Design and synthesis of thrombin receptor-derived nonpeptide mimetics utilizing a piperazine scaffold

Article abstract of DOI:10.1016/S0968-0896(99)00017-6

Focal thrombus formation and vasoconstriction serve to defend vessels when vascular damage occurs, but may be detrimental when an atherosclerotic plaque is disrupted. Recently, the identification of the platelet thrombin receptor opened a new area in the development of agents that may selectively inhibit the effects of thrombin on cells, without affecting fibrin formation. In this regard, we have synthesized a number of 1,4-disubstituted piperazines which are designed to be analogues of thrombin receptor activating peptides (TRAP) and carry the pharmacophoric features of Phe and Arg residues present in the active pentapeptide SFLLR. These compounds were tested in the rat aorta relaxation assay and in platelet aggregation studies and their biological activity was consistent with a direct action on thrombin receptor. Furthermore, the structure-activity relationships confirmed the importance of Phe and Arg for receptor activation and the molecular modeling revealed an intriguing relationship between their amphipathic similarity with SFLLR and their biological activity. Copyright (C) 1999 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(99)00017-6

Bioorganic & Medicinal Chemistry 7 (1999) 1033±1041
Design and Synthesis of Thrombin Receptor-derived Nonpeptide
Mimetics Utilizing a Piperazine Sca€old
Kostas Alexopoulos,a,y Panagiotis Fatseas,a,b E® Melissari,b Demetrios Vlahakos,b
Julian Smith,c Thomas Mavromoustakos,d Mahmoud Saifeddine,e Graham Moore,e
Morley Hollenberge and John Matsoukasa,*
aDepartment of Chemistry, University of Patras, Patras, Greece 26110
bOnassis Cardiac Surgery Center, 356 Sygrou Ave., Athens, Greece 17674
cDepartment of Chemistry, University of Exeter, Stocker Road, Exeter, Devon, UK, EX4 4QD
dNational Hellenic Research Foundation, Institute of Organic and Pharmaceutical Chemistry, Vas. Constantinou 48 Ave.,
Athens, Greece 11635
eDepartment of Pharmacology and Therapeutics, University of Calgary, Calgary, Alberta, Canada T2N 4N1
Received 21 May 1998; accepted 6 October 1998
AbstractÐFocal thrombus formation and vasoconstriction serve to defend vessels when vascular damage occurs, but may be det-
rimental when an atherosclerotic plaque is disrupted. Recently, the identi®cation of the platelet thrombin receptor opened a new
area in the development of agents that may selectively inhibit the e€ects of thrombin on cells, without a€ecting ®brin formation. In
this regard, we have synthesized a number of 1,4-disubstituted piperazines which are designed to be analogues of thrombin receptor
activating peptides (TRAP) and carry the pharmacophoric features of Phe and Arg residues present in the active pentapeptide
SFLLR. These compounds were tested in the rat aorta relaxation assay and in platelet aggregation studies and their biological
activity was consistent with a direct action on thrombin receptor. Furthermore, the structure±activity relationships con®rmed the
importance of Phe and Arg for receptor activation and the molecular modeling revealed an intriguing relationship between their
amphipathic similarity with SFLLR and their biological activity. # 1999 Elsevier Science Ltd. All rights reserved.
Introduction
corresponding to at least the ®ve residues of the new N-
terminus (S42FLLR45) are e€ective in mimicking the
action of thrombin.9±12 Structure±activity relationships
(SAR) and alanine scan experiments have indicated that
Phe43 and Arg45 are the most important amino acids of
the receptor-derived peptide SFLLR for its activity in
smooth muscle13,14 and for its ability to aggregate pla-
telets.9,10 We have recently shown, using cyclic TRAP,
the importance of the Phe/Arg relative conformation
for activity in a smooth muscle contractile assay.14
Considering the above information and based on pre-
vious NMR and SAR experiments,15±17 we proceeded in
the synthesis of non-peptide analogues of the minimal
SFLLR sequence using piperazine as the template to
hold groups that can bear the pharmacophoric features
of Phe and Arg (Scheme 1). The synthesized compounds
were initially tested for their biological activity in the rat
aorta relaxation assay,12,13 and three representative
compounds with low, medium and high biological
activity (compounds 11, 10 and 13, respectively) were
further tested to evaluate their ability to a€ect in vitro
the aggregation of human platelets induced by either
thrombin or the linear thrombin receptor activating
amide SFLLR-NH2. This peptidomimetic research is
Thrombin is a serine protease that in addition to being a
key enzyme in thrombus formation, can activate a vari-
ety of cells such as platelets, endothelial cells, ®broblasts
and vascular smooth muscle.1±3 The identi®cation of the
platelet thrombin receptor and the demonstration that
is a member of the super-family of G-protein-coupled
receptors4,5 opened a new area in the development of
agents that may selectively inhibit the e€ects of throm-
bin on cells, without a€ecting ®brin formation. The
activation mechanism of target cells by thrombin
involves a proteolytic cleavage of the extracellular N-
terminal bond between Arg41 and Ser42 of the thrombin
receptor. The newly generated N-terminus unmasked by
thrombin-induced cleavage serves as a tethered ligand6
which binds intramolecularly to e€ect receptor activa-
tion.7,8 In support of this model, synthetic peptides
(thrombin receptor activating peptides (TRAP))
Key words: Thrombin; TRAP; mimetics; peptides; piperazine.
* Corresponding author. Fax: +3061-997-118.
Parts of this work are incorporated in the Ph.D. Dissertations of
Kostas Alexopoulos and Panagiotis Fatseas.
y
0968-0896/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(99)00017-6

1034
K. Alexopoulos et al. / Bioorg. Med. Chem. 7 (1999) 1033±1041
(2, 3, 4, and 5) was accomplished with tri¯uoroacetic
acid giving the free-amine salts 6, 7, 8, and 9. Guanyla-
tion of these amines took place according to the proce-
dure described by Bernatowicz.18 For example, the
reaction of pyrazole with cyanamide in p-dioxane gave
1H-pyrazole-1-carboxamide hydrochloride in high
yield. This compound served as a guanylation reagent
which reacted with the previous amines to give 10, 11,
12, and 13 as ®nal products.
Scheme 1. Nonpeptide TRAP derived mimetics incorporating piper-
azine.
Rat aorta relaxation assay
Of the compounds synthesized, compound 13, with an
EC50 of about 150 mM (Fig. 2) was clearly the most
active, followed by compound 10. Compounds 7, 8, and
9 showed minimal activity; and compounds 6, 11, and
12 appeared to be devoid of activity. When tested at
concentrations higher than 500 mM, compounds 6, 7, 8,
9, 11, and 12 appeared to show toxicity, in that the
subsequent contractile responsiveness of the tissues to
phenylephrine was compromised. This was not the case
for compounds 10 and 13.
growing rapidly in an e€ort to improve the metabolic
stability, duration of action and bioavailability of the
endogenous peptides.16
Results
Chemistry
The general methodology for the synthesis of 1-4 disub-
stituted piperazine derivatives is depicted in Scheme 2.
1-Phenylacetyl-piperazine 1 was synthesized by coupling
piperazine and phenylacetic acid with DCC and HOBt
as coupling reagents. Boc-protected amino acids (gly-
cine, b-alanine, g-aminobutyric and e-aminohexanoic
acid) were then incorporated separately in 1 aided by
the use of DCC and HOBt under basic conditions
(DIEA). The Boc-protecting group was preferred over
the Fmoc group because of the problems that the latter
causes through its removal with piperidine during the
cleavage step. Boc-deprotection of each Boc-derivative
Because thrombin induces endothelium-dependent, ni-
tric-oxide-mediated vascular relaxations through the
activation of the tethered ligand receptor on endothelial
cells, the mechanism of vasorelaxation induced by the
most active compound 13, was further examined. As
shown in Figure 1, the relaxant activity of compound 13
was, like that of acetylcholine, dependent upon the pre-
sence of an intact endothelium (Tracing A) and it was
not observed after the endothelium had been denuded,
although the ability of the same preparation to relax
in response to isoproterenol via an endothelium-
independent mechanism remained intact (Tracing C).
Moreover, in the same endothelium-containing pre-
paration, the relaxant activity of compound 13 was
sustained, while that of the thrombin receptor activating
pentapeptide, P5-NH2 was rather transient (Tracings A
Scheme 2. Synthetic procedure for 1,4-disubstituted piperazine deri-
vatives.
Figure 1. Relaxant action of compound 13 (300 mM) compared with
P5-NH2 and Ach.

K. Alexopoulos et al. / Bioorg. Med. Chem. 7 (1999) 1033±1041
1035
and B). As shown previously with acetylcholine,
thrombin and the thrombin receptor activating peptide
(P5-NH2), the compound 13-induced vasorelaxation
was completely blocked by the nitric oxide synthase
inhibitor, l-NAME (Tracing B, and data not shown).
Finally, the relaxant activity of compound 13 could not
be attributed to the partial metabolism of its guanidino
side-chain to nitric oxide, since l-arginine itself, which
could be similarly metabolized, failed to cause a relax-
ant response (Tracing A).
Platelet aggregation studies
Incubation of platelet-rich plasma with varying con-
centrations (0±4.0 mM) of the nonpeptide TRAP ana-
logues (compounds 10, 11, and 13) failed to produce
platelet aggregation, indicating that these molecules
were devoid from any thrombin-like agonistic activity
on human platelets. In contrast, when platelet-rich
plasma was preincubated with each compound for 4 min
prior to the addition of thrombin or SFLLR-NH2, their
ability to behave like both thrombin and SFLLR-NH2
antagonists was brought to light. As shown in Figure
3(A), compound 10 was able to signi®cantly inhibit
platelet aggregation when a low concentration of
thrombin (40.2 NIU/mL) was used, but failed to inhibit
platelet aggregation when higher concentrations
(>0.3 NIU/mL) of thrombin were used. Compound 11
inhibited thrombin-induced platelet aggregation only at
high concentration (2.0 mM, Fig. 3(B)). Interestingly,
compound 13 was capable to produce a signi®cant
and dose-dependent inhibition of the thrombin-induced
platelet aggregation even when relatively low con-
centrations of this compound (0.25 mM) were utilized
(Fig. 3(C)).
The IC50 for compound 10 was 0.796 mM for thrombin
concentration 0.2 NIU/mL and 1.833 mM in the case of
thrombin range 0.25±0.3 NIU/mL. For compound 11
the IC50 was 0.867 mM when thrombin concentration
was 0.2 NIU/mL whereas for thrombin range 0.25±
0.3 NIU/mL was 0.895 mM. For compound 13 the IC50
was found to be 0.466 mM for thrombin concentration
0.2 NIU/mL and 0.481 for thrombin range 0.25±
0.3 NIU/mL. Finally, all three compounds were shown
to signi®cantly inhibit the SFLLR-NH2-induced (Fig. 4),
but not the collagen-induced platelet aggregation.
Figure 3. E€ect of nonpeptide thrombin mimetics (compounds 10, 11,
and 13) on inhibition of thrombin-induced platelet aggregation in
human PRP. These data are representative of nine experiments giving
similar results (p<0.001).
Figure 4. E€ect of nonpeptide thrombin mimetics (compounds 10, 11,
and 13) on inhibition of SFLLR-NH2 (1 mM)-induced platelet aggre-
gation in human PRP. Data are expressed, as mean‹SEM of four
experiments (*p<0.0001 versus control).
Figure 2. Concentration-e€ect curves for compounds 8±13.

1036
K. Alexopoulos et al. / Bioorg. Med. Chem. 7 (1999) 1033±1041
Molecular Modeling
showed the best NO-mediated endothelial cell-
dependent relaxation among the compounds tested
(Fig. 2) indicating that the alkyl guanidine chain length
may also be important for optimizing the activity.
Because the rat aorta relaxation assay is a complex sys-
tem the possibility could not be denied that the TRAP
analogues tested here produced vasorelaxation by
mechanism(s) not involving the thrombin receptor.
Indeed, other 1,4-disubstituted piperazines have been
previously shown to stimulate a variety of receptors,
including the receptor for the platelet-activating
factor24,25 and the adhesion receptor.26
Theoretical calculations were performed on a Silicon
Graphics 4-D/35 model using the Quanta 3.3 version of
molecular simulation incorporated (MSI) program. The
constructed structures of all eight molecules were ®rst
minimized. Constraints on the minimized structures were
imposed in order to achieve a close proximity between
Phe and Arg (or -NH2 moieties). The superimposition
of pairs of structures was performed using rigid body ®t
to target method of molecular similarity software. Rigid
body ®tting translates and rotates working structures
(6±13) to minimize the RMS (root mean square) of the
®t to the target structure (SFLLR). Details of this
approach was given in our previous publications.14,19
The e€ects of thrombin on platelets are mediated by
thrombin receptors. Since in our experience and that of
other investigators27 rat platelets are poorly reactive to
thrombin and to SFLLR-NH2, we chose human plate-
lets to evaluate the ability of three representative com-
pounds (compounds 10, 11, and 13) to induce platelet
aggregation or inhibit the thrombin- or SFLLR-NH2-
induced platelet aggregation. Although none of these
compounds showed any agonistic activity when incu-
bated with PRP, they were all able to signi®cantly inhi-
bit the thrombin- or SFLLR-NH2-induced platelet
aggregation. Of note, compound 13 was clearly the most
e€ective, as it was the case in the rat aorta relaxation
assay. Taking into consideration that (a) the e€ects of
thrombin on platelets are mediated through its speci®c
receptor, (b) SFLLR-NH2 directly deals with thrombin
receptor, and (c) the compounds had no e€ect on col-
lagen-induced platelet aggregation, we may come to the
conclusion that the compounds tested with the platelet
aggregation studies exert their biological activity
through speci®c interactions with the thrombin receptor
rather than by inhibiting thrombin's enzymatic activity
or stimulating other type of receptor(s).28
Discussion
The identi®cation of the human thrombin receptor has
generated signi®cant interest in the development of
receptor antagonists that can inhibit the e€ects of
thrombin on cells without a€ecting its enzymatic activ-
ity on ®brin formation. Synthetic peptides of which the
®rst ®ve residues contain the tethered ligand domain
SFLLR have been found to be capable of causing the
cellular e€ects of thrombin. It is well established that
the introduction of non-peptide molecules into peptides
can dramatically a€ect receptor binding anity.20,21
The design of non-peptide mimetics of TRAP, such as
the ones described here, has been based on the current
knowledge about the interactions of the tethered ligand
and the thrombin receptor.14±16 The fact that the cyclic
conformer of SFLLR observed by NMR is probably the
bioactive species, along with the observation that cer-
tain cyclic derivatives are biologically active22 prompted
us to design molecules that mimic the cyclic conforma-
tion adopted by SFLLR. In this conformation the Phe
and Arg residues are in a close proximity on the same
side of the cyclic ring. Replacing the aliphatic residues
of the active peptide core with piperazine would be one
way of stabilizing and testing compounds able to adopt
the above cyclic conformation. Indeed, piperazine has
been widely used as a template in medicinal chemistry23
and this and other templates are currently under inves-
tigation in our laboratory for their ability to maintain
the pharmacophoric features of Phe and Arg in a sui-
table conformation. Such mimetics could be valuable in
the development of a lead structure useful in drug opti-
mization. The design of mimetics was assembled
according to Scheme 1. The pharmacophoric character-
istic of Phe was incorporated by coupling phenylacetic
acid to piperazine, while the pharmacophoric feature of
Arg was built in through a variety of guanidino linear
chains attached to the other end of piperazine. The dif-
ferent chain lengths could give us information about the
hydrophobic core between the guanidino and phenyl
groups through structure±activity relationship studies.
Several tentative reasons to explain the di€erence between
the thrombin-like agonistic e€ect of these compounds in
the rat aorta relaxation assay and the anti-thrombin
e€ect in the plasma aggregation assay can be proposed
here. Firstly, di€erent types of thrombin receptor may
exist in di€erent species. In this regard, human but not
rat and dog platelets are activated by SFLLR pep-
tides.27 Secondly, di€erent types of thrombin receptor(s)
and di€erent activating pathways may be stimulated in
platelets and in endothelial cells. Indeed, similar results
to our ®ndings have been reported for Mpa-peptides.29,30
Nevertheless, the existence of a molecule that could
both inhibit platelet aggregation and augment vaso-
relaxation would be of paramount clinical signi®cance
and it could potentially be used to prevent vascular re-
occlusion after thrombolytic therapy of acute myocardial
infarction or after coronary artery angioplasty.31
To further examine the stereoelectronic factors that
determine the biological activity of these compounds
the experimental data were coupled with theoretical
calculations. As was mentioned above, we have recently
proposed a cyclic model shown in Figure 5 for SFLLR
based on a combination of results obtained from biolo-
gical assays, NMR NOE spectroscopy and molecular
modeling.14,15,17 The main feature of this model is that
Most of the compounds showed moderate activity in the
rat aorta relaxation assay. Compounds 7±9 had weaker
activity relative to 10 and 13 indicating that the guanidino
group is more essential than the amino group for acti-
vation of the receptor. In addition, compound 13

K. Alexopoulos et al. / Bioorg. Med. Chem. 7 (1999) 1033±1041
1037
Phe and Arg aminoacid residues are in close proximity
and their side chains are characterized by reduced
mobility. Interactions between Arg and Phe side
chains have also been observed in the crystal structures
of proteins.21 Synthesized structures 6±13 were con-
structed and minimized using distance constrains in
order to achieve a spatial vicinity between Phe and Arg
(or -NH2 terminal) moieties. The two sets of com-
pounds (6±9 and 10±13) were overlaid with SFLLR in
order to achieve best matching. Their superimposition
was based on the following group matches: (i) Phe of
SFLLR with Phe of compounds 6±13, (ii) Arg of
SFLLR and Arg of 10±13 or -NH2 terminal group of 6±
9. The best overlay between the two sets of the synthe-
sized molecules was achieved with compounds 9 of the
®rst set (Fig. 6) and 13 of the second set (Fig. 7). Both
compounds have the longer alkyl chain and di€er only
at one terminal end. Analogue 9 possess an amino
group and 13 a guanidino one. The RMS values for the
superimposition between SFLLR and 9 or 13 were 0.83
and 0.19, respectively. It appears that the super-
imposition between SFLLR and the most active syn-
thetic analogue 13 was better than SFLLR and 9. These
results point out the importance of the Arg aminoacid
in the synthesized peptidomimetic analogues. Arginine
contains more polar atoms (three nitrogens) versus one
nitrogen that is contained in the aminoacid lysine of 9.
This polar similarity between SFLLR and analogue 13
may account for its better biological pro®le when com-
pared with 9. Comparison between compound 13 and
the other compounds of the same data set (10±12) shows
that this di€ers only in the length of the alkyl chain
(Fig. 8). Compound 13 has the longest alkyl chain. This
gives an advantage to this molecule to maximize its
hydrophobic interactions with leucine of SFLLR. This
close proximity between the long alkyl chain of compound
13 and leucine of SFLLR may explain its higher bio-
logical activity vis a vis compounds 10±12. In general,
compound 13 is the most biologically active compound
because: (i) it contains both Phe and Arg, the key ami-
noacid residues present in SFLLR; (ii) possess a long
alkyl chain that allows it for a close proximity with the
hydrophobic Leu residue of SFLLR. Therefore, the
emerging picture for these comparisons is that amphi-
pathic interactions are responsible for optimum activity
in the studied peptidomimetic analogues. Synthesis of
new analogues will lead to more accurate determination
of the key stereoelectronic factors that determine their
biological activity.
Figure 6. Superimposition of the synthetic analogue 9 with SFLLR
using body ®t to target method.
Figure 5. Proposed model of SFLLR generated by imposing distance
geometry constraints between Phe and Arg aminoacids and use of
molecular dynamics.
Figure 7. Superimposition of the synthetic analogue 13 with SFLLR
using rigid body ®t to target method.

1038
K. Alexopoulos et al. / Bioorg. Med. Chem. 7 (1999) 1033±1041
Figure 8. Low energy conformers generated by imposing distance geometry constraints between Phe and Arg aminoacids of synthetic compounds 10
(top, left) 11 (top, right), 12 (bottom, left), and 13 (bottom, right).
Conclusion
were obtained in CDCl3 or DMSO-d6. All chemical
shifts were referenced to tetramethylsilane (TMS). The
accurate M+ of the synthesized compounds were
obtained on a Kratos Pro®le HV-3 mass spectrometer
operating with mass range among 0±1200 amu at 4 kV
and resolution variable from nominal mass (600) to
accurate mass (10,000). Most of the compounds were
puri®ed by ¯ash chromatography using Merck silica gel
60 (230±400 mesh). Thin-layer chromatography (TLC)
was performed using Merck pre-coated silica gel 60 F-
254 glass plates. Visualization was accomplished by
using a combination of UV-lamp and ninhydrin spray.
All synthetic reagents were purchased either from the
Aldrich Chemical Company or the Sigma Chemical
Company and were used without puri®cation. Solvents
were analytical reagent grade or better and were used as
supplied. All air- and moisture-sensitive reactions were
performed under nitrogen. A Buchi RE 111 rotavapor
was utilized for the removal of the solvents in vacuo.
Moisture from the hydroscopic arginine derivatives was
removed using a freeze-drying pump. Melting point
determinations were performed on a Buchi 530 melting
point apparatus.
We have described the synthesis of novel low weight
1,4-disubstituted piperazines derived from TRAP that
interact with the thrombin receptor. Synthesis of such
novel peptidomimetic analogues aims (i) to investigate
the role of piperazine as a template in drug design and
(ii) to examine the interactions between Phe and Arg
residues. Among the synthesized compounds, 10 and 13
were found to be the most active in the rat aorta
relaxation assay and in platelet aggregation studies,
con®rming our proposed model for the interaction
between TRAP and the receptor. Furthermore, the sig-
ni®cance of Arg residue of the tethered ligand for
activity is once more indicated. These are lead com-
pounds for further development. It is anticipated that
modi®cation of 13 will provide us with better mimetics.
For example, utilization of piperazine constraints, sub-
stitution of phenyl group or even employment of di€erent
templates could give us more information in our ongo-
ing e€ort to develop peptide mimetics for drug therapy.
Experimental
Chemistry
1-(Phenylacetyl)piperazine (1). Phenylacetic acid (1.65 g,
12.2 mmol) and HOBt (1.65 g, 12.2 mmol) were sus-
pended in CHCl3 (25 mL). DCC (2.51 g, 12.2 mmol) was
1
The H NMR spectra were recorded on a Bruker ACF-
300 spectrometer operating at 7.05 T. All NMR spectra

K. Alexopoulos et al. / Bioorg. Med. Chem. 7 (1999) 1033±1041
1039
then added followed by DMF (3 mL). The mixture was
stirred for ꢀ30 min at room temperature to give a white
suspension. The mixture was transferred slowly to a pre-
cold solution of piperazine (5.24 g, 60.9 mmol) in 60 mL
CHCl3. The reaction was stirred for 2 h at room tem-
perature and then the white suspension (DCU) was ®l-
tered and the ®ltrate was acidi®ed with 2 M HCl
(3Â30 mL). The HCl extracts (120 mL) were basi®ed
with 2 M NaOH, extracted with CHCl3 (4Â40 mL), dried
over anhydrous Na2SO4 and reduced in vacuo to give
2.27 g (11.1 mmol, 91%) of a pale-green liquid as a crude
product. Flash chromatography (7% MeOH:CHCl3)
yielded 1.87 g (9.2 mmol, 75%) of a colorless oil. Rf 0.38
(ButOH:H2O:AcOH, 4:1:1), 1H NMR (CDCl3): d 1.8 (s,
1H), 2.6 (t, 2H), 2.8 (t, 2H), 3.4 (t, 2H), 3.6 (t, 2H), 3.7 (s,
1-Phenylacetyl-4-[6-(tert-butoxycarbonylamino)hexanoyl]-
piperazine (5). The procedure described above for 2 was
followed except that Boc-e-Ahx-OH was used instead of
Boc-Gly-OH. The product was obtained pure in 64%
yield, after ¯ash chromatography (2% MeOH:CHCl3).
1
Rf 0.83 (30% MeOH:CHCl3), H NMR (CDCl3): d 1.3
(t, 2H), 1.4 (s, 9H), 1.5 (m, 2H), 1.6 (m, 2H), 2.3 (m,
2H), 3.1 (m, 2H), 3.2±3.6 (m, 8H), 3.7 (s, 2H), 4.6 (br s,
1H), 7.2±7.3 (m, 5H), MS (nominal): 417 (M+), 344
(M+ (CH3)3COꢁ ), MS (accurate): M+ calcd: 417.2627,
M+ found: 417.2619.
General procedure for Boc-deprotection with tri¯uoro-
acetic acid. Each of the Boc-derivatives (1.0 mmol) was
dissolved in 3 mL of a 30% TFA:CH2Cl2 solution and
the mixture was stirred for 1 h. The solution was eva-
porated to dryness in vacuo and the residue was tritu-
rated with dry ether giving a yellowish oil (0.9 mmol,
90%).
2H), 7.2±7.3 (m, 5H), MS (nominal): 204 (M+),113
ꢁ
(M+ C6H5CH2 ), MS (accurate): M+ calcd: 204.1263,
M
+ found: 204.1267.
1-Phenylacetyl-4-[2-(tert-butoxycarbonylamino)acetyl]-
piperazine (2). 1-(Phenylacetyl)piperazine (400 mg,
1.9 mmol), Boc-Gly-OH (2.0 mmol), HOBt (284 mg,
2.1 mmol) and DIEA (1 mL) were dissolved in 15 mL
CH2Cl2. The mixture was cooled to 0 ꢂC and DCC
(433 mg, 2.1 mmol) was added. The reaction was stirred
at this temperature for ꢀ30 min and then at room tem-
perature. The reaction was monitored by TLC (20%
MeOH:CHCl3) and showed completion after ꢀ18 h.
The suspension (DCU) was ®ltered and washed with
25 mL CH2Cl2. The ®ltrate was washed with 15% citric
acid (2Â20 mL), satd NaHCO3 solution (2Â20 mL) and
H2O (2Â20 mL), dried over Na2SO4 and concentrated
in vacuo to give an orange-yellow solid as a crude pro-
duct. Flash chromatography (2% MeOH:CHCI3) yiel-
ded a waxy white solid (397 mg, 1.1 mmol, 58%). Rf
1-Phenylacetyl-4-(2-aminoacetyl)piperazine tri¯uoroacetic
salt (6). Rf 0.22 (30% MeOH:CHCl3), 1H NMR
(DMSO-d6): d 3.2±3.7 (m, 8H), 3.5 (m, 2H), 3.8 (s, 2H),
7.2±7.3 (m, 5H), 8 (br s, 3H), MS (nominal): 261
(M+ TFA), MS (accurate): (M+ TFA) calcd:
261.1478, (M+ TFA) found: 261.1474.
1-Phenylacetyl-4-(3-aminopropionyl)piperazine tri¯uoro-
1
acetic salt (7). Rf 0.24 (30% MeOH:CHCl3), H NMR
(CDCl3): d 2.7 (br s, 2H), 3.2±3.6 (m, 8H), 3.4 (m, 2H),
3.7 (s, 2H), 7.2±7.3 (m, 5H), 8 (br s, 3H), MS (nominal):
275 (M+ TFA), MS (accurate): (M+ TFA) calcd:
275.1634, (M+ TFA) found: 275.1639.
1-Phenylacetyl-4-(4-aminobutyryl)piperazine tri¯uoroacetic
salt (8). Rf 0.29 (30% MeOH:CHCl3), 1H NMR
(CDCl3): d 1.9 (br s, 2H), 2.4 (br s, 2H), 3.0±3.6 (m,
8H), 3.4 (br s, 2H), 3.7 (s, 2H), 7.2±7.3 (m, 5H), 8 (br s,
3H), MS (nominal): 289 (M+ TFA), MS (accurate):
(M+ TFA) calcd: 289.1790, (M+ TFA) found:
289.1795.
1
0.79 (30% MeOH:CHCl3), H NMR (CDCl3): d 1.4 (s,
9H), 3.4 (d, 2H), 3.1±3.9 (m, 8H), 3.7 (s, 2H), 5.4 (br s,
1H), 7.2±7.3 (m, 5H), MS (nominal): 361 (M+), 288
(M+ (CH3)3COꢁ ), MS (accurate): M+ calcd: 361.2002,
M+ found: 361.2009.
1-Phenylacetyl-4-[3-(tert-butoxycarbonylamino)propionyl]-
piperazine (3). The same procedure described as above
for 2 was followed except that Boc-b-Ala-OH was used
instead of Boc-Gly-OH. The product was obtained pure
in 59% yield, after ¯ash chromatography (2% MeOH:
CHCl3). Rf 0.80 (30% MeOH:CHCl3), 1H NMR
(CDCl3): d 1.4 (s, 9H), 2.5 (t, 2H), 3.4 (m, 2H), 3.1±3.6
(m, 8H), 3.7 (s, 2H), 5.2 (br s, 1H), 7.2±7.3 (m, 5H), MS
(nominal): 375 (M+), 302 (M+ (CH3)3COꢁ ), MS
(accurate): M+ calcd: 375.2158, M+ found: 375.2157.
1-Phenylacetyl-4-(6-aminohexanoyl)piperazine tri¯uoro-
1
acetic salt (9). Rf 0.36 (30% MeOH:CHCl3), H NMR
(CDCl3): d 1.3 (t, 2H), 1.5 (m, 2H), 1.6 (m, 2H), 2.3 (m,
2H), 2.9±3.6 (m, 8H), 3.4 (br s, 2H), 3.7 (s, 2H), 7.2±7.3
(m, 5H), 8 (br s, 3H), MS (nominal): 317 (M+ TFA),
MS
(accurate):
(M+ TFA)
calcd:
317.2103,
(M+ TFA) found: 317.2108.
General procedure for the guanylation of primary
amine derivatives. TFA-salt (0.7 mmol), 1H-pyrazole-1-
carboxamidine hydrochloride (175 mg, 1.2 mmol) and
DIEA (0.7 mL) were dissolved in 2 mL DMF giving a
green-brown liquid. The mixture was stirred under
nitrogen for ꢀ24 h at room temperature. Then, ether
(15 mL) was added and the product appeared as gel at
the bottom of the ¯ask, while the supernatant liquid was
decanted. The crude product was triturated with acet-
one:Et2O and ®nally the remaining moisture was
removed in a freeze-drying pump to give a grey solid
which was recrystallized from MeOH:Acetone:Et2O to
give pure product (0.35 mmol, 50%).
1-Phenylacetyl-4-[4-(tert-butoxycarbonylamino)butyryl]-
piperazine (4). The same procedure described as above
for 2 was followed except Boc-g-Abu-OH was used
instead of Boc-Gly-OH. The product was obtained pure
in 69% yield, after ¯ash chromatography (2% MeOH:
CHCl3). Rf 0.81 (30% MeOH:CHCl3), 1H NMR
(CDCl3): d 1.4 (s, 9H), 1.8 (t, 2H), 2.4 (m, 2H), 3.1 (m,
2H), 3.1±3.6 (m, 8H), 3.7 (s, 2H), 4.7 (br s, 1H), 7.2±7.3 (m,
5H), MS (nominal): 389 (M+), 316 (M+ (CH3)3COꢁ ),
MS (accurate): M+ calcd: 389.2315, M+ found:
389.2324.

1040
K. Alexopoulos et al. / Bioorg. Med. Chem. 7 (1999) 1033±1041
1-Phenylacetyl-4-(2-guanidoacetyl)piperazine (10). Melt-
ing point: 190±191 ꢂC, Rf 0.40 (ButOH:H2O:AcOH,
4:1:1), H NMR (DMSO-d6): d 3.3±3.5 (m, 8H), 5.4 (m,
Platelet aggregation assay
1
A. Preparation of platelet-rich plasma
2H), 3.7 (s, 2H), 7.2±7.3 (m, 5H), 7.0±7.4 (br s, 3H), 7.5
(br s, 1H), MS (nominal): 303 (M+), MS (accurate):
M+ calcd: 303.1695, M+ found: 303.1695.
Blood was drawn between 8 and 11 a.m. from healthy
volunteers, who had not taken medications known to
a€ect platelet function for the past two weeks. Blood
venus (30 mL) was collected from a forearm vein,
without venous stasis, using a 19-gauge butter¯y needle,
into a 3.2% trisodium citrate bu€er, as anticoagulant
(®nal concentration 10 mM). Blood samples were then
centrifuged at 800 rpm at room temperature for 8 min
and the supernatant platelet-rich plasma (PRP) was
counted by a Coulter counter (Coulter Corp.) and dilu-
ted with platelet-poor plasma (PPP) to the desired ®nal
concentration (2Â108 platelets/mL).
1-Phenylacetyl-4-(3-guanidopropionyl)piperazine (11). Melt-
ing point: 184±185 ꢂC, Rf 0.41 (ButOH:H2O:AcOH,
1
4:1:1), H NMR (DMSO-d6): d 2.6 (t, 2H), 3.3±3.5 (m,
8H), 3.4 (m, 2H), 3.7 (s, 2H), 7.2±7.3 (m, 5H), 6.8±7.4
(br s, 3H), 7.5 (br s, 1H), MS (nominal): 316 (M+ 1),
258 [(M+ 1) H2NCNHNHꢁ ], MS (accurate): M+
calcd: 317.1852, M+ found: 317.1858.
1-Phenylacetyl-4-(4-guanidobutyryl)piperazine (12). Melt-
ing point: 169±170 ꢂC, Rf 0.43 (ButOH:H2O:AcOH,
4:1:1), 1H NMR (DMSO-d6): d 1.65 (t, 2H), 2.3 (m,
2H), 3.0±3.5 (m, 8H), 3.4 (m, 2H), 3.7 (s, 2H), 7.2±7.3
(m, 5H), 6.8±7.4 (br s, 3H), 7.7 (br s, 1H), MS (nom-
inal): 331 (M+), 315 (M+ H2Nꢁ ), MS (accurate): M+
calcd: 331.2008, M+ found: 331.2016.
B. Platelet aggregation study
Platelet aggregation was determined by the optical
method in a four-channel platelet aggregometer (Bio/
Data, USA) equipped with an aggregation module
PAP-4C. Samples of platelet suspensions (450 mL)
under stirring at 1100 rpm were activated with 50 mL
of either increasing doses of bovine thrombin (0.1,
0.2, and 0.3 NIU/mL) or 1 mM of SFLLR-NH2 at
37 ꢂC. Thrombin was obtained from Chromogenix,
Sweden and SFLLR-NH2 was synthesized in our Patras
Laboratory.
1-Phenylacetyl-4-(6-guanidohexanoyl)piperazine (13). Melt-
ing point: 112±114 ꢂC, Rf 0.47 (ButOH:H2O:AcOH,
1
4:1:1), H NMR (DMSO-d6): d 1.3 (t, 2H), 1.5 (m, 4H),
2.3 (m, 2H), 3.0±3.5 (m, 8H), 3.4 (m, 2H), 3.7 (s, 2H),
7.2±7.3 (m, 5H), 6.8±7.5 (br s, 3H), 7.8 (br.s, 1H), MS
(nominal): 359 (M+), 343 (M+ H2Nꢁ ), MS (accurate):
M+ calcd: 359.2321, M+ found: 359.2316.
Light transmittance was measured and recorded for
4 min at which time platelets from most platelet donors
had reached a maximal response to thrombin. The
results were expressed as the percent change in light
transmittance, with light transmittance of plasma alone
(PPP) taken as 100% and that of plasma with platelets
in suspension taken as 0%. In order to evaluate the
biological e€ects of the non-peptide thrombin receptor
mimetics under investigation on platelet aggregation,
platelet suspensions were incubated under stirring with
the desired concentrations of each compound for at
least 4 min prior to the addition of thrombin (0.1±
0.3 NIU/mL) or SFLLR-NH2 (1 mM). Each study was
completed within 1 h after blood collection.
Rat aorta relaxation assay
Compounds were dissolved in 25 mM Hepes bu€er, pH
7.4 and were diluted appropriately for evaluation in the
rat aorta relaxation assay as described in detail else-
where.12,13 The activity of the compounds was mani-
fested as a relaxation response in an aorta ring
preparation that had been precontracted by the addition
of 1 mM phenylephrine (PE). Aorta preparations were
used either with an intact endothelium, responsible
for the release of nitric oxide; or with the endothe-
lium denuded, in which case, no relaxant activity due to
nitric oxide release could be observed. The relaxant
activity of all compounds was monitored relative to that
of 1 mM acetylcholine (Ach), which requires the
presence of an intact endothelium for its nitric oxide-
dependent relaxant activity. The relaxant activity of
compound 13 was also compared with that of the
thrombin receptor activating peptide, SFLLR-NH2
(P5-NH2); the nitric oxide-dependent relaxant activity
of this peptide is also dependent upon the presence of
an intact endothelium. To construct concentration-
e€ect curves for all test compounds, the relaxant activity
of each compound was expressed as a percentage
(%Ach) of the relaxation caused in the same prepara-
tion by 1 mM acetylcholine. Values shown in the
concentration-response curves represent the averages
(‹SEM) of three or more independently measured
responses in di€erent tissue preparations. Those com-
pounds lacking activity were evaluated in at least two
separate experiments using tissues from di€erent
animals.
C. Calculation of IC50 values
The IC50 for the data regarding the response to the
three levels of compounds (10, 11, and 13) in di€erent
thrombin concentrations (0.2 NIU/mL and 0.25±
0.3 NIU/mL) was evaluated using the Random E€ects
Model (Repeated Maximum Likelihood, or REML
method).32 The independent variable was the percent
reductions in optical density (which itself was expressed
as a percent) from the baseline value, for various con-
centrations of the drug (from 0 to 2.0 mM). The depen-
dent variable was the drug concentration. Each patient
was considered as a separate variable nested within
the response variable and the analysis was performed
nine separate times, once for each type-thrombin
combination. The individual p-values of each estimate
indicate that all are statistically bigger than zero
(p<0.001).

K. Alexopoulos et al. / Bioorg. Med. Chem. 7 (1999) 1033±1041
1041
Acknowledgements
17. Panagiotopoulos, D.; Matsoukas, J.; Alexopoulos, K.;
Zebeki, A.; Mavromoustakos, T.; Saifeddine, M.; Hollenberg,
M. Let. Pept. Sci. 1996, 3, 233.
18. Bernatowicz, M.; Youling, W.; Matsueda, G. J. Org.
Chem. 1992, 57, 2497.
19. Mavromoustakos, T.; De-Ping Y.; Theodoropoulou E.;
Makriyannis, A. Eur. J. Med. Chem. 1995, 30, 227.
20. Hirschmann, R. Angew. Chem. Int. Ed. Engl. 1991, 30, 1278.
21. Giannis, A.; Kolter, T. Angew. Chem. Int. Ed. Engl. 1993,
32, 1244.
22. Panagiotopoulos, D.; Matsoukas, J.; Keramida, M.;
Gatos, D.; Monteiro, L.; Maia, H.; Yamdagni, R.; Wu, Q.;
Redstone, P.; Moore, G.; Laniyonu, A.; Saifeddine, M.; Hol-
lenberg, M. In Peptides 1994, Proceedings of the 23rd European
Peptide Symposium; Maia, H., Ed.; ESCOM: Leiden, 1995; pp
359±360.
23. (a) Williams, T., et. al. J. Med. Chem. 1996, 39, 1345. (b)
Mokrosz, J.; Paluchowska, M.; Chojnackawojcik, E.; Filip,
M.; Charakchievaminol, S.; Derenwesolek, A.; Mokrosz, M.
J. Med. Chem. 1994, 37, 2754. (c) Zhuang, Z.; Kung, M.;
Kung, H. J. Med. Chem. 1994, 37, 1406. (d) Seginko, J.;
Dobrocky, P.; Caganova, B. Pharmazie 1995, 50, 368. (e)
Kung, M.; Frederick, D.; Mu, M.; Zhuang, H. J. Pharmacol.
Exp. Ther. 1995, 272, 429.
This work was supported by European Community
(EC) grants (BIOMED program no. 920038, BIOMED-
PECO no. 930158, COPERNICUS No. 940238), and
the MRC of Canada and the Ministry of Energy and
Technology of Greece.
References
1. Bar-Shavit, R.; Kahn, A.; Mann, K.; Wilner, G. Proc. Natl.
Acad. Sci. USA 1986, 83, 976.
2. Chen, L.-B.; Buchanan, J. Proc. Natl. Acad. Sci. USA 1975,
72, 131.
3. White, R.; Chapleau, C.; Dugdale, M.; Robertson, J.
Stroke 1980, 11, 363.
4. Vu, T.-K.; Hung, D.; Wheaton, V.; Coughlin, S. Cell 1991,
64, 1057.
5. Rasmussen, U.; Vouret-Graviari, V.; Jallat, S.; Schlesinger,
Y.; Pages, G.; Pavirani, A.; Lecocq, J.-P.; Pouyssegur, J.; Van
Obberghen-Schilling, E. FEBS Lett. 1991, 288, 123.
6. Vu, T.-K.; Wheaton, V.; Hung, D.; Charo, I.; Coughlin, S.
Nature 1991, 353, 674.
24. Fukushi, H.; Mabuchi, H.; Itoh, K.; Tearshita, Z.;
Nishikawa, K.; Sugihara, H. Chem. Pharm. Bull. 1994, 42,
541.
7. Gerszten, R.; Chen, J.; Ishii, M.; Ishii, K.; Wang, L.;
Nanevicz, T.; Turck, C.; Vu, T.-K.; Coughlin, S. Nature 1994,
368, 648.
25. Fukushi, H.; Mabuchi, H.; Tearshita, Z.; Nishikawa, K.;
Sugihara, H. Chem. Pharm. Bull. 1994, 42, 551.
26. Gante, J.; Juraszyk, H.; Rattatz, P.; Wurziger, H.; Berno-
tat-Danielowski, S.; Melzer, G.; Rippmann, F. In Peptides
1996, Proceedings of the 24th European Peptide Symposium.
Ramage, R.; Epton, R. Eds; May¯ower, 1998; pp. 401±402.
27. Seiler, S.; Peluso, M.; Michel, I.; Goldenberg, H.; Fenton,
J.; Riexinger, D.; Natarajan, S. Biochem. Pharmacol. 1995, 49,
519.
28. Grand, J.; Turnell, A.; Grabham, P. Biochem. J. 1996,
313, 353.
29. Scarborough, R.; Teng, W.; Naughton, M.; Rose, J.;
Alves, V.; Arfsten, A. Circulation 1992, 86, 1.
30. Scarborough, R.; Teng, W.; Naughton, M.; Rose, J.;
Alves, V.; Arfsten, A.; Ramakrishnan, V.; Blackhart, B.
Abstract No. P632, 13th American Peptide Symposium,
Edmonton, Canada, 1993.
31. Raykal, D.; Schmedtje, J.; Runge, M. Am. J. Cardiol.
1995, 75, 82B.
32. Laird, N.; Ware, J. Biometrics 1982, 38, 963.
8. Chen, J.; Ishii, M.; Wang, L.; Ishii, K.; Coughlin, S. J. Biol.
Chem. 1994, 269, 16041.
9. Chao, B.; Kalkunte, S.; Maraganore, J.; Stone, S. Bio-
chemistry 1992, 31, 6175.
10. Vassallo, R.; Kieber-Emmons, T.; Cichowski, K.; Brass,
L. J. Biol. Chem. 1992, 267, 6081.
11. Scarborough, R.; Naughton, M.; Teng, W.; Hung, D. J.
Biol. Chem. 1992, 267, 13146.
12. Hollenberg, M.; Yang, S.-G.; Laniyonu, A.; Moore, G.;
Saifeddine, M. Mol. Pharmacol. 1992, 42, 186.
13. Hollenberg, M.; Laniyonu, A.; Saifeddine, M.; Moore, G.
Mol. Pharmacol. 1993, 43, 921.
14. Matsoukas, J.; Panagiotopoulos, D.; Keramida, M.;
Mavromoustakos, T.; Yamdagni, R.; Qiao, W.; Moore, G.;
Saifeddine, M.; Hollenberg, M. J. Med. Chem. 1996, 39, 3585.
15. Matsoukas, J.; Hollenberg, M.; Mavromoustakos, T.;
Panagiotopoulos, D.; Alexopoulos, K.; Yamdagni, R.; Wu,
Q.; Moore, G. J. Prot. Chem. 1997, 16, 113.
16. Moore, G.; Smith, J.; Baylis, B.; Matsoukas, J. Advances
in Pharmacology 1995, 33, 91.