Thiazole-based γ-building blocks as reverse-turn mimetic to design a gramicidin S analogue: Conformational and biological evaluation

Article abstract of DOI:10.1002/chem.201402190

This paper describes the ability of a new class of heterocyclic γ-amino acids named ATCs (4-amino(methyl)-1,3-thiazole-5-carboxylic acids) to induce turns when included in a tetrapeptide template. Both hybrid Ac-Val-(R or S)-ATC-Ile-Ala-NH₂ sequences were synthesized and their conformations were studied by circular dichroism, NMR spectroscopy, MD simulations, and DFT calculations. It was demonstrated that the ATCs induced highly stable C₉ pseudocycles in both compounds promoting a twist turn and a reverse turn conformation depending on their absolute configurations. As a proof of concept, a bioactive analogue of gramicidin S was successfully designed using an ATC building block as a turn inducer. The NMR solution structure of the analogue adopted an antiparallel β-pleated sheet conformation similar to that of the natural compound. The hybrid α,γ-cyclopeptide exhibited significant reduced haemotoxicity compared to gramicidin S, while maintaining strong antibacterial activity.

Full text of DOI:10.1002/chem.201402190

DOI: 10.1002/chem.201402190
Full Paper
&
Amino Acids
Thiazole-Based g-Building Blocks as Reverse-Turn Mimetic to
Design a Gramicidin S Analogue: Conformational and Biological
Evaluation
[a]
[a]
[a]
[a]
¨
Baptiste Legrand, Loıc Mathieu, Aurꢀlien Lebrun, Soahary Andriamanarivo,
Vincent Lisowski,[a] Nicolas Masurier,[a] Sꢀverine Zirah,[b] Young Kee Kang,[c] Jean Martinez,[a]
and Ludovic T. Maillard*[a]
Abstract: This paper describes the ability of a new class of
heterocyclic g-amino acids named ATCs (4-amino(methyl)-
1,3-thiazole-5-carboxylic acids) to induce turns when includ-
ed in a tetrapeptide template. Both hybrid Ac-Val-(R or S)-
ATC-Ile-Ala-NH2 sequences were synthesized and their con-
formations were studied by circular dichroism, NMR spec-
troscopy, MD simulations, and DFT calculations. It was dem-
onstrated that the ATCs induced highly stable C9 pseudo-
cycles in both compounds promoting a twist turn and a re-
verse turn conformation depending on their absolute config-
urations. As a proof of concept, a bioactive analogue of gra-
micidin S was successfully designed using an ATC building
block as a turn inducer. The NMR solution structure of the
analogue adopted an antiparallel b-pleated sheet conforma-
tion similar to that of the natural compound. The hybrid a,g-
cyclopeptide exhibited significant reduced haemotoxicity
compared to gramicidin S, while maintaining strong antibac-
terial activity.
have been characterized,[3] while homo- or heterogeneous se-
quences incorporating g-amino acids have received far less at-
tention due, in part, to the difficulty to access stereochemically
pure g-building blocks.[3e,f,4] Mann and Kessler reported the first
example of (hetero)aromatic-based g-amino acids in which the
ring is an oxazole. They suggested that when included into
a small peptide, such a g-amino acid building block induced
a C9 turn.[4d] However, the turn conformation was not fully
characterized.
Introduction
Among the various secondary structures, reverse turns are one
of the major structural elements in biologically active peptides
and globular proteins. They are key features for protein folding
by pre-organizing polypeptide chains and are involved in
many of the molecular recognition events in biological sys-
tems.[1] They are also an essential part of antimicrobial pep-
tides that contain b-hairpins or b-sheets, for example, prote-
grin, tachyplesin, defensins and gramicidin S.[2] In this context,
numerous efforts have been made to develop small-turn in-
ducers. A large variety of scaffolds derived from b-amino acids
We recently described a new class of constrained heterocy-
clic g-amino acids built around a thiazole ring, that is, 4-amino-
(methyl)-1,3-thiazole-5-carboxylic acids (ATCs, 1a and 2;
Figure 1). We demonstrated the high propensity of ATC oligo-
mers to adopt a C9 helical structure in solid and solution states
in both organic solvents and water.[5] We now report the syn-
thesis of a-a-ATC-a hybrid tetrapeptides and their structural
studies by circular dichroism (CD), NMR spectroscopy, molecu-
lar dynamic simulations, and DFT calculations. To investigate
the impact of the ATC configuration on the peptide shape, we
prepared the two diastereomers 3a and 3b consisting in Ac-
Val-(S or R)ATC-Ile-Ala-NH2 (Figure 1). Based on our structural
data, the (R)-ATC motif was then selected as a surrogate for
the d-Phe-Pro turn to synthesize a biologically active analogue
of the gramicidin S (GS; Figure 1), a highly haemolytic b-hairpin
cyclodecapeptide antibiotic. The antimicrobial and haemolytic
activities of analogue 4 (Figure 1) were tested and its NMR so-
lution structure was determined and compared to the charac-
teristic b-pleated sheet of the GS.
[a] Dr. B. Legrand,+ L. Mathieu,+ A. Lebrun, S. Andriamanarivo, Prof. V. Lisowski,
Dr. N. Masurier, Prof. J. Martinez, Dr. L. T. Maillard
Institut des Biomolꢀcules Max Mousseron
UMR 5247, CNRS, Universitꢀs Montpellier I et II
UFR des Sciences Pharmaceutiques et Biologiques
15 Avenue Charles Flahault, 34093 Montpellier Cedex 5 (France)
Fax: (+33) 4-67-54-86-54
E-mail: ludovic.maillard@univ-montp1.fr
[b] Dr. S. Zirah
Musꢀum National d’Histoire Naturelle (MNHN)
UMR 7245 CNRS/MNHN
Molꢀcules de Communication et Adaptation des Micro-organismes (MCAM)
57, rue Cuvier, 75005 Paris (France)
[c] Prof. Y. K. Kang
Department of Chemistry and BK21 PLUS Research Team
Chungbuk National University
Cheongju, Chungbuk 361-763 (Republic of Korea)
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201402190.
Chem. Eur. J. 2014, 20, 6713 – 6720
6713
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
proton resonances in [D6]DMSO
between 293 and 333 K (Figur-
es S2 and S3 in the Supporting
Information). For 3a, we could
dissociate two groups of HN.
While Ile, Ala-HN, and one
proton of the NH2 group exhibit-
ed small solvent chemical shifts
dependency (Dd<0.9 ppm), the
other amide protons were acces-
sible to the solvent according to
their significant resonance varia-
tions upon [D6]DMSO addition
(Dd>1.4 ppm). For 3b, the sol-
vent variations were globally less
significant. We observed that the
Ala- and Ile-HN resonances were
less sensitive to [D6]DMSO (Dd<
0.6 ppm) than those of Val- and
ATC-HN, supporting the idea
that they might be hydrogen
bonded. As formerly observed
for the ATC oligomers, the HN
resonances of both 3a and 3b
Figure 1. Structures of A) the ATC building blocks, B) of the a-a-(R or S)-ATC-a hybrid tetrapeptides 3a and 3b
and C) of gramicidin S (GS) and ATC-based gramicidin S analogue 4.
Results and Discussion
were highly temperature dependent in [D6]DMSO (À4.0ꢀDd/
DTꢀÀ4.6 ppbKÀ1, Table S8 in the Supporting Information),
suggesting all the HN protons were exposed to the solvent.
Nonetheless, temperature coefficient values could be highly in-
fluenced by deshielding effects from surrounding residues
and/or by a partial temperature dependent loss of structure in
small peptides.[6] The thiazole rings of the ATC backbone in-
duced strong magnetic anisotropic effects, which highly influ-
enced the neighboring proton chemical shifts (Ile-HN 9.72 and
9.94 ppm in CD3OH) and prevented the interpretation of the
variable-temperature experiments.
Syntheses and NMR studies of hybrid tetrapeptides 3a and
3b
Peptides 3a and 3b were synthesized by solid-phase peptide
techniques on rink amide resin starting from ATC building
blocks (S or R)-1a using the Fmoc strategy (Fmoc=9-fluorenyl-
methoxycarbonyl). NMR spectra of peptides 3a and 3b (5 mm)
were recorded in CD3OH and [D6]DMSO at 298 K. Whatever the
solvent considered, protons resonances of the C-terminal
moiety (Ile-Ala) of 3a and 3b were remarkably similar (DdH <
0.05 ppm), while significant variations were observed for the
Ile-HN, ATC-HN, ATC-Hg protons and the Val resonances, sug-
gesting the ATC-Cg configuration mainly affected the confor-
mation of the N-terminal part. The amide proton signals that
followed the ATCs (i.e., Ile) in 3a and 3b were strongly shifted
downfield at 9.72 and 9.94 ppm, respectively, in CD3OH (9.32
and 9.30 ppm, respectively, in [D6]DMSO) and both peptides
exhibited characteristic strong ROE correlations (ROE=rotating
frame Overhauser effect) between Ile-HN and ATC-Hg. Interest-
ingly, such unshielded amide resonances along with strong
HN(i+1), Hg(i) correlations were typical of the poly-ATC 9-helix,
supporting the presence of a (Ile)NH···OC(Val) hydrogen bond
forming a C9 pseudocycle. Moreover, the CD spectra of 3a and
3b in methanol exhibited quasi-symmetric signatures very
close to those previously reported for a (S,S)-ATC dimer, which
displayed a typical C9 turn.[5] Such a nine-membered ring was
fully characterized in the crystal structure and NMR spectra of
the poly-ATC compounds.[5]
ROEs were used as constraints for the NMR structure calcula-
tions of 3a and 3b in CD3OH using a typical simulated anneal-
ing protocol in vacuum with AMBER 11. Figure 2b shows a su-
perimposition of the 20 lowest-energy structures of 3a and
3b. The two diastereomers exhibited well-defined structures
stabilized by a C=O(i)···HN(i+2) hydrogen bond between the
Ile-HN and the Val-CO, forming an ATC-residue-driven C9 pseu-
docycle. Additionally, compound 3b exhibited a subsequent
hydrogen bond between the Ala-HN and the Val-CO to form
a C9/12 bidentate pseudocycle. The two peptides also displayed
lateral hydrogen bonds, which formed C10 or C7 pseudocycles
on either side of the ATC residue. The torsion angle values of
the ATC (Table S23 in the Supporting Information) were com-
parable to those measured in poly-(S)-ATC oligomers[5] ((S)-ATC
in 3a: f=À54Æ28, q=113 Æ58, z=À3Æ18, y=À14Æ38; (R)-
ATC in 3b: f=57Æ28, q=À105Æ28, z=3Æ28, y=10Æ28;
poly-(S)-ATC XRD structure: f=À78Æ38, q=127Æ148, z=0Æ
38, y=À41Æ48), while the orientation of the Val N-terminal
residue was dependent of the configuration of the ATC-Cg.
Thus, (S)-ATC promoted a “twist”-turn conformation, while (R)-
ATC induced a reverse turn like structure (Figure 2).
Hydrogen bonding was evaluated by DMSO-titration experi-
ments with gradual addition of [D6]DMSO to a solution of 3a
and 3b in CDCl3 (Figure S1B in the Supporting Information)
and by measuring the temperature coefficients of amide
Chem. Eur. J. 2014, 20, 6713 – 6720
www.chemeurj.org
6714
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
tions of the ATC residues were unimodal and narrow (Figure S5
in the Supporting Information). For the other residues, values
were relatively low dispersed with minor populations, which
explained the small variations of the RMSD. Only the y1 dihe-
dral angle distributions of Val were bimodal (Figure S5 in the
Supporting Information), reflecting the larger fluctuation of the
C-terminal atoms and explaining the small jumps on RMSDs
(Figure S4 in the Supporting Information). The conformational
behavior of 3a and 3b was studied by monitoring hydrogen-
bond patterns over the simulations. With 63 and 75% occur-
rence, respectively, the C9 pseudoring and the lateral C10 lateral
pseudocycle in the “twist”-turn conformation 3a imposed by
(S)-ATC were moderately stable (Figure 2d, Table S22 in the
Supporting Information). However, the C9/12 hydrogen bonds
stabilizing the reverse turn in 3b induced by the (R)-ATC resi-
due were more stable with a percentage of occupancy of 99
and 93%, respectively, during the 50 ns. The other hydrogen
bonds were transients with a low occurrence <16%.
During the unrestrained simulations, RMSFs were globally
larger. The initial NMR conformation of 3b was remarkably
maintained (RMSD=0.61Æ0.17 ꢂ) during the 50 ns MD, con-
firming the high stability of the folding. The C9/12 bidentate hy-
drogen bonds were preserved with 93 and 84% occupancy
over the whole simulation. The most significant change oc-
curred on the y1 dihedral angle distribution of Val (Figure S5 in
the Supporting Information) without impacting the peptide
global shape. By comparison, after 1.5 ns, 3a underwent a tran-
sition from a twist- to a reverse-turn structure stabilized by a bi-
dentate C9/12 pseudo cycle, mirroring the 3b conformation.
This structural rearrangement was associated with concomitant
rotations of y1, f2, and f3 dihedral angles (from 50 to À78,
À58 to À1128, and À59 to À1358, respectively), but was not
compatible with NOE correlations observed between ATC-Hg
and Ile side chain protons. Although NMR spectroscopy could
not ascertain the reverse-turn structure, since no characteristic
NOEs could be found, the fast interconversion between the
twist- and reverse-turn conformations in solution could not be
excluded. Finally, in order to estimate the relative stabilities of
the twist- and reverse-turn structures of 3a, the MD structures
were optimized at the B3LYP/6-31G(d) level of theory. The
single-point energies of these two turn structures were calcu-
lated at the same level of theory with the conductor-like polar-
izable continuum model (CPCM)[8] in methanol. The twist-turn
structure was found to be more stable by 3.40 kcalmolÀ1 than
the reverse-turn structure in methanol, which indicates the rel-
ative populations of 99.7 and 0.3% for twist- and reverse-turn
structures, respectively, at 298 K for 3a.
Figure 2. A) In black, typical NOE correlations stabilizing the C9 pseudo
cycle; in red, principal NOE correlations driving the folding. B) Overlay of the
20 lowest-energy structures of compounds 3a (in green) and 3b (in cyan).
C) Superimposition of the NMR lowest energy structures of 3a (in green)
and 3b (in cyan). D) Occurrence of the hydrogen bonding on 3a (in green)
and 3b over 50 ns rMD (cutoff: dOÀN <3.5 ꢂ and N-H-O angle >1208).
Molecular dynamic simulations
To get further insight on the stability of these structures and
to investigate the set of conformations that 3a and 3b could
adopt in solution, molecular dynamic simulations were ach-
ieved in a methanol box for 50 ns, starting from their lowest
NMR structures. Simulations were launched using the inter-res-
idue NOE-derived upper distances with (rMD) and without re-
straints (MD). The stability of each peptide folding during the
simulations was assessed by computing: 1) the atom-positional
root-mean-square deviation from the initial structure (RMSD);
2) the root-mean-square fluctuation (RMSF) values of the back-
bone heavy atoms (Figure S4 in the Supporting Information);
3) the occupancy of the hydrogen bonds along the simula-
tions; and 4) the backbone dihedral angle distributions (Fig-
ure S5 in the Supporting Information).[7] The backbone RMSD
were rather low over the restrained MD for both compounds
(i.e., (0.43Æ0.22) and (0.68Æ0.22) ꢂ for 3a and 3b, respective-
ly) showing the two peptides conserved their overall confor-
mations in methanol all along simulations. This was also re-
flected by the small RMSF values, which showed small oscilla-
tions on the central ATC and Ile residues, while the N and C
termini (Val and Ala residues) underwent higher mobility for all
trajectories. In both 3a and 3b the torsional angle distribu-
DFT calculations
The stability of the C9 and C9/12 hydrogen bond(s) in 3a and
3b were also investigated by DFT calculations by using Gaussi-
an 03.[9] The two lowest-energy NMR structures of 3a and 3b
were optimized at the B3LYP/6-31G(d) level of theory. After op-
timization, both 3a and 3b displayed the C9 hydrogen bond
around the g-amino acid. The C12 hydrogen bond suggested in
3b by NOE data and MD simulations disappeared after optimi-
Chem. Eur. J. 2014, 20, 6713 – 6720
www.chemeurj.org
6715
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
zation. As a consequence the C9 C=O(i)···HN(i+2) hydrogen and al.[15] A single set of signals was observed in 4 for the Val,
bond remained the major peptide folding driving force. The Orn, Leu, and (R)-ATC suggesting that it also exhibited a C2-
relative strength of the C9 hydrogen bonds of the optimized symmetric structure. The amide protons following the ATC resi-
structures was more deeply evaluated via the natural bond or- due (i.e., Val HN) were strongly unshielded (9.28 ppm) as previ-
bital (NBO) analysis[10] at the B3LYP/6-31+G(d,p) level of ously described for ATC oligomers and compounds 3a and 3b.
theory (Table 1). Compounds 3a and 3b shared similar values The Val/Orn/Leu had relatively high 3J(HNHa) coupling con-
stants (8.0, 8.0, and 8.5 Hz, respectively) compatible with an ex-
tended conformation, but significantly lower than those of the
Table 1. Relative strength of the hydrogen bonds calculated using the
NBO method at the B3LYP/6-31+G(d,p)//B3LYP/6-31G(d) level of theory
[units for atomic charges and Wiberg index in electrons, and the stabiliza-
cients exhibited comparable trends than in GS, but with small-
tion energies (DE2) are in kcalmolÀ1].
GS. The corresponding f torsional angles might display more
degrees of freedom than in the natural GS. Temperature coeffi-
er values. The Orn and ATC-NH shared the higher values (Dd/
[b]
DT=À3.8 ppbKÀ1
) compared to the Val HN (Dd/DT=
Natural charges
dO
Wiberg
index[a]
DE2
À2.5 ppbKÀ1) and Leu HN (Dd/DT=0 ppbKÀ1). This suggested
that these two first amide protons were solvent-shielded and
involved in hydrogen bonding. Nevertheless, as previously dis-
cussed for 3a and 3b, the measured values for the amide pro-
tons close to the thiazole rings should be interpreted with cau-
tion. The typical strong ROE correlations between ATC-Hg and
the amide proton of the following residue (i.e., Val-HN) sup-
ported the presence of the ATC-driven C9 pseudocycle. In addi-
tion, unambiguous long-range NOEs were observed between
dH
3a
3b
À0.690
À0.690
+0.465
+0.467
0.705
0.699
16.17
18.02
[a] Wiberg index indicates the strength of the Ile NÀH bond. [b] The
second perturbation energy of the lone pair orbitals of the oxygen with
the corresponding NÀH antibonding orbital, which is called the hyper-
conjugation is due to the charge transfer.[10]
of atomic charges and Wiberg index calculated for Ile NÀH the Val and the Leu on either side of the C9 turn (Table S19 in
bond.[11] However, in accordance to rMD simulations, calcula- the Supporting Information). The solution structures of 4 were
tion of hyperconjugation (DE2) term due to charge transfer solved using 90 distances and 6 torsion angles restraints and
showed that the C9 hydrogen bond was more stable by about exhibited a slightly distorted antiparallel b-pleated sheet stabi-
2 kcalmolÀ1 in the reverse turn 3b than in the “twist”-turn 3a. lized by (R)-ATCs acting as reverse turns with C=O(i)···HN(i+2)
hydrogen bonds (Figure 3a and b). As expected, the ATCs g-
amino acid adopted similar conformations in 3b and 4 (Fig-
Synthesis and structural studies of the ATC-containing gra-
ure 3c). The dissymmetry of the C9-turn of (R)-ATC may explain
micidin S analogue 4
the non-ideal geometry of the sheet compared to that of the
As a proof of concept, we present a successful application of GS stabilized by the d-Phe-Pro dipeptide (C10-turn). The major
the ATC C9 turn for the design of a new derivative of the difference between 4 and GS arose from the flip of the benzyl
highly effective bactericidal gramicidin S (GS).[12,13] GS is an am-
phiphilic C2-symmetrical cyclodecapeptide that adopts a highly
stable antiparallel b-pleated sheet conformation stabilized by
four intramolecular hydrogen bonds. The two strands are
linked by two hydrophobic type II’ b-turns formed by d-Phe-
Pro dipeptides. Detailed structure–activity relationship studies
revealed that a large number of amino acid substitutions are
tolerated,[13] while the stable, antiparallel b-sheet conformation
is considered essential and is maintained among a wide range
of derivatives.[14] Nevertheless, the major limitation in the clini-
cal use of GS is due to its high lysogenic activity.[12] Considering
(R)-ATC as a reverse-turn inducer, (R)-ATC 3c was used as a sur-
rogate for the d-Phe-Pro sequence to design the GS analogue
4. The linear peptide was obtained by classical Fmoc solid-
phase peptide synthesis then cyclized under high dilution con-
ditions with HBTU/HOBt as a coupling reagent, purified by
HPLC, and the side chains were deprotected (see the Support-
ing Information). GS was synthesized as a control under similar
conditions. The NMR resonances of GS and the analogue 4
were fully assigned in [D6]DMSO at 298 K. NMR observables Figure 3. A) Overlay of the 20 lowest-energy NMR structures of 4 (ATC resi-
(3J(HNHa), amide proton temperature coefficients) for GS were
dues are in blue, protons are omitted for clarity). B) Backbone of the ana-
logue 4, hydrogen bonds are represented in dashed lines and C) superimpo-
similar to those previously reported (see Table S14–S19 in the
sition of the backbone of 4 (white and blue for ATC) and the reverse-turn
Supporting Information). The entire set of ROE correlations
3b (cyan). D) Superimposition of the backbone of 4 (white and blue) and GS
was compatible with the XRD structure provided by Overhand (yellow).
Chem. Eur. J. 2014, 20, 6713 – 6720
www.chemeurj.org
6716
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
from 100:0 to 0:100 eluents A/B over 3 min, in which eluents A=
H2O/TFA 0.1% and B=CH3CN/TFA 0.1%. Detection was done at
214 and 254 nm using a photodiode array detector. Retention
times are reported as follows: tr (min). 1H and 13C NMR spectra
were recorded at room temperature in deuterated solvents. Chemi-
cal shifts (d) are given in parts per million relative to TMS or rela-
tive to the solvent (1H: d (CDCl3)=7.24 ppm; 13C: d (CDCl3)=
77.2 ppm). The following abbreviations were used to designate the
signal multiplicities: s (singlet), d (doublet), t (triplet), q (quartet), m
(multiplet), br (broad). Analytical thin-layer chromatography (TLC)
was performed using aluminium-backed silica gel plates coated
with a 0.2 mm thickness of silica gel or with aluminium oxide 60
F254, neutral. LC-MS spectra (ESI) were recorded on an HPLC using
an analytical Chromolith Speed Rod RP-C18 185 Pm column (50ꢃ
4.6 mm, 5 mm) using a flow rate of 3.0 mLminÀ1, and gradients of
100:0 to 0:100 eluents A/B over 3 min, in which eluents A=H2O/
HCOOH 0.1% and B, CH3CN/HCOOH 0.1%. High-resolution mass
spectrometric analyses were performed with a time-of-flight (TOF)
mass spectrometer fitted with an electrospray ionisation source
(ESI). All measurements were performed in the positive ion mode.
Melting points (m.p.) are uncorrected and were recorded on a capil-
lary melting point apparatus. Enantiomeric excesses were deter-
mined by Chiral HPLC analysis using a Chiracel OD-R column
(250 mm ꢃ 4.6 mm) with H2O, TFA 0.1%/ACN, TFA 0.1%. (40: 60) as
side chains from the charged face to the hydrophobic side of
the sheet thus strengthening the amphipathicity (Figure 3d).
Antimicrobial activities
GS analogue 4 was then tested for its antimicrobial activity
against a representative range of Gram-positive and Gram-neg-
ative bacteria strains. It conserved functional mimicry of the
natural product although the antibiotic activities were slightly
reduced. More importantly, the haemolytic activity of 4 was de-
creased sixfold (Table 2). Several studies support the haemo-
toxicity is highly correlated to the hydrophobicity and/or am-
Table 2. Antibacterial and haemolytic activities of 4 compared to GS.
4
GS
E. coli[a]
25
12.5
50[b]
3.125
1.56
Pseudomonas aeruginosa[a]
S. aureus[a]
>50[b]
4.68
Bacillus subtilis[a]
3.125
hemolysis [%][c]
14
80
eluent and a flow rate of 1 mLminÀ1
.
[a] Minimal inhibitory concentrations (MICs) in mgmLÀ1 were determined
after 48 h incubation. [b] 50 mm was the maximal tested concentration
compatible with the compound solubility in culture media. [c] The per-
centage of haemolytic activity of red blood human cells caused at
100 mm peptide concentration.
Synthesis of (S)-1a, (R)-1b and 2
Compounds 1a, 1b, and 2 were synthesized from Fmoc-(S)-ala-
nine, Fmoc-(R)-alanine and Fmoc-(S)-phenylalanine, according the
previously described procedure.[5]
phipathicity of the molecule.[16] As suggested by RP-HPLC anal-
yses and by the re-orientation of the benzyl side chains, the
haemotoxicity decrease could be explained in part by a higher
hydrophilicity and amphiphilicity of 4 compared to GS.
General procedure for the preparation of hybrid peptides
3a and 3b
Conclusion
Peptides 3a and 3b were synthesized by solid-phase methods
using standard Fmoc chemistry. The peptides were assembled on
ChemMatrixꢄ Rink Amide resin loaded at 0.54 mmolgÀ1. Chain
elongations were performed with Fmoc-Ala-OH, Fmoc-Ile-OH, (S)-
1a or (R)-1b, and Fmoc-Val-OH (0.3 mmol each), using 20% piperi-
dine/NMP for Fmoc deprotection, HBTU/HOBt for activation
(0.3 mmol each), N-methylmorpholine (NMM, 6 equiv) and NMP
(5 mL) as solvent. After completion of the synthesis, the linear pep-
tides were acetylated with Ac2O/DCM (1:1 vv; 3ꢃ5 min), then
cleaved from the resin with 100% CF3COOH. After removing the
solvent, the peptides were purified by preparative HPLC on a Delta
Pak, C18 column (15 mm, 40ꢃ100 mm; Solvent A: H2O/TFA vol/vol
0.1%; Solvent B: MeCN/TFA vol/vol 0.1%; flow 20 mLminÀ1; linear
gradient A/B: from 85:15 to 60:40 in 30 min). HPLC analyses were
performed on an analytical Chromolith Speed Rod RP-C18 185 Pm
column (50ꢃ4.6 mm, 5 mm) using a flow rate of 5.0 mLminÀ1, and
gradients from 100:0 to 0:100 eluents A/B over 3 min, in which elu-
ents A=H2O/TFA 0.1% and B=CH3CN/TFA 0.1%. Detections were
done at 214 and 254 nm using a photodiode array detector. the re-
tention times were determined to be 1.28 min for 3a and 1.26 min
for 3b. Compounds 3a and 3b were obtained in 35 and 32%
yields respectively. 3a: tr =1.28 min; LC-MS: (ESI+): m/z (%): 511.2
(100) [M+H]+, 533.2 (10) [M+Na]+. 3b: tr =1.26 min; LC-MS:
(ESI+): m/z (%): 511.2 (100) [M+H]+; for NMR data see Tables S1–
S4 in the Supporting Information.
We have demonstrated, using various techniques, that g-amino
acid ATC was able to induce a turn in solution when included
in a short peptide sequence. It stabilized a typical C9-turn with
a similar geometry to that adopted in the 9-helix ATC oligo-
mers. Depending on the configuration of the g-amino acid ste-
reocenter, ATCs could serve as turn mimetic. We successfully
synthesized a gramicidin S analogue exhibiting reduced hae-
motoxicity, while maintaining interesting antibacterial activi-
ties. Considering the high diversity and the unique conforma-
tional properties of these g-amino acid residues, we believe
that the development of such turn mimics and their incorpora-
tion into biologically active molecules to constraint active con-
formations is of interest.
Experimental Section
General procedures
Commercially available reagents and solvents were used without
any further purification. Reactions were monitored by HPLC using
an analytical Chromolith Speed Rod RP-C18 185 Pm column (50ꢃ
4.6 mm, 5 mm) using a flow rate of 5.0 mLminÀ1, and gradients
Chem. Eur. J. 2014, 20, 6713 – 6720
www.chemeurj.org
6717
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
and B=CH3CN/TFA 0.1%. Detections were done at 214 and
254 nm using a photodiode array detector. tr =1.78 min; LC-MS:
(ESI+): m/z (%): 571.4 (100) [M+2H]2+, 1141.7 (28) [M+H]+; for
NMR data see Tables S14–S15 in the Supporting Information.
Synthesis of gramicidin S
The backbone-cyclic peptide was assembled on 2-chlorotrityl chlo-
ride resin (Novabiochem). Fmoc-Pro-OH (0.25 mmol) was loaded
on 2-chlorotritylchloride resin (800 mg, loading 0.3 mmolgÀ1) in
the presence of N-methylmorpholine (NMM, 4 equiv) in CH2Cl2
(15 mL). The unreacted sites on the resin were capped by washing
with a mixture of CH2Cl2/MeOH/DIPEA (17:2:1) followed by MeOH.
Following removal of the Fmoc-group using 20% piperidine in N-
methyl-2-pyrrolidinone (NMP), chain elongation was performed
with Fmoc-d-Phe, Fmoc-Leu-OH, Fmoc-Orn(Boc)-OH, Fmoc-Val-OH,
Fmoc-Pro-OH, Fmoc-d-Phe, Fmoc-Leu-OH, Fmoc-Orn(Boc)-OH,
Fmoc-Val-OH (1 mmol each, 4 equiv), using 20% piperidine/NMP
for Fmoc deprotection, HBTU (4 equiv) for activation, DIPEA as
base and NMP as solvent. After completion of the synthesis, the
linear peptide was cleaved from the resin with CF3COOH/water
(1:99 v/v). After removing the solvent, the peptide was cyclized
overnight using HBTU (4 equiv) for activation, N-methylmorpholine
(11 equiv) as base in DMF (30 mL). The solvent was removed under
reduce pressure then the protected peptide was purified by prepa-
rative HPLC on a Delta Pak, C18 column (15 mm, 40ꢃ100 mm; Sol-
vent A: H2O/TFA vol/vol 0.1%; Solvent B: MeCN/TFA vol/vol 0.1%;
flow 20 mLminÀ1; linear gradient A/B: from 50:50 to 10:90 in
30 min) and lyophilized. Boc removal was done using TFA/Water
(9:1 v/v). Gramicidin S was recovered in 12% yield (39 mg). HPLC
analysis was performed on an analytical chromolith speed rod RP-
C18 185 Pm column (50ꢃ4.6 mm, 5 mm) using a flow rate of
5.0 mLminÀ1, and gradient from 100:0 to 0:100 eluents A/B over
3 min, in which eluents A=H2O/TFA 0.1% and B=CH3CN/TFA
0.1%. Detections were done at 214 and 254 nm using a photodiode
array detector. tr =2.14 min; LC-MS: (ESI+): m/z (%): 571.4 (100)
[M+2H]2+), 1141.8 (35) [M+H]+; for NMR data see Tables S14 and
S15 in the Supporting Information.
NMR experiments
The NMR samples contained 3a or 3b (5 mm) dissolved in
[D6]DMSO and in [D3]MeOH and gramicidine S or 4 (20 mm) dis-
solved in [D6]DMSO. All spectra were recorded on a Bruker Avance
III 600 equipped with a 5 mm quadruple-resonance (1H, 13C, 15N,
31P). Homonuclear 2D spectra DQF-COSY, TOCSY, and ROESY were
typically recorded in the phase-sensitive mode using the States-
TPPI method as data matrices of 256–512 real (t1)ꢃ2048 (t2) com-
plex data points; 8–64 scans per t1 increment with 1.5 s recovery
delay and spectral width of 6009 Hz in both dimensions were
used. The mixing times were 80 ms for TOCSY and 300 ms spinlock
for ROESY experiments. In addition, 2D heteronuclear spectra 15N-
1H, 13C-1H HSQC, and 13C-1H HMBC were acquired to fully assign
the oligomers (8–128 scans, 256 real (t1)ꢃ2048 (t2) complex data
points). Spectra were processed and visualized with Topspin 3.0
(Bruker Biospin) on a Linux Station. The matrices were zero-filled to
1024 (t1)ꢃ2048 (t2) points after apodization by shifted sine-square
multiplication and linear prediction in the F1 domain. Chemical
shifts were referenced to TMS.
Structure calculations
Parameter files for the ornithine residue were taken from the
Amber parameter database provided by the Bryce group at the
University of Manchester.[17] The ATC amino acid was built using
Sirius,[18] parameter files were then generated using Antechamber
and the R.E.D. server.[19] Leap was then used to create the topology
and coordinate files for 3a, 3b and 4.
Synthesis of gramicidine S analogue 4
Compound 4 was synthesized according to the general procedure
from Boc-b3-hAla-OH (1.0 g, 4.99). The backbone-cyclic peptide was
assembled on 2-chloro-trityl chloride resin (Novabiochem). Com-
pound 2 (0.25 mmol) was loaded on 2-chloro-tritylchloride resin
(800 mg, loading 0.3 mmolgÀ1) in the presence of N-methylmor-
pholine (NMM, 4 equiv) in CH2Cl2 (15 mL). The unreacted sites on
the resin were capped by washing with a mixture of CH2Cl2/MeOH/
DIPEA (17:2:1) followed by MeOH. Following removal of the Fmoc-
group using 20% piperidine in N-methyl-2-pyrrolidinone (NMP),
chain elongation was performed with Fmoc-Leu-OH, Fmoc-Orn-
(Boc)-OH, Fmoc-Val-OH, Fmoc-(R)-ATC 2, Fmoc-Leu-OH, Fmoc-Orn-
(Boc)-OH, Fmoc-Val-OH (1 mmol each, 4 equiv), using 20% piperi-
dine/NMP for Fmoc deprotection, HBTU (4 equiv) for activation,
DIPEA as base and NMP as solvent. After completion of the synthe-
sis, the linear peptide was cleaved from the resin with CF3COOH/
water (1:99 vv). After removing the solvent, the peptide was cy-
clized overnight using HBTU (4 equiv) for activation, N-methylmor-
pholine (11 equiv) as base in DMF (30 mL). The solvent was re-
moved under reduced pressure then the protected peptide was
purified by preparative HPLC on a Delta Pak, C18 column (15 mm,
40ꢃ100 mm; Solvent A: H2O/TFA vol/vol 0.1%; Solvent B: MeCN/
TFA vol/vol 0.1%; flow 20 mLminÀ1; linear gradient A/B: from 50:50
to 10:90 in 30 min) and lyophilized. Boc removal was done using
TFA/Water (9:1 v/v). The cyclopeptide 4 was recovered in 16%
yield (45 mg). HPLC analysis was performed on an analytical chro-
molith speed rod RP-C18 185 Pm column (50ꢃ4.6 mm, 5 mm)
using a flow rate of 5.0 mLminÀ1, and gradients from 100:0 to
0:100 eluents A/B over 3 min, in which eluents A=H2O/TFA 0.1%
1H chemical shifts were assigned according to classical procedures.
NOE cross-peaks were integrated and assigned within the
NMRView software.[20] The volume of a ROE between methylene
pair protons was used as a reference of 1.8 ꢂ. The lower bound for
all restraints was fixed at 1.8 ꢂ and upper bounds at 2.7, 3.3, and
5.0 ꢂ, for strong, medium, and weak correlations, respectively.
Pseudo-atom corrections of the upper bounds were applied for un-
resolved aromatic, methylene, and methyl protons signals as de-
scribed previously.[21] Structure calculations were performed with
AMBER 10[22] in two stages: cooking and simulated annealing (SA)
in vacuum. The cooking stage was performed at 500 K to generate
100 initial random structures. SA calculations were carried during
20 ps (20000 steps, 1 fs long) as described elsewhere. First, the
temperature was risen quickly and was maintained at 1000 K for
the first 5000 steps, then the system was cooled gradually from
1000 K to 100 K from step 5001 to 18000 and finally the tempera-
ture was brought to 0 K during the 2000 remaining steps. For the
3000 first steps, the force constant of the distance restraints was
increased gradually from 2.0 kcalmolÀ1 ꢂ to 20 kcalmolÀ1 ꢂ. For the
rest of the simulation (step 3001 to 20000), the force constant was
kept at 20 kcalmolÀ1 ꢂ. The calculations were launched in vacuum
for 3a, 3b, and 4. The 20 lowest energy structures with no viola-
tions >0.3 ꢂ were considered as representative of the peptide
structure. The representation and quantitative analysis were carried
out using Ptraj,[23] MOLMOL,[24] and PyMOL.[25] The 20 lowest energy
structures of 3a and 3b peptidomimetics were used for a molecu-
lar dynamic simulation during 50 ns in a methanol box (see the
MD section).
Chem. Eur. J. 2014, 20, 6713 – 6720
www.chemeurj.org
6718
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Molecular dynamic simulations in an explicit solvent box
Antimicrobial activities
The lowest energy structures of 3a and 3b were immersed in
a methanol box with a layer of 10 ꢂ for a molecular dynamic simu-
lation stage of 50 ns using the following protocol. First, a minimiza-
tion procedure consisting in two stages: 1) The solute was kept
fixed while the solvent methanol positions are optimized (10000
steps) and then 2) the entire system was minimized (2500 steps).
3) The system was then allowed to heat up from 0 to 300 K at con-
stant volume keeping the solute fixed (50000 steps, that is, 50 ps).
4) The whole system was equilibrated at a constant pressure of
1 atm and at 300 K over 50000 steps (50 ps). 5) Molecular dynamic
simulations were finally performed at 300 K and 1 atm during
50 ns (time steps of 0.1 fs). NMR inter-residues restraints were ap-
plied during the entire runs (Tables S11 and S12 in the Supporting
Information).
GS and 4 were dissolved at 1 mgmLÀ1 in DMSO then successively
diluted in the same solvent. DMSO was used as negative control.
The antibacterial activities of GS and 4 were monitored by liquid
growth inhibition assay performed in microtiter plates. Briefly,
a DMSO solution (10 mL) of GS or 4 at different concentrations
(ranging from 1 mgmLÀ1 to 1 mgmLÀ1; 100 to 0.1 mgmLÀ1 final
concentrations, respectively) was added to a suspension of a mid-
logarithmic phase culture of bacteria (90 mL; OD620nm =0.02) in
poor-broth nutrient medium. Cultures were carried out in triplicate
and microbial growths were assessed by measuring turbidity at
620 nm after incubation (48 h, 378C). The MIC was defined as the
lowest concentration of antibiotic where bacterial growth was not
detected. The MICs were determined from independent triplicate
assays and were based on a serial twofold plus or minus system.
To be considered valid, MIC determinations for each of the 3 repli-
cates had to be within plus or minus 1 dilution of each other.
DFT calculations
From MD simulations in methanol, compound 3a underwent
a transition after 1.5 ns from a twist-turn structure to a reverse-turn
structure stabilized by a bifurcated C9/12 pseudocycle. However, the
reverse-turn structure did not provide any characteristic NOE corre-
lation probed by NMR experiments. So, the twist-turn and reverse-
turn structures were reoptimized at the B3LYP/6–31G(d) level of
theory. Then, the single-point energies of this reverse-turn struc-
ture and the twist-turn structure obtained from NMR experiments
were calculated at the same level of theory with the conductor-like
polarizable continuum model (CPCM)[8] in methanol. The twist-turn
structure was found to be more stable by 3.40 kcalmolÀ1 than the
reverse-turn structure in methanol, which indicates the relative
populations of 99.7 and 0.3% for twist- and reverse-turn structures
at 298 K, respectively.
Haemolytic activities
GS and 4 (2 mm) were dissolved in DMSO then successively diluted
in the same solvent. The haemolytic activities of GS and 4 were
monitored in microtiter plates. PBS (45 mL) and sample (5 mL) were
added to a freshly prepared blood solution (50 mL; 5ꢃ105 erythro-
cytes per mL). The mixtures were incubated to the 378C during
30 min. After the incubation period, erythrocytes were separated
by centrifugation (300 g, 108C, 10 min). A 50 mL portion of each su-
pernatant was further transferred to a microwell plate to measure
the absorbance at 490 nm into a microplate reader. DMSO was
used as negative control while SDS (0.1% final concentration) was
used for 100% haemolysis.
The two NMR structures 3a and 3b were optimized at the B3LYP/
6–31G(d) level of theory. To figure out the relative strength of the
C9 hydrogen bond between the amide hydrogen of the third Ile
residue and the carbonyl oxygen of the first Val residue, a natural
bond orbital (NBO) analysis was performed for the optimized struc-
tures at the B3LYP/6–31+G(d,p) level of theory. The three kinds of
properties were compared to understand the relative strength of
the hydrogen bond and the calculated results are shown in
Table 1. The first values are the natural atomic charges of oxygen
(the carbonyl oxygen of the first Val residue) and hydrogen (the
amide hydrogen of the third Ile residue). The second value is the
Wiberg bond index of the NÀH bond of the third Ile residue, which
shows the strength of a bond. The third value is the second pertur-
bation energy (DE2) of the lone pair orbitals of the oxygen with
the corresponding NÀH antibonding orbital, which is called the hy-
perconjugation due to the charge transfer. We can say that the hy-
drogen bond is strong when the magnitudes of atomic charges
become larger, the Wiberg index becomes smaller, and DE2 be-
comes larger.
Acknowledgements
We thank the IBMM, the Laboratoire de Mesure Physique of
Montpellier, the CNRS and the University of Montpellier 1 for
financial support.
Keywords: amino acids · gramicidin S · hybrid peptide ·
thiazoles
[1] a) J. S. Richardson, Adv. Protein Chem. 1981, 34, 167–339; b) C. M.
Wilmot, J. M. Thornton, J. Mol. Biol. 1988, 203, 221–232; c) G. D. Rose,
L. M. Gierasch, J. A. Smith, Adv. Protein Chem. 1985, 37, 1–109; d) E.
Vass, M. Hollosi, F. Besson, R. Buchet, Chem. Rev. 2003, 103, 1917–1954.
[2] a) M. D. Seo, H. S. Won, J. H. Kim, T. Mishig-Ochir, B. J. Lee, Molecules
2012, 17, 12276–12286; b) R. M. Epand, H. J. Vogel, Biochim. Biophys.
Acta Biomembr. 1999, 1462, 11–28; c) L. T. Nguyen, E. F. Haney, H. J.
Vogel, Trends Biotechnol. 2011, 29, 464–472.
[3] a) Y. J. Chung, L. A. Christianson, H. E. Stanger, D. R. Powell, S. H. Gell-
man, J. Am. Chem. Soc. 1998, 120, 10555–10556; b) D. Seebach, A. K.
Beck, D. J. Bierbaum, Chem. Biodiversity 2004, 1, 1111–1239; c) W. S.
Horne, S. H. Gellman, Acc. Chem. Res. 2008, 41, 1399–1408; d) B. R.
Huck, J. D. Fisk, S. H. Gellman, Org. Lett. 2000, 2, 2607–2610; e) R. S.
Roy, P. Balaram, J. Pept. Res. 2004, 63, 279–289; f) S. Chatterjee, R. S.
Roy, P. Balaram, J. R. Soc. Interface 2007, 4, 587–606; g) K. Basuroy, A.
Rajagopal, S. Raghothama, N. Shamala, P. Balaram, Chem. Asian J. 2012,
7, 1671–1678; h) C. Andrꢀ, B. Legrand, C. Deng, C. Didierjean, G. Pick-
aert, J. Martinez, M. C. Averlant-Petit, M. Amblard, M. Calmes, Org. Lett.
2012, 14, 960–963.
Circular dichroism
Circular dichroism (CD) experiments were carried out using a Jasco
J815 spectropolarimeter. The spectra were recorded with of com-
pounds 3a (0.2 mm), 3b (0.2 mm) and 4 (0.1 mm) dissolved in
MeOH using a 1 mm path length CD cuvette at 208C, over a wave-
length range of 190–300 nm. Continuous scanning mode was
used, with a response of 1.0 s with 0.2 nm steps and a bandwidth
of 2 nm. The signal-to-noise ratio was improved by acquiring each
spectrum over an average of two scans. The baseline was correct-
ed by subtracting the background from the sample spectrum.
[4] a) F. Bouillꢅre, S. Thꢀtiot-Laurent, C. Kouklovsky, V. Alezra, Amino Acids
2011, 41, 687–707; b) A. Trabocchi, E. G. Occhiato, D. Potenza, A.
Guarna, J. Org. Chem. 2002, 67, 7483–7492; c) S. Aravinda, K. Ananda,
Chem. Eur. J. 2014, 20, 6713 – 6720
www.chemeurj.org
6719
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
N. Shamala, P. Balaram, Chem. Eur. J. 2003, 9, 4789–4795; d) E. Mann, H.
Kessler, Org. Lett. 2003, 5, 4567–4570; e) C. Grison, P. Coutrot, S.
Geneve, C. Didierjean, M. Marraud, J. Org. Chem. 2005, 70, 10753–
10764; f) A. Sengupta, S. Aravinda, N. Shamala, K. M. Raja, P. Balaram,
Org. Biomol. Chem. 2006, 4, 4214–4222; g) P. G. Vasudev, K. Ananda, S.
Chatterjee, S. Aravinda, N. Shamala, P. Balaram, J. Am. Chem. Soc. 2007,
129, 4039–4048; h) S. Chatterjee, P. G. Vasudev, K. Ananda, S. Raghotha-
ma, N. Shamala, P. Balaram, J. Org. Chem. 2008, 73, 6595–6606; i) S.
Chatterjee, P. G. Vasudev, S. Raghothama, C. Ramakrishnan, N. Shamala,
P. Balaram, J. Am. Chem. Soc. 2009, 131, 5956–5965; j) S. Thetiot-Lau-
rent, F. Bouillere, J. P. Baltaze, F. Brisset, D. Feytens, C. Kouklovsky, E.
Miclet, V. Alezra, Org. Biomol. Chem. 2012, 10, 9660–9663; k) S. K. Maji,
R. Banerjee, D. Velmurugan, A. Razak, H. K. Fun, A. Banerjee, J. Org.
Chem. 2002, 67, 633–639; l) Y. K. Kang, B. J. Byun, Biopolymers 2012, 97,
1018–1025.
[14] a) D. C. Hodgkin, B. M. Oughton, Biochem. J. 1957, 65, 752–756; b) M.
Doi, S. Fujita, Y. Katsuya, M. Sasaki, T. Taniguchi, H. Hasegawa, Arch. Bio-
chem. Biophys. 2001, 395, 85–93; c) K. Yamada, M. Unno, K. Kobayashi,
H. Oku, H. Yamamura, S. Araki, H. Matsumoto, R. Katakai, M. Kawai, J.
Am. Chem. Soc. 2002, 124, 12684–12688; d) M. Kawai, H. Yamamura, R.
Tanaka, H. Umemoto, C. Ohmizo, S. Higuchi, T. Katsu, J. Peptide Res.
2005, 65, 98–104; e) C. Solanas, B. G. de La Torre, M. Fernandez-Reyes,
C. M. Santiveri, M. A. Jimenez, L. Rivas, A. I. Jimenez, D. Andreu, C. Cati-
viela, J. Med. Chem. 2010, 53, 4119–4129; f) M. Tamaki, K. Takanashi, T.
Harada, K. Fujinuma, M. Shindo, M. Kimura, Y. Uchida, Chem. Pharm.
Bull. 2011, 59, 1481–1484; g) M. van der Knaap, L. T. Lageveen, H. J.
Busscher, R. Mars-Groenendijk, D. Noort, J. M. Otero, A. L. Llamas-Saiz,
M. J. van Raaij, G. A. van der Marel, H. S. Overkleeft, M. Overhand, Chem-
MedChem 2011, 6, 840–847; h) J. Xiao, B. Weisblum, P. Wipf, Org. Lett.
2006, 8, 4731–4734; i) E. Kamysz, B. Mickiewicz, W. Kamysz, S. Bielinska,
S. Rodziewicz-Motowidlo, J. Ciarkowski, J. Pept. Sci. 2011, 17, 211–217;
j) V. V. Kapoerchan, E. Spalburg, A. J. de Neeling, R. H. Mars-Groenendijk,
D. Noort, J. M. Otero, P. Ferraces-Casais, A. L. Llamas-Saiz, M. J. van Raaij,
J. van Doorn, G. A. van der Marel, H. S. Overkleeft, M. Overhand, Chem.
Eur. J. 2010, 16, 4259–4265; k) G. M. Grotenbreg, A. E. Buizert, A. L.
Llamas-Saiz, E. Spalburg, P. A. van Hooft, A. J. de Neeling, D. Noort, M. J.
van Raaij, G. A. van der Marel, H. S. Overkleeft, M. Overhand, J. Am.
Chem. Soc. 2006, 128, 7559–7565.
[5] L. Mathieu, B. Legrand, C. Deng, L. Vezenkov, E. Wenger, C. Didierjean,
M. Amblard, M. C. Averlant-Petit, N. Masurier, V. Lisowski, J. Martinez,
L. T. Maillard, Angew. Chem. 2013, 125, 6122–6126; Angew. Chem. Int.
Ed. 2013, 52, 6006–6010.
[6] N. H. Andersen, J. W. Neidigh, S. M. Harris, G. M. Lee, Z. Liu, H. Tong, J.
Am. Chem. Soc. 1997, 119, 8547–8561.
[7] a) X. Daura, K. Gademann, H. Schafer, B. Jaun, D. Seebach, W. F. van Gun-
steren, J. Am. Chem. Soc. 2001, 123, 2393–2404; b) J. Dolenc, J. H. Mis-
simer, M. O. Steinmetz, W. F. van Gunsteren, J. Biomol. NMR 2010, 47,
221–235; c) J. R. Allison, M. Muller, W. F. van Gunsteren, Protein Sci.
2010, 19, 2186–2195; d) Z. Gattin, A. Glattli, B. Jaun, W. F. van Gunste-
ren, Biopolymers 2007, 85, 318–332.
[8] V. Barone, M. Cossi, J. Phys. Chem. A 1998, 102, 1995–2001.
[9] Gaussian 03, Revision D.01, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.
Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomery Jr., T. Vreven,
K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone,
B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsu-
ji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Na-
kajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P.
Hratchian, J. B. Cross, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Strat-
mann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, P. Y.
Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakr-
zewski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick,
A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G.
Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P.
Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y.
Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc., Wallingford
CT, 2004.
[15] A. L. Llamas-Saiz, G. M. Grotenbreg, M. Overhand, M. J. van Raaij, Acta
Crystallogr. Sect. D 2007, 63, 401–407.
[16] a) L. H. Kondejewski, M. Jelokhani-Niaraki, S. W. Farmer, B. Lix, C. M. Kay,
B. D. Sykes, R. E. Hancock, R. S. Hodges, J. Biol. Chem. 1999, 274, 13181–
13192; b) C. McInnes, L. H. Kondejewski, R. S. Hodges, B. D. Sykes, J. Biol.
Chem. 2000, 275, 14287–14294.
[17] http://www.pharmacy.manchester.ac.uk/bryce/amber/.
[18] http://www.ngbw.org/sirius/.
[19] E. Vanquelef, S. Simon, G. Marquant, E. Garcia, G. Klimerak, J. C. Dele-
pine, P. Cieplak, F.-Y. Dupradeau, Nucl. Acids Res. 2011, 39, W511–W517.
[20] B. A. Johnson, R. A. Blevins, J. Biomol. NMR 1994, 4, 603–614.
[21] K. Wꢆthrich, NMR of Proteins and Nucleic Acids, Wiley-Interscience, New
York, 1986.
[22] AMBER 10, D. A. Case, T. A. Darden, T. E. Cheatham, III, C L. Simmerling,
J. Wang, R. E. Duke, R. Luo, M. Crowley, R. C. Walker, W. Zhang, K. M.
Merz, B. Wang, S. Hayik, A. Roitberg, G. Seabra, I. Kolossvary, K. F. Wong,
F. Paesani, J. Vanicek, X. Wu, S. R. Brozell, T. Steinbrecher, H. Gohlke, L.
Yang, C. Tan, J. Mongan, V. Hornak, G. Cui, D. H. Mathews, M. G. Seetin,
C. Sagui, V. Babin, P. A. Kollman, University of California, San Francisco,
2008.
[23] D. R. Roe and T.E. Cheatham, J. Chem. Theory Comput., 2013, 9, 3084–
3095.
[24] R. Koradi, M. Billeter, K. Wuthrich, J. Mol. Graph. 1996, 14, 51–55.
[25] Delano Scientific, http://www.pymol.org/
[10] F. Weinhold in Encyclopedia of Computational Chemistry, Vol. 3 (Eds.:
P. v. R. Schleyer, N. L. Allinger, T. Clark, J. Gasteiger, P. A. Kollman, H. F.
Schaefer, III, P. R. Schreiner), John Wiley
pp. 1792–1811.
& Sons, Chichester 1998,
[11] K. B. Wiberg, Tetrahedron 1968, 24, 1083.
[12] L. H. Kondejewski, S. W. Farmer, D. S. Wishart, R. E. Hancock, R. S.
Hodges, Int. J. Pept. Prot. Res. 1996, 47, 460–466.
Received: February 14, 2014
[13] E. J. Prenner, R. N. Lewis, R. N. McElhaney, Biochim. Biophys. Acta Bio-
membr. 1999, 1462, 201–221.
Published online on March 25, 2014
Chem. Eur. J. 2014, 20, 6713 – 6720
www.chemeurj.org
6720
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim