New synthesis of (-)- and (+)-actinobolin from D-glucose

Article abstract of DOI:10.1016/S0040-4039(03)01186-9

The total synthesis of (-)-actinobolin 2, an antipode of the natural product starting from D-glucose is described. A three-component coupling reaction of a functionalized cyclohexenone (+)-6, derived from D-glucose by way of Ferrier's carbocyclization, wi

Full text of DOI:10.1016/S0040-4039(03)01186-9

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 5047–5051
New synthesis of (−)- and (+)-actinobolin from
D-glucose
Satoshi Imuta, Shinya Ochiai, Miho Kuribayashi and Noritaka Chida*
Department of Applied Chemistry, Faculty of Science and Technology, Keio University, 3-14-1 Hiyoshi, Kohoku-ku,
Yokohama 223-8522, Japan
Received 9 April 2003; revised 26 April 2003; accepted 7 May 2003
Abstract—The total synthesis of (−)-actinobolin 2, an antipode of the natural product starting from
D-glucose is described. A
three-component coupling reaction of a functionalized cyclohexenone (+)-6, derived from -glucose by way of Ferrier’s
D
carbocyclization, with vinyl cuprate and an aldehyde (R)-5 effectively constructed the carbon framework of 2 in a highly
stereoselective manner. The formal synthesis of the natural enantiomer 1 from
Science Ltd. All rights reserved.
D-glucose was also achieved. © 2003 Elsevier
(+)-Actinobolin 1, isolated from the culture broths of
Streptomyces in 1959, has a broad antibacterial spec-
trum as well as moderate antitumor activity.1 The
structure elucidation study revealed that actinobolin
has a highly oxygenated bicyclic g-lactone (tetra-
hydroisochroman) framework with five contiguous chi-
attention from the synthetic community, and several
reports on total syntheses5 and chemical modification6
of actinobolin and bactobolin have been described. We
now report the new total synthesis of (−)-actinobolin 2,
the antipode of the natural product expected to show
some biological activity, starting from
D
-glucose. The
ral centers including an
L
-alanine residue.2 Later, in
formal synthesis starting from D-glucose of natural
1979, a structurally related natural product, bactobolin
was discovered and found to show more potent activi-
ties than actinobolin.3 It has also been reported that
they suppress antibody production and have a thera-
peutic effect on autoimmune encephalomyelitis.4 Such
interesting and challenging structures with potent bio-
logical properties have naturally received considerable
enantiomer 1 is also presented.
Our retrosynthetic analysis for (−)-actinobolin 2 sug-
gested that a bicyclic g-lactone possessing an azide
function 3 would be a promising intermediate for the
total synthesis (Fig. 1). The g-lactone 3 was expected to
arise from cyclohexanone derivative 4, which we
Figure 1. Structures of actinobolin and bactobolin, and retrosynthetic route to (−)-actinobolin. TBS=–SiMe2(t-Bu), MPM=
–CH2C6H4(p-OMe), Bn=–CH2Ph, MOM=–CH2OMe.
Keywords: actinobolin; Ferrier’s carbocyclization; three-component coupling reaction; carbohydrate.
* Corresponding author. Tel.: +81-45-566-1573; fax: +81-45-566-1573; e-mail: chida@applc.keio.ac.jp
0040-4039/03/$ - see front matter © 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0040-4039(03)01186-9

5048
S. Imuta et al. / Tetrahedron Letters 44 (2003) 5047–5051
planned to prepare by way of a one-pot three-compo-
nent coupling reaction7 of cyclohexenone (+)-6, alde-
hyde 5 and a vinyl metal species. The cyclohexenone
(+)-6, in turn, was envisioned to be synthesized in
racemic ( )-5 in 88% overall yield. On the other hand,
acid catalyzed p-methoxybenzylation of methyl (R)-lac-
tate,13 and subsequent reduction afforded (R)-5 in 66%
yield from methyl lactate. Similar treatment of (S)-lac-
tate gave (S)-5.
optically pure form starting from
D-glucose utilizing a
Ferrier’s carbocyclization8 as the key transformation.
With chiral cyclohexenone (+)-6 and aldehydes 5 in
hand, the crucial three-component coupling reaction
was investigated using vinyl cuprate as the nucleophile
(Scheme 3). Treatment of (+)-6 with higher order vinyl
cuprate in Et2O at −78°C caused the stereoselective
conjugate addition of the vinyl group to give an enolate
intermediate 6%, which was then reacted with racemic
aldehyde ( )-5 (excess amount) at −78°C, to provide 4
as the major isomer in 68% yield after chromatographic
separation. When chiral aldehyde (R)-5 was employed
as the electrophile, the same diastereomer 4 was
obtained in 85% yield. Interestingly, with another chiral
aldehyde (S)-5, the aldol process was found to proceed
much slower than the reaction with (R)-5, and a differ-
ent diastereomer 13, in which the stereochemistry of
substituents at C-2 and C-3 were cis, was formed in
75% yield. The predominant formation of 1%,2%-syn iso-
mers (4 and 13) suggested that the chelation control
(chelation by the alkoxy and aldehyde oxygen in (R)-
and (S)-5) should be an important factor in the aldol
process. Reaction of chelated (R)-5 and the intermedi-
ate enolate 6% would proceed in a ‘matched pair’ man-
ner (route a in Scheme 3) to give 4 smoothly, whereas
combination of (S)-5 and 6% would be ‘mismatched’.
The steric repulsion between chelated (S)-5 and 6% (both
routes b and c) rendered the aldol reaction sluggish, but
gave 2,3-cis-adduct 13 stereoselectively via route c.14
Synthesis of the cyclohexenone (+)-6 commenced from
3-deoxy-
D
-glucose derivative9 8 prepared from commer-
cially available methyl 4,6-O-benzylidene-a-
D
-glucopy-
ranoside 7 in two steps in 75% yield (Scheme 1).
Reduction of 8 with DIBAL-H gave 910 (73% yield),
whose primary hydroxy group was selectively iodinated
to afford 10 in 99% yield. After protection of the
remaining hydroxy function as a TBS ether, the result-
ing compound was treated with t-BuOK to give 5-
enopyranoside 11 in 81% yield. Catalytic Ferrier’s
carbocyclization11 of 11 in acetone–acetate buffer12 pro-
vided 12 as a mixture of diastereomers (a-OH:b-OH=
ca. 1:10). b-Elimination of the mixture cleanly gener-
ated cyclohexenone (+)-6 in 84% yield from 11. The
other requisite fragment, aldehyde 5 was synthesized
from 1,2-propanediol, and methyl (R)- and (S)-lac-
tates (Scheme 2). 1,2-Propanediol was converted into
a p-methoxybenzylidene acetal derivative, whose treat-
ment with DIBAL-H, followed by oxidation furnished
The stereoselective formation of three-components
adduct 4 led us to use 4 as the precursor for the
synthesis of (−)-actinobolin 2. Conversion of 4 to 2
required the following transformations: (1) reduction of
ketone carbonyl to b-alcohol; (2) introduction of a
Scheme 1.
Scheme 2.
Scheme 3.

S. Imuta et al. / Tetrahedron Letters 44 (2003) 5047–5051
5049
nitrogen function at C-1% via SN2 fashion; and (3)
formation of g-lactone with inversion of the configura-
tion at C-2% hydroxy group.
presence of HCl reduced the azide function and
removed the O-benzyl group to provide the amine
hydrochloride, which, without isolation, was condensed
with N-benzyloxycarbonyl-D-alanine (Z-D-alanine) in
Treatment of 4 with Me4NBH(OAc)315 at room temper-
ature stereoselectively reduced the carbonyl group to
give desired syn diol 14 in 70% yield (Scheme 4). The
hydroxy group at C-2 in 14 was anticipated to show
less reactivity than that at C-1% due to steric congestion.
Indeed, reaction of 14 with BuLi (3 equiv.) at 0°C,
followed by treatment with TsCl (2.8 equiv.) generated
the 1%-OTs derivative, quantitatively. The remaining
hydroxy function was then masked as a MOM ether to
afford 15 in 87% yield.
the presence of DCC to give protected actinobolin 22 in
57% yield. Finally, removal of benzyloxycarbonyl
group by hydrogenolysis in MeOH–AcOH–1 M aq.
HCl, followed by purification with Sephadex LH-20
(MeOH) furnished (−)-actinobolin hydrochloride
(2·HCl) in 76% yield. The spectral (1H and 13C NMR)
data of synthetic 2·HCl were fully identical with those
of natural (+)-actinobolin hydrochloride, kindly pro-
vided by Dr. Y. Nishimura, and the [h]D value of the
synthetic compound {[h]2D4 −47 (c 0.1, H2O): lit.,5f [h]D20
+48 (c 0.40, H2O)} confirmed its unnatural absolute
configuration.
The MPM protecting group was removed to give 16
(97% yield). Since attempted inversion of the hydroxy
function in 16 by Mitsunobu reaction proved fruitless,
we adopted an oxidation–reduction procedure. Dess–
Martin oxidation of 16 afforded methyl ketone 17 in
100% yield. Reduction of the ketone 17 with various
reducing reagents was attempted, however, the desired
inverted alcohol could not be obtained as the major
isomer. Fortunately, it was found that the stereoselec-
tive reduction successfully proceeded when carboxylic
acid 18 was employed as the substrate. Thus, ozonolysis
of 17, followed by further oxidation with sodium chlor-
ite afforded 18 in 83% yield. In this case, reduction of
18 with NaBH4 in MeOH provided desired product 19
as the major isomer (19: its 2%-epimer=82:18, 96%
yield), and bicyclic compound 20 was obtained in 72%
yield from 18 after lactonization followed by separation
with silica gel chromatography.16 Azidolysis of 20 with
NaN3 provided advanced intermediate 3 in 86% yield.
Deprotection of the O-TBS moiety in 3, followed by
Swern oxidation afforded b-ketoester 21 in 89% yield.
Interestingly, the methoxymethyl group was unexpect-
edly removed during the purification process with silica-
gel chromatography. Hydrogenation of 21 in the
Having established a new synthetic pathway to (−)-acti-
nobolin from D-glucose, we turned our attention to the
synthesis of the natural enantiomer 1, also starting
from -glucose. For this purpose, compound 8 was
D
again chosen as a building block, and its transforma-
tion into the enantiomer of (+)-6 was investigated
(Scheme 5). Benzylation of the hydroxy group in 8,
followed by acetal hydrolysis and selective iodination of
the resulting primary alcohol afforded 23 in 87% over-
all yield. Treatment of 23 with base and subsequent
O-silylation gave 24 in 60% yield. Catalytic Ferrier’s
carbocyclization of 24 in acetone–acetate buffer gener-
ated 25 as a diastereomeric mixture (a-OH: b-OH=ca.
6:1) in 83% yield. Protection of the hydroxy group in 25
as a THP ether and subsequent reduction of the ketone
carbonyl, followed by O-methanesulfonylation and
acidic workup afforded 26 as the major product in 66%
yield. The observed large coupling constants in 26
(J1,2=8.8, J4,5=9.0 Hz) clearly showed that both OMs
and OH groups were in the equatorial positions. Swern
oxidation of 26 was accompanied by the b-elimination
of the OMs group to furnish (−)-6 in 93% yield. The
Scheme 4.

5050
S. Imuta et al. / Tetrahedron Letters 44 (2003) 5047–5051
3. (a) Kondo, S.; Horiuchi, Y.; Hamada, M.; Takeuchi, T.;
Umezawa, H. J. Antibiot. 1979, 32, 1069–1071; (b) Hori,
M.; Suzukake, K.; Ishikawa, C.; Asakura, H.; Umezawa,
H. J. Antibiot. 1981, 34, 465–468.
4. (a) Ishizuka, M.; Fakasawa, S.; Masuda, T.; Sato, J.;
Kanbayashi, N.; Takeuchi, T.; Umezawa, H. J. Antibiot.
1980, 33, 1054–1062; (b) Tabira, T.; Da-Lin, Y.; Yama-
mura, T.; Aoyagi, T. Proc. Jpn. Acad. Ser. B 1987, 63,
127–130.
5. (a) Yoshioka, M.; Nakai, H.; Ohno, M. J. Am. Chem.
Soc. 1984, 106, 1133–1135; Heterocycles 1984, 21, 151–
165; (b) Askin D.; Angst, C.; Danishefsky, S. J. J. Org.
Chem. 1985, 50, 5005–5007; J. Org. Chem. 1987, 52,
622–635; (c) Rahman, M. A.; Fraser-Reid, B. J. Am.
Chem. Soc. 1985, 107, 5576–5578; (d) Kozikowski, A. P.;
Konoike, T.; Nieduzak, T. R. J. Chem. Soc., Chem.
Commun. 1986, 1350–1352; (e) Kozikowski, A. P.;
Nieduzak, T. R.; Konoike, T.; Springer, J. P. J. Am.
Chem. Soc. 1987, 109, 5167–5175; (f) Garigipati, R. S.;
Tschaen, D. M.; Weinreb, S. M. J. Am. Chem. Soc. 1985,
107, 7790–7792; J. Am. Chem. Soc. 1990, 112, 3475–3482;
(g) Ward, D. E.; Kaller, B. F. Tetrahedron Lett. 1993, 34,
407–410; J. Org. Chem. 1994, 59, 4230–4238; (h) Ward,
D. E.; Yuanzhu, G.; Kaller, B. F. J. Org. Chem. 1996, 61,
5498–5505.
6. (a) Munakata, T. Yakugaku Zasshi 1981, 101, 138–147;
(b) Munakata, T.; Okumoto, T. Chem. Pharm. Bull.
1981, 29, 891–894; (c) Nishimura, Y.; Kondo, S.;
Takeuchi, T. J. Antibiot. 1992, 45, 735–741; (d) Adachi,
M.; Nishimura, Y.; Kondo, S.; Takeuchi, T. J. Antibiot.
1998, 51, 202–209; (e) Adachi, H.; Nishimura, Y.;
Takeuchi, T. J. Antibiot. 2002, 55, 92–98.
7. For a review, see: (a) Taylor, R. J. K. Synthesis 1985,
364–392. We accomplished the synthesis of CD ring unit
of paclitaxel using a three-component coupling strategy,
see: (b) Momose, T.; Setoguchi, M.; Fujita, T.; Tamura,
H.; Chida, N. Chem. Commun. 2000, 2237–2238.
8. (a) Ferrier, R. J.; Middleton, S. Chem. Rev. 1993, 93,
2779–2831; Top. Curr. Chem. 2001, 215, 277–291. For
recent applications of Ferrier’s carbocyclization in natu-
ral product synthesis, see: (b) Takahashi, H.; Kittaka, H.;
Ikegami, S. J. Org. Chem. 2001, 66, 2705–2716; (c) Taille-
fumier, C.; Chapleur, Y. Can. J. Chem. 2000, 78, 708–
722; (d) Amano, S.; Ogawa, N.; Ohtsuka, M.; Ogawa, S.;
Chida, N. Tetrahedron 1999, 55, 3855–3870.
9. Vis, E.; Karrer, P. Helv. Chim. Acta 1954, 46, 378–381.
10. All new compounds described in this letter were isolated
as a single compound by chromatographic separation
and/or recrystallization, and fully characterized by 300
MHz 1H NMR, 75 MHz 13C NMR, IR, and mass
spectrometric and/or elemental analyses. The stereo-
chemistries were assigned based on NMR experiments
(COSY, decoupling, and/or NOE).
Scheme 5.
spectral data and the absolute value of [h]D of (−)-6
{[h]2D0 −22 (c 0.80, CHCl3)} were fully identical with
those of (+)-6 {[h]2D3 +22 (c 0.94, CHCl3)}, representing
a formal synthesis of (+)-actinobolin 1.
In summary, a new synthetic route to both (−)- and
(+)-actinobolin starting from
D-glucose has been estab-
lished. This work demonstrated that the methodology
involving the three-component coupling reaction on
chiral cyclohexenones, derived from carbohydrates by
way of Ferrier’s carbocyclization, is effective for the
chiral and stereoselective synthesis of natural products
possessing highly oxygenated cyclohexane moieties.
Acknowledgements
We thank Dr. Y. Nishimura (Institute of Microbial
Chemistry, Tokyo, Japan) for a generous gift of natural
actinobolin.
References
1. (a) Haskell, T. H.; Bartz, Q. R. Anibiot. Ann. 1958–1959,
505–509; (b) Fusari, S. A.; Machamer, H. E. Antibiot.
Ann. 1958–1959, 510–514; (c) Pittillo, R. F.; Fisher, M.
W.; McAlpine, R. J.; Thompson, P. E.; Ehrlich, J.
Anibiot. Ann. 1958–1959, 497–504; (d) Merker, P. C.;
Woolley, G. W. Antibiot. Ann. 1958–1959, 515–517; (e)
Teller, M. N.; Merker, P. C.; Palm, J. E.; Woolley, G. W.
Antibiot. Ann. 1958–1959, 518–521.
2. (a) Struck, R. F.; Thorpe, W. C.; Coburn, W. C., Jr.;
Shealy, Y. F. Tetrahedron Lett. 1967, 1589–1595; (b)
Antosz, F. J.; Nelson, D. B.; Herald, D. L., Jr.; Munk,
M. E. J. Am. Chem. Soc. 1970, 92, 4933–4942; (c)
Wetherington, J. B.; Moncrief, J. W. Acta Crystallogr.
1975, B31, 501–511.
11. Chida, N.; Ohtsuka, M.; Ogura, K.; Ogawa, S. Bull.
Chem. Soc. Jpn. 1991, 64, 2118–2121.
12. Use of acetate buffer (pH 4.8, 0.1 M solution) suppressed
the partial hydrolysis of O-TBS group during the reac-
tion, and greatly improved the cyclization yields.
13. Nakajima, N.; Horita, K.; Abe, R.; Yonemitsu, O. Tetra-
hedron Lett. 1988, 29, 4139–4142.
14. Further investigation of the interesting aldol process of
the intermediate enolate 6% with various aldehydes is
underway and the results will be reported in a full
account.

S. Imuta et al. / Tetrahedron Letters 44 (2003) 5047–5051
5051
15. Evans, D. A.; Chapman, K. T.; Carreira, E. M. J. Am.
Chem. Soc. 1988, 110, 3560–3578. It is noteworthy that
reduction of 4 with Me4NBH(OAc)3 afforded the syn
diol stereoselectively. This reagent is known to give 1,3-
anti diols preferentially in the reduction of b-hydroxy-
ketone systems.
THF), a g-lactone with retention of the 2% stereochem-
istry (3-epimer of 20) was formed in 80% yield. A simi-
lar phenomenon and mechanistic study on Mitsunobu
reaction of hydroxycarboxylic acids with hindered alco-
hols have been reported, see: Ahn, C.; Correia, R.;
DeShong, P. J. Org. Chem. 2002, 67, 1751–1753 and
Ahn, C.; Deshong, P. J. Org. Chem. 2002, 67, 1754–
1759.
16. When 2%-epimer of 19 was subjected to the intramolecu-
lar Mitsunobu reaction (Ph3P, diethyl azodicarboxylate,