Enzyme-like control of carbocation deprotonation regioselectivity in supramolecular catalysis of the nazarov cyclization

Article abstract of DOI:10.1002/anie.201105325

The kinetically controlled, regioselective deprotonation of cyclopentenyl cations is mediated by encapsulation within a metal-ligand assembly. The regiochemistry of the deprotonation step determines which one of two possible products is formed. Moreover, subtle differences in the stereochemistry of the encapsulated cation switch the selectivity of this process (see scheme). Copyright

Full text of DOI:10.1002/anie.201105325

Communications
DOI: 10.1002/anie.201105325
Supramolecular Catalysis
Enzyme-like Control of Carbocation Deprotonation Regioselectivity in
Supramolecular Catalysis of the Nazarov Cyclization**
Courtney J. Hastings, Mikael P. Backlund, Robert G. Bergman,* and Kenneth N. Raymond*
Acid–base reactions are among the fastest chemical reactions,
and as such they are typically reversible and under thermo-
dynamic control. However, under the appropriate conditions
some deprotonation reactions form kinetically favored prod-
ucts instead of the most thermodynamically stable species. A
sterically hindered base and low temperature are typically
necessary to ensure that the energetic difference between the
two competing transition states is large enough for complete
dehydration of geranyl pyrophosphate, and each exhibits a
very high degree of regioselectivity in producing either
myrcene or (E)-b-ocimene. The amino acid sequences of
the two enzymes are 93% identical, yet this small structural
difference completely switches the regioselectivity of geranyl
cation deprotonation.
[
3]
The high levels of selectivity achieved in enzymatic
catalysis are the result of precise control over the substrate
conformation and its interactions with catalytic functional
groups or other reactants within the active site. Supramolec-
ular encapsulation is similarly capable of enforcing a single
conformation of a bound guest molecule and the orientation
[1]
kinetic selectivity. Some enzymes involved in terpene
biosynthesis exert kinetic control over the deprotonation of
allyl cation intermediates, determining which products are
ultimately formed. The acid-catalyzed ionization of geraniol
or geranyl pyrophosphate produces the geranyl cation, which
can be deprotonated at one of two positions to form either
myrcene or b-ocimene (Scheme 1). In the absence of enzyme,
there is little selectivity for deprotonation of the geranyl
[
4]
of two co-encapsulated guests relative to one another. This
control over guest geometry can enhance the selectivity of
reactions that proceed inside a molecular host cavity by
[5]
favoring specific reaction pathways. We describe here the
kinetically controlled, regioselective deprotonation of cyclo-
pentenyl cations, the selectivity of which is governed by
encapsulation within the cavity of a self-assembled host. This
represents a rare example of a synthetic kinetic deprotonation
that does not rely on either low temperature or a bulky base
for its selectivity, and is the first example of supramolecular
control over a deprotonation reaction. Additionally, this
reactivity provides a completely synthetic analogue of the
regioselective, enzyme-controlled deprotonation of the ger-
anyl cation involved in the biosynthesis of myrcene and b-
ocimene.
[
2]
cation at one position over the other. Two enzymes isolated
from the snapdragon flower (Antirrhinum majus) catalyze the
1
2À
We recently disclosed the ability of the [Ga L ]
4
6
assembly (1, where L = N,N’-bis(2,3-dihydroxybenzoyl)-1,5-
[6]
diaminonaphthalene, Figure 1) to catalyze the Nazarov
cyclization of 1,3-pentadienols to form cyclopentadienes in
[
7]
aqueous solution. The ligand framework of 1 generates a
Scheme 1. Methyl deprotonation of the geranyl cation yields myrcene,
while methylene deprotonation produces either stereoisomer of b-
ocimene.
3
large, hydrophobic interior cavity (250–450 ꢀ ) that can
encapsulate suitably-sized cationic and neutral guest mole-
[*] C. J. Hastings, M. P. Backlund, Prof. R. G. Bergman,
Prof. K. N. Raymond
Department of Chemistry, University of California, Berkeley
Berkeley, CA 94720-1416 (USA)
and
Lawrence Berkeley National Laboratory
1
Cyclotron Road, Berkeley, CA 94720 (USA)
E-mail: rbergman@berkeley.edu
raymond@socrates.berkeley.edu
[
**] This work was supported by the Director of the Office of Energy
Research, Office of Basic Energy Sciences, Chemical Sciences
Division of the U.S. Department of Energy under Contract DE-AC02-
Figure 1. Left: Schematic view of 1 in which the bisbidentate ligands
are represented by blue lines and the gallium atoms by red circles.
Right: Space-filling model of 1 (C black, H white, O red, N blue, Ga
yellow).
0
5CH11231 and through a fellowship from Chevron (to C.J.H.).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201105325.
1
0570
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 10570 –10573

cules, and it is encapsulation within the host cavity that is
responsible for the catalysis of the Nazarov cyclization. The
quantitative formation of the Diels–Alder adduct of Cp*H
In considering the mechanism of the Nazarov cyclization
(Scheme 4a), the formation of 5 must occur through depro-
tonation of cyclopentenyl cation 8a at the appropriate methyl
group instead of at the cyclopentyl position. We suggest that
the outcome of the 1-catalyzed reactions of the three
stereoisomers of 2 was dictated by the stereochemistry of
the encapsulated cyclopentenyl cation intermediate (8a
versus 8b). The 4p electrocyclization of pentadienyl cations
(
pentamethylcyclopentadiene; 3) with maleimide, 4, was
1
observed by H NMR analysis in each of the reactions
examined in our rate and mechanism studies (Scheme 2).
[12]
occurs in a conrotatory fashion, so the alkene stereochem-
istry of the pentadienyl cations determines the stereochem-
istry of the resulting cyclopentenyl cation. Accordingly, the
electrocyclization of pentadienyl cations 7a and 7b (derived
from 2a and 2b, respectively) yields cation 8a with methyl
groups in the trans orientation, while the E,Z pentadienyl
cation 7c (derived from 2c) forms 8b with methyl groups in
cis orientation (Scheme 4b). This explanation is supported
by the observation of 5 as the sole product of 1-catalyzed
dehydration of alcohol 9a (Scheme 5a), whose dehydration
Scheme 2. Product of the 1-catalyzed Nazarov reaction in the presence
of a Diels–Alder trap.
[9]
The 1-catalyzed reaction of 2a or 2b in unbuffered D O at
2
room temperature, however, led to the formation of the
unexpected dihydrofulvene 5, which is isomeric to Cp*H
(
Scheme 3). The reaction of 2c conducted under identical
conditions yielded the expected product, Cp*H. In both
reactions, the selectivity for the observed product was greater
than 9:1. No report of either 5 or its diastereomer, 6, have
[
8]
been reported in the literature, although a mixture of the
two diastereomers was proposed as the minor product of a
[
9]
carbocation quenching reaction.
Scheme 4. a) Mechanism of the Nazarov cyclization, showing the
divergence that produces either 3 or 5. b) The stereochemistry of the
cyclopentenyl cation is determined by the olefin geometry of the
preceding pentadienyl cation.
Scheme 3. In unbuffered D O at room temperature, the 1-catalyzed
2
Nazarov reaction of symmetrical substrates 2a and 2b produces 5,
while the reaction of 2c forms Cp*H.
The observation of 5 as a reaction product was initially
puzzling, since it had not been detected when the reaction was
conducted under other conditions. Kinetic analysis of the 1-
catalyzed reaction of 2a or 2b in 1:1 D O/[D ]DMSO with
2
6
added maleimide (which cannot react with 5, due to the
forced trans orientation of its diene fragment) displays clean
conversion of reactant into 4 without the accumulation of
[
10]
significant quantities of any reaction intermediate.
Sub-
jecting 5 (in the absence of 1) to conditions similar to those
used to measure kinetic data (458C in 1:1 D O/[D ]DMSO)
2
6
caused quantitative conversion to Cp*H. This observation
implies that during our rate studies, 5 is initially produced as
the kinetic product from 2a and 2b, but is immediately
converted into the thermodynamic product Cp*H and
Scheme 5. a) The 1-catalyzed dehydration of cyclopentenyl alcohol 9a.
b) The acid-catalyzed dehydration reactions of cyclopentenyl alcohols
9a and 9b.
[11]
trapped by maleimide.
Angew. Chem. Int. Ed. 2011, 50, 10570 –10573
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 10571

Communications
proceeds through the intermediate trans cyclopentenyl cation
a.
We hypothesized that encapsulation in 1 was responsible
for the formation of either 5 or Cp*H, depending on the
stereochemistry of the encapsulated cyclopentenyl cation
deprotonating the geranyl cation in the biosynthesis of
myrcene and ocimene (Scheme 1). In both cases, deprotona-
tion of an allyl carbocation can potentially occur at multiple
positions to form diene products, and minor structural
changes are responsible for complete inversion of regioselec-
tivity at room temperature. In fact, the cyclopentenyl cations
from this study are constitutional isomers of the geranyl
cation, and the products 3 and 5 are constitutional isomers of
myrcene and ocimene. These similarities raise the possibility
that 1 could act as a mimic for some of the cyclization
reactions involved in terpene biosynthesis.
In conclusion, the first example of selective, kinetic
deprotonation mediated by supramolecular encapsulation
has been demonstrated in the 1-catalyzed Nazarov reaction of
1,4-pentadien-3-ols. The regiochemistry of deprotonation in
the host-catalyzed reaction is determined by the stereochem-
istry of an intermediate cyclopentenyl cation, the structure of
which is determined by the alkene stereochemistry of the
reactant. Changing the relative stereochemistry of two methyl
groups in the encapsulated carbocationic intermediate from
trans (8a) to cis (8b) completely switches the regioselectivity
of deprotonation, forming the corresponding diene regioiso-
mer with greater than 9:1 selectivity. In contrast to their host-
mediated reactivity, the deprotonation reactions of these
carbocations in free solution were not selective, leading to a
mixture of regioisomers in both cases. We propose that
supramolecular encapsulation within 1 forces deprotonation
to occur at a single position. This mimics the enzyme-
mediated deprotonation reactions involved in terpene bio-
synthesis.
8
(8aꢀ1 or 8bꢀ1, where ꢀ denotes encapsulation). To test this
notion, it was necessary to evaluate the products of the
deprotonation reaction of 8a and 8b in bulk solution, in case
the formation of 5 from 8a (or Cp*H from 8b) is an intrinsic
property of the cation unaffected by encapsulation in 1. The
dehydration of 9a conducted under benzoic acid catalysis in
the absence of 1 yielded a 2:3 ratio of 5 to 3 (Scheme 5b),
while the analogous reaction of 9b yielded a 1:3 ratio of 6 to 3.
The ratio of 5 (or 6) to 3 does not change over the course of
the acid-catalyzed reaction, indicating that little isomerization
of 5 or 6 occurs. Thus, the product ratio of this reaction
reflects the kinetic selectivity for deprotonating the cyclo-
pentenyl carbocation intermediate. These data indicate that
there is no significant kinetic preference for deprotonation at
either of the two positions of cyclopentenyl cations 8a and 8b
in free solution.
The above observations make clear that encapsulation of
these cations in 1 is directing the regiochemistry of deproto-
nation, producing Cp*H (3) from 8bꢀ1, and 5 from 8aꢀ1
(
Scheme 6). Although we were unable to obtain structural
data on the short-lived host–guest complexes 8bꢀ1 and 8aꢀ1,
the most likely explanation is that the specific orientation of
the carbocation within the cavity of 1 diminishes the
accessibility of one proton, forcing deprotonation to occur
exclusively at the other position. Changing from 8a to 8b
could require a different orientation within 1, switching the
accessibility of the two possible deprotonation sites. Given the
subtle structural difference between 8a and 8b, this encap-
sulation-mediated inversion of regioselectivity is remarkable,
especially when one considers the absence of functional
groups within the cavity of 1. This example of kinetically
controlled deprotonation in supramolecular catalysis is strik-
ingly similar to the enzymatic control of regiochemistry in
Received: July 28, 2011
Published online: September 20, 2011
Keywords: cage compounds · carbocations ·
.
electrocyclic reactions · homogeneous catalysis ·
supramolecular chemistry
[
1] J. dꢁAngelo, Tetrahedron 1976, 32, 2979.
[
2] a) C. Bunton, O. Cori, D. Hachey, J. Leresche, J. Org. Chem.
1979, 44, 3238; b) F. Cramer, W. Rittersdorf, Tetrahedron 1967,
23, 3015.
[
3] N. Dudareva, D. Martin, C. M. Kish, N. Kolosova, N. Gorenstein,
J. Fꢂldt, B. Miller, J. Bohlmann, Plant Cell 2003, 15, 1227.
4] a) J. Rebek, Jr., Angew. Chem. 2005, 117, 2104; Angew. Chem.
Int. Ed. 2005, 44, 2068; b) M. Yoshizawa, J. Nakagawa, K.
Kumazawa, M. Nagao, M. Kawano, T. Ozeki, M. Fujita, Angew.
Chem. 2005, 117, 1844; Angew. Chem. Int. Ed. 2005, 44, 1810;
c) M. D. Pluth, D. Fiedler, J. S. Mugridge, R. G. Bergman, K. N.
Raymond, Proc. Natl. Acad. Sci. USA 2009, 106, 10438.
[
[
5] a) R. Breslow, Acc. Chem. Res. 1995, 28, 146; b) C. J. Brown,
R. G. Bergman, K. N. Raymond, J. Am. Chem. Soc. 2009, 131,
17530; c) J. Chen, J. Rebek, Jr., Org. Lett. 2002, 4, 327; d) S. Das,
C. D. Incarvito, R. H. Crabtree, G. W. Brudvig, Science 2006,
312, 1941; e) T. Furusawa, M. Kawano, M. Fujita, Angew. Chem.
2007, 119, 5819; Angew. Chem. Int. Ed. 2007, 46, 5717; f) W. L.
Mock, T. A. Irra, J. P. Wepsiec, M. Adhya, J. Org. Chem. 1989,
54, 5302; g) Y. Nishioka, T. Yamaguchi, M. Kawano, M. Fujita, J.
Am. Chem. Soc. 2008, 130, 8160; h) F. R. Pinacho Crisꢃstomo, A.
Lledꢃ, S. R. Shenoy, T. Iwasawa, J. Rebek, Jr., J. Am. Chem. Soc.
2009, 131, 7402; i) V. F. Slagt, P. C. J. Kamer, P. W. N. M.
Scheme 6. The stereochemistry of the encapsulated cyclopentenyl
cation (8a versus 8b, drawn larger than scale to show the substrate
structure) determines the site of deprotonation, and the regiochemis-
try of the diene product (3 versus 5).
1
0572 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 10570 –10573

van Leeuwen, J. N. H. Reek, J. Am. Chem. Soc. 2004, 126, 1526;
j) T. Smejkal, B. Breit, Angew. Chem. 2008, 120, 317 – 321;
[7] C. J. Hastings, M. D. Pluth, R. G. Bergman, K. N. Raymond, J.
Am. Chem. Soc. 2010, 132, 6938.
ˇ
Angew. Chem. Int. Ed. 2008, 47, 311 – 315; k) M. Yoshizawa, M.
Tamura, M. Fujita, Science 2006, 312, 251.
[8] See Supporting Information for characterization data and
experimental details.
[
6] a) D. Caulder, R. Powers, T. Parac, K. Raymond, Angew. Chem.
[9] P. H. Campbell, N. W. K. Chiu, K. Deugau, I. J. Miller, T. S.
Sorensen, J. Am. Chem. Soc. 1969, 91, 6404.
10] Ref. [7].
11] A discussion of the isomerization of 5 is presented in the
Supporting Information.
1998, 110, 1940; Angew. Chem. Int. Ed. 1998, 37, 1840; b) D.
[
[
Fiedler, D. H. Leung, R. G. Bergman, K. N. Raymond, Acc.
Chem. Res. 2005, 38, 349; c) M. D. Pluth, R. G. Bergman, K. N.
Raymond, Acc. Chem. Res. 2009, 42, 1650.
[
12] R. B. Woodward, R. Hoffmann, J. Am. Chem. Soc. 1965, 87, 395.
Angew. Chem. Int. Ed. 2011, 50, 10570 –10573
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 10573