Minor groove hydrogen bonds and the replication of unnatural base pairs

Article abstract of DOI:10.1021/ja068282b

As part of an effort to expand the genetic alphabet, we examined the synthesis of DNA with six different unnatural nucleotides bearing methoxy-derivatized nucleobase analogues. Different nucleobase substitution patterns were used to systematically alter t

Full text of DOI:10.1021/ja068282b

Published on Web 04/06/2007
Minor Groove Hydrogen Bonds and the Replication of
Unnatural Base Pairs
Shigeo Matsuda, Aaron M. Leconte, and Floyd E. Romesberg*
Contribution from the Department of Chemistry, The Scripps Research Institute,
10550 N. Torrey Pines Road, La Jolla, California 92037
Received November 19, 2006; E-mail: floyd@scripps.edu
Abstract: As part of an effort to expand the genetic alphabet, we examined the synthesis of DNA with six
different unnatural nucleotides bearing methoxy-derivatized nucleobase analogues. Different nucleobase
substitution patterns were used to systematically alter the nucleobase electronics, sterics, and hydrogen-
bonding potential. We determined the ability of the Klenow fragment of E. coli DNA polymerase I to
synthesize and extend the different unnatural base pairs and mispairs under steady-state conditions. Unlike
other hydrogen-bond acceptors examined in the past, the methoxy groups do not facilitate mispairing,
implying that they are not recognized by any of the hydrogen-bond donors of the natural nucleobases;
however, they do facilitate replication. The more efficient replication results largely from an increase in the
rate of extension of primers terminating at the unnatural base pair and, interestingly, requires that the
methoxy group be at the ortho position where it is positioned in the developing minor groove and can form
a functionally important hydrogen bond with the polymerase. Thus, ortho methoxy groups should be generally
useful for the effort to expand the genetic alphabet.
1. Introduction
revealed that hydrophobic and packing forces were well suited
to mediate base pair stability and polymerase-mediated synthesis
(by insertion of the unnatural triphosphate opposite the unnatural
nucleotide in the template).10-12 However, the utility of these
unnatural base pairs has been consistently limited by insertion
of the next correct dNTP (i.e., extension).
More recently, a significant improvement in extension rate
has been achieved with several nucleobase analogues that have
relatively little aromatic surface area.6,9,22,23,25 Presumably, these
pairs form a more natural-like primer terminus, as opposed to
larger analogues that are likely to distort the primer terminus.
While the BEN self pair (formed between two identical BEN
analogues, Figure 1a) is not particularly stable24 or well
recognized by DNA polymerases,9 several derivatives have been
identified that form self pairs or heteropairs (formed between
two different analogues) with significantly improved properties.
For example, the 3FB,6 DM5,9,24 and TM9,11,12,24 nucleotides
(Figure 1a) form pairs that are reasonably stable and efficiently
synthesized by the exonuclease deficient Klenow fragment of
An unnatural base pair that is stable and replicable would
increase the biotechnological utility and information storage
potential of DNA.1-5 Toward this goal, we have examined
unnatural nucleotides that bear predominantly hydrophobic
nucleobase analogues.6-14 These analogues are expected to pair
with each other within duplex DNA via hydrophobic and
packing interactions but not with the natural nucleotides, due
to forced desolvation of the more hydrophilic natural nucleo-
bases. Our initial efforts focused on nucleobase analogues with
large aromatic surface areas and, along with other studies from
the Hirao,4,5,15 Kuchta,16,17 Kool,18,19 and Berdis20,21 labs,
(1) Henry, A. A.; Romesberg, F. E. Curr. Opin. Chem. Biol. 2003, 7, 727-
733.
(2) Piccirilli, J. A.; Krauch, T.; Moroney, S. E.; Benner, S. A. Nature 1990,
343, 33-37.
(3) Sismour, A. M.; Benner, S. A. Nucleic Acids Res. 2005, 33, 5640-5646.
(4) Hirao, I. Curr. Opin. Chem. Biol. 2006, 10, 622-627.
(5) Hirao, I.; Kimoto, M.; Mitsui, T.; Fujiwara, T.; Kawai, R.; Sato, A.; Harada,
Y.; Yokoyama, S. Nat. Methods 2006, 3, 729-735.
(6) Henry, A. A.; Olsen, A. G.; Matsuda, S.; Yu, C.; Geierstanger, B. H.;
Romseberg, F. E. J. Am. Chem. Soc. 2004, 126, 6923-6931.
(7) Henry, A. A.; Yu, C.; Romesberg, F. E. J. Am. Chem. Soc. 2003, 125,
9638-9646.
(16) Chiaramonte, M.; Moore, C. L.; Kincaid, K.; Kuchta, R. D. Biochemistry
2003, 42, 10472-10481.
(8) Hwang, G. T.; Romesberg, F. E. Nucleic Acids Res. 2006, 34, 2037-2045.
(9) Matsuda, S.; Henry, A. A.; Romesberg, F. E. J. Am. Chem. Soc. 2006,
128, 6369-6375.
(17) Kincaid, K.; Beckman, J.; Zivkovic, A.; Halcomb, R. L.; Engels, J. W.;
Kuchta, R. D. Nucleic Acids Res. 2005, 33, 2620-2628.
(18) Matray, T. J.; Kool, E. T. Nature 1999, 399, 704-708.
(19) Morales, J. C.; Kool, E. T. J. Am. Chem. Soc. 1999, 121, 2323-2324.
(20) Zhang, X.; Lee, I.; Berdis, A. J. Biochemistry 2005, 44, 13101-13110.
(21) Zhang, X.; Lee, I.; Zhou, X.; Berdis, A. J. J. Am. Chem. Soc. 2006, 128,
143-149.
(10) McMinn, D. L.; Ogawa, A. K.; Wu, Y.; Liu, J.; Schultz, P. G.; Romesberg,
F. E. J. Am. Chem. Soc. 1999, 121, 11585-11586.
(11) Ogawa, A. K.; Wu, Y.; Berger, M.; Schultz, P. G.; Romesberg, F. E. J.
Am. Chem. Soc. 2000, 122, 8803-8804.
(12) Ogawa, A. K.; Wu, Y.; McMinn, D. L.; Liu, J.; Schultz, P. G.; Romesberg,
F. E. J. Am. Chem. Soc. 2000, 122, 3274-3287.
(22) Kim, Y.; Leconte, A. M.; Hari, Y.; Romesberg, F. E. Angew. Chem., Int.
Ed. 2006, 45, 7809-7812.
(13) Tae, E. L.; Wu, Y. Q.; Xia, G.; Schultz, P. G.; Romesberg, F. E. J. Am.
Chem. Soc. 2001, 123, 7439-7440.
(23) Leconte, A. M.; Matsuda, S.; Hwang, G. T.; Romesberg, F. E. Angew.
Chem., Int. Ed. 2006, 45, 4326-4329.
(14) Yu, C.; Henry, A. A.; Romesberg, F. E.; Schultz, P. G. Angew. Chem., Int.
Ed. 2002, 41, 3841-3844.
(24) Matsuda, S.; Romesberg, F. E. J. Am. Chem. Soc. 2004, 126, 14419-
14427.
(15) Hirao, I.; Harada, Y.; Kimoto, M.; Mitsui, T.; Fujiwara, T.; Yokoyama, S.
J. Am. Chem. Soc. 2004, 126, 13298-12305.
(25) Adelfinskaya, O.; Nashine, V. C.; Bergstrom, D. E.; Davisson, V. J. J.
Am. Chem. Soc. 2005, 127, 16000-16001.
9
10.1021/ja068282b CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 5551-5557
5551

A R T I C L E S
Matsuda et al.
minor groove nitrogen substituents that may act as H-bond
acceptors.22 However, while these modifications increase exten-
sion rates, they also destabilize the base pair and appear to be
recognized by the H-bond donor of dG, resulting in reduced
thermal and replication fidelity.
To identify H-bond acceptors that may not be recognized by
the natural nucleobases but that could mediate important
primer-polymerase interactions, we have begun to characterize
phenyl-based nucleotides derivatized with methoxy groups.
Based on both theory31,32 and experiment,31-34 aryl methoxy
groups should be capable of forming H-bonds, although strong
H-bonds will only be possible when the methyl group is rotated
out of the plane of the aryl ring, localizing electron density at
the oxygen and making it a better H-bond acceptor. Indeed,
thermodynamic studies of the same derivatives suggest that the
methoxy groups are prone to desolvate upon duplex formation
(S.M. and F.R. unpublished results). Thus, it is unclear whether
these methoxy groups would engage the polymerase H-bond
donors and align the primer terminus for efficient extension.
We now report a detailed characterization of the Kf-mediated
synthesis and extension of the methoxy derivatized base pairs.
The analogues characterized were designed to systematically
explore the effects of methoxy groups at the ortho, meta, and/
or para-positions (Figure 1b). Generally, we find that the
addition of methoxy groups increases the selectivity of unnatural
base pair synthesis, and when the groups are present at the ortho
position, they also have a selective and significant effect on
extension, with some pairs being extended only ∼100-fold less
efficiently than a natural base pair. Using the R668A mutant of
Kf, we show that this efficient extension requires the presence
of the polymerase-based H-bond donor. The data suggest that
an appropriately positioned methoxy group can productively
engage the H-bond donor of the DNA polymerase, which
facilitates extension, but not the H-bond donors of any natural
nucleobase, which would stabilize or facilitate mispair forma-
tion.
Figure 1. (a) Unnatural nucleobases previously reported. (b) Methoxy
substituted benzene analogues used in this study.
E. coli DNA polymerase I (Kf) but that are also extended with
increased efficiency. Nonetheless, the extension of these pairs
remains significantly less efficient than that of a natural base
pair, and modifications that further increase extension rates are
still required.
2. Results
2.1. Unnatural Base Pair Synthesis Efficiency. The un-
natural nucleotides were synthesized and converted into the
corresponding phosphoramidites or triphosphates as described
in the Supporting Information. The phosphoramidites were used
to synthesize template DNA containing the unnatural nucleotides
at a single defined position. To begin to examine how methoxy
groups impact polymerase-mediated replication, we determined
the steady-state rates with which Kf extends a primer terminating
immediately 5′ to the unnatural base in the template by insertion
of an unnatural triphosphate (Table 1). For reference, Kf inserts
a natural dATP opposite dT with a second-order rate constant
(i.e., efficiency or kcat/KM) of 1.7 × 108 M-1 min-1. To compare
the ability of nucleobases with either an ortho methyl or an
ortho methoxy group to direct triphosphate insertion, we first
examined the insertion of the unnatural triphosphates opposite
either MM1 (Table 1, X ) MM1) or 2OMe (Table 1, X )
One possible reason for the poor extension of the predomi-
nantly hydrophobic base pairs is that, while optimized interbase
packing may obviate the need for interbase hydrogen bonds (H-
bonds),26 efficient extension may require H-bonding between
the nucleobase at the primer terminus and the polymerase.19,27,28
Indeed, the natural nucleobases all have an H-bond acceptor
oriented into the developing minor groove, and structural studies
have revealed a conserved H-bond between these acceptors at
the primer terminus and polymerase based H-bond donors,29
such as Arg668 in Kf. Biochemical studies have shown that
extension is significantly reduced when these H-bonds are
disrupted.27,30 Thus, we have explored the derivatization of the
phenyl-based nucleosides with either minor groove carbonyl
groups23 (by changing the C-nucleoside to an N-nucleoside) or
(31) Becucci, M.; Pietraperzia, G.; Pasquini, M.; Piani, G.; Zoppi, A.; Chelli,
R.; Castellucci, E.; Demtroeder, W. J. Chem. Phys. 2004, 120, 5601-5607.
(32) Reimann, K.; Buchhold, H.-D.; Brutschy, B.; Tarakeshwar, P.; Kim, K. S.
J. Chem. Phys. 2002, 117, 8805-8822.
(26) Moran, S.; Ren, R. X.-F.; Rumney, S. I.; Kool, E. T. J. Am. Chem. Soc.
1997, 119, 2056-2057.
(27) Spratt, T. E. Biochemistry 2001, 40, 2647-2652.
(28) Morales, J. C.; Kool, E. T. Biochemistry 2000, 39, 12979-12988.
(29) Li, Y.; Waksman, G. Protein Sci. 2001, 10, 1225-1233.
(30) Meyer, A. S.; Blandino, M.; Spratt, T. E. J. Biol. Chem. 2004, 279, 33043-
33046.
(33) Nobeli, I.; Yeoh, S. L.; Price, S. L.; Taylor, R. Chem. Phys. Lett. 1997,
280, 196-202.
(34) Ribblett, J. W.; Sinclair, W. E.; Borst, D. R.; Yi, J. T.; Pratt, D. W. J.
Phys. Chem. A 2006, 110, 1478-1483.
9
5552 J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007

Optimization of Unnatural Base Pair Replication
A R T I C L E S
Table 1. Rates of Unnatural Base Pair Synthesisa
Table 2. Incorporation of Unnatural and Natural Triphosphates
Opposite Unnatural Bases in the Templatea
kcat/Km
1
1
kcat/Km
X
triphosphate
kcat (min-
)
Km
(µM)
(M-1 min-
)
1
1
X
triphosphate
kcat (min-
)
Km
(
µ
M)
(M-1 min-
)
2OMe
2OMe
DMO
MMO2
MM1
DM5
0.69 ( 0.16
3.31 ( 0.73
3.47 ( 0.57
1.73 ( 0.16
14.0 ( 0.60
18.2 ( 1.3
104 ( 32
85 ( 4
6.6 × 103
3.9 × 104
4.8 × 104
1.5 × 104
1.4 × 105
7.0 × 105
2OMe
A
C
G
T
3.78 ( 0.17
44 ( 11
ndb
8.6 × 104
<1.0 × 103
<1.0 × 103
<1.0 × 103
ndb
ndb
ndb
72 ( 2
ndb
117 ( 16
103 ( 11
26 ( 4
ndb
TM
3OMe
4OMe
MMO1
MMO2
DMO
A
C
G
T
1.77 ( 0.22
0.29 ( 0.04
ndb
32 ( 10
168 ( 28
ndb
5.5 × 104
1.7 × 103
MMO1
MMO2
DMO
MM1
DM5
2OMe
DMO
MMO2
MM1
DM5
2.06 ( 0.51
6.16 ( 0.57
4.34 ( 1.1
2.89 ( 0.67
11.5 ( 1.30
7.18 ( 0.81
109 ( 8
42 ( 10
39 ( 6
78 ( 4
62 ( 17
15.4 ( 1.5
1.9 × 104
1.5 × 105
1.1 × 105
3.7 × 104
1.9 × 105
4.7 × 105
<1.0 × 103
2.5 × 103
0.33 ( 0.08
130 ( 12
A
C
G
T
2.96 ( 0.25
ndb
25 ( 7
ndb
1.2 × 105
<1.0 × 103
<1.0 × 103
3.2 × 104
TM
ndb
ndb
2OMe
DMO
MMO2
MM1
DM5
1.50 ( 0.35
5.14 ( 1.37
5.13 ( 0.67
4.38 ( 0.36
10.5 ( 2.8
22.0 ( 1.7
120 ( 20
51 ( 5
44 ( 4
109 ( 17
48 ( 15
13.8 ( 3.2
1.3 × 104
1.0 × 105
1.2 × 105
4.0 × 104
2.2 × 105
1.6 × 106
3.03 ( 0.23
95 ( 18
A
C
G
T
4.50 ( 0.40
ndb
25 ( 8
ndb
1.8 × 105
<1.0 × 103
<1.0 × 103
2.2 × 103
ndb
ndb
0.28 ( 0.02
128 ( 6
TM
A
C
G
T
3.27 ( 0.33
32 ( 4
ndb
1.0 × 105
<1.0 × 103
<1.0 × 103
<1.0 × 103
2OMe
DMO
MMO2
MM1
DM5
0.76 ( 0.20
1.85 ( 0.17
1.85 ( 0.38
1.68 ( 0.29
3.30 ( 0.29
4.17 ( 1.04
104 ( 29
26 ( 3
7.3 × 103
7.1 × 104
4.9 × 104
3.7 × 104
1.2 × 105
4.5 × 105
ndb
ndb
ndb
ndb
38 ( 2
ndb
46 ( 10
28 ( 9
A
C
G
T
1.08 ( 0.15
ndb
13 ( 1.3
ndb
8.3 × 104
<1.0 × 103
<1.0 × 103
2.5 × 103
TM
9.3 ( 0.3
ndb
ndb
2OMe
DMO
MMO2
MM1
DM5
1.12 ( 0.28
4.05 ( 0.42
5.24 ( 1.22
0.93 ( 0.31
7.11 ( 1.67
5.50 ( 0.91
128 ( 26
52 ( 13
39 ( 2
108 ( 11
82 ( 7
15.7 ( 2.6
8.8 × 103
7.8 × 104
1.3 × 105
8.6 × 103
8.7 × 104
3.5 × 105
0.14 ( 0.03
57 ( 15
a See Experimental Section for details. bReaction was too inefficient for
kcat and KM to be determined independently.
TM
2OMe
DMO
MMO2
MM1
DM5
2.54 ( 0.31
7.05 ( 0.45
9.63 ( 0.81
2.28 ( 0.34
50 ( 4.6
89 ( 4
52 ( 6
28 ( 3
102 ( 12
25 ( 6
12.6 ( 2.2
2.9 × 104
1.4 × 105
3.4 × 105
2.2 × 104
2.0 × 106
8.6 × 105
directs the insertion of the unnatural triphosphates with second-
order rate constants between 2.2 × 104 and 2.0 × 106 M-1 min-1
(Table 1, X ) DM5). Methoxy substitution of the template base
(i.e., Table 1, X ) MMO1 or MMO2) generally decreases the
rates of unnatural triphosphate insertion, most so when the
templating analogue has two methoxy substituents (i.e., Table
1, X ) DMO). With dMMO2TP the effects are slightly larger,
and they are largest with dDM5TP. However, the opposite effect
was observed with both dMM1TP and dTMTP, where changing
the ortho substituent of the template analogue from a methyl
to a methoxy group increases the rate of unnatural triphosphate
insertion. These data reinforce the conclusion that specific
structural and/or electrostatic effects contribute to efficient
unnatural base pair synthesis.
2.2. Unnatural Base Pair Synthesis Fidelity. To examine
the synthesis of mispairs we determined the efficiencies with
which Kf inserts natural dNTPs opposite the unnatural bases in
the template (Table 2). dATP is consistently the most efficiently
inserted natural triphosphate, with kcat/KM values falling between
5.5 × 104 M-1 min-1 and 1.8 × 105 M-1 min-1. While this is
more efficient than mispair synthesis among the natural nucle-
otides,35 it is significantly less efficient than dATP insertion
opposite any of the carbocyclic analogue MM1, MM2, MM3,
or DM5 (which template dATP insertion with second-order rate
constants of 3.9 × 105 to 2.9 × 106 M-1 min-1).9 After dATP,
the next most efficiently inserted triphosphate is dTTP, with
rates ranging from too slow to detect (<103 M-1min-1) to 2 ×
TM
10.8 ( 1.5
TM
2OMe
DMO
MMO2
MM1
DM5
4.24 ( 0.76
7.63 ( 1.30
9.47 ( 1.68
3.19 ( 0.09
18.1 ( 1.0
31 ( 1.8
39 ( 5
24 ( 6
1.1 × 105
3.2 × 105
8.1 × 105
5.8 × 104
6.2 × 105
2.2 × 106
11.7 ( 1.8
55 ( 18
29 ( 6
TM
14 ( 3
a See Experimental Section for details.
2OMe). MM1 in the template directs unnatural base pair
synthesis with second-order rate constants (i.e., efficiency, kcat/
KM) between 8.6 × 103 and 3.5 × 105 M-1 min-1. Interestingly,
methoxy substitution at the ortho position slightly decreases the
rate of insertion of the more hydrophilic triphosphates but
increases the rate of insertion of the more hydrophobic triph-
osphates. The most efficiently synthesized pair results from the
insertion of dTMTP opposite 2OMe, which proceeds with a
second-order rate constant of 7.0 × 105 M-1 min-1. Because
methoxy substitution increases the polarity of the nucleobase,
the data suggest that, in addition to hydrophobic forces, specific
structural and/or electrostatic interactions must contribute to base
pair synthesis.
We next examined the effect of substituting methoxy groups
for methyl groups within the DM5 scaffold in the template.
This scaffold was chosen because of the relative stability and
efficient replication of the DM5 self pair. The DM5 analogue
(35) Kuchta, R. D.; Benkovic, P.; Benkovic, S. J. Biochemistry 1988, 27, 6716-
6725.
9
J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007 5553

A R T I C L E S
Matsuda et al.
Table 3. Rates of Correct Extension of Unnatural Base Pairsa
1
1
1
1
X
Y
kcat (min-
)
Km
(µM)
kcat/Km (M-1 min-
)
X
Y
kcat (min-
)
Km
(
µM)
kcat/Km (M-1 min-
)
2OMe
2OMe
3OMe
4OMe
DMO
MMO1
MMO2
MM1
DM5
ndb
ndb
<1.0 × 103
1.6 × 103
4.6 × 103
1.7 × 103
1.3 × 103
2.2 × 103
<1.0 × 103
2.0 × 103
2.3 × 103
DMO
2OMe
3OMe
4OMe
DMO
MMO1
MMO2
MM1
DM5
ndb
0.35 ( 0.08
0.37 ( 0.04
ndb
0.17 ( 0.04
0.47 ( 0.11
ndb
ndb
152 ( 62
163 ( 42
ndb
143 ( 33
158 ( 22
ndb
<1.0 × 103
2.3 × 103
2.3 × 103
<1.0 × 103
1.2 × 103
3.0 × 103
<1.0 × 103
2.8 × 103
3.4 × 103
0.27 ( 0.02
0.96 ( 0.24
0.30 ( 0.10
0.21 ( 0.03
0.38 ( 0.14
ndb
173 ( 11
209 ( 61
178 ( 77
165 ( 51
170 ( 39
ndb
0.47 ( 0.15
0.40 ( 0.09
231 ( 48
173 ( 40
0.35 ( 0.03
0.34 ( 0.02
124 ( 13
99 ( 14
TM
TM
3OMe
4OMe
MMO1
MMO2
2OMe
3OMe
4OMe
DMO
MMO1
MMO2
MM1
DM5
3.12 ( 0.82
ndb
112 ( 29
ndb
2.8 × 104
<1.0 × 103
3.6 × 103
2.4 × 104
1.3 × 103
3.1 × 104
<1.0 × 103
2.3 × 103
1.3 × 103
MM1
DM5
TM
2OMe
3OMe
4OMe
DMO
MMO1
MMO2
MM1
DM5
9.34 ( 1.48
0.40 ( 0.02
0.92 ( 0.05
5.90 ( 1.19
0.25 ( 0.06
7.18 ( 0.39
0.25 ( 0.07
0.35 ( 0.08
0.54 ( 0.07
124 ( 14
153 ( 13
153 ( 25
86 ( 3
7.5 × 104
2.6 × 103
6.0 × 103
6.9 × 104
2.7 × 103
1.4 × 105
1.4 × 103
2.1 × 103
3.2 × 103
0.79 ( 0.15
3.14 ( 0.31
0.22 ( 0.08
4.95 ( 1.39
ndb
0.32 ( 0.07
0.18 ( 0.05
218 ( 53
132 ( 21
175 ( 74
160 ( 53
ndb
138 ( 43
144 ( 25
92 ( 10
50 ( 10
173 ( 100
163 ( 17
168 ( 46
TM
TM
2OMe
3OMe
4OMe
DMO
MMO1
MMO2
MM1
DM5
2.49 ( 0.83
162 ( 11
ndb
1.5 × 104
<1.0 × 103
<1.0 × 103
2.1 × 104
2OMe
3OMe
4OMe
DMO
MMO1
MMO2
MM1
DM5
7.30 ( 0.54
1.23 ( 0.33
1.29 ( 0.36
5.44 ( 0.49
0.48 ( 0.13
6.07 ( 0.85
0.23 ( 0.09
6.5 ( 1.1
57 ( 13
121 ( 30
104 ( 35
54 ( 5
121 ( 10
18.6 ( 2.5
220 ( 91
161 ( 17
95 ( 10
1.3 × 105
1.0 × 104
1.2 × 104
1.0 ×105
4.0 × 103
3.3 × 105
1.1 × 103
4.0 × 104
1.1 × 104
ndb
ndb
ndb
3.08 ( 0.73
147 ( 41
ndb
ndb
<1.0 × 103
4.8 × 104
6.04 ( 0.55
ndb
127 ( 30
ndb
<1.0 × 103
<1.0 × 103
<1.0 × 103
ndb
ndb
TM
ndb
ndb
TM
1.04 ( 0.18
2OMe
3OMe
4OMe
DMO
MMO1
MMO2
MM1
DM5
9.19 ( 1.24
0.68 ( 0.12
0.64 ( 0.14
6.52 ( 0.77
0.27 ( 0.04
6.88 ( 1.24
0.17 ( 0.07
0.56 ( 0.14
0.39 ( 0.04
86 ( 3
1.1 × 105
3.5 × 103
3.7 × 103
1.1 × 105
2.2 × 103
1.5 × 105
1.0 × 103
4.2 × 103
2.7 × 103
2OMe
3OMe
4OMe
DMO
MMO1
MMO2
MM1
DM5
12.0 ( 2.7
3.79 ( 0.09
3.37 ( 0.30
4.13 ( 0.60
1.51 ( 0.27
10.1 ( 1.4
0.62 ( 0.11
2.74 ( 0.54
7.9 ( 1.4
12.4 ( 1.5
92 ( 17
9.7 × 105
4.1 × 104
3.8 × 104
3.2 × 105
1.8 × 104
1.3 × 106
5.7 × 103
3.0 × 104
5.2 × 104
195 ( 74
172 ( 39
60 ( 18
89 ( 19
13.1 ( 2.9
85 ( 24
125 ( 35
46 ( 11
168 ( 43
132 ( 16
144 ( 34
7.9 ( 0.97
108 ( 3
92 ( 18
152 ( 32
TM
TM
2OMe
3OMe
4OMe
DMO
MMO1
MMO2
MM1
DM5
0.44 ( 0.11
1.07 ( 0.14
1.83 ( 0.66
0.55 ( 0.12
0.50 ( 0.13
0.87 ( 0.28
0.36 ( 0.20
1.48 ( 0.08
0.98 ( 0.23
201 ( 51
243 ( 37
184 ( 34
147 ( 24
247 ( 15
165 ( 47
184 ( 95
148 ( 24
155 ( 33
2.2 × 103
4.4 × 103
9.9 × 103
3.7 × 103
2.0 × 103
5.3 × 103
2.0 × 103
1.0 × 104
6.3 × 103
BEN
2OMe
3OMe
4OMe
DMO
MMO1
MMO2
MM1
DM5
1.49 ( 0.35
272 ( 28
ndb
5.5 × 103
<1.0 × 103
<1.0 × 103
1.3 × 104
ndb
ndb
ndb
2.42 ( 0.41
191 ( 44
ndb
ndb
<1.0 × 103
1.7 × 104
4.05 ( 0.65
ndb
233 ( 21
ndb
<1.0 × 103
<1.0 × 103
<1.0 × 103
ndb
ndb
TM
TM
ndb
ndb
a See Experimental Section for details. b Reaction was too inefficient for kcat and KM to be determined independently.
103 M-1min-1, for insertion opposite 3OMe, MMO1, and
DMO, or 3.2 × 104 M-1 min-1, for insertion opposite 4OMe.
Thus, dTTP insertion is strongly favored by para methoxy
substitution and disfavored by ortho methoxy substitution. dCTP
is only inserted with a detectable rate opposite 3OMe, and dGTP
is not detectably inserted opposite any of the analogues. Of
particular interest is the selectivity against mispairing of the
analogues, especially those with ortho methoxy H-bond accep-
tors (2OMe, MMO2, and DMO). The generally decreased rates
with which the natural triphosphates are inserted opposite the
methoxy derivatized analogues, relative to their fully carbocyclic
counterparts, are surprising considering their decreased hydro-
phobicity, and this again suggests that specific structural and/
or electrostatic effects are important.
examined the BEN nucleotide and nucleotides bearing single
substituents at the ortho, meta, or para position of the phenyl
nucleobase scaffold (Table 3). These analogues were incorpo-
rated into oligonucleotide templates and annealed to primers
containing one of the analogues shown in Figure 1, resulting in
the formation of an unnatural self-pair or heteropair at the primer
terminus. We determined the ability of Kf to extend each primer
by incorporation of the next correct triphosphate (dCTP). For
reference, a dA:dT pair is extended in the same sequence context
with a kcat/KM of 1.7 × 108 M-1 min-1. Pairs with BEN in the
template were recognized poorly by Kf (Table 3, X ) BEN).
Only primers terminating with 2OMe, DMO, or MMO2 are
extended with a detectable rate (kcat/KM > 1 × 103 M-1 min-1).
Pairs formed with 2OMe in the template are extended poorly
(Table 3, X ) 2OMe), while pairs with either 3OMe or 4OMe
in the template are extended between 1.5 × 104 and 4.8 × 104
2.3. Unnatural Base Pair Extension Efficiency. To examine
the contribution of methoxy groups to extension, we first
9
5554 J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007

Optimization of Unnatural Base Pair Replication
A R T I C L E S
Table 4. Rates of Mispair Extensiona
M-1 min-1, but again only when paired opposite 2OMe, DMO,
or MMO2 at the primer terminus (Table 3, X ) 3OMe or
4OMe). These data suggest that an ortho methoxy group at the
primer terminus facilitates extension.
kcat/Km
We next examined the ortho, para di-substituted analogues
MMO1, MMO2, and DMO in the template (Table 3). Kf does
not efficiently recognize primers that terminate opposite MMO2
or DMO, regardless of the analogue at the primer terminus
(Table 3, X ) MMO2 or DMO). In each case the second-
1
1
X
Y
kcat (min-
)
Km
(
µ
M)
(M-1 min-
)
MM1
DM5
MMO1
MMO2
TM
A
G
C
A
A
A
A
A
5.61 ( 0.38
7.29 ( 1.75
9.48 ( 0.69
8.65 ( 0.89
8.13 ( 1.59
ndb
29 ( 6
27 ( 6
23 ( 2
187 ( 17
15.5 ( 2.2
ndb
1.9 × 105
2.7 × 105
4.1 × 105
4.6 × 104
5.2 × 105
<1.0 × 103
<1.0 × 103
3.6 × 103
4.5 × 103
<1.0 × 103
<1.0 × 103
1.9 × 104
1.6 × 104
<1.0 × 103
<1.0 × 103
5.0 × 103
2.6 × 104
order rate constants are less than or equal to 1 × 104 M-1 min-1
.
2OMe
2OMe
2OMe
2OMe
DMO
DMO
DMO
DMO
MMO2
MMO2
MMO2
MMO2
Also, most pairs involving MMO1 in the template are not well
extended (Table 3, X ) MMO1). However, the pairs formed
with 2OMe, MMO2, or DMO in the primer and MMO1 in
the template are again extended more efficiently, with second-
order rate constants of ∼1 × 105 M-1 min-1. These data
reinforce the idea that ortho positioned methoxy groups at the
primer terminus increase the rates of extension but also suggest
that the ortho methyl group of MMO1 in the template
contributes to efficient extension as well.
ndb
ndb
0.40 ( 0.15
0.39 ( 0.09
ndb
111 ( 37
87 ( 13
ndb
T
A
G
C
ndb
ndb
1.31 ( 0.23
0.92 ( 0.04
ndb
69 ( 11
59 ( 6
ndb
T
A
G
C
ndb
ndb
0.67 ( 0.25
2.11 ( 0.63
134 ( 32
80 ( 13
T
To further explore the effect of ortho methyl substituents in
the template, the nucleobase analogue MM1, DM5, or TM was
incorporated into the template strand (Table 3, X ) MM1,
DM5, or TM). Generally, the pairs formed with each of the
primers are extended slightly better than the pairs without ortho
methyl groups in the template analogue, and the rates parallel
the extent of methyl group substitution (TM > DM5 > MM1).
Again, as with the monosubstituted analogues, the pairs formed
with 2OMe, DMO, or MMO2 in the primer are consistently
extended more efficiently. In fact, the pairs formed between
2OMe, DMO, or MMO2 at the primer terminus and TM in
the template are extended with rates between 3.2 × 105 M-1
min-1 and 1.3 × 106 M-1 min-1. The MMO2:TM heteropair
is extended ∼1000-fold faster than the unsubstituted BEN self
a See Experimental Section for details. b Reaction was too inefficient for
kcat and KM to be determined independently.
Table 5. Rates of Correct Extension by R668A Kf Mutanta
kcat/Km
1
1
X
Y
kcat (min-
)
Km
(µM)
(M-1 min-
)
T
TM
TM
A
9.09 ( 1.69
2.31 ( 0.79
0.60 ( 0.09
29 ( 6
63 ( 9
44 ( 3
3.1 × 105
3.7 × 104
1.4 × 104
MMO2
DM5
a See Experimental Section for details.
pair (which is extended with a rate of 1.6 × 103 M-1 min-1 9
)
and only ∼100-fold less efficiently than a natural base pair.
The increased rates of extension of the 2OMe:TM, DMO:TM,
and MMO2:TM heteropairs, relative to the BEN self pair, result
from both increases in the apparent kcat and decreases in the
apparent KM. The data clearly reveal that unnatural base pair
extension is facilitated by both an ortho methoxy group in the
primer nucleobase and an ortho methyl group in the template
nucleobase.
MMO2 opposite each natural nucleotide in the template and
determined the efficiency of mispair extension (Table 4). None
of the mispairs are extended efficiently, with rates varying from
too low to detect (<103 M-1 min-1) to 2.6 × 104 M-1 min-1
.
Interestingly, extension of the mispairs with dG is very
inefficient. These results clearly demonstrate that an H-bond is
not formed between guanine and the ortho methoxy substituent
or that if an H-bond is formed, it does not contribute to a
structure at the primer terminus that is recognized by Kf. In
fact, the mispairs between either purine and the unnatural
analogues are extended less efficiently than those with either
pyrimidine. Because the purine mispairs are more structurally
similar to a natural base pair, the more efficient extension of
the pyrimidine mispairs is surprising.
2.4. Unnatural Base Pair Extension Fidelity. To examine
extension fidelity, we first characterized the rate at which Kf
extends a primer terminating with dA paired opposite MMO1
or MMO2 (Table 4). To further elucidate the effect of the
methoxy group substituent, primers terminating with dA paired
opposite MM1, DM5, or TM were also characterized. The
mispairs with dA at the primer terminus were examined because,
in each case, they are the most efficiently synthesized (see
above). All mispairs except dA:MMO2 are extended by dCTP
insertion with rates between 1.9 × 105 and 5.2 × 105 M-1
min-1. In contrast, the dA:MMO2 mispair is extended less
efficiently, with a kcat/KM of 4.6 × 104 M-1 min-1. This rate is
only marginally more efficient than the extension of a natural
mispair35 and demonstrates that the ortho methoxy group
decreases the rate of mispair extension.
2.5. Primer-Polymerase Interactions and Unnatural Base
Pair Extension. To determine whether the minor groove
methoxy groups facilitate unnatural base pair extension through
formation of an H-bond with the polymerase, we examined
extension rates with the R668A mutant of Kf36 (Table 5). While
other residues are also important for minor groove recognition
of the primer terminus,36,37 Arg668 is the most thoroughly
characterized, and among the identified residues, it has the
(36) Minnick, D. T.; Bebenek, K.; Osheroff, W. P.; Turner, R. M., Jr.; Astatke,
M.; Liu, T.; Kunkel, T. A.; Joyce, C. M. J. Biol. Chem. 1999, 274, 3067-
3075.
(37) Summerer, D.; Rudinger, N. Z.; Detmer, I.; Marx, A. Angew. Chem., Int.
Ed. Engl. 2005, 44, 4712-4715.
To further explore whether ortho methoxy groups at the
primer terminus form mispairs with natural nucleotides that are
recognized and extended by Kf, we paired 2OMe, DMO, or
9
J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007 5555

A R T I C L E S
Matsuda et al.
largest effect during extension. Previous studies have shown
that either removal of the H-bond donor from the polymerase
(by mutation of Arg668 to Ala) or removal of the hydrogen
bond acceptor from a natural nucleobase at the primer terminus
(by substituting 3-deaza-dG for dG) significantly decreases the
rate of primer extension.27,30 Additionally, this mutant has been
used to show that the difference in extension rates of two natural
nucleobase analogues is due to H-bonding with the DNA
polymerase.38 Kf R668A extends the DM5:TM heteropair with
a rate only 2-fold reduced from that for wild-type Kf, demon-
strating that Arg668 does not form a functionally important
H-bond with DM5. However, extension of the MMO2:TM
heteropair is reduced 35-fold, to essentially the same rate as
that of the DM5:TM heteropair. The apparent KM increased
almost 8-fold with MMO2 at the primer terminus, while it
decreased 2-fold with DM5 at the primer terminus. In both
cases, kcat decreased by approximately 4-fold. While it is difficult
to interpret these changes mechanistically, as the kinetics are
run under steady-state conditions, it is clear that removing the
H-bond donor from the polymerase selectively decreases the
extension efficiency of the pair with the minor groove methoxy
group at the primer terminus. Thus, the ortho methoxy group
of MMO2 appears to favorably interact with Arg668, presum-
ably via H-bonding.
between different predominantly hydrophobic nucleobase
analogues.4-6,10,11 Interestingly, the effects of substituting meth-
oxy groups for methyl groups in unnatural base pair and mispair
synthesis is not consistent with simple changes in hydrophobic-
ity. For example, substitution of the ortho methyl group of MM1
with a methoxy group, to give 2OMe, results in the more
efficient insertion of more hydrophobic triphosphate analogues
but less efficient insertion of more hydrophilic triphosphate
analogues. In addition, 2OMe in the template is more selective
against insertion of the more hydrophilic natural triphosphates
than MM1. The effects can also be complex; for example, within
the DM5 scaffold, a minor groove methoxy group slightly
increases the rate of dMM1TP and dTMTP insertion but
significantly decreases the rate of dDM5TP insertion. These
triphosphate analogues differ only by single methyl substituents,
and the rates do not parallel the extent of substitution. In total,
these results suggest that the methoxy substituents are capable
of mediating specific structural and/or electrostatic effects that
contribute to efficient unnatural base pair synthesis.
As discussed above, continued primer extension after syn-
thesis of the unnatural base pair has traditionally limited the
replication of DNA containing unnatural base pairs. We find
that primers that terminate with a minor groove methoxy group
paired opposite a template analogue with a minor groove methyl
group are efficiently extended. Remarkably, the MMO2:TM
(primer:template) pair is extended only 100-fold slower than a
natural base pair in the same sequence context. It is not obvious
how the methyl group in the developing minor groove of the
template analogue contributes to efficient extension. However,
the fact that mispairs with natural pyrimidines in the template
are extended more efficiently than mispairs with purines
suggests that the role of both the methyl group and the carbonyl
group may be structural, perhaps helping to form a primer
terminus that is optimally packed and structured for continued
extension.
The effect of a minor groove methoxy group at the primer
terminus is consistent and significant. For example, when paired
opposite TM, MMO2 is extended more than 40-fold more
efficiently than its carbocyclic analogue DM5. Although this
difference is less than the approximately 3500-fold difference
observed with dG and 3-deazaG,30 it is similar to that observed
with other unnatural nucleobases where H-bond acceptors have
been introduced.4,19,22,23,39,40 This suggests that while H-bonding
between the primer terminus and the polymarase is important,
other factors also contribute.
Methoxy groups of anisole are typically only moderate
H-bond acceptors, as suggested by their pKa of -6.5.41 However,
both experimental and theoretical studies indicate that this is
due in part to conjugation of the lone pairs of electrons on
oxygen into the aromatic ring, which requires the methyl group
to be in the plane of the ring.31-34 The same studies also indicate
that when the methyl group rotates out of the plane, the electrons
localize on the oxygen atom and it becomes a significantly better
H-bond acceptor. Interestingly, molecular dynamics simulations
suggest that the minor groove methyl groups of the unnatural
base pairs rotate out of the plane of the phenyl ring to optimize
packing interactions with flanking nucleobases (S.M and F.R.
unpublished results). This, along with the “preordered” position-
ing of the polymerase-based H-bond donor, Arg668, appears
Discussion
Extension of unnatural base pairs by DNA polymerases
generally limits their replication, and thus understanding how
to facilitate extension is critical for developing viable unnatural
base pairs.1 Previously, we examined phenyl-based nucleotides
bearing H-bond acceptors that, when at the primer terminus,
are expected to be oriented in the developing minor groove
where they might engage a conserved polymerase-based H-bond
donor. Specifically, we examined both minor groove carbonyl
groups23 and nitrogen substituents.22 In addition, Hirao and co-
workers have positioned an aldehyde group in the minor
groove.5,39,40 While these modifications facilitate extension, they
also destabilize the base pair and, at least with the pyridone
and pyridine nucleobases, appear to be recognized by the
H-bond donor of dG, resulting in poor thermal and replication
fidelity.
We are interested in nucleobase modifications that might
facilitate H-bonding with the polymerase but not with a natural
nucleobase. Thermodynamic studies have indicated that, when
positioned ortho to the glycosidic linkage, methoxy groups are
desolvated upon duplex formation (S.M. and F.R. unpublished
results). This suggests that the H-bonding strength of the minor
groove methoxy groups is insufficient to maintain solvation.
Nonetheless, we were interested in examining whether these
minor groove methoxy groups can engage the “preordered”
H-bond donor of polymerases that is known to be required to
align the primer terminus for efficient extension. Thus, we
examined the effects of methoxy group substituents on poly-
merase-mediated replication.
Many previous studies have identified interbase hydrophobic
interactions as a major force underlying the synthesis of pairs
(38) Potapova, O.; Chan, C.; DeLucia, A. M.; Helquist, S. A.; Kool, E. T.;
Grindley, N. D.; Joyce, C. M. Biochemistry 2006, 45, 890-898.
(39) Mitsui, T.; Kimoto, M.; Sato, A.; Yokoyama, S.; Hirao, I. Bioorg. Med.
Chem. Lett. 2003, 13, 4515-4518.
(40) Mitsui, T.; Kitamura, A.; Kimoto, M.; To, T.; Sato, A.; Hirao, I.; Yokoyama,
S. J. Am. Chem. Soc. 2003, 125, 5298-5307.
(41) Arnett, E. M.; Wu, C. Y. J. Am. Chem. Soc. 1960, 82, 4999-5000.
9
5556 J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007

Optimization of Unnatural Base Pair Replication
A R T I C L E S
Tributylamine (5 equiv) was added, followed by a solution of
tributylammonium pyrophosphate (5 equiv) in DMF (final concentration
∼0.15 M). After 3 min, the reaction was quenched by addition of 1 M
aqueous triethylammonium bicarbonate (10 vol equiv). The resulting
crude solution was stirred for 30 min at 0 °C and then lyophilized.
The crude material was purified by reversed phase HPLC (C18 column,
1-35% CH3CN in 0.1 M NEt3-HCO3, pH 7.5) followed by lyophiliza-
tion to afford the triphosphate as a white solid. TM and DM5
triphosphates were synthesized as described previously.9,12
2OMe triphosphate: 31P NMR (162 MHz, D2O) δ -5.90 (d, J )
21.2 Hz), -10.55 (d, J ) 19.8 Hz), -22.16 (t, J ) 20.7 Hz). 3OMe
triphosphate: 31P NMR (162 MHz, D2O) δ -10.32 (d, J ) 19.9 Hz),
-10.76 (d, J ) 20.4 Hz), -22.83 (t, J ) 19.9 Hz). 4OMe triphos-
phate: 31P NMR (162 MHz, D2O) δ -10.39 (d, J ) 19.9 Hz), -10.71
(d, J ) 20.1 Hz), -22.82 (t, J ) 19.9 Hz). DMO triphosphate: 31P
NMR (162 MHz, D2O) δ -6.01 (d, J ) 21.2 Hz), -10.52 (d, J ) 19.9
Hz), -22.21 (t, J ) 20.6 Hz). MMO1 triphosphate: 31P NMR (162
MHz, D2O) δ -9.35 (d, J ) 19.3 Hz), -10.67 (d, J ) 20.1 Hz), -22.68
(t, J ) 20.1 Hz). MMO2 triphosphate: 31P NMR (162 MHz, D2O) δ
-5.91 (d, J ) 21.2 Hz), -10.56 (d, J ) 19.6 Hz), -22.17 (t, J ) 20.6
Hz). MM1 triphosphate: 31P NMR (162 MHz, D2O) δ -8.03 (bs),
-10.63 (d, J ) 19.9 Hz), -22.48 (t, J ) 20.4 Hz).
Gel-Based Kinetic Assay. Primer oligonucleotides were 5′-radio-
labeled with T4 polynucleotide kinase (New England Biolabs) and
[γ-33P]-ATP (GE Biosciences). Primers were annealed to template
oligonucleotides in the reaction buffer by heating to 90 °C followed
by slow cooling to ambient temperature. Assay conditions included 40
nM primer/template, 0.1-1.3 nM enzyme (either Kf or Kf R668A),
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 2 × 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 and 8 M urea
denaturing gel electrophoresis, and radioactivity was quantified using
a PhosphorImager (Molecular Dynamics) and ImageQuant software.
The Michaelis-Menten equation was fit to a plot kobsd versus triphos-
phate concentration using the program Kaleidagraph (Synergy Soft-
ware). The data presented are averages of three independent determi-
nations.
to be sufficient to form a functional H-bond between the primer
terminus and the polymerase. This conclusion is strongly
supported by the selective decrease in R668A Kf-mediated
extension efficiency of the MMO2:TM pair relative to the
DM5:TM pair. The “preordering” of the Arg668 donor within
the polymerase-DNA complex likely underlies the ability of
the minor groove methoxy group to act as an H-bond acceptor
despite its reduced ability to H-bond with minor groove waters
of solvation.
It appears that an ortho positioned methoxy group is able to
form a productive H-bond with the polymerase that appropriately
aligns the primer terminus for continued extension. This, and
the observation that these substituents do not increase mispair
recognition or stability, as has been observed with other H-bond
acceptors,22,23 suggests that suitably positioned methoxy groups
will be of great value in the design of new unnatural base pairs.
Experimental Section
General Methods. Chemical reagents were purchased from Sigma-
Aldrich and used without further purification, unless otherwise stated.
All unnatural nucleosides and nucleotides used in this study were
synthesized as described in the Supporting Information. All reagents
for oligonucleotide synthesis were purchased from Glen Research.
Oligonucleotides were synthesized using an Applied Biosystems Inc.
392 DNA/RNA synthesizer. 31P NMR spectra were recorded on a
Bruker AMX-400 spectrometer. Coupling constants (J values) are
reported in Hz. The chemical shifts are given in δ (ppm) using 85%
H3PO4 in D2O for 31P NMR as an external standard. T4 polynucleotide
kinase and Klenow fragment exo- were purchased from New England
Biolabs. [γ-33P]-ATP was purchased from Amersham Biosciences. The
R668A/D424A double mutant of Kf was a generous gift from Catherine
M. Joyce (Yale University). The D424A mutation renders the poly-
merase exonuclease deficient. For simplicity, the double mutant is
referred to as the R668A mutant of the exonuclease deficient poly-
merase.
Synthesis of Oligonucleotides. Oligonucleotides were prepared by
the â-cyanoethylphosphoramidite method on controlled pore glass
supports (1 µmol) using an Applied Biosystems Inc. 392 DNA/RNA
synthesizer as the standard method. After automated synthesis, the
oligonucleotides were cleaved from the support by concd aqueous
ammonia for 1 h at room temperature, deprotected by heating at 55 °C
for 12 h, and purified by denaturing polyacrylamide gel electrophoresis
(12-20%, 8 M urea). The primer oligonucleotides containing unnatural
bases at the 3′-end were obtained using a Universal Support, or 3′-
phosphate CPG, which was treated with alkaline phosphatase after
deprotection according to manufacturer’s protocols.
Acknowledgment. Funding was provided by the National
Institutes of Health (GM60005).
Supporting Information Available: Nucleoside and nucle-
otide synthesis and characterization. Representative kinetics plots
for heteropair synthesis, misincorporation, correct pair extension,
and mispair extension. This material is available free of charge
via the Internet at http://pubs.acs.org.
General Triphosphate Synthesis Procedure. Proton sponge (1.5
equiv) and nucleoside (1 equiv) were dissolved in trimethylphosphate
(final concentration ∼0.3 M) and cooled to 0 °C. POCl3 (1.05 equiv)
was added dropwise, and the mixture was stirred at 0 °C for 2 h.
JA068282B
9
J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007 5557