Total synthesis of bryostatin 2

Article abstract of DOI:10.1021/ja990860j

The total synthesis of the marine macrolide bryostatin 2 is described. The synthesis plan relies on aldol and directed reduction steps in order to construct the anti-1,3-diol array present in each of the principal subunits (A, B, and C). These fragments were coupled using a Julia olefination and subsequent sulfone alkylation. A series of functionalization reactions provided a bryopyran seco acid, which was macrolactonized under Yamaguchi conditions. Installation of the two enoate moieties took advantage of asymmetric phosphonate and aldol condensation strategies. Reduction of the C₂₀ ketone and simple protecting group operations then completed the synthesis of bryostatin 2. This flexible approach should provide access to a series of new analogues of this clinically important marine natural product.

Full text of DOI:10.1021/ja990860j

7540
J. Am. Chem. Soc. 1999, 121, 7540-7552
Total Synthesis of Bryostatin 2
David A. Evans,* Percy H. Carter,1 Erick M. Carreira,2a Andre´ B. Charette,2b
Joe1lle A. Prunet,2c and Mark Lautens2d
Contribution from the Department of Chemistry & Chemical Biology, HarVard UniVersity,
Cambridge, Massachusetts 02138
ReceiVed March 17, 1999
Abstract: The total synthesis of the marine macrolide bryostatin 2 is described. The synthesis plan relies on
aldol and directed reduction steps in order to construct the anti-1,3-diol array present in each of the principal
subunits (A, B, and C). These fragments were coupled using a Julia olefination and subsequent sulfone alkylation.
A series of functionalization reactions provided a bryopyran seco acid, which was macrolactonized under
Yamaguchi conditions. Installation of the two enoate moieties took advantage of asymmetric phosphonate and
aldol condensation strategies. Reduction of the C20 ketone and simple protecting group operations then completed
the synthesis of bryostatin 2. This flexible approach should provide access to a series of new analogues of this
clinically important marine natural product.
Introduction
Bryostatin 1 (1a) has captured the attention of the chemical,
biological, and medical communities since Pettit first reported
its structure determination in 1982.3 Following the discovery
of bryostatin 1, 17 other bryostatins have been isolated and fully
characterized (Figure 1). While some congeners (1b-d) differ
from bryostatin 1 (1a) only in the acyl residues appended to
the C7 and C20 hydroxyls, others (e.g., 1e and 2) embody
significant changes in the C19-C27 region of the bryostatin
nucleus. Our interest in the synthesis of bryostatin was motivated
by three circumstances: (i) the strict polyacetate origin of the
bryostatin macrolides offered an ideal forum for the extension
of our aldol-based approach to polypropionate natural products;4
(ii) the clinically relevant bryostatin 1 remains a scarce and
expensive commodity;5 and (iii) simpler analogues might prove
to be viable medicinal agents and useful biological probes. In
(1) This work is taken in part from the Ph.D. Thesis of P. H. Carter,
Harvard University, 1998.
(2) (a) Present address: Eidgeno¨ssischen Technische Hochschule, Labo-
ratorium Fu¨r Organische Chemie, ETH-Zentrum, Universitatstrasse 16, CH-
8092 Zu¨rich 91125, Switzerland. (b) Present address: De´partment de
Chimie, Universite´ de Montre´al, Montre´al, Que´bec, Canada H3C 3J7. (c)
Present address: Ecole Polytechnique, De´partement De Chimie, Laboratorie
De Synthese Organique, 91128 Palaiseau Cedex, France. (d) Present
address: Department of Chemistry, University of Toronto, Toronto, Ontario,
Canada M5S 3H6.
(3) Review: Pettit, G. R. The Bryostatins. In Progress in the Chemistry
of Organic Natural Products; Herz, W., Ed.; Springer-Verlag: New York,
1991; No. 57, pp 153-195.
(4) (a) Evans, D. A. Aldrichim. Acta 1982, 15, 23-32. (b) Evans, D.
A.; Kim, A. S. (S)-4-Benzyl-2-oxazolidinone. In Encyclopedia of Reagents
for Organic Synthesis; Paquette, L. A., Editor-in-Chief; John Wiley and
Sons: New York, 1995; Vol. 1, pp 345-356. (c) Kim, A. S. Ph.D. Thesis,
Harvard University, 1997.
Figure 1. Representative members of the bryostatin macrolides.
this study, we describe the details6 of our synthesis of bryostatin
2 (1b), a natural congener7 and synthetic precursor8 of 1a.
Structure. Bryostatin 1 was first isolated in 1968 from the
bryozoan Bugula neritina (Linnaeus), an innocuous filter feeder
with a propensity for attaching to ship hulls and boat docks.9
The structure of bryostatin 1 was determined through single-
crystal X-ray diffraction analysis (Figure 2) and confirmed using
1
a combination of H NMR, 13C NMR, IR, and high-resolution
(6) This work has been communicated in preliminary form: Evans, D.
A.; Carter, P. H.; Carreira, E. M.; Prunet, J. A.; Charette, A. B.; Lautens,
M. Angew. Chem., Int. Ed. 1998, 37, 2354-2359.
(7) Pettit, G. R.; Herald, C. L.; Kamano, Y.; Gust, D.; Aoyagi, R. J.
Nat. Prod. 1983, 46, 528-531.
(8) Pettit has previously described the conversion of bryostatin 2 (1b)
into bryostatin 1 (1a), cf.: Pettit, G. R.; Sengupta, D.; Herald, C. L.; Sharkey,
N. A.; Blumberg, P. M. Can. J. Chem. 1991, 69, 856-860.
(9) (a) Pettit, G. R.; Day, J. F.; Hartwell, J. L.; Wood, H. B. Nature
1970, 227, 962-963. (b) Pettit, G. R.; Herald, C. L.; Doubek, D. L.; Herald,
D. L.; Arnold, E.; Clardy, J. J. Am. Chem. Soc. 1982, 104, 6846-6848.
(5) (a) Extractive isolation proceeds in ca. 1.3 × 10-6% yield (18 g of
bryostatin from 14 000 kg of animal) on industrial scale. Schaufelberger,
D. E.; Koleck, M. P.; Beutler, J. A.; Vatakis, A. M.; Alvarado, A. B.;
Andrews, P.; Marzo, L. V.; Muschik, G. M.; Roach, J.; Ross, J. T.; Lebherz,
W. B.; Reeves, M. P.; Eberwein, R. M.; Rodgers, L. I.; Testerman, R. P.;
Snader, K. M.; Forenza, S. J. Nat. Prod. 1991, 54, 1265-1270. (b) The
current market price for bryostatin 1 is $1156 (U.S.)/500 µg: LC
Laboratories, Woburn, MA. (c) Marine aquaculture methods are currently
being implemented in an attempt to solve the supply problem; cf. New
Scientist 1996, 14 Sept, 38-42.
10.1021/ja990860j CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/07/1999

Total Synthesis of Bryostatin 2
J. Am. Chem. Soc., Vol. 121, No. 33, 1999 7541
ultrapotent (picomolar) binding to the phorbol ester binding
site,18 and disruption of phorbol ester-induced tumor promo-
tion.19 These effects are noteworthy for two reasons: (i)
antineoplastic agents typically antagonize the immune and
hematopoietogenic systems,17 but bryostatin potentiates them;
and (ii) exogenous agonists of PKC, such as the phorbol esters,
are usually tumor promoters, but bryostatin acts as an antitumor
agent.20
The combined antitumor, immunopotentiating, and hemato-
potentiating properties of bryostatin 1 made it a promising
candidate for clinical trials, and the results from the Phase I
study (which began in early 1991) have demonstrated that
bryostatin 1 may be administered safely (muscle soreness was
the dose-limiting toxicity) and can induce a tumor response with
a wide range of tumors.21 More recent tests have shown that
bryostatin may synergize with both tamoxifen (in vitro) and
paclitaxel (in vivo),22 and future clinical trials will probably
assess the use of these combination therapies. The U.S. National
Cancer Institute has moved bryostatin into Phase II trials against
non-Hodgkin’s lymphoma, melanoma, and renal cancer.14
Synthesis Plan. The unusual structure, potent biological
activity, and relative scarcity of the bryostatins make them
attractive targets for total synthesis.23 Accordingly, several
groups,23,24 including our own,25 have initiated programs directed
toward the total synthesis of the bryostatins. Most notable among
these efforts is the total synthesis of bryostatin 7 reported by
the Masamune group.26
Figure 2. X-ray crystal structure of bryostatin1.9b
mass spectrometry.9b,10 The absolute configuration of bryostatin
1 was tentatively assigned on the basis of the small anomalous
dispersion effects of oxygen and carbon in the crystal structure9b,11
and later confirmed via analysis of the heavy-atom dispersion
effects in the X-ray structure of a p-bromobenzoate derivative.12
As can be seen in Figure 2, bryostatin 1 displays several
distinctive architectural features. The macrolide houses three
pyran rings (two of which are hemiketals), a 20-membered
lactone, two exocyclic R,â-unsaturated esters (C13, C21), one
trans olefin (C16-C17), and an unusual octadienyl ester side
chain (C20). The three-dimensional structure is organized about
the C3 hydroxyl group, which serves as a hydrogen bond
acceptor for the C19 lactol (O3-H-O19 distance of 2.71 Å) and
a bifurcated hydrogen bond donor for the O5 and O11 ether
oxygens (O3-H-O distances of 3.00 and 2.84 Å, respectively).9b
The solution structure of the C20-desoxymacrolide bryostatin
10 (1e) has been determined recently using 2D-NMR ROESY
technique and shows close homology with the solid-state
structure depicted in Figure 2.13 This well-defined tertiary
structure has important ramifications for the solution reactivity
of the bryostatins (vide infra).
Our plan for the synthesis of bryostatin 1 is outlined in
Scheme 1. Initial scission (transform T1) of the O25-C1 lactone
bond (macrocyclization transform27), followed by further dis-
connection at the C9-C10 C-glycoside bond (sulfone alkylation
and hydrolysis28) and the C16-C17 trans olefin (Julia-Lythgoe
(17) (a) Tallant, E. A.; Smith, J. B.; Wallace, R. W. Biochim. Biophys.
Acta 1987, 929, 40-46. (b) May, W. S.; Sharkis, S. J.; Esa, A. H.; Gebbia,
V.; Kraft, A. S.; Pettit, G. R.; Sensenbrenner, L. L. Proc. Natl. Acad. Sci.
U.S.A. 1987, 84, 8483-8487. (c) Grant, S.; Pettit, G. R.; McCrady, C. Exp.
Hematol. 1992, 20, 34-42. (d) Leonard, J. P.; May, W. S.; Ihle, J. N.;
Pettit, G. R.; Sharkis, S. J. Blood 1988, 72, 1492-1496. (e) Sharkis, S. J.;
Jones, R. J.; Bellis, M. L.; Demetri, G. D.; Griffin, J. D.; Civin, C.; May,
W. S. Blood 1990, 76, 716-720. (f) McCrady, C. W.; Staniswalis, J.; Pettit,
G. R.; Howe, C.; Grant, S. Br. J. Haematol. 1991, 77, 5-15.
(18) (a) Smith, J. B.; Smith, L.; Pettit, G. R. Biochem. Biophys. Res.
Commun. 1985, 132, 939-945. (b) Berkow, R. L.; Kraft, A. S. Biochem.
Biophys. Res. Commun. 1985, 131, 1109-1116. (c) DeVries, D. J.; Herald,
C. L.; Pettit, G. R.; Blumberg, P. M. Biochem. Pharm. 1988, 37, 4069-
4073.
Biological Activity. The bryostatins are a family of potent
antitumor agents,14 and bryostatin 1 exhibits significant in vivo
antineoplastic activity against murine leukemia, B-cell lym-
phoma, reticulum cell sarcoma, ovarian carcinoma, and mela-
noma.3,14,15 The bryostatins also display a diverse range of other
biological effects in vitro and in vivo, including stimulation of
T-cells and the immune system,16 stimulation of the hemato-
poietic system,17 activation of protein kinase C (PKC) through
(19) Hennings, H.; Blumberg, P. M.; Pettit, G. R.; Herald, C. L.; Shores,
R.; Yuspa, S. Carcinogenesis 1987, 8, 1343-1346.
(20) It should be noted that 12-deoxy-13-acyl phorbol esters also bind
to PKC and induce an antitumor response. As with the bryostatins, a
mechanism for this discrepancy is under investigation, cf.: Szallasi, Z.;
Krsmanovi, L.; Blumberg, P. M. Cancer Res. 1993, 53, 2507-2512.
(21) (a) Prendiville, J.; Crowther, D.; Thatcher, N.; Woll, P. J.; Fox, B.
W.; McGown, A.; Testa, N.; Stern, P.; McDermott, R.; Potter, M.; Pettit,
G. R. Br. J. Cancer 1993, 68, 418-424. (b) Philip, P. A.; Rea, D.; Thavasu,
P.; Carmichael, J.; Stuart, N. S. A.; Rockett, H.; Talbot, D. C.; Ganesan,
T.; Pettit, G. R.; Balkwill, F.; Harris, A. L. J. Natl. Cancer Inst. 1993, 85,
1812-1818. (c) Scheid, J.; Prendiville, J.; Jayson, G.; Crowther, D.; Fox,
B.; Pettit, G. R. Cancer Immunol. Immunother. 1994, 39, 223-230. (d)
Jayson, G. C.; Crowther, D.; Prendiville, J.; McGown, A. T.; Scheid, C.;
Stern, P.; Young, R.; Brenchley, P.; Chang, J.; Owens, S.; Pettit, G. R. Br.
J. Cancer 1995, 72, 461-468. (e) Varterasian, M. L.; Mohammad, R. M.;
Eilender, D. S.; Hulburd, K.; Rodriguez, D. H.; Pemberton, P. A.; Pluda, J.
M.; Dan, M. D.; Pettit, G. R.; Chen, B. D.; Al-Katib, A. M. J. Clin. Oncol.
1998, 16, 56-62. (f) Grant, S.; Roberts, J.; Poplin, E.; Tombes, M. B.;
Kyle, B.; Welch, D.; Carr, M.; Bear, H. D. Clin. Cancer Res. 1944, 4,
611-618.
(10) For a revision of the 1H and 13C NMR assignments of bryostatin 1,
cf.: Schaufelberger, D. E.; Chmurny, G. N.; Koleck, M. P. Magn. Reson.
Chem. 1991, 29, 366-374.
(11) Engel, D. W. Acta Crystallogr., Sect. B 1972, B28, 1498-1509.
(12) Pettit, G. R.; Herald, D. L.; Gao, F.; Sengupta, D.; Herald, C. L. J.
Org. Chem. 1991, 56, 1337-1340.
(13) Kamano, Y.; Zhang, H.; Morita, H.; Itokawa, H.; Shirota, O.; Pettit,
G. R.; Herald, D. L.; Herald, C. L. Tetrahedron 1996, 52, 2369-2376.
(14) Reviews: (a) Kraft, A. S. J. Natl. Cancer Inst. 1993, 85, 1790-
1792. (b) Pettit, G. R. J. Nat. Prod. 1996, 59, 812-821. (c) Stone, R. M.
Leukemia Res. 1997, 21, 399-401. (d) Information on the status of
bryostatin clinical trials may be accessed via the Internet at http://
cancernet.nci.nih.gov/prot/protsrch.shtml.
(15) Schuchter, L. M.; Esa, A. H.; May, W.; Laulis, M. K.; Pettit, G.
R.; Hess, A. D. Cancer Res. 1991, 51, 682-687.
(16) (a) Trenn, G.; Pettit, G. R.; Takayama, H.; Hu-Li, J.; Sitkovsky,
M. V. J. Immunol. 1988, 140, 433-439. (b) Hess, A. D.; Silanksis, M. K.;
Esa, A. H.; Pettit, G. R.; May, W. S. J. Immunol. 1988, 141, 3263-3269.
(c) Tuttle, T. M.; Inge, T. H.; Bethke, K. P.; McCrady, C. W.; Pettit, G. R.;
Bear, H. D. Cancer Res. 1992, 52, 548-553. (d) Mohr, H.; Pettit, G. R.;
Plessing-Menze, A. Immunobiology 1987, 175, 420. (e) Drexler, H. G.;
Gignac, S. M.; Pettit, G. R.; Hoffbrand, A. V. Eur. J. Immunol. 1990, 20,
119-127.
(22) (a) McGown, A. T.; Jayson, G.; Pettit, G. R.; Haran, M. S.; Ward,
T. H.; Crowther, D. Br. J. Cancer 1998, 77, 216-220. (b) Koutcher, J. A.;
Matei, C.; Zakian, K.; Ballon, D.; Schwartz, G. K. Proc. Am. Assoc. Cancer
Res. 1998, 39, 191 (No. 1304).
(23) For a detailed review of the literature prior to early 1995, cf.:
Norcross, R. D.; Paterson, I. Chem. ReV. 1995, 95, 2041-2114.

7542 J. Am. Chem. Soc., Vol. 121, No. 33, 1999
EVans et al.
Scheme 1
olefination29), provided three subunits of comparable complexity.
Although each of these fragments could be readily synthesized,
the presence of the exocyclic enoate moieties complicated the
indicated fragment assemblage operations (vide infra). Accord-
ingly, we revised our plan in order to accommodate a late-stage
introduction of these groups (transform T2) while maintaining
the same fragment coupling strategy (transform T1*). Each of
these strategies will be discussed in some detail in the text.
Inspection of these fragments (Scheme 1) reveals that they each
contain a 1,3-anti diol unit, a fact which is consistent with the
polyacetate-based biosynthesis of the bryostatins.30 The presence
of this stereoregular array served as the impetus for the
development of two different stereoselective â-hydroxyketone
reductions in our laboratory (eqs 1 and 2),31 both of which have
been employed in the synthesis of a number of polyacetate and
polypropionate natural products. In parallel with these studies,
Both of the aforementioned approaches to the bryostatins
utilize the same set of acyclic precursors (transform T3).
(24) Recent work: (a) Kiyooka, S.; Maeda, H. Tetrahedron: Asymmetry
1997, 8, 3371-3374. (b) Weiss, J. M.; Hoffmann, H. M. R. Tetrahedron:
Asymmetry 1997, 8, 3913-3920. (c) DeBrabander, J.; Kulkarni, B. A.;
Garcia-Lopez, R.; Vandewalle, M. Tetrahedron: Asymmetry 1997, 8, 1721-
1724. (d) Lampe, T. F. J.; Hoffmann, H. M. R. Tetrahedron Lett. 1996, 37,
7695-7698. (e) Lampe, T. F. J.; Hoffmann, H. M. R. Chem. Commun.
1996, 1931-1932. (f) De Brabander, J.; Vandewalle, M. Pure Appl. Chem.
1996, 68, 715-718. (g) Kalesse, M.; Eh, M. Tetrahedron Lett. 1996, 37,
1767-1770. (h) Ohmori, K.; Suzuki, T.; Nishiyama, S.; Yamamura, S.
Tetrahedron Lett. 1995, 36, 6515-6518. (i) Ohmori, K.; Nishiyama, S.;
Yamamura, S. Tetrahedron Lett. 1995, 36, 6519-6522. (j) Hoffmann, R.
W.; Stiasny, H. C. Tetrahedron Lett. 1995, 36, 4595-4598.
(25) (a) Evans, D. A.; Gauchet-Prunet, J. A.; Carreira, E. M.; Charette,
A. B. J. Org. Chem. 1991, 56, 741-750. (b) Evans, D. A.; Carreira, E. M.
Tetrahedron Lett. 1990, 31, 4703-4706. (c) Evans, D. A.; Carreira, E. M.;
Charette, A. B.; Gauchet, J. A. Abstracts of Papers, 196th National Meeting
of the American Chemical Society, Los Angeles, CA, Fall 1998; American
Chemical Society: Washington, DC, 1988; ORGN 209.
(26) (a) Kageyama, M.; Tamura, T.; Nantz, M. H.; Roberts, J. C.; Somfai,
P.; Whritenour, D. C.; Masamune, S. J. Am. Chem. Soc. 1990, 112, 7407-
7408. (b) Blanchette, M. A.; Malamas, M. S.; Nantz, M. H.; Roberts, J. C.;
Somfai, P.; Whritenour, D. C.; Masamune, S.; Kageyama, M.; Tamura, T.
J. Org. Chem. 1989, 54, 2817-2825. (c) Masamune, S. Pure Appl. Chem.
1988, 60, 1587-1596.
(27) (a) Bartra, M.; Urp´ı, F.; Vilarrasa, J. In Recent Progress in the
Chemical Synthesis of Antibiotics and Related Microbial Products; Lukacs,
G., Ed.; Springer-Verlag: Berlin, 1993; Vol. 2. (b) Boeckman, R. K.;
Goldstein, S. W. In The Total Synthesis of Natural Products; Apsimon, J.,
Ed.; Wiley: New York, 1988; Vol. 7, pp 1-139. (c) Paterson, I.; Mansuri,
M. M. Tetrahedron 1985, 41, 3569-3624.
we have developed two general strategies for the enantioselec-
tive synthesis of â-hydroxyketones, and representative examples
of each are shown below (eqs 3 and 4).25a,32 The first approach
(eq 3) relies on Sharpless asymmetric epoxidation33 (or epoxi-
dation with kinetic resolution) in order to establish the absolute
asymmetry of an epoxystyrene precursor such as 3. Subsequent
protection, Birch reduction, and ozonolysis affords the â-hy-
droxyketoester. The second approach (eq 4) depends on aldol
methodology to incorporate the absolute stereochemistry in
â-hydroxyketoester 7. Although the illustrated reaction derives
its asymmetry from the chiral copper(II) complex 6,31 variants
of this procedure that utilize chiral methyl ketones34 and/or chiral
aldehydes35 to establish the stereochemistry of the â-hydrox-
yketone have been documented.
(30) Kerr, R. G.; Lawry, J.; Gush, K. A. Tetrahedron Lett. 1996, 37,
8305-8308.
(28) (a) Ley, S. V.; Lygo, B.; Sternfeld, F.; Wonnacott, A. Tetrahedron
1986, 42, 4333-4342. (b) For a description of the development of this
methodology, cf.: Ley, S. V.; Armstrong, A. In Strategies and Tactics in
Organic Synthesis; Lindberg, T., Ed.; Academic Press: San Diego, 1991;
Vol. 3, pp 237-292.
(29) (a) Julia, M.; Paris, J. M. Tetrahedron Lett. 1973, 4833-4836. (b)
Kocienski, P. J.; Lythgoe, B.; Ruston, S. J. Chem. Soc., Perkin Trans. 1
1978, 829. (c) Review: Kocienski, P. Phosphorus Sulfur 1985, 24, 97-
127.
(31) (a) Evans, D. A.; Chapman, K. T.; Carreira, E. M. J. Am. Chem.
Soc. 1988, 110, 3560-3578. (b) Evans, D. A.; Hoveyda, A. H. J. Am. Chem.
Soc. 1990, 112, 6447-6449.
(32) Evans, D. A.; Kozlowski, M. C.; Murry, J. A.; Burgey, C. S.;
Campos, K. R.; Connell, B. T.; Staples, R. J. J. Am. Chem. Soc. 1999, 121,
669-685.
(33) Gao, R.; Hanson, J.; Klunder, J. M.; Ko, S. Y.; Masamune, H.;
Sharpless, K. B. J. Am. Chem. Soc. 1987, 109, 5765-5780.

Total Synthesis of Bryostatin 2
J. Am. Chem. Soc., Vol. 121, No. 33, 1999 7543
aldol 11 as a 9:1 mixture of diastereomers. Removal of the
halide auxiliary was effected with zinc dust in glacial acetic
acid37 and afforded diastereomerically pure â-hydroxyimide 12
after silica gel chromatography. Subsequent reduction of the
imide with lithium borohydride40 gave the enantiopure diol 13
in 67% overall yield for the three steps and provided a 71%
yield of recovered (4S)-4-phenylmethyl-2-oxazolidinone. The
synthesis of aldehyde 15 was then completed in 90% yield with
a simple three-step sequence of acetalization, regioselective
reductive acetal cleavage (DIBAl-H, 0 °C),41 and hydroxyl
oxidation.
The aldol coupling of bis(trimethylsilyl)dienol ether 5b with
aldehyde 15 was only modestly stereoselective under standard
conditions (MgBr2‚OEt2, BF3‚OEt2). While strong Lewis acids
(TiCl4, TiCl3(Oi-Pr), SnCl4) did not effect a clean reaction with
15, the use of the mixed titanium species TiCl2(Oi-Pr)2 (CH2-
Cl2, -78 °C) delivered a high-yielding, stereoselective reaction
(93% yield, 6:1 diastereoselection).42 The selectivity for this
reaction could be improved with a modest reduction in chemical
yield by using toluene as the reaction solvent (83% yield, 15:1
diastereoselection), a result which is consistent with the opera-
tion of electrostatic effects as the stereochemical control
element.35 The subsequent hydroxyl-directed anti reduction of
aldol 16 (Me4NHB(OAc)3, MeCN/AcOH, -35 °C)31a proceeded
smoothly, affording the diol 17 in good yield and stereoselec-
tivity (ca. 10:1).43 Lactonization under acidic conditions dif-
ferentiated the diol functionality, and the remaining hydroxyl
group was then silylated to provide lactone 18 in diastereo-
merically pure form after silica gel chromatography.
To advance lactone 18 to the complete A-ring fragment 8a,
we needed only to perform operations on the two termini of
the fragment. Thus, lactone 18 was converted to anilide 19 in
acceptable yield using the aluminate derived from trimethyla-
luminum and aniline hydrochloride. Treatment of 19 with ozone
(10:1 CH2Cl2:MeOH, -78 °C) provided the cyclized lactols 20
as a 1.5:1 (â:R) mixture of anomers in 72% yield. This
transformation was plagued by two side reactionssepoxidation
of the C9 olefin (ca. 20% yield)44 and oxidation of the
p-methoxybenzyl group to a benzoate (ca. 5% yield)swhich
necessitated the use of minimal methanol in the solvent mixture
and a rapid quench to provide for optimal material throughput.
Preliminary experimental and computational results indicate that
a subtle conformational effect is probably responsible for the
spurious reactivity of 19 with ozone and its complete unreac-
tivity toward dihydroxylation agents.45 The product lactols 20
Both the epoxystyrene- and aldol-based strategies (eqs 3 and
4) have been combined with the reduction chemistry (eqs 1 and
2) in the syntheses of all three bryostatin acyclic building blocks
illustrated in Scheme 1. Since the aldol/reduction approach is
fundamentally more convergent and flexible, it is this method
that has been utilized in the fragment syntheses described in
the following three sections.36
Synthesis of the C1-C9 Subunit
In consideration of the basic conditions of the C9-C10 sulfone
alkylation reaction (vide infra), we elected to mask the C1
carboxyl terminus as a secondary amide,37 and thus formulated
8 as the C1-C9 synthon. Retrosynthetic removal of the C9
sulfone group and the C1 amide group from 8 provides a
protected 1,3,5-anti,anti triol, the synthesis of which has served
as a focal point for highlighting polyacetate synthesis methodol-
ogy.23 As shown in Scheme 2, we elected to address the
synthesis of this stereoarray using the aforementioned aldol/
reduction strategy. The successful execution of this plan is
illustrated in Scheme 3.
Scheme 2
(40) Penning, T. D.; Djuric, S. W.; Haack, R. A.; Kalish, V. J.; Miyashiro,
J. M.; Rowell, B. W.; Yu, S. S. Synth. Commun. 1990, 20, 307-312.
(41) (a) Johansson, R.; Samuelsson, B. J. Chem. Soc., Chem. Commun.
1984, 201. (b) Johansson, R.; Samuelsson, B. J. Chem. Soc., Perkin Trans.
1 1984, 2371-2374.
The stereoselective synthesis of synthon 8a began with the
dibutylboron triflate-mediated aldol reaction between excess
(4S)-3-chloroacetyl-4-phenylmethyl-2-oxazolidinone38 (10) and
the known39 2,2-dimethyl-4,4-diphenyl-3-butenal (9), providing
(42) (a) A similar result was obtained in the Felkin-selective addition to
an R-substituted aldehyde during our synthesis of Lepicidin, cf.: Evans,
D. A.; Black, W. C. J. Am. Chem. Soc. 1993, 115, 4497-4513. (b) The
(34) (a) Evans, D. A.; Coleman, P. J.; Coˆte´, B. J. Org. Chem. 1997, 62,
788-789. (b) Lagu, B. R.; Liotta, D. C. Tetrahedron Lett. 1994, 35, 4485-
4488. (c) Trost, B. M.; Urabe, H. J. Org. Chem. 1990, 55, 3982-3983. (d)
Paterson, I.; Goodman, J. M.; Isaka, M. Tetrahedron Lett. 1989, 30, 7121-
7124.
1
anti stereoselectivity for the aldol step was established by H NMR NOE
analysis of sulfone 8.
(43) The anti stereoselectivity for the reduction step was established by
1H NMR NOE analysis of lactone 18.
(35) Evans, D. A.; Dart, M. J.; Duffy, J. L.; Yang, M. G. J. Am. Chem.
Soc. 1996, 118, 4322-4343 and references therein.
(44) (a) Epoxidation of hindered olefins via the sterically less encumbered
“end-on” approach of ozone to the olefin has been well documented, cf.:
Bailey, P. S. Ozonation in Organic Chemistry; Academic: New York, 1978;
Vol. 1, Chapter 11. (b) While the oxirane byproduct could not be converged
to 20, it could be recycled back to the starting olefin 19 via reductive removal
of the oxygen (SmI2, 50% yield, unoptimized).
(45) Molecular mechanics calculations (MM2 force field) indicate that
the C3 TBS ether forces the C7 p-methoxybenzyl ether to rotate over the
C9 olefin. With the lactone 18, where the TBS ether is conformationally
constrained and the PMB ether can rotate away from the olefin (MM2), no
epoxidation was observed. This product aldehyde could not be advanced
to 20.
(36) For a discussion of the earlier approaches based on the epoxystyrene
synthon strategy, cf.: (a) Carreira, E. M. Ph.D. Thesis, Harvard University,
1990. (b) Gauchet-Prunet, J. A. Ph.D. Thesis, Harvard University, 1993.
(37) Evans, D. A.; Carter, P. H.; Dinsmore, C. J.; Barrow, J. C.; Katz,
J. L.; Kung, D. W. Tetrahedron Lett. 1997, 38, 4535-4538.
(38) Evans, D. A.; Sjogren, E. B.; Weber, A. E.; Conn, R. E. Tetrahedron
Lett. 1987, 28, 39-42.
(39) (a) Julia, M.; Baillarge, M. Bull. Soc. Chim. Fr. 1966, 734-742.
(b) Zimmerman, H. E.; Pratt, A. C. J. Am. Chem. Soc. 1970, 92, 6259-
6267.

7544 J. Am. Chem. Soc., Vol. 121, No. 33, 1999
EVans et al.
Scheme 3a
a Key: (a) 1.9 equiv of 10, Bu2BOTf, i-Pr2NEt, then 9, CH2Cl2, -78 f 0 °C, 20 h; (b) Zn, 2:1 THF/AcOH; (c) LiBH4, MeOH, THF, 0 °C; (d)
(i) PMPCH(OME)2, 10 mol % PPTS, CH2Cl2, (ii) DIBAl-H, CH2Cl2, 0 °C; (e) (COCl)2, DMSO, NEt3, CH2Cl2, -78 °C; (f) TiCl2(Oi-Pr)2, PhCH3,
-78 °C, then 5b, -78 °C; (g) Me4NHB(OAc)3, AcOH/MeCN, -35 °C; (h) PPTS, PhH, 80 °C; (i) TBSOTf, 2,6-lutidine, CH2Cl2, -10 °C; (j)
Me3Al, HCl‚H2NPh, CH2Cl2, 30 °C, then 18, 0 °C; (k) O3, 10:1 CH2Cl2/MeOH, -78 °C, then Me2S; (l) Ac2O, pyr; (m) PhSTMS, ZnI2, n-Bu4NI,
CH2Cl2; (n) m-CPBA, NaHCO3, EtOAc.
were selectively acetylated in the presence of the anilide using
acetic anhydride/pyridine (2 days, 23 °C) and then converted
into the R-sulfide (TMSSPh, ZnI2, n-Bu4NI)46 in good yield
and diastereoselectivity (86% yield, 93:7 R:â).47 Oxidation with
m-CPBA under buffered conditions (solid NaHCO3, EtOAc)
provided the analytically pure sulfone 8a in good yield.
afford an R-lactol as a single diastereomer.49 The subsequent
reduction of the lactol using the conditions of Kishi (BF3‚OEt2,
Et3SiH)50 was highly stereoselective (20:1 cis:trans)51 and was
accompanied by Lewis acid-catalyzed desilylation at C13 to
deliver the alcohol 24 in 89% yield. Removal of the bulky
triisopropylsilyl ether (TBAF, THF, 0 °C), global silylation
(TBSOTf, 2,6-lutidine, -78 °C), and hydrogenolysis of the
benzyl ether provided the cis,cis-2,4,6-tetrahydropyran 25 in
good yield. Oxidation to the labile C16 aldehyde 26 was best
accomplished using the method of Swern.
The illustrated route to 26 provided multigram quantities of
26 and allowed for the synthesis of the complete B-ring synthon
via alcohol 24 (cf. Scheme 6). Once our exploratory experiments
indicated that aldehyde 26 would be our target fragment (vide
infra), we opted to incorporate our newly discovered catalytic
aldol chemistry (eq 4)32 into the synthesis of this fragment
(Scheme 5, ent-7 f 26). Both of these approaches follow
parallel chemical strategies (Scheme 5), but the second route is
four steps shorter overall.52
Synthesis of the C10-C16 Subunit
Retrosynthetic excision of the exocyclic enoate from the
B-ring fragment 21 may be executed in either a diastereose-
lective (PG * PG′) or enantioselective (PG ) PG′) sense to
afford a â-C-glycoside (Scheme 4). Subsequent removal of the
C16 hydroxymethyl group affords a δ-lactone, which may be
obtained from the known25a 3,5,6-trishydroxy ester. The reduc-
tion of this plan to practice is shown in Schemes 5 and 6.
Scheme 4
Our initial efforts toward the bryostatin synthesis (Scheme
1, transform T1) required that we synthesize the complete C10-
C16 synthon 21, and our studies in this area are illustrated in
Scheme 6. Oxidation41 of alcohol 24 was followed by reductive
removal of the benzyl protecting group (H2, 10% Pd/C) to afford
hydroxyketone 31 in 77% overall yield. Preliminary attempts
to effect the intermolecular phosphonate condensation between
31 and trimethylphosphonoacetate under a variety of standard
conditions provided the exocyclic enoate E-34 in moderate yield
(70%) and poor stereoselection (ca. 70:30 E:Z),53 so we turned
(49) The stereochemical assignment was based on the presence of a 2-Hz
W-coupling between the lactol hydrogen and C14Hax. Given the propensity
for lactols to equilibrate in aqueous media, this stereochemistry may arise
from a postaddition isomerization, cf.: Dondoni, A.; Schurmann, M.-C. J.
Org. Chem. 1994, 59, 6404-6412.
At the time our work began, an enantioselective aldol-based
approach to δ-hydroxy-â-ketoesters (eq 4) was not available.
Thus, our initial synthesis of the fragment began with the
hydroxyl-directed anti reduction of 4,25a,31a which afforded diol
22 in good yield and stereoselectivity (Scheme 5). Cyclization
of the diol ester occurred readily under acidic conditions (PPTS,
PhH, 80 °C) and was followed by quantitative silylation of the
C13 alcohol to give lactone 23. One-carbon homologation of
23 using benzyloxymethyllithium48 proceeded smoothly to
(50) Lewis, M. D.; Cha, J. K.; Kishi, Y. J. Am. Chem. Soc. 1982, 104,
4976-4978.
(51) The stereochemical outcome of the silane reduction was apparent
1
from the high-field H NMR spectrum of tetrahydropyran 24 (400 MHz,
C6D6), which displayed identical axial (11 Hz) and equatorial (4.6 Hz)
couplings for both ring methylenes (C12 and C14).
(52) Three steps are saved by the incorporation of the aldol chemistry,
and one step is saved by the use of an in situ deprotection of the C10 PMB
ether.
(53) (a) The reaction of trimethylphosphonoacetate with ketone 31 was
screened under three conditions: LiCl, NEt3, MeCN, 23 °C, 2 days (66:33
E:Z, 50% conversion); NaH, PhCH3, 23 °C, 90 min. (70:30 E:Z); KHMDS,
THF, 0 °C, 30 min. (50:50 E:Z). (b) Recent results from the laboratory of
H. M. R. Hoffmann indicate that the exocyclic ethyl ester may be installed
in good stereoselectivity (9:1 E:Z, 72% yield); cf. ref 24e.
(46) (a) Hanessian, S.; Guindon, Y. Carbohydr. Res. 1980, 86, C3. (b)
Nicolaou, K. C.; Daines, R. A.; Ogawa, Y.; Chakraborty, T. K. J. Am. Chem.
Soc. 1988, 110, 4696-4705. (c) Evans, D. A.; Truesdale, L. K.; Grimm,
K. G.; Nesbitt, S. L. J. Am. Chem. Soc. 1977, 99, 5099-5116.
(47) Both the minor â-anomer and the C7 p-methoxybenzoate impurity
were removed at this stage by flash chromatography.
(48) Still, W. C. J. Am. Chem. Soc. 1978, 100, 1481-1486.

Total Synthesis of Bryostatin 2
J. Am. Chem. Soc., Vol. 121, No. 33, 1999 7545
Scheme 5a
a Key: (a) Me4NHB(OAc)3, AcOH/MeCN, -35 °C; (b) PPTS, 0.01 M PhH, 80 °C, -MeOH; (c) TESCl, imidazole, CH2Cl2, 0 °C; (d) BnOCH2Li,
THF, -78 f -50 °C; (e) BF3‚OEt2, Et3SiH, CH2Cl2, -20 °C; (f) TBAF, THF, 0 °C; (g) TBSOTf, 2,6-lutidine, CH2Cl2, -78 °C; (h) 1 atm H2,
catalytic 10% Pd/C, AcOH, EtOAc; (i) (COCl)2, DMSO, NEt3, CH2Cl2, -78 f -50 °C; (j) F3CCO2H, CH2Cl2; (k) PMBOCH2Li, THF, -78 f
-50 °C; (l) TBSCl, imidazole, DMAP, CH2Cl2.
Scheme 6a
E- and Z-34, which could be separated by silica gel chroma-
tography (Scheme 6).
The olefination of 30 with chiral phosphonate 35a was also
evaluated (eq 5). A careful study of reaction parameters58 led
to acceptable (E) selectivity (84:16 E:Z,), and this diastereose-
lection could be further improved (92:8 E:Z) by utilizing the
more hindered phosphonate57b 35b. Taken together with the fact
that phosphonate 35a is a (Z)-selective olefinating reagent with
aldehydes,59 these data point to a mechanism of stereoinduction
in which the kinetic facial bias exerted by 35a in the initial
aldol step is transferred to enoate geometry through kinetic
trapping of the intervening aldolate.60
a Key: (a) (COCl)2, DMSO, NEt3, -78 °C; (b) H2, 10% Pd/C,
catalytic AcOH, EtOAc; (c) (EtO)2P(O)CH2CO2(CH2)5CO2H, BOPCl,
DMAP, CH2Cl2; (d) 50 equiv of LiCl, 2 equiv of DBU, 1 mM MeCN;
(e) K2CO3, MeOH; (f) NaHMDS, 2 equiv of 35a, THF, -78 °C, then
31, -78 f 0 °C; (g) DMTrCl, i-Pr2NEt, DMAP, CH2Cl2; (h) (i) TBAF,
THF, 0 °C; (ii) Tf2O, 2,6-lutidine, CH2Cl2, -10 °C.
to the use of the tethered phosphonate strategy25b (Scheme 6).
Acylation of the C16 hydroxyl of 31 with 6-[(diethoxyphospho-
nyl)acetyl]oxyhexanoic acid25b (BOPCl, DMAP) provided phos-
phonate 32, which was readily cyclized under soft enolization
conditions (LiCl, syringe pump addition of 2 equiv of DBU)54
to afford enoate 33 in 63% yield and excellent stereoselectivity
Synthesis of the C17-C27 Subunit
In analogy with our previous work on cytovaricin,61 we
planned to carry the C-ring exocyclic enoate through the Julia
coupling as its exocyclic acid in order to avoid any problems
with proton transfer, and we thus formulated the C-ring synthon
as 38. Retrosynthetic removal of the exocyclic acid and C19-
C20 diol afforded glycal 39, which could be derived from the
acyclic precursor via cyclization/dehydration (Scheme 7). This
protected ketotriol was then dissected into ketone 40 and
ketoaldehyde 41 using our aldol/reduction strategy. The imple-
1
(>95:5 E:Z, H NMR). The subsequent cleavage of the tether
(K2CO3, MeOH, 23 °C)25b proceeded without incident to afford
enoate E-34 in 74% yield (43% yield overall).55 We speculated
that the use of a chiral phosphonate reagent56 might improve
the process and thus examined the use of methyl (R)-[(2,2′-
bishydroxy-1,1′-binaphthyl)phosphonyl]acetate (35a), a reagent
introduced by Fuji.57 We found that the condensation of the
sodium anion of 35a with hydroxyketone 31 (THF, -78 f 0
°C) afforded a 58% yield (unoptimized) of a 79:21 mixture of
(57) (a) Tanaka, K.; Ohta, Y.; Fuji, K.; Taga, T. Tetrahedron Lett. 1993,
34, 4071-4074. (b) Tanaka, K.; Otsubo, K.; Fuji, K. Tetrahedron Lett.
1996, 37, 3735-3738.
(58) The following parameters were examined: counterion (Li, Na, K,
and Zn), solvent (THF, Et2O, PhCH3, and THF/HMPA), and temperature
(0, -35, -65, and -78 °C).
(59) (a) The reaction of either 35a or 35b with dihydrocinnamaldehyde
(NaHMDS, THF, -78 °C) afforded an 82:18 Z:E mixture of olefins. (b)
The Z stereoselectivity of achiral phenolic phosphonates with simple
aldehydes has recently been documented, cf.: Ando, K. J. Org. Chem. 1997,
62, 1934-1939.
(60) For discussions regarding the mechanism of the Horner-Wad-
sworth-Emmons olefination, cf.: (a) Maryanoff, B. E.; Reitz, A. B. Chem.
ReV. 1989, 89, 863. (b) Vedejs, E.; Peterson, M. J. Top. Stereochem. 1994,
21, 1-157.
(61) Evans, D. A.; Kaldor, S. W.; Jones, T. K.; Clardy, J.; Stout, T. J. J.
Am. Chem. Soc. 1990, 112, 7001-7031.
(54) Blanchette, M. A.; Choy, W.; Davis, J. T.; Essenfeld, A. P.;
Masamune, S.; Roush, W. R.; Sakai, T. Tetrahedron Lett. 1984, 25, 2183-
2186.
(55) The geometry of the olefin in E-34 was evident from the NOE
enhancement between the exocyclic vinyl-H and the equatorial C12H and
the pronounced deshielding of the C14 equatorial proton relative to the C12
equatorial proton. The opposite deshielding pattern was seen for the Z
isomer.
(56) For leading references to the recent literature, cf.: (a) Abiko, A.;
Masamune, S. Tetrahedron Lett. 1996, 37, 1077-1080. (b) Denmark, S.
E.; Rivera, I. J. Org. Chem. 1994, 59, 6887-6889.

7546 J. Am. Chem. Soc., Vol. 121, No. 33, 1999
EVans et al.
Scheme 7
regio- and diastereoselective, but the resulting 1,3-diol was
difficult to differentiate in a high-yielding manner. Accordingly,
we turned to the samarium-catalyzed Tishchenko reduction31b
of aldol 46 (catalytic SmI2, excess p-NO2PhCHO), which
provided diastereomerically pure nitrobenzoate 47 in excellent
yield, regioselectivity, and diastereoselectivity (>95:5 ds).68 The
product diol monoester 47 was next silylated (TBSOTf, 2,6-
lutidine) and deprotected (LiOH, aqueous THF/MeOH) to afford
the target hydroxyketone 48 in 90% overall yield.69 The
synthesis of glycal 39a was completed by the dehydration of
48 under acidic conditions (catalytic CSA, PhH, azeotropic
removal of water at 80 °C, 92-96% yields).
The synthesis of the C17-C27 glycal 39 (Scheme 8) was very
reliable on large scale and was independently carried out with
a variety of functional groups at C17 (SO2Ph, SO2(N-Me)Im,70
3,4-(MeO)2PhCH2O) in order to assay different coupling strate-
gies (vide infra). For the sake of illustration, the elaboration of
the imidazolyl sulfone glycal 39b to the complete C17-C27
synthon is detailed below (Scheme 9). Epoxidation of glycal
39b with m-CPBA proceeded with in situ methanolysis to afford
the hydroxyketal as a mixture of four diastereomers (1H NMR
analysis), which was immediately oxidized with the Dess-
Martin periodinane71 to give the ketone 49 as an inseparable
4.5:1 mixture of diastereomers in 61% overall yield.72 Ketone
49 was readily condensed with glyoxylic acid monohydrate
under basic conditions (aqueous NaOH, THF, 23 °C)73 to give
the exocyclic carboxylic acid (single olefin isomer), which was
directly esterified by brief 74 treatment with dimethylformamide
dimethylacetal in refluxing benzene to provide the (E) olefin
50 in 77% yield for the two-step procedure. Since this compound
was prone to olefin isomerization, it was immediately reduced
under Luche conditions (NaBH4, CeCl3)75 to afford the axial
alcohol 51 as a single diastereomersan outcome which is
consistent with steric control of the approach of the borohydride
mentation of this plan for the synthesis of 38 and 39 is shown
in Schemes 8 and 9.
Conversion of 2,2-dimethyl-1,3-propanediol (42) into the
sulfonylaldehyde 43 proceeded over four standard steps (Scheme
8: monotosylation, sulfide displacement, sulfide oxidation, and
alcohol oxidation) in 76% overall yield and was routinely
performed on large (>500 mmol) scale. Reaction of aldehyde
43 with the Grignard reagent derived from 5-bromo-1-pentene
(CH2Cl2/Et2O, 0 °C) was followed by Swern oxidation (CH2-
Cl2, -78 f 0 °C) to provide a C19 ketone in 82% yield for the
two steps.62 Oxidation of the terminal alkene was best ac-
complished using an osmium-mediated dihydroxylation (cata-
lytic in K2OsO4‚(OH2)2)63 and was followed by diol cleavage
with sodium periodate to afford the desired keto aldehyde 41a
as a white solid in 95% yield overall. The requisite ketone
partner 40 for the ensuing aldol reaction was prepared using a
kinetic resolution strategy,33 as shown in Scheme 8. The
indicated route (resolution, protection, ozonolysis) is operation-
ally convenient (three steps, one purification) and proceeds well
on large scale (>250 mmol).
Reaction of 41a with the lithium enolate of 40 was both low
yielding (<40% yield) and poorly diastereoselective (<3:1).
After some optimization,64 it was found that the aldol reaction
of 40 and 41a could be efficiently mediated using the chiral
IPC enolates65 of Paterson and Brown (Scheme 8); reaction of
a slight excess of the chiral boron enolate of 40 with 41a
delivered aldol 46 in 86% yield and 93:7 diastereoselectivity.
As shown in eq 6, the success of the (-)-DIPCl66 reaction is a
result of double stereodifferentiation: reaction of 40 and
propionaldehyde with (-)-DIPCl afforded the desired 5S
diastereomer with enhanced diastereoselection, and (+)-DIPCl
afforded the undesired 5R diastereomer in lower diastereose-
lectivity and chemical yield.67
(64) (a) A study of the aldol reaction of 40 (and other protected
variants: TBDPS, MEM, BOM, Bn) with propionaldehyde demonstrated
that only moderate yields (25-50%) and low selectivities (<3:1 diaste-
reoselection in the desired sense) could be obtained using a wide array of
metal centers (Li, B, Sn, Zn, Mg, Ti) and enolization conditions. (b) For
the obtainment of the undesired 1,4-syn diastereomer in high stereoselec-
tivity, cf. ref 34c.
(65) (a) Ramachandran, P. V.; Xu, W.-C.; Brown, H. C. Tetrahedron
Lett. 1996, 37, 4911-4914. (b) Paterson, I.; Goodman, J. M.; Lister, M.
A.; Schumann, R. C.; McClure, C. K.; Norcross, R. D. Tetrahedron 1990,
46, 4663-4684.
(66) DIPCl is a trademark of the Aldrich Chemical Co. (DIP )
diisopinocampheyl).
(67) Assuming a base level selectivity of 89:11 for the reaction of
n-alkylaldehydes with these IPC enolates (ref 65b), our stereoselectivities
are roughly in accord with the precepts for double diastereodifferentiation:
calculated value ) 8/1 × 2.2/1 ) 17.8/1 ratio of S/R; actual value ) 19/1;
cf.: Masamune, S.; Choy, W.; Peterson, J. S.; Sita, L. R. Angew. Chem.,
Int. Ed. Engl. 1985, 24, 1-30.
(68) Analysis of the NOE spectrum of an ortho ester derivative of 47
(synthesized in two steps: (i) LiOH, (ii) DDQ, K2CO3) secured the
stereochemical assignment for both the aldol and reduction steps.
(69) Note that the use of p-nitrobenzoate as the C23 hydroxyl protecting
group was critical to the success of this sequence (46 f 48): the C23 acetate
was too labile and underwent migration during the silylation reaction, and
the C23 benzoate was too difficult to remove.
(70) The use of N-methyl imidazolyl sulfones has been reported to
improve the Julia olefination, cf.: Kende, A. S.; Mendoza, J. S. Tetrahedron
Lett. 1990, 31, 7105-7108.
(71) (a) Dess, D. B.; Martin, J. C. J. Am. Chem. Soc. 1991, 113, 7277-
7287. (b) For an improved preparation of the reagent, cf.: Ireland, R. E.;
Liu, L. J. Org. Chem. 1993, 58, 2899.
(72) Inexplicably, the yields for this two-step transformation were much
higher in the phenyl sulfone (81%) and dimethoxybenzyl (100%) series.
The diastereoselectivity was unaffected by the C17 substituent.
(73) Pettit, G.; Green, B.; Dunn, G. L. J. Org. Chem. 1970, 35, 1367-
1376.
With aldol adduct 46 in hand, we focused on completing the
synthesis of the 1,2,4-syn,anti stereochemical array (Scheme 8).
Directed reduction31a of 46 with Me4NHB(OAc)3 was both
(62) (a) The use of methylene chloride as a cosolvent was essential. (b)
Addition of this Grignard reagent to the C19 carboxylic acid or Weinreb
amide was low-yielding.
(63) (a) Minato, M.; Yamamoto, K.; Tsuji, J. J. Org. Chem. 1990, 55,
766-768. (b) Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless, K. B. Chem.
ReV. 1994, 94, 2483-2547.
(74) Reaction times longer than 20-30 min provided for considerably
reduced yields because of olefin isomerization (E f Z).
(75) Luche, J. L. J. Am. Chem. Soc. 1978, 100, 2226.

Total Synthesis of Bryostatin 2
J. Am. Chem. Soc., Vol. 121, No. 33, 1999 7547
Scheme 8a
a Key: (a) 0.2 equiv of TosCl, pyr, CH2Cl2, 0 f 23 °C; (b) PhSH, NaH, DMF, 80 °C; (c) m-CPBA, CH2Cl2, 0 f 23 °C; (d) (COCl)2, DMSO,
NEt3, CH2Cl2, -78 f 0 °C; (e) BrMg(CH2)3CHdCH2, 1:1 Et2O/CH2Cl2, 0 f 23 °C; (f) (COCl)2, DMSO, NEt3, CH2Cl2, -78 f -50 °C; (g) 2
mol % K2OsO4(OH2)2, 2 mol % quinuclidine, K3Fe(CN)6, K2CO3, 1:1 t-BuOH/H2O; (h) NaIO4, NaHCO3, 2:2:1 t-BuOH/H2O/THF; (i) 15 mol %
L-(+)-DIPT, 10 mol % Ti(Oi-Pr)4, 0.7 equiv of t-BuO2H, CH2Cl2, -20 °C; (j) NaH, PMBBr, catalytic n-Bu4NI, THF, 0 f 23 °C; (k) O3, CH2Cl2/
MeOH, -78 °C, then Me2S, -78 f 23 °C; (l) 40, (-)-DIPCl, NEt3, CH2Cl2, -78 °C; then 41a, -70 °C; (m) 20 mol % SmI2, p-NO2PhCHO, THF,
0 °C; (n) TBSOTf, 2,6-lutidine, CH2Cl2, -15 °C; (o) LiOH, 2:2:1 THF/MeOH/H2O; (p) 5 mol % CSA, PhH, 80 °C.
Scheme 9a
reagent. The desired exocyclic carboxylic acid 38b could be
synthesized along similar lines: introduction of the exocyclic
tert-butyl ester was accomplished using the aldol condensation
of tert-butylglyoxylate76 and the potassium enolate of 49 and
was followed immediately by Luche reduction to give the
diastereomerically pure alcohol 52 (54% yield, two steps).
Deprotection of the tert-butyl ester using trimethylsilyl triflate77
proceeded with concomitant protection of the axial alcohol to
provide the target acid 38b in 84% yield.78
Fragment Coupling
Our initial studies on the A-B fragment coupling utilized
tert-butyl amide79 8b and triflate80 53 and revealed that the
lithiated dianion of sulfone 8b could be condensed with 53 in
the presence of HMPA at -78 °C (eq 7). Upon warming of the
reaction mixture to room temperature,81 elimination of phenyl
sulfinate82 afforded enol ether 54 (45% yield), which could be
converted to the C9 mixed methyl ketal using trichloroacetic
acid in methanol (66% yield, 30% overall). However, when the
dianion of sulfone 8b was added to the fully functionalized
triflate 21, the enol ether 55 was only isolated in ca. 7% yield
(eq 8). The majority of the sulfone 8b was recovered unchanged
a Key: (a) m-CPBA, MeOH, -20 °C; (b) Dess-Martin periodinane,
from this reaction, and the bulk of triflate 21 had been converted
pyr, CH2Cl2; (c) OHCCO2H, aqueous NaOH, THF; (d) (MeO)2CHNMe2,
into a C10-C12 cyclopropane,83 presumably via a proton-transfer/
PhH, 80 °C; (e) NaBH4, CeCl3‚(OH2)7, MeOH, -78 °C; (f) KOt-Bu,
OHCCO2t-Bu, THF, -15 °C; (g) TMSOTf, 2,6-lutidine, CH2Cl2, 0 °C.
intramolecular alkylation sequence.
It was eventually discovered that the use of the tert-butyl
amide 8b was not synthetically viable,84 and so we turned to
the use of other amide residues. Interestingly, the use of a C1
(76) Subasinghe, N.; Schulte, M.; Chan, M. Y.-M.; Roon, R. J.; Koerner,
J. F.; Johnson, R. L. J. Med. Chem. 1990, 33, 2734-2744.
(77) (a) Jones, A. B.; Villalobos, A.; Linde, R. G.; Danishefsky, S. J. J.
Org. Chem. 1990, 55, 2786. (b) Evans, D. A.; Ng, H. P.; Rieger, D. L. J.
Am. Chem. Soc. 1993, 115, 11446-11459.
(78) Irradiation of either the exocyclic olefin resonance or the equatorial
proton at C20 produced a large NOE enhancement of the other resonance.
Likewise, the chemical shifts and coupling constants of C20Heq and C22Heq
were in accord with the assigned stereochemistry.
(79) The preparation of tert-butyl amide 8b was identical to that for
anilide 8a, with the exception of the transamidation of lactone 18 (t-BuNH2,
Me2AlCl, THF, room temperature, 16 h, 96% yield).
(80) (a) Triflate 53 was prepared from alcohol 24 in three steps and 89%
overall yield: (i) Bz2O, DMAP; (ii) TBAF, THF, 0 °C; (iii) Tf2O, 2,6-
lutidine, -10 °C. See the Supporting Information for details. (b) Note that
the analogous C10 iodide was not reactive toward C9 sulfonyl anions.
(81) If the reaction was quenched at lower temperatures (0 or -15 °C),
significant amounts of the C9 lactol were obtained, indicating that the
sulfinate elimination was slower than the anion alkylation.
(82) This is consistent with the results of both Ley and Beau; cf. ref 28
and the following: Beau, J.-M.; Sinay¨, P. Tetrahedron Lett. 1985, 26, 6189-
6192.

7548 J. Am. Chem. Soc., Vol. 121, No. 33, 1999
EVans et al.
benzyl (or PMB) amide resulted in the production of large
amounts of the N1,C9-bisalkylation product, even at low reaction
temperatures. This side reaction could be avoided by attenuating
the nucleophilicity of the amide residue; the dianion of N-phenyl
sulfone 8a alkylated exclusively at the carbon center, giving
the unstable85 ketal 56 in 46% overall yield after methanolysis
(eq 9).86 With this result, we delayed further attempts at
optimization until the real C9-C10 coupling was at hand (vide
infra).
good yield (LiBr,94 -78 °C95). Unfortunately, the fully elabo-
rated C-ring aldehyde96 59 proved to be completely unreactive
toward the B-ring sulfonylanion,97 and only the undesired ring-
opened product was obtained after prolonged reaction times
(eq 12).
Our initial study of the C16-C17 Julia olefination utilized
carboxylic acid 38b in order to supress C22 enolization of the
enoate moiety.87 Unfortunately, we found that reaction at C17
could not be achieved, regardless of the experimental condi-
tions88 employed. Rather, the use of strong bases promoted
deprotonation on the imidazole ring, as judged by quenching
of the orange dianion of 38b with deuterium chloride or simple
aldehydes (eq 10).89 In an attempt to override this preference
These exploratory results indicated that the steric congestion
at C17 was the dominant issue surrounding the C16-C17
olefination; accordingly, we explored the use of the less
sterically crowded C19-C20 glycal 39a (Scheme 10). Deproto-
nation of the sulfone 39a with n-butyllithium was followed by
addition of 1 equiv of aldehyde 26 to afford the hydroxysulfones
as a mixture of four diastereomers.98 Acetylation of the
unpurified mixture (Ac2O/DMAP) and purification by flash
chromatography afforded the acetoxysulfones in 87% yield for
the two steps. Acetoxysulfone elimination was most conve-
niently carried out according to the method of Pak,99 which
entailed forming a magnesium amalgam in situ with catalytic
amounts of mercuric chloride. Under these conditions, the
diastereomerically pure (E) olefin 60 (1H NMR, J ) 16 Hz)
was obtained in 74% yield (64% overall yield for the Julia
olefination).
for spurious deprotonation, we examined a variety of structural
modifications, including the use of the reduced oxidation state
at the exocyclic ester, an unprotected hydroxyl at C20,90 and
the phenyl sulfone at C17.91 Unfortunately, a high-yielding C16-
C17 olefination was never obtained.
Since the problems associated with the Julia olefination
stemmed from either the inability to enolize the sulfonylmeth-
ylene or the excessive basicity of the resultant anion, it was
reasoned that the reverse coupling approach (unhindered C16
sulfone and unenolizable C17 aldehyde) might be successful.92
Indeed, our first results with N-methyl imidazolyl sulfone93 (57)
were encouraging (eq 11): the lithium anion of 57 could be
coupled with 3-benzyloxy-2,2-dimethylpropanal to give 58 in
Based on these experiments, methods for forming both the
C9-C10 and C16-C17 bond were in hand; however, the
(83) The cyclopropane was isolated in 77% yield when the coupling was
carried out in the absence of HMPA.
(84) The tert-butyl amide could not be hydrolyzed in the presence of
the C9 ketal. For a discussion of amide hydrolysis, cf. ref 37.
(85) Ketal 56 was completely unstable when stored neat and reverted
back to the enol ether and lactols after even 10-30 min at room temperature.
In contrast, the benzyl, PMB, and t-Bu amide analogues were perfectly
stable, even after prolonged storage.
(86) An equivalent amount of product could be isolated from the two-
step, one-pot variant: Li2-8a and 53, THF, -78 f -15 °C, then MeOH,
Cl3CCO2H quench. The avoidance of HMPA required that a higher
temperature be utilized in the alkylation step.
(92) For the use of both ylide and nitronate anions generated â to a THP
oxygen, cf.: (a) McWhorter, W. W., Jr. Ph.D. Thesis, Harvard University,
1984. (b) Secrist, J. A., III; Wu, S.-R. J. Org. Chem. 1977, 42, 4084-
4088. (c) Kobertz, W. R.; Bertozzi, C. R.; Bednarski, M. D. J. Org. Chem.
1996, 61, 1894-1897. (d) Martin, O. R.; Lai, W. J. Org. Chem. 1990, 55,
5188.
(93) Sulfone 57 was synthesized from alcohol 24 in five steps and 76%
overall yield: (i) TBSOTf, 2,6-lutidine, -78 °C; (ii) H2, 10% Pd/C; (iii)
Tf2O, 2,6-lutidine, -10 °C; (iv) LiSIm, THF, 0 °C; (v) m-CPBA. See the
Supporting Information for details.
(87) Use of the methyl C20-hydroxyester 51 was complicated by
deprotonation at C22 (data not shown).
(94) For the use of LiBr with enolates, see: Seebach, D. Angew. Chem.,
Int. Ed. Engl. 1988, 27, 1624.
(88) (a) The following variables were screened: base (n-BuLi, PhLi,
LDA, LiNEt2); solvent (THF, THF/HMPA); temperature (-78 °C, -40
°C); and aldehyde (i-PrCHO, t-BuCHO, PhCHO). (b) In addition, the
following additives were evaluated: LiBr, MgBr2‚OEt2 (Danishefsky, S.
J.; Selnick, H. G.; Zelle, R. E.; DeNinno, M. P. J. Am. Chem. Soc. 1988,
110, 4368), and BF3‚OEt2 (Achmatowicz, B.; Baranowska, E.; Daniewski,
A. R.; Pankowski, J.; Wicha, J. Tetrahedron 1988, 44, 4989).
(89) The spurious deprotonation of phenyl sulfones is a common problem.
Although our model studies clearly indicated that the N-methyl imidazolyl
sulfone exerts a more powerful R-acidfying effect than the phenyl sulfone,
this is apparently not enough to override the tremendous steric compression
about C17.
(90) Low yield (<20%) of the desired product could be obtained when
the exocyclic ester was reduced and the C20 hydroxyl group was left
unprotected, provided LiNEt2 was used as the base.
(91) While enolization of the C20-hydroxy C17-phenyl sulfone could be
accomplished, the resultant anion (or modified variants thereof) proved to
be more basic than nucleophilic.
(95) The sulfone anion decomposed slowly at -78 °C, producing about
10% of ring-opened elimination product in 1 h. For comparison, the lithium
anion of a related phenyl sulfone anion instantly produced the undesired
open-chain product, cf.: Barnes, N. J.; Davidson, A. H.; Hughes, L. R.;
Procter, G. J. Chem. Soc., Chem. Commun. 1985, 1292-1294.
(96) Aldehyde 59 was made from the C17-ODMB analogue of ester 51
in three steps: (i) Ac2O, DMAP; (ii) DDQ, H2O/CH2Cl2, 5 °C; (iii) Dess-
Martin oxidation.
(97) The hindered aldehyde 59 was unreactive toward a number of
standard enolate nucleophiles but could be engaged with unhindered silyl
enol ethers under Lewis acid catalysis. For the reaction of a similar C-ring
aldehyde with an allylboronate, cf.: Wender, P. A.; DeBrabander, J.; Harran,
P.; Jimenez, J.-M.; Koehler, M.; Lippa, B.; Park, C.-M. J. Am. Chem. Soc.
1998, 120, 4534-4535.
(98) To obtain complete conversion, the addition step needed to be carried
out at -50 °C.
(99) Lee, G. H.; Lee, H. K.; Choi, E. B.; Kim, B. T.; Pak, C. S.
Tetrahedron Lett. 1995, 36, 5607-5610.

Total Synthesis of Bryostatin 2
J. Am. Chem. Soc., Vol. 121, No. 33, 1999 7549
Scheme 10a
a Key: (a) (i) n-BuLi, THF, -78 °C, then 26, -78 f -50 °C; (ii) Ac2O, DMAP, CH2Cl2; (b) Mg, 20 mol % HgCl2, EtOH; (c) TBAF, THF,
-15 °C; (d) Tf2O, 2,6-lutidine, CH2Cl2, -10 °C; (e) 8a, 2 equiv of n-BuLi, THF, -78 °C, then HMPA, then 61, -78 °C; (f) TESCl, imidazole,
MeCN; (g) Boc2O, DMAP, MeCN; (h) BnOLi, 1:1 THF/DMF, -30 °C; (i) (i) m-CPBA, MeOH, -20 °C, (ii) ClCH2CO2H, MeOH, 0 °C, (iii)
Dess-Martin periodinane, pyr, CH2Cl2; (j) HF‚pyr, 4:4:1 THF/MeOH/pyr; (k) TESCl, DMAP, CH2Cl2, 10 °C (65% + 15% each of the mono- and
tris-silylether); (l) 1,4-cyclohexadiene, 10% Pd/C (50 mol %), EtOAc; (m) 2,4,6-trichlorobenzoyl chloride, i-PrNEt2, PhH, then DMAP, 1.0 mM
PhH.
constraints of each coupling reaction (vide supra) required a
slightly less convergent approach to the synthesis of the tricycle
(Scheme 1, transform T1*). Following this new plan, we
investigated the two logical assemblage strategies (A + B f
AB + C f ABC and B + C f BC + A f ABC) and
discovered that the ABC tricycle could be accessed via either
pathway. Since the yields for both fragment-coupling processes
were higher with the second option (B + C f BC + A f
ABC), we utilized this approach in our synthesis (Scheme 10).
Due to the acid lability of the C19-C20 glycal, functionalization
of the C10 position could only be effected by selective
desilylation of 60 under basic conditions (TBAF, THF, -15
°C, 75% yield).100 Subsequent sulfonylation of the new hydroxyl
group with trifluoromethanesulfonic anhydride (2,6-lutidine,
-10 °C) delivered the unstable primary triflate 61 (95% yield),
which was immediately101 treated with 1.5 equiv of the dilithio
anion of sulfonylamide 8a (THF/HMPA, -78 °C) to deliver
the ABC lactol 62 in 84-90% yield.102 As expected on the basis
of our earlier model studies, overalkylation of the sulfonylamide
dianion Li2-8a was not detected. Utilizing the illustrated
sequence (Scheme 10), the ABC tricycle 62 was assembled in
19 linear steps (14% overall yield) from commercially available
starting materials (toluenesulfonyl chloride and 2,2-dimethyl-
1,3-propanediol) on a multigram scale.
Macrocyclization
The C9 lactol 62 (Scheme 10) was protected through a
selective103 methanolysis (PPTS, MeOH/(MeO)3CH, 62 f 63).
In accord with earlier results on the AB bicyclic systems, we
found that the N-aryl amide 63 was reverted to the starting lactol
after purification (t0.5 ≈ 2 h when neat). Although the exact
origin of this lability is unclear at the present time, the stability
of its N-benzyl analogue, the N-Boc-N-phenyl derivative 64,
and benzyl ester 68 all suggest that the C9 ketal of 63 may be
compromised by the acidity of the anilide N-H proton (pKa ≈
21.5, DMSO).104 Attempts to circumvent this problem by
making recourse to thioketalization or lactol acetylation were
unsuccessful, and initial attempts to advance the N-Boc deriva-
tive 64 were compromised by the poor electrophilicity and base
sensitivity of the Boc-anilide and the acid lability of the C19-
C20 glycal. After some experimentation, it was discovered that
the Boc-anilide was much more reactive when the A-ring was
locked in its open-chain form and that the sensitivity of the
C-ring glycal was eliminated upon oxidation. These observations
led to the development of the successful route to the seco acid
69. Functionalization of the A-ring began with silylation of the
open-chain hydroxyketone as its C5-OTES ether 65. When
trapped in this protected and extended tautomer, the C1 anilide
was now readily transesterified to the benzyl ester 66 (Boc2O,
then LiOBn).105 It is noteworthy that the use of DMF106 and
low temperature (-30 °C) was critical to the success of the
(100) The temperature control of this reaction was critical to its success,
as the lability of the secondary C13-OTBS was quite similar to that of the
primary C10-OTBS. Less than 30% of the primary alcohol was isolated
under a variety of acidic desilylation conditions.
(101) The triflate 61 decomposed rapidly (<10 min) when left in neat
form. See the Supporting Information for the details of the purification of
61.
(103) Our earlier results with acidic methanolysis of the A and C ring
enol ethers showed that the equilibrium lay on the side of either the ketal
(A ring) or the enol ether (C ring), depending on the position (exo or endo)
and steric environment of the enol ether.
(102) In practice, the lactol 62 was obtained in 60-75% yield after
chromatography. An additional 15-20% was isolated as the hydroxyketone
tautomer, which was silylated (vide infra) to reconverge the material. See
the Supporting Information for details.
(104) For the pKa of N-phenyl acetamide, cf.: Bordwell, F. G. Acc. Chem.
Res. 1988, 21, 456-463.
(105) Evans, D. A.; Ennis, M. D.; Mathre, D. J. J. Am. Chem. Soc. 1982,
104, 1737-1739.

7550 J. Am. Chem. Soc., Vol. 121, No. 33, 1999
EVans et al.
Scheme 11a
a Key: (a) 20 mol % PPTS, 2:1 MeOH/(MeO)3CH, CH2Cl2, -30 °C; (b) Dess-Martin periodinane, pyr, CH2Cl2; (c) 35a, NaHMDS, THF, -78
°C, then 71, -15 °C; (d) KHMDS, THF, -78 °C, then OHCCO2Me, -78 °C; (e) Et3NSO2NCO2Me, PhH; (f) 74, BH3‚SMe, CH2Cl2, then MeOH,
then MAc2O, pyr, DMAP; (g) (i) PPTS, 3:1 THF/H2O, (ii) Na2CO3, MeOH, (iii) pTsOH, 4:1 MeCN/H2O.
latter reaction (65 f 66) and resulted in a dramatic suppression
of the base-induced107 elimination of the C3-OTBS group. In
contrast to the oxidation of glycal 39, oxidation of the C19-
C20 glycal (m-CPBA/MeOH, then Dess-Martin oxidation)
delivered a 1:1 mixture of stereoisomers at C19, along with
unexpected substitution products (C19OH and/or C19O2CAr).
Incorporation of a thermodynamically controlled methanolysis
step (MeOH, ClCH2CO2H, 0 °C) into the double oxidation
sequence erased the effects of the unwanted substitution and
corrected the stereochemical deficiencies of the process, deliver-
ing the oxoketal 67 as a single diastereomer in 79% overall
yield.108 Complete desilylation and concomitant cyclization of
the A-ring was readily effected under buffered conditions (HF‚
Pyr, THF/MeOH/pyridine, 3 days) to afford the triol 68 in good
yield.109 Initial macrocyclization attempts focused on the
carboxylic acid derivative of 68 (R, R′ ) H); however, these
studies were frustrated by the unexpected reactivity of the C3
hydroxyl group, and only low-yielding cyclization reactions
could be obtained.110 These problems were circumvented via
selective protection of the C3 hydroxyl in the presence of the
C25 hydroxyl (TESCl, DMAP; 65% yield; 82% yield, based on
recovered starting material). Subsequent selective debenzylation
(cyclohexadiene, 10% Pd/C)111 afforded the hydroxyacid 69,
which was smoothly cyclized under modified Yamaguchi
conditions112 to afford macrocycle 70 in 81% yield (Scheme
10).
Synthesis of Bryostatin 2
With the macrocyclization accomplished, we addressed the
late-stage installation of the enoate moieties. Selective removal
of the C13-OTES group of 70 afforded the C13 alcohol, which
was readily oxidized to the diketone 71 using the Dess-Martin
procedure (Scheme 11).71 Condensation of the C13,C20 diketone
71 with 4 equiv of the sodium anion of trimethyl phosphonoac-
etate (THF, 23 °C) selectively113 derivatized the C13 ketone to
produce the C13-C30 enoate 72 in 100% yield and 64:36 Z:E
diastereoselectivity. Following the precedent previously estab-
lished in simple B-ring systems (eq 5), the use of the sodium
anion of Fuji’s chiral phosphonate 35a provided an 85:15
mixture of diastereomers in 93% combined yield (75% isolated
yield of 72114 after preparative TLC). Subsequent incorporation
of the C-ring enoate was best accomplished using a two-step
procedure (KHMDS, OHCCO2Me; then Burgess reagent).115
The ability of KHMDS to selectively enolize the C20 ketone in
the presence of the C1 ester is precedented116 and is apparently
(112) (a) Evans, D. A.; Kim, A. S. J. Am. Chem. Soc. 1996, 118, 11323-
11324. (b) Inanaga, J.; Hirata, K.; Saeki, H.; Katsuki, T.; Yamaguchi, M.
Bull. Chim. Soc. Jpn. 1979, 52, 1989-1993.
(113) The C20 ketone is both sterically crowded and electronically
deactivated. Model studies with the C-ring intermediate 49 showed that
this ketone failed to react with the sodium anion of trimethylphosphonoac-
etate, even over the course of several days.
(106) Jacobi has observed that DMF improves the ratio of exocyclic to
endocyclic cleavage in the LiOOH-mediated hydrolysis of N-acyloxazoli-
dinones, cf.: Jacobi, P. A.; Murphee, S.; Rupprecht, F.; Zheng, W. J. Org.
Chem. 1996, 61, 2413-2427.
(114) The olefin geometry of 73 was readily elucidated using 1H NMR
(COSY-90, 500 MHz, C3D6O): C14Heq (δ 3.81) was significantly downfield
of C12Heq (δ 2.37) in the 1H NMR spectrum of the Z diastereomer, and
C12Heq (δ 3.96) was significantly downfield of C14Heq (δ 2.10) in the E
diastereomer. In addition, the C30H resonance exhibited a strong enhance-
ment to the C12Heq resonance in the NOESY spectrum of 73.
(115) (a) Burgess, E. M.; Penton, H. R., Jr.; Taylor, E. A. J. Org. Chem.
1973, 38, 26-31. (b) The Burgess elimination needed to be carried out at
low temperatures and worked up under basic conditions, in contrast to the
literature reports. The acceleration in this reaction is probably attributable
to the acidifying properties of the C20 ketone.
(107) Interestingly, the less basic magnesio species (MgOBn, BnOH)
also inhibited the elimination reaction but resulted in the wrong cleavage
regioselectivity (the anilide 65 was the major product).
(108) Only the final product 67 was purified via flash chromatography.
(109) For another use of HF‚Pyr/MeOH with a mixed methyl ketal, cf.:
Nicolaou, K. C.; Daines, R. A.; Ogawa, Y.; Chakraborty, T. K. J. Am. Chem.
Soc. 1988, 110, 4696-4705.
(110) The best result obtained with the carboxylic acid derivative of 68
(R ) R′ ) H) utilized the thiopyridyl ester (cf. Corey, E. J.; Nicolaou, K.
C. J. Am. Chem. Soc. 1974, 96, 5614-5616) and provided the macrocycle
in 14% yield for the two steps (thioester formation and cyclization).
(111) Felix, A. M.; Heimer, E. P.; Lambros, T. J.; Tzougraki, C.;
Meienhofer, J. J. Org. Chem. 1978, 43, 4194-4196.
(116) (a) Ireland, R. E.; Gleason, J. L.; Gegnas, L. D.; Highsmith, T. K.
J. Org. Chem. 1996, 61, 6856-6872. (b) Ireland, R. E.; Wipf, P.; Armstrong,
J. D., III. J. Org. Chem. 1991, 56, 650-657.

Total Synthesis of Bryostatin 2
J. Am. Chem. Soc., Vol. 121, No. 33, 1999 7551
Scheme 12a
favored in the present system by the conformation of macro-
lactone 73.117
The reduction of ketone 73 proved quite troublesome, in
contrast to the analogous reaction with simple C-ring systems
(cf. 50 f 51, Scheme 9). Reduction of 73 under standard Luche
conditions (NaBH4, CeCl3) resulted in the formation of product
76 (30% yield, 65:35 diastereoselection), in which the C19 ketal
had been hydrolyzed. This reaction was not favorably influenced
by changes in the reaction temperature, time, or workup, and
all attempts to convert 76 into the deprotected 77 were
ineffective due to the unexpected118 stability of the C9 ketal. It
was reasoned that the C19 lactol must be the source of this
unusual stability at C9; accordingly, the in situ acylation of the
nascent hydroxyl C20 was explored. In the event, reduction of
73 with NaBH4 (10:1 THF/MeOH, -15 °C), followed by
anhydrous quenching of the reaction mixture with isobutyral-
dehyde and addition of excess methoxyacetic anhydride (pyri-
dine, DMAP), afforded compound 75 (ca. 50% yield, 60:40
diastereoselection, 25% isolated yield of 75), in which the C19
ketal had been maintained. Fortunately, the stereoselectivity of
the reduction step could be improved by employing the CBS
reagent119 74, which delivered the desired 75 in excellent yield
and selectivity120 (>90% yield, 10:1 diastereoselection, 70%
isolated 75). Methoxyacetate 75 was then transformed to 77
via the three-step sequence: (i) desilylation and hydrolysis of
C9 (PPTS, aqueous THF); (ii) saponification of the methoxy-
acetate protecting group (Na2CO3, MeOH); and (iii) hydrolysis
of C19 and equilibration of C9 (p-TsOH, aqueous MeCN). Not
surprisingly, tetraol 77 existed in a solution conformation
identical to that of bryostatin 10 (1e) according to high-
a Key: (a) (E,E)-2,4-octadienoic acid, DIC, DMAP, CH2Cl2; (b)
DDQ, 10:1 CH2Cl2/pH 7 buffer.
DMAP) for the monoacylation of 77 resulted in bisacylation to
form 78 (R′ ) Me or C7H11) (Scheme 12). This problem was
effectively circumvented by using carbodiimide activation:
exposure of 77 to octadienoic acid, DIC, and DMAP (CH2Cl2,
23 °C) provided a 62% yield of the monoacylated 79. Oxidative
deprotection of the two p-methoxybenzyl ethers at C7 and C26
was best effected (57% yield) under buffered conditions (DDQ,
pH 7 phosphate buffer, CH2Cl2) in order to avoid problems with
the sensitive octadienoate side chain. The bryostatin 2 (1b) thus
obtained was identical to a natural sample by several criteria:
500-MHz 1H NMR (CDCl3 or C6D6, including COSY-90), TLC
Rf, HPLC retention time (including co-injection), and EI mass
spectroscopy. The conversion of bryostatin 2 (1b) into bryostatin
1 (1a) via selective silylation of the C26 hydroxyl group has
been previously described by Pettit,8 and we have exploited
analogous chemistry to convert the C20 acetate analogue of 79
into bryostatin 6 (1c).122
1
resolution H NMR analysis.121
With compound 77 in hand, we focused our attention on the
synthesis of the C20-oxygenated series of bryostatins. Use of
anhydride chemistry (symmetric or Yamaguchi anhydride,
Conclusion
(117) Subjection of open-chain precursors of 70 to KHMDS conditions
resulted in instantaneous elimination of the C3-OTES group.
The total synthesis of bryostatin 2, a biologically potent
marine macrolide, has been accomplished. The synthesis
proceeds over 40 steps (longest linear sequence) and entails the
convergent coupling of the three intact tetrahydropyran rings,
followed by macrocyclization and late-stage installation of the
two enoate functional groups. The absolute stereochemistry of
each fragment was established using three distinct methods: an
oxazolidinone-mediated aldol reaction, a copper-catalyzed enol-
silane addition, and a catalytic epoxidation with kinetic resolu-
tion. The remaining hydroxyl stereocenters were established by
internal relay, with the exception of the C20 stereocenter, which
was set via an oxazaborolidine-based reduction. The synthesis
plan is flexible in the late stages and is thus well suited for
analogue development, a task which has the potential to shed
new light on the complex biological profile of the bryostatins.123
(118) The acid lability of the C9 ketal had been a liability throughout
the course of the synthesis, and all compounds (except 76) that contained
this group had to be handled carefully in order to avoid spurious hydrolysis.
(119) (a) Corey, E. J.; Bakshi, R. K.; Shibata, S.; Chen, C.-P.; Singh, V.
K. J. Am. Chem. Soc. 1987, 109, 7925-7926. (b) For a review, cf.: Corey,
E. J.; Helal, C. J. Angew. Chem., Int. Ed. 1998, 37, 1986-2012. (c)
Interestingly, the use of the B-Me catalyst was found to be superior to the
B-Bu catalyst, presumably due to the large degree of steric crowding
experienced in the transition state for the reduction. In addition, borane
methyl sulfide was the only suitable hydride source for the reaction.
(120) (a) The CBS reagent 74 was uniquely capable of delivering a high-
yielding, stereoselective reaction; all other chiral and achiral reducing agents
either delivered the wrong diastereomer or delivered the desired product in
low yields. (b) The stereochemical assignment was based on three factors:
(i) analogy to the reactions of a system lacking the B-ring enoate (cf. ref
1), in which the stereochemistry of both product C20 diastereomers was
determined by NMR; (ii) NOESY analysis of 77, which showed a cross-
peak between C20H and C34H.; and (iii) the synthesis of 1b and 1c (vide
infra).
(121) (a) With the hydroxyl groups at C3, C9, and C19 deprotected, tetraol
77 is capable of forming the bryostatin hydrogen bonding network (cf. Figure
2). Indeed, analysis of the NOESY spectrum of 77 (500 MHz, benzene-d6)
reveals an identical set of enhancements to that seen in the ROESY spectrum
of bryostatin 10 (cf. ref 13). Additional confirmation for the establishment
of the hydrogen bonding network comes from the chemical shift data for
the hydroxyl protons: the C19 lactol (δ 5.48) and C3 hydroxyl (δ 4.39)
protons are significantly deshielded relative to their C9 (δ 1.86) and C20 (δ
1.32) counterparts. Finally, the IR spectrum of 77 shows two distinct, sharp
bands for O-H stretching (3465 and 3363 cm-1). (b) It should be noted
that compound 77 is the first in the series of synthetic intermediates to
adopt the “natural” conformation depicted in Figure 2. A first-order
investigation of the solution structures of earlier intermediates has indicated
that it is the nature of the C19 ketal, and not the C3 hydroxyl group, that
dictates the preferred solution conformation of the molecule. For a more
detailed discussion, see: Carter, P. H. Ph.D. Thesis, Harvard University,
1998.
Acknowledgment. Support has been provided by the
National Institutes of Health. We thank Dr. Andrew Tyler of
(122) Intermediate 77 was converted into bryostatin 6 (1c) in 27% overall
yield via the following five-step sequence: (i) Ac2O, pyr; (ii) DDQ, 20:1
CH2Cl2:H2O; (iii) TESCl, DMAP; (iv) (n-PrCO)2O, pyr; (v) HF/MeCN/
H2O, 0 °C. The material exhibited spectral characteristics (exact mass FAB-
MS, 500-MHz 1H NMR) completely consistent with those reported by Pettit
et al.: Pettit, G. R.; Kamano, Y.; Herald, C. L.; Tozawa, M. Can. J. Chem.
1985, 63, 1204-1208.
(123) For a completely different approach to analogue development, cf.
the work of Wender et al.: (a) Wender, P. A.; DeBrabander, J.; Harran, P.;
Jimenez, J.-M.; Koehler, M.; Lippa, B.; Park, C.-M. J. Am. Chem. Soc.
1998, 120, 4534-4535. (b) Wender, P. A.; DeBrabander, J.; Harran, P. G.;
Jimenez, J.-M.; Koehler, M. F. T.; Lippa, B.; Park, C.-M.; Siedenbiedel,
C.; Pettit, G. R. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 6624-6629.

7552 J. Am. Chem. Soc., Vol. 121, No. 33, 1999
EVans et al.
the Harvard Mass Spectrometry Facility for providing mass
spectra; Prof. G. R. Pettit for providing an authentic sample of
bryostatin 2 and spectral data for bryostatin 6; and the NIH
BRS Shared Instrumentation Grant Program 1-S10-RR04870
and the NSF (CHE 88-14019) for providing NMR facilities.
Fellowship support from the NSF (P.H.C., E.M.C.), NSERC
(A.B.C., M.L.), Boehringer-Ingelheim (P.H.C.), and Hoffmann
LaRoche (P.H.C.) is also gratefully acknowledged.
Supporting Information Available: Experimental details
and analytical data regarding the preparation of all synthetic
intermediates, and comparison spectra of natural and synthetic
bryostatin 2 (1b) (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
JA990860J