Synthesis of difficult cyclic peptides by inclusion of a novel photolabile auxiliary in a ring contraction strategy

Article abstract of DOI:10.1021/ja992173y

Cyclic peptides comprise a large and important class of biologically active molecules. They are generally synthesized through amide bond-forming reactions of the C- and N- termini under high dilution conditions. Yields of such processes are highly dependent on the size of the ring being formed and on the particular amino acids of the linear precursor, giving rise to the well-known sequence-dependent effect of cyclization. To overcome this problem, we have developed a peptide cyclization strategy that proceeds through a ring closure/ring contraction process. The linear peptide Ala-Phe-Leu-Pro-Ala, which does not generate monocyclic product under conventional cyclization conditions, was used as a model to probe a range of auxiliaries. This has led to the development of a new photolabile peptide cyclization auxiliary. The 6-nitro-2-hydroxybenzyl group is readily and quantitatively introduced at the W-terminus via a reductive alkylation. Cyclization of the auxiliary-peptide initially proceeds through a cyclic nitrophenyl ester that preorganizes the peptide for lactamization. As the C- and N- termini are in close proximity, lactamization is achieved via an intramolecular O-N acyl transfer step to produce the N-substituted target cycle. The auxiliary is then removed by mild photolysis to produce the target cyclic peptide, cyclo-[Ala-Phe-Leu-Pro-Ala], in good yield. This strategy should find further useful applications in the assembly of libraries of small cyclic peptides.

Full text of DOI:10.1021/ja992173y

9790
J. Am. Chem. Soc. 1999, 121, 9790-9796
Synthesis of Difficult Cyclic Peptides by Inclusion of a Novel
Photolabile Auxiliary in a Ring Contraction Strategy
Wim D. F. Meutermans,* Simon W. Golding, Greg T. Bourne, Les P. Miranda,
Michael J. Dooley, Paul F. Alewood, and Mark L. Smythe*
Contribution from the Centre for Drug Design and DeVelopment, UniVersity of Queensland,
Brisbane, Queensland, 4072 Australia
ReceiVed June 24, 1999
Abstract: Cyclic peptides comprise a large and important class of biologically active molecules. They are
generally synthesized through amide bond-forming reactions of the C- and N- termini under high dilution
conditions. Yields of such processes are highly dependent on the size of the ring being formed and on the
particular amino acids of the linear precursor, giving rise to the well-known sequence-dependent effect of
cyclization. To overcome this problem, we have developed a peptide cyclization strategy that proceeds through
a ring closure/ring contraction process. The linear peptide Ala-Phe-Leu-Pro-Ala, which does not generate
monocyclic product under conventional cyclization conditions, was used as a model to probe a range of
auxiliaries. This has led to the development of a new photolabile peptide cyclization auxiliary. The 6-nitro-
2-hydroxybenzyl group is readily and quantitatively introduced at the N-terminus via a reductive alkylation.
Cyclization of the auxiliary-peptide initially proceeds through a cyclic nitrophenyl ester that preorganizes the
peptide for lactamization. As the C- and N- termini are in close proximity, lactamization is achieved via an
intramolecular O-N acyl transfer step to produce the N-substituted target cycle. The auxiliary is then removed
by mild photolysis to produce the target cyclic peptide, cyclo-[Ala-Phe-Leu-Pro-Ala], in good yield. This
strategy should find further useful applications in the assembly of libraries of small cyclic peptides.
Introduction
addition, cyclization generally promotes an increase of metabolic
stability and bioavailability of peptides.11,12
In the linear form, bioactive peptides can assume millions of
different conformations, very few of which are able to bind to
their receptor. To assess the important structural and dynamic
properties that are critical for the biological potency and
selectivity of peptides, conformational constraints are often
introduced, typically through cyclization.1-6 Such cyclic mol-
ecules may exist in more clearly defined conformations and are
appealing from a drug lead discovery perspective.7 If activity
is maintained or enhanced in these cyclic peptides, structural
information is sought, by NMR, X-ray or molecular modeling,
and used to guide the development of therapeutic drugs.8-10 In
As the side chains are generally considered to be the main
mediators for receptor interaction,1 cyclization is preferably
accomplished between the C- and N-termini. Whereas the
synthesis of linear peptides generally proceeds well, head-to-
tail cyclization is often troublesome, especially for small peptides
of less than seven residues in length.13 The primary reason for
ineffective cyclization originates from a sequence-related inef-
ficiency to bring the termini together for head-to-tail cyclization.
Since peptide bonds contain strong π-character and preferentially
adopt a trans conformation, linear peptides prefer more extended
conformations. This places the terminal carboxylic acid and
amine functional groups in remote positions that are unfavorable
for cyclization. Incorporation of turn-inducing elements such
as Gly, Pro, or D-amino acids are known to enhance cyclization
yields.14 For linear peptides of 4-6 residues that do not con-
tain amino acids that stabilize turn structures, slow cyclizations
lead to side reactions such as cyclodimerization and epimer-
ization.15
* Corresponding authors. Tel: (+61) 7 3365 1271. Fax: (+61) 7 3365
1990. E-mail: m.smythe@mailbox.uq.edu.au; w.meutermans@mailbox.
uq.edu.au.
(1) Hruby, V. J.; Al-Obeidi, F.; Kazmierski, W. Biochem. J. 1990, 268,
249.
(2) Marshall, G. Tetrahedron 1993, 49, 3547-3558.
(3) Jackson, S.; Degrado, 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.
(4) Morgan, B. P.; Bartlett, P. A.; Holland, D. R.; Matthews, B. W. J.
Am. Chem. Soc. 1994, 116, 3251-3260.
(5) Bitan, G.; Sukhotinsky, I.; Mashriki, Y.; Hanani, M.; Selinger, Z.;
Gilon, C. Int. J. Pept. Protein Res. 1997, 49, 421-426.
(6) Friedler, A.; Zakai, N.; Karni, O.; Broder, Y. C.; Baraz, L.; Kotler,
M.; Loyter, A.; Gilon, C. Biochemistry 1998, 37, 5616-5622.
(7) Fairlie, D. P.; Abbenante, G.; March, D. R. Curr. Med. Chem. 1995,
2, 654-686.
We are interested in developing chemical strategies that
facilitate the assembly of libraries of small cyclic peptides.16
To address the sequence-related inefficiency of small peptide
cyclizations, we have aimed at developing auxiliary strategies
(11) Milner-White, E. J. Trends Pharmacol. Sci. 1989, 10, 70-74.
(12) Verber, D. F.; Freidinger, R. M. Trends Neurosci. 1985, 8, 392-
396.
(8) Tian, Z.-P. A. Q.; Bartlett. J. Am. Chem. Soc. 1996, 118, 943-949.
(9) Pfaff, M.; Tangemann, K.; Muller, B.; Gurrath, M.; Muller, G.;
Kessler, H.; Timpl, R.; Engel, J. J. Biol. Chem. 1994, 269, 20233-20238.
(10) Burgle, M.; Koppitz, M.; Riemer, C.; Kessler, H.; Konig, B.; Weidle,
U. H.; Kellermann, J.; Lottspeich, F.; Graeff, H.; Schmitt, M.; Goretzki,
L.; Reuning, U.; Wilhelm, O.; Magdolen, V. Biol. Chem. 1997, 378, 231-
237.
(13) Ehrlich, A.; Heyne, H. U.; Winter, R.; Beyermann, M.; Haber, H.;
Carpino, L. A.; Bienert, M. J. Org. Chem. 1996, 61, 8831-8838.
(14) Kessler, H.; Haas, B. Int. J. Pept. Protein Res. 1993, 39, 36.
(15) Schmidt, U.; Langner, J. J. Pept. Res. 1997, 49, 67-73.
(16) Bourne, G.; Meutermans, W.; Alewood, P. F.; McGeary, R.;
Scanlon, M.; Watson, A.; Smythe, M. J. Org. Chem. 1999, 64, 3095-
3101.
10.1021/ja992173y CCC: $18.00 © 1999 American Chemical Society
Published on Web 10/08/1999

Synthesis of Difficult Cyclic Peptides
J. Am. Chem. Soc., Vol. 121, No. 42, 1999 9791
Scheme 1. A Peptide Cyclization Auxiliary Strategya
Scheme 2. Two Reaction Routes for the Introduction of the
Auxiliaries 1-3a
a The strategy involves (I) introduction of the auxiliary at the
N-terminus; (II) ring closure; (III) ring contraction; and (IV) removal
of the auxiliary. The auxiliaries 1, 2, and 3 were employed in this work.
a Route 1 via a Fukuyama approach using alcohol 5; route 2 via
reductive amination using the benzaldehydes 6 and 7.
Scheme 3. Cyclization of the N-Mercaptoethanyl Peptide 8a
that preorganize the N- and C-termini of a single chain for head-
to-tail cyclization. In one strategy (Scheme 1) an auxiliary (Aux)
carrying a hydroxy or thiol functionality (XH) is introduced at
the N-terminus of the peptide (Step I) and extends the se-
quence by several atoms. Activation of the C-terminus initially
results in the formation of a more accessible cyclic ester (Step
II), thereby bringing the N- and C-termini of the peptide closer
in space. An X-to-N acyl transfer (Step III) results in a ring
contraction to generate the target ring size, and removal of the
auxiliary (Step IV) produces the desired monocyclic product.
A suitable auxiliary must fulfil the following criteria: (i)
introduction of the auxiliary at the N-terminus of the peptide
should be facile and independent of the nature of the N-terminal
residue; (ii) to favor cyclization over oligomerization (step II),
the auxiliary must be sufficiently flexible to allow the N- and
C-termini, that is, the hydroxy/thiol functionality of the auxiliary
and the activated carbonyl at the C-terminus to meet (iii) as
there is no need to isolate the intermediate cyclic ester, rapid
ring contraction is preferred to avoid competitive ring opening
side reactions. Therefore the transition state geometry for this
acyl-transfer step should not be impeded by ring strain or steric
bulk caused by the extra atoms of the auxiliary. A more reactive
auxiliary-ester in the intermediate cyclic product should further
enhance the rate of the desired rearrangement but may increase
the rate of unwanted side reactions such as hydrolysis; (In-
tramolecular acyl transfers have been extensively studied by
Kemp et al.22,34-37 although not in a cyclization strategy. The
chief parameters that influence the rate of transfer are the ring
size of the transition state,34,37 and the correlated steric and
electronic effects of the auxiliary.35) (iv) removal of the auxiliary
should occur efficiently via orthogonal chemistry, not affecting
the peptide functionalities.
a (I) 3 equiv of BOP/5 equiv of DIEA, 3 h at room temperature; (II)
0.1 M NH4HCO3.
As a control experiment, we attempted to cyclize the unsub-
stituted linear peptide (Ala-Phe-Leu-Pro-Ala) using standard
cyclization conditions (1 mM in DMF [N,N-dimethylforma-
mide], 3 equiv of BOP [benzotriazol-1-yloxy-tris(dimethylami-
no)phosphonium hexafluorophosphate], 5 equiv of DIEA [di-
isopropylethylamine], 3 h at room temperature). As expected
from the previously reported results, only cyclic dimer and some
trimer were obtained but no target monocyclic product.
Evaluation of Ethanethiol Auxiliary 1. We initially evalu-
ated an ethanethiol auxiliary 1. This auxiliary was introduced
via an on-resin Fukuyama synthesis17-19 (Scheme 2, Route 1).
An O-nitrobenzene sulfonamide (ONBS), prepared from the
corresponding sulfonyl chloride and the primary amine 4, was
alkylated using S-(4-methylbenzyl)-2-thioethanol 5 under Mit-
sunobu-like conditions. The ONBS-group was then removed
using sodium thiophenoxide (PhSNa) in DMF, prior to HF
cleavage and deprotection of the thiol functionality.
Cyclization Using Ethanethiol Auxiliary 1. Whereas cy-
clization of the parent peptide only generates oligomeric
material, cyclization of the N-ethanethiol derivative 8 (Scheme
3) under the same conditions yielded only the monocyclic
product 11 (45% isolated yield), as determined by mass spectral
analysis (correct molecular weight and isotope distribution). No
dimeric (or other oligomeric) products were found in the crude
reaction mixture.
We now wish to report a novel photolabile auxiliary that
fulfils the criteria and has enabled the formation of small cyclic
peptides from “difficult” linear precursors in good yield and
purity.
Results
Cyclization of Ala-Phe-Leu-Pro-Ala. We selected the linear
peptide Ala-Phe-Leu-Pro-Ala for evaluating the ring contraction
strategy. Cyclization of this unsubstituted peptide has been
reported to yield cyclic dimers (decapeptide) and higher
oligomers but no cyclopentapeptide15 and thus serves as an
excellent model to examine new peptide cyclization auxiliaries.
(17) Fukuyama, T.; Jow, C.-K.; Cheung, M. Tetrahedron Lett. 1995, 36,
6373-6374.
(18) Reichwein, J. F.; Liskamp, R. M. J. Tetrahedron Lett. 1998, 39,
1243-1246.
(19) Yang, L.; Chiu, K. Tetrahedron Lett. 1997, 38, 7307-7310.

9792 J. Am. Chem. Soc., Vol. 121, No. 42, 1999
Meutermans et al.
Scheme 5. Cyclization of Auxiliary Containing Peptides
Scheme 4. Estimated Activation Energies for Lactam
Formation from the Linear Unsubstituted Peptide (a) and
from the Cyclic Auxiliary Esters (b,c)
9-10 (A) and Formation of the Target Cyclic Peptides 19
(B)a
The monocyclic product has the thioester structure, as
confirmed by saponification of the monocyclic product (11) in
NH4HCO3, which generated disulfides 12 of the linear peptide
amides and acids. Attempts to force ring contraction by heating
(65 °C) the isolated ester in organic solvents (DMF, dioxane)
in the presence of base (DIEA, DBU {1,8-diazabicyclo[5.4.0]-
undec-7-ene}), or heating in aqueous buffers (pH 4-8) failed.
The ester either remained unchanged or hydrolyzed to the linear
peptide.
a (I) 3 equiv of BOP/5 equiv of DIEA, 3 h at room temperature; (II)
1 equiv of BOP/2 equiv of DIEA, 3 h rt; 10 equiv of DIEA, 12 h rt or
1 h at 65°C; (III) hν (366 nm).
This failure to ring contract may be due to steric hindrance
in the transition state caused by the auxiliary. To evaluate this
we undertook a series of molecular modeling investigations.
the 2-hydroxynitrobenzyl auxiliaries 2 and 3 in the ring con-
traction approach, as initial cyclization would generate a more
reactive nitrophenylester with significantly improved acyl trans-
fer kinetics.22,23 Moreover, because of the similarity with pho-
tolabile ortho-nitrobenzyl linkers, we expected that photolysis
would enable removal of auxiliary 3.24
Evaluation of the Transition-State Geometry. The proposed
ring contraction approach involves a bicyclic transition state
comprised of a larger peptidic ring as determined by the length
of the peptide and a smaller ring determined by the size of the
auxiliary. To evaluate the ring strain and steric crowding of
this transition state for the proposed auxiliaries, we undertook
Monte Carlo conformational searching using the Amber* force
field in the Macromodel program, using a procedure similar to
Cavelier-Frontin.20 Tetra-alanine was selected as a model pep-
tide. The transition state of the cyclization reaction was
simulated by applying a series of distance constraints that have
been derived from experimental data.21 Bu¨rgi et al. had
previously used structural regularities in crystal structures to
map the reaction coordinates for the attack of a nucleophilic
nitrogen atom and a carbonyl group. Applying this transition-
state geometry to both unsubstituted peptides and auxiliary-
containing peptides followed by extensive exploration of their
conformational space allowed the calculation of the low-energy
conformations that are able to adopt the transition-state geom-
etry. The activation energy required for achieving this transition
geometry was then calculated by subtracting the energy of the
“lowest” conformation found in an unconstrained conformational
search from the energy of the “lowest” transition-state confor-
mation (Scheme 4).
The auxiliaries 2 and 3 were readily introduced by a two-
step reductive alkylation25 of the N-terminal primary amine 4
with aldehydes 6 or 7 (Scheme 2, Route 2). (5-Nitro-2-hydroxy
benzaldehyde is commercially available, while 6-nitro-2-hydroxy
benzaldehyde is readily synthesized via a Duff reaction.38) A
small excess of aldehyde was initially used to generate the
intermediate imine in quantitative yields (5 min). Reduction of
the imine after removal of the excess aldehyde was then
accomplished using NaBH4 (5 min). HF cleavage produced the
N-terminal substituted peptides 9 and 10 in high yield and
excellent purity. Racemization or incomplete alkylation of the
N-terminal amino acid was not observed.
Cyclization Using 5-Nitro-2-hydroxybenzyl Auxiliary 2.
Cyclization of peptide 9a under standard conditions initially
yielded two monocyclic products as well as significant amounts
of a side product 13a (Mr, 812 Da), caused by reaction of the
phenol functionality with excess BOP in the reaction mixture
(Scheme 5, A). By adjusting the amount of activating reagent
and base, formation of this side product was completely avoided.
The reaction conditions were further optimized by altering the
temperature and amount of base after an initial cyclization period
and monitoring the formation of monocyclic products by LC/
MS (liquid chromatography/mass spectrometry) analysis. The
best results were obtained when, after 3 h of reaction (1 mM in
DMF, 1 equiv of BOP, 2 equiv of DIEA, rt), excess DIEA (10
equiv) was added and the mixture was left standing for 24 h or
heated to 65 °C for 1 h. The HPLC profile of the crude product
is depicted in Figure 1B. The main product (50% isolated yield)
was characterized by NMR, ES-MS (electrospray mass spec-
The calculated activation energies for amide bond formation
in the cyclization of the unsubstituted linear peptide (Scheme
4a) and the auxiliary-linked equivalents (Scheme 4b,c) are very
similar. The results suggest that the additional five- or six-
membered ring system (from the ethanethiol or hydroxybenzyl
auxiliaries respectively) in the bicyclic transition state does not
introduce significant steric strain.
Nitrobenzyl Auxiliaries 2, 3. Following the above findings,
we concluded that the ring contraction of the cyclic alkylthio-
ester 11 (Scheme 3) was mostly impeded by a low reactivity of
the alkylthioester toward secondary amines rather than by a
constrained transition-state geometry. We decided to examine
(22) Kemp, D. S.; Choong, S.-L. H.; Pekaar, J. J. Org. Chem. 1974, 39,
3841-3847.
(23) Miranda, L. P.; Meutermans, W. D. F.; Smythe, M. L.; Alewood,
P. F. J. Org. Chem. 1999, Submitted for publication.
(24) Holmes, C. P. J. Org. Chem. 1997, 62, 2370-2380.
(25) Ede, N. J.; Ang, K. H.; James, I. W.; Bray, A. M. Tetrahedron
Lett. 1996, 37, 9097-9100.
(20) Cavelier-Frontin, F.; Pe`pe, G.; Verducci, J.; Siri, D.; Jacquier, R.
J. Am. Chem. Soc. 1992, 114, 8885-8890.
(21) Bu¨rgi, H.; Dunitz, J.; Shefter, E. J. Am. Chem. Soc. 1973, 95, 5065-
5067.

Synthesis of Difficult Cyclic Peptides
J. Am. Chem. Soc., Vol. 121, No. 42, 1999 9793
Figure 2. HPLC analysis of the photolysis of cyclic peptide 18a at
timed intervals. A 0.15 mM solution of peptide 18a in MeOH/1% AcOH
was photolyzed using a standard UV lamp, and at different time
intervals, small aliquots were injected onto a Zorbax reversed-phase
C-18 (3 µm, 300 Å, 0.21 × 5 cm) HPLC column. The products were
separated using a linear 0-80% buffer B gradient over 10 min at a
flow rate of 200 µL/min (detection at 214 nm).
Similarly N-(6-nitro-2-hydroxybenzyl)Phe-Leu-Pro-Ala-Ala
10c was assembled and cyclized as above. The all-L cyclo
pentapeptide 18c was isolated in 45% yield.
Removal of the Auxiliary. Cyclic peptide 18a was then
subjected to photolysis at 366 nm, using a standard UV lamp,
in a range of solvent conditions. In most solvents (MeOH,
MeOH/AcOH, THF/AcOH, dioxane) (AcOH ) acetic acid;
MeOH ) methanol; THF ) tetrahydrofuran) the nitrobenzyl
substituent on the backbone nitrogen is readily removed to
generate the target cyclic peptide 19a (Scheme 5, B). Figure 2
illustrates the clean and efficient conversion (18a to 19a). The
cyclic product was characterized by chiral amino acid analysis
Figure 1. HPLC analysis of cyclization of linear peptide 9a (A) after
3 h at room temperature and (B) 1 h heating to 65 °C in the presence
of excess DIEA. The solutions were dried under high vacuum, were
dissolved in 50% aqueous acetonitrile, and were loaded directly onto
a Vydac reversed-phase C-18 (5 µm, 300 Å, 0.46 × 25 cm) HPLC
column. The products were separated using a linear 0-80% buffer B
gradient over 40 min at a flow rate of 1 mL/min.
1
and H NMR. The spectral data are in good agreement with
the reported data.15 Further, an independent sample of cyclic
peptide, prepared from the cyclization of Phe-Leu-Pro-Ala-Ala
according to Schmidt et al.,15 coeluted with the product obtained
from photolysis.
trometry), and chiral amino acid analysis as the all-L target
1
monocyclic product 17a. A H NMR absorption at 11.5 ppm
confirmed that the product contained the free hydroxy substitu-
ent and thus did not have the ester structure but rather the target
cyclic amide structure. Further, a small amount of the C-
terminally racemized product 17b (see Figure 1B) was also
isolated. A chiral amino acid analysis of the product displayed
the presence of a D-Ala residue.
In an attempt to isolate the intermediate cyclic ester 15a, the
reaction mixture was analyzed after the initial 3 h cyclization
period by HPLC (Figure 1A) and LC/MS. The mixture
contained linear peptide 9a and monocyclic products 17a and
17b, but no monocyclic ester was found. The p-nitrophenyl ester
presumably hydrolyzes in the aqueous workup to the linear
peptide.
The same product 19a was obtained from photolysis of the
regio analogue 18c. The racemized cyclic product 18b was
photolyzed and similarly produced the unsubstituted D-Ala
containing product 19b, which coeluted with an independently
synthesized sample.
The rate of photolysis is highly dependent on the presence
of acid or base. In acidic environment (MeOH, MeOH/AcOH,
THF/AcOH), photolysis is fast and clean, whereas in basic
environment (THF dried on KOH, or in the presence of 10%
hydrazine), photolysis is very slow (Figure 3).
Discussion
Cyclization Using 6-Nitro-2-hydroxybenzyl Auxiliary 3.
As the 5-nitro-2-hydroxybenzyl auxiliary is not readily removed
after cyclization, we examined the 6-nitro-2-hydroxybenzyl
auxiliary peptide 10a toward cyclization. The ortho-nitro
substituent, while maintaining a similar activation effect on the
ring contraction of the cyclic intermediate 16a (compared to
15a), has the added benefit that it should render the auxiliary
photolabile. The linear peptide 10a was synthesized and treated
as described above for the 5-nitro-2-hydroxy derivative (Scheme
5A). Thus cyclization (at 1 mM in DMF, 1 equiv of BOP/
2 equiv of DIEA) was performed at room temperature for 3 h,
followed by addition of excess DIEA (10 equiv) and heating to
65 °C for 1 h. The major product was isolated in 39% yield
and characterized by NMR and chiral amino acid analysis as
the all-L cyclo-pentapeptide 18a. A small amount of the
C-terminal racemized cyclic product (containing a D-Ala) 18b
was also isolated.
The synthesis of cyclic peptides generally involves an
entropically disfavored head-to-tail cyclization. The choice of
linear precursor for cyclization can dictate the level of success
of the synthesis, and consequently, the cyclization position
should be carefully chosen according to a number of simple
guidelines: (i) the cyclization site should not be sterically
encumbered, (ii) if possible cyclization should occur between
a D and an L-residue,13 (iii) turn-inducing elements such as Gly
and Pro favor cyclization,15 and (iv) intrachain hydrogen bonds
can facilitate peptide cyclization.26 However, adhering to these
simple guidelines does not guarantee that any cyclic product
will be obtained. For example, the control sequence used in
this study Ala-Phe-Leu-Pro-Ala abides by these simple rules,
yet no cyclic monomeric product is obtained upon cyclization.
In addition, in the development of a backbone linker that is
(26) Wenger, R. M. HelV. Chim. Acta 1984, 67, 502.

9794 J. Am. Chem. Soc., Vol. 121, No. 42, 1999
Meutermans et al.
closure/ring contraction process is complete in 24 h at room
temperature, or by heating to 65 °C for 1 h. This is a remarkable
result, considering that no monocyclic product is obtained from
the same sequence without the auxiliary.
While the 5-nitro-2-hydroxybenzyl auxiliary (2) is not readily
removed after cyclization, we observed that photolysis at 366
nm efficiently removed the 6-nitro-2-hydroxybenzyl auxiliary
(3) for compounds 18a-c and generates the target product in
good yield and purity. The rate of photocleavage is highly
dependent on the selected conditions. In general a basic
environment significantly reduces the photolysis rate and could
be used to protect the N-substituted products 18 from premature
photolysis, if required. Similarly, the linear N-substituted
precursors 10 in their neutral form appear to be more stable to
photolysis presumably due to the presence of the internal
secondary amine. This simplifies the ring contraction strategy
as it is not necessary, at least in the studied cases, to protect
these compounds from light at any stage in the process. Further,
the standard laboratory light does not produce sufficient levels
of UV light at 366 nm to cause significant photolysis of
N-substituted products 18.
Consequently, we have developed an auxiliary that can be
easily introduced at the N-terminus of any amino acid (except
Pro) and that preorganizes the peptide prior to lactamization
by formation of a larger intermediate phenolic ester, thus
reducing entropic loss upon ring contraction and formation of
the target monocycle. This intermediate is suitably activated to
undergo rapid acyl transfer, even in the presence of steric bulk
at the cyclization site. The auxiliary is then readily removed by
photolysis. By using this photolabile auxiliary, we obtained
cyclo-[Ala-Phe-Leu-Pro-Ala] in 21% overall yield (after isola-
tion) from a linear precursor which in absence of the auxiliary
produced no target product at all.
Figure 3. Yield of photolysis of peptide 18a in varying solvent systems
as a function of time. The yield is calculated from the integration of
the product peak in the total HPLC chromatogram (214 nm): (squares)
MeOH/1% AcOH; (circles) MeOH; (triangles) Dioxane; (diamonds)
Dioxane/ 10% hydrazine;
useful for the synthesis of cyclic peptides, Bourne et al found
better cyclization yields between a sterically encumbered Ile,
Pro cyclization site.16
We wanted to overcome the extensive oligomerization that
troubles the cyclization of small cyclic peptides, by developing
a versatile ring closure/ring contraction strategy. We initially
selected the ethanethiol auxiliary for this purpose and had found
that, for larger peptides with unhindered cyclization sites (Gly-
Gly), ring closure and ring contraction proceeded satisfactorily.
These results were in good agreement with the work of Shao et
al.,27 who reported a similar approach for the synthesis of large
cyclic peptides using unhindered cyclization sites. Also Botti
et al.28,29 reported an elegant ring closure/ring contraction
process that involved an O-to-N acyl transfer via a tricyclic
transition state. However, for more hindered cyclization sites
or for “difficult” sequences such as our model peptide, we found
that the desired ring contraction does not proceed at all. Thus
the simple strategy of extending the peptide sequence with a
flexible ethanethiol moiety avoids oligomerization in the cy-
clization process, but for “difficult” sequences such as our model
peptide no ring contraction occurs even after prolonged heating.
On the basis of molecular modeling investigations, we
concluded that the ring contraction is impeded by a low
reactivity of the alkylthioester toward secondary amines, rather
than by a sterically constrained transition-state geometry as a
result of the extra atoms of the auxiliary. This is further
supported by the fact that in our cyclization experiments on
peptide 8 no intermolecular acyl transfer (i.e., oligomerization
and not requiring a bicyclic transition state) occurred upon
prolonged heating of the cyclic thioester 11. In the presence of
hydroxide anions, the thioester hydrolyzed followed by oxidation
of the thiol to the disulfide 12. This suggests that ring contraction
may be possible for difficult sequences when more reactive
intermediate esters are formed.
Conclusion
Cyclic peptides or cyclic peptidomimetics have been the
subject of intensive research, especially in the pharmaceutical
industry. A large number of these targeted molecules remain
inaccessible, despite the growing interest. In this work we have
successfully introduced a novel auxiliary strategy that will
significantly expand the repertoire of cyclic peptides and
peptidomimetics that are available synthetically.
The development of this activated and reversible auxiliary
has implications in other peptide research areas, such as native
ligation and backbone substitution, where current reversible
auxiliaries suffer from a rate-limiting acyl transfer step. We are
investigating the use of the auxiliary in these areas in addition
to the development of cyclic peptide libraries.
Experimental Section
Materials and Methods. Chlorotrityl resin (sv ) 0.92 mmol/g) was
purchased from PepChem (Tubingen, Germany). All Wang resins and
NR-tert-butoxycarbonyl-L-amino acids were peptide synthesis grade
purchased from Auspep (Melbourne Australia), Novabiochem (San
Diego) or Peptide Institute (Osaka, Japan). Pam (phenylacetamidom-
ethyl) resins were purchased from Applied Biosystems (Foster City,
CA). Dichloromethane, diisopropylethylamine, N,N-dimethylformamide,
and trifluoroacetic acid were obtained from Auspep (Melbourne,
Australia). p-Cresol, p-thiocresol, 3-nitrophenol, 5-nitro-2-hydroxy-
benzaldehyde, polyphosphoric acid, and hexamethylenetetramine were
purchased from Aldrich or Fluka (Sydney, Australia). HPLC grade
acetonitrile was purchased from BDH (Brisbane, Australia). 2-(1H-
benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluoro-phosphate
(HBTU) and benzotriazole-1-yl-oxy-tris(dimethylamino)-phosphonium
hexafluorophosphate (BOP) were purchased from Richelieu Biotech-
We thus focused on optimizing the acyl transfer step and
discovered that the more reactive nitrobenzyl auxiliaries (2 and
3) provide enhanced intermolecular acyl transfer rates that are
very tolerant to steric bulk at the acyl transfer site.23 When used
for the synthesis of cyclic peptides, the nitro-substituted
auxiliaries 2 and 3 yielded the desired N-substituted monocyclic
products in 50% and 39% isolated yield, respectively. The ring
(27) Shao, Y.; Lu, W.; Kent, S. B. H. Tetrahedron Lett. 1998, 39, 3911-
3914.
(28) Botti, P.; Pallin, T. D.; Tam, J. P. J. Am. Chem. Soc. 1996, 118,
10018-10024.
(29) Zhang, L. S.; Tam, J. P. J. Am. Chem. Soc. 1997, 119, 2363-2370.

Synthesis of Difficult Cyclic Peptides
J. Am. Chem. Soc., Vol. 121, No. 42, 1999 9795
nologies (Quebec, Canada). Deionized water was used throughout.
Screw-cap glass peptide synthesis reaction vessels (10 mL) with sintered
glass filter frit were obtained from Embell Scientific Glassware
(Queensland, Australia). Argon, helium, and nitrogen (all ultrapure
grade) were from BOC (tert-butoxycarbonyl) gases (Queensland,
methanol, 600 mL of acetonitrile, 300 mL of stock bufffer; stock buffer
was ammonium formate 10 mM pH 5.2).
Peptide Synthesis. All linear peptides were chemically synthesized
stepwise using either Fmoc (9-fluorenylmethyloxycarbonyl) or Boc
protecting groups and in situ HBTU activation protocols as previously
described.31,32 Coupling efficiencies were determined by the quantitative
ninhydrin test33 and recoupled where necessary to obtain >99.5%
efficiency.
Reductive Amination. The selected auxiliary-aldehyde (0.1 M) was
dissolved in MeOH/DMF (1:1) or DMF/AcOH (100:1) and added to
the amino peptide resin (2 equiv). After 5 min the resin was filtered
and a second portion of aldehyde (2 equiv) added. After another 5 min
the resin was filtered and washed with MeOH/DMF (1:1) or DMF/
NaBH4 (10 equiv) in MeOH/DMF (1:3) was then added and the reaction
mixture was left standing for 5 min. The resin was again filtered and
washed with MeOH/DMF (1:3), DMF, MeOH/DCM (1:1) (DCM )
dichloromethane) and air-dried prior to cleavage.
Cleavage. Peptides on chlorotrityl resin were cleaved using 1% TFA
in DCM (10 mL/500 mg resin, 1 h at room temperature). The resin
was filtered and washed with 50% buffer B (see HPLC section), the
filtrates were combined, and DCM was evaporated in vacuo. The
solution containing the crude product was then loaded onto a preparative
Vydac column, and the products were separated using a 1.5% linear
gradient from 100% A to 20% A. Fractions were subjected to MS
analysis.
Peptides on Wang resin were cleaved using TFA/water (95:5) [10
mL/500 mg resin, 1 h at room temperature]. The resin was filtered and
washed with neat TFA. TFA was then removed in vacuo and the residue
dissolved in 50% buffer B (see HPLC section) for HPLC purification
(as above).
1
Australia). H NMR and 13C NMR spectra were recorded on a 300
MHz instrument, and chemical shifts are reported in parts per million
(ppm) downfield from (CH3)4Si. Reversed-phase high-performance
liquid chromatography was performed on a microbore (C-18, 3 µm,
0.21 cm × 5 cm) column, an analytical (C-18, 5 µm, 0.46 cm × 25
cm) column, or a preparative (C-18, 10 µm, 2.2 cm × 25 cm) column.
Chromatographic separations were achieved using linear gradients of
buffer B in A (A ) 0.1% aqueous TFA [trifluoroacetic acid]; B )
90% CH3CN, 10% H2O, 0.09% TFA) at a flow rate of 0.25 mL/min
(microbore), 1 mL/min (analytical) and 8 mL/min (preparative).
Mass Spectrometry. Mass spectra were acquired on a triple
quadrupole mass spectrometer equipped with an Ionspray atmospheric
pressure ionization source. Samples (10 µL) were injected into a moving
solvent (30 µL/min; 50:50 CH3CN/0.05% TFA) coupled directly to
the ionization source via a fused silica capillary interface (50 mm i.d.
× 50 cm length). Sample droplets were ionized at a positive potential
of 5 kV and entered the analyzer through an interface plate and
subsequently through an orifice (100-120 mm diameter) at a potential
of 80 V. Full scan mass spectra were acquired over the mass range
400-2000 Da with a scan step size of 0.1 Da. Molecular masses were
derived from the observed m/z values. LC/MS runs were carried out
using a linear gradient on a dual syringe pump solvent delivery system
and a reversed-phase microbore (C-18, 3.5 µm, 0.21 cm × 5 cm)
column at a flow rate of 150 µL/min. Samples (typically 5 µL of 1
mg/mL solution) were loaded directly on the column, and the eluent
was directly connected to the mass spectrometer via a 30 cm, 75 mm
i.d., fused silica capillary. The application of Turbo Ionspray (5 L/min
N2 at 500 °C) allowed the introduction of the total eluent without
splitting and loss in sensitivity. Acquisition parameters were as
described above.
Peptides on PAM resin were cleaved as follows: 300 mg of resin
was mixed with 0.5 mL of p-cresol and 0.5 mL of p-thiocresol, 9 mL
of HF was added at 0 °C, and the mixture was stirred at 0 °C for 1 h.
After evaporation of the HF the crude product was precipitated and
washed with ether (2 × 10 mL). The precipitate was then dissolved in
50% CH3CN in water (0.095% TFA) for HPLC purification (as above).
S-(4-Methylbenzyl)-2-thioethanol (5). 2-Bromoethanol (25.0 g, 0.20
mol) was added dropwise to 4-methylbenzylmercaptan (27.6 g, 0.20
mol) in DMF (200 mL) at 0 °C. Triethylamine (29.3 mL, 0.21 mol)
was then added dropwise over a 1 h period. The solution was allowed
to stir for 3 h. Solids were filtered, H2O (400 mL) was added, and the
mixture was extracted with EtOAc (3 × 150 mL). The extracts were
combined and washed with H2O (2 × 100 mL). The organic layer was
dried over MgSO4, and the volatiles were removed in vacuo. The
product 5 was purified by distilation at 130-132 °C (4.6 mm Hg) to
Monte Carlo Conformational Search. Conformational searches
were performed both with and without transition-state constraints on
both unsubstituted tetra-alanine and auxiliary-containing tetra-alanine.
The difference in energy between the low-energy constrained and
unconstrained conformations was considered to be the activation energy
required to bring the molecule into the transition-state geometry. Each
conformational search was performed in two steps, using the AMBER*
force-field and the gb/sa continuum solvent model for water. Initially
15 000 iterations of Monte Carlo were performed where any backbone
bond was free to be included in the search. If a peptide bond was
chosen, it was always flipped. Only 250 minimization iterations were
performed per conformer, and if after 100 iterations the conformer was
>100kJ/mol from the lowest-energy conformer, it was rejected. If after
250 minimization iterations the conformer was >50 kJ/mol from the
lowest-energy conformer, it was rejected. In the second step, the
conformers found from the Monte Carlo search were fully minimized
using conjugate gradients. Identity checks against other conformers were
performed to exclude duplicates.
1
yield a colorless oil (26.2 g, 75%) H NMR (300 MHz, CDCl3, ppm)
δ 7.21 (d, J ) 8.2 Hz, 2H), 7.13 (d, J ) 8.2 Hz, 2H), 3.69 (s, 2H),
3.65 (t, J ) 6.0 Hz, 2H), 2.63 (t, J ) 6.0 Hz, 2H), 2.33 (s, 3H); 13C
NMR (75 MHz, CDCl3, ppm) δ 137.04, 135.11, 129.49, 128.92, 60.35,
35.58, 34.54, 21.29.
Ala-Phe-Leu-Pro-Ala. Synthesized on chlorotrityl resin (0.96 mmol/
g) on a 0.1 mmol scale. Yield after TFA cleavage was 23 mg (45%):
ES-MS Mr 517.1, calcd for C26H39N5O6, 517.3 (monoisotopic).
(D)Ala-Phe-Leu-Pro-Ala. Synthesized on chlorotrityl resin (0.96
mmol/g) on a 0.1 mmol scale. Yield after TFA cleavage was 16 mg
(31%): ES-MS Mr 517.1, calcd for C26H39N5O6, 517.3 (monoisotopic).
Phe-Leu-Pro-Ala-Ala. Synthesized on Boc-Ala-PAM resin (0.75
mmol/g) on a 0.1 mmol scale. Yield after HF cleavage was 47 mg
(91%): ES-MS Mr 517.2, calcd for C26H39N5O6, 517.3 (monoisotopic).
Chiral Amino Acid Analysis. Determination of the chirality of
peptides was performed according to the method of Goodlett30 with
the following modifications. Hydrolysate: samples were hydrolyzed
in 6 N HCl for 24 h at 105 °C using a Pico-Tag workstation. Trace
amounts of HCl were removed using a Savant speedivac. Derivatiza-
tion: hydrolysates were suspended in 100 µL of 1 M NaHCO3 and 10
µL of 38.7 mM 1-fluoro-2,4-dinitrophenyl-5-L-alanine amide in acetone.
Samples were then incubated at 40 °C for 60 min. The reaction was
quenched with the addition of approximately 300 µL of 1 N HCl. The
samples were injected directly into the HPLC with no dilution. HPLC
analysis was performed on a YMC basic column (0.46 cm × 250 cm)
using a gradient of 0-100% B over 45 min, with the column heated to
45 °C and with UV detection at 340 nm (buffer A, 100 mL of methanol,
50 mL of acetonitrile, 850 mL of stock buffer; buffer B, 100 mL of
(31) Schno¨lzer, M.; Alewood, P. F.; Jones, A.; Alewood, D.; Kent, S.
B. H. Int. J. Pept. Protein Res. 1992, 40, 180-193.
(32) Alewood, P. F.; Alewood, D.; Miranda, L. P.; Love, S.; Meutermans,
W. D. F.; Wilson, D. Methods Enzymol. 1997, 289, 14-29.
(33) Sarin, V. K.; Kent, S. B. H.; Tam, J. P.; Merrifield, R. B. Anal.
Biochem. 1981, 117, 147.
(34) Kemp, D. S.; Kerkman, D. J.; Leung, S.-L.; Hanson, G. J. Org.
Chem. 1981, 46, 490-498.
(35) Kemp, D.; Galakatos, N.; Dranginis, S.; Ashton, C.; Fotouhi, N.;
Curran, T. J. Org. Chem. 1986, 51, 3320-3324.
(36) Kemp, D. S.; Galakatos, N. G.; Bowen, B.; Tan, K. J. Org. Chem.
1986, 51, 1829-1838.
(30) Goodlett, D. R.; Abuaf, P. A.; Savage, P. A.; Kowalski, K. A.;
Mukherjee, T. K.; Tolan, J. W.; Goldstein, G.; Crowther, J. B. J.
Chromatogr., A 1995, 707, 233-244.
(37) Kemp, D. S.; Carey, R. I.; Dewan, J. C.; Galakatos, N. G.; Kerkman,
D.; Leung, S.-L. J. Org. Chem. 1989, 54, 1589-1603.

9796 J. Am. Chem. Soc., Vol. 121, No. 42, 1999
Meutermans et al.
N-(2-Mercaptoethanyl)-Ala-Phe-Leu-Pro-Ala (8). Ala-Phe-Leu-
Pro-Ala-Wang resin was prepared starting from Fmoc-Ala-Wang resin
(0.44 mmol/g) using standard Fmoc-SPPS (SPPS ) solid-phase peptide
synthesis) protocols.32 To 500 mg of resin was added o-nitrobenzene-
sulfonyl chloride (300 mg) in DMF (4 mL) containing DIEA (200 µL).
After 30 min, the resin was drained and washed with DMF (3 × 6
mL). The resin was mixed with a solution of S-(p-methylbenzyl)-2-
mercaptoethanol (270 mg, 1.5 mmol) in DCM (5 mL). Triphenylphos-
phine (393 mg, 1.5 mmol) and diethylazodicarboxylate (DEAD [diethyl
azodicarboxylate], 261 mg, 1.5 mmol) were premixed in DCM (5 mL),
and after 1 min, the solution was added to the resin and the reaction
was left for 30 min. The resin was then washed with DCM (3 × 6
mL) and DMF (3 × 6 mL) and further treated with a solution of PhSNa
(200 mg, 1.5 mmol) in DMF (4 mL) for 30 min. The resin was finally
washed with DMF (3 × 6 mL) and MeOH/DCM (3 × 6 mL) and
air-dried. After HF cleavage HS-(CH2)2-NH-CH(CH3)-CO-Phe-Leu-
Pro-Ala-OH was separated in 22% yield (25 mg): Mr 577.1, calcd for
C28H43N5O6S, 577.3.
Cyclo-[Phe-Leu-Pro-(D)Ala-Ala]. Cyclization of Phe-Leu-Pro-(D)-
Ala-Ala produced the cyclo-[Phe-Leu-Pro-(D)Ala-Ala] in 55% yield:
ES-MS Mr 499.3, calcd for C26H37N5O5, 499.3 (monoisotopic).
Cyclo-[N-(5-nitro-2-hydroxybenzyl)-Ala-Phe-Leu-Pro-Ala] (17a).
Cyclization of N-(5-nitro-2-hydroxybenzyl)-Ala-Phe-Leu-Pro-Ala 9a
(30 mg of the TFA salt, 0.038 mmol) produced 17a (12.5 mg, 0.019
mmol) in 51% yield: ES-MS Mr 650.2, calcd for C33H42N6O8, 650.3
(monoisotopic). 1H NMR (500 MHz, DMSO-d6, ppm) (DMSO )
dimethylsulfoxide) δ 11.5 (s, 1H, OH), 8.40 (d, 1H, NHLeu), 8.02 (dxd,
1H, H-ar), 7.70 (d, 1H, H-ar), 7.4 (d, 1H, HNPhe), 7.20-7.30 (m, 5H,
H-Phe), 6.99 (d, 1H, H-ar), 6.54 (d, 1H, HNAla), 5.00 (s, 1H, ArCHhN),
4.91 (m, 1H, R-Ala5), 4.75 (q, 1H, R-Ala1), 4.59 (m, 1H, R-Phe), 4.50
(m, 1H, R-Leu), 4.27 (t, 1H, R-Pro), 3.88 (d, 1H, ArCHhN-), 3.62 (m,
1H, δ-Pro), 3.37 (m, 1H, δ-Pro), 2.97 (m, 1H, â-Phe), 2.82 (m, 1H,
â-Phe), 2.04 (m, 2H, â-Pro), 1.88 (m, 1H, γ-Pro), 1.73 (m, 1H, â-Leu),
1.65 (m, 1H, γ-Pro), 1.44 (m, 1H, γ-Leu), 1.33 (m, 1H, â-Leu), 1.24
(d, 3H, â-Ala5), 0.91 (d, 3H, â-Ala1), 0.85 (m, 6H, δ-Leu). 13C NMR
(75 MHz, DMSO-d6, ppm) δ 172.61, 170.34, 170.07, 169.95, 169.47,
160.40, 139.73, 136.88, 129.31, 128.14, 126.50, 125.72, 124.21, 122.65,
115.00, 61.04, 56.50, 55.74, 48.70, 46.31, 44.34, 41.37, 38.28, 31.30,
24.20, 22.81, 22.68, 21.17, 18.97, 15.35.
Cyclo-[N-(6-nitro-2-hydroxybenzyl)-Ala-Phe-Leu-Pro-Ala] (18a).
From cyclization of N-(6-nitro-2-hydroxybenzyl)-Ala-Phe-Leu-Pro-Ala
10a (20 mg of the TFA salt, 0.025 mmol), 18a (6.5 mg, 0.010 mmol)
was obtained in 39% yield: ES-MS Mr 650.6, calcd for C33H42N6O8,
650.3 (monoisotopic). 13C NMR (75 MHz, CD3OD, ppm) δ 178.07,
176.95, 174.54, 174.32, 173.72, 159.11, 153.19, 140.41, 131.99, 129.96,
129.54, 127.57, 121.18, 116.57, 62.75, 60.67, 58.55, 54.05, 51.15, 44.54,
43.41, 34.85, 33.67, 25.03, 24.13, 22.30, 21.31, 15.49, 13.89.
Cyclo-[N-(6-nitro-2-hydroxybenzyl)-Phe-Leu-Pro-Ala-Ala] (18c).
From cyclization of the N-(6-nitro-2-hydroxybenzyl)-Phe-Leu-Pro-Ala-
Ala (20 mg of the TFA salt, 0.025 mmol), 18a (7.3 mg, 0.011 mmol)
was obtained in 44% yield: ES-MS Mr 650.2, calcd for C33H42N6O8,
650.3 (monoisotopic). 13C NMR (75 MHz, DMSO-d6, ppm) δ 171.43,
171.00, 169.46, 167.56, 156.65, 138.43, 129.24, 129.05, 128.32, 128.18,
126.08, 119.50, 115.87, 114.60, 62.18, 60.69, 51.07, 49.38, 46.57,
45.46, 41.54,38.17, 33.65, 31.43, 24.37, 22.73, 22.32, 21.06, 17.87,
16.92.
N-(5-Nitro-2-hydroxybenzyl)-Ala-Phe-Leu-Pro-Ala (9a). Synthe-
sized on chlorotrityl resin (0.96 mmol/g) on a 0.2 mmol scale. Yield
after TFA cleavage was 59 mg (45%): ES-MS Mr 668.2, calcd for
C33H44N6O9, 668.3 (monoisotopic).
N-(6-Nitro-2-hydroxybenzyl)-Ala-Phe-Leu-Pro-Ala (10a). Syn-
thesized on chlorotrityl resin (0.96 mmol/g) on a 0.1 mmol scale. Yield
after TFA cleavage was 22 mg (34%): ES-MS Mr 668.2, calcd for
C33H44N6O9, 668.3 (monoisotopic).
N-(6-Nitro-2-hydroxybenzyl)-Phe-Leu-Pro-Ala-Ala (10c). Syn-
thesized on chlorotrityl resin (0.96 mmol/g) on a 0.2 mmol scale. Yield
after TFA cleavage was 67 mg (51 %): ES-MS Mr 668.2, calcd for
C33H44N6O9, 668.3(monoisotopic).
Cyclization Experiments. Cyclization of auxiliary-containing pep-
tides 9 and 10: 1 equiv of BOP and 2 equiv of DIEA in DMF were
added to a 1 mM solution of the linear peptide in DMF and stirred for
3 h at room temperature. Ten equivalents of DIEA was then added
and the solution heated at 65° for 1 h. DMF was removed in vacuo
and the crude product dissolved in acetonitrile/water (1:1) and purified
by RP-HPLC (reversed-phase high-performance liquid chromatogra-
phy). Cyclization of other linear peptides: Cyclizations were performed
using a 1 mM solution of linear peptide in DMF. Three equivalents of
BOP and 5 equiv of DIEA were added, and the solution was stirred
for 3 h at room temperature. Workup was as described above.
Cyclo-[Ala-Phe-Leu-Pro-Ala] (19a). (a) Cyclo-[N-(6-nitro-2-hy-
droxybenzyl)-Ala-Phe-Leu-Pro-Ala] (1 mM MeOH) was purged with
nitrogen for 30 min and then photolyzed with a standard laboratory
UV lamp (366 nm, 0.25A) for 3 h. The MeOH was evaporated, the
residue was dissolved in 50% buffer B, and the solution was loaded
directly onto a Vydac C18 column (preparative) for HPLC purification.
Cyclo-[Ala-Phe-Leu-Pro-Ala] was isolated in 52% yield. The product
was coeluted with a independently synthesized sample: ES-MS Mr
499.4, calcd for C26H37N5O5, 499.3 (monoisotopic). (b) Photolysis of
purified cyclo-[N-(6-nitro-2-hydroxybenzyl)-Phe-Leu-Pro-Ala-Ala] was
perfomed as described above. Cyclo-[Phe-Leu-Pro-Ala-Ala] was iso-
lated in 28% yield. The product coeluted with a independently
synthesized sample: ES-MS Mr 499.1, calcd for C26H37N5O5, 499.3
(monoisotopic).
Cyclo-[S-(CH2)2-Ala-Phe-Leu-Pro-Ala] (11). Cyclization of HS-
(CH2)2-Ala-Phe-Leu-Pro-Ala 8 (10 mg of the TFA salt, 0.014 mmol)
produced the monocyclic thioester 11 (3.4 mg, 45% yield): Mr, 559.3,
calcd for C28H41N5O5S, 559.3. The thioester was hydrolyzed using
aqueous ammonium bicarbonate buffer (0.1 M, pH 8, 6h at 60 °C) to
form the C-terminal amides and acids. Under the mild base conditions
these thiol-products oxidized to the disulfides 12 which were character-
ized by ES-MS: [S-(CH2)2-NH-CH(CH3)-CO-Phe-Leu-Pro-Ala-NH2]2
Mr, 1150.8, calcd for C56H86N12O10S2, 1150.6; [S-(CH2)2-NH-CH(CH3)-
CO-Phe-Leu-Pro-Ala-NH2]-S-(CH2)2-NH-CH(CH3)-CO-Phe-Leu-Pro-
Ala-OH Mr, 1151.8, calcd for C56H85N11O11S2, 1151.6; [S-(CH2)2-
NH-CH(CH3)-CO-Phe-Leu-Pro-Ala-OH]2 Mr, 1152.8, calcd for
C56H84N10O12S2, 1152.6.
Acknowledgment. We acknowledge Glaxo Wellcome UK
and Glaxo Wellcome Australia for their financial support. Ms.
Trudy Bond, Mr. Alun Jones, Mr. Peter Seale, Dr. Martin
Scanlon, and Dr. Mike Hann are also acknowledged. L.P.M.
was supported by an Australian post-graduate research award
(APRA) scholarship from the Australian Government.
Cyclo-[Phe-Leu-Pro-Ala-Ala]. Cyclization of Phe-Leu-Pro-Ala-Ala
produced the cyclo-[Phe-Leu-Pro-Ala-Ala] in 6% yield: ES-MS Mr
499.4, calcd for C26H37N5O5, 499.3 (monoisotopic).
(38) Harayami, T.; Nakatsuka, K.; Nishioka, H.; Murakami, K.; Ohmori,
Y.; Takeuchi, Y.; Ishii, H.; Kenmotsu, K. Heterocycles 1994, 38, 2729-
2738.
JA992173Y