A.E. Golliher, A.J. Tenorio, N.O. Dimauro et al.
Tetrahedron Letters 67 (2021) 152891
ural (À)-CBDV is in early clinical development for the treatment of
autism spectrum disorders [16] and recently, (À)-CBDP has
emerged as a more potent cannabinoid than (À)-CBD itself, making
it an alternative to THC therapy without the signature psychoactiv-
ity of the latter [17].
At the onset of our synthetic campaign, we evaluated the cur-
rently known syntheses of (À)- and (+)-CBD, many of which
involve the acid-catalyzed union of a terpene derivative with olive-
tol, several of which are noteworthy here. The report by Petrzilka
utilized limonene-derived 5 as one of the coupling partners
(Scheme 1), uniting this compound with olivetol (10) under mildly
acidic conditions [18]. While this processes does permit access to
(À)-CBD, its key step suffers from a long reaction time (days), mod-
est yield, and the overall number of steps in which 5 was derived
from (+)-4 [19]. A separate approach, first pioneered by Cardillo
[8g] and later employed by Mechoulam [8c], utilized
isopiperitenone [(À)-6] as a starting material. From this terpene,
(+)-CBD could be accessed in two steps involving (1) LiAlH4
reduction, and (2) treatment of the resultant alcohol mix-ture (8
and 9) with 10 in the presence of BF3ꢀOEt2. Unfortun-ately, the
relatively high cost of isopiperitenone (in either enantio form, ~
$1000/g) challenged us to think of potential ways to synthesize
enantiopure 8 from more readily available starting compounds
(Scheme 1) [20]. Recognizing the structural similarity between
the southern hemisphere of 8 and (+)-carvone, we began to
envision strategies to convert this inexpensive ($0.15/g),
caraway-derived terpene into the requisite chiral, non-racemic
isopiperitenol.Scheme 2
Fig. 1. Cannabidiol and related analogs.
[( )-Cesamet] and Dronabinol [8b]. Also of significance, the
cannabinoid drug Dexanabinol (HU-211, Fig. 1), based on the (+)-
ent-cannabinoid skeletal structure, surprisingly has no affinity for
CB1 or CB2 receptors, yet has significant non-competitive
antagonist effects on N-methyl-D-aspartic acid [11]. This is notable
since it is based on HU-219, which is a synthetic and more potent
derivative of (À)-CBD [11].
While data suggests that (À)-CBD exhibits a low affinity for CB1
(found mainly in the brain) and CB2 (in peripheral cells), its non-
natural synthetic enantiomer ent-CBD [(+)-2] and related deriva-
tives are known to have a higher affinity for these same membrane
receptors [12]. We believe ent-CBD derivatives will continue to
prove valuable as novel derivatives of (À)-CBD continue to be
explored as potential new therapeutics. To help support this
statement, Table 1 shows the nM binding affinities of select
cannabinoids towards the CB1 and CB2 receptors, demonstrating
that (+)-ent-2 has increased binding when compared to its natural
stereoisomer [8c]. Interestingly, another trend that warrants
Results and discussion
attention is the increased binding affinity of
D
9-(À)-THC
In terms of retrosynthesis, based on literature precedent, we
believed it would be possible to access 8 from tosylhydrazone 12
by exploiting the McIntosh reduction/rearrangement chemistry,
which would effectively transpose the alkene from the D8 to the
D9 location (note: cannabinoid notation) [21]. Hydrazone 12, in
turn, could be easily derived from hydroxycarvone 11, which is
derivatives as their alkyl tails increase in length [13]; (À)-THCP,
which has a seven carbon tail, binds an order of magnitude tighter
to CB1 and CB2 than
D
9-(À)-THC (Table 1).
In 2018, the Maio laboratory reported a new synthetic method
that allowed for the expedient construction of non-natural CBD
derivatives via the Lewis Acid mediated union of (À)-carvone, a
readily available and inexpensive starting material, with resorcinol
derivatives [14]. Importantly, by using (+)-carvone, this protocol
also allowed access to enantiomers of the CBD scaffold in only
three synthetic operations, two of which are general and can be
carried out on gram scale, yielding a relatively stable epoxy-car-
OH
10, H
4 steps
*
)-CBD
)-2]
Petrzilka
non-stereo-
[
selective
vone silyl ether. However, difficulty in D8 to
D
9–alkene transposi-
tion forced us to explore an alternative route for converting our
scaffold into (+)-ent-CBD itself, as well as its C-3 and C-7 alkyl
chain isomers, (+)-ent-cannabidivarin [(+)-1, ent-CBDV] and (+)-
ent-cannabidiphorol [(+)-3, ent-CBDP], respect-ively, neither of
which have been previously prepared in their non-natural, enan-
tiomeric form. Our interest in these latter two derivatives stems
from structure activity relationship data that demonstrate the
importance of the alkyl chain length and how these derivatives
may bind to CB1 and CB2 receptors (Table 1) [15]. Also of note, nat-
(+)-4
)-6
5
OH
non-stereoselective
Cardillo
O
O
HO
4
10
10,
stereo-
Table 1
BF3 OEt2
selective
Previously reported binding affinities of select cannabinoids [8c-e,12].
(+)-CBD
[(+)-2]
*
basic
this work
Cannabinoid
CB1 Ki (nM)
CB2 Ki (nM)
OH
alumina
(À)-CBDV
(À)-CBD
(+)-CBD
>10,000
>10,000
842
>10,000
>10,000
203
)-8, trans
)-9, cis
(+)-7
(À)-THCV
(À)-THC
(À)-THCP
22–75
18–40
1.2
62–105
36–42
6.2
Scheme 1. Previous syntheses in the context of this work.
2