COMMUNICATION
A pyrimidine-like nickel(II) DNA base pair
Christopher Switzer* and Dongwon Shin
Received (in Cambridge, UK) 5th October 2004, Accepted 12th November 2004
First published as an Advance Article on the web 19th January 2005
DOI: 10.1039/b415426f
purified by PAGE, and their identities confirmed by MALDI-
4-(29-Pyridyl)-pyrimidinone deoxyriboside is synthesized and
characterized as a DNA metallo base-pair; this novel nucleoside
forms a self-pair in the presence of Ni(II) and stabilizes double
helical DNA to the same extent as a G?C pair.
TOF mass spectrometry.
Pyrp metallo base-pair formation was assayed by UV monitored
thermal denaturation of the 4/5 duplex in the presence of divalent
metal ions (Table 1). Denaturation profiles are displayed in Fig. 3.
Specifically, the assay consisted of comparing Tm’s of Pyrp/Pyrp
containing duplex 4/5 in the presence of the various divalent metal
ions to the Tm obtained in the absence of any divalent ion (bottom
of first column, Table 1). Of the six divalent metal ions screened,
Ni2+ led to far and away the greatest duplex stabilization—a
dramatic increase in Tm of 16.5 uC relative to the metal free control.
Also significant, the data in Table 1 indicate Pyrp?Ni2+?Pyrp is as
stabilizing to a double helix as C?G (41.2 versus 40.2 uC). Finally,
in control experiments, essentially no effect was observed on Tm
values of the T/A or C/G duplexes when denatured in the presence
or absence of Ni2+ (bottom of second region, Table 1), or the
Pyrp?Ni2+?Pyrp duplex when denatured at pH 8 rather than 7. The
latter result when taken with a mixing curve determined in earlier
work5 on the parent 11/9 duplex supports a duplex (as opposed to
a triplex) structure for 4/5.
All natural nucleobases support some level of self-pairing. In a
genomic context, A, G, C and T self-pairs are undesirable and can
lead to mutations. Nevertheless, from the standpoint of de novo
design, bases capable of high fidelity self-recognition hold a
theoretical advantage of fewer possible mispairs and less synthetic
overhead, balanced against their diminished informational capa-
city. Synthetic base-pairs have been devised whose recognition
depends on van der Waals interactions,1 metal-coordination2 and
hydrogen-bonds,3 some of which rely on self-recognition. Here
we report the successful realization of the most improbable of
naturally inspired self-pairs, one based on a pyrimidine scaffold.
4-(29-Pyridyl)-pyrimidinone (Pyrp, Fig. 1) is found to bind
nickel(II) selectively over other divalent ions, forming
a
Pyrp?Ni?Pyrp base-pair with stability and mismatch discrimination
rivaling natural Watson–Crick pairs.
Mismatch discrimination of Pyrp was assessed by measuring the
stability of the four natural bases against Pyrp in the presence of
Ni2+ (top of second region, Table 1). These data show Pyrp?Ni2+ is
a mismatch against all four natural bases as DTm values of
the mismatched pairs relative to Pyrp?Ni2+?Pyrp ranged from 18.3–
21.9 uC (these values are distinct from the D values in Table 1 that
are rooted to the C/G pair). As a reference, natural T/G and C/A
mismatches of the parent duplex under the same conditions show
DTm values of 7.4 and 18.5 uC.5 Therefore, all four Pyrp?Ni2+
mismatches are similar to severe natural nucleobase mismatches in
their instability.
Pyrp (Fig. 1) is formally derived from the natural nucleobase
cytosine by replacement of its 4-amino group with pyridine. This
transformation leads to Lewis basic nitrogen atoms in an optimal
1,4 relationship for metal ion coordination. The synthesis of Pyrp is
summarized in Fig. 2. The critical step was a modified Negishi
coupling4 of pyridyl zinc bromide with the chloropyrimidinone
deoxyriboside derived from 1 to provide pyridylpyrimidinone
deoxyriboside 2. Nucleoside 2 was then converted in three steps to
phosphoramidite 3. Two complementary DNA dodecamer strands
bearing single Pyrp residues, 59-d-CTTTCTPyrpTCCCT (4) and
59-d-AGGGAPyrPAGAAAG (5), were prepared using an ABI
394 synthesizer and phosphoramidite 3. Oligonucleotides were
Three coordination geometries are possible in principle for
d
Pyrp?Ni2+?Pyrp: square planar, D2 , and tetrahedral. All other
factors being equal, a square planar geometry should be preferred
to minimize disruption of base-stacking in the DNA double helix.
Square planar geometries are predicted to be accessible for Ni2+,
Co2+, and Cu2+, three of the metal ions screened for their ability to
stablize the Pyrp2 bearing helix. Ab initio geometry optimization of
Pyrp?Ni2+?Pyrp at the B3LYP/6-31G*(CHN)/LACVP*(Ni) level of
theory led to a square-planar geometry as a (local) minimum
as depicted in Fig. 4a. Remarkably, the N1–N19 (pyrimidine
numbering) distance in the optimized Pyrp?Ni2+?Pyrp structure
˚
spans only 4.9 A (also depicted in Fig. 4a). In contrast, the
corresponding N9–N19, Pur–Pyr, distance in natural B-DNA
˚
helices for both G/C and A/T base-pairs is 9.1 A (structure not
shown). Thus, despite a predicted base to base distance of
approximately half (i.e., 54%) of the corresponding distance of a
natural base-pair, Pyrp?Ni2+?Pyrp nonetheless confers stability to a
helix equivalent to a G?C base-pair. Bipyridyl-29-deoxyriboside2g,6
Fig. 1 4-(29-Pyridyl)-pyrimidinone (Pyrp) metallo base-pair.
*switzer@citrus.ucr.edu
1342 | Chem. Commun., 2005, 1342–1344
This journal is ß The Royal Society of Chemistry 2005
Switzer, Christopher
