Studies on the synthesis of nargenicin A1: Highly stereoselective synthesis of the complete carbon framework via the transannular Diels-Alder reaction of an 18-membered macrolide

Article abstract of DOI:10.1021/ja9607796

A synthesis of the complete carbon skeleton of the nargenicins, represented by tricyclic lactone 45, is described. The key step of the synthesis of 45 is the Yamaguchi macrolactonization of hydroxy acid 44 which is followed by the facile transannular Diels-Alder reaction of the 18-membered macrolide 22. This sequence provides tricycle 45 in 66% yield, along with a 14% yield of tricycle 46 which is epimeric at C(10). Macrolide 22 was obtained in 38% yield when the macrolactonization was performed at 80°C. The transannular Diels-Alder reaction of 22 at 80°C provided tricycle 45 as the exclusive product (85% yield). In contrast, the intramolecular Diels-Alder reaction of seco ester 43 provided a mixture of trans-fused 47 in 56% yield and the desired cis-fused cycloadduct 48 in only 27% yield. Two independent stereochemical control features determine the success of the transannular Diels-Alder reaction of 22: the C(6)-Br steric directing group that dictates that only one of the two faces of the diene is accessible to the dienophile in transition state 14 and allylic strain considerations involving the C(16)-Me substituent which enable only one face of the dienophile to be accessible to the diene in transition state 14. The latter effect is operational only in the transannular cycloaddition mode as indicated by the results with 43. Az added benefit of this strategy is that the 10-membered lactone is established by a formal ring contraction of the more easily synthesized 18-membered lactone. Attempts to extend this strategy to the transannular Diels-Alder reaction of the C(13)-hydroxyl substituted macrolide 13 have not been successful.

Full text of DOI:10.1021/ja9607796

+
+
7502
J. Am. Chem. Soc. 1996, 118, 7502-7512
Studies on the Synthesis of Nargenicin A1: Highly
Stereoselective Synthesis of the Complete Carbon Framework
via the Transannular Diels-Alder Reaction of an 18-Membered
Macrolide
William R. Roush,* Kazuo Koyama, Michael L. Curtin, and Kevin J. Moriarty
Contribution from the Department of Chemistry, Indiana UniVersity, Bloomington, Indiana 47405
ReceiVed March 11, 1996X
Abstract: A synthesis of the complete carbon skeleton of the nargenicins, represented by tricyclic lactone 45, is
described. The key step of the synthesis of 45 is the Yamaguchi macrolactonization of hydroxy acid 44 which is
followed by the facile transannular Diels-Alder reaction of the 18-membered macrolide 22. This sequence provides
tricycle 45 in 66% yield, along with a 14% yield of tricycle 46 which is epimeric at C(10). Macrolide 22 was
obtained in 38% yield when the macrolactonization was performed at 80 °C. The transannular Diels-Alder reaction
of 22 at 80 °C provided tricycle 45 as the exclusive product (85% yield). In contrast, the intramolecular Diels-
Alder reaction of seco ester 43 provided a mixture of trans-fused 47 in 56% yield and the desired cis-fused cycloadduct
48 in only 27% yield. Two independent stereochemical control features determine the success of the transannular
Diels-Alder reaction of 22: the C(6)-Br steric directing group that dictates that only one of the two faces of the
diene is accessible to the dienophile in transition state 14 and allylic strain considerations involving the C(16)-Me
substituent which enable only one face of the dienophile to be accessible to the diene in transition state 14. The
latter effect is operational only in the transannular cycloaddition mode as indicated by the results with 43. An added
benefit of this strategy is that the 10-membered lactone is established by a formal ring contraction of the more easily
synthesized 18-membered lactone. Attempts to extend this strategy to the transannular Diels-Alder reaction of the
C(13)-hydroxyl substituted macrolide 13 have not been successful.
Nargenicin A1 (1) and nodusmicin (2) are structurally novel
the intramolecular Diels-Alder reaction of 4 was exceptionally
diastereoselective, the stereochemistry of the sole cycloadduct
5 was not suitable for its use in the projected total synthesis.
Moreover, the stereoselectivity of the intramolecular Diels-
Alder reactions of all other substrates examined, including 6,
was sufficiently poor that this conventional intramolecular
Diels-Alder strategy for the synthesis of the nargenicins was
abandoned.12 An alternative approach to the nargenicins
involving an intramolecular Diels-Alder reaction has also
proven unsuccessful.6
antibiotics isolated from Nocardia argentineneas and Saccha-
ropolyspora hirsuta, respectively.1-3 The stereostructure of
nodusmicin was established by X-ray crystallographic studies,2
while that of nargenicin A1 was confirmed by its synthesis from
nodusmicin.3 The absolute configuration of the nargenicins,
originally assigned by Cane by the nonempirical CD exciton
method,4 was subsequently verified via Kallmerten’s enantio-
selective total synthesis of (+)-18-deoxynargenicin A1 (3).5
Several other approaches to the total synthesis of the nargenicins
have been reported.6,7
We have previously described an approach to the decalin
nucleus of 1 based on the intramolecular Diels-Alder
reactions8-10 of decatrienones 4 and 6.11 Unfortunately, while
X Abstract published in AdVance ACS Abstracts, July 15, 1996.
(1) Celmer, W. D.; Chmurny, G. N.; Moppett, C. E.; Ware, R. S.; Watts,
P. C.; Whipple, E. B. J. Am. Chem. Soc. 1980, 102, 4203.
(2) Whaley, H. A.; Chidester, C. G.; Mizsak, S. A.; Wnuk, R. J.
Tetrahedron Lett. 1980, 21, 3659.
(3) Magerlein, B. J.; Mizsak, S. A. J. Antibiot. 1982, 35, 111.
(4) Cane, D. E.; Yang, C.-C. J. Antibiot. 1985, 38, 423.
(5) Plata, D. J.; Kallmerten, J. J. Am. Chem. Soc. 1988, 110, 4041.
(6) Jones, R. C. F.; Tunnicliffe, J. H. Tetrahedron Lett. 1985, 26, 5845.
(7) Go¨ssinger, E.; Graupe, M.; Zimmermann, K. Monatsh. Chem. 1993,
124, 965.
(8) Ciganek, B. Org. React. 1984, 32, 1.
(9) Craig, D. Chem. Soc. ReV. 1987, 16, 187.
(10) Roush, W. R. In ComprehensiVe Organic Synthesis; Paquette, L.
A., Ed.; Pergamon Press: Oxford, 1991; Vol. 5, pp 513-550.
(11) Coe, J. W.; Roush, W. R. J. Org. Chem. 1989, 54, 915.
(12) Coe, J. W. Ph. D. Thesis, Massachusetts Institute of Technology,
1988.
S0002-7863(96)00779-2 CCC: $12 00 © 1996 American Chemical Society

+
+
Studies on the Synthesis of Nargenicin A1
J. Am. Chem. Soc., Vol. 118, No. 32, 1996 7503
Cane has suggested that the nargenicin biosynthesis involves
an intramolecular Diels-Alder reaction.4,13,14 Feeding experi-
ments with intact C(11)-C(19) and C(9)-C(19) acyclic frag-
ments establish that the nargenicin carbon framework derives
from the well established polypropionate-polyacetate biosyn-
thetic pathway common to macrolide antibiotics.13,14 Other
studies have established that the oxygen atoms of the C(2)
methoxy substituent, the C(8)-C(13) oxa bridge, and the C(18)
hydroxyl group are derived from O2, as determined by 18O2
labeling studies.4 This evidence supports the notion that 11 may
be a late stage biosynthetic intermediate, the stereostructure of
which logically implicates the acyclic tetraene 10 as a plausible
precursor.4,13,14
substrate must have additional stereochemical control elements
not present in 4 and 6 that serve to control the diastereoselec-
tivity as well as functionality that causes the Diels-Alder
reaction to occur rapidly under growth conditions of the
microorganism. One final consideration that led to the approach
described in this paper is that the “normal” products of the
polypropionate biosynthetic pathway are macrocyclic lactones.
Hence, we considered the possibility that the Diels-Alder
reaction could be performed in the transannular mode,20 with
the conformational preferences of the 18-membered lactone
serving to dictate the diastereoselectivity of the cycloaddition
step.21-24 We thus selected macrocyclic tetraenes 12 and 13
as targets for study.
Analysis of molecular models of 12 suggests that transition
state 14 should be highly favored over the three other possibili-
ties. We assume that the C(7)-C(12) unit will adopt the boat-
like conformation indicated in 14, which we have shown to be
highly favored in the intramolecular Diels-Alder reactions of
substituted 1,7,9-undecatrien-3-ones.11 Furthermore, we have
designed a C(6)-bromine steric directing group into 12 and 13,
which should induce the C(4-7) diene to react from a
conformation in which C(6) is syn to C(8)-H, as is the case in
14.25 The alternate conformation of the diene, as in 15, positions
the bulky C(6)-Br substituent syn to the much more sterically
demanding C(8,9)-acetonide unit; C(5)-H also interacts with
C(2)-H in transition state 15, but not in 14. In both transition
states 14 and 15, the C(2)-methoxy group is equatorially
positioned with respect to the macrocycle, the lactone is s-trans,
and the C(16)-C(18) unit adopts staggered conformations with
the C(16)-methyl group in an equatorial position on the
macrocycle. Importantly, the proton at C(16) rather than
C(16)-Me eclipses the C(14)-C(15) double bond, which
minimizes allylic strain at this position.26 Finally, as long as
an s-trans conformation is adopted by the C(13)-C(14) single
bond, then the conformation deduced for the C(1)-C(4) and
C(13)-C(19) segments of the macrocycle are very close to the
conformation of the ten membered ring of the natural product,2
with the exception that the C(4)-C(13) bond that is developing
in 14 is considerably longer than the fully formed bond at this
position in the natural product.
The preceding analysis indicates that transition state 14 should
be highly favored over the diastereomeric arrangement 15 owing
to the interactions of the diene with C(2)-H and the C(8)-
C(9) acetonide in 15. There are two additional transition states
(16 and 17) that should also be considered. These are obtained
simply by reversing the face of the dienophile that is exposed
to the diene in 14 and 15. The consequences of this seemingly
simple conformational change are profound. In order for the
C(4)-C(7) diene and the C(12)-C(13) dienophile to achieve
bonding interactions in these transition states, it is necessary
for the macrocycle to adopt a conformation that has the C(16)-
Me group eclipsing the C(14)-C(15) double bond, as indicated
in 16 and 17. We therefore anticipated from the outset that
transition states 16 and 17 would be noncompetitive with 14
and that allylic strain considerations26 involving the C(16)
Assuming that the biosynthesis does indeed involve an
intramolecular Diels-Alder reaction of the type formulated in
the proposed conversion of 10 to 11, how is it that Nature is
able to control the diastereoselectivity? The obvious answer is
that the biosynthetic reaction is promoted by an enzyme
catalyst.15 A second possibility is that the functionality that
we incorporated into the C(8)-C(14) segment of 4 and 6 is
sufficiently different from that present in the putative biosyn-
thetic Diels-Alder substrate that the reactions of 4, 6, and
related trienes are poor mimics of the biosynthetic conversion.
Although Diels-Alder reactions have been proposed as key
steps in the biosyntheses of several families of natural prod-
ucts,13,15,16 and while monoclonal antibodies have been devel-
oped that catalyze Diels-Alder reactions,17 there are as yet no
fully documented examples of naturally occurring “Diels-
Alderases”. To our knowledge, the closest example is the work
of Oikawa and Ichihara who have demonstrated that a labeled,
achiral triene precursor of the solanapyrones undergoes an
enantioselective intramolecular Diels-Alder reaction when fed
to Alternaria solani.18 More recently, the same group has
demonstrated that the stereoselectivity of the intramolecular
Diels-Alder reaction performed in the presence of an A. solani
cell-free extract is different from that of the thermal cycload-
dition.19 However, the presumed Diels-Alderase has not yet
been isolated or characterized.
It is irrelevant for our purposes whether the putative nar-
genicin Diels-Alder biosynthetic step is enzyme catalyzed or
not. Nevertheless, the conclusion that we reached is that if this
key biosynthetic step is not enzyme catalyzed, then the triene
(20) Deslongchamps, P. Aldrichimica Acta 1991, 24, 43.
(21) Still, W. C.; Galynker, I. Tetrahedron 1981, 37, 3981.
(22) Still, W. C.; Novack, V. J. J. Am. Chem. Soc. 1984, 106, 1148, and
references cited therein.
(23) Schreiber, S. L.; Sammakia, T.; Hulin, B.; Schulte, G. J. Am. Chem.
Soc. 1986, 108, 2106, and references cited therein.
(24) Porco, J. A., Jr.; Schoenen, F. J.; Stout, T. J.; Clardy, J.; Schreiber,
S. L. J. Am. Chem. Soc. 1990, 112, 7410. This paper describes a facile
macrolactonization and transannular Diels-Alder reaction that proceeds with
reversed regiochemistry compared to a related type II intramolecular Diels-
Alder reaction.
(25) Roush, W. R.; Kageyama, M.; Riva, R.; Brown, B. B.; Warmus, J.
S.; Moriarty, K. J. J. Org. Chem. 1991, 56, 1192.
(13) Cane, D. E.; Tan, W.; Ott, W. R. J. Am. Chem. Soc. 1993, 115,
527.
(14) Cane, D. E.; Luo, G. J. Am. Chem. Soc. 1995, 117, 6633.
(15) Laschat, S. Angew. Chem., Int. Ed. Engl. 1996, 35, 289.
(16) Roush, W. R.; Myers, A. G. J. Org. Chem. 1981, 46, 1509.
(17) Hilvert, D. Acc. Chem. Res. 1993, 26, 552, and references cited
therein.
(18) Oikawa, H.; Suzuki, Y.; Naya, A.; Katayama, K.; Ichihara, A. J.
Am. Chem. Soc. 1994, 116, 3605.
(19) Oikawa, H.; Katayama, K.; Suzuki, Y.; Ichihara, A. J. Chem. Soc.,
Chem. Commun. 1995, 1321.
(26) Hoffmann, R. W. Chem. ReV. 1989, 89, 1841.

+
+
7504 J. Am. Chem. Soc., Vol. 118, No. 32, 1996
Roush et al.
Results and Discussion
Retrosynthetic Analysis of 13-Deoxymacrolide 12. Analy-
sis of the structure of 12 suggested that this intermediate could
be assembled from three precursors: vinylboronic acid 23,
corresponding to C(1)-C(5); methyl ketone 25, corresponding
to C(6)-C(12); and enal 26 corresponding to the C(13)-C(19)
fragment of the macrolide target. We anticipated that the
fragment assembly sequence would involve an aldol condensa-
tion of 25 and 26, Suzuki cross coupling of the dibromoolefin
unit of 25 with vinylboronic acid 2332,33 followed by a final
macrolactonization step. However, at the time that this synthesis
was undertaken, vinylboronic acid 23 was unavailable and the
simpler vinylboronic acid 24 was used instead. Thus, the first
macrolide that we discuss in this paper is 22, and not 12.
Synthesis of Vinylboronic Acids 23 and 24. Vinylboronic
acid 23 was synthesized from benzyl (S)-glycidate (27),34 which
in turn was prepared from D-serine via potassium (S)-glycidate.35
stereocenter would play a major role in determining the
conformational preferences of 12 and 13 in the cycloaddition
transition states. As will be shown subsequently, allylic interac-
tions inVolVing the C(16) stereocenter constitute the additional
stereochemical control element required to control the diaste-
reoselectiVity of this key Diels-Alder reaction.
36,37
Thus, treatment of 27 with Et2AlCtCSiMe3
provided 28
in 83% yield. Conversion of the alcohol to a methyl ether by
using Ag2O and MeI followed by cleavage of the acetylenic
silane under mild conditions provided 29 in 58% yield. Finally,
hydroboration of 29 with catecholborane provided the sensitive
vinylboronic acid 23 in 75% yield.38
Several additional strategic considerations also add to the
attractiveness of this plan. First, one would anticipate that the
transannular Diels-Alder reactions of tetraenes 12 and 13 will
proceed at a much faster rate than the conventional intramo-
lecular Diels-Alder reactions of analogously functionalized
acyclic tetraenes.20,27,28 Second, this approach constitutes a very
simple ring contraction strategy for introducing the ten-
membered lactone of the natural product. Ten membered
lactones, in general, are difficult to prepare efficiently by
conventional lactonization methods.29,30 Kallmerten synthesized
the 10-membered lactone in his 18-deoxynargenicin synthesis
in 38% yield by macrolactonization of the seco acid.5 However,
Steliou was able to cyclize a seco acid prepared by degradation
of natural nargenicin A1 in only 8% yield.31 Consequently, since
18-membered lactones are generally easier to synthesize than
10-membered ones,30 this transannular Diels-Alder strategy
seemed to be an ideal way to control the stereochemistry of the
nargenicin nucleus as well as to introduce the highly problematic
10-membered lactone in a single operation.
Vinylboronic acid 24 was similarly prepared in 61% yield
by hydroboration of benzyl 4-pentynoate (30).
Synthesis of Methyl Ketone 25. Asymmetric (E)-crotylbo-
(32) Miyaura, N.; Yamada, K.; Suginome, H.; Suzuki, A. J. Am. Chem.
Soc. 1985, 107, 972.
(33) Roush, W. R.; Moriarty, K. J.; Brown, B. B. Tetrahedron Lett. 1990,
31, 6509.
(34) Larcheveque, M.; Petit, Y. Tetrahedron Lett. 1987, 28, 1993.
(35) Roush, W. R.; Brown, B. B. J. Org. Chem. 1992, 57, 3380.
(36) Roush, W. R.; Blizzard, T. A.; Basha, F. Z. Tetrahedron Lett. 1982,
23, 2331.
(37) Suzuki, T.; Saimoto, H.; Tomioka, H.; Oshima, K.; Nozaki, H.
Tetrahedron Lett. 1982, 23, 3597.
(27) Be´rube´, G.; Deslongchamps, P. Tetrahedron Lett. 1987, 28, 5255.
(28) Takahashi, T.; Shimizu, K.; Doi, T.; Tsuji, J. J. Am. Chem. Soc.
1988, 110, 2674.
(29) Meng, Q.; Hesse, M. Topics Curr. Chem. 1991, 161, 107-176.
(30) Rousseau, G. Tetrahedron 1995, 51, 2777.
(38) Kabalka, G. W.; Baker, J. D., Jr.; Neal, G. W. J. Org. Chem. 1977,
42, 512.
(31) Steliou, K.; Poupart, M.-A. J. Am. Chem. Soc. 1983, 105, 7130.

+
+
Studies on the Synthesis of Nargenicin A1
J. Am. Chem. Soc., Vol. 118, No. 32, 1996 7505
ration of L-glyceraldehyde pentylidene ketal 3139 with (E)-
crotylboronate (S,S)-32 provided 33 with excellent selectiv-
ity.40,41 Removal of the ketal by heating a methanolic solution
of 33 with Dowex 50 × 4-400 H+ resin followed by selective
silylation of the primary alcohol provided TBS ether 34 in 84%
overall yield from 31. The C(8,9) diol unit was then protected
as an acetonide, and the TBS ether was cleaved to give primary
alcohol 35 in 96% yield. It should be noted that it was not
possible to transform 33 directly into 35 since the acetonide at
the terminal C(7,8) position (as in 33) is more stable than at
the internal C(8,9) position in 35.11 Oxidation of 35 via the
Swern protocol42 and Corey-Fuchs olefination of the resulting
aldehyde provided dibromodiene 36.43 Finally, subjection of
36 to standard Wacker oxidation conditions provided methyl
ketone 25 in 65-70% yield.44
Synthesis and Transannular Diels-Alder Reaction of 13-
Deoxymacrolide 22. We began the synthesis of macrolide 22
with the aldol coupling of methyl ketone 25 and enal 26. Initial
attempts to prepare aldol 41 via the addition of the lithium
enolate of 25 to aldehyde 26 met with variable and poorly
reproducible results (0-50%, depending on the reaction scale).
Aldol 41 is very sensitive to retroaldol cleavage, which occurred
even during silica gel chromatographic purification. It was
possible to trap aldol 41 as the corresponding TBS ether in 69%
yield by adding TBS-OTf before the reaction workup. Simple
modification of this sequence involving treatment of the
intermediate lithium aldolate with Ac2O and DMAP followed
by addition of DBU provided the desired trienone 42 in 61%
yield. Suzuki coupling of 42 with vinylboronic acid 24 in the
presence of Pd2(dba)2 and TlOH provided seco ester 43 in 74%
yield.33,47 Finally, deprotection of the benzyl ester and the C(17)
triethylsilyl ether completed the synthesis of seco acid 44.
Synthesis of Enal 26. The synthesis of this fragment
originated with the highly diastereoselective (Z)-crotylboration
of (R)-lactaldehyde derivative 3745 and (R,R)-38,40,41 which
provided 39 with g99% diastereoselectivity (only one isomer
detected).46 Ozonolysis of 39, treatment of the resulting
â-hydroxy aldehyde with Ph3PdC(Me)CO2Et provided enoate
40 in 79% yield. Finally, DIBAL-H reduction of the carbo-
ethoxy group and MnO2 oxidation of the allylic alcohol provided
the targeted enal 26.
(39) Schmid, C. R.; Bradley, D. A. Synthesis 1992, 587.
(40) Roush, W. R.; Halterman, R. L. J. Am. Chem. Soc. 1986, 108, 294.
(41) Roush, W. R.; Hoong, L. K.; Palmer, M. A. J.; Straub, J. A.;
Palkowitz, A. D. J. Org. Chem. 1990, 55, 4117.
(42) Tidwell, T. T. Synthesis 1990, 857.
(43) Corey, E. J.; Fuchs, P. L. Tetrahedron Lett. 1972, 3769.
(44) Hegedus, L. S. In ComprehensiVe Organic Synthesis; B. M. Trost,
Ed.; Pergamon Press: Oxford, 1991; Vol. 4, p 551.
(45) Soai, K.; Shimada, C.; Takeuchi, M.; Itabashi, M. J. Chem. Soc.,
Chem. Commun. 1994, 567.
(46) Masamune, S.; Choy, W.; Petersen, J. S.; Sita, L. R. Angew. Chem.,
Int. Ed. Engl. 1985, 24, 1.
Addition of the mixed anhydride generated by treatment of
44 with trichlorobenzoyl chloride and Et3N in THF to a 100 °C
(47) Uenishi, J.; Beau, J.-M.; Armstrong, R. W.; Kishi, Y. J. Am. Chem.
Soc. 1987, 109, 4756.

+
+
7506 J. Am. Chem. Soc., Vol. 118, No. 32, 1996
Roush et al.
solution of DMAP in toluene provided a mixture of two
tetracyclic products, 45 and 46, in 79% combined yield.48 The
18-membered macrolide 22 was not observed under these
conditions. The stereochemistry of the major product, 45, which
was isolated in 66% yield, was assigned on the basis of 1H NMR
data. In addition to the J data summarized in the accompanying
this reaction possesses a trans ring fusion. This product arises
from an acyclic transition state analogous to 14. It is the minor
product in this series, 48, which derives from an acyclic
transition state analogous to 16, that has the correct nargenicin
stereochemistry. Interestingly, both 47 and 48 possess the same
relative stereochemistry at C(7)-C(8), indicating that the steric
directing group functioned as intended in allowing only one
face of the diene to interact with the dienophile in the reaction
transition states.25 The problem with the cyclization of 43 is
that both faces of the dienophile interact with the diene, leading
to the formation of two products with, unfortunately, the
incorrect stereoisomer 47 predominating. This provides retro-
spective support for our original hypothesis that allylic interac-
tions involving the C(16) stereocenter in the macrocyclic tetraene
would serve as an important stereochemical control element and
dictate the face of the dienophile that is presented to the diene
in the transannular Diels-Alder reaction.
figure, strong NOEs observed between H7-H12 (4.9%), H9-
H10 (4.6%), and H10-H13 (3.7%) are uniquely consistent with
the stereostructure depicted for 45. Interestingly, all stereo-
centers of 45 except C(8) correspond exactly to the stereo-
chemistry of the natural nargenicins. The minor product, 46,
isolated in 13% yield, is the C(10) epimer of 45. These results
verified that the stereoselectivity of this transannular Diels-
Alder reaction is strongly biased to produce cycloadduct 45,
exactly as predicted by our analysis of transition states 14-17.
In an effort to minimize the extent of epimerization of 45,
the macrolactonization of 44 was performed at 80 °C. Under
these conditions, macrolide 22 was obtained in 38% yield along
with 32% of tetracycle 45 and only 5% of the C(10) epimer
46. With a sample of 22 available, it was possible to assess
the diastereoselectivity of the transannular Diels-Alder reaction.
In the event, when a toluene solution of 22 was heated at 100
°C for 20 h, tetracycle 45 was obtained in 85% yield as the
sole reaction product. Therefore, it is clear that the transannular
Diels-Alder reaction of 22 is highly stereoselective, and that
the C(10)-epimer 46 derives from epimerization of either 44
(or the corresponding mixed anhydride), 22, or 45 under the
weakly basic macrolactonization conditions.
In order to verify our prediction that the success of the
synthesis of 45 depends on the conformational preferences of
macrocycle 22 in the transannular cycloaddition transition states,
and that a conventional intramolecular Diels-Alder (IMDA)
reaction would not be very stereoselective in this series,6,11,12
we examined the thermal cyclization of seco ester 43. When a
toluene solution of 43 was heated at 110 °C for 24 h, a ca. 2:1
mixture of cycloadducts 47 and 48 was obtained. As is clearly
illustrated in the following diagram, the major product (47) of
Synthesis of Macrolide 13. One problem not solved by the
successful transannular Diels-Alder reaction of 22 concerns
the introduction of the C(8)-C(13) oxa bridge present in the
natural products. Unfortunately, repeated attempts to introduce
the oxa bridge by application of remote oxidation procedures49-51
on intermediates derived from 45 were not successful.52 Ac-
cordingly, it was apparent that if the transannular Diels-Alder
strategy was to yield a productive route to the naturally occurring
(49) Barton, D. H. R.; Beaton, J. M.; Geller, L. E.; Pechet, M. M. J.
Am. Chem. Soc. 1961, 83, 4076.
(50) Heusler, K.; Kalvoda, J. HelV. Chim. Acta 1963, 46, 2020.
(51) Mihailovic, M. L.; Gojkovic, S.; Konstantinovic, S. Tetrahedron
1973, 29, 3675.
(48) Inanaga, J.; Hirata, K.; Saeki, H.; Katsuki, T.; Yamaguchi, M. Bull.
Chem. Soc. Jpn. 1979, 52, 1989.

+
+
Studies on the Synthesis of Nargenicin A1
J. Am. Chem. Soc., Vol. 118, No. 32, 1996 7507
nargenicins, it would be necessary to introduce an oxygen-
containing functionality at C(13) prior to the Diels-Alder
reaction. We therefore synthesized macrolide 13 in anticipation
that it would undergo a facile transannular cycloaddition reaction
to nargenicin precursor 18 (X ) OH).
by standard procedures via the corresponding carboxylic acid.54
The aldol coupling of 49 and 50 provided â-hydroxy ketone
51 in 73% yield as a 3:1 mixture of diastereomers. We presume
that the major diastereomer is the Felkin isomer, as formulated
in 51, but this stereochemical assignment was not established
rigorously. The diastereomeric mixture was not separated on
a routine basis, since both isomers behaved similarly in
subsequent transformations. Cross coupling of 51 (3:1 diaster-
eomeric mixture) with vinylboronic acid 23 provided 52 in 73%
yield. Standard deprotection of the benzyl ester and triethylsilyl
ether provided the seco acid 53, which when subjected to the
Yamaguchi macrocyclization protocol provided macrolide 54
in 55% yield.48 It is noteworthy that the potentially sensitive
â-hydroxy ketone functionality survived the sequence of reac-
tions from 51 to 54 without protection. Finally, oxidation of
54 with the Swern DMSO-trifluoroacetic anhydride reagent42
provided macrocycle 13, which exists exclusively as an enol
tautomer. While 13 is probably a rapidly equilibrating mixture
of the two isomeric enols, the regioisomer formulated as 13a is
expected to be the predominate one.55,56
Unfortunately, numerous attempts to induce 13 to undergo
the desired transannular Diels-Alder reaction have not been
successful. Macrocycle 13 was recovered unchanged when
heated to 200 °C in toluene; at higher temperatures (>250 °C)
substantial decomposition was observed. No reaction was
observed when solutions of 13 were exposed to ultrasound at
50 °C,57 or when 13 was dissolved in a 5 M solution of LiClO4
in Et2O.58 Similarly, macrocycle 13 was recovered unchanged
when treated with various Lewis acids including anhydrous
ZnCl2 in THF, Me2AlCl in THF, BF3‚Et2O in CH2Cl2 or LiBF4
in CH3CN at ambient temperature. Decomposition was ob-
served when 13 was treated with Bu3B and HOAc, in an attempt
to generate a cyclic borate ester.59
It was clear from these results that the â-hydroxy enone unit
in 13, which can be thought of as an “acac” complex of a proton,
is too electron rich to function as a dienophile in the transannular
Diels-Alder reaction and that the anticipated entropic assistance
of the transannular process is insufficient to overcome this elec-
tronic barrier.27 However, the literature contains several exam-
ples of intramolecular Diels-Alder reactions of â-alkoxy and
â-acetoxy enones.60,61 Accordingly, many attempts were made
to convert 13 or the analogous des-methoxy macrolide 55 into
the corresponding enol acetate derivatives (e.g., 56 from 13)
In view of the instability of aldol 41 (vide supra), the synthesis
of 13 that we developed proceeds by way of aldol 51 that is
considerably more stable. The required aldehyde component
49 was prepared by selective ozonolysis of the vinyl unit of
36.53 Aldehyde 49 was used without purification directly in
the aldol reaction with methyl ketone 50, which was prepared
(52) Numerous attempts to introduce the oxa bridge by remote oxidation
of alcohol i, which was prepared by an eight step sequence from 45, were
unsuccessful. Methods examined include Pb(OAc)4, CaCO3, benzene, 80
°C; Pb(OAc)4, I2, CaCO3; Ag2CO3, Br2; Ag2CO3, I2; DDQ, CH2Cl2; and
nitrite ester photolysis (see refs 49-51).
(53) Pryor, W. A.; Giamalva, D.; Church, D. F. J. Am. Chem. Soc. 1983,
105, 6858.
(54) Rubottom, G. M.; Kim, C. J. Org. Chem. 1983, 48, 1550.
(55) Bassetti, M.; Cerichelli, G.; Floris, B. Tetrahedron 1988, 44, 2997.
(56) Vila, A. J.; Lagier, C. M.; Oliveri, A. C. J. Phys. Chem. 1991, 95,
5069.
(57) Lee, J.; Snyder, J. K. J. Am. Chem. Soc. 1989, 111, 1522.
(58) Grieco, P. A.; Nunes, J. J.; Gaul, M. D. J. Am. Chem. Soc. 1990,
112, 4595.
(59) Narasaka, K.; Pai, F. Tetrahedron 1984, 40, 2233.

+
+
7508 J. Am. Chem. Soc., Vol. 118, No. 32, 1996
Roush et al.
by treatment with isopropenyl acetate or acetic anhydride in
the presence of p-TsOH, or with acetyl chloride and Et3N. How-
ever, no reaction occurred at ambient temperature to 90 °C,
while substantial decomposition was observed above 100 °C.
An enol acetate derivative of undetermined geometry was gener-
ated by treatment of 55 with LiN(TMS)2 and Ac2O, but this
material decomposed when heated at 110 °C in toluene. Simi-
larly, attempts to generate and cyclize enol diethyl phosphate
(LHDMS, (EtO)2POCl, THF) and enol sulfonate derivatives
(MsCl, Et3N) were unsuccessful. Attempts to generate enol
ether derivatives by treatment of 13 or 55 with CH2N2, TsOH
in MeOH, or MOM-Cl and i-Pr2NEt in CH2Cl2 were similarly
unproductive. A triethylsilyl enol ether (of undetermined regio-
and stereochemistry) was obtained by treatment of 13 with Et3-
SiOTf and lutidine at -78 °C, but this material also failed to
undergo the desired transannular Diels-Alder reaction.
To the best of our knowledge, the tandem macrolactonization
of 44 and transannular Diels-Alder reaction of 22 is the first
fully documented example in which the transition state confor-
mational preferences of the macrocycle play a decisive role in
determining the stereoselectivity of the cycloaddition event.24
One hurdle remains to be solved in order for this strategy to
be successfully applied to the total synthesis of the nargenicins,
namely establishment of a method for introduction of the oxa
bridge.52 Studies addressing this issue are in progress, and will
be the subject of future reports from our laboratory.
Experimental Section
General Methods. All reactions were conducted in flame-dried
glassware under dry argon or nitrogen. All solvents except DMF were
purified before use: diethyl ether, THF, and toluene were distilled from
sodium benzophenone ketyl; dichloromethane and triethylamine were
distilled from CaH2, and methanol was distilled from magnesium
turnings. DMF was used as received from commercial sources.
1H NMR spectra were measured at 300 and 400 MHz on com-
mericially available instruments. Chemical shifts are reported in δ units;
coupling constants are reported in Hz. Residual chloroform (δ 7.26)
and benzene (δ 7.15) were used as internal references for spectra
measured in these solvents. 13C NMR spectra were measured at 100.6
MHz, and residual chloroform (δ 77.0) and benzene (δ 128.0) were
used as internal references. High resolution mass spectra were
measured at 70 eV. Optical rotations were measured on a Rudolph
Autopol III polarimeter using a quartz cell with 1 mL capacity and a
10 cm path length. Elemental analyses were performed by Robertson
Laboratories, Florham Park, NJ.
Analytical thin layer chromatography (TLC) was performed with
the use of plates coated with a 0.25 mm thickness of silica gel containing
PF254 indicator (Analtech), and compounds were visualized with UV
light, iodine, p-anisaldehyde, ceric ammonium molybdate, or ninhydrin
stain. Preparative TLC was performed with the use of 20 cm × 20
cm plates coated with a 0.50 mm thickness of silica gel containing
PF254 indicator (Analtech). Flash chromatography was performed as
described by Still with the use of Kieselgel 60 (230-400 mesh).63
Benzyl (2S)-2-hydroxy-5-(trimethylsilyl)-4-pentynoate (28). To
a 0 °C solution of (trimethylsilyl)acetylene (20.3 mL, 144 mmol) in
toluene (120 mL) was added n-BuLi (80 mL, 2.5 M in hexanes, 144
mmol) by syringe over 30 min. The resulting mixture was stirred for
30 min at which time diethylaluminum chloride (80 mL, 1.8 M in
hexanes, 144 mmol) was added. The solution was warmed to room
temperature, stirred for 2 h and recooled to 0 °C; a solution of benzyl
(2S)-glycidate34 (27, 12.8 g, 71.7 mmol) in 40 mL toluene was then
added via cannula. The solution was stirred at 0 °C for 4 h at which
time it was diluted with ether (200 mL), acidified with 1 M HCl (pH
∼ 4) and extracted with ether (4 × 100 mL). The combined organic
layers were washed with water (100 mL), dried with MgSO4, filtered
and concentrated. Purification of the crude product by flash chroma-
tography [silica gel, 10 × 15 cm, gradient elution: 5% ethyl acetate/
hexanes (2 L); 10% ethyl acetate/hexanes (1 L); 15% ethyl acetate/
Conclusions
We have established that the complete carbon skeleton of
the nargenicins can be assembled with high stereoselectivity
from subunits 24, 25, and 26 by a sequence involving the
macrolactonization of seco acid 44 and the transannular Diels-
Alder reaction of the derived 18-membered macrolide 22. This
sequence permits all of the stereochemical features of the
nargenicin nucleus, represented by structure 45, to be controlled.
In contrast, stereocontrolled construction of the hydronaphtha-
lene nucleus is not possible in the conventional intramolecular
Diels-Alder mode as illustrated by the non-selective IMDA
reaction of seco ester 43. An added benefit of this strategy is
that the troublesome 10-membered lactone is established by a
formal ring contraction of the more easily synthesized 18-
membered lactone.
Two independent stereochemical control features determine
the success of the transannular Diels-Alder reaction of 22. First,
the C(6)-Br steric directing substituent dictates that only one
of the two faces of the diene is accessible to the dienophile in
transition state 14 (and potentially also in 16).25 Second, allylic
strain considerations involving the C(16)-Me substituent dictate
that only one face of the dienophile is accessible to the diene
in transition state 14 (and possibly also 15).26 The latter effect
is operational only in the transannular cycloaddition mode due
to conformational constraints posed by the 18-membered
macrocycle. The combination of these two effects lead to 14
being the most favorable of all the possible transition states,
and as a result the transannular Diels-Alder reaction is highly
stereoselective.
23
hexanes (2 L)] gave alcohol 28 (16.5 g, 83%) as a colorless oil: [R]D
1
-71.5° (c 0.9, hexanes); H NMR (300 MHz, CDCl3) δ 7.45-7.35
(m, 5 H), 5.27 and 5.21 (AB, J ) 12.1 Hz, 2 H), 4.36 (m, J ) 7.7, 4.9
Hz, 1 H), 3.05 (d, J ) 6.9 Hz, 1 H), 2.7-2.8 (m, 2 H), 0.13 (s, 9 H);
IR (neat) 3500, 2180, 1750 cm-1; HRMS calcd for C15H20O3Si (M+),
276.1182, found 276.1184. Anal. Calcd for C15H20O3Si: C, 65.17;
H, 7.31. Found: C, 65.00; H, 7.11.
Benzyl (2S)-2-methoxy-4-pentynoate (29). To a stirred solution
of alcohol 28 (3.0 g, 11 mmol) in CH2Cl2 (15 mL) in a 50 mL round
bottom flask with a screw cap seal was added silver(I) oxide (2.5 g, 11
mmol) and methyl iodide (2.0 mL, 32 mmol). The tube was sealed,
placed in an oil bath at 50 °C and stirred for 12 h at which time
additional silver(I) oxide (2.5 g, 22 mmol) and methyl iodide (2.0 mL,
65 mmol) were added. After being stirred for an additional 12 h at 50
°C the solution was cooled, filtered through Celite with CH2Cl2 and
concentrated in a fume hood by aspirator vacuum. Purification of the
crude product by flash chromatography (silica gel, 5 × 15 cm, 5%
ethyl acetate/hexanes) gave the methyl ether (2.28 g, 72%) as a colorless
(60) Schlessinger, R. H.; Wong, J.-W.; Poss, M. A.; Springer, J. P. J.
Org. Chem. 1985, 50, 3950.
(61) Shishido, K.; Takahashi, K.; Fukumoto, K.; Kametani, T.; Honda,
T. J. Org. Chem. 1987, 52, 5704.
(62) For two other examples of transannular Diels-Alder reactions that
are more selective than analogous intramolecular Diels-Alder reactions,
see: (a) Roush, W. R.; Warmus, J. S.; Works, A. B. Tetrahedron Lett.
1993, 34, 4427. (b) Jung, S. H.; Lee, Y. S.; Park, H.; Kwon, D.-S.
Tetrahedron Lett. 1995, 36, 1051.
(63) Still, W. C.; Kahn, M.; Mitra, A. J. Org. Chem. 1978, 43, 2923.

+
+
Studies on the Synthesis of Nargenicin A1
J. Am. Chem. Soc., Vol. 118, No. 32, 1996 7509
23
1
oil: [R]D -13.6° (c 0.5, hexanes); H NMR (300 MHz, CDCl3) δ
7.4-7.3 (m, 5 H), 5.21 (s, 2 H), 3.95 (m, 1 H), 3.45 (s, 3 H), 2.75-
2.65 (m, 2 H), 0.12 (s, 9 H); IR (neat) 2180, 1760 cm-1; HRMS calcd
for C16H22O3Si (M+), 290.1339, found 290.1354. Anal. Calcd for
C16H22O3Si: C, 66.22; H, 7.65. Found: C, 65.82; H, 7.41.
188.0833. Anal. Calcd for C12H12O2: C, 76.57; H, 6.43. Found: C,
76.29; H, 6.66.
(E)-4-(Benzyloxycarbonyl)but-1-enylboronic Acid (24). To a
flame-dried Carius tube was added benzyl 4-pentynoate (30) (686 mg,
3.65 mmol) and freshly distilled catecholborane (98 µL, 0.90 mmol).
After gas evolution ceased, the tube was sealed and heated to 100 °C.
An additional 3 aliquots of catecholborane (97 µL, 0.90 mmol, each)
were added over 9.5 h (total volume for four additions of catecholborane
was 390 µL, 3.68 mmol), and the mixture was stirred for an additional
20 h at 100 °C. The mixture was cooled to ambient temperature and
diluted with brine (5 mL) and EtOAc (5 mL). The resulting mixture
was vigorously stirred for 40 min. More EtOAc (10 mL) was added,
and the two-phase system was separated. The aqueous phase was
extracted with EtOAc (5 × 5 mL), and the combined organic extracts
were dried (MgSO4), filtered, and concentrated in Vacuo. The residue
was purified by flash chromatography on a short, wide column with
1:1 hexanes-ether to remove catechol, then quickly eluted with 95:5
CH2Cl2-CH3OH to provide 24 (519 mg, 61%) as a very viscous clear
oil: Rf 0.33 (95:5 CH2Cl2-CH3OH); 1H NMR (400 MHz, CD3OD) δ
7.40-7.28 (m, 5 H), 6.5 (dt, J ) 17.6, 5.8 Hz, 1 H), 5.6 (d, J ) 17.6
Hz, 1 H), 5.1 (s, 2 H), 2.5-2.4 (m, 4 H); IR (neat) 3430 (br), 1760,
1660 cm-1; HRMS (EI) calcd for C12H13O3 (M+ - B(OH)2) 189.0915,
found 189.0899. This compound was fully characterized as the pinacol
ester, which was prepared as described in the following procedure.
To a solution of vinyl boronic acid 24 (57 mg, 0.24 mmol) in dry
THF (1 mL) was added pinacol (26 mg, 0.22 mmol) and Na2SO4 (20
mg, 0.10 mmol). After being stirred at room temperature under N2
for 2 d, the mixture was filtered through a cotton plug and concentrated
in Vacuo. The crude product was purified by flash chromatography
using 5:1 hexanes-ether as eluant to provide the pinacol ester (47 mg,
74%) as a clear oil: 1H NMR (400 MHz, CDCl3) δ 7.40-7.30 (m, 5
H), 6.62 (m, 1 H), 5.47 (d, J ) 18.0 Hz, 1 H), 5.10 (s, 2 H), 2.51-
2.40 (m, 4 H), 1.26 (s, 12 H); 11B NMR (115 MHz, CDCl3) δ 29.9 (s);
IR (neat) 1735, 1635 cm-1; HRMS (CI) calcd for C18H25O4B (M+),
315.1882, found 315.1858. Anal. Calcd for C18H25O4B: C, 68.37; H,
7.97. Found: C, 68.41; H, 7.68.
To a stirred, 23 °C solution of the above methyl ether (1.74 g, 5.99
mmol) in N,N-dimethylformamide (20 mL) and water (7 mL) was added
potassium fluoride dihydrate (2.82 g, 30.0 mmol) in one portion. The
resulting mixture was stirred for 20 h at which time it was poured into
brine solution (20 mL). The aqueous layer was extracted with ether
(3 × 40 mL) and the combined organic layers were washed with brine
(40 mL), dried with MgSO4, filtered and concentrated. Purification of
the crude product by flash chromatography (silica gel, 5 × 15 cm,
15% ethyl acetate/hexanes) gave alkyne 29 (1.06 g, 81%) as a colorless
23
1
oil: [R]D -29.0° (c 0.4, hexanes); H NMR (300 MHz, CDCl3) δ
7.4-7.3 (m, 5 H), 5.26 and 5.20 (AB, J ) 12 Hz, 2 H), 3.96 (t, J )
6.1 Hz, 1 H), 3.46 (s, 3 H), 2.8-2.6 (m, 2 H), 2.68 (t, J ) 2.3 Hz, 1
H); IR (neat) 3300, 2220, 1750 cm-1; HRMS calcd for C13H14O3 (M+),
218.0943, found 218.0939. Anal. Calcd for C13H14O3: C, 71.54; H,
6.47. Found: C, 71.28; H, 6.72.
(E)-(4S)-4-(Benzyloxycarbonyl)-4-methoxybut-1-enylboronic Acid
(23). To a neat sample of alkyne 29 (800 mg, 3.66 mmol) in a tube
equipped with a screw cap seal was added catechol borane (98 µL,
0.92 mmol) via syringe. The tube was flushed with argon, sealed and
placed in an oil bath at 100 °C. After being stirred for 2.5 h the tube
was cooled and placed under a positive pressure of argon; additional
catechol borane (98 µL, 0.92 mmol) was added. This process was
repeated until a total of 1.5 equiv of catechol borane (586 µL, 5.5 mmol
total) had been added. After additional heating at 100 °C for 6 h the
solution was cooled and carefully quenched with a 1:1 ethyl acetate/
water mixture (4 mL). This mixture was stirred at room temperature
for 12 h, diluted with 20 mL ethyl acetate and the aqueous phase was
extracted with ethyl acetate (2 × 20 mL). The combined organic layers
were washed with brine (20 mL), dried with MgSO4, filtered and
concentrated. Purification of the crude product by flash chromatography
[silica gel, 3 × 15 cm, gradient elution: 50% ethyl acetate/hexanes
(0.5 L); 5% methanol/dichloromethane (0.5 L)] gave boronic acid 23
(722 mg, 75%) which was stored as a 0.5 M solution in methanol:
[R]D23 -39.3° (c 1.2, methanol); 1H NMR (400 MHz, CDCl3) δ 7.4-
7.3 (m, 5 H), 6.44 (dt, J ) 17.2, 7.0 Hz, 1 H), 5.59 (dt, J ) 17.2, 1.4
Hz, 1 H), 5.17 (AB, J ) 12.2 Hz, 2 H), 3.95 (dd, J ) 6.6, 5.4 Hz, 1
H), 3.34 (s, 3 H), 2.7-2.5 (m, 2 H); IR (CCl4) 3660, 3460, 1760, 1640
cm-1. This compound was fully characterized as the pinacol ester, as
described in the following procedure.
(2S,3R,4R)-1-[(tert-Butyldimethylsilyl)oxy]-4-methylhex-5-en-2,3-
diol (34). Chiral crotylboronate (S,S)-32 (73.6 mL of a 0.93 M solution
in toluene, 68 mmol)65 was diluted with toluene (200 mL) and treated
with powdered 4 Å molecular sieves (10 g, Aldrich) at room
temperature for 30 min. This dispersion was then cooled to -78 °C.
To this solution was then added a solution of 2,3-O-(3-pentylidene)-
L-glyceraldehyde (31) (7.22 g, 45.6 mmol)39 in toluene (10 mL) via
syringe pump over a 30 min period. The reaction mixture was stirred
for an additional 1.5 h at -78 °C at which time 1 M NaOH (135 mL)
was added; the resulting mixture was then stirred at room temperature
for 3 h. The organic phase was separated and the aqueous layer was
extracted with diethyl ether (3 × 100 mL). The combined organic
layers were washed with saturated NaHCO3 (100 mL) and brine (100
mL), dried with MgSO4, filtered and concentrated via a distillation
apparatus (3′′ Vigreux column; 40 °C at 30 mm Hg). Purification of
the resulting mixture of diastereomers (91:9 by GC analysis) by flash
chromatography [silica gel, 7 × 15 cm, gradient elution: hexanes (1
L); 10% ethyl ether/hexanes (1 L); 20% ethyl ether/hexanes (1 L)] gave
the desired homoallylic alcohol 33 as a colorless oil: [R]D23 -15.6° (c
To a mixture of boronic acid 23 (132 mg, 0.50 mmol) and Na2SO4
(85 mg) in THF (2 mL) at 23 °C was added pinacol (60 mg, 0.50
mmol) in one portion. This mixture was stirred at room temperature
for 1 h, filtered through a cotton plug and concentrated in Vacuo.
Purification of the crude product by flash chromatography (silica gel,
2 × 15 cm, 10% ethyl acetate/hexanes) gave the pinacol ester (111
mg, 67%) as a clear oil: [R]D23 -24.6° (c 1.4, CHCl3); 1H NMR (400
MHz, CDCl3) δ 7.4-7.3 (m, 5 H), 6.57 (dt, J ) 17.8, 6.6 Hz, 1 H),
5.52 (dd, J ) 17.8, 1.4 Hz, 1 H), 5.18 (s, 2 H), 3.90 (t, J ) 5.6 Hz, 1
H), 3.38 (s, 3 H), 2.65-2.55 (m, 2 H), 1.25 (s, 12 H); 11B NMR (115
MHz, CDCl3) δ 29.5; IR (CCl4) 1760, 1640 cm-1; HRMS calcd for
C19H27O5B (M+), 346.1951, found 346.1936.
1
0.9, CHCl3); H NMR (400 MHz, CDCl3) δ 5.9-5.8 (m, 1 H), 5.2-
5.1 (m, 2 H), 4.1-4.0 (m, 2 H), 3.9-3.8 (m, 1 H), 3.7-3.6 (m, 1 H),
2.45-2.35 (m, 1 H), 1.7-1.6 (m, 4 H), 1.09 (d, J ) 7.2 Hz, 3 H), 0.90
(q, J ) 7.2 Hz, 6 H); IR (CCl4) 3600, 3500, 1640 cm-1; HRMS calcd
for C10H17O3 (M+ - C2H5), 185.1178, found 185.1182. Anal. Calcd
for C12H22O3: C, 67.24; H, 10.37. Found: C, 67.40; H, 10.47.
This material was directly dissolved in a 2:1 mixture of methanol
and tetrahydrofuran (30 mL) and heated at reflux in the presence of
Dowex 50 × 4-400 H+ resin for 8 h. The solution was filtered, then
the solvent was removed in Vacuo and the crude triol was purified by
flash chromatography [silica gel, 3 × 15 cm, gradient elution: 50%
ethyl ether/hexanes (0.5 L); 50% methanol/ethyl acetate (0.5 L)], which
afforded (2S,3R,4R)-4-methylhex-5-en-1,2,3-triol (5.78 g, 86%) as a
Benzyl 4-Pentynoate (30). A solution of 4-pentynoic acid64 (3.15
g, 30 mmol) in CH3CN (100 mL) was treated with DBU (6.8 mL, 45
mmol) and benzyl bromide (9.5 mL, 80 mmol). This mixture was
stirred at room temperature overnight under N2. Et2O (100 mL) and
HCl (1 M, 50 mL) were added, the layers were separated, and the
organic layer was washed with aqueous HCl (1 M, 3 × 20 mL). The
combined aqueous layers were extracted with ether (3 × 20 mL). The
organic phases were combined, washed with brine, dried (MgSO4),
filtered, and concentrated in Vacuo. Residual CH3CN was removed
by distillation, and the pot residue was purified by flash chromatography
using 20:1 pentane-ether as eluant to yield 30 (4.63 g, 82%) as a yellow
1
oil: Rf 0.76 (95:5 CH2Cl2-CH3OH); H NMR (400 MHz, CDCl3) δ
23
1
white solid (mp 69-71 °C): [R]D +5.1° (c 0.5, CHCl3); H NMR
(400 MHz, CDCl3) δ 5.82 (ddd, J ) 18.0, 9.0, 8.0 Hz, 1 H), 5.18 (d,
J ) 12.0 Hz, 1 H), 5.17 (d, J ) 16.0 Hz, 1 H), 3.7-3.9 (m, 2 H),
7.40-7.28 (m, 5 H), 5.15 (s, 2 H), 2.64-2.59 (m, 2 H), 2.55-2.51
(m, 2 H), 1.98 (t, J ) 2.6 Hz, 1 H); IR (neat) 3290, 2120, 1740 (br),
1615 cm-1; HRMS (CI) calcd for C12H12O2 (M+) 188.0837, found
(65) Roush, W. R.; Palkowitz, A. D.; Ando, K. J. Am. Chem. Soc. 1990,
112, 6348.
(64) Schulte, K. E.; Reiss, K. P. Chem. Ber. 1954, 87, 964.

+
+
7510 J. Am. Chem. Soc., Vol. 118, No. 32, 1996
Roush et al.
3.7-3.6 (m, 1 H), 3.6-3.5 (m, 1 H), 2.5-2.4 (m, 1 H), 1.06 (d, J )
7.0 Hz, 3 H); IR (CH2Cl2) 3600, 1460 cm-1; HRMS (CI, NH3) calcd
for C7H15O3 (M++1), 147.1021, found 147.1009. Anal. Calcd for
C7H14O3Si: C, 57.51; H, 9.65. Found: C, 57.21; H, 9.47. The antipode
of this triol is a known compound.11
To a solution of triphenylphosphine (11.2 g, 42.9 mmol) in CH2Cl2
(16 mL) cooled at 0 °C was added a solution of carbon tetrabromide
(7.12 g, 21.5 mmol) in CH2Cl2 (8 mL) via cannula. After the solution
was stirred for 1 h at 0 °C, a solution of freshly prepared aldehyde
from the preceding experiment in CH2Cl2 (10 mL) was added via
cannula. The mixture was stirred for 3 h at 0 °C at which time hexanes
(40 mL) were added. The milky solution was filtered through a pad
of silica gel/Celite with 20% ethyl acetate/hexanes as the eluent; the
insoluble material left in the reaction flask was subjected to additional
(3×) cycles of CH2Cl2 extraction and hexanes precipitation to remove
any remaining product. Concentration of the filtrate followed by
purification of the crude product by flash chromatography [silica gel,
5 × 15 cm, gradient elution: hexanes (1 L); 2% ethyl acetate/hexanes
(1 L); 5% ethyl acetate/hexanes (1 L)] gave the dibromo diene 36 (2.93
To a solution of the above triol (5.78 g, 39.3 mmol) in dry N,N-
dimethylformamide (80 mL, Aldrich) at 0 °C was added tert-
butyldimethylsilyl chloride (5.92 g, 39.3 mmol), triethylamine (6.37
mL, 45.7 mmol) and 4-dimethylaminopyridine (DMAP, 467 mg, 3.82
mmol). The resulting mixture was warmed to room temperature and
stirred for 24 h at which time additional DMAP (467 mg, 3.82 mmol)
was added. After being stirred for an additional 12 h the solution was
poured into saturated NaHCO3 (100 mL). The aqueous phase was
extracted with diethyl ether (4 × 50 mL) and the combined organic
layers were washed with water (100 mL), dried with MgSO4, filtered
and concentrated. Purification of the residue by flash chromatography
[silica gel, 5 × 15 cm, gradient elution: hexanes (0.5 L); 10% ethyl
acetate/hexanes (0.5 L); 20% ethyl acetate/hexanes (1 L)] gave diol 34
23
1
g, 80%) as a colorless oil: [R]D +4.1° (c 0.4, hexanes); H NMR
(300 MHz, CDCl3) δ 6.52 (d, J ) 9.5 Hz, 1 H), 5.88 (ddd, J ) 17.8,
10.8, 7.0 Hz, 1 H), 6.0-5.8 (m, 2 H), 4.74 (dd, J ) 9.5, 5.4 Hz, 1 H),
3.97 (dd, J ) 9.0, 5.4 Hz, 1 H), 2.2-2.4 (m, 1 H), 1.47 (s, 3 H), 1.36
(s, 3 H), 0.99 (d, J ) 7.3 Hz, 3 H); IR (neat) 1640, 1620, 1460 cm-1
;
23
1
(10.2 g, 98%) as a colorless oil: [R]D +7.2° (c 1.0, hexanes); H
NMR (300 MHz, CDCl3) δ 6.0-5.8 (m, 1 H), 5.2-5.0 (m, 2 H) 3.9-
3.7 (m, 2 H), 3.6-3.5 (m, 1 H), 3.5-3.4 (m, 1 H), 2.6-2.5 (m, 1 H),
1.08 (d, J ) 6.0 Hz, 3 H), 0.91 (s, 9 H), 0.09 (s, 6 H); IR (neat) 3440,
1640 cm-1; HRMS (CI, NH3) calcd for C13H29O3Si (M+ + 1), 261.1886,
found 261.1877. The antipode of 34 is a known compound.11
HRMS (CI, NH3) calcd for C10H13Br2O2 (M+ - CH3), 323.9281, found
323.9277. Anal. Calcd for C11H16Br2O2: C, 38.85; H, 4.74. Found:
C, 38.86; H, 4.65.
(3S,4R,5R)-1,1-Dibromo-3,4-(O-isopropylidene)-5-methylhept-1-
en-6-one (25). A mixture of PdCl2 (80 mg, 0.45 mmol), CuCl (450
mg, 4.5 mmol), and dibromodiene 36 (1.53 g, 4.5 mmol) in DMF (22
mL) and H2O (3 mL) was stirred under an O2 atmosphere for 24 h.
The mixture was then poured into water (100 mL) and extracted with
Et2O (3 × 25 mL). The ethereal extracts were washed with brine
solution, dried (Na2SO4), filtered and concentrated in vacuo. Purifica-
tion of the crude product by flash chromatography [silica gel, 5 × 15
cm, gradient elution: 5% Et2O in hexanes (200 mL); 10% Et2O in
hexanes (200 mL); 15% Et2O in hexanes (200 mL); 20% Et2O in
hexanes (200 mL); then 25% Et2O in hexanes (200 mL)] provided
methyl ketone 25 (1.04 g, 65%): Rf (0.3, 3:1 hexanes-Et2O); 1H NMR
(400 MHz, CDCl3) δ 6.46 (d, J ) 10 Hz, 1 H), 4.76 (dd, J ) 9.3, 5.5
Hz, 1 H), 4.26 (dd, J ) 10.6, 5.5 Hz, 1 H), 2.71-2.65 (m, 1 H), 2.24
(s, 3 H), 1.34 (s, 3 H), 1.02 (d, J ) 7.1 Hz, 3 H); IR (neat) 1722, 1615,
1457 cm-1; HRMS (CI, CH4) calcd for C10H13O3Br2, 342.9192, found
342,9205.
(2S,3R,4R)-1-Hydroxy-2,3-(O-isopropylidene)-4-methylhex-5-
ene (35). To 0 °C a solution of diol 34 (10.2 g, 38.8 mmol) and
pyridinium p-toluenesulfonate (0.989, 3.9 mmol) in CH2Cl2 (100 mL)
was added 2-methoxypropene (7.44 mL, 77.7 mmol) in one portion
via syringe. The reaction mixture was allowed to warm to room
temperature and stirred for 2 h, then it was poured into saturated
NaHCO3 (100 mL). The aqueous layer was extracted with CH2Cl2 (2
× 50 mL) and the combined organic layers were dried with MgSO4,
filtered and concentrated to afford the intermediate acetonide which
was directly dissolved in THF (40 mL) and treated with tetra-n-
butylammonium fluoride (38.8 mL, 1.0 M solution in THF, 38.8 mmol)
at room temperature. After this mixture was stirred for 2 h, the solvent
was removed in Vacuo and the residue was purified by flash chroma-
tography [silica gel, 5 × 15 cm, gradient elution: hexanes (0.5 L);
20% ethyl acetate/hexanes (0.5 L); 30% ethyl acetate/hexanes (1 L)]
to afford alcohol 35 (6.92 g, 96%) as a colorless oil. Data for acetonide
(2R,3S,4R)-2-[(tert-Butyldiphenylsilyl)oxy]-4-methylhex-5-en-3-
ol (39). Crotylboronate (R,R)-38 (37 mL of a 0.90 M solution in
toluene, 33 mmol)65 was diluted with toluene (50 mL) and treated with
powdered 4-Å molecular sieves (10 g, Aldrich) at room temperature
for 30 min. This dispersion was then cooled to -78 °C. A solution
of (2R)-2-(tert-butyldiphenylsilyloxy)propanal (37)45 (7.0 g, 22.4 mmol)
in toluene (10 mL) was then added via syringe pump over a 30 min
period. The reaction mixture was stirred for an additional 1.5 h at
-78 °C at which time 1 M NaOH (135 mL) was added. The resulting
mixture was then stirred at room temperature for 3 h. The organic
phase was separated and the aqueous layer was extracted with diethyl
ether (3 × 50 mL). The combined organic layers were washed with
saturated NaHCO3 (50 mL) and brine (50 mL), dried over MgSO4,
filtered and concentrated. Filtration of the crude product through a 5
cm plug of silica gel with 5% ethyl acetate/hexanes afforded alcohol
39 (8.13 g, 98%) as a colorless oil: [R]D23 +31.1° (c 0.5, hexanes); 1H
NMR (300 MHz, CDCl3) δ 7.7-7.6 (m, 4 H), 7.5-7.3 (m, 6 H), 5.4-
5.2 (m, 1 H), 5.0-4.8 (m, 2 H), 3.9-3.8 (m, 1 H), 3.30 (dd, J ) 8.6,
3.4 Hz, 1 H), 2.2-2.1 (m, 1 H), 1.07 (s, 9 H), 1.04 (d, J ) 5.2 Hz, 3
H), 1.01 (d, J ) 4.9 Hz, 3 H); IR (neat) 3580, 3480, 1640, 1590, 1470,
1460 cm-1; HRMS (CI, NH3) calcd for C23H33O2Si (M+), 369.2251,
found 369.2256. Anal. Calcd for C23H32O2Si: C, 74.95; H, 8.75.
Found: C, 74.8; H, 8.65.
23
1
intermediate: [R]D +12.5° (c 1.4, hexanes); H NMR (400 MHz,
CDCl3) δ 5.91 (ddd, J ) 17.0, 10.0, 7.0 Hz, 1 H), 5.09 (d, J ) 17.0
Hz, 1 H), 5.04 (d, J ) 10.0 Hz, 1 H), 4.2-4.1 (m, 1 H), 3.89 (dd, J )
9.0, 5.5 Hz, 1 H), 3.78 (dd, J ) 10.5, 7.5 Hz, 1 H), 3.54 (dd, J ) 10.5,
5.0 Hz, 1 H), 2.6-2.5 (m, 1 H), 1.33 (s, 3 H), 1.41 (s, 3 H), 1.08 (d,
J ) 7.0 Hz, 3 H), 0.90 (s, 9 H), 0.07 (s, 6 H); IR (neat) 1640, 1470,
1460 cm-1; HRMS (CI, NH3) calcd for C16H33O3Si (M+ + 1), 301.2200,
found 301.2205. Anal. Calcd for C16H32O3Si: C, 63.96; H, 10.73.
Found: C, 63.98; H, 10.43.
23
1
Data for 35: [R]D -47.1° (c 0.6, hexanes); H NMR (400 MHz,
CDCl3) δ 5.88 (ddd, J ) 17.5, 10.2, 7.3 Hz, 1 H), 5.10 (d, J ) 17.2
Hz, 1 H), 5.07 (d, J ) 11.0 Hz, 1 H), 4.16 (apparent dd, J ) 12.4, 5.5
Hz, 1 H), 3.94 (dd, J ) 9.5, 5.5 Hz, 1 H), 3.7-3.6 (m, 2 H), 2.4-2.3
(m, 1 H), 1.37 (s, 3 H), 1.48 (s, 3 H), 1.04 (d, J ) 6.7 Hz, 3 H); IR
(neat) 3460, 1640, 1460 cm-1; HRMS calcd for C10H18O3, 186.1256,
found 186.1243. Anal. Calcd for C10H18O3: C, 64.53; H, 9.77.
Found: C, 64.45; H, 9.87.
(3S,4R,5R)-1,1-Dibromo-3,4-(O-isopropylidene)-5-methylhept-1,6-
diene (36). Oxalyl chloride (1.03 mL, 11.8 mmol) was dissolved in
CH2Cl2 (16 mL) and cooled to -78 °C. Dimethylsulfoxide (1.68 mL,
23.6 mmol) was added and the resulting mixture was stirred at -65
°C for 2 min at which time a solution of alcohol 35 (2.0 g, 11 mmol)
in CH2Cl2 (10 mL) was added via cannula. The solution was stirred
for 15 min; triethylamine (7.48 mL, 53.7 mmol) was added and the
mixture was allowed to warm to room temperature over 1 h. The
reaction mixture was poured into water (20 mL) and the aqueous phase
was extracted with CH2Cl2 (3 × 20 mL). The combined organic layers
were washed with 1 M sodium thiosulfate (20 mL) and brine (20 mL),
dried (MgSO4), filtered and concentrated. The resulting residue was
filtered through a plug of silica gel with 20% ethyl acetate/hexanes.
Concentration of the eluant afforded the intermediate aldehyde that was
used in the following experiment without any further purification.
Ethyl (E)-(4R,5S,6R)-6-[(tert-Butyldiphenylsilyl)oxy]-2,4-dimethyl-
5-[(triethylsilyl)oxy]hept-2-enoate (40). A -78 °C solution of
homoallylic alcohol 39 (2.29 g, 6.22 mmol) in CH2Cl2 (15 mL) was
treated with a stream of O3 in O2 until the solution turned blue; argon
was then bubbled through the solution until the color disappeared. The
resulting solution was treated with triphenylphosphine (3.26 g, 12.4
mmol), warmed to room temperature and stirred for 2 h. Removal of
the solvent gave the aldehyde which was immediately dissolved in CH2-
Cl2 (15 mL) and transferred to a tube equipped with a screw cap seal.
Ethyl 2-(triphenylphosphoranylidene)propionate (4.53 g, 12.4 mmol)
was added, the tube was flushed with argon, sealed and heated in an

+
+
Studies on the Synthesis of Nargenicin A1
J. Am. Chem. Soc., Vol. 118, No. 32, 1996 7511
oil bath at 50 °C for 26 h. The mixture was then triturated with hexanes,
filtered through Celite and concentrated in Vacuo. Purification of the
crude product by flash chromatography [silica gel, 5 × 15 cm, gradient
elution: 5% ethyl acetate/hexanes (1 L); 15% ethyl acetate/hexanes (1
L)] provided the intermediate hydroxy enoate (2.3 g, 82%) as a colorless
oil: [R]D23 -5.4° (c 0.8, CHCl3); 1H NMR (400 MHz, CDCl3) δ 7.7-
7.6 (m, 4 H), 7.5-7.3 (m, 6 H), 6.23 (dd, J ) 10.2, 1.4 Hz, 1 H), 4.14
(q, J ) 7.2 Hz, 2 H), 3.8-3.7 (m, 1 H), 3.38 (dd, J ) 9.0, 3.0 Hz, 1
H), 2.5-2.4 (m, 1 H), 1.75 (d, J ) 1.4 Hz, 3 H), 1.26 (t, J ) 7.2 Hz,
3 H), 1.07 (s, 9 H), 1.02 (d, J ) 6.8 Hz, 3 H), 0.97 (d, J ) 6.4 Hz, 3
H); IR (neat) 3510, 1710 cm-1; HRMS (CI, NH3) calcd for C27H38O4-
Si (M+), 454.2540, found 454.2501.
reaction mixture was then allowed to warm to room temperature and
DMAP (100 mg, 0.82 mmol) was added. After the solution was stirred
for 30 min at ambient temperature, DBU (625 µL, 4.2 mmol) was
added. The mixture was stirred for an additional 30 min, then it was
diluted with ether and poured into brine. The aqueous layer was
separated and extracted with ether. The combined organic extracts were
washed with 0.5 N HCl, brine, and dried over Na2SO4. Removal of
the solvent in Vacuo gave an oily residue, which was purified by silica
gel chromatography (10% ether-hexanes) to give dienone 42 (402 mg,
61%): [R]D20 +71.9° (c ) 1.01, CHCl3); 1H NMR (400 MHz, CDCl3)
δ 7.65-7.60 (m, 4 H), 7.45-7.30 (m, 6 H), 7.07 (d, J ) 15.6 Hz, 1
H), 6.54 (d, J ) 9.2 Hz, 1 H), 6.09 (d, J ) 15.6 Hz, 1 H), 5.45 (d, J
) 10.0 Hz, 1 H), 4.80 (dd, J ) 9.2, 5.2 Hz, 1 H), 4.48 (dd, J ) 10.2,
5.2 Hz, 1 H), 3.63 (dq, J ) 6.4, 2.2 Hz, 1 H), 3.51 (dd, J ) 7.4, 2.2
Hz, 1 H), 2.93 (dq, J ) 10.2, 6.4 Hz, 1 H), 2.64-2.50 (m, 1 H), 1.62
(s, 3 H), 1.44 (s, 3 H), 1.36 (s, 3 H), 1.09 (d, J ) 6.4 Hz, 3 H), 1.06
(s, 9 H), 0.99 (t, J ) 7.6 Hz, 9 H), 0.98 (d, J ) 6.4 Hz, 3 H), 0.91 (d,
J ) 6.4 Hz, 3 H), 0.74-0.66 (q, J ) 7.6 Hz, 6 H); IR (CHCl3) 1680,
1660, 1620, 1590, 1460 cm-1; HRMS calcd for C38H53O5Br2Si2 (M+
- tBu), 805.1758, found 805.1751. Anal. Calcd for C42H62O5Br2Si2:
C, 58.46; H, 7.24. Found: C, 58.20; H, 7.38.
To a solution of the hydroxy enoate (5.57 g, 12.2 mmol) in CH2Cl2
(30 mL) at -78 °C was added 2,4-lutidine (2.84 mL, 24.5 mmol)
followed by triethylsilyl trifluoromethanesulfonate (4.16 mL, 18.4
mmol). The resulting solution was stirred at -78 °C for 15 min and
at 0 °C for 10 min at which time it was quenched with water (5 mL).
The solution was then poured into water (20 mL) and the aqueous phase
was extracted with CH2Cl2 (3 × 25 mL). The combined organic layers
were washed with water (20 mL), dried (MgSO4), filtered and
concentrated in vacuo. Purification of the crude product by flash
chromatography (silica gel, 5 × 15 cm, 5% ethyl acetate/hexanes) gave
enoate 40 (6.95 g, 99%): [R]D23 -8.1° (c 1.0, hexanes); 1H NMR (400
MHz, CDCl3) δ 7.7-7.6 (m, 4 H), 7.5-7.3 (m, 6 H), 6.54 (dd, J )
10.4, 1.2 Hz, 1 H), 4.1-4.3 (m, 2 H), 3.65-3.75 (m, 1 H), 3.56 (dd,
J ) 6.4, 6.0 Hz, 1 H), 2.7-2.6 (m, 1 H), 1.71 (d, J ) 1.2 Hz, 3 H),
1.28 (t, J ) 7.2 Hz, 3 H), 1.06 (s, 9 H), 1.0-0.9 (m, 15 H), 0.67 (q,
J ) 8.0 Hz, 4 H), 0.52 (q, J ) 8.0 Hz, 2 H); IR (neat) 1720, 1650,
1460 cm-1; HRMS (CI, NH3) calcd for C33H52O4Si2 (M+), 568.3390,
found 568.3396. Anal. Calcd for C33H52O4Si2: C, 69.74; H, 9.24.
Found: C, 69.45; H, 9.41.
Benzyl (4E,6Z,12E,14E)-(8S,9R,10S,16R,17S,18R)-6-Bromo-18-
[(tert-butyldiphenylsilyl)oxy]-8,9-(O-isopropylidene)-17-[(triethylsi-
lyl)oxy]-10,14,16-trimethylnonadec-4,6,12,14-tetraen-11-onoate (43).
An aqueous solution of TlOH (1.0 mL, 0.52 M, 0.52 mmol) was added
to a solution of dibromide 42 (382 mg, 0.44 mmol), vinyl boronic acid
24 (165 mg, 0.71 mmol), Pd2(dba)3 (60 mg, 0.066 mmol) and Ph3P
(200 mg, 0.76 mmol) in THF (30 mL). After being stirred for 1 h at
room temperature, the mixture was diluted with ether and filtered
through Celite. Removal of the solvent in vacuo gave an oily residue,
which was purified by silica gel chromatography (20% ether-hexanes
as eluent) to afford tetraene 43 (318 mg, 74%): [R]D20 +84.6° (c 1.02,
(E)-(4R,5S,6R)-6-[(tert-Butyldiphenylsilyl)oxy]-2,4-dimethyl-5-
[(triethylsilyl)oxy]hept-2-enal (26). To a -78 °C solution of enoate
40 (1.0 g, 1.8 mmol) in CH2Cl2 (10 mL) was added DIBAL-H (4.4
mL, 1 M in THF, 4.4 mmol). The mixture was stirred at -78 °C for
3 h, then was allowed to warm to ambient temperature. The solution
was then diluted with MeOH (1 mL), aqueous Rochelle’s salt solution
(50 mL) and Et2O (50 mL). The aqueous phase was separated and
extracted with additional Et2O (2 × 25 mL). The combined extracts
were dried (Na2SO4), filtered and concentrated in vacuo. Purification
of the crude product by flash chromatography (30 mm silica gel column)
using a gradient of 5%-30% Et2O in hexanes provided the corre-
1
CHCl3); H NMR (400 MHz, CDCl3) δ 7.65-7.60 (m, 4 H), 7.50-
7.30 (m, 11 H), 7.07 (d, J ) 15.6 Hz, 1 H), 6.25-6.15 (m, 2 H), 6.10
(d, J ) 15.6 Hz, 1 H), 5.95 (d, J ) 9.4 Hz, 1 H), 5.43 (d, J ) 10.4 Hz,
1 H), 5.15 (s, 2 H), 5.14 (dd, J ) 9.4, 5.2 Hz, 1 H), 4.49 (dd, J ) 10.2,
5.2 Hz, 1 H), 3.64 (dq, J ) 6.4, 2.8 Hz, 1 H), 3.51 (dd, J ) 8.0, 2.8
Hz, 1 H), 2.94 (dq, J ) 10.2, 6.4 Hz, 1 H), 2.60-2.50 (m, 5 H), 1.62
(s, 3 H), 1.45 (s, 3 H), 1.38 (s, 3 H), 1.07 (s, 9 H), 1.05 (d, J ) 6.4 Hz,
3 H), 1.00 (t, J ) 8.0 Hz, 9 H), 0.98 (d, J ) 6.4 Hz, 3 H), 0.91 (d, J
) 6.4 Hz, 3 H), 0.69 (q, J ) 8.0 Hz, 6 H); IR (CHCl3) 1735, 1680,
1655, 1540, 1460 cm-1; HRMS calcd for C51H69O5Si2Br (M+
-
sponding allylic alcohol (808 mg, 84%): [R]22 -31.6° (c ) 1.00,
C3H6O2), 896.3866, found 896.3846. Anal. Calcd for C54H75O7-
BrSi2: C, 66.71; H, 7.78. Found: C, 66.46; H, 8.00.
D
1
CHCl3); H NMR (400 MHz, CDCl3) δ 7.66-7.65 (m, 4 H), 7.45-
7.43 (m, 2 H), 7.43-7.41 (m, 4 H), 4.92 (dd, J ) 10.2, 0.8 Hz, 1 H),
3.74-3.70 (m, 3 H), 3.51 (dd, J ) 8.2, 2.2 Hz, 1 H), 2.35-2.30 (m,
1 H), 1.45 (d, J ) 1.2 Hz, 3 H), 1.06 (s, 9 H), 0.99 (t, J ) 7.9 Hz, 9
H), 0.93 (d, 6.2 Hz, 3 H), 0.89 (d, 6.5 Hz, 3 H), 0.75-0.68 (m, 6 H);
IR (neat) 3070, 1430, 1260 cm-1; HRMS for C31H50O3Si2 (M+) calcd
526.3298, found 526.3339. Anal. Calcd for C31H50O3Si2: C, 70.67;
H, 9.56. Found: C, 70.63; H, 9.28.
(4E,6Z,12E,14E)-(8S,9R,10S,16R,17S,18R)-6-Bromo-18-[(tert-bu-
tyldiphenylsilyl)oxy]-17-hydroxy-8,9-(O-isopropylidene)-10,14,16-
trimethylnonadec-4,6,12,14-tetraen-11-onoic Acid (44). A solution
of 43 (280 mg, 0.29 mmol) and LiOH (14 mg, 0.58 mmol) in DME (4
mL) and water (1 mL) was stirred for 3 h at room temperature. The
reaction mixture was acidified with 0.5 N HCl, diluted with ether and
poured into brine. The aqueous layer was separated and extracted with
ether. The combined organic layers were washed with brine and dried
over Na2SO4. Removal of the solvent in Vacuo gave an oily residue,
which was purified by silica gel chromatography (80% ether-hexanes
as eluent) to afford the carboxylic acid (177 mg, 70%): [R]D20 +93.3°
A solution of the above allylic alcohol (436 mg, 0.83 mmol) in CH2-
Cl2 (3 mL) was treated with MnO2 (1.43 g, 16.5 mmol; added in small
portions over 1 h). The mixture was stirred at ambient temperature
for 48 h, then was filtered through Celite (3 × 20 mL of CH2Cl2).
Concentration of the filtrate in Vacuo, and purification of the crude
product by flash chromatography (silica gel, 5:1 hexanes-EtOAc)
provided R,â-unsaturated aldehyde 26 (407 mg, 94%): Rf 0.7 (5:1
1
(c 0.97, CHCl3); H NMR (400 MHz, CDCl3) δ 7.65-7.60 (m, 4 H),
7.50-7.30 (m, 6 H), 7.06 (d, J ) 15.4 Hz, 1 H), 6.25-6.15 (m, 2 H),
6.09 (d, J ) 15.4 Hz, 1 H), 5.96 (d, J ) 9.4 Hz, 1 H), 5.42 (d, J )10.0
Hz, 1 H), 5.15 (dd, J ) 9.4, 5.4 Hz, 1 H), 4.48 (dd, J ) 10.2, 5.4 Hz,
1 H), 3.63 (dq, J ) 6.6, 2.8 Hz, 1 H), 3.50 (dd, J ) 7.6, 2.8 Hz, 1 H),
2.89 (dq, J ) 10.2, 6.6 Hz, 1 H), 2.62-2.50 (m, 5 H), 1.61 (s, 3 H),
1.44 (s, 3 H), 1.37 (s, 3 H), 1.05 (s, 9 H), 1.04 (d, J ) 6.6 Hz, 3 H),
0.98 (t, J ) 7.8 Hz, 9 H), 0.97 (d, J ) 6.6 Hz, 3 H), 0.90 (d, J ) 6.6
Hz, 3 H), 0.69 (q, J ) 7.8 Hz, 6 H); IR (CHCl3) 3500, 1715, 1680,
1655, 1590 cm-1. Anal. Calcd for C47H69O7BrSi2: C, 63.99; H, 7.88.
Found: C, 64.07; H, 7.97.
22
1
hexanes-EtOAc); [R]D -19.8° (c 1.08, CHCl3); H NMR (CDCl3) δ
9.15 (s, 1 H), 7.65-7.60 (m, 4 H), 7.45-7.30 (m, 6 H), 5.98 (dd, J )
10.4, 1.2 Hz, 1 H), 3.56-3.60 (m, 2 H), 2.80-2.65 (m, 1 H), 1.58 (d,
J ) 1.2 Hz, 3 H), 1.06 (s, 9 H), 0.95-1.05 (m, 15 H), 0.75-0.65 (m,
6 H); IR (CHCl3) 3020, 2880, 1685, 1640 cm-1; HRMS calcd for
C27H39O3Si2 (M+ - tBu), 467.2437, found 467.2423. Anal. Calcd for
C31H48O3Si: C,70.94; H, 9.22. Found: C, 70.80; H, 9.37.
(5E,7E)-(2R,3S,4R,10S,11R,12S)-14,14-Dibromo-2-[(tert-butyl-
diphenylsilyl)oxy]-11,12-(O-isopropylidene)-4,6,10-trimethyl-3-[(tri-
ethylsilyl)oxy]tetradec-5,7,13-trien-9-one (42). A 1 M THF solution
of LiN(TMS)2 (1.0 mL, 1.0 mmol) was added to a -78 °C solution of
methyl ketone 25 (300 mg, 0.84 mmol) in THF (7 mL). The solution
was stirred for 15 min at -78 °C, then a solution of enal 26 (402 mg,
0.77 mmol) in THF (2 mL) was added. The mixture was stirred for
15 min at -78 °C before addition of Ac2O (160 µL, 1.7 mmol). The
A mixture of the above carboxylic acid (158 mg, 0.18 mmol) and
1M aq. HF (360 µL, 0.36 mmol) in acetonitrile (10 mL) was stirred in
an ice bath for 5 h. The reaction mixture was diluted with ether and
poured into brine. The aqueous layer was separated and extracted with
ether. The combined organic layers were washed with brine, dried
over Na2SO4, filtered, and concentrated in Vacuo. The crude product
was purified by silica gel chromatography (ether as eluent) to afford

+
+
7512 J. Am. Chem. Soc., Vol. 118, No. 32, 1996
Roush et al.
seco acid 44 (124 mg, 90%): [R]D20 +79.6° (c 1.04, CHCl3); 1H NMR
(400 MHz, CDCl3) δ 7.60-7.65 (m, 4 H), 7.45-7.30 (m, 6 H), 7.00
(d, J ) 15.6 Hz, 1 H), 6.25-6.15 (m, 2 H), 6.10 (d, J ) 15.6 Hz, 1
H), 5.95 (d, J ) 9.4 Hz, 1 H), 5.16 (d, J ) 11.6 Hz, 1 H), 5.13 (dd, )
9.4, 5.2 Hz, 1 H), 4.46 (dd, J ) 10.2, 5.2 Hz, 1 H), 3.69 (dq, J ) 6.4,
2.8 Hz, 1 H), 3.28 (dd, J ) 8.8, 2.8 Hz, 1 H), 2.85 (dq, J ) 10.2, 6.4
Hz, 1 H), 2.62-2.40 (m, 5 H), 1.68 (s, 3 H), 1.43 (s, 3 H), 1.36 (s, 3
H), 1.08 (s, 9 H), 1.02 (d, J ) 6.4 Hz, 3 H), 1.00 (d, J ) 6.4 Hz, 3 H),
0.97 (d, J ) 6.4 Hz, 3 H); IR (CHCl3) 3500, 1715, 1680, 1655, 1625,
1595 cm-1; HRMS calcd for C37H46O7BrSi (M+ - tBu) 709.2196, found
709.2215. Anal. Calcd for C41H55O7BrSi: C, 64.13; H, 7.22.
Found: C, 63.91; H, 7.11.
hexanes), affording 46 (2.0 mg, 5%), 45 (12.4 mg, 32%), and macrolide
22
1
22 (14.8 mg, 38%): [R]D +65.1° (c 1.00, CHCl3); H NMR (400
MHz, CDCl3) δ 7.70-7.760 (m, 4 H), 7.50-7.35 (m, 6 H), 6.90 (d, J
) 15.8 Hz, 1 H, H-13), 6.20 (d, J ) 15.8 Hz, 1 H, H-12), 5.95-6.09
(m, 3 H, H-4, H-5 and H-7), 5.46 (d, J ) 10.0 Hz, 1 H, H-15), 5.07
(dd, J ) 8.4, 5.4 Hz, 1 H, H-8), 4.98 (dd, J ) 8.8, 6.4 Hz, 1 H, H-17),
4.39 (dd, J ) 7.6, 5.4 Hz, 1 H, H-9), 3.84 (dq, J ) 8.8, 6.8 Hz, 1 H,
H-18), 3.38 (quint, J ) 6.8 Hz, 1 H, H-10), 3.27 (m, 1 H, H-16), 2.20-
2.70 (m, 4 H), 1.82 (s, 3 H), 1.51 (s, 3 H), 1.38 (s, 3 H), 1.21 (d, J )
6.8 Hz, 3 H), 1.06 (s, 9 H), 0.98 (d, J ) 6.8 Hz, 3 H), 0.86 (d, J ) 6.8
Hz, 3 H); IR (CHCl3) 2980, 2940, 2860, 1735, 1665, 1630, 1460, 1430,
1375, 1160, 1110, 1075, 1035, 910 cm-1; HRMS for C37H44O6BrSi
t
(M+ - Bu), calcd 691.2092, found 691.2090.
Transannular Diels-Alder Reaction of 22 via Macrolactonization
of 44: [1E,3R,4S(R),8aR,10aR,11S,12R,13S*,14R,14aS,14bR]-3,4,7,8,-
8a,10a,11,12,13,14a,14b-Undecahydro-10-bromo-4-[1-[(1,1-dimeth-
ylethyl)diphenylsilyl]oxyethyl]-11,12-(O-isopropylidene)-1,3,13-tri-
methylnaphth[2,1-e]oxecin-6(7H),14-dione (45) and [1E,3R,4S(R),-
8aR,10aR,11S,12R,13R*,14R,14aS,14bR]-3,4,7,8,8a,10a,11,12,13,-
14a,14b-Undecahydro-10-bromo-4-[1-[(1,1-dimethylethyl)diphenyl-
silyl]oxyethyl]-11,12-(O-isopropylidene)-1,3,13-trimethylnaphth[2,1-
e]oxecin-6(7H),14-dione (46). A mixture of seco acid 44 (50 mg, 0.065
mmol), trichlorobenzoyl chloride (33 µL, 0.21 mmol) and triethylamine
(63 µL, 0.45 mmol) in THF (1.0 mL) was stirred for 3 h at room
temperature. The solvent was removed in Vacuo, and the residue was
dissolved in toluene (20 mL). This solution was added over 15 h to a
100 °C solution of DMAP (50 mg, 0.41 mmol) in toluene (20 mL).
The mixture was stirred for an additional 5 h, then was diluted with
ether and poured into brine. The aqueous layer was separated and
extracted with ether. The combined organic layers were washed with
0.5 N HCl and brine, and dried over Na2SO4. Removal of the solvent
in Vacuo gave an oily residue, which was purified by silica gel
chromatography (20% ether-hexanes as eluent) to afford 46 (6.5 mg,
13%). Further elution with 25% ether-hexanes afforded the desired
transannular cycloadduct 45 (32.0 mg, 66%).
Transannular Diels-Alder Reaction of 22. A solution of 22 (13.0
mg) in toluene was heated at 100 °C for 20 h. Removal of the solvent
in vacuo gave an oily residue, which was purified by silica gel
chromatography(25% ether-hexanes as eluent) to afford 45 (11.1 mg,
85%) as the only observed product.
Intramolecular Diels-Alder Reaction of Seco Ester 43: [1S,*-
(1′E,3′R,4′S,5′R)-2R,4aR,5S,6R,7S,8aR*]-1,2,4a,5,6,7,8a-Heptahydro-
2-(3′′-benzyloxycabonylprop-1′′-yl)-4-bromo-1-[5′-[(1,1-dimethylethyl)-
diphenylsilyl]oxy-4′-(triethylsilyl)oxy-1′,3′-dimethyl-1′-hexenyl]-6,7-
(O-isopropylidene)-naphthalen-8-one (47) and [1R*(1′E,3′R,4′S,5′R)-
2R,4aR,5S,6R,7S,8aS*]-1,2,4a,5,6,7,8a-Heptahydro-2-(3′′-benzyl-
oxycabonylprop-1′′-yl)-4-bromo-1-[5′-[(1,1-dimethylethyl)diphenyl-
silyl]oxy-4′-(triethylsilyl)oxy-1′,3′-dimethyl-1′-hexenyl]-6,7-(O-iso-
propylidene)naphthalen-8-one (48). A solution of 43 (10.0 mg) in
toluene (10 mL) was heated at 110 °C for 24 h. Removal of the solvent
in Vacuo gave an oily residue, which was purified by silica gel
chromatography. Elution of the column with 10% ether-hexanes
afforded 48 (2.7 mg, 27%), and elution with 12% ether-hexanes
afforded 47 (5.6 mg, 56%).
23
1
Data for 47: [R]D +29.3° (c 0 .92, CHCl3); H NMR (400 MHz,
benzene-d6) δ 7.85-7.80 (m, 4 H), 7.25-7.15 (m, 11 H), 6.21 (dd, J
) 5.2, 1.8 Hz, 1 H, H-5), 5.02 (d, J ) 12.4 Hz, 1 H, benzyl), 4.98 (d,
J ) 12.4 Hz, 1 H, benzyl), 4.56 (d, J ) 9.2 Hz, 1 H, H-15), 4.23 (t, J
) 5.0 Hz, 1 H, H-9), 4.18 (dd, J ) 9.6, 5.0 Hz, 1 H, H-8), 4.09 (dq,
J ) 6.4, 2.2 Hz, 1 H, H-18), 3.67 (dd, J ) 7.2, 2.2 Hz, 1 H, H-17),
2.69 (ddt, J ) 12.0, 9.6, 1.8 Hz, 1 H, H-7), 2.58 (m, 1 H, H-16), 2.32
(t, J ) 12.0 Hz, 1 H, H-12), 2.30-1.85 (m, 7 H, H-4, H-2, H-3, H-10
and H-13), 1.43 (s, 3 H), 1.36 (s, 3 H), 1.30 (s, 3 H), 1.24 (d, J ) 6.4
Hz, 3 H), 1.22 (d, J ) 6.4 Hz, 3 H), 1.21 (s, 9 H), 1.10 (t, J ) 7.6 Hz,
9 H), 1.04 (d, J ) 6.4 Hz, 3 H), 0.83 (q, J ) 7.6 Hz, 6 H); IR (CHCl3)
1735, 1575, 1455 cm-1; HRMS for C50H66O7BrSi2 (M+ - tBu), calcd
915.3215, found 915.3526.
23
Data for 45: Rf 0.4 (2:1 hexanes-ether); [R]D +41.6° (c 0.99,
CHCl3); 1H NMR (400 MHz, benzene-d6) δ 7.65-7.85 (m, 4 H), 7.25-
7.15 (m, 6 H), 5.72 (d, J ) 3.4 Hz, 1 H, H-5), 5.35 (t, J ) 6.8 Hz, 1
H, H-17), 4.97 (d, J ) 6.8 Hz, 1 H, H-15), 4.13 (dd, J ) 8.2, 4.8 Hz,
1 H, H-8), 4.07 (quint, J ) 6.8 Hz, 1 H, H-18), 3.95 (t, J ) 4.8 Hz, 1
H, H-9), 3.10 (sextet, J ) 6.8 Hz, 1 H, H-16), 2.85 (dd, J ) 8.2, 4.8
Hz, 1 H, H-7), 2.65 (dd, J ) 11.6, 4.8 Hz, 1 H, H-12), 2.47 (dq, J )
6.8, 4.8 Hz, 1 H, H-10), 2.04 (ddd, J ) 11.6, 5.2, 3.4 Hz, 1 H, H-4),
1.98 (t, J ) 11.6 Hz, 1 H, H-13), 1.90 (m, 2 H, H-2), 1.84 (m, 2 H,
H-3), 1.67 (s, 3 H), 1.43 (s, 3 H), 1.22 (d, J ) 6.8 Hz, 3 H), 1.21 (s,
3 H), 1.20 (d, J ) 6.8 Hz, 3 H), 1.13 (s, 9 H), 1.04 (d, J ) 6.8 Hz, 3
H); nOe experiments in benzene-d6: irradiation at H-10 caused a 4.6%
enhancement of H-9 and a 3.7% enhancement of H-13; irradiation at
H-12 resulted in a 4.9% enhancement of H-7; IR (CHCl3) 3020, 2930,
2860, 1720, 1445, 1425, 1380, 1370, 1135, 1110, 1065, 995, 955, 860,
820 cm-1; HRMS for C37H44O6BrSi (M+ - tBu), calcd 693.2070, found
693.2109. Anal. Calcd for C41H53O6BrSi: C, 65.67; H, 7.12.
Found: C, 65.66; H 7.34.
23
1
Data for 48: [R]D +31.2° (c 0.42, CHCl3); H NMR (400 MHz,
benzene-d6) δ 7.85-7.80 (m, 4 H), 7.25-7.15 (m, 11 H), 6.01 (dd, J
) 3.2, 0.8 Hz, 1 H, H-5), 5.01 (d, J ) 12.4 Hz, 1 H, benzyl), 4.97 (d,
J ) 12.4 Hz, 1 H, benzyl), 4.72 (d, J ) 9.2 Hz, 1 H, H-15), 4.54 (dd,
J ) 7.2, 5.2 Hz, 1 H, H-8), 4.21 (t, J ) 5.2 Hz, 1 H, H-9), 3.98 (dq,
J ) 6.4, 2.8 Hz, 1 H, H-18), 3.64 (dd, J ) 7.2, 2.8 Hz, 1 H, H-17),
2.99 (dd, J ) 7.2, 5.2 Hz, 1 H, H-7), 2.62 (dd, J ) 6.4, 5.2 Hz, 1 H,
H-10), 2.58 (dd, J ) 9.6, 5.2 Hz, 1 H, H-12), 2.55 (ddd, J ) 9.6, 7.2,
3.2 Hz, 1 H, H-4), 2.18 (t, J ) 9.6 Hz, 1 H, H-13), 2.15-1.80 (m, 5
H, H-2, H-3 and H-10), 1.43 (s, 3 H), 1.36 (s, 6 H, 2 Me’s), 1.25 (d,
J ) 6.4 Hz, 3 H), 1.22 (d, J ) 6.4 Hz, 3 H), 1.19 (s, 9 H), 1.09 (t, J
) 7.6 Hz, 9 H), 0.92 (d, J ) 6.4 Hz, 3 H), 0.81 (q, J ) 7.6 Hz, 6 H);
23
Data for 46: Rf 0.5 (2:1 hexanes-ether); [R]D -30.8° (c 0.51
CHCl3); 1H NMR (400 MHz, benzene-d6) δ 7.85-7.65 (m, 4 H), 7.25-
7.15 (m, 6 H), 5.77 (d, J ) 3.4 Hz, 1 H, H-5), 5.42 (t, J ) 6.6 Hz, 1
H, H-17), 4.82 (d, J ) 6.6 Hz, 1 H, H-15), 4.20 (dd, J ) 8.8, 6.8 Hz,
1 H, H-8), 4.90 (quint, J ) 6.6 Hz, 1 H, H-18), 3.65 (dd, J ) 10.4, 6.8
Hz, 1 H, H-9), 3.23 (sextet, J ) 6.6 Hz, 1 H, H-16), 2.78 (dd, J ) 8.8,
5.2 Hz, 1 H, H-7), 2.22 (dd, J ) 12.8, 5.2 Hz, 1 H, H-12), 2.17 (dq,
J ) 10.4, 6.6 Hz, 1 H, H-10), 2.02 (ddd, J ) 12.8, 5.2, 3.4 Hz, 1 H,
H-4), 2.0-1.60 (m, 5 H, H-13, H-2, H-3), 1.85 (s, 3 H), 1.49 (s, 3 H),
1.23 (d, J ) 6.6 Hz, 3 H), 1.21 (s, 3 H), 1.20 (d, J ) 6.6 Hz, 3 H),
1.15 (s, 9 H), 1.02 (d, J ) 6.6 Hz, 3 H); IR (CHCl3) 2960, 2930, 2850,
1720, 1455, 1425, 1375, 1260, 1250, 1165, 1105, 1070, 1050, 980,
950, 870, 830, 710 cm-1; HRMS for C37H44O6BrSi (M+ - tBu), calcd
693.2070, found 693.2059.
t
IR (CHCl3) 1720, 1455 cm-1; HRMS for C50H66O7BrSi2 (M+ - Bu),
calcd 915.3215; found 915.3468.
Acknowledgment. This research was supported by the
National Institute of General Medical Sciences (GM 26782).
We are also grateful to the Sankyo Co. for providing financial
support for K.K.
Synthesis of [5E,7Z,13E,15E,9S,10R,11S,17R,18S(R)]-7-Bromo-
18-[1-[(1,1-dimethylethyl)diphenylsilyl]oxyethyl]-9,10-(O-isoprope-
nyl)-11,15,17-trimethyloxacyclooctadec-5,7,13,15-tetraen-2,12-di-
one (Macrolide 22). The macrolactonization of 44 (40 mg) was
performed as described in the preceding experiment, with the exception
that the reaction temperature was 80 °C rather than 100 °C. The crude
product was purified by silica gel chromatography (20-30% ether-
Supporting Information Available: Experimental proce-
dures for the synthesis of macrolide 13 from precursors 36, 40,
and 49 (9 pages). See any current masthead page for ordering
and Internet access instructions.
JA9607796