Total synthesis of a diastereomer of the marine natural product clavosolide A

Article abstract of DOI:10.1039/b509757f

The first total synthesis of the reported structure of the sponge metabolite clavosolide A is described using a Prins cyclisation to assemble the tetrahydropyran core followed by manipulation of the side-chain, dimerisation and finally glycosidation. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b509757f

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Total synthesis of a diastereomer of the marine natural product
clavosolide A
Conor S Barry,a Nick Bushby,b Jonathan P. H. Charmant,a Jon D. Elsworth,a John R. Hardingb and
Christine L. Willis*a
Received (in Cambridge, UK) 11th July 2005, Accepted 8th August 2005
First published as an Advance Article on the web 15th September 2005
DOI: 10.1039/b509757f
protecting group proved to be the best for later stages of the
synthesis and alcohol 2 was prepared in 86% overall yield from
enol ether 1.
The first total synthesis of the reported structure of the sponge
metabolite clavosolide A is described using a Prins cyclisation
to assemble the tetrahydropyran core followed by manipulation
of the side-chain, dimerisation and finally glycosidation.
To install the required side chain which would become C-9 to
C-13 of clavosolide A, a propenylation step to an (E)-allylic
alcohol followed by a directed cyclopropanation were favoured.
First alcohol 2 was oxidised to the known aldehyde 32b using
Dess–Martin periodinane.4 Nozaki–Hiyama–Kishi5 coupling of 3
with E-1-bromo-1-propene in the presence of CrCl2 and catalytic
NiCl2 gave a 1 : 1 mixture of allylic alcohols 4 and 5 which were
separated by column chromatography. This approach gave
exclusively the (E)-alkenes as apparent from the characteristic
trans coupling constant (J 15 Hz) of the signals assigned to the
The clavosolides A–D are a family of unusual diolides which have
been isolated from extracts of the marine sponge Myriastra clavosa
collected in the Phillipines.1 The structure of clavosolide A
was determined by use of extensive spectroscopic studies
combined with molecular modelling. It is a symmetrical dimeric
16-membered ring dilactone assembled on a functionalised
tetrahydropyran core with a permethylated D-xylose moiety. The
macrocycle is further adorned by two cyclopropyl containing side-
chains which were assigned 9S,99S,10S,109S,11S,119S. Herein the
first total synthesis of the reported structure of clavosolide A is
described. The approach involved an efficient assembly of the
tetrahydropyran core I via a Prins cyclisation followed by
elaboration of the side-chain to II, then dimerisation and finally
glycosidation (Scheme 1).
1
olefinic protons in the H-NMR spectrum of each. The lack of
stereocontrol was advantageous as it gave access to both
diastereomers required for confirmation of the structure of the
target compound.
The final carbon–carbon bond forming reaction in the synthesis
of clavosolide A required a stereoselective cyclopropanation.
Charette reported that reaction of acyclic (E)-allylic alcohols with
Et2Zn and CH2I2 in CH2Cl2 gives preferentially syn-directed
cyclopropanation.6 Thus allylic alcohols 4 and 5 were treated
separately under Charette’s conditions giving 6 and 7 in excellent
yields and diastereoselectivities. However, since none of the
compounds 4–7 or simple derivatives were crystalline, at this stage
it was not possible to unequivocally assign the structures of the
diastereomers 6 and 7. Thus each was converted separately to the
analogous dimeric lactones 11 and 15 respectively.
Clavosolide A shares a common tetrahydropyran core with the
marine metabolite, polycavernoside A2 and recently we have
reported the enantioselective synthesis of tetrahydropyran 2
(Scheme 2).3 The key steps were a stereoselective crotyl transfer
reaction en route to the (S)-enol ether 1 followed by a TFA
mediated cyclisation to create the 3 new asymmetric centres in the
tetrahydropyran with complete stereocontrol in a single-pot
process. This approach was adopted for the synthesis of the
tetrahydropyran core of clavosolide A reported herein. The TIPS
Scheme 1 Retrosynthesis of the reported structure of clavosolide A.
aSchool of Chemistry, University of Bristol, Cantock’s Close, Bristol,
UK BS8 1TS. E-mail: chris.willis@bristol.ac.uk
bAstraZeneca UK Ltd., Mereside, Alderley Park, Macclesfield, UK
SK10 4TG
Scheme 2 Preparation of alcohols 6 and 7.
This journal is ß The Royal Society of Chemistry 2005
Chem. Commun., 2005, 5097–5099 | 5097

View Article Online
Scheme 3 Synthesis of dimer 11.
First methyl ester 6 was hydrolysed to hydroxy acid 8 in 87%
yield under mild conditions using TMSONa (Scheme 3). It was
then necessary to effect the dimerisation of 8 which we
anticipated may be achieved without resorting to a multi-step
strategy requiring protecting groups, as intramolecular lactonisa-
tion of 8 would not be favoured. Several macrolactonisation
procedures were investigated and the Corey–Nicolaou protocol
met with the greatest success.7 Treatment of hydroxy acid 8
with 2,29-bipyridyl disulfide and Ph3P gave 2-pyridine thiol ester
9 which, on refluxing in toluene, led to the required dimer 10
in 56% yield based upon recovered starting material 9.
High resolution electrospray MS confirmed that 10 was indeed a
dimer. However none of compounds 8, 9 and 10 were crystalline
and so we were still faced with the issue of the unequivocal
assignment of the stereochemistry at C-9(99), C10(109) and
C-11(119).
Scheme 4 Synthesis of the reported structure of clavosolide A.
To complete the total synthesis of clavosolide A it was necessary
to submit 9S-allylic alcohol 7 to the same sequence of reactions as
used in the preparation of diolide 11. Hydrolysis of ester 7 gave
acid 12, macrolactonisation followed by deprotection of the silyl
ether 14 gave diolide 15 in 45% overall yield from 7 (Scheme 4).
The stage was now set for the final transformation—the
introduction of two permethylated D-xylose moieties. Due to
concerns about the stability of the potentially labile cyclopropyl
and ester functionalities, a mild method was selected involving
thioglycoside trimethyl ether 17 as the glycosyl donor. Although
Deprotection of 10 using TBAF gave the required diol 11 as
a crystalline solid in 73% yield. X-Ray analysis of 11 revealed a
folded structure with the C2 symmetric dimer adopting a
conformation in which the cyclopropyl groups point
towards each other (Fig. 1).8 Thus this dimer is the
9R,99R,10R,109R,11R,119R diastereomer confirming that the
this approach would give
a mixture of [a,a], [a,b] and
[b,b]-anomers, the target molecule 18 would be formed directly
with no need for further manipulation of this highly functionalised
dimer.
secondary alcohol
4 had the 9R configuration and that
the stereoselective cyclopropanation of allylic alcohol 4 had indeed
proceeded as expected with syn-stereocontrol giving cyclopropanol
6 (Scheme 2). The proposed structure of clavosolide A, with the
9S,99S,10S,109S,11S,119S stereochemistry, had been assigned from
spectroscopic studies alongside molecular modelling which
revealed that the macrolide was relatively flat.1 The 1H-NMR
spectrum of dimer 11 is significantly different from that of the
natural product consistent with the distinct conformation of each
macrodiolide.9
The known triol 16 was prepared from D-xylose following
literature procedures.10 Permethylation of 16 using NaH and MeI
gave the novel thioglycoside 17 in 92% yield. Finally reaction of 17
with diol 15 using Nicolaou’s NBS-mediated glycosylation
protocol11 gave the expected mixture of [a,a], [a,b] and
[b,b]-anomers (17%, 26%, 10% respectively) which were
separated by column chromatography and their structures
1
assigned by H-NMR spectroscopy. In the required [b,b]-anomer
18, the signal assigned to 15-H (and 159-H) appeared as a doublet
(J 7.5Hz) at d 4.27 ppm consistent with an axial–axial relationship
of 15-H and 16-H whereas in the [a,a]-anomer, the analogous
signal appeared further downfield at d 5.06 ppm as a doublet with
the expected smaller coupling constant (J 3Hz) in accord with an
axial–equatorial coupling.
Synthetic dimer 18 was isolated as an oil and its 13C-NMR
spectrum correlated well with that of the natural product (Table 1).
However whilst there was a very good correlation of the 1H-NMR
spectra of clavosolide A and 18 in the region d 1.0–4.5 ppm, the
signals assigned to the cyclopropyl protons were clearly different.
Fig. 1 X-Ray structure of 11.
5098 | Chem. Commun., 2005, 5097–5099
This journal is ß The Royal Society of Chemistry 2005

View Article Online
Table 1 Comparison of the 1H and 13C NMR spectra of synthetic
dimer 18 and clavosolide A in CDCl3 (coupling constants reported to
the nearest 0.5 Hz)
have the 9S,99S,10R,109R,11R,119R configuration rather than the
proposed1 9S,99S,10S,109S,11S,119S.
In conclusion we have completed the first total synthesis of the
reported structure for clavosolide A and found that it is in fact a
diastereomer of the natural product. A revised structure of the
natural product is proposed which awaits further verification by
total synthesis.12 This is currently under investigation.
dC (ppm)
dH (ppm)
Clav A1a
18
Clav A1a
18
1
2
170.7
39.3
171.0
39.3
2.41 dd 17.5, 7
2.55 dd 17.5, 3
3.42 m
2.41 dd 17.5, 6.5
2.55 dd 17.5, 4
3.48 m
1.37 tq 10, 6.5
3.25 m
1.39 q 11.5
2.05 dd 11.5, 5
3.50 m
1.71 br d, 15
1.92 dt 15, 9
4.41 td 9, 1
0.70 tt 9, 4
0.60 m
1.00 d 5.5
0.21 dt 8, 4
0.60 m
0.96 d 6.5
4.27 d 7.5
2.96 dd 9, 7.5
3.10 t 9
We are grateful to AstraZeneca UK Ltd and the University of
Bristol for funding to CSB and the EPSRC for funding to JDE.
3
4
5
6
77.0a
42.6
83.1
40.8
77.2
42.6
83.2
40.8
1.38 m
3.25 t 11.5
1.37 q 11.5
2.05 dd 11.5, 5
3.42 m
1.66 br d 15
1.87 dt 15, 9
4.41 br t 9
0.72 tt 9, 5
0.83 m
0.96 d 6.5
0.22 dt 8, 5
0.33 dt 8, 5
0.96 d 6.5
4.27 d 8
2.96 t 8
3.12 t 8
3.25 td 8, 5
3.10 dd 11, 8
3.96 dd 11, 5
3.57 s
Notes and references
1 (a) M. R. Rao and D. J. Faulkner, J. Nat. Prod., 2002, 65, 386; (b)
K. L. Erickson, K. R. Gustafson, L. K. Pannell, J. A. Beutler and
M. R. Boyd, J. Nat. Prod., 2002, 65, 1303.
7
8
74.8
41.1
74.8
41.8
2 For total syntheses of polycavernoside A see: (a) A. Fujiwara, M. Murai,
Y.-M Yamashita and T. Yasumoto, J. Am. Chem. Soc., 1998, 120,
10770; (b) J. D. White, P. R. Blakemore, C. C. Browder, J. Hong,
C. M. Lincoln, P. A. Nagornyy, L. A. Robarge and D. J. Wardrop,
J. Am. Chem. Soc., 2001, 123, 8593; (c) L. A. Paquette, L. Barriault,
D. Pissarnitski and J. N. Johnston, J. Am. Chem. Soc., 2000, 122, 619.
3 C. S. Barry, N. Bushby, J. R. Harding and C. L. Willis, Org. Lett., 2005,
7, 2683 and references cited therein.
9
77.1a
24.8
12.0
18.6
11.0
77.0
24.9
11.0
18.4
12.4
10
11
12
13
14
15
16
17
18
19
12.7
105.4
83.8
85.6
79.4
63.2
12.6
105.5
83.8
85.6
79.4
63.2
4 D. B. Dess and J. C. Martin, J. Am. Chem. Soc., 1991, 113, 7277.
5 (a) K. Takai, M. Tagashira, T. Kuroda, K. Oshima, K. Utimoto and
H. Nozaki, J. Am. Chem. Soc., 1986, 108, 6048; (b) H. Jin, J. Uenishi,
W. J. Christ and Y. Kishi, J. Am. Chem. Soc., 1986, 108, 5644.
6 A. B. Charette and H. Lebel, J. Org. Chem., 1995, 60, 2966.
7 E. J. Corey and K. C. Nicolaou, J. Am. Chem. Soc., 1974, 96, 5614.
8 Crystal data for 11: C28H48O10, M 5 544.66, monoclinic, a 5 20.351(4),
3.25 m
3.09 dd 11.5, 10
3.96 dd 11.5, 5
3.58 s
3.61 s
3.47 s
20
21
22
a
60.7
60.8
58.5
60.8
60.8
58.8
3
3.62 s
3.47 s
˚
b 5 8.8020(18), c 5 8.8390(18) A, b 5 100.07(3)u, V 5 1558.9(6) s ,
T 5 100(2), space group C2, Z 5 2, m 5 0.087 mm21, Rint 5 0.0249 (for
12594 measured reflections), R1 5 0.0589 [for 1896 unique reflections
with I.2s(I)], wR2 5 0.1447 (for all 1919 unique 10 reflections). The
absolute configuration of the molecule could not be determined from
the X-ray. However, the absolute configuration of tetrahydropyran 2,
used for the synthesis of 11, has been assigned previously3 by
comparison with the literature2b. CCDC 280355. See http://dx.doi.org/
10.1039/b509757f for crystallographic data in CIF or other electronic
format.
Assigment may be reversed.
The chemical shifts of the signals (d 0.22 and 0.33 ppm ) assigned
to 13(139)-H2 of clavosolide A were much closer than in the case of
the synthetic material (d 0.21 and 0.60 ppm). In addition 11(119)-H
in clavosolide A resonate 0.23 ppm downfield relative to the
corresponding signal in 18. Hence we propose that 18 is a
diastereomer of the natural product.
9 Diolide 11, mp: 175–177 uC; [a]D + 57.4 (c 1.5, CHCl3); dH (400 MHz,
CDCl3) 0.23 (2H, dt, J 8.5, 5, 13-HH and 139-HH), 0.57 (2H, tt, J 8.5, 5,
10-H and 109-H), 0.67 (2H, m, 11-H and 119-H), 0.88 (2H, dt, J 8.5, 4.5,
13-HH and 139-HH), 0.98 (6H, d, J 6, 12-H3 and 129-H3), 1.03 (6H, d,
J 6.5, 14-H3 and 149-H3), 1.23 (2H, m, 4-H and 49-H), 1.28 (2H, dt, J 12,
11, 6-Hax and 69-Hax), 1.48 (2H, d, J 5.5, 2 6 -OH), 1.65-1.78 (4H, m,
8-H2 and 89-H2), 1.95 (2H, ddd, J 12.5, 4.5, 2, 6-Heq and 69-Heq), 2.18
(2H, dd, J 11.5, 0.5, 2-HH and 29-HH), 2.62 (2H, dd, J 12, 1.5, 2-HH
and 29-HH), 3.12 (2H, td, J 11, 1.5, 3-H and 39-H), 3.33-3.43 (4H, m,
5-H and 59-H, 7-H and 79-H), 4.45 (2H, ddd, J 11.0, 8.5, 3, 9-H and
99-H); dC (100 MHz, CDCl3) 10.1 (C-11 and C-119), 12.4 (C-13
and C-139), 13.2 (C-14 and C-149), 18.6 (C-12 and C-129), 24.6 (C-10
and C-109), 40.2 (C-2 and C-29), 40.8 (C-8 and C-89), 41.6 (C-6 and
C-69), 44.0 (C-4 and C-49), 70.5 (C-5 and C-59), 72.9 (C-9 and C-99), 73.2
(C-7 and C-79), 79.2 (C-3 and C-39), 169.4 (C-1 and C-19); m/z (ESI):
531.2923 C28H44O8 requires 531.2928.
10 R. Lopez and A. Fernandez-Mayoralas, J. Org. Chem., 1994, 59, 737.
11 K. C. Nicolaou, S. P. Seitz and D. P. Papahatjis, J. Am. Chem. Soc.,
1983, 105, 2430.
12 Whilst clavosolide A is most likely derived from D-xylose and hence has
the structure and absolute configuration proposed herein, at this stage it
cannot be ruled out unequivocally that the natural product is the
enantiomer of that proposed (i.e. derived from L-xylose). Once synthetic
material is available, comparison of its optical rotation with that of the
natural product, [a]D 248.5 (c 1, CHCl3),1a will verify the absolute
configuration.
The original structure elucidation of clavosolide A was
conducted on a small quantity of material (extracted and purified
from the sponge M. clavosa) and was a challenging problem
considering the density of functionality within the molecule.1 The
trans stereochemistry about the cyclopropyl group was based on
NOE data and coupling constants (J10,11 5 Hz); as expected the
synthetic dimer 18 prepared from the E-allylic alcohol 5 gave a
similar coupling constant, J10,11 4 Hz. The absolute stereo-
chemistry of clavosolide A had been assigned on the assumption
that the xylose moiety has the usual D-configuration. The close
correlation of the 1H-NMR data of the natural product and 18 in
the region d 1.0–4.5 ppm as well as the 13C-NMR data are in
accord with the proposed relative stereochemistry of the sugar
moiety and macrocycle. Hence on the basis of the synthetic and
spectroscopic studies reported herein combined with further
analysis of molecular models, we propose that the structure of
the tetrahydropyran core of clavosolide A is as reported1
(i.e. 3S,39S,4R,49R,5S,59S,7S,79S) but the cyclopropyl side-chains
This journal is ß The Royal Society of Chemistry 2005
Chem. Commun., 2005, 5097–5099 | 5099