A Lasso-Inspired Bicyclic Peptide: Synthesis, Structure and Properties

Article abstract of DOI:10.1002/chem.201803899

The chemical synthesis of a bicycle inspired by the natural lasso peptide sungsanpin using a combination of solid-phase and in-solution chemistries is described. The bicyclic-derived topoisomer was designed by introducing a covalent linkage between the ri

Full text of DOI:10.1002/chem.201803899

A Journal of
Accepted Article
Title: A Lasso-Inspired Bicyclic Peptide: Synthesis, Structure and
Properties
Authors: Helena Martin-Gómez, Fernando Albericio, and Judit Tulla-
Puche
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201803899
Link to VoR: http://dx.doi.org/10.1002/chem.201803899
Supported by

10.1002/chem.201803899
Chemistry - A European Journal
FULL PAPER
A Lasso-Inspired Bicyclic Peptide: Synthesis, Structure and
Properties
Helena Martin-Gómez, [a] Fernando Albericio [b][c][d] and Judit Tulla-Puche* [b][c][e]
Abstract: The chemical synthesis of a bicycle inspired by the natural
lasso peptide sungsanpin using a combination of solid-phase and in-
solution chemistries is described. The bicyclic-derived topoisomer
was designed introducing a covalent linkage between the ring and the
loop, which allowed the tying of these two parts of the peptide
rendering the bicyclic structure. Several structural techniques, such
as MS fragmentation, ion-mobility and NMR were used to characterize
the bicycle. Ion-mobility spectroscopy studies revealed that it showed
lasso-like behavior. Its 3D structure was predicted on the basis of the
NMR restraints. In addition, the high proteolytic and thermal stability
encountered in building and maintaining the threaded lasso
structure. Recently, peptide-based [1]rotaxanes were synthesized,
which were used as molecular grafting scaffolds.[6] Here, we
focused on the class II lasso peptides sungsanpin[7] 1 and
chaxapeptin[8] 2. Both peptides contain 15 residues, and the
7-amino acid tail is threaded through the ring via hydrophobic
interactions. These peptides were isolated from two distinct
Streptomyces strains, which share 16S rRNA gene sequence
similarity of 94%.[8] An analysis of their sequences reveals that
they share the same number and chemical character of residues
of the bicycle potentially make it a suitable scaffold for epitope grafting. (Figure 1). They differ in only four residues, three of which are
located at the C-terminal tail. In this position, NFF residues are
replaced by SWL, thus maintaining hydrophobicity, and there is
also a conservative change in the ring, namely Leu instead of Ile.
Regarding their activity, both peptides show inhibitory effect on
the invasion of human lung cancer cells.[8]
Introduction
Lasso peptides are natural products that exhibit a broad spectrum
Here, we describe the first chemical synthesis of a bicyclic
peptide 5 in which the loop sequence is tied to the ring through a
covalent bond taking as basis the natural lasso peptide
sungsanpin 1 (Figure 2). For this purpose, we have designed a
synthetic route in which the bicyclic structure was achieved first,
through the ring and the loop, creating a bridged bicyclic
peptide.[9] The C-terminal tail was then introduced. It was known
that once the ring is closed the C-terminal tail cannot be trapped
due to the steric hindrance in these positions.[10] The synthetic
approach has been designed using the feasible and efficient solid-
phase peptide synthesis (SPPS),[11] in combination with solution
chemistry, which can facilitate the last synthetic steps that are
more demanding due to the steric rigidity already present in the
intermediates. Chaxapeptin 2, which was obtained via
homologous expression (see Supplementary information), was
used as a control for the characterization studies and structural
analysis.
of bioactivities. They consist of 15–26 proteinogenic amino acids
and share an N-terminal 7- to 9-residue macrolactam ring.[1] The
C-terminal tail threads through an N-terminal macrolactam ring in
a right-handed conformation, which provides them a unique 3D
conformation. The rigid and folded structure of these peptides is
the result of a steric lock, in which a bulky and often an aromatic
side chain is unable to pass through the macrolactam ring. This
constrained structure confers lasso peptides extraordinary
stability against chemical, thermal and proteolytic degradation.[1–
4]
They show a well-defined secondary structure; sometimes
presenting β-sheets associated with β-turns, forming a hairpin-
like structure despite their relatively small length.[5] These
distinctive features could make them promising candidates as
molecular scaffolds for drug design.
Importantly, no successful chemical synthesis approaches
to lasso peptides have been reported to date due to the difficulties
[a]
[b]
Dr. H. Martin-Gómez
Institute for Research in Biomedicine
Baldiri Reixac 10, 08028 Barcelona, Spain
Prof. F. Albericio, Dr. J. Tulla-Puche
Department of Inorganic and Organic Chemistry – Organic
Chemistry Section
University of Barcelona
Martí i Franquès 1-11, 08028 Barcelona, Spain
E-mail: judit.tulla@ub.edu
[c]
[d]
[e]
Prof. F. Albericio, Dr. J. Tulla-Puche
CIBER-BBN, Networking Centre on Bioengineering, Biomaterials
and Nanomedicine
Baldiri Reixac 10, 08028 Barcelona, Spain
Prof. F. Albericio
School of Chemistry and Physics
University of KwaZulu-Natal
Durban 4001, South Africa
Figure 1. Structures of the natural lasso peptides (A) sungsanpin 1 and (B)
chaxapeptin 2 (PDB code 2N5C). The ring residues are shown in green ribbon,
the loop amino acids in blue and the tail amino acids in red.
Dr. J. Tulla-Puche
Institut de Biomedicina de la Universitat de Barcelona (IBUB)
08028 – Barcelona, Catalonia, Spain
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/chem.201803899
Chemistry - A European Journal
FULL PAPER
O
i, ii
N
Boc
O
iii, ii
H2N
O
HO
H
N
Fmoc
N
H
O
TentaGel resin
HMPBA linker
O
Fmoc stepwise
peptideelongation
2-PhiPr
O
O
tBu
N
Boc
O
O
O
O
O
O
O
O
O
O
H
N
H
H
H
N
H
H
H
N
HN
Fmoc
N
N
N
N
N
H
N
H
N
N
H
N
H
N
H
N
H
O
O
O
O
O
O
O
O
R
O
tBu
tBu
HN
Boc
iv, v, vi
Figure 2. Schematic representation of the target peptide. Modification between
sungsanpin 1 and the target peptide. The ring residues are shown in green, the
loop amino in blue, the tail amino acids in red and Glu3 in yellow.
H
N
Boc
O
O
N
N
O
N
Boc
O
H
O
O
O
O
HN
NH
NH
H
N
H
N
H
N
O
HN
N
H
N
H
N
H
O
O
O
O
O
HN
O
O
O
tBu
O
H
N
tBu
O
tBu
Results and Discussion
NH
O
R
vii
Towards the synthesis of sungsanpin: design and synthetic
approach
H2
N
O
O
N
N
H
O
NH
O
O
O
O
O
As previously mentioned, once the macrolactam bond is closed,
the C-terminal tail cannot be trapped inside. This was confirmed
with the tentative synthesis of sungsanpin 3 using stepwise
Fmoc-SPPS (Scheme 1). The synthesis was performed on Wang-
HN
NH
NH
H
N
H
N
H
N
O
HN
HO
N
H
N
H
N
H
OH
O
O
O
O
HN
O
OH
HO
H
N
O
NH
R
O
type
resins
(TentaGel
S-NH2)
with
the
3: R
= Gly
3-(4-hydroxymethylphenoxy)propionic acid (HMPBA) linker. In
addition to the tBu based protecting group, the β-carboxylic acid
of Asp8 was protected in form of 2-phenylisopropyl ester (2-PhiPr).
This hindered protecting group can be removed with mild acid
conditions (5% TFA in DCM), and is compatible with Boc and tBu.
2-PhiPr was used to prevent aspartimide formation between Asp8
and Ser9.[12] The cyclization was carried out on resin between the
Nα-amino of the Gly1 and the side-chain carboxyl group of Asp8
released from its 2-PhiPr ester by treatment with a low percentage
of TFA. The main goal of this synthesis was to observe whether
the tail was able to thread into the ring unaided. Once the
cyclization was carried out, a treatment with TFA-TIS-H2O
(95:2.5:2.5) rendered the unprotected peptide. After several
attempts of folding only the structure corresponding to the
branched-cyclic sungsanpin 3 was obtained. Thus, we can
conclude that using a straightforward strategy, which involved
SPPS peptide elongation and cyclization, is not enough for the
obtention of this intriguing peptide.
4: R = Glu
Scheme 1. Solid-phase synthesis toward peptides 3 and 4. (i) Fmoc-L-Leu-OH,
DIPCDI, DMAP, CH2Cl2-DMF (9:1), 2 h + 3 h; (ii) piperidine-DMF (1:4) (1 x 1 min,
1 x 3 min); (iii) Fmoc-L-Trp(Boc)-OH, DIPCDI, OxymaPure, DMF, 30 min; (iv)
TFA-TIS-CH2Cl2 (5:1:94), 4 x 10 min; (v) piperidine-DMF (1:4) (1 x 1 min, 1 x
3 min, 1 x 5 min, 1 x 30 min); (vi) PyBOP, HOAt, DIEA, DMF, 1 h; and (vii) TFA-
TIS-H2O (95:2.5:2.5), 1 h.
Design and synthesis of the lasso-inspired bicyclic peptide
via ester bond
The synthesis of peptide 5 was designed to ensure that the
C-terminal tail was trapped inside the ring. We hypothesized that
this could be done by tying the peptide loop to the ring through a
covalent linkage, creating a bicycle. An ester bond was chosen
because it could be later in principle easily removed while
maintaining the lasso structure. The Gly at position 3 was
replaced by Glu to form the ester bond between the side chain of
Glu3 and the hydroxyl group of Ser13. Two branched peptide
chains—linked by the ester bond—were elongated. One of these
chains corresponded to the ring and the other to the loop. The
linkage of the other extreme of the loop with the ring was then
built, followed by a final cyclization step to close the ring (Figure 3).
For the next sets of experiments, when the bicyclic peptide was
planned to be synthesized having an extra ester bond between
the hydroxyl group of Ser13 and the γ-carboxylic acid of Glu at
position 3 instead of Gly, the corresponding branched-cyclic
analog 4, with the same topology of 3, was synthesized as a
control, following the same procedure described in Scheme 1.
This article is protected by copyright. All rights reserved.

10.1002/chem.201803899
Chemistry - A European Journal
FULL PAPER
The synthesis of this bicyclic peptide 5 represents a “tour de force”
as it will imply the use of several orthogonal protecting groups
(Scheme 2).[13] Furthermore, the sequence Asp-Gly present in the
sequence is very prone to give aspartimide side-reaction.[14] The
synthesis was carried out on HMPBA-Wang-TentaGel resin with
low loading to facilitate both the elongation of the two peptide
chains and the on-resin cyclization. The first peptide chain was
starting from the C-terminal Ile7 and finished at Asp8. Asp is a key
residue in the peptide sequence because it is the only residue
involved in the ring and the loop. For this reason, a unique
protection strategy was designed for this residue: (i) Boc was
used as protecting group of the amino group, because Asp8 was
located at the N-terminal of the first peptide chain elongated. This
position will pair with the carboxyl group of the C-terminal Ile7
which is anchored to the HMPBA linker. Both functions will be
released during the last TFA treatment; (ii) the 2-PhiPr was used
Figure 3. Schematic representation of the main important synthetic steps
toward the bicycle formation of the peptide 5. (i) Elongation of the second
branched peptide chain using Alloc-chemistry, (ii) First cyclization between the
loop and the ring and (iii) cleavage and macrolactamization to close the ring.
The ring residues are shown in green, the loop amino in blue and Glu3 in yellow.
Scheme 2. Synthetic strategy toward the bicyclic peptide 5. (i) Fmoc-L-Ile-OH, DIPCDI, DMAP, CH2Cl2-DMF (9:1), 2 h + 3 h; (ii) piperidine-DMF (1:4) (1 x 1 min,
1 x 3 min); (iii) Fmoc-L-Pro-OH, DIPCDI, OxymaPure, DMF, 30 min; (iv) Boc-L-Asp-O(2-PhiPr) 8, HATU, DIEA, DMF, 1 h; (v) Pd(PPh3)4, PhSiH3, CH2Cl2, 3 x
15 min; (vi) Alloc-L-Ser-OFm 10, DIPCDI, DMAP, CH2Cl2-DMF (9:1), 2 h + 3 h; (vii) Alloc-L-Leu-OH 11, DIPCDI, OxymaPure, DMF, 30 min; (viii) TFA-TIS-H2O
(5:1:94), 4 x 10 min; (ix) PyBOP, HOAt, DIEA, DMF, 2 h; (x) TFA-TIS-H2O (95:2.5:2.5), 1 h; (xi) PyBOP, HOAt, DIEA, CH2Cl2-DMF (1:1), 0 ºC  rT, 2 h; (xii)
DEA, DCM, 1 h; (xiii) H-Trp-Leu-OtBu 15, DIPCDI, OxymaPure, DMF, 3 h; (xiv) TFA-DCM (1:1), 30 min; and (xv) 6 M SnCl2 in DMF, 1.6 mM HCl/dioxane, 2 x
45 min. The ring residues are shown in green, the loop amino in blue, the tail amino acids in red and Glu3 in yellow.
This article is protected by copyright. All rights reserved.

10.1002/chem.201803899
Chemistry - A European Journal
FULL PAPER
to protect the α-carboxylic acid, thus allowing the β-carboxylic
acid to remain free. To prevent aspartimide formation at the
Asp8-Gly1 sequence, 2,4,6-trimethoxybenzyl (Tmb) backbone
protection was used for blocking the NH Gly1. In addition, to
ensure orthogonal deprotection before the formation of the ester
bond, allyl was selected as protecting group of the α-carboxylic
acid group of Glu3. Moreover, the standard side-chain Boc
(Figure S1-S3). The peptide was digested between Asp8-Ser9
and Phe10-Gly11, with the concomitant loss of the dipeptide
Ser9-Phe10 (m/z 1407.5). This result indicated the extraordinary
stability of the bicyclic peptide 5, in which long times were required
before degradation was observed. As expected, the natural lasso
peptide, chaxapeptin 2, remained intact after proteases
treatment.[8] However, there are several thermal sensitive lasso
peptides and also carboxypeptidase Y sensitive in which the
C-terminal residues are degraded. One such example is
sphingopyxin I,[17] in which the last two residues were digested
and the structure unfolds after thermal assay.
protecting
group
of
Lys5
was
substituted
by
p-nitrobenzyloxycarbonyl (pNZ), yielding compound 9.[15] This
modification was performed because the free amino group of the
Lys5 side chain may compete during the second cyclization in
solution, thus producing
a
non-desired cyclization. The
γ-carboxylic acid of Glu3 was protected with allyl, which was
removed before the esterification of that carboxylic acid with the
hydroxyl group of Ser13. Alloc-chemistry was then used for the
elongation of the second peptide chain to avoid the use of
piperidine treatment that can produce aminolysis of the ester bond,
serine N  O transacylation, and DKP formation. For this reason,
MS/MS analysis
Peptides 2-5 were extensively analyzed by tandem mass
spectrometry (Table 1). The fragmentation pattern of 3 and 4 was
very similar (Figure S4-S5), showing only fragmentation around
the tail, while the cycle remained intact. On the other hand,
chaxapeptin 2 showed a very similar fragmentation pattern to that
of the two aforementioned peptides, where b12 was also the most
intense signal (m/z 1188.60) (Figure S6). However, the signal
intensities are different, which can be attributed to the constrained
structures of native lasso peptides.[5],[18],[19]
new protecting scheme
for
Ser13 was
designed.
9-Fluorenylmethyl ester (Fm) was used as protecting group of the
α-carboxylic acid rather than of the amino group, which this time
was protected with Alloc, yielding compound 10.[16]
The formation of the ester bond proceeded via the
O-acylisourea activation.
A high excess of DIPCDI and
Alloc-L-Ser-OFm 10 was used to ensure its fast and efficient
formation. Short reaction times and high excesses were needed
to prevent the N  O transacylated product.
Table 1. Main MS2 fragmentation of peptides 2-5, being the first the most
abundant.
Peptide
2
3 and 4
5
After elongation of the second peptide chain, selective
deprotection of the Alloc and 2-PhiPr groups and addition of
PyBOP-HOAt-DIEA provided the first cyclization. The peptide
was then cleaved, and the second cyclization, corresponding to
the ring closure, was performed in solution in order to provide a
more flexible environment for the peptide backbone during
cyclization. At this point the ring and the loop sequences, linked
by the ester bond, set up a bicycle. The Fm from the α-carboxylic
acid of Ser13 was then removed and the C-terminal dipeptide 15
was introduced in a straightforward way after the bicycle formation.
The main advantage of this last coupling was that the steric
hindrance of the ring was avoided, and the tail could be introduced
through the middle of the ring. Finally, tBu and pNZ protecting
groups were removed using TFA and SnCl2, respectively. A
special beauty of this synthetic approach is that the integrity of the
ester bond during the synthesis was preserved by avoiding the
use of harsh basic conditions, thus using Alloc and pNZ
chemistries, which can be removed smoothly using Pd(0) and
SnCl2.
b12
b12
b13
Leu12-Asn13
Leu12-Ser13
Ser13-Trp14
b11
b13
b2 + y12
Gly11-Leu12
Ser13-Trp14
Ser4-Lys5/Lys5-Pro6
b13
b10
b2 + y11
Asn13-Phe14
Phe10-Gly11
Glu3-Ser4/Lys5-Pro6
b10
b12
Phe10-Gly11
Leu12-Ser13
A totally different fragmentation spectrum was obtained for
bicyclic peptide 5. In this case, fragmentation was observed in the
macrolactam ring (Figure S7). The fragmentation pattern
observed was more similar to class I and III lasso peptides, than
to class II. In this regard, class I and III show a weaker
fragmentation behavior than class II and also fragmentation in the
ring, thus hindering the distinction between a bicyclic and a lasso
peptide.[2],[20]
In conclusion, we were unable to distinguish between our
bicyclic peptide and a lasso-structured peptide based on the MS2
studies, as the weak fragmentation behavior could also be a
consequence of the fact that more than one bond breakage was
required for the fragmentation to be detected.
Thermal and proteolytic stability
Peptides 2-5 were subjected to proteases and thermal stability
assays. The bicyclic peptide 5 remained intact after both thermal
and carboxypeptidase Y treatment, and only the C-terminal Leu15
was digested. In contrast, the corresponding branched-cyclic
topoisomer 4 and 3 were readily degraded, showing cleavage
between Ser9 and Phe10. Regarding this result, 5 was tested
against a stronger protease, pepsin. After 2 h, the peptide
remained intact and after 24 h only 12% had been degraded
IM-MS analysis
Three characteristic IM-MS trends have been established for
class II lasso peptides.[21] Regarding the first trend when the
charge state was increased, the CCS remained relatively
This article is protected by copyright. All rights reserved.

10.1002/chem.201803899
Chemistry - A European Journal
FULL PAPER
constant for peptides 4 and 5 (Table 2). The [M + 3H]3+ was the
most abundant charge state, as it was present for all the peptides,
while the [M + 4H]4+, which is characteristic of unfolded branched
cyclic forms, was present only for 3.[21–23] The similar behavior of
4 and 5 can be related to the stabilization through hydrogen bonds
in the gas-phase of peptide 4.
to reported lasso peptides (Figure S26). The chemical shift
deviation values of αH of 5 did not show a clear tendency
(Figure 4). The negative values for the Gly1-Ser4 segment could
indicate an α-helical fold. These findings may indicate the
presence of α-turns.
The low range of CCS ∆Ω/Ω covered by all protonated
molecules of bicycle 5 is consistent with a rigid and constrained
structure, even the value (0.6 %) is lower value than the natural
lasso peptide 2 (3.9 %) (Table S1).[23]
Table 2. Number of conformations and CCS of the four peptides with 400 mM
of sulfolane.
Charge
state[a]
no. of
Peptide
CCS (Å) (Ratio(%))[b]
conformations
Considering the second trend, low intensity of the highly
protonated species is expected for lasso peptides (Figure S8).
The bicyclic peptide 5 showed a slightly different spectrum after
the addition of sulfolane compared to 2, in which the intensity of
the [M + H]+ peak decreased, being the doubly protonated ion the
most abundant. This decrease was more readily detected for 3, in
which the peak of [M + H]+ dropped considerably after the addition
of sulfolane. Regarding the parameter (ζ),[21],[23]] which provides a
comparison of both the intensity and charge state, the bicycle 5
showed the lowest values, indicating a rigid and constrained
structures (Table S2).
Regarding the width of the ion-mobility peak on the
[M + 3H]3+ charge state, peptides 4 and 5 showed a similar peak
profile than chaxapeptin 2 with narrow IM peaks (Figure S9).
Although the peptide sequence of 4 differed from that of 3 in only
one residue, the shape of the IM peak was more similar to that of
the natural lasso peptide 2 and to the bicyclic peptide 5. This
observation again reflects the importance of the structure
stabilization effect caused by the replacement of Gly3 by Glu3,
which produces a structure that resembles a bicyclic peptide
(Figure S10).
290 (10)/307 (53)/
[M + 2H]2+
[M + 3H]3+
[M + 2H]2+
4
4
3
330 (33)/367 (4)
2
281 (18)/295 (52)/
312 (23)/337 (8)
346 (39)/328 (18)/
312 (43)
388 (22)/352 (63)/
3
[M + 3H]3+
[M + 4H]4+
[M + 2H]2+
3
2
4
310 (15)
391 (25)/426 (75)
297 (24)/321 (43)/
346 (29)/372 (4)
4
5
305 (16)/323 (52)/
[M + 3H]3+
[M + 2H]2+
[M + 3H]3+
4
4
4
337 (23)/357 (9)
306 (10)/319 (61)/
340 (27)/373 (1)
296 (27)/317 (51)/
337 (18)/350 (5)
NMR analysis
[a] Charge state [M + H]+ is not shown because it is an unfolded structure with
the highest drift time values and is not relevant for the comparison. [b] Each
CCS was calculated based on the drift times.[26] When several conformations
are present, the major species is highlighted in bold.
On the basis of the previous results, an NMR analysis was carried
out to assess the 3D structure of peptides 3 and 5. The NMR data
for structural characterization of 3 was recorded in pyriridne-d5
and was compared with those of natural sungsanpin (Table S3).[7]
The general broadness of the spectral lines and the poor
dispersion of amide signals led us to conclude that 3 did not show
a
lasso structure (Figure S11-S14). This result was also
confirmed by the NMR-derived structure (Figure S15). The long-
range NOEs between the ring and the loop revealed that the linear
part was folded over the ring, but with no threading.
To achieve better resolution, we tested two solvents. To this
end, the NMR data of the bicyclic peptide 5 was recorded in
DMSO-d6
(Figure
S16-S22)
and
CD3CN-H2O
(1:5)
(Figure S23-S25). Duplicity in the residues Trp14, Leu15 and
Pro6 was observed, indicating that two conformations were
present. On the one hand, this duplicity could be assigned to the
cis/trans isomerization. [24] On the other hand, the duplicity of the
C-terminal residues (Trp14 and Leu15) pointed to high flexibility
on this part, which is not constrained by the bicyclic structure.
Exhaustive analysis of the fingerprint region of bicyclic
peptide 5 revealed poor 1H signal dispersion in the amide proton
region compared to lasso peptides.[25] In this regard, the chemical
shift dispersion (∆δ) of the bicyclic peptide 5 was low compared
Figure 4. Chemical shift deviations histogram from random coil for ΔHα in the
bicyclic peptide 5 in DMSO-d6. Random coil values were extracted from
Wüthrich et al.[27,28]
This article is protected by copyright. All rights reserved.

10.1002/chem.201803899
Chemistry - A European Journal
FULL PAPER
Regarding the long-range NOEs between the C-terminal tail
and the N-terminal macrolactam ring, bicyclic peptide 5 show
quite sequential NOEs and few of them were located between the
tail and the ring. A better dispersion was achieved in CD3CN-H2O
(1:5) than in DMSO-d6, thus allowing the measurement of some
long-range NOEs (Table 3). Contrary to what we expected, the
rigid structure of the bicycle was not enough to provide an
organized and structured spectrum.
fragmentation, ion-mobility and protease assays, confirmed that
bicyclic peptide
Table 1. Long-range NOEs between the C-terminal tail and the N-terminal ring
of the bicyclic peptide 5.
NOE cross-peaksa
NH Phe2 - Hβ, Hγ, Hδ Leu12
Hα Pro6 - Hγ, Hδ Leu12
NH2 Lys5 – Hδ Leu12
H2 Trp14 – Hβ Ser9; Hα, Hβ Phe2; Hγ Leu12
[a] From the NOESY spectrum in CD3CN-H2O (1:5) at 298 K.
Computational studies
To determine the 3D structure of the bicyclic peptide 5, a
conformational sampling algorithm was performed with the NMR-
derived distance restraints (see Experimental Section for more
details). To verify the compatibility of the NOE restraints, two
possible structural conformations (threaded and unthreaded)
were designed as starting point (Figure S27). Although the long-
range NOEs were not conclusive, they were enough to guide the
simulation of the bicyclic 3D structure. The initial and final
structures were different for the threaded bicycle (RMSD = 6.3 Å),
thereby revealing the incompatibility of the NOE-derived restraints
with the threaded structure. On the other hand, the low deviation
value (RMSD = 2.7 Å) showed that the starting structure with the
unthreaded tail was closer to the one modeled with restraints.
From the two starting conformations, the resulting minimal-energy
structures displayed similar features and high similarity
(RMSD = 4.2 Å) (Figure 5 b). Moreover, all the calculated
minimal-energy conformations led to the same bicyclic topology,
with the C-terminal tail outside the macrocyclic ring (Figure 5 a).
Figure 5. 3D structure of the bicyclic peptide 5. (A) Ensemble of the 20-lowest
energy structures derived from the NMR restraints in CD3CN-H2O (1:5): Above
- Starting structure with the threaded tail and below - with the unthreaded tail.
The maximum RMSD values between structures 1-20 are 0.6 and 0.4 Å,
respectively. (B) Lowest-energy structures for the threaded and unthreaded
starting conformation, from left to right. The ester bond is shown in yellow and
the macrolactam bond in pink. (C) Chemical structure and schematic
representation.
5 had the C-terminal tail outside the ring. The covalent linkage
between the tail and the ring is responsible for providing
proteolytic stability but was not sufficient to provide a well-defined
structure. Despite the topology, the low collision cross section
values and the narrow peak width observed for bicyclic peptide 5
indicated that it showed lasso-like behavior. Moreover,
preliminary biological assays demonstrated that bicycle 5 showed
similar cytotoxicity and invasion activity than chaxapeptin 2
(Figure S28-S29). On the other hand, the branched-cyclic peptide
4 displayed similar ion-mobility behavior as 5 with the electrostatic
interaction Glu3-Lys5 still present in the gas phase, thus
stabilizing the bicyclic structure. Despite this conformation,
peptide 4 did not show proteolytic stability. Given the singular
characteristics of 5, it may be a good candidate as molecular
scaffold for epitope grafting toward the generation of more potent
and more selective bioactive compounds. Finally, this synthesis
is expected to facilitate the future development of peptides as
therapeutics tools
Conclusions
In summary, a combination of a solid-phase/solution chemistry
hybrid approach allowed the synthesis of a bicyclic peptide
inspired by the natural lasso peptide sungsanpin. The key of the
synthetic approach relies in a proper selection of large number of
protecting groups: (i) Alloc, Boc, Fmoc and pNZ for the amino
function; (ii) allyl, Fm, 2-PhiPr, tBu and the oxymethyl from the
HMPBA linker for the carboxylic function and (iii) Tmb for the
amide of Gly. The combined use of these 10 protecting groups,
which removals require mostly orthogonal conditions (low and
high percentage of TFA, piperidine, Pd(0) and SnCl2), has allowed
the synthesis of this bicycle. The NMR-based experiments and
the structural calculations, together with the results of the MS2
This article is protected by copyright. All rights reserved.

10.1002/chem.201803899
Chemistry - A European Journal
FULL PAPER
500 μL of water (100 μM). Sample were diluted 1/10 in H2O-ACN (1:1) with
4% (FA) to obtain 10 μM, 1/10 in H2O-ACN (1:1) with 4% FA and 200 mM
sulfolane or 1/10 in H2O/ACN (1:1) with 4% FA and 200 mM or 400 mM
sulfolane. Sample solutions were directly injected and were infused by
automated chip-based nanoelectrospray using a Triversa Nanomate
system (AdvionBioSciences, Ithaca, NY, USA) as the interface. The
ionization was performed in positive mode using a spray voltage and a gas
pressure of 1.75 kV and 0.5 psi, respectively. Cone voltage, extraction
cone and source temperature were set to 40 V, 3 V and 20 ºC, respectively.
Trap and transfer collision energies were set to 6 V and 4 V, respectively.
The pressure in the Trap and Transfer T-Wave regions were
5.99·102 mbar of Ar and the pressure in the IMS T-Wave was 0.477 mbar
of N2. Trap gas and IMS gas flows were 8 and 25 mL/sec, respectively.
The travelling wave used in the IMS T-Wave for mobility separation was
operated at a velocity of 250 m/sec. The wave amplitude was a linear ramp
from 1.0 to 28.0 V. The bias voltage for entering in the T-wave cell was
15 V. The instrument was calibrated over the m/z range 200-3000 Da
using a solution of cesium iodide. MassLynx version 4.1 SCN 704 and Drift
scope version 2.4 software were used for data processing. Ion mobility
data analysis was performed with Driftscope software vs. 2.4 integrated in
MassLynx software.
Experimental Section
Materials and reagents
Dichloromethane, dimethylformamide, methanol, diethyl ether were
purchased from Scharlau; while acetone and acetonitrile were obtained
from SDS. All the reagents used for the synthesis were supplied by the
following: Bachem AG, Iris Biotech, Luxemburg Industries, Novabiochem,
Panreac and Sigma-Aldrich. Carboxypeptidase Y was purchased from Alfa
Aesar.
Analytics
Analytical RP-HPLC was performed on a Waters Alliance 2695 (Waters,
Milford, MA) with an automatic injector and a photodiode array detector
(Waters 2998 or Waters 996) and software Empower2. The columns used
were
a Xbridge™ C18 2.5 µm (4.6 mm x 75 mm) reversed-phase
analytical column and a Sunfire™ C18 3.5 µm (4.6 mm x 100 mm)
reversed-phase analytical column run with linear gradients of ACN
(0.036% TFA) into H2O (0.045% TFA) over 8 min, and a Phenomenex
Luna C18 5 µm (4.6 mm x 250 mm) reversed-phase analytical column run
with linear gradients over 40 min. UV detection was at 220 nm and the
system was run at a flow rate 1 mL/min. For semi-analytical HPLC, the
columns used were a Xbridge™ Prep BEH130 C18 5 µm (10 mm x 100
mm) reversed-phase analytical column and a Sunfire® C18 OBD™ Prep
Column 5 μm (10 x 150 mm) reversed-phase column. Linear gradients of
ACN (0.036% TFA) into H2O (0.045% TFA) were run over 20 min. UV
detection was at 220 nm and the system was run at a flow rate 3 mL/min.
Semi-preparative RP-HPLC was carried on Waters instrument comprising
a 2545 gradient module equipped with a sample manager module with
automatic injector and fraction collector (Waters Alliance 2767), a Waters
Alliance 2487 dual wavelength absorbance detector and a MassLynx v3.5
system controller. The column used was a XBridge® Prep C18 OBD™
5 μm (19 x 100 mm) reversed-phase column. UV detection was at 220 and
254 nm and linear gradients of ACN (0.1% TFA) into H2O (0.1% TFA) were
run at a flow rate of 16 mL/min over 30 min.
NMR studies
1H and 13C NMR spectra were recorded on a Varian MERCURY 400
(400 MHz for 1H NMR, 100 MHz for 13C NMR) spectrometer for organic
small molecules and on a Bruker 600 Avance II Ultrashield, provided with
a cryoprobe TCI (600 MHz for 1H NMR, 150 MHz for 13C NMR)
spectrometer for peptidic samples. Chemical shifts (δ) are expressed in
parts per million (ppm) downfield tetramethylsilane (TMS). Coupling
constants are expressed in Hertz (Hz). Xaa-Pro bond conformation was
determined using a previously described method based on the 13C
chemical shift statistics.[24]
Computational studies
They were carried out using the Macrocyclic Conformational Sampling
tool.[29] The coordinates for the peptides were extracted from the crystal
structure of chaxapeptin 2 (PDB code: 2N5C). Then, Leu7, Asn13, Phe14
and Phe15 were manually mutated to Ile, Ser, Trp and Leu, respectively
for all peptides. One additional mutation for 3 and 5 was required, the
exchange of Gly3 for Glu3. Bicyclic peptide 5 was manually cyclized
through the side chain of Asp8 and the ester bond through the side chains
of Glu3 and Ser13. Additionally, the tail was manually unthreaded from the
ring in the other bicyclic conformation, used as a starting structure. On the
other hand, for peptides 3 and 4, the tail was unthreaded and was cyclized
only through the side chain of Asp8. The resulting structures were used as
input for the Macrocycle Conformational Sampling tool of Schrödinger.
Constraints from the NOESY cross-peaks were applied to each structure
(Table S4). NOEs were classified as strong, medium and weak (upper
limits for structure calculation were set as 2.2 Å, 3.0 Å and 4.0 Å,
respectively). Conformations were generated using the OPLS-2005 force
field and GBSA for electrostatic treatment. Conformations were kept when
energies were below 10 kcal/mol, and redundant conformations (RMSD >
0.75 Å) were removed. For each peptide, 5000 simulation cycles were
used with 5000 LLMOD search steps.
MS-MS fragmentation
MS2 was performed on LTQ-FT Ultra mass spectrometer (Thermo
Scientific). 0.1 mg was reconstituted with 500 μL of water (100 μM). 10 μL
(100 μM) of sample was diluted 1/1 with ACN (1% FA) to obtain 20 μL of
50 μM. Sample solutions were directly injected and were infused by
automated chip-based nanoelectrospray using
a Nanomate system
(AdvionBioSciences, Ithaca, NY, USA) as the interface. The ionization was
performed in positive mode using a spray voltage and a gas pressure of
1.75 kV and 0.5 psi, respectively. Capillary temperature and voltage were
set up to 200 ºC and 44 V, respectively. The fragmentation technique used
was CID (collision induced dissociation). The instrument was calibrated
over the m/z range 200-2000. Data were acquired with Xcalibur software,
vs.2.0SR2 (ThermoScientific).
IM-MS
IM-MS experiments were performed using a Synapt G1-HDMS mass
spectrometer (Waters, Manchester, UK). 0.1 mg was reconstituted with
This article is protected by copyright. All rights reserved.

10.1002/chem.201803899
Chemistry - A European Journal
FULL PAPER
NMR 2002, 24, 149–154.
Acknowledgements
[25]
[26]
X. Xie, M. A. Marahiel, ChemBioChem 2012, 13, 621–625.
D. Smith, T. Knapman, I. Campuzano, R. Malham, J. Berryman, S.
Radford, A. Ashcroft, Eur. J. Mass Spectrom. 2009, 15, 113.
C. Grathwohl, K. Wüthrich, J. Magn. Reson. 1974, 13, 217–225.
A. Bundi, C. Grathwohl, J. Hochmann, R. M. Keller, G. Wagner, K.
Wüthrich, J. Magn. Reson. 1975, 18, 191–198.
The authors wish to thank Dr. Margarida Gairí and Dr. Mireia
Díaz-Lobo for helpful discussions on NMR and Ion Mobility data
analysis, respectively. This study was partially funded by the
CICYT (CTQ2015-67870-P to F.A., and CTQ2015-68677-R and
EUIN2017-88320 to J.T.-P, and the Institute for Research in
Biomedicine Barcelona (IRB Barcelona). H.M.-G. and J.T.-P.
thank a MINECO for a Severo Ochoa Predoctoral fellowship and
a Ramon y Cajal contract, respectively.
[27]
[28]
[29]
K. S. Watts, P. Dalal, A. J. Tebben, D. L. Cheney, J. C. Shelley, J.
Chem. Inf. Model. 2014, 54, 2680–2696.
Conflicts of interest
There are no conflicts to declare
Keywords: lasso peptides • bicyclic peptides • solid-phase
peptide synthesis • anticancer • orthogonal protection
[1]
[2]
H. Martin-Gómez, J. Tulla-Puche, Org. Biomol. Chem. 2018, 16,
5065–5080.
Y. Li, R. Ducasse, S. Zirah, A. Blond, C. Goulard, E. Lescop, C.
Giraud, A. Hartke, E. Guittet, J. L. Pernodet, et al., ACS Chem. Biol.
2015, 10, 2641–2649.
[3]
A. Blond, J. Péduzzi, C. Goulard, M. J. Chiuchiolo, M. Barthélémy,
Y. Prigent, R. A. Salomón, R. N. Farías, F. Moreno, S. Rebuffat,
Eur. J. Biochem. 1999, 259, 747–755.
[4]
[5]
M. Zimmermann, J. D. Hegemann, X. Xie, M. A. Marahiel, Chem.
Sci. 2014, 5, 4032–4043.
K. J. Rosengren, R. J. Clark, N. L. Daly, U. Göransson, A. Jones, D.
J. Craik, J. Am. Chem. Soc. 2003, 125, 12464–12474.
F. Saito, J. W. Bode, Chem. Sci. 2017, 8, 2878–2884.
S. Um, Y.-J. J. Kim, H. H. C. Kwon, H. Wen, S.-H. H. Kim, H. H. C.
Kwon, S. Park, J. Shin, D.-C. C. Oh, J. Nat. Prod. 2013, 76, 873–
879.
[6]
[7]
[8]
S. S. Elsayed, F. Trusch, H. Deng, A. Raab, I. Prokes, K.
Busarakam, J. A. Asenjo, B. A. Andrews, P. Van West, A. T. Bull, et
al., J. Org. Chem. 2015, 80, 10252–10260.
[9]
I. Di Bonaventura, X. Jin, R. Visini, D. Probst, S. Javor, B.-H. Gan,
G. Michaud, A. Natalello, S. M. Doglia, T. Köhler, et al., Chem. Sci.
2017, 8, 6784–6798.
[10]
A. L. Ferguson, S. Zhang, I. Dikiy, A. Z. Panagiotopoulos, P. G.
Debenedetti, A. James Link, A. J. Link, Biophys. J. 2010, 99, 3056–
65.
[11]
[12]
R. B. Merrifield, J. Am. Chem. Soc. 1963, 85, 2149–2154.
J. L. Lauer, C. G. Fields, G. B. Fields, Lett. Pept. Sci. 1995, 1, 197–
205.
[13]
[14]
F. Albericio, Biopolymers 2000, 55, 123–139.
E. Nicolás, E. Pedroso, E. Girald, Tetrahedron Lett. 1989, 30, 497–
500.
[15]
[16]
[17]
[18]
[19]
[20]
[21]
A. Isidro-Llobet, J. Guasch-Camell, M. Álvarez, F. Albericio,
European J. Org. Chem. 2005, 2005, 3031–3039.
L. J. Cruz, N. G. Beteta, A. Ewenson, F. Albericio, Org. Process
Res. Dev. 2004, 8, 920–924.
J. D. Hegemann, M. Zimmermann, S. Zhu, D. Klug, M. A. Marahiel,
Biopolymers 2013, 100, 527–542.
J. D. Hegemann, M. Zimmermann, X. Xie, M. A. Marahiel, J. Am.
Chem. Soc. 2013, 135, 210–222.
T. a. Knappe, U. Linne, S. Zirah, S. Rebuffat, X. Xie, M. a. Marahiel,
J. Am. Chem. Soc. 2008, 130, 11446–11454.
T. A. Knappe, U. Linne, X. Xie, M. A. Marahiel, FEBS Lett. 2010,
584, 785–789.
K. Jeanne Dit Fouque, C. Afonso, S. Zirah, J. D. Hegemann, M.
Zimmermann, M. A. Marahiel, S. Rebuffat, H. Lavanant, Anal.
Chem. 2015, 87, 1166–1172.
[22]
[23]
[24]
K. Jeanne, D. Fouque, H. Lavanant, S. Zirah, J. Lemoine, S.
Rebuffat, J. Tabet, A. Kulesza, C. Afonso, P. Dugourd, et al., Rapid
Commun. Mass Spectrom. 2015, 1411–1419.
K. J. D. Fouque, H. Lavanant, S. Zirah, J. D. Hegemann, M.
Zimmermann, M. A. Marahiel, S. Rebuffat, C. Afonso, J. Am. Soc.
Mass Spectrom. 2017, 28, 315–322.
M. Schubert, D. Labudde, H. Oschkinat, P. Schmieder, J. Biomol.
This article is protected by copyright. All rights reserved.

10.1002/chem.201803899
Chemistry - A European Journal
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
Layout 1:
FULL PAPER
The proper selection of large number of protecting
groups together with the combination of solid-
phase/solution chemistry hybrid approach,
allowed the chemical synthesis of a bicyclic
peptide inspired by the natural lasso peptide
sungsanpin.
Helena Martin-Gómez,
Fernando Albericio, Judit
Tulla-Puche*]
Page No. – Page No.
A Lasso-Inspired Bicyclic
Peptide: Synthesis,
Structure and Properties
[a]
[b]
Dr. H. Martin-Gómez
Institute for Research in Biomedicine
Baldiri Reixac 10, 08028 Barcelona, Sp
Prof. F. Albericio, Dr. J. Tulla-Puche
Department of Inorganic and Organic C
Chemistry Section
University of Barcelona
Martí i Franquès 1-11, 08028 Barcelona
E-mail: judit.tulla@ub.edu
This article is protected by copyright. All rights reserved.