Difficult macrocyclizations: New strategies for synthesizing highly strained cyclic tetrapeptides

Article abstract of DOI:10.1021/ol034907o

(Matrix presented) Cyclic tetrapeptides are an intriguing class of natural products. To synthesize highly strained cyclic tetrapeptides we developed a macrocyclization strategy that involves the inclusion of 2-hydroxy-6-nitrobenzyl (HnB) group at the N-te

Full text of DOI:10.1021/ol034907o

ORGANIC
LETTERS
2003
Vol. 5, No. 15
2711-2714
Difficult Macrocyclizations: New
Strategies for Synthesizing Highly
Strained Cyclic Tetrapeptides
Wim D. F. Meutermans,† Gregory T. Bourne,†,‡ Simon W. Golding,†
Douglas A. Horton,† Marc R. Campitelli,† David Craik,† Martin Scanlon,§ and
,†,‡
Mark L. Smythe*
Institute for Molecular Bioscience, The UniVersity of Queensland, St. Lucia,
QLD 4072, Australia, Protagonist Pty. Ltd., LeVel 1, 11 Lang Parade, Milton,
QLD 4064, Australia, and Victorian College of Pharmacy, Monash UniVersity,
381 Royal, Parade, ParkVille, VIC 3052, Australia
m.smythe@imb.uq.edu.au
Received May 22, 2003
ABSTRACT
Cyclic tetrapeptides are an intriguing class of natural products. To synthesize highly strained cyclic tetrapeptides we developed a macrocyclization
strategy that involves the inclusion of 2-hydroxy-6-nitrobenzyl (HnB) group at the N-terminus and in the “middle” of the sequence. The N-terminal
auxiliary performs a ring closure/ring contraction role, and the backbone auxiliary promotes cis amide bonds to facilitate the otherwise difficult
ring contraction. Following this route, the all-L cyclic tetrapeptide cyclo-[Tyr-Arg-Phe-Ala] was successfully prepared.
Historically natural products have played an important role
in drug development, as either chemical tools to validate
targets or as a rich source of drug or lead compounds.1 Cyclic
tetrapeptides2-11 are a class of natural products that have been
characterized as potent and highly selective molecules in a
diverse range of therapeutic areas (i.e., a privileged struc-
ture).12 Typical examples are the cytotoxic and antimitogenic
agents HC toxin,2,3 chlamydocin,2-4 the antitumor agent
(6) Singh, S. B.; Zink, D. L.; Polishook, J. D.; Dombrowski, A. W.;
Darkin-Rattray, S. J.; Schmatz, D. M.; Goetz, M. A. Tetrahedron Lett. 1996,
37, 8077-8080.
(7) Darkin-Rattray, S. J.; Gurnett, A. M.; Myers, R. W.; Dulski, P. M.;
Crumley, T. M.; Allocco, J. J.; Cannova, C.; Meinke, P. T.; Colletti, S. L.;
Bednarek, M. A.; Singh, S. B.; Goetz, M. A.; Dombrowski, A. W.;
Polishook, J. D.; Schmatz, D. M. Proc. Natl. Acad. Sci. U.S.A. 1996, 93,
13143-13147.
(8) (a) Rich, D. H. J. Am. Chem. Soc. 1978, 100, 2218-2224. (b) Rich,
D. H.; Bhatnagar, P. K.; Jasensky, R. D.; Steele, J. A.; Uchytil, T. F.; Durbin,
R. D. Bioorg. Chem. 1978, 7, 207-214.
(9) (a) Kopple, K. D.; Wang, Y. S. Int. J. Pept. Protein Res. 1989, 33,
82-85. (b) Rich, D. H.; Gardner, J. H. Tetrahedron Lett. 1983, 24, 5305-
5308. (c) Rich, D. H.; Bhatnagar, P.; Mathiaparanam, P. J. Org. Chem.
1978, 43, 296-302. (d) Rich, D. H.; Jasensky, R. D. J. Am. Chem. Soc.
1980, 102, 1112-1119.
† Institute for Molecular Bioscience.
‡ Protagonist Pty. Ltd.
§ Victorian College of Pharmacy.
(1) For example, see: (a) Cragg, G. M.; Newmann, D. J.; Snader, K.
M. J. Nat. Prod. 1997, 60, 52-60. (b) Shu, Y.-Z. J. Nat. Prod. 1998, 61,
1053-1071. (c) Newman, D. J.; Cragg, G. M.; Snader, K. M. Nat. Prod.
Rep. 2000, 17, 215-234.
(2) (a) Shute, R. E.; Kawai, M.; Rich, D. H. Tetrahedron 1988, 44, 685-
695. (b) Kawai, M.; Jasensky, R. D.; Rich, D. H. J. Am. Chem. Soc. 1983,
105, 4456-4462. (c) Liesch, J. M.; Sweeley, C. C.; Staffeld, G. D.;
Anderson, M. S.; Weber, D. J.; Scheffer, R. P. Tetrahedron 1982, 38, 45-
48.
(3) Bernardi, E.; Cros, S.; Viallefont, P.; Lazaro, R. Bull. Soc. Chim.
Fr. 1994, 131, 944-948.
(10) Jois, D. S. S.; Suresh, S.; Vijayan, M.; Easwaran, K. R. K. Int. J.
Pept. Protein Res. 1996, 48, 12-20. (b) Rich, D. H.; Bhatnagar, P. K. J.
Am. Chem. Soc. 1978, 100, 2212-2218. (c) Shimizu, T.; Tanaka, Y.; Tsuda,
K. Int. J. Pept. Protein Res. 1986, 27, 554-561.
(4) Schmidt, U.; Lieberkecht, A.; Griesser, H.; Roland, U.; Beuttler, T.;
Bartkowiak, F. Synthesis 1986, 361-366.
(5) Kijima, M.; Yoshida, M.; Sugita, K.; Horinouchi, S.; Beppu, T. J.
Biol. Chem. 1993, 268, 22429-22435.
(11) Rich, D. H. Tam, J. P. Tetrahedron Lett. 1977, 749-750.
10.1021/ol034907o CCC: $25.00 © 2003 American Chemical Society
Published on Web 07/01/2003

Scheme 1. Synthesis of Difficult Sequences through a Ring Contraction Strategya
a Reagents and conditions: (i) 1 equiv of BOP, 1 mM in DMF, 2 equiv of DIEA, 3 h/rt. (ii) 10 equiv of DIEA, 20 h/rt. (iii) hυ, 1% HOAc
in DMSO, 1:1, 2 h.
trapoxin,3,5 the tyrosinase inhibitor cyclo-[(L)Pro-(L)Tyr-(L)-
Pro-(L)Val], and the antimalarial apicidins.6 One compound
is reported to show in vivo activity by both parenteral and
oral administration in mice.7
monocyclic target, even when cyclization is performed at
high dilution or the sequence is rotated for optimal cycliza-
tion yield.14 The problem becomes more prominent for
shorter medium-sized ring peptides.14
Cyclic tetrapeptides are very rigid 12-membered-ring
structures. Different conformers of these molecules have been
isolated, and these have been shown to display differing
biological activities.8 Although isolation and structure de-
termination of these compounds has taken place over the
past three decades, there has been little success in synthesiz-
ing representative compounds. This is surprising, given that
cyclic tetrapeptides may be considered a rich source of drug
like molecules due to their wide-ranging biological activities,
low molecular weight, favorable pharmacokinetic charac-
teristics, and unique 12-atom cyclic backbone, providing a
rigid framework that can support a wide range of functional
groups.
The ring strain inherent in this framework makes the
synthesis of these molecules very difficult.13 The primary
reason for ineffective cyclization originates from a sequence-
related inefficiency to bring the termini together for cycliza-
tion.14 Because peptide bonds contain strong π-character and
preferentially adopt a trans conformation, linear peptides
prefer a more extended conformation. Incorporation of turn-
inducing elements such as Gly, Pro, or D-amino acids are
known to enhance cyclization yields.15 As a result, the few
cyclic tetrapeptides that have been synthesized contain either
D-residues or at least one tertiary amide in the sequence.2-11
Small head-to-tail cyclic peptides that do not contain turn-
inducing elements are known to be very difficult to
synthesize,14-16 as macrocyclization of the linear sequences
produce linear and cyclic oligomers in preference to the
In a previous report, the development of a novel ring-
contraction auxiliary that could be used to facilitate the room
temperature cyclization of difficult cyclic pentapeptide
sequences was described.17 It was anticipated that this
approach could also be applied to yield 12-membered cyclic
tetrapeptides. Cyclization of a 2-hydroxy-6-nitrobenzyl (HnB)
N-terminally substituted tetrapeptide 2 would initially gener-
ate a more accessible but reactive cyclic nitrophenylester
intermediate 3 (Scheme 1), which would ring contract
through an O-to-N acyl transfer to generate the desired,
substituted, target compound 4. Photolytic removal of the
HnB auxiliary would provide the target cyclic product 5.
When this strategy was applied to the tetrapeptide [HnB]-
Tyr-Arg-Phe-Gly-OH 2a none of the desired cyclic product
4a was obtained when cyclization was conducted at room
temperature. Fortunately, cyclization to 4a could be achieved
through formation of the nitrophenyl ester intermediate 3a,
followed by further addition of DIEA and heating at 70 °C
overnight. This cyclic product is not accessible from the
unsubstituted H-Tyr-Arg-Phe-Gly-OH 1 sequence. The extra
thermal energy must be required to overcome various ring
strain elements in the formation of the 12-membered-ring
products, despite the higher effective concentration of the
C- and N-termini and the entropic advantages when using
HnB for cyclization. This strategy has subsequently proven
to be useful in the synthesis of a variety of glycine-containing
cyclic tetrapeptides. However, glycine is a CR-unsubstituted
amino acid, and its presence in a cyclic peptide should greatly
(12) Horton, D. A.; Bourne, G. T.; Smythe, M. L. Chem. ReV. 2003,
103, 893-930.
(16) (a) Ehrlich, A.; Heyne, H. U.; Winter, R.; Beyermann, M.; Haber,
H.; Carpino, L. A.; Bienert, M. J. Org. Chem. 1996, 61, 8831-8838. (b)
Bourne, G. T.; Meutermans, W. D. F.; Alewood, P. F.; McGeary, R. P.;
Scanlon, M.; Watson, A. A.; Smythe, M. L. J. Org. Chem. 1999, 64, 3095-
3101.
(17) (a) Meutermans, W. D.; Golding, S. W.; Bourne, G. T.; Miranda,
L. P.; Dooley, M. J.; Alewood, P. F.; Smythe, M. L. J. Am. Chem. Soc.
1999, 121, 9790-9796. (b) Miranda, L. P.; Meutermans, W. D. F.; Smythe,
M. L.; Alewood, P. F. J. Org. Chem. 2000, 65, 5460-5468.
(13) Cavelier-Frontin, F.; Achmad, S.; Verducci, J.; Jacquier, R.; Pepe,
G. THEOCHEM 1993, 105, 125-130.
(14) (a) Schmidt, U.; Langner, J. J. Pept. Res. 1997, 49, 67-73. (b)
Pastuszak, J.; Gardner, J. H..; Singh, J.; Rich, D. H. J. Org. Chem. 1982,
47, 2982-2987. (c) Horton, D. A.; Bourne, G. T.; Smythe, M. L. J. Comput.-
Aided Mol. Des. 2002, 16, 415-430.
(15) Kessler, H.; Haase, B. Int. J. Pept. Protein Res. 1992, 39, 36-40.
2712
Org. Lett., Vol. 5, No. 15, 2003

Table 1. Cyclization of Linear Tetrapeptides 8 and Corresponding Yields of Cyclic Product 9 and Photolysis to Give Product 5
cyclic product (9)b
entry
linear peptide (8)a
% yield by HPLCc
% isolated materialc
% desired isomerd
photolysise (5)
a
b
c
d
[HnB]-Tyr-Arg-[HnB]-Phe-Gly-OH
[HnB]-Tyr-Arg-[HnB]-Phe-Ala-OH
[HnB]-Tyr-Arg-[HnB]-Phe-D-Ala-OH
[HnB]-Tyr(Bn)-Arg-[HnB]-Phe-Ala-OH
75
78
68
71
33
50%
59
51
48
26
22
23
0 (20%)f
42%
25%h
a All naturally occurring amino acids are the L-isomer unless otherwise stated. b Cyclization methods: (i) 1 equiv of BOP or HATU, 2 equiv of DIEA,
1 mM in DMSO, 3 h/rt; (ii) 10 equiv of DIEA, 20 h/rt. c Combined yield of L- and D-product. d % isolated yield of the correct isomer. e Photolysis method:
hυ, CH3CN/H2O or 1% HOAc in DMSO. f Photolysis conducted upon purified all-L MnB-isomer 10b. h Photolysis conducted upon crude all-L MnB-isomer
10d.
facilitate cyclization as a result of its flexibility. More
constrained cyclic peptides that do not contain glycine are
therefore far more challenging synthetic targets. As a result,
the next target was cyclo-[Tyr-Arg-Phe-Ala]. Unfortunately,
cyclization of the HnB-substituted linear peptide 2b at 70
°C overnight yielded none of the desired cyclic tetrapeptide
4b, even when the temperature, activating conditions, and
time were varied.
These results suggest that for non-glycine-containing all-L
tetrapeptides a significant barrier for ring contraction exists.
This barrier could be lowered by increasing the propensity
for cis-amide bonds in the linear precursor, an idea especially
attractive given that crystal and NMR structures of cyclic
tetrapeptides have at least one cis amide bond.10 As N-
alkylation of amide bonds have been previously reported to
lower the trans-cis amide bond barrier,11,16 this general
strategy was a promising route for cyclic tetrapeptide
synthesis.
approach in the synthesis of such difficult targets. On the
basis of this successful result, this combination approach was
applied to the synthesis of cyclo-[Tyr-Arg-Phe-Ala] 5b,
which had previously proven to be synthetically inaccessible.
Because of the increased ring strain, as a result of four CR-
alkyl amino acids, ring contraction was expected to proceed
at a slower rate than for glycine-containing analogues.
Therefore cyclization of [HnB]-Tyr-Arg-[HnB]-Phe-Ala-
OH 8b was examined at room temperature for 20 h, at 70
°C for 1 h, and at 70 °C for 20 h. Surprisingly, reasonable
yields of cyclic products (59%) were generated at room
temperature (Table 1). Chiral amino acid analysis, MS, and
NMR were used to characterize the isolated reaction
Scheme 2. Synthesis of Difficult Sequences via Combination
Strategya
It was reasoned that HnB should be an ideal backbone
substituent to aid cyclization as it is a bulky N-alkyl
substituent (increasing the propensity for cis amide bonds16),
the secondary amine substituted with HnB can be readily
acylated17 (even with bulky amino acids), and the substituent
is removed by photolysis. To this end, H-Tyr-Arg-[HnB]-
Phe-Gly-OH 6 was synthesized, which contained HnB in the
“middle” of the sequence. Cyclization of 6 yielded the
desired cyclic product cyclo-[Tyr-Arg-[HnB]-Phe-Gly] 7 in
15% isolated yield and significant dimer (10% isolated yield).
Previous observations indicated that an advantage of
including HnB at the N-terminus, followed by cyclization,
is the significant reduction of dimer formation.16,17
A
combination of HnB at the N-terminus (to increase the
effective concentration of the C- and N-termini, entropically
favoring cyclization and reducing oligomerization) and a
HnB substituent at the N2 position (to promote cis amide
conformations and facilitate ring contraction) should therefore
provide a suitable strategy for the synthesis of highly strained
macrocycles (Scheme 2). This strategy proved to be suc-
cessful. Synthesis and cyclization of the bis-HnB substituted
tetrapeptide [HnB]-Tyr-Arg-[HnB]-Phe-Gly-OH 8a yielded
the desired HnB-substitued cyclic product in 33% isolated
yield at room temperature, with no dimer formation (Table
1). This room temperature ring contraction confirmed the
previous hypothesis on the value of the HnB combination
a Reagents and conditions: (i) 1 equiv of BOP or HATU, 1 mM
in DMSO, 2 equiv of DIEA, 3 h/rt then 10 equiv of DIEA, 20 h/
rt. (ii) hυ, CH3CN/H2O or 1% HOAc in DMSO, 1:1, 2 h. (iii)
CH2N2 EtOH, 30 min. (iv) H2/Pd, MeOH, 1 h.
Org. Lett., Vol. 5, No. 15, 2003
2713

products. The constrained cyclo-([HnB]-Tyr-Arg-[HnB]-Phe-
Ala) 9b product was isolated in 26% yield.
generated the desired all-L-cyclic tetrapeptide cyclo-[Tyr-
Arg-Phe-Ala] 5b in 20% yield (after HPLC purification),
confirming our proposed hydrolysis pathway.
The only competing side reaction during cyclization was
racemization of Ala to yield cyclo-([HnB]-Tyr-Arg-[HnB]-
Phe-D-Ala) in 33% yield, which was not significantly
different when using other activating agents or bases.
Cyclization of both the bis-HnB L-Ala 8b and D-Ala 8c linear
peptides yielded similar ratios of cyclic L- and D-products,
suggesting that racemization is significantly faster than
cyclization and that cyclization rates for L-Ala 8b and D-Ala
8c are similar (Table 1).
Previously, it was demonstrated that the HnB group could
be removed from the amide backbone by photolysis.17
Photolysis of the glycine-containing cyclic tetrapeptides 4a
and 9a proceeded in reasonable yield and purity (Scheme
1), and products were isolated and characterized by MS,
chiral amino acid analysis, and NMR.
In contrast, whereas photolysis of the D-Ala cyclic product
9c yielded 5c cleanly, the more strained all-L-cyclic tetra-
peptide 9b yielded only hydrolyzed product. It may be
expected that the amide bonds of the highly strained all-L
analogues to be out of plane and more susceptible to
hydrolysis. This may occur via a ring-expansion reaction.
(Scheme 3).
To avoid the need for selective methylation a benzyl-
protected tyrosine was subsequently employed. Thus the all-L
monocyclic peptide 9d was isolated in 25% yield (the D-Ala
analogue was also isolated in 23% yield) from 8d. A 30 min
diazomethane treatment of 9d generated the bis-MnB-
substituted analogue 10d. Following photolysis and purifica-
tion, cyclo-[Tyr(Bn)-Arg-Phe-Ala] 5d was isolated in 25%
overall yield. Hydrogenolysis generated the target cyclo-[Tyr-
Arg-Phe-Ala] 5b.
In conclusion, despite three decades of discovery of
biologically active cyclic tetrapeptides and their reported oral
activity, they have been largely unexplored in the pharma-
ceutical industry. Primarily this is due to the difficulties in
preparing these rigid 12-membered-ring compounds. This
paper reports the development of a synthetic strategy that
uses a combination of two reversible auxiliaries, one at the
N-terminus and one at the middle backbone amide nitrogen
atom. The first auxiliary performs a ring closure/ring
contraction role, which increases the effective concentration
of the C- and N-termini, entropically favoring cyclization
and effectively preventing oligomerization from occurring,
and the latter serves as a cis-amide bond promotor to facilitate
the otherwise difficult ring contraction. HnB is suitable for
both auxiliary functions owing to its superior acyl transfer
properties, its ready accessibility, and its photolability.
Scheme 3. Proposed Reaction Pathway during Photolysis of
9b
Acknowledgment. We would like to acknowledge Glaxo-
SmithKline, UK; GlaxoSmithKline, Australia; and the Aus-
tralian Research Council (ARC) for their financial support.
Dr. B. Reid, Ms. T. Bond, Mr. A. Jones (ARC Special
Research Centre for Functional and Applied Genomics), Mr.
P. Seale, Dr. John Murray, and Dr. Mike Hann are also
acknowledged.
Alkylation of the 2-hydroxy functionality on HnB should
prevent these decomposition reactions. This was attempted
by treatment with excess diazomethane in ethanol, but
methylation of the HnB groups was slower than expected
and partial methylation of the tyrosine side chain was
inevitable. A 10 min diazomethane treatment provided the
best yield of the desired bis(2-methoxy-6-nitrobenzyl) (bis-
MnB) peptide cyclo-[[MnB]-Tyr-Arg-[MnB]-Phe-Ala] 10b
(10% yield after HPLC purification). Photolysis of 10b
Supporting Information Available: Full experimental
procedures including TOCSY data for cyclo-[Tyr-Arg-Phe-
Gly] showing the amide NH region. This material is available
free of charge via the Internet at http://pubs.acs.org.
OL034907O
2714
Org. Lett., Vol. 5, No. 15, 2003