Higher affinity quadruply hydrogen-bonded complexation with 7-deazaguanine urea

Article abstract of DOI:10.1021/ol0601803

UG forms a highly stable quadruply hydrogen-bonded heterocomplex with DAN, but the fidelity of the complex is lowered somewhat by the Hoogsteen-side oligomerization of UG (K_assoc ～ 230 M^-1, CDCl ₃). DeUG was prepared as a

Full text of DOI:10.1021/ol0601803

ORGANIC
LETTERS
2006
Vol. 8, No. 8
1589-1592
Higher Affinity Quadruply
Hydrogen-Bonded Complexation with
7-Deazaguanine Urea
Hugo C. Ong and Steven C. Zimmerman*
Department of Chemistry, 600 South Mathews AVenue,
UniVersity of Illinois, Urbana, Illinios 61801
sczimmer@uiuc.edu
Received January 22, 2006
ABSTRACT
UG forms a highly stable quadruply hydrogen-bonded heterocomplex with DAN, but the fidelity of the complex is lowered somewhat by the
-
Hoogsteen-side oligomerization of UG (Kassoc
∼
230 M 1, CDCl3). DeUG was prepared as a more robust analogue of UG lacking the Hoogsteen
nitrogen atom. Remarkably, the deaza analogue, DeUG, forms a much more stable complex with DAN (>10-fold higher Kassoc for DeUG
‚DAN
-
vs UG
‚DAN) but also dimerizes more strongly (Kdim
)
880
±
40 M 1, CDCl3) by adopting a conformation preorganized for both binding and
dimerization.
Recent efforts to expand the “supramolecular toolkit” have
focused on hydrogen-bonded complexes of tunable strength
in organic solvents.1 The quadruply hydrogen-bonded com-
plexes, in particular, are an attractive class as a result of their
high binding strengths and synthetic accessibility. Our DeAP
unit (1)2 and the Meijer-Sijbesma UPy unit (2)3 are
examples that dimerize strongly via AADD motifs (Figure
1). Both units also adopt an ADDA form (1′ and 2′) that
very tightly binds 2,7-diamido-1,8-naphthyridine (DAN, 3).2,4
Because 1 and 2 can present strongly self-associating AADD
hydrogen bonding arrays, the overall fidelity with which they
complex with 3 is lowered.5 We showed that in comparison
to 1 and 2, the AADD form of the UG unit 4 is much higher
in energy than the ADDA form so it complexes 3 with a
much higher fidelity. Thus, 4 tightly complexed 3 (Kassoc
∼
5 × 107 M-1, CHCl3) via form 4′ but did not dimerize
strongly via form 4′′.6 The UG‚DAN complex is well
optimized; however, two aspects of the UG unit make it less
than an ideal subunit. First, the UG weakly self-associates
(Kassoc ∼ 230 M-1) via oligomerization of form 4,6 the urea
group hydrogen bonding at the Hoogsteen side of the UG
(1) For reviews, see: (a) Krische, M. J.; Lehn, J.-M. Stuct. Bond. 2000,
96, 3-29. (b) Mele´ndez, R. E.; Carr, A. J.; Linton, B. R.; Hamilton, A. D.
Struct. Bond. 2000, 96, 31-61. (c) Zimmerman, S. C.; Corbin, P. S. Struct.
Bond. 2000, 96, 63-94. (d) Kato, T. Struct. Bond. 2000, 96, 95-146. (e)
Prins, L. J.; Reinhoudt, D. N.; Timmerman, P. Angew. Chem., Int. Ed. 2001,
40, 2382-2426. (f) Sherrington, D. C.; Taskinen, K. A. Chem. Soc. ReV.
2001, 30, 83-93. (g) Sijbesma, R. P.; Meijer, E. W. Chem. Commun. 2003,
5-16. (h) Sivakova, S.; Rowan, S. J. Chem. Soc. ReV. 2005, 34, 9-21. (i)
Sessler, J. L.; Jayawickramarajah, J. Chem. Commun. 2005, 1939-1949.
(2) (a) Corbin, P. S.; Zimmerman, S. C. J. Am. Chem. Soc. 1998, 120,
9710-9711. (b) Corbin, P. S.; Lawless, L. J.; Li, Z.-T.; Ma, Y.; Witmer,
M. J.; Zimmerman, S. C. Proc. Nat. Acad. Sci. U.S.A. 2002, 99, 5099-
5104.
(4) (a) Wang, X.-Z.; Li, X.-Q.; Shao, X.-B.; Zhao, X.; Deng, P.; Jiang,
X.-K.; Li, Z.-T.; Chen, Y.-Q. Chem. Eur. J. 2003, 9, 2904-2913. (b)
Ligthart, G. B. W. L.; Ohkawa, H.; Sijbesma, R. P.; Meijer, E. W. J. Am.
Chem. Soc. 2005, 127, 810-811.
(5) Todd, E. M.; Quinn, J. R.; Park, T.; Zimmerman, S. C. Isr. J. Chem.
2005, 45, 381-389.
(6) (a) Park, T.; Zimmerman, S. C.; Nakashima, S. J. Am. Chem. Soc.
2005, 127, 6520-6521. (b) Park, T.; Todd, E. M.; Nakashima, S.;
Zimmerman, S. C. J. Am. Chem. Soc. 2005, 127, 18133-18142.
(3) Sijbesma, R. P.; Beijer, F. H.; Brunsveld, L.; Folmer, B. J. B.;
Hirschberg, J. H. K. K.; Lange, R. F. M.; Lowe, J. K. L.; Meijer, E. W.
Science 1997, 278, 1601-1604.
10.1021/ol0601803 CCC: $33.50
© 2006 American Chemical Society
Published on Web 03/25/2006

chromatography was required until the installment of the
ureido group in the final step, allowing for a preparation that
could be readily performed on multigram scale.
In UG, the anti form (Figure 1) was dominant in all
solvents and concentrations examined. However, the strong
complexation of DAN indicated that the syn form was
energetically accessible. Unexpectedly, 1H NMR studies of
the conformation of DeUG 5a and 5b showed its conforma-
tion to be solvent dependent. In DMSO-d6 at 5 mM 5b, the
anti form is dominant; at the same concentration in CDCl3,
the NH(3) proton moves downfield by 1.71 ppm to 8.94 ppm,
in contrast to NH(1) and NH(2) which move upfield (see
the Supporting Information). The large downfield shift can
be rationalized by the intramolecular hydrogen bond that is
formed in the syn conformation. The NH(3) peak also
broadens, which we have found to be characteristic of
intramolecularly hydrogen-bonded pyridyl and naphthyridinyl
ureas.2 Further evidence for this tautomeric form10 and
Figure 1. Various tautomers/conformers of DeAP 1, UPy 2, UG
4, and DeUG 5 and complexes of 1′, 2′, 4′, and 5′ with DAN (3).
In 1 and 2, R ) Bu. In 3, R ) C6H13, and in 4 R ) Bu, R′ )
2′,3′,5′-O-acetylribose. See Scheme 1 for R in 5.
Scheme 1
unit. Additionally, the ribose unit is large, and its labile
linkage to the base limits its utility in some applications.
Herein, we describe the preparation and study of the ureido-
7-deazaguanine (DeUG) 5, an analogue of UG which lacks
both the ribose unit and the Hoogsteen-side nitrogen atom
N(7), thought to be essential for the oligomerization.
Various 7-deazaguanines were reported as replacements
for guanine in DNA and as enzyme inhibitors.7 The reported
synthesis was adapted to the synthesis of DeUG 38 which
began with the amination of 2-amino-4,6-dichloropyrimidine
(6). Although 6 can be obtained by commercial means, it is
more cheaply obtained by synthesis from 2-amino-4,6-
dihydroxypyrimidine in POCl3 and N,N-dimethylaniline,
which occurs in near quantitative yield (not shown).9 The
amination with cyclohexylmethylamine occurred smoothly
in DMF with Hu¨nig’s base to give 7 in good yields (80-
95%). The cyclohexylmethyl group was chosen to provide
organic solubility. Basic hydrolysis in ethylene glycol under
reflux gave 8 in 57% yield. Treatment of the pyrimidinone
with R-chloroacetaldehyde in a DMF/H2O mixture with
NaOAc gave the 7-deazaguanine derivative 9 in 55% yield.
We previously found the amino group of guanine and pterin
derivatives to be quite unreactive;2,5 however, the desired
ureido moiety in 5a could be prepared by treatment of 9
with 2.2 equiv of n-BuNCO in refluxing pyridine in 65%
yield. For added solubility, 5b was also prepared using the
corresponding isocyanate (6 equiv) in 55% yield. No
conformation in chloroform came from an NOE study,
summarized in Figure 2. Upon irradiation of NH(3) of 5b, a
key NOE was observed to the methylene of the cyclohexy-
lmethyl group (Ha). No such signal was observed when either
NH(1) or NH(2) was irradiated. These results suggest that
DeUG does not oligomerize in the same manner as UG. In
fact, DeUG can only self-associate through dimerization via
the syn form, which we find to be consistent with our
observations in CDCl3. The lactim (enol) tautomer of 5b
could not be ruled out by the NMR data. However, a similar
(7) (a) Fletcher, T. M.; Cathers, B. E.; Ravikumar K. S.; Mamiya, B.
M.; Kerwin, S. M. Bioorg. Chem. 2001, 29, 36-55. (b) Gangjee, A.;
Dubash, N. P.; Kisliuk, R. L. J. Heterocycl. Chem. 2001, 38, 349-354. (c)
Gibson, C. L.; Rosa, S. L.; Ohta, K.; Boyle, P. H.; Leurquin, F.; Lemac¸on,
A.; Suckling, C. J. Tetrahedron 2004, 60, 943-959.
(8) Park, T.; Mayer, M. F.; Nakashima, S.; Zimmerman, S. C. Synlett.
2005, 9, 1435-1436.
(9) Appleton, W. C.; Parziale, P. A. Eur. Pat. WO9507265, 1995.
(10) DeUG 5b showed small amounts of another tautomer (<3% by
NMR integration) which appeared to dimerize strongly in CDCl3. The
identity of this tautomer was not determined (see the Supporting Informa-
tion), and it was not observed with 5a.
1590
Org. Lett., Vol. 8, No. 8, 2006

The upfield shift of the key NH signals upon dilution
indicates that DeUG does not form strong UPy- or DeAP-
type dimers via (AADD)2 or (ADAD)2 motifs.2,3,11 Further-
more, DeUG does not oligomerize by the Hoogsteen-side
hydrogen bonding exhibited by UG. However, the removal
of the 7-aza group unexpectedly causes an equilibrium shift
from the anti-form 5 to the syn-form 5′. This conformer is
able to dimerize somewhat more strongly by virtue of its
pyridone-type moiety. The origin of this conformation shift
is not readily apparent. However, computational studies
indicate the two forms to be close in energy so that small
structural, electronic, or environmental changes can readily
effect the equilibrium.
Figure 2. Key NOE contacts observed for 5b (DeUG).
A 1:1 mixture of 5 (a and b, in separate experiments) and
3 in CDCl3 showed large downfield shifts of the NH protons
relative to the free components: NH(1), NH(2), and NH(3)
moved by 2.0, 1.8, and 0.6 ppm, respectively. The relatively
small shift exhibited by NH(3) is consistent with the syn
model in which it is intramolecularly hydrogen-bonded, and
not directly involved with heterocomplexation. No new peaks
were observed nor did any shift occur in the DeUG unit upon
dilution from 10 mM to 10 µM (see the Supporting
Information). Thus, assuming 95% complexation, a lower
limit of Kassoc > 107 M-1 can be estimated.
tautomer formed by UPy was shown by Meijer, Sijbesma,
and co-workers11 to very tightly dimerize (Kdim > 4.5 × 105
M-1) by its self-complementary DADA hydrogen bonding
array, and this is inconsistent with the self-association study
described below.
The self-association of DeUG was examined by NMR. A
dilution study with 5b was performed in CDCl3 from 50 µM
to 15 mM at 20 °C; the signal for NH(1) was followed
throughout the study, and the data was fit to a nonlinear
binding equation to provide a Kdim of 880 ( 40 M-1 (Figure
3). The signals for NH(2) and -(3) could not be followed
The shift in equilibrium from the anti- to syn-conformation
means 5 should be better preorganized for binding DAN (3).
Because of the difficulty in accurately determining such high
binding constants in CDCl3, a direct competition experiment
was performed. Thus, DAN was titrated into an equimolar
solution of DeUG and UG in CDCl3 (see Figure 4 and the
Supporting Information). As the concentration of DAN was
increased, the DeUG‚DAN complex appeared to form
preferentially over the UG‚DAN complex. This observation
Figure 3. Representative plot of ∆δ vs [DeUG, 5b] for the dilution
study in CDCl3 at 20 °C. Data for this proton was fit to a nonlinear
binding equation to calculate Kdim
.
due to poorly resolved signals at 50 µM. It should be noted
that, by virtue of it possessing a symmetric (palindromic)
ADDA hydrogen-bonding pattern, in principle, there are four
ways in which DeUG may self-associate. However, given
that pyridone dimers tend to be more stable than urea dimers
(Kdim ) 50-100 M-1 vs Kdim ) 5-15 M-1, respectively),12
it is likely that the lactam dimer and the lactam/urea dimers
contribute more than the urea dimer.
(11) Beijer, F. H.; Sijbesma, R. P.; Koojiman, H.; Spek, A. L.; Meijer,
E. W. J. Am. Chem. Soc. 1998, 120, 6761-6769.
Figure 4. Stacked NMR plots of 1:1 DeUG‚DAN, 5b‚3 (top), 1:1:1
UG‚DeUG‚DAN (middle), and UG (bottom). Each module is at
10 mM concentration in CDCl3 at 20 °C. The DeUG peaks are
shown in blue, DAN in green, and UG in pink. Top of CHCl3 peak
cut off.
(12) (a) Hammes, G. G.; Park, A. C. J. Am. Chem. Soc. 1969, 91, 956-
961. (b) Gallant, M.; Viet, M. T. P.; Wuest, J. D. J. Am. Chem. Soc. 1991,
113, 721-723. (c) Zimmerman, S. C.; Duerr, B. F. J. Org. Chem. 1992,
57, 2215-2217. (d) Corbin, P. S.; Zimmerman, S. C.; Thiessen, P. A.;
Hawryluk, N. A.; Murray, T. J. J. Am. Chem. Soc. 2001, 123, 10475-
10488.
Org. Lett., Vol. 8, No. 8, 2006
1591

is particularly evident in Figure 4 where a 1:1:1 mixture of
UG‚DeUG‚DAN (i.e., 4‚5b‚3) exhibits a 1H NMR spectrum
in CDCl3 that, with the exception of broadening and some
slight shifts, appears to be a composite of the spectrum of
UG (4) and the DeUG‚DAN (5b‚3) complex. It was not until
an excess of 1.0 equiv of DAN was added that the UG‚
DAN complex could be observed.
The binding strengths were also compared directly by
employing a more competitive solvent, namely 5% DMSO-
d6/CDCl3. Titration experiments showed that UG (4) binds
DAN (3) with Kassoc ) 3200 M-1 in this solvent.13 This
binding strength is comparable to DeAP‚DAN (1‚3), which
we have previously determined to be Kassoc ) 3030 M-1.2
The DeUG‚DAN (5b‚3) complex, however, gave a binding
constant of Kassoc ≈ 104-105 M-1. Only an approximate value
could be obtained for this complex, due to extreme broaden-
ing of the -NH peaks at <1.0 equiv of DAN. However, the
titration curve clearly indicates a stronger complex (see the
Supporting Information); even at 1.0 equiv of DAN at
micromolar concentrations, the NH(1) proton quickly reaches
65% of its maximum chemical shift change. Taken together
with the competition experiments, we conclude that the
DeUG‚DAN complex is indeed stronger than UG‚DAN,
perhaps >10-fold higher.
side oligomerization. However, unexpectedly, the urea group
of DeUG (5) adopts the syn-form whereas the urea in UG
(4) was exclusively in the anti-form. The syn conformer is
preorganized for binding, but also stronger dimerization via
the pyridone-type moiety. Thus, the DeUG unit dimerization
constant is about 3-4 times than that for UG but it
complexes DAN with a Kassoc that is g10-fold tighter. The
net effect is that DeUG is an overall better choice as a
complement for DAN, except in cases where it will be used
in excess. In such cases, the stronger dimerization of DeUG
will lead to a lowered fidelity in the overall assembly. Efforts
are currently directed toward developing DeUG analogues
for applications in nanotechnology and supramolecular
polymer chemistry.
Acknowledgment. We thank the ICI National Starch and
Chemical Company and the U.S. Department of Energy,
Division of Materials Science, under Award No. DEFG02-
91ER45439 through the Frederick Seitz Materials Research
Laboratory at the University of Illinois at Urbana-Cham-
paign, for funding this work. Taiho Park and Darrell
Kuykendall are gratefully acknowledged for their synthetic
contributions. Eric Todd is also acknowledged for his work
on examining the fidelity of DeUG.
In conclusion, we have developed a UG analogue, DeUG,
which lacks the large, labile ribose unit yet forms highly
stable, quadruply hydrogen-bonded complexes with DAN.
The deazaguanine unit lacking N(7) did prevent Hoogsteen-
Supporting Information Available: Detailed descrip-
tions of all experimental procedures along with characteriza-
1
tion data and H NMR spectra. This material is available
free of charge via the Internet at http://pubs.acs.org.
(13) Park, T. Unpublished results.
OL0601803
1592
Org. Lett., Vol. 8, No. 8, 2006