Structure-activity relationships of the estrogenic sesquiterpene ester ferutinin. Modification of the terpenoid core

Article abstract of DOI:10.1021/np049796w

Esterification of p-hydroxybenzoic acid, a very weak estrogenic compound, with the daucane alcohol jaeschkeanadiol (1b) leads to a spectacular magnification of the estrogenic activity. To identify the structural elements responsible for this effect, the terpenoid core of jaeschkeanadiol p-hydroxybenzoate (ferutinin, 1a) was modified, capitalizing on the presence of two functionalities, the monoacylated, hydrogen-bonded 1,3-diol system and the double bond. The hydrogen bonding, while possibly useful, was not critical for activity, while hydrogenation and cyclopropanation of the double bond were tolerated. Conversely, oxidative modifications of the double bond that placed a hydroxyl on the α-face of the molecule proved detrimental. Taken together, these observations identified the substitution at C-8/C-9 as critical for activity.

Full text of DOI:10.1021/np049796w

J . Nat. Prod. 2004, 67, 1557-1564
1557
Str u ctu r e-Activity Rela tion sh ip s of th e Estr ogen ic Sesqu iter p en e Ester
F er u tin in . Mod ifica tion of th e Ter p en oid Cor e
Giovanni Appendino,*,† Paola Spagliardi,†,‡ Olov Sterner,§ and Stuart Milligan
Dipartimento di Scienze Chimiche, Alimentari, Farmaceutiche e Farmacologiche, Universita` del Piemonte Orientale,
Via Bovio 6, 28100 Novara, Italy, Dipartimento di Scienza e Tecnologia del Farmaco, Universita` di Torino, Via Giuria 9,
10125 Torino, Italy, Department of Organic and Bioorganic Chemistry, Lund University, P.O. Box 124, 221 00 Lund, Sweden,
and Endocrinology and Reproductive Research Group, Kings College, SE1 1 UL, London, U.K.
Received J une 18, 2004
Esterification of p-hydroxybenzoic acid, a very weak estrogenic compound, with the daucane alcohol
jaeschkeanadiol (1b) leads to a spectacular magnification of the estrogenic activity. To identify the
structural elements responsible for this effect, the terpenoid core of jaeschkeanadiol p-hydroxybenzoate
(ferutinin, 1a ) was modified, capitalizing on the presence of two functionalities, the monoacylated,
hydrogen-bonded 1,3-diol system and the double bond. The hydrogen bonding, while possibly useful, was
not critical for activity, while hydrogenation and cyclopropanation of the double bond were tolerated.
Conversely, oxidative modifications of the double bond that placed a hydroxyl on the R-face of the molecule
proved detrimental. Taken together, these observations identified the substitution at C-8/C-9 as critical
for activity.
Estrogen receptors (ERs) are relatively indiscriminate
in their binding properties, to the point that “no single
chemical has yet been internationally agreed to be devoid
of estrogenicity”,1 but very few potent nonsteroid scaffolds
are known. Most of them are relatively flat molecules that
retain a certain degree of topological similarity with
estradiol (E2) and the capacity to position critical phenolic
hydroxyls into the hydrogen-bonding elements of ERs (the
diad Arg 291/Glu 324 and His 524 for the ER-R).2 The
discovery of potent estrogenic activity in the sesquiterpene
ester ferutinin (1a )3 and the occurrence of large amounts
of this compound in a Sardinian chemotype of giant fennel
Resu lts a n d Discu ssion
(Ferula communis L.)4 gave us the opportunity to investi-
Intramolecular hydrogen bonding is possible between the
tertiary hydroxyl and the ester carbonyl of jaeschkeanadiol
esters.11 This adds to the tendency of secondary esters to
maximize σfπ* conjugation by assuming a Z-conformation
around the ester C-O single bond,12 overall translating
into a geometry where the carbonyl oxygen is oriented
toward the â-face of the daucane skeleton and the aromatic
ring bisects its mean plane. The perturbation of this
conformational constraint seemed a good starting point to
explore the magnifying effect of the daucane moiety on the
hormonal activity of p-hydroxybenzoic acid. To this end,
the replacement of the ester link with an ether bond, the
epimerization of the secondary 6-hydroxyl, and the dehy-
dration of the 4-hydroxyl were pursued.
The ether analogue 1c was prepared by etherification
of jaeschkeanadiol (1b) with p-pivaloyloxybenzyl bromide,
in turn prepared from p-hydroxybenzaldehyde by reduc-
tion, Schotten-Baumann pivaloylation, and NBS-mediated
bromine-to-oxygen exchange. The etherification reaction
occurred in modest yield, delivering, after deprotection, 1c
as a foam. The secondary hydroxyl of jaeschkeanadiol
proved too hindered to act as a substrate for a Mitsunobu
reaction. To invert the configuration at C-6, we therefore
investigated an alternative strategy, based on an oxida-
tion-reduction protocol. The underlying rationale is that
the angular â-methyl at C-1 shields the upper face of the
daucane skeleton, directing attack from the opposite R-face.
In the event, the oxidation of jaeschkeanadiol under a
variety of conditions (PCC, PDC, TPAP, activated DMSO,
gate the structure-activity relationships of this novel
estrogenic prototype. Ferutinin is also the alleged active
principle of various preparations available in the health
food market as libido enhancers. These products have
recently featured prominently in the media and the popular
press, despite concerns about their safety.5
Ferutinin is a p-hydroxybenzoyl ester, and in a previous
study, the aromatic moiety was found essential for ER
binding.6 Esters of p-hydroxybenzoic acid are widely used
as preservatives in pharmaceutical formulations,7 while the
free acid is produced in the metabolic degradation of dietary
flavonoids such as apigenin and quercetin8 and is a
common component of human urine.9 J aeschkeanadiol (1b),
the parent diol of ferutinin, lacks any recognizable hor-
monal activity,6 while p-hydroxybenzoic acid and its simple
alkyl esters show only marginal estrogenic activity.10
Therefore, the daucane core contributes critically to the
activity of ferutinin, and this observation provided a
rationale for undertaking a study aimed at identifying
which functionalities of the terpenoid core are required,
and presumably involved, in ER binding.
* To whom correspondence should be addressed. Tel: +39 0321 375744
(G.A.); +34 957 218267+44 20 7848 6271 (S.M.). Fax: +39 0321 375821
(G.A.); +44 7848 6280 (S.M.). E-mail: Giovanni.Appendino@pharm.unipmn.it
(G.A.); Stuart.Milligan@kcl.ac.uk (S.M.).
† Universita` del Piemonte Orientale.
‡ Universita` di Torino.
§ Lund University.
Kings College.
10.1021/np049796w CCC: $27.50
© 2004 American Chemical Society and American Society of Pharmacognosy
Published on Web 08/25/2004

1558 J ournal of Natural Products, 2004, Vol. 67, No. 9
Appendino et al.
Sch em e 1. Synthesis of 6-Epiferutinin-R-epoxide (8, A) and Dehydroferutinin (9, B)
TEMPO) failed to afford the ketone 2, rather giving a
complex reaction mixture, from which, with Cr6+-based
reagents, a mixture of the enone 3 and its dehydration
product 4 could be characterized. Formation of a conjugated
enone is a foreseeable problem in the oxidation of homo-
allylic alcohols, but overoxidation to an ene-dione, presum-
ably via a concerted cyclic mechanism (Scheme 1, A), was
surprising. To overcome this problem, we “protected” the
double bond as an epoxide. Treatment of jaeschkeanadiol
with MCPBA occurred with remarkable diastereoselec-
tivity, affording a separable 30:1 mixture of epoxides (5a ,b).
The oxidation of the major R-epoxide (NOE correlation
between the C-1 and the C-8 methyls) was next screened
with the same set of oxidants assayed for 1b. Transition
metal oxidants (Cr6+, TPAP) afforded the expected epoxy-
ketone 6, but a certain degree of overoxidation to 4, pre-
sumably through a variant of the mechanism proposed for
its formation from 1b, could not be avoided. On the other
hand, no reaction occurred with milder oxidizers, like the
couple TEMPO/trichloroisocyanuric acid.13 The epoxy-
ketone 6 was then treated with NaBH4 to afford the 6â-
alcohol 7 as the only reaction product. Esterification of 7
with p-pivaloyloxybenzoic acid required harsh conditions
and gave a low yield (20%), but, after deprotection, 6-epi-
ferutinin epoxide (8) could eventually be obtained. Since
attempts to deoxygenate 8 failed (see infra), the effect of
C-6 epimerization could not be assessed by point mutation,
but only in an indirect way, that is, by comparing the
activity of the corresponding epoxides (8 and ferutinin
epoxide (12), available as the only reaction product from
the epoxidation of ferutinin). These unforeseen difficulties
were somewhat compensated by the observation that
ferutinin could be dehydrated in a highly regioselective
way. Thus, treatment with thionyl chloride afforded a ca.
3:1 mixture of ∆3,4- and ∆4,5-isomers (9 and 10, respec-
tively), which equilibrated to the pure ∆3,4-olefin 9 during
chromatography on silica gel or by standing in CDCl3
(Scheme 1, B). The ∆4,11-isomer could not be detected, either
in the reaction mixture or after its acidic equilibration.
The endocyclic double bond of ferutinin served as a basis
for a second set of analogues, where C-8 and C-9 are
modified in an oxidative or reductive way (Scheme 2A, B).
The minor epoxide from the epoxidation of jaeschkeanadiol
(5b ) was converted to its p-hydroxybenzoate (11) in the
usual way, complementing the stereoselectivity observed
during the epoxidation of ferutinin to 12. In a model
reaction for the synthesis of 6-epiferutinin from its epoxide
8, protected (pivalate) ferutinin epoxide was subjected to
various deoxygenation protocols. The couple zinc/titanocene
dichloride,14 while failing to materialize the expected olefin,
nevertheless led to a smooth isomerization to the allylic
alcohol 13, providing access to an analogue with an
exocyclic double bond. Dihydroxylation of ferutinin by the
Upjohn protocol afforded the R-glycol 14 as the only
reaction product, further oxidized with PCC to the ketol
15. Examination of the crude reaction mixture indicated
the formation of an aldehyde (singlet at δ 9.8, ca. 25% of
ketol 15). This compound proved too unstable on silica gel
for direct isolation, but could be characterized after reduc-
tion of the crude reaction mixture with NaBH4. The major
reaction product was the trans-diol 16, accompanied by the
pyrone 17, which could not be completely purified from
other minor reaction products even by HPLC. The structure
of 17 was unambiguously assigned by 1D and 2D NMR
measurements. A possible mechanism for the formation of
this compound is depicted in Scheme 3. Thus, overoxidation
of the ketol system could afford a 8,9-seco-daucane, featur-
ing two carbonyl-bearing groups bound to the cyclopentane
ring. The 8-ketone is located on the â-side of the molecule,
and formation of an acetal with the tertiary 4-hydroxyl
could take place, next followed by the hydrolytic loss of the
aromatic ester group and the oxidation of the 6-hydroxyl.
This could trigger the elimination of the semiacetal hy-
droxyl and the overall formation of a cyclopenta[b]pyrone
aldehyde, whose chemoselective reduction would afford the
primary alcohol 17. The characterization of a 8,9-cleaved
jaeschkeanadiol ester is interesting, since the cleavage of

Estrogenic Sesquiterpene Ester Ferutinin
J ournal of Natural Products, 2004, Vol. 67, No. 9 1559
Sch em e 2. Synthesis of Ferutinin-â-epoxide (11, A) and Oxidative Modification of the Double Bond of Ferutinin (B)
Sch em e 3. Possible Mechanism for the Oxidative
Fragmentation of the Daucane Core of 15 and the
Formation of 17
To complete the oxidative modifications, the double bond
of ferutinin acetate (1e) was allylically oxidized in a
regioselective way, affording a mixture of a 14-hydroxy and
a 14-oxo derivative, next converted by deprotection to the
analogues 18a and 18b, respectively. As for the nonoxida-
tive modifications, the double bond of ferutinin pivalate
(1d ) was cyclopropanated with a modified Simmons-Smith
protocol16 to a single cyclopropyl homologue, next depro-
tected to 19, while catalytic hydrogenation of ferutinin gave
an almost equimolecular mixture of epimeric dihydro
derivatives (20). The lack of stereoselectivity in the cata-
lytic hydrogenation is remarkable, since all the other
additions to the double bond of ferutinin occurred with
remarkable stereoselectivity.
the 8,9-bond of jaeschkeanadiol affords compounds of
relevance for the synthesis of prostanoids,15 but has been
reported to give complex mixtures with a variety of
reagents.15 Taken together, these observations suggest that
the 8,9-double bond of jaeschkeanadiol esters can indeed
be oxidatively cleaved, but that the cleavage products are
unstable and undergo further reactions, one of which is the
one leading to bicyclic pyrones.
The estrogenic activity of ferutinin (1a ) and its analogues
was investigated in a yeast screen containing the human
estrogen receptor alpha (ERa). This screen provides a rapid
evaluation of estrogenic agonist activity,17 and its results
are shown in Figure 1. Ferutinin (1a ) and its dihydro
derivatives (20) showed high estrogenic potency, ca. 2
orders of magnitude lower than estradiol. Significant

1560 J ournal of Natural Products, 2004, Vol. 67, No. 9
Appendino et al.
F igu r e 1. Relative estrogenic activity of estradiol, ferutinin (1a ), and its analogues 1c, 8, 9, 11, 12, 13, 14-16, 18a , 18b, 19, and 20. Estrogenic
activity was assessed in a yeast screen bearing the human estrogen receptor (ERR). Results are expressed as a percent of maximal stimulation
achieved with estradiol.
estrogenic activity was also detected in compounds 1c, 9,
11, 12, 18b, and 19, although 9, 12, and 19 showed clear
signs of cytotoxicity at the highest concentrations assayed.
Conversely, no significant activity was detected in all the
remaining compounds. Ferutinin and its analogues were
also assayed for antiestrogenic, androgenic, and antian-
drogenic activity, but no activity could be evidenced.
The estrogenic activity of the p-hydroxybenzyl analogue
1c and of the dehydro derivative 9 is somewhat unexpected
and shows that the conformational constraint of the acyl
and the sesquiterpene moieties via hydrogen bonding, while
possibly useful, is not critical for the estrogenic activity of
ferutinin. The modification of the double bond provided
several interesting insights. Thus, both diastereomeric
epoxides (11 and 12) maintained a significant share of the
activity of the natural product. Conversely, all oxidative
modifications that replaced the C-8/C-9 double bond with
oxygenated functions (glycol, ketol) having a hydroxyl on
the R-face of the molecule gave inactive analogues (com-
pounds 13-16). Taken together, these data suggest that
the R-face of the eastern part of ferutinin interacts with
the lipophilic domain of ER and that this critical interaction
is disrupted by the presence of a hydroxyl group. The
substantial retention of activity observed upon cyclopro-
panation and hydrogenation of the double bond supports
this view, since the lipophilic pocket of the ERs can bind a
wide variety of lipophilic substrates1 and should easily
accommodate a protruding methylene (as in 19) or methyl
(as in 20). The effect of epimerization at C-6 could not be
assessed by a point-mutation, since the synthesis of
6-epiferutinin proved elusive. Nevertheless, the marked
decrease of activity between R-epoxyferutinin (12) and its
6-epimer (8) suggests that the configuration at C-6 is
critical. The effect of the oxidation of the allylic methyl was
also interesting, since the introduction of a hydroxyl was
detrimental for activity, while methyl to formyl oxidation
was somewhat tolerated. A possible rationalization is that,
while rotation around the C-8/C-14 bond is possible for the
14-hydroxy derivative, electronic factors strongly favor a
coplanar orientation for the formyl and the endocyclic
double bond. Thus, while the hydroxymethyl derivative 18a
can attain inactive conformations having the hydroxyl
oriented toward the R-face of the molecule, resonance
effects maintain the formyl oxygen of 18b substantially
coplanar with the mean plane of the daucane skeleton,
preventing the perturbation of the critical lipophilic inter-
action.
In conclusion, capitalizing on the presence of two func-
tionalities, the monoacylated, hydrogen-bonded 1,3-diol
system and the double bond, a series of analogues of
ferutinin were synthesized, disclosing some interesting
trends in its structure-activity relationships. The 4-hy-
droxyl, the ester carbonyl, and their intramolecular hy-
drogen bonding were noncritical, while lipophilic interac-
tions involving the right-handed moiety of the molecule
were relevant. Further modifications of the terpenoid core
are conceivable, but could not be addressed starting from
the natural product. This limitation might be circumvented
by a total synthesis that incorporates the possibility to
generate focused structural diversity in the target, as
seminally outlined by Danishefsky in his manifesto of
“diverted” organic synthesis.18 The relevance of the ali-
phatic moiety is a unique feature of ferutinin that sets it
apart from all other phytoestrogens and justifies efforts
directed at its total synthesis.
Exp er im en ta l Section
Gen er a l Exp er im en ta l P r oced u r es. IR spectra were
1
obtained on a Shimadzu DR 8001 spectrophotometer. The H
and 13C NMR spectra were recorded at room temperature on
a Bruker DRX-300 spectrometer (300 and 75 MHz, respec-
tively) and on a Bruker DRX-500 (500 and 125 MHz). Chemical
shifts were referenced to the residual solvent signal (CDCl3,
δH 7.27, δC 76.9; CD3COCD3, δH 2.05, δC 206.26). HRMS were
recorded on a VG Prospect (Fisons) mass spectrometer. Silica
gel 60 (70-230 mesh) was used for gravity column chroma-
tography. HPLC separations were performed using a Micro-
porasil Waters (0.8 × 30 cm), equipped with a Waters 6000 A
pump and a Waters R 401 refraction index detector. CH2Cl2,
benzene, and toluene were dried by distillation from P4O10
,
and THF by distillation from Na-benzophenone. Ferutinin
esters (acetate, pivalate) were prepared by overnight treatment
of crude ferutinin (1a ) with an excess (5 molar equiv) of the
corresponding chloride in pyridine. Pivaloyl and acetyl depro-
tection of the p-hydroxybenzoyl (benzyl) moiety was affected
by overnight treatment with an excess (20 molar equiv) of
pyrrolidine (procedure A) or ethylenediamine (procedure B)

Estrogenic Sesquiterpene Ester Ferutinin
J ournal of Natural Products, 2004, Vol. 67, No. 9 1561
in dry CH2Cl2 or toluene (ca. 1.5 mL/mmol of substrate) and
workup by dilution with CH2Cl2 and washing with 1 N HCl
and brine. Organic solutions were dried with Na2SO4 prior to
evaporation. Zinc was activated by treatment with 1 N HCl,
filtration, and sequential washing with water, acetone, and
ether.
Deter m in a tion of Hor m on a l Activity in Vitr o. Estro-
genic activity was determined using an estrogen-inducible
yeast screen (Saccharomyces cerevisiae) expressing the human
estrogen receptor-R and containing expression plasmids car-
rying estrogen-responsive sequences controlling the reporter
gene lac-Z (encoding the enzyme â-galactosidase). Estrogenic
activity was determined from the metabolism of chlorophenol
red â-D-galactopyranoside by monitoring the absorbance at 540
nm.17a Antiestrogenic activity was investigated by determining
whether the compounds could inhibit the stimulatory response
to a submaximal concentration (1.56 nM) of estradiol.19
Androgenic and antiandrogenic activity were assessed using
a yeast screen expressing the human androgen receptor.19
J a esch k ea n a d iol p-Hyd r oxyben zyl Eth er (1c). Under
a nitrogen atmosphere, to a stirred solution of dimsyl anion
(obtained by reacting NaH (60% dispersion in mineral oil, 151
mg, 3.8 mmol, 2.2 molar equiv) with 5 mL of DMSO) was added
jaeschkeanadiol (1b, 208 mg, 0.87 mmol). After stirring 5 min
at room temperature, p-pivaloyloxybenzyl bromide20 (358 mg,
1,32 mmol, 1.5 molar equiv) was added. After stirring a further
60 h, the reaction was worked up by dilution with petroleum
ether-ether (3:1 v/v) and sequential washing with saturated
NH4Cl and brine. After drying and evaporation, the residue
was purified by column chromatography (15 g silica gel) eluted
with a petroleum ether-EtOAc gradient (from 98:2 to 95:5)
to afford the pure product (86 mg, 23%), which was deprotected
in the usual way (procedure B). After purification by column
chromatography (2.5 g silica gel) eluted with petroleum ether-
EtOAc gradient (from 9:1 to:8:2), 53 mg of 1c was obtained
(77%) as an amorphous foam: IR (KBr) νmax 3484, 1616, 1597,
1520, 1277, 1225, 1150, 1067, 1049 cm-1; 1H NMR (300 MHz,
CDCl3) δ 7.17 (2H, AA′, acyl), 6.74 (2H, BB′, acyl), 5.50 (1H,
br s, H-9), 4.65 (1H, d, J ) 12 Hz, H-1′), 4.20 (1H, d, J ) 12
Hz, H-1′), 3.55 (1H, br t, J ) 6.7 Hz, H-6), 1.88 (3H, br s, H-14),
0.94 (3H, br s, H-15), 0.87-0.85 (each 3H, d, J ) 6.7 Hz, H-12,
H-13); 13C NMR (75 MHz, CDCl3) δ 157.1 (s, C-5′), 133.3 (s,
C-8), 130.1 (d, C-3′), 128.1 (s, C-2′), 124.9 (d, C-9), 115.9 (d,
C-4′), 87.4 (s, C-4), 73.7 (d, C-6), 69.9 (t, C-1′), 61.5 (d, C-5),
43.4 (s, C-1), 41.3, 40.8, 39.6 (t, C-2, C-7, C-10), 37.4 (d, C-11),
30.6 (t, C-3), 20.7 (q, C-14), 19.7, 18.3, 17.1 (q, C-15, C-12, C-13;
HREIMS m/z 326.2271 [M - H2O]+ (calcd for C22H32O3 - H2O,
326.2246).
Oxid a tion of J a esch k ea n a d iol (1b). To a solution of 1b
(1.0 g, 4.20 mmol) in dry CH2Cl2 (10 mL) was added an excess
of PDC (6.32 g, 16.8 mmol, 4.0 molar equiv). After stirring at
room temperature for 24 h, the reaction was worked up by
dilution with ether and filtration through Celite. The filtrate
was washed with 2 N H2SO4 and brine, dried, and evaporated.
The residue was purified by column chromatography on silica
gel (25 g, petroleum ether-EtOAc, 95:5, as eluent) to afford
160 mg (15%) of 3 and 67 mg (7%) of 4.
1R,4R-4-Hyd r oxyd a u ca -7-en e-6,9-d ion e (3): yellow oil;
IR (liquid film) νmax 3497, 1651, 1616, 1468, 1445, 1379, 1196,
1171 cm-1; 1H NMR (300 MHz, CDCl3) δ 6.31 (1H, br s, H-7),
4.41 (1H, br s, OH), 2.85 (2H, br s, H-10), 2.79 (1H, s, H-5),
2.02 (3H, br s, H-14), 1.22 (3H, br s, H-15), 0.88-0.85 (each
3H, d, J ) 6.7 Hz, H-12, H-13, overlapped); 13C NMR (75 MHz,
CDCl3) δ 202.6, 200.7 (s, C-6 and C-9), 148.1 (s, C-8), 134.8 (d,
C-7), 86.3 (s, C-4), 63.4 (d, C-5), 58.4 (t, C-10), 42.5 (s, C-1),
40.5 (t, C-2), 37.6 (d, C-11), 34.0 (t, C-3), 22.5, 21.2 (q, C-14
and C-15), 18.0, 17.6 (q, C-12 and C-13); CIMS (isobutanol)
m/z 251 [M + 1]+ [C15H22O3 + H]+, 233 [M - H2O + H]+
[C15H22O3 - H2O + H]+.
br s, H-15), 1.05 (each 3H, d, J ) 6.8 Hz, H-12, H-13,
overlapped); 13C NMR (75 MHz, CDCl3) δ 199.7, 189.5 (s, C-9
and C-6), 167.0 (s, C-4), 144.0 (s, C-8), 139.8 (s, C-5), 138.2 (d,
C-7), 55.1 (t, C-10), 45.1 (s, C-1), 39.2 (t, C-3), 29.6 (t, C-2),
28.3 (d, C-11), 23.1 (q, C-14), 20.7, 20.6, 20.4 (q, C-12, C-13,
C-15); CIMS (isobutanol) m/z 233 [M + 1]+ [C15H20O2 + H]+.
Ep oxid a tion of J a esch k ea n a d iol. To a solution of 1b
(2.02 g, 8.48 mmol) in dry CH2Cl2 (8 mL) was added MCPBA
(77%, 2.93 g, 13.0 mmol, 1.5 molar equiv), resulting in the
formation of a milky suspension. After stirring 40 min at room
temperature, the reaction was worked up by dilution with CH2-
Cl2 and washing with 5% Na2S2O3, 5% NaOH, and brine. After
removal of the solvent, the residue was purified by column
chromatography on silica gel (65 g, packing with petroleum
ether-EtOAc, 9:1, elution with petroleum ether-EtOAc, 9:1
to 6:4). Fractions eluted with petroleum ether-EtOAc, 7:3,
afforded 65 mg (3%) of 5b, and fractions eluted with petroleum
ether-EtOAc, 6:4, afforded 1.87 g (87%) of 5a .
J a esch k ea n a d iol r-ep oxid e (5a ): amorphous foam; IR
(KBr) νmax 3416, 2959, 2863, 2361, 1379, 1051, 1030, 970, 878
cm-1; 1H NMR (300 MHz, CDCl3) δ 4.09 (1H, dt, J ) 10, 2 Hz,
H-6), 2.85 (1H, t, J ) 7 Hz, H-9), 1.41 (3H, br s, H-14), 1.15
(3H, br s, H-15), 0.95-0.89 (each 3H, d, J ) 6.7 Hz, H-12,
H-13); 13C NMR (75 MHz, CDCl3) δ 86.2 (s, C-4), 67.2 (d, C-6),
63.2, 61.0 (d, C-5 and C-9), 56.5 (s, C-8), 47.0 (s, C-1), 43.2,
41.0, 40.9 (t, C-2, C-7, C-10), 37.6 (d, C-11), 31.9 (t, C-3), 24.2
(q, C-14), 19.0, 18.1, 16.8 (q, C-15, C-12, C-13); HREIMS m/z
236.1713 [M - H2O]+ (calcd for C15H26O3 - H2O, 236.1776).
J a esch k ea n a d iol â-ep oxid e (5b): colorless gum; IR (KBr)
ν
max 3119, 3083, 2965, 2934, 1375, 1167, 1096, 1051, 1026 cm-1
;
1H NMR (300 MHz, CDCl3) δ 4.51 (1H, br s, OH), 4.01 (1H,
dt, J ) 10, 1 Hz, H-6), 2.91 (1H, d, J ) 5 Hz, H-9), 1.32 (3H,
br s, H-14), 1.24 (3H, br s, H-15), 0.93-0.89 (each 3H, d, J )
6.8 Hz, H-12, H-13); 13C NMR (75 MHz, CDCl3) δ 87.6 (s, C-4),
67.4 (d, C-6), 63.2, 57.9 (d, C-5 and C-9), 58.0 (s, C-8), 44.2 (s,
C-1), 43.9, 41.9, 40.2 (t, C-2, C-7, C-10), 37.5 (d, C-11), 32.9 (t,
C-3), 26.1 (q, C-14), 21.2 (q, C-15), 18.1, 16.9 (q, C-12, C-13);
HREIMS m/z 236.1719 [M - H2O]+ (calcd for C15H26O3 - H2O,
236.1776).
6-Deh yd r oja esch k ea n a d iol r-ep oxid e (6). To a solution
of 5a (1.60 g, 6.30 mmol) in dry CH2Cl2 (12 mL) was added
PCC (2.71 g, 12.6 mmol, 2.0 molar equiv). After stirring at
room temperature for 8 h, the reaction was worked up by
dilution with ether and filtration on Celite. The filtrate was
washed with 1 N HCl and brine, dried, and evaporated. The
residue was purified by column chromatography on silica gel
(40 g, petroleum ether-EtOAc, 95:5, as eluent) to afford 134
mg (9%) of 4 and 990 mg (62%) of 6. The latter was obtained
as a colorless powder: IR (KBr) νmax 3428, 2961, 2361, 1676,
1383, 1287, 1125, 1015, 884 cm-1; 1H NMR (300 MHz, CDCl3)
δ 4.83 (1H, br s, OH), 2.93 (1H, t, J ) 6.9 Hz, C-9), 2.76 (1H,
d, J ) 12 Hz, H-7R), 2.66 (1H, d, J ) 12 Hz, H-7â), 2.39 (1H,
br s, H-5), 1.35 (3H, br s, H-14), 1.17 (3H, br s, H-15), 0.86-
0.78 (each 3H, d, J ) 6.7 Hz, H-12, H-13); 13C NMR (75 MHz,
CDCl3) δ 210.8 (s, C-6), 85.7 (s, C-4), 67.4, 60.9 (d, C-5 and
C-9), 54.9 (s, C-8), 51.4 (t, C-7), 48.1 (s, C-1), 41.1, 39.2, 36.3
(t, C-2, C-3, C-10), 39.0 (q, C-11), 23.4, 20.1 (q, C-14 and C-15),
18.1, 17.6 (q, C-12, C-13); CIMS (isobutanol) m/z 253 [M + 1]+
[C15H24O3 + H]+, 235 [M - H2O + H]+ [C15H24O3 - H2O +
H]+.
6-Ep ija esch k ea n a d iol r-ep oxid e (7). To a solution of 6
(990 mg, 3.93 mmol) in MeOH (13 mL) was added an excess
of NaBH4 (223 mg, 5.89 mmol, 1.5 molar equiv) in ca. 1 min.
The reaction was then immediately worked up by dilution with
EtOAc, washing with saturated NH4Cl and brine, drying, and
evaporation. The residue was purified by column chromatog-
raphy on silica gel (25 g, petroleum ether-EtOAc, 9:1 to 8:2,
as eluent) to afford 890 mg (89%) of 7 as an amorphous foam,
which was used directly for the esterification step.
6-Ep ifer u tin in r-ep oxid e (8). To a solution of 7 (830 mg,
3.26 mmol) in dry toluene were added p-pivaloyloxybenxoic
acid21 (2.17 g, 9.78 mmol, 3.0 molar equiv), DCC (2.02 g, 9.78
mmol, 3.0 molar equiv), and DMAP (400 mg, 3.26 mmol, 1.0
molar equiv). The mixture was stirred at 80 °C overnight and
then worked up by dilution with EtOAc and washing with
1R-Da u ca -4,7-d ien e-6,9-d ion e (4): yellow oil; IR (liquid
film) νmax 1726, 1676, 1632, 1615, 1582, 1462, 1377, 1337, 1265
cm-1; 1H NMR (300 MHz, CDCl3) δ 6.65 (1H, s, H-7), 3.46 (1H,
hept, J ) 6.7 Hz, C-11), 2.91 (1H, d, J ) 15.1 Hz, H-10R), 2.77
(1H, d, J ) 15.1 Hz, H-10â), 2.02 (3H, br s, H-14), 1.19 (3H,

1562 J ournal of Natural Products, 2004, Vol. 67, No. 9
Appendino et al.
saturated Na2CO3 and brine. After drying and evaporation,
the residue was purified by column chromatography on silica
gel (30 g, petroleum ether-EtOAc, 9:1, as eluent) to afford 300
mg (20%) of the p-pivaloyloxybenzoic ester, a part of which
(100 mg) was deprotected according to procedure A. After
purification by column chromatography on silica gel (3 g,
petroleum ether-EtOAc, 9:1, as eluent), 56 mg of 8 was
obtained as a colorless gum (70%): IR (liquid film) νmax 3445,
ether-EtOAc, 8:2, elution with the same eluent) to afford 183
mg (87%) of 12 as an amorphous foam.22
8,9-Dih yd r o-8,14-d eh yd r o-9-h yd r oxyfer u tin in (13). To
a solution of ferutinin pivalate (250 mg, 0.566 mmol) in dry
CH2Cl2 was added MCPBA (77%, 190 mg, 0.848 mmol, 1.5
molar equiv). After stirring at room temperature for 2 h, the
reaction was worked up by dilution with CH2Cl2 and sequential
washing with 5% Na2S2O3, 5% NaOH, and brine. After drying
and removal of the solvent, the crude reaction mixture was
added to a solution of Cp2TiCl2 in THF (348 mg, 1.4 mmol, 2.5
equiv) previously treated with activated zinc (183 mg, 2.80
mmol, 5.0 molar equiv) for 45 min, until its color turned from
red to deep green. After stirring 15 min at room temperature,
the reaction was worked up by dilution with EtOAc and
sequential washing with 2 N H2SO4 and brine. After drying
and evaporation of the solvent, the residue was purified by
column chromatography on silica gel (8 g, elution with
petroleum ether-EtOAc, 9:1) to afford a pivaloyl derivative,
deprotected as usual (procedure A) to afford 60 mg of 13 as
an amorphous foam (overall 30%): IR (liquid film) νmax 3513,
1682, 1611, 1591, 1516, 1312, 1281, 1161, 1113, 1098 cm-1
;
1H NMR (300 MHz, CDCl3) δ 7.86 (2H, AA′, acyl), 6.88 (2H,
BB′, acyl), 5.42 (1H, br t, J ) 10 Hz, H-6), 2.92 (1H, t, J ) 7
Hz, H-9), 1.49 (3H, br s, H-14), 1.26 (3H, br s, H-15), 0.92-
0.84 (each 3H, d, J ) 6.7 Hz, H-12, H-13); 13C NMR (75 MHz,
CDCl3) δ 165.9 (s, C-1′), 161.2 (s, C-5′), 131.7 (d, C-3′), 121.1
(s, C-2′), 115.5 (d, C-4′), 86.7 (s, C-4), 70.7 (d, C-6), 61.0, 59.3
(d, C-5 and C-9), 58.6 (s, C-8), 44.8 (s, C-1), 43.7, 40.1, 39.5 (t,
C-2, C-7, C-10), 35.7 (d, C-11), 31.7 (t, C-3), 27.4, 25.9 (q, C-14
and C-15), 18.4, 17.1 (q, C-12, C-13); HREIMS m/z 356.205
[M - H2O]+ (calcd for C22H30O5 - H2O, 356.1988).
3,4-Deh yd r ofer u tin in (9). Under a nitrogen atmosphere,
a stirred and cooled (-50 °C) solution of ferutinin (1a ) (294
mg, 0.820 mmol) and DMAP (96 mg, 0.860 mmol, 1 molar
equiv) in dry pyridine (2 mL) was treated with thionyl chloride
(0.18 mL, 0.29 g, 2.5 mmol, 3.0 molar equiv). After stirring at
-50 °C for 2 h, the reaction was worked up by dilution with
CH2Cl2 and sequential washing with 2 N H2SO4, saturated
NaHCO3, and brine. After drying and evaporation, the residue
was purified by column chromatography on silica gel (8 g,
petroleum ether-EtOAc, 98:2 to 9:1, as eluent) to afford 253
mg (43%) of 9 as a yellow oil: IR (liquid film) νmax 3441, 1707,
1609, 1593, 1514, 1445, 1275, 1246, 1165, 1111 cm-1; 1H NMR
(300 MHz, CDCl3) δ 7.99 (2H, AA′, acyl), 7.53 (1H, br s, OH),
6.93 (2H, BB′, acyl), 5.55 (1H, br s, H-9), 5.40 (1H, br s, H-3),
5.30 (1H, m, H-6), 2.92 (1H, d, J ) 11 Hz, H-5), 1.93 (1H, dd,
J ) 15, 3 Hz, H-7â), 1.79 (3H, br s, H-14), 0.95 (3H, br s, H-15),
0.97-0.85 (each 3H, d, J ) 6.7 Hz, H-12, H-13); 13C NMR (75
MHz, CDCl3) δ 166.6 (s, C-1′), 160.7 (s, C-5′), 152.9 (s, C-4),
132.4 (s, C-8), 131.9 (d, C-3′), 125.0 (d, C-9), 122.1 (s, C-2′),
120,7 (d, C-3), 115.2 (d, C-4′), 71.8 (d, C-6), 60.7 (d, C-5), 45.8
(s, C-1), 46.9, 41.2, 39.2 (t, C-2, C-7, C-10), 28.1 (d, C-11), 26.7
(q, C-14), 22.7, 21.8, 20.8 (q, C-12, C-13, C-15); HREIMS m/z
340.2098 [M]+ (calcd for C22H28O3 340.2038).
1684, 1609, 1588, 1516, 1281, 1240, 1163, 1121, 1044 cm-1
;
1H NMR (300 MHz, CD3COCD3) δ 9.19 (1H, s, OH), 7.88 (2H,
AA′, acyl), 6.91 (2H, BB′, acyl), 5.46 (1H, m, H-6), 5.25-4.90
(each 1H, d, J ) 1.8 Hz, H-14, H-14′), 4,20 (1H, m, H-9), 3.50
(1H, s, OH), 1.22 (3H, br s, H-15), 0.92-0.84 (each 3H, d, J )
6.7 Hz, H-12, H-13); 13C NMR (75 MHz, CD3COCD3) δ 165.8
(s, C-1′), 162.1 (s, C-5′), 149.6 (s, C-8), 131.9 (d, C-3′), 122.8 (s,
C-2′), 115.6 (d, C-4′), 112.2 (t, C-14), 85.7 (s, C-4), 72.7, 72.2
(d, C-6 and C-9), 55.3 (d, C-5), 52.8 (s, C-1), 43.4, 42.5, 39.3 (t,
C-2, C-7, C-10), 36.5 (d, C-11), 32.1 (t, C-3), 20.8, 18.4, 17.5 (q,
C-15, C-12, C-13); HREIMS m/z 374.2112 [M]+ (calcd for
C
22H30O5, 374.2093).
8R,9S-8,9-Dih yd r o-8,9-d ih yd r oxyfer u tin in (14). To a
solution of ferutinin 1a (234 mg, 0.654 mmol) in CH3COCH3-
H2O, 4:1, were added OsO4 (0.2 M in toluene, 0.030 mL, 0.0060
mmol, 0.01 molar equiv) and NMMO (227 mg, 1.94 mmol; 3.0
molar equiv). After stirring at room temperature for 36 h, the
reaction was worked up by stirring with saturated Na2SO3 for
45 min; it was extracted with EtOAc and washed with brine.
After drying and removal of the solvent, the residue was
purified by column chromatography on silica gel (7 g, elution
with petroleum ether-EtOAc, 8:2 to 5:5) to afford 207 mg
(81%) of 14 as an amorphous foam: IR (KBr) νmax 3402, 1684,
1609, 1593, 1516, 1314, 1285, 1165, 1123, 1101 cm-1; 1H NMR
(300 MHz, CD3COCD3) δ 9.11 (1Η, s, ÃΗ), 7.91 (2H, AA′, acyl),
6.91 (2H, BB′, acyl), 5.59 (1H, dt, J ) 10, 3 Hz, H-6), 3.45 (1H,
m, H-9), 1.27 (3H, br s, H-14), 1.19 (3H, br s, H-15), 0.94-
0.84 (each 3H, d, J ) 6.7 Hz, H-12, H-13); 13C NMR (75 MHz,
CD3COCD3) δ 166.6 (s, C-1′), 162.4 (s, C-5′), 132.6 (d, C-3′),
123.7 (s, C-2′), 115.9 (d, C-4′), 86.3 (s, C-4), 75.1 (d, C-9), 75.0
(s, C-8), 71.3 (d, C-6), 53.5 (d, C-5), 50.1 (s, C-1), 44.2, 44.0,
42.8 (t, C-2, C-7, C-10), 37.0 (d, C-11), 32.9 (t, C-3), 30.4 (q,
C-14), 20.5, 19.0, 17.9 (q, C-15, C-12, C-13); HREIMS m/z
374.2066 [M - H2O]+ (calcd for C22H32O6 - H2O 374.2093).
8R,9H-8,9-Dih yd r o-8-h yd r oxy-9-oxofer u tin in (15). To a
solution of 14 (105 mg, 0.268 mmol) in EtOAc (3 mL) were
added PCC (373 mg, 1.73 mmol, 6.5 molar equiv) and Celite
(373 mg). After stirring at room temperature for 12 h, the
reaction was worked up by filtration over Celite and evapora-
tion. The residue was purified by column chromatography on
silica gel (3 g, petroleum ether-EtOAc, 8:2 to 6:4, as eluent)
to afford 46 mg (44%) of 15 as a yellow oil: IR (KBr) νmax 3402,
F er u tin in â-ep oxid e (11). To a solution of 5b (55 mg, 0.216
mmol) in dry toluene (2 mL) were added p-pivaloyloxybenzoic
acid (147 mg, 0.663 mmol, 3.0 molar equiv), DCC (134 mg,
0.648 mmol, 3.0 molar equiv), and catalytic DMAP (7.4 mg,
0.061 mmol, 0.28 molar equiv). After stirring overnight at
90 °C, the reaction was worked up by dilution with EtOAc and
washing with saturated NaHCO3 and brine. After drying and
evaporation, the residue was purified by column chromatog-
raphy on silica gel (2.5 g, petroleum ether-EtOAc, 9:1, as
eluent) to afford a p-pivaloyloxybenzoate (57 mg, 57%), depro-
tected following procedure B. The residue was purified by
column chromatography on silica gel (2 g, petroleum ether-
EtOAc, 8:2, as eluent) to afford 24 mg (52%) of 11 as an
amorphous foam: IR (KBr) νmax 3546, 3281, 2361, 1709, 1676,
1611, 1595, 1516, 1314, 1281, 1238, 1159, 1098 cm-1; 1H NMR
(300 MHz, CDCl3) δ 7.93 (2H, AA′, acyl), 7.09 (1H, s, OH), 6.91
(2H, BB′, acyl), 5.57 (1H, m, H-6), 2.99 (1H, t, J ) 6 Hz, H-9),
2.48 (1H, s, OH), 1.52 (3H, br s, H-14), 1.24 (3H, br s, H-15),
0.93-0.87 (each 3H, d, J ) 6.7 Hz, H-12, H-13); 13C NMR (75
MHz, CDCl3) δ 166.7 (s, C-1′), 160.8 (s, C-5′), 131.9 (d, C-3′),
121.8 (s, C-2′), 115.4 (d, C-4′), 85.2 (s, C-4), 71.4 (d, C-6), 59.9,
58.0 (d, C-5 and C-9), 52.7 (s, C-8), 42.7 (s, C-1), 42.2, 40.7,
40.0 (t, C-2, C-7, C-10), 36.8 (d, C-11), 31.9 (t, C-3), 24.2, 23.3
(q, C-14 and C-15), 18.4, 17.3 (q, C-12 and C-13); HREIMS
m/z 356.1999 [M - H2O]+ (calcd for C22H30O5 - H2O, 356.1988).
1705, 1609, 1593, 1514, 1312, 1283, 1165, 1121, 1100 cm-1
;
1H NMR (300 MHz, CDCl3) δ 7.87 (2H, AA′, acyl), 6.79 (2H,
BB′, acyl), 5.60 (1H, dt, J ) 10, 3 Hz, H-6), 1.35 (3H, br s,
H-14), 1.30 (3H, br s, H-15), 0.83-0.80 (each 3H, d, J ) 6.7
Hz, H-12, H-13); 13C NMR (75 MHz, CDCOCD3) δ 213.4 (s,
C-9), 167.0 (s, C-1′), 160.8 (s, C-5′), 132.1 (d, C-3′), 121.7 (s,
C-2′), 115.3 (d, C-4′), 84.4 (s, C-4), 77.4 (s, C-8), 69.7 (d, C-6),
50.9 (d, C-5), 42.0 (s, C-1), 48.7, 40.7, 40.2 (t, C-2, C-7, C-10),
36.6 (d, C-11), 31.7 (t, C-3), 27.9, 23.1 (q, C-14 and C-15), 18.1,
17.2 (q, C-12 and C-13); HREIMS m/z 390.2011 [M]+ (calcd
for C22H30O6, 390.2042).
F er u tin in r-ep oxid e (12). To a solution of 1a (200 mg,
0.559 mmol) in dry CH2Cl2 (2 mL) was added MCPBA (77%,
275 mg, 1.23 mmol, 2.2 molar equiv). After stirring 3 h at room
temperature, the reaction was worked up by dilution with CH2-
Cl2 and washing with 5% Na2S2O3, 5% NaOH, and brine. After
removal of the solvent, the residue was purified by column
chromatography on silica gel (3 g, packing with petroleum
8R,9R-8,9-Dih yd r o-8,9-d ih yd r oxyfer u tin in (16). To a
solution of 15 (166 mg, 0.423 mmol) in MeOH (2 mL) was

Estrogenic Sesquiterpene Ester Ferutinin
J ournal of Natural Products, 2004, Vol. 67, No. 9 1563
added portionwise an excess of NaBH4 (160 mg, 4.23 mmol,
10 molar equiv). After 2 h the reaction was worked up by
dilution with EtOAc, washing with saturated NH4Cl and brine,
drying, and evaporation. The residue was purified by column
chromatography on silica gel (3 g, petroleum ether-EtOAc,
7:3, as eluent) to afford a mixture. 16 (50%) was isolated only
after HPLC (hexane-EtOAc, 5:5) as an amorphous foam: IR
(KBr) νmax 3441, 1682, 1599, 1516, 1316, 1294, 1246, 1167,
1003 cm-1; 1H NMR (300 MHz, CDCl3) δ 7.89 (2H, AA′, acyl),
6.87 (2H, BB′, acyl), 5.73 (1H, br t, J ) 10 Hz, H-6), 3.85 (1H,
br s, H-9), 1.31, 1.28 (each 3H, br s, H-14 and H-15), 0.86-
0.84 (each 3H, d, J ) 6.7 Hz, H-12, H-13, overlapped); 13C NMR
(75 MHz, CDCl3) δ 166.6 (s, C-1′), 161.1 (s, C-5′), 131.8 (d, C-3′),
121.8 (s, C-2′), 115.4 (d, C-4′), 86.0 (s, C-4), 76.7 (d, C-9), 75.9
(s, C-8), 72.3 (d, C-6), 53.6 (d, C-5), 43.3 (s, C-1), 46.3, 43.1,
39.2 (t, C-2, C-7, C-10), 36.5 (d, C-11), 31.4 (t, C-3), 30.8 (q,
C-14), 21.5 (q, C-15), 18.4, 17.3 (q, C-12 and C-13); HREIMS
m/z 374.213 [M - H2O]+ (calcd for C22H32O6 - H2O, 374.2093).
When the reduction was carried out on crude 15, the seco-
daucane 17 could also be obtained as a minor reaction product.
This compound could not be completely purified from unfrag-
mented daucane impurities, but its structure elucidation could
nevertheless be clarified by 1D and 2D NMR experiments: 1H
NMR (500 MHz, CDCl3) δ 5.35 (1H, s, H-7), 4.57 (1H, dd, J )
3.8 and 7.4 Hz, 9-OH), 3.86 (1H, m, H-9R), 3.72 (1H, m, H-9â),
2.89 (1H, s, H-5), 2.06 (1H, ddd, J ) 5.3, 7.7, and 14.0 Hz,
H-3R), 2.04 (1H, m, H-11), 2.01 (3H, s, H-14), 1.96 (1H, ddd,
J ) 8.4, 8.6, and 14.0 Hz, H-3â), 1.93 (1H, ddd, J ) 4.2, 9.0,
and 14.7 Hz, H-10R), 1.70 (1H, ddd, J ) 7.7, 8.6, and 12.8 Hz,
H-2R), 1.66 (1H, ddd, J ) 5.3, 8.4, and 12.8 Hz, H-2â), 1.55
(1H, ddd, J ) 3.3, 5.8, and 14.7 Hz, H-10â), 0.97 (3H, s, H-15),
0.95 (3H, d, J ) 6.8, H-13), 0.94 (3H, d, J ) 6.8, H-12); 13C
NMR (125 MHz, CDCl3) δ 195.9 (s, C-6), 174.4 (s, C-8), 103.7
(s, C-7), 98.2 (s, C-4), 59.1 (t, C-9), 56.1 (d, C-5), 46.0 (t, C-10),
45.8 (s, C-1), 40.0 (t, C-2), 34.9 (d, C-11), 33.5 (t, C-3), 22.6 (q,
C-15), 21.7 (q, C-14), 18.1 (q, C-13), 17.6 (q, C-12); HREIMS
m/z 252.1734 [M]+ (calcd for C15H24O3, 252.1725).
8,9-Meth len fer u tin in (19). Under a nitrogen atmosphere,
a solution of ferutinin pivalate (250 mg, 0.565 mmol) in dry
benzene (20 mL) was treated with diethylzinc (1 N in hexanes,
9 mL, 6.53 g, 9.0 mmol, 16 molar equiv) and diiodomethane
(0.73 mL, 2.43 g, 9.0 mmol, 16 molar equiv). The mixture was
stirred at 65 °C for 12 h and then worked up by cooling and
sequential washing with 1 N HCl, saturated NaHCO3, and
finally brine. After drying and evaporation, the residue was
washed with petroleum ether and then deprotected (protocol
B). The final product was purified by column chromatography
over silica gel (5 g, petroleum ether-EtOAc, 8:2, as eluent) to
afford 74 mg of 19 as an amorphous foam (overall 35%): IR
(KBr) νmax 3349, 1703, 1669, 1609, 1595, 1512, 1314, 1277,
1248, 1167, 1100 cm-1; 1H NMR (300 MHz, CDCl3) δ 8.06 (1H,
s, OH), 7.91 (2H, AA′, acyl), 6.91 (2H, BB′, acyl), 5.51 (1H, dt,
J ) 10, 1 Hz, H-6), 2.68 (1H, s, OH), 1.30-1.21 (each 3H, br
s, H-14 and H-15), 0.90-0.86 (each 3H, d, J ) 6.7 Hz, H-12,
H-13), 0,70 (1H, m, H-9), 0,58 (1H, dd, J ) 8, 4 Hz, H-16),
0.21 (1H, t, J ) 4.4 Hz, H-16); 13C NMR (75 MHz, CDCl3) δ
167.8 (s, C-1′), 161.7 (s, C-5′), 132.4 (d, C-3′), 122.3 (s, C-2′),
116.0 (d, C-4′), 87.4 (s,C-4), 72.2 (d, C-6), 61.8 (d, C-5), 45.7 (s,
C-1), 45.7, 43.5, 41.0 (t, C-2, C-7, C-10), 37.4 (d, C-11), 32.3 (t,
C-3), 25.6, 19.9 (q, C-14 and 1C-5), 24.0 (q, C 16), 21.9 (d, C-9),
18.9, 18.0 (q, C-1 and C-13), 15.3 (s, C-8); HREIMS m/z
354.2119 [M - H2O]+ (calcd for C23H32O4 - H2O, 354.2195).
Dih yd r ofer u tin in (20). A methanol solution of ferutinin
(205 mg, 0.573 mmol) was hydrogenated (balloon) in the
presence of Pd(C) (20 mg). After 24 h, the reaction was worked
up by filtration over Celite to afford 179 mg (87%) of 20 as an
amorphous foam: IR (liquid film) νmax 2926, 2359, 1682, 1609,
1591, 1312, 1283, 1163, 1123, 1100 cm-1. Major isomer: 1H
NMR (500 MHz, CDCl3) δ 7.94 (2H, AA′, acyl), 6.89 (2H, BB′,
acyl), 5.52 (1H, ddd, J ) 4.1, 7.3 and 10.6 Hz, H-6), 2.09 (1H,
m, H-8), 2.07 (1H, d, J ) 10.5 Hz, H-5), 1.95 (1H, m, H-9R),
1.95 (1H, m, H-3R), 1.94 (1H, m, H-7R), 1.82 (1H, m, H-7â),
1.65 (1H, m, H-3â), 1.65 (1H, m, H-11), 1.64 (1H, m, H-2R),
1.61 (1H, m, H-9â), 1.57 (1H, m, H-2â), 1.47 (1H, m, H-10R),
1.22 (1H, m, H-10â), 1.19 (3H, s, H-15), 1.07 (3H, d, J ) 7.0,
H-14), 0.88 (3H, d, J ) 6.8 Hz, H-12), 0.85 (3H, d, J ) 6.8 Hz,
H-13); 13C NMR (125 MHz, CDCl3) δ 167.0 (s, C-1′), 160.4 (s,
C-5′), 132.0 (d, C-3′), 128.8 (s, C-2′), 115.4 (d, C-4′), 86.3 (s,
C-4), 74.6 (d, C-6), 57.7 (d, C-5), 44.0 (s, C-1), 42.7 (t, C-10),
41.5 (t, C-7), 38.8 (t, C-2), 36.7 (d, C-11), 31.9 (t, C-3), 31.7 (t,
C-9), 29.1 (d, C-8), 22.5 (q, C-14), 19.9 (q, C-15), 18.5 (q, C-12),
17.5 (q, C-13). Minor isomer: 1H NMR (500 MHz, CDCl3) δ
7.94 (2H, AA′, acyl), 6.89 (2H, BB′, acyl), 5.65 (1H, ddd, J )
1.3, 4.7 and 9.7 Hz, H-6), 2.23 (1H, d, J ) 9.7 Hz, H-5), 1.98
(1H, m, H-7R), 1.95 (1H, m, H-8), 1.94 (1H, m, H-2R), 1.82 (1H,
m, H-3R), 1.74 (1H, m, H-7â), 1.62 (1H, m, H-2â), 1.61 (2H,
m, H-9), 1.58 (1H, m, H-11), 1.52 (1H, m, H-10R), 1.34 (1H,
m, H-3â), 1.23 (1H, m, H-10â), 1.13 (3H, s, H-15), 0.94 (3H, d,
J ) 7.0 Hz, H-14), 0.88 (3H, d, J ) 6.8 Hz, H-12), 0.86 (3H, d,
J ) 6.8 Hz, H-13); 13C NMR (125 MHz, CDCl3) δ 166.8 (s, C-1′),
160.4 (s, C-5′), 132.0 (d, C-3′), 128.9 (s, C-2′), 115.3 (d, C-4′),
86.7 (s, C-4), 73.3 (d, C-6), 54.8 (d, C-5), 44.5 (s, C-1), 43.1 (t,
C-10), 44.1 (t, C-7), 43.1 (t, C-2), 36.6 (d, C-11), 31.7 (t, C-3),
36.1 (t, C-9), 29.6 (d, C-8), 25.2 (q, C-14), 20.2 (q, C-15), 18.5
(q, C-12), 17.5 (q, C-13); HREIMS m/z 360.2289 [M]+ (calcd
for C22H32O4, 360.2301).
Allylic Oxid a tion of F er u tin in Aceta te (1e). To a
suspension of SeO2 (60.1 mg, 0.542 mmol, 0.51 molar equiv)
in 2 mL of dry CH2Cl2 was added TBHP (3 M in toluene, 0.70
mL, 191 mg, 2.11 mmol, 2.0 molar equiv). After stirring at
room temperature for 20 min, a solution of ferutinin acetate
(422 mg, 1.06 mmol) in dry CH2Cl2 (3 mL) was added. Stirring
was continued at room temperature for 75 min, and then the
reaction was worked up by dilution with CH2Cl2 and washing
with 10% Na2SO3 and brine. After drying and evaporation, the
residue was purified by column chromatography on silica gel
(3 g, petroleum ether-EtOAc, 9:1 to 8:2, as eluent) to afford
15 mg (13%) of the 14-oxo derivative and 27 mg (22%) of the
14-hydroxy derivative, deprotected in the usual way (procedure
A) to afford, respectively, 18b and 18a .
14-Hyd r oxyfer u t in in (18a ): amorphous foam; IR (KBr)
ν
max 3402, 1686, 1609, 1593, 1516, 1312, 1277, 1165, 1119, 1099
cm-1; H NMR (300 MHz, CDCl3) δ 8.02 (2H, AA′, acyl), 7.17
(2H, BB′, acyl), 5.82 (1H, m, H-9), 5.26 (1H, dt, J ) 10, 3 Hz,
H-6), 4.09 (2H, br s, H-14), 1.11 (3H, br s, H-15), 0.99-0.88
(each 3H, d, J ) 6.7 Hz, H-12, H-13); 13C NMR (75 MHz,
CDCl3) δ 166.8 (s, C-1′), 160.7 (s, C-5′), 136.4 (s, C-8), 131.8
(d, C-3′), 127.8 (d, C-9), 121.9 (s, C-2′), 115.3 (d, C-4′), 86.6 (s,
C-4), 71.4 (d, C-6), 68.6 (t, C-14), 60.0 (d, C-5), 43.9 (s, C-1),
41.1, 40.5, 37.0 (t, C-2, C-7, C-10), 36.8 (d, C-11), 31.4 (t, C-3),
20.2, 18.5, 17.4 (q, C-15, C-12, C-13); HREIMS m/z 374.2074
[M]+ (calcd for C22H30O5, 374.2093).
1
Ack n ow led gm en t. We are grateful to Prof. M. Ballero
(Universita` di Cagliari, Italy) for a generous gift of the
nonpoisonous chemotype of giant fennel from Sardinia, and
to Prof. J . Sumpter (Brunel University, Uxbridge, UK) for
providing the estrogen-responsive yeast screen. Financial
support from MIUR (Fondi ex-40%, progetto Sostanze naturali
ed analoghi sintetici con attivita` antitumorale) is gratefully
acknowledged.
14-Oxofer u tin in (18b): amorphous foam; IR (KBr) νmax
3398, 1699, 1682, 1611, 1593, 1516, 1312, 1277, 1242, 1167,
1
1098 cm-1; H NMR (300 MHz, CDCl3) δ 9.41 (1H, s, H-14),
7.87 (2H, AA′, acyl), 6.93 (1H, m, H-9), 6.87 (2H, BB′, acyl),
5.26 (1H, dt, J ) 10, 3 Hz, H-6), 1.12 (3H, br s, H-15), 0.93-
0.85 (each 3H, d, J ) 6.7 Hz, H-12, H-13); 13C NMR (75 MHz,
CDCl3) δ 194.2 (d, C-14), 167.5 (s, C-1′), 161.9 (s, C-5′), 156.8
(d, C-9), 140.8 (s, C-8), 132.5 (d, C-3′), 121.8 (s, C-2′), 116.0 (d,
C-4′), 87.4 (s, C-4), 70.7 (d, C-6), 60.1 (d, C-5), 44.3 (s, C-1),
43.0, 41.8 (t, C-2 and C-10), 37.4 (d, C-11), 31.7, 31.3 (t, C-3
and C-7), 21.0 (q, C-15), 18.9, 17.9 (q, C-12 and C-13); HREIMS
m/z 372.1950 [M]+ (calcd for C22H28O5, 372.1937).
Refer en ces a n d Notes
(1) Elsby, R.; Ashby, J .; Sumpter, J . P.; Brooks, A. N.; Pennie, W. D.;
Maggs, J . L.; Lefevre, P. A.; Odum, J .; Beresford, N.; Paton, D.; Park,
B. K. Biochem. Pharmacol. 2000, 60, 1519-1530.
(2) Brzozowski, A. M.; Pike, A. C.; Dauter, Z.; Hubbard, R. E.; Bonn, T.;
Engstrom, O.; Ohman, L.; Greene, G. L.; Gustafsson, J .-A.; Carlquist,
M. Nature 1997, 389, 753-758.

1564 J ournal of Natural Products, 2004, Vol. 67, No. 9
Appendino et al.
(3) (a) Ignatkov, V. I.; Ahmedhodzjaeva, H. T.; Babichev, V. Farmakol.
Toksikol. 1990, 53, 37-38. (b) Ikeda, K.; Arao, Y.; Otsuka, H.; Nomato,
S.; Horiguchi, H.; Kato, S.; Kayama, F. Biochem. Biophys. Res.
Commun. 2002, 29, 354-360.
(Skiraman, M. C.; Nagasampagi, B. A.; Pandey, R. C.; Dev, S.
Tetrahedron 1973, 29, 985-991, and ref 6).
(12) Mathieson, A. L. Tetrahedron Lett. 1965, 4137-4140.
(13) De Luca, L.; Giacomelli, G.; Porcheddu, A. Org. Lett. 2001, 3, 3041-
3043.
(14) Schobert, R. Angew. Chem., Int. Ed. Engl. 1988, 27, 2693-2703.
(15) Tolstikov, G. A.; Akbutina, F. A.; Dembitskii, A. D.; Valeev, F. A.;
Miftakhov, M. S. Zh. Org. Chim. 1992, 28, 2081-2090.
(16) Furukawa, J .; Kawabata, N.; Nishimura, J . Tetrahedron 1968, 24,
53-58.
(17) (a) Routledge, E. J .; Sumpter, J . P. Environ. Toxicol. Chem. 1996,
15, 241-258. (b) Milligan, S. R.; Kalita, J . C.; Heyerick, A.; Rong,
H.; De Cooman, L.; De Keukeleire, D. J . Clin. Endocrinol. Med. 1999,
83, 2249-2252.
(18) Njardarson, J . T.; Gaul, C.; Shah, B.; Hjang, X.-Y.; Danishefsky, S.
J . J . Am. Chem. Soc. 2004, 126, 1038-1040.
(19) Sohoni, P.; Sumpter, J . P. J . Endocrinol. 1998, 158, 327-339.
(20) Prepared in overall 24% yield from 4-hydroxybenzaldehyde by reduc-
tion (4 molar equiv of NaBH4, MeOH), pivaloylation (1.6 molar equiv
of pivaloyl chloride, 1.6 molar equiv of KOH, dioxane), and treatment
with NBS (1.5 molar equiv, 1.2 molar equiv of TPP, CH2Cl2).
(21) Prepared in 85% yield by pivaloylation of p-hydroxybenzoic acid (4
molar equiv of pivaloyl chloride, 0.3 molar equiv of DMAP, pyridine).
(22) Diab, Y.; Dolmazon, R.; Bessie`re, J .-M. Flavour Fragr. J . 2001, 16,
120-122.
(4) Appendino, G.; Cravotto, G.; Sterner, O.; Ballero, M. J . Nat. Prod.
2001, 64, 393-395.
(5) (a) Hadidi, K. A.; Aburjai, T.; Battah, A. K. Fitoterapia 2003, 74, 242-
246. (b) El-Thaher, T. S.; Matalka, K. Z.; Taha, H. A.; Badwan, A. A.
Int. J . Impot. Res. 2001, 13, 247-251.
(6) Appendino, G.; Spagliardi, P.; Cravotto, G.; Pocock, V.; Milligan, S.
J . Nat. Prod. 2002, 65, 1612-1615.
(7) Nimmagudda, R. R.; Ramanathan, R.; Putcha, L. Biochem. Arch.
1997, 13, 171-178.
(8) Griffiths, L. A.; Smith, G. E. Biochem. J . 1972, 128, 901-911.
(9) Liebich, H. M.; Forst, C. J . Chromatogr. 1990, 525, 1-14.
(10) Routledge, E. J .; Parker, J .; Odum, J .; Ashby, J .; Sumpter, J . P.
Toxicol. Appl. Pharmacol. 1998, 153, 12-19. Alkyl p-hydroxyben-
zoates (parabens) from antiperspirants and deodorants were detected
in breast tumors (Darbre, P. D. J . Appl. Toxicol. 2003, 23, 89-95),
and frequent underarm shaving combined with deodorant use has
been linked to breast cancer (McGrath, K. G. Eur. J . Cancer Prev.
2003, 23, 479-485), though the matter is controversial (Whelan, J .
New Sci. 2004, J an 24, p 11).
(11) The presence of this eight-membered intramolecular hydrogen bond-
ing is in accordance with the sharp hydroxyl absorption (3300-3500
cm-1, unaffected by dilution) in the IR spectra of jaeschkeanadiol
esters and the low frequency of the ester carbonyl absorption (1730-
1710 cm-1 for aliphatic esters, <1700 cm-1 for the aromatic ones)
NP049796W