Synthesis and investigation of host-[2]rotaxanes that bind metal cations

Article abstract of DOI:10.1021/jo100330e

Materials that bind metal cations are highly sought after for new devices. In this report, we show that rotaxanes can transfer metal cations with picrate, perchlorate, or chloride counterions from an aqueous solution into chloroform. The rotaxanes contain a dibenzyl-24-crown-8 ether as the wheel with either a benzyl-18-crown-6 ether (CEBG-R1-3) or a 3,5-dimethylbenzyl moiety (ArBG-R) as one blocking group. Alkali and alkaline picrate salts were efficiently extracted from an aqueous solution, presented in the millimolar range, into chloroform. Large association constants were derived for the complexes in chloroform, especially for the divalent cation Mg²⁺. Switching the counterion to chloride greatly diminished the amount of salt extracted. To explore the transfer mechanism of the rotaxanes, a comparison was made in the amount of NaClO₄, KClO₄, NaCl, and KCl extracted by CEBG-R1, ArBG-R, benzyl-18-crown-6 ether (B18C6), and two model compounds, which were used to represent the crown-ether blocking group and the axle of a rotaxane. Two-dimensional NMR analysis was performed on the rotaxane-cation complexes in CDCl₃. We found that the host rotaxanes transfer the perchlorate salts poorly when compared to B18C6, but they transfer chloride salts from 1 M salt solutions, whereas B18C6 does not. The transfer of chloride salts appears to rely on an allosteric type relationship between the binding of the chloride ion and metal cation to a rotaxane. Accordingly, when chloride binds to the dialkylammonium ion of the axle, the wheel moves along the axle and forms a binding site for a metal cation. In this report we demonstrate that host rotaxanes can bind metal cations, change their geometries upon cation and anion association, and operate through allosteric mechanisms, making them promising candidates for molecular devices.

Full text of DOI:10.1021/jo100330e

pubs.acs.org/joc
Synthesis and Investigation of Host-[2]Rotaxanes
That Bind Metal Cations
Xiaoyang Wang, Jing Zhu,† and David B. Smithrud*
Department of Chemistry, University of Cincinnati, PO Box 210172, Cincinnati, OH 45221.
†Current address: Johns Hopkins University School of Medicine, Baltimore, MD 21205.
david.smithrud@uc.edu
Received March 2, 2010
Materials that bind metal cations are highly sought after for new devices. In this report, we show that
rotaxanes can transfer metal cations with picrate, perchlorate, or chloride counterions from an
aqueous solution into chloroform. The rotaxanes contain a dibenzyl-24-crown-8 ether as the wheel
with either a benzyl-18-crown-6 ether (CEBG-R1-3) or a 3,5-dimethylbenzyl moiety (ArBG-R) as
one blocking group. Alkali and alkaline picrate salts were efficiently extracted from an aqueous
solution, presented in the millimolar range, into chloroform. Large association constants were
derived for the complexes in chloroform, especially for the divalent cation Mg2þ. Switching the
counterion to chloride greatly diminished the amount of salt extracted. To explore the transfer
mechanism of the rotaxanes, a comparison was made in the amount of NaClO4, KClO4, NaCl, and
KCl extracted by CEBG-R1, ArBG-R, benzyl-18-crown-6 ether (B18C6), and two model com-
pounds, which were used to represent the crown-ether blocking group and the axle of a rotaxane.
Two-dimensional NMR analysis was performed on the rotaxane-cation complexes in CDCl3. We
found that the host rotaxanes transfer the perchlorate salts poorly when compared to B18C6, but they
transfer chloride salts from 1 M salt solutions, whereas B18C6 does not. The transfer of chloride salts
appears to rely on an allosteric type relationship between the binding of the chloride ion and metal
cation to a rotaxane. Accordingly, when chloride binds to the dialkylammonium ion of the axle, the
wheel moves along the axle and forms a binding site for a metal cation. In this report we demonstrate
that host rotaxanes can bind metal cations, change their geometries upon cation and anion
association, and operate through allosteric mechanisms, making them promising candidates for
molecular devices.
Introduction
that house metal cations in chemically active pockets.1-5
One of the many challenges in designing such hosts is to
arrange the functional groups in the correct spatial arrange-
ment to bind a metal cation, ostensibly by pointing their
dipole moments toward the complexation site.6 Proteins use
a complex arrangement of side chains to efficiently extract
(i.e., low concentration of the host and salt) metal cations out
of aqueous solutions, making them useful model systems to
New industrial and medical applications, such as artificial
enzymes or sensors, could arise with the creation of materials
(1) Palacios, M. A.; Wang, Z.; Montes, V. A.; Zyryanov, G. V.;
Anzenbacher, P. J. Am. Chem. Soc. 2008, 130, 10307–10314.
(2) Maeda, H.; Tierney, D. L.; Mariano, P. S.; Banerjee, M.; Cho, D. W.;
Yoon, U. C. Tetrahedron 2008, 64, 5268–5278.
(3) Creus, M.; Ward, T. R. Org. Biomol. Chem. 2007, 5, 1835–1844.
(4) Dhanasekaran, M.; Negi, S.; Sugiura, Y. Acc. Chem. Res. 2006, 39,
45–52.
(5) Wilkinson, M. J.; van Leeuwen, P. W. N. M.; Reek, J. N. H. Org.
Biomol. Chem. 2005, 3, 2371–2383.
(6) Hay, B. P.; Oliferenko, A. A.; Uddin, J.; Zhang, C. G.; Firman, T. K.
J. Am. Chem. Soc. 2005, 127, 17043–17053.
3358 J. Org. Chem. 2010, 75, 3358–3370
Published on Web 04/22/2010
DOI: 10.1021/jo100330e
r
2010 American Chemical Society

Wang et al.
JOCArticle
design hosts. Alkaline cations, e.g., Ca2þ and Mg2þ, are
generally bound within pockets composed of essentially two
domains. An inner shell contains carboxylates of several Asp
or Glu residues that coordinate with a cation. Mg2þ also
binds at least one water molecule.7,8 The outer shell is
composed of hydrophobic side chains.9 Alkali cations
(Naþ and Kþ), on the other hand, tend to reside at protein
surfaces and bind to the oxygen atoms of carbonyl amides.10
Coordination sites for Liþ are similar to the sites for Mg2þ 11
.
To construct protein mimetics that bind cations, crown
ethers appear to be an ideal building block. Crown ethers
come in a variety of sizes. To specifically bind or extract a
metal cation, a crown ether is chosen by simply matching its
volume to the size of the metal cation.12-14 Extensive in-
vestigations by Bartsch showed that the ability of crown
ethers to selectively extract metal cations can be improved by
derivatizing them with functional groups (e.g., carboxylic
acids, phosphonic acid monoesters, carboxamides, and car-
boxylic esters) that provide additional favorable interactions
with a cation.15-20 Another option is to combine crown
ethers with other synthetic hosts, such as calixarenes.21
Mechanically interlocked molecules, such as catenanes
and rotaxanes,22-27 are likely candidates to enhance the
binding properties of crown ethers and, for this study, to
make them more protein-like. Their components are com-
monly assembled around metal cations or dialkylammonium
ions, and many contain crown ethers. Rotaxanes comprise a
linear molecule (axle) threaded through a circular molecule
(wheel) with bulky molecules (blocking groups) on the ends
of the axle to keep the wheel threaded. The nature of the
wheel is one advantage the rotaxanes has over other hosts. It
can slide along the axle in response to internal cation pro-
duction or external cation presentation to give a host-guest
FIGURE 1. Host-rotaxanes that were constructed to bind and
transport metal cations.
complex. For example, Li recently reported on an impressive
rotaxane sensor that operates in response to the binding of
Liþ and Zn2þ 28
Other recent applications include a bipyr-
.
1
idine based rotaxane29 whose H NMR spectrum is highly
sensitive to the presence of Liþ, Naþ, Kþ, Mg2þ, and Ca2þ
and molecular switches that operate through changes in the
position of the wheel caused by the association of an alkali
cation to an oligo(ethylene glycol) axle30 or Zn2þ binding to a
pyridine containing axle.31 In a similar manner to the latter
-
study, Chiu demonstrated a switch of PF6 counterion to
Cl- resulted in the wheel being repositioned on the axle
though tight ionic bond formation to the dialkylammonium
ion of the axle.32
Our research group initially created host rotaxanes to bind
aromatic guests and deliver materials into cells.33-36 Host-
rotaxanes are rotaxanes with at least one blocking group
being a pocket for guest recognition (Figure 1).37,38 Addi-
tional groups for guest binding are attached to the wheel.
These groups can assemble with the pocket to envelop a guest
in a manner similar to protein pockets. Since the components
of the binding site are split between the pocket and the wheel
and the wheel can pirouette and slide along the axle, the
structure of the pocket can readily change.36 This is advan-
tageous for strong guest association since the pocket can
adjust its geometry to maximize the binding free energy.
Additionally, rotaxanes can undergo the structural changes
required for protein-like functions and the workings of
future devices. These new materials not only may become
new devices with a wide range of applications but could
(7) Dudev, T.; Lim, C. Chem. Rev. 2003, 103, 773–787.
(8) Dudev, T.; Cowan, J. A.; Lim, C. J. Am. Chem. Soc. 1999, 121,
7665–7673.
(9) Yamashita, M. M.; Wesson, L.; Eisenman, G.; Eisenberg, D. Proc.
Natl. Acad. Sci. U.S.A. 1990, 87, 5648–5652.
(10) Rodgers, M. T.; Armentrout, P. B. Acc. Chem. Res. 2004, 37,
989–998.
(11) De Freitas, D. M.; Castro, M. M. C. A.; Geraldes, C. F. G. C. Acc.
Chem. Res. 2006, 39, 283–291.
(12) Lee, E. K.; Cho, B. R.; Hu, H.; Bartsch, R. A. Anal. Chem. 2002, 74,
2177–2183.
(13) Izatt, R. M.; Pawlak, K.; Bradshaw, J. S.; Bruening, R. L. Chem. Rev.
1991, 91, 1721–1785.
(14) Izatt, R. M.;Bradshaw, J.S.;Nielsen, S.A.;Lamb, J. D.;Christensen, J.
J. Chem. Rev. 1985, 85, 271–339.
(15) Strzelbicki, J.; Bartsch, R. A. Anal. Chem. 1981, 53, 1894–1899.
(16) Walkowiak, W.; Ndip, G.; Bartsch, R. A. Anal. Chem. 1999, 71,
1021–1026.
(17) Ohki, A.; Lu, J. P.; Hallman, J. L.; Huang, X. W.; Bartsch, R. A.
Anal. Chem. 1995, 67, 2405–2408.
(28) Zhou, W. D.; Li, J. B.; He, X. R.; Li, C. H.; Lv, J.; Li, Y. L.; Wang, S.;
Liu, H. B.; Zhu, D. B. Chem.;Eur. J. 2008, 14, 754–763.
(29) Chen, N. C.; Huang, P. Y.; Lai, C. C.; Liu, Y. H.; Wang, Y.; Peng,
S. M.; Chiu, S. H. Chem. Commun. 2007, 4122–4124.
(30) Tokunaga, Y.; Nakamura, T.; Yoshioka, M.; Shimomura, Y. Tetra-
hedron Lett. 2006, 47, 5901–5904.
(31) Chen, N. C.; Lai, C. C.; Liu, Y. H.; Peng, S. M.; Chiu, S. H. Chem.-
Eur. J. 2008, 14, 2904–2908.
(32) Lin, C. F.; Lai, C. C.; Liu, Y. H.; Peng, S. M.; Chiu, S. H. Chem.-
Eur. J. 2007, 13, 4350–4355.
(33) Zhu, J.; House, B. E.; Fleck, E.; Isaacsohn, I.; Drew, A. F.; Smithrud,
D. B. Bioorg. Med. Chem. Lett. 2007, 17, 5058–5062.
(34) Wang, X.; Bao, X.; McFarland-Mancini, M.; Isaacsohn, I.; Drew,
A. F.; Smithrud, D. B. J. Am. Chem. Soc. 2007, 129, 7284–7293.
(35) Bao, X.; Isaacsohn, I.; Drew, A. F.; Smithrud, D. B. J. Org. Chem.
2007, 72, 3988–4000.
(36) Bao, X.; Isaacsohn, I.; Drew, A. F.; Smithrud, D. B. J. Am. Chem.
Soc. 2006, 128, 12229–12238.
(37) Smukste, I.; House, B. E.; Smithrud, D. B. J. Org. Chem. 2003, 68,
2559–2571.
(38) Smukste, I.; Smithrud, D. B. J. Org. Chem. 2003, 68, 2547–2558.
(18) Ohki, A.; Lu, J. P.; Huang, X. W.; Bartsch, R. A. Anal. Chem. 1994,
66, 4332–4336.
(19) Ohki, A.; Lu, J. P.; Bartsch, R. A. Anal. Chem. 1994, 66, 651–654.
(20) Walkowiak, W.; Charewicz, W. A.; Kang, S. I.; Yang, I. W.; Pugia,
M. J.; Bartsch, R. A. Anal. Chem. 1990, 62, 2018–2021.
(21) Gutsche, C. D. In Calixarenes. An Introduction, 2nd ed.; Royal
Society of Chemistry: Cambridge, U.K., 2008; pp 104-107.
(22) Kay, E. R.; Leigh, D. A. Pure Appl. Chem. 2008, 80, 17–29.
(23) Kay, E. R.; Leigh, D. A.; Zerbetto, F. Angew. Chem., Int. Ed. 2007,
46, 72–191.
(24) Menon, S. K.; Guha, T. B.; Agrawal, Y. K. Rev. Inorg. Chem. 2004,
24, 97–133.
(25) Balzani, V.; Gomez-Lopez, M.; Stoddart, J. F. Acc. Chem. Res. 1998,
31, 405–414.
(26) Molecular Catenanes, Rotaxanes and Knots: A Journey through the
World of Molecular Topology; Sauvage, J.-P., Dietrich-Buchecker, C. O.,
Eds.; Wiley-VCH: Weinheim, Germany, 1999 .
(27) Loeb, S. J. Chem. Soc. Rev. 2007, 36, 226–235.
J. Org. Chem. Vol. 75, No. 10, 2010 3359

Wang et al.
JOCArticle
SCHEME 1
potentially operate within cells. In this report, we describe
the synthesis and properties of the first host-rotaxanes
designed to bind metal cations.
rotaxanes (Scheme 1).37,39 For the host-[2]rotaxanes, their
synthesis began with the known amino-benzyl-18-crown-6-
ether 1.40 Boc-β-alanine was linked to the crown ether to
provide an attachment site for the DCC-rotaxane,39 once it
was deprotected. The DCC-rotaxane contains a Boc-pro-
tected diamino-B24C8 ether as the wheel. CEBG-R1 was
deprotected, and the wheel was derivatized with mono-
methylmalonate to give CEBG-R2 (once the ester was
hydrolyzed) or with glutaric anhydride to give CEBG-R3.
Although rotaxane synthesis is facilitated by the readily
formed amide bonds and the attraction between the oxygen
atoms of the wheel and the dialkylammonium ion of the axle,
the presence of these functional groups could provide for
more than one binding site for a metal cation. Metal cations
could bind to the B18C6-blocking group or one or more of
the carbonyl oxygen atoms. Because dibenzyl-24-crown-8
ether (DB24C8) can complex two Naþ or two Kþ,41,42 the
Results
Design and Synthesis of the Rotaxanes. As discussed pre-
viously, protein pockets that house metal cations contain
hydrophilic and hydrophobic domains. These first genera-
tion rotaxanes were designed to mimic the inner shell. A
CHCl3 solution was used to represent the hydrophobic shell.
Benzyl-18-crown-6 (B18C6) ether was chosen for the pre-
formed recognition pocket and attached to the rotaxanes as a
blocking group. It binds alkali cations, with Kþ being its
preferred guest. Carbonyl or carboxylic acid moieties were
attached to the wheel to bind the exposed portion of alka-
line and alkali cations, respectively, when housed within the
crown ether blocking group. The crown ether-blocking
group-rotaxanes (CEBG-Rs) and aromatic ring-blocking
group-rotaxane (ArBG-R) were readily synthesized using
methods previously developed for the construction of host-
(39) Zehnder, D. W.; Smithrud, D. B. Org. Lett. 2001, 3, 2485–2487.
(40) Chandler, J. C.; Deady, L. W.; Reiss, J. A. Aust. J. Chem. 1988, 41,
1051–1061.
(41) Hughes, D. L. J. Chem. Soc., Dalton Trans. 1975, 2374–2378.
3360 J. Org. Chem. Vol. 75, No. 10, 2010

Wang et al.
JOCArticle
wheel could be a binding site for a metal cation as well, even
though it is threaded onto the axle, which contains a dialky-
lammonium ion. Another possibility is that the wheel could
be positioned near the B18C6-blocking group, creating a
single binding pocket composed of two crown ethers. To help
elucidate the coordination sites, model compounds were
constructed and investigated as well (Scheme 1). Amide-
B18C6 contains the crown ether pocket and the two amide
moieties of the axle. It serves as a model for the portion of the
axle attached to the B18C6-blocking group. ArBG-R con-
tains the axle and wheel of the CEBG-R1, but not the B18C6-
blocking group. Instead it has a 3,5-dimethylbenzenoid ring
as one blocking group. Axle-1 was investigated to determine
the ability of the axle’s amides to bind metal cations. ArBG-
R was readily synthesized by coupling the DCC-rotaxane
with the commercially available 3-amino-N-(3,5-dimethyl-
phenyl)propanamide 5. The coupling of 5 with the Boc-
protected axle 6 gave Axle-1 after deprotection. We postu-
lated that the investigation of a potential binding pocket;
isolated and in combinations;would lead to the identifica-
tion of the coordinating site and the functional groups
necessary to strongly and selectively bind metal cations.
Association Constants Obtained from Extraction Assays.
Two-phase extraction assays (ions(aq)/CHCl3) were per-
formed to obtain transfer efficiencies and a measure of the
binding affinities.43 Extraction constants (Ke) were deter-
mined from the known concentration of the components and
the concentration of picrate ion extracted into the CHCl3
(chl) layer by a host, which is measured using UV-vis
absorption spectroscopic analysis. The distribution con-
stants (Kd) were taken from the literature.40,44 Association
constants for the binding of the metal cation by a host in
CHCl3 are the ratio of Ke/Kd. The percentage of metal cation
extraction (%T) is equal to the ratio of the concentration of
the picrate ion in the chloroform layer versus the total
amount of picrate ion measured in the chloroform and
aqueous layers. We found that all hosts used in this study
existed predominantly in the CHCl3 layer (>93%). Since the
concentration of the host is approximately 30-fold greater
than the concentration of the metal cation, the complexes are
likely in a 1:1 ratio of host to guest. Additionally, the amount
of picrate ion extracted in CHCl3 by the hosts, which was
determined through UV-vis analysis, was compared to the
amount of Naþ and Kþ extracted as determined through
atomic absorption analysis (see below). Plotting the concen-
trations in CHCl3 of Pic- versus the alkali cations produced
a correlation coefficient of 0.94. The amount of Mg2þ and
picrate extracted by the rotaxanes was found to be in a 1:2
ratio. These findings are consistent with a 1:1 ratio of a host
to a picrate salt in the CHCl3 layer and a picrate salt being
extracted instead of counterion exchange occurring between
FIGURE 2. Percent of picrate ion in CHCl3 versus the total con-
centration of picrate ion after extraction from an aqueous phase
into CHCl3 by a host (uncertainty %T e 10%). Experimental
conditions are given in Table 1 and %T values are given in
Supporting Information.
TABLE 1. Association Constants (Ka ꢁ 10-6, M-1) for Host-Metal
Cation Complexes in CHCl3 As Derived from Extraction Experiments of
Metal Cation Picrate Salts (% Transfer Values Shown in Figure 2)a
Liþ
Naþ
Kþ
Csþ
Mg2þ
B18C6
NDb
ND
1.6
14
14
3.0
16
2.0
0.12
9.8
35
7.9
3.1
13
35
1.6
ND
ND
Amide-B18C6
Axle-1
ArBG-R
CEBG-R1
CEBG-R2
CEBG-R3
3.2
1.8
9.1
52
3.2
40
0.58
0.68
2.8
7.4
1.0
7.9 ꢁ 103
9.0 ꢁ 103
2.6 ꢁ 106
1.4 ꢁ 106
2.7 ꢁ 105
11
a[Host] = 2 mM, [picric acid] = 70 μM, and [MþOH-] = 50 mM or
[Mg(OH)2] = 20 mM, stirred for 30 min at 25 °C, uncertainty in Ka’s e
10%, bND means not determined since %T < 1%.
Ka ¼ Ke=Kd
%transfer ð%TÞ ¼ ½Pic-
ꢀ
chl=ð½Pic- ꢀchl þ ½Pic- ꢀaqÞ 3 100%
The transfer efficiencies (%T) and the association con-
stants (Ka) in CHCl3 of the hosts interacting with picrate salts
are given in Figure 2 and Table 1, respectively. As expected,
B18C6 ether preferentially extracts Kþ and shows the fol-
lowing trend Kþ>Csþ>Naþ . Liþ ≈ Mg2þ. Amide-B18C6
shows the same trend for cation binding as B18C6, but its
transfer efficiencies and attraction for cations in CHCl3 are
diminished. Derivatizing the aromatic ring of B18C6 with an
amide and/or the existence of the additional amide on the
tether reduces the ability of the crown pocket to bind cations.
These results further show that the oxygen atoms of the
amides do not provide a stable binding pocket for a metal
cation. Combining a rotaxane wheel with Amide-B18C6 to
give CEBG-R1 results in a dramatic change in the observed
strength and preference of cation binding. CEBG-R1 trans-
fers all picrate salts into CHCl3 with a slight preference for
Mg2þ and Kþ (Mg2þ>Kþ>Csþ>Liþ>Naþ). The amount
of Kþ extracted is greater than seen for Amide-B18C6 and
approximately the same as obtained with B18C6. CEBG-R1
either diminishes the negative effect of the amide moieties on
the crown pocket or contains a different binding pocket. The
Pic- and TFA-.
n
Ke ¼ ½Host Mnþ nPic-
ꢀ
chl=½Hostꢀchl½Mnþꢀaq½Pic- ꢀaq
3
3
n
Kd ¼ ½Mnþ nPic-
ꢀ
chl=½Mnþꢀaq½Pic- ꢀaq
3
(42) Mercer, M.; Truter, M. R. J. Chem. Soc., Dalton Trans. 1973,
2469–2473.
(43) Helgeson, R. C.; Weisman, G. R.; Toner, J. L.; Tarnowski, T. L.;
Chao, Y.; Mayer, J. M.; Cram, D. J. J. Am. Chem. Soc. 1979, 101, 4928–4941.
(44) Koenig, K. E.; Lein, G. M.; Stuckler, P.; Kaneda, T.; Cram, D. J.
J. Am. Chem. Soc. 1979, 101, 3553–3566.
J. Org. Chem. Vol. 75, No. 10, 2010 3361

Wang et al.
JOCArticle
transfer percentage increases from approximately 1% for
Amide-B18C6 to 66%, 58%, and 97% for Liþ, Naþ, and
Mg2þ, respectively for CEBG-R1. There is also a modest
improvement in the transport of Csþ, as well. This dramatic
increase in the transfer of most metal cations by CEBG-R1,
as compared to Amide-B18C6, indicates an additional bind-
ing pocket exists in the rotaxane, which most likely involves
the wheel.
properties as CEBG-R1, whereas the short chain variant
CEBG-R2 extracts alkali cations less efficiently than CEBG-
R1. On the other hand, CEBG-R2 does show a clear pre-
ference for Mg2þ over the other metal cations.
Effect of Counter Ion on Transfer Efficiencies. The extrac-
tion of the metal cations is likely aided by the presence of the
picrate counterion(s). According to the hydrophobic ef-
fect,45 the aromatic ring of a picrate ion will favorably
associate with the hydrophobic groups of a host at the
interface between the CHCl3/H2O layers. Once in the CHCl3
layer, the strength of this interaction is diminished. Con-
sidering the large magnitude of the Ka values (Table 1), the
free energy for association in CHCl3 is driven by cation-host
interactions. The key step is the plucking of the metal cation
from bulk water. Transfer experiments were performed for
To find the other binding pocket(s) and to determine
whether the wheel or axle is part of this pocket, the properties
of ArBG-R and Axle-1 were determined and compared to
the other hosts. Axle-1 can extract all of the metal cations
tested, preferring Naþ and Mg2þ (Naþ ≈ Mg2þ > Kþ
≈
Csþ > Liþ). This result is surprising considering Amide-
B18C6 extracts only Kþ and Csþ. Both Axle-1 and Amide-
B18C6 contain the two amides, but Amide-B18C6 has
the known crown ether binding pocket. ArBG-R shows a
preference for Naþ (Naþ>Kþ ≈ Csþ ≈ Mg2þ ≈ Liþ) and
outperforms Axle-1, except in the extraction of Mg2þ. This
shows that the wheel contributes to the transfer process,
likely through the formation of a new binding pocket. Cob-
bling all the pieces together: B18C6 ether, wheel, and axle
produces CEBG-R1. It demonstrates a greater extraction of
the metal cations than Amide-B18C6, which was also ob-
served with ArBG-R. CEBG-R1, however, outperforms
ArBG-R in the extraction of Mg2þ, Kþ, and Csþ, giving it
a unique preference for metal cations (Mg2þ ≈ Kþ>Csþ>
Liþ ≈ Naþ). Surprisingly, CEBG-R1 demonstrates a nearly
quantitative transfer of Mg2þ and forms a very stable com-
-
metal cations with ClO4 and Cl- counterions to discover
the importance of the counterion. The percent transfer values
(%T) for these experiments were calculated by dividing the
concentration of a metal cation detected in the CHCl3 layer
by the concentration of the host (%T = [Mþ]chl/[Host]chl).
Atomic absorption spectroscopy was used to determine the
concentration of Kþ and Naþ in the CHCl3 layer after the
extraction experiments.46 The measured absorbance values
were converted to concentrations using calibration curves,
which were constructed from the analysis of solutions that
contained a known amount of metal cations. Extraction
assays were also performed without salt and host to account
for any solvent effect on the absorption. The very low
concentration of free cations in CHCl3, i.e., unbound ca-
tions, were undetectable using this assay. The carboxylic
acids of CEBG-R2 and CEBG-R3 were deprotonated using
Me4NOH. Being a large cation, Me4Nþ should not compete
with the metal cations for the binding sites.
We found the hosts do not extract NaCl or KCl at the same
concentration used in the picrate assay (2 mM), which shows
the importance of the picrate ion in the transfer of metal
picrates. To more fully investigate the effect of the counter-
ion, the concentration of the model hosts, ArBG-R, and
CEBG-R1 were raised to 20 mM and extraction assays were
performed with NaClO4 and KClO4 at 100 mM or with NaCl
and KCl at 1 M. We note that the maximum solubility of
KClO4 in water is around 100 mM, and metal cation transfer
was not observed from aqueous solutions containing 100 or
200 mM chloride salts. Because of the differences in the
concentrations of the components, only the trends of %T
will be compared and discussed. Switching the counterion
from picrate to a perchlorate produces small differences in
the preference of Kþ over Naþ for B18C6 and for Amide-
B18C6 (Figure 2 and Table 2). A much greater effect on %T
is seen for Axle-1, ArBG-R, and CEBG-R1 upon switching
the counterion. Whereas these hosts outperformed B18C6
and Amide-B18C6 in the %T of NaPic and in some cases
KPic, this advantage is lost with the perchlorate salts. Axle-1,
ArBG-R, and CEBG-R1 either did not extract the perchlo-
rate salts or performed as well as Amide-B18C6. This result is
consistent with the ability of picrate to interact favorably
with the aromatic surfaces of the hosts, which drives host-
guest association in an aqueous environment. Because Axle-1,
plex with Mg2þ in CHCl3 (ΔG° = -17 kcal mol-1). Its
3
extraction of Kþ is similar to the level displayed by B18C6,
which suggests that the presence of the B18C6-blocking
group of CEBG-R1 is responsible for improved extraction
of Kþ and likely Csþ as well when compared to ArBG-R.
ArBG-R extracts only Naþ better than CEBG-R1. The
binding pocket for Naþ available in ArBG-R appears not
to exist in CEBG-R1. A comparison of the experimental
results for the hosts shows that the components B18C6-
blocking group, wheel, the amide(s) of the axle, and even
the aromatic blocking group can enhance or diminish the
binding of picrate salts, which depends on the number of
binding pockets that assemble and how they combine.
To potentially obtain a single binding site and more stable
complexes, the wheel of the rotaxane was derivatized with
carboxylic acids. We envisioned the carboxylates extending
outward and covering the B18C6 blocking group in a fashion
similar to protein binding sites. Since the rotaxane can adopt
a variety of conformations, linkers with different lengths
were used in CEBG-R2 and CEBG-R3 to attach the car-
boxylic acid to the wheel. The wheel tends to reside over the
dialkylammonium ion of the axle in CHCl3 or in an apolar
environment, whereas in an aqueous environment, the wheel
resides closer to a hydrophobic end of an axle.36 Metal cation
extraction from an aqueous solution into CHCl3 will likely
involve the wheel being positioned on the axle to maximize
the favorable binding free energy. The carboxylic acid of
CEBG-R2 or CEBG-R3 could be better positioned depend-
ing on which step, aqueous sequestration of the cation or
cation complexation in CHCl3, drives the transfer event.
Attaching a carboxylic acid to the wheel unfortunately does
not significantly improve the ability of a CEBG-rotaxane to
bind or transfer metal cations. CEBG-R3 displays similar
(45) Xu, H. F.; Dill, K. A. J. Phys. Chem. B 2005, 109, 23611–23617.
(46) Borrebaeck, C. A. K.; Lonnerdal, B.; Etzler, M. E. Biochemistry
1981, 20, 4119–4122.
3362 J. Org. Chem. Vol. 75, No. 10, 2010

Wang et al.
JOCArticle
TABLE 2. Percent of Host Bound to a Metal Chloride Salt in CHCl3 Obtained through Extraction Experimentsa
KCl
(1 M)
NaCl
(1 M)
KCl or NaCl
(0.2 M)
KClO4
(0.1 M)
NaClO4
(0.1 M)
KClO4 (0.1 M);
KCl (0.1 M)
NaClO4 (0.1 M);
NaCl (0.1 M)
B18C6
Amide B18C6
Axle-1
ArBG-R
CEBG-R1
NDb
ND
ND
74
ND
5
ND
32
ND
ND
ND
ND
ND
60
26
ND
ND
22
40
9
ND
7
52
22
ND
17
13
ND
ND
12
40
14
ND
65
ND
a[Host] = 20 mM, stirred for 24 h at 25 °C, uncertainty in %T is 3-10%. bND not detected (%T < 1%).
ArBG-R, and CEBG-R1 contain a greater number of aro-
matic rings than B18C6 and Amide-B18C6, they should be
more greatly affected by a switch to a nonaromatic counter-
ion. The fact that Amide-B18C6 and CEBG-R1 extract
KClO4 approximately in the same amount (26% Amide-
B18C6 Kþ and 22% CEBG-R1 Kþ), whereas Axle-1 and
The importance of the chloride ion for the transfer event is
realized upon comparing %T of perchlorate salts versus
chloride salts. A switch in preference occurs for ArBG-R
-
from Naþ to Kþ with a switch from ClO4 to Cl- (7%
ArBG-R Naþ/ND ArBG-R Kþ for ClO4- to 32% ArBG-
3
3
R Naþ/70% ArBG-R Kþ for Cl-). CEBG-R1 demon-
3
3
3
3
strates a smaller preference for Kþ over Naþ with a switch
ArBG-R do not detectably transfer KClO4, suggests that
Kþ resides in the ether pocket of the B18C6-blocking group
-
of ClO4 to Cl- (>22-fold preference for KClO4 over
-
when it carries a ClO4 counterion. NaClO4 is not extr-
NaClO4 to a 3-fold preference for KCl over NaCl). To obtain
a direct measure of the effect of Cl- on the transfer of metal
cations, extraction assays were performed with a mixture of
NaCl (100 mM)/NaClO4 (100 mM) or KCl (100 mM)/
KClO4 (100 mM). The total cation concentration of 200 mM
was chosen because we found that Naþ and Kþ are not
noticeably extracted from an aqueous solution containing
chloride salts at 200 mM. An enhancement in the %T under
these conditions, as compared to the perchlorate salts
alone at 100 mM, would show that Cl- is not only actively
participating in the transfer event by binding to a host, but its
binding results in an allosteric type of phenomenon, giving
improved cation binding. There is a slight decrease in the
amount of Kþ extracted when 100 mM of Cl- is added to an
aqueous solution of KClO4 for B18C6 and Amide-B18C6. A
large reduction in the amount of Naþ extracted by these
hosts occurs using the mixed aqueous solution (NaCl/
NaClO4) as compared to the NaClO4 solution. Contrarily,
the rotaxanes extract a significantly greater amount of Kþ
with the addition of 100 mM KCl to the aqueous solution
containing 100 mM KClO4 (22% without Cl- to 65% with
Cl- of bound host for CEBG-R1 and ND without Cl- to
17% with Cl- of bound host for ArBG-R). Only ArBG-R
extracts a greater amount of Naþ from the mixed aqueous
solution of NaCl and NaClO4 (7% without Cl- to 12% with
Cl- of bound host). As seen with the other extraction assays,
CEBG-R1 extracts less Naþ under these conditions than
ArBG-R.
acted by Axle-1 or CEBG-R1, and only a small amount is
extracted by Amide-B18C6 and ArBG-R. The inability of
Axle-1 to transfer NaClO4 indicates that a crown ether,
either free or as a wheel, is required for these hosts to extract
NaClO4. CEBG-R1 does not transfer Naþ as well as ArBG-
R independent of whether the counterion is perchlorate or
picrate.
Because of a large desolvation penalty, the transfer of
chloride salts from water into CHCl3 is a very unfavorable
process. This explains the lack of observable extraction of
KCl or NaCl from a 1 M aqueous solution into CHCl3 by
B18C6, Amide-B18C6, and Axle-1, except for the 5% of
Amide-B18C6 bound to Naþ (Table 2). Considering the
poor performance of the rotaxanes in the extraction of the
perchlorate salts, we were surprised to find that they extr-
act chloride salts. KCl is extracted preferably over NaCl.
Seventy-four percent of ArBG-R and 40% of CEBG-R1 are
bound to Kþ in CHCl3, assuming a 1:1 complex of cation per
host. The percentage of bound host is approximately halved
in the extraction of NaCl. Interestingly, CEBG-R1 is not as
efficient in extracting KCl as ArBG-R, even though it
contains the B18C6-binding pocket. This fact combined with
the observation that B18C6, Axle-1, and Amide-B18C6
do not noticeably extract KCl suggests that the binding
pocket of the host rotaxanes for Kþ and Naþ is constructed
from the axle’s amides and the wheel. Competition assays
were performed to determine whether Naþ and Kþ bind to
the same pocket or different pockets. Presenting CEBG-R1
to KCl (1 M) and NaCl (1 M) resulted in 24% of the host
being bound to Kþ and 10% bound to Naþ in the CHCl3
layer. This is a reduction in the amount of both cations being
1H NMR Analysis of the Coordination Sites of the Host
Rotaxanes. To further investigate the coordination sites for
the metal cations and the counterions, 1D and 2D 1H NMR
analyses were performed on solutions containing CEBG-R1
and ArBG-R with and without the various salts. The “metal-
cation-free” condition is defined as the state of the host
obtained from HPLC purification. Because the eluent con-
tained 0.1% TFA, we assume the dialkylamines of the
rotaxanes are protonated with TFA- counterions. Extrac-
tion assays were performed, and the materials in the chloro-
form layer were analyzed via 1H NMR spectroscopy.
Substantial changes were observable in the 1H NMR spectra
of the hosts (Figure 3). The effect of sodium salts on the 1H
NMR spectra of ArBG-R is shown in Figure 4 as an
example. A few general trends observed in the plots are
that the aromatic signals (C, F and G, H) and the axle
extracted when presented separately (40% CEBG-R1 Kþ
3
and 14% CEBG-R1 Naþ). This is consistent with the cat-
3
ions binding to the same site on the rotaxane, with Kþ being
selected over Naþ. Presenting ArBG-R to NaCl (1 M) and
KCl (1 M), however, resulted in an increase in the amount of
Naþ extracted and a modest reduction in the amount of Kþ
extracted (57% ArBG-R Kþ and 54% ArBG-R Naþ; salts
3
þ 3
presented separately gave 74% ArBG-R K and 32%
3
ArBG-R Naþ). Apparently, Naþ and Kþ bind to different
3
pockets of ArBG-R. The binding of Kþ enhances the asso-
ciation of Naþ, giving an allosteric type relationship for
cation binding and a concomitant loss of preference for Kþ.
J. Org. Chem. Vol. 75, No. 10, 2010 3363

Wang et al.
JOCArticle
FIGURE 3. Changes to the 1H NMR spectra of hosts after the extraction assays. (A) Red bonds highlight the protons that show a change in
their chemical shift (>0.05 ppm) in the spectra of CEBG-R1 after it is extracted with ClO4- (Naþ or Kþ) at 100 mM compared to when it is
extracted with Cl- (Naþ or Kþ) at 200 mM. Also shown are the differences in the chemical shifts for CEBG-R1 after extracting it with a low
concentration of Cl- (Naþ or Kþ at 200 mM) compared to after extracting with a higher concentration of KCl (green) or NaCl (blue) at 1 M. (B)
The same comparisons were made for ArBG-R.
protons O become more separated and the protons of amide
linker (D, E) converge when the plots are viewed starting
from TFA and going up to 1 M NaCl. The dialkylam-
monium ion protons L collapse and are positioned upfield
1H NMR analysis was performed on the hosts after they
were extracted with the chloride salts at a concentration
where cation extraction does not occur (200 mM) to deter-
mine if Cl- binds to them without concomitant cation
binding. Observed changes include the protons of the dia-
lkylammonium group of CEBG-R1 and ArBG-R, which are
shifted downfield to 7.9 and 8.2 ppm, respectively. A chemi-
cal shift of 9.3 ppm is observed for these protons when the
axle is protonated with HCl. At least a portion of the hosts
had Cl- as the counterion, and the wheel is off the dialky-
lammonium ion. The spectra of hosts with Cl- as the
counterion were compared to the corresponding host with
ClO4- as the counterion to potentially discover how Cl- aids
in the transfer of the cations. As discussed previously, the
chemical shifts of these spectra are not influenced by a metal
cation. For CEBG-R1, shifts were observed in the Ar-H’s of
the wheel, B18C6-blocking group, and the dialkylammo-
nium protons (Figure 3). Changes to the aromatic environ-
ments suggest that the wheel is positioned closer to the
B18C6-blocking group with Cl- as the counterion. Cl-
forms a tight ion pair with the dialkylammonium ion in
CHCl3, which forces the wheel off the dialkylammonium ion.
This is why pseudorotaxane formation (wheel threaded onto
axle) requires a swap of Cl- for PF6-, which forms a weak or
loose ion pair. For ArBG-R, substantial changes are not only
observed for the Ar-H protons of both blocking groups
(CEBG-R1 changes occurred with only the 18C6-blocking
group) and the wheel, but changes are also observed for the
protons of the axle and several of the ether linkages of the
wheel. Apparently, the wheel shifts a greater distance away
from the dialkylammonium ion of ArBG-R, as compared to
CEBG-R1, with a swap of Cl- for ClO4-. Most protons
-
with Cl- as the counterion instead of ClO4 or TFA-. A
greater separation is clearly seen for protons N for the 1 M
NaCl extraction as compared to the other conditions. 2D 1H
NMR spectra of the hosts are available in Supporting
Information.
The hosts were extracted with 100 mM NaClO4 or 100 mM
KClO4. At this concentration, ArBG-R transfers Naþ, albeit
poorly (7% of host bound), but does not transfer Kþ (not
detected (ND)), whereas CEBG-R1 transfers Kþ (22% of
host bound) but not Naþ (ND). Substantial changes are
1
observable in the H NMR spectra of the hosts after the
extraction assays (Supporting Information), which suggest
that TFA- was replaced with ClO4-. Unexpectedly, the
chemical shifts of the dialkylammonium protons of the axle
(7.6 ppm, 7.7 ppm for ArBG-R and 7.2 ppm, 8.1 ppm for
CEBG-R1) are unchanged after the extraction assay. A
change in counterion should produce a change in the chemi-
cal shift of these protons. This unexpected result can be
explained with the observation that similar chemical shifts
and coupling pattern are observed for the pseudorotaxane
with the PF6- counterion (7.5 ppm, 7.7 ppm). This indicates
the wheel is over the dialkylammonium ion with either
-
TFA-, ClO4-, or PF6 as the counterion. The spectra of
each host after extracting with NaClO4 or KClO4 appear to
be identical (Supporting Information). Thus, changes to the
spectra that occur with a change in counterion (TFA- to
ClO4-) are independent of the nature of the cation. Either the
geometry of a host is unchanged when bound to a cation with
perchlorate as the counterion or the amount of host bound to
a cation is too low to produce an observable change in the
spectra.
-
affected by the swap of Cl- for ClO4 also show different
chemical shifts when the 1H NMR spectra of the hosts
with TFA- as the counterion are compared to hosts after
3364 J. Org. Chem. Vol. 75, No. 10, 2010

Wang et al.
JOCArticle
FIGURE 5. Proposed binding pockets for a metal cation are shown
combined in CEBG-R1.
Discussion
The hosts were designed to bind cations in performed
pockets, e.g., the B18C6-blocking group, with the wheel
providing additional contacting groups. We did not design
a binding site for a counterion. The counterion would play
the traditional role of enhancing the solubility of a salt in
chloroform with the order of picrate>perchlorate . chloride.
As expected, picrate salts are extracted at a lower concentra-
tion of components than the other salts. For B18C6 and
Amide-B18C6, the predicted solubility trend continues with
perchlorate salts being extracted more efficiently than the
chloride salts. The rotaxanes, however, break from the
pattern of the other hosts, showing the ability to transfer
chloride salts from modest to high levels. Advantages for the
rotaxanes include more oxygen atoms to bind a cation and
more aromatic surfaces to remove water molecules from the
oxygen atoms, which reduces the desolvation penalty that
occurs with transfer. The trick is to get these components to
work together to bind a metal cation, which is not a trivial
endeavor. For example, we found that not all oxygen atoms
embedded within a rotaxane contribute to cation binding,
and in some cases, their presence lowers the amount of a
metal cation that is extracted. To find which coordination
sites are responsible for cation transfer, the properties of the
hosts were compared. For this discussion, we will refer to the
potential coordination sites for the cations as the B18C6-
pocket, the amide-pocket, and the wheel-pocket (all pockets
are shown in CEBG-R1, Figure 5). These sites were chosen
on the basis of the known binding properties of crown ethers
and proteins.
Aromatic surfaces clearly aid in the transfer of picrate
salts, likely through the formation of stacking interactions
between the aromatic rings of a host with the picrate. Bartsch
found Pic- stacks in a face-to-face orientation with DB24C8
ether in the crystallographic structure of DB24C8-NaPic
complexes.47 Aromatic interactions should be the strongest
during the transfer event from water to chloroform, but once
in CHCl3, the favorable charge-dipole interactions between a
cation and the oxygen atoms of a host will provide the
majority of the binding free energy. Binding modes were
constructed to explain the observed results for the extraction
of picrate salts by placing a cation in one of the three pockets
while maintaining stacking aromatic interactions (Supporting
Information). Axle-1 contains only the amide-pocket. Thus,
FIGURE 4. Partial 1H NMR spectra in CDCl3 for ArBG-R (a) as
the TFA- salt and after extracting with an aqueous solution
containing (b) 100 mM NaClO4, (c) 200 mM NaCl, and (d) 1 M
NaCl.
extracting with Cl- (Naþ or Kþ) at 200 mM (Supporting
-
Information). This not surprising considering that ClO4
and TFA- form weaker contact ion pairs than Cl- in CHCl3.
1H NMR analysis was performed on the hosts after
extracting them with a high concentration of salt (1 M NaCl
or KCl), where metal cation binding occurs, to discover the
nature of the coordinating site for Naþ or Kþ with chloride
as the counterion. These spectra were compared to the spec-
tra of the corresponding salt presented at the lower concen-
tration of 200 mM. In the case of CEBG-R1 binding to Kþ,
there is very little change in the conformation of the host
except for the axle protons as indicated in green (Figure 3).
Thus, the host appears to be optimized for binding to Kþ
with Cl- present. Binding to Naþ, however, produces sub-
stantial changes in the chemical shifts of the wheel’s protons,
protons of the dialkylammonium ion, and a couple of
protons of the axle. This suggests that either the wheel shifts
its position upon Naþ binding and/or the wheel binds to
Naþ. A different pattern is seen for the binding of Naþ and
K
þ to ArBG-R. Binding of Kþ or Naþ produces the chan-
ges in the chemical shifts of the wheel as seen with CEBG-
R1; however, changes are also observed for the 3,5-dimethyl-
benzyl blocking group and the portion of the axle that
contains the amides. Furthermore, nearly identical spectra
are observed for ArBG-R bound to Naþ as Kþ, except for
one unique axle proton (shown in green) seen with the
binding of Kþ.
(47) Talanova, G. G.; Elkarim, N. S. A.; Hanes, R. E.; Hwang, H. S.;
Rogers, R. D.; Bartsch, R. A. Anal. Chem. 1999, 71, 672–677.
J. Org. Chem. Vol. 75, No. 10, 2010 3365

Wang et al.
JOCArticle
the cation likely binds to this pocket and its picrate ion
interacts with the aromatic surface of a blocking group. For
the binding of MgPic2, each picrate ion likely interacts with
one of the aromatic blocking groups. The addition of an
aromatic crown ether, i.e., the B18C6-blocking group, to a
host does not guarantee that it will extract a greater amount
of salt into chloroform. For example, linking B18C6 to Axle-1,
giving Amide-B18C6, diminishes the ability of the resulting
host to transfer Naþ, Liþ, and Mg2þ. Although Amide-B18C6
contains an aromatic blocking group, aromatic interactions
between this blocking group and the picrate counterion are
probably weak. The crown ether moiety of Amide-B18C6
forms favorable interactions with water molecules, and their
presence diminishes the hydrophobic effect afforded aro-
matic surfaces. Amide-B18C6 shows the same transfer trend
as B18C6, suggesting that Kþ and Csþ bind to its B18C6-
pocket and not the amide-pocket.
Mg2þ in CHCl3 suggests that a carboxylate of CEBG-R2 and
CEBG-R3 does not replace a picrate from Mg2þ. A possible
reason for the inability of a carboxylate to remove picrate
and for the poorer performance of CEBG-R2 is that the
carboxylic acids remain protonated in the CHCl3 layer, even
though 2 equiv of base was added to the aqueous phase. The
close proximity of a carboxylic acid of CEBG-R2 to a cation
binding site could result in H-bond formation between the
carboxylic acid proton and an oxygen atom or atoms used to
bind the metal cations, e.g., in the B18C6-pocket, amide-
pocket, or even the wheel-pocket. This H-bond would com-
pete with a metal cation for the binding domain. Mg2þ
apparently does not bind to the functional group or pocket
that interacts with the carboxylic acid of CEBG-R2. The
picrate ions could force the wheel away from the amide-
pocket and B18C6-pocket to maximize aromatic inter-
actions, which would reduce the negative effect of the carbo-
xylic acid.
ArBG-R contains the amide-pocket and the wheel-pocket.
Unlike the combination of B18C6-pocket and amide-pocket
(i.e., Amide-B18C6), this combination produces more effi-
cient transfer of all the alkali cations than the amide-pocket
alone (i.e., Axle-1). This suggests that the amide-pocket and
the wheel-pocket can converge to make a favorable binding
site. The wheel can slide over the amide-pocket, giving a
convergent arrangement of functional groups, in a manner
similar to protein binding sites. The similar level of Mg2þ
extracted by ArBG-R and Axle-1 is consistent with the
requirement that both picrates need to form aromatic inter-
actions with a host for Mg2þ to be efficiently extracted into
CHCl3. Dual stacking interactions between ArBG-R and the
picrate ions likely forces the wheel to move away from the
amide-pocket. Not only are aromatic interactions important
in the efficient transfer of picrate salts by these hosts, but
these interactions can also control which coordination site a
metal cation will bind to.
Perchlorate salts were extracted less efficiently into CHCl3
than the picrate salts. The perchlorate ion cannot form
aromatic interactions with a host and is less hydrophobic
than the picrate ion. The importance of aromatic interac-
tions is clearly seen with the inability of Axle-1 to transfer
either NaClO4 or KClO4. Furthermore, Amide-B18C6 out-
performs ArBG-R and CEBG-R1 in the transfer of NaClO4
and KClO4, whereas these rotaxanes transfer NaPic and
KPic to a greater extent than Amide-B18C6. The only useful
binding site for the perchlorate salts appears to be the
B18C6-pocket. Hosts with this pocket transfer KClO4
around 25% of bound host, except for B18C6, which shows
60% of bound host. ArBG-R, which contains the amide-
pocket and wheel-pocket, does not transfer KClO4 but does
transfer NaClO4, albeit weakly. Surprisingly, CEBG-R1,
which contains all three pockets, does not transfer NaClO4.
Unlike the picrate ions, ClO4- does not bind to the rotaxanes
in a manner that promotes metal cation transfer.
CEBG-R1 and CEBG-R3 contain all three binding pock-
ets and outperform ArBG-R in the transfer of Kþ, Csþ, and
Mg2þ. The improved extraction of Kþ and Csþ and the
availability of the B18C6-pocket, which extracts these ca-
tions, suggest that Kþ and Csþ bind in this pocket of the
hosts. Because CEBG-R1 and -R3 extracts these cations to a
greater extant than Amide-B18C6, the wheel must be in-
volved in the process, likely through favorable interactions
between the picrate ion and an aromatic ring of a wheel. In
the extraction of Mg2þ, the cation could bind in the amide-
pocket in a fashion similar to the predicted coordination site
of ArBG-R. CEBG-R1 and -R3, however, transfer Mg2þ
better than ArBG-R, which suggests Mg2þ also binds in the
B18C6-pocket. The wheel must be part of the coordinating
site since Amide-B18C6 does not transfer Mg2þ. The binding
site must also accommodate both aromatic rings of MgPic2,
which appears to be necessary for the observed high %T of
Mg2þ. We propose that Mg2þ is housed within the B18C6-
pocket and its picrate ions interact with both aromatic rings
of the wheel by one being positioned above and the other
positioned below the plane of the crown ether. CEBG-R2
transfers Mg2þ as well as CEBG-R1 and -R3, but it transfers
alkali cations less efficiently than these hosts. Its short -CH2-
linker, which attaches the carboxylic acid to the wheel,
appears to be the reason for its poorer performance.
CEBG-R3, which contains longer -(CH2)3- linkers, performs
as well as or better than CEBG-R1. The 2:1 ratio of picrate to
The transfer of NaCl and KCl from a 1 M aqueous phase
could be simply a result of the high concentration of alkali
cation, which is possible for assays with chloride salts. In this
case, Cl- would just go along with the cation into CHCl3 as
the counterion. Another possibility is that Cl- binds to a
host, and this association results in an allosteric-like relation-
ship promoting the binding of a cation. Although Pic- aids in
the extraction of picrate salts, it likely stays associated with
the metal cation and the entire salt binds as a single unit. An
allosteric relationship of proteins involves at least two bind-
ing sites and the binding of an agent to one site alters the
protein’s conformation in such a manner to promote the
binding of the second agent. The results of the experiments
presented herein are consistent with Cl- and Kþ or Naþ
binding to different sites on the host rotaxanes and the
binding of Cl- promotes the binding of the metal cation
through a conformational change. Replacing TFA- or
-
ClO4 with Cl- results in significant changes to its 1H
NMR spectra to a host. The likely binding site for Cl- is
the dialkylammonium ion of the axle since this will form a
stable, tight ion pair and anions are known to bind to the side
chains of proteins that are positively charged or form strong
1
H-bonds.48 Since the changes to the H NMR spectra are
not localized at the dialkylammonium ion, the binding of
(48) Kubik, S. Chem. Soc. Rev. 2009, 38, 585–605.
3366 J. Org. Chem. Vol. 75, No. 10, 2010

Wang et al.
JOCArticle
FIGURE 6. Proposed binding modes of the host rotaxanes with the perchlorate and chloride salts. For ArBG-R Cl-, Kþ, and Naþ can be
bound simultaneously within the crown ether-amide pocket, but not necessarily in the positions shown.
3
Cl- produces a global change in a host’s conformation most
likely through a repositioning of the wheel. Evidence for a
cation binding in a site being separate from the Cl- site is
obtained by the observation that different protons are
require the wheel as much for binding as Naþ, resulting in Kþ
being the preferred guest. The binding site of ArBG-R for
alkali metals of chloride salts was drawn with the wheel-
pocket sitting over the amide-pocket. Both amides of the
amide-pocket can combine with the wheel to produce a
binding site that can house Kþ, Naþ, or both cations
simultaneously. Kþ is the preferred guest, but once bound,
it may help to “lock” the wheel promoting the binding of
Naþ, producing the observed allosteric relationship in the
extraction assay.
1
affected in the H NMR spectra when Cl- replaces TFA-
as compared to when both a Mþ and Cl- bind. The last piece
of the allosteric puzzle is the fact that the binding of Cl-
promotes that association of a metal cation, which was
observed in extraction assays of mixed Cl-/ClO4- solutions.
We postulate that the binding of Cl- to the dialkylammo-
nium ion of the axle forces the wheel to shift toward the
amide-pocket, as shown in Figure 3. This establishes the
binding site of the metal cation, which is a combination of the
wheel-pocket and the amide-pocket. Even though the bind-
ing site exists, complexation requires either a high concen-
tration of a metal cation in the aqueous phase (available with
1 M NaCl or 1 M KCl) or the presence of a salt that is more
soluble in CHCl3, e.g. a perchlorate salt. ClO4-, on the other
hand, does not form a tight ion pair. The wheel either stays
on the dialkylammonium ion of the axle or slides rapidly
along the axle, but in either case, the binding site for a metal
cation is not formed. The coordinating sites of CEBG-R1
and ArBG-R formed by the chloride ion are not equivalent.
First, Naþ is bound more favorably by ArBG-R than
CEBG-R1 independent of the counterion. A second differ-
ence was observed in the competition assay. Naþ and Kþ
bind to the same site in CEBG-R1, whereas Naþ and Kþ can
bind simultaneously to different sites of ArBG-R and the
Conclusion
The goal of this research project was to create synthetic
mimetics of protein binding domains that combine their
functional groups in an appropriate arrangement to extract
metal chloride salts from aqueous solutions into chloroform.
Rotaxanes can perform this extraction, albeit it requires a
high concentration of NaCl or KCl in the aqueous phase.
The rotaxanes contain a variety of coordination sites for a
cation and aromatic surfaces to reduce the desolvation
penalty that occurs upon transfer. To obtain efficient trans-
fer, the various pockets need to be organized to bind a
particular salt. The extraction of picrate salts benefits from
the favorable interactions with aromatic surfaces of a host,
whereas the extraction of chloride salts require a host to
present an apolar pocket containing a convergent arrange-
ment of oxygen atoms. The interlocked nature of rotaxanes
enables the wheel to adjust its position along the axle to
accommodate the various salts. For picrate salts, the wheel
appears to provide an aromatic surface to bind to the picrate
ion of alkali cations and possibly both picrates of MgPic2
when the cation binds to the B18C6-pocket or amide-pocket.
In the extraction of NaCl and KCl, the wheel likely combines
with the amide-pocket to bind Naþ and Kþ. Formation of
this binding site, however, requires tight binding of Cl- to
the rotaxane. Since ClO4- does not bind tightly to a rotaxane
and thus position its wheel, the wheel does not participate
favorably in the binding of its cation. In our search for hosts
1
binding of Kþ promotes the binding of Naþ. Finally, H
NMR analysis revealed that the wheel is positioned further
away from the dialkylammonium ion for ArBG-R than
CEBG-R1. On the basis of these findings, binding modes
were constructed as shown in Figure 6. The wheel-pocket
was positioned on the edge of the amide-pocket for CEBG-
R1, according to the 1H NMR analysis. Possibly, the B18C6-
pocket keeps the wheel away from the amide-pocket through
steric hindrance or repulsion between the electron-rich
crown ethers. Kþ, being larger, appears to fit between the
amides of the amide-pocket better than Naþ. It would not
J. Org. Chem. Vol. 75, No. 10, 2010 3367

Wang et al.
JOCArticle
that bind cations, we constructed hosts that not only bind
cations but also can form a tight ion pair with Cl- in chlo-
roform, giving an allosteric relationship that drives cation
binding. Allosteric relationships between cations and anions
have recently been found to exist with some ligand-gated ion
channels.49 The next generation of rotaxanes is being desig-
ned to further investigate allosteric relationships and to selec-
tively bind targeted cations. Materials that operate through
allosteric relationships could become integral components
in future protein mimetics, cation sensors, or mechanical
devices.
for Naþ and Kþ, respectively. These assays were duplicated.
Any solvent effects were eliminated by subtracting background
levels of absorbance obtained from assays performed without
the presence of a host or metal cation. The absorbance values
were converted to concentrations using a series of calibration
curves, which were constructed prior to the assays. Percent
transfer (%T) was determined by solving the following equa-
tion. The values give the percentage of hosts that contained a
metal cation.
½metal cationꢀCHCl ,t ¼24h
3
%T ¼
ꢁ 100%
½hostꢀ
CHCl3,t ¼24h
Procedure for Determining the Effect of a Salt on the 1H NMR
Chemical Shifts of a Host. Stock solutions were made of
CEBG-R1 and ArBG-R (20 mM) in CHCl3, the metal chloride
salts (200 mM or 1.0 M) in distilled water, and the metal
perchlorate salts (100 mM) in distilled water. Extraction assays
were performed by combining 0.20 mL of a host solution and
0.20 mL of a guest solution in a vial and then vigorously stirring
the solution for 24 h. The CHCl3 layer was carefully removed
and evaporated. The residue was dissolved in 0.4 mL CDCl3 and
Experimental Section
General Experimental. Solvents, reagents, and materials
used to create the materials were purchased. Moisture-sensitive
reactions were carried out under positive argon pressure. All
organic solvents were freshly distilled under vacuum over a
suitable drying agent. Some rotaxanes were purified via HPLC
using a C18 reversed phase column with water (0.1% TFA)/
CH3CN as the eluent. Concentrations of metal ions were
determined using an atomic absorption spectrometer. 1H
NMR and 13C NMR spectra were measured in CDCl3 or
DMSO-d6 using a spectrometer operating at 400.14 MHz
for proton and 100.23 MHz for carbon nuclei. Chemical
shifts are in ppm and are referenced against an internal TMS
standard.
Determining Association Constants and %T for Host and
Metal Picrate Complexes. Stock solutions were made of the
hosts (2.0 ꢁ 10-3 M) in CHCl3 and the metal hydroxides or
metal oxides (5.0ꢁ10-2 M) with picric acid (7.0ꢁ10-5 M) in
water. Extraction assays were performed by combining 1.0 mL
of each solution in a vial and then vigorously stirring the
solutions for 2 h. We determined the system had reached
equilibrium within 2 h. A small aliquot of each CHCl3 layer
(0.10 mL) was carefully removed, and the CHCl3 was re-
moved in vacuo. The resulting material was dissolved in CH3CN
(1.0 mL), according to the procedure developed by Cram,43
and the absorption of the solution at 380 nm was measured
via UV-vis spectroscopy. The concentrations of the metal
picrates were determined by solving the Beer-Lambert equa-
tion using these absorption values and the known extinction
coefficients. The concentrations of the picrate ion in the aqueous
phases were also determined using the same method to verify
their concentrations. Each experiment was performed three
times, and the observed absorbance values were averaged. Kd
values were taken from the literature.40,44 Ka and %T values
were obtained by solving the equations presented earlier in the
paper.
%T Experiments for Metal Chloride salts, Metal Perchlorate
Salts, and Metal Chloride/Perchlorate Salts. Stock solutions
were made for the hosts (5 mM) in CHCl3, the metal chloride
salts (200 mM or 1.0 M) in distilled water, the metal perchlorate
salts (100 mM) in distilled water, and the metal chloride/
perchlorate saltes (100 mM/100 mM) in distilled water. For
hosts with carboxylic acids, 2 equiv of (Me)4NOH was added
from a stock solution. Extraction assays were performed by
combining 0.20 mL of a host solution and 0.20 mL of a guest
solution in a vial and then vigorously stirring the solution for
24 h. A small aliquot of each CHCl3 layer (0.10 mL) was carefully
removed, extracted with 1.0 mL of 0.1 N HCl for 2 h, and finally
diluted with distilled water to 10 mL. The amount of metal ion in
this final solution was determined by atomic absorption spec-
troscopy. The absorbance maxima were read at 589 and 766 nm
1
its 1D and 2D (COSY and ROESY) H NMR spectra were
recorded.
Synthesis of Materials. Amide-B18C6. Boc-β-alanine (1.0 g,
5.3 mmol) and 1,10-carbonyldiimidazole (1.0 g, 6.1 mmol) were
added to CH2Cl2 (15 mL). The reaction mixture was refluxed
for 3 h and then exposed to NH2-benzyl-18-crown-6 ether 143
(1.73 g, 5.25 mmol) dissolved in CH2Cl2 (5 mL). The reaction
mixture was refluxed for 0.5 h and then stirred at room
temperature for 24 h. The solvent was removed under reduced
pressure, and the residue was extracted with EtOAc/H2O. The
organic layer was collected, dried with MgSO4, and evaporated
under vacuum. The crude material was separated by flash
chromatography on silica gel with EtOAc as the eluent.
Amide-B18C6 (2.5 g, 5.0 mmol) was obtained as a white solid
in a 95% yield. 1H NMR (CDCl3): δ 1.43 (9H, s), 2.56 (2H, t),
3.46 (2H, t), 3.62-3.80 (12H, m), 3.89 (4H, m), 4.12 (4H, m),
5.32 (1H, s), 6.75 (1H, d), 6.92 (1H, d), 7.33(1H, s), 8.24 (1H, s).
13C NMR (CDCl3): δ 168.9, 155.2, 147.8, 144.1, 131.5, 113.6,
111.4, 105.9, 78.5, 69.7, 69.6, 69.5, 68.6, 68.4, 67.7, 35.9, 35.5,
27.3. MS: found 499.2730, calcd for C24H39N2O9 [M þ Hþ]
499.2656.
β-Alanine-amide-benzyl-18-crown-6 Ether 2. Amide-B18C6
(2.5 g, 5.0 mmol) was dissolved in 1:4 TFA/CH2Cl2 (15 mL).
After the above solution was stirred for 2 h, the volatile materials
were removed under reduced pressure. The residue was sepa-
rated by flash chromatography on silica gel with 99:5 CH2Cl2/
MeOH as the eluent. β-Alanine-amide-benzyl-18-crown-6 ether
2 (1.9 g, 4.8 mmol) was obtained as a pink oil in a 96% yield. 1H
NMR (CDCl3 þ DMSO-d6): δ 2.50 (1H,s), 2.66 (2H, t), 3.09
(2H, t), 3.41-3.65 (12H, m), 3.74 (4H, m), 4.35 (4H, s), 6.87 (1H,
d), 7.10 (1H, d), 7.30 (1H, s), 7.87 (1H, s). 13C NMR (DMSO-d6):
δ 168.9, 147.9, 143.4, 135.2, 113.3, 112.9, 105.6, 72.3, 71.9, 71.6,
71.5, 71.1, 68.5, 37.0, 34.0. MS: found 399.2186, calcd for
C19H31N2O7 [M þ Hþ] 399.2131.
CEBG-R1. To a solution containing DCC-[2]rotaxane39
(0.64 mmol in 1 mL CHCl3) cooled to 0 °C was rapidly added
β-alanine-amide-benzyl-18-crown-6 ether 2 (26 mg, 0.64 mmol)
dissolved in 1:1 DMSO/CHCl3 (1 mL). The reaction mixture
was brought to room temperature and stirred overnight. The
solvent was removed under vacuum. CH3CN (20 mL) was added
to the solution, and the resulting precipitated dicyclohexy-
lurea (DCU) was removed by filtration. After the solvent was
removed under reduced pressure, the residue was dissolved in
CH2Cl2 (1 mL). Diethyl ether (20 mL) was added to remove
N-(3,5-di-tert-butylbenzyl)valerolactam, a byproduct, as a
precipitate. After decanting the ether layer, the precipitation
(49) Chaudhry, C.; Plested, A. J. R.; Schuck, P.; Mayer, M. L. Proc. Natl.
Acad. Sci. U.S.A. 2009, 106, 12329–12334.
3368 J. Org. Chem. Vol. 75, No. 10, 2010

Wang et al.
JOCArticle
procedure was repeated. The ether layers were collected and
dried over MgSO4. After the solvent was removed in vacuum,
the crude material was separated via column chromatography
with 95:5 CH2Cl2/MeOH as the eluent.. CEBG-R1 3 (470 mg,
0.34 mmol) was obtained as a yellow glass in a 53% yield. Note:
to be consistent with the other rotaxanes, CEBG-R1 was sub-
jected to HPLC conditions prior to performing extraction
assays. 1H NMR (CDCl3): δ 1.30 (18H, s), 1.50 (18H, s),
1.51-2.00 (10H, m), 2.01-2.20 (12H, m), 2.30-2.40 (4H, m),
2.80-2.95 (2H, m), 3.00-3.5 (10H, m), 3.52-3.78 (16H, m),
3.80 (24H, s), 3.81-4.15 (16H, m), 4.59 (2H, s), 6.40 (8H, s),
6.60-6.65 (2H, m), 6.80-6.85 (2H, m), 7.01-7.10 (2H, m),
7.23-7.46 (5H, m). 13C NMR (CDCl3): δ 172.5, 171.5, 162.2,
153.3, 151.2, 151.5, 147.5, 142.9, 142.7, 133.8, 133.5, 131.7,
124.4, 123.2, 118.6, 115.5, 113.1, 111.6, 104.9, 79.9, 70.3, 70.2,
70.0, 69.7, 69.0, 68.8, 68.5,68.1, 53.5, 45.8, 45.4, 34.7, 33.9, 31.3,
28.2, 27.8, 25.6, 22.0. MS: found 1378.7896, calcd for
C73H112N5O20þ [Mþ Hþ] 1378.7943.
116.5, 115.0, 111.7 106.3, 71.1, 71.0, 70.4, 70.2, 70.0, 69.6,
68.8, 66.4, 55.5, 49.4, 47.2, 40.9, 34.8, 34.7, 33.7,þ31.4, 24.9 23.1.
MS: found 1369.8661, calcd for C69H102N5O23 [M þ H3Oþ]
1369.8547.
CEBG-R3. NH2-CEBG-R 3 (120 mg, 0.10 mmol) was dis-
solved in 1:1 CH2Cl2/DMSO (15 mL). To this solution was
added glutaric anhydride (23 mg, 0.20 mmol) followed by Et3N
(15 μL, 0.11 mmol). The reaction mixture was stirred overnight
at room temperature. After the solvent mixture was removed
under reduced pressure, the residue was separated by HPLC to
yield CEBG-R3 (110 mg, 0.04 mmol) as a yellow oil in a 80%
1
yield. H NMR (CDCl3): δ 1.14 (18H, s), 1.31-1.51 (2H, m),
1.74-1.85 (6H, t), 2.21-2.35 (8H, m), 3.08 (8H, m), 3.17-3.35
(4H, m), 3.36-3.7 (30H, m), 3.71-3.89 (8H, m), 3.98-4.2 (10H,
m), 4.57 (2H, s), 6.84-7.01 (3H, m), 7.04-7.16 (3H, m),
7.15-7.27 (2H, m), 7.30-7.36 (2H, m) 7.4-7.51 (2H, m) . 13
C
NMR (CDCl3): δ 175.7, 172.8, 171.9, 169.9, 151.9, 148.2, 148.1,
144.7, 144.4, 134.9, 134.8, 134.1, 132.88, 125.3, 124.1, 113.2,
112.7, 112.5, 106.0, 105.4, 71.5, 71.1, 70.9, 70.5, 69.7, 69.6, 69.1,
67.9, 53.4, 49.6, 47.0, 37.2, 36.7, 36.5, 34.7, 34.4, 32.6, 3þ2.4,26.8
22.0. MS: found 1406.7620, calcd for C73H110N5O23 [M þ
H3Oþ] 1406.7487.
NH2-CEBG-R 3. CEBG-R1 (150 mg, 0.11 mmol) was dissol-
ved in a solution of 1:5 TFA/CH2Cl2 (5 mL). The reaction mix-
ture was stirred at room temperature for 2 h. After the solvent
was removed in vacuum, NH2-CEBG-R 3 was obtained as a
yellow glass in a 99% yield. 1H NMR (CDCl3): δ 1.09 (18H, s),
1.43-1.78 (8H, m), 3.02-3.31 (12H, m), 3.38-3.89 (24H, m),
4.00-4.23 (10H, m), 4.60 (2H, s), 6.88-6.97 (3H, m), 6.98-7.09
(4H, m), 7.10-7.41 (5H, m). 13C NMR (DMSO-d6): 174.2,
170.4, 154.9, 152.0, 150.1, 149.4, 148.9, 148.0, 132.6, 126.5, 126.0,
125.3, 124.3, 116.9, 115.9, 114.0, 110.2, 109.2, 71.7, 71.4, 71.1, 70.7,
70.5, 70.3, 70.1, 69.5, 55.9, 49.6, 46.8, 35.7, 34.5, 33.7, 32.0, 26.5,
22.9. MS: found 1178.6954, calcd for C63H96N5O16þ [M þ Hþ]
1178.6847.
MeO2C-CEBG-R 4. Monomethylmalonate potassium salt
(32 mg, 0.20 mmol) and 1,10-carbonyldiimidazole (33 mg,
0.20 mmol) were dissolved in CH2Cl2 (15 mL). The reaction
mixture was refluxed for 3 h and then treated with a solution
of NH2-CEBG-R 3 (120 mg, 0.10 mmol) in DMSO (2 mL).
The reaction mixture was refluxed for 20 min and then stirred
at room temperature for 24 h. The solvent was evaporated
under reduced pressure, and the crude material was extracted
with EtOAc/H2O. The organic layer was collected and dried
with MgSO4, and then the solvent was removed under vacuum.
The residue was separated by flash chromatography on silica
gel using 8:92 MeOH/CH2Cl2 as the eluent. MeO2C-CEBG-
R 4 (110 mg, 0.080 mmol) was obtained as a yellow oil in a
78% yield. 1H NMR (CDCl3): δ 1.17 (18H, s), 1.22-1.44
(6H, m), 1.80-1.93 (2H, m), 3.18-3.82 (44H, m), 3.92-
4.19 (20H, m), 4.57 (2H, s), 6.29-6.45 (2H, m), 6.5-6.62
(2H, m), 6.79-6.94 (2H, m), 7.11-7.42 (6H, m). 13C NMR
(CDCl3): δ 170.4, 168.7, 157.9, 154.6, 151.6, 148.7, 147.5, 144.0,
137.0, 134.9, 132.7, 131.7, 130.3, 124.4, 123.4, 121.1, 117.2,
112.3, 107.4, 105.8, 70.7, 70.6, 70.0, 69.7, 69.4, 68.7, 68.5, 68.3,
52.6, 52.5, 49.0, 41.3, 40.9, 36.4, 34.9, 31.4, 26.þ6, 25.3, 21.8. MS:
found 1397.7711, calcd for C71H106N5O23 [M þ H2Oþ]
1397.7357.
Boc-Axle. 5-[tert-Butoxycarbonyl(3,5-di-tert-butylbenzyl)-
amino]pentanoic acid 638 (0.92 g, 2.2 mmol) was combined
with DCC (0.87 g, 4.2 mmol) in CHCl3 (5 mL). The resulting
solution was stirred for 2 h at room temperature. 3-Amino-
N-(3,5-dimethylphenyl) propanamide 5 (0.40 g, 2.1 mmol)
was added to the solution, which was stirred overnight. CHCl3
was partially evaporated under reduced pressure. DCU was
precipitated with the addition of CH3CN, which was removed
by filtration. The mother liquor was removed in vacuum, and
the residue was purified via flash chromatography on silica gel
with 3:97 MeOH/CH2Cl2 as the eluent. Boc-Axle (1.24 g, 2.10
mmol) was obtained as a white glass in a 95% yield. NMR
(CDCl3): δ 1.31 (18H, s), 1.45 (9H, s), 1.68 (2H, m), 1.78 (2H, m),
2.26 (6H, s), 2.48 (2H, t), 2.58 (2H, t), 3.21 (2H, t), 3.57 (2H, t),
4.60 (2H, s), 6.73 (1H, s), 7.03 (2H, s), 7.19 (2H, s), 7.30 (1, s),
8.22 (1H, s), 8.47 (1H, s). 13C NMR (CDCl3): δ 172.6, 168.9,
155.1, 150.0, 137.5, 135.1, 125.0, 121.0, 120.4, 120.2, 116.7, 78.7,
49.4, 46.2, 36.2, 33.8, 32.9, 31.4, 30.5, 27.5, 23.9, 22.0, 20.4. MS:
found 616.4095, calcd for C36H55N3O4Naþ [M þ Naþ]
616.4091.
Axle-1. Boc-Axle (1.4 g, 2.1 mmol) was treated with 1:4 TFA/
CH2Cl2 (15 mL) for 2 h. After the volatile materials were
removed via evaporation under reduced pressure, the material
was separated via column chromatography using a 2:98 MeOH/
CH2Cl2 as the eluent. Axle-1 (0.89 mg, 1.8 mmol) was obtained
1
as a colorless oil in an 86% yield. H NMR (CDCl3): δ 1.28
(18H, s), 1.63 (2H, m), 1.76 (2H, m), 2.14 (1H, s), 2.23 (6H, s),
2.60 (2H, t), 2.86 (2H, t), 3.32 (2H, t), 3.51 (2H, t), 4.61(2H, s),
6.75 (1H, s), 7.10 (2H, s), 7.22 (2H, s), 7.36 (1, s), 8.90 (1H,s), 9.00
(1H, s). 13CNMR(CDCl3):δ172.5, 170.9, 152.1, 138.7, 137.2, 134.9,
126.7, 123.9, 121.8, 118.5, 51.5, 47.5, 36.1, 34.8, 33.1, 31.4, 31.2,
þ
25.2, 24.6, 21.2. MS: found 494.3734, calcd for C31H48N3O2
CEBG-R2. MeO2C-CEBG-R 4 (110 mg, 0.080 mmol) was
dissolved in MeOH (20 mL), and (CH3)4NOH 5H2O (40 mg,
[M þ Hþ] 494.3747.
ArBG-R. ArBG-R was synthesized and purified following
the methods used to obtain CEBG-R1 except that 3-amino-
N-(3,5-dimethylphenyl)propanamide (38 mg, 0.20 mmol)
was used as the starting material. ArBG-R (99 mg, 0.084 mmol)
was obtained as a colorless oil in a 42% yield. Note: to be
consistent with the other rotaxanes, ArBG-R was subjected
to HPLC conditions prior to performing extraction assays. 1H
NMR: 1.27 (18H, s), 1.50 (18H, s), 1.71-2.03 (6H, m), 2.56-
2.69 (2H, m), 2.79-2.96 (3H, m), 3.08-3.27 (3H, m), 3.28-3.83
(15H, m), 3.95-4.13 (3H, m), 4.16-4.42 (4H, m), 4.57 (2H, m),
6.69 (1H, m), 6.85 (2H, m), 6.87 (1H, m), 7.17 (1H, m), 7.22 (2H,
m), 7.30 (2H, m), 7.39 (1H, m), 7.40 (1H, m), 7.56 (1H, m), 7.74
(1H, m), 8.06 (1H, s), 8.35 (2H, s), 8.37 (1H, m). 13C NMR:
3
0.2 mmol) was added. The reaction mixture was stirred over-
night at 50 °C. The solvent was evaporated under reduced
pressure, and the residue was extracted with EtOAc/1 N HCl.
The organic layer was collected, dried with MgSO4, and evapo-
rated under vacuum. The residue was separated by HPLC to
yield CEBG-R2 (105 mg, 0.076 mmol) as an orange foam in a
95% yield. 1H NMR (CDCl3): δ 1.17 (18H, s), 1.22-1.45 (6H,
m), 1.81-1.94 (2H, m), 3.32-3.58 (8H, m), 3.62-3.87 (38H, m),
3.93-4.21 (14H,m), 4.58 (2H,s), 6.51-6.63 (2H,m), 6.66-6.81
(2H,m), 6.88-7.05 (2H,m), 7.13-7.43 (6H,m). 13C NMR
(CDCl3): δ 171.1, 169.9, 157.8, 154.5, 151.0, 148.7, 143.9,
142.7, 136.4, 136.1, 135.4, 133.5, 130.9, 122.2, 121.8, 117.1,
J. Org. Chem. Vol. 75, No. 10, 2010 3369

Wang et al.
JOCArticle
171.6, 169.1, 167.0, 156.2, 150.2, 146.3, 142.0, 137.2, 137.1,
132.1, 127.7, 124.5, 123.3, 122.2, 116.6, 112.2, 110.5, 79.2,
69.3, 68.9, 68.7, 51.6, 47.6, 35.9, 35.4, 34.1, 33.7, 31.4, 30.3,
27.2, 24.8, 23.1. MS: found 1172.7792, calcd for C65H98N5O14
[M þ Hþ] 1172.7111.
(Grant No. CHE-0400539) for funding this research
program.
þ
Supporting Information Available: 1D and 2D NMR spec-
tra, postulated binding modes, chemical shift differences for
hosts with TFA- versus Cl- as counterions, %T values. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Acknowledgment. The authors thank the University
of Cincinnati and the National Science Foundation
3370 J. Org. Chem. Vol. 75, No. 10, 2010