An approach to the tricyclic lactone core of brasoside and related natural products

Article abstract of DOI:10.1055/s-2004-835637

A one-pot iodine atom transfer/cyclisation and intermolecular olefination is described for the efficient, stereoselective construction of bicyclic lactone intermediates en route to compound 21, a potential precursor to the brasoside (1) skeleton.

Full text of DOI:10.1055/s-2004-835637

2788
LETTER
An Approach to the Tricyclic Lactone Core of Brasoside and Related Natural
Products
An
Appro
e
achtothe
T
r
ricyclic
e
Lactone Co
m
re of Brasoside y Robertson,*a Morgan Ménard,a Rhonan Fordb
a
Department of Chemistry, University of Oxford, Chemistry Research Laboratory, Mansfield Road, Oxford, OX1 3TA, UK
E-mail: jeremy.robertson@chem.ox.ac.uk
b
AstraZeneca R & D Charnwood, Medicinal Chemistry, Bakewell Road, Loughborough, LE11 5RH, UK
Received 5 August 2004
at the X-bearing position should be of no consequence.
Thus, synthesis of aldehyde 3, by functional group inter-
conversion and formylation, was expected to yield the
aglucone of brasoside (2) by cyclisation3 in situ.
Abstract: A one-pot iodine atom transfer/cyclisation and inter-
molecular olefination is described for the efficient, stereoselective
construction of bicyclic lactone intermediates en route to compound
21, a potential precursor to the brasoside (1) skeleton.
Key words: cyclisations, domino reactions, natural products, radi- Compounds of general structure 5 have featured in a vari-
cal reactions, stereoselective synthesis
ety of total synthesis endeavours and there is plenty of
precedent from which to draw in planning their synthesis;
however, in the light of separate results within our group,4
an attractive possibility was a free-radical group transfer
cyclisation of a suitable cyclopentyl ester (Scheme 2,
X–R¢ being the starting cyclopentenyl ester).
Littoralisone is a neuritogenic heptacyclic iridolactone
that may be considered to be derived biosynthetically
from brasoside (1) purely on the basis of their related
structures and their co-occurrence in Verbena littoralis.1
As part of our general interest in developing new synthetic
methodology within the context of complex natural
product synthesis, we initiated a programme of research to
effect a total synthesis of brasoside as a prelude to littoral-
isone. This letter summarises our first approach to braso-
side which was based, in part, on the conversion 6 → 7 in
Vidari’s formal synthesis of ( )-loganin (Scheme 1).2
O
O
O
O
O
O
H
H
H
X–R′
H
H
R
R
R
X
Scheme 2
O
Studies initiated with application to model iodoacetate 8
(Scheme 3) of the iodine atom transfer general method de-
scribed by Curran.5 In this process, it was expected that
the slow cyclisation of the a-acetyl radical would proceed
in synthetically useful yield because the more populated s-
trans ester conformation cannot suffer direct reduction in
the absence of a suitable H-atom source. In the event, ex-
posure of a mixture of ester 8 and hexabutyldistannane in
benzene to a 300 W tungsten filament lamp, at various
concentrations (0.02–0.3 M), and in either Pyrex or quartz
vessels, led to generally disappointing results, with a mix-
ture of products, apparently deriving from trapping of the
a-acetyl radical with either the stannane or the solvent, ac-
companying the desired product 9, which was isolated in
a maximum yield of 63% as a ca. 2:1 ratio of exo:endo
iodo-diastereomers.
O
O
O
H
H
O
O
H
H
O
H
H
OH
H
Me
OR
CHO
X
R
R
1: brasoside, R = Glu
2: R = H
3: R = CH3
4: R = H
5
OH
CO2Me
H
H
H
O
5% aq HCl,
THF (1:1), 20 h
(ref. 2)
CO2Me
O
O
O
O
6
7
O
OH
Scheme 1
In seeking a more practical alternative, attention then
turned to Zard’s elegant dithiocarbonate group transfer
method6 in which the intermediate a-acetyl radical could
only suffer degenerate intermolecular group transfer, or
group transfer following cyclisation. Thus, xanthate 12,
prepared in essentially quantitative yield (98%) from the
iodide 8, was heated in the presence of dilauroyl peroxide
(DLP) to yield exo-xanthate derivative 13 in 66% yield
(Scheme 4). This approach was operationally straight-
A key feature of the formation of hydroxydihydropyran 7
is that acetal hydrolysis under acidic conditions was
accompanied by trapping of the epimerised aldehyde; in
other words, in a projected synthesis of brasoside from an
intermediate of the form 5 (R = CH3), the stereochemistry
SYNLETT 2004, No. 15, pp 2788–2790
0
7
.1
2
.2
0
0
4
Advanced online publication: 08.11.2004
DOI: 10.1055/s-2004-835637; Art ID: D22504ST
© Georg Thieme Verlag Stuttgart · New York

LETTER
An Approach to the Tricyclic Lactone Core of Brasoside
2789
O
A convenient solution was found in another contribution
by Zard.9 Thus, heating iodide 9a in chlorobenzene in the
presence of ethyl 2-phenylvinyl sulfone (EPVS) and tert-
butyl peroxide gave, in a slow reaction, the exo-b-styryl
derivative 18 as a single diastereomer (67%) after recrys-
tallisation from diethyl ether.
I
O
O
O
1) NaBH4,
CeCl3·7H2O, MeOH
8
2) ICH2CO2H, PPh3,
DEAD, toluene (47%)
O
Given the experimental similarity of the two homolytic
steps leading to lactone 18, it was an attractive possibility
to run the two processes – atom transfer cyclisation and
olefination – as a single step. In practice, this worked very
well: iodide 8 was heated with DLP in chlorobenzene at
95 °C until formation of lactone 9 was complete then
EPVS and tert-butyl peroxide were added and the temper-
ature raised to 145 °C to complete the reaction; lactone 18
was then isolated in 57% yield as a single diastereomer
(Scheme 5).10
O
X
H
hν, Bu6Sn2,
O
PhH
H
10, X = SnBu3
11, X = Ph
9
I
Scheme 3
forward and provided the bicyclic product in diastereo-
merically pure form.
Application of DLP-mediation to related iodine atom
transfer reactions has been reported by Renaud et al.7 and,
pleasingly, heating iodide 8 in benzene in the presence of
DLP led to clean cyclisation and atom transfer to give a
5.9:1 ratio of exo:endo iodo-diastereomers 9a and 9b that
were separated for analytical purposes. Because this reac-
tion was both practical and tin-free, it became our method
of choice for the production of bicyclic products of gener-
al structure 5. When the procedure was applied to an
alkylated cyclopentenyl ester [(+)-14],4 the reaction was
less efficient and markedly less stereoselective (15:16,
2.2:1; Scheme 4).
O
O
O
KCN, DMSO
O
EPVS (see text)
H
H
9a
or
(t-BuO)2, PhCl
145 °C (67%)
DBU, MeCN
H
H
17
18
O
Ph
I
O
DLP, PhCl, 95 °C
then EPVS, (t-BuO)2,
145 °C (57%)
8
Scheme 5
O
From this point (18), completion of the route to the braso-
side core structure appeared to be readily achievable
merely by oxidative cleavage of the styryl group, formy-
lation of the lactone, and cyclisation (following the Vidari
precedent). The first two of these objectives were reduced
to practice without complication; thus, ozonolysis of alk-
ene 18 afforded aldehyde 19 that was immediately pro-
tected as its ethylene acetal 20 and formylated (→ 21).
Interestingly, the formyl lactone 21 existed almost entire-
O
O
X
O
O
O
H
H
DLP, PhH
80 °C, 6 h
H
H
EtOC(S)SK
X
X
8, X = I
9a, X = I (65%)
13, X = SC(S)OEt (66%)
9b, X = I (11%)
12, X = SC(S)OEt
O
1
O
O
I
ly in the enol form (according to H NMR analysis) and
O
O
O
H
R
H
NOE experiments confirmed the E-geometry of the hy-
droxymethylene group, suggesting the presence of a hy-
drogen bond between the enolic hydrogen and an acetal
oxygen (Scheme 6).11
DLP, PhH
80 °C, 6 h
H
H
R
R
I
I
16 (17%)
14, R = (CH2)6Cl
15 (37%)
All of our attempts to effect hydrolysis, epimerisation and
cyclisation of functionalised lactone 21 have, to date,
proven unsuccessful. Application of Vidari’s conditions
as well as a range of alternative acidic reagents showed
that acetal hydrolysis proceeded quickly but the so-
formed dialdehyde showed no propensity to cyclise, de-
composing instead to intractable materials. On the as-
sumption that the lactone present in 21 could be
responsible for problems with this reaction, substrates
22–24 were prepared and subjected to a similar investiga-
tion. In no case were we able to observe formation of a
hydroxy- or alkoxy-dihydropyran.12
Scheme 4
Introduction of the two formyl groups present in com-
pound 4 proved to be non-trivial. An initial attempt to dis-
place iodide in an SN2 manner, by treating compound 9a
with carbon-based nucleophiles, was not successful, pre-
sumably as a result of a hindered endo-approach to the bi-
cyclic structure; in most cases (e.g. with KCN in DMSO
at 80 °C) tricyclic lactone 178 was obtained as a major
product. More success was obtained under free-radical
olefination conditions and, although xanthate 13 could be
a useful precursor in this regard, the ease of access to
iodide 9a led us to adopt this compound for further study.
Whilst it remains disappointing that we have been unable
to complete this first-generation route to the brasoside
Synlett 2004, No. 15, 2788–2790 © Thieme Stuttgart · New York

2790
J. Robertson et al.
LETTER
(5) Curran, D. P.; Chang, C. T. J. Org. Chem. 1989, 54, 3140.
(6) Forbes, J. E.; Tailhan, C.; Zard, S. Z. Tetrahedron Lett.
1990, 31, 2565.
(7) Ollivier, C.; Bark, T.; Renaud, P. Synthesis 2000, 1598.
(8) (a) Doyle, M. P.; Dyatkin, A. B.; Kalinin, A. V.; Ruppar, D.
A.; Martin, S. F.; Spaller, M. R.; Liras, S. J. Am. Chem. Soc.
1995, 117, 11021. (b) Ali, S. M.; Chapleo, C. B.; Finch, M.
A. W.; Roberts, S. M.; Woolley, G. T.; Cave, R. J.; Newton,
R. F. J. Chem. Soc., Perkin Trans. 1 1980, 2093.
O
O
O
O
H
H
O3, CH2Cl2;
DMS
(HOCH2)2,
TsOH, THF
18
H
H
CHO
O
19 (99%)
20 (90%)
O
O
O
(9) Bertrand, F.; Quiclet-Sire, B.; Zard, S. Z. Angew. Chem. Int.
Ed. 1999, 38, 1943.
O
O
H
acidic
treatment
H
NaH,
H
(10) DLP (100 mg, 0.25 mmol) was added every 2 h (3×) to a
MeOCHO,
THF
H
solution of iodide 8 (750 mg, 2.98 mmol) in dry
OH
O
chlorobenzene (40 mL) at 95 °C. After 7 h in total, a solution
of ethyl 2-phenylvinyl sulfone15 (1.75 g, 8.93 mmol) in
chlorobenzene (5 mL) was added and the reaction mixture
was heated at 145 °C. tert-Butyl peroxide was added every 2
h (3 × 100 mL) and the solution was heated at 145 °C for 16
h. The cooled solution was poured directly onto the top of a
silica gel column, and the column was flushed with
petroleum ether to remove chlorobenzene. Elution was then
continued with a 4:1 mixture of petroleum ether–EtOAc to
afford lactone 18 (390 mg, 57%) as colourless crystals after
recrystallisation from Et2O. Rf = 0.42 (3:1 petroleum ether–
EtOAc); mp 76–81 °C. IR (KBr disc): nmax = 2971 (m), 1779
(s), 1493 (w), 1448 (w), 1421 (w), 1357 (w), 1299 (w), 1248
(w), 1176 (s), 1042 (s), 974 (s), 898 (w), 752 (m), 696 (m)
cm–1. 1H NMR (400 MHz, CDCl3): d = 1.57–1.66 (1 H, m),
1.93–2.01 (1 H, m), 2.04–2.11 (1 H, m), 2.23–2.31 (1 H, m),
2.42–2.50 (1 H, m), 2.44 (1 H, dd, J = 18.8, 1.4 Hz), 2.57 (1
H, br q, apparent J = 8.0 Hz), 2.76 (1 H, dd, J = 18.8, 8.8 Hz),
4.99–5.08 (1 H, m), 6.08 (1 H, dd, J = 15.6, 7.6 Hz), 6.44 (1
H, d, J = 15.6), 7.21–7.36 (5 H, m, Ph). 13C NMR (100.6
MHz, CDCl3): d = 32.0, 34.1, 45.7, 49.6, 77.2, 85.8, 126.1,
127.5, 128.6, 130.6, 130.8, 136.8, 176.9. MS (CI, NH3):
O
O
H
OR
21 (51%)
TBDPSO
MeO
CO2Me
CO2Me
CO2Et
OH
OH
O
OH
O
O
22
23
24
O
O
O
Scheme 6
core structure, this investigation generated an efficient
synthesis of bicyclic compounds of general structure 5
[R = H, (CH2)6Cl; X = I, SC(S)OEt, CH=CHPh, CHO,
CH(OCH2CH2O)]. This chemistry, represented by the
retrosynthetic disconnection shown in Scheme 7 is a prac-
tical and efficient alternative to, for example, Stork’s rad-
ical cyclisation and trapping with tert-butylisocyanide,13
and tandem organometallic approaches.14
+
m/z (%) = 246 (100) [MNH4 ], 229 (25) [MH+], 169 (10), 87
(20), 70 (20). HRMS (CI): m/z calcd for C15H20NO2
+
[MNH4 ]: 246.1494; found: 246.1491.
(11) Compounds analogous to lactone 21 that lack oxygen-
containing functionality in the cyclopentane ring (cf. the
dioxolane in 21) exist, in general, as tautomeric mixtures.
Spectroscopic data for 21: Rf = 0.35 (EtOAc); mp 81–84 °C.
IR (KBr disc): nmax = 2967 (s), 2883 (m), 2740 (s), 1718 (s),
1682 (s), 1618 (s), 1420 (s), 1373 (m), 1198 (s), 1128 (s),
1074 (s), 1014 (s), 966 (m), 943 (m) cm–1. 1H NMR (500
MHz, CDCl3): d = 1.51–1.63 (1 H, m), 1.68–1.80 (1 H, m),
2.01–2.10 (1 H, m), 2.24–2.39 (2 H, m), 3.27 (1 H, ddd, J =
9.1, 6.8, 2.3 Hz), 3.95–4.03 and 4.09–4.16 (4 H, m), 4.86–
4.93 (1 H, m) overlaying 4.89 (1 H, d, J = 5.2 Hz), 7.44 (1 H,
dd, J = 14.1, 2.3 Hz), 9.13 (1 H, d, J = 14.1 Hz). 13C NMR
(125.7 MHz, CDCl3): d = 27.0, 33.5, 39.1, 49.0, 65.0, 65.1,
82.3, 105.6, 105.9, 154.6, 172.8. HRMS (ES): m/z calcd for
C11H13O5 [M – H]: 225.0763; found: 225.0754. NOE
experiment: Irradiation of the enolic hydroxyl proton at d =
9.13 ppm led to enhancement of the resonances at d = 3.27
(-CHC=), 3.95–4.03 and 4.09–4.16 (-OC2H4O-), 4.89
[-CH(OR)2], and 7.44 (=CHOH) ppm; irradiation of the
olefinic proton at d = 7.44 ppm led only to an enhancement
at d = 9.13 (=CHOH) ppm.
CHO
H
2 steps
O
O
I
O
O
H
Scheme 7
Acknowledgement
We are very grateful to AstraZeneca for an Industrial CASE Award
to MM, to the EPSRC Mass Spectrometry Service, Swansea for
high-resolution mass spectra, and to both AstraZeneca and Pfizer
for unrestricted research donations.
References
(1) Li, Y. S.; Matsunaga, K.; Ishibashi, M.; Ohizumi, Y. J. Org.
Chem. 2001, 66, 2165.
(2) Garlaschelli, L.; Vidari, G.; Zanoni, G. Tetrahedron 1992,
48, 9495.
(3) (a) This step is related to the proposed biosynthesis of the
iridoids; for a review of iridoid chemistry, including
biosynthetic aspects, see: El-Naggar, L. J.; Beal, J. L. J. Nat.
Prod. 1980, 43, 649. (b) For a pioneering total synthesis
based on this mode of dialdehyde cyclisation see: Büchi, G.;
Gubler, B.; Schneider, R. S.; Wild, J. J. Am. Chem. Soc.
1967, 89, 2776.
(12) Conditions attempted: CSA, THF; HCl, THF; HCl, acetone;
TsOH, MeOH; BF3·OEt2, CD2Cl2; Dowex, MeOH; CSA,
CD3OD. For further details, see ref. 4.
(13) Stork, G.; Sher, P. M. J. Am. Chem. Soc. 1983, 105, 6765.
(14) For example: Larock, R. C.; Lee, N. H. J. Am. Chem. Soc.
1991, 113, 7815.
(15) Truce, W. E.; Goralski, C. T. J. Org. Chem. 1971, 36, 2536.
(4) Ménard, M. DPhil Thesis; Oxford: UK, 2003.
Synlett 2004, No. 15, 2788–2790 © Thieme Stuttgart · New York