Synthesis of cyclic RGD derivatives via solid phase macrocyclization using the Heck reaction

Article abstract of DOI:10.1016/S0040-4020(01)00113-2

A novel intramolecular macrocyclization reaction on a solid support using the Heck reaction has been achieved. For head to tail cyclization on a solid support, the linear precursor was anchored to a chlorotrityl chloride resin via an ester linkage using the β-carboxyl group of Asp. The Heck coupling of acrylic acid amide to 3-iodobenzylamine on the solid support proceeds smoothly to yield a cyclic tetrapeptide derivative, which contains a new 3-substituted cinnamic acid template and Arg-Gly-Asp sequence. The macrocyclization reaction takes place considerably more rapidly on a solid support than in solution. The solid phase procedure was successfully used for the construction of cyclic RGD libraries having diverse side chain structures, combined with a variety of ring sizes.

Full text of DOI:10.1016/S0040-4020(01)00113-2

TETRAHEDRON
Pergamon
Tetrahedron 57 (2001) 2293±2303
Synthesis of cyclic RGD derivatives via solid phase
macrocyclization using the Heck reaction
Kenichi Akaji,a,p Kenta Teruya,a Masako Akajib and Saburou Aimotoa
aInstitute for Protein Research, Osaka University, Suita, Yamadaoka, Osaka 565-0871, Japan
bFaculty of Pharmaceutical Sciences, Kyoto University, Sakyo-ku, Yoshida, Kyoto 606-8304, Japan
Received 20 December 2000; accepted 24 January 2001
AbstractÐA novel intramolecular macrocyclization reaction on a solid support using the Heck reaction has been achieved. For head to tail
cyclization on a solid support, the linear precursor was anchored to a chlorotrityl chloride resin via an ester linkage using the b-carboxyl
group of Asp. The Heck coupling of acrylic acid amide to 3-iodobenzylamine on the solid support proceeds smoothly to yield a cyclic
tetrapeptide derivative, which contains a new 3-substituted cinnamic acid template and Arg-Gly-Asp sequence. The macrocyclization
reaction takes place considerably more rapidly on a solid support than in solution. The solid phase procedure was successfully used for
the construction of cyclic RGD libraries having diverse side chain structures, combined with a variety of ring sizes. q 2001 Elsevier Science
Ltd. All rights reserved.
The application of solid support chemical synthesis has
recently been extended beyond the traditional preparation
of biopolymers to the synthesis of general organic mol-
ecules, for the preparation of chemical libraries in com-
binatorial synthesis.1 In order for solid phase synthesis to
be useful in generating diversity, however, the repertoire of
carbon±carbon bond forming reactions that can be used in
the preparation of biologically relevant molecules need to
be increased. The Heck reaction2 provides a very attractive
method for the preparation of such bonds on solid support.
The reaction does not require a stoichiometric amount of
Pd(0) nor inert anhydrous conditions and proceeds readily at
room temperature, even on solid support.3 The intramol-
ecular Heck reaction is a useful method for the preparation
of ®ve, six, or seven-membered rings fused to aromatic
rings. Although the reaction has been applied to the solid
phase synthesis of indole derivatives,4 benzofuran deriva-
tives,5 and isoquinolinone derivatives,6 the application of
these synthetic strategies to the production of more complex
macrocyclic molecules have, to date, been rather limited in
scope.7
We selected intramolecular macrocyclization as a suitable
reaction that could be carried out ef®ciently on a solid
support rather than in solution, because of the `pseudo-
dilution' effect.8 This speci®c effect, which is the result of
the polymeric solid support, has been used successfully for
macrocyclization via disul®de bond formation in peptide
chemistry.9 In contrast to these solid phase results, in®nite
dilution is generally inevitable in solution phase reactions,
Scheme 1.
Keywords: RGD derivatives; Heck reaction; cyclic tetrapeptide; solid phase macrocyclization; tert-alcohol resin.
Abbreviations: But, tert-butyl; CIP, 2-chloro-1,3-dimethylimidazolidium hexa¯uorophosphate; DIPCDI, diisopropylcarbodiimide; HRFAB, high-resolution
fast atom bombardment; Fmoc, ¯uoren-9-ylmethyloxycarbonyl; HOAt, 1-hydroxy-7-azabenzotrizazole; HOBt, 1-hydroxybenzotriazole; MALDI-TOF,
matrix-assisted laser desorption ionization time of ¯ight; MD, molecular dynamics; MS, mass spectrometry; Mtr, 4-methoxy-2,3,6-trimethylbenzenesulfonyl;
SPPS, solid phase peptide synthesis; TFA, tri¯uoroacetic acid; TFE, tri¯uoroethanol.
p
Corresponding author. Tel.: 181-6-6879-8608; fax: 181-6-6879-8609; e-mail: akaji@protein.osaka-u.ac.jp
0040±4020/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(01)00113-2

2294
K. Akaji et al. / Tetrahedron 57 (2001) 2293±2303
for obtaining a reasonable reaction ef®ciency of the mono-
meric product. Based on the above information, we report
here a novel intramolecular macrocyclization on a solid
support using the Heck reaction.10
1. Results and discussion
The synthetic scheme for the cyclic RGD derivative is
outlined in Scheme 2. For the construction of the linear
precursor peptide, a base-labile Fmoc group was employed
as the Na-protecting group.13 The linear precursor is
anchored to a solid support through the b-carboxyl group
of Asp; this makes intramolecular head to tail cyclization on
the solid support feasible. After the cyclization, the desired
product can be cleaved from the solid support by treatment
with mild acid. Based on this scheme, we ®rst examined a
suitable solid support for precursor synthesis and reaction
ef®ciency of solid phase intramolecular cyclization.
A cyclic tetrapeptide derivative was synthesized using the
Heck coupling of acrylic acid amide to a 3-iodobenzyl
amine moiety, on a solid support. The cyclic derivative
consists of a new 3-substituted cinnamic acid template
for the construction of the rigid cyclic structure containing
Arg-Gly-Asp (RGD) sequence (Scheme 1). The RGD
sequence is a common recognition motif for the integrin
family of receptors, which are involved in cell±cell and
cell±matrix adhesion.11 Glycoprotein IIb/IIIa (GP IIb/
IIIa), a member of the integrin family, is expressed on the
surface of activated platelets which binds to ®brinogen to
cause platelet aggregation. Thus, the solid phase preparation
of cyclic RGD-derivatives with af®nities for GPIIb/IIIa
could be effective approach for the discovery of drug candi-
dates which could prevent the formation of platelet-rich
clots.12
For anchoring, a C-terminal Asp derivative, 2, was prepared
via the CIP-mediated coupling14 of commercially available
Fmoc-Asp(OBut)-OH and 3-iodobenzylamine followed by
cleavage of the side chain But group with TFA. The use of
the conventional Wang resin15 and more acid-labile 2-
chlorotrityl resin (Clt resin)16 were examined as potential
solid supports. To the Wang resin, at ®rst, the Asp derivative
Scheme 2.

K. Akaji et al. / Tetrahedron 57 (2001) 2293±2303
2295
Figure 1. Dipeptide cleavage by intramolecular aminolysis of the b-benzyl ester linked to the Wang resin. Cleavage yield was estimated by amino acid
analysis of the recovered resin.
prepared above was anchored according to literature pro-
cedures. The successive introduction of Fmoc-Gly-OH and
Fmoc-Arg(Mtr)-OH17 was then conducted by the combi-
nation of piperidine mediated Fmoc deprotection and
DIPCDI mediated coupling. However, no increment of the
resin weight was obtained because of the dipeptide cleavage
by intramolecular aminolysis of b-benzyl ester linked to the
solid support (Fig. 1). After treating with piperidine, the
diketodiazepin 11 and N-(¯uoren-9-ylmethyl)piperidine 12
were isolated and characterized from the ®ltrate. The
cleavage yield was estimated at 63% by amino acid analysis
of the recovered Wang resin.
same reaction conditions as described in the solid phase
reaction. An aliquot was periodically removed from the
reaction mixture and the progress of the reaction was moni-
tored in a similar manner by HPLC.
The intramolecular cyclization in solution proceeded in
proportion to the reaction time, but was relatively slow.
Most of precursor 6 was converted to the cyclized product
8 after 8 h of reaction (Fig. 2b). In contrast, the same cycli-
zation reaction occurred rapidly on the solid support as
shown in Fig. 2a. Most of the precursor was converted to
product within 2 h and the time course of the reaction was
quite different with that of solution phase reaction. The
results clearly show that the Pd(0)-mediated intramolecular
macrocyclization described in this paper is one of the reac-
tions suitable for solid phase organic synthesis.
We then examined the construction of the desired precursor
starting from the Clt resin. The Asp derivative 2 was
anchored to the Clt resin according to the published pro-
cedure. The desired precursor was then constructed by the
successive introduction of Fmoc-Gly-OH, Fmoc-Arg(Mtr)-
OH, Fmoc-Gly-OH, and acrylic acid monomer by the
combination of piperidine mediated Fmoc deprotection
and DIPCDI mediated coupling. Each condensation reac-
tion was monitored by the Kaiser-ninhydrin test.18 The
quantitative increment of the resin weight was obtained
and the incorporation of amino acids was con®rmed by
amino acid analysis of the product resin. In addition, the
linear precursor 6 obtained by a brief treatment of the
precursor resin with a mixture of AcOH±TFE±CH2Cl2
(1:1:2) showed a sharp main peak on analytical HPLC.
Thus, in the present synthesis, the linear precursor for
head to tail cyclization was successfully synthesized from
the sterically hindered Clt resin.
Thus, for the preparation of a cyclic RGD derivative 9, the
Pd(0)-mediated macrocyclization of 5 employing Pd(OAc)2
with Ph3P and Bu4NCl in a DMF/H2O/Et3N solvent system
was carried out at 378C for 4 h to provide 7. The cyclized
peptide resin 7 was then treated with AcOH±TFE±CH2Cl2
to cleave the protected cyclic tetrapeptide 8 from the resin
support. The cyclic product 8 was treated with TFA±thio-
anisole at 258C for 3 h without any further puri®cation to
remove the Mtr group from the Arg side chain. The depro-
tected product 9 was puri®ed by preparative HPLC, and the
homogeneous product was obtained in 20% overall yield
(calculated from the starting resin). The puri®ed cyclic
derivative 9 exhibited a single sharp peak on analytical
HPLC and was shown to be a monomer by HRFAB MS.
The trans-con®guration of the product 9 was de®ned by 1H
NMR analysis.
The ef®ciency of solid phase macrocyclization was then
investigated and the results were compared with that in
solution phase. The linear precursor resin 5 obtained
above was treated with Pd(OAc)2 in the presence of Ph3P
and Bu4NCl using DMF/H2O/Et3N as the solvent. The cycli-
zation reactions were conducted for 2, 4, and 8 h in parallel
at 378C. Each resin was then treated with AcOH±TFE±
CH2Cl2 (1:1:8) and the detached product was analyzed
using HPLC to examine the progress of the reaction. For
comparison, a cyclization reaction in solution was also
conducted. The linear precursor 6, obtained by AcOH±
TFE±CH2Cl2 treatment of the precursor resin 5, was treated
with Pd(OAc)2/Ph3P/Bu4NCl in DMF/H2O/Et3N under the
The puri®ed cyclic RGD derivative inhibited ®brinogen
binding to immobilized GPIIb/IIIa with IC50 2.0£1025 M,
but showed no inhibitory activity for the vitronectin recep-
tor, another member of the integrin family of receptors.
According to the works on active conformation of RGD
peptides,19 the distance between the guanidino group of
Arg and the b-carboxylic acid of Asp should be within
Ê
12±18 A to show platelet aggregation activity. From a
ROESY (rotating-frame Overhauser enhancement spectro-
scopy) experiment of the puri®ed cyclic product and mol-
ecular dynamics (MD) simulations, the distance between the

2296
K. Akaji et al. / Tetrahedron 57 (2001) 2293±2303
guanidino group and b-carboxyl group in 9 was determined
Ê
to be within 12±15 A in its typical lower energy conform-
ation. Thus, presumably, conformational restriction of the
cyclized product 9 contributed, to some extent, to its selec-
tivity for integrin receptors.
The above solid phase macrocyclization reaction was then
applied to library construction by the combinatorial
approach because the 3-substituted cinnamic acid template
containing the RGD sequence showed a weak but speci®c
binding to GPIIb/IIIa. Two sites were selected for intro-
ducing structural diversity: one at the side chain of the
N-terminal amino acid residues to alter the steric or ionic
effect and the second at the residue between Arg and Asp, to
alter the ring size. To facilitate the synthesis, the split and
mix procedure20 was selected for the preparations of these
libraries.
Based on the above scheme, we ®rst constructed an alkyl
library, which consisted of 15 components (Fig. 3). To intro-
duce diversity regarding steric effect in this library, proton,
methyl, ethyl, propyl, and cyclic propyl structures were
incorporated at the side chain of the N-terminal amino
acid. For this purpose, each of the corresponding amino
acids, Gly, Ala, a-aminobutylic acid (aAbu), norvaline
(Nva), or Pro was introduced. The ring size was altered by
inserting different numbers of methylene groups between
the Arg and Asp residues, i.e. by introducing Gly, g-amino-
butylic acid (gAbu), or e-aminohexanoic acid (eAhx) resi-
dues at this position.
For multi component macrocyclization, the alkyl library
precursor resin was treated with Pd(OAc)2/Ph3P/Bu4NCl
under the same reaction conditions as described for the
monomeric resin. The solid phase reaction was easily moni-
tored on HPLC by analyzing the acid extract of the starting
and product resins. Each component prior to and after cycli-
zation was unambiguously identi®ed by MALDI±TOF MS
Figure 2. HPLC pro®les of the cyclized products on solid support: (a) or in
solution, (b) pderived from Ph3P. HPLC on YMC AM302 [CH3CN (20±
30%/30 min) in 0.1% aq. TFA, 0.9 ml/min]
Figure 3. Alkyl library.

K. Akaji et al. / Tetrahedron 57 (2001) 2293±2303
2297
Figure 4. HPLC pro®les of the alkyl library prior to (a) and after (b) cyclization. Each component was identi®ed by MS-analysis: 1 [Gly1, gAbu3], 2 [Gly1,
Gly3], 3 [Ala1, gAbu3], 4 [Gly1, eAhx3], 5 [Ala1, Gly3], 6 [Pro1, gAbu3], 7 [aAbu1, gAbu3], 8 [Ala1, eAhx3], 9 [Pro1, Gly3], 10 [aAbu1, Gly3], 11 [Pro1, eAhx3],
12 [aAbu1, eAhx3], 13 [Nva1, gAbu3], 14 [Nva1, Gly3], and 15 [Nva1, eAhx3] derivative.
Figure 5. HPLC pro®le of the alkyl library after deprotection. For identi®cation of each compound, see the caption of Fig. 4.
analysis of each peak, although both chromatograms
showed relatively complex elution patterns (Fig. 4). No
other peptide components were detected from the chromato-
gram. The cyclized product was cleaved from the resin with
AcOH/TFE/DCM and the Mtr group of the product was
removed by TFA/thioanisole as described for cyclic RGD
derivative 9. The deprotected mixture was quickly passed
through a reversed phase HPLC column to remove hydro-
phobic contaminants such as scavengers, which are eluted at
high concentrations of CH3CN. The desired library product,
which consisted of 15 components, was obtained as white
powder in 10% overall yield calculated from the starting
resin. Each component of the library was easily identi®ed
by HPLC (Fig. 5) and MS analysis.
The same approach was then applied to the construction of a
functional group library (Fig. 6). Instead of alkyl group in
the above described alkyl library, four types of functional
groups, hydroxyl, amino, carboxyl, and amido groups,
were incorporated to introduce diversity regarding ionic
Figure 6. Functional group library.

2298
K. Akaji et al. / Tetrahedron 57 (2001) 2293±2303
Figure 7. HPLC pro®le of the functional group library. Each component was identi®ed by HRFAB MS analysis: 1 [Dap1, gAbu3], 2 [Dap1, Gly3], 3 [Lys1,
gAbu3], 4 [Lys1, Gly3], 5 [Dap1, eAhx3], 6 [Asn1, gAbu3], 7 [Ser1, gAbu3], 8 [Asn1, Gly3], 9 [Ser1, Gly3], 10 [Asp1, gAbu3], 11 [Lys1, eAhx3], 12 [Asp1, Gly3],
13 [Asn1, eAhx3], 14 [Ser1, eAhx3], and 15 [Asp1, eAhx3] derivative.
effects. For this, each of the corresponding amino acids, Ser,
diaminopropionic acid (Dpr), Lys, Asp, or Asn were intro-
duced. Combined with ring size diversity, this library also
consists of 15 components. The construction of the precur-
sor resin by the split and mix procedure and the following
solid phase macrocyclization were conducted under the
same conditions as described for the alkyl library. All
components prior to and after cyclization were similarly
identi®ed by HPLC-MS analysis without dif®culties. After
partial puri®cation by HPLC, the desired product, which
consisted of 15 components was obtained as white powder
in 9% overall yield. Each component in the library was
similarly identi®ed by HPLC-MS analysis without dif®culty
(Fig. 7).
chloride resin were obtained from Calbiochem Nova-
biochem and were used without further puri®cation.
Fmoc-Dpr(Boc)-OH was prepared according to the pro-
cedure described by Waki et al.21 For quanti®cation of
Fmoc amino acids on the resin, the absorbance at 301 nm
after cleavage of the Fmoc group with piperidine
was measured according to the procedure described by
Meienhofer et al.22
1
Melting points were uncorrected. The H and 13C NMR
spectra were recorded on a 270 MHz (JEOL), 400 MHz
(Bruker), or 600 MHz (Bruker) spectrometer with TMS as
an internal standard. HRFAB MS was obtained on a JEOL
JMS-SX102A spectrometer equipped with a JMA-DA7000
data system. MALDI-TOF MS was obtained on a Voyager-
DE Biospectrometry Workstation of PerSeptive Bio-
systems. The binding of ®brinogen to immobilized human
GP IIb/IIIa was performed according to the method of
Hayashi et al.23 Molecular modeling studies were performed
using Insight and Discover programs. Molecular dynamics
(MD) simulations and energy minimization (EM) were
conducted using consistence valence force ®eld (CVFF)
parameters according to the method of Siahaan et al.24
HPLC was carried out on a reversed phase column which
was eluted with CH3CN in 0.1% aqueous TFA and detected
at OD 220 nm: Rt1±Rt4, YMC AM302 (4.6£150 mm), ¯ow
rate 0.9 ml/min; Rt5, YMC-Pack ProC18 AS 323
(10£250 mm), ¯ow rate 2.5 ml/min.
2. Conclusion
The ®nding here clearly demonstrate that the Heck reaction
can be used for solid phase macrocyclization reaction.
The tert-alcohol based Clt resin was more suitable than
the benzyl alcohol based Wang resin for the preparation
of the linear precursor used for head to tail cyclization. In
addition, the acid lability of the Clt resin was especially
advantageous for monitoring the solid phase reaction by
HPLC-MS analysis. The monitoring method showed that
the reaction proceeded more ef®ciently on the solid support
than in solution, presumably because of the pseudo-dilution
effect. Thus, the cyclic RGD derivative and its libraries were
easily prepared, isolated, and identi®ed by applying the
solid phase macrocyclization combined with HPLC-MS
analysis. Preparation of a variety of libraries followed by
an assessment of ®brinogen inhibitory activity will allow the
application of this solid phase procedure to the design of a
high af®nity ligand of GPIIb/IIIa.
3.1.1. Fmoc-Asp(OBut)±NH±CH2±C6H4±I, 1. To a solu-
tion of Fmoc-Asp(OBut)±OH (0.50 g, 1.2 mmol) in THF
(5.0 ml) were added DIEA (0.63 ml, 3.6 mmol), HOAt
(0.10 g, 0.74 mmol), CIP (0.41 g, 1.5 mmol), and 2-iodo-
benzylamine (0.19 g, 1.4 mmol), and the mixture was
stirred at 258C for 45 min. AcOEt (20 ml) was then added
to the mixture and the organic phase was washed with 5%
citric acid, H2O, dried over MgSO4, and evaporated. The
crude product was puri®ed by silica gel column chroma-
tography using CHCl3 followed by precipitation with
AcOEt/hexane to yield 0.66 g (87%) of 1 as a solid: Mp
3. Experimental section
3.1. General
89±908C,[a]D ˆ2148 (c 0.50, DMF), 1H NMR
25
(270 MHz, CDCl3) d 7.76±7.73 (m, 2H), 7.60±7.55 (m,
4H), 7.43±7.26 (m, 4H), 7.20 (d, Jˆ7.6 Hz, 1H), 7.02 (t,
Jˆ7.6 Hz, 1H), 6.83 (br t, Jˆ6.3 Hz, 1H), 5.96 (d,
Solvents were reagent grade and dried prior to use. Fmoc
amino acid derivatives, except for the diaminopropionic
acid (Dpr) derivative, the Wang resin, and the 2-chlorotrityl

K. Akaji et al. / Tetrahedron 57 (2001) 2293±2303
2299
Jˆ8.0 Hz, 1H), 4.53 (m, 1H), 4.45 (br d, Jˆ6.3 Hz, 2H),
4.38 (m, 2H), 4.20 (t, Jˆ6.6 Hz, 1H), 2.96 (dd,
Jˆ17.2 Hz, Jˆ3.8 Hz, 1H), 2.61 (dd, Jˆ17.2 Hz,
Jˆ6.6 Hz, 1H), 1.43 (s, 9H). 13C NMR (67.8 MHz,
CDCl3) d 171.28, 170.49, 156.13, 143.61, 141.35, 140.34,
136.57, 136.40, 130.35, 127.80, 127.10, 126.74, 124.99,
120.07, 94.55, 82.07, 67.17, 51.18, 47.17, 42,75, 37.34,
28.07. Anal. Calcd for C30H31N2O5I: C, 57.51; H, 4.99; N,
4.47. Found: C, 57.67; H, 5.12; N, 4.31. HRFAB MS, m/z
627.1359 for [M1H]1 (Calcd 627.1356 for C30H32N2O5I).
deprotection and condensation procedure was repeated for
the incorporation of Fmoc-Arg(Mtr)-OH, Fmoc-Gly-OH,
and acrylic acid monomer. At each condensation step, the
results of a Kaiser ninhydrin test were negative after a single
coupling reaction. The resin obtained was ®ltered, washed
with DMF and MeOH, and dried to yield 2.3 g (98%) of
desired tetrapeptide±Clt resin 5. Amino acid analysis after
12 M HCl-propionic acid hydrolysis at 1108C for 18 h: Asp
0.97, Gly 1.98, Arg 1.00 (recovery of Arg 80%).
3.1.6. CH2vCHCO±Gly-Arg(Mtr)-Gly-Asp(Wang
resin)±NH±CH2±C6H4±I (condensation on Wang
resin). Starting from 2.1 g of the anchored resin 4, the
same procedure as described for Clt resin was used for
construction of the peptide chain to yield 2.1 g of product
resin. The recovery of Arg after 12M HCl±propionic acid
hydrolysis (1108C, 18 h) of the product resin was 5.0%.
3.1.2. Fmoc-Asp±NH±CH2±C6H4±I, 2. TFA/anisole (22±
2.2 ml) was added to Fmoc-Asp(OBut)±NH±CH2±C6H4±I
(2.5 g, 4.0 mmol) and the mixture was stirred at 258C.
During the stirring, a ¯occulent precipitate separated out.
After 90 min of stirring, TFA of the mixture was removed
by evaporation at 208C. Hexane (50 ml) was added to the
residue and the precipitate was ®ltered and dried to yield
2.0 g (88%) of 2 as a white powder. The product was used
immediately for anchoring without puri®cation. Mp 181±
1828C, [a]D ˆ2128 (c 0.50, DMF), H NMR (400 MHz,
DMSO-d6) d8.47 (t, Jˆ5.9 Hz, 1H), 7.89 (d, Jˆ7.5 Hz,
2H), 7.72 (d, Jˆ7.7 Hz, 3H), 7.62 (s, 1H), 7.58 (d,
Jˆ7.8 Hz, 1H), 7.42 (t, Jˆ7.4 Hz, 2H), 7.33 (t, Jˆ7.3 Hz,
1H), 7.32 (t, Jˆ7.4 Hz, 1H), 7.25 (d, Jˆ7.7 Hz, 1H), 7.09 (t,
Jˆ7.7 Hz, 1H), 4.38 (m, 1H), 4.34±4.21 (m, 5H), 2.70 (dd,
Jˆ16.5, 5.3 Hz, 1H), 2.53 (dd, Jˆ16.5, 8.8 Hz, 1H). 13C
NMR (100 MHz, DMSO-d6) d171.72, 170.88, 155.84,
143.77, 142.12, 140.69, 135.56, 135.37, 130.37, 127.64,
127.09, 126.42, 125.34, 120.10, 94.70, 65.81, 51.49,
46.63, 41.48, 36.29. HRFAB MS, m/z 571.0734 for
[M1H]1 (Calcd 571.0730 for C26H24N2O5I).
3.1.7. Diketodiazepin, 11 and N-(¯uoren-9-ylmethyl)-
piperidine, 12. To 0.50 g of Fmoc-Gly-Asp(Wang resin)±
NH±CH2±C6H4±I, 10 (0.27 mmol/g) was added 20% piper-
idine in DMF (6.0 ml), and the mixture agitated for 40 min
at 258C. The resin was ®ltered, and washed with DMF. The
cleavage yield (63%) was estimated by amino acid analysis
after 12 M HCl±propionic acid hydrolysis (66 h at 1108C)
of the resin: the ratios of amino acids in the hydrolysate of
the resin; Asp 1.00, Gly 0.98 (recovery 33%), and of the
starting dipeptide resin; Asp 1.00, Gly 0.95 (recovery 88%).
25
1
The solvent of the ®ltrate was removed by evaporation and
the residue was suspended in CHCl3 (3.0 ml). The precipi-
tate of the suspension was washed with CHCl3, Et2O, and
then dried in vacuo to yield 21 mg (62% recovery) of 11 as a
25
3.1.3. Fmoc-Asp(Clt resin)±NH±CH2±C6H4±I (anchor-
ing to Clt resin), 3. To 2.0 g (2.6 mmol) of 2-chlorotrityl
chloride (Clt) resin were added Fmoc-Asp±NH±CH2±
C6H4±I (0.37 g, 0.65 mmol) in CH2Cl2/DMF (2:1, 5.0 ml),
and DIEA (0.90 ml, 5.2 mmol). The mixture was agitated
for 2 h at 378C and the solvent was then removed by ®ltra-
tion. MeOH/DIEA (9:1, 8 ml) was added to the resin and the
mixture was further agitated for 30 min at 258C. The resin
was ®ltered, washed with MeOH, and then dried in vacuo to
yield 2.2 g of 3, substituted at a level of 0.22 mmol/g.
solid: Mp 2078C (decomp.) [a]D ˆ8.68 (c 0.30, DMSO),
1H NMR (400 MHz, DMSO-d6) ( 8.45 (br t, Jˆ5.6 Hz, 1H),
7.99 (s, 1H), 7.95 (s, 1H), 7.63 (s, 1H), 7.60 (d, Jˆ7.7 Hz,
1H), 7.28 (d, Jˆ7.7 Hz, 1H), 7.11 (t, Jˆ7.7 Hz, 1H), 4.23
(br d, Jˆ5.5 Hz, 2H), 4.09 (br t, Jˆ5.0 Hz, 1H), 3.72 (s,
2H), 2.63 (d, Jˆ5.0 Hz, 2H). 13C NMR (100 MHz, DMSO-
d6) d169.18, 167.83, 166.02, 142.06, 135.81, 135.49,
130.47, 126.73, 94.76, 51.16, 44.59, 41.38, 37.60. HRFAB
MS, m/z 388.0162 for [M1H]1 (Calcd 388.0158 for
C13H15N3O3I).
3.1.4. Fmoc-Asp(Wang resin)±NH±CH2 ±C6H4 ±I
(anchoring to Wang resin), 4. To 1.0 g (0.70 mmol) of
Wang resin were added Fmoc-Asp±NH±CH2±C6H4±I
(1.0 g, 1.8 mmol) in DMF (5.0 ml), DIEA (0.12 ml,
0.70 mmol), DMAP (86 mg, 0.70 mmol), and DIPCDI
(0.28 ml, 1.8 mmol), and the mixture was agitated for 2 h
at 378C. The resin was ®ltered, washed with DMF and
MeOH, and then dried in vacuo to yield 1.1 g of 4, sub-
stituted at a level of 0.21 mmol/g.
The solvent of the supernatant was removed by evaporation
and the residue was puri®ed by silica gel column chroma-
tography using CHCl3 to yield 18 mg (recovery 80%) of 12
1
as a solid: Mp. 107±1098C, H NMR (400 MHz, CDCl3) d
7.74 (d, Jˆ7.5 Hz, 2H), 7.69 (d, Jˆ7.5 Hz, 2H), 7.36 (t,
Jˆ7.4 Hz, 2H), 7.284 (t, Jˆ7.4 Hz, 1H), 7.281 (t,
Jˆ7.4 Hz, 1H), 4.04 (t, Jˆ8.0 Hz, 1H), 2.57 (d, Jˆ8.0 Hz,
2H), 2.56 (br t, Jˆ5.7 Hz, 4H), 1.68 (tt, Jˆ5.5, 5.7 Hz, 4H),
1.51 (tt, Jˆ5.6 Hz, Jˆ5.6 Hz, 2H). 13C NMR (100 MHz,
CDCl3) d 146.68, 140.96, 126.94, 126.64, 125.42, 119.58,
63.08, 54.91, 44.82, 26.25, 24.63. HRFAB MS, m/z
264.1757 for [M1H]1 (Calcd 264.1752 for C19H22N).
3.1.5. CH2vCHCO±Gly-Arg(Mtr)-Gly-Asp(Clt resin)±
NH±CH2±C6H4±I (condensation on Clt resin), 5. Piper-
idine (20%) in DMF was added to the starting resin 3 and the
mixture was agitated for 20 min at 258C. The Na-depro-
tected resin was ®ltered and washed with DMF. To this
resin, Fmoc-Gly-OH (2.5 equiv.) in DMF, HOBt
(2.5 equiv.), DIEA (2.5 equiv.), and DIPCDI (2.5 equiv.)
were added and the resin mixture was agitated for 90 min
at 258C. After washing the resin with DMF, the same
3.1.8. CH2vCHCO±Gly-Arg(Mtr)-Gly-Asp±NH±CH2±
C6H4±I, 6. To 0.10 g of RGD±Clt resin 5 (0.21 mmol/g)
was added AcOH/TFE/CH2Cl2 (1:1:8, 3.0 ml) and the
mixture was agitated for 45 min at 258C. The resin was
®ltered, washed with AcOH/TFE/CH2Cl2 (1:1:8, 3.0 ml).
The solvent of the ®ltrate was removed by evaporation

2300
K. Akaji et al. / Tetrahedron 57 (2001) 2293±2303
and the crude product was isolated by precipitation with
H2O to yield 15 mg (80%) of 6 as a solid: Rt1 17.6 min
[CH3CN (20±80%/30 min)], HRFAB MS, m/z 885.2091
for [M1H]1 (Calcd 885.2102 for C34H46N8O10SI). Amino
acid analysis after 6 M HCl (1108C, 18 h) hydrolysis; Asp
1.07, Gly 1.88, Arg 1.00.
2.86 (dd, Jˆ16.8, 9.1 Hz, 1H), 1.94 (tdd, Jˆ7.0, 13.9,
8.5 Hz, 1H), 1.71 (tdd, Jˆ8.5, 13.9, 5.8 Hz, 1H), 1.59
(tdd, Jˆ6.9, 8.5, 7.0 Hz, 2H). HRFAB MS, m/z 545.2469
for [M1H]1 (Calcd 545.2472 for C24H33N8O7). Amino acid
analysis after 6 M HCl hydrolysis (1108C, 22 h): Asp 1.00,
Gly 2.03, Arg 0.96.
3.4. Alkyl library
3.2. Cyclization in solution
9.0 mg (10 mmol) of linear precursor 6 in DMF/H2O/
CH2Cl2 (9:1:1, 0.30 ml) was added in one portion to a solu-
tion of Pd(OAc)2 (1.0 mg, 4.0 mmol), PPh3 (1.0 mg,
4.0 mmol), and Bu4NCl (1.0 mg, 4.0 mmol) in DMF/H2O/
CH2Cl2 (9:1:1, 0.50 ml) and the mixture (13 mmol/ml
concentration) was stirred at 378C. A 50 ml aliquot of the
solution, taken from the reaction mixture after 2, 4, and 8 h,
was added to H2O (250 ml) and the solution was subjected to
analytical HPLC. The results are shown in Fig. 2b.
Starting from three batches of Fmoc-Asp(Clt resin)±NH±
CH2±C6H4±I (0.40 g each of 0.23 mmol/g resin), Fmoc-
Gly±OH, Fmoc-gAbu±OH, and Fmoc-eAhx±OH were
introduced in parallel by piperidine treatment, followed by
DIPCDI coupling, as described for linear precursor resin 5.
The three batches were combined and Fmoc-Arg(Mtr)±OH
was introduced to the resin, as described above. The result-
ing resin was divided to ®ve portions and Fmoc-Gly±OH,
Fmoc-Ala±OH, Fmoc-aAbu±OH, Fmoc-Nva±OH, and
Fmoc-Pro±OH were introduced to the ®ve portions of the
resin in parallel. These resins were again combined and
acrylic acid was introduced to yield 1.3 g of library pre-
cursor resin.
3.3. Cyclization on solid support
0.30 g (63 mmmol) of CH2vCHCO±Gly-Arg(Mtr)-Gly-
Asp(Clt resin)±NH±CH2±C6H4±I, 5 (0.21 mmol/g) was
added in one portion to the solution of Pd(OAc)2 (5.6 mg,
25 mmol), PPh3 (6.6 mg, 25 mmol), and Bu4NCl (5.7 mg,
25 mmol) in DMF/H2O/Et3N (9:1:1, 3.0 ml) and the mixture
(21 mmol/ml concentration) was agitated for 2 h at 378C.
Two other batches of the resin (0.30 g each) were also
treated with the same reagent mixture for 4 and 8 h, respect-
ively, at 378C. Each resin was ®ltered, washed with DMF
and MeOH, and dried in vacuo. To 15 mg of each cyclized
resin 7 was added AcOH/TFE/CH2Cl2 (1:1:8, 0.50 ml) and
the mixture was agitated for 45 min at 258C. The resin was
®ltered and the ®ltrate was evaporated to dryness. The resi-
due was dissolved in 50% AcOH and analyzed by analytical
HPLC (Fig. 2a). The desired cyclized product 8 was eluted
at 13.7 min [CH3CN (20±80%/30 min)]: MALDI-TOF MS,
757.407 for [M1H]1 (Calcd 757.297 for C34H45N8O10S).
Amino acid analysis after 6 M HCl (1108C, 22 h) hydroly-
sis: Asp 1.00, Gly 2.05, Arg 0.88.
20 mg of the library resin was treated with AcOH/TFE/
CH2Cl2 (1:1:8, 0.50 ml) for 30 min at 258C. The resin was
®ltered and the ®ltrate was evaporated to dryness. The resi-
due was dissolved in DMF/H2O (2:1, 150 ml) and a 20 ml
aliquot of the solution was subjected to analytical HPLC
[CH3CN (30±70%/30 min)] (Fig. 4a). 1 ml of the eluate of
each peak was analyzed by MALDI-TOF MS. [Gly1,
gAbu3] derivative (Rt3 16.43 min) 913.086 for [M1H]1
(Calcd 913.792 for C36H50N8O10SI); [Gly1, Gly3] derivative
(Rt3 16.9 min) 885.336 for [M1H]1 (Calcd 885.740 for
C34H46N8O10SI); [Ala1, gAbu3] derivative (Rt3 17.20 min)
927.573 for [M1H]1 (Calcd 927.818 for C37H52N8O10SI);
[Gly1, eAhx3] derivative (Rt3 17.56 min) 941.54 for
[M1H]1 (Calcd 941.844 for C38H54N8O10SI); [Ala1, Gly3]
derivative (Rt3 17.76 min) 899.074 for [M1H]1 (Calcd
899.766 for C35H48N8O10SI); [Pro1, gAbu3] derivative (Rt3
17.98 min) 953.32 for [M1H]1 (Calcd 953.854 for
C39H54N8O10SI); [aAbu1, gAbu3] derivative (Rt3
18.32 min) 941.457 for [M1H]1 (Calcd 941.844 for
C38H54N8O10SI); [Ala1, eAhx3] derivative (Rt3 18.32 min)
955.462 for [M1H]1 (Calcd 955.870 for C39H56N8O10SI);
[Pro1, Gly3] derivative (Rt3 18.32 min) 925.457 for
[M1H]1 (Calcd 925.802 for C37H50N8O10SI); [aAbu1,
Gly3] derivative (Rt3 18.94 min) 913.451 for [M1H]1
(Calcd 913.792 for C36H50N8O10SI); [Pro1, eAhx3] deriva-
tive (Rt3 18.94 min) 981.544 for [M1H]1 (Calcd 981.906
for C41H58N8O10SI); [aAbu1, eAhx3] derivative (Rt3
19.38 min) 969.52 for [M1H]1 (Calcd 969.896 for
C40H58N8O10SI); [Nva1, gAbu3] derivative (Rt3 20.07 min)
954.968 for [M1H]1 (Calcd 955.870 for C39H56N8O10SI);
[Nva1, Gly3] derivative (Rt3 20.68 min) 927.477 for
[M1H]1 (Calcd 927.818 for C37H52N8O10SI); [Nva1,
eAhx3] derivative (Rt3 21.07 min) 983.511 for [M1H]1
(Calcd 983.922 for C41H60N8O10SI).
3.3.1. Cyclic RGD derivative, 9. AcOH/TFE/CH2Cl2
(1:1:8, 15 ml) was added to 0.56 g (0.12 mmol) of the
cyclized resin 7 (0.22 mmol/g), and the resin mixture was
stirred for 45 min at 258C. The mixture was ®ltered and the
solvent of the ®ltrate was removed by evaporation. To
the residue was added TFA/thioanisole (7.0±0.70 ml) and
the mixture was stirred for 3 h at 258C. The TFA of the
mixture was removed by evaporation at 258C and the resi-
due was extracted with 6 M guanidine HCl (5.0 ml). The
product in the extract was puri®ed by preparative HPLC
[YMC SH-343-5AM column (20£250 mm), CH3CN (10±
20%/60 min)] to yield 13 mg (overall 20%) of 9 as a white
powder: Rt2 10.4 min [CH3CN (10±20%)/30 min], 1H NMR
(600 MHz, D2O) d7.72 (d, Jˆ8.8 Hz, 1H), 7.54 (d,
Jˆ7.9 Hz, 1H), 7.50 (d, Jˆ15.8 Hz, 1H), 7.44 (t,
Jˆ7.9 Hz, 1H), 7.41 (d, Jˆ8.8 Hz, 1H), 7.36 (s, 1H), 7.35
(d, Jˆ7.9 Hz, 1H), 6.71 (d, Jˆ15.8 Hz, 1H), 4.78 (dd,
Jˆ9.1, 5.1 Hz, 1H), 4.63 (d, Jˆ16.9 Hz, 1H), 4.54 (dd,
Jˆ5.8, 8.5 Hz, 1H), 4.39 (d, Jˆ16.1 Hz, 1H), 4.37 (d,
Jˆ16.9 Hz, 1H), 4.09 (d, Jˆ17.6 Hz, 1H), 4.01 (d,
Jˆ17.6 Hz, 1H), 3.80 (d, Jˆ16.1 Hz, 1H), 3.19 (dd,
Jˆ6.9, 4.1 Hz, 2H), 2.98 (dd, Jˆ16.8 Hz, Jˆ5.1 Hz, 1H),
0.40 g (76 mmol) of the above library precursor resin was
added in one portion to a solution of Pd(OAc)2 (17 mg,
76 mmol), Ph3P (20 mg, 76 mmol), and Bu4NCl (17 mg,
76 mmol) in DMF/H2O/Et3N (9:1:1, 1.5 ml) and the mixture
(50 mmol/ml concentration) was agitated for 5 h at 378C.

K. Akaji et al. / Tetrahedron 57 (2001) 2293±2303
2301
The resin was ®ltered, washed with DMF and MeOH, and
dried in vacuo. 20 mg of the cyclized resin was treated with
AcOH/TFE/CH2Cl2 and analyzed by HPLC- MALDI-TOF
MS, as described for library precursor resin (Fig. 4b). [Gly1,
gAbu3] derivative (Rt3 8.17 min) 785.23 for [M1H]1
(Calcd 785.88 for C36H49N8O10S); [Gly1, Gly3] derivative
(Rt3 8.91 min) 757.36 for [M1H]1 (Calcd 757.828 for
C34H45N8O10S); [Gly1, eAhx3] derivative (Rt3 9.62 min)
813.652 for [M1H]1 (Calcd 813.932 for C38H53N8O10S);
[Ala1, gAbu3] derivative (Rt3 9.62 min) 799.593 for
[M1H]1 (Calcd 799.906 for C37H51N8O10S); [Ala1, Gly3]
derivative (Rt3 10.00 min) 771.443 for [M1H]1 (Calcd
771.854 for C35H47N8O10S); [Ala1, eAhx3] derivative (Rt3
10.07 min) 827.569 for [M1H]1 (Calcd 827.958 for
C39H55N8O10S); [Pro1, Gly3] derivative (Rt3 10.07 min)
797.519 for [M1H]1 (Calcd 797.890 for C37H49N8O10S);
[aAbu1, gAbu3] derivative (Rt3 11.91 min) 813.543 for
[M1H]1 (Calcd 813.932 for C38H53N8O10S); [aAbu1,
Gly3] derivative (Rt3 12.48 min) 785.136 for [M1H]1
(Calcd 785.880 for C36H49N8O10S); [Pro1, gAbu3] derivative
(Rt3 12.48 min) 825.186 for [M1H]1 (Calcd 825.942 for
C39H53N8O10S); [aAbu1, eAhx3] derivative (Rt3 12.48 min)
841.222 for [M1H]1 (Calcd 841.984 for C40H57N8O10S);
[Pro1, eAhx3] derivative (Rt3 12.48 min) 853.255 for
[M1H]1 (Calcd 853.994 for C41H57N8O10S); [Nva1,
gAbu3] derivative (Rt3 14.72 min) 827.58 for [M1H]1
(Calcd 827.958 for C39H55N8O10S); [Nva1, eAhx3] deriva-
tive (Rt3 15.07 min) 855.866 for [M1H]1 (Calcd 856.01 for
C41H59N8O10S); [Nva1, Gly3] derivative (Rt3 15.24 min)
799.526 for [M1H]1 (Calcd 799.906 for C37H51N8O10S).
C26H37N8O7); [Pro1, eAhx3] derivative (Rt4 22.87 min)
641.451 for [M1H]1 (Calcd 641.734 for C31H45N8O7);
[aAbu1. eAhx3] derivative (Rt4 22.87 min) 627.849 for
[M1H]1 (Calcd 628.716 for C30H44N8O7); [Nva1, gAbu3]
derivative (Rt4 24.64 min) 615.19 for [M1H]1 (Calcd
615.698 for C29H43N8O7); [Nva1, Gly3] derivative (Rt4
26.78 min) 587.225 for [M1H]1 (Calcd 587.646 for
C27H39N8O7); [Nva1, eAhx3] derivative (Rt4 26.78 min)
643.247 for [M1H]1 (Calcd 643.750 for C31N47N8O7).
3.5. Functional group library
Three batches of Fmoc-Asp(Clt resin)±NH±CH2±C6H4±I
(0.40 g each, 0.23 mmol/g resin) were used as the starting
resins. Fmoc-Gly-OH, Fmoc-gAbu-OH, and Fmoc-eAhx-
OH at position 4, and Fmoc-Ser(But)-OH, Fmoc-
Dpr(Boc)-OH, Fmoc-Lys(Boc)-OH, Fmoc-Asp(OBut)-OH,
and Fmoc-Asn(Trt)-OH at position 1 were incorporated by
the split and mix method as described for the alkyl library to
yield 1.3 g of the precursor library resin. An aliquot of the
library resin was cleaved with AcOH/TFE/CH2Cl2 (1:1:8)
and each component was identi®ed by HPLC[CH3CN (30±
70%/30 min)]-HRFAB MS analysis. [Ser(But)1, gAbu3]
derivative (Rt3 21.91 min) 999.3152 for [M1H]1 (Calcd
999.3147 for C41H60O11N8SI); [Dpr(Boc)1, gAbu3] deriva-
tive (Rt3 22.27 min) 1042.3196 for [M1H]1 (Calcd
1042.3205 for C42H61O12N9SI); [Ser(But)1, Gly3] derivative
(Rt3 22.50 min) 971.2841 for [M1H]1 (Calcd 971.2834 for
C39H56O11N8SI); [Ser(But)1, eAhx3] derivative (Rt3
22.83 min) 1027.3472 for [M1H]1 (Calcd 1027.3460 for
C43H64O11N8SI); [Dpr(Boc)1, Gly3] derivative (Rt3
22.83 min) 1014.2876 for [M1H]1 (Calcd 1014.2892 for
C40H57O12N9SI); [Asp(OBut)1, gAbu3] derivative (Rt3
22.93 min) 1027.3088 for [M1H]1 (Calcd 1027.3096 for
C42H60O12N8SI); [Dpr(Boc)1, eAhx3] derivative (Rt3
23.16 min) 1070.3514 for [M1H]1 (Calcd 1070.3518 for
C44H65O12N9SI); [Lys(Boc)1, gAbu3] drivative (Rt3
23.27 min) 1084.3665 for [M1H]1 (Calcd 1084.3675 for
C45H67O12N9SI); [Asp(OBut)1, Gly3] derivative (Rt3
23.50 min) 999.2765 for [M1H]1 (Calcd 999.2783 for
C40H56O12N8SI); [Lys(Boc)1, Gly3] derivative (Rt3
23.92 min) 1056.3373 for [M1H]1 (Calcd 1056.3362 for
C43H63O12N9SI); [Asp(OBut)1, eAhx3] derivative (Rt3
23.92 min) 1055.3414 for [M1H]1 (Calcd 1055.3409 for
C44H64O12N8SI); [Lys(Boc)1, eAhx3] derivative (Rt3
24.14 min) 1112.4023 for [M1H]1 (Calcd 1112.3988 for
C47H71O12N9SI); [Asn(Trt)1, gAbu3] derivative (Rt3
31.39 min) 1212.3714 for [M1H]1 (Calcd 1212.3726 for
C57H67O11N9SI); [Asn(Trt)1, Gly3] derivative (Rt3
31.39 min) 1184. 3420 for [M1H]1 (Calcd 1184.3413 for
C55H63O11N9SI); [Asn(Trt)1, eAhx3] derivative (Rt3
31.82 min) 1240.4065 for [M1H]1 (Calcd 1240.4039 for
C59H71O11N9SI).
To 80 mg (15 mmol) of the above cyclized library resin was
added AcOH/TFE/CH2Cl2 (1:1:8, 2.0 ml), and the mixture
was stirred for 40 min at 258C. The mixture was ®ltered and
the ®ltrate was evaporated to dryness. To the residue was
added TFA/thioanisole (2.0±0.20 ml), and the mixture was
stirred for 3 h at 258C. The TFA of the mixture was removed
by evaporation at 258C and the residue was extracted with
4 M AcOH (2.0 ml). The aqueous phase was washed with
Et2O (2.0 ml), ®ltered, and then lyophilized. The resulting
powdered product was partially puri®ed by semipreparative
HPLC [CH3CN (10±30%/60 min)]: the product which
eluted below 25% concentration of CH3CN was pooled
and lyophilized to yield 0.90 mg (10%) of a white powder.
Each component was identi®ed by HPLC[CH3CN (10±
30%/30 min)]- MALDI-TOF MS analysis (Fig. 5). [Gly1,
gAbu3] derivative (Rt4 15.22 min) 573.353 for [M1H]1
(Calcd 573.620 for C26H37N8O7); [Ala1, gAbu3] derivative
(Rt4 16.59 min) 587.452 for [M1H]1 (Calcd 587.646 for
C27H39N8O7); [Gly1, Gly3] derivative (Rt4 16.92 min)
545.418 for [M1H]1 (Calcd 545.568 for C24H33N8O7);
[Gly1, eAhx3] derivative (Rt4 18.47 min) 601.503 for
[M1H]1 (Calcd 601.672 for C28H41N8O7); [Ala1, Gly3]
derivative (Rt4 18.47 min) 559.426 for [M1H]1 (Calcd
559.594 for C25H35N8O7); [Pro1, Gly3] derivative (Rt4
19.68 min) 584.957 for [M1H]1 (Calcd 585.630 for
C27H37N8O7); [Ala1, eAhx3] derivative (Rt4 20.08 min)
615.228 for [M1H]1 (Calcd 615.698 for C29H43N8O7);
[aAbu1, gAbu3] derivative (Rt4 20.08 min) 601.433 for
[M1H]1 (Calcd 601.672 for C28H41N8O7); [Pro1, gAbu3]
derivative (Rt4 21.43 min) 613.386 for [M1H]1 (Calcd
613.682 for C29H41N8O7); [aAbu1, Gly3] derivative (Rt4
22.18 min) 573.28 for [M1H]1 (Calcd 573.620 for
Cyclization of the precursor resin (0.40 g, 76 mmol) and the
HPLC[CH3CN (30±70%/30 min)]-HRFAB MS analysis
were carried out using the same procedure as described
for the alkyl library. [Ser(But)1, gAbu3] derivative (Rt3
17.24 min) 871.4006 for [M1H]1 (Calcd 871.4024 for
C41H59O11N8S); [Ser(But)1, eAhx3] derivative (Rt3
17.60 min) 899.4359 for [M1H]1 (Calcd 899.4337 for

2302
K. Akaji et al. / Tetrahedron 57 (2001) 2293±2303
C43H63O11N8S); [Ser(But)1, Gly3] derivative (Rt3 17.92 min)
843.3719 for [M1H]1 (Calcd 843.3711 for C39H55O11N8S);
[Dpr(Boc)1, Gly3] derivative (Rt3 17.92 min) 886.3779 for
[M1H]1 (Calcd 886.3769 for C40H56O12N9S); [Dpr(Boc)1,
gAbu3] derivative (Rt3 17.92 min) 914.4070 for [M1H]1
(Calcd 914.4082 for C42H60O12N9S); [Dpr(Boc)1, eAhx3]
derivative (Rt3 17.92 min) 942.4366 for [M1H]1 (Calcd
942.4395 for C44H64O12N9S); [Asp(OBut)1, Gly3] derivative
(Rt3 18.38 min) 871.3668 for [M1H]1 (Calcd 871.3660 for
C40H55O12N8S); [Asp(OBut)1, gAbu3] derivative (Rt3
18.38 min) 899.3984 for [M1H]1 (Calcd 899.3973 for
C42H59O12N8S); [Asp(OBut)1, eAhx3] derivative (Rt3
18.58 min) 927.4303 for [M1H]1 (Calcd 927.4286 for
C44H63O12N8S); [Lys(Boc)1, gAbu3] derivative (Rt3
18.94 min) 956.4570 for [M1H]1 (Calcd 956.4552 for
C45H66O12N9S); [Asn(Trt)1, eAhx3] derivative (Rt3
18.94 min) 1112.4366 for [M1H]1 (Calcd 1112.4916 for
C59H70O11N9S); [Lys(Boc)1, Gly3] derivative (Rt3
19.22 min) 928.4230 for [M1H]1 (Calcd 928.4239 for
C43H62O12N9S); [Lys(Boc)1, eAhx3] derivative (Rt3
19.22 min) 984.4874 for [M1H]1 (Calcd 984.4865 for
C47H70O12N9S); [Asn(Trt)1, Gly3] derivative (Rt3
19.22 min) 1056.5254 for [M1H]1 (Calcd 1056.4290 for
C55H62O11N9S); [Asn(Trt)1, gAbu3] derivative (Rt3
19.22 min) 1084.3840 for [M1H]1 (Calcd 1084.4603 for
C57H66O11N9S).
3.6. Inhibition assay of ®brinogen binding to
immobilized GPIIb/IIIa
Human GpIIb/IIIa (Enzyme Research Lab.) was diluted
with Tris buffer containing 20 mM Tris-HCl, 150 mM
NaCl, 1 mM CaCl2 and 0.02% NaN3 (®nal concentration
was 5 mg/ml). The solution was then immediately added
to 96-well microtiter plates at 100 ml per well. Each well
was washed with 50 mM Tris, 100 mM NaCl, 2 mM CaCl2,
and 0.02% NaN3 (pH 7.4) and then incubated with EIA
buffer containing 35 mg/ml bovine serum albumin for 3 h
at room temperature. After each well was washed with EIA
buffer containing 1 mg/ml BSA, ®brinogen solution (30 mg/
ml) was added to each well and incubated for 3 h at room
temperature in the presence or absence of peptides. Follow-
ing the incubation each well was washed twice with 400 ml
of EIA buffer. Bound ®brinogen was quantitated by the
addition of 100 ml of an anti-human ®brinogen antibody
conjugated to alkaline phosphatase (1:1000 dilution, EY
laboratories), followed by a wash with EIA buffer and the
addition of p-nitrophenyl phosphate (pNPP Tablets, Sigma).
The resulting IC50 value was 2.0£1025 M.
3.7. Molecular dynamics simulations
The ROESY experiments were carried out using mixing
time 300 ms with a spin locking ®eld 30 dB. The spectral
width was 4085 Hz. Molecular dynamics (MD) simulation
were done using the InsightII/Discover program (Biosym
Technologies) on an INDY3 computer. The simulations
were done using the consistent valence force ®eld
(CVFF). MD simulation for 50 ps at 1000 K was carried
out. Every 10 fs (20 steps) a structure was stored in the
history ®le for analysis. The history was analyzed by plot-
ting the energy against time. Twelve conformers those
provided local minimum in the plot were sampled as a start-
ing conformer of next MD simulation. These conformers
were subjected to the 300 K MD simulation with 17 ROE
restrains for 10 ps. Every 1 ps (1000 steps) a structure was
stored in the history ®le. Then the 132 conformers were
subjected to unrestrained minimization using the VA09A
algorithm until the maximum derivative was less than
0.001 kcal/mol. An error of ^308 was tolerated for the phi
angles calculated from J(NH±Ha) of 2Arg and 4Asp. The
conformers that were inconsistent with this threshold were
discarded. Ninety-seven conformers were analyzed. These
conformers were clustered into three groups in their
2Arg±4Asp phi torsion angle plot. Each group has another
total energies (81.3, 82.1,85.6 averaged, relatively) with
small standard deviation (SDˆ1.33, 1.87, 1.28). In low-
energy group, conformers have phi angle around 21208
and 2808 in 2Arg and 4Asp. Distance between guanidino
carbon of 2Arg and carbonyl carbon of 4Asp was 11.77 A
(averaged) with SDˆ1.54. Details of the procedure and
results will be published elsewhere.
Cleavage and deprotection of the cyclized resin (50 mg,
10 mmol) were carried out using the same procedure as
described for the alkyl library. The crude product was
partially puri®ed by semipreparative HPLC [CH3CN (10±
30%/60 min)]: the product, which eluted below 20%
concentration of CH3CN was pooled and lyophilized to
yield 0.50 mg (9.0%) of a white powder. Each component
was identi®ed by HPLC [CH3CN (10±30%/60 min)]-
HRFAB MS analysis (Fig. 7). [Dpr1, gAbu3] derivative
(Rt5 11.32 min) 602.3055 for [M1H]1 (Calcd 602.3051
for C27H40O7N9); [Dpr1, Gly3] derivative (Rt5 11.32 min)
574.2744 for [M1H]1 (Calcd 574.2738 for C25H36O7N9);
[Lys1, gAbu3] derivative (Rt5 14.51 min) 644.3526 for
[M1H]1 (Calcd 644.3520 for C30H46O7N9); [Lys1, Gly3]
derivative (Rt5 15.70 min) 616.3212 for [M1H]1 (Calcd
616.3207 for C28H42O7N9); [Dpr1, eAhx3] derivative (Rt5
16.51 min) 630.3358 for [M1H]1 (Calcd 630.3364 for
C29H44O7N9); [Asn1, gAbu3] derivative (Rt5 17.15 min)
630.3005 for [M1H]1 (Calcd 630.3000 for C28H40O8N9);
[Ser1, gAbu3] derivative (Rt5 18.12 min) 603.2885 for
[M1H]1 (Calcd 603.2891 for C27H39O8N8); [Asn1, Gly3]
derivative (Rt5 18.83 min) 602.2675 for [M1H]1 (Calcd
602.2687 for C26H36O8N9); [Ser1, Gly3] derivative (Rt5
19.42 min) 575.2588 for [M1H]1 (Calcd 575.2578 for
C25H35O8N8); [Asp1, gAbu3] derivative (Rt5 20.18 min)
631.2833 for [M1H]1 (Calcd 631.2840 for C28H39O9N8);
[Lys1, eAhx3] derivative (Rt5 20.18 min) 672.3843 for
[M1H]1 (Calcd 672.3833 for C32H50O7N9); [Asp1, Gly3]
derivative (Rt5 22.51 min) 603.2517 for [M1H]1 (Calcd
603.2527 for C26H35O9N8); [Asn1, eAhx3] derivative (Rt5
23.35 min) 658.3320 for [M1H]1 (Calcd 658.3313 for
C30H44O8N9); [Ser1, eAhx3] derivative (Rt5 24.34 min)
631.3212 for [M1H]1 (Calcd 631.3204 for C29H43O8N8);
[Asp1, eAhx3] derivative (Rt5 26.44 min) 659.3163 for
[M1H]1 (Calcd 659.3153 for C30H43O9N8).
Acknowledgements
The authors are grateful to Ms Kayoko Oda (Kyoto
Pharmaceutical University) for measuring the HRFAB MS
spectra.

K. Akaji et al. / Tetrahedron 57 (2001) 2293±2303
2303
Uno, I.; Ojima, I.; Hayashi, Y. Bioorg. Med. Chem. Lett. 1997,
7, 209±212. (b) Jackson, S.; DeGrade, W.; Dwivedi, A.;
Parthasarathy, A.; Higley, A.; Krywko, J.; Rockwell, A.;
Markwalder, J.; Wells, G.; Wexler, R.; Mousa, S.; Harlow,
R. J. Am. Chem. Soc. 1994, 116, 3220±3230.
References
1. (a) Lorsbach, B. A.; Kurth, M. J. Chem. Rev. 1999, 99, 1549±
1581. (b) Thompson, L. A.; Ellman, J. A. Chem. Rev. 1996,
96, 555±600. (c) Gordon, E. M.; Gallop, M. A.; Patel, D. V.
Acc. Chem. Res. 1996, 29, 144±154.
13. Carpino, L. A.; Han, G. Y. J. Am. Chem. Soc. 1970, 92, 5748±
5749.
2. Heck, R. F. Org. React. 1982, 27, 345±390.
14. Akaji, K.; Kuriyama, N.; Kiso, Y. J. Org. Chem. 1996, 61,
3350±3357.
3. (a) Hiroshige, M.; Hauske, J. R.; Zhou, P. Tetrahedron Lett.
1995, 36, 4567±4570. (b) Yu, K.-L.; Deshpande, M. S.; Vyas,
D. M. Tetrahedron Lett. 1994, 35, 8919±8922. (c) Jeffery, T.
Tetrahedron Lett. 1994, 35, 3051±3054.
15. Wang, S. S. J. Am. Chem. Soc. 1973, 95, 1328±1333.
16. (a) Barlos, K.; Gatos, D.; Kallitsis, J.; Papaphotiu, G.; Sotiriu,
P.; Wenqing, Y.; Schafer, W. Tetrahedron Lett. 1989, 30,
3943±3946. (b) Barlos, K.; Gatos, D.; Kapolos, S.;
Papaphoutiu, G.; Schafer, W.; Wenqing, Y. Tetrahedron
Lett. 1989, 30, 3947±3950.
4. (a) Fagnola, M. C.; Candiani, I.; Visentin, G.; Cabri, W.;
Zarini, F.; Mongelli, N.; Bedeschi, A. Tetrahedron Lett.
1997, 38, 2307±2310. (b) Arumugam, V.; Routledge, A.;
Abell, C.; Balasubramanian, S. Tetrahedron Lett. 1997, 38,
6473±6476. (c) Yun, W.; Mohan, R. Tetrahedron Lett.
1996, 37, 7189±7192.
17. Fujino, M.; Wakimasu, M.; Kitada, C. Chem. Pharm. Bull.
1981, 29, 2825±2831.
18. Kaiser, E.; Colescott, R. L.; Bossinger, D. C.; Cook, P. I. Anal.
Biochem. 1970, 34, 595±598.
5. Fanelli, D.; Fagnola, M. C.; Severino, D.; Bedeschi, A.
Tetrahedron Lett. 1997, 38, 2311±2314.
19. (a) Peyman, A.; Wehner, V.; Knolle, J.; Stilz, H. V.;
Breipohl, G.; Scheunemann, K-H.; Carniato, D.;
Ruxer, J-M.; Gourvest, J-F.; Gadek, T. R.; Bodary, S. Bioorg.
Med. Chem. 2000, 10, 179±182. (b) McDowell, R. S.;
Gadek, T. R.; Barker, P. L.; Burdick, D. J.; Chan, K. S.;
Quan, C. L.; Skelton, N.; Struble, M.; Thorsett, E. D.;
Tischler, M.; Tom, J. Y. K.; Webb, T. R.; Burnier, J. P.
J. Am. Chem. Soc. 1994, 116, 5069±5076.
6. Goff, D. A.; Zuckermann, R. N. J. Org. Chem. 1995, 60,
5748±5749.
7. Hiroshige, M.; Hauske, J. R.; Zhou, P. J. Am. Chem. Soc.
1995, 117, 11590±11591.
8. Mazur, S.; Jayalekshmy, P. J. Am. Chem. Soc. 1979, 101,
677±683.
9. (a) Chen, L.; Barany, G. Lett. Peptide Sci. 1996, 3, 283±292.
(b) Al-Obeidi, F.; Castrucci, A. M. de L.; Hadley, M. E.;
Hruby, V. J. J. Med. Chem. 1989, 32, 2555±2561.
(c) Eritja, R.; Ziehler-Martin, J. P.; Walker, P. A.;
Lee, T. D.; Legesse, K.; Albericio, F.; Kaplan, B. E.
Tetrahedron 1987, 43, 2675±2680. (d) Lebl, M.; Hruby,
V. J. Tetrahedron Lett. 1984, 25, 2067±2068.
 Â
20. Furka, A.; Sebestyen, F.; Asgedom, M.; Dibo, G. Int. J.
Peptide Protein Res. 1991, 37, 487±493.
21. Waki, M.; Kitajima, Y.; Izumiya, N. Synthesis 1981, 266±268.
22. Meienhofer, J.; Waki, M.; Heimer, E. P.; Lambros, T. J.;
Makofske, R. C.; Lhang, C-D. Int. J. Peptide Protein Res.
1979, 13, 35±42.
10. A part of this work was preliminary reported; Akaji, K.;
Kiso, Y. Tetrahedron Lett. 1997, 38, 5185±5188.
11. Ruoslahti, E.; Pierschbacher, M. D. Science 1987, 238, 491±
497.
23. Hayashi, Y.; Katada, J.; Sato, Y.; Igarashi, K.; Takiguchi, Y.;
Harada, T.; Muramatsu, M.; Yasuda, E.; Uno, I. Bioorg. Med.
Chem. 1998, 6, 355±364.
24. Bogdanowich-Knipp, S. J.; Jois, D. S. S.; Siahaan, T. J.
J. Peptide Res. 1999, 53, 523±529.
12. (a) Harada, T.; Katada, J.; Tachiki, A.; Asari, T.; Iijima, K.;