Spiroperoxy lactones from furans in one pot: Synthesis of (+)-premnalane A

Article abstract of DOI:10.1021/ol702575a

A [4+2]-cycloaddition between singlet oxygen and a furan, followed by an ene reaction and ketalization, in one synthetic operation, was used for the synthesis of (+)-premnalane A. The first example of a singlet oxygen ene reaction that furnishes exclusively a Z-double bond is noted.

Full text of DOI:10.1021/ol702575a

ORGANIC
LETTERS
2007
Vol. 9, No. 26
5585-5588
Spiroperoxy Lactones from Furans in
One Pot: Synthesis of ( )-Premnalane A
+
Ioannis Margaros, Tamsyn Montagnon, and Georgios Vassilikogiannakis*
Department of Chemistry, UniVersity of Crete, Vasilika Vouton,
71003 Iraklion, Crete, Greece
Vasil@chemistry.uoc.gr
Received October 23, 2007
ABSTRACT
A [4+2]-cycloaddition between singlet oxygen and a furan, followed by an ene reaction and ketalization, in one synthetic operation, was used
for the synthesis of (
+
)-premnalane A. The first example of a singlet oxygen ene reaction that furnishes exclusively a Z-double bond is noted.
Singlet oxygen (1O2) has considerable unmined potential as
a synthetic tool. It is a highly reactive yet selective oxidant
(protecting groups are rendered essentially redundant), which
is readily generated. Furthermore, it is an excellent mediator
of cascade reaction sequences. Recent work from our group
has focused on designing and implementing such sequences,
which use singlet oxygen’s unique reactivity, to target a
selection of different natural products.1 It was in this context
that we developed an interest in the rare γ-spiroperoxy
lactone motif F (Scheme 1) of premnalane A,2 an antibacte-
rial natural product isolated from the shrub, Premna oligtricha,
found in the Sidamo province of Ethiopia.3
our strategy for the synthesis of premnalane A’s unique
γ-spiroperoxy lactone core was updated (Scheme 1). In the
new proposal, two different modes of 1O2 reaction, a [4+2]-
cycloaddition and an ene reaction, were to be called upon
to stitch up the spirolactone unit in an efficient one-pot
operation. Thus, a [4+2]-cycloaddition5 between the initial
1
substrate’s furan moiety and O2, furnishing a 4-hydroxy-
butenolide portion (A f B, Scheme 1), would initiate the
reaction sequence with an ene reaction following immediately
afterward. There are a number of possible outcomes to the
ene reaction resulting from the different orientations and
pathways open to the intermediates. We deconvoluted these
issues in our planning, by applying known mechanistic details
regarding ene reactions (such as the cis-effect6 and the large
group effect7) to the premnalane A system. From hydroxy-
butenolide B, there are two possible sites for hydrogen
abstraction, sites a and b. Precedent suggests that in the
simplest systems abstraction from site a predominates;7
Our first approach to the synthesis of premnalane A had,
1
at its heart, a [4+2]-addition between O2 and a diene;
however, this approach failed.4 In the light of this result,
(1) (a) Vassilikogiannakis, G.; Stratakis, M. Angew. Chem., Int. Ed. 2003,
42, 5465-5468. (b) Vassilikogiannakis, G.; Margaros, I.; Montagnon, T.;
Stratakis, M. Chem. Eur. J. 2005, 11, 5899-5907. (c) Sofikiti, N.; Tofi,
M.; Montagnon, T.; Vassilikogiannakis, G.; Stratakis, M. Org. Lett. 2005,
7, 2357-2359. (d) Georgiou, T.; Tofi, M.; Montagnon, T.; Vassilikogian-
nakis, G. Org. Lett. 2006, 8, 1945-1948. (e) Tofi, M.; Montagnon, T.;
Georgiou, T.; Vassilikogiannakis, G. Org. Biomol. Chem. 2007, 5, 772-
777.
(2) Premnalane A is our designation for the natural product 1 that was
given only an IUPAC descriptor by the isolation team (see ref 3).
(3) Habtemariam, S.; Gray, A. I.; Lavaud, C.; Massiot, G.; Skelton, B.
W.; Waterman, P. G.; White, A. H. J. Chem. Soc., Perkin Trans. 1 1991,
893-896.
(4) Margaros, I.; Montagnon, T.; Tofi, M.; Pavlakos, E.; Vassilikogian-
nakis, G. Tetrahedron 2006, 62, 5308-5317.
(5) For general reviews of furan photooxygenation, see: (a) Gollnick,
K.; Griesbeck, A. Tetrahedron 1985, 41, 2057-2068. (b) Feringa, B. L.
Recl. TraV. Chim. Pays-Bas 1987, 106, 469-488.
(6) Stephenson, L.; Grdina, M. J.; Orfanopoulos, M. Acc. Chem. Res.
1980, 13, 419-425.
(7) For a review on regioselectivity in the ene reaction of 1O2 with
alkenes, see: Stratakis, M.; Orfanopoulos, M. Tetrahedron 2000, 56, 1595-
1615.
10.1021/ol702575a CCC: $37.00
© 2007 American Chemical Society
Published on Web 11/21/2007

Scheme 1. The Concept
Scheme 2. Synthesis of the Key Precursor 5
Our first task was to install the requisite furan moiety. To
this end, we began by reducing the lactone functionality of
2 to the corresponding lactol using Dibal-H (94%, 8:1 lactol/
open hydroxyaldehyde). Elimination of the tertiary alcohol
using BF3‚OEt2 (fast addition) furnished aldehyde 3 (con-
taminated with minor amounts of the ∆7,8 regioisomer, ∆8,9/
∆
7,8 18:1, Scheme 2) in 93% combined yield. The furan was
then constructed in three steps using the method of Marshall9
as modified by Winkler et al.10 The organostannane, derived
from 1-bromo-2-butyne,11 was added to aldehyde 3 to afford
a diastereomeric mixture of allenic alcohols in 75% yield.
These alcohols were oxidized to the corresponding allenic
ketone 4 using IBX (95% yield). Treatment of allenic ketone
4 with AgNO3 on silica gel furnished the desired furan (90%
yield). Before attempting the singlet oxygen cascade se-
quence, a silyl group was smoothly introduced onto the furan
(5, 95%). This functionality has been shown to dramatically
improve the yield of the transformation of furan-derived
endoperoxides into 4-hydroxybutenolides (the first steps of
the proposed reaction sequence).12 The substrate for the
singlet oxygen-mediated reaction sequence had now been
successfully synthesized in a very straightforward manner
(53% overall yield for six steps).
however, when the starting double bond is substituted with
a large group (LG), abstraction of type b has been shown to
eclipse site a abstraction (the so-called large group effect).8
In the case of premnalane A, there is more than one large
group (the decalin ring and the 4-hydroxybutenolide) as well
as considerable restriction on the conformational and rota-
tional freedom of various key bonds and positions. As a
result, we were unsure of exactly how the premnalane A
system might react. Finally, a very unusual feature was
required from the sequence if success were to be ensured;
the double bond formed in the ene reaction would need to
have Z-geometry (intermediate D, Scheme 1) for the final
cyclization step to be accomplished. This geometrical prefer-
ence has not been reported previously, but we felt that the
large steric demands of the 4-hydroxybutenolide and the
decalin ring might conspire to help us attain this unprec-
edented geometry. Despite its risky nature, we felt the
sequence held enough potential to be worth investigating in
the laboratory.
The stage was now set to investigate if we could assemble
the entire endoperoxy-spirolactone unit in one-pot. We began
by validating each key step. Neutralized CDCl3 (K2CO3) was
initially employed as the solvent to monitor the reaction’s
progress by 1H NMR (Scheme 3). When furan 5 was treated
1
with O2 [generated by bubbling air gently through the
The decalin system of (+)-sclareolide (2, Scheme 2), the
chosen starting point for our synthetic investigation, is
enantiomeric to that reported for premnalane A, but this
difference in absolute stereochemistry was not relevant to
our primary goal which was a thorough investigation of the
proposed singlet oxygen-mediated reaction sequence.
(9) (a) Marshall, J. A.; Wang, X.-J. J. Org. Chem. 1991, 56, 960-969.
(b) Marshall, J. A.; Bartley, G. S. J. Org. Chem. 1994, 59, 7169-7171. (c)
Marshall, J. A.; Sehon, C. A. J. Org. Chem. 1995, 60, 5966-5968.
(10) Winkler, J. D.; Quinn, K. J.; MacKinnon, C. H.; Hiscock, S. D.;
McLaughlin, E. C. Org. Lett. 2003, 5, 1805-1808.
(11) Mukaiyama, T.; Harada, T. Chem. Lett. 1981, 621-624.
(12) For the introduction of the stabilizing effect of the silyl substituent,
see: Adam, W.; Rodriguez, A. Tetrahedron Lett. 1981, 22, 3505-3508.
For direct comparisons between photooxygenations of silyl- and unsubsti-
tuted furans, see: (a) Bury, P.; Hareau, G.; Kocien˜ski, P. J.; Dhanak, D.
Tetrahedron 1994, 50, 8793-8808. (b) Shiraki, R.; Sumino, A.; Tadano,
K.; Ogawa, S. J. Org. Chem. 1996, 61, 2845-2852.
(8) Orfanopoulos, M.; Stratakis, M.; Elemes, Y. J. Am. Chem. Soc. 1990,
112, 6417-6419.
5586
Org. Lett., Vol. 9, No. 26, 2007

Scheme 3. Validation of the Proposed Reaction Sequence and
Synthesis of (+)-premnalene A (8b)
Scheme 4. Mechanistic Rationale
reaction solution containing catalytic amounts (10-4 M) of
the sensitizer methylene blue, at 8 °C, in the presence of
visible spectrum light] for 30 s we observed the smooth
1
formation of the 4-hydroxybutenolide 6 (by H NMR). If
the irradiation time was extended to 3 min we observed the
formation of allylic hydroperoxides from an ene reaction,
(I, J, and L, Scheme 4). These hydroperoxides were cyclized
in the desired manner on treatment with catalytic PTSA (7/
8, 2:1; overall yield for three steps, 62%). The two different
γ-spiroperoxy lactones, 7 and 8, could be separated readily
by column chromatography. Thus, we had validated each
individual step of our proposed cascade reaction sequence,
and, in so-doing, synthesized (+)-premnalane A (8b). That
one of the diastereoisomers of 8 was indeed (+)-premnalane
A was confirmed by X-ray crystallography (Figure 1).13
Furthermore, extensive NOE studies clarified that in the case
of 7, both spirocycle isomers at C-12 were present, whereas,
with 8, only one spirocycle stereochemistry was observed,
but both stereoisomers of the C-8 center were present (Figure
1).
reaction conditions on the one-pot operation. Changes in
solvent could indeed engineer the desired changes in the
We next attempted to complete the sequence as a one-pot
operation. This objective was readily achieved; if neutralized
CDCl3 was replaced with untreated CHCl3, furan 5 could
be transformed directly into the endoperoxy spirolactone
products 7 and 8 (after 2.5 min of irradiation followed by
treatment for 30 min with catalytic amount of SiO2) in an
overall yield of 73% (7/8, 2:1).
With hope of improvements in the yield of premnalane A
(8b), we sought to investigate the effect of changes to the
Figure 1. Selected NOE’s observed for 7a, 7b, and 8a and ORTEP
representation obtained for (+)-premnalene A (8b).
(13) CCDC 658553 contains the supplementary crystallographic data.
These data can be obtained free of charge from The Cambridge Crystal-
lographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
Org. Lett., Vol. 9, No. 26, 2007
5587

product distribution. Thus, when CHCl3 was exchanged with
highly polar MeOH, the undesired [5,5]-spirocycle 7 was
strongly favored over the desired [6,5]-spirocyclic compound
8 (7/8, 4:1); however, when a nonpolar solvent such as
toluene was employed the ratio improved significantly in
favor of the desired [6,5]-spirocyclic compound 8 (7/8, 1.3:
1). Increasing the reaction temperature (from 8 to 65 °C)
made no significant change in the [5,5]-spirocyclic to [6,5]-
spirocyclic product ratio. Within the product distribution, the
two diastereomeric relationships are also of interest. The 8a/
8b ratio remains unchanged, independent of reaction condi-
tions; this feature is in stark contrast to the 7a/7b ratio, which
varies according to the reaction conditions. Presumably, in
the case of 7a/b, the spirocyclic center may equilibrate so
that the ratio is dependent on the conditions, whereas, with
8a/b the C-8 center is fixed during the course of the reaction.
Tracing the source of each of these four different reaction
products provides an interesting analysis of the competing
factors at work in this series of reactions (Scheme 4). Thus,
singlet oxygen may approach either one of the two faces of
the double bond in the ene reaction (approach a or b). The
face opposite the C-10 methyl group is favored because it
avoids a steric clash between the incoming electrophile and
the aforementioned methyl group (see Scheme 4, approach
a, 6 f G). The resultant perepoxide sits so that stabilization
from its interaction with allylic hydrogens is maximized, as
shown in intermediate G.6 From the final product distribution
we know that hydrogen abstraction then occurs mainly from
the C-8 methyl group (G f I). In other words, in this case
the decalin ring system (C-10) is not behaving like a large
group and the large group effect8 is not observed (minor
amounts of the product wherein Ha is abstracted are also
seen, G f J). Cyclization of I then furnishes a mixture of
spirocycles 7a,b (variable ratios depending on reaction
conditions) at the newly formed spirocycle center, while
cyclization of the minor isomer J furnishes 8-epi-premnalane
A (8a) only. Returning to the original approach of singlet
oxygen to the double bond, when approach b is followed
we see a different set of influences. Here, the C-10 methyl
group on the decalin bridge now sits adjacent to, and on the
same face as, the perepoxide (oriented as shown in inter-
mediate H to maximize allylic interactions6). Thus it forces
the decalin substituent (C-10) to act as a large group, and
the large group effect is observed in the subsequent ene
reaction.8 In other words, only hydrogen Hb is abstracted
(K f L). Furthermore, to avoid steric clashes between the
4-hydroxybutenolide and the decalin skeleton (shown in
intermediate H), Hb is abstracted in such a manner that the
double bond formed from this reaction has exclusively the
Z-geometry (L but also G f J). A double bond with
Z-geometry arises from the ene reaction of 1O2, and a double
bond is, to the best of our knowledge, without precedent7
and as such provides an appropriate climax to this ambitious
one-pot reaction sequence. The hydroperoxide product L then
cyclizes to give a single spirocycle stereoisomer, (+)-
premnalane A (8b).
In conclusion, we have synthesized (+)-premnalane A
using a remarkable 1O2-mediated one-pot reaction sequence
which first employs a [4+2]-cycloaddition between 1O2 and
1
a furan, then a O2 ene reaction, and, finally ketalization.
Furthermore, we have also noted the first example of an ene
reaction of this type that furnishes exclusively a Z-double
bond.
Acknowledgment. This research was supported by a
Marie Curie Intra-European Fellowship (T.M.) within the
sixth European Community Framework Programme and was
co-funded by the European Social Fund and National
Resources (B EPEAEK, Pythagoras Program). We thank the
National Fellowship Institute (IKY) for a fellowship (I.M.).
We are grateful to Prof. P. Trikalitis (University of Crete)
for obtaining the X-ray crystallographic data.
Supporting Information Available: Experimental pro-
cedures and full spectroscopic data for all new compounds.
This material is available free of charge via the Internet at
http://pubs.acs.org.
OL702575A
5588
Org. Lett., Vol. 9, No. 26, 2007