Oxa-adamantyl cannabinoids

Article abstract of DOI:10.1016/j.bmcl.2021.127882

As a continuation of earlier work on classical cannabinoids bearing bulky side chains we report here the design, synthesis, and biological evaluation of 3′-functionalized oxa-adamantyl cannabinoids as a novel class of cannabinergic ligands. Key synthetic steps involve nucleophilic addition/transannular cyclization of aryllithium to epoxyketone in the presence of cerium chloride and stereoselective construction of the tricyclic cannabinoid nucleus. The synthesis of the oxa-adamantyl cannabinoids is convenient, and amenable to scale up allowing the preparation of these analogs in sufficient quantities for detailed in vitro evaluation. The novel oxa-adamantyl cannabinoids reported here were found to be high affinity ligands for the CB1 and CB2 cannabinoid receptors. In the cyclase assay these compounds were found to behave as potent and efficacious CB1 receptor agonists. Isothiocyanate analog AM10504 is capable of irreversibly labeling both the CB1 and CB2 receptors.

Full text of DOI:10.1016/j.bmcl.2021.127882

Bioorg. Med. Chem. Lett. 38 (2021) 127882
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Oxa-adamantyl cannabinoids
,
Thanh C. Hoa, Marcus A. Tiusa *, Spyros P. Nikasb, Ngan K. Tranb, Fei Tongb, Han Zhoub,
,
*
Nikolai Zvonokb, Alexandros Makriyannisb
a Department of Chemistry, University of Hawaii at Manoa, 2545 The Mall, Honolulu, HI 96822, United States
b Center for Drug Discovery, Department of Chemistry and Chemical Biology, and Department of Pharmaceutical Sciences, Northeastern University, Boston, MA 02115,
United States
A R T I C L E I N F O
A B S T R A C T
Keywords:
As a continuation of earlier work on classical cannabinoids bearing bulky side chains we report here the design,
synthesis, and biological evaluation of 3′-functionalized oxa-adamantyl cannabinoids as a novel class of can-
nabinergic ligands. Key synthetic steps involve nucleophilic addition/transannular cyclization of aryllithium to
epoxyketone in the presence of cerium chloride and stereoselective construction of the tricyclic cannabinoid
nucleus. The synthesis of the oxa-adamantyl cannabinoids is convenient, and amenable to scale up allowing the
preparation of these analogs in sufficient quantities for detailed in vitro evaluation. The novel oxa-adamantyl
cannabinoids reported here were found to be high affinity ligands for the CB1 and CB2 cannabinoid re-
ceptors. In the cyclase assay these compounds were found to behave as potent and efficacious CB1 receptor
agonists. Isothiocyanate analog AM10504 is capable of irreversibly labeling both the CB1 and CB2 receptors.
Oxa-adamantyl cannabinoid
Classical cannabinoid
Hexahydrocannabinol
Synthesis
Design
Novel class
CB1 cannabinoid receptor
CB2 cannabinoid receptor
Binding affinity
The adamantyl group has attracted the interest of medicinal chemists
in both academia and industry because of its unique structural proper-
ties, namely, steric bulk, and conformational rigidity. Incorporation of
the adamantyl group into the structure of biologically active molecules
has the risk of destroying the biological activity of the molecule. How-
ever, addition of an adamantyl group to the chemical structure of an
active bioactive molecule may result in analogs with novel chemical,
biological, and pharmaceutical properties. Over the years, our studies on
the incorporation of the adamantyl group in the structure of cannabi-
noid ligands have led to interesting observations that can be summarized
as follows: a) Replacement of the n-pentyl chain with the adamantyl
group in the prototypic cannabinoid, tetrahydrocannabinol (THC), the
psychoactive constituent in cannabis, led to a crystalline analog
(AM411; Fig. 1) that enabled for first time the study of the tricyclic THC
structure by X-ray crystallography.1 AM411, shows preference in bind-
ing to the CB1 receptor over the CB2 receptor.1 CB1 and CB2 are
currently the two well-characterized cannabinoid receptors belonging to
the superfamily of G-protein-coupled receptors (GPCRs).2–8 The CB1
receptor retains a high (≥97%) degree of amino acid sequence identity
across mouse, rat, and human.9 In contrast, the human CB2 receptor
displays only 82% and 81% amino acid identity with mouse10 and rat11
respectively. Another intriguing pharmacological property of AM411 is
its ability to function as a CB1 partial agonist.12 This CB1 partial agonist
profile is also observed in the 9-hydroxymethyl cannabinol analog
AM408913,14 but not in the hexahydrocannabinol (HHC) analog
AM4054 that has been found to behave as a potent and relatively long-
lasting CB1 full agonist in monkeys (Fig. 1).13,15 b) More recently we
reported functionalized adamantyl cannabinoid receptor probes car-
rying substituents such as the electrophilic isothiocyanate and photo-
activatable azido groups, both of which are capable of covalent
attachment with the target protein.16 Worthy of note is the in vitro
functional profile of key adamantyl substituted cannabinoid probes
within this class of compounds. Thus, compounds AM993 and AM994,
carrying the adamantyl group were found to behave as potent, high
efficacy, agonists at CB1 and low efficacy, inverse agonists/antagonists
at CB2 in the cyclase assay. This is a functional profile that is not
observed in classical cannabinoid compounds carrying alkyl side chain
and therefore seems to be unique for HHC analogs carrying the
Abbreviations: CB1, cannabinoid 1 receptor; CB2, cannabinoid 2 receptor; DMAP, 4-dimethylaminopyridine; dppp, 1,3-bis(diphenylphosphino)propane; EDCI, 1-
ethyl-3-(3-dimethylaminopropyl)carbodiimide; EI, electron impact; ESIMS, electrospray ionization mass spectrometry; HRMS, high resolution mass spectrometry; m-
CPBA, meta-chloroperoxybenzoic acid; NMR, nuclear magnetic resonance; MOMCl, methyl chloromethyl ether; MS, molecular sieves; PMHS, poly-
methylhydrosiloxane; Tf, trifluoromethanesulfonyl (triflyl); TFAA, trifluoroacetic anhydride; HOBT, hydroxybenzotriazole.
* Corresponding authors.
E-mail addresses: tius@hawaii.edu (M.A. Tius), a.makriyannis@neu.edu (A. Makriyannis).
https://doi.org/10.1016/j.bmcl.2021.127882
Received 11 November 2020; Received in revised form 1 January 2021; Accepted 12 February 2021
Available online 24 February 2021
0960-894X/Published by Elsevier Ltd.

T.C. Ho et al.
Bioorganic & Medicinal Chemistry Letters 38 (2021) 127882
OH
O
OH
11
9
OH
OH
OH
OH
O
O
O
1d
1a
9-THC
1b
AM411
1c
AM4054
AM4089
OH
R1
9
O
OH
OH
N
H
R2
3'
R
N
N
O
O
O
2
1e
R1 = OH, CH2OH
1f
AM993
AM994
: R = CH2N3
: R = CH2NCS
tPSA: 62.05
CLogP: 7.78
R2 = CH2NCS, CH2N3, CN ...
AM10257
OH
: R1 = OH, R2 = CH2NCS
AM10504
tPSA = 71.28, CLogP = 4.71
Oxa-adamantyl cannabinoids
Figure 1. Structures of adamantyl and oxa-adamantyl cannabinoids.
adamantyl side chain. c) The purposeful design of the adamantyl
substituted CB2 antagonist AM10257 led to the appropriate fit of the
ligand within the CB2 receptor binding domain, and thus facilitated the
formation of crystals of the AM10257-human CB2 complex for X-ray
analysis.2
The oxa-adamantyl HHCs reported here are also endowed with
increased polarity when compared to the more lipophilic adamantyl
substituted analogs (the adamantyl group adds 10 lipophilic carbon
atoms to the parent molecule, see for example the cLogP and the tPSA
values25 of the adamantyl and oxa-adamantyl analogs AM994 and
AM10504 in Fig. 1). Enhancement of the polar characteristics of these
cannabinoids is expected to make them more suitable not only for in
vitro experiments, but also for in vivo studies because higher polarity of
a molecule generally results in lower plasma protein binding and better
absorption.
In an effort to explore in more detail the stereoelectronic re-
quirements for activity of these novel, bulky, and conformationally fixed
side chains in classical cannabinoids, we have extended our earlier work
to include the oxa-adamantyl substituted hexahydrocannabinols (HHCs)
depicted in Fig. 1. In terms of stereochemical features the adamantyl and
oxa-adamantyl groups are similar, however, the presence of the elec-
tronegative oxygen atom in the oxa-adamantyl group changes the
characteristics of the electronic density of this group and makes it
electronically dissimilar to the adamantyl moiety.
Moreover, in the earlier work with the adamantyl HHCs we have
chosen the β-C9-hydroxy-HHC template.16 In the current work, we
maintain the successful β-C9-hydroxy-HHC template, and we also
extended our SAR to include the β-C9-hydroxymethyl (C11-hydroxy)
HHC template. Inclusion of a second tricyclic template in our study was
based on: a) recent data from the crystal structure of the hCB1-AM841
complex, where it appears that the C11-hydroxy group in the HHC
template forms a hydrogen bond with Ile267 at the extracellular loop 2.5
and b) earlier in vitro and in vivo data showing very high potency and
efficacy of C11-hydroxy-HHC analogs.13 Importantly, such observations
have been seen in studies with both rodents and non-human primates.26
An additional goal during the development of the synthesis of 3′-
functionalized oxa-adamantyl cannabinoids (2; Fig. 1) was to circum-
vent issues arising from the synthesis of 3′-functionalized adamantyl
cannabinoids 1e. Scheme 1 summarizes the early steps in the synthesis
of compounds 1e. Commercially available 2,6-dimethoxyphenol 3 and
We recently reported our studies on the synthesis of oxaza adamantyl
HHCs17 as well as our preliminary work on the synthesis of oxa ada-
mantane18 and model 1-aryl-2-oxa adamantanes.19 In this work we have
designed, synthesized, and in preliminary work we have assessed a new
class of HHC analogs in vitro, namely 3′-functionalized oxa-adamantyl
cannabinoids (2; Fig. 1).
Following design approaches we described earlier for the adamantyl
substituted HHCs we have incorporated in the oxa-adamantyl HHC
scaffold electrophilic (NCS and CN) and photoactivatable (N3) groups to
generate probe molecules that are anticipated to be useful in exploring
the structural and functional properties of the ligand-receptor complex
using our LAPS approach as well as crystallography methods.2,4,5,20–24
CO2H
OCH3
OCH3
OCH3
CO2R2
CO2CH3
HO
R1O
H
a
d
+
HO
H3CO
H3CO
H3CO
3
4
7
1 = R2 = H
R1 = Tf, R2 = CH3
5
6
R
b,c
Scheme 1. The Friedel-Crafts synthesis of carboxyadamantanyl resorcinol ether 7a Note: aReagents and conditions: (a) MeSO3H, 50 ◦C; 100%; (b) TMSCHN2, PhH,
MeOH, RT; (c) Tf2NPh, Et3N, DMAP (cat), CH2Cl2, reflux; 91% (two steps); (d) PdCl2(Ph3P)2, dppp, n-Bu3N, HCOOH, PMHS (cat), DMF, reflux; 89%.
2

T.C. Ho et al.
Bioorganic & Medicinal Chemistry Letters 38 (2021) 127882
O
OCH3
OH
CH2
a
b
O
O
+
OH
Li
OCH3
11
10
8
9
OH
O
CO2H
O
c
O
OCH3
d
OR
HO
12
15
OCH3
OR
13
14
R = CH3
R = H
e
Scheme 2. Synthesis of resorcinol 14a, Note: aReagents and conditions: (a) p-TsCl, benzene/Pyr, 75 ◦C, 40 h, 81%; (b) m-CPBA, NaHCO3, CH2Cl2/H2O, rt, 3 h, 85%;
(c) CeCl3, ꢀ 50 ◦C, 1 h, 0 ◦C, 30 min, 83%; (d) Jones reagent, (CH3)2CO, 0 ◦C, 2 h, 87%; (e) 47% HI(aq), 120 ◦C, 21 h, 92%.
hydroxyadamantanecarboxylic acid 4 were combined in a Friedel-Crafts
alkylation reaction that led to 5. The phenolic hydroxyl group in 5 was
converted to the triflate and the carboxylic acid was exposed to trime-
thylsilyl diazomethane to give methyl ester 6. Palladium-catalyzed
reductive cleavage of the triflate led to resorcinol dimethyl ether 7.
Although the chemistry in Scheme 1 worked well, the Friedel-Crafts
reaction in the first step had to be performed in neat methanesulfonic
acid and was therefore inconvenient and impractical to scale up. The
high cost of 4 was also an issue. We hypothesized that the oxa-
adamantyl series of compounds typified by structure 2 could be
accessed through a more practical synthesis that would be easier to scale
up, while retaining the potent activity seen in compounds 1e. This
would facilitate the in vivo studies of these compounds. We had reported
the synthesis of 2 (R1 = OH, R2 = H) and had determined its high affinity
for CB1 and for CB2,18 thereby providing some validation of our
hypothesis.
Scheme 2 summarizes the preparation of the functionalized resor-
cinol fragment 14. Although 1,3-adamantanediol is commercially
available, it is quite expensive. Fortunately, it is easily prepared in high
yield from cheap 1-adamantanol by oxidation with catalytic ruthenium
trichloride using a small excess of sodium periodate as the stoichio-
metric oxidant.27 Grob fragmentation of the monotosylate derived from
828 led to 7-methylenebicyclo[3.3.1] nonan-3-one 9 in 81% yield.
Epoxidation of 9 with m-CPBA in a dichloromethane-aqueous sodium
O
9
O
CO2H
OAc
OH
OH
AcO OAc
CO2H
10a
CO2H
O
6a
OH
b
a
+
+
O
O
OAc
O
HO
20
19
14
22
21
OH
OH
OH
OH
OH
OH
OH
CO2H
CONH2
CN
c
d
e
O
O
O
O
O
O
25
24
23
f
OH
OH
OH
OH
OH
OH
N3
NCS
NH2
h
g
O
O
O
O
O
O
27
28
26
Scheme 3. Synthesis of functionalized oxa-adamantyl cannabinoidsa, Note: aReagents and conditions: (a) p-TsOH⋅H2O, (CH3)2CO/CHCl3, 50 ◦C, 22 h, 52%; (b)
TMSOTf, CH3NO2, 0 ◦C to rt, 3 h, 87%; (c) NaBH4, CH3OH, ꢀ 5 ◦C, 3 h, 94%; (d) i. HOBT, EDCI, THF, rt, 16 h, ii. NH4OH (aq), rt, 1 h, 81%; (e) i. TFAA, Pyr, 1,4-
dioxane, rt, 16 h, ii. K2CO3, CH3OH, rt, 16 h, 84%; (f) BH3⋅Me2S, THF, 0 ◦C, 30 min, rt, 3 h, 81%; (g) TfN3, Et3N, CuSO4, CH2Cl2/CH3OH/H2O, rt, 14 h, 74%. (h) i.
CS2, Et3N, THF, 0 ◦C, 2 h, ii. p-TsCl, 0 ◦C to rt, 1.5 h, 82%.
3

T.C. Ho et al.
Bioorganic & Medicinal Chemistry Letters 38 (2021) 127882
H3CO
O
CHO
OH
OH
OH
CO2H
CO2H
CO2H
a
b
c
O
O
O
O
O
O
22
29
30
OH
OH
O
11
CHO
OH
OH
OH
CO2H
CO2H
33
34
35
36
X = CONH2
X = CH2NH2
X = CH2N3
X
e
d
O
O
O
O
O
X = CH2NCS
32
31
Scheme 4. Synthesis of oxa-adamantyl 9β-hydroxymethyl hexahydrocannabinolsa, Note: aReagents and conditions: (a) Ph3PCH2OCH3Cl, t-BuOK, THF, rt, 1.5 h; (b)
CCl3COOH, H2O, CH2Cl2, rt, 45 min; (c) K2CO3, CH3OH, rt, 4 h; (d) NaBH4, CH3OH, rt, 30 min, 77% from 22; (e) details for the synthesis are reported in Supporting
Information section.
bicarbonate reaction medium led to a single isomer of epoxy ketone 10
in 85% yield. It is known that exposure of 9 to peracids or to hydrogen
peroxide in benzonitrile leads to exo-epoxide 10 as the sole product29
and that no lactone from a competing Baeyer-Villiger process was
observed.30 The preferred twin chair conformer of 9 effectively prevents
endo approach by the peracid, and also suppresses the formation of the
Criegee intermediate of the Baeyer-Villiger reaction. Epoxidation of 9 in
neat dichloromethane led to a lower yield of 10 (70–75%), presumably
due to competing acid-catalyzed epoxide cleavage.
47% aqueous HI at reflux for 21 h led to the desired resorcinol 14 in 92%
yield. The reason why HI, a strong Bronsted acid with a nucleophilic
counterion, succeeded in cleaving the methyl ether groups in 13
whereas BBr3 failed is not obvious. Having developed an effective and
scale-able route to 14 we proceeded to the condensation-cyclization
reaction sequence to form the tricyclic cannabinoid nucleus.
We utilized our modification of the strategy developed by Archer and
co-workers for their pioneering synthesis of Nabilone.35 Resorcinol 14
was condensed with the mixture of diacetates 19 and 20 that are derived
in three steps from (ꢀ )-β-pinene (Scheme 3). This led to the desired
(6aR,10aR) absolute configuration of the cannabinoid nucleus. As we
have noted in earlier work18 Archer and co-workers’ procedure works
beautifully for Nabilone, but any deviation from their conditions for the
condensation of diacetates 19 and 20 with the resorcinol results in a
poor outcome for the reaction. The reaction is especially sensitive to any
changes to the preferred solvent, chloroform. Resorcinol 14 is poorly
soluble in chloroform, leading to very long reaction times, incomplete
conversion to product and the appearance of numerous byproducts. This
made it necessary to add the minimum amount of acetone as a co-solvent
to ensure that 14 was completely soluble in the reaction medium. When
the reaction was performed under these conditions at 50 ◦C 21 was
formed in 52% yield. Higher reaction temperatures led to the formation
of numerous byproducts. The relative stereochemistry of asymmetric
carbon atoms in 21 was determined through NOE experiments on 23 as
described in our earlier work.36,37 Rearrangement-cyclization of 21
promoted by TMSOTf in dry nitromethane led to tricyclic cannabinoid
22 in 87% yield. Treatment of 21 with less than 3 equiv TMSOTf led only
to traces of product, whereas ca. 5 equiv led to clean conversion to 22.
The use of TMSOTf to promote the rearrangement-cyclization is pref-
3,5-Dimethoxyphenyllithium 11 was generated in situ from lithium-
bromine exchange of commercially available 1-bromo-3,5-dimethoxy-
benzene and n-butyllithium.31 Nucleophilic addition of 11 to 10 in the
32
presence of anhydrous CeCl3 led to the formation of 3-(3,5-dime-
thoxyphenyl)-2-oxaadamantan-1-yl methanol 12 in 83% yield through
transannular epoxide cleavage and cyclization. In the absence of CeCl3,
an insignificant amount of 12 was formed. Instead ca. 30% of epoxy
ketone 10 was recovered along with ca. 50% of keto alcohol 15 (Scheme
2). The formation of 15 is consistent with the earlier observation by
Mlinaric-Majerski and co-workers33 who found that in the absence of
CeCl3 the addition of methyllithium to 10 led to diastereomers 16 and
17 in 69% yield while producing only 10% of 18 (eq 1). It is therefore
important to attenuate the basicity of 11 while retaining its nucleophi-
licity by transmetallation to cerium. Jones oxidation of the hydrox-
ymethyl group in 12 led to carboxylic acid 13 in 87% yield.34
Equation 1
17
erable to SnCl4 that was used by Archer and co-workers and which
leads to the formation of emulsions during work up that erode the iso-
lated yield of product.18
The C9 ketone in 22 was reduced at ꢀ 5 ◦C with excess NaBH4 in 94%
yield to give a ca. 95:5 mixture of C9-β equatorial alcohol 23 and its C9-
α
axial diastereomer. Alcohol 23 served as the common intermediate for
the entire series of functional oxa-adamantyl cannabinoids. Amidation
of 23 with ammonium hydroxide in the presence of EDCI and HOBT took
place in 81% yield to give 24. Dehydration of amide 24 by exposure to
trifluoroacetic anhydride and pyridine in dioxane38 led to nitrile 25 in
84% yield following methanolysis of the two trifluoroacetate esters that
had formed at the C9 alcohol and the phenol. Reduction of the amide
with BH3⋅Me2S in THF led to primary amine 26 in 81% yield. Azide 27
was prepared by diazo transfer to amine 26 following methodology that
was originally developed by Cavender and Shiner39 and later by Wong
Cleavage of the methyl ether groups in 13 to reveal resorcinol 14 was
unexpectedly challenging. Whereas treatment of 7 with BBr3 in
dichloromethane at 0 ◦C followed by warming to room temperature led
to cleavage of both methoxy groups and of the methyl ester in 89%
yield,16 exposure of 13 to the same conditions led to a multitude of
products. We eventually determined that exposure of 13 to an excess of
4

T.C. Ho et al.
Bioorganic & Medicinal Chemistry Letters 38 (2021) 127882
Table 1
Table 2
Affinities of oxa-adamantyl cannabinoids for rCB1, mCB2, and hCB2 cannabi-
CB1/CB2 Functional potencies (EC50) of selected oxa-adamantyl cannabinoids.
noid receptors (±95% confidence limits).
Compd
Compd
EC50 (nM),a E(max) (%),b classification
R1
R2
Ki (nM)a
rCB1
hCB2
rCB1
mCB2
hCB2
CP,55940
7.3 nM, 73%, agonist
2.4 nM, 45%, agonistc
0.8 nM, 94%, agonistc
3.1 nM, 87%, agonist
0.8 nM, 88%, agonist
0.9 nM, 93%, agonist
0.9 nM, 94%, agonist
1.4 nM, 76%, agonist
AM993
NRc
(¡)Δ9-THC
AM993
39.5b
4.4b
3.0b
28.7
26.6
2.3
40b
36.4b
9.6b
10.3b
369
31.7
3.1
AM994
>400 nM, ꢀ 91%, inverse agonistc
26.4b
34.6b
865
1.6
27 AM10509
28 AM10504
35 AM10512
36 AM10511
NR
NR
NR
NR
AM994
24 AM10507
25 AM10508
27 AM10509
28 AM10504
33 AM10510
35 AM10512
36 AM10511
OH
CONH2
CN
OH
OH
CH2N3
CH2NCS
CONH2
CH2N3
CH2NCS
2.7
OH
4.8
9.5
3.6
a
Functional potencies at rCB1 and hCB2 receptors were determined by
measuring the decrease in forskolin-stimulated cAMP levels.16,48,50 EC50 values
were calculated using nonlinear regression analysis. Data are the average of two
independent experiments run in triplicate.
CH2OH
CH2OH
CH2OH
10.8
2.0
34.9
0.7
91.8
2.2
8.4
6.6
4.7
a
b
Affinities were determined using rat brain (CB1) preparations or membranes
from HEK293 cells expressing mouse or human CB2 receptors and [3H]CP-
55,940 as the radioligand following previously described procedures.16,47,48
Data were analyzed using nonlinear regression analysis. Ki values were obtained
from three independent experiments performed in triplicate and are expressed as
the mean of the three values.
Forskolin stimulated cAMP levels were normalized to 100%. E(max) is the
maximum inhibition of forskolin stimulated cAMP levels and is presented as the
percentage of CP-55,940 response at 500 nM. NR: no response up to 2.3
μ
M.
c
Reported previously.16
b
28 (Scheme 3).
Reported previously.16,47
The abilities of compounds 24, 25, 27, 28, 33, 35, and 36 to displace
the radiolabeled CB1/CB2 agonist CP-55,940 from membranes prepared
from rat brain (source of CB1) and HEK 293 cells expressing either
mouse CB2 or human CB2 were determined as described earlier,16,47,48
and inhibition constant values (Ki) from the respective competition
binding curves are listed in Table 1. Two CB2 receptor preparations were
used in order to address species differences that we observed earlier.49 It
should also be noted that the rat, mouse, and human CB1 receptors have
97–99% sequence identity across species and, as shown earlier,4,26,50,51
are not expected to exhibit variations in their Ki values. However, the
CB2 receptor shows less homology (~82%) between species than does
CB1 (97–99%), and that variability could cause species-related differ-
ences in affinity.48
and co-workers.40,41 Cupric sulfate-catalyzed diazo transfer from freshly
prepared triflyl azide at 0 ◦C led to 27 in 74% yield. The presence of
triflic acid in the reaction medium is likely deleterious to both starting
material and to product, therefore sodium azide was used in excess in
order to consume all the triflic anhydride, and the dichloromethane
solution of the crude triflyl azide was washed with cold saturated
aqueous bicarbonate prior to use. Following the protocol of Wong and
co-workers, the reaction was performed in a dichloromethane/meth-
anol/water mixed solvent. The ratio was carefully optimized to 6/6/1 so
as to maintain a homogeneous reaction medium, since this significantly
affected the yield of the diazo transfer reaction.40 It is also noteworthy
that stoichiometric cupric sulfate contributed to the decomposition of
the product, whereas the reaction was very clean with 4 mol % Cu(II).
Treatment of primary amine 26 with CS2 in the presence of trie-
thylamine in THF, followed by tosyl chloride mediated decomposition of
the derived dithiocarbamate gave isothiocyanate 28 in 82% yield.42 The
use of excess CS2 to improve the yield of isothiocyanate has been re-
ported.43 More than 3.0 equiv of triethylamine was required to consume
amine 26 completely, however, use of a large excess (>10 equiv) of
triethylamine with a similarly large excess of tosyl chloride led to un-
desired products, including the tosylate of the phenolic hydroxyl group.
The use of no more than 2.0 equiv of tosyl chloride led to clean con-
version of 26 to 28.
Our binding affinity data show that oxa-adamantyl cannabinoids
have very high binding affinities for both the CB1 and CB2 cannabinoid
receptors. Optimal affinities are observed in compounds with –CH2N3,
and –CH2NCS groups at the C3′-position (27, 28, 35, and 36) where the
respective Ki values are at the nanomolar or sub-nanomolar range.
Worthy of note, the binding affinities of the oxa-adamantyl cannabi-
noids are much higher than those of the major psychoactive constituent
of Cannabis sativa, (ꢀ )-Δ9-THC, and are comparable to those of the
corresponding adamantyl analogs (e.g., AM993, AM994),16 an obser-
vation that is in agreement with our design hypothesis.
A cursory examination of the binding data in Table 1 indicates that
analogs carrying the carboxamide and the cyano groups (24, 25, and 33)
exhibit slightly lower binding affinities when compared to their azido
and isothiocyanato counterparts (27, 28, 35, and 36). Notably, the
cyano analog (25) shows species differences between mouse CB2 and
human CB2 in binding affinity (20-fold difference) when compared with
Scheme 4 summarizes the preparation of C11-hydroxy cannabinoids.
The first task in this synthesis was the stereospecific introduction of the
9β-hydroxymethyl group into 22. This conversion was reported by our
group several years ago.36,37,44 The Wittig reagent that was generated in
situ from excess (methoxymethyl)triphenylphosphonium chloride and
potassium tert-butoxide in THF at room temperature led to a very clean
mixture of methyl enol ether geometric isomers 29. Commercially
available potassium tert-butoxide45 could be used effectively in the place
of n-BuLi,13 or sodium tert-amylate, which was prepared from sodium
hydride and tert-amyl alcohol in benzene.46 Hydrolysis of crude 29 with
wet trichloroacetic acid in dichloromethane afforded aldehyde di-
astereomers 30. Epimerization of the diasteromeric mixture with po-
tassium carbonate in methanol led to the more stable β-equatorial
aldehyde 31 as the major diastereomer. Reduction in situ with sodium
borohydride provided 9β-hydroxymethyl intermediate 32 in 77% yield
120
CP55940
AM10504
AM10509
AM10511
AM10512
110
100
90
80
70
60
50
40
30
20
10
0
110
from ketone 22, accompanied by ca. 3% of the 9α-hydroxymethyl dia-
-3
-2
-1
0
1
2
3
4
stereomer. It is noteworthy that 32 could be obtained from 22 in high
yield in only 4 steps without purification of any intermediates and
without the use of protecting groups. Intermediate 32 was converted to
amide 33, amine 34, azido 35, and isothiocyanate 36 through functional
group transformations as described for the synthesis of 24, 26, 27, and
Log [Ligand] (nM)
Figure 2. Concentration-dependent inhibition of forskolin-stimulated cAMP
accumulation in HEK293 cells expressing rCB1 receptors by representative li-
gands. All compounds behave as potent agonists.
5

T.C. Ho et al.
Bioorganic & Medicinal Chemistry Letters 38 (2021) 127882
Figure 3. Saturation binding curves using [3H]CP-55,940 for CB1 (left panel) and CB2 (right panel) receptors preincubated with isothiocyanate analog 28
(AM10504). Control membranes processed in parallel but without prior exposure to irreversible ligand are shown in blue. Data represent the mean values ± SEM of
at least two independent experiments performed in triplicate. Compound 28 inhibits the specific binding of the radioligand to CB1 and CB2 receptors.
other analogs (ca. 3-fold difference). Currently, we do not have an im-
mediate explanation for this observation, and future investigation is
required.
and practical to scale up compared to that of the adamantyl analogs. Our
biological evaluation shows that oxa-adamantyl cannabinoids are high
affinity ligands to both CB1 and CB2 cannabinoid receptors. In vitro
functional assays, key compounds behave as highly potent and effica-
cious CB1 receptor agonists similar to the adamantyl analogs we re-
ported earlier. In addition, the electrophilic isothiocyanate analog 28
can interact irreversibly with both the CB1 and CB2 receptors. Within
the class of tricyclic cannabinoids, these indicate that the more polar
oxa-adamantyl group behaves in vitro as a bioisostere to the more
lipophilic adamantyl moiety.
Additionally, a comparison of the binding affinity data of the C9-
hydroxy-HHC analogs 24, 27, and 28 with those of the C11-hydroxy-
HHC analogs 33, 35, and 36 indicate that both C9 and C11-hydroxy
HHC templates are well recognized by the CB receptors and thus, they
lead to oxa-adamantyl analogs with very high binding affinities for CB1
and CB2.
Focusing on those analogs possessing the highest CB1/CB2 affinities,
experiments were carried out by measuring changes in forskolin-
stimulated cAMP, as detailed earlier.16,48,50 Data are listed in Table 2
in which the standard cannabinoid agonist CP-55,940 is included for
comparison. The dose–response curves of representative analogs for the
CB1 receptors are depicted in Figure 2.
Declaration of Competing Interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Our testing results for the CB1 receptor showed that oxa-adamantyl
cannabinoids with the azido and isothiocyanate group (compounds 27,
28, 35, and 36) potently decreased the levels of cAMP, indicating that
within this signaling mechanism these compounds behaved as potent
agonists at the CB1 receptor (Table 2). When tested with hCB2 prepa-
rations, these oxa-adamantyl cannabinoids showed no responses in
Acknowledgements
This work was supported by grants from the National Institutes of
Health to A.M., DA009158, DA041435, DA041307, DA045020,
DA03801, DA023142, DA007215, and DA026795.
concentrations up to 2.3 μM in the cAMP functional test (dose–response
curves are not shown). These data parallel those reported earlier for the
3′-functionalized adamantyl courterparts.16
The ability of a representative compound carrying the covalently
reacting isothiocyanate group (28, AM10504) to label each of the two
receptors was obtained by measuring reductions in the binding for the
radioligand [3H]CP-55,940 when membrane preparation was pretreated
with the irreversibly bound ligand, compared to the untreated sample
(details in Experimental Section). Following earlier work from our lab-
oratory,16,20,52 the experiment was conducted by pretreating the mem-
brane with a concentration equal to 10-fold the Ki value of compound for
the receptor in question and subsequently measuring the decrease of
Bmax obtained from a saturation curve using [3H]CP-55,940. Percent
covalent labeling was calculated as {[Bmax(control) ꢀ Bmax(ligand)]/
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.bmcl.2021.127882.
References
1 Lu D, Meng Z, Thakur GA, et al. Adamantyl cannabinoids: a novel class of
cannabinergic ligands. J Med Chem. 2005;48:4576.
2 Li XT, Hua T, Vemuri K, et al. Crystal structure of the human cannabinoid receptor
CB2. Cell. 2019;176:459.
3 Shao ZH, Yin J, Chapman K, et al. High-resolution crystal structure of the human CB1
cannabinoid receptor. Nature. 2016;540:602.
B
max(control)} × 100. Bmax(ligand) as described in Experimental Sec-
4 Hua T, Vemuri K, Pu MC, et al. Crystal structure of the human cannabinoid receptor
CB1. Cell. 2016;167:750.
tion. The Bmax(control) for [3H]CP-55,940 was determined by con-
ducting the assay as described but in the absence of the test ligand. Our
results show that isothiocyanate 28 (AM10504) binds tightly to both the
rCB1 and hCB2 cannabinoid receptors and decreases Bmax by 65% and
32% respectively (Fig. 3).
5 Hua T, Vemuri K, Nikas SP, et al. Crystal structures of agonist-bound human
cannabinoid receptor CB1. Nature. 2017;547:468.
6 Devan WA, Dysarz III FA, Johnson MR, Melvin LS, Howlett AC. Determination and
characterization of a cannabinoid receptor in rat brain. Mol Pharmacol. 1988;34:605.
7 Munro S, Thomas KL, Abu-Shaar M. Molecular characterization of a peripheral
receptor for cannabinoids. Nature. 1993;365:61.
In summary, as an extension of our earlier work on cannabinoid li-
gands bearing bulky side chain pharmacophores we report here the
design, synthesis, and biological evaluation of a novel class of canna-
binergic ligands namely 3′-functionalized oxa-adamantyl cannabinoids.
These compounds were designed as alternative polar surrogates of 3′-
functionalized adamantyl cannabinoids. Oxa-adamantyl cannabinoids
have been prepared via nucleophilic addition/transannular cyclization
of aryllithium to an epoxyketone in the presence of cerium chloride
followed by the stereoselective construction of the tricyclic cannabinoid
nucleus. The synthesis of oxa-adamantyl cannabinoids was convenient
8 Iliopoulos-Tsoutsouvas C, Georgiadis MO, Ji P, Nikas SP, Makriyannis A. Natural
compounds and synthetic drugs to target type-1 cannabinoid (CB1) receptor. Royal
Society of Chemistry; 2020:48–88.
9 Console-Bram L, Marcu J, Abood ME. Cannabinoid receptors: Nomenclature and
pharmacological principles. Prog Neuro-Psychopharmacol Biol Psychiatry. 2012;38:4.
10 Shire D, Calandra B, Rinaldi-Carmona M, et al. Molecular cloning, expression and
function of the murine CB2 peripheral cannabinoid receptor. Biochim Biophys Acta,
Gen Subj. 1996;1307:132.
11 Griffin G, Tao Q, Abood ME. Cloning and pharmacological characterization of the rat
CB2 cannabinoid receptor. J Pharm Exp. 2000;292:886.
12 Desai RI, Thakur GA, Vemuri VK, Bajaj S, Makriyannis A, Bergman J. Analysis of
tolerance and behavioral/physical dependence during chronic CB1 agonist
6

T.C. Ho et al.
Bioorganic & Medicinal Chemistry Letters 38 (2021) 127882
treatment: effects of CB1 agonists, antagonists, and noncannabinoid drugs.
34 (a) Jiang Z-X, Yu YB. The synthesis of a geminally perfluoro-tert-butylated β-amino
acid and its protected forms as a potential pharmacokinetic modulator and reporter
for peptide-based pharmaceuticals. J Org Chem. 2007;72:1464.(b) For the
preparation of Jones reagent see: Eisenbraun EJ. Cyclooctanone Org Synth. 1965;45:
28.(c) For the original research article see: Bowden K, Heilbron IM, Jones ERH,
Weedon BCL. Researches on acetylenic compounds. Part I. The preparation of
acetylenic ketones by oxidation of acetylenic carbinols and glycols J Chem Soc. 1946;
39.
J Pharmacol Exp Ther. 2013;344:319.
13 Thakur GA, Bajaj S, Paronis C, et al. Novel adamantyl cannabinoids as CB1 receptor
probes. J Med Chem. 2013;56:3904.
14 Liu Y, Ho TC, Baradwan M, et al. Synthesis of functionalized cannabilactones.
Molecules. 2020;25:684.
15 Paronis CA, Chopda GR, Vemuri K, Zakarian AS, Makriyannis A, Bergman J. Long-
lasting in vivo effects of the cannabinoid CB1 antagonist AM6538. J Pharmacol Exp
Ther. 2018;364:485.
35 Archer RA, Blanchard WB, Day WA, et al. 3. Synthesis approaches to 9-ketocanna-
binoids. Total synthesis of nabilone. J Org Chem. 1977;42:2277.
36 Harrington PE, Stergiades IA, Erickson J, Makriyannis A, Tius MA. Synthesis of
functionalized cannabinoids. J Org Chem. 2000;65:6576.
16 Ogawa O, Tius MA, Zhou H, et al. 3’-functionalized adamantyl cannabinoid receptor
probes. J Med Chem. 2015;58:3104.
17 Goanvic DL, Tius MA. Oxaza adamantyl cannabinoids. A new class of cannabinoid
receptor probes. J Org Chem. 2006;71:7800.
37 Busch-Petersen J, Hill WA, Fan P, et al. Unsaturated side chain β -11-
hydroxyhexahydrocannabinol analogs. J Med Chem. 1996;39:3790.
38 Campagna F, Carotti A, Casini G. A convenient synthesis of nitriles from primary
amides under mild conditions. Tetrahedron Lett. 1813;1977:18.
18 Dixon DD, Sethumadhavan D, Benneche T, et al. J Med Chem. 2010;53:5656.
19 Benneche T, Tius MA. Synthesis of 2-oxaadamantane and 1-aryl-2-oxaadamantanes.
Tetrahedron Lett. 2016;57:3150.
20 Makriyannis A. Division of Medicinal Chemistry Award Address. 2012 Trekking the
cannabinoid road: a personal perspective. J. Med. Chem. 2014;57:3891.
21 Janero DR, Yaddanapudi S, Zvonok N, et al. Molecular-interaction and signaling
profiles of AM3677, a novel covalent agonist selective for the cannabinoid 1
receptor. ACS Chem Neurosci. 2015;6:1400.
39 Cavender CJ, Shiner Jr VJ. Trifluoromethanesulfonyl azide. Its reaction with alkyl
amines to form alkyl azides. J Org Chem. 1972;37:3567.
40 Alper PB, Hung S-C, Wong C-H. Metal catalyzed diazo transfer for the synthesis of
azides from amines. Tetrahedron Lett. 1996;37:6029.
41 Nyffeler PT, Liang C-H, Koeller KM, Wong C-H. The chemistry of amine-azide
interconversion: catalytic diazotransfer and regioselective azide reduction. J Am
Chem Soc. 2002;124:10773.
22 Mallipeddi S, Kreimer S, Zvonok N, et al. Binding site characterization of AM1336, a
novel covalent inverse agonist at human cannabinoid 2 receptor, using mass
spectrometric analysis. J Proteome Res. 2017;16:2419.
42 Wong R, Dolmabn SJ. Isothiocyanates from tosyl chloride mediated decomposition of
in situ generated dithiocarbamic acid salts. J Org Chem. 2007;72:3969.
43 For excess CS2 or excess TEA affects the reaction see: Li G, Tajima H, Ohtani T. An
improved procedure for the preparation of isothiocyantes from primary amine by
using hydrogen peroxide as the dehydrosulfurization reagent J Org Chem. 1997;62:
4539.
23 Pei Y, Mercier RW, Anday JK, et al. Ligand-binding architecture of human CB2
cannabinoid receptor: evidence for receptor subtype-specific binding motif and
modeling GPCR activation. Chem Biol. 2008;15:1207.
24 Szymanski DW, Papanastasiou M, Melchior K, et al. Mass spectrometry-based
proteomics of human cannabinoid receptor 2: covalent cysteine 6.47 (257)-ligand
interaction affording megagonist receptor activation. J Proteome Res. 2011;10:4789.
25 The cLogP and tPSA values were calculated using ChemDraw Professional, version
16.0.1.4 (77), PerkinElmer Informatics, Inc.
44 Tius MA, Hill WAG, Zou XL, et al. Classical/non-classical cannabinoid hybrids:
stereochemical requirements for the southern hydroxyalkyl chain. Life Sci. 2007;
1995:56.
26 Kulkarni S, Nikas SP, Sharma R, et al. Novel C-ring-hydroxy-substituted controlled
deactivation cannabinergic analogues. J Med Chem. 2016;59:6903.
27 Muraoka O, Wang Y, Okumura M, Nishiura S, Tanabe G, Momose T. A facile
synthesis of 7-methylenebicyclo-[3.3.1]nonan-3-one and its transformation leading
to the novel tricylic system, protoadamantane. Synth Commun. 1996;26:1555.
28 Shibuya M, Tomizawa M, Suzuki I, Iwabuchi Y. 2-Azaadamantane N-oxyl (AZADO)
and 1-Me-AZADO: highly efficient organocatalysts for oxidation of alcohols. J Am
Chem Soc. 2006;128:8412.
45 Fitjer L, Quabeck U. The Wittig reaction using potassium-tert-butoxide high yield
methylenations of sterically hindered ketones. Synth Commun. 1985;15:855.
́
46 Conia J-M, Limasset J-C. L’utilisation du t-amylate de sodium dans les reactions de
Wittig. Bull Soc Chim Fr. 1936;1967:6.
47 Ho TC, Shimada N, Tius MA, Nikas SP, Zhang W, Makriyannis A. C1’-azacycloalkyl
hexahydrocannabinols. J Org Chem. 2017;82:7839.
48 Alapafuja SO, Nikas SP, Ho TC, Tong F, Benchama O, Makriyannis A. Chain
substituted cannabilactones with selectivity for the CB2 cannabinoid receptor.
Molecules. 2019;24:3559.
29 Mori T, Yang KH, Kimoto K, Nozaki H. Photochemistry of bicyclo [3.3.1] nonanes
having two functional groups at 3,7-positions. Tetrahedron Lett. 1970;28:2419.
30 Momose T, Atarashi S. Bicyclo[3.3.1]nonanes as synthetic intermediates. V. The
Baeyer-Villiger oxidation of bicyclo[3.3.1]nonane-3,7-dione and its congeners. Chem
Pharm Bull. 1979;27:824.
49 Khanolkar AD, Lu D, Ibrahim M, et al. Cannabilactones: a novel class of CB2 selective
agonists with peripheral analgesic activity. J Med Chem. 2007;50:6493.
50 Nikas SP, Alapafuja SO, Papanastasiou I, et al. Novel 1’,1’-chain substituted
hexahydrocannabinols: 9beta-1-hydroxy-3-(1-hexyl-cyclobut-l-yl)hexahydrocanna-
binol (AM2389) a highly potent cannabinoid receptor 1 (CB1) agonist. J Med Chem.
2010;53:6996.
31 For the preparation of aryllithium by bromine/lithium exchange see: Muller D,
Guenee L, Alexakis A. Practical synthesis of simplephos ligands: further development
of alkyl-substituted phosphanamines Eur. J. Org. Chem. 2013;28:6335.
32 Imamoto T, Kusumoto T, Tawarayama Y, et al. Carbon-carbon bond-forming
reactions using cerium metal or organocerium(III) reagents. J Org Chem. 1984;49:
3904.
51 Nikas SP, Sharma R, Paronis CA, et al. Probing the carboxyester side chain in
controlled deactivation (-)-delta(8)-tetrahydrocannabinols. J Med Chem. 2015;58:
665.
52 Picone RP, Khanolkar AD, Xu W, et al. (ꢀ )-7′-Isothiocyanato-11-hydroxy-1′,1′-
dimethylheptylhexahydrocannabinol (AM841), a high-affinity electrophilic ligand,
interacts covalently with a cysteine in helix six and activates the CB1 cannabinoid
receptor. Mol Pharmacol. 2005;68:1623.
33 Mlinaric-Majerski K, Kragol G, Ramljak TS. Transannular cyclization with Grignard
reagents: facile synthetic routes to oxaadamantane and protoadamantane
derivatives. Synlett. 2008;3:405.
7