Synthetic Studies toward the Total Synthesis of Swinholide. 1. Stereoselective Construction of the C19-C35 Subunit

Article abstract of DOI:10.1021/jo990291y

The development of an approach directed at the total synthesis of the complex cytotoxic marine macrodiolide swinholide is described. The present study focuses on the development of a synthetic route for the preparation of the C₁₉-C₃₅ segment of the structure, which is composed of a trisubstituted pyran moiety with a contrathermodynamic anti arrangement of the C₂ and C₆ pyran substituents (swinholide C₂₇ and C₃₁) which is joined by an ethano linker to an acyclic array containing five contiguous stereocenters. The pyran subunit was constructed using a stereoselective allylation of a β-alkoxy aldehyde with 1,3-asymmetric induction and a second stereoselective allylation to prepare the C-glycosidic type of linkage. Use of the Hafner-Duthaler reagent was investigated as a potential means of constructing the anti vicinal hydroxyl-methyl relationships found in the C₁₉-C₂₄ segment but was found not to be practical in this instance. The Evans bis propionate methodology was used to introduce a four-carbon unit, and a Mukaiyama aldol was used for chain extension to incorporate the remaining two carbons and two stereocenters of this segment. Attempted use of the Hanessian benzylidene acetal fragmentation reaction in this sequence was thwarted by neighboring group participation of an oxazolidinone in one case and an unexpected regiochemical outcome in another. The approach developed affords the C₁₉-C₃₅ substructure in 18 steps overall from ethyl acetoacetate and in adequate quantities (10% overall yield) to support the projected total synthesis.

Full text of DOI:10.1021/jo990291y

4482
J . Org. Chem. 1999, 64, 4482-4491
Syn th etic Stu d ies tow a r d th e Tota l Syn th esis of Sw in h olid e. 1.
Ster eoselective Con str u ction of th e C19-C35 Su bu n it
Gary E. Keck* and Gregory D. Lundquist
Department of Chemistry, University of Utah, Salt Lake City, Utah 84102
Received February 17, 1999
The development of an approach directed at the total synthesis of the complex cytotoxic marine
macrodiolide swinholide is described. The present study focuses on the development of a synthetic
route for the preparation of the C19-C35 segment of the structure, which is composed of a
trisubstituted pyran moiety with a contrathermodynamic anti arrangement of the C2 and C6 pyran
substituents (swinholide C27 and C31) which is joined by an ethano linker to an acyclic array
containing five contiguous stereocenters. The pyran subunit was constructed using a stereoselective
allylation of a â-alkoxy aldehyde with 1,3-asymmetric induction and a second stereoselective
allylation to prepare the C-glycosidic type of linkage. Use of the Hafner-Duthaler reagent was
investigated as a potential means of constructing the anti vicinal hydroxyl-methyl relationships
found in the C19-C24 segment but was found not to be practical in this instance. The Evans bis
propionate methodology was used to introduce a four-carbon unit, and a Mukaiyama aldol was
used for chain extension to incorporate the remaining two carbons and two stereocenters of this
segment. Attempted use of the Hanessian benzylidene acetal fragmentation reaction in this sequence
was thwarted by neighboring group participation of an oxazolidinone in one case and an unexpected
regiochemical outcome in another. The approach developed affords the C19-C35 substructure in 18
steps overall from ethyl acetoacetate and in adequate quantities (10% overall yield) to support the
projected total synthesis.
In tr od u ction
have each completed the synthesis of this large marine
macrolide while Nakata10 has synthesized the monomer
preswinholide A and Mulzer11 has synthesized the C25-
C32 tetrapyran subunit.
Syn th etic P la n . Our approach to swinholide A is
similar to that of Paterson and Nicolaou in that the
molecule is first broken down into its monomer 1 and
Swinholide A, a potent cytotoxic macrolide, was first
isolated from Okinawan marine sponges of genus The-
onella swinhoei by Carmely and Koshman in 1985.1
Although originally assigned as monomeric,2 FABMS3
and X-ray crystallography4-6 have shown that swinholide
A is a dimer containing a 44-membered dilactone ring
which exists in the solid phase in the shape of a twisted
saddle. Swinholide A possesses in vitro antifungal activ-
(1) Carmely, S.; Kashman, Y. Tetrahedron Lett. 1985, 26, 511.
(2) Tanaka, J .-I.; Katori, T.; Matsuura, M.; Kitagawa, I. Tetrahedron
Lett. 1989, 30, 2963.
(3) Doi, M.; Ishida, T.; Kobayashi, M.; Kitagawa, I. J . Org. Chem.
1991, 56, 3629.
(4) Kitagawa, I.; Kobayashi, M.; Katori, T.; Yamashita, M.; Tanaka,
J .-I.; Doi, M.; Ishida, T. J . Am. Chem. Soc. 1990, 112, 3710.
(5) Kobayashi, M.; Tanako, J .-I.; Katori, T.; Matsuura, M.; Ya-
mashita, M.; Kitagawa, I. Chem. Pharm. Bull. 1990, 38, 2409.
(6) Rotem, M.; Kashman, Y. Magn. Reson. Chem. 1986, 24, 343.
(7) Bubb, M. R.; Spector, I.; Bershadsky, A. D.; Korn, E. D. J . Biol.
Chem. 1995, 270, 3463.
ity and exhibits high cytotoxic activity for L1210 (IC50
)
0.03 µg/mL) and KB (IC50 ) 0.04 µg/mL) tumor cells.3,4
Recently it has been shown that swinholide A dimerizes
actin and disrupts the actin skeleton.7 It is because of
this potent activity and unique structure that we under-
took the asymmetric total synthesis of swinholide A.
(8) (a) Paterson, I.; Cumming, J . Tetrahedron Lett. 1992, 33, 2847.
(b) Paterson, I.; Smith, J . D. J . Org. Chem. 1992, 57, 3261. (c) Paterson,
I.; Yeung, K. Tetrahedron Lett. 1993, 34, 5347. (d) Paterson, I.; Smith,
J . D. Tetrahedron Lett. 1993, 34, 5354. (e) Paterson, I.; Cumming, J .
G.; Smith, J . D.; Ward, R. A. Tetrahedron Lett. 1994, 35, 3405. (f)
Paterson, I.; Smith, J . D.; Ward, R. A.; Cumming, J . G. J . Am. Chem.
Soc. 1994, 116, 2615. (g) Paterson, I.; Yeung, K.; Ward, R. A.;
Cumming, J . G.; Smith, J . D. J . Am. Chem. Soc. 1994, 116, 9391. (h)
Paterson, I.; Cumming, J . G.; Ward, R. A.; Lamboley, S. Tetrahedron
1995, 51, 9393. (i) Paterson, I.; Smith, J . D.; Ward, R. A. Tetrahedron
1995, 51, 9413. (j) Paterson, I.; Ward, R. A.; Smith, J . D.; Cumming,
J . G.; Yeung, K. Tetrahedron 1995, 51, 9437. (k) Paterson, I.; Yeung,
K.; Ward, R. A.; Smith, J . D.; Cumming, J . G.; Lamboley, S. Tetrahe-
dron 1995, 51, 9467.
(9) (a) Patron, A. P.; Richter, P. K.; Tomaszewski, M. J .; Miller, R.
A.; Nicolaou, K. C. J . Chem. Soc., Chem. Commun. 1994, 1147. (b)
Richter, P. K.; Tomaszewski, M. J .; Miller, R. A.; Patron, A. P.;
Nicolaou, K. J . Chem. Soc., Chem. Commun. 1994, 1151. (c) Nicolaou,
K. C.; Ajito, K.; Patron, A. P.; Khatuya, H.; Richter, P. K.; Bertinato,
P. J . Am. Chem. Soc. 1996, 118, 3059.
(10) (a) Nakata, T.; Komatsu, T.; Nagasawa, K. Chem. Rev. Bull.
1994, 42, 2403. (b) Nakata, T.; Komatsu, T.; Nagasawa, K.; Yamada,
H.; Takahashi, T. Tetrahedron Lett. 1994, 35, 8225. (c) Nagasawa, K.;
Shimizu, I.; Nakata, T. Tetrahedron Lett. 1996, 37, 6881. (d) Nagasawa,
K.; Shimizu, I.; Nakata, T. Tetrahedron Lett. 1996, 37, 6885.
Swinholide A has attracted much interest in the field
of synthetic organic chemistry. Paterson8 and Nicolaou9
10.1021/jo990291y CCC: $18.00 © 1999 American Chemical Society
Published on Web 05/19/1999

Total Synthesis of Swinholide
J . Org. Chem., Vol. 64, No. 12, 1999 4483
further disconnected into the C1-C18 (2) and C19-C32 (3)
pyran subunits, which we planned to couple together
using a Mukaiyama aldol reaction. The focus of this
discussion will be on the synthesis of 3.
etate was reduced using Noyori conditions12 to provide
the â-hydroxy ester 6 in 98% yield and 98% ee. After
benzyl protection13 and a low-temperature half-reduction
using DIBAL, aldehyde 7 was subjected to a chelation-
controlled allylstannane addition. This allylation process,
previously developed in our laboratories and used in the
synthesis of (-)-colletol,14 produced 5 in 75% yield and
with 29:1 anti:syn diastereoselectivity.
After conversion of the hydroxyl group to the methyl
ether using KH/MeI and removal of the benzyl group
using dissolving metal conditions, 8 was treated with
ozone and worked up with DMS followed by acidic
methanol to give tetrahydropyran 9. This material was
isolated as a 3:1 mixture of anomers which did not
require separation, since the planned installation of the
allyl side chain utilized conditions which would allow
kinetically controlled axial attack on the oxonium ion
derived from either anomer of 9.
This was first attempted using a variety of conditions
with allyltributylstannane as the three-carbon nucleo-
phile which should produce a surrogate of the desired
aldehyde intermediate 4. A quick survey of Lewis acids
(Bu3SnOTf, TMSOTf, BF3‚OEt2) and solvents (CH3CN,
CH2Cl2) provided the desired allylpyran 10 but with only
moderate levels of diastereoselectivity. At the same time
this work was in progress, the Paterson group published
their synthesis of the C19-C32 segment of swinholide A
and reported the use of allyltrimethylsilane and TMSOTf
for this conversion.3a These conditions gave the desired
product 10 as one diastereomer with a yield greater than
90%, perhaps due to a later transition state for nucleo-
philic addition with the less reactive allylsilane. Thus we
decided, for the sake of time and resources, to use these
conditions in our synthesis. Ozonolysis of 10 gave alde-
hyde 4 in 95% yield. Interestingly, ozonolysis of 10 in
CH2Cl2/MeOH and workup with DMS gave a 30% yield
of the pyran aldehyde dimethyl acetal 11. The formation
The original plan for assembling 3 involved a series of
iterative polypropionate additions to pyran aldehyde 4
which could be made from homoallylic alcohol 5 and,
ultimately, from ethyl acetoacetate through a series of
highly selective reagent- and substrate-controlled reac-
tions. This approach (outlined above) describes the
overall basis of our plan. More specifically, if one intro-
duces an additional hydroxyl substituent at C25, then a
reasonable means for coupling of the pyran subunit with
the C19-C24 acyclic segment can be envisioned by nu-
cleophilic addition to 4. Moreover, if this hydroxyl is of â
orientation, then it can be seen that three iterative
disconnections of anti vicinal hydroxyl-methyl units can
be envisioned, as indicated below in structure 3a . We
hoped to accomplish these anti bond construction events
through the use of the Hafner-Duthaler reagent (vide
infra). Thus the plan becomes one of iterative chain
extension from pyran aldehyde 4 involving anti bond
constructions under “reagent control”. Crotyl addition to
aldehyde 4 should give the desired anti-homoallylic
alcohol using the appropriate Hafner reagent. After
deoxygenation and aldehyde formation, a second crotyl
addition would be performed. Finally, conversion of the
terminal olefin to the aldehyde, a third crotyl addition,
and protection and aldehyde formation would give the
desired C19-C32 segment 14. The pyran aldehyde 4 itself
was to be prepared by two sequential allylstannane
additions previously studied in our laboratories.
(11) Mulzer, J .; Meyer, F.; Buschmann, J .; Luger, P. Tetrahedron
Lett. 1995, 36, 3503.
(12) Noyori, R.; Ohkuma, T.; Kitamura, M.; Takaya, H.; Sayo, N.;
Kumbayashi, H.; Akutagawa, S. J . Am. Chem. Soc. 1987, 109, 5856.
(13) (a) Iverson, T.; Bundle, D. R. J . Chem. Soc., Chem. Commun.
1981, 1240. (b) Wessel, H.-P.; Iverson, T.; Bundle, D. R. J . Chem. Soc.,
Perkin Trans. 1 1985, 2247. (c) Widmer, J . Synthesis 1987, 568. (d)
Clizbe, A.; Overman, L. E. Org. Synth. 1978, 8, 4.
Con str u ction of th e Dih yd r op yr a n Un it a n d In -
vestiga tion of th e Iter a tive Ap p r oa ch . Ethyl acetoac-
(14) Keck, G. E.; Murry, J . A. J . Org. Chem. 1991, 56, 6606.

4484 J . Org. Chem., Vol. 64, No. 12, 1999
Keck and Lundquist
tremely high with the anti bond construction product
formed preferentially.
Examination of this synthetic pathway began by
ensuring the Hafner-Duthaler reaction would work in
our hands. After working out the intricacies of this reac-
tion (there were many), we were able to successfully re-
peat the crotyl addition to hydrocinnamaldehyde as re-
ported in the original paper. Unfortunately, translation
of this method to pyran aldehyde 4 proved to be futile.
In small-scale (25-50 mg) reactions, addition product 12
could be obtained in 70% yield as a single diastereomer.
However, these results could not be reliably repeated, and
increasing the scale of the reaction required several
equivalents of the titanium reagent. The low yields ob-
tained on scale-up, the large amounts of auxiliary re-
quired, and the overalll technical difficulty of the reaction
(the crotyl Grignard reagent must be freshly prepared
and titrated prior to use and the magnesium salts had
to be filtered off by running the Grignard reagent directly
into the reaction mixture through oven-dried pipets
packed with Celite) led us to reexamine our synthetic
route.
Con str u ction of th e Acyclic Segm en t Usin g th e
Eva n s Bis P r op ion a te Syn th on . In 1990 Evans and
co-workers17 reported that the readily available â-ke-
toimide 15 could be used in aldol reactions to provide a
variety of diastereomeric aldol products. The product
stereochemistry was shown to depend on the type of
Lewis acid and amine base used. Treatment of 15 with
TiCl4 and Hu¨nig’s base and reaction of the resulting
enolate with a variety of aldehydes provide syn,syn aldol
products. Likewise, the use of Sn(OTf)2/triethylamine or
(c-hex)2BCl/EtNMe2 gives predominantly anti,syn and
anti,anti aldol products, respectively.
of this undesired side product was easily quelled by
omitting MeOH and working up the reaction with PPh3.
Unfortunately, the workup and chromatography are
somewhat more tedious using the latter method.
With aldehyde 4 in hand, we were ready to extend the
pyran side chain using a series of anti-selective crotyl
additions with the Hafner-Duthaler reagent. In 1992,
Hafner and Duthaler reported16 the use of a new tartrate-
derived titanium reagent, prepared using an allyl or
crotyl Grignard reagent, in the allylation or crotylation
of prochiral and chiral aldehydes. Generally, the yields
and selectivities reported for these reactions were ex-
Looking at the C19-C32 segment it becomes obvious
that the TiCl4/Hu¨nig’s base reaction conditions will give
(15) (a) Wiley, M. Ph.D. Thesis, University of Utah, 1988. (b) Keck,
G. E.; Castellino, S.; Wiley, M. R. J . Org. Chem. 1986, 51, 5478.

Total Synthesis of Swinholide
J . Org. Chem., Vol. 64, No. 12, 1999 4485
the desired 1,3-syn relationship between the two methyl
substituents. The configuration of the resulting hydroxyl
group is relatively inconsequential, since it must ulti-
mately be removed. When 4 was subjected to these
reaction conditions, aldol product 16 was obtained in 86%
yield as a single diastereomer. Moreover, this reaction
was reproducible and could be done on a large scale with
no compromise in diastereoselectivity and only a slight
decrease in yield (80% using 750 mg of aldehyde).
Treatment of 18 under these conditions proceeded very
smoothly; however, the product obtained was not the
desired deoxygenated compound 19 but was, in fact, 20.
This product can be envisioned to arise as indicated from
attack of the oxazolidinone at the C23 position with
bromide ultimately relieving the resulting positive charge
on the oxazolidinone. While this was an interesting
result, it obviously did not help to meet the final synthetic
goal. In an attempt to rectify this problem, the oxazoli-
dinone was removed with LiBH4 and the resulting alcohol
was protected as the TBS ether. Subjecting 21 to the
same reaction conditions led to a mixture of TBS-
protected compounds 22 and 23 along with the depro-
tected compounds 24 and 25. Unfortunately, the major
compounds present in these mixtures were the products
arising from bromide attack at the seemingly less acces-
sible position, between the two methyl-bearing carbons.
The seemingly extraneous hydroxyl group could now
be used as a handle to selectively reduce the ketone to
give an anti 1,3-diol. This was accomplished using
18
Me4NBH(OAc)3 which provided 17 exclusively. Diol 17
was then protected as the benzylidene acetal with the
intention of using this acetal as a means of removing the
C25 oxygen functionality. Hannesian has shown,19 pri-
marily in the context of carbohydrate chemistry, that
treatment of a benzylidine acetal with NBS provides a
labile intermediate (in which bromine is positioned at the
acetal carbon) from free radical bromination. Upon in situ
solvolysis of the benzylic bromide, the dioxolane ring is
opened, with bromide attacking at the less sterically
encumbered carbon. The resulting bromobenzoate can,
of course, be reduced with Bu3SnH which would lead to,
in our case, the desired deoxygenated product 19.
Because these variations of the Hannesian reaction did
not work, it became clear that other options had to be
examined. It was decided that the triol unit would have
to be differentially protected to allow for deoxygenation
at C25. Initial attempts toward accomplishing this goal
using the Tischenko-Evans reaction20 (SmI2, RCHO) on
16 did not succeed and led to a complex mixture of
(18) Evans, D. A.; Chapman, K. T.; Carreira, E. M. J . Am. Chem.
Soc. 1988, 110, 3560.
(19) Hannesian, S.; Plessas, N. R. J . Org. Chem. 1969, 34, 1035,
1045, 1053.
(20) Evans, D. A.; Hoveyda, A. H. J . Am. Chem. Soc. 1990, 112,
6447.
(16) Hafner, A.; Duthaler, R. O.; Marti, R.; Rihs, G.; Rothe-Streit,
P.; Schwarzenbach, F. J . Am. Chem. Soc. 1992, 114, 2321.
(17) Evans, D. A.; Clark, J . S.; Matternich, R.; Novack, V. J .;
Sheppard, G. S. J . Am. Chem. Soc. 1990, 112, 866.

4486 J . Org. Chem., Vol. 64, No. 12, 1999
Keck and Lundquist
products. Therefore, the oxazolidinone of 17 was removed
using LiBH4 to give triol 26 as a solid which could be
recrystallized from 5% ethyl acetate/hexanes. Although
this is a seemingly straightforward reaction, it was
discovered that the outcome of the reaction was ex-
tremely sensitive to the conditions employed. Large
amounts of the side product 27, in which the oxazolidi-
none carbonyl had been removed, were produced using
Evans’ published method of LiBH4 in THF/MeOH. It was
found that the yield of this side product could be
significantly reduced by excluding MeOH and using wet
THF or Et2O.
DDQ,21 though, led to a >10:1 ratio of 30 to 31, and the
remaining alcohol was free for deoxygenation.
As larger amounts of triol 26 were brought through
the synthesis, a concern arose regarding the Me4NBH-
1
(OAc)3 reduction of 16. The H NMR of 26 began to show
significant amounts of a previously unnoticed “side
product” which would occasionally disappear after silica
gel chromatography and which disappeared entirely after
the LiBH4 reaction. It was feared that either a large
amount of the syn 1,3-diol was produced in the Me4NBH-
(OAc)3 reaction and then removed upon chromatography
or the product could be epimerizing upon silica gel
chromatography. It should be pointed out that the yield
for the two-step sequence, Me4NBH(OAc)3 followed by
LiBH4, was quite high (81% on a l g scale) and very clean
(one diastereomer). To gain insight into this problem, a
known method for making syn 1,3-diols was used. Thus,
16 was treated with Zn(BH4)2 which presumably provided
syn diol 28. The NMR spectrum obtained for this diol
Previously, during the investigation of the Hafner-
Duthaler methodology, deoxygenation of the crotyl ad-
dition product 12 was attempted via hydride reductions
of the corresponding mesylate and phosphoroamidate
derivatives. Because these reactions did not succeed, it
was decided to use standard radical conditions to remove
the hydroxyl substituent in 30. Initially, thiocarbamate
32 was formed using TCDI and DMAP in dichloroethane
at reflux,22 and then 32 was treated with Bu3SnH and
AIBN in toluene at 110 °C. These two reactions appeared
to proceed reasonably well, and the product 33 was taken
onto the next reaction; however, regioselective reductive
ring opening of 33 using DIBAL23 at 0 °C gave <50% yield
(three steps) of alcohol 34 with no detection of any other
products and complete consumption of 33. Because the
TCDI and DIBAL reactions were so clean, the Bu3SnH
reaction was investigated more thoroughly. Paterson
reported in his synthesis of the C19-C32 segment that
difficulties were encountered in the deoxygenation of a
very similar intermediate. They found that a side product
was produced in which the radical intermediate ab-
stracted hydrogen from the C29 position, scrambling that
center. To overcome this problem, excess Bu3SnH (10
equiv) was added by syringe to a solution of the radical
precursor in toluene at reflux, and the desired deoxy-
1
did not match either set of H NMR signals observed in
the Me4NBH(OAc)3 reaction and gave, itself, a very clean
set of signals. Diol 28 was further reduced to triol 29
using LiBH4, which gave a completely different set of
NMR signals than those seen for 26 and, again, was very
clean. We now felt confident that 17 was indeed the de-
sired anti diol and that the extra NMR signals observed
on occasion were a result of our inability to completely
remove some sort of inorganic material from the triol.
As a result, we carried the material through to 26 without
purification.
Differentiation of the two secondary alcohols was still
a concern, and the problem was addressed as follows.
Triol 26 was treated with PMPCHO or PMPCH(OMe)2
under a variety of acid-catalyzed acetalization conditions;
however, it was generally found that, although the
desired “external” acetal 30 was formed preferentially
over the “internal” acetal 31, this occurred in approxi-
mately a 3:1 ratio. Subjection of 26 to PMBOMe and
(21) Oikawa, Y.; Nishi, T.; Yonemitsu, O. Tetrahedron Lett. 1983,
24, 4037.
(22) (a) Rasmussen, J . R.; Slinger, C. J ..; Kordish, R. J .; Newman-
Evans, D. D. J . Org. Chem. 1981, 46, 4843. (b) Barton, D. H. R.;
McCombie, S. W. J . Chem. Soc., Perkin Trans. 1 1975, 1574.
(23) Takano, S.; Akiyan, M.; Sato, S.; Ogasawara, K. Chem. Lett.
1983, 1593.

Total Synthesis of Swinholide
J . Org. Chem., Vol. 64, No. 12, 1999 4487
genated product was produced in high yield. When 32
was subjected to the same conditions (notice the absence
of a radical initiator) using a freshly opened bottle of
Bu3SnH, 33 was isolated in 95% yield and the ensuing
DIBAL reaction worked very well (84%).
clear whether this was a result of the small-scale
synthesis using strongly basic reagents or was due to the
presence of a thiolester in place of a “normal” ester;
alkylations of this type of thiolester are not well prece-
dented in the literature. In any event the two-step
process was not successful, and attention was turned to
the one-step method.
Oxidation of the primary alcohol to aldehyde 35 with
TPAP24 seemed, at first, to be quite straightforward, with
only a small amount of an unidentified side product
observed in the reaction. As this reaction was repeated,
this side product occasionally became more prevalent and
ultimately was identified as the elimination product 36.
It appeared that the amount of 36 formed depended
largely on the amount of TPAP used in the reaction. At
this late stage in the synthesis it was very difficult to
measure the precise amount of a catalytic reagent such
as TPAP, and alternative reagents were investigated.
Fortunately, the Dess-Martin periodinane25 oxidation
proceeded quite well with none of the elimination product
detected. Although this reaction required 10 equiv of
periodinane, an aqueous workup, and a chromatography
(all of which are unnecessary when TPAP is used), it was
deemed the superior procedure for this transformation.
Gennari has reported that aldol reactions between 40
and various aldehydes proceed with moderate selectivity
for the anti product independent of the E to Z ratio of
40.28 The Woodward group has also reported the use of
this type of chemistry in the synthesis of erythromycin
with very high diastereoselectivity.29 Using the conditions
set forth by Gennari, 35 and 40 (1.5 equiv) were stirred
in CH2Cl2 at -78 °C and BF3‚OEt2 (2 equiv) was added
dropwise to the solution. This led to a 5:1 mixture of
1
diastereomers by H NMR which have thus far proven
inseparable by chromatography. The major component
of this mixture has been characterized with a fair degree
of certainty as the desired product 41. Homonuclear
decoupling experiments revealed a J H18-H19 coupling
constant of 9.5 Hz, which would lead one to believe H18
and H19 were anti to one another (assuming 41 exists
largely in the hydrogen-bonded conformation shown).
This would be consistent with the NMR data reported
by Gennari for the anti products but not with the
coupling constants found for the syn products. While the
absolute stereochemistry of this reaction has not yet been
proven, the structure will ultimately be confirmed by the
completion of the synthesis of preswinholide A. Finally,
41 has been protected as the PMB ether30 and the
thiolester reduced to aldehyde 42 with DIBAL at -78
°C. Aldehyde 42 has been produced on a very small scale
(ca. 5 mg) so as not to deplete the resources on hand.
Thus, a yield for this step is not meaningful, although,
in accord with our previous experiences with the DIBAL
reduction of thiolesters, this reaction was very clean.
Alcohol 41 was deemed a suitable storage point for
material in this portion of the route on the basis that it
still allows for variations in strategy should the proposed
coupling reaction not proceed as planned.
With sufficient quantities of aldehyde 35 in hand, the
final three carbons and two stereocenters could be
installed. Initially this was attempted in a two-step
process using well-known, seemingly reliable chemistry
with the idea of then working out a one-step procedure.
Thus, aldehyde 35 was treated with thiosilylketene acetal
37 in the presence of BF3‚OEt2 and â-hydroxy thiolester
38 was formed in good yield and high diastereoselectiv-
ity.26 Unfortunately, the following reaction, an attempted
Frater-type alkylation of the dianion of 38 with MeI,27
did not give any of the alkylation product 39. It is not
In summary, the C19-C32 segment of swinholide A (in
the form of aldehyde 42) has been made in 16 steps from
the known compound 5 (20 steps from ethyl acetoacetate).
(24) Ley, S. V.; Norman, J .; Griffith, W. P.; Mardsen, S. P. Synthesis
1994, 639.
(25) Dess, D. B.; Martin, J . C. J . Am. Chem. Soc. 1991, 113, 7277.
(26) (a) Mukaiyama, T.; Narasaka, K.; Banno, K. Chem. Lett. 1973,
1011. (b) Evans, D. A.; Dart, M. J .; Duffy, J . L.; Yang, M. G. J . Am.
Chem. Soc. 1996, 118, 4322.
(28) Gennari, C.; Beretta, M. G.; Bernardi, A.; Moro, G.; Scolastico,
C.; Todeschini, R. Tetrahedron 1986, 42, 893.
(29) Woodward, R. B.; et al. J . Am. Chem. Soc. 1981, 103, 3210.
(30) Nakajima, N.; Horita, K.; Abe, R.; Yonemitsu, O. Tetrahedron
Lett. 1988, 29, 4139.
(27) Frater, G. Helv. Chim. Acta 1979, 62, 2825.

4488 J . Org. Chem., Vol. 64, No. 12, 1999
Keck and Lundquist
10.4, 16.9 Hz, 1H), 5.1-5.0 (m, 2H), 4.60 (d, J AB ) 11.7 Hz,
1H), 4.41 (d, J AB ) 11.7 Hz, 1H), 3.74 (dqd, J ) 3.3, 6.2, 9.0
Hz, 1H), 3.48 (m, 1H), 3.27 (s, 3H), 2.3-2.2 (m, 2H), 1.64 (ddd,
J ) 3.2, 9.4, 14.5 Hz, 1H), 1.51 (ddd, J ) 3.3, 9.5, 14.5 Hz,
1H), 1.20 (d, J ) 6.2 Hz, 3H); 75-MHz 13C NMR (CDCl3) δ
138.9, 134.4, 128.3(2), 127.8(2), 127.4, 117.1, 76.8, 71.6, 70.6,
56.8, 42.4, 37.9, 20.0. Anal. Calcd for C15H22O2: C, 76.88; H,
9.46. Found: 77.02; H, 9.41.
P r ep a r a tion of (4S,6S)-4-(Meth oxy)-6-h yd r oxy-1-h ep -
ten e (8). To a 500 mL, three-necked flask equipped with a
mechanical stirrer, Dewar condenser, and an ammonia inlet
was added (4S,6S)-4-(methoxy)-6-(benzyloxy)-1-heptene (2.00
g, 8.53 mmol) dissolved in dry THF (75 mL). The flask was
cooled to -78 °C, and ammonia was condensed until a final
volume of 150 mL was achieved. Small pieces of lithium wire
were added to the solution until a deep blue color persisted.
At this time, solid ammonium chloride was added until the
solution was white. The solution was allowed to warm to room
temperature over a period of 4 h and then diluted with Et2O
(75 mL) and H2O (75 mL). The organic layer was separated
and the aqueous layer extracted two times with CH2Cl2 (75
mL). The organic layers were combined, washed with satu-
rated aqueous NaHCO3 solution (100 mL), then dried over
MgSO4, filtered, and concentrated in vacuo. The resulting oil
was purified by chromatography over a 3.8 × 19 cm silica gel
column (slurry packed with hexanes) eluting with 30% ethyl
acetate/hexanes. The product-containing fractions were con-
centrated to yield 8 (1.18 g, 96% yield) as a clear volatile
liquid: [R]D +40.7° (c 0.425, CHCl3); Rf ) 0.29 (35% ethyl
Excluding the final two steps of this synthesis (yields for
these transformations have not been determined), 41 was
synthesized from ethyl acetoacetate in 18 steps. The
overall yield of 41 was 10% with 98% ee and 62% ds. Our
synthesis provides 41 in sufficient quantity to support
the overall synthesis and allow for coupling studies at a
later time.
1
acetate/hexanes); 300-MHz H NMR (CDCl3) δ 5.79 (ddt, J )
7.0, 10.2, 17.1 Hz, 1H), 5.12 (dd, J ) 2.1, 10.2 Hz, 1H), 5.07
(dd, J ) 2.1, 17.1 Hz, 1H), 4.10 (m, 1H), 3.56 (m, 1H), 3.38 (s,
3H), 2.74 (d, J ) 3.9 Hz, 1H), 2.5-2.2 (m, 2H), 1.7-1.6 (m,
2H), 1.20 (d, J ) 6.3 Hz, 3H); 75-MHz 13C NMR (CDCl3) δ
134.3, 117.3, 78.6, 64.8, 56.8, 41.1, 37.4, 23.7; IR (neat) 3420
(br) cm-1. Anal. Calcd for C8H16O: C, 66.63; H, 11.18. Found:
C, 66.62; H, 11.14.
Exp er im en ta l Section
Gen er a l. Solvents were purified according to the guidelines
in Purification of Common Laboratory Chemicals (Perrin and
Armarego, Pergamon: Oxford, 1966). Yields were calculated
for material judged homogeneous by thin-layer chromatogra-
phy and NMR. Thin-layer chromatography was performed on
Merck Kieselgel 60 F254 plates eluting with the solvent
indicated, visualized with a 254 nm UV lamp, and stained with
an ethanolic solution of 12-molybdophosphoric acid or p-
anisaldehyde. Flash column chromatography was performed
with Davisil 62 silica gel slurry packed with the solvents
indicated. Nuclear magnetic resonance spectra were acquired
at 300 MHz for 1H and 75 MHz for 13C. The abbreviations s,
d, t, q, m, and br stand for singlet, doublet, triplet, quartet,
multiplet, and broad, respectively. Melting points were ob-
tained on a Mel-Temp electrochemical melting point apparatus
and are uncorrected. Optical rotations were obtained (Na D
line) using a microcell with a 1 dm path length. Analytical C
and H analyses were performed by Atlantic Microlab, Inc.,
Norcross, GA. Glassware for all reactions was oven-dried at
125 °C and cooled in a desiccator prior to use. Liquid reagents
and solvents were introduced by oven-dried syringes through
septa-sealed flasks under a nitrogen atmosphere.
P r ep a r a tion of (4R,6S)-2,4-(Dim eth oxy)-6-m eth yltet-
r a h yd r op yr a n (9). To a stirring solution of 8 (0.800 g, 5.55
mmol) in MeOH (35 mL) at -78 °C was bubbled in ozone until
a light blue color persisted. Oxygen was then bubbled in until
the solution turned clear. DMS (10 mL) was added, and the
solution was allowed to warm to room temperature. After a
period of 12 h, 1 M aqueous HCl (10 mL) was added. The
resulting solution was allowed to stir for 4 h, then quenched
with saturated aqueous NaHCO3 solution, and diluted with
CH2Cl2 (30 mL). The organic layer was separated and the
aqueous layer extracted three times with CH2Cl2 (25 mL). The
organic layers were combined, dried over Na2SO4, and con-
centrated in vacuo. The resulting oil was purified by chroma-
tography over a 3.5 × 21 cm silica gel column (slurry packed
with pentanes), eluting with a solvent gradient from pentanes
through 30% ether/pentanes. The product-containing fractions
were combined and concentrated to yield 9 (0.844 g, 95% yield)
as a colorless volatile liquid.
P r ep a r a tion of (4S,6S)-4-(Meth oxy)-6-(ben zyloxy)-1-
h ep ten e. Into a 250 mL round-bottom flask with stirbar was
placed potassium hydride (2.48 g, 21.8 mmol, 35% oil disper-
sion) washed with hexanes (3 × 20 mL) followed by dry THF
(100 mL). The stirring solution was cooled to 0 °C, and 5 (2.40
g, 10.9 mmol), dissolved in THF (6 mL), was added dropwise
via cannula. After 20 min, MeI (4.64 g, 32.7 mmol) was added
to the solution via syringe. This solution was warmed to room
temperature, allowed to stir for 1 h, and then quenched by
cautious addition of a saturated aqueous NaHCO3 solution (50
mL). The organic layer was separated and the aqueous layer
extracted twice with CH2Cl2 (50 mL). The organic layers were
combined, washed with saturated aqueous NaHCO3 solution
(75 mL) and brine (75 mL), dried over Na2SO4, filtered,
concentrated, and purified by chromatography over a 4 × 15
cm silica gel column (slurry packed with hexanes) eluting with
a solvent gradient from hexanes through 12% ethyl acetate/
hexanes. The product-containing fractions were concentrated
to yield 2.37 g (93% yield) of a clear, colorless oil: [R]D +72.1°
(c 2.60, CHCl3); Rf ) 0.47 (20% ethyl acetate/hexanes); 300-
P r ep a r a t ion of (2R,4R,6S)-4-(Met h oxy)-6-m et h yl-2-
(p r op -1-en yl)tetr a h yd r op yr a n (10). To a stirring solution
of 9 (0.890 g, 55.6 mmol) and allyltrimethylsilane (1.27 g, 11.1
mmol) in acetonitrile (50 mL) at 0 °C was added trimethylsilyl
trifluoromethanesulfonate (0.103 g, 0.556 mmol). After being
stirred for 15 min, the reaction was quenched with a saturated
aqueous NaHCO3 solution (30 mL) and diluted with CH2Cl2
(30 mL) and H2O (30 mL). The organic layer was separated
and the aqueous layer extracted three times with CH2Cl2 (50
mL). The organic layers were combined, dried over MgSO4,
and concentrated in vacuo. The resulting oil was purified by
chromatography over a 3.5 × 17.5 cm silica gel column (slurry
packed with pentanes), eluting with 30% ether/pentanes. The
product-containing fractions were combined and concentrated
to yield 0.803 g (84%) of 10 as a colorless volatile liquid: [R]D
-59.5° (c 2.91, CHCl3); Rf ) 0.49 (35% ethyl acetate/hexanes);
1
300-MHz H NMR (CDCl3) δ 5.79 (ddt, J ) 7.0, 10.2, 17.1 Hz,
1H), 5.09 (dd, J ) 1.8, 10.2 Hz, 1H), 5.04 (dd, J ) 1.8, 17.1
Hz, 1H), 4.08 (m, 1H), 3.76 (dqd, J ) 2.9, 6.2, 9.7 Hz, 1H),
3.53 (m, 1H), 3.33 (s, 3H), 2.5-2.2 (m, 2H), 1.98 (m, 1H), 1.86
1
MHz H NMR (CDCl3) δ 7.2-7.4 (m, 5H), 5.80 (ddt, J ) 7.1,

Total Synthesis of Swinholide
J . Org. Chem., Vol. 64, No. 12, 1999 4489
(m, 1H), 1.55 (ddd, J ) 5.4, 10.1, 12.9 Hz, 1H), 1.21 (buried
m, 1H), 1.20 (d, J ) 6.2 Hz, 3H); 75-MHz 13C NMR (CDCl3) δ
135.0, 116.7, 73.0, 71.5, 65.1, 55.3, 38.5, 36.8, 33.8, 21.7. Anal.
Calcd for C10H18O2: C, 70.55; H, 10.68. Found: C, 70.43; H,
10.68.
mmol) was dissolved in 40 mL of acetonitrile (20 mL rinse)
and added via cannula. After 45 h, H2O (16 mL) was added
rapidly followed by MeOH (60 mL), and the solution was
concentrated. This was repeated four times. The resulting
residue was quenched carefully by the addition of 5% aqueous
NaHCO3 solution (120 mL) and extracted three times with
CH2Cl2 (80 mL). The combined organic phases were dried over
MgSO4, filtered through Celite, and concentrated to yield 17
as a foam which was used without further purification.
To a stirring solution of 17 in wet Et2O (23 mL) at 0 °C was
added 2.3 mL of lithium borohydride (2.0 M in THF) dropwise
via syringe. After 1 h the cloudy white solution was quenched
with 0.5 N Rochelle salts (10 mL) and stirred for 1 h. This
solution was poured into brine (50 mL) and extracted four
times with CH2Cl2 (75 mL). The combined organic phases were
dried over MgSO4, filtered through Celite, and concentrated.
The resulting oil was purified by chromatography over a 5 ×
13 cm silica gel column (slurry packed with 3% MeOH/CHCl3)
eluting with 500 mL portions of 3% and 5% MeOH/CHCl3 and
collected in 25 mL fractions. The product-containing fractions
(22-33) were combined and concentrated to yield 0.533 g (81%,
two steps) of 26 as a colorless solid. This solid could be
recrystallized from 5% ethyl acetate/hexanes (mp ) 73 °C) but
was generally used without further purification. [R]D -27.8°
P r ep a r a tion of (2R,4R,6S)-2-(Eth a n -1-a l)-4-(m eth oxy)-
6-m eth yltetr a h yd r op yr a n (4). To a stirring solution of 10
(0.612 g, 3.59 mmol) in CH2Cl2 (40 mL) at -78 °C was bubbled
in O3 until a light blue color persisted. Oxygen was then
bubbled in until the solution turned clear. PPh3 (1.85 g, 7.05
mmol) was added, and the solution was allowed to warm to
room temperature. After a period of 2.5 h, the solution was
concentrated in vacuo. The resulting oil was purified by
chromatography over a 3.6 × 20 cm silica gel column (slurry
packed with 10% acetone/hexanes), eluting with 10% acetone/
hexanes (500 mL) and 25% acetone/hexanes (200 mL) and
collected in 9 mL fractions. The product-containing fractions
(43-77) were combined and concentrated to yield 0.591 g (95%)
of 4 as a colorless oil: [R]D -28.9° (c 1.06, CHCl3); Rf ) 0.33
(25% acetone/hexanes); 300-MHz 1H NMR (CDCl3) δ 9.75 (dd,
J ) 3.4, 1.6 Hz, 1H), 4.70 (dt, J ) 5.2, 9.2 Hz, 1H), 3.78 (dqd,
J ) 9.4, 6.4, 3.1 Hz, 1H), 3.52 (m, 1H), 3.34 (s, 3H), 2.82 (ddd,
J ) 3.4, 9.3, 15.9 Hz, 1H), 2.50 (ddd, J ) 1.6, 5.3, 15.9 Hz,
1H), 1.99 (m, 1H), 1.82 (m, 1H), 1.70 (ddd, J ) 5.3, 9.4, 13.1
Hz, 1H), 1.27 (buried m, 1H), 1.23 (d, J ) 6.4 Hz, 3H); 75-
MHz 13C NMR (CDCl3) δ 200.5, 72.6, 66.3, 65.8, 55.3, 46.4,
1
(c 0.670, CHCl3); Rf ) 0.33 (10% MeOH/CHCl3); 300-MHz H
NMR (CDCl3) δ 4.3-4.2 (m, 2H), 4.0-3.5 (br s, 3H), 3.97 (m,
1H), 3.80 (dd, J ) 3.3, 11.0 Hz, 1H), 3.66 (dd, J ) 7.3, 11.0
Hz, 1H), 3.59 (m, 1H), 3.54 (dd, J ) 4.6, 7.8 Hz, 1H), 3.35 (s,
3H), 2.05 (ddd, J ) 9.8, 11.2, 14.4 Hz, 1H), 2.0-1.9 (m, 2H),
1.85 (m, 1H), 1.8-1.6 (m, 2H), 1.40 (ddd, J ) 7.8, 7.8, 13.2
Hz, 1H), 1.28 (buried m, 1H), 1.28 (d, J ) 6.6 Hz, 3H), 1.03 (d,
J ) 7.1 Hz, 3H), 0.90 (d, J ) 6.8 Hz, 3H); 75-MHz 13C NMR
(CDCl3) δ 82.7, 73.2, 72.8, 71.0, 68.0, 66.2, 55.5, 37.5, 37.2,
36.3, 35.7, 35.5, 21.1, 14.0, 11.6; 75-MHz DEPT NMR (CDCl3)
CH3 δ 55.5, 21.1, 14.0, 11.6; CH2 δ 68.0, 36.3, 35.7, 35.5; CH
δ 82.7, 73.2, 72.8, 71.0, 66.2, 37.5, 37.2; IR (CHCl3) 3500 (br),
3130 cm-1. Anal. Calcd for C15H30O5: C, 62.04; H, 10.41.
Found: C, 62.11; H, 10.32.
37.5, 34.4, 21.1; IR (neat) 2970, 2935, 2825, 2730, 1725 cm-1
.
Anal. Calcd for C9H16O3: C, 62.77; H, 9.36. Found: C, 62.80;
H, 9.35.
P r epar ation of (4S)-4-Ben zyl-3-[(2S,4S,5R)-2,4-dim eth yl-
5-h yd r oxy-6-((2S,4R, 6S)-4-(m eth oxy)-6-m eth yltetr a h y-
d r op yr a n -2-yl)-3-oxoh exa n oyl]-2-oxa zolid in on e (16). To
a solution of 1517 (0.218 g, 0.755 mmol) in CH2Cl2 (1 mL) at
-5 °C was added titanium tetrachloride (0.150 g, 0.790 mmol,
0.087 mL) dropwise by syringe. After 5 min, diisopropyleth-
ylamine (0.113 g, 0.871 mmol, 0.152 mL) was added dropwise
to the yellow solution, resulting in a deep red color. The
solution was allowed to stir at -5 °C for 1 h and then cooled
to -78 °C. A solution of aldehyde 4 (0.100 g, 0.581 mmol) in
CH2Cl2 (1 mL) was added dropwise by cannula and rinsed over
with an additional 0.5 mL of CH2Cl2. After a period of 2.5 h,
the solution was warmed to -40 °C and allowed to stir for an
additional 1.5 h. The solution was allowed to warm to 0 °C
and was quenched with pH 7 phosphate buffer. The mixture
was diluted with CH2Cl2 (2 mL) and H2O (2 mL). The organic
layer was separated and the aqueous layer extracted three
times with CH2Cl2 (5 mL). The organic layers were combined,
washed with saturated aqueous NaHCO3 solution (10 mL) and
brine (10 mL), dried over MgSO4, filtered, and concentrated.
The resulting oil was purified by chromatography over a 2 ×
17 cm silica gel column (slurry packed with 30% ethyl acetate/
hexanes), eluting with a solvent gradient of 30%-50% ethyl
acetate/hexanes. The product-containing fractions were com-
bined and concentrated to yield 0.231 g (86%) of 16 as a foam:
[R]D +54.3° (c 0.910, CHCl3); Rf ) 0.17 (50% ethyl acetate/
P r ep a r a tion of 18. To a stirring solution of 17 (0.157 g,
0.339 mmol) and PhCH(OMe)2 (0.254 mL, 0.258 g, 1.69 mmol)
in N,N-dimethylformamide (7 mL) was added HBF4 (0.462 mL,
3.39 mmol, 54% solution in Et2O). After 3 d, the solution was
quenched with a saturated aqueous NaHCO3 solution (10 mL)
and diluted with H2O (10 mL) and Et2O (10 mL), and the
aqueous layer extracted three times with Et2O (25 mL). The
combined organic layers were dried over MgSO4, filtered
through Celite, concentrated, and purified using radial chro-
matography (2 mm plate, 30% ethyl acetate/hexanes). The
product-containing fractions were combined and concentrated
to yield 0.187 g (78%) of 18 as a colorless oil: Rf ) 0.52 (50%
1
ethyl acetate/hexanes); 300-MHz H NMR (CDCl3) δ 7.4-7.2
(m, 10H), 5.86 (s, 1H), 5.00 (m, 1H), 4.61 (m, 1H), 4.24 (d, J )
11.0 Hz, 1H), 4.2-4.1 (m, 2H), 4.12 (dd, J ) 2.3, 11.0 Hz, 1H),
3.97 (dd, J ) 7.8, 7.8 Hz, 1H), 3.80 (m, 1H), 3.55 (m, 1H), 3.34
(s, 3H), 3.28 (dd, J ) 3.2, 13.3 Hz, 1H), 2.77 (dd, J ) 9.8, 13.3
Hz, 1H), 2.18 (ddd, J ) 5.8, 11.2, 14.3 Hz, 1H), 2.0-1.8 (m,
3H), 1.64 (ddd, J ) 5.6, 10.3, 12.9 Hz, 1H), 1.54 (ddd, J ) 4.9,
11.0, 13.9 Hz, 1H), 1.31 (d, J ) 7.1 Hz, 3H), 1.3-1.2 (buried
m, 1H), 1.25 (d, J ) 6.5 Hz, 3H), 1.24 (d, J ) 5.9 Hz, 3H);
75-MHz 13C NMR (CDCl3) δ 175.6, 153.5, 138.7, 135.1, 129.3-
(2), 128.9(3), 128.2(2), 127.3, 126.1(2), 96.7, 82.9, 73.1, 71.8,
67.7, 66.2, 65.3, 55.7, 55.4, 38.4, 37.8, 36.0, 34.8, 33.9, 28.6,
21.7, 14.5, 13.1. Anal. Calcd for C32H41NO7: C, 69.67; H, 7.49;
N, 2.54. Found: C, 69.40; H, 7.41; N, 2.62.
1
hexanes); 300-MHz H NMR (CDCl3) δ 7.3-7.1 (m, 5H), 4.92
(q, J ) 7.3 Hz, 1H), 4.73 (m, 1H), 4.3-4.1 (m, 4H), 3.87 (m,
1H), 3.53 (m, 1H), 3.54 (d, J ) 2.1 Hz, 1H), 3.31 (s, 3H), 3.27
(dd, J ) 3.2, 13.4 Hz, 1H), 2.93 (dq, J ) 4.0, 7.1 Hz, 1H), 2.77
(dd, J ) 9.5, 13.4 Hz, 1H), 2.0-1.9 (m, 2H), 1.78 (m, 1H), 1.61
(ddd, J ) 5.2, 9.5, 13.1 Hz, 1H), 1.47 (d, J ) 7.3 Hz, 3H), 1.44
(m, 1H), 1.22 (m, 1H), 1.21 (d, J ) 6.2 Hz, 3H), 1.12 (d, J )
7.1 Hz, 3H); 75-MHz 13C NMR (CDCl3) δ 210.1, 170.4, 153.6,
134.9, 129.2(2), 128.8(2), 127.2, 72.7, 71.1, 70.7, 66.3, 65.6, 55.3,
55.2, 51.7, 48.5, 37.7, 37.6, 35.0(2), 21.3, 13.0, 10.5; 75-MHz
DEPT NMR (CDCl3) CH3 δ 55.3, 21.3, 13.0, 10.5; CH2 δ 66.3,
37.7, 37.6, 35.0(2); CH δ 72.7, 71.1, 70.7, 65.6, 55.2, 51.7, 48.5;
IR (CHCl3) 3555 (br), 1780, 1715, 1700 cm-1. Anal. Calcd for
P r ep a r a tion of 20. A stirring solution of 18 (0.054 g, 0.098
mmol), barium carbonate (0.435 g, 0.220 mmol), and N-
bromosuccinimide (0.020 g, 0.113 mmol) in CH2Cl2 (1 mL) was
heated to reflux. After 2.5 h the solution was cooled to room
temperature, diluted with CH2Cl2 (5 mL) and a saturated
aqueous NaHCO3 solution (5 mL), and filtered. The aqueous
layer was extracted three times with CH2Cl2 (10 mL), and the
combined organic layers were dried over MgSO4, filtered
through Celite, and concentrated to yield 0.059 g (95%) of a
slightly yellow foam which was used without further purifica-
tion: Rf ) 0.29 (35% ethyl acetate/hexanes); 300-MHz 1H NMR
C
25H35NO7: C, 65.06; H, 7.64; N, 3.03. Found: C, 64.91; H,
7.62; N, 3.05.
P r epar ation of (2R,3S,4S,5R)-2,4-Dim eth yl-6-((2S,4R,6S)-
4-(m eth oxy)-6-m eth yltetr a h yd r op yr a n -2-yl)h exa n e-1,3,5-
tr iol (26). Tetramethylammonium triacetoxyborohydride (8.98
g, 34.1 mmol) was added to acetic acid (34 mL), and the
resulting solution was stirred for 1 h. Ketone 16 (1.05 g, 2.27

4490 J . Org. Chem., Vol. 64, No. 12, 1999
Keck and Lundquist
(CDCl3) δ 8.0-7.9 (m, 2H), 7.64 (m, 1H), 7.5-7.4 (m, 2H), 7.1-
7.0 (m, 5H), 6.71 (m, 1H), 5.18 (m, 1H), 5.06 (m, 1H), 4.1-4.0
(buried m, 1H), 4.09 (dd, J ) 10.0, 10.0 Hz, 1H), 3.89 (m, 1H),
3.63 (dd, J ) 5.6, 10.2 Hz, 1H), 3.6-3.4 (m, 2H), 3.31 (s, 3H),
3.11 (m, 1H), 2.72 (dq, J ) 2.4, 7.4 Hz, 1H), 2.31 (ddd, J )
4.9, 12.3, 13.9 Hz, 1H), 2.20 (ddd, J ) 1.5, 6.7, 10.2 Hz, 1H),
2.01 (m, 1H), 1.79 (m, 1H), 1.7-1.5 (m, 2H), 1.24 (d, J ) 6.8
Hz, 3H), 1.19 (d, J ) 6.2 Hz, 3H), 1.19 (buried d, 3H); 75-MHz
13C NMR (CDCl3) δ 172.2, 165.4, 150.6, 136.6, 133.4, 129.6,
129.5(2), 128.8(2), 128.4(2), 128.2(2), 126.7, 78.3, 72.7, 69.6,
68.4, 65.2, 55.3, 38.4, 38.0, 36.0, 34.7, 33.8, 32.2, 21.7, 9.7, 9.2.
A deoxygenated solution of this yellow foam (0.135 g, 0.214
mmol), tributyltin hydride (0.115 mL, 0.125 g, 0.428 mmol),
and AIBN (0.007 g, 0.043 mmol) in benzene (2 mL) was heated
to reflux. After 2 h, this solution was cooled to room temper-
ature, concentrated, and purified using radial chromatography
(2 mm plate, 30% ethyl acetate/hexanes). The product-contain-
ing fractions were combined and concentrated to yield 0.114
g (97%) of 20 as a glass: Rf ) 0.28 (35% ethyl acetate/hexanes);
300-MHz 1H NMR (CDCl3) δ 8.0-7.9 (m, 2H), 7.65 (m, 1H),
7.6-7.5 (m, 2H), 7.1-7.0 (m, 5H), 6.76 (m, 1H), 5.07 (ddd, J )
1.7, 5.1, 9.7 Hz, 1H), 4.95 (m, 1H), 4.11 (m, 1H), 3.88 (m, 1H),
3.51 (m, 1H), 3.32 (s, 3H), 3.18 (dd, J ) 10.7, 13.7 Hz, 1H),
2.96 (dd, J ) 6.4, 13.7 Hz, 1H), 2.67 (dq, J ) 2.4, 7.4 Hz, 1H),
2.31 (ddd, J ) 4.9, 12.2, 13.9 Hz, 1H), 2.16 (m, 1H), 2.02 (m,
1H), 1.80 (m, 1H), 1.7-1.5 (m, 3H), 1.47 (d, J ) 6.8 Hz, 3H),
1.21 (d, J ) 6.6 Hz, 3H), 1.96 (d, J ) 6.1 Hz, 3H), 1.15 (d, J )
7.4 Hz, 3H); 75-MHz 13C NMR (CDCl3) δ 172.5, 165.3, 150.5,
138.2, 133.4, 129.6, 129.5(2), 128.8(2), 128.4(2), 128.1(2), 126.2,
78.0, 72.7, 69.6, 68.4, 65.2, 55.3, 50.2, 38.6, 38.5, 38.0, 34.7,
33.8, 32.3, 21.7, 18.1, 9.6, 9.1.
P r epar ation of (2R,3S,4S,5R)-2,4-Dim eth yl-1,3-(p-m eth -
oxyben zylid en ed ioxy)-6-((2S,4R,6S)-4-m eth oxy-6-m eth -
yltetr a h yd r op yr a n -2-yl)h exa n -5-ol (30). To a stirring so-
lution of 26 (0.270 g, 0.930 mmol) and 4-methoxybenzyl meth-
yl ether (0.566 g, 3.72 mmol) in CH2Cl2 (9.5 mL) was added
2,3-dichloro-5,6-dicyano-1,4-benzoquinone (0.464 g, 2.05
mmol). After being stirred for 40 min, the cloudy brown
solution was filtered through Celite, concentrated, and diluted
with CH2Cl2 (20 mL). The organic layer was washed with a
5% aqueous NaHCO3 solution (30 mL) and the aqueous layer
extracted two times with CH2Cl2 (30 mL). The combined
organic layers were dried over MgSO4, filtered through Celite,
and concentrated to yield a slightly yellow solid. This yellow
solid was recrystallized from 10% ethyl acetate/hexanes (10
mL) to yield 0.234 g of 30 (64%) as colorless needles (mp )
111 °C). The mother liquor was purified by chromatography
over a 3 × 14 cm silica gel column (slurry packed with 10%
acetone/hexanes) eluting with 200 mL each of 10% and 25%
acetone/hexanes and 100 mL of 50% acetone/hexanes, collected
in 9 mL fractions. The product-containing fractions (31-38)
were combined and concentrated to yield an additional 0.062
g (16%) of a white solid: [R]D -17.5° (c 0.800, CHCl3); Rf )
0.56 (50% acetone/hexanes); 300-MHz 1H NMR (CDCl3) δ 7.4-
7.3 (m, 2H), 6.9-6.8 (m, 2H), 5.40 (s, 1H), 4.22 (dd, J ) 6.5,
8.0 Hz, 1H), 4.15 (dd, J ) 4.6, 11.2 Hz, 1H), 4.12 (m, 1H), 3.80
(m, 1H), 3.80 (s, 3H), 3.6-3.5 (m, 2H), 3.51 (dd, J ) 11.2, 11.2
Hz, 1H), 3.35 (s, 3H), 3.23 (br s, 1H), 2.27 (m, 1H), 2.13 (ddd,
J ) 5.9, 11.2, 14.2 Hz, 1H), 2.1-2.0 (m, 2H), 1.85 (m, 1H),
1.62 (ddd, J ) 5.6, 10.4, 12.8 Hz, 1H), 1.43 (ddd, J ) 4.6, 8.8,
13.8 Hz, 1H), 1.20 (d, J ) 6.1 Hz, 3H), 1.19 (ddd, J ) 10.3,
10.3, 12.5 Hz, 1H), 1.10 (d, J ) 7.1 Hz, 3H), 0.82 (d, J ) 6.6
Hz, 3H); 75-MHz 13C NMR (CDCl3) δ 159.9, 130.7, 127.1(2),
113.6(2), 101.9, 89.2, 73.1, 73.0, 69.1, 66.9, 64.9, 55.3, 55.2,
38.6, 35.3, 34.8, 34.4, 31.0, 21.7, 12.1, 10.6; 75-MHz DEPT
NMR (CDCl3) CH3 δ 55.3, 55.2, 21.7, 12.1, 10.6; CH2 δ 73.1,
38.6, 35.3, 34.8; CH δ 101.9, 89.2, 73.0, 69.1, 66.9, 64.9, 34.4,
31.0; IR (CHCl3) 3470 (br) cm-1. Anal. Calcd for C23H36O6: C,
67.62; H, 8.88. Found: C, 67.47; H, 8.95.
collected in 9 mL fractions. The product-containing fractions
(19-31) were concentrated to yield 0.082 g (86%) of 32 as a
foam: [R]D -52.3° (c 0.555, CHCl3); Rf ) 0.38 (50% acetone/
1
hexanes); 300-MHz H NMR (CDCl3) δ 8.17 (s, 1H), 7.39 (dd,
J ) 1.5, 1.5 Hz, 1H), 7.1-7.0 (m, 2H), 6.82 (dd, J ) 0.9, 1.5
Hz, 1H), 6.7-6.6 (m, 2H), 6.10 (br dd, J ) 4.8, 10.2 Hz, 1H),
5.28 (s, 1H), 4.20 (m, 1H), 4.10 (dd, J ) 4.5, 11.4 Hz, 1H), 4.05
(m, 1H), 3.76 (s, 3H), 3.53 (m, 1H), 3.44 (dd, J ) 11.1, 11.1
Hz, 1H), 3.5-3.4 (buried m, 1H), 3.34 (s, 3H), 2.48 (ddd, J )
4.9, 12.4, 13.9 Hz, 1H), 2.38 (m, 1H), 2.23 (m, 1H), 2.06 (m,
1H), 1.85 (m, 1H), 1.7-1.6 (m, 2H), 1.3-1.1 (buried m, 1H),
1.28 (d, J ) 7.1 Hz, 3H), 1.23 (d, J ) 6.4 Hz, 3H), 0.83 (d, J )
6.6 Hz, 3H); 75-MHz 13C NMR (CDCl3) δ 183.7, 159.6, 136.8,
130.8, 130.0, 126.8(2), 117.7, 113.2(2), 102.1, 86.8, 79.7, 72.9,
72.9, 69.0, 65.4, 55.3, 55.2, 38.8, 34.7, 33.8, 32.3, 31.1, 21.5,
12.0, 11.9; 75-MHz DEPT NMR (CDCl3) CH3 δ 55.3, 55.2, 21.5,
12.0, 11.9; CH2 δ 72.9, 38.8, 34.7, 32.3; CH δ 102.1, 86.8, 79.7,
72.9, 69.0, 65.4, 33.8, 31.1; Anal. Calcd for C27H38N2O6S: C,
62.52; H, 7.38; N, 5.40; S, 6.18. Found: C, 62.39; H, 7.43; N,
5.29; S, 6.10.
P r ep a r a tion of (2R,3S,4S)-2,4-Dim eth yl-1,3-(p-m eth -
oxyben zylid en ed ioxy)-6-((2S,4R,6S)-4-m eth oxy-6-m eth -
yltetr a h yd r op yr a n -2-yl)h exa n e (33). To a deoxygenated
solution of 32 (0.070 g, 0.135 mmol) in toluene (2 mL) at reflux
was added tributyltin hydride (0.363 mL, 0.393 g, 1.35 mmol)
via syringe. After 30 min the solution was cooled to room
temperature, concentrated, purified by chromatography over
a 2.5 × 13 cm silica gel column (slurry packed with 5% ethyl
acetate/hexanes), eluting with 100 mL each of hexanes and
15%, 15%, 25%, and 25% ethyl acetate/hexanes, and collected
in 8 mL fractions. The product-containing fractions (42-49)
were collected and concentrated to yield 0.053 g (100%) of 33
as a colorless oil: [R]D -34.8° (c 2.48, CHCl3); Rf ) 0.34 (35%
1
ethyl acetate/hexanes); 300-MHz H NMR (CDCl3) δ 7.4-7.3
(m, 2H), 6.9-6.8 (m, 2H), 5.42 (s, 1H), 4.09 (dd, J ) 4.8, 11.2
Hz, 1H), 4.01 (m, 1H), 3.80 (s, 3H), 3.71 (dddd, J ) 2.7, 6.3,
12.7, 16.1 Hz, 1H), 3.52 (m, 1H), 3.48 (dd, J ) 11.2, 11.2 Hz,
1H), 3.35 (dd, J ) 2.0, 9.8 Hz, 1H), 3.34 (s, 3H), 2.09 (m, 1H),
2.0-1.8 (m, 4H), 1.60 (ddd, J ) 5.6, 10.3, 12.7 Hz, 1H), 1.6-
1.2 (m, 3H), 1.2-1.1 (buried m, 1H), 1.21 (d, J ) 6.4 Hz, 3H),
1.06 (d, J ) 6.8 Hz, 3H), 0.78 (d, J ) 6.6 Hz, 3H); 75-MHz 13C
NMR (CDCl3) δ 159.6, 131.6, 127.2(2), 113.4(2), 101.1, 87.7,
73.2, 73.1, 71.5, 64.5, 55.2(2), 38.6, 34.9, 32.8, 30.6, 29.1, 25.5,
21.7, 16.9, 12.2; 75-MHz DEPT NMR CH3 δ 55.2(2), 21.7, 16.9,
12.2; CH2 δ 73.1, 38.6, 34.9, 29.1, 25.5; CH δ 101.1, 87.7, 73.2,
71.5, 64.5, 32.8, 30.6; Anal. Calcd for C23H36O5: C, 70.38; H,
9.24. Found: C, 70.22; H, 9.24.
P r ep a r a tion of (2R,3S,4S)-2,4-Dim eth yl-3-(p-m eth oxy-
ben zyloxy)-6-((2S,4R,6S)-4-m eth oxy-6-m eth yltetr a h yd r o-
p yr a n -2-yl)h exa n ol (34). To a stirring solution of 33 (0.041
g, 0.104 mmol) in CH2Cl2 (1.5 mL) at 0 °C was added DIBAL
(0.210 mL, 0.313 mmol, 1.5 M solution in toluene). After 1.5
h, a saturated solution of Rochelle salts (3 mL) was added and
the solution stirred for 5 h. The mixture was diluted with
CH2Cl2 (5 mL) and H2O (5 mL) and the aqueous layer
extracted three times with CH2Cl2 (10 mL). The combined
organic layers were dried over MgSO4, filtered through Celite,
and concentrated. The resulting oil was purified by chroma-
tography over a 2.5 × 13 cm silica gel column (slurry packed
with 20% acetone/hexanes) eluting with 20% acetone/hexanes
and collected in 9 mL fractions. The product-containing
fractions (11-19) were combined and concentrated to yield
0.038 g (91%) of 34 as a colorless oil: [R]D -18.9° (c 1.48,
1
CHCl3); Rf ) 0.51 (50% acetone/hexanes); 300-MHz H NMR
(CDCl3) δ 7.3-7.2 (m, 2H), 6.9-6.8 (m, 2H), 4.58 (d, J AB
)
10.5 Hz, 1H), 4.50 (d, J AB ) 10.5 Hz, 1H), 3.99 (m, 1H), 3.80
(s, 3H), 3.71 (dd, J ) 3.4, 11.0 Hz, 1H), 3.70 (m, 1H), 3.59 (dd,
J ) 5.7, 11.0 Hz, 1H), 3.51 (m, 1H), 3.33 (s, 3H), 3.22 (dd, J )
5.2, 6.1 Hz, 1H), 2.84 (br s, 1H), 2.0-1.8 (m, 5H), 1.59 (ddd, J
) 5.3, 10.0, 12.8 Hz, 1H), 1.59 (m, 1H), 1.4-1.1 (m, 3H), 1.21
(d, J ) 6.2 Hz, 3H), 1.02 (d, J ) 7.1 Hz, 3H), 1.00 (d, J ) 6.8
Hz, 3H); 75-MHz 13C NMR (CDCl3) δ 154.2, 130.4, 129.3(2),
113.8(2), 89.4, 74.8, 73.2, 71.8, 66.1, 64.8, 55.3, 55.2, 38.5, 36.9,
35.7, 34.8, 29.7, 28.1, 21.7, 16.9, 15.8; 75-MHz DEPT NMR
(CDCl3) CH3 δ 55.3, 55.2, 21.7, 16.9, 15.8; CH2 δ 74.8, 66.1,
P r ep a r a tion of 32. A stirring solution of 30 (0.075 g, 0.184
mmol) and 1,1′-thiocarbonyldiimidazole (0.065 g, 0.367 mmol)
in THF was heated at reflux for 40 h. The solution was cooled
to room temperature, concentrated, purified by chromatogra-
phy over a 2.5 × 14 cm silica gel column (slurry packed with
20% acetone/hexanes) eluting with 20% acetone/hexanes, and

Total Synthesis of Swinholide
J . Org. Chem., Vol. 64, No. 12, 1999 4491
38.5, 34.8, 29.7, 28.1; CH δ 89.4, 73.2, 71.8, 64.7, 36.9, 35.6;
IR (neat) 3380 (br) cm-1. Anal. Calcd for C23H38O5: C, 70.02;
H, 9.71. Found: C, 69.88; H, 9.63.
P r ep a r a tion of 41. To a stirring solution of 35 (0.023 g,
0.059 mmol) and 4028 (0.023 g, 0.088 mmol) in CH2Cl2 (1.2 mL)
at -78 °C was added BF3‚OEt2 (0.014 mL, 0.017 g, 0.117 mmol)
dropwise via syringe. After 45 min the solution was quenched
with pH 7 phosphate buffer (0.5 mL), warmed to room
temperature, and diluted with CH2Cl2 (5 mL) and H2O (5 mL).
The aqueous layer was extracted three times with CH2Cl2 (10
mL), and the combined organic layers were dried over MgSO4,
filtered through Celite, and concentrated. The resulting oil was
purified by chromatography over a 2.5 × 13 cm silica gel
column (slurry packed with 20% acetone/hexanes) eluting with
20% acetone/hexanes and collected in 7 mL fractions. The
product-containing fractions (9-13) were combined and con-
centrated to yield 0.029 g (92%) of 44 as a colorless oil and a
5:1 ratio of inseparable diastereomers. Major diastereomer: Rf
) 0.52 (35% acetone/hexanes); 300-MHz 1H NMR (CDCl3) δ
7.3-7.2 (m, 2H), 6.9-6.8 (m, 2H), 4.54 (s, 2H), 4.20 (d, J )
9.8 Hz, 1H), 4.0-3.9 (m, 1H), 3.79 (s, 3H), 3.7-3.6 (m, 1H),
3.6-3.5 (m, 1H), 3.58 (br s, 1H), 3.34 (s, 3H), 3.24 (dd, J )
2.9, 8.3 Hz, 1H), 2.66 (m, 1H), 2.0-1.5 (m, 8H), 1.47 (s, 9H),
1.4-1.1 (buried m, 2H), 1.21 (d, J ) 6.1 Hz, 3H), 1.06 (d, J )
7.1 Hz, 3H), 1.03 (d, J ) 7.1 Hz, 3H), 0.90 (d, J ) 6.6 Hz, 3H);
75-MHz 13C NMR (CDCl3) δ 204.0, 159.2, 130.0, 129.4(2),
113.7(2), 90.0, 75.9, 73.1, 72.3, 72.2, 64.8, 55.3, 55.2, 55.2, 47.9,
38.6, 35.8, 34.6, 34.4, 29.7(4), 29.2, 21.8, 16.3, 14.5, 11.3; 75-
MHz DEPT NMR (CDCl3) CH3 δ 55.3, 55.2, 29.7(3), 21.8, 16.3,
14.5, 11.3; CH2 δ 75.9, 38.6, 34.6, 29.7, 29.2; CH δ 90.0, 73.1,
72.3, 72.2, 64.8, 52.2, 35.8, 34.4.
P r ep a r a tion of (2R,3S,4S)-2,4-Dim eth yl-3-(p-m eth oxy-
ben zyloxy)-6-((2S,4R,6S)-4-m eth oxy-6-m eth yltetr a h yd r o-
p yr a n -2-yl)h exa n a l (35). To a stirring solution of 34 (0.190
g, 0.482 mmol) in CH2Cl2 (10 mL) was added the Dess-Martin
periodinane (2.00 g, 4.82 mmol). After 45 min the solution was
diluted with Et2O (25 mL) and poured into a 1:1 mixture of
saturated aqueous NaHCO3/Na2S2O3 (50 mL). This solution
was stirred for 20 min and the aqueous layer extracted with
Et2O (3 × 50 mL). The combined organic layers were washed
with saturated aqueous NaHCO3 solution (100 mL) and brine
(100 mL) and dried over MgSO4. The solution was filtered
through Celite and concentrated. The resulting oil was purified
by chromatography over a 2.5 × 13 cm silica gel column (slurry
packed with 15% acetone/hexanes) eluting with 15% acetone/
hexanes and collected in 9 mL fractions. The product-contain-
ing fractions (12-17) were combined and concentrated to yield
0.176 g (93%) of 38 as a colorless oil: [R]D -5.8° (c 1.54, CHCl3);
1
Rf ) 0.33 (20% acetone/hexanes); 300-MHz H NMR (CDCl3)
δ 9.78 (d, J ) 2.4 Hz, 1H), 7.3-7.2 (m, 2H), 6.9-6.8 (m, 2H),
4.53 (d, J AB ) 10.9 Hz, 1H), 4.47 (d, J AB ) 10.9 Hz, 1H), 3.98
(m, 1H), 3.80 (s, 3H), 3.68 (dddd, J ) 2.9, 6.2, 12.6, 15.6 Hz,
1H), 3.50 (m, 1H), 3.48 (dd, J ) 5.4, 5.4 Hz, 1H), 3.33 (s, 3H),
2.70 (m, 1H), 2.0-1.8 (m, 4H), 1.59 (ddd, J ) 5.6, 10.3, 12.9
Hz, 1H), 1.56 (m, 1H), 1.4-1.2 (m, 2H), 1.21 (d, J ) 6.1 Hz,
3H), 1.13 (d, J ) 7.1 Hz, 3H), 0.94 (d, J ) 6.8 Hz, 3H); 75-
MHz 13C NMR (CDCl3) δ 204.7, 159.2, 130.3, 129.2(2), 113.7-
(2), 85.0, 73.4, 73.2, 71.7, 64.7, 55.3, 55.2, 48.5, 38.5, 35.2, 34.8,
29.4, 28.3, 21.7, 16.1, 11.5; 75-MHz DEPT NMR (CDCl3) CH3
δ 55.3, 55.2, 21.7, 16.1, 11.5; CH2 δ 73.4, 38.5, 34.8, 29.4, 28.3;
CH δ 85.0, 73.2, 71.7, 64.7, 48.5, 35.2; IR (neat) 2940, 2870,
2725, 1720 cm-1. Anal. Calcd for C23H36O5: C, 70.38; H, 9.24.
Found: C, 69.83; H, 9.21.
Ack n ow led gm en t. Financial support for this re-
search provided by the National Institutes of Health
(through Grant GM28961) and by Pfizer Inc. is grate-
fully acknowledeged.
J O990291Y