A molecular oyster: A neutral anion receptor containing two cyclopeptide subunits with a remarkable sulfate affinity in aqueous solution

Article abstract of DOI:10.1021/ja026996q

An artificial anion receptor is presented, in which two cyclohexapeptide subunits containing L-proline and 6-aminopicolinic acid subunits in an alternating sequence are connected via an adipinic acid spacer. This compound was devised to stabilize the 2:1

Full text of DOI:10.1021/ja026996q

Published on Web 10/08/2002
A Molecular Oyster: A Neutral Anion Receptor Containing
Two Cyclopeptide Subunits with a Remarkable Sulfate Affinity
in Aqueous Solution
Stefan Kubik,*,† R. Kirchner,‡ D. Nolting,§ and J. Seidel‡
Contribution from the Institut fu¨r Organische Chemie und Makromolekulare Chemie,
Heinrich-Heine-UniVersita¨t, UniVersita¨tsstrasse 1, D-40225 Du¨sseldorf, Germany, Institut fu¨r
Physikalische Chemie, Technische UniVersita¨t Bergakademie Freiberg, Leipziger Strasse 29,
D-09596 Freiberg, Germany, and Institut fu¨r Physikalische Chemie und Elektrochemie,
Heinrich-Heine-UniVersita¨t, UniVersita¨tsstrasse 1, D-40225 Du¨sseldorf, Germany
Received May 22, 2002
Abstract: An artificial anion receptor is presented, in which two cyclohexapeptide subunits containing
L-proline and 6-aminopicolinic acid subunits in an alternating sequence are connected via an adipinic acid
spacer. This compound was devised to stabilize the 2:1 sandwich-type anion complexes that are observed
when the two cyclopeptide moieties are not covalently connected and to obtain a 1:1 stoichiometry for
these aggregates. Electrospray ionization mass spectrometry and NMR spectroscopic investigations showed
that the bridged bis(cyclopeptide) does indeed form defined 1:1 complexes with halides, sulfate, and nitrate.
ROESY NMR spectroscopy and molecular modeling allowed a structural assignment of the sulfate complex
in solution. The stabilities of various anion complexes were determined by means of NMR titrations and
isothermal titration microcalorimetry in 50% water/methanol. Both methods gave essentially the same
quantitative results, namely stability constants that varied in the range 105-102 M-1 and decreased in the
order SO4 > I- > Br- > Cl- > NO3-. This order was rationalized in terms of the size of the anions with
2-
the larger anions forming the more stable complexes because they better fit into the cavity of the host. The
ability of sulfate to form stronger hydrogen bonds to the NH groups of the receptor, in addition to its slightly
larger ionic radius with respect to iodide, causes the higher stability of the sulfate complex. No significant
effect of the countercation on complex stability was observed. Furthermore, complex stability is enthalpically
as well as entropically favored. A comparison of the iodide and sulfate complex stabilities of the ditopic
receptor with those of a cyclopeptide that forms 1:1 anion complexes in solution showed that the presence
of a second binding site increases complex stability by a factor of 100-350.
Introduction
array of hydrogen bonds between the anion and NH groups of
the protein backbone, serine OH, or tryptophane NH groups.3
Many vitally important biochemical processes rest upon the
specific interaction between proteins and anionic substrates such
as carbonate, sulfate, or phosphate.1 To achieve the high
substrate affinity and selectivity necessary in these interactions,
Nature has devised a number of efficient binding motifs. In the
phosphate binding protein (PBP), for example, a protonated
guanidinium group suitably placed inside the active center is
an important element of substrate recognition.2 Other natural
systems, such as the sulfate binding protein (SBP), do not require
strong electrostatic interactions for substrate binding. In these
systems, affinity and selectivity is mainly controlled by a defined
These basic principles of anion complexation have been used
extensively in the context of supramolecular chemistry for the
design of artificial anion receptors.4 Host molecules have been
described, for example, that bind anions efficiently in polar,
even aqueous, solution by means of strong electrostatic (or
coordinative) interactions. Other receptors only use hydrogen
bonds for substrate recognition.5 Although these neutral systems
often form stable anion complexes in competitive solvents such
as DMSO or acetonitrile, complex formation in water generally
proved to be weak if present at all, one reason for this being
the high hydration energies of many anions. An exception in
this respect is the cyclic hexapeptide with alternating L-proline
* To whom correspondence should be addressed. E-mail: kubik@uni-
duesseldorf.de. Fax: +49-211-81-14788.
† Institut fu¨r Organische Chemie und Makromolekulare Chemie, Hein-
rich-Heine-Universita¨t.
(3) Pflugrath, J. W.; Quiocho, F. A. Nature 1985, 314, 257-260. Quiocho, F.
A.; Sack, J. S.; Vyas, N. K. Nature 1987, 329, 561-564. Pflugrath, J. W.;
Quiocho, F. A. J. Mol. Biol. 1988, 200, 163-180. He, J. J.; Quiocho, F.
A. Science 1991, 251, 1479-1481.
(4) For recent reviews, see: Bianchi, A., Bowman-James, K., Garc´ıa-Espan˜a,
E., Eds. Supramolecular Chemistry of Anions; Wiley-VCH: New York,
1997. Schmidtchen, F. P.; Berger, M. Chem. ReV. 1997, 97, 1609-1646.
Antonisse, M. M. G.; Reinhoudt, D. N. Chem. Commun. 1998, 443-448.
Beer, P. D.; Gale, P. A. Angew. Chem. 2001, 113, 502-532; Angew. Chem.,
Int. Ed. 2001, 40, 417-516.
‡ Institut fu¨r Physikalische Chemie, Technische Universita¨t Bergakademie
Freiberg.
§ Institut fu¨r Physikalische Chemie und Elektrochemie, Heinrich-Heine-
Universita¨t.
(1) Mangani, S.; Ferraroni, M. In Supramolecular Chemistry of Anions; Bianchi,
A., Bowman-James, K., Garc´ıa-Espan˜a, E., Eds.; Wiley-VCH: New York,
1997; pp 63-78.
(2) Luecke, H.; Quiocho, F. A. Nature 1990, 347, 402-406.
9
12752
J. AM. CHEM. SOC. 2002, 124, 12752-12760
10.1021/ja026996q CCC: $22.00 © 2002 American Chemical Society

A Molecular Oyster
A R T I C L E S
and 6-aminopicolinic acid subunits (1) that has recently been
described by us.6 This compound interacts with anions such as
halides or sulfate by hydrogen bond formation even in water/
methanol mixtures and thus represents a promising candidate
for the development of new biomimetic anion hosts. On the
basis of 1, we have now developed a receptor of the next
generation, whose synthesis and properties we describe here.
Our new host 2 binds sulfate with a high affinity and selectivity
in 50% water/methanol and can thus be regarded as a model
for the sulfate binding protein.
of the 1:1 complex, is in every case significantly smaller than
K2, the stability constant that describes the equilibrium between
the 1:1 and the 1:2 complex. Although higher aggregates are
obviously preferred over 1:1 complexes, the composition of a
2:1 anion complex of 1 still depends markedly on the ratio and
the concentration of receptor and substrate in solution. In
general, a 2:1 complex predominates when an excess of receptor
is present, whereas, in an excess of guest, a 1:1 complex is
preferentially formed. Since the high anion affinity of 1 is mainly
caused by the ability of the peptide to form sandwich-type
complexes, we looked for ways to stabilize these aggregates.
A possible approach is based upon a strategy that has been
successfully applied in supramolecular chemistry to a number
of artificial receptors, more recently, e.g., to cyclodextrins,10
calixarenes,11 or, in the context of anion receptors, Schmidt-
chen’s macrotricyclic host12 or sapphyrines,13 namely the
covalent linkage of two receptor subunits. Increasing the number
of binding sites usually causes an improved affinity or selectivity
of a host toward a substrate of interest, but in the case of 1, we
expected that connecting two cyclopeptide rings also provides
a means to convert the observed 2:1 anion complexes into 1:1
complexes, whose binding equilibria are easier to describe.
Success of this approach requires that the two cyclopeptide
moieties are connected via a spacer that does not prevent their
cooperative action in anion binding. Chosen appropriately, the
spacer should allow a certain control over the binding properties
of the corresponding host, however.
The proline rings in 1 are the most suitable positions for the
introduction of a linker unit since they come in close contact in
the anion complexes, and by replacement of one proline subunit
of the peptide with a hydroxyproline residue, a functionalization
in the form of a hydroxyl group can easily be introduced into
the ring. Hydroxyl groups have the disadvantage, however, that
they usually possess a low reactivity in hydroxyproline residues.
Furthermore, we have shown that 4R-configured hydroxyl
groups in the proline residues of 1 prevent the formation of 2:1
complexes possibly because of steric effects.9 We therefore
decided to use a cyclopeptide as a building block for a ditopic
receptor, in which one proline subunit of 1 carries an S-
configured amino group in the 4 position (3). This compound
was prepared from a derivative of 1 containing one tosylated
4R-configured hydroxyproline by nucleophilic substitution of
the tosyl group with azide followed by hydrogenation of the
resulting product (Scheme 1).
Results and Discussion
Design and Synthesis. The unusual receptor properties of 1
are mainly a consequence of the special geometry of the anion
complexes formed. Our structural assignment has shown that 1
preferentially forms sandwich-type 2:1 complexes with anions,7
in which the guests reside in a cavity formed by the aggregation
of two almost perfectly shape-complementary molecules of 1.6
In this cavity, the anions can interact with all six NH groups of
the peptide moieties simultaneously and are effectively shielded
from the surrounding solvent, a combination of effects that could
provide a possible explanation for the stability of such com-
plexes in aqueous solution.8
Recently, we succeeded in a quantitative determination of
the stability of some anion complexes of 1.9 We found high
overall stability constants for halide and sulfate complexes
ranging from 104 to 105 M-2 in 80% D2O/CD3OD. Moreover,
complex formation is cooperative; i.e., K1, the stability constant
We then used molecular modeling to identify a dicarboxylic
acid able to connect two molecules of 3 while still enabling
them to bind anions in a cooperative manner.14 These calcula-
(8) The 2:1 sandwich complexes between 1 and anions are structurally
somewhat related to the molecular capsules described by, e.g., Rebek. For
an excellent overview on structural aspects of such capsules and the unusual
effects that are caused by an inclusion of guests into their cavities, see:
Hof, F.; Craig, S. L.; Nuckolls, C.; Rebek, J., Jr. Angew. Chem. 2002, 114,
1556-1578; Angew. Chem., Int. Ed. 2002, 41, 1488-1508.
(9) Kubik, S.; Goddard, R. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5127-
5132.
(10) Breslow, R. Pure Appl. Chem. 1994, 66, 1573-1582. Breslow, R. Acc.
Chem. Res. 1995, 28, 146-153. Venema, F.; Rowan, A. E.; Nolte, R. J.
M. J. Am. Chem. Soc. 1996, 118, 257-258. Venema, F.; Nelissen, H. F.
M.; Berthault, P.; Birlirakis, N.; Rowan, A. E.; Feiters, M. C.; Nolte, R. J.
M. Chem.sEur. J. 1998, 4, 2237-2250.
(11) Arduini, A.; Pochini, A.; Secchi, A. Eur. J. Org. Chem. 2000, 2325-2334.
(12) Schmidtchen, F. P. J. Am. Chem. Soc. 1986, 108, 8249-8255. Schmidtchen,
F. P. Tetrahedron Lett. 1986, 27, 1987-1990.
(13) Kra´l, V.; Andrievsky, A.; Sessler, J. L. J. Am. Chem. Soc. 1995, 117, 2953-
2954. Sessler, J. L.; Andrievsky, A.; Kra´l, V.; Lynch, V. J. Am. Chem.
Soc. 1997, 119, 9385-9392.
(5) Ishida, H.; Suga, M.; Donowaki, K.; Ohkubo, K. J. Org. Chem. 1995, 60,
5374-5375. Raposo, C.; Pe´rez, N.; Almaraz, M.; Mussons, L.; Caballero,
C.; Mora´n, J. R. Tetrahedron Lett. 1995, 36, 3255-3258. Bisson, A. P.;
Lynch, V. M.; Monahan, M.-K. C.; Anslyn, E. V. Angew. Chem. 1997,
109, 2435-2437; Angew. Chem, Int. Ed. Engl. 1997, 36, 2340-2342.
Anzenbacher, P., Jr.; Try, A. C.; Miyaji, H.; Jurs´ıkova´, K.; Lynch, V. M.;
Marquez, M.; Sessler, J. L. J. Am. Chem. Soc. 2000, 122, 10268-10272.
Snellink-Rue¨l, B. H. M.; Antonisse, M. M. G.; Engbersen, J. F. J.;
Timmerman, P.; Reinhoudt, D. N. Eur. J. Org. Chem. 2000, 165-170.
(6) Kubik, S.; Goddard, R.; Kirchner, R.; Nolting, D.; Seidel, J. Angew. Chem.
2001, 113, 2722-2725; Angew. Chem., Int. Ed. 2001, 40, 2648-2651.
(7) For examples of other anion receptors that form sandwich complexes, see:
Choi, K.; Hamilton, A. D. J. Am. Chem. Soc. 2001, 123, 2456-2457.
Hossain, M. A.; Llinares, J. M.; Powell, D.; Bowman-James, K. Inorg.
Chem. 2001, 40, 2936-2937.
9
J. AM. CHEM. SOC. VOL. 124, NO. 43, 2002 12753

A R T I C L E S
Kubik et al.
Scheme 1. Synthesis of Compounds 3 and 2
Figure 1. Top (above) and side (below) view of the calculated structure
of the iodide (purple) complex of 2. All hydrogen atoms except the ones at
nitrogens have been omitted for reasons of clarity. In the side view, H(R)
protons of a substituted and an unsubstituted proline ring are marked in
yellow. The distance of these atoms amounts to 2.5 Å.
Figure 2. ESI mass spectrum of 2 in 80% H2O/CH3OH (0.1 mM) after
addition of 0.33 equiv of each NaI, NaBr, and NaBr. The inset shows the
fine structure of the signal of the 1:1 bromide complex of 2 with the black
line representing the observed isotope distribution and the red line the
calculated one.
tions were based on the crystal structure of the iodide complex
of 1. They indicated that adipinic acid should have the
appropriate length and flexibility, making 2 a promising receptor
candidate. The optimized structure of the iodide complex of 2
is depicted in Figure 1. Its structural relationship with the
corresponding complex of 1 is high.6 In both aggregates, the
anion binds to all six peptide NH groups via hydrogen bonds,
the two cyclopeptide subunits interlock perfectly like gears, and
pairs of proline rings from different receptor moieties come in
close contact. The overall symmetry of the iodide complex of
2 is C2, however, because of the presence of the adipinic acid
spacer.
Receptor 2 was obtained by reaction of 3 with adipinic acid
under standard peptide coupling conditions in good yields and
analytically pure after recrystallization (Scheme 1). Compound
2 is somewhat less soluble than 1. Still, for electrospray mass
spectrometric investigations, the same solvent mixture could be
used as for 1 (80% H2O/CH3OH). NMR spectroscopic and
microcalorimetric experiments had to be carried out in 50%
D2O/CD3OD or 50% H2O/CH3OH, respectively.
investigations on the complex formation between anions and
macrocyclic host molecules as recently demonstrated by us and
others.6,15 In the negative ion mode, ESI MS not only allows a
direct determination of the m/z ratio of the anion complexes
formed but can also give information on their relative stabili-
ties.16
The mass spectrum of a 0.1 mM solution of 2 in 80% H2O/
CH3OH, to which 0.33 equiv of each NaCl, NaBr, and NaI was
added, is depicted in Figure 2.
In addition to the signal of the free receptor [2 - H+] (m/z
1441.6), three other prominent signals are visible in this
spectrum, whose m/z ratios correspond to complexes of the
composition [2 + Cl-] (m/z 1477.5), [2 + Br-] (m/z 1523.5),
and [2 + I-] (m/z 1569.5).17 Thus, no significant amounts of
complexes were detected, in which 2 binds two anions. Higher
complexes with, e.g., two anions bound by two molecules of 2
have the same m/z ratio as simple 1:1 complexes. Such
complexes are distinguishable from 1:1 complexes in the MS,
(15) Szumna, A.; Jurczak, J. Eur. J. Org. Chem. 2001, 4031-4039.
(16) Brady, P. A.; Sanders, J. K. M. New. J. Chem. 1998, 411-417. Cheng,
X.; Chen, R.; Bruce, J. E.; Schwartz, B. L.; Anderson, G. A.; Hofstadler,
S. A.; Gale, D. C.; Smith, R. D.; Gao, J.; Sigal, G. B.; Mammen, M.;
Whitesides, G. M. J. Am. Chem. Soc. 1995, 117, 8859-8860.
(17) The smaller signals visible in the mass spectrum can be assigned to
complexes of the composition [2 + NO2-] and [2 + NO3-]. Small amounts
of these anions were obviously present in the solutions investigated. They
originate either from the salts or the solvents used.
Complex Stoichiometry. First indications of the anion
affinity of 2 were obtained by using electrospray ionization mass
spectrometry (ESI MS). This method is well suited for
(14) The calculations have been performed at the AM1 level without considering
solvent molecules since only the influence of the spacer on the geometry
of the sandwich-type complexes of 1 should be evaluated.
9
12754 J. AM. CHEM. SOC. VOL. 124, NO. 43, 2002

A Molecular Oyster
A R T I C L E S
however, by analyzing the fine structures of the signals, which
reflect the isotope distribution of an ion and thus differ distinctly
for ions of different composition. Since the fine structures of
the signals in Figure 2 are only consistent with 1:1 complexes
between 2 and the different halides, a formation of higher
aggregates can be ruled out (inset in Figure 2 and Supporting
Information).18 This result is further corroborated by the lack
of signals in the spectrum that can be attributed to possible
complexes, in which two or more molecules of 2 bind different
anions.
The stoichiometry of the anion complexes of 2 could also be
determined by NMR spectroscopy. For 1 we have shown that
anion binding results in a downfield shift of the H(R) signal of
1
the peptide in the H NMR spectrum because of the spatial
proximity of the corresponding protons to the negative charge
density of the guest in the complex.6,9 The reduced symmetry
of 2 causes the three H(R) protons in a cyclopeptide subunit to
be nonequivalent, and the 1H NMR spectrum thus contains three
H(R) signals, with the signal of the proton located in the
substituted proline unit clearly separated from the other two
that possess very similar chemical shifts. Upon addition of
anions such as halides, sulfate, or nitrate to solutions of 2 in
50% D2O/CD3OD, all three H(R) signals shift downfield in the
1H NMR spectrum, which clearly indicates that complex
formation occurs. In addition, signals corresponding to aromatic
protons at the aminopicolinic acid subunits of 2 are also shifted,
an effect that has analogously been observed for 1.6,9 All of
these signal shifts conveniently allow an evaluation of complex
stoichiometry and stability by means of Job’s method of
continuous variation19,20 and host-guest titrations,19,21 respec-
tively.
1
Figure 3. Job plot of the iodide complex of 2 (a) and H NMR titration
curves (b) of the iodide (squares), bromide (circles), and chloride (diamonds)
complexes of 2 in 50% D2O/CD3OD. Experimental results are depicted as
symbols, and the lines represent the calculated saturation curves.
the Job plots that have been obtained from analyzing the shifts
of the aromatic protons of 2. Taken together, both the ESI MS
and the NMR investigations strongly indicate the formation of
defined 1:1 complexes between anions and 2, a stoichiometry
that is in fact entropically the most favorable one.
The Job plot of the iodide complex of 2 possesses a sharp
maximum at an equimolar ratio of host and guest, which, taking
the results of the mass spectrometric experiments also into
consideration, indicates a 1:1 complex stoichiometry (Figure
3a).19,20 In the case of the sulfate complex of 2, a similar
straightforward analysis is more difficult because an addition
of less than 1 equiv of Na2SO4 to a solution of 2 in 50% D2O/
CD3OD results in a significant line broadening of the H(R)
Complex Structure. The high affinity of 2 toward sulfate
allowed a structural investigation of the corresponding complex
in solution by using ROESY NMR spectroscopy. In the ROESY
NMR spectrum of uncomplexed 2 in 50% D2O/ CD3OD (1
mM), pronounced cross-peaks are visible between the H(R) and
H(γ) signals of the substituted prolines and between H(R) and
H(δ) signals of the unsubstituted prolines, all of which indicate
a spatial proximity of the corresponding pairs of protons within
the five membered rings. Interactions between the two cyclo-
peptide moieties of 2 cannot be observed. Such effects are
clearly visible in the spectrum of a solution of 2 (1 mM),
however, that contains 5 equiv of Na2SO4. In particular,
characteristic cross-peaks between the H(R) signals of a
substituted and an unsubstituted proline unit and ones between
the H(R) of a substituted and the H(â) of an unsubstituted proline
unit (and vice versa) occur that account for a close contact of
a substituted and an unsubstituted proline ring in the sulfate
complex of 2. The corresponding arrangement of the two
cyclopeptide subunits of 2 is consistent with the one we
calculated for the iodide complex (Figure 1). In this optimized
structure, the distance between H(R) protons of substituted and
unsubstituted proline rings (marked yellow in Figure 1) amounts
to 2.5 Å, and that between H(R) and H(â) protons, to 3.3 Å.
Assuming that on average 2 preferentially adopts a similar
solution conformation in the sulfate complex, both distances
are small enough to allow for cross-peaks in ROESY or NOESY
NMR spectra. The NMR spectroscopic results thus indicate that
the calculated structure in Figure 1 reflects characteristic aspects
of the anion complexes of 2. In the absence of suitable anions
1
signals in the H NMR spectrum that prevents a quantitative
evaluation of their shift. This effect is less pronounced for
signals of the aromatic protons of 2, which can be followed
over the whole concentration range of a Job plot. All receptor
signals become sharp again when the solution contains more
than 1 equiv of Na2SO4, and the resonances of the H(R) signals
in the corresponding NMR spectra are located downfield by
ca. 1.1 ppm with respect to their positions in the spectrum of
the free receptor. The observed line broadening in the NMR
spectrum indicates a strong interaction between 2 and sulfate,
and the fact that complex formation is almost complete after
addition of only 1 equiv of guest is consistent with a 1:1
complex stoichiometry. This interpretation is corroborated by
(18) Besides the expected [2 + SO42-] complex, a much more intensive signal
is observed in our electrospray mass spectrometric investigations of the
sulfate complex of 2, whose m/z ratio can be assigned to a complex of the
composition [22 + SO42-]. Upon collision-induced dissociation (CID) of
this ion the three fragments [2 + SO42-], [2 + HSO4-], and [2 - H+] are
obtained in almost equal amounts. A similar 2:1 complex could not be
detected in solution. Its presence in the mass spectrometric investigations
thus has to be attributed to the solvent-free conditions of the method.
(19) Connors, K. A. Binding Constants; Wiley: New York, 1987.
(20) Gil, V. M. S.; Oliveira, N. C. J. Chem. Educ. 1990, 67, 473-478.
(21) Macomber, R. S. J. Chem. Educ. 1992, 69, 375-378.
9
J. AM. CHEM. SOC. VOL. 124, NO. 43, 2002 12755

A R T I C L E S
Kubik et al.
Table 1. Stability of Various Anion Complexes of 2 in 50%
D2O/CD3OD Determined by NMR Titrations at T ) 298 Ka
slightly larger size of the sulfate ion with respect to iodide most
probably cannot account alone for the high stability of the sulfate
complex. Important additional factors that favorably contribute
to complex stability are the higher charge of sulfate and its
ability to form strong hydrogen bonds to NH groups via the
oxygens.26 Table 1 also shows that, within the error limits, the
stability of the iodide complex is not significantly affected by
the countercation. To the best of our knowledge, no other neutral
anion receptor that forms similar stable complexes in protic
solvents has been described so far.
K
a
∆δmax
radius23
Na2SO4
NaI
KI
(CH3)4NI
NaBr
NaCl
3.5 × 105
0.20
0.94
0.93
0.93
0.48
0.19
0.61
230
220
220
220
196
181
179
8900
11000
11300
5300
710
130
NaNO3
a Ka stability constants in M-1. Errors in Ka <15%. ∆δmax ) maximum
chemical shifts of the receptor signal followed during the titration in ppm:
In the case of the sulfate complex, the signal of protons in the 3 position
of an aromatic subunit were used, in all other cases, the signal of the protons
in R position of the substituted proline rings. Ionic radii of the anions are
in pm.
To support these results, we also investigated the anion
affinity of 2 by using isothermal titration calorimetry (ITC).27
This method directly measures the heat evolved or absorbed
during complex formation in dependence on the molar ratio of
receptor to ion. It thus provides a means to not only determine
the enthalpy of complex formation (∆H) but also to evaluate
the stability constant (Ka) and the free energy of complex
formation (∆G) as well as the entropic contribution (∆S). A
number of different anion receptors were characterized micro-
calorimetrically, and it was shown that, in accordance with the
simple electrostatic model of ion pairing,28 the interactions
between anions and charged receptors in polar solvents are often
endothermic making entropic factors, i.e., the release of solvent
molecules from the solvation spheres of host and guest, the
major driving force of complex formation.29 In less polar
solvents or with neutral hosts also exothermic anion complex-
ation has been observed with either a favorable30 or unfavor-
able31 contribution of entropy. In a preliminary microcalorimetric
investigation, we observed a small endothermic effect when
iodide interacted with 1 in 80% H2O/CH3OH, indicating that
complex formation is mainly driven by entropy.6 By using
receptor 2 instead of 1 and changing the solvent mixture to 50%
H2O/CH3OH (as well as using a more sensitive microcalorim-
eter), we were now able to investigate complex formation much
more accurately (Figure 4). The results of our measurements
are summarized in Table 2.
in solution, 2 seems to prefer more open conformations as
indicated by the lack of cross-peaks between the cyclopeptide
moieties in the ROESY NMR spectrum.
Complex Stability. We first used NMR titrations to deter-
mine the stability of some anion complexes of 2 (Figure 3b).19,21
Only salts of strong acids were used as guests in these
measurements to avoid protonation equilibria during the titra-
tions and the need to use buffers. The stability constant of the
sulfate complex of 2 could only be determined by using a
dilution titration and by analyzing the signal shifts of the
aromatic receptor protons because of the high stability of this
complex and the line broadening of the H(R) signals in the NMR
spectrum. The resulting Ka approaches the upper limit that can
be determined reliably by NMR titrations, and its experimental
error is therefore higher than that in the other stability constants
determined.22 The order of magnitude of the stability of the
sulfate complex should be correctly reflected in the Ka obtained,
however. All stability constants were independent of the proton
that was followed during the titration, which is a further
indication that the model of a 1:1 complex stoichiometry is
correct and that no higher equilibria have to be considered. The
results of our investigations are summarized in Table 1.
For every anion complex, the stability constant determined
by ITC is slightly smaller than the corresponding one determined
by NMR titration. Nevertheless, the agreement between the
spectroscopically and calorimetrically determined log Ka values
is quite good. The only case in which a larger difference between
the two stability constants is observed is the sulfate complex,
most probably because of the experimental error that has to be
Table 1 shows that the stability constants of the anion
complexes investigated cover a range of 3 orders of magnitude
between 102 and 105 M-1, and as a consequence, the binding
selectivity of 2 is quite good. The only stability constants that
differ by a relatively small factor of 1.7 are the ones of the
iodide and the bromide complex of 2. For all other pairs of
anion complexes, a ratio of at least 5 is observed in their Ka.24
The observed trend in complex stability seems to depend mainly
on the size of the bound anion with larger anions forming more
stable complexes most probably because they better fit into
available space between the two cyclopeptide moieties of 2.
Thus, despite the fact that, e.g., the smaller halides can form
stronger hydrogen bonds to the NH groups of 2, the better size
complementarity between iodide and the receptor cavity, which
is evident in Figure 1, is responsible for the higher complex
stability. A similar trend has also been observed for one of
Schmidtchen’s macrotricyclic anion hosts in methanol.25 The
considered for the spectroscopically determined Ka. The general
2-
trend of anion complex stabilities of 2, namely SO4 > I-
>
Br- > Cl-, is clearly confirmed by ITC, however.
Table 2 shows that complex formation is in every case
enthalpically as well as entropically favored. The fact that all
(26) Due to fast H/D exchange in 50% D2O/CD3OD, anion binding in this solvent
mixture is due to D-binding interactions with the cyclopeptide amide groups.
(27) Wadso¨, I. Chem. Soc. ReV. 1997, 26, 79-86.
(28) Bianchi, A.; Garc´ıa-Espan˜a, E. In Supramolecular Chemistry of Anions;
Bianchi, A., Bowman-James, K., Garc´ıa-Espan˜a, E., Eds.; Wiley-VCH:
New York, 1997; pp 217-275.
(29) Berger, M.; Schmidtchen, F. P. Angew. Chem. 1998, 110, 2840-2842;
Angew. Chem., Int. Ed. 1998, 37, 2694-2696. Berger, M.; Schmidtchen,
F. P. J. Am. Chem. Soc. 1999, 121, 9986-9993. Linton, B.; Hamilton, A.
D. Tetrahedron 1999, 55, 6027-6038. Arranz, P.; Bencini, A.; Bianchi,
A.; Diaz, P.; Garc´ıa-Espan˜a, E.; Giorgi, C.; Luis, S. V.; Querol, M.;
Valtancoli, B. J. Chem. Soc., Perkin Trans. 2 2001, 1765-1770. Linton,
B. R.; Goodman, M. S.; Fan, E.; van Arman, S. A.; Hamilton, A. D. J.
Org. Chem. 2001, 66, 7313-7319.
(22) Fielding, L. Tetrahedron 2000, 56, 6151-6170.
(23) Marcus, Y. Ion Properties; Marcel Dekker: New York, 1997.
(24) The increasing stability of the halide complexes from chloride through
bromide to iodide is in accordance with the mass spectrometric investiga-
tions that show an increase of signal intensity in the same order.
(25) Schmidtchen, F. P. Angew. Chem. 1977, 89, 751-753; Angew. Chem., Int.
Ed. Engl. 1977, 16, 720-721. Schmidtchen, F. P. Chem. Ber. 1981, 114,
597-607.
(30) Haj-Zaroubi, M.; Mitzel, N. W.; Schmidtchen, F. P. Angew. Chem. 2002,
114, 111-114; Angew. Chem., Int. Ed. 2002, 41, 104-107.
(31) Schmidtchen, F. P. Org. Lett. 2002, 4, 431-434.
9
12756 J. AM. CHEM. SOC. VOL. 124, NO. 43, 2002

A Molecular Oyster
A R T I C L E S
Table 3. Comparison of the Stabilities of the Sulfate and Iodide
Complexes of 1 and 2 at T ) 298 Ka
50% D O/ CD OD
80% D O/CD OD
2
3
2
3
1
4
1
K for 2
a
K
a
∆δmax
K
a
∆δmax
K 9
∆δmax
a
Na2SO4 K1 35500 360
0.59 260
0.96
0.59 96
0.44
0.77
K2
KT
8760
1270
3.15 × 106
1.22 × 105
∆G -26.0 -37.1
K1 8900 30
-13.8
0.64 26
0.82
-29.0
NaI
0.69 22
0.55
0.84
K2
KT
7670
7380
2.30 × 105
1.62 × 105
∆G -22.6 -30.6
-8.1
-29.7
a K1 and K2 in M-1. KT in M-2. Errors in stability constants for the
complexes of 1 < 40%. ∆G in kJ‚mol-1. ∆δmax ) maximum chemical
shifts of a H(R) signal of 1 or 4 in ppm.
Monotopic Receptor vs Ditopic Receptor. So far, we have
shown that a receptor composed of two covalently linked
molecules of 1 can bind anions in protic solvents with a defined
1:1 stoichiometry and high affinity. How does the anion affinity
of this compound compare to that of the parent peptide 1,
however?
Complex stability of 1 is more difficult to determine because
of the higher stoichiometry of the anion complexes of this
peptide. Saturation curves obtained from NMR titrations, e.g.,
have to be fitted to four parameters, namely K1, K2, ∆δmax1
,
Figure 4. Isothermal calorimetric titration of 2 (0.45 mM) with Na2SO4
(14 mM) in 50% H2O/CH3OH at 298 K. The upper graph shows the
measured heat pulses. The molar heats/pulse are depicted along with the
curve fit (solid line) in the lower diagram.
and ∆δmax2, to describe the underlying equilibrium correctly.
The required mathematical treatment very often leads to more
than one solution, which complicates a straightforward analysis.
We recently solved this problem by comparing the anion affinity
of 1 with the one of a structurally related cyclopeptide 4, in
which all three proline subunits are replaced by 4R-configured
hydroxyprolines.9
Table 2. Stability of Various Anion Complexes of 2 in 50%
H2O/CH3OH Determined by Isothermal Microcalorimetry at T )
298 K
ITC
NMR
log K
log K
∆G (kJ‚mol-1
)
∆H (kJ‚mol-1
)
T∆S (kJ‚mol-1
)
a
a
Na2SO4 5.54 4.55 ( 0.23 -26.0 ( 1.3 -15.0 ( 0.9 11.0 ( 2.2
NaI
KI
NaBr
NaCl
3.95 3.79 ( 0.26 -21.6 ( 1.5 -13.2 ( 1.0
4.04 3.51 ( 0.34 -20.0 ( 1.9 -14.7 ( 1.0
3.72 3.45 ( 0.39 -19.7 ( 2.2 -11.2 ( 0.9
2.85 2.51 ( 0.86 -14.3 ( 4.9 -10.5 ( 1.5
8.4 ( 2.5
5.3 ( 2.8
8.5 ( 3.1
3.8 ( 6.4
reactions are exothermic demonstrates that the inclusion of an
anion into the cavity of 2 represents an energetically more stable
situation for the system than the separated hydrated components
even in the protic solvent mixture used. The dependence of
complex stability on the type of anion is mainly determined by
∆H, and the more negative ∆H observed for the complexation
of the larger anions again reflects the importance of an optimal
fit of the guests in the available cavity. The difference in ∆H
upon iodide and sulfate complexation is consistent with the fact
that sulfate can form stronger hydrogen bonds to the NH groups
of 2. Entropic factors are also important for anion binding and
constitute ca. 40% of the observed free enthalpy of complex
formation in every case. The dependence of ∆S on the type of
anion is less obvious, however, possibly because of superimpos-
ing effects of desolvation, not necessarily complete desolvation,
of both host and guest. Our ITC measurements thus nicely
complement the NMR experiments as well as provide valuable
information on the thermodynamics of anion complexation by
2. They support the view that, by the right choice of receptor
structure, artificial receptors can be devised that strongly bind
to suitable substrates by weak interactions even in highly
competitive solvents.
This peptide adopts a similar conformation in solution as 1.
Because of the additional hydroxyl groups, 4 can only form
1:1 complexes with anions, however, whose stabilities are of
the same order of magnitude as those of the 1:1 complexes of
1. The stability constants of the anion complexes of 4 thus
provide a good quantitative estimation for the first binding step
in the interaction of 1 with anions and as a consequence allow
a reliable evaluation of the whole complexation equilibrium by
using NMR titrations. We used this strategy to determine the
stability constants of a number of anion complexes of 1 in 80%
D2O/CD3OD.9 To be able to compare the anion affinities of 1
and 2, we now repeated these measurements in 50% D2O/
CD3OD for the iodide and the sulfate complex. The results
obtained are summarized in Table 3, which also contains the
complex stabilities in 80% D2O/CD3OD for comparison.
Table 3 shows that the change of the solvent has only a minor
effect on the stability of the iodide complex of 1. The sulfate
9
J. AM. CHEM. SOC. VOL. 124, NO. 43, 2002 12757

A R T I C L E S
Kubik et al.
complex, however, is significantly more stable in 50% D2O/
CD3OD than in 80% D2O/CD3OD. This result is consistent with
our assumption that hydrogen bonds are an important factor in
the stabilization of the sulfate complexes of 1 or 2. These
interactions become considerably stronger when the polarity of
the solvent decreases, thus causing the observed increase in
complex stability. The strong dependence of sulfate complex
stability on the solvent composition also suggests, however, that
the binding selectivity of 2 for sulfate over iodide in 50% D2O/
CD3OD could be an effect of the solvent and not an intrinsic
property of the receptor. This assumption is supported by the
fact that the selectivity observed for 1 in 50% D2O/CD3OD is
not only reduced but even reversed in favor of iodide in 80%
D2O/CD3OD. To be able to study whether the same trend is
observed for 2 we will now synthesize a better water-soluble
derivative of this receptor.
A comparison of the anion complex stabilities of receptors 1
and 2 in 50% D2O/CD3OD on the basis of ∆G shows that the
anion affinity of the ditopic receptor is smaller, possibly because
the adipinic acid spacer in 2 affects the optimal complex
geometry of the 2:1 complexes of 1 unfavorably. To gain a
deeper insight into which thermodynamic factor is responsible
for the reduced anion affinity of 2, we characterized the iodide
and the sulfate complex of 1 also calorimetrically. For the sulfate
complex, we obtained an enthalpy of complex formation ∆H
of -19.3 ( 1.6 kJ‚mol-1 and a complex stability of log KT )
6.48 ( 0.08, which is in excellent agreement with the stability
constant determined by NMR titration (log KT ) 6.50). Entropy
T∆S thus amounts to 17.7 kJ‚mol-1. Moreover, the microcalo-
rimetric titration clearly confirmed the 2:1 complex stoichiom-
etry of the sulfate complex of 1.
anion complex stabilities of 2 are compared with those of 4 (or
with K1 of the complexes of 1). Doubling the number of receptor
sites in 2 with respect to 4 causes a 2-fold increase in the free
energy of complex formation, which translates into a 140-fold
increase of complex stability (100-fold increase of K1 of 1) in
the case of the sulfate complex. The receptor properties of 2
thus almost exactly reflect the individual contributions of the
two cyclopeptide subunits. A larger cooperative effect is
observed for iodide binding. In this case, ∆G is 2.7 times larger
than that of the iodide complex of 4 (and 2.6 times larger than
the ∆G corresponding to K1 of 1), which translates into a ca.
340 times increase in the stability constant. This comparison
thus clearly demonstrates the cooperativity of the two cyclo-
peptide moieties of 2 in anion complexation. An additional
stabilizing effect that seems to operate in the sandwich-type
anion complexes of 1 is most probably compensated by an
unfavorable influence of the adipinic acid spacer in 2. Future
investigations will show if cooperativity can be further improved
by a structural optimization of the spacer between the two
cyclopeptide subunits.
Conclusion
In conclusion, we have shown that by covalently linking two
cyclopeptide molecules derived from 1 with a suitable spacer,
an artificial receptor is accessible that strongly binds anions in
50% water/methanol. In this solvent, we observed a particularly
high affinity and selectivity for sulfate. Microcalorimetric
investigations showed that anion binding is enthalpically as well
as entropically driven. We use the term molecular oyster for
this type of receptor because the two perfectly interlocking
receptor subunits that are connected via a flexible hinge resemble
the two halves of a shell. In the presence of suitable anions,
this shell closes and binds the guests in an inner cavity like
pearls. We expect that receptor properties and solubility of such
anion hosts can be controlled by varying the nature (and the
number) of the spacer unit. Investigations in this context are
currently underway.
The more exothermic reaction observed in the interaction
between sulfate and 1 (for 2, ∆H ) -15.0 ( 0.9 kJ‚mol-1
)
supports our assumption that the sulfate complex of 1 is more
stable than that of 2. The fact that also entropy favors the
complex formation of the monotopic receptor was surprising
because we expected that a chelate effect, which is usually
discussed in terms of entropy, would act in favor of the bis-
(cyclopeptide). A similar entropic disadvantage for complex
formation has also been reported for cyclodextrin dimers,
however, and this result was interpreted in terms of a reduced
flexibility of the ditopic hosts in the complex.32 The large
increase in stability observed for the complexes of cyclodextrin
dimers with respect to their monotopic counterparts has therefore
been ascribed to an enthalpic chelate effect.32 In comparing the
sulfate affinity of receptors 1 and 2, we could detect neither an
entropic nor an enthalpic chelate effect, however.
For iodide complexation of 1, the microcalorimetric titration
gave a ∆H for complex formation in 50% H2O/CH3OH, which
is of the same order of magnitude as that of the iodide complex
of 2 (1, ∆H ) -14.8 ( 2.8 kJ‚mol-1; 2, ∆H ) -13.2 ( 1.0
kJ‚mol-1). The smaller stability of the iodide complex of 2 must
therefore also be caused by an unfavorable entropic effect.
Unfortunately, we were not able to get a good fit for the
calorimetric binding curve in this case so that the exact
contribution of entropy could not be quantified.
Experimental Section
General Methods. Analyses were carried out as follows: melting
points, Bu¨chi 510 apparatus; optical rotation, Perkin-Elmer 241 MC
digital polarimeter (d ) 10 cm); NMR, Varian VXR 300, Bruker DRX
500 equipped with an automatic sampler; FT IR, Bruker Vector 22
FT-IR spectrometer; FAB MS, Finnigan MAT 8200; ESI MS, Bruker
Esquire 3000; microcalorimetry, Thermal Activity Monitor with 4 mL
high performance calorimetric unit and titration microreaction cell,
Thermometric AB; elemental analysis, Pharmaceutical Institute of the
Heinrich-Heine-University, Du¨sseldorf, Germany; RP chromatography,
MERCK LiChroprep RP-8 (40-63 µm) prepacked column size B (310-
25). The following abbreviations are used: Bn, benzyl; BOC, tert-
butoxycarbonyl; DIEA, N-ethyldiisopropylamine; TFA, trifluoroacetic
acid; PyCloP, chlorotripyrrolidinophosphonium hexafluorophosphate;
TBTU, O-(1H-benzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium tet-
rafluoroborate; Hyp, L-hydroxyproline; Pro, L-proline; APA, 6-ami-
nopicolinic acid; Adi, adipinic acid.
Job Plot. Equimolar solutions of 2 (0.5 mM) and of the guest in
50% D2O/CD3OD were prepared and mixed in various ratios. 1H NMR
spectra of the solutions were recorded, and the change in chemical shifts
of the proline H(R), APAH(3), and APAH(5) protons were analyzed.
NMR Titrations. Stock solutions of the guest (50 µmol/500 µL for
titrations with 1 or 4, 2.5 µmol/500 µL for titrations with 2) in D2O
and a peptide (1 µmol/500µL for titrations with 1 or 4, 0.5 µmol/500µL
That the covalent linkage of two subunits of 1 does have a
pronounced effect on receptor properties is evident when the
(32) Zhang, B.; Breslow, R. J. Am. Chem. Soc. 1993, 115, 9353-9354. Breslow,
R.; Halfon, S.; Zhang, B. Tetrahedron 1995, 51, 377-388.
9
12758 J. AM. CHEM. SOC. VOL. 124, NO. 43, 2002

A Molecular Oyster
A R T I C L E S
for titrations with 2) in 0.002% TMS/CD3OD were prepared. In total,
11 NMR tubes were set up for titrations with 2 or 4 and 21 for titrations
with 1 by adding increasing amounts of the guest solution (0-500 µL)
to 500 µL aliquots of the host solution. All samples were made up to
a volume of 1 mL with D2O, and the respective 1H NMR spectra were
recorded. The chemical shifts of the H(R), APAH(3), or APAH(5)
protons of the peptide were referenced against the internal standard
used, and plotted against the ratio of guest/host concentration. From
the resulting saturation curves, Ka and ∆δmax were calculated by the
suitable nonlinear least-squares fitting method described for 1:119,21 or
1:233 complexes using the SIGMA Plot 3.0 (Jandel Scientific) software
package.
DMSO) δ 1. 24 + 1.34 (2 s, 9H; tBuCH3), 2.14 (m, 1H; HypC(â)H),
2.37 (m, 1H; HypC(â)H), 2.44 (s, 3H; TsCH3), 3.46 (dt, 2J ) 12.3 Hz,
1H; HypC(δ)H), 3.54 (dd, 2J ) 12.4 Hz, 3J ) 3.5 Hz, 1H; HypC(δ)H),
4.55 (m, 1H; HypC(R)H), 5.11 (m, 1H; HypC(γ)H), 5.39 (s, 2H;
PhCH2), 7.35-7.53 (m, b, 7H; PhH + TsH(3)), 7.83 (m, 3H; APAH(3)
+ TsH(4)), 7.99-8.02 (2 t, 3J ) 8.2 Hz, 1H; APAH(4)), 8.29 + 8.34
(2 d, 3J ) 8.2 Hz, 1H; APAH(5)), 11.04 + 11.11 (2 s, 1H; APANH).
Anal. Calcd for C30H33N3O8S (Mr ) 595.7): C, 60.49; H, 5.58; N,
7.05. Found: C, 60.42; H, 5.60; N, 6.94.
Tosylated Cyclopeptide. For the synthesis of the linear hexapeptide
precursor of a cyclopeptide moiety of 2, the BOC deprotected dipeptide
H-(4R)-4TsHyp-APA-OBn and a tetrapeptide with proline subunits
BOC-[Pro-6APA]2-OH was coupled as described previously by us for
similar compounds (see for example, supplementary information of ref
6). For the cyclization, this hexapeptide (1.03 g, 1 mmol) was
deprotected at both ends of the peptide chain and dissolved in a mixture
of degassed DMF (40 mL) and DIEA (1.04 mL, 6 mmol). The resulting
solution was added dropwise over the course of 4 h to a solution of
TBTU (1.60 g, 5 mmol) and DIEA (0.42 mL, 2.4 mmol) in degassed
DMF (200 mL) at 80 °C. If necessary, the pH of the reaction mixture
was adjusted afterward to ca. 9 by adding more DIEA, and stirring
was continued for 1 h at 80 °C. The solvent was then evaporated in
vacuo, and the product was isolated from the residue by chromato-
graphic workup. An initial purification step was carried out using a
silica gel column (acetone). The material recovered was further purified
on an RP-8 column. For this, it was dissolved in a small amount of
DMF and applied to a column conditioned with 1,4-dioxane/H2O, 1:10.
The eluent composition was gradually changed until the pure product
eluted (1,4-dioxane/H2O, 1:1). The material thus obtained was dissolved
in acetone (20 mL), and the resulting solution was poured slowly in
diethyl ether (200 mL) under stirring. Stirring was continued for 15
min, and the precipitate was filtered off and dried in vacuo: yield 0.33
g (40%); mp 205 °C (dec); [R]25D ) -516.5 (c ) 2, DMF); 1H NMR
(300 MHz, [d6]DMSO, 25 °C, TMS) δ 1.84 (m, 4H; ProC(γ)H2), 2.06
(m, 2H; ProC(â)H), 2.23 (m, 1H; HypC(â)H), 2.45 (s, 3H; TsCH3),
2.62 (m, 2H; ProC(â)CH), 2.96 (m, 1H; HypC(â)H), 3.55-3.75 (m,
b, 6H; ProC(δ)H2 + HypC(δ)H2), 5.14 (m, 1H; HypC(γ)H), 5.54 (m,
2H; ProC(R)H), 5.66 (dd, 3J(Hax,Hax) ) 7.0 Hz, 3J(Hax,Heq) ) 6.2 Hz,
Microcalorimetry. In a typical experiment, solutions of the guest
(15 mM) in 50% H2O/CH3OH were injected stepwise (8 µL per step)
at a rate of 1 µL/s to a solution of 1 or 2 in the same solvent mixture
(0.75 mL, 0.5 mM). The next titration step was always started after
reaching chemical and thermal equilibrium. The measured heat flow
was recorded as function of time and converted into enthalpies by
integration of the appropriate reaction peaks. Dilution effects were
corrected by subtracting the results of a blank experiment with a solution
of 50% H2O/CH3OH in place of the receptor solution under identical
experimental conditions. The association parameters were evaluated
by means of a computer program generated using the MATLAB
programming package (Mathworks Inc., Natick, MA).
Dipeptide BOC-(4R)-Hyp-APA-OBn. 6-Amino-2-picolinic acid
benzyl ester (1.71 g, 7.50 mmol) (supplementary information of ref
6), BOC-(4R)-L-hydroxyproline (2.61 g, 11.3 mmol), and PyCloP (4.76
g, 11.3 mmol) were dissolved in CH2Cl2 (150 mL). At room temper-
ature, DIEA (3.9 mL, 22.6 mmol) was added dropwise, and then the
reaction mixture was stirred for 5 d. The solvent was subsequently
evaporated in vacuo, and the residue was subjected to a silica gel
column (hexane/ethyl acetate, 3:1; ethyl acetate). All fractions contain-
ing the product were collected and evaporated to dryness in vacuo.
The remaining residue was dissolved in CH2Cl2 (100 mL), and this
solution was poured into diethyl ether (800 mL) under stirring. After
10 min, the precipitate was filtered off, and the filtrate was again
evaporated to dryness. From the residue, pure product was isolated by
another chromatographic purification step on a silica gel column
3
3
(hexane/ethyl acetate, 1:15): yield 2.41 g (73%); mp 78-82 °C; [R]25
) -39.0 (c ) 2, MeOH); H NMR (300 MHz, [d6]DMSO, 100 °C,
1H; HypC(R)H), 7.14 (d, J ) 8.2 Hz, 1H; APAH(3)), 7.22 (d, J )
D
8.2 Hz, 1H; APAH(3)), 7.27 (d, 3J ) 8.2 Hz, 1H; APAH(3)), 7.45 (m,
b, 3H; APAH(5)), 7.54 (d, J ) 7.9 Hz, 2H; TsH(3)), 7.69-7.78 (m,
b, 3H; APAH(4)), 7.88 (d, J ) 8.2 Hz, 2H; TsH(2)), 9.70 + 9.72 +
1
3
TMS) δ 1.32 (s, 9H; tBuCH3), 1.97 (m, 1H; HypC(â)H), 2.15 (m, 1H;
3
2
2
HypC(â)H), 3.30 (d, J ) 11.0 Hz, 1H; HypC(δ)H), 3.49 (dd, J )
11.0 Hz, 3J ) 4.8 Hz, 1H; HypC(δ)H), 4.30 (m, 1H; HypC(γ)H), 4.55
(t, 3J ) 7.6 Hz, 1H; HypC(R)H), 4.71 (d, 3J ) 3.8 Hz, 1H; OH), 5.38
(s, 2H; PhCH2), 7.39 (m, b, 5H; PhH), 7.76 (dd, 3J ) 7.5 Hz, 4J ) 1.0
9.83 (3 s, 3 × 1H; APANH); 13C NMR (75 MHz, [d6]DMSO, 25 °C,
TMS) δ 22.3 + 22.4 (ProC(γ)), 30.6 (TsCH3), 32.3 + 32.5 (ProC(â)),
38.2 (HypC(â)), 47.9 + 48.1 (ProC(δ)), 54.0 (HypC(δ)), 59.9
(HypC(R)), 61.4 + 61.5 (ProC(R)), 79.0 (HypC(γ)), 115.4 + 115.6 +
116.2 (APAC(3)), 119.5 + 119.6 + 119.7 (APAC(5)), 127.5 (TsC(2)),
130.3 (TsC(3)), 132.6 (TsC(4)), 138.8 + 138.9 (APAC(4)), 145.3
(TsC(1)), 148.3 + 148.5 + 148.6 (APAC(2)), 151.3 + 151.7 + 152.0
(APAC(6)), 165.7 + 166.1 + 166.3 (APACO), 169.6 + 170.8 + 171.1
(HypCO/ProCO). Anal. Calcd for C40H39N9O9S‚2H2O (Mr ) 857.9):
C, 56.00; H, 5.05; N, 14.69. Found: C, 55.87; H, 5.07; N, 14.42. FAB-
MS [m/z (relative intensity)]: 822 (30) [M + H+].
3
3
Hz, 1H; APAH(3)), 7.94 (t, J ) 7.7 Hz, 1H; APAH(4)), 8.26 (dd, J
) 7.7 Hz, 4J ) 1.0 Hz, 1H; APAH(5)), 10.47 (s, 1H; APANH). Anal.
Calcd for C23H27N3O6‚2H2O (Mr ) 477.5): C, 57.85; H, 6.54; N, 8.80.
Found: C, 58.16; H, 6.21; N, 8.64.
Tosylated Dipeptide BOC-(4R)-4TsHyp-APA-OBn. BOC-(4R)-
Hyp-APA-OBn (2.21 g, 5.00 mmol) was dissolved in a mixture of
methylene chloride/pyridine, 1:1 (10 mL). Toluenesulfonyl chloride
(4.77 g, 25 mmol) was added in one portion, and the reaction mixture
was stirred at room temperature overnight. Afterward, the solvent was
evaporated in vacuo, and the residue was dissolved in ethyl acetate.
The resulting solution was washed three times with 10% aqueous
Na2CO3 and three times with water, and the solvent was again removed.
The product was isolated from the residue chromatographically (hexane/
Cyclopeptide Azide. The tosylated cyclopeptide (0.45 g, 0.55 mmol)
was dissolved in DMF (10 mL). After the addition of sodium azide
(0.18 g, 2.75 mmol), the reaction mixture was heated to 80 °C for 6 h.
The solvent was then removed in vacuo, and the product was isolated
from the residue by chromatographic workup using a silica gel column
(CH2Cl2/MeOH, 5:1). Trituration with diethyl ether of the material
recovered afforded an off white solid: yield 0.33 g (87%); mp 184 °C
(dec); [R]25D ) -530.0 (c ) 2, DMF); 1H NMR (500 MHz, [d6]DMSO,
25 °C, TMS) δ 1.84 (m, 4H; ProC(γ)H2), 2.05 (m, 2H; ProC(â)H),
2.41 (td, 2J ) 13.6 Hz, 1H; HypC(â)H), 2.57 (m, 2H; ProC(â)H), 2.86
(m, 1H; HypC(â)H), 3.59 (m, 2H; ProC(δ)H), 3.71 (m, 3H; ProC(δ)H
+ HypC(δ)H), 3.83 (dd, 3J ) 6.0 Hz, 2J ) 12.9 Hz, 1H; HypC(δ)H),
ethyl acetate, 1:1). The product solidified upon drying in vacuo: yield
1
2.74 g (92%); mp 67-73 °C; [R]25 ) -26.6 (c ) 2, MeOH); H
D
NMR (300 MHz, [d6]DMSO, 25 °C, TMS) (some signals split in the
spectrum due to a slow rotation around the tertiary amide bond; at
elevated temperatures, decomposition of the product occurred in
(33) Bonar-Law, R. P.; Sanders, J. K. M. J. Am. Chem. Soc. 1995, 117, 259-
3
271.
4.48 (m, 1H; HypC(γ)H), 5.60 (t, J ) 7.0 Hz, 1H; ProC(R)H), 5.64
9
J. AM. CHEM. SOC. VOL. 124, NO. 43, 2002 12759

A R T I C L E S
Kubik et al.
(t, 3J ) 7.1 Hz, 1H; ProC(R)H), 5.85 (dd, 3J(Hax,Hax) ) 8.7 Hz,
3J(Hax,Heq) ) 3.6 Hz, 1H; HypC(R)H), 7.15 (d, 3J ) 8.2 Hz, 1H;
APAH(3)), 7.34 (d, 3J ) 8.2 Hz, 1H; APAH(3)), 7.43 (m, 3H; APAH(3)
+ 2 APAH(5)), 7.49 (d, 3J ) 7.3 Hz, 1H; APAH(5)), 7.71 (t, 3J ) 7.9
Hz, 1H; APAH(4)), 7.75 (t, 3J ) 7.6 Hz, 2H; APAH(4)), 9.58 + 9.61
+ 9.69 (3 s, 3 × 1H; APANH); 13C NMR (125 MHz, [d6]DMSO, 25
°C, TMS) δ 22.2 + 22.3 (ProC(γ)), 32.3 + 32.4 (ProC(â)), 36.9
(HypC(â)), 48.0 (ProC(δ)), 52.2 (HypC(δ)), 56.4 (HypC(γ)), 60.1
(HypC(R)), 61.3 + 61.5 (ProC(R)), 114.9 + 115.6 + 116.1 (APAC(3)),
119.4 + 119.7 (APAC(5)), 138.8 + 139.0 + 139.1 (APAC(4)), 148.4
+ 148.5 + 148.7 (APAC(2)), 151.4 + 151.7 + 151.9 (APAC(6)), 165.7
+ 165.8 (APACO), 169.6 + 170.9 + 171.0 (HypCO/ProCO); IR (KBr)
0.21 g (73%); mp > 250°C; [R]25D ) -538.3 (c ) 2, DMF); 1H NMR
(500 MHz, [d6]DMSO, 25 °C, TMS) δ 1.27 (m, 4H; AdiC(â)H2), 1.74-
2.09 (m, b, 18H; AdiC(R)H2 + ProC(γ)H2 + ProC(â)H + HypC(â)H),
2.57 (m, 4H; ProC(â)H), 2.84 (m, 2H; HypC(â)H), 3.43 (m, 2H;
HypC(δ)H), 3.59 (m, 4H; ProC(δ)H), 3.68 (m, 4H; ProC(δ)H), 3.87
3
(m, 2H; HypC(δ)H), 4.28 (m, 2H; HypC(γ)H), 5.54 (t, J ) 6.9 Hz,
3
2H; HypC(R)H), 5.60 (m, 4H; ProC(R)H), 7.20 (d, J ) 8.2 Hz, 2H;
APAH(3)), 7.22 (d, 3J ) 8.2 Hz, 2H; APAH(3)), 7.27 (d, 3J ) 8.2 Hz,
2H; APAH(3)), 7.41 (d, 3J ) 7.6 Hz, 2H; APAH(5)), 7.42 (d, 3J ) 7.6
3
Hz, 2H; APAH(5)), 7.45 (d, J ) 7.6 Hz, 2H; APAH(5)), 7.74 (m,
3
6H; APAH(4)), 7.89 (d, J ) 6.3 Hz, 2H; AdiNH), 9.60 + 9.62 +
9.68 (3 s, 3 × 1H; APANH); 13C NMR (125 MHz, [d6]DMSO, 25 °C,
TMS) δ 22.2 + 22.3 (ProC(γ)), 24.6 (AdiC(â)), 32.3 + 32.4 (ProC(â)),
34.9 (AdiC(R)), 37.2 (HypC(â)), 46.0 (HypC(γ)), 48.0 (ProC(δ)), 51.9
(HypC(δ)), 60.4 (HypC(R)), 61.3 + 61.4 (ProC(R)), 115.5 + 115.7 +
115.8 (APAC(3)), 119.6 (APAC(5)), 138.8 + 138.9 + 139.0 (APAC(4)),
148.4 + 148.5 (APAC(2)), 151.4 + 151.8 + 151.9 (APAC(6)), 165.8
+ 165.9 + 166.0 (APACO), 170.5 + 170.8 + 170.9 (HypCO/ProCO/
AdiCO). Anal. Calcd for C72H74N20O14‚6H2O (Mr ) 1551.6): C, 55.74;
H, 5.59; N, 18.05. Found: C, 55.47; H, 5.55; N, 17.95. FAB-MS: [m/z
(relative intensity)]: 1443 (3) [M + H+].
2107 cm-1 (azide). Anal. Calcd for C33H32N12O6‚2.5H2O (Mr
)
737.7): C, 53.73; H, 5.06; N, 22.78. Found: C, 54.02; H, 5.17; N,
22.51. FAB-MS [m/z (relative intensity)]: 693 (30) [M + H+].
Bis(cyclopeptide) (2). The azide prepared in the previous step (310
mg, 0.45 mmol) was dissolved in CH2Cl2/MeOH, 1:1 (75 mL). After
the addition of 1 N HCl (0.45 mL) and 10% Pd/C (100 mg), the
resulting reaction mixture was subjected to hydrogenation at 1 atm for
18 h. The catalyst was then filtered off by passage through a layer of
Celite and washed with methanol. The combined filtrate and washings
were evaporated to dryness in vacuo, the residue was dissolved in
CH2Cl2, and the solvent was evaporated again. The crude product (3)
was dried in vacuo and was used for the coupling reaction without
further purification.
Cyclopeptide 3 (270 mg, 0.4 mmol) and adipinic acid (30 mg, 0.2
mmol) were dissolved in DMF (20 mL). In succession, TBTU (140
mg, 0.44 mmol) and DIEA (220 µL, 1.28 mmol) were added, and the
resulting mixture was stirred 3 h at room temperature. Afterward, the
solvent was evaporated in vacuo, and the product was isolated by
chromatographic workup. For this, the residue was dissolved in a small
amount of DMF. The resulting solution was applied to a RP-8 column
conditioned with 1,4-dioxane/H2O, 1:10, and the eluent composition
was gradually changed to 1,4-dioxane/H2O, 1:2, with which pure
product eluted. The material recovered was dissolved in boiling
methanol (50 mL). Under reflux, water (200 mL) was slowly added,
and the solution was kept at room-temperature overnight. The precipi-
tated product was filtered off, washed with water, and dried: yield
Acknowledgment. This work was sponsored by the Deutsche
Forschungsgemeinschaft. S.K. thanks Mrs. D. Kubik for her
invaluable assistance in the synthetic work, Dr. K. Schaper for
his help in the AM1 calculations, and Prof. G. Wulff as well as
Prof. H. Ritter for their support. The mass spectrometric
investigations were funded by the Volkswagen Foundation.
Supporting Information Available: 1H and ROESY NMR
spectra of 2 and its sulfate complex, observed and calculated
isotope distribution of the complexes between 2 and iodide,
bromide, and chloride, and Job plots and NMR titration curves
of anion complexes of 1 and 2. This material is available free
of charge via the Internet at http://pubs.acs.org.
JA026996Q
9
12760 J. AM. CHEM. SOC. VOL. 124, NO. 43, 2002