Guest covalent capture by a host: A biomimetic strategy for the selective functionalization of a cavity

Article abstract of DOI:10.1002/chem.201202391

A biomimetic strategy for the monofunctionalization of a calix[6]arene core is described. It is based on host-guest chemistry (mimicking the Michaelis-Menten adduct in enzymes) and allows the finely tuned pre-organization of the substrate (an alkyne) with respect to the reactant (three azido groups introduced at the calixarene large rim). It is shown that the thermal Huisgen reaction implemented in this work proceeds under very mild conditions with total regioselectivity of the cycloaddition process. The scope of the reaction was explored and the results suggest that such a supramolecular strategy is quite versatile and could be applied to the selective functionalization of other cavitands bearing different recognition patterns. A detailed structural, thermodynamic, and kinetic study is also reported, highlighting interesting biomimetic features: The importance of the host-guest adduct strength, the high sensitivity of the reaction to the pre-organization of the reactive partners (alkyne vs. azide), and a significant impact of the embedment on the transition state. The self-coordination of the monofunctionalized products was also studied and an "endo/exo" switch of the internal side-chain could be triggered by adding competitive ligands. Copyright

Full text of DOI:10.1002/chem.201202391

DOI: 10.1002/chem.201202391
Guest Covalent Capture by a Host: A Biomimetic Strategy for the Selective
Functionalization of a Cavity
Nicolas Menard, Olivia Reinaud, and Benoit Colasson*[a]
Abstract: A biomimetic strategy for
the monofunctionalization of a calix[6]-
arene core is described. It is based on
host–guest chemistry (mimicking the
Michaelis–Menten adduct in enzymes)
and allows the finely tuned pre-organi-
zation of the substrate (an alkyne) with
respect to the reactant (three azido
groups introduced at the calixarene
large rim). It is shown that the thermal
Huisgen reaction implemented in this
work proceeds under very mild condi-
tions with total regioselectivity of the
cycloaddition process. The scope of the
reaction was explored and the results
study is also reported, highlighting in-
teresting biomimetic features: The im-
portance of the host–guest adduct
strength, the high sensitivity of the re-
action to the pre-organization of the
reactive partners (alkyne vs. azide),
and a significant impact of the embed-
ment on the transition state. The self-
coordination of the monofunctionalized
products was also studied and an
“endo/exo” switch of the internal side-
chain could be triggered by adding
competitive ligands.
suggest that such
a supramolecular
strategy is quite versatile and could be
applied to the selective functionaliza-
tion of other cavitands bearing differ-
ent recognition patterns. A detailed
structural, thermodynamic, and kinetic
Keywords: biomimetics
· calixar-
enes · host–guest systems · regiose-
lectivity · zinc
Introduction
proved to be more selective.[8] Only a few procedures for
the monofunctionalization of cucurbiturils are available.[9]
As far as resorcinarene receptors are concerned, the modifi-
cation of one wall out of four has been developed and ex-
tensively used by Rebek and co-workers.[10]
Selective desymmetrization is a classic challenge in organic
synthesis. The selective transformation of one functional
group among other identical groups is a difficult task and
often mixtures are generated.[1] Consequently, the desired
desymmetrized molecule is obtained in low-to-moderate
yield. The synthesis and purification of larger desymme-
trized molecules may be even more tedious. Nevertheless,
the need for such structures is often required when one
wants to introduce different reactive functionalities, add co-
operativity, or explore multivalency in host–guest systems.[2]
Highly symmetrical parent macrocycles such as resorcinar-
enes,[3] cucurbiturils,[4] cyclodextrins,[5] and calix[n]arenes[6]
are often encountered in supramolecular chemistry as mo-
lecular scaffolds for hosts, sensors, or catalysts. Reliable and
useful synthetic procedures allowing for the selective trans-
formation of one and only one reactive site of these macro-
molecules are scarce. The monofunctionalization of cyclo-
dextrins is often achieved in low-to-moderate yield because
of the need for chromatographic separation.[7] Very recently,
one procedure based on the covalent capture strategy
Synthetic procedures for the monofunctionalization of ca-
AHCTUNGTREGlNNNU ix[6]arene macrocycles are rare. We previously reported a
unique method for the selective monofunctionalization of a
tris-azidocalix[6]arene[11] through a combination of the ther-
mal Huisgen cycloaddition of azido compounds to acety-
lenes[12] and the host–guest properties of ZnII–calix[6]arene-
based funnel complexes.[13] 4-Pentyn-1-amine ((C3)NH2) can
coordinate to the ZnII ion, placing the acetylene in close vi-
cinity to the three azido groups, thus pre-organizing the re-
actants. The cycloaddition occurs upon heating the complex
2+
to give the monofunctionalized complex [ZnM(C )NH
]
3
2
(Scheme 1).[14] Subsequent cycloaddition reactions are pre-
vented for two reasons: 1) Under such conditions, the inter-
molecular cycloaddition reaction is much slower as it re-
quires much more drastic conditions (higher temperature
and concentrations) and 2) the cavity of the calixarene is
now occupied, preventing a second molecule of 4-pentyn-1-
amine from displacing the now covalently linked amino
chain and reacting intramolecularly.[15] This reaction is
100% selective towards monofunctionalization and creates
exclusively the 1,5-triazole regioisomer.[11]
[a] Dr. N. Menard, Prof. Dr. O. Reinaud, Dr. B. Colasson
Laboratoire de Chimie et Biochimie Pharmacologiques et
Toxicologiques, CNRS UMR 8601
In this example, the “monoclick” reaction occurs because
of the geometric fit between the acetylene and the azido
groups. We intended to explore the scope of this methodolo-
gy. To do so, we studied a variety of substrates that can be
used to monofunctionalize the calixarene precursor. The
acetylene guest can be constructed from three modifiable
Universitꢀ Paris Descartes, PRES Sorbonne Paris Citꢀ
45 rue des Saints Pꢁres, 75006 Paris (France)
Fax : (+33)142-868-387
E-mail: benoit.colasson@parisdescartes.fr
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201202391.
642
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 642 – 653

FULL PAPER
al modifications. In the second
part, this synthetic methodology
has been applied to alcohols.
The differences and similarities
between the reactions with the
amines and alcohols will be em-
phasized. In particular, the ther-
modynamics of the intramolec-
ular coordination of the hy-
droxy arm will be discussed.
Scheme 1. Reaction pathway for the monoclick reaction with 4-pentyn-1-amine (the structures of the com-
plexes have been simplified for clarity).
Monoclick reactions with pri-
mary amines:
Effect of spacer length: The
spatial proximity of the acety-
lene and the three azido groups
appears to be one of the keys
to monofunctionalization. The
rate of reaction should be en-
hanced if the acetylene of the
coordinated guest faces precise-
ly the azido groups present on
the large rim. Thus, we antici-
pated the existence of an opti-
mal length for the spacer. To
Scheme 2. Schematic representation of the monoclick reaction with a reactive guest bearing a coordinating
function X, a spacer, the reactive acetylenic group, and a substituent R. The coordinating chain of the mono-
functionalized complex can be either endo or exo with respect to the cavity, depending on the nature of X, R,
the spacer, and the nature of the competitive ligand L.
parts: The coordinating group X (primary amine, alcohol), a
spacer that can vary in length and in shape between X and
the reactive acetylene function, and finally a terminal func-
tionality R that can vary in size or in electronic properties
(Scheme 2). We then investigated the relationship between
the thermodynamically controlled complexation step and
the kinetics of the cycloaddition step.
optimize our system, we studied three different alkylamines
that differ only in the number n of methylene groups in the
ꢀ
spacer: HC C
amine ((C3)NH2), 5-hexyn-1-amine ((C4)NH2), and 6-
heptyn-1-amine ((C5)NH2).
G
Synthesis and characterization of the three monoclick com-
The monoclick reaction of such a guest provides a novel
calix[6]arene-based ligand MCnX (Scheme 2), which differs
from classical “funnel complexes” as an additional coordi-
nating group X is covalently linked to the large rim of the
calixarene. Such a system is prone to undergo self-coordina-
tion and offers an interesting example of an “ouroboros-
like” molecule as defined by Durola and Rebek.[16] Hence,
in addition to methodological studies yielding these mono-
functionalized calixarene complexes, we wish to describe
and discuss their specific host–guest properties. In this re-
spect, we have studied the structural and conformational
changes accompanying self-coordination and the factors
controlling the intramolecular coordination to the zinc
cation, that is, the “endo/exo” equilibrium of the internal co-
ordinating arm.
plexes: The three pre-click host–guest adducts HGn were
synthesized in situ from [ZnX6N3]ACHTNUTRGNEG(UN ClO4)2 and the corre-
sponding amines (see Figure 1) and characterized by
1H NMR spectroscopy (see Figure 2A, C, and E) in CD3CN.
1
All the H NMR spectra exhibit the same pattern and chem-
ical shifts except for the resonances in the high-field region,
which attests to the coordination of the amine to the metal
center and its encapsulation in the cavity of the complex.
The monoclick reactions were conducted in toluene at
reflux with two equivalents of amine to provide the corre-
sponding monofunctionalized ZnII complexes in high yields
(95, 93, and 92% for [ZnM(C )NH ]AHCNUTGETRNGNNU(ClO4)2, [ZnM(C )NH ]-
3
2
4
2
ACHNUTTGRENN(GUN ClO4)2, and [ZnM(C )NH ]ACHTUNGERTG(NNUN ClO4)2, respectively). The reac-
2
5
tions were completed within 2 hours for n=3 and 4, and
5 hours for n=5. The products were isolated by precipita-
tion with Et2O and centrifugation or filtration, depending
on the scale of the reaction.
Results and Discussion
The products were all characterized by positive electro-
spray mass spectroscopy (EIMS) after demetalation of the
samples by addition of trifluoroacetic acid (TFA). In all
cases, a single peak was observed corresponding to the pro-
tonated monofunctionalized ligands [MC NH +H]+. Neither
the bis- nor tris-adduct was ever observed, which proves
that the monofunctionalization reaction is extremely selec-
tive in all three cases. The products were characterized by
First, we evaluated the scope of the reaction by using pri-
mary amine ligands, which are the best guests for Zn–calix-
arene funnel complexes. The strong association to the zinc
metal center provides unambiguous characterization of the
intermediates and final compounds and facilitates discussion
about the factors controlling reaction efficiency and structur-
n
2
Chem. Eur. J. 2013, 19, 642 – 653
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
643

B. Colasson et al.
the aromatic region attested to
the formation of the triazole
moiety. New upfield resonances
appeared; the integration of
these resonances indicates
guest/calixarene ratio of 1:1,
which is consistent with
monofunctionalized calixarene
a
a
AHCTUNGTRENNUNG
in which the amino chain is co-
ordinated to the ZnII ion
through the cavity. The reso-
nances of the monofunctional-
2+
ized complexes [ZnM(C )NH
]
3
2
2+
and [ZnM(C )NH
]
are sharp at
4
2
300 K, whereas the resonances
of the monofunctionalized com-
plex [ZnM(C )NH ]
2+, especially
5
2
the resonances of the imidazole
arms ImCH2, are broad at
300 K. However, at 353 K, the
resonances sharpen. This indi-
2+
cates that [ZnM(C )NH
]
has a
5
2
restricted conformational mo-
bility at 300 K that is slow on
A
ACHTUNGTRENNUNG
the NMR timescale. Because
this phenomenon was not ob-
2+
served for [ZnM(C )NH
]
or
3
2
[ZnM(C )NH ]
2+, we can conclude
4
2
Figure 1. Top: Reaction pathway for the monoclick reactions of 4-pentyn-1-amine (n=3), 5-hexyn-1-amine
(n=4), and 6-heptyn-1-amine (n=5). The corresponding complexation constant and the rate constant of the
that the longer alkyl chain im-
poses some distortion on the
calixarene that hinders its con-
intramolecular cycloaddition are denoted as Kn and kn, respectively. Bottom: Representation of the planar pro-
2+
jection of the monofunctionalized complex [ZnM(C )NH
]
emphasizing the vertical symmetry plane sv. For
4
2
clarity, the bond between the amino group and the ZnII ion is not represented and the alkyl chain is represent-
formational
mobility
(see
ed out of the symmetry plane.
below). In comparison with the
pre-click inclusion complexes,
the 1H NMR spectra display a
higher number of signals, con-
sistent with
a symmetry de-
crease
(Figure 1).
from
C3v
Although
to
Cs
the
ImCH2 protons resonate as a
singlet in the inclusion com-
plexes, after reaction they split
into three signals: One singlet
and two doublets experiencing
2
a J coupling. This is consistent
with the existence of a sv sym-
metry plane, which differenti-
ates the ImCH2 that belongs to
the sv plane (one singlet inte-
grating for 2H) and two pairs
of protons (two doublets inte-
grating for 4H). Likewise, the
initial HArN3 signal is a singlet.
After desymmetrization, the six
protons are differentiated into
Figure 2. 1H NMR spectra in CD3CN of inclusion complexes HGn and the corresponding monofunctionalized
2+
2+
complexes [ZnMCnNH2
(500 MHz, 300 K), D) [ZnM(C )NH
(500 MHz, 300 K), and G) [ZnM(C )NH
]
(n=3, 4, 5): A) HG3 (500 MHz, 300 K), B) [ZnM(C )NH
]
(250 MHz, 300 K), C) HG4
3
2
2+
2+
]
(500 MHz, 300 K), E) HG5 (500 MHz, 300 K), F) [ZnM(C )NH ]
(500 MHz, 353 K) (f: free amino-acetylene; s: residual solvent; the
4
2
5 2
2+
]
5
2
methylene groups of the alkyl chain are denoted as a, b, g, d, and e, starting from the NH2 group).
1H NMR spectroscopy (Figure 2B, D, and G), evidencing
further the high selectivity of the reaction. A new singlet in
three singlets integrating for two protons each: One singlet
for HArtria and two singlets for the remaining HArN3. The
644
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 642 – 653

Guest Covalent Capture by a Host
FULL PAPER
initial tBu singlet is also split into two singlets integrating
for 9H (for the tBu that belongs to the symmetry plane)
and 18H (for the two tBu on either side of the symmetry
plane). In the same manner, the methoxy groups (OCH3)
are split from one singlet (9H) into two singlets integrating
for 3H and 6H, respectively. This split is a general trend for
all the other groups of protons (NCH3, ImH) for which one
set of protons belongs to the symmetry plane and the other
two sets are on each side of the plane.
thalpically driven by the formation of a bond between NH2
and the metal center.[18] The moderate selectivity of the host
complex for (C4)NH2 is probably controlled by secondary
CH–p interactions between the methylene groups and the
calixarene cavity.
We then studied how the difference in the pre-organiza-
tion of the host–guest inclusion complexes HGn influences
the rate of the cycloaddition reaction by means of competi-
tive monoclick reactions. A known mixture of host–guest in-
clusion complexes HG3 and HG4 and a large excess of the
corresponding amines in toluene were heated at 1108C for
2 hours and, after isolation, the resulting mixture of mono-
Optimization of the chain length: All three monofunctional-
ized complexes were obtained in high yields after reaction
for 2 hours for (C3)NH2 and (C4)NH2, and 5 hours for
(C5)NH2. We wondered what was the impact of the length
of the spacer on the kinetics of the reaction. To answer this
question, two competitive monoclick reactions were carried
out: (C3)NH2 versus (C4)NH2 and (C4)NH2 versus (C5)NH2.
The monoclick reaction is a two-step process: Inclusion of
the amino guest followed by intramolecular Huisgen cyclo-
addition. To fully understand the origin of the anticipated
difference in the reaction rates, the thermodynamic and ki-
netic parameters of these two steps were analyzed (see
Figure 1 for the definition of the parameters Kn and kn).
The competitive complexation studies were carried out in
CD3CN at 300 K. A known mixture of 4-pentyn-1-amine
(n=3) and 5-hexyn-1-amine (n=4) was added to a solution
of precursor complex [ZnX6N3]2+ in CD3CN to yield quanti-
tatively a mixture of inclusion complexes HG3 and HG4
(Figure 3C), identified by comparison with the pure inclu-
2+
2+
functionalized calixarenes [ZnM(C )NH
]
and [ZnM(C )NH ]
3
2
4
2
was analyzed by 1H NMR spectroscopy in CD3CN (Fig-
ure 3D–F). The kinetic law for each intramolecular reaction
is given by Equation (1).
vn ¼ kn½HGnꢂ ¼ knKn½CnNH2ꢂ½ZnX6N3ꢂ=½H2Oꢂ
ð1Þ
During the reaction, the ratio [(C3)NH2]/[(C4)NH2] varies
little and thus can be considered as a constant. Hence, we
can write Equation (2).
v4=v3 ¼ ½ZnMðC ÞNH ꢂ1=½ZnMðC ÞNH ꢂ1 ¼ k4½HG4ꢂ0=k3½HG3ꢂ0
4
2
3
2
ð2Þ
From this we calculate Equation (3) (for more details, see
the Supporting Information).
k4=k3¼ ð½ZnMðC ÞNH ꢂ1=½ZnMðC ÞNH ꢂ1Þð½HG3ꢂ0=½HG4ꢂ0Þ
4
2
3
2
ð3Þ
¼ ð1:25 ꢁ 0:10Þ
A
similar experiment was conducted with amines
(C4)NH2 and (C5)NH2, which provided k4/k5 =(2.28ꢁ0.20).
Therefore, the highest rate constant for the intramolecular
cycloaddition was reached with 5-hexyn-1-amine, which indi-
cates that the geometry of the corresponding host–guest
complex HG4 is the closest to the transition state. In this
case, the acetylene faces best the three azido groups. The
effect is more significant when (C4)NH2 is compared with
the longest chain (C5)NH2.
The trends in the values of the rate constants and the
complexation constants follow the same pattern. Hence, the
best ligand leads coincidentally to the highest reaction rate.
5-Hexyn-1-amine is unambiguously the best substrate for
this monoclick reaction due to the synergy between the two
phenomena. The synergy between thermodynamics and ki-
netics is often at work in enzymatic systems. It can be quan-
tified in the Michaelis–Menten model by the ratio kcat/KM,
kcat being the turnover number and KM the Michaelis con-
stant. When the assumption of a rapid equilibrium is fulfil-
led, KM is equal to the dissociation constant of the enzyme–
substrate complex.[19] In the system presented herein, the
ratio kcat/KM is equivalent to knKn. Calculation gives k4K4/
k3K3 =2.2 and k4K4/k5K5 =4.3. The highest value for k4K4 in-
dicates a better specificity of the transformation of the tris-
azido–calixarene complex with the (C4)NH2 substrate. Ac-
Figure 3. 1H NMR (CD3CN, 500 MHz, 300 K) spectra in the high-field
region. From bottom to top: A) HG3, B) HG4, C) HG4 +HG3 with
2+
10 equiv of
a
1:1 mixture of (C3)NH2/(C4)NH2, D) [ZnM(C )NH
]
+
3
2
2+
[ZnM(C )NH
]
after heating of mixture
C
in toluene at 1108C,
The resonances of the b- and g-
4
2
2+
E) [ZnM(C )NH
]
2+, and F) [ZnM(C )NH
]
.
3
2
4
2
methylene groups of the coordinated alkyl chain are indicated.
sion complexes (Figure 3A and B). Relative integration of
the upfield resonances of the coordinated guests provided
the relative ratio of the inclusion complexes, which, correct-
ed by the ratio of the free amines, provided the ratio of the
binding constants: K4/K3 =(1.81ꢁ0.20).[17] A similar compe-
tition experiment was conducted with (C4)NH2 and
(C5)NH2, which provided the ratio K4/K5 =(1.87ꢁ0.20) (for
more details, see the Supporting Information). Therefore
(C4)NH2 is the best amino-acetylene ligand of all three for
inclusion. The formation of the host–guest complexes is en-
Chem. Eur. J. 2013, 19, 642 – 653
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
645

B. Colasson et al.
Table 1. Chemical shifts (CD3CN, 500 MHz, 300 K[a]) of species investigated in this study.[b]
d [ppm]
a
b
g
d
e
tBu
OCH3
NCH3
3.65
ImCH2
5.04
HArN
5.91
ImH
6.88
ACHTUNGTRENNUNG
[ZnX6N3]2+
1.38
3.65
(C3)NH2
HG3
CIS3
2.79
0.74
1.64
2.24
ꢄ0.82
ꢄ2.46
ꢄ0.66
ꢄ0.13
ꢄ2.37
1.15
1.40
0.02
1.43 (9H)
1.34 (18H)
3.58
3.79
0.14
3.80 (3H)
5.31
0.27
6.17
0.26
6.00 (N3)
6.22 (N3)
6.03 (tria)
ꢄ0.17 (N3)
0.05 (N3)
ꢄ0.14 (tria)
6.93
0.05
6.84 (1H)
6.85 (2H)
ꢄ2.05
ꢄ0.07
2+
2+
2+
[ZnM(C )NH
]
]
]
0.64
3.80 (3H)
5.34 (s)
5.02 (d1)
5.23 (d2)
0.03 (s)
ꢄ0.29 (d1)
ꢄ0.08 (d2)
2
2
2
3.58 (6H)
3.76 (6H)
Dd3
ꢄ0.10
0.16
1.28
0.03 (9H)
ꢄ0.06 (18H)
0.22 (3H)
0.00 (6H)
0.01 (3H)
-0.03 (6H)
ꢄ0.09 (1H)
ꢄ0.08 (2H)
(C4)NH2
HG4
CIS4
2.7
0.75
1.55
ꢄ0.97
ꢄ2.52
ꢄ1.09
1.55
ꢄ0.73
ꢄ2.28
ꢄ0.25
2.2
1.22
1.39
0.01
1.41 (9H)
1.34 (18H)
3.57
3.79
0.14
3.81 (3H)
3.77 (6H)
5.34
0.30
6.21
0.30
6.93
0.05
6.87 (1H)
6.90 (2H)
ꢄ1.95
ꢄ0.98
ꢄ0.08
[ZnM(C )NH
0.78
1.57
3.76 (3H)
3.57 (6H)
5.37 (s)
5.18 (d1)
5.28 (d2)
0.03 (s)
ꢄ0.16 (d1)
ꢄ0.06 (d2)
6.10 (N3)
6.24 (N3)
6.28 (tria)
ꢄ0.11 (N3)
0.03 (N3)
0.07 (tria)
Dd4
0.03
ꢄ0.12
0.48
0.35
0.02 (9H)
ꢄ0.05 (18H)
0.19 (3H)
0.00 (6H)
0.02 (3H)
ꢄ0.02 (6H)
ꢄ0.06 (1H)
ꢄ0.03 (2H)
(C5)NH2
HG5
CIS5
2.71
0.75
1.55
ꢄ1.08
ꢄ2.63
ꢄ1.11
1.55
ꢄ0.92
ꢄ2.47
ꢄ0.66
1.55
0.52
2.17
1.80
1.40
0.02
1.42 (9H)
1.38 (18H)
3.56
3.80
0.15
3.82 (3H)
3.83 (6H)
5.34
0.30
6.22
0.31
6.92
0.04
6.92 (1H)
7.02 (2H)
ꢄ1.96
ꢄ1.03
ꢄ0.37
ꢄ0.09
[ZnM(C )NH
0.70
0.70
2.09
3.71 (3H)
3.62 (6H)
5.37[a] (s)
5.35[a] (d1)
5.42[a] (d2)
6.16[a] (N3)
6.18[a] (N3)
6.70[a] (tria)
Dd5
ꢄ0.05
ꢄ0.03
0.26
0.18
0.29
0.02 (9H)
ꢄ0.02 (18H)
0.15 (3H)
0.06 (6H)
0.02 (3H)
0.03 (6H)
0.00 (1H)
0.10 (2H)
[a] With the exception of the peaks corresponding to CH2Im, HArN3, and HArtria of [ZnM(C )NH ]
2+, which are broad at 300 K and were recorded at
5
2
353 K. [b] First row is precursor complex [ZnX6N3]2+. Each of the other three rows (n=3, 4, 5) is divided into five sub-rows: A) Free amine (n=3, 4, 5);
B) HGn (n=3, 4, 5); C) CIS (difference between dFreeAmine in and out, or for the calixarene: difference between HGn=3,4,5 and the initial complex;
2+
D) Monofunctionalized complex [ZnMCnNH2
]
(n=3, 4, 5); E) Ddn (difference between the chemical shifts of the monofunctionalized complexes
2+
[ZnMCnNH2
]
and the corresponding inclusion complex HGn. The methylene groups of the alkyl chain are denoted as a, b, g, d, and e, starting from the
NH2 group).
cordingly, n=4 corresponds to the optimized length for the
spacer.
to another, the difference in CIS is 0.05 ppm at most,
which indicates that the shielding effects are very similar.
Therefore all three guests are positioned at the same
height inside the calixarene cavity. Upon inclusion of the
guests, some resonances of the calixarene unit remain
almost unchanged. For instance, there is little effect of
the guest complexation on the tBu (ca. 0.01 ppm) or
ImH CIS (ca. 0.04 ppm). However, the resonances of
ImCH2 are different to those of the precursor complex,
but are almost identical for the three amino guests (CIS
ꢃ0.28 ppm). The same is true for methoxy groups
(CISꢃꢄ0.08 ppm) and HArN3 (CISꢃ0.28 ppm). These
data highlight a slight change in the conformation of the
calixarene upon complexation of the guests. However,
for all three, the conformational change is almost identi-
cal, which suggests that the difference in the reactivities
of the three amino-alkynes does not come from the dis-
tortion of the calixarene at the complexation step.
Origin of the observed specificity based on a conformational
analysis: As pre-organization is key for this reaction, we
studied the host–guest symmetry changes and the adaptation
process that both the calixarene host and the guest alkyl
chain must undergo to yield the monofunctionalized com-
plex. To this end we analyzed the conformations of both the
inclusion complexes and the monofunctionalized complexes
in CD3CN (see Figure 2 and Table 1).
1) Conformational comparison of the inclusion complexes
HGn: In all cases, the aromatic walls of the calixarene
cavity shield the amino guest upon complexation and
lead to an upfield shift of the resonances of the alkyl
chain.[13a,20] We defined the complex-induced shift (CIS)
value as the difference between the chemical shifts of a
group of protons in the inclusion complex HGn and in
the free ligand. The CIS values are reported in Table 1.
In all cases, Hb has the greatest CIS value, which means
that it is the most shielded. Nevertheless, from one guest
2) Conformational analysis of the monoclick products
[ZnMCnNH2]
2+: Unlike the inclusion complexes, there are
clear differences between the spectra of the monoclick
products, which indicates that the three products have
646
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 642 – 653

Guest Covalent Capture by a Host
FULL PAPER
significantly different conformations (Figure 2). These
differences can be observed by measuring the differences
between the chemical shifts of the monoclick product
and the corresponding inclusion complex (Ddn). To pic-
ture completely the influence of the chain length on the
structural properties, two conformational features need
to be discussed: Distortion of the calixarene cone and re-
stricted conformational mobility.
tively) and the Htria chemical shifts (7.33, 7.24, and 7.15 ppm
for n=5, 4, and 3, respectively) decrease significantly when
the spacer is shorter, which also indicates that the Artria aro-
matic unit rotates more towards the inside of the cavity
when the chain is shorter.
The rotation of the two remaining ArN3 units can be meas-
ured by the average chemical shift of their aromatic protons
HArN3. In all cases, the ArN3 units are in very similar posi-
tions (the average chemical shift for the HArN3 protons is
2+
6.11 ppm for [ZnM(C )NH
]
and 6.17 ppm for both
3
2
2+
Distortion of the calixarene cone: After the click reaction,
the ArtBu units are differentiated but in all three cases the
Dd values are very close to zero for all complexes. Therefore
the ArtBu units see little conformational change following
the cycloaddition step, which is consistent with the fact that
these aromatic units are projected away from the calixarene
cavity and are little affected by modifications to the inside
of the cavity (Figure 4). The aromatic unit that sees the
[ZnM(C )NH
]
and [ZnM(C )NH
]
2+). Therefore there is no sig-
4
2
5
2
nificant impact of the length of the spacer on the pivot of
the remaining ArN3 units.
The symmetry decrease from C3v to Cs gives rise to newly
differentiated protons that no longer resonate at the same
frequency because they perceive different environments.
The difference in the chemical shifts of protons that are no
longer equivalent is an indication of the difference in the en-
vironment of these protons. Therefore a greater splitting in
1
the H NMR spectrum is indicative of a greater distortion of
the calixarene cone. The splitting between the two sets of
methoxy groups is greater when the chain is shorter (0.09,
0.19, and 0.22 ppm for n=5, 4, and 3, respectively). This
phenomenon can also be observed for the two HArN3 signals
(0.02, 0.14, and 0.22 ppm for n=5, 4, and 3, respectively),
the two ImCH2 doublets (0.07, 0.10, and 0.21 ppm for n=5,
4, and 3, respectively), and, to a lesser extent, the two tBu
signals (0.04, 0.07, and 0.09 ppm for n=5, 4, and 3, respec-
tively). Therefore shorter chains accentuate the distortion of
the cone of the calixarene.
Resonance broadening and restricted conformational mobili-
ty: With n=5 only, a new phenomenon occurs: At 300 K,
the resonances of ImCH2, HArtria
,
and HArN3 of
2+
[ZnM(C )NH
for [ZnM(C )NH
]
are broad (Figure 2F) whereas they are sharp
5
2
2+
2+
Figure 4. Schematic representations of A) the pivot of the Artria unit by
an ap angle (side-view, the tBu and N3 groups have been omitted for
clarity) and B) the equilibrium of the helix inversion around the metal
center and the distortion of the calixarene structure (top view, the N3
groups have been omitted for clarity).
]
and [ZnM(C )NH
]
at the same tempera-
3
2
4
2
ture. In addition, the resonances of the ArCHax protons,
2+
2+
which are broad for [ZnM(C )NH
]
and [ZnM(C )NH ] ,
3
2
4
2
become even broader and can barely be seen in the spec-
trum (Figure 2F), which is indicative of a different confor-
mational mobility of [ZnM(C )NH ]
2+. Indeed, at 353 K, all
5
2
greatest environmental change is the Artria unit because it
becomes covalently linked to the coordinated amino chain.
In all three cases, the Ddn value for its methoxy group
(Table 1, column OCH3 (3H)) is positive. This indicates that
the methoxy groups are pushed away from the cavity forcing
the Artria unit to pivot towards the inside of the shielding
cavity (Figure 4A). Interestingly, Ddn for the methoxy group
increases as the alkyl chain becomes shorter (0.15, 0.19, and
0.22 ppm for n=5, 4, and 3, respectively), which indicates
that with a shorter chain, the aromatic unit pivots more to-
wards the inside of the cavity. Although the signs of the Ddn
values for HArtria can hardly be interpreted in terms of con-
formational change because the chemical transformation of
azido to triazole has a great influence on the chemical shifts,
their relative values can be compared. The Ddn values for
HArtria (0.48, 0.07, and ꢄ0.14 ppm for n=5, 4, and 3, respec-
these resonances are sharper (Figure 2G). This reflects the
fact that the longer chain imposes a constrained conforma-
tion on the calixarene structure and more precisely on the
coordination arms (through the coordination of the amino
group to the ZnII ion) and the Artria unit (through the tri-
2+
azole group). Thus, [ZnM(C )NH
]
has a restricted conforma-
5
2
tional mobility that impacts on the helix inversion (see
below).
Dynamics of the helix inversion at the metal center: We fo-
cused then on the influence of self-coordination on helix in-
version, a conformational change experienced at the ZnII
cation coordination sphere (Figure 4B).[21] To study the dy-
namics of such a process, we conducted 1H NMR experi-
ments at various temperatures in CD3CN and in CDCl3, a
noncoordinating solvent (see the Supporting Information).
Chem. Eur. J. 2013, 19, 642 – 653
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
647

B. Colasson et al.
ment for the monoclick reaction. The optimal four-atom
chain length ensures the best inclusion of the guest as well
the highest reactivity of the host–substrate complex. In addi-
tion, the nature of the spacer has a great impact on the con-
formation of the monofunctionalized complex and on its
conformational mobility.
Monofunctionalization with other amino guests: Having de-
termined that the optimal length for the spacer is n=4, we
decided to test other amino guests with four carbon atoms
between the amino and the acetylene groups. The benzyl-
AHCTUNGTREaGUNNN mine-based guest (ArCH2NH2) presents a more rigid
spacer. The other two substrates have the same aliphatic
spacer as 5-hexyn-1-amine and the acetylene is substituted
by p-methoxyphenyl (pMeOAr(C4)NH2) or p-nitrophenyl
(pNO2Ar(C4)NH2; Figure 6).
2+
Figure 5. 1H NMR spectra of monofunctionalized complex [ZnM(C )NH
(CD3CN, 500 MHz) at various temperatures (s: residual solvent).
]
2
4
2+
The 1H NMR spectrum of [ZnM(C )NH
]
in CD3CN (Fig-
4
2
ure 2D) is that of a Cs-symmetric species, which means that
at 300 K the helicoidal inversion is fast on the NMR time-
scale. The loss of Cs symmetry in favor of the C1-symmetric
species can be observed when the solution is cooled to
240 K (Figure 5). Indeed, there are six signals for the six
[HArN3 +HArtria] protons. There is also a clear splitting of
the b- and g-protons, which indicates that they have become
diastereotopic. The ImCH2 protons are divided into three
AB systems. As previously observed for the tris-imidazole
funnel complexes, the NMR data are consistent with the
helicoidal structure of the coordination arms around the
ZnII ion.[21a] This helicoidal chirality is transferred to the
whole calixarene structure that twists around its axis. In
CD3CN, the inversion barrier between the two helices is
Figure 6. Aminoalkynes with a C4 spacer.
Monofunctionalization with ArCH2NH2: The monoclick re-
action proceeds in high yield with ArCH2NH2 (95%). The
reaction is complete within 2 hours, like the corresponding
aliphatic 5-hexyn-1-amine. The inclusion complex was ob-
tained quantitatively by adding 1.1 equivalents of 3-ethynyl-
benzylamine to the [ZnX6N3]2+ complex. The resonances of
the guest protons are all shifted upfield, which is consistent
with the inclusion of the guest in the cavity. The strongest
shielding effects can be observed on Ha (CIS=ꢄ2.84 ppm)
and Hd (CIS=ꢄ2.50 ppm), which is stronger than with an
alkyl spacer (see Figure 7 for assignment). Protons Hb and
Hc are also shielded, but to a lesser extent, which indicates
that their positions are lower in the calixarene cavity. How-
ever, the CIS values for the calixarene protons are all very
close to the CIS values for HG4 (see the Supporting Infor-
mation, Table 4.1), which means that the rigidity of the
guest ligand does not have an impact on the conformation
of the host–guest complex.
¼
DG =(50.7ꢁ0.4) kJmolꢄ1 and was calculated for the
CH2(g) protons at their coalescence temperature (Tc =
2+
(260ꢁ2) K). The spectrum of [ZnM(C )NH
]
in CD3CN is
5
2
broad at 300 K (Figure 2F) and greatly resembles the spec-
2+
trum of [ZnM(C )NH
]
in CD3CN at 280 K. This indicates
4
2
2+
that the helicoidal inversion is slower for [ZnM(C )NH
]
5
2
than for [ZnM(C )NH ]
2+. Indeed, the inversion barrier be-
4
2
tween the two helices is DG =(54.9ꢁ0.4) kJmolꢄ1 for
¼
2+
[ZnM(C )NH
]
in CD3CN (calculated for the CH2(g) protons
5
2
at their coalescence temperature Tc =(265ꢁ2) K), which is
greater than that for [ZnM(C )NH
]
2
2+. This result is also con-
4
sistent with the competition study. In this case the spacer is
too long so the chain is packed in the cavity and impacts on
the mobility of the whole structure. Thus, the competition
experiments are plainly corroborated by the conformational
analysis of the monofunctionalized products by reasoning on
a late transition state. A C3 spacer is too short and imposes
a distortion on the calixarene cone whereas a C5 spacer is
too long and imposes constrained conformations on the
global structure. Both spacers compel energetically unfavor-
able constraints on their corresponding transition states,
which explains their high energies relative to the optimal C4
spacer.
The 1H NMR (CD3CN, 500 MHz, 300 K) spectrum of
2+
[ZnMArCH NH
]
is consistent with a Cs-symmetrical com-
2
2
plex. All the chemical shifts of the calixarene structure are
2+
similar to those of the [ZnM(C )NH
]
complex. In this case,
4
2
the signals from the ArCHax protons are very sharp dou-
blets, whereas they give broad resonances in [ZnM(C )NH
2+
] .
4
2
In conclusion to this extensive conformational analysis,
we have shown that the length of the spacer is a key ele-
This indicates that they waver in a more restricted space in
2+
[ZnMArCH NH
] . Consequently, we can conclude that
2
2
648
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 642 – 653

Guest Covalent Capture by a Host
FULL PAPER
tion of the calixarene structure.
The functionalization of the tri-
azole at the 5-position has little
impact on this structure.
Conclusions on the monoclick
reactions with amine substrates:
The monofunctionalized reac-
tion developed with amine sub-
strates is an easy, clean, and ef-
ficient method for selectively
monofunctionalizing calix[6]ar-
enes in high yields. We have es-
tablished that there is an opti-
mal length of four carbon
atoms for the spacer, which
originates from the synergy be-
tween the thermodynamically
controlled complexation step
and the kinetically controlled
intramolecular
cycloaddition
step. For the same optimal
length, different structures for
the spacer and for the alkyne
substituent can be used for the
monoclick reaction provided
Figure 7. Top: Reaction pathway for the monoclick reaction with 3-ethynylbenzylamine. Bottom: 1H NMR
(CD3CN, 500 MHz, 300 K) spectra of the inclusion complex HGArCH2NH2 and the monofunctionalized complex
2+
[ZnMArCH2NH2
]
(f: free ligand; s: residual solvent; w: water).
they are not too bulky. The
monoclick reaction works with
2+
[ZnMArCH2NH2
]
is more rigid than [ZnM(C )NH
]
2+. The ri-
different amines, despite distorting the calixarene structure.
The distortion has a low energetic cost, which indicates that
we are dealing with a very flexible cavity system. The con-
formation of the monofunctionalized calixarene is imposed
by the nature of the spacer. The spacer has a steric limit be-
cause it should fit into the cavity and not differ too much
from the optimum n=4. There is also a steric limit for the
group functionalizing the acetylene. Too much hindrance
prevents coordination of the substrates and inhibits the in-
tramolecular reaction.
4
2
gidity of the guest is then transferred to the monofunctional-
ized calixarene structure. H NMR studies indicate that all
six aromatic units are oriented very similarly to
but the calixarene structure is much more dis-
torted (see the Supporting Information for more details).
1
2+
[ZnM(C )NH
]
4
2
Monofunctionalization with substituted acetylenes: The
monoACHTUNGTRENNUNGclick reaction also proceeds in high yields with
pNO2Ar(C4)NH2 (91%) and pMeOAr(C4)NH2 (92%) sub-
stituted acetylenes but the reaction requires 5 hours to go to
completion. This indicates that the substitution of the acety-
lene has a strong effect on the kinetics of the reaction.[22]
This difference originates from a less favorable coordination
step (K4/KpNO Ar(C )NH =(1.60ꢁ0.20), see the Supporting In-
Monoclick reactions with primary alcohols: Amines proved
to be suitable for structural analysis because the intramolec-
ular coordination is fully observed both in coordinating and
noncoordinating solvents, thus avoiding a mixture of endo
and exo isomers. Primary alcohols are weaker ligands than
primary amines but their coordination to the ZnII ion can
still be observed in noncoordinating solvents.[14a] Starting
with the conditions found for the amine substrates, the syn-
thesis was optimized for the alcohols. This optimization is
discussed below. The synthesis of a calixarene complex
2
4
2
formation for competition experiments) because of the
bulky phenyl group in this position, as well as a less favora-
ble intramolecular reaction (k4/kpNO Ar(C )NH =(2.60ꢁ0.20)).
2
4
2
The discrimination based on host–guest adduct formation
can be even more drastic: A bulkier group (anthracene) pre-
vents the coordination of the amino guest and no conversion
for the monoclick reaction is observed. In both cases the
spacer is the same (n=4) but the acetylene is differently
functionalized. The chemical shifts of the monofunctional-
ized complexes are all very similar to the chemical shifts of
monoACTHUNRTGNEUNGfunctionalized by a hydroxy group enabled us to study
the “endo/exo” dynamic equilibrium that operates under
certain conditions.
[ZnM(C )NH
]
2+. This indicates that the three monofunctional-
ized complexes [ZnMpOCH (C )NH
2+, [ZnMpNO Ar(C )NH 2+, and
have very similar conformations. Therefore
the nature of the spacer is what determines the conforma-
Synthesis and characterization of the monoclick product
4
2
]
]
[ZnM(C )OH]
2+: 5-Hexyn-1-ol is the strict equivalent to 5-
3
4
2
2
4
2
4
2+
[ZnM(C )NH
]
hexyn-1-amine as the only modified parameter is the coordi-
nating group. Because water competes with the alcohol for
4
2
Chem. Eur. J. 2013, 19, 642 – 653
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
649

B. Colasson et al.
coordination at the metal center, the reaction was conducted
in dry toluene at reflux. The optimized conditions required
the use of ten equivalents of 5-hexyn-1-ol (vs. 2 equiv for 5-
hexyn-1-amine) and 4 h of reflux (vs. 2 h for 5-hexyn-1-
placed by one molecule of CD3CN (exo situation for the
chain). A second minor species can also be detected.
Indeed, shielded resonances were also observed at ꢄ0.25
and ꢄ1.03 ppm, attesting to the existence of a self-coordi-
nated complex (endo situation for the chain). At 300 K, the
exo to endo ratio, evaluated by integration of the signals of
the alkyl chain, is 93:7. This ratio was evaluated at different
temperatures between 270 and 340 K (see the Supporting
Information). Self-coordination of the hydroxy chain is fa-
vored at lower temperatures. The vanꢃt Hoff plot for the
equilibrium of the displacement of the coordinated hydroxy
chain by an acetonitrile molecule (endo/exo equilibrium)
gave the following thermodynamic parameters: DrH8=
11 kJmolꢄ1 and DrS8=59 JKꢄ1 molꢄ1. The positive enthalpic
change reflects the better affinity of the zinc cation for an
alcohol compared with a nitrile.[13a] The favorable change in
entropy can be rationalized by the expulsion of the alkyl
chain outside of the cavity due to the decoordination and
gain of conformational flexibility of both the chain and the
calixarene core. In CD3CN, the addition of one equivalent
of a primary amine like propylamine yields quantitatively
amine). Unlike with primary amines, the resulting mono-
2+
functionalized complex [ZnM(C )OH
]
was generally isolated
4
with some starting material [ZnX6N3]2+ and a difunctional-
ized complex, which shows that the monoclick reaction with
5-hexyn-1-ol is less selective than with 5-hexyn-1-amine.
After purification by column chromatography (during which
decoordination of the metal cation occurred), pure mono-
functionalized ligand M(C )OH was obtained. Upon recom-
4
plexation with Zn
monofunctionalized calixarene [ZnM(C )OH
ACHTUNGTRENNUNG(ClO4)2·ACHUTNGTREN(NNGU H2O)6 in CH2Cl2/CH3CN, pure
2+
]
was isolated in
4
a yield of 70–95% (over two steps) depending on the run.
The monofunctionalized ligand M(C )OH was characterized
4
by EIMS spectrometry (a single peak was observed at m/z=
1350.7 for [MC4OH +H]+). The 1H NMR spectrum of the
ligand is broad in the usual solvents (CDCl3, CD3CN,
[D6]DMSO) at 300 K but the resonances are sharper at
360 K in [D6]DMSO (Figure 8). The appearance of a singlet
the
[ZnM
inclusion complex
(C )OHA , which
(PrNH2)]2+
4
means that both the self-coordi-
nated hydroxy chain and the
acetonitrile molecule are dis-
placed. This displacement is
enthalACTHNUGRTENUNGpically driven by the
ꢄ
strength of the Zn N bond.
The behavior of complex
2+
[ZnM(C )OH
]
is very similar in
4
CD3OD because the two com-
plexes (endo and exo) coexist at
300 K, the exo complex being
largely favored (see the Sup-
porting Information). In the
noncoordinating solvent CDCl3
at 325 K, the upfield resonances
of the b- and g-methylenes at
ꢄ1.08 and ꢄ0.21 ppm, respec-
tively, as well as the absence of
Figure 8. Bottom: 1H NMR (500 MHz) spectrum of ligand MC4OH ([D6]DMSO, 360 K). Top: 1H NMR
2+
(500 MHz) spectrum of complex [ZnMC4OH
]
with zoom of the signals of the encapsulated b- and g-methylene
groups in the upfield region (CD3CN, 300 K) (s: residual solvent; w: water).
resonances
at
3.40
and
2.70 ppm for the a- and d-meth-
ylenes, suggest the presence of only the self-coordinated
complex. The only competing ligand in the media is residual
water, which is not strong enough to displace the intramo-
at 7.64 ppm (integrating for 1H) for Htria on the newly
formed triazole group, as well as the resonances for the one
clicked alkyl chain attest to the monofunctionalization of
the calixarene. In addition, the splitting of the tBu, OCH3,
and ImCH2 groups is consistent with a Cs-symmetrical spe-
lecularly coordinated hydroxy chain.
Thus, the host–guest properties of the [ZnM(C )OH
2+
]
com-
4
cies.
plex are very different to those of its amino analogue. The
hydroxy chain can be chased out of the cavity by an exoge-
nous alcohol, nitrile, or primary amine (Scheme 3). Never-
2+
The monofunctionalized complex [ZnM(C )OH
]
was char-
4
acterized by EIMS (m/z=1512.6 for [ZnM(C )OH +ClO4]+).
4
1
The H NMR spectrum in CD3CN at 300 K displays sharp
theless, compared with classic funnel complexes like
2+
resonances and evidences one major product, consistent
with a monofunctionalized calixarene (Figure 8). In contrast
[ZnX6tBu6]2+, the affinity of [ZnM(C )OH
]
for alcohols and
4
nitriles is greatly reduced due to the control of the accessi-
bility of the metal center by the self-coordination of the in-
ternal arm. In contrast, the amino-functionalized calixarene
remains self-coordinated in all tested conditions.
to [ZnM(C )NH ]
2+, the resonances of the alkyl chain in
2
42+
[ZnM(C )OH
]
are not shielded. This indicates that the hy-
4
droxy group is not coordinated to the zinc cation and is dis-
650
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 642 – 653

Guest Covalent Capture by a Host
FULL PAPER
plexation with Zn
functionalized complexes [ZnM(C )OH
ACHTUGTNRENNU(G ClO4)2·ACHTUNTGERN(NUGN H2O)6, the corresponding mono-
2+
2+
]
and [ZnMArCH OH]
3
2
were obtained in good yields of 75–95%. Both complexes
were characterized by EIMS (m/z=1498.6 for [ZnM(C )OH
+
3
ClO4]+ and m/z=1546.6 for [ZnMArCH OH +ClO4]+). In con-
2
1
2+
trast to [ZnM(C )OH
]
2+, the H NMR spectra of [ZnM(C )OH
]
4
3
2+
and [ZnMArCH OH
]
in CD3CN evidence only one Cs-sym-
2
metrical species with the pendant hydroxy chain out of the
cavity and an acetonitrile molecule coordinated to the ZnII
ion. No self-coordinated complexes were observed.
Scheme 3. Favorable displacement of the self-coordinated hydroxy chain
of complex [ZnMC4OH]ACHTUNGTRENNUNG(ClO4)2 by coordinating solvents MeCN, MeOH,
or PrNH2 in stoichiometric proportions in any solvent.
The reaction conducted with pNO2ArC4OH was much
cleaner and much more selective. Indeed, the pure mono-
2+
functionalized complex [ZnMpNO Ar(C) OH
]
was easily isolat-
2
4
ed after precipitation. It was characterized by EIMS (m/z=
1
Variable-temperature (VT) NMR experiments were con-
1633.6 for [ZnMpNO ArC OH +ClO4]+). Its H NMR spectrum
2
4
2+
ducted on [ZnM(C )OH
]
in CDCl3 (see the Supporting Infor-
in CD3CN at 300 K evidences a 75:25 mixture of the exo
and endo complexes. VT NMR experiments in CD3CN in
the range 270–325 K gave the following thermodynamic pa-
rameters for the endo/exo equilibrium: DrH8=3.8 kJmolꢄ1
4
mation). The results are very similar to those obtained with
2+
2 2+
[ZnM(C )NH
]
]
.
Indeed, at 325 K, the spectrum of
4
[ZnM(C )OH
is characteristic of a Cs-symmetrical species in
4
which the helicoidal inversion of the imidazole arms is fast
on the NMR timescale. At 300 K, the signals broaden and at
266 K the spectrum is characteristic of a C1-symmetrical spe-
cies in which the helicoidal inversion has become very slow.
and DrS8=22 JKꢄ1 molꢄ1. Compared with [ZnM(C )OH
]
2+, the
4
self-coordination is enthalpically less favored, likely caused
by the steric hindrance exerted by the p-nitrophenyl group
at the large rim. The change in entropy is also much less
positive, which can hypothetically be explained by the re-
duced mobility of the alkyl chain in the vicinity of the p-ni-
trophenyl group. These results evidence a stronger self-coor-
dination effect with the p-nitrophenyl group compared with
other alcohols, which diminishes the possibility of the dis-
The inversion barrier was calculated for protons Hg at their
¼
coalescence
temperature
(302 K):
DG =(59.7ꢁ
0.5) kJmolꢄ1. The inversion barrier is the same as for
2+
¼
[ZnM(C )NH
]
in the same solvent (DG =(59.3ꢁ
4
2
0.5) kJmolꢄ1). Although we have observed that the inver-
sion barrier of the helicoidal equilibrium is sensitive to the
length of the spacer, there is little impact of the nature of
the coordinating group.
placement of the self-coordinated hydroxy chain in
2+
[ZnMpNO Ar(C) OH
]
by a second pNO2ArC4OH molecule
2
4
and thus rationalizes the better selectivity of the monoclick
reaction. In CD3CN, both the coordinated acetonitrile mole-
cule and the self-coordinated hydroxy chain were classically
displaced by the addition of one equivalent of propylamine
Monoclick reactions with other alcohol substrates: As in the
case of the amines, the spacer can be slightly modified.
Other alcohols such as 3-ethynylbenzyl alcohol (ArCH2OH)
and 4-pentyn-1-ol ((C3)OH) allow a change in the spacer
(Figure 9). Acetylene can also be substituted by a 4-nitro-
phenyl group, for example, pNO2ArC4OH. The monoclick
reaction was conducted under the same conditions (10 equiv
of 4-pentyn-1-ol or 3-ethynylbenzyl alcohol at reflux in dry
toluene for 4 h). In both cases, the major product was the
monofunctionalized product, but the precursor complex and
the difunctionalized product were also usually observed.
After purification by column chromatography and recom-
to
[ZnM
yield
pNO ArC OHA
quantitatively
the
inclusion
complex
G
.
2
4
Conclusion
Selective functionalization of a macrocyclic structure pre-
senting multiple and equally reactive sites is a real synthetic
challenge. It generally represents a limiting step for append-
ing valuable and spatially well-defined functions. In this
paper we have presented a strategy based on host–guest
chemistry that allows for the selective reaction of one out of
three equivalent units of a calix[6]arene core. The strategy
was inspired by natural systems, that is, enzymes that are
well known to perform highly selective reactions under very
mild conditions. The key points for enzymatic transforma-
tions are the formation of an enzyme–substrate complex
and preorganization of the reactive partners (reactant/sub-
strate) in the active-site pocket, which lead to low activation
energy and high selectivity. The reaction implemented here
is the thermal Huisgen cycloaddition, which, when noncata-
lyzed, requires high temperatures and leads to low regiose-
lectivities. To take advantage of the host properties of the
Figure 9. Hydroxy-acetylenes bearing different spacers and functionali-
ties.
Chem. Eur. J. 2013, 19, 642 – 653
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
651

B. Colasson et al.
calix[6]arene-based ZnII funnel complexes, three azido
groups were introduced at the large rim of the cavity (the
reactant) and a coordinating moiety was appended to the
alkyne substrates through a spacer (substrate). To explore
the potentialities of this supramolecular strategy, four pa-
rameters were varied: 1) The nature of the coordinating
moiety (primary amine vs. alcohol), 2) the length of the
spacer from (CH2)3 to (CH2)5, 3) the shape of the spacer
(linear alkane vs. phenyl nucleus), and 4) the substituent at
the terminal site of the alkyne substrate.
Detailed structural and kinetic studies have evidenced in-
teresting biomimetic features. The better the guest, the
higher the reaction rate, which is related to thermodynamics
through the host–guest binding constant K. The less sterical-
ly constrained the product, the higher the rate constant,
which proves that embedment also has a significant impact
on the transition state.
The high sensitivity of the system is nicely illustrated by
the fact that a change of only one methylene group in the
spacer length has a significant impact. The optimal C4 length
allows both a better preorganization and a lower-energy
transition state in comparison with C3 and C5. Nevertheless,
the flexibility of the host makes possible the efficient em-
bedment of either a linear alkyl chain or a cyclic moiety.
The efficiency of the confined reaction also permits alkynes
bearing a sterically encumbered aromatic substituent at
their terminal position to be grafted. All these results show
that such a pre-association mimicking the Michaelis–Menten
enzyme–substrate complex allows for the selective orienta-
tion of reactant versus substrate and a good pre-organiza-
tion leading to lower activation energies. This is illustrated
by the remarkably mild conditions under which this thermal
Huisgen reaction proceeds compared with classical bimolec-
ular reactions and also by the remarkable regioselectivity of
the cycloaddition.
Finally, it is important to note that another key point of
this synthetic strategy is the self-inhibition of the process.
After covalent capture of a single guest, the intramolecular
coordination link lowers the affinity of the metal center for
exogenous ligands. With very good donors such as amines,
the endo coordination is very strong and quasi-quantitative
yields of monofunctionalized product are obtained. With the
weaker donor alcohols, the transformation is less selective
because the formation of around 5–10% of dialkylated
products was observed. This is attributable to the more
labile alcohol–ZnII link that leads to more facile competition
for the endo binding of a second equivalent of guest alcohol.
Interestingly, the study related to such endo–exo competitive
binding revealed that the self-coordination process of the
appended alcohol function can be switched on and off by
the presence of a more or less competing ligand. The equi-
the redox state of a copper ion is under investigation in our
laboratory.
In conclusion, an in-depth analysis of the main factors
governing the selectivity and efficiency of the supramolecu-
lar covalent capture process has shown that such a strategy
is quite versatile and should be applicable to other cavitands
associated with other recognition patterns. Also, this strat-
egy could possibly be extended to other covalent capture re-
actions. Interestingly, such a covalent capture strategy has
been recently applied to the synthesis of a carbonic anhy-
drase II mutant.[23]
Experimental Section
General procedure for the monoclick reaction with amines: Precursor
complex [ZnX6N3]ACTHNUTRGEN(UNG ClO4)2 (200 mg, 0.13 mmol, 1.0 equiv) was added to a
solution of the appropriate amine (2 to 3 equiv) in dry toluene (10 mL).
The resulting mixture was heated at reflux under argon for 2 h (for
(C3)NH2 and (C4)NH2) and 5 h (for (C5)NH2). Toluene was evaporated
under reduced pressure and the residue was redissolved in a minimum
volume of acetonitrile (ca. 3 mL). The calixarene was then precipitated
with diethyl ether (30 mL). The solid was filtered off, washed with diethyl
ether (3ꢄ10 mL), and dried under vacuum.
General procedure for the monoclick reaction with alcohols: Precursor
complex [ZnX6N3]ACTHNUTRGEN(UNG ClO4)2 (200 mg, 0.13 mmol, 1.0 equiv) was added to a
solution of the appropriate alcohol (10 equiv) in dry toluene (10 mL).
The resulting mixture was heated at reflux under argon for 4 h. Toluene
was evaporated under reduced pressure and the residue was redissolved
in a minimum volume of acetonitrile (ca. 3 mL). The calixarene was then
precipitated with diethyl ether (30 mL). The solid was filtered off,
washed with diethyl ether (3ꢄ10 mL), and dried under vacuum. In the
case of 6-(4-nitrophenyl)-5-hexyn-1-ol, the pure monofunctionalized com-
plex was obtained without further purification. In the case of 5-hexyn-1-
ol, 4-pentyn-1-ol, and 3-ethynylbenzyl alcohol, the crude product was pu-
rified by column chromatography on basic alumina (CH2Cl2/CH3OH/NH3
(27%), 99:1:0.01) to obtain the pure monofunctionalized calixarene
ligand. The monofunctionalized ligand (20 mg, 1.0 equiv) was dissolved
in a 1:1 mixture of CH2Cl2 and CH3CN (1 mL). A solution of zinc per-
chlorate hexahydrate (1.0 equiv) predissolved in a minimum amount of
THF was added to a solution of the ligand. After stirring for 15 min, di-
ethyl ether was added (10 mL). The resulting precipitate was filtered off
and dried under vacuum to provide the corresponding ZnII complex as a
white solid in 94–98% yield.
Full characterization data for all the monoclick products are presented in
the Supporting Information.
Acknowledgements
This project was supported by the CNRS (Institut de Chimie), the Minis-
tꢁre de lꢃEnseignement Supꢀrieur et de la Recherche, and the Agence
Nationale pour la Recherche (Cavity-zyme(Cu) Project ANR-2010-
BLAN-7141).
[1] a) M. B. Baker, I. Ghiviriga, R. K. Castellano, Chem. Sci. 2012,
DOI: 10.1039/c2sc00943a; b) M. Lehmann, Chem. Eur. J. 2009, 15,
3638–3651; c) M. Lehmann, M. Jahr, Chem. Mater. 2008, 20, 5453–
5456; d) D. G. Rivera, L. A. Wessjohann, J. Am. Chem. Soc. 2009,
131, 3721–3732; e) P. Gamez, J. Reedijk, Eur. J. Inorg. Chem. 2006,
29–42.
ACHTUNGTRENNUNGlibACHTUNGTRENNUNGrium is shifted towards the endo adduct in pure CDCl3
whereas the exo situation is obtained in the presence of one
equivalent of a primary amine. The “ouroboros-like” behav-
ior of this system actually offers the possibility of controlling
the access to the cavity by using different kinds of stimuli.
In this context, the use of an electrochemical input tuning
[2] a) C. Ornelas, R. Pennell, L. F. Liebes, M. Weck, Org. Lett. 2011, 13,
976–979; b) L. Rçglin, E. H. M. Lempens, E. W. Meijer, Angew.
652
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 642 – 653

Guest Covalent Capture by a Host
FULL PAPER
Chem. 2011, 123, 106–117; Angew. Chem. Int. Ed. 2011, 50, 102–
112; c) K. Sivanandan, B. S. Sandanaraj, S. Thayumanavan, J. Org.
Chem. 2004, 69, 2937–2944; d) L. J. Prins, F. Mancin, P. Scrimin,
Curr. Org. Chem. 2009, 13, 1050–1064; e) H. B. Albada, R. M. J. Lis-
kamp, J. Comb. Chem. 2008, 10, 814–824; f) A. P. Davis, Molecules
2007, 12, 2106–2122; g) A. Gea, N. Farcy, N. Roquꢀ i Rossell, J. C.
Martins, P. J. De Clercq, A. Madder, Eur. J. Org. Chem. 2006, 4135–
4146; h) G. J. Rowlands, Tetrahedron 2001, 57, 1865–1882;
i) M. J. W. Ludden, D. N. Reinhoudt, J. Huskens, Chem. Soc. Rev.
2006, 35, 1122–1134; j) L. Baldini, A. Casnati, F. Sansone, R.
J. Rebek Jr., J. Am. Chem. Soc. 2004, 126, 16280–16281; h) A.
Gissot, J. Rebek Jr., J. Am. Chem. Soc. 2004, 126, 7424–7425.
[11] B. Colasson, O. Reinaud, J. Am. Chem. Soc. 2008, 130, 15226–
15227.
[12] R. Huisgen, Angew. Chem. 1963, 75, 604–637; Angew. Chem. Int.
Ed. Engl. 1963, 2, 565–598.
[13] a) O. Sꢀnꢁque, M.-N. Rager, M. Giorgi, O. Reinaud, J. Am. Chem.
Soc. 2000, 122, 6183–6189; b) D. Coquiꢁre, S. Le Gac, U. Darbost,
O. Sꢀnꢁque, I. Jabin, O. Reinaud, Org. Biomol. Chem. 2009, 7,
2485–2500.
´
Ungaro, Chem. Soc. Rev. 2007, 36, 254–266; k) J. D. Badjic, A.
[14] M represents the monofunctionalized calix[6]arene, C3 indicates that
a three methylene chain connects the amino and acetylene groups,
and NH2 indicates that the coordinating unit is an amino group. The
descriptor MCnX is used throughout the manuscript.
Nelson, S. J. Cantrill, W. B. Turnbull, J. F. Stoddart, Acc. Chem. Res.
2005, 38, 723–732.
[3] P. Timmerman, W. Verboom, D. N. Reinhoudt, Tetrahedron 1996, 52,
2663–2704.
[4] J. Lagona, P. Mukhopadhyay, S. Chakrabarti, L. Isaacs, Angew.
Chem. 2005, 117, 4922–4949; Angew. Chem. Int. Ed. 2005, 44, 4844–
4870.
[15] In the absence of ZnII, the reaction does not show any conversion.
The addition of a CuI catalyst leads to a mixture of triazole prod-
ucts.
[16] Ouroboros (“tail-eater” in Greek) is an ancient symbol representing
a serpent swallowing its own tail, see: a) F. Durola, J. Rebek Jr.,
Angew. Chem. 2010, 122, 3257–3259; Angew. Chem. Int. Ed. 2010,
49, 3189–3191; b) S. Nygaard, Y. Liu, P. C. Stein, A. H. Flood, J. O.
Jeppesen, Adv. Funct. Mater. 2007, 17, 751–762; c) Y. Enmi, K. Ko-
bayashi, H. Konishi, M. Morimoto, Tetrahedron Lett. 2011, 52,
3367–3370.
[17] We assume that the calculated ratio of equilibrium constants re-
mains valid under the reaction conditions (toluene, 1108C) because
the structures of the three amines are very similar and no solvent ef-
fects are expected to favor the complexation of one particular guest.
[18] K. Kano, M. Kondo, H. Inoue, H. Kitagishi, B. Colasson, O. Rein-
aud, Inorg. Chem. 2011, 50, 6353–6360.
[5] J. Szejtli, Chem. Rev. 1998, 98, 1743–1753.
[6] a) C. D. Gutsche, in Calixarenes An introduction, Monographs in:
Supramolecular Chemistry, (Ed: J. F. Stoddart, The Royal Society of
Chemistry, Cambridge), 2nd ed., 2008; b) C. D. Gutsche, in Calixar-
enes Revisited, Monographs in: Supramolecular Chemistry, (Ed: J. F.
Stoddart, The Royal Society of Chemistry, Cambridge), 1998.
[7] a) K. Teranishi, Chem. Commun. 2000, 1255–1256; b) K. Teranishi,
Tetrahedron 2003, 59, 2519–2538; c) K. Teranishi, J. Inclusion
Phenom. Macrocyclic Chem. 2002, 44, 313–316; d) D. Rong, T.
DꢃSouza, Tetrahedron Lett. 1990, 31, 4275–4278; e) H. Cousin, P.
Cardinael, H. Oulyadi, X. Pannecoucke, J. C. Combret, Tetrahedron:
Asymmetry 2001, 12, 81–88; f) P. Fꢅgedi, P. Nꢆnꢆsi, Carbohydr. Res.
1988, 175, 173–118; g) Z. Chen, J. S. Bradshaw, M. L. Lee, Tetrahe-
dron Lett. 1996, 37, 6831–6834; h) A. R. Khan, P. Forgo, K. J. Stein,
V. T. DꢃSouza, Chem. Rev. 1998, 98, 1977–1996.
[19] At room temperature, the exchange of two amine guests in the re-
ceptor takes place within less than a minute. At 1108C, the time-
scale for the cycloaddition reaction is around 2 h.
[8] R. F. Gꢇmez-Biagi, M. Nitz, Chem. Commun. 2011, 47, 8614–8616.
[9] a) M. Ahmed, K. Koga, M. Fukudome, H. Sasaki, D. Yuan, Tetrahe-
dron Lett. 2011, 52, 4646–4649; b) D. Lucas, T. Minami, G. Iannuzzi,
L. Cao, J. Wittenberg, P. Anzenbacher, L. Isaacs, J. Am. Chem. Soc.
2011, 133, 17966–17976; c) N. Zhao, G. O. Lloyd, O. A. Scherman,
Chem. Commun. 2012, 48, 3070–3072.
[20] B. Colasson, M. Save, P. Milko, J. Roithovꢆ, D. Schrçder, O. Rein-
aud, Org. Lett. 2007, 9, 4987–4990.
[21] a) O. Sꢀnꢁque, Y. Rondelez, L. Le Clainche, C. Inisan, M.-N. Rager,
M. Giorgi, O. Reinaud, Eur. J. Inorg. Chem. 2001, 2597–2604; b) O.
Sꢀnꢁque, M. Giorgi, O. Reinaud, Supramol. Chem. 2003, 15, 573–
580.
[10] a) A. R. Renslo, J. Rebek Jr., Angew. Chem. 2000, 112, 3419–3421;
Angew. Chem. Int. Ed. 2000, 39, 3281–3283; b) U. Lꢅcking, F. C.
Tucci, D. M. Rudkevich, J. Rebek Jr., J. Am. Chem. Soc. 2000, 122,
8880–8889; A. R. Renslo, D. M. Rudkevich, J. Rebek Jr., J. Am.
Chem. Soc. 1999, 121, 7459–7460; c) A. R. Renslo, F. C. Tucci, D. M.
Rudkevich, J. Rebek Jr., J. Am. Chem. Soc. 2000, 122, 4573–4582;
d) U. Lꢅcking, D. M. Rudkevich, J. Rebek Jr., Tetrahedron Lett.
2000, 41, 9547–9551; e) S. D. Starnes, D. M. Rudkevich, J. Rebek Jr.,
Org. Lett. 2000, 2, 1995–1998; f) S. D. Starnes, D. M. Rudkevich, J.
Rebek Jr., J. Am. Chem. Soc. 2001, 123, 4659–4669; g) S. Richeter,
[22] Because the reactions with the two substrates pNO2Ar(C4)NH2 and
pOMeAr(C4)NH2 need 5 h to go to completion, we privilege steric
rather than electronic factors for the discussion about the reaction
rates.
[23] J. S. Wischeler, D. Sun, N. U. Sandner, U. Linne, A. Heine, U. Koert,
G. Klebe, Chem. Eur. J. 2011, 17, 5842–5851.
Received: July 5, 2012
Revised: September 24, 2012
Published online: November 20, 2012
Chem. Eur. J. 2013, 19, 642 – 653
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
653