Thrombin receptor-activating peptides (TRAPs): Investigation of bioactive conformations via structure-activity, spectroscopic, and computational studies

Bioorganic & Medicinal Chemistry 7 (1999) 2353±2371

Thrombin Receptor-Activating Peptides (TRAPs): Investigation of

Bioactive Conformations via Structure±Activity, Spectroscopic,

and Computational Studies

Marco A. Ceruso, ^aDavid F. McComsey, ^aGregory C. Leo, ^aPatricia Andrade-Gordon, ^a

Michael F. Addo, ^aRobert M. Scarborough, ^bDonna Oksenberg ^band

Bruce E. Maryano ^a,*

^aThe R. W. Johnson Pharmaceutical Research Institute, Spring House, PA 19477, USA

^bCOR Therapeutics, Inc., South San Francisco, CA 94080, USA

Received 11 December 1998; accepted 10 May 1999

AbstractÐThe thrombin receptor (PAR-1) is an unusual transmembrane G-protein coupled receptor in that it is activated by serine

protease cleavage of its extracellular N-terminus to expose an agonist peptide ligand, which is tethered to the receptor itself. Synthetic

peptides containing the agonist motif, such as SFLLRN for human PAR-1, are capable of causing full receptor activation. We have

probed the possible bioactive conformations of thrombin receptor-activating peptides (TRAPs) by systematic introduction of certain

conformational perturbations, involving a-methyl, ester É(COO), and reduced-amide É(CH₂N) scans, into the minimum-essential

agonist sequence (SFLLR) to probe the importance of the backbone conformation and amide NH hydrogen bonding. We performed

extensive conformational searches of representative pentapeptides to derive families of putative bioactive structures. In addition, we

1

employed H NMR and circular dichroism (CD) to characterize the conformational disposition of certain pentapeptide analogues

experimentally. Activation of platelet aggregation by our pentapeptide analogues aorded a structure-function correlation for PAR-1

agonist activity. This correlation was assisted by PAR-1 receptor binding data, which gauged the anity of peptide ligands for the

thrombin receptor independent of a functional cellular response derived from receptor activation (i.e. a pure molecular recognition

event). Series of alanine-, proline-, and N-methyl-scan peptides were also evaluated for comparison. Along with the known structural

features for PAR-1 agonist peptides, our work adds to the understanding of peptide topography relative to platelet functional activity

and PAR-1 binding. The absolute requirement of a positively charged N-terminus for strong agonist activity was contradicted by the

N-terminal hydroxyl peptide É(HO)S-FLLR-NH₂. The amide nitrogen between residues 1 and 2 was found to be a determinant of

receptor recognition and the carbonyl groups along the backbone may be involved in hydrogen bonding with the receptor. Position 3

(P3) of TRAP-5 is known to tolerate a wide variety of side chains, but we also found that the amide nitrogen at this position can be

substituted by an oxygen, as in SF-É(COO)-LLR-NH₂, without diminishing activity. However, this peptide bond is sensitive to

conformational changes in that SFPLR-NH₂was active, whereas SF-NMeL-LR-NH₂was not. Additionally, we found that position

3 does not tolerate rigid spacers, such as 3-aminocyclohexane-1-carboxylic acid and 2-aminocycloalkane-1-carboxylic acid, as ana-

logues 1A, 1B, 2A, 2B, 3, 4, 5A and 5B lack agonist activity. On the basis of our results, we suggest that an extended structure of the

agonist peptide is principally responsible for receptor recognition (i.e. binding) and that hydrophobic contact may occur between the

Introduction

Abbreviations: Nonstandard chemical abbreviations not de®ned in

text: Boc, tert-butyloxycarbonyl; BOP-Cl, N,N⁰-bis(2-oxo-3-oxazolidi-

nyl)phosphinyl chloride; DMS, dimethylsul®de; DMAP, 4-(dimethyl-

Many biological processes rely on receptor-based

intracellular communication involving endogenous

amino)pyridine; DCE, 1,2-dichloroethane; DIC, N,N⁰-diisopropyl-

hormones, neurotransmitters, or growth factors. Of

carbodiimide; PMC, 2,2,5,7,8-pentamethylchroman-6-sulfonyl; Mts,

particular signi®cance are the G-protein coupled recep-

2,4,6-trimethylbenzenesulfonyl; Fmoc, 9-¯uorenylmethoxycarbonyl;

tors (GPCRs), a large superfamily of cell-surface glyco-

proteins. GPCRs are activated by diverse small and

large molecules, ranging from biogenic amines, such as

dopamine and histamine, to peptides, such as substance

P and cholecystokinin, to proteins, such as gonado-

tropin-releasing hormone (GnRH), and are also targets

HOBT, 1-hydroxybenzotriazole; TFA, tri¯uoroacetic acid; Har=

homoarginine, Aib=l-2-aminoisobutyric acid; Cha=l-cyclohexyl-

alanine.

Key words: Peptides; receptor agonists; NMR; circular dichroism;

molecular modeling.

* Corresponding author. Tel.: +1-215-628-5530; fax: +1-215-628-

4985; e-mail: bmaryano@ prius.jnj.com

PII: S0968-0896(99)00180-7

2354

M. A. Ceruso et al. / Bioorg. Med. Chem. 7 (1999) 2353±2371

responsible for the action of a broad selection of drugs.^1±6

For a long time, all GPCRs were thought to be activated

by native ligands, which are biosynthesized at a site

remote from the receptor and compelled to travel some

distance before reaching their target. However, this pic-

ture changed dramatically with the cloning of the

``platelet'' thrombin receptor (TR; PAR-1) from human

megakaryocytoblastoma cells.⁷This unusual GPCR

mediates the cellular actions of the serine protease

a-thrombin via proteolytic cleavage of its lengthy, extra-

cellular, N-terminal peptide chain, between Arg-41 and

Ser-42, to expose a new N-terminus bearing the sequence

SFLLRN (TR 42-47).^7±13Remarkably, synthetic pep-

tides containing this motif, such as the hexapeptide

SFLLRN-NH₂(TRAP-6), exhibit full agonist properties

for thrombin receptor activation independent of throm-

bin-induced proteolysis, indicating that the receptor-

linked SFLLRN sequence does serve as an activating

ligand.^7,14Thus, the thrombin receptor became the ®rst

known GPCR to be activated by a protease (``protease-

activated receptor'' or ``PAR'') and the ®rst GPCR to

have an intramolecular mechanism for ligand activation, in

this case via a ``tethered-peptide'' epitope. Subsequently,

three other members of this unique class of GPCRs,

PAR-2, PAR-3, and PAR-4, have been identi®ed.^15±20

peptides, the 14-mer SFLLRNPNDKYEPF-NH₂(TR

42-55; TRAP-14) was found to contain a stabilized

turn between positions 6 and 9 (NPND), but the ®ve

N-terminal residues were unstructured in solution.³¹An

NMR study of SFLLRNPNDKY-NH₂(TR 42-52;

TRAP-11) showed a central type-I b-turn around

NPND that is controlled by the rigidity of the proline

and the charge of the aspartyl residue.³²Additionally,

an unconstrained computational search of TRAP-6

revealed that this small peptide is capable of adopting a

folded conformation involving a g-turn around Leu-3;³²

however, this secondary structure type would not be

stable in a polar hydrogen-bonding environment. A

structural investigation of TRAP-6 analogues with

various conformational perturbations at positions 1±3

(P1±P3) illustrated some salient points for a bioactive

structure: (1) the Phe-2 side chain has a limited con-

formation with backbone torsional angles of an exten-

ded conformation; (2) the Leu-3 residue has a f

torsional angle similar to that of proline and a É tor-

sional angle close to that of a b-sheet; and (3) a trans

con®guration is likely for the amide bonds of S-F and

F-L.²⁸These studies, although quite valuable, have

mainly addressed the conformational preference of the

unbound ligand, rather than the bound bioactive form.

Systematic structure±activity studies in dierent labora-

tories have established several critical components of

the thrombin receptor-activating peptides (TRAPs),

relative to PAR-1.^13,14,21±27The minimum sequence

length is ®ve amino acids, with SFLLR-NH₂(TRAP-5)

being a good agonist but SFLL-NH₂being quite

weak.^14,21±24The phenylalanine (F) residue at position 2

is crucial for activity, while an N-terminal free amino

group, a basic or aromatic residue at position 5, and a

bulky aliphatic residue at position 4 are moderately

important.^14,22±26The amino acids at positions 1, 3, and

6 (S, L, and N, respectively) can be replaced by alanine

without signi®cantly attenuating agonist potency (``ala-

nine scan''), positions 3 and 6 are highly tolerant of

diverse substitution, and positions 1 and 3 tolerate sub-

stitution with proline (``proline scan'').^14,22±27By con-

trast, the scanning of positions 1 through 5 with d- or

N-Me amino acids caused a major loss in agonist

potency.^25,26,28Interestingly, modi®cation of certain

side chains, especially by having p-¯uorophenylalanine

at position 2, yielded agonist peptides with a signi®cant

enhancement of potency, such as A-(p-¯uoro-F)-R-

Cha-Har-Y-NH₂and S-(p-¯uoro-F)-Cha-Cha-RK-

NH2.^25,27,29,30The body of existing structure±activity

data for TRAPs is based on functional biological activ-

ity mediated by the PAR-1 thrombin receptor, rather

than on ligand binding anity, and focuses mainly on

the nature of the amino acid side chains of the peptides.

However, for such peptide ligands it would be important

to have receptor binding data to better correlate biologi-

cal activity with structural features, such as backbone

conformation and amide hydrogen-bonding. Thus, we

were intent on obtaining PAR-1 receptor binding data.

We have been interested in analyzing the possible

bioactive conformations of TRAPs as part of a project

directed toward the discovery of thrombin receptor

(PAR-1) antagonists, a class of agents that could serve

as useful antithrombotic drugs. Consequently, we have

examined the eects of systematic introduction of con-

formational perturbations into the native agonist peptide

sequence. In particular, we carried out a-methyl, ester

É(COO), and reduced amide É(CH₂N) scans with

SFLLR-NH₂to assess the importance of the backbone

conformation and amide NH hydrogen bonding. By

using spatial constraints established for Phe-2 in the

work of Shimamoto et al.²⁸we performed extensive

conformational searches of these and related pentapeptide

structures to derive families of putative bioactive con-

formations. We employed ¹H NMR and circular dichroism

(CD) measurements to characterize the conformational

disposition of certain agonist peptide analogues experi-

mentally. Activation of platelet aggregation by our

pentapeptide analogues aorded a structure±function

correlation for agonist activity, which was assisted by

data from a novel PAR-1 receptor binding assay.³³This

allowed us to assess the anity of agonist peptide ligands

for PAR-1 independent of a functional cellular response

derived from receptor activation, which is preferable for

interpreting molecular recognition events. Our work

adds to the body of information on the structure-func-

tion properties of TRAPs, and could help in the design of

new agonist and antagonist molecules for PAR-1.

Results

Considerable work with TRAPs, as noted above, has

principally focused on the role of the individual side

chains and of the termini. In this paper, we have taken a

systematic approach to evaluating the spatial relation-

ships between dierent elements of TRAP-5, by using a

A limited number of studies have been conducted to

de®ne the favored or bioactive conformations of

TRAPs. In an NMR study of SFLLRN-containing

M. A. Ceruso et al. / Bioorg. Med. Chem. 7 (1999) 2353±2371

2355

three-pronged strategy. (1) TRAP-5 analogues that con-

tain peptide-bond surrogates and local conformational

perturbations were synthesized and biologically eval-

uated. Thus, we have analyzed a-methyl, N-methyl, and

proline scans of SFLLR-NH₂to determine the in¯uence

of the peptide backbone conformation. Additionally,

the importance of amide NH and carbonyl groups,

potentially re¯ective of hydrogen bonding capabilities,

was probed by replacing each peptide bond of SFLLR-

NH₂with ester (COO) and secondary amine (CH₂NH)

linkages. (2) The conformational space of relevant

TRAP-5 mimetics was characterized computationally

by using molecular dynamics methods. Thus, we per-

formed extensive conformational searches of TRAP-5

and related pentapeptide structures to derive families of

putative bioactive conformations. (3) The proposed

three-dimensional models were tested by the synthesis,

biological evaluation, and physical (¹H NMR and CD)

characterization of additional analogues. The biological

responses of the synthetic peptides were measured by

two independent methods: a functional cellular assay

involving platelet aggregation and a PAR-1 binding

assay³³based on the very potent agonist ligand [³H]S-

(p-¯uoro-F)-Har-L-Har-KY-NH₂. This dual evaluation

is important in that the dierential analysis of responses

can distinguish those features in TRAP-5 analogues

required for receptor recognition (binding) and signal

transduction from those required for just receptor

recognition. Reported structure-function relationships

for diverse TRAP analogues have relied mainly on biolo-

gical function mediated by PAR-1, which re¯ects both

binding and subsequent signal transduction. Hence, our

structure±activity results should be amenable for inter-

preting molecular recognition events of the PAR-1 ligands.

Table 1. Biological data for TRAP-5 analogues containing peptide

bond surrogates

Entry

Peptide

Platelet

aggregation^a

EC₅₀(mM)

Receptor

binding^b

IC₅₀(mM)

1

2

3

4

5

6

7

8

SFLLR-NH₂

AFLLR-NH₂

SALLR-NH₂

SFALR-NH₂

SFLAR-NH₂

SFLLA-NH₂

0.490.08

0.740.16

>50 (10%)

1.20.3

1.50.5

0.660.15

>100

0.750.2

1.80.8

1.70.7

14

3.40.3

347^c

3.30.7

3.70.4

É(HO)A-FLLR-NH₂

É(HO)S-FLLR-NH₂

S-É(COO)-FLLR-NH₂

SF-É(COO)-LLR-NH₂

SFL-É(COO)-LR-NH₂

SFLL-É(COO)-R-NH₂

S-É(CH₂NH)-FLLR-NH₂

SF-É(CH₂NH)-LLR-NH₂

SFL-É(CH₂NH)-LR-NH

SFLL-É(CH₂NH)-R-NH

3.6

9

>50 (20%)

0.240.18

>50 (38%)

>50 (31%)

>50 (5%)

>50 (7%)

>50 (11%)

>50 (17%)

>100

8.94.1

20

10

11

12

13

14

15

16

2.5

>100

a

Activation of human platelet aggregation, expressed as an EC₅₀

value in mM, or percent aggregation induced at 50 mM. PAR-1

antagonist activity was not signi®cant (<20% at 50 mM), unless noted

otherwise.

b

Inhibition of radiolabeled agonist peptide binding to PAR-1,

expressed as an IC₅₀value in mM. IC₅₀values are the concentration of

test ligand required for 50% inhibition of the radioligand (n=3 when

the error limits are given; otherwise, n=1).

c

Antagonist IC₅₀value of 20 mM.

and geometrical properties associated with ligand

recognition and receptor activation. In our structure-

function studies, topographical features were probed by

replacing backbone atoms (Table 1) and conformational

attributes were probed by impacting the ¯exibility of the

peptide backbone (Table 2). To investigate the back-

bone elements possibly involved in hydrogen-bonding

interactions, either within TRAP-5 (intramolecular) or

with the receptor (intermolecular), we synthesized and

bioassayed two series of pentapeptides: an ester linkage

scan series, employing É(COO) for each amide linkage

For SFLLR-NH₂(TRAP-5), PAR-1 binding (IC₅₀=

1.5 mM) and platelet activation data (EC₅₀=0.49 mM)

are reasonably consistent (Table 1, entry 1). To check the

binding assay against the functional assay further, we

obtained a set of data for the alanine-scan peptides of

TRAP-5 (entries 2±6). The results for alanine substitution

at positions 1±4 are reasonably consistent with a correla-

tion between receptor binding and platelet activation,

although the agonist activity for SFLLA-NH₂(entry 6)

is weaker than one would expect from the level of recep-

tor binding. In some cases, a correspondence between the

receptor binding and agonist activity may be confounded

by the presence of antagonist activity, which is deter-

mined as an IC₅₀value for inhibition of thrombin-

induced platelet aggregation or percent inhibition at

50mM. Indeed, SFLLA-NH₂exhibits PAR-1 antagonism

with an IC₅₀value of 20mM. Consequently, it is probably

(depsipeptides) and

a reduced-amide scan series,

employing the amide-bond surrogate É(CH₂NH).

Deletion of carbonyl groups along the backbone pro-

duced compounds that did not bind to PAR-1 and,

correspondingly, there was no functional response

observed (Table 1). On the other hand, most of the

depsipeptides were capable of eliciting at least one of the

two biological responses. Of particular note, SF-

É(COO)-LLR-NH₂(Table 1, entry 10) showed response

levels comparable to those of TRAP-5. However, SFL-

É(COO)-LR-NH₂and SFLL-É(COO)-R-NH₂(entries

11 and 12), although capable of binding, failed to

induce platelet aggregation; also, S-É(COO)-FLLR-

NH₂(entry 9) was devoid of activity. Among the depsi-

peptide analogues, we prepared a derivative with l-

glyceric acid in place of Ser-1, É(HO)S-FLLR-NH₂

(entry 8), and this compound was capable of PAR-1

binding and activation. This is surprising because the

positively charged N-terminus is now replaced by a

neutral hydroxyl. To test this observation further, the

l-lactic acid analogue, É(HO)A-FLLR-NH₂(entry 7),

was prepared and found to be comparable in activity to

the l-glyceric acid compound. The data for the É(COO)

unreasonable to expect

a precise correspondence

between the receptor binding and PAR-1 agonist data,

although the binding data would generally be reliable in

re¯ecting agonist or antagonist activity. Certainly, the

functional result for replacement of F by A is consistent

with the literature, and our binding result follows suit.

Synthetic peptides and biological activity

The rational design of peptide ligands with speci®c biolo-

gical activity requires knowledge about the topographical

2356

M. A. Ceruso et al. / Bioorg. Med. Chem. 7 (1999) 2353±2371

Table 2. Biological data for conformationally restricted TRAP-5

analogues

(entries 7±11) are the least tolerated modi®cations. The

dierence in functional activity observed between

SFPLR-NH₂(entry 14) and SF-NMeL-LR-NH₂(entry

4) is rather striking and will be discussed later on.

Entry

Peptide

Platelet

aggregation^a

EC₅₀(mM)

Receptor

binding^b

IC₅₀(mM)

Conformational analysis

1

2

3

4

5

6

7

8

SFLLR-NH₂

0.490.08

164^c

1.50.5

0.70.3

629

5.30.7

160.8

114

>100

0.520.15

>100

1.70.2

528

144

NMeS-FLLR-NH₂

S-NMeF-LLR-NH₂

SF-NMeL-LR-NH₂

SFL-NMeL-R-NH₂

SFLL-NMeR-NH₂

ꢀMeS-FLLR-NH₂

S-ꢀMeF-LLR-NH₂

SF-ꢀMeL-LR-NH₂

SFL-ꢀMeL-R-NH₂

SFLL-ꢀMeR-NH₂

PFLLR-NH₂

Computational techniques were used to develop a qua-

litative impression of the three-dimensional bioactive

conformation of TRAP-5. As a starting point, an

unrestrained search of the conformational space of

TRAP-5 was performed by using a protocol based on

the quenched high-temperature molecular-dynamics

methodology of Karplus and co-workers.^38,39This pro-

tocol was, in turn, applied to several of the TRAP-5

analogues. Our objective was to develop a picture of the

backbone fold and the spatial relationships among the

side chains of TRAP-5 that were required for biological

activity by correlating the computational results with

the biological data.

>50 (3%)

>50 (10%)

>50 (12%)

>50 (37%)

>50 (10%)

>50 (18%)

>50 (17%)

0.590.14

>50 (3%)

4.10.5^e

d

9

d

10

11

12

13

14

15

16

SPLLR-NH₂

SFPLR-NH₂

SFLPR-NH₂

SFLLP-NH₂

>50 (4%)

>50 (11%)

a

See Table 1.

b

c

As a measure of the ecacy of the searches, the dis-

tribution of end-to-end distances (d_ij) and the displace-

ment autocorrelation function of the backbone atoms³⁹

were computed for each peptide under study. In all

cases, the displacement autocorrelation function

reached a plateau (Fig. 1a), suggesting that all of the

conformational space accessible at 1000 K had been

explored. The distribution of end-to-end distances,

which ranged from 3.5 to 16 A in most simulations

(Fig. 1b), agrees with what is expected for a ¯exible

pentapeptide. Since the maximum distance between two

consecutive residues (e.g. two a-carbons) is ca. 3.6 A,

a fully extended pentamer (b-sheet structure) would

have an end-to-end distance of ca. 18 A. In most cases

the distribution of end-to-end distances was uniform in

this range, but slightly bimodal distributions, where

extended conformations could be distinguished from

more collapsed ones, were also observed (Fig. 1b). In

any event, the energy histograms always showed a

unimodal Gaussian-like pattern (data not shown), as

previously reported for peptides.³⁸

Antagonist IC₅₀value of 392.5 mM.

1:1 Mixture of diastereomers due to racemic aMeL.

Antagonist IC₅₀value of ca. 50 mM.

d

e

and É(CH₂NH) analogues suggest that all of the back-

bone carbonyls are essential for recognition at the

receptor surface of PAR-1 and that at least a subset of

amide protons participate in intramolecular or inter-

molecular hydrogen bonding (see Discussion).

Next, we deliberately perturbed the conformation of

each residue along the sequence to investigate the back-

bone geometry of TRAP-5 as it binds to or activates the

receptor. Addition of a methyl group to the a-carbon

(a-methyl scan) or to the amide nitrogen (N-methyl

scan) is expected to reorient the local conformation at

the modi®ed residue.^34±37In the case of N-methyl sub-

stitution, the conformation of the preceding residue (i.e.

that toward the N-terminal) will also be aected. How-

ever, in such alterations conformational control is

gained through introduction of steric bulk, which needs

to be taken into account when interpreting the results.

Besides conducting a-methyl and N-methyl scans with

TRAP-5, we also conducted a proline scan, although

this one may have a disadvantage because of mutating

the side chain of the residue under study (Table 2). The

only conformationally perturbed analogues with activ-

ity in the functional assay are PFLLR-NH₂(Table 2,

entry 12), SFPLR-NH₂(entry 14), and NMe-SFLLR-

NH₂(entry 2), in agreement with known topographical

requirements (see Introduction); however, receptor

recognition (i.e. binding) was only partially aected. It

is interesting to note that N-Me substitution at position

1 (entry 2) gave a mixed agonist-antagonist, whereas

proline substitution at position 1 (entry 12) gave a

potent agonist. Proline and N-methyl substitutions at

position 2 are not tolerated, while the N-terminal and

P3 through P5 modi®cations conserve varying degrees

of binding anity. There is a parallelism in receptor

binding trends between the ester, N-methyl, and proline

scans at positions 1 through 5. The a-methyl substitutions

The results of conformational searches involving a-

methyl or N-methyl substituted residues were in good

agreement with literature data.^34±37Addition of a

methyl group to an amide nitrogen (N-Me scan) con-

®ned the backbone conformation of the modi®ed

residue to four regions of the Ramachandran map (Fig.

2A). The backbone conformation of the preceding resi-

due was, in turn, limited to the upper left-hand quad-

rant of the map (Fig. 2B). a-Methyl-substituted residues

had narrow distributions along the ꢀ 60^ꢀand the

ꢀ 60^ꢀlines in the Ramachandran plane (Fig. 2C).

No perturbation of the backbone conformation of

neighboring residues was observed.

Among the proline-containing peptides, the case of

SFPLR-NH₂is quite interesting because this peptide is

one of the few constrained mimetics with functional

activity. As expected, the proline residue has a ꢀ dihe-

dral angle centered around 60^ꢀ(Fig. 2d). In addition,

the prolines, just as the N-methyl groups, in¯uence the

M. A. Ceruso et al. / Bioorg. Med. Chem. 7 (1999) 2353±2371

2357

residue preceding them (in this case Phe-2), con®ning

the backbone conformation to the same upper left-hand

quadrant of the Ramachandran map (Fig. 2B). The

dierences in the maps of the residue in third position

(P3) in SFPLR-NH₂and SF-NMeL-LR-NH₂shed

some light on the respective biological activity of these

two peptides, both of which contain a secondary amide

linkage between P2 and P3.

To ®nd a basis for selecting the bioactive range, we

elected to perform computations on two recently

reported²⁸b-methylphenylalanine-containing peptides,

namely [(3S)-3-Me-Phe²]TRAP-5 (``erythro'') and [(3R)-

3-Me-Phe²]TRAP-5 (``threo''). These isomers present an

approximately 100-fold dierence in functional activity,

with the threo analogue having virtually the same activ-

ity as TRAP-5.²⁸

On the basis of these results, a three-dimensional picture

of TRAP-5's backbone, in its functional conformation,

began to emerge. The main-chain dihedral angles of the

Phe residue and the ꢀ torsion angle of the third residue

(Leu-3) had been circumscribed (Fig. 2B and 2D). The

remaining dihedral angle, , at P3 was still unde®ned.

For all searches, the Ramachandran maps of individual

residues in the erythro peptide and threo peptide were

essentially identical, indicating that the local backbone

folds of the two peptides are very similar (Fig. 3). The

features of these maps agree with the results from the

other TRAP-5 analogues. The skewed distributions and

Figure 1. Validation of the search protocol. (a) Displacement autocorrelation function obtained for the conformational search of TRAP-5: mean-

square atomic displacement (MSD) in time. (b) Distribution of end-to end distances in the case of SFPLR-NH₂.

2358

M. A. Ceruso et al. / Bioorg. Med. Chem. 7 (1999) 2353±2371

Figure 2. Typical phi-psi maps of several locally constrained residues: N-Me-Leu-3 (A) and Phe-2 (B) within SF-(NMeLeu)-LR-NH₂; Aib-3 (C)

within SF-Aib-LR-NH₂; Pro-3 (D) within SFPLR-NH₂.

mean values of the end-to-end distances indicate that

both peptides preferentially adopt extended conforma-

tions (Table 3), but more compact conformations were

also observed, as for SFPLR-NH₂(Fig. 1b). In general,

folded/compact structures occurred through formation

group of Arg-5. On the other hand, the mean distance

between the phenyl ring and the isopropyl group of

Leu-4 was found to be signi®cantly dierent (at the 95%

con®dence level). The threo peptide displays a stronger

contact between these two side chains than the erythro

peptide. The ꢁ₁dihedral angle distributions at P2 (b-

methylphenylalanine) showed that the threo peptide

adopts mostly anti (180^ꢀ) and gauche (+60^ꢀ) con®gura-

tions, while the erythro peptide principally adopts

gauche, g⁺( 60^ꢀ) and g (+60^ꢀ) orientations. This

can serve to explain the greater proximity of the

phenyl ring to the isopropyl group of Leu-4 in the

threo peptide.

of a bend at the C-terminus of the peptides (i.e. <d_2,5

was less than <d_1,4>).

>

``Folded'' backbone conformations can be identi®ed by

the distance (d_i;i3) between two Ca carbons that are

three residues apart: a value of d_i;i3of less than 7 A

indicates the presence of a bend.⁴⁰Comparison of the

means, by using a Student's t-test, indicated that the

erythro peptide tended to adopt a more compact

conformation at its C-terminus than the threo peptide:

erythro <d_2,5> less than threo <d_2,5>. The opposite

was true when comparing the N-terminus folds: threo

<d_1,4> less than erythro <d_1,4>. To gain insight into

the structural characteristics of the threo and erythro

peptides, three side-chain-to-side-chain distances were

monitored (Table 3). The average distance between the

geometric centers of the phenyl ring (at P2) and the

guanidinium group (at P5) are statistically similar (95%

con®dence level) for both diastereomers. The same con-

clusion was reached for the average distance between

the isopropyl group of Leu-4 and the guanidinium

To visualize these interactions, the energy-®ltered con-

®gurations (within 10 kcal/mol of the energy minimum)

of the two peptides were superimposed along the back-

bone atoms of the three central residues: b-MePhe-2,

Leu-3, and Leu-4. Hierarchical cluster analysis (average

linkage method) was used to classify the con®gurations

of each peptide into distinct families. The results for the

erythro and threo peptides are comparable. In each case,

two major families are distinguishable (Figs 4 and 5),

and these families adopt similar backbone conforma-

tions across enantiomers (Figs 6 and 7). For each pep-

tide, the dierence between the two principal families

M. A. Ceruso et al. / Bioorg. Med. Chem. 7 (1999) 2353±2371

2359

Figure 3. Conformational preferences of individual residues in the b-methylphenylalnine-containing peptides. Typical plots of the conformational

space sampled during simulation of the ``erythro'' (top) and ``threo'' peptides (bottom). Only maps for residues 2±5 are shown (residue number above

plot).

Table 3. Characteristic distances for the ``threo'' and ``erythro'' pep-

tides^a

the side chain of Arg-5 was very ¯exible, without pre-

ferential orientation with respect to the rest of the chain

(Figs 4 and 5).

Distances

``Erythro'' form

``Threo'' form

End-to-end distance

11.66 (2.59)

8.94 (1.07)

7.14 (1.60)

8.69 (2.21)

8.67 (3.59)

9.46 (2.28)

11.78 (3.01)

8.62 (1.41)

7.64 (1.54)

7.74 (2.16)

8.75 (3.28)

9.37 (2.12)

Presently, in terms of favored bioactive conformations

for TRAP-5, the data highlight the importance of the

carbonyl groups as hydrogen-bond acceptors, whether

they are involved in an intermolecular (ligand-receptor)

or an intramolecular sense. Independently of how

folded or extended TRAP-5 may be, the spatial place-

ment of the Phe-2 and Leu-4 side chains, which is governed

by the local conformation of Leu-3, appears to be a key

factor for determining biological activity.

C1_a-C4_a: <d_1,4

C2_a-C5_a: <d_2,5

>

Side chain-2<-> Side chain-4

Side chain-2<-> Side chain-5

Side chain-4<-> Side chain-5

a

Average distance in A with standard deviation in parentheses.

resides mostly in the local backbone conformation of

the third residue in the chain (i.e. Leu-3). In the ®rst

family, Leu-3 adopts an a-helical fold, while in the sec-

Spectroscopic studies

eq

ond family, Leu-3 adopts a more extended C₇

(fꢁ 80^ꢀ, Éꢁ80^ꢀ) conformation, essentially an inverse

g-turn, seven-membered ring structure.^41±43In both

families, the second residue assumes a C₅(fꢁ 150^ꢀ,

É=+150^ꢀ), extended conformation, the fourth residue

assumes a C₇^eqconformation, and the last residue (Arg-

5) does not appear to favor any particular orientation

(i.e. all allowed regions of the Ramachandran map are

explored almost equally).

To test the validity of the above model and to discern

the local conformation at Leu-3, physical experiments

were conducted. Because short linear peptides are

extremely ¯exible in aqueous solution, the conforma-

tional properties were studied in the presence of 2,2,2-

tri¯uoroethanol (TFE). It was hoped that this more

acidic, polar solvent would stabilize transiently ordered

structures (e.g. a turn at Leu-3) by competing less eec-

tively with intramolecular hydrogen bonding.

In summary, both isomers were found to exist mainly in

extended-type conformations stabilized by at least one

backbone hydrogen bond, involving the last three resi-

dues in the chain as well as the C-terminal amide group.

More compact backbone folds of the C-terminal resi-

dues were detected in the case of the virtually inactive

erythro peptide,²⁸while the bioactive threo peptide²⁸

showed close hydrophobic contacts between the P2 and

P4 side chains. The organization of the core residues

Phe-2, Leu-3, and Leu-4 was mostly determined by the

local conformation of Leu-3. The N-terminal serine was

not a structural determinant for the core residues and

Circular dichroism (CD) is a powerful tool for mon-

itoring conformational changes and estimating second-

ary structures in peptides and proteins. The CD

absorption of TRAP-5 showed that, on titration with

TFE, the peptide was capable of populating nonran-

dom-coil conformations (Fig. 8A), as particularly

re¯ected by an intense positive absorption band around

190 nm. No eort was made to interpret the spectrum

further in terms of speci®c secondary structure, but

it was important to establish that under adequate

conditions TRAP-5 could exist in disordered and

2360

M. A. Ceruso et al. / Bioorg. Med. Chem. 7 (1999) 2353±2371

Figure 4. Stereoviews of the most populated backbone conformations of the ``threo'' peptide: (A) second family: backbone-only representation

(top), backbone plus side chains of residues at P2 and P4 (bottom); (B) ®fth family: corresponding views.

ordered (not necessarily with a bend in the sequence)

conformations, as the computer simulations suggested.

(1A/B, 2A/B), showed some anity for PAR-1, it did

not activate the receptor nor act as an antagonist. These

subunits introduce one or two additional carbons into

the backbone chain relative to the TRAP-5, which

probably separates the P2 and P4 side chains at too

great a distance to permit agonist action.

To unveil the hydrogen-bonding patterns responsible

for the stability of such conformations, all ester and

reduced-amide derivatives were examined under similar

conditions (Fig. 8B and C). Substantial changes in the

CD absorption for TRAP-5 occurred when either cen-

tral amide bond, É_2,3or É_3,4, was replaced with an ester

or reduced-amide linkage. Besides changes at shorter

wavelength, signi®cant variations in the sign and inten-

sity of the Cotton eect at longer wavelengths (i.e. at ca.

230 nm) were observed. Such patterns have been shown

to re¯ect the handedness of the turn or helical structure

in a peptide.^44,45Conversely, replacement of the ®rst or

last amide bond of TRAP-5, É_1,2or É_4,5, led to much

less dramatic changes in the CD absorption spectra of

those peptides. These data con®rm the importance of

the central residue, Leu-3, for the three-dimensional

organization and biological activity of the agonist

peptides.

Compounds 1A, 2 and 3 (Table 4) and TRAP-5 were

studied by ¹H NMR in 50% TFE at 278 K. As indicated

3

by the J_HN-Hacoupling constants of ca. 7 Hz and the

amide temperature coecients in the range of 6 and

8 ppb/K, all of the peptides showed substantial confor-

mational averaging. In 1A and TRAP-5 there was a

nuclear Overhauser eect (NOE) contact between the

side chains of residues 2 and 4. This result is interesting

because, despite the presence of two extra carbons in the

backbone, 1A possesses sucient conformational ¯ex-

ibility to establish the native NOE contact seen in

TRAP-5 (which is absent with the 1,2-disubstituted

case). However, this capacity of 1A was not sucient to

elicit the native biological response.

Rigidi®ed peptides 1±5 were synthesized in an attempt

to probe the role of the P3 residue in TRAP-5's struc-

ture (Table 4). These peptides, which contain cyclic

amino acids with 1,2- and 1,3-disubstituted rings

(1±5), were not very biologically active as agonists, or

antagonists. Although 1A, one of the four peptide

diastereomers with the 1,3-substituted cyclohexyl ring

To gauge the signi®cance of the contact between Phe-2

and Leu-4, we can consider TRAP-5 analogues with the

native leucine at position 4 mutated to valine, alanine,

or glycine. The platelet activation EC₅₀values for

SFLVR-NH₂(L4V), SFLAR-NH₂(L4A), and SFLGR-

NH₂(L4G) were determined to be 130, 3.4, and 120 mM,

respectively, compared with 0.49 mM for TRAP-5, while

M. A. Ceruso et al. / Bioorg. Med. Chem. 7 (1999) 2353±2371

2361

Figure 5. Stereoviews of the two most populated backbone conformations of the ``erthro'' peptide: (A) third family: backbone-only representation

(top), backbone plus side chains of residues at P2 and P4 (bottom); (B) ®fth family: corresponding views.

the receptor binding IC₅₀values were 9.3, 1.8, and

7.4 mM, respectively, compared with 1.5 mM for TRAP-

5. The functional activity was attenuated for all three P4

mutants, to the extent of 25-fold, 7-fold, and 25-fold,

respectively. However, the receptor binding was atte-

nuated for just two of the P4 mutants: 6-fold for L4V

and 5-fold for L4G. The loss in functional activity

agrees with results reported by others,^25,26who found

that substitution with other aliphatic residues usually

gives diminished platelet agonist activity, although nor-

leucine (Nle) was slightly better than leucine. It was also

reported that aromatic residues, such as F, W and Y,

are well tolerated at position 4, lending support to the

idea of hydrophobic interaction between positions 2

and 4.²⁶

binding and signal transduction events, whereas

antagonist ligands would bind eectively but not trans-

duce a signal into the cell. For a ligand with many pos-

sible conformations in solution, it is likely that an

``induced-®t'' model of ligand-receptor interaction, in

which there is mutual conformational adjustment of

both the ligand and the receptor macromolecule, is

operative, rather than a ``lock-and-key'' model.^46,47In

the induced-®t model, the initial recognition event is

followed by a cascade of transformations where the

ligand and receptor cooperatively evolve to produce a

biologically relevant (agonist or antagonist) three-

dimensional complex. The following discussion pertains

to structure-function analysis of the topographical

(speci®c chemical groups) and geometrical (speci®c spa-

tial arrangements) features that characterize our TRAP-

5 analogues.

Overall, the peptide series that we studied from a con-

formational aspect point to a pivotal role for P3 in

controlling the backbone conformation, the orientation

of critical side chains, and the nature of the biological

response.

Polarity, orientation, and steric bulk at the N-terminus

Recently, Bernatowitz et al.⁴⁸reported the existence of

an active TRAP-5 mimic with an uncharged N-termi-

nus: Ac-b-Ala-FLLR-NH₂. This is an unusual result

since a positively charged N-terminus has generally been

presented as a key feature for agonist activity (see

Introduction). In fact, for this analogue the importance

of the additional methylene group (introduced within

Discussion

The biological activity of agonist peptide ligands inter-

acting with the thrombin receptor (PAR-1) results from

2362

M. A. Ceruso et al. / Bioorg. Med. Chem. 7 (1999) 2353±2371

Figure 6. Ramachandran plots of individual residues in the ``threo'' peptide (residues P2±P5 only), grouped by family clusters (columns A±F). Most

populated clusters appear in the second family (column B) and ®fth family (column E). Low population families are not presented.

the b-alanine residue) should not be underestimated

since acetylation of the N-terminus of standard agonist

peptide sequences has led to a complete loss of activ-

ity.^14,22,25,49This suggests that Ac-b-alanine analogues

are forming new interactions with the receptor that are

capable of compensating for the absence of the original

charged N-terminus, probably via the N-acetyl group.

This may explain why the shorter Ac-Ala peptide is

inactive, and why the de-acetylated version, b-Ala-

FLLR-NH₂, is inactive.⁴⁹However, it is possible to

retrieve biological activity in the native pocket with just a

polar N-terminus according to our results with É(HO)S-

FLLR-NH₂and É(HO)A-FLLR-NH₂(Table 1, entries

8 and 7). In this case, because of their similarity in size,

the N-termini are probably interacting within the native

pocket, rather than farther away. For this to occur, it

appears that there must be strict limitations on steric

bulk (see Introduction) and strict requirements on

polarity and orientation. With respect to alteration of

the residue at the N-terminus, it is worth recounting

that proline substitution at position 1 (Table 2, entry 12)

resulted in a potent agonist, and N-Me substitution at

position 1 (Table 2, entry 2) resulted in moderated ago-

nist activity along with an antagonist component.

activity of a-methyl-substituted residues such as a-MeS-

FLLR-NH₂(Table 2) and Aib-FLLR-NH₂(data not

shown).²⁵The steric bulk introduced by the methyl

substitution could be responsible for the ineectiveness

of these analogues, but for S-É(COO)-FLLR-NH₂

(Table 1) the change in conformational preferences,

which is apparent in CD studies (Fig. 8B), might sup-

port a stereospeci®c requirement rather than a regio-

speci®c one.

Backbone requirements for recognition and activity

Substitutions involving peptide bond modi®cations had

dramatic structural (and biological) eects. From a

conformational point of view, CD studies showed the

existence of a certain parallelism between ester bond

replacements and their corresponding reduced-amide

bond homologues. That is, a modi®cation at a given

amide bond produced similar eects on replacement of

either the amide nitrogen or the carbonyl oxygen (Fig.

8B and C). In the biological studies, this correspondence

disappeared, as only one TRAP-5 analogue, SF-

É(COO)-LLR-NH₂, conserved functional activity. The

peptide bond replacements seemed to impact on dier-

ent stages of the recognition process. Designating the

NH group between residues i and j as NH_ij, the data in

Table 1 suggest that NH₁₂and NH₃₄are needed at the

inception of ligand-receptor recognition, that NH₄₅is

While the eect of steric bulk has already been descri-

bed in the literature, the orientational preferences at

position 1 start to become apparent with the lack of

M. A. Ceruso et al. / Bioorg. Med. Chem. 7 (1999) 2353±2371

2363

Figure 7. Ramachandran plots of individual residues in the ``erythro'' peptide (residues P2±P5 only), grouped by family clusters (columns A±F).

Most populated clusters appear in the third family (column C) and ®fth family (column E). Low population families are not presented.

only involved at the activation stage, and that NH₂₃is

not involved at all in ligand-receptor interaction. The

case of NH₃₄is not clear since its binding constant

adopts an intermediate value which has to be appre-

ciated in combination with those of SFL-NMeL-R-NH₂

and SFPLR-NH₂(Table 2). This peptide bond could

have partial involvement both at the recognition and

activation stages, since the lack of biological activity

here could arise from conformational rather than topo-

graphical (i.e. absence of an amide hydrogen) sources.

On the other hand, carbonyl mutation at any position

prevented recognition of the peptide by the receptor,

which is not uncommon with reduced-amide scans.^50±53

Although these reduced-amide homologues may be

expected to possess greater conformational ¯exibility,

the loss in recognition/activity could be attributed to an

unfavorably biased conformational distribution. On

protonation, the presence of a strong NH⁺hydrogen

bond with a neighboring carbonyl (ten-membered ring)

would stabilize a b-turn-like conformation,^52,53which

may preclude formation of the putative bioactive con-

formation, i.e. one with a more extended backbone.

Alternatively, one or more of these groups could be

used by the peptide to ``lock'' itself into a conformation

that would then bind to the receptor (intramolecular

role). Finally, a combination of intermolecular and

intramolecular components may also pertain. Several

pieces of information weaken the intramolecular hydro-

gen bonding interpretation. First, given the reciprocity

of hydrogen bond donors and acceptors, the fact that

all of the amine-for-amide substitutions resulted in

a loss receptor binding, whereas four of the ®ve ester-

for-amide substitutions did not (Table 1, entries 3, 5, 6,

and 7), suggests that only the NH of the Phe residue

and one of the carbonyls are candidates for intra-

molecular hydrogen bonding. However, this view is not

supported by the Phe amide ¹H NMR temperature

coecient. Secondly, the loss of the conformational/

biological correspondence between the ester and reduced-

amide peptides indicates that, although the peptides

populate similar conformational ensembles, their capa-

city to be recognized is dierent. These observations

serve to distinguish the recognition-stage conformation

of the ligand from its receptor-bound conformation.

Consequently, with the large number of backbone

interactions that are necessary, an extended conforma-

tion of the peptide appears most probable at an initial

stage of interaction with the receptor. Extended con-

formations are not uncommon for peptide ligands^54,55

and, more speci®cally, Siligardi and Drake⁵⁶have

Six out of eight NH and CO groups are required at

the inception of recognition. This could re¯ect the fact

that one or more of the NH/carbonyl groups needs to

interact directly with the receptor to stabilize the

bound conformation of TRAP-5 (intermolecular role).

2364

M. A. Ceruso et al. / Bioorg. Med. Chem. 7 (1999) 2353±2371

Table 4. Biological data for some P3-constrained analogues of

TRAP-5

Compound

Residue at P3

SF-X-LR-NH₂

Platelet

aggregation^a

EC₅₀(mM)

Receptor

binding^b

IC₅₀(mM)

1A^c

1B^c

>50 (5%)

>50 (7%)

8

>100

2A^c

2B^c

>50 (6%)

3^d

>50 (5%)

>50 (9%)

66

4^d

>100

5A^c

5B^c

>50 (40%)

>50 (39%)

96

40

a

See Table 1.

Two diastereomers, A and B, were isolated and evaluated indepen-

b

c

dently.

d

Mixture of diastereomers.

2A). Similarly, SF-aMeL-LR-NH₂is also inactive;

however, in this case steric bulk, as well as conforma-

tional constraint, could be a problem. Computational,

¹H NMR, and P4 mutation data (vide supra) suggest a

possible cause for this conformational sensitivity: the

presence of hydrophobic contact between the Phe-2

and Leu-4 side chains. This interaction is reminiscent of

a hydrophobic staple motif, which applies to these same

residues in some instances.⁵⁷This structural motif,

which has an important role in the stabilization of pep-

tide helices, is based on hydrophobic contact between

the i and i+4 residues in the helix, and the distance

between such residues is comparable to that identi®ed in

our modeling studies. However, such hydrophobic con-

tact is not a sucient condition for inducing a biological

response, as indicated by the 1,3-disubstituted cyclo-

hexyl analogue, 1A, which should be capable of the

necessary contacts but is devoid of agonist activity.

Figure 8. CD spectra of TRAP-5 (t5) and peptide-bond surrogate

É(CH₂NH) and É(COO) analogues: (A) CD of TRAP-5 in increasing

amounts of TFE; (B) comparison of TRAP-5 and É(COO) analogues

in 50% TFE; (C) comparison of TRAP-5 and É(CH₂NH) analogues

in 50% TFE. The parameters e_ijand pn_ijdesignate the pentapeptide

for which the amide bond between residues i and j was replaced,

respectively, by an ester or a secondary amine.

underscored the importance of the P_IIconformation in

molecular recognition.

In summary, an extended conformation appears to be a

probable structure for agonist peptide binding to PAR-

1 (i.e. receptor recognition), although intermolecular

hydrogen bonding between the peptide backbone and

receptor may not be uniform throughout the

sequence. Intramolecular hydrogen bonding may also

give rise to a locally folded conformation that plays a

role in ligand-induced signal transduction (i.e. receptor

activation).

Another notable trait of P3 is the intolerance for a rigid

spacer, such as a 3-aminocyclohexyl-1-carboxylic acid,

which introduces two more carbon atoms into the

backbone chain (Table 4). Also, the 1,2-disubstituted

analogues, which introduce only one such carbon, are

inactive (Table 4). In 1±5 there may be a likelihood of

reorientation of the peptide backbone, because of the

conformational constraint at position 3, to generate a

dierent type of peptide chain.

Position 3 as a key structural element

Position 3 (P3) appears to play a special role in TRAP-

5. A wide variety of side chains can be tolerated in this

position and we determined herein that the amide

nitrogen can be substituted by an oxygen, as in SF-

É(COO)-LLR-NH₂, without loss of bioactivity. On the

other hand, this peptide bond is sensitive to conforma-

tional changes. For example, while SFPLR-NH₂is

active, SF-NMeL-LR-NH₂is not. Our computational

studies point out the slightly dierent conformational

preferences of P3 in these two peptides (cf. Figs 2D and

In any event, we suggest that the conformational sensi-

tivity at P3 is linked with hydrophobic contact between

Phe-2 and Leu-4. The P3 region might act as a con-

formational switch to enable or disable the formation of

either the agonist complex with PAR-1 by orchestrating

M. A. Ceruso et al. / Bioorg. Med. Chem. 7 (1999) 2353±2371

2365

an interaction between Phe-2 and Leu-4, and possibly

involving Arg-5.

or antagonist nature of the ligand-receptor complex and

thus be useful in the design of novel agonist or antago-

nist molecules for the thrombin receptor (PAR-1).

Conclusion

Experimental

In addition to the well-known structural features of

PAR-1 agonist peptides, this work has presented an

analysis of the topographical requirements of TRAPs in

conjunction with PAR-1 binding and functional activ-

ity. The absolute requirement of a positive N-terminal

charge of NH⁺₃to exhibit strong agonist activity was

contradicted, consistent with another recent result in the

literature,⁴⁸as a hydroxyl group can suce. The amide

nitrogen between residues 1 and 2 was identi®ed as a

key determinant for receptor recognition. Also, carbon-

yl groups along the backbone appear to be involved in

hydrogen bonding with the receptor. We elucidated

conformational and local geometrical features, leading

to our proposal that an extended form of the agonist

peptide is required for receptor recognition (i.e. PAR-1

binding). This conformational viewpoint generally

agrees with that proposed by Shimamoto et al.²⁸

Materials. Synthesis solvents were HPLC grade and

obtained from Fisher or VWR Scienti®c. Commercially

available reagents, unless otherwise noted, were used

without further puri®cation. Rink-amide resin was pur-

chased from Advanced ChemTech with a measured

loading of 0.46±0.67 mmol/g; 4-methylbenzhydrylamine

resin (MBHA) was purchased from Bachem Bioscience

with a loading of 0.66 mmol/g. Fmoc- and Boc-pro-

tected amino acids were purchased from Novabiochem

and Bachem Bioscience. DIC, HOBT, BOP-Cl, DMAP,

Fmoc-Cl, piperidine, TFA, l-2-hydroxy-isocaproic

acid, l-3-phenyllactic acid, and calcium l-glycerate

were purchased from Aldrich Chemical Co. l-Argininic

acid and a-methyl-d,l-leucine were purchased from

Sigma, cis-3-aminocyclohexanecarboxylic acid was pur-

chased from TCI, and l-a-methylphenylalanine was

purchased from Novabiochem. (S)-2-Amino-2-methyl-

3-hydroxypropanoic acid, cis- and trans-2-amino-1-

cyclohexanecarboxylic acids, and cis-2-amino-1-cyclo-

pentanecarboxylic acid were purchased from Acros.

Purchased amino acids that required Fmoc protection

were acylated by using Fmoc-Cl and sodium carbonate

following the procedure of Carpino and Han.⁶³l-a-

Methylarginine was prepared from rac-2-methyl-

ornithine (Aldrich) according to the method of Roeske

et al.⁶⁴The Boc-protected a-amino aldehydes were pre-

pared from the corresponding alcohols (Bachem

Bioscience) via sulfur trioxide/pyridine oxidation.⁶⁵

Hour-glass bubbler reactors were purchased from Pep-

tides International.

During the course of our studies, Matsoukas et al.⁵⁸

published a report on the conformation of SFLLR-NH₂

in dimethylsulfoxide (polar, aprotic conditions), in

which they proposed a curved (``cyclic'') backbone for

the active form of TRAP-5 because of weak NOE con-

tacts between the Arg-5 side chain and the Ser-1 and

Phe-2 residues. We found no side-chain contacts of this

kind under our experimental conditions (polar, protic),

nor in our computational searches. With SFPGR-NH₂,

where the PG sequence should encourage a b-turn geo-

metry, we did observe an NOE between Phe-2 and Arg-

5; however, this analogue failed to bind to PAR-1

(IC₅₀>100 mM).

Matsoukas et al. gleaned further support for a ``cyclic''

backbone for agonist peptides from macrocyclic analo-

gues of SFLLR-NH₂,⁵⁹since macrocyclic peptides tend

to adopt folded conformations, such as b and/or g-

turns, ®xed by intramolecular hydrogen bonding

(although this in¯uence can vary with ring size). Since a

19-membered-ring macrocyclic SFLLR analogue linked

from the P1 side chain to the C-terminus was nearly

equipotent with SFLLR-NH₂in inducing contractions

of gastric smooth muscle,^60,61they proposed that the

agonist motif approaches the receptor in a ``cyclic''

conformation with spatial proximity of the Phe-2 and

Arg-5 side chains. However, one should keep in mind

that a macrocyclic peptide can still experience an

induced conformation on complexation with a receptor,

or present a linear (extended) motif if the macrocycle is

suciently large and ¯exible.⁶²

Methods. ¹H NMR spectra for analytical purposes were

recorded on a Bruker AC-300 (300 MHz) or a Bruker

AM-400 (400 MHz) spectrometer with D₂O as solvent,

locked on the solvent, unless otherwise noted (s=sing-

let, d=doublet, t=triplet, m=multiplet, dd=doublet

of doublets, br=broadened); Me₄Si was used as internal

standard for nonaqueous solutions. Standard fast-atom

bombardment (FAB) mass spectra were obtained on a

VG 7070E high-resolution mass spectrometer and elec-

trospray (ES) mass spectra were obtained on a VG Trio

2000 mass spectrometer. Accurate mass measurements

(HRMS) were obtained by using a VG ZAB 2-SE spec-

trometer in the FAB mode. Peptide puri®cation was

accomplished by semi-preparative HPLC on a Waters

600E instrument (Waters 486 detector) with a reverse-

phase column [Waters Bondapak C18, 40Â100 mm

(three in series, 125 A, 15±20 mm), Bondapak C18,

3.9Â300 mm (100 A, 10 mm), or Delta-Pak C18,

8Â100 mm (100 A, 15 mm)], eluting with a mixture of

0.16% TFA in acetonitrile and 0.20% TFA/water in

isocratic or gradient modes.

An important aspect of our conformational studies of

the SFLLR motif is the indication of hydrophobic con-

tact between the side chains of the second and fourth

residues (P2-P4 interaction), which diers from the

P2-P5 side chain interaction proposed by Matsoukas et

al.⁵⁸The spatial arrangement of the P2 and P4 side

chains might be responsible for determining the agonist

Synthesis of conventional peptides. Conventional pep-

tides involving native amino acids and N-methyl pep-

tides (Tables 1 and 2) were synthesized by stepwise

2366

M. A. Ceruso et al. / Bioorg. Med. Chem. 7 (1999) 2353±2371

solid-phase synthesis employing Applied Biosystems

Model 431A or Model 430A peptide synthesizers with

Boc/benzyl chemistry on MBHA or PAM (phenyl-

acetamidomethyl) resins according to the manufac-

turer's instructions. The assembled peptide resins were

deprotected and the peptides were released from the

resin by using standard liquid HF procedures to aord

the crude peptides. Each peptide was puri®ed to homo-

geneity by reverse-phase HPLC on a Waters Delta-Pak

C18 column with gradient elution using acetonitrile in

water containing 0.1% tri¯uoroacetic acid. Puri®ed pep-

tides were reanalyzed by reverse-phase HPLC on a Vydac

C18 analytical column to ensure purity (>95%). FABMS

was used to con®rm the structure of each peptide.

CH₂Cl₂(3Â), DMF (3Â), and CH₂Cl₂(3Â) and dried

under an argon stream; this peptido-resin was stored at

0^ꢀC. Fmoc deprotection was repeated as above and the

resin was combined with Fmoc-Phe-OH (387 mg,

1 mmol), HOBT (150 mg, 1 mmol), and DIC (126 mg,

1 mmol) in DMF (10 mL) and agitated for 2 h. The

solution was drained o and the resin was washed with

DMF (5Â). Fmoc deprotection was repeated and the

functionalized resin was combined with Fmoc-Ser(O-t-

Bu)-OH (0.38 g, 1 mmol), HOBT (150 mg, 1 mmol), and

DIC (126 mg, 1 mmol) in DMF (15 mL) and agitated for

4 h. After washing, the resin was carefully deprotected

(to avoid reaction of the terminal amine with the ester)

with 20% piperidine in DMF (15 mL) at 23^ꢀC for 12

min and washed with DMF (3Â) and dry CH₂Cl₂(3Â),

then dried under a nitrogen stream. Peptide cleavage

was accomplished by agitation of the resin with 99%

TFA (20 mL) at 23^ꢀC for 2 h. The red solution was

drained and evaporated in vacuo to an oil, which was

triturated with ethyl ether (2Â, 25 mL) to a white solid

(140 mg). The peptide was puri®ed by reverse-phase

HPLC by using the Waters Bondapak C18 column

(40Â100 mm) and 0.16% TFA in acetonitrile:0.20%

TFA in water (30:70) as an eluent. The solution was

evaporated in vacuo and the solid was redissolved in

water. This aqueous solution was frozen and lyophilized

to aord a white amorphous solid (42mg). FAB-MS m/z

General procedure for the preparation of depsipeptides.

The esters were prepared via solid-phase synthesis

employing the Rink-amide resin and an Fmoc protocol

in the customary fashion.^66±68Piperidine in DMF (20%

v/v, 10mL) was used to cleave the Fmoc protection from

the growing peptide (0.1 mmol), and couplings between

Fmoc protected amino acids (0.3mmol) or the a-hydroxy

acid (0.3 mmol) and the peptide on resin were performed

with DIC (0.3 mmol) and HOBT (0.3 mmol) in DMF.

BOP-Cl (0.3 mmol) and DMAP⁶⁹(0.6 mmol) in CH₂Cl₂

(10 mL) was used to couple the resin-bound hydroxy

peptide (0.1 mmol) to the next Fmoc-protected amino

acid (0.3 mmol) to form the ester bond. After ®nal Fmoc

deprotection, cleavage from the resin and concomitant

arginine side-chain PMC deprotection was accomplished

with 99% TFA at 23^ꢀC to aord the desired product as

the tri¯uoroacetate salt. These materials were puri®ed by

reverse-phase semi-preparative HPLC, followed by lyo-

philization to give the depsipeptides as white solids.

1

635.5 (MH⁺); H NMR (400 MHz, CD₃OD) d 0.88±

1.00 (2 pairs of d, 12H), 1.58±1.96 (m, 10H), 2.89 (dd,

J=10.3, 12.8 Hz, 1H), 3.15±3.40 (m, 3H), 3.78±3.97 (m,

3H), 4.36±4.53 (m, 2H), 4.74 (dd, J=4.0, 8.0 Hz, 1H),

5.05 (dd, J=4.0, 9.5 Hz, 1H), 7.18±7.35 (m, 5H, arom.),

8.10 (d, exch. H), 8.40 (d, exch. H). Anal. (C₃₀H₅₀

.

N₈O₇3.0 CF₃CO₂H): C, H, N.

B. Other depsipeptides. In the same manner, the follow-

ing depsipeptides (oxygen isosteres) were prepared. (1)

É(HO)Ser-Phe-Leu-Leu-Arg-NH₂: ES-MS m/z 635.4

Examples A

A. Synthesis of H₂N-Ser-Phe-Leu-É(COO)-Leu-Arg-NH₂.

Rink-amide resin (4.98 g, 0.46 mmol/g, 2.3 mmol) was

placed in an hour-glass bubbler reactor, washed twice

with DMF (30 mL) and agitated (nitrogen) with 20%

piperidine in DMF (30 mL) for 1 h at 23^ꢀC and the

solution drawn o. The resin was washed with fresh

DMF (4Â), CH₂Cl₂(3Â), and DMF (3Â). Fmoc-Arg-

(PMC)-OH (5.36 g, 7 mmol), HOBT (1.2 g, 7.8 mmol)

and DIC (0.88 g, 7 mmol) were added and slurried with

DMF (30 mL) and the mixture was agitated for 16 h.

The solution was drawn o and the resin was washed

with DMF (4Â), CH₂Cl₂(4Â) and dried in vacuo at

25^ꢀC for 24 h. The Fmoc arginine-loaded resin (0.75 g,

0.33 mmol) was deprotected with 20% piperidine in

DMF (15 mL) and washed as above, then combined

with l-2-hydroxyisocaproic acid (132 mg, 1 mmol),

HOBT (150 mg, 1 mmol), and DIC (126 mg, 1 mmol) in

DMF (10 mL) and agitated for 3 h. After the solution

was removed, the resin was washed with DMF (4Â) and

dry (4A sieves) CH₂Cl₂(4Â) and then combined with

Fmoc-Leu-OH (370 mg, 1.05 mmol) and DMAP (0.26 g,

2.1 mmol) in CH₂Cl₂(10 mL) and agitated (with argon)

as BOP-Cl (0.26 g, 1 mmol) was added to the mixture.

Agitation was continued for 7 h at 23^ꢀC with periodic

addition of dry CH₂Cl₂to replace evaporated solvent.

Solution was drained o and the resin was washed with

1

(MH⁺), H NMR (400 MHz, DMSO-d₆); (2) É(HO)

Ala-Phe-Leu-Leu-Arg-NH₂: ES-MS m/z 619.6 (MH⁺),

¹H NMR (400 MHz); (3) H₂N-Ser-É(COO)-Phe-Leu-

Leu-Arg-NH₂: ES-MS m/z 635.4 (MH⁺), ¹H NMR

(300 MHz, DMSO-d₆); (4) H₂N-Ser-Phe-É(COO)-Leu-

1

Leu-Arg-NH₂: FAB±MS m/z 635.5 (MH⁺), H NMR

(400 MHz, CD₃OD); (5) H₂N-Ser-Phe-Leu-Leu-É(COO)-

Arg-NH₂: ES-MS m/z 635.4 (MH⁺), ¹H NMR (400

MHz, DMSO-d₆).⁷⁰

General procedure for the preparation of ꢀ-methyl-sub-

stituted peptides. The peptides were prepared via solid-

phase synthesis on the Rink-amide resin as described

above for the depsipeptides. However, when coupling to

the a-methyl amino acid or the a-methyl peptide on the

solid phase, the BOP-Cl reaction was repeated to

insure complete coupling, on account of increased

steric hindrance.

Examples

A. Synthesis of H₂N-Ser-(ꢀ-Me)Phe-Leu-Leu-Arg-NH₂.

Fmoc-arginine(PMC)-functionalized Rink-amide resin,

prepared above, (1 g, 0.46 mmol) was placed in an

hour-glass bubbler with 20% piperidine in DMF

M. A. Ceruso et al. / Bioorg. Med. Chem. 7 (1999) 2353±2371

2367

(20 mL) and agitated (nitrogen) for 1 h at 23^ꢀC. The

solution was drawn o and the resin was washed with

DMF (4Â), CH₂Cl₂(2Â), and DMF (3Â), then com-

bined with Fmoc-Leu-OH (480 mg, 1.35 mmol), HOBT

(200 mg, 1.35 mmol), and DIC (170 mg, 1.35 mmol) in

DMF (20 mL) and agitated for 3 h. After washing, the

steps of deprotection and coupling with Fmoc-Leu-OH

were repeated, followed by deprotection with 20%

piperidine in DMF (20 mL) for 1 h. The tripeptido-resin

was washed with DMF (5Â) and CH₂Cl₂(3Â), dried

under a nitrogen stream, and stored at 0^ꢀC. Half of

this resin (0.50 g, 0.23 mmol) was combined with Fmoc-

(S)-2-amino-2-methyl-3-phenylpropionic acid (0.44 g,

1 mmol), DMAP (0.25 g, 2 mmol) in CH₂Cl₂(25 mL)

and BOP-Cl (0.26 g, 1 mmol) were added, and it was

agitated for 5 h at 23^ꢀC with periodic replacement of the

evaporated solvent. The solution was drained o and

the resin was washed with CH₂Cl₂(4Â), DMF (3Â) and

CH₂Cl₂(3Â), then resubjected to the same reaction

procedure to insure complete coupling. Removal of the

Fmoc group with 20% piperidine in DMF (20 mL) was

followed by washing with DMF (5Â) and CH₂Cl₂(5Â).

The resin was combined with Fmoc-Ser(t-Bu)-OH

(0.42 g, 1.1 mmol), DMAP (0.24 g, 2 mmol), and

BOP-Cl (0.26 g, 1 mmol) in CH₂Cl₂(20 mL) and agi-

tated at 23^ꢀC for 5 h. After washing, the resin was

resubjected to the same reaction conditions, followed by

washing with CH₂Cl₂(3Â), DMF (3Â), CH₂Cl₂(3Â),

and DMF (2Â). The Fmoc group was removed with

20% piperidine in DMF (20 mL) and the resin was

washed with DMF (3Â) and CH₂Cl₂(3Â) and dried

under a nitrogen stream. TFA (99%, 25 mL) was added

and the mixture agitated for 3 h; the red solution was

collected and evaporated in vacuo to an oil, which was

triturated with ether (25 mL) to obtain a white solid

(90 mg). The crude peptide was puri®ed by reverse-

phase HPLC by using the Waters Bondapak C18 col-

umn (40Â100 mm) and 0.16% TFA in acetonitrile:

0.2% TFA in water (30:70). The solution was evapo-

rated in vacuo to a solid, which was dissolved in water.

This aqueous solution was frozen and lyophilized to

give the titled compound (24 mg). ES-MS m/z 648.6

and 1,4-dioxane (15 mL) was added. (S)-2-Amino-2-

methyl-3-phenylpropionic acid (1.02 g, 5.69 mmol) was

added and the solution was cooled to 0^ꢀC. 9-Fluor-

enylmethylchloroformate (1.47 g, 5.69 mmol) in 1,4-

dioxane (15 mL) was added over 10 min and the solu-

tion was stirred at 0^ꢀC for 30 min, then at 23^ꢀC for 16 h.

The reaction was diluted with water, extracted with

ether (2Â), and acidi®ed with aqueous HCl (pH=1).

The milky solution was extracted with chloroform:

2-propanol (4:1, 2Â). The organic solution was washed

with brine (2Â), dried (Na₂SO₄), and evaporated in

vacuo to a clear oil (2.22 g), which was used without

further puri®cation. FABMS m/z 402.3 (MH⁺); ¹H

NMR (300 MHz, DMSO-d₆) d 1.17 (s, Me), 2.93 (d,

J=13.1 Hz, 1H), 3.26 (d, J=13.3 Hz, 1H), 4.25±4.40 (m,

2H), 4.51 (dd, J=6.2, 10.2 Hz, 1H), 7.00±7.45 (m, 9H,

arom.), 7.73 (d, J=7.5 Hz, 2H), 7.91 (d, J=7.3 Hz, 2H).

Synthesis of Boc-Ser(Bz)-CHO.⁶⁵Triethylamine (1.58 g,

15.6 mmol) was added to Boc-serinol(OBz) (1.1 g,

3.9 mmol) in dry DMSO (10 mL) under argon, fol-

lowed by slow addition (10 min) of a DMSO (15 mL)

solution of sulfur trioxide:pyridine complex (2.5 g,

15.6 mmol) at 23^ꢀC. After stirring for 3 h, the reaction

mixture was poured into water (300 mL) and extracted

with ether (125 mL, 3Â). The ether solution was washed

with 10% citric acid (2Â), water (2Â), saturated

NaHCO₃(2Â), then dried (MgSO₄) and evaporated in

vacuo to give a clear oil (1.04 g), which was used with-

1

out further puri®cation. ES-MS m/z 280.2 (MH⁺); H

NMR (300 MHz, CDCl₃) d 1.48 (s, Me, 9 H), 3.68 (dd,

J=12, 2.0 Hz; 1H), 4.0 (dd, J=12, 1.5 Hz; 1H), 4.31 (m,

1H), 4.51 (br s, CH₂Ph), 4.56 (m, CH), 7.22±7.40 (m,

5H, arom.), 9.65 (s, CHO).

General procedure for the preparation of reduced-amide

peptides. These peptides were prepared on the MBHA

resin by using a classical Boc protocol. A solution of

TFA:CH₂Cl₂(1:1, 15 mL) containing DMS (2%) was

used to deprotect the Boc group from the growing pep-

tide (0.33 mmol) followed by CH₂Cl₂washing (6Â),

then a CH₂Cl₂:triethylamine (10:1, 15 mL) wash (2Â),

again followed by CH₂Cl₂washing (6Â). Coupling with

the next Boc-protected amino acid (1 mmol) was

achieved with DIC (1 mmol) and HOBT (1 mmol) in

DMF:CH₂Cl₂(1:1, 10 mL). Reductive amination of the

aminopeptide on resin with the Boc-protected l-amino

aldehyde (1.33 mmol) employed the method of Sasaki

and Coy⁷¹by using NaBH₃CN (1.33 mmol) in DMF:

DCE (1:2). The product was cleaved from the resin,

along with side-chain deprotection, by using liquid HF:

anisole (10:1) at 0^ꢀC for 4 h. After evaporation of the

HF, the mixture was triturated with ether and the

product was dissolved in methanol. The solution was

evaporated to a white solid, which was puri®ed by reverse-

phase HPLC and lyophilized to aord desired target.

1

(MH⁺); H NMR (400 MHz, CD₃OD) d 0.89±1.00 (2

pairs of d, 12H), 1.48 (s, Me, 2.7H), 1.40±1.95 (m, 11H),

3.09 (d, J=13.4 Hz, 1H), 3.18 (m, 3H), 3.89 (dd, J=3.9,

11.1 Hz, 1H), 3.97 (m, 2H), 4.15 (m, 1H), 4.30±4.45 (m,

3H), 7.15±7.33 (m, 5H, arom.), 7.70 (d, exch. H), 7.80

(d, exch. H), 8.10 (d, exch. H). FAB-HRMS calcd for

C₃₁H₅₃N₉O₆+H⁺, 648.4197; found, 648.4211.

B. Other ꢀ-methyl-substituted peptides. In the same

manner, the following a-methyl peptides were prepared.

(1) H₂N-(a-Me)Ser-Phe-Leu-Leu-Arg-NH₂: ES-MS m/z

1

648.6 (MH⁺), H NMR (400 MHz, CD₃OD); (2) H₂N-

Ser-Phe-(a-Me)Leu-Leu-Arg-NH₂: ES-MS m/z 648.3

(MH⁺), ¹H NMR (300 MHz); (3) H₂N-Ser-Phe-Leu-(a-

Me)Leu-Arg-NH₂: ES-MS m/z 648.6 (MH⁺), ¹H NMR

(300 MHz); (4) H₂N-Ser-Phe-Leu-Leu-(a-Me)Arg-NH₂:

FAB-MS m/z 648.7 (MH⁺), ¹H NMR (400 MHz,

DMSO-d₆).⁷⁰

Examples

A. Synthesis of H₂N-Ser-Phe-(CH₂NH)-Leu-Leu-Arg-

NH₂. MBHA resin (1g, 0.66mmol) was washed with

CH₂Cl₂(15 mL, 2Â) and combined with Boc-Arg(Mts)-

OH (0.9 g, 2 mmol) and HOBT (0.31 g, 2 mmol) in

Synthesis of Fmoc-(ꢀ-Me)-Phe-OH.⁶³Sodium carbon-

ate (1.81 g, 17.07 mmol) was dissolved in water (30 mL)

2368

M. A. Ceruso et al. / Bioorg. Med. Chem. 7 (1999) 2353±2371

DMF:CH₂Cl₂(1:1, 30 mL). The DIC (0.25 g, 2 mmol)

was added, the mixture was agitated for 2 h (negative

ninhydrin test), and the solution was drained. The resin

was washed with CH₂Cl₂(6Â, 30 mL), deprotected by

treating it twice with TFA:CH₂Cl₂:DMS (49:49:2,

30 mL, 2 min/30 min), washed with CH₂Cl₂(6Â,

30 mL), washed twice with CH₂Cl₂:triethylamine (10:1,

30 mL, 2 min), and washed with CH₂Cl₂(6Â, 30 mL).

The resin was combined with Boc-Leu-OH (0.50 g,

2 mmol), and HOBT (0.31 g, 2 mmol) in DMF:CH₂Cl₂

(1:1, 30 mL) followed by DIC (0.25 g, 2 mmol) and agi-

tated for 2 h at 23^ꢀC. As above, the resin was washed,

deprotected, washed, and resubjected to the same cou-

pling conditions to add the second leucine. A portion of

the resin (0.25 g, 0.165 mmol) was deprotected and

washed as above and combined with Boc-Phe-CHO²⁹

(165 mg, 0.66 mmol) in DMF:DCE (1:2, 5 mL) contain-

ing 1% HOAc and agitated for 1 h. NaBH₃CN (42 mg,

0.67 mmol) was added in small portions over the next

1 h and agitated at 23^ꢀC for 16 h. The solution was

removed and the resin was washed with DMF (6Â) and

CH₂Cl₂(6Â). The Boc group was cleaved in the usual

manner and, following the washings, the resin was cou-

pled to Boc-Ser-OH (102 mg, 0.5 mmol) with HOBT

(79 mg, 0.5 mmol) and DIC (93 mg, 0.75 mmol) as

above. The resin was washed with DMF (6Â) and

CH₂Cl₂(6Â), and dried in vacuo. The resin was com-

bined with liquid HF:anisole (10:1, 10 mL) at 0^ꢀC and

stirred for 4 h. The HF was evaporated and the residue

was diluted with ether (20 mL) and stored at 0^ꢀC for

16 h. The resin and solids were collected, mixed with

methanol, and ®ltered. The ®ltrate was evaporated to a

solid, which was puri®ed by using the Bondapak C18

3.9Â300 mm HPLC column with 0.16% TFA in aceto-

nitrile:0.20% TFA in water (35:65) to yield the title

compound (3.3 mg). ES-MS m/z 620.6 (MH⁺); ¹H

NMR (300 MHz) d 0.84 (d, J=6.0 Hz, 3H), 0.89 (d,

J=6.0 Hz, 3H), 0.96 (m, 6H), 1.30±1.95 (m, 10H), 2.75±

2.95 (m, 2H), 3.15 (m, 2H), 3.28 (dd, J=7.0, 7.0 Hz,

2H), 3.82±4.18 (m, 5H), 4.42 (dd, J=7.0, 7.0 Hz, 1H),

4.69 (dd, J=8.0, 8.0 Hz, 1H), 7.25±7.48 (m, 5H, arom.).

FAB-HRMS calcd for C₃₀H₅₃N₉O₅+H⁺, 620.4248;

found, 620.4218.

(20 mL) for 1 h. The resin was washed [DMF (5Â);

CH₂Cl₂(5Â)] and coupled with Fmoc-Phe-OH (426 mg,

1.1 mmol), DMAP (244 mg, 2 mmol) and BOP-Cl

(255 mg, 1 mmol) in CH₂Cl₂(20 mL), as above. Depro-

tection of the Fmoc group and the washings were

repeated. The ®nal coupling was accomplished with

Fmoc-Ser(O-t-Bu)-OH (230 mg, 0.6 mmol), HOBT

(95 mg, 0.6 mmol), and DIC (76 mg, 0.6 mmol) in DMF

(20 mL) for 3 h. The resin was washed with DMF (5Â),

deprotected with 20% piperidine:DMF (45 min),

washed with DMF (5Â) and CH₂Cl₂(5Â), and dried

in vacuo. The product was cleaved with TFA:anisole

(99:1, 20 mL) and the collected solution was con-

centrated in vacuo to a residue, which was triturated

with ether and dissolved in water (20 mL). This solution

was frozen and lyophilized to white solid. The crude

product was puri®ed on a Bondapak C18 (3.9Â300 mm)

HPLC column with 0.16% TFA in acetonitrile:0.2%

TFA in water (gradient: 10:90 to 80:20 over 40 min) as

the eluant to give a white solid (38 mg). ES-MS m/z

1

632.6 (MH⁺); H NMR (400 MHz, CD₃OD) d 0.92 (d,

J=6.0 Hz, 3H), 0.95 (d, J=6.0 Hz, 3H), 1.30 (m, 1H),

1.48±1.96 (m, 12H), 2.94 (dd, J=8, 12 Hz, 1H), 3.0±3.25

(m, 3H), 4.22±4.48 (m, 6H), 4.60 (m, 1H), 7.22±7.40 (m,

5H, arom.). FAB-HRMS calcd for C₃₀H₄₉N₉O₆+H⁺,

632.3884; found, 632.3877.

Synthesis of peptides 1, 2, 4, and 5. These analogues

were prepared the same manner as for 3. (1) H₂N-Ser-

Phe-(cis-3-aminocyclohexylcarboxy)-Leu-Arg-NH₂(1)

from

Fmoc-protected

cis-3-amino-1-cyclohexane-

carboxylic acid [HPLC: Waters Bondapak C18

(40Â100 mm), 0.16% TFA in acetonitrile:0.2% TFA in

water (22:78)]: diastereomer A, t_R=28.0±30.5 min, ES-

MS m/z 646.3 (MH⁺), ¹H NMR (400 MHz); dia-

stereomer B, t_R=31.0±36.0 min), ES-MS m/z 646.3

(MH⁺), ¹H NMR (400 MHz).⁷⁰(2) H₂N-Ser-Phe-

(trans-3-aminocyclohexyl-carboxy)-Leu-Arg-NH₂

(2)

from Fmoc-protected trans-3-amino-1-cyclohexane-

carboxylic acid⁷²[HPLC: Waters Bondapak C18

(40Â100 mm), 0.16% TFA in acetonitrile:0.2% TFA in

water (25:75)]: diastereomer A, t_R=20.5±24.0 min, ES-

1

MS m/z 646.3 (MH⁺), H NMR (400 MHz); diastereo-

mer B, t_R=37.2±41.1 min, ES-MS m/z 646.3 (MH⁺),

¹H NMR (400 MHz).⁷⁰(3) H₂N-Ser-Phe-(trans-2-ami-

nocyclohexylcarboxy)-Leu-Arg-NH₂(4) from Fmoc-

protected trans-2-amino-1-cyclohexanecarboxylic acid:

mixture of diastereomers, ES-MS m/z 646.6 (MH⁺), ¹H

NMR (300 MHz).⁷⁰(4) H₂N-Ser-Phe-(cis-2-aminocyclo-

hexylcarboxy)-Leu-Arg-NH₂(5) from Fmoc-protected

cis-2-amino-1-cyclohexanecarboxylic acid: diastereomer

A, HPLC t_R=17.0±18.5 min, ES-MS m/z 646.6 (MH⁺),

¹H NMR (300 MHz); diastereomer B, HPLC t_R=

21.0±22.5 min, ES-MS m/z 646.6 (MH⁺), ¹H NMR

(300 MHz).⁷⁰

B. Other reduced-amide peptides. In the same manner,

the following reduced-amide peptides were synthesized.

(1) H₂N-Ser-É(CH₂NH)-Phe-Leu-Leu-Arg-NH₂: ES-

1

MS m/z 620.5 (MH⁺), H NMR (300 MHz); (2) H₂N-

Ser-Phe-Leu-É(CH₂NH)-Leu-Arg-NH₂: ES-MS m/z

1

620.6 (MH⁺), H NMR (300 MHz); (3) H₂N-Ser-Phe-

Leu-Leu-É(CH₂NH)-Arg-NH₂: ES-MS m/z 620.6

1

(MH⁺), H NMR (300 MHz).⁷⁰

Synthesis of H₂N-Ser-Phe-(cis-2-aminocyclopentylcar-

boxy)-Leu-Arg-NH₂(3). The H₂N-Leu-Arg(PMC)-

functionalized Rink-amide resin (0.45 g, 0.2 mmol), pre-

pared above, was placed in an hour-glass bubbler with

dry CH₂Cl₂(20 mL), DMAP (171 mg, 1.45 mmol), and

Fmoc-protected cis-2-amino-1-cyclopentanecarboxylic

acid (245 mg, 0.7 mmol) and agitated while BOC-Cl

(153, 0.6 mmol) was added. After 5 h, the solution was

separated. The resin was washed [CH₂Cl₂(5Â); DMF

(5Â)] and deprotected with 20% piperidine:DMF

Gel-®ltered platelet aggregation agonist activity. Human

platelet-rich plasma (PRP) concentrate (Biological Spe-

cialties, Inc) was gel ®ltered (Sepharose 2B, Pharmacia)

in Tyrode's buer (140 mM NaCl, 2.7 mM KCl, 12 mM

NaHCO₃, 0.76 mM Na₂HPO₄, 5.5 mM dextrose, 5 mM

Hepes, and 2 mg/ml bovine serum albumen at pH 7.4).

The gel-®ltered platelets were diluted with Tyrode's

M. A. Ceruso et al. / Bioorg. Med. Chem. 7 (1999) 2353±2371

2369

buer (143,000 platelets/mL, ®nal platelet count per

well) and 2 mM CaCl₂in a 96-well microtiter plate.

Platelet aggregation was initiated by the addition of a

®xed concentration of the test compound dissolved in

buer or by a concentration of human a-thrombin

(American Diagnostica) sucient to achieve 80%

aggregation (0.015±0.075 NIH units/mL). The assay

plate was stirred constantly. Platelet aggregation was

monitored by intermittently placing the plate in a

microplate reader (Molecular Devices) to read optical

density (650 nm, DSOFT) at 0 and 5 min after thrombin

addition. Aggregation was calculated to be the decrease

in optical density between 0 and 5-min time measure-

ments. The percentage aggregation for each test com-

pound relative to thrombin was determined by dividing

the aggregation of the test sample by the aggregation of

thrombin (maximal response) and multiplying the result

by 100. All samples were tested in duplicate wells on the

same plate. For EC₅₀determinations, a percentage of

aggregation relative to thrombin was used.

Hepes/138 mM NaCl (pH 7.5) and immediate ®ltration

on Whatman GF/C ®lters, followed by four washes in a

Brandel harvester (Brandel, Gaithersburg, MD).

Radioactivity bound to the membranes on the ®lters

was measured in a scintillation counter. Speci®c binding

was determined by subtraction of nonspeci®c binding

from total binding. The GF/C ®lters were presoaked for

2±3 h in 10 mM Hepes containing 0.5% poly-

ethylenimine and 0.1 M N-acetylglucosamine to reduce

nonspeci®c binding. Competition binding data were

analyzed by using the Prism software (GraphPad, San

Diego, CA).

Circular dichroism studies. Spectra were obtained using

a Jasco model J-710 instrument calibrated with a stan-

dard solution of ammonium d-(+)-camphorsulfonate

at 192 nm. Spectra were measured from 250 to 190 nm

at 0.2-nm intervals. A scan speed of 20 nm/min with a

response time of 8 s was used. Two scans were accu-

mulated per sample. The path length of the quartz cell

was 0.1 cm. Samples were run at room temperature

(23^ꢀC). The peptide concentration in samples ranged

from 50 to 300 mM. The solvent (pH 7.2 phosphate

buer/TFE) baseline was recorded and subtracted from

sample spectra. The ®nal concentration in phosphate

buer was 12.5 mM in all TFE/aqueous solution mix-

tures. Peptide molar ellipticity is expressed in [Â] (deg-

cm²/dmol).

Inhibition of gel-®ltered platelet aggregation induced by

thrombin. Gel-®ltered platelets (vide supra) were diluted

with Tyrode's buer (143,000 platelets/mL, ®nal platelet

count per well), compound solution in buer, and 2 mM

CaCl₂in a 96-well microtiter plate. Platelet aggregation

was initiated by addition of a concentration of human

a-thrombin (American Diagnostica) sucient to achieve

80% aggregation (0.015±0.075 NIH units/mL). The

assay plate was stirred constantly. Platelet aggregation

was monitored and calculated as described above. All

samples were tested in duplicate wells on the same plate.

The test compound was dissolved in DMSO at a con-

centration of 10 mM, diluted to the desired concentra-

tion with deionized water, and then added to the assay

well containing platelets, prior to the initiation of

aggregation by a-thrombin. The percent inhibition of

platelet aggregation was measured by comparing aggre-

gation in the presence and absence of test compound; it

Structural NMR studies. Proton spectra were recorded

at 600 MHz on a Bruker DMX 600 spectrometer. Sam-

ples were 5±8 mg peptide dissolved in 400 mL of 100 mM

sodium phosphate buer, pH 3.8 and 400 mL of 2,2,2-

tri¯uoroethanol-d₃. Spectra were referenced by the

addition of a small amount of internal standard tri-

methylsilylpropanoate d 0.00). Assignments were made

using standard TOCSY (DIPSI-2 spin-lock) and

NOESY experiments. These spectra were recorded at

0^ꢀC using presaturation to suppress the water peak.

Spin-lock mix times were between 50 and 100 ms or less

than 20 ms for observation of only vicinal coupling

partners. Cross-relaxation mix times were between 100

and 400 ms.

was calculated using the following equation:

%

inhibition=100 {(% aggregation with test compound/

% aggregation without test compound)Â100}. IC₅₀

values are de®ned as the concentration of the test

compound that results in 50% inhibition of platelet

aggregation.

Computational methods. Molecular structures were gen-

erated in a neutral form and calculations were con-

ducted in vacuum with a dielectric constant of 1. For a

typical run, all bonded potential force constants were

scaled by 0.4, a quadratic potential was used for cova-

lent bond interactions, and no coulomb potential was

introduced. A given starting conformation was partially

minimized and heated to 1000 K. This elevated tem-

perature, coupled with force ®eld modi®cations to

facilitate bond rotations and atom ``tunneling'', con-

siderably randomized the peptide prior to quenching

and regularization. Structures were selected every pico-

second and quenched by successive minimizations, by

using a combination of steepest descent and conjugate

gradients methods. The quenched structures were reg-

ularized under a more complete force ®eld (no re-scal-

ing; Morse bond potential and cross-term potentials,

but no coulombic potential). Finally each conformation

was completely minimized with the complete force ®eld

Thrombin receptor (PAR-1) binding assay.⁷³The radio-

ligand Ser-pFPhe-Har-Leu-Har-Lys-(³H-Tyr)-NH₂was

prepared by Amersham Pharmacia Biotech by catalytic

tritiation of an iodotyrosyl-containing precursor to a

speci®c activity level of 46 Ci/mmol. Binding assays

were performed in 24-well plates in a ®nal volume of

200 mL. The assay buer consisted of 50 mM Hepes pH

7.5, 5 mM MgCl₂, and 0.1% bovine serum albumin. The

membranes of CHRF-288 megakaryocytic cells^74±76

(25 mg in 100 mL) were incubated with 10 nM of the

9

radioligand and dierent concentrations (10

to

10 ⁴M) of the various test compounds. The total bind-

ing was determined in the absence of test compounds

and nonspeci®c binding was determined in the presence

of unlabeled S-pFPhe-Har-L-Har-KY-NH₂at 10 mM.

Incubations were carried out for 30 min at 23^ꢀC. Reac-

tions were arrested by addition of 1 mL of cold 20 mM

2370

M. A. Ceruso et al. / Bioorg. Med. Chem. 7 (1999) 2353±2371

until the maximum derivative was less than 0.001 kcal/

mol/A. All calculations were carried out using the

CVFF force ®eld⁷⁷as implemented in Insight II version

2.95 (Biosym Technologies, CA). Hierarchical cluster

analysis (average linkage method) was carried out using

the S-Plus package (MathSoft, Data Analysis Products

Division, 1700 Westlake Ave. N., Suite 500, Seattle,

WA 98109).

22. Sabo, T.; Gurwitz, D.; Motola, L.; Brodt, P.; Barak, R.;

Elhanaty, E. Biochem. Biophys. Res. Commun. 1992, 188, 604.

23. Chao, B. H.; Kalkunte, S.; Maraganore, J. M.; Stone, S.

R. Biochemistry 1992, 31, 6175.

24. Vassallo, R. R., Jr.; Kieber-Emmons, T.; Cichowski, K.;

Brass, L. F. J. Biol. Chem. 1992, 267, 6081.

25. Feng, D.-M.; Veber, D. F.; Connolly, T. M.; Condra, C.;

Tang, M.-J.; Nutt, R. F. J. Med. Chem. 1995, 38, 4125.

26. Natarajan, S.; Reixinger, D.; Peluso, M.; Seiler, S. M. Int.

J. Pept. Protein Res. 1995, 45, 145.

27. Seiler, S. M.; Peluso, M.; Tuttle, J. G.; Pryor, K.; Klimas,

C.; Matsueda, G. R.; Bernatowicz, M. S. Mol. Pharmacol.

1996, 49, 190.

Acknowledgements

28. Shimamoto, T.; Tsuboi, H.; Kitajima, Y.; Miyazaki, T.;

Oyama, Y.; Imajo, S.; Ishiguro, M. Bioorg. Med. Chem. Lett.

1995, 5, 2417.

29. Nose, T.; Shimohigashi, Y.; Ohno, M.; Costa, T.; Shimizu,

N.; Ogino, Y. Biochem. Biophys. Res. Commun. 1993, 193, 694.

30. Scarborough, R. M., personal communication.

31. Smith, K. J.; Trayer, I. P.; Grand, R. J. A. Biochemistry

1994, 33, 6063.

We thank members of the spectroscopy group for sci-

enti®c discussions and for making their facilities very

accessible; the molecular modeling group (especially

Dr. Harold Almond, Jr.) for hardware and software

assistance; and Dr. William Hoekstra for advice and

assistance.

32. Ceruso, M. Ph.D. thesis, State University of New York at

Stony Brook, 1994.

References and Notes

33. Oksenberg, D.; Scarborough, R. M., unpublished results.

34. Marshall, G. R.; Bosshard H. E.; Eilers N. C.; Needleman

P. In Chemistry and Biology of Peptides: Meienhofer, J., Ed.;

Ann Arbor Science: Ann Arbor, MI, 1972, pp 571±578.

35. Marshall, G. R.; Bosshard, H. E. Circulation Res. 1972, 31

(suppl. 2), 143.

1. Jackson, T. Pharmac. Ther. 1991, 50, 425.

2. Strader, C. D.; Fong, T. M.; Tota, M. R.; Underwood, D.;

Dixon, R. A. F. Annu. Rev. Biochem. 1994, 63, 101.

3. Strader, C. D.; Fong, T. M.; Graziano, M. P.; Tota, M. R.

FASEB J. 1995, 9, 745.

36. Ramnarayan, K.; Chan, M. F.; Balaji, V. K.; Profeta, S.,

Jr.; Rao, S. N. Int. J. Pept. Protein Res. 1995, 45, 366.

37. Note: Incorporation of branched, a,a-dialkyl amino acids

into a peptide tends to induce folded structures, such as 3₁₀-

helical, a-helical, and b-turn conformations (Mutter, M.

Angew. Chem., Int. Ed. Engl. 1985, 24, 639; Prasad, B. V. V.;

Balaram, P. CRC Critical Rev. Biochem. 1984, 16, 307; Karle,

I. L.; Balaram, P. Biochemistry 1990, 29, 6747; Di Blasio, B.;

Pavone, V.; Lombardi, A.; Pedone, C.; Benedetti, E. Biopoly-

mers 1993, 33, 1037; Wipf, P.; Heimgartner, H. Helv. Chim.

Acta 1988, 71, 258; Wipf, P.; Kunz, R. W.; Prewo, R.; Heim-

gartner, H. Helv. Chim. Acta 1988, 71, 268).

38. Pettitt, B. M.; Matsunaga, T.; Al-Obeidi, F.; Gehig, C.;

Hruby, V. J.; Karplus, M. Biophys. J. 1991, 60, 1540.

39. Bruccoleri, R. E.; Karplus, M. Biopolymers 1990, 29,

1847.

40. Lewis, P. N.; Momany, F. A.; Scheraga, H. A. Biochim.

Biophys. Acta 1973, 303, 211.

41. Milner-White, E. J.; Ross, B. M.; Ismall, R.; Belhadj-

Mostefa, K.; Poet, R. J. Mol. Biol. 1988, 204, 777.

42. Vass, E.; Kurz, M.; Konat, R. K.; Hollosi, M. Spectro-

chim. Acta, Part A 1998, 54A, 773.

43. Hagler, A. T.; Osguthorpe, D. J.; Dauber-Osguthorpe, P.;

Hempel, J. C. Science 1985, 227, 1309.

44. Toniolo, C; Formaggio, F.; Crisma, M.; Schoemaker, H.

E.; Kamphuis, J. Tetrahedron: Asymmetry 1994, 5, 507.

45. Banerjee, A.; Ragothama, S. R.; Karle, I. S.; Balaram, P.

Biopolymers 1996, 39, 279.

46. Burgen, A. S. V.; Roberts, G. C. K.; Feeney, J. Nature

(London) 1975, 253, 753.

4. Moereels, H.; Lewi, P. J.; Koymans, L. M. H.; Janssen, P.

A. J. Recept. Channels 1996, 4, 19.

5. Gudermann, T.; Nurnberg, B.; Schultz, G. J. Mol. Med.

1995, 73, 51.

6. Trumpp-Kallmeyer, S.; Ho¯ack, J.; Bruinvels, A.; Hibert,

M. J. Med. Chem. 1992, 35, 3448.

7. Vu, T.-K. H.; Wheaton, V. I.; Hung, D. T.; Coughlin, S. R.

Cell 1991, 64, 1057.

8. Van Obberghen-Schilling, E.; Chambard, J.-C.; Vouret-

Craviari, V.; Chen, Y.-H.; Grall, D.; Pouyssegur, J. Eur. J.

Med. Chem. 1995, 30S, 117.

9. Dennington, P. M.; Berndt, M. C. Clin. Exp. Pharmacol.

Physiol. 1994, 21, 349.

10. Coughlin, S. R.; Scarborough, R. M.; Vu, T.-K. H.; Hung,

D. T. Cold Spring Harbor Symp. Quant. Biol. 1992, 57, 149.

11. Coughlin, S. R. Thromb. Haemostasis 1993, 66, 184.

12. Coughlin, S. R. Trends Cardiovasc. Med. 1994, 4, 77.

13. Ogletree, M. L.; Natarajan, S.; Seiler, S. M. Persp. Drug

Discov. Des. 1994, 1, 527.

14. Scarborough, R. M.; Naughton, M.; Teng, W.; Hung, D.

T.; Rose, J.; Vu, T.-K. H.; Wheaton, V. I.; Turck, C. W.;

Coughlin, S. R. J. Biol. Chem. 1992, 267, 13146.

15. For PAR-2: Nystedt, S.; Emilsson, K.; Wahlestedt, C.;

Sundelin, J. Proc. Natl. Acad. Sci. USA 1994, 91, 9208.

16. For PAR-2: Nystedt, S.; Emilsson, K.; Larsson, A.-K.;

Stroembeck, B.; Sundelin, J. Eur. J. Biochem. 1995, 232, 84.

17. For PAR-2: Nystedt, S.; Larsson, A.-K.; Aberg, H.; Sun-

delin, J. J. Biol. Chem. 1995, 270, 5950.

18. For PAR-3: Ishihara, H.; Connolly, A. J.; Zeng, D.; Kahn,

M. L.; Zheng, Y. W.; Timmons, C.; Tram, T.; Coughlin, S. R.

Nature (London) 1997, 386, 502.

19. For PAR-4: Xu, W.-F.; Andersen, H.; Whitmore, T. E.;

Presnell, S. R.; Yee, D. P.; Ching, A.; Gilbert, T.; Davie, E.

W.; Foster, D. C. Proc. Natl. Acad. Sci. USA 1998, 95, 6642.

20. Kahn, M. L.; Zheng, Y.-W.; Huang, W.; Bigornia, V.;

Zeng, D.; Mo, S.; Farese, R. V., Jr.; Tam, C.; Coughlin, S.

R. Nature 1998, 394, 690.

47. Silverman, R. B. The Organic Chemistry of Drug Design

and Drug Action. Academic: New York, 1992, pp 71±74.

48. Bernatowitz, M. S.; Kllimas, C. E.; Hartl, K. S.; Peluso,

M.; Allegretto, N. J.; Seiler, S. M. J. Med. Chem. 1996, 39,

4879.

49. Coller, B. S.; Ward, P.; Ceruso, M.; Scudder, L. E.;

Springer, K.; Kutok, J.; Prestwich, G. D. Biochemistry 1992,

31, 11713.

21. Hui, K. Y.; Jakubowski, J. A.; Wyss, V. L.; Angleton, E.

L. Biochem. Biophys. Res. Commun. 1992, 184, 790.

50. Boulanger, Y.; Khiat, A.; Chen, Y.; Gagnon, D.; Poitras,

P.; St.-Pierre, S. Int. J. Pept. Protein Res. 1995, 46, 527.

M. A. Ceruso et al. / Bioorg. Med. Chem. 7 (1999) 2353±2371

2371

51. Hudson, D.; Kenner, G.; Sharp, R.; Szelke, M. Int. J.

Pept. Protein Res. 1979, 14, 177.

52. Dauber-Osguthorpe, P.; Campbell, M.; Osguthorpe, D.

Int. J. Pept. Protein Res. 1991, 38, 357.

53. Masdouri, L.; Aubry, A.; Sakarellos, C.; Gomez, E.; Cung,

M.; Marraud, M. Int. J. Pept. Protein Res. 1988, 31, 420.

54. Ghosh, P.; Amaya, M.; Mellins, E.; Wiley, D. C. Nature

(London) 1995, 378, 457.

55. Swain, A. L.; Miller, M. M.; Green, J.; Rich, D. H.;

Schneider, J.; Kent, S. B. H.; Wlodawer, A. Proc. Natl. Acad.

Sci. USA 1990, 87, 8805.

K. C.; Liu, A.; Brungs, P. H.; Fusetani, N. Proc. Natl. Acad.

Sci. USA 1993, 90, 8048. A synthetic 19-membered-ring

thrombin inhibitor peptide: Greco, M. N.; Powell, E. T.;

Hecker, L. R.; Andrade-Gordon, P.; Kauman, J. A.; Lewis,

J. M.; Ganesh, V.; Tulinsky, A.; Maryano, B. E. Bioorg.

Med. Chem. Lett. 1996, 6, 2947.

63. Carpino, L. A.; Han, G. Y. J. Org. Chem. 1972, 37, 3404.

64. Roeske, R. W.; Tian, Z.; Edwards, P. Int. J. Pept. Protein

Res. 1992, 40, 119.

65. Hamada, Y.; Shioiri, T. Chem. Pharm. Bull. 1982, 30, 1921.

66. Fields, G. B.; Noble, R. L. Int. J. Pept. Protein Res. 1990,

35, 161.

67. Atherton, E.; Logan, C. J.; Sheppard, R. C. J. Chem. Soc.,

Perkin Trans. 1981, 1, 538.

56. Siligardi, G.; Drake, A. F. Biopolymers 1995, 37, 281.

57. Munoz, V.; Blanco, F. J.; Serrano, L. Nature Struct. Biol.

1995, 2, 380.

58. Matsoukas, J.; Hollenberg, M. D.; Mavromoustakos, T.;

Panagiotopoulos, D.; Alexopoulos, K.; Yamdagni, R.; Wu,

Q.; Moore, G. J. J. Protein Chem. 1997, 16, 113.

68. Penke, B.; Rivier, J. J. Org. Chem. 1987, 52, 1197.

69. Diaza-Meseguer, J.; Palomo-Coll, A. L.; Fernandez-

Lizarbe, J. R.; Zugaza-Bilbao, A. Synthesis 1980, 547.

70. ¹H NMR data indicated a homogeneous material and

supported the assigned structure; spectra are available on

request to the corresponding author.

71. Sasaki, Y.; Coy, D. H. Peptides 1987, 8, 119.

72. Hewgill, F. R.; Jeeries, P. R. J. Chem. Soc. 1955, 2767.

73. This PAR-1 binding assay was developed in our labora-

tories. For other PAR-1 binding assays, see: ref 25; Ahn,

H.-S.; Foster, C.; Boykow, G.; Arik, L.; Smith-Torhan, A.;

Hesk, D.; Chatterjee, M. Molec. Pharmacol.. 1997, 51, 350.

74. Jones, C. L. A.; Witte, D. P.; Feller, M. J.; Fugman, D. A.;

Dorn II, G. W.; Lieberman, M. A. Biochim. Biophys. Acta

1992, 1136, 272.

59. Matsoukas, J.; Panagiotopoulos, D.; Keramida, M.; Mav-

romoustakos, T.; Yamdagni, R.; Wu, Q.; Moore, G. J.; Sai-

feddine, M.; Hollenberg, M. D. J. Med. Chem. 1996, 39, 3585.

60. We have studied macrocyclic analogues of SFLLRN, with

a cyclization linkage between the P1 and P6 or the P3 and P6

side chains, for PAR-1 agonist activity in platelet activation,

and found that some compounds with the P3/P6 linkage exhi-

bit PAR-1 agonist activity at the level of TRAP-6.⁶¹

61. McComsey, D. F.; Hecker, L. R.; Andrade-Gordon, P.;

Addo, M. F.; Maryano, B. E. Bioorg. Med. Chem. Lett.

1999, 9, 255.

62. For example, we have observed an extended conformation

for peptide recognition motifs in a 19-membered-ring macro-

cyclic environment during complexation of inhibitor ligands

with human a-thrombin. Cyclotheonamide A, a 19-mem-

bered-ring thrombin inhibitor peptide: Maryano, B. E.; Qiu,

X.; Padmanabhan, K. P.; Tulinsky, A.; Almond, Jr., H. R.;

Andrade-Gordon, P.; Greco, M. N.; Kauman, J. A.; Nicolaou,

75. Fugman, D. A.; Witte, D. P.; Jones, C. L. A.; Aronow, B.

J.; Lieberman, M. A. Blood 1990, 75, 1252.

76. Witte, D. P.; Stambrook, P. J.; Feliciano, E.; Jones, C. L.

A.; Lieberman, M. A. J. Cell. Physiol. 1988, 137, 86.

77. Hagler, A. T.; Dauber, P.; Lifson, S. J. Am. Chem. Soc.

1979, 101, 5131.

Article Doi

Thrombin receptor-activating peptides (TRAPs): Investigation of bioactive conformations via structure-activity, spectroscopic, and computational studies

DOI: 10.1016/S0968-0896(99)00180-7

Source and publish data:

Authors:

Article abstract of DOI:10.1016/S0968-0896(99)00180-7

Full text of DOI:10.1016/S0968-0896(99)00180-7

Products guided by the article

R&D Labs maybe for 152436-04-9

Relevant to this article

Hot Product