Cyclic hexapeptides with free carboxylate groups as new receptors for monosaccharides

Article abstract of DOI:10.1021/ol016158l

(Equation presented) Cyclic hexapeptides composed of alternating L-proline and 3-aminobenzoic acid subunits with substituents on the aromatic subunits that contain free carboxylate groups are able to bind monosaccharides in 4% CD₃OD/CDCl_3<

Full text of DOI:10.1021/ol016158l

ORGANIC
LETTERS
2001
Vol. 3, No. 17
2637-2640
Cyclic Hexapeptides with Free
Carboxylate Groups as New Receptors
for Monosaccharides
Joachim Bitta and Stefan Kubik*
Institut fu¨r Organische Chemie und Makromolekulare Chemie,
Heinrich-Heine-UniVersita¨t, UniVersita¨tsstr. 1, D-40225 Du¨sseldorf, Germany
kubik@uni-duesseldorf.de
Received May 23, 2001
ABSTRACT
Cyclic hexapeptides composed of alternating L-proline and 3-aminobenzoic acid subunits with substituents on the aromatic subunits that
contain free carboxylate groups are able to bind monosaccharides in 4% CD OD/CDCl3. The binding selectivity of these peptides depends on
3
the structure of the substituents on the aromatic subunits.
Because of the importance of carbohydrate recognition in
nature, the design of artificial receptors for saccharide
complexation is currently intensively being pursued in
supramolecular chemistry. The various approaches in this
area of research have recently been reviewed by A. P. Davis.1
In this paper, we add a new type of receptor to the number
of systems studied so far that is based on a cyclic hexapeptide
composed of alternating L-proline and 3-aminobenzoic acid
residues. This macrocyclic host has currently only been used
for the complexation of anions and cations,2 but our structural
assignment has shown that its cavity dimension should also
be well suited for the inclusion of monosaccharides.2a The
peptide certainly lacks functional groups with which sugars
could interact, but these groups can be introduced either in
the aromatic or the natural amino acid subunits along the
peptide cavity.
In natural carbohydrate receptors, carboxylate groups in
the side chains of glutamic acid or aspartic acid are important
factors in the binding of the substrate by hydrogen bonding,3
and we therefore decided to use the same functional groups
to induce a carbohydrate affinity in our cyclopeptides. All
derivatives tested (2a-c) contain a cyclic array of carboxy-
late groups around their cavity that stems from amino acid
substituents attached to the aromatic cyclopeptide subunits.
These macrocycles are thus related to the carbohydrate
receptors, in which phosphonate groups are used for guest
binding.4
To facilitate the synthesis of cyclopeptides with peripheral
substituents, we used the cyclic hexapeptide 1 as a molecular
scaffold, in which additional carboxyl groups on the 3-ami-
nobenzoic acid subunits are available for further modifica-
(1) Davis, A. P.; Wareham, R. S. Angew. Chem., Int. Ed. Engl. 1999,
38, 2978-2996.
(2) (a) Kubik, S.; Goddard, R. J. Org. Chem. 1999, 64, 9475-9486. (b)
Kubik, S.; Goddard, R. Eur. J. Org. Chem. 2001, 311-322. Ishida et al.
have reported on the phosphate binding of a similar peptide in DMSO:
Ishida, H.; Suga, M.; Donowaki, K.; Ohkubo, K. J. Org. Chem. 1995, 60,
5374-5375.
(3) Lemieux, R. U. Chem. Soc. ReV. 1989, 18, 347-374. Quiocho, F.
A. Pure Appl. Chem. 1989, 61, 1293-1306. Weis, W. I.; Drickamer, K.
Annu. ReV. Biochem. 1996, 65, 441-473.
10.1021/ol016158l CCC: $20.00 © 2001 American Chemical Society
Published on Web 07/25/2001

by deprotection and deprotonation of the carboxyl groups
designated for substrate binding (Scheme 1). All products
were fully characterized by 1H and 13C NMR spectroscopy,
mass spectrometry, and elemental analysis.
Scheme 1. Synthesis of Cyclopeptides 1 and 2a-ca
The 1H NMR spectra of 2a-c in d6-DMSO are consistent
with averaged C3-symmetrical conformations of the macro-
cycles. Interestingly, deprotonation of the carboxylic acids
in the final step of the synthesis of 2c caused a significant
downfield shift of the signal of the glutamic acid NH protons,
whereas the chemical shift of the ring NH protons is almost
unaffected. The same effect is to a lesser extent also observed
for 2b (Figure 1).
a (a) BOC2O; (b) AllBr/NaHCO3; (c) BnBr/NaHCO3; (d) TFA;
(e) BOC-L-proline/PyCloP/DIEA; (f) [Pd(P(4-Me2NPh)Ph2)4]/
morpholine; (g) HCl/1,4-dioxane; (h) TBTU/DIEA; (i) H2/10% Pd/
C; (j) 2a: n ) 2, R1 ) Bn, R2 ) i-Pr; 2b: n ) 1, R1 ) i-Pr, R2 )
Bn; 2c: n ) 2, R1 ) i-Pr, R2 ) Bn, PyCloP/DIEA; (k)
N(n-Bu)4+OH-.
tions. Because these groups line the wide rim of the peptide
cavity, the overall conformation of 1 should be similar to
the one we have determined for the unsubstituted parent
compound.2a
Figure 1. Possible cyclic arrangements of the amino acid substit-
uents on the aromatic subunits of 2a-c. The resonance of the NH
protons before (a) and after (b) deprotonation of the carboxyl
groups, the temperature dependency of the NH resonances in d6-
DMSO (c), and the calculated lengths (d) and angles (e) of the
N-H‚‚‚O hydrogen bonds are indicated.
Downfield shifts of NH signals usually indicate that the
corresponding protons are involved in hydrogen bonding.
The small temperature dependency of the NH resonance in
the aromatic substituents of 2a-c indicate that this type of
interaction occurs in fact intramolecularly.5 The correspond-
ing hydrogen bonds are presumably formed between the NH
groups and the neighboring carboxylate groups in which case
they lock the aromatic substituents in a cyclic conformation.
In Figure 1, energy minimized cyclic arrangements of the
substituents in 2a-c are depicted. These structures are in
accordance with calculated conformations of R,ω aminocar-
boxylic acids that contain intramolecular hydrogen bonds.6
Peptide 1 was prepared according to Scheme 1. For the
chain elongation of the orthogonally protected dipeptide in
solution, standard peptide methodology could be used, since
the allyl ester and the N-BOC group can be cleaved
selectively without affecting the benzyl ester. In the final
step of the synthesis of 1, the remaining benzyl esters were
removed cleanly by hydrogenation.
The syntheses of peptides 2a-c were completed by
coupling suitably substituted amino acids with 1, followed
(4) Das, G.; Hamilton, A. D. J. Am. Chem. Soc. 1994, 116, 11139-
11140. Anderson, S.; Neidlein, U.; Gramlich, V.; Diederich, F. Angew.
Chem., Int. Ed. Engl. 1995, 34, 1596-1600. Cotero´n, J. M.; Hacket, F.;
Schneider, H.-J. J. Org. Chem. 1996, 61, 1429-1435. Neidlein, U.;
Diederich, F. Chem. Commun. 1996, 1493-1494. Das, G.; Hamilton, A.
D. Tetrahedron Lett. 1997, 38, 3675-3678. Ba¨hr, A.; Felber, B.; Schneider,
K.; Diederich, F. HelV. Chim. Acta 2000, 83, 1346-1376.
(5) Kessler, H. Angew. Chem., Int. Ed. Engl. 1982, 21, 512-523.
(6) Ramek, M. Int. J. Quantum Chem., Quantum Biol. Symp. 1990, 17,
45-53. Seebacher, U.; Ramek, M. Amino Acids 1994, 7, 223-230.
2638
Org. Lett., Vol. 3, No. 17, 2001

The shortest hydrogen bond occurs in the seven-membered
ring of the glutamic acid derivative 2c, which explains the
largest downfield shift of the NH signal upon deprotonation
in this peptide. The intramolecular hydrogen bonds in the
peripheral substituents reduce the flexibility of 2a-c and
stabilize certain orientations of the carboxylate groups. Both
effects could contribute to receptor selectivity, but overall
peptides 2a-c should still be flexible enough to adapt to
the steric demand of different guest molecules.
glucosides as potential guests, saccharides that are insoluble
in pure chloroform.
Even in this competitive solvent mixture, interactions
between the cyclopeptides and carbohydrates are still detect-
able as illustrated by the effects of, e.g., 2c on prominent
signals of R-D-methylglucopyranoside in the 1H NMR
spectrum (Figure 2). Again, a small upfield shift of the signal
In the 1H NMR spectra of 2a-c in CDCl3, the signals of
all peptide protons are significantly broadened, indicating
either a slow conformational equilibrium or intermolecular
association. We still carried out initial investigations on the
monosaccharide affinity of the peptides in this solvent. A
comparison of the 1H NMR spectrum of an equimolar
mixture of R-D-octylglucopyranoside and 2c with the spectra
of the individual components at the same concentration
clearly revealed an interaction of the cyclopeptide with the
sugar. Most importantly, the prominent doublet of the
anomeric proton of the glucoside is slightly but reproducibly
shifted upfield in the presence of 2c (ca. -0.015 ppm). Such
upfield shifts are usually observed when guest molecules are
included into a receptor cavity that is lined by aromatic
subunits,7 and we believe that they also indicate complex
formation in the present case. No influence of sugar binding
on the spectrum of 2c could be detected. However, the signals
of the tetra-n-butylammonium protons are also shifted upfield
upon complex formation. This effect is strongest for the
R-CH2 protons of the cation and decreases with increasing
distance of the protons from the ammonium nitrogen. It is
reasonable to assume that in CDCl3 the anionic receptor and
the quaternary ammonium ion form a close ion pair with
the R-CH2 protons of the cation located in close proximity,
possibly even hydrogen bonded to the receptor carboxylate
groups. Hydrogen bonding capabilities of methylene groups
adjacent to ammonium centers have in fact been demon-
strated in solution and in the solid state.8 Such hydrogen
bonding would lead to a downfield shift of the R-CH2 protons
in the 1H NMR spectrum of 2c, and the upfield shift that is
observed when the guest is added therefore illustrates the
dissociation of the cation-carboxylate ion pair that precedes
sugar binding.
Figure 2. 1H NMR spectra of 2c (a), R-D-methylglucopyranoside
(c), and a 1:1 mixture of both compounds (b) in 4% CD3OD/CDCl3
(c ) 1 mM, 25 °C).
of the anomeric proton can be observed, the upfield shift of
the resonance of the sugar OCH3 group is significantly larger,
however.
Receptor 2c (and also the other two peptides) thus binds
monosaccharides in the presence of a ca. 10 000-fold excess
of methanol molecules that compete in binding to the
carboxylate groups. A Job plot furthermore revealed a
defined 1:1 stoichiometry for the complex between 2c and
R-D-methylglucopyranoside.9 This indicates that the monosac-
charide is included into the peptide cavity upon complex
formation where it can interact with all three carboxylate
groups simultaneously. Linear analogues of 2c such as the
hexapeptide 3 and the dipeptide 4 lack the preorganization
of the cyclic derivative, and their interaction with monosac-
charides is thus less specific. In the case of 4, the interaction
is in fact too weak to be determined quantitatively (∆δ <
0.001 ppm for the sugar protons), and in the case of 3, a
A competition between cation and glucoside complicates
the complex equilibrium in CDCl3, and we therefore tried
to suppress the interaction of the ammonium cation with the
peptides by increasing the polarity of the solvent. Indeed,
no significant interactions between the ammonium ions and
the carboxylate groups of the peptides were detected in
CDCl3 containing 4% CD3OD. This solvent mixture has the
additional advantages that the receptor signals in the 1H NMR
spectra become sharper, and it is possible to use methyl
(7) Ma, J. C.; Dougherty, D. A. Chem. ReV. 1997, 97, 1303-1324.
Lhota´k, P.; Shinkai, S. J. Phys. Org. Chem. 1997, 10, 273-285.
(8) Reetz, M. T.; Hu¨tte, S.; Goddard, R. J. Am. Chem. Soc. 1993, 115,
9339-9340. Deakyne, C. A.; Meot-Ner (Mautner), M. J. Am. Chem. Soc.
1999, 121, 1546-1557. Houk, K. N.; Menzer, S.; Newton, S. P.; Raymo,
F. M.; Stoddart, J. F., Williams, D. J. J. Am. Chem. Soc. 1999, 121, 1479-
1487. Cousins, G. R. L.; Furlan, R. L. E., Ng, Y.-F., Redman, J. E., Sanders,
J. K. M. Angew. Chem., Int. Ed. Engl. 2001, 40, 423-428.
(9) We determined a 1:1 complex stoichiometry also for other receptor/
monosaccharide combinations, e.g., 2b and â-D-methylglucopyranoside. In
addition, the excellent agreement of the saturation curves obtained
experimentally in our NMR titrations with the ones calculated on the basis
of 1:1 complex formation indicates that this stoichiometry holds for all
complexes investigated (see Supporting Information).
Org. Lett., Vol. 3, No. 17, 2001
2639

peptide/sugar ratio of 2:3 was determined in the aggregate.
This result clearly demonstrates the importance of the
macrocyclic structure of peptides 2a-c in carbohydrate
complexation.
The exact nature of the role of water in the interactions
between carbohydrates and our peptides has still to be
clarified, however.
Table 1 shows that the complexes between peptides 2a-c
and all glycosides tested possess stability constants (and
maximum chemical shifts) of the same order of magnitude.
The selectivity of the receptors with respect to a specific
epimer or anomer is obviously only moderate. However,
certain trends are evident. In general, peptide 2b forms more
stable complexes than the other two receptors. This indicates
that the orientation of the carboxylate groups in the aspartic
acid residues of 2b seem to be somewhat better suited for
monosaccharide binding than those in the substituents of the
other two peptides. Furthermore, the hexopyranosides are
significantly better bound than R-D-methylribofuranoside.
This could be due to the larger ring size of the pyranosides
but also to the additional hydroxyl group in hexoses with
which the peptides can interact. In contrast to the OCH3
protons of the methyl glycosides, the OCH2 protons in the
octyl chain of â-octylglucopyranoside are shifted downfield
upon complex formation with 2b. This most probably
indicates that, as a result of the steric bulk of the octyl chain,
the geometry of this complex differs from the ones of the
methyl glycosides.
The shift of the methylglycoside OCH3 protons upon
complex formation allows the quantitative determination of
1
complex stability by NMR titrations. Since the H NMR
spectra of 2a-c are independent of the concentration in the
region relevant for such quantitative investigations (0.2-
2.0 mM), an intermolecular association of the peptides has
not to be considered. The stability constants of peptides 2a-c
with a variety of glycosides are summarized in Table 1. They
Sanders et al. have shown that the ability of sugars to form
intermolecular hydrogen bonds to the solvent (methanol in
CDCl3) or to receptor binding sites increases in the order
R-galactose < â-galactose ≈ R-glucose < â-glucose <
â-mannose < R-mannose.11 Carbohydrate receptors whose
complex stabilities increase in the same order are therefore
essentially nonselective. Table 1 shows that the complex
stabilities of peptides 2a-c deviate from the order proposed
by Sanders in some instances, e.g., the â-methylglucopyra-
noside is bound best by all receptors. This suggests that the
peptides do possess a certain selectivity in carbohydrate
recognition.
Table 1. D-Glycoside Association Constants of Complexes of
Peptides 2a-c in 4% CD3OD/CDCl3 at 298 Ka
Ka/∆δmax
2a
2b
2c
R-methylglucopyranoside
â-methylglucopyranoside
â-octylglucopyranoside
420/-0.06 660/-0.07 550/-0.05
550/-0.04 810/-0.07 650/-0.06
560/+0.05
R-methylmannopyranoside 450/-0.04 700/-0.04 440/-0.06
R-methylgalactopyranoside 300/-0.05 790/-0.06 390/-0.04
â-methylgalactopyranoside 400/-0.06 540/-0.08 430/-0.05
R-methylribofuranoside
160/-0.07 290/-0.08 190/-0.08
A more detailed knowledge of the structure of the
complexes would certainly help in the evaluation of the
receptor properties of 2a-c. However, in our eyes these
initial investigations already show that the substituted cy-
clopeptides presented in this paper represent a promising
basis for the design of a new type of monosaccharide
receptors. By better preorganizing the peptides for substrate
complexation, it should in principle be possible to increase
the binding selectivity. Also, other functional groups can be
used for the interactions with carbohydrates. Investigations
in these directions are currently underway.
a Ka stability constant in M-1, error limits for Ka < 20%; ∆δmax is the
maximum chemical shift in ppm for the sugar OCH3 protons except in the
case of â-octylglucopyranoside, where the shift of the glucose C(3)H has
been followed.
were calculated from the shift of prominent sugar signals
by using the mathematical treatment for 1:1 complex
equilibria.10
It has to be noted that reproducible results in the NMR
titrations could only be obtained when the water content of
the solvent mixture used was at least 0.002%. The presence
of water is obviously important for complex formation. We
can currently only speculate that water molecules contribute
to the hydrogen bonding network between the hydroxyl
groups of the monosaccharides and the peptide carboxyl
groups and thus stabilize the whole aggregate. A similar
effect of water has also been observed by Sanders et al. in
the carbohydrate binding of another macrocyclic receptor.11
Acknowledgment. The funding of this project by the
Deutsche Forschungsgemeinschaft is gratefully acknowl-
edged. S.K. also thanks Prof. G. Wulff for his support.
Supporting Information Available: Syntheses of 2a-
c, 3, and 4. Selected 1H NMR spectra of compounds 2a-c,
saturation curves of NMR titrations, and Job plots. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(10) Connors, K. A. Binding Constants, Wiley: New York, 1987.
(11) Bonar-Law, R. P.; Sanders, J. K. M. J. Am. Chem. Soc. 1995, 117,
259-271.
OL016158L
2640
Org. Lett., Vol. 3, No. 17, 2001