Atom-economic and stereoselective syntheses of the ring A and B subunits of the bryostatins

Article abstract of DOI:10.1002/chem.201002930

This article describes chemoselective and atom-economic methods for the stereoselective assembly of the ring A and B subunits of bryostatins. A Ru-catalyzed tandem alkene-alkyne coupling/Michael addition reaction was developed and applied to the synthesis

Full text of DOI:10.1002/chem.201002930

FULL PAPER
DOI: 10.1002/chem.201002930
Atom-Economic and Stereoselective Syntheses of the Ring A and B Subunits
of the Bryostatins
Barry M. Trost,* Hanbiao Yang, Cheyenne S. Brindle, and Guangbin Dong[a]
Abstract: This article describes chemo-
selective and atom-economic methods
for the stereoselective assembly of the
ring A and B subunits of bryostatins. A
Ru-catalyzed tandem alkene–alkyne
coupling/Michael addition reaction was
developed and applied to the synthesis
of bryostatin ring B. We explored an
acetylide-mediated epoxide-opening/6-
exo-dig cyclization route to access the
bryostatin ring A, although ring A was
eventually furnished through an acid-
catalyzed tandem transketalization/ke-
talization sequence. In addition, a dinu-
clear zinc-catalyzed methyl vinyl
ketone (MVK) aldol strategy was eval-
uated for the construction of the poly-
acetate moiety. Utilization of these
methods ultimately led to the rapid as-
sembly of the northern bryostatin frag-
ment containing both the ring A and B
subunits.
Keywords: aldol reaction · asym-
metric synthesis · chemoselectivity ·
diastereoselectivity · enantioselec-
tivity · enynes · Michael addition ·
natural products
Introduction
As part of our continuing efforts towards the total synthesis
of bryostatins,[1] we explored new methods and strategies for
assembling rings A and B (Scheme 1). The structure of
bryostatin ring B features a 2,6-cis-disubstituted tetrahydro-
pyran with a geometrically defined exocyclic conjugated
methyl ester at the C13 position (bryostatin numbering).
Controlling the olefin geometry of this a,b-unsaturated ester
has been one of the major challenges in syntheses of ring B
of bryostatins.[2] The use of a chiral Horner–Wadsworth–
Emmons (HWE) reagent developed by Evans et al.[2a]
proved to be effective in the syntheses of bryostatins 2 and
3; however, a stoichiometric amount of the expensive chiral
reagent was required for this non-asymmetric transforma-
tion and the E/Z ratio was only 6:1 to 8:1 (Scheme 2). We
envisioned that by utilizing our Ru-catalyzed enyne coupling
reaction with multi-functionalized substrates, we would be
able to generate the required ring-B motif in a highly regio-
and chemoselective fashion. A single olefin isomer is expect-
ed and this approach would also be catalytic and atom eco-
nomic.
Scheme 1. Retrosynthetic analysis of the ring A and B subunits of bryo-
AHCTUNGTERGsNNUN tatins. Pg=protecting group.
The bryostatin ring-A region contains multiple 1,3-anti
diol units, and thus an aldol approach seems to be one of
the most straightforward ways to provide this motif. Indeed,
in all previous total syntheses aldol reactions have been em-
ployed for the synthesis of ring A; moreover, these aldol re-
actions all relied on the generation of a stoichiometric
amount of metal enolates.[3] For example, in the synthesis by
Masamune et al., boron enolates were utilized to form the
C10–C11 and C2–C3 bonds;[3a,b] in the synthesis by Evans
et al., a boron enolate was used to form the C6–C7 bond
and a titanium enolate was used to form the C4–C5 bond;[2a]
and in the synthesis of bryostatin 3 by Ohmori et al., a lithi-
[a] Prof. B. M. Trost, Dr. H. Yang, Dr. C. S. Brindle, Dr. G. Dong
Department of Chemistry, Stanford University
Stanford, California, 94305-5080 (USA)
Fax : (+1)650-725-0002
E-mail: bmtrost@stanford.edu
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201002930. It contains experi-
mental details for the synthesis of compounds 5, 16–19, 24, 25, 27, 28,
46, 49a–g, 50a, 53–60, 62, 63, and 67, as well as spectral data.
Chem. Eur. J. 2011, 17, 9777 – 9788
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
9777

for further elaboration and union with the other fragments.
Herein, we describe our detailed efforts towards developing
methods for the atom-economic and stereoselective synthe-
ses of bryostatin rings A and B.[5]
Results and Discussion
First generation strategy: Our initial strategy for the synthe-
sis of the northern rings A and B is shown in Scheme 3. We
envisioned that the C9 methyl ketal in compound 2 could
arise from the addition of CH3OH to the C9–C10 enol ether
in 3 under acidic conditions. Enol ether 3 could be obtained
through a 6-exo-dig oxypalladation cyclization from homo-
propargyl alcohol 4. Inspection of intermediate 4 suggested
disconnection at the C8–C9 bond, resulting in fragments 5
and 6. In the forward direction, by using a nucleophile de-
rived from alkyne 5, attack at the C8 terminus of epoxide 6
was expected to furnish secondary alcohol 4 in the presence
of a Lewis acid. Alkyne fragment 5, containing the ring B
subunit would be derived from alkene 7, which can be ulti-
mately prepared from a Ru-catalyzed alkene–alkyne cou-
pling followed by a Pd-catalyzed cyclization. Epoxide frag-
ment 6 could be prepared from a dinuclear zinc–ProPhenol-
catalyzed aldol addition between aldehyde 8 and MVK (9).
Scheme 2. Installation of the geometrically defined exo-enoate of bryo-
statins. i) [CpRu(CH3CN)3][PF6] (cat.; Cp=cyclopentadienyl). TMS=tri-
methylsilyl.
G
E
ACHTUNGTRENNUNG
um enolate was employed to form the C4–C5 bond.[3c,d]
From an atom-economy viewpoint, the use of a catalytic
asymmetric direct aldol reaction would allow access to these
poly-1,3-diol units in a more efficient manner. Recently, we
have developed a highly enantioselective dinuclear Zn-cata-
lyzed direct aldol reaction in which methyl vinyl ketone
(MVK) was used as a bifunctional nucleophile (Scheme 3).[4]
With this method, we envisioned that the C3–C9 fragment
along with the C5 and C7 stereocenters would be construct-
ed in a rapid fashion; meanwhile, an olefin as a precursor
for C3–C4 as in enone 9 could provide a functional handle
Synthesis of alkyne fragment 5: The synthesis of terminal
alkyne 5 commenced with (S)-glycidol 12, which was con-
verted into alkynyl alcohol 10 in two straightforward steps
consisting of PMB protection followed by epoxide opening
Scheme 3. First generation approach for the construction of rings A and B of bryostatins. TBDPS=tert-butyldiphenylsilyl, PMB=para-methoxybenzyl.
9778
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 9777 – 9788

Synthesis of Bryostatin Rings A and B
FULL PAPER
with lithium (trimethylsilyl)acetylide (Scheme 4).[6] The cou-
pling between alkyne 10 and alkene 11 by using the tandem
procedure[7] of alkene–alkyne coupling followed by Pd-cata-
unable to achieve the selective oxidation of the terminal
olefin with OsO4 (7!14).[8] Given the electron-rich nature
of the vinylsilane in diene 7 and relative insensitivity of the
epoxidation reaction with peracids towards steric factors, we
decided to protect the vinylsilane as an epoxysilane. Treat-
ment of vinylsilane 7 with mCPBA (1.5 equiv) at 08C gave
epoxysilane 16 in 95% yield as a 3:1 diastereomeric mixture
(Scheme 6). Oxidative cleavage of the monosubstituted
olefin followed by alkyne formation[9] went smoothly to de-
liver alkyne 17 in 83% yield over three steps.
Scheme 4. A tandem Ru-catalyzed ene–yne coupling/Pd-catalyzed allylic
alklylation strategy for the formation of ring B. i) [CpRu
(10 mol%), acetone, RT; ii) [Pd2A(dba)3]·CHCl3 (dba=dibenzylideneace-
tone; 2 mol%), (S,S)-LST (6 mol%), triethylamine (TEA), CH2Cl2.
ACHTUNGTREN(NUG CH3CN)3]ACHTUGNTREN[NUGN PF6]
CTHUNGTRENNUNG
lyzed asymmetric allylic alkylACTHNUGTRNEUNGation (AAA) proceeded un-
Scheme 6. Protection of the vinyl silane as an epoxide to achieve selec-
tive functionalization of the terminal olefin. i) meta-Chloroperoxybenzoic
acid (mCPBA), Li2CO3 (30%), CH2Cl2, 08C ; ii) OsO4, N-methylmorpho-
line N-oxide (NMO), acetone/H2O; iii) NaIO4, THF/H2O; iv) Ohira–
Bestmann reagent, K2CO3, CH3OH; v) Sodium hexamethyldisilazane
(NaHMDS), triethylsilyl chloride (TESCl), THF, ꢀ788C; vi) [Rh2-
eventfully to give tetrahydropyran 7. However, an excess of
alkene 11 (5 equiv) was needed for complete consumption
of alkyne 10, and tetrahydropyran 7 had the same Rf value
as diene 11. Thus, a two-step procedure was used to facili-
tate product purification. It is noteworthy that the stereo-
chemistry at C11 is under ligand control: if the standard
Trost ligand (R,R)-LST was employed, the product was ob-
tained as a 97:3 diastereomeric mixture favoring the thermo-
dynamically disfavored 2,6-trans-tetrahydropyran, whereas
(S,S)-LST gave a 98:2 ratio favoring cis-tetrahydropyran 7.
Interestingly, the use of 1,2-bis(diphenylphosphino)ethane
(dppe) as a ligand gave a nearly 1:1 ratio of a cis/trans iso-
meric mixture. By following this procedure, tetrahydropyran
7 was obtained on a gram scale.
An efficient way to convert alkene 7 into alkyne 5 would
be to oxidatively cleave the terminal alkene, followed by al-
kynylation of the resulting aldehyde, and then transforma-
tion of the OPMB ether into an olefin (Scheme 5). Unfortu-
nately, the electron-rich disubstituted vinylsilane was also re-
active under the dihydroxylation conditions. Despite the use
of a variety of oxidative cleavage conditions, we were
AHCTUNGTREUNNG(N OAc)4], dimethyl diazomelonate, C6H6, 808C; vii) 2,3-dichloro-5,6-dicya-
no-1,4-benzoquinone (DDQ), CH2Cl2/H2O; viii) 2-iodoxybenzoic acid
(IBX), CH3CN, 808C; ix) Ph3P=CH2, toluene, 08C; x) tetra-n-butylammo-
nium fluoride (TBAF), THF/H2O (95:5), RT.
A deoxygenation reaction at this stage to unmask the exo-
cyclic vinylsilane proved to be nontrivial (Table 1). Martin
and Ganem reported a mild Rh-catalyzed deoxygenation of
epoxides with dimethyl diazomalonate in refluxing benzene
with retention of olefin geometry.[10] Under their conditions
the deoxygenation of 17 gave the desired vinylsilane 21 as a
single isomer (Table 1, entry 1), albeit in low yield (15%).
Switching the solvent to chlorobenzene gave only trace
amounts of product (Table 1, entry 2). The use of alumina
with a catalytic amount of HgCl2 in refluxing toluene/isopro-
panol[11] led to clean deoxygenation by both TLC and crude
1H NMR analysis (Table 1,
entry 3). Unfortunately, the
product was obtained as an in-
separable mixture of exocyclic
vinylsilane isomers. Deoxygena-
tion employing low valent tung-
sten, originally reported by
Scheme 5. Attempted direct functionalization of the terminal olefin proved non-selective.
Sharpless and Umbreit,[12] gave
Chem. Eur. J. 2011, 17, 9777 – 9788
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9779

B. M. Trost et al.
Table 1. Deoxygenation studies for compounds 17 and 20.
but with the opposite stereo-
chemistry.[15]
Changing
the
starting glycidol and Trost
ligand would allow access to
the desired enantiomer.
Efforts towards the synthesis of
epoxide fragment 6: We envi-
sioned that epoxide fragment 6
could be accessed by an ester
aldol or Reformatsky reaction
from aldehyde 26 (Scheme 9).
Aldehyde 26 could be derived
from allyl alcohol intermediate
27, which can ultimately be pre-
pared from diol 28.
X
Conditions
[Rh2A(OAc)4] (10 mol%), C6H6, 80 8C
[Rh2A(OAc)4] (10 mol%), C6H5Cl, 80 8C
Product
Yield [%]
1
2
3
4
5
6
H
H
H
H
H
TES
N
G
21
21
15
trace
clean reaction
0
0
52
Al/HgCl2, C6H6/iPrOH, reflux
21+double bond isomer
baseline material
recovered 17
WCl6, nBuLi (3 equiv), ꢀ78 8C!RT
(PhO)3PCH3I, BF3·OEt2, RT
A
N
18
no reaction at ꢀ788C and decomposition upon warming
(Table 1, entry 4). Treatment with (PhO)3PCH3I[13] in the
presence of BF3·OEt2 resulted in no reaction at room tem-
perature (Table 1, entry 5). As the low yield in Table 1,
entry 1 was likely caused by reaction between the terminal
alkyne and a Rh carbenoid, the terminal alkyne was protect-
ed with a TES group and the resulting silylalkyne 20 was
Scheme 8. Inversion of the olefin geometry through manipulation of the
epoxide. i) HBr, Et2O, ꢀ508C; ii) BF3·Et2O, RT; iii) IBX, CH3CN, 608C;
iv) Ohira–Bestmann reagent, K2CO3, CH3OH.
subjected to the Rh-catalyzed deoxygenation ([Rh2ACTHNUTRGNEN(UG OAc)4],
benzene, 808C). Gratifyingly, the desired vinylsilane (18)
was obtained in 52% yield. With compound 18 in hand,
alkyne 5 was obtained smoothly by using a four-step se-
quence of PMB cleavage, IBX oxidation, Wittig olefination,
and TES removal.
Alternatively, the ring B alkyne subunit could be synthe-
sized by taking advantage of the internal symmetry present
in pyran 7 (Scheme 7). If the exocyclic double-bond geome-
try is inverted, the terminal alkene can be mapped onto the
product. To obtain alkyne 5 with the desired absolute ste-
reochemistry for the bryostatin synthesis, (R)-glycidol and
the (R,R)-Trost ligand would be needed.
Scheme 9. Retrosynthetic analysis of the 1,3-anti-diol 6 region of bryosta-
tins. TBS=tert-butyldimethylsilyl.
In the forward direction, by utilizing the dinuclear zinc-
catalyzed direct aldol addition reaction[4] followed by cis-re-
duction of the crude aldol adduct, diol 28 was rapidly ob-
tained from aldehyde 8 and MVK 9 (Table 2). The yield, d.r.
and enantioselectivity were carefully optimized as shown in
Table 2. Initially, the effect of temperature was examined
(Table 2, entries 1–3). The reaction showed a small inverse
relationship between temperature and enantioselectivity, a
trait that has been observed previously with this catalyst
system.[16] The reaction gave a better yield at higher temper-
atures, though at ambient temperature the product was ob-
tained impurely (Table 2, entry 3). Longer reaction times
caused increased decomposition and lower isolated yields
(Table 2, entries 3–5). Increasing the reaction temperature
to 508C only led to decomposition (Table 2, entry 6). Sur-
prisingly, the use of undistilled MVK at 48C led to an im-
provement in the isolated yield (Table 2, entry 7). As com-
mercial MVK contains a small amount of acetic acid and hy-
Scheme 7. Route to the desired enyne through the enantiomer of diene 7.
This route was explored by use of the previously prepared
epoxysilane 16 (Scheme 8). Bromide 24 was obtained from
epoxide 16 in a two-step sequence (bromohydrin formation
and siloxy elimination) in 81% yield.[14] Subsequent oxida-
tion followed by alkyne formation gave alkyne 25, which
possessed all the required functionalities for the epoxide
coupling strategy (see Equations (1) and (2), given later)
9780
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 9777 – 9788

Synthesis of Bryostatin Rings A and B
FULL PAPER
Table 2. Optimization of the dinuclear zinc-catalyzed direct aldol addition reaction.[a]
fortunately,
chemoselective
PMB protection of the allylic
alcohol proved problematic.
For example, treatment of 28
with PMBCl and sodium hy-
dride in DMF, or use of PMB
trichloroACTHNUGRTENUNGacetimidate with cata-
lytic trifluoromethanesulfonic
acid, gave only recovered start-
ing material. Attempted forma-
tion of the tin acetal was also
unsuccessful, even after reflux
for 36 h. The use of catalytic
Catalyst
[mol%]
T
[8C]
T
[h]
Additives
Yield
[%][b]
ee
d.r.[d]
G
[%][c]
1
2
3
4
5
6
7
8
9
10
11
12
13
10
10
10
10
10
10
10
10
5
10
10
10
10
ꢀ15
4
7
7
7
9
24
9
7
7
7
7
iPrOH (5 equiv)
iPrOH (5 equiv)
iPrOH (5 equiv)
iPrOH (5 equiv)
iPrOH (5 equiv)
iPrOH (5 equiv)
iPrOH (5 equiv), AcOH (0.6%), hydroquinone (3%)
iPrOH (5 equiv), AcOH (1.6%), hydroquinone (3%)
iPrOH (5 equiv), AcOH (1.6%), hydroquinone (3%)
iPrOH (5 equiv), AcOH (1.6%), hydroquinone (3%)
iPrOH (5 equiv), AcOH (1.6%), hydroquinone (3%)
iPrOH (5 equiv), AcOH (1.6%), hydroquinone (3%)
methyl b-hydroxypropionate (16 equiv)
20
25
41
25
25
0
45
52
50
32
40
trace
16
93
97
97
96
98
–
98
97
90
98
92
–
24:1
22:1
18:1
38:1
26:1
–
35:1
31:1
25:1
17:1
50:1
–
25
25
25
50
4
4
4
25
25
4
tetrabutylammonium
iodide
(TBAI) to further activate the
PMBCl in situ gave no desired
product at ambient tempera-
ture, but did give the desired
product at reflux, though as a
mixture of the two isomeric
products.
Consequently, an alternative
approach was undertaken in
which the diol was transformed
into the diastereomeric cyclic
sulfites 33, thus activating the
4
20
7
4
84
25:1
[a] Conditions: (S,S)-29, toluene, 4 ꢁ molecular sieves, then Et2BOMe, NaBH4. [b] Isolated yield. [c] ee was
determined by HPLC. [d] d.r. was determined by H NMR spectroscopy.
1
droquinone to prevent polymerization, acetic acid is proba-
bly responsible for the increase in yield. Indeed, adding an
additional 1% acetic acid gave an even better result
(Table 2, entry 8). It is likely that acetic acid could serve as
a buffer for the system, thus mitigating the undesired elimi-
nation pathway. Lowering the catalyst loading gave a slight-
ly reduced yield, with only a modest drop in enantioselectiv-
ity (Table 2, entry 9). Raising the temperature gave a poor
yield, even after shorter reaction times (Table 2, entries 10–
11). Increasing the amount of added acetic acid to 2% re-
sulted in the observation of only trace amounts of product
(Table 2, entry 12). The use of methyl b-hydroxypropionate,
an achiral additive that found some success in a similar
system, gave little conversion.[16] From these studies, the
most practical set of conditions emerged as those shown in
Table 2, entry 8, and the desired reduced adduct 28 was ob-
tained in 52% yield over two steps with 97% ee and 31:1
d.r. in favor of the desired diastereomer. This method pro-
vides a facile route for the installation of two of the stereo-
centers in epoxide fragment 6.
C7 alcohol for intramolecular SN2 inversion to form the ep-
oxide, while differentiating the two alcohols by tying them
into a ring and thus preventing attack of the tertiary alcohol
at the allylic C5 site (Scheme 11). After generation of the
Scheme 11. Epoxide formation. i) SOCl2, Et3N, CH2Cl2, 08C; ii) TBAF,
THF, reflux.
diastereomeric sulfites, a variety of conditions were attempt-
ed for the desilylation to form desired epoxide 27. Treat-
ment with an anhydrous fluoride source, tetrabutylammoni-
um triphenyldifluorosilicate (TBAT) or 3HF·NEt3, did not
give the desired product or recovered starting material. If
TBAF was used as the fluoride source in THF, the desired
product was isolated in an 18% yield at ambient tempera-
ture, and 48% yield under reflux. The reaction was shown
to be successful by TLC, but the isolated yield was low. This
is probably caused by the volatility of product 27, or the in-
stability of the compound, as it was found to decompose
upon storage at ꢀ158C. In an
Protection of the less hindered allylic alcohol as a PMB
ether was envisioned to differentiate the two alcohols, which
vary in their steric environment (Scheme 10). After PMB in-
stallation, transformation of the more hindered alcohol into
a leaving group would allow for epoxide formation. Un-
effort to solve both of these
problems, the epoxide was sub-
jected to silylation conditions
with TBDPSCl and imidazole
in DMF, but this resulted in de-
Scheme 10. Failed elaboration of diol 28. Lg=leaving group.
composition.
Chem. Eur. J. 2011, 17, 9777 – 9788
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9781

B. M. Trost et al.
Meanwhile, we conducted a parallel study to investigate
the feasibility of the proposed alkyne–epoxide coupling re-
action (see Scheme 3). Sterically, although the C8 position is
more substituted and therefore more hindered than C7, nu-
cleophilic attack at C7 would also be difficult due to its
pseudoneopentyl nature (next to C8). Accordingly, the
steric difference between C7 and C8 during nucleophilic
attack could be small. On the other hand, when Lewis acids
are employed in the reaction, their coordination to the ep-
ꢀ
ꢀ
oxide will weaken the C8 O bond more than the C7 O
bond because of the relief of steric congestion around C8,
and the ability of a tertiary carbon to stabilize a positive
charge. Therefore, electronically, nucleophilic attack at the
C8 terminus of the epoxide would be favored. A recent pub-
lication by Zhao and Pagenkopf provides further support
for this analysis.[17] To this end, model epoxide 37 was syn-
thesized from 3-pentyne-1-ol 34 in three steps
(Scheme 12).[18] Carbometallation of 3-pentyne-1-ol 34
ZnACHTUNGERTN(GNUN OTf)2, AlAHCTUNEGTRGN(NNU OTf)3, or BACTHNUGTRENN(UNG C6F5)3), or the order of addition
did not provide promising results.
Second generation strategy: Although there are other possi-
ble ways to carry out the epoxide-opening strategy, the diffi-
culties associated with the ring-opening reaction prompted
us to re-evaluate our approach to the northern fragment. At
the outset, we decided to form the hindered C8–C9 bond
early in the synthesis. Our alternative, second-generation
strategy is shown in Scheme 13. We envisioned that the ring
A moiety, along with the C9 methyl ketal in 2, could be in-
stalled through an acid-catalyzed ketalization reaction with
Scheme 12. Preparation of model epoxide 37. i) Al
CH2Cl2, ꢀ788C; ii) TBDPSCl, imidazole, DMF; iii) mCPBA, CH2Cl2.
ACHTNUGTNER(NUGN CH3)3, TiCl4, THF,
(TiCl4, Al
G
complete. The product was obtained as a mixture of starting
material 34 and olefin 35 in approximately a 1:2 ratio, which
was carried over two steps (TBDPS protection and epoxida-
tion) to provide epoxide 37.
ꢀ
CH3OH from hydroxyketone 42. The presence of the C O
bond at C11, which is b to the C9 ketone, suggested that a
ꢀ
Michael addition would construct the C11 O bond and ring
This set the stage for examination of the proposed epox-
ide-opening reaction. Following the procedure by Zhao and
Pagenkopf,[17] lithium acetylide trimethylaluminum ate com-
plex 38 (derived from pentyne) reacted with trisubstituted
epoxide 37 in the presence of BF3·OEt2, providing secondary
alcohol 39, which indicates that the attack at the more hin-
dered position was indeed favored [Eq. (1)]. Formation of
the other regioisomer could not be completely excluded
B simultaneously. The requisite a,b-unsaturated ketone 43
contains a 1,4-diene, which is the characteristic functionality
obtained from our Ru-catalyzed alkene–alkyne coupling re-
action. Enone fragment 44 could be derived from aldehyde
45 through functional group manipulations in which the C2–
C3 bond could be formed by either an ester aldol or a Re-
formatsky reaction. In analogy with our first-generation
strategy, the aldehyde intermediate (45) would be prepared
from diol 46, which could be efficiently synthesized through
a dinuclear Zn-catalyzed direct aldol reaction from aldehyde
47 and MVK 9.
This new strategy revealed a novel enyne coupling reac-
tion to construct ring B of bryostatins because one of the
coupling partners, the b,g-enone, had never before been
used in the Ru-catalyzed ene reaction. One immediate con-
cern was the stability of b,g-enones in the presence of the
cationic Ru catalyst since they are prone to isomerization in
the presence of Brønsted or Lewis acids. Thus, we decided
to address these issues with a model system [Eq. (3)].[15] The
initial model studies gave surprising but gratifying results
(Table 3). If a 1:1 mixture of alkyne 10 and b,g-enone 48a
1
since the H NMR spectrum was not very clean; however,
the major signal at 3.64 ppm (CDCl3, ddd, J=1.5, 4.0,
10.5 Hz, assigned to the proton next to the hydroxyl group
in 39) indicated that secondary alcohol 39 was the major
component (>85%). The use of the trimethylaluminum ate
complex is crucial since lithiated pentyne resulted in no re-
action.
Encouraged by this model study, the coupling between
alkyne 5 and epoxide 37 was subsequently attempted. Un-
fortunately, under the Zhao and Pagenkopf conditions, no
reaction occurred. Upon warming, epoxide 37 underwent re-
arrangement to give ketone 41 [Eq. (2)]. Changing the sol-
vent (THF, hexane, or toluene), the Lewis acid (AlACHTNUGTRNEG(UN CH3)3,
9782
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 9777 – 9788

Synthesis of Bryostatin Rings A and B
FULL PAPER
thylsilyl)acetylide (method A;
Scheme 14); b,g-enones were
accessed either through a two-
step sequence (method B: allyl-
A
then
oxidation;
Scheme 14) or a BiCl3–NaI or
BiCl3–ZnI2 catalyzed Friedel–
Crafts acylation reported by
Le Roux and Dubac[19] (method
C; Scheme 14).
The propargylic alcohols and
b,g-enones were then subjected
to the optimized coupling/cycli-
zation conditions. The results
are summarized in Table 4. A
variety of functional groups, in-
cluding PMB, TBS, and acetyl
Scheme 13. Retrosynthetic analysis featuring an ene–yne coupling/Michael addition reaction.
was treated with [CpRuACTHNUTRGENN(UG CH3CN)3]ACHTUNGTRNEN[GUN PF6] (10 mol%) in ace-
tone at room temperature, diene 51a was not observed. In-
stead the cyclized product, 2,6-cis-dihydropyran 49a, was
isolated in a 31% yield (Table 3, entry 1). The cis configura-
tion was confirmed by nOe studies. The potential minor
product 50a (2,6-trans) was not isolated in this case. To ach-
ieve both reasonable conversion and yield, excess enone was
employed (Table 3, entry 2). Higher concentrations led to
increased conversion, as typically expected for a bimolecular
reaction (Table 3, entry 3). Interestingly, increased catalyst
loading did not improve the yield substantially (Table 3,
entry 4). The best yield was achieved by using 3 equivalents
of enone with acetone as the solvent at 0.4m concentration
of alkyne (Table 3, entry 5; recovery of enone 48a was not
attempted due to its volatility.)
Encouraged by this model study, we decided to probe the
scope and limitations of this reaction, with an emphasis on
substrates that may prove useful for the construction of
rings A and B of bryostatins. Accordingly, a variety of ho-
mopropargylic alcohols and b,g-enones were prepared
(Scheme 14). In general, the homopropargylic alcohols were
synthesized through epoxide opening with lithium (trime-
are tolerated in this reaction. Since excess enone was neces-
sary to achieve a reasonable conversion and yield, the ability
to recover the unreacted enone was crucial. To our delight,
enones with tert-alkyl groups could be mostly recovered
(Table 4, entries 3, 5, and 6), which boded well for our
bryostatin synthesis. On the other hand, if dec-1-ene-4-one
Table 3. Optimization of the Ru-catalyzed ene–yne/Michael addition re-
action.
Concentration
[m]
Catalyst loading
[mol%]
48a:10
Yield 49a [%]
(50a [%])
N
ACHTUNGTRENNUNG
1
2
3
4
5
0.1
0.1
0.4
0.1
0.4
10
10
10
30
10
1:1
1.8:1
1.2:1
1:1
31
43 (6)
41 (6)
36
Scheme 14. Synthesis of substrates for the ene–yne coupling reaction.
i) BF3·Et2O, THF, ꢀ788C; ii) Dess–Martin Oxidation; iii) BiCl3
(2.5 mol%), NaI (7.5 mol%), CH2Cl2, RT (nHex=n-hexyl).
3:1
68 (15)
Chem. Eur. J. 2011, 17, 9777 – 9788
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9783

B. M. Trost et al.
Table 4. Tandem alkyne–enone coupling/Michael addition reaction.[a]
ratio were thwarted by acid- and base-mediated decomposi-
tion.
The geometrically defined exocyclic vinylsilane resulting
from the ruthenium-catalyzed enyne-coupling/Michael addi-
tion provides a convenient handle for further functionaliza-
tion (Scheme 15). Protiodesilylation of 49c went smoothly
to give terminal alkene 53 without double-bond migration.
Vinyl iodide formation also proceeded efficiently and the
product was subsequently carbonylated to give a,b-unsatu-
rated methyl ester 54. Furthermore, we were able to invert
the double-bond geometry through an epoxidation, bromo-
hydrin formation, and siloxy elimination sequence to give
inverted bromide 56 (Scheme 16), which was then carbonyl-
R1
R2
Yield 49 [%]
(recovered 48 [equiv])
1
2
tBu
nHex
Et
Et
80
62
3
4
5
CH2OPMB (10)
CH2OPMB (10)
CH2OPMB (10)
69 (1.8)
58
AHCTUNGTREGaNNNU ted to give methyl enoate 57 with high efficiency. Thus,
from one geometrically defined vinylsilane, either geometric
isomer of the exocyclic enoate is cleanly available.
77 (1.9)
6
CH2OPMB (10)
39 (1.9)
Synthesis of b,g-enone fragment 44: Efforts were next direct-
ed towards the synthesis of the b,g-enone fragment for the
real system. Aldol adduct 58 is available in good yield and
excellent enantioselectivity from the dinuclear zinc-cata-
lyzed MVK aldol reaction with aldehyde 47 (containing one
more methylene unit than aldehyde 8 used for the epoxide-
opening strategy).[20] The opposite enantiomer of the cata-
lyst [(R,R)-29] was used for this reaction, as the C7 stereo-
center would not be inverted in this strategy (Scheme 17).
To obtain the desired stereochemistry at the C5 position,
a trans reduction of b-hydroxyl ketone 58 was required. The
Evans acetoxyborohydride reagent is a commonly employed
strategy for the synthesis of anti diols from b-hydroxyke-
tones.[21] However, this reaction proceeds slowly at low tem-
perature, but gives poorer selectivities at higher tempera-
ture. To balance reactivity and selectivity, a brief survey of
reaction conditions was undertaken (Table 5). The use of
sodium acetoxyborohydride at ꢀ108C gave incomplete con-
version and moderate selectivity (Table 5, entry 1). Switch-
ing to the more reactive tetramethylammonium counterion,
the reaction temperature could be lowered to ꢀ208C, and
both the yield and diastereoselectivity were improved
(Table 5, entry 2). Increasing the amount of the reducing
agent increased the conversion and yield, though it lowered
the diastereoselectivity (Table 5, entry 3).
[a] 48 (3 equiv), 52 (1 equiv), [CpRu
(0.5m), RT, 40 h.
ACHTUTGNRENNU(G CH3CN)3]ACHTUGNTREN[NUGN PF6] (10 mol%), acetone
was used, it was recovered together with 15–20% of the
a,b-enone isomer as an inseparable mixture (Table 4,
entry 2). Interestingly, complete chemoselectivity was ob-
served in the reaction of a compound with two different
types of double bond (Table 4, entry 6). No product derived
from the coupling of the alkyne with the double bond bear-
ing an allylic oxygen was detected. The cis/trans ratios
ranged from 5:1–8:1. Preliminary studies suggested that the
cis- and trans-isomers were not equilibrating under the reac-
tion conditions. Attempts to establish the thermodynamic
Diol 46 was then protected with a cyclopentanone-derived
ketal (Scheme 18). Subsequent removal of the silyl group
provided free alcohol 60, which set the stage for the installa-
tion of the b,g-unsaturated ketone. Unexpectedly, oxidation
of the primary alcohol by using IBX or TPAP/NMO caused
rearrangement of the cyclic acetal, and the resultant allylic
Scheme 15. Functionalization of the vinyl silane. i) Trifluoroacetic acid
(TFA), toluene, 08C to RT; ii) N-iodosuccinimide (NIS), CH3CN;
iii) [PdCl
(PhCN)2],
1,1’-bis(diphenylphosphino)ferrocene
(dppf),
CH3OH, CO.
Scheme 16. Inversion of the olefin geometry. i) mCPBA, Li2CO3, CH2Cl2; ii) aqueous HBr, CH3OH, ꢀ788C; iii) BF3·OEt2, CH2Cl2, ꢀ50 to 108C; iv) [Pd-
(PPh3)4], DMF, CH3OH, CO, 858C.
ACHTUNGTRENNUNG
9784
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 9777 – 9788

Synthesis of Bryostatin Rings A and B
FULL PAPER
Scheme 18. i) Camforsulfonic acid (CSA; 1 mol%), CH2Cl2, 08C;
ii) TBAF, THF; iii) IBX, AcCN; iv) tetrapropylammonium perruthenate
(TPAP), NMO; v) DMSO, (COCl)2, Et3N, CH2Cl2, ꢀ788C, then allyl-
Scheme 17. ProPhenol–zinc-catalyzed asymmetric aldol reaction. i) (R,R)-
29, toluene, 4 ꢁ molecular sieves, iPrOH (5 equiv), ꢀ158C.
AHCTUNGTERNNGMUN gBr, Et2O; vi) Dess–Martin periodinane (DMP), NaHCO3, CH2Cl2.
Table 5. The trans reduction of b-hydroxyl ketone.
theless, the success of this method in uniting the
two components and forming the desired bryostatin
ring B proved the utility of the Ru-catalyzed
tandem coupling/cyclization reaction as a viable
strategy for bryostatin synthesis. Subsequently,
treatment of tetrahydropyran 63 with a catalytic
amount of CSA in MeOH resulted in a tandem
transketalization/ketalization sequence that ulti-
mately provided compound 67, containing both
Conditions
Yield 46 [%]
d.r.[b]
(recovered 58 [%])[a]
1
2
3
NaB
NMe4B
NMe4B
G
56 (11)
54 (26)
5:1
9:1
5:1
(OAc)3H (5 equiv), AcOH/acetone, ꢀ20 8C
[a] Isolated yield. [b] d.r. was determined by H NMR spectroscopy.
alcohol was oxidized to give enone 61. Fortunately, oxida-
tion under Moffat–Swern conditions, followed by quenching
with allylmagnesium bromide, provided the desired homoal-
lylic alcohol. Oxidation of the resulting diastereomeric alco-
hols with DMP furnished the desired b,g-unsaturated ketone
62 in 71% yield over the two steps.
rings A and B of bryoACHTGNUTERNNUsG tatins.
Conclusion
During our continuing efforts towards the total synthesis of
the bryostatin family, we have developed methods for the
stereoselective assembly of the ring A and B subunits. For
the synthesis of bryostatin ring B, a new method for the ste-
Proof of principle: Synthesis of fragment 67, containing
both the ring A and B subunits: At this stage, we investigat-
ed the ene–yne coupling/1,4-ad-
dition reaction to construct
bryostatin ring B (Scheme 19).
Gratifyingly, submission of b,g-
unsaturated ketone 62 and
alkyne 10 to the reaction condi-
tions developed earlier (see
Tables 3 and 4) afforded desired
tetrahydropyran 63, in spite of
the presence of a second termi-
nal olefin. A portion of starting
alkene 62 could be recovered.
The low yield of this reaction
was attributed to the instability
of the ketal moiety to the
Lewis acidic ruthenium catalyst,
as well as competitive binding
of the allyl ether double bond
with the catalyst. Since this was
only
effort was made to optimize the
yield for this example. Never- ii) CSA (10 mol%), MeOH.
a proof-of-principle, no
Scheme 19. Completion of rings A and B of bryostatins. i) [CpRuACTHNUTRGENN(UG CH3CN)3]ACHUTGNTREN[NUNG PF6] (10 mol%), acetone, RT;
Chem. Eur. J. 2011, 17, 9777 – 9788
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9785

B. M. Trost et al.
The combined organic layers were dried (Na2SO4), filtered, and concen-
trated in vacuo to give the aldol adduct as a yellow oil, which was used
directly in the following reaction. 1H NMR (400 MHz, CDCl3): d=6.38
(dd, J=10.4, 17.6 Hz, 1H), 6.23 (dd, J=1.2, 17.6 Hz, 1H), 5.86 (dd, J=
1.2, 10.4 Hz, 1H), 3.82 (dd, J=2.4, 9.6 Hz, 1H), 2.83 (dd, J=2.4, 16.4 Hz,
1H), 2.64 (dd, J=9.6, 16.4 Hz, 1H), 1.23 (s, 3H), 1.21 (s, 3H), 0.85 (s,
9H), 0.10 (s, 3H), 0.09 ppm (s, 3H); IR (thin film): n˜ =3423, 2952, 2929,
2860, 1677, 1590, 1460, 1404, 1363, 1252, 1155, 1122, 1090, 1044, 831,
reoselective synthesis of tetrahydropyran rings was dis-
closed, which features a Ru-catalyzed tandem alkene–
alkyne coupling/Michael addition sequence. For the assem-
bly of the bryostatin ring A subunit, a dinuclear zinc-cata-
lyzed MVK aldol strategy was explored for the construction
of the polyacetate entities. Furthermore, an acid-catalyzed
cascade transketalization/ketalization sequence provided an
advanced intermediate containing both rings A and B. All
of these highly chemoselective and/or atom-economical
methods set the stage for further advances towards the total
synthesis of the bryostatins.[22]
771 cmꢀ1
.
Reduction reaction: The crude aldol adduct from above was dissolved in
20% methanol/THF (10.0 mL) and then diethylmethoxyborane (0.8 mL,
6.09 mmol) was added. The reaction vessel was cooled to ꢀ788C and
sodium borohydride (49.2 mg, 1.30 mmol) was added. The reaction mix-
ture was stirred for 2.5 h, quenched with glacial acetic acid (0.1 mL) and
warmed to ambient temperature. The mixture was then poured into di-
ethyl ether and washed four times with saturated aqueous NaHCO3. The
organic layer was dried (MgSO4), filtered, concentrated in vacuo, and
then chromatographed with a solvent gradient (0–1% v/v diethyl ether/
petroleum ether) to give 28 (141.7 mg, 52% yield over two steps, 97%
ee, 31:1 d.r.) as a yellow oil. Chiral GC (Cyclosil B): 1008C for 20 min
then ramped up at 58Cminꢀ1 to 2008C: tr (major enantiomer, minor dia-
stereomer)=35.81 min, tr (minor enantiomer, minor diastereomer)=
35.91 min, tr (minor enantiomer, major diastereomer)=36.09 min, tr
(major enantiomer, major diastereomer)=36.25 min; Rf =0.8 (10% v/v
ethyl acetate/petroleum ether); [a]2D5 =ꢀ1.2 (c=1.0 in CHCl3); 1H NMR
(500 MHz, CDCl3): d=5.85 (ddd, J=5.4, 10.5, 17.1 Hz, 1H) 5.30 (dt, J=
17.2, 1.5 Hz, 1H), 5.11 (dt, J=10.5, 1.3 Hz, 1H), 4.37–4.42 (m, 1H), 3.67
(dd, J=2.8, 11.4 Hz, 1H), 2.01 (dt, J=13.8, 2.8 Hz, 1H), 1.45 (ddd, J=
2.4, 11.4, 13.8 Hz, 1H), 1.23 (s, 3H), 1.16 (s, 3H), 0.83 (s, 9H), 0.07 ppm
(s, 6H); 13C NMR (100 MHz, CDCl3): d=139.33, 114.49, 77.92, 74.70,
71.84, 33.00, 27.24, 25.78, 24.11, 18.11, 7.77, ꢀ2.16 ppm; IR (thin film):
n˜ =2958, 2931, 2884, 2858, 1740, 1673, 1648, 1614, 1472, 1463, 1392, 1354,
1326, 1292, 1254, 1216, 1169, 1138, 1091, 1044, 1006, 989, 925, 903, 835,
812, 774, 708, 694, 680, 653 cmꢀ1; elemental analysis calcd (%) for: C
61.26, H 11.02; found: C 61.47 H 11.01.
Experimental Section
General procedure for carrying out the Ru-catalyzed alkyne–b,g-enone
coupling reaction: A point-shaped vial equipped with a stirring bar and
capped with a rubber septum was flame-dried and cooled under an argon
atmosphere. The septum was temporarily removed to allow the addition
of the alkyne (0.25 mmol), enone[22] (0.75 mmol), and acetone (0.5 mL).
After stirring at room temperature for 5 min to ensure the formation of a
homogeneous solution, the septum was again removed and [CpRu-
ACHTUNGTRENNUNG(CH3CN)3]ACHTUNGTRENNUNG[PF6] (11 mg, 0.025 mmol) was added quickly. The vial was
sealed and the septum was wrapped with black electric tape. The reaction
mixture was stirred at room temperature for 40 h and the solvent re-
moved under reduced pressure. The residue was purified by flash column
chromatography on silica gel to give the desired product.
Compound 49a: Rf =0.38 (6% EtOAc/petroleum ether); 1H NMR
(500 MHz, CDCl3): d=7.25 (d, J=8.6 Hz, 2H), 6.87 (d, J=8.8 Hz, 2H),
5.27 (s, 1H), 4.49 (dd, J=11.8, 15.5 Hz, 2H), 3.88–3.83 (m, 1H), 3.80 (s,
3H), 3.53–3.49 (m, 1H), 3.49–3.42 (m, 2H), 2.91 (dd, J=5.7, 16.7 Hz,
1H), 2.54 (dd, J=7.0, 17.0 Hz, 1H), 2.38 (dt, J=2.0, 13.5 Hz, 1H), 2.24
(dt, J=2.0, 13.0 Hz, 1H), 1.98–2.02 (m, 2H), 1.12 (s, 9H), 0.08 ppm (s,
9H); 13C NMR (125 MHz, CCl3): d=213.5, 159.1, 152.4, 130.3, 129.2,
123.9, 113.7, 77.4, 75.1, 72.9, 72.8, 55.2, 45.0, 44.2, 43.1, 36.4, 26.1,
0.2 ppm; IR (neat film): n˜ =1701, 1615, 1585, 1512, 1473, 1460, 1362,
Synthesis of compound 27: Diol 28 (119.3 mg, 0.435 mmol) was dissolved
in CH2Cl2 (3.3 mL, 0.1m) and triethylamine (0.3 mL, 2.15 mmol) was
then added. The reaction was cooled to 08C and then thionyl chloride
(0.75m in CH2Cl2, 1.2 mL, 0.9 mmol) was added. The reaction was stirred
for 10 min and then quenched with water, diluted with CH2Cl2, washed
with brine, dried (Na2SO4), filtered, and concentrated in vacuo to give
the crude diastereomeric cyclic sulfites, which were used without purifica-
tion. The crude cyclic sulfites were dissolved in THF (6.5 mL) and then
TBAF (1m in THF, 2.2 mL, 2.2 mmol) was added. The solution was
heated to reflux for 15 min, after which it was cooled to ambient temper-
ature, diluted with diethyl ether, washed with water and brine, dried
(Na2SO4), filtered, concentrated in vacuo, and chromatographed with a
solvent gradient (10–20% v/v ethyl acetate/petroleum ether) to give ep-
1245, 1099, 1035, 837 cmꢀ1
; HRMS calcd for C20H29O4Si: 361.1835
[MꢀC4H9]+; found: 361.1835.
Compound 50a: Rf =0.28 (6% EtOAc/petroleum ether); 1H NMR
(500 MHz, CDCl3): d=7.26 (d, J=8.7 Hz, 2H), 6.87 (d, J=8.7 Hz, 2H),
5.27 (s, 1H), 4.47 (dd, J=11.6, 13.9 Hz, 2H), 4.36–4.32 (m, 1H), 3.94–
3.89 (m, 1H), 3.80 (s, 3H), 3.52 (dd, J=5.9, 9.8 Hz, 1H), 3.44 (dd, J=5.5,
9.8 Hz, 1H), 2.72 (d, J=6.5 Hz, 2H), 2.43 (dd, J=3.3, 13.1 Hz, 1H), 2.38
(dd, J=4.4, 13.7 Hz, 1H), 2.21 (dd, J=6.5, 13.4 Hz, 1H), 2.02 (dd, J=6.2,
13.2 Hz, 1H), 1.12 (s, 9H), 0.09 ppm (s, 9H); 13C NMR (125 MHz, CCl3):
d=213.4, 159.2, 150.3, 129.4, 125.9, 113.7, 73.1, 72.1, 71.2, 69.6, 55.3, 44.3,
43.8, 40.2, 35.6, 26.2, 26.1, 0.2 ppm; IR (film): n˜ =1703, 1611, 1582, 1509,
oxide 27 (29.8 mg, 48% yield over 2 steps) as a dark orange oil. [a]D25
=
ꢀ3.2 (c=0.56 in CHCl3); 1H NMR (500 MHz, CDCl3): d=5.92 (ddd, J=
5.5, 10.5, 17.5 Hz, 1H), 5.27 (d, J=17.5 Hz, 1H), 5.12 (d, J=10.5 Hz,
1H), 4.34 (brq, J=5.5 Hz, 1H), 2.93 (dd, J=4.5 Hz, 7.5 Hz, 1H), 2.12
(brs, 1H), 1.81 (ddd, J=4.5, 7.5, 14.5 Hz, 1H), 1.67 (ddd, J=4.5, 7.5,
14.5 Hz, 1H), 1.29 (s, 3H), 1.25 ppm (s, 3H); 13C NMR (500 MHz,
CDCl3): d=140.5, 114.7, 70.9, 61.2, 58.3, 35.8, 24.7, 18.9 ppm; IR (thin
film): n˜ =3426, 3082, 2964, 2926, 1708, 1645, 1459, 1427, 1380, 1325, 1252,
1206, 1123, 1053, 993, 923, 898, 848, 784, 760, 679 cmꢀ1; LRMS calcd for
C8H15O2: 143.1; found: 143.1.
1461, 1364, 1301, 1242, 1170, 1097, 1029, 835 cmꢀ1; HRMS calcd for
+
ꢀ
C20H29O4Si: 361.1835 [M C4H9] ; found: 361.1833.
Synthesis of compound 28:
Catalyst preparation: S,S-ProPhenol ligand 29 (70.2 mg, 0.11 mmol) was
added to a flame-dried test tube containing a stirring bar. The ligand was
then azeotroped three times with benzene (1 mL each) and dissolved in
toluene (0.40 mL). Diethylzinc (1m in hexanes, 0.20 mL, 0.20 mmol) was
added slowly. Isopropanol (0.4 mL, 5.22 mmol) and acetic acid (0.1 mL,
1m in toluene, 0.01 mmol) were then added and the catalyst was allowed
to stir at ambient temperature for 30 min.
Aldol reaction: Aldehyde 8[23] (202.8 mg, 1.00 mmol) and methyl vinyl
ketone 9 (undistilled, 0.50 mL, 6.2 mmol) were added to a flame-dried
test tube equipped with molecular sieves (4 ꢁ; 203 mg) and a stirring bar.
The catalyst solution was then added and the reaction was stirred at 48C
for 7 h. The reaction mixture was then poured into phosphate buffer (1m,
pH 7) and extracted with petroleum ether/diethyl ether (1:1). The organic
layer was then washed twice with phosphate buffer (1m, pH 7) and once
with brine. The aqueous layers were extracted twice with diethyl ether.
Synthesis of compound 58:
Catalyst preparation: (R,R)-ProPhenol ligand 29 (35.9 mg, 0.056 mmol)
was added to a flame-dried test tube containing a stirring bar. The ligand
was azeotroped three times with benzene (1 mL each) and then dissolved
in toluene (0.20 mL), and diethylzinc was added slowly (1m in hexanes,
0.10 mL, 0.10 mmol). Isopropanol (0.2 mL, 2.6 mmol) was then added
and the catalyst solution was allowed to stir at ambient temperature for
30 min.
Aldol reaction: Aldehyde 47[24] (111.7 mg, 0.52 mmol) and methyl vinyl
ketone 9 (freshly distilled, 0.50 mL, 6.2 mmol) were added to a flame-
9786
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 9777 – 9788

Synthesis of Bryostatin Rings A and B
FULL PAPER
dried test tube equipped with molecular sieves (4 ꢁ; 110.7 mg) and a stir-
ring bar. The catalyst was then added and the reaction was stirred at 48C
for 7 h. The reaction mixture was then poured into phosphate buffer (1m,
pH 7) and extracted with petroleum ether/diethyl ether (1:1). The organic
layer was washed twice with phosphate buffer (1m, pH 7) and once with
brine. The aqueous layers were extracted twice with diethyl ether. The
combined organic layers were dried (Na2SO4), filtered, concentrated in
vacuo, and chromatographed with a solvent gradient (2–10% v/v diethyl
ether/petroleum ether) to give compound 58 (93.1 mg, 63% yield, 94%
1092, 1037, 991, 923, 840, 768, 717, 689 cmꢀ1
C33H50O6Si: 570.3377; found: 570.3388.
; HRMS calcd for
Synthesis of compound 67: Ketal 63 (16.1 mg, 0.028 mmol) was dissolved
in methanol (0.5 mL). Camphor sulfonic acid (0.8 mg, 0.003 mmol) was
added at ambient temperature and the reaction was stirred for 3 h. The
solution was then poured over aqueous NaHCO3, diluted with diethyl
ether, washed with brine, dried (MgSO4), filtered, concentrated in vacuo,
and isocratically chromatographed (25% v/v diethyl ether/petroleum
ether) to give compound 67 (10.2 mg, 70% yield) as a yellow oil. Rf =
0.17 (10% v/v ethyl acetate/petroleum ether) [a]2D5 =9.8 (c=0.98 in
CH2Cl2); 1H NMR (500 MHz, CDCl3): d=7.24 (d, J=8.5 Hz, 2H), 6.85
(d, J=8.7 Hz, 2H), 5.83 (ddd, J=17.3, 10.6, 5.1 Hz, 1H), 5.25 (dt, J=
17.2, 1.6 Hz, 1H), 5.21 (brs, 1H), 5.06 (dt, J=10.6, 1.5 Hz, 1H), 4.53 (d,
J=11.7 Hz, 1H), 4.49 (d, J=11.7 Hz, 1H), 3.98–4.02 (m, 2H), 3.79 (s,
3H), 3.59 (ddt, J=11.0, 2.2, 5.0 Hz, 1H), 3.53–3.40 (m, 3H), 3.15 (s, 3H),
2.35–2.42 (m, 2H), 2.09 (dd, J=15.9, 5.1 Hz, 1H), 2.01–2.13 (m, 2H),
1.90–1.98 (m, 2H), 1.78 (dd, J=16.0, 5.2 Hz, 1H), 1.74 (ddd, J=12.4, 4.5,
3.1 Hz, 1H), 1.03 (s, 3H), 0.95 (s, 3H), 0.07 ppm (s, 9H); 13C NMR
(125 MHz, CDCl3): d=153.7, 138.4, 130.6, 129.1, 123.0, 114.2, 113.7,
104.1, 77.5, 75.5, 73.0, 72.9, 71.3, 68.9, 55.3, 47.9, 47.0, 43.1, 39.2, 36.6,
35.6, 29.7, 20.6, 15.5, 0.3 ppm; IR (thin film): n˜ =3456, 2952, 2925, 2854,
1618, 1514, 1465, 1362, 1302, 1248, 1114, 1038, 843 cmꢀ1; HRMS calcd for
C28H43O6Si: 503.2829 [MꢀCH3]; found: 503.2829.
ee) as a light yellow oil. HPLC: AD column, 230 nm, 1.0 mLminꢀ1
,
99.5:0.5 heptane/isopropanol, tr (minor)=11.124 min, tr (major)=
12.918 min; Rf =0.65 (20% ethyl acetate/petroleum ether); [a]2D4 =ꢀ26.0
(c=2.69 in CH2Cl2,); 1H NMR (400 MHz, CDCl3): d=6.37 (dd, J=10.5,
17.5 Hz, 1H), 6.21 (dd, J=1.1, 17.5 Hz, 1H), 5.81 (dd, J=1.1, 10.5 Hz,
1H), 4.03 (dt, J=9.6, 2.9 Hz, 1H), 3.61 (d, J=3.2 Hz, 1H), 3.44–3.46 (m,
2H), 2.72 (dd, J=9.6, 16.0 Hz, 1H), 2.63 (dd, J=2.7, 15.9 Hz, 1H), 0.88
(s, 3H), 0.85 (s, 9H), 0.81 (s, 3H), 0.020 (s, 3H), 0.019 ppm (s, 3H);
13C NMR (100 MHz, CDCl3): d=201.0, 136.9, 128.5, 74.2, 71.9, 41.9, 38.4,
25.8, 21.8, 19.2, 18.2, ꢀ5.6, ꢀ5.7 ppm; IR (thin film) n˜ =3500, 2926, 2855,
1746, 1681, 1619, 1590, 1464, 1391, 1363, 1252, 1098, 1006, 984, 837, 776,
669 cmꢀ1
;
HRMS calcd for C15H31O3Si: 287.2042 [M+H]+; found:
287.2014.
Synthesis of compound 46: Tetramethylammonium triacetoxyborohy-
dride (2.8919 g, 10.99 mmol) was dissolved in acetic acid (5.5 mL) and
then cooled to ꢀ208C. Enone 58 (311.9 mg, 1.089 mmol) was then added
as a solution in acetone (5.5 mL) and the reaction was stirred for 17 h.
Sodium potassium tartrate (saturated aqueous solution, ꢁ100 mL) was
then added, as well as CH2Cl2 (ꢁ100 mL) and the thick mixture was
stirred for 2.5 h. The layers were then separated and the aqueous layer
was further extracted three times with CH2Cl2. The combined organic
layers were dried (MgSO4), filtered, concentrated in vacuo, and chroma-
Acknowledgements
We thank the National Institutes of Health (GM33049) and NSF for
their generous support of our programs. We also thank Amgen (HY),
Bristol–Myers Squibb (HY), NSF (CSB), and ARCS (CSB) for their gen-
erous financial support and fellowships. GD is a Stanford Graduate
Fellow. We appreciate the assistance of G. Wuitschik in the model study
of the bryostatin ring-B synthesis. Mass spectra were provided by the
Mass Spectrometry Regional Center of the University of California, San
Francisco, supported by the NIH Division of Research Resources. Palla-
dium and ruthenium salts were generously supplied by Johnson–Matthey.
tographed with
a solvent gradient (10–20% diethyl ether/petroleum
ether) to give diol 46 (215.6 mg, 69% yield, 5:1 d.r., as judged by integra-
tion of the olefin peak at 5.11 vs. 5.06 ppm for the diastereomer) as a
yellow oil. Rf =0.43 (20% v/v ethyl acetate/petroleum ether); [a]D24
=
ꢀ3.62 (c=2.98 in CH2Cl2); 1H NMR (400 MHz, CDCl3): d=5.93 (ddd,
J=5.0, 10.5, 17.2 Hz, 1H), 5.30 (dt, J=17.1, 1.7 Hz, 1H), 5.11 (dt, J=
10.5, 1.5 Hz, 1H), 4.45 (brs, 1H), 4.09 (d, J=2.9 Hz, 1H), 3.82 (dt, J=
12.0, 2.7 Hz, 1H), 3.50 (d, J=9.8 Hz, 1H), 3.46 (d, J=9.8 Hz, 1H), 3.12
(d, J=6.4 Hz, 1H), 1.72 (ddd, J=3.5, 11.0, 14.0 Hz, 1H), 1.56 (ddd, J=
1.4, 6.9, 14.0 Hz, 1H), 0.88 (s, 9H), 0.86 (s, 3H), 0.80 (s, 3H), 0.06 ppm (s,
6H); 13C NMR (100 MHz, CDCl3): d=141.1, 113.9, 76.3, 73.7, 70.5, 37.8,
37.1, 25.8, 22.3, 18.9, 18.1, ꢀ5.71, ꢀ5.69 ppm; IR (thin film): n˜ =3452,
2927, 2856, 1744, 1717, 1472, 1404, 1363, 1327, 1293, 1254, 1161, 1094,
1006, 990, 922, 837, 777, 669 cmꢀ1; elemental analysis calcd (%) for
C15H32O3Si: C 62.45, H 11.18; found: C 62.66, H 10.97.
[1] For bryostatin 1 isolation and structure determination, see: a) G. R.
Pettit, C. L. Herald, D. L. Doubek, D. L. Herald, E. Arnold, J.
Clardy, J. Am. Chem. Soc. 1982, 104, 6846; for recent reviews on
bryostatins, see: b) K. J. Hale, S. Manaviazar, Chem. Asian J. 2010,
5, 704; c) K. J. Hale, M. G. Hummersone, S. Manaviazar, M. Friger-
io, Nat. Prod. Rep. 2002, 19, 413; d) P. A. Wender, J. L. Baryza,
M. K. Hilinski, J. C. Horan, C. Kan, V. A. Verma in Drug Discovery
Research: New Frontiers in the Post-Genomic Era (Ed.: Z. Huang),
Wiley, New York, 2007, 127–162; e) R. Mutter, M. Wills, Bioorg.
Med. Chem. 2000, 8, 1841; for some of our efforts, see: B. M. Trost,
H. Yang, G. Dong, Chem. Eur. J. 2011, 17, 9789–9805; B. M. Trost,
H. Yang, O. R. Thiel, A. J. Frontier, C. S. Brindle, J. Am. Chem. Soc.
2007, 129, 2206–2207; for studies related to ring C construction, see:
B. M. Trost, A. J. Frontier, O. R. Thiel, H. Yang, G. Dong, Chem.
Eur. J. 2011, 17, 9762–9776.
[2] a) D. A. Evans, P. H. Carter, E. M. Carreira, A. B. Charrette, J. A.
Prunet, M. Lautens, J. Am. Chem. Soc. 1999, 121, 7540, and referen-
ces therein; for other examples of stereoselective formation of the
exocyclic enoate of ring B, including with chiral phosphonate re-
agents, see: b) M. OꢂBrien, N. H. Taylor, E. J. Thomas, Tetrahedron
Lett. 2002, 43, 5491; for cuprate additions to form enoates, see: c) A.
Baron, M. Ball, B. Bradshaw, S. Donnelly, O. Germay, P. C. Oller, N.
Kumar, N. Martin, M. OꢂBrien, H. Omori, C. Moore, E. J. Thomas,
Pure Appl. Chem. 2005, 77, 103; d) R. J. Maguire, S. P. Munt, E. J.
Thomas, J. Chem. Soc. Perkin Trans. 1 1998, 2853; for intramolecular
radical additions to enoates, see: e) S. P. Munt, E. J. Thomas, J.
Chem. Soc. Chem. Commun. 1989, 480; for sterically-controlled
phosphonate additions, see: f) T. F. J. Lampe, H. M. R. Hoffmann,
Tetrahedron Lett. 1996, 37, 7695; for tethered phosphonate reagents,
Synthesis of compound 63: Alkene 62 (134.7 mg, 0.484 mmol) and alkyne
10 (49.7 mg, 0.17 mmol)[6] were dissolved in acetone (0.4 mL) in a flame-
dried microwave vial. [CpRuACHTUNTRGENNUG(MeCN)3]ACHTUTGNREN[NGUN PF6] (7.7 mg, 0.018 mmol) was
then added and the reaction was stirred at ambient temperature for 22 h.
The reaction was then concentrated in vacuo and chromatographed with
a solvent gradient (5–30% v/v diethyl ether/petroleum ether) to give 63
(17.6 mg, 18% yield, 2.4:1 d.r.), recovered alkene 62 (39.5 mg, 31% re-
covery) and recovered alkyne 10 (2.8 mg, 7% recovery). Rf =0.55 (10%
v/v ethyl acetate/petroleum ether); [a]2D4 =ꢀ3.2 (c=1.24 in CH2Cl2);
1H NMR (500 MHz, CDCl3): d=7.23 (d, J=7.7 Hz, 2H), 6.85 (d, J=
8.7 Hz, 2H), 5.83 (ddd, J=17.3, 10.5, 5.9 Hz, 1H), 5.25 (brs, 1H), 5.19
(dt, J=17.3, 1.6 Hz, 1H), 5.10 (ddd, J=10.5, 3.3, 1.5 Hz, 1H), 4.49 (d, J=
11.8 Hz, 1H), 4.46 (d, J=11.8 Hz, 1H), 4.19–4.23 (m, 1H), 3.97 (dd, J=
9.5, 6.3 Hz, 1H), 3.78–3.85 (m, 1H), 3.78 (s, 3H), 3.45–3.50 (m, 1H), 3.45
(dd, J=10.1, 5.2 Hz, 1H), 3.42 (dd, J=10.0, 4.3 Hz, 1H), 2.94 (ddd, J=
17.5, 10.0, 5.2 Hz, 1H), 2.58 (dd, J=17.3, 7.3 Hz, 1H), 2.37 (brd, J=
13.4 Hz, 1H), 2.20–2.24 (m, 2H), 1.95–2.00 (m, 3H), 1.86–1.42 (m, 8H),
1.302 (s, 3H), 1.298 (s, 3H), 0.06 ppm (s, 9H); 13C NMR (125 MHz,
CDCl3): d=212.5, 159.1, 152.5, 138.4, 130.3, 129.2, 123.8, 115.2, 113.7,
100.5, 77.4, 75.0, 72.9, 72.8, 70.9, 69.8, 68.1, 55.3, 50.6, 45.6, 45.1, 36.5,
32.2, 30.3, 25.3, 24.2, 20.3, 18.9, 0.3 ppm; IR (thin film): n˜ =2954, 2895,
2854, 1706, 1618, 1587, 1514, 1466, 1364, 1331, 1302, 1248, 1224, 1173,
Chem. Eur. J. 2011, 17, 9777 – 9788
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9787

B. M. Trost et al.
see: g) D. A. Evans, E. M. Carreira, Tetrahedron Lett. 1990, 31,
4703; for the use of a symmetrical substrate to ensure one geometri-
cal outcome, see: h) K. J. Hale, M. G. Hummersone, G. S. Bhatia,
Org. Lett. 2000, 2, 2189.
[11] P. W. D. Mitchell, Org. Prep. Proced. 1990, 22, 534.
[12] M. A. Umbreit, K. B. Sharpless, Org. Synth. 1981, 60, 29.
[13] K. Yamada, S. Goto, H. Nagase, Y. Kyotani, Y. Hirata, J. Org.
Chem. 1978, 43, 2076.
[3] a) M. A. Blanchette, M. S. Malamas, M. H. Nantz, J. C. Roberts, P.
Somfai, D. C. Whritenour, S. Masamune, M. Kageyama, T. Tamura,
J. Org. Chem. 1989, 54, 2817; b) M. Kageyama, T. Tamura, M. H.
Nantz, J. C. Roberts, P. Somfai, D. C. Whritenour, S. Masamune, J.
Am. Chem. Soc. 1990, 112, 7407; c) K. Ohmori, Y. Ogawa, T.
Obitsu, Y. Ishikawa, S. Nishiyama, S. Yamamura, Angew. Chem.
2000, 112, 2376; Angew. Chem. Int. Ed. 2000, 39, 2290, and referen-
ces therein; for a review of this synthesis by Professor K. Ohmori
and the Keio team, see: d) K. Ohmori, Bull. Chem. Soc. Jpn. 2004,
77, 875–885; for a formal synthesis, see: e) S. Manaviazar, M. Friger-
io, G. S. Bhatia, M. G. Hummersone, A. E. Aliev, K. J. Hale, Org.
Lett. 2006, 8, 4477, and references therein.
[14] S. S. P. Chou, H. L. Kuo, C. J. Wang, C. Y. Tsai, C. M. Sun, J. Org.
Chem. 1989, 54, 868.
[15] For a preliminary report of a portion of our work, see: B. M. Trost,
H. Yang, G. Wuitschik, Org. Lett. 2005, 7, 4761.
[16] B. M. Trost, A. Fettes, B. T. Shireman, J. Am. Chem. Soc. 2004, 126,
2660.
[17] H. Zhao, B. L. Pagenkopf, Chem. Commun. 2003, 2592: in the pres-
ence of BF3·OEt2, the lithium trimethylaluminum ate complex
added to the trisubstituted epoxide at the more hindered position in
a stereo- and regiospecific manner to give a single stereoisomer.
[4] B. M. Trost, S. Shin, J. A. Sclafani, J. Am. Chem. Soc. 2005, 127,
8602.
[5] For other notable synthetic approaches to rings A and B, see: a) Y.
Lu, M. J. Krische, Org. Lett. 2009, 11, 3108; b) G. E. Keck, M. B.
Kraft, A. P. Truong. W. Li, C. C. Sanchez, N. Kedei. N. E. Lewin,
P. M. Blumberg, J. Am. Chem. Soc. 2008, 130, 6660; c) P. A. Wender,
B. A. DeChristopher, A. J. Schrier, J. Am. Chem. Soc. 2008, 130,
6658; d) M. Ball, B. J. Bradshaw, R. Dumeunier, T. J. Gregson, S.
MacCormick, H. Omuri, E. J. Thomas, Tetrahedron Lett. 2006, 47,
2223; e) E. A. Voight, H. Seradj, P. A. Roethles, S. D. Burke, Org.
Lett. 2004, 6, 4045; f) A. Vakalopoulos, T. F. J. Lampe, H. M. R.
Hoffmann, Org. Lett. 2001, 3, 929; g) S. Kiyooka, H. Maeda, Tetra-
hedron: Asymmetry 1997, 8, 3371.
[6] B. M. Trost, M. R. Machacek, B. D. Faulk, J. Am. Chem. Soc. 2006,
128, 6745.
[7] For a procedure that has combined both the Ru and Pd catalysis in
one pot, see: B. M. Trost, M. R. Machacek, Angew. Chem. 2002, 114,
4887; Angew. Chem. Int. Ed. 2002, 41, 4693.
[8] This is in striking contrast to what has been accomplished in our my-
calamide A synthesis wherein the a,b-unsaturated double bond was
selectively dihydroxylated (OsO4, acetone/H2O, 5 8C) in the pres-
ence of a similarly disubstituted vinylsilane in high yield: B. M.
Trost, H. Yang, G. D. Probst, J. Am. Chem. Soc. 2004, 126, 48.
[9] G. J. Roth, B. Liepold, S. G. Mꢃller, H. J. Bestmann, Synthesis 2004,
59.
[18] Made in analogy to the TBS epoxide in: Y. F. Fang, A. C. Oehlschl-
ager, N. H. Gerogopapadakou, P. G. Hartman, P. Scheliga, J. Am.
Chem. Soc. 1995, 117, 6727.
[19] C. Le Roux, J. Dubac, Organometallics 1996, 15, 4646.
[20] The use of the conditions employing acetic acid, which were devel-
oped for aldehyde 8, resulted in a lower yield (40%) of aldol adduct
58.
[21] D. A. Evans, K. T. Chapman, E. M. Carreira, J. Am. Chem. Soc.
1988, 110, 3560.
[22] B. M. Trost, G. Dong, Nature 2008, 456, 485.
[23] S. E. Denmark, R. A. Stavenger, J. Am. Chem. Soc. 2000, 122, 8837.
[24] S. Sano, K. Mori, Eur. J. Org. Chem. 1999, 1679.
[25] E. Napolitano, R. Fiaschi, E. Mastrorilli, Synthesis 1986, 122.
Received: October 11, 2010
Revised: April 21, 2011
[10] M. G. Martin, B. Ganem, Tetrahedron Lett. 1984, 25, 251.
Published online: July 19, 2011
9788
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 9777 – 9788