Self-sorting molecular clips

Article abstract of DOI:10.1021/jo8009424

(Figure Presented) We report the synthesis and characterization of 12 C-shaped methylene-bridged glycoluril dimers (1-12) bearing H-bonding groups on their aromatic rings. Compounds 1, 2, (±)-4a, (±)-5, (±)-7, and 8 form tightly associated homodimers in CDCl₃, due to the combined driving force of π,π and H-bonding interactions. Compounds 2, (±)-5, and 8, having disparate spatial distribution of their H-bonding groups, display the ability to efficiently distinguish between self and nonself even within three-component mixtures in CDCl₃. When the spatial distributions of the H-bonding groups of the molecular clips are similar (e.g., 1 and 2), a mixture of homodimers and heterodimers is formed. The effect of various structural modifications (e.g., chirality, side chain steric bulk, number and pattern of H-bonds) on the strength of self-assembly and the fidelity of self-sorting are presented. On the basis of these results we prepared self-sorting systems comprising three (e.g., 1, (±)-5, and (±)-7) and even four (2, (±)-5, 9, and 10) components. The potential of molecular clips 1-12 as robust, functionalizable, self-sorting modules to control the noncovalent interaction network in systems chemistry studies is described.

Full text of DOI:10.1021/jo8009424

Self-Sorting Molecular Clips
Soumyadip Ghosh,† Anxin Wu,*,‡ James C. Fettinger,† Peter Y. Zavalij,† and Lyle Isaacs*,†
Department of Chemistry and Biochemistry, UniVersity of Maryland, College Park, Maryland 20742, and
Key Laboratory of Pesticide and Chemical Biology, Ministry of Education, Central China Normal
UniVersity, Wuhan 430079, People’s Republic of China
LIsaacs@umd.edu
ReceiVed May 1, 2008
We report the synthesis and characterization of 12 C-shaped methylene-bridged glycoluril dimers (1-12)
bearing H-bonding groups on their aromatic rings. Compounds 1, 2, (()-4a, (()-5, (()-7, and 8 form
tightly associated homodimers in CDCl3, due to the combined driving force of π-π and H-bonding
interactions. Compounds 2, (()-5, and 8, having disparate spatial distribution of their H-bonding groups,
display the ability to efficiently distinguish between self and nonself even within three-component mixtures
in CDCl3. When the spatial distributions of the H-bonding groups of the molecular clips are similar (e.g.,
1 and 2), a mixture of homodimers and heterodimers is formed. The effect of various structural
modifications (e.g., chirality, side chain steric bulk, number and pattern of H-bonds) on the strength of
self-assembly and the fidelity of self-sorting are presented. On the basis of these results we prepared
self-sorting systems comprising three (e.g., 1, (()-5, and (()-7) and even four (2, (()-5, 9, and 10)
components. The potential of molecular clips 1-12 as robust, functionalizable, self-sorting modules to
control the noncovalent interaction network in systems chemistry studies is described.
Introduction
via dynamic combinatorial chemistry, drug delivery, membrane
transport agents, DNA nanotechnology, control of otherwise
unfavorable chemical reactions, and supramolecular polymers.2,3
In all of these applications, the ready availability of robust and
easily functionalized modules capable of specific noncovalent
interactions (e.g., π-π, metal-ligand, and H-bonding modules)
with their targets in organic or aqueous solution has proved
invaluable.4–6 For example, quadruple hydrogen-bonding mod-
ules with high values of Ka have been used in the formation of
high-molecular-weight supramolecular polymers.6 To date,
however, the majority of such functional systems have been
demonstrated to operate only in isolation and not as components
of more complex chemical or biological systems.
As supramolecular chemistry has continued to develop
following the pioneering work of Cram, Lehn, and Pedersen,
the focus of much research in the area has shifted from an
understanding of the fundamental aspects of noncovalent
interactions, molecular recognition, and self-assembly and
toward the use of such information to construct systems with
function.1 For example, supramolecular chemistry has provided
new approaches toward chemical sensing ensembles, discovery
† University of Maryland.
‡ Central China Normal University.
(1) Cram, D. J. Angew. Chem., Int. Ed. Engl. 1988, 27, 1009–1020. Lehn,
J. M. Angew. Chem., Int. Ed. Engl. 1988, 27, 89–112. Pedersen, C. J. Angew.
Chem., Int. Ed. Engl. 1988, 27, 1021–1027.
(2) Corbett, P. T.; Leclaire, J.; Vial, L.; West, K. R.; Wietor, J.-L.; Sanders,
J. K. M.; Otto, S. Chem. ReV. 2006, 106, 3652–3711. Lehn, J.-M. Chem. Eur.
J. 1999, 5, 2455–2463. Wright, A. T.; Anslyn, E. V. Chem. Soc. ReV. 2006, 35,
14–28. Davis, A. P.; Sheppard, D. N.; Smith, B. D. Chem. Soc. ReV. 2007, 36,
348–357. Seeman, N. C. Mol. Biotechnol. 2007, 37, 246–257. Jon, S. Y.; Ko,
Y. H.; Park, S. H.; Kim, H.-J.; Kim, K. Chem. Commun. 2001, 1938–1939.
Sisson, A. L.; Shah, M. R.; Bhosale, S.; Matile, S. Chem. Soc. ReV. 2006, 35,
1269–1286. Haag, R.; Kratz, F. Angew. Chem., Int. Ed. 2006, 45, 1198–1215.
Yoshizawa, M.; Tamura, M.; Fujita, M. Science 2006, 312, 251–254. Pluth,
M. D.; Bergman, R. G.; Raymond, K. N. Science 2007, 316, 85–88.
(3) Pollino, J. M.; Weck, M. Chem. Soc. ReV. 2005, 34, 193–207.
(4) Fujita, M.; Umemoto, K.; Yoshizawa, M.; Fujita, N.; Kusukawa, T.;
Biradha, K. Chem. Commun. 2001, 50, 9–518. Seidel, S. R.; Stang, P. J. Acc.
Chem. Res. 2002, 35, 972–983. Corbin, P. S.; Lawless, L. J.; Li, Z.; Ma, Y.;
Witmer, M. J.; Zimmerman, S. C. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5099–
5104. Pollino, J. M.; Stubbs, L. P.; Weck, M. J. Am. Chem. Soc. 2004, 126,
563–567.
(5) Balzani, V.; Credi, A.; Raymo, F. M.; Stoddart, J. F. Angew. Chem., Int.
Ed. 2000, 39, 3348–3391.
(6) Sijbesma, R. P.; Meijer, E. W. Chem. Commun. 2003, 5–16.
10.1021/jo8009424 CCC: $40.75
Published on Web 06/28/2008
2008 American Chemical Society
J. Org. Chem. 2008, 73, 5915–5925 5915

Ghosh et al.
Incontrasttobiologists,computerscientists,andengineersswho
embrace the need to study complex systems and appreciate the
insightsthatmaybegainedfromasystemslevelapproachschemists
have been comparably slow to approach the study of complex
multicomponent systems. This difference has begun to subside
in recent years with the development of powerful new analytical
tools that provide approaches to previously intractable problems.
This group of scientistssdrawn from numerous fieldssare
beginning to define an area now known as systems chemistry.7
Systems chemists study complex multicomponent systems that
exhibit emergent properties that go beyond those of their
components. Some systems chemists take their inspiration from
biology and design systems that exhibit properties such as self-
replication,8 whereas others add new components to existing
biological systems in an attempt to control biological function.
Other systems chemists take inspiration from technology and
aim to integrate individual molecular devices into more complex
molecular machines.5,9
We have based our entry into systems chemistry on the
realization that all of these biological or technological systems
depend critically on the availability of robust, easily function-
alized supramolecular modules (vide supra) that operate not only
in isolation but also as components of more complex systems.
The behavior of such systems is controlled by the network of
binary and higher order interactions between the constituent
molecules. Rather than being governed by all possible (e.g.,
random) sets of interactions, such systems tend to organize
themselves into a smaller number of sets of interacting molecules
with interconnections between these sets that respond to external
stimuli from their environment. As such, the ability of the
constituents of such systems to efficiently distinguish between
self and nonself is critical. We previously showed that a mixture
comprising the components of eight well-known aggregates from
the literature is capable of efficiently distinguishing between
self and nonself even within the mixtureson the basis of H-bond
pattern and geometrical distributionsand undergoes a high-
fidelity self-sorting process.10 Subsequently, we have developed
self-sorting systems based on host-guest interactions (social
self-sorting), those that display well-defined kinetic and ther-
modynamic self-sorted states, and shown how such systems can
be made to respond to chemical stimuli (e.g., guest addition or
pH change).11 In this paper we investigate the ability of a series
of glycoluril-derived molecular clips12–14 (1-12)sseveral of
which undergo tight dimerization (Ka g 106 M-1) in CDCl3
solutionsto act as robust, functionalizable, self-sorting su-
pramolecular modules.15,16 We envision that such supramo-
lecular modules will greatly expand the toolbox available to
systems chemists for the construction of complex and functional
systems.
Results and Discussion
This section is organized as follows. First, we present the
design and synthesis of molecular clips 1-12. Next, we detail
the dimeric aggregates formed by these molecular clips in CDCl3
solution and the solid state. Subsequently, we describe the ability
of these compounds to undergo self-sorting within two-,13 three-,
and four-component mixtures. Within these complex mixtures,
we investigate the influence of chirality, side chain steric bulk,
and electronic effects on the strength of dimerization and
ultimately the fidelity of self-sorting.
Design and Synthesis of the Chemical Components
Used in This Study. Chart 1 shows the chemical structures of
the 12 C-shaped methylene-bridged glycoluril dimers used in
this paper. The key structural features in the design of molecular
clips 1-12 were as follows. First, all 12 clips possess two
roughly parallel aromatic walls separated by ∼7 Å which define
a cleft that promotes dimerization in chloroform driven by π-π
interactions.13,14,17 Even though π-π interactions have preferred
geometries (e.g., edge-to-face or offset face-to-face),18 we
thought that dimeric molecular clips driven solely by π-π
interactions might exist as an ensemble of conformational
isomers that differ in the relative orientation of the molecular
clips with respect to each other (Figure 1). Second, the rigid
(12) For leading examples of molecular clips and tweezers, see: Rowan, A. E.;
Elemans, J. A. A. W.; Nolte, R. J. M. Acc. Chem. Res. 1999, 32, 995–1006.
Zimmerman, S. C. Top. Curr. Chem. 1993, 165, 71–102. Klaerner, F.-G.; Kahlert,
B. Acc. Chem. Res. 2003, 36, 919–932. Harmata, M. Acc. Chem. Res. 2004, 37,
862–873. Jiang, H.; Leger, J.-M.; Guionneau, P.; Huc, I. Org. Lett. 2004, 6,
2985–2988. Chiang, P.-T.; Cheng, P.-N.; Lin, C.-F.; Liu, Y.-H.; Lai, C.-C.; Peng,
S.-M.; Chiu, S.-H. Chem. Eur. J. 2006, 12, 865–876. Petitjean, A.; Khoury,
R. G.; Kryritsakas, N.; Lehn, J.-M. J. Am. Chem. Soc. 2004, 126, 6637–6647.
Artacho, J.; Nilsson, P.; Bergquist, K.-E.; Wendt, O. F.; Wa¨rnmark, K. Chem.
Eur. J. 2006, 12, 2692–2701. Hu, S.-L.; She, N.-F.; Yin, G.-D.; Guo, H.-Z.;
Wu, A.-X.; Yang, C.-L. Tetrahedron Lett. 2007, 48, 1591–1594.
(13) Wu, A.; Chakraborty, A.; Fettinger, J. C.; Flowers, R. A., II; Isaacs, L.
Angew. Chem., Int. Ed. 2002, 41, 4028–4031.
(14) Wu, A.; Mukhopadhyay, P.; Chakraborty, A.; Fettinger, J. C.; Isaacs,
L. J. Am. Chem. Soc. 2004, 126, 10035–10043.
(15) For leading examples of self-sorting, see: Rowan, S. J.; Hamilton, D. G.;
Brady, P. A.; Sanders, J. K. M. J. Am. Chem. Soc. 1997, 119, 2578–2579. Jolliffe,
K. A.; Timmerman, P.; Reinhoudt, D. N. Angew. Chem., Int. Ed. 1999, 38, 933–
937. Taylor, P. N.; Anderson, H. L. J. Am. Chem. Soc. 1999, 121, 11538–11545.
Bilgicer, B.; Xing, X.; Kumar, K. J. Am. Chem. Soc. 2001, 123, 11815–11816.
Hwang, I.-W.; Kamada, T.; Ahn, T. K.; Ko, D. M.; Nakamura, T.; Tsuda, A.;
Osuka, A.; Kim, D. J. Am. Chem. Soc. 2004, 126, 16187–16198. Burd, C.; Weck,
M. Macromolecules 2005, 38, 7225–7230. Uzun, O.; Xu, H.; Jeoung, E.; Thibault,
R. J.; Rotello, V. M. Chem. Eur. J. 2005, 11, 6916–6920. Kamada, T.; Aratani,
N.; Ikeda, T.; Shibata, N.; Higuchi, Y.; Wakamiya, A.; Yamaguchi, S.; Kim,
K. S.; Yoon, Z. S.; Kim, D.; Osuka, A. J. Am. Chem. Soc. 2006, 128, 7670–
7678. Xu, H.; Hong, R.; Lu, T.; Uzun, O.; Rotello, V. M. J. Am. Chem. Soc.
2006, 128, 3162–3163. He, Y.; Tian, Y.; Chen, Y.; Ribbe, A. E.; Mao, C. Chem.
Commun. 2007, 165–167. Legrand, Y.-M.; Van der Lee, A.; Barboiu, M. Inorg.
Chem. 2007, 46, 9540–9547. Plutowski, U.; Jester, S. S.; Lenhert, S.; Kappes,
M. M.; Richert, C. AdV. Mater. 2007, 19, 1951–1956. Hutin, M.; Cramer, C. J.;
Gagliardi, L.; Shahi, A. R. M.; Bernardinelli, G.; Cerny, R.; Nitschke, J. R.
J. Am. Chem. Soc. 2007, 129, 8774–8780. Braekers, D.; Peters, C.; Bogdan, A.;
Rudzevich, Y.; Boehmer, V.; Desreux, J. F. J. Org. Chem. 2008, 73, 701–706.
Barrett, E. S.; Dale, T. J.; Rebek, J., Jr. J. Am. Chem. Soc. 2008, 130, 2344–
2350. Shi, X.; Fettinger, J. C.; Davis, J. T. J. Am. Chem. Soc. 2001, 123, 6738–
6739. Cai, M.; Shi, X.; Sidorov, V.; Fabris, D.; Lam, Y.-F.; Davis, J. T.
Tetrahedron 2002, 58, 661–671. Schnarr, N. A.; Kennan, A. J. J. Am. Chem.
Soc. 2002, 124, 9779–9783. Hirst, A. R.; Huang, B.; Castelletto, V.; Hamley,
I. W.; Smith, D. K. Chem. Eur. J. 2007, 13, 2180–2188.
(7) Ludlow, R. F.; Otto, S. Chem. Soc. ReV. 2008, 37, 101–108. Corbett,
P. T.; Sanders, J. K. M.; Otto, S. Angew. Chem., Int. Ed. 2007, 46, 8858–8861.
Kindermann, M.; Stahl, I.; Reimold, M.; Pankau Wolf, M.; von Kiedrowski, G.
Angew. Chem., Int. Ed. 2005, 44, 6750–5. Sarma, R. J.; Nitschke, J. R. Angew.
Chem., Int. Ed. 2008, 47, 377–380. Nitschke, J. R. Acc. Chem. Res. 2007, 40,
103–112.
(8) Wintner, E. A.; Conn, M. M.; Rebek, J. J. Acc. Chem. Res. 1994, 27,
198–203. Bachmann, P. A.; Luisi, P. L.; Lang, J. Nature 1992, 357, 57–9. Terfort,
A.; Von Kiedrowski, G. Angew. Chem., Int. Ed. Engl. 1992, 31, 654–656. Lee,
D. H.; Severin, K.; Yokobayashi, Y.; Ghadiri, M. R. Nature 1997, 390, 591–
594. Sievers, D.; von Kiedrowski, G. Chem. Eur. J. 1998, 4, 629–641. Yao, S.;
Ghosh, I.; Zutshi, R.; Chmielewski, J. Nature 1998, 396, 447–450. Issac, R.;
Chmielewski, J. J. Am. Chem. Soc. 2002, 124, 6808–6809. Kassianidis, E.; Philp,
D. Angew. Chem., Int. Ed. 2006, 45, 6344–6348.
(9) Kay, E. R.; Leigh, D. A.; Zerbetto, F. Angew. Chem., Int. Ed. 2007, 46,
72–191.
(10) Wu, A.; Isaacs, L. J. Am. Chem. Soc. 2003, 125, 4831–4835.
(11) Mukhopadhyay, P.; Wu, A.; Isaacs, L. J. Org. Chem. 2004, 69, 6157–
6164. Liu, S.; Ruspic, C.; Mukhopadhyay, P.; Chakrabarti, S.; Zavalij, P. Y.;
Isaacs, L. J. Am. Chem. Soc. 2005, 127, 15959–15967. Mukhopadhyay, P.;
Zavalij, P. Y.; Isaacs, L. J. Am. Chem. Soc. 2006, 128, 14093–14102. Liu, S.;
Zavalij, P. Y.; Lam, Y.-F.; Isaacs, L. J. Am. Chem. Soc. 2007, 129, 11232–
11241. Chakrabarti, S.; Mukhopadhyay, P.; Lin, S.; Isaacs, L. Org. Lett. 2007,
9, 2349–2352.
(16) Saur, I.; Scopelliti, R.; Severin, K. Chem. Eur. J. 2006, 12, 1058–1066.
(17) Isaacs, L.; Witt, D. Angew. Chem., Int. Ed. 2002, 41, 1905–1907.
(18) Meyer, E. A.; Castellano, R. K.; Diederich, F. Angew. Chem., Int. Ed.
2003, 42, 1210–1250.
5916 J. Org. Chem. Vol. 73, No. 15, 2008

Self-Sorting Molecular Clips
CHART 1. Chemical Structures of Molecular Clips Used in
This Paper
FIGURE 1. Schematic representation of three members of the con-
formational ensemble open to dimeric molecular clips driven solely
by π-π interactions: (a) in register geometry; (b) in skewed geometry;
(c) in perpendicular geometry.
(()-17. For the synthesis of (()-17, we first performed the free
radical bromination of 4-nitro-o-xylene using N-bromosuccin-
imide in 1,2-dichloroethane, which gave 15 in 47% yield.20
Alkylation of 16 with compound 15 (t-BuOK, DMSO) gave
(()-17 in 22% yield. Compound (()-17 was dimerized by acid-
catalyzed condensation with paraformaldehyde in the presence
of p-toluenesulfonic acid (PTSA) in ClCH2CH2Cl to give 18
and (()-19 in 23% and 20% yields, respectively.19 After
chromatographic separation, regioisomers 18 and (()-19 were
separately reduced and then acylated with benzoyl chloride to
give 8 and (() 11 in 73% and 76% yields, respectively.
Compounds 9, 10, (()-12, and (()-7 were synthesized by
related procedures. Somewhat surprisingly, we found that
heating 20 with PTSA in p-xylene gave 21 by extrusion of one
of the CH2OCH2 bridges in 36% yield. Acid-catalyzed conden-
sation of compound 21 with paraformaldehyde gave dimer 22
in 78% yield. Compound 22 was nitrated (TFA, HNO3) to give
a mixture of regioisomers 23 and (()-24 in 42% and 38% yields,
respectively. After chromatographic separation, regioisomers 23
and (()-24 were separately reduced and then acylated with the
appropriate acid chloride to give 9, 10, and (()-12 in 65-75%
yields. Condensation of (()-25 with cyclic ether (()-2621 in
presence of PTSA in ClCH2CH2Cl gave a mixture of two dinitro
geometry of the molecular clips and the straightforward func-
tionalization of their o-xylylene sidewalls was utilized to display
H-bonding functional groups (e.g., amide, urea, oxamide) in
well-defined relative orientations. We envisioned that interac-
tions between these H-bonding arms could be used to drive
dimerization in CDCl3 solution and also result in the selection
of a single well-defined conformation from the conformational
ensemble (Figure 1) that satisfied the geometric requirements
of both the H-bonding and π-π interactions.14
The synthesis of compounds 1-12 is shown in Scheme 1.
For the preparation of compounds 1-6, we used the previously
reporteddinitrocompounds13and(()-14asstartingmaterials.17,19
The separate catalytic hydrogenation of 13 and (()-14 to the
corresponding air-sensitive diamines proceeded smoothly in
dimethylformamide (DMF) as solvent. The crude diamines were
transformed into the corresponding amides and ureas by reaction
with acid chlorides and isocyanates in CH2Cl2 at room tem-
perature to deliver 1-6 in 43-90% yield. To prepare com-
pounds 8 and 11, which contain amide substituents at the
4-position of the o-xylylene sidewalls, we needed to prepare
(20) Wang, Z.-G.; Zhou, B.-H.; Chen, Y.-F.; Yin, G.-D.; Li, Y.-T.; Wu, A.-
X.; Isaacs, L. J. Org. Chem. 2006, 71, 4502–4508.
(21) Qin, S.; Yin, G. Acta Crystallogr., Sect. E: Struct. Rep. Online 2005,
E61, o3164–o3165.
(19) Wu, A.; Chakraborty, A.; Witt, D.; Lagona, J.; Damkaci, F.; Ofori, M. A.;
Chiles, J. K.; Fettinger, J. C.; Isaacs, L. J. Org. Chem. 2002, 67, 5817–5830.
J. Org. Chem. Vol. 73, No. 15, 2008 5917

Ghosh et al.
SCHEME 1. Synthesis of Molecular Clips 1-12a
a Conditions: (a) Pd/C, H2, DMF; (b) PhNCO or R′COCl, Et3N, CH2Cl2; (c) 15, t-BuOK, DMSO; (d) PTSA, (CH2O)n, ClCH2CH2Cl, reflux; (e) PTSA,
p-xylene, reflux; (f) TFA, HNO3; (g) PTSA, ClCH2CH2Cl, reflux.
compounds. Chromatographic purification gave (()-27 in low
yield (13%), which was reduced (Pd/C, H2, DMF) to the
corresponding diamine followed by reaction with phenyl iso-
cyanate to give (()-7 in 39% yield.
1
Molecular Clips Form Homodimers in CDCl3. The H
NMR spectra of 1, 2, (()-4a, and (()-5 in DMSO-d6 show a
single set of resonances due to the presence of symmetry-
equivalent halves, which indicates that these molecules are
monomeric in DMSO-d6.22 In CDCl3, however, we observe a
1
doubling of resonances in their H NMR spectra along with
significant anisotropic effects, which suggests that these mol-
ecules exist as dimers (Figure 2). For example, although only
two N-H resonances (Ha and Hb) were observed in DMSO-d6
for 2, we observed four resonances in CDCl3 (Ha, Ha′, Hb, and
Hb′), which correspond to the two different N-H· · ·O H-bonds
in homodimer 2·2. Similarly, in the monomeric form the
aromatic protons of 2 (Hc and Hd) give rise to a single set of
resonances due to the presence of a mirror plane in the molecule.
1
When 2 undergoes dimerization to yield 2·2, this mirror plane
no longer exists. One set of aromatic rings resides in the interior
of the dimer, whereas the other set is on the exterior. The
FIGURE 2. Portions of the H NMR (500 MHz, room temperature)
spectra recorded for (a) 2 in DMSO-d6, (b) 2·2 in CDCl3, (c) (()-5 in
DMSO-d6, and (d) (+)-5·(-)-5 in CDCl3.
chemical shift of the aromatic protons (Hc′ and Hd′) on the
internal aromatic wall of 2·2 resonate upfield relative to the
protons of the external aromatic ring (Hc and Hd) because they
(22) We did not observe any changes in chemical shifts over the 10-0.1
mM concentration range. This rules out the possibility of a fast monomer-
dimer exchange as the cause of a single set of resonances.
5918 J. Org. Chem. Vol. 73, No. 15, 2008

Self-Sorting Molecular Clips
FIGURE 4. Portions of the 1H NMR (2.5 mM, 500 MHz, CDCl3, room
temperature) spectra recorded for (a) 1·1, (b) 2·2, and (c) a mixture
of 1·1, 2·2, and (()-1·2. Resonances for the amide N-H groups are
color coded as follows: 1·1, red; 2·2, green; (()-1·2, blue.
mixture of 1 and 2 with common spatial distribution of their
H-bonding arms was prepared in CDCl3, we observed a roughly
statistical mixture of homodimers 1·1 and 2·2 and heterodimer
1
(()-1·2 by H NMR (Figure 4). Similarly, when we analyzed
1
a mixture of (()-4a and (()-5 by H NMR, a mixture of two
FIGURE 3. Schematic representation of the concept of geometrical
match versus geometrical mismatch: (a) 1·1; (b) (+)-4a·(-)-4a; (c)
hypothetical diastereomeric aggregate formed by (()-1·4a; (d) racemic
heterodimer (()-1·2.
heterochiral homodimers ((+)-4a·(-)-4a and (+)-5·(-)-5) and
a racemic mixture of heterochiral heterodimers ((+)-4a·(-)-5
and (-)-4a·(+)-5) were observed (Supporting Information). On
the basis of this result, we hypothesized that subtle differences
in the spatial orientation of H-bonding groups on the edges of
the o-xylylene rings of these molecular clips might endow these
compounds with the ability to efficiently distinguish between
self and nonself within binary and higher order mixtures. As
such, we anticipated that this class of molecular clips based on
methylene-bridged glycoluril dimers might be robust, function-
alizable, and self-sorting (e.g., orthogonal) components for the
construction of complex chemical systems.
are in the shielding region of two neighboring aromatic rings.
1
Similar diagnostic features were observed in the H NMR of
1·1, (+)-4a·(-)-4a, and (+)-5·(-)-5 in CDCl3.
1
We performed H NMR dilution experiments to determine
the self-association constant (Ks) for the dimeric aggregation.
Remarkably, the chemical shifts observed for dimers 1·1, 2·2,
(+)-4a·(-)-4a, and (+)-5·(-)-5 do not change when the
concentration is decreased from 10 mM and 50 µM, indicating
their high thermodynamic stability. If we assume that we could
detect changes in the chemical shifts due to 10% monomer,
then we can place a lower limit on the of the value of Ks (Ks >
9 × 105 M-1, ∆G < -8.1 kcal mol-1). Such a high thermo-
dynamic stability was not anticipated, given that these dimeric
aggregates benefit from a mere two H-bonds. We surmise that
the high thermodynamic stability of these aggregates is due to
the cumulative effect of π-π interactions and H-bonds. The
concentration invariance and distinct pattern of anisotropic
effects in the 1H NMR spectra of dimers 1·1, 2·2, (+)-4a·(-)-
4a, and (+)-5·(-)-5 indicate that the combined driving force
and geometrical constraints of π-π interactions and H-bonds
are capable of selecting a single member of the complex
conformational ensemble open to dimeric molecular clips
(Figure 1).
X-ray Crystal Structures of 1, 2, (()-4b, and (()-5.
Although the 1H NMR experiments mentioned in the previous
section clearly support the formation of dimeric aggregates,
additional evidence of dimer formation was obtained from the
X-ray crystal structures of 1, 2, (()-4b, and (()-5. Interestingly,
all four molecular clips form dimers in the solid state. All four
dimerss1 · 1, 2 · 2, (+)-4b · (-)-4b, and (+)-5 · (-)-5sbenefit
from π-π interactions between the substituted o-xylylene walls.
All of them also benefit from two H-bonds between the external
amide (urea) N-H groups and internal amide (urea) CdO
groups and secondary electrostatic interactions. Several other
features of the geometries of these dimers in the solid state are
noteworthy. For example, the pendant benzoyl groups in dimers
1·1 and 2·2 are displayed in a nearly collinear orientation on
a single face of the dimer (Figure 5a,b). An additional level of
complexity is present for the case of chiral but racemic
molecular clips (()-4b and (()-5, where both homochiral
dimerization (e.g., (+)-4b·(+)-4b and (-)-4b·(-)-4b) and
heterochiral dimerization (e.g., (+)-4b·(-)-4b) are conceiv-
able.23 In the crystal we only observe the formation of
heterochiral dimers (Figure 5c,d). We attribute this result to the
geometrical mismatch between the H-bonding groups in the
hypothetical homochiral homodimer. All these observations in
the solid state determined by X-ray crystallography are consis-
Concept of Geometrical Match and Mismatch. As de-
scribed above, compounds 1, 2, (()-4a, and (()-5 undergo
strong homodimerization in CDCl3 driven by π-π interactions
and H-bonds. We previously investigated their recognition
behavior in binary mixtures of 1 and (()-4a or 2 and (()-5
and observed the phenomenon of self-sorting.13 We attribute
this high preference for self-sortingsheterodimers ((()-1·4a
1
or (()-2·5) were not observed by H NMRsto a geometrical
mismatch between the H-bonding arms 1 and (()-4a or 2 and
(()-5 in the hypothetical heterodimers (Figure 3). The hypo-
thetical heterodimer (()-1·4a, for example, benefits from π-π
interactions and only a single H-bond, as opposed to π-π
interactions and at least two H-bonds in homodimers 1·1 and
(+)-4a·(-)-4a (Figure 3). Conversely, when an equimolar
(23) Rudi, A.; Kopilov, J.; Goldberg, I.; Kol, M. J. Am. Chem. Soc. 2002,
124, 5449–5456. Marguly, E.; McDonald, R.; Branda, N. R. Org. Lett. 2000, 2,
3169–3172. Maeda, C.; Kamada, T.; Aratani, N.; Osuka, A. Coord. Chem. ReV.
2007, 251, 2743–2752. Telfer, S. G.; Sato, T.; Kuroda, R.; Lefebvre, J.; Leznoff,
D. B. Inorg. Chem. 2004, 43, 421–429.
J. Org. Chem. Vol. 73, No. 15, 2008 5919

Ghosh et al.
FIGURE 5. Cross-eyed stereoviews of the molecular structures of (a) 1·1, (b) 2·2, (c) (+)-4b·(-)-4b, and (d) (+)-5·(-)-5 in the crystal. Color
coding: C, gray; H, white; N, blue; O, red; H-bonds, yellow-red striped.
tent with their 1H NMR spectra in CDCl3, and we suggest that
1, 2, (()-4a, and (()-5 are isostructural in the solid state and
in solution.14 The X-ray crystal structures of 1, 2, (()-4b, and
(()-5 also helps us to rationalize the behavior of these molecules
within binary and higher order mixtures (vide infra).
Design of Molecular Clips for Three- and Four-Com-
ponent Self-Sorting Systems. The high selectivity we observed
during the self-recognition of 2 and (()-5 in a binary mixture
of molecular clips and the lack of self-selectivity for mixtures
of 1 and 2 or (()-4a and (()-5 inspired us to consider situations
that fall between these two extreme cases guided by a series of
questions. What happens to the strength of dimeric aggregation
and fidelity of self-sorting when the H-bonding groups are
moved around the edges of the substituted o-xylylene walls?
Will more complex mixtures of molecular clips (e.g., three or
four components) constitute self-sorting systems if each clip
has a distinct spatial arrangement of its H-bonding arms? How
efficiently can a chiral molecular clip distinguish between itself,
1
FIGURE 6. Portion of the H NMR (500 MHz, room temperature)
spectrum recorded for (a) 8 in DMSO-d6, (b) 8·8 (10 mM in CDCl3),
and (c) 8·8 (50 µM in CDCl3). × ) 13CHCl3.
its enantiomer, and achiral molecular clips? We were also
interested to investigate the effect of various structural modifica-
tions (e.g., substituents, steric bulk, number and pattern of
H-bonds) on self-assembly and self- versus nonself recognition
behavior in a multicomponent mixture that would impact their
use as components for advanced applications (e.g., stimuli
responsive supramolecular polymers) and the topology of their
network interactions16 when utilized as parts of more complex
(bio)molecular systems. To address these questions, we syn-
thesized compounds 3, 6, and 7-12 and studied their self-
association and self-sorting properties in CDCl3 solution.
Three-Component Self-Sorting System. As a first attempt
to assess the ability of more subtle geometrical changes to direct
self-sorting processes, we prepared 8, which as opposed to 1-6
has PhCONH groups on the tips of its o-xylylene sidewalls.
Similar to the cases for 1, 2, (()-4a and (()-5, compound 8
Interestingly, the dimeric aggregate 8·8 does not undergo
significant dissociation down to 50 µM, which establishes a high
thermodynamic stability (Ka > 106 M-1) for this dimer. We
were fortunate to obtain an X-ray crystal structure of 8·8, which
shows a number of interesting features (Figure 7a). In the
dimeric aggregate 8·8, the clips are skewed relative to one
another (see Figure 1b) in such a way that the aggregate benefits
from π-π interactions as well as four H-bonds (Figure 8). The
amide N-H groups on the internal aromatic rings form H-bonds
with the ureidyl CdO of a glycoluril ring on the opposing clip,
whereas the external amide N-H groups form H-bonds with
the internal amide CdO groups. On the basis of the distinct
geometries observed for 2 · 2, (+)-5 · (-)-5, and 8 · 8, we
wondered whether a mixture of these three molecular clips
would undergo high-fidelity self-sorting.
The high level of enantioselectivity exhibited by (+)-5·(-)-5
and skewing observed in the X-ray structure of 8·8 led us to
combine these clips with 2·2 for the next set of experiments.
Initially we examined the behavior of the three binary mixtures
1
exhibits a single set of resonances in its H NMR spectrum in
DMSO-d6sindicative of a monomer sbut two sets of reso-
nances in CDCl3 indicative of dimer formation (Figure 6).
5920 J. Org. Chem. Vol. 73, No. 15, 2008

Self-Sorting Molecular Clips
FIGURE 9. 1H NMR spectra (500 MHz, 298 K) for (a) 2·2, (b) (+)-
5·(-)-5, (c) 8·8, and (d) a mixture of 2·2, (+)-5·(-)-5, and 8·8.
FIGURE 7. Cross-eyed stereoviews of the molecular structures of (a)
8·8 and (b) (+)-7·(+)-7 in the crystal. Color coding: C, gray; H, white;
N, blue; O, red; H-bonds, yellow-red striped.
FIGURE 8. Schematic representations of the shapes of 8·8: (a) steric
interaction in the head-to-head form; (b) skewed geometry that avoids
steric interactions.
(e.g., 2·2 and (+)-5·(-)-5 or 2·2 and 8·8 or (+)-5·(-)-5 and
8·8) and were delighted to observe a simple superimposition
of the 1H NMR spectra of the constituent homodimeric
aggregates (Supporting Information) in each case. These results
indicate that these three molecular clips possess the ability to
efficiently distinguish between self and nonself. Accordingly,
we next constructed the three-component mixture comprising
FIGURE 10. 1H NMR spectra (500 MHz, 298 K) for (a) (+)-7·(+)-
7, (b) (+)-5·(-)-5, (c) 1·1, and (d) a mixture of 1·1, (+)-5·(-)-5,
and (+)-7·(+)-7.
1
ment. The H NMR spectrum of (()-7 in CDCl3 shows two
sets of resonances corresponding to a clear dimeric aggregate
(Figure 10a). As predicted on the basis of CPK models, the
X-ray structure reveals that a racemic mixture of homochiral
homodimers (+)-7·(+)-7 and (-)-7·(-)-7 (Figure 7b) is formed
and adopts a skewed orientation to maximize the π-π interac-
tions and form four H-bonds.
1
2, (()-5, and 8 and observed an H NMR spectrum consistent
with the formation of a self-sorted mixture of the three
homodimeric aggregates 2·2, (+)-5·(-)-5, and 8·8 (Figure 9).
These results demonstrate that simply by choosing molecular
clips that possess different geometrical arrangements of their
H-bonding arms it is possible to achieve high-fidelity self-sorting
even within a complex mixture. We expect that such mutually
orthogonal H-bonding modules have broad potential as triggered
components of more complex molecular machines.5,9
Inspired by the highly enantioseletive recognition behavior
imprinted in the chiral backbone of (()-7, we decided to study
the behavior of this new molecular clip in the presence of other
molecular clips. We prepared various binary mixtures of (()-7
with achiral and chiral molecular clips and observed the outcome
Designed Homochiral Dimerization and
a Second
1
Three-Component Self-Sorting System. Given our success in
the preparation of a three-component self-sorting system based
on 2·2, (+)-5·(-)-5, and 8·8, we wondered whether we could
expand the scope of the system even further. For this purpose
we designed (()-7, which is a hybrid molecule whose halves
resemble 2 and 8, respectively. Before synthesizing (()-7, we
examined CPK models of 7·7 which suggested that a skewed
geometryssimilar to 8·8sbased on homochiral dimerization
would retain π-π interactions and maximize the number of
H-bonding interactions. This prediction is borne out by experi-
of the resulting mixture by H NMR spectroscopy. When we
mixed equimolar amounts of (()-7 and achiral 1 or 2, we found
that they form a binary self-sorting mixture comprising (+)-
7·(+)-7 and (-)-7·(-)-7, and 1·1 or 2·2. This result demon-
strates that the cleft of chiral (()-7 is not complementary to
either 1 or 2, due to a geometric mismatch between their
H-bonding arms. Similarly, in an equimolar mixture of (()-7
with (()-4a or (()-5, we observed the coexistence of homo-
chiral ((+)-7·(+)-7 and (-)-7·(-)-7) and heterochiral ((+)-
4a·(-)-4a or (+)-5·(-)-5) dimeric molecular clipssa result
J. Org. Chem. Vol. 73, No. 15, 2008 5921

Ghosh et al.
1
FIGURE 11. Portiona of the 1H NMR (1 mM, 500 MHz, CDCl3, room
temperature) spectra recorded for (a) a racemic mixture of (+)-7·(+)-7
and (-)-7·(-)-7, (b) 8·8, and (c) a mixture of (+)-7·(+)-7 and (-)-
7·(-)-7, 8·8, (+)-7·8, and (-)-7·8.
FIGURE 12. Portions of the H NMR spectra recorded for 9·9 (400
MHz, CDCl3, room temperature) upon dilution: (a) 11 mM; (b) 1.1
mM; (c) 0.11 mM. Resonances are color-coded as follows: dimeric
9·9, red; monomeric 9, blue.
that demonstrates self-sorting based on chirality (Supporting
Information). In accord with these observations, when we mixed
1, (()-5, and (()-7 we observed an NMR spectrum that was
the sum of those measured for 1·1, heterochiral (+)-5·(-)-5,
and a racemic mixture of homochiral (+)-7 · (+)-7 and (-)-
7·(-)-7 dimers (Figure 10), which indicates a high-fidelity self-
sorting event.
Although we were delighted to observe such high levels of
homochiral and heterochiral recognition within the three-
component self-sorting systems, we wanted to extend the system
toward larger numbers of simultaneously self-sorting molecular
clips. When an equimolar mixture of (()-7 and 8 was prepared,
we observed a mixture of homodimers and heterodimers by 1H
NMR (Figure 11). The lack of self-selectivity in mixture of
(()-7 and 8 can be rationalized from their X-ray crystal
structures. Both molecular clips form dimeric aggregates in
which the opposing clips are skewed with respect to each other.
Similarities in the spatial distribution of the H-bonding arm of
molecular clips (()-7 and 8 reduce their ability to discriminate
between self and nonself within this binary mixture.
formation of 1 · 2 (Figure S14, Supporting Information). We
envision that the ability to selectively alter the interaction
network of one component (or one subset of components) of a
complex self-sorting mixture by addition of a new component
will enable the construction of functional stimuli responsive
systems.
Steric Effects Influence the Thermodynamics and
Kinetics of Dimerization. In the next phase of our search for
self-sorting systems comprising larger numbers of molecular
clips, we decided to introduce steric congestion onto the
o-xylylene sidewalls. We hypothesized that increased steric
interactions between two molecular clips would destabilize their
homodimers more dramatically than their heterodimers. This
hypothesis is supported by simulations reported previously14
which show that the system comprising one tight and one weak
homodimer is thermodynamically more favorable than the
alternative arrangement featuring two heterodimers of modest
strength. Therefore, we anticipated that a mixture of two
homodimerssonestericallyhinderedandoneunsubstitutedsmight
constitute a self-sorting system.
Selective Dissociation of One Member of a Self-Sorted
Mixture. The heterodimerization of (()-7 and 8 was disap-
pointing, in that we were precluded from preparing a four-
component self-sorting system comprising 2, 5, (()-7, and 8.
When considered from a different viewpoint, it enables a useful
type of stimuli responsiveness that will be particularly useful
in the engineering of complex and functional chemical systems.
Consider, for example, a system comprising four members
(A-D). Of the 10 homodimeric and heterodimeric species (AA,
BB, CC, DD, AB, AC, AD, BC, BD, CD) that would comprise
an unbiased dynamic combinatorial library, a high-fidelity self-
sorting system is only able to access the four homodimeric states
(e.g., AA, BB, CC, DD). The ability to selectively access
heterodimeric states (e.g., BD) by suitable chemical stimuli (e.g.,
addition of molecular clip D) would allow fine tuning of the
topology of the interaction network that characterizes this four-
component mixture.16 In this context, we examined the addition
of a fourth component to pre-existing three-component self-
sorting systems. As expected, addition of a racemic mixture of
(+)-7·(+)-7 and (-)-7·(-)-7 to a self-sorted mixture compris-
ing 2·2, (+)-5·(-)-5, and 8·8 results in the selective het-
erodimerization of 8·8 under the formation of (()-7·8 (Figure
S13, Supporting Information). Similarly, the addition of 1·1 to
a self-sorting mixture comprising 2·2, (+)-5·(-)-5, and 8·8
results in the selective heterodimerization of 2·2 under the
We choose compound 8 as the base systemsbecause it forms
a tight dimer and we thought it would be straightforward to
modify syntheticallysand gradually increased the number of
substituents in order to test the influence of steric congestion
on dimeric aggregation. Accordingly, we prepared tetramethyl
benzanilide (9) and tetramethyl pivalanilide (10). When a 2.5
mM solution of 9 in CDCl3 was examined by 1H NMR
spectroscopy at room temperature, we observed two sets of
peaks. After warming the same solution gradually to 55 °C, we
observed a decrease in the intensity of one set of peaks
concomitant with an increase in the intensity of the other set
(Supporting Information). An indistinguishable spectrum was
obtained when this solution was cooled slowly to room
temperature. A similar change in the system was observed when
an 11 mM solution of 9·9 was diluted 100-fold to 0.11 mM
(Figure 12). These observations suggest that 9 exists in
equilibrium with 9·9 with slow exchange kinetics relative to
the chemical shift time scale at room temperature. Apparently,
steric interactions between the methyl substituents on the
aromatic walls of 9·9 thermodynamically destabilize the dimer.
Interestingly, however, the kinetics of exchange (e.g., koff)
remain slow (as observed for 8 · 8) even as Ka is decreased.
Conveniently, the observed slow exchange between monomer
and dimer on the chemical shift time scale allows us to follow
1
this equilibrium by H NMR spectroscopy and calculate the
5922 J. Org. Chem. Vol. 73, No. 15, 2008

Self-Sorting Molecular Clips
FIGURE 13. Molecular clips that remain monomeric in solution: (a)
steric bulk preventing dimeric aggregation of 10; (b) crystal structure
of (()-12 showing the display of H-bonding groups in front of the
cleft, which leads to steric hindrance to dimerization.
association constant by measuring the integrals for monomer
versus dimer at various concentrations (Ks ) 2900 ( 600
mol-1). We incorporated additional steric congestion in 10 in
the form of t-Bu(CdO)NH- H-bonding arms (Figure 13a). In
this case, the chemical shifts of the aromatic as well as the amide
protons do not change in the dilution experiment, which suggests
that 10 exists as a monomer (Supporting Information). Interest-
ingly, when we constructed an equimolar binary mixture of 10
and 9, an NMR spectrum was a superposition of the spectra of
the components. The presence of severe steric congestion around
the cleft of the hypothetical homodimer 10·10 and heterodimer
(()-9·10 dictates that compound 10 must remain monomeric
(Supporting Information).
FIGURE 14. Schematic representation of the aggregates formed by
(a) 1·1, (b) 3·3, and (c) (()-6.
and chirality on the ability of molecular clips to undergo self-
assembly and self-sorting, we decided to investigate the influ-
ence of the number and pattern of H-bonding donors and
acceptors in the pendant arm on the fidelity of self-sorting. We
synthesized 3, which contains additional NH(CO)CH3 H-
bonding groups in the para position of the pendant aromatic
rings. We thought that 3 might form a very robust aggregate,
due to an increased number of H-bonding interactions upon
dimerization. We further predicted that 3 might undergo self-
sorting with 1 or 2 on the basis of differences in the number of
H-bonds in the aggregate (Figure 14a,b). When 3 was dissolved
in CDCl3, we observed upfield shifting, severe broadening, and
multiple resonances for some protons, which indicates the
formation of multiple ill-defined species in solution. Upon
further examination of a CPK model of 3, we realized that the
flexibility of the pendant arm of 3 increases as the arm gets
longer. Due to this lack of rigidity, unsatisfied H-bond donors
in the amide group farthest from the clip can interact with
unsatisfied H-bond acceptors on other molecules of 3 or
assemblies of 3, resulting in uncontrolled higher order aggrega-
tion.24
Compounds 2 and (()-5 contain an H-bond donor-acceptor-
donor arrangement in the ureidyl group. We hypothesized that
a molecular clip with a specific donor-acceptor arrangement
in the pendant H-bonding arms might undergo self-sorting with
another clip that possessed a different donor-acceptor arrange-
ment. In order to test that hypothesis, we synthesized compound
(()-6, containing H-bonding groups having a donor-acceptor-
acceptor pattern. When (()-6 was dissolved in CDCl3, we
observed an upfield shift of the proton attached to its o-xylylene
sidewall, indicating that (()-6 forms an aggregate in solution.
As described earlier, the relative orientation of H-bonding
sidearms in Cs- or C2-symmetrical molecular clips (e.g., parallel
in 1 and 2 versus antiparallel in (()-4a and (()-5) endow them
with the ability to efficiently distinguish between self versus
nonself. Accordingly, we prepared (()-11 and (()-12, which
are the C2-symmetric regioisomers of Cs-symmetric 8 and 9.
Interestingly, although not unexpectedly, we did not find any
1
evidence of discrete dimer formation in the H NMR spectra
of (()-11 and (()-12. In the molecular structures of (()-11
and (()-12 the H-bonding pendant arms are displayed in
divergent directions on the tips of the aromatic rings and,
therefore, block the face of the cleft, create steric hindrance to
dimerization, and prevent association with a second molecular
clip. The molecular clips (()-11 and (()-12 do not form well-
defined discrete dimers such as those formed by 8 or 9. Rather,
1
we observe broadening and upfield shifting in their H NMR
spectra, which indicates that (()-11 and (()-12 form poorly
structured oligomeric species in solution (Supporting Informa-
tion). The solid-state behavior is similarly affected; unlike 8,
which exists as a dimer, compound (()-12 remains monomeric
in the crystal (Figure 13b). These resultsswhich show that both
π-π and H-bonds are required to form discrete dimeric
aggregatessfurther validate the use of molecular clips as a
platform for the construction of complex self-sorting systems.
For example, the transition from monomer to discrete dimer to
poorly defined oligomeric states can be programmed into the
system by the identity and spatial orientation of the substituents
on the o-xylylene sidewalls of the constituent molecular clips.
Effect of Additional H-Bonding Groups on the Ability
of Molecular Clips To Undergo Self-Assembly and
Self-Sorting. After exploring the influence of geometry, sterics,
(24) Interestingly, examination of the X-ray crystal structures of 1 and 2
shows that the two internal amide protons remain free in the dimeric aggregate.
Upon careful inspection, we conclude that these amide protons are actually
involved in weak electrostatic interactions with the ureidyl CdO group and also
that they are not accessible to any external H-bond acceptors because they are
buried inside the molecular architecture.
J. Org. Chem. Vol. 73, No. 15, 2008 5923

Ghosh et al.
FIGURE 15. Portions of the 1H NMR (2 mM, 500 MHz, CDCl3, room
temperatue) spectrum recorded for (a) (+)-5·(-)-5, (b) (()-6, and (c)
a mixture of (+)-5·(-)-5, and (()-6.
Unlike (+)-4a·(-)-4a and (+)-5·(-)-5, we did not observe any
doubling of resonances due to aggregation, as the molecule
undergoes fast exchange on the chemical shift time scale.
Interestingly, when a solution of (()-6 was diluted, we observed
a downfield shift of the amide protons (Supporting Information).
Usually the amide protons move upfield in dilution experiments
as the H-bonds are disrupted with increasing dilution. This result
indicates the presence of two strong intramolecular H-bonds in
monomer (()-6 which are sacrificed to form two weak
intermolecular H-bonds and π-π interactions during dimeriza-
tion. Accordingly, compound (()-6 forms a weak dimer (Ks )
17.5 ( 2.2 mol-1) in CDCl3 solution (Figure 14c). When we
mixed (()-6 with an equimolar amount of (()-5, we mainly
observed the separate coexistence of (()-6 and (+)-5·(-)-5 in
the solution with a very small amount of heteromeric assemblies
(Figure 15). This observation suggests that the tendency of
molecular clips with a common geometrical arrangement of their
H-bonding arms toward formation of mixtures of homo- and
heterodimers can be suppressed with concomitant enhancement
of self-selectivity by incorporating an electronic mismatch
between their H-bonding arms.
Four-Component Self-Sorting System. If we consider that
only monomeric and dimeric aggregates can exist in solution,
then a mixture of 4 molecular clips (A-D) can give rise to 14
different species (A, B, C, D, A2-D2, AB, AC, AD, BC, BD,
CD). In our quest for a four-component self-sorting systemsfor
use as orthogonal supramolecular modules for studies of systems
chemistryswe screened a large number of different combina-
tions guided by the principles discussed above. Initially, we
examined an equimolar mixture of 2, (()-5, and 9 and observed
a self-sorted mixture by 1H NMR spectroscopy due to the
disparate spatial orientation of their H-bonding arms. On the
basis of the fact that 9 and 10 constitute a self-sorting system
(vide supra), we decided to prepare a mixture comprising
molecular clips 2, (()-5, 9, and 10. Remarkably, this set of 4
molecular clips exists in a self-sorted state (Figure 16) compris-
ing 5 different speciess2·2, (+)-5·(-)-5, 9 in equilibrium with
9·9, and monomeric 10sout of 14 possibilities.
FIGURE 16. 1H NMR spectra (400 MHz, 298 K) for (a) 2·2, (b) (+)-
5·(-)-5, (c) 9·9, (d) 10, and (e) a mixture of 2·2, (+)-5·(-)-5, 9·9,
and 10.
around the o-xylylene rings. For example, 1·1 and 2·2 form
homodimers and (+)-4a·(-)-4a and (+)-5·(-)-5 undergo
heterochiral recognition, both with an in register type geometry.
When H-bonding groups were moved to the tips of the aromatic
rings (e.g., 8·8 and 9·9), a skewed geometry was observed that
maximizes H-bonding and π-π interactions. Most interesting
is (()-7, which forms the homochiral homodimers (+)-7·(+)-7
and (-)-7·(-)-7 based on a skewed geometry. Steric bulkiness
around the cleft of these molecular clips (e.g., 9 and 10) leads
to reduced values of Ka, but slow exchange kinetics relative to
the 1H NMR time scale are maintained. The high levels of self-
selectivity of these molecular clips allowed us to prepare three-
component (2, (()-5, and 8; 1, (()-5, and (()-7) and even a
four-component self-sorted system (2, (()-5, 9, and 10).
The implications of this research go beyond the system-
specific considerations described above. The availability of a
series of robust, easily functionalized, and orthogonal H-bonding
modules for assembly in nonpolar solvents such as CDCl3 can
be utilized for numerous applications. For example, the pat-
terning of surfaces using two orthogonal recognition units has
recently been demonstrated; additional levels of patterning could
be added along with stimuli control based on molecular clips
reported herein. Such modules also enable the noncovalent
derivatization of polymer backbones in solution to optimize their
properties for specific applications (e.g., light emitting diodes,
drug delivery, tissue engineering).3 Although the molecular clips
described herein assemble in nonpolar solvents, water-soluble
versions of these compounds also assemble in water. When such
water-soluble molecular clips also exhibit self-sorting, it would
be possible to use them as tags to promote dimerization of
appended (bio)molecules within complex cellular environments.
Perhaps most significantly, the development of dynamic
combinatorial chemistry and self-sorting systems has helped
stimulate the development of systems chemistry. In unbiased
dynamic combinatorial libraries all constituent states are equally
populated and application of a chemical stimulus leads to
enhanced formation of one or more members of the library due
Conclusions
In summary, we presented the synthesis of methylene-bridged
glycoluril dimers 1-12, which undergo strong dimerization in
CDCl3 due to a combination of H-bonding and π-π interactions.
The structure of these dimeric aggregates changes as the position
and relative orientation of the two H-bonding groups move
5924 J. Org. Chem. Vol. 73, No. 15, 2008

Self-Sorting Molecular Clips
spectra were recorded at a series of concentrations (10-0.05 mM).
Spectra for dimeric clips at higher than 10 mM concentration were
recorded in 12 mm microtubes matched with CDCl3. Spectra were
referenced relative to residual solvent resonances.
X-ray Crystal Structures for 2 ·2, (+)-5·(-)-5, (+)-7·
(+)-7, 8 ·8, and 12. Descriptions of the data collection, solution,
and refinement of these structures can be found in the Supporting
Information. The crystal structures of 1·1 and (+)-4a·(-)-4a have
been reported previously13 and deposited with the Cambridge
Crystallographic Data Center (Nos. CCDC-187956 and CCDC-
187957).
to favorable noncovalent interactions. Similar to natural
systemsswhose noncovalent interaction networks are tightly
controlledsself-sorting systems typically comprise only a small
fraction of the conceivable noncovalent aggregates. In this paper,
we demonstrated that the addition of a new molecular clip (e.g.,
1) to a self-sorting mixture (e.g., 2·2, (+)-5·(-)-5, and 8·8)
results in the selective heterodimerization of only one member
of the system. By extension, it should be possible to prepare
larger systems comprising sets of molecular clips with common
spatial distribution of their H-bonding arms (e.g., 1-3, 4-6,
and 8-10) that display self-sorting between sets but not within
sets. The development of stimuli responsive versions of such
self-sorting systemssthat result in new connection between
setsspromises the development of complex systems that exhibit
behaviors typically reserved for natural systems.
Acknowledgment. We thank the National Science Founda-
tion (Grant No. CHE-0615049) and the National Institutes of
Health (Grant No. GM61854) for financial support of this work.
Supporting Information Available: Text, tables, and figures
giving synthetic procedures and characterization data, selected
1H NMR spectra, 1H and 13C NMR spectra for all new
compounds and details of the X-ray structures of 2·2, (+)-
5 · (-)-5, (+)-7 · (+)-7, 8 · 8, and (()-12; X-ray data are also
given as CIF files. This material is available free of charge via
the Internet at http://pubs.acs.org.
Experimental Section
General experimental details have been reported previously.14
NMR Experiments. NMR spectra were measured on spectrom-
eters operating at 400 or 500 MHz for 1H and 100 or 125 MHz for
13C. Temperature was maintained ((0.5 K) with a temperature
control module that has been calibrated using the separation of the
resonances of methanol. For the self-association measurements,
JO8009424
J. Org. Chem. Vol. 73, No. 15, 2008 5925