Total synthesis of photoactivatable or fluorescent anandamide probes: Novel bioactive compounds with angiogenic activity

Article abstract of DOI:10.1021/jm8011382

Endocannabinoids are endogenous polyunsaturated fatty acids involved in a multitude of health and disease processes. Recently, several lines of evidence suggest the presence of a novel non-CB1/CB2 anandamide receptor in endothelial cells. Thus, we synthes

Full text of DOI:10.1021/jm8011382

J. Med. Chem. 2009, 52, 1005–1017
1005
Total Synthesis of Photoactivatable or Fluorescent Anandamide Probes: Novel Bioactive
Compounds with Angiogenic Activity
Laurence Balas,*,† Thierry Durand,† Sattyabrata Saha,‡ Inneke Johnson,‡ and Somnath Mukhopadhyay‡,§
Institut des Biomole´cules Max Mousseron (IBMM), UMR 5247 CNRS, UniVersite´ de Montpellier I and UniVersite´ de Montpellier II,
15, AV. Ch. Flahault, F-34093 Montpellier Cedex 05, France, Neuroscience Research Program, JLC-Biomedical Biotechnology Research
Institute, North Carolina Central UniVersity, 700 George Street, Durham, North Carolina 27707, and Department of Chemistry, DiVision of
Biochemistry, North Carolina Central UniVersity, Durham, North Carolina 27707
ReceiVed September 11, 2008
Endocannabinoids are endogenous polyunsaturated fatty acids involved in a multitude of health and disease
processes. Recently, several lines of evidence suggest the presence of a novel non-CB1/CB2 anandamide
receptor in endothelial cells. Thus, we synthesized two types of photoaffinity probes that contain either an
arylazide group or a diazirin moiety, together with a fluorescent analogue. The key steps rely on selective
hydrogenation of skipped tetrayne backbones and on copper-mediated cross-coupling reactions between
diynic precursors. Three synthetic routes were investigated. In biological functional assays, we found that
both the arylazide and the fluorescent probes induced robust increases in matrix metalloprotease activity
and produced positive angiogenic responses in in vitro endothelial cell tube formation assays. Irradiation of
the arylazide probe nicely enhanced this effect in both HUVEC and CB1-KO HUVEC. These results suggest
that the arylazide and the fluorescent probes can be used to identify “non-CB1/CB2 anandamide receptor”
from endothelial cells.
Introduction
deficits, anorexia) as well as cardiovascular disease states
(hypertension, atherosclerosis), AIDS, and intestinal disorders.
It is well established that the eCBs bind to the same GPCR
receptors (i.e., CB1 and CB2)10 as phytocannabinoid drugs such
as ∆9-tetrahydrocannabinol (THC), the principal psychoactive
agent in the Cannabis satiVa plant. Depending on their structures
and localization, the eCBs may exert very different effects on
CB1 and CB2 receptors, acting11 as partial agonists, full
agonists, or antagonists. Moreover, these receptors have a
widespread distribution in various tissues and organ systems.12
Several excellent reviews on the biochemistry and pharmacology
of eCBs have appeared in recent years12-17 Nevertheless, many
questions remain unanswered. In fact, the more the eCB system
is studied, the more complex it appears,18 leading some authors19
to call it “the eCB soup”.
Endocannabinoids (eCBsa) are a family of endogenous
polyunsaturated long-chain fatty acid (PUFA) derivatives,1
regarded as promising templates for the development of new
pharmacological agents potentially useful for various conditions.
The first eCB to be discovered2 was the ethanolamide of
arachidonic acid (also called anandamide). The seven eCBs
isolated to date are shown in Figure 1, together with four
structurally related endogenous lipoamino acids (also called
elmiric acids3), i.e., NAala, NAgly, NAGABA, ARA-S, and a
taurine derivative NAT.4-6 All of these newly isolated PUFA
derivatives are known to exist in brain tissue and appear to
exhibit properties similar to anandamide. However, little or no
information is currently available concerning their natural
occurrence or biological roles.
Interestingly, in transgenic CB1 knockout and CB1/CB2
double knockout mice, anandamide produces biological re-
sponses that must be attributed to alternative pharmacological
targets other than the identified cannabinoid receptors. Multiple
lines of evidence suggest the existence of several types of non-
CB1/CB2 receptors in different tissues.20-25
To date, none of these putative receptors have been isolated
or cloned. However, recent findings from our laboratories have
shown that anandamide analogue methanandamide acting on
this non-CB1/CB2 anandamide receptor can activate angiogenic
responses in endothelial cells.26,27 In those studies we have also
illustrated that methanandamide acting on non-CB1/CB2 anan-
damide receptors stimulates eNOS protein to produce nitric
oxide in endothelial cells.28,26
The therapeutic potential of agents that influence the eCB
system is impressive. Abundant evidence indicates that eCBs
participate as neuromodulators in numerous physiological and/
or pathological processes.7-9 Besides their involvement in
chronic pain, other conditions where abnormal activity of the
eCB system has been implicated include neurological/psycho-
logical disorders (multiple sclerosis, Alzheimer’s disease, Par-
kinsonism, schizophrenia, epilepsy, learning and memory
* To whom correspondence should be addressed. Phone: +33-4-67-54-
86-24. Fax: 33-4-67-54-86-25. E-mail: balas@univ-montp1.fr.
†
Universite´ de Montpellier.
JLC-Biomedical Biotechnology Research Institute, North Carolina
‡
Central University.
§
Division of Biochemistry, North Carolina Central University.
a Abbreviations: AcOH, acetic acid; AIDS, acquired immune deficiency
syndrome; bFGF, fibroblast growth factor; BSA, bovine serum albumin;
CB, cannabinoid; DMF, dimethylformamide; eCB, endocannabinoid; EDTA,
ethylenediaminetetraacetic acid; FBS, fetal bovine serum; GPCR, G-protein-
coupled receptor; HFIP, hexafluoroisopropanol; HUVEC, human umbilical
vein endothelial cells; KO, knockout; MMP, matrix metalloproteinase;
PUFA, polyunsaturated fatty acid; SN, nucleophilic substitution; TBAF,
tetrabutylammonium fluoride; TBDPS, tert-butyldiphenylsilyl; TBMB,
2-tert-butyl 2-methyl-1,3-benzodioxole-4-carboxylate; THC, ∆9-tetrahy-
drocannabinol;THF,tetrahydrofuran;TIMP,tissueinhibitorofmetalloprotease.
Clearly, pharmacological studies of the eCB system would
benefit from the availability of tools that would assist in the
detection, isolation, and characterization of these putative
receptors. Fluorescent29 and photolabeling30 techniques are
widely used for mechanistic studies of receptors and their
binding properties and might provide interesting information
on the eCB system. Fluorescence microscopy can reveal the
exact location of the target protein. Photoaffinity labeling of
10.1021/jm8011382 CCC: $40.75
2009 American Chemical Society
Published on Web 01/22/2009

1006 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4
Balas et al.
Figure 1. Endogenous ligands discovered to date.
Figure 2. Targeted compounds.
receptors is particularly useful for the isolation of putative
receptors. In this case, the receptor is tagged with an irreversi-
ble ligand, formed when the ligand is converted to a highly
reactive species, e.g., a nitrene or carbene31 by photoactivation.
These reactive “suicide” agonists or antagonists insert covalently
into C-H or O-H bonds of the closest amino acids in, or near,
the binding pocket of the putative receptor protein.
We synthesized a replica of the anandamide structure with
C-methylation, as a relatively minor structural modification of
the polar headgroup and the inclusion of a photoreactive probe
on the tail region of the analogue 1 (Figure 2). It is noteworthy
that one of the shortcomings of anandamide as an effective
pharmacological tool is its rapid in vivo and in vitro degradation.
Methylation at the R-position of the nitrogen atom is known to
result in a compound more resistant to enzymatic degradation,
i.e., the analogue R-(+)-methanandamide.32 Concerned with
possible steric hindrance33 and interference with the “horseshoe”
conformation of the PUFA backbone, the pentyl chain was
shortened to an ethylene group. In our preliminary paper,34 we
reported the synthesis of a new anandamide arylazide probe.
Herein, we report the details of its synthesis as well as those of
three novel probes 1 that bear either a diazirin photoactivatable
or a fluorescent group at the terminus of the anandamide
lipophilic chain.
strategy that allows the facile modification of head and/or tail
groups at a later stage (analogues 2).
Results and Discussion
Our retrosynthetic strategy (Scheme 1) relies on the esteri-
fication of the hydroxy function on the tail of the key tetraene
intermediate 3b with photoreactive carboxylic acids. The four
skipped cis-double bonds come from selective hydrogenation
of a tetrayne derivative 4. Instead of the previously described
linear strategy35,36 that built consecutively the diyne, triyne and
then tetrayne backbone, we developed a convergent route to
tetrayne 4 by coupling the two skipped diynes 5 and 6. This
has the advantage, over the linear strategy, of requiring fewer
experimental steps with these unstable structures. Each skipped
diyne comes from organometallic cross-coupling reactions
between commercially available terminal alkynes and propar-
gylic halides.
As outlined in Scheme 1, three routes were investigated: the
eCB (and elmiric acid) headgroup may be introduced after
sensitive elaboration of the tetraene backbone (pathway A) or
from the very beginning of the synthesis (pathway B). Therefore,
tetraene 3a (pathway A) will serve as a pivotal intermediate to
obtain analogues 2 with different eCB head groups and different
photoreactive tail groups whereas pathway B focuses on one
eCB series only (herein, anandamide analogues 1).
We further tested the ability of these compounds to stimulate
angiogenesis as a biological function triggered by the activation
of non-CB1/CB2 anandamide receptors in endothelial cells to
verify the validity of these probes as specific non-CB1/CB2
anandamide receptors. Our design represents a flexible synthetic
Replacement of one alkyne function in the tetrayne key
intermediate 4 with a Z-alkene function was investigated
(pathway C) to confer greater chemical stability to the inter-
mediate 7.

Synthesis of Anandamide Probes
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4 1007
Scheme 1. Retrosynthetic Analysis
Scheme 2. Synthesis of Terminal Alkyne
Pathway A. The diyne precursor 5 was prepared in two steps
with excellent yield (95%, Scheme 2) by means of a copper-
mediated cross-coupling reaction between the commercially
available 3-butyn-1-ol 9 and 3-bromo-1-(trimethylsilyl)-1-
propyne 10, followed by deprotection of the masked terminal
alkyne. Upon flash chromatography (required to remove salts,
starting materials, and remaining DMF), the silylated diyne 11
is recovered as a mixture with its deprotected analogue 5 (about
5-10%). The partial deprotection at this stage was not
considered a problem, since the next step resulted in complete
conversion of 11 to 5. This last step was initially problematic,
since the terminal diyne 5 readily rearranges to its corresponding
allene 12 on silica or Florisil solid phase columns and under
basic conditions. Thus, standard C-Si deprotection with fluoride
ions, aqueous alcoholic potassium hydroxide, or silver nitrate
mainly led to the allene structure 12 (Scheme 2). Addition of
acetic acid simultaneously to the tetrabutylammonium fluoride
in order to reach complete neutrality was required in order to
obtain pure terminal diyne 5 in high yield (95-99%).
Starting from the commercially available 5-hexynoic acid 13,
the bis-propargylic bromide 6a was obtained in three steps
(Scheme 3) according to the procedure reported by Razdan.37
As depicted in Scheme 3, a subsequent cross-coupling
reaction between the two precursors 5 and 6a using Caruso
conditions38 afforded the tetrayne 4a. It is advisable to carry
out the transformation of tetrayne 4a to tetraene 3a on the same
experimental day, as 4a deteriorates on standing. Attempts to
further repurify the sample will then be pointless. The selective
hydrogenation step was performed under the Ni-P2 conditions
with only moderate yield, reflecting the poor stability of the
starting tetrayne 4a. After protection of the homoallylic alcohol
with a dimethoxytrityl group and saponification with lithium
hydroxide in methanol, the resulting carboxylic acid 15 was
coupled with di-O-tert-butyldiphenylsilyl-L-alaninol 16 to give
the amide 17 using the anhydride activation procedure. We
suggest that ethanolamine 16 can be replaced with any primary
amine function to incorporate the other eCB or elmiric acid head
groups, provided their functionalities are suitably protected.
Subsequently, removal of the dimethoxytrityl group using the
acidic properties of hexafluoroisopropanol (HFIP)39 afforded the
expected key alcohol 3b in good yields.
Pathway B. As explained above, the aim of our program is
to investigate the existence of new anandamide receptors. Thus,
we initially focused our attention on the synthesis of various
anandamide analogues. To this end, the ethanolamide headgroup
was introduced at the very beginning of the synthesis, which
avoided the homoallylic alcohol protection-deprotection steps.
The shortened strategy is depicted in Scheme 4. Thus, the
skipped tetrayne backbone 4b was reached in four steps only,
consisting of amide bond formation between the amine 16 and
5-hexynoic acid 13, followed by a copper cross-coupling
reaction with propargylic chloride40 19, bromination, and a
copper cross-coupling reaction between the resultant bis-
propargylic bromide 6b and the terminal diyne 5. As observed

1008 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4
Scheme 3. Synthesis of Tetraene 3b via Tetraene 3a
Balas et al.
Scheme 4. Shorter Synthesis of Tetraene 3b
for the tetrayne ester 4a, the tetrayneamide 4b is highly unstable
and was subsequently hydrogenated to its corresponding tetraene
3b using Ni-P2 catalyst. The rate of hydrogenation of each
alkyne functionality is not the same, with that closest to the
carbonyl group being less reactive. Consequently, depending
on the conditions (extent of nickel catalyst poisoning and
reaction time), the reaction led to a mixture of the desired
tetraene 3b, partially reduced alkenes, together with the trieneyne
intermediate 21. Conditions41 were optimized in order to obtain
pure tetraene 3b. This optimization was important, since
separation of 3b and 21 was difficult because of their similar
chromatographic mobilities. Unfortunately, we could not in-
crease the yield, which remained moderate (40% to 56%) mainly
because of the instability of the starting tetrayne structure 4b.
Pathway C. The synthetic pathways A and B described above
are based on the synthesis of a tetrayne backbone 4. However,
these tetrayne intermediates 4 were highly unstable and difficult
to work with. Therefore, a strategy encompassing a partially
reduced alkyne-alkene backbone 7 was developed in order to
confer stability to the precursors prior to the hydrogenation steps
(Scheme 1).
Synthesis of the yne-ene-yne derivative 7 via a one-pot
linchpin approach using 1,4-dichoro-2-cis-butene 22 was ap-
pealing. However, no information was found in the literature
with regard to monoselectivity of the copper-mediated cross-
coupling reaction between terminal alkynes and 1,4-dichoro-
2-cis-butene 22 using the standard coupling conditions (NaI,
CuI, base, DMF). A pilot experiment was undertaken using
methyl 5-hexynoate37 23 and 3-butyne-1-ol 9 in place of the
actual desired reactants 18 and 5, respectively. The reaction was
carried out with a stoichiometric amount of alkynes. In addition,
a reaction was run with 2.0 equiv of terminal alkyne 23 in order
to obtain a sample of the double substitution product for
comparison in TLC and NMR analyses. However, in both cases,
dichorobutene 22 had exhibited unexpected reactivity at the
olefin carbon rather than the methylene chloride group, affording
the chloride 24 as main product (81% yield, Scheme 5). It must
be noted that this reaction was highly regioselective, with only
one regioisomer (compound 24a) formed. Unfortunately, as this
was not the desired regioisomer 24b, it was felt that no more
progress could be made using this organometallic reaction.

Synthesis of Anandamide Probes
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4 1009
Scheme 5. Regioselectivity of the Reaction of 1,4-cis-Dichlorobutene 22 with Terminal Alkyne 23
Scheme 6. Obtention of the Diamide 25
Scheme 7. Undesired Regioselectivity of the Cross-Coupling Reaction
Alternative coupling reactions with dichloride 22 were then
investigated under Taber coupling conditions. Using organo-
magnesium reagents with copper salts as catalysts, Taber was
able to achieve42 monoselectivity, good yields, and high
regioselectivity in favor of a desired linear yne-ene-CH2Cl
substructure by carrying out the reaction at a temperature of 55
°C. As our alkyne precursor 18 is a secondary amide, an amount
of 2 equiv of EtMgBr was added in order to counteract the
reactivity of the amide structure. However, the amide function
dramatically suffered from these organomagnesium conditions.
In addition, both regioisomers were obtained as a 1:1 mixture
(10%), the main isolated product being a symmetric disubstituted
cis-butene 25 (Scheme 6, 27%).
under the previous standard conditions. The ratio of the two
regioisomers 27 and 28 was greatly improved, i.e., a 4-fold
greater production of the desired product 27.
When we initiated our studies of pathway C, we were rather
optimistic44 because the literature suggested that copper cross-
coupling reactions were not associated with any regioselectivity
problems.45,35 The poor yields they obtained could be explained
on the basis of the predictable instability of the polyskipped
backbone. On the other hand, these SN/SN′ regioselectivity
problems have been encountered by some authors, using
somewhat different copper reaction conditions.46
Since pathways A and B were successful, we did not persevere
in our efforts to obtain the yne-ene-yne structure 7.
Subsequently, a strategy based on the precursor 4-chloro-
cis-buten-1-ol43 26 was examined. Unfortunately, poor regi-
oselectivity was again observed, leading to an inseparable 1:1
mixture of both regioisomers 27 and 28 (Scheme 7).
In view of these coupling difficulties, we considered the
possibility that the reactivity at the olefin carbon could be
significantly decreased by substituting the methylene chloride
with a more electrophilic iodine atom, so the reaction could
be more directed. This conversion was undertaken following
the Finkelstein reaction (yield not optimized). The resulting
4-iodo-cis-buten-1-ol 29 was then coupled to the alkyne 18
Synthesis of Probes 35, 37, 39, and 41. As reported in our
preliminary study,34 introduction of an arylazide group at the
terminal pentyl chain of methanandamide does not disturb the
CB2 binding affinity while its CB1 binding is somewhat
affected. There is no way to know how this structural change
would affect the interaction with the putative anandamide
receptor. Binding or activation of the receptor could be disturbed
because of the bulkiness of the aromatic moiety. Accordingly,
we elected to use neighboring photophore groups no larger (and
preferably smaller) than the 2-azido-5-iodobenzoate group. In
addition, we believe that introduction of polar functionalities

1010 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4
Balas et al.
Scheme 8. Synthesis of Probes 35, 37, 39, and 41 from the Key Tetraene 3b
should be avoided, since the anandamide pentyl chain probably
fits in a hydrophobic region of the receptor active site.
These two requirements drastically reduce the possibilities
for the design of fluorescent probes, since most of the well-
known fluorescent groups47 (e.g., Alexa, BODIPY) are bulky
and polar. We elected the 2-tert-butyl-2-methyl-1,3-benzodiox-
ole-4-carboxylate (TBMB)48 group. This particular structure,
although not widely used for receptor-ligand studies, was
attractive because of its relatively small size and reported
fluorescence properties.
receptor in comparison with anandamide. Replacement of the
pentyl chain tail of anandamide with the fluorescent TBMB
group (a slightly bulkier group than the arylazide group) resulted
in the complete loss of CB1 and CB2 affinities. Effectively,
probe 37 exhibited a Ki (µM) of 2.80 ( 0.05 and 1.02 ( 0.03
for CB1 and CB2 receptors, respectively (mean ( SE, N )
3).51 Under the same conditions, anandamide shows a Ki (µM)
of 0.07 ( 0.03 and 0.18 ( 0.02, respectively.
Concerning the design of novel photoactivatable probes, we
turned our attention to short alkyl diazirins which are easier to
handle than their alkyl azide homologues. Two chain lengths
were tried, an approach expected to clarify SAR involving the
tail chain binding region, at least for the CB1 and CB2 receptors.
Subsequently, a series of ester-linked probes were syn-
thesized. Thus, Steglich esterification of the pivotal tetraene
alcohol 3b with photoreactive carboxylic acids such as
2-azido-5-iodobenzoic acid49 30, TBMB carboxylic acid48
31, and the C5 and C6 diazirin50 precursors 32 and 33,
followed by fluoride cleavage of the silyl group, afforded
the previously described probes 35 and then the three novel
probes 37, 39, 41 respectively (Scheme 8).
As part of preliminary studies to optimize the photoirradiation
conditions, UV spectra were recorded in methanol. Thus, upon
irradiation at 254 nm, the arylazide function (at 265 nm)
disappears within a few minutes (Figure 3). The diazirin
absorption is weaker in comparison to the skipped bis-allylic
system. The diazirin group was stable when irradiated at 254
nm (two shoulders at 348 and 367 nm, in agreement with
literature data), regardless of the distance from the lamp.
Furthermore, we found that upon irradiation of the diazirin 41
at 365 nm, the spectral scans obtained before, and various times
after, were virtually superimposable over the entire UV range
except in the diazirin absorption region (see expanded spectra
in Figure 4).
Figure 3. UV spectra of the arylazide 35 in methanol after 0, 30 s, 1
min, and 2 min of irradiation (lamp 254 nm, 15 W, d ) 5 cm).
Effect of Probes 35 and 37 on Angiogenesis. In our
preliminary biological experiments,34 the arylazide probe 35 was
shown to bind the CB2 receptor while it poorly binds the CB1
Figure 4. UV spectra of the diazirin 41 in methanol after 0, 10, 25,
and 40 min of irradiation (lamp 365 nm, 6 W, d ) 2.5 cm).

Synthesis of Anandamide Probes
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4 1011
the angiogenic response in tube formation assay (Figure 6B),
producing strong dose-dependent increases (21%, 30%, 47%
with 10 nM, 100 nM, and 1 µM, respectively, while they did
not exceed 8% before irradiation). The degree of angiogenic
response was statistically similar using CB1-KO HUVEC or
HUVEC.
Studies from Kunos laboratories showed28 that endothelial
non-CB1/CB2 anandamide receptor-mediated vasorelaxation
was inhibited by the selective non CB1/CB2 “anandamide
receptor” antagonist O-1918. In the current study we found that
probe 35-induced increase in tube formation in CB1-KO
HUVEC was significantly attenuated when cells were treated
with probe 35 (after irradiation) in the presence of the anan-
damide receptor antagonist O-1918 (1 µM, Figure 6C). HUVEC
does not express CB2 receptors (Mukhopadhyay et al., unpub-
lished observations), and blockade of probe 35-induced increase
in angiogenic response in CB1-KO HUVEC by O-1918 strongly
suggests that, in HUVEC, probe 35 produced angiogenesis by
acting on “non-CB1/CB2 anandamide” receptors and these
receptors are functionally active in HUVEC.
We tested the effect of probe 37 on HUVEC and CB1-
KO HUVEC cells before and after irradiation under similar
experimental conditions as probe 35. First of all, as described
for the probe 35, the probe 37 also produced robust
angiogenic response on HUVEC and CB1-KO HUVEC
(Figure 7), showing statistically similar levels of response
(p > 0.05). In addition, as expected because of its fluorescent
properties (different from the photolabeling properties such
as those of probe 35 or probe 41), probe 37 did not exhibit
any difference for angiogenic response before (data not
shown) and after irradiation (Figure 7).
Figure 5. Effect of probe 35 on angiogenesis in HUVEC. (A) HUVEC
cells were grown in 100 mm2 plate up to 90% confluence and then
harvested with versene (1× PBS-0.05% EDTA), and cell pellet was
obtained by gentle centrifugation at 600g for 5 min. The cell pellet
was resuspended in serum-free RPMI 1640 media and mixed with
vehicle or probe 35 at the indicated final concentration. The cell mixture
was then placed on Matrigel with or without UV irradiation as explained
in Experimental Section and incubated at 37 °C for 36 h. The increase
in cord length (tube formation) was assessed under microscope (10×),
and images were captured as described in Experimental Section. The
result shown here is a representative example of three similar results
obtained from three separate experiments. (B) Quantitation of angio-
genic response of probe 35 on HUVEC. After 36 h of incubation, the
plates were fixed with Diff-Quick and images were captured using
Nikon camera and analyzed using ImagePro software. Total length of
the tubes formed by the endothelial cells in Matrigel was measured
from three randomly chosen fields (10×) for each sample and analyzed
by the ImagePro software (version 6.0) program. The results are
analyzed as percent increase in cord length (tube formation) with respect
to the corresponding vehicle-treated control. The results are expressed
as the mean ( SEM of percent increase from three separate experiments.
Effects of Probes 35 and 37 on MMP Activity. MMP
(matrix metalloproteinase) plays an important role in the
regulation of angiogenesis in relation to the degradation of the
extracellular matrix so that endothelial cells can detach and
migrate. Since probes 35 and 37 produce strong angiogenic
responses in endothelial cells, we tested their effects on MMP
(gelatinase; MMP2 and MMP9) activity. As shown in Figures
8 and 9, similar to the effect observed in tube formation assay,
probes 35 and 37 produced dose-dependent increases in MMP
activity in HUVEC and CB1-KO HUVEC. However, probe 37-
induced increase in MMP activity (Figure 9) was stronger
compared to that produced by probe 35 (Figure 8). It is
important to note that methanandamide produced relatively less
increase in MMP activity (5%, 20%, and 28% increases with
respect to vehicle-treated control at 10 nM, 100 nM, and 1 µM,
respectively, in HUVEC; 10%, 25%, and 33% increases
with respect to vehicle-treated control at 10 nM, 100 nM, and
1 µM, respectively, in CB1-KO HUVEC) compared to increase
in MMP activity produced by probe 37.
As shown in Figure 5, probe 35 produces a dose-dependent
increase in angiogenesis as evident by the increase in the length
of tube formed by the human umbilical vein endothelial cells
(HUVEC) in Matrigel. Interestingly, upon irradiation probe 35
produced a dramatic increase of the angiogenic response
compared to the effect observed before irradiation (Figure 5B).
This dose-dependent increase in the angiogenic response fol-
lowing irradiation strongly indicates that irradiation produced
covalent binding of probe 35 to a specific site.
The involvement of “non-CB1/CB2 anandamide receptor” in
probe 35-induced angiogenesis was further confirmed from the
results with CB1 receptor knockout HUVEC cells (CB1-KO
HUVEC; Figure 6A shows that the CB1 receptor is reduced in
the siRNA treated cells). Here again, probe 35 induced some
nice tube length increases and irradiation dramatically stimulated
Further, as shown in Figure 10, in CB1-KO HUVEC cells
probe 35- and probe 37-induced increases in MMP activity were
significantly reduced by non-CB1/CB2 anandamide receptor
antagonist O-1918. This suggests that probe 35 and probe 37
simulate MMP activity by acting on non-CB1/CB2 anandamide
receptor.
MMP2 (also known as gelatinase A) is a constitutive enzyme
that exists in pro- (72 kDa) and active (64 kDa) forms, and
MMP9 is an inducible enzyme (92 kDa latent form). When
secreted by cells, these enzymes act on their substrate in
extracellular matrix proteins (gelatin for MMP2 and MMP9)
to evoke cellular response. The secretion and activation mech-
anisms of MMPs are multistep processes and are mediated by
the coordinated interactions of various factors including cell

1012 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4
Balas et al.
Figure 6. siRNA-mediated knocking down of CB1 receptor and effect of probe 35 on angiogenesis in CB1-KO HUVEC. (A) Western blot analysis
of CB1 receptor expression in HUVEC and CB1-KO HUVEC. Cells were transfected with siRNA against CB1R or nonsilencing siRNA as described
in Experimental Section. Thirty-six hours after transfection, cells were harvested and cell lysates were tested for CB1 receptor immunoreactivity
as described in Experimental Section. (B) HUVEC and CB1-KO HUVEC were seeded in 100 mm2 plates such that they grow to 90% confluence
within 36 h after siRNA transfection and then harvested with versene (1× PBS-0.05% EDTA), and a cell pellet was obtained by gentle centrifugation
at 600g for 5 min. The cell pellet was resuspended in serum-free RPMI 1640 media and mixed with vehicle or probe 35 at the indicated final
concentration. The cell mixture was then placed on Matrigel with or without UV irradiation as explained in Experimental Section and incubated at
37 °C for 36 h. The formation of the tube was assessed under microscope, and images were captured and analyzed as described above. The results
are expressed as the mean ( SEM percent increase in cord length (tube formation) with respect to the corresponding vehicle-treated control from
three separate experiments. (C) Effect of anandamide receptor antagonist O-1918 on probe 35-induced angiogenesis. CB1-KO HUVEC cells were
treated with probe 35 (1 µM) in the presence and absence of O-1918 (1 µM). The extent of the tube formation was assessed as described above.
Results are representative of three separate experiments with similar results.
surface MT1-MMP and tissue inhibitors of metalloproteases
(TIMP 2 for MMP2 and TIMP1 for MMP9) with latent forms
of MMPs (MMP2 and MMP9). Since probes 35 and 37
increased MMP activities, it is expected that activation of non-
CB1/CB2 anandamide receptors evokes signal transduction that
regulates TIMP-MMP interaction. We are currently testing the
effect of probe 35 and probe 37 on TIMP-MMP interaction in
endothelial cells. Since MMPs (MMP2 and MMP9) acting on
extracellular matrix proteins produce cell detachment and evoke
cell motility, we have also tested the effect of probes 35 and
37 on HUVEC cell motility. In cell migration assay probe 35
(following irradiation) produced dose-dependent 18%, 35%, and
54% increases with 10 nM, 100 nM, and 1 µM, respectively.
Under similar assay conditions, probe 37 produced a relatively
higher degree of increase in cell migration (27%, 40%, and 55%
increases in cell migration with 10 nM, 100 nM, and 1 µM,
respectively). Since increase in cell motility is an integral
component of the angiogenesis process, the ability of the
compounds to stimulate the cell migration strongly suggests that
probes 35 and 37 bear strong positive angiogenic properties in
endothelial cells.
In conclusion, we have accomplished a new convergent total
synthesis of an arachidonic carbon skeleton using copper-
mediated cross-coupling reactions between terminal alkynes and
propargylic halides and using selective hydrogenation of skipped
tetraynes 4. Several synthetic routes were investigated to reach
the pivotal tetraene 3b. The latter compound was successfully
transformed into four novel anandamide probes containing an
arylazide, a diazirin, or a fluorescent group. We plan to extend
this work by applying the same strategy to the preparation of
other probes 2 with different eCB and elmiric acid head groups.
These tools may be expected to advance significantly our
understanding of the molecular events underlying receptor
activation and signal transduction in the eCB system. The ability
of probes 35 and 37 (a) to dose-dependently increase capillary-
like tube formation in endothelial Matrigel assay in CB1-KO

Synthesis of Anandamide Probes
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4 1013
Figure 9. Effect of probe 37 on MMP activity. HUVEC or CB1-KO
HUVEC cells were grown in six-well plates to 90% confluence and
treated with vehicle or probe 37 at the indicated final concentration.
The cells were either directly placed in an incubator at 37 °C or
irradiated with UV (as described in the Experimental Section) and then
placed in incubator for 24 h. The MMP activity was measured from
conditioned media as described in the Experimental Section. The results
are analyzed as percent increase in enzyme activity with respect to the
corresponding vehicle treatment (basal activity) in the same cell line.
The results are expressed as the mean ( SEM of percent increase from
three separate experiments.
Figure 7. Effects of probe 37 on angiogenesis in HUVEC and CB1-
KO HUVEC. The cells were grown in 100 mm2 plates to 90%
confluence and then harvested with versene (1× PBS-0.05% EDTA),
and a cell pellet was obtained by gentle centrifugation at 600g for 5
min. The cell pellet was resuspended in serum-free RPMI 1640 media
and mixed with vehicle or probe 37 at the indicated final concentration.
The cell mixture was then placed on Matrigel as explained in
Experimental Section and incubated at 37 °C for 36 h. The formation
of the tube was assessed under microscope, and images were captured
and analyzed as described above.
bright-orange color appears quickly, characteristic of the dimethox-
ytrityl carbocation release. The completion of the reaction was
monitored by TLC (cyclohexane/AcOEt, 6/4, Rf ) 0.45). The
reaction was quenched by adding methanol (500 µL, discoloration)
at room temperature. The solvents were removed under reduced
pressure (without heating). The orange residue was dissolved in
cyclohexane/ethyl acetate, 80/20) and plugged through a pad of
Florisil to afford the deprotected homoallylic alcohol 3b in 65%
yield from acid 15.
Method B. Nickel(II) acetate tetrahydrate (293.62 mg, 1.18
mmol, 0.64 equiv) was placed in a 50 mL two-necked round-bottom
flask. Ethanol (16 mL) was added before the flask was placed under
vacuum and then hydrogen to ensure no oxygen was present. A 1
M solution of NaBH4 in ethanol (2.48 mL, 2.48 mmol, 1.35 equiv)
was added. The resulting black solution was evacuated 3 times and
then maintained under hydrogen atmosphere for 20 min. Ethylene-
diamine (878 µL, 13.14 mmol, 7.14 equiv) was added quickly via
syringe, and the reaction mixture was stirred to stabilize for 15
min. The flask was evacuated 3 times. The tetrayne 4b (1.04 g,
1.84 mmol, 1.0 equiv) was dissolved in ethanol (1.6 mL) and added
quickly to the reaction flask. The mixture was evacuated 3 times
and then maintained under hydrogen atmosphere for 8.5 h at 13
°C. The reaction mixture was quenched by addition of wet Et2O
(20 mL), followed by NH4Cl (a pinch) to poison the catalyst. The
reaction mixture was filtered through Celite which was washed with
Et2O. The solvents were removed under reduced pressure, and the
residue was partitioned between Et2O (100 mL) and sequentially
NaHSO4 (0.1 M, 5 mL), saturated aqueous NH4Cl solution (2 × 5
mL), and brine (2.5 mL). The aqueous phases were extracted 4
times (100 mL), and the resulting organic phases were dried over
MgSO4, filtered, and concentrated under reduced pressure. The
orange residue was chromatographed over Florisil (cyclohexane/
ethyl acetate:, 99/1 to 92/8) to give the expected tetraene 3b (588.16
mg, 56%) as a colorless oil. Rf ) 0.36 (cyclohexane/ethyl acetate,
6/4). IR (cm-1): 3574-3159 (OH/NH), 3012 (C-H), 2931 (C-H),
2857 (C-H), 1647 (NCdO), 1548 (CHdCH), 1428 (CHdCH),
Figure 8. Effect of probe 35 on MMP activity. HUVEC or CB1-KO
HUVEC cells were grown in six-well plates to 90% confluence and
treated with vehicle or probe 35 at the indicated final concentration.
The cells were either directly placed in an incubator at 37 °C or
irradiated with UV (as described in the Experimental Section) and then
placed in an incubator at 37 °C for 24 h. The MMP activity was
measured from conditioned media as described in the Experimental
Section. The results are analyzed as percent increase in enzyme activity
with respect to corresponding vehicle treatment (basal activity) in the
same cell line. The results are expressed as the mean ( SEM of percent
increase from three separate experiments.
HUVEC in an O-1918-sensitive manner, (b) to stimulate MMP
activity, and (c) to evoke an increase in cell migration strongly
indicate that these compounds are acting as an agonist on non-
CB1/CB2 anandamide receptors and the activation of the
receptor functionally coupled to angiogenic response in endot-
helial cells. We will use a modified fluorescent probe (probe
37) to develop a pharmacological ligand-binding assay as well
as to localize the non-CB1/CB2 anandamide receptor in
endothelial cells.
1
1112 (C-O). H NMR (300 MHz, CDCl3) δ 7.80-7.51 (m, 4H,
H arom), 7.50-7.25 (m, 6H, H arom), 5.67 (d, J ) 8.3 Hz, NH),
5.57-5.20 (m, 8H, CH olefin), 4.15-4.00 (m, 1H, CHN), 3.75-3.49
(m, 4H, CH2O and CH2OSi), 2.90-2.71 (m, 6H, CH2 bis-allylic),
2.42-2.27 (m, 2H, CH2 allyl), 2.20-1.95 (m, 4H, CH2 and CH2
allyl), 1.75-1.57 (m, 2H, CH2), 1.17 (d, J ) 6.7 Hz, 3H, CH3),
1.06 (s, 9H, t-Bu). 13 C NMR (75 MHz, CDCl3) δ 172.14 (NCdO),
135.52 (CH arom), 133.32 (Cquat arom), 133.16 (Cquat arom),
130.79 (CH olefin), 129.82 (CH arom), 129.10 (CH olefin), 128.86
Experimental Section
Tetraeneamide 3b. Method A. Dimethoxytrityl ether 17 (50.07
mg, 57.19 mmol) was dissolved in 1,1,1,3,3,3-hexafluoro-2-propanol
(HFIP, 500 µL, 100 mg/mL) and stirred overnight at 19 °C. A

1014 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4
Balas et al.
Figure 10. Effect of O-1918 on probe 35- and probe 37-induced MMP activity in CB1-KO HUVEC. CB1-KO HUVEC cells were grown in
six-well plates to 90% confluence and treated with vehicle or probe 35 or probe 37 at the indicated final concentration in the presence and absence
of O-1918 (1 µM). The cells were either directly placed in an incubator at 37 °C or irradiated with UV (as described in the Experimental Section)
and then placed in an incubator for 24 h. The MMP activity was measured from conditioned media as described in the Experimental Section. The
results are analyzed as percent increase in enzyme activity with respect to the corresponding vehicle treatment (basal activity) in the same cell line.
The results are expressed as mean ( SEM of percent increase from three separate experiments.
(CH olefin), 128.68 (CH olefin), 128.22 (CH olefin), 128.00 (CH
olefin), 127.76 (2 × CH olefin and CH arom), 125.73 (CH olefin),
66.83 (CH2OH), 62.09 (CH2OSi), 46.29 (CHN), 36.23 (CH2), 30.90
(CH2 allyl), 26.75 (t-Bu), 26.70 (CH2 allyl), 25.76 (CH2 bis-allyl),
25.65 (CH2 bis-allyl), 25.55 (CH2), 19.32 (Cquat t-Bu), 17.56
(CH3). MS (electrospray, positive mode): 596.5 (M + Na). Anal.
Calcd for C36H51NO3Si: C, 75.34; H, 8.96. Found: C, 75.42; H,
9.06.
General Procedure for the Deprotection of the TBDPS
Group. TBAF (1.5 equiv, 1 M in THF) was added dropwise to a
solution of silylated ether 34, 36, 38, or 40 (1.0 equiv) in anhydrous
THF (10 mL/mmol) at 0 °C. The reaction mixture was directly
plugged through a pad of Florisil using cyclohexane/CH2Cl2, 1/1,
then CH2Cl2, then increasing concentration of methanol in cyclo-
hexane to provide the corresponding pure alcohol.
Fluorescent Probe 37. White powder. 80% yield. Rf ) 0.57
(CH2CH2/MeOH, 96/4). IR (cm-1): 3719-3101 (OH/NH), 2952
(C-H), 2931 (C-H), 2851 (C-H), 1709 (COOR), 1649 (NCdO),
1546 (CHdCH), 1468 (CHdCH), 1428 (CHdCH), 1376 (CHdCH),
1297 (C-O), 1250, 1234 (C-O), 1149 (C-O), 1113, 1068, 1051.
1H NMR (500 MHz, CDCl3) δ 7.29 (d, J ) 7.9 Hz, 1H, H arom),
6.83 (d, J ) 7.9 Hz, 1H, H arom), 6.74 (t, J ) 7.9 Hz, 1H, H
arom), 5.70 (br s, 1H, NH), 5.55-5.42 (m, 2H, H olefin), 5.42-5.26
(m, 6H, H olefin), 4.39-4.18 (m, 2H, CH2O), 4.14-3.95 (m, 1H,
CHN), 3.64 (dd, J ) 10.9 Hz and J ) 3.2 Hz, 1H, CHOSi), 3.51
(dd, J ) 10.9 Hz and J ) 6.1 Hz, 1H, CHOSi), 3.02-2.66 (m, 6H,
CH2 bis-allyl), 2.58-2.42 (m, 2H, CH2 allyl), 2.25-2.13 (m, 2H,
CH2), 2.12-1.94 (m, 2H, CH2 allyl), 1.80-1.64 (m, 2H, CH2), 1.58
(s, 3H, CH3), 1.14 (d, J ) 6.8 Hz, 3H, CH3), 1.06 (s, 9H, t-Bu).
13C NMR (75 MHz, CDCl3) δ 173.67 (NCdO), 164.84 (COOR),
149.37 (Cquat arom), 148.81 (Cquat arom), 130.42 (CH olefin),
129.11 (CH olefin), 128.75 (CH olefin), 128.29 (CH olefin), 128.07
(CH olefin), 127.98 (CH olefin), 127.70 (CH olefin), 125.26 (CH
olefin), 124.69 (Cquat), 121.74 (CH arom), 120.15 (CH arom),
112.1 (Cquat), 111.32 (CH arom), 67.42 (CH2OH), 64.12 (CH2O),
47.84 (CHN), 39.56 (Cquat), 36.08 (CH2), 26.93 (CH2 allyl), 26.65
(CH2 allyl), 25.75 (CH2 bis-allyl), 25.65 (CH2 bis-allyl), 25.48
(CH2), 24.46 (t-Bu), 20.26 (CH3), 17.04 (CH3). MS (electrospray,
positive mode): 576.20 (M + Na), 1130.53 (M × 2 + Na). Anal.
Calcd for C33H47NO6: C, 71.58; H, 8.56. Found: C, 71.63; H, 8.42.
Diazirin Probe 39. Colorless oil. 75% yield. Rf ) 0.38 (CH2Cl2/
MeOH: 96/4). IR (cm-1): 3602-3137 (OH/NH), 3012 (C-H), 2930
(C-H), 2862 (C-H), 1736 (COOR), 1643 (NCdO), 1545
(CHdCH), 1450 (CHdCH), 1386 (CHdCH), 1252 (C-O), 1177
(C-O), 1051. 1H NMR (400 MHz, CDCl3) δ 5.61 (br s, 1H, NH),
5.52-5.41 (m, 1H, H olefin), 5.41-5.26 (m, 7H, H olefin),
4.14-3.99 (m, 3H, CHN and CH2OC), 3.64 (dd, J ) 3.4 Hz and
J ) 10.9 Hz, 1H, CHO), 3.51 (dd, J ) 6.1 Hz and J ) 10.9 Hz,
1H, CHO), 2.87-2.71 (m, 6H, CH2 bis-allyl), 2.45-2.33 (m, 2H,
CH2 allyl), 2.17 (t, J ) 7.7 Hz, 2H, CH2), 2.15 (t, J ) 7.8 Hz, 2H,
CH2), 2.13-2.06 (m, 2H, CH2 allyl), 1.75-1.64 (m, 4H, CH2), 1.60
(br s, 1H, OH), 1.15 (d, J ) 6.8 Hz, 3H, CH3), 1.00 (s, 3H, CH3).
13C NMR (100 MHz, CDCl3) δ 173.54 (NCdO), 172.32 (COOR),
130.75 (CH olefin), 129.12 (CH olefin), 128.76 (CH olefin), 128.36
(CH olefin), 128.30 (CH olefin), 128.05 (CH olefin), 127.88 (CH
olefin), 124.81 (CH olefin), 67.39 (CH2OH), 64.07 (CH2O), 47.82
Tetrayneamide 4b. The reaction flask was protected from light
with aluminum foil, and workup procedure was run under dim light.
A mixture of anhydrous Cs2CO3 (1.77 g, 5.43 mmol, 1.0 equiv),
CuI (1.034 g, 5.43 mmol, 1.0 equiv), NaI (1.63 g, 10.86 mmol, 2.0
equiv), terminal diyne 5 (644.12 mg, 5.97 mmol, 1.1 equiv), and
bromide 6b (2.92 g, 5.43 mmol, 1.0 equiv) in DMF (11 mL) was
stirred overnight at 40 °C. Upon cooling to 0 °C, the mixture was
diluted with isopropyl ether (40 mL) and quenched with saturated
NH4Cl (6 mL). The two layers were separated, and the etheral layer
was washed with brine (6 mL), followed by a 10% aqueous Na2S2O3
solution (6 mL) and water (2 × 6 mL). The combined aqueous
phases were re-extracted thrice with a mixture isopropyl ether/ether
(1v/1v, 150 mL), dried over MgSO4, and filtered. The solvents were
removed in vacuo, avoiding heating and light. The residue was
purified by flash chromatography on a pad of Florisil (cyclohexane/
ethyl acetate, 95/5 to 70/30) to give 1.59 g (52%) of the expected
tetrayne 4b as a yellow oil (which darkened quickly on storage).
Attempts to further purify it resulted in partial decomposition.
Hence, it was used as such in the subsequent reaction as soon as
possible. Rf ) 0.48 (cyclohexane/ethyl acetate, 4/6). IR (cm-1):
3432-3100 (OH, NH), 3071 (C-H), 2931 (C-H), 2858 (C-H),
2244 (CtC), 2207 (CtC), 1644 (NCdO), 1543 (CHdCH), 1471
(CHdCH), 1462 (CHdCH), 1427 (CHdCH), 1362 (CHdCH),
1318 (CHdCH), 1262 (C-O), 1110 (C-O), 1044 (C-O), 998.
1H NMR (300 MHz, CDCl3) δ 7.75-7.51 (m, 4H, H arom),
7.50-7.25 (m, 6H, H arom) 5.65 (br d, J ) 8.1 Hz, 1H, NH),
4.17-4.00 (m, 1H, CHN), 3.75-3.50 (m, 3H, CHOSi and CH2O),
3.58 (dd, J ) 3.9 Hz, J ) 10.1 Hz, 1H, CHOSi), 3.25-3.00 (m,
6H, CH2 bis-propargylic), 2.57-2.31 (m, 2H, CH2 propargylic),
2.31-2.05 (m, 4H, CH2 and CH2 propargylic), 1.85-1.65 (m, 2H,
CH2), 1.16 (d, J ) 6.7 Hz, 3H, CH3), 1.05 (s, 9H, t-Bu). 13C NMR
(75 MHz, CDCl3) δ 171.64 (NCdO), 135.50 (CH arom), 133.29
(Cquat arom), 133.15 (Cquat arom), 129.82 (CH arom), 127.76 (CH
arom), 79.73 (CtC), 77.31 (CtC), 75.87 (CtC), 75.06 (CtC),
74.77 (CtC), 74.65 (CtC), 74.35 (CtC), 74.14 (CtC), 66.78
(CH2OSi), 61.00 (CH2OH), 46.36 (CHN), 35.47 (CH2), 26.86 (t-
Bu), 24.49 (CH2), 23.02 (CH2 propargylic), 19.31 (Cquat t-Bu),
18.14 (CH2 propargylic), 17.52 (CH3), 9.73 (CH2 bis-propargylic).
MS (electrospray, positive mode): 566.4 (M + H), 588.6 (M +
Na).

Synthesis of Anandamide Probes
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4 1015
(CHN), 36.07 (CH2), 29.68 (CH2), 28.76 (CH2), 26.82 (CH2 allyl),
26.64 (CH2 allyl), 25.65 (CH2 bis-allyl), 25.48 (CH2), 24.89 (Cquat
diazirin), 19.66 (CH3), 17.04 (CH3). MS (electrospray, positive
mode): 468.13 (M + Na), 440.20 (M - N2 + Na). Anal. Calcd for
C25H39N3O4: C, 67.39; H, 8.82. Found: C, 67.15; H, 8.97.
Measurement of MMP Activity. The MMP (matrix metallo-
proteinase) activity was measured using Enzchek gelatinase/
collagenase assay kit (Molecular Probes) following the manufac-
turer’s protocol using a multiwell microplate reader. The cells were
plated in a six-well plate and at 90% confluence treated with vehicle
or test drugs for 24 h. Conditioned media samples (100 µL) were
collected (following treatment) and added into the wells of a 96-
well plate in triplicate followed by 80 µL of 1× reaction buffer
(from Molecular Probes kit) in each assay well. Then 20 µL of
“DQ gelatin” (6 µg/mL, Molecular Probes) was added to each well
and the mixture was incubated in the plate for 24 h at 37 °C in the
dark. After the incubation period, fluorescence is measured using
a microplate reader (Victor S23-271-05) at excitation/emission of
485/530 nm. A standard curve is simultaneously developed incubat-
ing the same amount of substrate “DQ gelatin” (6 µg/mL) with
varying amounts (0.01-0.04 µg/mL) of purified Clostridium
collagenase enzyme (Molecular Probe kit). The digestion of the
substrate with the enzyme will remove the quenching of heavily
loaded fluorescence labeling in gelation, and the increase in the
fluorescence reading is directly proportional to the enzyme activity.
The experiment is performed in triplicate, and data are analyzed
using GraphPad Prism software. For experiments with probe 37,
background fluorescence reading was subtracted from the experi-
mental readings. The background fluorescence of probe 37 was
obtained by adding probe 37 at different concentrations (as used
to treat the cells) into complete media and then measuring
fluorescence after 24 h as described above.
Diazirin Probe 41. Pale-yellow oil. 81% yield. Rf ) 0.51
(CH2Cl2/MeOH, 96/4). IR (cm-1): 3669-3109 (OH/NH), 3012
(C-H), 2931 (C-H), 2857 (C-H), 1734 (COOR), 1643 (NCdO),
1545 (CHdCH), 1453 (CHdCH), 1386 (CHdCH), 1256 (C-O),
1
1173 (C-O), 1051, 994. H NMR (300 MHz, CDCl3) δ 5.66 (br
s, 1H, NH), 5.55-5.41 (m, 1H, H olefin), 5.41-5.23 (m, 7H, H
olefin), 4.17-3.97 (m, 3H, CHN and CH2O), 2.88 (t, J ) 5.5 Hz,
1H, OH), 2.85-2.67 (m, 6H, CH2 bis-allyl), 2.49-2.32 (m, 2H,
CH2 allyl), 2.26- (t, J ) 7.3 Hz, 2H, CH2), 2.17 (t, J ) 7.6 Hz, 2H,
CH2), 2.12-1.94 (m, 2H, CH2 allyl), 1.80-1.61 (m, 2H, CH2),
1.57-1.43 (m, 2H, CH2), 1.42-1.26 (m, 2H, CH2), 1.14 (d, J )
6.8 Hz, 3H, CH3), 0.98 (s, 3H, CH3). 13 C NMR (75 MHz, CDCl3)
δ 173.61 (NCdO), 173.02 (COOR), 130.70 (CH olefin), 129.10
(CH olefin), 128.74 (CH olefin), 128.34 (CH olefin), 128.29 (CH
olefin), 128.04 (CH olefin), 127.89 (CH olefin), 124.87 (CH olefin),
67.29 (CH2OH), 63.80 (CH2O), 47.77 (CHN), 36.06 (CH2), 33.69
(CH2), 33.51 (CH2), 26.85 (CH2 allyl), 26.63 (CH2 allyl), 25.63
(br, CH2 bis-allyl), 25.47 (Cquat diazirin), 19.68 (CH3), 19.46
(CH2), 17.03 (CH3). MS (electrospray, positive mode): 482.1 (M
+ Na), 454.2 (M - N2 + Na). Anal. Calcd for C26H41N3O4: C,
67.94; H, 8.99. Found: C, 68.09; H, 9.12.
siRNA and Transfection. Several siRNA against the CB1
cannabinoid receptors were constructed and synthesized by Ambion
(Austin, TX) using proprietary methods. The specificity of the
siRNA sequence was checked via BLAST search. For silencing
CB1 receptors, HUVEC cells were grown to 70% confluence.
Thereafter, cells were transfected with CB1 cannabinoid receptor-
specific silencing siRNA or nonsilencing siRNA using siPORT
siRNA electroporation kit (Ambion, TX) following the manufac-
turer’s instructions. In brief, a cell pellet (approximately 7 × 104)
was collected, suspended in 75 µL of SiRNA electroporation buffer
(Ambion) and mixed with 1.5 µg of siRNA or nonsilencing siRNA.
The mixture was subjected to electroporation in an electroporation
cuvette (1 mm pass) under conditions identified in optimizing
experiments. Following electroporation, samples in the cuvette were
incubated at 37 °C for 10 min and cells were transferred to a culture
dish with prewarmed growth media. Cells were incubated at 37 °C
in a 95% O2/5% CO2 incubator until they were analyzed.
In Vitro Angiogenesis Assay (Tube Formation Assay). The
tube formation assay was performed as described previously.52
Briefly, 24-well plates were coated with 300 µL of 1:1 mixture of
RPMI 1640 and Matrigel (10 mg/mL, Invitrogen) and incubated at
37 °C for 1 h to promote gelling. HUVECs (40 000) in 0.5 mL of
serum-free RPMI 1640 medium with varying concentrations of drug
were added to each well. For a positive control, bFGF was added
at a concentration of 5 ng/mL. All test samples were performed in
triplicate. After 36 h of incubation, the plates were fixed with Diff-
Quick and images were captured using a Nikon camera and
analyzed using ImagePro software. The total length of the tubes
formed by the endothelial cells was measured from three randomly
chosen fields (10×) for each sample and analyzed by the ImagePro
software program. This experiment was performed in triplicate for
each sample three times.
Cell Migration Assay. The cell migration assay was performed
as described previously53 in a 48-well microchemotaxis chamber
(Neuro Probe, Inc., MD). Briefly, polyester membranes with 10
µm pores (Neuro Probe, Inc., MD.) were coated with a 0.1 mg/mL
of collagen IV (Invitrogen) in doubly distilled water and then dried
for 1 h. HUVECs and CB1-KO HUVEC were harvested using
Versene (Sigma; St Louis, MO) and resuspended in RPMI 1640
containing 0.1% BSA. The bottom chamber was loaded with 30 000
cells, and the filter was laid over the cells. The microchamber was
then inverted and incubated at 37 °C for 2 h. After reinversion of
the chamber to its upright position, the upper wells were then loaded
with RPMI 1640 containing 0.1% BSA and test drugs. bFGF was
added at a concentration of 5 ng/mL as a positive control. The
chamber was then reincubated at 37 °C for 6 h, and the filters were
fixed and stained using Diff-Quick (Baxter Healthcare Corp., IL).
The cells that migrated through the filter were quantitated by
counting the center of each well at ×20 using an Olympus CK 40
microscope. Each condition was studied in triplicate wells, and each
experiment was performed three times.
Acknowledgment. We gratefully acknowledge Pr. Joseph
Vercauteren (Laboratory of Pharmacognosy, Montpellier, France)
for allowing us to record ESI mass spectra in his laboratory.
The authors are deeply grateful to Pr. Jean-Yves Lallemand and
the ICSN for their generous financial support. We thank Dr.
Eric G. Spokas (Rider University, Lawrenceville, NJ) for a
critical reading of the manuscript. This work was supported by
a NC Biotech grant to S.M., NIH Grant U24-DA 12385, and
U.S. Army Medical Research and Material Command (Grant
07-1-0418).
Supporting Information Available: General methods and
experimental procedures together with 1H NMR and 13C NMR data
and spectra for every new compound. This material is available
free of charge via the Internet at http://pubs.acs.org.
Irradiation of Ligands. The endothelial cells were grown in
six-well plates at 90% confluence and then placed in a C-65 cabinet
attached with one 8-W EL lamp [UVLMS-38 8-W UV lamp (14.8
in. length × 3.8 in. width × 2.5 in. depth) equipped with low,
medium, and long range (365/302/254 nm) UV-generating wave-
length 365/302/254 nm; UVP, Upland, CA]. The cells were treated
with test drugs and irradiated for 1-15 min at 280-315 nm range.
The distance between the light source and the cells was 10 in. The
cell viability was assessed before and after irradiation, and no
significant difference in cell viability was observed up to 10 min
of exposure. Therefore, the experiment described herein was
performed after 5 min of excitation.
References
(1) Hanus, L. O. Discovery and isolation of anandamide and other
endocannabinoids. Chem. BiodiVersity 2007, 4, 1828–1841.
(2) Devane, W. A.; Hanus, L.; Breuer, A.; Pertwee, R. G.; Stevenson,
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, 1946–1949.

1016 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4
Balas et al.
(3) (a) Burstein, S. H.; Adams, J. K.; Bradshaw, H. B.; Fraioli, C.; Rossetti,
R. G.; Salmonsen, R. A.; Shaw, J. W.; Walker, J. M.; Zipkin, R. E.;
Zurier, R. B. Potential anti-inflammatory actions of the elmiric
(lipoamino) acids. Bioorg. Med. Chem. 2007, 15, 3345–3355. (b)
Burstein, S. The elmiric acids: biologically active anandamide analogs.
Neuropharmacology 2007, doi:10.1016/j.neuropharm.2007.11.011.
(4) Kogan, N. M.; Mechoulam, R. J. The chemistry of endocannabinoids.
Endocrinol. InVest. 2006, 29, 3–14.
(5) Milman, G.; Maor, Y.; Abu-Lafi, S.; Horowitz, M.; Gallily, R.; Batkai,
S.; Mo, F. M.; Offertaler, L.; Pacher, P.; Kunos, G.; Mechoulam, R.
N-Arachidonoyl L-serine, an endocannabinoid-like brain constituent
with vasodilatory properties. Proc. Natl. Acad. Sci. U.S.A. 2006, 103,
2428–2433.
(6) Saghatelian, A.; McKinney, M. K.; Bandell, M.; Patapoutian, A.;
Cravatt, B. F. A FAAH-regulated class of N-acyl taurines that activates
TRP ion channels. Biochemistry 2006, 45, 9007–9015.
(7) Mechoulam, R. Cannabinoids as Therapeutics; Mechoulam, R., Ed.;
Milestones in Drug Therapy; Birkhauser Publisher: Basel, Switzerland,
2005.
(8) Grimaldi, C.; Bibulco, M. Anandamide, an endogenous ligand of
cannabinoids receptors, inhibits human breast cancer cell proliferation
through a lipid rafts mediated mechanism. Pharmacologyonline 2007,
1, 1–45.
(26) Perry, S. N.; Johnson, I. J.; Mukhopadhyay, S. Anandamide Mediated
Angiogenesis: Interplay between CB1 Receptor and Non CB1/CB2
Anandamide Receptor. Proceedings of the 16th Annual Symposium
of Cannabinoids, Burlington, VT, 2006; International Cannabinoid
Research Society, 2006; p 61.
(27) Hilderbrandt, S.; Johnson, I. J.; Mukhopadhyay, S. CB1 Receptor and
Non-CB1/CB2 Annadmaide Receptor Mediated Differential S-Ni-
trosylation of MMP: A Novel Angiogenic Switch for the Regulation
of Angiogenesis (2008). Proceedings of the 18th Annual Symposium
of Cannbinoids, Burlington, VT, 2008; International Cannabinoid
Research Society, 2008; p 17.
(28) McCollum, L.; Howlett; Mukhopadhyay, S. Anandamide-mediated
CB1/CB2 cannabinoid receptorsindependent nitric oxide production
in rabbit aortic endothelial cells. J. Pharmacol. Exp. Ther. 2007, 321,
930–937.
(29) Daly, C. J.; McGrath, J. C. Fluorescent ligands, antibodies, and proteins
for the study of receptors. Pharmacol. Ther. 2003, 100, 101–118.
(30) Vodovozova, E. L. Photoaffinity labeling and its application in
structural biology. Biochemistry (Moscow) 2007, 72, 1–20.
(31) (a) Brase, S.; Gil, C.; Knepper, K.; Zimmermann, V. Organic azides:
an exploding diversity of a unique class of compounds. Angew. Chem.,
Int. Ed. 2005, 44, 5188–240. (b) Bucher, G. Photochemical Reactivity
of Azides. In CRC Handbook of Organic Photochemistry and
Photobiology, 2nd ed.; Horspool, W. M., Lenci, F., Eds.; CRC Press:
Boca Raton, FL, 2004; Chapter 44. (c) Blencowe, A.; Hayes, W.
Development and application of diazirines in biological and synthetic
macromolecular systems. Soft Matter 2005, 1, 178–205. (d) Sadakane,
Y.; Hatanaka, Y. Photochemical fishing approaches for identifying
target proteins and elucidating the structure of a ligand-binding region
using carbene-generating photoreactive probes. Anal. Sci. 2006, 22,
209–218.
(9) Di Marzo, V.; Petrosino, S. Endocannabinoids and the regulation of
their levels in health and disease. Curr. Opin. Lipidol. 2007, 18, 129–
140.
(10) Howlett, A. C.; Barth, F.; Bonner, T. I.; Cabral, G.; Casellas, P.;
Devane, W. A.; Felder, C. C.; Herkenham, M.; Mackie, K.; Martin,
B. R.; Mechoulam, R.; Pertwee, R. G. International Union of
Pharmacology. XXVII. Classification of cannabinoid receptors. Phar-
macol. ReV 2002, 54, 161–202.
(32) (a) Abadji, V.; Lin, S.; Taha, G.; Griffin, G.; Stevenson, L. A.; Pertwee,
R. G.; Makriyannis, A. (R)-Methanandamide: a chiral novel ananda-
mide possessing higher potency and metabolic stability. J. Med. Chem.
1994, 37, 1889–1893. (b) Haller, V. L.; Cichewicz, D. L.; Welch,
S. P. Non-cannabinoid CB1, non-cannabinoid CB2 antinociceptive
effects of several novel compounds in the PPQ stretch test in mice.
Eur. J. Pharmacol. 2006, 546, 60–68.
(33) It must be realized that we are not looking for a specific ligand for
CB1 or CB2 receptors but for a mimic of eCBs, mainly anandamide,
to search for novel putative receptors. However, very little is known
about these targeted new receptors. It is therefore rather difficult to
anticipate what structural requirements would be needed for a synthetic
ligand.
(34) Balas, L.; Cascio, M. G.; Di Marzo, V.; Durand, T. Synthesis of a
potential photoactivatable anandamide analog. Bioorg. Med. Chem.
Lett. 2006, 16, 3765–3768.
(35) Altundas, R.; Mahadevan, A.; Razdan, R. K. A synthetic route to
anandamide analogues carrying a substituent at the terminal carbon
and an acetylene group in the end pentyl chain. Tetrahedron Lett. 2004,
45, 5449–5451.
(36) Li, C.; Xu, W.; Vadivel, S. K.; Fan, P.; Makriyannis, A. High affinity
electrophilic and photoactivatable covalent endocannabinoid probes
for the CB1 receptor. J. Med. Chem. 2005, 48, 6423–6429.
(37) Dasse, O.; Mahadevan, A.; Han, L.; Martin, B. R.; Marzo, V. D.;
Razdan, R. K. The synthesis of N-vanillyl-arachidonoyl-amide (arvanil)
and its analogs: an improved procedure for the synthesis of the key
synthon methyl 14-hydroxy-(all-cis)-5.8.11-tetradecatrienoate. Tetra-
hedron 2000, 56, 9195–9202.
(38) Caruso, T.; Spinella, A. Cs2CO3 promoted coupling reactions for
the preparation of skipped diynes. Tetrahedron 2003, 59, 7787–7790.
(39) Leonard, N. J.; Neelima. 1,1,1,3,3,3-Hexafluoro-2-propanol for the
removal of the 4,4′-dimethoxytrityl protecting group from the 5′-
hydroxyl of acid-sensitive nucleosides and nucleotides. Tetrahedron
Lett. 1995, 36, 7833–7336.
(40) Bailey, W. J.; Fujiwara, E. Acetylenes. I. Mixed dihalides and
halohydrins from butynediol. J. Am. Chem. Soc. 1955, 1, 165–166.
(41) In our first attempts, there were both side products with over-reduced
double bond(s) and compounds with not-yet-reduced alkyne(s) and
they migrated with the same Rf (hampered by the presence of amine
and boron and/or nickel derivatives salts in the ethanolic reaction
mixture). Thus, unfortunately, development and completion of the
hydrogenation reaction could not be monitored by hydrogen volume
absorption measurement or TLC. We had to use a trial and error
approach to find the level of poisoning of the catalyst and reaction
time.
(11) (a) Steffens, M.; Zentner, J.; Honegger, J.; Feuerstein, T. J. Binding
affinity and agonist activity of putative endogenous cannabinoids at
the human neocortical CB1 receptor. Biochem. Pharmacol. 2005, 69,
169–178. (b) Porter, A. C.; Sauer, J. M.; Knierman, M. D.; Becker,
G. W.; Berna, M. J.; Bao, J.; Nomikos, G. G.; Carter, P.; Bymaster,
F. P.; Leese, A. B.; Felder, C. C. Characterization of a novel
endocannabinoid, virodhamine, with antagonist activity at the CB1
receptor. J. Pharmacol. Exp. Ther. 2002, 301, 1020–1024.
(12) Fride, E.; Gobshtis, N. Endocannabinoids and their receptors: physiol-
ogy, pathology and pharmacology. Immunol., Endocr. Metab. Agents
Med. Chem. 2007, 7, 157–173.
(13) A full issue of the British Journal of Pharmacology (2008, 153 (2))
is dedicated to reviews and topics on the CB2 receptor, stemming
from a meeting “CB2 Receptors: New Vistas” that was held in Canada
in June 2007.
(14) (a) Liu, J.; Wang, L.; Harvey-White, J.; Huang, B. X.; Kim, H. Y.;
Luquet, S.; Palmiter, R. D.; Krystal, G.; Rai, R.; Mahadevan, A.;
Razdan, R. K.; Kunos, G. Multiple pathways involved in the
biosynthesis of anandamide. Neuropharmacology 2008, 54, 1–7. (b)
McPartland, J. M.; Norris, R. W.; Kilpatrick, C. W. Coevolution
between cannabinoid receptors and endocannabinoid ligands. Gene
2007, 397, 126–135.
(15) Demuth, D. G.; Molleman, A. Cannabinoid signalling. Life Sci. 2006,
78, 549–563.
(16) Felder, C. C.; Dickason-Chesterfield, A. K.; Moore, S. A. Cannabinoids
biology: the search for new therapeutic targets. Mol. InterVentions
2006, 6, 149–161.
(17) Smita, K.; Sushil Kumar, V.; Premendran, J. S. Anandamide: an update.
Fundam. Clin. Pharmacol. 2007, 21, 1–8.
(18) Alexander, S. P.; Kendall, D. A. The complications of promiscuity:
endocannabinoid action and metabolism. Br. J. Pharmacol. 2007, 152,
602–623.
(19) Fowler, C. J. The contribution of cyclooxygenase-2 to endocannabinoid
metabolism and action. Br. J. Pharmacol. 2007, 152, 594–601.
(20) Brown, A. J. Novel cannabinoid receptors. Br. J. Pharmacol. 2007,
152, 567–575.
(21) Di Marzo, V.; De Petrocellis, L. Non-CB1 Non CB2 Receptors for
Endocannabinoids. In Endocannabinoids: The Brain and Body’s
Marijuana and Beyond; Onaivi, E. S., Sugiura, T., Di Marzo, V., Eds.;
Taylor & Francis: Boca Raton, FL, 2006; Vol. 15, pp 1-174.
(22) Begg, M.; Pacher, P.; Batkai, S.; Osei-Hyiaman, D.; Offertaler, L.;
Mo, F. M.; Liu, J.; Kunos, G. Evidence for novel cannabinoid
receptors. Pharmacol. Ther. 2005, 106, 133–145.
(23) Pertwee, R. G. Novel pharmacological targets for cannabinoids. Curr.
Neuropharmacol. 2004, 2, 9–29.
(24) Fride, E.; Foox, A.; Rosenberg, E.; Faigenboim, M.; Cohen, V.; Barda,
L.; Blau, H.; Mechoulam, R. Milk intake and survival in newborn
cannabinoid CB1 receptor knockout mice: evidence for a “CB3”
receptor. Eur. J. Pharmacol. 2003, 461, 27–34.
(25) Di Marzo, V.; De Petrocellis, L.; Fezza, F.; Ligresti, A.; Bisogno, T.
Anandamide receptors. Prostaglandins, Leukotrienes Essent. Fatty
Acids 2002, 66, 377–391.
(42) Taber, D. F.; Zhang, Z. J. A linpchin approach to unsaturatted fatty
acids 11,12-epoxyeicosatrienoic acid and 11S,12S-dihydroxyeicosa-
trienoic acid ethyl esters. Org. Chem. 2005, 70, 8093–8095.
(43) (a) Colonge, J.; Poilane, G. Utilisation du chloro-4 bute`ne-2 ol-1
en synthe`se organiqueaction sur les compose´ organomanesiens. Bull.
Soc. Chim. Fr. 1955, 953, 955. (b) Imai, T.; Nishida, S. A mild and
convenient Barbier-type allylation od aldehyddes to homoallylic

Synthesis of Anandamide Probes
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 4 1017
alcohol via iodide ion promoted stannylation of allylic bromides
and chlorides with tin(II) chloride. Synthesis 1993, 4, 395–399.
(44) We were mostly preoccupied with double selectivity versus monose-
(49) Perrier, H.; Prasit, P.; Wang, Z. Synthesis of a radioactive photoaffinity
arachidonic acid analog. Tetrahedron Lett. 1994, 35, 1501–1502.
(50) (a) Church, R. F. R.; Conn, R.; Weiss, M. J. Substituted Diaziridines
and Diazirines. U.S. Patent 3,525,736 1970; American Cyanamid
Compagny. (b) Church, R. F. R.; Weiss, M. J. Diazirines. II. Synthesis
and properties of small functionalized diazirine molecules. Some
observations on the reaction of a diaziridine with the iodine-iodide
ion system. J. Org. Chem. 1970, 35 (8), 2465–2471.
(51) These experiments were run by Pr. Vincenzo DiMarzo team (Napoli)
with the same conditions as the ones reported in our preliminary paper
(see ref 34).
(52) Grant, D. S.; Kinsella, J. L.; Fridman, R.; Auerbach, R.; Piasecki,
B. A.; Yamada, Y.; Zain, M.; Kleinman, H. K. Interaction of
endothelial cells with a laminin A chain peptide (SIKVAV) in vitro
and induction of angiogenic behavior in vivo. J. Cell Physiol. 1992,
153, 614–625.
lectivity on the dichloride 22.
(45) Hansen, T. V.; Stenstrom, Y. First total synthesis of (-)-aplyolide A.
Tetrahedron: Asymmetry 2001, 12, 1407–1417.
(46) (a) Kwok, P.-Y.; Muellner, F. W.; Chen, C.-K.; Fried, J. Total synthesis
of 7,7-, 10,10-, and 13,13-difluoroarachidonic acids. J. Am. Chem.
Soc. 1987, 109, 3684–3692. (b) Smith, L. M.; Smith, R. G.; Loehr,
T. M.; Daves, G. D. Douglas Fir Tusock Moth pheromone: identifica-
tion of a diene analogue of the principal attractant and synthesis of
stereochemically defined 1,6-, 2,6- and 3,6-heneicosadien-11-ones. J.
Org. Chem. 1978, 43, 2361–2366.
(47) Blackburn, C.; Neumeyer, J. L. Receptor-selective fluorescent probes
for use in neuroscience. Med. Chem. Res. 1992, 2, 257–275.
(48) Dohi, H.; Nishida, Y.; Mizuno, M.; Shinkai, M.; Kobayashi, T.;
Takeda, T.; Uzawa, H.; Kobayashi, K. Synthesis of an artificial
glycoconjugate polymer carrying Pk antigenic trisaccharide and its
potent neutralization activity against Shiga-like toxin. Bioorg. Med.
Chem. 1999, 7, 2053–2062.
(53) Koch, A. E.; Halloran, M. M.; Haskell, C. J.; Shah; Polverini.
Angiogenesis mediated by soluble forms of E-selectin and vascular
cell adhesionmolecule-1. Nature 1995, 376, 517–519.
JM8011382