An efficient synthesis of cyclic RGD peptides as antithrombotic agents

Full text of DOI:10.1021/jo9603284

5180
J . Org. Chem. 1996, 61, 5180-5185
synthesis and purification.10 This paper reports our
An Efficien t Syn th esis of Cyclic RGD
successful large scale synthesis of cyclic peptide 1.
The original 14-step synthesis, similar to Scheme 1,
was employed to allow quick preparations of many
structurally similar cyclic peptides. In this procedure,
the linear peptide 7 was prepared on a polymeric oxime
support then cleaved from the resin by an intramolecular
aminolysis. The guanidino group of arginine was pro-
tected as the tosylate, and the carboxyl group of aspartic
acid was protected as the cyclohexyl ester.8 Both protect-
ing groups were removed by anhydrous HF, and the
product was purified by reverse-phase HPLC in 11%
overall yield.
Scheme 1 represents our attempt to adapt the same
14-step reaction sequence in solution phase by substitut-
ing the 4-nitrobenzophenone oxime ester in place of the
solid support. In this preparation, the cyclic peptide 8
must be prepared under dilute conditions to avoid
intermolecular coupling of the linear peptide 7. The tosyl
and cyclohexyl protecting groups of 8 were removed by a
combination of trifluoroacetic acid (TFA) and triflic acid,
over 3-5 h, under carefully controlled temperature
constraints (-6 to -8 °C). At higher temperatures (>
-6 °C) ring-opened side products were formed and at
lower temperatures (<-8 °C) the reaction was too slow
to be practical. Furthermore, a large volume of anhy-
drous diethyl ether was used to precipitate the product
from the acidic media. This synthesis provided sufficient
material for safety assessment and early clinical studies;
however, it was neither economical nor practical for large
scale synthesis.
P ep tid es a s An tith r om botic Agen ts
Lin-hua Zhang,* J . A. Pesti, T. D. Costello,
P. J . Sheeran, R. Uyeda, P. Ma, G. S. Kauffman,
R. Ward, and J . L. McMillan
The DuPont Merck Pharmaceutical Company,
Deepwater, New J ersey 08023
Received February 16, 1996
The use of glycoprotein (GP) IIb/IIIa antagonists as
antithrombotic agents has been an active area of research
in recent years.1-7 When linear peptides containing the
Arg-Gly-Asp (RGD) sequence bind to the IIb/IIIa receptor,
their dissociation constants are about 10 mM. It has
been shown that the conformation of the peptides has
an important impact on the binding affinities. For
instance, the binding was considerably improved when
cyclic RGD peptides, such as SK&F 106760 (2), were
studied as IIb/IIIa antagonists.7,8 We found that the
cyclic peptide linked with d-2-aminobutyric acid (D-Abu)
and m-(aminomethyl)benzoic acid (Mamb) had an IC50
value of 20 nM. After an extensive study of structure-
activity relationships, cyclic (D)-Abu-N-MeArg-Gly-Asp-
Mamb (1), was selected as a development candidate.
Compound 1 has a dissociation constant in the picomolar
range when binding to GP IIb/IIIa and exhibits potent
antiplatelet activities both in vitro and in vivo.8
Since target 1 is a cyclic pentapeptide, retrosynthetic
analysis indicated that any of the five peptide bonds
provided the opportunity to form the cycle. We knew that
successfully closing the cycle onto the secondary R-nitro-
gen of the arginine was unlikely. We also preferred a
convergent synthesis that minimized the number of
deprotection steps. Finally, we wanted to be able to use
as many commercially available peptide derivatives as
possible. The combination of these considerations di-
rected us to the retrosynthetic analysis shown in Scheme
2. Models indicated that ring closure between the
C-terminus of glycine and the N-terminus of aspartic acid
was a good choice because of minimal steric hindrance
and favorable conformation. Furthermore, after coupling
of intermediates 10 and 11, the orthogonality of the
remaining four protecting groups permits unmasking of
the cyclization sites in a single step.
The tripeptide 11 was produced in four steps without
isolation of intermediates. The coupling of commercially
available N-Boc-N-Me-Ng-p-tosyl-L-arginine (5) with gly-
cine benzyl ester (12) was performed with HBTU and
diisopropylethylamine (DIEA) as shown in Scheme 3.
Using ethyl acetate, an atypical choice of solvent, pro-
duced a homogenous reaction and an easy aqueous
workup.11 The unreacted glycine was removed by aque-
ous citric acid washes. The organic portion was then
concentrated followed by the addition of TFA to remove
the Boc group and produce 13.
In support of a clinical study of this antithrombotic
agent, an effective large scale preparation of 1 was
required. However, there was little known about prepa-
rations of cyclic peptides on larger than a laboratory
scale.9 The presence of arginine and aspartic acid in the
peptide was expected to increase the difficulties of
(1) Philips, D. R.; Charo, I. F.; Parise, L. V.; Fitzgerald, L. A. Blood
1988, 71, 831.
(2) Gould, R. J .; Polokoff, M. A.; Friedman, P. A.; Huang, T.-F.; Holt,
J . C.; Cook, J . J .; Niewiarowski, S. Proc. Soc. Exp. Biol. Med. 1990,
195, 168.
(3) Coller, B. S.; Peerschke, E. I.; Scudder, L. E.; Sullivan, C. A. A.
J . Clin. Invest. 1983, 72, 325.
(4) Pidard, D.; Montgomery, R. R.; Bennett, J . S.; Kunicki, T. J . J .
Biol. Chem. 1983, 258, 12582.
(5) Hung, T.-F.; Holt, J . C.; Lukasiewicz, H.; Niewiarowski, S. J .
Biol. Chem. 1987, 262, 16157.
(6) Nutt, R. F.; Brady, S. F.; Sisko, J . T.; Ciccarone, T. M.; Colton,
C. D.; Levy, M. R.; Gould, R. J .; Zhang, G.; Friedman, P. A.; Veber, D.
A. Proc. Eur. Pept. Symp. 21st 1990, 784.
(7) Semanen, J .; Ali, F.; Romoff, T.; Calvo, R.; Sorenson, E.; Vasko,
J .; Storer, B.; Berry, D.; Bennett, D.; Strohsacker, M.; Powers, D.;
Stadel, J .; Nichols, A. J . Med. Chem. 1991, 34, 3114.
(8) (a) Bach, A. C.; Eyermann, C. J .; Gross, J . D.; Bower, M. J .;
Harlow, R. L.; Weber, P. C.; Degrado, W. F. J . Am. Chem. Soc. 1994,
116, 3207. (b) J ackson, S. A.; Degrado W.; Dwivedi, A.; Parthasarathy,
A.; Higley, A.; Krywko, J .; Rockwell, A.; Markwalder, J .; Wells, G.;
Wexler, R.; Mousa, S. A.; Harlow, R. J . Am. Chem. Soc. 1994, 116,
3220.
(10) Bodanszky, M.; Bodanszky, A. The Practice of Peptide Synthesis;
Springer-Verlag, 1984. (b) Barker, P. L.; Bullens, S.; Bunting, S.;
Burdick, D.; Chan, K. S.; Deisher, T.; Eigenbrot, C.; Gadek, T. R.;
Gantzos, R.; Lipari, M. T.; Muir, C. D.; Napier, M. A.; Pitti, R. M.;
Padua, A.; Quan, C.; Stanley, M.; Struble, M.; Tom, J . Y. K.; Burnier,
J . J . Med. Chem. 1992, 35, 2040.
(9) (a) Bodanszky, M. Principles of Peptide Synthesis, 2nd ed.;
Springer-Verlag, 1993. (b) Zhang, L. H.; Ma, P. Tetrahedron Lett. 1994,
35 (32), 8387.
(11) THF or methylene chloride are typically used solvents.
S0022-3263(96)00328-3 CCC: $12.00 © 1996 American Chemical Society

Notes
J . Org. Chem., Vol. 61, No. 15, 1996 5181
Sch em e 1
The physical characteristics of 13 were not favorable,
The remaining portion (10) of the cyclopeptide was
prepared by either the reaction of N-(carbobenzyloxy)-
tert-butyl-L-aspartic acid 15 with 1,1-carbonyldiimidazole
followed by the addition of m-(aminomethyl)benzoic acid
(16), or by the reaction of N-(carbobenzyloxy)-tert-butyl-
L-aspartic N-hydroxysuccinimide ester with the sodium
salt of m-(aminomethyl)benzoic acid (Scheme 4). The
yield was higher (82% versus 73%) for the latter method
and the use of N-hydroxysuccinimide esters proved to be
most suitable for our large scale preparations. Further-
more, the major byproduct from the reaction, N-hydroxy-
succinimide, was water soluble and easily removed. The
peptide 10 obtained by this approach had a purity of
>96%.
Coupling of intermediate 11 with 10 in the presence
of HBTU and DIEA in acetonitrile provided the penta-
peptide 17. Because this linear peptide was oily and
difficult to handle, it was best to replace ethyl acetate
via the methanol azeotrope and proceed directly to the
following deprotection step without isolation of 17.
preventing its useful isolation. The product tended to
be gummy and was very hygroscopic.12 After concentra-
tion and workup, the crude product solution was used
directly in the next step. Peptide coupling on a substi-
tuted nitrogen is known to be difficult,9 but HBTU
functioned well to add N-Boc-d-amino butyric acid (D-
Abu) to the arginine. Accordingly, an ethyl acetate
solution of the free acid of D-Abu was added to the
solution containing 13 and dried via toluene azeotrope
at reduced pressure. A small amount of acetonitrile was
added to homogenize the solution, HBTU/DIEA were
charged, and coupling was complete in about 4 h. At a
kilogram scale, tripeptide 14 was deprotected and pre-
cipitated as the HCl salt of 11 using anhydrous HCl in
dioxane.13 The tripeptide prepared by this procedure was
greater than 97% pure by HPLC analysis. We later
learned that a similar deprotection/precipitation protocol
could be used to isolate dipeptide 13, as long as all solvent
was replaced by dioxane.
At this point, four protecting groups remain on the
pentapeptide. Catalytic hydrogenation at 1 atm at 25
°C over wetted 10% palladium/carbon simultaneously
removed both protecting groups from the ends of the
linear pentapeptide to form 18. The reduction proceeded
in 2 h and was nonselective in rate between the two
protecting groups. The product was precipitated as a
(12) Isolation of a dry solid was only possible by addition of an ethyl
acetate solution of 13 to a large excess of tert-butyl methyl ether at
low temperature (-15 to 20 °C).
(13) Initially, the HCl salt precipitation was unsuccessful in either
dioxane or toluene due to the residual DCHA, because the free base of
D-Abu had not been used. Once this was recognized and rectified, the
HCl salt was easily produced as long as both water and ethyl acetate
were removed via initial azeotropic distillation with dioxane.

5182 J . Org. Chem., Vol. 61, No. 15, 1996
Notes
Sch em e 2
Sch em e 3a
a
white solid by the addition of the filtered reaction solution
into 1-butanol followed by the distillation of methanol
and water.
(a) HBTU, DIEA. (b) TFA. (c) HBTU, DIEA. (d) HCl, 82% yield
from 5.
peptides.15 Adding back sufficient methanol to break the
gel and restarting the distillation always eventually
resulted in solids. On a preparative scale, the yield of
19 was 58% from crude intermediate 11. No further
purification was required to use the cyclopeptide in the
subsequent tandem deprotection step.
Typically, removal of tosyl protection from guanidine
requires relatively forcing conditions.16 There are a
number of examples reported for the detosylation of
arginine- and arginine-containing peptides.17,18 However,
these methods did not produce satisfactory results when
applied to deprotection of 19. In this case, we discovered
that trifluoroacetic acid/triflic acid/anisole in combination
was a good detosylation system. The tert-butyl group of
aspartic acid was first removed by neat trifluoroacetic
acid to give intermediate 20. Subsequent addition of
triflic acid and anisole cleaved the tosyl group. In
contrast to the slow detosylation of the N-tosyl, C-
cyclohexyl congener 8, the tosyl group of cyclic peptide
20 was removed within 1 h between -20 to -6 °C. This
Once again, cyclization of 18 made use of the coupling
agent HBTU/DIEA.14 Dicyclohexylcarbodiimide (DCC)
in acetonitrile could be used but the yields were half that
using HBTU. Remarkably, in spite of not using high
dilution conditions, excellent yields of the cyclopenta-
peptide were obtained. The product was contaminated
with trace quantities of impurities, tentatively character-
ized as oligomers by 1H NMR spectroscopy. To minimize
the putative oligomer formation, the solution of 18 and
DIEA in acetonitrile was added to the solution of HBTU
in acetonitrile over 5-6 h. However, even when the
addition was shortened to 1 hour, minor, if any, ad-
ditional impurities formed.
The cyclized product 19 precipitated as it formed,
leaving in solution most of the impurities from the
previous several steps. If insufficient base was initially
charged to complete the reaction, a second charge com-
pleted the cyclization with no effect upon yield or purity.
The only purification necessary was to wash the filtered
solids with acetonitrile/DIEA.
The cyclic peptide 19 was sparingly soluble in most
solvents. We could purify 19 by slow reprecipitation from
acetonitrile/dimethylacetamide solution but found the
following process to be more efficient. Pure (>98%) 19
was obtained by passing a methanolic chloroform solution
through a pad of silica gel. Solids were precipitated by
replacement of the solvent with acetonitrile via vacuum
distillation. Occasionally, a gel would begin to form
rather than dry solids, a known propensity of certain
(15) Hanadusa, K.; Natsumoto, Y.; Miki, T.; Koyama, T.; Shirai, H.
J . Chem. Soc., Chem. Commun. 1994, 1401.
(16) Bodanszky, M. Principles of Peptide Synthesis; Reactivity and
Structure Concepts in Organic Chemistry, Hafner, K., et al., Eds.;
Springler-Verlag, Berling, 1984; Vol. 16, pp 137-141.
(17) (a) Fujii, N.; Otaka, A.; Ikemura, O.; Akaji, K.; Funakoshi, S.;
Hayashi, Y.; Kuroda, Y.; Yajima, H. J . Chem. Soc., Chem. Commun.
1987, 274. (b) Petrakis, K. S.; Kaiser, E. T.; Osapay, G. J . Chem. Soc.,
Chem. Commun. 1991, 530.
(18) (a) Yoshino, H.; Kaneko, T.; Arakawa, Y.; Nakazawa, T.;
Yamatsu, K.; Tachibana, S. Chem. Pharm. Bull. 1990, 38, 404. (b)
Huguenin, R. L.; Boissonnas, R. A. Helv. Chim. Acta 1962, 91, 1629.
(c) Nishino, N.; Xu, M.; Mihara, H.; Fujimoto, T. Tetrahedron Lett.
1992, 33, 1479.
(14) Dourloglov, V.; Gross, B. Synthesis 1984, 22, 572.

Notes
J . Org. Chem., Vol. 61, No. 15, 1996 5183
Sch em e 4a
acidic medium degraded the cyclic peptide. Accordingly,
1 was precipitated out of solution by butyl ether dilution.
However the particles were slow to filter, and as these
solids warmed on the filter, the residual acids led to the
formation of impurities. The use of large filter beds
minimized this, but ultimately this problem was resolved
by controlling the particle size via the stirring speed and
the addition rate of butyl ether.
The crude solid was dissolved into aqueous acetone19
and treated with an anion exchange resin until the pH
reached 4.55-4.95. This neutralization was very efficient
for the removal of any residual trifluoroacetic acid and
trifluoromethanesulfonic acid.20 Following evaporation
of the solvent, recrystallization from warm aqueous
acetone or simply stirring the solids in water would
usually reduce the remaining impurities to acceptable
levels. The isolated product 1 often contained ap-
preciable water (up to 20%) which did not affect the
subsequent step. On a large scale, the yield of the
tandem deprotection was 79%.
The preparation of the methanesulfonic acid salt of 1
was done by the addition of 1 equiv of acid to a
suspension of 1 in refluxing 2-propanol (IPA). Usually,
no change in the appearance of the slurry was seen
during the acid addition as the solids transformed from
zwitterion to the salt. After cooling, filtration and an IPA
wash of the white solids, clean drug was obtained in
nearly quantitative yield by simply drying the powder
on a Buchner funnel under nitrogen. The product
prepared by this synthetic route has >99% optical purity
based on chiral HPLC analysis and >98.5% weight
percent purity exclusive of solvents.
In summary, we have developed an efficient route to
cyclic peptide 1, a potent antithrombotic agent. This is
a nine-step synthesis from commercially available start-
ing materials, involving no chromatographic purification.
The economy of deprotection, efficiency of cyclization and
the enhanced detosylation method provided an attractive
route to manufacture cyclic peptide 1 in bulk.
Exp er im en ta l Section
Gen er a l. All melting points are uncorrected. The 1H and
13C NMR spectra were recorded at 300 and 75.4 MHz, respec-
tively, and are reported in DMSO-d6 (except where noted). The
coupling constants (J ) of NMR spectra were reported in hertz.
Mass spectra were recorded by the analytical laboratory, Chemi-
cal Process R&D, DuPont Merck Research Laboratories. The
elemental analyses were determined by Quantitative Technolo-
gies Inc., Whitehouse, NJ . Solvents and reagents were used as
purchased without further purification. The protected amino
acids were purchased from Bachem Bioscience Inc.
a
(a) N-Hydroxysuccinimide ester of 15, NaOH, 82%. (b) HBTU,
DIEA. (c) H2, Pd/C, >99%. (d) HBTU, DIEA, 58% from 11.
Sch em e 5a
HCl Sa lt of Boc-N-Me-Ar g(Tos)-Gly-OBzl (13). Ethyl
acetate (3.2 L), N-Boc-N-Me-Ng-p-tosyl-L-arginine (5) (500 g, 1.13
mol), glycine benzyl ester (12) toluenesulfonate salt (450 g, 1.33
mol), and HBTU (500 g, 1.32 mol) were combined under nitrogen
and stirred at room temperature. To this was added diiso-
propylethylamine (480 mL) at <30 °C. The reaction was stirred
at room temperature for 6 h. Solvent (2 L) was removed in vacuo
at <30 °C, and trifluoroacetic acid (1.7 L) was added at <40 °C.
The reaction was stirred for 6 h, and the solution was concen-
trated in vacuo until 1.1 L of solvent had been removed. This
crude 13 in solution was used directly in the subsequent step.
The spectral data of 13 were identical to those reported in the
literature.8
a
(a) TFA. (b) Triflic acid, anisole, 79% from 19.
(19) Occasionally it was necessary to dissolve 1 by the addition of a
few drops of acetic acid to a water slurry prior to resin treatment.
(20) The extent of exchange could be determined by measuring the
19F NMR spectrum of a portion of evaporated solution versus an
internal standard of trifluoroethanol. In instances of incomplete
removal, further treatment with more resin would remove the last
traces.
rate difference suggests that the free carboxylic acid
function might facilitate the cleavage of the tosyl group
by neighboring group participation (Scheme 5).
A variety of workup procedures were surveyed in order
to determine conditions to isolate 1 before the strongly

5184 J . Org. Chem., Vol. 61, No. 15, 1996
Notes
Boc-D-Abu -N-MeAr g(Tos)-Gly-OBzl (14). Ethyl acetate (2
L) and water (2 L) were added to the above reaction mixture,
and its pH was adjusted to 7.6 with 25% NaOH. Additional
water (4 L) and ethyl acetate (2 L) were added and the layers
separated. The aqueous phase was extracted with ethyl acetate
(2 × 4 L), and the combined organic layer was washed with
saturated aqueous NaHCO3 (2 × 3 L). The organic phase was
concentrated in vacuo until 8.8 L of distillate was removed.
In a separate flask, Boc-d-aminobutyric acid dicyclohexyl-
amine salt (551 g, 1.43 mol) was suspended in ethyl acetate (2.5
L) and 10% aqueous KHSO4 (4 L). The suspension was stirred
until all the solids dissolved (∼40 min). The phases were
separated, and the aqueous phase extracted with ethyl acetate
(2 × 1 L). The combined ethyl acetate layers were washed with
water (3 × 1 L) until the aqueous portion no longer gave a
precipitate with 2% BaCl2 solution.
This amine-free solution of Boc-d-aminobutyric acid was
combined with the solution containing dipeptide 13 and toluene
(4 L). The solution was concentrated in vacuo until 6 L had
been removed (water content of solution ) 0.05% at this point).
Acetonitrile (1 L), HBTU (501 g, 1.32 mol), and DIEA (480 mL)
were added successively at <30 °C. The reaction was stirred at
room temperature overnight, the solution was concentrated in
vacuo until 2.5 L of solvent had been removed, and the solution
was diluted with ethyl acetate (2.5 L). The solution was washed
with 5% citric acid solution (3 × 1.7 L) followed by saturated
aqueous NaCHO3 (2 × 800 mL). The solution was dried by
azeotropic distillation with dioxane until 6.7 L had distilled and
5.5 L of dioxane had been added (water content ) 0.16% at this
point). This solution containing 14 was used directly in the next
step. The spectral data of 14 were identical to those reported
in the literature.8
for C24H28N2O7: C, 63.14; H 6.18; N 6.13. Found C, 62.90; H,
6.31; N, 6.09.
Z-Asp (OBu t)-Ma m b-D-Abu -N-MeAr g(Tos)-Gly-OBzl(17).
(Carbobenzyloxy)aspartic
acid(tert-butyl)-m-(aminomethyl)-
benzoic acid 10 (305 g, 0.654 mol) and D-R-aminobutyric acid-
N-methylarginine(tosyl)-glycine benzyl ester hydrochloride salt
11 (400 g, 0.654 mol) were mixed with HBTU (273 g, 0.720 mol)
in acetonitrile (2.4 L) and chilled to 4 °C. DIEA (490 mL, 2.81
mol) was added slowly over 90 m at <9 °C. The mixture was
stirred for 16 h at 25 °C. Ethyl acetate (4.8 L) was added, and
the organic layer was washed sequentially with 5% aqueous
citric acid, saturated aqueous NaHCO3, and brine. The solvent
was azeotropically replaced with methanol and used as is in the
next step (>99% yield). An analytical sample of 17 was purified
by chromatography. 1H NMR δ 0.97 (m, 1H), 1.20 (m, 1H), 1.30
(m, 2H), 1.42 (s, 5H), 1.66 (m, 1H), 1.80 (m, 3H), 2.05 (m, 1H),
2.41 (s, 1H), 2.77 (s, 1H), 2.81 (s, 1H), 2.99 (s, 1H), 3.12 (s, 1H),
3.39 (s, 1H), 3.99 (m, 1H), 4.40 (d, J ) 6.0 Hz, 1H), 4.50 (m,
1H), 4.85 (m, 1H), 5.15 (m, 3H), 7.40 (m, 8H), 7.70 (m, 1H), 7.82
(m, 1H), 8.23 (m, 1H), 8.60 (m, 1H). MS (CI) m/z 1013 (M +
H)+. Anal. Calcd for C51H64N8O12S (1013.2): C, 60.46; H, 6.37;
N, 11.06. Found: C, 60.86; H, 6.31; N, 10.64.
Asp (OBu t)-Ma m b-D-Abu -N-MeAr g(Tos)-Gly (18). A slurry
of 17 (658g, 0.65 mol) in methanol (5.0 L) and 10% palladium
on carbon (316 g of 50% aqueous mixture, 0.15 mol) was
degassed, and hydrogen gas was sparged subsurface at atmo-
spheric pressure for 2 h. The catalyst was filtered off over Celite.
The filtrate was stripped to 1.2 L of viscous oil and added
dropwise to 1-butanol (7.5 L) at 30 °C under 7-20 mm Hg of
vacuum. The resulting slurry was cooled to 5 °C, filtered, and
dried in vacuo to yield 721 g of 18/1-butanol complex (∼100%
yield). An analytical sample was produced by flash chromatog-
raphy. 1H NMR (DMSO-d6) δ 8.93 (br s, 1H), 8.63 (d, 1H), 7.85
(s, 1H), 7.71-7.81 (m, 2H), 7.63 (d, J ) 8.9, 3H), 7.32-7.50 (m,
3H), 7.21-7.32 (m, 3H), 5.01(dd, J ) 6.0, 9.0; 1H), 4.77 (q, J )
6.8, 1H), 4.33 (br s, 2H), 3.63 (dd, J ) 5.2, 7.4; 2H), 3.31-3.57
(m, 12H), 3.06 (br s, 3H), 2.91 (s, 3H), 2.60 (dd, J ) 3.0, 15.0;
2H), 2.34 (s, 3H), 2.29-2.47 (m, 2H), 1.47-1.91 (m, 6H), 1.38
(s, 9H), 1.05 (t, J ) 6.9, 2H), 1.04 (d, J ) 6.2, 1H), 0.90 (t, J )
7.7, 3H). HRMS (ESI) Calcd for C36H52N8O10S 789.3597 (M +
H)+, found 789.3605 (M + H)+.
D-Abu -N-MeAr g(Tos)-Gly-OBzl (11). The above solution of
14 was combined with 4.0 M HCl in dioxane (2 L) at <40 °C.
The solution was stirred over 12 h during which time the product
precipitated. Acetonitrile (3 L) was added and the mixture
cooled to <15 °C. The white solids were filtered under nitrogen.
The wet weight of the product was 860.5 g (66.1 weight percent
purity), corresponding to an overall yield over four steps of 82.4%
from 5. Preparation of an analytical standard of 11 was
accomplished by a slurry of the crude product in THF. The
product was collected by a filtration and dried: mp 120-121
Cyclo-D-Abu -N-MeAr g(Tos)-Gly-Asp (OBu t)-Ma m b (19).
A solution of 18 (393 g, ∼0.35 mol), DIEA (100 mL, 0.574 mol),
and acetonitrile (3.4 L) was added dropwise to a stirred solution
of HBTU (136 g, 0.359 mol) in acetonitrile (3.0 L) over 5.5 h.
The reaction mixture was stirred overnight. The mixture was
cooled to 0 °C over 2 h, and the solids were collected by filtration.
The solids were washed with a solution of acetonitrile/DIEA (400
mL/4 mL) and dried in vacuo. The crude 19 was dissolved into
a solution of 18% methanolic chloroform (1215 mL) at 34 °C.
Silica gel (527 g) was added to the solution. This slurry was
filtered, and the filter cake was further extracted with 12%
methanolic chloroform. The filtrate was concentrated in vacuo.
The solvent was replaced with acetonitrile until GC analysis of
the pot contents indicated the chloroform content was <0.5%.
The slurry was stirred at 34 °C overnight, and the purified 19
was collected by filtration. The solids were washed with
acetonitrile (200 mL) and dried under vacuum to 158.6 g (58%
overall yield from 11) of white solids: mp 200.6-202.3 °C; 1H
NMR δ 8.86 (br d, J ) 5.1; 1H, NH), 8.61-8.46 (m, 2H, NH),
7.76-7.22 (m, 8H), 7.02 (v br s, NH), 6.82 (v br s, NH), 6.55 (v
br s, NH), 5.11 (dd, J ) 4.6,10.8; 1H), 4.66 (dd, J ) 7.8,16.7;
1H), 4.60-4.48 (m, 2H), 4.19 (dd, J ) 8.4,16.5; 1H), 4.00 (dd, J
) 4.0,16.5; 1H), 3.68 (br s, NH), 3.65-3.54 (m, 1H), 3.14-2.96
(m, 2H), 2.89 (s, 3H), 2.68 (dd, J ) 7.1,15.9; 1H), 2.33 (s, 3H),
2.02-1.84 (m, 1H), 1.84-1.62 (m, 3H), 1.60-1.42 (m, 1H), 1.38
(s, 9H), 1.34-1.16 (m, 2H), 0.93 (t, J ) 7.3, 3H); 13C NMR (DMF-
d7 + TFA) δ 174.27, 171.42, 170.33, 170.00, 169.89, 168.85,
157.92, 143.16, 141.86, 140.55, 134.59, 130.31, 129.61, 128.27,
126.34, 125.96, 125.83, 80.98, 56.65, 53.01, 51.66, 41.94, 41.90,
40.98, 37.34, 30.51, 28.00, 26.52, 24.83, 24.61, 21.05, 10.52.
MS(ESI), m/z 771 (M + H)+. Anal. Calcd for C36H50N8O9S: C,
56.09; H, 6.54; N, 14.53; S, 4.16. Found: C, 56.10; H, 6.71; N,
14.45; S, 4.00.
1
°C; H ΝMR δ 0.93 (t, 3H, J ) 7.6), 1.9 (m, 2H), 1.6-2.0 (br m,
4H), 2.42 (s, 3H), 2.95 (s, 3H), 3.15 (br s, 2H), 3.87 (br m, 2H),
4.25 (br m, 1H), 5.05 (dd, J ) 4.0, 11.2, 1H), 5.37 (s, 2H), 6.75
(br s, 1H), 7.30 (d, J ) 7.8, 2H), 7.35 (m, 5H), 7.65 (d, J ) 7.8,
2H), 8.28 (br s, 3H), 8.59 (t, J ) 5.8, 1H). HRMS (M + H): Calcd
for C27H39N6O6S: 575.2652. Found: 575.2635.
Z-Asp(OBu t)-m-(Am in om eth yl)ben zoic Acid (Mam b) (10).
A slurry of N-(carbobenzyloxy)-â-tert-butyl aspartic succinate
(756.7 g, 1.8 mol) in acetonitrile (2.7 L) was prepared. To this
slurry was added over 1-2 h a solution of the sodium salt of
m-aminomethylbenzoic acid (327.3 g, 1.89 mol) at 20-25 °C.
(This solution of the sodium salt was made from the HCl salt
by dissolving m-(aminomethyl)benzoic acid hydrochloride (354.6
g, 1.89 mol) in water (2.4 L). The pH of this solution was
adjusted to 8.8 ( 0.2 with 30% NaOH at 20-30 °C.) Sodium
bicarbonate (151.2 g, 1.8 mol) was added in five portions, and
small amounts of insolubles were removed by filtration. The
reaction was stirred overnight and the pH of the solution was
adjusted to 3.4-3.7 with 60% aqueous citric acid. The slurry
was stirred at 20-25 °C for 2 h while maintaining this pH range.
The slurry was cooled to 5-10 °C, held for 1 h, and filtered, and
the wet cake washed successively with water (1.2 L) and a cold
solution of 1/1 volume of acetonitrile/water (1.2 L). The yield of
10 after drying in vacuo at 40-50 °C, was 681 g (82.8%).
Recrystallization of a sample from MeOH/water gave an analyti-
cal sample as a white solid: mp 149-151 °C dec; 1H NMR δ
1.34 (s, 9H), 2.45 (dd, J ) 16.1, 8.8; 1H), 2.68 (dd, J ) 16.1, 5.5;
1H), 4.31 (d, J ) 5.9; 2H), 4.44-4.34 (m, 1H), 5.07-4.99 (ABq,
J ) 12.4; 2H), 7.36-7.26 (m, 5H), 7.41 (t, J ) 7.7, 1H), 7.47 (dt,
J ) 7.7, 1.5; 1H), 7.60 (d, J ) 8.4, 1H), 7.80 (dt, J ) 7.7, 1.8;
1H), 7.84 (t, J ) 1.8, 1H), 8.54 (t, J ) 5.9, 1H), 12.91 (bs, 1H);
13C NMR (DMSO-d6, 75.4 MHz) 36.22, 46.29, 47.32, 47.59, 47.87,
48.15, 48.42, 48.70, 48.98, 50.61, 60.2, 74.12, 76.26, 86.67, 88.67,
128.42, 136.27, 136.33, 136.74, 136.85, 136.99, 139.36, 140.21,
145.48, 148.42, 164.32, 175.81, 177.87, 179.1; IR (nujol) 3307,
3248, 1713, 1660 cm-1; ms (CI) m/z 457.1 [M + H]+. Anal. Calcd
Cyclo-D-Abu -N-MeAr g-Gly-Asp -Ma m b (1). Trifluoroacetic
acid (1043 g, 9.15 mol) was added to ice cooled 19 (140.0 g, 0.182
mol) with stirring. The ice bath was removed and stirring
continued for 20 min after the solids dissolved. Trifluoro-

Notes
J . Org. Chem., Vol. 61, No. 15, 1996 5185
methanesulfonic acid (1196 g, 7.97 mol) was added to this
solution at -18 to -13 °C. The solution was cooled to -30 °C,
anisole (139 g, 1.3 mol) was added at -30 to -23 °C, and then
the reaction was warmed to -14 °C. The deep purple mixture
was held at -14 to -8 °C for 1 h and cooled to -30 °C. Butyl
ether (4150 mL) was added at -30 to -12 °C to produce bright
yellow solids. The slurry was cooled to -30 °C and held there
without stirring for 30 min. The solids were collected on
Whatman no. 24 filter paper using vacuum and under nitrogen.
The resulting solids were washed with additional butyl ether
and dried under a stream of nitrogen. The dry solids were
dissolved into 1:1 aqueous acetone (2.0 L). Residual butyl ether
as a top layer at this point was separated and extracted with
water (2 × 100 mL). These extracts were added to the aqueous
layer. Anion exchange resin (Biorad AG1 resin, acetate form;
1.03 kg) was added in portions until the pH entered the range
4.55-4.95. The slurry was stirred for an additional 40 min and
filtered. The resin was washed with 1:1 aqueous acetone (1.6
L) and the filtrate evaporated to 117.8 g of white solids (crude
yield 115%). These were slurried into water (950 mL), heated
to 50 °C, and diluted slowly with acetone (160 mL). Most of the
acetone was removed in vacuo at 45 °C, and 1 was allowed to
further recrystallize overnight with stirring. The next day the
slurry was cooled to 0 °C over 45 min, held 5 h, and collected by
filtration. The solids were washed with enough water to cover
the solids, allowed a 10-min soak, and dried under a nitrogen
stream. These solids were further dried to constant weight in
vacuo to produce 106.8 g of white solids (79% yield): mp 270-5
°C dec; 1H NMR (D2O/DCl with water repression, reference )
TSP) δ 8.89 (m, NH), 8.00 (br dd, J ) 2.4, 8.6; NH), 7.64 (br dt,
J ) 2.2, 6.2; 1H), 7.51-7.42 (m, 2H), 7.36 (br s, 1H), 5.38 (dd, J
) 4.4, 11.4; 1H), 4.82-4.65 (m, 3H), 4.59 (dt, J ) 4.4, 17.2; 1H),
4.31 (br d, J ) 16.8, 1H), 3.84 (br d, J ) 17.2, 1H), 3.23 (br t, J
) 6.6, 2H), 3.17 (s, 3H), 2.99 (dd, J ) 7.3,16.8, 1H), 2.84 (dd, J
) 7.3,16.8, 1H), 2.26-2.08 (m, 1H), 2.04-1.74 (m, 3H), 1.70-
1.42 (m, 2H), 1.02 (t, J ) 7.3, 3H); 13C NMR (CD3OD + TFA) δ
176.29, 173.49, 173.42, 172.08, 171.15, 171.03, 158.70, 140.44,
134.89, 131.26, 129.31, 126.52, 126.11, 57.71, 54.01, 52.36, 42.76,
42.48, 41.90, 36.05, 31.28, 26.49, 25.22, 25.09, 10.68. MS (ESI)
m/z 561 (M + H)+. Anal. Calcd for C25H36N8O7: C, 53.56; H,
6.47; N, 19.99. Found: C, 53.31; H, 6.41; N, 19.85.
Cyclo-D-Abu -N-MeAr g-Gly-Asp -Ma m b, MeSO3H Sa lt. A
solution of methanesulfonic acid (11.66g, 0.121 mol) in 2-pro-
panol (0.2 L) was added to a refluxing suspension of 1 (80.0 g,
0.121 moles, 15% water by weight) in 2-propanol (3 L) over 16
m. The slurry was cooled to 1 °C over 29 min and stirred for 1
h. The salt was collected, washed with cold 2-propanol (600 mL),
and dried to constant weight under a stream of dry nitrogen.
The final weight was 79.0 g (99%) of white solids: mp 260-1
1
°C; H NMR (400 MHz) δ 12.33 (br s, D2O exchanged), 8.85 (d,
J ) 5.5, 1H), 8.52 (d, J ) 7.5, 1H), 8.44 (dd, J ) 7.6, 4.9; 1H),
7.75-7.66 (m, 1H), 7.63 (dd, J ) 8.2, 3.1; 1H), 7.57 (br t, J )
5.6, 1H), 7.50 (br s, 1H), 7.40-7.32 (m, 2H), 5.15 (dd, J ) 4.6,
11.0; 1H), 4.68-4.54 (m, 2H), 4.17 (dd, J ) 16.7, 8.2; 1H), 4.05
(dd, J ) 16.5, 4.9; 1H), 3.68 (dd, J ) 16.7, 3.1; 1H), 3.31 (br s,
D2O exchanged), 3.18-3.04 (m, 2H), 2.95 (s, 3H), 2.71 (dd, J )
16.7, 6.6; 1H), 2.58-2.50 (m, 1H), 2.37 (s, 3H), 2.08-1.94 (m, 1H),
1.77 (dq, J ) 7.5, 7.5; 2H), 1.66-1.52 (m, 1H), 1.44-1.26 (m,
2H), 0.95 (t, J ) 7.5, 3H); 13C NMR (100.6 MHz) δ 10.17, 23.52,
23.69, 25.11, 30.20, 35.35, 39.66, 40.27, 40.80, 41.04, 50.14, 51.99,
55.64, 125.18, 125.24, 127.46, 129.65, 133.23, 139.51, 156.64,
167.99, 168.40, 169.21, 170.61, 171.44, 173.43. MS (FAB) m/z
561.3 (M + H)+. Anal. Calcd for C26H40N8O10S: C, 47.55; H,
6.14; N, 17.06; O, 24.36; S, 4.88. Found: C, 47.49; H, 6.27; N,
16.72; O, 24.51; S, 4.78.
Ack n ow led gm en t. We thank Mr. Thomas Scholz
and Dr. Gregory Nemeth for spectroscopy assistance.
Su p p or tin g In for m a tion Ava ila ble: Copies of 1H and 13C
NMR spectra of 1, 11, 10, 17, 18, and 19 (12 pages). This
material is contained in libraries on microfiche, immediately
follows this article in the microfilm version of the journal, and
can be ordered from the ACS; see any current masthead page
for ordering information.
J O9603284