Potent anandamide analogs: The effect of changing the length and branching of the end pentyl chain

Article abstract of DOI:10.1021/jm970212f

To examine the effect of changing the length and branching of the end pentyl chain (C₅H₁₁) of anandamide (AN), various analogs la-h and 2a-f were synthesized from either the known aldehyde ester 6a or from the alcohol 6b and tested for their pharmacological activity. A reproducible procedure was developed for the conversion of arachidonic acid to 6a or 6b in gram quantities (overall yield 15%). The appropriate tetraene esters 7 were prepared by carrying out a Wittig reaction, between 6a and the ylide generated from the phosphonium salt of the appropriate alkyl halide or between the ylide of 6d (prepared from 6a → 6b → 6c → 6d) and the appropriate alkyl aldehydes. They were then hydrolyzed to the corresponding acids and transformed into AN analogs 1 via their acid chlorides then treated with excess ethanolamine. α-Alkylation of esters 7 gave compounds 8 which were hydrolyzed to the corresponding acids. These acids via their acid chlorides and subsequent treatment with excess fluoroethylamine gave the target compounds 2. In this way analogs 1e and 2a-c were synthesized from 6d while all the remaining analogs were prepared from 6a. In order to assess the optimal length of the alkyl terminus, analogs la-d were prepared and showed moderately high affinities(18-55 nM). However analogs la-c failed to produce significant pharmacological effects at doses up to 30 mg/kg. Analog 1d was found to be a weak partial agonist. The reason for the lack of activity in la-c is presently not clear. Like the THCs, the branching of the end pentyl chain in AN (1e-h) increased potency both in in vitro and in vive activities; the dimethylheptyl (DMH) analog 1e was the most potent in the series. Similar alkyl substitutions were carried out in the fluoro-2-methylanandamide series (2a-f), and all of these analogs had high receptor affinities (1-14 nM), the DMH analog 2a being the most potent. With a few exceptions they showed robust pharmacological effects, and AN-like profiles. It was shown that the SAR of the end pentyl chain in AN is very similar to that of THCs. However, the magnitude of enhanced potency observed when the side chain of THC was changed from straight to branched was not observed when the end chain of AN was similarly changed.

Full text of DOI:10.1021/jm970212f

J . Med. Chem. 1997, 40, 3617-3625
3617
P oten t An a n d a m id e An a logs: Th e Effect of Ch a n gin g th e Len gth a n d
Br a n ch in g of th e En d P en tyl Ch a in
†
†
William J . Ryan, W. Kenneth Banner, J enny L. Wiley, Billy R. Martin, and Raj K. Razdan*
Organix, Inc., 65 Cummings Park, Woburn, Massachusetts 01801
X
Received March 31, 1997
5
To examine the effect of changing the length and branching of the end pentyl chain (C H11) of
anandamide (AN), various analogs 1a -h and 2a -f were synthesized from either the known
aldehyde ester 6a or from the alcohol 6b and tested for their pharmacological activity. A
reproducible procedure was developed for the conversion of arachidonic acid to 6a or 6b in
gram quantities (overall yield 15%). The appropriate tetraene esters 7 were prepared by
carrying out a Wittig reaction, between 6a and the ylide generated from the phosphonium salt
of the appropriate alkyl halide or between the ylide of 6d (prepared from 6a f 6b f 6c f 6d )
and the appropriate alkyl aldehydes. They were then hydrolyzed to the corresponding acids
and transformed into AN analogs 1 via their acid chlorides then treated with excess
ethanolamine. R-Alkylation of esters 7 gave compounds 8 which were hydrolyzed to the
corresponding acids. These acids via their acid chlorides and subsequent treatment with excess
fluoroethylamine gave the target compounds 2. In this way analogs 1e and 2a -c were
synthesized from 6d while all the remaining analogs were prepared from 6a . In order to assess
the optimal length of the alkyl terminus, analogs 1a -d were prepared and showed moderately
high affinities (18-55 nM). However analogs 1a -c failed to produce significant pharmacological
effects at doses up to 30 mg/kg. Analog 1d was found to be a weak partial agonist. The reason
for the lack of activity in 1a -c is presently not clear. Like the THCs, the branching of the end
pentyl chain in AN (1e-h ) increased potency both in in vitro and in vivo activities; the
dimethylheptyl (DMH) analog 1e was the most potent in the series. Similar alkyl substitutions
were carried out in the fluoro-2-methylanandamide series (2a -f), and all of these analogs had
high receptor affinities (1-14 nM), the DMH analog 2a being the most potent. With a few
exceptions they showed robust pharmacological effects, and AN-like profiles. It was shown
that the SAR of the end pentyl chain in AN is very similar to that of THCs. However, the
magnitude of enhanced potency observed when the side chain of THC was changed from straight
to branched was not observed when the end chain of AN was similarly changed.
In tr od u ction
Since the discovery of anandamide (AN; 1, R ) C5H11),
a putative endogenous ligand for the cannabinoid recep-
1
tor (CB1) in the brain, several groups including our
own2
-11
have initiated the synthesis and biological
evaluation of AN analogs. AN shares biological activi-
9
ties exhibited by ∆ -THC, including decreased sponta-
neous motor activity, immobility, and production of
hypothermia and analgesia. However, it has a faster
9
6
onset and shorter duration of action than ∆ -THC.
Binding studies have to be carried out in the presence
of the enzyme inhibitor phenylmethylsulfonyl fluoride
(
PMSF) since AN is susceptible to metabolic degradation
5
,12,13
due to an amidase.
Since the discovery of AN, two
additional brain constituents (homo-γ-linolenylethano-
lamide and docosatetraenylethanolamide) that are struc-
turally related to AN and exhibit cannabinoid activity
have been reported.14 Although several analogs have
been reported with a variation in the ethanolamine part
of AN, only limited work has been carried out in the
arachidonic acid part of AN, indicating that the presence
of three double bonds or two more methylene groups
had little effect on activity.3 However, the activity was
greatly reduced with the introduction of an additional
3,5,10
double bond in the end pentyl chain (C17-C18).
3
This observation led Felder et al. to suggest that “the
end pentyl chain may play a role similar to the pentyl
9,10
side chain of THC”. Results from our group
have
revealed that the introduction of substituents on the
carbon in the 2-position of the arachidonic acid part of
AN enhances metabolic stability, presumably by steri-
cally hindering the amidase(s) responsible for AN
degradation. Of this series the methyl substituent was
found to be optimal, and it was found that (()-2-
methylarachidonyl-2′-fluoroethylamide (2, R ) C5H11;
,5
*
Author to whom inquiries about this paper should be sent.
Department of Pharmacology and Toxicology, Medical College of
†
Virginia, Richmond, VA 23298.
X
Abstract published in Advance ACS Abstracts, September 15, 1997.
S0022-2623(97)00212-4 CCC: $14.00 © 1997 American Chemical Society

3
618 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 22
Ryan et al.
Sch em e 1
O-689) was the most potent both in binding (Ki ) 5.7 (
synthesized from 6d while all the remaining target
compounds were prepared from 6a . It is noteworthy
that the Wittig reaction between 6a and the ylide of the
phosphonium salts to form 7e-g failed under a variety
of conditions.
2
.1 nM) and in in vivo activity. Additionally, our
molecular modeling studies also suggested that the end
pentyl chain of AN is mimicking the side chain in
THCs.15 Our studies further showed that in various AN
analogs, the tetraene analogs, had an optimal pharma-
To synthesize the above compounds we required a
9
cophore overlay with ∆ -THC, and the potencies were
large supply of the aldehyde 6a . Its synthesis was
predictable by QSAR with a reasonable degree of
accuracy.15 With this background we designed com-
pounds 1e and 2a as they both carry a dimethylheptyl
17
18
reported by Corey et al. and Manna et al. who both
used essentially the same procedure. In 1979 Corey et
19
al. reported a procedure using 90% H2O2 for the
(
DMH) chain which is known to impart very potent
selective epoxidation of arachidonic acid 3a to give 4
via the peroxy acid 3b. As 90% hydrogen peroxide is
hazardous to use and no longer commercially available,
we developed a modified procedure for the synthesis of
1
6
agonist activity in THCs. To test this hypothesis and
lend support to our alignment strategy for AN and
THCs, we synthesized compounds 1a -h and 2a -f and
examined the effect of changing the length and branch-
ing of the end pentyl chain (C5H11) of AN.
4
using 50% hydrogen peroxide. We also tried the
20
modified procedure described by Perrier et al. using
0% aqueous H2O2 but in our hands it could not be
3
Ch em istr y
scaled up successfully.
Our general strategy for the synthesis of compounds
In our modified procedure, the acid chloride of arachi-
donic acid (oxalyl chloride in benzene at 0 °C f 30 °C,
24 h) was treated with 0.5 equiv of pyridine in dry THF
at 0 °C, followed by a solution of LiOH‚H2O in 50%
H2O2. After stirring for 20 min, the reaction was
quenched with water and then extracted with CH2Cl2.
The resulting solution was left standing over Na2SO4
for 1.5 to 3 h while the peracid 3b rearranged to the
epoxide acid (followed by TLC). It was isolated as the
methyl ester 4 (diazomethane/ether at 0 °C) after
chromatography. This procedure gave the epoxide 4 in
reproducible yields of 30-35% from 3a in addition to
50% recovered methyl arachidonate. Conversion of 4
to the diol 5 (53%) was achieved with 1.2 M solution of
HClO4 in THF by stirring at 30 °C for 16 h. Oxidation
with lead tetraacetate/CH2Cl2 at -18 °C for 0.5 h gave
the unstable aldehyde 6a , which was quickly filtered
through a thin pad of Celite/silica gel and washed with
hexanes. After removal of the solvent and drying in
1
a -h and 2a -f is based on the conversion of the known
1
7,18
aldehyde ester 6a
(Scheme 1) to the appropriate
tetraene esters 7. This was achieved by carrying out a
Wittig reaction between 6a and the ylide generated from
the phosphonium salt of the appropriate alkyl halide
or between 6d (prepared from 6a by reduction f
mesylation/sodium iodide f phosphonium salt) and the
appropriate alkyl aldehyde. The tetraene esters 7 were
9
then hydrolyzed to the corresponding acids and trans-
formed into AN analogs 1 via their acid chlorides and
treated with excess ethanolamine as in the synthesis
of AN.1
,9
The racemic fluorinated analogs 2 were
synthesized from 7 by methylation in the 2-position
using LDA and CH3I in THF at -40 °C to give the esters
and then hydrolyzed to give the appropriate acids.
8
These acids via their acid chlorides and subsequent
treatment with excess fluoroethylamine gave target
9
,10
compounds 2.
Thus compounds 1e, and 2a -c were

Potent Anandamide Analogs
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 22 3619
vacuo the aldehyde 6a was either used as such for the
Wittig reaction or reduced to the alcohol with NaBH4
in methanol to form 6b (43%).
7g. Improved yields (∼70%) were obtained in the Wittig
reaction when lithiohexamethyldisilazane (LiHMDS)
was used in a 3:2 HMPA/THF mixture at -78 °C and a
solution of 6d in THF was added at -78 °C to form the
2
2
ylide (see Experimental Section).
In summary, arachidonic acid was converted to the
aldehyde ester 6a or the alcohol 6b in an overall yield
of 15% (based on recovered starting material). The
yields in the Wittig reactions varied from 30-70%.
P h a r m a cology a n d Discu ssion of Resu lts
In previous studies,6,10 we have demonstrated that ∆9-
THC binds to the CB1 receptor with a Kd of 41 nM and
is very effective in the four behavioral tests presented
in Table 1. Anandamide (AN) is not effective in
competing for binding to the CB1 receptor unless the
enzyme inhibitor, PMSF, is added. In the presence of
PMSF, AN binds to the receptor with an affinity
For the synthesis of 7a -d , the appropriate alkyl
bromides (commercially available) were refluxed in
xylene with triphenylphosphine to form the phospho-
nium salts21 which on treatment with butyllithium/
hexane in THF at -78 °C formed the ylide. Further
2
2
treatment with the aldehyde 6a in a Wittig reaction
9
approximately half that of ∆ -THC. AN is also effective
gave the tetraenes 7. For the synthesis of 7h the
phosphonium salt 13c was prepared from 11 which was
synthesized from 3,3-dimethylacrylic acid and sec-
in all four mouse pharmacological assays, although it
9
is considerably less potent than ∆ -THC. In an effort
to retard the enzymatic hydrolysis, a methyl group was
added at C-2 in a fluoro derivative of AN to form fluoro-
2
3
butanol. A Grignard reaction of 11 in the presence of
CuCl and trimethylsilyl chloride at 0 °C with n-pentyl-
magnesium chloride gave the ester 12 (97%).24 Reduc-
tion with LAH formed the alcohol 13a which on treat-
ment with triethylamine/mesyl chloride25 followed by
treatment with tetrabutylammonium iodide gave 13b.
2
-methylanandamide (F-2-Me-AN) as depicted in struc-
ture 2 (Table 1). As shown in Table 1, F-2-Me-AN was
capable of binding to the CB1 receptor either in the
10
presence or absence of PMSF. Despite the dramatic
increase in binding affinity that occurred with the
addition of this C-2 methyl, pharmacological potency
between AN and F-2-Me-AN was similar for antinoci-
ceptive activity and only fourfold different for spontane-
ous activity. However it showed the expected increase
in potency for hypothermia (9-fold) and ring immobility
(12-fold).
In order to assess the optimal length of the alkyl
terminus to AN, analogs 1a -d were prepared. Addi-
tions of one and two methylenes to AN resulted in 1a
and 1b, respectively, which failed to bind effectively in
the absence of PMSF. On the other hand, additions of
three and four methylenes produced analogs 1c and 1d ,
respectively, that exhibited much greater affinity and
metabolic stability as evidenced by their binding affini-
ties with and without PMSF. In the presence of PMSF,
all four of these analogs bound with moderately high
affinities in the range of 18-55 nM. Based upon these
affinities, it was anticipated that all four analogs would
be pharmacologically more potent than AN. Analogs
1a -c failed to produce significant pharmacological
effects at doses up to 30 mg/kg, with the exception of
analog 1b that was a partial agonist in decreasing
spontaneous activity. Analog 1d was reasonably potent
in reducing spontaneous activity and in producing
antinociception and ring immobility; however, it acted
only as a partial agonist with maximal effects of 57, 50,
and 42% in the respective assays. It was also unique
in that it failed to produce hypothermia at doses up to
30 mg/kg.
The phosphonium salt 13c was similarly prepared as
described above and then condensed as the ylide with
aldehyde 6a to form 7h . Compound 7i was synthesized
in the same fashion. For the synthesis of 7j, the same
sequence as described for 7h was used except that
3
-methylacrylic acid was used in place of 3,3-dimethy-
lacrylic acid.
Compounds 7e to 7g were synthesized by treatment
of the ylide from 6d and the appropriate aldehydes. The
phosphonium salt 6d was prepared from 6b via 6c
following the same sequence as described for 13c. For
the synthesis of aldehyde 9c, ethyl isobutyrate was
treated with LDA followed by n-bromohexane to form
the ester 9a . Reduction with LAH converted 9a to 9b,
which on oxidation with pyridinium chlorochromate
(
The addition of a dimethylheptyl (DMH) side chain
to ∆ -THC dramatically enhanced pharmacological ac-
PCC)2 gave the aldehyde 9c. Wittig reaction with this
6
9
aldehyde and the ylide of 6d furnished 7e. The homolog
f was similarly prepared from 2,2-dimethylnonanal.
tivity from 10-75 fold, depending upon the pharmaco-
2
7
7
logical assay.
A similar structural change in AN
The synthesis of 7g was achieved in a similar way.
Methyl caprylate was monoalkylated to give 10a which
was then transformed to the aldehyde 10c as in the case
of 9c. A Wittig reaction with the ylide of 6d furnished
resulted in analog 1e that had a receptor affinity 12
times greater than that of AN. It was also more potent
than AN in producing hypoactivity (8 times), antinoci-
ception (3 times), hypothermia (3 times) and ring

3
620 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 22
Ryan et al.
Ta ble 1. Receptor Affinity and Pharmacological Effects of Analogsa
a
Ki’s (nM) were determined in the presence and absence of PMSF and are expressed as means ( SE for at least three experiments.
The pharmacological measures included inhibition of spontaneous activity (SA), antinociception as measured by the tail-flick response
(
TF), hypothermia as changes in rectal temperature (RT), and ring immobility (RI). The pharmacological data are expressed as the ED50
b
32
c
Not tested. d Reported previously.6 e Reported previously.
10 f
(mg/kg). Inactive (IA); not determined (ND). Reported previously.
Inactive
g
because maximal effects at 30 mg/kg were either less than 30% or, in the case of hypothermic, less than 1 °C. Partial agonist. Maximum
effect of 40-60% (30-40% for immobility) at 30 mg/kg.
immobility (11 times). 1e is also unique in that it
the dimethylheptyl derivative 2a and the least affinity
by the dimethyloctyl analog 2b. All of these compounds
produced robust pharmacological effects with a few
exceptions. 2b was only a partial agonist and in analog
2d , where the gem-dimethyl is moved to the adjacent
carbon, a similar profile (i.e. active in all four tests) was
obtained in pharmacological measures as found in 1f
except that partial agonism was found in the ring
immobility test rather than in the antinociceptive test.
This is interesting, since it could provide a lead for
separation of pharmacological activities. These analogs
were similar to AN in that they failed to produce
hypothermia greater than 3.0 °C. However, their
potencies were not always consistent with potencies in
the other pharmacological assays. For example, analogs
2c and 2e, which are diastereomeric mixtures, were
considerably more potent in reducing rectal temperature
than in producing the other effects, whereas the degree
of hypothermia caused by 2d and 2f was considerably
less than their other effects.
produced 84% immobility making it more efficacious
9
than ∆ -THC (60% maximal effect). Moving the dim-
ethyl groups to the adjacent C atom to form 1f had
relatively little effect on activity since both 1e and 1f
were active in all four tests. The fact that subtle
changes in the terminal alkyl position can be influential
is amply illustrated by reducing the side chain of 1f by
one methylene to produce 1g. Analog 1g was a weak
partial agonist with regard to reduction of spontaneous
activity and was inactive in the other three tests. We
were able to evaluate a dose of 100 mg/kg in SA and
TF and no greater effects were obtained, supporting our
contention of partial agonistic effects. Insufficient
quantities precluded testing this higher dose in the
other two pharmacological measures. 1h , the monom-
ethyl derivative of 1f, has a profile quite similar to that
of 1f.
Alkyl substitutions similar to those described above
were also prepared in the fluoro-2-methylanandamide
series, and all of these analogs had high receptor
affinity. The highest receptor affinity was exhibited by
These findings provide additional evidence that AN
9
and ∆ -THC share common pharmacological properties

Potent Anandamide Analogs
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 22 3621
in that these analogs bound to the cannabinoid receptor
and produced pharmacological effects in all four mouse
pharmacological procedures. These data also reveal
some important pharmacological differences between
(ethanol) at -20 °C and analyzed after about 2 years in the 1
series and 1 year in the 2 series. All reactions were carried
out under nitrogen.
Meth yl 14,15-Ep oxya r a ch id on a te (4). To a solution of
arachidonic acid (6.59 g, 21.6 mmol) in anhydrous benzene (60
mL) cooled to 0 °C was added oxalyl chloride (3.78 mL, 43.3
mmol). The mixture was stirred overnight allowing it to come
to room temperature. Evidence of the formation of the acid
9
AN and ∆ -THC. It is unclear why several of the
analogs in the 1 series failed to produce pharmacological
effects despite relatively good receptor affinity. The
receptor binding studies demonstrate that these analogs
are metabolized by brain tissue, as it is for anandamide.
While it is possible that they are degraded to such an
extent in vivo that they are inactive, it seems unlikely
that metabolism would differ between the potent dim-
ethylheptyl analog 1e and the other analogs. In addi-
tion, the pharmacological evaluations were made rapidly
after iv administration which has been shown to ef-
fectively measure AN’s effects as well as a wide range
-1
chloride was determined by IR (CdO, 1790 cm ). The solvent
was removed in vacuo followed by addition of benzene and
evaporation on a rotary evaporator to remove traces of oxalyl
chloride. To a solution of this acid chloride in anhyd THF (90
mL) was added pyridine (0.87 mL, 10.8 mmol), and the mixture
was stirred for 10 min at 0 °C. A solution of LiOH‚H
2
O (0.91
g, 21.6 mmol) in 50% H O (29.4 mL) was added to the mixture
2
2
in one portion at 0 °C, and the mixture was stirred for 20 min.
The reaction was quenched with pH 7 buffer (50 mL) and brine
(
50 mL) and then extracted with CH
layer was saturated with NaCl and extracted further with CH
Cl
(4 × 50 mL). The combined organic extracts were washed
with brine (50 mL) and dried (Na SO ). The organic extracts
were dried for 1.5 to 3 h while following the formation of the
epoxy acid by TLC (30% ethyl acetate/hexanes; R ) 0.40 for
epoxy acid). The solvent was decanted, and the drying agent
was washed with CH Cl After removal of the solvent in
2 2
Cl (50 mL). The aqueous
2
-
1
0
of AN analogs that are not metabolically stable. We
cannot rule out the possibility that the inactive analogs
in series 1 lack intrinsic activity or they may have some
antagonist activity. On the other hand, analogs in the
metabolically stable 2 series were quite potent in most
of the pharmacological assays. It is intriguing that the
alterations in the putative “side-chain” did not have the
anticipated impact on pharmacological potency. These
findings reinforce the notion that AN’s interactions with
2
2
4
f
2
2
.
vacuo, the resultant oil was taken up in anhyd ether (100 mL)
and cooled to 0 °C. The solution was then treated with an
excess of diazomethane solution in ether and stirred for 15
min. After evaporation of the unreacted diazomethane in the
fume hood at room temperature, the solvent was removed in
vacuo to yield a yellow oil which was purified by flash
9
the cannabinoid receptor is not identical to that of ∆ -
THC, in consideration of the weak hypothermic effects
of the anandamides.
2
chromatography (225 g SiO , 5% ethyl acetate/hexanes): Yield
2
8
3
.30 g (32%); NMR δ 0.89 (t, J ) 6.6 Hz, 3H), 1.27-1.58 (m,
H), 1.63-1.85 (m, 2H), 2.02-2.43 (m, 6H), 2.75-3.00 (m, 6H),
.68 (s, 3H), 5.33-5.58 (m, 6H).
Con clu sion s
We conclude that the SAR of the end pentyl chain in
AN is very similar to that of THCs with certain
exceptions. In anandamides like the THCs, increasing
the length of the end pentyl chain increases the binding
affinity; however, unlike the THCs, the in vivo activity
does not parallel the binding affinities. Additionally,
increasing the chain length in anandamides appears to
impart some metabolic stability toward amidases.
Branching the chain enhances both binding affinity and
in vivo activity as in THCs. However, by branching the
chain, the effect in pharmacological measures is not as
dramatic in the AN series as in THCs. This dissimilar-
ity might be a reflection of the differences in chemical
structure and in the biodisposition of THCs and anan-
damides.
Meth yl 14,15-Dih yd r oxya r a ch id on a te (5). To a solution
of epoxide 4 (1.22 g, 3.66 mmol) in THF (60 mL) were added
O (30 mL) and 1.2 M HClO (15 mL). The reaction was
stirred overnight at room temperature and quenched with pH
buffer (50 mL). Ethyl acetate (30 mL) was added, and the
H
2
4
7
layers were separated. The aqueous layer was saturated with
NaCl and extracted with ethyl acetate (3 × 30 mL). The
combined organic extracts were washed with brine (30 mL)
and dried (MgSO
brown oil which was purified by flash chromatography (100 g
SiO , 20% ethyl acetate/hexanes) to yield 0.68 g (53%) of a
4
). Removal of the solvent in vacuo gave a
2
colorless oil: NMR δ 0.89 (t, J ) 6.6 Hz, 3H), 1.20-1.60 (m,
8
3
H), 1.63-1.85 (m, 2H), 2.05-2.40 (m, 6H), 2.75-2.92 (m, 4H),
.45-3.58 (m, 2H), 3.68 (s, 3H), 5.21-5.71 (m, 6H).
Met h yl 14-H yd r oxy-5,8,11-a ll-cis-t et r a d eca t r ien oa t e
(6b). To a stirred solution of 5 (1.12 g, 3.17 mmol) in CH
2
Cl
20 mL) at -18 °C (ice/salt bath) was added a solution of lead-
IV) acetate (1.42 g, 3.21 mmol) in CH Cl (20 mL). The
2
(
(
2
2
solution was stirred for 30 min at -18 °C. The reaction was
filtered through a thin pad of Celite/silica and the filter cake
was washed with anhyd hexanes (5 × 10 mL). The solvent
was removed in vacuo, and the unstable aldehyde 6a was
Exp er im en ta l Section
1H NMR spectra were recorded on either a Bruker 100 or a
3
Varian XL400 spectrophotometer using CDCl as the solvent
with trimethylsilane as an internal standard. Thin layer
chromatography (TLC) was carried out on Baker Si 250F
plates. Visualization was accomplished with either iodine
vapor, UV exposure, or treatment with phosphomolybdic acid
dissolved in MeOH (12 mL) and cooled to 0 °C. NaBH
g, 3.8 mmol) was added in one portion, and the reaction was
stirred for 10 min. The reaction was quenched with H O (20
4
(0.14
2
mL), and most of the MeOH was removed in vacuo. The
aqueous layer was saturated with NaCl and extracted with
ethyl acetate (3 × 20 mL). The combined organic layers were
(PMA). Flash chromatography was carried out on EM Science
Silica Gel 60. Elemental analyses were performed by Atlantic
Microlab, Atlanta, GA. The purity of the products on which
the high-resolution mass spectral data are reported were
4
washed with brine (15 mL) and dried (MgSO ). Removal of
the solvent in vacuo gave a cloudy oil which was purified by
flash chromatography (120 g SiO , 15% ethyl acetate/hexanes)
1
determined by GC, TLC, and H NMR analysis. The analyses
2
by gas chromatography (GC) were performed on a Perkin
Elmer 8500 instrument equipped with a 25 m fused silica
capillary column, 0.53 mm i.d. with 0.25 µm film thickness
to yield 0.34 g (43%) of alchol 6b as a yellow oil: NMR δ 1.62-
1.82 (m, 3H), 2.02-2.50 (m, 6H), 2.75-2.90 (m, 4H), 3.59-
3.72 (m, 5H), 5.32-5.56 (m, 6H).
2
2
(
007 Methyl Phenyl (5%) Silicone; Quadrex Corp.). For anan-
Meth yl 14-Iod o-5,8,11-a ll-cis-tetr a d eca tr ien oa te (6c).
To a stirred solution of alcohol 6b (0.19 g, 0.75 mmol) and
triethylamine (0.17 mL, 1.2 mmol) in CH Cl (1.5 mL) cooled
damide analogs 1a -h the column temperature was pro-
grammed to increase from 60 to 280 °C at a rate of 20 °C/min.
For analogs in the 2 series, the temperature ranged from 140
to 280 °C at a rate of 20 °C/min. The injector and detector
temperatures were set at 300 °C for both. The GC purity is
indicated in parenthesis of samples kept as stock solutions
2
2
to -20 °C was added methanesulfonyl chloride (0.08 mL, 0.97
mmol). The reaction was stirred for 15 min at -20 °C, and
2
then H O (1.25 mL) was added and the reaction was allowed
to warm to room temperature. The reaction was diluted with

3
622 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 22
Ryan et al.
CHCl
phase was dried (MgSO
vacuo. A solution of the crude mesylate and sodium iodide
0.45 g, 3.0 mmol) dissolved in acetone (5 mL) was refluxed
under nitrogen for 1.5 h. The cooled reaction was diluted with
hexanes (20 mL) and filtered through a pad of Celite. The
solvent was removed in vacuo. The resultant yellow oil was
dissolved in 10% ethyl acetate/hexanes and filtered through a
bed of silica gel. The solvent was removed in vacuo to yield
the product (0.26 g, 95%) as a colorless oil which was used
without further purification: NMR δ 1.61-1.88 (m, 2H), 2.03-
3
(10 mL), and the layers were separated. The organic
2,2-Dim eth ylocta n a l (9c).26 To a suspension of PCC (1.54
4
), and the solvent was removed in
g, 7.1 mmol) in CH
g, 4.8 mmol) in CH
2
Cl
Cl
2
(10 mL) was added dropwise 9b (0.75
(5 mL). The reaction was stirred at
2
2
(
room temperature for 2 h followed by diluting with ether (35
mL) and decanting the solvent. The black residue was washed
with ether (3 × 10 mL) followed by decanting the ether. The
combined ether washings were filtered through a bed of
Florisil, and the solvent was removed in vacuo to give the
product as a yellow/green oil which was used directly: Yield
0.65 g (87%); NMR δ 0.88 (t, J ) 6.1 Hz, 3H), 1.04 (s, 6H),
1.20-1.50 (m, 10H), 9.44 (s, 1H).
Meth yl 2-Meth ylca p r yla te (10a ). The product was pre-
pared according to the method described for 9a and was used
without futher purification: Yield 97%; NMR δ 0.88 (t, J )
6.0 Hz, 3H), 1.16 (d, J ) 7.2 Hz, 3H), 1.20-1.72 (m, 10H), 3.67
(s, 3H).
2
5
.40 (m, 6H), 2.56-2.85 (m, 4H), 3.15 (m, 2H), 3.67 (s, 3H),
.29-5.48 (m, 6H).
Met h yl 5,8,11-a ll-cis-Tet r a d eca t r ien oa t e-14-t r ip h e-
n ylp h osp h on iu m Iod id e (6d ).22 A solution of 6c (0.26 g,
0
.71 mmol) and triphenylphosphine (0.21 g, 0.82 mmol) in
acetonitrile (3 mL) was heated at reflux overnight. While the
reaction was still warm, the condenser was removed and a
2
-Meth ylocta n ol (10b). The product was prepared ac-
cording to the method described for 9b: Yield 83%; NMR δ
.75-0.99 (m, 6H), 1.15-1.70 (m, 12H), 3.30-3.60 (m, 2H).
-Meth ylocta n a l (10c). The product was prepared accord-
ing to the method described for 9c: Yield 83%; NMR δ 0.89
t, J ) 5.5 Hz, 3H), 1.09 (d, J ) 7.0 Hz, 3H), 1.16-1.70 (m,
0H), 2.25-2.45 (m, 1H), 9.61 (d, J ) 1.9 Hz, 1H).
,2-Dim eth yln on a n a l. The product was prepared accord-
stream of N
evaporated. After cooling, the excess Ph
repeated washing and decanting with hexanes/benzene (1/1,
mL). The solvent was decanted, and the procedure was
2
was blown over the reaction until the solvent
0
3
P was removed by
2
5
(
1
repeated. The remaining solvent was removed in vacuo, and
the orange gum was heated in a vacuum oven overnight at 90
2
°C and used directly in the Wittig reaction: Yield 0.42 g (95%).
ing to the method described for 9c: Yield 56%; NMR δ 0.88
Met h yl 16,16-Dim et h yl-5,8,11,14-a ll-cis-d ocosa t et r a -
2
2
(t, J ) 6.0 Hz, 3H), 1.05 (s, 6H), 1.15-1.45 (m, 12H), 9.44 (s,
en oa te (7e). To a solution of HMDS (0.14 mL, 0.68 mmol)
and HMPA (0.47 mL, 2.7 mmol) in THF (1.9 mL) at -78 °C
was added n-butyllithium (2.2 M in hexanes, 0.29 mL, 0.64
mmol), and the reaction was warmed for 3 min to 0 °C followed
by cooling back to -78 °C. A solution of phosphonium iodide
1
H).
(
()-Met h yl 2,16,16-Tr im et h yl-5,8,11,14-a ll-cis-d oco-
sa tetr a en oa te (8e). The product was prepared according to
the alkylation method described for 9a using MeI in the place
of n-bromohexane. The product was purified by flash chro-
matography (0.5% ethyl acetate/hexanes): Yield 66%; NMR δ
6
d (0.42 g, 0.68 mmol) in THF (2.5 mL) was added dropwise
to the LiHMDS solution, and the reaction was stirred for 30
min at -78 °C. Then a solution of 9c (0.15 g, 0.95 mmol) in
THF (0.75 mL) was added to the ylide and the dry ice/acetone
bath was removed. The reaction was allowed to warm to room
temperature and stirred for 1 h. The reaction was diluted with
hexanes (12 mL), and the precipitate was allowed to settle.
The solvent was decanted, and the residue was extracted with
another portion of hexanes (6 mL). The combined organic
extracts were evaporated to give the crude product which was
0
1
(
.88 (t, J ) 6.4 Hz, 3H), 1.09 (s, 6H), 1.16 (d, J ) 7.0 Hz, 3H),
.27 (br s, 10H), 1.40-2.20 (m, 4H), 2.32 (m, 1H), 2.80-2.98
m, 6H), 3.67 (s, 3H), 5.13-5.42 (m, 8H).
(()-2,16,16-Tr im e t h yl-5,8,11,14-a ll-cis-d ocosa t e t r a -
en oic Acid . To a solution of 8e (0.07 g, 0.18 mmol) in MeOH
(9 mL) and H
mmol), and the reaction was heated to 55 °C overnight. After
cooling, the reaction was diluted with H O (10 mL) and ether
2
2
O (3 mL) was added LiOH‚H O (0.05 g, 1.27
2
purified by flash chromatography (60 g SiO
% ethyl acetate/hexanes): Yield 0.18 g (70%); NMR δ 0.88
2
, gradient 0% to
(10 mL). The layers were separated, and the aqueous layer
was acidified with 1 N HCl and saturated with NaCl. The
aqueous layer was extracted with 3 × 10 mL ether, and the
combined organic extracts were washed with brine (5 mL) and
2
(
t, J ) 5.5 Hz, 3H), 1.09 (s, 6H), 1.26 (br s, 10H), 1.60-1.85
m, 2H), 2.02-2.45 (m, 6H), 2.75-2.98 (m, 6H), 3.66 (s, 3H),
(
5
.13-5.41 (m, 8H).
4
dried (MgSO ). The solvent was removed in vacuo, and the
Eth yl 2,2-Dim eth ylocta n oa te (9a ). To a solution of LDA
resultant oil was dried in vacuo and used directly: Yield 0.06
g (94%); NMR δ 0.88 (t, J ) 6.4 Hz, 3H), 1.09 (s, 6H), 1.16 (d,
J ) 7.0 Hz, 3H), 1.26 (br s, 10H), 1.37-2.55 (m, 5H), 2.81-
2.98 (m, 6H), 5.13-5.42 (m, 8H), 10.11 (br s, 1H).
(
2.0 M in THF/ethylbenzene/heptane, 13.0 mL, 26 mmol) in
THF (60 mL) at -45 °C was added ethyl isobutyrate (1.01 g,
.7 mmol), and the reaction was maintained at -45 °C for 45
8
min with stirring. The reaction was then cooled to -60 °C,
and n-bromohexane (14.38 g, 87 mmol) was added in one batch.
The reaction was stirred for 3.5 h, slowly coming to room
temperature, followed by quenching with H
2
O (50 mL) and
diluting with ether (100 mL). The layers were separated, the
aqueous layer was saturated with NaCl and extracted with 3
(()-2,16,16-Tr im eth yl-5,8,11,14-a ll-cis-docosatetr aen oyl-
2′-flu or oeth yla m id e (2a ). To a solution of 2,16,16-trimethyl-
5,8,11,14-all-cis-docosatetraenoic acid (0.06 g, 0.17 mmol) in
benzene (5 mL) cooled to 0 °C were added oxalyl chloride (0.03
mL, 0.34 mmol) and a drop of DMF. The reaction was slowly
brought to room temperature while stirring for 2 h. The
solvent was removed in vacuo, and the oil was treated twice
with fresh benzene (10 mL) and evaporated to remove mois-
×
25 mL of ether, and the combined ether extracts were
washed with 1 N HCl (15 mL) and brine (2 × 20 mL), and
dried (MgSO ). Removal of the solvent in vacuo gave an oil
4
ture. A solution of the acid chloride in CH
added to a solution of fluoroethylamine hydrochloride (0.17 g,
1.71 mmol) in CH Cl (10 mL) pretreated (stirred for 15 min
2 2
Cl (3 mL) was
which was fractionally distilled at reduced pressure to give
the product as a colorless oil: Yield 1.09 g (63%); bp 64-66
2
2
°
1
C /1.3 mm; NMR δ 0.88 (t, J ) 6.2 Hz, 3H), 1.15 (s, 6H), 1.17-
at 0 °C to make the free base) with triethylamine (0.24 mL,
1.73 mmol) at 0 °C. The reaction was stirred for 30 min at 0
°C followed by quenching with brine (10 mL). The reaction
.49 (m, 13H), 4.11 (q, J ) 7.1 Hz, 3H).
,2-Dim eth ylocta n ol (9b). To a suspension of LAH (0.62
2
g, 16.3 mmol) in THF (15 mL) cooled to 0 °C was added
dropwise a solution of ester 9a in THF (5 mL). The reaction
was stirred overnight allowing it to come slowly to room
temperature. The reaction was quenched by the sequential
addition of water (0.61 mL), 15% sodium hydroxide solution
was diluted with CH
separated. The aqueous layer was extracted with CH
× 10 mL), the combined organic extracts were dried (MgSO
and the solvent was removed in vacuo. The crude product was
purified by flash chromatography (15 g of SiO , 15% ethyl
2
Cl
2
(10 mL), and the layers were
Cl (3
),
2
2
4
2
(
0.61 mL), and water (1.83 mL). The white suspension was
filtered through a pad of Celite, and the solid was washed with
ether. The filtrate was washed with H
O (2 × 20 mL), brine
10 mL), and dried (MgSO ). The solvent was removed in
acetate/hexanes) to give the product as a yellow oil: Yield 0.05
g (65%); NMR δ 0.88 (t, J ) 4.0 Hz, 3H), 1.09 (s, 6H), 1.16 (d,
J ) 6.9 Hz, 3H), 1.26 (br s, 10H), 1.39-2.34 (m, 5H), 2.72-
2.98 (m, 6H), 3.58 (ddt, J ) 29.0, 5.0, 4.5 Hz, 2H), 4.49 (dt, J
) 47.5, 4.6 Hz, 2H), 5.13-5.53 (m, 8H), 5.82 (br s, 1H); HRMS
2
(
4
vacuo to give the product as a colorless oil which was used
directly: Yield 0.78 g (91%); NMR δ 0.85-0.93 (m, 9H), 1.23-
1
calcd for C27
H47FNO (M + 1), 420.3641; found 420.3591; GC,
.54 (m, 11H), 3.30 (s, 2H).
R
t 6.13 min (97%).

Potent Anandamide Analogs
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 22 3623
H yd r olysis of 7e. 16,16-Dim et h yl-5,8,11,14-a ll-cis-
d ocosa tetr a en oic Acid . The product was prepared according
to the method described for 2,16,16-trimethyl-5,8,11,14-all-cis-
docosatetraenoic acid: Yield 99%; NMR δ 0.88 (t, J ) 6.4 Hz,
to remove water. The residual oil was then dissolved in
benzene and refluxed with excess tetra-n-butylammonium
iodide (45.0 g, 123 mmol, 2.0 equiv) for 1.5 h. The reaction
mixture was then filtered and evaporated, and the residue was
taken up in pentane. It was passed through a plug of silica
and evaporated to leave the product as an oil (10.32 g, 62%);
NMR δ 1.86 (s, 6H), 1.90 (t, J ) 7.5 Hz, 3H), 1.30 (br s, 8H),
1.80-2.00 (m, 2H), 3.05-3.25 (m, 2H).
3
2
H), 1.09 (s, 6H), 1.26 (br s, 10H), 1.63-1.86 (m, 2H), 2.04-
.44 (m, 4H), 2.81-2.98 (m, 6H), 5.19-5.53 (m, 8H), 10.20 (br
s, 1H).
1
6,16-Dim eth yl-5,8,11,14-a ll-cis-d ocosa tetr a en oyleth a -
2
1
n ola m id e (1e). To a solution of 16,16-dimethyl-5,8,11,14-all-
cis-docosatetraenoic acid (0.14 g, 0.38 mmol) in benzene (4 mL)
cooled to 0 °C were added oxalyl chloride (0.07 mL, 0.75 mmol)
and DMF (0.0029 mL, 0.04 mmol). The reaction was slowly
brought to room temperature while stirring for 2 h. The
solvent was removed in vacuo, and the oil was treated twice
with fresh benzene (10 mL) and evaporated to remove mois-
3,3-Dim eth yloctyltr iph en ylph osph on iu m Iodide (13c).
Triphenylphosphine (13.8 g, 52.7 mmol, 1.4 equiv) was dis-
solved in xylene (150 mL) with 13b (10.2 g, 38.1 mmol) and
refluxed overnight. The xylene was decanted from the oil and
the residue was covered with ether and the flask was put in a
sonic bath for 1 h. After decanting the ether to remove
triphenylphosphine oxide, the residue was similarly treated
again. The residue was the desired product isolated as a
hygroscopic solid (4.28 g, 21%); mp 66.1-68.5 °C; NMR δ 0.85
(t, J ) 6.9 Hz, 3H), 0.97 (s, 6H), 1.10-1.70 (m, 10 H), 3.20-
3.60 (m, 2H), 7.65-7.95 (m, 15H).
Met h yl 17,17-Dim et h yl-5,8,11,14-a ll-cis-d ocosa t et r a -
en oa te (7h ). A solution of the Wittig reagent was prepared
by the dropwise addition of a solution of n-butyllithium in
hexanes (1.0 mL of 2.5 M solution, 2.5 mmol, 1.15 equiv) to a
solution of 13c (1.20 g, 2.26 mmol, 1.04 equiv) in THF at -78
°C and stirred at that temperature for 0.5 h. The aldehyde
6a (prepared from 0.766 g, 2.17 mmol of 5) was dissolved in
dry THF (10 mL) and added dropwise to the solution of the
Wittig reagent. The reaction mixture was stirred, allowing it
to warm to room temperature over 2.5 h. The reaction was
quenched with brine, extracted with ethyl acetate, dried,
evaporated, and then chromatographed on silica gel eluting
with 2.5% ethyl acetate/hexanes to yield the product as an oil
(0.814 g, 30%); NMR δ 0.85 (s, 6H), 0.95 (t, J ) 7.5 Hz, 3H),
1.20-2.45 (m, 16H), 2.70-2.90 (m, 6H), 3.65 (s, 3H), 5.20-
5.50 (m, 8H).
ture. A solution of the acid chloride in CH
added to a solution of ethanolamine (0.27 mL, 3.77 mmol) in
CH Cl (2 mL) at 0 °C. The reaction was stirred for 10 min
and poured into H O (10 mL). The layers were separated, and
the aqueous layer was saturated with NaCl followed by further
extractions with CH Cl
(3 × 10 mL). The combined organic
extracts were washed with brine (10 mL) and dried (Na SO ).
After removal of the solvent in vacuo, the crude product was
dissolved in 10% MeOH/CHCl and passed through a pad of
2 2
Cl (2 mL) was
2
2
2
2
2
2
4
3
silica gel. The solvent was removed in vacuo, and the oil was
taken up in pentane and filtered through a tightly packed
cotton plug. Removal of the solvent gave pure product as a
yellow oil: Yield 0.14 g (93%); NMR δ 0.88 (t, J ) 5.4 Hz, 3H),
1
.09 (s, 6H), 1.27 (br s, 10H), 1.62-1.86 (m, 2H), 2.03-2.30
(
(
(
m, 4H), 2.68-2.98 (m, 6H), 3.40 (dt, J ) 4.0, 6.3 Hz, 2H), 3.71
t, J ) 5.0 Hz, 2H), 5.13-5.41 (m, 8H), 6.19 (br s, 1H). Anal.
C
26
H
45NO
2
2
‚0.2H O) C, H, N.
2
4
sec-Bu tyl 3,3-Dim eth ylh ep ta n oa te (12). sec-Butyl 3,3-
dimethylacrylate (2.0 g, 12.80 mmol), prepared from 3,3-
dimethylacrylic acid and 2-methylpropyl alcohol according to
2
3
a literature procedure, was dissolved in THF (5 mL) and
treated with solid CuCl (0.039 g, 0.39 mmol, 0.031 equiv).
Trimethylsilyl chloride (1.95 mL, 15.4 mmol, 1.2 equiv) was
added dropwise via a syringe and the reaction mixture cooled
to 0 °C in an ice bath.
Compounds 7a -d , 7i, and 7j were similarly prepared from
6a and the appropriate ylide of the phosphonium salts as
described for the preparation of 7h . They were then hydro-
lyzed and transformed into 1a -d and 1f-h using the same
procedure as described for the conversion of 7e to 1e.
5,8,11,14-a ll-cis-Hen eicosatetr aen oyleth an olam ide (1a).
Obtained as an oil (76% from 7a ); NMR δ 0.88 (t, J ) 7.5 Hz,
3H), 1.3 (br s, 8H), 1.60-2.30.(m, 8H), 2.82 (br s, 7H), 3.30-
3.50 (m, 2H), 3.65-3.80 (m, 2H), 5.20-5.55 (m, 8H), 6.05 (br
s, 1H); MS m/z 362 (M + 1) (100), 322 (83), 282 (81); HRMS
n-Pentylmagnesium chloride (6.7 mL of a 2.0 M solution in
ether; 13.4 mmol, 1.05 equiv) was added dropwise via a
syringe, and the reaction was stirred, allowing it to warm to
room temperature over 3 h. The reaction was quenched by
4
the addition of sat. NH Cl solution (5 mL) and filtered, and
the product was extracted with ether. After drying, the ether
was evaporated to yield the product as an oil (2.83 g, 97%);
NMR δ 0.85-1.70 (m, 19H), 1.00 (s, 6H), 2.20 (s, 2H), 4.75-
calcd for C23
10.99 min (86%).
5,8,11,14-a ll-cis-Docosa tetr a en oyleth a n ola m id e (1b).
2
H40NO (M + 1), 362.3059; found 362.3077; GC,
t
R
4
.95 (m, 1H).
,3-Dim eth ylocta n ol (13a ). Lithium aluminum hydride
7.30 g, 192 mmol, 3.0 equiv) was placed in a dry flask cooled
Obtained as an oil (35% from 7b); NMR δ 0.88 (t, J )7.5 Hz,
3H), 1.29 (br s, 10H), 1.60-1.90 (m, 2H), 2.0-2.30 (m, 6H),
2.65-2.95 (m, 7H), 3.30-3.55 (m, 2H), 3.65-3.80 (m, 2H),
5.20-5.55 (m, 8H), 5.9 (br s, 1H); MS m/z 376 (M + 1) (100),
3
(
in an ice bath. Dry THF (80 mL) was added via a cannula
followed by the dropwise addition of a solution of 12 in THF
336 (40), 296 (23); HRMS calcd for C24
H
42NO
2
(M + 1),
(40 mL, 64 mmol). The reaction mixture was allowed to stir
376.3215; found 376.3200; GC, t 11.24 min (57%).
R
overnight, and after cooling it was quenched cautiously by the
sequential addition of water (16 mL), 15% sodium hydroxide
solution (16 mL), and water (32 mL). After stirring for 1 h,
the mixture was filtered, and the precipitate was triturated
with warm THF. The THF washings were combined with the
filtrate and then concentrated. The product was partitioned
between brine and ether, and after separating the layers, the
organic layer was dried. After evaporation the product was
isolated as an oil (9.77 g, 96%); NMR δ 0.90 (s, 6H), 0.90 (t, J
5,8,11,14-a ll-cis-Tr icosa tetr a en oyleth a n ola m id e (1c).
Obtained as an oil (57% from 7c); NMR δ 0.89 (t, J ) 7.5 Hz,
3H), 1.29 (br s, 12H), 1.60-1.90 (m, 2H), 2.05-3.50 (m, 6H),
2.65-2.95 (m, 7H), 3.30-3.50 (m, 2H), 3.60-3.85 (m, 2H),
5.20-5.55 (m, 8H), 5.95 (br s, 1H); MS m/z 390 (M + 1) (100),
372 (10); HRMS calcd for C25
2
H44NO (M + 1), 390.3372; found
390.3339; GC, t 11.70 min (66%).
R
5,8,11,14-a ll-cis-Tetr a cosa n oyleth a n ola m id e (1d ). Ob-
tained as an oil (66% from 7d ); NMR δ 0.88 (t, J ) 7.5 Hz,
3H), 1.28 (br s, 14 H), 1.60-1.90 (m, 2H), 2.05-2.30 (m, 6H),
2.75-2.95 (m, 7H), 3.35-3.55 (m, 2H), 3.65-3.80 (m, 2H),
5.20-5.55 (m, 8H), 5.85 (br s, 1H); MS m/z 404 (M + 1) (100),
)
5.0 Hz, 3H), 1.20 (br s, 8H), 1.40-1.60 (m, 3H), 3.70 (t, J )
.5 Hz, 2H).
,3-Dim eth yl-1-iod oocta n e (13b).25 Compound 13a (9.77
g, 61.7 mmol) was dissolved in CH Cl (150 mL), cooled to 0
C, and treated dropwise with an excess of triethylamine (1.2
mL, 92.6 mmol). A solution of methanesulfonyl chloride (5.27
mL, 68.0 mmol, 1.1 equiv) in CH Cl (150 mL) was added via
8
3
2
2
364 (70), 324 (57); HRMS calcd for C26
H
46NO
2
(M + 1),
°
404.3528; found 404.3523; GC, t 12.10 min (87%).
R
17,17-Dim eth yl-5,8,11,14-a ll-cis-d ocosa tetr a en oyleth a -
n ola m id e (1f). Obtained as an oil (61% from 7h ); NMR δ
0.88 (s, 6H), 0.90 (t, J ) 7.5 Hz, 3H), 1.20 (br s, 8H), 1.60-
2.30 (m, 8H), 2.70-2.95 (m, 7H), 3.35-3.55 (m, 2H), 3.65-
3.80 (m, 2H), 5.20-5.60 (m, 8H), 6.05 (br s, 1H); MS m/z 404
(M + 1) (97), 364 (72), 324 (100), 294 (14); HRMS calcd for
2
2
a syringe, and the reaction mixture was stirred 1 h, allowing
it to warm to room temprature. The mesylate was isolated
by quenching with brine and extracting into CH Cl . It was
2 2
filtered and evaporated to an oil. The product was taken into
hexanes, filtered through Celite, and evaporated, and the
residue was treated with dry benzene and again evaporated
C
26
H
46NO
2
R
(M + 1), 404.3528; found 404.3498; GC, t 11.70
min (61%).

3
624 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 22
Ryan et al.
1
7,17-Dim eth yl-5,8,11,14-a ll-cis-h en eicosa n oyleth a n o-
sulfonyl fluoride (PMSF, 50 µM). After a 1-h incubation at
30 °C, the reaction was terminated with the addition of 2 mL
of ice-cold buffer (50 mM Tris-HCl, 1 mg/mL bovine serum
albumin) followed by rapid filtration through PEI-treated
filters. The assays were performed in triplicate, and the
results represent the combined data from three to six inde-
pendent experiments.
la m id e (1g). Obtained as an oil (40% from 7i); NMR δ 0.85
(
s, 6H), 0.90 (t, J ) 7.5 Hz, 3H), 1.25 (br s, 6H), 1.65-2.30 (m,
8
5
2
3
H), 2.65-2.95 (m, 7H), 3.30-3.50 (m, 2H), 3.65-3.80 (m, 2H),
.20-5.50 (m, 8H), 6.05 (br s, 1H); MS m/z 390 (M + 1) (98),
90 (42), 210 (27); HRMS calcd for C25
90.3372; found 390.3335; GC, t 11.42 min (79%).
7-Meth yl-5,8,11,14-a ll-cis-d ocosa tetr a en oyleth a n ola -
m id e (1h ). Obtained as an oil (64% from 7j); NMR δ 0.90
m, 6H), 1.28 (br s, 9H), 1.60-2.35 (m, 8H), 2.65-2.95 (m, 7H),
.30-3.50 (m, 2H), 3.65-3.85 (m, 2H), 5.20-5.60 (m, 8H), 5.95
br s, 1H); MS m/z 390 (M + 1) (38), 350 (91), 310 (100); HRMS
calcd for C25 (M + 1), 390.3372; found 390.3365; GC,
11.53 min (71%).
()-2,16,16-Tr im eth yl-5,8,11,14-a ll-cis-tr icosatetr aen oyl-
′-flu or oeth yla m id e (2b). The product was prepared ac-
cording to the method described for 2a : Yield 57%; NMR δ
H
44NO
2
(M + 1),
R
1
Beh a vior a l Eva lu a tion s. Mice were acclimated to the
laboratory overnight. Depression of locomotor activity and
antinociception, as determined by the tail-flick (TF) response
(
3
(
28
to a heat stimulus, were measured in the same animal.
Control tail-flick latencies of 2 to 4 s were measured for each
animal with a standard tail-flick apparatus prior to drug or
vehicle administration. Four minutes following an iv injection
of either vehicle or drug, mice were tested for tail-flick
response. Immediately thereafter, the mice were placed into
individual photocell activity cages (11 × 6.5 in.) for assessment
of spontaneous activity. For the next 10 min the total number
of beam interruptions in the 16 photocell beams per cage were
recorded using a Digiscan Animal Activity Monitor (Omnitech
Electronics Inc., Columbus, OH). Spontaneous activity was
expressed as percent of control activity. Antinociception was
expressed as the % maximum possible effect (MPE) using a
H
44NO
2
t
R
(
2
0
1
(
.88 (t, J ) 6.0 Hz, 3H), 1.09 (s, 6H), 1.16 (d, J ) 6.8 Hz, 3H),
.25 (br s, 10H), 1.39-2.40 (m, 5H), 2.81-2.97 (m, 6H), 3.57
ddt, J ) 28.5, 5.2, 4.5 Hz, 2H), 4.49 (dt, J ) 47.4, 4.6 Hz,
H), 5.13-5.52 (m, 8H), 5.86 (br s, 1H). Anal. (C28
NOF‚0.4CHCl ) C, H, N. The presence of chloroform was
confirmed by NMR in DMSO-d
2
48
H -
3
6
.
(
()-2,16-Dim eth yl-5,8,11,14-a ll-cis-d ocosa tetr a en oyl-2′-
1
0-s maximum test latency as described earlier.28 Each dose
flu or oeth yla m id e (2c). The product was prepared according
to the method described for 2a : Yield 37%; NMR δ 0.88 (t, J
tested in the antinociception and hypomotility assays repre-
sents one group of animals (six mice per group). Cannabinoid-
induced hypothermia and immobility were determined in a
separate group of animals. Prior to vehicle or drug adminis-
tration, rectal temperature was determined by a thermistor
probe (inserted 25 mm) and a telethermometer (Yellow Springs
Instrument Co., Yellow Springs, OH). Four minutes after the
iv injection of the drug, mice were tested for body temperature.
The difference between pre- and postinjection rectal temper-
atures were calculated. Immediately after measure of body
temperature, the mice were placed on a 5.5 cm ring attached
at a height of 16 cm to a ring stand, and amount of time the
animals remained motionless during a 5-min period was
recorded.29 The time that each animal remained motionless
on the ring was divided by 300 s and multiplied by 100 to
obtain a % immobility rating.
)
6.0 Hz, 3H), 0.94 (d, J ) 6.2 Hz, 2H), 1.16 (d, J ) 6.8 Hz,
3
3
2
H), 1.25 (br s, 10H), 1.36-2.60 (m, 6H), 2.70-2.90 (m, 6H),
.60 (ddt, J ) 28.5, 5.1, 4.6 Hz, 2H), 4.49 (dt, J ) 47.4, 4.8 Hz,
H), 5.15-5.60 (m, 8H), 5.81 (br s, 1H); HRMS calcd for C26
45
H -
FNO (M + 1), 406.3485; found 406.3480; GC, t
()-2,17,17-Tr im eth yl-5,8,11,14-a ll-cis-docosatetr aen oyl-
′-flu or oeth yla m id e (2d ). It was obtained as an oil (72%
from 7h ); NMR δ 0.85 (s, 6H), 0.90 (t, J ) 6.0 Hz, 3H), 1.15
d, J ) 7.0 Hz, 3H), 1.25 (br s, 8H), 1.40-2.40 (m, 7H), 2.70-
.95 (m, 6H), 3.58 (ddt, J ) 28.5, 5.2, 4.5 Hz, 2H), 4.50 (dt, J
R
9.79 min (73%).
(
2
(
2
)
47.4, 4.6, Hz, 2H), 5.20-5.55 (m, 8H), 5.85 (br s, 1H); MS
m/z (M + 1) (100), 272 (37), 2328 (19); HRMS calcd for C27
FNO (M + 1), 420.3641; found 420.3608. Anal. (C27
H
H
47
-
-
46
2
NOF‚0.9H O) C, H, N.
(
()-2,17-Dim eth yl-5,8,11,14-a ll-cis-d ocosa tetr a en oyl-2′-
flu or oeth yla m id e (2e). It was obtained as an oil (69% from
j); NMR δ 0.80-0.98 (m, 6H), 1.15 (d, J ) 6.5 Hz, 3H), 1.25
br s, 8H), 1.38-2.40 (m, 8H), 2.65-3.0 (m, 6H), 3.60 (ddt, J
28.5, 5.2, 4.6 Hz, 2H), 4.50 (dt, J ) 47.5, 4.8 Hz, 2H), 5.20-
.55 (m, 8H), 5.80 (br s, 1H). Anal. (C26 44FNO‚0.15CHCl
C, H, N. The presence of chloroform was confirmed by NMR
in DMSO-d
values.30
Da ta An a lysis. IC50 values were converted to K
i
7
(
)
5
Statistical evaluation of parallelism between displacement
curves generated in the presence and absence of PMSF were
31
performed using ALLFIT. Dose-response relationships were
H
3
)
determined for each analog in the pharmacological assays.
Percent effect was determined based upon the maximal effects
6
.
9
that are produced by ∆ -THC and anandamide which are 90,
(
()-2-Meth yl-5,8,11,14-a ll-cis-tr icosa tetr a en oyl-2′-flu o-
1
00, and 60% for spontaneous activity, antinociception and
r oeth yla m id e (2f). It was obtained as an oil (72% from 7c);
NMR δ 0.88 (t, J ) 6.0 Hz, 3H), 1.13 (d, J ) 7.5 Hz, 3H), 1.25
ring immobility, respectively. The percent effect for hypoth-
ermia was based upon the maximal effect produced by anan-
(
br s, 12H), 1.38-2.40 (m, 7H), 2.70-2.95 (m, 6H), 3.60 (ddt,
9
damide (-3.0 °C) rather than that produced by ∆ -THC (-6.0
J ) 28.5, 5.2, 4.5 Hz, 2H), 4.50 (dt, J ) 47.5, 4.5 Hz, 2H), 5.20-
°C). Antinociception, hypomotility, and immobility data were
5
(
.60 (m, 8H), 5.80 (br s, 1H); MS m/z 406 (M + 1) (100), 326
converted to probit values, and ED50’s were calculated by
unweighted least-squares linear regression analysis of the log
dose versus the probit values. Several analogs were classified
as having partial agonist effects because they produced dose-
responsive effects that failed to exceed 60% of maximal effect.
Analogs producing effects less than 30% or hypothermia less
than 1 °C at the highest doses tested were considered to be
inactive.
27), 119 (96); HRMS calcd for C26
H
45FNO (M + 1), 406.3485;
found 406.3459; GC, t
R
6.10 min (81%).
P h a r m a cology. Dr u g P r ep a r a tion a n d Ad m in istr a -
tion . For binding assays, compounds were prepared as 1 mg/
mL stock solutions in absolute ethanol and were stored at -20
C. For behavioral assays, drugs were dissolved in a 1:1:18
mixture of ethanol, emulphor (GAF Corporation, Linden, NJ ),
and saline (0.9% NaCl) and were administered intravenously
°
(
iv) in the mouse tail vein in volumes of 0.1 mL/10 g of body
weight.
Bin d in g Assa ys. Radioligand binding to P
Ack n ow led gm en t. Acknowledgment is made to
NIDA grants DA-08904, DA-09789, and DA-03672 for
the support of this work.
2
membrane
1
0
preparations were performed as described elsewhere. Etha-
nolic 1 mg/mL stock solutions of anandamide analogs were
2
diluted in buffer (50 mM Tris-HCl, 1 mM EDTA, 3 mM MgCl ,
Refer en ces
and 5 mg/mL bovine serum albumin) without evaporation of
the ethanol (final concentration not exceeding 0.4%). Analog
concentrations ranging from 1 nM to 10 µM and a 1 nM
(1) Devane, W. A.; Hannus, L.; Breuer, A.; Pertwee, R. G.; Steven-
son, L. A.; Griffin, G.; Gibson, D.; Mandelbaum, A.; Etinger, A.;
Mechoulam, R. Isolation and Structure of a Brain Constituent
that Binds to the Cannabinoid Receptor. Science 1992, 258,
3
concentration of [ H]CP-55,940 were used. Binding was
initiated with the addition of 150 µg of membrane preparation
protein. Nonspecific binding was determined in the presence
of 1 µM CP-55,940. The binding experiments were performed
either with or without the amidase inhibitor phenylmethyl-
1
946-1949.
(
2) Fride, E.; Mechoulam, R. Pharmacological activity of the can-
nabinoid receptor agonist, anandamide, a brain constituent. Eur.
J . Pharmacol. 1993, 231, 313-314.

Potent Anandamide Analogs
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 22 3625
(
3) Felder, C. C.; Briley, E. M.; Axelrod, J .; Simpson, J . T.; Mackie,
K.; Devane, W. A. Anandamide, an Endogenous Cannabimimetic
Eicosanoid, Binds to the Cloned Human Cannabinoid Receptor
and Stimulates Receptor-mediated Signal Transduction. Proc.
Natl. Acad. Sci. 1993, 90, 7656-7660.
(18) Manna, S.; Falck, J . R.; Chacos, N.; Capdevila, J . Synthesis of
Arachidonic Acid Metabolites Produced by Purified Kidney
Cortex Microsomal Cytochrome P-450. Tetrahedron Lett. 1983,
2
4, 33-36.
(
19) Corey, E. J .; Niwa, H.; Falck, J . R. Selective Epoxidation of
Eicosa-cis-5,8,11,14-tetraenoic (Arachidonic) Acid and Eicosa-
cis-8,11,14-trienoic Acid. J . Am. Chem. Soc. 1979, 101, 1586-
(
4) Vogel, Z.; Barg, J .; Levy, R.; Saya, D.; Heldman, E.; Mechoulam,
R. Anandamide, a brain endogenous compound, interacts specif-
ically with cannabinoid receptors and inhibits adenylate cyclase.
J . Neurochem. 1993, 61, 352-355.
1
587.
(
20) Perrier, H.; Prasit, P.; Wang, Z. Synthesis of a Radioactive
Photoaffinity Arachidonic Acid Analog. Tetrahedron Lett. 1994,
35, 1501-1502.
(5) Mechoulam, R.; Hanus, L.; Martin, B. R. Search for endogenous
ligands of the cannabinoid receptor. Biochem. Pharm. 1994, 48,
1
537-1544.
(
6) Smith, P. B.; Compton, D. R.; Welch, S. P.; Razdan, R. K.;
Mechoulam, R.; Martin, B. R. The Pharmacological Activity of
Anandamide, a Putative Endogenous Cannabinoid, in Mice. J .
Pharmacol. Exp. Ther. 1994, 270, 219-227.
(21) McDonald, R. N.; Campbell, T. W. 1,4-Diphenyl-1,3-Butadiene.
Organic Syntheses; Wiley: New York, 1973; Coll. Vol. 5, pp 499-
501.
(22) Labelle, M.; Falgueyret, J . P.; Riendeau, D.; Rokach, J . Synthesis
of Two Analogues of Arachidonic Acid and their Reactions with
(
7) Pinto, J . C.; Potie, F.; Rice, K. C.; Boring, D.; J ohnson, M. R.;
Evans, D. M.; Wilken, G.; Cantrell, C. H.; Howlett, A. C.
Cannabinoid Receptor Binding and Agonist Activity of Amides
and Esters of Arachidonic Acid. Mol. Pharmacol. 1994, 46, 516-
1
2-Lipoxygenase. Tetrahedron 1990, 46, 6301-6310.
(
23) Munch-Petersen, J . 3-Methylheptanoic Acid. Organic Syntheses;
Wiley: New York, 1973; Coll. Vol. 5, pp 762-766.
5
22.
(
24) Cahiez, G.; Alami, M. Copper-Catalyzed Conjugate Addition of
Organomagnesium Reagents to R,â-Ethylenic Esters: A Simple
high yield Procedure. Tetrahedron Lett. 1990, 31, 7425-7428.
25) Crossland, R. K.; Servis, K. L. A Facile Synthesis of Methane-
sulfonate Esters. J . Org. Chem. 1970, 35, 3195-3196.
(
8) Abadji, V.; Lin, S.; Taha, G.; Griffin, G.; Stevenson, L. A.;
Pertwee, R. G.; Makriyannis, A. (R)-Methanandamide: A Chiral
Novel Anandamide Possessing Higher Potency and Metabolic
Stability. J . Med. Chem. 1994, 37, 1889-1893.
(
(
9) Adams, I.; Ryan, W.; Singer, M.; Razdan, R. K.; Compton, D.
R.; Martin, B. R. Pharmacological and Behavioral Evaluation
of Alkylated Anandamide Analogs. Life Sci. 1995, 56, 2041-
(26) Corey, E. J .; Suggs, J . W. Pyridinium Chlorochromate. An
efficient reagent for oxidation of primary and secondary alcohols
to carbonyl compounds. Tetrahedron Lett. 1975, 2647-2650.
(27) Martin, B. R.; Compton, D. R.; Prescott, W. R.; Barret, R. L.;
Razdan, R. K. Pharmacological Evaluation of Dimethylheptyl
2
048.
(
10) Adams, I. B.; Ryan, W.; Singer, M.; Thomas, B. F.; Compton, D.
R.; Razdan, R. K.; Martin, B. R. Evaluation of Cannabinoid
Receptor binding and in vivo Activities of Anandamide Analogs.
J . Pharmacol. Exp. Ther. 1995, 273, 1172-1181.
11) (a) Khanolkar, A. D.; Abadji, V.; Lin, S.; Hill, W. A. G.; Taha,
G.; Abouzid, K.; Meng, Z.; Fan, P.; Makriyannis, A. Head Group
Analogs of Arachidonylethanolamide, the Endogenous Cannab-
inoid Ligand. J . Med. Chem. 1996, 39, 4515-4519. (b) Sheskin,
T.; Hanus, L.; Shager, J .; Vogel, J .; Mechoulam, R. Structural
Requirements for Binding of Anandamide-Type Compounds to
the Brain Cannabinoid Receptor. J . Med. Chem. 1997, 40, 659-
9
Analogs of ∆ -THC: Reassesment of the Putative Three Point
Cannabinoid-Receptor Interaction. Drug Alcohol Depend. 1995,
(
3
7, 231-240.
(
28) Dewey, W. L.; Harris, L. S.; Howews, J . F.; Nuite, J . A. The effect
of various neurohumoral modulators on the activity of morphine
and the narcotic antagonists in the tail-flick and phenylquinone
tests. J . Pharmacol. Exp. Ther. 1970, 175, 435-442.
(29) Pertwee, R. G. The ring test: A quantitative method for
assessing the cataleptic effect of cannabis in mice. Br. J .
Pharmacol. 1972, 46, 753-763.
6
67
(
(
(
(
12) Deutsch, D. G.; Chin, S. A. Enzymatic Synthesis and Degrada-
tion of Anandamide, A Cannabinoid Receptor Agonist. Biochem.
Pharm. 1993, 46, 791-796.
13) Childers, S. R.; Sexton, T.; Roy, M. B. Effects of anandamide on
cannabinoid receptors in rat brain membranes. Biochem. Phar-
macol. 1994, 47, 711-715.
14) Hanus, L.; Gopher, A.; Almog, S.; Mechoulam, R. Two New
Unsaturated Fatty Acid Ethanolamides in Brain that Bind to
the Cannabinoid Receptor. J . Med. Chem. 1993, 36, 3032-3034.
15) Thomas, B. F.; Adams, I. B.; Mascarella, S. W.; Martin, B. R.;
Razdan, R. K. Structure-Activity Analysis of Anandamide
Analogs: Relationship to a Cannabinoid Pharmacophore. J . Med.
Chem. 1996, 39, 471-479.
(30) Cheng, Y. C.; Prusoff, W. H. Relationship between the inhibtion
constant (K ) and the concentration of inhibitor which causes
I
5
0 percent inhibition (I50) on an enzymatic reaction. Biochem.
Pharmacol. 1973, 22, 3099-3108.
(
31) DeLean, A.; Munson, P. J .; Rodbard, D. Simultaneous analysis
of families of sigmoidal curves: Application to bioassay, radio-
ligand assay, and physiological dose-response curves. Am. J .
Physiol. 1978, 235, E97-E102.
(
32) Compton, D. R.; J ohnson, M. R.; Melvin, L. S.; Martin, B. R.
Pharmacological profile of a series of bicyclic cannabinoid
analogs: Classification as cannabimimetic agents. J . Pharmacol.
Exp. Ther. 1992, 260, 201-209.
(
16) Razdan, R. K. Structure-Activity Relationships in Cannabinoids.
Pharmacol. Rev. 1986, 38, 21-43.
17) Corey, E. J .; Iguchi, S.; Albright, J . O.; De, B. Studies on the
Conformational Mobility of Arachidonic Acid. Facile Macrolac-
tonization of 20-Hydroxyarachidonic Acid. Tetrahedron Lett.
(
1
983, 24, 37-40.
J M970212F