Halichondrin B: Synthesis of the C1-C22 subunit

Halichondrin B: Synthesis of the C1-C22 Subunit

William T. Lambert, Gregory H. Hanson, Farid Benayoud, and Steven D. Burke*

Department of Chemistry, University of Wisconsin-Madison, 1101 University Avenue,

Madison, Wisconsin 53706-1396

burke@chem.wisc.edu

Received July 16, 2005

Two efficient routes to the C1-C22 subunit of halichondrin B are described. The cage ketal 7,

which contains 11 asymmetric centers embedded within the ABCDEF-ring framework, was

assembled from (+)-conduritol E (27) in 18 steps and 4% overall yield. In a separate route, 7 was

also synthesized in 18 steps and 2% overall yield from a derivative of R-^D-glucoheptonic acid γ-lactone

(62). While the former route installs the fully elaborated C-ring endowed with the correct C12

stereochemistry early in the synthesis, the latter features a late-stage introduction of the C12

stereocenter during the ultimate one-pot Michael addition/ketalization cascade to form the CDE-

ring system of the cage. The importance of the C12 stereocenter to the crucial ketalization event

is discussed through comparison of these two strategies.

Introduction

Halichondria okadai is a common marine sponge found

off the coast of Japan. In 1986, Hirata and Uemura

reported that extracts from this black-colored animal

displayed compelling in vivo antitumor activity.^1aEight

active compounds were separated from these extracts and

purified and were subsequently named the halichondrins.

The most potent of these was halichondrin B (Figure 1),

isolated in low yield (1.8 × 10^-6%), which displayed an

in vitro IC₅₀value for B-16 melanoma of 0.093 ng/mL. It

was also found to have significant in vivo activity against

B-16 melanoma, P-388 leukemia, and L-1210 leukemia.

Halichondrin B features 32 asymmetric centers that

decorate a contiguous 54-carbon backbone. Structural

features of note include a macrolactone spanning C1-

C30, exo-methylene groups at C19 and C26, spiroketals

at C38 and C44, and a 2,6,9-trioxatricyclo[3.2.2.0^3,7]-

decane ketal (hereafter referred to as the “cage”) embed-

FIGURE 1. Halichondrin B.

ded within the C8-C14 region of the molecule. The latter

is an interesting structural motif unique to the halichon-

drins and is thought to be largely responsible for its

remarkable biological activity. From their original data,

Uemura and Hirata noted a strong correlation between

the lipophilicity of the cage portion of the halichondrins

and antitumor potency. Indeed, a C1-C38 diol (Figure

2) prepared by the Kishi group during the course of their

synthetic studies²was found to display the same activity

profile as that of halichondrin B over the 60 cell lines

screened by the National Cancer Institute (NCI) and gave

IC₅₀values within 1 order of magnitude of the natural

product itself. However, the in vivo potency of this

compound was substantially lower than that of halichon-

(1) (a) Hirata, Y.; Uemura, D. Pure Appl. Chem. 1986, 58, 701-

710. (b) Pettit, G. R.; Herald, C. L.; Boyd, M. R.; Leet, J. E.; Dufresne,

C.; Doubek, D. L.; Schmidt, J. M.; Cerny, R. L.; Hooper, J. N. A.;

Ru¨tzler, K. C. J. Med. Chem. 1991, 34, 3339-3340. (c) Pettit, G. R.;

Gao, F.; Doubek, D. L.; Boyd, M. R.; Hamel, E.; Bai, R.; Schmidt, J.

M.; Tackett, L. P.; Ru¨tzler, K. Gazz. Chim. Ital. 1993, 123, 371-377.

(d) Pettit, G. R.; Tan, R.; Gao, F.; Williams, M. D.; Doubek, D. L.; Boyd,

M. R.; Schmidt, J. M.; Chapuis, J.-C.; Hamel, E.; Bai, R.; Hooper, J.

N. A.; Tackett, L. P. J. Org. Chem. 1993, 58, 2538-2543. (e) Litaudon,

M.; Hart, J. B.; Blunt, J. W.; Lake, R. J.; Munro, M. H. G. Tetrahedron

Lett. 1994, 35, 9435-9438.

(2) Stamos, D. P.; Chen, S. S.; Kishi, Y. J. Org. Chem. 1997, 62,

7552-7553.

Published on Web 10/15/2005

9382

J. Org. Chem. 2005, 70, 9382-9398

Halichondrin B: Synthesis of the C1-C22 Subunit

thwarted progress toward this end. Although halichon-

drin B has since been isolated from three other sponge

genera, the yields from these sources, which include

Axinella (1.8 × 10^-6%),^1b,cPhakellia (3.5 × 10^-6%),^1dand

Lissodendoryx (4.0 × 10^-5%),^1ehave also been disappoint-

ingly low. The isolation of the halichondrins from four

different sources has led Munro et al. to suggest that the

natural products are synthesized by a microbial sponge

symbiont rather than by the sponges themselves.^7aThere

are reports detailing attempts at aquacultural production

of the most prolific sponge, Lissodendoryx n. sp. 1,^7a,bbut

this work is in the very early stages of development, and

the authors note that this would require a yearly output

of nearly 5 million kg of sponge in order to satisfy the

anticipated clinical demand of 5 kg per year of halichon-

drin B.

Chemical synthesis has proven to be a viable method

by which scarce natural products of considerable struc-

tural complexity may be accessed in larger quantities.⁸

The exceedingly small amounts of halichondrin B avail-

able from natural sources provides a strong impetus for

its total synthesis. Efforts toward this end have been

published by four groups, including Kishi,⁹Salomon,¹⁰

Yonemitsu,¹¹and our own.¹²Although a total synthesis

of halichondrin B was published in 1992 by the Kishi

group,^9aits length (ca. 53 linear steps/ca. 115 total steps)

limits its scalability. Among several reports describing

halichondrin subunit syntheses,¹²we have previously

published syntheses of a C1-C15 subunit,^12grequiring

16 steps and proceeding in 0.6% overall yield from R-^D-

glucoheptonic acid γ-lactone, as well as a C1-C14

subunit model,^12jwhich was obtained in 3% overall yield

over 18 steps from (+)-conduritol E.

FIGURE 2. Biologically active halichondrin B analogues.

drin B. Recent drug discovery efforts at Eisai Research

Institute have yielded many different halichondrin B

analogues;³the two most highly active ones (ER-086526

and ER-076349, Figure 2) feature a truncated H-ring

bearing a hydrophilic side chain. The replacement of the

macrolactone functionality with a hydrolytically stable

ketone unit at C1 was done in order to prolong the in

vivo activity of these compounds. The primary amine ER-

086526 (E7389), which displays greater aqueous solubil-

ity than ER-076349, is currently in phase I and phase II

trials at the National Cancer Institute.⁴

The initial findings by Hirata and Uemura prompted

further investigation of the biological activity displayed

by halichondrin B. It was determined by the NCI that

halichondrin B is a mitotic inhibitor which binds to the

vinca domain of tubulin, resulting in inhibition of mi-

crotubule formation and tubulin-dependent GTP hydroly-

sis.⁵Halichondrin B was found to be extremely effective

in vivo against human solid tumors which had been

xenografted into immune-deficient mice.⁵It also exhibited

an IC₅₀value of 0.3 nM against L-1210 murine leukemia

cells, thereby surpassing in potency the best previously

known agents dolastatin 10 (0.5 nM), rhizoxin (1 nM),

and vinblastine (20 nM).⁵

(7) (a) Munro, M. H. G.; Blunt, J. W.; Dumdei, E. J.; Hickford, S. J.

H.; Lill, R. E.; Li, S.; Battershill, C. N.; Duckworth, A. R. J. Biotech.

1999, 70, 15-25. (b) Sipkema, D.; Osinga, R.; Schatton, W.; Mendola,

D.; Tramper, J.; Wijffels, R. H. Biotechnol. Bioeng. 2005, 90, 201-

222.

Halichondrin B is currently in preclinical development

by the NCI based on the data detailing its remarkable

biological activity.^5,6However, the extremely low yield

of natural product obtained from the original sponge has

(8) Several efforts illustrating this point have recently appeared.

For a recent synthesis of spongistatin, see: (a) Hubbs, J. L.; Heathcock,

C. H. J. Am. Chem. Soc. 2003, 125, 12836-12843. (b) Heathcock, C.

H.; McLaughlin, M.; Medina, J.; Hubbs, J. L.; Wallace, G. A.; Scott,

R.; Claffey, M. M.; Hayes, C. J.; Ott, G. R. J. Am. Chem. Soc. 2003,

125, 12844-12849. For (+)-discodermolide, see: (c) Smith, A. B., III;

Freeze, B. S.; Brouard, I.; Hirose, T. Org. Lett. 2003, 5, 4405-4408.

(d) Smith, A. B., III; Kaufman, M. D.; Beauchamp, T. J.; LaMarche,

M. J.; Arimoto, H. Org. Lett. 1999, 1, 1823-1826. (e) Mickel, S. J.;

Sedelmeier, G. H.; Niederer, D.; Daeffler, R.; Osmani, A.; Schreiner,

K.; Seeger-Weibel, M.; Be´rod, B.; Schaer, K.; Gamboni, R.; Chen, S.;

Chen, W.; Jagoe, C. T.; Kinder, F. R., Jr.; Loo, M.; Prasad, K.; Repic,

O.; Shieh, W.-C.; Wang, R.-M.; Waykole, L.; Xu, D. D.; Xue, S. Org.

Process Res. Dev. 2004, 8, 92-100. (f) Mickel, S. J.; Sedelmeier, G. H.;

Niederer, D.; Schuerch, F.; Grimler, D.; Koch, G.; Daeffler, R.; Osmani,

A.; Hirni, A.; Schaer, K.; Gamboni, R.; Bach, A.; Chaudhary, A.; Chen,

S.; Chen, W.; Hu, B.; Jagoe, C. T.; Kim, H.-Y.; Kinder, F. R., Jr.; Liu,

Y.; Lu, Y.; McKenna, J.; Prashad, M.; Ramsey, T. M.; Repic, O.; Rogers,

L.; Shieh, W.-C.; Wang, R.-M.; Waykole, L. Org. Process Res. Dev. 2004,

8, 101-106. (g) Mickel, S. J.; Sedelmeier, G. H.; Niederer, D.; Schuerch,

F.; Koch, G.; Kuesters, E.; Daeffler, R.; Osmani, A.; Seeger-Weibel, M.;

Schmid, E.; Hirni, A.; Schaer, K.; Gamboni, R.; Bach, A.; Chen, S.;

Chen, W.; Geng, P.; Jagoe, C. T.; Kinder, F. R., Jr.; Lee, G. T.;

McKenna, J.; Ramsey, T. M.; Repic, O.; Rogers, L.; Shieh, W.-C.; Wang,

R.-M.; Waykole, L. Org. Process Res. Dev. 2004, 8, 107-112. (h) Mickel,

S. J.; Sedelmeier, G. H.; Niederer, D.; Schuerch, F.; Seger, M.;

Schreiner, K.; Daeffler, R.; Osmani, A.; Bixel, D.; Loiseleur, O.; Cercus,

J.; Stettler, H.; Schaer, K.; Gamboni, R.; Bach, A.; Chen, G.-P.; Chen,

W.; Geng, P.; Lee, G. T.; Loeser, E.; McKenna, J.; Kinder, F. R., Jr.;

Konigsberger, K.; Prasad, K.; Ramsey, T. M.; Reel, N.; Repic, O.;

Rogers, L.; Shieh, W.-C.; Wang, R.-M.; Waykole, L.; Xue, S. Org. Process

Res. Dev. 2004, 8, 113-121. (i) Mickel, S. J.; Niederer, D.; Daeffler,

R.; Osmani, A.; Kuesters, E.; Schmid, E.; Schaer, K.; Gamboni, R.;

Chen, W.; Loeser, E.; Kinder, F. R., Jr.; Konigsberger, K.; Prasad, K.;

Ramsey, T. M.; Repic, O.; Wang, R.-M.; Florence, G.; Lyothier, I.;

Paterson, I. Org. Process Res. Dev. 2004, 8, 122-130.

(3) (a) Zheng, W.; Seletsky, B. M.; Palme, M. H.; Lydon, P. J.; Singer,

L. A.; Chase, C. E.; Lemelin, C. A.; Shen, Y.; Davis, H.; Tremblay, L.;

Towle, M. J.; Salvato, K. A.; Wels, B. F.; Aalfs, K. K.; Kishi, Y.;

Littlefield, B. A.; Yu, M. J. Bioorg. Med. Chem. Lett. 2004, 14, 5551-

5554. (b) Seletsky, B. M.; Wang, Y.; Hawkins, L. D.; Palme, M. H.;

Habgood, G. J.; DiPietro, L. V.; Towle, M. J.; Salvato, K. A.; Wels, B.

F.; Aalfs, K. K.; Kishi, Y.; Littlefield, B. A.; Yu, M. J. Bioorg. Med.

Chem. Lett. 2004, 14, 5547-5550. (c) Towle, M. J.; Salvato, K. A.;

Budrow, J.; Wels, B. F.; Kuznetsov, G.; Aalfs, K. K.; Welsh, S.; Zheng,

W.; Seletsky, B. M.; Palme, M. H.; Habgood, G. J.; Singer, L. A.;

DiPietro, L. V.; Wang, Y.; Chen, J. J.; Quincy, D. A.; Davis, A.;

Yoshimatsu, K.; Kishi, Y.; Yu, M. J.; Littlefield, B. A. Cancer Res. 2001,

61, 1013-1021. (d) Wang, Y.; Habgood, G. J.; Christ, W. J.; Kishi, Y.;

Littlefield, B. A.; Yu, M. J. Bioorg. Med. Chem. Lett. 2000, 10, 1029-

1032.

(4) Compound E7389 is involved in three clinical trials: National

Cancer Institute-Clinical Trials Results at http://www.cancer.gov/

search/ResultsClinicalTrialsAdvanced.aspx?protocol-search-

id)1667610, accessed 10/7/05.

(5) (a) Hamel, E. Pharmacol. Ther. 1992, 55, 31-51. (b) Bai, R.;

Paull, K. D.; Herald, C. L.; Malspeis, L.; Pettit, G. R.; Hamel, E. J.

Biol. Chem. 1991, 266, 15882-15889. (c) Paull, K. D.; Lin, C. M.;

Malspeis, L.; Hamel, E. Cancer Res. 1992, 52, 3892-3900. (d) Luduen˜a,

R. F.; Roach, M. C.; Prasad, V.; Pettit, G. R. Biochem. Pharmacol. 1993,

45, 421-427. (e) Roberson, R. W.; Tucker, B.; Pettit, G. R. Mycol. Res.

1998, 102, 378-382.

(6) Halichondrin B has been in preclinical development since 5/1/

2000, see: Questions and Answers About NCI’s Natural Products

Branch, Cancer Facts 7.33 at http://cis.nci.nih.gov/fact/7_33.htm),

accessed 10/7/05.

J. Org. Chem, Vol. 70, No. 23, 2005 9383

Lambert et al.

quent synthesis of the C1-C14 subunit model 6,^12jwe

introduced the C12 stereochemistry in a controlled man-

ner early in the sequence by constructing the densely

functionalized C-ring at the outset. Subjection of the enol

ether substrate 4 to the acidic hydrolysis conditions

shown led to clean formation of 6, presumably proceeding

through the intermediate ketone 5. This favorable result

appeared to validate our assertion that the C12 stereo-

chemistry played a key role in facilitating cage formation

by enforcing proximity between the reactive functionality.

Furthermore, we were able to employ milder conditions

in this case (p-TsOH, CH₂Cl₂, MeOH, H₂O) due to the

absence of any robust protecting groups that resisted

hydrolysis. The presence of the C8-C9 pentylidene ketal

and the C11 TBS ether in our first synthesis necessitated

the use of more forcing conditions (52% aq HF, CH₃CN)

to induce the deprotection of the C8 and C9 alcohols,

which preceded the Michael addition/ketalization cas-

cade. We sought to modify the protecting group strategy

of our first-generation synthesis in order to allow for the

use of a milder acid promoter, thereby enabling a better

comparison of the two ketalization protocols. We also

desired to incorporate the adjacent F-ring into our

synthesis in order to access the C1-C22 fragment of

halichondrin B, incorporating our results in recently

reported syntheses of the C14-C22 subunit.^12i,kIn this

paper, we report the synthesis of the C1-C22 subunit 7

by two routes, which feature alternative protocols for

setting the C12 stereocenter and for effecting the critical

ketalization event. Our work leading to the synthesis of

6 is also discussed in full detail, including results not

described in our earlier communication.^12j

SCHEME 1

Results and Discussion

A key issue in both of our reported syntheses of the

ABCDE-ring system was the late-stage construction of

the acid-sensitive caged ketal. As shown in Scheme 1,

C8, C9, and C11 hydroxyl unmasking in 1 gave the enone

triol 2, which then underwent Michael addition of the

C9 hydroxyl to set the C12 stereocenter and ketalization

of the C14 ketone with the C8 and C11 hydroxyl groups

to provide the C1-C15 subunit model 3. We hypothesized

that poor control over the C12 stereochemistry during

the Michael addition step led to low and variable yields

of 3 in this first-generation synthesis.^12gIn our subse-

The retrosynthesis for our two routes to the C1-C22

subunit 7 are shown in Scheme 2.

(11) (a) Horita, K.; Hachiya, S.; Nagasawa, M.; Hikota, M.; Yone-

mitsu, O. Synlett 1994, 38-40. (b) Horita, K.; Nagasawa, M.; Hachiya,

S.; Yonemitsu, O. Synlett 1994, 40-43. (c) Horita, K.; Sakurai, Y.;

Nagasawa, M.; Hachiya, S.; Yonemitsu, O. Synlett 1994, 43-45. (d)

Horita, K.; Sakurai, Y.; Nagasawa, M.; Maeno, K.; Hachiya, S.;

Yonemitsu, O. Synlett 1994, 46-48. (e) Horita, K.; Hachiya, S.-i.;

Ogihara, K.; Yoshida, Y.; Nagasawa, M.; Yonemitsu, O. Heterocycles

1996, 42, 99-104. (f) Horita, K.; Nagasawa, M.; Hachiya, S.-i.; Sakurai,

Y.; Yamazaki, T.; Uenishi, J.; Yonemitsu, O. Tetrahedron Lett. 1997,

38, 8965-8968. (g) Horita, K.; Hachiya, S.-i.; Yamazaki, T.; Naitou,

T.; Uenishi, J.; Yonemitsu, O. Chem. Pharm. Bull. 1997, 45, 1265-

1281. (h) Horita, K.; Sakurai, Y.; Nagasawa, M.; Yonemitsu, O. Chem.

Pharm. Bull. 1997, 45, 1558-1572. (i) Yonemitsu, O.; Yamazaki, T.;

Uenishi, J.-i. Heterocycles 1998, 49, 89-92. (j) Horita, K.; Nagasawa,

M.; Sakurai, Y.; Yonemitsu, O. Chem. Pharm. Bull. 1998, 46, 1199-

1216. (k) Horita, K.; Nishibe, S.; Yonemitsu, O. Phytochem. Phyto-

pharm. 2000, 386-397.

(12) (a) Burke, S. D.; Buchanan, J. L.; Rovin, J. D. Tetrahedron Lett.

1991, 32, 3961-3964. (b) Burke, S. D.; Jung, K. W.; Phillips, J. R.;

Perri, R. E. Tetrahedron Lett. 1994, 35, 703-706. (c) Burke, S. D.;

Zhang, G.; Buchanan, J. L. Tetrahedron Lett. 1995, 36, 7023-7026.

(d) Burke, S. D.; Phillips, J. R.; Quinn, K. J.; Zhang, G.; Jung, K. W.;

Buchanan, J. L.; Perri, R. E. Synthetic Studies Toward Complex

Polyether Macrolides of Marine Origin. In Anti-Infectives: Recent

Advances in Chemistry and Structure-Activity Relationships; Bentley,

P. H., O’Hanlon, P. J., Eds.; The Royal Society of Chemistry: Cam-

bridge, 1997; pp 73-85. (e) Burke, S. D.; Austad, B. C.; Hart, A. C. J.

Org. Chem. 1998, 63, 6770-6771. (f) Burke, S. D.; Quinn, K. J.; Chen,

V. J. J. Org. Chem. 1998, 63, 8626-8627. (g) Burke, S. D.; Jung, K.

W.; Lambert, W. T.; Phillips, J. R.; Klovning, J. J. J. Org. Chem. 2000,

65, 4070-4087. (h) Austad, B. C.; Hart, A. C.; Burke, S. D. Tetrahedron

2002, 58, 2011-2026. (i) Jiang, L.; Burke, S. D. Org. Lett. 2002, 4,

3411-3414. (j) Lambert, W. T.; Burke, S. D. Org. Lett. 2003, 5, 515-

518. (k) Jiang, L.; Martinelli, J. R.; Burke, S. D. J. Org. Chem. 2003,

68, 1150-1153. (l) Keller, V. A.; Kim, I.; Burke, S. D. Org. Lett. 2005,

7, 737-740.

(9) (a) Aicher, T. D.; Buszek, K. R.; Fang, F. G.; Forsyth, C. J.; Jung,

S. H.; Kishi, Y.; Matelich, M. C.; Scola, P. M.; Spero, D. M.; Yoon, S.

K. J. Am. Chem. Soc. 1992, 114, 3162-3164. (b) Aicher, T. D.; Kishi,

Y. Tetrahedron Lett. 1987, 28, 3463-3466. (c) Aicher, T. D.; Buszek,

K. R.; Fang, F. G.; Forsyth, C. J.; Jung, S. H.; Kishi, Y.; Scola, P. M.

Tetrahedron Lett. 1992, 33, 1549-1552. (d) Buszek, K. R.; Fang, F.

G.; Forsyth, C. J.; Jung, S. H.; Kishi, Y.; Scola, P. M.; Yoon, S. K.

Tetrahedron Lett. 1992, 33, 1553-1556. (e) Fang, F. G.; Kishi, Y.;

Matelich, M. C.; Scola, P. M. Tetrahedron Lett. 1992, 33, 1557-1560.

(f) Duan, J. J.-W.; Kishi, Y. Tetrahedron Lett. 1993, 34, 7541-7544.

(g) Stamos, D. P.; Kishi, Y. Tetrahedron Lett. 1996, 37, 8643-8646.

(h) Xie, C.; Nowak, P.; Kishi, Y. Org. Lett. 2002, 4, 4427-4429. (i) Wan,

Z.-K.; Choi, H.-w.; Kang, F.-A.; Nakajima, K.; Demeke, D.; Kishi, Y.

Org. Lett. 2002, 4, 4431-4434. (j) Choi, H.-w.; Nakajima, K.; Demeke,

D.; Kang, F.-A.; Jun, H.-S.; Wan, Z.-K.; Kishi, Y. Org. Lett. 2002, 4,

4435-4438. (k) Choi, H.-w.; Demeke, D.; Kang, F.-A.; Kishi, Y.;

Nakajima, K.; Nowak, P.; Wan, Z.-K.; Xie, C. Pure Appl. Chem. 2003,

75, 1-17. (l) Namba, K.; Jun, H.-S.; Kishi, Y. J. Am. Chem. Soc. 2004,

126, 7770-7771.

(10) (a) Kim, S.; Salomon, R. G. Tetrahedron Lett. 1989, 30, 6279-

6282. (b) Cooper, A. J.; Salomon, R. G. Tetrahedron Lett. 1990, 31,

3813-3816. (c) DiFranco, E.; Ravikumar, V. T.; Salomon, R. G.

Tetrahedron Lett. 1993, 34, 3247-3250. (d) Cooper, A. J.; Pan, W.;

Salomon, R. G. Tetrahedron Lett. 1993, 34, 8193-8196.

9384 J. Org. Chem., Vol. 70, No. 23, 2005

Halichondrin B: Synthesis of the C1-C22 Subunit

SCHEME 2

SCHEME 3

of some significance, as we had also considered the

analogous C₂-symmetric bis(lactone) as a potentially

attractive intermediate. At issue was the point at which

desymmetrization could most effectively be accomplished.

Our decision to pursue 13 was based on the proposal that

it could be prepared via a symmetry-breaking ozonolytic

cleavage of a C₂-symmetric cyclohexene diol endowed

with the ^L-mannose configuration. Compound 14, a

protected derivative of (+)-conduritol E,¹³meets these

criteria and was therefore seen as an attractive precursor

to 13. Ozonolytic cleavage of the carbon-carbon double

bond in 14 was expected to give rise to the dissymmetric

intermediate 15, which should form upon retro-[3 + 2]

cycloaddition of the initially formed primary ozonide.

Intramolecular trapping of the carbonyl and carbonyl

oxide moieties in 15 with the C9 and C10 hydroxyls

should then lead to formation of the dioxabicyclo[3.3.0]-

octane 16. Although the carbonyl oxide and carbonyl

groups normally undergo intramolecular [3 + 2] cycload-

dition to afford a secondary ozonide, the carbonyl oxide

residues can be trapped intermolecularly by methanol to

form peroxyacetals. Schreiber et al. has used this strategy

for the desymmetrization of simple cycloalkenes.¹⁴More-

over, Criegee et al. has demonstrated the intramolecular

trapping of carbonyl oxides with tethered oxygen nucleo-

philes.¹⁵Our prediction that 16 would be formed upon

ozonolysis of 14 was predicated on this body of work. We

In route A, acidic hydrolysis of the enol ether mixture

8 is expected to lead to directly to the targeted subunit.

Subunit convergence to form 8 will be realized via Wittig

coupling of the C1-C13 aldehyde 9 with the ylide derived

from the C14-C22 R-methoxyphosphonium salt 12a. As

mentioned previously, a potential advantage of this route

is the preset configuration at C12, which is expected to

facilitate the conversion of 8 to 7. Our route B synthesis

employs enone 10 as the penultimate intermediate, which

is anticipated to arise via Horner-Wadsworth-Emmons

coupling between C1-C12 aldehyde 11 and the C13-

C22 â-ketophosphonate 12b. As with our earlier C1-C15

subunit synthesis, this route relies upon the late-stage

assembly of the C-ring via Michael addition of the C9-

hydroxyl with C12 of the enone moiety, with the at-

tendant issue of control at the C12 stereocenter prior to

ketal formation.

Route A: Synthesis of the C1-C22 Subunit 7 from

(+)-Conduritol E (27). Our retrosynthetic analysis of

the pivotal route A intermediate 9 is depicted in Scheme

3. Recognizing an element of local C₂-symmetry about

the C8-C11 region of the ABC-ring aldehyde 9, we

sought to preserve this substructure in our retrosynthetic

simplification. We arrived at the ^L-mannose-configured

γ-lactone 13, which contains the C8-C11 stereocenters

and features differentiated termini at C7 and C12 for

further elaboration. The dissymmetric nature of 13 was

(13) For syntheses of (+)-conduritol E via RCM, see: (a) Jørgensen,

M.; Iversen, E. H.; Paulsen, A. L.; Madsen, R. J. Org. Chem. 2001, 66,

4630-4634. (b) Ackermann, A.; Tom, D. E.; Fu¨rstner, A. Tetrahedron

2000, 56, 2195-2202. (c) Lee, W.-W.; Chang, S. Tetrahedron: Asym-

metry 1999, 10, 4473-4475. For other syntheses of (+)-conduritol E,

see: (d) Hudlicky, T.; Luna, H.; Olivo, H. F.; Andersen, C.; Nugent,

T.; Price, J. D. J. Chem. Soc., Perkin Trans. 1 1991, 2907-2917. (e)

Takano, S.; Yoshimitsu, T.; Ogasawara, K. J. Org. Chem. 1994, 59,

54-57. (f) Sanfilippo, C.; Patti, A.; Piattelli, M.; Nicolosi, G. Tetrahe-

dron: Asymmetry 1997, 8, 1569-1573. (g) Cere`, V.; Mantovani, G.;

Peri, F.; Pollicino, S.; Ricci, A. Tetrahedron 2000, 56, 1225-1231.

(14) Schrieber, S. L.; Claus, R. E.; Reagan, J. Tetrahedron Lett. 1982,

23, 3867-3870.

J. Org. Chem, Vol. 70, No. 23, 2005 9385

Lambert et al.

SCHEME 4

SCHEME 5

quent treatment of 21 with K₂CO₃in MeOH provided the

expected diol 14a in 96% yield.

With 14a in hand, the stage was now set for ozonolytic

desymmetrization. However, treatment of a solution of

14a in CH₂Cl₂(-78 °C) with O₃followed by quenching

with Ac₂O and Et₃N provided none of the expected

lactone. We reasoned that the desired γ-lactone product

may suffer decomposition in the presence of Et₃N via a

â-elimination pathway. Fortunately, the use of AcCl

(TsCl and MsCl were also effective) in conjunction with

a heterogeneous base (NaOAc) effected the desired

dehydration reaction without inducing subsequent prod-

uct decomposition, and lactone 13a was obtained in ca.

25% yield following treatment of the crude product

mixture with Amberlist H⁺resin in refluxing MeOH

(Scheme 5). The methanolysis step was necessary since

the methoxy acetals so obtained proved more stable to

purification by silica gel chromatography than the inter-

mediate hemiacetals. However, the bis(methoxy acetal)

22a was isolated as the major product of this reaction in

ca. 50% yield (ca. 21% yield of the C₂-symmetric isomer),

which pointed to a deleterious reduction pathway. This

side reaction temporarily foiled our goal of desymmetri-

zation, since the termini at C7 and C12 were rendered

equivalent with respect to oxidation state.

A plausible mechanistic rationale for the formation of

22a is depicted in Scheme 6. Since it is known that

carbonyl oxides can undergo dimerization followed by

extrusion of a molecule of O₂,²⁰we reasoned that a related

process might be occurring in our system. The peroxy-

acetal 15a, which contains a highly nucleophilic hydro-

peroxide residue, could compete with intramolecular

carbonyl oxide capture via an intermolecular process with

the hemiacetal-carbonyl oxide species 23. Since the

desired intramolecular closure results in the formation

of a strained bicyclic system, the intermolecular process

(which involves an inherently more potent nucleophile)

may have a comparable rate, thus allowing it to compete

effectively. This would result in the formation of the

geminal bis(alkylperoxide) dimer 24, which can lose O₂

as shown to afford the hydroxy aldehyde 25 and the bis-

(hemiacetal) 22b. Rapid closure of 25 to the hemiacetal

form would then give another molecule of 22b. Since O₂

have found only one isolated example in which these

cyclization and desymmetrization tactics have been ap-

plied in tandem.¹⁶Dehydration of the cyclic peroxyacetal

in 16 with Ac₂O/Et₃N in line with Schreiber’s method

would then give rise to lactone 13. In this way, the C8-

C11 stereochemical array of halichondrin B would be

assembled from a readily available C₂-symmetric precur-

sor. By virtue of its symmetry, the C8 and C11 stereo-

centers in 14 are homotopic, as are the C9 and C10

stereocenters. This reduces the problem of its synthesis

to one involving the formation of two asymmetric centers

instead of four. Desymmetrization of 14 effectively pro-

duces two new stereocenters by breaking this degeneracy,

resulting in a substantial increase in molecular complex-

ity.¹⁷

The synthesis of an appropriate (+)-conduritol E

derivative for this purpose is shown in Scheme 4. Subjec-

tion of dimethyl ^L-tartrate acetonide 17¹⁸to a one-pot

DIBAl-H reduction/Grignard addition sequence provided

diol 18 as a 67:23:10 mixture of diastereomers in 80%

yield.^13a-cBenzylation and acetonide hydrolysis was then

followed by acetylation to provide a chromatographically

separable mixture of diacetates, and the ^L-mannitol-

configured diene 19 was isolated as the major product

in 51% overall yield over the three steps. Slow addition

of the Grubbs second-generation ruthenium benzylidene

catalyst 20¹⁹(3 mol %) to a refluxing solution of 19 in

benzene induced ring-closing metathesis (RCM) to afford

the (+)-conduritol E derivative 21 in 98% yield. Subse-

(15) Criegee, R.; Banciu, A.; Keul, H. Chem. Ber. 1975, 108, 1642-

1654.

(16) Wu, H.-J.; Chao, C.-S.; Lin, C.-C. J. Org. Chem. 1998, 63, 7687-

7693.

(17) (a) Bertz, S. H. New J. Chem. 2003, 27, 860-869. (b) Bertz, S.

H. New J. Chem. 2003, 27, 870-879.

(18) Mash, E. A.; Nelson, K. A.; Van Deusen, S.; Hemperly, S. B.

Org. Synth. 1990, 68, 92-103.

(19) (a) Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Org. Lett.

1999, 1, 953-956. (b) Morgan, J. P.; Grubbs, R. H. Org. Lett. 2000, 2,

3153-3155. (c) Trnka, T. M.; Morgan, J. P.; Sanford, M. S.; Wilhelm,

T. E.; Scholl, M.; Choi, T.-L.; Ding, S.; Day, M. W.; Grubbs, R. H. J.

Am. Chem. Soc. 2003, 125, 2546-2558.

(20) (a) Flisza´r, S.; Gravel, D.; Cavalieri, E. Can. J. Chem. 1966,

44, 1013-1019. (b) Flisza´r, S.; Chylin˜ska, J. B. Can. J. Chem. 1967,

45, 29-31. (c) Flisza´r, S.; Chylin˜ska, J. B. Can. J. Chem. 1968, 46,

783-788. For a review on carbonyl oxides, see: (d) Bunnelle, W. H.

Chem. Rev. 1991, 91, 335-362.

9386 J. Org. Chem., Vol. 70, No. 23, 2005

Halichondrin B: Synthesis of the C1-C22 Subunit

SCHEME 6

SCHEME 8

employed for the hydrolysis of the acetonide group, the

use of the more chemoselective FeCl₃reagent was

required; however, the reaction could not be driven to

completion in this case. Moreover, these yields were

obtained only after recovering and resubjecting the

acetonide-protected bis(TBDPS) ether mixture to FeCl₃‚

SiO₂for a total of three cycles. Subjection of 26 to RCM

with 10 mol % of the Grubbs catalyst 20¹⁹in refluxing

benzene provided diol 14b in 55% yield along with 25%

of recovered 26. The increased steric bulk adjacent to the

vinyl groups apparently hinders the RCM reaction, as a

substantial portion of the starting material failed to react

even at high catalyst loadings. This setback, coupled with

the inefficiency of the acetonide cleavage step, ultimately

led us to develop a more satisfactory synthesis of 14b.

We were hopeful that the TBDPS groups could be

introduced after RCM had been achieved, which sug-

gested that (+)-conduritol E itself might be employed as

an intermediate. Although several syntheses of this

biologically active natural product have been reported,¹⁴

there has been much recent activity in the preparation

of conduritol derivatives (including the E-subtype) via

RCM.^13a-cAccordingly, (+)-conduritol E (27) was pre-

pared in four steps and 46% overall yield from diol 18

with only minor modification of the published procedures

(Scheme 8). Treatment of 27 with 2.5 equiv of TBDPSCl

in DMF in the presence of pyridine for 2 days at room

temperature provided the desired diol 14b in 83% yield,

accompanied by 16% of the undesired regioisomer 28,

which could be recycled to 27 with TBAF. Interestingly,

substitution of imidazole for pyridine greatly increased

the reaction rate; however, 14b and 28 were formed in a

ratio of ca. 1:1 in this case.

SCHEME 7

is lost, this fragmentation process does indeed represent

a net reduction and would account for the formation of

the bis(acetal) product 22a. Conducting the ozonolysis

under high dilution conditions in an attempt to thwart

this dimerization/reductive fragmentation pathway did

not noticeably improve the ratio of lactone (13) to bis-

(acetal) (22) products.

Although disappointed by the interference of this

unanticipated side reaction, we were nonetheless encour-

aged by the isolation of lactone 13a, which validated our

proposal that the desymmetrization of C₂-symmetric

cyclohexene diols typified by 14a could be achieved by

ozonolysis. We reasoned that an increase in steric bulk

adjacent to the reactive functionality might retard in-

termolecular processes such as the dimerization/reductive

fragmentation pathway that plagued our earlier studies.

The easiest way to achieve this would be to modify the

protecting groups at the C8 and C11 alcohols. Accord-

ingly, protection of diol 18 (Scheme 7) as its bis(TBDPS

ether) followed by acetonide cleavage with FeCl₃adsorbed

on silica gel²¹provided the diastereomerically pure diene

diol 26 in 23% yield after chromatography, with 47% of

recovered acetonide isolated as well. Since the TBDPS

ethers suffered cleavage with other acidic conditions

With ample quantities of 14b in hand, we turned our

attention once again to the ozonolytic desymmetrization

step. We were delighted to find that passing ozone

through a -78 °C solution of 14b in EtOAc followed by

addition of Ac₂O/Et₃N and warming to 0 °C resulted in

the formation of the desired lactone 13b as a ca. 2:1

mixture of anomers as evidenced by ¹H NMR analysis of

the crude product mixture (Scheme 9). Subsequent treat-

ment with TESCl/imidazole provided lactone 13c as a

single diastereomer in 80% yield. A small amount of the

bis(TES acetal) was also formed, indicating that the

dimerization/reductive fragmentation pathway had not

(21) Kim, K. S.; Song, Y. H.; Lee, B. H.; Hahn, C. S. J. Org. Chem.

1986, 51, 404-407.

J. Org. Chem, Vol. 70, No. 23, 2005 9387

Lambert et al.

SCHEME 9

SCHEME 10

been suppressed entirely (cf. 22b, Scheme 6). Another

noteworthy feature of this reaction is the stability of 13b

and 13c to Et₃N, demonstrating that the bulky TBDPS

groups effectively shield the base-sensitive γ-lactone from

decomposition.

Having differentiated the two termini of the ^L-manni-

tol-configured backbone, we sought to construct the

highly oxygenated C-ring of the natural product, thereby

setting the stereochemical configuration at C12 at an

early stage. We envisioned hydroboration-oxidation of

an exocyclic enol ether^12e,h,22derived from the C12 car-

bonyl in 13c as the most expedient method for achieving

this goal. Engagement of the enol ether double bond by

a borane reagent would be expected to occur on the

convex face of the bicyclic system, thereby providing the

desired 12R stereochemistry. The highly polarized nature

of such olefins leads to exceptionally high levels of

regioselectivity as well, leading to exclusive production

of the anti-Markovnikov net hydration product.

SCHEME 11

Treatment of lactone 13c with a slight excess of

dimethyltitanocene (Petasis’ reagent)²³in PhCH₃at 85

°C cleanly afforded enol ether 29 in quantitative yield

(Scheme 10). This compound could be purified on silica

gel without decomposition or isomerization of the double

bond to the endocyclic position, a process that plagued

the isolation of a similar exocyclic enol ether in our

work.^12e,hAlthough the steric bulk afforded by the

adjacent TBDPS ethers may explain the exceptional

stability of 29, another possible reason is that the

adjacent oxygen substituent at C11 destabilizes any

developing positive charge at C12, thereby attenuating

the reactivity of the double bond. Although we had

success with the Tebbe reagent²⁴in this transformation

as well, we found the preparation and handling of the

Petasis reagent^23bmore convenient for large-scale ap-

plications. Treatment of 29 with BH₃‚SMe₂followed by

the addition of H₂O₂and NaOH gave the desired C-ring

alcohol 30 in 72% yield accompanied by 13% of its C12

epimer 31. We found the use of bulkier borane reagents

unsuccessful, with catecholborane giving lower yields and

9-BBN failing to react with 29 altogether. The imposing

bulk of the TBDPS groups seems a likely cause of this

sluggish reactivity, as the reaction is complete within 1

h at room temperature with the comparatively small BH₃

reagent. The undesired C12 epimer 31 was recycled to

enol ether 29 by base-induced elimination of the derived

iodide.

While 29 proved reasonably stable to acidic media such

as silica gel, we encountered an acid decomposition

pathway for this compound during the course of this

work. Upon dissolution in CDCl₃(of somewhat poor

quality) that contained HCl, we observed the clean

formation of the acyclic hemiacetal 33 as a single dias-

1

tereomer by H NMR analysis, with presumed interme-

diacy of the endocyclic enol ether 32 (Scheme 11). After

concentration and chromatography on silica gel, the

hemiacetal collapsed with loss of TESOH to afford

aldehyde 34 in 96% yield. While acyclic hemiacetals are

not normally sufficiently stable to be observed, 33 may

be stabilized by an intramolecular hydrogen bond with

the furan oxygen. The acidic medium in which 33 was

formed may also prevent the dissociation of the hemiac-

etal. The fact that only one diastereomer of 33 was

observed spectroscopically suggests that the hemiacetal

stereochemistry was retained under these conditions.

Exposure of 29 to an excess of PPTS in CH₂Cl₂, on the

(22) Untersteller, E.; Xin, Y. C.; Sinay¨, P. Tetrahedron Lett. 1994,

35, 2537-2540.

(23) (a) Petasis, N. A.; Bzowej, E. I. J. Am. Chem. Soc. 1990, 112,

6392-6394. For the preparation of dimethyltitanocene, see: (b) Payack,

J. F.; Hughes, D. L.; Cai, D.; Cottrell, I. F.; Verhoeven, T. R. Org. Synth.

2002, 79, 19-26.

(24) (a) Tebbe, F. N.; Parshall, G. W.; Reddy, G. S. J. Am. Chem.

Soc. 1978, 100, 3611-3613. (b) Pine, S. H.; Zahler, R.; Evans, D. A.;

Grubbs, R. H. J. Am. Chem. Soc. 1980, 102, 3270-3272.

9388 J. Org. Chem., Vol. 70, No. 23, 2005

Halichondrin B: Synthesis of the C1-C22 Subunit

SCHEME 12

SCHEME 13

other hand, gave only a mixture of enol ether products,

of which 32 was a major component. This indicates that

fragmentation of the bicyclic system to give furan byprod-

ucts requires more strongly acidic conditions.

stereoselectivity observed for this reaction is consistent

with Kishi’s empirical rule,²⁶a rationale formulated to

explain the high levels of diastereofacial selectivity

exhibited by chiral allylic ethers in addition reactions.

It was possible to recycle 39 by oxidative cleavage

followed by hydrolysis of the intermediate formate ester

with Amberlist-IRA-400 (OH) resin, affording 36 in 70%

yield. The overall sequence from 36 to 38 represents a

net ring expansion in which the C7 carbinol unit has been

formally inserted in a stereoselective manner.

Attachment of a suitably functionalized C1-C5 chain

at C6 would simultaneously complete construction of the

B-ring and set the stage for A-ring closure. The Lewis-

acid-promoted C-glycosidation with an allylic silane

nucleophile constitutes an appropriate tool for this task

and has been employed by Kishi during his synthesis of

this fragment of the halichondrins.^9g,jAccordingly, we

sought to activate 38 as a glycosyl donor via acetylation

of the two hydroxyl groups. Treatment of 38 with Ac₂O/

DMAP/py²⁷at room temperature provided diacetate 40

in 68% yield as a single diastereomer (Scheme 14).

Approximately 30% of a monoacetate product mixture

was also isolated, which could not be converted to 40

upon resubjection to the reaction conditions. Fortunately,

use of the more powerful combination of Ac₂O/TMSOTf²⁸

gave 40 in 65% yield²⁹along with its C6 anomer, 41, in

35% yield after ca. 10 min at 0 °C. Apparently, the C7

hydroxyl of the â-hemiacetal 38 is quite difficult to

acetylate, as 41 was not isolated from the milder reaction

conditions.

The tentative C7 and C12 stereochemical assignments

were verified by NOE studies of pivalate 35, obtained in

87% yield upon treatment of 30 with PivCl/DMAP

(Scheme 12). Irradiation of the anomeric proton at C7

led to enhancement of the C12 methylene proton signals,

an observation that is only possible if these protons reside

within the concave face of the bicyclic framework. Since

we sought to carry the C12 pivaloate ester throughout

the synthesis, and the C7 hemiacetal had to be unmasked

prior to further elaboration, we did not normally isolate

35 except for the purpose of the NOE study. A one-pot

procedure was developed from 30 in which the pivalation

reaction mixture was simply quenched with HCl in

MeOH, thereby inducing chemoselective cleavage of the

TES group to provide hemiacetal 36 in 98% yield.

Having confirmed the correct stereochemistry at C12,

we turned our attention to the C7 hemiacetal moiety at

the other end of the molecule, which would serve as the

starting point for the assembly of the A- and B-rings.

Kishi and co-workers have developed an efficient protocol

for constructing the C1-C7 portion of halichondrin B

from a C7 hemiacetal,^9f,gand we sought to employ a

modification of this method for our purpose. Treatment

of 36 with an excess of (methoxymethylene)triphenylphos-

phorane at 0 °C followed by stirring the crude product

mixture in a biphasic mixture of THF/10% aq KOH

provided 37 in 87% yield (Scheme 13), with only trace

With anomeric epimers 40 and 41 in hand, the stage

was now set for the stereoselective installation of an

A-ring handle at C6 via C-glycosidation followed by

assembly of the complete AB-ring system. In the event,

exposure of 40 to BF₃‚OEt₂in the presence of Kishi’s

1

amounts of the corresponding Z isomer detected by H

NMR analysis. The aqueous base treatment was neces-

sary due to the observation that silyl ether migration had

occurred during the Wittig step. We found that KOH

induced equilibration of the crude product mixture to give

37, with only a trace of its silyl ether migration regio-

isomer remaining. Treatment of 37 with a catalytic

amount of OsO₄in aqueous acetone in the presence of

N-methylmorpholine-N-oxide (NMO)²⁵yielded hemiacetal

38 in 73% yield, which features the desired 7R stereo-

chemistry, along with 22% of its C7 epimer 39. The

(26) (a) Cha, J. K.; Christ, W. J.; Kishi, Y. Tetrahedron 1984, 40,

2247-2255. (b) Cha, J. K.; Christ, W. J.; Kishi, Y. Tetrahedron Lett.

1983, 24, 3943-3946. (c) Christ, W. J.; Cha, J. K.; Kishi, Y. Tetrahedron

Lett. 1983, 24, 3947-3950.

(27) (a) Steglich, W.; Ho¨fle, G. Angew. Chem., Int. Ed. Engl. 1969,

8, 981. For a review, see: (b) Ho¨fle, G.; Steglich, W.; Vorbru¨ggen, H.

Angew. Chem., Int. Ed. Engl. 1978, 17, 569-583.

(28) Procopiou, P. A.; Baugh, S. P. D.; Flack, S. S.; Inglis, G. G. A.

J. Org. Chem. 1998, 63, 2342-2347.

(25) (a) VanRheenen, V.; Cha, D. Y.; Hartley, W. M. Org. Synth.,

Coll. Vol. 1988, 6, 342. (b) VanRheenen, V.; Kelly, R. C.; Cha, D. Y.

Tetrahedron Lett. 1976, 17, 1973-1976.

(29) Recrystallization of 40 from heptane/CH₂Cl₂provided crystals

suitable for X-ray crystallographic analysis. See the Supporting

Information.

J. Org. Chem, Vol. 70, No. 23, 2005 9389

Lambert et al.

SCHEME 14

SCHEME 16

SCHEME 15

for another 16 h resulted in exclusive formation of 46 in

95% yield. A mixture (ca. 3:1) of 46 and the corresponding

3S isomer could be isolated in 94% yield when the

reaction was conducted in refluxing benzene for 16 h,

although no equilibration of the two was observed in this

case. Apparently, the retro-Michael addition step that

initiates the C3 epimerization requires the higher tem-

perature afforded by the use of refluxing toluene. In like

fashion, triol 45 was also converted to 46 in 95% yield,

with migration of the pivalate group to the C13 hydroxyl

occurring in this case. Although we had initially explored

A-ring assembly directly from 43 using Triton-B meth-

oxide, the C7 acetate proved resistant to cleavage in this

case, prompting us to remove the bulky TBDPS ethers

at C8 and C11 prior to cyclization.

To access the targeted C13 aldehyde 9, we had planned

to first mask the C8 and C11 alcohols in 46 followed by

pivalate cleavage and oxidation of the resultant C13 al-

cohol. We encountered a procedure,^30a,bhowever, whereby

a primary TMS or TES ether may be chemoselectively

oxidized to the corresponding aldehyde in the presence

of secondary TMS or TES ethers under Swern condi-

tions.^30cThis prompted us to modify our plans for

oxidation at C13 accordingly. Cleavage of the pivalate

ester in 46 was accomplished with a basic resin in

refluxing methanol to provide triol 47 in 93% yield, along

with 7% of recovered 46 (Scheme 16). It should be noted

that 1.64 g (5.15 mmol) of 47 has been prepared by the

present route. Subsequent treatment of 47 with an excess

of TESCl afforded the tris(TES ether) 48 in 91% yield.

Oxidation of 48 under standard Swern conditions^30a,cat

-40 °C proceeded cleanly, providing aldehyde 9 in 89%

yield, with recovery of unreacted 48 in 6% yield. No

products from oxidation at either C8 or C11 were

detected. Since 9 is an R-alkoxy aldehyde, and hence

prone to epimerization at C12 to a presumably more

stable product, it was purified by chromatography on

Florisil and used in subsequent reactions without delay.

allylic silane 42^9gprovided the â,γ-unsaturated ester 43

in 81% yield as a ca. 3:1 (E)- to (Z)-mixture of isomers

(Scheme 15). Although the analogous conversion of 41

to 43 proceeded sluggishly at room temperature, heating

the reaction mixture to 60 °C resulted in complete

conversion within 5 h, and 43 was isolated in 86% yield.

The lower reactivity of 41 relative to 40 is probably

stereoelectronic in nature, since the â-C6 acetate in 41

is equatorially disposed and hence more difficult to ionize

due to the absence of an anomeric effect and the lack of

anchimeric assistance by the C7 acetate. The construction

of the B-ring had now been completed, and all of the

carbons necessary for subsequent assembly of the fully

elaborated A-ring were in place. Cleavage of the TBDPS

ethers with TBAF buffered with AcOH followed by

acetate cleavage with a basic resin provided triol 44 in

74% yield for two steps, with 8% yield of the pivalate

migration product 45 also being isolated. Treatment of

44 with excess DBU in refluxing toluene resulted in

isomerization of the C3-C4 double bond to the R,â-

unsaturated isomer, which then underwent Michael

addition by the C7 hydroxyl. A mixture of 46 (desired)

and its C3 epimer was observed by TLC analysis upon

consumption of starting material (ca. 1 h). Equilibration

(30) (a) Rodr´ıguez, A.; Nomen, M.; Spur, B. W.; Godfroid, J. J.

Tetrahedron Lett. 1999, 40, 5161-5164. For a review on the oxidative

deprotection of silyl ethers, see: (b) Muzart, J. Synthesis 1993, 11-

27. (c) Omura, K.; Swern, D. Tetrahedron 1978, 34, 1651-1660.

9390 J. Org. Chem., Vol. 70, No. 23, 2005

Halichondrin B: Synthesis of the C1-C22 Subunit

SCHEME 17

SCHEME 18

We had now arrived at the point of investigating the

Wittig reaction of 9 with a suitable (R-methoxyalkyl)-

triphenylphosphorane.³¹This reaction would serve a dual

purpose, as subunit convergence would occur simulta-

neously with the installation of a C14-ketone surrogate

(in the form of a C13-C14 enol ether) as a prelude to

the ultimate cage-forming event. The use of (R-methoxy-

alkyl)triphenylphosphoranes as Wittig partners for frag-

ment coupling in complex natural products synthesis is,

to the best of our knowledge, unprecedented. Therefore,

we sought to explore its viability through the synthesis

of a model system that would incorporate a simple alkyl

chain at C14 in the cage product.

The preparation and use of a Wittig reagent appropri-

ate for this task is shown in Scheme 17. Treatment of

commercially available 1,1-dimethoxyhexane (49) with

31a

PPh₃and BF₃‚OEt₂

provided the (R-methoxyalkyl)-

triphenylphosphonium tetrafluoroborate salt 50 in 67%

yield after recrystallization from THF. Since (R-meth-

oxyalkyl)triphenylphosphoranes are known to be un-

stable at temperatures above -40 °C,^31btheir generation

from the corresponding phosphonium salts must be

conducted at or below this temperature. Accordingly,

deprotonation of 50 with n-BuLi at -40 °C resulted in

the formation of the blood-red (R-methoxyalkyl)triphen-

ylphosphorane 51. Cooling this solution to -78 °C fol-

lowed by addition of aldehyde 9 and warming to room

temperature resulted in the isolation of enol ethers 4E

and 4Z in 41% and 37% yields, respectively. Since these

sensitive compounds suffered decomposition on silica gel,

Florisil was again employed for their chromatographic

separation and purification.

All of the functionality required for formation of the

caged ketal is present in 4E and 4Z. Since hydrolysis of

the enol ether moiety will provide a C14 ketone, both 4E

and 4Z are potentially viable precursors to the desired

cage product. For this reason, the subsequent step in the

synthetic sequence was most conveniently conducted on

the mixture of geometrical isomers without chromato-

graphic purification. We were pleased to observe that

treatment of the crude Wittig product mixture with

p-TsOH in CH₂Cl₂-H₂O for 1 h followed by the addition

of MeOH provided the cage ketal 6 in 67% yield over two

steps from aldehyde 9 via the intermediate ketone 52

1

(Scheme 18). The H NMR data we obtained for 6 (vide

infra) were in excellent agreement with the literature

data for its enantiomer, ent-6, which was synthesized

by Aicher and Kishi in 1987.^9bIt was determined (TLC

analysis) that enol ether hydrolysis in the biphasic CH₂-

Cl₂-H₂O solvent system was complete within 1 h,

although silyl ether cleavage and subsequent ketalization

were not realized until the mixture was homogenized

with MeOH. The choice of conditions for this reaction was

critical, as the C11-C14 furans 53a and 53b, epimeric

at C10, were observed as side products with other acid

promoters. For example, 53a was isolated in ca. 30% yield

upon treatment of a 4Z/4E mixture with 52% aq HF in

CH₃CN (5:95 v/v), while the use of a biphasic aq HCl-

THF system with the same substrate resulted in forma-

tion of 53b and 6 in 21% and 41% yields, respectively.

In these cases, the desired ketal 6 was observed to form

alongside 53a/53b, and TLC analysis of the reaction

mixture revealed that the latter were slowly formed at

the expense of the former. A likely mechanism for furan

formation involves protonation of the C8 oxygen in 6 and

(31) (a) Tu¨ckmantel, W.; Oshima, K.; Utimoto, K. Tetrahedron Lett.

1986, 27, 5617-5618. (b) Coulsen, D. R. Tetrahedron Lett. 1964, 5,

3323-3326. (c) Ley, S. V.; Lygo, B.; Organ, H. M.; Wonnacott, A.

Tetrahedron 1985, 41, 3825-3836.

J. Org. Chem, Vol. 70, No. 23, 2005 9391

Lambert et al.

SCHEME 19

SCHEME 20

ionization to form the oxocarbenium ion 54. Loss of a

proton at C13 would give the endocyclic enol ether 55

which, upon protonation of the allylic C9 oxygen, would

suffer C-ring rupture via elimination to give the delocal-

ized oxocarbenium species 56. Loss of the C11 proton

would then give furan 53a. Epimerization at C10 to give

53b, in which the furanyl side chain is deployed equa-

torially on the tetrahydropyranyl ring, most likely pro-

ceeds via the intermediacy of a furanylic C10 carbocation.

This cage-to-furan pathway is suppressed under the

conditions given above (p-TsOH, CH₂Cl₂-H₂O; MeOH),

an observation initially noted by Hart et al. on their work

detailing the acid lability of the halichondrins.³²They

reported an analogous furan-forming process upon ex-

posure of homohalichondrin B (which contains the same

cage unit as halichondrin B) to various acids (TFA, CD₃-

OD; PPTS, CDCl₃). Their observation that no furan was

formed in the presence of a large excess of p-TsOH in

either CH₂Cl₂or CH₂Cl₂-H₂O prompted us to employ

these conditions.

The cage-to-furan transformation deserves further

comment. The C1-C15 cage model 3 (Scheme 1) does not

give rise to any of the corresponding furan 57 (Scheme

19) upon exposure to a variety of acids, an observation

made by both us^12gand Cooper and Salomon^10bin our

respective syntheses of 3. It is well-known that R-alkoxy

substituents destabilize carbenium ion species through

the inductive withdrawal of electron density, thereby

rendering the oxocarbenium ion derived from 3 less

stable than 54. The acid lability of 6 relative to that of 3

strongly suggests a correlation between the facility of

furan formation and C14-oxocarbenium ion stability.

gel chromatography or recrystallization. Without delay,

a THF solution of 12a in a base-washed round-bottom

flask was treated with a slight deficiency of n-BuLi at

-78 °C. We observed immediate formation of a blood-

red solution of the (R-methoxyalkyl)triphenylphospho-

rane 59 under these conditions. Subsequent addition of

aldehyde 9 in THF at -78 °C was followed by slow

warming to room temperature. Since we were unsuccess-

ful in isolating the intermediate enol ether 8 in pure form

by chromatography on Florisil, the crude mixture was

simply exposed to the hydrolysis/ketalization conditions

we had developed earlier for 4E/Z. We were delighted to

find that the expected C1-C22 subunit of halichondrin

B 7, which contains the complete ABCDEF-ring system

of the natural product, was formed to the exclusion of

any furan products in a two-step yield of 29%. It should

be noted that the transformation of 8 into cage ketal 7

requires four distinct reactions, including hydrolysis of

the enol ether and two silyl ethers, and a dehydrative

ketalization of the resulting keto diol. Thus, despite the

poorer efficiency relative to our model system (possibly

a consequence of the purity of 12a relative to that of 50),

overall efficiency of subunit convergence and cage as-

sembly was still acceptable. The overall yield of 7 from

(+)-conduritol E (27) was 4% over 18 linear steps.

Route B: Synthesis of the C1-C22 Subunit 7 from

R-^D-Glucoheptonic γ-Lactone (62). The fully elabo-

rated C1-C22 subunit (7) was also constructed as

depicted through Route B in Scheme 2. The retrosynthe-

sis for the formation of 11 is shown in Scheme 21.

Tris(TES)-protected aldehyde 11 is available from â,γ-

unsaturated ester 60, through construction of the A-ring,

followed by inversion at C11 and protecting group

manipulations. â,γ-Unsaturated ester 60 comes from 61,

available in two steps from commercially available and

relatively inexpensive R-^D-glucoheptonic γ-lactone (62)

Having established a reliable set of reaction conditions

for the introduction of the sensitive cage functionality,

we sought to apply this Wittig coupling/ketalization

strategy toward the synthesis of a complete C1-C22

subunit. This would entail the preparation of a C14-C22

fragment endowed with (R-methoxyalkyl)triphenylphos-

phonium salt functionality at the C14 terminus. We have

previously described the synthesis of the C14-C22

subunit 12c (Scheme 20), which bears an aldehyde at

C14, via two different symmetry-based routes.^12i,kTo

install the requisite phosphonium salt at C14, 12c was

first treated with PPTS in MeOH/HC(OMe)₃to provide

the dimethyl acetal 58 in 95% yield. Without further

31a

purification, 58 was treated with PPh₃/BF₃‚OEt₂

as

done previously with 49, affording the (R-methoxyalkyl)-

triphenylphosphonium tetrafluoroborate salt 12a in 80%

yield as a mixture of C14 epimers. This compound proved

somewhat moisture-sensitive, giving aldehyde 12c and

triphenylphosphine as decomposition products. Addition-

ally, 12a could not be rigorously purified either by silica

(32) Hart, J. B.; Blunt, J. W.; Munro, M. H. G. J. Org. Chem. 1996,

61, 2888-2890.

9392 J. Org. Chem., Vol. 70, No. 23, 2005

Halichondrin B: Synthesis of the C1-C22 Subunit

SCHEME 21

SCHEME 22

SCHEME 23

($0.07/g, Aldrich) through a ring expansion from the

γ-lactone to form the B-ring. This starting material is

extremely attractive for this synthesis, as it contains the

correct stereochemistry for the C8, C9, and C10 stereo-

centers and the opposite stereochemistry for C11. This

is also the same starting material that was used in our

initial synthesis of the C1-C15 subunit.^12gSeveral

problems arose in the previous synthesis using this

starting material, and modifications to address these

issues will be elaborated below.

Lactone 62 was regioselectively protected as the bis-

(pentylidene) ketal,^12b,33which left the C11 hydroxyl open

for orthogonal protection as TBS ether 61 (Scheme 22).

This fully protected lactone was then treated with

DIBAl-H at -78 °C to yield the crude hemiacetal, which

was treated with an excess of vinylmagnesium bromide³⁴

to give triol 63 after cleavage of the TBS group with

TBAF (85% yield over the three steps). The selective

formation of the C7 stereocenter is consistent with

observations on related reactions of carbohydrate deriva-

tives.³⁵

lowed by pinacol ring expansion of the corresponding diol

to form the B-ring. This method suffered from lack of

stereoselectivity and regioselectivity in the formation and

rearrangement of the diol, which thus resulted in low

yield of the target product. Additionally, the instability

of the R-alkoxyorganostannane and its R-hydroxystan-

nane precursor made that route unsuitable for supplying

enough material for further chemical studies. This cur-

rent sequence avoids those problems and forms the

B-ring in high yield and high stereocontrol.

With the B-ring now installed, attention turned to the

construction of the A-ring of halichondrin B. As predicted,

glycoside triacetate 64 undergoes axial C-allylation³⁷at

the C6 anomeric center with Kishi’s allylic silane 42^9g

and an excess of BF₃‚OEt₂in acetonitrile to yield â,γ-

unsaturated ester 60, as a mixture of E and Z isomers

in 70% yield (Scheme 23). The ratio of the E and Z

isomers could not be determined, but proved to be

inconsequential, as treatment of with benzyltrimethy-

lammonium (Triton B) methoxide^9bleads to concomitant

acetate cleavage, double bond migration, and intramo-

lecular Michael addition from the unprotected C7 hy-

droxyl group to the enoate moiety. This initially produces

a mixture of C3 epimers, which then equilibrates to the

more stable equatorial isomer 65, thus completing the

A-ring in excellent yield.

Ring expansion to form the B-ring was achieved

through ozonolysis of 63 to yield the corresponding

aldehyde, which was intramolecularly trapped with the

C10 hydroxyl to form the B-ring pyran.³⁶This was

followed by protection of the remaining hydroxyls as

acetates by treatment with acetic anhydride in pyridine

to yield glycoside triacetate 64 as the major diastereomer

1

by H NMR in 81% yield over the two steps.

In the early stages of our previously published syn-

thesis of the C1-C15 subunit from 61,^12gan R-alkoxy-

organolithium reagent was added to the γ-lactone, fol-

Once the A-ring had been assembled, the next step was

to invert the C11 hydroxyl to the desired halichondrin

stereochemistry (Scheme 24). The oxidation/reduction

protocol utilized by us developed previously was revisit-

ed.^12gOxidation of the C11 hydroxyl with TPAP/NMO³⁸

(33) Masamune, S.; Ma, P.; Okumoto, H.; Ellingboe, J. W.; Ito, Y.

J. Org. Chem. 1984, 49, 2834-2837.

(34) Mekki, B.; Singh, G.; Wightman, R. H. Tetrahedron Lett. 1991,

32, 5143-5146.

(35) (a) Lygo, B.; Swiatyj, M.; Trabsa, H.; Voyle, M. Tetrahedron

Lett. 1994, 35, 4197-4200. (b) Shing, T. K. M.; Elsley, D. A.; Gillhouley,

J. G. J. Chem. Soc., Chem. Commun. 1989, 1280-1282.

(36) Lay, H.; Lehmann, J.; Ziser, L.; Reutter, W. Carbohydr. Res.

1989, 195, 145-149.

(37) (a) Giannis, A.; Sandhoff, K. Tetrahedron Lett. 1985, 26, 1479-

1482. (b) Lewis, M. D.; Cha, J. K.; Kishi, Y. J. Am. Chem. Soc. 1982,

104, 4976-4978. (c) Kozikowski, A. P.; Sorgi, K. L. Tetrahedron Lett.

1982, 23, 2281-2284.

J. Org. Chem, Vol. 70, No. 23, 2005 9393

Lambert et al.

SCHEME 24

SCHEME 25

delivers ketone 66, which was followed by a chelation-

controlled reduction with Zn(BH₄)₂in pentane at 0 °C.

Our previous work^12gcalled for reaction at room temper-

ature, which yielded the inverted C11 alcohol with 14:1

stereoselectivity. Performing the reaction at 0 °C lead to

exclusive formation of the desired C11 epimer 67 in 98%

yield (two steps), with no inhibition of the reaction rate

(the reaction was done in 10 min), and no trace of 65 by

¹H NMR spectroscopy. The inverted alcohol was then

protected as the TBS ether (68) in excellent yield.

Installation of a terminal alkene as a C12 aldehyde

precursor preceded the C8, C9, C11 protecting group

switch and was synthesized from diol 69 using the

Corey-Winter protocol⁴⁰to yield alkene 71 via thiono-

carbonate 70 (Scheme 25).⁴¹

Protecting group removal was facilitated with acidic

DOWEX 50WX4-400 resin in refluxing MeOH yielding

70% of triol 72 after one resubjection of starting material.

This triol was then protected as the tris-TES ether (73)

in quantitative yield.

Ozonolysis of 73 at -78 °C in CH₂Cl₂and quenching

with PPh₃yielded aldehyde 11, which was concentrated,

redissolved in DME, and combined with the anion formed

from commercially available dimethyl-2-oxoheptyl phos-

phonate to form R,â-unsaturated ketone 74 in 64% yield

over the two steps (Scheme 26). Several other Horner-

Wadsworth-Emmons conditions were also attempted⁴²

but were either low yielding or led to decomposition.

The next step required selective cleavage of the ter-

minal pentylidene ketal, which was achieved utilizing

21

FeCl₃on SiO₂in CHCl₃and optimized to yield exclu-

sively 1,2 diol 69 in 90% yield after one recycle of 68.³⁹

Because of what we had learned about the acid-induced

decomposition of the CDE-ring cage ketal to a furan

(Scheme 18), it was clear that the protecting groups for

the C8, C9, and C11 hydroxyls in 69 were too robust. To

address this issue and to establish a C12 aldehyde for

subunit coupling, the conversion of 69 to 73 (Scheme 25)

was executed.

(40) Winter, R. A. E.; Corey, E. J. J. Am. Chem. Soc. 1963, 85, 2677-

2678.

(41) Gratifyingly, thionocarbonate 70 produced crystals suitable for

X-ray crystallography, which allowed us to unambiguously determine

the C11 stereochemistry. See the Supporting Information.

(42) (a) Blanchette, M. A.; Choy, W.; Davis, J. T.; Essenfeld, A. P.;

Masamune, S.; Roush, W. R.; Sakai, T. Tetrahedron Lett. 1984, 25,

2183-2186. (b) Paterson, I.; Keown, L. E. Tetrahedron Lett. 1997, 38,

5727-5730. (c) Rathke, M. W.; Nowak, M. J. Org. Chem. 1985, 50,

2624-2626. (d) Watanabe, R.; Kita, M.; Uemura, D. Tetrahedron Lett.

2002, 43, 6501-6504.

(38) Griffith, W. P.; Ley, S. V.; Whitcombe, G. P.; White, A. D. J.

Chem. Soc., Chem. Commun. 1987, 1625-1627.

(39) Use of 80% aqueous AcOH^12gled to extended reaction times

(20-24 h), some TBS cleavage, and lower yields (maximum yields were

75%, also after one recycle of recovered SM).

9394 J. Org. Chem., Vol. 70, No. 23, 2005

Halichondrin B: Synthesis of the C1-C22 Subunit

SCHEME 26

SCHEME 27

SCHEME 28

The final step for the production of the ABCDE-ring

model involved treatment of 74 under mildly acidic

conditions (p-TsOH in CH₂Cl₂/MeOH/H₂O) to afford cage

product 6 in 47% yield. All spectral data matched that

of 6 synthesized as shown in Scheme 18, and Kishi’s

subunit model ent-6.^9b

Once the model study was complete, coupling of the

ABC-ring and F-ring subunits followed by caged ketal

formation to complete the C1-C22 subunit was explored

(Scheme 27). Oxidation of F-ring alcohol 12d^12kwith

freshly prepared Jones’ reagent⁴³followed by esterifica-

tion with TMSCHN₂yielded methyl ester 75 in 73% over

the two steps. â-Ketophosphonate 12b was synthesized

in 93% yield by addition of the lithium anion of dimethyl

methylphosphonate to 75 in THF at -78 °C.

Aldehyde 11 was formed in the same manner as before

and was subjected to a quick purification on silica gel to

remove triphenylphosphine oxide, which we discovered

was interfering with recovery of excess 12b after the

Horner-Wadsworth-Emmons coupling step. This alde-

hyde was then coupled to F-ring â-ketophosphonate 12b

through a Horner-Wadsworth-Emmons reaction in

refluxing DME to yield enone 10 (Scheme 28). Unfortu-

nately, coupling of the C13-C22 subunit to the C1-C12

subunit proved to be problematic using the conditions

employed for the synthesis of enone 74.

amount; therefore, a more easily handled base, KHMDS,

was employed. Enone 10 was also found to be extremely

difficult to isolate, as an unknown byproduct constantly

coeluted with 10 during chromatography. Thus, enone

10 and this byproduct were carried on crude through the

Michael addition/ketalization cascade to form ketal 7 in

10% yield over the three steps. Again, the complexity of

this transformation deserves comment, in that three silyl

ether cleavages, a Michael addition, and a ketalization

constitute the conversion of 10 into 7. This sample of 7

was identical in all respects to the one synthesized via

the less problematic Route A as shown in Scheme 20.⁴⁴

The use of sodium hydride proved not to be feasible

due to the uncertainties in weighing out such a small

Efforts are underway to complete the total synthesis

of halichondrin B.

(43) Guanti, G.; Banfi, L.; Riva, R. Tetrahedron 1995, 51, 10343-

10360.

J. Org. Chem, Vol. 70, No. 23, 2005 9395

Lambert et al.

4E (27 mg, 0.042 mmol, 42%) were obtained as oils by FCC on

Florisil (5% to 10% to 20% EtOAc in hexanes). Data for 4Z:

R_f0.33 (20% EtOAc in hexanes); [R]²³_D-62 (c 1.1, CHCl₃); IR

Experimental Section

Hemiacetal 13b. A solution of diol 14b (11.6 g, 18.7 mmol)

in EtOAc (746 mL, 0.025 M) was cooled to -78 °C in a three-

necked round-bottom flask fitted with a gas inlet tube. The

solution was purged with ozone until a blue color was seen

(40 min) that persisted for 2-3 min. At this point the solution

was purged with N₂until the blue color dissipated. To the

resulting colorless solution was added Ac₂O (2.64 mL, 28.0

mmol, 1.5 equiv) and Et₃N (5.16 mL, 37.3 mmol, 2 equiv), and

the solution was warmed to 0 °C and stirred for 2 h. The

solution was quenched with water (300 mL) and warmed to

room temperature, where it stirred for 1 h. After extraction

with EtOAc (3×), the combined organic layers were washed

sequentially with 5% aq HCl and brine, dried over MgSO₄,

filtered, and concentrated in vacuo. The crude hemiacetal 13b

(ca. 2:1 mixture of anomers) was isolated as a white foam and

was used directly in the next reaction without further purifica-

tion. Data for 13b: R_f0.27 (20% EtOAc in hexanes); [R]²³_D-85

(thin film) 1743, 1653 cm^-^{1 1}H NMR (300 MHz, CDCl₃) δ

;

0.58-0.69 (m, 12 H), 0.87 (t, J ) 6.8 Hz, 3 H), 0.94-1.02 (m,

18 H), 1.15-1.55 (m, 8 H), 1.69-1.78 (m, 1 H), 2.03-2.28 (m,

3 H), 2.38 (ABX, J_ab) 15.2 Hz, J_ax) 4.9 Hz, 1 H), 2.55 (ABX,

J_ab) 15.2 Hz, J_ax) 8.2 Hz, 1 H), 2.96 (dd, J ) 9.4, 2.2 Hz, 1

H), 3.56 (s, 3 H), 3.68 (s, 3 H), 3.78-3.79 (m, 1 H), 3.93 (dd, J

) 5.2, 3.6, 1 H), 4.00 (dd, J ) 9.8, 5.3 Hz, 1 H), 4.06-4.17 (m,

2 H), 4.30 (dd, J ) 9.9, 5.2 Hz, 1 H), 4.45 (dd, J ) 9.3, 3.5 Hz,

1 H), 4.74 (d, J ) 9.4 Hz, 1 H); ¹³C NMR (75 MHz, CHCl₃) δ

5.1 (CH₂), 5.2 (CH₂), 7.1 (CH₃), 7.2 (CH₃), 14.0 (CH₃), 22.5

(CH₂), 27.8 (CH₂), 29.8 (CH₂), 30.2 (CH₂), 31.0 (CH₂), 31.6

(CH₂), 40.7 (CH₂), 51.6 (CH₃), 54.2 (CH₃), 64.8 (CH), 67.1 (CH),

73.6 (CH), 74.0 (CH), 74.9 (CH), 76.1 (CH), 77.4 (CH), 78.2

(CH), 93.8 (CH), 160.5 (C), 171.9 (C); HRMS calculated for

C₃₃H₆₂O₈Si₂(M + Na)⁺665.3881, found 665.3911 (5.0 ppm).

Data for 4E: R_f0.18 (20% EtOAc in hexanes); [R]²³-63 (c

D

(c 1.8, CHCl₃); IR (thin film) 3631-3134 (br), 1803 cm^-1; H

1

1.1, CHCl₃); IR (thin film) 1743, 1675 cm^-1; ¹H NMR (300 MHz,

CDCl₃) δ 0.57-0.68 (m, 12 H); 0.75-1.06 (m, 21 H), 1.24-

1.53 (m, 8 H), 1.68-1.80 (m, 1 H), 2.03-2.20 (m, 3 H), 2.38

(ABX, J_ab) 15.4 Hz, J_ax) 5.0 Hz, 1 H), 2.55 (ABX, J_ab) 15.2

Hz, J_ax) 8.2 Hz, 1 H), 2.94 (dd, J ) 9.3, 1.9 Hz, 1 H), 3.54 (s,

3 H), 3.68 (s, 3 H), 3.75-3.89 (m, 1 H), 3.97 (dd, J ) 5.3, 1.2

Hz, 1 H), 3.99 (dd, J ) 7.6, 5.4 Hz, 1 H), 4.04-4.17 (m, 2 H),

4.27 (dd, J ) 9.5, 5.5 Hz, 1 H), 4.78 (dd, J ) 9.3, 3.4 Hz, 1 H),

4.88 (d, J ) 8.9 Hz, 1 H); ¹³C NMR (75 MHz, CHCl₃) δ 5.0

(CH₂), 5.2 (CH₂), 7.0 (CH₃), 7.2 (CH₃), 13.9 (CH₃), 22.6 (CH₂),

26.7 (CH₂), 29.7 (CH₂), 30.2 (CH₂), 31.2 (CH₂), 31.7 (CH₂), 40.7

(CH₂), 51.6 (CH₃), 56.0 (CH₃), 64.8 (CH), 67.1 (CH), 74.0 (CH),

74.4 (CH), 75.2 (CH), 76.1 (CH), 77.5 (CH), 104.8 (CH), 158.2

(C), 171.9 (C); HRMS calculated for C₃₃H₆₂O₈Si₂(M + Na)⁺

665.3881, found 665.3911 (5.0 ppm).

NMR (300 MHz, CDCl₃) see S48; ¹³C NMR (75 MHz, CHCl₃)

see S49; HRMS calculated for C₃₈H₄₄O₆Si₂(M + Na)⁺675.2574,

found 675.2588 (2.1 ppm).

Aldehyde 9. To a solution of oxalyl chloride (0.263 mL of a

2 M solution in CH₂Cl₂, 0.526 mmol, 2.2 equiv) in CH₂Cl₂(1.2

mL) was added DMSO (0.075 mL, 1.1 mmol, 4.4 equiv) at -50

°C. After stirring for 10 min at this temperature, 48 (158 mg,

0.239 mmol) in CH₂Cl₂(1.2 mL) was added via cannula. The

mixture was warmed to -40 °C and stirred for 1 h, at which

point Et₃N (0.291 mL, 2.10 mmol, 8.8 equiv) was added. After

stirring at -40 °C for 5 min, the reaction mixture was warmed

to room temperature. The mixture was diluted with water (8

mL) and extracted with CH₂Cl₂. The combined organic extracts

were washed with brine, dried over Na₂SO₄, and concentrated

in vacuo. Chromatography of the crude product on Florisil

(10% to 20% EtOAc in hexanes) provided aldehyde 9 (116 mg,

0.213 mmol, 89%) as a colorless oil along with recovered 48

(10 mg, 0.015 mmol, 6%). Data for 9: R_f0.18 (20% EtOAc in

hexanes); [R]²³_D-124 (c 1.38, CHCl₃); IR (thin film) 2732, 1734

cm^-1; ¹H NMR (300 MHz, CDCl₃) δ 0.56-0.69 (m, 12 H), 0.92-

1.00 (m, 18 H), 1.26-1.51 (m, 2 H), 1.70-1.80 (m, 1 H), 2.05-

2.14 (m, 1 H), 2.41 (ABX, J_ab) 15.4 Hz, J_ax) 5.0 Hz, 1 H),

2.58 (ABX, J_ab) 15.4 Hz, J_bx) 8.25 Hz, 1 H), 3.01 (dd, J )

9.6, 2.1 Hz, 1 H), 3.69 (s, 3 H), 3.81-3.92 (m, 1 H), 4.06-4.15

(m, 2 H), 4.20-4.30 (m, 3 H), 4.45-4.50 (m, 1 H), 9.75 (d, J )

2.8 Hz, 1 H); ¹³C NMR (75 MHz, CHCl₃) δ 4.6 (CH₂), 5.0 (CH₂),

6.8 (CH₃), 7.0 (CH₃), 29.5 (CH₂), 30.1 (CH₂), 40.5 (CH₂), 51.6

(CH₃), 65.2 (CH), 67.3 (CH), 74.3 (CH), 74.4 (CH), 75.2 (CH),

76.1 (CH), 77.1 (CH), 85.4 (CH), 171.7 (C), 203.7 (CH); HRMS

calculated for C₂₆H₄₈O₈Si₂(M + Na)⁺567.2785, found 567.2773

(2.1 ppm).

ABCDE Cage Ketal 6, Route A. A crude mixture of 4Z

and 4E (ca. 1:1) was prepared as described above from

aldehyde 9 (35 mg, 0.064 mmol) and dissolved in CH₂Cl₂(2.1

mL). Water (0.5 mL) was added, followed by p-TsOH‚H₂O (12

mg, 0.063 mmol, 1 equiv). The biphasic mixture was stirred

vigorously at room temperature for 1 h before MeOH (1.5 mL)

was added. The mixture became homogeneous, and stirring

was continued at room temperature for another 2 h. The

reaction was quenched by the addition of saturated aqueous

NaHCO₃, and the reaction mixture was extracted with CH₂-

Cl₂. The combined organic layers were washed with brine,

dried over Na₂SO₄, and concentrated in vacuo. FCC (20% to

30% EtOAc in hexanes) provided the desired ketal 6 (17 mg,

0.043 mmol, 67%) as a yellow oil. Data for 6: R_f0.25 (50%

EtOAc in hexanes); [R]²³-46 (c 1.0, CHCl₃); IR (thin film)

D

1

1740, 1437, 1189, 1140, 1082 cm^-1; H NMR (500 MHz, CD₃-

OD) δ 0.92 (t, J ) 7.3 Hz, 3 H), 1.25-1.45 (m, 7 H), 1.46-1.55

(m, 1 H), 1.55-1.65 (m, 1 H), 1.69 (ddd, J ) 13.5, 11.2, 5.1

Hz, 1 H), 1.75-1.85 (m, 2 H), 1.98 (dd, J ) 13.2, 5.0 Hz, 1 H),

2.04 (d, J ) 13.2 Hz, 1 H), 2.45 (ABX, J_ab) 15.7 Hz, J_ax) 5.0

Hz, 1 H), 2.52 (ABX, J_ab) 15.7 Hz, J_bx) 7.8 Hz, 1 H), 2.96

(dd, J ) 9.5, 1.9 Hz, 1 H), 3.66 (s, 3 H), 3.79-3.86 (m, 1 H),

4.11 (dd, J ) 6.6, 4.3 Hz, 1 H), 4.17 (dd, J ) 6.6, 4.6 Hz, 1 H),

4.29 (app td, J ) 10.0, 4.8 Hz, 1 H), 4.33 (dd, J ) 4.0, 2.0 Hz,

1 H), 4.58 (app t, J ) 4.5 Hz, 1 H), 4.68 (app t, J ) 4.6 Hz, 1

H); ¹³C NMR (75 MHz, CD₃OD) δ 14.5 (CH₃), 24.0 (CH₂), 25.1

(CH₂), 31.3 (CH₂), 31.8 (CH₂), 33.3 (CH₂), 40.0 (CH₂), 41.4

(CH₂), 48.2 (CH₂), 52.3 (CH₃), 69.7 (CH), 75.4 (CH), 76.1 (CH),

76.2 (CH), 78.1 (CH), 79.6 (CH), 82.4 (CH), 83.8 (CH), 112.1

(C), 173.5 (C); HRMS calculated for C₂₀H₃₀O₇(M + Na)⁺

405.1889, found 405.1896 (1.728 ppm).

ABCDEF Cage Ketal 7, Route A. A solution of phospho-

nium salt 12a (29 mg, 0.037 mmol, 2 equiv) in THF (0.4 mL)

was cooled to -78 °C for the dropwise addition of n-BuLi (15

µL of a 2.3 M solution in hexanes, 0.035 mmol, 1.9 equiv). The

resulting blood-red solution was stirred at this temperature

for 10 min before aldehyde 9 (10 mg, 0.018 mmol, 1 equiv) in

THF (0.3 mL) was added via cannula. After stirring at -78

Enol ethers 4Z and 4E. A slurry of phosphonium salt 50

(112 mg, 0.242 mmol, 2.4 equiv) in THF (1 mL) was cooled to

-40 °C for the dropwise addition of n-BuLi (0.125 mL of a 1.8

M solution in hexanes, 0.222 mmol, 2.2 equiv). The resulting

blood-red solution was stirred at this temperature for 10 min

before being cooled to -78 °C, and aldehyde 9 (55 mg, 0.10

mmol) in THF (1 mL) was added via cannula. After stirring

at -78 °C for another 5 min, the mixture was warmed to room

temperature and quenched with brine (4 mL). The layers were

separated, and the aqueous later was extracted with Et₂O. The

combined ether layers were then dried over Na₂SO₄and

concentrated in vacuo. The residue was triturated with hex-

anes to separate the product from excess phosphonium salt

and triphenylphosphine oxide, and the combined hexane

washings were concentrated in vacuo. Although the crude

product could be used directly in the next step, analytically

pure samples of enol ethers 4Z (24 mg, 0.037 mmol, 37%) and

(44) The ¹H and ¹³C NMR data for 6 and 7 prepared via both routes

A and B described here (Schemes 18, 20, 26, and 28) were in excellent

agreement with that for halichondrin B itself.^1aSee the Supporting

Information.

9396 J. Org. Chem., Vol. 70, No. 23, 2005

Halichondrin B: Synthesis of the C1-C22 Subunit

°C for another 5 min, the mixture was warmed to room

temperature and quenched with brine (4 mL). The layers were

separated and the aqueous later was extracted with Et₂O. The

combined ether layers were then dried over Na₂SO₄and

concentrated in vacuo. The resultant residue was passed

through a plug of Florisil to give a crude enol ether mixture

(ca. 11 mg, ca. 0.011 mmol, ca. 62%). As further purification

was not feasible, this material was carried directly into the

subsequent acid hydrolysis procedure. To this mixture was

added water (0.1 mL) and CH₂Cl₂(0.4 mL) followed by

p-TsOH‚H₂O (ca. 2 mg, ca. 0.010 mmol). The biphasic mixture

was stirred vigorously at room temperature for 1 h before

MeOH (0.3 mL) was added. The mixture became homogeneous,

and stirring was continued at room temperature for another

3 h. The reaction was quenched by the addition of saturated

aqueous NaHCO₃, and the reaction mixture was extracted with

CH₂Cl₂. The combined organic layers were washed with brine,

dried over Na₂SO₄, and concentrated in vacuo. FCC (40% to

50% EtOAc in hexanes) provided the desired C1-C22 ketal 7

(3.7 mg, 0.0053 mmol, 29% from 9) as a yellow oil. Data for 7:

) 8.1 Hz, 6 H), 0.66 (q, J ) 8.2 Hz, 6 H), 0.92 (t, J ) 8.0 Hz,

9 H), 0.98 (t, J ) 8.1 Hz, 9 H), 0.99 (t, J ) 7.9 Hz, 9 H), 1.21-

1.34 (m, 9 H), 1.61-1.68 (m, 3 H), 1.80-1.85 (m, 1 H), 2.35

(ABX, J_ab) 15.3 Hz, J_ax) 8.4 Hz, 1 H), 2.50 (ABX, J_ab)15.3

Hz, J_bx) 4.9 Hz, 1 H), 2.56 (dt, J ) 7.4, 5.8 Hz, 2 H), 2.91 (dd,

J ) 9.6, 2.2 Hz, 1 H), 3.38 (app dt, J ) 9.9, 4.1 Hz, 1 H), 3.68

(s, 3 H), 3.81 (dd, J ) 7.1, 2.5 Hz, 2 H), 3.94 (dd, J ) 6.9, 3.5

Hz, 1 H), 4.00 (m, 1 H), 5.07 (dd, J ) 7.3, 3.0 Hz, 1 H), 6.28

(dd, J ) 16.1, 0.8 Hz, 1 H), 7.09 (dd, J ) 16.2, 7.5 Hz, 1 H);

¹³C NMR (75 MHz, CDCl₃) δ 4.7 (CH₂× 3), 5.1 (CH₂× 3), 5.3

(CH₂× 3), 6.9 (CH₃× 3), 6.99 (CH₃× 3), 7.04 (CH₃× 3), 13.9

(CH₃), 22.5 (CH₂), 24.1 (CH₂), 28.7 (CH₂), 30.5 (CH₂), 31.5

(CH₂), 39.8 (CH₂), 40.5 (CH₂), 51.6 (CH₃), 64.6 (CH), 69.9 (CH),

71.5 (CH₂), 72.2 (CH), 74.0 (CH), 78.7 (CH), 82.3 (CH), 130.3

(CH), 146.1 (CH), 171.7 (C), 200.8 (C); HRMS calculated for

C₃₈H₇₄O₈Si₃(M + Na)⁺765.4589, found 765.4626 (5 ppm).

ABCDE Cage Ketal 6, Route B. To a solution of enone

74 (35 mg, 0.047 mmol) in a mixture of CH₂Cl₂/water (0.5 mL/

0.25 mL) was added p-TsOH (8.96 mg, 0.047 mmol, 1 equiv),

and this was stirred at room temperature for 30 min, followed

by the addition of MeOH (0.25 mL). The reaction mixture was

stirred for 5 min and was then quenched by addition of

saturated aqueous NaHCO₃(ca. 2 mL), followed by extraction

with CH₂Cl₂. The combined organic extracts were dried over

Na₂SO₄, filtered, and concentrated in vacuo. FCC (10% to 20%

to 50% hexanes in EtOAc to 100% EtOAc) yielded desired

ABCDE ketal 6 (8.5 mg, 47%) as an oil. Data for 6 are given

above.

R, â-Unsaturated Enone 10. To a cooled (0 °C) orange-

colored solution of dry KHMDS (32 mg, 0.160 mmol, 3.6 equiv)

in DME (1 mL) was added â-ketophosphonate 12b (99.7 mg,

0.183 mmol, 4.1 equiv, dissolved in 1 mL of DME) via cannula.

An additional 1 mL of DME was used to assist in the transfer.

This mixture stirred at 0 °C for 30 min. In another flask,

aldehyde 11 (based on 28.5 mg, or 0.0462 mmol of alkene 73)

was dissolved in DME (0.5 mL) and transferred via cannula

to the reaction flask. An additional 0.5 mL of DME was used

to assist in the transfer. This was stirred at 0 °C for 10 min,

warmed to room temperature, and then heated at reflux for

10 h. The reaction was then cooled and quenched with addition

of water and extracted with EtOAc (4×) and CH₂Cl₂(2×). The

combined organic extracts were dried over Na₂SO₄, filtered,

and concentrated in vacuo. The desired enone could not be

separated from an impurity that coelutes, so FCC (10% to 20%

Et₂O in hexanes followed by 50% EtOAc in hexanes) gave a

mixture of enone 10 plus the unknown impurity (∼35% pure),

which was carried on without further purification.

R_f0.25 (50% EtOAc in hexanes); [R]²³-36 (c 0.50, CHCl₃);

D

IR (thin film) 3070, 2928, 2855, 1739, 1428, 1360, 1177, 1111,

1078, 996, 918 cm^-1; ¹H NMR (500 MHz, CD₃OD) δ 1.04 (s, 9

H), 1.32-1.42 (m, 2 H), 1.58-1.77 (m, 4 H), 1.80-1.92 (m, 4

H), 1.96-2.03 (m, 2 H), 1.97 (dd, J ) 13.3, 4.8 Hz, 1 H), 2.05

(d, J ) 13.0 Hz, 1 H), 2.24-2.31 (m, 1 H), 2.40 (dd, J ) 15.8,

5.1 Hz, 1 H), 2.51 (dd, J ) 15.8, 7.7 Hz, 1 H), 2.64-2.70 (m, 1

H), 2.95 (dd, J ) 9.6, 1.9 Hz, 1 H), 3.62 (s, 3 H), 3.75-3.83 (m,

2 H), 3.83-3.91 (m, 2 H), 4.01-4.07 (m, 1 H), 4.11 (J ) 6.3,

3.8 Hz, 1 H), 4.17 (dd, J ) 6.7, 4.3 Hz, 1 H), 4.28 (ddd, J )

9.7, 9.7, 4.4, 1 H), 4.33 (dd, J ) 3.7, 2.0 Hz, 1 H), 4.53-4.62

(m, 2 H), 4.59 (app t, J ) 4.6 Hz, 1 H), 4.67 (app t, J ) 4.6 Hz,

1 H), 4.89 (aq, J ) 1.8 Hz, 1 H), 4.99 (aq, J ) 2.1 Hz, 1 H),

7.38-7.46 (m, 6 H), 7.66-7.71 (m, 4 H); ¹³C NMR (125 MHz,

CD₃OD) δ 20.5 (C), 27.9 (CH₃× 3), 31.3 (CH₂), 31.6 (CH₂),

32.1 (CH₂), 36.6 (CH₂), 39.9 (CH₂), 40.0 (CH₂), 41.8 (CH₂), 48.8

(CH₂), 52.6 (CH₃), 62.4 (CH₂), 70.1 (CH), 75.7 (CH), 76.3 (CH),

76.5 (CH), 78.0 (CH), 78.5 (CH), 79.0 (CH), 79.8 (CH), 82.8

(CH), 84.2 (CH), 105.8 (CH₂), 112.1 (C), 129.3 (CH × 4), 131.31

(CH), 131.33 (CH), 135.5 (C), 135.6 (C), 137.18 (CH × 2), 137.20

(CH × 2), 153.8 (C), 173.4 (C); HRMS calculated for C₄₀H₅₂O₉-

Si (M + Na)⁺727.3278, found 727.3281 (0.41 ppm).

Aldehyde 11. Alkene 73 (60 mg, 0.093 mmol for synthesis

of enone 74; 28.5 mg, 0.0462 mmol for synthesis of enone 10)

was dissolved in CH₂Cl₂(3 mL) and cooled to -78 °C. Ozone

was bubbled through the reaction mixture for 5 min, and then

the mixture was purged with N₂for 3 min. Triphenylphosphine

(24 mg, 1.5 equiv for synthesis of enone 74; 12 mg, 1.5 equiv

for synthesis of enone 10) was added to the reaction mixture

at -78 °C, and then, the bath was removed and reaction

mixture was stirred at room temperature for 30 min. The

contents of the flask were concentrated under a stream of N₂

and then were put under high vacuum for 1 h. The resulting

aldehyde was used without further purification.

ABCDEF Cage Ketal 7, Route B. To a solution of crude

enone 10 from above in CH₂Cl₂(2 mL) and 5 drops of water

was added p-TsOH (4.8 mg, 0.025 mmol, 1 equiv), and the

reaction mixture was stirred for 5 min. MeOH (2 mL) was

added, and the reaction mixture was stirred at room temper-

ature for 2 h. The reaction was quenched by addition of

saturated aqueous NaHCO₃, followed by extraction with CH₂-

Cl₂(4×). The combined organic extracts were dried over Na₂-

SO₄, filtered, and concentrated in vacuo. FCC (50% EtOAc in

hexanes) yielded desired ketal 7 (3.4 mg, 10% from 73). Data

for 7 are given above.

R, â-Unsaturated Enone 74. To a suspension of dry NaH

(8.7 mg, 0.3627 mmol, 3.9 equiv), in DME (1 mL) was added

dimethyl-2-oxoheptyl phosphonate (0.077 mL, 0.372 mmol, 4

equiv). The reaction mixture bubbled slightly, and then went

clear. This was stirred at room temperature for 5 min with

gentle heating by a heating mantle. Crude aldehyde 11 (based

on 60 mg, or 0.093 mmol of alkene 73) from the previous

reaction was dissolved in DME (1 mL) and transferred via

cannula with an additional 1 mL of DME to assist in the

transfer. The reaction mixture was then heated at reflux for

3 h. After this time, the reaction was cooled and quenched by

addition of water, and the reaction mixture was extracted with

ether. The combined organic extracts were dried over MgSO₄,

filtered, and concentrated in vacuo. FCC (9% Et₂O in hexanes)

yielded enone 74 as an oil (44.2 mg, 64%). Data for 74: R_f

Acknowledgment. We thank the NIH [Grant CA

74394] for generous support of this research. An NIH

CBI Training Grant [5 T32 GM08505 (W.T.L.)] and an

Abbott Fellowship [133-4-CL09 (W.T.L.)] are also ac-

knowledged. The NIH [1S10 RR08389-01] and NSF

[CHE-9208463 and CHE-0342998] are acknowledged for

their support of the NMR facilities at the University of

Wisconsin-Madison Department of Chemistry. Dr. Ilia

Guzei is gratefully acknowledged for solving the crystal

structures for compounds 40 and 70. Mr. Brian S. Lucas

is acknowledged for performing high-field NMR experi-

ments on 6 and 7.

0.62 (50% hexanes in Et₂O); [R]²³-34.9 (c 0.00235, CHCl₃);

D

IR (thin film) 2954, 2876, 1743, 1676, 1630, 1081 cm^-1

;

¹H

NMR (300 MHz, CDCl₃) δ 0.57 (q, J ) 7.7 Hz, 6 H), 0.65 (q, J

J. Org. Chem, Vol. 70, No. 23, 2005 9397

Lambert et al.

Supporting Information Available: Description of gen-

eral procedures, additional experimental data, ¹H and ¹³C

NMR spectra for all new compounds, and crystallographic

information for 40 and 70. This material is available free of

charge via the Internet at http://pubs.acs.org.

JO051479M

9398 J. Org. Chem., Vol. 70, No. 23, 2005

Article Doi

Halichondrin B: Synthesis of the C1-C22 subunit

DOI: 10.1021/jo051479m

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jo051479m

Full text of DOI:10.1021/jo051479m

Products guided by the article

R&D Labs maybe for 352695-01-3

Relevant to this article

Hot Product