Efforts to expand the genetic alphabet: Identification of a replicable unnatural DNA self-pair

Article abstract of DOI:10.1021/ja049961u

Six unnatural nucleotides featuring fluorine-substituted phenyl nucleobase analogues have been synthesized, incorporated into DNA, and characterized in terms of the structure and replication properties of the self-pairs they form. Each unnatural self-pair

Full text of DOI:10.1021/ja049961u

Published on Web 05/18/2004
Efforts To Expand the Genetic Alphabet: Identification of a
Replicable Unnatural DNA Self-Pair
Allison A. Henry,† Anne Goldbech Olsen,† Shigeo Matsuda,† Chengzhi Yu,†
Bernhard H. Geierstanger,‡ and Floyd E. Romesberg*,†
Contribution from the Department of Chemistry, The Scripps Research Institute,
10550 North Torrey Pines Road, La Jolla, California, 92037, and Genomics Institute of the
NoVartis Research Foundation, 10675 John-Jay-Hopkins DriVe,
San Diego, California 92121-1125
Received January 3, 2004; E-mail: floyd@scripps.edu
Abstract: Six unnatural nucleotides featuring fluorine-substituted phenyl nucleobase analogues have been
synthesized, incorporated into DNA, and characterized in terms of the structure and replication properties
of the self-pairs they form. Each unnatural self-pair is accommodated in B-form DNA without detectable
structural perturbation, and all are thermally stable and selective to roughly the same degree. Furthermore,
the efficiency of polymerase-mediated mispair synthesis is similar for each unnatural nucleotide in the
template. In contrast, the efficiency of polymerase-mediated self-pair extension is highly dependent on the
specific fluorine substitution pattern. The most promising unnatural base pair candidate of this series is the
3-fluorobenzene self-pair, which is replicated with reasonable efficiency and selectivity.
1. Introduction
limits the number of available unnatural nucleobases to fewer
than can be constructed by synthetic chemists. Furthermore, it
is unclear at present that H-bonding is the only intermolecular
force suitable for the storage of genetic information. In fact, it
has been known for nearly a decade that unnatural nucleobases
without H-bonding functionality can be substrates for DNA
polymerases.15-19 Among the alternative intermolecular forces
that could mediate interbase interactions to a degree sufficient
for information storage is the hydrophobic force, whose
contribution to protein structure and folding has been appreciated
for decades.20 Thus, we have explored the utility of hydrophobic
and van der Waals forces in developing additional base pairs
for the storage and replication of increased genetic informa-
tion.21-27
Genetic information is encoded by a four-letter alphabet.
Faithful replication of this information is ensured by the selective
interactions between natural nucleobases: adenine with thymine
and guanine with cytosine. The identification of unnatural base
pairs that are enzymatically replicated and that have thermal
stability comparable to the natural pairs would allow for the
introduction of additional information, biological or abiological,
within a cell. Ongoing efforts to design unnatural base pairs
that interact via nonnatural hydrogen-bonding (H-bonding)
patterns have yielded some success,1-7 but have generally
proven difficult, possibly due in part to facile tautomerization
which converts the unnatural nucleobase to a form that readily
pairs with a natural nucleobase.8-14 In addition, this approach
(13) Lutz, M. J.; Horlacher, J.; Benner, S. A. Bioorg. Med. Chem. Lett. 1998,
8, 499-504.
† The Scripps Research Institute.
‡ Genomics Institute of the Novartis Research Foundation.
(1) Switzer, C.; Moroney, S. E.; Benner, S. A. J. Am. Chem. Soc. 1989, 111,
8322-8323.
(14) Roberts, C.; Bandaru, R.; Switzer, C. Tetrahedron Lett. 1995, 36, 3601-
3604.
(15) Moran, S.; Ren, R. X.-F.; Rumney, S. I.; Kool, E. T. J. Am. Chem. Soc.
1997, 119, 2056-2057.
(2) Piccirilli, J. A.; Krauch, T.; Moroney, S. E.; Benner, S. A. Nature 1990,
343, 33-37.
(16) Moran, S.; Ren, R. X.-F.; Kool, E. T. Proc. Natl. Acad. Sci. U.S.A. 1997,
94, 10506-10511.
(3) Switzer, C. Y.; Moroney, S. E.; Benner, S. A. Biochemistry 1993, 32,
10489-10496.
(17) Morales, J. C.; Kool, E. T. Nat. Struct. Biol. 1998, 5, 950-954.
(18) Kool, E. T. Biopolymers 1998, 48, 3-17.
(4) Piccirilli, J. A.; Moroney, S. E.; Benner, S. A. Biochemistry 1991, 30,
10350-10356.
(19) Kool, E. T. Curr. Opin. Chem. Biol. 2000, 4, 602-608.
(20) Rose, G. D.; Wolfenden, R. Annu. ReV. Biophys. Biomol. Struct. 1993, 22,
381-415.
(5) Moser, M. J.; Prudent, J. R. Nucleic Acids Res. 2003, 31, 5048-5053.
(6) Moser, M. J.; Marshall, D. J.; Greiner, J. K.; Kieffer, C. D.; Killeen, A.
A.; Ptacin, J. L.; Richmond, C. S.; Roesch, E. B.; Scherrer, C. W.; Sherrill,
C. B.; Van Hout, C. V.; Zanton, S. J.; Prudent, J. R. Clin. Chem. 2003, 49,
407-414.
(21) McMinn, D. L.; Ogawa, A. K.; Wu, Y.; Liu, J.; Schultz, P. G.; Romesberg,
F. E. J. Am. Chem. Soc. 1999, 121, 11585-11586.
(22) Ogawa, A. K.; Wu, Y.; McMinn, D. L.; Liu, J.; Schultz, P. G.; Romesberg,
F. E. J. Am. Chem. Soc. 2000, 122, 3274-3287.
(7) Tor, Y.; Dervan, P. B. J. Am. Chem. Soc. 1993, 115, 4461-4467.
(8) Bain, J. D.; Switzer, C.; Chamberlin, A. R.; Benner, S. A. Nature 1992,
356, 537-539.
(23) Wu, Y.; Ogawa, A. K.; Berger, M.; McMinn, D. L.; Schultz, P. G.;
Romesberg, F. E. J. Am. Chem. Soc. 2000, 122, 7621-7632.
(24) Ogawa, A. K.; Wu, Y.; Berger, M.; Schultz, P. G.; Romesberg, F. E. J.
Am. Chem. Soc. 2000, 122, 8803-8804.
(9) Horlacher, J.; Hottiger, M.; Podust, V. N.; Hu¨bscher, U.; Benner, S. A.
Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 6329-6333.
(10) Lutz, M. J.; Held, H. A.; Hottiger, M.; Hu¨bscher, U.; Benner, S. A. Nucleic
Acids Res. 1996, 24, 1308-1313.
(25) Tae, E. L.; Wu, Y. Q.; Xia, G.; Schultz, P. G.; Romesberg, F. E. J. Am.
Chem. Soc. 2001, 123, 7439-7440.
(11) Lutz, M. J.; Horlacher, J.; Benner, S. A. Bioorg. Med. Chem. Lett. 1998,
8, 1149-1152.
(26) Berger, M.; Luzzi, S. D.; Henry, A. A.; Romesberg, F. E. J. Am. Chem.
Soc. 2002, 124, 1222-1226.
(12) Lutz, S.; Burgstaller, P.; Benner, S. A. Nucleic Acids Res. 1999, 27, 2972-
2798.
(27) Yu, C.; Henry, A. A.; Romesberg, F. E.; Schultz, P. G. Angew. Chem., Int.
Ed. 2002, 41, 3841-3844.
9
10.1021/ja049961u CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 6923-6931
6923

A R T I C L E S
Henry et al.
Chart 1
Scheme 1
In the course of characterizing these non-H-bonding unnatural
nucleobases, we identified several promising pairs formed
between two of the same analogues. The use of such “self-
pairs” is not a limitation since the addition of a single self-pair
to the genetic alphabet would create 61 new codons. Moreover,
a self-pair is advantageous because the potential for mispairing
with the natural nucleobases is significantly reduced. Unfortu-
nately, though several self-pairs have been identified that are
thermally selective in duplex DNA and selectively synthesized
by the Klenow fragment of Escherichia coli DNA polymerase
I (Kf), the nascent self-pairs have been in general poor substrates
for Kf, and synthesis continues very inefficiently. Apparently,
the physiochemical properties of the nucleobase analogues
necessary for stability and enzymatic base pair synthesis are
not sufficient for extension.
The unnatural nucleobase properties that may be important
for polymerase recognition and efficient replication include
aromatic surface area and heteroatom substitution. For example,
the large surface area of many previously reported unnatural
nucleobases may cause the nascent base pair to adopt a thermally
stable non-Watson-Crick-like structure, favoring synthesis but
also leading to a primer terminus with a geometry that is not
recognized by the polymerase. Heteroatom substitution may be
critical for continued synthesis by affecting the dipole moment
and polarizability of the nucleobase. Recently, we found that
heteroatom-substituted nucleobases of intermediate size may
form thermally stable pairs in duplex DNA, despite their reduced
aromatic surface area, and that these unnatural base pairs may
be synthesized and extended by DNA polymerases with
improved efficiency.28,29 Here we further explore this approach
using a simpler scaffold and report the synthesis and charac-
terization of six novel fluorine-substituted benzene nucleobase
analogues (Chart 1). The aromatic carbon-fluorine bonds are
expected to impart a significant dipole moment to each
nucleobase without introducing H-bonding functionality that
would be complementary to the natural bases. One of the self-
pairs, formed between two 3-fluorobenzene nucleobase ana-
logues, is found to be thermally selective and efficiently
synthesized and extended by Kf, making it the most promising
unnatural base pair identified to date.
coupling to the disiloxane-protected lactone (Scheme 1), based
on a procedure in the literature for simple aromates not
containing fluorine atoms.30,31 In the case of the o-fluoro-
bromobenzene derivatives, it was essential to add the haloaro-
mate to the n-butyllithium, as the conventional reverse manner
gave very low yields. Treatment with triethylsilane and strong
Lewis acid resulted in reduction to give the protected aryl
C-nucleoside in moderate yield; the kinetic preference for
hydride attack on the planar C1′ carbocation from below the
ring30,31 was manifested in the â-anomer being the major product
of the reaction. The minor R-anomer was readily removed by
flash column chromatography. Deprotection was achieved by
1
treatment with TBAF. In all cases, the 1-D H NMR splitting
pattern for H1′ conformed to literature precedents in which
â-stereochemistry is indicated by a doublet of doublets for
C-nucleosides.30 In each case, the conversion of free nucleoside
to phosphoramidite or triphosphate was accomplished according
to standard literature procedures.32,33
2.2. Preliminary Structural Characterization of DNA
Containing Unnatural Self-Pairs. To determine whether the
presence of any unnatural self-pair in a short DNA duplex
perturbed its overall structure, each unnatural nucleotide was
incorporated, at position X, into the complementary oligonucle-
otides 5′-d(GCGATGXGTAGCG) and 5′-d(CGCTACXCAT-
CGC). Circular dichroism (CD) spectra were collected for each
unnatural duplex, as well as for the duplexes containing the
correct pairs dA:dT or dG:dC or the mispair dT:dT. The spectra
of the unnatural duplexes are virtually identical and are within
the extrema set by the fully natural duplexes, indicating that
the unnatural base pairs do not significantly distort the B-form
duplex (Figure 1).
2. Results
More detailed structural characterization of a short DNA
duplex containing the 3FB self-pair has also been initiated. This
particular base pair was chosen because of its favorable
2.1. Nucleobase Analogue Synthesis. Tetraisopropyl di-
siloxane-protected nucleosides were synthesized by aryllithium
(28) Matsuda, S.; Henry, A. A.; Schultz, P. G.; Romesberg, F. E. J. Am. Chem.
Soc. 2003, 125, 6134-6139.
(29) Henry, A. A.; Yu, C.; Romesberg, F. E. J. Am. Chem. Soc. 2003, 125,
9638-9646.
(30) Wichai, U.; Woski, S. A. Org. Lett. 1999, 1, 1173-1175.
(31) Wichai, U.; Woski, S. A. Bioorg. Med. Chem. Lett. 1998, 8, 3465-3468.
(32) Schweitzer, B. A.; Kool, E. T. J. Am. Chem. Soc. 1995, 117, 1863-1872.
(33) Kovacs, T.; Otvos, L. Tetrahedron Lett. 1988, 29, 4525-4528.
9
6924 J. AM. CHEM. SOC. VOL. 126, NO. 22, 2004

A Replicable Unnatural Base Pair
A R T I C L E S
Figure 1. CD spectra of duplex DNA [5′-d(GCGATGXGTAGCG):3′-
d(CGCTACXCATCGC)] containing each unnatural self-pair, each natural
correct pair, or a dT:dT mismatch at the position labeled X. See
Experimental Section for details.
replication properties (see below). The oligonucleotides 5′-
d(G1C2A3G4A53FB6C7G8A9T10C11C12) and 5′-d(G13G14A15-
T16C17G183FB19T20C21T22G23C24) were synthesized, and the
resulting duplex was characterized by NMR spectroscopy. This
sequence context was chosen for its similarity to that used in
the kinetic assays. All proton resonances associated with the
base moieties, as well as H1′ and H2′/H2′′ of the deoxyribose
rings, were assigned by 2D NOESY and DQF-COSY spectra
recorded in D2O and H2O at several temperatures and buffer
conditions. Assignments of the resonances of the two 3FB bases
Figure 2. NMR characterization of duplex DNA containing the 3FB self-
pair. (a) DNA duplex sequence with residue numbering, (b) unnatural
nucleotide proton numbering guide, and (c) aromatic proton to H1′ region
of a 2D NOESY of 0.6 mM duplex in 8.5 mM sodium phosphate, pH 7.0,
134 mM NaCl, 0.25 mM EDTA, and 90%/10% H2O/D2O recorded at 295
K with a mixing time of 200 ms. Boxes represent NOE cross-peaks between
H2 of dA5 and H6, H2 and H5 of d3FB6. Intranucleotide connectivities are
labeled by the residue number. d3FB and dA H2 resonances are shown as
broken lines. See Experimental Section for details.
1
were further confirmed by H-13C HSQC spectra.
In general, all observations are consistent with an overall
standard B-form DNA structure.34,35 This conclusion is sup-
ported by characteristic aromatic to H1′ (Figure 2), aromatic to
H2′/H2′′, and aromatic to dT methyl NOE connectivities, as well
as the intensities of the H1′ to H2′/H2′′ NOE cross-peaks (data
not shown). Imino to imino NOE connectivities and imino to
adenine H2 connectivites are observed along the center of the
DNA helix from base pair dC2-dG23 to dA5-dT20, and from dG8-
dC17 to dC11-dG14 (data not shown). The amino protons of all
dC residues were assigned, and NOE cross-peaks to the imino
protons (dG residues) confirm Watson-Crick base pairing along
the entire length of the DNA duplex, including the dA5-dT20
base pair adjacent to the d3FB6-d3FB19 base pair. The imino
resonances of the other adjacent base pair, dC7-dG18, are broad
presumably because of enhanced solvent exchange. This may
be the result of a base pair stacking defect, or it may simply be
due to the two relatively small 3FB nucleobases incompletely
occupying the space that is normally occupied by a purine-
pyrimidine base pair. The 3FB nucleobases do not interrupt the
aromatic to H1′ (Figure 2) and aromatic to H2′/H2′′ (data not
shown) walk along the B-DNA helix. These observations and
the NOE cross-peaks observed between the dT20 CH3 protons
and the d3FB19 H6 or H2 proton (data not shown; H6 and H2
resonances are degenerate in chemical shift), between dA5 H2
and d3FB6 H6 and H2 (Figure 2, boxes), and between dA5 H2
and d3FB6 H5 (Figure 2, boxes), suggest that both fluorobenzene
bases are positioned within the DNA helix. However, NOE
cross-peaks from d3FB6 H2 and H6 to d3FB6 H2′/H2′′ and to
dA5 H2′/H2′′ suggest that the 3FB6 nucleobase base exists within
the duplex in two orientations that are related by rotation about
the glycosidic bond.
2.3. Stability of Unnatural Self-Pairs. To evaluate the
thermodynamic stability of DNA duplexes containing an un-
natural self-pair or mispair, each unnatural and natural nucleotide
was incorporated into the complementary oligonucleotides 5′-
d(GCGATGXGTAGCG) and 5′-d(CGCTACYCATCGC) at
positions X and Y. The melting temperature of each duplex
was determined by thermal denaturation experiments. As used
in this text, the term “stability” is intended to refer only to the
duplex melting temperature (Tm). Each self-pair-containing
duplex is about as stable as a pyrimidine-pyrimidine mispair
(Table 1). There is little dependence of the Tm value on the
number of fluorine substituents or on the substitution pattern:
all self-pair-containing duplexes melt between 52 and 54 °C.
For reference, the stability of this duplex containing dA:dT is
59.2 °C, and the stability of this duplex containing dT:dT is
53.3 °C (Table 1, footnote). The stabilities of all possible
mispairs are less than any self-pair, resulting in a range of
thermal selectivities from 1.9 to 6.7 °C for the 3FB self-pair
and from 4.1 to 10.0 °C for the 3,5DFB self-pair (Table 1).
Mispairs with dA are consistently the most stable, followed
closely by those with dT (except 3,4DFB for which the dT
mispair is slightly more stable). Mispairs involving dG and dC
(34) Hare, D. R.; Wemmer, D. E.; Chou, S.-H.; Drobny, G.; Reid, B. R. J. Mol.
Biol. 1983, 171, 319-3368
(35) Wuethrich, K. NMR of Proteins and Nucleic Acids; John Wiley & Sons:
New York, 1986.
9
J. AM. CHEM. SOC. VOL. 126, NO. 22, 2004 6925

A R T I C L E S
Henry et al.
Table 1. Tm Values for Duplexes Containing Unnatural Base
Table 3. Incorporation of Natural and Unnatural Triphosphates
Pairsa
Opposite Unnatural Bases in the Templatea
5′-d(TAATACGACTCACTATAGGGAGA)
5′-d(GCGTACXCATGCG)
3′-d(CGCATGYGTACGC)
3′-d(ATTATGCTGAGTGATATCCCTCTXGCTAGGTTACGGCAGGATCGC)
X
Y
Tm (°C)
X
Y
Tm (°C)
kcat
(min-1
KM
(µM)
k
cat/KM
X
triphosphate
)
(min-1M-1
)
2FB
2FB
54.0
52.0
46.0
46.0
49.0
52.0
50.2
45.4
45.3
48.4
53.2
49.0
43.2
44.3
47.1
2,3DFB
2,3DFB
53.0
49.1
45.2
45.1
48.3
53.0
50.1
47.2
45.3
51.1
53.0
49.1
46.2
45.2
48.3
2.1 × 106
5.4 × 105
2.1 × 103
1.8 × 103
9.5 × 104
1.1 × 105
1.6 × 105
1.3 × 105
3.2 × 103
8.6 × 104
2.0 × 105
1.5 × 103
5.7 × 103
2.2 × 105
1.4 × 105
6.5 × 104
1.2 × 104
1.8 × 105
1.0 × 106
<1.0 × 103
2.1 × 103
8.6 × 103
4.8 × 105
3.1 × 103
1.5 × 103
1.2 × 105
A
C
G
T
A
C
G
T
3FB
3FB
A
C
G
T
3,5DFB
A
C
G
T
A
C
G
T
A
C
G
T
A
C
G
T
A
C
G
T
25 ( 4
26 ( 4
12 ( 2
48 ( 10
195 ( 35
168 ( 80
158 ( 23
218 ( 8
31 ( 16
4 ( 3
0.4 ( 0.1
0.3 ( 0.1
15 ( 8
3FB
3FB
3,4DFB
3,4DFB
A
C
G
T
A
C
G
T
3,5DFB
25 ( 5
5 ( 2
0.5 ( 0.2
0.34 ( 0.31
16 ( 7
106 ( 33
187 ( 94
41 ( 13
334 ( 153
123 ( 43
79 ( 13
51 ( 10
6 ( 4
3,5DFB
3,5DFB
2,3,4TFB
2,3,4TFB
A
C
G
T
A
C
G
T
2,3,4TFB
2,3DFB
2FB
8 ( 2
0.5 ( 0.2
0.7 ( 0.3
17 ( 8
7 ( 2
a
0.39 ( 0.02
1.6 ( 0.5
9 ( 3
Uncertainty in values is less than 0.1 °C. The duplex containing a
dA:dT correct pair has a Tm value of 59.2 °C, and the one containing a
dT:dT mismatch has a Tm value of 53.3 °C. See Experimental Section for
details.
131 ( 28
49 ( 17
24 ( 5
25 ( 9
ndb
ndb
Table 2. Rates of Correct Extension of Unnatural Self-Pairs and
3FB Mispairsa
0.16 ( 0.01
1.4 ( 0.9
30 ( 9
0.8 ( 0.2
0.27 ( 0.20
20 ( 14
77 ( 22
163 ( 39
63 ( 23
255 ( 93
184 ( 95
164 ( 52
3,4DFB
5′-d(TAATACGACTCACTATAGGGAGAX)
3′-d(ATTATGCTGAGTGATATCCCTCTYGCTAGGTTACGGCAGGATCGC)
kcat
KM
k
cat/KM
X
Y
(min-1
)
(µM)
(min-1 M-1
)
a
b
See Experimental Section for details. Reaction was too inefficient
for kcat and KM to be determined independently.
3FB
3FB
28 ( 4
25 ( 6
3.5 ( 0.7
4.1 ( 1.5
5 ( 1
0.5 ( 0.3
17 ( 4
4.2 ( 0.7
ndb
85 ( 24
125 ( 29
289 ( 59
127 ( 31
254 ( 55
242 ( 75
268 ( 60
310 ( 80
ndb
132 ( 60
237 ( 48
219 ( 54
ndb
3.3 × 105
2.0 × 105
1.2 × 104
3.2 × 104
2.0 × 104
2.1 × 103
6.3 × 104
1.4 × 104
<1.0 × 103
1.9 × 105
1.1 × 104
3.7 × 104
<1.0 × 103
1.1 × 106
3,5DFB
2,3,4TFB
2,3DFB
2FB
3,4DFB
A
3,5DFB
2,3,4TFB
2,3DFB
2FB
3,4DFB
3FB
3FB
3FB
3FB
A
103 M-1 min-1, due to both a low kcat and a high KM. The three
nucleobases having a fluorine substitutent ortho to the glycosidic
bond (2FB, 2,3DFB, and 2,3,4TFB) form self-pairs that are
extended with intermediate efficiency, approximately 104 M-1
min-1, due largely to an increase in kcat, relative to the 3,4DFB
self-pair. The nucleobases having only meta-fluorine substituents
(3FB and 3,5FB) are extended very efficiently because of both
an increase in kcat and a decrease in KM. The 3FB and 3,5DFB
C
G
T
25 ( 4
2.5 ( 0.4
8 ( 1
3FB
3FB
3FB
3FB
C
G
T
ndb
self-pairs are extended with an efficiency (>105 M-1 min-1
that is only approximately 100-fold reduced relative to natural
synthesis in the same sequence context.
)
42 ( 1
40 ( 14
a
b
See Experimental Section for details. Reaction was too inefficient
for kcat and KM to be determined independently.
Because mispair extension rates also contribute to overall
replication fidelity, we determined the rate at which each
possible mispair between 3FB and a natural base is extended,
using the same assay and the appropriate primer-template
combinations (Table 2). The most efficiently extended mispairs
involve dT. With dT at the primer terminus, paired opposite
3FB in the template, dCTP is incorporated opposite dG with
an efficiency of 1.1 × 106 M-1 min-1; the mispair formed with
3FB at the primer terminus and dT in the template is extended
with an efficiency of 1.9 × 105 M-1 min-1. All other mispairs
in either context are extended with efficiencies less than 6.4 ×
104 M-1 min-1 (Table 2).
Because of the efficient and selective extension of the 3FB
self-pair, we characterized the efficiency and selectivity of its
Kf-mediated synthesis. For comparison, we also characterized
the efficiency and fidelity of 3,5FB self-pair synthesis, which
is also efficiently extended (Table 3). Both self-pairs are
synthesized with a kcat value of 25 min-1. However, the d3FBTP
substrate is bound 20-fold more tightly by the polymerase-
DNA binary complex, resulting in more efficient self-pair
result in duplexes of similar stability that have significantly
reduced Tm values relative to duplexes containing mispairs
involving dA and dT.
2.4. Polymerase-Mediated Replication of the Unnatural
Self-Pairs. As mentioned above, continued synthesis after
incorporation of the unnatural base pair has consistently limited
the replication of unnatural DNA. Here we refer to this step as
self-pair extension. Thus, to determine the effect of fluorine
substitution on self-pair extension step, each unnatural nucleotide
was first incorporated at positions X and Y, respectively, into
the primer oligonucleotide 5′-d(TAATACGACTCACTAT-
AGGGAGAX) and the template oligonucleotide 5′-d(CGCT-
AGGACGGCATTGGATCGYTCTCCCTATAGTGAGTCG-
TATTA). Each primer-template assembly was used as a
substrate in a single nucleotide (dCTP) incorporation assay with
Kf polymerase. Unlike base pair stability, fluorine substitution
had a pronounced effect on the self-pair extension rate, which
varies by more than 2 orders of magnitude (Table 2). The
3,4DFB self-pair is extended with the lowest efficiency, 2.1 ×
9
6926 J. AM. CHEM. SOC. VOL. 126, NO. 22, 2004

A Replicable Unnatural Base Pair
A R T I C L E S
Table 4. Incorporation of Unnatural Triphosphates Opposite
for the poor self-pair extension rates: one is interstrand
intercalation due to large aromatic surface area, and the other
is the absence of suitably positioned heteroatoms to provide
the nucleobases with optimal physiochemical properties such
as H-bond acceptors, dipole moment, or polarizability. For
example, heteroatom derivatization of the isocarbostyril scaffold
changed the dipole moment of the nucleobase and lead to an
improved self-pair extension rate.27
To explore the potential of heteroatom derivatization of a
small nucleobase scaffold, we have begun a systematic analysis
of the stability and polymerase recognition of self-pairs and
heteropairs formed between two derivatized phenyl rings.
Importantly, this scaffold is expected to abrogate interstrand
intercalation by virtue of its reduced size. In this article, we
report the synthesis and characterization of six unnatural self-
pairs formed between nucleotides bearing fluorine-substituted
benzene nucleobase analogues.
3.1. Stability and Thermal Selectivity of Unnatural Self-
Pairs. Each self-pair is accommodated in B-form duplex DNA
with reasonable stability; however, there is a surprising lack of
dependence on the extent and pattern of fluorine substitution.
All of the self-pair-containing duplexes denature between 52
and 54 °C, approximately 6 °C lower than a duplex with
dA:dT in the same sequence context, and similar to a duplex
containing a dT:dT mismatch. The stability of the parent phenyl
self-pair is 52.8 °C,36 demonstrating that the addition of fluorine
substituents results in little or no stabilization of the self-pairs.
Despite their only moderate stability, the self-pairs are formed
with good selectivity because of the destabilization of all
possible mispairs with natural bases. In this respect, the most
discriminating unnatural nucleobase analogue is 2FB as the Tm
values range from 46 to 52 °C. The least discriminating
nucleobase analogue is 3FB, which nonetheless retains thermal
selectivity of 1.9 to 6.7 °C, typical of that observed among the
natural pairs.
It is interesting to speculate about the nucleobase character-
istics that may contribute to the self-pair and mispair stabilities,
including structure, permanent electrostatic moments, polariz-
ability, and hydrophobicity. Neither the specific fluorine
substitution pattern nor the extent of fluorine substitution appears
to affect the stability of either the self-pairs or the mispairs.
Thus, it seems unlikely that the unnatural bases engage in
specific electronic or structural interbase interactions within
duplex DNA or that the polarizability of π-electron density
within the unnatural nucleobase analogue is important. However,
the mispair stabilities do roughly follow the hydrophobicity of
the natural nucleobases as thymine is the most hydrophobic,
followed closely by adenine, and next cytosine, then guanine.
Thus, forced desolvation of the natural nucleobases appears to
be the dominant factor in the thermal selectivity of the unnatural
self-pairs, which do not require desolvation.
Natural Bases in the Templatea
5′-d(TAATACGACTCACTATAGGGAGA)
3′-d(ATTATGCTGAGTGATATCCCTCTNGCTAGGTTACGGCAGGATCGC)
kcat
(min-1
KM
(µM)
k
cat/KM
N
triphosphate
)
(min-1M-1
)
A
3FB
3,5DFB
3FB
3,5DFB
3FB
3,5DFB
3FB
6 ( 1
14 ( 3
4.3 × 105
2.3 × 104
<1.0 × 103
5.5 × 103
7.4 × 105
2.2 × 105
1.8 × 105
5.0 × 104
5 ( 2
215 ( 4
C
G
T
ndb
ndb
11 ( 3
19 ( 11
45 ( 27
33 ( 24
241 ( 106
0.06 ( 0.01
14 ( 5
10 ( 5
6 ( 2
3,5DFB
12 ( 3
a
b
See Experimental Section for details. Reaction was too inefficient
for kcat and KM to be determined independently.
synthesis: 2.1 × 106 M-1 min-1, compared to 1.1 × 105 M-1
min-1 for the 3,5DFB self-pair. Remarkably, the 3FB self-pair
is synthesized only 20-fold less efficiently than is a natural base
pair in the same sequence context. The fidelity of self-pair
synthesis was determined by measuring the rate of incorporation
of each natural dNTP opposite the unnatural base in the
template, and comparing these with the value for self-pair
synthesis efficiency. In the template, 3,5DFB is somewhat
promiscuous, directing Kf to insert dATP, dCTP, and dTTP with
an efficiency of roughly 105 M-1 min-1. In contrast, 3FB in
the template is less prone to mispair synthesis, directing Kf to
insert dATP and dTTP with efficiencies of 5.4 × 105 M-1 min-1
and 9.5 × 104 M-1 min-1, respectively, and dCTP and dGTP
with efficiencies of only 2.1 × 103 M-1 min-1 and 1.8 × 103
M-1 min-1, respectively.
Neither d3FBTP nor d3,5DFBTP are inserted opposite a
natural base in the template at a rate competitive with that of
natural Watson-Crick pair synthesis (Table 4). d3FBTP is
inserted opposite the natural bases in the template with rates
between 1.8 × 105 and 7.4 × 105 M-1 min-1, except for dC in
which case insertion is not detectable (<103 M-1 min-1).
Insertion of d3,5DFBTP is also least efficient opposite dC (5.5
× 103 M-1 min-1) and otherwise varies between 2.3 × 104 M-1
min-1 and 2.2 × 105 M-1 min-1. Thus, each unnatural
nucleotide is selective for self-pair synthesis both in the template
and as a triphosphate.
3. Discussion
Efforts to expand the genetic alphabet are based on developing
a thermally stable and replicable unnatural base pair that is
orthogonal to the natural dG:dC and dA:dT base pairs. There
is no reason why such an unnatural base pair must resemble a
natural base pair in any way. The only constraints are
functional: the unnatural base pair must not significantly
destabilize duplex DNA, and it must be efficiently and
selectively replicated by a DNA polymerase. With this in mind,
we have synthesized and characterized a wide variety of the
self-pairs and heteropairs formed between nucleobase analogues
with different structures and substituents.22,23,26-29
Several promising self-pairs based on the azaindole, isocar-
bostyril, or naphthyl scaffolds have been identified.21,24,25 These
self-pairs are thermally selective in duplex DNA and are
enzymatically synthesized with reasonable efficiency and
selectivity. However, after insertion of the unnatural triphosphate
opposite its partner in the template, continued primer extension
is inefficient in each case. There are at least two possible reasons
3.2. Replication Properties of Unnatural Self-Pairs. Kinetic
Orthogonality of Unnatural Bases in Template DNA. Rep-
lication fidelity requires that the unnatural base in the template
not direct the efficient misinsertion of any natural substrate
(dNTP). The mispair synthesis rates for each unnatural base in
the template reveal that not only are the unnatural mispairs with
dA and dT the most thermally stable (see above), but they are
also the most efficiently synthesized. In each case, efficient
(36) Matsuda, S.; Romesberg, F. E. Unpublished results.
9
J. AM. CHEM. SOC. VOL. 126, NO. 22, 2004 6927

A R T I C L E S
Henry et al.
Table 5. Overall Fidelity of 3FB Self-Pair Replication
synthesis is the result of a large kcat value. The correlation
between mispair stability and kcat implies that the energetic cost
of desolvation, argued above to be the most important deter-
minant of stability in this series, is also a significant component
of the activation barrier for dNTP insertion.
correct pair/mispair
(primer:template)
synthesis
fidelitya,c
extension
fidelityb,c
total
fidelityd
dA:dT/d3FB:dT
dT:dA/d3FB:dA
dC:dG/d3FB:dG
dG:dC/d3FB:dC
d3FB:d3FB/dA:d3FB
d3FB:d3FB/dT:d3FB
d3FB:d3FB/dC:d3FB
d3FB:d3FB/dG:d3FB
222
93
54
36
3636
>40 000
1081
5
2
24
>330
7992
338 148
>2 160 000
>43 240 000
20
The only unnatural nucleotide that does not direct the
preferential insertion of both dATP and dTTP is 2FB. In the
template, 2FB directs Kf to insert dTTP with a kcat/KM of only
8.6 × 103 M-1 min-1. While this is significantly more efficient
than the insertion of dGTP and dCTP, it is more than 100-fold
less efficient than the insertion of dATP. The reduced insertion
efficiency of dTTP results from both a decrease in kcat and an
increase in KM. The preferential insertion of dATP opposite 2FB
is reminiscent of the efficiency with which other unnatural
nucleobase scaffolds bearing methyl groups alpha to the
glycosidic bond direct Kf to insert dATP.24 Previously, it was
speculated that the methyl group was oriented in the developing
minor groove, where it participated in stabilizing van der Waals
interactions with the adenine methine. Perhaps the ortho-fluorine
atom of 2FB is also well accommodated both in the ternary
complex with dATP and its corresponding transition state.
>40 000
4
22
44
1000
1167
24 000
>385 110
a
Ratio of synthesis efficiencies (Tables 3 and 4) for correct pair and
b
mispair shown in first column of current table. Ratio of extension
efficiencies (Table 2) for correct pair and mispair shown in first column of
c
current table. Using a specificity constant for natural synthesis of 4.7 ×
d
107 min-1 M-1 21
.
Product of values in second and third columns of
current table.
by an ortho- and/or para-fluorine, the turnover rate is modest
(0.5-4.1 min-1). However, when not accompanied by an ortho-
or para-fluorine, as in 3FB and 3,5DFB, the turnover rate is
remarkably high (∼25 min-1). In fact, these kcat values are only
approximately 6-fold reduced relative to natural synthesis
(approximately 160 min-1). These data support two possible
hypotheses. First, that there exist strong anisotropic interactions
between the incoming dNTP and the polymerase-DNA com-
plex that develop upon proceeding from the polymerase-
duplex-dNTP ternary complex to its transition state. Second,
that there exist subtle yet functionally important differences in
the structure of the primer terminus for the different self-pairs.
3.4. The 3FB Self-Pair and the Effort To Expand the
Genetic Alphabet. On the basis of thermal stability data and
CD spectra, the unnatural 3FB nucleobase appears to selectively
form a self-pair in an undistorted B-form DNA duplex of a given
sequence. This interpretation is further supported by the NMR
characterization of a different 3FB self-pair-containing DNA
duplex, one that closely resembles the sequence used in the
kinetic assays. Preliminary data indicate a B-form conformation
with Watson-Crick base pairing along the entire length of the
DNA duplex, including the dA:dT base pair adjacent to the 3FB
self-pair, although the other flanking base pair may be somewhat
more dynamic. The molecular details of the effect of incorporat-
ing 3FB self-pairs into additional sequences is currently being
studied by both NMR and X-ray crystallography. These studies
will be of paramount importance in defining structure-function
relationships that would facilitate optimization of the unnatural
base pairs. For now, it is interesting to speculate that efficient
single nucleotide extension of the 3FB self-pair is related to
the fact that it does not perturb a DNA duplex or, by extension,
the primer-template terminus where it is recognized by the DNA
polymerase.
A second exception to the preferential synthesis of mispairs
involving dA and dT is the relatively efficient insertion of dCTP
opposite 2,3DFB (6.5 × 104 M-1 min-1), and especially 3,-
5DFB (1.3 × 105 M-1 min-1). The latter is as efficient as dATP
incorporation opposite 3,5DFB (1.6 × 105 M-1 min-1) and more
efficient than dTTP incorporation opposite 3,5DFB (8.6 × 104
M-1 min-1). Efficient insertion of dCTP in these cases results
largely from a decrease in KM which, at approximately 5 µM,
is similar to that of natural base pair synthesis in the same
sequence context. 2,3DFB and 3,5DFB each have meta-fluorine
substitution that may facilitate dCTP binding. However, 3FB,
3,4DFB, and 2,3,4DFB, which also possess meta-fluorine
substitutents, direct dCTP insertion very inefficiently (∼103 M-1
min-1), indicating that there is a complex role for the substitution
pattern in dCTP insertion.
Overall, the unnatural nucleobases in the template resist
insertion of dGTP and dCTP. Each member of the series directs
the insertion of dATP and dTTP more efficiently. Nonetheless,
all of the mispair synthesis rates are slow relative to natural
synthesis and would not be expected to compromise the
replication fidelity of an efficiently synthesized and extended
unnatural self-pair.
3.3. Extension of Unnatural Self-Pairs. As discussed above,
continued enzymatic polymerization after unnatural self-pair
synthesis has limited replication of the previously described
unnatural base pairs. The KM values for extension of the
fluorobenzene self-pairs vary by only 3.4-fold. dCTP is bound
most tightly (85 µM) by the enzyme-DNA complex containing
the 3FB self-pair. This value indicates only 20-fold weaker
binding than in natural DNA synthesis. The addition of a second
meta-fluorine substitutent results in a slight decrease in dCTP
binding (125 µM). Both ortho- and para-fluorine substitution
within the self-pair also reduced the affinity of the polymerase-
duplex complex for dCTP. There appears to be no correlation
between these KM values and self-pair stability.
The 3FB self-pair is well-replicated by Kf polymerase,
including efficient and selective synthesis, and most notably
efficient and selective extension. In fact, the efficiency of self-
pair synthesis and extension are only approximately 20-fold and
100-fold reduced, respectively, relative to natural DNA syn-
thesis. The most significant fidelity problems result from the
insertion of dATP and dTTP opposite 3FB in the template and
extension of these mispairs. This occurs with overall efficiencies
that are 20- and 44-fold reduced, respectively, relative to self-
pair replication (Table 5). Compared to mispairs involving dG
and dC, the fidelity of 3FB self-pair replication is in excess of
24 000-fold (Table 5). We also point out that enzymatic
In contrast to its effect on substrate binding, the nature of
the unnatural base has a pronounced effect on the turnover rate
associated with self-pair extension: the kcat values vary by more
than 50-fold. When a meta-fluorine substitutent is accompanied
9
6928 J. AM. CHEM. SOC. VOL. 126, NO. 22, 2004

A Replicable Unnatural Base Pair
A R T I C L E S
1
Compound 3a. H NMR (CDCl3): δ 7.99 (2H, d, J ) 6.6 Hz),
7.70 (2H, d, J ) 6.6 Hz), 7.27-7.16 (4H, m), 7.01-6.96 (2H, m),
6.77-6.65 (1H, m), 6.45-6.40 (1H, m), 5.61-5.58 (1H, m), 5.35 (1H,
dd, J ) 7.8, 5.2 Hz), 4.75-4.62 (1H, m), 4.59 (2H, s), 3.06-2.82 (1H,
m), 2.41 (3H, s), 2.39 (3H, s), 2.40-2.30 (1H, m). 13C NMR (CDCl3):
δ 166.5, 166.3, 165.9, 165.6, 160.9, 160.7, 145.3, 144.5, 144.1, 129.9,
129.4, 127.2, 127.1, 109.1, 108.5, 103.2, 83.5, 79.9, 64.7, 41.8, 21.9.
HRMS (MALDI-FTMS): calcd for C27H25F2O5 (MH+), 466.1593;
found, 466.1588.
proofreading, absent in the Kf mutant used in these experiments,
is predicted to contribute to fidelity because of the lower thermal
stability of the mispairs with dA and dT relative to the unnatural
self-pair.25
Despite its lack of H-bonding and shape complementarity,
the 3FB self-pair appears to have all of the properties required
of a third base pair, including those associated with structure,
stability, and replication. This argues that shape complementarity
and H-bonding are not unique in their ability to control the
specific interbase interactions required for DNA stability and
replication. It seems likely that through continued derivatization,
optimal hydrophobic and van der Waals forces will result in a
third base pair with stability and polymerase recognition
sufficient to expand the genetic alphabet.
1
Compound 4a. H NMR (CDCl3): δ 7.22-7.16 (1H, m), 7.01-
6.92 (2H, m), 5.25 (1H, t, J ) 7.0 Hz), 4.47-4.37 (1H, m), 4.05 (1H,
dd, J ) 11.5, 2.9 Hz), 3.91-3.74 (2H, m), 2.43-2.33 (1H, m), 2.07-
1.97 (1H, m), 1.07-0.94 (28H, m). 19F NMR (CDCl3): δ -139.8 (1F,
m), -144.5 (1F, m). 13C NMR (CDCl3): δ 124.3, 124.1, 122.0, 122.0,
116.3, 116.0, 85.9, 73.0, 72.4, 63.1, 41.9, 17.8, 17.7, 17.6, 17.5, 17.3,
17.2, 13.7, 13.6, 13.2, 12.8. HRMS (MALDI-FTMS): calcd for
C23H38F2O4Si2Na (MNa+), 495.2169; found, 495.2169.
4. Experimental Section
Compound 5a. 1H NMR (CDCl3): δ 7.26-7.07 (3H, m), 5.04 (1H,
t, J ) 7.1 Hz), 4.54-4.46 (1H, m), 4.09 (1H, t, J ) 4.4 Hz), 3.89 (1H,
dd, J ) 14.6, 7.0 Hz), 2.44-2.31 (1H, m), 2.17-1.92 (1H, m), 1.08-
0.87 (28H, m). 13C NMR (CDCl3): δ 121.9, 117.4, 117.1, 115.2, 114.8,
86.6, 63.4, 43.3, 17.9, 17.7, 17.6, 17.5, 17.3, 17.2, 13.8, 13.6, 13.3,
12.8. 19F NMR (CDCl3): δ -138.11 (1F, ddd, J ) 20.7, 11.6, 6.9 Hz),
-140.28 (1F, ddd, J ) 17.1, 7.9, 5.3 Hz). HRMS (MALDI-FTMS):
calcd for C23H38F2NaO4Si2 (MNa+), 495.2174; found, 495.2169.
General Synthetic Methods. All reactions were carried out in oven-
dried glassware under inert atmosphere unless otherwise stated. All
solvents were dried over 4 Å molecular sieves except dicloromethane
(distilled from CaH2), tetrahydrofuran (distilled from sodium and
potassium metal), and diethyl ether (distilled from LiAlH4). All other
reagents were purchased from Aldrich. High-resolution mass spectro-
scopic data were obtained from the facilities at The Scripps Research
Institute. All small molecule NMR spectra were collected by means of
a Varian Mercury 200 MHz spectrometer (13C 50.3 MHz; 19F 188.2
1
Compound 6a. H NMR (CDCl3): δ 7.19-7.11 (1H, m), 6.90-
1
6.76 (1H, m), 5.29 (1H, t, J ) 6.8 Hz), 4.46-4.36 (1H, m), 4.03 (1H,
dd, J ) 11.7, 3.1 Hz), 3.91-3.75 (2H, m), 2.45-2.32 (1H, m), 2.04-
1.91 (1H, m), 1.00-0.94 (28H, m). 19F NMR (CDCl3): δ -137.6 (1F,
m), -140.2 (1F, m), -161.7 (1F, m). 13C NMR (CDCl3): δ 120.6,
112.2, 111.7, 85.7, 72.6, 72.0, 62.9, 41.7, 31.9, 22.9, 17.8, 17.6, 17.6,
17.4, 17.3, 17.2, 13.7, 13.6, 13.2, 12.8. HRMS (MALDI-FTMS): calcd
for C23H37F3O4Si2Na (MNa+), 513.2080; found, 513.2077.
MHz; 31P 81 MHz). The H and 13C chemical shifts are referenced
relative to TMS, and the 31P and 19F chemical shifts are referenced
relative to 85% phosphoric acid in D2O and neat trichlorofluoromethane,
respectively.
Representative Procedure for Synthesis of Tetraisopropyldisi-
loxanediyl-Protected Nucleosides. To a solution of n-butyllithium (1.6
M in hexane, 2.8 mL) at -78 °C was added dropwise a solution of
1-bromo-2-fluorobenzene (0.467 mL, 4.27 mmol) in THF (7.5 mL).
After being stirred at this temperature for 40 min, a solution of 3′,5′-
O-((1,1,3,3-tetraisopropyl)disiloxanediyl)-2′-deoxy-D-ribono-1,4-lac-
tone (1.0 g, 2.67 mmol) in THF (7.5 mL) was added via cannula. After
being stirred for 1 h, the reaction was quenched with saturated aqueous
NH4Cl, extracted with diethyl ether, washed with saturated aqueous
NH4Cl and brine, and dried over MgSO4. After concentration in vacuo,
the resulting crude oily product (0.8 g, 1.69 mmol) was dissolved in
CH2Cl2 (6 mL) and cooled to -78 °C after which was added Et3SiH
(0.81 mL, 5.06 mmol) followed by BF3-OEt2 (0.64 mL, 5.06 mmol),
added dropwise. After being stirred at -78 °C for 7 h, the reaction
was quenched by the addition of aqueous NaHCO3 (10 mL). Additional
water was added, the solution was extracted with diethyl ether, and
after washing with aqueous NaHCO3, water, and brine, it was dried
over MgSO4 and evaporated to dryness. Purification by column
chromatography on silica gel (0-20% ethyl acetate in hexane) afforded
the â-anomer of tetraisopropyl disiloxane-protected nucleoside 1a in
16% yield over two steps.
Representative Procedure for Deprotection of Tetraisopropyl-
disiloxanediyl-Protected Nucleosides. To a stirred solution of 4a
(0.134 g, 0.283 mmol) in THF (6 mL) was added dropwise TBAF (1
M in THF, 0.848 mL). After being stirred at room temperature for 45
min, the reaction was quenched with saturated aqueous NH4Cl (5 mL)
and extracted with ether/THF (1:1). NaCl was added to the aqueous
phase, which was further extracted, dried over MgSO4, and evaporated
to dryness. Purification by column chromatography on silica gel (50-
100% ethyl acetate in hexane) afforded free nucleoside 4b in 92% yield.
Compound 1b. 1H NMR (CDCl3): δ 7.58 (1H, dt, J ) 7.6, 1.8 Hz),
7.26-6.98 (3H, m), 5.38 (1H, dd, J ) 10.3, 5.2 Hz), 4.33 (1H, t, J )
3.5 Hz), 3.99-3.74 (1H, m), 3.74 (2H, br s), 2.33-2.22 (1H, m), 1.99-
1.82 (1H, m). 19F NMR (CDCl3): δ -121.64 (1F, t, J ) 5.3 Hz). 13C
NMR (CDCl3): δ 129.0, 127.3, 124.2, 114.9, 114.5, 87.7, 74.2, 74.1,
73.2, 62.8, 42.4. HRMS (MALDI-FTMS): calcd for C11H14FO3
(MH+), 213.0927; found, 213.0924.
1
Compound 2b. H NMR (CDCl3): δ 7.32 (1H, dd, J ) 7.9, 5.9
Hz), 7.15 (2H, d, J ) 7.4 Hz), 7.01-6.89 (1H, m), 5.12 (1H, dd, J )
10.2, 5.4 Hz), 4.32 (1H, dd, J ) 4.2, 1.8 Hz), 4.01-3.93 (1H, m), 3.66
(2H, d, J ) 5.0 Hz), 2.22 (1H, ddd, J ) 13.0, 5.6, 1.6 Hz), 2.00-1.82
(1H, m). 13C NMR (CDCl3): δ 165.6, 145.3, 130.0, 121.7, 114.2, 112.7,
88.1, 79.7, 73.2, 62.9, 43.8. 19F NMR (CDCl3): δ -115.54 (1F, t, J )
5.3 Hz). HRMS (MALDI-FTMS): calcd for C11H13FNaO3 (MNa+),
235.0746; found, 235.0738.
Compound 1a. 1H NMR (CDCl3): δ 8.69 (1H, d, J ) 3.5 Hz), 7.70-
6.99 (3H, m), 5.38 (1H, t, J ) 6.8 Hz), 4.60-4.54 (1H, m), 4.18 (1H,
br d, J ) 10.5 Hz), 4.03-3.94 (2H, m), 2.53-2.47 (1H, m), 2.15-
2.09 (1H, m), 1.13-1.06 (28H, m). 19F NMR (CDCl3): δ -119.1.
HRMS (MALDI-FTMS): calcd for C23H39FO4Si2Na (MNa+), 477.2263;
found, 477.2261.
1
1
Compound 3b. H NMR (CDCl3): δ 7.01-6.80 (2H, m), 6.78-
Compound 2a. H NMR (CDCl3): δ 7.21 (1H, dd, J ) 6.9, 1.3
6.73 (1H, m), 5.11 (1H, t, J ) 7.4 Hz), 4.40 (1H, dd, J ) 6.8, 5.2 Hz),
4.02 (1H, dd, J ) 5.2, 3.4 Hz), 3.70-3.55 (2H, m), 2.82-2.61 (1H,
m), 1.95-1.78 (1H, m). 13C NMR (CDCl3): δ 165.9, 148.1, 136.2,
113.5, 128.9, 102.7, 86.7, 74.8, 64.8, 55.5, 43.3. HRMS (MALDI-
FTMS): calcd for C11H13F2O3 (MH+), 231.0755; found, 231.0750.
Compound 4b. H NMR (CD3OD): δ 7.40-7.34 (1H, m), 7.20-
7.08 (2H, m), 5.37 (1H, dd, J ) 10.2, 5.5 Hz), 4.34-4.31 (1H, m),
3.99-3.92 (1H, m), 3.73-3.64 (2H, m), 2.34-2.25 (1H, m), 2.00-
Hz), 7.06-6.89 (1H, m), 6.84 (1H, t, J ) 7.6 Hz), 4.98 (1H, t, J ) 7.2
Hz), 4.47-4.40 (1H, m), 4.03 (1H, t, J ) 7.4 Hz), 3.82 (2H, br s),
2.31-2.20 (1H, m), 2.01-1.86 (1H, m), 1.13-1.06 (28H, m). 19F NMR
(CDCl3): δ -113.3 (1F, t, J ) 3.9 Hz). 13C NMR (CDCl3): δ 165.6,
160.7, 145.3, 130.1, 121.5, 114.6, 113.1, 86.6, 73.1, 63.7, 43.4, 18.2,
17.8, 17.7, 17.6, 17.5, 17.3, 17.2, 13.8, 13.7, 13.3, 13.2, 12.8. HRMS
(MALDI-FTMS): calcd for C23H39FO4Si2Na (MNa+), 477.2263;
found, 477.2260.
1
9
J. AM. CHEM. SOC. VOL. 126, NO. 22, 2004 6929

A R T I C L E S
Henry et al.
1.84 (1H, m). 19F NMR (CD3OD): δ -138.5 (1F, m), -143.6 (1F, m).
13C NMR (CD3OD): δ 131.9, 124.3, 122.1, 115.9, 115.6, 87.9, 73.8,
73.1, 72.1, 62.7, 42.4. HRMS (MALDI-FTMS): calcd for C11H12F2O3-
Na (MNa+), 253.0647; found, 253.0644.
121.9, 114.7, 114.3, 113.4, 112.9, 86.9, 86.5, 79.6, 74.5, 64.9, 55.5,
44.2. HRMS (MALDI-FTMS): calcd for C32H31FO5Na (MNa+),
537.2048; found, 537.2045.
Compound 2d. 1H NMR (CDCl3): δ 7.51-7.02 (12H, m), 6.98 (1H,
t, J ) 6.2 Hz), 6.87 (4H, d, J ) 8.4 Hz), 5.10 (1H, dd, J ) 10.0, 5.2
Hz), 4.41 (1H, dd, J ) 9.8, 5.5 Hz), 4.16 (1H, br s), 3.71 (6H, s),
3.65-3.40 (4H, m), 3.21 (2H, t, J ) 4.0 Hz), 2.46 (1H, t, J ) 6.2 Hz),
2.32 (1H, t, J ) 6.2 Hz), 2.28-2.01 (1H, m), 2.01-1.81 (1H, m), 1.25-
1
Compound 5b. H NMR (CDCl3): δ 7.31 (1H, dd, J ) 12.0, 8.6
Hz), 7.22-7.11 (2H, m), 5.08 (1H, dd, J ) 10.6, 5.4 Hz), 4.82 (1H, br
s), 4.32 (1H, d, J ) 6.0 Hz), 4.03-3.81 (1H, m), 3.67 (2H, d, J ) 4.6
Hz), 2.25-2.15 (1H, m), 1.95-1.80 (1H, m). 13C NMR (CDCl3): δ
139.9, 122.3, 117.0, 116.7, 115.0, 114.6, 88.1, 79.2, 73.2, 62.8, 43.9.
19F NMR (CDCl3): δ -140.53 (1F, ddd, J ) 23.8, 14.0, 5.3 Hz),
-142.36 (1F, dd, J ) 13.3, 6.6 Hz). HRMS (MALDI-FTMS): calcd
for C11H12F2NaO3 (MNa+), 253.0647; found, 253.0648.
0.95 (14H, m). 19F NMR (CDCl3): δ -113.42 (1F, d, J ) 5.6 Hz). 31
P
NMR (CDCl3): δ 149.16, 149.03. 13C NMR (CDCl3): δ 165.6, 160.8,
158.7, 145.2, 136.2, 130.4, 130.2, 130.0, 128.5, 128.1, 127.0, 121.8,
117.9, 114.7, 114.3, 113.4, 112.9, 86.4, 86.1, 86.0, 79.9, 76.2, 76.1,
75.8, 64.5, 58.8, 58.4, 55.4, 43.6, 43.4, 26.0, 24.8, 24.7, 20.7, 20.6,
20.4. HRMS (MALDI-FTMS): calcd for C41H49FN2O6P (MH+),
715.3312; found, 715.3310.
1
Compound 6b. H NMR (CD3OD): δ 7.40-7.32 (1H, m), 7.13-
6.99 (1H, m), 5.32 (1H, dd, J ) 10.2, 5.9 Hz), 4.34-4.30 (1H, m),
3.97-3.91 (1H, m), 3.71-3.64 (2H, m), 2.33-2.22 (1H, m), 2.00-
1.83 (1H, m). 19F NMR (CD3OD): δ -139.4 (1F, m), -143.2 (1F, m),
-161.1 (1F, m). 13C NMR (CD3OD): δ 121.2, 121.1, 112.2, 112.1,
111.8, 111.7, 87.9, 73.5, 73.0, 62.7, 42.3. HRMS (MALDI-FTMS):
calcd for C11H11F3O3Na (MNa+), 271.0552; found, 271.0559.
General Procedure for Phosphoramidite Synthesis. To a solution
of free nucleoside (0.6 mmol) coevaporated with toluene (10 mL) was
added pyridine (2.5 mL), followed by DMTr-Cl (1.5 equiv) in pyridine
(1.2 mL) over 10 min. After being stirred for an additional 30 min at
room temperature, ethyl acetate was added (50 mL), and the organic
phase was washed with saturated NaHCO3 (20 mL) and brine (20 mL)
and dried over Na2SO4. The solvent was removed in vacuo, and flash
column chromatography on silica gel (20-50% ethyl acetate in hexane)
yielded a light yellow foam. The tritylated nucleoside (0.2 mmol) was
dissolved in CH2Cl2 (2 mL), diisopropylethylamine (4 equiv) was added,
and the solution was cooled to 0 °C. 2-Cyanoethyl diisopropyl-
aminochloro phosphoramidite (1.3 equiv) was added dropwise, and the
reaction was allowed to reach room temperature over 15 min. The
solution was transferred to ethyl acetate (50 mL), and the organic phase
was washed with saturated NaHCO3 (20 mL) and brine (20 mL) and
dried over Na2SO4. The solvent was removed in vacuo and flash column
chromatography on silica gel washed with 2% triethylamine in hexane
(20% ethyl acetate in hexane with 2% triethylamine) afforded a white
foam. The phosphoramidite was generally obtained in 70% yield over
two steps, as a mixture of two diastereoisomers.
1
Compound 3c. H NMR (CDCl3): δ 7.51-7.20 (10H, m), 6.97-
6.70 (6H, m), 6.67-6.66 (1H, m), 5.12 (1H, t, J ) 7.4 Hz), 4.45 (1H,
br s), 4.19 (1H, d, J ) 6.4 Hz), 3.80 (6H, s), 3.40-3.20 (2H, m), 2.80-
2.61 (1H, m), 2.06-1.94 (1H, m). HRMS (MALDI-FTMS): calcd
for C32H31F2O5 (MH+), 533.2061; found, 533.2056.
1
Compound 3d. H NMR: 7.40-7.18 (12H, m), 6.89-6.62 (7H,
m), 5.11 (1H, m), 4.25-4.18 (1H, m), 4.18 (1H, br s), 3.81-3.40 (12H,
m), 3.20 (2H, t, J ) 6.6 Hz), 2.60-2.42 (2H, m), 2.32 (2H, t, J ) 6.5
Hz), 1.40 (1H, d, J ) 6.2 Hz), 1.19-0.96 (6H, m). 31P NMR (CDCl3):
δ 149.15, 149.06. HRMS (MALDI-FTMS): calcd for C41H48F2N2O5P
(MH+), 717.3191; found, 717.3187.
Compound 4c. 1H NMR (CDCl3): δ 7.47 (2H, d, J ) 7.8 Hz), 7.40-
7.18 (9H, m), 7.05 (2H, t, J ) 3.5 Hz), 6.83 (4H, d, J ) 8.6 Hz), 5.44
(1H, dd, J ) 9.7, 4.9 Hz), 4.43 (1H, t, J ) 3.0 Hz), 4.12 (1H, dd, J )
7.5, 3.0 Hz), 3.79 (6H, s), 3.36 (2H, t, J ) 4.8 Hz), 2.58 (1H, br s),
2.38 (1H, dd, J ) 12.8, 5.8 Hz), 2.10-1.90 (1H, m). 19F NMR (CDCl3):
δ -144.30 (1F, t, J ) 8.9 Hz), -139.77 (1F, dt, J ) 10.9, 5.3 Hz). 13
C
NMR (CDCl3): δ 158.7, 145.0, 136.2, 132.2, 130.3, 128.4, 128.1, 127.0,
124.2, 122.0, 116.3, 115.9, 113.4, 86.5, 86.4, 74.4, 74.1, 64.5, 55.5,
42.8. HRMS (MALDI-FTMS): calcd for C32H31F2O5 (MH+), 533.2140;
found, 533.2138.
Compound 4d. 1H NMR (CDCl3): δ 7.38 (2H, d, J ) 7.8 Hz), 7.26-
7.12 (9H, m), 6.97 (2H, dd, J ) 12.2, 5.6 Hz), 6.76-6.70 (4H, m),
5.34 (1H, dd, J ) 12.7, 5.2 Hz), 4.44 (1H, br s), 4.15 (1H, br s), 3.70
(6H, s), 3.69-3.40 (5H, m), 3.22 (2H, dd, J ) 9.8, 5.0 Hz), 2.52 (1H,
t, J ) 6.2 Hz), 2.37 (1H, t, J ) 6.2 Hz), 2.35-2.25 (1H, m), 1.98-
1.84 (1H, m), 1.25-0.96 (14H, m). 19F NMR (CDCl3): -144.46 (1F,
t, J ) 8.8 Hz), -139.88 (1F, dt, J ) 10.9, 5.3 Hz). 13C NMR (CDCl3):
δ 158.7, 145.0, 136.2, 130.4, 128.5, 128.0, 127.0, 124.3, 122.0, 116.3,
115.9, 113.3, 86.4, 85.8, 74.3, 64.1, 58.8, 58.4, 55.5, 43.7, 43.4, 42.1,
25.0, 24.8, 24.7, 20.6, 20.4. 31P NMR (CDCl3): δ 149.54, 149.01.
HRMS (MALDI-FTMS): calcd for C41H47F2N2NaO6P (MNa+),
755.3037; found, 755.3032.
Compound 1c. 1H NMR (CDCl3): δ 7.64-7.02 (13H, m), 6.85 (4H,
d, J ) 8.6 Hz), 5.48 (1H, dd, J ) 9.4, 6.0 Hz), 4.44 (1H, t, J ) 2.8
Hz), 4.13 (1H, dd, J ) 7.4, 5.0 Hz), 3.79 (6H, s), 3.34 (2H, t, J ) 5.0
Hz), 2.82 (1H, br s), 2.38 (1H, dd, J ) 13.2, 5.8 Hz), 2.18-1.95 (1H,
m). 19F NMR (CDCl3): δ -118.93 (1F, t, J ) 5.3 Hz). 13C NMR
(CDCl3): δ 162.5, 158.7, 157.6, 149.7, 145.2, 136.6, 136.3, 130.8, 129.7,
129.5, 129.0, 128.9, 128.5, 128.1, 127.5, 127.4, 127.1, 124.4, 124.3,
115.5, 115.1, 113.4, 110.0, 86.5, 86.3, 85.0, 75.0, 74.5, 74.4, 64.7, 55.5,
42.9. HRMS (MALDI-FTMS): calcd for C32H31FNaO5 (MNa+),
537.2053; found, 537.2051.
Compound 5c. 1H NMR (CDCl3): δ 7.51 (2H, d, J ) 8.2 Hz), 7.40
(4H, d, J ) 7.8 Hz), 7.38-7.22 (4H, m), 7.18-7.07 (2H, m), 6.86
(4H, d, J ) 7.4 Hz), 5.17 (1H, dd, J ) 9.8, 5.4 Hz), 4.43 (1H, d, J )
2.2 Hz), 4.19 (1H, br s), 3.79 (6H, s), 3.34 (2H, t, J ) 4.2 Hz), 2.27
(1H, dd, J ) 13.1, 5.6 Hz), 2.06-1.93 (1H, m). 19F NMR (CDCl3): δ
-138.15 (1F, dd, J ) 17.9, 7.3 Hz), -140.21 (1F, dd, J ) 14.5, 7.9
Hz). 13C NMR (CDCl3): δ 158.8, 149.4, 145.1, 139.6, 139.4, 136.2,
130.4, 130.2, 128.4, 128.1, 127.1, 122.3, 122.2, 122.2, 122.1, 117.4,
117.1, 115.4, 115.0, 113.4, 87.0, 86.6, 79.2, 74.4, 64.8, 55.5, 44.3.
HRMS (MALDI-FTMS): calcd for C32H30F2O5Na (MNa+), 555.1953;
found, 555.1948.
Compound 1d. 1H NMR (CDCl3): δ 7.51-7.02 (13H, m), 6.73 (4H,
dd, J ) 8.8, 2.6 Hz), 5.33 (1H, dd, J ) 10.3, 5.3 Hz), 4.46 (1H, br s),
4.15 (1H, br s), 3.74 (6H, s), 3.72-3.41 (4H, m), 3.23 (2H, t, J ) 4.6
Hz), 2.46 (1H, t, J ) 6.2 Hz), 2.44-2.39 (1H, m), 2.38 (1H, t, J ) 6.2
Hz), 1.97-1.81 (1H, m), 1.17-0.81 (14H, m). 19F NMR (CDCl3): δ
-119.12 (1F, d, J ) 5.3 Hz). 13C NMR (CDCl3): δ 158.7, 157.6, 151.1,
145.2, 136.3, 130.4, 129.2, 128.9, 128.5, 128.0, 127.5, 127.0, 124.4,
115.5, 115.1, 113.3, 110.0, 86.4, 85.8, 74.6, 64.4, 58.8, 58.4, 55.5, 43.6,
43.4, 42.2, 24.8, 24.3, 20.6. 31P NMR (CDCl3): δ 149.40, 148.87.
HRMS (MALDI-FTMS): calcd for C41H48FN2NaO6P (MNa+),
737.3132; found, 737.3127.
Compound 2c. 1H NMR (CDCl3): δ 7.56-7.28 (12H, m), 6.98 (1H,
t, J ) 6.4 Hz), 6.87 (4H, d, J ) 8.6 Hz), 5.23 (1H, dd, J ) 10.0, 5.6
Hz), 4.43 (1H, br s), 4.25-4.10 (1H, m), 3.79 (6H, s), 3.36 (2H, t, J
) 4.8 Hz), 2.33-2.30 (1H, m), 2.05-1.98 (1H, m). 19F NMR (CDCl3):
δ -113.32 (1F, d, J ) 5.3 Hz). 13C NMR (CDCl3): δ 165.6, 160.8,
158.7, 149.5, 145.2, 136.3, 130.4, 130.2, 128.5, 128.1, 127.1, 124.3,
1
Compound 5d. H NMR (CDCl3): δ 7.40-7.15 (10H, m), 7.04-
6.97 (2H, m), 6.76-6.71 (4H, m), 5.04 (1H, dd, J ) 10.5, 5.1 Hz),
4.47-4.39 (1H, m), 4.15-4.06 (1H, m), 3.79-3.40 (4H, m), 3.70 (3H,
s), 3.23-3.18 (2H, m), 2.52 (1H, t, J ) 6.6 Hz), 2.37 (1H, t, J ) 6.6
Hz), 2.30-2.21 (1H, m), 1.93-1.82 (1H, m), 1.13-1.00 (12H, m).
19F NMR (CDCl3): δ -138.2 (1F, m), -140.3 (1F, m). 31P NMR
9
6930 J. AM. CHEM. SOC. VOL. 126, NO. 22, 2004

A Replicable Unnatural Base Pair
A R T I C L E S
(CDCl3): δ 149.2, 149.1. ESI-MS: calcd for C41H47F2N2NaO6P (MNa+),
nucleotide kinase (New England Biolabs) and [γ-33P]-ATP (Amersham
Biosciences). Primers were annealed to template oligonucleotides in
the reaction buffer by heating to 95 °C followed by slow cooling to
ambient temperature. Assay conditions included 40 nM primer template,
0.1-1.3 nM enzyme, 50 mM Tris-HCl, pH 7.5, 10 mM MgCl2, 1 mM
DTT, and 50 µg/mL acetylated BSA. The reactions were carried out
by combining the DNA-enzyme mixture with an equal volume (5 µL)
of 2X dNTP stock solution, incubating at 25 °C for 1-10 min, and
quenching by the addition of 20 µL of loading dye (95% formamide,
20 mM EDTA, and sufficient amounts of bromophenol blue and xylene
cyanole). The reaction mixtures were resolved by 15% polyacrylamide
gel electrophoresis, and the radioactivity was quantified by means of
a PhosphorImager (Molecular Dynamics) and ImageQuant software.
A plot of kobs versus triphosphate concentration was fit to a Michaelis-
Menten equation using the program Kaleidagraph (Synergy Software).
The data presented are averages of three independent determinations.
Procedure for Preparation of 3FB Self-Pair-Containing DNA
Duplex Used in NMR Experiments. Trityl-on oligonucleotides were
cleaved and deprotected by a 12 h reaction with concentrated ammonia
at 60 °C, then dried with centrifugation in vacuo (Savant SpeedVac
Plus SC110A). Detritylation was achieved by adding 80% acetic acid
(2 mL per 1 µmol scale synthesis), vortexing, and incubating at room
temperature for 20 min. The mixture was divided into two aliquots
and transferred to an ice bath, and ice cold concentrated aqueous
ammonium hydroxide (1 mL per one-half 1 µmol scale synthesis) was
added dropwise with continuous mixing by inversion of the capped
tube. The mixture was dried with centrifugation in vacuo, dissolved in
50 mM ammonium formate (0.5 mL per 1 µmol scale synthesis), and
clarified with 0.65 µm centrifugal filtration devices (Ultrafree-MC,
Millipore). After purification by reverse-phase (C18) HPLC (3-15%
CH3CN in 50 mM ammonium formate, pH 7.0 over 30 min) followed
by lyophilization, the oligonucleotides were redissolved in water (0.5
mL per 1 µmol scale synthesis), and their concentrations were
determined by using the value for absorbance at 260 nm in the
Biopolymer Calculator; note that for oligonucleotides containing an
unnatural residue, the position of the unnatural in the sequence was
replaced by dC. Duplexes were formed by mixing equimolar amounts
of each single strand, heating at 90 °C for 2 min, cooling on ice, and
lyophilizing.
755; found, 755.
1
Compound 6c. H NMR (CDCl3): δ 7.50-7.22 (10H, m), 6.94-
6.82 (5H, m), 5.38 (1H, dd, J ) 9.6, 6.0 Hz), 4.44-4.41 (1H, m),
4.12-4.06 (1H, m), 3.79 (6H, s), 3.52-3.29 (2H, m), 2.40-2.31 (1H,
m), 2.06-1.97 (1H, m). 19F NMR (CDCl3): δ -136.4 (1F, m), -140.1
(1F, m), -161.6 (1F, m). 13C NMR (CDCl3): δ 158.7, 145.0, 136.1,
130.3, 128.4, 128.1, 127.1, 113.4, 86.6, 86.4, 74.4, 73.8, 64.4, 55.5,
42.8. HRMS (MALDI-FTMS): calcd for C32H29F3O5Na (MNa+),
573.1859; found, 573.1844.
1
Compound 6d. H NMR (CDCl3): δ 7.39-7.11 (10H, m), 6.89-
6.70 (5H, m), 5.27 (1H, dd, J ) 10.2, 5.1 Hz), 4.47-4.41 (1H, m),
4.14-4.08 (1H, m), 3.75-3.44 (4H, m), 3.70 (6H, s), 3.25-3.13 (1H,
m), 2.55-2.13 (1H, m), 2.52 (1H, t, J ) 6.3 Hz), 2.49-2.29 (1H, m),
2.37 (1H, t, J ) 6.3 Hz), 1.91-1.85 (1H, m), 1.20-0.92 (12H, m). 19
F
NMR (CDCl3): δ -136.6 (1F, m), -140.1 (1F, m), -161.7 (1F, m).
31P NMR (CDCl3): δ 149.5, 149.0. ESI-MS: calcd for C41H47F3N2O6P
(MH+), 751; found, 751.
General Procedure for Triphosphate Synthesis. Proton sponge
(1.5 equiv) and the free nucleoside (1 equiv) were dissolved in trimethyl
phosphate (0.3 M) and cooled to 0 °C. POCl3 (1.05 equiv) was added
dropwise, and the purple slurry was stirred at 0 °C for 2 h.
Tributylamine (4 equiv) was added, followed by a solution of
tributylammonium pyrophosphate (2.5 equiv) in DMF (0.15 M). After
1 min, the reaction was quenched by addition of 1 M aqueous Et3N-
HCO3 (20 vol-equiv). The resulting solution stood for 40 min at 0 °C
and was then lyophilized. Purification by reverse-phase (C18) HPLC
(4-35% CH3CN in 0.1 M Et3N-HCO3, pH 7.5) followed by lyo-
philization afforded the triphosphate as a white solid.
Procedure for Synthesis and Purification of Oligonucleotides
Used in Thermal Denaturation, CD, and Replication Experiments.
Oligonucleotides were synthesized using standard â-cyanoethylphos-
phoramidite chemistry by means of an Applied Biosystems Inc. 392
DNA/RNA synthesizer and reagents that were either synthesized in-
house or purchased from Glen Research, Sterling, VA. Trityl-off
oligonucleotides were cleaved from the CPG support and deprotected
by a 12 h reaction with concentrated ammonia at 60 °C. Purification
by denaturing polyacrylamide gel electrophoresis (10-20%, 8 M urea)
and visualization of the ssDNA bands by UV shadowing was followed
by electroelution and desalting with Sep-Pak (C18) cartridges. The pure
oligonucleotides were dissolved in water, and their concentration was
determined by using the value for absorbance at 260 nm in the
Biopolymer Calculator (http://paris.chem.yale.edu/extinct.html); note
that for oligonucleotides containing an unnatural residue, dC replaced
it in the sequence.
Thermal Denaturation. Oligonucleotide duplex denaturation tem-
perature measurements were made in buffer containing 10 mM PIPES,
pH 7.0, 10 mM MgCl2, and 100 mM NaCl at an ssDNA concentration
of 3 µM using a Cary 200 Bio UV-visible spectrometer. Measurements
of absorbance at 260 nm were taken over the temperature range 16-
80 °C at 0.5 °C/min intervals after rapid annealing and cooling. Melting
temperatures were obtained from one scan by means of the derivative
method in the Cary Win UV software.
CD Spectroscopy. CD experiments were performed by means of
an AVIV model 61 DS spectropolarimeter equipped with a Peltier
thermoelectric temperature control unit. The duplex concentration was
3 µM in buffer containing 10 mM PIPES, pH 7.0, 10 mM MgCl2, and
100 mM NaCl. The data were collected at 25 °C using a quartz cuvette
of 1 cm path length with scanning from 360 to 220 nm, a time constant
of 3 s, and a step size 0.5 nm. After subtracting the buffer reference
from the sample spectra, we converted the observed signal (in units of
millidegrees) to molar ellipticity (in units of degree cm2/dmol) and
plotted it against wavelength.
NMR Spectroscopic Studies of 3FB Self-Pair-Containing DNA
Duplex. Lyophilized duplex DNA containing the 3FB unnatural self-
pair was dissolved in buffer containing 8.5 mM sodium phosphate, pH
7.0, 134 mM NaCl, and 0.25 mM EDTA in either 100% D2O or in
90%/10% H2O/D2O at a final analyte concentration of 0.6 mM. All
NMR spectra were acquired on a Bruker Avance 600 MHz spectrometer
equipped with a 1H/13C/15N-TXI CryoProbe or on an Avance 400 MHz
spectrometer equipped with a 1H/13C/31P/19F-QNP probe (Bruker
Biospin, Billerica, MA). Spectra were acquired at temperatures of 295,
303, and 310 K. Proton resonance assignments were made according
to established procedures.34,35,37-39 The 2D NOESY spectrum shown
in Figure 2 was acquired by collecting 96 transients in each of the 700
T1 experiments using a spectral width of 13 227 Hz, a mixing time of
200 ms, a recycle delay of 2 s, and excitation sculpting with gradients
for water suppression40 using the release pulse program noesyesgpph.
Acknowledgment. Funding was provided by the National
Institutes of Health (No. GM 60005 to F.E.R.) and the JSPS
Research Fellowships for Young Scientists (S.M.). We thank
Peter Wright and Linda Tennant for granting access to the CD
spectropolarimeter.
JA049961U
(37) Chou, S.-H.; Wemmer, D. R.; Reid, B. R. Biochemistry 1983, 22, 3037-
3041.
(38) Boelens, R.; Scheek, R. M.; Dijkstra, K.; Kaptein, R. J. Magn. Reson. 1985,
62, 378-386.
Gel-Based Kinetic Assay. DNA polymerase I (large fragment,
exonuclease deficient) from E. coli was purchased from New England
Biolabs. Primer oligonucleotides were 5′-radiolabeled with T4 poly-
(39) Sklenar, V.; Brooks, B. R.; Zon, G.; Bax, A. FEBS Lett. 1987, 216, 249-
252.
(40) Hwang, T.-L.; Shaka, A. J. J. Magn. Reson., Ser. A 1995, 112, 275-279.
9
J. AM. CHEM. SOC. VOL. 126, NO. 22, 2004 6931