Organic Letters
Letter
1
6
upon GC−MS analysis. The H NMR spectrum (in CDCl )
3
of the minor product exhibited seven singlet methyl signals (δH
0
0
3
1
.664 (s, 3H), 0.764 (s, 3H), 0.796 (s, 3H), 0.826 (s, 6H),
.961 (s, 3H), 0.967 (s, 3H)), two hydroxymethine signals (δH
.185 (dd, J = 11.2, 4.8 Hz, 1H), 3.231 (dd, J = 11.5, 4.1 Hz,
H)), and an olefinic signal (δ 5.331 (m, 1H)), indicating a
H
13
presence of an olefin. The C NMR spectrum showed the
presence of two carbons bearing hydroxyl groups (δ 78.81,
C
7
9.14) and an olefin (δ 122.14, 138.18). These data were in
C
complete agreement with the literature data as well as the data
18
for the authentic sample of 6.
Therefore, LCE was demonstrated to be a novel onoceroid
synthase producing serratane-type triterpenes with 5 as the
major product together with 6 as the minor product in a ∼10:1
ratio. The structures of these LCE products suggested that the
LCE-catalyzed reaction of 4 was initiated by a proton attack on
the epoxide moiety, followed by bicyclization to give a C8
cation, a common cationic intermediate en route to 3 (Scheme
1). Subsequent cyclization from the neighboring C14−C27 exo-
methylene bridged the two decalin structures to construct a
seven-membered C-ring and gave a 6/6/7/6/6-fused pentacy-
clic C14 carbocation intermediate. This cation underwent a
nucleophilic addition of a water to produce 5, or alternatively,
deprotonation of H15 to yield 6. As with the case of LCD, LCE
preferably utilized 4 among other substrates such as 1 and 2
and catalyzed the formation of the fused cyclic structure on the
linear side chain of 4. These results indicated that in ferns,
onocerane and serratane, two of the major onoceroid, were
each individually produced by a different cyclase using 4 as a
common substrate. It is noteworthy that the serratane skeleton
was biosynthesized from a bicyclic triterpene such as 4. This
finding may disprove the previously accepted proposal that the
serratane formation involves protonation of one of the exo-
methylenes of 3 followed by a cyclization into a seven-
membered ring. Our novel finding proposed that serratane
formation did not require the formation of an intermediary 3
and was directly produced from 4.
Figure 2. GC−MS analysis of each enzymatic product: (i) standard
sample of 6, (ii) hexane extracts of samples expressing both LCC and
LCE, (iii) LCC, (iv) LCE, (v) empty vector. Compounds 5 and 6
were only seen in (ii).
hydrated 4. On the other hand, the minor product gave a
molecular ion peak at m/z 442 that was equivalent to 4. In
addition, the characteristic peaks at m/z 207 and 189 due to
retro Diels−Alder fragmentation of the D-ring were observed
1
6
for this compound.
To clarify the structures of these products, from a large-scale
culture (4 L), these products were purified by silica gel column
to yield 10 mg of the major product and 1 mg of the minor
1
product and then subjected to NMR analyses. The H NMR
spectrum (in CDCl ) of the major product showed the
3
presence of seven singlet methyls (δ 0.751 (s, 3H), 0.781 (s,
H
3
3
1
H), 0.786 (s, 3H), 0.932 (s, 6H), 0.946 (s, 3H), 0.973 (s,
H)) and two hydroxymethine protons (δH 3.195 (dd, (J =
1.5, 4.5 Hz, 1H), 3.202 (dd, J = 11.0, 5.0 Hz, 1H). The 13
C
NMR spectrum showed 30 carbon signals in which three of
them bear a hydroxyl group (δ 75.64, 78.76, 78.85). Two of
C
them were methines while the other was a quaternary carbon.
Among serratane-type triterpenes, 5 possesses a structure that
agrees with this data. The literature data for 5 were reported
With both onocerin and serratane synthases in hand, we can
now study the difference in the catalytic mechanism of the two
enzymes. As mentioned, the two enzymes catalyze the same
cyclization cascade up to the C8 cation intermediate, from
which an immediate deprotonation will give rise to onocerane
while an additional cyclization from the neighboring exo-
methylene will result in the formation of serratane skeleton.
What factor of these enzymes is responsible for the difference
between these reaction paths remains to be elucidated. In
addition, the LCD-catalyzed cyclization into 3 is terminated by
deprotonation, while the LCE-catalyzed reaction is terminated
by water addition. The difference of termination mechanism
between LCD and LCE is intriguing. We speculate that LCE
may have an additional space around C−D rings in its active
site cavity, which allows a solvent water molecule to enter.
Mutational studies would clarify such questions and elucidate
the molecular basis for the formation of two of the most
unusual triterpenes found in nature.
Our studies also pointed out that an evolutionary divergence
of OSC-like genes such as LCD and LCE was responsible for a
production of diverse onoceroid structures in ferns. While the
LCC gene would produce the bicyclic triterpene 4, a
combination of LCC and LCD would produce onoceranes,
while that of LCC and LCE would produce serratanes. Such a
switch in production of these onoceroids can be controlled by
an expression of either gene in the cells. Here we also
17
with pyridine-d as a solvent, while the data for diacetylated 5
5
8
were reported with CDCl . Therefore, our sample of the major
3
product was acetylated and analyzed by NMR, which showed
complete agreement with the literature values. We also
1
13
measured H and C NMR in pyridine-d as solvent, which
5
1
7
showed complete agreement with the literature data.
To further clarify the structure including the stereochemistry
of the tertiary hydroxyl group at C14, the diacetylated sample
was subjected to dehydration conditions using SOCl . Under
2
such conditions, 5 having a β-hydroxyl group at C14 would
result in production of serratenediol bearing an olefin between
C14−C15, while the α-epimer would give a mixture of
serratenediol and its olefinic isomer isoserratenediol having
8
an olefin between C13−C14. The configuration of the
hydroxyl group may result in such a difference since trans-
elimination toward C13 to produce isoserratenediol is only
possible with the α-configuration. Our result after GC−MS
analysis showed almost exclusive production of diacetylated
serratenediol, which agreed well with the results from the
literature and further confirmed that our major product was
8
indeed 5.
On the other hand, the minor product showed identical
retention time and a MS spectrum with an authentic sample of
C
Org. Lett. XXXX, XXX, XXX−XXX