Terpene cyclization catalysed inside a self-assembled cavity

Article abstract of DOI:10.1038/nchem.2181

In nature, complex terpene natural products are formed by the so-called tail-to-head terpene (THT) cyclization. The cationic reaction cascade is promoted efficiently in complex enzyme pockets, in which cationic intermediates and transition states are stabilized. In solution, the reaction is hard to control and man-made catalysts able to perform selective THT cyclizations are lacking. We herein report the first example of a successful THT cyclization inside a supramolecular structure. The basic mode of operation in cyclase enzymes was mimicked successfully and a catalytic non-stop THT was achieved with geranyl acetate as the substrate. The results presented have implications for the postulated reaction mechanism in cyclase enzymes. Evidence indicates that the direct isomerization of a geranyl cation to the cisoid isomer, which so far was considered unlikely, is feasible.

Full text of DOI:10.1038/nchem.2181

ARTICLES
PUBLISHED ONLINE: 16 FEBRUARY 2015 | DOI: 10.1038/NCHEM.2181
Terpene cyclization catalysed inside
a self-assembled cavity
Q. Zhang and K. Tiefenbacher
*
In nature, complex terpene natural products are formed by the so-called tail-to-head terpene (THT) cyclization. The
cationic reaction cascade is promoted efficiently in complex enzyme pockets, in which cationic intermediates and transition
states are stabilized. In solution, the reaction is hard to control and man-made catalysts able to perform selective THT
cyclizations are lacking. We herein report the first example of a successful THT cyclization inside a supramolecular
structure. The basic mode of operation in cyclase enzymes was mimicked successfully and a catalytic non-stop THT was
achieved with geranyl acetate as the substrate. The results presented have implications for the postulated reaction
mechanism in cyclase enzymes. Evidence indicates that the direct isomerization of a geranyl cation to the cisoid isomer,
which so far was considered unlikely, is feasible.
ature’s extraordinary elegance when performing chemical control, and give mixtures of acyclic and, to a much lesser extent,
reactions has fascinated and inspired chemists for a long cyclic products. Especially, the conversion of geranyl substrates in
4
N
time. Arguably, one of the most complex organic transform- solution is problematic .
1
ations performed in living organisms is the tail-to-head terpene
As the acyclic terpene alcohols geraniol (GOH), farnesol and ger-
THT) cyclization . It allows the construction of the most diverse anylgeraniol are readily available, a direct selective conversion into
2,3
(
class of natural products, namely terpenes, through nature’s cyclic terpenes is highly desirable. What are the main differences
version of combinatorial chemical synthesis. Thousands of different between THT cyclizations in solution and those in the hydrophobic
natural products are formed from just a handful of simple, acyclic enzyme pocket? (1) The initial ionization requires harsh conditions
starting materials: geranyl pyrophosphate (GPP) (1 (Fig. 1a), mono- (a strong Lewis or Brønsted acid) in solution. In the cyclase,
terpenes), farnesyl pyrophosphate (2, sesquiterpenes) and geranyl- however, the pyrophosphate leaving group is removed under
4
geranyl pyrophosphate (3, diterpenes). Nature utilizes enzymes, neutral conditions (optimum pH 6–7) with the help of divalent
termed cyclases or terpene synthases, to bind the acyclic terpene metal ions. (2) In solution, high-energy cationic intermediates are
diphosphate in a hydrophobic reaction pocket. The reaction susceptible to undesired side reactions (elimination or quenching
cascade is initiated by a metal(II)-triggered departure of the by nucleophiles). In the enzyme pocket, cationic key intermediates
diphosphate leaving group, generating, in the case of GPP (1), the are stabilized by cation–π and cation–dipole interactions.
first carbocationic intermediate 4a (Fig. 1b)2 . A direct cyclization Specifically, it has been proposed that the aromatic side chains of
of cation 4a into the α-terpinyl cation 6 is not possible; it has to phenylalanine, tyrosine and tryptophan play a key role in this
–8
1
0,11
be preceded by isomerization into the cisoid conformer 4b. It was regard . In addition, the enzyme pocket blocks the access of
postulated that cation 4a has to collapse to linalyl pyrophosphate undesired nucleophiles, and thereby prevents premature quenching.
(linalyl-PP) (5a) to allow the isomerization to the cisoid isomer
Therefore, we explored the possibility of mimicking the complex
5
b, which is suitable for cyclization. It was argued that the free- cyclase reaction pocket using relatively simple aromatic supramole-
1
2–19
energy barrier for the direct rotation of the allylic cation 4a is rela- cular capsules
tively high (about 55 kJ mol ) and therefore unlikely . We herein active (for recent examples, see Bocokić et al. , Dydio et al. ,
present evidence that such a direct isomerization is possible in an Wang et al. , Zhao et al.
. Interestingly, although the field is quite
−
1
4,9
20
21
2
2
23,24
25
and Salles et al. ), no THT cyclizations
enzyme-like catalyst. The α-terpinyl cation 6, formed by the cycliza- in a supramolecular structure have been reported to date. The only
tion of the cisoid cation 4b, is a key intermediate in the biosynthesis. reported example of a monoterpene-like cyclization was based on a
2
6
It is converted into a variety of different monoterpene products, Prins reaction, which is reproducible in solution . The hexameric
2
7
depending on the properties of the cyclase. If a water molecule is resorcinarene capsule I (Fig. 2) (for catalysis within I, see
2
8
29
30
present in the reaction pocket, α-terpineol (7) can be formed, Cavarzan et al. , Bianchini et al. and Zhang and Tiefenbacher ),
2
,4,5,8
which may be cyclized further into eucalyptol (8)
.
which is readily available by self-assembly of the monomer 13 in
Alternatively, a direct proton elimination of 6 would deliver terpino- apolar solvents, seemed promising for this study. (1) It forms
3
lene (9) or limonene (10). If, however, cation 6 is stabilized enough a large hydrophobic cavity (∼1,400 Å ) that is suitable for accom-
and nucleophiles excluded within the reaction pocket, it is able to modating acyclic terpenes. (2) Owing to its aromatic walls, I is
undergo alternative intramolecular cyclizations or rearrangements. known to complex cationic guests, such as alkylammonium ions,
3
1,32
For example, it can rearrange into cation 11 via a hydride shift via cation–π interactions . In analogy to the aromatic residues
and then deliver α-terpinene (12). In any case, a wide selection of in a terpene cyclase enzyme pocket, the aromatic walls of I could
reaction pathways is available for the high-energy cationic inter- potentially stabilize cationic intermediates and transition states.
mediates: intra- or intermolecular substitution reactions, elimin- (3) We recently reported that capsule I is a mild Brønsted acid
ations, rearrangements, oligomerizations and polymerizations. Not (pKa ∼5.5–6) and can function as a catalyst for simple acetal
3
0
surprisingly, in solution THT reactions prove to be hard to hydrolysis . (4) Guest exchange is facile and is believed to occur
Department Chemie, Technische Universität München, Lichtenbergstraße 4, D-85747 Garching, Germany. *e-mail: konrad.tiefenbacher@tum.de
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
1
©
2015 Macmillan Publishers Limited. All rights reserved

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2181
ARTICLES
a
OPP
OPP
OPP
Geranyl-PP (1)
Farnesyl-PP (2)
Geranylgeranyl-PP (3)
b
Direct
isomerization?
O
OH
Eucalyptol (8)
α-Terpineol (7)
2
,3-Transoid
2,3-Cisoid
cation (4b)
α-Terpinyl
cation (6)
cation (4a)
OPP
OPP
Terpinolene (9) Limonene (10)
Linalyl-PP (5a)
transoid)
Linalyl-PP (5b)
(cisoid)
Terpinen-
4-yl-cation (11)
α-Terpinene (12)
(
Figure 1 | Biosynthesis of cyclic terpene natural products. a, Structures of the natural terpene substrates. These are converted into cyclic monoterpenes
geranyl-PP), sesquiterpenes (farnesyl-PP) and diterpenes (geranylgeranyl-PP). b, Reaction pathways for the biosynthesis of a selection of monoterpene
(
natural products via THT cyclization. Cleavage of the PP-leaving group forms the transoid cation 4a, which has to isomerize to the cisoid form 4b before
cyclization can occur. A direct isomerization is considered unlikely in the literature and therefore a step-wise isomerization that involves reattachment of the
pyrophosphate group was proposed. After cyclization to the α-terpinyl cation (6), a variety of different reaction paths are available. An attack of water forms
α-terpineol (7), which can further cyclize to eucalyptol (8). Alternatively, elimination delivers terpinolene (9) or limonene (10). Rearrangements can also
occur. A hydride shift, for instance, delivers cation 11, which yields α-terpinene (12) after elimination.
33
via dissociation of one resorcinarene unit . We speculated that it Supplementary Tables 5 and 6). These experiments provide strong
could potentially activate a suitable leaving group on an acyclic evidence that the cyclization reaction is, indeed, only taking place
terpene by protonation, and thereby initiate a THT cyclization inside the cavity. Additionally, substrate internalization was
1
cascade inside the cavity. Indeed, the system is able to biomimeti- evidenced by H- and diffusion-ordered NMR spectroscopy (see
cally catalyse THT cyclizations and herein we report these findings. Supplementary Figs 6–8).
Conversion of nerol (NOH) (16) under similar conditions pro-
Results and discussion
duced contrasting results (Fig. 3b): the conversion was much faster
We first investigated the cyclization of monoterpene alcohols. and the dominating product within the first 20 hours was α-terpineol
Treatment of a solution of GOH (14) in CDCl (33 mM) with (7), which was then converted into eucalyptol (8) with good selectiv-
3
1
0 mol% capsule I at 30 °C led to complete conversion of the starting ity. Besides α-terpinene (12), terpinolene (9) was also formed in
material within 28 hours (Fig. 3a). Identification of the products considerable amounts. Side reactions were greatly reduced.
formed was achieved by comparison with standard solutions of
The product profile of LOH (15) (Fig. 3c) could be interpreted
1
monoterpenes (gas chromatography, H NMR spectroscopy). as a combination of the GOH (14) and NOH (16) results to give a
Initially, in analogy to the proposed biosynthesis (Fig. 1b), linalool
(LOH) (15) was formed as the main product (for the initial rates
HO
OH
of formation, see Supplementary Table 9), which indicates that,
after protonation of GOH (14) and ionization to give cation 4a,
the cleaved water molecule attacked the allylic carbocation. The
other two main products formed within the first ten hours were
the cyclic α-terpinene (12) and α-terpineol (7). With progressing
reaction time, the concentrations of LOH (15) and α-terpineol (7)
decreased, and eucalyptol (8) started to form (for the final yields
after 72 hours, see Supplementary Table 4). Alkylation of the
phenol groups of monomer 13 (for details, see Supplementary
Figs 13–15) and dimerization were observed as side reactions, as evi-
denced by a multitude of smaller gas chromatography peaks in the
diterpene region. For dimerization inside the cavity, two substrates
have to be encapsulated simultaneously. It was reported that six
chloroform molecules occupy the cavity space if no other guests
R
R
HO
HO
OH
OH
Self-
assembly
R
R
HO
OH
1
3 (R = C11H23)
= OH group
=
=
H O molecule
2
Intermolecular
hydrogen bonding
I = 136(H2O)8
34
are present . According to size considerations (see
Supplementary Table 8), one substrate would replace two or three Figure 2 | Structure of the resorcinarene capsule I. This self-assembles
of the six chloroform molecules. Several control experiments were via a hydrogen-bond network from six monomer units of 13 and eight water
3
performed. When catalyst I was omitted, no conversion was molecules in apolar solvents and encloses a volume of ∼1,400 Å . Cationic
observed (12 days). When the catalyst’s cavity was blocked with a guests, such as alkylammonium ions, are bound inside the cavity via cation–π
strongly binding guest (nBu NBr), no cyclization was observed, interactions. Guest exchange is facile and probably occurs via dissociation of
4
although the acidity of the system was thereby increased (see one monomer unit. Capsule I is a mild Brønsted acid (pK ∼5.5–6).
a
2
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
©
2015 Macmillan Publishers Limited. All rights reserved

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2181
ARTICLES
a
45
c
45
40
35
30
7
4
3
3
2
2
1
1
0
5
0
5
0
5
0
5
0
10 mol%
O
OH
I
10 mol%
I
15
OH
12
8
14
12
2
2
1
1
5
0
5
0
5
0
12
0
10
20
30
40
Time (h)
50
60
70
80
0
10
20
30
40
Time (h)
50
60
70
80
b
O
1
0 mol%
I
O
OH
4
4
3
3
2
2
1
1
5
0
5
0
5
0
5
0
5
0
16
OH
Eucalyptol (8)
Terpinolene (9)
α-Terpineol (7)
OH
7
8
OH
Limonene (10)
α-Terpinene (12)
Linalool (15)
0
10
20
30
40
Time (h)
50
60
70
80
γ-Terpinene (17) Isoterpinolene (18) Substrate
1
Figure 3 | Results of the THT cyclizations with catalyst I for the alcohol substrates. Products were identified by gas chromatography and H NMR
spectroscopy and quantified by gas chromatography. a, The cyclization of GOH (14) is rather unselective. Initially, in analogy to the proposed biosynthesis,
LOH (15) was formed. The other main products formed within the first ten hours were α-terpinene (12) and α-terpineol (7). With progressing reaction time,
the concentrations of LOH (15) and α-terpineol (7) decreased as eucalyptol (8) started to form. b, NOH (16) is first cyclized mainly to α-terpineol (7), which
is then converted into eucalyptol (8). c, The reaction profile of LOH (15) appears to be a composite of the results from GOH (14) and NOH (16), and gives a
less-pronounced initial maximum of α-terpineol (7) and a better selectivity of α-terpinene (12) over terpinolene (9) than in the case of NOH (16). Error bars
indicate typical maximal errors (±5%) as evidenced by the experiment, which was run in triplicate (Fig. 4a).
less-pronounced initial maximum of α-terpineol (7) and a better solution (for a different approach to ‘non-stop’ THT cyclizations,
1
selectivity of α-terpinene (12) over terpinolene (9) than in the case see Pronin and Shenvi ). As described for GOH (14), a series of
of NOH (16). In addition, side reactions were between the control experiments were performed with 14a and confirmed that
two extremes.
the reaction did, indeed, only take place inside the cavity (see
The cyclization results are remarkable; for example, eucalyptol Supplementary Fig. 2 and Supplementary Table 6). Additionally,
(8) has not been formed in a cascade reaction from acyclic terpenes 14a was converted selectively inside I in the presence of the
before. It was only accessible from cyclic α-terpineol utilizing a extended derivative 19 (Fig. 4d) (for the details, see
strong Brønsted acid35 or Lewis acid , or in a two-step procedure Supplementary Fig. 4 and Supplementary Scheme 1). For compari-
36
37
via phenylselenoetherification . Nevertheless, we wanted to sup- son, several Brønsted and Lewis acids were tested for the cyclization
press the interception of cationic intermediates by the cleaved of GOAc (14a) and showed only traces of cyclic monoterpene
leaving group, which resulted, for example, in the formation of products (see Supplementary Fig. 3).
α-terpineol (7) as an intermediate. We therefore tested leaving
groups with a reduced nucleophilicity and found that the acetate duct profile also changed dramatically. As expected, only trace
proved to be well suited in that regard. amounts of α-terpineol (7) were observed. Nevertheless, a small
In the case of neryl acetate (NOAc) (16a) (Fig. 4b), the pro-
Submitting geranyl acetate (GOAc) (14a) to the described reac- fraction of α-terpinyl acetate (7a) was formed. More significantly,
tion conditions resulted in the selective formation of α-terpinene the initial product ratio between α-terpinene (12) and terpinolene
(12) (Fig. 4a). The interception of reactive cationic intermediates (9) changed dramatically, with the latter dominating in the first
was suppressed successfully, as evidenced by the disappearance of 36 hours of the reaction. The product spectrum of linalyl
LOH and α-terpineol derivatives. The α-terpinyl cation formed acetate (LOAc) (15a), as in the case of the free alcohols,
(6) (Fig. 1b) is not quenched by nucleophiles or elimination, but appears to be a composite of the results of GOAc and NOAc
is able to propagate via a 1,2-hydride shift to form cation 11 in (Fig. 4c). Additionally, GOAc was formed initially, which could
good selectivity. Therefore, this reaction represents a true indicate an equilibration to the thermodynamically more
‘
non-stop’38 THT cyclization cascade, which is hard to achieve in stable isomer.
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
3
©
2015 Macmillan Publishers Limited. All rights reserved

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2181
ARTICLES
a
c
1
0 mol%
I
45
40
10 mol%
OAc
I
OAc
‘Non-stop’
4
4
3
3
2
2
1
1
5
0
5
0
5
0
5
0
5
0
14a
15a
3
3
5
0
12
12
9
25
2
1
1
0
5
0
5
0
0
10
20
30
40
50
60
70
80
0
10
20
30
40
50
60
70
80
Time (h)
Time (h)
b
10 mol%
I
O
OAc
16a
OAc
4
5
12
9
α-Terpinyl acetate (7a) Eucalyptol (8)
Terpinolene (9)
4
3
3
2
2
1
1
0
5
0
5
0
5
0
5
0
OAc
Limonene (10)
α-Terpinene (12)
Geranyl acetate (14a)
0
10
20
30
40
50
60
70
80
Time (h)
γ-Terpinene (17)
Isoterpinolene (18) Substrate
d
C10H21
e
f
F
F
1
0 mol%
I
F
OAc
OAc
19
F-14a
F
F-16a
OAc
F-9
F-10
Figure 4 | Results of the THT cyclizations with catalyst I for the acetate substrates. a, The cyclization of GOAc (14a) led to α-terpinene (12) in good
selectivity. It represents a ‘non-stop’ THT cyclization because the cationic intermediates were not quenched by nucleophiles or elimination—as evidenced by
the disappearance of LOH and α-terpineol derivatives. b, NOAc (16a) cyclized mainly to α-terpinene (12) and terpinolene (9), with the latter dominating in
the first 36 hours. c, The product profile of LOAc (15a) appears to be a composite of the results from GOAc (14a) and NOAc (16a). Initially, GOAc was
formed, which could indicate an equilibration to the thermodynamically more stable isomer. d, The extended geranyl substrate 19 used for the control
experiments. e,f, Fluoro derivatives explored in the cyclization. 2-Fluorogeranyl acetate (F-14a) did not undergo cyclization within I (e). 2-Fluoroneryl acetate
(F-16a) produced fluoroterpinolene (F-9) and fluorolimonene (F-10) as the only main products (f), which indicates that these products are formed
predominantly via a concerted (S 2) mechanism, whereas α-terpinene (12) is formed predominantly via an S 1-mechanism within I. Error bars in a represent
N
N
the standard deviation from the mean value (the experiment was run in triplicate). Error bars in b and c indicate typical maximal errors (±5%), as evidenced
by the results in a.
The dramatically different product profiles for GOAc (14a) and (half-life < 0.5 milliseconds, short compared to typical turnovers
4
LOAc (15a) were striking. If GOAc (14a) formed the cisoid cation of every 1–100 seconds in cyclase enzymes ), a direct isomerization
4
b (Fig. 1b) via LOAc (15a) as an intermediate, as proposed in should no longer be excluded in cyclase enzymes.
the biosynthesis, such a difference would not be expected. Compared to NOH (16) and LOH (15), the reactions of NOAc
Additionally, no LOAc (15a) was observed as an intermediate in (16a) and LOAc (15a) led to an increased initial formation of
the cyclization of GOAc (14a). In contrast to natural enzymes, inter- terpinolene (9) and limonene (10), respectively, as compared to
mediates can be detected during cyclization in I, as can be seen in α-terpinene (12). This may be attributed to the presence of the
Figs 3a–c and 4b,c. Based on these results, we propose that the acetate/acetic acid leaving group in the reaction pocket, which
allylic cation 4a is, indeed, directly undergoing the transoid–cisoid could function as a general base to deprotonate the α-terpinyl
1
isomerization into 4b inside I. As there is no direct evidence cation 6 (Fig. 1b) . The high degree of deprotonation in the cases
for the formation of linalyl-PP as an intermediate in cyclase of NOAc (16a) and LOAc (15a), as compared with the results of
4
−1
enzymes and the isomerization energy barrier of 55 kJ mol (gas GOAc (14a), might seem surprising at first because all the sub-
9
phase) can be overcome relatively quickly at ambient conditions strates have to converge to the common intermediate α-terpinyl
4
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
©
2015 Macmillan Publishers Limited. All rights reserved

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2181
ARTICLES
cation 6. We speculated that this may indicate a different reaction of natural substrates to cyclase enzymes may be the strong inter-
mechanism. In the case of GOAc (14a), because of its 2E-alkene action of the pyrophosphate group with magnesium ions in the
4
geometry, ring formation has to be preceded by ionization and iso- enzyme pocket .
merization. Thereby, the leaving group is cleaved in a distinct step
before the α-terpinyl cation 6 is generated. As the isomerization is Conclusions
4,9
considered a slow step , the leaving group may be able to diffuse Two main conclusions can be drawn from this work. (1) The
away from the reaction centre. LOAc and NOAc, however, can presented evidence indicates that direct isomerization of the
react in a concerted fashion (S 2′/S 2-like). In these cases the transoid cation 4a into the cisoid form 4b, considered unlikely in
N
N
4
,9
leaving groups would be liberated simultaneously with the for- the biosynthesis , is possible inside cavity I. Therefore, a direct iso-
mation of the α-terpinyl cation 6, and thus they have a greater merization should also be considered in the biosynthesis. (2) It was
probability to assist in deprotonation. Evidence in this direction demonstrated for the first time that a relatively simple aromatic
was obtained by the investigation of the cyclization behaviour of cavity is catalytically competent in complex THT cyclizations. The
2-fluoro derivatives.
catalytic power of the system relies, as with enzyme pockets, on
The electron-withdrawing fluoro substituent suppresses the the stabilization of the cationic intermediates and transition states.
formation of neighbouring cationic species. Thereby, ionization- These findings set the stage for the development of catalysts able
dependent reaction pathways are efficiently slowed down, whereas to revolutionize the total synthesis of complex terpene
concerted (S 2-type) displacements remain viable. In cyclases, natural products.
N
2
-fluorogeranyl-pyrophosphate and 2-fluorolinalyl-pyrophosphate
39
were utilized to elucidate the reaction mechanism .
Received 15 September 2014; accepted 12 January 2015;
When submitting 2-fluorogeranyl acetate (F-14a) (Fig. 4e) to published online 16 February 2015
catalyst I, no formation of the corresponding cyclic products was
observed (19 days), which confirms that, indeed, ionization is References
required to initiate the reaction cascade in this case. 2-Fluoroneryl 1. Pronin, S. V. & Shenvi, R. A. Synthesis of highly strained terpenes by non-stop
tail-to-head polycyclization. Nature Chem. 4, 915–920 (2012).
Christianson, D. W. Structural biology and chemistry of the terpenoid cyclases.
Chem. Rev. 106, 3412–3442 (2006).
Miller, D. J. & Allemann, R. K. Sesquiterpene synthases: passive catalysts or
active players? Nat. Prod. Rep. 29, 60–71 (2012).
acetate (F-16a), however, showed conversion, although at a
reduced rate (8%; see Supplementary Table 7), to produce fluoro-
terpinolene F-9 and fluorolimonene F-10 as the only main products
2
.
.
3
(see Supplementary Fig. 5). The reduced rate can be explained by
the decreased uptake in the catalyst (fluoro-containing substrates 4. Croteau, R. Biosynthesis and catabolism of monoterpenoids. Chem. Rev. 87,
929–954 (1987).
have generally shown a lower degree of encapsulation) and by the
5.
Degenhardt, J., Köllner, T. G. & Gershenzon, J. Monoterpene and sesquiterpene
synthases and the origin of terpene skeletal diversity in plants. Phytochemistry
inhibition of the S 1 reaction path. The observed rate reduction,
N
both in absolute terms as well as in comparison with 2-fluorogeranyl
70, 1621–1637 (2009).
acetate (F-14a), does not indicate an ionization-dependent 6. Cane, D. E. Enzymic formation of sesquiterpenes. Chem. Rev. 90,
(
SN1-type) reaction path. In contrast to regular NOAc (16a), only
1089–1103 (1990).
7
8
.
.
Dickschat, J. S. Isoprenoids in three-dimensional space: the stereochemistry of
terpene biosynthesis. Nat. Prod. Rep. 28, 1917–1936 (2011).
Wise, M. L. & Croteau, R. in Comprehensive Natural Products Chemistry Vol. 2
(eds Barton, D., Nakanishi, K. & Meth-Cohn, O.) 97–153 (Pergamon, 1999).
traces of α-terpinene (12) were formed. These results indicate that
terpinolene (9) and limonene (10) are formed predominantly via
a concerted mechanism (S 2′ in the case of LOAc, and S 2 in the
N
N
case of NOAc), whereas α-terpinene (12) is formed predominantly 9. Allinger, N. L. & Siefert, J. H. Organic quantum chemistry. XXXIII. Electronic
spectra and rotational barriers of vinylborane, allyl cation, and related
via an S 1 mechanism within I. These results are in line with the
hypothesis that the local concentration of the leaving group influ-
ences product selectivity. Interestingly, these results are in contrast
with solvolysis studies of neryl mesylate, in which a 160-fold rate
N
compounds. J. Am. Chem. Soc. 97, 752–760 (1975).
1
0. Lesburg, C. A., Zhai, G., Cane, D. E. & Christianson, D. W. Crystal structure of
pentalenene synthase: mechanistic insights on terpenoid cyclization reactions in
biology. Science 277, 1820–1824 (1997).
4
0
reduction was observed for the 2-fluoroneryl substrate . As, in 11. Starks, C. M., Back, K., Chappell, J. & Noel, J. P. Structural basis for cyclic
terpene biosynthesis by tobacco 5-epi-aristolochene synthase. Science 277,
that case, the rate decrease was comparable to that observed
with the 2-fluorogeranyl substrate, it was argued that both react
via an ionization (S 1) mechanism. As expected, 2-fluorolinalyl
acetate (F-15a) also produced fluoroterpinolene (F-9) and
fluorolimonene (F-10) as the only main products. Additionally,
these results could explain the different degrees of side reactions
1815–1820 (1997).
1
1
1
1
2. Hof, F., Craig, S. L., Nuckolls, C. & Rebek, J. J. Molecular encapsulation. Angew.
Chem. Int. Ed. 41, 1488–1508 (2002).
N
3. Palmer, L. C. & Rebek, J. J. The ins and outs of molecular encapsulation. Org.
Biomol. Chem. 2, 3051–3059 (2004).
4. Koblenz, T. S., Wassenaar, J. & Reek, J. N. H. Reactivity within a confined self-
assembled nanospace. Chem. Soc. Rev. 37, 247–262 (2008).
5. Rebek, J. Molecular behavior in small spaces. Acc. Chem. Res. 42,
(
alkylation of phenolic groups of monomer 13 and dimerization)
observed during cyclization. Geranyl derivatives, which require
1
660–1668 (2009).
the formation of acyclic cations, displayed the highest degree. 16. Yoshizawa, M., Klosterman, J. K. & Fujita, M. Functional molecular flasks:
Linalyl and neryl derivatives, able to react in a concerted fashion
and thereby avoid acyclic cationic intermediates, displayed a much
lower tendency.
new properties and reactions within discrete, self-assembled hosts.
Angew. Chem. Int. Ed. 48, 3418–3438 (2009).
7. Wiester, M. J., Ulmann, P. A. & Mirkin, C. A. Enzyme mimics based upon
1
supramolecular coordination chemistry. Angew. Chem. Int. Ed. 50,
To compare the catalytic efficiency of the system to that of
114–137 (2011).
natural cyclase enzymes, the kinetics of the conversion of GOAc 18. Ajami, D. & Rebek, J. More chemistry in small spaces. Acc. Chem. Res.
(
14a) inside capsule I were investigated (see Supplementary
46, 990–999 (2012).
9. Raynal, M., Ballester, P., Vidal-Ferran, A. & van Leeuwen, P. W. N. M.
Supramolecular catalysis. Part 2: artificial enzyme mimics. Chem. Soc. Rev.
1
Figs 9–12 and Supplementary Table 10). The obtained k value
of 0.00079 ± 0.00006 s is only about two orders of magnitude
lower than that observed in natural cyclase enzymes (0.01–1 s ) .
However, a large difference can be seen when comparing the
cat
−1
43, 1734–1787 (2014).
−1
4
2
0. Bocokić, V. et al. Capsule-controlled selectivity of a rhodium hydroformylation
catalyst. Nature Commun. 4, 2670 (2013).
K values (0.078 ± 0.007 M inside I, as compared to low micromo- 21. Dydio, P., Detz, R. J. & Reek, J. N. H. Precise supramolecular control of
m
4
selectivity in the Rh-catalyzed hydroformylation of terminal and internal
alkenes. J. Am. Chem. Soc. 135, 10817–10828 (2013).
lar values in natural cyclase enzymes ). As K is related to
m
the inverse of the binding constant, this translates into a much
weaker binding of the substrate in the presented capsule catalyst
than in natural enzymes. One reason for the much higher affinity
2
2. Wang, Z. J., Clary, K. N., Bergman, R. G., Raymond, K. N. & Toste, F. D.
A supramolecular approach to combining enzymatic and transition metal
catalysis. Nature Chem. 5, 100–103 (2013).
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
5
©
2015 Macmillan Publishers Limited. All rights reserved

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2181
ARTICLES
2
3. Zhao, C. et al. Chiral amide directed assembly of a diastereo- and enantiopure
supramolecular host and its application to enantioselective catalysis of
neutral substrates. J. Am. Chem. Soc. 135, 18802–18805 (2013).
4. Zhao, C., Toste, F. D., Raymond, K. N. & Bergman, R. G. Nucleophilic
substitution catalyzed by a supramolecular cavity proceeds with retention of
absolute stereochemistry. J. Am. Chem. Soc. 136, 14409–14412 (2014).
5. Salles, A. G., Zarra, S., Turner, R. M. & Nitschke, J. R. A self-organizing chemical
assembly line. J. Am. Chem. Soc. 135, 19143–19146 (2013).
6. Hart-Cooper, W. M., Clary, K. N., Toste, F. D., Bergman, R. G. &
Raymond, K. N. Selective monoterpene-like cyclization reactions achieved by
water exclusion from reactive intermediates in a supramolecular catalyst. J. Am.
Chem. Soc. 134, 17873–17876 (2012).
7. MacGillivray, L. R. & Atwood, J. L. A chiral spherical molecular assembly held
together by 60 hydrogen bonds. Nature 389, 469–472 (1997).
36. Kelly, B. D., Allen, J. M., Tundel, R. E. & Lambert, T. H. Multicatalytic synthesis
of complex tetrahydrofurans involving bismuth(III) triflate catalyzed
intramolecular hydroalkoxylation of unactivated olefins. Org. Lett.
11, 1381–1383 (2009).
37. Bugarčić, Z. M., Dunkić, J. D. & Mojsilović, B. M. A simple, convenient and
expeditious approach to cineol. Heteroatom Chem. 15, 468–470 (2004).
38. Eschenmoser, A., Ruzicka, L., Jeger, O. & Arigoni, D. Zur Kenntnis der
Triterpene. 190. Mitteilung. Eine stereochemische Interpretation der
biogenetischen Isoprenregel bei den Triterpenen. Helv. Chim. Acta 38,
1890–1904 (1955).
39. Croteau, R. Evidence for the ionization steps in monoterpene cyclization
reactions using 2-fluorogeranyl and 2-fluorolinalyl pyrophosphates as substrates.
Arch. Biochem. Biophys. 251, 777–782 (1986).
40. Poulter, C. D. & King, C. H. R. Model studies of terpene biosynthesis. A stepwise
mechanism for cyclization of nerol to α-terpineol. J. Am. Chem. Soc. 104,
1422–1424 (1982).
2
2
2
2
2
8. Cavarzan, A., Scarso, A., Sgarbossa, P., Strukul, G. & Reek, J. N. H.
Supramolecular control on chemo- and regioselectivity via encapsulation of
(NHC)-Au catalyst within a hexameric self-assembled host. J. Am. Chem. Soc.
1
33, 2848–2851 (2011).
Acknowledgements
2
3
9. Bianchini, G., Sorella, G. L., Canever, N., Scarso, A. & Strukul, G. Efficient
isonitrile hydration through encapsulation within a hexameric self-assembled
capsule and selective inhibition by a photo-controllable competitive guest.
Chem. Commun. 49, 5322–5324 (2013).
0. Zhang, Q. & Tiefenbacher, K. Hexameric resorcinarene capsule is a Brønsted
acid: investigation and application to synthesis and catalysis. J. Am. Chem. Soc.
This project was supported by the Bayerischen Akademie der Wissenschaften (Junges
Kolleg), Fonds der Chemischen Industrie (Sachkostenzuschuss), the TUM Junior Fellow
Fund and the Dr.-Ing. Leonhard-Lorenz-Stiftung. The help of J. Richers with the graphical
design is gratefully acknowledged.
1
35, 16213–16219 (2013).
Author contributions
3
1. Shivanyuk, A. & Rebek, J. Reversible encapsulation by self-assembling
resorcinarene subunits. Proc. Natl Acad. Sci. USA 98, 7662–7665 (2001).
2. Avram, L. & Cohen, Y. Spontaneous formation of hexameric resorcinarene
capsule in chloroform solution as detected by diffusion NMR. J. Am. Chem. Soc.
K.T. conceived the project and wrote the manuscript with Q.Z. Q.Z. planned and carried
out the experiments. K.T. and Q.Z. discussed the experiments and results.
3
Additional information
1
24, 15148–15149 (2002).
Supplementary information and chemical compound information are available in the
online version of the paper. Reprints and permissions information is available online at
www.nature.com/reprints. Correspondence and requests for materials should be
addressed to K.T.
3
3
3
3. Yamanaka, M., Shivanyuk, A. & Rebek, J. Kinetics and thermodynamics of
hexameric capsule formation. J. Am. Chem. Soc. 126, 2939–2943 (2004).
4. Shivanyuk, A. & Rebek, J. Assembly of resorcinarene capsules in wet solvents.
J. Am. Chem. Soc. 125, 3432–3433 (2003).
5. Leão Lana, E. J., da Silva Rocha, K. A., Kozhevnikov, I. V. & Gusevskaya, E. V.
Synthesis of 1,8-cineole and 1,4-cineole by isomerization of α-terpineol
catalyzed by heteropoly acid. J. Mol. Catal. A 259, 99–102 (2006).
Competing financial interests
The authors declare no competing financial interests.
6
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
©
2015 Macmillan Publishers Limited. All rights reserved