Total synthesis of bafilomycin V1: A methanolysis product of the macrolide bafilomycin C2

Article abstract of DOI:10.1021/jo015864x

A synthesis of bafilomycin V₁, a methanolysis product of the macrolide natural product bafilomycin C₂; is described. The route utilizes chiral nonracemic allenylzinc reagents, prepared in situ from propargylic mesylates, to access key segments of this methyl ester. The acetylenic moieties of the derived homopropargylic alcohol adducts play an important role in further elaboration of these subunits. Final assemblage of the 25-carbon chain, containing 12 stereocenters, an α-methoxy Z,E 1,3-dienic ester, and an additional E,E 1,3-diene, was achieved through Stille coupling of an acetylene-derived vinyl stannane and vinyl iodide of approximate equal complexity. Attempted cyclization of several C15 hydroxy acid derivatives to the 16-membered lactone bafilomycin A₁, a potent inhibitor of vacuolar ATPases, could not be achieved.

Full text of DOI:10.1021/jo015864x

J . Org. Chem. 2002, 67, 733-740
733
Tota l Syn th esis of Ba filom ycin V1: A Meth a n olysis P r od u ct of th e
Ma cr olid e Ba filom ycin C2
J ames A. Marshall* and Nicholas D. Adams
Department of Chemistry, University of Virginia, McCormick Road, P.O. Box 400319,
Charlottesville, Virginia 22904
jam5x@virginia.edu
Received J une 25, 2001
A synthesis of bafilomycin V1, a methanolysis product of the macrolide natural product bafilomycin
C2, is described. The route utilizes chiral nonracemic allenylzinc reagents, prepared in situ from
propargylic mesylates, to access key segments of this methyl ester. The acetylenic moieties of the
derived homopropargylic alcohol adducts play an important role in further elaboration of these
subunits. Final assemblage of the 25-carbon chain, containing 12 stereocenters, an R-methoxy Z,E
1,3-dienic ester, and an additional E,E 1,3-diene, was achieved through Stille coupling of an
acetylene-derived vinyl stannane and vinyl iodide of approximate equal complexity. Attempted
cyclization of several C15 hydroxy acid derivatives to the 16-membered lactone bafilomycin A1, a
potent inhibitor of vacuolar ATPases, could not be achieved.
Bafilomycin A1, a polyketide isolated from the fermen-
tation broth of Streptomyces griseus by Werner et al. in
1984,1 is a member of a macrolide family, which includes
the bafilomycins (Figure 1), the concanamycins,2 and the
hygrolidins.3 Structural features of the bafilomycins
include a 16-membered macrolide with a tetraene core,
an R-methoxy Z,E dienoate, 12 stereocenters, and a
â-hydroxy hemiacetal appendage.
The bafilomycins differ in substitution on the C21-
oxygen. Accordingly, bafilomycin A1 is unsubstituted
while bafilomycin B1 contains a flavensomycinic ester and
bafilomycin C1 has a fumaric ester appendage (Figure
1). Bafilomycins A2, B2, and C2 possess a methoxy
substituent at C19. The stereochemistry of bafilomycin
A1 was initially proposed by Corey and Ponder4 on the
basis of molecular modeling and NMR data and was later
confirmed by X-ray crystallography.5 The X-ray crystal
structure revealed an interesting hydrogen-bonding net-
work between the C17-hydroxyl and both the lactone
carbonyl and the C19-hydroxyl of the hemiacetal. The
solution conformation has been shown to be indistin-
guishable from that in the solid state.5b
The bafilomycins are structurally related to the con-
canamycins and hygrolidins, differing mainly in the C22-
substituent. Accordingly, the bafilomycins are substituted
with a C22-isopropyl group, the hygrolidins are substi-
tuted with an ethyl group, and the concanamycins
contain a 2-propenyl appendage. The concanamycins are
18-membered lactones incorporating a “biosynthetic”
F igu r e 1. Structures of the bafilomycins.
butyrate moiety at C8-C9 of the carbon backbone. All
of these macrolides share the hydrogen bonding motif.
Biologica l Activity. Bafilomycin A1 is a potent and
specific inhibitor of vacuolar ATPases (V-ATPases) in
vitro and in vivo.6 Given that V-ATPases are known to
participate in bone resorption (i.e, the loss of bone),6a
inhibitors of such enzymes may serve as potential treat-
ments for osteoporosis. Structure-activity relationships
(SAR) of bafilomycin A1 have shown that derivatization
of the C21-hydroxyl through acylation or alkylation has
no effect, while the C7-hydroxyl is essential in preserving
activity.6b The tetraene core of the lactone ring is also
essential for activity as both partial and complete hy-
drogenation result in marked loss of activity. Several
functional group modifications on the hemiacetal ring did
not have a major effect on activity, but replacing the ring
with an acyclic side-chain resulted in complete loss of
activity. Several degradation products of bafilomycin A1
(1) Werner, G.; Hagenmaier, H.; Drautz, H.; Baumgartener, A.;
Za¨hner, H. J . Antibiot. 1984, 37, 110.
(2) (a) Kinashi, H.; Someno, K.; Higashijima, T.; Miyazawa Tetra-
hedron Lett. 1981, 22, 3857. (b) Westley, J . W.; Liu, C.-M.; Sello, L.
H.; Evans, R. H.; Troup, N.; Blount, J . F.; Chiu, A. M.; Todaro, L. J .;
Miller, P. A. J . Antibiot. 1984, 37, 1738.
(3) Seto, H.; Tajima, I.; Akao, H.; Furihata, K.; Otake, N. J . Antibiot.
1984, 37, 610.
(4) Corey, E. J .; Ponder S. W. Tetrahedron Lett. 1984, 25, 4325.
(5) (a) Baker, G. H.; Brown, P. J .; Dorgan, R. J . J .; Everett, J . R.;
Ley, S. V.; Slawin, A. M. Z.; Williams, D. J . Tetrahedron Lett. 1987,
28, 5565. (b) Baker, G. H.; Brown, P. J .; Dorgan, R. J . J .; Everett, J .
R. J . Chem Soc., Perkin Trans. 2 1989, 1073.
(6) (a) Sundquist, K. T.; Marks, S. C. J . Bone Miner. Res. 1994, 9,
1575. (b) Gagliardi, S.; Gatti, P. A.; Belfiore, P.; Zocchetti, A.; Clarke,
G. D.; Farina, C. J . Med. Chem. 1998, 41, 1883.
10.1021/jo015864x CCC: $22.00 © 2002 American Chemical Society
Published on Web 01/04/2002

734 J . Org. Chem., Vol. 67, No. 3, 2002
Marshall and Adams
Syn th esis of th e C1-C11 Su bu n it. Our synthesis
of the C1-C11 segment began with formation of the
elusive anti,anti stereotriad. It was desirable to adopt
an orthogonal protecting group strategy to permit an
eventual selective deprotection. For this, we envisioned
an addition of the allenylzinc reagent derived from
mesylate (R)-1 to (R)-R-methyl-â-OPMB (PMB ) p-
methoxybenzyl) propanal (2),11 an aldehyde that had yet
to be examined in allenylzinc stereotriad synthesis (eq
1).12 The PMB protecting group seemed appropriate
because it can be selectively cleaved in the presence of
several other protecting groups such as silyl ethers and
esters. Surprisingly, the addition was extremely sluggish
and additional catalyst and warming to 0 °C were
necessary for product formation. A disappointing 79:19
mixture of anti,anti and anti,syn isomers 3 was formed
as an inseparable mixture. The selectivity of this addition
is considerably lower than that involving the previously
examined â-ODPS (DPS ) diphenyl-tert-butylsilyl) al-
dehyde.12 The anti,syn isomer derives from partial race-
mization of the allenylzinc intermediate, a consequence
of the unreactivity of aldehyde 2, which necessitates
higher temperatures and additional catalyst for adduct
formation.
F igu r e 2. Synthetic plan for bafilomycin A1.
have also been prepared; however, their SAR has not
been reported to our knowledge.7
Syn th esis. To date, three total syntheses of bafilomy-
cin A1 have been reported. The first, by Evans and Calter
in 1993, took advantage of the versatility of the aldol
reaction to introduce stereocenters of subunits and a
Stille coupling to join key fragments.8 The second, by
Toshima and co-workers, employed Stille and aldol
reactions, analogous to those of Evans and Calter, as
major coupling events.9 They assembled the subunits in
a linear fashion from chiral pool materials. The third and
most recent total synthesis of bafilomycin A1 was achieved
by Roush and co-workers in 1999.10 They effected a
stereoselective aldol condensation for fragment coupling
and diastereoselective aldehyde crotylboration methodol-
ogy to set subunit stereocenters.
Formation of a significant amount of the anti,syn
diastereomer with the â-OPMB aldehyde prompted us
to substitute an alternative protecting group in an effort
to increase the selectivity. Accordingly, we examined the
R-methyl-â-OTBS (TBS ) tert-butyldimethylsilyl) alde-
hyde (R)-413 (eq 2). Addition of the allenylzinc reagent
derived from mesylate (R)-1 to aldehyde (R)-4 proceeded
at -20 °C to give the anti,anti triad 5 in 70% yield along
with a small amount of the anti,syn isomer, which was
separated by silica gel chromatography.
Syn th etic Ap p r oa ch . Our synthetic approach, il-
lustrated in Figure 2, utilizes stereoselective additions
of chiral nonracemic allenylzinc reagents to aldehydes
F , G, and I for assembly of the anti,anti stereotriad D
and the anti subunits C and F . The acetylenic moiety
plays a key role in elaborations to segments B and C as
well as the joining of these segments, or their equivalents,
through Stille coupling of iodide B with a C12 vinylstan-
nane derived from an analogue of C.
(7) Deeg, M.; Hagenmaier, H.; Kretschmer, A. J . Antibiot. 1987, 40,
320.
The resulting hindered alcohol was protected as the
TBS ether 6 with TBSOTf in the presence of 2,6-lutidine
in near quantitative yield (Scheme 1). At this stage, we
wanted to perform a one carbon homologation of the
terminal alkyne. For this we employed a two-step hy-
(8) (a) Evans, D. A.; Calter, M. A. Tetrahedron Lett. 1993, 34, 6871.
(b) Only the late stage aldol and subsequent conversion to the natural
product have been reported (ref 8a). The details of the remainder of
the synthesis were taken from the Ph.D. Thesis of Michael A. Calter,
Harvard University, Cambridge, Massachusetts, 1993.
(9) Toshima, K.; J yojima, T.; Yamaguchi, H.; Noguchi, Y.; Yoshida,
T.; Murase, H.; Nakada, M.; Matsumura, S. J . Org. Chem. 1997, 62,
3271 and references therein.
(10) Total synthesis: (a) Scheidt, K.; Tasaka, A.; Bannister, T. D.;
Wendt, M. D.; Roush, W. R. Angew. Chem., Int. Ed. 1999, 38, 1652.
Subunit synthesis: (b) Roush, W. R.; Bannister, T. D. Tetrahedron Lett.
1992, 33, 3587. (c) Roush, W. R.; Bannister, T. D.; Wendt, M. D.
Tetrahedron Lett. 1993, 34, 8387.
(11) Smith, A. B., III; Kaufman, M. D.; Beauchamp, T. J .; LaMarche,
M. J .; Arimoto, H. Org. Lett. 1998, 1, 1823.
(12) Marshall, J . A.; Adams, N. D. J . Org. Chem. 1999, 64, 5201.
(13) Roush, W. R.; Palkowitz, A. D.; Palmer, M. A. J . J . Org. Chem.
1987, 52, 316.

Synthesis of Bafilomycin V1
Sch em e 1a
J . Org. Chem., Vol. 67, No. 3, 2002 735
Sch em e 2a
a
Key: (a) (c-C6H11)2BH, DME, 0 °C to room temperature, and
then H2O2, NaOH; (b) TBSOTf, 2,6-lutidine, CH2Cl2, 0 °C (96%);
(c) Dess-Martin periodinane, CH2Cl2, room temperature (95%);
(d) (EtO)2P(O)CHN2, KOt-Bu, THF, -78 to 0 °C (96%); (e)
Cp2ZrCl2, AlMe3, DCE, 60 °C, and then I2, THF, -30 to 0 °C (37-
48%).
droboration-homologation sequence. Several reagents
and conditions were examined, including borane and
catecholborane, and it was found that dicyclohexylborane
was the most effective. Accordingly, treatment of alkyne
6 with freshly prepared (c-C6H11)2BH in 1,2-dimethoxy-
ethane from 0 °C to room temperature followed by
treatment with basic H2O2 provided aldehyde 7 in 81%
yield along with 12% of alcohol 8.14 Formation of the
alcohol was somewhat surprising and is presumably a
result of reduction of the aldehyde by a borohydride
species (formally R2B(H)OH-) that may be formed during
the oxidation step.14b Alcohol 8 could be recycled through
oxidation with the Dess-Martin periodinane reagent15
to give aldehyde 7 in near quantitative yield. Attempts
to convert this aldehyde to alkyne 9 by utilizing either
the Corey-Fuchs16 protocol or Savignac’s dichlorophos-
phonate reagent17 were unsuccessful due to extensive
decomposition and cleavage of the TBS ethers. However,
homologation occurred cleanly with the Seyferth-Gilbert
diazophosphonate reagent18 to provide alkyne 9 in 96%
yield.
The primary TBS ether was then removed chemose-
lectively with catalytic PPTS in MeOH to give alcohol
10 in 79% yield (95% based on recovered starting mate-
rial). With alcohol 10 in hand, we turned our attention
to the formation of the vinyl iodide, which we envisioned
obtaining by Negishi carbozirconation.19 Accordingly,
alkyne 10 was treated with Cp2ZrCl2 and AlMe3 in
dichloroethane at 60 °C and the resultant vinylalane was
quenched by addition of an I2 solution in THF at -30
°C. Unfortunately, conversion to vinyl iodide 11 was
a
Key: (a) (COCl)2, DMSO, TEA, CH2Cl2, -78 °C (91%); (b)
DIBAL-H, CH2Cl2, -78 °C (92%); (c) MnO2, CH2Cl2, room tem-
perature (95%); (d) TBAF, THF, room temperature (81%).
inefficient (37-48% yield) despite several attempts at
optimization. Given the low yield of the Negishi carbozir-
conation/iodination of alkyne 10 and the potential insta-
bility of the vinyl iodide product, we decided to further
functionalize alcohol 10 and carry out the vinyl iodide
transformation at a later stage.
Oxidation of alcohol 10 under Swern conditions20 and
treatment of the resulting aldehyde 12 with the tri-
phenylphosphorylidene propionate Wittig reagent in
toluene provided ester 13 in high yield (Scheme 2). It is
noteworthy that a high temperature (100-110 °C) was
necessary for complete reaction within a reasonable time.
Reactions were incomplete in refluxing CH2Cl2 or 1,2-
dichloroethane, presumably due to the steric environment
at the R-position of the aldehyde. A small amount of the
Z isomer (∼5%) was formed as evidenced by the appear-
ance of a vinylic proton doublet at 6.7 ppm (vs 6.9 ppm
1
for 13) in the H NMR spectrum of the conjugated ester.
Ester reduction with DIBAL-H gave the requisite allylic
alcohol 14 in high overall yield. Stereoselective installa-
tion of the C10-C11 trisubstituted olefin was accom-
plished under Negishi’s carbozirconation conditions, which
provided vinyl iodide 15 in 65% yield. We later found that
the yield of the Negishi carbozirconation could be in-
creased to 71% by the addition of water to the reaction
mixture, a modification developed by Wipf and Lim.21
Allylic alcohol 15 was oxidized with MnO2 in CH2Cl2.
The resulting aldehyde 16 was treated with an R-OMe
phosphonate reagent22 in the presence of KHMDS and a
catalytic amount of 18-crown-6 in THF from 0 °C to room
temperature, which afforded a 94:6 2Z/2E mixture of
(14) (a) Hoffmann, R. W.; Dresely, S. Synthesis 1988, 103. (b) Cf.:
Marshall, J . A.; J ohns, B. A. J . Org. Chem. 2000, 65, 1501.
(15) (a) Dess, B. D.; Martin, J . C. J . Org. Chem. 1983, 48, 4155. (b)
The Dess-Martin reagent in this study was prepared by a modified
procedure using oxone; Cf.: Frigerio, M.; Santagostino, M.; Sputore,
S. J . Org. Chem. 1999, 64, 4537.
(16) Corey, E. J .; Fuchs, P. L. Tetrahedron Lett. 1972, 12, 3769.
(17) Marinetti, A.; Savignac, P. Org. Syn. 1974, 108.
(18) (a) Seyferth, D.; Marmor, R. S.; Hilbert, P. J . Org. Chem. 1971,
36, 1379. (b) Gilbert, J . C.; Weerasooriya, U. J . Org. Chem. 1979, 44,
4997. (c) Cf.: Kabat, M. M.; Kiegiel, J .; Cohen, N.; Toth, K.; Wovkulich,
P. M.; Uskokovi′c, M. R. J . Org. Chem. 1996, 61, 118.
(19) Negishi, E.-i.; Van Horn, D. E.; Yoshida, T. J . Am. Chem. Soc.
1985, 107, 6639.
(20) Mancuso, A. J .; Huang, S.-L.; Swern, D. J . Org. Chem. 1978,
43, 2480.
(21) Wipf, P.; Lim, S. Angew. Chem., Int. Ed. Engl. 1993, 32, 1068.
(22) Paterson, I.; McLeod, M. D. Tetrahedron Lett. 1997, 38, 4183.

736 J . Org. Chem., Vol. 67, No. 3, 2002
Marshall and Adams
isomers. These esters could be distinguished by signals
at 5.8 (major) and 5.6 ppm (minor) for the â-vinylic proton
1
in the H NMR spectrum. The hindered TBS ether was
removed with TBAF to complete the synthesis of the C1-
C11 subunit 18.
Syn th esis of th e C15-C25 Su bu n it 36. We have
previously described the assembly of this segment, so a
detailed discussion is not required.23 However, it is worth
pointing out that the stereochemically defining steps
involving addition of the allenylzinc reagent derived from
mesylate 1924 to isobutyraldehyde (20), Sharpless epoxi-
dation25 of the derived allylic alcohols 22 and 31, as well
as the methylcuprate addition leading to alcohol 33 all
proceed with excellent diastereoselectivity. The one
exception, addition of the allenylzinc reagent derived
from mesylate 19 to the unbranched aldehyde 27, is of
no consequence as the alcohol product is converted to
ketone 29, the precursor of glycoside 30. The structure
of triol 33 was confirmed through single-crystal X-ray
structure analysis.23
Unable to effect a kinetic resolution, we decided to
pursue an alternative strategy. Some years ago, we
reported a method for the synthesis of differentially
protected anti-1,2-diols by treatment of enantioenriched
R-oxygenated allylic stannanes with InCl3 to form an
intermediate allylic indium reagent through an anti SE2′
transmetalation process (eq 4).31
With aldehyde 36 in hand, we turned our attention to
the anti-1,2-diol array. A direct approach would simul-
taneously install the differentially protected anti-1,2-diol
and the alkyne, the latter of which could be hydrometa-
lated prior to coupling to the vinyl iodide partner. At the
time, there were no methods for the formation of enan-
tioenriched methoxy-substituted allenylmetal reagents.26
Racemic or achiral reagents of this type can be formed
by deprotonation of propargylic ethers with s- or t-BuLi
and subsequent transmetalation with ZnBr2.26,27 Follow-
ing this procedure, we prepared the allenylzinc reagent
from TMS-propargyl methyl ether,28 which was allowed
to react with a model aldehyde, 37,12 affording a 47:47:6
inseparable mixture of diastereomers in high yield (eq
3). As these additions are known to occur through a cyclic
transition state, the major isomers are presumed to be
the anti adducts and the minor is presumed to be a syn
isomer. The addition of external chiral additives such as
(-)-sparteine29 or (-)-N-methylephedrine30 resulted in
lower yields and exerted no beneficial effect on the
selectivity.
These reagents undergo highly anti-selective additions
to aldehydes, presumably through a six-membered chair
transition state. In principle, the MeO substituent could
be installed prior to the addition reaction through the
use of a chiral R-OMe allylic stannane (eq 4, R1 ) Me).
In an attempt to access this unknown reagent, hydroxy
stannane 39 was prepared as previously reported by
addition of Bu3SnLi to crotonaldehyde. Unfortunately, all
attempts to methylate the derived alkoxide with MeI
resulted in extensive decomposition of the hydroxy stan-
nane. Typical etherifications of R-hydroxy stannanes
utilize MOM or BOM chloride in the presence of an amine
base (i.e., i-Pr2NEt or Et3N). These reactions are believed
to proceed through an SN1 pathway. In an attempt to
prevent decomposition of the hydroxy stannane, we
employed MeOTf in the presence of pyridine-type bases
for this methylation. However, none of the expected
methyl ether 41 was detected. Instead, an isomerized
product whose 1H NMR spectrum is consistent with
aldehyde 40 was produced as the sole product. Interest-
ingly, this isomerization could also be effected with 10
mol % MeOTf or Yb(OTf)2.31
(23) Marshall, J . A.; Adams, N. D. Org. Lett. 2000, 2, 2897.
(24) Cf.: Marshall, J . A.; Yanik, M. M. J . Org. Chem. 2001, 66, 1373.
(25) (a) Katsuki, T.; Sharpless, K. B. J . Am. Chem. Soc. 1980, 102,
5974. For the catalytic asymmetric epoxidation, see: (b) Hanson, R.
M.; Sharpless, K. B. J . Org. Chem. 1986, 51, 1922.
(26) Recently, Poisson and Normant have reported a kinetic resolu-
tion of racemic allenylzinc bromides through selective reaction with
0.5 equiv of the N-benzyl imine of lactic or mandelic aldehyde to con-
sume the “matched” allenylzinc reagent as the homopropargylic amine
adduct. Subsequent addition of pivalic aldehyde to the reaction mixture
afforded the homopropargylic alcohol adduct of the less reactive
(“mismatched”) allenylzinc enantiomer: Poisson, J .-F.; Normant, J .-
F. J . Am. Chem. Soc. 2001, 123, 4639. These workers have also shown
that racemic MOMO-substituted allenylzinc reagents, which can be
formed from the TMS derivative of propargyl methoxymethoxy ether
by treatment with s-butyllithium at -40 °C followed by addition of
ZnBr2 at 0 °C, react with the same imines to afford homopropargyl-
amine adducts of the matched allenylzinc pairing. However, they have
not determined if the unreactive residual mismatched allenylzinc
reagent will afford adducts of aldehydes with useful diastereoselec-
tivity. The addition of the racemates of such reagents to a silyl
derivative of mandelic aldehyde proceeds with low diastereoselectiv-
ity: Poisson, J .-F.; Normant, J .-F. Org. Lett. 2001, 3, 1889.
(27) Poisson, J .-F.; Normant, J .-F. J . Org. Chem. 2000, 65, 6553.
(28) Labaudiniere, L.; Hanaizi, J .; Normant, J .-F. J . Org. Chem.
1992, 57, 6903.
With the synthesis of an R-OMe allylic stannane
looking unfeasible, we decided to pursue an alternative
method for installation of the anti-1,2-diol. The use of
an R-OMOM or R-BOM stannane was considered. How-
ever, we foresaw protecting group orthogonality prob-
lems. Given the prior results of Roush10 and Toshima9
with Takai’s in situ generated γ-methoxyallylchromium
reagent33 we decided to follow this approach. Accordingly,
(31) (a) Marshall, J . A.; Hinkle, K. W. J . Org. Chem. 1996, 61, 105.
(b) For a recent review on additions of allylic indium reagents to
aldehydes, see: Marshall, J . A. Chemtracts-Organic Chemistry 1997,
481.
(32) For a recent mechanistic study on the 1,3-isomerization of allylic
stannanes promoted by BF3‚OEt2, see: Marshall, J . A.; Gill, K. J .
Organomet. Chem. 2001, 624, 294.
(29) Kerrick, S. T.; Beak, P. J . Am. Chem. Soc. 1991, 113, 9708.
(30) For recent reports utilizing N-methylephedrine in enantiose-
lective additions of zinc acetylides to aldehydes, see: (a) Frantz, D.
E.; Fa¨ssler, R.; Carriera, E. M. J . Am. Chem Soc. 2000, 122, 1806. (b)
Boyall, D.; Lo´pez, F.; Frantz, D. E.; Carriera, E. M. Org. Lett. 2000, 2,
4233.

Synthesis of Bafilomycin V1
J . Org. Chem., Vol. 67, No. 3, 2002 737
Sch em e 3a
a
Key: (a) LiAlH4, THF (87%); (b) (+)-DIPT, TBHP, TIP (86%); (c) Red-Al, THF (93%); (d) TESOTf, 2,6-lutidine (99%); (e) H2O, HOAc,
THF (95%); (f) Swern (99%); (g) Dess-Martin; (h) PPTS, MeOH, room temperature, 1 h (84%, 2 steps).
Sch em e 4a
a
Key: (a) L-(+)-DIPT, TBHP, TIP, 4 Å MS, CH2Cl2, -20 °C, (80%, 2 steps); (b)TESOTf, 2,6-lutidine, CH2Cl2, -78 °C, (92%); (c) DIBAl-
H, CH2Cl2, -78 °C (92%); (d) Dess-Martin, NaHCO3, CH2Cl2, (quant).
the model aldehyde 37 was treated with CrCl2 and
acrolein dimethyl acetal in the presence of TMSI at -42
°C in THF to provide alcohol 42 in 76% yield along with
two minor diastereomers with an overall selectivity of
87:8:4 (eq 6). The absolute configuration of the carbinyl
center of the major isomer was confirmed by conversion
to the (S)- and (R)-mandelic esters (MPA).34
Before applying this methodology to our aldehyde
intermediate 36, we decided to test the proposed conver-
sion of the allylic ether adduct to the vinylstannane and
carry through to the Stille coupling of this model system.
To that end, hydroxylation of alkene 42 with a catalytic
amount of OsO4 in the presence of NMO and cleavage of
the resultant 1,2-diol with NaIO4 provided aldehyde 44
in 83% yield for the two steps (Scheme 5). Oxidative
cleavage of alkene 42 with O3 and ozonide reduction with
PPh3 or Me2S was less efficient. Conversion of aldehyde
44 to the terminal alkyne was best accomplished with
the Ohira diazophosphonate reagent35 in the presence of
K2CO3 in MeOH. Cleavage of the TES ether was mini-
mized to 10% by employing 1.5 equiv of both the phos-
phonate reagent and K2CO3 as compared to Ohira’s
original conditions (2.0 equiv) which resulted in ca. 20%
of the desilated adduct. Use of the Seyferth-Gilbert
reagent resulted in the formation of several unidentifi-
able side products. Palladium-catalyzed hydrostannation
(34) Trost, B. M.; Belletire, J . L.; Godleski, S.; McDougal, P. G.;
Balkovec, J . M.; Baldwin, J . J .; Christy, M. E.; Ponticello, G. S.; Varga,
S. L.; Springer, J . D. J . Org. Chem. 1986, 51, 2370.
(33) Takai, K.; Nitta, K.; Utimoto, K. Tetrahedron Lett. 1988, 29,
5263.
(35) Ohira, S. Synth. Commun. 1989, 19, 561.

738 J . Org. Chem., Vol. 67, No. 3, 2002
Marshall and Adams
Sch em e 5a
Sch em e 6a
a
Key: (a) OsO4 (5 mol %), NMO, THF/pH 7 buffer (4:1) (86%);
(b) Bu3SnH, Pd(PPh3)2Cl2, CH2Cl2 (73%).
of alkyne 45 provided vinyl stannane 46 with >90:10 E/ Z
selectivity.36
a
Key: (a) Dess-Martin, NaHCO3, CH2Cl2,(99%, crude); (b)
OsO4 (5 mol %), NMO, THF, pH 7 buffer (4:1) (85%); (c) NaIO4,
THF/H2O (1:1) (99%).
For the Stille coupling, we initially tried Liebeskind’s
copper(I) thiophene catalyst.37 However, this method led
only to recovered starting materials after several hours
at room temperature (eq 7). The use of Pd2dba3 in the
presence of AsPh3 in N-methylpyrrolidinone (NMP) at 55
°C produced only a small amount of product.38,39 Some
homocoupling of the iodide was observed, which could be
attributed to the sluggishness of the cross coupling
reaction. We reasoned that if the transmetalation process
could be accelerated, the cross coupling rate would
increase. Added LiCl provided the requisite acceleration
as the reaction proceeded to completion in 5 h at 55 °C.
tivity was comparable to that of the model system. Alkene
48 was converted to aldehyde 50 in high yield by
treatment with OsO4 and cleavage of the resultant diol
49 with NaIO4. Homologation of aldehyde 50 to alkyne
51 and hydrostannation completed the synthesis of the
C12-C25 subunit 52.40 Stille coupling between stannane
52 and iodide 18 was effected with Pd2dba3 (0.25 equiv)
in the presence of AsPh3 (2.0 equiv) and dry LiCl (3.0
equiv) in freshly distilled NMP (Scheme 7). The coupling
reaction proceeded in only 5 h at room temperature to
give diene 53 in 76% yield. Elevated temperatures were
not required.
For the macrolactonization, we were attracted to the
Yamaguchi method41 in view of the previously reported
successful lactonization of a C1-C17 seco acid by this
mixed anhydride approach.8 Following the precedent of
Evans and Calter,8 we saponified the methyl ester 53
with potassium trimethylsilanolate42 to give hydroxy acid
54 in quantitative yield (crude) (Scheme 7). Formation
of the mixed anhydride was found to be slow with 3 equiv
of the 2,4,6-trichlorobenzoyl chloride reagent. However,
as previously reported by Calter,8 increasing the amount
of this reagent to 10 equiv resulted in complete consump-
tion of the acid within 6 h at room temperature. Unfor-
tunately, the resulting mixed anhydride decomposed
upon treatment with DMAP in refluxing in toluene. None
of the lactone adduct was observed as judged by TLC and
NMR analysis of the crude product.
With the Stille coupling seemingly in control, the fully
elaborated C12-C25 subunit was assembled along the
same lines as our model system. Accordingly, alcohol 35
was oxidized with the Dess-Martin periodinane reagent
and the resultant aldehyde 36 was subjected to the Takai
methoxyallylation conditions (Scheme 6).
The allylation reaction on this substrate was signifi-
cantly slower than on the model aldehyde, taking several
days at -42 °C, which ultimately led to epimerization of
the aldehyde. However, if the reaction mixture was
stirred at -78 °C with gradual warming to 0 °C over the
course of 1 h, the reaction rate significantly increased
with a corresponding decreased epimerization of the
aldehyde (Scheme 6). Furthermore, the diastereoselec-
We believe that the mixed anhydride was formed but
that lactonization was the problematic step, possibly
(40) After completion of this work, a more direct approach to
propargyl methyl ethers involving addition of a chiral methoxyalle-
nylstannane reagent to aldehydes was reported: Savall, B. M.; Powell,
N. A.; Roush, W. R. Org. Lett. 2001, 3, 3057.
(41) (a) Inanaga, J .; Hirata, K.; Saeki, H.; Katsuki, T.; Yamaguchi,
M. Bull. Chem. Soc. J pn. 1979, 52, 1989. (b) For a selected application,
see: Hikota, M.; Tone, H.; Horita, K.; Yonemitsu, O. J . Org. Chem.
1990, 55, 7.
(36) Zhang, H. X.; Gaube´, F.; Balavoine, G. J . Org. Chem. 1990, 55,
1867.
(37) (a) Allred, G.; Liebeskind, L. S. J . Am. Chem. Soc. 1996, 118,
2748. (b) Farina, V.; Kapadia, S.; Krishnan, B.; Wang, C.; Liebeskind,
L. S. J . Org. Chem. 1994, 59, 5905.
(38) Farina, V.; Krishnan, B. J . Am. Chem. Soc. 1991, 113, 9585.
(39) Cf.: Toshima, K.; J yojima, T.; Miyamoto, N.; Katohno, M.;
Nakata, M.; Matsumura, S. J . Org. Chem. 2001, 66, 1708.
(42) Laganis, E. D.; Chenard, B. L. Tetrahedron Lett. 1984, 51, 5831.

Synthesis of Bafilomycin V1
Sch em e 7a
J . Org. Chem., Vol. 67, No. 3, 2002 739
a
Key: (a) KOTMS, THF (quant, crude); (b) 2,4,6-trichloroben-
zoyl chloride, i-Pr2NEt, THF (0.01 M), room temperature, and then
DMAP, toluene, 110 °C.
Sch em e 8a
F igu r e 3. Calculated structures for the TES-protected seco
acid 54 and the derived triol 59 (MacroModel 5.5) showing the
relationship between the C15-OH and the carboxylic group.
Only relevant hydrogen atoms are shown.
a
Key: (a) KOTMS, THF, room temperature; (b) TESCl, imid,
CH2Cl2, room temperature (65%, 2 steps); (c) 2,4,6-trichlorobenzoyl
chloride, TEA, (-)-menthol, THF (0.01 M), room temperature, and
then DMAP, toluene, 110 °C (34%).
allow for the formation of the mixed anhydride in the
absence of the ketal fragment and thus circumvent the
decomposition. Accordingly, acid 57 was treated with
2,4,6-trichlorobenzoyl chloride and TEA in the absence
of the stannane 52. Upon complete consumption of
starting material, the anhydride solution was cannulated
into a mixture of hydroxy stannane 52 and DMAP in
toluene. After 18 h at room temperature, none of the
desired ester had been formed. In fact, most of the
hydroxy stannane 52 was recovered unchanged. A major-
ity of the acid starting material was converted to a less
because the C15-OH was unreactive. We reasoned that
activation of the this OH might be achieved with a
proximal basic site. One of the more commonly employed
of the so-called “double activation” methods utilizes
commercially available 2,2′-dipyridyl disulfide, which is
activated by PPh3.43 Unfortunately, acid 54 underwent
rapid decomposition under the thioester forming condi-
tions.
To evaluate the effect of the C12-C25 ketal segment,
acid 55 was converted to the mixed anhydride and
condensed with (-)-menthol under the previously noted
modified Yamaguchi conditions (Scheme 8). The reaction
occurred with minimal decomposition as evidenced by
TLC analysis. The low yield (34%) of menthyl ester 56
actually isolated can be attributed to a premature
quenching of the reaction mixture. This result, together
with the successful lactonization of Calter,8 suggests that
the C18-C25 (pyranoside) portion of the molecule is
responsible for the failure of the lactonization step and
the eventual decomposition of the substrate.
1
polar product with a H NMR spectrum consistent with
symmetrical anhydride 58. Attempted esterification of
acid 57 and alcohol 52 with DCC and DMAP also resulted
in the formation of anhydride 58. An analogous anhy-
dride was produced in several of Calter’s abortive at-
tempts at macrolactonization.8 Esterification between
stannane 52 and the symmetrical anhydride 58 in the
presence of DMAP under thermal conditions was also
examined to no avail.
At this point, we began to suspect that conformational
factors might be responsible for the lack of success in
attempted macrolactonizations of seco acid 54. In fact,
molecular mechanics calculations (MacroModel 5.5) in-
dicated that the C15-OH and the carboxylic group were
on “opposite sides of the fence” (Figure 3). Suspecting that
the bulky TES groups might be responsible, at least in
part, for this conformational incompatiblity, we carried
out analogous calculations on the free seco acid tetrol 59
We then considered a reversal of the coupling steps
by attempting an intermolecular esterification followed
by an intramolecular Stille coupling. This strategy would
(43) (a) Corey, E. J .; Nicolaou, K. C. J . Am. Chem. Soc. 1974, 96,
5614. (b) Corey, E. J .; Brunelle, D. J .; Stork, P. J . Tetrahedron Lett.
1976, 38, 3405. (c) Corey, E. J .; Brunelle, D. J . Tetrahedron Lett. 1976,
38, 3409.

740 J . Org. Chem., Vol. 67, No. 3, 2002
Marshall and Adams
Sch em e 9a
Sch em e 10
a
Key: (a) TMSCl, imid, CH2Cl2, room temperature; (b) KOTMS,
THF, room temperature, (64%, 2 steps); (c) 2,4,6-trichlorobenzoyl
chloride, TEA, THF (0.01 M), room temperature, and then 52,
DMAP, toluene, room temperature.
(Figure 3). The results were quite astounding! Not only
were the C15-OH and the acid moiety in close (hydrogen
bonding) proximity, the structure also showed a poten-
tially beneficial hydrogen bond between the C17-OH and
the pyranoside ring oxygen. This interaction could serve
to disfavor an alternative cyclization pathway at the C-17
OH to afford an 18-membered lactone. Accordingly, the
TES protecting groups were removed by treatment with
buffered TBAF and the resulting seco acid tetrol was
subjected to the modified Yamaguchi cyclization condi-
tions. Unfortunately, the promise of the calculated struc-
ture was not fulfilled by the experiment. TLC analysis
of the cyclization reaction mixture showed no spot
attributable to the lactone product.
the deprotection resulted in hydrolysis of the methyl
pyranoside leading to 61. The H NMR spectrum of our
1
synthetic bafilomycin V1 was identical to that of the
spectrum provided to us by the SmithKline-Beecham
group with the exception of additional -OH signals at
2.93 and 2.45 ppm in our spectrum. The 13C NMR
spectrum, optical rotation, and melting point of this
substance were not available. However, the close agree-
ment of the 1H NMR spectrum and the ample stereo-
chemical precedent for our stereodefining steps, not to
mention the X-ray crystal structure determination for
triol 33, leave little doubt as to the structural integrity
of our synthetic material.
Some years following the published isolation and
structure elucidation of the bafilomycins, a group from
SmithKline-Beecham reported structure-activity rela-
tionships of several derivatives.6b One such derivative,
designated bafilomycin V1 (60), was prepared through
methanolysis of bafilomycin C2 to the open chain seco
ester (Scheme 10). Remarkably, this compound retained
biological activity as an inhibitor of vacuolar ATPases,
although with less potency than bafilomycin A1.
Because the final lactonization step to complete the
synthesis of bafilomycin A1 was evidently unfeasible, we
turned our attention to a synthesis of bafilomycin V1. This
was achieved in 80% yield through treatment of the Stille
adduct 53 with buffered TBAF in THF at room temper-
ature. It is worth noting that the use of HF‚pyridine for
Ack n ow led gm en t. Support for this research was
provided by NIH Grant RO1 CA 90383 and NSF Grant
CHE-9901319. We thank Dr. Carlo Farina, SmithKline-
1
Beecham SpA, for providing detailed H NMR spectra
of bafilomycin V1 derived from natural sources. We
gratefully acknowledge Dr. Brian A. J ohns for insightful
discussions and suggestions.
Su p p or tin g In for m a tion Ava ila ble: Complete experi-
1
mental procedures and H NMR spectra for key intermediates
and the final product, except for the compounds in Schemes 3
and 4, which can be found in the Supporting Information for
ref 23. This material is available free of charge via the Internet
at http://pubs.acs.org.
J O015864X