Binding of Imidazole-Derived Nucleosides to a CG Base Pair

Article abstract of DOI:10.1021/jo035597q

Novel imidazole nucleosides with substituents of different flexibility were studied for their binding to a CG Watson-Crick base pair by ¹H NMR spectroscopy in an aprotic solvent. Thermodynamic data as determined by titration experiments at diff

Full text of DOI:10.1021/jo035597q

Bin d in g of Im id a zole-Der ived Nu cleosid es
to a CG Ba se P a ir
Maria G. M. Purwanto and Klaus Weisz*
Institut fu¨r Chemie und Biochemie,
Ernst-Moritz-Arndt-Universita¨t Greifswald,
Soldmannstrasse 16, D-17489 Greifswald, Germany
weisz@uni-greifswald.de
Received October 30, 2003
F IGURE 1. Hydrogen bond donor and acceptor sites of a CG
base pair and binding of a potential ligand with a complemen-
tary acceptor (A)-donor (D) motif; R ) O-protected ribose
sugar.
Abstr a ct: Novel imidazole nucleosides with substituents
of different flexibility were studied for their binding to a CG
Watson-Crick base pair by 1H NMR spectroscopy in an
aprotic solvent. Thermodynamic data as determined by
titration experiments at different temperatures reveal the
influence of the substituent on the enthalpy and entropy of
complex formation and thus on the strength of binding.
SCHEME 1. Syn th esis of 2 a n d 3
DNA triplex formation through Hoogsteen hydrogen
bonds between third-strand bases and the bases of the
duplex target is mostly limited to homopurine sequences.
However, excluding the effective recognition of pyrimi-
dine interruptions within a purine tract of the duplex
generally results in triple helices of low stability, thus
restricting potential applications. To overcome these
limitations, various nucleobase analogues have been
designed in the past as potential ligands for any given
base pair, but the reliable prediction of their interactions
within a triple helix has not yet been achieved.1,2 Re-
cently, we have reported the specific binding of a novel
urocanamide derived nucleoside 1 toward a CG Watson-
Crick base pair at its major groove side in an organic
solvent.3 Based on these results, we now describe binding
studies involving different structural variants 1-4.
All the nucleobases are expected to facilitate the
formation of two hydrogen bonds to a CG base pair, i.e.,
to NH2 of cytosine and the O6 carbonyl oxygen of guanine
(see Figure 1). To understand the effect of conformational
flexibility on base triple formation, nucleobases 3 and 4
were designed with increased and decreased side chain
flexibility, whereas 2 was additionally dedicated to study
the effect of an N-propyl substituent at the amide
function.
As was described for nucleoside 13, 2 was synthesized
starting with trans-urocanic acid, which after activation
using N-hydroxysuccinimide (NHS) and N-(3-dimethyl-
aminopropyl)-N′-ethylcarbodiimide hydrochloride (DCI)
reacted with propylamine to afford the corresponding
N-propylurocanamide (Scheme 1). This was glycosylated
to the O-protected nucleoside with 1-O-acetyl-2,3,5-tri-
O-benzoyl-â-D-ribofuranose under Vorbru¨ggen conditions.
Compound 1 was smoothly converted to 3 by a Pd-
mediated reduction of the double bond which did not
affect the sugar or the imidazole moiety. Compound 4
was synthesized based on the Bredereck imidazole syn-
thesis4 recently also employed by Griffin et al.5 This
procedure affords 2-unsubstituted imidazoles from an
R-hydroxy or an R-halogenated ketone (in this case
ω-bromo-m-nitroacetophenone) by using formamide as a
* Corresponding author.
(1) (a) Huang, C.-Y.; Miller, P. S. J . Am. Chem. Soc. 1993, 115,
10456-10457. (b) Sasaki, S.; Nakashima, S.; Nagatsugi, F.; Tanaka,
Y.; Hisatome, M.; Maeda, M. Tetrahedron Lett. 1995, 36, 9521-9524.
(c) Zimmerman, S. C.; Schmitt, P. J . Am. Chem. Soc. 1995, 117, 10769-
10770. (d) Lehmann, T. E.; Greenberg, W. A.; Liberles, D. A.; Wada,
C. K.; Dervan, P. B. Helv. Chim. Acta 1997, 80, 2002-2022. (e) Lecubin,
F.; Benhida, R.; Fourrey, J .-L.; Sun, J .-S. Tetrahedron Lett. 1999, 40,
8085-8088. (f) Lengeler, D.; Weisz, K. Tetrahedron Lett. 2001, 42,
1479-1481. (g) Guianvarc’h, D.; Benhida, R.; Fourrey, J .-L.; Maurisse,
R.; Sun, J .-S. J . Chem. Soc., Chem. Commun. 2001, 1814-1815.
(2) For reviews, see: (a) Doronina, S. O.; Behr, J .-P. Chem. Soc. Rev.
1997, 63-71. (b) Gowers, D. M.; Fox, K. R. Nucleic Acids Res. 1999,
27, 1569-1577. (c) Purwanto, M. G. M.; Weisz, K. Curr. Org. Chem.
2003, 7, 427-446.
(3) Purwanto, M. G. M.; Lengeler, D.; Weisz, K. Tetrahedron Lett.
2002, 43, 61-64.
10.1021/jo035597q CCC: $27.50 © 2004 American Chemical Society
Published on Web 12/05/2003
J . Org. Chem. 2004, 69, 195-197
195

reagent as well as a solvent and a source of ammonia. A
Vorbru¨ggen reaction and a subsequent reduction of the
nitro group with Raney nickel and hydrazine as hydrogen
source yielded the amino-substituted nucleoside.
By employing concentration- and temperature-depend-
ent 1H NMR measurements, the thermodynamics of
binding for each nucleoside analogue toward a CG base
pair was examined in methylene chloride and analyzed
in terms of enthalpic and entropic contributions. As a first
step of our binding studies, we have examined the self-
association of the nucleoside analogues through concen-
1
tration dependent H NMR chemical shifts. None of them
was found to significantly dimerize (Kassoc e 7 M-1 for T
g 200 K) under the given experimental conditions,
prerequisite for the reliability of using a simple 1:1
association model in the case of CG binding. Also, the
strong binding between C and G in apolar solvents (Kassoc
∼105 M-1) allowed a 1:1 C:G mixture to be treated as a
single species.1c,f Correspondingly, NMR measurements
were performed in CD2Cl2 by titrating the nucleoside
analogue (∼2 mM) with a 1:1 mixture of di-O-triisopro-
pylsilyl-2′-deoxycytidine and di-O-triisopropylsilyl-2′-
deoxyguanosine. A representative titration experiment
is shown for 3 in Figure 2a. Progressive downfield shifts
of the hydrogen-bonded NH protons of the analogues
were subjected to a nonlinear regression analysis for a
1:1 complex to give an association constant between the
nucleoside analogue and the CG base pair (Table 1).6
These results were further confirmed by reverse titra-
tions observing the non-Watson-Crick bound cytosine
amino resonance (not shown).7
A van’t Hoff analysis of the temperature-dependent
association constants as shown in Figure 2b for the
3‚CG asociation yielded enthalpies and entropies of
association. As can be seen from Table 1, a similar heat
of complex formation -∆H was obtained for 1 and 3
complexes, i.e., 14-15 kJ /mol. With the formation of two
hydrogen bonds this gives about 7 kJ /mol per hydrogen
bond. A smaller value obtained for 4 (4.4 kJ /mol per
hydrogen bond) can be attributed to the amino group
being a weaker hydrogen bond donor compared to an
amide function but also suggests that the rigidity of the
base hampers conformational readjustments to optimize
the two hydrogen bond contacts.8
Overall, the values determined are considerably smaller
than the heat of formation of adenine-uracil (AU) as-
sociation measured in chloroform by Kyogoku et al. (13
kJ /mol per hydrogen bond).9 This might be attributed to
the cooperative nature of the two formed resonance
assisted hydrogen bonds and an optimal geometric fit
F IGURE 2. (a) NH proton chemical shift of 3 as a function of
the CG concentration in CD2Cl2 at 233 K; (b) van’t Hoff plot
of the association constant Kassoc for 3‚CG base triple formation
in CD2Cl2. Lines represent the least-squares fit.
TABLE 1. Su m m a r y of Ka ssoc, ∆H, a n d ∆S for th e
Associa tion of Nu cleosid e An a logu es 1-4 w ith a CG Ba se
P a ir
a
T (K)
Kassoc (M-1
)
∆H (kJ /mol)
-14.1 ( 1
∆S (J /K‚mol)
-17.9 ( 2
1
2
226
257
283
296
207
220
229
254
233
262
288
212
219
242
212
88
43
38
84
62
66
31
30
13
7
b
b
3
4
-14.7 ( 2
-8.7 ( 2
-35.0 ( 4
-13.5 ( 4
28
23
15
(4) (a) Bredereck, H.; Theilig, G. Chem. Ber. 1953, 86, 88-96. (b)
Theilig, G. Chem. Ber. 1953, 86, 96-109.
a
b
Estimated error about 10%. Not determined due to nonlin-
(5) Griffin, L. C.; Kiessling, L. L.; Beal, P. A.; Gillespie, P.; Dervan,
P. B. J . Am. Chem. Soc. 1992, 114, 7976-7982.
earity of van’t Hoff plot.
(6) For 1 the RH vinyl resonance, which undergoes a downfield shift
upon titration with CG in agreement with the proposed binding mode,
between the monomers in the AU base pair. It is also
important to note that the enthalpy determined here not
only reflects the formation of hydrogen bonds but also
includes the enthalpy of solvation, i.e., the entire en-
thalpy difference between solvated complex and solvated
was used for the determination of Kassoc
.
(7) Because of severe signal overlap in the spectra, we used
specifically 15N amino labeled cytidine for the reverse titration experi-
ments and an 1H-15N filter as implemented in a one-dimensional
HMQC experiment for cytosine amino proton observation.
(8) The weak NHO hydrogen bond formed by 4 is also manifested
in the small 4 amino downfield shift upon complex formation, i.e., the
average limiting chemical shift difference between monomer and
complex as obtained by curve fitting amounts to only 0.18 ppm as
compared to 0.78 ppm for the 3 amide resonance.
(9) Kyogoku, Y.; Lord, R. C.; Rich, A. J . Am. Chem. Soc. 1967, 89,
496-504.
196 J . Org. Chem., Vol. 69, No. 1, 2004

monomers. However, assuming a small difference be-
tween the enthalpy of solvation for the dimer and
monomers, -∆H might reasonably well reflect the hy-
drogen bond energy.
As expected, the entropy upon complex formation -∆S
closely follows the conformational flexibility of the base
analogue, ranging from -13.5 J /K‚mol for the most rigid
nucleoside analogue 4 to -35.0 J /K‚mol for the nucleoside
3 with the most flexible side chain. Consequently, smaller
association constants for 4 and 3 compared to 1 arise from
less favorable enthalpic and entropic contributions to the
free energy of association, respectively.
For the N-alkylated analogue 2, values of Kassoc are
smaller compared to 1 over the temperature range
studied. Moreover, a nonlinearity of the van’t Hoff plot
precluded the determination of reliable enthalpy and
entropy values for the 2‚CG association. Although the
Z-conformation 2a is expected to predominate in solution,
an additional temperature dependent Z-E conformational
equilibrium through the hindered rotation about the
C-N amide bond might attribute to these observations
(Figure 3).10 Clearly, only the 2a conformation can
selectively bind via two hydrogen bonds to the CG base
pair in the proposed binding mode.
F IGURE 3. Possible equilibria in a 2 + CG mixture. R )
O-protected ribose sugar.
more flexible 3 and the more rigid 4 imidazole derivative
significantly compromise complexation with a CG base
pair, respectively. Nucleoside 1 shows the highest affinity
toward a CG base pair and seems to be a promising
candidate for its incorporation into a triplex-forming
oligonucleotide targeting a double-helical DNA with a CG
interruption. Corresponding experiments on triplex for-
mation will be reported in due course.
Ack n ow led gm en t. We thank the Fonds der Chem-
ischen Industrie for financial support.
In summary, less favorable entropic and enthalpic
contributions to the hydrogen bond formation for the
Su p p or tin g In for m a tion Ava ila ble: Analytical and NMR
spectroscopic data including 1H and 13C NMR spectra for
compounds 2-4. This material is available free of charge via
the Internet at http://pubs.acs.org.
(10) Similar conformational equilibria have also been discussed by
Zimmerman et al. in the case of ligands with N-alkylureido substitu-
ents. Mertz, E.; Mattei, S.; Zimmerman, S. C. Org. Lett. 2000, 2, 2931-
2934.
J O035597Q
J . Org. Chem, Vol. 69, No. 1, 2004 197