Molecular clips based on propanediurea. Exceptionally high binding affinities for resorcinol guests

Article abstract of DOI:10.1021/jo001317k

A series of new receptor molecules derived from 2,4,6,8-tetraazabicyclo[3.3.1]nonane-3,7-dione (propanediurea) is described. These molecules possess a cavity which is defined by two nearly parallel aromatic side walls positioned on top of a bis-urea frame

Full text of DOI:10.1021/jo001317k

J . Org. Chem. 2001, 66, 2643-2653
2643
Molecu la r Clip s Ba sed on P r op a n ed iu r ea . Excep tion a lly High
Bin d in g Affin ities for Resor cin ol Gu ests
Rob J . J ansen,† Rene´ de Gelder,‡ Alan E. Rowan,† Hans W. Scheeren,*,† and
Roeland J . M. Nolte*,†
Department of Organic Chemistry, NSR Center, and Department of Inorganic Chemistry,
University of Nijmegen, Toernooiveld, 6525 ED Nijmegen, The Netherlands
nolte@sci.kun.nl
Received September 2, 2000
A series of new receptor molecules derived from 2,4,6,8-tetraazabicyclo[3.3.1]nonane-3,7-dione
(propanediurea) is described. These molecules possess a cavity which is defined by two nearly
parallel aromatic side walls positioned on top of a bis-urea framework. The resulting “U-shaped”
clip molecules are ideal hosts for the complexation of flat aromatic guest molecules. The affinity of
these new propanediurea based molecular clips for dihydroxybenzene derivatives is exceptionally
high, with association constants up to Ka ) 2 400 000 L mol-1. Comparison of the binding mechanism
of a variety of clip and half clip hosts, in conjunction with NMR, IR, and X-ray studies, has enabled
the reason for this high binding to be elucidated. It is shown that subtle sub-angstrom changes in
the geometry of the clip molecules have a great impact on their binding properties.
Ch a r t 1
In tr od u ction
The design and synthesis of host molecules for the
binding of neutral guest molecules continues to be an
area of interest in supramolecular chemistry.1 Research
in our laboratory has for some time been focused on the
development of receptors derived from the concave
molecule diphenylglycoluril (Chart 1, 1, R1 ) R2 ) Ph).
These “U-shaped” clips bind dihydroxybenzenes by means
of hydrogen bonding between the hydroxyl groups of the
guest and the urea carbonyl groups of the host and by
π-π stacking interactions between the guest and the host
side walls.2 Other groups have synthesized and studied
receptors having a similar U-shape, which are composed
of linked aromatic residues.3 Molecular clips with a large
variety of side walls have been synthesized, and the
supramolecular chemistry of these clips has been exten-
sively studied.2,4 Some effort has been directed toward
modifying the glycoluril framework 1 of the hosts. The
urea functionalities in 1 have been replaced by thiourea
and guanidine groups,5 and clips have been synthesized
starting from dimethylglycoluril (1, R1 ) R2 ) Me),6
dipyridylglycoluril (1, R1 ) R2 ) 2-pyridyl), and di-
toluylglycoluril (1, R1 ) R2 ) 4-toluyl), with subsequent
functionalization of the toluyl groups.7,8,9 As could be
expected, these modifications did not significantly alter
the overall shape of the clips. The binding properties
changed considerably in the clips with guanidine and
thiourea hydrogen bond acceptor sites, the guest binding
affinities being higher in the former clips and much lower
in the latter ones.5 More recently clip molecules have
been synthesized from bipyridylglycoluril (1, R1, R2 ) 2,2′-
bipyridyl), which displayed slightly “squeezed” cavities,
resulting in rather poor binding properties.10 In a pre-
liminary communication we recently published the syn-
thesis and binding properties of a modified molecular clip
* To whom correspondence should be sent. Fax: +31-24-3652929.
† Department of Organic Chemistry.
‡ Department of Inorganic Chemistry.
(1) For an overview see: Lehn, J .-M. Comprehensive Supramolecular
Chemistry; Elsevier Science Ltd.: Oxford, 1996; Vol. 2.
(2) (a) Sijbesma, R. P.; Nolte, R. J . M. Top. Curr. Chem. 1995, 179,
25. (b) Rowan, A. E.; Elemans, J . A. A. W.; Nolte, R. J . M. Acc. Chem.
Res. 1999, 32, 995.
(3) (a) Fleischhauer, J .; Harmata, M.; Kahraman, M.; Koslowski,
A.; Welch, C. J . Tetrahedron Lett. 1997, 38, 8655. (b) Harmata, M.;
Barnes, C. L.; Rao Karra, S.; Elahmad, S. J . Am. Chem. Soc. 1994,
116, 8392. (c) Kurebayashi, H.; Fukazawa, Y. Chem. Lett. 2000, 530.
(d) Kla¨rner, F.-G.; Burkert, U.; Kamieth, M.; Boese, R.; Benet-
Buchholz, J . Chem.sEur. J . 1999, 5, 1700. (e) Kamieth, M.; Burkert,
U.; Corbin, P. S.; Dell, S. J .; Zimmerman, S. C.; Kla¨rner, F.-G. Eur. J .
Org. Chem. 1999, 2741.
(4) Reek, J . N. H.; Priem, A. H.; Engelkamp, H.; Rowan, A. E.;
Elemans, J . A. A. W.; Nolte, R. J . M. J . Am. Chem. Soc. 1997, 119,
9956.
(5) Gieling, G. T. W.; Scheeren, H. W.; Israe¨l, R.; Nolte, R. J . M.
Chem. Commun. 1996, 241.
(6) Van Nunen, J . L. M.; Nolte, R. J . M. J . Chem. Soc., Perkin Trans.
2 1997, 1473.
(7) Reek, J . N. H.; Kros, A.; Nolte, R. J . M. Chem. Commun. 1996,
245.
(8) Holder, S. J .; Elemans, J . A. A. W.; Barbera´, J .; Rowan, A. E.;
Nolte, R. J . M. Chem. Commun. 2000, 355.
(9) Rebek’s group has synthesized glycoluril derivatives with various
groups R which have been used for the construction of molecular
capsules: (a) Conn, M. M.; Rebek, J ., J r. Chem. Rev. 1997, 97, 1647.
(b) Tokunaga, Y.; Rudkevich, D. M.; Rebek, J ., J r. Angew. Chem., Int.
Ed. Engl. 1997, 36, 2656. (c) Szabo, T.; Hilmersson, G.; Rebek, J ., J r.
J . Am. Chem. Soc. 1998, 120, 6193.
10.1021/jo001317k CCC: $20.00 © 2001 American Chemical Society
Published on Web 03/22/2001

2644 J . Org. Chem., Vol. 66, No. 8, 2001
J ansen et al.
Ch a r t 2
Sch em e 1a
based on 9,9-dimethyl-2,4,6,8-tetraazabicyclo[3.3.1]nonane-
3,7-dione (2, R1 ) R2 ) Me; R3 ) R4 ) H, dimethylpro-
panediurea).11 Molecules of type 2 show strong similari-
ties with diphenylglycoluril and were first described in
1908 (R1,2 ) H or Me; R3,4 ) Me).12 Both 1 and 2 possess
a shallow cavity with two strong hydrogen bond acceptor
sites. The ureido nitrogen atoms allow for functionaliza-
tion of the framework. The main difference between 1
and 2, which triggered us to synthesize and study clip
molecules from the latter type, is the slightly sharper
angle between the carbonyl groups in 2. This structural
feature results in a cavity that is somewhat deeper with
the carbonyl oxygen atoms being closer together. A
further reason to study 2 is the fact that the substituents
R3 and R4 can be varied in order to influence the com-
plexation properties of clips derived from this compound.
The groups R1 and R2 allow for functionalization of the
clips in a similar way as reported for 1.7,8,10 In addition,
molecular modeling indicated that the side walls of clips
derived from 2 would be more parallel than those of clips
derived from 1, which might enhance the π-π stacking
interactions between a bound guest and the host.
In this paper we give a full account of the synthesis,
structure, and binding properties of a series of clip
molecules derived from 2, which show exceptionally high
affinities for resorcinol derivatives.
a
Reagents and conditions. i: H2O, H2SO4, 100 °C, 4 h, 90%. ii:
H2O, (CH2O)n, NaOH, 60 °C, 3 h. iii: H2SO4, 1,4-dimethoxyben-
zene, rt, 16 h, 23%. iv: Ac2O, HCl, 100 °C, 30 min, 17%. v: MeOH,
HCl, 50 °C, 2 h. vi: MeOH, HCl, rt, 16 h. vii: Ac2O, TFA, 95 °C,
30 min, 1,4-dimethoxybenzene, 100 °C, 30 min, 78%. viii: CH2Cl2,
SOCl2, rt, 16 h. ix: Ac2O, TFA, 1,4-dimethoxynaphthalene, 95 °C,
30 min, 58%. x: Benzene, AlCl3, reflux, 16 h. xi: 6 M HCl, reflux,
30 min, 67%.
Resu lts a n d Discu ssion
comparable to 3 could not be obtained, neither by heating
with acid nor by using dehydrating agents such as
triethyl orthoformate or molecular sieves. Since purifica-
tion of 10 turned out to be difficult, this compound was
used directly in the subsequent reaction with 1,4-
dimethoxybenzene in sulfuric acid to afford clip molecule
9 in 23% yield. Alternatively, 10 can be transformed into
the corresponding tetra(methoxymethyl) derivative 11 by
stirring in methanol in the presence of HCl at 50 °C.13
Compound 11 was reacted with 1,4-dimethoxybenzene
in an amidoalkylation reaction in acetic anhydride/TFA
to give clip molecule 9 in high yield. The intermediate in
this reaction is the tetra(acetoxymethyl) derivative 12.
This compound could be obtained directly from 10 by
heating with acetic anhydride/HCl. It was possible to
transform 12 into 11 in high yield by stirring in methanol
containing a trace of HCl. A clip molecule with 1,4-
dimethoxynaphthalene walls (13) was prepared in rea-
sonable yields by reaction of 11 with 1,4-dimethoxynaph-
thalene in acetic anhydride/TFA (Scheme 1).
Syn th esis. Clip molecules derived from diphenylgly-
coluril are usually synthesized from the cyclic ether 3
(Chart 2).2 This compound is readily available from the
base catalyzed reaction of 1 (R1 ) R2 ) Ph) with for-
maldehyde, yielding tetra(hydroxymethyl)diphenylgly-
coluril, followed by acid catalyzed dehydration.2 The re-
action of formaldehyde with 9,9-dimethyl-2,4,6,8-tetra-
azabicyclo[3.3.1]nonane-3,7-dione (8) (Scheme 1), which
was obtained by treating dimethylmalonaldehyde tetra-
ethylacetal (7) with two molecules of urea, has been
described in a patent.13 Thus, 8 was reacted with
paraformaldehyde in a slightly basic solution to give the
tetra(hydroxymethyl) derivative 10. A ring closed product
(10) Elemans, J . A. A. W.; de Gelder, R.; Rowan, A. E.; Nolte, R. J .
M. Chem. Commun. 1998, 1553.
(11) J ansen, R. J .; Rowan, A. E.; de Gelder, R.; Scheeren, H. W.;
Nolte, R. J . M. Chem. Commun. 1998, 121.
(12) De Haan, T. Recl. Trav. Chim. Pays-Bas 1908, 27, 162.
(13) Petersen, H.; Steimmig, A.; Scho¨nleben, W.; Bille, H. Ger.
Patent 2011703, 1971; Chem. Abstr. 1972, 76, 60867s.

Molecular Clips Based on Propanediurea
J . Org. Chem., Vol. 66, No. 8, 2001 2645
Sch em e 2a
Sch em e 3a
a
Reagents and conditions. i: DMSO, NaH, rt, 24 h. ii: 21, rt,
16 h, ∼30%.
stable under the reaction conditions or are sterically
blocked for reaction, leading to deformylation instead of
substitution.15
When 1,5,9,9-tetramethyl-2,4,6,8-tetraazabicyclo[3.3.1]-
nonane-3,7-dione 20 was subjected to the reaction condi-
tions described above, no clip molecule could be isolated.
To access clip molecules derived from 9,9-dimethyl-
2,4,6,8-tetraazabicyclo[3.3.1]nonane-3,7-diones substi-
tuted at the 1 and 5 positions of the skeleton (17a ,b and
20), we developed a different synthetic route (Scheme 3).
Compounds 17a ,b and 20 were treated with NaH and
1,4-dimethoxy-2,3-bis(bromomethyl)benzene (21) to give
clip molecules 22a , 22b, and 23 in moderate yields.
Compound 21 was obtained either by benzylic bromina-
tion of 1,4-dimethoxy-2,3-dimethylbenzene with NBS or
in two steps by reduction of 3,6-dimethoxyphthalic
anhydride with LiAlH4 followed by treatment with PBr3.
When 9,9-dimethyl-2,4,6,8-tetraazabicyclo[3.3.1]nonane-
3,7-dione (8) was subjected to the reaction conditions
described above, clip molecule 9 could be obtained
analogously in 20% yield.
Str u ctu r e a n d Bin d in g P r op er ties. The binding
affinities of the new host molecules 9, 22a , and 23 for a
number of hydroxybenzene derivatives (24 and 25) were
measured by NMR titration experiments in CDCl3 using
the aromatic wall protons of the host and the aromatic
protons of the guest as probes.18,19 To allow for reliable
comparisons of the binding energies of different clips, we
preferred to use competition experiments as opposed to
standard titration techniques to determine association
constants (see Experimental Section). The results are
summarized in Table 1. Of all the hosts studied, com-
pound 9 has the highest affinity for resorcinol derivatives
(Ka values up to 2.4 × 106 L mol-1). If a methyl
substituent is present at the 1-position of the clip (22a )
the association constants drop slightly (approximately
10% (∆∆G ≈ 0.3 kJ mol-1)). If methyl substituents are
present at both the 1 and 5 positions (clip 23), the
measured Ka values are approximately 30% (∆∆G ≈ 0.9
kJ mol-1) lower than those of the parent compound
a
Reagents and conditions. i: Toluene, TFA, reflux, 4 h, ∼95%.
ii: H2O, CH3CN, (CH2O)n, NaOH, 60 °C, 3 h. iii: H2SO4,
1,4-dimethoxybenzene, rt, 16 h, ∼25%.
The tetra(methoxymethyl) and tetra(acetoxymethyl)
substituted propanediurea compounds 11 and 12 did not
react with unactivated aromatic molecules such as
benzene. To synthesize a clip molecule with benzene side
walls as has been described for clips of type 1, we changed
these substituents into chlorines, thus creating better
leaving groups.14 This could be accomplished by reacting
11 or 12 with SOCl2 at room temperature, giving the
tetra(chloromethyl) derivative 14 in high yield (Scheme
1). The reaction of 14 with benzene in the presence of
AlCl3, as has been described for 4b,14 afforded however
2,6-dibenzyl-9,9-dimethyl-2,4,6,8-tetraazabicyclo[3.3.1]-
nonane-3,7-dione 15 as the sole product. Apparently,
when one of the chloromethyl groups on one side of 14
has reacted, the remaining chloromethyl group is not
reactive enough to give the bis-substituted benzene
derivative. In the subsequent hydrolytic workup the
remaining chloromethyl group is converted back to a
hydroxymethyl group, which loses formaldehyde to give
the free HNC(O) function (Scheme 1).15
An obvious way of preparing clip molecules function-
alized at R3 or R4 (see 2, Chart 1) seemed to be by starting
from 9,9-dimethyl-2,4,6,8-tetraazabicyclo[3.3.1]nonane-
3,7-dione derivatives 17. These compounds have been
synthesized by heating 2,2-dimethyl-3-oxoaldehydes 16
with urea.16 We found that reaction of compounds 16 with
2 equiv of urea in toluene at reflux temperature in the
presence of TFA as a catalyst with azeotropic removal of
water produced compounds 17 in very high yield (over
90%). The reaction product is virtually insoluble in most
solvents and precipitates from the reaction mixture.
Washing with various solvents and subsequent drying
yielded a material of high purity. Reaction of compounds
17 with paraformaldehyde in the presence of base,13
followed by reaction with 1,4-dimethoxybenzene in sul-
furic acid, yielded exclusively “half clip” compounds 19
(Scheme 2). Apparently, the hydroxymethyl groups at the
side of 18 where R is located are either not sufficiently
(16) Spa¨nig, H.; Scho¨nleben, W. Belg. Patent 645094, 1963; Chem.
Abstr. 1965, 63, 13299f.
(17) The observed ∆∆G values are very small, but we believe that
they are indicative of a change in the structure of the binding pocket.
Most competition experiments were carried out in duplo and where
possible compared with direct binding measurements. Wilcox et al.
have shown that self-association of one of the components significantly
reduces the accuracy of a titration experiment: Wilcox, C. S.; Adrian,
J . C., J r.; Webb, T. H.; Zawacki, F. J . J . Am. Chem. Soc. 1992, 114,
10189. The self-association of the resorcinol guests and the clips was
therefore investigated but was found to be too small to be measured.
(18) Sijbesma, R. P.; Kentgens, A. P. M.; Lutz, E. T. G.; Van der
Maas, J . H.; Nolte, R. J . M. J . Am. Chem. Soc. 1993, 115, 8999.
(19) (a) Whitlock, B. J .; Whitlock, H. W. J . Am. Chem. Soc. 1994,
116, 2301. (b) Alper, J . S.; Gelb, R. I.; Laufer, D. A.; Schwartz, L. M.
Anal. Chim. Acta 1989, 220, 171.
(14) (a) Sijbesma, R. P.; Kentgens, A. P. M.; Nolte, R. J . M. J . Org.
Chem. 1991, 56, 3199. (b) Sijbesma, R. P.; Nolte, R. J . M.; Recl. Trav.
Chim. Pays-Bas 1993, 112, 643.
(15) The acid hydrolysis of hydroxymethylene ureas to yield form-
aldehyde and the corresponding urea compound has been described:
(a) Nordhøy, F.; Ugelstad, J . Acta Chem. Scand. 1959, 13, 864. (b)
Petersen, H. Textile Res. J . 1971, 239.

2646 J . Org. Chem., Vol. 66, No. 8, 2001
J ansen et al.
Ch a r t 3
observed for the clip side wall protons. A shift of ∆δ -0.60
ppm was observed for the flexible clip 9, and a slightly
smaller shift of ∆δ -0.44 ppm for the rigid clip. The
aromatic wall protons of clip 22a showed shifts of ∆δ
-0.53 and -0.49 ppm, respectively. The observed ∆δ’s
indicate that upon binding of resorcinol, the side walls
of clip 9 move closer to the guest. As a result of this, the
aromatic protons of the host experience a shielding effect
caused by the guest. This effect is more pronounced when
the distance between the side wall protons and the
aromatic ring of the guest is smaller. This leads to slight
differences in shift: ∆δ -0.60 ppm for 9, ∆δ -0.44 ppm
for 23, and two values between these extremes for 22a .
The difference in binding affinity for resorcinol deriva-
tives, as found for clips 9 and 23, is mainly due to a
difference in flexibility of the cavity side walls. As can
be seen in Table 1, the difference in ∆G° is a steady 0.9
( 0.1 kJ mol-1 for all resorcinol derivatives bound in 9
and 23. Since the position of the carbonyl groups with
respect to the cavity walls is expected to be the same in
9 and 23, differences in hydrogen bonding cannot account
for the observed differences in binding affinities. This
assumption has been validated by determining the bind-
ing affinities of clips 9 and 23 for a number of 4-substi-
tuted phenol derivatives (25a -e, Table 1). In this case,
only one hydrogen bond can be formed, which is optimal
in all clip molecules. Any observed differences in binding
energies must be caused by the cavity side walls. Since
the binding affinities of 9 and 23 for 25 are relatively
low, the ∆∆G° values could not be determined as ac-
curately as in the case of the resorcinol guests. The aver-
age difference in ∆G° values between 9 and 23 again was
found to be approximately 0.9 kJ mol-1. Noteworthy in
this respect are also the CIS values of the aromatic wall
protons of the clip molecules, which are a measure of the
position of the guest with respect to the cavity side walls.
These values are higher for clip molecule 9 than for clip
23. For example, with 4-nitrophenol (25e) as a guest the
CIS values are -0.66 and -0.48 ppm for hosts 9 and 23,
respectively. This suggests that for the former host the
distance between the aromatic side walls and the guest
in the complex is smaller than for the latter host.
Ta ble 1. Associa tion Con sta n ts (Ka in L m ol-1) a n d (in
br a ck ets) F r ee En er gies of Bin d in g (k J m ol-1) of Va r iou s
Com bin a tion s of Clip Com p ou n d s a n d Su bstitu ted
Resor cin ol a n d P h en ol Gu est Com p ou n d sa
host
guest
24a c
9
22a
9 500
23
8 800
6b
13 000
(-23.5)
31 000
(-25.6)
34 000
(-25.8)
280 000
(-31.1)
2 400 000
(-36.4)
23
1 900
(-18.7)
2 600
(-19.5)
4 400
(-20.8)
16 000
(-24.0)
100 000
(-28.5)
20
(-22.7)
25 000
(-25.1)
33 100
(-25.8)
270 000
(-31.0)
2 200 000
(-36.2)
f
(-22.5)
22 600
(-24.8)
23 500
(-24.9)
200 000
(-30.2)
1 600 000
(-35.4)
16
(-6.9)
18
(-7.2)
160
24bc
24cc
24d d
24ee
25a d
25bd
25cd
25d c
25ec
(-7.8)
(-7.4)
22
f
f
f
f
29
(-7.7)
180
(-12.9)
2000
(-18.8)
4400
(-20.8)
(-8.3)
80
(-10.9)
415
(-14.9)
1200
(-17.6)
(-12.6)
1450
(-18.0)
2000
(-18.8)
a
The association constants were determined at 298 K in CDCl3.
b Values taken from ref 4. c Approximate error 20%. Approximate
d
error 30%. e Approximate error 40%. f Not determined.
without these substituents (9).17 Molecular modeling
indicates that the methyl substituent directly under the
side wall of the clip “locks” this wall in a position which
is slightly bent out due to van der Waals contacts with
the xylylene methylene protons. Once the side wall is
locked it is unable to adopt the optimal geometry for
guest binding, which results in a lower binding constant.
In the case of clip 22a , the decreased flexibility of one of
the side walls can largely be compensated for by the other
side wall, which is still flexible. This results in only a
small decrease in binding affinity. When both side walls
are “locked”, however, as is the case for clip 23, the
binding constant decreases slightly more because no
adjustment is possible. Additional evidence for this
hypothesis comes from the NMR titration data. The
resonances due to the o-xylylene protons of the different
clip molecules shift upon the addition of a guest molecule
(see Table 2). Such a shift was not observed for host
molecules of type 6. It indicates that the side walls of
the clip undergo a conformational change upon binding
of a guest (induced fit, see Figure 1). For clip 9 the
complexation induced shifts (∆δmax, CIS) were consider-
ably larger than for clip 23. Due to its asymmetry, clip
22a has two sets of signals for the o-xylylene protons:
one set for the protons on the “methyl side” and one set
for the protons on the “hydrogen side”. As expected, one
set showed large shifts, as in clip 9, and one set showed
relatively minor shifts, as in clip 23. The same trend was
From the above data it is clear that binding of guests
in clip molecules 9 and 22a , and to a much lesser extent
in clip molecule 23 takes place via an induced fit
mechanism (see Figure 1). Increased flexibility of the
host, and thus a greater ability to adopt its conformation
to a guest bound in its cavity, results in a higher binding
constant. With a strongly binding guest (5-cyanoresor-
cinol, 24e), association constants of up to 2.4 × 106
L
mol-1 can be reached.
To find additional support for the mechanism of guest
binding, we solved the X-ray structures of clips 9, 22a ,
and 22b. The structure of 9 has been published in the
preliminary communication.11 Single crystals from 22a
were obtained by the vapor diffusion technique using
chloroform/acetone as the solvent and diethyl ether as
the precipitant. Single crystals of 22b were obtained by
the liquid diffusion procedure using dichloromethane as
the solvent and hexane/acetone as the precipitant. De-
spite repeated attempts, single crystals suitable for X-ray
analysis could not be obtained for clip molecule 23. The
structures of clips 9 and 22a turned out to be very similar
(Figure 2a,b). Both form dimers in the solid state, with
the wall of one clip molecule being buried in the cavity
of another clip. Clip molecule 22b on the contrary did

Molecular Clips Based on Propanediurea
J . Org. Chem., Vol. 66, No. 8, 2001 2647
Ta ble 2. Ch em ica l Sh ifts of Selected P r oton s in F r ee Clip Molecu les 9, 22a , 22b, a n d 23 a n d Com p lex In d u ced Sh ift
(CIS) Va lu es (p p m , in p a r en th eses) of Th ese P r oton s in Host-Gu est Com p lexes of th e Clip s w ith Resor cin ol in CDCl3 a t
298 Ka ,b
HW1,2
HA1,2
5.42
(+0.45)
HB1,2
MeW
Meconv
9
6.74
(-0.60)
3.91
d
1.35
(+0.16)
1.33
(-0.28)c
22a
23
6.72
(-0.53)
6.63
5.55
5.95
3.85
3.87
1.71
(-0.49)
(+0.31)
(+0.06)
(-0.19)
(+0.01)
(+0.09)
1.82
(+0.13)
1.34
6.63
6.05
3.93
(-0.44)
(-0.04)
(+0.05)
(+0.10)
d
(+0.12)
22b
6.66
(-0.45)
6.53
(-0.43)
5.54
(+0.43)
5.42
(+0.14)
3.81
(-0.10)
3.90
1.06
(+0.28)c
(+0.15)
a
b
See structural depiction (above right). A “-” sign denotes an upfield shift and a “+” sign denotes a downfield shift. c On binding of
a resorcinol guest molecule the xylylene protons HA2 and HB2 of clips 9 and 22b change positions as depicted in the above left illustration.
Proton is not present in the molecule.
d
a size two times this distance (6.84 Å) is ideal for binding
of an aromatic guest.19a It is therefore evident that on
binding of a guest (or the side wall of a second clip mole-
cule), the walls of 22b are forced to bend outward. This
is hampered by the phenyl group under one of these side
walls. This will result in a higher barrier for guest bind-
ing. This explanation is validated by the results of the
binding constant determinations. The association con-
stant of clip 22b with resorcinol (24b) was measured to
be 2800 L mol-1, which is only 9% of the value measured
for clip 9 (Ka ) 31 000 L mol-1). For 5-chlororesorcinol
(24d ) the binding constant was 25 000 L mol-1, again 9%
of the value measured for clip 9 (Ka ) 280 000 L mol-1).
The NMR spectra recorded during the determination of
the binding constants for 22b showed some interesting
features. The pattern of the proton signals for the phenyl
group under the cavity wall of 22b changed considerably
upon addition of a guest to the solution, in line with the
idea that a guest bound in the cavity forces the side walls
to bend outward. The side wall directly above the phenyl
group when forced outward restricts the free rotation of
this phenyl group, resulting in a splitting of the signals
in the NMR spectrum. In Table 2 the CIS values of some
of the protons of clip 22b with resorcinol as a guest as
well as the chemical shifts of these protons in the
F igu r e 1. Schematic representation of the formation of host-
guest complexes from host molecules 9, 22a , and 23 and
resorcinol.
uncomplexed state are listed. Upon complexation of a
guest the side walls bend away, pushing the xylylene CH2
B2 protons out of the shielding region of one of these side
walls and the phenyl group. As a consequence a downfield
shift of +0.28 ppm results. In the case of clip molecule 9
the side walls bend inward upon complexation of a guest,
moving the xylylene CH2 B2 protons in their shielding
zones. This results in an upfield shift of -0.28 ppm (see
also drawing in note of Table 2).
not pack in this geometry (Figure 2c). This suggests that
filling up the cavity of this clip molecule is energetically
unfavorable. The X-ray structure of 22b indicates that
the side walls of this clip are in a more parallel orienta-
tion than the side walls of clip 9. The distance between
the centers of the aromatic side walls of clip 22b is 4.84
Å, as opposed to 6.43 and 6.50 Å for clips 9 and 22a ,
respectively. Since the minimum energy contact distance
between two parallel π systems is 3.42 Å, a cavity with
Clip molecules of type 9 have a considerably higher
affinity for resorcinol derivatives than clip molecules of
type 6 (Table 1). One of the reasons for this increased

2648 J . Org. Chem., Vol. 66, No. 8, 2001
J ansen et al.
F igu r e 2. Crystal structures of different host molecules: (a) 9; (b) 22a ; (c) 22b; (d) 19a ; (e) 19b; (f) 13. The drawings were made
using the PLUTON program.27
affinity is the flexibility of the side walls in the former
molecules. Hosts 6 are quite rigid molecules; their
xylylene CH2 protons exhibit no CIS upon binding of a
guest. The same ridigity is observed for clip molecule 23,
which otherwise has the same structure as clip 9. The
CIS values of the side wall aromatic protons in the case
of resorcinol as the guest are -0.60 ppm for the flexible
clip 9 and -0.44 and -0.45 ppm for the rigid clips 23
and 6, respectively. This suggests that in complexes of
resorcinol with the latter two clips, the position of the
guest with respect to the side walls is similar; the
F igu r e 3. Binding free energies of guests 24a -e in clips 6
(a) and 9 (b) as a function of the Hammett parameter (σm (R))
geometry is considerably different for the flexible clip 9.
For clips of type 6 it has been shown that the major
contribution to guest binding comes from hydrogen
bonding.4 To investigate the relative importance of
hydrogen bonding in clip 9, we plotted the binding free
energies of guests 24a -e in 9 and 6 as a function of the
Hammett [σm(R)] parameter of the substituent R of the
guest (Figure 3). The plot reveals a linear correlation,
for both clip 6 and 9. This increase in binding as the σm
increases is the result of both an increase in the hydrogen
bonding interaction and as well as an increase in the π-π
stacking as the polarization of the guest increases. The
slopes of the plots also indicate that binding of resorcinol
derivatives in clip 9 is more sensitive to the substituent
on the guest than binding in clip 6. The increase in
sensitivity of host 9 over host 6 can be attributed to two
factors, namely, a greater sensitivity of the hydrogen
bonding in host 94 and the induced fit mechanism that
operates upon guest binding to 9. As a result of the
induced fit mechanism the π-π interaction is stronger
of the substituent R in the guest.
for the more polarized guests and hence the gradient is
larger for 9 than for 6 (Figure 3).
The binding of resorcinol in the cavity molecules
described here is due to three factors, namely hydrogen
bonding between the phenolic OH groups and the urea
carbonyl groups, π-π stacking interactions between the
guest and the cavity walls, and a cavity effect.4 By com-
paring the binding affinities of resorcinol with hosts 3,
5, 6, 9, 19c, and 26 (Figure 4), the individual importance
of these factors can be determined, both for the glycoluril
based host-guest system (A) and for the propanediurea
based system (B). This additive approach, however, is
more a qualitative than a quantitative analysis. The
comparaison of the different host-guest binding energies
gives only an indication that the binding increases and
cannot be directly attributed to one specific interaction
because the sum of the different components is always
less than the whole, as has been demonstrated by

Molecular Clips Based on Propanediurea
J . Org. Chem., Vol. 66, No. 8, 2001 2649
F igu r e 4. Association constants and Gibbs free energies of binding of resorcinol in clips, half clips, and frameworks of type 3 and
of type 26 and schematic structures of the clips, half clips, and frameworks.
J encks.22 Half clip 19c was used instead of 19a for
solubility reasons; the binding affinities of the two half
clips are supposed to be comparable. The structures of
hosts 19a and 19b were established by X-ray structure
determinations (Figure 2d,e). The shapes of the different
hosts are represented schematically in Figure 4.
receptor 3: (hydrogen bonding)
-∆G ) 8.0 kJ mol-1 (1a)
receptor 26: (hydrogen bonding)
-∆G ) 12.9 kJ mol-1 (1b)
receptor 5: (hydrogen bonding + 1 ×
Receptors 3 and 26 can bind resorcinol on the basis of
hydrogen bonding only (see eqs 1a and 1b, respectively).
In receptors 5 and 19c the binding is a result of hydrogen
bonding and π-π stacking interactions of one side wall
with the resorcinol guest (see eqs 2a and 2b, respectively).
By subtracting the values in eqs 2a and 2b from the
values in equations (1a) and (1b) the energy involved in
π-π stacking interactions of one side wall with the
resorcinol guest can be calculated (see eqs 4a and 4b. In
the case of receptor molecules 6 and 9 the interaction
energy is made up of hydrogen bonding, twice the π-π
stacking interaction of the guest with one cavity wall,
and the cavity effect (see eqs 3a and 3b). The energy
attributable to the cavity effect can be calculated by sub-
tracting twice the values in eqs 4a and 4b and the values
in eqs 1a and 1b from the values in eqs 3a and 3b (see
eqs 5a and 5b). It should be noted that the hydrogen
bonding energy upon binding to 3 and 26 has a significant
unfavorable entropic component. Hence subtracting eq
1 from eq 2 overestimates the contribution of one π-π
interaction; the relative values are more an indication
of the enthalpy of the π-π interaction rather than the
free energy. However, one can say that the relative effect
of adding one wall and then a second is the same for both
cavities.
π-π interaction)
-∆G ) 10.3 kJ mol-1 (2a)
receptor 19c: (hydrogen bonding + 1 ×
π-π interaction)
-∆G ) 15.4 kJ mol-1 (2b)
receptor 6: (hydrogen bonding + 2 ×
π-π interaction + cavity effect)
-∆G ) 19.5 kJ mol-1 (3a)
receptor 9: (hydrogen bonding + 2 ×
π-π interaction + cavity effect)
-∆G ) 25.6 kJ mol-1 (3b)
1 × π-π stacking interaction (host-guest
system A)
-∆G ) 2.3 kJ mol-1 (4a)
1 × π-π stacking interaction (host-guest
system B)
-∆G ) 2.5 kJ mol-1 (4b)
cavity effect (host-guest system A)
-∆G ) 6.9 kJ mol-1 (5a)
cavity effect (host-guest system B)
-∆G ) 7.7 kJ mol-1 (5b)
(cavity effect ) -∆G for binding in a clip molecule
(eqs 3a and b) minus 2(-∆G) for π-π interaction
(eqs 4a and b) minus -∆G for hydrogen bonding
(eqs 1a and b))
(20) Hadzˇi, D.; Bratos, S. In The Hydrogen Bond; Schuster, P.,
Zundel, G., Sandorfy, C., Eds.; North-Holland Publishing Company:
Amsterdam, 1976; Vol. II, pp 565-611.
(21) Olovsson, I.; J o¨nsson P.-G. In ref 20, pp 393-456.
(22) J encks, W. P. Proc. Natl. Acad. Sci. U.S.A. 1981, 78, 4046.

2650 J . Org. Chem., Vol. 66, No. 8, 2001
J ansen et al.
F igu r e 5. Hydrogen bonding of a phenolic OH group to a
carbonyl group: (a) to the π-electrons; (b) to the n-electrons.
By comparing eqs 1a-5a with eqs 1b-5b the following
conclusions can be drawn with respect to the differences
in binding affinity of resorcinol in clip molecules 6 and
9: (i) the difference in binding energies is mainly due to
hydrogen bonding effects which are stronger for the
propanediurea framework than for the glycoluril frame-
work (∆∆G ) 4.9 kJ mol-1, see eqs 1a and 1b); (ii) the
π-π stacking interactions are approximately equal (see
eqs 4a and 4b); (iii) the cavity effects are approximately
equal (see eqs 5a and 5b).
F igu r e 6. Hydrogen bonding interactions between the guest
molecules resorcinol (top) and 2,7-dihydroxynaphthalene (bot-
tom) and host molecules of 9 (left) and 6 (right). For clarity,
only the frameworks with the hydrogen bonding sites of the
host molecules are shown.
It is clear from the above analysis that the difference
in hydrogen bonding is the major factor governing the
increase in binding affinity for resorcinol derivatives of
clip 9 compared to clip 6. Therefore, we decided to
investigate differences in hydrogen bond geometry be-
tween both host-guest complexes. A hydrogen bond with
the oxygen atom of a carbonyl group can be formed with
either the π-electrons (Figure 5a) or with the n-electrons
(Figure 5b). It was shown previously that in clip mol-
ecules derived from glycoluril, resorcinol forms hydrogen
bonds with the π-electrons. The reason for this is that
resorcinol that is being bound in the cavity prefers to
form hydrogen bonds with both carbonyl groups of the
clip molecule, which leads to geometrical constraints.18
To investigate the importance of the hydrogen bonds in
more detail we monitored the length of the hydrogen
bonds between these clips and several guests by means
of IR spectroscopy. For compounds with comparable
structural features the frequency difference between the
OH vibration in the free state and in the hydrogen
bonded state is a measure of the length of the hydrogen
bond.20 Experiments carried out with receptor 3 have
shown that the length of the hydrogen bonds is indepen-
dent of the guest and consequently of the energy involved
in hydrogen bonding.4 In the case of clip molecules of type
6 the length is dependent on the type of guest because
the binding geometry is a compromise between optimum
hydrogen bonding and optimum π-π interactions: guests
that form strong hydrogen bonds are pulled into the
cavity in order to optimize hydrogen bonding, whereas
guests that form weaker hydrogen bonds are pushed out
in order to optimize π-π stacking. IR spectra of chloro-
form solutions of mixtures of guests and hosts showed
two absorptions for the OH vibration: one for the free
species and one for the hydrogen bonded complex. The
difference between these values can thus be determined
easily. The measured OH vibration differences (∆ν) were
observed to be equal for clip molecules 6 and 9 for each
guest. For example, for resorcinol (24b) ∆ν ) 233 cm-1
with clip 6, and ∆ν ) 239 cm-1 with clip 9. For cyanore-
sorcinol (24e) ∆ν values of 301 and 302 cm-1 were
measured. (The ∆ν is larger for the latter guest because
the hydrogen bond is shorter and stronger.) This indi-
cates that for a specific guest the hydrogen bonds between
the hydroxy groups and the carbonyl oxygen atoms of the
clip have almost exactly the same length in the complexes
with both clip molecules. At a first glance this observed
similarity seems to be contradictory to the observed
differences in guest binding strength. An additional
important parameter in binding of resorcinol, however,
is the distance between the oxygen atoms of the carbonyl
groups of the host molecules which determines the
geometry of the hydrogen bond. This distance, as deter-
mined from the X-ray structures, is 5.2 Å for clip 910 and
5.5 Å for clip 6.14 If one assumes that the hydrogen bonds
are linear and have an O-H-O distance of 2.7 Å,21 the
optimal distance between the carbonyl oxygen atoms for
binding of resorcinol is 3.9 Å. Although both 9 and 6 do
not approach this value, the former clip molecule still is
better suited for resorcinol binding than the latter clip
molecule, in line with the measured binding constants.
To prove this concept the binding of 2,7-dihydroxynaph-
thalene was studied. For this guest the optimal distance
between the carbonyl oxygen atoms is larger, viz 6.3 Å.
This means that host 6 should have a higher affinity for
this guest than host 9 (see Figure 6). This could indeed
be confirmed by NMR titration experiments. Host 9 binds
2,7-dihydroxynaphthalene with an association constant
of Ka ) 2300 ( 200 L mol-1, whereas for host 6 the
association constant is Ka ) 7100 ( 500 L mol-1 18
We
.
also measured the binding affinities of 9 and 6 for
catechol. The association constants amounted to Ka ) 130
( 15 L mol-1 and Ka ) 60 ( 10 L mol-1, respectively.
These low numbers indicate that the two clips are not
very well equipped to complex this guest. Catechol forms
an intramolecular hydrogen bond that needs to be broken
to allow formation of two hydrogen bonds between this
guest and the urea carbonyl groups of the clips. Never-
theless, clip 9 is better suited than clip 6 to match the
hydrogen bond donor sites of catechol, in line with the
ideas presented above.
It has been previously shown that resorcinol binds in
an asymmetric geometry in a clip that possesses only one
hydrogen bond acceptor group.4 It is assumed that in clips
with two hydrogen bond acceptor groups resorcinol

Molecular Clips Based on Propanediurea
J . Org. Chem., Vol. 66, No. 8, 2001 2651
solution, and paraformaldehyde (0.250 g, 8.33 mmol) were
stirred at 60 °C for 3 h. The resulting clear solution was
neutralized by bubbling CO2 through it for a few minutes. The
solution was concentrated in vacuo, yielding impure 10 (0.602
g). A solution of 1,4-dimethoxybenzene (1.38 g, 10.0 mmol) in
16 mL of concentrated sulfuric acid was then added and the
mixture was stirred for 16 h at room temperature. The
resulting solution was poured onto 200 g of crushed ice, made
cannot form two perfect hydrogen bonds simultaneously
because of the geometry of the host carbonyl groups. We
postulate given the experimental evidence that in hosts
of type 9 the hydrogen bond geometry is closer to the
optimum than for hosts of type 6 and hence stronger
hydrogen bonding occurs. It is quite remarkable how such
slight changes in geometry can induce such significant
increases in binding.
basic with
a concentrated aqueous NaOH solution, and
extracted with CH2Cl2 (3 × 100 mL). The organic layer was
washed once with water (50 mL) and dried (Na2SO4). The
solution was concentrated in vacuo, and the resulting yellowish
powder was purified through a short column (EtOH/CH2Cl2,
2:98 v/v) yielding 9 as a white powder (0.242 g, 0.476 mmol,
23%). (B) From 11. To a mixture of TFA (1 mL) and acetic
anhydride (1 mL) was added 11 (400 mg, 1.11 mmol). The
resulting suspension was heated at 95 °C for 30 min. Then
1,4-dimethoxybenzene (337 mg, 2.44 mmol) was added and the
resulting solution was heated at 100 °C for 30 min. The
reaction mixture was cooled on ice and 6 mL of methanol was
cautiously added. The precipitate was filtered off, washed with
cold methanol, and dried in vacuo, yielding 9 as a white powder
(439 mg, 0.864 mmol, 78%). (C) From 8 and 21. In 50 mL of
degassed DMSO were suspended 8 (331 mg, 1.80 mmol) and
NaH (191 mg, 7.96 mmol), and the mixture was stirred under
a nitrogen atmosphere at room temperature for 24 h. Then,
21 (1.20 g, 3.70 mmol) was added, and the reaction mixture
was stirred for another 16 h at room temperature. The
resulting solution was poured onto 150 g of crushed ice. The
resulting precipitate was filtered off, washed with water and
ethanol, and purified by column chromatography (EtOH/CH2-
Cl2, 2:98 v/v) yielding 9 as a white powder (183 mg, 0.360
mmol, 20%). An analytically pure sample was obtained by
recrystallization from MeOH/CHCl3. Crystals suitable for
X-ray analysis were obtained by slow diffusion of diethyl ether
in a CH2Cl2 solution of the compound: mp 348 °C subl; EIMS
m/z 508 M+ (100); 1H NMR (CDCl3) δ 1.35 (6H, s), 3.78 (12H,
s), 3.90 (4H, d, J ) 15.2 Hz), 4.32 (2H, s), 5.42 (4H, d, J )
15.2 Hz), 6.74 (4H, s); 13C NMR (CDCl3) δ 23.04, 32.18, 44.73,
57.44, 79.10, 112.57, 128.31, 151.85, 154.42; IR (KBr) ν ) 1638,
1659 cm-1 (CdO). Anal. Calcd for C28H364O7 (9‚MeOH): C,
62.21; H, 6.71; N, 10.36. Found: C, 62.21; H, 6.66; N, 10.28.
Com p ou n d 12. Paraformaldehyde (1.86 g, 62.0 mmol) was
dissolved by heating in 7.5 mL water containing 5 drops of an
aqueous 2 N NaOH solution. Then, 9,9-dimethyl-2,4,6,8-
tetraazabicyclo[3.3.1]nonane-3,7-dione (8) (2.85 g, 15.5 mmol)
was added, and the mixture was heated at 60 °C until it
became clear. The resulting solution was stirred for 30 min at
room temperature and was neutralized by bubbling CO2
through it for a few minutes. The solution was concentrated
in vacuo and 25 mL of acetic anhydride, and 5 drops of aqueous
concentrated HCl was added. The mixture was stirred at 100
°C for 30 min and poured onto 300 g of crushed ice. The
solution was neutralized with aqueous NaOH and extracted
with CH2Cl2 (2 × 200 mL). The organic layer was washed with
water (50 mL), dried (MgSO4), and concentrated in vacuo. The
crude product was passed through a short column (EtOH/CH2-
Cl2, 2:98 v/v). The first fraction was collected and recrystallized
twice from 1-propanol, yielding white crystalline 12 (1.25 g,
2.65 mmol, 17%): mp 164-166 °C; EIMS m/z 267 (100), 442
Con clu sion s
The synthesis of a series of new receptors derived from
propanediurea has been presented. These molecules are
U-shaped and show a high affinity for resorcinol deriva-
tives. The binding properties of these hosts have been
studied in detail and have been compared to those of
previously developed clip molecules derived from diphen-
ylglycoluril. The enhanced binding properties of the new
hosts can be accounted for by the smaller distance
between the hydrogen bond acceptor sites in these clip
molecules. Although the differences in geometry are sub-
angstrom (∼0.3 Å), the effect on the binding properties
is very large, highlighting the necessity for host-guest
complementarity. It was found that the affinity of the
propanediurea receptors for resorcinol can be fine-tuned
by introducing substituents under the side walls of the
molecules.
The very high binding affinities of the new hosts (Ka
up to 2 400 000 L mol-1) can be used in the construction
of supramolecular aggregates. Work in this direction is
in progress.
Exp er im en ta l Section
Gen er a l. Thionyl chloride was distilled prior to use. Sodium
hydride was a commercially available 60% suspension in
mineral oil. Benzene, diethyl ether, and THF were distilled
under nitrogen from sodium benzophenone ketyl. Carbon
tetrachloride was dried on molecular sieves (4 Å). All other
solvents and chemicals were commercially available materials
and were used as received. Merck silica gel (60H) was used
for column chromatography. Melting points were determined
on a J eneval polarization microscope THMS 600 hot stage and
are uncorrected. IR spectra were recorded on a Perkin-Elmer
FTIR 1720-X spectrometer. 1H and 13C NMR spectra were
recorded on a Bruker AM-300 instrument with (CH3)4Si as the
1
internal standard (δ 0.00 ppm) for the H spectra and CDCl3
as the internal standard (δ 77.0 ppm) for the 13C spectra.
Abbreviations used are as follows: s, singlet; d, doublet; t,
triplet; m, multiplet. MS spectra were recorded on a VG 7070E
instrument. Elemental analyses was determined with a Carbo
Erba EA 1108 instrument.
Com p ou n d s. The syntheses of compounds 3,14 6,14 and 523
have been reported elsewhere. 1,4-Dimethoxynaphthalene and
1,4-dimethoxy-2,3-dimethylbenzene were prepared by methy-
lation of 1,4-dihydroxynaphthalene and 2,3-dimethylhydro-
quinone, respectively, using dimethyl sulfate and KOH as
reagents. Compounds 8,24 11,13 20,12 26,25 and 3,6-dimethoxy-
phthalic anhydride26 were prepared according to literature
procedures.
1
[M - 2CH3]+; H NMR (CDCl3) δ 1.12 (6H, s), 2.09 (12H, s),
4.98 (2H, s), 5.44 (4H, d, J ) 10.8 Hz), 5.54 (4H, d, J ) 10.8
Hz); 13C NMR (CDCl3) δ 21.27, 21.82, 31.73, 72.02, 152.41,
171.61; IR (KBr)R ν ) 1669, 1745 cm-1 (CdO). Anal. Calcd
for C19H284O10: C, 48.30; H, 5.97; N, 11.86. Found: C, 48.50;
H, 5.94; N, 11.69.
Com p ou n d 13. Prepared as described for 9, method B, from
11 (116 mg, 0.322 mmol) and 1,4-dimethoxynaphthalene (140
mg, 0.745 mmol) yielding 13 as a white powder (113 mg, 0.186
mmol, 58%). An analytically pure sample was obtained by slow
diffusion of diethyl ether in a dichloromethane solution of the
compound to give 13 as white needles. Crystals suitable for
Com p ou n d 9. (A) From 8. A mixture of 9,9-dimethyl-
2,4,6,8-tetraazabicyclo[3.3.1]nonane-3,7-dione (8) (0.380 g, 2.07
mmol), water (0.9 mL), 3 drops of an aqueous 1 N NaOH
(23) Reek, J . N. H.; Elemans, J . A. A. W.; Nolte, R. J . M. J . Org.
Chem. 1997, 62, 2234.
(24) Khasapov, B. N.; Novikova, T. S.; Lebedev, O. V.; Khmel′nitskii,
L. I.; Novikov, S. S. Zh. Organ. Khim. 1973, 9, 23.
(25) Eres’ko, V. A.; Epishina, L. V.; Lebedev, O. V.; Khmel′nitskii,
L. I.; Novikov, S. S.; Povstyanoi, M. V.; Kulik, A. F. Bull. Acad. Sci.
USSR, Chem. Sec. 1978, 5, 1073.
(27) Spek, A. L. PLUTON. A program for plotting molecular and
crystal structures; University of Utrecht: The Netherlands, 1995.
(26) Cardani, C.; Piozzi, F. Rend. Sc. Fis. Mat. Nat. 1952, 12, 719.

2652 J . Org. Chem., Vol. 66, No. 8, 2001
J ansen et al.
X-ray analysis were prepared by slow diffusion of hexane in a
neutralized with a concentrated aqueous solution of NaOH and
NaHCO3, and extracted with CH2Cl2 (4 × 50 mL). The
combined organic layers were washed with water (2 × 25 mL),
dried (MgSO4), and concentrated in vacuo yielding 19a as a
white powder (78 mg, 0.22 mmol, 20%). Crystals suitable for
X-ray analysis were prepared by slow diffusion of diethyl ether
in a MeOH/CHCl3 solution of the compound: mp 320-322
°C dec; EIMS m/z 360 M+ (100); 1H NMR (CDCl3) δ 1.27
(3H, s), 1.30 (6H, s), 3.72 (2H, d, J ) 15.5 Hz), 3.86 (s, 6H),
4.41 (1H, s), 4.90 (2H, s), 5.82 (2H, d, J ) 15.5 Hz), 6.82 (2H,
s); 13C NMR (CDCl3:MeOD-d4 4:1): δ 19.43, 20.35, 34.72,
44.08, 56.77, 69.08, 81.25, 112.14, 128.05, 151.40, 154.92; IR
(KBr) ν ) 1631, 1676 cm-1 (CdO). Anal. Calcd for C19H284O5
(19a .MeOH): C, 58.15; H, 7.19; N, 14.28. Found: C, 58.16; H,
7.13; N, 14.10.
Com p ou n d 19b. Prepared as decribed for 19a from 17b
(1.42 g, 5.46 mmol), paraformaldehyde (660 mg, 22 mmol), and
1,4-dimethoxybenzene (3.0 g, 22 mmol). The crude product was
purified through a short column (EtOH/CH2Cl2, 5:95 v/v)
yielding 19b (580 mg, 1.41 mmol, 26%). Crystals suitable for
X-ray analysis were prepared by slow diffusion of hexane in a
MeOH/CH2Cl2 solution of the compound, followed by slow
evaporation: mp 333-334 °C; EIMS m/z 422 M+ (100); 1H
NMR (CDCl3) δ 1.15 (6H, s), 3.78 (2H, d, J ) 15.4 Hz), 3.89
(6H, s), 4.49 (1H, s), 5.20 (2H, s), 5.8 (2H, d, J ) 15.4 Hz),
6.85 (2H, s), 7.40-7.58 (5H, m); 13C NMR (CDCl3) δ 21.29,
36.52, 44.54, 57.38, 74.88, 81.90, 112.76, 127.78, 128.35,
128.57, 129.32, 151.79, 154.71; IR (KBr) ν ) 1636, 1644, 1674
cm-1 (CdO). Anal. Calcd for C24H284O4Cl2 (19b.CH2Cl2): C,
56.90; H, 5.58; N, 11.07. Found: C, 56.88; H, 5.48; N, 11.18.
CH2Cl2 solution of the compound: mp 326-328 °C; EIMS m/z
1
608 M+ (100); H NMR (CDCl3) δ 1.45 (6H, s), 4.03 (12H, s),
4.05 (4H, d, J ) 15.6 Hz), 4.42 (2H, s), 5.77 (4H, d, J ) 15.6
Hz), 7.43-7.47 (4H, m), 7.99-8.02 (4H, m); 13C NMR (CDCl3)
δ 23.01, 32.01, 46.47, 62.64, 79.80, 123.05, 126.28, 127.13,
128.34, 150.38, 154.18; IR (KBr) ν ) 1663 cm-1 (CdO). Anal.
Calcd for C36H384O6Cl2 (13.CH2Cl2): C, 62.41; H, 5.53; N, 8.09.
Found: C, 62.27; H, 5.48; N, 8.04.
Com p ou n d 14. (A) From 12. A solution of 12 (321 mg, 0.680
mmol) in 2.5 mL of thionyl chloride was stirred under a
nitrogen atmosphere for 16 h at room temperature. The solvent
was removed in vacuo yielding 14 as a white, very hygroscopic
powder (255 mg, 0.675 mmol, 99%). (B) From 11. A solution
of 11 (1.00 g, 2.78 mmol) in a mixture of CH2Cl2 (4 mL), thionyl
chloride (9 mL), and one drop of water was stirred under a
nitrogen atmosphere for 16 h at room temperature. Diethyl
ether (40 mL) was added, and the reaction mixure was cooled
to 0 °C. The mixture was filtered, and the residue was washed
with dry diethyl ether. The resulting white solid was redis-
solved in dry CH2Cl2 and filtered. The solvent was removed
in vacuo yielding 14 as a white powder (0.950 g, 2.51 mmol,
1
90%): FAB-MS m/z 329 [M- CH2Cl]+ (100); H NMR (CDCl3)
δ 1.28 (6H, s), 4.65 (2H, s), 5.35 (4H, d, J ) 10.7 Hz), 5.67
(4H, d, J ) 10.7 Hz); 13C NMR (CDCl3) δ 22.16, 31.73, 58.95,
75.62, 149.66. Due to the instability of the compound, no
reproducible elemental analysis could be obtained.
Com p ou n d 15. A mixture of 14 (399 mg, 1.06 mmol), AlCl3
(920 mg, 6.89 mmol), and 6 mL of dry benzene was refluxed
under a nitrogen atmosphere for 16 h. The reaction mixture
was allowed to cool to room temperature, and an aqueous 6 N
HCl solution (20 mL) was added. The resulting mixture was
refluxed for 30 min, cooled, and extracted with CH2Cl2 (2 ×
100 mL). The organic layer was washed with aqueous 2 N
NaOH (50 mL) and water (50 mL), dried (MgSO4), and
concentrated in vacuo. The resulting brown powder was
washed with diethyl ether to give (()-15 as a white powder
(260 mg, 713 mmol, 67%). A sample was recrystallized for
elemental analysis by slow diffusion of diethyl ether in a CH2-
Cl2/MeOH solution of the compound, giving pure 15 as white
crystals: mp 305-306 °C; EIMS m/z 365 [M+H]+ (100); 1H
NMR (CDCl3) δ 1.08 (6H, s), 3.91 (2H, d, J ) 3.9 Hz), 4.23
(2H, d, J ) 15 1 Hz), 4.98 (2H d, J ) 15.0 Hz), 5.63 (2H, d, J
) 3.6 Hz), 7.27-7.37 (10H, m); 13C NMR (CDCl3) δ 22.55,
31.74, 48.76, 68.74, 127.87, 128.51, 128.85, 137.03, 153.81; IR
(KBr) ν ) 1650, 1672 cm-1 (CdO). Anal. Calcd for C21H244O2:
C, 69.21; H, 6.64; N, 15.37. Found: C, 69.39; H, 6.60; N, 15.23.
Com p ou n d 17a . To toluene (90 mL) were added 16a (2.41
g, 21.1 mmol), urea (2.54 g, 42.3 mmol), and TFA (2 mL). The
mixture was refluxed in an inert atmosphere for 4 h with
azeotropic removal of water. Subsequently it was cooled and
filtered. The precipitate was washed with ethanol and dried
in vacuo, yielding 3.93 g (19.8 mmol, 94%) of 17a as a white
Com p ou n d 19c. Prepared as decribed for 19a from 17c
(501 mg, 2.22 mmol), paraformaldehyde (273 mg, 9.10 mmol),
and 1,4-dimethoxybenzene (790 mg, 5.72 mmol). Yield (210
mg, 0.54 mmol, 24%) 19c as a white powder: mp 322 °C dec;
1
EIMS m/z 388 M+ (100); H NMR (CDCl3) δ 0.99 (3H, t, J )
7.1 Hz), 1.23-1.34 (8H, m), 1.57-1.63 (2H, m), 3.72 (2H, d, J
) 15.4 Hz), 3.86 (6H, s), 4.38 (1H, s), 5.04 (2H, s), 5.82 (2H, d,
J ) 15.4 Hz), 6.82 (2H, s); 13C NMR (CDCl3) δ 14.23, 14.44,
20.79, 34.65, 35.69, 44.26, 57.34, 70.71, 81.69, 112.63, 128.38,
151.68, 154.19; IR (KBr) ν ) 1646, 1675 cm-1 (CdO). Anal.
Calcd for C21H304O4Cl2 (19c‚CH2Cl2): C, 53.28; H, 6.39; N,
11.83. Found: C, 53.40; H, 6.21; N, 11.89.
Com p ou n d 22a . Prepared as described for 9 (method C)
from 17a (554 mg, 2.80 mmol) and 21 (1.95 mg, 6.02 mmol).
The crude product was purified by column chromatography
(EtOH/CH2Cl2 2/98 v/v) yielding 22a as a white powder (366
mg, 0.700 mmol, 25%). Crystals suitable for X-ray analysis
were prepared by the vapor diffusion technique using acetone/
CHCl3 as the solvent and diethyl ether as the precipitant: mp
352-353 °C subl; EIMS m/z 522 M+ (100); 1H NMR (CDCl3) δ
1.33 (6H, s), 1.71 (3H, s), 3.74 (6H, s), 3.77 (6H, s), 3.85 (2H,
d, J ) 15.2 Hz), 3.87 (2H, d, J ) 17.1 Hz), 4.27 (1H, s), 5.55
(2H, d, J ) 15.2 Hz), 5.97 (2H, d, J ) 17.2 Hz), 6.63 (2H, s),
6.72 (2H, s); 13C NMR (CDCl3) δ 18.30, 22.62, 35.23, 39.35,
45.24, 56.85, 57.34, 82.14, 110.38, 112.36, 128.51, 129.38,
151.58, 151.74, 154.29; IR (KBr) ν ) 1624, 1652 cm-1 (CdO).
Anal. Calcd for C29H354O6Cl3 (22a ‚CHCl3): C, 54.26; H, 5.50;
N, 8.73. Found: C, 54.44; H, 5.44; N, 8.62.
powder. Recrystallization from boiling water (∼1 g L-1
)
provided white crystals.
Com p ou n d s 17b. Prepared as described for 17a starting
from 16b (8.00 g, 48.8 mmol) and urea (5.86 g, 97.7 mmol).
Yield 11.8 g (45.4 mmol, 93%) of 17b as a white powder.
Crystalline material was obtained by recrystallization from
boiling water.
Com p ou n d s 17c. Prepared as described for 17a starting
from 16c (2.75 g, 19.4 mmol) and urea (2.40 g, 40.0 mmol).
Yield 4.35 g (19.2 mmol, 99%) of 17c as a white powder.
Crystalline material was obtained by recrystallization from
boiling water.
Com p ou n d 19a . In a mixture of acetonitrile (7 mL) and
water (2 mL) were suspended 17a (213 mg, 1.08 mmol) and
paraformaldehyde (230 mg, 7.67 mmol). Two drops of an
aqueous 2 N NaOH solution were added and the mixture was
stirred for 16 h at 60 °C. The resulting solution was cooled
and CO2 was bubbled through it for a few minutes. The solvent
was removed in vacuo and 1,4-dimethoxybenzene (1.01 g, 7.31
mmol) in concentrated sulfuric acid (12 mL) was added. The
reaction mixture was stirred for 16 h at room temperature.
The resulting solution was poured onto 100 g of crushed ice,
Com p ou n d 22b. Prepared as described for 9 (method C)
from 17b (350 mg, 1.35 mmol) and 21 (923 mg, 2.85 mmol).
The crude product was purified through a short column
(MeOH/CH2Cl2 2/98 v/v) yielding 22b as a white powder (245
mg, 0.419 mmol, 30%). An analytically pure sample was
obtained by liquid diffusion using CH2Cl2 as the solvent and
diethyl ether as the precipitant. Crystals suitable for X-ray
analysis were prepared by slow diffusion of hexane/acetone
in a CH2Cl2 solution of the compound: mp >350 °C; EIMS
1
m/z 584 M+ (100); H NMR (CDCl3) δ 1.06 (6H, s), 3.73 (6H,
s), 3.77 (6H, s), 3.81 (2H, d, J ) 15.0 Hz), 3.89 (2H, d, J )
15.2 Hz), 4.35 (1H, s), 5.42 (2H, d, J ) 15.2 Hz), 5.54 (2H, d,
J ) 15.0 Hz), 6.53 (2H, s), 6.66 (2H, s), 7.27-7.43 (5H, m); 13
C
NMR (CDCl3) δ 23.03, 36.63, 41.76, 45.00, 57.39, 57,73, 79.49,
84.58, 112.00, 112.84, 127.84, 128.14, 128.59, 128.82, 129.61,
135.55, 151.92, 155.65; IR (KBr) ν ) 1654, 1675 cm-1 (CdO).

Molecular Clips Based on Propanediurea
J . Org. Chem., Vol. 66, No. 8, 2001 2653
Anal. Calcd for C33H364O6: C, 67.79; H, 6.21; N, 9.58. Found:
C, 67.91; H, 6.26; N, 9.42.
(CDCl3) δ 23.92, 56.24, 112.12, 126.37, 151.74. Anal. Calcd for
C
10H12O2Br2: C, 37.19; H, 3.43. Found: C, 37.19; H, 3.58.
Bin d in g Exp er im en ts.17 The association constants of
Com p ou n d 23. Prepared as described for 9 (method C) from
20 (326 mg, 1.54 mmol) and 21 (1.02 g, 3.15 mmol). The crude
product was recrystallized from CHCl3/EtOH yielding 23 as a
white crystalline solid (234 mg, 0.436 mmol, 28%): mp 340-
341 °C subl; EIMS m/z 536 M+ (100); 1H NMR (CDCl3) δ 1.35
(6H, s), 1.83 (6H, s), 3.73 (12H, s), 3.93 (4H, d, J ) 17.4 Hz),
6.04 (4H, d, J ) 17.2 Hz), 6.59 (4H, s); 13C NMR: δ 19.02,
20.96, 39.21, 40.06, 56.74, 74.22, 110.20, 129.35, 151.51,
complexes with all guests except 24d ,e were determined by
1H NMR shift titrations at 298 K on a Bruker AM-300
instrument with (CH3)4Si as the internal standard, using fresh
CDCl3, predried on molecular sieves (4 Å). The concentration
of the fixed component was approximately 0.5 mM when Ka
values were greater than 2000 M-1, 4 mM when Ka values were
smaller than 250 M-1, and Ka in other cases. The concentra-
-1
154.44; IR (KBr) ν ) 1652 cm-1 (CdO). Anal. Calcd for C30H37
-
tion of the other component was varied between the concentra-
tion of the fixed component and 10 times this value (or 15 times
for Ka values lower than 150). Since for guest 24c very high
binding constants were measured, these values were double-
checked in competition experiments with 24b. For complexes
with guests 24d ,e no accurate values could be obtained from
1H NMR shift titrations since the binding constants are too
large to be measured by this technique. Therefore binding
constants of clips with guests 24d and 24e were determined
in competition experiments with 24c and 24d , respectively.
For comparison of binding properties of clips 9 and 22a and
clips 9 and 23 we also used competition experiments. For
competition experiments a 1 mM solution of a guest (or host)
was prepared. This solution was used to prepare a solution
containing 1 mM guest (or host) and 1 mM host A (or guest
A) and a solution containing 1 mM guest (or host) and 1 mM
host B (or guest B). Aliquots of these two solutions were mixed
to a total volume of 0.6 mL and measured. Due to the
experimental setup of the competition experiments, the relative
association constant Krel can be determined very accurately.
Generally the error margins are smaller than 5%.19b This
method is therefore very useful when comparing the affinity
of two hosts for a certain guest molecule. If this affinity is
similar, comparison of the binding constants determined by
standard titrations is impossible because of the relatively large
error margins. Direct competition experiments do allow for
reliable comparison.
4O6Cl3 (23‚CHCl3): C, 54.93; H, 5.68; N, 8.54. Found: C, 54.98;
H, 5.69; N, 8.48.
1,4-Dim eth oxy-2,3-bis-(h yd r oxym eth yl)ben zen e. To a
cooled suspension of LiAlH4 (4 g, 0.11 mol) in THF (250 mL)
was slowly added a suspension of 3,6-dimethoxyphthalic
anhydride (2.0 g, 9.6 mmol) in THF (300 mL). The reaction
mixture was refluxed for 48 h in an inert atmosphere. The
mixture was cooled on ice, and water was added dropwise until
evolution of hydrogen stopped. Water (200 mL) was added, and
the mixture was extracted with CHCl3 (4 × 100 mL). The
combined organic layers were washed with water (150 mL),
dried (MgSO4), and concentrated in vacuo, yielding the crude
product. Recrystallization from toluene/methanol gave a white
crystalline solid (1.58 g, 7.98 mmol, 83%): mp 145 °C; EIMS
1
m/z 198 M+ (100); H NMR (CDCl3) δ 2.87 (2H, s), 3.81 (6H,
s), 4.82 (4H, s), 6.83 (2H, s); 13C NMR: δ 56.15, 56.42, 111.04,
129.45, 151.81. Anal. Calcd for C10H14O4: C, 60.60; H, 7.12.
Found: C, 60.70; H, 7.02.
1,4-Dim eth oxy-2,3-bis-(br om om eth yl)ben zen e (21). (A)
From 1,4-dimethoxy-2,3-dimethylbenzene. A mixture of N-
bromosuccinimide (4.98 g, 28 mmol), 1,4-dimethoxy-2,3-di-
methylbenzene (2.32 g, 14 mmol), and 50 mL of dry CCl4 was
irradiated with a 200 W lamp at reflux temperature in a
nitrogen atmosphere for 1 h. The solution was cooled to room
temperature, filtered, and concentrated in vacuo yielding 21
as a white powder (4.36 g, 13.5 mmol, 96%). (B) From 1,4-
dimethoxy-2,3-bis-(hydroxymethyl)benzene. To a cooled sus-
pension of 1,4-dimethoxy-2,3-bis-(hydroxymethyl)benzene (6.26
g, 31.6 mmol) in 200 mL of dry diethyl ether was slowly added
a solution of PBr3 (20 g, 74 mmol) in diethyl ether (10 mL).
The suspension was then refluxed for 2 h and allowed to cool
to room temperature. The mixture was cautiously poured onto
200 g of crushed ice and extracted with diethyl ether (3 × 250
mL). The combined organic layers were washed with water
(50 mL), dried (Na2SO4), and concentrated in vacuo yielding
21 as a white powder (9.97 g, 30.8 mmol, 97%). An analytically
pure sample was prepared by crystallization from toluene: mp
149 °C; EIMS m/z 164 [M - 2Br]+ (100), 324 M+; 1H NMR
(CDCl3) δ 3.86 (6H, s), 4.74 (4H, s), 6.84 (2H, s); 13C NMR
Ack n ow led gm en t. The authors thank Mr. E. T. H.
Lutz of the Department of Analytical Molecular Spec-
troscopy of the University of Utrecht for his help with
the IR measurements.
Su p p or tin g In for m a tion Ava ila ble: X-ray structural
information on 9, 13, 19a , 19b, 22a , and 22b (PDF) and
ORTEP drawings of these structures. This material is avail-
able free of charge via the Internet at http://pubs.acs.org.
J O001317K