Rediscovering an endothelin antagonist (BQ-123): A self-deconvoluting cyclic pentapeptide library

Article abstract of DOI:10.1021/jm9604078

A 'self-deconvoluting' cyclic pentapeptide library, designed to produce 82 944 head-to-tail-linked peptides in 48 vials, has been prepared. The mixture included amine acids found in a recently optimized endothelin antagonist, BQ-123, originally isolated from microbial sources by Banyu investigators. Using a positional scan approach, the most potent of 12 residues at each of the four variable positions uniquely rediscovered the BQ- 123 sequence or cyclo(L-Pro-D-Val-L-Leu-D-Trp-D-Asp). Resynthesis of the four most potent amine acid combinations gave the following values of relative potency: cyclo(L-Pro-D-Val-L-Leu-D-Trp-D-Asp) or BQ-123 = 1.0, cyclo(L-Pro- D-Pro-L-Leu-D-Trp-D-Asp) = 0.0, cyclo(L-Pro-D-Pro-L-Trp-D-Trp-D-Asp) = 0.0, and cyclo(L-Pro-D-Val-L-Trp-D-Trp-D-Asp) = 0.1. This study reflects the first time that the positional scan approach has been applied to cyclic peptide libraries using a known target. Although no analogs more potent than BQ-123 were discovered, our results provide verification of our synthetic methods for preparing head-to-tail cyclic peptide libraries and also lend support to the use of carefully designed sublibraries for the rapid elucidation of potential leads within a relatively constrained set of peptide macrocycles.

Full text of DOI:10.1021/jm9604078

3842
J . Med. Chem. 1996, 39, 3842-3846
Red iscover in g a n En d oth elin An ta gon ist (BQ-123): A Self-Decon volu tin g Cyclic
P en ta p ep tid e Libr a r y
Arno F. Spatola* and Yvon Crozet
Department of Chemistry, University of Louisville, Louisville, Kentucky 40292
Damiane deWit and Masashi Yanagisawa
Howard Hughes Medical Institute, Department of Molecular Genetics, University of Texas Southwestern Medical Center,
Dallas, Texas 75235
Received J une 7, 1996X
A “self-deconvoluting” cyclic pentapeptide library, designed to produce 82 944 head-to-tail-
linked peptides in 48 vials, has been prepared. The mixture included amino acids found in a
recently optimized endothelin antagonist, BQ-123, originally isolated from microbial sources
by Banyu investigators. Using a positional scan approach, the most potent of 12 residues at
each of the four variable positions uniquely rediscovered the BQ-123 sequence or cyclo(L-Pro-
D-Val-L-Leu-D-Trp-D-Asp). Resynthesis of the four most potent amino acid combinations gave
the following values of relative potency: cyclo(L-Pro-D-Val-L-Leu-D-Trp-D-Asp) or BQ-123 )
1.0, cyclo(L-Pro-D-Pro-L-Leu-D-Trp-D-Asp) ) 0.0, cyclo(L-Pro-D-Pro-L-Trp-D-Trp-D-Asp) ) 0.0, and
cyclo(L-Pro-D-Val-L-Trp-D-Trp-D-Asp) ) 0.1. This study reflects the first time that the positional
scan approach has been applied to cyclic peptide libraries using a known target. Although no
analogs more potent than BQ-123 were discovered, our results provide verification of our
synthetic methods for preparing head-to-tail cyclic peptide libraries and also lend support to
the use of carefully designed sublibraries for the rapid elucidation of potential leads within a
relatively constrained set of peptide macrocycles.
In tr od u ction
efficiency of moderately large mixtures combined with
a positional scan approach as applied to cyclic peptides
and related compounds.
A variety of peptide and peptide-related structures
have been prepared as combinatorial mixtures for drug
lead discovery.1-3 Most of these have been linear
structures, although there is increasing interest in
nonpeptide organics as more rigid functional group
templates.4-7
Cyclic peptide mixtures represent a middle ground
between constrained systems and linear peptide librar-
ies.8 Naturally occurring cyclic peptide “libraries” in the
form of disulfide-bridged peptide toxins have been found
in many of the 500 species of Conus, each of which may
contain 100 or more constrained peptides.9 The first
examples of synthetic disulfide-bridged libraries were
prepared by De Grado and co-workers, who oxidized
phage display libraries containing pairs of cysteines to
form their cyclic counterparts.10 But these compounds
still retain free amine and carboxyl termini and are thus
susceptible to proteolytic attack.
Head-to-tail cyclization confers partial enzyme resis-
tance (toward exopeptidases) but preserves the rich
diversity of stereochemically defined side chains char-
acteristic of amino acid-derived analogs. A particularly
facile method for preparing head-to-tail cyclic peptides
using aspartic acid linked to a polystyrene resin was
recently reported.11 A variety of other methods for
preparing cyclic peptides has also been described.12,13
We have previously described the utility of side chain
attachment and resin-bound cyclization for the prepara-
tion of relatively clean cyclic peptide mixtures.14,15 We
now wish to report the preparation and bioassay of a
cyclic pentapeptide library designed to validate the
Resu lts a n d Discu ssion
The starting point for our synthesis was a single
D-aspartic acid linked to a solid phase support through
its side chain â-carboxylic acid function. Boc-D-Asp-
OFm (1) was initially attached to a p-hydroxymethyl
poly(styrene) resin with a functionalization of 0.8 mequiv/
g. The substitution yield was evaluated by cleavage and
UV quantification of fluorenylmethanol on a known
amount of acylated resin16 (2) and was found to be 0.52
mmol/g. Synthesis of the libraries (Table 1) was per-
formed in a synthesis block17 (Coshisoft, Tucson) com-
posed of 42 poly(propylene) syringes equipped with a
plastic frit. Each of the syringes was loaded with 150
mg of 2 and acted as an individual reaction vessel. Boc
chemistry18,19 was used to elongate the peptide and
included 50% trifluoroacetic acid cleavage and a phos-
phonium-based20 condensing agent (BOP/HOBt). The
resin was kept in suspension through shaking of the
synthesis block with a rotary shaker (3D rotator, CMS).
Completions of the couplings were monitored by the
Kaiser ninhydrin test21 on four syringes chosen at
random at each step. In the event of a positive (blue)
color, the couplings for all vials were repeated.
At each coupling step, either a mixture of protected
amino acids was used, consisting of six D- and six
L-configuration amino acids, or a single amino acid of
the composite was introduced. The amino acids selected
were D- and L-versions of Boc-Trp, Boc-Leu, Boc-Glu-
(OBzl), Boc-Arg(Tos), Boc-Val, and Boc-Pro. These
choices were based on a desire both to provide diver-
sity22 in the classic Hansch characteristics23 (steric,
electronic, and hydrophobic factors) as well as to include
* Address correspondence to this author. Tel: 502-852-5979. Fax:
502-852-8149. E-mail: afspat01@ulkyvm.louisville.edu.
X Abstract published in Advance ACS Abstracts, August 15, 1996.
S0022-2623(96)00407-4 CCC: $12.00 © 1996 American Chemical Society

Notes
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 19 3843
Ta ble 1. Structure and Endothelin Binding Activities of the 48 Cyclic Peptide Libraries, where Xxx is a Nearly Equimolar Mixture
of the 12 Following Amino Acids: D,L-Proline, D,L-Valine, D,L-Leucine, D,L-Tryptophan, D,L-Arginine, and D,L-Glutamic Acida
% binding
inhibition
% binding
inhibition
code
structure
code
structure
L1
L2
L3
L4
L5
L6
L7
L8
cyclo(Xxx-Xxx-Xxx-L-Pro-D-Asp)
cyclo(Xxx-Xxx-Xxx-L-Val-D-Asp)
cyclo(Xxx-Xxx-Xxx-L-Leu-D-Asp)
cyclo(Xxx-Xxx-Xxx-L-Trp-D-Asp)
cyclo(Xxx-Xxx-Xxx-L-Arg-D-Asp)
cyclo(Xxx-Xxx-Xxx-L-Glu-D-Asp)
cyclo(Xxx-Xxx-Xxx-D-Pro-D-Asp)
cyclo(Xxx-Xxx-Xxx-D-Val-D-Asp)
cyclo(Xxx-Xxx-Xxx-D-Leu-D-Asp)
cyclo(Xxx-Xxx-Xxx-D-Trp-D-Asp)
cyclo(Xxx-Xxx-Xxx-D-Arg-D-Asp)
cyclo(Xxx-Xxx-Xxx-D-Glu-D-Asp)
cyclo(Xxx-Xxx-L-Pro-Xxx-D-Asp)
cyclo(Xxx-Xxx-L-Val-Xxx-D-Asp)
cyclo(Xxx-Xxx-L-Leu-Xxx-D-Asp)
cyclo(Xxx-Xxx-L-Trp-Xxx-D-Asp)
cyclo(Xxx-Xxx-L-Arg-Xxx-D-Asp)
cyclo(Xxx-Xxx-L-Glu-Xxx-D-Asp)
cyclo(Xxx-Xxx-D-Pro-Xxx-D-Asp)
cyclo(Xxx-Xxx-D-Val-Xxx-D-Asp)
cyclo(Xxx-Xxx-D-Leu-Xxx-D-Asp)
cyclo(Xxx-Xxx-D-Trp-Xxx-D-Asp)
cyclo(Xxx-Xxx-D-Arg-Xxx-D-Asp)
cyclo(Xxx-Xxx-D-Glu-Xxx-D-Asp)
0
0
0
L25
L26
L27
L28
L29
L30
L31
L32
L33
L34
L35
L36
L37
L38
L39
L40
L41
L42
L43
L44
L45
L46
L47
L48
cyclo(Xxx-L-Pro-Xxx-Xxx-D-Asp)
cyclo(Xxx-L-Val-Xxx-Xxx-D-Asp)
cyclo(Xxx-L-Leu-Xxx-Xxx-D-Asp)
cyclo(Xxx-L-Trp-Xxx-Xxx-D-Asp)
cyclo(Xxx-L-Arg-Xxx-Xxx-D-Asp)
cyclo(Xxx-L-Glu-Xxx-Xxx-D-Asp)
cyclo(Xxx-D-Pro-Xxx-Xxx-D-Asp)
cyclo(Xxx-D-Val-Xxx-Xxx-D-Asp)
cyclo(Xxx-D-Leu-Xxx-Xxx-D-Asp)
cyclo(Xxx-D-Trp)-Xxx-Xxx-D-Asp)
cyclo(Xxx-D-Arg-Xxx-Xxx-D-Asp)
cyclo(Xxx-S-Glu-Xxx-Xxx-D-Asp)
cyclo(L-Pro-Xxx-Xxx-Xxx-D-Asp)
cyclo(L-Val-Xxx-Xxx-Xxx-D-Asp)
cyclo(L-Leu-Xxx-Xxx-Xxx-D-Asp)
cyclo(L-Trp-Xxx-Xxx-Xxx-D-Asp)
cyclo(L-Arg-Xxx-Xxx-Xxx-D-Asp)
cyclo(L-Glu-Xxx-Xxx-Xxx-D-Asp)
cyclo(D-Pro-Xxx-Xxx-Xxx-D-Asp)
cyclo(D-Val-Xxx-Xxx-Xxx-D-Asp)
cyclo(D-Leu-Xxx-Xxx-Xxx-D-Asp)
cyclo(D-Trp-Xxx-Xxx-Xxx-D-Asp)
cyclo(D-Arg-Xxx-Xxx-Xxx-D-Asp)
cyclo(D-Glu-Xxx-Xxx-Xxx-D-Asp)
10.8
11.6
7.0
3.0
0.3
0
31.2
37.3
11.5
4.9
0
8.5
43.6
0
0
0
0
0
8.4
0
0
10.3
15.2
0
0
4.0
7.7
0
0
0
72.3
0
0
0
0
76.7
45.4
15.1
1.2
0
0
0
14.8
0
0
L9
L10
L11
L12
L13
L14
L15
L16
L17
L18
L19
L20
L21
L22
L23
L24
a
The data reflect percent inhibition of ET1 binding, corrected for nonspecific binding.
the residues found in the potent cyclic peptide antago-
nist BQ-123 or cyclo(L-Pro-D-Val-L-Leu-D-Trp-D-Asp).24
When using a mixture, the total excess of amino acids
was 4-fold, and the individual ratios were adjusted to
values comparable to those reported by Geysen et al.,25
empirically determined to produce nearly equivalent
product ratios.
The cyclic peptides were then prepared in groups of
12 × 12 × 12 × 1 or 1728-component mixtures. In each
of 48 vials, aspartic acid and a second amino acid in
one of the remaining four positions were held as
constants. This approach is predicted to result in the
redundant formation of 20 736 × 4 or 82 944 peptides.
Among the 48 vials, any one particular sequence is
represented four times, thus providing a possibility of
discovering optimized lead candidates in a single series
of bioassays, without the need for extensive deconvolu-
tion from repeated synthesis. This novel approach was
first described by Houghten and colleagues26,27 and used
to rediscover the sequences of leucine and methionine
enkephalin among a series of linear pentapeptides.
Following the completion of the synthesis of the linear
pentapeptides, the N-terminal Boc protecting group was
removed with trifluoroacetic acid and the C-terminal
fluorenylmethyl ester R-aspartyl protecting function was
removed with 20% piperidine. On-resin cyclization was
carried out using BOP/HOBt for 2 h and was repeated
three times to obtain a negative Kaiser test. A multiple
HF apparatus28 permitted the simultaneous treatment
of 24 libraries using anisole as a scavenger. After
removal of HF, the peptide products were extracted with
a 30% acetic acid solution. The latter was washed with
diethyl ether and lyophilized. Salt free peptides were
afforded after reversed phase solid phase extraction
using Varian Bond Elut C18. Model studies, monitored
by mass spectrometry, suggested that both hydrophobic
and hydrophilic cyclic peptides are recoverable intact
following the procedure (unpublished observation). The
48 mixtures were lyophilized individually to yield a final
weight of peptide product ranging from 21.2 to 34.8 mg.
F igu r e 1. Rediscovering Banyu cyclic pentapeptide endo-
thelin antagonist cyclo(Pro-D-Val-Leu-D-Trp-D-Asp). Relative
potencies of 48 vials containing 1728 cyclic peptides are given
as reciprocals of ETA binding concentrations, after correction
for nonspecific binding.
Biological assays for endothelin receptor binding were
performed in vitro using a cell culture assay. In order
to assess biological activity of peptide libraries, we
performed competitive radioligand binding assay with
[
125I]ET-1 in CHO cell monolayers expressing recombi-
nant human endothelin-A (ETA) receptors. Each library
was added to the bioassay at a total peptide concentra-
tion of 800 µM. This resulted in a nominal concentra-
tion of individual cyclic peptide species of approximately
0.5 µM in the bioassay, assuming a uniform combina-
torial synthesis. The results of the screening from a
single-point assay allowed the determination of the most
active amino acid for each position. Maximal enhance-
ment of activity (Figure 1) was observed for L-Pro at
the first position, D-Val at the second, L-Leu at the third,
and D-Trp at the fourth. This combination in fact
corresponds exactly to the residues found in the Banyu

3844 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 19
Notes
antagonist24 BQ-123. Nevertheless, a more modest
activity was also observed for D-Pro at the second
position n and for L-Trp at the third. We were aware
that the activity of a library containing two slightly
active molecules could have a global potency higher than
the activity measured for a mixture containing the most
active peptide. Furthermore, the occurrence in a library
of several molecules with different activity intensity is
very probable since it is commonly observed that only
a few amino acids in a peptide sequence are truly critical
for the biological response. In this respect, the peptide
with the apparently preferred combination of amino
acids in a positional scan assay may not be the optimum
combination for maximum potency.
Accordingly, four new peptides (P 1-P 4) were syn-
thesized based upon the four apparently best combina-
tions. The four analogs were synthesized using solid
phase methods using the same approach as with the
peptide mixtures. Purification by reversed phase high-
performance liquid chromatography (RP-HPLC) af-
forded four compounds whose structures were further
defined by electrospray mass spectrometry (ES-MS) (see
Experimental Section) and by comparison to the au-
thentic compound in the case of BQ-123.
Next, dose-activity response studies were carried out
on each of these four single compounds. The dose-
activity response curves (Figure 2) from a seven-point
assay indicated the highest endothelin receptor binding
activity for P 3 which corresponds to the original Banyu
compound. P 4, which substitutes an L-tryptophan for
the L-leucine in BQ-123, possessed only a modest
endothelin receptor binding activity, while P 1 and P 2,
both of which contain two prolines, were totally inactive.
The lack of activity for P 1 and P 2 either could reflect
false positives in the binding assay or might suggest
possible synergism among these complex mixtures.
Con clu sion s
The major objective of this study was to validate our
synthetic approach for the preparation of relatively pure
cyclic peptide libraries. The major synthesis features,
including side chain attachment of the first amino acid
to the resin, and on-resin cyclization appear as reason-
able and effective choices for the preparation of head-
to-tail cyclic peptide libraries.
Our study presented two secondary goals: on the one
hand, validation of the positional scanning approach for
medium size cyclic peptide libraries and, on the other
hand, an attempt to find a cyclic peptide with higher
endothelin receptor binding activity than BQ-123. While
no improved BQ-123 analogs were unveiled in the
process, the rediscovery of the Banyu endothelin an-
tagonist compound among the 82 944 cyclic peptides of
the “self-deconvoluting” library confirms that this strat-
egy might prove useful for the rapid identification of
suitable targets for new receptor classes where no prior
constrained leads are available.
F igu r e 2. Dose-activity response curves where inhibitory
activity of authentic cyclo(Pro-D-Val-Leu-D-Trp-D-Asp) (BQ-
123) is compared to four synthetic analogs with purification
protocols comparable to library components.
on a Dionex D-300 amino acid analyzer and a Dionex CP-3
programmer at the Medical Center of the University of
Louisville after the peptide mixtures were hydrolyzed in 6 N
HCl at 110 °C for 20-24 h in sealed, evacuated hydrolysis
tubes. Molecular weight determinations were made by ES-
MS at the University of Michigan. HPLC was performed on
a Hitachi 655A-22 instrument with a Vydac C18 column using
a linear gradient of (A) water containing 0.05% TFA and (B)
acetonitrile containing 0.05% TFA, at a flow rate of 1 mL/min
with UV detection at 254 nm unless otherwise specified. UV
spectra were recorded on a Hewlett Packard 8452A diode array
spectrometer.
Exp er im en ta l Section
In str u m en ts a n d Ma ter ia ls. 1H-NMR spectra (300 MHz,
CDCl3) were recorded on a Varian Inova-300 spectrometer.
TLC were run on Merck silica gel 60 F254 plates. Melting
points were determined on a Unimelt Thomas Hoover capillary
melting point apparatus and are uncorrected. Specific optical
rotations were determined on a J asco Model 700 instrument
at the sodium D line. Amino acid analyses were performed

Notes
Nr-(ter t-Bu t yloxyca r b on yl)-D-Asp a r t ic Acid â-F lu o-
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 19 3845
filtered, and washed again with DCM. The Boc group was
cleaved with TFA and the resin neutralized as described above.
Next the condensing agents were preactivated as follows: BOP
(3.31 g, 7.48 mmol), HOBt (1.012 g, 7.48 mmol), and DIEA
(1.3 mL, 7.48 mmol) in 24 mL of DMF. The mixture was
stirred for 1 min and distributed to the 24 syringes, and the
synthesis block was shaken for 2 h. The cyclization step was
repeated two more times to obtain a negative Kaiser test in
four randomly chosen syringes. The resin was washed with
DCM and dried in vacuo overnight.
Cyclic P ep tid e Libr a r y Clea va ge fr om Resin Usin g
An h yd r ou s H yd r ogen F lu or id e. The cleavages were
achieved in a 24-chamber multiple HF apparatus. L1-L24
and L25-L48 were cleaved separately. For each syringe, the
resin was placed in a cleavage vessel with 200 µL of anisole
and 5 mL of anhydrous HF. The mixture was stirred at 0 °C
for 1 h before removal of HF. For each reaction vessel, the
resin was transferred to a fritted glass funnel, and the peptide
products were extracted with 50 mL of 30% acetic acid each.
The acetic acid solution was washed twice with 25 mL of
diethyl ether and lyophilized to obtain salty flakes. The
peptide libraries were obtained after solid phase extraction
on reversed phase Varian Bond Elut C18; the remaining salts
were washed out with water, and the peptides were extracted
with 50 mL of acetonitrile-water (1:1) and lyophilized to afford
white or off-white powders. Yields ranged from 21 to 35 mg
for the individual syringe reactions (theory ) 47.6 mg based
on an average MW of 610 for BQ-123 or 49.6 mg if calculated
on the number-average MW of 636). The amino acid composi-
tions of L10, L17, L30, and L45 were randomly checked by
amino acid analysis.
Cyclo(P r o-D-P r o-Leu -D-Tr p -D-Asp ) (P 1), Cyclo(P r o-D-
P r o-Tr p -D-Tr p -D-Asp ) (P 2), Cyclo(P r o-D-Va l-Leu -D-Tr p -D-
Asp ) (P 3), a n d Cyclo(P r o-D-Va l-Tr p -D-Tr p -D-Asp ) (P 4).
The synthesis, cleavage, and purification of these four com-
pounds were performed on 150 mg of resin for each peptide in
a fashion identical with that reported for the libraries. HPLC
analysis of the crude cyclic peptide products following solid
phase extraction varied 45-60%, and no further purification
of these products was attempted prior to bioassay. The
compounds were characterized using high-resolution electro-
spray mass spectroscopy and HPLC. Compound P 3, corre-
sponding to authentic BQ-123, was cochromatographed against
authentic BQ-123 (kindly provided by Dr. K. Darlak).
For purposes of more complete characterization, each of the
four crude samples was further purified as follows. Reversed
phase HPLC purifications of samples of P 1-P 4 were per-
formed using a linear gradient of (A) water containing 0.05%
TFA and (B) acetonitrile containing 0.05% TFA (30-90% B,
30 min) at a flow rate of 1 mL/min on an analytical Vydac
C18 reversed phase column (4 mm × 250 mm) with UV
monitoring at 254 nm. Samples were submitted for mass
spectral analysis and amino acid analysis and provided the
following results.
P 1: Asp 1.2 (1), Pro 2.3 (2), Leu 0.7 (1), Trp 0.8 (1); ES-MS
(positive mode) M + Na+ ) 631.7; ES-MS (negative mode)
607.3 (calcd for C31H40N6O7 608.28).
P 2: Asp 1.0 (1), Pro 2.2 (2), Trp ND (2); ES-MS (positive
mode) M + Na+ ) 706.6; ES-MS (negative mode) 680.5 (calcd
for C36H39N7O7 681.28).
P 3: Asp 1.0 (1), Pro 1.1 (1), Leu 0.8 (1), Trp ND (1), Val 1.0
(1); ES-MS (positive mode) M + Na ) 632.3; ES-MS (negative
mode) 609.5 (calcd for C31H42N6O7 610.30).
P 4: Asp 1.0 (1), Pro 1.1 (1), Trp ND (2), Val 0.9 (1); ES-MS
(positive mode) M + Na ) 706.6; ES-MS (negative mode) 682.4
(calcd for C36H41N7O7 683.30).
P 1: TLC Rf (nBuOH-AcOH-H2O ) 8:1:1) 0.51; HPLC tR
17.46 min (linear gradient 5-90% B, 30 min), 8.21 min (linear
gradient 30-90% B, 30 min). P 2: TLC Rf (nBuOH-AcOH-
H2O ) 8:1:1) 0.49; HPLC tR 18.20 min (linear gradient 5-90%
B, 30 min), 9.47 min (linear gradient 30-90% B, 30 min). P 3:
TLC Rf (nBuOH-AcOH-H2O ) 8:1:1) 0.87; HPLC tR 19.36
min (linear gradient 5-90% B, 30 min), 10.36 min (linear
gradient 30-90% B, 30 min). P 4: TLC Rf (nBuOH-AcOH-
H2O ) 8:1:1) 0.88; HPLC tR 19.95 min (linear gradient 5-90%
B, 30 min), 11.27 min (linear gradient 30-90% B, 30 min).
r en ylm eth yl Ester (1).29 DCC (12.4 g, 60.1 mmol) was added
portionwise to a stirred solution of NR-(tert-butyloxycarbonyl)-
D-aspartic acid30 (14 g, 60 mmol) in 100 mL of ethyl acetate at
0 °C. The reaction mixture was kept at 0 °C for 1 h and at
room temperature for 1 h. The solution was filtered and the
solvent evaporated to dryness to obtain NR-(tert-butyloxy-
carbonyl)-D-aspartic anhydride. The latter was dissolved in
100 mL of dry THF with 12.4 g (63.2 mmol) of 9-fluorenyl-
methanol. Next 11.5 mL (66 mmol) of DIEA was added, and
the solution was stirred overnight. The solvent was evapo-
rated and the remaining solid partitioned between AcOEt and
a 5% citric acid solution. The aqueous solution was acidified
to pH 4 with 4 N HCl. The organic phase was washed with a
5% citric acid solution and then water, dried (MgSO4), and
concentrated until the appearance of the first crystals. The
product 1 was crystallized from ethyl acetate/hexane, filtered,
washed with hexane, and dried in vacuo: yield 13.4 g (54%);
TLC Rf (CHCl3-EtOH-AcOH ) 90:8:2) 0.63; mp 156°C; [R]20
D
+15.2° (c 1.00, THF); 1H-NMR (300 MHz, CDCl3) δ 1.50 (s,
9H, Me3C), 2.95 (ddd, 2H, CH2COOH), 4.24 (t, 1H,
COOCH2CH), 4.50 (multiplet, 2H, COOCH2CH), 4.68 (t, 1H,
NHCH), 5.55 (d, 1H, NH), 7.30-7.79 (cluster, 8H, aryl). Calcd
for C23H25NO6, MW ) 411.47; C, H, N.
Boc-Asp (O-ben zyl p oly(styr en e) r esin )-OF m -Acyla ted
Resin (2). Hydroxymethyl resin (9.0 g, 0.8 mequiv/g) (Pen-
insula; 1% cross-linked) was swollen in 75 mL of distilled
pyridine. Next 4.7 mL (30 mmol) of diisopropylcarbodiimide
was added to a solution of 12.0 g (29.1 mmol) of 1 and 4.0 g
(29.6 mmol) of HOBt in 100 mL of dry THF. The mixture was
stirred for 20 min before addition to the resin. After 6 h of
reaction, the resin was washed with 3 × 40 mL of methanol
followed by 3 × 40 mL of DCM. The resin was capped by
stirring for 90 min in 100 mL of DCM containing 5.5 mL (58.2
mmol) of acetic anhydride and 4.5 mL (55.7 mmol) of pyridine.
The resin was then filtered, washed as described previously,
and dried in vacuo: yield 11.2 g; substitution yield 0.52 mmol/
g.
Gen er a l P r oced u r e for th e Syn th esis of Libr a r ies L1-
L48 Cyclo(Xxx-Xxx-Xxx-Xxx-D-Asp ), w h er e Xxx ) D,L-P r o,
D,L-Va l, D,L-Leu , D,L-Tr p , D,L-Ar g, a n d D,L-Glu . The libraries
were synthesized in two separate batches (L1-L24 followed
by L25-L48) on 150 mg (0.078 mmol of aspartic acid) of resin
2 in each reactor.
P ep tid e ch a in elon ga tion : The resin was washed with
DCM (3 × 2 min × 1 mL/syringe) and shaken with 50% TFA
in DCM (5 min × 1 mL/syringe). The TFA treatment was
repeated for an additional 25 min; the resin was washed with
DCM (5 min × 1 mL/syringe), neutralized with DIEA-DCM
(1:9) (2 min × 1 mL/syringe), and washed again with DCM.
For each coupling, the protected amino acids were individually
weighed, combined, and divided into individual syringes
according to the amino acid ratios given in the Supporting
Information. For each syringe, the proportionate values would
be Boc-L-Trp (7.97 mg, 0.026 mmol), Boc-L-Leu (5.68 mg, 0.023
mmol), Boc-L-Glu(OBzl) (9.97 mg, 0.030 mmol), Boc-L-Arg(Tos)
(16.4 mg, 0.038 mmol), Boc-L-Val (4.34 mg, 0.020 mmol), Boc-
L-Pro (4.30 mg, 0.020 mmol), Boc-D-Trp (7.97 mg, 0.026 mmol),
Boc-D-Leu (5.68 mg, 0.023 mmol), Boc-D-Glu(OBzl) (9.97 mg,
0.030 mmol), Boc-D-Arg(Tos) (16.4 mg, 0.038 mmol), Boc-D-
Val (4.34 mg, 0..020 mmol), and Boc-D-Pro (4.30 mg, 0.020
mmol). Coupling reagent amounts used were BOP (138.1 mg,
0.312 mmol) and HOBt (42.2 mg, 0.312 mmol), dissolved with
the amino acids in 1 mL of DMF/syringe. Each milliliter of
amino acid solution also contained 54.5 µL of DIEA (0.312
mmol). The mixture was stirred for 1 min prior to transfer to
individual syringes. The coupling of single amino acid was
achieved with a 4-fold excess of amino acid, BOP (138.06 mg,
0.312 mmol), HOBt (42.15 mg, 0.312 mmol), and DIEA 54.5
µL (0.312 mmol) in 1 mL of DMF. The block was shaken for
1 h, and the completion of the couplings was monitored by
testing of four randomly chosen syringes using the Kaiser
ninhydrin test.
Cyclic P ep tid e Libr a r y Cycliza tion s Usin g BOP Re-
a gen t. The resin in each syringe was washed with DCM,
shaken for 20 min with 1 mL of 20% piperidine in DMF,

3846 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 19
Notes
(11) Rovero, P.; Quartara, L.; Fabbri, G. Synthesis of cyclic peptides
on solid support. Tetrahedron Lett. 1991, 32, 2639-2642.
(12) Kates, S. A.; Sole´, N. A.; J ohnson, C.; Hudson, D.; Barany, G.;
Albericio, F. A novel, convenient, three phase synthesis of cyclic
peptides. Tetrahedron Lett. 1993, 34, 1549-1552.
(13) Kates, S. A.; Sole´, N. A.; Albericio, F.; Barany, G. Solid phase
synthesis of cyclic peptides. In Peptides: Design, Synthesis and
Biological Activity; Basava, C., Anantharamaiah, G. M., Eds.;
Birkha¨user: Boston, 1994; pp 39-58.
(14) Darlak, K.; Romanovskis, P.; Spatola, A. Cyclic peptide libraries.
In Peptides: Chemistry, Structure, and Biology; Hodges, R.,
Smith, J ., Eds.; ESCOM: Leiden, 1994; pp 981-983.
(15) Spatola, A. F.; Darlak, K.; Romanovskis, P. An approach to cyclic
peptide libraries: reducing epimerization in medium size during
solid phase synthesis. Tetrahedron Lett. 1996, 37, 591-594.
(16) Green, J .; Bradley, K. Studies on the acylation of hydroxy-
functionalized resins using Fmoc amino acids activated using
diisopropylcarbodiimide/HOBt or as acid fluorides. Tetrahedron
1993, 49, 4141-4146.
Ra d ioliga n d Bin d in g Assa y. CHO cells stably expressing
human endothelin-A receptor cDNA were cultured in mono-
layers on 24-well plates as previously described.31,32 Confluent
monolayers of cells were preincubated with cyclic peptide
analogs at 37 °C for 10 min in a binding buffer (DMEM/0.3%
bovine serum albumin) containing 2% DMSO. Then [125I-
Tyr13]ET-1 was added to each well at 20 pM, and the cells
were incubated at 37 °C for an additional 1 h. Cells were
washed four times with ice-cold binding buffer and then lysed
with 0.1 N NaOH. Cell-bound radioactivity was determined
using a γ-counter. Nonspecific binding was defined in the
presence of 1 µM unlabeled ET-1 and was approximately 1%
of the levels of the specific binding in the absence of competitor
peptides. Assays were carried out in triplicate. Synthetic
cyclic peptide libraries were solubilized in a small volume of
DMSO and then diluted with binding buffer to achieve a final
concentration of total peptides of approximately 0.8 mM in
assay wells.
(17) Krchnak, V.; Vagner, J . Color monitored solid-phase multiple
peptide synthesis under low-pressure continuous-flow conditions.
Pept. Res. 1990, 3, 182.
Ack n ow led gm en t. This work was supported by
NIH-GM33376 and by grants from Eli Lilly and Torrey
Pines Institute for Molecular Studies. Mass spectral
determinations were provided by Drs. Philip Andrews
and Rachel Loo at the Protein and Carbohydrate
Structure Facility, University of Michigan, and amino
acid analyses were performed by Ben Van Osdal at the
University of Louisville. M.Y. is an Associate Investiga-
tor of the Howard Hughes Medical Institute.
(18) Fields, G.; Tian, Z.; Barany, G. In Synthetic Peptides: A User’s
Guide; Grant, G., Ed.; W. H. Freeman and Co.: New York, 1992;
Chapter 3.
(19) Gairi, M.; Lloyd-Williams, P.; Albericio, F.; Giralt, E. Use of BOP
reagent for the suppression of diketopiperazine formation in Boc/
Bzl solid-phase peptide synthesis. Tetrahedron Lett. 1990, 31,
7363-7366.
(20) Castro, B.; Dormoy, J .; Evin, G.; Selve, C. Reactif de couplage
peptidique IV (1)-L’hexafluorophosphate de benzotriazolyl N-
oxytrisdimethylamino phosphonium (B.O.P.). (Peptide coupling
reagents. IV. N-[Oxytris(dimethylamino)phosphonium]benzo-
triazole hexafluorophosphate.) Tetrahedron Lett. 1975, 1219-
1222.
Su p p or tin g In for m a tion Ava ila ble: Tables S1 and S2
containing the ratio of amino acids used in mixtures and amino
acid analysis results for six randomly selected samples of cyclic
peptide mixtures (2 pages). Ordering information is given on
any current masthead page.
(21) Kaiser, E.; Colescott, R. L.; Bossinger, C. D.; Cook, P. I. Color
test for detection of free terminal amino groups in the solid-phase
synthesis of peptides. Anal. Biochem. 1970, 34, 595-598.
(22) Hellberg, S.; Sjo¨stro¨m, M.; Skagerberg, B.; Wold, S. Peptide
quantitative structure-activity relationship: a multivariate ap-
proach. J . Med. Chem. 1987, 30, 1126-1135.
(23) Hansch, C. In Drug Design; Ariens, E., Ed.; Academic Press:
New York and London, 1971; Chapter 2.
Refer en ces
(1) Gallop, M. A.; Barrett, R. W.; Dower, W. J .; Fodor, S. P. A.;
Gordon, E. M. Applications of combinatorial technologies to drug
discovery. 1. Background and peptide combinatorial libraries.
J . Med. Chem. 1994, 37, 1233-1251.
(2) Gordon, E. M.; Barrett, R.; Dower, W. J .; Fodor, S. P. A.; Gallop,
M. A. Applications of combinatorial technologies to drug discov-
ery. 2. Combinatorial organic synthesis, library screening
strategies, and future directions. J . Med. Chem. 1994, 37, 1385-
1401.
(3) Terrett, N. K.; Gardner, M.; Gordon, D. W.; Kobylecki, R. J .;
Steele, J . Combinatorial synthesis - The design of compound
libraries and their application to drug discovery. Tetrahedron
1995, 51, 8135-8173.
(4) Thompson, L. A.; Ellman, J . A. Synthesis and applications of
small molecule libraries. Chem. Rev. 1996, 96, 555-600.
(5) DeWitt, S.; Kiely, J .; Stankovic, C.; Schroeder, M.; Cody, D.;
Pavia, M. “Diversomers”: an approach to nonpeptide, nonoligo-
meric chemical diversity. Proc. Natl. Acad. Sci. U.S.A. 1993,
90, 6909-6913.
(24) Ihara, M.; Noguchi, K.; Saeki, T.; Fukuroda, T.; Tsuchida, S.;
Kimura, S.; Fukami, T.; Ishikawa, K.; Nishikibe, M.; Yano, M.
Biological profiles of highly potent novel endothelin antagonists
selective for ETA receptor. Life Sci. 1992, 50, 247-255.
(25) Geysen, H. M. U.S. Patent 5 194 392, 1993.
(26) Houghten, R.; Pinilla, C.; Blondelle, S.; Appel, J .; Dooley, C.;
Cuervo, J . Generation and use of synthetic peptide combinatorial
libraries for basic research and drug discovery. Nature (London)
1991, 354, 84-86.
(27) Dooley, C. T.; Houghten, R. A. The use of positional scanning
synthetic peptide combinatorial libraries for the rapid determi-
nation of opioid receptor ligands. Life Sci. 1993, 52, 1509-1517
and references therein.
(28) Houghten, R.; Bray, M.; Degraw, S.; Kirby, C. Simplified
procedure for carrying out simultaneous multiple hydrogen
fluoride cleavages of protected peptide resins. Int. J . Pept.
Protein Res. 1986, 27, 673-768.
(29) Al-Obeidi, F.; Sanderson, D.; Hruby, V. Synthesis of â- and
γ-fluorenylmethyl esters of respectively NR-Boc-L-aspartic acid
and NR-Boc-L-glutamic acid. Int. J . Pept. Protein Res. 1990, 35,
215-218.
(30) Stewart, J .; Young, J . Solid Phase Peptide Synthesis; Pierce
Chemical Co.: Rockford, 1984; pp 61-62.
(31) Xu, D.; Emoto, N.; Giaid, A.; Slaughter, C.; Kaw, S.; deWit, D.;
Yanagisawa, M. ECE-1: a membrane-bound metalloprotease
that catalyzes the proteolytic activation of big endothelin-1. Cell
1994, 78, 473-485.
(32) Sakurai, T.; Yanagisawa, M.; Takuwa, Y.; Miyazaki, H.; Kimura,
S.; Goto, K.; Masaki, T. Cloning of a cDNA encoding a non-
isopeptide-selective subtype of the endothelin receptor. Nature
1990, 348, 732-735.
(6) Bunin, B.; Plunkett, M.; Ellmann, J . The combinatorial synthesis
and chemical and biological evaluation of 1,4-benzodiazepine
library. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 4708-4712.
(7) Cho, C.; Moran, E.; Stephans, J.; Fodor, S.; Adams, C.; Sundaram,
A.; J acobs, J .; Schultz, P. An unnatural biopolymer. Science
1993, 261, 1303-1305.
(8) Ladner, R. C. Constrained peptides as binding entities. Tibtech
1995, 13, 426-430.
(9) Olivera, B.; Rivier, J .; Clark, C.; Ramilo, C.; Corpuz, G.;
Abogadie, F.; Mena, E.; Woodward, S.; Hillyard, D.; Cruz, L.
Diversity of conus neuropeptides. Science 1990, 249, 257-263.
(10) O’Neil, K. T.; Hoess, R. H.; J ackson, S. A.; Ramachandran, N.;
Mousa, S. A.; DeGrado, W. F. Identification of GP-IIB/IIIA
Binding Peptides Using a Conformationally Constrained Phage
Peptide Library. Proteins 1992, 14, 509-515.
J M9604078