Grafting aminocyclopentane carboxylic acids onto the RGD tripeptide sequence generates low nanomolar αvβ3/ αvβ5 integrin dual binders

Article abstract of DOI:10.1021/jm050698x

Eleven γ-aminocyclopentane carboxylic acid (Acpca) platforms, including four dihydroxy representatives (19-22), three hydroxy analogues (34-36), and four deoxy derivatives (30-33), were prepared in a chiral nonracemic format. These simple units were then

Full text of DOI:10.1021/jm050698x

J. Med. Chem. 2005, 48, 7675-7687
7675
Grafting Aminocyclopentane Carboxylic Acids onto the RGD Tripeptide
Sequence Generates Low Nanomolar rVâ3/rVâ5 Integrin Dual Binders
Giovanni Casiraghi,*,‡ Gloria Rassu,† Luciana Auzzas,† Paola Burreddu,† Enrico Gaetani,‡ Lucia Battistini,‡
Franca Zanardi,‡ Claudio Curti,‡ Giuseppe Nicastro,# Laura Belvisi,§ Ilaria Motto,§ Massimo Castorina,|
Giuseppe Giannini,| and Claudio Pisano|
Dipartimento Farmaceutico, Universita` di Parma, Parco Area delle Scienze 27A, I-43100 Parma, Italy, Istituto di Chimica
Biomolecolare del CNR, Traversa La Crucca 3, I-07040 Li Punti, Sassari, Italy, Centro Interfacolta` Misure “G. Casnati”,
Universita` di Parma, Parco Area delle Scienze 23A, I-43100 Parma, Italy, Dipartimento di Chimica Organica e Industriale,
Universita` di Milano, Via Venezian 21, I-20133 Milano, Italy, and Sigma-Tau, Research & Development, Via Pontina Km
30.400, 00040 Pomezia, Roma, Italy
Received July 21, 2005
Eleven γ-aminocyclopentane carboxylic acid (Acpca) platforms, including four dihydroxy
representatives (19-22), three hydroxy analogues (34-36), and four deoxy derivatives (30-
33), were prepared in a chiral nonracemic format. These simple units were then grafted onto
an Arg-Gly-Asp (RGD) tripeptide framework by a mixed solid phase/solution protocol delivering
an ensemble of 11 macrocyclic analogues of type cyclo-[-Arg-Gly-Asp-Acpca-], 1-11. The
individual compounds were evaluated for their binding affinity toward the RVâ3 and RVâ5
integrin receptors. The analogue 10 exhibited a very interesting activity profile (IC50/RVâ3 )
1.5 nM; IC50/RVâ5 ) 0.59 nM), comparable to that of reference compounds EMD121974 and
ST1646. Closely related congeners 6, 8, and 9 also proved to be excellent dual binders with
activity levels in the low nanomolar range. The three-dimensional (3D) NMR solution structures
were determined, and docking studies to X-ray crystal structure of the extracellular segment
of integrin RVâ3 in complex with the reference compound EMD121974 were performed on
selected analogues to elucidate the interplay between structure and function in these systems
and to evidence the subtle bases for receptorial recognition. The results prove that the principle
of isosteric dipeptide replacement for peptidomimetics design and synthesis can be violated,
without detriment to the development of highly effective integrin binders.
Introduction
proteins through the Arg-Gly-Asp (RGD) tripeptide
sequence.5,13 Today, several major classes and com-
pounds are presented as effective binders of RVâ3 and
RVâ5, where proper inducers of secondary peptide struc-
ture are grafted onto the RGD sequence to prevent free
rotation of the amide linkages and project peripheral
substituents in proper spatial orientations.2b,3b,7e,14
The aim of our project was to construct novel macro-
cyclic, RGD-carrying binders of RVâ3 and RVâ5 by utiliz-
ing hydroxylated and nonhydroxylated γ-aminocyclo-
pentane carboxylic acid (Acpca) platforms that can be
seen as γ-aminobutanoic acid (GABA) surrogates locked
in a W-conformation by an R,γ-ethylene (or hydroxy-
ethylene) bridge (Figure 1, structure of type A).
Although grafting of these small scaffolds onto the
RGD consensus motif results in one-atom restriction of
the macrocycle size in the formed compounds as com-
pared to previously investigated analogues (14 vs 15
atoms) and violates the classical principle of isosteric
dipeptide replacement for peptidomimetics design and
synthesis,15,16 we envisioned that the Acpca motifs
would equally provide the requisite scaffolds upon which
to design and develop potent integrin ligands for their
RVâ3 and RVâ5 receptors. In other words, we surmised
that the global conformational constraint of such chi-
meric structures arising from the local restriction by the
cyclopentane platforms combined with the bias imposed
by macrocyclization would result in a favorable spatial
orientation of the integrin recognition tripeptide.
In the recent past, small molecular templates that
mimic or induce loop and turn secondary structural
features of peptides and proteins have gained ample
popularity due to the important role they play in
molecular recognition and biological events.1 In par-
ticular, great effort is focused on the design and
synthesis of a number of molecular platforms encom-
passing dipeptide core structures,2 constrained dipeptide
surrogates,3 and nonpeptide motifs,4 whose incorpora-
tion into known peptide sequences has resulted in the
creation of promising cyclic and acyclic entities targeting
diverse, therapeutically relevant receptors and enzymes.
Among the most sought after receptors, the integrin
family represents an enterprising biological target,5 for
their diverse components play a prime role in physio-
pathological processes that span from coagulation of
blood,6 tumor-induced angiogenesis,7 and control of the
immune system8 to the initiation and progression of
tumor metastasis,9 osteoporosis,10 restenosis,11 and
inflammatory diseases.12
In the present work, we focused on RVâ3 and RVâ5
integrin receptors, which recognize several matrix
* To whom correspondence should be addressed. Phone: +39-0521-
905080. Fax: +39-0521-905006. E-mail: giovanni.casiraghi@unipr.it.
‡ Dipartimento Farmaceutico.
† Istituto di Chimica Biomolecolare del CNR.
# Centro Interfacolta` Misure “G. Casnati”.
§ Dipartimento di Chimica Organica e Industriale.
| Sigma-Tau, Research & Development.
10.1021/jm050698x CCC: $30.25 © 2005 American Chemical Society
Published on Web 11/04/2005

7676 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24
Casiraghi et al.
After benzylation of the lactam nitrogen, exposure of
the resulting fully protected compound to solid periodic
acid in ethyl acetate at room temperature yielded
aldehyde 16 directly (89% yield, two steps) via simul-
taneous deacetonidation and oxidative diol cleavage.20
The creation of the cyclopentane moiety utilized a highly
productive, intramolecular silylative aldol maneuver
driven by the Lewis acid-Lewis base reagent system
TBSOTf-DIEA, arriving at an isolable 80:20 mixture
of two cycloadducts, 17 and 18, in 98% combined yield.
The finale of the synthesis involved four operations
(three steps), N-debenzylation, hydrolytic fragmentation
of the lactam bond with full deprotection, and selective
N-Fmoc functionalization, yielding individual amino
acids 19 and 20 in 89% and 67% yields, respectively.
With the optimized conditions in hand, the 10-step
synthesis was exactly duplicated to prepare the enan-
tiomerical couple 21 and 22 by simply changing the
chirality of the glyceraldehyde precursor, S-13 in lieu
of R-13.
The dideoxy derivatives 30-33, as well as monohy-
droxylated congeners 34-36, were derived from com-
mercially available sources 23-26 and 27-29, respec-
tively, via simple functional group manipulations, as
shown in Scheme 2.21
All of the components of this precursor collection were
prepared as g98% pure materials by suitable chromato-
graphic operations and then used to build up our 11-
member ensemble of macrocyclic pseudopeptides, 1-11.
Figure 1. Structural comparison between a generic Acpca-
bearing RGD ligand A (14-membered cycle) and selected
known congeners B-E (15-membered cycles), highlighting the
principle of isosteric dipeptide replacement. As compared to
that of the lead cyclopentapeptide B, structures of type C-E
satisfy this principle, and the structure of the ring-contracted
A does not.
Synthesis of Cyclopeptides 1-11. To assemble the
targeted 11-component collection of cyclopeptides, 1-11,
we first planned to realize linear tetrapeptide precursors
of type H-Asp(But)-Acpca-Arg(Pmc)-Gly-OH and then
complete the macrocycle via head-to-tail juncture. The
Acpca units, 19-22 and 30-36, were critically posi-
tioned along the peptide chain, flanked by the aspartate
and arginine residues, to create the local constraint that
preorganizes the chain terminals toward macrocycliza-
tion.22
Eleven tetrapeptides of general formula H-Asp(But)-
Acpca-Arg(Pmc)-Gly-OH were synthesized in parallel on
solid phase, utilizing a conventional polystyrene-based
2-chlorotrityl chloride resin (Scheme 3). Although this
type of linker is quite acid labile, its stability toward
bases makes it ideal for the various coupling base
promoters as well as for the resident Arg- and Asp-
protecting groups using the Fmoc strategy, the product
being readily cleaved from the resin under mild acidic
conditions (e.g., AcOH in trifluoroethanol).
The Acpca-containing cyclopeptides 1-11 in Figure
2 were synthesized and biochemically evaluated in vitro
for their ability to compete with [125I]-echistatin for
binding to RVâ3 and RVâ5 receptors.17 The influence of
the hydroxyl substituents on the cyclopentane units as
well as the stereodisposition of peripheral functionality
were explored. Finally, to understand the differences
in the binding ability for closely related compounds,
thorough investigations, including in-solution NMR
structural analysis and docking studies of selected
analogues to X-ray crystal structure of the extracellular
segment of integrin RVâ3 in complex with the reference
compound EMD121974,18 were performed.
Results and Discussion
Synthesis of Cyclopentane Amino Acid Building
Blocks 19-22 and 30-36. The N-Fmoc-protected,
dihydroxylated cyclopentane amino acids 19-22 utilized
to assemble the RGD constructs 1-4 were synthesized
as enantiomerically pure materials by a modification of
our previously reported procedure (Scheme 1).19 In the
first step, silyloxypyrrole 12 was reacted with glycer-
aldehyde R-13 in diethyl ether under the guidance of
SnCl4. The reaction conditions were optimized (slow
addition of a 1.0 M solution of SnCl4 in CH2Cl2 at -90
°C is critical) to obtain an 82% yield and g98% de of
unsaturated adduct 14, which was isolated by simple
crystallization of the crude reaction product. Next, the
double bond within 14 was saturated and the free
hydroxyl group protected as TBS ether. Exposure of this
intermediate to ammonium cerium(IV) nitrate in ac-
etonitrile ensured clean removal of the N-Boc protection,
resulting in the formation of lactam 15 in 87% isolated
yield in three steps.
In the resin-loading step, N-Fmoc-glycine was at-
tached to the 2-chlorotrityl linker using a 2.5-fold excess
amino acid to ensure maximum yield (g1.55 mmol/g
actual loading). After deprotection of the amino group
(20% piperidine in DMF), condensation of the Fmoc-Arg-
(Pmc)-OH residue was attained using the TBTU-HOBt
system in the presence of DIEA. Next, the second amino
group within the resin-bound dipeptide Fmoc-Arg(Pmc)-
Gly-OcTrt was liberated as usual, and an Acpca plat-
form, chosen among the 11 amino acids of this study
(19-22 and 30-36, Schemes 1 and 2), was coupled and
the resin-bound tripeptide deprotected.
The Fmoc-Asp(But)-OH residue was then connected
and the amino group liberated to afford a resin-bound
tetrapeptide, which was cleaved from the resin with a

Cyclopentane-Based RGD Integrin Binders
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7677
Figure 2. Structures of the targeted RGD-containing cyclopeptides 1-11.
Scheme 1. Synthesis of Dihydroxylated Cyclopentane Amino Acid Building Blocks 19-22a
a Reagents: (a) SnCl4, Et2O, -90 °C (82%); (b) (i) H2, Pd/C, EtOAc; (ii) TBSOTf, 2,6-lutidine, CH2Cl2; (iii) CAN, CH3CN (87%, three
steps); (c) (i) BnBr, NaH, THF; (ii) H5IO6, EtOAc (89%, two steps); (d) TBSOTf, DIEA, CH2Cl2 (98%); (e) (i) Na, liq NH3, THF; (ii) 6 N aq
HCl, 80 °C; (iii) FmocOSu, Et3N, MeCN (89% for 19, 67% for 20, three steps).
1:1:3 v/v mixture of AcOH, TFE, CH2Cl2. The crude
linear peptides were thus obtained in yields ranging
from 27% to 43% for the entire solid-phase sequence.
The 11 individual linear peptides were cyclized in
solution (2.5 mM in DMF) at room temperature using
diphenylphosphoryl azidate (DPPA) activation in the
presence of solid sodium hydrogen carbonate. After
chromatographic purification, the side-chain deprotec-
tion was carried out under acidic conditions in the
presence of scavengers, with the reagent system TFA,
phenol, H2O, thioanisole, and 1,2-ethanedithiol (EDT).
The resulting compounds were first purified by prepara-
tive reversed-phase HPLC and finally transformed into
stabile hydrochloride salts by exposing the solid materi-
als thus obtained to anhydrous HCl (gas) until a
constant weight was reached. The global yields of the
cyclization/deprotection step ranged from good (68-
98%) for compounds 1-6 and 8-9, carrying a cyclopen-
tane motif with a CO/NH cis disposition, to moderate
(34-70%) for compounds 7, 10, and 11 that bear a trans-
configured amino/carbonyl cyclopentane substitution.
The purity of each macrocycle in the 11-component
collection 1-11 was checked by two independent RP-
HPLC analyses and judged to be g98% for all materials.
The synthesis scale was in the range of 22 mg for 1 to
68 mg for 3. All RGD macrocycles were characterized
by high-resolution ESI mass spectrometry as well as
various NMR techniques (see infra).
Biological Results. The ability of cyclopeptides
1-11 to compete with [125I]-echistatin for binding to the
isolated, purified RVâ3 and RVâ5 receptors was evaluated
in a solid-phase receptor assay17 and compared with

7678 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24
Casiraghi et al.
Scheme 2. Synthesis of Monohydroxylated and
Dideoxycyclopentane Amino Acid Building Blocks
30-36a
C3 stereocenters was evaluated. The general trends in
binding affinities indicate that within the subset of
dideoxy Acpca-containing ligands, the systems 9 and 10
bearing an R-located amino functionality (3R absolute
configuration) show a higher affinity than 8 and 11 that
bear the amino group â-disposed (3S configuration). The
dihydroxylated congeners 1-4 were also compared, and
additional support for the beneficial effect of the R
stereochemistry at the C3 aminated position of the
Acpca ring was found. Thus, both 3 and 4 (3S-
configured) have significantly higher affinity than 1 and
2, with an Acpca motif carrying â-located amino func-
tions (3R-configured). One notable exception is mono-
hydroxylated compound 7, which shows significantly
lower RVâ3/RVâ5 binding over the corresponding conge-
ners 5 and 6.
Among the best candidates, 6, 9, and 10, the most
distinctive compound is 10, carrying a robust, all-carbon
1,3-trans 3R-configured (1R,3R) aminocyclopentane car-
boxylic moiety. In binding to RVâ3, it is 3-fold superior
to 9 and 4-fold superior to 6. For RVâ5, it is 6-fold
superior to 9 and 8-fold superior to 6. When compared
to reference compounds EMD121974 and ST1646, our
simplified lead compound 10 proved to be of comparable
or even superior binding efficiency.
Overall, this study highlighted that in this series of
ligands containing simple Acpca moieties, the vacancy
at the C5 carbon along with the 3R-positioning of the
amino group proved beneficial for activity, whereas the
presence of a hydroxyl group at the same C5 position
invariably resulted in a marked loss of binding capabil-
ity of the ligands. On the other hand, the relative
stereodisposition of the amino and carboxyl substituents
of the Acpca motif (1,3-cis vs 1,3-trans) manifests a
minor impact on the activity of the corresponding RGD
ligand. Possible reasons for these findings will be
discussed in the next paragraphs dealing with the
solution structures and computational analysis of the
selected candidates.
NMR Spectroscopy and Solution Structural
Analysis. The complete chemical shift assignment of
proton and carbon resonances (Supporting Information,
Tables S1-S3) for 1-11 has been carried out for water
solutions (90% H2O/10% D2O) using one- and two-
dimensional (1D and 2D) NMR experiments, following
the standard procedure described in the Experimental
Section.
Six analogues from this 11-member family of RVâ3/
RVâ5 dual binders, compounds 1, 3, 6, 7, 9, and 10, were
selected and their NMR three-dimensional (3D) solution
structures determined. These analogues differ as fol-
lows: compounds 1 and 3 possess a dihydroxylated
Acpca “residue”, whereas the absolute configuration of
the stereocenters within the ring is opposite. Further-
more, the amino and carboxyl substituents lie in a cis
disposition, with trans-located hydroxyl groups. Com-
pounds 6 and 7 lack the C5 hydroxyl but have a 1,3-cis
and a 1,3-trans orientation, respectively. The two
members of the dideoxycyclopentane family, 9 and 10,
differ from each other at position C1 with 1S and 1R
absolute configuration, respectively. Although all these
analogues showed a dual binding ability to RVâ3 and
RVâ5 receptors, their affinities vary from 132.9 and 411.5
nM for 1 to 1.5 and 0.59 nM for 10.
a Reagents: (a) FmocOSu, Et3N, MeCN (75% for 30, 38% for
31); (b) (i) 6 N aq HCl, 80 °C; (ii) FmocOSu, Et3N, MeCN (50% for
32, 68% for 33, 86% for 34, 88% for 35, 83% for 36, two steps).
that of known standard binders, EMD121974,14a
c(-RGDfV-),2b and ST1646.3b Integrin-coated 96-well
plates were incubated with the ligand in the presence
of serially diluted competing compounds. The IC50 ( SD
values (nM) were calculated as the concentration of
compound required for 50% inhibition of ligand binding,
as estimated by the Allfit program, and are given in
Table 1.
For this strictly related compound series, the influ-
ence of the nature of the cyclopentane scaffold grafted
into the RGD residue was especially analyzed. The first
trend to note in the series is that the binding affinities
of the RGD ligand family embodying 4,5-dihydroxylated
Acpca units are invariably reduced as compared to those
of the 4-hydroxy and 4,5-dideoxy analogues (e.g., 1-4
vs 5-7, vs 8-11). This indicates that the scaffold
disubstitution is not as well tolerated for both the RVâ3
and RVâ5 receptor binding. In particular, comparison of
the dihydroxylated derivatives 1 and 2 to 1,3-stere-
ochemically related monohydroxylated compound 5 and
dideoxy derivative 8 shows that there is a 18-31-fold
increase in activity when going from two hydroxyl
groups at C4 and C5 (e.g., 1 and 2) to one hydroxyl at
C4 (e.g., 5), whereas there is only a negligible variation
when going from one hydroxyl at C4 (e.g., 5) to no
hydroxyls (e.g., 8). A similar trend is also observed for
the dihydroxylated couple 3/4 as compared to monohy-
droxylated derivative 6 and dideoxy analogue 9. This
seems to indicate that though placement of one OH-
group at the C4 position of the Acpca ring only displays
a marginal effect, the positioning of an extra hydroxyl
substituent at C5 results in a significant decrease in
binding affinity.
In addition to examination of hydroxyl substituents
at the C4 and C5 positions of the cyclopentane motifs,
the effect of the absolute configuration at the C1 and

Cyclopentane-Based RGD Integrin Binders
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7679
Scheme 3. Synthesis of Cyclopeptides 1-11a
a Reagents: (a) DIEA, CH2Cl2, MeOH; (b) piperidine/DMF 20%; (c) Fmoc-Arg(Pmc)-OH, TBTU/HOBt, DIEA, NMP; (d) Fmoc-Acpca-
OH, TBTU/HOBt, DIEA, NMP; (e) Fmoc-Asp(But)-OH, TBTU/HOBt, DIEA, NMP; (f) AcOH/TFE/CH2Cl2 (1:1:3) (27-43%, nine steps); (g)
DPPA, NaHCO3, DMF; (h) TFA, PhOH/H2O/thioanisole/EDT (85.5:5:2:5:2.5) (34-98%, two steps).
Table 1. Binding Affinity of Acpca-Containing Ligands 1-11 in Radioligand Assays at RVâ3 and RVâ5 Integrin Receptors
a For comparison purposes, a uniformed atom numbering for the cyclopentane residue is adopted. b For the reference ligands, the structure
of the corresponding residue is reported. c Average of three or more independent determinations. d Compounds 1-11, tested as hydrochloride
salts. e This assay. The literature values are 3.7 ( 0.6 and 3.0 ( 0.5, respectively (ref 3b).
These analogues were selected to cover the chemical
diversity of the Acpca “residues” embodied in the RGD
sequence and judged sufficient for clarifying the impact
that such simple cyclopentane motifs exert on the entire
conformational behavior of these macrocycles as well as
on their binding affinities. For the six selected com-
pounds, the nuclear Overhauser enhancement spectros-
copy (NOESY) spectra, and the JNH-RH coupling con-
stants, led to approximately 50 restraints per analogue
used as input for the structure calculations (for more
3

7680 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24
Casiraghi et al.
Figure 3. NMR structures of analogues 1, 3, 6, 7, 9, and 10. For each analogue, 20 energy-minimized conformers with the
lowest energy are used to represent the 3D NMR structure. The bundle is obtained by superposing the backbone atoms of the
RGD residues. The following color code is used: green - carbon; white - hydrogen; blue - nitrogen; red - oxygen. H-bonds are
indicated by broken lines.
details, see the Experimental Section). Of the 100
computed structures for each compound, 20 models per
analogue were selected, on the basis of the lowest-energy
criterion, to represent the solution’s 3D structure (Fig-
ure 3). The small residual distance violations (less than
0.3 Å) and the low dihedral angles violations (less than
3°) indicate that the input data represent a self-
consistent set and the restraints are well-satisfied in
the calculated conformers for each analogue.
Compound 1 binds to both receptors with low affinity
(IC50 ) 132.9 and 411.5 nM). The quality of the
structure, shown in Figure 3, is reflected by the low
average backbone rmsd value (0.07 Å). On the basis of
the backbone torsion angle analysis (see Table S4 in
Supporting Information), a family of structures (30%
occurrence), characterized by an inverse γ-turn with the
Asp residue in the i+1 position, has been identified.23
The γ-turn motif is stabilized by a hydrogen bond (H
bond) between HN-Acpca and CO-Gly, as supported by
the temperature coefficient ∆δ/∆T ) -3.5 ppb/K ex-
perimentally observed for the amide proton of Acpca.
Noteworthy, a H-bonding network responsible for the
reduced flexibility of the side chains is present, that
involves the Acpca C5 hydroxyl with the backbone and
guanidinium moiety of Arg (50% occurrence) as well as
the Acpca C4 hydroxyl with the Asp carboxylate group
(23% occurrence). In addition, a significant Arg intra-
residual H-bonding further reduces the flexibility of this
part of the cyclopeptide. The Acpca template keeps the
distance between the Câ atoms of Arg and Asp at 8.0 Å.
Analogue 3 has a 3-fold higher binding affinity than
1 and differs from it in the absolute configuration of all
four stereocenters of the Acpca scaffold. The backbone
torsion angles indicate that the structure contains an
inverse γ-turn centered on Gly(i)-Asp(i+1)-Acpca(i+2)
(>60% occurrence), which is highly stabilized by the
presence of a H bond between the amide proton of Acpca
and the carbonyl of Gly. The very low-temperature
coefficient ∆δ/∆T ) -1.0 ppb/K observed for the amide
proton of Acpca, indicative of the expected H-bond
AcpcaNH-COGly, along with the scarce scattering of
the torsion angle values of the minimized conformers
indicates a reduced conformational freedom of the
macrocycle backbone as compared to that for 1. Of the
two Acpca hydroxyls at C4 and C5, only the latter is
partially involved in H-bonding with the Arg side chain.

Cyclopentane-Based RGD Integrin Binders
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7681
This feature results in Asp and Arg side chains expe-
riencing an enhanced conformational freedom as com-
pared to that of 1, even if the distance between the Câ
atoms of Arg and Asp is still kept at 8.0 Å by the
presence of the template.
structured cyclopeptide backbone, be it an inverse
γ-turn motif or an organized kink centered at the Asp
residue, associated with a great conformational freedom
of the Asp and Arg pharmacophoric side chains is a
stringent requisite for activity.
In summary, from the biological and structural results
of this study, it can be assumed that an organized
arrangement of the cyclopeptide backbone, that ensures
a conformation of the macrocycle in the Arg-Gly-Asp
region with Câ distances Asp-Arg in the range 8.0-8.4
Å, is essential for an efficient RVâ3 and RVâ5 integrin
binding. Furthermore, a crucial contribution is given by
the conformational freedom of the Asp and Arg side
chains. When their spatial arrangement is constrained
by inter- and intraresidual H-bonding as, for example,
in hydroxylated compounds 1, 3 and 7, the binding
capability is markedly eroded.
In all of the active compounds of the series, the amide
proton of the Asp moiety points outward from the ring,
exposed to the solvent. A remarkable exception is the
worst representative of the series, dihydroxylated com-
pound 2 (IC50 ) 712.7 and 1300.0 nM), whose 3D NMR
solution structure (data not shown) highlights a marked
preference for an inverse γ-turn centered at the Gly
residue. This causes the macrocycle backbone to assume
a “closed” conformation with shorter CR and Câ Asp-Arg
distances (5.8 and 7.9 Å respectively). Different from
the previous cases, the Asp amide proton points inside
the macrocycle ring, presumably preventing the forma-
tion of crucial interactions with the binding pocket of
the receptors.
Monohydroxylated representatives 6 and 7 bind with
high affinity to both integrin receptors, with the former
being 6× more active than the latter. Their chemical
structures differ from each other in the orientation of
the carboxylic residue, cis to the amino group in 6 and
trans in 7. The calculated NMR structure of 6 possesses,
again, an inverse γ-turn centered at Asp (90% popu-
lated) that is stabilized by the H bond between AcpcaNH
and COGly. This is justified by the low-temperature
gradient (-2.0 ppb/K) of the Acpca amide proton. As for
compound 3, no H-bonding is observed involving the C4
hydroxyl, and the Asp and Arg side chains are confor-
mationally flexible. The Câ distance between Arg and
Asp is 8.4 Å. In contrast, compound 7 does not present
any defined secondary structures consistently with all
3
three coupling constants JNH-RH of ca. 7.0 Hz that
suggest conformational averaging. Also, a highly popu-
lated family of conformers (51%) is present, dominated
by a strong H-bond between the C4 hydroxyl and the
Asp residue. Again, the distance Câ Asp-Arg is 8.2 Å.
The comparison of compounds 6 and 7 seems to indicate
that a somewhat rigid macrocycle backbone conforma-
tion associated with the absence of H-bonding involving
the C4 hydroxyl represents a key element for the
biological activity of 6.
Analogues 9 and 10 are the most active compounds
of this series, but the binding affinity of 10 is 3-fold
higher for RVâ3 and 6-fold higher for RVâ5 than 9. The
chemical structures differ from those of compounds 1,
3, 6, and 7 by the absence of hydroxyl substituents in
the Acpca motifs and differ each other in the cis (9) vs
trans (10) location of the Acpca scaffold substituents.
The NMR data and the results of SA calculations of 9
revealed a highly rigidified and well-defined structure
of the backbone, showing an inverse γ-turn with Asp in
the i+1 position (75%). A low-temperature coefficient
of -3.3 ppb/K for the amide hydrogen of Acpca is further
indicative of the H-bond formation between AcpcaNH
and COGly. Also, there is a consistent presence of
intraresidual H-bonding of the Asp moiety, which is
supported by the temperature gradient of -2.4 ppb/K
for the Asp amide proton. The distance Câ Arg-Asp
remains almost unchanged at 8.1 Å.
Docking Studies. To understand the experimental
results on a structural basis, molecular docking studies
were undertaken for six representative compounds,
namely dihydroxylated derivatives 1 and 3, monohy-
droxylated derivatives 6 and 7, and dideoxy derivatives
9 and 10. The protein binding site was derived from the
X-ray crystal structure of the extracellular segment of
integrin RVâ3 in complex with the cyclic pentapeptide
ligand EMD121974.18b For each analogue, representa-
tive cyclopeptide backbone conformations were selected
among the energy-minimized conformers derived from
the NMR data as starting structures for docking studies.
The crystal structure of the peptide/integrin complex
provides the actual conformation of EMD121974 bound
to the RVâ3 integrin active site and can serve as a basis
for understanding the general mode of interaction of
integrins with other RGD-containing ligands. Examina-
tion of the 3D structure of the cyclic pentapeptide ligand
EMD121974 bound to the RVâ3 integrin receptor (Pro-
tein Data Bank entry code ) 1L5G)18b reveals a con-
formation characterized by an inverse γ-turn with Asp
at position (i+1) and by a distorted âII′-turn with Gly
and Asp at the (i+1) and (i+2) positions, respectively.
An 8.9 Å distance between the Asp and Arg Câ atoms
and an almost extended conformation of the RGD
sequence are observed in this pentapeptide bound
conformation.
Strictly speaking, the NMR structure of the most
active compound 10 does not evidence any classical
backbone secondary structures, as indicated by the
distribution of the torsion angle values and by the
3
vicinal coupling constants JNH-RH (ca. 7.0 Hz). How-
ever, a kinked conformation is present, populated to a
degree of 90% in the 100 structures generated, centered
at the Asp moiety (see Figure 3), which is also substan-
tiated by a strong NOE contact between the Acpca NH
and the RH and âH2 Asp protons as well as by a low-
temperature gradient of -2.88 ppb/K for the Acpca-
amide proton. In addition, both of the side-chain resi-
dues freely rotate with no definite spatial region occupied.
The Câ distance of Arg and Asp is 8.3 Å. Thus, a
deduction can be drawn from the results involving
compounds 9 and 10; it seems that the existence of a
Remarkably, solution NMR-derived structures of
selected Acpca-containing ligands show preferred cy-
clopeptide backbone arrangements featuring a γ-turn
or a kink centered at the Asp residue and an extended
conformation of the RGD region with average Arg-
Asp Câ distances in the range of 8.0-8.4 Å. These geo-

7682 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24
Casiraghi et al.
Figure 4. Results of the docking studies. Top-ranking binding modes of RGD ligands 1, 3, 7, and 10 (carbon atoms are shown
in gray, nitrogen atoms in blue, oxygen atoms in red, and polar hydrogen atoms in white) into the crystal structure of the
extracellular domain of RVâ3 integrin overlayed on the bound conformation of EMD121974 (green). Selected integrin residues
involved in the interactions with the Acpca ligands (residues of the R subunit are shown in cyan, those of the â subunit in orange)
and crucial H bonds are shown. The Mn2+ ion at MIDAS (metal-ion-dependent adhesion site) is shown in magenta. Nonpolar
hydrogens were removed for clarity.
metries are, therefore, remarkably similar to the bound
EMD121974 conformation and also ensure an orienta-
tion of the Asp NH group suitable to form the H-bond
ligand-receptor interactions revealed in the crystal
structure.
central Gly residue is in close contact with the integrin
surface. Automated docking calculations of the most
active compound 10 as well as 6 and 9 (not shown in
Figure 4) starting from different representative cyclo-
peptide backbone conformations always produced top-
ranked poses conserving all these important interactions
and the ligand in the same location as that for
EMD121974 in the crystal structure. On the contrary,
only one representative macrocycle conformation of the
less active compound 7 resulted in best-score binding
modes reproducing the X-ray pose. Also, the hydroxyl
substituent at C4 of this monohydroxylated derivative
may form a H bond with the (R) Tyr 178 OH-group, and
alternative poses lacking the major interactions of the
crystallographic binding geometry may occur.
Moreover, as the cleft in which the ligands bind is
rather shallow, alternative binding modes differing in
the orientation of the Acpca motif were found by the
automated docking calculations. The guanidine and
carboxy groups of the ligands seem to be essential for
binding to the integrin subunits R and â, respectively,
acting like an electrostatic clamp in attaching to charged
regions of the protein.25,26
Starting from this X-ray complex, structural models
for the interactions of the selected compounds with the
ligand-binding site of the RVâ3 integrin receptor were
generated by automated computational docking using
the Glide program24 after removal of the peptide ligand.
The obtained complexes appeared to maintain almost
the same ligand-receptor distances and interactions
observed in the crystalline complex of EMD121974 with
RVâ3.18b The most important interactions involve the
positively charged Arg guanidinium group of the ligand
and the negatively charged side chains of Asp 150 and
Asp 218 in the R subunit as well as one of the Asp
carboxylate oxygens of the ligand and the metal cation
in the metal-ion-dependent adhesion site (MIDAS)
region of the â subunit (Figure 4). Further stabilization
could occur through H bonds between the backbone NH
of the Asp residue and the backbone carbonyl oxygen
of Arg 216 in the â subunit as well as between the other
Asp carboxylate oxygen and the backbone amide hy-
drogen of Asn 215 in the â subunit. Moreover, the
In addition to the calculated poses resembling the
experimentally observed binding mode of EMD121974,

Cyclopentane-Based RGD Integrin Binders
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7683
automated docking of the least active dihydroxylated
derivatives 1 and 3 revealed high-score alternative
binding modes differing in the H-bond pattern and in
the position of the ligand within the binding pocket. In
the alternative poses, the hydroxyl substituent at C5
of compound 1 forms a H bond with the (R) Tyr 178 OH-
group, and both the Acpca hydroxyls at C4 and C5 of
derivative 3 form H-bond interactions with the guani-
dinium group of the Arg 214 residue in the â subunit.
As a consequence of the H bonds engaged by the
functional groups of the Acpca ring, a rotation of the
ligand in the binding site is observed, which prevents
the formation of several important interactions disclosed
by the crystal complex. Assuming that the X-ray pose
describes the best interaction mode with the RVâ3
integrin receptor, this finding could explain the decrease
in binding affinity observed when going from dideoxy
analogues to RGD ligands containing 4,5-dihydroxylated
Acpca units.27
In summary, the models built for the interaction of
representative Acpca derivatives with RVâ3 integrin
through docking studies confirmed that the macrocycle
conformations of these ligands enable them to fit proper-
ly in the shallow cleft of the receptor, sharing the bind-
ing features of the crystal structure of the EMD121974-
RVâ3 complex. The polar interactions and the H-bonding
network governing the recognition process may be
partially destroyed and modified by the Acpca scaffold
substitution, frequently resulting in alternative binding
modes.
and H-bonding networks involving these substituents
and alternative binding sites of the receptor may occur,
that prevent these ligands from fitting properly into the
shallow cleft of the binding pocket.
All in all, these results establish γ-aminocyclopentane
carboxylic acids, conformationally restricted GABA sur-
rogates, as privileged scaffolds capable of generating
functional molecular binders of useful peptidal second-
ary structures. Efforts to develop more potent integrin
blocking agents and conjugates29 embodying GABA-
reminiscent pentacyclic scaffolds equipped with tunable
anchoring functionalities continue in our laboratories.
Experimental Section
For routine experimental procedures and spectroscopic data,
see the Supporting Information.
c[-Arg-Gly-Asp-Acpca19-] (1) (General Procedure for
Cyclopeptides Synthesis). Resin Loading. All number of
equivalents of reagents are given relative to the resin loading
(mmol/g). In a solid-phase reaction vessel, to the cTrt resin
(101 mg, 0.181 mmol) preswollen in CH2Cl2 (30 min), was
added a solution of Fmoc-Gly-OH (135 mg, 0.453 mmol) and
DIEA (72 µL, 0.416 mmol) in CH2Cl2 (2.5 mL). The reaction
mixture was stirred under a flow of nitrogen for 1 h. After
adding another 72 µL of DIEA (0.416 mmol) and 430 µL of
MeOH, the mixture was shaken for a further 30 min and
drained, and the resin was washed several times with DMF
(3 × 3 mL), CH2Cl2 (5 × 3 mL), PriOH (2 × 3 mL), MeOH (5
× 3 mL), Et2O (2 × 3 mL), and CH2Cl2 (3 × 3 mL). The Fmoc-
Gly resin, swollen in DMF (5 × 3 mL), was treated with 5%
v/v piperidine in DMF/CH2Cl2 1:1 (5 mL, 5 min). The solution
was drained and the resin treated with 20% v/v piperidine in
DMF (5 mL × 5 min × 6 cycles). The resin was washed with
DMF (3 × 3 mL), PriOH (3 × 3 mL), Et2O (3 × 3 mL), CH2Cl2
(3 × 3 mL), and DMF (2 × 3 mL), and the presence of the free
amino groups was checked with the TNBS test.
Conclusions
One of the most striking features in the design of
peptidomimetics is the embodiment of secondary struc-
ture-inducing scaffolds into the amino acid sequences
with the perspective to force peptides into bioactive
conformations and enhance stability toward degradation
by enzymes.1 In keeping with this reasoning, we syn-
thesized 11 RGD cyclopeptides, 1-11, incorporating
simple, readily available aminocyclopentane carboxylic
acid motifs (Acpca) and assessed their binding capabili-
ties toward RVâ3 and RVâ5 integrin receptors. Of the
compound repertoire prepared covering a wide range of
binding affinities, four representatives, 6, 8, 9, and 10,
displayed one-digit nanomolar dual binding levels, with
analogue 10 being the best candidate (IC50 ) 1.5 nM vs
RVâ3 and 0.59 nM vs RVâ5). These values are comparable
or even superior to those of several RGD-based high-
affinity binders reported in the literature, including the
reference compounds EMD121974, c[-RGDfV-], and
ST1646.
The data from this study, supported by in-solution 3D
NMR structural results and docking evidences, high-
light several points: first, the nonisosteric (yet bioiso-
steric) substitution15 of a cyclopentane γ-amino acid
motif spanning five atoms for a six-atom dipeptidic/
pseudodipeptidic spacer proved beneficial for the devel-
opment of high affinity integrin binders;28 second, a
close similarity of the ligand backbone conformation was
found for all the active compounds in this series, which
ideally matches the receptor-bound conformation; third,
the presence of hydroxyl substituents at the Acpca
nucleus heavily depletes the binding affinity of the
ligands for parasitic, noncooperative polar interactions
Peptide Coupling. A preformed solution of Fmoc-Arg-
(Pmc)-OH (300 mg, 0.453 mmol), TBTU (116 mg, 0.362 mmol),
HOBt (61 mg, 0.453 mmol), and DIEA (158 µL, 0.905 mmol)
in NMP (2 mL) was added to the deprotected peptidyl resin.
The mixture was shaken at room temperature for 1.5 h.
Reaction completion was checked with the TNBS test. The
solution was drained and the resin washed several times with
DMF (3 × 3 mL), PriOH (3 × 3 mL), Et2O (2 × 3 mL), and
CH2Cl2 (3 × 3 mL). The resin was washed again with DMF (5
× 3 mL) and then treated with 20% v/v piperidine in DMF (5
mL × 5 min × 6 cycles). The solution was drained and the
resin washed with DMF (5 × 3 mL), PriOH (3 × 3 mL), Et2O
(2 × 3 mL), CH2Cl2 (2 × 3 mL), and DMF (2 × 3 mL), and the
presence of the free amino groups was checked with the TNBS
test.
The coupling of the Fmoc-Acpca19-OH (104 mg, 0.271
mmol) and Fmoc-Asp(But)-OH (186 mg, 0.453 mmol) residues
was carried out under the same conditions.
Acetic Acid Cleavage. The resin-bound peptide, H-Asp-
(But)-Acpca19-Arg(Pmc)-Gly-O-cTrt, was treated with 5
mL of a mixture of acetic acid, 2,2,2-trifluoroethanol (TFE),
and CH2Cl2 (1:1:3) for 10 min at ambient temperature. The
solution was recovered and the resin carefully washed with
the above AcOH/TFE/CH2Cl2 mixture (2 × 5 mL × 10 min).
The combined solution was coevaporated in vacuo with hexane
several times, furnishing 42 mg of the linear tetrapeptide
H-Asp(But)-Acpca19-Arg(Pmc)-Gly-OH as a white solid,
used as such in the subsequent synthesis step.
Cyclization. The linear tetrapeptide, H-Asp(But)-Acpca19-
Arg(Pmc)-Gly-OH (42 mg, 0.048 mmol) was dissolved in
DMF (19 mL, 2.5 mM) under nitrogen, and the solution was
cooled at 0 °C. Solid NaHCO3 (20 mg, 0.24 mmol) and
diphenylphosphoryl azide (DPPA, 31 µL, 0.144 mmol) were
added, and the reaction mixture was stirred at ambient
temperature for 36 h. After filtration of excess solid NaHCO3,
the mixture was diluted with water and evaporated in vacuo.

7684 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24
Casiraghi et al.
The solid residue was purified by flash chromatography
(EtOAc/MeOH 8:2), furnishing the protected cyclic tetrapeptide
(90%) as a glassy white solid.
c[-Arg-Gly-Asp-Acpca32-] (10). Resin loading: 1.6 mmol/
g. Overall yield: 26% (27 mg). A glassy white solid; [R]20D -17.5
(c 0.3, H2O) (HCl salt); HPLC purity: g98%; HPLC tR ) 14.5
min; 1H and 13C NMR: see Table S3. HRMS (ES+) C18H30N7O6
calcd for [MH]+ 440.2258, found 440.2251.
c[-Arg-Gly-Asp-Acpca33-] (11). Resin loading: 1.55 mmol/
g. Overall yield: 24% (24 mg). A glassy white solid; [R]20D -20.3
(c 0.04, H2O) (HCl salt); HPLC purity: g98%; HPLC tR ) 11.0
min; 1H and 13C NMR: see Table S3. HRMS (ES+) C18H30N7O6
calcd for [MH]+ 440.2258, found 440.2263.
Solid-Phase Receptor Binding Assay. The receptor
binding assays were performed as described previously.17
Purified RVâ3 and RVâ5 receptors (Chemicon International Inc.,
Temecula, CA) were diluted to 500 and 1000 ng/mL, respec-
tively, in coating buffer [20 mM Tris-HCl (pH 7.4), 150 mM
NaCl, 2 mM CaCl2, 1 mM MgCl2, and 1 mM MnCl2]. An aliquot
of diluted receptors (100 µL/well) was added to 96-well
microtiter plates and incubated overnight at 4 °C. The coating
solution was removed, and 200 µL of blocking solution (coating
buffer plus 1% BSA) was added to the wells and incubated for
an additional 2 h at room temperature. After incubation, the
plates were rinsed three times with 200 µL of blocking-
binding solution and incubated for 3 h at room temperature
with 0.05 and 0.1 nM [125I]-echistatin (Amersham Pharmacia)
for RVâ3 and RVâ5 receptor binding assays, respectively. After
incubation, the plates were sealed and counted in the γ-counter
(Packard).
Side-Chain Deprotection. The protected cyclic tetrapep-
tide (34 mg, 0.043 mmol) was treated with 5 mL of a solution
of TFA (85.5%), phenol (5%), water (2%), thioanisole (5%), and
1,2-ethanedithiol (2.5%) at ambient temperature. After 24 h,
the solvent was evaporated in vacuo and the residue dissolved
in 5 mL of 50% aq AcOH. The mixture was diluted with 5 mL
of Et2O and extracted with water (3 × 5 mL). The combined
aqueous layers were washed with Et2O (3 × 3 mL) and
concentrated in vacuo. The crude residue was dissolved in 5
mL of 3 N aq HCl and washed again with Et2O (3 × 3 mL).
The aqueous phase was concentrated in vacuo, furnishing the
deprotected cyclic tetrapeptide, c[-Arg-Gly-Asp-Acpca19-] (1)
(22 mg, quant., corresponding to an overall yield of 24%), as a
hydrochloride salt.
Peptide Purification. Cyclic tetrapeptide was purified by
semipreparative RP-HPLC (RP C18-10 µm, 250 mm × 10 mm)
using acetonitrile (0.05% TFA) in H2O (0.05% TFA), with a
0-25% linear gradient over 25 min. A flow rate of 5.0 mL/
min was used, and detection was at 220 nm. HPLC tR ) 10.2
min. The purity of the final cyclopeptide was checked with
analytical HPLC (Discovery C18-10 µm column, 250 mm ×
4.6 mm) in two different solvent systems (methanol/water and
acetonitrile/water) using a gradient program and found to be
g98% pure. The HPLC sample was evaporated under vacuum
and finally transformed into the hydrochloride salt by exposing
the solid material to anhydrous gaseous HCl until constant
weight was reached, ready for biological assay. A light white
Sample Preparation and NMR Experiments. NMR
samples of cyclopeptides 1-11 were prepared by dissolving
3.0-5.0 mg of the analogue in 0.6 mL of 90/10 H2O/D2O at pH
5.5. All experiments were recorded at 300 ( 0.5 K on a Varian
Inova 600 or 800 spectrometer equipped with the triple
resonance probe. The chemical shifts were referenced to the
H2O signal located at 4.754 ppm. To assign the proton
resonances, conventional 2D experiments, such as DQF-COSY,
TOCSY, NOESY, and ROESY were recorded. The TOCSY
spectrum was acquired using an MLEV-17 spin-lock se-
quence30 at a field strength of 10 kHz and an evolution time
of 80 ms. NOESY experiments were carried out with mixing
times of 200 and 400 ms. ROESY spectra were also collected,
using a continuous wave mixing of 200 and 400 ms (3.0 kHz
spin-locking field strength). For the temperature coefficients
of amide protons, six 1D and 2D TOCSY experiments were
acquired in a temperature range from 278 to 313 K. All 2D
experiments were acquired in the phase-sensitive mode using
the method of States et al.,31 recording 512-1024 t1 experi-
ments with a variable number of transients of 2048 complex
points for each free induction decay. The spectral width in both
dimensions was typically 10 ppm. In all spectra, water
suppression was achieved using the WATERGATE technique.
The transmitter offset was placed at the water resonance. Prior
to Fourier transformation, the time-domain data were zero-
filled in both dimensions to yield 4k × 2k matrixes and
apodized by a shifted-squared sine bell window function to
improve resolution. When required, a fourth-order polynomial
baseline correction algorithm was applied after transformation
and phasing. The cross-peak intensities were measured from
the NOESY spectrum with a mixing time of 200 ms. No
differences were observed in the NOESY spectra with different
mixing times. To relate the NOEs data with the interproton
distances, the calibration was made using the distance of 1.8
Å for the well-defined geminal â-protons. The NOE intensities
were classified as strong, medium, and weak, corresponding
to upper bound distance constraints of 2.7, 3.5, and 5.0 Å,
respectively. Lower bounds between nonbonded atoms were
set to the sum of their van der Waals radii (1.8 Å). Pseudoatom
corrections were added to interproton distance restraints when
necessary.
1
solid; [R]20 -10.6 (c 0.4, H2O); H and 13C NMR: see Table
D
S1. HRMS (ES+) C18H30N7O8 calcd for [MH]+ 472.2156, found
472.2162.
c[-Arg-Gly-Asp-Acpca20-] (2). Resin loading: 1.8 mmol/
g. Overall yield: 26% (43 mg). A white solid; [R]20 +27.1 (c
D
0.6, H2O) (HCl salt); HPLC purity: g98%; HPLC tR ) 9.3 min;
1H and 13C NMR: see Table S1. HRMS (ES+) C18H30N7O8 calcd
for [MH]+ 472.2156, found 472.2167.
c[-Arg-Gly-Asp-Acpca21-] (3). Resin loading: 1.8 mmol/
g. Overall yield: 36% (68 mg). A colorless glassy solid; [R]20
D
+13.3 (c 0.7, H2O) (HCl salt); HPLC purity: g98%; HPLC tR
) 7.5 min; 1H and 13C NMR: see Table S1. HRMS (ES+)
C18H30N7O8 calcd for [MH]+ 472.2156, found 472.2151.
c[-Arg-Gly-Asp-Acpca22-] (4). Resin loading: 1.8 mmol/
g. Overall yield: 36% (66 mg). A white solid; [R]20 +5.7 (c
D
0.6, H2O) (HCl salt); HPLC purity: g98%; HPLC tR ) 8.6 min;
1H and 13C NMR: see Table S1. HRMS (ES+) C18H30N7O8 calcd
for [MH]+ 472.2156, found 472.2165.
c[-Arg-Gly-Asp-Acpca35-] (5). Resin loading: 1.55 mmol/
g. Overall yield: 30% (35 mg). A white solid; [R]20 +25.2 (c
D
0.7, H2O) (HCl salt); HPLC purity: g98%; HPLC tR ) 9.6 min;
1H and 13C NMR: see Table S2. HRMS (ES+) C18H30N7O7 calcd
for [MH]+ 456.2207, found 456.2218.
c[-Arg-Gly-Asp-Acpca36-] (6). Resin loading: 1.55 mmol/
g. Overall yield: 37% (34 mg). A glassy solid; [R]20 +9.2 (c
D
0.3, H2O) (HCl salt); HPLC purity: g98%; HPLC tR ) 8.5 min;
1H and 13C NMR: see Table S2. HRMS (ES+) C18H30N7O7 calcd
for [MH]+ 456.2207, found 456.2196.
c[-Arg-Gly-Asp-Acpca34-] (7). Resin loading: 1.55 mmol/
g. Overall yield: 15% (24 mg). A white solid; [R]20 -14.7 (c
D
0.1, H2O) (HCl salt); HPLC purity: g98%; HPLC tR ) 9.2 min;
1H and 13C NMR: see Table S2. HRMS (ES+) C18H30N7O7 calcd
for [MH]+ 456.2207, found 456.2193.
c[-Arg-Gly-Asp-Acpca30-] (8). Resin loading: 1.6 mmol/
g. Overall yield: 25% (30 mg). A white solid; [R]20 +12.6 (c
D
0.5, H2O) (HCl salt); HPLC purity: g98%; HPLC tR ) 11.4
min; 1H and 13C NMR: see Table S3. HRMS (ES+) C18H30N7O6
calcd for [MH]+ 440.2258, found 440.2267.
3
Vicinal proton coupling constants JNH-RH were measured
c[-Arg-Gly-Asp-Acpca31-] (9). Resin loading: 1.6 mmol/
g. Overall yield: 25% (33 mg). A light white solid; [R]20D +10.0
(c 0.7, H2O) (HCl salt); HPLC purity: g98%; HPLC tR ) 11.1
min; 1H and 13C NMR: see Table S3. HRMS (ES+) C18H30N7O6
calcd for [MH]+ 440.2258, found 440.2244.
from well-digitized 1D proton spectra and from DQF-COSY
spectra, when in the presence of resonance overlap. The
3
dihedral angles φ have been estimated from the JNH-RH
coupling constants using the Karplus relation.

Cyclopentane-Based RGD Integrin Binders
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7685
Structure Determination. The Acpca modules in the
selected ligands were built using the BIOPOLYMER module
of the INSIGHTII package. The linear RGD peptides were
added to the template models. All calculations were carried
out on a Silicon Graphics Octane workstation using the CVFF,
implemented in the DISCOVER (Accelrys, San Diego, CA)
software package, together with the INSIGHTII program as
graphic interface. The structures of the compounds were
computed using the simulated annealing (SA) method in the
NMR refine module. The simulated annealing calculations
included some distinct phases. Phase 1 involved randomization
of all atomic coordinates, followed by minimization of the
starting structures using a quadratic potential and very low
force constants for each term of the pseudoenergy function,
including chiral and NOE constraints. Phase 2 involved
simulated annealing with a progressive increment of the force
constants up to their full values. Phase 3 involved the cooling
of the molecule from 1000 to 300 K over 10 ps. At the end of
this protocol, the structures were energy minimized with full
CVFF (Morse and Lennard-Jones potentials, Coulombic term)
by steepest descents and conjugate gradients using several
thousand iterations, until the maximum derivative was less
than 0.001 kcal/Å. Amide bonds were kept in the more stable
trans configuration. The quality of the final structures was
analyzed on the basis of the number of NOE distance violations
(lower than 0.3 Å) and energy values. The 3D structures of
analogues 1, 3, 6, 7, 9, and 10 are available on request to us.
Hydrogen bonds were searched using the Measure Hbond
facility of INSIGHT and were regarded as present if the
following criteria were satisfied simultaneously: (1) the
distance between the donor H and the acceptor O was less than
2.5 Å; (2) the angle between the heavy-atom donor, the
hydrogen, and the heavy-atom acceptor (NH-O) was greater
than 120°.
A family of 100 structures was calculated for all compounds,
and 20 models were selected on the basis of their lowest energy
to represent a possible 3D structure of the compounds.
Docking Protocol. The automated molecular docking
calculations were carried out using the Glide (grid-based ligand
docking with energetics) module implemented in the Schro¨d-
inger’s FirstDiscovery Suite, version 2.7.24 Any docking tool
needs to combine a docking engine with a fast-scoring function.
Glide uses a hierarchical series of filters to search for possible
locations of the ligand in the active-site region of the receptor.
The shape and properties of the receptor are represented on a
grid by several different sets of fields that provide progressively
more accurate scoring of the ligand poses. A pose is a complete
specification of the ligand: its position and orientation.
Although the protein is required to be rigid, the program
allows torsional flexibility in the ligand. Conformational
flexibility is handled in Glide via an extensive conformational
search, augmented by a heuristic screen that rapidily elimi-
nates conformations deemed unsuitable for binding to a
receptor. During the docking process, the macrocycle backbone
conformation of the ligands was held fixed, whereas the side-
chain dihedral angles were free to rotate.
347 for chain â. Because of the lack of parameters, the Mn2+
ions in the experimental protein structure were modeled by
replacing them with Ca2+ ions. The protein-charged groups
that were neither located in the ligand-binding pocket nor
involved in salt bridges were neutralized using the Schro¨dinger
pprep script. The hydrogens were added using the Schro¨dinger
graphical user interface Maestro,24 and the resulting structure
was optimized using the Schro¨dinger impref script.
The grid-generation step requires mae input files of both
ligand and active site, including hydrogen atoms. The center
of the grid-enclosing box was defined by the center of the bound
ligand, as described in the original PDB entry. The enclosing
box dimensions, which are automatically deduced from the
ligand size, fit the entire active site. For the docking step, the
size of the bounding box for placing the ligand center was set
to 12 Å. No further modifications were applied to the default
settings. The GlideScore scoring function was used to select
30 poses for each ligand.33
The Glide program was initially tested for its ability to
reproduce the crystallized binding geometry of EMD121974.
The program was successful in reproducing the experimentally
found binding mode of this compound, as it corresponds to the
best-scored pose.
Acknowledgment. This work was supported by
Ministero dell’Istruzione, dell’Universita` e della Ricerca
(MIUR, COFIN 2004, and FIRB 2001). We thank Dr.
Emanuela Azara for LC-MS spectrometric analyses,
the Centro Interfacolta` Misure “G. Casnati” (Universita`
di Parma), and the NMR Centre of Mill Hill at the
National Institute for Medical Research, London (UK)
for instrumental facilities. A doctoral fellowship to P.B.
from the University of Sassari is gratefully acknowl-
edged.
Supporting Information Available: Experimental pro-
cedures and spectroscopic data for compounds 14-22 and 30-
36, NMR data (Tables S1-S3), and copies of 1H NMR spectra
of cyclopeptides 1-11, torsion angles of energy minimized
conformers of 1, 3, 6, 7, 9, and 10 (Table S4), as well as table
of elemental analyses (Table S5) of relevant intermediary
compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
References
(1) (a) Giannis, A.; Kolter, T. Peptidomimetics for Receptor Ligands.
Discovery, Development, and Medical Perspectives. Angew.
Chem., Int. Ed. Engl. 1993, 32, 1244-1267. (b) Advances in
Amino Acid Mimetics and Peptidomimetics; Abell, A., Ed.; JAI
Press: Greenwich, 1997; Vol. 1. (c) Advances in Amino Acid
Mimetics and Peptidomimetics; Abell, A., Ed.; JAI Press: Stan-
ford, CT, 1999; Vol. 2. (d) Randolph, J. T.; DeGoey, D. A.
Peptidomimetic Inhibitors of HIV Protease. Curr. Top. Med.
Chem. 2004, 4, 1079-1095. (e) Lohof, E.; Born, M. A.; Kessler,
H. In Synthesis of Peptides and Peptidomimetics; Goodman, M.,
Felix, A., Moroder, L., Toniolo, C., Eds.; Houben-Weyl Methods
of Organic Chemistry, 4th ed.; Georg Thieme Verlag: Stuttgart,
2003; Vol. E22b. (f) Fairlie, D. P.; Abbenante, G.; March, D. R.
Macrocyclic Peptidomimetics - Forcing Peptides into Bioactive
Conformations. Curr. Med. Chem. 1995, 2, 654-686.
(2) (a) Fujii, N.; Oishi, S.; Hiramatsu, K.; Araki, T.; Ueda, S.;
Tamamura, H.; Otaka, A.; Kusano, S.; Terakubo, S.; Nakashima,
H.; Broach, J. A.; Trent, J. O.; Wang, Z.; Peiper, S. C. Molecular-
Size Reduction of a Potent CXCR4-chemokine Antagonist using
Orthogonal Combination of Conformation- and Sequence-Based
Libraries. Angew. Chem., Int. Ed. 2003, 42, 3251-3253. (b)
Wermuth, J.; Goodman, S. L.; Jonczyk, A.; Kessler, H. Stereo-
isomerism and Biological Activity of the Selective and Super-
active RVâ3 Integrin Inhibitor cyclo(-RGDfV-) and its Retro-
Inverso Peptide. J. Am. Chem. Soc. 1997, 119, 1328-1335. (c)
Nikiforovich, G. V.; Kover, K. E.; Zhang, W.-J.; Marshall, G. R.
Cyclopentapeptides as Flexible Conformational Templates. J.
Am. Chem. Soc. 2000, 122, 3262-3273. (d) Porcelli, M.; Casu,
M.; Lai, A.; Saba, G.; Pinori, M.; Cappelletti, S.; Mascagni, P.
Cyclic Pentapeptides of Chiral Sequence DLDDL as Scaffold for
Antagonism of G-protein Coupled Receptors: Synthesis, Activity
and Conformational Analysis by NMR and Molecular Dynamics
of ITF 1565 a Substance P Inhibitor. Biopolymers 1999, 50, 211-
219.
The refined poses are scored using Schro¨dinger’s proprietary
GlideScore scoring function. GlideScore is based on Chem-
Score32 but includes a steric-clash term and adds buried polar
terms devised by Schro¨dinger to penalize electrostatic mis-
matches.
For the selected Acpca-containing ligands, representative
energy-minimized conformers derived from the NMR data and
featuring different cyclopeptide backbone geometries were
used as starting conformations.
The recently solved crystal structure of the extracellular
domain of the RVâ3 integrin receptor in complex with
EMD121974 and in the presence of the proadhesive ion Mn2+
(PDB entry code ) 1L5G) was used for docking studies.
Because the headgroup of RVâ3 integrin has been identified in
the X-ray structure as the ligand binding region, the docking
was performed only on the globular head. The protein struc-
ture was setup for docking as follows. The protein was
truncated to residue sequences 41-342 for chain R and 114-

7686 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24
Casiraghi et al.
(3) (a) Hanessian, S.; McNaughton-Smith, G. A Versatile Synthesis
of a â-turn Peptidomimetic Scaffold: An Approach Towards a
Designed Model Antagonist of the Tachykinin NK-2 Receptor.
Bioorg. Med. Chem. Lett. 1996, 6, 1567-1572. (b) Belvisi, L.;
Bernardi, A.; Checchia, A.; Manzoni, L.; Potenza, D.; Scolastico,
C.; Castorina, M.; Capelli, A.; Giannini, G.; Carminati, P.;
Pisano, C. Potent Integrin Antagonists from a Small Library of
RGD-Including Cyclic Pseudopeptides. Org. Lett. 2001, 3, 1001-
1004. (c) Gosselin, F.; Tourwe, D.; Ceusters, M.; Meert, T.;
Heylen, L.; Jurzak, M.; Lubell, W. D. Probing Opioid Receptor-
Ligand Interactions by Employment of Indolizidin-9-One Amino
Acid as a Constrained Gly2-Gly3 Surrogate in a Leucine-
Enkephalin Mimic. J. Pept. Res. 2001, 57, 337-344. (d) Andreu,
D.; Ruiz, S.; Carreno, C.; Alsina, J.; Albericio, F.; Jime´nez, M.
A.; De La Figuera, N.; Herranz, R.; Garcia-Lopez, M. T.;
(12) Bazan-Socha, S.; Bukiej, A.; Marcinkiewicz, C.; Musial, J.
Integrins in Pulmonary Inflammatory Diseases. Curr. Pharm.
Des. 2005, 11, 893-901.
(13) (a) Pytela, R.; Pierschbacher, M. D.; Argraves, S.; Suzuki, S.;
Ruoslahti, E. Arginine-Glycine-Aspartic Acid Adhesion Recep-
tors. Methods Enzymol. 1987, 144, 475-489. (b) Goodman, S.
L.; Holzemann, G.; Sulyok, G. A. G.; Kessler, H. Nanomolar
Small Molecule Inhibitors for RVâ6, RVâ5, and RVâ3 Integrins. J.
Med. Chem. 2002, 45, 1045-1051.
(14) (a) Dechantsreiter, M. A.; Planker, E.; Matha¨, B.; Lohof, E.;
Ho¨lzemann, G.; Jonczyk, A.; Goodman, S. L.; Kessler, H.
N-Methylated Cyclic RGD Peptides as Highly Active and Selec-
tive RVâ3 Integrin Antagonists. J. Med. Chem. 1999, 42, 3033-
3040. (b) Belvisi, L.; Caporale, A.; Colombo, M.; Manzoni, L.;
Potenza, D.; Scolastico, C.; Castorina, M.; Cati, M.; Giannini,
G.; Pisano, C. Cyclic RGD Peptides Containing Azabicycloalkane
Reverse-Turn Mimics. Helv. Chim. Acta 2002, 85, 4353-4368.
(c) Haubner, R.; Schmitt, W.; Ho¨lzemann, G.; Goodman, S. L.;
Jonczyk, A.; Kessler, H. Cyclic RGD Peptides Containing â-Turn
Mimetics. J. Am. Chem. Soc. 1996, 118, 7881-7891. (d) Lohof,
E.; Planker, E.; Mang, C.; Burkhart, F.; Dechantsreiter, M. A.;
Haubner, R.; Wester, H. J.; Schwaiger, M.; Ho¨lzemann, G.;
Goodman, S. L.; Kessler, H. Carbohydrate Derivatives for Use
in Drug Design: Cyclic RV-Selective RGD Peptides. Angew.
Chem., Int. Ed. 2000, 39, 2761-2764.
Gonzalez-Muniz, R. IBTM-Containing Gramicidin
S Ana-
logues: Evidence for IBTM as a Suitable Type II′ â-Turn
Mimetic. J. Am. Chem. Soc. 1997, 119, 10579-10586.
(4) (a) Vianello, P.; Cozzi, P.; Galvani, A.; Meroni, M.; Varasi, M.;
Volpi, D.; Bandiera, T. Solid-Phase Synthesis of a Small Library
of 3-Phenylthio-3-nicotinyl Propionic Acid Derivatives Acting as
Antagonists of the Integrin RVâ3. Bioorg. Med. Chem. Lett. 2004,
14, 657-661. (b) Gruner, S. A. W.; Ke´ri, G.; Schwab, R.;
Venetianer, A.; Kessler, H. Sugar Amino Acid Containing
Somatostatin Analogues that Induce Apoptosis in Both Drug-
Sensitive and Multidrug-Resistant Tumor Cells. Org. Lett. 2001,
3, 3723-3725. (c) Chakraborty, T. K.; Ghosh, S.; Jayaprakash,
S.; Sharma, J. A. R. P.; Ravikanth, V.; Diwan, P. V.; Nagaraj,
R.; Kunwar, A. C. Synthesis and Conformational Studies of
Peptidomimetics Containing Furanoid Sugar Amino Acids and
a Sugar Diacid. J. Org. Chem. 2000, 65, 6441-6457.
(5) Cox, D.; Aoki, T.; Seki, J.; Motoyama, Y.; Yoshida, K. The
Pharmacology of the Integrins. Med. Res. Rev. 1994, 14, 195-
228.
(6) (a) Ojima, I.; Chakravarty, S.; Dong, Q. Antithrombotic Agents:
From RGD to Peptide Mimetics. Bioorg. Med. Chem. 1995, 3,
337-360. (b) Samanen, J. GPIIb/IIIa Antagonists. Annu. Rep.
Med. Chem. 1996, 31, 91-100.
(7) (a) Maruga´n, J. J.; Manthey, C.; Anaclerio, B.; Lafrance, L.; Lu,
T.; Markotan, T.; Leonard, K. A.; Crysler, C.; Eisennagel, S.;
Dasgupta, M.; Tomczuk, B. Design, Synthesis, and Biological
Evaluation of Novel Potent and Selective RVâ3/RVâ5 Integrin Dual
Inhibitors with Improved Bioavailability. Selection of the Mo-
lecular Core. J. Med. Chem. 2005, 48, 926-934. (b) Cardo-Vila,
M.; Arap, W.; Pasqualini, R. RVâ5 Integrin-Dependent Pro-
grammed Cell Death Triggered by a Peptide Mimic of Annexin
V. Mol. Cell 2003, 11, 1151-1162. (c) Eskens, F. A. L. M.;
Dumez, H.; Hoekstra, R.; Perschl, A.; Brindley, C.; Bottcher, S.;
Wynendaele, W.; Drevs, J.; Verweij, J.; van Oosterom, A. T.
Phase I and Pharmacokinetic Study of Continuous Twice Weekly
Intravenous Administration of Cilengitide (EMD 121974), a
Novel Inhibitor of the Integrins RVâ3 and RVâ5 in Patients with
Advanced Solid Tumours. Eur. J. Cancer 2003, 39, 917-926.
(d) Stro¨mblad, S.; Cheresh, D. A. Integrins, Angiogenesis and
Vascular Cell Survival. Chem. Biol. 1996, 3, 881-885. (e) Chen,
X.; Plasencia, C.; Hou, Y.; Neamati, N. Synthesis and Biological
Evaluation of Dimeric RGD Peptide-Paclitaxel Conjugate as a
Model for Integrin-Targeted Drug Delivery. J. Med. Chem. 2005,
48, 1098-1106 (corigendum, J. Med. Chem. 2005, 48, 5874).
(8) (a) Wang, J.; Sprinter, T. A. Structural Specializations of
Immunoglobulin Superfamily Members for Adhesion to Integrins
and Viruses. Immunol. Rev. 1998, 163, 197-215. (b) Shimizu,
Y.; Rose, D. M.; Ginsberg, M. H. Integrins in the Immune
System. Adv. Immunol. 1999, 72, 325-380.
(9) (a) Danen, E. H. J. Integrins: Regulators of Tissue Function and
Cancer Progression. Curr. Pharm. Des. 2005, 11, 881-891. (b)
Yun, Z.; Menter, D. G.; Nicolson, G. L. Involvment of Integrin
RVâ3 in Cell Adhesion, Motility, and Liver Metastasis of Murine
RAW117 Large Cell Lymphoma. Cancer Res. 1996, 56, 3103-
3111.
(10) (a) Duong, L. T.; Rodan, G. A. Regulation of Osteoclast Formation
and Function. Rev. Endocr. Metab. Disord. 2001, 2, 95-104. (b)
Robey, P. G. RGD-Containing Proteins and Bone. Annu. Rep.
Med. Chem. 1993, 28, 227-236. (c) Engleman, V. W.; Nickols,
G. A.; Ross, F. P.; Horton, M. A.; Griggs, D. W.; Settle, S. L.;
Ruminski, P. G.; Teitelbaum, S. L. A Peptidomimetic Antagonist
of the RVâ3 Integrin Inhibits Bone Resorption in vitro and
Prevents Osteoporosis in vivo. J. Clin. Invest. 1997, 99, 2284-
2292.
(15) (a) Thornber, C. W. Isosterism and Molecular Modification in
Drug Design. Chem. Soc. Rev. 1979, 8, 563-580. (b) Chen, X.;
Wang, W. The Use of Bioisosteric Groups in Lead Optimization.
In Annual Reports in Medicinal Chemistry; Doherty, A. M., Ed.;
Elsevier Academic Press: Amsterdam, 2003; Vol. 38, pp 333-
346.
(16) For recent examples of nonisosteric replacement in RGD-
containing cyclic ligands, see: (a) van Well, R. M.; Overkleeft,
H. S.; van der Marel, G: A.; Bruss, D.; Thibault, G.; de Groot,
P. G.; van Boom, J. H.; Overhand, M. Solid-Phase Synthesis of
Cyclic RGD-Furanoid Sugar Amino Acid Peptides as Integrin
Inhibitors. Bioorg. Med. Chem. Lett. 2003, 13, 331-334. (b) Royo,
M.; Van Den Nest, W.; del Fresno, M.; Frieden, A.; Yahalom,
D.; Rosenblatt, M.; Chorev, M.; Albericio, F. Solid-Phase Syn-
theses of Constrained RGD Scaffolds and Their Binding to the
RVâ3 Integrin Receptor. Tetrahedron Lett. 2001, 42, 7387-7391.
(c) Schumann, F.; Mu¨ller, A.; Koksch, M.; Mu¨ller, G.; Sewald,
N. Are â-Amino Acids γ-Turn Mimetics? Exploring a New Design
Principle for Bioactive Cyclopeptides. J. Am. Chem. Soc. 2000,
122, 12009-12010. (d) Annis, D. A.; Helluin, O.; Jacobsen, E.
N. Stereochemistry as a Diversity Element: Solid-Phase Syn-
thesis of Cyclic RGD Peptide Derivatives by Asymmetric Ca-
talysis. Angew. Chem., Int. Ed. 1998, 37, 1907-1909.
(17) Kumar, C. C.; Nie, H.; Rogers, C. P.; Malkowski, M.; Maxwell,
E.; Catino, J. J.; Armstrong, L. Biochemical Characterization of
the Binding of Echistatin to Integrin RVâ3 Receptor. J. Phar-
macol. Exp. Ther. 1997, 283, 843-853.
(18) (a) Xiong, J.-P.; Stehle, T.; Diefenbach, B.; Zhang, R.; Dunker,
R.; Scott, D. L.; Joachimiak, A.; Goodman, S. L.; Arnaout, M. A.
Crystal Structure of the Extracellular Segment of Integrin RVâ3.
Science 2001, 294, 339-345. (b) Xiong, J.-P.; Stehle, T.; Zhang,
R.; Joachimiak, A.; Frech, M.; Goodman, S. L.; Arnaout, M. A.
Crystal Structure of the Extracellular Segment of Integrin RVâ3
in Complex with an Arg-Gly-Asp Ligand. Science 2002, 296,
151-155.
(19) Rassu, G.; Auzzas, L.; Pinna, L.; Zambrano, V.; Zanardi, F.;
Battistini, L.; Marzocchi, L.; Acquotti, D.; Casiraghi, G. Variable
Strategy toward Carbasugars and Relatives. 4. Viable Access
to (4a-Carbapentofuranosyl)amines, (5a-Carbahexopyranosyl)-
amines, and Amino Acids Thereof. J. Org. Chem. 2002, 67,
5338-5342.
(20) Wu, W.-L.; Wu, Y.-L. Chemoselective Hydrolysis of Terminal
Isopropylidene Acetals and Subsequent Glycol Cleavage by
Periodic Acid in One Pot. J. Org. Chem. 1993, 58, 3586-3588.
(21) For previous syntheses of 3-aminocyclopentane carboxylic acids
see, for example: (a) Nakamura, S.; Karasawa, K.; Tanaka, N.;
Yonehara, H.; Umezawa, H. Structure of Amidinomycin. J.
Antibiot. Ser. A 1960, 13, 362-365. (b) Allan, R. D.; Johnston,
G. A. R.; Twitchin, B. Synthesis of Analogues of GABA. III. All
Four Stereoisomers of 3-Aminocyclopentanecarboxylic Acid and
a Stereochemical Correlation with Amidinomycin. Aust. J. Chem.
1979, 32, 2517-2521. (c) Evans, C.; McCague, R.; Roberts, S.
M.; Sutherland, A. G. Synthesis of Either Enantiomer of cis-3-
Aminocyclopentanecarboxylic Acid from Both Enantiomers of
Racemic 2-azabicyclo[2.2.1]hept-5-en-3-one. J. Chem. Soc., Per-
kin Trans. 1 1991, 656-657. (d) Cheˆnevert, R.; Martin, R.
Enantioselective Synthesis of (+) and (-)-cis-3-Aminocyclopen-
tanecarboxylic Acids by Enzymatic Asymmetrization. Tetrahe-
dron: Asymmetry 1992, 3, 199-200.
(11) (a) Matsuno, H.; Stassen, J. M.; Vermylen, J.; Deckmyn, H.
Inhibition of Integrin Function by a Cyclic RGD-Containing
Peptide Prevents Neointima Formation. Circulation 1994, 90,
2203-2206. (b) Choi, E. T.; Engel, L.; Callow, A. D.; Sun, S.;
Trachtenberg, J.; Santoro, S.; Ryan, U. S. Inhibition of Neoin-
timal Hyperplasia by Blocking RVâ3 Integrin with a Small
Peptide Antagonist GpenGRGDSPCA. J. Vasc. Surg. 1994, 19,
125-134.
(22) Blankenstein, J.; Zhu, J. Conformation-Directed Macrocycliza-
tion Reactions. Eur. J. Org. Chem. 2005, 1949-1964.

Cyclopentane-Based RGD Integrin Binders
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7687
(23) Herrero, S.; Garc´ıa-Lo´pez, M. T.; Latorre, M.; Cenarruzabeitia,
E.; Del R´ıo, J.; Herranz, R. 2-Oxopiperazine-Based γ-Turn
Conformationally Constrained Peptides: Synthesis of CCK-4
Analogues. J. Org. Chem. 2002, 67, 3866-3873.
S.; Wang, Z.; Oishi, S.; Esaka, A.; Trent, J. O.; Nakashima, H.;
Yamamoto, N.; Peiper, S. C.; Otaka, A.; Fujii, N. Identification
of Novel Low Molecular Weight CXCR4 Antagonists by Struc-
tural Tuning of Cyclic Tetrapeptide Scaffolds. J. Med. Chem.
2005, 48, 3280-3289.
(24) Schro¨dinger, http://www.schrodinger.com
(25) Gottschalk, K.-E.; Kessler, H. The Structures of Integrins and
Integrin-Ligand Complexes: Implications for Drug Design and
Signal Transduction. Angew. Chem., Int. Ed. 2002, 41, 3767-
3774.
(26) Marinelli, L.; Lavecchia, A.; Gottschalk, K.-E.; Novellino, E.;
Kessler, H. Docking Studies on Rvâ3 Integrin Ligands: Phar-
macophore Refinement and Implications for Drug Design. J.
Med. Chem. 2003, 46, 4393-4404.
(27) The hydrophobicity of the peptidomimetic scaffold grafted into
cyclic RGD ligands could affect the binding affinity to the
integrin receptors (see ref 25). Lowering the hydrophobic char-
acter of the ligands by hydroxyl substituents could render
dihydroxy analogues in this series less active than the nonhy-
droxylated counterparts. We thank the reviewers for drawing
this point to our attention.
(29) For a recent example of RGD peptide-paclitaxel conjugate, see
ref 7e.
(30) (a) Davis, D. G.; Bax, A. Assignment of Complex 1H NMR Spectra
via Two-Dimensional Homonuclear Hartmann-Hahn Spectros-
copy. J. Am. Chem. Soc. 1985, 107, 2820-2821. (b) Bax, A.;
Davis, D. G. MLEV-17-Based Two-Dimensional Transverse NOE
Spectroscopy. J. Magn. Reson. 1985, 63, 207-213.
(31) States, D. J.; Haberkorn, R. A.; Ruben, D. J. A Two-Dimensional
Nuclear Overhauser Experiment with Pure Absorption Phase
in Four Quadrants. J. Magn. Reson. 1982, 48, 286-292.
(32) Eldridge, M. D.; Murray, C. W.; Auton, T. R.; Paolini, G. V.; Mee,
R. P. Empirical Scoring Functions. 1. The Development of a Fast
Empirical Scoring Function to Estimate the Binding Affinity of
Ligands in Receptor Complexes. J. Comput.-Aided Mol. Des.
1997, 11, 425-445.
(28) Similar ring contractions in cyclopeptide constructs were recently
reported: (a) Sukopp, M.; Schwab, R.; Marinelli, L.; Biron, E.;
Heller, M.; Va´rkondi, E.; Pap, AÄ .; Novellino, E.; Ke´ri, G.; Kessler,
H. Structure-Activity Relationship Studies Optimizing the
Antiproliferative Activity of Novel Cyclic Somatostatin Ana-
logues Containing a Restrained Cyclic â-Amino Acid. J. Med.
Chem. 2005, 48, 2916-2926. (b) Tamamura, H.; Araki, T.; Ueda,
(33) Kellenberger, E.; Rodrigo, J.; Muller, P.; Rognan, D. Comparative
Evaluation of Eight Docking Tools for Docking and Virtual
Screening Accuracy. Proteins: Struct., Funct., Bioinf. 2004, 57,
225-242.
JM050698X