Enzyme-labile protecting groups for the synthesis of natural products: Solid-phase synthesis of thiocoraline

Article abstract of DOI:10.1002/anie.201301708

Another (orthogonal) dimension: The solid-phase synthesis of thiocoraline was accomplished for the first time by a combined approach involving chemical and enzymatic methods. One-pot cleavage of the phenylacetamidomethyl protecting group using immobilized penicillinG acylase enzyme (see picture) and disulfide formation are the key steps of the synthetic strategy. Copyright

Full text of DOI:10.1002/anie.201301708

Angewandte
Chemie
DOI: 10.1002/anie.201301708
Natural Products
Enzyme-Labile Protecting Groups for the Synthesis of Natural
Products: Solid-Phase Synthesis of Thiocoraline**
Judit Tulla-Puche,* Miriam Gꢀngora-Benꢁtez, Nfflria Bayꢀ-Puxan, AndrØs M. Francesch,
Carmen Cuevas, and Fernando Albericio*
Nature creates extremely complex molecules which com-
monly present pharmaceutical properties of interest.[1]
A
particularly intriguing class of these molecules are cyclic
peptides isolated from marine species.[2] The development of
these peptides as drug candidates is hampered by the
difficulty in isolating sufficient amounts of material from
the natural sources. Despite the huge development of
chemical methods during the past decade, the synthesis of
this kind of compound remains a challenge, with numerous
reaction and purification steps. In the case of peptides, many
of these steps correspond to protection/deprotection steps,
and limitations in a synthetic strategy often arise as a result of
the incompatibility of protecting groups. Protecting groups
should preferably be orthogonal,[3] meaning that they can be
removed in any order and in the presence of the other
protecting groups. Such orthogonal removal is required in the
case of cyclic peptides, where an intramolecular reaction
takes place once the peptide chain has been constructed. It is
best to make these kinds of molecules using solid-phase
synthesis, although very often it is necessary to carry out a few
solution reactions at the end of the synthesis. Enzyme-labile
protecting groups add a new orthogonal level to the synthesis
of complex molecules for which strategies based only on
chemical methods have shown limitations. This is exemplified
herein by the first solid-phase synthesis of the complex
cyclothiodepsipeptide thiocoraline[4,5] (Figure 1). Isolated in
Figure 1. Chemical structure of thiocoraline.
1997 off the coast of Mozambique, this unique and potent
antitumor agent has several features that make its structure
extremely complex, namely a sequence rich in Cys, the
presence of consecutive N-methylamino acids, and a bicyclic
structure formed by a disulfide bridge flanked by two
thioester moieties.
In addition to protection of the a-amino group, the
presence of six Cys residues (four of them N-methylated and
the remaining two in the d configuration, which mask a DNA
bisintercalating chromophore) requires different protecting
groups for the thiol groups and mild removal conditions for all
of them. In this regard, the phenylacetamidomethyl[6]
(Phacm) group is suitable, because it can be cleaved by the
enzyme penicillin G acylase (PGA). Therefore, an orches-
trated scheme of protecting groups becomes the cornerstone
for the synthesis of thiocoraline.[7]
Although attempts were made to apply three distinct
strategies developed by our group to other analogues of
thiocoraline that do not have thioester moieties, our efforts
were unsuccessful.[8] The formation of these extremely
delicate functional groups needs to be postponed until the
final steps of the synthesis, and usually thioester formation is
incompatible with the formation of the disulfide bridge. As
a result, the NMe-Cys(Me) residue, which is the C-carboxyl
component of the thioester, was chosen as a starting point for
the synthesis. One drawback of this approach was the b-
elimination side reaction that this residue undergoes when
placed at the anchoring position and piperidine is used to
remove the 9-fluorenylmethoxycarbonyl (Fmoc) group. This
problem was overcome using the milder allyl-based chemistry
during most of the elongation of the peptide chain. Synthesis
of the allyloxycarbonyl (Alloc)-protected N-methylated Cys
was accomplished starting from Boc-Cys(Me)-OH and Boc-
Cys(trityl,Trt)-OH, by carrying out an N-methylation reaction
with NaH and MeI, then introducing the Phacm moiety in the
[*] Dr. J. Tulla-Puche, Dr. M. Gꢀngora-Benꢁtez, Dr. N. Bayꢀ-Puxan,
Prof. F. Albericio
Institute for Research in Biomedicine Barcelona, CIBER-BBN
Baldiri Reixac 10, 08028 Barcelona (Spain)
E-mail: judit.tulla@irbbarcelona.org
albericio@irbbarcelona.org
Prof. F. Albericio
Department of Organic Chemistry, University of Barcelona
Martꢁ i Franquꢂs 1, 08028 Barcelona (Spain)
Prof. F. Albericio
University of KwaZulu Natal
4001-Durban (South Africa)
Dr. A. M. Francesch, Dr. C. Cuevas
PharmaMar S. A.
Avda. de los Reyes 1, 28770 Colmenar Viejo, (Spain)
[**] This study was funded by the CICYT (CTQ2012-30930), the
Generalitat de Catalunya (2009SGR1024), and the Institute for
Research in Biomedicine Barcelona (IRB Barcelona). We thank Dr.
M. Gairꢁ and Dr. M. Royo from the Barcelona Science Park for NMR
technical support and for helpful discussions, respectively, and Dr.
T. Bruckdorfer (Iris Biotech) for encouraging us to use Phacm.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201301708.
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
Scheme 1. Synthetic strategy for the synthesis of thiocoraline; a) Alloc-NMe-Cys(Me)-OH, DIEA, CH2Cl2; MeOH; b) [Pd(PPh3)4], PhSiH3, CH2Cl2;
c) Alloc-NMe-Cys(Phacm)-OH, HATU, HOAt, DIEA, DMF; d) Alloc-Gly-OH, HATU, HOAt, DIEA, DMF; e) Fmoc-d-Cys(Trt)-OH, DIPCDI, HOBt,
DMF; f) piperidine-DMF (1:4); g) 3-hydroxyquinaldic acid, DIPCDI, HOBt, CH2Cl2; h) TFA/Me2S/TIS/CH2Cl2 (2:0.5:0.5:97); i) TFA/TIS/CH2Cl2
(10:2.5:87.5); j) DIPCDI, HOAt, DMF/CH2Cl2; k) immobilized PGA enzyme, H2O/DMSO (9:1), pH 6.7.
next step, and finally incorporating the Alloc protecting
group.[9] Following this strategy, Alloc-NMe-Cys(Me)-OH
and Alloc-NMe-Cys(Phacm)-OH were obtained in good
purities (> 95%) and were used without further purification.
The use of 2-chlorotrityl chloride (2-CTC) resin[10] as solid
support was compulsory, because the two first residues are
NMe-amino acids, which are extremely prone to producing
diketopiperazine (DKP).[11] Therefore, the synthesis began by
anchoring Alloc-NMe-Cys(Me)-OH onto the 2-CTC resin
(Scheme 1).
subsequent formation of the thioester. Therefore, after
removing the Alloc group of Gly, Fmoc-d-Cys(Trt)-OH was
easily coupled using DIPCDI/HOBt in DMF for one hour. To
minimize dehydration of the NMe-Cys(Me) residue, the
Fmoc group was removed with two treatments of piperidine/
N,N-dimethylformamide (DMF) (1:4), each lasting two
minutes. Coupling of 3-hydroxyquinaldic acid was achieved
under mild conditions [N,N’-diisopropylcarbodiimide
(DIPCDI)/1-hydroxybenzotriazole (HOBt)] and few equiv-
alents (1.1 equiv) to prevent overacylation at the free hydroxy
group. After one hour, a ninhydrin test showed completion of
the coupling. With the tetrapeptide of thiocoraline fully
assembled, we detached it from the solid support. Owing to
a large amount of oxidation on the Cys residues, especially
NMe-Cys(Me), alternative methods to the standard cleavage
conditions for 2-CTC resin [trifluoroacetic acid (TFA)/
CH2Cl2, 1:99], were examined. In this regard, cleavage of
the tetrapeptide under mild reductive conditions [TFA/Me2S/
triisopropylsilane (TIS)/CH2Cl2, 2:0.5:0.5:97] afforded the
protected tetrapeptide in 91% purity as shown by analytical
HPLC (see the Supporting Information).
Next, two possible methods were devised (Figure 2). The
first consisted of forming an intermolecular disulfide dimer,
which led to synthesis of the analogue oxathiocoraline.[8b]
However, although the formation of the dimer was complete,
as indicated by HPLC and HPLC-MS, removal of the Trt
groups to perform the double cyclization destabilized the
dimer because of the presence of free thiol groups. This
strategy was thus abandoned.
The second option, which a priori was more difficult to
handle, involved removing the Trt groups and then perform-
ing the double thioesterification in a concentrated (18 mm)
solution. To this end, the Trt groups were removed with
a TFA/TIS/CH2Cl2 (10:2.5:87.5) solution at 48C for ten
minutes, and then co-evaporated with TBME at low temper-
Following removal of the Alloc group by treatment with
[Pd(PPh3)4] and PhSiH3, Alloc-NMe-Cys(Phacm)-OH was
coupled using 1-[bis(dimethylamino)-methylene]-1H-1,2,3-
triazolo-[4,5-b]pyridinium hexafluorophosphate 3-oxide
(HATU)/1-hydroxy-7-azabenzotriazole (HOAt)/diisopropy-
lethylamine (DIEA) for one hour. The coupling was repeated
to ensure complete reaction of the amino acid. Although 2-
CTC had been used, removal of the protecting group was
accompanied by DKP formation. We therefore optimized the
method to minimize this side reaction, which consisted of
performing the coupling under almost neutral conditions by
adding fewer equivalents of base (4.9 equiv) with respect to
the incoming amino acid and coupling reagents (5.0 equiv).
At this point, the purity (89%) and the absorbance spectra of
the protected tripeptide showed no indication of DKP
formation.[12] To introduce the next residue (d-Cys), which
has a 3-hydroxyquinaldic unit, either Boc or Fmoc could be
used as the protecting group for the a-amino group of the Cys
residue to later introduce the heterocycle. There is an
advantage in the use of the Fmoc group in that the hetero-
cycle can be introduced during the solid-phase synthesis,
where coupling proceeds faster than in solution[13] and gives
better yields, even when using few equivalents of the precious
chromophore. A Trt group is suitable to protect the thiol,
because it can be removed under mild conditions and allows
2
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
mation). Finally, the
in vitro activity of the
synthetic compound in
four human tumor cell
lines showed very simi-
lar activity to thiocora-
line isolated from its
natural
source
(Table 1).
In summary, the
enzyme-labile Phacm
protecting group has
proven to be key in
synthesizing the com-
plex antitumor agent,
thiocoraline. Addition-
ally, all synthetic steps
were carried out using
mild
conditions
to
remove the protecting
groups. The elongation
Figure 2. Possible routes from the tetrapeptide: dimer formation and double thioesterification.
ature to remove the remaining acid. The unprotected
tetrapeptide was dissolved in CH2Cl2 and the solution was
cooled to 48C. DIEAwas added until the pH was neutral. The
addition of DIPCDI and HOAt started the cyclization
reaction,[14] which was monitored by HPLC. After two
hours, HPLC showed that the reaction had gone to comple-
tion, and no monomeric thioesterification was detected. The
crude solution was then washed twice with H2O and pre-
purified with an ISCO instrument equipped with a reversed-
phase C18 column to remove coupling reagents and other
impurities. The cyclic thiodepsipeptide was obtained in 85%
purity. To remove the Phacm protecting groups and form the
disulfide bond, a portion of this crude product (5 mg) was
dissolved in dimethylsulfoxide (DMSO; 1.5 mL), and H2O
was added to a final volume of 15 mL. The immobilized PGA
enzyme was then added, resulting in a final pH of 6.7, and the
suspension was incubated at 378C and monitored by HPLC.
After 56 hours, the immobilized enzyme was filtered and the
reaction solution was lyophilized.[15] Final purification of
Figure 3. Analytical HPLC of a) natural thiocoraline; b) synthetic thio-
coraline; and c) co-elution of natural and synthetic thiocoraline. HPLC
conditions: linear gradient of ACN (0.036% TFA)/H2O (0.045% TFA)
thiocoraline was performed using semipreparative HPLC.
Co-elution with the natural product using a smooth analytical
HPLC gradient (Figure 3), HR-ESMS, and NMR confirmed
the identity of the synthetic thiocoraline (Supporting Infor-
62:38 to 65:35 over 45 min at a flow rate of 1 mLminÀ1
.
of the monomer was done on a solid-phase support using 2-
CTC resin and Alloc as the protecting group for the a-amino
group of the first three residues, which prevented two side-
reactions: b-elimination to form a dehydroAla residue and
DKP formation. Macrocyclization was carried out success-
fully through the thioester without racemization, taking
advantage of the high reactivity of the thiols, in a double
reaction. This unprecedented reaction could be favored by
the presence of the two N-methyl residues, which tend to
favor the cis-conformation of the peptide bond. Finally, the
Phacm groups were removed with concomitant formation of
Table 1: In vitro results for natural and synthetic thiocoraline.
Cell line
Breast MDA- NSCLC[a] Pancreas
Colon
HT-29
MB-231
A549
PSN1
[b]
GI50
2.07E-9
6.22E-9
2.42E-8
9.50E-8
4.29E-9
1.90E-8
9.50E-8
1.47E-9
4.23E-9
1.30E-8
1.56E-9
3.63E-9
9.50E-9
4.15E-8
2.68E-8
1.38E-7
1.90E-9
1.73E-8
4.06E-7
Natural
thiocoraline
TGI[c] 9.50E-9
[d]
LC50 3.46E-8
GI50
TGI
LC50 1.73E-8
1.04E-9
4.75E-9
Synthetic
thiocoraline
[a] NSCLC=non-small-cell lung cancer. [b] GI50 =growth inhibition 50%.
[c] TGI=total growth inhibition. [d] LC50 =lethal concentration 50%.
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
Ichikawa, W. C. Tse, M. P. Hedrick, Q. Jin, J. Am. Chem. Soc.
2001, 123, 561 – 568.
the disulfide bridge to give the synthetically difficult thiocora-
line in just a few steps and with only two purification steps.
[8] Strategies developed for azathiocoraline, oxathiocoraline, or
NMe-azathiocoraline could not be applied to the synthesis of
thiocoraline: a) N. Bayꢃ-Puxan, A. Fernꢂndez, J. Tulla-Puche, E.
Riego, C. Cuevas, M. ꢄlvarez, F. Albericio, Chem. Eur. J. 2006,
12, 9001 – 9009; b) J. Tulla-Puche, N. Bayꢃ-Puxan, J. A. Moreno,
A. M. Francesch, C. Cuevas, M. ꢄlvarez, F. Albericio, J. Am.
Chem. Soc. 2007, 129, 5322 – 5323; c) J. Tulla-Puche, E. Marcucci,
E. Prats-Alfonso, N. Bayꢃ-Puxan, F. Albericio, J. Med. Chem.
2009, 52, 834 – 839.
[9] E. Marcucci, N. Bayꢃ-Puxan, J. Tulla-Puche, J. Spengler, F.
Albericio, J. Comb. Chem. 2008, 10, 69 – 78.
[10] K. Barlos, D. Gatos, J. Kallitsis, G. Papaphotiu, P. Sotiriu, W. Yao,
W. Schaefer, Tetrahedron Lett. 1989, 30, 3943 – 3946.
[11] M. C. Khosla, R. R. Smeby, F. M. Bumpus, J. Am. Chem. Soc.
1972, 94, 4721 – 4724.
Received: February 27, 2013
Published online: && &&, &&&&
Keywords: antitumor agents · cysteine · peptides ·
.
protecting groups · thiocoraline
[1] B. Mishra, V. K. Tiwari, Eur. J. Med. Chem. 2011, 46, 4769 – 4807.
[2] G.-M. Suarez-Jimenez, A. Burgos-Hernandez, J.-M. Ezquerra-
Brauer, Mar. Drugs 2012, 10, 963 – 986.
[3] a) G. Barany, R. B. Merrifield, J. Am. Chem. Soc. 1977, 99, 7363 –
7365; b) G. Barany, F. Albericio, J. Am. Chem. Soc. 1985, 107,
4936 – 4942.
[4] a) F. Romero, F. Espliego, J. Pꢀrez Baz, T. Garcꢁa de Quesada,
D. Gravalos, F. De La Calle, J. L. Fernꢂndez-Puentes, J. Antibiot.
1997, 50, 734 – 737; b) J. Pꢀrez Baz, L. M. Canedo, J. L. Fernꢂn-
dez Puentes, M. V. Silva Elipe, J. Antibiot. 1997, 50, 738 – 741.
[5] A. Negri, E. Marco, V. Garcꢁa-Hernꢂndez, A. Domingo, A. L.
Llamas-Saiz, S. Porto-Sandꢂ, R. Riguera, W. Laine, M.-H.
David-Cordonnier, C. Bailly, L. F. Garcꢁa-Fernꢂndez, J. J.
Vaquero, F. Gago, J. Med. Chem. 2007, 50, 3322 – 3333.
[6] a) M. Royo, J. Alsina, E. Giralt, U. Slomcyznska, F. Albericio, J.
Chem. Soc. Perkin Trans. 1 1995, 1095 – 1102; b) M. Gꢃngora-
Benꢁtez, A. Basso, T. Bruckdorfer, M. Royo, J. Tulla-Puche, F.
Albericio, Chem. Eur. J. 2012, 18, 16166 – 16176.
[12] The same strategy carried out with the less hindered acetami-
domethyl (Acm) group showed some diketopiperazine forma-
tion.
[13] Coupling of the heterocycle in solution may take up to four days.
[14] The low temperature and DIPCDI in the presence of HOAt in
a rather apolar solvent such as CH2Cl2 accompanied by high
efficiency of the double macrothiolactamization assured the
absence of racemization.
[15] The use of the Acm instead of Phacm in the diBoc-cyclic
analogue was tested, but the I2 treatment led to total destruction
of the compound. Interestingly, similar results were found by
Boger et al.[7] in their synthesis trials.
[7] An elegant synthesis in solution of thiocoraline with 18 linear
steps and 12 purifications was developed: D. L. Boger, S.
4
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Communications
Natural Products
Another (orthogonal) dimension: The
solid-phase synthesis of thiocoraline was
accomplished for the first time by a com-
bined approach involving chemical and
enzymatic methods. One-pot cleavage of
the phenylacetamidomethyl protecting
group using immobilized penicillin G
acylase enzyme (see picture) and disul-
fide formation are the key steps of the
synthetic strategy.
J. Tulla-Puche,* M. Gꢁngora-Benꢂtez,
N. Bayꢁ-Puxan, A. M. Francesch,
C. Cuevas, F. Albericio*
&&&& — &&&&
Enzyme-Labile Protecting Groups for the
Synthesis of Natural Products: Solid-
Phase Synthesis of Thiocoraline
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!