Synthesis and CB1 receptor activities of dimethylheptyl derivatives of 2-arachidonoyl glycerol (2-AG) and 2-arachidonyl glyceryl ether (2-AGE)

Article abstract of DOI:10.1016/j.bmc.2005.12.007

Results from a factor analysis and activity studies of commercially available endocannabinoid-type compounds set the starting point for the current study where dimethylheptyl (DMH) analogues of two endocannabinoids, 2-arachidonoyl glycerol (2-AG) and 2-ar

Full text of DOI:10.1016/j.bmc.2005.12.007

Bioorganic & Medicinal Chemistry 14 (2006) 2850–2858
Synthesis and CB1 receptor activities of dimethylheptyl
derivatives of 2-arachidonoyl glycerol (2-AG) and
2-arachidonyl glyceryl ether (2-AGE)
Teija Parkkari,a,* Outi M. H. Salo,a Kristiina M. Huttunen,a Juha R. Savinainen,b
Jarmo T. Laitinen,b Antti Poso,a Tapio Nevalainena and Tomi Ja¨rvinena
aDepartment of Pharmaceutical Chemistry, University of Kuopio, PO Box 1627, FIN-70211 Kuopio, Finland
bDepartment of Physiology, University of Kuopio, PO Box 1627, FIN-70211 Kuopio, Finland
Received 10 October 2005; revised 1 December 2005; accepted 2 December 2005
Available online 5 January 2006
Abstract—Results from a factor analysis and activity studies of commercially available endocannabinoid-type compounds set the
starting point for the current study where dimethylheptyl (DMH) analogues of two endocannabinoids, 2-arachidonoyl glycerol
(2-AG) and 2-arachidonyl glyceryl ether (2-AGE), were synthesized and their ability to activate the CB1 receptors was determined
by the [35S]GTPcS binding assay using rat cerebellar membranes. The main goal of the study was to examine how the DMH end tail
affects the activity properties of 2-AG (1) and its stable ether (2) and urea analogues (5). The importance of the chain length was also
explored by synthesizing 2-AG and 2-AGE derivatives (3 and 4) possessing the chain length C21 instead of C22. Replacement of the
pentyl end chain with the DMH resulted in distinct potency decrease as compared to the reference compounds. The modification did
not have such a strong impact on the efficacy values. In fact, the efficacy of the derivatives of 2-AGE (2 and 4) was comparable or
even improved. Introducing a more stable and hydrophilic urea bond led to a dramatic decrease in biological activity.
ꢀ 2005 Elsevier Ltd. All rights reserved.
1. Introduction
recognized earlier. After modifying experimental
conditions with serine esterase inhibitors, first with
diisopropyl fluorophosphate (DFP) and later with the
more potent phenylmethylsulfonyl fluoride (PMSF)
and methyl arachidonoyl fluorophosphonate (MAFP),3
more reliable and promising results of 2-AG and its role
in the endocannabinoid system have been gained.
Although the first CB1 receptor binding studies gave a
reason to postulate that the role of 2-AG as an endocan-
nabinoid is not very significant, it has been later shown
that 2-AG is a potent, full efficacy agonist and most
probably the main endogenous ligand for both the
CB1 and CB2 receptors.4–7
In 1995, two independent research groups published sig-
nificant findings about a new endogenous compound
which possessed clear cannabimimetic activity but had
an ester linkage and a glycerol head, distinguishing it
from the well-known endogenous cannabinoid, N-
arachidonoyl ethanol amide (AEA).1,2 It was notewor-
thy that the concentration of this novel cannabinoid
receptor ligand, 2-arachidonoyl glycerol (2-AG), in the
rat brain was almost 800 times higher than that of
AEA. However, at that time, this new endocannabinoid
did not receive the attention it would have deserved,
probably because of its poor receptor binding affinity
properties as compared to those of AEA (Ki (CB1) for
2-AG 2.4 lM, Ki (CB1) for AEA 99 nM).1 Possibly,
2-AG would have been found earlier if its susceptibility
for enzymatic degradation and acyl migration were
HU-310 (2-arachidonyl glyceryl ether, 2-AGE), the
ether analogue of 2-AG, was first reported in 1998 by
Mechoulam et al. as a stable chemical tool for various
cannabinoid related studies both in vitro and in vivo.8
At the same time, this metabolically stable compound
was synthesized and tested by Sugiura et al.5,9 Later
on, 2-AGE was postulated as an endocannabinoid,10
even though its status as the third endocannabinoid
has been under debate since its discovery because it
has taken a while for other research groups to repeat this
Keywords: Cannabinoid; Synthesis; CB1; 2-AG; 2-AGE; Dimethyl-
heptyl; [35S]GTPcS binding assay.
*
Corresponding author. Tel.: +358 17 162431; fax +358 17
162456; e-mail: Teija.Parkkari@uku.fi
0968-0896/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2005.12.007

T. Parkkari et al. / Bioorg. Med. Chem. 14 (2006) 2850–2858
2851
finding. However, Di Marzoꢀs and co-workers showed in
2003 that the compound possessing the same mass and
similar chemical properties (i.e., chromatographic prop-
erties) as 2-AGE can be found in the brain.11 Neverthe-
less, 2-AGEꢀs occurrence in the brain did not fully
overlap the regional distribution of the CB1 receptors.
In the same study, it was also postulated that 2-AGE
can be taken up into the cells by the same transporter
protein as the other endocannabinoids. Quite opposite
findings were reported by Oka et al.12 They were not
able to detect appreciable amounts of 2-AGE from the
distinct mammalian brain tissues using GC–MS and
fluorometric HPLC. Since the contradictory findings
and unclear biosynthetic route of 2-AGE, perhaps it
is too early to consider 2-AGE as the third
endocannabinoid.
DMH derivative of 2-AG (1) was synthesized in order
to study whether the potency and efficacy can be in-
creased. In addition, we synthesized a DMH analogue
of 2-AGE (2), aimed at an enzymatically more stable,
and potent, high efficacy cannabinoid for in vitro stud-
ies. Finally, we studied whether the shortening of the
chain length has any effect on the activity properties (3
and 4), and if the activity properties of 2-AG can be
retained by substituting the unstable ester linkage with
a more stable and more polar urea linkage (5).
2. Results and discussion
2.1. Selection of compounds for activity tests
The structures of 59 commercially available endocanna-
binnoid-type compounds were examined. To select a
representative group of molecules, factor analysis was
performed on the calculated steric and electrostatic
fields of the molecules. The first three components were
graphed and 11 molecules from the different areas of the
field-property space were chosen to be tested for their
CB1 receptor activity (Fig. 2, see Supporting informa-
tion). The molecular structures and the CB1 receptor-
mediated G-protein activation data are presented in
Table 2 (see Supporting information). During this study,
the [35S]GTPcS binding assay protocol was optimized,
and the activity data of these compounds have been
produced with a slightly different method than those
of the synthesized compounds.
2-AG is known as the most efficacious CB1 receptor li-
gand so far. In the [35S]GTPcS binding assays with the
rat cerebellar membranes, Emax value for 2-AG was
reported to be 620 5 (% basal SEM), while the
respective figures for AEA, 2-AGE and CP55,940 were
484 7, 415 3 and 510 4.3 The classical cannabinoid
HU-210 has the potency value higher than any other
cannabinoid found so far. In our [35S]GTPcS binding as-
says, pEC50 for HU-210 was found to be as high as
8.3 0.1 (unpublished data), whereas the corresponding
value for 2-AG is 6.0 0.0.3 The high potency of HU-
210 results most probably from the optimal positioning
of the hydroxyl groups in relation to the aliphatic dim-
ethylheptyl (DMH) side chain, which already in 1949
was shown to be favoured in the ligand–receptor recog-
nition of tetrahydrocannabinol-type compounds.13,14
Molecule 34 (Table 2, see Supporting information) was
found to be as effective and potent as 2-AG. Therefore,
four additional molecules were chosen from the struc-
tural space surrounding that compound and tested for
the CB1 receptor activity as well (Table 2, see Support-
ing information). Another molecule with a similar
DMH side chain to that of 34 was found to have a
comparable potency and efficacy with 34 and 2-AG.
Consequently, we focussed on the design and synthesis
of new endocannabinoid derivatives with DMH side
chain in order to evaluate structure–activity relation-
ships and to discover novel active CB1 lead structures.
The importance of the end pentyl chain on the activity
properties of AEA and its derivatives has also been
extensively examined by branching and changing the
length of the chain.15,16 Concurrent study of Ryan
et al. showed similar results of a DHM derivative of
AEA; the activity was enhanced both in vitro and
in vivo.16 In the same study, methyl groups attached
to the adjacent carbon produced only minor activity
differences. The corresponding monomethyl derivative
had also a very similar pharmacological profile. The
study also provided further information on an optimal
end chain length of AEA. The addition of one to four
methylenes into AEA backbone resulted in higher
affinity properties; however, such analogues did not
show significant in vivo pharmacological effects at
doses up to 30 mg/kg.
2.2. Chemistry
A synthesis route for the final products 1 and 3 is illus-
trated in Scheme 1. The key synthons 6 and 17 were pre-
pared as described in the literature.17 The synthesis of
endogenous 2-AG and its analogues is very challenging.
First, 2-AG is very easily isomerized due to the acid,
base and heat promoted migration of the acyl group,
and second, the double bond system in fatty acids is
extremely sensitive to auto-oxidation limiting the num-
ber of suitable synthesis methods. Stamatov and Stawin-
ski17,18 developed an efficient synthetic strategy to 2-AG,
which was utilized in the synthesis of final products 1
and 3. The corresponding carboxylic acids 6 and 17 were
esterified with ( )-glycidol in the presence of 1-(3-dim-
ethylaminopropyl)-3-ethylcarbodiimide hydrochloride
(EDC) and 4-dimethylaminopyridine (DMAP) to give
the glycidyl derivatives 7 and 18, which were treated
While the activity properties for the DMH derivatives of
AEA have been well studied, the effect of DMH modifi-
cations on the other endocannabinoids, 2-AG and 2-
AGE, has not yet been reported. In the present study,
we explored the CB1 activating properties of endocan-
nabinoid derivatives in the rat cerebellar membranes
using the [35S]GTPcS binding assay. As a starting point,
a representative group of molecules was tested for the
CB1 receptor activation. Based on the 2-AGꢀs activity
properties, and the knowledge of the SARs compounds
with DMH modification, we decided to explore more
closely the DMH derivatives of endocannabinoids. A

2852
T. Parkkari et al. / Bioorg. Med. Chem. 14 (2006) 2850–2858
O
O
O
O
a
b
OH
n
n
7
n = 1
6
n = 1
18 n = 0
17 n = 0
OH
OCOCF3
OCOCF3
O
O
O
OH
c
n
O
n
1 n = 1
3 n = 0
8
n = 1
19 n = 0
Scheme 1. Reagents: (a) EDC, DMAP, CH2Cl2; (b) TFAA, CH2Cl2; (c) pyridine, MeOH, CH2Cl2/hexane.
with TFAA giving trifluoroacetate esters 8 and 19.
These were further converted into the desired products
by transesterification using pyridine and methanol in
dichloromethane/hexane.
which enabled a selective protection of hydroxyl groups
with benzaldehyde.21 The benzyl carbamate 15 was
deprotected by catalytic hydrogenation at elevated pres-
sure and then treated with isocyanate 24.20 Finally, the
protecting group of the 1,3-diol was removed with
concentrated HCl in methanol.
Scheme 2 illustrates the synthesis method for the 2-AGE
analogues 2 and 4. The key synthons 9 and 20 were pre-
pared as described in the literature.17 In the first step,
corresponding ester (9 or 20) was reduced to alcohol
with lithium aluminium hydride (LAH), and alcohol
was activated by methane sulfonyl chloride (mesyl chlo-
ride, MsCl) prior to the addition of 2-phenyl-[1,3]-diox-
an-5-ol. Finally, the benzylidene protective group was
cleaved by concentrated HCl in methanol.19
2.3. [35S]GTPcS membrane binding studies
Ability of the synthesized compounds to activate the
CB1 receptors was determined by an optimized
[35S]GTPcS binding assay protocol where noise due to
tonic adenosine A1 receptor activity and enzymatic deg-
radation of endocannabinoids have been eliminated
(Method G in Experimental).3 The efficacy and potency
values are presented in Table 1, and dose–response
curves for compounds 1–4 are shown in Figure 1. All
the synthesized compounds showed dose-dependent
CB1 activity, and their responses at 5 · 10À5 M were
reversed by the CB1 receptor antagonist, AM251
(10À6 M, data not shown).
The urea 5 was synthesized by converting the acid 6 to a
corresponding azide with diphenylphosphorylazide
(DPPA), which was further rearranged by heating to
an isocyanate 2420 (Scheme 3). 2-Phenyl-[1,3]-dioxan-5-
yl-amine 16 was produced by protecting an amino group
of 2-amino-1,3-propanediol with benzyl chloroformate,
O
O
OH
b
a
n
n
9
n = 1
10 n = 1
21 n = 0
O
20 n = 0
O
O
S
O
n
O
c
O
n
O
O
HO
12 n = 1
23 n = 0
d
11 n = 1
22 n = 0
OH
OH
O
n
2 n = 1
4 n = 0
Scheme 2. Reagents: (a) LAH, Et2O; (b) MsCl, Et3N, CH2Cl2; (c) KOH, benzene; (d) HCl/MeOH.

T. Parkkari et al. / Bioorg. Med. Chem. 14 (2006) 2850–2858
2853
OH
OH
O
O
a
H
H2N
O
N
b
O
14
15
24
O
O
16
H2N
OH
N C O
d
c
O
6
OH
OH
O
O
O
N
N
N
H
N
e
O
H
H
H
13
5
Scheme 3. Reagents: (a) 1—CbzCl, EtOH, Et3N, 2—benzaldehyde, toluene, p-TSA; (b) Pd/C, H2, 300 psi, 60 ꢁC; (c) 1—DPPA, Et3N, benzene,
2—toluene, D; (d) toluene; (e) HCl/MeOH.
Surprisingly, the DMH derivatives of 2-AG (1 and 3)
acted as much weaker CB1 receptor agonists than
2-AG. Although good efficacy values were achieved,
potency values remained relatively poor. It was also
interesting to notice that the derivative having a shorter
chain length (3) was more efficient than compound 1
Table 1. Comparison of efficacy (Emax) and potency (pEC50) values of 2-AG, 2-AGE and 1–5
Compound
Structure
CB1 activation (n = 3)
Emax (% basal SEM)
pEC50 SEM
O
OH
OH
2-AG
620
484
5
7
6.0 0.0
O
OH
2-AGE
5.2 0.0
O
OH
OH
O
OH
1
521 35
4.6 0.1
O
OH
OH
2
3
4
5
475
2
4.6 0.0
4.3 0.1
4.5 0.1
4.0 0.1a
O
OH
OH
O
590 76
548 27
309 8a
O
OH
OH
O
O
OH
OH
N
H
N
H
The activation data have been produced using the method G described in Section 4.
a n = 2.

2854
T. Parkkari et al. / Bioorg. Med. Chem. 14 (2006) 2850–2858
addition of the urea bond (5) decreased the activation
650
550
450
350
250
150
50
parameters dramatically. The weak activity properties
of compound 5 are most likely a result of its hydrophilic
structure; in flash chromatography, the compound re-
quires clearly more polar solvents for elution, and the
state of the product is more like wax than oil. In addi-
tion, the weak activation properties may also be a con-
sequence of the strong intramolecular hydrogen bonds
2-AG
2-AGE
1
2
3
1
4
which can be observed in the H NMR spectrum.
3. Conclusions
In the present study, dimethylheptyl derivatives of the
endocannabinoids, 2-AG and 2-AGE, have been synthe-
sized and their ability to activate the CB1 receptor have
been determined by the [35S]GTPcS binding assay using
the rat cerebellar membranes. Modification of a pentyl
tail of 2-AG with dimethylheptyl led to a dramatic
potency decrease, while the impact on efficacy was much
weaker. Replacement of the pentyl chain of 2-AGE with
DMH resulted in similar loss of potency, whereas
the efficacy remained comparable to that of 2-AGE.
Shortening of the chain length from C22 to C21 did not
improve the potency values but, interestingly, led to
agonists with increased efficacy. Introducing a more
hydrophilic and stable urea bond produced only weak
agonistic activity. Based on these results, we conclude
that unlike AEA-type compounds and classical cannab-
inoids, the activity properties of 2-AG and 2-AGE
cannot be improved by replacement of the end pentyl
chain with the DMH structure.
-7
-6
-5
-4
log [ligand]
Figure 1. Dose–response curves for 2-AG, 2-AGE and 1–4 for the G-
protein binding in rat cerebellar membranes (mean SEM, n = 3–4).
having the structure closer to that of 2-AG. 2-AGE
derivatives 2 and 4 also acted as very low potency ago-
nists for the CB1 receptor, however, they were equally
or even more efficacious than 2-AGE. In this series, a
similar effect of the chain length on the CB1 receptor
activities was observed; shorter chain length gives higher
efficacy values. It seems that the ligand–receptor interac-
tions of ester derivatives of endocannabinoids are sub-
stantially more sensitive to structural changes than
amide derivatives. This is in good agreement with the
previous findings of Suhara et al. who synthesized met-
abolically stable analogues of 2-AG and determined
their CB1 receptor activities by measuring the ability
of the compounds to induce Ca2+ transients in
NG108-15 cells.22 Few of their compounds showed com-
parable activity to that of 2-AGE, however, their ago-
nistic activities were approximately 100 times lower
than that of 2-AG. 2-AG analogues with different degree
of double bonds have also been investigated and arachi-
donic backbone with four double bonds seemed to be
favoured, although the effect on the activity was not very
drastic.5
4. Materials and experimental procedures
4.1. Factor analysis
All computations were carried out on SGI O2 R5000 or
R12000 workstations. Ligand structures were construct-
ed using the molecular modelling package SYBYL 6.7.24
The applied force field was Tripos force field,25 and the
atomic partial charges for the ligands were calculated by
the Gasteiger–Huckel26 method. A crystallographic tem-
¨
plate chosen for building the ligands was taken from the
Protein Data Bank:27 arachidonic acid complexed with
adipocyte lipid–binding protein (PDB 1adl28). After a
light geometry optimization, all the constructed ligands
were aligned according to the template and the factor
analysis of the electrostatic and steric fields of the com-
pounds was performed as implemented in SYBYL 6.7.
Razdan and co-workers published results on carbamate
and urea analogues of endocannabinoids aimed at better
enzymatic stability.20 In this series, the carbamate deriv-
atives showed only weak activity properties, whereas
some urea derivatives were clearly more potent than
AEA both in vitro and in vivo. Important finding of this
study was that the enzymatic stability can indeed be in-
creased by substituting the amide bond with the urea
bond. Interestingly, a compound where the amide bond
of AEA was replaced for urea bond showed only weak
binding affinity (Ki = 347 33 nM) but in several phar-
macological in vivo studies (tail flick test, spontaneous
activity and immobility) it was more potent than AEA
or other urea derivatives synthesized. The authors were
not able to explain this observation but it resembles the
data published on 2-AG.1,4,5,23 In our DMH series, the
4.2. Chemistry
1H NMR and 13C NMR were recorded on a Bruker
Avance 500 spectrometer operating on 500.1 and
125.8 MHz, respectively. CDCl3 was used as a solvent,
and tetramethylsilane (TMS) was used as an internal
standard. The spectra were processed from the recorded
FID files with MestRe-C software (version 2.3a,
`
`
Departemento Quımica Organica, Universidade de San-
tiago de Compostela, Spain). Chemical shifts (d) are
reported in parts per million (ppm) downfield from

T. Parkkari et al. / Bioorg. Med. Chem. 14 (2006) 2850–2858
2855
TMS. Following abbreviations are used: s = singlet,
d = doublet, t = triplet, dd = doublet of doublets,
dq = doublet of quartets, br s = broad singlet,
qn = quintet and m = multiplet. Coupling constants
1.25–1.38 (m, 10H), 1.70 (qn, 2H, J = 7.5), 1.98–2.24
(m, 3H), 2.33–2.38 (m, 2H), 2.79–2.95 (m, 5H), 4.46
(dd, 2H, 2J = 11.9, J = 5.5), 4.63 (dd, 2H, 2J = 12.0,
J = 4.3), 5.13–5.27 (m, 2H), 5.32–5.44 (m, 7H).
3
are reported in hertz and the letter J indicates J if not
otherwise noted. ESI-MS spectra were acquired using
a LCQ ion trap mass spectrometer equipped with an
electrospray ionization source (Finnigan MAT, San
Jose, CA, USA). Gas chromatography mass spectrum
was obtained on a HP6890 GC mass spectrometer with
electron-ionisation detector. The free hydroxyl groups
of the sample (0.1 mg/ml) in methanol were coated with
silicon groups. Elemental analyses for C, H and N were
performed on a ThermoQuest CE Instruments EA1110-
CHNS-O elemental analysator (ThermoQuest, Italy.)
TLC was performed using silica gel (60 F254) coated alu-
minium sheets (Merck, Germany).
4.4.2. 1,3-Bis-(trifluoroacetyl)-2-(15,15-dimethyl-henico-
sa-all-cis-4,7,10,13-tetraenoic acid) propyl ester (19).
Light yellow oil; yield 81%. H NMR: d 0.88 (t, 3H,
J = 6.9), 1.09 (s, 6H), 1.23–1.34 (m, 10H), 2.33–2.45
(m, 4H), 2.81–2.95 (m, 6H), 4.46 (dd, 2H, 2J = 11.9,
J = 5.4), 4.63 (dd, 2H, 2J = 11.9, J = 4.2), 5.13–5.46
(m, 9H).
1
4.5. Method C. Cleavage of trifluoroacetyl groups
Corresponding trifluoroacetyl propyl ester (8 or 19)
(1 equiv) was dissolved in dry hexane/CH2Cl2 (2/1,
15 ml/mmol) solution and the mixture was cooled to
À20 ꢁC. Dry pyridine (12.5 equiv) and methanol
(18.5 equiv) in dry hexane/CH2Cl2 solution (2/1, 15 ml/
mmol) were added slowly into the reaction mixture.
The cooling bath was removed and the solution was stir-
red at room temperature for 2 h. The reaction was con-
cluded by evaporating solvents under reduced pressure
without a heating bath.
4.3. Method A. EDC coupling
Corresponding acid (6 or 17) (1 equiv), glycidol
(1.2 equiv), EDC (1.8 equiv) and DMAP (0.3 equiv) in
dry CH2Cl2 (15 ml/mmol) were stirred at room temper-
ature for 20 h. The solvent was evaporated under re-
duced pressure.
4.3.1. 16,16-Dimethyl-docosa-all-cis-5,8,11,14-tetraenoic
acid oxiranylmethyl ester (7). The crude product was
purified by flash chromatography eluting with petro-
4.5.1. 1,3-Bis-(hydroxy)-2-(16,16-dimethyl-docosa-all-cis-
5,8,11,14-tetraenoic acid) propyl ester (1). H NMR: d
1
0.88 (t, 3H, J = 6.8), 1.09 (s, 6H), 1.23–1.36 (m, 10H),
2.04–2.14 (br s, 2H), 2.36–2.53 (m, 4H), 2.82–2.96 (m,
6H), 3.82 (d, 4H, J = 4.7), 4.93 (qn, 1H, J = 4.7), 5.13–
5.46 (m, 8H). 13C NMR: d 14.1, 17.6, 22.7, 24.8 (2C),
25.6, 25.7, 26.5, 26.8, 29.0, 30.2, 31.9, 33.7, 44.3 (2C),
62.6 (2C), 75.2, 127.0, 127.8, 128.1, 128.3, 128.8, 129.1
(2C), 139.22, 173.3. Elemental analysis. Calculated for
C27H46O4Æ1/4H2O: C, 73.84; H, 10.62. Found: C,
73.50; H, 10.80.
1
leum ether/ethyl acetate (8/1). H NMR: d 0.88 (t, 3H,
J = 7.0), 1.09 (s, 6H), 1.26–1.38 (m, 10H), 1.73 (qn,
2H, J = 7.5), 2.00–2.17 (m, 3H), 2.34–2.39 (m, 2H),
2.65 (dd, 1H, 2J = 4.8, J = 2.6), 2.79–2.86 (m, 4H),
2.88–2.95 (m, 2H), 3.19–3.22 (m, 1H), 3.92 (dd, 1H,
2J = 12.3, J = 6.3), 4.41 (dd, 1H, 2J = 12.3, J = 3.1),
5.13–5.27 (m, 2H), 5.31–5.43 (6H).
4.3.2. 15,15-Dimethyl-henicosa-all-cis-4,7,10,13-tetrae-
noic acid oxiranylmethyl ester (18). The crude product
was purified by flash chromatography eluting with
petroleum ether/ethyl acetate (10/1) yielding a colour-
less oil (52%). 1H NMR: d 0.88 (t, 3H, J = 7.0),
1.09 (s, 6H), 1.26–1.36 (m, 10H), 2.36–2.42 (m,
4H), 2.64 (dd, 1H, 2J = 4.9, J = 2.6), 2.82–2.95 (m,
7H), 3.19–3.22 (m, 1H), 3.92 (dd, 1H, 2J = 12.3,
J = 6.1), 4.41 (dd, 1H, 2J = 12.3, J = 3.1), 5.14–5.45
(m, 8H).
4.5.2. 1,3-Bis-(hydroxy)-2-(15,15-dimethyl-henicosa-all-
cis-4,7,10,13-tetraenoic acid) propyl ester (3). Yellowish
oil; yield 95%. H NMR: d 0.88 (t, 3H, J = 6.8), 1.09
1
(s, 6H), 1.23–1.36 (m, 10H), 2.04–2.14 (br s, 2H),
2.36–2.53 (m, 4H), 2.82–2.96 (m, 6H), 3.82 (d, 4H,
J = 4.7), 4.93 (qn, 1H, J = 4.7), 5.13–5.46 (m, 8H). 13C
NMR: d 14.1, 22.7, 22.8, 24.8, 25.6, 25.7, 26.7, 29.0
(2C), 30.2, 31.9, 34.2, 36.4, 44.3, 62.5 (2C), 75.3, 127.0,
127.7, 127.8, 127.9, 128.5, 129.1, 129.7, 139.2, 173.3.
GC–MS (EI): m/z = 565. Elemental analysis. Calculated
for C26H44O4Æ1/7H2O: C, 73.79; H, 10.55. Found: C,
73.77; H, 10.82.
4.4. Method B. TFAA promoted stereoselective transfor-
mation of glycidyl ester
Corresponding oxiranylmethyl ester (7 or 18) (1 equiv)
in dry CH2Cl2 (5 ml/mmol) was cooled to À20 ꢁC and
TFAA (6 equiv) in dry CH2Cl2 (5 ml/mmol) was added.
The cooling bath was removed, and the mixture was al-
lowed to warm to room temperature. After 2 h, the reac-
tion was concluded by evaporating the solvent under
vacuum. The residue was purified by flash chromatogra-
phy eluting with toluene.
4.6. Method D. LiAlH4 reduction and alcohol activation
with methanesulfonyl chloride
LiAlH4 (1 equiv) was added several portions into dry
THF (1.7 ml/mmol), and the solution was cooled to
0 ꢁC. Corresponding ester (9 or 20) (1 equiv) in dry
diethyl ether (1.7 ml/mmol) was added slowly into the
reaction mixture, and stirring was continued overnight
at room temperature. The solution was cooled to 0 ꢁC
and quenched sequentially with water, 15% NaOH solu-
tion and water. The cooling bath was removed, and the
mixture was allowed to precipitate at room temperature
4.4.1. 1,3-Bis-(trifluoroacetyl)-2-(16,16-dimethyl-docosa-
all-cis-5,8,11,14-tetraenoic acid) propyl ester (8). Yield
1
83%. H NMR: d 0.88 (t, 3H, J = 6.9), 1.09 (s, 6H),

2856
T. Parkkari et al. / Bioorg. Med. Chem. 14 (2006) 2850–2858
for half an hour. The solid was filtered and washed with
ether. The aqueous layer was separated from the filtrate
and extracted several times with ether. The combined
organic layers were dried over MgSO4 and the solvent
was evaporated under reduced pressure to obtain a yel-
lowish, oily alcohol. Alcohol (10 or 21) (1 equiv) and tri-
ethylamine (2.7 equiv) were dissolved in dry CH2Cl2
(10 ml/mmol). The reaction mixture was cooled to 0 ꢁC
and methanesulfonyl chloride (1.3 equiv) was added to
the mixture. The cooling bath was removed and the mix-
ture was stirred for 2 h at room temperature. The reac-
tion mixture was quenched with cold water and the
separated aqueous layer was extracted with ether. The
combined organic layers were acidified with 0.5 M
H2SO4, washed with saturated NaHCO3 and finally
dried over MgSO4. The solvent was evaporated under
vacuum. The product was used in the next reaction with-
out further purification.
4.7.2. 5-(15,15-Dimethyl-henicosa-all-cis-4,7,10,13-tet-
raenyloxy)-2-phenyl-[1,3]-dioxane (23). The crude prod-
uct was purified by flash chromatography eluting with
10% diethyl ether in petroleum ether yielding a yellow
oil (50%). H NMR: d 0.88 (t, 3H, J = 6.8), 1.10 (s,
6H), 1.24–1.32 (m, 10H), 1.74 (qn, 2H, J = 7.0), 2.13–
1
2.21 (m, 2H), 2.79–2.94 (m, 6H), 3.24–3.27 (m, 1H),
3.56 (t, 2H, J = 6.6) 4.04 (dd, 2H, J = 12.4, J = 1.8),
2
4.33 (dd, 2H, 2J = 12.4, J = 1.5), 5.13–5.45 (m, 8H),
5.55 (s, 1H), 7.30–7.36 (m, 3H), 7.47–7.52 (m, 2H).
4.8. Method F. Deprotection of a diol
Corresponding 2-phenyl-[1,3]-dioxane derivative (12 or
23) was dissolved in a solution of concentrated HCl/
MeOH (1/2) and stirred at room temperature for 3 h.
MeOH was evaporated under reduced pressure, and
the residue was diluted with ethyl acetate. The separated
organic phase was washed with 5% NaHCO3 followed
by brine and finally dried over MgSO4, and the solvent
was evaporated under reduced pressure.
4.6.1. Methanesulfonic acid 16,16-dimethyl-docosa-all-
1
cis-5,8,11,14-tetraenyl ester (11). H NMR: d 0.88 (t, 3H,
J = 6.9), 1.10 (s, 6H), 1.22–1.43 (m, 10H), 1.45–1.57 (m,
2H), 1.72–1.80 (m, 2H), 2.02–2.24 (m, 3H), 2.80–2.95
(m, 5H), 3.00 (s, 3H), 4.23 (t, 2H, J = 6.5), 5.13–5.43 (m,
8H).
4.8.1. 2-(16,16-Dimethyl-docosa-all-cis-5,8,11,14-tetrae-
nyloxy)-propane-1,3-diol (2). The crude product was
purified by flash chromatography eluting with 30–50%
ethyl acetate in petroleum ether yielding a yellowish oil
1
4.6.2. Methanesulfonic acid 15,15-dimethyl-henicosa-all-
cis-4,7,10,13-tetraenyl ester (22). 1H NMR: d 0.88 (t, 3H,
J = 6.9), 1.10 (s, 6H), 1.24–1.30 (m, 10H), 1.83 (qn, 2H,
J = 6.9), 2.14–2.23 (m, 2H), 2.80–2.95 (m, 6H), 3.00 (s,
3H), 4.23 (t, 2H, J = 6.5), 5.13–5.47 (m, 8H).
(68%). H NMR: d 0.88 (t, 3H, J = 6.8), 1.10 (s, 6H),
1.26–1.39 (m, 10H), 1.44 (qn, 2H, J = 7.6), 1.59–1.66
(m, 2H), 1.93 (br s, 2H), 2.02–2.21 (m, 3H), 2.74–2.95
(m, 5H), 3.46 (qn, 1H, J = 4.8), 3.58 (t, 2H, J = 6.6)
3.68 (dd, 2H, 2J = 11.5, J = 5.1), 3.76 (dd, 2H,
2J = 11.6, J = 4.4), 5.13–5.42 (m, 8H). 13C NMR: d
14.1, 22.7, 24.8, 25.7 (2C), 26.2, 26.7, 27.0 (2C), 29.0,
29.7, 30.2, 31.9, 36.4, 44.3 (2C), 62.3, 70.0, 79.6, 127.0,
127.9, 128.1, 128.2, 128.3, 129.0, 129.8, 139.2. Elemental
analysis. Calculated for C27H48O3Æ1/10H2O: C, 76.76; H,
11.50. Found: C, 76.62; H, 11.80.
4.7. Method E. Addition of 5-hydroxy-2-phenyl-[1,3]-
dioxane
KOH (12 equiv) and 5-hydroxy-2-phenyl-1,3-dioxane
(12 equiv) were crushed together and dissolved in dry
benzene (50 ml/mmol). The mixture was stirred at
50 ꢁC for half an hour. Corresponding alcohol activated
with methanesulfonyl group (11 or 22) (1 equiv) in dry
benzene (50 ml/mmol) was added slowly to the reaction
mixture and the solution was stirred at 50 ꢁC for three
days. The reaction was concluded by adding crushed
KOH (6 equiv) and 5-hydroxy-2-phenyl-1,3-dioxane
(6 equiv) to the reaction mixture and stirring it overnight
at 50 ꢁC. After the heating bath was removed, the mix-
ture was diluted with ether and acidified with 10%
HCl to pH 1. The separated aqueous layer was extracted
several times with ether, and the combined organic lay-
ers were washed with brine and dried over MgSO4. The
solvent was evaporated under vacuum.
4.8.2. 2-(15,15-Dimethyl-henicosa-all-cis-4,7,10,13-tet-
raenyloxy)-propane-1,3-diol (4). The crude product was
purified by flash chromatography eluting with ethyl ace-
tate/petroleum ether (1/1) solution yielding a brownish
1
oil (37%). H NMR: d 0.88 (t, 3H, J = 6.9), 1.10 (s,
6H), 1.24–1.32 (m, 10H), 1.69 (qn, 2H, J = 6.8), 2.01–
2.30 (m, 4H), 2.78–2.96 (m, 6H), 3.46 (qn, 1H,
2
J = 4.7), 3.59 (t, 2H, J = 6.6) 3.69 (dd, 2H, J = 11.6,
J = 5.0), 3.77 (dd, 2H, 2J = 11.6, J = 4.4), 5.12–5.43
(m, 8H). 13C NMR: d 14.1, 22.7, 23.8, 24.8, 25.6, 25.7,
26.8, 29.0 (2C), 29.3, 29.9, 30.2, 31.9, 44.3, 62.3 (2C),
69.5, 79.6, 127.0, 127.8, 128.2, 128.2, 128.6, 129.1,
129.3, 139.2. Elemental analysis. Calculated for
C26H46O3Æ1/4H2O: C, 75.95; H, 11.40. Found: C,
75.87; H, 11.36.
4.7.1. 5-(16,16-Dimethyl-docosa-all-cis-5,8,11,14-tetrae-
nyloxy)-2-phenyl-[1,3]-dioxane (12). The crude product
was purified by flash chromatography eluting with 10–
20% diethyl ether in petroleum ether yielding a colour-
4.8.3. (2-Phenyl-[1,3]-dioxan-5-yl) carbamic acid benzyl
ester (15). 2-Amino-1,3-propanediol 14 (1.0 g,
11.0 mmol) and Et3N (2.3 ml, 16.6 mmol) were dissolved
in dry EtOH (40 ml). The reaction mixture was cooled to
0 ꢁC and benzyl chloroformate (4.7 ml, 32.9 mmol) was
added slowly to the mixture. The cooling bath was re-
moved and the mixture was stirred for 1 h at room tem-
perature. The reaction was concluded by evaporating
the solvent under vacuum, and the residue was purified
1
less oil (60%). H NMR: d 0.88 (t, 3H, J = 6.7), 1.09
(s, 6H), 1.26–1.40 (m, 10H), 1.43–1.52 (m, 2H), 1.64–
1.71 (m, 2H), 1.97–2.17 (m, 3H), 2.80–2.98 (m, 5H),
3.25–3.26 (m, 1H), 3.56 (t, 2H, J = 6.6) 4.04 (dd, 2H,
2J = 12.6, J = 1.7), 4.33 (dd, 2H, 2J = 12.5, J = 1.4),
5.13–5.43 (m, 8H), 5.55 (s, 1H), 7.30–7.36 (m, 3H),
7.49–7.52 (m, 2H).

T. Parkkari et al. / Bioorg. Med. Chem. 14 (2006) 2850–2858
2857
by flash chromatography eluting with 2% MeOH in
CH2Cl2 to yield N-benzyloxycarbonyl-2-amino-1,3-pro-
panediol as a white solid compound (2.0 g, 81%).
1H NMR: d 3.36–3.46 (m, 5H), 4.53 (t, 2H, J = 5.5),
5.01 (s, 2H), 6.79 (d, 1H, J = 7.8), 7.28–7.36 (m, 5H).
N-Benzyloxycarbonyl-2-amino-1,3-propanediol (2.0 g,
8.9 mmol), benzaldehyde (1.2 ml, 11.9 mmol) and
p-TSA (20.0 mg, 0.1 mmol) in dry toluene (40 ml) were
refluxed under a Dean–Stark water separator for 2 h.
After enough water (0.16 ml) was separated from the
reaction mixture, the mixture was allowed to cool to
room temperature and the solvent was evaporated under
reduced pressure. The product was purified by flash
chromatography eluting with ethyl acetate/petroleum
ether (1/5) solution to yield a mixture of the cis- and
4.8.6. 1-(14,14-Dimethyl-icosa-all-cis-3,6,9,12-tetraenyl)-
3-(2-hydroxy-1-hydroxymethyl-ethyl)-urea (5). 1-(14,14-
Dimethyl-eicosa-all-cis-4,7,10,13-tetraenyl)-3-(2-phenyl-
[1,3]-dioxan-5-yl)-urea 13 (0.19 g, 0.4 mmol) was dis-
solved in a solution of concentrated HCl/MeOH (11/
25 ml) and stirred at room temperature for 4 h. MeOH
was evaporated under reduced pressure and the residue
was diluted with ethyl acetate. The separated organic
phase was washed with brine and dried over MgSO4.
The solvent was evaporated and the product was purified
by flash chromatography eluting with 1% MeOH in
CH2Cl2 to yield a whitish waxy product (38.0 mg, 46%).
1H NMR: d 0.88 (t, 3H, J = 6.9 Hz), 1.09 (s, 6H),
1.22–1.35 (m, 10H), 2.05–2.29 (m, 4H), 2.75–2.94
(m, 6H), 3.18 (qn, 1H, J = 6.9 Hz), 3.65–3.78 (m, 5H),
4.04 (br s, 1H), 5.12–5.55 (m, 10H). 13C NMR: d 14.1,
22.7, 24.8, 25.7, 26.8 (2C), 27.9, 29.0 (2C), 30.2, 31.9,
36.4, 40.2, 44.4, 53.4, 63.4 (2C), 126.5, 126.9, 127.7,
128.0, 128.5, 129.2, 130.5, 139.3, 159.3. ESI-MS: m/
z = 435.3 [M+H]+. Elemental analysis. Calculated for
C26H46N2O3Æ1/3H2O: C, 70.87; H, 10.67; N, 6.36. Found:
C, 70.91; H, 10.52; N, 6.00.
1
trans-isomers as a white solid product (2.0 g, 72%). H
NMR: (trans) d 3.56–3.60 (m, 2H ax), 4.07 (br s, 1H),
4.34 (dd, 2H eq, 2J = 10.9, J = 4.8), 4.60 (br s, 1H),
5.13 (s, 2H), 5.43 (s, 1H), 7.32–7.47 (m, 10H). 1H
NMR: (cis) d 3.74 (d, 1H, J = 8.9 Hz), 4.13–4.18 (m,
4H), 5.13 (s, 2H), 5.54 (s, 1H), 5.81 (d, 1H, J = 9.5),
7.30–7.47 (m, 10H).
4.8.4. 2-Phenyl-[1,3]-dioxan-5-yl-amine (16). N-Benzyl-
oxycarbonyl-(2-phenyl-[1,3]-dioxan-5-yl)-amine 15 (2.0 g,
6.4 mmol) and 10% Pd/C (0.1 g) were dissolved in dry
EtOH (150 ml). The reaction mixture was placed in a
high-pressure hydrogenation reactor (300 psi, 60 ꢁC).
After 3 h, the mixture was filtered through a pad of Cel-
ite and the solvent was evaporated under vacuum. The
product was purified by flash chromatography eluting
with 5% MeOH in CH2Cl2 to obtain a white solid amine
containing both isomers (0.95 g, 92%). 1H NMR: (trans)
d 1.01 (br s, 2H), 3.21 (m, 1H, J = 5.1), 3.43–3.48 (m, 2H
4.9. CB1 receptor activity based on [35S]GTPcS mem-
brane binding studies
4.9.1. Method G. The [35S]GTPcS membrane binding
studies were preformed as previously described.3 Maxi-
mal agonist responses (Emax, % basal) and potencies
(pEC50) were determined from dose–response curves.
The CB1-dependent activity was confirmed by antagoniz-
ing half-maximal responses with the CB1 selective antag-
onist AM251 (10À6 M). Results are presented as
means SEM of at least three independent experiments,
performed in duplicate. Data-analysis was calculated as
non-linear regression by GraphPad Prism 4.0 for
Windows.
2
ax), 4.28 (dd, 2H, eq, J = 11.3, J = 4.8), 5.39 (s, 1H),
1
7.32–7.49 (m, 5H). H NMR: (cis) d 1.79 (s, 2H), 2.81
(qn, 1H, J = 1.8), 4.05 (dq, 2H ax, J = 11.7, J = 1.5),
2
2
4.17 (dq, 2H, eq, J = 11.7, J = 1.6), 5.52 (s, 1H), 7.33–
7.51 (m, 5H).
4.9.2. Method H. The [35S]GTPcS membrane binding
studies were preformed as previously described.4 For ago-
nist dose–response and antagonist experiments, results
are presented as mean SEM of at least three indepen-
dent experiments, performed in duplicate. The CB1
dependent activity was confirmed by antagonizing half-
maximal responses with the CB1 selective antagonist
AM251 (10À6 M). Data-analysis was calculated as non-
linear regression by GraphPad Prism 3.0 for Windows.
4.8.5. 1-(14,14-Dimethyl-icosa-all-cis-3,6,9,12-tetraenyl)-
3-(2-phenyl-[1,3]-dioxan-5-yl)-urea (13). 15,15-Dimethyl-
heneicosa-all-cis-4,7,10,13-tetraenoic acid
6
(0.2 g,
0.6 mmol) and Et3N (0.12 ml, 0.87 mmol) were stirred
in dry benzene (3 ml) at room temperature for 10 min.
DPPA (0.24 g, 0.9 mmol) was added slowly and the mix-
ture was stirred for 2 h. The solvent was evaporated un-
der vacuum and the product was purified by flash
chromatography eluting with 5% ethyl acetate in petro-
leum ether to yield a yellowish oil (0.11 g, 51%). 15,15-
Acknowledgments
Dimethyl-heneicosa-all-cis-4,7,10,13-tetraenyl
azide
(0.11 g, 0.3 mmol) was dissolved in dry toluene (2 ml)
and stirred at 65 ꢁC for 2 h. 2-Phenyl-[1,3]-dioxan-5-yl-
amine 16 (60.0 mg, 0.3 mmol, cis-isomer) in dry toluene
(2 ml) was added to the reaction mixture and the stirring
was continued at 65 ꢁC for another 2 h. The reaction
was concluded by evaporating the solvent under reduced
pressure to obtain an impure whitish waxy urea (0.19 g).
1H NMR: d 0.88 (t, 3H, J = 6.8), 1.09 (s, 6H), 1.25–1.38
(m, 10H), 2.10–2.40 (m, 4H), 2.74–2.95 (m, 6H), 3.23
(qn, 1H, J = 6.1), 3.89 (d, 1H, J = 9.1), 4.03 (d, 1H,
J = 11.5), 4.11–4.18 (m, 4H), 5.12–5.51 (m, 8H), 5.53
(s, 1H), 7.34–7.51 (m, 5H).
The authors are grateful to Mrs. Miia Reponen, Mrs.
Tiina Koivunen, Mrs. Minna Glad and Mrs. Helly Rissa-
nen for their skilful technical help. The study was support-
ed by a grant from the National Technology Agency of
Finland and Academy of Finland (Grant No. 107300).
Supplementary material
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmc.2005.12.007.

2858
T. Parkkari et al. / Bioorg. Med. Chem. 14 (2006) 2850–2858
13. Huffman, J. W.; Yu, S.; Showalter, V.; Abood, M. E.;
References and notes
Wiley, J. L.; Compton, D. R.; Martin, B. R.; Bramblett,
R. D.; Reggio, P. H. J. Med. Chem. 1996, 39, 3875–3877.
14. Adams, R.; Harfenist, M.; Loewe, S. J. Am. Chem. Soc.
1949, 71, 1624–1628.
1. Sugiura, T.; Kondo, S.; Sukagawa, A.; Nakane, S.;
Shinoda, A.; Itoh, K.; Yamashita, A.; Waku, K. Biochem.
Biophys. Res. Commun. 1995, 215, 89–97.
15. Selzman, H. H.; Fleming, K. N.; Thomas, B. F.;
Gilliam, A. F.; McCallion, D. S.; Pertwee, R. G.;
Compton, D. R.; Martin, B. R. J. Med. Chem. 1997, 40,
3623–3634.
16. Ryan, W. J.; Banner, W. K.; Wiley, J. L.; Martin, B. R.;
Razdan, R. K. J. Med. Chem. 1997, 40, 3617–3625.
17. Dasse, O.; Mahadevan, A.; Han, L.; Martin, B. R.; Di
Marzo, V.; Razdan, R. K. Tetrahedron 2000, 56, 9195–9202.
18. Stamatov, S. D.; Stawinski, J. Tetrahedron Lett. 2002, 43,
1759–1761.
19. Mechoulam, R.; Fride, E.; Ben Shabat, S.; Sheskin, T.;
Breuer, A. Synthetic endogenous cannabinoids analogues
and uses threof. PCT Patent WO 99/51560, 1999.
20. Ng, E. W.; Aung, M. M.; Abood, M. E.; Martin, B. R.;
Razdan, R. K. J. Med. Chem. 1999, 42, 1975–1981.
21. Wharton, T.; Wilson, L. J. Tetrahedron Lett. 2002, 43,
561–564.
22. Suhara, Y.; Nakane, S.; Arai, S.; Takayama, H.; Waku,
K.; Ishima, Y.; Sugiura, T. Bioorg. Med. Chem. Lett. 2001,
11, 1985–1988.
2. Mechoulam, R.; Ben-Shabat, S.; Hanus, L.; Ligumsky,
M.; Kaminski, N. E.; Schatz, A. R.; Gopher, A.; Almog,
S.; Martin, B. R.; Compton, D. R.; Pertwee, R. G.;
Griffin, G.; Bayewitch, M.; Barg, J.; Vogel, Z. Biochem.
Pharmacol. 1995, 50, 83–90.
3. Savinainen, J. R.; Saario, S. M.; Niemi, R.; Jarvinen, T.;
Laitinen, J. T. Br. J. Pharmacol. 2003, 140, 1451–1459.
4. Savinainen, J. R.; Jarvinen, T.; Laine, K.; Laitinen, J. T.
Br. J. Pharmcol. 2001, 134, 664–672.
5. Sugiura, T.; Kodaka, T.; Nakane, S.; Miyashita, T.; Kondo,
S.; Suhara, Y.; Takayama, H.; Waku, K.; Seki, C.; Baba, N.;
Ishima, Y. J. Biol. Chem. 1999, 274, 2794–2801.
6. Sugiura, T.; Kondo, S.; Kishimoto, S.; Miyashita, T.;
Nakane, S.; Kodaka, T.; Suhara, Y.; Takayama, H.;
Waku, K. J. Biol. Chem. 2000, 275, 605–612.
7. Gonsiorek, W.; Lunn, C.; Fan, X.; Narula, S.; Lundell,
D.; Hipkin, R. W. Mol. Pharmacol. 2000, 57, 1045–1050.
8. Mechoulam, R.; Fride, E.; Ben-Shabat, S.; Meiri, U.;
Horowitz, M. Eur. J. Pharmacol. 1998, 362, R1–R3.
9. Suhara, Y.; Takayama, H.; Nakane, S.; Miyashita, T.;
Waku, K.; Sugiura, T. Chem. Pharm. Bull. (Tokyo) 2000,
48, 903–907.
23. Hillard, C. J. Prostaglandins Other Lipid. Mediat. 2000,
61, 3–18.
24. Sybyl v. 6.7; Tripos Associates, Inc.: St. Louis, MO, USA.
25. Clark, M.; Cramer, R. D., III; Van Opdenbosch, N. J.
Comput. Chem. 1989, 10, 982–1012.
10. Hanus, L.; Abu-Lafi, S.; Fride, E.; Breuer, A.; Vogel, Z.;
Shalev, D. E.; Kustanovich, I.; Mechoulam, R. Proc. Natl.
Acad. Sci. U.S.A. 2001, 98, 3662–3665.
26. Gasteiger, J.; Marsili, M. Tetrahedron 1980, 36, 3219–
3222.
27. Berman, H. M.; Westbrook, J.; Feng, Z.; Gilliland, G.;
Bhat, T. N. Nucleic Acids Res. 2000, 235–242.
28. LaLonde, J. M.; Levenson, M. A.; Roe, J. J.; Bernlohr, D.
A.; Banaszak, L. J. J. Biol. Chem. 1994, 269, 25339–25347.
11. Fezza, F.; Bisogno, T.; Minassi, A.; Appendino, G.;
Mechoulam, R.; Di Marzo, V. FEBS Lett. 2002, 513, 294–
298.
12. Oka, S.; Tsuchie, A.; Tokumura, A.; Muramatsu, M.;
Suhara, Y.; Takayama, H.; Waku, K.; Sugiura, T.
J. Neurochem. 2003, 85, 1374–1381.