Epimerization-free block synthesis of peptides from thioacids and amines with the sanger and mukaiyama reagents

Full text of DOI:10.1002/anie.200805782

Angewandte
Chemie
DOI: 10.1002/anie.200805782
Chemical Ligation
Epimerization-Free Block Synthesis of Peptides from Thioacids and
Amines with the Sanger and Mukaiyama Reagents
David Crich* and Indrajeet Sharma
Dedicated to Professor Carl R. Johnson
The reaction of 2,4-dinitrobenzenesulfonamides with thioa-
cids to give amides was described in 1998 by Tomkinson and
co-workers.[1] We extended this reaction to the synthesis of
peptides and their glycoconjugates by combining amino
thioacids and their C-terminal-peptide congeners with N-
terminal sulfonamides.[2a] In a three-component coupling, we
used this process to capture thioacids generated in situ from
the nucleophilic ring-opening of a variety of cyclic thioanhy-
drides.[2b] These reactions involve nucleophilic aromatic
substitution of the thiocarboxylate group on the electron-
deficient sulfonamide with the formation of a highly reactive
S-(2,4-dinitrophenyl) thioester, and its subsequent condensa-
tion with the released amine.[2a,3–6] We show herein that amide
bonds can be formed by the coupling of thioacids with amines
under the aegis of 2,4-dinitrofluorobenzene and other elec-
tron-deficient arenes and heteroarenes at ambient temper-
ature.
The Sanger reagent, 2,4-dinitrofluorobenzene, was tradi-
tionally used for the degradation of peptides and proteins
through its reaction with N-terminal amines and for the
identification of N-terminal amino acids.[7] Sanger noted,
however, that thiols and other nucleophiles, including the side
chains of histidine and tyrosine, competed with amines for
capture by 2,4-dinitrofluorobenzene.[7a] We surmised that a
thiocarboxylate salt would react significantly more rapidly
than an amine with this reagent and thus enable the formation
of the thioester and ultimately peptide synthesis. In this way, a
peptide degradation reagent could effectively be turned into a
reagent for their synthesis (Scheme 1).
Scheme 1. Amide-forming reaction with the Sanger reagent.
Table 1: Preparation of thioesters from protected amino acids and
thiols.[a]
Reagent
Thioester
Yield [%]
DCC
DCC
DCC
DIC
DIC
DIC
Boc-l-Val-SFm
Boc-Aib-SFm
Boc-l-Asp-(a-OBn)-g-SFm
Fmoc-l-Ala-STmob
Boc-l-Phe-STmob
Z-l-Val-STrt
99
82
97
98
96
95
72
82
PyBop
PyBop
Z-l-Ala-l-Phe-SFm
Boc-l-Lys(Boc)-l-Arg(Pbf)-l-Asn(Trt)-
-l-Arg(Pbf)-SFm
[a] Aib=2-aminoisobutyric acid, Bn=benzyl, Boc=tert-butoxycarbonyl,
DCC=dicyclohexylcarbodiimide,
Fmoc=9-fluorenylmethoxycarbonyl,
DIC=diisopropylcarbodiimide,
FmSH=9-fluorenylmethylthiol,
Pbf=2,2,4,6,7-pentamethyldihydrobenzofuran-5-sulfonyl, PyBop=(1H-
benzotriazol-1-yloxy)tris(pyrrolidino)phosphonium hexafluorophos-
phate, TmobSH=2,4,6-trimethoxybenzylthiol, TrtSH=triphenylmetha-
nethiol, Z=benzyloxycarbonyl.
A series of thioesters was prepared by the coupling of
suitably protected amino acids with 9-fluorenylmethane-
thiol,[2a] 2,4,6-trimethoxybenzyl thiol,[8] or triphenylmethane-
thiol (Table 1). With simple carbamate-protected amino acids,
carbodiimide coupling reagents were employed for these
condensations; however, with peptides (Table 1, last two
entries) we preferred to use the reagent (1H-benzotriazol-1-
yloxy)tris(pyrrolidino)phosphonium
hexafluorophosphate
(PyBop) recommended by Kajihara and co-workers[9] to
avoid the problem of epimerization in the thioesterification of
all but simple amino acids. The final entry in Table 1 is
noteworthy for both the sterically hindered nature of the
reaction site and the fact that no epimerization was observed
under the PyBop coupling conditions. When the same
thioester was prepared under the carbodiimide–hydroxyben-
zotriazole conditions reported[10] for epimerization-free thio-
ester synthesis, the product was obtained in good yield (77%)
but as a 3:1 mixture of epimers. The requisite thioacids were
released cleanly from the thioesters with piperidine[2a] in the
case of the fluorenylmethyl thioesters, or with trifluoroacetic
acid for the trimethoxybenzyl thioesters and triphenylmethyl
thioesters,[11] and used immediately in the peptide-bond-
forming reactions.
[*] Prof. Dr. D. Crich, I. Sharma
Department of Chemistry, Wayne State University
5101 Cass Avenue, Detroit, MI 48202 (USA)
Prof. Dr. D. Crich
Centre de Recherche CNRS de Gif-sur-Yvette
Institut de Chimie des Substances Naturelles
Avenue de la Terrasse, 91198 Gif-sur-Yvette (France)
Fax : (+33)1-6907-7752
E-mail: dcrich@icsn.cnrs-gif.fr
The coupling of thioacids with amines in the presence of
the Sanger reagent was first investigated by using thioacetic
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200805782.
Angew. Chem. Int. Ed. 2009, 48, 2355 –2358
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2355

Communications
Table 2: Coupling of thioacids with amines with the Sanger reagent.[a]
Entry
Thioester
Release
Thioacid
Amine
Product
Yield [%]
reagent[b]
1
2
3
4
5
6
7
8
9
–
–
–
–
AcSH
AcSH
Boc-l-Val-SH
Boc-l-Val-SH
Boc-l-Val-SH
Boc-Aib-SH
l-Trp
Ac-l-Trp
Ac-l-Trp-OMe
76
99
84
82
76
95
96
93
60
59
64
66
l-Trp-OMe·HCl
l-Phe-OMe·HCl
d-Phe-OMe·HCl
l-Trp-OMe·HCl
l-Trp-OMe·HCl
Gly-OEt·HCl
l-Phe-OMe·HCl
Gly-OEt·HCl
l-Phe-OMe·HCl
l-Tyr-OMe·HCl
Gly-Gly-OMe·HCl
Boc-l-Val-SFm
Boc-l-Val-SFm
Boc-l-Val-SFm
Boc-Aib-SFm
Pip
Pip
Pip
Pip
Pip
Pip
TFA
TFA
TFA
Pip
Boc-l-Val-l-Phe-OMe
Boc-l-Val-d-Phe-OMe
Boc-l-Val-l-Trp-OMe
Boc-Aib-l-Trp-OMe
Boc-Aib-Gly-OEt
Boc-l-Asp-(a-OBn)-g-l-Phe-OMe
Fmoc-l-Ala-Gly-OEt
Fmoc-l-Ala-l-Phe-OMe
Fmoc-l-Ala-l-Tyr-OMe
Z-l-Ala-l-Phe-Gly-Gly-OMe
Boc-Aib-SFm
Boc-Aib-SH
Boc-l-Asp-(a-OBn)-g-SFm
Fmoc-l-Ala-STmob
Fmoc-l-Ala-STmob
Fmoc-l-Ala-STmob
Z-l-Ala-l-Phe-SFm
Boc-l-Asp-(a-OBn)-g-SH
Fmoc-l-Ala-SH
Fmoc-l-Ala-SH
Fmoc-l-Ala-SH
Z-l-Ala-l-Phe-SH
10
11
12
[a] In all reactions, the amine hydrochloride (0.1m) was used in N,N-dimethylformamide (DMF) with the thioacid (1.2 equiv), and with Cs2CO3
(1.5 equiv) as the base (see the Supporting Information). [b] Pip=piperidine, TFA=trifluoroacetic acid.
Table 3: Comparison of the Sanger reagent with 2,4-dinitroiodobenzene
and with the Mukaiyama reagent for the construction of amide bonds
between amines and Fmoc-l-Ala-SH.[a]
acid with l-tryptophan and its methyl ester (Table 2, entries 1
and 2). The acetamide was obtained in excellent yield when
the three components were mixed at room temperature in the
presence of cesium carbonate. A series of dipeptides was then
accessed by coupling of the released thioacids with other
amino esters and peptide esters. Our premise that deproton-
ated thioacids would be more nucleophilic toward the Sanger
reagent than amines was validated by the yields of the
dipeptides (Table 2); 2,4-dinitrofluorobenzene is an effective
reagent for the formation of amide bonds between thioacids
and amines. The preparation of both the l,l and l,d isomers
of Boc-Val-Phe-OMe (Table 2, entries 3 and 4) verified the
absence of epimerization, as also deduced by the inspection of
“high-field” NMR spectra. The suitability of this method for
the formation of hindered peptide bonds is evident from
entries 6 and 7 of Table 2.
Dipeptide yield [%]
Amine
Sanger
reagent
2,4-dinitroiodo-
benzene
Mukaiyama
reagent
Gly-OEt·HCl
60
59
64
–
83
79
86
78
82
80
86
85
82
88
l-Phe-OMe·HCl
l-Tyr-OMe·HCl
l-Ala-OMe·HCl
l-Cys(Trt)-OEt·HCl
–
[a] In all reactions, the amine hydrochloride (0.1m) was used in DMF
with the thioacid (1.2 equiv), and with Cs2CO3 (1.5 equiv) as the base
(see the Supporting Information).
The yields observed with the Fmoc-protected thioacids
(Table 2, entries 9–11) were lower than those observed with
Boc-protected thioacids as a result of partial cleavage of the
Fmoc group by the fluoride anion released in the course of the
nucleophilic aromatic substitution step.[12] Accordingly, we
turned to 2,4-dinitroiodobenzene and 2-chloro-1-methylpyr-
idinium iodide (the Mukaiyama reagent)[13a] as condensing
agents (Scheme 2, Table 3).
The Mukaiyama reagent and its analogues have been
employed previously to effect peptide coupling reactions
between amines and simple carboxylic acids.[13b,c] However,
the yields of these coupling reactions are modest, except when
secondary and sterically hindered amines are used.[13b,c,14] In
contrast, the method described herein is applicable to all
amines owing to the greater nucleophilicity of the thiocar-
boxylate moiety. The increased reactivity of thiocarboxylate
groups over simple carboxylate groups is illustrated by a
competition reaction (Scheme 3),[15] in which the reactivity of
a carboxylic acid was compared directly with that of its thio
analogue. The tripeptide product isolated from this reaction
was shown by mass spectrometry to be fully deuterium-
labeled in the glycine residue, a result indicating the
superiority of the thioacid in this chemistry.
Scheme 3. Carboxylate/thiocarboxylate competition reaction.
DMAP=4-dimethylaminopyridine, EDCI=1-ethyl-3-(3’-dimethylamino-
propyl)carbodiimide hydrochloride.
Scheme 2. Amide-forming reaction with the Mukaiyama reagent.
2356
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 2355 –2358

Angewandte
Chemie
The coupling of a valine-derived thioacid with Val-OtBu
in the presence of Z-l-Arg-OH resulted in the formation of a
single dipeptide in high yield. This result reinforces the
superior reactivity of the thioacid over the simple acid and
also demonstrates the compatibility of the method with the
guanidine and carboxylic acid moieties of the arginine
derivative (Scheme 4).[16]
Scheme 6. Comparative block syntheses of an octapeptide (see Table 4
for conditions and yields for the coupling step).
Scheme 4. Functional-group compatibility.
Table 4: Comparative block syntheses of an octapeptide (coupling step
in Scheme 6).[a]
A dipeptide was also formed in good yield in a coupling
reaction of a valine-derived thioacid with the Mukaiyama
reagent in aqueous buffer (Scheme 5).[21]
X
Coupling reagent[b]
t [h]
Yield [%]
Epimer ratio[c]
SFm
OH
OH
OH
OH
1) Pip; 2) Mukaiyama
PyBop
HATU
EDCI
EDCI-HOBt
1
8
6
14
8
66
69
70
54
63
>99:1
93.7:6.3
93.1:6.9
83:17
94.2:5.8
[a] All coupling reactions were conducted in DMF under standard
conditions (temperature and concentration) with epimerization-free, all-
l
tetrapeptides. [b] HATU=N-[(dimethylamino)-1H-1,2,3-triazolo[4,5-
Scheme 5. Coupling under aqueous conditions. HEPES=4-(2-hydrox-
yethyl)-1-piperazineethanesulfonic acid, NMP=N-methylpyrrolidone,
Gn=guanidine.
b]pyridine-1-ylmethylene]-N-methylmethanaminium hexafluorophos-
phate N-oxide, HOBT=N-hydroxybenzotriazole. [c] Epimerization
ratios were determined by comparison with an authentic sample of
Boc-Lys(Boc)-Arg(Pbf)-Asn(Trt)-d-Arg(Pbf)-Asn(Trt)-Asn(Trt)-Ile-Ala-
OEt.
Finally, we describe the application of this new peptide-
bond-forming reaction to the 4+4 block synthesis of l-Lys-l-
Arg-l-Asn-l-Arg-l-Asn-l-Asn-l-Ile-l-Ala. This octapeptide
is the C-terminal sequence of oxyntomodulin from porcine
jejunoileum and is responsible for the inhibitory properties of
this extended version of glucagon towards the secretion of
gastric acid.[17] This octapeptide was selected as a target
because of the widespread current interest in oxyntomodulin
derivatives as appetite suppressants and as potential ther-
apeutics for obesity,[18] and because the concatenation of its
multiple functionalized side chains provides a true test of the
methodology.
Two tetrapeptides were synthesized by standard manual
solution-phase techniques (see the Supporting Informa-
tion).[19] An allyl ester protecting group was then removed
from the C-terminal arginine residue, and the resulting acid
was converted into the 9-fluorenylmethyl thioester, as set out
in the final entry of Table 1. The 4+4 fragment coupling was
carried out by liberation of the thioacid functionality with
piperidine, followed by combination with the N-terminal
tetrapeptide in the presence of the Mukaiyama reagent. The
protected target octapeptide was isolated in 66% yield as a
single epimer (Scheme 6, Table 4). This block synthesis
provided an opportunity for the comparison of the thioacid
methodology with more common methods. To this end, the
coupling of the tetrapeptide carboxylic acid was effected with
PyBop, HATU, and EDCI[20] under standard conditions with
the results indicated in the Table 4. The PyBop and HATU
methods gave the protected octapeptide in comparable yields
to that observed with the thioacid method but required
significantly longer reaction times for a similar result, and
some epimerization occurred. The reaction was not complete
with the carbodiimide method after 14 h and resulted in
significantly more epimerization.[21,22] Finally, the treatment
of this octapeptide with trifluoroacetic acid in dichloro-
methane removed the complete suite of acid-labile protecting
groups and afforded the octapeptide as a single epimer in
63% yield after purification by reversed-phase HPLC.
The high nucleophilicity of thiocarboxylates enables their
preferential reaction with electron-deficient aromatic halides
by nucleophilic substitution in the presence of an amine. This
method generates highly reactive thioesters in situ, which
then react with the amine to form peptide bonds, and tolerates
the presence of all proteinogenic amino acid side chains,
except for cysteine and lysine,[2a] which must be protected. No
particular amino acid is required for coupling, in contrast to
native chemical ligation and its variants,[6a,23] and peptide-
bond formation at hindered residues functions efficiently,
unlike peptide ligation by direct aminolysis.[24]
Received: November 27, 2008
Published online: February 19, 2009
Keywords: aromatic substitution · chemical ligation · peptides ·
.
thioacids · thioesters
[1] a) T. Messeri, D. D. Sternbach, N. C. O. Tomkinson, Tetrahedron
Lett. 1998, 39, 1669 – 1672; b) T. Messeri, D. D. Sternbach,
N. C. O. Tomkinson, Tetrahedron Lett. 1998, 39, 1673 – 1676.
[2] a) D. Crich, K. Sana, S. Guo, Org. Lett. 2007, 9, 4423 – 4426; b) D.
Crich, A. A. Bowers, Org. Lett. 2007, 9, 5323 – 5325.
Angew. Chem. Int. Ed. 2009, 48, 2355 –2358
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2357

Communications
[3] For peptide synthesis by “superarmed” esters synthesized by
[15] In the preparation of the dideuterated dipeptide thioacid, a Boc
group was removed in the presence of the Tmob thioester.
[16] Parallel reactions of Val-OtBu with Z-Gly-Phe-SH and Z-Gly-
Phe-OH were also conducted with the Mukaiyama reagent. The
tripeptide (Z-Gly-Phe-Val-OtBu) was formed in 76% yield from
the reaction of the thioacid, whereas the corresponding carbox-
ylic acid gave the product in only 23% yield under identical
conditions. No epimerization was observed in either experiment,
as determined by comparison with an authentic sample of Z-
Gly-d-Phe-Val-OtBu (see the Supporting Information).
[17] C. Carles-Bonnet, J. Martinez, C. Jarrousse, A. Aumelas, H. Niel,
D. Bataille, Peptides 1996, 17, 557 – 561.
[18] a) R. Nogueiras, S. J. Caton, D. Perez-Tilve, M. Bidlingmaier,
M. H. Tschꢁp, Drug Discovery Today Dis. Mech. 2006, 3, 463 –
470; b) K. Wynne, S. R. Bloom, Nat. Clin. Pract. Endocrinol.
Metab. 2006, 2, 612 – 620.
[19] a) Houben-Weyl, Methods in Organic Chemistry: Synthesis of
Peptides and Peptidomimetics, Vol. E22a–e (Eds.: M. Goodman,
A. F. Mahwah, L. Moroder, C. Toniolo), Thieme, Stuttgart,
2002–2003; b) M. Bodanszky, Peptide Chemistry: A Practical
Textbook, 2nd ed., Springer, Berlin, 1993; c) M. Bodanszky, A.
Bodanszky, The Practice of Peptide Synthesis, 2nd ed., Springer,
Heidelberg, 1994.
[20] a) S. Y. Han, Y. A. Kim, Tetrahedron 2004, 60, 2447 – 2467;
b) C. A. G. N. Montalbetti, V. Falque, Tetrahedron 2005, 61,
10827 – 10852, and references therein.
[21] Although we have not conducted any confirmatory experiments,
the rapidity of the in situ activation and coupling of thioacids
reported herein stands in contrast to the synthesis of peptides by
the two-step “superarmed”-ester protocol, which requires a
relatively slow reaction of the acid with a 4-(4,6-dimethoxy-1,3,5-
triazin-2-yl)-4-methylmorpholinium tetrafluoroborate salt, fol-
lowed by addition of the amine.[3]
´
nucleophilic aromatic substitution, see: Z. J. Kaminski, B.
´
´
Kolesinska, J. Kolesinska, G. Sabatino, M. Chelli, P. Rovero,
M. Błaszczyk, M. L. Głꢀwka, A. M. Papini, J. Am. Chem. Soc.
2005, 127, 16912 – 16920.
[4] For peptide synthesis with nitrophenyl thioesters formed by the
conventional condensation of thiols with acids, see: A. Ishiwata,
T. Ichiyanagi, M. Takatani, Y. Ito, Tetrahedron Lett. 2003, 44,
3187 – 3190.
[5] For previous applications of thioacids in peptide synthesis, see:
a) D. Yamashiro, J. F. Blake, Int. J. Pept. Protein Res. 1981, 18,
383 – 392; b) J. Blake, Int. J. Pept. Protein Res. 1981, 17, 273 – 274;
c) Y. V. Mitin, N. P. Zapevalova, Int. J. Pept. Protein Res. 1990,
35, 352 – 356.
[6] For the use of thioacids as precursors to thioesters for native
chemical ligation, see: a) P. E. Dawson, T. W. Muir, I. Clark-
Lewis, S. B. H. Kent, Science 1994, 266, 776 – 779; b) J. P. Tam,
Y.-A. Lu, C. F. Liu, J. Shao, Proc. Natl. Acad. Sci. USA 1995, 92,
12485 – 12489; c) P. E. Dawson, S. B. H. Kent, Annu. Rev.
Biochem. 2000, 69, 923 – 960.
[7] a) F. Sanger, Biochem. J. 1945, 39, 507 – 515; b) H. N. Eisen, S.
Belman, M. E. Carston, J. Am. Chem. Soc. 1953, 75, 4583 – 4585;
c) G. T. Hermanson, Bioconjugate Techniques, Academic Press,
San Diego, 1996.
[8] S. Vetter, Synth. Commun. 1998, 28, 3219 – 3223.
[9] T. J. Hogenauer, Q. Wang, A. K. Sanki, A. J. Gammon, C. H. L.
Chu, C. Kaneshiro, Y. Kajihara, K. Michael, Org. Biomol. Chem.
2007, 5, 759 – 762.
[10] a) P. Wang, L. P. Miranda, Int. J. Pept. Res. Ther. 2005, 11, 117 –
123; b) H. Yang, H. Li, R. Wittenburg, M. Egi, W. Huang, L. S.
Liebeskind, J. Am. Chem. Soc. 2007, 129, 1132 – 1140; c) H.
Yang, L. S. Liebeskind, Org. Lett. 2007, 9, 2993 – 2995; d) S.
Ficht, R. J. Payne, R. T. Guy, C.-H. Wong, Chem. Eur. J. 2008, 14,
3620 – 3629.
[22] As acids are activated, albeit more slowly, by the Mukaiyama
reagent, protection of side-chain acid groups is advisable when
the thioacid coupling method is used.
[11] N. Shangguan, S. Katukojvala, R. Greenberg, L. J. Williams, J.
Am. Chem. Soc. 2003, 125, 7754 – 7755.
[12] For the cleavage of Fmoc groups with tetrabutylammonium
fluoride, see: a) Y. Rew, D. Shin, I. Hwang, D. L. Boger, J. Am.
Chem. Soc. 2004, 126, 1041 – 1043; b) W. R. Li, J. Jiang, M. M.
Joullie, Synlett 1993, 362; c) M. Ueki, M. Amemiya, Tetrahedron
Lett. 1987, 28, 6617 – 6620.
[13] a) T. Mukaiyama, Angew. Chem. 1979, 91, 798 – 812; Angew.
Chem. Int. Ed. Engl. 1979, 18, 707 – 808; b) P. Li, J. C. Xu,
Tetrahedron 2000, 56, 8119 – 8131; c) S. Crosignani, J. Gonzalez,
D. Swinnen, Org. Lett. 2004, 6, 4579 – 4582.
[14] As is implicit in the results of Sanger, simple peptide-based
primary amines are more nucleophilic than carboxylic acids
toward electron-deficient arenes. Steric bulk reduces the nucle-
ophilicity of the amine and enables the carboxylate group to
compete. The poor yields observed in the coupling of less
hindered amines with carboxylic acids in the presence of the
Mukaiyama reagent are readily shown by mass spectrometry to
arise from competing attack of the amine on the reagent.
[23] a) D. Crich, A. Banerjee, J. Am. Chem. Soc. 2007, 129, 10064 –
10065; b) P. Botti, S. Tchertchian, WO 2006/133962, 2006; c) R.
Okamoto, Y. Kajihara, Angew. Chem. 2008, 120, 5482 – 5486;
Angew. Chem. Int. Ed. 2008, 47, 5402 – 5406; d) R. Quaderer, D.
Hilvert, Chem. Commun. 2002, 2620 – 2621; e) R. J. Hondal,
B. L. Nilsson, R. T. Raines, J. Am. Chem. Soc. 2001, 123, 5140 –
5141; f) C. Haase, H. Rohde, O. Seitz, Angew. Chem. 2008, 120,
6912 – 6915; Angew. Chem. Int. Ed. 2008, 47, 6807 – 6810; g) J.
Chen, Q. Wan, Y. Yuan, J. Zhu, S. J. Danishefsky, Angew. Chem.
2008, 120, 8649 – 8652; Angew. Chem. Int. Ed. 2008, 47, 8521 –
8524; h) J. W. Bode, Curr. Opin. Drug Discovery Dev. 2006, 9,
765 – 775; i) D. Macmillan, Angew. Chem. 2006, 118, 7830 – 7834;
Angew. Chem. Int. Ed. 2006, 45, 7668 – 7672.
[24] R. J. Payne, S. Ficht, W. A. Greenberg, C.-H. Wong, Angew.
Chem. 2008, 120, 4483 – 4487; Angew. Chem. Int. Ed. 2008, 47,
4411 – 4415.
2358
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 2355 –2358