Braces for the peptide backbone: Insights into structure-activity relationships of protease inhibitor mimics with locked amide conformations

Article abstract of DOI:10.1002/anie.201108983

Flower power: Potent protease inhibitors containing triazolyl mimics of cis and trans backbone amides were engineered based on the structure of the sunflower trypsin inhibitor 1. The biologically relevant cis-Pro motif was successfully replaced with a non

Full text of DOI:10.1002/anie.201108983

.
Angewandte
Communications
DOI: 10.1002/anie.201108983
Amide Mimics
Braces for the Peptide Backbone: Insights into Structure–Activity
Relationships of Protease Inhibitor Mimics with Locked Amide
Conformations**
Marco Tischler, Daichi Nasu, Martin Empting, Stefan Schmelz, Dirk W. Heinz,
Philipp Rottmann, Harald Kolmar, Gerd Buntkowsky,* Daniel Tietze,* and Olga Avrutina*
The architecture of protein macromolecules fundamentally
depends on the sequential arrangement of peptide backbone
bonds in defined conformations. Among the three torsion
angles (f, y, and w) present at each amino acid, it is the amide
bond (w) which is intrinsically hindered as a result of its
partial double-bond character and it is thus more or less
restricted to either a trans or a cis conformation (Figure 1).[1]
Generally, amide bonds occur predominantly in the trans
conformation as it minimizes unfavorable contacts between
adjacent amino acids.[1d] Nevertheless, cis amides, if present,
are usually an imperative for bioactivity.[2]
The distribution of cis/trans conformations drastically
changes from about 0.03% for non-prolyl peptide bonds to
about 6% for Xaa-Pro motifs.[1a–c,e] This is directly linked to
the propensity of proline with its cyclic side chain to form
a sterically hindered backbone motif which strongly affects
the preceding peptide bond.[3]
Beyond the structural features that cis amide bonds
Figure 1. Representation of triazolyl amide mimics incorporated into
induce in proteins, cis–trans isomerization (CTI) adds a kinetic
a generic peptide backbone. a) trans amide mimic based on a 1,4-
dimension to biomolecular systems. In fact, it is not only one
disubstituted 1,2,3-triazole, b) cis amide mimic based on a 1,5-disub-
stituted 1,2,3-triazole. Backbone atoms are depicted as sticks and side-
of the rate-determining steps in protein folding,[4] but CTI
also provides the possibility for a time-dependent conforma-
tional switch, thus allowing for a dynamic modulation of
structure and activity.[3]
During the last decade, a few structural mimics of cis
amides in peptides and proteins have been reported, among
chain moieties as balls. An overlay of the Lewis structure with stereo
information is given. N blue, C light gray, O red, side chains yellow;
hydrogen atoms omitted for clarity. The dihedral angles defining the
trans and cis amide bonds (w torsion angles) and the respective
triazole-based mimics are highlighted by black outlines.
[*] M. Tischler,[+] P. Rottmann, Prof. Dr. G. Buntkowsky, Dr. D. Tietze
Eduard-Zintl-Institut fꢀr Anorganische und Physikalische Chemie
Technische Universitꢁt Darmstadt
them pseudo-prolines,[2f,5] disubstituted tetrazoles,[6] and tri-
azoles.[2a,e,7] Thus, an Asn-cis-Pro bond in bovine pancreatic
ribonuclease was successfully replaced by a triazolyl unit
through expressed protein ligation[8] without loss of catalytic
activity.[2e] In fact, disubstituted 1,2,3-triazoles have been
shown to be viable surrogates for cis and trans- amide bonds,
depending on their substitution pattern (Figure 1).[2e,9] To our
knowledge, no detailed insights into the molecular architec-
ture and structural requirements of a bioactive cis amide
mimic containing a 1,5-disubstituted 1,2,3-triazole upon bind-
ing to its target have been reported.
Petersenstrasse 22, 64287 Darmstadt (Germany)
E-mail: Gerd.Buntkowsky@chemie.tu-darmstadt.de
tietze@chemie.tu-darmstadt.de
Homepage: http://www.tu-darmstadt.de/fb/ch/akbuntkowsky
D. Nasu,[+] M. Empting,[+] Prof. Dr. H. Kolmar, Dr. O. Avrutina
Clemens-Schçpf-Institut fꢀr Organische Chemie und Biochemie
Technische Universitꢁt Darmstadt
Petersenstrasse 22, 64287 Darmstadt (Germany)
E-mail: Avrutina@Biochemie-TUD.de
Dr. S. Schmelz, Prof. Dr. D. W. Heinz
Herein, we demonstrate the applicability of modular
triazole-based backbone elements for locking cis and trans
amides within the functional loop of a Bowman–Birk
protease inhibitor (BBI).[10] High-resolution crystal structures
revealed the detailed structural features of both cis and trans
triazolyl amide surrogates in highly potent peptidomimetics
bound to bovine trypsin. Our results provide new information
on the behavior of triazolyl units within active biomolecules,
Abteilung Molekulare Strukturbiologie
Helmholtz-Zentrum fꢀr Infektionsforschung (HZI)
Inhoffenstrasse 7, 38124 Braunschweig (Germany)
[+] These authors contributed equally to this work.
[**] This research was supported by the Deutsche Forschungsgemein-
schaft (grant Ko 1390/9-1) and LOEWE—Soft Control.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201108983.
3708
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 3708 –3712

Angewandte
Chemie
and lead to a re-evaluation of their steric and electronic
homology to the native peptide bond.
validate our results and to assure consistency within the test
system used, we included the corresponding Pro!Ala
mutants 1 and 2 in this work. Efforts were made to exchange
the subsequent trans amide bond with a corresponding 1,4-
disubstituted counterpart 6 and to synthesize SFTI-1[1,14]
derivatives bearing the respective mismatched substitution
patterns of the triazolyl moiety at each position (4 and 5). To
precisely refer to each peptidomimetic compound, we
assigned a three-letter code for the introduced dipeptide
surrogates. Thus, the one-letter code was used to reflect the
mimicked amino acid sequence. A “c” or “t” indicates the
conformation of the amide bond locked by the triazole
counterfeit.
Based on 1,5-disubstituted 1,2,3-triazoles, conformation-
ally defined non-prolyl backbone motifs are easily accessible
by solid-phase peptide synthesis. All peptides were assembled
on a solid support either using commercially available
building blocks or synthetic precursors (Scheme 1 and the
Supporting Information). (S)-2-Azidopropanoic acid (7) and
the Fmoc-protected alkyne components 8 and 9 were
synthesized according to previously reported procedures.[15]
Azide 7 was introduced at position 8 or 9 in the growing
peptide chain using in situ activation under gentle microwave
irradiation. The subsequent generation of 1,4- or 1,5-disub-
stituted 1,2,3-triazoles 14–17 was achieved by CuI- or RuII-
catalyzed azide–alkyne cycloaddition (CuAAC or RuAAC,
respectively) on the solid support.[2a,7b,13,16] Microwave-
assisted Fmoc-SPPS was continued until the peptide chain
was assembled. After acidolytic cleavage, precipitation, and
DMSO- or air-mediated oxidation of crude products, chro-
matographic purification yielded the cyclic target compounds
1–6 on a multi-milligram scale.
As a model peptide the highly potent sunflower trypsin
inhibitor 1 (SFTI-1) from Helianthus annuus was chosen.[11] In
the native form, its 14 amino acid backbone GRCTKSIP-
PICFPD is cystine bridged and head-to-tail cyclized, and
contains an Ile-cis-Pro amide bond as an indispensible
prerequisite for bioactivity.[12] Recently we demonstrated
the utility of its open-chain variant (SFTI-1[1,14]) to study the
influence of subtle changes within shape-defining regions, like
the disulfide bridge, on structure–activity relationships.[13]
Because of to the highly conserved canonical conformation
of the functional loop of protease inhibitors, even minor steric
alterations on the sub-ꢀngstrom range have significantly
affected the bioactivity of the investigated peptidomimet-
ics.[13]
Besides the cystine motif, the cis amide bond between Ile7
and Pro8 is a prominent structural element within the
inhibitor loop of SFTI-1[1,14] crucial for its bioactivity
(Figure 2).[12] Consequently, the replacement of this intriguing
conformational archetype by a 1,5-disubstituted 1,2,3-triazole
mimic appeared to be a promising concept, although previous
attempts to mimic this particular motif with non-natural
surrogates were not successful.[14] Thus, a cis amide mimic
must not only match the steric requirements of the parent
perfectly, but also provide a similar chemical environment.
It has been shown that replacement of Pro8 by an alanine
residue (Figure 2, compound 1) led to a drastic decrease of
inhibitory activity against trypsin.[12] This was apparently
caused by an induced conformational heterogeneity resulting
from the loss of the cis amide stabilizing effect of proline. To
In order to gain insights into the spatial aspects of the
active loop conformation upon binding to a serine protease, in
silico calculations (Supporting Information) and crystallo-
graphic analysis were conducted using a modified proce-
dure,[17] which resulted in high-resolution crystal structures
(1.45–1.55 ꢀ) for the complex of bovine trypsin with
[IcA7,8]SFTI-1[1,14] 4 and [PtA8,9]SFTI-1[1,14] 6 (Figure 3).
The inhibitory potency of all synthesized SFTI derivatives was
evaluated in enzyme kinetic studies with active-site-titrated
trypsin using the chromogenic substrate Boc-QAR-pNA, and
the apparent and substrate-independent inhibitory constants
(Kiapp and Ki) were determined as previously reported.[10b,13]
As seen from Table 1, the presence of the undesirable
conformation at each of the two examined positions led to
a drastic decrease of inhibitory activity. Thus, the alanine
exchange at position 8 (SFTI derivative 1) resulted in the loss
of the cis-stabilizing effect of proline, hence, dynamic disorder
of the inhibitory loop by CTI. Locking the trans conformation
at this position by a 1,4-disubstituted 1,2,3-triazole had an
even more dramatic effect: peptidomimetic 4 possessed the
lowest activity of all studied compounds. Vice versa, the
variant with a locked cis conformation through the use of
a 1,5-disubstituted triazolyl building block between Ile7 and
Ala8 (compound 3) showed a significant improvement in
bioactivity over compound 1, although not in the range of the
parent peptide SFTI-1[1,14]. This can be explained consider-
ing the crystal structures of the wild-type inhibitor and 3 in
Figure 2. Overview of synthesized compounds: [Ala8]SFTI-1[1,14] (1),
[Ala9]SFTI-1[1,14] (2), [IcA7,8]SFTI-1[1,14] (3), [ItA7,8]SFTI-1[1,14] (4),
[PcA8,9]SFTI-1[1,14] (5), [PtA8,9]SFTI-1[1,14] (6). The investigated cis
(between Ile7 and Pro8) and trans amides (between Pro8 and Pro9)
are highlighted in bold.
Angew. Chem. Int. Ed. 2012, 51, 3708 –3712
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 3709

.
Angewandte
Communications
Scheme 1. a) Azide- and alkyne-bearing building blocks (2S)-2-azido-
propanoic acid (7), (2S,3S)-N-(9-fluorenylmethyloxycarbonyl)-1-ethynyl-
2-methylbutylamine (8), and (2S)-N-(9-fluorenylmethyloxycarbonyl)-2-
ethynylpyrrolidine (9). b) Synthesis of azide-bearing peptide resins 12
and 13. c) Synthesis of oxidized peptidomimetics 3–6. Fmoc-SPPS:
Fmoc-assisted solid-phase peptide synthesis; HATU: 2-(1H-7-azaben-
zotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate; DIEA:
N,N-diisopropylethylamine; MW: microwave; RuAAC: ruthenium(II)-
catalyzed azide–alkyne cycloaddition using [Cp*Ru(cod)Cl]
(Cp*=C5Me5, cod=cyclooctadiene) as catalyst; CuAAC: copper(I)-
catalyzed azide–alkyne cycloaddition using CuSO4, sodium ascorbate,
and DIEA for catalysis; multiple reaction arrows at the end of the
synthetic sequence: Fmoc-SPPS, acidolytic cleavage, precipitation,
DMSO- or air-mediated oxidation, and chromatographic isolation.
Figure 3. Comparison of the inhibitor loops of SFTI-1 and peptido-
mimetics 3 and 6 in the corresponing protease/inhibitor complexes.
Structures aligned at Lys5 of inhibitors. The orientation of the side
chain of Tyr39 from trypsin (surface) before (dashed outline) and after
energy minimization (solid outline) and the formed hydrogen bonds
are shown (Software: YASARA structure with AMBER03 force field,
POVRay). Inhibitors shown as stick representations; N blue, O red;
hydrogen atoms omitted for clarity. a) Native bicyclic SFTI-1 (C white,
PDB ID code: 1SFI). b) [IcA7,8]SFTI-1[1,14] 3 (solid outlines, C lemon
yellow, PDB ID code: 4ABJ) in an overlay with 1SFI (thin white).
c) [PtA8,9]SFTI-1[1,14] 6 (solid outlines, C orange, PDB ID code: 4ABI)
in an overlay with 1SFI (thin white).
complex with trypsin (Figure 3). It is evident that SFTI-1 is
able to form a hydrogen bond with the phenolic hydroxy
group of Tyr39 near the binding pocket of the protease
through the carbonyl oxygen of Ile7. This finding is supported
by in silico experiments, as an energy-minimization procedure
(AMBER force field) applied to the crystal structure (1SFI)
resulted in formation of the aforementioned hydrogen bond,
which may contribute to the binding enthalpy of SFTI-1[1,14].
Indeed, the N2 and N3 atoms of 1,2,3-triazoles have been
reported to possess hydrogen-bond-accepting abilities[18] and,
thus, may partly compensate for the missing interactions.
In full agreement with our expectations, the alanine
exchange at residue 9 (2) locked the trans conformation with
a 1,4-disubstituted 1,2,3-triazole between Pro8 and Ala9 (6)
and also had no pronounced effect on bioactivity. This finding
can easily be explained taking into consideration that no
alterations have been made relative to the native conforma-
3710
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 3708 –3712

Angewandte
Chemie
Table 1: Inhibitory activity of compounds 1–6 and monocyclic
SFTI-1[1,14] (wild-type, wt).[13]
To conclude, our facile preparation of peptidomimetics
bearing triazolyl backbone units may find utility in various
biochemical applications, as, in contrast to the native amide
bond, a triazole does not undergo proteolytic hydrolysis.
Therefore, enhanced metabolic stability can be expected.[21]
Gathered crystallographic data provide fundamental infor-
mation important for ongoing research aimed at the improve-
ment of peptidomimetic protease inhibitors through imple-
mentation of non-natural (triazole) units in their activity-
defining regions.
Entry
Sequence
Ki [nm][a]
Relative activity
1
2
3
4
5
6
wt
GRCTKSIAPICFPD
GRCTKSIPAICFPD
GRCTKS[IcA]PICFPD
GRCTKS[ItA]PICFPD
GRCTKSI[PcA]ICFPD
GRCTKSI[PtA]ICFPD
GRCTKSIPPICFPD
178ꢀ25
3.2ꢀ0.5
34ꢀ5
1[b]
1[c]
0.2[b]
1.7[b]
80[c]
2[c]
302ꢀ50
255ꢀ42
6.3ꢀ1.8
0.2ꢀ0.03
–
[a] Error calculated by propagation of error for Kiapp and KM (see the
Supporting Information); relative activities are calculated as the ratios of
the Ki values for the respective compound to that of [b] 1 or [c] 2.
Received: December 20, 2011
Published online: February 28, 2012
Keywords: amide mimics · Bowman–Birk inhibitors ·
.
protease inhibitors · structure elucidation
tion, and no relevant contacts between residue 9 and trypsin
were observed in the crystal structure (Figure 3c). In contrast,
an induced conformational mismatch (5) caused a drastic
reduction of binding affinity.
[1] a) J. Song, K. Burrage, Z. Yuan, T. Huber, BMC Bioinf. 2006, 7,
124; b) A. Jabs, M. S. Weiss, R. Hilgenfeld, J. Mol. Biol. 1999,
286, 291 – 304; c) D. Pal, P. Chakrabarti, J. Mol. Biol. 1999, 294,
271 – 288; d) D. E. Stewart, A. Sarkar, J. E. Wampler, J. Mol.
Biol. 1990, 214, 253 – 260; e) M. S. Weiss, A. Jabs, R. Hilgenfeld,
Nat. Struct. Biol. 1998, 5, 676.
[2] a) W. S. Horne, C. A. Olsen, J. M. Beierle, A. Montero, M. R.
Ghadiri, Angew. Chem. 2009, 121, 4812 – 4818; Angew. Chem.
Int. Ed. 2009, 48, 4718 – 4724; b) K. P. Exarchos, T. P. Exarchos,
G. Rigas, C. Papaloukas, D. I. Fotiadis, BMC Bioinf. 2011, 12,
142; c) C. Freund, P. Gehrig, A. Baici, T. A. Holak, A.
Plꢁckthun, Folding Des. 1998, 3, 39 – 49; d) M. Keller, C.
Boissard, L. Patiny, N. N. Chung, C. Lemieux, M. Mutter, P. W.
Schiller, J. Med. Chem. 2001, 44, 3896 – 3903; e) A. Tam, U.
Arnold, M. B. Soellner, R. T. Raines, J. Am. Chem. Soc. 2007,
129, 12670 – 12671; f) S. Chierici, M. Jourdan, M. Figuet, P.
Dumy, Org. Biomol. Chem. 2004, 2, 2437 – 2441.
[3] K. P. Lu, G. Finn, T. H. Lee, L. K. Nicholson, Nat. Chem. Biol.
2007, 3, 619 – 629.
[4] a) R. L. Stein, Adv. Protein Chem. 1993, 44, 1 – 24; b) F. X.
Schmid, L. M. Mayr, M. Mꢁcke, E. R. Schçnbrunner, Adv.
Protein Chem. 1993, 44, 25 – 66.
[5] a) Y. Che, G. R. Marshall, Biopolymers 2006, 81, 392 – 406; b) M.
Keller, C. Sager, P. Dumy, M. Schutkowski, G. S. Fischer, M.
Mutter, J. Am. Chem. Soc. 1998, 120, 2714 – 2720.
[6] a) J. Zabrocki, G. D. Smith, J. B. Dunbar, H. Iijima, G. R.
Marshall, J. Am. Chem. Soc. 1988, 110, 5875 – 5880; b) G. D.
Smith, J. Zabrocki, T. A. Flak, G. R. Marshall, Int. J. Pept.
Protein Res. 1991, 37, 191 – 197; c) J. Zabrocki, J. B. Dunbar,
K. W. Marshall, M. V. Toth, G. R. Marshall, J. Org. Chem. 1992,
57, 202 – 209; d) R. J. Nachman, J. Zabrocki, J. Olczak, H. J.
Williams, G. Moyna, A. Ian Scott, G. M. Coast, Peptides 2002, 23,
709 – 716.
[7] a) Y. Hitotsuyanagi, S. Motegi, H. Fukaya, K. Takeya, J. Org.
Chem. 2002, 67, 3266 – 3271; b) D. Tietze, M. Tischler, S. Voigt,
D. Imhof, O. Ohlenschlꢂger, M. Gçrlach, G. Buntkowsky, Chem.
Eur. J. 2010, 16, 7572 – 7578.
[8] T. W. Muir, Annu. Rev. Biochem. 2003, 72, 249 – 289.
[9] a) M. Nahrwold, T. Bogner, S. Eissler, S. Verma, N. Sewald, Org.
Lett. 2010, 12, 1064 – 1067; b) E. Valverde, F. Lecaille, G.
Lalmanach, V. Aucagne, F. Delmas, Angew. Chem. 2012, 124,
742 – 746; Angew. Chem. Int. Ed. 2012, 51, 718 – 722.
[10] a) A. Clemente, G. Sonnante, C. Domoney, Curr. Protein Pept.
Sci. 2011, 12, 358 – 373; b) J. D. McBride, A. B. Brauer, M. Nievo,
R. J. Leatherbarrow, J. Mol. Biol. 1998, 282, 447 – 458.
[11] a) P. Li, S. Jiang, S. L. Lee, C. Y. Lin, M. D. Johnson, R. B.
Dickson, C. J. Michejda, P. P. Roller, J. Med. Chem. 2007, 50,
Nevertheless, it must be mentioned that as a result of its
intrinsic architecture, the 1,4-disubstituted 1,2,3-triazole unit
as a trans amide backbone element increases the distance
between the a-carbons of adjacent residues by 1.3 ꢀ. In
contrast, the 1,5-disubstituted 1,2,3-triazolyl cis amide mimic
preserves bond length and spatial arrangement almost
perfectly (increase of distance between a-carbons: 0.376 ꢀ).
Interestingly, the crystal structure of peptidomimetic 6
revealed that the predominant orientation of the incorpo-
rated 1,4-disubstituted 1,2,3-triazole significantly differed
from the geometry predicted from the hydrogen-bond profile
(see the Supporting Information).[9b]
In summary, we have demonstrated that 1,4- and 1,5-
disubstituted 1,2,3-triazoles can easily be introduced into the
backbone of synthetic peptides as viable amide surrogates for
the generation of locked trans or cis conformations. Only
peptidomimetics resembling the native-like amide isomers
showed high inhibition potency (two-digit nanomolar Ki
values). Our approach allowed for the establishment of the
rather uncommon cis conformation in sensitive, structure-
defining regions of the functional loop of BBI SFTI-1 through
the introduction of non-prolyl backbone motifs. The exchange
of a proline residue following a cis amide bond by any of the
other 19 natural amino acids usually leads to a formation of
the undesired trans isomer. The presented synthetic strategy,
in principle, enables the installation of each desired side chain
without violating the backbone structure, thus, providing
access to tailor-made cis amide mimics previously excluded
from rational design.
Generally, SFTI-1 is a valuable framework for the
development of selective protease inhibitors of diagnostic
and therapeutic relevance.[11a,19] Since all BBIs share a con-
served cis-Pro motif in their functional loop, data and
structural insights validated for the SFTI-1 scaffold may be
extrapolated to the whole inhibitor family.[20] Nevertheless, as
other trypsin-like proteases of pharmaceutical relevance may
differ in their amino acid composition close to the enzyme
active pocket, the loss of binding affinity for the cis amide
mimic 3, presumably caused by changes in the intermolecular
hydrogen-bonding pattern, might not be a general problem.
Angew. Chem. Int. Ed. 2012, 51, 3708 –3712
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3711

.
Angewandte
Communications
5976 – 5983; b) U. C. Marx, M. L. Korsinczky, H. J. Schirra, A.
Jones, B. Condie, L. Otvos, Jr., D. J. Craik, J. Biol. Chem. 2003,
278, 21782 – 21789; c) M. L. Korsinczky, H. J. Schirra, K. J.
Rosengren, J. West, B. A. Condie, L. Otvos, M. A. Anderson,
D. J. Craik, J. Mol. Biol. 2001, 311, 579 – 591; d) M. L. Kor-
sinczky, H. J. Schirra, D. J. Craik, Curr. Protein Pept. Sci. 2004, 5,
351 – 364.
[16] a) K. Holland-Nell, M. Meldal, Angew. Chem. 2011, 123, 5310 –
5312; Angew. Chem. Int. Ed. 2011, 50, 5204 – 5206; b) M. Roice,
I. Johannsen, M. Meldal, QSAR Comb. Sci. 2004, 23, 662 – 673;
c) B. C. Boren, S. Narayan, L. K. Rasmussen, L. Zhang, H. Zhao,
Z. Lin, G. Jia, V. V. Fokin, J. Am. Chem. Soc. 2008, 130, 8923 –
8930.
[17] S. Luckett, R. S. Garcia, J. J. Barker, A. V. Konarev, P. R.
Shewry, A. R. Clarke, R. L. Brady, J. Mol. Biol. 1999, 290,
525 – 533.
[18] a) M. Juricek, P. H. J. Kouwer, A. E. Rowan, Chem. Commun.
2011, 47, 8740 – 8749; b) R. M. Meudtner, M. Ostermeier, R.
Goddard, C. Limberg, S. Hecht, Chem. Eur. J. 2007, 13, 9834 –
9840.
[12] A. B. E. Brauer, G. J. Domingo, R. M. Cooke, S. J. Matthews,
R. J. Leatherbarrow, Biochemistry 2002, 41, 10608 – 10615.
[13] M. Empting, O. Avrutina, R. Meusinger, S. Fabritz, M. Rein-
warth, M. Biesalski, S. Voigt, G. Buntkowsky, H. Kolmar, Angew.
Chem. 2011, 123, 5313 – 5317; Angew. Chem. Int. Ed. 2011, 50,
5207 – 5211.
[14] D. Scarpi, E. G. Occhiato, A. Trabocchi, R. J. Leatherbarrow,
A. B. Brauer, M. Nievo, A. Guarna, Bioorg. Med. Chem. 2001, 9,
1625 – 1632.
[15] a) H. D. Dickson, S. C. Smith, K. W. Hinkle, Tetrahedron Lett.
2004, 45, 5597 – 5599; b) J. T. Lundquist IV, J. C. Pelletier, Org.
Lett. 2001, 3, 781 – 783; c) S. Ohira, Synth. Commun. 1989, 19,
561 – 564.
[19] C. Yuan, L. Chen, E. J. Meehan, N. Daly, D. J. Craik, M. Huang,
J. C. Ngo, BMC Struct. Biol. 2011, 11, 30.
[20] A. B. Brauer, J. D. McBride, G. Kelly, S. J. Matthews, R. J.
Leatherbarrow, Bioorg. Med. Chem. 2007, 15, 4618 – 4628.
[21] O. Hartley, H. Gaertner, J. Wilken, et al., Proc. Natl. Acad. Sci.
USA 2004, 101, 16460 – 16465.
3712
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 3708 –3712