Enantioselective access to key intermediates for salvinorin a and analogues

Article abstract of DOI:10.1002/ejoc.201100207

Access to enantiopure synthetic platforms that can generate key intermediates for salvinorin A analogues through diastereoselective Diels-Alder cycloaddition between an enantiopure sulfinylquinone and semicyclic dienes is described.

Full text of DOI:10.1002/ejoc.201100207

FULL PAPER
DOI: 10.1002/ejoc.201100207
Enantioselective Access to Key Intermediates for Salvinorin A and Analogues[‡]
Don Antoine Lanfranchi,[a] Christophe Bour,[a] and Gilles Hanquet*[a]
Keywords: Natural products / Total synthesis / Synthesis design / Cycloaddition / Enantioselectivity / Polycycles
Access to enantiopure synthetic platforms that can generate
key intermediates for salvinorin A analogues through dia-
stereoselective Diels–Alder cycloaddition between an
enantiopure sulfinylquinone and semicyclic dienes is de-
scribed.
methoxycarbonyl group, iii. the 17-position carbonyl, and
iv. the 12-position furan ring.[2,7] Figure 1 summarizes the
structure–activity relationships for salvinorin A analogues.
Despite the fact that the nature of the interactions at the
C(2) and C(4) positions are well-understood, those at the
C(12) position remain unclear. Two salvinorin A···KOR in-
teraction models have been proposed so far.[8] The hydro-
phobic nature of salvinorin A inherently lends itself to re-
cognition through hydrophobic interactions.[8b] According
to the second model,[8a] the furan oxygen may act as a hy-
drogen-bond acceptor to tyrosine moieties of KOR.
Introduction
Among the isolated natural products exhibiting psychodys-
leptic activities, salvinorin A,[1] which displays strong neu-
robiological activity (highly specific kappa opoid receptor –
KOR – ligand) with high medicinal potential and a chal-
lenging chemical structure,[2] has emerged as a target of par-
ticular interest because of the difficulties involved in mod-
ifying the structure of the extracted natural product for
SAR studies. Salvinorin A is now recognized as a lead
structure for the design of novel opioid receptor-specific li-
gands.[3] Interestingly, although salvinorin A (1) has low af-
However, both the electron density of the heteroatom
finity for mu opoid receptors (MOR),[4] it has been shown acting as a hydrogen-bond acceptor at C(12), and the
to be an allosteric modulator of MOR.[5] To date, structure– hydrophobic interactions between the C(12) moiety and
activity relationships of 1 have focused on several main KOR are critical to the binding affinity of salvinorin A tem-
areas: i. the 2-position acetoxy group,[6] ii. the 4-position plate analogs to KOR. This hypothesis has been recently
Figure 1. Summary of structure–activity relationships for salvinorin analogues.
supported by the preparation of 12-epi-salvinorin A, which
exhibits partial agonist properties at KOR with high selec-
tivity.[7] Furthermore, the authors demonstrated that inver-
sion at C(12) has little effect on the position of the furan
ring: superimposition with the crystal structure of 1 shows
that C(13) and the furan oxygen atom are almost coincident
in the two structures. That this subtle change in position
[‡] Total Synthesis of Salvinorin A and Analogues, 1 (Part 1 of a
new article series)
[a] Laboratoire de Stéréochimie Associé au CNRS, ECPM, UdS,
25 rue Becquerel, 67087 Strasbourg, France
Fax: +33-3-68852442
E-mail: gilhanquet@unistra.fr
ghanquet@chimie.u-strasbg.fr
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201100207.
2818
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 2818–2826

Total Synthesis of Salvinorin A and Analogues
Scheme 1. Diastereoselective Diels–Alder reaction as a key step in the proposed strategy.
and orientation of the furan ring should convert a full ago- positions for pharmacological purposes (R1, R2, X, Y, Z)
nist into a partial one perhaps represents a key to the mode and allows access to numerous analogues as well as dia-
of action of 1. Furthermore, no studies have evaluated the stereoisomers that are not available by chemical modifica-
importance, if any, of the C(19) and C(20) methyl groups. tion of the natural product. This method is also an ideal
From their locations on the salvinorin skeleton, modifica- tool to introduce both angular methyl groups (R3, R4),
tions to these two quaternary centers are clearly dificult. either simultaneously or sequentially, by choosing the ap-
Finally, the influence of the size of the C-ring has not been propriate sulfinylquinone and diene precursors, therefore
investigated. All these parameters should influence the posi- potentially giving access to nor-salvinorins or aromatic ana-
tioning of the furan ring at C(12) within the receptor, which logues: cycle A and/or A/B (Scheme 1). Finally, this strategy
is critical to the binding affinity at KOR.
allows a modification of the size of the C-ring by modifica-
Four years ago, we embarked on studies towards the to- tion of the dihydropyran skeleton of B (five- or seven-mem-
tal synthesis of salvinorin A through a highly convergent bered ring).
strategy that allows easy modifications at C(2), C(4), and
C(12), as well as allowing modifications to both quaternary demonstrate that the strategy depicted in Scheme 1 (C; R3
centers, C-ring size, and relative configuration at C(12).[9] = Me and B; R4 = H, Y = H) can be used to open a route
In this paper, we report our preliminary results, which
Two total syntheses of salvinorin A have been reported to enantiomerically pure synthetic platforms that represent
so far.[10a] One approach, which was developed by Evans the building blocks of many analogues of salvinorin A.
and co-workers,[10b] required 29 steps and involved a trans- Modifications of the aromatic part at C(12) (i.e., 2-furyl) of
annular Michael reaction cascade, whereas the second ap- the C-ring size (five- or seven-membered rings) and of the
proach, developed in two phases by Hagiwara’s group,[11] quaternary centre at C(5) and/or C(9), should lead to inter-
required 20[11a] or 15[11b] steps from a Wieland–Miescher esting analogues for SAR studies that would give a better
ketone.[12]
understanding of interactions with KOR. We have recently
Our own strategy is based on a highly diastereoselective published the preparation of dienes B with cyclic O,O-
Diels–Alder cycloaddition of a semicyclic dihydropyran- acetals (X = O, Y = OEt) – as precursors of the C(17)
based diene B with an enantiomerically pure sulfinylquin- carbonyl group of salvinorin A or hemicetal group of salvi-
one C (Scheme 1). The strength of this strategy is the si- norin J[13] – by ring-closing metathesis of the corresponding
multaneous creation of multiple stereogenic centers, and the 1,7- and 1,8-enynes of propargylic O,O-acetals.[14] Cycload-
whole tricyclic skeleton A (on a multigram scale) of salvino- dition of these dienes with C (R3 = Me), as well as of dienes
rin A with the required functional groups for its total syn- B with a vinylic methyl group (R4 = Me) with C (R3 = H)
thesis. It also allows a variation of the different important will be disclosed in due course.
Eur. J. Org. Chem. 2011, 2818–2826
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2819

G. Hanquet et al.
FULL PAPER
loss of affinity for KOR,[7,20] the previously unknown diene
(Ϯ)-7d was chosen for its aromatic moiety because it could
be derivatized into a quinone and also because it bears the
same aromatic substitution pattern as 2,5-dimethoxy-4-
methylamphetamine (DOM), which is a serotoninegic psy-
chedelic and 5-HT2 neuroligand.[21] RCEM reactions lead-
ing to dienes 7a–d were performed using Grubbs 1st genera-
tion catalyst (1 mol-%), however, these conditions were to-
tally ineffective for the RCEM of the precursors of 8a, 8b,
9a, and 9b. Finally, seven-membered ring based dienes 8a
and 8b were obtained in good yields (80 and 70%, respec-
tively),[14] as well as the corresponding five-membered rings
9a and 9b (72 and 68%, respectively) by reaction with
Grubbs 2nd generation catalyst (5 mol-%) in dichloro-
ethane under an ethylene atmosphere at 80 °C.[22]
Results and Discussion
For initial studies, we selected the model 1,3-diene 3[15]
as a suitable substrate for reaction development with enan-
tiomerically pure sulfinylquinone 2 (Scheme 2). We have al-
ready demonstrated that (SS)-6-methoxy-2-(p-tolylsulfinyl)-
3-methyl-1,4-benzoquinone (2) is a very efficient building
block for the construction of functionalized decalins with
an angular methyl group, and we have disclosed a new ac-
cess to Wieland–Miescher diketone analogues.[16] (SS)-Sulf-
inylquinone 2 was prepared on large scale within 10 days
from methylhydroquione.
By analogy with our previous study,[16] we anticipated
that precomplexation of the sulfinylquinone 2 with ZnBr2
would be necessary to obtain good conversion in reasonable
time. Thus, by mixing (SS)-sulfinylquinone 2 with 2 equiv.
of ZnBr2 at –20 °C in dichloromethane and adding 2 equiv.
of diene 3 after a few minutes, the cycloaddition occurred
smoothly and complete conversion was observed after 10 h
(Scheme 2).
During our first attempts, we obtained, after work up
and purification on demetalated silica gel,[17] very low yields
of cycloadduct 4 (15%) and a quantitative amount of thio-
sulfinate 5. The quantitative isolation of the latter showed
clearly that the cycloaddition occurred completely, because
it resulted from the spontaneous desulfinylation reaction of
the cycloadduct at room temperature. Finally, we found that
rapid filtration at –20 °C of the crude material on standard
silica gel, using a mixture of CH2Cl2/Et2O (1:1), delivered
the cycloadduct 6 in 95% isolated yield. This first set of
experiments, together with a recent paper by the group of
Perlmutter[18] describing access to a racemic precursor of
20-nor-salvinorin using a similar strategy, confirmed that
rapid access to synthetic platforms for the preparation of
salvinorin A and analogues as depicted in Scheme 1, was
viable.
Figure 2. Semicyclic dienes 7, 8 and 9 prepared as substrates for
cycloaddition reactions with dienophile (S)-2.
To demonstrate the versatility of our strategy for the
preparation of a range of salvinorin A analogues, we pre-
Returning to our model reaction (Scheme 2), we ob-
pared different semicyclic dienes 7a–d, 8a, 8b, 9a, and 9b served that cycloadduct 6 was stable at –20 °C for several
using ring-closing enyne metathesis (RCEM) under condi- months (even years) and that it was possible to perform a
tions developed by Mori[19] (Figure 2). Although replacing desulfinylation using Raney-Ni in cold MeOH (–20 °C) to
the furan ring with another heteroaromatic ring results in give the corresponding cis-decalin 10 in 70% yield. Interest-
Scheme 2. Diels–Alder cycloaddition of sulfinylquinone 2 with model diene 3.
2820
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 2818–2826

Total Synthesis of Salvinorin A and Analogues
ingly, applying Luche reduction conditions[23] to cycload-
duct 6 stereoselectively afforded alcohol 11 in 81% isolated
yield as pure crystals (Scheme 3), showing that the two
carbonyl groups at positions C(2) and C(4) can be easily
differentiated.
Figure 3. X-ray structures of 11, 12β-12b, and 12α-12a.
Simple crystallization in cold diethyl ether of a dia-
stereomeric mixture of 12 afforded pure 12α-12a–c, which
were stable at –20 °C, and which were characterized by
NMR analysis.
Scheme 3. Desulfinylations and Luche reduction of cycloadduct 6
at –20 °C.
Crystal structures of 12β-12b and 12α-12a were obtained
A crystal structure of 11 was obtained that confirmed its and their X-ray analyses confirmed their absolute configu-
relative and absolute configuration (Figure 3) as rations to be (SS,5S,9S,10R,12R) and (SS,5S,9S,10R,12S),
(SS,1R,5S,9S,10R), which is in agreement with our pre- respectively (see Figure 3).
viously proposed model.[16] The decomposition observed
Similar results were observed with dienes 7d, 8a and 9a.
when cycloadduct 6 was warmed to room temperature The corresponding cycloadducts were obtained in 1 d, 10 h,
(Scheme 2) was attributed to a reaction between the sulfenic and 7 d, respectively, with α/β ratios of 95:5, 80:20 and
acid (p-TolSOH) – resulting from the syn-elimination of the 65:35 (Scheme 5).[25]
sulfoxide moiety – and the newly formed cycloadduct 4.
Compounds 12a–c were desulfinylated with Raney-Ni at
Classical sulfenic acid scavengers (Michael acceptors, PPh3) –20 °C, and the resulting trans-decalins 13a–c were ob-
failed to trap p-TolSOH. However, when cycloadduct 6 was tained in 60–67% yields (Scheme 6). In contrast to the cy-
warmed (room temp.) in pyridine, compound 4 was ob- cloadduct 6, which was desulfinylated with retention of
tained in good yield (see Scheme 3).
configuration to the cis-decalin 10, a total epimerization at
With these preliminary results in hand, we added racemic C(10) was observed for cycloadducts 12, leading to trans-
dienes 7a–d, 8a and 9a (2 equiv.) to (SS)-sulfinylquinone 2 decalins 13 during the desulfinylation process. The relative
(1 equiv.) under conditions similar to those described above configuration of the decalin skeleton in cycloadducts 13
for 3. In the case of 7a–c, the resulting cycloadducts 12a–c (trans) and 10 (cis) was unambiguously assigned by
were isolated with high yields (85–90%). As could be ex- NOESY experiments, and by measuring the coupling con-
pected from previous work,[24] we observed a kinetic resolu- stants between protons at C(10) and C(9) (10.2 Hz for 13
tion at the benzylic position of the diene during the cyclo- and 5 Hz for 10).
addition, which depended on the nature of the substituents.
To obtain the required configuration at C(12), we pre-
Excess of diene 7 was recovered as an S-(–)-enantioenriched pared enantiomerically enriched (S)-7a–c and (R)-7a–c
mixture, showing that the R-(+)-enantiomer of 7 was more using enantioselective enzymatic hydrolysis of the corre-
reactive; this is in complete agreement with known models sponding homoallylic acetate.[26] As the cycloaddition pro-
(see Scheme 4).
ceeded smoothly within one day with the (R)-enantiomers
Eur. J. Org. Chem. 2011, 2818–2826
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2821

G. Hanquet et al.
FULL PAPER
Scheme 4. Matched and mismatched approaches between enantiomerically pure dienes 7 and (+)-(S)-sulfinylquinone 2.
In the second case, the resulting 12β-12a–c cycloadducts
[“natural C(12) epimers”] were much less stable compared
to their C(12) epimers, and the crude reaction mixtures
were directly treated with Raney-Ni at –20 °C to give de-
sulfinylated compounds (+)-12β-13a–c[28] in enantiomer-
ically enriched form in 63–67% yield. Conducting an elimi-
nation reaction of the same intermediates lead to unsatu-
rated dienones 12β-14a–c in 52–63% yield. In contrast to
their 12α epimers, these 12β dienones were much more
stable. Finally, (+)-12β-13a,b were obtained in 45% and
55% overall yield, which compares favorably with Perlmut-
ter’s racemic sequence [32% for (Ϯ)12β-13a, and 44% for
(Ϯ)12β-13b].[18]
Conclusions
We have reported a short route to enantiomerically pure
cycloadducts (+)12β- or (–)12α-13a–c, with (+)12β-13a be-
ing a key intermediate in Perlmutter’s racemic approach
towards 20-nor-salvinorin A. The stereoselective cycload-
dition of (SS)-Sulfinylquinone 2 with various semicyclic
dienes 7 leads to the controlled formation of four ste-
reogenic centers in one step. This paper describes the first
example of a kinetic resolution at the C(12) position.[29] The
introduction of a substituent (such as 3-Furyl) at this posi-
Scheme 5. Diels–Alder cycloaddition of sulfinylquinone 2 with
dienes 8a and 9a.
of 7a–c (matched pair), the reaction performed with (S)-7a– tion is usually stereochemically uncontrolled in diter-
c (mismatched pair) was slow (6–10 d), but lead smoothly to penoids synthesis, particularly in the clerodane series.[11,30]
the expected enantiopure cycloadducts 12β-12a–c
Accessing the dienone (+)12β-14a opens a route to salvi-
(Scheme 4 and Scheme 6).[27] In the first case, the cycload- norin A, because it represents a useful precursor for the
ducts 12α-12a,b [“unnatural C(12) epimers”] were isolated introduction of the C(20) quaternary center after the stereo-
at –20 °C in high yield (Ͼ85%) and subsequent desulfin- selective reduction of the C(7)–C(8) double bond. This
ylation using Raney-Ni at –20 °C or standing at room tem- strategy also opens the way to the synthesis of many ana-
perature lead to enantiomerically enriched compounds 12α- logues of salvinorin A for the evaluation of their KOR
13a,b in 49–62% and 12α-14a,b and 52% yields, respectively selectivity. Accordingly, possible variations in the size of the
(see Scheme 6).
C-cycle (seven- and five-membered ring) have been assessed
2822
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 2818–2826

Total Synthesis of Salvinorin A and Analogues
Scheme 6. Preparation of epimeric C(12) compounds (α or β) 12, 13 and 14. Reagents and conditions: (i) S-(+)-2 (1 equiv.), ZnBr2
(2 equiv.), CH2Cl2, –20 °C; (ii) Raney-Ni, MeOH, –20 °C; (iii) CH2Cl2, –20 °C, 5 d or pyridine*, room temp., overnight. * Conversion
1
determined from the H NMR spectra for 14a, which decomposed on silica gel.
stant in Hz, integration). 13C NMR spectra were also run at various
and the transformation of the resulting cycloadducts into
field strengths as indicated; spectra were recorded in CDCl3 using
salvinorin A analogues will be published in due course.
residual undeuterated solvent (δ = 77 ppm) as internal reference.
Infra red (IR) spectra were recorded with a diamond ATR spec-
trometer using neat simples. Infra red frequencies are reported in
Experimental Section
wavenumbers (cm–1), intensities were determined qualitatively and
are reported as strong (s), medium (m), or weak (w). IR measure-
ments were obtained with Universal ATR sampling accessories.
General Remarks: Reagents and solvents were purchased as reagent
grade and used without further purification. THF was distilled
from sodium benzophenone ketyl. Dichloromethane was distilled
from CaH2. ZnBr2 was flamed-dried in the reaction flask under
vacuum and under Argon before use. Flash column chromatog-
raphy (FC) was performed using silica gel 60 for preparative col-
umn chromatography (40–63 mm), unless specifically noted other-
wise. Demetalled silica gel was prepared according to a published
procedure.[17] Thin-layer chromatography (TLC) was performed on
glass sheets coated with silica gel 60 F254 (otherwise stated), visual-
ization by UV light or through staining with phosphomolybdic
acid, KMnO4 or vanillin. Optical rotations were measured with a
polarimeter with a sodium lamp and are reported as [α]2D0 (c is given
in g/100 mL, solvent). 1H and 13C NMR spectra were recorded
with a 200, 300, or 400 MHz spectrometer. Chemical shifts are re-
ported in ppm relative to the solvent (CDCl3) resonance δ =
7.26 ppm (s = singlet, d = doublet, t = triplet, q = quartet, m =
multiplet, app s = apparent singlet, m = multiplet, coupling con-
Melting points of solid compounds bearing a sulfoxide were not
recorded due to their high instability (even at room temp.).
The following known compounds were isolated as pure samples
and showed data matching those of reported compounds: (S)-2,[16]
3,[15] (Ϯ)-7b,[19a] 8a and 8b,[14] 9b.[31]
General Procedure for ZnBr2-Catalyzed Diels–Alder Reactions: To a
solution of freshly prepared sulfinylquinone (S)-2 (0.75–1.00 mmol,
1 equiv.) in anhydrous CH2Cl2 (10 mL), was added ZnBr2 (2 equiv.)
at –78 °C under argon, and the mixture was stirred for 10 min.
The appropriate 1,3-diene (2 equiv.) was added and the resulting
solution was stored in a fridge at –20 or –40 °C (dry ice/MeCN).
After completion of the reaction, a decoloration of the solution
from dark-orange to pale-yellow was observed and the mixture was
quenched with cold aqueous NH4Cl. The cold aqueous phase was
extracted with of cold (–20 °C) CH2Cl2 (3ϫ10 mL). The resulting
Eur. J. Org. Chem. 2011, 2818–2826
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2823

G. Hanquet et al.
FULL PAPER
organic solution was vacuum filtered through a short pad of silica
The relative stereochemistry was confirmed by NOESY experi-
gel (diameter: 6 cm, height: 3–4 cm) with: i. 150 mL of cold CH2Cl2 ments.
(–70 °C) to recover optically active diene (enantiopure or enantio-
(2S,6aS,10aR,10bS)-2-(Furan-3-yl)-9-methoxy-6a-methyl-10a-[(S)-
merically enriched if a racemic diene has been used); ii. 200 mL of a
1:1 CH2Cl2/Et2O mixture (–70 °C) to isolate the cycloadduct (Note:
some ice crystals may appear if the filtrate is left at –70 °C, which
may be filtered).
(p-tolylsulfinyl)]-6,6a,10a,10b-tetrahydro-1H-benzo[f]isochromene-
7,10(2H,4H)-dione (12β-12a): The reaction of quinone (S)-2 with
(S)-7a (2 equiv.) gave pure 12β-12a (340 mg, 72%). Reaction time:
6 d, temperature: –20 °C. Yellow, waxy oil. 1H NMR (300 MHz,
CD2Cl2, –20 °C): δ = 7.50–7.20 (m, 6 H, ArH), 5.94 (s, 1 H, Me-
O=CH), 5.39 (m, 1 H, =CH), 5.26 (m, 1 H, CHO), 4.30 (d, J =
14.2 Hz, 1 H, OCH2), 4.23 (d, J = 14.2 Hz, 1 H, OCH2), 3.53 (s, 3
H, OCH3), 3.16 (td, J = 12.8, 5.5 Hz, 1 H), 2.42 [s, 3 H, CH3
(pTol)], 2.37–2.34 (m, 3 H, CH2), 2.04 (d, J = 19.8 Hz, 1 H, CH2),
1.52 (s, 3 H, CH3) ppm. 13C NMR (100 MHz, CD2Cl2, –20 °C): δ
= 197.1 (C=O), 187.1 (C=O), 160.9 (Cquat), 142.9 (Cquat), 140.4
(Cquat), 135.0 (Cquat), 134.1 (Cquat), 129.7 (CH), 128.7 (CH), 127.4
(CH), 126.3 (2ϫ CHarom), 116.3 (CH), 110.0 (CH), 82.6 (Cquat),
73.8 (CH), 65.4 (CH2), 56.3 (OCH3), 52.1 (Cquat), 39.6 (CH2), 28.6
(CH2), 26.9 (CH2); 21.4 [CH3 (pTol)], 16.5 (CH3) ppm. IR (neat):
To avoid spontaneous elimination of the sulfoxide moiety from the
cycloadduct, the distillation of solvents must be performed at low
temperature (Ͻ0 °C). It is possible to obtain white crystals by tritu-
ration of the resulting solid in cold Et2O (–20 °C). When racemic
1,3-diene was used, the diastereomeric purification was performed
as follow: the crude diastereomeric mixture (350 mg) was dissolved
in cold (–20 °C) Et2O (50 mL) and placed in the freezer overnight.
The resulting crystals were filtered and the mother liquor was evap-
orated (same precaution as indicated above), leaving the pure major
diastereoisomer.
These compounds could be preserved for several months at –20 °C
ν = 3313 (br), 3142 (w), 1712 (w), 1662 (s), 1607 (s), 1504 (m), 1453
˜
(even for three years for 6).
(m), 1358 (m), 1237 (s), 1205 (s), 1080 (s), 1037 (br), 955 (m), 875
(m), 810 (m), 709 (m) cm–1.
(–)-(R)-3-Methoxy-10a-methyl-5,6,7,8,10,10a-hexahydro-phenanth-
rene-1,4-dione (4): To a solution of sulfinylquinone (S)-2 (290 mg,
1.00 mmol, 1.0 equiv.) in anhydrous CH2Cl2 (10 mL), under argon,
2-vinyl-1-cyclohexene (3; 216 mg, 2.00 mmol, 2.0 equiv.) was
added. After 10–12 d at room temp., the solvent was evaporated to
give the crude cycloadduct 6. Flash chromatography on demetall-
ated silica gel (gradient elution cyclohexane/CH2Cl2/acetone, 2:1:1)
afforded pure (S)-4 as a yellow oil (15%). The yield could be im-
proved by dissolving 6 in pyridine and allowing the sulfinyl group
to spontaneously eliminate at room temp. (194 mg, 75%). Rf = 0.43
(cyclohexane/CH2Cl2/acetone, 2:2:1); the absolute stereochemistry
was deduced from the known models and the X-ray analysis of
cycloadduct 6. 1H NMR (300 MHz, CDCl3): δ = 5.97 (s, 1 H, Me-
OC=CH), 5.92 (m, 1 H, C=CH), 3.85 (s, 3 H, OCH3), 2.95–2.30
(m, 8 H, 4 ϫ CH2), 2.10–1.81 (m, 2 H, CH2), 1.25 (s, 3 H,
CH3) ppm. 13C NMR (75 MHz, CDCl3): δ = 201.7 (C=O), 199.1
(C=O), 181.9 (Cquat), 164.0 (Cquat), 143.7(Cquat), 136.6 (Cquat),
135.0 (Cquat), 124.7 (MeOC=CH), 109.7 (C=CH), 56.6 (OCH3),
40.5 (CH2), 30.9 (CH2), 29.7 (CH2), 26.9 (CH2), 23.8 (CH3), 22.2
(CH2) ppm. [α]2D0 = –23.2 (c = 0.42, CH2Cl2). C16H18O3 (258.32):
calcd. C 74.40, H 7.02; found C 74.31, H 6.89.
(+)-(2S,6aR,10aR,10bR)-2-(Furan-3-yl)-9-methoxy-6a-methyl-
6,6a,10a,10b-tetrahydro-1H-benzo[f]isochromene-7,10(2H,4H)-di-
one (12β-13a): To a solution of compound 12β-12a (100 mg,
0.25 mmol) in MeOH (10 mL) at –20 °C, was added a wet slurry
of Raney-Ni (2800 purchased from Aldrich, ca. 10 mg), and the
resulting suspension was stored in a fridge at –20 °C for 12 h. The
mixture was then carefully filtered through Celite, and the solvent
was evaporated. Purification by flash chromatography (gradient
elution cyclohexane/Et2O, 1:1) afforded pure 12β-13a (61 mg, 61%)
as a pale-yellow amorphous powder. Rf = 0.5 (Et2O). 1H NMR
(400 MHz, CDCl3): δ = 7.37 (m, 1 H), 7.35 (t, J = 1.8 Hz, 1 H,
ArH), 6.37 (m, 1 H, ArH), 5.81 (s, 1 H, MeOC=CH), 5.62 (m, 1
H, =CH), 4.63 (dd, J = 11.3, 1.6 Hz, 1 H, OCH), 4.25 (d, J =
12.5 Hz, 1 H, OCH2), 4.16 (dm, J = 12.4 Hz, 1 H, OCH2), 3.79 (s,
3 H, OCH3), 2.95 (m, 1 H), 2.78 (d, J = 9.8 Hz, 1 H, OCCH), 2.57
(ddd, J = 12.7, 4.6, 2.0 Hz, 1 H, CH2), 2.51 (d, J = 20.0 Hz, 1 H,
CH2), 2.28 (ddt, J = 18.5, 5.7, 2.0 Hz, 1 H, CH2), 1.30 (dt, J = 12.7,
11.4 Hz, 1 H, CH2), 1.14 (s, 3 H, CH3) ppm. 13C NMR (100 MHz,
CDCl3): δ = 202.4 (C=O), 194.7 (C=O), 163.1 (Cquat), 143.2 (Cquat),
139.1 (Cquat), 132.8 (Cquat), 126.7 (=CH), 119.3 (MeOC=CH),
109.1 (CHarom), 108.7 (CH), 72.4 (COH), 72.3 (COH2), 56.5
(OCH3), 56.4 (=CCH), 48.2 (Cquat), 39.8 (CH2), 33.1 (CH2), 21.1
(–)-(4aS,4bR,10aR)-3-Methoxy-10a-methyl-5,6,7,8,10,10a-hexahydro-
phenanthrene-1,4(4aH,4bH)-dione (10): To a solution of 6 (100 mg,
0.25 mmol) in MeOH (10 mL) at –20 °C, was added a wet slurry
of Raney-Ni (2800 purchased from Aldrich, ca. 10 mg), and the
resulting suspension was stored in a fridge at –20 °C for 12 h. The
mixture was then carefully filtered through Celite, and the solvent
was evaporated. Purification by flash chromatography (gradient
(CH3) ppm. [α]2D0 = +67 (c = 1.15, CH Cl ). IR (neat): ν = 2962
˜
2
2
(m), 2844 (m), 1711 (m), 1661 (s), 1605 (s) cm–1.
The spectroscopic data are identical to those previously reported
for the racemic form.[18]
(+)-(2S,6aR)-2-(Furan-3-yl)-9-methoxy-6a-methyl-6,6a-dihydro-1H-
benzo[f]isochromene-7,10(2H,4H)-dione (12β-14a): A syn-desulfin-
elution cyclohexane/Et2O, 1:1) afforded pure (S)-10 as a colorless
1
foam (43 mg, 70%). Rf = 0.5 (Et2O). H NMR (400 MHz, C6D6): ation was observed when the cycloadduct 12β-12a was stored in
δ = 5.42 (s, 1 H, MeOC=CH), 5.14 (m, 1 H, C=CH), 2.55 (d, J =
[D5]pyridine at 4 °C after 2 d. Purification of the crude material by
8.2 Hz, 1 H, COCH), 2.47 (ddd, J = 12, 5, 4 Hz, 1 H, =CCH), flash chromatography on silica gel (gradient elution cyclohexane to
2.35–2.21 (m, 3 H, CH2), 2.00 (m, 1 H, CH2), 1.78 (m, 1 H, CH2), cyclohexane/Et2O, 1:1) afforded pure 12β-14b (31 mg, 60%) as a
1.60–1.54 (m, 2 H, CH2), 1.35 (m, 1 H, CH2), 1.13 (m, 1 H, CH2), white amorphous solid. Rf = 0.4 (cyclohexane/Et2O, 1:1). 1H NMR
0.98 (s, 3 H, CH3), 0.62 (dddd, J = 28, 12, 8, 1.5 Hz, 1 H) ppm.
13C NMR (100 MHz, C6D6): δ = 201.8 (C=O), 194.1 (C=O), 163.3
(MeO-Cquat), 138.3 (HC=Cquat), 116.3 (=CH), 108.7 (=CH), 59.7
(300 MHz, [D5]pyridine): δ = 7.65–7.54 (m, 2 H, ArH), 7.19 (m, 1
H, ArH), 5.66 (br. s, 1 H, MeOC=CH), 5.22 (app br. s, 1 H, =CH),
4.46 (dd, J = 15.3, 4.3 Hz, 1 H, COH), 4.25–4.19 (m, 2 H, COH2),
(COCH), 55.3 (OCH3), 48.4 (Cquat), 35.6 (CH2), 34.9 (CH), 34.4 3.41 (s, 3 H, OCH3), 2.59–2.48 (m, 2 H, CH2), 2.33 (m, 1 H, CH2),
(CH2), 33.7 (CH2), 28.2 (CH2), 26.3 (CH2), 20.5 (CH3) ppm. [α]2D0
= –1 (c = 1.15, CH2Cl2). HRMS (EI): calcd. for C16H20O3Na [M]
283.130; found 283.130. C16H20O3 (260.33): calcd. C 73.82, H
230–2.25 (m, 1 H, CH2), 1.22 (s, 3 H, CH3) ppm. 13C NMR
(75 MHz, [D5]pyridine): δ = 202.2 (C=O), 194.3 (C=O), 167.3
(Cquat), 142.3 (Cquat), 133.1 (Cquat), 128.7 (CHarom), 126.5
(CHarom), 118.4 (=CH), 108.9 (=CH), 79.4 (OCH), 72.8 (OCH2),
+·
7.74; found C 73.91, H 7.88.
2824
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 2818–2826

Total Synthesis of Salvinorin A and Analogues
56.4 (OCH3), 48.2 (Cquat), 41.0 (CH2), 33.1 (CH2), 18.0 (CH3) ppm. CH Cl ). IR (neat): ν = 3294, 3077, 2907 (s), 2857, 1642, 1502,
˜
2
2
[α]2D0 = +51 (c = 1, CHCl3). C19H18O5 (326.35): calcd. C 69.93, H
1159, 1072, 1020, 917 cm–1. MS (EI): m/z (%) = 176 (7) [M], 135
5.56; found C 69.99, H 5.43.
(100), 95 (41), 91 (17), 77 (43).
General Procedure for Enyne Metathesis Using Grubbs’ 1st Genera-
tion Catalyst: To a stirring solution of Grubbs’ 1st generation cata-
lyst (0.1–0.2 mmol, 2 mol-%) in distilled and degassed (argon,
10 min and ethylene, 5 min) CH2Cl2 (10–20 mL), was added enyne
(1–2 g, 5.8–12.5 mmol) dissolved in CH2Cl2 (3 mL). The solution
was purged with ethylene gas for 10 min and stirred under an ethyl-
ene atmosphere at 100 °C (2–4 atm) for 10 min under microwave
irradiation (20 mL sealed vial) or overnight at room temperature.
The catalyst was quenched with the addition of ethyl vinyl ether
(2 mL), and the solvent was removed in vacuo. The resulting brown
oil was purified by flash chromatography (EtOAc/hexanes, 1:30) to
afford the semicyclic 1,3-diene as a colorless oil. Most of the time,
the dienes proved to be unstable to air/water upon prolonged stor-
age and were, therefore, refrigerated (–15 °C) under an inert atmo-
sphere, which allowed storage for several months.
General Method for Lipase-Catalyzed Hydrolysis: The substrate
(40 mmol) was dissolved in a mixture of pH 7.0 phosphate buffer
(200 mL) and DMSO (20 mL) containing 5% (w/v) silica gel, at
40 °C. The enzyme (Pseudomonas cepacia) was added, and the pH
was maintained at 7.0 by titration with 1 n aqueous NaOH. The
progress of the reaction was followed by gas chromatographic
analysis of aliquots taken from the reaction mixture at intervals.
When the reaction was complete or the desired conversion was
reached, the reaction mixture was filtered through silica gel and
extracted three times with diethyl ether. The combined organic ex-
tracts were washed with saturated aqueous NaHCO3 and brine,
dried with MgSO4 and concentrated in vacuo. The acetate and
alcohol were separated by flash chromatography over silica gel
(pentane/diethyl ether, 1:1).
General Method for Lipase-Catalyzed Transesterification:[32] THF
(0.1 m), vinylacetate (10 equiv.) and substrate (90 mmol, 1 equiv.)
were placed in a 500 mL screw-cap bottle and stirred gently at
35 °C. The enzyme (Pseudomonas cepacia) was added (1 equiv. w/
w), and the reaction vessel was closed. The progress of the reaction
was followed by gas chromatographic analysis of samples taken
from the reaction mixture at intervals. When the reaction was com-
plete or the desired conversion was reached, the reaction mixture
was filtered through silica gel and the solvent was removed in
vacuo. The acetate and alcohol were separated by flash chromatog-
raphy over silica gel (cyclohexane/diethyl ether, 1:1).
(–)-(S)-2-(Furan-3-yl)-5-vinyl-3,6-dihydro-2H-pyran [(S)-7a]: Fol-
lowing the general procedure, the corresponding diene (S)-7a
(163 mg, 93%) was obtained as a yellow oil. The absolute stereo-
chemistry was determined from the sign of the optical rotation,
which was opposite to that reported in the literature.[12] 1H NMR
(300 MHz, CDCl3): δ = 7.43 (m, 1 H, ArH), 7.40 (t, J = 1.6 Hz, 1
H, ArH), 6.44 (s, 1 H, ArH), 6.29 (dd, J = 11.2, 10.8 Hz, 1 H,
CH=CH2), 5.90 (m, 1 H, =CH), 4.99 (d, J = 4.0 Hz, 1 H,
CH=CH2), 4.95 (d, J = 10.8 Hz, 1 H, CH=CH2), 4.54 (dd, J =
10.0, 4.0 Hz, 1 H, CHO), 4.43 (m, 2 H, CH2O), 2.48 (m, 1 H, CH2),
2.37 (d, J = 18.4 Hz, 1 H, CH2). [α]2D0 = –163 (c = 1.16, CHCl3)
{lit.[19b] [α]2D0 = +167.9 (c = 1.16, CHCl3)}.
(–)-(S)-1-(Furan-3-yl)but-3-en-1-ol (a): Following the general pro-
cedure described above, compound a was obtained as a clear, color-
less oil (2.88 g, 46%). The S absolute stereochemistry was deter-
mined by comparison of the sign of optical rotation with reported
Supporting Information (see footnote on the first page of this arti-
cle): Experimental procedure for the preparation of dienes for (Ϯ)-
9a,b; (Ϯ)-7d enantio-enriched 7a–c; full analytical data for com-
pounds 4, 6, 10, 11, 12a–c, 12α-13a,b, 12β-13a–c, 12α-14a,b, and
12α-14a–c; 1H and 13C NMR spectra for 12d and X-ray crystal
structure of compounds 6, 11, 12α-12a, and 12β-12b.
literature values.[7] [α]2D0 = –30 (c = 1.72, CH2Cl2) {ref.[32] [α]2D0
–30.7 (c = 1.72, CH2Cl2)}; Ͼ95% e.e.
=
(–)-(S)-3-[1-(Prop-2-ynyloxy)but-3-enyl]furan (d): To a flask charged
with NaH (60% dispersion in mineral oil; 0.984 g, 24.6 mmol) in
anhydrous toluene (30 mL) under an atmosphere of argon, was
added
a solution of (S)-1-(furan-3-yl)but-3-en-1-ol (a; 2 g,
Acknowledgments
14.5 mmol) in toluene (10 mL) at 0 °C under vigorous stirring. The
mixture was warmed to room temperature and stirred for 1.5 h
before being cooled to 0 °C and 15-crown-5 (1.4 mL, 7.2 mmol)
was added slowly. The reaction mixture was warmed to room tem-
perature and stirred for 0.5 h. After cooling again to 0 °C, propar-
gyl bromide (80 wt.-% in toluene, 7.5 mL, 50.6 mmol) was added
dropwise and the reaction was stirred for 2 h at room temperature.
The reaction was quenched at 0 °C with ice-water (30 mL) and di-
luted with Et2O (30 mL). The aqueous layer was extracted with
Et2O (2ϫ30 mL) and the combined organic layers were washed
with saturated aqueous Na2S2O3·5H2O (20 mL), water (20 mL),
and brine (20 mL), dried with MgSO4, filtered, and concentrated
in vacuo. The resulting brown oil was purified by flash chromatog-
raphy (toluene/cyclohexane, 5:95) to afford d (2.18 g, 85%) as a
colorless oil. 1H NMR (300 MHz, CDCl3): δ = 7.40 (m, 2 H, ArH),
6.39 (t, J = 1.1 Hz, 1 H, ArH), 5.77 (ddt, J = 17.2, 10.2, 6.9 Hz,
CH=CH2), 5.08 (dm, J = 17.2 Hz, 1 H, CH=CH2), 5.04 (dm, J =
10.2 Hz, 1 H, CH=CH2), 4.56 (t, J = 6.8 Hz, 1 H, CHO), 4.13 (dd,
J = 15.8, 2.4 Hz, 1 H, CH2-Cϵ), 3.93 (dd, J = 15.8, 2.4 Hz, 1 H,
CH2-Cϵ), 2.57–2.65 (m, 1 H, CH2), 2.41–2.49 (m, 1 H, CH2), 2.40
(t, J = 2.4 Hz, 1 H, CϵCH) ppm. 13C NMR (75 MHz, CDCl3): δ
= 143.6 (CHarom.), 140.8 (CHarom.), 134.3 (C=CH2), 124.6 (CArom),
117.3 (C=CH2), 108.7 (CHarom.), 79.9 (CHO), 74.2 (CϵCH), 72.1
(CϵCH), 55.2 (CH2), 40.6 (CH2) ppm. [α]2D0 = –107 (c = 1.45,
We thank Prof. M. Carmen Carreño (UAM Madrid) for useful
discussions, the Collectivité Territoriale de Corse (DAL) and the
Centre National de la Recherche Scientifique (CNRS) for financial
support and Mr. Bastien Boff for the preparation of dienes 9a and
9b.
[1] B. L. Roth, K. Baner, R. Westkaemper, D. Siebert, K. C. Rice,
S. Steinberg, P. Ernsberger, R. B. Rothman, Proc. Natl. Acad.
Sci. USA 2002, 99, 11934–11939.
[2] T. E. Prisinzano, R. B. Rothman, Chem. Rev. 2007, 108, 1732–
1743.
[3] a) T. A. Vorthems, B. L. Roth, Mol. Int. 2006, 6, 257–265; b)
A three-dimentional pharmacophore of κ-opioid receptor ago-
nists has been recently identified, see: N. Yamaotsu, H. Fujii,
H. Nagase, S. Hirono, Bioorg. Med. Chem. Lett. 2010, 20,
4446–4452.
[4] M. Waldhoer, S. E. Bartlett, J. L. Whistler, Annu. Rev. Biochem.
2004, 73, 953–990.
[5] R. B. Rothman, D. L. Murphy, H. Xu, J. A. Godin, C. M.
Dersch, J. S. Partilla, K. Tidgewell, M. Schmidt, T. E. Prisin-
zano, J. Pharmacol. Exp. Ther. 2007, 320, 801–810.
[6] Y. W. D. Lee, L. Yang, W. Xu, G. Deng, L. Guo, L. Y. Liu-
Chen, Bioorg. Med. Chem. Lett. 2010, 20, 5749–5752, and ref-
erences cited therein. In this article, we will use the neo-clerod-
Eur. J. Org. Chem. 2011, 2818–2826
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2825

G. Hanquet et al.
A. Kinoshita, M. Mori, Synlett 1994, 1020–1022.
A. L. Gemal, J. L. Luche, J. Am. Chem. Soc. 1981, 103, 5454–
5459.
FULL PAPER
ane nomenclature to define the atom numbering of the pre-
pared compounds.
[7] C. Béguin, K. K. Duncan, T. A. Munro, D. M. Ho, W. Xu,
L. Y. L. Chen, W. A. Carlezon, B. M. Cohen, Bioorg. Med.
Chem. 2009, 17, 1370–1380.
[8] a) F. Yan, P. D. Mosier, R. B. Westkaemper, J. Stewart, J. K.
Zjawiony, T. A. Vorthems, D. J. Scheffler, B. L. Roth, Biochem-
istry 2005, 44, 8643–8651; b) B. E. Kane, C. R. McCurdy,
D. M. Ferguson, J. Med. Chem. 2008, 51, 1824–1830; B. E.
Kane, M. J. Nieto, C. R. McCurdy, D. M. Ferguson, FEBS J.
2006, 273, 1966–1974.
[22]
[23]
[24]
[25]
C. M. Carreño, A. Urbano, C. Di Vitta, J. Org. Chem. 1998,
63, 8320–8330.
Reaction time required for quantitative conversion of the start-
ing (+)-sulfinylquinone 2. The corresponding cycloadducts
were not isolated and the diastereomeric ratio was determined
from the 1H NMR spectra of the crude material after rapid
filtration through silica gel with cold CH2Cl2 (–40 °C) to elim-
inate excess diene.
C. Held, R. Fröhlich, P. Metz, Adv. Synth. Catal. 2002, 344,
720–727.
It should be noted that excess of enantiopure dienes 7 (i.e.,
1 equiv.) was readily separated from the crude mixture and
could therefore be used in further batches.
[9] D. A. Lanfranchi, Ph. D. Thesis: Vers la synthèse totale de la
Salvinorine A, Université de Strasbourg, 2006.
[26]
[27]
[10] a) A. Lozama, T. E. Prisinzano, Bioorg. Med. Chem. Lett. 2009,
19, 5490–5495; b) J. R. Scheerer, J. F. Lawrence, C. Grace, D.
Wang, A. Evans, J. Am. Chem. Soc. 2007, 129, 8968–8969.
[11] a) M. Nozawa, Y. Suka, T. Hoshi, T. Suzuki, H. Hagiwara,
Org. Lett. 2008, 10, 1365–1368; b) H. Hagiwara, Y. Suka, T.
Nojima, T. Hoshi, T. Suzuki, Tetrahedron 2009, 65, 4820–4825.
[12] D. A. Lanfranchi, N. Baldovini, G. Hanquet, Synthesis 2008,
3775–3778.
[28] Analytical data for compounds 13a and 13b match those de-
scribed by Perlmutter and co-workers, see ref.[18]
[29] a) For a review on the C(12) position of neo-cleorodane, see: J.
Jiménez-Barbero, Tetrahedron 1993, 49, 6921–6930. The C(12)
position has been very recently controlled through diastereo-
selective reduction using K-selectride before C-ring closure, see
ref.[11b]
[30] Examples; for clerocidin, see: a) A. X. Xiang, D. A. Watson, T.
Ling, E. A. Theodorakis, J. Org. Chem. 1998, 63, 6774–6775;
b) for tanabalin, see: H. Watanabe, T. Onoda, T. Kitahara, Tet-
rahedron Lett. 1999, 40, 2545–2548; c) 15-for desoxoclerocidin,
see: A. S. Kende, J. J. Rustenhoven, K. Zimmermann, Tetrahe-
dron Lett. 2000, 41, 843–846; d) for methyl barbascoate, see:
H. Hagiwara, K. Hamano, M. Nozawa, T. Hoshi, T. Suzuki,
F. Kido, J. Org. Chem. 2005, 70, 2250–2255; e) for a review on
the total synthesis of clerodanes, see: T. Tokoroyama, Synthesis
2000, 611–633.
[13] L. M. Kutrzeba, D. Ferreira, J. K. Zjawiony, J. Nat. Prod. 2009,
72, 1361–1363.
[14] D. A. Lanfranchi, C. Bour, B. Boff, G. Hanquet, Eur. J. Org.
Chem. 2010, 5232–5247.
[15] W. Herz, R. R. Juo, J. Org. Chem. 1985, 50, 618–627.
[16] D. A. Lanfranchi, G. Hanquet, J. Org. Chem. 2006, 71, 4854–
4861.
[17] J. S. Hubbard, T. M. Harris, J. Org. Chem. 1981, 46, 2566–2570.
[18] Y. E. Bergman, R. Mulder, P. Perlmutter, J. Org. Chem. 2009,
74, 2589–2591.
[19] a) For (Ϯ)-7b, see: M. Mori, N. Sakakibara, J. Kinoshista, Org.
Chem. 1998, 63, 6082–6083; b) For (+)-7a, see: A. K. Cheung,
M. Murelli, M. L. Snapper, J. Org. Chem. 2004, 69, 5712–5719;
c) For (Ϯ)-8a,b, see ref.[14]
[20] L. Yang, W. Xu, F. Chen, L. Y. Liu-Chen, Z. Ma, D. Y. W. Lee,
Bioorg. Med. Chem. Lett. 2009, 19, 1301–1304.
[21] a) S. H. Snyder, L. A. Faillace, H. Weingartner, Am. J. Psychia-
try 1968, 125, 357–364; b) E. Sanders-Bush, K. D. Burris, K.
Knoth, J. Pharmocol. Exp. Ther. 1988, 246, 924–928.
[31] Y.-J. Lee, R. R. Schrock, A. H. Hoveyda, J. Am. Chem. Soc.
2009, 131, 10652–10661.
[32] A. Bierstedt, J. Stölting, R. Fröhlich, P. Metz, Tetrahedron:
Asymmetry 2002, 12, 3399–3407.
Received: February 15, 2011
Published Online: April 7, 2011
2826
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 2818–2826