Entropy is key to the formation of pentacyclic terpenoids by enzyme-catalyzed polycyclization

Article abstract of DOI:10.1002/anie.201402087

Polycyclizations constitute a cornerstone of chemistry and biology. Multicyclic scaffolds are generated by terpene cyclase enzymes in nature through a carbocationic polycyclization cascade of a prefolded polyisoprene backbone, for which electrostatic stabilization of transient carbocationic species is believed to drive catalysis. Computational studies and site-directed mutagenesis were used to assess the contribution of entropy to the polycyclization cascade catalyzed by the triterpene cyclase from A. acidocaldarius. Our results show that entropy contributes significantly to the rate enhancement through the release of water molecules through specific channels. A single rational point mutation that results in the disruption of one of these water channels decreased the entropic contribution to catalysis by 60 kcalmol^-1. This work demonstrates that entropy is the key to enzyme-catalyzed polycyclizations, which are highly relevant in biology since 90% of all natural products contain a cyclic subunit.

Full text of DOI:10.1002/anie.201402087

Angewandte
Chemie
DOI: 10.1002/anie.201402087
Enzyme Catalysis
Entropy is Key to the Formation of Pentacyclic Terpenoids by Enzyme-
Catalyzed Polycyclization**
Per-Olof Syrꢀn,* Stephan C. Hammer, Birgit Claasen, and Bernhard Hauer
Abstract: Polycyclizations constitute a cornerstone of chemis-
try and biology. Multicyclic scaffolds are generated by terpene
cyclase enzymes in nature through a carbocationic polycycli-
zation cascade of a prefolded polyisoprene backbone, for
which electrostatic stabilization of transient carbocationic
species is believed to drive catalysis. Computational studies
and site-directed mutagenesis were used to assess the contri-
bution of entropy to the polycyclization cascade catalyzed by
the triterpene cyclase from A. acidocaldarius. Our results show
that entropy contributes significantly to the rate enhancement
through the release of water molecules through specific
channels. A single rational point mutation that results in the
disruption of one of these water channels decreased the
entropic contribution to catalysis by 60 kcalmolÀ1. This work
demonstrates that entropy is the key to enzyme-catalyzed
polycyclizations, which are highly relevant in biology since
90% of all natural products contain a cyclic subunit.
therapeutics for cancer[4] and malaria[5] and they display
a high synthetic potential[6] for the generation of novel
terpenoid-based compounds. Hence it is of uttermost impor-
tance to enhance our understanding of the reaction mecha-
nism displayed by these enzymes, which catalyze one of the
most important and complex transformations known in
biochemistry, one that has fascinated researchers for dec-
ades.[7,8] It is known that these enzymes bind and stabilize the
flexible substrate in the precise orientation required for
catalysis through a unique cyclization sequence, trigger
carbocation formation, and stabilize carbocations against
premature quenching of water molecules present in the active
site.[8–11] Pre-arranging the substrate for ring closure will
inevitably be associated with a high entropic penalty because
of the lost internal rotational degrees of freedom.[12] It is well
known that this entropic cost plays an important role in
intramolecular cyclization reactions, such as macrolactoniza-
tions,[13] that have highly ordered transition-state structures.
We were puzzled by how enzyme-catalyzed concerted poly-
cyclizations can overcome the high entropic cost of prefolding
the substrate for polycyclization. This issue is addressed in the
present investigation, which is based on experimental and
theoretical work with a thermophilic triterpene cyclase
(Scheme 1). These enzymes function by the class II mecha-
nism in analogy to Brønsted acid catalysis, which is the
T
erpenoids have central functions for all living organisms
and are a diverse class of biologically active compounds that
includes potent anticancer, antiviral, antimicrobial, and anti-
fungal agents. The structural diversity of terpenoids originates
from an enzyme-catalyzed carbocationic polycyclization cas-
cade of linear polyisoprenes that are all derived from
isopentenyl diphosphate and its regioisomer dimethylallyl
diphosphate.[1] The cyclization cascade is initiated either by
the metal-dependent ionization of a labile allylic diphosphate
group (class I mechanism) or by the protonation of the
terminal isoprene double bond (class II mechanism). Penta-
cyclic chemistry based on the latter was already established
during the Archean eon[2] by triterpene cyclases. Terpene
cyclases have provided access to new potential biofuels[3] and
[*] Dr. P.-O. Syrꢀn,[+] S. C. Hammer, Prof. Dr. B. Hauer
Institute of Technical Biochemistry, University of Stuttgart
Allmandring 31, 70569 Stuttgart (Germany)
E-mail: per-olof.syren@biotech.kth.se
Dr. B. Claasen
Institute for Organic Chemistry
Pfaffenwaldring 55, 70569 Stuttgart (Germany)
[+] Present address: Division of Proteomics
Scheme 1. The cyclization of polyisoprene backbones of different
lengths by Alicyclobacillus acidocaldarius triterpene cyclase. A) The
polycyclization of squalene (C30) yields the pentacyclic hopenyl cation
(possibly via a tricyclic core and subsequent ring expansion[9]). This
reacts to give hopene through pathway 1 (blue) or hopanol through
termination by water addition (pathway 2, red). B) The cyclization of
homofarnesol (C16) yields ambroxan with a tricyclic C6–C6–C5 skel-
eton. C) The cyclization of geranyl octyl ether yields a monocyclic
water addition product. Enz–H refers to the catalytic acid in the
triterpene cyclase active site that initiates the polycyclization cascade
by protonating the terminal isoprene group.
KTH Royal Institute of Technology, AlbaNova University Center
Roslagstullsbacken 21, 10691 Stockholm (Sweden)
[**] The Alexander von Humboldt foundation, the Royal Swedish
Academy of Sciences, Fonds der Chemischen Industrie, and the
European Union’s Seventh Framework Programme FP7/2007-2013
under grant agreement no 289646 are gratefully acknowledged for
financial support.
Supporting information for this article (including experimental
details) is available on the WWW under http://dx.doi.org/10.1002/
anie.201402087.
Angew. Chem. Int. Ed. 2014, 53, 4845 –4849
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4845

.
Angewandte
Communications
foundation of fundamental organic transformations[14] includ-
ing biomimetic cyclization reactions.[15] Hence the mechanis-
tic study herein could have important implications for
synthetic chemistry.
The entropic penalty associated with forming a productive
precyclic substrate conformation is well recognized and
amounts up to 43 calmolÀ1 KÀ1 (14 kcalmolÀ1 at 328 K) even
for monocycles.[13,16] This energetic cost would slow down the
rate 109-fold compared to that of a reaction with an activation
entropy of zero. Nonetheless, the impact of entropy on
enzyme-catalyzed concerted polycyclization cascades remains
unknown. The importance of addressing entropy is corrobo-
rated by examining the enzyme-catalyzed generation of
a tricyclic core (Figure S1a in the Supporting Information).
An attractive way to obtain an enzymatic polycyclization
catalyst that display biologically relevant rates would be to
make the entropy of catalysis more favorable in addition to
stabilizing the carbocations. This would compensate for the
entropic penalty associated with substrate prefolding that
would otherwise significantly reduce the reaction rate. Such
an entropic catalysis could be achieved by expelling ordered
water molecules from the active site (Figure S1a, top) in
a manner analogous to the binding of some ligands to
receptors.[17,18] In support of this idea, 1H STD-NMR revealed
that the binding of a small substrate-like probe to the active
site of the triterpene cyclase from Alicyclobacillus acido-
caldarius was driven by entropy (Figure S1b and S1c), as
would be expected from the expulsion of water upon
association. Moreover, seminal crystallographic work by
Reinert et al.[19] and Wendt et al.[20] allowed us to perform
an analysis of the crystal structures of this thermophilic
triterpene cyclase complexed with ligands of two different
sizes (PDB 1UMP[19] and 2SQC[20]). The results showed that
the active site contained nine additional water molecules
when complexed with the less bulky ligand (Figure S2 in the
Supporting Information). We were faced with the problem of
how such fixed water molecules in the active site could be
expelled since the substrate and corresponding pentacyclic
products enter and exit through the hydrophobic cell
membrane.[21] Furthermore, no conformational change
occurs upon substrate binding (see the Supporting Informa-
tion). The work with the extremophilic triterpene cyclase
(PDB 1UMP[19]) started with a structural analysis for water
channels by using the software Caver.[22] Interestingly,
a number of different tunnels were found and we focused
on the three highest-ranked channels that connect the binding
site to the cytosol (Figure 1a). Water molecules were found to
reside within the identified channels and 20 ns MD-simula-
tions in a waterbox were sufficient to observe movement of
water molecules in the three tunnels in support of entropic
catalysis (Figure S3).
Figure 1. The existence of specific tunnels implies a central role for
water in promoting the terpene cyclase catalyzed generation of multi-
cyclic products. A) Water channel analysis using Caver and snapshots
from MD simulations of the triterpene cyclase (PDB 1UMP).[19] The
three tunnels with highest ranking according to the software are
schematically drawn with the water molecules confined within the
channels shown. Key residues that were changed to block the channels
are shown as stick models and the corresponding change in bulk upon
introducing mutations are shown as gray space-filling representations.
The natural C30 substrate squalene is shown in magenta. The catalytic
Asp that initiates the cyclization cascade is located at the back of the
figure and is not shown. B) Suggested reaction mechanism of entropi-
cally favored enzyme-catalyzed polycyclization. The release of ordered
water molecules (red) makes the prefolding of the polyisoprenoid
substrate (purple) favorable.
they blocked the water channels (Table S4). The experimental
study was initiated by analyzing the temperature dependence
of the second-order rate constant (apparent kcat/KM) for the
enzyme-catalyzed generation of pentacycles. It was not
possible to achieve saturation of the enzyme under our
experimental conditions (0.2% TritonX-100 micelles, 60 mm
citrate, pH 6). This is perhaps not surprising since the
triterpene cyclase is a monotopic membrane “receptor
protein” and the phospholipid bilayer is different from the
aqueous cellular environment that many enzymes operate in.
The thermodynamic consequences of blocking the channels
by the introduced mutations were studied experimentally
(Figure 2) and by using transition-state theory (see Equa-
tion (1) in the Supporting Information).
To find experimental support for the entropically favored
mechanism depicted in Figure 1b, kinetic experiments were
performed using three substrates (for mono-, tri- and
pentacyclization; Scheme 1) with wild-type enzyme and
several tunnel variants. The tunnel variants, in which amino
acid residues in the walls of the tunnels were mutated
(Figure 1a), were initially constructed in silico and MD
simulations combined with Caver analysis confirmed that
It was found that the rate of pentacycle formation
catalyzed by the wild-type enzyme increased two orders of
magnitude when the temperature increased from 308C to
558C. The data corresponds to an activation entropy of
16 kcalmolÀ1 at 328 K (or 50 calmolÀ1 KÀ1) favoring the
polycyclization reaction, whereas the high enthalpy of 31 kcal
molÀ1 disfavors catalysis. To corroborate the very large
entropic contribution to enzymatic polycyclization catalysis,
4846 www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 4845 –4849

Angewandte
Chemie
tunnel variant displayed up to
100-fold higher apparent kcat
/
KM values for pentacycle for-
mation compared to that of
the wild-type enzyme at tem-
peratures lower than 698C.
This is a consequence of a re-
duced activation enthalpy
(favoring
catalysis)
and
entropy (disfavoring cataly-
sis). By analogy, the S168F
tunnel variant displayed an
absolute apparent kcat/KM
value that was approximately
an order of magnitude higher
than that of the wild-type
enzyme at 408C (Figure 2a).
These facts demonstrate the
important consequences of
the enthalpy–entropy com-
pensation[23] and show how
a different arrangement of
water channels can affect the
synthetic capabilities of triter-
pene cyclases at different tem-
peratures. The S38W muta-
tion (located 8 ꢀ away from
Figure 2. Thermodynamic aspects of enzyme-catalyzed polycyclization. A) Data for the generation of the
natural pentacyclic scaffold (shown next to the graph with bonds formed/broken shown as dashed lines).
The intersection of the extrapolated dashed blue line (wild type) and red line (F605W) on the graph
corresponds to the temperature for which the two proteins have equal activities (698C). Experiments were
not conducted above 558C because of micelle stability issues. Error bars for the wild-type enzyme from four the polyisoprene substrate) is
independent experiments are shown for reference. B) Thermodynamic data from (a) were obtained by using
Equation (1) in the Supporting Information. C) A competition experiment with all three substrates in one-
pot to assess the relative thermodynamic consequences for different extents of cyclization. Data for
pentacyclic over monocyclic formation is shown in the Supporting Information. The generation of
unusual in that it displays
lower rates of polycyclization
at higher temperatures (with
linear kinetics) with a concom-
monocycles was associated with low activity and data could only be obtained for the wild-type enzyme.
itant massive decrease in
entropy of almost 60 kcal
molÀ1 compared to that of
the wild-type enzyme at
Bonds formed/broken are shown as dashed lines in the structures shown to the left of the graph.
D) Thermodynamic data obtained by using the data in (c) and Equation (2) in the Supporting Information.
we performed an additional thermodynamic analysis using
previously published kinetic parameters for pentacyclization
catalyzed by the same triterpene cyclase.[10c] From that study,
in which no thermodynamic analysis was performed (and for
which kcat/KM values were only available at two temper-
atures), we derived an activation enthalpy of 31 kcalmolÀ1
and an activation entropy of 19 kcalmolÀ1 (at 328 K). The
excellent agreement between the activation parameters that
we obtain for the two independent systems is noteworthy.
The favorable activation entropy of 16 kcalmolÀ1 is
surprisingly high when compared to the estimated entropic
cost (of opposite sign) of prefolding the C30 polyisoprene
backbone (at least 22 kcalmolÀ1 at 328 K according to the
prevention of free rotation around 15 bonds).[12] It should be
noted that the entropic cost of “freezing” the rotation around
a chemical bond is known experimentally.[12] Hence the
entropic contribution to enzyme-catalyzed pentacyclization
corresponds to a rate enhancement of up to 1028-fold,
assuming that all fifteen bonds are completely frozen. The
mutants harboring blocked channels were all active and
displayed a different temperature dependence of catalysis. By
contrast, control mutations in and around the active site had
small effects (see the Supporting Information). The F605W
328 K. The entropic contribution to the rate enhance-
ment was thus decreased 1040-fold by one point mutation.
This value corresponds to “freezing” 17 water molecules in
a salt crystal.[24] Such very large entropic effects, which are
furthermore not associated with large structural rearrange-
ments,[25] were previously not acknowledged in enzyme
catalysis.[26,27] Moreover, the association of ligands with
receptors and enzymes is not driven by entropy per se.[18]
The negative activation enthalpies are in accordance with
a difference in exposure of the hydrophobic active site to
water going from the ground state to the transition state and
have previously been observed for the kinetics of protein
folding.[28]
Kinetic isotope effect analysis in D2O (Figure S8) resulted
in an isotope effect of around two, which demonstrates that
protonation of the terminal isoprene unit of the substrate is
rate limiting (except for in S168W, which gave no isotope
effect). The measured entropic effect is thus concomitant with
the actual chemical reaction since cyclization occurs in
concert with the protonation that generates a transient
reactive carbocationic species.[11] A close to perfect linear
(Figure 3a, R2 0.9986) enthalpy–entropy compensation[23] for
the four proteins that display a kinetic isotope effect indicates
Angew. Chem. Int. Ed. 2014, 53, 4845 –4849
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 4847

.
Angewandte
Communications
the pentacyclic and the tricyclic product is calculated to be
only around 200 ꢀ2 (by using the modeling software
YASARA[30]). This result indicates that the entropic contri-
bution to catalysis is more complicated than what would
perhaps be expected, a fact that highlights the prerequisite of
prefolding of the substrate for cyclization. Moreover, the
formation of pentacycles was entropically disfavored com-
pared to the generation of tricycles for the S168F and F605W
tunnel variants (Figure 2d). The largest effect was observed
for the S168F variant, which showed a massive change in the
relative activation entropy for pentacycle over tricycle
formation of 84 kcalmolÀ1 when compared to that of the
wild-type enzyme (DS168F-wildtypeD5-3DS° at 328 K; Figure 2d).
These effects together indicate that the large entropic
contribution to catalysis is more complex than what would
be expected. The enthalpy–entropy compensation between
DDH° and DDS° was found to be close to perfect, which
indicates that different extents of cyclization could operate by
the same reaction mechanism (Figure 3b). Moreover, the
analyzed tunnel variants displayed a shift in the multicyclic
product spectrum in relation to that of the wild-type enzyme
(Figure 2c and Table S10 in the Supporting Information).
In summary, this study on a thermophilic cyclase provides
new mechanistic insight for a very important enzyme class.
The expulsion of water molecules through specific channels
assists in the prefolding of the linear polyisoprene chain and
entropy is the key to polycyclizations catalyzed by terpene
cyclases. This process allows remarkable chemistry through
a carbocationic polycyclization cascade initiated by the rate-
limiting transfer of a proton to the terminal isoprene group of
the prefolded substrate. Our results demonstrate how an
important class of enzymes has evolved to benefit from
entropy rather than enthalpy to perform efficient catalysis in
contrast to virtually all other described enzymes.[25,27,31] The
findings herein indicate that the use of entropy as a strategy in
enzyme catalysis likely appeared early[2] and could be of high
importance and more general than previously thought.[32]
Moreover, the entropy-driven cyclization strategy described
herein is expected to be of high importance in biology since
90% of all natural products contain a cyclic subunit.[33]
Figure 3. Enthalpy–entropy compensations with the entropic contribu-
tions given at 328 K. A) The activation parameters for pentacycle
formation were taken from Figure 2b. B) The T*DDS° values are
plotted against the corresponding DDH° values (taken from Fig-
ure 2d) for the competition experiments performed using three differ-
ent substrates with the wild-type enzyme and the variants.
that they could operate by the same reaction mechanism and
share the same transition state.
Mono-, tri-, and pentacyclic ring closures (Scheme 1) were
analyzed in a one-pot reaction in order to investigate the
impact of entropy on different extents of cyclization (Fig-
ure 2c,d and Equation (2) in the Supporting Information).
The rate for the one-ring closure was 2000-fold lower than for
pentacycle formation and could only be measured for the wild
type cyclase. Our results show that pentacyclization catalyzed
by the wild-type enzyme is entropically favored over mono-
cyclization (by 16 kcalmolÀ1 at 328 K), although the inherent
entropic constraints of formation of a productive precyclic
conformation for the latter are much less severe. By analogy,
the formation of the pentacyclic scaffold is entropically
favored over the tricyclic core by close to 30 kcalmolÀ1 for
the wild-type enzyme, as illustrated by an increase in the
relative amount of pentacyclic product with temperature
(Figure 2c,d). In terms of the hydrophobic effect,[29] this
energy difference would correspond to a massive change in
accessible surface area of around 1500 ꢀ2 for the two
substrates that differ by 13 atoms. By contrast, the difference
in the accessible surface area between the substrates that yield
Received: February 4, 2014
Revised: February 21, 2014
Published online: April 7, 2014
Keywords: biosynthesis · enzyme catalysis · polycyclization ·
.
reaction mechanisms · thermodynamics
[1] Y. Gao, R. B. Honzatko, R. J. Peters, Nat. Prod. Rep. 2012, 29,
1153 – 1175.
[2] J. J. Brocks, G. A. Logan, R. Buick, R. E. Summons, Science
1999, 285, 1033 – 1036.
[3] P. P. Peralta-Yahya, F. Zhang, S. B. del Cardayre, J. D. Keasling,
Nature 2012, 488, 320 – 328.
[4] P. K. Ajikumar, W.-H. Xiao, K. E. J. Tyo, Y. Wang, F. Simeon, E.
Leonard, O. Mucha, T. H. Phon, B. Pfeifer, G. Stephanopoulos,
Science 2010, 330, 70 – 74.
[5] D.-K. Ro, et al., Nature 2006, 440, 940 – 943.
[6] a) S. C. Hammer, P.-O. Syren, M. Seitz, B. M. Nestl, B. Hauer,
Curr. Opin. Chem. Biol. 2013, 17, 293 – 300; b) M. Seitz, P.-O.
4848 www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 4845 –4849

Angewandte
Chemie
Syren, L. Steiner, J. Klebensberger, B. M. Nestl, B. Hauer,
ChemBioChem 2013, 14, 436 – 439.
[17] a) T. Young, R. Abel, B. Kim, B. J. Berne, R. A. Friesner, Proc.
Natl. Acad. Sci. USA 2007, 104, 808 – 813; b) P. Ball, Chem. Rev.
2008, 108, 74 – 108; c) P. C. Weber, J. J. Wendoloski, M. W.
Pantoliano, F. R. Salemme, J. Am. Chem. Soc. 1992, 114, 3197 –
3200.
[7] a) D. C. Hibbit, R. P. Linstead, J. Chem. Soc. 1936, 470 – 476;
b) C. D. Upper, C. A. West, J. Biol. Chem. 1967, 242, 3285 – 3292;
c) D. J. Miller, R. K. Allemann, Nat. Prod. Rep. 2012, 29, 60 – 71;
d) B. T. Greenhagen, P. E. OꢁMaille, J. P. Noel, J. Chappell, Proc.
Natl. Acad. Sci. USA 2006, 103, 9826 – 9831.
[8] a) L. Ruzicka, Experientia 1953, 9, 357 – 367; b) G. Stork, A. W.
Burgstahler, J. Am. Chem. Soc. 1955, 77, 5068 – 5077; c) R.
Rajamani, J. Gao, J. Am. Chem. Soc. 2003, 125, 12768 – 12781;
d) Y. Jin, D. C. Williams, R. Croteau, R. M. Coates, J. Am. Chem.
Soc. 2005, 127, 7834 – 7842.
[18] P. W. Snyder, J. Mecinoviv, D. T. Moustakas, S. W. Thomas III,
M. Harder, E. T. Mack, M. R. Lockett, A. Heroux, W. Sherman,
G. M. Whitesides, Proc. Natl. Acad. Sci. USA 2011, 108, 17889 –
17894.
[19] D. J. Reinert, G. Balliano, G. E. Schulz, Chem. Biol. 2004, 11,
121 – 126.
[20] K. U. Wendt, A. Lenhart, G. E. Schulz, J. Mol. Biol. 1999, 286,
175 – 187.
[21] K. U. Wendt, K. Poralla, G. E. Schulz, Science 1997, 277, 1811 –
1815.
[22] E. Chovancova et al., PLoS Comput. Biol. 2012, 8, e1002708.
[23] L. Liu, Q.-X. Guo, Chem. Rev. 2001, 101, 673 – 695.
[24] J. D. Dunitz, Science 1994, 264, 670.
[25] M. Wang, K. Warncke, J. Am. Chem. Soc. 2013, 135, 15077 –
15084.
[26] a) M. Garcia-Viloca, J. Gao, M. Karplus, D. G. Truhlar, Science
2004, 303, 186 – 195; b) D. Ringe, G. A. Petsko, Science 2008, 320,
1428 – 1429.
[27] A. Sievers, M. Beringer, M. V. Rodnina, R. Wolfenden, Proc.
Natl. Acad. Sci. USA 2004, 101, 7897 – 7901.
[9] K. U. Wendt, Angew. Chem. 2005, 117, 4032 – 4037; Angew.
Chem. Int. Ed. 2005, 44, 3966 – 3971.
[10] a) E. E. Van Tamelen, J. D. Willett, R. B. Clayton, J. Am. Chem.
Soc. 1967, 89, 3371 – 3373; b) T. Hoshino, T. Sato, Chem.
Commun. 2002, 291 – 301; c) N. Morikubo, Y. Fukuda, K.
Ohtake, N. Shinya, D. Kiga, K. Sakamoto, M. Asanuma, H.
Hirota, S. Yokoyama, T. Hoshino, J. Am. Chem. Soc. 2006, 128,
13184 – 13194; d) D. W. Christianson, Curr. Opin. Chem. Biol.
2008, 12, 141 – 150; e) E. Oldfield, F.-Y. Lin, Angew. Chem. 2012,
124, 1150 – 1163; Angew. Chem. Int. Ed. 2012, 51, 1124 – 1137;
f) D. J. Tantillo, Nat. Prod. Rep. 2011, 28, 1035 – 1053.
[11] a) L. Smentek, B. A. Hess, J. Am. Chem. Soc. 2010, 132, 17111 –
17117; b) D. H. Aue, Wiley Interdiscip. Rev. Comput. Mol. Sci.
2011, 1, 487 – 508.
[28] M. Oliveberg, Y.-J. Tan, A. R. Fersht, Proc. Natl. Acad. Sci. USA
1995, 92, 8926 – 8929.
[12] M. I. Page, W. P. Jencks, Proc. Natl. Acad. Sci. USA 1971, 68,
1678 – 1683.
[29] D. Chandler, Nature 2005, 437, 640 – 647.
[30] E. Krieger, T. Darden, S. B. Nabuurs, A. Finkelstein, G. Vriend,
Proteins Struct. Funct. Bioinf. 2004, 57, 678 – 683.
[31] a) R. Wolfenden, Annu. Rev. Biochem. 2011, 80, 645 – 667;
b) S. S. Licht, C. C. Lawrence, J. Stubbe, Biochemistry 1999, 38,
1234 – 1242.
[32] R. B. Stockbridge, C. A. Lewis, Jr., Y. Yuan, R. Wolfenden,
Proc. Natl. Acad. Sci. USA 2010, 107, 22102 – 22105.
[33] K. Gilmore, I. V. Alabugin, Chem. Rev. 2011, 111, 6513 – 6556.
[13] G. Illuminati, L. Mandolini, Acc. Chem. Res. 1981, 14, 95 – 102.
ˇ
´
[14] I. Coric, B. List, Nature 2012, 483, 315 – 319.
[15] K. Ishihara, S. Nakamura, H. Yamamoto, J. Am. Chem. Soc.
1999, 121, 4906 – 4907.
[16] a) K. Manna, W. C. Everett, G. Schoendorff, A. Ellern, T. L.
Windus, A. D. Sadow, J. Am. Chem. Soc. 2013, 135, 7235 – 7250;
b) W. D. Huntsman, T. H. Curry, J. Am. Chem. Soc. 1958, 80,
2252 – 2254.
Angew. Chem. Int. Ed. 2014, 53, 4845 –4849
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 4849