A multiproduct terpene synthase from medicago truncatula generates cadalane sesquiterpenes via two different mechanisms

Article abstract of DOI:10.1021/jo100917c

Terpene synthases are responsible for a large diversity of terpene carbon skeletons found in nature. The multiproduct sesquiterpene synthase MtTPS5 isolated from Medicago truncatula produces 27 products from farnesyl diphosphate (1, FDP). In this paper, we report the reaction steps involved in the formation of these products using incubation experiments with deuterium-containing substrates; we determined the absolute configuration of individual products to establish the stereochemical course of the reaction cascade and the initial conformation of the cycling substrate. Additional labeling experiments conducted with deuterium oxide showed that cadalane sesquiterpenes are mainly produced via the protonation of the neutral intermediate germacrene D (5). These findings provide an alternative route to the general accepted pathway via nerolidyl diphosphate (2, NDP) en route to sesquiterpenes with a cadalane skeleton. Mutational analysis of the enzyme demonstrated that a tyrosine residue is important for the protonation process.

Full text of DOI:10.1021/jo100917c

pubs.acs.org/joc
A Multiproduct Terpene Synthase from Medicago truncatula Generates
Cadalane Sesquiterpenes via Two Different Mechanisms
†
Stefan Garms, Tobias G. Kollner, and Wilhelm Boland*
‡
,†
€
†
€
Department of Bioorganic Chemistry, Max Planck Institute for Chemical Ecology, Hans-Knoll-Strasse 8,
D-07745 Jena, Germany, and ‡Institute for Pharmacy, Martin Luther University Halle-Wittenberg,
Hoher Weg 8, D-06120 Halle, Germany
boland@ice.mpg.de
Received May 11, 2010
Terpene synthases are responsible for a large diversity of terpene carbon skeletons found in nature.
The multiproduct sesquiterpene synthase MtTPS5 isolated from Medicago truncatula produces 27
products from farnesyl diphosphate (1, FDP). In this paper, we report the reaction steps involved in
the formation of these products using incubation experiments with deuterium-containing substrates;
we determined the absolute configuration of individual products to establish the stereochemical
course of the reaction cascade and the initial conformation of the cycling substrate. Additional
labeling experiments conducted with deuterium oxide showed that cadalane sesquiterpenes are
mainly produced via the protonation of the neutral intermediate germacrene D (5). These findings
provide an alternative route to the general accepted pathway via nerolidyl diphosphate (2, NDP) en
route to sesquiterpenes with a cadalane skeleton. Mutational analysis of the enzyme demonstrated
that a tyrosine residue is important for the protonation process.
Introduction
there are multiproduct enzymes, which are known to synthe-
size up to 52 different products.5,6 The structural basis of
varying selectivity is believed to be a consequence of the
three-dimensional contour of the active site that allows or
refuses different conformations of the reactive cationic
intermediates.7,8 Nevertheless, the structure and stereo-
chemistry of the products are suggested to be mainly deter-
mined by the correct folding of the substrate and the
electrostatic interactions of the cationic intermediates with
amino acids of the active site.9 The reaction cascade is
initiated by a metal-ion-mediated cleavage of the dipho-
sphate group from the substrate, resulting in a highly
reactive carbocation. This reactive intermediate can under-
go different cyclizations and rearrangements, such as
hydride and methyl shifts. The products are released by
elimination of a proton or after reaction with water, afford-
ing sesquiterpene alcohols.10 Terpene synthases are also able
Sesquiterpenes are one of the structurally most diverse
class of natural products isolated from plants, fungi, bacteria,
and marine invertebrates.1,2 They are known to play crucial
roles when plants defend themselves against enemies, attract
pollinators, and signal within and to each other. All the tens
of thousands of sesquiterpenoids known to date are derived
from 300 basic hydrocarbon skeletons generated by sesqui-
terpene cyclases (synthases) using farnesyl diphosphate
(1, FDP) as substrate.3 The variety of products generated
by these enzymes can differ dramatically. In addition to
high-fidelity enzymes, for example, the δ-cadinene synthase
from Gossypium arboreum4, which releases a single product,
(1) Connolly, J. D.; Hill, R. A. Dictionary of Terpenes, Chapman & Hall:
New York, 1992.
(2) Cane, D. E. Chem. Rev. 1990, 90, 1089.
(3) Lesburg, C. A.; Caruthers, J. M.; Paschall, C. M.; Christianson, D. W.
Curr. Opin. Struct. Biol. 1998, 8, 695.
(4) Chen, X. Y.; Chen, Y.; Heinstein, P.; Davisson, V. J. Arch. Biochem.
Biophys. 1995, 324, 255.
(5) Steele, C. L.; Crock, J.; Bohlmann, J.; Croteau, R. J. Biol. Chem. 1998,
273, 2078.
(7) Vedula, L. S.; Rynkiewicz, M. J.; Pyun, H. J.; Coates, R. M.; Cane,
D. E.; Christianson, D. W. Biochemistry 2005, 44, 6153.
(8) Aaron, J. A.; Lin, X.; Cane, D. E.; Christianson, D. W. Biochemistry
2010, 49, 1787.
€
(6) Degenhardt, J.; Kollner, T. G.; Gershenzon, J. Phytochemistry 2009,
70, 1621.
(9) Christianson, D. W. Chem. Rev. 2006, 106, 3412.
(10) Cane, D. E. Compr. Nat. Prod. Chem. 1999, 2, 155.
5590 J. Org. Chem. 2010, 75, 5590–5600
Published on Web 07/20/2010
DOI: 10.1021/jo100917c
r
2010 American Chemical Society

Garms et al.
JOCArticle
SCHEME 1. Proposed Pathways to Cadalane-Type Sesquiterpenes
to produce enzyme-bound neutral intermediates that are
further cyclized after reprotonation. Germacrene A, for
example, formed by a sequence consisting of the ionization
of FDP (1), C1-C10 cyclyzation, and deprotonation at C12, is
an intermediate in the reaction catalyzed by the 5-epi-aristo-
lochene synthase (TEAS) from tobacco.11 Reprotonation of
the neutral intermediate initiates the reaction sequence starting
with a C1-C7 cyclization, followed by a Wagner-Meerwein
rearrangement of a methyl group and the loss of a proton that
affords the main product 5-epi-aristolochene. Analysis of the
crystal structure of TEAS cocrystallized with the artificial
substrate farnesyl hydroxyphosphonate suggests that a cata-
lytic triad consisting of a tyrosine residue (Tyr520) and two
aspartate residues (Asp444 and Asp525) are responsible for the
reprotonation of germacrene A.12 The important role of
Tyr520 in the proton-transfer reaction has been demonstrated
by the exchange of the tyrosine residue with phenylalanine; this
results in a protein that released germacrene A as its sole
product.13 Other germacrene-type sesquiterpenes have also
been proposed as biogenic intermediates for certain groups
of sesquiterpenes.14,15 Arigoni suggested a pathway for the
biosynthesis of cadalane skeletons involving germacrene D (5)
as a neutral intermediate (pathway a, Scheme 1); such a
pathway would overcome the geometric restriction hindering
the direct cyclization of FDP (1) to 10-membered ring systems
with cis-configured C2-C3-double bonds.16 The key step in the
reaction sequence is the conformational change of the neutral
intermediate germacrene D (5) from the cisoid to the transoid
structure, affording a cis-configured C2-C3-double bond after
reprotonation (cation 4, Scheme 1). In fact, the existence of
this pathway has never been experimentally demonstrated.
However, the enzymatic cyclization of other cadalane-type
sesquiterpenes such as cubenene,17 1-epi-cubenol,17,18 and
δ-cadinene19 has been shown to proceed via a nerolidyl dipho-
sphate (2, NDP) intermediate (pathway b, Scheme 1), which
allows rotation about the 2,3-single bond and ring closure to
give the helmintogermacrenyl cation (3). In these cases, the
cationic intermediate 4originated from 3byahydrideshift17-19
or by protonation of the cyclopropane moiety of previously
formed isobicyclogermacrene (6)16 intermediate (pathway c,
Scheme 1).
We recently reported a multiproduct terpene synthase from
Medicago truncatula that catalyzes the formation of different
hydrocarbon skeletons such as cadinane, copaane, cubebane,
and germacrane.20 Using an approach that combines classical
labeling experiments, the determination of the absolute config-
uration and the enantiomeric excess of individual products by
chiral gas chromatography and site-directed mutagenesis, we
established the reaction cascade and the mode by which the
products are cyclized. In addition, we report an alternative
pathway for the formation of cadinane, cubebane, muurolane,
and copaane skeletons, which proceeds via the protonation of
(-)-germacrene D (5), proving a prior hypothesis of Arigoni.16
Knowledge of the exact configuration of products and their
optical purity allows us to elucidate the extent of enzymatic
control over certain reaction steps and the initial folding of the
cycling substrate.
Results
(11) Cane, D. E.; Prabhakaran, P. C.; Oliver, J. S.; McIlwaine, D. B. J.
Am. Chem. Soc. 1990, 112, 3209.
(12) Starks, C. M.; Back, K. W.; Chappell, J.; Noel, J. P. Science 1997,
277, 1815.
MtTPS5 Is a Multiproduct Enzyme Generating 27 Differ-
ent Products. Recombinant MtTPS5 incubated with FDP (1)
(13) Rising, K. A.; Starks, C. M.; Noel, J. P.; Chappell, J. J. Am. Chem.
Soc. 2000, 122, 1861.
(14) Nishimura, K.; Shinoda, N.; Hirose, Y. Tetrahedron Lett. 1969, 10,
3097.
(15) Yoshihara, K.; Ohta, Y.; Sakai, T.; Hirose, Y. Tetrahedron Lett.
1969, 10, 2263.
(16) Arigoni, D. Pure Appl. Chem. 1975, 41, 219.
(17) Nabeta, K.; Kigure, K.; Fujita, M.; Nagoya, T.; Ishikawa, T.;
Okuyama, H.; Takasawa, T. J. Chem. Soc., Perkin Trans. 1 1995, 1935.
(18) Cane, D. E.; Tandon, M. J. Am. Chem. Soc. 1995, 117, 5602.
(19) Benedict, C. R.; Lu, J. L.; Pettigrew, D. W.; Liu, J. G.; Stipanovic,
R. D.; Williams, H. J. Plant Physiol. 2001, 125, 1754.
(20) Arimura, G. I.; Garms, S.; Maffei, M.; Bossi, S.; Schulze, B.; Leitner,
M.; Mithoefer, A.; Boland, W. Planta 2008, 227, 453.
J. Org. Chem. Vol. 75, No. 16, 2010 5591

Garms et al.
JOCArticle
FIGURE 1. GC-FID chromatograms of enzymatic products of recombinant MtTPS5 wild type (A) and Y526F mutant (B) from incubation
ꢀ
with (2E,6E)-FDP (1). The products were identified by Kovats indices and mass spectra in comparison to authentic samples: 7, R-cubebene; 8,
R-copaene; 9, β-cubebene; 10, (E)-β-caryophyllene; 11, β-copaene; 12, cadina-3,5-diene; 13, R-humulene; 14, allo-aromadendrene; 15, trans-
cadina-1(6),4-diene; 16, γ-muurolene; 5, germacrene D; 17, bicyclosesquiphellandrene; 18, bicyclogermacrene; 19, R-muurolene; 20, δ-
amorphene; 21, cubebol; 22, δ-cadinene; 23, cadina-1,4-diene; 24, nerolidol;a 25, copan-3-ol; 26, 4R-hydroxygermacra-1(10),5-diene; 27,
copaborneol;b 28, 1-epi-cubenol; 29, T-cadinol; 30, cubenol;b 31, torreyol; 32, kunzeaol;b 33, C15H26O; 34, bicycloelemene.c aNonenzymatic
c
hydrolysis product of the substrate (1). bNo reference compound available. Thermal rearrangement product of bicyclogermacrene (18), X
BHT-like antioxidant; peaks not indicated were contaminants from the GC column or the solvent used.
has been reported to result in 15 sesquiterpenoid products.20
A thorough reevaluation of the GC-MS data revealed that
at least 27 sesquiterpenes are formed in the presence of Mg2þ
ions, 25 of which could unambiguously be identified by their
alcohol (33, C15H26O > 0.1%) were generated. The acyclic
nerolidol (24) resulted from nonenzymatic hydrolysis of the
substrate as demonstrated by control experiments with the heat-
inactivated enzyme (data not shown).
ꢀ
Kovats indices and mass fragmentation in comparison with
Stereochemical Analysis of the Enzyme Products. To ex-
amine the stereochemical course of product formation, the
absolute configuration of a subset of enzyme products was
determined by GC-MS using chiral stationary phases. The
stereochemistry of the two major alcohols, cubebol (21) and
copan-3-ol (25), was inferred from the acid rearrangement
products δ-cadinene (22) and R-muurolene (19) (Figure SI-1,
Supporting Information), respectively. No information
about the absolute configuration at the carbon C3 of 25
from literature was available.21 However, the enantiomeric
analysis of 17 products of the MtTPS5 (Table 1, Figure 2)
revealed that germacrene D (5), allo-aromadendrene (14),
bicyclogermacrene (18), and all analyzed cadalanes exhibited
the (S)-configuration at C10. Additionally, the relative con-
figuration of all other cadalane stereocenters turned out to be
the same. These results enabled us to deduce the absolute
configurations of all other products. The evaluation of
the enantiomeric composition revealed that all substances were
optically pure (g98% ee). The enantiomeric excess of (-)-16
could not be determined since the (þ)-enantiomer elutes with
the same retention time as (-)-14 generated by the enzyme. In
addition, the identical enantiomers of products are formed by
the Y526F mutant discussed below (Table 1, Figure 2).
The identical configuration in compounds sharing the
same skeleton is in agreement with a common cationic
precursor and a largely predetermined conformation of the
acyclic substrate in the active site. Only a limited number of
conformations of (2E,6E)-FDP (1) allow the formation of
10- or 11-membered ring systems.22 In fact, four (Scheme 2)
authentic references (Figure 1, Table SI-1 (Supporting Informa-
tion)). In addition to the sesquiterpene hydrocarbons (31.7%),
the recombinant protein generates predominantly sesquiterpene
alcohols (68.3%) with cubebol (21) (38.6%) as the major
product. The second most abundant product, 25 (14.9%),
showed a mass spectrum very similar to that of 21 with a
molecular ion at m/z 222 and a base peak at m/z 161, typical
for the loss of an isopropyl radical. The attempted catalytic
hydrogenation (5% rhodium on alumina, 1 h) of 25 showed no
hydrogen uptake, proving a tricyclic backbone for 25. More-
over, dehydratization of 20 with trifluoroacetic anhydride
yielded predominantly R-copaene (8). A compound matching
all experimental observations is copan-3-ol, previously reported
from Ocimum americanum.21 Like the two major products, a
number of substances are formed containing a 3,7-dimethyl-10-
isopropyloctaline skeleton (so-called cadalane skeleton), such as
torreyol (31) (9.3%), γ-muurolene (16) (4.9%), R-muurolene
(19) (3.1%), δ-cadinene (22) (2.5%), R-copaene (8) (2.1%),
copaborneol (27) (1.4%), β-cubebene (9) (1.3%), bicyclosesqui-
phellandrene (17) (1.1%), R-cubebene (7) (0.8%), β-copaene
(11) (0.7%), cadina-1,4-diene (23) (0.5%), and six other com-
pounds (each of which >0.5%). A second group comprises the
germacranes, namely germacrene D (5) (5.8%), 4R-hydoxyger-
macra-1(10),5-diene (26) (2.8%), kunzeaol (32) (0.7%), and
bicyclogermacrene (18) (0.3%). Additionally, R-humulene (13)
(2.0%), (E)-β-caryophyllene (10) (2.6%), and allo-aromaden-
drene (14) (3.5%) along with an unidentified sesquiterpene
(21) Upadhyay, R. K.; Misra, L. N.; Singh, G. Phytochemistry 1991, 30,
691.
(22) Cane, D. E. Acc. Chem. Res. 1985, 18, 220.
5592 J. Org. Chem. Vol. 75, No. 16, 2010

Garms et al.
JOCArticle
TABLE 1. Absolute Configuration of Enzyme Products
bridgehead atoms of (E)-β-caryophyllene (10) (2S,10R), allo-
aromadendrene (14) (2R,6S), and the cadalanes (cis-fused
decalin system).
enzyme product
(-)-R-cubebene (7)
(-)-R-copaene (8)a
Hydride Shifts and 1,3-Deprotonation Steps in the Forma-
tion of MtTPS5 Products. Incubation of the MtTPS5 with
selectively labeled FDPs, such as (1S)-[1-2H]-FDP, (1R)-
[1-2H]-FDP, and [1,1-2H2]-FDP, generated products with
deuterium atoms in positions that allowed the reconstruction
of mechanistic details of the reaction cascade. Most impor-
tant are the molecular ion ([M]þ, number of deuterium
atoms) and the cation (fragment a) resulting from the loss
of the isopropyl side chain (Table 2, Figure 3, Table SI-2
(Supporting Information)).17,23,24
(-)-β-cubebene (9)
(-)-(E)-β-caryophyllene (10)
(-)-β-copaene (11)
(-)-cadina-3,5-diene (12)
(-)-allo-aromadendrene (14)
(-)-γ-muurolene (16)
(-)-germacrene D (5)
(-)-bicyclosesquiphellandrene (17)b
(-)-bicyclogermacrene (18)
(-)-R-muurolene (19)
(-)-cubebol (21)a
Incubating the enzyme with [1,1-2H2]-FDP resulted in a
molecular ion (m/z 206 and m/z 224, respectively) for both
germacrene D (5) and cubebol (21), confirming that both
deuterium atoms remained in the molecules. Additional
fragments at m/z 162 indicated the loss of a labeled isopropyl
group, supporting that a single deuterium atom migrated
from position C1 to C11. The stereochemical course of the
1,3-hydride shift was elucidated by the use of (1R)- and
(1S)-[1-2H]-FDP as substrates. Germacrene D (5) and cube-
bol (21) showed ions at m/z 205 and 161 ([M - C3H6D]þ)
after incubation with (1R)-[1-2H]-FDP, while incubation
with (1S)-[1-2H]-FDP the m/z of the corresponding ions were
shifted to m/z 205 and 162, confirming the loss of an unlabeled
isopropyl group. The results are consistent with a mechanism
by which the C1-HRe of the substrate is transferred via a
suprafacial 1,3-hydride shift to the tertiary carbon atom of
the isopropyl group (Scheme 3). The same result was observed
for all other compounds with a germacrane or cadalane
skeleton (Table SI-2, Supporting Information). The mass
spectra of bicyclogermacrene (18) and allo-aromadendrene
(14) obtained by the same series of incubations are con-
sistent with a stereospecific 1,3-deprotonation of the (2E,6E)-
germacren-11-yl cation (35) (Scheme 3) to give 18. With
(1R)-[1-2H]-FDP as substrate, the C1-2HRe is lost (m/z 204),
while in the case of (1S)-[1-2H]-FDP as substrate the deuterium
atom remained in the product (m/z 205).
(þ)-δ-cadinene (22)
(-)-cadina-1,4-diene (23)
(þ)-copan-3-ol (25)a
(-)-torreyol (31)a
aNot determined for Y526F mutant. bNot detected in Y526F.
The mass spectra of trans-cadina-1(6),4-diene (15) and
cadina-3,5-diene (12) also provided information on later
hydride shifts in the reaction cascade. For example, the
cadinanyl cation (37) (Schemes 4 and Scheme 6) suffers
two consecutive 1,2-hydrogen shifts to the cations 39 and
38, which serve as precursors for 12, 15, and 17. An alter-
native 1,3-hydrogen shift from cation 37 to the cation 38
could be excluded since incubation of (1S)-[1-2H]-FDP gave
15 with complete loss of deuterium. In case of a direct 1,3-
shift, the deuterium atom of 37 would migrate to C7 and
remain in the molecule (Scheme 4).
Formation of Cadalanes and allo-Aromadendrene Proceed
via Protonation of Neutral Intermediates. To confirm the
existence of protonation steps in the reaction cascade, we
analyzed the enzyme activity in the presence of D2O. Mass
spectrometric analysis of the compounds revealed different
incorporation rates of deuterium in the enzyme products
(Table 3 and Table SI-3 (Supporting Information)). The germa-
cranes (5, 26, and 32), R-humulene (13), (E)-β-caryophyllene
FIGURE 2. Identification of the absolute configuration of R-cube-
bene (7) (A) and germacrene D (5) (B) produced by the wild type
MtTPS5 and the mutant Y526F enzyme.
of the eight possible diastereomeric conformations of 1 result
in a (R)-configuration at C10 after ring closure between C1
and C10 of the substrate. Following the concept of the least
motion during the course of cyclization, conformation 1a
(Scheme 2) can be inferred as the starting conformation for
the majority of the enzyme’s products. This conclusion is
supported by relative and absolute configurations of the
(23) Schmidt, C. O.; Bouwmeester, H. J.; Franke, S.; Konig, W. A.
Chirality 1999, 11, 353.
(24) Benedict, C. R.; Alchanati, I.; Harvey, P. J.; Liu, J. G.; Stipanovic,
R. D.; Bell, A. A. Phytochemistry 1995, 39, 327.
J. Org. Chem. Vol. 75, No. 16, 2010 5593

Garms et al.
JOCArticle
TABLE 2. Mass Spectrometric Parent and Fragment Ions from Some Nonlabeled and Deuterated Products of Recombinant MtTPS5
ionsa (m/z)
nonlabeled FDP
(1S)-[1-2H]-FDP
(1R)-[1-2H]-FDP
[1,1-2H2]-FDP
Mþ
a
Mþ
a
Mþ
a
Mþ
a
germacrene D (5)
cubebol (21)
trans-cadina-1(6),4-diene (15)
allo-aromadendrene (14)
bicyclogermacrene (18)
204
204b
204
204
204
161
161
161
161
161
205
205b
204
205
205
162
162
161
162
162
205
205b
205
204
204
161
161
161
161
161
206
206b
205
205
205
162
162
161
162
162
aFormation of ions shown in Figure 3. b[M - H2O]þ.
SCHEME 2. Possible Starting Conformations of FDP (1) Leading
to (R)-Configuration at C10 after C1 to C10 Cyclization
525 which delivers a proton to the neutral intermediate
germacrene A.12 The crucial role of the phenolic hydroxyl
group of Tyr520 in the proton transfer reaction was con-
firmed by site-directed mutagenesis.13 A comparison of the
amino acid sequence of the MtTPS5 and TEAS revealed
in MtTPS5 a tyrosine in an equivalent position. The ex-
change of this tyrosine 526 by phenylalanine (Y526F
mutant) resulted in a recombinant protein showing only
10% of the activity of the wild type (WT) and a dramatic
change in the product profile (Figure 1). The Y526F mutant
generated almost exclusively germacranes, namely germa-
crene D (5), 4R-hydroxygermacra-1(10),5-diene (26), and
bicyclogermacrene (18), accounting for 75% of product
profile (Figure 1). In contrast, the cadalane-type sesquiter-
penes are reduced from 82% (WT) to 15% (Y526F mutant)
supporting a crucial role of the OH-group from Tyr526 for
the protonation of the germacrene D precursor. To under-
stand the origin of the remaining cadalanes, we performed
the same incubation experiments with the Y526F mutant in
D2O as described above. In these experiments the cadalanes
displayed a low degree of deuteration (0% to 30% d1) which
suggests that an additional and independent reaction chan-
nel for their formation should exist. The most plausible
alternative appears to be the intermediacy of NDP (2) and
the nerolidyl cation which delivers the correct orientation for
the C1 to C10 ring closure to the cation 3 (Scheme 6). This
bypasses the protonation by Tyr526 en route to the cation 4a
(Scheme 6) and explains the formation of cadalanes with low
deuterium labeling.
(10), and bicyclogermacrene (18) did not incorporate deuter-
ium, which is in line with mechanistic pathways reported in
the literature.2,25 In contrast, the mass of the molecular ion of
allo-aromadendrene (14) was increased by 1 amu to m/z 205,
indicating that a single deuterium atom was most likely
transferred to bicyclogermacrene (18) at C3 (Scheme 3).
These results validate the suggestion that the formation of
14 is achieved by the loss of the hydrogen at C1 with
subsequent closure to the three-membered ring system, as
was originally proposed for the formation of kelsoene and
prespatane proceeding via an allo-aromadendranyl cation.26
However, because of hydrogen scrambling during the frag-
mentation process (cf. Schmidt et al.27), the exact deuterium
position in 14 cannot be determined. Interestingly, all cada-
lanes except δ-amorphene (20) and T-cadinol (29) also in-
corporated deuterium, with labeling degrees up to 92% d1
(Table 3 and Table SI-3 (Supporting Information)). These
results confirm a previous hypothesis of Arigoni, that the
formation of cadalane-type sesquiterpenes could be initiated
by a proton transfer to the exocyclic double bond of transoid
germacrene D (5) (Scheme 1, pathway a).16 Indirect evidence
for the site of protonation is provided by the lower degree of
deuterium labeling of β-copaene (11) (60% d1), β-cubebene
(9) (74% d1), and bicyclosesquiphellandrene (17) (73% d1)
that is consistent with the partial loss of deuterium during the
final deprotonation step to the exocyclic double bond of 9, 11,
and 17 (Scheme 5).
Discussion
Proposed Mechanism of the MtTPS5 Enzyme Product
Formation. A general scheme integrating the results of different
incubation experiments with H2O and D2O obtained from
unlabeled FDP and chirally labeled substrates is given in
Scheme 6. The mechanistic concept also takes into account
the identical absolute configuration of products sharing the
same skeleton and the existence of parallel and independent
pathways initiated by protonation of neutral intermediates.
The cyclization cascade starts by ionization of FDP prefolded
according to conformation 1a (Scheme 2) yielding the highly
reactive transoid farnesyl cation (42). The C1 to C11 ring
closure of 42 affords the humulyl cation (43), which can either
lose a proton from C9 to form the achiral R-humulene (13) or
undergo a second C2 to C10 closure by electrophilic attack on
the re-face of the C2-C3 double bond. Final loss of a proton
generates (-)-(E)-β-caryophyllene (10). The majority of the
products (95%) requires the initial C1 to C10 closure and is
achieved by the electrophilic attack on the si-face of the distal
Identification of Tyrosine 526 as a Key Amino Acid in
Proton-Transfer Reactions. The crystal structure of the
5-epi-aristolochene synthase from tobacco (TEAS) contains
a catalytic triad of aspartate 444, tyrosine 520, and aspartate
(25) Matsuo, A.; Nozaki, H.; Kubota, N.; Uto, S.; Nakayama, M. J.
Chem. Soc., Perkin Trans. 1 1984, 203.
(26) Nabeta, K.; Yamamoto, M.; Fukushima, K.; Katoh, K. J. Chem.
Soc., Perkin Trans. 1 2000, 2703.
(27) Schmidt, C. O.; Bouwmeester, H. J.; Bulow, N.; Konig, W. A. Arch.
Biochem. Biophys. 1999, 364, 167.
5594 J. Org. Chem. Vol. 75, No. 16, 2010

Garms et al.
JOCArticle
FIGURE 3. Formation of fragment ions of some products of MtTPS5.
SCHEME 3. Stereochemical Course of the 1,3-Hydride Shift and the 1,3-Deprotonation Leading to the (2E,6E)-Germacren-1-yl Cation
(36), 14, and 18, Respectively
SCHEME 4. Possible Labeling Pattern of trans-Cadina-1(6),4-
diene (15) Obtained by Incubation with (1S)-[1-2H]-FDP along
Different Types of Hydride Shifts
TABLE 3. Deuterium Content of a Subset of Enzyme Products from
Incubating Recombinant MtTPS5 and Y526F Mutant in Highly Enriched
D2O Medium (Mean Values and Standard Deviation of Four Independent
Experiments)
deuterium content d1 (%)
wild type
mutant Y526F
R-cubebene (7)
92 ( 1
74 ( 1
73 ( 5
92 ( 1
60 ( 3
91 ( 1
-4 ( 0
98 ( 0
11 ( 10
29 ( 4
24 ( 3
n.d.
30 ( 6
11 ( 3
2 ( 5
-1 ( 1
64 ( 4
0 ( 1
β-cubebene (9)
bicyclosesquiphellandrene (17)
R-copaene (8)
β-copaene (11)
R-muurolene (19)
δ-amorphene (20)a
allo-aromadendrene (14)
bicyclogermacrene (18)a
aNegative values or large standard deviations were the consequence
of the low concentration of products (<0.5%), resulting in an increased
influence of the background on the calculated values.
double bond of 42, which establishes the (R)-configuration at
C10 of the (2E,6E)-germacren-11-yl cation (35). This con-
figuration brings the C1-HRe in a favorable position to the
unoccupied p-orbital of the isopropyl side chain and initiates
the 1,3-hydride shift to the (2E,6E)-germacren-1-yl cation (36).
The stereoelectronics of this hydride migration are in line with a
1,3-hydride shift previously reported in the biosynthesis of (-)-
germacrene D in Solidago canadensis.23 However, a different
geometry of the (2E,6E)-germacren-1-yl cation (36) favors the
transfer of the C1-HSi rather than the C1-HRe in the formation
of (-)-germacrene D mediated by Streptomyces coelicolor
A3(2).28 Alternatively, the cation 35 may suffer a 1,3-depoto-
noation (loss of the C1-HRe) to(þ)-bicyclogermacrene (18) asa
minor product. The majority of 18 remains enzyme-bound and
is further cyclized to (-)-allo-aromadendrene (14) by reproto-
nation at C3 on the si-face followed by concomitant C6 to C2
closure and the elimination of a proton at C14. These results
give the first evidence of the important role of 18 as biogenetic
precursor for sesquiterpenes with a dimethylcyclopropane
ring.14,25,29 Approximately 91% of the overall carbon flux is
shuttled through the 10-membered monocyclic cation 36 that is
quenched either by deprotonation to (-)-germacrene D (5) or
by reaction with a water molecule to the germacrenyl alcohols
4R-hydroxygermacra-1(10),5-diene (26) and kunzeol (32). As
evident from labeling experiments conducted in D2O, a large
fraction of (-)-5 remains enzyme-bound and after a conforma-
tional change from the cisoid to the transoid structure its
reprotonation at C15 affords the (2Z,6E)-germacren-1-yl ca-
tion 4a along with the cadinan-7-yl cation 37a. These results
support a hypothesis of Arigoni16 suggested for the biosynthe-
sis of cadalane skeletons and provide an alternative to the
generally accepted and well-studied pathway proceeding via
(28) He, X. F.; Cane, D. E. J. Am. Chem. Soc. 2004, 126, 2678.
(29) Adio, A. M. Tetrahedron 2009, 65, 1533.
J. Org. Chem. Vol. 75, No. 16, 2010 5595

Garms et al.
JOCArticle
SCHEME 5. Proposed Labeling Pattern of Products with an Exocyclic Double Bond Obtained by Incubation of Recombinant MtTPS5
in Highly Enriched D2O Medium
the intermediacy of NDP (2).17-19,30 The cation 37a represents
the key branching point for all cadalanes with a cis-decalin ring
system (ca. 80% of the product profile) and stabilizes itself by
an unspecific loss of a vicinal proton to δ-cadinene (22),
γ-muurolene (16), andR-muurolene (19) orbycaptureofawater
molecule affording torreyol (31). After ring closure between C2
and C7 of 37a, the new cation eliminates a proton or adds a
water molecule to afford the copaane-type sesquiterpenes R-,
β-copaene (8, 11) and copan-3-ol (25). This occurs in concurrence
withaC3toC7ringclosureandadditionofwatertocopaborneol
(27). Another reaction channel via a syn 1,2-hydride shift trans-
forms 37a into the cation 39, which is either deprotonated to (-)-
cadina-1,4-diene (23) or reacts with water to the epimeric cubenols
28 and 30. The same hydride shift has been previously docu-
mented for (þ)-epi-cubenol produced by Heteroscyphus planus
and Streptomyces sp LL-B7.31 A second syn 1,2-shift of cation 39
produces cation 38 that leads to either (-)-bicyclosesquiphell-
andrene (17), (-)-cadina-3,5-diene (12), or trans-cadina-1(6),4-
diene (15); the latter one could also be formed by direct
deprotonation from 39. The principal product (-)-cubebol
(21), as well as (-)-R-cubebene (7) and (-)-β-cubebene (9), are
formed from cation 39 after C2 to C6 ring closure followed by
deprotonation or the capture of a water molecule. Cadalane
structures originating from cadinanyl cations with different
relative configurations of the angular hydrogen atoms and the
isopropyl group (cations 37b-d), such as T-cadinol (29) and
δ-amorphene (20), seem not to be biosynthesized via protonation
of the transoid (-)-5 as suggested by the labeling experiments
with D2O. In these cases, the cationic key intermediate 4b is
generated by a sequence starting from farnesyl diphosphate (1)
with subsequent isomerization to the tertiary allylic intermedi-
ate, nerolidyl diphosphate (2). After rotation around the
C2-C3 single bond and the cleavage of the diphosphate
group, the stereospecific C10 to C1 closure establishes the
(R)-configuration at C10 of the cation 3. Finally, a stereospe-
cific 1,3-hydride shift of the C1-HRe affords cation 4b. The
different conformations of 4b determine the configurations of
the angular hydrogens, which are realized after the electrophilic
attack on the C6-C7 double bond. Both cations 37b and 37c,
which exhibit a trans-configuration of the C1-H and the
isopropyl group, are potential precursors of 20, which is
released after deprotonation. The formation of 29 involves
the electrophilic attack on the C6-C7 double bond of 4b,
establishing the trans-configuration (1R,6S) (cation 37d) of the
bridgehead hydrogen atoms and quenching of the positive
charge by a water molecule. Labeling experiments with the
Y526F mutant conducted in D2O revealed that all other
cadalane products, such as 19 and 31, could also arise from
the nerolidyl diphosphate pathway, as indicated by the absence
of deuterium in the products. This may also explain why the
labeling degrees of the cadalanes obtained by D2O experiments
with the wild type do not exceed 91% d1 (although 98% d1 is
possible, as obtained for 14). These results suggest that about
92% of the key intermediate cation 4a is derived by the
protonation of (-)-5 (highlighted blue in Scheme 6) and about
8% arises from the nerolidyl diphosphate pathway (highlighted
orange in Scheme 6). The MtTPS5 enzyme seems to use both
pathways to produce the same compounds, reflecting the high
flexibility of multiproduct enzymes. In addition, 5-epi-aristolo-
chene synthase from tobacco is able to cyclize both FDP (1) and
its tertiary allylic isomer NDP (2), but different products were
generated by both pathways.32,33
Enzymatic Control Exerted on Reaction Channels. The
sesquiterpene synthase gene MtTPS5, isolated from Medicago
truncatula, encodes an enzyme that generates from FDP (1) at
least 27 different sesquiterpenes. The stereochemical analysis
of 17 enzyme products reveals the conservation of configurations
(32) O’Maille, P. E.; Chappell, J.; Noel, J. P. Arch. Biochem. Biophys.
2006, 448, 73.
(33) Faraldos, J. A.; O’Maille, P. E.; Dellas, N.; Noel, J. P.; Coates, R. M.
J. Am. Chem. Soc. 2010, 132, 4281.
(30) Nabeta, K.; Fujita, M.; Komuro, K.; Katayama, K.; Takasawa, T.
J. Chem. Soc., Perkin Trans. 1 1997, 2065.
(31) Cane, D. E.; Tandon, M. Tetrahedron Lett. 1994, 35, 5351.
5596 J. Org. Chem. Vol. 75, No. 16, 2010

Garms et al.
JOCArticle
SCHEME 6. Proposed Reaction Mechanism for the Formation of Sesquiterpene Products by MtTPS5a
aStereocenters with the same absolute configuration are marked in green, red, and blue. Reaction channels leading to 4a,b by protonation of germacrene
D (5) (light blue) or by the intermediacy of NDP (2) (orange) are highlighted. Absolute configuration determined (gray highlighted boxes).
among compounds with the same hydrocarbon skeleton as
well as between related structures. This allows the deduction
of common cationic precursors and the initial conformation
of the cyclizing substrate in the active site. In addition,
compounds belonging to a common structural group are
almost optically pure (g98% ee). For example, the initial
cyclization of FDP (1) to the 10-membered ring system of
(-)-germacrene D (5) and the follow-up product R-cubebene
(7) (Scheme 6) leads to enantiomerically pure products
including the alcohols 26, 32, and 21, originating from
cationic intermediates en route to cadalane skeletons
(Figure 2). In contrast, the multiproduct terpene synthase
TPS4 from Zea mays and the trichodiene synthase from
Fusarium sporotrichioides generate racemic and diastereomeric
products that differ in configuration at C6 or C7, which
is explained by invoking the intermediacy of both (R) and
(S)-bisabolyl cation enantiomers.34,35 In TPS4 this has been
attributed to two different pockets at the active site of the
enzyme that are populated by different conformations of the
farnesyl cation, each of which produces a single enantiomer of
the initially formed bisabolyl cation.36 MtTPS5 apparently exerts
a stronger control on the initial conformation of the farnesyl
cation (42) allowing only the formation of the (R)-enantiomer
of the (2E,6E)-germacren-11-yl cation (35) and its follow-up
products (germacranes, cadalanes). Most importantly, the
configuration of the product spectrum discussed so far is
basically determined by the starting conformation of FDP (1)
and the subsequently formed farnesyl cation (42). Nevertheless,
the correct folding of the FDP (1) does not alone dictate the
structure of the cyclization products. The enzyme-active site
(35) Vedula, L. S.; Jiang, J.; Zakharian, T.; Cane, D. E.; Christianson,
D. W. Arch. Biochem. Biophys. 2008, 469, 184.
(34) Kollner, T. G.; Schnee, C.; Gershenzon, J.; Degenhardt, J. Plant Cell
2004, 16, 1115.
(36) Kollner, T. G.; O’Maille, P. E.; Gatto, N.; Boland, W.; Gershenzon,
J.; Degenhardt, J. Arch. Biochem. Biophys. 2006, 448, 83.
J. Org. Chem. Vol. 75, No. 16, 2010 5597

Garms et al.
JOCArticle
seems to be spacious or flexible enough to enable conformational
changes and rearrangements of the initial cation 35 to give 37a
after protonation of the neutral intermediate (-)-5. These results
are in line with theoretical studies on cyclization mechanisms of
several sesquiterpenes, which suggest that certain productive
conformations of the bisabolyl cation could not be directly
derived from FDP (1). Only after conformational changes from
accessible to initially inaccessible bisabolyl cation conformers
certain sesquiterpenes are formed.37 Once the central cation 37a
is formed it stabilizes by a multitude of independent reaction
channels (hydride shifts, cyclizations, deprotonations, or addi-
tion of water) to hydrocarbons or alcohols as the main products.
Apparently, the enzyme exerts less control on the pathways
diverting from cation 37a, which is consistent with a lack of
properly positioned functional groups that could direct the
final cyclizations and deprotonations to defined cadalane end
products.
deuterium content of (-)-allo-aromadendrene (14) released
from the mutant enzyme (98% d1 WT to 64% d1 Y526F
mutant). However, in MtTPS5 the influence of the tyrosine is
in fact smaller compared to that observed for cadalane type
sesquiterpenes, leading to the suggestion that other amino
acids are more important for this proton transfer. Notably,
the production of (-)-18 is strongly enhanced in the mutant
enzyme, accounting for 12.9% of the overall product out-
come (0.3% WT). These results demonstrate that the tyro-
sine residue might influence the equilibrium of the competing
pathways branching from (2E,6E)-germacren-11-yl cation
(35, Scheme 6). The proportion of the 1,3-deprotonation
affording (-)-18 competing with the 1,3-hydride shift is
slightly enhanced in the mutant enzyme. A possible explana-
tion is a disturbed hydrogen bonding network, activating the
active site base which facilitate the removal of the C1-HRe
.
This remains more speculative, since no information about
the structural bases favoring hydride shifts are available.
Role of Tyrosine 526 in the Catalysis Mediated by MtTPS5.
Introduction of the Y526F mutation in the active site of
MtTPS5 results in almost exclusive formation of germa-
cranes (80% of the product profile) accompanied by strong
reduction of cadalane type sesquiterpenes (82% WT f 15%
Y526F mutant). The enhanced accumulation of the precur-
sor related products verifies the intermediacy of germacrene
D (5) in MtTPS5 catalysis and strongly supports the crucial
role of the hydroxyl group of Y526 in the proton transfer to
the exocyclic double bond of (-)-5. Further evidence is given
by the low incorporation rates of deuterium in cadalane
products obtained by incubation experiments with D2O
(0% to 30% d1). Nevertheless, the exchange of tyrosine by
a phenylalanine does not fully disable the reprotonation of
(-)-5 as indicated by the noticeable incorporation of deuter-
ium in cubebanes and copaanes (11-30% d1), which in turn
suggests that Tyr526 is not the direct proton donor. It is more
likely that a water molecule trapped in the active site might
serve as the requisite Lewis acid. A similar function for a
solvent molecule has been proposed for the aristolochene
synthase from Penicillium roqueforti.38 Evidence for the
presence of water in the active site of MtTPS5 comes from
the high amounts of sesquiterpene alcohols formed by the
wild type (68%) and the Y526F mutant (53%). The Lewis
acid properties of water might be adjusted by hydrogen
bonds to the Tyr526 and other polar amino acid side chains.
The study of the crystal structure of the aristolochene
synthase from Aspergillus terreus revealed a diphosphate
moiety derived from the substrate FDP (1) as another
possible acid driving the reprotonation of (-)-5.39 The
reported strong cofactor-dependence of the macrocarpene
biosynthesis mediated by TPS6 and TPS11 from Z. mays,
which includes the protonation of β-bisabolene, provides
further evidence for this hypothesis.40 Neither this hypoth-
esis nor that which claims another amino acid is the actual
proton donor could be excluded without more detailed
structural information and additional experiments.
Conclusions
Our studies using the multiproduct enzyme MtTPS5 from
Medicago truncatula demonstrate viable mechanistic path-
ways for the formation of a structural diverse group of
sesquiterpene hydrocarbon skeletons sharing the 3,7-di-
methyl-10-isopropyloctalin skeleton as their common struc-
tural feature (cadalane skeleton). We validated a proposal of
Arigoni16 that includes a pathway for the production of
compounds containing a cis-configured C2-C3-double
bond representing an alternative to the generally accepted
and well-studied pathway proceeding via initial FDP-NDP
isomerization. Furthermore, our results underline the im-
portance of proton transfer reactions in sesquiterpene bio-
synthesis and show how these reactions expand the
evolutionary flexibility of this highly versatile class of en-
zymes. The alteration of only a single amino acid can have a
dramatic effect on the product profile; as we could show, the
alteration of tyrosine to phenylalanine in MtTPS5 prevents
the formation of a key intermediate via protonation of germ-
acrene D (5). This gives the plant the opportunity to adapt
very quickly to environmental changes. Furthermore, know-
ing the key amino acids that govern the carbon flux in
distinct directions, we can tailor enzymes accordingly. Our
analysis of the stereochemical course of the reaction cascade
by incubation experiments with chirally labeled FDPs and
the determination of the absolute configuration of enzyme
products gave us information about the timing and the
enzymatic control of the reaction steps. The cyclization
steps, as well as the hydride shifts, establish new stereocen-
ters under tight control of the enzyme, resulting in optically
pure products. In contrast, the last steps, involving depro-
tonation and reaction with the nucleophile water, are less
controlled, as evident from the formation of double bond
isomers and epimeric alcohols. In summary, every cationic
species led to a direct product, indicating that the enzyme
was unable to prevent the premature quenching of the
emerging cationic intermediates.
The Tyr526 is also involved in the proton transfer to
(-)-bicyclogermacrene (18) as evident from the reduced
(37) Hong, Y. J.; Tantillo, D. J. J. Am. Chem. Soc. 2009, 131, 7999.
(38) Felicetti, B.; Cane, D. E. J. Am. Chem. Soc. 2004, 126, 7212.
(39) Shishova, E. Y.; Di Costanzo, L.; Cane, D. E.; Christianson, D. W.
Biochemistry 2007, 46, 1941.
(40) Kollner, T. G.; Schnee, C.; Li, S.; Svatos, A.; Schneider, B.;
Gershenzon, J.; Degenhardt, J. J. Biol. Chem. 2008, 283, 20779.
Experimental Section
Protein Expression. Strains of E. coli (BL21-CodonPlus(DE3))
harboring the recombinant vectors of MtTPS520 and mutant Y526F
carrying an N-terminal His8-tag were grown to A600 = 0.5 at 37 °C
5598 J. Org. Chem. Vol. 75, No. 16, 2010

Garms et al.
JOCArticle
in LB-medium with kanamycin at 50 μg mL-1. After induction with
isopropyl β-D-1-thiogalactopyranoside (IPTG, final concentration
1 mM), cultures were shaken overnight at 16 °C and 200 rpm. Cells
were harvested by centrifugation for 20 min at 4000 rpm, and the
pellet was resuspended in lysis buffer (50 mM NaH2PO4, 300 mM
NaCl, 10 mM imidazole, pH 8.0) and incubated with lysozyme
(1 mg mL-1) for 1 h at 4 °C. Disruption of the cells was achieved by
sonication for 2 ꢀ 2 min. The cell debris was removed by centrifuga-
tion at 10000g for 30 min. The supernatant was passed over a
column of Ni2þ-NTA-Agarose (QIAGEN, Hilden, Germany),
equilibrated with eight bed volumes of lysis buffer. After being
washed twice with four bed volumes of washing buffer (50 mM
NaH2PO4, 300 mM NaCl, 20 mM imidazole, pH 8.0), the protein
was eluted with elution buffer (50 mM NaH2PO4, 300 mM NaCl,
250 mM imidazole, pH 8.0). The purified protein was desalted into a
TRIS-buffer (50 mM TRIS, pH 7.5, 10 mM NaCl, 10% glycerol) by
passing through a NAP 25 column (Amersham Biosciences,
Uppsala, Sweden), diluted to reach a concentration of 1 mg mL-1
and stored at -20 °C. Protein quantification was performed by
using a method of Bradford et al.41
Site-Directed Mutagenesis. For site-directed mutagenesis, the
QuickChange site-directed mutagenesis kit was used according
to the manufacturer’s instructions. The PCR-based mutagenesis
protocol was performed with the cDNA of MtTPS5 cloned into
the pHis8-3 expression vector42 using primers containing the
desired mutations (Y526Ffwd GTTTTATGGATGTTATTTT-
CAAAAACAAAGATAAC, Y526Frev GTTATCTTTGTTTT-
TGAAAATAACATCCATAAAAC). The mutagenized construct
was fully sequenced before expression.
For quantification of enzyme products, the compounds were
first separated on a gas chromatograph (H2 carrier gas 1.5 mL
min-1, injection volume 2 μL) under the conditions described
above and subsequently analyzed on a flame ionization detector
(FID) (250 °C). Correction of the different response factors of
sesquiterpene hydrocarbons and alcohols was achieved using
calibration curves obtained from samples with different con-
centrations of (E)-β-caryophyllene (10) and torreyol (31). The
average and standard deviations of relative ratios were deter-
mined by at least four independent samples setting the sum of
identified compounds to 100%.
The enantiomers of the enzyme products were separated and
identified by GC-MS using a heptakis(2,3-di-O-methyl-O-tert-
butyldimethylsilyl)-β-cyclodextrin column (50% in OV1701, w/w)
(FS-Hydrodex β-6TBDM) (0.25 mm i.d. ꢀ 25 m ꢀ 0.25 μm film)
operated with helium at 1 mL min-1 as carrier gas, a splitless
injection of 1 μL sample at 220 °C and a temperature program
starting from 60 °C kept for 5 min, followed by a ramp of 2 °C
min-1 (for R-copaene (8) 1 °C min-1) to 160 °C followed by an
additional ramp of 30 °C min-1 to 220 °C with 2 min hold. The
separation of torreyl enantiomers was achieved by using a heptakis-
(2,3,6-tri-O-methyl)-β-cyclodextrin column (50% in OV1701, w/w)
(FS-Hydrodex β-PM) (0.25 mm i.d. ꢀ 25 m ꢀ 0.25 μm film).
Helium was used as carrier gas at a constant flow rate of 1 mL min-1
,
and samples (1 μL) were injected at 220 °C. The GC was pro-
grammed with an initial oven temperature of 110 °C (15-min hold),
which was then increased 2 °C min-1 up to 160 °C followed by a
30 °C min-1 ramp until 220 °C (2-min hold). Samples containing
both enantiomeres were either gifts from von Reuss or were prepared
by mixing pentane extracts of the liverwort Preissia quadrata
(containing generally the opposite enantiomer of the relevant
sesquiterpene46) with the corresponding reference (Table SI-4, Sup-
porting Information) or obtained by acid-catalyzed rearrangements
of racemic germacrene D (5) and R-copaene (8) (see above). No
information about the absolute configuration of the torreyol (31)
isolated from Cortinarius odorifer Britz (gift from Simon Egli) was
available.47 Therefore, we determined the optical rotation with a
polarimeter. The value of [R]D = þ100.6 (c = 0.9, CHCl3) estab-
lishes the configuration of 31 isolated from the fungi to be the
opposite of that reported by Borg-Karlson et al.48
Identification and Determination of the Stereochemistry of
Enzyme Products. GC-MS analysis was performed on an
instrument equipped with a ZB-5 capillary column (0.25 mm i.
d. ꢀ 15 m with 0.25 μm film). One microliter of the sample was
injected in splitless mode at an injection port temperature of 220 °C.
The oven temperature was kept at 50 °C for 2 min followed by a
ramp of 10 °C min-1 to 240 °C followed by an additional ramp of
30 °C min-1 to 280 °C and finally kept for 2 min. Helium at a flow
rate of 1.5 mL min-1 served as carrier gas. Ionization potential was
set to 70 eV, and scanning was performed from 40 to 250 amu.
Compounds were identified by comparing their mass spectra and
Enzyme Assays for Product Analysis. Standard assays con-
tained 600 nM purified protein in assay buffer (25 mM HEPES,
pH 7.5, 10% glycerol, 10 mM MgCl2, 1 mM DTT) with 50 μM
substrate (FDP, (1S)-[1-2H]-FDP, (1R)-[1-2H]-FDP, or [1,1-2H2]-
FDP) in a final volume of 1 mL. Deuterated FDPs were syn-
thesized as previously described.49,50 The reaction mixture was
covered with 100 μL of pentane containing 1 ng μL-1 of dodecane
as an internal standard to trap the reaction products. After being
incubated for 90 min at 30 °C, the reaction was stopped by
vortexing for 20 s. The whole mixture was frozen in liquid nitrogen,
and the pentane layer was removed after thawing and analyzed by
GC-MS as described above.
To analyze the protonation reaction, 50 μL of the purified
enzyme (1 mg mL-1) was lyophilized and redissolved in 50 μL
of D2O and incubated for 30 min on ice to ensure proper H-D
exchange of the enzyme. An aliquot of 20 μL of the pro-
tein sample was analyzed in assay buffer prepared with D2O
(>99% d1) containing 50 μM FDP. The assays were incubated
ꢀ
Kovats indices (retention indices) with those of published reference
spectra in MassFinders’ (software version 3.5) and Adams’43
terpene library and in the NIST database. In addition, retention
indices (RI) of sesquiterpene peaks derived by calibrating GC runs
with a C8-C20 alkane standard were compared with RI values
of authentic reference compounds (Table SI-1, Supporting
Information). Many of the references not commercially available
were either obtained as described below or kindly provided by
Stefan von Reuss, Hamburg, Germany; Charles Fehr, Fimenich,
€
€
Geneva, Switzerland; Alois Furstner, Muhlheim an der Ruhr,
Germany; and Simon Egli, Birmensdorf, Switzerland. Essential
oils with known composition containing relevant sesquiterpenoids
were purchased from a commercial supplier or were generously
provided by Stefan von Reuss, Hamburg, Germany. Additional
references were generated by acid-catalyzed rearrangements of
€
germacrene D (5) (W. A. Konig) and R-copanene (8) using condi-
tions reported in the literature.44,45 The liverwort Preissia quadrata
was collected by H. J. Zuendorf (Herbarium Haussknecht, Jena,
€
Germany) near Talburgel, Germany.
€
€
(46) Konig, W. A.; Bulow, N.; Fricke, C.; Melching, S.; Rieck, A.; Muhle,
H. Phytochemistry 1996, 43, 629.
(41) Bradford, M. M. Anal. Biochem. 1976, 72, 248.
(42) Jez, J. M.; Ferrer, J.-L.; Bowman, M. E.; Dixon, R. A.; Noel, J. P.
Biochemistry 2000, 39, 890.
(43) Adams, R. P. Identification of Essential Oil Components by Gas
Chromatography/Mass Spectrometry, 4th ed.; Allured Publishing Corp.:
Carol Stream, IL, 2007.
(47) Egli, S.; Gfeller, H.; Bigler, P.; Schlunegger, U. P. Eur. J. Forest
Pathol. 1988, 18, 351.
(48) Borg-Karlson, A.-K.; Norin, T.; Talvitie, A. Tetrahedron 1981, 37,
425.
(49) Cane, D. E.; Oliver, J. S.; Harrison, P. H. M.; Abell, C.; Hubbard,
B. R.; Kane, C. T.; Lattman, R. J. Am. Chem. Soc. 1990, 112, 4513.
(50) Cane, D. E.; Iyengar, R.; Shiao, M. S. J. Am. Chem. Soc. 1981, 103,
914.
(44) Bulow, N.; Konig, W. A. Phytochemistry 2000, 55, 141.
(45) Ohta, Y.; Ohara, K.; Hirose, Y. Tetrahedron Lett. 1968, 9, 4181.
J. Org. Chem. Vol. 75, No. 16, 2010 5599

Garms et al.
JOCArticle
for 1 h at 30 °C. The reaction products were collected by a solid-
phase microextration fiber (SPME) consisting of 100 μm poly-
dimethylsiloxane and analyzed by GC-MS. The labeling degree of
each product was calculated from the intensities of the respective
[M]þ, [M þ 1]þ for hydrocarbons and [M - H2O]þ, [M -
H2Oþ1]þ for alcohols after correcting for the abundance of their
13C satellite peaks.
For identifying and determining the stereochemistry of copan-
3-ol (25) and cubebol (21), two 10-mL assays containing 600 nM
purified enzyme in assay buffer with 50 μM FDP were covered
with 10 mL of pentane. After being incubated overnight at 30 °C,
the mixture was extracted three times with 5 mL of pentane. The
combined organic phases were passed through a Pasteur pipet
containing Na2SO4, and the volume was reduced to ∼100 μL.
Alcohols were separated on silica (1 g, Pasteur pipet) using
pentane/ether (6:1, v/v) for elution. Two fractionshighly enriched
in the desired compounds were obtained, besides fractions con-
taining both compounds.
Acid-Catalyzed Dehydratization and Rearrangement of Copaen-
3-ol (25) and Cubebol (21). The solvent of samples containing ∼500 ng
(GC) of the sesquiterpene alcohol was removed under an argon
stream. The compounds were redissolved under argon atmosphere
in 200 μL of dry CH2Cl2, and 5 μL of trifluoroacetic anhydride
(derivatization grade) was added. The solution was stirred for
30 min. Subsequently, after the addition of 1 mL of pentane, the
reaction was quenched with 1 mL of 10% NaHCO3. The aqueous
phase was extracted twice with 1 mL of pentane. The unified organic
extracts were dried (Na2SO4) and reduced to a volume of ∼100 μL
under a stream of argon. Samples were analyzed by GC-MS as
described above.
Dehydratization of copan-3-ol (25) resulted in R-copaene (8)
(60%) and in the three rearrangement products R-muurolene
(19) (21%), γ-muurolene (16) (9%), and δ-cadinene (22) (9%).
The rearrangement of cubebol (21) afforded δ-cadinene (22)
(70%), cadina-1,4-diene (23) (18%), and trans-cadina-1(6),4-
diene (15) (12%) (Figure SI-1, Supporting Information).
Acknowledgment. Financial support of this work by the
Max Planck Society is gratefully acknowledged.
Supporting Information Available: Complete tables of
retention indices, relative ratios of products, sources of refer-
ences, mass spectra of all enzyme products in comparison to
authentic references, mass spectrometric data from incubation
experiments with deuterium-containing substrates, deuterium
content of products obtained by experiments conducted in
highly enriched D2O medium, and gas chromatographic data
of the chiral separation of enzyme products. Total ion chroma-
tograms of catalyzed dehydratization products of copan-3-ol
(25). This material is available free of charge via the Internet at
http://pubs.acs.org.
5600 J. Org. Chem. Vol. 75, No. 16, 2010