Synthesis, X-ray structure and binding properties of molecular clips based on dimethylpropanediurea

Article abstract of DOI:10.1039/a706012b

New concave host molecules show strong noncovalent binding of hydroxybenzene derivatives by an induced fit mechanism (K_a up to 3.4 × 10⁶ dm³ mol^-1).

Full text of DOI:10.1039/a706012b

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis, X-ray structure and binding properties of molecular clips based on
dimethylpropanediurea
Rob J. Jansen,a Alan E. Rowan,a Rene´ de Gelder,b Hans W. Scheeren*a and Roeland J. M. Nolte*a†
a Department of Organic Chemistry, NSR Center, University of Nijmegen, Toernooiveld, 6525 ED Nijmegen, The Netherlands
b Crystallography Laboratory, NSR Center, University of Nijmegen, Toernooiveld, 6525 ED Nijmegen, The Netherlands
New concave host molecules show strong noncovalent
binding of hydroxybenzene derivatives by an induced fit
mechanism (Ka up to 3.4 3 106 dm3 mol21).
of TFA with azeotropic removal of water. Clip molecule 2c
could be obtained in 20% yield by reacting 3a with 2,3-bis-
(bromomethyl)-1,4-dimethoxybenzene in DMSO in the pres-
ence of NaH. The latter compound was prepared from
3,6-dimethoxyphthalic anhydride4 by reduction with LiAlH4
followed by reaction with PBr3. Full experimental details will
be reported in a forthcoming paper.‡
Single crystals of 2a were obtained by vapour diffusion using
CHCl3 as the solvent and hexane as the precipitant and of 2c by
vapour diffusion using CH2Cl2 as the solvent and Et2O as the
precipitant. The crystal structure of 2a revealed that this clip
molecule has a U-shaped cavity similar to that of di-
phenylglycoluril derived clips.2a In contrast to previous clip
molecules of type 1 the crystal structure of 2c§ shows an
asymmetric geometry with respect to the side walls (Fig. 1),
which is attributed to its greater flexibility. The clip molecule
dimerizes to give a ‘head-to-head’ packing, with the wall of one
clip molecule being buried in the cavity of another clip. This
dimerization was not observed in CDCl3 solution.
The design and synthesis of host molecules for neutral guests
continues to be an area of great interest in supramolecular
chemistry.1 In recent years a series of receptors derived from the
concave molecule glycoluril (see 1) have been developed in our
O
O
R2
R2
R2
R2
N
N
N
N
N
N
N
N
R1
R1
R1
R2
R2
R2
R2
O
O
1a R1 = Ph, R2 = H
b R1 = Ph, R2 = OMe
c R1 = Me, R2 = H
2a R1 = H, R2 = H
b R1 = Me, R2 = H
c R1 = H, R2 = OMe
The binding properties of hosts 2 and, for comparison, hosts
1 with a number of hydroxybenzene derivatives (4a–f, 5–7),
were measured by NMR titration experiments in CDCl3 using
the outer wall protons of the host and the aromatic protons of the
guest as probes.¶ The results are summarized in Table 1. The
new clip molecules bind resorcinol derivatives significantly
more strongly than diphenylglycoluril molecular clips. The
binding constant is a factor of three higher for hosts 2a,b when
compared to host 1a.∑ In the case of 2c the increase in binding
constant strongly depended on the type of guest used (see Table
1). For guests 4a–c the binding was increased by a factor of
three, five and twelve respectively when compared to host 1b.
For guests 4d–f the binding constants were so large that they
could not be measured by standard NMR titrations.5 We
therefore determined the binding constants of 4d,e by competi-
tion experiments with 4c, and of 4f by a competition experiment
with 4e.6 A plot of the binding free energies of guests 4a–f as a
function of the Hammett parameter [sm (R)] of the guest’s
substituent (Fig. 2) reveals a linear correlation which has been
previously observed for clip 1b.2b From the gradients of the
R
O
HN
HN
NH
R1
NH
HO
4a R = Me
b R = H
c R = OMe
OH
O
3a R1 = H
b R1 = Me
d R = CO2C6H13
e R = Cl
f R = CN
laboratory.2 These receptors, which are U-shaped, bind dihy-
droxybenzenes by hydrogen bonding interactions between the
hydroxy groups of the guest and the urea carbonyl groups of the
host and by p–p stacking interactions between the guest and the
host side-walls. Although the supramolecular chemistry of
glycoluril-based clips has been widely explored, and their
sidewalls extensively varied,2 relatively little attention has been
given to variations in the diphenylglycoluril part of the clip
molecules. Here we describe the synthesis, X-ray structures and
binding properties of a new type of related molecular clips
derived from 2,4,6,8-tetraazabicyclo[3.3.1]nonane-3,7-dione
(propanediurea, see 2). Molecular modelling studies suggested
that the o-xylylene side walls of these new clips would be more
parallel than the walls of glycoluril derived clips, which would
result in better p–p stacking interactions with an aromatic guest
molecule sandwiched between the side walls of the clip and
hence lead to increased binding affinities for aromatic mole-
cules. Here we show that compound 2 indeed binds aromatic
guest molecules with very high association constants (Ka > 106
dm3 mol21).
Clip molecules 2a,b were prepared in ca. 20% yield by
reacting the respective propanediurea derivatives 3a,b with
a,aA-dibromo-o-xylene in DMSO in the presence of NaH.
Compound 3a was accessible following a literature procedure,3
and compound 3b was synthesized in 90% yield by refluxing
2,2-dimethyl-3-oxobutanal with urea in toluene in the presence
Fig. 1 X-Ray structure showing a dimer of clip 2c
Chem. Commun., 1998
121

View Article Online
CO C H
6 13
modelling and suggested in the asymmetry observed in the
X-ray crystal structures. The signals due to the benzylic protons
of clip 2c were found to shift considerably upon binding of a
guest (4d: up to +0.44 ppm for the upfield benzylic proton
signals and 20.29 ppm for the downfield benzylic proton
signals), in contrast to those of clips 1, for which virtually no
shifts were observed. These shifts indicate that the conforma-
tion of the clip’s side walls changes upon binding of a guest and
are consistent with a tightening of the cavity upon binding of an
aromatic guest.
2
R = Me
–15
H
OMe
Cl
CN
–20
–25
–30
–35
(a)
(b)
G
In conclusion, a new type of molecular clips is presented
which show enhanced binding of aromatic guest molecules.
Applications of these receptor molecules in the construction of
new supramolecular architectures are under investigation.
–40
–0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
s
m
Fig. 2 Binding free energies of guests 4a–f in clips (a) 1b and (b) 2c as a
function of the Hammett parameter [sm(R)] of the guest’s substituent
Footnotes and References
† E-mail: tijdink@sci.kun.nl
‡ All new compounds were fully characterized by 1H and 13C NMR and
mass spectroscopy and elemental analysis. Selected data for 2c: dH (300
MHz) 6.74 (s, 4 H), 5.43 (d, 2 H, J 15.1), 4.32 (s, 2 H), 3.90 (d, 4 H, J
15.2), 3.78 (s, 12 H), 1.35 (s, 6 H).
§ Crystal data and data collection parameters for 2c: C27H32N4O6, M =
508.57, monoclinic, a = 11.760(2), b = 15.300(3), c = 14.574(7) Å, b =
106.849(12)°, V = 2509.7(14) Å3, T = 293(2) K, space group P21/a, l =
1.54184 Å, Z = 4, Dc = 1.346 Mg m23, F(000) = 1080, colourless crystal
with dimensions 0.29 3 0.19 3 0.16 mm, m(Cu-Ka) = 0.791 mm21, Enraf-
Nonius CAD4 diffractometer, q–2q scans, 3.17 < q < 62.19°, +h, +k, ±l,
plots it is clear that binding of guests in clip 2c is more sensitive
to the substituent on the guest than binding in clip 1b. Previous
analysis of the binding properties of clips of type 1 showed that
the steeper the gradient the greater the hydrogen bonding
contribution is to the overall binding.2b One of the reasons for
the stronger binding of 2c is the fact that the carbonyl oxygens
atoms are situated slightly higher with respect to the cavity
walls than the carbonyl oxygens atoms in clip 1b, as is clear
from the X-ray structures. Previously it has been shown that the
optimal distance for p–p interaction of a guest in clips 1 is at a
position further out of the cavity than that for optimal hydrogen
bonding.2b This means that in clip 2c the complexation
geometry is more ideal for optimum p–p interactions than in
clip 1b. Furthermore, the carbonyl–carbonyl distance in clip 2c
(5.2 Å) is closer to the ideal value for resorcinol binding (3.9
Å)** than this distance in clip 1b (5.5 Å). Although the
difference is small, it is significant since 1b is shown to be better
suited for binding of guest molecules with relatively large OH–
OH distances such as 2,7-dihydroxynaphthalene 5 than 2c
(Table 1); this guest prefers a carbonyl–carbonyl distance of 6.3
Å.** The enhanced binding of catechol 6 in 2c can be explained
by the smaller carbonyl–carbonyl distance.∑ Other factors, apart
from the position of the carbonyl groups, also contribute to the
difference in binding properties of the clips. The binding of
4-nitrophenol 7, which has only one hydroxy group and hence
forms one strong optimum hydrogen bond, is stronger in clip 2c
than in clip 1b (Table 1), suggesting that additional factors, e.g.
the possibility of a guest to adopt a more parallel orientation
with respect to the cavity walls, play a role in the enhanced
binding.††
2
2
maximum drift 14.526%, 4186 reflections measured, 3967 unique (Rint
=
0.0098). The structure was solved using the program CRUNCH (ref. 8), and
refined anisotropically, by full-matrix least squares on F2 [program
SHELXL (ref. 9)]. The final wR(F2) was 0.1991, with conventional R(F)
0.0549.
¶ NMR titration experiments were performed as described in ref. 2(a). The
chloroform used was standard NMR grade and predried on molecular sieves
(4 Å) before use. The NMR spectra used for the determination of the binding
constants showed only a small water peak.
∑ For example: for 4d: Ka = 2000 m21 with clip 2a and 600 m21 with clip
1c; for 4b: Ka = 550 m21 with clip 2b and 165 m21 with clip 1c.
** Assuming that the hydrogen bonds are linear and that the O–H–O
distance is 2.7 Å (ref. 7).
†† Since clip molecules 1a and 1c have different groups at their convex
sides, we compared the binding affinities of 1a and 1c with guests 4b, 4d
and 4f. The binding constants were the same within the experimental error
(e.g. Ka = 165 and 175 m21, respectively, for 1a and 1c with guest 4b and
Ka = 3600 and 3500 m21, respectively, with guest 4f), which indicates that
the groups at the convex side of clips 1 do not contribute significantly to the
binding. Clip molecule 1c was prepared in 60% yield from dimethylglyco-
luril and a,aA-dibromo-o-xylene in DMF at room temperature in the
presence of NaH.
1 For an overview, see R. M. Izatt, J. S. Bradshaw, K. Pawlak, R. L.
Bruening and B. J. Tarbet, Chem. Rev., 1992, 92, 1261.
2 (a) R. P. Sijbesma, A. P. M. Kentgens, E. T. G. Lutz, J. H. van der Maas
and R. J. M. Nolte, J. Am. Chem. Soc., 1993, 115, 8999; (b) J. N. H. Reek,
A. H. Priem, H. Engelkamp, A. E. Rowan, J. A. A. W. Elemans and
R. J. M. Nolte, J. Am. Chem. Soc., 1997, 119, 9956; (c) J. N. H. Reek,
J. A. A. W. Elemans and R. J. M. Nolte, J. Org. Chem., 1997, 62,
2234.
The NMR data suggest that the binding of a guest in clips of
type 2 takes place via an induced fit mechanism, which is in
agreement with the increased flexibility predicted by molecular
Table 1 Association constants of complexes between various host and guest
molecules in CDCl3, T = 25 °C
3 B. N. Khasapov, T. S. Novikova, O. V. Lebedev, L. I. Khmel’nitskii and
S. S. Novikov, Zh. Org. Khim., 1973, 9, 23.
Host
4 C. Cardani and F. Piozzi, Lincei-Rend. Sc. Fis. Mat. Nat., 1952, 12,
719.
5 B. J. Whitlock and H. W. Whitlock, J. Am. Chem. Soc., 1990, 112,
3910.
6 J. S. Alper, R. I. Gelb, D. A. Laufer and L. M. Schwartz, Anal. Chim.
Acta, 1989, 220, 171.
7 I. Olovsson and P.-G. Jo¨nsson, in The Hydrogen Bond, ed. P. Schuster, G.
Zundel and C. Sandorfy, North Holland Publishing Company, Am-
sterdam, New York, Oxford, 1976, vol. II, pp. 393–456.
8 R. de Gelder, R. A. G. de Graaff and H. Schenk, Acta Crystallogr., Sect.
A, 1993, 49, 287.
Guest
2c
5500a
1b
1900b
4a
4b
4c
4d
4e
4f
5
14000a
53000a
2.7·105d
4.2·105d
3.4·106e
2300a
2600c
4400b
16500b
16000b
1·105b
7100b
60c
6
7
130d
4400a
1200c
9 G. M. Sheldrick, SHELXL-97, Program for the refinement of crystal
structures, University of Go¨ttingen, Germany, 1997.
a Estimated error, 20%. b Values taken from ref. 2(a). c Values taken from
ref. 2(b). d Estimated error, 30%. e Estimated error, 40%.
Received in Cambridge, UK, 15th August 1997; 7/06012B
122
Chem. Commun., 1998