Stability and polymerase recognition of pyridine nucleobase analogues: Role of minor-groove H-bond acceptors

Article abstract of DOI:10.1002/anie.200602579

In the groove: In an effort to develop unnatural base pairs, pyridyl nucleoside analogues were synthesized and characterized (see structures). An α-glycosidic nitrogen atom provides an H-bond acceptor that does not significantly facilitate pairing with na

Full text of DOI:10.1002/anie.200602579

Angewandte
Chemie
which would expand the genetic alphabet and lay the
foundations of an expanded genetic code. Studies from the
Benner and Kool groups have elegantly demonstrated that
the interbase hydrogen bonds (H-bonds)may be rear-
ranged,[1] or removed altogether,[2] without complete loss of
stability or polymerase recognition; this information has
inspired design strategies based on the use of predominantly
hydrophobic nucleobase analogues.[3]
While base-pair stability and efficient replication are
required for any unnatural base-pair candidate, predomi-
nantly hydrophobic analogues are consistently limited by
extension, that is, efficient primer elongation after synthesis
of the unnatural base pair at a primer terminus.[3a,c] One
possible explanation for inefficient extension is that, while
properly positioned H-bonds are not required to mediate
interbase recognition, they are required at the primer
terminus to appropriately engage the polymerase for efficient
DNA synthesis.[4,5] We have previously optimized nucleobase
analogues by the inclusion of natural-nucleobase-like car-
bonyl groups in a position b to an N-glycosidic linkage.[6]
However, this natural-like H-bond acceptor also reintroduces
interbase H-bonds and facilitates mispairing. By contrast, H-
bond acceptors at the a position of C-nucleosides should be
less recognized by the natural nucleobases but may still be
able to engage the polymerase appropriately for DNA
synthesis.
Here, we report the synthesis and analysis of three pyridyl
nucleotide analogues that are designed to explore the effect of
positioning a nitrogen atom in the minor groove (2Py), the
interbase interface (3Py), or the major groove (4Py)
(Scheme 1). We focus on the evaluation of the corresponding
Nonnatural Base Pairs
DOI: 10.1002/anie.200602579
Stability and Polymerase Recognition of Pyridine
Nucleobase Analogues: Role of Minor-Groove
H-Bond Acceptors**
Scheme 1. Unnatural nucleobases. BEN[3c,7b] and PYR[6] have been
previously reported and are included for comparison (see the text).
Yoonkyung Kim, Aaron M. Leconte, Yoshiyuki Hari,
and Floyd E. Romesberg*
self pairs (that is, pairs formed between two identical
analogues). Each nucleoside was synthesized, converted into
the corresponding phosphoramidite, and incorporated into
DNA as described in the Supporting Information.
The information potential of a genome is limited by the four-
letter, two-base-pair genetic alphabet. Significant efforts have
been directed towards developing a third, unnatural base pair,
To explore the effect of aza substitution on thermal
stability and selectivity, the melting temperatures (Tm)of
duplexes containing the unnatural base pairs or mispairs were
determined (Table 1). The 2Py self pair was the most stable
(Tm = 52.28C), followed by the 3Py and 4Py self pairs (Tm =
50.2 and 48.18C, respectively). Thus, the introduction of a
single nitrogen substituent can alter self-pair stability by more
than 48C. This is in contrast to the much smaller variations
associated with fluoro, bromo, cyano, or methyl derivatiza-
tion, which were each less than 28C.[3b,7] For all but 2Py, the
self pairs are significantly less stable than the parental,
unsubstituted BEN self pair (Tm = 52.88C; Scheme 1). We
also examined the stability of heteropairs formed between the
pyridyl nucleotides; heteropairs containing 2Py were again
[*] Y. Kim, A. M. Leconte, Y. Hari, Prof. Dr. F. E. Romesberg
Department of Chemistry
The Scripps Research Institute
10550 NorthTorrey Pines Road, La Jolla, CA 92037 (USA)
Fax: (+1)858-784-7472
E-mail: floyd@scripps.edu
[**] We acknowledge the National Institutes of Health (GM60005 to
F.E.R.) and the Uehara Memorial Foundation (Y.H.) for financial
support.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2006, 45, 7809 –7812
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7809

Communications
Table 1: Denaturation temperatures for duplex DNA containing 2Py, 3Py, and 4Py.[a]
sion of the fully carbocyclic BEN self pair,
by incorporation of deoxycytidine triphos-
phate (dCTP)opposite deoxyguanosine
5’-dGCGTACXCATGCG
3’-dCGCATGYGTACGC
X
Y
Tm [8C]
X
Y
Tm [8C]
X
Y
Tm [8C]
(dG)in the template, occurs with
a
second-order rate constant of 1.6 ”
103 mÀ1 minÀ1.[3c] The 3Py and 4Py self pairs
are extended with rates that are too low to
detect (kcat/KM < 103 mÀ1 minÀ1). However,
the 2Py self pair is extended by Kf with a
kcat/KM value of 3.4 ” 104 mÀ1 minÀ1, at least
34-fold more efficiently than the other
pyridyl analogues. Thus, as with base-pair
stability, aza substitution has a large effect
2Py
2Py
2Py
2Py
2Py
2Py
2Py
2Py
3Py
4Py
A
C
G
52.2
51.1
49.6
50.2
44.2
49.2
47.7
3Py
3Py
3Py
3Py
3Py
3Py
3Py
2Py
3Py
4Py
A
C
G
51.1
50.2
49.1
50.2
44.7
52.7
49.2
4Py
4Py
4Py
4Py
4Py
4Py
4Py
2Py
3Py
4Py
A
C
G
50.1
49.6
48.1
49.1
45.1
49.6
50.1
T
T
T
[a] Determined in 10 mm 1,4-piperazinediethanesulfonic acid (PIPES), 10 mm MgCl2, and 100 mm NaCl
(pH 7). The error in the Tm values is Æ0.18C
the most stable, although they were not as stable as the 2Py
self pair (Table 1).
Table 2: Correct extension of unnatural self pairs.[a]
5’-dTAATACGACTCACTATAGGGAGAX
The stability of the pyridyl base pairs probably results
from a compromise between stabilizing hydrophobic forces,
due to transfer of the hydrophobic analogues into the less
polar duplex interior, and destabilizing desolvation effects.
Forced desolvation will destabilize base-pair formation as the
pyridyl analogues are expected to be more favorably solvated
in single-stranded DNA.[8] Forced desolvation also explains
the relative stability of the pairs involving 2Py, as only its
nitrogen atom will be oriented into the duplex minor groove
where stabilizing H-bonds with water molecules from the
spine of solvation[9] can offset the cost of desolvation.
Interestingly, as described above, 2Py formed more stable
unnatural base pairs than 3Py and 4Py; however, it did not
form more stable mispairs (Table 1). Thus, the 2Py self pair is
thermally selective by 2.0–8.08C. This compares favorably to
the thermal selectivity observed among natural base pairs.[10]
The thermal selectivity of the 3Py and 4Py self pairs is limited
by stable mispairing with dG or dT, respectively. In each case,
the stability of the mispair may result from H-bonding
(Scheme 2).[9b,11] While only speculative, these arguments
3’-dATTATGCTGAGTGATATCCCTCTXGCTAGGTTACGGCAGGATCGC
fldCTP
5’-dTAATACGACTCACTATAGGGAGAXC
3’-dATTATGCTGAGTGATATCCCTCTXGCTAGGTTACGGCAGGATCGC
[b]
[b]
Unnatural self pair kcat[b] [minÀ1
]
KM [mm] kcat/KM [mÀ1 minÀ1
]
2Py
3Py
4Py
2.1Æ0.8
61Æ17
3.4”104
<1.0”103
<1.0”103
n.d.[c]
n.d.[c]
n.d.[c]
n.d.[c]
[a] See the Supporting Information for experimental details. [b] kcat =rate
of catalysis, KM =Michaelis constant. [c] n.d.=not determined. Rates
were too slow for the determination of kcat and KM independently.
on self-pair extension, with substitution at the ortho position
substantially increasing the rate of extension and substitution
at the meta or para positions decreasing the rate of extension.
The fidelity of 2Py self-pair extension was also high; the rates
of self-pair extension by deoxyguanosine triphosphate
(dGTP), deoxyadenosine triphosphate (dATP), or deoxythy-
midine triphosphate (dTTP)insertion opposite dG were all
too low to detect (kcat/KM < 103 mÀ1 minÀ1).
The efficiency and fidelity of 2Py self-pair extension
suggests that the ortho-pyridyl nitrogen atom is positioned in
the developing minor groove and productively engages the
polymerase. Interestingly, extension of the 2Py self pair is
only about twofold less efficient than the pyridone scaffold
(PYR, Scheme 1; extended at 8.2 ” 104 mÀ1 minÀ1).[6] This small
difference is somewhat surprising when it is considered that
the minor-groove H-bond acceptor of 2Py is located at the a
position relative to the glycosidic bond, whereas the H-bond
acceptor of PYR is located at a position b to the glycosidic
bond, which is more analogous to the natural nucleobases
(Scheme 3).
Scheme 2. H-bonding interactions that may underlie the relative stabil-
ity of the 3Py:dG and 4Py:dT mispairs.
are consistent with the idea that desolvation is important for
both mispair and unnatural base-pair stability, as H-bonds
within the duplex environment would again compensate for
the H-bonds with water that are lost upon duplex formation.
Regardless of the interpretation, it appears that the minor-
groove H-bond acceptor at the a position of 2Py is not
appropriately positioned to interact with the H-bond donors
of the natural nucleobases.
The ability of the exonuclease-deficient Klenow fragment
of Escherichia coli DNA polymerase I (Kf)to extend each self
pair was examined by using single-nucleotide-incorporation
steady-state kinetics (Table 2). As reported previously, exten-
Scheme 3. Minor-groove H-bond acceptors within natural and unnatu-
ral scaffolds.
7810
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 7809 –7812

Angewandte
Chemie
Based on the efficiency and fidelity of 2Py self-pair
extension, we synthesized the corresponding nucleoside
triphosphate (see the Supporting Information)and examined
the efficiency of Kf-mediated self-pair synthesis (Table 3).
to its efficient synthesis. dCTP was not efficiently inserted
opposite any of the pyridyl analogues. Overall, the rates at
which 2Py directs the synthesis of mispairs are similar to those
of BEN,[3c] a result suggesting that the minor-groove H-bond
acceptor has only a minor effect on mispair synthesis.
The data demonstrate that the phenyl-ring nucleobase
scaffold may be optimized for both base-pair stability and
replication by derivatization with an ortho nitrogen atom.
This H-bond acceptor, which is at the a position relative to
the glycosidic linkage, appears to be well positioned to
mediate critical minor-groove interactions with water mole-
cules of solvation or with polymerase-based H-bond donors.
In contrast to H-bond acceptors introduced at the b posi-
tion,[6] this H-bond acceptor does not stabilize mispairs in
duplex DNA. While the ortho nitrogen atom does appear to
somewhat facilitate dGTP mispair synthesis, the mispair is
synthesized only inefficiently and, in fact, not significantly
faster than mispairs among the natural nucleotides.[12] More
problematic is the insertion of dATP opposite 2Py, which is
probably mediated by hydrophobic interactions. However,
dATP insertion opposite 2Py is less efficient than with other
scaffolds that have already been successfully optimized by
derivatization.[7,3b,c] Combination of these modifications with
ortho aza substitution may result in the simultaneous
optimization of interbase interactions and nucleobase–poly-
merase interactions and it should result in unnatural base
pairs that are both more stable and efficiently replicated.
Table 3: Triphosphate insertion opposite pyridyl analogues in the
template.[a]
5’-dTAATACGACTCACTATAGGGAGA
3’-dATTATGCTGAGTGATATCCCTCTXGCTAGGTTACGGCAGGATCGC
fldYTP
5’-dTAATACGACTCACTATAGGGAGAY
3’-dATTATGCTGAGTGATATCCCTCTXGCTAGGTTACGGCAGGATCGC
X
Y
kcat [minÀ1
]
KM [mm]
kcat/KM [mÀ1 minÀ1
]
2Py
2Py
A
C
G
T
A
C
G
T
A
C
0.92Æ0.16
3.6Æ1.2
0.16Æ0.02
0.8Æ0.3
0.16Æ0.06
6.8Æ0.9
n.d.[b]
149Æ12
18Æ2
98Æ30
41Æ9
39Æ9
25Æ0.5
n.d.[b]
6.2”103
2.0”105
3.2”103
2.0”104
4.1”103
2.7”105
<1.0”103
1.9”105
6.5”103
5.5”105
<1.0”103
<1.0”103
3.3”104
3Py
4Py
11Æ3
58Æ13
83Æ15
20Æ6
n.d.[b]
0.54Æ0.05
11Æ3
n.d.[b]
G
T
n.d.[b]
n.d.[b]
1.6Æ0.3
49Æ16
[a] See the Supporting Information for experimental details. [b] Rates
were too slow for the determination of kcat and KM independently.
Received: June 28, 2006
The self pair is synthesized with a second-order rate constant
of 6.2 ” 103 mÀ1 minÀ1. While this synthesis is less efficient than
that of self pairs comprised of nucleobase analogues with a
large, extended aromatic surface area[3a] or more highly
substituted phenyl rings,[3b,c] it is actually high compared to
other minimally substituted scaffolds. For example, the BEN
and PYR self pairs are both synthesized at rates less than
103 mÀ1 minÀ1.[3c,6]
Keywords: DNA replication · hydrophobic effects · nucleobases ·
.
polymerases · pyridines
[1] J. L. Piccirilli, T. Krauch, S. E. Moroney, S. A. Benner, Nature
1990, 343, 33 – 37.
[2] S. Moran, R. X.-F. Ren, S. Rumney IV, E. T. Kool, J. Am. Chem.
Soc. 1997, 119, 2056 – 2057.
We also examined the rates at which each natural
triphosphate is inserted opposite each pyridyl analogue
(Table 3). As with most small hydrophobic analogues, dATP
was the most efficiently incorporated deoxynucleoside tri-
phosphate (dNTP)opposite the pyridyl analogues. The rate of
dATP insertion showed little dependence on aza substitution
and ranged from only 2 ” 105–5.5 ” 105 mÀ1 minÀ1. The magni-
tude and template independence of the rate of dATP
insertion suggests that it results from nonspecific hydrophobic
interactions. dGTP was the next most efficiently inserted
natural triphosphate opposite 2Py and 3Py, but its insertion
was not detectable opposite 4Py. The rates are at least roughly
correlated with the expected ability of the pyridine moiety to
H-bond with guanosine (Scheme 2). The absolute rate at
which the 2Py:dG, PYR:dG (1.2 ” 104 mÀ1 minÀ1), and
BEN:dG (1.6 ” 103 mÀ1 minÀ1)mispairs are formed suggests
that while the ortho nitrogen atom does not form stable H-
bonds with dG in duplex DNA, it facilitates dGTP insertion,
although less efficiently than the meta nitrogen substituent of
3Py. dTTP misincorporation opposite 2Py and 3Py proceeded
with rates of ꢀ 5 ” 103 mÀ1 minÀ1; however, its insertion was an
order of magnitude more efficient opposite 4Py. Perhaps the
same factors that stabilize the 4Py:dT mispair also contribute
[3] a)D. L. McMinn, A. K. Ogawa, Y. Wu, J. Liu, P. G. Schultz, F. E.
Romesberg, J. Am. Chem. Soc. 1999, 121, 11585 – 11586;
b)A. A. Henry, A. G. Olsen, S. Matsuda, C. Yu, B. H. Geier-
stanger, F. E. Romesberg, J. Am. Chem. Soc. 2004, 126, 6923 –
6931; c)S. Matsuda, A. A. Henry, F. E. Romesberg, J. Am.
Chem. Soc. 2006, 128, 6369 – 6375; d)T. Mitsui, A. Kitamura, M.
Kimoto, T. To, A. Sato, I. Hirao, S. Yokoyama, J. Am. Chem. Soc.
2003, 125, 5298 – 5307; e)K. Kinkaid, J. Beckman, A. Zivkovic,
R. L. Halcomb, J. W. Engels, R. D. Kuchta, Nucleic Acids Res.
2005, 33, 2620 – 2628.
[4] Y. Li, G. Waksman, Protein Sci. 2001, 10, 1225 – 1233.
[5] A. S. Meyer, M. Blandolino, T. E. Spratt, J. Biol. Chem. 2004,
279, 33043 – 33046.
[6] A. M. Leconte, S. Matsuda, G. T. Hwang, F. E. Romesberg,
Angew. Chem. 2006, 118, 4432 – 4435; Angew. Chem. Int. Ed.
2006, 45, 4326 – 4329.
[7] a)G. T. Hwang, F. E. Romesberg, Nucleic Acids Res. 2006, 34,
2037 – 2045; b)S. Matsuda, F. E. Romesberg, J. Am. Chem. Soc.
2004, 126, 14419 – 14427.
[8] a)I. Papai, G. Jansco, J. Phys. Chem. A 2000, 104, 2132 – 2137;
b)L. Partay, P. Jedlovszky, G. Jansco, J. Phys. Chem. B 2005, 109,
8097 – 8102.
[9] a)A. N. Lane, T. C. Jenkins, T. A. Frenkiel, Biochim. Biophys.
Acta 1997, 1350, 205 – 220; b)F. Eisenhaber, V. G. Tumanyan, F.
Eisenmenger, W. Gunia, Biopolymers 1989, 28, 741 – 761; c)F.-Y.
Angew. Chem. Int. Ed. 2006, 45, 7809 –7812
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7811

Communications
Dupradeau, D. A. Case, C. Yu, R. Jimenez, F. E. Romesberg, J.
Am. Chem. Soc. 2005, 127, 15612 – 15617.
[10] M. Berger, A. K. Ogawa, D. L. McMinn, Y. Wu, P. G. Schultz,
F. E. Romesberg, Angew. Chem. 2000, 112, 3069 – 3071; Angew.
Chem. Int. Ed. 2000, 39, 2940 – 2942.
[11] a)Y. X. Wang, S. Huang, D. E. Draper, Nucleic Acids Res. 1996,
24, 2666 – 2672; b)L. Zendlov, P. Hobza, M. Kabel, ChemPhys-
Chem 2006, 7, 439 – 447.
[12] K. A. Eckert, T. A. Kunkel, J. Biol. Chem. 1993, 268, 13462 –
13471.
7812
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 7809 –7812