Cyclotetrapeptide mimics based on a 13-membered, partially modified retro-inverso structure

Article abstract of DOI:10.1002/ejoc.200700932

Cyclic tetrapeptides (CTP) incorporating a distinct β-amino acid represent realistic conformationally homogeneous CTP analogs, and they can find application in medicinal chemistry as turn mimics and scaffolds. In this respect, we synthesized a small libra

Full text of DOI:10.1002/ejoc.200700932

FULL PAPER
DOI: 10.1002/ejoc.200700932
Cyclotetrapeptide Mimics Based on a 13-Membered, Partially Modified Retro-
Inverso Structure
Luca Gentilucci,*[a] Giuliana Cardillo,[a] Alessandra Tolomelli,[a] Santi Spampinato,[b]
Antonino Sparta,[b] and Federico Squassabia[a]
Keywords: Medicinal chemistry / Inhibitors / Peptidomimetics / Conformation analysis / Cell adhesion
Cyclic tetrapeptides (CTP) incorporating a distinct β-amino
acid represent realistic conformationally homogeneous CTP
analogs, and they can find application in medicinal chemistry
as turn mimics and scaffolds. In this respect, we synthesized
tures. The comparison with a β-amino acid containing CTP
revealed that the PMRI–CTPs can be regarded as alternative
β-turn scaffolds. To validate the structural deductions based
on conformational analysis, we synthesized a couple of dia-
a small library of CTP-mimics based on a 13-membered, par- stereomeric RGD analogs (5, 6) as integrin inhibitors, and we
tially modified retro-inverso (PMRI) structure containing a
1,2-diamine as a β-amino acid mimetic and malonic acid as
a glycine surrogate. Conformational analysis revealed that
the PMRI–CTP models containing an unsubstituted diamine
(1, 2) exhibited a certain flexibility. On the contrary, the intro-
duction of a chiral, substituted 1,2-diamine rendered the
backbone more rigid (3, 4) and induced type I β-turn struc-
found a clear relationship between the 3D display of the
pharmacophores and biological activity. In particular, 6
showed good efficacy in inhibiting fibronectin adhesion to
the αvβ3 integrin-expressing human melanoma cells SK-
MEL-24 (EC50 = 3.7ϫ10–7).
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2008)
c[βAla-ψ(NHCO)-Ala-ψ(NHCO)-Gly--Phe] (2), c[βPhe-
ψ(NHCO)-Ala-ψ(NHCO)-Gly-Phe] (3), c[βPhe-ψ(NHCO)-
Ala-ψ(NHCO)-Gly--Phe] (4), and related compounds
c[βAla-ψ(NHCO)-Asp-ψ(NHCO)-Gly-Arg] (5), c[βAla-
ψ(NHCO)-Asp-ψ(NHCO)-Gly--Arg] (6) (Figure 1).
Introduction
Many biologically active peptides and proteins exert their
functions by means of relatively small regions of their
folded surfaces.[1] As a consequence, much effort has been
dedicated to the design of smaller, conformationally defined
peptides or analogs[2] that mimic these localized folded re-
gions.[1,3]
Within the family of cyclic peptides, cyclotetrapeptides
(CTPs) composed of all α-amino acids can be considered
truly minimalist turn mimics,[4] but their potential utility in
medicinal chemistry is often hampered by difficult synthesis
and scarce conformational definition in a polar environ-
ment. On the other hand, the incorporation of a distinct β3-
or β2-amino acid residue in the CTP sequence may result in
the stabilization of the overall secondary structure.[4b,4c,5]
As a continuation of our recent investigations on the use
of cyclic peptides and analogs as restricted mimics of biolo-
gically active, naturally occurring peptides,[6] we report the
partially modified retro-inverso[7] CTP (PMRI–CTP)
models c[βAla-ψ(NHCO)-Ala-ψ(NHCO)-Gly-Phe] (1),
Figure 1. Structures of PMRI–CTPs 1–6.
These 13-membered PMRI–CTPs containing a 1,2-di-
amine as a β-amino acid analog represent easily available
CTP mimetics and show promising applications in medici-
nal chemistry and biochemistry as topologically defined
scaffolds. Indeed, the low molecular weight, the increased
lipophilicity, and the presence of modified peptide bonds is
expected to improve the bioavailability and ADMET pro-
file.[2]
[a] Department of Chemistry “G. Ciamician”, University of Bolo-
gna,
Via Selmi 2, 40126 Bologna, Italy
Fax: +39-051209456
E-mail: luca.gentilucci@unibo.it
[b] Department of Pharmacology, University of Bologna,
Via Irnerio 42, 40126 Bologna, Italy
E-mail: spampi@biocfarm.unibo.it
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Eur. J. Org. Chem. 2008, 729–735
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
729

L. Gentilucci et al.
FULL PAPER
Table 1. ∆δ/∆t values (ppb/K) of amide protons for 1–4, 13, 14, as
determined by VT-1H-NMR analysis in [D6]DMSO at 400 MHz
over the range 298–348°K.[a]
Results and Discussion
Peptides 1–6 were easily obtained in a few steps by cycli-
zation of the corresponding linear precursors obtained in
turn by divergent synthesis in solution (see also Supporting
Information).
As a representative example, the preparation of 3 (and 4)
is depicted in Scheme 1 (Experimental Section). Chiral Cbz-
Compound
AlaNH
/-PheNH
NHb[a]
NHa[a]
1
2
3
4
–5.0
–4.4
–5.3
–4.1
–0.05
–0.5[b]
–1.2
–0.7
–1.3
+1.4
–1.3
–5.5[c]
–4.1
–5.0[c]
–3.0
–0.2
[8]
1,2-diamine 7 was obtained by reduction of Cbz-Phe-NH2
AspNH
/-ArgNH
NHb
NHa
with BH3.[9] Coupling of 7 under standard conditions with
Fmoc-Phe-OH, and deprotection of the resulting dipeptide
by treatment with 2  dimethylamine in THF afforded 8 in
good yield. Dipeptide 9 was readily prepared from Meld-
rum’s acid and NH2-Ala-OBz. The coupling of dipeptides
8 and 9 gave the fully protected linear tetrapeptide. The
removal of the protecting groups with H2 and Pd/C, and
finally cyclization with DPPA, gave cyclopeptide 3 in good
yield and purity after semipreparative RP-HPLC.
13
14
–6.3
–4.8
–5.7
–4.9
–0.2
–0.8
–1.5
+0.6
[a] See Figure 1. [b] Determined by gCOSY experiments recorded
at 298 and 313°K. [c] PheNH and AlaNH in 3 are superimposed.
in
organic
polar
environments,
in
particular
DMSO.[4b,4c,7f,12] Accordingly, the NMR experiments on li-
pophilic cyclopeptides 1–4 were performed at 400 MHz in
[D6]DMSO.
The 2D-ROESY analysis furnished details on cyclopep-
tide backbone conformations (Supporting Information,
Tables S1–S4). Apart from the inversion of configuration at
Phe, the most significant difference between 1 and 2 (and
between 3 and 4) is limited to the /-PheNH region (Fig-
ure 3), as revealed by the comparison of the ROESY cross
peaks between the COCH2CO protons and /-PheNH or
AlaNH (Figure 2). In 1, AlaNH is in close proximity of one
of the two COCH2CO protons (at 3.2 ppm), whereas
PheNH is close to the other (δ = 3.1 ppm). In 2, both
AlaNH and -PheNH are close to the same COCH2CO
proton (δ = 3.3 ppm), and -PheNH shows also a medium
intensity cross peak with the second COCH2CO (δ =
2.8 ppm). The behavior of 3 is similar to that of 1, whereas
the behavior of 4 reproduces that of 2.
The occurrence of ROESY cross peaks between -
PheNH and both the COCH2CO protons, which should
presumably occupy opposite sides of molecular plane,
seems to suggest that in 2 and 4 -PheNH could flip above
and below the molecular plane.
Further, the existence in the ROESYs of 1 and 2 of
strong cross peaks for NHb-PheNH and NHb-PheHα, and
The H-NMR analysis of models 1–4 revealed a single for NHa-AlaNH and NHa-AlaHα, is indicative that NHa
set of resonances (Supporting Information). However, the and NHb (Figure 1) can flip above and below the backbone
amide protons generally showed modest JNH,CH values plane, as PheNH and PheHα, AlaNH and AlaHα are trans
(Ͻ9 Hz), which suggests a fast equilibrium rather than (Supporting Information, Tables S1–S4).
Scheme 1. Synthesis of PMRI–CTPs 3, 4.
1
complete conformational rigidity.[10]
Low-energy conformations consistent with spectroscopic
Nevertheless, variable temperature (VT) 1H-NMR analy- analyses were obtained by restrained MD by using the dis-
sis denoted that the NHa and NHb protons (Figure 1) ap- tances deduced from ROESY as constraints (see also Sup-
pear to be involved in H-bonds (|∆δ/∆t| lower than or close porting Information). For both 1 and 2, the computations
to 1 ppb/K,[11] Table 1), which supports the existence of a gave different structures, each of which showed some viola-
population of ordered structures with definite secondary tions of constraints. The structures having the lowest in-
structural elements.
ternal energy and the least number of violations are shown
The conformational features of the PMRI–CTPs were in- in Figure 3.
vestigated by spectroscopic and molecular dynamics (MD)
The structure proposed for 1 can be characterized by two
analyses. We could not perform experiments in water, for inverse γ-turns, one centered on Phe (φ = –101, ψ = +51)
the peptides were practically insoluble. Many peptides or and the other on Ala (φ = –98, ψ = +50), whereas the repre-
peptidomimetics of interest described in the literature are sentative structure of 2 is roughly compatible with an in-
not highly soluble in water and were studied experimentally verse γ-turn centered on Ala (φ = –130, ψ = +68).
730
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 729–735

Cyclotetrapeptide Mimics with a Partially Modified Retro-Inverso Structure
Figure 3. Representative lowest-energy structures of 1–4 calculated
by restrained MD with the least number of violations of ROESY
data (Supporting Information).
NMR-derived geometry as a starting structure. During this
period, the analysis of the trajectories showed modest back-
bone flexibility (Figure 4).
Figure 2. Inset of ROESY observed for 1 (A) and 2 (B) in [D6]-
DMSO at 400 MHz (room temp.) showing the cross peaks between
COCH2CO and /-PheNH/AlaNH protons.
On the contrary, more than 90% of the structures of 3
calculated by restrained MD did not show any violation of
restraints, and all were well ordered. Apparently, the intro-
duction of a further substituent on the diamine is sufficient
to confer upon the structure conformational homogeneity.
The representative structure of 3 in Figure 3 is compati-
ble with a type I β-turn centered on Phe–diamine and a
second one centered on Ala–diamine. However, this struc-
ture does not confirm the presence of H-bonds as predicted
by VT-1H-NMR analysis, probably for the occurrence of a
fast equilibrium between slightly different geometries,
whose average in the NMR timescale gives the NMR-de-
rived structure.
Figure 4. Overlay of minimized structures of 3 sampled every
500 ps during the unrestrained 5.0 ns MD simulation in explicit
water, and representative structures A, B characterized by explicit
H-bonds and secondary structural elements.
Moreover, the simulation clearly revealed the occurrence
In order to investigate the dynamic behavior of 3 in water of backbone structures A and B with NHa and NHb alter-
to determine the eventual secondary structures stabilized by natively participating in explicit H-bonds (Figure 4). The
H-bonds, unrestrained MD simulations were conducted in dihedral angles of Phe–diamine (for A: φi+1 = –54, ψi+1
=
a box of explicit water molecules for 5.0 ns by using the –40, φi+2 = –96, ψi+2 = –28; for B: φi+1 = –83, ψi+1 = –38,
Eur. J. Org. Chem. 2008, 729–735
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
731

L. Gentilucci et al.
FULL PAPER
φi+2 = –92, ψi+2 = –37), and Ala–diamine (for A: φi+1
=
the recently reported CTP containing a β2-amino acid c[(S)-
–65, ψi+1 = –33, φi+2 = –116, ψi+2 = –28; for B: φi+1 = –73, β2hPhe--Pro-Lys-Phe][5a] (E, Figure 6). In normal cyclo-
ψi+1 = –29, φi+2 = –89, ψi+2 = –37) fragments are consistent peptides, the 3D structure is generally determined for the
with two symmetric type I β-turns,[4b,4c] although only the most part by the specific array of the stereogenic centers,
latter fragment can be considered a realistic type I β-turn whereas the nature of the residues is less important.[10,13]
mimic, for the position of the benzyl group.
Concerning 4, restrained MD essentially gave two fami-
lies of structures, represented by 4a and 4b (Figure 3), with
nearly the same energy, differing exclusively in the opposite
orientation of -PheNH.
Both 4a and 4b are compatible with a type I β-turn cen-
tered on the Ala–diamine fragment, but 4b also shows a
type I β-turn on the Phe–diamine moiety. The two struc-
tures reasonably represent distinct conformers in equilib-
rium.
However, 4b shows a smaller number of distance viola-
tions, and a structure more compatible with VT-NMR spec-
troscopic data, which are suggestive of two H-bonds on
NHa and NHb.
Unrestrained MD simulations performed on 4a evi-
denced the formation of a single explicit H-bond involving
Figure 6. Comparison of the backbones of representative structures
of c[βPhe-ψ(NHCO)-Ala-ψ(NHCO)-Gly-Phe] (3) and c[(S)-
β2hPhe--Pro-Lys-Phe] E (some side chains have been omitted for
NHb, whereas unrestrained MD performed on 4b showed
the occurrence of H-bonds involving NHa and/or NHb
(Figure 5). The simulations failed to reproduce the inver-
clarity).
Peptides 3 and E share a common stereochemistry
sion of -PheNH; evidently, this rotation is slow relative to pattern. The (S)-diamine residue in 3 matches the (S)-β2Phe
the time selected for the simulation (5.0 ns).
in E; the  configuration of Ala in partially retro-inverso 3
corresponds to the  configuration of Pro in the normal
peptide; the malonyl moiety, ψ(NHCO)-Gly, can act as a
-amino acid; and the last residue is the same for 3 and E.
Comparison of the two structures proves that the two scaf-
folds show distinct features (Figure 6), in particular for the
presence of a cis β-Phe--Pro ω bond in E, and for the
position of the β-Phe side chain, which in E is placed below
the plane, whereas in 3 it remains above.
Finally, in order to validate the structural deductions
based on spectroscopic and MD analyses, we made a pre-
liminary test of the potential utility of the PMRI–CTP scaf-
folds in medicinal chemistry. We observed that the main
structural difference between 1 and 2, and between 3 and
4, is the distance between the β carbons of Ala and /-
Figure 5. Representative structures C, D of 4b calculated by unre-
strained MD, characterized by explicit H-bonds and secondary
structural elements.
Overall, the conformational analysis of 1–4 revealed that Phe; for example this distance is 8.4 Å in 1 and 7.6 Å in 2.
the PMRI–CTPs containing an unsubstituted diamine The latter value nicely fits the requisites generally accepted
show some flexibility, especially in the diamine region, for the antagonists of αvβ3 integrins. Peptidomimetic antago-
whereas the PMRI–CTPs that contain a substituted di- nists based on cyclic cores carrying a guanidino and car-
amine are conformationally much more stable. On the other boxylic-type functionality have been very often adopted for
hand, the 3D display of the side chains is well-defined for investigating how the 3D side chains affect biological ac-
all the models.
tivity.[4b,4c,13] In particular, it was exhaustively demonstrated
This observation is not unexpected; for instance, it has that a comparatively short distance between the β carbons
been well documented that cyclic penta- or hexapeptides (7–8 Å) favored selective binding to αvβ3 integrins over
can adopt defined 3D structures and clear side-chain posi- other RGD-binding integrins.[13] Furthermore, different ad-
tions, depending on the amino acid stereochemistry, but the ditional lipophilic flanking groups[14] at the N- and C-ter-
presence of a glycine unit generally introduces a fast local minus very often proved to be crucial for high affinity and
backbone switch.[10,13]
selectivity.[15] To avoid any effect of such flanking groups,
The cyclic structures described constitute distinct, alter- we synthesized RGD mimetics 5 and 6, which are based
native 13-membered scaffolds with respect to those com- on the structures of 1 and 2, containing the unsubstituted
posed of α-amino acids plus a β-residue (see Introduction). diamine, and we assayed their biological activity as integrin
For instance, peptidomimetic 3, c[βPhe-ψ(NHCO)-Ala- inhibitors. The lack of full rigidity does not constitute a
ψ(NHCO)-Gly-Phe], can be considered a PMRI analog of insurmountable impediment in terms of biological activity,
732
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 729–735

Cyclotetrapeptide Mimics with a Partially Modified Retro-Inverso Structure
for the side-chain topology with respect to the peptide bond
orientation is of higher relevance.[5b,13c,13e]
In addition, the retro-peptide bonds can be regarded as
true peptide-bond surrogates. The presence of these modi-
The structural analysis of 13 and 14, which are the fully fied peptide bonds is expected to increases the biological
protected precursors of 5 and 6 (Supporting Information, half lives of the compounds. Further, the cyclic structure
Tables S5 and S7), gave unexpected results. Comparison of and the presence of intramolecular H-bonds increase the
1
the H NMR, VT-NMR, and ROESY analyses indicated molecular lipophilicity. The combination of increased sta-
that although 1 and 13 are structurally analogous, 2 [-Ala1, bility and increased lipophilicity can strongly impact the
-Phe3] and 14 [-Asp1, -Arg3] differ in the mirror disposi- overall bioavailability of compound based on PMRI–CTP
tion of the backbone (see Figures 3 and 7). In particular, in scaffolds.
2 the residue having its NH positioned on the molecular
plane has the  configuration, whereas in 14 the NH that
lies on the molecular plane belongs to the  residue. Apart
Experimental Section
from this difference, the expected distance between the β
carbon atoms is retained.
General Methods: Unless stated otherwise, standard chemicals were
obtained from commercial sources and used without further purifi-
cation. The SK-MEL-24 cell line (human malignant melanoma)
was obtained from the American Tissue Type Collection (ATCC);
BSS from Invitrogen; fetal bovine serum from Cambrex; essential
medium from Eagle; PBS from Cambrex; Fibronectin (FN) from
Sigma–Aldrich. Flash chromatography was performed on silica gel
(230–400 mesh) with mixtures of distilled EtOAc and MeOH. Ana-
lytical RP-HPLC was performed with an ODS column, 4.6-µm par-
ticle size, 100 Å pore diameter, 250 mm, DAD 210 nm, with H2O/
CH3CN (9:1Ǟ2:8) in 20 min at a 1.0 mLmin–1 flow, followed by
10 min at the same composition. Semipreparative RP-HPLC was
performed on a C18 column, 7-µm particle size, 21.2ϫ150 mm,
from H2O/CH3CN (9:1Ǟ3:7) in 15 min at a 12 mLmin–1 flow. Flu-
orometry to evaluate the number of adherent cells was performed
with a multilabel counter.
1H NMR spectra were recorded by using 5-mm tubes by using
0.01  peptide at room temperature. Chemical shifts are reported
as δ values relative to the solvent peak. VT-1H-NMR experiments
were performed over the range 298–348°K. 2D spectra were ac-
quired in the phase-sensitive mode and processed by using a 90°
shifted, squared sine-bell apodization. 1H-NMR resonances were
assigned from gCOSY and ROESY spectra. gCOSY experiments
were recorded with a proton spectral width of 3103 Hz. ROESY
experiments were recorded with a 300 ms mixing time with a pro-
ton spectral width of 3088 Hz.
Figure 7. Representative structures of 13 and 14 calculated by re-
strained MD with the lowest internal energy and the least number
of violations of ROESY data (tBu and Mtr protective groups have
been omitted for clarity).
To evaluate the potential activity of 5 and 6 as αvβ3 inte-
grin inhibitors, we tested their ability to inhibit the adhesion
of an αvβ3 integrin-expressing cell line, SK-MEL-24, to the
specific ligand fibronectin.[16] Fibronectin was immobilized
on each well of a 96-assay plate. SK-MEL-24 cells were
preincubated with various peptide concentrations, and dis-
pensed on the wells. After removal of nonadherent cells, the
number of adherent cells was quantified by fluorometry.
The activity of potential antagonists was determined by the
number of cells adhered relative to the control.
RGD mimetics 5 and 6 showed IC50 values of 5ϫ10–4
and 3.7ϫ10–7 , respectively. The inhibitory activity dis-
played by 6 is attractive,[16] and the different performance
of 5 seems to confirm that in 6 the pharmacophores actu-
ally adopt spatial positions compatible with an efficient li-
gand–receptor interaction, which substantiates the struc-
tures proposed on the basis of spectroscopic and computa-
tional analyses.
Representative Procedure for the Synthesis of PMRI–CTPs: Com-
pound 3 was prepared by cyclization of the corresponding linear
precursor (Scheme 1), obtained in turn by in solution coupling of
dipeptides 8 and 9.
Benzyl [2-(2-Amino-3-phenylpropionylamino)-1-benzylethyl]carba-
mate (8): HOBt (0.16 g, 1.2 mmol) was added to a stirred solution
of Fmoc-Phe-OH (0.39 g, 1.0 mmol) in DCM/DMF (9:1, 15 mL)
at room temp. After 10 min, 7[8,9] (0.28 g, 1.0 mmol), EDCI·HCl
salt (0.24 g, 1.2 mmol), and TEA (0.40 mL, 3.0 mmol) were added
while stirring at room temp. After 4 h, the mixture was diluted with
DCM, and the solution was washed with HCl (0.5 ), and a satu-
rated aqueous solution of Na2CO3. The organic layer was dried
with Na2SO4, and the solvent was removed under reduced pressure.
The dipeptide was isolated by crystallization from DCM/Et2O
(85%). Fmoc-group deprotection was performed by treatment with
5 mL of 2  dimethylamine in THF at room temp. After 20 min
the solution was evaporated under reduced pressure, and the treat-
ment was repeated. After final evaporation of the solution, the resi-
due was triturated twice with n-pentane. Crude 8 (0.36 g, 85%, 83%
pure by RP-HPLC) was used without further purification. MS
(ES): m/z = 431.1 [M + 1].
Conclusions
The easily available PMRI–CTPs constitute a novel kind
of 13-membered cyclotetrapeptide β-turn mimics. The in-
troduction of the retro modification by insertion of a substi-
tuted 1,2-diamine and a diacid induced a peculiar backbone
conformation. This conformation can be an alternative to
that of cyclotetrapeptides containing a normal sequence of
amino acids. Therefore, the PMRI–CTPs can find applica-
tions as effective scaffolds for designing biologically active
molecules with a predictable, atypical 3D display of the
pharmacophores, which is useful for spatial screening of di-
verse bioactive conformations.
Benzyl 2-(2-Carboxyacetylamino)propionate (9): A solution of Mel-
drum’s acid (0.85 g, 6 mmol) and NH2-Ala-OBz (0.90 g, 5 mmol)
Eur. J. Org. Chem. 2008, 729–735
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
733

L. Gentilucci et al.
FULL PAPER
in CH3CN (10 mL) was warmed to 70 °C under an inert atmo-
sphere. After 5 h, cyclohexane/Et2O (4:1, 40 mL) was added, and
the oily residue that precipitated was separated. This residue was
washed with a hexane/Et2O mixture (4:1, 2ϫ20 mL), and the re-
sulting dense oil was diluted with EtOAc (40 mL) and washed with
HCl (0.1 , 2ϫ5 mL). The organic layer was dried with Na2SO4,
and the solvent was evaporated under reduced pressure (temp.
Ͻ40 °C) to give 9 as a waxy solid, which was used for the following
step without further purification (0.91 g, 69%, 80% pure by NMR
analysis). 1H NMR (300 MHz, [D6]DMSO): δ = 1.40 (d, J =
7.2 Hz, 3 H, CH3), 3.40 (s, 2 H, COCH2CO), 4.59–4.66 (m, 1 H,
AlaHα), 5.20 (s, 2 H, CH2Ph), 7.25–7.41 (m, 5 H, Ph), 7.80 (br. d,
1 H, NH), 9.40–10.1 (br. s, 1 H, COOH) ppm.
scent and 3000 cycles of conjugated gradient (convergence:
0.01 kcalmol–1 Å–1). The structures that showed the lowest internal
energy and the least number of violations of the experimental data
were selected and analyzed.
MD simulation in explicit water was performed for 5.0 ns at 298°K
by using the AMBER force field in a 30ϫ30ϫ30 Å box of stan-
dard TIP3P models of equilibrated water,[18] with a minimum sol-
vent–solute distance of 2.3 Å, at constant temperature and pressure
(Berendsen Scheme,[19] bath relaxation constant 0.2).
Cell Adhesion Assays
Inhibition of the Adhesion of an αvβ3 Integrin-Expressing Cell Line,
SK-MEL-24, to Fibronectin: As a reference compound we tested
the αvβ3 integrin antagonist AcDRGDS, which showed an IC50 of
0.2ϫ10–7 .[16]
c[βPhe-ψ(NHCO)-Ala-ψ(NHCO)-Gly-Phe] (3): Dipeptides
9
(0.13 g, 0.5 mmol) and 8 (0.22 g, 0.5 mmol) were coupled under the
same conditions used for the synthesis of 8 in the presence of HOBt
(0.082 g, 0.6 mmol), EDCI·HCl salt (0.12 g, 0.6 mmol), and TEA
(0.20 mL, 1.5 mmol), in DCM/DMF (9:1, 10 mL) at room temp.
After 6 h, the usual workup afforded the protected tetrapeptide,
which was isolated by crystallization from DCM/Et2O (0.23 g,
68%, 86% pure by RP-HPLC). MS (ES): m/z = 679.2 [M + 1].
Experimental Details: SK-MEL-24 cells were routinely grown as
adherent monolayers in minimum essential medium in Earle’s BSS
supplemented with 10% fetal bovine serum, nonessential amino
acids and 1 m sodium pyruvate in a humidified environment con-
taining 5% CO2 and 95% air at 37 °C.
Assays were performed as described in the literature,[16] with the
following modifications. Cells were harvested with 0.05% trypsin/
0.53 m EDTA and washed three times with 0.5 mgmL–1 trypsin
inhibitor in PBS. The final cell pellet was resuspended at a concen-
tration of 106 cells per mL in PBS. FN was diluted in PBS
(10 µgmL–1) and 100 µL of this solution was added to each well of
a 96-assay plate and incubated for 2 h at 37 °C. The wells were
aspirated and washed once with PBS. Wells were blocked with
100 µL of 1% BSA in MEM for 1 h at 37 °C and washed once
with PBS. Cells were preincubated with various concentrations of
peptides 5, 6 for 30 min at room temp. agitating, and 50 µL of this
suspension were added to FN-coated wells (in quadruplicate). The
plate was incubated for 90 min at room temp. Following the incu-
bation period, the wells were aspirated and washed three times (or
until no more cells were seen in the background wells) with PBS.
The number of cells attached was determined by a hexosaminidase
reaction.[20] Hexosaminidase substrate (p-nitrophenol-N-acetyl-β-
-glucosaminide; 50 µL, at 7.5 m in 0.1  citrate buffer, pH = 5,
mixed with an equal volume of 0.5% triton X-100 in water) was
added to each well. The plate was incubated 1 h at 25 °C followed
by addition of 100 µL stopping solution (50 m glycine buffer, pH
= 10.4, containing 5 m EDTA). Standard curves prepared by add-
ing a different number of cells were run in parallel in the same
plate. The number of adherent cells was quantified by fluorometry.
The activity of potential antagonists was determined by the number
of cell adhered relative to the control. Aliquots of this suspension
were added to fibronectin-coated wells. The concentration–re-
sponse curves were analyzed by a least-square curve-fitting com-
puter program and IC50 values were determined (GraphPad, San
Diego, CA, USA).
Protecting group removal was performed by treatment with H2 and
cat. Pd/C in EtOH (10 mL) at room temp. After 6 h, the mixture
was filtered through Celite, and the solvent was evaporated under
reduced pressure to give the deprotected linear tetrapeptide (0.15 g,
96%, 82% pure by RP-HPLC), which was used without further
purification. MS (ES): m/z = 454.2 [M + 1].
Peptide Cyclization: A mixture of the deprotected tetrapeptide
(0.15 g, 0.32 mmol), DPPA (0.15 mL, 0.68 mmol), and NaHCO3
(0.43 g, 5.1 mmol) in DMF (70 mL) was stirred at room temp. Af-
ter 48 h, the solvent was distilled at reduced pressure, the residue
was diluted with water, and the mixture was extracted three times
with DCM. The solution was evaporated at reduced pressure, and
the residue was precipitated from DCM/Et2O. Semipreparative RP-
1
HPLC gave 3 (0.092 g, 68%, 95% pure by RP-HPLC). H NMR
(400 MHz, [D6]DMSO): δ = 1.03 (d, J = 7.2 Hz, 3 H, CH3), 2.57
(dd, J = 10.0, 14.0 Hz, 1 H, PhCH2CHNHa), 2.75 (dd, J = 5.6,
14.0 Hz, 1 H, PhCH2CHNHa), 2.89 (dd, J = 8.7, 15.0 Hz, 1 H,
PheHβ), 2.94–3.11 (m, 2 H, NHbCH2), 3.09 (dd, J = 3.6, 15.0 Hz,
1 H, PheHβ), 3.13 (d, J = 10.3 Hz, 1 H, COCH2CO), 3.17 (d, J =
10.3 Hz, 1 H, COCH2CO), 3.86 (dq, J = 5.7, 7.3 Hz, 1 H, AlaHα),
3.88–3.98 (m, 1 H, NHaCH), 4.30–4.39 (m, 1 H, PheHα), 6.16 (d,
J = 8.4 Hz, 1 H, NHa), 6.82 (br. t, 1 H, NHb), 7.03–7.16 (m, 10
H, ArH), 9.00 (d, J = 6.0 Hz, 2 H, PheNH + AlaNH) ppm. MS
(ES): m/z = 437.1 [M + 1].
Conformational Analysis: The restrained MD simulations were con-
ducted by using AMBER[17] force field with a distance dependent
ε = 4.0xr. A 50-ps simulation at 1200°K was used for generating
100 random structures that were subsequently subjected to a 20 ps
restrained MD with a 50% scaled force field at the same tempera-
ture, followed by 20 ps with full restraints (distance force constant:
7 kcalmol–1 Å–2), after which the system was cooled in 5 ps to
50°K. For the absence of Hα-COCH2CO or Hα-NHCH (i, i + 1)
ROESY cross peaks, the ω bonds were set at 180° (force constant:
16 kcalmol–1 Å–2). Only ROESY-derived constraints were included
in the restrained molecular dynamics. H-bond interactions were not
included as well as torsion angle restraints. ROESY intensities were
classified according to a calibration against the intensity of geminal
protons. Very strong, strong, medium, and weak signals were asso-
ciated to distances of 2.3, 2.6, 3.0, and 4.0 Å, respectively. Geminal
couplings and other obvious correlations were discarded. The re-
sulting structures were minimized with 3000 cycles of steepest de-
Supporting Information (see footnote on the first page of this arti-
cle): Syntheses of the cyclopeptides, NMR characterization, and
conformational analyses.
Acknowledgments
We gratefully thank MIUR (Ministero dell’Università e della
Ricerca), Bologna University, and Fondazione CARISBO (Cassa
di Risparmio in Bologna), Bologna, for providing financial sup-
port. We also thank A. Garelli, Dr. E. Florio, and Dr. R. Artali.
734
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 729–735

Cyclotetrapeptide Mimics with a Partially Modified Retro-Inverso Structure
[9] T. Morie, S. Kato, H. Harada, I. Fujiwara, K. Watanabe, J.-I.
Matsumoto, J. Chem. Soc. Perkin Trans. 1 1994, 2565–2569.
[10] a) S. J. Stradley, J. Rizo, M. D. Bruch, A. N. Stroup, L. M. Gi-
erasch, Biopolymers 1990, 29, 263–287; b) H. Kessler, Angew.
Chem. Int. Ed. Engl. 1982, 21, 512–523.
[11] C. Toniolo, CRC Crit. Rev. Biochem. 1980, 9, 1–44.
[12] For a leading reference on the use of DMSO as a biomimetic
medium, see: P. A. Temussi, D. Picone, G. Saviano, P. Amodeo,
A. Motta, T. Tancredi, S. Salvadori, R. Tomatis, Biopolymers
1992, 32, 367–372. For other selected examples of recently pub-
lished papers, see: a) J. Chatterjee, D. Mierke, H. Kessler, J.
Am. Chem. Soc. 2006, 128, 15164–15172; b) E. Locardi, D. G.
Mullen, R.-H. Mattern, M. Goodman, J. Pept. Sci. 1999, 5,
491–506; c) H. B. Lee, M. Pattarawarapan, S. Roy, K. Burgess,
Chem. Commun. 2003, 14, 1674–1675.
[13] a) C. Henry, N. Moitessier, Y. Chapleur, Mini-Rev. Med. Chem.
2002, 2, 531–542; b) M. A. Dechantsreiter, E. Planker, B. Ma-
tha, E. Lohof, G. Holzemann, A. Jonczyk, S. L. Goodman, H.
Kessler, J. Med. Chem. 1999, 42, 3033–3040; c) R. Haubner, R.
Gratias, B. Diefenbach, S. L. Goodman, A. Jonczyk, H.
Kessler, J. Am. Chem. Soc. 1996, 118, 7461–7472; d) K. Bur-
gess, D. Lim, S. A. Mousa, J. Med. Chem. 1996, 39, 4520–4526;
e) G. Casiraghi, G. Rassu, L. Auzzas, P. Burreddu, E. Gaetani,
L. Battistini, F. Zanardi, C. Curti, G. Nicastro, L. Belvisi, I.
Motto, M. Castorina, G. Giannini, C. Pisano, J. Med. Chem.
2005, 48, 7675–7687.
[1]
a) J. D. A. Tyndall, B. Pfeiffer, G. Abbenante, D. P. Fairlie,
Chem. Rev. 2005, 105, 793–826; b) V. J. Hruby, P. M. Balse,
Curr. Med. Chem. 2000, 7, 945–970.
a) V. J. Hruby, R. S. Agnes, Biopolymers 2000, 51, 391–410; b)
for a recent review on the use of peptidomimetics in biochemis-
try, medicine, pharmacology, etc., see: L. Gentilucci, A. Tolo-
melli, F. Squassabia, Curr. Med. Chem. 2006, 13, 2449–2466.
A. Carotenuto, A. M. D’Ursi, B. Mulinacci, I. Paolini, F. Lolli,
A. M. Papini, E. Novellino, P. Rovero, J. Med. Chem. 2006, 49,
5072–5079.
a) N. Loiseau, J.-M. Gomis, J. Santolini, M. Delaforge, F. An-
dre, Biopolymers 2003, 69, 363–385; b) F. Schumann, A.
Muller, M. Koksch, G. Muller, N. Sewald, J. Am. Chem. Soc.
2000, 122, 12009–12010; c) M. P. Glenn, M. J. Kelso, J. D. A.
Tyndall, D. P. Fairlie, J. Am. Chem. Soc. 2003, 125, 640–641,
and references herein; d) P. Mora, C. Mas-Moruno, S. Tambor-
ero, L. J. Cruz, E. Perez-Paya, F. Albericio, J. Pept. Sci. 2006,
12, 491–496.
[2]
[3]
[4]
[5] a) A. S. Norgren, F. Buttner, S. Prabpai, P. Kongsaeree, P. I.
Arvidsson, J. Org. Chem. 2006, 71, 6814–6821; b) N. Maulucci,
M. G. Chini, S. Di Micco, I. Izzo, E. Cafaro, A. Russo, P. Galli-
nari, C. Paolini, M. C. Nardi, A. Casapullo, R. Riccio, G. Bi-
fulco, F. De Riccardis, J. Am. Chem. Soc. 2007, 129, 3007–
3012.
[6] a) G. Cardillo, L. Gentilucci, A. Tolomelli, R. Spinosa, M.
Calienni, A. R. Qasem, S. Spampinato, J. Med. Chem. 2004,
47, 5198–5203; b) F. Benfatti, G. Cardillo, S. Fabbroni, L. Gen-
tilucci, R. Perciaccante, A. Tolomelli, M. Baiula, S. Spampin-
ato, Tetrahedron: Asymmetry 2006, 17, 167–170; c) L. Gentil-
ucci, F. Squassabia, R. Artali, Curr. Drug Targets 2007, 8, 185–
196.
[7] a) M. Chorev, M. Goodman, Acc. Chem. Res. 1993, 26, 266–
273; b) M. D. Fletcher, M. M. Campbell, Chem. Rev. 1998, 98,
763–795; c) M. Chorev, Biopolymers 2005, 80, 67–84; d) Y. S.
Lee, R. S. Agnes, P. Davis, S.-w. Ma, H. Badghisi, J. Lai, F.
Porreca, V. J. Hruby, J. Med. Chem. 2007, 50, 165–168; e) K.-
J. Kim, S.-W. Park, S. S. Yoon, J. Kor. Chem. Soc. 2000, 44,
286–289; f) for the use of a 10-membered ethylene-bridged
PMRI peptide as β-turn mimetic, see: Y. Han, C. Giragossian,
D. F. Mierke, M. Chorev, J. Org. Chem. 2002, 67, 5085–5097.
[8] S. Nakabayashi, C. D. Warren, R. W. Jeanloz, Carbohydr. Res.
1988, 174, 279–289.
[14] M. Gurrath, G. Muller, H. Kessler, M. Aumailley, R. Timpl,
Eur. J. Biochem. 1992, 210, 911–921.
[15] B. Cacciari, G. Spalluto, Curr. Med. Chem. 2005, 12, 51–70.
[16] a) S. Caltabiano, W. T. Hum, G. J. Attwell, D. N. Gralnick, L. J.
Budman, A. M. Cannistraci, F. J. Bex, Biochem. Pharm. 1999,
58, 1567–1578; b) H. Fujii, H. Komazawa, H. Mori, M. Ko-
jima, I. Itoh, J. Murata, I. Azuma, I. Saiki, Biol. Pharm. Bull.
1995, 18, 1681–1688.
[17] W. D. Cornell, P. Cieplak, C. I. Bayly, I. R. Gould, K. M.
Merz, D. M. Ferguson, D. C. Spellmeyer, T. Fox, J. W. Cald-
well, P. A. Kollman, J. Am. Chem. Soc. 1995, 117, 5179–5197.
[18] W. L. Jorgensen, J. Chandrasekhar, J. Madura, R. W. Impey,
M. L. Klein, J. Chem. Phys. 1983, 79, 926–933.
[19] H. J. C. Berendsen, J. P. M. Postma, W. F. van Gunsteren, A.
DiNola, J. R. Haak, J. Chem. Phys. 1984, 81, 3684–3690.
[20] U. Landegren, J. Immunol. Methods 1984, 67, 379–388.
Received: October 1, 2007
Published Online: November 28, 2007
Eur. J. Org. Chem. 2008, 729–735
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
735