Germacrene A is a product of the aristolochene synthase-mediated conversion of farnesylpyrophosphate to aristolochene

Article abstract of DOI:10.1021/ja020762p

The biosynthesis of several sesquiterpenes has been proposed to proceed via germacrene A. However, to date, the production of germacrene A has not been proven directly for any of the sesquiterpene synthases for which it was postulated as an intermediate. We demonstrate here for the first time that significant amounts of germacrene A (7.5% of the total amount of products) are indeed released from wild-type aristolochene synthase (AS) from Penicillium roqueforti. Germacrene A was identified through direct GC-MS comparison to an authentic sample and through production of β-elemene in a thermal Cope rearrangement. AS also produced a small amount of valencene through deprotonation of C6 rather than C8 in the final step of the reaction. On the basis of the X-ray structure of AS, Tyr 92 was postulated to be the active-site acid responsible for protonation of germacrene A (Caruthers, J. M.; Kang, I.; Rynkiewicz, M. J.; Cane, D. E.; Christianson, D. W. J. Biol. Chem. 2000, 275, 25533-25539). The CD spectra of a mutant protein, ASY92F, in which Tyr 92 was replaced by Phe, and of AS were very similar. ASY92F was approximately 0.1% as active as nonmutated recombinant AS. The steady-state kinetic parameters were measured as 0.138 min^-1 and 0.189 mM for k_cat and K_M, respectively. Similar to a mutant protein of 5-epiaristolochene (Rising, K. A.; Starks, C. M.; Noel, J. P.; Chappell, J. J. Am. Chem. Soc. 2000, 122, 1861-1866), the mutant released significant amounts of germacrene A (～29%). ASY92F also produced various amounts of a further five hydrocarbons of molecular weight 204, valencene, β-(E)-farnesene, α- and β-selinene, and selina-4, 11-diene.

Full text of DOI:10.1021/ja020762p

Published on Web 09/06/2002
Germacrene A Is a Product of the Aristolochene
Synthase-Mediated Conversion of Farnesylpyrophosphate to
Aristolochene
Melanie J. Calvert, Peter R. Ashton, and Rudolf K. Allemann*
Contribution from the School of Chemical Sciences, UniVersity of Birmingham, Edgbaston,
Birmingham, B15 2TT, U.K.
Received May 31, 2002
Abstract: The biosynthesis of several sesquiterpenes has been proposed to proceed via germacrene A.
However, to date, the production of germacrene A has not been proven directly for any of the sesquiterpene
synthases for which it was postulated as an intermediate. We demonstrate here for the first time that
significant amounts of germacrene A (7.5% of the total amount of products) are indeed released from
wild-type aristolochene synthase (AS) from Penicillium roqueforti. Germacrene A was identified through
direct GC-MS comparison to an authentic sample and through production of â-elemene in a thermal Cope
rearrangement. AS also produced a small amount of valencene through deprotonation of C6 rather than
C8 in the final step of the reaction. On the basis of the X-ray structure of AS, Tyr 92 was postulated to be
the active-site acid responsible for protonation of germacrene A (Caruthers, J. M.; Kang, I.; Rynkiewicz, M.
J.; Cane, D. E.; Christianson, D. W. J. Biol. Chem. 2000, 275, 25533-25539). The CD spectra of a mutant
protein, ASY92F, in which Tyr 92 was replaced by Phe, and of AS were very similar. ASY92F was
approximately 0.1% as active as nonmutated recombinant AS. The steady-state kinetic parameters were
measured as 0.138 min-1 and 0.189 mM for kcat and KM, respectively. Similar to a mutant protein of 5-epi-
aristolochene (Rising, K. A.; Starks, C. M.; Noel, J. P.; Chappell, J. J. Am. Chem. Soc. 2000, 122, 1861-
1866), the mutant released significant amounts of germacrene A (∼29%). ASY92F also produced various
amounts of a further five hydrocarbons of molecular weight 204, valencene, â-(E)-farnesene, R- and
â-selinene, and selina-4,11-diene.
Scheme 1
Introduction
Aristolochene synthase (AS) from Penicillium roqueforti is
a member of the family of sesquiterpene synthases, which
catalyze the cyclization of the universal acyclic precursor
farnesyl pyrophosphate (FPP, 1) (Scheme 1) in marine and
terrestrial plants, fungi, bacteria, and insects, to generate more
than 300 distinct sesquiterpene products, many of which have
important antibiotic, antifungal, or neurotoxic activities.1,2
Ionization of the substrate through loss of pyrophosphate triggers
an exquisitely regio- and stereospecific sequence of intramo-
lecular addition reactions and rearrangements. The enzymes
control solvent access to prevent premature quenching of the
extremely reactive cationic reaction intermediates by water. The
starting conformation of 1 in the enzyme’s active site is known
to be a critical determinant of product diversity.2 Sesquiterpene
synthases are an extreme example of product maximization from
a minimal substrate pool. Many mechanistic studies of FPP
cyclizations have been reported.3-6 However, little is known
about the molecular details of catalysis and stereochemical
control by sesquiterpene synthases, mainly because their genes
have only recently been isolated.4,7-11
* To whom correspondence should be addressed. E-mail: r.k.allemann@
bham.ac.uk.
(1) Glasby, J. S. Encyclopedia of Terpenoids; Wiley: Chichester, 1982.
(2) Cane, D. E. Chem. ReV. 1990, 90, 1089-1103.
(3) Croteau, R. B.; Wheeler, C. J.; Cane, D. E.; Ebert, R.; Ha, H. J. Biochemistry
1987, 26, 5383-9.
(4) Cane, D. E.; Sohng, J.-K.; Lamberson, C. R.; Rudnicki, S. M.; Wu, Z.;
Lloyd, M. D.; Oliver, J. S.; Hubbard, B. R. Biochemistry 1994, 33, 5846-
5857.
(5) Schmidt, C. O.; Bouwmeester, H. J.; Bulow, N.; Konig, W. A. Arch.
Biochem. Biophys. 1999, 364, 167-77.
(6) Benedict, C. R.; Lu, J. L.; Pettigrew, D. W.; Liu, J.; Stipanovic, R. D.;
Williams, H. J. Plant Physiol. 2001, 125, 1754-65.
(7) Proctor, R. H.; Hohn, T. M. J. Biol. Chem. 1993, 268, 4543-4548.
9
11636
J. AM. CHEM. SOC. 2002, 124, 11636-11641
10.1021/ja020762p CCC: $22.00 © 2002 American Chemical Society

Germacrene A Production
A R T I C L E S
Materials and Methods
AS has previously been cloned and overproduced in E.
coli.7,11 The momomeric enzyme of 39 kD catalyses the bivalent
metal-dependent cyclization of FPP to (+)-aristolochene (4),
the precursor of several fungal toxins including the potentially
lethal PR toxin.7 Studies with labeled FPP,12,13 the mechanism-
based inhibitor 12-methylidenefarnesyl pyrophosphate,14 and
with (7R)-6,7-dihydrofarnesyl pyrophosphate15 led to the pro-
posal that the cyclization of FPP to aristolochene proceeds
through at least two discrete intermediates, S-germacrene A (2)
and eudesmane cation (3) (Scheme 1). However, so far
germacrene A could not be detected directly in AS-catalyzed
conversions of FPP to aristolochene.15
Materials. Oligonucleotide primers were purchased from Alta
Bioscience, University of Birmingham. The pZW04 plasmid encoding
wild-type AS from P. roqueforti was kindly donated by Prof. David
Cane, Brown University. Germacrene A from soldier cephalic secretion
of a subterranean termite species and R-and â-selinene from Albies
magnifica oleoresin were gifts from Larry Cool (Forest Products
Laboratory, University of California, Berkeley). Valencene was a gift
from De Monchy Aromatics Ltd. â-(E)-Farnesene was a gift from John
A. Pickett, FRS, and Lynda Ireland (BBSRC - Institute for Arable
Crops, Rothamsted). Aristolochene was enzymatically generated from
FPP using AS. The Quick-change mutagenesis kit was obtained from
Stratagene, and miniprep kits and Big Dye sequencing reagents were
obtained form Qiagen and Applied Biosystems, respectively. [1-3H]-
FPP (21.5 mCi/mmol) was purchased from Sigma and was diluted with
cold FPP to a working specific activity of 50 µCi/µmol. Q-Sepharose
was obtained from Amersham Pharmacia Biotech. Ultracell Amicon
YM3 membranes were purchased from Millipore, and BCA protein
acid reagent was obtained from the Pierce Chemical Co. EcoScint
scintillation fluid was from National Diagnostics. All other chemicals
were from Fluka or Sigma.
Site-Directed Mutagenesis of Recombinant AS cDNA. The Quick-
change site-directed mutagenesis kit (Stratagene) was used to introduce
the Y92F mutation (TAC f TTC) into the AS encoding plasmid,
pZW04, according to the manufacturer’s instructions. The mutagenic
primers were as follows: 5′-CAGAGGTTACTTGTCTTTTCTTC-
CCTCTTGCACTGG-3′ and 5′-CCAGTGCAAGAGGGAAGAAAAG
ACAAGTAACCTCTG-3′ (altered bases shown in bold). Plasmids were
purified from overnight LB/amp cultures (5 mL) using the Qiagen
miniprep kit as described by the manufacturer. Mutations were
confirmed by DNA sequence analysis, using an Applied Biosystems
3700 automated DNA sequencer, Functional Genomics Laboratory,
University of Birmingham.
Expression in E. coli and Purification of Wild-type and Mutant
Aristolochene Synthases. Wild-type AS and ASY92F constructs were
transformed into and expressed in E. coli BL21(DE3) cells. The cells
were grown at 37 °C in LB medium with 0.3 mM ampicillin until they
reached an A600 of 0.8-1 and were then induced with 0.5 mM isopropyl-
â-D-1-thiogalactopyranoside and incubated for a further 2 h. Cells were
harvested by centrifugation at 5000g for 10 min and resuspended in
20 mM Tris, pH 8, 5 mM EDTA, 5 mM â-mercaptoethanol. AS and
ASY92F were extracted from the insoluble inclusion bodies using the
base extraction procedure previously described16 and purified on a
Q-Sepharose Fast Flow column (2.5 × 20 cm), equilibrated with 20
mM Tris, pH 8, 5 mM EDTA, 5 mM 2-mercaptoethanol. Enzyme was
eluted using a linear 0.1-0.6 M NaCl gradient, and the fractions
containing protein were identified by absorption at 280 nm and SDS-
polyacrylamide gel electrophoresis. Fractions were assayed for enzyme
activity, and pooled active enzyme was dialyzed against 10 mM Tris,
pH 7, 5 mM â-mercaptoethanol (3 × 3 L). AS and ASY92F were
concentrated using an Amicon 8050 with an Ultracell amicon YM3
membrane and were judged to be >98% pure by SDS-PAGE. Protein
concentrations were determined by the BCA (bicinchoninic acid)
method calibrated with bovine serum albumin per the manufacturer’s
instructions (Pierce Chemical Co.).
Assay for Enzymatic Activity. AS and ASY92F were assayed in a
total volume of 250 µL containing 10 mM Tris, pH 7.5, 5 mM MgCl2,
5 mM 2-mercaptoethanol, and 15% glycerol (v/v) and 0.25-15 µM or
2.5-100 µM of [1-3H]-FPP for AS or ASY92F assays, respectively.
Reaction mixtures were prewarmed to 30 °C and initiated by addition
of 50 µL of enzyme, to give a final concentration of 0.4 µM AS or 4
µM ASY92F. Reactions were terminated by the addition of 100 µL of
100 mM EDTA, pH 7.25, and overlaid with 0.5 mL of n-hexane. The
samples were vortex mixed for 15 s, and the hexane layer was then
vortexed with 0.5 g of silica in 1 mL of n-hexane. The samples were
extracted with an additional 2 × 1 mL of hexane, and the combined
In the proposed reaction sequence, the active site of aris-
tolochene synthase orientates the C10-C11 bond of FPP for
attack at C-1 after (or concurrent with) metal triggered diphos-
phate departure (Scheme 1). Deprotonation of H-12 results in
the formation of the cis-fused Decalin S-germacrene A (2).
Protonation of germacrene A at C-6 and cyclization through
electron flow from the double bond at C2-C3 yield the bicyclic
eudesmane cation (3) with an internal bond between C2 and
C7. A hydride shift from C2 followed by a methyl shift from
C7 to C2 and deprotonation at C8 results in the formation of
aristolochene.16
Proof that germacrene A was an intermediate in the reaction
has been precluded by the inability to demonstrate its release
from any of the sesquiterpene synthases for which it was
postulated as an intermediate.17 Investigations into the mech-
anism of 5-epi-aristolochene synthase (EAS) from Nicotiana
tabacum, which shows only 16% sequence identity with AS,
have provided evidence for a mechanism similar to that proposed
for AS.17,18 The two synthases provide different templates for
binding FPP and the reaction intermediates to produce the
correct stereoisomer. Analysis of the X-ray structure of EAS
and of the catalytic properties of a mutant EAS showed that
Tyr 520 is the active-site acid which protonates germacrene A.19
Mutation of Tyr 520 to phenylalanine led to the accumulation
of germacrene A as the sole reaction product.17
On the basis of the X-ray structure of AS which has been
solved at 2.5 Å in the absence of either substrate or a substrate
analogue,16 and on molecular modeling studies, it was suggested
that the hydroxyl group of Tyr 92, which does not align with
Tyr 520 of EAS, could act as the proton donor to C-6 of
germacrene A in AS.16 Here we report results from studies
addressing the intermediacy of germacrene A in AS catalysis
and the specific role of Tyr 92 during the conversion of FPP to
aristolochene.
(8) Cane, D. E.; Kang, I. Arch. Biochem. Biophys. 2000, 376, 354-364.
(9) Fekete, C.; Logrieco, A.; Giczey, G.; Hornok, L. Mycopathologia 1997,
138, 91-97.
(10) Trapp, S. C.; Croteau, R. B. Genetics 2001, 158, 811-832.
(11) Cane, D. E.; Wu, Z.; Proctor, R. H.; Hohn, R. H. Arch. Biochem. Biophys.
1993, 304, 415-419.
(12) Cane, D. E.; Prabhakaran, P. C.; Salaski, E. J.; Harrison, P. H. M.; Noguchi,
H.; Rawlings, B. J. J. Am. Chem. Soc. 1989, 111, 8914-8916.
(13) Cane, D. E.; Prabhakaran, P. C.; Oliver, J. S.; McIlwaine, D. B. J. Am.
Chem. Soc. 1990, 112, 3209-3210.
(14) Cane, D. E.; Bryant, C. J. Am. Chem. Soc. 1994, 116, 12063-12064.
(15) Cane, D. E.; Tsantrizos, Y. S. J. Am. Chem. Soc. 1996, 118, 10037-10040.
(16) Caruthers, J. M.; Kang, I.; Rynkiewicz, M. J.; Cane, D. E.; Christianson,
D. W. J. Biol. Chem. 2000, 275, 25533-25539.
(17) Rising, K. A.; Starks, C. M.; Noel, J. P.; Chappell, J. J. Am. Chem. Soc.
2000, 122, 1861-1866.
(18) Whitehead, I. M.; Threlfall, D. R.; Ewing, D. F. Phytochemistry 1989, 28,
775-779.
(19) Starks, C. M.; Back, K.; Chappell, J.; Noel, J. P. Science 1997, 277, 1815-
1820.
9
J. AM. CHEM. SOC. VOL. 124, NO. 39, 2002 11637

A R T I C L E S
Calvert et al.
Table 1. Kinetic Parameters for AS and ASY92F
kinetic parameters
enzyme E
KM (µM)
kcat (s-1
)
kcat/KM (M-1 s-1
)
AS
ASY92F
2.3 ( 0.5
188.7 ( 52.8
0.03 ( 0.01
0.0023 ( 0.0005
13 858 ( 2989
12.36 ( 0.8
hexane extracts were added to a scintillation vial containing 4 mL of
EcoScint and analyzed for radioactivity. The percentage conversion of
[1-3H]-FPP to hexane extractable products was determined by compar-
ing the observed radioactivity to that observed with a known concentra-
tion of [1-3H]-FPP. Background levels of radioactivity were observed
in control samples without the enzyme.
CD Spectroscopy of AS and ASY92F. Circular dichroism experi-
ments were run for AS and ASY92F at protein concentrations of 6
µM in 10 mM Tris, pH 7.5, 5 mM MgCl2, 0.2 mM dithiothreitol in the
wavelength interval 190-300 nm using a Jasco J-810 spectropolarim-
eter. The CD spectra were measured at 20 °C using 1 mm quartz
cuvettes.
Characterization of Products from Aristolochene Synthase
Mutant by GC-MS. Purified ASY92F (50 µM) was incubated with
230 µM FPP in 10 mM Tris, pH 7.5, 5 mM MgCl2, 5 mM
2-mercaptoethanol, and 15% glycerol in a final volume of 500 µL at
30 °C for 15 min, 1 h, and 12 h. Reactions were terminated by the
addition of 200 µL of 100 mM EDTA, pH 7.25, and the products were
extracted by vortexing against hexane (3 × 3 mL) and purified using
a silica column (containing 1.5 g of silica). Pooled hexane extracts
were concentrated by rotary evaporation on ice. Identical incubations
were performed with wild-type AS.
Figure 1. CD spectra of 6 µM AS (O) and 6 µM ASY92F (9) in 10 mM
Tris, pH 7.5, 5 mM MgCl2, 0.2 mM dithiothreitol at 20 °C.
(produced by incubation of wild-type AS with FPP) and
germacrene A standards yielded distinct peaks with retention
times of 20.68 and 21.37 min, respectively. The corresponding
mass spectra exhibited molecular ions at 204 mass units as well
as characteristic fragmentation patterns (Figure 2).17
The GC analysis of the products formed by AS revealed the
formation of two further hydrocarbons in addition to aris-
tolochene, which according to the total ion chromatogram
represented approximately 92% of the hexane extractable
material, and traces of selina-4,11-diene (8) and R-selinene (9)
(vide infra) (Table 2). One of them made up approximately 7.5%
of the products and showed the same retention time and
fragmentation pattern as germacrene A (4) (Figure 2). Co-
injection of a germacrene A standard and the products produced
by AS led to an increase in the intensity of the peak corre-
sponding to 4.
Germacrene A is known to undergo a thermal Cope rear-
rangement to â-elemene.21 When the hexane extractable prod-
ucts formed by AS were analyzed by GC-MS with an injector
temperature of 200 °C, the peak corresponding to aristolochene
remained unaffected, while the retention time of the major
byproduct was significantly shortened (Figure 3). The new peak
was identified through its mass spectrum and comparison with
the known spectrum of â-elemene.22
The amount of germacene A formed by AS relative to
aristolochene was not dependent on the extent of turnover,
indicating that a small amount of germacrene A was indeed
released from the active site of the wild-type enzyme. Once
released, the enzyme appeared not to rebind the intermediate
in good agreement with the observation that germacrene A is
not a substrate for AS. Alternatively, germacrene A could be a
side product of the AS-catalyzed conversion of FPP rather than
an intermediate. While our results do not allow us to distinguish
between these two possible mechanisms, they nevertheless
represent the first demonstration that germacrene A is indeed
produced at significant levels by AS and released from the
enzyme for which it was postulated as an intermediate.
Interestingly, an additional hydrocarbon of molecular mass
204 was observed in the gas chromatogram (retention time 20.92
min) (Figure 2). It made up only 0.4% of the total amount of
hexane extractable material (Table 2). Its mass spectrum closely
matched that of valencene (5) in the Wiley and NIST98 libraries
The hexane extractable products were analyzed by GC-MS using a
Carlo Erba CE8000 gas chromatograph (equipped with a BPX5 30 m
column) and a Prospec Mass spectrometer. High-resolution GC-MS
data were obtained on a Micromass GCT with an integrated GC
(HP6890) (equipped with a 30 m DB5 column). Splitless injections (5
µL) were generally performed at the relatively low injection port
temperature of 100 °C to prevent the Cope rearrangement of germacrene
A to â-elemene, which occurs at higher injection port temperatures.
The column temperature was maintained at 60 °C for 3 min and was
then increased to 150 °C at 4 °C min-1
.
Results and Discussion
Because several lines of indirect evidence (vide supra) had
suggested that germacrene A was formed from FPP (Scheme
1), but had never been observed during AS catalysis, we have
investigated the kinetics and the product distribution of AS with
the use of modern mass spectrometer detectors with increased
sensitivity.
AS was produced to high levels in E. coli and purified to
apparent homogeneity. The steady-state kinetic parameters of
purified recombinant AS were measured by incubation with
[1-3H]-FPP and monitoring the formation of tritiated, hexane
extractable products.11 The KM and kcat values for AS were
determined as 2.3 ( 0.5 µM and 0.03 ( 0.01 s-1, respectively
(Table 1), which was in good agreement with those reported
for AS isolated from P. roqueforti (KM ) 0.55 µM and kcat
)
0.04 s-1 20
)
and suggested that the recombinant protein was
properly folded. The CD spectrum of AS showed that its
secondary structure was mainly R-helical as was expected from
the X-ray structure (Figure 1).
The products obtained from incubation of 1 with AS were
analyzed by GC-MS. Gas chromatography of aristolochene
(21) De Kraaker, J. W.; Franssen, M. C. R.; de Groot, A.; Koenig, W. A.;
Bouwmeester, H. J. Plant. Physiol. 1998, 117, 1381-1392.
(22) Wiley Mass Spectral Database # 130986.
(20) Hohn, T. M.; Plattner, R. D. Arch. Biochem. Biophys. 1989, 272, 137-
143.
9
11638 J. AM. CHEM. SOC. VOL. 124, NO. 39, 2002

Germacrene A Production
A R T I C L E S
Figure 3. Analysis of the temperature-induced reaction of the hexane
extractable products of FPP utilization by AS. (a) GC analysis of the
hydrocarbons produced by AS from FPP at an injection port temperature
of 200 °C. Note that the material corresponding to 4 is not affected by the
increased temperature, while the retention time of the byproduct is
significantly shortened; (b) mass spectrum of the material eluting at 17.3
min in (a); (c) mass spectrum of GC fraction from authentic â-elemene.
X-ray crystallography19 and mutational studies17 have sug-
gested that a tyrosine residue of EAS served as the active-site
acid required for the protonation of germacrene A. In the final
step of 5-epi-aristolochene production, the conjugate base
abstracts a proton from C8. While Tyr 92 of AS is not in a
homologous position to Tyr 520 of EAS, inspection of the
crystal structure of AS16 suggested that it could nevertheless
fulfill a similar function during AS catalysis. Similarly, a
phenolate anion could abstract a proton from C8 to produce
aristolochene. Occasionally, a proton appeared to be abstracted
from C6 to produce 5 (Scheme 1). To investigate the role of
Tyr 92 within the active site of AS, we have replaced Tyr 92
by phenylalanine to create the mutant ASY92F.
ASY92F was produced in E. coli and purified to apparent
homogeneity. The steady-state kinetic parameters were deter-
mined as described for AS. Relative to the wild-type enzyme,
ASY92F displayed an 80-fold increased value for KM, while
the turnover number kcat was reduced to 0.0023 s-1, resulting
in an overall decrease of the catalytic efficiency of more than
3 orders of magnitude (Table 1). While the replacement of Tyr
Figure 2. GC-MS analysis of hexane extractable products of FPP utilization
by AS. (a) Total ion chromatogram of the hydrocarbons produced by AS
from FPP; (b) mass spectrum of product 2 of AS catalysis; (c) mass spectrum
of GC fraction from authentic germacrene A (2); (d) mass spectrum of
product 5 of AS catalysis; (e) mass spectrum of GC fraction from authentic
valencene.
of mass spectra (Figure 2). Co-injection with a reference sample
of valencene confirmed the identity of this hydrocarbon that is
most likely produced by the abstraction of a proton from C6
rather than from C8 as is observed in the production of
aristolochene (Scheme 1). Inspection of the crystal structure of
AS indicated that the hydroxyl of Tyr 92 was approximately
equidistant to C6 and C8 of eudesmane cation (3), suggesting
that the phenolate anion could act as the base abstracting either
of these protons. The yield of valencene produced was, however,
very low (Table 2), and a residue other than Tyr 92 might have
acted as the base (vide infra).
9
J. AM. CHEM. SOC. VOL. 124, NO. 39, 2002 11639

A R T I C L E S
Calvert et al.
Table 2. Relative Amounts of Hexane Extractable Hydrocarbon Products from Incubations with FPP Determined by GC Analysis
product
2
4
5
6
7
8
9
germacrene A
aristolochene
valencene
â-farnesene
â-selinene
selina-4,11-diene
R-selinene
retention time (min)
WTAS
Y92F
21.37
7.5
28.6
20.68
91.5
56.2
20.92
0.4
5.9
19.78
0
1.8
20.81
0
1.8
20.37
0
2.1
21.03
0
3.6
92 with Phe reduced the catalytic power of the enzyme
significantly, it is important to keep in mind that the catalytic
efficiencies of wild-type AS and other terpene cyclases are
intrinsically rather low. These enzymes have evolved to produce
cyclic hydrocarbons with often exquisite specificity rather than
with high speed. The significant decrease in the catalytic power
of the mutant was not a consequence of a grossly altered
structure of the enzyme. The CD spectra of AS and ASY92F
were very similar, indicating that they had similar secondary
structure (Figure 1). Given the similar hydrophobicities of
germacrene A and aristolochene, the decreased kcat observed
for ASY92F was most likely due to a decrease of the rate of
the chemical steps rather than of the rate of product release
which has been shown to be rate limiting in several WT-
sesquiterpene cyclases.23,24
The hexane extractable materials produced from 1 by
ASY92F were analyzed by GC-MS (Figure 4). While the mutant
enzyme produced a significantly increased amount of germa-
crene A (28.6%) when compared to WT-AS (Table 2), it still
produced aristolochene. However, substantial amounts of further
five hydrocarbon products, valencene (5), â-(E)-farnesene (6),
â-selinene (7), selina-4,11-diene (8), and R-selinene (9) (Scheme
2), were also observed in the GC trace (Figure 4). The major
one of these (5.9%) was valencene (5) (Table 2), which had
already been identified as a minor product of the reaction of 1
with AS. Three further bicyclic hydrocarbons were generated
through proton loss from C2, C4, and the methyl group on C2
of eudesmane cation (3) to produce selina-4,11-diene (8),
R-selinene (9), and â-selinene (7) (Scheme 2). The relative
amounts for 8, 9, and 7 were 2.1, 3.6, and 1.8% of the total
hexane extractable products (Table 2). Germacrene A is known
to undergo a slow and spontaneous acid-catalyzed rearrangement
to form selinenes in chloroform. However, no evidence for such
a rearrangement was obtained in hexane, which was the solvent
in our studies. Hydrocarbons 7, 8, and 9 appeared therefore to
be true enzymatic products of ASY92F.
Compounds 7 and 9 were identified from their MS spectra
and comparison with reference spectra of the Wiley and the
NIST 98 libraries of mass spectra. Co-injections with reference
samples of R- and â-selinene confirmed their identity. The mass
spectrum of selina-4,11-diene was identical to the reference
spectra in the NIST 98 mass spectral library.
The final hydrocarbon which was generated in 1.8% yield
by ASY92F could be identified as â-(E)-farnesene (6) (Scheme
2) by co-injection with authentic material and comparison of
the mass spectra. The production of 6 by this mutant suggested
strongly that the initial cyclization of FPP to germacrene A was
not concerted as had been suggested previously.16 Instead, it
Figure 4. GC-MS analysis of hexane extractable products formed from
FPP by ASY92F. (a) Gas chromatogram of the products of ASY92F
catalysis; mass spectra of 9 (b), 7 (c), 8 (d), and 6 (e) produced from FPP
by ASY92F.
appeared to follow a stepwise mechanism in which the departure
of the pyrophosphate group preceded deprotonation of the cis-
methyl group (C12) (Scheme 1).
(23) Mathis, J. R.; Back, K.; Starks, C.; Noel, J.; Poulter, C. D.; Chappell, J.
Biochemistry 1997, 36, 8340-8348.
(24) Cane, D. E.; Chiu, H.-T.; Liang, P.-O.; Anderon, K. S. Biochemistry 1997,
36, 8332-8339.
9
11640 J. AM. CHEM. SOC. VOL. 124, NO. 39, 2002

Germacrene A Production
A R T I C L E S
Scheme 2
catalysis. The possibility that 2 was a side product in AS
catalysis could not be ruled out completely, in which case
changing Tyr 92 to Phe would have affected the relative rates
of formation of germacrene A and aristolochene. However, our
results together with previous studies using mechanism-based
inhibitors and labeled FPP as well as the high structural
similarity of AS with EAS are consistent with the proposal that
germacrene A is indeed an intermediate during the AS-catalyzed
formation of aristolochene from FPP.
The results presented above indicated that Tyr 92 might act
as the critical proton donor responsible for the activation of 2
to eudesmane cation (3), its absence severely impairing the
enzyme’s catalytic ability. The production of a significant
amount of 2 by ASY92F suggested that if Tyr 92 was indeed
the active-site acid during AS catalysis, then some other group
must have served as the surrogate acid in the mutant. Inspection
of the X-ray structure of AS revealed that the side chain of Lys
206,16 which is involved in a hydrogen-bonding network with
Asp 203 and Arg 200, might act as the proton donor. It is ideally
placed to serve as the active-site acid in ASY92F to protonate
the 6,7-double bond of germacrene A. Subsequently, because
the ꢀ-amino group of Lys 206 is approximately equidistant to
carbons 2, 4, 6, 8, and the carbon of the methyl group on C3 of
eudesmane cation (3), it could also abstract a proton from 3 to
produce 7, 8, and 9 or from C6 and C8 to produce valencene
(5) and aristolochene (4), respectively (Scheme 2).
The ability of terpene synthases to produce multiple products
has been well documented and may be a consequence of the
reaction mechanism employed by these enzymes. The most
extreme cases reported so far are δ-selinene and γ-humulene
synthase for which 34 and 52 discrete products were identified.25
The recent publication of the X-ray structures of EAS from
Nicotiana tabacum,19 AS from Penicillium roqueforti,16 pen-
talenene synthase from Streptomyces UC5319,26 and trichodiene
synthase from Fusarium sporotrichioides27 revealed that the
three-dimensional structures of the active-site domains of these
and most likely many other terpene cyclases are very similar.
Only relatively few active-site residues appear to account for
the formation of different products. It is therefore not surprising
that a relatively small change in the active site of AS can lead
to a significant increase in product diversity. Terpene cyclases,
which have evolved to form tightly fitting templates for their
substrates, subtly channel conformation and stereochemistry
during the cyclization reactions. Alterations of active-site
residues, introduced by either nature or man, therefore often
result in the formation of desired or aberrant cyclization
products, thereby shedding light on the molecular mechanisms
that control these exquisitely specific reactions.
In the wild-type enzyme, Lys 206 is also hydrogen bonded
to Tyr 92. This hydrogen-bonding network provides a proton
shuttle from Arg 200, which lies at the top of the active-site
cleft and is exposed to the solvent, to the double bond at C6-
C7 of germacrene A, which is deeply buried in the active site
of the enzyme. This hydrogen-bonding network might enhance
the acidity of Tyr 92 sufficiently to allow protonation of the
C6-C7 double bond in 2.
Despite minimal sequence identity, P. roqueforti AS and N.
tabacum EAS appear to share some common active-site features.
While the amino acid residues that stabilize Tyr 520 in EAS
are different from those involved in the hydrogen-bonding
network observed in AS, a tyrosine residue appears to act as
the general acid responsible for protonating C6 of germacrene
A in both enzymes. It may be possible that such a tyrosine is
the common proton donor in all sesquiterpene synthases whose
mechanism proceeds via germacrene A.
Acknowledgment. We thank David E. Cane for the pZW04
plasmid,2b Larry Cool for germacrene A and R- and â-selinene,
John A. Pickett, FRS, and Lynda Ireland for â-(E)-farnesene,
and Mr. T. Cannon of De Monchy Aromatics Ltd. for valencene.
The expert technical assistance of Graham Burns, David Lewis,
and Lianne Hill and Peter Hancock at Micromass for help with
GC and mass spectrometry is gratefully acknowledged. This
work was supported by the BBSRC.
The identification of a substantial amount of germacrene A
during AS catalysis and the formation of large amounts of 2 by
ASY92F confirmed that 2 was indeed formed during AS
(25) Steele, C. L.; Crock, J.; Bohlmann, J.; Croteau, R. J. Biol. Chem. 1998,
273, 2078-89.
(26) Lesburg, C. A.; Zhai, G.; Cane, D. E.; Christianson, D. W. Science 1997,
277, 1820-4.
(27) Rynkiewicz, M. J.; Cane, D. E.; Christianson, D. W. Proc. Natl. Acad.
Sci. U.S.A. 2001, 98, 13543-13548.
JA020762P
9
J. AM. CHEM. SOC. VOL. 124, NO. 39, 2002 11641