Rescuing biological activity from synthetic phakellistatin 19

Article abstract of DOI:10.1021/jm401520x

Phakellistatins is one of the families of Pro-rich cyclic peptides whose synthetic counterparts have revealed cytotoxicities that differ greatly from those displayed by their corresponding natural ones. This is also the case of the last member isolated fr

Full text of DOI:10.1021/jm401520x

Article
pubs.acs.org/jmc
Rescuing Biological Activity from Synthetic Phakellistatin 19
Marta Pelay-Gimeno,†,‡ Alessandra Meli,†,§ Judit Tulla-Puche,*,†,‡ and Fernando Albericio*,†,‡,∥,⊥
†Institute for Research in BiomedicineBarcelona, Baldiri Reixac 10, 08028 Barcelona, Spain
‡CIBER-BBN, Networking Centre for on Bioengineering, Biomaterials, and Nanomedicine, Baldiri Reixac 10, 08028 Barcelona, Spain
§Universita
̀
degli Studi di Salerno, Via Giovanni Paolo II, 132-84084, Fisciano, Salerno, Italy
∥Department of Organic Chemistry, University of Barcelona, Martí i Franques
̀
1, 08028 Barcelona, Spain
⊥University of KwaZulu Natal, 4001 Durban, South Africa
S
* Supporting Information
ABSTRACT: Phakellistatins is one of the families of Pro-rich cyclic peptides whose synthetic counterparts have revealed
cytotoxicities that differ greatly from those displayed by their corresponding natural ones. This is also the case of the last member
isolated from this family, phakellistatin19, an octacyclopeptide containing three Pro moieties and a high percentage of apolar
residues. Exhaustive NMR studies on the synthetic and natural phakellistatin 19 have been performed in order to find a plausible
explanation for this intriguing behavior. Moreover, taking advantage of phakellistatin’s framework, analogues with different cis/
trans geometry at the key prolyl peptide bonds were designed, covering a promising conformational space that could not be
reached by the natural peptide. By introduction of proline surrogates (ΨMe,Mepro residues) in phakellistatin 19, which effectively
increases the percentage of cis conformation in the final peptides, this translates into enhanced biological activity, therefore
“rescuing” an otherwise inactive cyclopeptide.
INTRODUCTION
■
Several families of proline-rich peptides isolated from marine
sponges and displaying significant cytotoxicity, such as
hymenamides,1 stylopeptides,2 axinellins,3 axinastatins,4 and
phakellistatins,5 have been isolated. All of these peptides share
a number of structural features. In general, they comprise
homodetic hepta- or octapeptides with an unusual percentage of
Pro moieties, with high contents of apolar amino acids, including
one or two aromatic residues, and with a significant structural
analogy.
Nineteen cyclopeptides have been isolated so far from marine
sponges of the genus Phakellia. They all consist of seven to 10
amino acids, including at least one Pro moiety, most of them
having more than one. Of all the peptides described, four
comprise the distinctive Pro-Pro track,6 which represents a
considerable synthetic challenge.
Phakellistatin 19 (Figure 1) is an octacyclopeptide containing
three Pro moieties and a high percentage of apolar residues,
including a Leu, an Ile, and a Phe residue. Its amino acid sequence
greatly resembles that of phakellistatin 10 7 but with one single
Figure 1. Chemical structure of phakellistatin 19.
modification; the Val moiety is replaced by a Phe residue.
Biological evaluation of natural phakellistatin 19 showed
promising cytotoxic and antimitotic activity (see Table 2).
Received: September 30, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jm401520x | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Scheme 1. Synthetic Scheme for Phakellistatin 19
Surprising behavior has been associated with phakellistatins
and other proline-rich cyclopeptides such as axinastins8 and
stylopeptides.2 Thus, after chemical and spectral validation by
means of nuclear magnetic resonance (NMR), high performance
liquid chromatography (HPLC), high-resolution mass spec-
trometry (HRMS), and Marfey’s techniques, biological evalua-
tion of synthetic peptides has revealed cytotoxicities that differ
greatly from those displayed by their natural counterparts. This
phenomenon has been widely reported by several groups
working in this field and represents a scientific puzzle.2,8,9 This
is also the case of our synthetic phakellistatin 19.
formation and to perform the “head-to-tail” cyclization in
solution.
As the macrolactamization step poses the biggest synthetic
challenge of an all-L-cyclopeptide synthesis, two cyclization/
starting points were evaluated, namely, Leu(C)-(N)Thr and
Pro(C)-(N)Leu. The first amide bond involved a β-branched
residue, while the second linkage had a Pro at the C-terminus,
which minimized racemization during cyclization but increased
the risk of DKP formation during the assembly of the linear
peptide. The two approaches were carefully examined. Use of the
Pro-Leu linkage as the cyclization point rendered the desired
product with overall better yields (Scheme 1).
1-[Bis(dimethylamino)methylene]-1H-benzotriazolium hexa-
fluorophosphate 3-oxide (HBTU)/1-hydroxybenzotriazole
(HOBt)/N,N-diisopropylethylamine (DIEA) (HBTU/HOBt/
DIEA) was used as the coupling system to form the amide bonds.
The use of 4 equiv of aa during 1 h guaranteed quantitative
coupling in all cases. Fmoc group removal was accomplished by
treatment with piperidine−N,N-dimethylformamide (DMF)
(1:4) (2 × 1 min; 2 × 5 min). After incorporation of the third
residue, Fmoc quantification proved the absence of DKP
formation.
RESULTS AND DISCUSSION
■
To date, two main hypotheses have been proposed to explain this
biological incongruity. On the one hand, a number of authors
argue that the presence of trace amounts of a highly cytotoxic
contaminant that binds noncovalently to natural phakellistatins
(or peptides from other families) would account for the
biological activity of these compounds.9 This cytotoxic agent
would be present in a low percentage and would thus prevent its
detection by NMR spectroscopy. In this sense, Pettit et al.9c
proved that, indeed, nondetectable NMR contamination can
cause biological activity. On the other hand, a conformational
issue caused by the presence of a high percentage of Pro residues
in quite a small cyclopeptide may also explain this intriguing
biological behavior.
Finally, in our case, two other possible causes were discarded.
Thus, the hypothesis of a stereochemical misassignment was
abandoned after the synthesis and biological evaluation of 10
possible epimers of phakellistatin 19 (see Supporting Informa-
tion).10 Neither did preliminary chelation experiments account
for the differences in cytotoxicity, as no effective chelation was
detected by matrix-assisted laser desorption/ionization
(MALDI) analysis (see Supporting Information).
Thus, our efforts were focused on performing a thorough
NMR study of both the synthetic and natural samples in the
search of conformational differences, without disregarding the
possibility of an impurity present in the natural sample, which
would be difficult to prove.
Synthesis of Phakellistatin 19. The linear precursor of
phakellistatin 19 was synthesized on solid phase following the 9-
fluorenylmethoxycarbonyl/tert-butyl (Fmoc/tBu) protection
scheme and using the 2-chlorotrityl11 chloride resin (2-CTC)
as the polymeric support to minimize diketopiperazine (DKP)
Once the linear precursor was fully assembled, it was cleaved
from the resin (dichloromethane (DCM)−trifluoroacetic acid
(TFA) (98:2), 5 × 2 min) and collected in water to prevent the
loss of the side chain protecting groups (tert-butyloxycarbonyl
t
(Boc) and Bu). After lyophilization of the precursor, the
macrolactamization reaction was undertaken for 3 h by means of
benzotriazol-1-yloxytris(pyrrolidino)phosphonium hexafluoro-
phosphate (PyBOP)/1-hydroxy-7-azabenzotriazole (HOAt)/
DIEA (2:1:4) in DCM−DMF (95:5) at pH 8 and at diluted
conditions (10−4 M) to prevent oligomerization. Finally, all the
side chain protecting groups were removed by treatment with
TFA−H2O (95:5) for 1 h, and the crude product was purified by
reversed-phase semipreparative HPLC to obtain phakellistatin
19 in 22% overall yield (see Supporting Information).
Chemical and Spectral Validation. To verify the chemical
identity of the product obtained, samples of synthetic and natural
phakellistatin 19 were dissolved in H2O−acetonitrile (ACN)
(1:1) and analyzed by reversed-phase high performance liquid
chromatography photodiode array (HPLC-PDA) using a C18
analytical column (Figure 2a,b). The two samples were then
coeluted using a flat long gradient (Figure 2c). A single peak was
obtained, indicating the chemical equivalence of the two samples.
B
dx.doi.org/10.1021/jm401520x | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 3. Energetically minimized structure of phakellistatin 19
obtained after application of a restricted SA protocol. The two checked
hydrogen bonds are highlighted in dashed green lines.
(ROESY) cross-peaks between Hα-AAi+1 and Hδ-Proi were
found for all the Pro moieties in both natural and synthetic
phakellistatin 19, confirming the trans geometry of the prolyl
peptide bonds in all cases. Moreover, ROE cross-peaks involving
the side chains of Phe and Ile proved the presence of key
hydrophobic interactions between these two residues in both
peptides. Finally, the existence of minor conformers is detected
for synthetic and natural phakellistatin 19. The occurrence of
these conformers in such a small percentage prevents their
assignment.
However, as expected, biological evaluation of synthetic
phakellistatin 19 against three human cancer cell lines did not
provide the same cytotoxicity as for its natural counterpart (see
Table 2).
Structural Elucidation. Exhaustive NMR studies further
proved that the three Pro residues in synthetic phakellistatin 19
adopted the trans geometry in all solvents (dimethyl sulfoxide
(DMSO)-d6, CDCl3, CD3OD). The structure of phakellistatin
19 was calculated by applying a restricted simulated annealing
(SA) protocol. ROESY cross-peaks volumes (in DMSO-d6) were
suitably corrected and converted into interatomic distances
which were used as experimental restrictions in the SA.
Furthermore, variable-temperature nuclear magnetic resonance
(VTNMR) analysis strongly pointed out the likely participation
of the Phe and Thr amide protons in two intramolecular
hydrogen bonds. They were also taken into account when
undertaking the SA. The minimized structure of synthetic
phakellistatin 19 (Figure 3) presents a β-turn stabilized by the
hydrogen bond ThrNH−OCPhe (the measured distance is 2.06
Å) and with the residues Pro1 and Leu placed at the positions i +
1 and i + 2 respectively.12 A second hydrogen bond is observed,
PheNH−OCPro4 (1.95 Å), that would stabilize a γ-turn
involving the residues Phe, Trp, and Pro4 in the positions i, i +
1, and i + 2, respectively. The ϕi+1 and ψi+1 angles measure 78°
and −56°, respectively, on the minimized structure. These values
perfectly match the ones established for a γ-turn: 70 to 95 for ϕi+1
and −75 to −45 for ψi+1. The region around the γ-turn resembles
a hairpin motif. It comprises two antiparallel strains linked by
means of hydrogen bonds and orienting the hydrophobic side
chains of the Phe and Ile residues into the same direction. Van
der Waals interactions between these two side chains help to
stabilize the structural motif. Moreover, Trp’s side chain points to
the upper region of the γ-turn and is completely exposed to the
solvent, suggesting its possible participation in the pharmaco-
phore. The whole structure is highly folded and, somehow, draws
a characteristic chair shape. Remarkably, the key ROESY cross-
peaks between Hβ-Ile and Hβ-Phe and between Hγ-Ile and Hβ-
Phe are found in both synthetic and natural phakellistatin 19.
Figure 2. HPLC-PDA analysis of (a) synthetic phakellistatin 19, (b)
natural phakellistatin 19, and (c) both natural and synthetic
phakellistatin 19, coelution. Conditions were the following: Symmetry
C18 reversed-phase analytical column (5 μm × 4.6 mm × 150 mm);
linear gradient from 45% to 75% of ACN over 15 min for (a) and (b) and
from 35% to 55% of ACN over 30 min for the coelution experiment.
HRMS data for the natural and synthetic phakellistatin 19 also
matched perfectly. In addition, comparison of monodimensional
1H and homonuclear bidimensional spectra of natural and
synthetic phakellistatin 19 in CD3OD at 298 K proved their
spectral equivalence (Table 1). Only minor differences were
1
detected when comparing complete H assignment of the two
peptides. Rotating frame Overhauser effect spectroscopy
C
dx.doi.org/10.1021/jm401520x | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 1. 1H NMR Spectral Assignment of Natural and Synthetic Phakellistatin 19 in CD3OD
1H (ppm)
natural
peptide
1H (ppm)
synthetic
peptide
1H (ppm)
natural
peptide
1H (ppm)
synthetic
peptide
1H (ppm)
natural
peptide
1H (ppm)
synthetic
peptide
Pro1
Hα
H2
6.72 (d, 1.7)
7.48 (d, 7.9)
7.09 (t, 7.2)
7.00 (m)
6.73 (s)
Pro6
Hα
4.22
4.22
H4
7.49 (d, 7.9)
7.09 (t, 7.1)
7.00 (m)
4.56 (dd, 8.8, 5.5)
2.31, 1.96′
1.95
4.57 (dd, 8.7, 5.4)
2.33, 1.97′
1.95
Hβ
2.30, 1.89′
2.07
2.30, 1.90′
2.10
H5/6
H6/5
H7
Hβ
Hγ
Hγ
Hδ
Phe2 NH
3.96, 3.75′
3.97 (t, 8.1), 3.75′
8.19 (d, 9.8)
5.27 (td, 9.4, 5.0)
3.02
7.35−7.24
10.29
7.35−7.24
Hδ
Thr7 NH
3.83, 3.67′
3.83, 3.68′
7.93 (d, 9.6)
5.07 (dd, 9.7, 4.0)
4.36
NHind
Pro4
Hα
Hα
5.27
Hα
5.06
Hβ
3.01
3.98
3.98
Hβ
4.36
H2/6
H3/5
H4
Trp3 NH
Hα
7.27−7.24
7.35−7.30
7.35−7.24
7.28−7.24
7.35−7.31
7.35−7.24
Hβ
2.04, 1.77′
1.94, 1.87′
4.05, 3.41′
2.04, 1.79′
1.96, 1.87′
4.05, 3.41′
8.56 (d, 8.7)
4.26
Hγ
1.13 (d, 6.3)
1.14 (d, 6.3)
8.74 (d, 6.7)
3.76
Hγ
Leu8 NH
Hδ
Hα
3.76
Ile5 NH
Hα
Hβ
2.36, 1.77′
1.58
2.37, 1.78′
1.58
4.24
4.24
4.25
2.06
Hγ
Hβ
3.40, 3.16′
3.42 (dt, 14.9, 7.7), 3.17′
Hβ
2.06
Hδ
0.96 (d, 6.7)
0.97 (d, 6.7)
(14.9, 4.8)
Hγ
Hδ
1.47, 1.13′
0.60 (d, 6.8)
1.47, 1.13′
0.61 (d, 6.8)
0.93 (d, 6.5)
0.93 (d, 6.5)
0.80 (t, 7.4)
0.81 (t, 7.4)
Figure 4. Chemical structures of Cys(ΨMe,Mepro) analogues of phakellistatin 19. The residues Cys(ΨMe,Mepro) are highlighted in dark red.
Influence of the pH on the Geometries of Prolyl
Peptide Bonds. To discard a possible influence of the pH in the
biological behavior (in vitro assays are carried out in aqueous
media), a spectrally comparable and H2O-soluble analogue was
designed on the basis of the previous described minimized
structure, with a positively charged Orn at the Leu’s position. Its
synthesis was successfully achieved following the synthetic
strategy developed for phakellistatin 19, and the biological assays
revealed no significant cytotoxic properties. Once the Orn
analogue was proven to be H2O-soluble, its spectral equivalence
with synthetic phakellistatin 19 was checked by comparison of
1H NMR spectral assignments of the two peptides in DMSO-d6,
confirming small differences. After complete NMR character-
ization at pH 5.95 and 8.12, trans isomerism was confirmed for
D
dx.doi.org/10.1021/jm401520x | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Scheme 2. Synthetic Strategy To Obtain Cys(ΨMe,Mepro)-
present in very small amounts, would be mainly responsible for
the high biological activity.
a
Containing Dipeptides
Phakellistatin 19 as a Hit for a Medicinal Chemistry
Program. Despite the low biological activity of synthetic
phakellistatin 19, we believe that its structure is enough
interesting as a starting point for the development of a medicinal
chemistry program. As a first step, we decided to cover a new
conformational scenario in the search of an increased biological
activity. To this end, chemical modification was undertaken to
access analogues of phakellistatin 19 with induced cis-isomerism
by replacing Pro by Cys(ΨMe,Mepro). In 1992, Mutter et al.
described new Pro surrogates easily accessed by means of
cyclocondensation of the amino acids Ser, Thr, or Cys with
aldehydes or ketones.13 These pseudo-prolines (ΨPro or
Xaa(ΨR,R′pro)) acted as structure-disrupting agents, preventing
peptide aggregation and self-association and therefore increasing
the efficiency of peptide synthesis.14 Furthermore, they can be
used as removable turn inducers, facilitating the cyclization of
linear peptides.15 Finally, their introduction into a peptide
sequence contributes to the modulation of the peptide’s
biological and pharmacokinetic properties.16
a
Fmoc-Ile-Cys(ΨMe,Mepro)-OH synthesis is represented.
Table 2. In Vitro Results for Natural Phakellistatin 19,
Synthetic Phakellistatin 19, and Thz Analogues
a
Interestingly, the cis to trans ratio at the Xaai+1-Xaa(ΨR,R′pro)i
amide bond, as well as the lability of pseudo-prolines to acid, can
be modulated. Thus, Cys-derived pseudo-prolines exhibit a larger
Pro effect enhancement in comparison to the ΨPro obtained
from Thr and Ser.16 Moreover, both the stereochemistry and the
degree of substitution at the 2-C ΨPro position are crucial factors
determining the cis content along the imidic bond.17,18
Detailed NMR studies confirmed that 2,2-dimethylated
derivatives show a higher percentage of cis geometry at the
Xaa-ΨPro peptide bond.17,18 Thus, it was finally decided that
Cys(ΨMe,Mepro) would replace Pro to strongly enhance the cis
conformer at the Xaai+1-Cys(ΨMe,Mepro)i peptide bonds in
synthetic phakellistatin 19. With this purpose in mind, a small
library of seven peptides with the Pro moieties replaced by
Cys(ΨMe,Mepro), covering all the possibilities, was designed and
biologically tested (Figure 4).
The synthetic strategy previously validated for phakellistatin
19 could not be directly applied to obtain the pseudo-proline-
containing analogues, since the extremely hindered Cys-
(ΨMe,Mepro) was not acylated under any conditions (1-[(1-
cyano-2-ethoxy-2-oxoethylideneaminooxy)dimethylamino-
morpholinomethylene)]methanaminium hexafluorophosphate
(COMU)/OxymaPure/DIEA, 1-[bis(dimethylamino)-
methylene]-1H-1,2,3-triazolo-[4,5-b]pyridinium hexafluoro-
phosphate 3-oxide (HATU)/HOAt/DIEA, PyBOP/HOAt/
DIEA, N,N,N′,N′-tetramethylchloroformamidinium hexafluor-
ophosphate (TCFH)/HOAt/DIEA, or Fmoc-aa-F) applying
microwave and other reagents and techniques. Taking this into
consideration, all approaches including solid-phase acylation of
Cys(ΨMe,Mepro) were discarded, and the synthesis in solution of
dipeptides containing the Cys(ΨMe,Mepro) at the C-terminus was
faced.19 Although coupling of Fmoc-aa-F on Cys(ΨMe,Mepro)
was not achieved on solid phase, it did work in solution to form
the corresponding dipeptides (Scheme 2).
a
n.d. means not detected. GI50, compound concentration that
produces 50% of cell growth inhibition compared to control cultures.
TGI, total cell growth inhibition. LC50, compound concentration that
produces 50% of net cell killing. IC50, compound concentration that
produces 50% inhibition of biological activity of cancer cells.
Antimitotic activity for natural phakellistatin 19 has also been
described: IC50 = 4.20 × 10−7 to 8.40 × 10−8 M (see Supporting
Information).
the three prolines also in these conditions, proving that no
conformational change at the Pro linkages occurred in acidic or
basic media.
In view of all these results and considering the presence of
minor peaks in the NMR spectra and the HPLC-PDA
chromatogram of natural phakellistatin 19 (see Figure 2a and
Supporting Information), it is plausible that a cytotoxic agent,
With all dipeptides in hand, a modified synthetic scheme on
solid phase was achieved to access all the Cys(ΨMe,Mepro)-
containing analogues. For the synthesis of Thz1, Thz1,4, Thz1,6,
and Thz1,4,6 analogues, the starting/cyclization point was
changed to the Thr-Leu linkage to prevent direct incorporation
of the dipeptide onto the resin. The synthesis of all the
E
dx.doi.org/10.1021/jm401520x | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Scheme 3. Synthetic Strategy for the Thz4 Phakellistatin 19 Analogue as Example for the Rest of Cys(ΨMe,Mepro)-Containing
Peptides
Cys(ΨMe,Mepro)-containing analogues was successfully achieved.
A representative scheme for Thz4 analogue is shown (Scheme 3).
Data obtained from the biological assays showed that
replacement of Pro6 by a Cys(ΨMe,Mepro) caused a significant
increase in cytotoxicity (Table 2). Moreover, a more rigid
structure produced by the presence of an increasing number of
Cys(ΨMe,Mepro) residues also increased the bioactivity of the
compound. For a better understanding of these results, we
performed a structural study of the library by means of NMR.
CD3OH was chosen as the working solvent because it favors the
coexistence of conformers less than CDCl3 does.
(gCOSY), total correlation spectroscopy (TOCSY), nuclear
Overhauser effect spectroscopy (NOESY), ROESY, and
heteronuclear single-quantum correlation spectroscopy
(gHSQC) experiments allowed H NMR spectral assignment
of both conformers.
1
The gHSQC experiment enabled us to assign the key carbon
atoms of the Pro moiety Cβ and Cγ. For Pro6 of conformer 1, δCβ
= 33.095 ppm and δCγ = 22.643 ppm, meaning that ΔδCβ‑Cγ
=
10.452 ppm. For Pro6 of conformer 2, δCβ = 29.943 ppm and δCγ
= 26.791 ppm, meaning that ΔδCβ‑Cγ = 3.152 ppm. According to
these data, Pro6 adopted cis isomerism in conformer 1 and trans
isomerism in conformer 2. A large NOE cross-peak between Hα-
Thr and Hδ-Pro6 in conformer 2 and ROE cross-peak between
Hα-Thr and Hα-Pro6 in conformer 1 also supported the assigned
isomerism at the Thr-Pro6 linkage for the two conformers.
As expected, both Cys(ΨMe,Mepro)1 and Cys(ΨMe,Mepro)4
adopted cis isomerism in all conformers. Large NOE cross-
peaks between Hα-Phe and Hα-Cys(ΨMe,Mepro)1 and between
Hα-Ile and Hα-Cys(ΨMe,Mepro)4 served as confirmation.
The Thz1,6 analogue showed more confusing spectroscopic
data. Again, two major distinct conformers were detected. 1H and
13C NMR spectral assignment of the 16 residues was
accomplished, but the cross-peaks of the ROESY spectra did
not provide enough information to perform complete sequential
assignment of the two conformers.20 However, the two Pro
moieties were fully assigned and their geometry was identified
(see Supporting Information).
1H NMR spectra of the most active monosubstituted analogue
Thz6 in CD3OH were recorded at two temperatures: 278 and
308 K. Analysis of the indolic region (between 10 and 11 ppm) of
the spectra provided an idea of the complex conformational
equilibrium in which this analogue was involved. Replacement of
one single Pro residue by a Cys(ΨMe,Mepro) dramatically
modified the cis−trans isomerism at the Pro linkages, thus
altering the conformational panorama. In CD3OD at 298 K,
synthetic phakellistatin 19 appeared as a single major conformer
with all Pro in trans, while Thz6 presented in CD3OH at least
seven conformers (five major conformers and two minor
conformers) at 278 K and at least four conformers (two of
them in a fast equilibrium) at 308 K. Any attempt to assign 1H
appeared extremely challenging. Thus, two presumably less
flexible analogues, Thz1,4 and Thz1,6, were examined.
As detected by NMR, Thz1,4 in CD3OH at 298 K presented
two major conformers, in a conformer 1 to conformer 2 ratio of
58:42. Exhaustive analysis of 1H, correlation spectroscopy
For conformer 1, it was found that ΔδCβ‑Cγ = 9.54 ppm,
suggesting a cis geometry of the amide bond Ile-Pro4. On the
F
dx.doi.org/10.1021/jm401520x | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 5. 1H NMR spectrum of Thz1,4,6 analogue in CD3OH at 298 K. The region between 10 and 11 ppm is enlarged.
contrary, for conformer 2, ΔδCβ‑Cγ = 2.03 ppm strongly pointed
to a trans prolyl peptide bond. The ROESY cross-peaks detected
between Hα-Ile′ and Hδ/Hδ′-Pro4′ for conformer 2 also
supported the trans isomerism. For conformer 1, no ROE
cross-peak between Hα-Ile and Hα-Pro4 was detected, probably
because of the low quality of the ROESY spectra. The cis
geometry of the Xaa-Cys(ΨMe,Mepro) linkages was confirmed by
the cross-peaks between Hα-Thr and Hα-Cys(ΨMe,Mepro) and
between Hα-Phe and Hα-Cys(ΨMe,Mepro). All the NMR
experiments were recorded in CD3OH at 273 K.
two extra Me groups, and also an alteration of the hydrogen
donors and acceptors pattern, as a sulfur atom (a hydrogen
acceptor) is introduced. However, the bioactivity results showed
a noticeable trend from the monosubstituted analogue Thz6
(lower activity) toward the trisubstituted analogue Thz1,4,6
(highest activity). Moreover, only small differences are observed
between Thz1 and Thz6. Thus, we propose that this difference is
caused by a structural issue (gain of cis geometry) rather than by
the aa6 being the pharmacophore of phakellistatin 19 analogues.
Moreover, as confirmed by NMR analysis, an increasing
number of Cys(ΨMe,Mepro) in the phakellistatin 19 structure
entails a gain of structural rigidity. The Thz1,4,6 analogue not only
was the most active one but also was the most rigid, with a major
conformer accounting for 89% of the mixture with all the Xaai+1-
Cys(ΨMe,Mepro)i linkages adopting the cis isomerism. A
significantly more active all-cis analogue of phakellistatin 19
strongly suggests that the cis−trans isomerism at the Pro linkages
makes a crucial contribution to the bioactivity displayed by the
ΨPro-containing analogues of phakellistatin 19.
Finally, the Thz1,4,6 analogue, with the three Pro residues
replaced by Cys(ΨMe,Mepro), was the most conformationally
restricted peptide, as confirmed by the presence of one major
(89%) and three minor conformers (Figure 5). NMR studies
showed that all Cys(ΨMe,Mepro) adopted cis isomerism, as proved
by the NOE cross-peaks detected between the protons Hα-Phe
and Hα-Cys(ΨMe,Mepro)1, Hα-Ile and Hα-Cys(ΨMe,Mepro)4, Hα-
Thr and Hα-Cys(ΨMe,Mepro)6, Hβ-Phe and Hα-Cys(ΨMe,Mepro)1,
Hβ-Ile and Hα-Cys(ΨMe,Mepro)4, and Hβ-Thr and Hα-Cys-
(ΨMe,Mepro)6.
Together, the biological and structural data suggest that Pro6
plays a crucial role in the structure of phakellistatin 19 analogues
and has a direct effect on the bioactivity exerted by the ΨPro-
containing peptides. Pro replacement by Cys(ΨMe,Mepro) causes
a significant gain of steric hindrance due to the presence of the
CONCLUSIONS
■
The introduction of ΨMe,Mepro residues in a cyclic peptide such
as phakellistatin 19 increases the percentage of cis conformation
in the final peptides, and this translates into enhanced biological
activity. A correlation between the number of ΨMe,Mepro units
G
dx.doi.org/10.1021/jm401520x | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
introduced and the enhanced cytotoxic activity was also
observed, the peptide containing the three ΨMe,Mepro residues
showing the highest activity. In this regard, we envisage that the
use of ΨMe,Mepro moieties will be widely adopted to increase the
biological activity of cyclic peptides. Furthermore, other families
of Pro-rich cyclic peptides should be revised under this new
perspective.
correlation spectroscopy; VTNMR, variable-temperature nu-
clear magnetic resonance
REFERENCES
■
(1) (a) Napolitano, A.; Bruno, I.; Rovero, P.; Lucas, R.; Peris, M. P.;
Gomez-Paloma, L.; Riccio, R. Synthesis, Structural Aspects and
Bioactivity of the Marine Cyclopeptide Hymenamide C. Tetrahedron
2001, 57, 6249−6255. (b) Kobayashi, J.; Nakamura, T.; Tsuda, M.
Hymenamide F, New Cyclic Heptapeptide from Marine Sponge
Hymeniacidon sp. Tetrahedron 1996, 52, 6355−6360.
ASSOCIATED CONTENT
■
S
* Supporting Information
(2) (a) Pettit, G. R.; Srirangam, J. K.; Herald, D. L.; Xu, J.; Boyd, M. R.;
Cichacz, Z.; Kamano, Y.; Schmidt, J. M.; Erickson, K. L. Isolation and
Crystal Structure of Stylopeptide 1, a New Marine Porifera Cyclo-
heptapeptide. J. Org. Chem. 1995, 60, 8257−8261. (b) Pettit, G. R.;
Taylor, S. R. Synthesis of the Marine Sponge Cycloheptapeptide
Stylopeptide 1. J. Org. Chem. 1996, 61, 2322−2325.
(3) (a) Randazzo, A.; Dal Piaz, F.; Orru, S.; Debitus, C.; Roussakis, C.;
Pucci, P.; Gomez-Paloma, L. Axinellins A and B. New Proline-
Containing Antiproliferative Cyclopeptides from the Vanuatu Sponge
Axinella carteri. Eur. J. Org. Chem. 1998, 11, 2659−2665. (b) Tabudravu,
J. N.; Morris, L. A.; Kettenes-van den Bosch, J. J.; Jaspars, M. Axinellin C,
a Proline-Rich Cyclic Octapeptide Isolated from the Fijian Marines
Sponge Stylotella Aurantium. Tetrahedron 2002, 58, 7863−7868.
(4) (a) Pettit, G. R.; Herald, C. L.; Boyd, M. R.; Leet, J. E.; Dufresne,
C.; Doubek, D. L.; Schmidt, J. M.; Cerny, R. L.; Hooper, J. N.; Rutzler, K.
C. Antineoplastic Agents. 219. Isolation and Structure of the Cell
Growth Inhibitory Constituents from the Western Pacific Marine
Sponge Axinella sp. J. Med. Chem. 1991, 34, 3339−3340. (b) Pettit, G.
R.; Gao, F.; Schmidt, J. M.; Chapuis, J.-C.; Cerny, R. L. Isolation and
Structure of Axinastatin 5 from a Republic of Comoros Marine Sponge.
Bioorg. Med. Chem. Lett. 1994, 4, 2935−2940.
(5) (a) Pettit, G. R.; Tan, R.; Williams, M. D.; Tackett, L.; Schmidt, J.
M.; Cerny, R. L.; Hooper, J. N. A. Isolation and Structure of
Phakellistatin 2 from the Eastern Indian Ocean Marine Sponge
Phakellia carteri. Bioorg. Med. Chem. Lett. 1993, 3, 2869−2874. (b) Pettit,
G. R.; Cichacz, Z.; Barkoczy, J.; Dorsaz, A.-C.; Herald, D. L.; Williams,
M. D.; Doubek, D. L.; Schmidt, J. M.; Tackett, L. P.; Brune, D. C.; Cerny,
R. L.; Hooper, J. N. A.; Bakus, G. J. Isolation and Structure of the Marine
Sponge Cell Growth Inhibitory Cyclic Peptide Phakellistatin 1. J. Nat.
Prod. 1993, 56, 260−267.
(6) (a) Pettit, G. R.; Xu, J.; Dorsaz, A.-C.; Williams, M. D.; Boyd, M. R.;
Cerny, R. L. Antineoplastic agents. 324. Isolation and Structure of the
Human Cancer Cell Growth Inhibitory Cyclic Decapeptides
Phakellistatins 7, 8 and 9. Bioorg. Med. Chem. Lett. 1995, 5, 1339−
1344. (b) Pettit, G. R.; Tan, R. Antineoplastic Agents 390. Isolation and
Structure of Phakellistatin 12 from a Chuuk Archipelago Marine
Sponge. Bioorg. Med. Chem. Lett. 2003, 13, 685−688.
(7) Pettit, G. R.; Tan, R.; Ichihara, Y.; Williams, M. D.; Doubek, D. L.;
Tackett, L. P.; Schmidt, J. M.; Cerny, R. L.; Boyd, M. R.; Hooper, J. N. A.
Antineoplastic Agents. 325. Isolation and Structure of the Human
Cancer Cell Growth Inhibitory Cyclic Octapeptides Phakellistatin 10
and 11 from Phakellia sp. J. Nat. Prod. 1995, 58, 961−965.
(8) (a) Bates, R. B.; Caldera, S.; Ruane, M. D. Synthesis and
Stereochemistry of Axinastatin 4. J. Nat. Prod. 1998, 61, 405.
(b) Mechnich, O.; Hessler, G.; Kessler, H.; Bernd, M.; Kutscher, B.
Cyclic Heptapeptides Axinastatin 2, 3, and 4. Conformational Analysis
and Evaluation of the Biological Potential. Helv. Chim. Acta 1997, 80,
1338−1354.
(9) (a) Pettit, G. R.; Rhodes, M. R.; Tan, R. Antineoplastic Agents. 400.
Synthesis of the Indian Ocean Marine Sponge Cyclic Heptapeptide
Phakellistatin 2. J. Nat. Prod. 1999, 62, 409−414. (b) Pettit, G. R.; Toki,
B. E.; Xu, J.-P.; Brune, D. C. Synthesis of the Marine Sponge
Cycloheptapeptide Phakellistatin 5. J. Nat. Prod. 2000, 63, 22−28.
(c) Pettit, G. R.; Lippert, J. W.; Taylor, S. R.; Tan, R.; Williams, M. D.
Synthesis of Phakellistatin 11: A Micronesia (Chuuk) Marine Sponge
Cyclooctapeptide. J. Nat. Prod. 2001, 64, 883−891. (d) Napolitano, A.;
Rodriguez, M.; Bruno, I.; Marzocco, S.; Autore, G.; Riccio, R.; Gomez-
Paloma, L. Synthesis, Structural Aspects and Cytotoxicity of the Natural
Synthetic protocols and characterization data of phakellistatin 19,
epimers, and analogues. This material is available free of charge
via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Authors
*J.T.-P.: phone, +34 93 4037127; e-mail, judit.tulla@
irbbarcelona.org.
*F.A.: phone, +34 93 4037088; fax, +34 93 4037126; e-mail,
albericio@irbbarcelona.org.
Author Contributions
The manuscript was written through contributions of all authors.
All authors have given approval to the final version of the
manuscript.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This study was funded by the CICYT (Grant CTQ2012-30930),
the Generalitat de Catalunya (Grant 2009SGR1024), and the
Institute for Research in Biomedicine Barcelona (IRB
Barcelona). M.P.-G. thanks MICINN CTQ 2009-07758 for a
FPU Ph.D. fellowship. A.M. thanks the Italian Agency for the
LifeLong Learning Programme Erasmus.
ABBREVIATIONS USED
■
ACN, acetonitrile; Alloc, allyloxycarbonyl; Boc, tert-butylox-
ycarbonyl; 2-CTC, 2-chlorotrityl chloride (Barlos) resin;
COMU, 1-[(1-(cyano-2-ethoxy-2-oxoethylideneaminooxy)-
dimethylaminomorpholinomethylene)]methanaminium hexa-
fluorophosphate; COSY, correlation spectroscopy; DCM,
dichloromethane; DIEA, N,N-diisopropylethylamine; DIPCDI,
N,N′-diisopropylcarbodiimide; DKP, diketopiperazine; DMF,
N,N-dimethylformamide; DMSO, dimethyl sulfoxide; ESMS,
electrospray mass spectrometry; Fmoc, 9-fluorenylmethoxycar-
bonyl; HATU, 1-[bis(dimethylamino)methylene]-1H-1,2,3-
triazolo[4,5-b]pyridinium hexafluorophosphate 3-oxide;
HBTU, 1-[bis(dimethylamino)methylene]-1H-benzotriazolium
hexafluorophosphate 3-oxide; HOAt, 1-hydroxy-7-azabenzotria-
zole; HOBt, 1-hydroxybenzotriazole; HPLC, high performance
liquid chromatography; HRMS, high-resolution mass spectrom-
etry; HSQC, heteronuclear single-quantum correlation spec-
troscopy; MALDI, matrix-assisted laser desorption/ionization;
MS, mass spectrometry; NMR, nuclear magnetic resonance;
NOESY, nuclear Overhauser effect spectroscopy; PDA, photo-
diode array; PyBOP, benzotriazol-1-yloxytris(pyrrolidino)-
phosphonium hexafluorophosphate; ROESY, rotating frame
Overhauser effect spectroscopy; ΨPro, pseudo-proline; TBME,
t
tert-butyl methyl ether; Bu, tert-butyl; TCFH, N,N,N′,N′-
tetramethylchloroformamidinium hexafluorophosphate; TFA,
trifluoroacetic acid; TIS, triisopropylsilane; TOCSY, total
H
dx.doi.org/10.1021/jm401520x | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Cyclopeptides Yunnanins A, C and Phakellistatins 1, 10. Tetrahedron
Lett. 2003, 59, 10203−10211.
(10) Analogues designed in order to increase activity (e.g., with
fluorine-containing amino acids) also failed (see Supporting Informa-
tion).
(11) Barlos, K.; Gatos, D.; Kallitsis, J.; Papaphotiu, G.; Sotiriu, P.; Yao,
W.; Schaefer, W. Preparation of Protected Peptide Fragments Using
Triphenylmethyl Resins. Tetrahedron Lett. 1989, 30, 3943−3946.
(12) Numbering of the residues involved in β- and γ-turns follows the
criteria used for proteins. In the rest of the article, aa is numbered from
C-terminal to N-terminal of the linear precursor.
(13) (a) Haack, T.; Mutter, M. Serine Derived Oxazolidines as
Secondary Structure Disrupting, Solubilizing Building Blocks in Peptide
Synthesis. Tetrahedron Lett. 1992, 33, 1589−1592. (b) Wohr, T.;
̈
Mutter, M. Pseudo-Prolines in Peptide Synthesis: Direct Insertion of
Serine and Threonine Derived Oxazolidines in Dipeptides. Tetrahedron
Lett. 1995, 36, 3847−3848.
(14) (a) Mutter, M.; Nefzi, A.; Sato, T.; Sun, X.; Wahl, F.; Wohr, T.
̈
Pseudo-prolines (ψPro) for Accessing Inaccessible Peptides. Pept. Res.
1995, 8, 145−153. (b) Wohr, T.; Wahl, F.; Nefzi, A.; Rohwedder, B.;
̈
Sato, T.; Sun, X.; Mutter, M. Pseudo-Prolines as a Solubilizing,
Structure-Disrupting Protection Technique in Peptide Synthesis. J. Am.
Chem. Soc. 1996, 118, 9218−9227.
(15) (a) Skropeta, D.; Jolliffe, K.; Turner, P. Pseudoprolines as
Removable Turn Inducers: Tools for the Cyclization of Small Peptides.
J. Org. Chem. 2004, 69, 8804−8809. (b) Sayyadi, N.; Skropeta, D.;
Jolliffe, K. N,O-Isopropylidenated Threonines as Tools for Peptide
Cyclization: Application to the Synthesis of Mahafacyclin B. Org. Lett.
2005, 7, 5497−5499. (c) Wong, M. S. Y.; Jolliffe, K. Synthesis of
Cyclogossine B Using a Traceless Pseudoproline Turn-Inducer. Aust. J.
Chem. 2010, 63, 797−801.
(16) Wittelsberger, A.; Keller, M.; Scarpellino, L.; Patiny, L.; Acha-
Orbea, H.; Mutter, M. Pseudoprolines: Targeting a Cis Conformation in
a Mimetic of the gp120 V3 Loop of HIV-1. Angew. Chem., Int. Ed. 2000,
39, 1111−1115.
(17) Keller, M.; Sager, C.; Dumy, P.; Schutkowski, M.; Fischer, G. S.;
Mutter, M. Enhancing the Proline Effect: Pseudo-Prolines for Tailoring
Cis/Trans Isomerization. J. Am. Chem. Soc. 1998, 120, 2714−2720.
(18) Dumy, P.; Keller, M.; Ryan, D. E.; Rohwedder, B.; Wohr, T.;
̈
Mutter, M. Pseudo-Prolines as a Molecular Hinge: Reversible Induction
of Cis Amide Bonds into Peptide Backbones. J. Am. Chem. Soc. 1997,
119, 918−925.
(19) Wittelsberger, A.; Patiny, L.; Slaninova, J.; Barberis, C.; Mutter, M.
Introduction of a cis-Prolyl Mimic in Position 7 of the Peptide Hormone
Oxytocin Does Not Result in Antagonistic Activity. J. Med. Chem. 2005,
48, 6553−6562.
(20) The partial assignment prevented the finding of the ratio between
the cis and the trans conformers.
I
dx.doi.org/10.1021/jm401520x | J. Med. Chem. XXXX, XXX, XXX−XXX