Mechanism-inspired engineering of phenylalanine aminomutase for enhanced β-regioselective asymmetric amination of cinnamates

Article abstract of DOI:10.1002/anie.201106372

Turn to switch: A mutant of phenylalanine aminomutase was engineered that can catalyze the regioselective amination of cinnamate derivatives (see scheme, red) to, for example, β-amino acids. This regioselectivity, along with the X-ray crystal structures, suggests two distinct carboxylate binding modes differentiated by C_βi￡C_ipso bond rotation, which determines if β- (see scheme) or α-addition takes place. Copyright

Full text of DOI:10.1002/anie.201106372

.
Angewandte
Communications
DOI: 10.1002/anie.201106372
Enzyme Catalysis
Mechanism-Inspired Engineering of Phenylalanine Aminomutase for
Enhanced b-Regioselective Asymmetric Amination of Cinnamates**
´
Bian Wu, Wiktor Szymanski, Gjalt G. Wybenga, Matthew M. Heberling, Sebastian Bartsch,
Stefaan de Wildeman, Gerrit J. Poelarends, Ben L. Feringa, Bauke W. Dijkstra, and
Dick B. Janssen*
In the last decade, enormous progress has been made in the
synthesis of optically pure b-amino acids, which occur as
components of numerous bioactive compounds.[1] Although
biocatalytic routes have been explored, they tend to use
kinetic resolutions, implying a maximum yield of 50%.[2] Only
a few enzymes, including aminotransferases[3] and amino-
mutases,[4] have been used in asymmetric synthesis of this
class of compounds. The asymmetric addition of ammonia to
cinnamic acid and its derivatives catalyzed by phenylalanine
aminomutase (PAM) is an attractive route to enantiopure
aromatic b-amino acids, because the substrates are readily
available and PAM exhibits excellent enantioselectivity.[5]
However, the final product mixture consists of both a- and
b-amino acids owing to low regioselectivity.[6] Although
structures of aminomutases in complex with substrate ana-
logues,[9] suggests that in the mutase reaction MIO reacts with
the amine group of the substrate. Kinetic studies of PAM with
different cinnamates indicated that the regioselectivity is
dominated by electronic effects,[5c,d] with either the aromatic
ring or the carboxylate group acting as an electron sink to
facilitate a- or b-addition, respectively. The delocalization of
negative charge to these groups must be stabilized by the
protein environment (Scheme 1). Therefore we hypothesized
that mutations that diminish the ability of the phenyl ring
binding pocket to accommodate negative charge or mutations
that enhance the electron deficiency of the carboxylate
binding pocket will likely promote b-addition.
selective separation procedures have been developed,[6]
a
much better option would be to employ a highly b-regiose-
lective enzyme. Unfortunately, no such enzyme has been
discovered to date. Herein we describe the successful alter-
ation of Taxus chinensis PAM into a selective b-lyase through
structure- and mechanism-based protein engineering.
The activity of PAM relies on the internal cofactor 4-
methylideneimidazole-5-one (MIO).[7] The role of MIO has
been debated,[8] but most recent evidence, including X-ray
[*] Dr. B. Wu, M. M. Heberling, Dr. S. Bartsch, Prof. D. B. Janssen
Department of Biochemistry, Groningen Biomolecular Sciences and
Biotechnology Institute, University of Groningen
Nijenborgh 4, 9747 AG, Groningen (The Netherlands)
E-mail: d.b.janssen@rug.nl
Scheme 1. Proposed electronic effects that influence the regioselectiv-
ity of PAM-catalyzed amination reactions.
G. G. Wybenga, Prof. B. W. Dijkstra
Laboratory of Biophysical Chemistry, Groningen Biomolecular
Sciences and Biotechnology Institute
University of Groningen (The Netherlands)
To identify targets for mutations, we inspected structures
of cinnamic acid bound PAM (PDB 3NZ4)[10] and the
apoenzyme (PDB 2YII, see the Supporting Information).
Overall, PAM is structurally similar to the Streptomyces
globisporus tyrosine aminomutase (sgTAM).[9] The two loops
of PAM covering the active site, which are also present in
other MIO-dependent enzymes,[9,11] are in a closed confor-
mation. The active-site residues can be divided in two distinct
sets, which form the aromatic ring binding pocket (F86, L104,
C107, L108, L179, Q459; Figure 1A) and the carboxylate
binding pocket (N231, Q319, Y322, R325, N355, F371;
Figure 1B), respectively.
The binding pockets were separately mutated using a
pairwise saturation strategy.[12] The carboxylate-pocket resi-
dues were divided into three sets, each comprising two
structurally neighboring residues. Three saturation libraries
were constructed (Q319X/R325X, N231X/N355X, Y322X/
´
Dr. W. Szymanski, Prof. B. L. Feringa
Center for Systems Chemistry, Stratingh Institute for Chemistry
University of Groningen (The Netherlands)
Dr. S. de Wildeman
DSM Pharmaceutical Products
Geleen (The Netherlands)
Dr. G. J. Poelarends
Department of Pharmaceutical Biology
Groningen Research Institute of Pharmacy
University of Groningen (The Netherlands)
[**] The work was supported through B-Basic, a public–private NWO–
ACTS program. The authors thank Dr. O. May and Dr. B. Kaptein
from DSM and Hans Raj from the Department of Pharmaceutical
Biology, University of Groningen, for helpful discussions.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201106372.
482
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 482 –486

Angewandte
Chemie
Figure 1. The binding pockets of PAM. A) Superposition of the aromatic ring binding pocket of PAM with that of sgTAM; B) Superposition of the
carboxylate binding pocket of PAM with that of sgTAM. Blue: sgTAM in complex with substrate analogue a-di-F-b-Tyr (gray); green: apo structure
of PAM; red: PAM in complex with trans-cinnamic acid (brown). C=Cys =cysteine, F=Phe=phenylalanine, H=His=histidine, L=Leu=leucine,
N=Asn=asparagine, Q=Gln=glutamine, R=Arg=arginine, Y=Tyr =tyrosine.
F371X) and expressed in E. coli as N-terminal His-
Table 1: Kinetic parameters and regio- and enantioselectivity of PAM mutants in
tag fusion proteins. The mutants were purified in a
96-well format using cobalt affinity resin plates (see
the Supporting Information) and prescreened for
deamination activity towards b-Phe by monitoring
the formation of cinnamic acid at 290 nm. Whereas
the mutants in libraries N231X/N355X and Y322X/
F371X showed no detectable activity, two active
mutants (Q319M and R325K; M = Met = methio-
nine, K = Lys = lysine) were identified in library
Q319X/R325X. The R325K mutant showed a 2.5-
fold increase in deamination activity towards b-Phe
and its b-regioselectivity (percentage of b-Phe in the
amination reactions.[a]
Enzyme
kcat
Km
k
cat/Km
b
eeb (R) eea (S)
[sꢀ1ꢀ103] [mm]
[sꢀ1 mꢀ1
]
[%]
wt PAM
R325K
Q319M
R325K+Q459E
12ꢁ2
8ꢁ1
2.6ꢁ0.6
5.2ꢁ0.4
6.3ꢁ1.8 49ꢁ1 >99% >99%
1.5ꢁ0.3 85ꢁ2 >99% >99%
15ꢁ1
>0.15[b]
1.2ꢁ0.2 12ꢁ1
n.d. n.d.
88ꢁ2 >99% >99%
ca. 95 >99% n.d.
[a] Reactions were performed in saturated ammonium carbonate buffer at pH 9, and
the cinnamate concentration was varied. [b] No saturation was obtained. Number
represents the catalytic rate at the conditions used.
product mixture) in the amination reaction of cinnamic acid
had shifted from 49% (wild-type (wt) PAM) to 85%. The
Q319M mutant exhibited fivefold enhanced b-Phe deamina-
tion activity and increased b-regioselectivity (88%), but an
enzyme in which both mutations (Q319M and R325K) are
combined was inactive.
We then mutated the aromatic ring binding pocket for
improved b-regioselectivity. Mutagenesis work was focused
on residues C107 and Q459, to avoid changing the conserved
nonpolar residues. From library C107X/Q459X/R325K
mutant R325K/Q459E was discovered. This mutant retained
30% of the deamination activity towards b-Phe, and its
deamination activity towards a-Phe was almost completely
abolished (Vrel < 1% of wt PAM). In an amination reaction
with cinnamic acid, the mutant appeared to be highly selective
towards the b-position (up to 95% b-addition, Table 1),
although the activity was low.
The enantioselectivity of the enzyme, both at the Ca and
Cb carbon atom of the substrate, was retained in all mutants.
The products of the amination reaction were enantiopure (S)-
a-Phe and (R)-b-Phe. A time-course experiment with the
efficient Q319M mutant showed that nearly 90% b-selectivity
was maintained until at least 70% conversion (see the
Supporting Information). Furthermore, we set up a labora-
tory-scale preparative experiment by incubating trans-cin-
namic acid (20 mg) with Q319M mutant enzyme (5 mg) in
ammonia solution (2m, pH 9, 10% glycerol). After four days
the conversion reached 60%, and after purification with a
cation exchange resin (Dowex 50W X8) column, a 1:7 mixture
of (S)-a-Phe (ee > 99%) and (R)-b-Phe (ee > 99%) was
isolated with 50% yield. This outcome represents an almost
twofold increase in the yield of (R)-b-Phe compared to the
process in which wt PAM was used as a catalyst.[5d]
To investigate the applicability of the regioselective
mutant for preparation of other aromatic b-amino acids, we
used the Q319M mutant to convert several cinnamates. The
results, compared to those obtained with wt PAM (Table 2),
showed that in all cases the mutant is superior in terms of b-
selectivity, thus allowing the preparation of various b-amino
acids. Notably, ammonia addition to a cinnamic acid deriv-
ative with an electron-donating p-methyl substituent gave
exclusively the b-product. Furthermore, the Q319M mutant
has an improved affinity (lower Km value) for all compounds
described in entries 1–5 (Table 2); the improved affinity
results in significantly higher catalytic efficiency (kcat/Km) than
for wt PAM.[5c]
As expected, the introduction of an electron-rich, charged
residue in the aromatic binding pocket (Q459E) enhanced b-
selectivity, because the mutation reduces the electron-with-
drawing effect of the aromatic ring. However, it was
Angew. Chem. Int. Ed. 2012, 51, 482 –486
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
483

.
Angewandte
Communications
amine group and pro-3S hydrogen atom (HA, Scheme 2A) are
oriented in an anti-periplanar geometry that is suitable for the
elimination step. A nucleophilic attack of the amine group of
the substrate on the MIO group preceeds abstraction of the
proton (HA) by the general base Y80 (Scheme 2 B!C),
possibly assisted by another basic group, and elimination of
MIO-NH2. This sequence leads to formation of trans-cin-
namic acid, with the carboxylate group still bound to residue
R325 (Scheme 2 C). The next step is the readdition of MIO-
NH2 to the b-position and of proton HA to the a-position. This
step requires exposure of the Re face of Cb to the MIO-NH2
and the carboxylate group to function as a good electron sink.
This 1,4-addition is hindered when the substrate makes a salt
bridge with residue R325, owing to the high energy level of
the LUMO (lowest unoccupied molecular orbital) of this
ionized electrophile.[15] However, trans-cinnamic acid can
Table 2: Kinetic parameters and regioselectivity of amination reactions
catalyzed by the Q319M mutant PAM.
Entry
R
kcat
Km [mm]
1.2ꢁ0.2
kcat/Km
b
b
[sꢀ1ꢀ103]
[sꢀ1 mꢀ1
]
[%]
[%]wt
PAM[5c]
1
2
3
4
5
H
F
Cl
15ꢁ1
12ꢁ1
48ꢁ1
88ꢁ2 49
10ꢁ0.03 0.21ꢁ0.003
91ꢁ2 65
86ꢁ2 59
8.7ꢁ0.4 0.022ꢁ0.003 395ꢁ72
Me 12ꢁ0.3
0.072ꢁ0.004 167ꢁ14 >99
0.3ꢁ0.06 326ꢁ92
96
2
NO2 98ꢁ8
8ꢁ1
unexpected that the Q319M or the R325K mutations in the
carboxylate binding pocket also enhanced b-regioselectivity,
because electrostatic interactions that stabilize negative
charge on the carboxylate group during conjugate addition
would be weakened by these mutations. To unveil the
mechanistic basis of the enhanced b-regioselectivity induced
by the mutations in the carboxylate binding pocket, we
analyzed the structural and biochemical data on MIO-
dependent aminomutases.
It has been demonstrated that PAM catalyzes the
conversion of (S)-a-Phe into (R)-b-Phe by exchanging the
amine group (Ca!Cb) and the pro-3S hydrogen atom (Cb!
Ca) with retention of configuration at the reaction termini.[13]
A possible explanation is that the phenyl ring and carboxylate
group of (S)-a-Phe are arranged in a syn-periplanar orienta-
tion with the amine group and the leaving pro-3S hydrogen
atom positioned on the same side of the substrate, leading to
the formation of cis-cinnamic acid. However, the evidence
listed below suggests that trans-cinnamic acid is the real
intermediate: 1) trans-cinnamic acid is released from the
active center during the mutase reaction;[14] 2) only the trans
form of cinnamic acid is a substrate for PAM;[5c] 3) the X-ray
structure shows bound trans-cinnamic acid in the active site of
PAM.[10] Accordingly, the stereochemistry requires a reorien-
tation after deamination of (S)-a-Phe to cinnamic acid such
that the Re face of the Cb and the Si face of the Ca carbon
atom are positioned for amine addition and protonation,
respectively.[10,13] In the structure of PAM with trans-cinnamic
acid bound, the carboxylate group makes a strong salt bridge
with residue R325.[10] Feng et al. postulated that this salt
ꢀ
undergo an internal rotation around the Cb Cipso bond to
position the carboxylate group into an alternative binding
mode where each oxygen atom is involved in two hydrogen
bonds, that is, with residues N231, R325, and N355 (Scheme 2
D, binding mode 2). The sharing of hydrogen atoms between
these H-bond donors and the carboxylate group of the
substrate lowers the LUMO of the latter, thereby rendering
the carboxylate a better electron sink and promoting con-
jugate addition at the b-position.[15b] Similar LUMO-lowering
activation through hydrogen-bond interactions has been
observed in other conjugate additions[16] and in Diels–
Alder[17] enzymatic reactions. The re-added proton apparently
is not always the pro-3S hydrogen atom from the substrate,
because some hydrogen exchange was observed during the
isomerization reaction.[13] The mutations (Q319M and
R325K) in the carboxylate binding pocket would disturb
binding mode 1, and consequently make binding mode 2
more favorable and promote b-addition.
Additional evidence for this proposal is provided by a
sgTAM structure (PDB 2RJR),[9b] the active site of which is
similar to that of PAM. The alternative binding mode 2 of the
carboxylate group proposed for PAM has already been
observed in the sgTAM structure with a substrate analogue
bound in the active center. The carboxylate oxygen atoms are
involved in a hydrogen-bonding network with residues N205,
R311, and N341 (corresponding to N231, R325, and N355 in
PAM).[9b] The aromatic hydroxy group forms hydrogen bonds
with residues H93 and Y415, thereby triggering a different
binding mode for the ring as compared to trans-cinnamate in
PAM and placing the carboxylate group in an orientation that
is suitable for b-readdition without rotation, leading to
formation of (S)-b-Tyr instead of (R)-b-Tyr.
In summary, we have shown that the regioselectivity of a
MIO-dependent enzyme can be tailored towards b-regiose-
lectivity. A redesigned enzyme allows the synthesis of almost
pure (R)-b-Phe and its derivatives by one-step asymmetric
amination of cinnamic acid. Further engineering of PAM,
aimed at improving activity while retaining regioselectivity
and enantioselectivity, is underway. In combination with
stereochemical,[13] synthetic,[5] and structural investiga-
tions,[9,10] the results provide a framework for explaining and
engineering the regioselective and stereochemical properties
of MIO-dependent aminomutases.
ꢀ
ꢀ
bridge is maintained, while both the C1 Ca and the Cb Cipso
bonds rotate to fit the stereochemical course of the PAM
reaction.[10] However, this rotation is energetically burden-
some, because it requires the loss of olefinic bond conjugation
with both the aromatic ring and the carboxylate group.
Furthermore, our mutagenesis work shows that interfering
with the proposed stable salt bridge enhances b-addition.
We propose a mechanism for the mutase reaction that
explains both the stereochemical features and the results of
our mutagenesis experiments. When (S)-a-Phe enters the
active site of PAM, the carboxylate group forms a bidentate
salt bridge with residue R325 (Scheme 2A, binding mode 1),
as present in the cinnamic acid bound PAM structure. The
carboxylate group further interacts with residue Q319. The
484
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 482 –486

Angewandte
Chemie
Scheme 2. Representation of the stereochemical course and reaction mechanism of PAM. The conformation of trans-cinnamic acid and the
orientation of its carboxylate in (C) are based on the structure of PAM with bound trans-cinnamic acid (PDB 3NZ4). Enz-B:=enzyme base.
Received: September 8, 2011
Revised: October 12, 2011
Published online: November 23, 2011
´
Szymanski, P. Wietzes, S. de Wildeman, G. J. Poelarends, B. L.
Feringa, D. B. Janssen, ChemBioChem 2009, 10, 338 – 344; d) W.
´
Szymanski, B. Wu, B. Weiner, S. de Wildeman, B. L. Feringa,
D. B. Janssen, J. Org. Chem. 2009, 74, 9152 – 9157; e) B. Wu, W.
´
Szymanski, S. de Wildeman, G. J. Poelarends, B. L. Feringa, D. B.
Keywords: aminomutase · b-amino acids · directed evolution ·
enzyme catalysis · regioselectivity
.
Janssen, Adv. Synth. Catal. 2010, 352, 1409 – 1412; f) B. Wu, W.
´
Szymanski, H. J. Wijma, C. G. Crismaru, S. de Wildeman S, G. J.
Poelarends, B. L. Feringa, D. B. Janssen, Chem. Commun. 2010,
46, 8157 – 8159.
´
[1] a) B. Weiner, W. Szymanski, D. B. Janssen, A. J. Minnaard, B. L.
´
[6] a) B. J. V. Verkuijl, W. Szymanski, B. Wu, A. J. Minnaard, D. B.
Feringa, Chem. Soc. Rev. 2010, 39, 1656 – 1691; b) D. Seebach, T.
Janssen, J. G. de Vries, B. L. Feringa, Chem. Commun. 2010, 46,
901 – 903; b) S. Wijeratne, N. A. Byrne, K. D. Walker, J. Sep. Sci.
2010, 33, 1279 – 1282.
ˇ
Kimmerlin, R. Sebesta, M. A. Campo, A. K. Beck, Tetrahedron
2004, 60, 7455 – 7506; c) D. Seebach, J. Gardiner, Acc. Chem.
Res. 2008, 41, 1366 – 1375.
[7] a) T. F. Schwede, J. Rꢁtey, G. E. Schulz, Biochemistry 1999, 38,
5355 – 5361; b) D. Rçther, D. Merkel, J. Rꢁtey, Angew. Chem.
2000, 112, 2592 – 2594; Angew. Chem. Int. Ed. 2000, 39, 2462 –
2464; c) S. D. Christenson, W. Liu, M. D. Toney, B. Shen, J. Am.
Chem. Soc. 2003, 125, 6062 – 6063.
[8] a) L. Poppe, J. Rꢁtey, Angew. Chem. 2005, 117, 3734 – 3754;
Angew. Chem. Int. Ed. 2005, 44, 3668 – 3688; b) H. A. Cooke,
C. V. Christianson, S. D. Bruner, Curr. Opin. Chem. Biol. 2009,
13, 460 – 468; c) S. Bartsch, U. T. Bornscheuer, Angew. Chem.
2009, 121, 3412 – 3415; Angew. Chem. Int. Ed. 2009, 48, 3362 –
3365; d) S. Bartsch, U. T. Bornscheuer, Protein Eng. Des. Sel.
2010, 23, 929 – 933.
[2] a) A. Liljeblad, L. T. Kanerva, Tetrahedron 2006, 62, 5831 – 5854;
b) V. Gotor-Fernꢀndez, V. Gotor, Curr. Opin. Drug Discovery
Dev. 2009, 12, 784 – 797.
[3] J. Kim, D. Kyung, D. H. Yun, B. K. Cho, J. H. Seo, M. Cha, B. G.
Kim, Appl. Environ. Microbiol. 2007, 73, 1772 – 1782.
[4] a) N. J. Turner, Curr. Opin. Chem. Biol. 2011, 15, 234 – 240;
b) B. M. Nestl, B. A. Nebel, B. Hauer, Curr. Opin. Chem. Biol.
´
2011, 15, 187 – 193; c) B. Wu, W. Szymanski, M. M. Heberling,
B. L. Feringa, D. B. Janssen, Trends Biotechnol. 2011, 29, 352 –
362; d) C. L. Steele, Y. Chen, B. A. Dougherty, W. Li, S.
Hofstead, K. S. Lam, Z. Xing, S. J. Chiang, Arch. Biochem.
Biophys. 2005, 438, 1 – 10.
[9] a) C. V. Christianson, T. J. Montavon, S. G. Van Lanen, B. Shen,
S. D. Bruner, Biochemistry 2007, 46, 7205 – 7214; b) C. V. Chris-
tianson, T. J. Montavon, G. M. Festin, H. A. Cooke, B. Shen,
[5] a) K. L. Klettke, S. Sanyal, W. Mutatu, K. D. Walker, J. Am.
Chem. Soc. 2007, 129, 6988 – 6989; b) B. M. Cox, J. B. Bilsborrow,
K. D. Walker, J. Org. Chem. 2009, 74, 6953 – 6959; c) B. Wu, W.
Angew. Chem. Int. Ed. 2012, 51, 482 –486
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
485

.
Angewandte
Communications
S. D. Bruner, J. Am. Chem. Soc. 2007, 129, 15744 – 15745;
c) H. A. Cooke, S. D. Bruner, Biopolymers 2010, 93, 802 – 810.
[10] L. Feng, U. Wanninayake, S. Strom, J. Geiger, K. D. Walker,
Biochemistry 2011, 50, 2919 – 2930.
[13] W. Mutatu, K. L. Klettke, C. Foster, K. D. Walker, Biochemistry
2007, 46, 9785 – 9794.
[14] a) See Ref. [4d]; b) K. D. Walker, K. Klettke, T. Akiyama, R.
Croteau, J. Biol. Chem. 2004, 279, 53947 – 53954.
[11] a) H. Ritter, G. E. Schulz, Plant Cell 2004, 16, 3426 – 3436;
b) J. C. Calabrese, D. B. Jordan, A. Boodhoo, S. Sariaslani, T.
Vannelli, Biochemistry 2004, 43, 11403 – 11416; c) M. C. Moffitt,
G. V. Louie, M. E. Bowman, J. Pence, J. P. Noel, B. S. Moore,
Biochemistry 2007, 46, 1004 – 1012; d) L. Wang, A. Gamez, H.
Archer, E. E. Abola, C. N. Sarkissian, P. Fitzpatrick, D. Wendt,
Y. Zhang, M. Vellard, J. Bliesath, S. M. Bell, J. F. Lemontt, C. R.
Scriver, R. C. Stevens, J. Mol. Biol. 2008, 380, 623 – 635.
[15] a) R. G. Pearson, J. Chem. Sci. 2005, 117, 369 – 377; b) D. C.
Chatfield, A. Augsten, C. DꢂCunha, E. Lewandoska, S. F. Wnuk,
Eur. J. Org. Chem. 2004, 313 – 322.
[16] a) X. Q. Dong, H. L. Teng, C. J. Wang, Org. Lett. 2009, 11, 1265 –
1268; b) A. Lu, T. Liu, R. Wu, Y. Wang, G. Wu, Z. Zhou, J. Fang,
C. Tang, J. Org. Chem. 2011, 76, 3872 – 3879.
[17] J. B. Siegel, A. Zanghellini, H. M. Lovick, G. Kiss, A. R.
Lambert, J. L. St Clair, J. L. Gallaher, D. Hilvert, M. H. Gelb,
B. L. Stoddard, K. N. Houk, F. E. Michael, D. Baker, Science
2010, 329, 309 – 313.
[12] M. T. Reetz, J. D. Carballeira, A. Vogel, Angew. Chem. 2006,
118, 7909 – 7915; Angew. Chem. Int. Ed. 2006, 45, 7745 – 7751.
486
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 482 –486