Protection by conformationally restricted mobility: First solid-phase synthesis of triostin A

Article abstract of DOI:10.1002/chem.200800372

A study was conducted to demonstrate the first solid-phase synthesis of triostin A, applying a new concept in protection during synthesis. The strategy allowed fast and efficient access to the natural product and demonstrated that it can be used to prepare libraries of analogues with modifications on the sequence and/or on the heterocyclic unit. It was found that the solid-phase synthesis employed several coupling reagents, a number of orthogonal protecting groups, and the new concept of protection, called conformationally restricted mobility. It was also demonstrated the new conformationally restricted mobility was essential for the minimization of DKP formation and for the success of the approach. The conformationally restricted mobility strategy was used to synthesize triostin A, as it reduced the number of steps and minimized the risk of side-reactions.

Full text of DOI:10.1002/chem.200800372

COMMUNICATION
DOI: 10.1002/chem.200800372
Protection by Conformationally Restricted Mobility:
First Solid-Phase Synthesis of Triostin A
Judit Tulla-Puche,*[a] Eleonora Marcucci,[a] Manuel Fermin,[a] Nfflria Bayó-Puxan,[a] and
Fernando Albericio*[a, b, c]
Triostin A (Figure 1) belongs to a family of peptide antibi-
otics produced by marine bacteria. Triostins (A, B, and C)
were isolated in 1961 from a strain of Streptomyces aureus.[1]
Differing only in one of the NMe residues, triostin A bears
an NMe-Val, whereas triostin B and C contain an NMe-Ile
and an N,g-dimethyl-allo-Ile, respectively.[2] Triostin A, as
well as other members of this family, which include echino-
mycin,[3] BE-22179,[4] and thiocoraline[5,6] show antibacterial
and cytotoxic activity.[7]
logue results in a change of selectivity to TpA sequences.[10]
Later studies have shown that binding to specific bases is
not only driven by hydrogen bonding, but that these com-
pounds also search for the most favourable stacking interac-
tions.[11]
The difficulty to synthesize triostin A is exemplified by
the fact that since its discovery more than 30 years ago, only
two elegant, but tedious, solution chemistry syntheses, which
involved purification at every intermediate, have been per-
formed more than 20 years ago.[12] To elucidate the mode of
action of this family of molecules, further biological studies
require an efficient synthetic protocol.[13]
The presence of consecutive NMe amino acids in triostin
A, which intrinsically implies synthetic difficulty, and of the
two ester bonds, which favour the formation of diketopiper-
azines (DKPs) as the main problem to be solved, makes the
solid-phase approach of this peptide a challenge.[14] Herein
we describe the first solid-phase synthesis of triostin A, ap-
plying a new concept in protection during synthesis. The
present strategy allows fast and efficient access to the natu-
ral product, and could also be used to prepare libraries of
analogues with modifications on the sequence and/or on the
heterocyclic unit.
Figure 1. Structure of triostin A.
Triostin A binds to DNA by bisintercalation[8] through its
two quinoxaline units and shows CpG selectivity.[9] The re-
moval of the four N-methyl groups in the TANDEM ana-
The solid-phase synthesis employs several coupling re-
agents, a number of orthogonal protecting groups, and the
new concept of protection, referred to as conformationally
restricted mobility, which is crucial for the minimization of
DKP formation and for the success of the approach.
Stepwise elongation: Taking advantage of the C2 symme-
try of triostin and given its synthetic difficulty, the strategy
of choice is a [4+4] fragment coupling, because it reduces
the number of steps and therefore minimizes the risk of
side-reactions. The 4-amino acid fragment contains two
ester bonds (between the first amino acid and the resin and
between the Val and the d-Ser) and therefore two points
where DKPs could form. This risk is increasing because the
two remaining amino acids are NMe-Val and NMe-Cys. The
NMe increases the presence of the cis-configuration in the
dipeptide, thereby making it prone to DKP formation.[15]
[a] Dr. J. Tulla-Puche, Dr. E. Marcucci, M. Fermin, Dr. N. Bayó-Puxan,
Prof. F. Albericio
Institute for Research in Biomedicine, Barcelona Science Park
Josep Samitier 1, 08028 Barcelona (Spain)
Fax : (+34)93-403-7126
E-mail: judit.tulla@irbbarcelona.org
albericio@irbbarcelona.org
[b] Prof. F. Albericio
Department of Organic Chemistry, University of Barcelona
Martí i FranquØs 1, 08028 Barcelona (Spain)
[c] Prof. F. Albericio
CIBER-BBN, Networking Centre on Bioengineering
Biomaterials and Nanomedicine
Barcelona Science Park, Josep Samitier 1, 08028 Barcelona (Spain)
Supporting information for this article is available on the WWW
under http://www.chemistry.org or from the author.
Chem. Eur. J. 2008, 14, 4475 – 4478
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4475

Previous experience with analogues of this family prompt-
ed the selection of the 4-amino acid fragment shown in
Figure 2 as a target intermediate and the following approach
for its synthesis.
Qxc was introduced by reaction with PyBOP/HOAt/DIEA
in DMF. The Trt was removed by washings with TFA/TIS/
CH2Cl2 2:2.5:95.5 until colourless filtrates were obtained.
After washing the resin with CH2Cl2, introduction of Alloc-
NMe-Val-OH was accomplished by treatment with
DIPCDI/DMAP in CH2Cl2/DMF 9:1. Removal of the
Alloc-protecting group with [Pd
ACHTREUNG
CH2Cl2, followed by introduction of Boc-NMe-Cys
ACHTREUNG
OH afforded the target tetradepsipeptide with 85% purity.
Coupling of this second consecutive NMe amino acid was
particularly difficult, therefore the coupling reaction using
HATU/HOAt/DIEA was repeated.
Figure 2. Sequence of the tetrapeptide moiety.
i) A Wang-type resin was used for the synthesis because
the Cl-TrtCl-resin,[16] known to minimize DKP formation, is
incompatible with Trt protection, used for temporary pro-
tecting group of the hydroxyl function of the d-Ser; ii) Ala
was used as the starting point for the synthesis. Since the
second amino acid is d-Ser, which is then elongated through
its side-chain, the risk of DKP formation is kept to mini-
Protection by conformationally restricted mobility:[18] For-
mation of the disulfide bridge: Synthetic organic chemistry
is based on the concourse of reagents and catalysts to
achieve clean formation of new bonds and appropriate
protecting groups to prevent the formation of undesired
bonds as well as side-reactions.
mum. After introducing Fmoc-d-Ser
G
Conventional protection chemistry masks the properties
of the reactive chemical groups. Thus, carbamates such as
Fmoc, pNZ, or Boc, used in this study, reduce the nucleo-
philicity and basicity of the amines. A different approach for
protection is by the introduction of steric congestion around
the reactive group. A good example of this is the Trt group.
Thus, the Cl-TrtCl-resin, in addition to linking the carboxyl
group of the C-terminal amino acid to the resin, reduces
DKP formation (attack of the amine of the second amino
acid at the carboxyl group of the first amino acid) by steric
hindrance. In this regard, Barlos has reported that the Cl at
position 2 is crucial because simply by removing the Cl from
the Cl-TrtCl-resin, this support is more susceptible to DKP
formation.[19] Furthermore, Trt-amino acids, although diffi-
cult to couple to the peptide sequence, are practically free
of racemization, again due to the steric hindrance induced
around the a-H.[20] Moreover, backbone protection, such as
the introduction of the 2-hydroxy-4-methoxy benzyl (Hmb)
group[21] and the use of pseudoprolines[22] have enabled the
synthesis of difficult peptides
group was removed and the 2-quinoxaline carboxylic unit
(Qxc) was introduced on solid-phase at the dipeptide level,
prior to ester formation.[17] iii) To introduce the two remain-
ing residues NMe-Val and NMe-Cys, different combinations
of protecting groups were investigated. Removal of the
Fmoc and pNZ groups by piperidine and SnCl2, respectively,
at the NMe-Val position resulted in partial cleavage of the
ester bond, with recovery of only the dipeptide (Qxc-d-Ser-
Ala-OH). Thus, at this position the Alloc derivative was in-
troduced, which was removed under practically neutral con-
ditions. For the introduction of NMe-Cys(Acm), the Boc
group was selected, to allow concomitant cleavage with the
release of the peptide from the support (Scheme 1).
To further detail the construction of the tetradepsipeptide,
Fmoc-Ala-OH was coupled onto Wang resin by treatment
with DIPCDI/DMAP. The Fmoc group was removed with
piperidine-DMF (1:4) and Fmoc-d-Ser
A
with HATU/HOAt/DIEA. After removing the Fmoc group,
by disrupting secondary struc-
ture formation. Herein, we de-
scribe a new concept of protec-
tion based on conformational
mobility restriction.
In previous syntheses of ana-
logues, the removal of the pro-
tecting group of NMe-Cys (Boc
in this case) was accompanied
by immediate DKP formation.
Regardless of the protecting
groups used (Fmoc removed
with a base, pNZ and Alloc re-
moved under almost neutral
conditions, and the acid labile
Trt and Boc, removed by acid
Scheme 1. Synthesis of the tetradepsipeptide of triostin A on solid-phase. HATU = O-(7-azabenzotriazol-1-
yl)tetramethyluronium hexafluorophosphate; DIPCDI = N,N’-diisopropylcarbodiimide; Qxc = 2-quinoxaline
and neutralized), the formation
of DKPs was extensive, and in
most cases quantitative (Fig-
carboxylic unit; PyBOP
= benzotriazol-1-yloxy-tris (dimethylamino) phosphonium hexafluorophosphate;
HOAt = 1-hydroxy-7-azabenzotriazole.
4476
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 4475 – 4478

Solid-Phase Synthesis of Triostin A
COMMUNICATION
ure 3A). At this point, we considered that the only solution
to solve this problem was to restrict the mobility of the pep-
tide chain, by making the amino function of the NMe-Cys
Scheme 2. Dimerization and cyclization to obtain triostin A.
Figure 3. Protection by conformationally restricted mobility prevents
DKP formation.
rative reversed HPLC (17% overall yield) (Figure 4C, D).
The presence of two discrete conformers of triostin A in
slow exchange, which may be separated chromatographical-
ly, has been reported.[23] 1H NMR spectra in CDCl3 of the
two compounds were consistent with previously reported
data,[24,25] where the NMe-Val adopts a trans conformation
unable to reach the carboxyl of the NMe-Val. In this case,
this was achieved by the formation of the inter-chain disul-
fide bridge (Figure 3B) which restricts the mobility of the
peptide chain and thereby effectively prevents DKP forma-
tion.
Scheme 2 shows the on-resin
disulfide formation, followed by
the cleavage, and the final
double cyclization, which ren-
dered triostin A (Scheme 2).
The tetrapeptidyl resin was
subjected to two 10 min treat-
ments of oxidation with iodine.
The dimer was cleaved by treat-
ment with TFA/H2O/CH2Cl2
25:5:70, evaporated, redissolved
in H2O/CH3CN 1:1 and lyophi-
lized. The challenging final
macrocyclization (double and
with an NMe-amino as a nucleo-
phile) was performed by using
PyBOP/HOAt/DIEA in CH2Cl2
for 4 h. After workup, 83 mg of
crude triostin A was obtained
with excellent purity (51%, Fig-
ure 4B) for
a crude peptide
after 12 steps, using six protect-
ing groups (Fmoc, Alloc and
Boc for the a-amino, Trt for hy-
droxyl, Acm for Cys, and sup-
ported p-alkoxybenzyl for the
C-terminal), an on-resin disul-
fide formation, and a double
cyclization.
Analytical HPLC and HPLC-
MS analysis revealed the pres-
ence of two peaks of the same
mass in a 65:35 ratio, which
Figure 4. Analytical HPLC of A) crude dimer (tR =7.34 min), B) crude triostin A (tR =10.44 min), C) most
were separated by semi-prepa- polar and D) least polar conformer.
Chem. Eur. J. 2008, 14, 4475 – 4478
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4477

F. Albericio, J. Tulla-Puche et al.
[10] M. A. Viswamitra, O. Kennard, W. Cruse, G. Egert, G. Sheldrick, P.
Jones, M. J. Waring, L. P. G. Wakelin, R. K. Olsen, Nature 1981, 289,
817–819; K. R. Fox, R. K. Olsen, M. J. Waring, Biochim. Biophys.
Acta Gene Struct. Expression 1982, 696, 315–322; M. M. van Dyke,
P. B. Dervan, Science 1984, 225, 1122–1127; C. M. L. Low, R. K.
Olsen, M. J. Waring, FEBS Lett. 1984, 176, 414–419; K. J. Addess, J.
Feigon, Biochemistry 1994, 33, 12386–12396; K. J. Addess, J. Feigon,
Biochemistry 1994, 33, 12397–12404.
[11] E. Marco, A. Negri, F. J. Luque, F. Gago, Nucleic Acids Res. 2005,
33, 6214–6224.
[12] P. K. Chakravarty, R. K. Olsen, Tetrahedron Lett. 1978, 19, 1613–
1616; M. Shin, K. Inouye, H. Otsuka, Bull. Chem. Soc. Jpn. 1984,
57, 2203–2210.
[13] In the course of studying its mode of action, analogues with less
degree of N-methylation were synthesized. Thus, considerable atten-
tion has received the desmethylated triostin A (TANDEM), [T. L.
Ciardelli, R. K. Olsen, J. Am. Chem. Soc. 1977, 99, 2806–2807] and
modifications on it such as the bis l-Ser TANDEM [T. L. Ciardelli,
P. K. Chakravarty, R. K. Olsen, J. Am. Chem. Soc. 1978, 100, 7684–
7690] and the TANDEM with nucleobases [B. K. Lorenz, ; U. Die-
derichsen, J. Org. Chem. 2004, 69, 3917–3927; U. Diederichsen, J.
Org. Chem. 2004, 69, 3917–3927]. Furthermore, an increase in con-
formational rigidity has been shown by the aza analogue of triostin
A, azatriostin [D. L. Boger, J. K. Lee, J. Org. Chem. 2000, 65, 5996–
6000]. All these synthesis had been carried in solution.
[14] Syntheses on the solid-phase have been limited up to now to the
desmethylated analogue of triostin (TANDEM) and to its aza deriv-
ative: a) J. P. Malkinson, M. K. Anim, M. Zloh, M. Searcey, J. Org.
Chem. 2005, 70, 7654–7661; b) B. Dietrich, U. Diederichsen, Eur. J.
Org. Chem. 2005, 147–153.
and the presence of the two conformers is a result of the iso-
merization cis/trans of the NMe groups of Cys.
The in vitro activity of the two conformers of triostin A
was evaluated in three tumour cell lines (MDA-MB-231,
A549, HT29). To the best of our knowledge, this is the first
time that these two conformers have been assayed separate-
ly. In the three cell lines, the most polar conformer showed
a higher inhibitory effect on cell growth (10À7 m) than the
least polar one, which was one order of magnitude less
active.
In conclusion, we have demonstrated that an optimized
synthetic approach should allow the synthesis of any natural
product, as the two triostins A. Furthermore, the new con-
cept of conformationally restricted mobility can be key for
the synthesis of other peptides as well as other complex nat-
ural products.
Acknowledgements
This study was partially supported by CICYT (CTQ2006-03794/BQU),
the ISCIII (CIBER, nanomedicine), the Institute for Research in Bio-
medicine, and the Barcelona Science Park. J.T.P. is a Juan de la Cierva
fellow (MEC). We gratefully acknowledge PharmaMar S.L. for perform-
ing the biological tests.
[15] M. C. Khosla, R. R. Smeby, F. Bumpus, J. Am. Chem. Soc. 1972, 94,
4721–4724.
[16] K. Barlos, D. Gatos, J. Kallitsis, G. Papaphotiu, P. Sotiriu, W. Yao,
W. Schaefer, Tetrahedron Lett. 1989, 30, 3943–3946.
Keywords: conformation analysis · peptides · protecting
groups · solid-phase synthesis · triostin
[17] Previous studies on N-methylated analogues had shown that the
ester bond between the NMe-Val and the d-Ser was not stable to pi-
peridine.
[18] KociØnsky [P. J. KociØnsky, Protecting groups, 3rd edition, Thieme,
Stuttgart (Germany), 2004, pp. 29–32] describes several examples
showing that a conformation change can produce new hindrance re-
sulting in a modification of the regioselectivity or the stereospecific-
ity, but to the best of our knowledge this is the first example of the
new concept of protection.
[1] J. Shoji, K. Katagiri, J. Antibiot. 1961, 14, 335–339; H. Otsuka, J.
Shoji, Tetrahedron 1965, 21, 2931–2938; H. Otsuka, J. Shoji, K.
Kawano, Y. Kyogoku, J. Antibiot. 1976, 29, 107–110; S. Dawson,
J. P. Malkinson, D. Paumier, M. Searcey, Nat. Prod. Rep. 2007, 24,
109–126.
[2] J. Shoji, H. Otsuka, J. Antibiot. 1965, 18, 134; J. Shoji, K. Tori, H.
Otsuka, J. Org. Chem. 1965, 30, 2772–2776.
[19] Personal communication.
[20] K. Barlos, D. Papaioannou, S. Patrianakou, T. Tsegenidis, Liebigs
Ann. Chem. 1986, 1950–1955.
[21] T. Johnson, M. Quibell, Tetrahedron Lett. 1994, 35, 463–466.
[22] T. Haack, M. Mutter, Tetrahedron Lett. 1992, 33, 1589–1592.
[23] T. V. Alfredson, A. H. Maki, M. E. Adaskaveg, J.-L. Excoffier, M. J.
Waring, J. Chrom. 1990, 507, 277–292.
[24] T. L. Blake, J. R. Kalman, D. H. Williams, Tetrahedron Lett. 1977,
18, 2621–2624.
[3] A. Dell, D. H. Williams, H. R. Morris, G. A. Smith, J. Feeney,
C. G. K. Roberts, J. Am. Chem. Soc. 1975, 97, 2497–2502.
[4] H. Okada, H. Suzuki, T. Yoshimari, H. Arakawa, A. Ocurra, H.
Suda, J. Antibiot. 1994, 47, 129–135.
[5] F. Romero, F. Espliego, J. PØrez-Baz, T. García de Quesada, D.
Gravalos, F. De La Calle, J. L. Fernµndez-Puentes, J. Antibiot. 1997,
50, 734–737; J. PØrez-Baz, L. M. CaÇedo, J. L. Fernµndez-Puentes,
M. V. Silva Elipe J. Antibiot. 1997, 50, 738–741.
[6] J. Tulla-Puche, N. Bayo-Puxan, Juan A. Moreno, Andres M. Fran-
cesc, C. Cuevas, M. Alvarez, F. Albericio, J. Am. Chem. Soc. 2007,
129, 5322–5323.
[25] Kyogoku, N. Higuchi, M. Watanabe, K. Kawano, Biopolymers 1981,
20, 1959–1970.
[7] S. Matsuura, J. Antibiot. 1965, 18, 43–46.
[8] M. J. Waring, L. P. G. Wakelin, Nature 1974, 252, 653–657.
[9] J. S. Lee, M. J. Waring, Biochem. J. 1978, 173, 115–128; J. S. Lee,
M. J. Waring, Biochem. J. 1978, 173, 129–144.
Received: March 2, 2008
Published online: April 11, 2008
4478
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 4475 – 4478