Synthesis and DNA duplex recognition of a triplex-forming oligonucleotide with an ureido-substituted 4-phenylimidazole nucleoside

Article abstract of DOI:10.1016/j.tetlet.2008.10.015

A 4-(3-n-butylureidophenyl)imidazole nucleoside was successfully incorporated into a triplex-forming oligonucleotide (TFO). Binding affinity and base pair selectivity of the TFO containing this non-natural nucleoside were studied with various duplex targets containing all four possible Watson-Crick base pairs opposite the nucleoside analog in the third strand. Triplex thermal stabilities indicate that the synthetic nucleoside acts as a universal base in binding to all four possible Watson-Crick base pairs with moderate affinity but poor selectivity. Based on an analysis of its binding thermodynamics, this can be rationalized by the absence of strong specific interactions and more favorable entropic contributions upon triplex formation.

Full text of DOI:10.1016/j.tetlet.2008.10.015

Tetrahedron Letters 49 (2008) 7264–7267
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Synthesis and DNA duplex recognition of a triplex-forming oligonucleotide
with an ureido-substituted 4-phenylimidazole nucleoside
*
Falk Wachowius, Michael Rettig, Gottfried Palm, Klaus Weisz
Institut für Biochemie, Ernst-Moritz-Arndt-Universität Greifswald, Felix-Hausdorff-Str. 4, D-17487 Greifswald, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
A 4-(3-n-butylureidophenyl)imidazole nucleoside was successfully incorporated into a triplex-forming
oligonucleotide (TFO). Binding affinity and base pair selectivity of the TFO containing this non-natural
nucleoside were studied with various duplex targets containing all four possible Watson–Crick base pairs
opposite the nucleoside analog in the third strand. Triplex thermal stabilities indicate that the synthetic
nucleoside acts as a universal base in binding to all four possible Watson–Crick base pairs with moderate
affinity but poor selectivity. Based on an analysis of its binding thermodynamics, this can be rationalized
by the absence of strong specific interactions and more favorable entropic contributions upon triplex
formation.
Received 31 August 2008
Revised 1 October 2008
Accepted 3 October 2008
Available online 8 October 2008
Keywords:
Nucleoside analog
Triple helix
Ó 2008 Elsevier Ltd. All rights reserved.
UV melting
4-Phenylimidazole
The use of triple-helix forming oligonucleotides in targeting de-
fined stretches of double-helical DNA with high affinity and selec-
tivity to generate three-stranded triplex structures has attracted
much interest in past years due to their various potential applica-
tions. Thus, triplex formation may be used to block the expression
of particular genes,1,2 for site-directed mutagenesis3,4 or for the
mapping of genomic DNA.5,6 Pyrimidine TFOs can bind parallel to
the purine strand of a Watson–Crick duplex in its major groove
through TꢀA or C+ꢀG Hoogsteen base pairing, yet binding suffers
from the limited base pair recognition code with poor binding at
inversion sites within a homopurineꢀhomopyrimidine tract of the
duplex. In order to overcome these limitations, substantial efforts
aimed at a better affinity and selectivity toward a mixed target se-
quence have been made through base, sugar, and backbone modi-
fications of the TFO.7 However, despite some success in achieving a
better base pair recognition, there is currently a lack of sufficient
structural and thermodynamic data that would allow a more reli-
able prediction for the binding of newly designed nucleotides
within a triple-helical complex.
Recently, we have synthesized various organic-soluble nucleo-
side receptors based on the 4-(3-aminophenyl)imidazole struc-
ture8 that was initially employed by the Dervan group in a
nucleoside surrogate.9 Their ability to recognize a free CG Wat-
son–Crick base pair primarily through specific hydrogen bond
interactions was tested in aprotic organic solvents.8 Based on 1H
NMR measurements in methylene chloride, strong binding was ob-
served for the 3-n-butylureido-substituted phenylimidazole deriv-
ative. Such alkylated ureido substituents were first introduced in a
naphth[1,2-d]imidazole base analog,10 and have also been subse-
quently employed for other non-natural bases to provide for two
potential hydrogen bonds to the guanine of a targeted CG base pair.
Indeed, subsequent 2D NOE experiments clearly indicated forma-
tion of a base triple with the novel N-alkylated butylureidophenyl
imidazole base associated via three hydrogen bonds to the CG base
pair (Fig. 1). In contrast, we only observed a weak and less specific
hydrogen bond-mediated complexation under such aprotic condi-
tions for the 4-(3-benzamidophenyl)imidazole receptor which had
previously been incorporated by Dervan and co-workers into a tri-
plex-forming oligonucleotide and has been shown to preferentially
bind opposite TA and CG base pairs, albeit through an unexpected
intercalative mode of binding.9,11 In order to shed more light on the
O
dRib
N
N
H
N
H
N
H
O
N
N
N
N
O
H
N
N
H
H
dRib
N
N
dRib
H
Figure 1. Hydrogen bond-mediated recognition of a CG base pair by a 4-(3-n-
butylureidophenyl)imidazole nucleoside receptor; the filled circle indicates the
mean C10 position of protonated third strand cytosine in a non-visible superim-
posed C+ꢀGC base triad.
* Corresponding author. Tel.: +49 (0) 3834 864426; fax: +49 (0) 3834 864427.
E-mail address: weisz@uni-greifswald.de (K. Weisz).
0040-4039/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2008.10.015

F. Wachowius et al. / Tetrahedron Letters 49 (2008) 7264–7267
7265
forces that govern base–base interactions in the two different
environments and to expand our model studies in aprotic solvents
to the formation of triple-helical complexes under physiological
conditions, we herein report on the incorporation of the 4-(3-n-
butylureidophenyl)imidazole-20-deoxynucleoside D4 into a TFO
and on its binding to various duplex targets through UV melting
experiments.
differ in their Watson–Crick base pair XY located opposite the non-
natural D4 analog in a triplex:
TFO15: 5’- C T T C D4 T T T T T C T T T T -3’
5’- G A A G X A A A A A G A A A A -3’
3’- C T T C Y T T T T T C T T T T -5’
A scheme for the preparation of the 50-tritylated and 30-phos-
phitylated 20-deoxyribonucleoside D4, which was used as syn-
thon for the oligonucleotide synthesis, is depicted in Scheme 1.
The sodium salt of 4-(3-nitrophenyl)imidazole, prepared from
1-nitro-3-(bromoacetyl)benzene and formamide according to
the Bredereck imidazole synthesis,8,12 was condensed with 2-
In addition, thermal stabilities of corresponding triplexes with
the internal D4ꢀXY triplet substituted for canonical C+ꢀGC and TꢀAT
as well as GꢀTA and TꢀCG triads were studied for comparison.
Although less stable than C+ꢀGC and TꢀAT, GꢀTA and TꢀCG base triads
have been found to be the best combinations for recognizing TA
and CG by natural bases.18,19 A typical UV melting experiment at
pH 7.0 on a 1:1 mixture of duplex and TFO with D4 opposite CG
(D4ꢀCG triplex) is shown in Figure 2. Each mixture displays a sig-
moidal high-temperature transition which corresponds to the du-
plex to single strand transition. Except for the C+ꢀGC and TꢀAT
triplexes that only exhibit one resolved transition at lower pH,
additional hyperchromicity effects indicate the dissociation of the
TFO from the duplex at lower temperatures.
Melting temperatures determined for the triplexes at pH 6.0,
6.5, and 7.0 are summarized in Table 1. Due to the presence of
at least three protonated C+ꢀGC base triplets within the triplexes,
thermal stabilities decrease with increasing pH as expected. Hav-
ing an additional C+ꢀGC base triplet, the canonical C+ꢀGC triplex
exhibits the largest drop (23 °C) in triplex melting temperature
Tm upon increasing the pH from 6.0 to 7.0. Interestingly, how-
ever, triplexes carrying the D4 analog and in particular triplexes
with D4ꢀGC and D4ꢀAT triads are less sensitive to pH, and triplex
melting temperatures only decrease by about 2 °C when going
from pH 6 to 7. Obviously, the non-natural base triplet counter-
acts the regular pH dependence of the triplex indicating that an
unprotonated imidazole base in the D4 analog is strongly
deoxy-3,5-di-O-p-toluoyl-
trile to yield the 4-(3-nitrophenyl)imidazole nucleoside in a b:
a-D
-ribofuranosyl chloride13 in acetoni-
a
anomeric molar ratio of 4:1.9,14 Following separation by flash
chromatography, the b-configuration was unambiguously as-
signed based on 2D NOE NMR experiments. The reduction of
the nitro group was accomplished under mild conditions with
Fe3(CO)12 and methanol,15 and the resulting free amine subse-
quently N-substituted through reaction with n-butylisocyanate
in THF.8 After deprotection under alkaline conditions, the nucle-
oside analog D4 was 50-tritylated using extended reaction times
of 18 h at 40 °C.16,17 Following conversion to its phosphorami-
dite, 6 was directly used as synthon for the synthesis of the oli-
gonucleotide
TFO15
using
the
standard
b-cyanoethyl
phosphoramidite method. The 15-mer oligonucleotide containing
the nucleoside analog was finally characterized by MALDI-TOF
mass spectrometry, giving a mass peak at m/z 4584.9 within
experimental error of the calculated molecular weight of
TFO15 (m/z 4588.1).
In order to examine the affinity and selectivity toward a duplex
target, equimolar amounts of four AT-rich double-helical oligonu-
cleotides were mixed with the D4 containing oligonucleotide
TFO15, which was designed to bind in a parallel motif to the com-
plementary duplex. Sequences of the 15-mer double-helical targets
ii
i
pTolO
pTolO
O
O
NH2
NO2
NO2
N
N
HN
N
O
N
pTolO
N
pTolO
1
3
2
O
pTolO
pTolO
HO
HO
O
O
N
H
N
H
N
N
H
iv
vi
iii
N
N
H
N
N
D4
4
O
O
DMTrO
O
DMTrO
HO
O
O
v
N
H
N
H
N
N
N
H
N
H
N
N
O
P
NC
N(iPr)2
6
5
Scheme 1. Reagents and conditions: (i) 2-deoxy-3,5-di-O-p-toluoyl-a-
D-ribofuranosyl chloride, NaH, CH3CN, 43%; (ii) Fe3(CO)12, toluene, 63%; (iii) C4H9NCO, THF, 79%; (iv) 1%
NaOH in MeOH, 80%; (v) DMTCl, DMAP, Et3N, pyridine, 56%; (vi) [(CH3)2CH]2NP(Cl)O(CH2)2CN, i-Pr2EtN, CH2Cl2.

7266
F. Wachowius et al. / Tetrahedron Letters 49 (2008) 7264–7267
Table 2
1.05
1.00
0.95
0.90
0.85
0.80
0.75
Thermodynamic parameters for the triplex–duplex transition from the fitting of UV
melting curves obtained from triplexes with different ZꢀXY base triplets at pH 7
D
Gb (kJ/mol)
a
ZꢀXY base triplet
Tm (°C)
D
H (kJ/mol)
D
S (J/mol K)
C+ꢀGC
26.8
31.3
22.3
19.4
23.9
24.2
ꢁ339
ꢁ322
ꢁ169
ꢁ191
ꢁ188
ꢁ158
ꢁ1018
ꢁ947
ꢁ460
ꢁ541
ꢁ521
ꢁ420
ꢁ35.6
ꢁ39.8
ꢁ31.9
ꢁ29.8
ꢁ32.7
ꢁ32.8
TꢀAT
D4ꢀCG
D4ꢀTA
D4ꢀGC
D4ꢀAT
a
From curve fitting.
At 25 °C.
b
assumes that extinction coefficients of the individual DNA species
are linear functions of temperature and enthalpy H as well as en-
tropy S of triplex formation are constant over the temperature
273
293
313
T (oC)
333
353
373
D
D
Figure 2. Temperature-dependent absorbance at 260 nm of the D4ꢀCG triplex; the
range of triplex melting.20 As a consequence of the two-state
approximation employed, a reliable fitting is restricted to UV melt-
ing curves with two clearly separated transitions for triplex and
duplex melting. Even with a rather conservative evaluation of the
fitted parameters to possibly avoid their overinterpretation consid-
ering the various simplifications of the analysis, thermodynamic
data from the curve fitting as summarized in Table 2 provide for
obvious trends in the thermodynamics of third strand binding with
solid line corresponds to the fitted curve. [Triplex] = 2.5
sodium cacodylate, 20 mM MgCl2, pH 7.
lM, 100 mM NaCl, 20 mM
Table 1
Summary of triplex–duplex melting temperatures Tm (°C) of triplexes with different
ZꢀXY base triplets
ZꢀXY base triplet
Tm (°C), pH 6.0a
Tm (°C), pH 6.5a
Tm (°C), pH 7.0b
the non-natural TFOs. With a
DH of 320–340 kJ/mol, that is, with
C+ꢀGC
49.0f
48.4f
29.6c
29.6d
29.0c
27.3c
26.2d
27.5d
35.7d
49.3f
25.7d
21.4d
23.8d
23.0d
25.1e
27.1e
25.7c
31.6e
20.1d
18.6e
21.8e
18.6e
24.9e
25.1e
21–23 kJ/mol per base, the van’t Hoff enthalpy change of triplex
formation for the two canonical C+ꢀGC and TꢀAT triplexes falls with-
in the range determined from previous UV melting experiments on
DNA triplexes.21 Although still driven by a negative enthalpy
change, third strand association with a D4 analog is accompanied
by less exothermic binding amounting to only ꢁ158 kJ/mol for
the D4ꢀAT triplex up to ꢁ191 kJ/mol for the D4ꢀTA triplex. How-
ever, as seen from Table 2, such a less favorable enthalpic contribu-
tion to triplex formation is partially balanced by a significantly less
unfavorable negative entropy change determined for all D4 tri-
plexes. Such thermodynamic profiles indicate that the D4 analog
only possesses a limited number of strong and specific interactions
with the duplex target. Also, binding through the hydrophilic ure-
ido functionality may be enthalpically compromised by the need of
its initial desolvation. On the other hand, less specific interactions
are expected to result in an enhanced conformational flexibility
associated with a less negative entropy change. Also, a major driv-
ing force for the D4 association when compared to natural bases
may arise from the release of water from the hydrophobic surface
of the analog upon binding and possibly from additional release of
structural water upon accommodating the N-butyl substituent
within the DNA major groove. The poor base pair discrimination
exhibited previously by a corresponding N-alkylureido-substituted
isoindolin-1-one receptor within a triple-helical complex, being in
contrast to its ability for effective complexation of a free CG base
pair in organic solvents, may likewise originate from similar
enthalpic and entropic contributions to binding in an aqueous
medium.16
TꢀAT
GꢀTA
TꢀCG
D4ꢀCG
D4ꢀTA
D4ꢀGC
D4ꢀAT
a
[Triplex] = 2.5
[Triplex] = 2.5
Uncertainty 0.5 °C.
Uncertainty 1 °C.
Uncertainty 2 °C.
Only one transition from triplex to single strands is observed.
l
l
M, 100 mM NaCl, 20 mM NaAc, 20 mM MgCl2.
M, 100 mM NaCl, 20 mM sodium cacodylate, 20 mM MgCl2.
b
c
d
e
f
favored upon triplex formation especially when opposite a GC or
AT base pair.
All D4 triplexes fail to achieve thermal stabilities of Tm >30 °C
observed for the canonical C+ꢀGC and TꢀAT triplexes under favor-
able conditions. Thus, the D4ꢀCG triplex being the most stable
non-natural triplex at low pH exhibits a Tm of 29 °C at pH 6 com-
parable to the melting of the GꢀTA and TꢀCG triplexes under equal
conditions. On the other hand and as a result of their small pH
dependence, the D4ꢀGC and D4ꢀAT triplexes exhibit the highest tri-
plex melting temperature of all D4 containing triple helices at pH 7
with a Tm of about 25 °C and compare favorably with the lower
melting GꢀTA and TꢀCG triplexes as well as with the canonical
C+ꢀGC triplex that shows about the same thermal stability at phys-
iological pH.
It has to be noted that there is no single most stable D4 triplex
within a pH range of 6 6 pH 6 7. Rather, relative stabilities of D4
triplexes vary according to their different pH dependence, but
selectivities toward the four target base pairs are rather poor at
any given pH value. Thus, only a maximum difference in Tm of
6.5 °C is observed between the most stable D4ꢀAT and the least sta-
ble D4ꢀTA triplex at pH 7. Consequently, the D4 analog must be re-
garded as a universal base binding opposite all four base pairs with
moderately high affinity at physiological pH.
In summary, our thermodynamic data on triplex formation with
the D4 containing TFO in aqueous solution indicate less specific
binding of the analog with a major driving force arising from
hydrophobic effects. As a consequence, moderately strong binding
affinities toward all possible base pairs with a poor base pair dis-
crimination are expected and observed based on triplex thermal
stabilities. This is in contrast to our previous model system of free
nucleosides in aprotic solvents showing the formation of a well-de-
fined hydrogen-bonded D4ꢀCG base triplet closely isomorphous to
the canonical base triads, and highlights the importance of addi-
tional contributions to binding within an aqueous environment.
For the future design of novel nucleoside surrogates,
Thermodynamic parameters for triplex formation at pH 7 were
determined by fitting the melting curves with a non-linear least-
squares program. The analysis is based on a two-state model and

F. Wachowius et al. / Tetrahedron Letters 49 (2008) 7264–7267
7267
3032; for reviews, see: (e) Gowers, D. M.; Fox, K. R. Nucleic Acids Res. 1999, 27,
1569–1577; (f) Doronina, S. O.; Behr, J.-P. Chem. Soc. Rev. 1997, 63–71; (g)
Purwanto, M. G. M.; Weisz, K. Curr. Org. Chem. 2003, 7, 427–446.
8. Wang, W.; Purwanto, M. G. M.; Weisz, K. Org. Biomol. Chem. 2004, 2, 1194–
1198.
thermodynamic data together with structural information will be
indispensible in achieving a more predictable base pair recognition
accompanied by efficient complexation.
9. Griffin, L. C.; Kiessling, L. L.; Beal, P. A.; Gillespie, P.; Dervan, P. B. J. Am. Chem.
Soc. 1992, 114, 7976–7982.
Acknowledgment
10. Zimmerman, S. C.; Schmitt, P. J. Am. Chem. Soc. 1995, 117, 10769–10770.
11. Koshlap, K. M.; Gillespie, P.; Dervan, P. B.; Feigon, J. J. Am. Chem. Soc. 1993, 115,
7908–7909.
We thank the Deutsche Forschungsgemeinschaft, Bonn, for
financial support.
12. Bredereck, H.; Theilig, G. Chem. Ber. 1953, 86, 88–96.
13. (a) Hoffer, M. Chem. Ber. 1960, 93, 2777–2781; (b) Rolland, V.; Kotera, M.;
Lhomme, J. Synth. Commun. 1997, 27, 3505–3511.
14. Kazimierczuk, Z.; Cottam, H. B.; Revankar, G. R.; Robins, R. K. J. Am. Chem. Soc.
1984, 106, 6379–6382.
References and notes
1. Duval-Valentin, G.; Thuong, N. T.; Hélène, C. Proc. Natl. Acad. Sci. U.S.A. 1992, 89,
504–508.
2. Kim, H.-G.; Miller, D. M. Biochemistry 1998, 37, 2666–2672.
3. Havre, P. A.; Gunther, E. J.; Gasparro, F. P.; Glazer, P. M. Proc. Natl. Acad. Sci.
U.S.A. 1993, 90, 7879–7883.
4. Jones, W. R.; Stone, M. P. Nucleic Acids Res. 1998, 26, 1070–1075.
5. Dervan, P. B. Nature 1992, 359, 87–88.
6. Strobel, S. A.; Doucette-Stamm, L. A.; Riba, L.; Housman, D. E.; Dervan, P. B.
Science 1991, 254, 1639–1642.
7. (a) Guianvarc’h, D.; Fourrey, J. L.; Maurisse, R.; Sun, H. S.; Benhida, R. Bioorg.
Med. Chem. 2003, 11, 2751–2759; (b) Buchini, S.; Leumann, C. J. Angew. Chem.,
Int. Ed. 2004, 43, 3925–3928; (c) Eldrup, A. B.; Nielson, P. E. J. Am. Chem. Soc.
1997, 119, 11116–11117; (d) Rusling, D. A.; Powers, V. E. C.; Ranasinghe, R. T.;
Wang, Y.; Osborne, S. D.; Brown, T.; Fox, K. R. Nucleic Acids Res. 2005, 33, 3025–
15. Landesberg, J. M.; Katz, L.; Olsen, C. J. Org. Chem. 1972, 37, 930–936.
16. Mertz, E.; Mattei, S.; Zimmerman, S. C. Bioorg. Med. Chem. 2004, 12, 1517–1526.
17. Spectroscopic data of 5: 1H NMR (600 MHz, 293 K, CDCl3): d (ppm) 0.94 (t, 3H,
CH3), 1.37 (m, 2H, CH2), 1.50 (m, 2H, CH2), 2.45 (m, 2H, H20, H200), 3.25 (m, 2H,
CH2), 3.37 (m, 2H, H50, H500), 3.79 (s, 6H, OCH3), 4.18 (m, 1H, H40), 4.59 (m, 1H,
H30), 5.98 (t, J = 6.5 Hz, 1H, H10), 6.85 (dd, J1 = 8.7, J2 = 8.9 Hz, 4H, ArH), 7.20–
7.40 (m, 10H, ArH), 7.46 (m, 2H, ArH), 7.59 (m, 2H, ArH), 7.68 (s, 1H, ArH).
18. Griffin, L. C.; Dervan, P. B. Science 1989, 245, 967–971.
19. Yoon, K.; Hobbs, C. A.; Koch, J.; Sardaro, M.; Kutny, R.; Weiss, A. L. Proc. Natl.
Acad. Sci. U.S.A. 1992, 89, 3840–3844.
20. Longfellow, C. E.; Kierzek, R.; Turner, D. H. Biochemistry 1990, 29, 278–285.
21. Kamiya, M.; Torigoe, H.; Shindo, H.; Sarai, A. J. Am. Chem. Soc. 1996, 118, 4532–
4538.