Bifunctional abietadiene synthase: Free diffusive transfer of the (+)-copalyl diphosphate intermediate between two distinct active sites

Article abstract of DOI:10.1021/ja010670k

Abietadiene synthase (AS) catalyzes two sequential, mechanistically distinct cyclizations in the conversion of geranylgeranyl diphosphate to a mixture of abietadiene double bond isomers as the initial step of resin acid biosynthesis in grand fir (Abies grandis). The first reaction converts geranylgeranyl diphosphate to the stable bicyclic intermediate (+)-copalyl diphosphate via protonation-initiated cyclization. In the second reaction, diphosphate ester ionization-initiated cyclization generates the tricyclic perhydrophenanthrene-type backbone, and is directly coupled to a 1,2-methyl migration that generates the C13 isopropyl group characteristic of the abietane family of diterpenes. Using the transition-state analogue inhibitor 14,15-dihydro-15-azageranylgeranyl diphosphate, it was demonstrated that each reaction of abietadiene synthase is carried out at a distinct active site. Mutations in two aspartate-rich motifs specifically delete one or the other activity and the location of these motifs suggests that the two active sites reside in separate domains. These mutants effectively complement each other, suggesting that the copalyl diphosphate intermediate diffuses between the two active sites in this monomeric enzyme. Free copalyl diphosphate was detected in steady-state kinetic reactions, thus conclusively demonstrating a free diffusion transfer mechanism. In addition, both mutant enzymes enhance the activity of wild-type abietadiene synthase with geranylgeranyl diphosphate as substrate. The implications of these results for the kinetic mechanism of abietadiene synthase are discussed.

Full text of DOI:10.1021/ja010670k

8974
J. Am. Chem. Soc. 2001, 123, 8974-8978
Bifunctional Abietadiene Synthase: Free Diffusive Transfer of the
(+)-Copalyl Diphosphate Intermediate between Two Distinct Active
Sites
Reuben J. Peters,† Matthew M. Ravn,‡ Robert M. Coates,‡ and Rodney B. Croteau*,†
Contribution from the Institute of Biological Chemistry, Washington State UniVersity, Pullman, Washington
99164-6340, and Department of Chemistry, UniVersity of Illinois, 600 South Matthews AVenue, Urbana,
Illinois 61801
ReceiVed March 13, 2001
Abstract: Abietadiene synthase (AS) catalyzes two sequential, mechanistically distinct cyclizations in the
conversion of geranylgeranyl diphosphate to a mixture of abietadiene double bond isomers as the initial step
of resin acid biosynthesis in grand fir (Abies grandis). The first reaction converts geranylgeranyl diphosphate
to the stable bicyclic intermediate (+)-copalyl diphosphate via protonation-initiated cyclization. In the second
reaction, diphosphate ester ionization-initiated cyclization generates the tricyclic perhydrophenanthrene-type
backbone, and is directly coupled to a 1,2-methyl migration that generates the C13 isopropyl group characteristic
of the abietane family of diterpenes. Using the transition-state analogue inhibitor 14,15-dihydro-15-
azageranylgeranyl diphosphate, it was demonstrated that each reaction of abietadiene synthase is carried out
at a distinct active site. Mutations in two aspartate-rich motifs specifically delete one or the other activity and
the location of these motifs suggests that the two active sites reside in separate domains. These mutants effectively
complement each other, suggesting that the copalyl diphosphate intermediate diffuses between the two active
sites in this monomeric enzyme. Free copalyl diphosphate was detected in steady-state kinetic reactions, thus
conclusively demonstrating a free diffusion transfer mechanism. In addition, both mutant enzymes enhance
the activity of wild-type abietadiene synthase with geranylgeranyl diphosphate as substrate. The implications
of these results for the kinetic mechanism of abietadiene synthase are discussed.
Introduction
Scheme 1. AS-Catalyzed Reactions and Aza Analogues of
Carbocation Intermediatesa
The secretion of oleoresin (a roughly equal mixture of
monoterpene olefins (turpentine) and diterpene resin acids
(rosin)1) is the primary response of conifers to wounding,
following which evaporation of the turpentine carrier results in
crystallization and polymerization of the nonvolatile resin acids
to seal the wound site.2 Abietadiene synthase (AS) catalyzes
the committed step of resin acid biosynthesis in grand fir (Abies
grandis) by cyclization of the universal diterpene precursor
geranylgeranyl diphosphate (GGPP, 1) to a mixture of abieta-
diene double bond isomers, which are subsequently oxidized
to the corresponding resin acids.3,4 To produce the tricyclic
perhydrophenanthrene (abietane) skeletal structure, AS carries
out two consecutive, mechanistically distinct cyclizations (Scheme
1).4 The first is initiated by protonation across the terminal 14,-
15-double bond of GGPP, producing the stable bicyclic
intermediate (+)-copalyl diphosphate [(+)-CPP, 2], in a reaction
analogous to that catalyzed by kaurene synthase A [(-)-copalyl
diphosphate synthase, CPS] of the gibberellin biosynthetic
pathway.5,6 In the second cyclization mediated by AS, diphos-
* Corresponding author: (phone) (509) 335-1790; (fax) (509) 335-7643;
(e-mail) croteau@wsu.edu.
† Washington State University.
‡ University of Illinois.
(1) Lewinsohn, E.; Savage, T. J.; Gijzen, M.; Croteau, R. Phytochem.
Anal. 1993, 4, 220-225.
a Aspartate-rich motifs involved in each cyclization reaction are also
depicted; amino acid (aa) numbering is based on the preprotein (i.e.
rAS starts with Met then Val 85 f Ala 868).
(2) Phillips, M. A.; Croteau, R. B. Trends Plant Sci. 1999, 4, 184-190.
(3) LaFever, R. E.; Stofer Vogel, B. S.; Croteau, R. Arch. Biochem.
Biophys. 1994, 131, 139-149.
phate ester ionization of (+)-CPP (2) forms the third ring in a
reaction corresponding to that mediated by kaurene synthase B
(4) Peters, R. J.; Flory, J. E.; Jetter, R.; Ravn, M. M.; Lee, H.-J.; Coates,
R. M.; Croteau, R. B. Biochemistry 2000, 39, 15592-15602.
10.1021/ja010670k CCC: $20.00 © 2001 American Chemical Society
Published on Web 08/25/2001

Bifunctional Abietadiene Synthase
J. Am. Chem. Soc., Vol. 123, No. 37, 2001 8975
(kaurene synthase, KS).5,6 However, in AS this cyclization is
immediately coupled, via intramolecular proton transfer7 within
a sandaracopimarenyl intermediate (3),8 to a 1,2-methyl migra-
tion that creates the C13 isopropyl group characteristic of
abietanes, such as abietadiene (4).
Scheme 2. Synthesis of 13,14-Dihydro-13-azacopalyl
Diphosphate from Manool in Eight Steps
Both native and recombinant AS (rAS)9 are bifunctional in
catalyzing both reactions, in contrast to the formation of (-)-
kaurene in higher plants which requires the two distinct
enzymes, CPS and KS.5,6 However, an unrelated fungal
bifunctional kaurene synthase (FCPS/KS) catalyzing both cy-
clizations has been identified and shown to contain two distinct
active sites.10 AS contains a DXDD motif conserved in CPS11
and triterpene cyclases12,13 that is involved in the protonation-
initiated cyclization reaction, as well as a DDXXD motif also
found in KS14 and other terpene synthases15,16 that is involved
in diphosphate ionization-initiated cyclization. FCPS/KS con-
tains similar motifs and mutation of either aspartate-rich motif
prevents the corresponding cyclization.17
In the formation of kaurene, the intermediate (-)-CPP is
channeled between CPS and KS, which form a functional
heterodimer.6 FCPS/KS also seems to exhibit intermediate
channeling; (E,E,E)-geranylgeranyl diphosphate (GGPP) reacts
to form kaurene at a faster rate than does (-)-CPP.17 In this
study, we demonstrate that AS also contains distinct active sites
responsible for each cyclization step that are defined, at least
in part, by the two respective aspartate-rich DXDD and DDXXD
motifs. The location of these motifs in a modeled AS structure
suggests that the two active sites reside in separate domains
(see Figure 4), similar to the organization of FCPS/KS.17
However, unlike the kaurene synthases, the (+)-CPP intermedi-
ate of AS is not channeled between the two active sites in this
monomeric enzyme, but rather is transferred by a free diffusion
mechanism. Previous kinetic analysis of AS demonstrated that
the intrinsic rate-limiting step of the overall reaction resides
after formation of (+)-CPP, although substrate inhibition
prevents direct observation of this limiting step with GGPP as
substrate.4 Based in part on the mistaken assumption that FCPS/
KS exhibited similar GGPP substrate inhibition,17 we had
suggested that nonproductive binding of GGPP at the second
active site limited the overall reaction.4 Here we demonstrate
that GGPP inhibits its own cyclization, not that of (+)-CPP.
described.3,9 The preparations of (E,E,E)-[1-3H]geranylgeranyl diphos-
phate (120 Ci/mol)3 and (+)-[1-3H]CPP (120 Ci/mol)4 have been
previously described. Kinetic assays with freshly prepared enzyme were
performed and the data analyzed as previously described.4 Micromolar
inhibition constants were determined by analysis of the change in KM
upon inclusion of inhibitor. However, because the kinetic assay requires
a minimum of 3 nM rAS, it is not possible to satisfy the constraints of
the Michaelis-Menten derivation, so determination of nanomolar
inhibition constants required use of an alternative experimental protocol.
A Dixon-type experiment was utilized in which reaction rates were
determined at various inhibitor concentrations in the presence of a fixed
concentration of substrate. The relative rates were fitted to a tight
binding inhibitor equation18 to obtain apparent Ki. Acid-catalyzed
solvolysis of reaction products was achieved by the addition of HCl to
0.3 M to the reaction mixture, followed by incubation for 15 min at
room temperature and pentane extraction.
Synthesis of 13,14-Dihydro-13-azacopalyl Diphosphate (Scheme
2). Manool (7) was degraded to acid 8 by a three-step literature
procedure that involved oxidation (pyridinium chlorochromate, CH2-
Cl2)19 to a 2:1 mixture of (E)- and (Z)-copalal followed by a retro-
aldol cleavage (K2CO3, H2O/tetrahydrofuran, reflux, 2d, 66% for two
steps)20 and iodoform oxidation of the resulting methyl ketone (KOH,
KI, I2, 53%). Modified Curtius rearrangement21 to the isocyanate
(diphenylphosphoryl azide, benzene, reflux, 73%) followed by reduction
provided methylamine 9 (LiAlH4, tetrahydrofuran, reflux and room
temperature, 62%), which was alkylated in quantitative yield with tert-
butyl bromoacetate (tetrahydrofuran, Et3N, 0 °C, 96%) and reduced to
amino alcohol 10 (LiAlH4, Et2O, 86%). Conversion to the diphosphate
ester followed published procedures developed for the diphosphorylation
of 1,2-amino alcohols.22 Thus, alcohol 10 was converted to the mesylate
at low temperatures (mesyl chloride, Et3N, CH2Cl2, -13 °C), followed
by displacement with tris(tetrabutylammonium) hydrogen pyrophos-
phate (CH3CN, 8.5 h). Counterion exchange on Dowex 50W-X8 (NH4+
form) gave a 10:3:1 mixture of ammonium salts containing inorganic
pyrophosphate, diphosphate 6, and an unknown product that exhibited
a singlet in the 31P NMR spectrum. Purification by preparative reverse
phase HPLC afforded 13-aza-CPP (6, 40% overall yield), which was
Experimental Procedures
Materials and General Procedures. Liquid scintillation counting
and product analysis by GC-MS were carried out as previously
(5) Saito, T.; Abe, H.; Yamane, H.; Sakurai, A.; Murofushi, N.; Takio,
K.; Takahashi, N.; Kamiya, Y. Plant Physiol. 1995, 109, 1239-1245.
(6) Duncan, J. D.; West, C. A. Plant Physiol. 1981, 68, 1128-1134.
(7) Ravn, M. M.; Coates, R. M.; Jetter, R.; Croteau, R. J. Chem. Soc.,
Chem. Commun. 1998, 1998, 21-22.
(8) Ravn, M. M.; Coates, R. M.; Flory, J.; Peters, R. J.; Croteau, R.
Org. Lett. 2000, 2, 573-576.
1
characterized by its H and 31P NMR spectra, the latter showing the
(9) Stofer Vogel, B.; Wildung, M. R.; Vogel, G.; Croteau, R. J. Biol.
Chem. 1996, 271, 23262-23268.
distinctive pair of doublets associated with monoalkyl diphosphates.
The Supporting Information contains a more complete description of
the synthetic procedures and characterization data.
Mutant Construction and Preparation. Point mutations of rAS
were generated by a variation on a previously described polymerase
chain reaction (PCR)-based site directed mutagenesis protocol.23 Two
mutated fragments were produced by PCR (40 cycles with annealing
(10) Kawaide, H.; Imai, R.; Sassa, T.; Kamiya, Y. J. Biol. Chem. 1997,
272, 21706-21712.
(11) Sun, T.-P.; Kamiya, Y. Plant Cell 1994, 6, 1509-1518.
(12) Poralla, K. In ComprehensiVe Natural Products Chemistry: Iso-
prenoids Including Carotenoids and Steroids; Cane, D. E., Ed.; Elsevier:
Oxford, 1999; Vol. 2, pp 307-319.
(13) Abe, I.; Prestwich, G. D. In ComprehensiVe Natural Products
Chemistry: Isoprenoids Including Carotenoids and Steroids; Cane, D. E.,
Ed.; Elsevier: Oxford, 1999; Vol. 2, pp 267-298.
(14) Yamaguchi, S.; Saito, T.; Abe, H.; Yamane, H.; Murofushi, N.;
Kamiya, Y. Plant J. 1996, 10, 101-111.
(15) Davis, E. M.; Croteau, R. In Topics in Current Chemistry; Leeper,
F., Verderas, J., Eds.; Springer-Verlag: Berlin, 2000; Vol. 209, pp 53-95.
(16) Bohlmann, J.; Meyer-Gauen, G.; Croteau, R. Proc. Natl. Acad. Sci.
U.S.A. 1998, 95, 4126-4133.
(17) Kawaide, H.; Sassa, T.; Kamiya, Y. J. Biol. Chem. 2000, 275, 2276-
2280.
(18) Williams, J. W.; Morrison, J. F. Methods Enzymol. 1979, 63, 437-
467.
(19) Sundararaman, P.; Herz, W. J. Org. Chem. 1977, 42, 813-819.
(20) Drengler, K. Ph.D. Thesis, University of Illinois, Urbana, 1980.
(21) Ninomiya, K.; Shiori, T.; Yamada, S. Tetrahedron 1974, 30, 2151-
2157.
(22) Steiger, A.; Pyun, H. J.; Coates, R. M. J. Org. Chem. 1992, 57,
3444-3449.
(23) Shalk, M.; Croteau, R. Proc. Natl. Acad. Sci. U.S.A. 2000, 97,
11948-11953.

8976 J. Am. Chem. Soc., Vol. 123, No. 37, 2001
Peters et al.
Figure 1. Differential inhibition of rAS by 15-aza-GGPP (5) deter-
mined at 5 µM GGPP (9) or 3 µM CPP (b). Data are the averages of
duplicate assays and the error bars represent the standard deviations.
at 55 °C), using an external 5′-primer (T7 promoter) with a reverse
mutagenic primer, and a complementary forward mutagenic primer with
an external 3′-primer (T7 terminator). Full-length mutagenized rAS was
produced using an overlapping fragment PCR scheme; the fragments
overlap in the mutagenic primer region and were joined in an initial
unprimed PCR reaction (25 cycles with annealing at 60 °C) and
amplified by dilution (1:1000) into a T7 promoter/T7 terminator primed
PCR reaction (40 cycles with annealing at 55 °C). NdeI/BamHI
digestion allowed ligation back into the original expression vector
pSBETa.24 The resulting sequence verified constructs were expressed
in E. coli BLR and the resulting recombinant enzymes were purified
as previously described.4 The concentration of purified rAS was
determined by absorbance at 280 nm using the calculated extinction
coefficient (138 350 M-1 cm-1).16
Figure 2. GC-MS analysis (total ion chromatogram) of the reaction
products (following acid-catalyzed solvolysis) of rAS assayed at 4 µM
GGPP under steady-state kinetic conditions. The products indicated
are geranyllinalool (1, from the rearrangement of GGPP), manool (2,
from rearrangement of (+)-CPP), abieta-7(8),13(14)-diene (3, all
isomeric olefins rearrange to abieta-7(8),13(14)-diene), and geranylge-
raniol (4a (cis-isomer) and 4b (trans-isomer) from solvolysis of GGPP).
corresponding D404A point mutant of rAS exhibits similar
behavior in abolishing the conversion of GGPP to (+)-CPP and
leaving the reaction of (+)-CPP to products unaffected (see
Figure 3a). In a similar fashion, alanine substitution of the first
aspartate in the DDXXD motif has been shown to selectively
eliminate diphosphate ionization-dependent reactions in FCPS/
KS17 as well as in other terpene synthases.15 Again, the
analogous point mutant of rAS, D621A, selectively eliminates
the diphosphate ionization-dependent reaction but not the
conversion of GGPP to (+)-CPP (i.e., phosphatase hydrolysis
of the reaction product of D621A with GGPP as substrate yields
copalol by GC-MS comparison to an authentic sample), which
occurs with comparable kinetics to the wild-type enzyme (see
Figure 3b). Co-incubation of D404A and D621A with GGPP
results in the formation of abietadienes at a rate equal to that of
wild-type rAS. Since AS is monomeric (native AS has been
previously shown to be monomeric,3 and gel filtration chroma-
tography (Superdex 200, Pharmacia) indicates that rAS (cal-
culated size ) 94 kDa) has a size of 90 ( 8 kDa, demonstrating
that the recombinant enzyme is also monomeric), this kinetically
effective complementation between mutants defective in the
partial reactions suggests that (+)-CPP diffuses freely between
the two active sites.
Detection of Free (+)-CPP in the Steady-State Reaction.
If (+)-CPP freely diffuses between the two active sites of AS,
the solution concentration of this intermediate must increase
sufficiently to promote productive rebinding at the second active
site. Therefore, under steady-state reaction conditions, free (+)-
CPP should be detectable in reactions with the wild-type
enzyme. For this analysis, acid hydrolysis was utilized both to
stop the enzymatic reaction and to hydrolyze the resulting
products, which were verified by GC-MS comparison with
authentic standards. Free (+)-CPP was easily detectable under
kinetic steady-state conditions at low initial concentrations of
GGPP (Figure 2). Quantitative analysis further indicated that
∼2% of the GGPP had been converted to (+)-CPP under these
conditions, representing a 20-fold excess over the amount of
rAS present. Failure to detect appreciable amounts of free (+)-
CPP at higher initial concentrations of GGPP was instructive;
this result indicated that the observed substrate inhibition with
GGPP arises from suppression of GGPP cyclization, rather than
by GGPP binding to and blocking the second active site as
previously suggested (since this would lead to accumulation of
(+)-CPP).4
Results
Transition-State Aza Analogues. To discriminate between
possibly discrete active sites responsible for protonation-initiated
versus diphosphate ionization-initiated cyclization, we synthe-
sized aza analogues corresponding to carbocation transition state
intermediates predicted to occur in the AS-mediated reactions
(Scheme 1). Synthesis of 14,15-dihydro-15-azageranylgeranyl
diphosphate (15-aza-GGPP, 5) has been previously reported.25
13,14-Dihydro-13-azacopalyl diphosphate (13-aza-CPP, 6) was
synthesized from manool in eight steps with an overall yield of
5.1% (Scheme 2).
13-Aza-CPP (6) inhibits the rAS-catalyzed (rAS ) the
recombinant “pseudomature” enzyme of AS) conversion of both
[3H]GGPP (KiGGPP ) 0.5 ( 0.3 µM) and [3H]CPP (KiCPP ) 0.2
( 0.05 µM) to [3H]abietane olefins to an approximately equal
extent. This result is consistent with the AS sequential cycliza-
tion mechanism, as 6 is expected to inhibit the second cyclization
and thus equally inhibit both reactions, irrespective of the
relationship of active sites. In contrast, 15-aza-GGPP (5) mimics
the initial carbocation intermediate of the protonation-initiated
cyclization and should preferentially inhibit the first reaction.
In fact, 5 does inhibit the reaction with GGPP dramatically more
effectively than that with (+)-CPP (Figure 1). At concentrations
where rAS is completely inhibited (with GGPP as substrate),
the reaction with (+)-CPP is unaffected. The 5 orders of
magnitude difference in inhibition (KiGGPP ) 0.25 ( 0.1 nM;
KiCPP ) 20 ( 4 µM) clearly demonstrates that each cyclization
is carried out at indiVidually accessible, and therefore separate,
actiVe sites.
Site-Directed Mutagenesis of Aspartate-Rich Motifs De-
fines Active Sites. To define each active site, PCR site-directed
mutagenesis was employed to create point mutants in either
aspartate-rich motif thought to be involved in each specific
reaction (Scheme 1). Alanine substitution for the second
aspartate in the DXDD motif selectively abolishes protonation-
initiated cyclization in FCPS/KS17 and triterpene cyclases.12 The
(24) Schenk, P. M.; Baumann, S.; Mattes, R.; Steinbiss, H.-H. BioTech-
niques 1995, 19, 196-200.
Kinetic Analysis. Freshly prepared recombinant AS (rAS)
(25) Ravn, M. M.; Jin, Q.; Coates, R. M. Eur. J. Org. Chem. 2000, 1401-
GGPP
1410.
was employed for kinetic analyses, which provided KM
)

Bifunctional Abietadiene Synthase
J. Am. Chem. Soc., Vol. 123, No. 37, 2001 8977
at ∼100 nM D621A, just short of the maximum rate obtained
with (+)-CPP, and decreases thereafter.
The ability of D404A to increase the observed rate of the
overall reaction with GGPP as substrate simply reflects the
increased capacity provided by the mutant to utilize excess free
(+)-CPP produced by wild-type rAS. Thus, increasing the
amount of D404A ultimately exceeds the ability of the first
active site (in either wild type or D621A) to produce (+)-CPP.
Again, this coupling effect saturates at ∼100 nM D404A,
seemingly just short of the maximum rate. This coupled assay
allows approximation of the true rate of the protonation-initiated
cyclization as kcat ≈ 3 s-1 (Figure 3b), which is ∼35% faster
than the rate of the reaction with (+)-CPP and which confirms
that the intrinsic rate-limiting step in the overall reaction resides
after formation of (+)-CPP. This approach also reveals that
D621A exhibits very similar kinetic behavior to wild-type rAS
under the same conditions (i.e., KM ≈ 5 µM and Ksi ≈ 6 µM;
Figure 3b), demonstrating the specific involvement of the
DDXXD motif in the diphosphate ester ionization step. The
coupled assays also exhibit substrate inhibition by GGPP,
thereby directly demonstrating that GGPP inhibits its own
cyclization. Consistent with this interpretation, the reaction of
D404A with (+)-CPP is only weakly inhibited by GGPP, and
the inhibition constants for GGPP and 15-aza-GGPP are similar
(KiCPP ∼23 and ∼20 µM, respectively), and are substantially
higher than the observed substrate inhibition constant (Ksi ∼5
µM) for the overall reaction. Finally, if GGPP also inhibited
the diphosphate ionization-dependent reaction, then (+)-CPP
levels would be expected to increase with time, resulting in a
reaction rate increase that is contrary to the observed kinetic
behavior.
Figure 3. Steady-state kinetic analyses of rAS. (A) Comparison of
wild-type rAS (b) and D404A (0) with (+)-CPP as substrate. (B)
Kinetics for GGPP as substrate with wild-type rAS alone (b) and for
wild-type (O) or D621A (0) with saturating concentrations (100 nM)
of D404A. Data are the averages of duplicate assays and the error bars
represent the standard deviations.
CPP
3 ( 2 µM and KM
) 0.4 ( 0.2 µM, kcat of 2.2 ( 0.3 s-1
(for both substrates) and a GGPP substrate inhibition constant
(Ksi) of 5 ( 2 µM (these kinetic constants differ somewhat from
those previously reported4 due to our earlier failure to appreciate
the instability of rAS stored under cryogenic conditions).
Although the calculated kcat is the same for either substrate, the
obserVable reaction rate with GGPP never approaches the level
attainable with (+)-CPP due to substrate inhibition as previously
noted (Figure 3).4
Discussion
Each active site of AS was independently characterized by
the use of the generated mutants that lack one or the other
activity. D404A is completely unreactive with GGPP (at least
104 decrease in kcat) while retaining the ability to react with
(+)-CPP. Kinetic analysis of D404A with (+)-CPP revealed
essentially identical behavior (kcat ) 2.3 ( 0.2 s-1; KM ) 0.4
( 0.2 µM) to wild-type rAS (Figure 3a), demonstrating a
specific role for the DXDD motif in the AS-catalyzed proto-
nation-initiated cyclization reaction. D621A cannot react with
(+)-CPP (again at least 104 decrease in kcat) but clearly cyclizes
GGPP to (+)-CPP at kinetically competent rates, as demon-
strated by the ability of this mutant to effectively complement
the activity of D404A with GGPP as substrate.
Bifunctional abietadiene synthase catalyzes two sequential
cyclization reactions to produce a mixture of abietadiene isomers
(Scheme 1).4 Protonation-initiated cyclization of the universal
diterpene precursor GGPP generates the stable bicyclic inter-
mediate (+)-CPP, which further undergoes diphosphate ester
ionization-initiated cyclization to a sandaracopimarenyl inter-
mediate that, via intramolecular proton transfer, is coupled to a
1,2-methyl migration;7,8 several alternative deprotonations then
yield the final mixture of products. The selective inhibition
displayed by the transition-state analogue 15-aza-GGPP (Scheme
1, 5), corresponding to the first carbocationic intermediate of
the catalytic cycle, demonstrated that each cyclization reaction
is carried out at a separate active site. Thus, at concentrations
of 5 where the cyclization of GGPP is completely inhibited,
(+)-CPP is readily converted to abietadienes (Figure 1), clearly
demonstrating the distinct nature of the two active sites.
The potent inhibition by 15-aza-GGPP (5) indicates a large
increase in the affinity for this transition-state analogue relative
to the substrate GGPP (KM/Ki ≈ 104, ∆∆Gq ) 5.5 kcal/mol),
consistent with transition-state stabilization by AS in the
protonation-initiated reaction. The magnitude of this effect
approximates that reported for inhibition by a similar transition-
state analogue (2-(dimethylamino)ethyl diphosphate) of isopen-
tenyl diphosphate isomerase, which also catalyzes a protonation-
initiated reaction.26,27 It is noteworthy that 5 is a much more
potent inhibitor than any of the numerous transition-state
analogues evaluated to date with triterpene cyclases which
catalyze similar protonation-initiated cyclization reactions, and
The addition of either D404A or D621A to wild-type rAS
increased the obserVed overall reaction rate with GGPP as
substrate. This seemingly contradictory ability of both mutants
to enhance overall catalysis (i.e., with GGPP as substrate) is a
consequence of AS containing two separate active sites, coupled
with substrate inhibition by GGPP and the free diffusive transfer
of the intermediate (+)-CPP. The intrinsic rate-limiting step of
the overall reaction resides after the protonation-initiated
cyclization of GGPP to (+)-CPP (i.e., the calculated kcat for
formation of abietadienes is the same for either substrate).
However, owing to substrate inhibition, the obserVed reaction
rate with GGPP never reaches the maximum rate attainable with
(+)-CPP (Figure 3). D621A simply increases the overall reaction
rate by converting GGPP to free (+)-CPP, thereby relieving
substrate inhibition and allowing the obserVed rate to approach
that with (+)-CPP as substrate. Thus, increasing the amount of
D621A in the presence of a constant amount of wild-type rAS
allows the rate of conversion of GGPP to approach that for an
equivalent concentration of (+)-CPP; this enhancement saturates
(26) Reardon, J. E.; Abeles, R. H. Biochemistry 1986, 25, 5609-5616.
(27) Muehlbacher, M.; Poulter, C. D. Biochemistry 1988, 27, 7315-
7328.

8978 J. Am. Chem. Soc., Vol. 123, No. 37, 2001
Peters et al.
in the observed overall rate suppression. First, the steady-state
concentration of free (+)-CPP decreases with increasing con-
centrations of GGPP and parallels the observed rate decrease
for the overall reaction. If GGPP inhibited the overall reaction
by binding at the second active site, then (+)-CPP would be
expected to accumulate over time. As an additional consequence,
such accumulation would result in the reaction rate increasing
with time as GGPP concentrations decreased and inhibition was
relieved; such kinetic behavior was not observed. Second, GGPP
and the 15-aza-GGPP analogue only weakly inhibit the reaction
of AS with (+)-CPP, at a concentration much higher (Ki ∼20
µM) than that observed for substrate inhibition of the overall
reaction (Ksi ∼5 µM). Third, kinetic analysis of the cyclization
of GGPP, utilizing a coupled assay approach with the mutant
enzyme (D404A) deficient in this reaction, directly demonstrated
that GGPP inhibition arises from interaction at the first active
site (Figure 3b).
Figure 4. Structural schematic of AS, as modeled from 5-epi-
aristolochene synthase,28 illustrating the locations of the aspartate-rich
motifs, transit sequence (T.S.), and the unusual insertion element that
is also found in a few other terpenoid synthases.16
which also contain a catalytically important DXDD motif.12,13
In contrast, the inhibition exhibited by 13-aza-CPP (6) on the
conversion of CPP to abietadienes is essentially the same as
the binding affinity of (+)-CPP (KMCPP/KiCPP ) 2). This bicyclic
aminodiphosphate (6) behaves as an inert substrate mimic rather
than as a true transition-state analogue.
The involvement of two distinct active sites in bifunctional
AS of grand fir is similar to the arrangement of active sites in
the bifunctional kaurene synthase of fungal origin10,17 and
reminiscent of the requirement for two separate enzymes in the
production of (-)-kaurene in higher plants.5,6 However, in
contrast to the observed channeling of the (-)-CPP intermediate
in the biosynthesis of kaurene by these enzymes, 5,6,17 the (+)-
CPP intermediate of AS freely diffuses between active sites.
This latter phenomenon was initially suggested by the surprising
finding that mutants specifically lacking in one or the other
cyclization activity could effectively complement each other
with GGPP as substrate, and was confirmed by the demonstra-
tion of excess free (+)-CPP intermediate under steady-state
kinetic conditions (Figure 2).
The two aspartate-rich motifs of AS were previously sug-
gested to be involved in the two cyclization reactions cata-
lyzed,4,9 based upon the occurrence of similar sequence elements
in mechanistically related terpenoid synthases.12,13,15 Mutations
in each aspartate cluster confirmed the specific requirement for
each in one of the two cyclizations catalyzed by AS. Sequence
alignments16 and comparison to the crystal structure of 5-epi-
aristolochene synthase28 indicate that the two active sites of AS
occur in structurally distinct domains (Figure 4). Since the active
site of most terpenoid synthases is considered to reside primarily
in the C-terminal domain, the present results with AS provide
an additional role29 for the N-terminal (glycosyl hydrolase-like)
domain that is found in all plant terpene synthases, although its
function is largely unknown. These structural and functional
observations with AS may be of evolutionary significance
because AS appears to be a close relative of the ancestral plant
terpenoid synthase based upon both protein sequence phylog-
eny16 and genomic organization.30 Thus, while the N-terminal,
glycosyl hydrolyase-like, domain is of functional utility in the
seemingly ancient AS, its retention in other plant terpenoid
synthase family members may be largely relectual.
Conclusions
The potent selective inhibition exhibited by the transition-
state analogue, 15-aza-GGPP, demonstrated that abietadiene
synthase carries out the two sequential, mechanistically distinct
reactions at separate active sites. The tight binding of 15-aza-
GGPP strongly indicates a role for transition-state stabilization
in the protonation-initiated (first) cyclization step. Surprisingly,
it was found that the (+)-CPP intermediate of the reaction
scheme is not channeled within this monomeric enzyme, but
freely diffuses between the two active sites. Mutational analysis
of the two aspartate-rich motifs has specifically defined the two
active sites and indicated that they are contained in structurally
distinct domains. Kinetic analysis with these mutants verified
that the intrinsic rate-limiting step in the overall reaction resides
after formation of (+)-CPP and demonstrated that GGPP inhibits
its own cyclization. Thus, AS binds GGPP at an active site, in
part defined by the DXDD motif, in the N-terminal domain
where it undergoes protonation-initiated cyclization to form (+)-
CPP. This intermediate must then diffuse out of this active site
and rebind at a distinct second active site, in part defined by
the DDXXD motif, in the C-terminal domain (either of the same
or a different molecule of AS). At this site, diphosphate ester
ionization initiates the second cyclization and the reaction cycle
is completed to yield the olefinic products.
Acknowledgment. This work was supported by National
Institutes of Health Grants GM31354 to R.B.C. and GM13956
to R.M.C., and by a Postdoctoral Fellowship from the Jane
Coffin Childs Memorial Fund for Medical Research to R.J.P.
We thank C. Burke for performing size analysis of rAS by gel
filtration chromatography.
A number of lines of evidence have been presented to
demonstrate that GGPP inhibits its own cyclization, resulting
Supporting Information Available: Experimental details
of 13,14-dihydro-13-azacopalyl diphosphate synthesis (PDF).
This material is available free of charge via the Internet at
http://pubs.acs.org.
(28) Starks, C. M.; Back, K.; Chappell, J.; Noel, J. P. Science 1997,
277, 1815-1820.
(29) Williams, D. C.; McGarvey, D. J.; Katahira, E. J.; Croteau, R.
Biochemistry 1998, 37, 12213-12220.
(30) Trapp, S. C.; Croteau, R. B. Genetics 2001, 158, 811-832.
JA010670K