Opioid receptor probes derived from cycloaddition of the hallucinogen natural product salvinorin A

Article abstract of DOI:10.1021/np1007872

As part of our continuing efforts toward more fully understanding the structure-activity relationships of the neoclerodane diterpene salvinorin A, we report the synthesis and biological characterization of unique cycloadducts through [4+2] Diels-Alder cyc

Full text of DOI:10.1021/np1007872

ARTICLE
pubs.acs.org/jnp
Opioid Receptor Probes Derived from Cycloaddition of the
Hallucinogen Natural Product Salvinorin A
Anthony Lozama,‡ Christopher W. Cunningham,§ Michael J. Caspers,§ Justin T. Douglas,^
Christina M. Dersch,|| Richard B. Rothman,|| and Thomas E. Prisinzano*,§
‡Division of Medicinal & Natural Products Chemistry, University of Iowa, Iowa City, Iowa 52242, United States
§Department of Medicinal Chemistry, University of Kansas, Lawrence, Kansas 66045, United States
^Molecular Structures Group, University of Kansas, Lawrence, Kansas
Clinical Psychopharmacology Section, IRP, NIDA, DHHS, Baltimore, Maryland 21224, United States
S Supporting Information
b
ABSTRACT: As part of our continuing efforts toward more
fully understanding the structure-activity relationships of the
neoclerodane diterpene salvinorin A, we report the synthesis
and biological characterization of unique cycloadducts
through [4þ2] Diels-Alder cycloaddition. Microwave-as-
sisted methods were developed and successfully employed,
aiding in functionalizing the chemically sensitive salvinorin A
scaffold. This demonstrates the first reported results for both cycloaddition of the furan ring and functionalization via microwave-
assisted methodology of the salvinorin A skeleton. The cycloadducts yielded herein introduce electron-withdrawing substituents
and bulky aromatic groups into the C-12 position. Kappa opioid (KOP) receptor space was explored through aromatization of the
bent oxanorbornadiene system possessed by the cycloadducts to a planar phenyl ring system. Although dimethyl- and
diethylcarboxylate analogues 5 and 6 retain some affinity and selectivity for KOP receptors and are full agonists, their aromatized
counterparts 13 and 14 have reduced affinity for KOP receptors. The methods developed herein signify a novel approach toward
rapidly probing the structure-activity relationships of furan-containing natural products.
alvia divinorum Epling & Jꢀativa (Lamiaceae) is a hallucino-
genic mint species found in Oaxaca, Mexico, and recently in
therapeutics and analgesics. On the basis of the potential
therapeutic applications, we and others have begun to investigate
the structure-activity relationships (SAR) of 1.10-12
S
parts of California. Traditionally, S. divinorum has been used by
the natives of Oaxaca for their divination ceremonies, along with
the treatment of headaches, rheumatism, and abdominal swelling.1
The main active component of S. divinorum has been identified as
the neoclerodane diterpene salvinorin A (1).2,3 While many
neoclerodane diterpenes have been found to possess biological
activity, 1 was found to be a potent hallucinogen in humans with
an active dose between 200 and 500 μg.4 Furthermore, 1 was
found to be a full agonist at κ-opioid (KOP) receptors despite
having no structural similarity to other known KOP receptor
ligands, such as the benzomorphinan cyclazocine or the arylace-
tamide U50,488.5 The potential for abuse has led to the regula-
tion of S. divinorum and 1 in several U.S. states and a growing
number of countries abroad. However, the unique characteristics
of 1 have also helped spark significant scientific interest as well.
Even though 1 is a potent hallucinogen, it does not have any
affinity for the 5-HT2A receptor,5 which is the target receptor for
classical hallucinogens such as LSD, mescaline, and psilocin.
Experiments with 1 in rats have demonstrated that 1 can block
the locomotor effects of cocaine6,7 along with having opioid-
mediated antinociceptive properties,8,9 providing evidence for
the potential utility of 1 and related analogues as stimulant abuse
To date, the majority of analogues of 1 prepared and evaluated
have explored the role of the C-2 acetoxy group. These investiga-
tions have found that the appropriate substituent in this position
may alter selectivity for opioid receptors (2), as well as increase
potency and extend duration of action in vivo (3 and 4).13-15 By
comparison, the role of the furan ring is less understood.
Additional studies are necessary due to the implication of the
furan ring in binding at KOP receptors by several modeling
studies,16,17 as well as the potential for furan-containing natural
products to be hepatotoxic upon bioactivation by various CYP450
enzymes.18-23 In an effort to circumvent potential hepatotoxicity
and provide further insight into the SAR of 1, furan-modified
analogues were sought.
In the development of additional SAR for the furan ring
present in 1, two approaches were considered. The first approach
considered was the incorporation of conformational constraint.
This strategy is often a fruitful technique and has been used
previously for the interactions of natural products at receptors.24
In our case this approach is synthetically challenging and would
Received: October 30, 2010
Published: February 21, 2011
r
Copyright
2011 American Chemical Society and
718
dx.doi.org/10.1021/np1007872 J. Nat. Prod. 2011, 74, 718–726
American Society of Pharmacognosy
|

Journal of Natural Products
ARTICLE
Scheme 1a
a Reagents and conditions: (a) Appropriate alkyne, toluene, reflux; (b)
2-(trimethylsilyl)phenyl trifluoromethane sulfonate, CsF, CH3CN,
room temperature.
require the development of a new or modification of an existing
synthesis of 1;25,26 thus this approach was not undertaken. The
alternative approach followed was to incorporate additional
steric bulk directly to the natural product. This approach has
been used successfully with opium alkaloids27 and is particularly
informative in cases where the X-ray crystal structure of a given
molecular target, such as the KOP receptor, is unavailable.
Furthermore, this approach was expected to provide details in
a more timely manner.
Furan rings have demonstrated themselves as versatile build-
ing blocks for the construction of molecules in semi- and total
synthesis.28 One of the more popular reactions that furan rings
undergo is the Diels-Alder reaction where furan rings act as the
4π diene. Thus, we sought to explore the binding requirements
of the KOP receptor through the incorporation of steric bulk into
1 through the utilization of Diels-Alder reactions. Our results
are presented below.
very well be due to the considerable difference in oxygen func-
tionality between these highly structurally dissimilar secondary
metabolites.
With the failure of the alkene-type dienophiles, alkynes were
explored, as they have been reported to successfully form
cycloadducts with furan rings.33 Additionally, an advantage to
using alkynes in the Diels-Alder reaction is the inability to form
additional stereoisomers with respect to the substituents on the
existing dienophile (exo vs endo). The reaction of diethyl
acetylenedicarboxylate with 1 at reflux over the course of two
days produced cycloadduct 5 as the major product (Scheme 1).
While Diels-Alder reactions have been employed for the total
synthesis of neoclerodane diterpenes, to our knowledge, this is
the first example of a neoclerodane diterpene undergoing chemical
modification via the Diels-Alder reaction.
’ RESULTS AND DISCUSSION
1
Investigation of the NMR spectra illustrated that the H
Synthesis. Initially, maleic anhydride and maleimide were
reacted with 1 in an attempt to form the corresponding adducts,
as both dienophiles have been found in the literature to react readily
with furan rings.28 However, after employing several different
solvents (THF, ether, CH2Cl2, toluene) no reaction with the
dienophiles maleic anhydride and maleimide was noted. Several
other well-known dienophiles including benzoquinone, dimethyl
maleate, dimethyl fumarate, diethyl maleate, diethyl fumarate,
methyl vinyl ketone, di-tert-butyl diazo-dicarboxylate, diethyl
diazo-dicarboxylate, and dibenzyl diazo dicarboxylate were ex-
plored in a variety of reaction conditions. However, none proved
successful in generating the desired cycloadducts.
Failure of the cycloaddition reaction to proceed may be
attributed to reduced reactivity of the furan ring in 1. Although
a 1999 publication reported a noncatalyzed Diels-Alder cycliza-
tion between maleic anhydride and the furan-containing labdane
diterpenoid hedychenone,29 a subsequent detailed study30 de-
proton signals, which correspond to C-15 (dd at δH 7.39) and
H-16 (multiplet at δH 7.41) and 13C signals for C-15 (δC 143.7)
and C-16 (δC 139.4) of 1, were no longer present in cycloadduct
5 (Scheme 1). Signals for the ethoxy groups were readily apparent
from the six-proton multiplet at δH 1.32 and four-proton multi-
plet at δH 4.28. Additionally, two new signals representing
bridgehead protons were found at δH 5.66 (dd) and 5.60
(doublet) along with their corresponding 13C shifts at δC
85.55 and 85.46. The remaining signals in the 13C spectra were
found to be indicative of two carbonyls and four sp2 carbons,
which fit the expected shifts of the newly introduced R,β-
unsaturated esters via the Diels-Alder mechanism. These data
are consistent with formation of a new oxanorbornadiene ring
system in 5. The regiochemistry of the oxygen bridge is depen-
dent on the facial orientation between the alkyne and the furan
ring of 1 (Figure 1). The production of diastereomers is possible,
and while HPLC analysis shows a single peak, we are unable to
determine the relative configuration at this time. However,
additional efforts toward this goal are underway and will be
reported in due course.
When reacted under similar conditions, dimethyl acetylene
dicarboxylate and 1 also successfully form cycloadduct 6. There
remained a caveat with the formation of cycloadducts 5 and 6,
specifically that prolonged exposure to heat would cause the
cycloadduct to undergo the retro Diels-Alder reaction. This was
proven by heating the pure compound 5 in fresh toluene, which
yielded a distribution of 5 and 1 upon isolation. Interestingly,
retro Diels-Alder reactions did not result in the salvinorin-
alkyne and 3,4-substituted furan ring, as might have been
predicted; rather, the retro [4þ2] resulted in a return of 1 and
scribed the necessity for Lewis acid catalysis using BF3 OEt2 and
3
high temperatures, conditions that we anticipated would surely
result in prominent degradation of 1.31 Attempts using AlCl3 and
TiCl4 to further activate the dienophiles still failed to produce
cyclization, with degradation seeming to dominate. One possible
explanation for this may be the presence of multiple oxygen
moieties in 1 competing for coordination with the Lewis acid,
resulting in degradation. It was thought that while HfCl4 is not as
reactive as other Lewis acids,32 it may be less likely to coordinate
with the oxygens of 1, thus circumventing the issues seen with
AlCl3; however, HfCl4 also failed to catalyze cycloadduct formation.
The reason that 1 failed to cyclize under the Lewis acid catalyzed
conditions required to induce cyclization in hedychenone could
719
dx.doi.org/10.1021/np1007872 |J. Nat. Prod. 2011, 74, 718–726

Journal of Natural Products
ARTICLE
Scheme 2a
Figure 1. Formation of oxanorbornadiene analogues.
Table 1. Microwave Reaction Conditions for the Synthesis of
Cycloadducts 5 and 6
time (min) absorbance temp (ꢀC)
solvent
product yield %
a Reagents and conditions: (a) Br2, CH2Cl2; (b) KOt-Bu, THF, furan or
2-methyl furan; (c) 1, toluene, reflux.
30
60
30
60
30
60
30
60
30
60
30
60
normal
normal
low
50
50
dioxane
dioxane
dioxane
dioxane
dioxane
dioxane
toluene
toluene
toluene
toluene
toluene
toluene
5 or 6
5 or 6
5 or 6
5 or 6
5 or 6
5 or 6
5 or 6
5 or 6
5 or 6
5 or 6
5 or 6
5 or 6
21
23
the reduced activity of the furan ring of 1, allowing cycloadduct
formation to occur.
100
100
100
100
50
62
low
64
To further explore the reactivity of the furan ring in 1, its
reaction with benzyne was investigated. There are several studies
that have demonstrated that benzyne can be effectively trapped
by furan rings to form the corresponding cycloadduct.34-36
Initial attempts to form the benzyne in situ using anthranilic
acid and isoamyl nitrite and trap it with the furan of 1 were
unsuccessful. However, it was found that when 1 was treated with
benzyne generated from 2-(trimethylsilyl)phenyl trifluorometh-
anesulfonate and cesium fluoride in acetonitrile at room tempera-
ture, cycloadduct 7 was formed in 24% overall yield (Scheme 1).35
The structure was confirmed through HRESIMS along with
NMR experiments including 1H NMR, 13C NMR, and
135DEPT. As with compounds 5 and 6, the relative config-
uration of the newly formed oxygen bridge in 7 has not been
determined.
While the furan of 1 can participate in Diels-Alder reactions,
it seems to react optimally only with very reactive/electron-poor
dienophiles, which allows for proper orbital overlap. Potentially
the linearity of the alkynes is allowing for the required p-orbital
overlap to overcome the activation energy required for reaction,
whereas the alkenes would be hindered from the proper overlap
due to the positioning over the salvinorin core for the favored
endo conformation to form. To further investigate this phenom-
enon, we investigated the reaction of 1 with 3,6-epoxy-3,6-
dihydrotribenzocycloheptatrienone (10a).37,38 Trienone 10a
and similar compounds have been shown to readily react in
Diels-Alder reaction with furans with the trienone serving as the
dienophile (Scheme 2). The added advantage of selecting these
compounds is that they exhibit mild blue light-emitting effects
and are direct precursors to a recently described class of fluor-
escent ligands.39 In addition to serving as dienophiles that would
add steric bulk to 1, they may also provide new avenues toward
fluorescent neoclerodanes, potentially helping elucidate how 1
interacts at opioid receptors.
normal
normal
normal
normal
low
70
70
24/15
30
50
100
100
100
100
55
low
55
normal
normal
67
67
the starting alkyne. With the prolonged times needed for cyclo-
adduct formation to occur, it was thought that this phenomenon
may be affecting overall yields. In an effort to avoid this retro
reaction and to optimize Diels-Alder reactions with 1 and alkynes,
microwave irradiation was utilized. Several solvents were screened
for reaction in the microwave based on their ability to solubilize
reagents and polarity, including toluene, xylene, benzene, and
dioxane. While various times and absorbance levels were explored,
optimal conditions were found to involve either the use of dry
and degassed toluene or dioxane at 100 ꢀC for 30 min (Table 1).
With optimal microwave conditions found, several other
alkynes were reacted with 1 in an attempt to form cycloadducts
including methyl propiolate, methyl 2-butynoate, methyl phenyl
propiolate, and acetylene dicarboxylate. Additionally, the alkene-
type dienophiles (maleic anhydride, maleimide, diazo dicarbox-
ylates, benzoquinone, dimethyl maleate, dimethyl fumarate,
diethyl maleate, diethyl fumarate) were reacted with 1 under
the defined microwave conditions; unfortunately, no cycload-
ducts incorporating the desired moieties were observed. Failure
of the alkynes to form cycloadducts may be due to the presence of
only one electron-withdrawing group. Diethyl and dimethyl
acetylenedicarboxylate have their sp carbons flanked by electron-
withdrawing groups, making the alkyne dienophile increasingly
electron poor, which may contribute to its ability to overcome
720
dx.doi.org/10.1021/np1007872 |J. Nat. Prod. 2011, 74, 718–726

Journal of Natural Products
ARTICLE
The reaction of dibenzosuberenone (8) with Br2 in CH2Cl2 at
0 ꢀC yielded the dibromo intermediate (9), which was treated
with KOt-Bu and either furan or 2-methylfuran in THF to
produce compounds 10a and 10b (Scheme 2). Reaction of 1
with 10a under standard reflux conditions yielded cycloadduct
11 as a mixture of C-15/C-16 exo and endo isomers. The
suggestion of a mixture of endo and exo isomers was corroborated
by the presence of a doubling of signals in both the 1H NMR and
13C NMR spectra as well as the corresponding mass of cycload-
duct 11 observed by HRESIMS. While cycloadduct 11 was
obtained as a mixture of exo and endo products, cycloadduct 12
was obtained as a single compound as seen from the lack of signal
approach from the face of the furan opposite the tricyclic core
of 1, as depicted by the red arrow in Figure 1, causing the methyl
group to be oriented as shown in Scheme 2. Although we were
unable to verify this with a crystal structure, analysis of the 3JH,H
coupling values (for details, see the Supporting Information)
seen in the 1H NMR spectrum of 12 supports the hypothesis that
the final product exhibits an endo,exo configuration. The results
seen with the reactions of 10a and 10b with 1 appear to be in
agreement with those described previously,40 wherein the inclu-
sion of alkyl substituents at the bridgehead carbons caused a shift
in endo/exo preference (as seen in 12), with nonsubstituted
reagents showing no selectivity (as seen in 11).
1
doubling in the H NMR and 13C NMR spectra. HRESIMS
The synthesis of cycloadducts 5-7 not only provided analo-
gues that introduced steric bulk at the furan position of 1, but also
provided useful synthetic intermediates. Treatment of com-
pound 5 with Fe2(CO)9 caused a reductive elimination of water
and consequent aromatization to produce compound 13
(Scheme 3).41 Analysis of HRESIMS data verified the expected
analysis corresponded to a molecular formula that matched the
predicted formula of 12.
To determine the location of the methyl group and if the
cycloadduct was endo or exo regarding its orientation, extensive
NMR experiments were conducted. The first step was to
determine whether C-43 was proximal or distal (cis or trans) to
1
mass of the new compound, and examination of H and 13C
the core of 1. HMBC experiments were able to show 1H and 13
C
spectra verified that the oxygen bridge present in 5 was indeed
couplings to trace the carbon framework of 12, showing that
C-43 is in fact proximal or cis to the salvinorin core (Figure S4).
This also helped to establish the carbon framework of 12, which
was found to be in agreement with the proposed product of the
Diels-Alder reaction. With regard to the assignment of endo or
exo configuration, it was anticipated that steric effects would force
cycloaddition to occur through an endo,exo approach. As de-
scribed previously,40 an endo,exo approach eliminates unfavor-
able steric interactions between dienophile and diene, resulting
in formation of a single product. This would cause 10b to
absent. Furthermore, new aromatic signals were observed in both
1
the H and 13C spectra that were in good agreement with the
proposed new structure. This reaction also proved successful in
transforming compounds 6 and 7 to their corresponding aro-
matic rings at the C-12 position, 14 and 15, respectively. These
compounds are significant from a synthetic standpoint, as they
are the first reported examples of non-heterocyclic aromatic rings
directly attached to the core of neoclerodane diterpenes.
Finally, we sought to explore the reactivity of the furan ring
with a nonsymmetrical alkyne. The reaction of ethyl 4,4,4-trifluoro-
2-butynoate with 1 at 95 ꢀC gave a mixture of 16a and 16b
(Scheme 4) in a 91:9 ratio, as determined by analytical HPLC.
While our initial predictions assumed a mixture of regioisomers
would form, this hypothesis was realized upon careful examina-
tion of multiple NMR spectra, including HSQC, DEPT135, and
COSY NMR. Particularly, there was a doubling of signals realized
for the sp2-hybridized methine at C-14, which appeared as a
doublet of doublets instead of an anticipated doublet. These
signals were coupled with two carbon signals at δC 84.42 and
85.01, which corresponds to the western bridgehead carbon (C-15).
Scheme 3a
3
Additionally, simultaneous JCF coupling was seen with both
bridgehead carbons C-15 and C-16 in the fluorine-coupled
carbon spectrum, which suggests that the CF3 group was attached
to both C-26 and C-27 at the same time. This could be explained
by the sample containing a mixture of regioisomers 16a and 16b.
Our attempts to separate 16a and 16b on a preparative scale
a Reagents and conditions: (a) Fe2(CO)9, toluene, reflux.
Scheme 4a
a Reagents and conditions: (a) ethyl 4,4,4-trifluoro-2-butynoate, toluene, reflux; (b) Fe2(CO)9, toluene, reflux.
721
dx.doi.org/10.1021/np1007872 |J. Nat. Prod. 2011, 74, 718–726

Journal of Natural Products
ARTICLE
using column chromatography under various solvent conditions
were unsuccessful, meaning that the unambiguous determination
of 16a and 16b was not possible. The preference to form
cycloadduct 16a over 16b can potentially be explained by the
difference in electron-withdrawing groups that flank the alkyne.
We propose that the stronger electron-withdrawing CF3 group
promotes a more electron-rich end to the alkyne that pairs to
C-15 due to the inductive push of electron density to C-16 from
the core of 1.
Given our success with converting oxanorbornadienes to the
corresponding phenyl rings, we decided to convert the mixture of
16a and 16b to their corresponding phenyl derivatives. It was
envisioned that a mixture of the phenyl derivatives would be
more easily separated than the mixture of oxanorbornadienes. As
expected, the treatment of the mixture of 16a and 16b with
Fe2(CO)9 in toluene gave the corresponding phenyl rings. To
our delight, 17a and 17b (Scheme 4) were isolated in 62% and
4% yield, respectively. The structure of 17a was elucidated in a
fashion similar to 16a and 16b by first identifying the three
aromatic protons at δ 7.72, 7.69, and 7.50, which all correlate to
aromatic carbons in the HSQC. Additionally the two protons at δ
7.72 and 7.50 are correlated to each other in the COSY spectrum,
which would represent C-14 and C-15 (Figure 2). Further
analysis of the COSY spectrum showed that the proton signal
at δ 7.69 was not correlated to any other protons, indicating that
it was attached to the C-16 position. Examination of the HMBC
spectrum showed that the proton signal at δ 7.72 was correlated
to C-25, C-27, and C-13, while the proton signal at δ 7.69 was
correlated to C-28, C-26, and C-12. This would suggest that the
COSY uncoupled proton signal from δ 7.69 is attached to C-16
and the COSY coupled proton signal from δ 7.72 is attached to
C-15. The structure of 17b was elucidated in a similar manner to
17a. The success of our methodology suggests that this approach
may be applicable to other furan ring containing natural products
in efforts to rapidly explore their SAR.
Biological Results. The synthesized compounds were eval-
uated for affinity and efficacy at opioid receptors.42 It was the
thought that the cycloadducts prepared would give greater
insight into the position of the oxygen atom and its ability to
participate in hydrogen bonding. Additionally, these compounds
would provide some measure of the amount of steric bulk that
was tolerated around the furan ring. Cycloadduct 6 had de-
creased but still appreciable affinity for KOP receptors in
comparison to 1 (Ki = 60 nM vs Ki = 7.4 nM) (Table 2). The
diethyl analogue 6 (5) saw a decrease in affinity at KOP receptors
when compared to its methyl counterpart, 6 (Ki = 120 nM vs Ki =
60 nM). While affinity for KOP receptors was diminished in 5,
affinity for delta opioid (DOP) receptors increased over 4-fold in
comparison to 1 (Ki = 2,260 nM vs Ki > 10 000 nM). As
mentioned above, we were unable to assign the relative config-
uration of the oxygen of the oxanorbornadiene in 5 and 6. It is
therefore possible that one regioisomer of the oxanorbornadiene
is preferred for interacting with opioid receptors. However,
additional synthesis and biological testing will be required to
test this hypothesis.
Cycloadduct 7 also explored the role of steric bulk in activity at
opioid receptors along with increasing the lipophilic/hydropho-
bic characteristics of 1 through the introduction of a benzene ring
instead of esters, allowing us to investigate how these character-
istics factor into activity at opioid receptors. Cycloadduct 7 was
found to have decreased affinity at KOP receptors (Ki = 790 nM)
in comparison to 1. This decreased affinity may be the result of 7
not having the ability to form hydrogen bonds like its ester-
containing cycloadduct counterparts and/or the regiochemistry
of the oxanorbornadiene is not optimal. The installation of
substituents that may partake in hydrogen bonding on the
benzene ring of 7 may improve affinity at opioid receptors and
warrants further investigation.
Figure 2. Partial numbering of compounds 16a, 16b, and 17a with
selected HMBC correlations.
Table 2. Binding Affinities of Cycloadduct Analoguesa
Ki ( SD (nM)
compound
MOP
DOP
KOP
MOP/KOP
N.D.c
DOP/KOP
>1351
1
EC50 = 2860 ( 980b
Emax = 75 ( 8%
>3200
>10 000
7.4 ( 0.7
5
2260 ( 280
>5000
120 ( 10
60 ( 10
790 ( 200
>13 000
>25
>80
>3
19
6
>4800
>83
>13
>17
>17
>22
>17
>0.6
>2
7
>2500
>10 000
>5200
11
>2700
>0.2
>6
12
>1700
>5000
290 ( 20
228 ( 12
286 ( 19
>8000
13
>3000
>5000
>13
>8
14
>3000
>5000
15
>3000
>5000
>0.4
16a/16b
1670 ( 150
>5000
1970 ( 80
0.85
a [3H]DAMGO was used for MOP, [3H]DADLE for DOP, and [3H]U69,593 for KOP receptors. b A more complete analysis provided this value, which
corrects our previously reported value.12 c Not determined.
722
dx.doi.org/10.1021/np1007872 |J. Nat. Prod. 2011, 74, 718–726

Journal of Natural Products
ARTICLE
Table 3. [35S]GTP-γ-S Activity Assay of Cycloadducts
In conclusion, a series of Diels-Alder cycloadduct analogues
of salvinorin A were synthesized in an effort to explore the effects
of steric bulk and the position of the oxygen ring on binding and
activity at opioid receptors. Microwave conditions were opti-
mized to further enhance the amenability of the furan ring in
salvinorin A to serve as a 4π diene in the [4þ2] Diels-Alder
reaction. This work signifies the first reported example of the
Diels-Alder reaction to modify a neoclerodane diterpene. Further-
more, several of the cycloadduct analogues were themselves
useful as synthetic intermediates, as they were able to undergo
reductive elimination to produce their phenyl ring counterparts.
In vitro evaluation found that steric bulk around the furan ring of
salvinorin A seems to be tolerated for binding at the KOP
receptor, as cycloadducts 5, 6, and 12 maintained appreciable
affinity. Although all compounds explored had reduced affinity
relative to salvinorin A, this indicates that the furan ring is not a
stringent requirement for KOP receptor affinity, as previously
believed. Additionally, cycloadducts 5 and 6 were found to be full
agonists at KOP receptors as compared to the known KOP agonist
U50,488. Furthermore, phenyl derivatives 13 and 14 maintained
modest affinity at KOP receptors, which further indicates that the
furan ring is not essential for affinity at opioid receptors. Further
investigation into the role of steric bulk at the furan ring will yield
valuable information into the nature of how 1 interacts at opioid
receptors. This information can then be used to aid in the
development of opioid therapeutics with enhanced pharmacolo-
gical properties.
a
compound
KOP ED50 ( SD, nM
KOP Emaxb ( SD, nM
1
5
6
40 ( 10
2150 ( 500
980 ( 200
120 ( 2
90 ( 10
100 ( 10
a ED50 = effective dose for 50% maximal response. b Emax is % at which
compound stimulates in comparison to (-)-U50,488 (500 nM) at KOP
receptors.
Further probing how steric bulk affects the interactions of 1 at
opioid receptors, the incorporation of large appendages to the
core structure of 1 resulted in the synthesis of 11 and 12. Given
that the core structures of 11 and 12 had fluorescent properties, it
was the hope that these analogues of 1 could be used as potential
visualizing agents for opioid receptors, along with exploring the
effect of increased steric bulk. Unfortunately, cycloadduct 11 had
no affinity at KOP receptors, as it had a Ki of >13 000 nM and no
appreciable affinity for mu opioid (MOP) or DOP receptors as
well. Interestingly, the introduction of a methyl group into the
bridgehead of 11 (12) increased affinity for KOP receptors 44-
fold (Ki = 290 nM vs Ki > 13 000 nM). The rationale behind this
observed result is not immediately apparent but may be attrib-
uted to 11 being tested as a mixture of endo and exo isomers,
while 12 was tested as a single isomer. Further studies are being
explored to evaluate this hypothesis.
The phenyl ring analogues were also evaluated to explore
steric bulk and the effects of non-heterocyclic aromatics at the
C-12 position. The napthyl derivative of 7 (15) had no affinity at
opioid receptors (Ki > 10 000 nM), indicating that the lack of any
groups that may hydrogen bond is detrimental for affinity at
opioid receptors. This possible explanation seems to be corro-
borated by phenyl derivatives of 5 (13) and 6 (14), as they
retained some affinity for KOP receptors though less than 1 (14:
Ki = 228 nM; 15: Ki = 286 nM vs Ki = 7.4 nM).
Along with the bridgehead oxygen, it was thought that the
carbonyls of the additional esters might be hydrogen bonding at
opioid receptors. Investigation of this hypothesis was possible
with 16, where one of the ethyl esters of 5 was replaced with a
CF3 group. Introduction of the CF3 group led to a 16-fold
decrease in affinity at KOP receptors in comparison to 5 (Ki =
1970 nM vs 120 nM). This suggested that an additional carbonyl
in this position assists in binding; however, a clear explanation
has yet to be determined, which is in part due to 16 being tested
as a mixture of regioisomers. Evaluation of each individual
regioisomer will be necessary to further investigate the impact
of the carbonyl moiety as well as the position of the CF3. This will
aid in the investigation of our initial hypothesis pertaining to the
hydrogen-bonding capabilities of the esters and their role in
binding at opioid receptors.
With cycloadducts 5 and 6 having relatively high affinity for
KOP receptors (Ki e 150 nM), they were evaluated for efficacy
at opioid receptors in the [35S]GTP-γ-S assay. Cycloadduct 6
was found to be a full agonist at KOP receptors compared to (-)-
U50,488 (Emax = 100 ( 10) (Table 3). Cycloadduct 5 was also
found to have high efficacy in this assay (Emax = 90); however both
compounds wereless potentthan 1 (6 ED50 = 980 nMvs 5 ED50 =
2150 vs ED50 = 40 nM). These compounds show that extension of
the cycloadduct esters has more of an effect on potency than
overall efficacy, and both compounds further illustrate that the
furan ring 1 of is not essential for binding or efficacy at opioid
receptors and some steric bulk is tolerated at this position.
’ EXPERIMENTAL SECTION
General Experimental Procedures. Unless otherwise indi-
cated, all reagents were purchased from commercial suppliers and were
used without further purification. Melting points were determined on a
Thomas-Hoover capillary melting apparatus. NMR spectra were re-
corded on either a Bruker Advance-300 spectrometer, a Bruker DRX-
400 with qnp probe, or a Bruker AV-500 with cryoprobe using δ values
in ppm (TMS as internal standard) and J(Hz) assignments of 1Hresonance
coupling. High-resolution mass spectrometry data were collected on
either a LCT Premier (Waters Corp.) time-of-flight mass spectrometer
or an Agilent 6890 N gas chromatograph in conjunction with a Quattro
Micro GC mass spectrometer (Micromass Ltd.). Thin-layer chroma-
tography (TLC) was performed on 0.25 mm Analtech GHLF silica gel
plates using EtOAc/n-hexanes, in 1:1 ratio, as the solvent system unless
otherwise noted. Spots on TLC were visualized by UV (254 or 365 nm),
phosphomolybdic acid in EtOH, or vanillin/H2SO4 in EtOH. Column
chromatography was performed with silica gel (32-63 μm particle size)
from MP Biomedicals. Analytical HPLC was carried out on an Agilent
1100 Series Capillary HPLC system with diode array detection at 254.8 nm
on an Agilent Eclipse XDB-C18 column (4.6 ꢀ 150 mm, 5 mm) with
isocratic elution in 60% CH3CN/40% H2O at a flow rate of 5.0 mL/min
unless otherwise noted. The systematic name for salvinorin A (1) is
(2S,4aR,6aR,7R,9S,10aS,10bR)-methyl 9-acetoxy-2-(furan-3-yl)-6a,10b-
dimethyl-4,10-dioxododecahydro-1H-benzo[f]isochromene-7-carboxylate.
Salvinorin A was isolated from S. divinorum as previously described.43
Diethyl 5-((2S,4aR,6aR,7R,9S,10aS,10bR)-9-acetoxy-7-(meth-
oxycarbonyl)-6a,10b-dimethyl-4,10-dioxododecahydro-
1H-benzo[f]isochromen-2-yl)-7-oxabicyclo[2.2.1]hepta-2,
5-diene-2,3-dicarboxylate (5). A solution of 1 (200 mg, 0.462
mmol), diethyl acetylene dicarboxylate (85 mg, 0.500 mmol), and
toluene (20 mL) was allowed to stir at room temperature and
gradually heated to reflux over 45 min. The solution was heated at
reflux for two days. The solvent was removed under reduced pressure,
and the resulting residue was purified by column chromatography
723
dx.doi.org/10.1021/np1007872 |J. Nat. Prod. 2011, 74, 718–726

Journal of Natural Products
ARTICLE
(eluent: EtOAc/n-hexanes, 2:3) to afford 262 mg of 5 (70%) as a
white powder: mp 84 - 86 ꢀC; 1H NMR (500 MHz, CDCl3) δ 6.92
(1H, ddd, J = 1.9, 3.6 Hz), 5.66 (1H, t, J = 1.8 Hz), 5.60 (1H, d, J = 1.6
Hz), 5.39 (1H, dd, J = 3.8, 11.7 Hz), 5.29 (1H, d, J = 7.7 Hz), 5.22-
5.08 (2H, m), 4.33-4.23 (4H, m), 3.73 (3H, s), 2.78-2.72 (1H, m),
2.43 (1H, d, J = 5.6 Hz), 2.41 (1H, d, J = 5.5 Hz), 2.31 (2H, dd, J = 3.2,
11.0 Hz), 2.17 (4H, d, J = 3.1 Hz), 1.80-1.75 (1H, m), 1.50 (2H, dd,
J = 11.3, 23.9 Hz), 1.40 (3H, d, J = 1.6 Hz), 1.32 (6H, ddd, J = 2.8, 6.3,
10.1 Hz), 1.10 (3H, s); 13C NMR (126 MHz, CDCl3) δ 202.0, 171.6,
170.7, 169.9, 163.1, 157.8, 152.8, 151.9, 138.0, 136.4, 85.6, 85.3, 74.9, 73.9,
64.1, 61.8, 61.6, 53.5, 52.0, 51.2, 42.0, 41.2, 40.3, 38.0, 35.4, 30.8, 20.6, 18.1,
16.3, 15.3, 14.1; HRESIMS m/z 625.2203 [M þ Na] (calcd for C31H38-
O12Na, 625.2261); HPLC tR = 6.802 min; purity = 96%.
Dimethyl 5-((2S,4aR,6aR,7R,9S,10aS,10bR)-9-acetoxy-7-(meth-
oxycarbonyl)-6a,10b-dimethyl-4,10-dioxododecahydro-
1H-benzo[f]isochromen-2-yl)-7-oxabicyclo[2.2.1]hepta-2,
5-diene-2,3-dicarboxylate (6). Method A. Compound 6was prepared
from 1 and dimethyl acetylene dicarboxylate using a similar procedure to that
for 5 to afford 152 mg (70%) as a white solid.
Method B. A solution of 1 (100 mg, 0.23 mmol), dimethyl acetylene di-
carboxylate (35 mg, 0.25 mmol), and toluene (25 mL) was placed in a
sealed 25 mL quartz tube and irradiated in a microwave reactor (Biotage
Initiator) at 100 ꢀC for 30 min with normal absorbance levels. Solvent
was removed under reduced pressure, and the resulting residue was
purified using column chromatography (EtOAc/n-hexanes, 2:3) to afford
69 mg (63%) as a white solid: mp 107-110 ꢀC; 1H NMR (500 MHz,
CDCl3) δ 6.91 (1H, d, J = 12.2 Hz), 5.71 (1H, d, J = 1.6 Hz), 5.67 (1H,
s), 5.63 (1H, d, J = 1.6 Hz), 5.40-5.35 (1H, m), 5.32-5.26 (1H, m),
3.85 (3H, s), 3.82 (3H, d, J = 1.4 Hz), 3.73 (3H, s), 2.74 (1H, s), 2.42
(1H, s), 2.31 (2H, d, J = 7.4 Hz), 2.17 (4H, d, J = 3.7 Hz), 2.05 (2H, s),
1.80-1.75 (1H, m), 1.47 (2H, s), 1.40 (3H, s), 1.10 (3H, s); 13C NMR
(126 MHz, CDCl3) δ 200.8, 171.5, 170.6, 170.0, 162.8, 157.8, 137.9,
136.5, 84.3, 84.2, 84.1, 73.6, 72.5, 71.7, 62.7, 62.7, 52.2, 51.4, 51.2, 50.8,
49.7, 40.7, 39.8, 36.7, 34.1, 19.4, 16.8, 15.1, 14.0; HRESIMS m/z
597.1921 [M þ Na] (calcd for C29H34O12Na, 597.1948); HPLC tR =
4.757 min; purity = 97%.
KOt-Bu (815 mg, 7.3 mmol) and 2-methylfuran (200 mg, 2.4 mmol).
TLC indicated that the reaction was complete after approximately 3 h.
The solvent was removed under reduced pressure, and CH3OH was
added to the residue, resulting in an off-white precipitate. The precipitate
was filtered to afford 416 mg of 8b (60%) as an off-white powder: mp
130-132 ꢀC; 1H NMR (400 MHz, CDCl3) δ 8.18 (2H, ddd, J = 1.1, 4.6,
7.9 Hz), 7.84-7.76 (2H, m), 7.73-7.63 (3H, m), 7.60-7.53 (2H, m),
7.43 (1H, s), 6.02 (1H, d, J = 1.8 Hz), 2.22 (3H, s); 13C NMR (126 MHz,
CDCl3) δ 194.0, 149.8, 148.4, 144.1, 142.0, 138.1, 137.0, 130.3, 130.2,
130.2, 129.4, 128.3, 128.0, 127.5, 127.1, 121.8, 121.6, 92.5, 82.5, 27.7;
HRESIMS m/z 287.1055 [M þ H] (calcd for C20H15O2, 287.1072);
HPLC tR = 32.15 min; purity = 98%.
(2S,4aR,6aR,7R,9S,10aS,10bR)-Methyl 9-acetoxy-2-(8,8a,9,
12,12a,13-hexahydro-8,13:9,12-diepoxy-5H-dibenzo[3,4:6,7]-
cyclohept[1,2]naphthanen-5-on-10-yl)-6a,10b-dimethyl-4,
10-dioxododecahydro-1H-benzo[f]isochromene-7-carbox-
ylate (11). Compound 11 was synthesized as described for 5 from 1
using 8a38 to afford 43.2 mg of the exo and endo isomers of 11 (32%) as a
white powder: mp 211-214 ꢀC; 1H NMR (500 NMR, CDCl3) δ 8.12
(5H, ddd, J = 1.2, 6.8, 12.3 Hz), 7.74-7.64 (6H, m), 7.55 (4H, d, J =
7.5 Hz), 7.51-7.44 (3H, m), 7.40 (3H, d, J = 7.9 Hz), 6.18 (1H, d, J =
2.1 Hz), 6.11 (2H, t, J = 2.0 Hz), 5.32 (2H, s), 5.23 - 5.21 (3H, m), 5.14
(2H, s), 3.73 (5H, s), 3.69 (3H, s), 3.06-2.97 (5H, m), 2.82-2.57 (7H,
m), 2.31 (5H, dd, J = 7.5, 14.8 Hz), 2.20 (4H, d, J = 2.5 Hz), 2.18 (4H, s),
2.14-2.08 (3H, m), 2.05-2.00 (2H, m), 1.45 (3H, s), 1.43 (4H, s), 1.11
(2H, s), 1.10 (2H, s); 13C NMR (126 MHz, CDCl3) δ 199.9, 199.6,
191.7, 191.6, 169.20, 169.19, 168.9, 168.7, 167.7, 145.9, 145.6, 142.72,
142.69, 142.2, 136.2 (2C), 136.04, 136.01, 129.7 (2C), 129.6, 129.44,
129.38, 128.24, 128.22, 128.1, 128.0, 127.7, 127.6(2C), 126.72, 126.70,
126.65, 126.6, 125.6, 125.2, 122.7, 122.6, 122.3, 122.2, 78.5, 78.3, 78.1,
77.7, 77.2, 76.9, 76.5, 74.8, 74.6, 74.4, 72.7, 72.64, 72.60, 70.8, 61.9, 61.6,
51.1, 51.0, 49.61, 49.56, 48.70, 48.69, 47.9, 47.8, 47.4, 47.2, 39.7, 39.6,
39.0, 38.3, 35.7, 35.5, 33.0, 32.7, 28.4, 18.3, 18.2, 15.8, 15.7, 14.0, 13.9,
13.3, 12.9; HRESIMS m/z 727.2549 [M þ Na] (calcd for C42H40O10-
Na, 727.2519); HPLC tR = 5.657 min; purity = >99%.
(2S,4aR,6aR,7R,9S,10aS,10bR)-Methyl 9-acetoxy-2-(8,8a,
9,12,12a,13-hexahydro-8-methyl-8,13:9,12-diepoxy-5H-di-
benzo[3,4:6,7]cyclohept[1,2]naphthanen-5-on-10-yl)-6a,10b-
dimethyl-4,10-dioxododecahydro-1H-benzo[f]isochromene-
7-carboxylate (12). Compound 12 was synthesized as described for
5from 1using 9b to afford 14.2 mg of 12 (26%) as a white powder, mp 198-
200 ꢀC (dec); 1H NMR (500 MHz, CDCl3) δ 7.97 (2H, d, J = 7.8 Hz),
7.65 (1H, dt, J = 1.4, 7.5 Hz), 7.59 (1H, m), 7.52 (3H, d, J = 8.7 Hz), 7.43
(1H, d, J = 7.32 Hz), 6.17 (2H, t, J = 1.9 Hz), 5.21 (3H, m), 5.05 (2H, s),
3.73 (3H, s), 3.17 (3H, m), 2.76 (2H, dd, J = 10.8, 6.1 Hz), 2.63 (1H, dd,
13.3, 5.8 Hz), 2.31 (2H, m), 2.18 (3H, s), 1.81 (1H, s), 1.70 (3H, s), 1.59
(1H, m), 1.42 (3H, s), 1.26 (2H, s), 1.11 (3H, s); 13C NMR (126 MHz,
CDCl3) δ 201.4, 194.9, 171.6, 170.7, 169.5, 148.0, 147.2, 146.2, 138.8,
131.7, 130.8, 129.8, 129.6, 129.3, 129.1, 129.0, 128.5, 128.4, 124.7, 123.9,
88.8, 79.5, 79.3, 78.5, 77.6, 75.01, 74.98, 64.4, 53.5, 53.0, 52.5, 52.0, 51.0,
42.1, 41.5, 38.1, 35.4, 30.8, 20.6, 18.2, 17.5, 16.3, 15.7; HRESIMS m/z
741.2693 [M þ Na] (calcd for C43H42O10Na, 741.2676); HPLC tR =
10.938 min; purity = 99%.
(2S,4aR,6aR,7R,9S,10aS,10bR)-Methyl 9-acetoxy-2-(7-benzo-
oxabicyclo[2.2.1]hepta-2,5-dien-2-yl)-6a,10b-dimethyl-4,
10-dioxododecahydro-1H-benzo[f]isochromene-7-carbox-
ylate (7). To a solution of 1 (150 mg, 0.35 mmol), 2-(trimethylsilyl)-
phenyl trifluoromethanesulfonate (310.6 mg, 1.04 mmol), and CH3CN
(20 mL) was added CsF (319 mg, 2.1 mmol), and the solution was
allowed to stir at room temperature overnight. Upon completion, the
reaction was diluted with H2O (25 mL) and ether (25 mL). The aqueous
layer was extracted with ether and dried (Na2SO4). The compound was
purified by column chromatography (eluent: EtOAc/n-hexanes, 1:1) to
afford 42.4 mg of 7 (24%) as a white powder: mp 242 - 245 ꢀC (dec);
1H NMR (500 MHz, CDCl3) δ 7.33-7.31 (1H, m), 7.25-7.22 (1H,
m), 7.01 (2H, dd, J = 3.0, 5.1 Hz), 6.69 (1H, t, J = 2.0 Hz), 5.72 (2H, d,
J = 11.9 Hz), 5.17-5.09 (2H, m), 3.72 (3H, s), 2.73-2.68 (1H, m),
2.37-2.33 (1H, m), 2.31-2.26 (2H, m), 2.17 (3H, s), 2.11 (1H, s), 1.95
(1H, d, J = 11.9 Hz), 1.76 (1H, d, J = 13.4 Hz), 1.66-1.58 (1H, m),
1.53-1.45 (2H, m), 1.37 (3H, s), 1.31 (1H, s), 1.07 (3H, s); 13C NMR
(126 MHz, CDCl3) δ 199.7, 169.1, 168.5, 167.6, 153.9, 146.2, 145.7,
134.1, 123.1, 122.9, 118.3, 117.8, 80.5, 79.8, 72.7, 71.4, 61.8, 51.1, 49.6,
48.5, 39.8, 37.6, 35.6, 32.8, 28.3, 18.2, 15.6, 13.9, 13.2; HRESIMS m/z
531.1961 [M þ Na] (calcd for C29H32O8Na, 531.1995); HPLC tR =
8.260 min; purity = 98%.
Diethyl 4-((2S,4aR,6aR,7R,9S,10aS,10bR)-9-acetoxy-7-(meth-
oxycarbonyl)-6a,10b-dimethyl-4,10-dioxododecahydro-
1H-benzo[f]isochromen-2-yl)phthalate (13). A solution of 5
(100 mg, 0.17 mmol), Fe2(CO)9 (150 mg, 0.41 mmol), and toluene (15 mL)
was allowed to stir at 60 ꢀC for 20 min. Once the solution turned black, it
was gradually heated to reflux and allowed to stir for 2 h. The solution
was filtered through a pad of Celite, and the solvent was removed under
reduced pressure. The residue was purified by column chromatography
(eluent: 50% EtOAc/50% n-hexanes) to afford 55 mg of 13 (70%)
3,6-Epoxy-3-methyl-3,6-dihydrotribenzocycloheptatrienone
(10b). A solution of dibenzosuberenone (500 mg, 2.4 mmol) and Br2
(774 mg, 4.8 mmol) in CH2Cl2 (100 mL) was allowed to stir at -30 ꢀC
for 2 h, during which time a precipitate was formed. The precipitate was
collected by filtration and placed in a sealed tube containing THF (100 mL).
The resulting suspension was cooled to 0 ꢀC and then treated with
1
as a white powder: mp 110-112 ꢀC; H NMR (500 NMR MHz,
CDCl3) δ 7.73 (1H, d, J = 8.0 Hz), 7.62 (1H, d, J = 1.8 Hz), 7.43 (1H, dd,
724
dx.doi.org/10.1021/np1007872 |J. Nat. Prod. 2011, 74, 718–726

Journal of Natural Products
ARTICLE
J = 1.7, 8.1 Hz), 5.60 (1H, dd, J = 5.1, 11.9 Hz), 5.12-5.06 (1H, m),
4.40-4.33 (4H, m), 3.73 (3H, s), 2.73 (1H, dd, J = 6.3, 10.5 Hz), 2.55
(1H, dd, J = 5.1, 13.6 Hz), 2.33-2.26 (2H, m), 2.19 (1H, s), 2.16 (3H,
s), 2.14 (1H, s), 2.12 (1H, d, J = 3.1 Hz), 1.81 (1H, d, J = 13.2 Hz), 1.72-
1.62 (1H, m), 1.58 (1H, s), 1.50 (3H, d, J = 8.0 Hz), 1.47-1.42 (1H, m),
1.37 (6H, q, J = 7.1 Hz), 1.13 (3H, s); 13C NMR (126 MHz, CDCl3)
δ 201.7, 171.2, 170.6, 169.5, 167.1, 166.7, 143.3, 132.8, 131.4, 129.2,
127.3, 125.3, 77.5, 74.6, 63.5, 61.5, 61.4, 53.2, 51.7, 51.3, 44.8, 41.8, 37.8,
35.5, 30.4, 22.4, 20.2, 17.8, 16.1, 13.80, 13.78; HRESIMS m/z 609.2332
[M þ Na] (calcd for C31H38O11Na, 609.2314); HPLC tR = 11.978 min;
purity = >99%.
Dimethyl 4-((2S,4aR,6aR,7R,9S,10aS,10bR)-9-acetoxy-7-(meth-
oxycarbonyl)-6a,10b-dimethyl-4,10-dioxododecahydro-
1H-benzo[f]isochromen-2-yl)phthalate (14). Compound 14
was synthesized as described for 13 from 6 to afford 55.2 mg of 14 (77%)
as a white powder: mp 116-119 ꢀC; 1H NMR (500 MHz, CDCl3) δ
7.73 (1H, d, J = 8.0 Hz), 7.64 (1H, s), 7.45 (1H, d, J = 7.9 Hz), 5.60 (1H,
d, J = 6.9 Hz), 5.13-5.06 (1H, m), 3.91 (3H, s), 3.91 (3H, s), 3.73 (3H,
s), 2.72 (1H, d, J = 6.3 Hz), 2.55 (1H, d, J = 8.5 Hz), 2.31 (2H, d, J = 9.8
Hz), 2.22 (1H, s), 2.19 (1H, s), 2.16 (3H, s), 2.12 (1H, s), 1.81 (1H, d,
J = 13.2 Hz), 1.66 (1H, s), 1.59 (2H, s), 1.51 (3H, s), 1.13 (3H, s); 13C
NMR δ 200.1, 169.6, 169.0, 168.0, 165.8, 165.5, 141.9, 130.9, 129.5,
127.6, 126.0, 123.8, 75.9, 73.1, 62.0, 51.7, 50.9, 50.8, 50.1, 49.7, 43.2,
40.2, 36.2, 33.9, 28.8, 18.7, 16.2, 14.5, 13.3; HRESIMS m/z 581.1994 [M
þ Na] (calcd for C29H34O11Na, 581.1999); HPLC tR = 4.103 min;
purity = 96%.
(2S,4aR,6aR,7R,9S,10aS,10bR)-Methyl 9-acetoxy-6a,10b-
dimethyl-2-(naphthalen-2-yl)-4,10-dioxododecahydro-1H-
benzo[f]isochromene-7-carboxylate (15). Compound 15 was
synthesized as described for 15 from 7 to afford 41 mg of 15 (70%) as a
white powder: mp 240-242 ꢀC; 1H NMR (500 MHz, CDCl3) δ 7.82
(3H, dd, J = 7.2, 12.6 Hz), 7.75 (1H, s), 7.51-7.46 (2H, m), 7.38 (1H,
d, J = 1.8 Hz), 5.72 (1H, dd, J = 5.2, 11.8 Hz), 5.11-5.06 (1H, m), 3.73
(3H, s), 2.76-2.70 (1H, m), 2.62 (1H, dd, J = 5.3, 13.6 Hz), 2.30 (2H,
dd, J = 5.8, 13.1 Hz), 2.22 (3H, s), 2.16 (2H, d, J = 4.2 Hz), 2.15 (3H, s),
1.81 (1H, s), 1.63 (1H, s), 1.55 (3H, s), 1.14 (3H, s); 13C NMR (126
MHz, CDCl3) δ 200.2, 169.7, 169.6, 168.1, 135.6, 131.3, 131.2, 126.8,
126.2, 125.9, 124.6, 124.5, 122.6, 121.3, 74.9, 73.2, 62.2, 51.8, 50.2, 49.8,
43.6, 40.3, 36.4, 34.0, 28.9, 27.9, 18.7, 16.4, 14.6; HRESIMS m/z
515.1949 [M þ Na] (calcd for C29H32O7Na, 515.1945); HPLC tR =
9.175 min; purity = >99%.
20.58, 18.12, 18.08, 16.4, 15.3, 15.1, 13.9; HRESIMS m/z 621.1986 [M þ
Na] (calcd for C29H33F3O10Na, 621.1924); analytical HPLC 16a tR
12.858 min; ratio = 91%; 16b tR = 11.759 min; ratio = 9%.
=
(2S,4aR,6aR,7R,9S,10aS,10bR)-Methyl 9-acetoxy-2-(3-(eth-
oxycarbonyl)-4-(trifluoromethyl)phenyl)-6a,10b-dimethyl-4,
10-dioxododecahydro-1H-benzo[f]isochromene-7-carbox-
ylate (17a) and (2S,4aR,6aR,7R,9S,10aS,10bR)-Methyl
9-acetoxy-2-(4-(ethoxycarbonyl)-3-(trifluoromethyl)phenyl)-
6a,10b-dimethyl-4,10-dioxododecahydro-1H-benzo[f]iso-
chromene-7-carboxylate (17b). Asolution of16a and 16b (200 mg,
0.33 mmol), Fe2(CO)9 (150 mg, 0.41 mmol), and toluene (15 mL) was
allowed to stir at 60 ꢀC for 20 min. Once the solution turned black, it was
gradually heated to reflux and allowed to stir for 2 h. The solution was
filtered through a pad of Celite, and solvent was removed under reduced
pressure. The residue was purified by column chromatography (gradient
eluent: 10% EtOAc/90% n-hexanes to 30% EtOAc/70% n-hexanes) to
afford 121 mg of 17a (62%) and 8.2 mg of 17b (4.2%).
17a: white powder; mp 110-112 ꢀC; 1H NMR (500 MHz, CDCl3) δ
7.72 (1H, d, J = 8.2 Hz), 7.69 (1H, s), 7.50 (1H, d, J = 8.0 Hz), 5.62 (1H,
dd, J = 5.0, 12.0 Hz), 5.13 - 5.06 (1H, m), 4.40 (2H, q, J = 6.9 Hz), 3.73
(3H, s), 2.75-2.71 (1H, m), 2.56 (1H, dd, J = 5.1, 13.6 Hz), 2.30 (2H,
dd, J = 7.8, 13.5 Hz), 2.24-2.18 (1H, m), 2.16 (3H, s), 2.13 (2H, d, J =
8.3 Hz), 1.82 (1H, d, J = 13.2 Hz), 1.66 (2H, dd, J = 14.2, 26.0 Hz), 1.52
(3H, s), 1.49-1.43 (1H, m), 1.39 (3H, t, J = 7.2 Hz), 1.13 (3H, s); 13C
NMR (126 MHz, CDCl3) δ 202.0, 171.5, 170.8, 169.9, 166.5, 144.4,
132.2 (q, 3JCF = 1.80), 128.4 (q, 2JCF = 32.99), 127.8, 127.2 (q, 3JCF
=
5.20), 127.0, 123.2 (q, 1JCF = 273.41), 77.6, 75.0, 63.8, 62.3, 53.6, 52.0,
51.6, 45.1, 42.1, 38.1, 35.8, 30.7, 20.6, 18.1, 16.4, 15.2, 13.9; HRESIMS
m/z 605.1959 [M þ Na] (calcd for C29H33F3O9, 605.1974); HPLC tR =
16.767 min; purity = >99%.
17b: white powder; mp 100-102 ꢀC; 1H NMR (500 MHz, CDCl3)
δ 7.78 (1H, d, J = 8.0 Hz), 7.66 (1H, s), 7.51 (1H, d, J = 8.0 Hz), 5.62
(1H, dd, J = 5.0, 12.0 Hz), 5.12-5.07 (1H, m), 4.39 (2H, q, J = 7.1 Hz),
3.73 (3H, s), 2.78-2.71 (1H, m), 2.55 (1H, dd, J = 5.1, 13.6 Hz), 2.30
(2H, dd, J = 7.7, 13.5 Hz), 2.21 (1H, d, J = 13.9 Hz), 2.17 (3H, d, J =
5.8 Hz), 1.82 (1H, d, J = 13.2 Hz), 1.69 (1H, d, J = 14.4 Hz), 1.60 (3H, s),
1.52 (3H, s), 1.46 (1H, t, J = 12.8 Hz), 1.38 (3H, t, J = 7.1 Hz), 1.14 (3H,
s); 13C NMR (126 MHz, CDCl3) δ 202.0, 171.5, 170.8, 169.9, 166.4,
143.7, 131.3 (q, 3JCF = 1.83), 130.8, 129.4 (q, 2JCF = 32.66), 128.6, 123.7
3
(q, JCF = 5.38), 123.1 (q, 1JCF = 273.83), 77.7, 75.0, 63.8, 62.2, 53.6,
52.1, 51.7, 45.1, 42.1, 38.1, 35.9, 30.7, 20.6, 18.1, 16.5, 15.2, 13.9;
HRESIMS m/z 605.1979 [M þ Na] (calcd for C29H33F3O9, 605.1974);
HPLC tR = 16.705 min; purity = 90%.
(2S,4aR,6aR,7R,9S,10aS,10bR)-Methyl 9-acetoxy-2-(5-(eth-
oxycarbonyl)-6-(trifluoromethyl)-7-oxabicyclo[2.2.1]hepta-2,
5-dien-2-yl)-6a,10b-dimethyl-4,10-dioxododecahydro-1H-
benzo[f]isochromene-7-carboxylate (16a) and (2S,4aR,6aR,
7R,9S,10aS,10bR)-Methyl 9-acetoxy-2-(6-(ethoxycarbonyl)-
5-(trifluoromethyl)-7-oxabicyclo[2.2.1]hepta-2,5-dien-2-yl)-
6a,10b-dimethyl-4,10-dioxododecahydro-1H-benzo[f]iso-
chromene-7-carboxylate (16b). Compounds 16a and 16b were syn-
thesized as described for 5 from 1 using ethyl 4,4,4-trifluoro-2-butynoate to
afford 402 mg of 16a and 16b (48%) as a white powder: mp 119-121 ꢀC;
1H NMR (500 MHz, CDCl3) δ 6.91 (1H, dd, J = 1.9, 3.9 Hz), 5.76 (1H, s),
5.64 (1H, dt, J = 1.8, 5.5 Hz), 5.61 (1H, s), 5.39 (1H, dd, J = 4.7, 11.0 Hz),
5.33-5.27 (1H, m), 5.15 (1H, dd, J = 11.0, 19.9 Hz), 4.35-4.23 (3H, m),
3.73 (4H, s), 2.79-2.72 (1H, m), 2.43 (1H, td, J = 5.4, 13.3 Hz), 2.31 (2H,
dd, J = 4.3, 10.6 Hz), 2.17 (5H, d, J = 2.5 Hz), 2.13 (1H, s), 2.09 (1H, d, J =
8.7 Hz), 2.03 (1H, d, J = 12.3 Hz), 1.78 (1H, d, J = 10.1 Hz), 1.61 (5H, dd,
J = 12.4, 21.2 Hz), 1.54-1.46 (2H, m), 1.41 (4H, s), 1.37-1.29 (4H, m),
1.11 (4H, s); 13C NMR (126 MHz, CDCl3) δ 202.0, 171.54, 171.53, 170.7,
170.4, 170.0, 169.9, 162.0, 161.5, 158.5, 157.9, 151.3 (q, 3JCF = 4.66), 151.0
’ ASSOCIATED CONTENT
Supporting Information. NMR (1H, 13C) and HPLC
S
b
analysis of compounds 5-7 and 11-17. This material is available
free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*Tel: (785) 864-3267. Fax: (785) 864-5326. E-mail: prisinza@
ku.edu.
’ ACKNOWLEDGMENT
The authors would like to thank the National Institute on
Drug Abuse (DA018151) for financial support of ongoing
research. Portions of this work were supported by the Intramural
Research Program, National Institute on Drug Abuse, NIH,
DHHS. The content is the sole responsibility of the authors
and does not necessarily represent the official views of the
2
3
1
(q, JCF = 35.57), 150.8 (q, JCF = 4.87), 137.4, 135.8, 121.5 (q, JCF
=
=
269.57), 121.4 (q, 1JCF = 269.63), 85.7 (d, 3JCF = 2.67), 84.5 (d, 3JCF
2.52), 75.0, 74.9, 73.9, 72.7, 64.0, 64.0, 62.2, 62.1, 53.50, 53.47, 52.0, 51.2,
51.0, 42.04, 41.99, 41.1, 40.4, 38.02, 37.96, 35.5, 35.3, 30.8, 30.7, 20.61,
725
dx.doi.org/10.1021/np1007872 |J. Nat. Prod. 2011, 74, 718–726

Journal of Natural Products
ARTICLE
National Institute on Drug Abuse or the National Institutes
of Health.
(28) Kappe, C. O.; Murphree, S. S.; Padwa, A. Tetrahedron 1997, 53,
14179–14233.
(29) Zhao, Q.; Zou, C.; Hao, X. J.; Chen, Y. Z. Chin. Chem. Lett.
1999, 10, 531–532.
(30) Reddy, P. P.; Lavekar, A. G.; Babu, K. S.; Rao, R. R.; Shashidhar,
J.; Shashikiran, G.; Rao, J. M. Bioorg. Med. Chem. Lett. 2010, 20, 2525–
2528.
’ DEDICATION
Dedicated to Professor James M. Cook on the occasion of his
65th birthday.
(31) Harding, W. W.; Schmidt, M.; Tidgewell, K.; Kannan, P.;
Holden, K. G.; Dersch, C. M.; Rothman, R. B.; Prisinzano, T. E. Bioorg.
Med. Chem. Lett. 2006, 16, 3170–3174.
(32) Shoji, M.; Inoue, T.; Nakao, S.; Nakamura, M.; Hayashi, Y.
Angew. Chem., Int. Ed. 2002, 41, 4079–4082.
(33) Hong, V.; Kislukhin, A. A.; Finn, M. G. J. Am. Chem. Soc. 2009,
131, 9986–9994.
(34) Padwa, A. Tetrahedron 1997, 53, 14179–14233.
(35) Lautens, M.; Webster, R. Org. Lett. 2009, 11, 4688–4691.
(36) Ma, C.; Ding, H.; Zhang, Y.; Bian, M.; Yao, W. J. Org. Chem.
2008, 73, 578–584.
(37) Schafer, D.; Franke, C.; Tochtermann, W. Chem. Ber. 1968,
101, 3122–3137.
’ REFERENCES
(1) Valdes, L. J., 3rd; Diaz, J. L.; Paul, A. G. J. Ethnopharmacol. 1983,
7, 287–312.
(2) Valdes, L. J., III; Butler, W. M.; Hatfield, G. M.; Paul, A. G.;
Koreeda, M. J. Org. Chem. 1984, 49, 4716–4720.
(3) Ortega, A.; Blount, J. F.; Manchand, P. S. J. Chem. Soc., Perkin
Trans. 1 1982, 2505–2508.
(4) Siebert, D. J. J. Ethnopharmacol. 1994, 43, 53–56.
(5) Roth, B. L.; Baner, K.; Westkaemper, R.; Siebert, D.; Rice, K. C.;
Steinberg, S.; Ernsberger, P.; Rothman, R. B. Proc. Natl. Acad. Sci. U. S. A.
2002, 99, 11934–11939.
(6) Chartoff, E. H.; Potter, D.; Damez-Werno, D.; Cohen, B. M.;
Carlezon, W. A., Jr. Neuropsychopharmacology 2008, 33, 2676–2687.
(7) Morani, A. S.; Kivell, B.; Prisinzano, T. E.; Schenk, S. Pharmacol.,
Biochem. Behav. 2009, 94, 244–249.
(38) Tochtermann, W.; Oppenlaender, K.; Walter, U. Chem. Ber.
1964, 97, 1318–1328.
(39) Shao, H.; Chen, X.; Wang, Z.; Lu, P. J. Phys. Chem. B 2007, 111,
10386–10396.
(8) McCurdy, C. R.; Sufka, K. J.; Smith, G. H.; Warnick, J. E.; Nieto,
M. J. Pharmacol., Biochem. Behav. 2006, 83, 109–113.
(9) Ansonoff, M. A.; Zhang, J.; Czyzyk, T.; Rothman, R. B.; Stewart,
J.; Xu, H.; Zjwiony, J.; Siebert, D. J.; Yang, F.; Roth, B. L.; Pintar, J. E.
J. Pharmacol. Exp. Ther. 2006, 318, 641–648.
(10) Prisinzano,T. E.;Rothman, R.B.Chem. Rev. 2008, 108, 1732–1743.
(11) Bꢀeguin, C.; Duncan, K. K.; Munro, T. A.; Ho, D. M.; Xu, W.;
Liu-Chen, L.-Y.; Carlezon, W. A., Jr.; Cohen, B. M. Bioorg. Med. Chem.
2009, 17, 1370–1380.
(40) Sasaki, T.; Kanamatsu, K.; Izizuka, K.; Ando, I. J. Org. Chem.
1976, 41, 1425–1429.
(41) Best, W.; Collins, P.; McCulloch, R.; Wege, D. Aust. J. Chem.
1982, 35, 843–848.
(42) Fontana, G.; Savona, G.; Rodríguez, B.; Dersch, C. M.; Roth-
man, R. B.; Prisinzano, T. E. Tetrahedron 2008, 64, 10041–10048.
(43) Tidgewell, K.; Harding, W. W.; Schmidt, M.; Holden, K. G.;
Murry, D. J.; Prisinzano, T. E. Bioorg. Med. Chem. Lett. 2004, 14, 5099–
5102.
(12) Simpson, D. S.; Lovell, K. M.; Lozama, A.; Han, N.; Day, V. W.;
Dersch, C. M.; Rothman, R. B.; Prisinzano, T. E. Org. Biomol. Chem.
2009, 7, 3748–3756.
(13) Wang, Y.; Chen, Y.; Xu, W.; Lee, D. Y.; Ma, Z.; Rawls, S. M.;
Cowan, A.; Liu-Chen, L. Y. J. Pharmacol. Exp. Ther. 2008, 324, 1073–
1083.
(14) Hooker, J. M.; Munro, T. A.; Beguin, C.; Alexoff, D.; Shea, C.;
Xu, Y.; Cohen, B. M. Neuropharmacology 2009, 57, 386–391.
(15) Beguin, C.; Potter, D. N.; Dinieri, J. A.; Munro, T. A.; Richards,
M. R.; Paine, T. A.; Berry, L.; Zhao, Z.; Roth, B. L.; Xu, W.; Liu-Chen,
L. Y.; Carlezon, W. A., Jr.; Cohen, B. M. J. Pharmacol. Exp. Ther. 2008,
324, 188–195.
(16) Kane, B. E.; McCurdy, C. R.; Ferguson, D. M. J. Med. Chem.
2008, 51, 1824–1830.
(17) Vortherms, T. A.; Mosier, P. D.; Westkaemper, R. B.; Roth,
B. L. J. Biol. Chem. 2007, 282, 3146–3156.
(18) Peterson, L. A. Drug Metab. Rev. 2006, 38, 615–626.
(19) Zhou, S.; Koh, H.-L.; Gao, Y.; Gong, Z.-y.; Lee, E. J. D. Life Sci.
2004, 74, 935–968.
(20) Kouzi, S. A.; McMurtry, R. J.; Nelson, S. D. Chem. Res. Toxicol.
1994, 7, 850–856.
(21) Druckova, A.; Marnett, L. J. Chem. Res. Toxicol. 2006, 19, 1330–
1340.
(22) Druckova, A.; Mernaugh, R. L.; Ham, A. J.; Marnett, L. J. Chem.
Res. Toxicol. 2007, 20, 1393–1408.
(23) Zhou, S. F.; Xue, C. C.; Yu, X. Q.; Wang, G. Curr. Drug. Metab.
2007, 8, 526–553.
(24) Smissman, E. E.; Nelson, W. L.; LaPidus, J. B.; Day, J. L. J. Med.
Chem. 1966, 9, 458–465.
(25) Scheerer, J. R.; Lawrence, J. F.; Wang, G. C.; Evans, D. A. J. Am.
Chem. Soc. 2007, 129, 8968–8969.
(26) Hagiwara, H.; Suka, Y.; Nojima, T.; Hoshi, T.; Suzuki, T.
Tetrahedron 2009, 65, 4820–4825.
(27) Bentley, K. W.; Boura, A. L.; Fitzgerald, A. E.; Hardy, D. G.;
McCoubrey, A.; Aikman, M. L.; Lister, R. E. Nature 1965, 206, 102–103.
726
dx.doi.org/10.1021/np1007872 |J. Nat. Prod. 2011, 74, 718–726