N-acylvanillamides: Development of an expeditious synthesis and discovery of new acyl templates for powerful activation of the vanilloid receptor

Article abstract of DOI:10.1021/jm020844o

A simple and general synthesis of vanillamides was developed and employed to screen acids from the fatty and isoprenoid pools for new acyl templates of biological relevance as capsaicin analogues. Potent activation of the human vanilloid receptor 1 (VR1) was observed for the vanillamides of certain polyfunctional acids from both pools, showing that the vanilloid activity of capsaicinoids can be substantially improved by introducing polar groups and/or unsaturations on the acyl moiety. The activity of the unsaturated analogues was maintained or even increased by cyclopropanation, while ω dimerization led to a substantial increase of activity. Because of the wide structural diversity of the library of compounds screened, these observations could not be translated into a single framework of structure-activity relationships. Nevertheless, a series of new highly active leads was identified, validating the pharmacological potential of the unnatural combination of natural building blocks to provide new bioactive compounds.

Full text of DOI:10.1021/jm020844o

J . Med. Chem. 2002, 45, 3739-3745
3739
N-Acylva n illa m id es: Develop m en t of a n Exp ed itiou s Syn th esis a n d Discover y of
New Acyl Tem p la tes for P ow er fu l Activa tion of th e Va n illoid Recep tor
Giovanni Appendino* and Alberto Minassi
Dipartimento di Scienze Chimiche, Alimentari, Farmaceutiche e Farmacologiche, Viale Ferrucci 33, 28100 Novara, Italy
Aniello Schiano Morello and Luciano De Petrocellis
Istituto di Cibernetica del CNR, Endocannabinoid Research Group, Via Campi Flegrei 34, Comprensorio A. Olivetti,
Building 70, 80078 Pozzuoli (NA), Italy
Vincenzo Di Marzo*
Endocannabinoid Research Group, Istituto per la Chimica di Molecole di Interesse Biologico, CNR, Via Campi Flegrei 34,
Comprensorio A. Olivetti, Building 70, 80078 Pozzuoli (NA), Italy
Received February 6, 2002
A simple and general synthesis of vanillamides was developed and employed to screen acids
from the fatty and isoprenoid pools for new acyl templates of biological relevance as capsaicin
analogues. Potent activation of the human vanilloid receptor 1 (VR1) was observed for the
vanillamides of certain polyfunctional acids from both pools, showing that the vanilloid activity
of capsaicinoids can be substantially improved by introducing polar groups and/or unsaturations
on the acyl moiety. The activity of the unsaturated analogues was maintained or even increased
by cyclopropanation, while ω dimerization led to a substantial increase of activity. Because of
the wide structural diversity of the library of compounds screened, these observations could
not be translated into a single framework of structure-activity relationships. Nevertheless, a
series of new highly active leads was identified, validating the pharmacological potential of
the unnatural combination of natural building blocks to provide new bioactive compounds.
In tr od u ction
that the lipophilic domain of the capsaicin receptor has
element(s) capable of binding polar groups or, alterna-
tively, that certain structural elements can preorganize,
in conformational terms, the lipophilic acyl moiety of
N-AVAM for binding. Thus, while long-chain analogues
of CPS like N-stearoylvanillamine (1b) are inactive, the
introduction of unsaturations fully restores activity, as
cogently demonstrated by N-oleylvanillamine (olvanil,
1c).6 Both alternatives hint at the possibility of sub-
stantially increasing the activity of CPS by suitable
modification of the acyl chain, but this opportunity has
remained, surprisingly, largely untested.
N-Acylvanillamines (N-AVAM), a class of compounds
unique to the genus Capsicum, are exemplified by
capsaicin (CPS, 1a ),1 the major pungent principle of hot
pepper and the archetypal vanilloid.2 N-AVAM have
been at the center of intense research activity aimed at
elucidating the basis of their antinociceptive properties
and exploiting their therapeutic potential. These studies
provided evidence of definite structure-activity rela-
tionships; however, the potency of the parent compound
could not be increased significantly, and no better
clinical lead emerged from these investigations.3 This,
and the discovery of alternative templates with im-
proved potency and/or pharmacological profiles (resinifer-
Long lipophilic sequences resembling the acyl moiety
of CPS but constrained in a covalent or conformational
way and occasionally interrupted by polar groups (hy-
droxyl, ketone carbonyl, double bond(s)) are present in
a host of compounds from the natural products pool,
mostly fatty and terpenoid acids. Many of them are
easily available and could provide interesting probes for
characterizing the lipophilic domain of the capsaicin
binding site. These compounds might also expand the
pharmacological profile of N-AVAM as shown by the
peculiar activity of vanillamides derived from polyun-
saturated fatty acids, like N-arachidonoylvanillamine
(arvanil, 1d ).7 This hybrid between the vanilloid CPS
and the endocannabinoid anandamide can activate
vanilloid and cannabinoid receptors more powerfully
than related pure agonists of either receptors,7 a finding
which places N-AVAM in a much broader perspective
and validates them as interesting tools for targeting
multiple receptors involved in pain perception and
inflammation.
onoids,4a
N-vanilly-N′-(3-acyloxy-2-benzylpropyl)-
thioureas,4b N-alkylhomovanillamides,4c and N-vanillyl-
N′-benzylureas4d), substantially shifted the focus of
vanilloid research from N-AVAM.
Modification of the amide region and the vanillyl
moiety of CPS has been pursued in a systematic and
extensive way,3 but information on the acyl moiety is
still scanty. This element has a high degree of confor-
mational freedom because six torsion angles can be
varied in the range of 0-360°, and the resulting rota-
meric distribution spans a wide range of geometries,
with hundreds of low-energy conformations.5 Many of
these might not be appropriate for binding and, thus,
are essentially inactive, but a systematic conformational
searching with constrained probes of CPS would be
prohibitive because of the number of torsional variables
involved. Furthermore, certain observations suggest
These observations provided a strong rationale for
screening the natural products pool for new acyl tem-
plates of biological significance as N-AVAM constitu-
* Corresponding authors. G.A.: phone +39 011 6707684, fax +39
011 6707687, e-mail appendino@pharm.unipmn.it. V.D.M.: phone +39
081 8675093, fax +39 081 8041770, e-mail vdimarzo@icmib.na.cnr.it.
10.1021/jm020844o CCC: $22.00 © 2002 American Chemical Society
Published on Web 07/09/2002

3740 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 17
Appendino et al.
Ta ble 1. Amidation Yields and Biological Evaluation of the
Vanillamides 1a -4ba
amidation
procedure
(yield)b
percentage of
maximum
response
compound
1a
pEC50
65.4 ( 4.6
7.7 ( 0.1
(capsaicin)
1b
1c
A (76%)
A (81%), B (77%)
32.4 ( 2.3
67.2 ( 7.1
N Mc
9.3 ( 0.2
(olvanil)
(arvanil)
1d
A (95%), B (79%)
74.8 ( 3.5
9.3 ( 0.2
ents. To this aim, we have developed a new and
expeditious procedure for the synthesis of these com-
pounds and report its application to the discovery of
several new and highly active templates from both the
terpenoid and fatty acid series.
1e
1f
A (78%)
A (73%)
70.9 ( 3.2
79.9 ( 1.2
7.4 ( 0.1
9.3 ( 0.3
(farvanil)
1g
1h
1i
A (80%)
85.0 ( 4.8
85.0 ( 5.1
74.5 ( 1.0
92.6 ( 8.2
8.7 ( 0.1
6.7 ( 0.3
6.1 ( 0.1
8.4 ( 0.1
A (78%)
A (56%)
1j
A (98%), B (86%)
(retvanil)
Ch em istr y
1k
1l
B (88%)
A (85%)
B (75%)
A (86%)
B (84%)
34.6 ( 1.5
91.4 ( 5.3
73.7 ( 2.2
70.7 ( 0.3
85.0 ( 4.9
N M
Despite the pharmacological relevance of vanilla-
mides, there is still a need for a simple and mild method
to prepare these compounds, especially in a single step
and with an easy isolation procedure. The major prob-
lem is the chemoselectivity of the acylation. In situ
activation with carbodiimide condensing agents (DCC,
EDC) requires protection of the phenolic hydroxyl of
vanillamine8 because discrimination between the amino
group and the phenolic hydroxyl is poor.9 N,O-Diacyl
derivatives formed in the reaction can be selectively
O-deacylated with pyrrolidine,10 but this strategy is
untenable with expensive acids, with one full equivalent
being wasted, and problems can arise with base-labile
functionalities. Protection of vanillamine is also neces-
sary when the acid is activated ex situ as a hydroxysuc-
cinimide derivative or as a mixed anhydride.8 Acylation
under Schotten-Baumann conditions shows excellent
selectivity for N- versus O-acylation;6 however, activa-
tion of the acid as a chloride is required, and formation
of emulsions can take place with long-chain acids.
Furthermore, while formation of acyl chlorides is a
routine operation for simple unfunctionalized acids,
carefully controlled conditions are required with poly-
unsaturated acids, and protection on the acyl moiety is
necessary for hydroxy acids.
To overcome these limitations, we have developed a
protocol for acylation of vanillamine via mixed phos-
phoric anhydrides. Acyl activation with the commer-
cially available reagents diethyl phosphorocyanidate
(DEPC)11 and propane phosphonic acid anhydride
(PPAA)12 could be attained in situ under mild condi-
tions, and the mixed phosphoric anhydrides obtained
in this way displayed the same excellent selectivity of
acyl chlorides for N- versus O-(phenolic) acylation.
Formation of mixed anhydrides from carboxylic acids
was faster than phosphorylation of the nucleophilic sites
of vanillamine, and a reduced reactivity toward alcoholic
hydroxyls was also observed, thus making hydroxyl
protection on both vanillamine and hydroxy acids un-
necessary. This reactivity pattern, and the possibility
of generating vanillamine in situ from its air-stable
hydrochloride, eventually reduced the whole process into
a one-pot and experimentally simple operation, while
the formation of emulsions could be prevented by using
a dry workup, an option impossible under Schotten-
Baumann conditions. The mixed phosphoric anhydride
protocol could also be applied successfully to the prepa-
ration of other bioactive phenolic amides (tyramides,
dopamides, and p-hydroxyanilides).13
8.5 ( 0.1
7.9 ( 0.1
8.9 ( 0.1
8.5 ( 0.1
1m
1n
1o
(rinvanil)
1p
1q
1r
1s
1t
A (90%), B (91%)
85.0 ( 4.1
93.5 ( 5.9
69.7 ( 0.9
67.0 ( 1.2
86.7 ( 3.2
98.0 ( 7.1
91.0 ( 4.9
96.2 ( 6.9
91.2 ( 4.4
37.5 ( 2.5
14.3 ( 2.3
4.0 ( 3.0
8.7 ( 0.1
9.4 ( 0.5
8.3 ( 0.2
8.2 ( 0.1
8.9 ( 0.1
8.3 ( 0.2
7.1 ( 0.3
8.5 ( 0.1
8.6 ( 0.1
N M
B (85%)
A (82%)
A (80%)
ad
1u
1v
1w
1x
2a
2b
2c
2d
2e
2f
a
a
a
a
A (68%)
A (31%)
A (80%), B (7%)
B (50%)
A (36%)
A (91%)
B (46%)
A (71%)
be
N M
N M
5.8 ( 2.2
N M
88.6 ( 4.5
78.9 ( 3.7
63.6 ( 4.2
59.3 ( 5.1
83.4 ( 8.9
85.7 ( 5.3
16.4 ( 2.5
37.2 ( 3.1
6.9 ( 0.1
6.7 ( 0.1
6.1 ( 0.1
7.7 ( 0.1
6.3 ( 0.1
8.8 ( 0.1
N M
3a
3b
3c
3d
4a
4b
b
A (58%)
A (52%)
N M
a
Data are means ( SEM of three separate determinations. The
maximum response was measured at 10 µM for the less-potent
compounds and at 1 µM for the most-potent ones. The doses tested
were 0.1, 1, 10, 100, and 1000 nM for the most-potent compounds
b
and 0.1, 1, and 10 µM for the less-potent compounds. A: DEPC
protocol. B: PPAA protocol. c N M ) non measurable. Cyclopro-
d
panation yield from the corresponding monounsaturated VA: 1t,
45%; 1u , 36%; 1v, 43%; 1w , 40%; and 1x, 42%. e Yield from 1p :
21% for 3c; 25% for 3d .
With simple, saturated fatty acids, amidation of
vanillamine via mixed phosphoric anhydrides gave
yields comparable to those obtained with acyl chlorides,
but the milder reaction conditions dramatically in-
creased yields with polyenoic acids (Table 1). Thus, the
reaction between arachidonic acid and vanillamine
hydrochloride in the presence of DEPC and triethy-
lamine gave arvanil (1d ) in 95% yield. The yield with
PPAA was lower (79%) but still considerably higher
compared to that observed when arachidonoyl chloride
was reacted with vanillamine hydrochloride under
Schotten-Baumann conditions (34%).6 Similarly, the
vanillamide of retinoic acid (1j) could be prepared in
essentially quantitative yield, and ximeninic acid, an
unstable enine, could also be amidated in satisfactory
yield to 1n (63%). The mixed anhydride protocol could
also be extended successfully to the preparation of the
vanillamides of hydroxy (1o, 1r , 2b-f) and bicarboxylic
acids (4a ,b ). Compared to DEPC, PPAA has several

Synthesis and Discovery of New Acyl Templates
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 17 3741
desirable attributes with respect to cost, toxicity, and
environmental factors, but its inherent lower reactivity
made it unsuitable with hindered acids, such as those
from the triterpenoid series. These could be amidated,
generally in rather low yield, only under DEPC promo-
tion (Table 1, entries 2b-e). Remarkably, reaction of
these acids with vanillamine using peptide coupling
reagents (e.g., BOP) failed completely, an observation
which underscores the synthetic potential of the mixed
anhydride protocol.
Monounsaturated N-AVAM were cyclopropanated us-
ing a diethylzinc-based Simmons-Smith protocol,14
while the vanillamide of eicosandioic acid (3c) was
prepared from the vanillamide of undecylenic acid (1p )
by homodimerization via cross metathesis15 and hydro-
genation.
acyl analogues. The development of an expeditious
protocol for acylating vanillamine was central to our
program, and in situ activation via mixed phosphoric
anhydrides proved to be successful in solving the
chemoselectivity of the reaction and avoiding unproduc-
tive protection/deprotection steps. An amidation protocol
based on commercially available and cheap reagents
(PPAA or DEPC) was optimized using unstable poly-
unsaturated acids (arachidonic and ximeninic acids) as
a probe and was then extended to a series of substrates
from the isoprenoid and fatty acid pools (Table 1).
The choice of the acids was inspired by the possibility
of evaluating conformational and covalent constraints
within the lipophilic domain of the acyl moiety of CPS.
This structural element has a high degree of conforma-
tional freedom, allowing the attainment of self-coiled
conformations.5 Instead, linear isoprenoid residues show
a strong tendency to adopt an extended conformation,18
a bias which is essentially solvent-insensitive,18 while
their lipophilic terpenoid skeleton can be covalently
constrained in a variety of cyclic frameworks. In the
event, replacement of the isodecenoic acyl residue of
CPS with geranic acid had little effect on activity (Table
1) despite the presence of further unsaturation and the
different connectivity of the C-10 chain (cf. 1a and 1e).
Homologation to farnesic acid caused a marked increase
in activity, affording “farvanil” (1f), one of the two most-
potent compounds observed in this study. Potency was
not increased further by additional prenylogation to
phytoic acid, a C-20 compound (cf. 1g). Reshaping of the
C-10 chain into a cyclopropane ring, as in the vanilla-
mide of chrysanthemic acid (1h ), decreased the potency
as compared to that of geranylvanillamine (1e) and so
did covalent constraint of the isoprenoid residue into a
cyclohexane ring, as exemplified by the vanillamide of
(S)-perillic acid (1i). Polyunsaturation within the iso-
prenoid moiety had a dramatic effect on activity because
the vanillamide of retinoic acid (retvanil, 1j) was more
efficacious as well as 1 order of magnitude more potent
than CPS. Polyunsaturation alone was not sufficient to
impart vanilloid properties as shown by the inactivity
of the VA of sorbic acid (1k ), an observation suggesting
that allylic methyl groups play an important role in the
interaction of linear lipophilic moieties with VR1.
Biologica l Eva lu a tion
The activity of N-AVAM on the CPS binding sites of
vanilloid receptors was evaluated by measuring the
entry of Ca2+ into human embryonic kidney HEK-293
cells overexpressing the human VR1,16 a typical VR1-
mediated effect. In this assay, CPS induces a dose-
dependent increase of intracellular calcium (pEC50
)
7.7) which is not observed in wild-type HEK-293 cells
and is blocked by the vanilloid receptor antagonist
capsazepine.17 This method allows the combined evalu-
ation of potency (as pEC50, where EC50 is the molar
concentration necessary to induce a relative half-
maximal response for each compound) and efficacy (as
a percent of the maximum possible absolute effect
obtained with ionomycin). These two parameters reflect
the affinity of a compound for the CPS binding site on
VR1 and its capacity to induce a conformational change
eventually translating into a functional response (i.e.,
calcium efflux), respectively. Although compound solu-
bility and permeability through the cell membrane can
affect efficacy and potency, respectively, at least the
former of these factors was kept under control by testing
the compounds in both methanol and dimethylsulfoxide
(DMSO). This did not lead to any observed significant
differences in efficacy, except for some of the polycyclic
di- and triterpenoid derivatives which were more ef-
ficacious in DMSO. This method is not biased by the
possible heterogenicity of native vanilloid receptors, a
typical drawback of the rodent preparations used in the
structure-activity studies carried out before the cloning
of VR1.4,6,8,10 Furthermore, this method can be carried
out under accurately defined conditions where param-
eters that normally influence the VR1-mediated calcium
response, such as cell number, degree of confluency,
presence of carrier protein, etc., can be tightly controlled
(Experimental Section). Finally, to demonstrate un-
equivocally the involvement of VR1 receptors in their
pharmacological actions, the most-potent N-AVAM in
this assay, that is, farvanil, rinvanil, and acetyl-rinvanil,
were also tested in untransfected wild-type HEK-293
cells and were found to be inactive on Ca2+ influx.
Resu lts a n d Discu ssion
Most in vitro studies of vanilloid activity were carried
out on rat dorsal root ganglia where a heterogeneous
population of binding sites for CPS exists;2 therefore, it
was interesting to revisit and expand the structure-
activity relationships that emerged in these studies
employing the cloned human vanilloid receptor and new
Constraint of the CPS acyl moiety into a polycyclic
structure gave puzzling results. Thus, the vanillamide
of abietic acid was almost inactive (2a ), as were those
of ursolic, 18â-glycyrrhetic, and cholic acids (2b-d ,
respectively). Nevertheless, the vanillamide of betulinic

3742 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 17
Appendino et al.
acid (2e), a lupane derivative, and that of ilicic acid (2f),
a bicyclic sesquiterpene, were quite efficacious, albeit 1
order of magnitude less potent than CPS.
lation (1q), while hydrogenation of the double bond, as
in 1r and its corresponding ketone 1s, reduced efficacy
but not potency. Taken together, these observations
suggest that a second pocket, capable of interacting with
ester groups, might also be present in the apolar cleft
accommodating the acyl residue of N-AVAM.
With the aim of dissecting the conformational and
polar contributions of unsaturations to the activity of
unsaturated fatty acid amides, a selection of monoun-
saturated acids were cyclopropanated. With all com-
pounds tested (cyclopropanated VA of oleic (1t), ricino-
leic (1u ), erucic (1v), and undecylenic (1w ) acids) an
increase of efficacy, but not of potency, was observed
(Table 1), supporting a conformational role for the
double bond of long-chain fatty acids. The increase in
efficacy induced by cyclopropanation was especially
marked for the vanillamide of undecylenic acid (1w ).
Cyclopropanation also mildly increased the activity of
phytoic acid vanillamide, a C-20 monounsaturated
linear terpenoid acid (cf. 1g and 1x).
Next, we investigated the activity of the vanillamides
of a series of fatty acids. Previous studies had estab-
lished that the vanilloid activity of linear fatty acids
peaks at C-9 and then fades with a further increase of
the chain length.6 Within C-18 acids, a double bond at
C-9 was shown to fully restore activity.6 The archetypal
vanillamide of fatty acids is olvanil (1b).6 Its efficacy,
but not its potency, could be increased by extension of
the side chain (VA of erucic acid, 1l), an effect that
cannot be observed by replacement of the double bond
with a triple bond (VA of stearolic acid, 1m ) or an enine
moiety (VA of ximeninc acid, 1n ) (Table 1). Introduction
of an allylic hydroxyl on the distal homoallylic position
(VA of ricinoleic acid, 1o) and truncation of the acyl
moiety at the distal olefin terminus (VA of undecylenic
acid, 1p ) also increased the efficacy without greatly
altering potency. The vanillamide of ricinoleic acid
(rinvanil, 1o) had potency comparable to that of retvanil
(1j),19 a finding that suggests the presence of polar
elements within the apolar pocket accommodating the
acyl residue of CPS. However, this polar element(s) does
not establish a hydrogen bond with the hydroxyl group
of rinvanil because the activity of the latter was
significantly increased, rather than reduced, by acety-
Because the VR is a multimeric receptor,2 we finally
investigated a series of dimeric vanillamides obtained
either by the bidirectional amidation of various dicar-
boxylic acids or by the dimerizative metathesis of
ω-unsaturated acids. The VA of both azelaic acid (3a )
and sebacic acid (3b) showed activity lower or compa-
rable to that of CPS, but increasing the chain length
led to a remarkable increase of efficacy, as exemplified
by the bis-vanillamide 3c (Table 1).

Synthesis and Discovery of New Acyl Templates
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 17 3743
Indeed, the olefinic precursor of 3c (compound 3d )
was also higly active, while the bis-vanillamides of exo-
cis- and trans-norbornene dicarboxylic acids (4a and 4b,
respectively) were almost inactive, showing that a long,
and possibly monounsaturated, aliphatic tether is nec-
essary for vanilloid activity of bicarboxylic acids.
particular, the VA of farnesic, retinoic, and ricinoleic
acids (farvanil, retvanil, and rinvanil; 1f, 1j, and 1o,
respectively) as well as acety-rinvanil (1q) qualify as
novel and powerful vanilloid leads, worth an in-depth
investigation of their structure-activity relationships.
Like olvanil (1c)6 and arvanil (1d ),7 these compounds
can be viewed as “unnatural natural products”,21 the
result of the artificial combination of natural building
blocks (vanillamine and fatty or isoprenoid acids) into
unnatural assemblies. Their interesting biomedical
profile validates the still largely untapped potential of
the expansion of the natural products pool by the
artificial combinations of natural building blocks and/
or the triggering of reactivity patterns apparently
overlooked by Nature.22
Exp er im en ta l Section
The results achieved from evaluating a series of
isoprenoid and fatty acids can be summarized as follows:
(1) Linear isoprenoic acids are excellent replacements
for the isodecenoyl moiety of CPS, and reshaping of the
isoprenoid residue to accommodate a monocyclic system
is tolerated. The polyunsaturated sesquiterpenoids far-
nesic acid and retinoic acid afforded the most-potent and
efficacious vanillamides, respectively (farvanil, 1f and
retvanil, 1j). The substantial inactivity of the VA of
sorbic acid (1k ) points to an important role for chain
length and the presence of methyl branchings.
(2) The VA of polycyclic di- and triterpenoid acids were
almost inactive with the notable exception of the VA of
betulinic acid (2e), a remarkable observation because
the VA of a closely related compound (ursolic acid, 2b)
was almost inactive.
(3) Within C-18 fatty acids, the substantial inactivity
on the VA of stearic acid (1b), previously established
using an in vivo assay,6 was also confirmed at the
receptor level. On the other hand, vanilloid activity
could be induced by midchain unsaturation, as in olvanil
(1c), and was tolerant to elongation (VA or erucic acid,
1l) and to the presence of an oxygen function on the
distal side of the double bond, as in rinvanil (1o).
Cyclopropanation of olvanil (1c) and other monounsat-
urated vanillamides (1o, 1l, and 1p ) led to increased
efficacy, suggesting that the double bond essentially has
a conformational effect, preorganizing the long chain for
bonding. Nevertheless, the increased efficacy and/or
potency of the oxygenated olvanil analogues rinvanil
(1o) and acetylrinvanil (1q) is a strong indication that
the lipophilic binding site of vanilloids also has elements
capable of interacting with a polar function, as con-
firmed by the activity of the non-olefinic analogues of
rinvanil 1r and 1s.
Ma ter ia ls. Commercially available reagents and solvents
were used without further purification. CH2Cl2 was dried by
distillation from CaH2, and THF was dried by distillation from
sodium benzophenone. Silica gel 60 (70-230 mesh, Macherey-
Nagel) was used for open column chromatography (CC). Unless
stated otherwise, the carboxylic acids used in this investigation
are commercially available in satisfactory purity (>95%) and
were purchased from Fluka-Aldrich or Alexis Biochemicals and
used as such. Commercial ricinoleic acid (ca. 80%, Fluka) was
purified by CC (hexanes/EtOAc gradient, from 9:1 to 8:2).
Farnesic,23 phytoic,24 perillic,25 dehydroricinoleic,26 and dihy-
drodehydroricinoleic26 acids were prepared from the corre-
sponding commercially available alcohols or aldehydes accord-
ing to literature. Chrysanthemic acid was prepared by the
hydrolysis (15 equiv of LiOH, 2:1 ratio of THF/water) of the
corresponding commercially available ethyl ester. Ximeninic,
ursolic, 18â-glycyrrhetic, betulinic, and abietic acids were a
generous gift from Indena, Milano, Italy. Ilicic acid was
available from previous phytochemical work.27
Biologica l Assa ys. Overexpression of human VR1 cDNA
into human embryonic kidney HEK-293 cells was carried out
as described previously.16 Cells were grown as monolayers in
a minimum essential medium supplemented with nonessential
amino acids, 10% fetal calf serum, and 0.2 mM glutamine, and
were maintained under 95%/5% O2/CO2 at 37 °C. The effect of
the substances on [Ca2+]I was determined by using Fluo-3, a
selective intracellular fluorescent probe for Ca2+ 7b
One day
.
prior to the experiments, almost confluent (70%) cells were
transferred into six-well dishes coated with poly-L-lysine
(Sigma) and grown in the culture medium mentioned above.
On the day of the experiment, the cells (50-60000 per well)
were loaded for 2 h at 25 °C with 4 µM Fluo-3 methyl ester
(molecular probes) in DMSO containing 0.04% Pleuronic. After
the cells were loaded, the cells were washed with Tyrode (pH
7.4), trypsinized, resuspended in Tyrode, and transferred into
the cuvette of the fluorescence detector (Perkin-Elmer LS50B)
under continuous stirring. The fact that fluorescence was
measured in a quartz cuvette allowed us to conduct experi-
ments in the absence of bovine serum albumin, which inhibits
the calcium response of N-AVAM as determined in previous
studies. Experiments were carried out by measuring cell
fluorescence at 25 °C (λEX ) 488 nm, λEM ) 540 nm) before
and after the addition of the test compounds at various
concentrations. Potency data are expressed as the -log of the
concentration exerting a half-maximal effect (pEC50), whereas
the efficacy of the effect was determined by comparing it to
the analogous effect observed with ionomycin (4 µM).
(4) ω-Dimerization of capsaicinoids leads to increased
activity, provided that the alkyl tether linking the two
vanillyl moieties is long enough, as in 3c and 3d . This
might be due to the simultaneous activation of two of
the four monomers that constitute the quaternary
structure of VR1 in expression systems,20 an interaction
requiring a definite distance between the vanillyl moi-
eties.
Acyla tion of Va n illa m in e. (A) DEP C P r otocol. Synthe-
sis of arvanil (1d ) as exemplificative. To a solution of high-
purity arachidonic acid (>98%, 262 mg, 0.86 mmol, Alexis
Biochemicals) in dry THF (2.5 mL) were added vanillamine
hydrochloride (196 mg, 1.03 mmol, 1.2 molar equiv), DEPC
(216 mg, 198 µL, 1.3 mmol, 1.5 molar equiv), and triethylamine
(309 µL, 224 mg, 2.06 mmol, 2.4 molar equiv). The solution
was stirred under nitrogen for 1.5 h. Volatiles were removed
Con clu sion s
Certain acids from the isoprenoid and unsaturated
fatty acid pools can provide highly active vanillamides,
whose activity on VR1 is comparable to that of other
vanilloid prototypes,4b-d including resiniferonoids.4a In

3744 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 17
Appendino et al.
(2) For recent reviews on vanilloids, see: (a) Szallasi, A. Vanilloid
Receptor Ligands. Hopes and Realities for the Future. Drugs
Aging 2001, 18, 561-573. (b) Szallasi, A.; Blumberg, P. M.
Vanilloid (Capsaicin) Receptors and Mechanisms. Pharmacol.
Rev. 1999, 51, 159-211.
under reduced pressure, and the residue was adsorbed on silica
gel 60 (ca. 1 g) and purified by open column chromatography
(5 g silica gel, 8:2 ratio of hexanes/EtOAc as eluant), affording
360 mg (95%) of arvanil (1d ) as a colorless oil. (B) P P AA
P r otocol. Synthesis of rinvanil (1o) as exemplificative. To a
cooled (0 °C) solution of ricinoleic acid (500 mg, 1.67 mmol) in
dry CH2Cl2 (5 mL) were added vanillamide hydrochloride (318
mg, 1.67 mmol, 1 molar equiv), triethylamine (950 µL, 690 mg,
3.3 mmol, 2 molar equiv), and PPAA (50% EtOAc solution, 1.2
mL, 3.3 mmol, 2 molar equiv). After the mixture was stirred
for 15 min at 0 °C, the ice bath was removed, and the reaction
mixture was allowed to warm to room temperature in ca. 1 h.
Volatiles were removed under reduced pressure, and the
residue was adsorbed on silica gel 60 (ca. 2 g) and purified by
open column chromatography (15 g silica gel, 3:7 ratio of
petroleum ether/EtOAc as eluant) to afford 509 mg (70%) of
rinvanil (1o) as a colorless oil.
Cyclop r op a n a tion of Mon ou n sa tu r a ted Va n illa m id es.
Synthesis of 1t as exemplificative. To a solution of olvanil (1c,
200 mg, 0.48 mmol) in dry benzene (20 mL) were added Et2-
Zn (1 M in THF, 7.2 mL, 7.2 mmol, 15 molar equiv) and
diiodomethane (0.58 mL, 7.2 mmol, 15 molar equiv) dropwise
at room temperature. The reaction mixture was then heated
at 65 °C for 6 h and worked up by cooling, acidification with
2 N H2SO4, and extraction with EtOAc. The organic phase was
washed with saturated NaHCO3 and brine, dried (Na2SO4),
and evaporated. The residue was purified by CC (10 g silica
gel, petroleum ether/EtOAc gradient). Fractions eluted with
a 6:4 ratio of petroleum ether/EtOAc afforded 94 mg (45%) of
1t as a colorless oil.
Dim er iza tive Meta th esis of Un d ecylen yl Va n illa m id e.
To a solution of 1p (310 mg, 0.97 mmol) in dry CH2Cl2 (3.5
mL) was added Grubbs catalyst (8 mg, 0.0097 mmol, 0.1 molar
equiv). The reaction mixture was stirred under nitrogen at
room temperature for 3 days, and then worked up by evapora-
tion and chromatography (8 g silica gel, petroleum ether/EtOAc
gradient). Fractions eluted with a 2:8 ratio of petroleum ether/
EtOAc gave 100 mg (25%) of 3d as a powder.
Ca ta lytic Hyd r ogen a tion of th e Dim er 3d . The crude,
unchromatographed residue from the dimerizative metathesis
of the vanillamide of undecylenic acid (310 mg) was dissolved
in MeOH (3 mL), and 10% Pd(C) (10 mg) was added. The
suspension was stirred for 3 h under a hydrogen atmosphere
(balloon), and then worked up by filtration. The residue was
purified by CC (8 mL silica gel, 1:9 ratio of petroleum ether/
EtOAc as eluant) to afford 85 mg of 3c (21% from 1p ) as an
off-white powder.
(3) For a review, see: Wrigglesworth, R.; Walpole, C. S. J . Capsaicin-
Like Agonists. Drugs Future 1998, 23, 531-538.
(4) (a) Walpole, C. S.; Bevan, S.; Bloomfield, G.; Breckenridge, R.;
J ames, I. F.; Richie, T.; Szallasi, A.; Winter, J .; Wrigglesworth,
R. Similarities and Differences in the Structure-Activity Rela-
tionships of Capsaicin and Resiniferatoxin Analogues. J . Med.
Chem. 1996, 39, 2939-2952. (b) Lee, J .; Lee, J .; Kim, J .; Kim,
S. Y.; Chun, M. W.; Cho, H.; Hwang, S. W.; Oh, U.; Park, Y. H.;
Marquez, V. E.; Beheshti, M.; Szabo, T.; Blumberg, P. M. N-(3-
Acyloxy-2-benzylpropyl)-N′-(4-hydroxy-3-methoxybenzyl)thio-
urea Derivatives as Potent Vanilloid Receptor Agonists and
Analgesics. Bioorg. Med. Chem. 2001, 9, 19-32. (c) Park, N. S.;
Seong, C. M.; J ung, Y. S.; Kim, W. B.; Kim, S. H. DA-5018
(Capsavanil). Drugs Future 2000, 25, 1131-1137. (d) Urban, L.;
Campbell, E. A.; Panesar, M.; Patel, S.; Chaudhry, N.; Kane,
S.; Buchheit, K. H.; Sandells, B.; J ames, I. F. In Vivo Pharma-
cology of SDZ 249-665, a Novel, Non-pungent Capsaicin Ana-
logue. Pain 2000, 89, 65-74.
(5) For
a diffractometric study on CPS, see: David, W. I. F.;
Shankland, K.; Shankland, N. Routine Determination of Molec-
ular Crystal Structures from Powder Diffraction Data. Chem.
Commun. 1998, 931-932.
(6) J anusz, J . M.; Buckwalter, B. L.; Young, P. A.; LaHann, T. R.;
Farmer, R. W.; Kasting, G. B.; Loomans, M. E.; Kerchaert, G.
A.; Maddiin, C. S.; Berman, E. F.; Bohne, R. L.; Cupps, T. L.;
Milstein, J . R. Vanilloids. 1. Analogs of Capsaicin with Anti-
nociceptive and Antiinflammatory Activity. J . Med. Chem. 1993,
36, 2595-2604.
(7) (a) Melck, D.; Bisogno, T.; De Petrocellis, L.; Chuang, H.-H.;
J ulius, D.; Bifulco, M.; Di Marzo, V. Unsaturated Long-Chain
N-Acyl-vanillyl-amides (N-AVAM): Vanilloid Receptor Ligands
that Inhibit Anandamide-Facilitated Transport and Bind to CB1
Cannabinoid Receptors. Biochem. Biophys. Res. Commun. 1999,
262, 275-284. (b) De Petrocellis, L.; Bisogno, T.; Davis, J . B.;
Pertwee, R. G.; Di Marzo, V. Overlap Between the Ligand
Recognition Properties of the Anandamide Transporter and the
VR1 Vanilloid Receptor: Inhibitors of Anandamide Uptake with
Negligible Capsaicin-like Activity. FEBS Lett. 2000, 483, 52-
56.
(8) Walpole, C. S. J .; Wrigglesworth, R.; Bevan, S.; Campbell, E.
A.; Dray, A.; J ames, I. F.; Masdin, K. J .; Perkins, M. N.; Winter,
J . Analogues of Capsaicin with Agonist Activity as Novel
Analgesic Agents: Structure-Activity Studies. 3. The hydro-
phobic Side-Chain “C-Region”. J . Med. Chem. 1993, 36, 2381-
2389.
(9) Reaction of equimolecular amounts of vanillamine and oleic acid
in the presence of DCC (1 molar equiv) and DMAP (0.2 molar
equiv) gave, apart from unreacted vanillamine, a ca. 1.5:1
mixture (
1H NMR analysis) of olvanil and O-oleylolvanil.
(10) Walpole, C. S. J .; Wrigglesworth, R.; Bevaqn, S.; Campbell, E.
A.; Dray, A.; J ames, I. F.; Masdin, K. J .; Perkins, M. N.; Winter,
J . Analogues of Capsaicin with Agonist Activity as Novel
Analgesic Agents: Structure-Activity Studies. 2. The Amide
Bond “B-Region”. J . Med. Chem. 1993, 36, 2373-2380.
(11) Shioiri, T.; Yokoyama, Y.; Kasai, Y.; Yamada, S. Reaction of
Diethyl Phosphorocyanidate (DEPC) with Carboxylic Acids. A
New Synthesis of Carboxylic Esters and Amides. Tetrahedron
1976, 32, 2211-2217.
Ack n ow led gm en t. We thank MURST (Fondi ex
40% to G.A. and Grant 3933 to V.D.M.) and INDENA
for financial support. We are grateful to Dr. J ohn Davis
(GlaxoSmithKline, Harlow, U.K.) for the gift of HEK
cells transfected with human VR1 cDNA and to Dr.
Bruno Gabetta (Indena, Milano, Italy) for a generous
gift of abietic, ximeninic, ursolic, 18â-glycyrrhetic, and
betulinc acids. We are grateful to Dr. Gian Cesare Tron
(Universita` del Piemonte Orientale) for his help in the
early steps of this investigation.
(12) Schwarz, M. n-Propane Phosphonic Acid Anhydride - A Con-
densation Reagent. Synlett 2000, 1369.
(13) Appendino, G.; Minassi, A. Unpublished data.
(14) Furukawa, J .; Kawabata, N.; Nishimura, J . Synthesis of Cyclo-
propanes by the Reaction of Olefins with Dialkylzinc and
Methylene Iodide. Tetrahedron 1968, 24, 53-58.
(15) For a review article on olefin metathesis, see: Fu¨rstner, A. Olefin
Metathesis and Beyond. Angew. Chem., Int. Ed. 2000, 39, 3012-
3043.
Su p p or t in g In for m a t ion Ava ila b le: General methods
1
and characterization (mp, IR, MS, and H and 13C NMR data)
(16) Hayes, P.; Meadows, H. J .; Gunthorpe, M. J .; Harries, M. H.;
Duckworth, D. M.; Cairns, W.; Harrison, D. C.; Clarke, C. E.;
Ellington, K.; Prinjha, R. K.; Barton, A. J . L.; Medhurst, A. D.;
Smith, G. D.; Topp, S.; Murdock, P.; Sanger, G. J .; Terrett, J .;
J enkins, O.; Benham, C. D.; Randall, A. D.; Gloger, I. S.; Davis,
J . B. Cloning and Functional Expression of a human Orthologue
of Rat Vanilloid Receptor-1. Pain 2000, 88, 205-215.
(17) Walpole, C. S. J .; Bevan, S.; Bovermann, G.; Boelsterli, J . J .;
Breckenridge, R.; Davies, J . W.; Hughes, G. A.; J ames, I.; Oberer,
L.; Winter, J .; Wrigglesworth, R. The Discovery of Capsazepine,
the First Competitive Antagonist of the Sensory Neuron Exci-
tants Capsaicin and Resiniferatoxin. J . Med. Chem. 1994, 37,
1942-1954.
of all new compounds. This material is available free of charge
via the Internet at http://pubs.acs.org.
Note Ad d ed a fter ASAP P ostin g. This manuscript
was released ASAP on 7/9/2002 with minor errors in
the structures of 1d and 1k . The correct version was
posted on 8/8/2002.
Refer en ces
(1) Suzuki, T.; Iwai, K. Constituents of Red-Pepper Species: Chem-
istry, Biochemistry, Pharmacology, and Food Science of the
Pungent Principle of Capsicum Species. In The Alkaloids; Brossi,
A., Ed.; Academic Press: Orlando, FL, 1984; Vol. 23, pp 227-
299.
(18) (a) Zahn, T. J .; Eilers, M.; Guo, Z.; Ksebati, M. B.; Kimon, M.;
Scholten, J . D.; Smith, S. O.; Gibbs, R. A. Evaluation of
Isoprenoid Conformation in Solution and in the Active Site of

Synthesis and Discovery of New Acyl Templates
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 17 3745
Protein-Farnesyl Transferase Using Carbon-13-labeling in Con-
junction with Solution and Solid-State NMR. J . Am. Chem. Soc.
2000, 122, 7153. (b) Green, N. S.; Miller, M. M.; Houk, K. N.
Chemtracts 2000, 13, 749-756.
machines to build completely new ones, can be traced back to
Franc¸ois J acob (J acob, F. Evolution and Tinkering. Science 1977,
196, 1161-1166; see also: Keeling, P. J . Parasites go to the Full
Monty. Nature 2001, 414, 401).
(19) Ricinoleic acid shows antinociceptive properties resembling those
of CPS, but direct binding to VR1 was not reported: Vieira, C.;
Evangelista, S.; Cirillo, R.; Terracciano, R.; Lippi, A.; Maggi, C.
A.; Manzini, S. Antinociceptive Activity of Ricinoleic Acid, a
Capsaicin-like Compound Devoid of Pungent Properties. Eur.
J . Pharmacol. 2000, 407, 109-116.
(20) (a) Kedei, N.; Szabo, T.; Lile, J . D.; Treanor, J . J .; Olah, Z.;
Iadarola, M. J .; Blumberg, P. M. Analysis of the Native Qua-
ternary Structure of Vanilloid Receptor 1. J . Biol. Chem. 2001,
276, 28613-28619. (b) Kuzhikandathil, E. V.; Wang, H.; Szabo,
T.; Morozova, N.; Blumberg, P. M.; Oxford, G. S. Functional
Analysis of Capsaicin Receptor (Vanilloid Receptor Subtype 1)
Multimerization and Agonist Responsiveness Using a Dominant
Negative Mutation. J . Neurosci. 2001, 21, 8697-8706.
(21) Schreiber, S. L. Target-Oriented and Diversity-Oriented Organic
Synthesis in Drug Discovery. Science 2000, 287, 1964-1969.
(22) Appendino, G.; Tron, G. C.; J arevang, T.; Sterner, O. Unnatural
Natural Products from the Transannular Cyclization of Lathyr-
ane Diterpenes. Org. Lett. 2001, 3, 1609-1612. The germane
image of evolution as a tinkerer, using bits and pieces of existing
(23) Caliezi, A.; Schinz, H. Zur Kenntnis der Sesquiterpene und
Azulene. Die Cyclisation der Farnesylsa¨ure. Helv. Chim. Acta
1949, 32, 2556-2560.
(24) Ronayne, J .; Williams, D. H. Solvent Effects in Nuclear Magnetic
Resonance Spectroscopy. Part XIII. Solvent Shifts Induced by
Benzene in Some R,â-Unsaturated Compounds. An Aid to the
Determination of Stereochemistry. J . Chem. Soc. C 1967, 2642-
2645.
(25) Hardcastle, I. R.; Rowlands, M. G.; Barber, A. M.; Grimshaw,
R. M.; Mohan, M. K.; Nutley, B. P.; J arman, M. Inhibition of
Protein Prenylation by Metabolites of Limonene. Biochem.
Pharmacol. 1999, 57, 801-809.
(26) Nichols, J .; Schipper, E. Keto Fatty Acids Derived from Castor
Oil. I. Unsaturated Acids. J . Am. Chem. Soc. 1958, 80, 5705-
5710.
(27) Appendino, G.; Calleri, M.; Chiari, G.; Gariboldi, P.; Menichini,
F. Structure of Graveolide and Conformational Aspects of
Related Helenanolides. Gazz. Chim. Ital. 1986, 116, 637-641.
J M020844O