Free radical alkylation of tri-o-benzoyl-6-exo-bromo-levoglucosan: Approaches to the synthesis of okadaic acid

Article abstract of DOI:10.1016/S0040-4039(01)02181-5

Tri-O-benzoyl-6-exo-bromo-levoglucosan undergoes stereoselective free radical alkylation to give 6-exo-substituted-1,6-anhydrosugars with six contiguous chiral centres.

Full text of DOI:10.1016/S0040-4039(01)02181-5

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 511–514
Free radical alkylation of
tri-O-benzoyl-6-exo-bromo-levoglucosan: approaches to the
synthesis of okadaic acid
David R. Kelly* and Jassem G. Mahdi
Department of Chemistry, Cardiff University, PO Box 912, Cardiff CF10 3TB, UK
Received 25 September 2001; revised 8 November 2001; accepted 14 November 2001
Abstract—Tri-O-benzoyl-6-exo-bromo-levoglucosan undergoes stereoselective free radical alkylation to give 6-exo-substituted-1,6-
anhydrosugars with six contiguous chiral centres. © 2002 Elsevier Science Ltd. All rights reserved.
Total syntheses of okadaic acid 1 have been reported by
Isobe,1 Forsyth2 and Ley.3 In this letter we describe
model studies related to a strategy for the synthesis of
the C-19 to C-30 segment. This segment encompasses the
most densely functionalised region (C-22 to C-27) which
contains five asymmetric centres. Disconnection of
okadaic acid 1 between C-18 and C-19 yields nucleophilic
2 and electrophilic 3 fragments which would be viable
synthons if suitably protected and/or activated. A syn-
thetic equivalent of the majority of fragment 3 can in
principle be constructed by using the two major modes
of reactivity of allylstannanes. The C-27 centre is formed
by free radical4 allylation using 6 on the exo-face of the
1,6-anhydrosugar 5 and the C-22 centre by electrophilic
cleavage using 4 of the 1,6-anhydro ring with inversion
of configuration. The order of these steps is obligatory
because the stereoselectivity of allylation at C-27 is
dependant on the presence of the 1,6-anhydro linkage.5
The Isobe1 and Forsyth2 syntheses, create the C27ꢀ28
bond by addition of an anion to a C-27 aldehyde.
18
O
OH
H
H
H
H
O
O
19
HO
O
28
O
O
30
OH
22
21
27
H
OH
O
O
H
H
OH
1 Okadaic acid
OH
O
18
H
O
HO
H
O
19
HO
O
28
30
27
O
OH
OH
H
O
O
OH
H
H
H
2
3
OH
OR'
28
30
R'O
OR'
R3M
19
OR''
SnBu3
22
H
O
21
27
O
Br
4 M = Sn or Si
5
6
* Corresponding author. Tel.: +44 (0)29 20874063; fax: +44 (0)29 20874030; e-mail: kellydr@cardiff.ac.uk
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)02181-5

512
D. R. Kelly, J. G. Mahdi / Tetrahedron Letters 43 (2002) 511–514
In both cases the C-27 hydroxyl group was formed with
predominantly the wrong stereochemistry, which was
remedied by oxidation to the ketone and reduction. In
the Ley synthesis,3 construction of the C26ꢀ27 bond
yields a comparable ketone which is reduced as in the
other two syntheses. Thus the proposed free radical
alkylation potentially represents a distinct improvement,
over previous syntheses. The electrophilic allylation
parallels comparable steps in the Isobe and Forsyth
syntheses, which construct the same bond using allyl-
trimethylsilane in a Ferrier rearrangement or by dis-
placement of an anomeric methoxy group, respectively.
glucopyranoside (tri-O-benzoyl-levoglucosan) 7.6 The
stereochemistry required for okadaic acid synthon 5 is
b-D
-altro- or -ido-; however, tri-O-benzoyl-mannosan7
and -galactosan8 also undergo photobromination exclu-
sively at the 6-exo-position, which provides some reas-
surance that other stereoisomers will react similarly.
Free radical alkylation of 8 with allyltri-n-butylstan-
nane 10 gave the desired exo-adduct 9a9 in fair yield.
The stereochemistry at C-6 was inferred from the chem-
ical shift and coupling pattern of 6-H. The 6-endo-pro-
ton in tri-O-benzoyl-levoglucosan 7 (l 4.4) is downfield
relative to the 6-exo-proton (l 4.0, 3J5,6=6 Hz) and has
no appreciable coupling to 5-H. Similarly the 6-protons
in the adducts 9a–f have chemical shifts in the range l
4.5–4.8 and no measurable coupling to 5-H.
Model studies commenced with the photobromination
of readily available tri-O-benzoyl-1,6-anhydro-b-D-
Table 1. Free radical alkylation of tri-O-benzoyl-6-exo-bromo-levoglucosan 8
Br
9
R
a R = Allyl
O
OBz
O
O
OBz
O
O
Reagents 10-15
Br2, CCl4,
b R = 2-Methylallyl
c R = (2-Benzyloxymethyl)allyl
d R = Methyl β-acrylyl
e R = Ethyl β-acrylyl
f R = 2-Chloroallyl
OBz
O
hν, ∆, (85%)
7
8
OBz OBz
OBz OBz
OBz OBz
Entry
1
Reagent
Reagent prep. ref.
10
Method
Product
Yield (%)
47
Mp (°C)
A
9a
120–121
Bu3Sn
10
11
2
3
10
11
A, B
9b
48
46
121–123
Bu3Sn
A, B
9c
89–93
BnO
12
Bu3Sn
4
12
A
9d
9e
35, 24
31, 23
trans,cis
85–87, oil
77–79, oil
CO2R'
13
Bu3Sn
R' = Me or Et
5
6
13
13
C
C
9a
7
67
3
As above
197
TolSO2
TolSO2
14
Cl
9f
7
71
4
119–121
As above
15
Reagents and conditions: Method A: 8 (1 equiv.), stannane 10–13 (2.5 equiv.), AIBN (0.16 equiv. every 8 h), PhMe, 96–100°C, 36 h,
hexane/CH3CN partition,14 method B if applicable, column chromatography, crystallisation; Method B: BzCl (4 equiv.), pyr. (15 equiv.), CH2Cl2;
Method C 8 (1 equiv.), sulphone 14 and 15 (3 equiv.), PhMe, 96–100°C, Bu3SnH by syringe pump over 20 h, 7 h, column chromatography,
crystallisation. All yields refer to analytically pure materials.

D. R. Kelly, J. G. Mahdi / Tetrahedron Letters 43 (2002) 511–514
513
HO
BnO
BnO
O
O
O
O
OH
O
BnO
16
17
18
OBz OBz
19
O
SnBu3
O
OBz
O
10
O
OBz
O
O
OBz
O
a BF3.MeOH, MeOH, ∆
N
MeO
BF3.OEt2, CH3CN, ∆
b BzCl, pyr., CH2Cl2
(37%)
H
(28%)
9b
20
21
BzO BzO
BzO BzO
BzO BzO
To our surprise, free radical alkylation with the 2-
methylallylstannane 11 gave a complex mixture of the
anticipated adduct 9b and debenzoylated adducts and
this also occurred with the benzyloxymethylallylstan-
nane 12, but to a much lesser extent. In one run, a
small amount of a crystalline product 16 (mp 144–
147°C) was isolated and the structure determined by
X-ray crystallography. Although this material was part
of a mixture of ill defined composition, it is the product
expected on the basis of a nucleophilic cleavage mecha-
nism.15 Rebenzoylation of the mixture (Table 1, method
B) before final purification afforded adequate yields of
the desired adducts 9b and 9c. We have been unable to
find any precedents for the cleavage, but plausibly, the
high nucleophilicity of the stannanes 11 and 12, cataly-
sis by tri-n-butyltin bromide and high temperatures
suffice to promote this otherwise unfavourable reaction.
Both the stannane and tri-n-butyltin bromide seem to
be essential for the cleavage. Refluxing tri-O-benzoyl-
levoglucosan 7 with 2-methylallylstannane 11 for 48 h
at 96°C occurred without change. However, methyl
benzoate under the same conditions for 37 h gave an
intractable mixture, but was unchanged when refluxed
(80°C, C6D6) with tri-n-butyltin bromide for 5 days. We
were not able to identify conclusively the anticipated
by-product 17 or the conjugated analogue 18 in any of
the reactions.
ion, participation of the 2-benzoate group would likely
lead to substitution with overall retention of configura-
tion at C-1, rather than the desired inversion (cf. 3).16
Moreover the discovery of ester cleavage during free
radical alkylation did not augur well for the proposed
reaction. Tri-O-benzoyl-levoglucosan 7 was unchanged
upon treatment with the boron trifluoride methanol
adduct in methanol or with boron trifluoride etherate in
acetonitrile and allyltri-n-butylstannane 10 at 0°C to
room temperature, whereas at reflux the allylmethyl
derivative 9b, gave the addition products 20 and 21.
Attempts to functionalise stereoselectively the side
chain by epoxidation were unrewarding. Treatment of
the alkenes 9a–c with mCPBA in methylene chloride at
−10 or −78°C gave mixtures of diastereomeric epoxides
in the range 50:50 to 42:58 (75–78% yield).
Benzylated 1,6-anhydrosugars have been cleaved with
allyltrimethylsilane17,18 and acyloxy-19 or methoxy-20
allylsilanes. We envisaged that the greater nucleophilic-
ity of allylstannanes would promote the reaction, how-
ever, despite the use of a wide range of conditions we
were not able to achieve cleavage in satisfactory yields.
However, tri-O-benzyl-levoglucosan was cleaved by
allyltrimethylsilane and boron trifluoride etherate in
acetonitrile to give the a-C-glycoside 19 as reported
previously by Kishi (17%).17
Conclusions: Free radical alkylation of 6-bromolevoglu-
cosan tribenzoate is highly stereoselective and delivers a
product with six contiguous chiral centres. Benzoyl
ester cleavage by the more nucleophilic allylstannanes
11 and 12 is an undesirable side reaction. Plausibly,
electrophilic cleavage of the 1,6-anhydro linkage of
substituted levo-altrosans or -idosans by allylstannanes
may be more facile than with -glucosans and future
work will be directed towards these compounds.
The cis-b-stannyl acrylates 13a and 13b yielded 70:30,
trans,cis crude mixtures of the adducts 9d and 9e. When
the reaction was run in toluene-d8, and monitored
directly by 1H NMR, cis:trans isomerisation of the
reagents 13a and 13b occurred more rapidly than alky-
lation and the equilibrium value was circa 70:30,
trans:cis. Isomerisation can presumably be attributed to
reversible addition of the tri-n-butyltin radical. The
sulfones 14 and 15 gave reaction mixtures which were
easier to purify, albeit with some concomitant reduction
of tri-O-benzoyl-6-exo-bromo-levoglucosan 8, but there
was no detectable reduction of the vinyl chloride
substituent.
References
1. (a) Isobe, M.; Ichikawa, Y.; Bail, D.; Maski, H.; Goto, T.
Tetrahedron 1987, 43, 4767–4776; (b) For related
advanced intermediates, see: Dankwardt, J. W.;
Dankwardt, S. M.; Schlessinger, R. H. Tetrahedron Lett.
1998, 39, 4983–4986.
Cleavage of the 1,6-anhydro-linkage of a tri-O-benzoyl-
levoglucosan derivative was anticipated to be challeng-
ing. The 2-O-substituent lies anti to the bond
undergoing cleavage and, with a fully formed oxonium

514
2. Forsyth, C. J.; Sabes
D. R. Kelly, J. G. Mahdi / Tetrahedron Letters 43 (2002) 511–514
, S. F.; Urbanek, R. A. J. Am.
THF gave ratios of the di- and monobenzyl ethers 23 and
24 in the range 34:66 to 12:88.21 The best yield of the
monobenzyl ether 24 (68%) was obtained with potassium
t-butoxide and 18-crown-6 in THF on a 20 g scale.
Chem. Soc. 1997, 119, 8381–8382.
3. Ley, S. V.; Humphries, A. C.; Eick, H.; Downham, R.;
Ross, A. R.; Boyce, R. J.; Pavey, J. B. J.; Pietruszka, J. J.
Chem. Soc., Perkin Trans. 1 1998, 3907–3911.
4. For previous free radical reactions of sugars, see: Kelly,
D. R.; Picton, M. J. Chem. Soc., Perkin Trans. 1 2000,
1571–1586.
Cl
BnO
BnO
BnO
a
b
+
Cl
BnO
Cl
Bu3Sn
22
23
24
12
Reagents and conditions: (a) BnOH (0.9 equiv.), KOtBu (1
equiv.), 18-crown-6, THF, 0°C to rt, 12 h, chromatogra-
phy (23 12%, 24 68%); (b) Mg, THF, reflux, nBu3SnCl, 24
(Barbier conditions), 12 h (12, 85%).
5. Ohrui, H.; Nishida, Y.; Itoh, H.; Meguro, H. J. Org.
Chem. 1991, 56, 1726–1731.
6. (a) Ferrier, R. J.; Furneaux, R. H. Aust. J. Chem. 1980,
33, 1025–1036; (b) Coleman, G. H. Methods Carbohydr.
Chem. 1963, 2, 397–399.
12. Kelly, D. R.; Baldwin, J. E. J. Chem. Soc., Chem. Com-
7. Hori, H.; Nishida, Y.; Ohrui, H.; Meguro, H. J. Carbo-
mun. 1985, 682–684.
hydr. Chem. 1990, 9, 601–618.
8. Ohrui, H.; Nishida, Y.; Higuchi, H.; Hori, H.; Meguro,
H. Can. J. Chem. 1987, 65, 1145–1153.
9. A similar reaction with tri-O-acetyl-6-exo-bromo-
levoglucosan has been reported. See: Lopez, J. C.;
Alonso, R.; Fraser-Reid, B. J. Am. Chem. Soc. 1989, 111,
6471–6473.
10. Prepared by Barbier reaction of the allyl chloride, tri-n-
butyltin chloride and magnesium in refluxing THF.
11. We were surprised to find that no stannane bearing this
substituent has been reported, although analogous com-
pounds are known, e.g. the parent alcohol. See: (a)
Baldwin, J. E.; Fieldhouse, R.; Russell, A. T. Tetrahedron
Lett. 1993, 34, 5491–5494; Acetate: (b) Trost, B. M.;
Bonk, P. J. J. Am. Chem. Soc. 1985, 107, 8277–8279;
tBuPh2Si ether: (c) Davidson, B. S.; Plavcan, K. A.;
Meinwald, J. J. Org. Chem. 1990, 55, 3912–3917; Alky-
lated derivatives: (d) Nishigaichi, Y.; Kuramoto, H.;
Takuwa, A. Tetrahedron Lett. 1995, 36, 2253–3356; (e)
Williams, D. R.; Brooks, D. A.; Meyer, K. G.; Clark, M.
P. Tetrahedron Lett. 1998, 39, 7251–7254.
13. (a) Baldwin, J. E.; Adlington, R. M.; Lowe, C.; O’Neil, I.
A.; Sanders, G. A.; Schofield, C. J.; Sweeney, J. B. J.
Chem. Soc., Chem. Commun. 1988, 1030–1031. We thank
the authors for providing experimental conditions. For
related uses of this and comparable reagents, see: (b)
Chatgilialoglu, C.; Alberti, A.; Ballestri, M.; Macciantelli,
D.; Curran, D. P. Tetrahedron Lett. 1996, 37, 6391–6394.
14. Berge, J. M.; Roberts, S. M. Synthesis 1979, 471–472.
Separation of the phases is improved if a trace of water is
present (ca. 0.5 ml in 100 ml CH3CN).
15. Kucar, S.; Zamocky, J.; Zemek, J.; Anderle, D.; Mat-
ulova, M. Coll. Czech. Chem. Commun. 1984, 49, 1780–
1787.
16. Zottola, M.; Rao, V.; Fraser-Reid, B. J. Chem. Soc.,
Chem. Commun. 1991, 969–970.
17. Lewis, M. D.; Cho, J. K.; Kishi, Y. J. Am. Chem. Soc.
1982, 104, 4976–4978.
18. McDevitt, J. P.; Lansbury, P. T., Jr. J. Am. Chem. Soc.
1996, 118, 3818–3828.
19. Panek, J. S.; Sparks, M. A. J. Org. Chem. 1989, 54,
2034–2038.
20. Pirrung, M. C.; Kenney, P. M. J. Org. Chem. 1987, 52,
2335–2336.
21. Ramon, D. J.; Yus, M. Tetrahedron 1993, 49, 10103–
10110.
The benzyl ether stannane 12 was prepared from com-
mercially available 3-chloro-2-chloromethyl-1-propene
22. Investigation of a range of conditions with sodium or
potassium hydride and potassium t-butoxide in ether or