Protein chemical synthesis by ligation of peptide hydrazides

Article abstract of DOI:10.1002/anie.201100996

pH determines selectivity: The ligation of peptide hydrazides is a new method for protein chemical synthesis that is complementary to native chemical ligation. Peptide hydrazides may be the long-sought reagent equivalent to a "thioester synthon", one that

Full text of DOI:10.1002/anie.201100996

DOI: 10.1002/anie.201100996
Protein Chemical Synthesis
Protein Chemical Synthesis by Ligation of Peptide Hydrazides**
Ge-Min Fang, Yi-Ming Li, Fei Shen, Yi-Chao Huang, Jia-Bin Li, Yun Lin, Hong-Kui Cui, and
Lei Liu*
Dedicated to Professor Ronald Breslow on the occasion of his 80th birthday
Protein chemical synthesis can overcome the potential
limitations of protein expression and produce proteins with
predesigned changes and modifications with atomic preci-
sion.[1] Key to the success of modern protein chemical
synthesis is the chemoselective ligation reaction. The most
successful ligation method is the native chemical ligation
developed by Kent et al.[2] It involves a chemoselective
reaction between a C-terminal peptide thioester and an N-
terminal cysteine (Cys).[3] Both synthetic and recombinant
Scheme 1. Ligation of peptide hydrazides.
peptides can be used in native chemical ligation.[2,4] The utility
of native chemical ligation has been demonstrated by the total
and semisynthesis of a variety of proteins, which enables the
broad application of synthetic chemistry to the study of
protein biology.[5] Although native chemical ligation is a
transformative method, its applicability can sometimes be
limited in two respects: 1) peptide thioesters remain challeng-
ing to synthesize with Fmoc chemistry;[6] 2) convergent
synthesis of larger proteins without using protecting groups
requires an “orthogonal” amide-forming ligation chemistry,
that is, one that is compatible with the use of native chemical
ligation.[7]
In the present study we describe the ligation of peptide
hydrazides that is complementary to native chemical ligation
(Scheme 1). This method involves a chemoselective reaction
between a C-terminal peptide hydrazide and a Cys-peptide to
yield a native peptide bond. Notably, peptide hydrazides can
be readily prepared through either Boc- or Fmoc-based solid-
phase peptide synthesis (SPPS).[8] In addition, peptide hydra-
zide can be obtained through biological expression. Thus both
total and semisynthesis of proteins can be achieved with
ligation of peptide hydrazides. More importantly, the ligation
of peptide hydrazides enables sequential ligation in the N-to-
C direction.[9] It should be pointed out that peptide hydrazides
have been used in protein chemistry since the beginning of the
field.[10] However, the previous methods with peptide hydra-
zides are not chemoselective and necessitate protections at
Lys and Cys residues. The new ligation method described here
bypasses the requirement of protection groups, which is a
critical advance.[10]
Our study began with the ligation between model peptides
H-Leu-Tyr-Arg-Ala-Tyr-NHNH2 (1a) and H-Cys-Lys-Tyr-
Met-His-OH (2).[11] The ligation involves two steps that are
carried out in a one-pot fashion. In the first step the two
peptides (1.5 and 2.0 mm in final concentration, respectively)
were added together to the aqueous phosphate (0.2m) buffer
containing 6.0m guanidinium chloride (Gn·HCl). At a low pH
(3.0–7.0) and À108C, the oxidant (10 mm in final concen-
tration) was added to the ligation mixture presumably
producing a peptide azide. After 20 min, a thiol compound
was added (100 mm in final concentration) and the pH value
was adjusted to 7.0 to initiate the second step. The second step
was allowed to proceed for 2 h at RT before the yield was
determined by HPLC. As shown in Table 1, organic oxidants
(entries 1 and 2) do not provide good results, whereas NaNO2
can successfully stimulate the ligation to reach a high yield
(entries 3 and 5). The optimal pH value for the oxidation is
3.0–4.0 and MPAA (4-mercaptophenylacetic acid)[3] is impor-
tant for mediating the ligation.
With the optimized conditions in hand, we investigated
the scope of the ligation for 17 amino acids (Table 2). For
three other amino acids (Xaa = Gln, Asp, and Asn) we have
not been able to prepare the peptide hydrazides because
intramolecular cyclization occurs between the hydrazide and
C-terminal side-chain amide or acid group in SPPS (see the
Supporting Information).[12] In all cases, the first oxidation
step was conducted at pH 3.0 for 20 min, while for most of the
C-terminal amino acids the second step can reach completion
with a high yield (84–99%) in 2 h. Notably, unprotected Ser,
Thr, Tyr, His, Lys, and even Cys are well compatible with the
ligation. We did not observe any oxidized byproducts for Met
and Trp either. For more sterically hindered amino acids (e.g.
[*] G.-M. Fang, Y.-M. Li, F. Shen, Y.-C. Huang, J.-B. Li, Y. Lin, H.-K. Cui,
Prof. Dr. L. Liu
Key Laboratory of Bioorganic Phosphorus Chemistry & Chemical
Biology (Ministry of Education)
Department of Chemistry, Tsinghua University
Beijing 100084 (China)
Fax: (+86)10-6277-1149
E-mail: lliu@mail.tsinghua.edu.cn
Homepage: http://chem.tsinghua.edu.cn/liugroup/
[**] This work was supported by NSFC (no. 20802040 and no. 20932006)
and the national “973” grants from the Ministry of Science and
Technology (no. 2011CB965300).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201100996.
Angew. Chem. Int. Ed. 2011, 50, 7645 –7649
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7645

Communications
Table 1: Optimization of the ligation conditions.[a]
This experiment further confirmed that the ligation of peptide
hydrazides is chemoselective towards the N-terminal Cys.
To understand the mechanism of the ligation of peptide
hydrazides, we analyzed the intermediates of the ligation
between 1a and H-Cys-OH (Figure 1). We found that 1a was
Entry
pH of oxidation
Oxidant
Thiol
Yield [%]
1
2
3
4
5
6
7
8
3.0
3.0
3.0
3.0
4.0
5.0
6.0
7.0
tBuONO
isopentyl nitrite
NaNO2
NaNO2
NaNO2
NaNO2
NaNO2
NaNO2
MPAA
MPAA
MPAA
MESNa
MPAA
MPAA
MPAA
MPAA
57
32
96
36
95
0
0
0
[a] The first oxidation step was conducted at pH 3.0–7.0 and À108C in
6m Gn·HCl aqueous solution for 20 min. The second step was
conducted at pH 7.0 and RT after the addition of thiol (100 mm).
MES=2-mercaptoethane sulfonate.
Table 2: Scope of the ligation of peptide hydrazides.[a]
Figure 1. Intermediates of the ligation peptide hydrazides. Analytical
HPLC traces (215 nm) for the reaction solution: a) before addition of
NaNO2; b) after addition of NaNO2, t=20 min, pH 3.0; c) t=20 min,
pH has been adjusted to 7.0 and MPAA added; d) after addition of H-
Cys-OH, 2 h, pH 7.0.
Entry
Xaa (1)
Ligation time [h]
Yield [%]
1
2
3
4
5
6
7
8
Tyr (1a)
Ala (1b)
Met (1c)
Phe (1d)
Gly (1e)
His (1 f)
Arg (1g)
Ser (1h)
Thr (1i)
Cys (1j)
Glu (1k)
Lys (1l)
Trp (1m)
Leu (1n)
Val (1o)
Ile (1p)
Pro (1q)
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
5.0
2.0
2.0
2.0
2.0
3.0
7.0
7.0
9.0
94
94
91
96
99
93
95
91
95
96
84
90
92
90
86
86
84
cleanly oxidized by NaNO2 to the peptide azide 1a’ in less
than 20 min at pH 3.0. After MPAA had been added and the
pH value adjusted to 7, 1a’ was immediately converted to the
peptide thioester 1a’’. The thioester then reacted with H-Cys-
OH chemoselectively by native chemical ligation to give the
desired ligation product 4. Note that the N-terminal amino
group of 1a was not protected during the process. Similarly,
the Lys side chain does not require protection (Table 2).
These amino groups do not react with either the peptide azide
or peptide thioester at pH 3.0–7.0. Note that the Cys side
chain also does not need protection in the ligation of peptide
hydrazides. It has been known that the thiol group (RSH) may
be oxidized by HNO2 to form nitrothioite (RSNO) and
disulfide (RSSR).[13] Fortunately, both nitrothioite and disul-
fide can be cleanly reduced by other thiols to the original
thiol.[14] In the ligation of peptide hydrazides, the addition of
excess MPAA not only promotes the thioester exchange
reaction, but also reduces any oxidized Cys residues. Thus,
although the acyl azide condensation method is known,[10] the
present protocol advances it into a truly chemoselective
ligation method.
Like peptide thioesters used in native chemical ligation,
peptide hydrazides can be readily prepared with the Boc-
based SPPS.[8] However, unlike peptide thioesters that are not
stable to piperidine,[6] peptide hydrazides can be directly
synthesized with the Fmoc-based SPPS without any diffi-
culty.[11] Thus we expect the ligation of peptide hydrazides to
be particularly useful for the synthesis of phosphorylated and
glycosylated proteins,[15] where the peptide thioesters have to
be prepared through Fmoc chemistry. As an initial test of this
idea, we successfully synthesized H-Leu-Tyr-Arg-Ser[b-d-
Glc(OAc)4]-Ala-NHNH2 (5). This peptide was readily ligated
9
10
11
12
13
14
15
16
17
[a] The oxidation step was conducted at pH 3.0 and À108C in 6m
Gn·HCl aqueous solution for 20 min. The second step was conducted at
pH 7.0 and RT after the addition of MPAA (100 mm).
Thr, Val, Ile, and Pro), the second step requires longer time
(5–9 h) but the ligation yields are still good (84–95%). To test
whether the ligation involves racemization at the C-terminal
amino acid, we compared the ligations of benzyl-protected
Bz-Gly-(d)Xaa-NHNH2 and Bz-Gly-(l)Xaa-NHNH2 with H-
Cys-OH for three C-terminal amino acids (Xaa = Ala, Phe,
Cys) (see the Supporting Information). We also compared the
ligations of H-Leu-Tyr-Ala-Ala-(d)Tyr-NHNH2 and H-Leu-
Tyr-Ala-Ala-(l)Tyr-NHNH2 with H-Cys-OH (see the Sup-
porting Information). The results showed that the extent of
racemization is always less than 1%. Moreover, we conducted
competition experiments by ligating 1a with a mixture of H-
Lys-OH (20 equiv) and H-Cys-OH (1 equiv). Only H-Cys-
OH was found to participate the ligation reaction, whereas H-
Lys-OH remained inert (see the Supporting Information).
7646 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 7645 –7649

with 2 showing that the ligation of peptide hydrazides is
compatible with glycopeptides.
Application of the ligation of peptide hydrazides to model
protein synthesis is shown in Figure 2. In this example, the 42-
Figure 2. Synthesis of trifolitoxin[Ala23Cys]. Analytical HPLC traces
(215 nm) for the reaction mixture a) before ligation and b) after
ligation. The final product was characterized by MALDI-TOF mass
spectrometry.
Figure 3. Total synthesis of CssII. Analytical HPLC traces (215 nm) for
the reaction mixture a) after the first ligation, b) after the second
ligation, and c) after the third ligation. More details are available in the
Supporting Information.
mer peptide antibiotic[16] trifolitoxin[Ala23Cys] (i.e. Ala23
was mutated to Cys) was made by ligating two synthetic
peptide fragments (Figure 2). The ligation site was chosen to
be Ala22–Cys23. By using the optimized one-pot protocol
shown in Table 2, we found that the ligation proceeds cleanly
to provide the desired product in a high yield (97%). Note
that trifolitoxin[Ala23Cys] contains nine Lys residues and one
Cys residue. These residues do not interfere with the ligation
of peptide hydrazides.
We want to point out that the ligation of peptide
hydrazides provides complementary utility that is not easily
achievable with native chemical ligation, that is, the sequen-
tial ligation in the N-to-C direction.[9] In the previous studies
the sequential ligation of multiple peptide fragments was
usually conducted in the C-to-N direction by introducing a
protecting group at the N-terminal Cys group.[17] The N-to-C
sequential ligation, on the other hand, is mostly carried out by
using the kinetically controlled ligation (KCL) strategy.[9]
Using the ligation of peptide hydrazides we designed the
new N-to-C sequential ligation approach shown in Figure 3.
The key idea is to conduct the oxidation step for the first
peptide hydrazide before adding the second peptide that
possesses both an N-terminal Cys and a C-terminal hydrazide.
After the oxidation, the addition of MPAA not only converts
the peptide azide to peptide thioester, but also eliminates the
remaining NaNO2 in the ligation mixture. Then, addition of
the second peptide into the reaction mixture will only cause
ligation to its N-terminal Cys residue, while its C-terminal
hydrazide will survive the ligation.
In our synthesis the protein was divided into four fragments
(8–11) and each fragment was prepared using Fmoc-based
SPPS. To assemble the target, we first oxidized peptide 8 with
NaNO2 at pH 3.0 and À108C for 20 min. Then, MPAA was
added to the reaction mixture, followed by the addition of
peptide 9. After the ligation was completed, the product
CssII[Lys1–Tyr24] was purified from the reaction mixture.
The purified CssII[Lys1–Tyr24] was used for the next rounds
of ligation until the final product was obtained. The yields of
the three ligations at the Gly11–Cys12, Tyr24–Cys25, and
Tyr40–Cys41 sites are 92%, 85%, and 50%, respectively. The
synthetic 66residue polypeptide was characterized by
MALDI-TOF mass spectrometry (Figure 3).
Finally, another important feature of the ligation of
peptide hydrazides is that the key component of the method
(i.e. peptide hydrazide) can be obtained through biological
expression. This task is readily achieved by hydrazinolysis of
the protein thioester intermediate in the standard expressed
protein ligation (EPL) process.[4] To test the application of the
ligation of peptide hydrazides in protein semisynthesis, we
chose the microtubule-associated protein light chain 3 (LC3)
as the synthetic target.[19] Native LC3 does not contain Cys,
but to facilitate the ligation Phe119 was mutated to Cys. To
obtain the protein hydrazide, the DNA sequence for the first
118 amino residues of LC3 was cloned into pTXB1 bacterial
expression vector. The protein fused to C-terminal gyrA
intein-chitin binding domain (CBD) was expressed in E. coli
ER2566 cells and further loaded on chitin affinity column for
The new N-to-C sequential ligation strategy was tested for
the synthesis of the model protein CssII.[18] This protein has 66
amino acid residues including six Lys and eight Cys residues.
Angew. Chem. Int. Ed. 2011, 50, 7645 –7649
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 7647

Communications
1–2 h for purification. LC3[Pro2–Thr118]-NHNH2 was
terminal hydrazide is compatible with native chemical
ligation unless it is activated through oxidation. Therefore, a
protein hydrazide may be chemically modified first at its
N terminus or side chains before the final ligation at its
C terminus.
cleaved from the column during incubation with the normal
MESNa cleavage buffer (pH 8.0) containing 8% NH2NH2.
Then, the target protein was eluted with the cleavage buffer,
concentrated by membrane ultrafiltration, and purified by
RP-HPLC. The product LC3[Pro2–Thr118]-NHNH2 was
characterized by SDS-PAGE and ESI mass spectrometry
(Figure 4).
In summary, we have developed a ligation of peptide
hydrazides that is complementary to the native chemical
ligation. Auseful advantage of the new ligation method is that
peptide hydrazides can be easily prepared through both Boc-
and Fmoc-based SPPS. In addition, peptide hydrazides can be
readily obtained through recombinant expression. Both total
and semisynthesis of proteins can be achieved with the
ligation of peptide hydrazides. Moreover, the ligation of
peptide hydrazides enables sequential ligation in the N-to-C
direction. It should be pointed out that the the ligation of
peptide hydrazides is, in essence, a modified version of native
chemical ligation with an in situ generation of a peptide-
thioester from a peptide-hydrazide. We expect that the
combination of native chemical ligation and the ligation of
peptide hydrazides may enable more convergent synthesis of
proteins.
Received: February 9, 2011
Published online: June 6, 2011
Keywords: native chemical ligation · peptide hydrazides ·
.
protein chemical synthesis · protein design · proteins
[1] For a recent comprehensive review, see: S. B. H. Kent, Chem.
Soc. Rev. 2009, 38, 338.
[2] P. E. Dawson, T. W. Muir, I. Clark-Lewis, S. B. H. Kent, Science
1994, 266, 776.
[3] E. C. B. Johnson, S. B. H. Kent, J. Am. Chem. Soc. 2006, 128,
6640.
[4] M. Vila-Perellꢀ, T. W. Muir, Cell 2010, 143, 191.
[5] Recent examples: a) Y. Sohma, Q. X. Hua, J. Whittaker, M. A.
Weiss, S. B. H. Kent, Angew. Chem. 2010, 122, 5621; Angew.
Chem. Int. Ed. 2010, 49, 5489; b) Y. W. Wu, L. K. Oesterlin, K. T.
Tan, H. Waldmann, K. Alexandrov, R. S. Goody, Nat. Chem.
Biol. 2010, 6, 534; c) J. C. Scheuermann, A. G. D. Alonso, K.
Oktaba, N. Ly-Hartig, R. K. McGinty, S. Fraterman, T. W. Muir,
J. Muller, Nature 2010, 465, 243.
Figure 4. Semisynthesis of LC3[Phe119Cys]. a) SDS-PAGE analysis of
the expression and cleavage of the protein hydrazide LC3[Pro2–
Thr118]-NHNH2. b,c) Analytical HPLC traces (215 nm) for the reaction
mixture before (b) and after ligation (c). d) ESI mass spectrum for
LC3[Pro2–Thr118]-NHNH2. e) ESI mass spectrum for LC3[Phe119Cys].
More details can be found in the Supporting Information.
[6] For a recent review, see: F. Mende, M. Beisswenger, O. Seitz,
Angew. Chem. 2011, 123, 1266; Angew. Chem. Int. Ed. 2011, 50,
1232.
Further experiments showed that MESNa is not necessary
in the cleavage buffer. A neutral aqueous buffer (pH 7.0)
containing 8% NH2NH2 was sufficient for the hydrazinolysis
of the LC3[Pro2–Thr118]-intein-CBD fusion protein afford-
ing the hydrazide product in less than 3 h. LC3[Pro2–Thr118]-
NHNH2 was then ligated with the heptapeptide H-Cys-Gly-
Thr-Ala-Leu-Ala-Val-OH (i.e. LC3[Cys119–Val125]) using
the one-pot protocol shown in Table 2 to give the full-length
LC3[Phe119Cys] in 80% yield within 5 h. Thus, by changing
the thiol cleavage step to hydrazine cleavage, the intein-based
EPL method can be readily adapted to produce recombinant
protein hydrazides that are suitable for the ligation of peptide
hydrazides. In comparison to protein thioesters, we found that
protein hydrazides can be produced more quickly (cleavage
time 3 h) and in higher yields.[20] A protein hydrazide is also
chemically more stable than a protein thioester and therefore,
easier to handle. Finally, unlike a protein thioester, the C-
[7] a) L. E. Canne, A. R. Ferrꢁ-DꢂAmarꢁ, S. K. Burley, S. B. H.
Kent, J. Am. Chem. Soc. 1995, 117, 2998; b) C. F. Liu, C. Rao,
J. P. Tam, J. Am. Chem. Soc. 1996, 118, 307; c) B. L. Nilsson, L. L.
Kiessling, R. T. Raines, Org. Lett. 2000, 2, 1939; d) J. W. Bode,
R. M. Fox, K. D. Baucom, Angew. Chem. 2006, 118, 1270;
Angew. Chem. Int. Ed. 2006, 45, 1248; e) C. P. R. Hackenberger,
D. Schwarzer, Angew. Chem. 2008, 120, 10182; Angew. Chem.
Int. Ed. 2008, 47, 10030.
[8] a) S. S. Wang, J. Am. Chem. Soc. 1973, 95, 1328; b) J. K. Chang,
M. Shimizu, J. Org. Chem. 1976, 41, 3255.
[9] a) L. E. Canne, P. Botti, R. J. Simon, Y. Y. Chen, E. A. Dennis,
S. B. H. Kent, J. Am. Chem. Soc. 1999, 121, 8720; b) D. Bang,
B. L. Pentelute, S. B. H. Kent, Angew. Chem. 2006, 118, 4089;
Angew. Chem. Int. Ed. 2006, 45, 3985; c) J. S. Zheng, H. K. Cui,
G. M. Fang, W. X. Xi, L. Liu, ChemBioChem 2010, 11, 511;
d) Z. P. Tan, S. Y. Shang, S. J. Danishefsky, Angew. Chem. 2010,
122, 9690; Angew. Chem. Int. Ed. 2010, 49, 9500.
7648 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 7645 –7649

[10] a) T. Curtius, J. Prakt. Chem. 1904, 70, 57; b) N. Yanaihara, C.
Yanaihara, G. Dupuis, J. Beacham, R. Camble, K. Hofmann, J.
Am. Chem. Soc. 1969, 91, 2184; c) A. M. Felix, R. B. Merrifield,
J. Am. Chem. Soc. 1970, 92, 1385; d) H. Romovacek, S. R. Dowd,
K. Kawasaki, N. Nishi, K. Hofmann, J. Am. Chem. Soc. 1979,
101, 6081; e) For a related study on the Fmoc-based peptide
thioester synthesis from an aryl hydrazine support, see: J. A.
Camarero, B. J. Hackel, J. J. de Yoreo, A. R. Mitchell, J. Org.
Chem. 2004, 69, 4145.
[11] To synthesize peptide hydrazides, we treated Wang resin with 4-
nitrophenyl carbonochloridate. The resin was then reacted with
hydrazine to load the hydrazinecarboxylate moiety. Next, by
using the Fmoc protocols we prepared peptide hydrazides in
good quality.
[16] B. Y. Breil, P. W. Ludden, E. W. Triplett, J. Bacteriol. 1993, 175,
3693.
[17] a) A. Brik, E. Keinan, P. E. Dawson, J. Org. Chem. 2000, 65,
3829; b) C. F. W. Becker, C. L. Hunter, R. Seidel, S. B. H. Kent,
R. S. Goddy, M. Engelhard, Proc. Natl. Acad. Sci. USA 2003,
100, 5075; c) D. Bang, N. Chopra, S. B. H. Kent, J. Am. Chem.
Soc. 2004, 126, 1377; d) E. C. B. Johnson, T. Durek, S. B. H.
Kent, Angew. Chem. 2006, 118, 3361; Angew. Chem. Int. Ed.
2006, 45, 3283; e) D. Bang, S. B. H. Kent, Proc. Natl. Acad. Sci.
USA 2005, 102, 5014; f) E. C. B. Johnson, E. Malito, Y. Shen, D.
Richi, W. J. Tang, S. B. H. Kent, J. Am. Chem. Soc. 2007, 129,
11480; g) B. L. Pentelute, Z. P. Gates, J. L. Dashnau, J. M.
Vanderkooi, S. B. H. Kent, J. Am. Chem. Soc. 2008, 130, 9702.
[18] M. F. Martin, L. G. G. Perez, M. E. Ayeb, C. Kopeyan, G. Bechis,
E. Jover, H. Rochat, J. Biol. Chem. 1987, 262, 4452.
[19] Y. Kabeya, N. Mizushima, T. Ueno, A. Yamamoto, T. Kirisako,
T. Noda, E. Kominami, Y. Ohsumi, T. Yoshimori, EMBO J. 2000,
19, 5720.
[20] Through the MESNa cleavage we obtained 3–4 mg of LC3-
thioester from 1 L of LB medium, whereas through the
hydrazine cleavage we obtained 6–7 mg of LC3-hydrazide
from 1 L of LB medium.
[12] L. Campayo, B. Jimenez, T. Manzano, P. Navarro, Synthesis 1985,
197.
[13] L. Grossi, P. C. Montevecchi, J. Org. Chem. 2002, 67, 8625.
[14] M. Kashiba-Iwatsuki, M. Yamagushi, M. Inoue, FEBS Lett. 1996,
389, 149.
[15] D. P. Gamblin, E. M. Scanlan, B. G. Davis, Chem. Rev. 2009, 109,
131.
Angew. Chem. Int. Ed. 2011, 50, 7645 –7649
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 7649