Salvinorin A, an Active Component of the Hallucinogenic Sage Salvia divinorum Is a Highly Efficacious κ-Opioid Receptor Agonist: Structural and Functional Considerations

Article abstract of DOI:10.1124/jpet.103.059394

The diterpene salvinorin A from Salvia divinorum has recently been reported to be a high-affinity and selective κ-opioid receptor agonist (Roth et al., 2002). Salvinorin A and selected derivatives were found to be potent and efficacious agonists in severa

Full text of DOI:10.1124/jpet.103.059394

0022-3565/04/3083-1197–1203$20.00
THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS
Copyright © 2004 by The American Society for Pharmacology and Experimental Therapeutics
JPET 308:1197–1203, 2004
Vol. 308, No. 3
59394/1130181
Printed in U.S.A.
Salvinorin A, an Active Component of the Hallucinogenic Sage
Salvia divinorum Is a Highly Efficacious ␬-Opioid Receptor
Agonist: Structural and Functional Considerations
Charles Chavkin, Sumit Sud, Wenzhen Jin, Jeremy Stewart, Jordan K. Zjawiony,
Daniel J. Siebert, Beth Ann Toth, Sandra J. Hufeisen, and Bryan L. Roth
Department of Pharmacology, University of Washington School of Medicine, Seattle, Washington (C.C., S.S., W.J.); Salvia
divinorum Research and Information Center, Los Angeles, California (D.S.); Department of Pharmacognosy, University of
Mississippi, Oxford, Mississippi (J.S., J.K.Z.); and National Institute of Mental Health Psychoactive Drug Screening Program
and Department of Biochemistry, Case Western Reserve University Medical School, Cleveland, Ohio (B.A.T., S.J.H., B.L.R.)
Received August 31, 2003; accepted November 7, 2003
ABSTRACT
The diterpene salvinorin A from Salvia divinorum has recently salvinorin A was critical for ␬-opioid receptor binding and ac-
been reported to be a high-affinity and selective ␬-opioid re- tivation. Because issues of receptor reserve complicate esti-
ceptor agonist (Roth et al., 2002). Salvinorin A and selected mates of agonist efficacy and potency, we also examined the
derivatives were found to be potent and efficacious agonists in agonist actions of salvinorin A by measuring potassium con-
several measures of agonist activity using cloned human ␬-opi- ductance through G protein-gated Kϩ channels coexpressed in
oid receptors expressed in human embryonic kidney-293 cells. Xenopus oocytes, a system in which receptor reserve is mini-
Thus, salvinorin A, salvinorinyl-2-propionate, and salvinorinyl- mal. Salvinorin A was found to be a full agonist, being signifi-
2-heptanoate were found to be either full (salvinorin A) or partial cantly more efficacious than (trans)-3,4-dichloro-N-methyl-N-
(2-propionate, 2-heptanoate) agonists for inhibition of forskolin- [2-(1-pyrrolidinyl)-cyclohexyl] benzeneacetamide methane-
stimulated cAMP production. Additional studies of agonist po- sulfonate hydrate (U50488) or (trans)-3,4-dichloro-N-methyl-N-
tency and efficacy of salvinorin A, performed by cotransfecting [2-(1-pyrrolidinyl)-cyclohexyl] benzeneacetamide methane-
either the chimeric G proteins Gaq-i5 or the universal G protein sulfonate hydrate (U69593) (two standard ␬-opioid agonists)
Ga16 and quantification of agonist-evoked intracellular calcium and similar in efficacy to dynorphin A (the naturally occurring
mobilization, affirmed that salvinorin A was a potent and effec- peptide ligand for ␬-opioid receptors). Salvinorin A thus repre-
tive ␬-opioid agonist. Results from structure-function studies sents the first known naturally occurring non-nitrogenous full
suggested that the nature of the substituent at the 2-position of agonist at ␬-opioid receptors.
Salvia divinorum, a member of the Lamiaceae family, has Salvinorin A induces an intense, short-lived hallucinogenic
been used by the Mazatec Indians of northeastern Oaxaca,
Mexico, primarily for its psychoactive effects (Wasson, 1962,
1963) for many hundreds of years (for reviews, see Valdes et
al., 1983; Sheffler and Roth, 2003). The active ingredient of S.
divinorum is salvinorin A, a non-nitrogenous neoclerodane
diterpene that represents the most potent naturally occur-
ring hallucinogen known (Valdes et al., 1984; Siebert, 1994).
experience qualitatively distinct from that induced by the
classical hallucinogens lysergic acid diethylamide, psilocy-
bin, and mescaline (Siebert, 1994). Both S. divinorum and
salvinorin A have been used recreationally for their halluci-
nogenic properties (Giroud et al., 2000). Intriguingly, an an-
ecdotal case report has suggested that S. divinorum may
have antidepressant properties as well (Hanes, 2001).
Quite recently, we discovered that salvinorin A has high
affinity and selectivity for the cloned ␬-opioid receptor (KOR)
and suggested, based on limited functional studies, that
salvinorin A was a KOR agonist (Roth et al., 2002). We now
present a detailed report on the agonist properties of salvi-
norin A and selected derivatives. We discovered that salvi-
norin A is an extraordinarily efficacious and potent ␬-opioid
The work was supported by U.S. Public Health Service Grant RO1 DA04123
from National Institute on Drug Abuse (to C.C.) and by the National Institute
of Mental Health Psychoactive Drug Screening Program and KO2MH01366
and RO1DA017204 (to B.L.R.).
Article, publication date, and citation information can be found at
http://jpet.aspetjournals.org.
DOI: 10.1124/jpet.103.059394.
ABBREVIATIONS: KOR, ␬-opioid receptor; hKOR, human ␬-opioid receptor; nor-BNI, nor-binaltorphimine; U50488, (trans)-3,4-dichloro-N-
methyl-N-[2-(1-pyrrolidinyl)-cyclohexyl] benzeneacetamide methane-sulfonate hydrate; U69593, (ϩ)-(5␣,7␣,8␤)-N-methyl-N-[7-(1-pyrrolidinyl)-1-
oxaspiro[4.5]dec-8-yl]-benzeneacetamide.
1197

1198
Chavkin et al.
agonist. We also found, based on structure-function studies, normal oocyte saline buffer was modified to increase the KCl con-
centration to 24 mM Kϩ. Microelectrodes were filled with 3 M KCl
and had resistances of 0.4 to 2.0 M⍀.
Radioligand Binding and Functional Studies. Radioligand
that the nature of the substituent on the 2-position of salvi-
norin profoundly affects functional activity. Together, these
results support the hypothesis that the unique effects of
binding studies were performed as described previously (Roth et al.,
salvinorin A on human perception are due to selective acti-
2002) with the exception that 150 mM NaCl was added to the
vation of KOR.
standard binding buffer to mimic physiological sodium concentra-
tions. In brief, membranes (10–50 ␮g) were incubated together with
[3H]bremazocine in a final volume of 0.5 ml with a buffer of the
following composition: 50 mM Tris-HCl, 150 mM NaCl, pH 7.40
along with test agents for 90 min at room temperature. Incubations
were terminated by rapid filtration and collection on GF/C glass fiber
filters and washing with ice-cold binding buffer. Dried filters were
put into sample vials, scintillation fluid was added, and dpm were
measured by liquid scintillation spectroscopy. Measurements of the
Materials and Methods
Materials. U50488, U69593, dynorphin A, norbinaltorphimine
(nor-BNI) were obtained from Sigma-Aldrich (St. Louis, MO).
[3H]Bremazocine was from PerkinElmer Life Sciences (Boston, MA).
Complementary DNA Clones and cRNA Synthesis for Oo-
cyte Studies. The rat KOR was obtained from Dr. David Grandy
(GenBank accession no. D16829). The human KOR cDNA was ob-
ability of KOR agonists to inhibit forksolin-stimulated adenylate
tained from the Guthrie Research Foundation (GenBank accession
cyclase activity were performed as detailed previously (Roth et al.,
no. NM000912) and subcloned into the eukaryotic expression vector
2002). For studies involving measurements of intracellular calcium
pIRESNEO (Invitrogen, Carlsbad, CA); cDNAs for KIR3.1 (accession
mobilization, a Molecular Devices FLEXSTATION was used as re-
no. U01071) and KIR3.2 (accession no. U11859) were obtained from
cently detailed (Rothman et al., 2003). For these studies, hKOR were
Drs. Cesar Lebarca and Henry Lester, respectively. The chimeric G
cotransfected with the chimeric G protein Gaq-i5 (Conklin et al.,
protein Gq-i5 was obtained from Bruce Conklin (University of Cali-
1993) or the “universal” G protein Ga16 (Offermanns and Simon,
fornia, San Francisco), whereas G␣16 was obtained from the Guthrie
1995). Measurements of intracellular calcium mobilization and
Research Foundation; both constructs were verified by automated
quantification of agonist efficacy and potency were performed as
dsDNA sequencing (Cleveland Genomics, Inc., Cleveland, OH) before
described in Rothman et al. (2003).
use. Plasmid templates for all constructs were linearized before
Data Analysis. EC50 values and curve fitting were determined
cRNA synthesis, and the mMESSAGE MACHINE kit (Ambion, Aus-
using Nfit software (Island Products, Galveston, TX) or GraphPad
tin, TX) was used to generate capped cRNA.
Prism (GraphPad Software, Inc., San Diego, CA). Student’s t test
Cell Lines and Maintenance. A stable line expressing the hu-
was used for comparison of independent means, with values reported
man KOR (hKOR-293) was obtained by transfecting an hKOR ex-
pression vector (hKOR-pIRESNEO) into human embryonic kidney-
293 cells (maintained and transfected as previously detailed; Roth et
al., 2002) and selecting in 600 ␮g/ml G418. Surviving clones were
as two-tailed p values.
Chemistry. Salvinorin A was isolated from dried leaves of S.
divinorum by the method reported previously (Valdes et al., 1984).
Salvinorin A was hydrolyzed using potassium carbonate in methanol
expanded and characterized with one (hKOR-293) that expressed
to yield salvinorin B. The reported esters were formed using salvi-
high levels of hKOR (ca. 1 pmol/mg) used for further studies.
Oocyte Maintenance and Injection. Healthy stage V and VI
oocytes were harvested from mature anesthetized Xenopus laevis
(Nasco, Ft. Atkinson, WI) and defolliculated enzymatically as de-
norin B, dimethylaminopyridine, and the corresponding acid chlo-
ride in methylene chloride.
Salvinorin B was characterized by 1H NMR, 13C NMR, and high-
resolution mass spectrometry (HRMS) and found to be authentic by
scribed previously (Snutch, 1988). The oocytes were maintained at
comparison with literature values (Valdes et al., 1984). The reported
18°C in standard oocyte buffer, ND96 (96 mM NaCl, 2 mM KCl, 1
esters were purified by high-performance liquid chromatography and
characterized by HRMS. 1H and 13C NMR spectra were recorded on
a Bruker AMX 500 MHz NMR spectrometer in CDCl3. The HRMS
were measured using a Bioapex FT mass spectrometer with electro-
spray ionization. High-performance liquid chromatography was con-
ducted on a Waters Deltaprep 4000 system using a Waters Xterra
RP18, 5 ␮m, 4.6 ϫ 150-mm column, with mobile phase H2O/acetoni-
trile (1:1). Thin layer chromatography analyses were carried out on
precoated Si gel G254, 250-␮m plates, with the developing system
hexane/ethyl acetate (2:1) and visualized with vanillin/H2SO4 in
ethanol.
mM CaCl2, 1 mM MgCl2, and 5 mM HEPES, pH 7.5), supplemented
with 2.5 mM sodium pyruvate and 50 ␮g/ml gentamicin (Sigma-
Aldrich). One day after harvest, cRNAs were injected (50 nl/oocyte)
with a Drummond microinjector. Each oocyte was injected with 0.5
ng of KOR cRNA and 0.1 ng of KIR3.1 and KIR3.2 cRNA. Recordings
were made at least 48 h after injection.
Electrophysiological Studies. An Axon Geneclamp 500 ampli-
fier was used for standard two-electrode voltage-clamp experiments.
The FETCHEX program (Axon Instruments, Foster City, CA) and
recorded data traces were used for data acquisition and analysis.
Oocytes were then removed from incubation medium, placed in the
recording chamber containing ND96 medium, and clamped at –80
mV. Recordings were made in hK buffer (72.5 mM NaCl, 24 mM KCl,
1 mM CaCl2, 1 mM MgCl2, and 5 mM HEPES, pH 7.5). To facilitate
Preparation of Esters. Salvinorin B (10 mg, 26 nmol) and 4-dim-
ethylaminopyridine (catalytic amount) were dissolved in methylene
chloride (3 ml). The corresponding acid chloride (130 nmol) was
the recording of inward Kϩ currents through the KIR3 channels, the added, and the reaction stirred at room temperature overnight. The
TABLE 1
Calculated molecular weights were obtained using ChemDraw software
Yields and Masses of Salvinorinyl Esters
Calculated
Found(M ϩ 23)for sodium
1) Propionate
9.0 mg, 78.5%
446.1941
502.2567
474.2254
572.1046
564.0721
462.1890
538.1839
544.2097
458.1941
469.1917
525.2566
497.2215
595.1009
587.0689
485.1833
561.1834
567.2087
481.1952
2) Heptanoate
10.5 mg, 81.6%
11.1 mg, 91.4%
12.4 mg, 84.4%
11.5 mg, 79.4%
9.8 mg, 82.7%
1.6 mg, 11.6%
2.1 mg, 15.1%
10.5 mg, 89.4%
3) Pivalate
4) p-Bromobenzoate
5) 2,2,2-Trichloroethylcarbonate
6) Ethylcarbonate
7) Piperonylate
8) 1-Naphthoate
9) Cyclopropanecarboxylate

Salvinorin A Activates ␬-Opioid Receptors
1199
Fig. 1. Structures of salvinorin A, B, and 2-salvinorinyl esters. Shown are the structures of the compounds used in this study.
mixture was quenched with methanol, loaded onto silica, and puri- no significant activity at other receptors, including various
fied by vacuum liquid chromatography using Si gel (230–400-mesh)
with hexane/ethyl acetate (3:1) solvent system. Calculated molecular
weights were obtained using ChemDraw software (Table 1).
serotonergic, dopaminergic, muscarinic, adrenergic, cannabi-
noid, and ␴ receptors (see Table 2 for details)
We next evaluated the ability of salvinorin A and the
propionate and heptanoate derivatives to activate hKORs by
measuring the ability to inhibit forskolin-stimulated cAMP
production using U69593 as the comparator. As shown in
Table 2, salvinorin A and salvinorinyl-2-propionate were po-
tent and full agonists compared with U69593, whereas salvi-
norinyl-2-heptanoate was a partial agonist.
Results
In initial studies, we examined the abilities of salvinorin A
and selected derivatives (see Fig. 1 for structures) for their
ability to bind to hKORs. As can be seen, the synthetic
derivatives of salvinorin A differ solely in the nature of the
substituent in the 2-position. As is shown in Table 2, salvi-
norinyl-2-propionate was the only derivative with submicro-
molar affinity for hKORs; also of note is that salvinorin B was
inactive at hKORs. A screen of a number of other receptor
subtypes showed that the salvinorin A derivatives tested had
We also evaluated the ability of U69593, dynorphin A,
salvinorin A, and the propionate derivative of salvinorin A to
activate hKORs using
a
fluorescent-microplate-reader
(FLEXSTATION) wherein hKORs were cotransfected with
either the chimeric G protein Gqi5 or the universal G protein

1200
Chavkin et al.
TABLE 2
represents the mean response measured in four to seven
different oocytes. Data were collected from multiple batches
of oocytes and merged by normalizing the responses to the
average maximal response produced by salvinorin A on that
recording day. Based on these results, salvinorin A was not
significantly more potent (EC50 ϭ 69 nM; confidence inter-
vals 50–94 nM) than U69593 (EC50 ϭ 224 nM; confidence
intervals 51–157 nM) or U50488 (EC50 150 nM; confidence
intervals 50–194 nM).
Under these expression conditions, there was an apparent
lack of spare ␬-receptors. Increasing the ␬-receptor cRNA
from 0.5 ng/oocte to 1.0 ng increased the average U69593
response from 1.63 Ϯ 0.57 to 2.76 Ϯ 1.04 ␮A (n ϭ 7 or 8).
Based on the lack of spare receptors, we directly compared
the maximal responses evoked by 10 ␮M each of the ␬-ago-
nists (Fig. 5) with that of dynorphin A. In this assay, propio-
nyl-salvinorin also acted as a partial agonist whose maximal
activity was less than salvinorin A. The response to salvi-
norin A was significantly greater than that to U69593 and
U50488 (p Ͻ 0.05), but not significantly greater to that of
dynorphin A.
Effect of salvinorin A derivatives on KOR binding and inhibition of
forskolin-stimulated adenylate cyclase in KOR-293 cells
Shown are the mean values Ϯ S.D. from n ϭ 2–4 separate experiments in which Ki
values for inhibition of [3H]bremazocine binding and EC50 and Emax values for
inhibition of adenylate cyclase in KOR-293 cells were performed as detailed under
Materials and Methods with the response induced by U69593 defined as 100%. The
salvinorin A derivatives listed above were also screened at a large number of cloned
receptors and found to have no significant activity, when tested at 10 ␮M at the
following receptors: serotonin (5-HT1A, 5-HT1B, 5-HT1D, 5-HT1E, 5-HT2A,
5-HT2B, 5-HT2C, 5-HT3, 5-HT5a, 5-HT6, 5-HT7), dopamine (D1, D2, D3, D4, D5),
muscarinic (m1, m2, m3, m4, m5), ␮, ␦, and ORL-1 opioid receptors, ␴1, ␴2, ␣1-
adrenergic (1a, 1b, 1d), ␣2-adrenergic (2A, 2B, 2C) ␤2-adrenergic, and CB-1 canna-
binoid receptors [assayed as previously detailed (Shi et al., 2003)].
EC50 in nM
Ki Ϯ S.E.M. (nM)
Emax
(pEC50 Ϯ S.E.M.)
Salvinorin A
18.74 Ϯ 3.38
32.63 Ϯ 15.7
3199 Ϯ 961.2
Ͼ10,000
0.63 (Ϫ0.2 Ϯ 0.07)
100
100
Propionate
4.7 (0.7 Ϯ 0.3)
Heptanoate
40 (1.6 Ϯ 0.4)
34 Ϯ 11
NA
Privalate
NA
NA
NA
NA
NA
NA
NA
NA
p-Bromobenzoate
2,2,2-Triethylcarbonate
Ethylcarbonate
Piperonylate
Ͼ10,000
NA
Ͼ10,000
NA
Ͼ10,000
NA
Ͼ10,000
NA
1-Napthoate
Cyclopropanecarboxylate
Salvinorin B
Ͼ10,000
NA
Ͼ10,000
NA
Ͼ10,000
NA
NA, not active at 10,000 nM.
Discussion
Ga16 as detailed previously (Rothman et al., 2003). Figure 2
shows representative results for U69593 and salvinorin A
The principal finding of this study is that salvinorin A is an
using either G␣16 (A and B) or Gq-i5 (C and D). No responses extraordinarily potent full agonist at hKORs. Additionally,
were seen in untransfected cells or in cells transfected with we report that salvinorinyl-2-propionate is a potent partial
hKOR alone (data not shown). Figure 2 also shows a repre- agonist at KORs and also demonstrate that the nature of the
sentative dose-response study using Gq-i5 as the chimeric G 2-substituent of the salvinorin scaffold is critically important
protein. Because both methods seemed to yield equivalent for agonist efficacy and potency. We also have obtained data
results, further studies were performed with Gq-i5. Table 3 with KOR-knockout and wild-type mice that the actions of
shows representative EC50 and Emax values for a variety of salvinorin A are mediated by KOR in vivo (J. Pintar, personal
KOR agonists using Gq-i5. In these studies, salvinorin A was communication). Together, these results imply that the pro-
more potent than any other of the tested KOR agonists (Table found effects of salvinorin A on human consciousness are
3). In terms of maximal response, all of the active compounds mediated by potent and highly efficacious activation of
gave similar responses.
It is well known that overexpression systems tend to pro-
KORs.
In prior reports, we have suggested that because salvinorin
vide inaccurate estimates of agonist potencies and efficacies A is a potent hallucinogen that is apparently selective for
because of issues of receptor reserve (Kenakin, 2002). As KORs, and that targeting KORs might lead to novel medica-
well, it has been well described that unnatural expression tions for the treatment of diseases manifested by hallucina-
systems wherein chimeric or “universal” G proteins are used tory experiences (e.g., schizophrenia, affective disorders, and
also lead to misleading estimates of agonist potencies and dementia) (Roth et al., 2002; Sheffler and Roth, 2003). In this
maximal responses (Woolf et al., 2001; Kenakin, 2002). Ac- regard, studies with nonselective opioid antagonists that pos-
cordingly, we next determined the maximal agonist re- sess KOR actions in schizophrenia have been mixed (Rapa-
sponses (Emax) and potencies (EC50 values) of selected com- port et al., 1993; Sernyak et al., 1998), although there are no
pounds using a system without receptor reserve.
studies in which selective KOR antagonists have been tested.
Salvinorin A Is a Full agonist. For these studies, Xeno- Because of anecdotal reports that extracts of S. divinorum
pus oocytes were coinjected with inwardly rectifying Kϩ may possess antidepressant actions (Hanes, 2001), and pub-
channels and KORs. In the experiment shown, a representa- lished studies in rodents that KOR antagonists block stress-
tive oocyte voltage clamped at –80 mV was first perfused induced responses (McLaughlin et al., 2003), KOR antago-
with hK buffer (containing 24 mM KCl) to shift the reversal nists could possess antianxiety/antidepressant actions as
potential of potassium and facilitate Kϩ current through well. Indeed, a recent study (Mague et al., 2003) suggested
Kir3 (Fig. 3). Perfusion with 1 ␮M salvinorin A significantly that ␬-selective antagonists might have intrinsic antidepres-
increased the inward current, and the activation was re- sant actions. Our current studies suggest that novel KOR-
versed by 100 nM nor-BNI. Similarly, 1 ␮M U69593 in- selective agents might be obtained by selective modification
creased the inward current in a different oocyte, and the of the salvinorin scaffold. Whether such agents might possess
effect was also blocked by 100 nM nor-BNI (Fig. 1B). Neither antidepressant or antipsychotic activity is unknown.
10 ␮M salvinorin A nor U69593 increased the membrane
conductance of oocytes expressing Kir3 without KOR (data 2-propionate are potent agonists at KORs with salvinorin A
not shown). being a full agonist in most assay systems, whereas salvi-
As shown in these studies, salvinorin A and salvinorinyl-
Concentration-response curves of salvinorin-A and ␬-ago- norinyl-2-propionate is likely a partial agonist. Salvinorin B
nists U69593 and U50488 were compared (Fig. 4). Each point and all other tested salvinorin derivatives were devoid of

Salvinorin A Activates ␬-Opioid Receptors
1201
Fig. 2. Salvinorin A mobilizes intracellular Ca2ϩ when hKORs are cotransfected with the universal G protein G16 or the chimeric G protein. For these
studies human embryonic kidney-293 cells were transfected with hKOR and either Gqi5 or G16 and the mobilization of intracellular calcium
quantified as described previously (Rothman et al., 2004) using a 96-well FLEXSTATION. A and B, representative results with increasing doses of
U69593 or salvinorin A (0, 10, and 100 nM), whereas hKORs were cotransfected with G16. C and D, results obtained when hKORs were cotransfected
with Gq-i5. E, average for n ϭ 3 separate experiments for dose-response studies to salvinorin A and U69593.
significant activity. One potential complication of the studies actions of salvinorin A and other compounds at KORs ex-
performed on recombinant, overexpressed receptors relates pressed in Xenopus oocytes.
to the issue of receptor reserve. Thus, it is widely appreciated
KOR expressed in Xenopus oocytes activate intrinsic G
that overexpressing G proteins and/or receptors in heterolo- proteins that then increase the conductance of coexpressed
gous expression systems leads to inaccurate estimates of G protein-coupled inwardly rectifying potassium channels
agonist potencies and maximal responses (for review, see (GIRK and Kir3) (Henry et al., 1995). Injection of cRNAs
Kenakin, 1997). Accordingly, we also evaluated the agonist coding for the mammalian receptor and channel has been

1202
Chavkin et al.
TABLE 3
Salvinorin A and salvinorinyl-2-propionate are agonist at hKOR-
stimulated intracellular Ca2ϩ mobilization: comparison with reference
compounds
Data represent mean Ϯ S.D. of quadruplicate determinations for EC50 and Emax for
mobilization of intracellular calcium.
Emax
Drug
EC50 in nM (pEC50 Ϯ SD)
(Relative to U69593)
U69593
13 (1.14 Ϯ 0.2)
24 (1.39 Ϯ 0.14)
7 (0.84 Ϯ 0.07)
83 (1.92 Ϯ 0.17)
17.3 (1.23 Ϯ 0.18)
100
U50488
102 Ϯ 4
104 Ϯ 7
107 Ϯ 8
102 Ϯ 8
Salvinorin A
Dynorphin A
Salvinorinyl-2-
propionate
Salvinorin B
No activity
No activity
demonstrated to faithfully reconstitute opioid signaling in Fig. 4. Concentration-response curve of salvinorin A, and ␬-agonists
U69593 and U50488. Cumulatively higher concentrations of salvinorin A
oocytes equivalent to that observed in guinea pig substan-
tia gelatinosa neurons (Grudt and Williams, 1993). In ad-
and the ␬-agonists were applied to the bath. The agonist response at each
concentration was normalized as a percentage of the maximal salvinorin
dition, by controlling the levels of receptor and channel
expression, spare receptors can be avoided and the peak
responses produced by different drugs can be a direct
measure of agonist efficacy. The in vitro bioassay also
eliminates pharmacokinetic barriers, and the electrophys-
iological recording of channel activation provides a rapid
measure of receptor activation. In this study, we compared
the relative activity of salvinorin A with three compounds
having established ␬-opioid receptor agonist activity.
Salvinorin A was found to be more potent and have higher
efficacy than either U50488 and U69593. The agonist ef-
ficacy of salvinorin A was not significantly different from
dynorphin A(1-17), an endogenous neurotransmitter of the
␬-opioid receptor (Chavkin et al., 1982).
Structure-activity relationship studies show that the KOR
agonistic activity of salvinorin derivatives depend largely on
the size and character of the substituent on the 2-ester moi-
ety. Generally, the studied derivatives have either lower
affinity for KOR than salvinorin A or are completely devoid of
activity. The two active derivatives, the propionate and the
heptanoate, demonstrate that as the alkyl chain is length-
A response. Each point represents the mean response measured in four to
seven different oocytes.
length must not be the only factor, because the short-chain
ethylcarbonate derivative is absent of activity.
The current results support the conclusion that just as
morphine is a natural plant product able to activate the
␮-opioid receptor, salvinorin A is a natural plant product able
to activate the KOR. The strongly psychotomimetic actions of
salvinorin A suggest that the dynorphin/␬-opioid system may
have a role in the regulation of cognition and perception and
support earlier proposals that some forms of schizophrenic
hallucinations may be caused by hyperactive endogenous
opioid systems (Gunne et al., 1977). Recent data implicating
the KOR-dynorphinergic system in modulating stress and
anxiety responses in rodents suggest that targeting KORs
might also lead to novel antidepressant and anxiolytic med-
ications. Salvinorin A, by virtue of its potency, efficacy, and
selectivity as a KOR agonist will be an important tool for
discovering the role that the KOR-dynorphinergic system has
ened, KOR affinity diminishes. Interestingly however, chain in health and disease.
Fig. 3. Salvinorin A is a highly efficacious ␬-receptor agonist. Representative traces showing the change in current during a typical experiment. A large
inward current was apparent as the Kϩ concentration was increased from 2 to 24 mM in normal oocyte saline buffer. Salvinorin A (1 ␮M) and U69593
(1 ␮M) in the buffer (24 mM Kϩ) further increased Kir3 currents, and the response was reversed by nor-BNI (100 nM), a ␬-antagonist.

Salvinorin A Activates ␬-Opioid Receptors
1203
receptors couple to inwardly rectifying potassium channels when coexpressed by
Xenopus oocytes. Mol Pharmacol 47:551–557.
Kenakin T (1997) Differences between natural and recombinant G protein-coupled
receptor systems with varying receptor/G protein stoichiometry. Trends Pharma-
col Sci 18:456–464.
Kenakin T (2002) Efficacy at G-protein-coupled receptors. Nat Rev Drug Discov
1:103–110.
Mague SD, Pliakas AM, Todtenkopf MS, Tomasiewicz HC, Zhang Y, Stevens WC Jr,
Jones RM, Portoghese PS, and Carlezon WA Jr (2003) Antidepressant-like effects
of kappa-opioid receptor antagonists in the forced swim test in rats. J Pharmacol
Exp Ther 305:323–330.
McLaughlin JP, Marton-Popovici M, and Chavkin C (2003) Kappa opioid receptor
antagonism and prodynorphin gene disruption block stress-induced behavioral
responses. J Neurosci 23:5674–5683.
Offermanns S and Simon MI (1995) G alpha 15 and G alpha 16 couple a wide variety
of receptors to phospholipase C. J Biol Chem 270:15175–15180.
Rapaport MH, Wolkowitz O, Kelsoe JR, Pato C, Konicki PE, and Pickar D (1993)
Beneficial effects of nalmefene augmentation in neuroleptic-stabilized schizo-
phrenic patients. Neuropsychopharmacology 9:111–115.
Roth BL, Baner K, Westkaemper R, Siebert D, Rice KC, Steinberg S, Ernsberger P,
and Rothman RB (2002) Salvinorin A: a potent naturally occurring nonnitrog-
enous kappa opioid selective agonist. Proc Natl Acad Sci USA 99:11934–11939.
Rothman RB, Vu N, Partilla JS, Roth BL, Hufeisen SJ, Compton-Toth BA, Birkes J,
Young R, and Glennon RA (2003) Related in vitro characterization of ephedrine-
related stereoisomers at biogenic amine transporters and the receptorome reveals
selective actions as norepinephrine transporter substrates. J Pharmacol Exp Ther
307:138–145.
Sernyak MJ, Glazer WM, Heninger GR, Charney DS, Woods SW, Petrakis IL,
Krystal JH, and Price LH (1998) Naltrexone augmentation of neuroleptics in
schizophrenia. J Clin Psychopharmacol 18:248–251.
Sheffler DJ and Roth BL (2003) Salvinorin A: the “magic mint” hallucinogen finds a
molecular target in the kappa opioid receptor. Trends Pharmacol Sci 24:107–109.
Shi Q, Savage JE, Hufeisen SJ, Rauser L, Grajkowska E, Ernsberger P, Wroblewski
JT, Nadeau JH, and Roth BL (2003) L-Homocysteine sulfinic acid and other acidic
homocysteine derivatives are potent and selective metabotropic glutamate recep-
tor agonists. J Pharmacol Exp Ther 305:131–142.
Siebert DJ (1994) Salvia divinorum and salvinorin A: new pharmacologic findings. J
Ethnopharmacol 43:53–56.
Snutch TP (1988) The use of Xenopus oocytes to probe synaptic communication.
Trends Neurosci 11:250–256.
Valdes LJ, Butler WM, Hatfield GM, Paul AG, and Koreeda M (1984) Divinorin A: a
psychotropic terpenoid and divnorin B from the hallucinogenic Mexican mint
Salvia divinorum. J Org Chem 49:4716–4720.
Valdes LJ 3rd, Diaz JL, and Paul AG (1983) Ethnopharmacology of ska Maria
Pastora (Salvia divinorum, Epling and Jativa-M.). J Ethnopharmacol 7:287–312.
Wasson RG (1962) A new Mexican psychotropic drug from the mint family. Bot Mus
Leaflets Harvard Univ 20:77–84.
Wasson RG (1963) Notes on the present status of ololuiqui and the other hallucino-
gens of Mexico. Bot Mus Leaflets Harvard Univ 20:163–193.
Woolf PJ, Kenakin TP, and Linderman JJ (2001) Uncovering biases in high through-
put screens of G-protein coupled receptors. J Theor Biol 208:403–418.
Fig. 5. Salvinorin A is more efficacious than U69593 and U50488 in
␬-receptor-mediated activation of Kir3 currents. At saturating concentra-
tion, salvinorin A (10 ␮M) evoked a large Kir3 currents, which were
significantly higher than the response evoked by U69593 (10 ␮M) or
U50488 (10 ␮M). Data are mean Ϯ S.E.M.; ءء, p Ͻ 0.05. Dynorphin A (10
␮M) produced a response that was not significantly different from salvi-
norin A.
References
Chavkin C, James IF, and Goldstein A (1982) Dynorphin is a specific endogenous
ligand of the kappa opioid receptor. Science (Wash DC) 215:413–415.
Conklin BR, Farfel Z, Lustig KD, Julius D, and Bourne HR (1993) Substitution of
three amino acids switches receptor specificity of Gq alpha to that of Gi alpha.
Nature (Lond) 363:274–276.
Giroud C, Felber F, Augsburger M, Horisberger B, Rivier L, and Mangin P (2000)
Salvia divinorum: an hallucinogenic mint which might become a new recreational
drug in Switzerland. Forensic Sci Int 112:143–150.
Grudt TJ and Williams JT (1993) kappa-Opioid receptors also increase potassium
conductance. Proc Natl Acad Sci USA 90:11429–11432.
Gunne LM, Lindstrom L, and Terenius L (1977) Naloxone-induced reversal of schizo-
phrenic hallucinations. J Neural Transm 40:13–19.
Hanes KR (2001) Antidepressant effects of the herb Salvia divinorum: a case report.
J Clin Psychopharmacol 21:634–635.
Henry DJ, Grandy DK, Lester HA, Davidson N, and Chavkin C (1995) Kappa opioid
Address correspondence to: Dr. Bryan Roth, Department of Biochemistry;
Room RT500-9, Case Western Reserve University Medical School, 2109 Adel-
bert Rd., Cleveland, OH 44106. E-mail: roth@biocserver.cwru.edu