Solid-phase synthesis of N Me-IB-01212, a highly N -methylated cyclic peptide

Article abstract of DOI:10.1021/ol203231q

N-Methylation of peptides is an important synthetic tool in peptide-based medicinal chemistry. Herein, an optimized strategy for solid-phase synthesis of small but highly N-methylated cyclic peptides is described. The proposed route addresses several prob

Full text of DOI:10.1021/ol203231q

ORGANIC
LETTERS
2012
Vol. 14, No. 2
612–615
Solid-Phase Synthesis of NMe-IB-01212, a
Highly N-Methylated Cyclic Peptide
Eleonora Marcucci,†,‡ Judit Tulla-Puche,*,†,‡ and Fernando Albericio*,†,‡
Institute for Research in Biomedicine, Barcelona Science Park, Baldiri Reixac 10, 08028
Barcelona, Spain, and CIBER-BBN Networking Centre on Bioengineering,
Biomaterials, and Nanomedicine, and Department of Organic Chemistry,
University of Barcelona, 08028 Barcelona, Spain
judit.tulla@irbbarcelona.org; albericio@irbbarcelona.org
Received December 4, 2011
ABSTRACT
N-Methylation of peptides is an important synthetic tool in peptide-based medicinal chemistry. Herein, an optimized strategy for solid-phase
synthesis of small but highly N-methylated cyclic peptides is described. The proposed route addresses several problems associated with the
synthesis of peptides containing several sequential N-methyl-amino acids, such as in situ N-methylation, difficulty of acylation, epimerization,
diketopiperazine formation, and stability at the NMe sites under trifluoroacetic acid exposure. The resulting NMe-IB-01212 exhibits micromolar
activity and considerable stability.
IB-01212 (Figure 1) is a cyclodepsipeptide isolated from
the marine fungi Clonostachys sp. ESNA-A009.1 The com-
pound shows micromolar activity in various tumor cell
lines, including LNCaP (prostate cancer), SK-BR-3 (breast
cancer), HT-29 (colon cancer), and HELA (cervical
cancer).2 Like thiocoraline,3 an antitumor cyclothiodepsi-
peptide belonging to the same chemical family, IB-01212
features a C2-symmetrical structure, but rather than the
heterocyclic bisintercalator chromophores of the former, it
contains N,NMe2-Leu groups. Furthermore, IB-01212
contains two additional NMe-aa residues (Leu and Phe).
Inthe course of our researchon thiocoraline, wefoundthat
replacing the thioester moieties with bridged NMe amides
confers an analogue (NMe-azathiocoraline) that is more
stable yet retains the activity of the parent compound.4 This
finding is consistent with the work of Gilon, Kessler, and co-
workers, who proposed multiple N-methylation of peptides
as a new paradigm in peptide-based medicinal chemistry.5
Synthesis of peptides featuring several sequential NMe-
amino acids is hindered by a limited number of cyclization
points in the linear intermediates, poor acylation rates, a
risk of epimerization, lability of the NMe-amide bond
under acidic conditions, and formation of C-terminal and
ꢀ
(1) (a) Cruz, L. J.; Martınez Insua, M.; Perez Baz, J.; Trujillo, M.;
Rodriguez-Mias, R. A.; Oliveira, E.; Giralt, E.; Albericio, F.; Canedo,
~
ꢀ
(4) Tulla-Puche, J.; Marcucci, E.; Prats-Alfonso, E.; Bayo-Puxan,
L. M. J. Org. Chem. 2006, 71, 3335. (b) Cruz, L. J.; Cuevas, C.; Giralt, E.;
Canedo, L. M.; Albericio, F. J. Org. Chem. 2006, 71, 3339.
(2) Cruz, L. J.; Francesch, A.; Cuevas, C.; Albericio, F. ChemMed-
N.; Albericio, F. J. Med. Chem. 2009, 52, 834.
(5) Chatterjee, J.; Gilon, C.; Hoffman, A.; Kessler, H. Acc. Chem.
Res. 2008, 41, 1331.
~
Chem 2007, 2, 1076.
(3) (a) Romero, F.; Espliego, F.; Perez-Baz, J.; Garcıa de Quesada,
(6) (a) McDermott, J. R.; Benoiton, N. L. Can. J. Chem. 1973,
51, 2555. (b) McDermott, J. R.; Benoiton, N. L. Can. J. Chem. 1973,
51, 2562. (c) Humphrey, J. M.; Chamberlin, A. R. Chem. Rev. 1997, 97,
2243. (d) Teixido, M; Albericio, F.; Giralt, E. J. Peptide Res. 2005, 65,
153–166.
ꢀ
ꢀ
T.; Gravalos, D.; De la Calle, F.; Fernandez-Puentes, J. L. J. Antibiot.
ꢀ
ꢀ
~
ꢀ
1997, 50, 734. (b) Perez-Baz, J.; Canedo, L. M.; Fernandez-Puentes,
J. L.; Silva Elipe, M. V. J. Antibiot. 1997, 50, 738.
r
10.1021/ol203231q
Published on Web 12/22/2011
2011 American Chemical Society

internal diketopiperazines (DKPs).6 Moreover, the lack of
commercial availability for these NMe-amino acids means
that there is a requirement for in situ methylation. Thus,
the strategies reported here for the synthesis of NMe-
IB-01212 could be extrapolated to the synthesis of other
highly methylated peptides.
The synthesis of the natural IB-01212 was accomplished
via a 3 þ 3 fragment coupling on solid-phase through ester
formation because the linear strategy gave very low yields
due to the instability of the ester bond.7
Figure 2. Starting points for the synthesis of NMe-IB-01212.
each starting point (Figure 2b,c) led to substantial degrees
of epimerization,11 thereby making it impossible to con-
tinue in this direction (see the Supporting Information).
In contrast, the successful strategy to reach NMe-IB-
01212 is depicted in Scheme 1, and it involves a com-
pletely stepwise approach. The starting point for the syn-
thesis was the Fmoc-NMe-Phe-OH residue, which was
anchored onto 2-CTC resin (3). Here, we have faced the
challenge of coupling Fmoc-NMe-Leu-OH as the second
amino acid.12 In a first attempt using 1- [bis(dimethylamino)-
methylene]-1H-1,2,3-triazolo-[4,5-b]pyridinium hexafluoro-
phosphate 3-oxide (HATU), 1-hydroxy-7-azabenzotriazole
(HOAt), and diisopropylethylamine (DIEA) in DMF as
coupling reagent, significant deletion was observed, reaching
40%.13 DKP formation was also detected. To further opti-
mize the preparation of the initial tripeptide, several changes
were introduced: the synthesis was performed on 2-CTC
resin, but the loading was lowered to ca. 0.5 mmol/g in order
to favor coupling of NMe-Leu; second, the less-hindered
Alloc-NMe-Leu-OH was used with the double objective of
facilitating the coupling and simultaneously preventing DKP
formation due to the rather neutral conditions to remove the
Alloc group; and finally, the coupling system used was
changed to (1-cyano-2-ethoxy-2-oxoethylidenaminooxy) di-
methylamino-morpholino-carbenium hexafluorophosphate
Figure 1. Structure of natural IB-01212 and NMe-IB-01212.
For the synthesis of NMe-IB-01212, as no clear point for
cyclization exists (no Gly or Pro at the C-terminus is present,
for example, that would avoid epimerization), the three dif-
ferent options were initially studied (Figure 2).
Strategy (a) was not pursued because we had previously
observed that N-methylation of the side chain of the Dap
residue, when it is directly anchored to the 2-chlorotrityl
(2-CTC) resin,8 results in a severe loss of peptide.4 Strategy
(b) has the advantage that the 6-membered ring DKP can-
not be formed because the second residue, Dap, is elongated
through its side chain; but, on the other hand, the cycliza-
tion between the bulky NMe-Phe and the β-branched
NMe-Leu residue may prove complicated. In strategy (c),
NMe-Phe has the drawback of DKP formation due to the
consecutive NMe-amino acids, which was already observed
during the earlier synthesis of natural IB-01212.9
The first attempts at preparing NMe-IB-01212 entaileda
fragment approach10 to avoid a double solid-phase N-meth-
ylation step, which could affect the purity of the final chain,
and to circumvent the coupling of so many consecutive
NMe-amino acids. Nevertheless, fragment couplings at
~
(7) Cruz, L. J.; Cuevas, C.; Canedo, L. M.; Giralt, E.; Albericio, F.
J. Org. Chem. 2006, 71, 3339.
(11) With NMe-Leu (b) at the C-terminus, the epimerization ratio
was 50:50, whereas with NMe-Leu (c) at the same position, the ratio was
65:35.
(12) In previous work with the 2-CTC resin, which contains bulky
trityl groups, we observed that when the first amino acid coupled to the
resin is sterically hindered, the coupling of the second amino acid is very
difficult.
(13) To test a less hindered solid support, the same tripeptide was
elongated on Wang resin; however, the deletion of NMe-Leu was still
present. Moreover and as expected, a greater amount of DKP was
formed.
(8) Barlos, K.; Gatos, D.; Kallitsis, J.; Papaphotiu, G.; Sotiriu, P.;
Yao, W.; Schaefer, W. Tetrahedron Lett. 1989, 30, 3943.
(9) Although our group (Chiva, C.; Vilaseca, M.; Giralt, E.
Albericio, F. J. Peptide Sci. 1999, 5, 131.) and others (Rovero, P.;
Vigano, S.; Pegoraro, S.; Quartara, L. Lett. Peptide Sci. 1996, 2, 319.)
have shown that CTC resins avoid DKP formation during synthesis of
C-terminal Pro peptides, CTC resin can not prevent DKP formation in
sequences that are highly prone to this side reaction.
(10) The tripeptide, which constitutes half of the molecule, was
assembled on solid-phase. At this point, one-third of the resin-bound
peptide was deprotected, and the remaining two-thirds was cleaved from
the resin. After lyophilization, the 3 þ 3 fragment coupling was carried
out on solid phase.
ꢀ
(14) Subiros-Funosas, R.; Prohens, R.; Barbas, R.; El-Faham, A.;
Albericio, F. Chem.;Eur. J. 2009, 15, 9394.
Org. Lett., Vol. 14, No. 2, 2012
613

Scheme 1. Final Approach to Assemble NMe-IB-01212
(COMU),14 ethyl (hydroxyimino)cyanoacetate (oxyma),15
and DIEA in DMF. The excellent solubility in DMF of this
coupling system over HATU and HOAt allows the use of
several equivalents of amino acid in minimal solvent, result-
ing in a far better coupling yield. Therefore, after perfoming
the coupling with Alloc-NMe-Leu-OH (3 equiv), COMU
(3 equiv), oxyma (3 equiv), and DIEA (6 equiv) for one hour
and stirring one further hour after adding COMU (1 equiv)
and oxyma 1 (equiv), the protected dipeptide (4) was ob-
tained in 93% purity.
After removing the Alloc group from NMe-Leu with
Pd(PPh3)4 and PhSiH3 in DCM, the Dap residue was intro-
duced, using Fmoc-L-Dap(Alloc)-OH (4 equiv) in the pre-
sence of COMU (4 equiv), oxyma (4 equiv), and DIEA
(8 equiv) in DMF for 1 h (5).16 After removing the Fmoc
group with piperidine, N,NMe2-Leu was coupled over the
free R-amino functionality of the Dap group with HATU/
HOAt/DIEA in DMF for 1 h (6).
deprotected (removal of 2-NBS) by treating the resin with
2-mercaptoethanol and 1,8-diazabicyclo[5.4.0]undec-7-ene
(DBU) in DMF. With the tetrapeptide (8) in hand, the
same set of reactions was repeated in order to achieve the
complete linear skeleton of NMe-IB-01212. The octapep-
tide (9) was cleaved from the resin with a TFAꢀDCM
solution (2:98, 5 ꢁ 1 min) and then lyophilized. The crude
peptide was obtained in 52% yield with a purity of 56% as
shown by HPLC analysis.
In earlier work using the hexapeptide {[Boc-L-Dap
(Me&1)-NMe-Leu-NMe-Phe&2][Boc-L-Dap(Me&2)-
NMe-Leu-NMe-Phe&1]}, various cyclization conditions
were investigated that would imply minimal epimeriza-
tion (Table 1).
Table 1. Cyclization Studies
The Alloc group was again removed using the same
conditions described above, and the side chain of the Dap
residue was N-methylated on solid-phase using Kessler’s
conditions:17 first, the free amino functionality was re-
protected with the 2-nitrobenzenesulfonyl (2-NBS) group
(7), then methylated via a Mitsunobu reaction [PPh3,
MeOH, diisopropyl azodicarboxylate (DIAD)], and finally
coupling
reagent
diastereomeric
entry
base
solvent pH conv ratio (D:L)
1
2
3
4
PyBOP/HOAt DIEA DMFꢀDCM 8 low
PyBOP/HOAt collidine DMFꢀDCM 8 low
15:85
67:33
13:87
PyBOP/HOAt DIEA DMF
8 high
DIPCDI/HOAt
DMFꢀDCM 7
Cyclization under neutral conditions (entry 4) did not
take place and only proceeded when performed at pH 8.
The other three experiments were designed to examine the
influence of the solvent and base on the epimerization
side-reaction. When using the milder base 2,4,6-collidine
(entry 2), low conversion was observed, whereas a high
degree of epimerization was detected. The combination of
DIEA as base and DMF as solvent gave the highest con-
version and the lowest epimerization percentage (13%,
entry 3). Therefore, the octapeptide precursor was dissolved
ꢀ
(15) El-Faham, A.; Subiros-Funosas, R.; Prohens, R.; Albericio, F.
Chem.;Eur. J. 2009, 15, 9404.
(16) Alternatively, when we used Boc-L-Dap(Fmoc)-OH to construct
the hexapeptide skeleton and then removed the Boc groups in solution
after cyclization to later introduce the N,NMe2-Leu residues, use of
TFAꢀDCM (1:1) caused the peptide to break at its NMe sites (as
evidenced by a complex chromatogram). Other conditions that were
tested, such as lowering the percentage of TFA, minimizing the time of
exposure, or using HCl instead of TFA, did not offer any improve-
ments; therefore, this strategy was abandoned (see the Supporting
Information).
(17) Biron, E.; Chatterjee, J.; Kessler, H. J. Peptide Sci. 2006, 12, 213
and references cited therein.
614
Org. Lett., Vol. 14, No. 2, 2012

In summary, a highly N-methylated cyclopeptide, which
showed micromolar activity in four cell lines and consider-
able stability in human serum, was assembled by optimized
stepwise solid-phase mode.
The synthetic strategy described herein could be ex-
ploited for the preparation of other highly N-methylated
cyclic peptides.5,17,18 The principal features to consider are
(i) the use of anorthogonalprotecting groupschemewould
avoid the use of Boc/tBu groups, which can cause breakage
at NMe sites;19 (ii) a low loading of the resin is required to
obtain satisfactory purities of the linear chain; (iii) the use of
the Alloc group to mask the R-amino functionality of the
second amino acid provides higher yields in the coupling step
and minimizes DKP formation; and (iv) COMU and oxyma
allow carrying out the demanding couplings with a higher
concentration when compared to other coupling reagents
and additives, thus favoring high-yielding couplings.
Figure 3. Analytical HPLC of pure NMe-IB-01212. Linear
gradient from 30:70 to 50:50 (0.036% TFA in ACN/0.045%
TFA in H2O) in 8 min.
in DMF and cyclized with PyBOP and HOAt by main-
taining the pH at 8 with the addition of DIEA. The
cyclization proceeded smoothly despite the rigidity im-
posed by the high number of N-methyl amino acids and
was complete after 3 h. A low degree of epimerization
(9%) was observed.
The crude cyclic peptide was purified by semipreparative
HPLC (linear gradient from 30:70 to 40:60 (0.036% TFA
in ACN/0.045% TFA in H2O) to afford the pure NMe-IB-
01212 (99% purity, 4% overall yield, Figure 3). The pure
peptide showed cytotoxic activity in four cell lines (HT-29
5.8 μM; SK-BR-3 15μM; A-549 23μM; HeLa 11.3 μM), as
well as high stability in human serum, with a half-life time
of 60 h (see the Supporting Information), much higher
than that of NMe-azathiocoraline.
ꢀ
Acknowledgment. We thank Drs. Jose Pastor and
Miriam Royo from the Barcelona Science Park for kindly
performing the biological assays. This work was partially
supported by CICYT (CTQ2009-07758), the Generalitat
de Catalunya (2009SGR 1024), the Institute for Research
in Biomedicine, and the Barcelona Science Park. E.M. and
J.T.-P. thank MICINN for a Torres y Quevedo and a Juan
de la Cierva contract, respectively.
Supporting Information Available. Experimental pro-
cedures, characterization, data, and cytotoxic and stabi-
lity assays. This material is available free of charge via the
Internet at http://pubs.acs.org
(18) (a) Plaza, A.; Bifulco, G.; Masullo, M.; Lloyd, J. R.; Keffer, J. L.;
Colin, P. L.; Hooper, J. N. A.; Bell, L. J.; Bewley, C. A. J. Org. Chem.
2010, 75, 4344. (b) Mas-Moruno, C.; Beck, J. G.; Doedens, L.; Frank,
A. O.; Marinelli, L.; Cosconati, S.; Novellino, E.; Kessler, H. Angew.
Chem. 2011, 50, 9496.
(19) Peptide skeletons that contain NMe-amino acids tend to be
unstable, depending of the number of consecutive NMe-amino acids
and/or the rigidity of the peptide structure, as demostrated in this case,
whereby a relatively smaller cycle (compared with the thiocoraline
family) cannot tolerate even a moderate amount of acid.
Org. Lett., Vol. 14, No. 2, 2012
615