The N-carbamoyl squaramide dimer: A compact, strongly associated H- bonding motif

Article abstract of DOI:10.1039/a907179b

N-Carbamoyl squaramides 3, readily available from dialkyl squarate 5, form hydrogen-bonded dimers in solution and in non-polar organic solvents; in CDCl₃ they show no significant dissociation down to concentrations of ~0.5 mM.

Full text of DOI:10.1039/a907179b

The N-carbamoyl squaramide dimer: a compact, strongly associated
H-bonding motif
Anthony P. Davis,*a Sylvia M. Draper,a Gavin Dunnea and Peter Ashtonb
a
Department of Chemistry, Trinity College, Dublin 2, Ireland. E-mail: adavis@tcd.ie
School of Chemistry, University of Birmingham, Edgbaston, Birmingham, UK B15 2TT
b
Received (in Cambridge, UK) 6th September 1999, Accepted 1st October 1999
N-Carbamoyl squaramides 3, readily available from dialkyl
squarate 5, form hydrogen-bonded dimers in solution and in
non-polar organic solvents; in CDCl3 they show no sig-
nificant dissociation down to concentrations of ~ 0.5 mM.
The continuing development of supramolecular self-assembly
requires accessible subunits which can be relied upon to
associate strongly and predictably through non-covalent inter-
actions. Particular attention has recently been paid to systems
capable of forming four intermolecular hydrogen bonds,1–4
such as the self-complementary ‘DDAA’ units 1 and 2 studied
Scheme 1 Reagents and conditions: i, NH3, MeOH–CH2Cl2, room temp.; ii,
R1NCO, Et3N, MeCN, room temp.; iii, R2NH2, CH2Cl2, room temp.
However, dilution studies in this solvent showed no substantial
changes in the spectra of 3a or 3b (Dd < 0.1 ppm) over a wide
range of concentrations ( ~ 25–0.5 mM), suggesting either that
association was not taking place or that it was very strong
indeed.
Suspecting the latter, we obtained crystals of 3a from
CH2Cl2–light petroleum and subjected them to X-ray crys-
tallography.† As shown in Fig. 1, 3a exists as a centrosym-
metric dimer in the crystal, apparently maintained by four
NH…ONC hydrogen bonds. The carbamoylsquaramide units
are essentially planar, and exhibit the expected intramolecular
hydrogen bond.
The hypothesis that 3 dimerises strongly in CDCl3 was tested
by NMR studies in the presence of ‘competitive’ co-solvents.
Addition of CD3CN (up to 40% v/v) to CDCl3 solutions of 3a
resulted in upfield movements of all three NH signals. For H(1)
and H(3) the motions were moderate (Dd ~ 20.7 ppm), while
by the groups of Meijer1 and Zimmerman,2 respectively. These
units self-associate essentially completely in solvents such as
CDCl3, and are only disrupted by significant quantities of polar
solvents such as DMSO.
Herein we report a new homodimer which is related to 1 and
2 but with significant geometric differences. It is nearly as
strongly bound, while being exceptionally compact and easy to
construct. The new system 3·3 is based on the squaramide unit,
previously shown by Ballester, Costa and co-workers to provide
a distinctive array of H-bonding sites for molecular recogni-
tion.5,6 Unlike the earlier squaramides, monomer 3 possesses a
ureidyl substituent capable in principle of intramolecular
hydrogen bonding to the neighbouring NH. The result should be
a well-defined, roughly planar structure with an intriguing
resemblance to guanine derivatives 4.
Planning to investigate the interaction of 3 with nucleobases,
we synthesized examples 3a and 3b from diisopropyl squarate
5 as shown in Scheme 1. Preliminary 1H NMR binding
experiments in CDCl3 gave unexpected results, leading us to
suspect that the squaramides self-associated to some degree.
Fig. 1 Crystal structure of 3a showing the relationship between adjacent H-
bonded molecules. Selected distances (Å) and angles (°) are H(1)…O(5)
2.120(2), H(2)…O(3) 1.936(1), H(3)…O(3) 2.204(6), H(1)–N(1) 0.869(6),
H(2)–N(2) 0.900(3), H(3)–N(3) 0.832(2); N(1)–H(1)…O(5) 136.65(6),
N(2)–H(2)…O(3)
160.96(8),
N(3)–H(3)…O(3)
149.03(9),
H(2)…O(3)…H(3) 60.81(9).
Chem. Commun., 1999, 2265–2266
This journal is © The Royal Society of Chemistry 1999
2265

Table 1 1H NMR dilution studies for compounds 3 in CDCl3–CD3CN (95+5)a
ddimer 2 dmonomer
(ppm)
H(3)
Compound
R1, R2
Kd/M21b
H(2)
3a
3b
3c
Bu, cyclohexyl
Bu, Bu
2,6-difluorophenyl, Bu 9800 ± 900
5800 ± 1100
6100 ± 1300
2.32
2.42
2.99
b
0.59
c
0.62
a
Experiments performed at 303 K. Concentrations of 3 were varied between ~ 25 and ~ 0.2 mM. Dimerisation constants Kd and error limits (standard
deviations) were estimated from the movements of the signal due to H(2) using HOSTEST version 5.0 (ref. 7). Not calculated, due to signal
broadening.
c
for H(2) a substantial Dd of 22.2 ppm was recorded. In the
absence of self-association, downfield motions would have
been expected due to the formation of NH…N·C hydrogen
bonds.‡ The observed upfield movements presumably result
from the break-up of the aggregate, with the loss or weakening
of the stronger NH…ONC hydrogen bonding.
values of ca. 103 M21 are reported for both 1 and 2 in CDCl3–
(CD3)2SO (95+5).1,2 However, 3·3 is more tightly bound than
other systems joined by four hydrogen bonds,3,4 being main-
tained down to very low concentrations in CDCl3. Given the
accessibility, variability and compact size of 3, this moiety
should be capable of playing a distinctive rôle in the design of
self-assembling systems.
The solvent system CDCl3–CD3CN (95:5) proved suitable
for quantitative 1H NMR dilution studies. Results are sum-
marised in Table 1. In addition to 3a and 3b, the difluorophenyl-
substituted example 3c was synthesized (cf. Scheme 1) and
studied under these conditions.§ In each case the signal due to
H(2), which appeared between d 10.3 and 10.9 at high
concentrations, moved upfield upon dilution. The data were
consistent with simple 1+1 dimerisations, yielding dimerisation
constants (Kd) in the range 5 3 103 to 104 M21, and limiting
chemical shift differences (ddimer 2 dmonomer) in the range 2.3–3
ppm. The signals due to H(3), starting at d ~ 6.5 for 3a/b and d
~ 8.1 for 3c, also moved upfield during the experiments. Their
motions were smaller and difficult to follow accurately due to
broadening, but generally supported the analyses of H(2). A
study of 3a in CDCl3–(CD3)2SO (99+1) also suggested
dimerisation, but with the lower Kd of ~ 180 M21. A re-
examination of 3a in CDCl3 gave data consistent with Kd ~ 106
M21, on the assumption that (ddimer–dmonomer) is likely to be
between 3 and 4 ppm in this solvent. The rôle of H(3) in
maintaining the dimer was confirmed by a study of N-
acetylsquaramide 6, for which a Kd of just 120 M21 was
measured in CDCl3–CD3CN (95+5).
Financial support for this work was provided by Enterprise
Ireland, the Irish Health Research Board and the Wellcome
Trust. We are grateful to Professor C. S. Wilcox for access to the
HOSTEST binding analysis programme.
Notes and references
† Crystal data for 3a: C15H23N3O, M = 293.36, triclinic, space group P1,
¯
a = 5.9800(4), b = 9.5090(5), c = 14.1543(8) Å, a = 89.694(6), b =
81.471(5), g = 87.322(5)°, U = 795.10(8) Å3, Z = 2, Dc = 1.225 g cm23
,
T = 293(2) K, m(Mo-Ka) = 0.086 mm21, wR2 = 0.1250 (3009 reflections
collected, 2726 unique), R = 0.0510 [I > 2s(I)], ENRAF NONIUS CAD4
diffractometer with graphite monochromator, w-scans, structure solved by
automatic direct methods using SHELXS-86 and refined using full-matrix
least-squares on F2 using SHELXL-93 (ref. 8). All the non-hydrogen atoms
were refined anisotropically and the hydrogen atoms were located from
subsequent difference Fourier maps. CCDC 182/1435. See http:/
/www.rsc.org/suppdata/cc/1999/2265/ for crystallographic data in .cif
format.
‡ The NH signal for N,N’-dibutylurea in CDCl3 moves downfield by ~ 0.13
ppm on addition of 5% CD3CN.
§ A number of other aryl-substituted N-carbamoylsquaramides were
prepared but found unsuitable for study, mainly due to solubility
problems.
1 R. P. Sijbesma, F. H. Beijer, L. Brunsveld, B. J. B. Folmer, J. H. K. K.
Hirschberg, R. F. M. Lange, J. K. L. Lowe and E. W. Meijer, Science,
1997, 278, 1601; F. H. Beijer, R. P. Sijbesma, H. Kooijman, A. L. Spek
and E. W. Meijer, J. Am. Chem. Soc., 1998, 120, 6761.
2 P. S. Corbin and S. C. Zimmerman, J. Am. Chem. Soc., 1998, 120,
9710.
3 U. Lüning and C. Kühl, Tetrahedron Lett., 1998, 39, 5735.
4 B. Gong, Y. F. Yan, H. Q. Zeng, E. Skrzypczak-Jankunn, Y. W. Kim, J.
Zhu and H. Ickes, J. Am. Chem. Soc., 1999, 121, 5607.
5 R. Prohens, S. Tomàs, J. Morey, P. M. Deyà, P. Ballester and A. Costa,
Tetrahedron Lett., 1998, 39, 1063.
6 S. Tomàs, R. Prohens, M. Vega, M. C. Rotger, P. M. Deyà, P. Ballester
and A. Costa, J. Org. Chem., 1996, 61, 9394.
7 C. S. Wilcox, in Frontiers in Supramolecular Chemistry and Photo-
chemistry, ed. H.-J. Schneider and H. Durr, VCH, Weinheim, 1990, p.
123.
The tendency of 3 to dimerise was also revealed in the FAB+
mass spectrum of 3b. In addition to the monomer at m/z 268
(MH+), there was a significant ( ~ 30%) signal for the dimer at
m/z 535 (M2H+). Signals for higher aggregates MnH+ (n = 3-5)
appeared at < 1.5%. Interestingly the addition of AcONa
promoted not only the MnNa+ signals as expected, but also a
series of general formula [Mn 2 mH + (m + 1)Na]+ (1 < m < n).
These ions presumably arise from substitution of the relatively
acidic H(2) by Na, followed by clustering in various combina-
tions with 3b itself.
8 G. M. Sheldrick, Acta Crystallogr., Sect. A, 1990, 46, 467; SHELXL-93,
University of Gottingen, Germany.
When compared with 1 and 2, the present system 3·3 would
seem to be somewhat less strongly associated. For example, Kd
Communication 9/07179B
2266
Chem. Commun., 1999, 2265–2266