-aminobenzoic acid 8 (4.4 g, 29.1 mmol) was dissolved in
80 mL of a 6% HCl (aq) solution. The mixture was warmed to 50 °C.
To this was added ICl (14.17 g, 87.3 mmol) dissolved in 20 mL of a
6% HCl (aq) solution in a single portion. The mixture was stirred briskly
for 30 min and then was allowed to cool to room temperature. Sodium
metabisulfate (16.6 g, 87.3 mmol) was added in a single portion. The
solution was subsequently neutralized with potassium hydroxide pellets.
The precipitate was filtered, washed with ice-cold water, and dried to
1
foam (1.22 g, 64%). H NMR (CDCl3, 400 MHz): δ 7.83-7.60 (m,
4H), 7.71 (m, 1H), 7.46-7.37 (m, 7H), 6.58 (dd, 1H, J ) 4 Hz, 1.6
Hz), 4.86 (m, 1H), 4.54 (m, 1H), 4.11 (s, 3H), 3.91 (s, 1H), 3.90 (m,
2H), 2.38 (s, 3H), 1.08 (s, 9H). 13C NMR (CDCl3, 100 MHz): δ 150.14,
137.26, 135.95, 135.67, 133.82, 133.44, 131.82, 131.58, 130.42, 128.12,
128.05, 122.86, 113.55, 103.46, 83.46, 79.63, 62.96, 54.81, 54.67, 26.62,
1
give a brown powder (7.8 g, 96%). H NMR (DMSO-d6, 500 MHz):
δ 7.70 (s, 1H), 7.61 (s, 1H), 2.51 (s, 2H), 2.15 (s, 3H). 13C NMR
(DMSO-d6, 125 MHz): δ 169.67, 148.76, 145.18, 132.50, 126.42,
9
1108 J. AM. CHEM. SOC. VOL. 126, NO. 4, 2004

A Genetic System with Expanded Size
A R T I C L E S
21.77, 19.48. HRMS m/z: [M + H]+ (FAB+) calcd for C32H37N2O5-
Si, 557.2471; found, 557.2453.
mixture was stirred at room temperature for 6 h. Volatiles were removed
under vacuum, and the crude product was purified by silica column
chromatography (ethyl acetate to ethyl acetate/methanol 10:1) to give
155 mg of product 16 as yellowish foam. Free nucleoside (62 mg)
was recovered from the reaction, which was allowed to react with DMT
chloride again following the procedure above, and 78 mg of additional
2,4-Dimethoxy-8-(â-D-glycero-pentofuran-3′-ulos-1′yl)-6-meth-
ylquinazoline (14). Compound 13 (540 mg, 0.97 mmol) was dissolved
in 20 mL of THF, and the solution was cooled to 0 °C by an ice-
water bath. Glacial acetic acid (0.24 mL) and TBAF solution (1 M
solution in THF, 1.46 mL, 1.46 mmol) were added sequentially via
syringe. The reaction mixture was allowed to stir 10 min at 0 °C.
Volatiles were removed under vacuum, and silica column chromatog-
raphy (hexanes/ethyl acetate 1:1) gave product 14 as white solid (260
mg, 84%). 1H NMR (CDCl3, 400 MHz): δ 7.82 (s, 1H), 7.78 (s, 1H),
5.84 (dd, 1H, J ) 10.8 Hz, 6.0 Hz), 4.15 (s, 3H), 4.03 (s, 5H), 3.18
(m, 1H), 2.68 (m, 1H), 2.49 (s, 3H). 13C NMR (CDCl3, 100 MHz): δ
215.06, 169.44, 161.40, 147.99, 134.93, 134.02, 132.85, 123.05, 113.70,
82.37, 75.49, 62.11, 55.04, 54.82, 45.00, 21.70. HRMS m/z: [M]+ (EI+)
calcd for C16H18N2O5, 318.1216; found, 318.1203.
1′-â-[8-(2,4-dimethoxy-6-methylquinazoline)]-2′-deoxyribofura-
nose (15). Compound 14 (123 mg, 0.39 mmol) was dissolved in a
mixture of 10 mL of CH3CN and 10 mL of glacial acetic acid. The
solution was chilled to -15 °C by an ice-salt bath. Sodium tri-
acetoxyborohydride (110 mg, 0.52 mmol) was added in one portion.
Then the mixture was stirred at -15 °C for 30 min. Solvents were
removed under vacuum, and the residue was purified by silica column
chromatography (ethyl acetate to ethyl acetate/methanol 95:5) to give
product 15 as a yellowish solid (116 mg, 93%). 1H NMR (CDCl3, 400
MHz): δ 7.77 (s, 1H), 7.67 (s, 1H), 5.79 (dd, 1H, J ) 11.2 Hz, 6 Hz),
4.53 (m, 1H), 4.16 (s, 3H), 4.06 (s, 3H), 3.93 (m, 1H), 3.84 (m, 1H),
2.53 (m, 1H), 2.46 (s, 3H), 2.26 (m, 1H). 13C NMR (CDCl3, 100
MHz): δ 169.19, 160.94, 147.90, 136.01, 133.61, 132.88, 122.17,
113.34, 87.46, 78.12, 74.36, 63.66, 54.77, 54.49, 43.07, 21.45. HRMS
m/z: [M]+ (EI+) calcd for C16H20N2O5, 320.1372; found, 320.1385.
1′-â-[8-(6-Methylquinazoline-2,4-dione)]-2′-deoxyribofuranose (2).
Compound 15 (123 g, 0.39 mmol) was dissolved in 8 mL of glacial
acetic acid, to which was added sodium iodide (288 mg, 1.92 mmol).
The reaction mixture was heated to 60 °C and then was stirred for 45
min. Volatiles were removed under vacuum. The residue was suspended
in 10 mL of saturated sodium bicarbonate solution, which was extracted
with 5 × 30 mL ethyl acetate. The organic layers were combined, dried
over anhydrous sodium sulfate, and concentrated. Silica column
chromatography (ethyl acetate/methanol 10:1) gave the product 2 as
white solid (113 mg, 100%). 1H NMR (CD3OD, 500 MHz): δ 7.83 (s,
1H), 7.47 (s, 1H), 5.31 (dd, 1H, J ) 11.5 Hz, 5 Hz), 4.49 (m, 1H),
4.07 (m, 1H), 3.85 (m, 2H), 2.40 (s, 3H), 2.23 (m, 1H), 2.14 (m, 1H).
13C NMR (CD3OD, 125 MHz): δ 163.96, 150.91, 136.52, 135.57,
132.63, 126.76, 126.15, 115.65, 88.73, 80.83, 73.45, 61.75, 42.16, 19.37.
HRMS [M]+ m/z: (EI+) calcd for C14H16N2O5, 292.1059; found,
292.1066.
1
product was obtained. Total yield was 84%. H NMR (CDCl3, 400
MHz): δ 7.88 (s, 1H), 7.41 (s, 1H), 7.37 (m, 2H), 7.28-7.17 (m, 7H),
6.78-6.75 (m, 4H), 5.44 (dd, 1H, J ) 11.2 Hz, 4.8 Hz), 4.46 (d, 1H,
J ) 5.2 Hz), 4.18 (m, 1H), 3.75 (s, 6H), 3.29 (m, 2H), 2.44 (m, 1H),
2.34 (s, 3H), 2.23 (m, 1H). 13C NMR (CDCl3, 125 MHz): δ 163.50,
158.70, 151.21, 145.00, 136.14, 136.05, 133.63, 133.31, 130.30, 128.36,
128.06, 127.43, 127.25, 127.06, 115.23, 113.34, 87.45, 86.52, 76.35,
74.50, 64.37, 55.43, 34.78, 21.04. HRMS m/z: [M + Na]+ (FAB+)
calcd for C35H34N2O7Na, 617.2264; found, 617.2278.
1′-â-[8-(6-Methylquinazoline-2,4-dione)]-5′-O-(4,4′-dimethoxy-
trityl)-2′-deoxyribofuranose-3′-(2-cyanoethyl phosphoramidite) (17).
5′-DMT-protected nucleoside 16 (233 mg, 0.39 mmol) and N,N-
diisopropylammonium tetrazolide (40 mg, 0.23 mmol) were dissolved
in 10 mL of dry dichloromethane. 2-Cyanoethyl tetraisopropylphos-
phoramidite (137 µL, 0.43 mmol) was added via syringe in one portion.
The reaction mixture was stirred at room temperature for 3 h. Volatiles
were then removed under vacuum, and the crude product was purified
by silica column chromatography to give product 17 as white foam
1
(254 mg, 82%, as a mixture of two diastereomers). H NMR (CDCl3,
400 MHz): δ 7.91 (s, 2H), 7.37-7.16 (m, 20H), 6.77-6.73 (m, 8H),
5.38 (m, 2H), 4.56 (m, 2H), 4.30 (m, 2H), 3.76 (s, 6H), 3.75 (s, 6H),
3.60 (m, 4H), 3.23 (m, 4H), 2.64 (t, 4H, J ) 6 Hz), 2.53 (m, 2H), 2.36
(s, 6H), 2.32 (m, 2H), 1.31-1.13 (m, 28H). 13C NMR (CDCl3, 125
MHz): δ 162.13, 157.40, 149.16, 143.64, 135.58, 134.71, 134.62,
132.61, 132.40, 131.61, 131.57, 129.01, 127.03, 126.74, 126.22, 125.74,
125.23, 116.66, 114.15, 112.02, 85.22, 76.64, 62.68, 57.24, 57.08, 54.13,
44.56, 42.24, 42.14, 39.12, 23.61, 23.56, 22.45, 22.10, 21.80, 21.19.
HRMS m/z: [M + Na]+ (FAB+) calcd for C44H51N4O8NaP, 817.3342;
found, 817.3338.
Fluorescence Measurements. Samples were measured in methanol.
UV-vis absorption spectra were recorded on a Varian Cary 1 UV-
vis spectrometer. Fluorescence excitation and emission spectra were
measured on a Spex Fluorolog 3 spectrometer. To prevent aggregation
and reabsorption of light, samples were diluted to an absorption at λmax
of less than 0.05. Quantum yields were calculated with fluorescein in
0.1 N NaOH solution as a reference.
Acknowledgment. We are grateful for support from the
National Institutes of Health (GM63587). J.G. and H.L. are
recipients of Stanford Graduate Fellowships.
Supporting Information Available: 1H and 13C NMR spectra
of all dxA and dxT synthetic intermediates, 2D ROESY data
for intermediates 4a, 4b, and 6, and COSY data for 6 (PDF).
This material is available free of charge via the Internet at
http://pubs.acs.org.
1′-â-[8-(6-Methylquinazoline-2,4-dione)]-5′-O-(4,4′-dimethoxy-
trityl)-2′-deoxyribofuranose (16). The nucleoside 2 (150 mg, 0.51
mmol) was coevaporated with dry pyridine (3 × 10 mL), and the residue
was dissolved in 4 mL of pyridine. N,N-diisopropylethylamine (266
µL, 1.53 mmol) was added via syringe in one portion. 4,4′-Dimeth-
oxytrityl chloride (227 mg, 0.67 mmol), dissolved in 6 mL of dry
pyridine, was transferred to the nucleoside solution via syringe. The
JA038384R
9
J. AM. CHEM. SOC. VOL. 126, NO. 4, 2004 1109