Tyrosine 92 of aristolochene synthase directs cyclisation of farnesyl pyrophosphate

Article abstract of DOI:10.1039/b206517g

A mutant of Aristolochene Synthase (AS), in which Tyr 92 was replaced by Val, produced the alicyclic β-(E)-farnesene as the major product, indicating that cyclisation of FPP is controlled by Tyr 92 in AS.

Full text of DOI:10.1039/b206517g

Tyrosine 92 of aristolochene synthase directs cyclisation of farnesyl
pyrophosphate†
Melanie J. Calvert, Susan E. Taylor and Rudolf K. Allemann*
School of Chemical Sciences, University of Birmingham, Edgbaston, Birmingham, UK B15 2TT.
E-mail: r.k.allemann@bham.ac.uk; Fax: 44 121 414 4446; Tel: 44 121 414 4359
Received (in Cambridge, UK) 8th July 2002, Accepted 19th August 2002
First published as an Advance Article on the web 19th September 2002
A mutant of Aristolochene Synthase (AS), in which Tyr 92
was replaced by Val, produced the alicyclic b-(E)-farnesene
as the major product, indicating that cyclisation of FPP is
controlled by Tyr 92 in AS.
to use 4 as a pheromone.11 Interestingly, b-(E)-farnesene
synthase from M. piperita is homologous to many sesquiterpene
cyclases including 5-epi-aristolochene synthase (EAS) from N.
tabacum,10,12 the tertiary structure of which closely resembles
that of AS.5,13
Sesquiterpene cyclases are a ubiquitously expressed family of
proteins capable of converting the universal acyclic precursor
farnesyl pyrophosphate (FPP, 1) into more than 300 different
cyclic sesquiterpene products.1,2 Product diversity is to a large
extent controlled by the proper folding of 1 in the active sites of
the enzymes. After substrate binding and folding, the reactions
are often initiated by metal cation triggered expulsion of
pyrophosphate followed by a sequence of cyclisations and
rearrangements of exquisite regio- and stereo-specificity. The
reaction sequence is terminated by deprotonation of or water
capture by the final cation.
Aristolochene synthase (AS) from P. roqueforti is a mono-
meric enzyme of 39 kD that catalyses the cyclisation of 1 to
(+)-aristolochene (3) (Scheme 1), the biosynthetic precursor of
several toxins in filamentous fungi including PR-toxin, giga-
ntenone, sporogen-AO1, phaseolinone, phomenone, and bipo-
laroxin.3,4 The structure of AS has been solved to 2.5 Å by X-
ray crystallography.5 Investigations of the mechanism of AS
catalysis6–9 and X-ray crystallography5 suggested that 1 was
bound in the active site of AS in a conformation, which,
subsequent to (or concurrent with) metal triggered pyr-
ophosphate expulsion, favoured attack of C1 by the C10–C11
bond (Scheme 1). Reprotonation of germacrene A (2) at C6,
cyclisation and hydride and methyl migrations generated 3. The
precise folding of 1 leading to the proper spatial orientation of
the C10–C11 double bond and the breaking carbon–oxygen
bond ensures optimal reactivity for cyclisation and prevents the
competing reaction of pyrophosphate elimination by deprotona-
tion from the C3 methyl-group to produce the alicyclic
sesquiterpene b-(E)-farnesene (4). 4 is used extensively by both
plants and insects for communication10 while mammals appear
We have investigated the mechanisms by which AS controls
the specificity of the initial cyclisation of 1 leading to 2 and
prevents the formation of 4. Inspection of the X-ray structure of
AS5 suggested that the relative bulkiness of Tyr 92 might be a
key determinant of proper folding of 1 in the active site of the
enzyme. Tyr 92 is positioned close to the C6–C7 double bond
and might direct folding of 1 leading to cyclisation. Previous
results have shown that Tyr 92 acts as the active site acid which
protonates the C6–C7 double bond of germacrene A to induce
the generation of aristolochene.14 In order to determine whether
Tyr 92 also directed proper folding of 1, we have produced
ASY92V, in which a valine residue replaced Tyr 92 in
aristolochene synthase. Substitution of the bulky side chain of
Tyr with the much smaller Val was predicted to reduce the
efficiency of the cyclisation to 2. As a consequence the mutant
enzyme was predicted to act as a b-(E)-farnesene synthase.
ASY92V was constructed by site directed mutagenesis from
a cDNA of AS in plasmid pZW04.15 ASY92V was efficiently
expressed in E. coli and purified to apparent homogeneity.5 The
CD-spectra of WT-AS and ASY92V matched closely indicating
that the two proteins had similar secondary structures (data not
shown). The steady state kinetic parameters of ASY92V were
determined by incubation with [1-3H]-FPP and monitoring the
formation of tritiated, hexane extractable products.16 The KM
and kcat values for ASY92V were determined as 4.86 mM and
2.51 3 1024 s21, respectively. Relative to the wild-type
enzyme, ASY92V displayed a 2-fold increase of KM, while the
turnover number kcat was reduced by approximately two orders
of magnitude.14 The catalytic efficiency of ASY92V was
relatively low (51.65 M21 s21). However, even wild-type AS is
not a fast enzyme (kcat/KM = 13,043 M21 s21 14
) and it appears
that terpene cyclases have been optimised by nature for product
specificity rather than speed.
The hexane extractable materials produced by ASY92V were
also analysed by GC–MS, which indicated that despite its low
activity ASY92V produced terpenes specifically.17 The GC-
trace revealed that the mutant enzyme produced several
products of mass 204 (Fig. 1). The minor products 5, 6, 7 and 8,
which had already been observed in the case of ASY92F,14
constituted only 18% of the total amount of the hexane
extractable material (Table 1). They were identified by
Scheme 1 Formation of 3 and 4 during ASY92V catalysis.
† Electronic supplementary information (ESI) available: GC-profiles of co-
injections of 4 from ASY92V and authentic material and mass spectra of 4
produced by ASY92V and an authentic sample. See http://www.rsc.org/
suppdata/cc/b2/b206517g/
Fig. 1 GC-trace of hydrocarbons produced during ASY92V catalysis.
2384
CHEM. COMMUN., 2002, 2384–2385
This journal is © The Royal Society of Chemistry 2002

Table 1 Relative amounts of products formed from 1 by wild-type AS and ASY92V17
Product number
Name
2
3
4
5
6
7
8
Germacrene A
Aristolochene
b-(E)-Farnesene Valencene
b-Selinene
Selina-4,11-diene a-Selinene
Retention time /min
AS14
ASY92V
21.37
7.5
11.2
20.68
91.5
28.2
19.78
0
42.6
20.92
0.4
6.1
20.81
0
6.9
20.37
0
4.0
21.03
0
1.0
unaltered DDVIE motif located at the top of the active site. This
motif binds the pyrophosphate group of the substrate through
the catalytically important Mg2+ ion. However, the conforma-
tion of the hydrophobic portion of 1 within the active site of
ASY92V appeared to be altered leading to preferred deprotona-
tion of the methyl group on C3 rather than cyclisation through
the double bond at C10–C11. Tyr 92 of AS therefore controls
the reactivity of 1 by forcing C1 and C10 into close proximity
thereby favouring cyclisation to 2.
This work was financially supported by the BBSRC
(6/B17177).
Notes and references
‡ We thank Prof. John A. Pickett, FRS and Dr. Lynda Ireland (Institute of
Arable Crops Research, Rothamsted, Harpenden, Hertfordshire) for 4; Dr.
Larry Cool (Forest Products Laboratory, University of California, Berkeley)
for 2, 6 and 8 and Mr T. Cannon of De Monchy Aromatics Ltd. for 5.
Scheme 2 Formation of products 5, 6, 7 and 8 during ASY92V catalysis.
comparison of their mass spectra with those in the Wiley and
NIST libraries18 and by co-injections with authentic valencene
(5) and a- (8) and b-selinene (6).†‡ They are the products of
erroneous deprotonations (Scheme 2). Germacrene A is known
to undergo relatively facile acid-catalysed cyclisation to
selinenes in solvents such as chloroform.19,20 However, no
evidence for such a rearrangement was obtained in hexane,
which was the solvent in our studies.21 Hydrocarbons 6, 7, and
8 appeared therefore to be true enzymatic products of ASY92V.
The relative amount of 2 was slightly increased when compared
to the wild-type enzyme (Table 1).14 2 was identified from its
mass-spectrum and by comparison with authentic mate-
1 J. S. Glasby, Encyclopedia of Terpenoids, Wiley, Chichester, 1982.
2 D. E. Cane, Chem. Rev., 1990, 9, 1089–1103.
3 T. M. Hohn and R. D. Plattner, Arch. Biochem. Biophys., 1989, 272,
137–143.
4 R. H. Proctor and T. M. Hohn, J. Biol. Chem., 1993, 268, 4543–4548.
5 J. M. Caruthers, I. Kang, M. J. Rynkiewicz, D. E. Cane and D. W.
Christianson, J. Biol. Chem., 2000, 275, 25533–25539.
6 D. E. Cane, P. C. Prabhakaran, E. J. Salaski, P. H. M. Harrison, H.
Noguchi and B. J. Rawlings, J. Am. Chem. Soc., 1989, 111,
8914–8916.
7 D. E. Cane, P. C. Prabhakaran, J. S. Oliver and D. B. McIlwaine, J. Am.
Chem. Soc., 1990, 112, 3209–3210.
8 D. E. Cane and C. Bryant, J. Am. Chem. Soc., 1994, 116,
12063–12064.
rial.14,22
‡
While aristolochene still constituted approximately 28% of
the total amount of hexane extractable products a new major
product was formed with a retention time of 19.78 min.
Comparison of the mass-spectrum of this product with the
entries in the Wiley and NIST libraries18 suggested that this
material was b-(E)-farnesene. Co-injection of authentic b-(E)-
farnesene‡ and the products produced by ASY92V led to an
increase in the intensity of the peak corresponding to 4 which
made up almost 43% of the products generated by ASV92V.
The replacement of a single amino acid in the active site of AS
had therefore converted the enzyme into a b-(E)-farnesene
synthase. Both the catalytic activity and the specificity of the
mutant enzyme were reduced. However, it is noteworthy in this
context that many natural terpene cyclases also have relatively
low specificities. Multiple product formation may be a
consequence of the electrophilic reaction mechanism employed
by these enzymes in which highly reactive carbocationic
intermediates are generated. The most extreme cases reported
so far are d-selinene synthase and g-humulene synthase, for
which the major name-giving products represent only just over
25% of the total amount of products formed.23 b-(E)-farnesene
synthase isolated from peppermint has been shown to produce 4
as a major product.10 However, it also produced approximately
7% of cyclic sesquiterpenes of the cadinene-type.
9 D. E. Cane and Y. S. Tsantrizos, J. Am. Chem. Soc., 1996, 118,
10037–10040.
10 J. Crock, M. Wildung and R. Croteau, Proc. Natl. Acad. Sci USA, 1997,
94, 12833–12838.
11 M. Nonotny, S. Harvey and B. Jemiolo, Experentia, 1990, 46,
109–113.
12 J. Bohlmann, G. Meyer-Gauen and R. Croteau, Proc. Natl. Acad. Sci.
USA, 1998, 95, 4126–4133.
13 C. M. Starks, K. Back, J. Chappell and J. P. Noel, Science, 1997, 277,
1815–1820.
14 M. J. Calvert, P. R. Ashton and R. K. Allemann, J. Am. Chem. Soc., in
press.
15 Plasmid pZW04 was a gift from Professor David E. Cane, Brown
University, Rhode Island.
16 ASY92V was assayed in a total volume of 250 ml (10 mM Tris, pH 7.5,
5 mM MgCl2, 5 mM 2-mercaptoethanol, 15% glycerol (v/v) and
0.8–160 mM of [1-3H]-FPP). The reaction was initiated by addition of 50
ml of enzyme (2.5 mM final concentration), terminated by the addition of
200 ml of 100 mM EDTA, pH 7.25 and extracted with 1.8 ml of n-
hexane in the presence of 0.5 g of silica.
17 GC–MS analysis was performed as described14
.
18 Wiley Mass Spectral Database # 130986.
19 A. J. Weinheimer, W. W. Youngblood, P. H. Washecheck, T. K. B.
Karns and L. S. Ciereszko, Tetrahedron Lett., 1970, 7, 497–500.
20 W. S. Bowers, C. Nishino, M. E. Montgomery, L. R. Nault and M. W.
Nielson, Science, 1977, 196, 680–681.
21 M. Seemann, G. Zhai, K. Umezawa and D. Cane, J. Am. Chem. Soc.,
1999, 121, 591–592.
22 K. A. Rising, C. M. Starks, J. P. Noel and J. Chappell, J. Am. Chem.
Soc., 2000, 122, 1861–1866.
23 C. L. Steele, J. Crock, J. Bohlman and R. Croteau, J. Biol. Chem., 1998,
273, 2078–2089.
The results reported here establish that Tyr 92 in aris-
tolochene synthase plays a pivotal role in forcing 1 into the
reactive conformation necessary for efficient cyclisation to 2.
When the relatively bulky residue is replaced with the smaller
isopropyl group of valine, 4 is produced as the major product.
The mutant enzyme still bound FPP efficiently through the
CHEM. COMMUN., 2002, 2384–2385
2385