Tailoring of integrin ligands: Probing the charge capability of the metal ion-dependent adhesion site

Article abstract of DOI:10.1021/jm2013826

Intervention in integrin-mediated cell adhesion and integrin signaling pathways is an ongoing area of research in medicinal chemistry and drug development. One key element in integrin-ligand interaction is the coordination of the bivalent cation at the me

Full text of DOI:10.1021/jm2013826

Article
pubs.acs.org/jmc
Tailoring of Integrin Ligands: Probing the Charge Capability of the
Metal Ion-Dependent Adhesion Site
Markus Bollinger,† Florian Manzenrieder,† Roman Kolb,† Alexander Bochen,† Stefanie Neubauer,†
Luciana Marinelli,‡ Vittorio Limongelli,‡ Ettore Novellino,‡ Georg Moessmer,§ Reinhard Pell,∥
Wolfgang Lindner,∥ Joseph Fanous,⊥ Amnon Hoffman,⊥ and Horst Kessler*,†,@
†Institute for Advanced Study and Center of Integrated Protein Science, Department Chemie, Technische Universitat Munchen,
̈
̈
Lichtenbergstrasse 4, 85747 Garching, Germany
‡Dipartimento di Chimica Farmaceutica e Tossicologica, Universita
̀
di Napoli “Federico II”, Via D. Montesano, 49-80131 Napoli, Italy
§Institut fur Klinische Chemie und Pathobiochemie, Klinikum rechts der Isar, Technische Universitat Munchen, Ismaninger Strasse
̈
̈
̈
22, 81675 Munchen, Germany
̈
∥Institute of Analytical Chemistry, University of Vienna, Wahringer Strasse 38, A-1090 Vienna, Austria
̈
⊥School of Pharmacy, Faculty of Medicine, The Hebrew University of Jerusalem, P.O. Box 12065, Jerusalem 91120, Israel
@Chemistry Department, Faculty of Science, King Abdulaziz University, P.O. Box 80203, Jeddah 21589, Saudi Arabia
S
* Supporting Information
ABSTRACT: Intervention in integrin-mediated cell adhesion and
integrin signaling pathways is an ongoing area of research in
medicinal chemistry and drug development. One key element in
integrin−ligand interaction is the coordination of the bivalent
cation at the metal ion-dependent adhesion site (MIDAS) by a
carboxylic acid function, a consistent feature of all integrin ligands.
With the exception of the recently discovered hydroxamic acids, all
bioisosteric attempts to replace the carboxylic acid of integrin
ligands failed. We report that phosphinates as well as monomethyl
phosphonates represent excellent isosters, when introduced into
integrin antagonists for the platelet integrin αIIbβ3. The novel
inhibitors exhibit in vitro and ex vivo activities in the low
nanomolar range. Steric and charge requirements of the MIDAS
region were unraveled, thus paving the way for an in silico
prediction of ligand activity and in turn the rational design of the next generation of integrin antagonists.
(FMD).1,6,7 Constraining and mimicking RGD has successfully
been used to develop thousands of integrin ligands, all
containing an essential carboxylate moiety as a metal-
INTRODUCTION
Integrin signaling is profoundly implicated in numerous
physiological processes, such as tissue remodeling or angio-
■
coordinating group in the MIDAS. Recently, we found
hydroxamic acids as a first potent isosteric replacement of the
carboxylic group in integrins αvβ3 and α5β1.8 The known
tendency of phosphate groups to coordinate bivalent metal ions
(calcium, manganese, or magnesium) prompted us to develop
new phosphorus-containing integrin ligands with the aim of
unraveling the steric and electrostatic requirements of the
MIDAS region and obtaining significant new insights into the
binding mode of integrins for further optimizing integrin
ligands.
genesis, as well as in important pathological disorders such as
thrombosis, cancer, osteoporosis, and autoimmune diseases.
Because of their biological relevance in many diseases, integrins
represent highly important targets for medicinal chemistry.1,2
From a structural point of view, integrins are heterodimers of a
noncovalently linked α-subunit and a β-subunit. Each domain is
composed of an extracellular domain, a single membrane-
spanning helical domain, and a short cytoplasmic tail. The β-
subunit contains a metal ion-dependent adhesion site (MIDAS)
in the ligand binding domain.3−5 Among the 24 known
integrins, a number of them, known as the RGD-dependent
integrins, recognize the tripeptide sequence arginine-glycine-
aspartic acid (RGD) of extracellular matrix (ECM) proteins
(e.g., fibronectin for α5β1, fibrinogen for platelet integrin
αIIbβ3, and vitronectin for αvβ3 and other αv integrins) or
other ligands, such as ADAMs, snake venoms, or viruses
To date, extensive efforts have been made to discover and
develop integrin antagonists for clinical applications. However,
only for three integrins9 (αIIbβ3, α4β1, and αLβ2), of the 24
Received: October 14, 2011
Published: December 19, 2011
© 2011 American Chemical Society
871
dx.doi.org/10.1021/jm2013826 | J. Med. Chem. 2012, 55, 871−882

Journal of Medicinal Chemistry
Article
Figure 1. Bioisosteric replacement of the carboxylic MIDAS binding motif for the development of new αIIbβ3 integrin antagonists: phosphinic acid
1, phosphonic acid 2, monomethyl phosphonic acid 3, thiophosphonic acid 4, tirofiban analogue 5, and tirofiban 6.
a
Scheme 1. Synthesis of Integrin Ligands 1−4
a
(a) Fmoc-Cl, 10% Na2CO3, room temperature (RT), 2 h; (b) IBX, DMSO, RT, 16 h; (c) NH2OH·HCl, DIEA, DCM, RT, 16 h; (d) anhydrous
crystalline H3PO2, MeOH, reflux, 4 h; (e) PhSO2Cl, 10% Na2CO3, dioxane, RT, 3 h; (f) 20% piperidine/DMF, RT, 2 h; (g) HATU, DIEA, DMF,
RT, 16 h; (h) TFA/H2O/TIPS, RT, 16 h; (i) (1) BSA, air, DCM, RT, 1 h; (2) TFA/H2O/TIPS, 16 h; (j) (1) BSA, air, DCM, RT, 1 h; (2)
EDC·HCl, DMAP, MeOH, RT, 1 h; (3) TFA/H2O/TIPS, 16 h; (k) BSA, sulfur, DCM, RT, 1 h. All compounds were purified by RP-HPLC (for
more details, see the Supporting Information).
known, have small molecular ligands been approved as drugs.
Candidates for other integrins, such as αvβ3, are in clinical
phase III trials.10 In contrast, significant advances have been
made in targeting platelet integrin αIIbβ3.9,11 In fact, the
αIIbβ3 integrin inhibitor tirofiban12 is being successfully used
in acute antithrombotic therapy;12−14 thus, inhibition of this
receptor is now a validated way of inhibiting fibrinogen-
dependent platelet aggregation.
In this work, integrin antagonists 1−4 (Figure 1) were
developed on the basis of tirofiban analogue 5 (Figure 1)
previously described by Duggan et al.15 Compounds 1−4 allow
us to explore the sensitivity of the MIDAS region for
coordination of differently charged groups. Synthetic methods
for obtaining the desired compounds 1−4 have been developed
(Scheme 1). Enantiomers due to the chiral center in α-position
of the phosphorus-containing group were resolved via
chromatography on chiral columns and tested independently.
The ability of the novel ligands to inhibit the binding of integrin
αIIbβ3 to its corresponding ECM protein fibrinogen was tested
in a competitive ELISA, and the efficacy of the most active
compounds was proven in ex vivo experiments. To understand
the dependency of the charge of the metal-coordinating group
and the corresponding biological activities, we studied the
protonation states of compounds 1−4 by means of extensive
theoretical calculations and 31P NMR titration experiments.
The combination of ab initio calculations and molecular
872
dx.doi.org/10.1021/jm2013826 | J. Med. Chem. 2012, 55, 871−882

Journal of Medicinal Chemistry
Article
Table 1. Biological Evaluation of Activities of Ligands 1−6 on Integrin αIIbβ3
a
b
compd
R
IC50 (nM) (αIIbβ3)
EC50 (nM) (αIIbβ3)
(R)-1
(S)-1
(R)-2
(S)-2
(R)-3
(S)-3
(R)-4
(S)-4
(S)-5
(R)-5
6
PHOOH
1.2 0.06
6.3 0.73
22.7 2.7
136 10.7
3.3 0.4
7.8 0.9
PO(OH)2
PO(OH)(OMe)
PS(OH)2
276 15.6
40.8 1.9
1154 272
62.5 6.0
1322 498
0.81 0.05
4.4 0.3
COOH
tirofiban
0.95 0.09
13.6 3.3
a
b
IC50 values were derived from a competitive ELISA using immobilized fibrinogen and soluble integrin αIIbβ3. Effective concentrations (EC50) of
some key compounds for inhibition of platelet aggregation were derived from aggregation measurements using multiple-electrode aggregometry in
hirudin-anticoagulated TRAP-6-activated blood.
docking simulations has defined the binding modes of the new
ligands 1 and 2 identifying the molecular requisites for
achieving a high inhibitory activity, and on the basis of these,
a prediction of the activity of 3 and 4 was successfully executed.
this process was applied to 1 and an excess of sulfur was added,
thiophosphonic acid 4 was obtained. Furthermore, Cbz-
protected phosphonic acid 2 undergoes rapid monoesterifica-
tion at room temperature in the presence of N-[3-
(dimethylamino)propyl]-N′-ethylcarbodiimide (EDC) and 2
equiv of 4-dimethylaminopyridine (DMAP) in dry methanol.
Without DMAP, the reaction results in exclusive formation of
the dimethyl ester and no desired product 3 is observed.
We decided to synthesize all molecular probes as racemic
mixtures, as “libraries of enantiomers”. However, for a more
detailed biological evaluation, it was necessary to determine the
activity of the optically pure compounds.
Enantiomer separation was successfully performed on
quinine-based chiral zwitterionic ion-exchange-type stationary
phases developed by Hoffmann et al.22 Here, a sulfonic acid
derivative of quinine served as a zwitterionic chiral selector.23
The concept of the resolution of the racemic ampholytes, as
compounds 1−3 can be classified, is based on a simultaneous
ion pair formation of the zwitterionic chiral selector motif of
the chiral stationary phase with the individual enantiomers of
the analytes. These two diastereomerically behaving selector−
(R)-enantiomer and selector−(S)-enantiomer associates are the
basis for the enantiomer separations.22 The preference and
magnitude of molecular recognition and chiral discrimination
are based on the stereochemical properties of the chiral selector
and the ampholytes that include additional intermolecular
interaction sites like hydrogen bonding and π−π interactions
that determine the overall chromatographic enantioselectivity
and elution order.
CHEMISTRY
■
To prepare compounds 1−4, two major fragments (12 and 15)
were connected via an amide bond (Scheme 1). For the
synthesis of 12, commercially available 2-aminoethanol was
Fmoc-protected followed by oxidation with IBX.16,17 The
corresponding aldehyde 8 was converted with hydroxylamine
hydrochloride in DCM to oxime 9 in high yield.18 Refluxing
oxime 9 with anhydrous crystalline phosphinic acid in dry
methanol resulted in racemic phosphinate intermediate 10,18,19
which was sulfonylated upon treatment with phenylsulfonyl
chloride in aqueous Na2CO3 to provide Fmoc-protected
derivative 11. Fmoc deprotection and purification by reverse-
phase high-performance liquid chromatography (RP-HPLC)
afforded fragment 12 as a racemic mixture.
The second fragment containing the arginine mimic as a
piperidine moiety had already been described by Duggan et
al.15 Herein, Cbz protection was used instead of the more acid
labile Boc group, to avoid any deprotection of the secondary
amine during synthesis. Furthermore, during purification of 13,
the Cbz group allows UV detection. 2-[N-(Benzyloxycarbonyl)-
piperidin-4-yl]ethanol (13) was coupled to methyl-4-hydrox-
ybenzoate via Mitsunobu reaction with tributylphosphine and
1,1′-(azodicarbonyl)dipiperidine.20,21 Saponification of methyl
ester 14 gave benzoic acid derivative 15. Activation of 15 by use
of O-(7-azabenzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium
hexafluorophosphate (HATU) and subsequent condensation
with 12 gave Cbz-protected precursor 16.
For N-acyl-protected amino acid-type analytes with known
absolute configurations, we formulated a general model of
intermolecular interactions and chromatographic elution order
that correlates with the absolute configuration of the α-carbon
of an α-amino acid.24,25 Accordingly, we assigned the
stereocenter of compounds 1−3, taking into account the
isosteric behavior of the analytes and the Cahn−Ingold−Prelog
Deprotection of 16 with trifluoroacetic acid yielded
molecular probe 1, while the conversion of phosphinic acid
16 to the bis(trimethylsilyl)phosphonite intermediate, which
was oxidized by atmospheric oxygen, afforded after Cbz
deprotection phosphonic acid ligand 2 in high yield. When
873
dx.doi.org/10.1021/jm2013826 | J. Med. Chem. 2012, 55, 871−882

Journal of Medicinal Chemistry
Article
Figure 2. Assessing the protonation state of the molecular probes. Comparison of 31P NMR titration curves (top) and computed microspecies
distributions (middle) for molecular probes 1−4 (a−d, respectively). The physiological pH of 7.2 is shown as a dotted line to highlight the presence
of either a single protonation state [horizontal line for 1 (a), 3 (c), and 4 (d)] or two protonation states [descending line for 2 (b)]. In the calculated
microspecies distribution (bottom), structures of the most dominant protonation states are shown. The dotted line and the intersection with the
presentation of microspecies represent the protonation states at the physiological pH of 7.2.
(CIP) convention (for more details, see the Supporting
Information).
Platelet Aggregation in Whole Blood. In an attempt to
compare the potencies of binding of various compounds to
isolated platelet integrin αIIbβ3 (IC50) with their predicted
inhibitory potency for the complex biological process of platelet
aggregation, the effects of synthetic ligands (R)-1, (R)-2, (R)-3,
and tirofiban on ex vivo platelet aggregation were assessed
using impedance-based platelet aggregometry28,29 in hirudin-
anticoagulated TRAP-6-activated whole blood (Table 1, EC50).
Phosphinic acid (R)-1 was found to behave like tirofiban in
terms of platelet aggregation inhibition, whereas phosphonic
acid derivative (R)-2 was ∼20 times less potent. Phosphonic
methyl ester (R)-3 showed intermediate efficacy. There is a
very good correlation between the inhibitory potencies of
compounds in the ELISA (IC50) and their inhibitory potencies
in the platelet aggregation assay (EC50).
In Vitro Permeability Study. We have tested pharmaco-
kinetic properties (including disposition and membrane
permeability) of ligands 1−3 and tirofiban (6) by evaluating
their permeability properties with an enterocyte monolayer
derived from human colonic carcinoma cells (Caco-2 model).30
This model is commonly used to predict the degree of
intestinal permeability of therapeutic compounds as well as to
gain a certain indication regarding their likelihood of
penetrating the brain. The results could not differentiate
between the permeability properties of the four compounds in
cases where all of them exhibit poor permeability properties
(Papp < 1 × 10−8 cm/s). It should be noted that the permeability
of mannitol (positive control) used in this study was 2.4 × 10−6
cm/s. The permeability mechanism of mannitol is paracellular,
using the pores between the cells (tight junctions), with no
transcellular component. The fact that all of the tested tirofiban
analogues had significantly lower permeability values indicates
that the charged moiety of these analogues (at physiological
pH, 7.2) restricted the paracellular transport properties in an
effective manner. However, to validate this argument, we
prepared the corresponding dimethyl ester of 2 [inactive in the
RESULTS
■
In Vitro Inhibition of Integrin αIIbβ3. To validate the
inhibition properties of the compounds against αIIbβ3, a
competitive ELISA (αIIbβ3 fibrinogen assay) was performed
using soluble integrin αIIbβ3 and the immobilized natural
ligand fibrinogen.26,27 The clinically used αIIbβ3 inhibitor
tirofiban was used as an internal control (Table 1, IC50).
Carboxylic acid (S)-5 is similar in potency to tirofiban (6),
whereas the (R)-configuration is 6 times less active (0.8 nM vs
4.4 nM). Isosteric replacement of the carboxylic acid with a
phosphinic acid results in retained activity depending on its
relative configuration [1.2 nM for (R)-1 and 6.3 nM for (S)-1].
However, the phosphonic acid derivatives 2 are 20 times less
active [22.7 nM for (R)-2 and 136 nM for (S)-2] than
phosphinic acid 1, clearly indicating that additional negative
charge is not well tolerated in the MIDAS. Being aware that
phosphonic acid 2 might exist in several different protonation
states at the given pH, we used thiophosphonic acid 4 as a
molecular probe existing only in the double-negative
protonation state (Figure 2). The higher IC50 of 4 [62.5 nM
for (R)-4 and 1322 nM for the less favored (S)-4 enantiomer]
indicates that the additional charge in phosphonates and
thiophosphonates reduces the binding affinity for the MIDAS.
To further prove and underline this concept, we evaluated
monomethyl ester 3 of phosphonic acid 2, resulting in the
elimination of the additional negative charge. The (R)-3
phosphonic methyl ester regains nearly all of the activity (3.3
nM) compared to the phosphinic acid (R)-1, whereas the (S)-3
methyl ester is far less potent (1154 nM). This could indicate a
steric hindrance by the methyl group pointing toward the
binding pocket. Similar indications were obtained from the
docking calculations (for more details, see the Supporting
Information).
874
dx.doi.org/10.1021/jm2013826 | J. Med. Chem. 2012, 55, 871−882

Journal of Medicinal Chemistry
Article
Figure 3. Modes of binding of tirofiban (A), (R)-1 (B), (R)-2 in the overall neutral form (C) and in the overall anionic form (D), (R)-3 (E), and
(R)-4 (F). The αIIb domain is displayed as a yellow cartoon, while the β3 domain is colored green. The interacting residues and the ligands are
shown as licorice, while the magnesium ion is represented as a red sphere. For the sake of clarity, only the polar hydrogens are displayed. The
stereoview version of each complex is provided in Figure S3 of the Supporting Information.
αIIbβ3 fibrinogen assay (data not shown] and tested its
permeability. Unfortunately, this compound also exhibited poor
permeability.
found at physiological pH, both having a protonated piperidine
unit; however, in one case, the phosphonic acid is monoanionic
(∼86% in silico and 63% in vitro), and in the other case, it is in
the dianionic form (∼14% in silico and 37% in vitro) (Figure
2b). To further investigate the effect of the charge of the
different substituents on the phosphorus atom, two more
molecular probes, 3 and 4, were studied. In particular, the
thiophosphonate group of compound 4, like the phosphonate
group of 2, coexists at neutral pH as mono- and dianionic
species, but a shift toward the dianionic form (∼82% in silico
and 99% in vitro) was observed (Figure 2d). The in vitro
measurements showed that the second acid functionality (pKa2)
of compounds 2 and 4 is more acidic than calculated. For 2 a
pKa2 of 6.97 (7.99) and for 4 a pKa2 of 5.27 (6.53) can be
extracted from the obtained data points. Compound 3, like 1,
possesses only one single microspecies (Figure 2c).
Assessment of the Protonation State of 1−4. To
elucidate the molecular properties responsible for the different
activity profiles of the investigated compounds (Table 1) and
the mechanism of recognition between the inhibitors and the
αIIbβ3 integrin receptor (see the docking section), we
conducted an extensive computational study. First, the
protonation states of compounds 1−4 were computed for
each compound, and the distribution of microspecies between
pH 0 and 14 was calculated with the Marvin Sketch package.
The results were validated via titration experiments monitored
by 31P NMR chemical shifts that are very sensitive to ionization
state31−35 (Figure 2; for more details, see the Supporting
Information).
While 1 and 3 are characterized by two pKa values,
corresponding to the acidities of the phosphorus acid moiety
(pKa1) and the piperidinium group (pKa3), for 2 and 4 a total of
three pKa values are found [phosphonic and thiophosphonic
acid (pKa1 and pKa2, respectively) and piperidinium group
(pKa3)]. The experimental as well as the calculated (in
parentheses) pKa values show that at the physiological pH of
7.2 all piperidine moieties are protonated with pKa3 values of
10.44 (10.48) for 1, 10.06 (10.48) for 2, 11.02 (10.57) for 3,
and 12.03 (10.58) for 4. Furthermore, the first acid
functionality of the phosphorus unit (pKa1) of all compounds
1−4 is present in the completely deprotonated form at pH 7.2
with pKa1 values of greater than 1.56 (1.90) for 1, 1.01 (1.63)
for 2, 1.13 (1.63) for 3, and 1.27 (1.94) for 4. All in all, we
found that the experimental NMR results fully corroborate the
trends in the computational calculations.
Molecular Docking and Electrostatic Potential Calcu-
lations of 1−4. Compounds 1−4 were docked into the
αIIbβ3 RGD binding site with the aid of AutoDock version
4.0.36,37 Both enantiomers were studied. Results of the (S)-
enantiomers are reported in the Supporting Information along
with an explanation of the lower inhibitory potency observed
for (S)-1, (S)-2, and (S)-3 with respect to those of the
corresponding (R)-enantiomers.
As for (R)-1, in accordance with the protonation state
assessment (Figure 2a), only the neutral ligand form
(piperidine moiety protonated, phosphinate deprotonated)
was used for docking simulations. As a result, a binding
mode highly similar to that of tirofiban was observed (Figure
3A,B and Figure S2 of the Supporting Information). In fact, the
phosphinate group coordinates the magnesium ion occupying
in the MIDAS the same region that hosts the carboxylate group
of tirofiban. In particular, one of the oxygen atoms coordinates
the metal, while the other one engages two H-bonds with the
backbone NH groups of (β3)-Asn215 and (β3)-Tyr122, like
At the physiological pH of 7.2, the piperidine moiety (pKa3)
of 1 is protonated and the phosphinate group (pKa1) is fully
deprotonated (Figure 2a). For 2, two microspecies can be
875
dx.doi.org/10.1021/jm2013826 | J. Med. Chem. 2012, 55, 871−882

Journal of Medicinal Chemistry
Article
Figure 4. Electrostatic potential mapped onto the molecular surface of tirofiban, 1, 2 in the overall neutral form and in the anionic form, 3, and 4. At
the top right, the electrostatic surface of the integrin αIIbβ3 binding site complexed with compound (R)-1 is shown. Both for ligands and for the
protein, the scale of the electrostatic potential ranges from −294.178 (red) to 294.178 (blue) kbT/ec.
tirofiban in αIIbβ3 and Cilengetide in αvβ3 [Protein Data Bank
(PDB) entries 2vdm and 1L5G]. The benzenesulfonamide
moiety forms H-bonds with the (β3)-Asn215 backbone CO
group, while the phenyl ring occupies the aromatic pocket
formed by (αIIb)-Phe160, (αIIb)-Tyr190, and (β3)-Tyr166.
The aromatic ring of the p-hydroxybenzoate scaffold is close
enough to form a π−π interaction with (αIIb)-Tyr190, the
hydroxyl group of which is involved in an H-bond interaction
with the inhibitor amide CO group. This allows the piperidine
moiety to point toward the αIIb subunit, where the protonated
nitrogen group makes a salt bridge interaction with the (αIIb)-
Asp224 side chain and an H-bond with the (αIIb)-Ser225
backbone CO group. All these favorable interactions are
certainly responsible for the low nanomolar activity of (R)-1
toward the αIIbβ3 receptor.
Because phosphonic acid 2 is present at the physiological pH
of 7.2 both in the monoanionic form and in the dianionic form
(Figure 2b), docking calculations for (R)-2 were performed for
both protonation states, and the results showed a binding
conformation similar to that of tirofiban and (R)-1 (Figure
3C,D). This suggests that although the phosphonic group is
bulkier than the phosphinate moiety, it can fit in the MIDAS
region. Nevertheless, because (R)-2 shows reduced activity
compared to that of (R)-1 (Table 1), reasons different from the
steric hindrance should be responsible for the experimentally
observed lower activity. In this regard, it has to be clarified that
all docking algorithms, either based on classical force fields38 or
based on empirical free energy scoring functions39 or
knowledge-based scoring functions,40 cannot accurately predict
properties like the exact metal coordination geometry or
particular charge effects. Thus, to overcome this limitation and
to shed light on the different activity profiles of (R)-1 and (R)-
2, more accurate computational techniques must be used.
Recent progress made in the force field parametrization of
bivalent ions41 and some theoretical studies on a Mg2+
protein42,43 would suggest the use of molecular dynamics-
based approaches to sample more accurately the specific
ligand−protein interaction of each complex. Another possibility
might be to perform QM/MM calculations to accurately
describe the ligand−protein interaction at the binding site.
Unfortunately, both these approaches are computationally time
intensive and are useful for the study of only a few compounds.
To perform calculations on most of the ligands of the series in a
reasonable computational time, we decided to conduct
quantum mechanical calculations only on the ligands, thus
elucidating their different electrostatic potential profiles.
Quantum mechanical calculations revealed that (R)-1 has an
electrostatic potential profile highly similar to that of tirofiban,
particularly with regard to the metal-coordinating group, while
(R)-2 is similar to tirofiban only in its neutral form (Figure 4).
In fact, with regard to (R)-2, the electrostatic potential of the
phosphonate group in the dianionic form, existing at
physiological pH, is highly negative. Indeed, if on one hand
this improves the coordination of the metal ion, on the other
hand, it causes electrostatic repulsion effects with the
surrounding atoms in the MIDAS such as the backbone CO
groups of (β3)-Asn215 and (β3)-Asp217 or the (β3)-Glu220
side chain (Figures 3D and 4). Thus, our computational study
876
dx.doi.org/10.1021/jm2013826 | J. Med. Chem. 2012, 55, 871−882

Journal of Medicinal Chemistry
Article
of (R)-1 and (R)-2 suggests that electrostatic properties, more
than steric features, are the crucial factors for the different
activities of the two compounds.
with the ex vivo prevention of platelet aggregation measured by
multiple-electrode aggregometry in hirudin-anticoagulated
TRAP-6-activated blood. Both assays yield the same order of
activities for all tested compounds. Approximately 10-fold
higher EC50 values are required compared to inhibition of
binding of integrin to fibrinogen (Table 1, IC50).
Also in the case of (R)-3 and (R)-4, it was not possible to
predict the αIIbβ3 activities solely on the basis of the docking
calculations (see Figure 3E,F for the predicted binding modes
and the Supporting Information for further details about
docking). As shown in Figure 4, the electrostatic potential
profile calculated for (R)-3 is highly similar to those of tirofiban
and (R)-1, while compound (R)-4, like (R)-2, has a highly
negative potential localized on the thiophosphonate group
(dianionic form). Those results are in perfect line with the
ELISA data reported in Table 1. In fact, (R)-4 possesses a
reduced activity because of the preponderance (∼82% in silico
and 99% in vitro) of its dianionic form, which has the less
favorable electrostatic profile for its interaction with αIIbβ3.
Accordingly, (R)-3 has an electrostatic potential profile and an
activity comparable to those of (R)-1.
The pharmacotherapeutic activity of tirofiban is provided
following intravenous administration and is confined to the
systemic blood circulation. The results of the permeability
studies can be regarded as an indication that the new ligands
will also be restricted to the central compartment. Thus, once
administered by a parenteral route, the molecules will be
distributed within the central compartment where they produce
their antiplatelet aggregation activity. In case the novel
analogues will be further developed for clinical use (e.g., for
molecular imaging), the restricted distribution predicted for
these analogues ensures minimal side effects that could be
derived by interaction with peripheral tissues, including the
central nervous system. Thus, the poor membrane permeability
contributes to the safety profile of these analogues.
Pharmacokinetically wise, the polar moieties of these analogues
at the physiological pH of 7.2 did not allow passive diffusion
across the biological membrane, like enterocytes. It also
inhibited the transport via the tight junctions pores as
evidenced by the significantly lower permeability compared to
that of mannitol.
In summary, we here report a successful case of drug design
on the αIIbβ3 integrin that allowed us to identify the molecular
and electrostatic requisites for achieving strong αIIbβ3
inhibition. In particular, we have found that even a space-
demanding group such as methyl phosphonate (R)-3 can bind
to the αIIbβ3 MIDAS without strongly affecting the activity.
We have found that highly negatively charged metal-
coordinating groups are not well tolerated in the αIIbβ3
MIDAS. All these findings, on one hand, are extremely useful
for an easy and quick tuning of both the steric and electrostatic
features of αIIbβ3 inhibitors; on the other hand, they
demonstrate that docking calculations together with more
rigorous computational procedures, such as ab initio studies,
can be successfully used in rational design of new inhibitors
active on integrin receptors different from αIIbβ3.
DISCUSSION
■
The strong binding affinity of phosphorus ligands for bivalent
metal ions stimulated us to investigate those groups as
pharmacophores for binding to the MIDAS region of integrins.
As a proof of principle, we investigated analogues of αIIbβ3
integrin inhibitor 5, previously described by Duggan et al.,15
which is structurally related to the drug tirofiban (6) and
exhibits a similar binding affinity. The essential carboxyl group
of 5 was modified by different phosphorus-containing groups.
We found that phosphinate groups as well as a phosphonate
monomethyl ester are suitable isosteres for the carboxyl group
in integrin ligands while phosphonate and thiophosphonate
groups could not be used for this purpose. Docking of ligands
1−4 in the αIIbβ3 receptor did not reveal substantial
differences with respect to the tirofiban binding mode (Figure
3 and Figure S2 of the Supporting Information), thus excluding
steric effects as a reason for different binding potencies. On the
basis of ab initio studies, the high affinity of phosphinate 1 and
phosphonate monomethyl ester 3 can be attributed to the
lower negative charge of the ligand metal binding groups, if
compared to those of 2 and 4 (Figure 4). The differences in
charges among compounds 1 and 4 have been investigated
both by theoretical calculations and by 31P NMR measure-
ments. In phosphonate 2 and thiophosphonate 4 at neutral pH,
there are considerable amounts of dianionic species in the
equilibrium (Figure 2). Obviously, the high negative charge is
not tolerated by the MIDAS region, which is already negatively
charged (Figure 4).
The modification of the carboxylate group into a phosphinate
or a phosphonate monomethyl ester yields an attractive new
way of optimizing integrin ligands. The successful design of
non-carboxylate-containing ligands as integrin antagonists
surely opens a new era in the design and finding of novel
integrin ligands.
The dominating charge effect is accompanied by second-
order steric effects induced by substituents in α-position to the
phosphorus atom expressed in the different acceptance of
stereoisomers. Synthesis and the biological assays of the pure
(R)- and (S)-stereoisomers of 5 reveal that the (R)-enantiomer
is ∼5 times less active than the (S)-enantiomer. The same
general trend was found for compounds 1−4, bearing in mind
the fact that according to the CIP rules the notation in the
phosphorus compounds is reversed (Table 1). Molecular
docking revealed that in the case of compounds 1−4, although
the (S)-enantiomers are also able to bind the receptor, they lose
a number of favorable interactions within the binding pocket,
particularly in the case of phosphonic compounds (S)-2 and
(S)-3 (see the Supporting Information for details).
EXPERIMENTAL SECTION
■
All technical solvents were distilled prior to use. Dry solvents were
purchased from Aldrich, Fluka, or Merck. Reactions sensitive to
oxygen or water were performed in flame-dried reaction vessels under
an argon atmosphere (99.996%). Fmoc-protected amino acids and
coupling reagents were purchased from Novabiochem (Schwalbach,
Germany), Iris Biotech GmbH (Marktredwitz, Germany), and
Medalchemy (Alicante, Spain). All other chemicals and organic
solvents were purchased from commercial suppliers at the highest
purity available and used without further purification.
TLC monitoring was performed on Merck DC silica gel plates (60
F-254 on aluminum foil). Spots were detected by UV absorption at
254 nm and/or by staining with a 5% solution of ninhydrine in ethanol
or mostain [6.25 g of phosphomolybdic acid, 2.5 g of cerium(IV)
sulfate, and 15 mL of sulfuric acid in 235 mL of water] or potassium
permanganate (5% in 1 N aqueous NaOH).
In the biological evaluation of the new αIIbβ3 antagonists,
we have shown that the in vitro binding affinity fully correlates
877
dx.doi.org/10.1021/jm2013826 | J. Med. Chem. 2012, 55, 871−882

Journal of Medicinal Chemistry
Article
3
3
Flash column chromatography was performed using silica gel 60
(40−63 μm) from Merck at a pressure of 1−1.5 atm.
H), 7.84 (d, J = 7.4 Hz, 2 H), 7.70 (d, J = 8.9 Hz, 2 H), 7.49−7.32
(m, 3 H), 6.95 (d, 3J = 8.9 Hz, 2 H), 4.13 (t, 3J = 6.0 Hz, 2 H), 3.96−
3.85 (m, 1 H), 3.73−3.62 (m, 1 H), 3.55−3.44 (m, 1 H), 3.43−3.35
(m, 2 H), 3.07−2.93 (m, 2 H), 2.08−1.99 (m, 2 H), 1.98−1.87 (m, 1
H), 1.82 (dd, 3J = 12.4 Hz, 3J = 6.1 Hz, 2 H), 1.51 (m, 2 H); 13C NMR
(126 MHz, D2O/CD3CN/NaOH, RT) δ 167.7, 160.9, 141.4, 131.5,
128.6, 128.5, 125.6, 125.2, 113.7, 65.5, 52.1 (d, 1J = 134 Hz), 44.8, 41.5
(d, 2J = 5 Hz), 35.1, 32.1, 31.8; 31P NMR (101 MHz, D2O/H2O, RT)
δ 13.7 (pH 7.2); MS (EI) m/z 430.1 [M − PO(OH)2]+, 494.3 [M −
OH]+, 512.2 [M + H]+, 1023.1 [2M + H]+, 1045.1 [2M + Na]+; RP-
HPLC tR = 10.2 min (10−90% in 30 min); HRMS (ESI) m/z calcd for
C22H31N3O7P32S 512.1615 [M + H]+, found 512.1616.
Analytical HPLC was performed using Amersham Pharmacia
̈
Biotech Akta Basic 10F equipment, with a P-900 pump system, a
reversed-phase YMC-ODS-A C18 column (12 nm pore size, 5 μm
particle size, 250 mm × 4.6 mm), and UV detection (UV-900; 220 and
254 nm). The system was run at a flow rate of 1.0 mL/min over 30
min using H2O (0.1% TFA) and MeCN (0.1% TFA) as solvents.
Semipreparative HPLC was performed using Waters equipment:
System Breeze; pump system 1525, UV detector 2487 dual (220 and
254 nm); Driver Software Breeze version 3.20; column material, YMC-
ODS-A C18 (12 nm pore size, 5 μm particle size, 250 mm × 20 mm),
YMC-ODS-AQ C18 (12 nm pore size, 5 μm particle size, 250 mm ×
20 mm), or YMCbasic (proprietary, 5 μm particle size, 250 mm × 20
mm).
HPLC−ESI-MS analyses were performed on a Hewlett-Packard
Series HP 1100 system with a Finnigan LCQ mass spectrometer using
a YMC-Hydrosphere C18 column (12 nm pore size, 3 μm particle size,
125 mm × 2.1 mm) or a YMC-Octyl C8 column (20 nm pore size, 5
μm particle size, 250 mm × 2.1 mm). The system uses H2O (0.1%
formic acid) and MeCN (0.1% formic acid) as eluents.
Methyl-1-(phenylsulfonamido)-2-{4-[2-(piperidin-4-yl)-
ethoxy]benzamido}ethylphosphonate (3). A mixture of TFA,
TIPS, and water [5 mL, 95:2.5:2.5 (v/v/v) TFA/TIPS/H2O] was
added to 18 (16.7 mg, 0.025 mmol) and the mixture stirred at RT for
16 h. Purification by semipreparative RP-HPLC and lyophilization
1
gave 3 (10.4 mg, 0.020 mmol, 78%) as a white solid: H NMR (500
3
MHz, MeOH-d3, RT) δ 8.60−8.40 (m, 1 H), 8.18 (t, J = 5.2 Hz, 1
H), 7.88 (d, 3J = 8.0 Hz, 2 H), 7.74 (d, 3J = 9.0 Hz, 2 H), 7.49 (t, 3J =
7.2 Hz, 1 H), 7.44 (t, 3J = 7.5 Hz, 2 H), 7.18−7.11 (m, 1 H), 6.95 (d,
3J = 9.0 Hz, 2 H), 4.11 (t, 3J = 6.2 Hz, 2 H), 3.73 (ddd, 3J = 16.8 Hz, 3J
= 13.3 Hz, 3J = 8.1 Hz, 1 H), 3.65 (ddd, 2J = 19.2 Hz, 3J = 9.8 Hz, 3J =
High-resolution mass spectrometry was conducted on a Thermo
Finnigan LTQ-FT (ESI-ICR) spectrometer.
3
1H NMR,13C NMR, and 31P NMR spectra were recorded at 295 K
on 500 MHz Bruker DMX, 360 MHz Bruker AV, and a 250 MHz
Bruker AV spectrometers, respectively (Bruker, Karlsruhe, Germany).
Chemical shifts (δ) are given in parts per million. The following
solvent peaks were used as internal standards: DMSO-d5, 2.50 ppm
(1H NMR) and 39.52 ppm (13C NMR); CHCl3, 7.26 ppm (1H NMR)
and 77.16 ppm (13C NMR); MeOH-d3, 3.31 ppm (1H NMR) and
49.00 ppm (13C NMR).44 With MeOH-d3 as the solvent, standard
pulse sequences provided by Bruker were used to eliminate the solvent
peak (watergate, P3919GP; presaturation, ZGPR). For 31P NMR
spectra, 85% phosphoric acid was used as an external standard.
Duggan ligand (S)-5 was synthesized according to literature
procedures,15 starting from Nα-Fmoc-L-2,3-diaminopropionic acid.
Synthesis of the corresponding (R)-enantiomer was conducted as
4.9 Hz, 1 H), 3.53−3.49 (m, 1 H), 3.42−3.35 (m, 2 H), 3.35 (d, J =
10.5 Hz, 3 H), 3.06−2.32 (m, 2 H), 2.06−1.97 (m, 2 H), 1.97−1.86
(m, 1 H), 1.80 (dt, 3J = 6.5 Hz, 3J = 6.3 Hz, 2 H), 1.52−1.39 (m, 2 H);
13C NMR (126 MHz, MeOH-d3, RT) δ 169.7, 163.1, 142.9, 133.5,
130.3, 130.1, 128.1, 127.7, 115.2, 66.5, 52.7 (d, 2J = 6.3 Hz), 50.9 (d, 1J
= 147.1 Hz), 45.5, 42.4 (d, J = 6.1 Hz), 36.2, 32.4, 30.1; 31P NMR
2
(101 MHz, D2O/H2O, RT) δ 17.6 (pH 6.91); MS (ESI) m/z 494.2
[M − OMe]+, 526.2 [M + H]+, 548.2 [M + Na]+, 1051.1 [2M + H]+,
1073.1 [2M + Na]+, 1576.0 [3M + H]+; RP-HPLC tR = 10.8 min (10−
90% in 30 min); HRMS (ESI) m/z calcd for C23H33N3O7P32S
526.1777 [M + H]+, found 526.1769.
1-(Phenylsulfonamido)-2-{4-[2-(piperidin-4-yl)ethoxy]-
benzamido}ethylthiophosphonic Acid (4). N,O-Bis-
(trimethylsilyl)acetamide (BSA, 18.6 μL, 0.076 mmol) was added to
a mixture of 1 (9.60 mg, 0.019 mmol) and sulfur powder (1.86 mg,
0.058 mmol) in DCM (dry, 5 mL) at 0 °C under an argon
atmosphere.34 The mixture was allowed to warm to RT and stirred for
1 h. Concentration in vacuo and purification by semipreparative RP-
HPLC gave 4 (7.34 mg, 0.014 mmol, 72%) as a white solid: 1H NMR
(500 MHz, MeOD-d4, RT) δ 7.87 (d, 3J = 7.8 Hz, 2 H), 7.75 (d, 3J =
8.6 Hz, 2 H), 7.46 (dd, 3J = 8.9 Hz, 3J = 15.9 Hz, 1 H), 7.42 (t, 3J = 7.4
β
described here, starting from Nα-Boc-N -Fmoc-D-2,3-diaminopropionic
acid. Standard peptide coupling techniques were employed.
All yields are not optimized. The analytical data of compounds 7−
18 are listed in the Supporting Information. All tested compounds
were ≥95% pure as determined by RP-HPLC-(MS).
1-(Phenylsulfonamido)-2-{4-[2-(piperidin-4-yl)ethoxy]-
benzamido}ethylphosphinic Acid (1). A mixture of TFA, TIPS,
and water [5 mL, 95:2.5:2.5 (v/v/v) TFA/TIPS/H2O] was added to
16 (36.0 mg, 0.057 mmol) and the mixture stirred at room
temperature (RT) for 16 h. Purification by semi-preparative RP-
HPLC and lyophilization gave 1 (23.5 mg, 0.047 mmol, 83%) as a
3
3
Hz, 2 H), 6.96 (d, J = 8.6 Hz, 2 H), 4.13 (t, J = 6.0 Hz, 2 H), 3.80
3
3
(dt, J = 11.4 Hz, J = 4.5 Hz, 2 H), 3.65−3.56 (m, 1 H), 3.42−3.36
(m, 2 H), 3.04−2.95 (m, 2 H), 2.07−1.99 (m, 2 H), 1.98−1.87 (m, 1
H), 1.85−1.79 (m, 1 H), 1.53−1.42 (m, 2 H); 13C NMR (126 MHz,
MeOD-d4, RT) δ 169.7, 163.1, 142.6, 133.4, 130.3, 129.9, 128.2, 127.7,
115.1, 66.5, 56.8 (d, 1J = 114.7 Hz), 45.3, 42.3 (d, 2J = 11.3 Hz), 36.2,
32.3, 30.0; 31P NMR (101 MHz, D2O/H2O, RT) δ 51.8 (pH 6.73);
MS (ESI) m/z 494.2 [M − 2OH]+, 510.2 [M − OH]+, 528.1 [M +
H]+, 550.2 [M + Na]+, 1055.0 [2M + H]+; RP-HPLC tR = 19.1 min
(10−50% in 30 min); HRMS (ESI) m/z calcd for C22H31N3O6P32S2
528.1392 [M + H]+, found 528.1374.
1
white solid: H NMR (500 MHz, MeOH-d3, RT) δ 8.53 (br s, 1 H),
8.24 (br s, 1 H), 8.10 (t, 3J = 5.0 Hz, 1 H), 7.83 (d, 3J = 7.6 Hz, 2 H),
3
7.66 (d, 3J = 9.0 Hz, 2 H), 7.42 (t, J = 7.3 Hz, 1 H), 7.37 (t, 3J = 7.5
Hz, 2 H), 6.93 (d, 3J = 9.0 Hz, 2 H), 6.83 (d, 1J = 531.3 Hz, 1 H), 4.11
(t, 3J = 6.2 Hz, 2 H), 3.66−3.48 (m, 3 H), 3.43−3.34 (m, 2 H), 3.07−
2.87 (m, 2 H), 2.11−1.94 (m, 2 H), 1.96−1.84 (m, 1 H), 1.80 (dt, 3J =
6.3 Hz, 3J = 6.3 Hz, 2 H), 1.46 (dt, 3J = 14.9 Hz, 3J = 3.9 Hz, 2 H); 13
C
(9H-Fluoren-9-yl)methyl (2-Hydroxyethyl)carbamate (7). 9-
Fluorenylmethoxycarbonyl chloride (Fmoc-Cl, 1.55 g, 6.00 mmol) was
added to a solution of 2-aminoethanol (0.330 g, 5.40 mmol) in 10%
aqueous Na2CO3 (50 mL) and the mixture stirred for 2 h at RT. The
reaction mixture was extracted with ethyl acetate (3 × 50 mL). The
organic phases were combined, washed with aqueous HCl (1 M, 2 ×
50 mL) and brine (1 × 50 mL), and dried over MgSO4. Concentration
in vacuo and purification by column chromatography (silica gel, 5:1
ethyl acetate/hexane) gave 7 as a white solid (1.48 g, 5.22 mmol,
NMR (126 MHz, MeOH-d3, RT) δ 169.9, 163.2, 142.6, 133.6, 130.4,
1
130.2, 128.0, 127.5, 115.2, 66.6, 54.9 (d, JPC = 99 Hz), 45.5, 39.4 (d,
2JPC = 5 Hz), 36.2, 32.4, 30.1; 31P NMR (101 MHz, D2O/H2O, RT) δ
25.4 (pH 7.3); MS (ESI) m/z 430.2 [M − PHOOH]+, 496.2 [M +
H]+, 518.2 [M + Na]+, 991.1 [2M + H]+, 1013 [2M + Na]+, 1029.2
[2M + K]+; RP-HPLC tR = 10.5 min (10−90% in 30 min); HRMS
(ESI) m/z calcd for C22H31N3O6P32S 496.1671 [M + H]+, found
496.1665.
1-(Phenylsulfonamido)-2-{4-[2-(piperidin-4-yl)ethoxy]-
benzamido}ethylphosphonic Acid (2). A mixture of TFA, TIPS,
and water [5 mL, 95:2.5:2.5 (v/v/v) TFA/TIPS/H2O] was added to
17 (24.1 mg, 0.037 mmol) and the mixture stirred at RT for 16 h.
Purification by semipreparative RP-HPLC and lyophilization gave 2
1
97%): TLC Rf = 0.5 (5:1 ethyl acetate/hexane) (UV). H NMR and
13C NMR spectra were identical to those previously reported.45
(9H-Fluoren-9-yl)methyl (2-Oxoethyl)carbamate (8). IBX
(7.78 g, 27.8 mmol) was added to a solution of 7 (6.06 g, 21.4
mmol) in DMSO (20 mL) and the mixture stirred at RT for 16 h.17
DCM (1 L) was added to the reaction mixture, and the resulting white
1
(15.0 mg, 0.029 mmol, 79%) as a white solid: H NMR (500 MHz,
MeOH-d3, RT) δ 8.47 (br s, 1 H), 8.18 (br s, 1 H), 8.12−8.05 (m, 1
878
dx.doi.org/10.1021/jm2013826 | J. Med. Chem. 2012, 55, 871−882

Journal of Medicinal Chemistry
Article
suspension was stirred for 0.5 h, prior to filtration through Celite. The
organic layer was washed with aqueous saturated Na2CO3 (3 × 300
mL) and brine (2 × 300 mL). Each extraction was followed by
filtration through Celite, if necessary. Drying over MgSO4 and
concentration in vacuo resulted in a yellow crude product, which was
purified by column chromatography (silica gel, 5:1 ethyl acetate/
hexane) to give 8 (4.81 g, 17.1 mmol, 80%) as a white solid: TLC Rf =
0.8 (5:1 ethyl acetate/hexane) (UV). 1H NMR and 13C NMR spectra
were identical to those previously reported.45
(9H-Fluoren-9-yl)methyl [2-(Hydroxyimino)ethyl]carbamate
(9). DIEA (15.3 mL, 89.7 mmol) was added to a mixture of 8 (4.21 g,
15.0 mmol) and hydroxylammonium chloride (3.12 g, 44.9 mmol) in
DCM (dry, 40 mL) and the mixture stirred at RT for 16 h.18
Concentration in vacuo and purification by column chromatography
(silica gel, 3:1 ethyl acetate/hexane) gave 9 (2.45 g, 8.27 mmol, 55%)
as a slightly yellow colored solid.
2-{N-[(9H-Fluoren-9-yl)methoxy]carbonylamino}-1-aminoe-
thylphosphinic Acid (10). Commercial 60 wt % aqueous phosphinic
acid (40.0 g, 364 mmol) was lyophilized to obtain anhydrous
crystalline H3PO2, which was subsequently added to a solution of 9
(2.10 g, 7.09 mmol) in methanol (dry, 100 mL). The reaction mixture
was heated to reflux for 4 h and then concentrated in vacuo.18,19 The
residue was dissolved in aqueous HCl (3 M, 200 mL) and washed with
diethyl ether (3 × 100 mL). The pH was adjusted to 1.5 by addition of
solid Na2CO3. The resulting precipitate was isolated and purified by
RP-HPLC to give 10 (1.25 g, 3.61 mmol, 51%) as a colorless solid.
2-{N-[(9H-Fluoren-9-yl)methoxy]carbonylamino}-1-
(phenylsulfonamido)ethylphosphinic Acid (11). Benzenesulfonyl
chloride (0.507 mL, 3.97 mmol) was added to a solution of 10 (0.310
g, 0.895 mmol) in dioxane and aqueous Na2CO3 [50 mL, 1:1 dioxane/
aqueous Na2CO3 (10 wt %)] and stirred at RT for 3 h. The solvent
was removed in vacuo and the residue dissolved in water (100 mL).
The aqueous phase was acidified (pH 1) by addition of aqueous HCl
(3 M) and extracted with ethyl acetate (3 × 50 mL). The combined
organic phases were washed with brine (3 × 50 mL) and dried over
Na2SO4. Concentration in vacuo and lyophilization from dioxane gave
11 (0.382 g, 0.785 mmol, 88%) as a white solid.
2-Amino-1-(phenylsulfonamido)ethylphosphinic Acid (12).
A solution of piperidine (20%) in DMF (v/v, 50 mL) was added to 11
(0.672 g, 1.38 mmol) and the mixture stirred at RT for 2 h.
Concentration in vacuo and purification by RP-HPLC and
lyophilization gave 12 (0.253 g, 0.96 mmol, 69%) as a white solid.
2-[N-(Benzyloxycarbonyl)piperidin-4-yl]ethanol (13). Benzyl
chloroformate (Cbz-Cl, 6.05 mL, 42.38 mmol) was added to a solution
of 2-(piperidin-4-yl)ethanol (5.00 g, 38.7 mmol) in a dioxane/aqueous
10% Na2CO3 mixture (1:1, 250 mL) and the mixture stirred at RT for
1 h. The reaction mixture was concentrated in vacuo and the residue
dissolved with ethyl acetate (100 mL). The organic phase was washed
with aqueous saturated NaHCO3 (2 × 50 mL), aqueous HCl (1 M, 2
× 50 mL), and brine (1 × 50 mL). Drying over MgSO4 was followed
by column chromatography (silica gel, 5:1 ethyl acetate/hexane) to
give 13 (7.46 g, 28.2 mmol, 73%) as a colorless liquid.
was dissolved in water (100 mL) and acidified (pH 1) with HCl (12
M, 10 mL). A white precipitate formed, which was extracted with ethyl
acetate (3 × 100 mL). The combined organic phases were washed
with brine (1 × 100 mL) and dried over MgSO4. Concentration in
vacuo and lyophilization from dioxane gave 15 (0.47 g, 1.23 mmol,
97%) as a white solid.15
2-({4-[2-N-(Benzyloxycarbonyl)piperidin-4-yl]ethoxy}-
benzamido)-1-(phenylsulfonamido)ethylphosphinic Acid (16).
HATU (168 mg, 0.442 mmol), 15 (170 mg, 0.443 mmol), and DIEA
(452 μL, 2.66 mmol) were dissolved in DMF (dry, 10 mL), and the
mixture was stirred at RT for 15 min. A solution of 12 (117 mg, 0.443
mmol) in DMF (dry, 3 mL) was added and the mixture stirred at RT
for 16 h. Concentration in vacuo and purification by RP-HPLC gave
16 (234 mg, 0.372 mmol, 84%) as a white solid.
2-({4-[2-N-(Benzyloxycarbonyl)piperidin-4-yl]ethoxy}-
benzamido)-1-(phenylsulfonamido)ethylphosphonic Acid
(17). N,O-Bis(trimethylsilyl)acetamide (BSA, 98 μL, 0.400 mmol)
was added to a solution of 16 (10.1 mg, 0.016 mmol) in DCM (5 mL)
and the mixture stirred at RT for 1 h under an ambient atmosphere.
Concentration in vacuo, purification by semipreparative HPLC, and
lyophilization gave 17 (9.68 mg, 0.015 mmol, 93%) as a white solid.
Methyl 2-(4-{2-[1-(Benzyloxycarbonyl)piperidin-4-yl]-
ethoxy}benzamido)-1-(phenylsulfonamido)ethylphosphonate
(18). EDC·HCl (25.8 mg, 0.134 mmol) and DMAP (8.3 mg, 0.068
mmol) were added to a solution of 17 (21.7 mg, 0.034 mmol) in
methanol (dry, 5 mL), and the mixture was stirred at RT for 2 h.
Concentration in vacuo and purification via semipreparative RP-HPLC
gave 18 (17.4 mg, 0.026 mmol, 78%) as a white solid.
Integrin Binding Assay (Fibrinogen−αIIbβ3 Assay). The
inhibiting activity of the integrin antagonists was determined in a
solid-phase binding assay using coated extracellular matrix protein
fibrinogen and soluble αIIbβ3 integrin.26 The assay was based on a
previously reported method with some modifications.27 Flat-bottom
96-well ELISA plates (BRAND, Wertheim, Germany) were coated
overnight at 4 °C with 100 μL of 10 μg/mL fibrinogen per well
(Calbiochem, Darmstadt, Germany) in carbonate buffer [15 mM
Na2CO3 and 35 mM NaHCO3 (pH 9.6)]. Wells were then washed
three times with PBST buffer [137 mM NaCl, 2.7 mM KCl, 10 mM
Na2HPO4, 2 mM KH2PO4, and 0.01% Tween 20 (pH 7.4)] and
blocked for 1 h at room temperature with 150 μL of TSB buffer [20
mM Tris-HCl, 150 mM NaCl, 1 mM CaCl2, 1 mM MgCl2, 1 mM
MnCl2 (pH 7.5), and 1% BSA] per well. After being washed three
times with PBST, equal amounts of controls (tirofiban, Sigma-Aldrich)
or test compounds were mixed with 5.0 μg/mL human integrin αIIbβ3
(Enzyme Research Laboratory, Swansea, U.K.), resulting in a final TSB
buffer dilution of 0.00013 to 10 μM for the inhibitors and 2.5 μg/mL
for integrin αIIbβ3; 100 μL of these solutions was incubated per well
for 1 h at room temperature. The plate was washed three times with
PBST buffer, and 100 μL of 2.0 μg/mL primary antibody (mouse anti-
human CD41b, BD Biosciences, Heidelberg, Germany) per well was
added to the plate. After incubation for 1 h at room temperature, the
plate was washed three times with PBST, and 100 μL of 1.0 μg/mL
secondary peroxidase-labeled antibody (anti-mouse IgG-POD, Sigma-
Aldrich) per well was added to the plate and the plate incubated for 1
h at room temperature. After the plate had been washed three times
with PBST, the plate was developed by addition of 50 μL of
SeramunBlau fast (Seramun Diagnostic GmbH, Heidesee, Germany)
per well and incubated for 5 min at room temperature. The reaction
was stopped with 50 μL of 3 M H2SO4 per well, and the absorbance
was measured at 450 nm with a plate reader (POLARstar Galaxy,
BMG Labtechnologies). Each compound concentration was tested in
duplicate, and the resulting inhibition curves were analyzed using
OriginPro version 7.5G; the inflection point describes the IC50 value.
Each plate contained tirofiban as an internal control.
Methyl 4-[2-N-(Benzyloxycarbonyl)piperidin-4-ylethyloxy]-
benzoate (14). 13 (1.17 g, 4.43 mmol) was added to a solution of
methyl 4-hydroxybenzoate (0.62 g, 4.08 mmol) and tributylphosphine
(1.31 mL, 5.25 mmol) in THF (dry, 40 mL) at 0 °C under an argon
atmosphere. A solution of 1,1-(azodicarbonyl)dipiperidine (ADDP,
1.32 g, 5.23 mmol) in THF (dry, 15 mL) was added within 5 h by the
help of a syringe pump, and the reaction mixture was stirred at RT for
16 h.20,21 The white precipitate was removed by filtration and
destroyed. The filtrate was concentrated in vacuo and the residue
dissolved in ethyl acetate (100 mL). The organic phase was washed
with saturated aqueous Na2CO3 (3 × 50 mL), dried over MgSO4, and
concentrated in vacuo. Purification by column chromatography (silica
gel, 1:2 ethyl acetate/hexane) and crystallization from methanol gave
14 (1.39 g, 3.50 mmol, 86%) as a white solid.15
In Vitro Permeability Study (Caco-2 test). Growth and
Maintenance of Cells. Caco-2 cells were obtained from ATCC
(Manassas, VA) and then grown in 75 cm2 flasks with approximately
0.5 × 106 cells/flask at 37 °C in a 5% CO2 atmosphere at a relative
humidity of 95%. The culture growth medium consisted of Dulbecco’s
modified Eagle's medium (DMEM) supplemented with 10% heat-
4-[2-N-(Benzyloxycarbonyl)piperidin-4-ylethyloxy]benzoic
Acid (15). Aqueous NaOH (1 M, 100 mL) was added to a solution of
14 (0.50 g, 1.26 mmol) in ethanol (100 mL) and stirred at RT for 16
h. The reaction mixture was concentrated in vacuo, and the residue
879
dx.doi.org/10.1021/jm2013826 | J. Med. Chem. 2012, 55, 871−882

Journal of Medicinal Chemistry
Article
inactivated fetal bovine serum (FBS), 1% nonessential amino acids, 2
mM L-glutamine, 2 mM sodium pyruvate, and a 2 mM penicillin/
streptomycin solution.
calculated using the parm99 Amber force field48,49 through apbs,
which is an adaptive Poisson−Boltzmann solver.50
Docking Simulations. Redocking Experiment. The reliability of
AutoDock for this system was assessed through the redocking of the
αIIbβ3 cocrystallized ligand, tirofiban (PDB entry 2vdm). The X-ray
binding conformation of tirofiban has been clearly predicted by
AutoDock among the poses with the best scoring function values
(Figure 3A). One may notice that for tirofiban the (S)-enantiomer is
the bioactive one, while for the phosphorus-containing compounds,
because of the change in priority according to the Cahn−Ingold−
Prelog rules, the (R)-enantiomers are the bioactive ones. Molecular
docking calculations for tirofiban and compounds 1−4 were
conducted using the three-dimensional X-ray structure of αIIbβ3 in
the apo form (PDB entry 2vdm) through AutoDock (version 4.0).36,37
The apo form of αIIbβ3 was obtained via removal of tirofiban from the
X-ray complex.
Preparation of Cells. For transport studies, cells in a passage range
of 52−60 were seeded at a density of 25 × 105 cells/cm2 on untreated
culture inserts of a polycarbonate membrane with 0.4 μm pores and a
surface area of 1.1 cm2. The culture inserts containing the Caco-2
monolayer were placed in 24 transwell plates (12 mm, Costar). The
culture medium was changed every other day. Transport studies were
performed 21−23 days after seeding, when the cells were fully
differentiated and the TEER values were stable (300−500 Ω cm2).
Caco-2 Assay. The transport study (apical to basolateral, A to B)
was initiated by removal of medium from both sides of the monolayer
and replacement with apical buffer (600 μL) and basolateral buffer
(1500 μL), both warmed to 37 °C. The cells were incubated for 30
min at 37 °C with shaking (100 cycles/min). After the incubation
period, the buffers were removed and replaced with 1500 μL of
basolateral buffer at the basolateral side. Test solutions were preheated
to 37 °C and added (600 μL) to the apical side of the monolayer; 50
μL samples were taken from the apical side immediately at the
beginning of the experiment, resulting in an apical volume of 550 μL
during the experiment. For the period of the experiment, cells were
kept at 37 °C with shaking. At predetermined times (30, 60, 90, 120,
and 150 min), 200 μL samples were taken from the basolateral side
and replaced with the same volume of fresh basolateral buffer to
maintain a constant volume. A mass balance was performed for each
tested compound to detect instability and/or nonspecific binding of
the peptides. For the basolateral to apical study (B to A), compounds
were placed in the basolateral chamber, followed by sampling of the
apical side, in the same manner used for the A to B protocol.
Blood Samples. Venous blood was collected from a healthy
volunteer who had refrained from taking any medication affecting
platelet function for the two preceding weeks. Blood was drawn by
peripheral venipuncture into 4.5 mL plastic tubes containing
recombinant hirudin as an anticoagulant (specified final concentration
of 25 μg/mL of blood). Measurements were performed 0.5−4 h after
venipuncture.
Platelet Aggregation Assay. Platelet aggregation in hirudin-
anticoagulated whole blood with thrombin receptor-activating peptide
(TRAP-6, final concentration of 33 μM) as an activator was measured
using an impedance-based multiple-electrode platelet aggregometer
(Multiplate, Dynabyte Informationssysteme GmbH, Munich, Ger-
many)28,29 according to the manufacturer’s instructions, i.e., at 37 °C,
minicuvettes with 175 μL of blood, 175 μL of isotonic saline, and 12
μL of TRAP-6 reagent (1 mM). The only modification was the use of
a serial saline dilution of integrin inhibitors instead of pure saline. The
increase in electrical impedance was recorded for 6 min and
transformed into arbitrary aggregation units, and the area under the
curve (AUC) was calculated. Reported AUC values represent the
average from two electrode pairs per cuvette. Inhibition curves were
analyzed using OriginPro version 7.5G; the inflection point indicates
the half-maximal effective concentration (EC50).
Ligand Setup. The structures of the inhibitors were first generated
using the PRODRG server.51 Then the ligands and the protein were
charged using the Gasteiger partial charge52 and converted to
AutoDock format files using AutoDockTools (ADT 1.5.4).
Docking Setup. The docking area was defined by a box, centered
approximately on the center of mass of tirofiban cocrystallized with the
protein. Grid points (60 × 60 × 60) with 0.375 Å spacing were
calculated around the docking area for all the ligand atom types using
AutoGrid4. For each ligand, 100 separate docking calculations were
performed. Each docking calculation consisted of 2.5 × 106 energy
evaluations using the Lamarckian genetic algorithm local search
(GALS) method. Otherwise, default docking parameters were applied.
The docking conformations were clustered on the basis of the root-
mean-square deviation (rmsd of 2 Å) calculated for the Cartesian
coordinates of the ligand atoms and then were ranked on the basis of
AutoDock scoring function. The binding mode figures were generated
using PyMOL (http://www.pymol.org), while the molecular electro-
static potential surfaces were rendered using GaussView.
ASSOCIATED CONTENT
* Supporting Information
■
S
Analytical data of compounds 7−18, HPLC enantiomer
separation, 31P NMR titration of compounds 1−4, docking of
molecular probes (R)-3 and (R)-4, docking of (S)-enantiomers,
and supporting figures. This material is available free of charge
via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*Institute for Advanced Study and Center of Integrated Protein
Science, Department Chemie, Technische Universitat Mun-
chen, Lichtenbergstrasse 4, 85747 Garching, Germany. Phone:
+49 (0) 89 289 13300. Fax: +49 (0) 89 289 13210. E-mail:
kessler@tum.de.
■
̈
̈
Protonation State Calculations. The estimation of pKa values of
compounds 1−4 was conducted using the calculator plugin Marvin
5.3, 2010, from ChemAxon (http://www.chemaxon.com). The
estimation is computed through an algorithm that uses the empirically
calculated atomic charges for each protonation state of a subset of
molecules.46 Each atom of the query molecule is identified in one of
the atom's subsets, and via the algorithm, the pKa is finally calculated.
The concentration of the different microspecies formed by a molecule
at a given pH is predicted using the distribution coefficient (D),
calculated using the previously computed pKa values.47
ACKNOWLEDGMENTS
■
We gratefully acknowledge financial support from the Interna-
tional Graduate School of Science and Engineering (IGSSE)
and from the Studienstiftung des deutschen Volkes, Werner
Spahl for recording high-resolution mass spectra, Renate Reher
for performing the platelet aggregation measurements, and
Timo Huber for initial support.
Molecular Electrostatic Potential Calculations. For tirofiban (6)
and compounds 1−4, the electrostatic potential was calculated by
means of Gaussian03 and mapped onto the electron density surface for
each compound. The isovalue of 0.0004 electron/Bhor3 was chosen
for the definition of the density surface, while the electrostatic
potential was computed at the Hartree−Fock level of theory using the
6-31G* basis set with a scale of −294.178 (red) to 294.178 (blue)
KbT/ec. The electrostatic potential of the αIIbβ3 receptor was
ABBREVIATIONS
■
ADAMs, disintegrin and metalloprotease; AUC, area under the
curve; CIP, Cahn−Ingold−Prelog; DMAP, 4-(dimethylamino)-
pyridine; ECM, extracellular matrix; EDC, N-[3-
(dimethylamino)propyl]-N′-ethylcarbodiimide; ELISA, en-
zyme-linked immunosorbent assay; FMD, foot-and-mouth
disease; HATU, O-(7-azabenzotriazol-1-yl)-N,N,N′,N′-tetrame-
880
dx.doi.org/10.1021/jm2013826 | J. Med. Chem. 2012, 55, 871−882

Journal of Medicinal Chemistry
Article
(19) Zhukov, Y. N.; Khomutov, A. R.; Osipova, T. I.; Khomutov, R.
M. Synthesis of phosphinic analogs of sulfur-containing amino acids.
Russ. Chem. Bull. 1999, 48, 1348−1351.
thyluronium hexafluorophosphate; IBX, o-iodoxybenzoic acid;
MIDAS, metal ion-dependent adhesion site; RGD, arginine-
glycine-aspartic acid
(20) Heckmann, D.; Meyer, A.; Laufer, B.; Zahn, G.; Stragies, R.;
Kessler, H. Rational design of highly active and selective ligands for the
α5β1 integrin receptor. ChemBioChem 2008, 9, 1397−1407.
REFERENCES
■
́
(21) Marugan, J. J.; Manthey, C.; Anaclerio, B.; Lafrance, L.; Lu, T.;
(1) Meyer, A.; Auernheimer, J.; Modlinger, A.; Kessler, H. Targeting
RGD recognizing integrins: Drug development, biomaterial research,
tumor imaging and targeting. Curr. Pharm. Des. 2006, 12, 2723−2747.
(2) Mousa, S. A. Anti-integrin as novel drug-discovery targets:
Potential therapeutic and diagnostic implications. Curr. Opin. Chem.
Biol. 2002, 6, 534−541.
Markotan, T.; Leonard, K. A.; Crysler, C.; Eisennagel, S.; Dasgupta,
M.; Tomczuk, B. Design, synthesis, and biological evaluation of novel
potent and selective αvβ3/αvβ5 integrin dual inhibitors with improved
bioavailability. Selection of the molecular core. J. Med. Chem. 2005, 48,
926−934.
(22) Hoffmann, C. V.; Pell, R.; Lammerhofer, M.; Lindner, W.
̈
(3) Humphries, M. J. Integrin structure. Biochem. Soc. Trans. 2000,
28, 311−339.
Synergistic effects on enantioselectivity of zwitterionic chiral stationary
phases for separations of chiral acids, bases, and amino acids by HPLC.
Anal. Chem. 2008, 80, 8780−8789.
(4) Arnaout, M. A.; Mahalingam, B.; Xiong, J. P. Integrin structure,
allostery, and bidirectional signaling. Annu. Rev. Cell Dev. Biol. 2005,
21, 381−410.
(23) Hoffmann, C. V.; Reischl, R.; Maier, N. M.; Lammerhofer, M.;
̈
Lindner, W. Stationary phase-related investigations of quinine-based
zwitterionic chiral stationary phases operated in anion-, cation-, and
zwitterion-exchange modes. J. Chromatogr., A 2009, 1216, 1147−1156.
(24) Maier, N. M.; Schefzick, S.; Lombardo, G. M.; Feliz, M.;
Rissanen, K.; Lindner, W.; Lipkowitz, K. B. Elucidation of the chiral
recognition mechanism of cinchona alkaloid carbamate-type receptors
for 3,5-dinitrobenzoyl amino acids. J. Am. Chem. Soc. 2002, 124, 8611−
8629.
(5) Hynes, R. O. Integrins: Bidirectional, allosteric signaling
machines. Cell 2002, 110, 673−687.
(6) Humphries, J. D.; Byron, A.; Humphries, M. J. Integrin ligands at
a glance. J. Cell Sci. 2006, 119, 3901−3903.
(7) Ruoslahti, E. RGD and other recognition sequences for integrins.
Annu. Rev. Cell Dev. Biol. 1996, 12, 697−715.
(8) Heckmann, D.; Laufer, B.; Marinelli, L.; Limongelli, V.;
Novellino, E.; Zahn, G.; Stragies, R.; Kessler, H. Breaking the
dogma of the metal-coordinating carboxylate group in integrin ligands:
Introducing hydroxamic acids to the MIDAS to tune potency and
selectivity. Angew. Chem., Int. Ed. 2009, 48, 4436−4440.
(9) Cox, D.; Brennan, M.; Moran, N. Integrins as therapeutic targets:
Lessons and opportunities. Nat. Rev. Drug Discovery 2010, 9, 804−820.
(10) Mas-Moruno, C.; Rechenmacher, F.; Kessler, H. Cilengitide:
The first anti-angiogenic small molecule drug candidate. Design,
synthesis and clinical evaluation. Anti-Cancer Agents Med. Chem. 2011,
10, 753−768.
(11) Shimaoka, M.; Springer, T. A. Therapeutic antagonists and
conformational regulation of integrin function. Nat. Rev. Drug
Discovery 2003, 2, 703−716.
(12) Hartman, G. D.; Egbertson, M. S.; Halczenko, W.; Laswell, W.
L.; Duggan, M. E.; Smith, R. L.; Naylor, A. M.; Manno, P. D.; Lynch,
R. J.; Zhang, G.; Chang, C. T.-C.; Gould, R. J. Non-peptide fibrinogen
receptor antagonists. 1. Discovery and design of exosite inhibitors. J.
Med. Chem. 1992, 35, 4640−4642.
(13) Topol, E. J.; Moliterno, D. J.; Herrmann, H. C.; Powers, E. R.;
Grines, C. L.; Cohen, D. J.; Cohen, E. A.; Bertrand, M.; Neumann, F.
J.; Stone, G. W.; DiBattiste, P. M.; Demopoulos, L. Comparison of two
platelet glycoprotein IIb/IIIa inhibitors, tirofiban and abciximab, for
the prevention of ischemic events with percutaneous coronary
revascularization. N. Engl. J. Med. 2001, 344, 1888−1894.
(14) Cannon, C. P.; Weintraub, W. S.; Demopoulos, L. A.; Vicari, R.;
Frey, M. J.; Lakkis, N.; Neumann, F. J.; Robertson, D. H.; DeLucca, P.
T.; DiBattiste, P. M.; Gibson, C. M.; Braunwald, E. Comparison of
early invasive and conservative strategies in patients with unstable
coronary syndromes treated with the glycoprotein IIb/IIIa inhibitor
tirofiban. N. Engl. J. Med. 2001, 344, 1879−1887.
(25) Lammerhofer, M.; Hebenstreit, D.; Gavioli, E.; Lindner, W.;
̈
Mucha, A.; Kafarski, P.; Wieczorek, P. High-performance liquid
chromatographic enantiomer separation and determination of absolute
configurations of phosphinic acid analogues of dipeptides and their α-
aminophosphinic acid precursors. Tetrahedron: Asymmetry 2003, 14,
2557−2565.
(26) Mas-Moruno, C.; Beck, J. G.; Doedens, L.; Frank, A. O.;
Marinelli, L.; Cosconati, S.; Novellino, E.; Kessler, H. Increasing αvβ3
selectivity of the anti-angiogenic drug cilengitide by N-methylation.
Angew. Chem., Int. Ed. 2011, 50, 9496−9500.
(27) Chatterjee, J.; Ovadia, O.; Zahn, G.; Marinelli, L.; Hoffman, A.;
Gilon, C.; Kessler, H. Multiple N-methylation by a designed approach
enhances receptor selectivity. J. Med. Chem. 2007, 50, 5878−5881.
́
(28) Toth, O.; Calatzis, A.; Penz, S.; Losonczy, H.; Siess, W. Multiple
electrode aggregometry: A new device to measure platelet aggregation
in whole blood. Thromb. Haemostasis 2006, 96, 781−788.
(29) Halimeh, S.; Angelis, G.; Sander, A.; Edelbusch, C.; Rott, H.;
Thedieck, S.; Mesters, R.; Schlegel, N.; Nowak-Gottl, U. Multiplate
whole blood impedance point of care aggregometry: Preliminary
reference values in healthy infants, children and adolescents. Klin.
Paediatr. 2010, 222, 158−163.
(30) Artursson, P.; Karlsson, J. Correlation between oral drug
absorption in humans and apparent drug permeability coefficients in
human intestinal epithelial (Caco-2) cells. Biochem. Biophys. Res.
Commun. 1991, 175, 880−885.
(31) Kelley, J. L.; McLean, E. W.; Crouch, R. C.; Averett, D. R.;
Tuttle, J. V. [[(Guaninylalkyl)phosphinico]methyl]phosphonic acids.
Multisubstrate analogue inhibitors of human erythrocyte purine
nucleoside phosphorylase. J. Med. Chem. 1995, 38, 1005−1014.
(32) Manzenrieder, F.; Frank, A. O.; Kessler, H. Phosphorus NMR
spectroscopy as a versatile tool for compound library screening. Angew.
Chem., Int. Ed. 2008, 47, 2608−2611.
(33) Stirtan, W. G.; Withers, S. G. Phosphonate and α-
fluorophosphonate analogue probes of the ionization state of pyridoxal
5′-phosphate (PLP) in glycogen phosphorylase. Biochemistry (Moscow,
Russ. Fed.) 1996, 35, 15057−15064.
(34) Selvam, C.; Goudet, C.; Oueslati, N.; Pin, J. P.; Acher, F. C. L-
(+)-2-Amino-4-thiophosphonobutyric acid (L-thioAP4), a new potent
agonist of group III metabotropic glutamate receptors: Increased distal
acidity affords enhanced potency. J. Med. Chem. 2007, 50, 4656−4664.
(35) Castellino, S.; Leo, G. C.; Sammons, R. D.; Sikorski, J. A. 31P,
15N, and 13C NMR of glyphosate: Comparison of pH titrations to the
herbicidal dead-end complex with 5-enolpyruvoylshikimate-3-phos-
(15) Duggan, M. E.; Duong, L. T.; Fisher, J. E.; Hamill, T. G.;
Hoffman, W. F.; Huff, J. R.; Ihle, N. C.; Leu, C. T.; Nagy, R. M.;
Perkins, J. J.; Rodan, S. B.; Wesolowski, G.; Whitman, D. B.; Zartman,
A. E.; Rodan, G. A.; Hartman, G. D. Nonpeptide αvβ3 antagonists. 1.
Transformation of a potent, integrin-selective αIIbβ3 antagonist into a
potent αvβ3 antagonist. J. Med. Chem. 2000, 43, 3736−3745.
(16) Frigerio, M.; Santagostino, M.; Sputore, S. A user-friendly entry
to 2-iodoxybenzoic acid (IBX). J. Org. Chem. 1999, 64, 4537−4538.
(17) Frigerio, M.; Santagostino, M. A mild oxidizing reagent for
alcohols and 1,2-diols: o-Iodoxybenzoic acid (IBX) in DMSO.
Tetrahedron Lett. 1994, 35, 8019−8022.
(18) Liboska, R.; Pícha, J.; Hanclova,
Jiracek, J. Synthesis of methionine- and norleucine-derived phosphi-
nopeptides. Tetrahedron Lett. 2008, 49, 5629−5631.
́
́
I.; Budesínsky, M.; Sanda, M.;
́
881
dx.doi.org/10.1021/jm2013826 | J. Med. Chem. 2012, 55, 871−882

Journal of Medicinal Chemistry
Article
phate synthase. Biochemistry (Moscow, Russ. Fed.) 1989, 28, 3856−
3868.
(36) Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.; Hart,
W. E.; Belew, R. K.; Olson, A. J. Automated docking using a
Lamarckian genetic algorithm and an empirical binding free energy
function. J. Comput. Chem. 1998, 19, 1639−1662.
(37) Huey, R.; Morris, G. M.; Olson, A. J.; Goodsell, D. S. A
semiempirical free energy force field with charge-based desolvation. J.
Comput. Chem. 2007, 28, 1145−1152.
(38) Jones, G.; Willett, P.; Glen, R. C.; Leach, A. R.; Taylor, R.
Development and validation of a genetic algorithm for flexible docking.
J. Mol. Biol. 1997, 267, 727−748.
(39) Rarey, M.; Kramer, B.; Lengauer, T.; Klebe, G. A fast flexible
docking method using an incremental construction algorithm. J. Mol.
Biol. 1996, 261, 470−489.
(40) Gohlke, H.; Hendlich, M.; Klebe, G. Knowledge-based scoring
function to predict protein-ligand interactions. J. Mol. Biol. 2000, 295,
337−356.
(41) Oelschlaeger, P.; Klahn, M.; Beard, W. A.; Wilson, S. H.;
Warshel, A. Magnesium-cationic dummy atom molecules enhance
representation of DNA polymerase β in molecular dynamics
simulations: Improved accuracy in studies of structural features and
mutational effects. J. Mol. Biol. 2007, 366, 687−701.
(42) Rucker, R.; Oelschlaeger, P.; Warshel, A. A binding free energy
decomposition approach for accurate calculations of the fidelity of
DNA polymerases. Proteins 2010, 78, 671−680.
(43) Xiang, Y.; Oelschlaeger, P.; Florian, J.; Goodman, M. F.;
Warshel, A. Simulating the effect of DNA polymerase mutations on
transition-state energetics and fidelity: Evaluating amino acid group
contribution and allosteric coupling for ionized residues in human Pol
β. Biochemistry (Moscow, Russ. Fed.) 2006, 45, 7036−7048.
(44) Gottlieb, H. E.; Kotlyar, V.; Nudelman, A. NMR chemical shifts
of common laboratory solvents as trace impurities. J. Org. Chem. 1997,
62, 7512−7515.
(45) Porcheddu, A.; Giacomelli, G.; Piredda, I.; Carta, M.; Nieddu, G.
A practical and efficient approach to PNA monomers compatible with
Fmoc-mediated solid-phase synthesis protocols. Eur. J. Org. Chem.
2008, 2008, 5786−5797.
(46) Dixon, S. L.; Jurs, P. C. Estimation of pKa for organic oxyacids
using calculated atomic charges. J. Comput. Chem. 1993, 14, 1460−
1467.
(47) Csizmadia, F.; Tsantili-Kakoulidou, A.; Panderi, I.; Darvas, F.
Prediction of distribution coefficient from structure. 1. Estimation
method. J. Pharm. Sci. 1997, 86, 865−871.
(48) Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K.
M.; Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.; Caldwell, J. W.;
Kollman, P. A. A second generation force field for the simulation of
proteins, nucleic acids, and organic molecules. J. Am. Chem. Soc. 1995,
117, 5179−5197.
(49) Wang, J.; Cieplak, P.; Kollman, P. A. How well does a restrained
electrostatic potential (RESP) model perform in calculating conforma-
tional energies of organic and biological molecules? J. Comput. Chem.
2000, 21, 1049−1074.
(50) Baker, N. A.; Sept, D.; Joseph, S.; Holst, M. J.; McCammon, J.
A. Electrostatics of nanosystems: Application to microtubules and the
ribosome. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 10037−10041.
(51) Schuttelkopf, A. W.; van Aalten, D. M. F. PRODRG: A tool for
̈
high-throughput crystallography of protein-ligand complexes. Acta
Crystallogr. 2004, D60, 1355−1363.
(52) Gasteiger, J.; Marsili, M. Iterative partial equalization of orbital
electronegativity: A rapid access to atomic charges. Tetrahedron 1980,
36, 3219−3228.
882
dx.doi.org/10.1021/jm2013826 | J. Med. Chem. 2012, 55, 871−882