Synthesis of a model DEF-ring core of hexacyclinic acid

Article abstract of DOI:10.1039/b303706a

The first synthesis of a DEF-ring system of hexacyclinic acid is reported. The key steps being an intramolecular Pd(0) π-allyl substitution reaction, followed by a transannular iodocyclisation with acetyl hypoiodite.

Full text of DOI:10.1039/b303706a

Synthesis of a model DEF-ring core of hexacyclinic acid†
Paul A. Clarke,*a Matthew Grist,a Mark Ebdenb and Claire Wilsona
a School of Chemistry, University of Nottingham, University Park, Nottingham, Notts, UK NG7 2RD.
E-mail: paul.clarke@nottingham.ac.uk; Fax: +44 115 9513564; Tel: +44 115 9513566
b Medicinal Chemistry, AstraZeneca R&D Charnwood, Bakewell Road, Loughborough, Leics, UK LE11
5RH
Received (in Cambridge, UK) 3rd April 2003, Accepted 13th May 2003
First published as an Advance Article on the web 3rd June 2003
The first synthesis of a DEF-ring system of hexacyclinic acid
is reported. The key steps being an intramolecular Pd(0) p-
allyl substitution reaction, followed by a transannular
iodocyclisation with acetyl hypoiodite.
of the b-ketoester in one step from an aldol reaction of a b-
ketoester with an appropriate aldehyde and ii) obviate the
potential problem of the relative stereochemistry between the
C4 methyl group and the C5 hydroxyl group inherent in the
Mukaiyama aldol reaction of g-substituted b-ketoesters.‡
Our DEF-ring synthesis (Scheme 2) started from nerol, which
was acylated with acetic anhydride in pyridine with catalytic
DMAP and then epoxidised at the most electron rich double
bond with mCPBA; this was followed by oxidative cleavage of
the epoxide with periodic acid to give aldehyde 5.3 Mukaiyama
aldol reaction of the bis-TMS enol ether of tert-butyl acet-
oacetate with 5 followed by TBS protection of the resulting
alcohol yielded cyclisation precursor 6.
Recently the structure of hexacyclinic acid 1, a new polyketide
natural product, was reported (Fig. 1).1 This compound was
isolated from Streptomyces cellulosa subsp. griserubiginosus
(strain S1013) and shown to have some cytotoxic activity when
tested in 3 cell lines.1 It was due to its complex hexacyclic ring
system with its unusual ‘knotted’ DEF-rings, coupled with the
biological potential, that we became interested in developing
chemistry for the synthesis of the hexacyclinic acid ring
systems. In this communication we disclose the first successful
synthesis of a DEF-ring system of hexacyclinic acid and report
an interesting solvent effect on the key iodocyclisation
reaction.
Scheme 2 Reagents and conditions: i) Ac2O, py, DMAP, 100%; ii)
mCPBA, CH2Cl2, 89%; iii) HIO4, THF, H2O, 76%; iv) tBuO-
C(OTMS)NCHC(OTMS)NCH2, TiCl4, CH2Cl2, 278 °C, 79%; v) TBSOTf,
py, 230 °C, 78%.
Formation of the 9-membered carbocycle 7 was achieved
using a Pd(0) catalysed intramolecular p-allyl substitution
reaction.4 We found that the efficiency of this reaction was
dependent upon the additional phosphine ligands employed
(Table 1), with the optimum being 2 equiv. of dppe, which
provided 7 in a 61% isolated yield.5 The relative ster-
eochemistry of the substituents on the 9-membered ring was
determined as cis by gradient nOe experiments. Interestingly,
this also highlighted that the 9-membered ring existed in a
chair–boat conformation (Fig. 2).
Fig. 1 Hexacyclinic acid 1.
We envisaged that the DEF-rings of hexacyclinic acid would
be accessible from the nucleophilic addition to an oxonium ion
intermediate 2, which in turn would arise from the transannular
cyclisation of the ketone carbonyl of a b-ketoester onto a cation
equivalent generated from a double bond containing 9-mem-
bered ring 3. Carbocycle 3 was to be obtained by a cyclisation
of an acyclic precursor such as 4 (Scheme 1).
As can be seen from Table 1, two other products could also
be formed under the reaction conditions, depending on the
Table 1 Optimisation of the Pd(0) p-allyl cyclisation reaction
Scheme 1 Retrosynthesis of the DEF rings.
Entry
Ligand
Yield (%)a
Ratio 7 : 8 : 9b
Before, however, we embarked upon an asymmetric total
synthesis of hexacyclinic acid we desired to test our DEF-ring
forming hypothesis. To this end we elected to synthesise a
hexacyclinic acid model system without the methyl group at C4
in a racemic series. This would expedite the synthesis of our key
9-membered ring precursor as it would i) allow the installation
1
2
3
4
5
Ph3P
dppm
dppe
dppp
dppf
91
97
93c
94
90
0 : 1 : 4
0 : 1 : 8
2 : 0 : 1
1 : 1 : 2
3 : 2 : 8
a Combined yield. All compounds co-ran and exhaustive flash column
chromatography provided only small amounts of purified products. b By 1H
NMR (400 MHz). c Isolation of 7 (61%) and 9 (32%) proved possible by
flash column chromatography on AgNO3 impregnated silica gel.5,6
† Electronic supplementary information (ESI) available: experimental
procedures and characterisation data. See http://www.rsc.org/suppdata/cc/
b3/b303706a/
1560
CHEM. COMMUN., 2003, 1560–1561
This journal is © The Royal Society of Chemistry 2003

1
from CHCl3 to AcOH. However, H NMR (400 MHz) and
gradient nOe experiments in d3-acetic acid showed that not only
was there no enol form detectable, but the 9-membered ring
remained in the same chair–boat conformation as indicated by
the 1H NMR experiments conducted previously in CDCl3. We
rationalise that, as neither of the reaction pathways occurs via
the detected chair–boat conformation, and iodonium ion
formation is reversible, the change in the reaction pathway must
therefore be due to a change in the reacting iodonium ion (either
a or b), brought about by solvent effects.∑ At present we are
unable to provide any further explanation of this remarkable
solvent effect, but we are investigating it further.
Treatment of 11 with aqueous HF in MeCN resulted in the
removal of both the TBS and acetate groups to give 12 in 97%
yield.5 The structure of 12 was confirmed by single crystal X-
ray analysis,** which showed that cyclisation had occurred
through the ketone oxygen of the b-ketoester onto the b-
iodonium ion. Removal of the tert-butyl ester and lactonisation
was achieved by reaction of 12 with TFA in CH2Cl2, which
furnished a model DEF-ring system of hexacyclinic acid 13 in
a yield of 63% (Scheme 4).5
In summary, we have developed a new route for the synthesis
of the DEF-ring system of hexacyclinic acid utilising a
transannular iodocyclisation approach. Application of this route
to a total synthesis of hexacyclinic acid is underway.
We thank the EPSRC and AstraZeneca for studentship
funding under the CASE Award for New Academics Scheme,
AstraZeneca for an unrestricted research support grant and the
EPSRC National Mass Spectrometry Service, Swansea for
accurate mass determination. We also thank Dr Adrienne Davis
(Nottingham) for NMR technical advice and support and Dr
Hitesh Sanganee (AstraZeneca) for helpful discussions.
Fig. 2 Conformation of 9-membered ring 7.
diphosphine ligands used: the 7-membered ring 8, arising from
attack of the acetoacetate anion on the other side of the p-allyl
complex,§ and the diene 9, resulting from elimination of the p-
allyl complex.
With carbocycle 7 in hand we next studied the key
iodocyclisation reaction. We decided to use AcOI for the
formation of the iodonium ion,7 as previous studies in the group
showed that use of I2 resulted in the deprotection of pendant
silyl ethers and competing modes of cyclisation. We were also
optimistic that the acetate ion generated would be sufficiently
nucleophilic to add to the oxonium ion and thus introduce the
masked hemi-ketal needed for a synthesis of the hexacyclinic
acid DEF-ring system.
When 7 was treated with AcOI in CHCl3 a single product was
formed, which was identified as iodolactone 10 (Scheme 3).
Iodolactone 10 is obviously generated by the cyclisation of the
ester carbonyl onto the a-iodonium ion via a chair–chair
conformation of the 9-membered ring (Fig. 3), with loss of a
tert-butyl cation as isobutene.
Scheme 3 Reagents and conditions: i) AcOI, CHCl3, 49%.
Notes and references
Gratifyingly, treatment of 7 with AcOI in AcOH resulted in
the cyclisation of the ketone carbonyl onto the desired b-
iodonium ion via a boat–boat conformation (Fig. 3), and led to
the formation of hexacyclinic acid DF-ring unit 11 (Scheme 4),
where the intermediate oxonium ion was trapped by the AcOH
solvent.5¶ We first thought that the differential modes of
cyclisation were due to either formation of some of the enol
form of the b-ketoester or due to a change in the conformation
of the 9-membered ring brought about by the change in solvent
‡ Formation of the syn or anti aldol product is controlled by the geometry
of the silyl enolether. At present there are no reliable methods for the
exclusive formation of the desired Z-silyl enolether.2 Therefore, while this
strategy may be employed for the synthesis of the actual DEF-ring system,
separation of the syn and anti aldol product diastereomers would be
required.
§ It is a feature of these reactions that when differently substituted p-allyl
systems are employed, it is the larger ring system which is preferred over the
smaller one, especially when there is an increase in the steric encumbrance
of the system. See reference 4a.
¶ Studies using d3-acetic acid as solvent showed that d3-acetate was
introduced exclusively into the cyclised product.
∑ These other conformations are presumably within about 2 kcal mol21 of
the ground state chair–boat conformation.
** C15H25IO5, M = 412.25, triclinic, a = 6.5249(5), b = 11.3828(9), c =
13.3018(10) Å, a = 65.415(1), b = 77.402(1), g = 81.897(1)°, U =
875.35(12) Å3, T = 150 K, space group P-1 (no. 2), Z = 2, m(Mo-Ka) =
1.846 mm21, 7666 reflections measured, 3876 unique (Rint = 0.028) which
were used in all calculations. The final wR(F2) was 0.077 for all data, R1(F)
was 0.035 for 3298 observed data where I > 2s(I). CCDC 207876. See
http://www.rsc.org/suppdata/cc/b3/b303706a/ for crystallographic data in
.cif or other electronic format.
1 R. Hofs, M. Walker and A. Zeeck, Angew. Chem., Int. Ed., 2000, 39,
3258.
2 T. H. Chan and P. Brownbridge, Tetrahedron, 1981, 37, 387.
3 For use of this route in the synthesis of a geraniol based aldehyde see: K.
Tago, M. Arai and H. Kogen, J. Chem. Soc., Perkin Trans. 1, 2000,
2073.
Fig. 3 Proposed reactive conformations of 7.
4 (a) B. M. Trost and T. R. Verhoeven, J. Am. Chem. Soc., 1979, 101, 1595;
(b) For a review see: B. M. Trost, Angew. Chem., Int. Ed. Engl., 1989, 28,
1173.
5 For experimental procedures and characterisation data refer to supporting
information.
6 For a review of the use of AgNO3 in chromatography see: C. M. Williams
and L. N. Mander, Tetrahedron, 2001, 57, 425.
7 R. C. Cambie, P. S. Rutledge, G. M. Stewart, P. D. Woodgate and S. D.
Woodgate, Aust. J. Chem., 1984, 37, 1689.
Scheme 4 Reagents and conditions: i) AcOI, AcOH, 61%; ii) 40% HF (aq.),
MeCN, 97%; iii) TFA, CH2Cl2, 63%.
CHEM. COMMUN., 2003, 1560–1561
1561