Purification and characterization of cannabichromenic acid synthase from Cannabis sativa

Article abstract of DOI:10.1016/S0031-9422(98)00278-7

Cannabichromenic acid synthase was purified to apparent homogeneity by sequential column chromatography including DEAE-cellulose, phenyl-Sepharose CL-4B, and hydroxylapatite. The enzyme catalysed the oxidocyclization of cannabigerolic acid and cannabinerolic acid to cannabichromenic acid. The K(m) values for both substrates were in the same order of magnitude although the V(max) value for the former was higher than that for the latter. These results suggested that cannabichromenic acid is predominantly formed from cannabigerolic acid rather than cannabinerolic acid. The enzyme required neither molecular oxygen nor hydrogen peroxide, indicating that the cannabichromenic acid synthase reaction proceeds through direct dehydrogenation without hydroxylation.

Full text of DOI:10.1016/S0031-9422(98)00278-7

Phytochemistry Vol. 49, No. 6, pp. 1525±1529, 1998
# 1998 Elsevier Science Ltd. All rights reserved
Printed in Great Britain
0031-9422/98/$ - see front matter
PII: S0031-9422(98)00278-7
PURIFICATION AND CHARACTERIZATION OF
CANNABICHROMENIC ACID SYNTHASE FROM
CANNABIS SATIVA
SATOSHI MORIMOTO*, KENGO KOMATSU, FUTOSHI TAURA and YUKIHIRO SHOYAMA
Faculty of Pharmaceutical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka, 812-82,
Japan
(Received 16 March 1998)
Key Word IndexÐCannabis sativa; Moraceae; biosynthesis; cannabichromenic acid synthase;
cannabichromenic acid; cannabigerolic acid.
AbstractÐCannabichromenic acid synthase was puri®ed to apparent homogeneity by sequential column
chromatography including DEAE-cellulose, phenyl-Sepharose CL-4B, and hydroxylapatite. The enzyme
catalysed the oxidocyclization of cannabigerolic acid and cannabinerolic acid to cannabichromenic acid.
The Km values for both substrates were in the same order of magnitude although the Vmax value for the
former was higher than that for the latter. These results suggested that cannabichromenic acid is predomi-
nantly formed from cannabigerolic acid rather than cannabinerolic acid. The enzyme required neither mol-
ecular oxygen nor hydrogen peroxide, indicating that the cannabichromenic acid synthase reaction
proceeds through direct dehydrogenation without hydroxylation. # 1998 Elsevier Science Ltd. All rights
reserved
INTRODUCTION
order to evaluate the precise properties of CBCA
synthase, we attempted to purify the enzyme from
young leaves of C. sativa and characterize its bio-
chemical properties. In this report, we describe
the puri®cation and characterization of CBCA
synthase.
Cannabinoids are novel secondary metabolites
found only in marijuana (Cannabis sativa L.), and
more than sixty cannabinoids have been isolated
and structurally elucidated (Fig. 1). Among
them, the biological activities of the psychoactive
cannabinoid, D1-tetrahydrocannabinol have been
thoroughly investigated [1]. Other cannabinoids also
show a variety of interesting pharmacological activi-
ties. For example, cannabichromene (CBC) and its
related cannabinoids possess anti-in¯ammatory,
anti-fungal and anti-microbial e€ects [2], while the
analgetic e€ect induced by D1-tetrahydrocannabinol
is potentiated by CBC [3]. Thus, CBC has attracted
considerable attention, whereas little was known
about the biosynthesis of cannabichromenic acid
(CBCA), the precursor of CBC.
RESULTS AND DISCUSSION
Puri®cation of CBCA synthase
The puri®cation of CBCA synthase from the
cannabidiolic acid (CBDA) strain was carried out
using the extraction and assay methods as pre-
viously established [4]. Consequently, we developed
a four-step procedure that resulted in the puri®-
cation of CBCA synthase to homogeneity. The typi-
cal data from the puri®cation of CBCA synthase
are shown in Table 1. Since the presence of CBDA
synthase, which catalyses the formation of CBDA
from CBGA, was con®rmed in the CBDA strain [5],
its activity as well as CBCA synthase activity was
assayed to avoid the contamination of CBDA
synthase.
Recently, we have identi®ed an enzyme (CBCA
synthase) catalysing the highly stereoselective con-
version of cannabigerolic acid (CBGA) to CBCA in
young leaves of C. sativa [4]. However, since a
crude enzyme preparation was used to characterize
the properties of CBCA synthase, the mechanism of
CBCA biosynthesis has remained ambiguous. In
As a ®rst step, the soluble fraction from young
leaves of the CBDA strain was fractionated with
ammonium sulfate. More than 80% of the enzyme
*Author to whom correspondence should be addressed.
1525

1526
S. MORIMOTO et al.
Fig. 1. Structures of cannabinoids.
activity precipitated between 30 and 75% saturation Subunit mass, native mass and isoelectric point of
CBCA synthase
of ammonium sulfate, resulting in a 3-fold puri®-
cation. The CBCA-synthase-active fraction was
applied to a DEAE-cellulose column. The enzyme
activity did not bind to the matrix, and most of the
activity was recovered in the void volume. The void
volume fractions still displayed weak CBDA
synthase activity, but this chromatography yielded a
28-fold puri®cation. CBCA synthase was further
puri®ed by applying the DEAE-cellulose eluate
onto a phenyl-Sepharose CL-4B hydrophobic inter-
action column. Elution of the bound enzyme by a
SDS-PAGE of the puri®ed CBCA synthase
a subunit molecular mass of 71 kDa
showed
(Fig. 2). The native molecular mass was estimated
from the elution volume of CBCA synthase on
Sephacryl S-200 HR chromatography, where the
enzyme eluted as a single molecular species with a
molecular mass of 136 kDa. These results suggested
that CBCA synthase exists as a homodimer com-
posed of two identical 71 kDa subunits. On the
other hand, the pI for CBCA synthase was deter-
descending gradient of ammonium sulfate led to mined to be 7.3 by comparison with marker pro-
teins of known pI on isoelectrofocusing gels.
a 226-fold puri®cation. The CBCA-synthase-active
fractions showed no CBDA synthase activity,
although they were still contaminated with several
proteins (Fig. 2). Finally, the enzyme was puri®ed
to homogeneity by hydroxylapatite chromatography
using an increasing gradient of NaCl. At individual
steps of the puri®cation procedure, the intensity of
the 71 kDa band increased on SDS-PAGE as the
preparation was enriched in activity, and the hydro-
Substrate speci®city
When the puri®ed CBCA synthase was tested
with CBGA and cannabinerolic acid (CBNRA), the
Z-isomer of CBGA, it became clear that the enzyme
catalyses the formation of CBCA from both sub-
strates. Since these reactions followed Michaelis-
Menten kinetics in response to changes in the con-
centrations of CBGA and CBNRA, the Vmax and
Km values were determined by Lineweaver-Burk
double reciprocal plots (Fig. 3). The Vmax value
xylapatite fraction migrated as
a single band
(Fig. 2). CBCA synthase was the ®rst enzyme
involved in CBCA biosynthesis to be puri®ed.
Table 1. Puri®cation of CBCA synthase from Cannabis sativa (CBDA strain)*
Step
Protein
(mg)
Total activity
(pkat/mg)
Speci®c activity
(pkat/mg)
Recovery
(%)
Puri®cation
(fold)
100 000 g Supernatant
30±75% (NH4)2SO4
DEAE-cellulose (DE-52)
Pheny-Sepharose CL-4B
Hydroxylapatite
267
77.6
2.1
0.09
0.02
508
435
113
38
1.9
5.6
54
430
607
100
86
22
5.6
2.6
1
3
28
226
319
13
*Data are based on the extraction of 30 g of rapidly expanding leaves.

CBCA synthase from C. sativa
1527
kinetic data indicate that CBCA synthase has
higher anity for CBGA than CBDA synthase
and D1-THCA synthase. In contrast, the turnover
1
number (kcat=0.04 s
) was lower than those
reported for D1-THCA synthase (kcat=0.20 s 1) and
CBDA synthase (kcat=0.19 s
judging from the kcat/Km values, we concluded
1
) [5, 6]. However,
that CBGA is an ecient substrate for CBCA
biosynthesis because CBCA synthase displays
1
a
higher value (1717 M 1S
)
for CBGA than
CBDA (1492 M 1S
(1382 M 1S 1).
)
and D1-THCA synthases
1
Biochemical properties of CBCA synthase
CBCA synthase is an oxidoreductase that cata-
lyses the cyclization of CBGA. Hence the properties
of CBCA synthase were investigated using cofac-
tors, coenzymes and inhibitors of oxidoreductases.
As shown in Table 2, hydrogen peroxide apparently
inhibits the enzyme activity, indicating that CBCA
synthase is not a peroxidase. On the other hand,
CBCA synthase did not require molecular oxygen
for the oxidocyclization of CBGA, suggesting that
it is not an oxygenase like P-450. This was further
con®rmed by the fact that P-450 inhibitors such as
CO, KSCN [7] and triazole little a€ect CBCA
synthase activity. From these results, it became evi-
dent that the oxidocyclization of CBGA to CBCA
is not accompanied by hydroxylation. In addition,
radical formation was not observed during the
enzymatic oxidation, because the radical scavenger,
2-methyl-2-nitrosopropane did not signi®cantly
a€ect the enzyme activity [8]. Thus, the enzymatic
reaction was considered to proceed through direct
dehydrogenation without hydroxylation and radical
formation, although CBCA synthase did not require
the typical coenzymes (NAD, NADP, FAD and
FMN) of dehydrogenases. CBCA biosynthesis has
Fig. 2. SDS-PAGE analysis of CBCA synthase. Samples
from each puri®cation step were resolved by electrophor-
esis on a 12.5% acrylamide gel. Proteins were stained with
Coomassie Brilliant Blue. Lane 1, soluble fraction; lane 2,
ammonium sulfate precipitation; lane 3, DE-52 column;
lane 4, phenyl-Sepharose CL-4B column; lane 5, hydroxy-
lapatite column; lane 6, molecular standards with the indi-
cated molecular masses.
(0.63 nkat/mg) for CBGA was higher than that
(0.14 nkat/mg) for CBNRA, suggesting that CBCA
is predominantly biosynthesized from CBGA rather
than CBNRA. On the other hand, the Km values
for CBGA (23 mM) and CBNRA (33 mM) were
similar to each other, but lower than those of
CBDA synthase (134 mM for CBGA and 254 mM
for CBNRA) and D1-tetrahydrocannabinolic acid
(D1-THCA) synthase (137 mM for CBGA and
206 mM for CBNRA) that catalyse the formation
of CBDA and D1-THCA, respectively [5, 6]. These
Table 2. CBCA synthase activity under various
conditions*
Conditions
Relative activity
(%)
Standard
100
54
98
103
97
108
103
99
94
97
99
32
1 mM H2O2
N2
CO
2 mM Triazole
2 mM KSCN
2 mM 2-Methyl-2-nitrosopropane
1 mM NADPH
1 mM NAD
1 mM NADP
1 mM FAD
1 mM FMN
*
All assay mixtures contained 0.11 mg of puri®ed CBCA
Fig. 3. Kinetic analyses of CBCA synthase. CBCA
synthase activity was measured using various concen-
trations of CBGA (.) and CBNRA ( ). Double-reciprocal
plots of the velocity data are shown. The velocity data
were ®tted by the nonlinear least squares test to determine
the apparent Km and Vmax for both substrates.
synthase and 200 mM CBGA. The relative activity of 100%
is 0.6 nkat/mg, and the SD was always within 5% of the mean of
triplicate determinations.
Molecular oxygen in the assay mixture was removed by nitro-
gen purge for 5 min.
Assay solution was saturated with CO gas (purged for 5 min).

1528
S. MORIMOTO et al.
Fig. 4. Mechanism of CBCA biosynthesis by CBCA synthase.
hitherto been assumed to proceed via a hydroxy or analytical HPLC under the same conditions as
a dienone intermediate [9], but the properties of described previously [4]. The e€ects of metal ions
CBCA synthase apparently indicate that CBGA
does not cyclise to CBCA via the hydroxy inter-
mediate. Therefore, it seems likely that the conver-
sion of CBGA to CBCA involves the dienone
intermediate (Fig. 4). The formation of a chromen
ring by dienone cyclisation has also been reported
for the biosynthesis of the rotenoid deguelin, and it
was demonstrated that deguelin cyclase catalyses
the formation of deguelin from rot-2'-enoic acid via
a dienone intermediate [10]. It is notable that like
CBCA synthase deguelin cyclase is assumed not to
be a P450-type enzyme [8].
The metal ion requirement of CBCA synthase
was also examined. However, none of the metal
ions tested (Mg2+, Zn2+, Ca2+ and Cu2+, 2 mM
each) stimulated the enzyme activity, and Hg2+
(1 mM) completely inhibited the reaction. The
chelating agent EDTA slightly increased the enzyme
activity (10% stimulation), suggesting that the
enzymatic reaction does not require metal ions.
Oxidoreductases that require neither cofactors
nor coenzymes are rarely found in plants, but these
biochemical properties of CBCA synthase resemble
those of CBDA and D1-THCA synthases [5, 6].
Thus, these cannabinoid synthases are unique
oxidoreductases catalysing the cyclization of terpene
moiety. Although various monoterpene cyclases
that catalyse the cyclization of geranyl pyro-
phosphate have been con®rmed in several
plants [11±15], they are not oxidoreductases.
and EDTA were determined by dialysing an enzyme
preparation against bu€er A (10 mM Na-Pi bu€er
(pH 7.0), 3 mM 2-mercaptoethanol) containing
2 mM EDTA (12 hr), then against bu€er A alone
(12 hr). CBDA synthase activity was measured as
described previously [5].
Extraction and puri®cation of CBCA synthase.
Unless otherwise indicated, all extraction and puri®-
cation procedures were performed at 48. Young
leaves (30 g) of CBDA strain were homogenized in
a Waring blender at high speed together with
300 ml of extraction bu€er (10 mM Na-Pi bu€er
(pH 7), 1 M NaCl, 10 mM 2-mercaptoethanol).
After the homogenate was ®ltered through a Nylon
screen, the ®ltrate was centrifuged at 100 000 g for
1 hr. The supernatant was then fractionated with
(NH4)2SO4. Proteins precipitating at 30±75% sat
were collected by centrifugation at 20 000 g for
15 min, re-suspended in ca 30 ml of bu€er A and
dialysed overnight against three changes of the
same bu€er. Insoluble materials were removed by
centrifugation at 20 000 g for 15 min. The super-
natant was applied at a ¯ow rate of 1 ml/min to a
1.5 Â15 cm column of DE-52 cellulose equilibrated
with bu€er A. After washing with 150 ml of bu€er
A, the column was developed with bu€er A con-
taining 0.5 M NaCl. Elution was monitored photo-
metrically at 280 nm, and the eluent was collected
in 15 ml fractions. Most of CBCA synthase activity
was found in the fractions eluted with bu€er A.
The active fractions (fractions 2±3) from the DE-52
column were concd to 6 ml and brought to 0.7 M
(NH4)2SO4 by adding 6 ml of bu€er A containing
1.4 M (NH4)2SO4. The sample was passed over a
1 Â10 cm column containing phenyl-Sepharose CL-
4B equilibrated with bu€er B (bu€er A containing
0.7 M (NH4)2SO4). The column was rinsed with 3
column vols of the same bu€er, and bound proteins
were then eluted by a 300 ml linear gradient of
bu€er B to bu€er A at a ¯ow rate of 1 ml/min. The
most active fractions (fractions 10±12, each 15 ml)
were pooled, concd and dialysed overnight against
EXPERIMENTAL
Plant materials. Cannabis sativa (CBDA strain)
was grown in the herbal garden and greenhouse of
the Faculty of Pharmaceutical Sciences, Kyushu
University. CBCA synthase was extracted from the
young leaves of the 2-week-old CBDA strain.
Assays of enzyme activities. CBCA synthase ac-
tivity was assessed by incubating enzyme (50 ml) in
substrate bu€er (0.4 ml) containing 0.2 mM CBGA,
0.05% (w/v) SDS and 0.1 M Na-Pi (pH 6.5) at 308
for 2 hr [4]. After terminating the reaction with three changes of bu€er A. The dialysed sample was
450 ml of MeOH, a 50-ml aliquot was applied to applied to a 1 Â10 cm column containing hydroxyl-

CBCA synthase from C. sativa
1529
apatite equilibrated with bu€er A. After washing
with 3 column vols of the same bu€er, bound pro-
teins were eluted with a 200 ml linear gradient of
NaCl (0±1.5 M) at a ¯ow rate of 1 ml/min. The
most active fractions (fractions 14±16, each 8 ml)
were pooled, concd and dialysed overnight against
two changes of bu€er A. The purity of CBCA
synthase was con®rmed by SDS-PAGE analysis of
the dialysate.
Protein assay. Protein concentrations were
measured according to ref. [18] using bovine serum
albumin as the standard.
REFERENCES
1. Gaoni, R. and Mechoulam, R., Journal of the
American Chemical Society, 1964, 86, 1946.
2. ElSohly, H. N., Turner, C. E., Clark, A. M.
and ElSohly, M. A., Journal of Pharmaceutical
Sciences, 1982, 71, 1319.
Analysis of proteins by SDS-PAGE and isoelectric
focusing. SDS-PAGE was carried out with the sys-
tem of ref. [16] in a 12.5% acrylamide gel of
0.75 mm thickness. Samples were dialysed against
H2O, lyophilized, resuspended in SDS sample bu€er
and denatured in a steam bath (958) for 2 min
before loading. Gels were run at a constant current
of 30 mA for about 2 hr. Proteins were visualized
by Coomassie Brilliant Blue R-250 and silver stain-
ing. The subunit molecular mass of the enzyme was
determined by comparison with high molecular
mass protein standards. Isoelectric focusing was
conducted according to ref. [17] using 7.5 mm glass
tubes. The ampholyte gradient ranged from pH 4.0
to pH 7.0. After electrophoresis at 750 V for 2 hr,
proteins were visualized by Coomassie Brilliant
Blue R-250 staining. The pI of the puri®ed enzyme
was determined by comparison with marker
proteins.
3. Davis, M. W. and Hatoum, N. S., General
Pharmacology, 1983, 14, 247.
4. Morimoto, S., Komatsu, K., Taura, F. and
Shoyama, Y., Journal of Natural Products,
1997, 60, 854.
5. Taura, F., Morimoto, S. and Shoyama, Y.,
Journal of Biological Chemistry, 1996, 271,
17411.
6. Taura, F., Morimoto, S., Shoyama, Y. and
Mechoulam, R., Journal of the American
Chemical Society, 1995, 117, 9766.
7. Imai, Y. and Sato, R., European Journal of
Biochemistry, 1967, 1, 419.
8. Crombie, L., Rossiter, J. T., Bruggen Van, N.
and Whiting, D. A., Phytochemistry, 1992, 31,
451.
9. Mechoulam, R., Science, 1970, 168, 1159.
10. Bhandari, P., Crombie, L., Harper, M. F.,
Rossiter, J. T., Sanders, M. and Whiting, D. A.,
Journal of the Chemical Society Perkin
Transactions I, 1992, 0, 1685.
Determination of native molecular mass. The
native molecular mass of CBCA synthase was deter-
mined by gel ®ltration chromatography on
a
11. Croteau, R., Chemical Reviews, 1987, 87, 929.
12. Cane, D. E. and Pargellis, C., Archives of
Biochemistry and Biophysics, 1987, 254, 421.
13. Hohn, T. M. and Plattner, R. D., Archives of
Biochemistry and Biophysics, 1989, 272, 137.
14. Alonso, W. R. and Croteau, R., Archives of
Biochemistry and Biophysics, 1991, 286, 511.
15. Rajaonarivony, J. I. M., Gershenzon, J. and
Croteau, R., Archives of Biochemistry and
Biophysics, 1992, 296, 49.
16. Laemmli, U. K., Nature, 1970, 224, 680.
17. O'Farrell, P. H., Journal of Biological
Chemistry, 1975, 250, 4007.
18. Bradford, M. M., Analytical Biochemistry,
1976, 72, 248.
1.5 Â75 cm column of Sephacryl S-200 HR equili-
brated with bu€er A at a ¯ow rate of 0.3 ml/min.
Fractions of 15 ml were collected. Molecular mass
markers from 29 to 700 kDa were resolved under
the same conditions prior to running the fraction
containing CBCA synthase activity.
Substrate speci®city and kinetic parameters.
CBGA and CBNRA were tested as potential sub-
strates of CBCA synthase using the standard assay,
unless otherwise stated. The Km and Vmax values
for both substrates were determined by Lineweaver-
Burk double-reciprocal plots of the velocity curves
of the CBCA-producing reaction with an increasing
concn of substrates.