Highly Selective, Amine-Derived Cannabinoid Receptor 2 Probes

Article abstract of DOI:10.1002/chem.201904584

The endocannabinoid (eCB) system is implied in various human diseases ranging from central nervous system to autoimmune disorders. Cannabinoid receptor 2 (CB₂R) is an integral component of the eCB system. Yet, the downstream effects elicited by this G protein-coupled receptor upon binding of endogenous or synthetic ligands are insufficiently understood—likely due to the limited arsenal of reliable biological and chemical tools. Herein, we report the design and synthesis of CB₂R-selective cannabinoids along with their in vitro pharmacological characterization (binding and functional studies). They combine structural features of HU-308 and AM841 to give chimeric ligands that emerge as potent CB₂R agonists with high selectivity over the closely related cannabinoid receptor 1 (CB₁R). The synthesis work includes convenient preparation of substituted resorcinols often found in cannabinoids. The utility of the synthetic cannabinoids in this study is showcased by preparation of the most selective high-affinity fluorescent probe for CB₂R to date.

Full text of DOI:10.1002/chem.201904584

DOI: 10.1002/chem.201904584
Full Paper
&
Chemical Probes |Hot Paper|
Highly Selective, Amine-Derived Cannabinoid Receptor 2 Probes**
[
a]
[a]
[b]
[b]
Matthias V. Westphal, Roman C. Sarott, Elisabeth A. Zirwes, Anja Osterwald,
[b]
[b]
[b]
[a]
Wolfgang Guba, Christoph Ullmer, Uwe Grether, and Erick M. Carreira*
Abstract: The endocannabinoid (eCB) system is implied in
various human diseases ranging from central nervous
system to autoimmune disorders. Cannabinoid receptor 2
and functional studies). They combine structural features of
HU-308 and AM841 to give chimeric ligands that emerge as
potent CB R agonists with high selectivity over the closely
2
(
CB R) is an integral component of the eCB system. Yet, the
related cannabinoid receptor 1 (CB R). The synthesis work in-
2
1
downstream effects elicited by this G protein-coupled recep-
tor upon binding of endogenous or synthetic ligands are in-
sufficiently understood—likely due to the limited arsenal of
reliable biological and chemical tools. Herein, we report the
cludes convenient preparation of substituted resorcinols
often found in cannabinoids. The utility of the synthetic can-
nabinoids in this study is showcased by preparation of the
most selective high-affinity fluorescent probe for CB R to
2
design and synthesis of CB R-selective cannabinoids along
date.
2
with their in vitro pharmacological characterization (binding
Introduction
shown to be upregulated in brain microglia during neuroin-
[7]
flammation, while its expression in cells of the healthy central
9
[8]
The recognition of (À)-D -trans-tetrahydrocannabinol as the
nervous system is still under debate. Investigations of CB2R
[1]
main active ingredient in Cannabis sativa preparations marks
the beginning of a series of important discoveries that unrav-
eled a lipid signaling system referred to as endocannabinoid
biology are made difficult due to low expression levels (if pres-
ent at all), the inducible nature of CB R and the lack of reliable
2
[9]
and selective protein detection tools, such as antibodies. In
[
2]
[10]
(
eCB) system. The eCB system is found in all vertebrates and
general, functional derivatives of small molecules are an al-
comprises (at least) two G protein-coupled receptors known as
cannabinoid receptors 1 and 2 (CB R and CB R), endogenous
ternative that may specifically engage targets and enable ap-
plications such as fluorescence-activated cell sorting and con-
focal microscopy (fluorescent probes), activity/affinity-based
1
2
lipidic ligands including N-arachidonoylethanolamine (also
[11,12]
known as anandamide) and 2-arachidonoyl glycerol, as well as
protein profiling (electrophilic or photoactivatable ligands)
[
3]
[13]
enzymes responsible for ligand metabolism. The eCB system
is implied in numerous processes such as memory, nociception
and targeted protein degradation (PROTACs), to name just a
few.
[
4]
and immune response, and virtually all components of the
eCB system have been considered as promising targets for the
treatment of psychiatric disorders, autoimmune diseases,
The search for an irreversible CB R-selective ligand
2
[
5,6]
cancer and various other conditions.
Yet, the underlying
In combination with protein engineering, small molecule li-
mechanism of CB R- and CB R-mediated effects upon receptor
gands for CB R could help identifying a structural basis for
1
2
2
[14,15]
modulation is insufficiently understood, especially in a tissue-
CB R’s pronounced biased signaling.
Chemical probes that
2
and disease-dependent context. CB R in particular, commonly
bind irreversibly would be particularly useful to investigate bio-
logical effects following continued receptor modulation, as sta-
bilizing agents for crystallographic studies, and for antibody
generation. However, despite decades of cannabinoid receptor
2
referred to as peripheral cannabinoid receptor, has been
[a] Dr. M. V. Westphal, R. C. Sarott, Prof. E. M. Carreira
Laboratorium fꢀr Organische Chemie
research, no selective, irreversible binder for CB R has been re-
2
Eidgençssische Technische Hochschule Zꢀrich
Vladimir-Prelog-Weg 3, 8093 Zꢀrich (Switzerland)
E-mail: erickm.carreira@org.chem.ethz.ch
ported.
We hypothesized that hybrid structures derived from two
prominent cannabinoid receptor ligands, namely HU-308 and
[
b] E. A. Zirwes, A. Osterwald, Dr. W. Guba, Dr. C. Ullmer, Dr. U. Grether
Roche Innovation Center Basel, F. Hoffmann-La Roche Ltd.
Grenzacherstrasse 124, 4070 Basel (Switzerland)
[16,17]
AM841,
could serve as leads towards CB R-selective cova-
2
lent ligands (Figure 1). HU-308 is a full agonist of CB R (cAMP
2
[
**] A previous version of this manuscript has been deposited on a preprint
assay: hCB R EC ꢀ6 nm) and has been appreciated for its
2
50
server (https://doi.org/10.26434/chemrxiv.9919022.v1).
high affinity (hCB R K ꢀ20 nm) and pronounced selectivity
2
i
Supporting information and the ORCID identification number(s) for the au-
thor(s) of this article can be found under:
https://doi.org/10.1002/chem.201904584.
[
15,16]
over CB R.
The compound features an aliphatic, primary
1
[18]
hydroxyl group and an arene bearing an aliphatic sidechain.
Chem. Eur. J. 2020, 26, 1 – 9
1
ꢀ 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
side chain of AM841 commenced with preparation of known
resorcinol derivative 7 (Scheme 1). The published route to-
wards 7 involves organolithium addition to Weinreb amide 3
(
62% yield), TiMe Cl -mediated conversion of ketone 6 into the
2 2
Scheme 1. Synthesis of resorcinol 7. Reagents and conditions: a) Fe(acac)
3 mol%), THF, À788C, 90%. b) TiCl , ZnMe , CH Cl , À308C to rt, 34%.
c) BBr , CH Cl
3
(
4
2
2
2
Figure 1. A hybrid of HU-308 and AM841 featuring structural elements re-
sponsible for CB R selectivity (pinene core, capped phenol/resorcinol) and ir-
3
2
2
, À788C to rt, quantitative as inseparable mixture of 7 and sec-
ondary bromide 7’ (ca. 9:1).
2
reversible binding (isothiocyanate).
In contrast to classical cannabinoids as defined by the pres-
ence of a tricyclic benzochromene motif, HU-308 is a pinene
derivative lacking the central pyran. HU-308 features two
methyl ethers that were shown to be an important contribu-
corresponding geminal dimethyl derivative as described by
[26]
Reetz and subsequent global ether cleavage with BBr to
3
[27]
afford 7 (84% yield over two steps). We found that 6 could
be conveniently prepared by Fe(acac) -catalyzed (acac=acety-
3
[
6]
[28]
ting factor to high CB R selectivity. As a classical cannabinoid,
lacetonate) cross-coupling of readily available acyl chloride 4
2
[
29]
AM841 includes a phenol along with a primary aliphatic alco-
hol. Its most notable feature is the electrophilic isothiocyanate
at the terminus of the aliphatic sidechain, which can enable
and Grignard reagent 5
in 90% yield (ꢀ12.3 g). In our
hands, subsequent conversion of ketone 6 into its geminal di-
methyl derivative using TiMe Cl (generated in situ from TiCl
4
2
2
[
21]
[30]
cross-linking of protein targets. AM841 is a full agonist of
and ZnMe ) proceeded in 34% yield. Global ether cleavage
2
[
27]
both CB R and CB R (cAMP assay: hCB R EC ꢀ1 nm, hCB R
according to Tius’ BBr protocol proceeded cleanly as judged
1
2
1
50
2
3
1
EC <1 nm) and has been described as a ligand that binds ir-
by thin layer chromatography. However, inspection of H and
50
13
reversibly to both receptors on the basis of radioligand satura-
C NMR spectra revealed the presence of a side product iden-
[
4,8]
tion binding studies (CB R and CB R)
as well as mass spec-
tified as isomeric secondary alkyl bromide 7’ amounting to ca.
10% of the material. The outcome of the last two steps
prompted us to devise an alternative synthetic strategy to-
wards resorcinol derivatives that would avoid the observed iso-
merization (vide infra).
1
2
[
9]
trometry-based proteomics (CB R). In these studies, the au-
2
thors concluded that Cys6.47 (Ballesteros–Weinstein number-
[
24]
ing), a conserved amino acid present in both CB R and CB R,
1
2
forms covalent adducts with the electrophilic isothiocyanate of
AM841. Surprisingly, a recent structure resulting from co-crys-
At this stage of the project, although the synthesis of 7 was
suboptimal, we decided to venture ahead to collect initial
pharmacological data. In analogy to the published synthesis of
tallization of CB R with AM841 shows the aliphatic sidechain of
1
[
25]
the latter pointing away from Cys6.47. While this observation
renders a putative covalent adduct geometrically unlikely for
[16]
HU-308, pTsOH-mediated Friedel–Crafts alkylation of 7 using
allylic alcohol 8 derived from (+)-a-pinene followed by double
methylation afforded 9 (Scheme 2). Ester reduction (by means
of diisobutylaluminum hydride, DIBAL-H) and nucleophilic dis-
placement of the derived alkyl bromide with sodium azide de-
livered 11 (93% yield over both steps), which was further ela-
borated into 1. Alternatively, 11 was first propargylated prior
to conversion of the azide to the corresponding isothiocya-
nate. The latter two-step sequence afforded bifunctional
ligand 13 as putatively irreversible probe amenable to further
derivatization for use in activity-based protein profiling experi-
ments.
CB R, it is important to note that the situation may be different
1
in the case of CB R, for which a covalent interaction with
2
[
22]
AM841 had been shown by mass spectrometry. Herein, we
describe the design, synthesis and in vitro pharmacological
evaluation of 1 (X=OH) and various derivatives, such as ethers
and esters (X=OR) as well as amides (X=HNR). Our studies
culminate in the identification of a privileged amine platform
(
2) for the modular preparation of highly selective CB R ago-
2
nists.
Results and Discussion
In a comparative study of HU-308 and its enantiomer HU-
The synthesis of hybrid cannabinoid 1 featuring the pinene
433, the authors noted that the latter exhibited higher CB R-
2
core and methyl ethers of HU-308 as well as the electrophilic
mediated biological activity than HU-308. This contrasted the
&
&
Chem. Eur. J. 2020, 26, 1 – 9
www.chemeurj.org
2
ꢀ 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Scheme 2. Synthesis of electrophilic hybrid cannabinoids 1 and 13. Reagents
and conditions: a) pTsOH·H O (0.28 equiv), CH Cl , rt, 86%; b) Me SO , K CO
acetone, rt, 73%; c) DIBAL-H, CH Cl , 08C, 94%; d) NaN , DMF, rt, 99%;
e) PPh , CS , THF, rt, 84%; f) propargyl bromide, Bu
aq. NaOH-PhMe, rt, 36%; g) PPh , CS , THF, rt, 81%.
2
2
2
2
4
2
3
,
2
2
3
3
2
4
4
NHSO (0.2 equiv), 50%
3
2
Scheme 3. Synthesis of amine 2 and fluorescent derivatives 15–17. Reagents
and conditions: a) DIAD, PPh
8%; c) N ·H O, crotyl alcohol, EtOH, 758C, 73%; d) for 15: 1, EDC·HCl,
DMAP, NEt , THF, rt, 90%; for 16: 2, EDC·HCl, NEt , THF, rt, 56%; e) PPh , CS
THF, rt, 74%.
3 3
, phthalimide, THF, rt, 62%; b) NaN , DMF, rt,
8
2
H
4
2
expectation based on radioligand binding studies, which at-
3
3
3
2
,
tested HU-308’s higher CB R affinity when compared to HU-
2
[
31]
4
33. Therefore, ent-1 was included in the present study in
order to maximize the chance of identifying irreversible and se-
lective CB R ligands. Compound ent-1 was accessed by Frie-
plored amide linkage prompted us to devise a concise synthet-
ic strategy towards amide derivatives of HU-308. In addition,
an expedient route to resorcinol derivatives that a) avoids con-
tamination with constitutional isomers and b) allows for intro-
duction of other functionalities in the side chain was devel-
oped (Scheme 4). The latter is of special interest regarding in-
corporation of polar functional groups to effect reduction of
non-specific binding, an issue known to correlate with lipophi-
2
del–Crafts alkylation with ent-8 derived from (1R)-(À)-myrtenol
and subsequent elaboration analogous to the sequence shown
in Scheme 2 (see Supporting Information).
Drawing from a report on biotin-conjugated HU-308, in
which elongation along the allylic alcohol was tolerated by
[
32]
CB R, fluorescent probe 15 was prepared by esterification of
2
1
with nitrobenzoxadiazole-derived acid 14 (Scheme 3). Since
[38]
esters are prone to hydrolysis in biological environments,
amide analogs of 15 were synthesized. To this end, amine 2
was prepared by a three-step sequence involving Mitsunobu
reaction of allylic alcohol 10 with potassium phthalimide, bro-
mide displacement with sodium azide and, finally, treatment
with hydrazine. Subsequent amide formation and conversion
of azide to isothiocyanate afforded fluorescent compounds 16
and 17 (see Supporting Information for fluorescence spectra).
In vitro pharmacological assessment of prepared com-
pounds (vide infra) showed that amide derivatives 16 and 17
licity (logD).
Following a patent procedure, (+)-a-pinene (18) was oxi-
dized to verbenone by CrO in presence of N-hydroxyphthali-
3
[39]
mide.
Subsequent allylic bromination under Wohl-Ziegler
conditions afforded 19. The two-step sequence involving bro-
mide displacement and ketone reduction could be carried out
in either order, although it proved convenient to first reduce
ketone 19 and then subject the product bromide to displace-
ment by azide. This order of events yielded allyl alcohol 20 in
77% yield over two steps and avoided capricious chemoselec-
tive ketone reduction in presence of phthalimide under Luche
conditions.
exhibited significantly increased binding affinity to hCB R com-
2
pared to ester 15, as well as sharply increased selectivity over
hCB R. Interestingly, cannabinoids in which the primary alcohol
The revised resorcinol synthesis commenced with addition
1
[40]
is replaced by amines and their derivatives have not been
studied in detail. The few examples include an acetamide de-
of Grignard reagent 22 to ketoester 21. Alternatively, 25 was
added to ketone 24. The resulting tertiary alcohols 23 and 26
were converted into their gem-dimethyl derivatives by modifi-
8
8
rivative of D -tetrahydrocannabinol (D -THC) reported to be
[
33]
[41]
analgesically inactive and more recent studies of fluorescent
cation of a reported procedure. First, treatment with SOCl2
[
34,35]
probes for the cannabinoid receptors.
A small number of
(for 23) or conc. HCl (for 26) resulted in formation of the corre-
sponding tertiary chlorides, which without purification were
[
36]
[37]
amine analogs was evaluated in mice and baboons, but
since these studies predate the discovery and cloning of can-
nabinoid receptors, no data on binding affinities are available.
The apparent lack of structure–activity information and the
treated with AlMe to install the gem-dimethyl group. The un-
3
purified mixtures were subjected to oxone/cat. OsO to effect
4
oxidative degradation of the otherwise inseparable olefin by-
products (ca. 10%) formed by elimination of the tertiary alco-
observed high CB R affinity of ligands bearing the underex-
2
Chem. Eur. J. 2020, 26, 1 – 9
www.chemeurj.org
3
ꢀ 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
hol over the course of installation of the gem-dimethyl
[42]
group. This procedure allowed for convenient isolation of
the desired dimethyl resorcinols 27 and 28 (67% and 42%
yield, respectively). Cleavage of the methyl ethers with BBr3
cleanly afforded resorcinols 29 (93% yield) and 30 (97% yield)
with no detectable formation of constitutional isomers. Each of
these was then allowed to react with 20 and further elaborat-
ed into amines 2, 31 and 32 following the logic outlined in
Scheme 2 (see Supporting Information). Additionally, known re-
sorcinol 30b was subjected to the same sequence to afford
32. It is worth noting that THC derivatives bearing a terminal
carboxylate in the side chain have been employed as haptens
[43]
for antibody generation.
With allylic amines 2, 31, and 32 in hand, straightforward de-
rivatization reactions afforded a number of aza-HU-308 deriva-
tives shown in Figure 2. Probes 33–35, which may be activated
[44]
upon irradiation, were synthesized by amidation of the cor-
responding amines (2, 31, and 32) and recently applied in a
[45]
study on CB R photoaffinity-labeling. In this work, 33 suc-
2
cessfully labelled the receptor upon irradiation (350 nm) as
shown by in-gel fluorescence following fluorophore attach-
ment.
Conjugation of 4-pentynoic acid and subsequent triazole for-
mation with 1-azidoadamantane afforded compounds 36 and
37. Amide 38 and sulfonamide 39 were prepared by condensa-
tion reactions with biotin and 1-butanesulfonyl chloride, re-
spectively. Amide formation with tetra-ortho-chloro-azoben-
[
46]
Scheme 4. A) Synthesis of N-protected allylic alcohol. Reagents and condi-
tions: a) CrO , N-hydroxyphthalimide, acetone, rt, 44%; b) NBS, dibenzoyl
peroxide, CCl , reflux, 62%; c) potassium phthalimide, DMF, rt, 85%;
d) CeCl ·7H O, NaBH O-hexanes; then
zene containing fatty acid FAAzo4 afforded photoswitch 40
and completed the set of amine-derived compounds tested in
this study.
3
4
3
2
4
, MeOH, À788C, 93%; e) DIBAL-H, Et
2
acidic workup; then potassium phthalimide, DMSO, rt, 77% over 2 steps.
B) Synthesis of resorcinol derivatives and further elaboration into allyl
amines. Reagents and conditions: f) 22, THF, À788C to rt, 92%; g) 25, THF,
In vitro pharmacological characterization
À788C to rt, 60%; h) for 22: SOCl
CH Cl
, À788C to rt; j) OsO
8: 67% from 26; k) BBr , CH
2 3
, 08C; for 28: HCl conc., rt; i) AlMe ,
All synthesized compounds were evaluated in radioligand
binding studies with tritiated CP55,940 using membrane prep-
arations of Chinese hamster ovary (CHO) cells overexpressing
2
2
4
(cat.), oxone, DMF, rt, for 27: 42% from 23, for
Cl
, À788C to rt, 97% (29), 93% (28).
2
3
2
2
Figure 2. Additional aza-HU-308 derivatives tested in this study. Compounds were prepared by condensation reactions of the corresponding allylic amines.
7 was synthesized from 36 by copper-catalyzed cycloaddition. See Supporting Information for experimental details.
3
&
&
Chem. Eur. J. 2020, 26, 1 – 9
www.chemeurj.org
4
ꢀ 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
[
a]
Table 1. In vitro pharmacological assessment of HU-308 derivatives.
#
R
1
R
2
hCB
nm]
2
R K
i
mCB
[nm]
2
R K
i
1
hCB R K
[nm]
i
ratio K
hCB R/hCB
i
hCB
[nm]
2
R EC50
mCB
[nm]
2
R EC50
hCB
[nm]
1
R EC50
[
1
2
R
(
%efficacy)
(%efficacy)
(%efficacy)
1
13.1
121
67
7.8
56
3.7
158
2670
>10000
>10000
670
204
>83
>149
86
0.6 (95)
11.6 (95)
n/a
1.7 (95)
65 (97)
n/a
>10000
>10000
n/a
ent-1
1320
1780
59
2
10-Br
0.9 (99)
7.5 (98)
0.6 (99)
3.0 (95)
0.7 (98)
9.0 (98)
0.5 (96)
5.5 (97)
178 (80)
>10000
100 (79)
>10000
ent-10-Cl
760
40
3150
56
11
1130
305
117
ent-11
33
428
3860
1
2
3
116
1080
530
>10000
>86
0.5 (101)
0.2 (95)
1.9 (97)
1.8 (94)
28.4 (41)
1380 (58)
1
45
6160
137
1
1
1
5
6
7
900
1580
n/a
>10000
>10000
>10000
>11
>2381
>2128
41 (95)
n/a
>10000
n/a
>10000
n/a
4.2
4.7
9.3
78
0.5 (97)
2.6 (97)
>10000
33
34
35
36
37
88
417
3890
357
418
4
0.9 (98)
17.0 (96)
1.9 (97)
0.7 (100)
168 (97)
13.9 (93)
0.3 (100)
128 (85)
>10000
120 (106)
>10000
>10000
>10000
96
151
7.8
3220
6430
780
43
100
1
128
0.5 (101)
197 (91)
1380
>10000
1340
3
8
9
88
314
2760
31
1.6 (100)
1.8 (99)
1.2 (100)
>10000
>10000
3
147
5400
>10000
>68
16.2 (102)
40
41
389
3630
89
9.3 (99)
12.9 (96)
650 (82)
[
[
a] Dissociation constants (K
i
) were determined by radioligand binding assay using membrane preparations of CB
1 2
R/CB R overexpressing CHO cells with
3
H]-CP55,940 as described previously and are given as average values from one up to six independent experiments each performed in triplicate. cAMP
[
47]
1 2
assays were performed with CHO cells expressing human CB R and CB R receptor variants as described. Efficacies are expressed as percentages relative
to the effect evoked by CP55,940 (1 mm). EC50 values are averages from one to two independent experiments each performed in triplicate.
CB R (human and mouse) or human CB R, respectively. All ex-
functional activity of prepared HU-308 derivatives. Results are
listed in Table 1.
2
1
periments were performed using aliquots of the same batches
of isolated CHO-membranes expressing the respective recep-
tor. In addition, widely used cyclic adenosine monophosphate
Compounds 1, 10 and 11 share the free allylic alcohol and
exhibit distinct groups terminating the dimethylheptyl chain.
[
47]
(
cAMP) assay was employed as described earlier to assess
These compounds emerged as high affinity hCB R binders
2
Chem. Eur. J. 2020, 26, 1 – 9
www.chemeurj.org
5
ꢀ 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
(Ki hCB R<13 nm) with at least 86-fold selectivity over hCB R.
2
1
[
31]
In line with Mechoulam’s report, compounds ent-1, ent-10-Cl
and ent-11 showed lower hCB R affinity than their enantiomers
2
by a factor of ca. 10 (hCB R K <121 nm) but retained high se-
2
i
lectivity over hCB R (hCB R K >3.1 mm). Compared to alco-
1
1
i
hol 11, free amine 2 exhibited reduced affinity to human and
mouse CB R (hCB R K =67 nm, mCB R K =1780 nm) but
2
2
i
2
i
showed no detectable interaction with hCB R. Nitrobenzofura-
1
zan-derived compounds 15–17 were identified as highly prom-
ising hCB R-selective fluorescent probes given their complete
2
selectivity over CB R. Importantly, exchanging the potentially
1
vulnerable ester functionality in 15 for an amide linkage result-
ed in analogs 16 and 17 with significantly increased CB R affin-
2
ity and remarkable selectivity over CB R (ratio hCB R K/hCB R
1
1
i
2
[
48]
Ki >2000). The primary amide in 34 proved detrimental to
receptor selectivity. Comparing compounds 33–35 differing
only in the side chain, the terminal azide emerged as most fa-
Figure 3. Docking pose of 17 within the crystal structure of CB
in complex with antagonist AM10257 (PDB code: 5ZTY). The model suggests
formation of a hydrogen bond (magenta) between 17 and Ser72 (bold
ribbon).
2
R determined
vorable for achieving high hCB R-selectivity. Introduction of
2
the terminal azide (33) resulted in ca. 10-fold increase in selec-
[
35]
tivity compared to its saturated analog 35. Adamantyl-sub-
stituted triazole and biotin derivatives 37 and 38 exhibited de-
creased hCB R affinity along with low selectivity over hCB R.
creased binding affinity of 17 (and other amide analogs) com-
pared to ester 15. The hypothesis is further corroborated by
pairwise comparison of compounds 11/12 and 1/13. In both
cases, analogs bearing a free hydroxy group as potential H-
bond donor (1, 11) exhibit higher binding affinity than the cor-
responding propargyl ether analogs (12, 13).
2
1
Sulfonamide 39 showed affinity comparable to 35, while azo-
benzene derivative 40 (as undefined mixture of trans- and cis-
isomers) was found to be a hCB R-selective high-affinity ligand.
2
Interestingly, all compounds tested showed a preference for
human CB R over the mouse homolog. Human and mouse
2
CB R share 86% sequence identity within the ligand binding
2
domain differing only in amino acids 72 and 261. While hCB2R
Probing for irreversible binding
contains amino acids Ser72 and Val261, mCB R displays Asn72
2
and Ala261 residues. These alterations seem to create a bind-
ing pocket less effective in accommodating HU-308 deriva-
tives.
Earlier reports describe saturation binding experiments to
show AM841’s ability to covalently bind both CB R and CB R
1
2
[17,23]
via Cys6.47.
In these experiments, membranes expressing
In the functional cAMP assay, all compounds were identified
wildtype CB R and CB R were preincubated with AM841 or
1
2
as full agonists of both human and mouse CB R with relative
DMSO control. Following excessive washing steps, saturation
binding experiments with tritiated CP55,940 revealed de-
creased receptor density (Bmax) for pretreated membranes com-
pared to control. In addition, performing the same experiment
with receptor variants in which Cys6.47 was mutated to serine,
alanine or leucine did not show reduction of Bmax when com-
paring pretreated membranes to control. As orthosteric cova-
lent binders are expected to reduce the number of available
binding sites, these observations led to the conclusion that
AM841 is an irreversible binder of both CB R and CB R.
2
EC50 values resembling the trends observed in the radioligand
binding assay, albeit not as pronounced. With the exception of
10, 11 and to some extent 40, all compounds bearing a termi-
nal NCS or N group were functionally inactive at hCB R (12:
3
1
partial hCB R agonist with 58% efficacy and hCB R EC
50
1
1
ꢀ
1.4 mm). Comparison of compounds 15 and 17 shows that
the higher binding affinity of amide 17 also translates into
higher potency in cAMP assay for both human and mouse
CB R. For the newly prepared compounds, potential off-target
2
1
2
effects are important to consider. A recent study of prototypi-
cal CB R and CB R ligands details extensive efforts towards off-
In analogy, we proceeded to evaluate compounds 1 and ent-
1 for their ability to reduce Bmax. To this end, membrane prepa-
1
2
[
15]
target identification.
The work concluded that HU-308,
rations of hCB R-overexpressing CHO-cells were incubated with
2
which shares key structural features with the compounds in
Table 1, exhibits one of the cleanest profiles in the test set.
DMSO control or putatively covalent compounds 1 (90 nm)
and ent-1 (790 nm) at concentrations corresponding to ca. 6-
fold K for 60 minutes. Following excessive washing steps to
i
remove non-covalently bound material, determination of re-
ceptor density (Bmax) using tritiated CP55,940 would allow for
indirect proof of irreversible bond formation for orthosteric li-
gands. As shown in Figure 4, saturation binding of
Docking studies
Docking agonist 17 into the recently published X-ray crystal
structure of antagonist-bound CB R (PDB code: 5ZTY) sug-
[49]
2
3
gests a favorable hydrogen bond between the amide group of
[ H]-CP55,940 using membranes preincubated with 1 or DMSO
1
7 (NÀH as H-bond donor) and the carbonyl of Ser72 (H-bond
(Figure 4A), and ent-1 or DMSO (Figure 4B) revealed similar
maximum specific binding (Bmax) within either set of experi-
acceptor) (Figure 3). This finding may explain the observed in-
&
&
Chem. Eur. J. 2020, 26, 1 – 9
www.chemeurj.org
6
ꢀ 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
ester derivatives. Interestingly, the chemotype presented dis-
plays intrinsic preference for human CB R over the mouse ho-
2
molog. This finding is important in the light of transferability
of preclinical in vivo data generated using commonly em-
ployed mouse models.
We believe the in vitro pharmacological data presented for
amide derivatives of HU-308 is a valuable addition to the struc-
ture–activity-relationship knowledge of cannabinoids. Our
work serves as blueprint for the preparation of selective CB2R
ligands, incorporating functional elements tailored to a given
biological experiment. Ongoing work in our laboratories focus-
es on additional applications of the ligands described and will
be reported as results become available.
Acknowledgements
E.M.C. is grateful to ETH-Zꢁrich and F. Hoffmann-La Roche for
support of the research program. R.C.S. acknowledges the
Scholarship Fund of the Swiss Chemical Industry (SSCI). We
thank Prof. James Frank for providing a sample of FAAzo4 and
Dr. Marjolein Soethoudt for evaluation of electrophilic fluorop-
robe 17. Dr. Sylwia Huber, Eric Kusznir and Dr. Arne Rufer are
gratefully acknowledged for the generation of absorption, exci-
tation and emission spectra.
Figure 4. Washout experiments probing for irreversible binding. Saturation
binding of [ H]-CP55,940. Specific binding in dependence on radioligand
3
concentration determined using membranes pretreated with 6ꢆ K
i
of li-
gands 1, ent-1 or DMSO.
ments. Thus, 1 and ent-1 are not considered irreversible, or-
Conflict of interest
thosteric ligands of hCB R. This outcome is in line with a sepa-
2
rate experiment aiming to detect covalent bond formation of
The authors declare no conflict of interest.
fluorescent electrophilic ligand 17. In this experiment, mem-
branes of hCB R overexpressing CHO-cells were incubated with
2
17 (up to 10 mm). When proteins of the sample were resolved
Keywords: cannabinoid receptor 2 · CB R · chemical probe ·
2
by polyacrylamide gel electrophoresis, no fluorescent band
electrophilic ligands · endocannabinoid system · fluorescent
probe
[
50]
corresponding to CB R protein was detected.
2
[
[
[
1] Y. Gaoni, R. Mechoulam, J. Am. Chem. Soc. 1964, 86, 1646.
2] D. Piomelli, Nat. Rev. Neurosci. 2003, 4, 873.
3] V. D. Marzo, M. Bifulco, L. D. Petrocellis, Nat. Rev. Drug Discovery 2004, 3,
Conclusions
Enabled by a practical synthesis of functionalized resorcinols,
the amine derivatives identified in this study allow modular
771.
[
4] R. Mechoulam, L. A. Parker, Annu. Rev. Psychol. 2013, 64, 21.
and convenient access to highly CB R-selective ligands. Conju-
2
[5] P. Pacher, S. Bꢂtkai, G. Kunos, Pharmacol. Rev. 2006, 58, 389.
[6] M. Maccarrone, I. Bab, T. Bꢃrꢄ, G. A. Cabral, S. K. Dey, V. Di Marzo, J. C.
Konje, G. Kunos, R. Mechoulam, P. Pacher, K. A. Sharkey, A. Zimmer,
Trends Pharmacol. Sci. 2015, 36, 277.
gates prepared from amine 2 in particular, bearing a terminal
azide in the dimethylheptyl side chain, emerged as ligands
with the highest affinity and selectivity for hCB R across the
2
[
[
7] B. K. Atwood, K. Mackie, Br. J. Pharmacol. 2010, 160, 467.
8] A. Lꢄpez, N. Aparicio, M. R. Pazos, M. T. Grande, M. A. Barreda-Manso, I.
Benito-Cuesta, C. Vꢂzquez, M. Amores, G. Ruiz-Pꢅrez, E. Garcꢃa-Garcꢃa,
M. Beatka, R. M. Tolꢄn, B. N. Dittel, C. J. Hillard, J. Romero, J. Neuroin-
flammation 2018, 15, 158.
series. The promise of this approach is showcased by the syn-
thesis of various functional CB R ligands, most notably fluores-
2
cent probe 16. This exceptionally selective fluoroprobe binds
hCB R with a K of 4 nm as determined by radioligand binding
2
i
[
9] Y. Marchalant, P. W. Brownjohn, A. Bonnet, T. Kleffmann, J. C. Ashton, J.
Histochem. Cytochem. 2014, 62, 395.
assay and shows no detectable interaction with hCB R. Thus,
1
our initial quest for selective CB R ligands that bind irreversibly
[10] C. H. Arrowsmith, J. E. Audia, C. Austin, J. Baell, J. Bennett, J. Blagg, C.
Bountra, P. E. Brennan, P. J. Brown, M. E. Bunnage, C. Buser-Doepner,
R. M. Campbell, A. J. Carter, P. Cohen, R. A. Copeland, B. Cravatt, J. L.
Dahlin, D. Dhanak, A. M. Edwards, M. Frederiksen, S. V. Frye, N. Gray,
C. E. Grimshaw, D. Hepworth, T. Howe, K. V. M. Huber, J. Jin, S. Knapp,
J. D. Kotz, R. G. Kruger, D. Lowe, M. M. Mader, B. Marsden, A. Mueller-
Fahrnow, S. Mꢁller, R. C. O’Hagan, J. P. Overington, D. R. Owen, S. H.
Rosenberg, R. Ross, B. Roth, M. Schapira, S. L. Schreiber, B. Shoichet, M.
Sundstrçm, G. Superti-Furga, J. Taunton, L. Toledo-Sherman, C. Walpole,
M. A. Walters, T. M. Willson, P. Workman, R. N. Young, W. J. Zuercher, Nat.
Chem. Biol. 2015, 11, 536.
2
by merging structural motifs found in HU-308 and AM841 af-
forded useful hybrid cannabinoids of type 1 and its enantiomer
ent-1. However, studies of these ligands and their derivatives
do not support the formation of covalent adducts to CB R.
2
Docking studies using the recently published crystal struc-
ture of inactive state hCB R suggest the marked affinity of the
2
amide analogs is due to a hydrogen bond between the amide
NÀH and Ser72, which is not available to the corresponding
Chem. Eur. J. 2020, 26, 1 – 9
www.chemeurj.org
7
ꢀ 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
[
[
[
11] B. F. Cravatt, A. T. Wright, J. W. Kozarich, Annu. Rev. Biochem. 2008, 77,
83.
12] P. Kleiner, W. Heydenreuter, M. Stahl, V. S. Korotkov, S. A. Sieber, Angew.
Chem. Int. Ed. 2017, 56, 1396; Angew. Chem. 2017, 129, 1417.
13] M. Toure, C. M. Crews, Angew. Chem. Int. Ed. 2016, 55, 1966; Angew.
Chem. 2016, 128, 2002.
[32] L. Martꢃn-Couce, M. Martꢃn-Fontecha, ꢈ. Palomares, L. Mestre, A. Cordo-
mꢃ, M. Hernangomez, S. Palma, L. Pardo, C. Guaza, M. L. Lꢄpez-Rodrꢃ-
guez, S. Ortega-Gutiꢅrrez, Angew. Chem. Int. Ed. 2012, 51, 6896; Angew.
Chem. 2012, 124, 7002.
3
[33] R. S. Wilson, B. R. Martin, W. L. Dewey, J. Med. Chem. 1979, 22, 879.
[34] M. Martꢃn-Fontecha, A. Angelina, B. Rꢁckert, A. Rueda-Zubiaurre, L.
Martꢃn-Cruz, W. van de Veen, M. Akdis, S. Ortega-Gutiꢅrrez, M. L. Lꢄpez-
Rodrꢃguez, C. A. Akdis, O. Palomares, Bioconjugate Chem. 2018, 29, 382.
[35] A. R. Zubiaurre, Chemical Probes for the Study of the Endogenous Canna-
binoid System. PhD Thesis, Universidad Complutense de Madrid, 2016.
[36] D. R. Compton, P. J. Little, B. R. Martin, J. W. Gilman, J. K. Saha, V. S. Jora-
pur, H. P. Sard, R. K. Razdan, J. Med. Chem. 1990, 33, 1437.
[37] H. Edery, G. Porath, R. Mechoulam, N. Lander, M. Srebnik, N. Lewis, J.
Med. Chem. 1984, 27, 1370.
[38] S. Nagar, K. Korzekwa, Drug. Metabol. Dispos. 2012, 40, 1649.
[39] P. Marwah, H. A. Lardy, US Patent US6384251 B1.
[40] S. Biswas, S. Maiti, U. Jana, Eur. J. Org. Chem. 2010, 2861.
[41] J. A. Hartsel, D. T. Craft, Q.-H. Chen, M. Ma, P. R. Carlier, J. Org. Chem.
2012, 77, 3127.
[14] L. Zhou, L. M. Bohn, Curr. Opin. Cell Biol. 2014, 27, 102.
[
15] Screening against numerous proteins of potential clinical significance
included Trp channels, GPR55, metabolizing enyzmes of the eCB
system and a customized panel of 64 proteins associated with adverse
side effects in humans. See: M. Soethoudt, U. Grether, J. Fingerle, T. W.
Grim, F. Fezza, L. de Petrocellis, C. Ullmer, B. Rothenhꢇusler, C. Perret, N.
van Gils, D. Finlay, C. MacDonald, A. Chicca, M. D. Gens, J. Stuart, H.
de Vries, N. Mastrangelo, L. Xia, G. Alachouzos, M. P. Baggelaar, A. Mar-
tella, E. D. Mock, H. Deng, L. H. Heitman, M. Connor, V. Di Marzo, J.
Gertsch, A. H. Lichtman, M. Maccarrone, P. Pacher, M. Glass, M. van der
Stelt, Nat. Commun. 2017, 8, 13958.
[
[
[
16] L. Hanu sˇ , A. Breuer, S. Tchilibon, S. Shiloah, D. Goldenberg, M. Horowitz,
R. G. Pertwee, R. A. Ross, R. Mechoulam, E. Fride, Proc. Natl. Acad. Sci.
USA 1999, 96, 14228.
17] R. P. Picone, A. D. Khanolkar, W. Xu, L. A. Ayotte, G. A. Thakur, D. P. Hurst,
M. E. Abood, P. H. Reggio, D. J. Fournier, A. Makriyannis, Mol. Pharmacol.
[42] B. R. Travis, R. S. Narayan, B. Borhan, J. Am. Chem. Soc. 2002, 124, 3824.
[43] L. Qi, N. Yamamoto, M. M. Meijler, L. J. Altobell, G. F. Koob, P. Wirsching,
K. D. Janda, J. Med. Chem. 2005, 48, 7389.
[44] Z. Li, P. Hao, L. Li, C. Y. J. Tan, X. Cheng, G. Y. J. Chen, S. K. Sze, H.-M.
Shen, S. Q. Yao, Angew. Chem. Int. Ed. 2013, 52, 8551; Angew. Chem.
2013, 125, 8713.
2005, 68, 1623.
18] G. A. Thakur, S. P. Nikas, C. Li, A. Makriyannis, in Cannabinoids, ed. by
Professor Dr Roger G. Pertwee, Springer Berlin Heidelberg, 2005,
pp. 209–246.
[45] For detailed experimental data concerning irreversible CB R photoaffini-
2
[
[
19] J. W. Huffman, J. Liddle, S. Yu, M. M. Aung, M. E. Abood, J. L. Wiley, B. R.
Martin, Bioorg. Med. Chem. 1999, 7, 2905.
20] J. W. Huffman, S. Yu, V. Showalter, M. E. Abood, J. L. Wiley, D. R. Comp-
ton, B. R. Martin, R. D. Bramblett, P. H. Reggio, J. Med. Chem. 1996, 39,
ty-labeling using 33 and visualization by in-gel fluorescence, see: M.
Soethoudt, S. C. Stolze, M. V. Westphal, L. van Stralen, A. Martella, E. J.
van Rooden, W. Guba, Z. V. Varga, H. Deng, S. I. van Kasteren, U. Grether,
A. P. IJzerman, P. Pacher, E. M. Carreira, H. S. Overkleeft, A. Ioan-Facsinay,
L. H. Heitman, M. van der Stelt, J. Am. Chem. Soc. 2018, 140, 6067–
6075.
3875.
[
[
21] D. Weichert, P. Gmeiner, ACS Chem. Biol. 2015, 10, 1376.
22] D. W. Szymanski, M. Papanastasiou, K. Melchior, N. Zvonok, R. W. Mer-
cier, D. R. Janero, G. A. Thakur, S. Cha, B. Wu, B. Karger, A. Makriyannis, J.
Proteome Res. 2011, 10, 4789.
[46] J. A. Frank, M. Moroni, R. Moshourab, M. Sumser, G. R. Lewin, D. Trauner,
Nat. Commun. 2015, 6, 7118.
[47] C. Ullmer, S. Zoffmann, B. Bohrmann, H. Matile, L. Lindemann, P. J. Flor,
P. Malherbe, Br. J. Pharmacol. 2012, 167, 1448.
[
[
[
23] Y. Pei, R. W. Mercier, J. K. Anday, G. A. Thakur, A. M. Zvonok, D. Hurst,
P. H. Reggio, D. R. Janero, A. Makriyannis, Chem. Biol. 2008, 15, 1207.
24] J. A. Ballesteros, H. Weinstein, in Methods in Neurosciences, ed. by Stuart
C. Sealfon, Academic Press, 1995, Vol. 25, pp. 366–428.
25] T. Hua, K. Vemuri, S. P. Nikas, R. B. Laprairie, Y. Wu, L. Qu, M. Pu, A.
Korde, S. Jiang, J.-H. Ho, G. W. Han, K. Ding, X. Li, H. Liu, M. A. Hanson,
S. Zhao, L. M. Bohn, A. Makriyannis, R. C. Stevens, Z.-J. Liu, Nature 2017,
[48] NBD dye is not ideal for imaging CB R both in terms of photophysical
2
and physicochemical properties, as a reviewer has noted. In this work,
NBD serves as inexpensive fluorescent dye for initial in vitro pharmaco-
logical evaluation of fluorescent probes derived from amine 2, and as
such a point of departure. CB R imaging studies employing more suita-
2
ble fluorescent dyes will be reported in due course.
5
47, 468.
[49] X. Li, T. Hua, K. Vemuri, J.-H. Ho, Y. Wu, L. Wu, P. Popov, O. Benchama, N.
Zvonok, K. Locke, L. Qu, G. W. Han, M. R. Iyer, R. Cinar, N. J. Coffey, J.
Wang, M. Wu, V. Katritch, S. Zhao, G. Kunos, L. M. Bohn, A. Makriyannis,
R. C. Stevens, Z.-J. Liu, Cell 2019, 176, 459.
[
[
[
26] M. T. Reetz, J. Westermann, S.-H. Kyung, Chem. Ber. 1985, 118, 1050.
27] M. A. Tius, Chem. Commun. 1997, 1867.
28] B. Scheiper, M. Bonnekessel, H. Krause, A. Fꢁrstner, J. Org. Chem. 2004,
6
9, 3943.
[50] This work has been done in collaboration with Prof. M. van der Stelt at
Leiden University, and details will be reported in due course.
[
[
[
29] F. M. Piller, P. Appukkuttan, A. Gavryushin, M. Helm, P. Knochel, Angew.
Chem. Int. Ed. 2008, 47, 6802; Angew. Chem. 2008, 120, 6907.
30] C. Chu, A. Ramamurthy, A. Makriyannis, M. A. Tius, J. Org. Chem. 2003,
68, 55.
31] R. Smoum, S. Baraghithy, M. Chourasia, A. Breuer, N. Mussai, M. Attar-
Namdar, N. M. Kogan, B. Raphael, D. Bolognini, M. G. Cascio, P. Marini,
R. G. Pertwee, A. Shurki, R. Mechoulam, I. Bab, Proc. Natl. Acad. Sci. USA
Manuscript received: October 7, 2019
Revised manuscript received: November 19, 2019
Version of record online: && &&, 0000
2015, 112, 8774.
&
&
Chem. Eur. J. 2020, 26, 1 – 9
www.chemeurj.org
8
ꢀ 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
FULL PAPER
The design and synthesis of CB R-se-
lective cannabinoids along with their in
vitro pharmacological characterization
2
&
Chemical Probes
M. V. Westphal, R. C. Sarott, E. A. Zirwes,
A. Osterwald, W. Guba, C. Ullmer,
U. Grether, E. M. Carreira*
(binding and functional studies) is re-
ported. Chimeric ligands emerge as
potent CB R agonists with high selectivi-
2
&& – &&
ty over the closely related cannabinoid
receptor 1 (CB R).
Highly Selective, Amine-Derived
Cannabinoid Receptor 2 Probes
1
Chem. Eur. J. 2020, 26, 1 – 9
www.chemeurj.org
9
ꢀ 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ