Fully stereocontrolled total syntheses of the prostacyclin analogues 16S-iloprost and 16S-3-oxa-iloprost by a common route, using alkenylcopper-azoalkene conjugate addition, asymmetric olefination, and allylic alkylation

Article abstract of DOI:10.1021/ja0558037

In this article we describe fully stereocontrolled total syntheses of 16S-iloprost (16S-2), the most active component of the drugs Ilomedin and Ventavis, and of 16S-3-oxa-iloprost (16S-3), a close analogue of 16S-2 having the potential for a high oral act

Full text of DOI:10.1021/ja0558037

Published on Web 11/25/2005
Fully Stereocontrolled Total Syntheses of the Prostacyclin
Analogues 16S-Iloprost and 16S-3-Oxa-Iloprost by a Common
Route, Using Alkenylcopper-Azoalkene Conjugate Addition,
Asymmetric Olefination, and Allylic Alkylation
Guido J. Kramp, Mikhail Kim, Hans-Joachim Gais,* and Cornelia Vermeeren
Contribution from the Institut fu¨r Organische Chemie der Rheinisch-Westfa¨lischen Technischen
Hochschule (RWTH) Aachen, Landoltweg 1, D-52056 Aachen, Germany
Received August 24, 2005; E-mail: Gais@RWTH-Aachen.de
Abstract: In this article we describe fully stereocontrolled total syntheses of 16S-iloprost (16S-2), the most
active component of the drugs Ilomedin and Ventavis, and of 16S-3-oxa-iloprost (16S-3), a close analogue
of 16S-2 having the potential for a high oral activity, by a new and common route. The key steps of this
route are (1) the establishment of the complete C13-C20 ω side chain of the target molecules through a
stereoselective conjugate addition of the alkenylcopper derivative 9 to the bicyclic C6-C12 azoalkene 10
with formation of hydrazone 8, (2) the diastereoselective olefination of ketone 7 with the chiral phosphoryl
acetate 39, and (3) the regio- and stereoselective alkylation of the allylic acetate 43 with cuprate 42. These
measures allowed the 5E,15S,16S-stereoselective synthesis of 16S-2 and 16S-3, a goal which had
previously not been achieved. Azoalkene 10 was obtained from the achiral bicyclic C6-C12 ketone 11 as
previously described by using as key step an enantioselective deprotonation. The configuration at C16 of
ω-side chain building block 9 has been installed with high stereoselectivity by the oxazolidinone method
and that at C15 by a diastereoselective oxazaborolidine-catalyzed reduction of the C13-C20 ketone 23
with catecholborane. Surprisingly, a high diastereoselectivity in the reduction of 23 was only obtained by
using 2 equiv of oxazaborolidine 24. Application of substoichiometric amounts of 24 resulted in irreproducible
diastereoselectivities ranging from very high to nil.
cal activity, respectively, to iloprost.4 Iloprost has already been
approved as Ilomedin for the treatment of peripheral arterial
occlusive disease, severe thrombo-angiitis obliterans involving
a high risk of amputation, and Raynaud’s disease.5 Moreover,
iloprost has recently found approval as Ventavis for the
treatment of pulmonary arterial hypertension, a highly debilitat-
ing and potentially fatal disease.6,7 However, Ilomedin and
Ventavis are not single isomer drugs. They contain ap-
proximately 1:1 mixtures of 16S-2 and its R-configured isomer
16R-2, which has, however, a much lower biological activity
than 16S-2.1c,8,9 For example, 16R-2 is 20 times less potent than
16S-2 in inhibiting collagen-induced platelet aggregation.8,9
Although the synthesis of iloprost developed by the Schering
group provides access to the drug in sufficient quantities, it is
1. Introduction
Prostacyclin (1) (Figure 1) is the most potent endogenous
inhibitor of blood platelet aggregation and a strong vasodilator.1
These features make prostacyclin per se an attractive drug for
a therapy of cardiovascular diseases. However, the medicinal
application of prostacyclin is severely hampered by its chemical
and metabolic instability as indicated by a half-life of only 3
min under physiological conditions.
Intensive efforts to find stable and potent analogues2 led the
Schering group to the imaginative and highly successful design
of the carbocyclic prostacyclin analogue iloprost (2).1c,3 Re-
placement of the O atom at the 6a position of 1 by a methylene
group, the introduction of a methyl group at C16, and a triple
bond at C18,C19 conveyed high chemical stability and biologi-
(3) (a) Skuballa, W.; Vorbru¨ggen, H. Angew. Chem. 1981, 93, 1080-1081;
Angew. Chem., Int. Ed. Engl. 1981, 20, 1046-1047. (b) Schenker, K. V.;
Philipsborn, von, W.; Evans, C. A.; Skuballa, W.; Hoyer, G.-A. HelV. Chim.
Acta 1986, 69, 1718-1727. (c) Skuballa, W.; Scha¨fer, M. Nachr. Chem.
Tech. Lab. 1989, 37, 584-588. (d) Petzold, K.; Dahl, H.; Skuballa, W.;
Gottwald, M. Liebigs Ann. Chem. 1990, 1087-1091.
(1) (a) Moncada, S.; Gryglewski, R. J.; Bunting, S.; Vane, J. R. Nature 1976,
263, 663-665. (b) Prostacyclin; Vane, J. R., Bergstro¨m, S., Eds.; Raven
Press: New York, 1979. (c) Prostacyclin and its Stable Analogue Iloprost;
Gryglewski, R. J., Stock, G., Eds.; Springer-Verlag: Berlin, 1987. (d)
Platelets and Their Factors; Bruchhausen, F., Walter, U., Eds.; Springer-
Verlag: Berlin, 1997. (e) De Leval, X.; Hanson, J.; David, J.-L.; Masereel,
B.; Pitorre, B.; Dogne, J.-M. Curr. Med. Chem. 2004, 11, 1243-1252. (f)
Hildebrand, M. Prostaglandins 1992, 44, 431-442. (g) Janssen, M. C. H.;
Wollersheim, H.; Kraus, C.; Hildebrand, M.; Watson, H. R.; Thien, T.
Prostaglandins Other Lipid Mediators 2000, 60, 153-160. (h) Schermuly,
R. T.; Schulz, A.; Ghofrani, H. A.; Meidow, A.; Rose, F.; Roehl, A.;
Weissmann, N.; Hildebrand, M.; Kurz, J.; Grimminger, F.; Walmrath, D.;
Seeger, W. J. Pharmacol. Exp. Therap. 2002, 303, 741-745.
(4) Because of clarity the prostacyclin numbering is used for iloprost, 3-oxa-
iloprost, and all building blocks through out this paper except in the
experimental part where the numbering of the compounds follows the
nomenclature rules.
(5) (a) Wilhelm, W.; Grundmann, U. Anaesthesist 2004, 53, 745-747. (b)
Veroux, P.; Veroux, M.; Macarone, M.; Bonanno, M. G.; Tumminelli, M.
G. Curr. Ther. Res. 2004, 65, 255-265. (c) Marasini, B.; Massarotti, M.;
Bottasso, B.; Coppola, R.; Del Papa, N.; Maglione, W.; Comina, D. P.;
Maioli, C. Scand. J. Rheumatol. 2004, 33, 253-256. (d) De Leval, X.;
Hanson, J.; David, J.-L.; Masereel, B.; Pitorre, B.; Dogne, J.-M. Curr. Med.
Chem. 2004, 11, 1243-1252.
(2) (a) Nickolson, R. C.; Town, M. H.; Vorbru¨ggen, H. Med. Res. ReV. 1985,
5, 1-53. (b) Schinzer, D. Nachr. Chem. Tech. Lab. 1989, 37, 734-738.
(c) Collins, P. W.; Djuric, S. W. Chem. ReV. 1993, 93, 1533-1564.
9
17910
J. AM. CHEM. SOC. 2005, 127, 17910-17920
10.1021/ja0558037 CCC: $30.25 © 2005 American Chemical Society

Total Syntheses of the Prostacyclin Analogue
A R T I C L E S
of the R-side chain leading to the formation of the inactive
tetranoriloprost. Therefore, Ilomedin and Ventavis have to be
administered by intravenous infusion and inhalation, respec-
tively. To alleviate the considerable costs, efforts, and possible
complications associated with an intravenous delivery it would
be highly desirable to have in addition to iloprost an analogue
thereof at the disposal, which not only has a similar biological
profile but also a long lasting oral activity. A most attractive
candidate that should meet these criteria is 16S-3-oxa-iloprost
(16S-3). The O atom in the 3 position of 16S-3 will provide for
a higher metabolic stability, since the â-oxidation of the R-side
chain is impeded and the 16S-configuration is expected to ensure
a high biological activity. These expectations are born out by
the high metabolic stability1f of the antimetastatic 3-oxa-
carbacyclin derivative cicaprost (4)10,11 and the much higher
biological activity of the 16S-configured iloprost (16S-2) as
compared to the 16R-configured isomer 16R-2.8,9 In 1982, a
synthesis of 3-oxa-iloprost had been disclosed in patents.12 This
synthesis, however, which closely parallels that of iloprost until
a late stage, is nonstereoselective concerning C5, C15, and C16.
It finally yields a nearly 1:1 mixture of 16S-3 and 16R-3, both
of which are difficult to separate. Because of (1) the considerable
medicinal importance of iloprost, (2) the potential medicinal
prospects of 3-oxa-iloprost,13 and (3) the deficiencies of the
existing syntheses, we were seeking fully stereocontrolled
syntheses of 16S-2 and 16S-3, a goal which had not been
achieved yet, by a new route which would give, by a diversion
at a late stage, access to both prostacyclin analogues. Herein
we describe fully stereocontrolled total syntheses of 16S-2 and
16S-3 by a new and common route,14 the key elements of which
are the attachment of the complete ω-side chain to the bicyclo-
[3.3.0]octane skeleton through the conjugate addition of an
alkenylcopper compound to an azoalkene and the construction
of the R-side chains via an asymmetric olefination and a
stereoselective allylic alkylation.
Figure 1. Prostacyclin and its carbocyclic analogues 16S-iloprost, 16S-3-
oxa-iloprost, 16R-iloprost, and cicaprost.
nonstereoselective in regard to the configurations at C5, C15,
and C16, all three of which are crucial to its biological
activity.1c,3 It finally gives after the separation and discarding
of the less active C5 and C15 isomers a practically nonseparable
1:1 mixture of 16S-2 and 16R-2.1c,3 Thus the development of a
fully stereocontrolled synthesis of 16S-2, the most active
component of the drugs Ilomedin and Ventavis, would be of
considerable importance. Although iloprost is orally active, it
has only a relatively short duration of action because of a rapid
metabolization,1c,f-h the key step of which is the â-oxidation
2. Results and Discussion
2.1. Retrosynthetic Analysis. The first key feature of our
retrosynthetic analysis of 16S-2 and 16S-3 is the coupling of a
bicyclic C6-C12 building block, the azoalkene 10,14 with a
C13-C20 ω-side chain building block, the alkenylcopper
compound 9, with the formation of the C12-C13 bond of the
(6) (a) Badesch, D. B.; McLaughlin, V. V.; Delcroix, M.; Vizza, C. D.;
Olschweski, H.; Sitbon, O.; Barst, R. J. J. Am. Col. Cardiol. 2004, 43,
56S-61S. (b) Olscheski, H.; Rose, F.; Schermuly, R.; Ghofrani, H. A.; Enke,
B.; Olschewski, A.; Seeger, W. Pharmacol. Therapeut. 2004, 102, 139-
153. (c) Nagaya, N. Am. J. CardioVasc. Drugs 2004, 4, 75-85. (d)
Olscheski, H.; Rose, F.; Schermuly, R.; Ghofrani, H. A.; Enke, B.;
Olschewski, A.; Seeger, W. Pharmacol. Therapeut. 2004, 102, 139-153.
(e) Goldsmith, D. R.; Wagstaff, A. J. Drugs 2004, 64, 763-773. (f)
Ghofrani, H. A.; Friese, G.; Discher, T.; Olschewski, H.; Schermuly, R.
T.; Weissmann, N.; Seeger, W.; Grimminger, F.; Lohmeyer, J. Eur. Respir.
J. 2004, 23, 321-326.
(10) (a) Skuballa, W.; Schillinger, E.; Stu¨rzebecher, C.-S., Vorbru¨ggen, H. J.
Med. Chem. 1986, 29, 313-315. (b) Stuerzebecher, C. S.; Haberey, M.;
Mueller, B.; Schillinger, E.; Schroeder, G.; Skuballa, W.; Stock, G.;
Vorgru¨ggen, H.; Witt, W. Prostaglandins 1986, 31, 95-109. (c) Stuerze-
becher, C. S.; Loge, O.; Schroeder, G.; Mueller, B.; Witt, W. Prog. Clin.
Biol. Res. 1987, 242, 425-432. (d) Schneider, M. R.; Schirner, M.; Lichtner,
R. B.; Graf, H. Breast Cancer Res. Treat. 1996, 38, 133-141. (e) Schirner,
M.; Kraus, C.; Lichtner, R. B.; Schneider. M. R.; Hildebrand, M. Prostagl.
Leukotr. Ess. Fatty Acids 1998, 58, 311-317. (f) Sava, G.; Bergamo, A.
Anticancer Res. 1999, 19, 1117-1124.
(11) Lerm, M.; Gais, H.-J.; Cheng, K.; Vermeeren, C. J. Am. Chem. Soc. 2003,
125, 9653-9667.
(12) (a) Skuballa, W.; Raduechel, B.; Schwarz, N.; Vorbru¨ggen, H.; Casals-
Stenzel, J.; Schillinger, E.; Town, M. H., 1982, EP 55208; Chem. Abstr.
1983, 98, 53513. (b) Skuballa, W.; Raduechel, B.; Vorbru¨ggen, H.;
Mannesmann, G.; Nieuweboer, B.; Town, M. H., 1983, DE 3221193; Chem.
Abstr. 1984, 101, 6931.
(7) Further potential applications of iloprost, which have been studied: (a)
Acute thromboebolic eVents: Cretel, E.; Disdier, P.; Chagmaud, C.;
Tournigand, P.; Harle, J. R.; Plette, J. C.; Weiller, P. J. Angiology 1998,
49, 929-936. (b) Bone marrow oedema syndrome: Aigner, N.; Petje, G.;
Steinboeck, G.; Schneider, W.; Krasny, C.; Landsiedl, F. J. Bone Joint
Surg. 2001, 83, 855-858. (c) Systemic sclerosis: Biasi, D.; Carletto, A.;
Caramaschi, P.; Zeminian, S.; Pacor, M. L. ReV. Rheum. Engl. Ed. 1998,
65, 745-750. (d) Dry macular degeneration of the eye: Geradino, L.;
Santoliquido, A.; Flore, R.; Dal Lago, A.; Gaetani, E.; Gasbarrini, A.;
Papaleo, P.; Abed, A.; Pola, R. J. Am. Geriatr. Soc. 2000, 48, 1350-1351.
(e) Diabetic neuropathy: Scindo, H.; Tarata, M.; Aida, K.; Onaya, T.
Prostaglandins 1991, 41, 85-96. (f) Diabetic nephropathy: Shindo, H.;
Tawata, M.; Yohomori, N.; Hosaka, Y.; Ohtaka, M. Diabetes Res. Clin.
Pract. 1993, 21, 115-122.
(13) Biological studies of 16S-3 have not been reported. It has been briefly stated,
without giving any details, that 3-oxa-iloprost shows a decreased receptor
affinity as compared to iloprost (Stu¨rzebecher, S.; Haberey, M.; Mu¨ller,
B.; Schillinger, E.; Schro¨der, G.; Skuballa, W.; Stock, G.; Vorbru¨ggen,
H.; Witt, W. Prostaglandins 1986, 31, 95-109). However, it appears that
the underlying experiments were done with the 1:1 mixture of 16S-3 and
16R-3, the 16R-isomer of which is expected to have a significantly lower
receptor affinity than the 16S-isomer.8,9
(8) Tsai, A.-l., Vijjeswarapu, H.; Wu, K. K. Biochim. Biophys. Acta 1988, 942,
220-226.
(9) Ashby, B. Prostaglandins 1992, 43, 255-261.
(14) Gais, H.-J.; van Bergen, M. J. Am. Chem. Soc. 2002, 124, 4321-4328.
9
J. AM. CHEM. SOC. VOL. 127, NO. 50, 2005 17911

A R T I C L E S
Kramp et al.
Scheme 1. Retrosynthetic Analysis of 16S-Iloprost and 16S-3-Oxa-Iloprost
target molecules (Scheme 1). This measure should not only
allow the stereoselective attachment of the ω-side chain to the
bicyclo[3.3.0]octane skeleton from the convex side in one step
but also provide a high degree of flexibility in regard to the
synthesis of derivatives with other ω-side chains which are
crucial to the biological activity of prostacyclin analogues.2
The known syntheses of 16S/R-312 and 16S/R-21c,3 start from
a C6-C13 building block and feature a stepwise construction
of the ω-side chain, which renders the stereoselective generation
of C15 and C16 difficult. We had previously carried out an
asymmetric synthesis of 13,14-dinor-inter-p-phenylene carba-
cyclin, a prostacyclin (1) analogue which carries a 1,4-
disubstituted phenyl group instead of the C13-C14 double bond
by connecting the complete ω-side chain to the bicyclo[3.3.0]-
octane skeleton through the conjugate addition of the corre-
sponding arylcopper compound to azoalkene 10.14 Thus it was
hoped that a stereoselective conjugate addition of the alkenyl-
copper derivative 9 to azoalkene 10 could also be achieved with
formation of hydrazone 8. However, nothing was known about
the reactivity and stereoselectivity of azoalkene 10 and azoalk-
enes in general toward alkenylcopper reagents. Despite their
potential for a nucleophilic derivatization of ketones at the
R-position only little is known about the reactivity of azoalkenes
toward organocopper reagents.15 Furthermore, since 9 has to
be enantio- and diastereomerically pure, it was of considerable
importance to see whether such a conjugate addition could be
efficiently carried out by using the two building blocks in a
ratio of approximately 1:1. A critical step of Scheme 1 could
be the conversion of hydrazone 8 to ketone 7 because of the
requirement of a selective cleavage of the hydrazone group in
the presence of the acetal group. The second key feature of the
retrosynthesis of 16S-2 and 16S-3 is the Horner-Wadsworth-
Emmons (HWE) olefination of 7 with a chiral phosphoryl
acetate, which should stereoselectively give ester E-6. An
asymmetric HWE olefination of this type16a,b,17 had been
successfully applied in the synthesis of cicaprost (4),11,17 3-oxa-
carbacyclin,16c,d and 3-oxa-isocarbacyclin.16d
The allylic alcohol 5 was planned to be the point of digression
of the retrosynthesis of 16S-2 and 16S-3. An etherification of 5
should give 16S-3.10a,11,16 We envisioned as the third new key
feature of the retrosynthesis of 16S-2 the establishment of the
R-side chain through a copper-mediated allylic alkylation of a
derivative of 5 with a suitable functionalized C1-C3-organo-
copper building block. Here the critical point could be the
achievement of both a high R-regioselectivity and a complete
retention of configuration of the double bond in the alkylation.
Literature was ambiguous as to the feasibility of such a
transformation.18 Generally the regio- and stereoselectivity of
the allylic alkylation with organocopper reagents are strongly
dependent on the structure of the allylic substrate, the leaving
(16) (a) Gais, H.-J.; Schmiedl, G.; Ball, W. A.; Bund, J.; Hellmann, G.;
Erdelmeier, I. Tetrahedron Lett. 1988, 29, 1773-1774. (b) Gais, H.-J.;
Schmiedl, G.; Ossenkamp, R. K. L. Liebigs Ann. Recl. 1992, 2419-2431.
(c) Ossenkamp, R. K. L.; Gais, H.-J. Liebigs Ann. Recl. 1997, 2433-2441.
(d) Vaulont, I.; Gais, H.-J.; Reuter N.; Schmitz, E.; Ossenkamp, R. K. L.
Eur. J. Org. Chem. 1998, 805-826.
(17) Rehwinkel, H.; Skupsch, J.; Vorbru¨ggen, H. Tetrahedron Lett. 1988, 29,
1775-1776.
(18) (a) Lipshutz, B. H.; Sengupta, S. Org. React. 1992, 41, 135-631. (b) Spino,
C.; Godbout, C.; Beaulieu, C.; Harter, M.; Mwene-Mbeja, T. M. Boisvert,
L. J. Am. Chem. Soc. 2004, 126, 13312-13319. (c) Delay, F.; Ohloff, G.
HelV. Chim. Acta 1979, 62, 369-377. (d) Suzuki, S.; Mori, F.; Takigawa,
T.; Ibata, K.; Ninagawa, Y.; Nishida, T.; Mizuno, M.; Tanaka, Y.
Tetrahedron Lett. 1983, 24, 5103-5106. (e) Tanis, S. P.; Chuang, Y.-H.;
Head, D. B. J. Org. Chem. 1988, 53, 4929-4938.
(15) (a) Sacks, C. E.; Fuchs, P. L. J. Am. Chem. Soc. 1975, 97, 7372-7374. (b)
Cacchi, S.; Felici, M.; Rosini, G. J. Chem. Soc., Perkin I 1977, 1260-
1263. (c) Cacchi, S.; La Torre, F.; Misiti, D. Chem. Ind. (Milan) 1978, 60,
715-716.
9
17912 J. AM. CHEM. SOC. VOL. 127, NO. 50, 2005

Total Syntheses of the Prostacyclin Analogue
A R T I C L E S
Scheme 2. Asymmetric Synthesis of Azoalkene 10
Scheme 3. Asymmetric Synthesis of Amide 19a
a Reagents and conditions: (I) (a) 1.35 equiv of NaN(SiMe3)2, THF,
15, -78 °C; (b) 1.35 equiv of 16, -78 °C, 2 h; (c) AcOH, -78 °C. (II)
EtOH, Ti(OEt)4, reflux, 10 h. (III) (a) 1.55 equiv of [Me(OMe)NH2]Cl, 18,
3.1 equiv of i-PrMgCl, THF, -20 °C; (b) H2O, NH4Cl, -20 °C; (c)
preparative HPLC.
Scheme 4. Synthesis and Chromatographic Resolution of Amide
rac-19a
group, and the reagent thus making predictions in the present
case difficult.
2.2. Asymmetric Synthesis of the C6-C12 Azoalkene.
Azoalkene 10 with 95% enantiomeric excess (ee) was obtained
from ketone 11, which is readily available on large scale,19,20
in four steps via 12, 13, and 14 in 55% overall yield (Scheme
2) as described previously.14
The key step of the synthesis is the enantioselective depro-
tonation of the meso-configured ketone 11 with the R,R-
configured chiral LiCl-complexed lithium amide, which is
readily available on large scale.21 The synthesis of the chloro
ketone 13 through a catalytic enantioselective chlorination of
ketone 11 by using the R,R-configured imidazolidine derivative
(20 mol %), o-NO2-C6H4CO2H (50 mol %) and N-chlorosuc-
cinimide (NCS) (2 equiv) according to a recently reported
method22 was not possible. Formation of 13 could not be
detected and ketone 11 was recovered.
a Reagents and conditions: (I) (a) LDA, THF, -78 °C; (b) 16. (II) See
Scheme 3. (III) HPLC, Daicel Chiralcel AD column (250 × 50 mm),
n-hexane/i-PrOH, 95:5, 500 mg of rac-19/per injection.
now of particular importance. It was planned to generate the
alkenylcopper compound 9 from stannane 27 (Scheme 5, vide
infra) by using the well-established tin-lithium exchange.23
Thus, a retrosynthetic analysis of 27 identified a C15-C20
subunit, the chiral N-methoxyamide 19, and a C13-C14 subunit,
the lithiated stannane 22. The joining of both subunits should
give ketone 23 without effecting its stereochemical integrity.11,24
A stereoselective reduction of 23 would deliver alcohol 26.
Previous results suggested that a highly diastereoselective
reduction of ketone 23 with an achiral reducing reagent would
be difficult to achieve.25 Thus a chiral reducing reagent had to
be applied. An oxazaborolidine-catalyzed reduction with a
borane26,27 was chosen since this method had already been
successfully employed in our and other laboratories for the
2.3. Asymmetric Synthesis of the C13-C20 Alkenylstan-
nane. The development of an efficient synthesis of the enantio-
and diastereomerically pure ω-side chain building block 9 was
(23) Johnson, C. R.; Penning, T. D. J. Am. Chem. Soc. 1988, 110, 4726-4735.
(24) For a review on the synthesis of ketones from N-methoxy-N-methylamides,
see: Mentzel, M.; Hoffmann, H. M. R. J. Prakt. Chem. 1997, 339, 517-
524.
(19) (a) Carceller, E.; Moyana, A.; Serratosa, F. Tetrahedron Lett. 1984, 25,
2031-2034. (b) Piers, E.; Karunaratne, V. Can. J. Chem. 1989, 67, 160-
164.
(20) (a) Dahl, H. 1989, DE 3816801; Chem. Abstr. 1989, 113, 23512. (b) Bertz,
S. H.; Cook, J. M.; Gawish, A.; Weiss, U. Org. Synth. 1986, 64, 27-38.
(21) (a) Hemmerle, H.; Gais, H.-J. Angew. Chem. 1989, 101, 362-365; Angew.
Chem., Int. Ed. Engl. 1989, 28, 349-351. (b) Vaulont, I.; Gais, H.-J.; Reuter
N.; Schmitz, E.; Ossenkamp, R. K. L. Eur. J. Org. Chem. 1998, 805-826.
(c) Gais, H.-J.; Ossenkamp, R. K. L. Liebigs Ann. Recl. 1997, 2433-2441.
(22) Marigo, M.; Bachmann, S.; Halland, N.; Bruanton, A.; Jørgensen, K. A.
Angew. Chem. 2004, 116, 5623-5626; Angew. Chem., Int. Ed. 2004, 43,
5507-5510.
(25) We had previously developed a less yielding synthesis of ketone 13 starting
from methyl (S)-3-hydroxy-2-methylpropanoate via methyl (S)-3-iodo-2-
methylpropanoate,26a the methyl analogue of 18, and amide 19. The
reduction of 23 with Li[BH(CH(Me)CH2Me] in THF gave a mixture of 26
and 30 in a ratio of 1:3: Cheng, K.; Gais, H.-J., unpublished results.
(26) (a) Corey, E. J.; Helal, C. J. Tetrahedron Lett. 1997, 38, 7511-7514. (b)
Corey, E. J.; Helal, C. J. Angew. Chem. 1998, 110, 2092-2118; Angew.
Chem., Int. Ed. Engl. 1998, 37, 1986-2012.
(27) Itsuno, M. Org. React. 1998, 52, 395-576.
9
J. AM. CHEM. SOC. VOL. 127, NO. 50, 2005 17913

A R T I C L E S
Kramp et al.
Scheme 5. Synthesis of Alkenylstannane 27a
92% ee on a Daicel Chiralcel AD column (250 mm × 50 mm)
allowed the ready separation of ent-19 and 19 (500 mg of 19/
ent-19 per injection) and furnished amide 19 with g99% ee in
95% yield.
The direct transamidation of 17 upon treatment with [MeO-
28c
(Me)NH2]Cl/AlMe3 with formation of 19 was not efficient.
Besides 19 (34%), the corresponding amide resulting from an
attack of the hydroxylamine derivative at the endocyclic
carbonyl group of 17 was also obtained (30%) (for details, see
Scheme S1 Supporting Information Available).28c,d
Although amide 19 could be efficiently synthesized from
oxazolidinone 15 and iodide 16, the ready preparative scale
separation of 19 and ent-19 by HPLC on a chiral stationary
phase led us to consider, as an alternative, a synthesis of rac-
19 and its chromatographic resolution.11 The racemic ester rac-
18 was prepared from ester 20 and iodide 16 in 50% yield
(Scheme 4).32 Amidation of rac-18 with MeO(Me)NMgCl gave
amide rac-19 in 90% yield. The resolution of rac-19 on a 3.6-g
scale by HPLC on a Daicel Chiralcel AD column (250 mm ×
50 mm) readily (500 mg of rac-19 per injection) afforded 19
with g99% ee in 47% yield and ent-19 with g99% ee in 47%
yield.33
Chain elongation of amide 19 upon reaction with the
alkenyllithium derivative 22,34 which was prepared through Sn/
Li exchange of bisstannane 21 with n-BuLi, afforded ketone
23 with g99% ee in 75% yield (Scheme 5). HPLC analysis of
23 showed that the synthesis of the ketone was not accompanied
by a partial racemization.
a Reagents and conditions: (I) 21, n-BuLi, THF, -78 °C. (II) (a) 22,
19, THF, -78 °C; (b) NH4Cl/H2O, -78 °C. (III) (a) 2 equiv of 24, 2 equiv
of 25, toluene, -78 °C, 12 h; (b) MeOH, -78 °C to room temperature; (c)
HPLC. (IV) (a) 2,6-Lutidine, CH2Cl2, -10 °C; (b) t-BuMe2SiOSO2CF3,
-10 °C.
highly diastereoselective reduction of structurally related un-
saturated ketones.11,26a
Oxazaborolidine 2426 was selected as catalyst for the dias-
tereoselective reduction of ketone 23 with catecholborane (25).
The reduction was carried out by the slow addition of a toluene
solution of the ketone, which was treated with molecular sieve
(4 Å) prior to the reduction experiment, to a solution of 2 equiv
of freshly distilled 25 and 2 equiv of freshly prepared catalyst
24 in toluene at -78 °C. Thereby a mixture of the diastereomeric
alcohols 26 and 30 (Scheme 6, vide infra) in a ratio of 95:5
could be reproducibly isolated in 90% yield after chromatog-
raphy. For comparison purposes, a 1:1-mixture of 26 and 30
(vide infra) was prepared through reduction of ketone 23 with
NaBH4 in EtOH at room temperature. The diastereomeric
alcohols were readily and quantitatively separated by preparative
HPLC. Thus HPLC of alcohol 26 of 90% de gave the alcohol
of g99% de in 81% yield. The relative and absolute configura-
tions of alcohols 26 and 30 had been previously determined by
calculation of their CD spectra and comparison with the
experimental CD spectra.35 Inversion of the configuration of
alcohol 30 upon treatment with PPh3, PhCO2H, and diethyl
azodicarboxylate36 afforded a mixture of the corresponding
regioisomeric allylic C15 and C13 benzoates in a ratio of 80:
2.1.1. Synthesis of the C15-C20 Amide. The synthesis of
the C15-C20 subunit 19 of the C13-C20 building block 27
was carried out by using the oxazolidinone method28 for the
selective generation of the stereogenic center (Scheme 3). The
benzyl-substituted oxazolidinone 1528,29 was selected as the
starting material. It had given a high selectivity in the synthesis
of the homologous C15-C21 amide, which served as the side
chain building in the synthesis of cicaprost (4).11 Deprotonation
of 15 with NaN(SiMe3)2 and treatment of the corresponding
sodium enolate with iodide 16,30 which was prepared from the
commercially available 2-butyne-1-ol,31 gave the substituted
oxazolidinone 17 with 92% diastereomeric excess (de) in 70%
yield. Treatment of oxazolidinone 17 with Ti(OEt)4 in refluxing
EtOH11,28b furnished ester 18 with 92% ee (gas chromatography,
GC) in 68% yield. Finally, amidation of ester 18 with MeO-
(Me)NMgCl, which was prepared in situ from [MeO(Me)NH2]-
Cl and 2 equiv of i-PrMgCl in THF,24 afforded amide 19 in
93% yield. High-performance liquid chromatography (HPLC)
and GC analysis showed the amide to have an ee value of 92%.
Thus no partial racemization of 19 had occurred during the
reaction sequence. Preparative HPLC of amide 19 (10 g) with
(32) For a synthesis of the corresponding racemic acid, see: Wakita, H.;
Matsumoto, K.; Yoshiwara, H.; Hosono, Y.; Hayashi, R.; Nishiyama, H.;
Nagase, H. Tetrahedron 1999, 55, 2449-2474.
(33) For laborious resolutions of the corresponding acid with chiral amines,
see: (a) Wakita, H.; Yoshiwara, H.; Kitano, Y.; Nishiyama, H.; Nagase,
H. Tetrahedron: Asymmetry 2000, 11, 2981-1989. (b) Skuballa, W.;
Vorbru¨ggen, H. In AdVances in Prostaglandin, Thromboxane, and Leu-
kotriene Research; Samuelsson, B., Paoletti, R., Ramwell, P., Eds.; Raven
Press: New York, 1983; Vol. 11, pp 299-305.
(34) Corey, E. J.; Wollenberg, R. H. J. Am. Chem. Soc. 1974, 96, 5581-5583.
(b) Renaldo, A. F.; Labadie, L. W.; Stille, J. K. Org. Synth. 1989, 67, 86-
97.
(35) Voloshina, E. N.; Raabe, G.; Fleishhauer, J.; Kramp, G. J.; Gais, H.-J. Z.
Naturforsch. 2004, 59a, 124-132.
(36) Mitsunobu, O. Synthesis 1981, 1-28.
(28) (a) Evans, D. A.; Starr, J. T. J. Am. Chem. Soc. 2003, 125, 13531-13540.
(b) Evans, D. A.; Ellman, J. A.; Dorow, R. L. Tetrahedron Lett. 1987, 23,
1123-1126. (c) Evans, D. A.; Bender, S. L. Tetrahedron Lett. 1986, 27,
799-802. (d) Romo, D.; Rzasa, R. M.; Shea, H. A.; Park, K.; Langenhan,
J. M.; Sun, L.; Akhiezer, A.; Liu, J. O. J. Am. Chem. Soc. 1998, 120,
12237-12254.
(29) (a) Tyrell, E.; Tsang, M. W. H.; Skinner, G. A.; Fawcett, J. Tetrahedron
1996, 52, 9841-9852. (b) Ho, G.-J.; Mature, D. J. J. Org. Chem. 1995,
60, 2271-2273.
(30) For a report on the alkylation of ent-16 with 1-bromo-2-butyne without
giving details about the diastereoselectivity, see: Savignac, M.; Durand,
J.-O.; Geneˆt, J.-P. Tetrahedron: Asymmetry 1994, 5, 717-722.
(31) Lange, G. L.; Gottardo, C. Synth. Commun. 1990, 20, 1473-1479.
9
17914 J. AM. CHEM. SOC. VOL. 127, NO. 50, 2005

Total Syntheses of the Prostacyclin Analogue
A R T I C L E S
followed by a distillation.37b According to 1H, 13C, and 11B NMR
spectroscopic analyses the thus obtained oxazaborolidine 24 was
pure and contained neither starting material nor n-butylboronic
acid or any intermediate of the synthesis of 24, all of which
have been previously identified as possible sources of low
enantioselectivity in reductions catalyzed by 24.37 Furthermore
all solvents were rigorously dried including a final treatment
with activated molecular sieve (4 Å). Similarly, the solution of
ketone 23 was dried with activated molecular sieve, which was
removed prior to the reaction with 24 and 25. Finally the
commercial catecholborane (25), the 11B NMR spectrum of
which indicated the presence of two further unidentified BH
group containing boranes, was purified by distillation.
In all experiments with substoichiometric amounts of 24, the
borane 25 was slowly added to a solution of 23 and 24 in order
to suppress a noncatalyzed reduction. However, all of the above
measures that were taken to obtain a reproducible high diaste-
reoselectivity in the reduction of ketone 23 in the presence of
substoichiometric amounts of 25 were unsuccessful. At present
we have no rationalization for the apparently irregular variation
of the selectivities in the substoichiometric reductions of 23 with
24 and 25. It seems interesting to note that in a recently
described oxazaborolidine-catalyzed reduction of a chiral alkynone
with BH3 a significantly lower diastereoselectivity was observed
when substoichiometric amounts of the catalyst were em-
ployed.38,39
2.4. Conjugate Addition of the C13-C20 Alkenylcopper
Derivative 9 to the C6-C12 Azoalkene 10. With the required
building blocks 10 and 27 in hand, their coupling was the next
and crucial step according to the synthetic plan. Special attention
had to be paid to the generation of a reactive derivative of the
alkenylcopper compound 9 and its conjugate addition to
azoalkene 10 by using nearly equimolar amounts of the two
building blocks. Because of the easiness of preparation, we opted
for the alkenylcopper-n-Bu3P complex 9a (Scheme 7) instead
of a corresponding cuprate containing a nontransferable organic
ligand.18a
Scheme 6. Oxazaborolidine-Catalyzed Reduction of Ketone 23
with Borane 25
20. Hydrolysis of the mixture of the benzoates without prior
separation gave alcohol 26 with g99% ee in 56% overall yield
(for details, see Scheme S2 Supporting Information Available).
Silylation of alcohol 26 with t-BuMe2SiOSO2CF3 furnished
the silyl ether 27 in 98% yield. Interestingly, the silylation of
26 with t-BuMe2SiCl in DMF in the presence of imidazol was
not only much slower but was also accompanied by a partial
destannylation of 27, which led to the formation of the derivative
of 27, carrying a H-atom instead of the stannyl group, as a side
product.
2.1.2. Substoichiometric Oxazaborolidine-Catalyzed Re-
duction of Ketone 23. Much to our surprise, reproducible high
diastereoselectivities in the reduction of ketone 23 with 24 and
25 were only obtained when 2 equiv of catalyst 24 were applied.
The use of sub stoichiometric amounts of 24 (0.10-0.35 equiv)
and 1.1-2.0 equiv of 25 in toluene or CH2Cl2 at -78 °C gave
in our hands widely differing and nonreproducible diastereo-
selectivities ranging from very high (98:2) to almost nil (55:
45) (for details, see Table S1 and Scheme S3, Supporting
Information Available). In all reductions experiments of 23 with
less than stoichiometric amounts of catalyst 24 to various extend
the formation of the Z-configured ketone 31, and the saturated
ketone 32 was observed (Scheme 6). Reduction of 23 with 25
in the absence of 24 was much slower and gave a nearly 1:1
mixture of 26 and 30.
Treatment of alkenylstannane 27 with n-BuLi afforded the
alkenyllithium derivative 33 which upon reaction with 0.275
equiv of [CuI(n-Bu3P)]440 furnished the alkenylcopper derivative
9a. The conjugate addition was achieved through reaction
azoalkene 10 with 1.1 equiv of 9a in the presence of 0.275 equiv
of [CuI(n-Bu3P)]4, which afforded hydrazone 8 in 84% yield
as a single diastereomer at C12. Quenching of the reaction
mixture with n-Bu3SnCl allowed a recovery of the 0.1 equiv of
excess 9a as stannane 27 in 60% yield. The configuration of 8
at the CN-double bond, which has not been determined, is
arbitrary depicted as E.
It had been described that the enantioselectivity of oxazaboro-
lidine-catalyzed reduction of ketones with boranes can be
crucially dependent on the water content of the system and the
presence of starting materials or intermediates of the synthesis
of the catalyst.37 Thus in all reduction experiments of 23, catalyst
24 was freshly prepared from 1 equiv of the amino alcohol 28
and 0.333 equiv of the pure borinane 29 in refluxing toluene
The crucial chemoselective cleavage of hydrazone 8 was
accomplished upon treatment with 1.05 equiv of (PhSeO)2O in
the presence of 20 equiv of cyclohexene.41 Thereby ketone 34
was isolated in 70% yield. The reduction of ketone 34 with
(38) (a) Trost, B. M.; Gunzner, J. L.; Dirat, O.; Rhee, Y. H. J. Am. Chem. Soc.
2002, 124, 10396-10415.
(39) For further examples, see: (a) Parker, K. A.; Ledeboer, M. W. J. Org.
Chem. 1996, 61, 3214-3217. (b) Rodriguez, A.; Nomen, M.; Spur, B. W.;
Godfroid, J.-J. Eur. J. Org. Chem. 1999, 2655-2662. (c) Gais, H.-J.;
Kramp, G. J., unpublished results.
(40) Suzuki, M.; Noyori, R. In Organocopper Reagents; Taylor, R. J. K., Ed.;
Oxford University Press: Oxford, 1994; pp. 182-216.
(41) (a) Barton, D. H. R.; Okano, T.; Parekh, S. I. Tetrahedron 1991, 47, 1823-
1836. (b) Barton, D. H. R.; Choi, S.-Y.; Liu, W.; Smith, J. A. Mol. Online
1998, 2, 22-28.
(37) (a) Jones, K. T.; Mohan, J. J.; Xavier, L. C.; Blacklock, T. J.; Mathre, D.
J.; Sohar, P.; Jones, E. T. T.; Reamer, R. A.; Roberts, F. E.; Grabowski, E.
J. J. J. Org. Chem. 1991, 56, 763-769. (b) Mathre, D. J.; Jones, T. K.;
Xavier, L, C.; Blacklock, T. J.; Reamer, R. A.; Mohan, J. J.; Jones, E. T.
T.; Hoogsteen, K.; Baum, M. W.; Grabowski, E. J. J. J. Org. Chem. 1991,
56, 751-762. (c) Tschaen, D. M.; Abramson, L.; Cai, D.; Desmond, R.;
Dolling, U.-H.; Frey, L.; Karady, S.; Shi, Y.-J.; Verhoeven, T. R. J. Org.
Chem. 1995, 60, 4324-4330.
9
J. AM. CHEM. SOC. VOL. 127, NO. 50, 2005 17915

A R T I C L E S
Kramp et al.
Scheme 7. Conjugate Addition of the Alkenylcopper Derivative 9
to Azoalkene 10a
Scheme 8. Cleavage of Hydrazone 8 in the Absence of
Cyclohexene
1
revealed by a HPLC and H NMR spectroscopic analysis of
the bissilyl ether 36.
Crucial for the attainment of a high yield of ketone 34 in the
cleavage of hydrazone 8 with (PhSeO)2O proved to be the
presence of a large excess of cyclohexene. In the absence of
cyclohexene the selanyl derivative 38 (Scheme 8) was obtained
as a single diastereomer (32%) in addition to ketone 34 (30%).
The configuration at C12 of 38, which could not be determined
by nuclear Overhauser effect (NOE) experiments, is arbitrarily
assigned as depicted. Evidence had been previously presented
showing that the cleavage of tosyl hydrazones with (PhSeO)2O
involves the intermediate formation of the tolylsulfonyl and
phenylselanyl radicals, which can be trapped with cyclohexene.41
Thus the formation of 38 can perhaps be ascribed to a reaction
of ketone 34 with these radicals which are otherwise efficiently
trapped by cyclohexene with formation of the corresponding
phenylselenotosylate derivative.
2.5. Asymmetric HWE Olefination of the C6-C20 Ketone
7 and Completion of the Synthesis of 16S-3-Oxa-Iloprost.
Now the synthetic plan called for a diastereoselective HWE
olefination of ketone 7 with formation of the E-configured R,â-
unsaturated ester E-6 (Scheme 9). We selected the dimeth-
ylphosphoryl acetate derived from (1S,2R)-8-phenylnormen-
thol11,16 because of the ready availability of this alcohol in
enantiomerically pure form.42 Treatment of ketone 7 with 6
equiv of the lithium salt 39 in THF at -62 °C for 6 days resulted
in a highly selective olefination and gave after aqueous work
up and chromatography a mixture of the diastereomeric esters
E-6 and Z-6 in a ratio of 98:2 in 89% yield. The excess of the
phosphoryl acetate was recovered in almost quantitative yield.
Preparative HPLC of the mixture of esters E-6 and Z-6 afforded
ester E-6 with g99% de in 78% yield and ester Z-6 with g99%
de in 1% yield.
The high diastereoselectivity of the reaction of 7 with 39 can
be rationalized by assuming that (1) a selectivity determining
step involving the addition of phosphonate 39 to ketone 7 and
(2) a preferred trajectory of approach from the Re side of 39
and the convex side of 7.16b Reduction of ester E-6 with (i-
Bu)2AlH in THF gave the allylic alcohol E-5 in 94% yield and
8-phenylnormenthol in 85% yield. The E-configuration of the
double bond of E-5 was secured by NOE experiments, which
revealed strong NOEs between 6a-H and 4-H and between 5-H
and 7-H.
a Reagents and conditions: (I) 1.1 equiv of n-BuLi, THF, -78 °C, 1 h.
(II) 33, 0.275 equiv of [CuI(n-Bu3P)]4, THF, -78 °C, 30 min. (III) 33, 1.2
equiv of CuCN, 2.4 equiv of LiCl, THF, -78 °C. (IV) (a) 1.1 equiv of 9b,
10, THF, -78 °C; (b) 2 equiv of Bu3SnCl, THF, -78 °C; (c) H2O, NH4Cl,
THF, -78 °C to room temperature; (d) 1.05 equiv of (PhSeO)2O, 20 equiv
of cyclohexene, THF, room temperature; (e) 6 equiv of NaBH4, EtOH,
0 °C; (f) H2O, NH4Cl, 0 °C to room temperature. (V) (a) 1.1 equiv of 9a,
10, 0.275 equiv of [CuI(n-Bu3P)]4, THF, -78 °C, 1 h; (b) 2 equiv of
Bu3SnCl, -78 °C, 30 min; (c) NH4Cl/NH3 (9:1), -78 °C to room
temperature. (VI) 20 equiv of cyclohexene, 1.05 equiv of (PhSeO)2O, THF,
room temperature, 1 h. (VII) 2 equiv of NaBH4, EtOH, 0 °C, 3 h. (VIII)
(a) 2,6-Lutidine, CH2Cl2, -10 °C; (b) t-BuMe2SiOSO2CF3, -10 °C. (IX)
p-TsOH, acetone/water, room temperature, 24 h. (X) t-BuMe2SiCl, imidazol,
DMF, room temperature.
NaBH4 furnished alcohol 35 as a single diastereomer at C11 in
76% yield. The synthesis of alcohol 35 has also been carried
out without the isolation of hydrazone 8 and ketone 34 in 51%
overall yield starting from 10 and 27 by using the cyano cuprate
9b which was prepared by treatment of 33 with 1.2 equiv of
CuCN and 2.4 equiv of LiCl.
Now a cleavage of the protecting groups of 35 with formation
of ketone-diol 37 was required. Surprisingly, the reaction of 35
with p-TsOH in acetone/water was sluggish. Therefore alcohol
35 was first converted to the bissilyl ether 36 in 98% yield.
Treatment of 36 with p-TsOH in acetone/water proceeded
uneventful and gave diol-ketone 37, the hydroxy groups of
which were silylated to afford ketone 7 in 98% yield. Isolation
of 37 is not required for the synthesis of 7 from 36.
A 1H NMR spectroscopic analysis of hydrazone 8 indicated
the presence of a minor amount of the diastereomer of 8
stemming from the reaction of ent-10, contained in azoalkene
10 of 95% ee, with 9a. A chromatographic separation of the
minor diasteromer was achieved at the stage of alcohol 35 as
The synthesis of 16S-3 was completed following the route
previously used for the synthesis of cicaprost (4)10a,11 and other
3-oxa-analogues of carbacyclins.16c,d Thus etherification of
alcohol E-5 through treatment with excess BrCH2COOt-Bu
under phase transfer conditions in the presence of 50% aqueous
(42) Commins, D. L.; Salvador, J. M. J. Org. Chem. 1993, 58, 4656-4661.
9
17916 J. AM. CHEM. SOC. VOL. 127, NO. 50, 2005

Total Syntheses of the Prostacyclin Analogue
A R T I C L E S
Scheme 9. Asymmetric HWE Olefination of the Bicyclic Ketone 7a
Scheme 11. Allylic Alkylation of Acetate 43 with Cuprate 42 and
Completion of the Synthesis of 16S-2a
a Reagents and conditions: (I) CuI, Et2O, Me2S. (II) Ac2O, THF. (III)
42, Et2O. (IV) Alumina, hexanes, H2O. (V) DMSO, SO3-pyridine, NEt3.
(VI) AgNO3, EtOH, H2O, NaOH. (VII) n-Bu4NF, THF.
a Reagents and conditions: (I) (a) 6.0 equiv of 39, THF, -62 °C, 6 d;
(b) NH4Cl, -62 °C to room temperature; (c) HPLC. (II) (i-Bu)2AlH, THF,
0 °C.
alkylation of the allylic alcohol 5 or a derivative thereof with
an organocopper C1-C3 building block (Scheme 11). In
principle ketones 7 and 37 (cf. Scheme 7) could also serve as
starting materials for the synthesis of 16S-2. However, the Wittig
olefination of ketones 7 and 37 with the achiral ylide
Ph3PdCH(CH2)3CO2-M+ proceeds only with low E-stereose-
lectivity because of the insufficient asymmetric induction
provided by the $-side chain.43 For example reaction of the
hydroxy ketone 37 with Ph3PdCH(CH2)3CO2-M+, which had
been reported to be more selective than that of the protected
ketone 7 (E:Z ) 78:22) and to deliver a mixture of 16S-2 and
its Z isomer in a ratio of 90:10,43 gave in our hands a mixture
of 16S-2 and its Z isomer in a ratio of only 62:38, the separation
of which by HPLC afforded 16S-2 in 45% yield and its Z isomer
in 12% yield (for details, see Scheme S4 Supporting Information
Available).
Scheme 10. Etherification of the Allylic Alcohol 5 and Completion
of the Synthesis of 16S-3a
a Reagents and conditions: (I) (a) n-Bu4NHSO4, 50% NaOH, BrCH2COOt-
Bu, CH2Cl2, room temperature; (b) n-Bu4NF, THF, room temperature. (II)
MeOH, 1 N NaOH, NaH2PO4, pH 4-5, room temperature.
NaOH and a subsequent desilylation of the corresponding
bissilyl ether, which was not isolated, with n-Bu4NF gave the
dihydroxy ester 40 in 90% overall yield (Scheme 10). Finally,
the hydrolysis of ester 40 with NaOH in MeOH and acidification
with NaH2PO4 to pH 4-5 furnished the enantio- and diaste-
reomerically pure 16S-3-oxa-iloprost (16S-3) in 90% yield. An
acidification of the solution of 16S-3 to pH 1 resulted in its
partial conversion to a mixture of the corresponding 6-vinyl
substituted C6-C6a- and C6-C7-dienes and 2-hydroxyacetic
acid.
To solve the problem of the stereoselective olefination of
ketone 37 the sulfoximine method could be applied.44 Applica-
tion of this method would involve a conversion of the ketone
to the corresponding 5E-configured 6-(phenylsulfonimidoyl)-
methylene derivative and its subsequent Ni-catalyzed cross-
coupling reaction with a suitably functionalized C1-C4-
(43) Westermann, J.; Harre, M.; Nickisch, K. Tetrahedron Lett. 1992, 33, 8055-
8056.
2.7. Allylic Alkylation of the C4-C20 Alcohol 5 and
Completion of the Synthesis of 16S-Iloprost. The retrosyn-
thesis of 16S-2 now demanded for a regio- and stereoselective
(44) (a) Erdelmeier, I.; Gais, H.-J.; Lindner, H. J. Angew. Chem. 1986, 98, 912-
914; Angew. Chem., Int. Ed. Engl. 1986, 25, 935-937. (b) Erdelmeier, I.;
Gais, H.-J. J. Am. Chem. Soc. 1989, 111, 1125-1126. (c) Gais, H.-J.;
Bu¨low, G. Tetrahedron Lett. 1992, 33, 465-468.
9
J. AM. CHEM. SOC. VOL. 127, NO. 50, 2005 17917

A R T I C L E S
Kramp et al.
Scheme 12. Mechanistic Rationalization of the Allylic Alkylation of
Acetate 43 with Cuprate 42
complex 48a is expected to give via conformer 48b the
Z-configured π-allylcopper(III) complex 49b, the isomerization
of which would generate the Z-configured R-σ-allylcopper(III)
complex 50b. A final reductive elimination of 50b would lead
to the formation of the Z isomer of alkene 44, an event which
was, however, not observed.
The completion of the synthesis of 16S-2 demanded a
chemoselective deprotection of of the primary silyl ether group
of 44 in the presence of the two secondary silyl ether groups
(cf. Scheme 11). This was achieved upon treatment of ether 44
with water containing neutral alumina in hexanes,47 which
furnished the primary alcohol 45 in 80% yield. Formation of a
corresponding diol was not observed. The oxidation of the
primary alcohol 45 to acid 47 was carried out without purifica-
tion of intermediates 46 and 47 by the two-step procedure we48
and others49 had successfully used in previous isocarbacylin
syntheses. Thus oxidation of alcohol 45 with SO3/pyridine/
DMSO afforded aldehyde 46, which was oxidized with Ag2O
to acid 47. The subsequent deprotection of the bissilyl ether 47
upon treatment with n-Bu4NF in THF finally gave the enantio-
and diastereomerically pure 16S-iloprost (16S-2) in 59% overall
yield based on alcohol 5.
3. Conclusion
In this article, we report asymmetric total syntheses of 16S-
iloprost (16S-2) and 16S-3-oxa-iloprost (16S-3) from the chiral
bicyclic azoalkene 10 (4 steps from 11, 55% yield) and the chiral
alkenylstannane 27 (6 steps from 15, 25% yield) in 11 steps in
12% overall yield and in 7 steps in 25% overall yield,
respectively. A diversion at the late-stage intermediate, the allylic
alcohol 5, allowed the synthesis of both 16S-2 and 16S-3. The
key steps of the syntheses are (1) the coupling of a bicyclic
C6-C12 building block with a C13-C20 building block
through a stereoselective conjugate addition of an alkenylcopper
derivative to an azoalkene, (2) the stereoselective olefination
of ketone 7 with a chiral phosphoryl acetate, and (3) the regio-
and stereoselective allylic alkylation of acetate 43. The synthesis
of 16S-2 from ketone 7 via the ester 6 and the allylic alcohol 5
represents a solution of the long-standing problem of the
stereoselective olefination of 7 and related bicyclic ketones. The
successful syntheses of 16S-2 and 16S-3 clearly demonstrate
the viability of the alkenylcopper-azoalkene route for the
synthesis of carbocyclic prostacyclin analogues.
zinkorganyl. Although the synthesis of the 5E-configured
alkenylsulfoximine should proceed with high stereoselectivity
and yield, the feasibility of a stereoselective cross-coupling
reaction has thus far only been demonstrated for bicyclic
alkenylsulfoximines of this type carrying an incomplete ω-side
chain. Therefore, an allylic alkylation of the late-stage inter-
mediate 5 seemed to be much more attractive provided this
transformation can be made to proceed with high regio- and
stereoselectivities. Acetate 43 and homocuprate 4245a which was
prepared from the lithiumorganyl 4145b were selected as starting
materials. Gratifyingly, the allylic alkylation of acetate 43, which
was obtained from alcohol 5 in 96% yield, with homocuprate
42 proceeded with high regio- and diastereoselectivities at C4
1
and gave the E-configured alkene 44 in 82% yield. H NMR
The facile conjugate addition of alkenyl- and arylcopper
derivatives to the azoalkene 10 points to an unexplored potential
of azoalkenes for the alkylation, arylation and alkenylation of
ketones at the R-position,50 which warrants further studies to
determine the scope and limitation of this transformation. Fuchs
et al. had already shown that azoalkenes derived from simple
cyclic and acyclic ketones readily react with aryl- and alkyl-
copper reagents.15a
spectroscopy and HPLC analysis showed alkene 44 to be free
of the corresponding Z isomer and the regioisomers derived from
a substitution of 43 at C6.
The complete retention of the configuration of the exocyclic
double bond in the allylic alkylation of acetate 43 is noteworthy.
It suggests that the oxidative addition/substitution of acetate 43
upon reaction with cuprate 42 directly leads to the E-configured
π-allylcopper(III) complex 49a which suffers an isomerization
to the E-configured R-σ-allylcopper(III) complex 50a, the
reductive elimination of which gives 4446 (Scheme 12). The
normally expected oxidative addition/substitution of 43 upon
reaction with 42 under formation of the γ-σ-allylcopper(III)
complex 48a18a seems not to have taken place. Once formed
The enantio- and diastereomerically pure building blocks 18,
19, and 27, for which efficient syntheses have been developed,
could also find application in asymmetric syntheses of beraprost.
(47) Feixas, J.; Capdevila, A.; Guerrero, A. Tetrahedron 1994, 50, 8539-8550.
(48) (a) Hemmerle, H.; Gais, H.-J. Angew. Chem. 1989, 101, 362-365; Angew.
Chem., Int. Ed. Engl. 1989, 28, 349-351. (b) Bund, J.; Gais, H.-J.; Schmitz,
E.; Erdelmeier, I.; Raabe, G. Eur. J. Org. Chem. 1998, 1319-1335.
(49) Mase, T.; Sodeoka, M.; Shibasaki, M. Tetrahedron Lett. 1984, 25, 5087-
5090.
(45) (a) Hua, D. H.; Verma, A. Tetrahedron Lett. 1985, 26, 547-550. (b) Tanner,
D.; Hagberg, L. Tetrahedron 1998, 54, 7907-7918.
(46) (a) Corey, E. J.; Boaz, N. W. Tetrahedron Lett. 1984, 25, 3063-3066. (b)
Underiner, T. L.; Paisley, S. D.; Schmitter, J.; Lesheski, L.; Goering, H.
L. J. Org. Chem. 1989, 54, 2369-2374.
(50) For a recent example, see: Chae, J.; Yun, J.; Buchwald, S. L. Org. Lett.
2004, 6, 4809-4812.
9
17918 J. AM. CHEM. SOC. VOL. 127, NO. 50, 2005

Total Syntheses of the Prostacyclin Analogue
A R T I C L E S
41.3 (u), 43.9 (d), 57.6 (d), 71.7 (u), 71.7 (u), 76.4 (u), 76.76 (d), 78.0
(u), 78.2 (d), 110.2 (u), 132.8 (u), 133.4 (u); IR (neat) ν 3457 (m),
2955 (s), 2858 (s), 2278 (m), 1467 (m), 1393 (m), 1361 (m), 1329
(m), 1253 (m), 1219 (m), 1176 (w), 1113 (s), 1062 (s), 974 (m), 938
(w), 837 (s) cm-1; MS (EI, 70 eV) m/z (relative intensity, %) 476 (M+,
0.7), 419 (29), 378 (28), 377 (100), 333 (17), 291 (14), 263 (10), 241
(15), 159 (14); HRMS calcd for C28H48O4Si+ 476.332189, found
476.332258.
Beraprost which is an orally active benzoprostacyclin derivative
having the same ω-side chain as iloprost has already been
approved as Procyclin and Dorner for the treatment of peripheral
vascular diseases and primary pulmonary hypertension.51 How-
ever, the Procyclin and Dorner formulations of beraprost consist
of mixtures of the four C15/C16 isomers, three of which are
much less active.51a
The diastereoselective reduction of the chiral C13-C20
ketone 23 with substoichiometric amounts of the oxazaboroli-
dine catalyst and catecholborane gave widely differing diaste-
reoselectivities ranging from very high to nil. Reproducible high
selectivities were recorded only in experiments with 2 equiv of
the chiral catalyst. Thus, it seems that with certain substrates
the stereoselectivity of the oxazaborolidine catalyzed reduction
with substoichiometric amounts of the catalyst is influenced by
factors which have not yet been fully identified.
(+)-(E)-((1S,2R)-2-(2-Phenylpropan-2-yl)cyclohexyl)-2-((3aS,4R,5R,-
6aS)-5-(tert-butyldimethylsilyloxy)-4-((3S,4S,E)-3-(tert-butyldimethyl-
silyloxy)-4-methyl-oct-1-en-6-ynyl)hexahydropentalen-2(1H)-yl-
idene)acetate (E-6) and (Z)-((1S,2R)-2-(2-Phenylpropan-2-yl)cyclo-
hexyl)-2-((3aS,4R,5R,6aS)-5-(tert-butyldimethylsilyloxy)-4-((3S,4S,E)-
3-(tert-butyl-dimethylsilyloxy)-4-methyl-oct-1-en-6-ynyl)hexahydro-
pentalen-2(1H)-ylidene)acetate (Z-6). To a solution of (1S,2R)-
dimethoxyphosphanyl-2-(2-phenylpropan-2-yl)cyclohexyl 2-(dimethoxy-
phosphoryl)acetate (1.313 g, 3.56 mmol) in THF (8 mL) was added
n-BuLi (2.08 mL of 1.6 M in hexanes, 3.32 mmol) at -78 °C. The
resulting solution of the lithium salt 39 was warmed to ambient
temperature for 15 min and cooled to -62 °C. Then a solution of ketone
7 (300 mg, 0.594 mmol) in THF (3 mL) was added within 10 min.
Subsequently the mixture was stirred at -62 °C by using a cryostat
for 144 h. Then aqueous NH4Cl (15 mL) was added at -62 °C, and
the mixture was warmed to ambient temperature. The aqueous phase
was separated, diluted with water until a clear solution was formed,
and extracted with Et2O (4 × 30 mL). The combined organic phases
were dried (MgSO4) and concentrated in vacuo. Chromatography
(hexanes/EtOAC, 10:1) afforded a mixture of esters E-6 and Z-6 (394
mg, 89%) in a ratio of 98:2 (1H NMR δ (CHdCHCHOSi) 4.15, δ
(CHOCO) 5.38 (E-6); δ (CHdCHCHOSi) 3.95, δ (CHOCO) 5.42 (Z-
6) as a colorless oil, ketone 7 (25 mg, 8%) and the phosphoryl acetate
(1.06 g, 2.9 mmol). Preparative HPLC (Kromasil Si-100, 30 mm;
n-hexane/EtOAc, 95:5; UV, 254 nm, RI) gave ester E-6 (348 mg, 78%)
with g99% de and ester Z-6 (6 mg, 1%) as colorless oils.
4. Experimental Section
(+)-(3′aS,4′R,5′R,6′aR)-4′-((3S,4S,E)-3-(tert-Butyldimethylsilyloxy)-
4-methyl-oct-1-en-6-ynyl)-5,5-dimethylhexahydro-1′H-spiro[[1,3]-
dioxane-2,2′-pentalen]-5′-ol (35). From azoalkene 10 and stannane 27
without isolation of 8 and 34. n-BuLi (0.88 mL, 1.60 M in hexanes,
1.40 mmol) was added at -78 °C to a solution of stannane 27 (762
mg, 1.40 mmol) in THF (4 mL). After the mixture was stirred at -78
°C for 1 h, it was added to a cold solution of CuCN (275 mg, 3.08
mmol) and LiCl (260 mg, 6.13 mmol) in THF (3 mL) at -78 °C via
a double-ended needle. The resulting yellow solution was stirred at
-78 °C for 30 min. Then a cold solution of azoalkene 10 (500 mg,
1.28 mmol, 95% ee) in THF (4 mL) was added via a double-ended
needle. After the mixture was stirred at -78 °C for 1 h, n-Bu3SnCl
(0.76 mL, 2.8 mmol) was added. Then the mixture was stirred at -78
°C for 30 min, water (3 mL) was added, and the mixture was warmed
to ambient temperature. Subsequently the mixture was diluted with Et2O
(100 mL) and washed with a mixture of saturated aqueous NH4Cl and
concentrated aqueous NH3 (10:1) (3 × 20 mL). The combined aqueous
phases were extracted with Et2O (3 × 20 mL), and the combined
organic layers were dried (MgSO4) and concentrated in vacuo. The
residue was dissolved in THF (20 mL), cyclohexene (2.60 mL, 25.61
mmol) was added, and the solution was treated with (PhSeO)2O (461
mg, 1.28 mmol) at ambient temperature, whereby a gas evolution
occurred. The mixture was stirred at ambient temperature for 40 min,
cooled to 0 °C, and EtOH (30 mL) added. Then NaBH4 (291 mg, 7.68
mmol) was added at 0 °C. After the mixture was stirred at 0 °C for 2
h, saturated aqueous NH4Cl (3 mL) was added, and the mixture was
warmed to ambient temperature. Then the mixture was concentrated
in vacuo, and the residue was dissolved in a mixture of Et2O (100 mL)
and water (10 mL). The aqueous layer was extracted with Et2O (3 ×
20 mL), and the combined organic layers were dried (MgSO4) and
concentrated in vacuo. Purification by chromatography (hexanes/EtOAc,
6:1) gave alcohol 35 (330 mg, 51%) as a colorless oil. [R]D +16.5 (c
1.71, CDCl3); Rf 0.36 (hexanes/EtOAc, 3:1); 1H NMR (400 MHz, C6D6)
δ 0.00 (s, 3 H, SiCH3), 0.02 (s, 3 H, SiCH3), 0.62 (s, 3 H, CH(CH3)2),
0.67 (s, 3 H, CH(CH3)2), 0.88 (s, 9 H, SiC(CH3)3), 0.92 (d, J ) 6.9
Hz, 3 H, CHCH3), 1.22 (bs, 1 H, OH), 1.29-1.37 (m, 1 H), 1.48 (t, J
) 2.5 Hz, 3 H, CH2C°CCH3), 1.64-1.70 (m, 2 H), 1.80-1.87 (m, 1
H), 1.93-2.02 (m, 4 H), 2.08-2.28 (m, 4 H), 3.15-3.18 (m, 4 H,
OCH2), 3.45 (dt, J ) 6.6, J ) 9.3 Hz, 1 H, CHOH), 3.92 (t, J ) 6.3
Hz, 1 H, CHOSi), 5.32-5.36 (m, 2 H, CHdCH); 13C NMR (100 MHz,
C6D6) δ -4.8 (d), -3.9 (d), 3.2 (d), 15.7 (d), 18.2 (u), 22.2 (d), 22.3
(d), 22.3 (u), 25.9 (d), 29.7 (u), 35.9 (d), 38.8 (u), 39.7 (d), 40.3 (u),
E-6: [R]D +17.6 (c 0.83, CDCl3); Rf 0.84 (hexanes/EtOAc, 5:1);
1H NMR (400 MHz, C6D6) δ 0.04 (s, 3 H, SiCH3), 0.08 (s, 3 H, SiCH3),
0.15 (s, 3 H, SiCH3), 0.17 (s, 3 H, SiCH3), 0.85-1.58 (m, 7 H), 0.97
(s, 9 H, SiC(CH3)3), 1.04 (s, 9 H, SiC(CH3)3), 1.12 (d, J ) 6.6 Hz, 3
H, CHCH3), 1.17 (s, 3 H, CCH3), 1.38 (s, 3 H, CCH3), 1.62 (t, J ) 2.5
Hz, 3 H, C°CCH3), 1.80-1.90 (m, 2 H), 2.00-2.45 (m, 10 H), 2.80-
3.08 (m, 2 H), 3.70 (dt, J ) 7.1, J ) 8.5 Hz, 1 H, CHOSi), 4.14 (t, J
) 6.0 Hz, 1 H, CHdCHCHOSi), 5.07 (dt, J ) 4.4, J ) 10.7 Hz, 1 H,
CHOCO), 5.38 (m, 1 H, COCHdC), 5.50 (dd, J ) 6.0, J ) 15.7 Hz,
1 H, CH)CH-CHOSi), 5.58 (dd, J ) 6.0, J ) 15.7 Hz, 1 H, CH-
CH)CH), 7.14-7.30 (m, 5 H, Ph); 13C NMR (100 MHz, C6D6) δ -4.8
(d), -4.6 (d), -4.3 (d), -3.8 (d), 3.2 (d), 15.6 (d), 18.0 (u), 18.2 (u),
22.4 (u), 24.8 (u), 25.4 (d), 25.9 (d), 26.0 (d), 26.1 (u), 27.1 (u), 27.6
(d), 33.9 (u), 38.8 (d), 39.4 (u), 39.7 (u), 39.9 (u), 40.1 (d), 42.6 (u),
44.4 (d), 51.2 (d), 56.3 (d), 73.3 (d), 76.2 (d), 76.4 (u), 77.9 (u), 78.7
(d), 113.8 (d), 124.7 (d), 125.5 (d), 127.9 (d), 132.0 (d), 132.6 (d),
151.5 (u), 165.0 (u), 165.5 (u); IR (CDCl3) ν 2933 (s), 2858 (s), 1706
(s), 1658 (m), 1466 (m), 1368 (m), 1253 (m), 1214 (m), 1124 (s), 1066
(m), 1032 (m), 910 (m), 839 (s) cm-1; MS (CI, CH4) m/z (relative
intensity, %) 748 (3), 747 (M+ + 1, 4), 746 (2), 689 (16), 615 (22),
531 (8), 489 (15), 416 (16), 415 (68), 397 (8), 283 (26), 201 (40), 119
(33), 105 (18), 101 (10), 85 (29), 83 (45); HRMS calcd for C46H74O4-
+
Si2 C4H9 689.442143, found 689.442163.
1
Z-6: Rf 0.84 (hexanes/EtOAc, 5:1); H NMR (300 MHz, C6D6) δ
0.03 (s, 3 H, SiCH3), 0.09 (s, 3 H, SiCH3), 0.14 (s, 3 H, SiCH3), 0.15
(s, 3 H, SiCH3), 0.85-1.58 (m, 7 H), 0.98 (s, 9 H, SiC(CH3)3), 1.03 (s,
9 H, SiC(CH3)3), 1.13 (d, J ) 6.6 Hz, 3 H, CHCH3), 1.21 (s, 3 H,
CCH3), 1.42 (s, 3 H, CCH3), 1.60 (t, J ) 2.5 Hz, 3 H, C°CCH3), 1.80-
1.90 (m, 2 H), 2.00-2.45 (m, 10 H), 2.80-3.08 (m, 2 H), 3.64-3.74
(m, 1 H, CHOSi), 4.08 (t, J ) 6.0 Hz, 1 H, CHdCHCHOSi), 5.02 (dt,
J ) 6.2, J ) 10.4 Hz, 1 H, CHOCO), 5.43-5.54 (m, 3 H, COCHdC,
CHdCHsCHOSi, CHsCHdCH), 7.15-7.27 (m, 5 H, Ph); 13C NMR
(51) (a) Wakita, H.; Yoshiwara, H.; Nishiyama, H.; Nagase, H. Heterocycles
2000, 53, 1085-1110. (b) Kim, Y. H.; Lee, Y. S. 2004, WO 2004026224;
Chem. Abstr. 2004, 140, 270668. (d) Melian, E.; Balmori, G.; Karen, L.
Drugs 2002, 62, 107-133.
9
J. AM. CHEM. SOC. VOL. 127, NO. 50, 2005 17919

A R T I C L E S
Kramp et al.
(75 MHz, C6D6) δ -4.7 (d), -4.4 (d), -4.2 (d), -3.6 (d), 3.4 (d),
15.8 (d), 18.2 (u), 18.4 (u), 22.5 (u), 25.0 (u), 26.1 (d), 26.2 (d), 26.3
(d), 27.0 (u), 27.3 (u), 27.6 (d), 34.1 (u), 36.3 (u), 37.5 (d), 40.3 (d),
40.4 (u), 42.1 (u), 43.5 (u), 47.0 (d), 51.6 (d), 57.2 (d), 73.8 (d), 76.4
(d), 76.6 (u), 78.3 (u), 79.2 (d), 113.9 (d), 125.2 (d), 126.0 (d), 128.2
(d), 132.6 (d), 132.7 (d), 151.7 (u), 165.5 (u), 166.7 (u).
discarded. Then the alumina was washed subsequently with EtOAc (6
× 20 mL) and MeOH (2 × 20 mL). The combined organic phases
were concentrated in vacuo. Purification by column chromatography
(hexanes/EtOAc, 9:1) afforded alcohol 45 (78 mg, 80%) as a colorless
oil. [R]D +31.5 (c 0.6, CDCl3); Rf 0.20 (hexanes/EtOAc, 6:1); 1H NMR
(400 MHz, CDCl3) δ 0.01 (s, 3 H, SiCH3), 0.03 (s, 3 H, SiCH3), 0.03
(s, 3 H, SiCH3), 0.05 (s, 3 H, SiCH3), 0.87 (s, 9 H, SiC(CH3)3), 0.89
(s, 9 H, SiC(CH3)3), 0.91 (d, J ) 6.9 Hz, 3 H, CHCH3), 1.16-1.28
(m, 1 H), 1.35-1.46 (m, 2 H), 1.52 (bs, 1 H, OH), 1.53-1.72 (m, 3
H), 1.78 (t, J ) 2.5 Hz, 3 H, CtCCH3), 1.90-2.11 (m, 7 H), 2.16-
2.25 (m, 2 H), 2.30-2.42 (m, 3 H), 3.64 (t, J ) 6.6 Hz, 2 H, CH2OH),
3.75 (dt, J ) 6.9, J ) 9.1 Hz, 1 H, CHOSi), 3.97 (t, J ) 6.3 Hz, 1 H,
CHOSi), 5.18-5.26 (m, 1 H, CdCHsCH2), 5.39 (dd, J ) 6.6, J )
15.7 Hz, 1 H, CHdCHsCHOSi), 5.51 (dd, J ) 7.1, J ) 15.7 Hz, 1 H,
CHdCHsCHOSi); 13C NMR (100 MHz, CDCl3) δ -4.9 (d), -4.6
(d), -4.4 (d), -3.9 (d), 3.5 (d), 15.4 (d), 18.1 (u), 18.1 (u), 22.0 (u),
25.9 (d), 25.9 (u), 29.1 (u), 32.3 (u), 36.0 (u), 37.6 (d), 38.2 (u), 39.8
(d), 42.6 (u), 44.4 (d), 56.0 (d), 62.8 (u), 76.1 (d), 76.1 (u), 78.0 (u),
78.2 (d), 121.2 (d), 131.4 (d), 132.5 (d), 142.0 (u); IR (CHCl3) ν 3375
(bw), 2931 (s), 2858 (s), 1667 (w), 1467 (m), 1382 (m), 1362 (m),
1254 (m), 1216 (m), 1107 (m), 1008 (m), 975 (m), 938 (w), 909 (w),
838 (s) cm-1; MS (EI, 70 eV) m/z (relative intensity, %) 574 (M+, 1),
517 (16), 494 (10), 493 (24), 386 (32), 385 (100), 362 (30), 361 (96),
335 (27), 311 (11), 310 (19), 293 (13), 229 (11), 197 (11), 187 (12),
185 (16), 183 (13), 171 (60), 159 (22), 157 (10), 147 (18), 145 (17),
133 (16), 131 (11), 119 (11), 109 (11), 107 (10), 105 (15), 91 (15), 81
(11); HRMS calcd for C34H62O3Si2+-C4H9 517.353328, found
517.353430.
(+)-tert-Butyl((E)-5-((3aS,4R,5R,6aS)-5-(tert-butyldimethylsilyl-
oxy)-4-((3S,4S,E)-3-(tert-butyldimethylsilyloxy)-4-methyl-oct-1-en-6-yn-
yl)hexahydropentalen-2(1H)-ylidene))-pentyloxy)dimethylsilane (44).
To a solution of t-BuMe2SiOCH2CH2CH2I (780 mg, 2.60 mmol) in
Et2O (12 mL) was added t-BuLi (3.24 mL of 1.60 M in hexanes, 5.18
mmol) at -78 °C. After the mixture was stirred at this temperature for
0.5 h, it was transferred via a double-ended needle to a solution of CuI
(246 mg, 1.29 mmol) in Et2O (10 mL) and Et2S (2 mL) at -78 °C.
The mixture was stirred at -45 °C for 1 h whereby the color of the
mixture changed from yellow to greenish-brown. Then a solution of
acetate 43 (124 mg, 0.216 mmol) in Et2O (2 mL) was added within 10
min at -45 °C. The reaction mixture was stirred at -45 °C for 3 h.
Then a saturated aqueous NH4Cl (5 mL) was added, and the mixture
was warmed to ambient temperature. The aqueous phase was extracted
with Et2O (3 × 30 mL), and the combined organic phases were dried
(MgSO4) and concentrated in vacuo. Column chromatography (hexanes/
EtOAc, 40:1) gave a mixture of alkene 44 and tBuMe2SiO(CH2)6-
OSitBuMe2. Preparative HPLC (Kromasil Si-100, 30 mm, n-hexane/
EtOAc, 98:2, RI) afforded alkene 44 (122 mg, 82%) as a colorless oil.
[R]D +27.8 (c 1.05, CDCl3); Rf 0.29 (hexanes/ EtOAc, 50:1); 1H NMR
(300 MHz, CDCl3) δ 0.00-0.07 (m, 18 H, 3 × Si(CH3)2), 0.87 (s, 9
H, SiC(CH3)3), 0.89 (s, 9 H, SiC(CH3)3), 0.90 (s, 9 H, SiC(CH3)3),
0.91 (d, J ) 6.9 Hz, 3 H, CHCH3), 1.14-1.31 (m, 1 H), 1.31-1.45
(m, 2 H), 1.45-1.59 (m, 2 H), 1.59-1.76 (m, 1 H), 1.78 (t, J ) 2.5
Hz, 3 H, CtCCH3), 1.88-2.13 (m, 7 H), 2.13-2.28 (m, 2 H), 2.28-
2.45 (m, 3 H), 3.61 (t, J ) 6.2 Hz, 2 H, CH2OSi), 3.74 (dt, J ) 6.9, J
) 8.9 Hz, 1 H, CHOSi), 3.97 (t, J ) 6.2 Hz, 1 H, CHOSi), 5.18-5.27
(m, 1 H, CdCHsCH2OCdO), 5.39 (dd, J ) 6.2, J ) 15.8 Hz, 1 H,
CHdCHsCHOSi), 5.52 (dd, J ) 7.0, J ) 15.8 Hz, 1 H, CHdCHs
CHOSi); 13C NMR (75 MHz, CDCl3) δ -5.2 (d), -4.9 (d), -4.6 (d),
-4.4 (d), -3.9 (d), 3.5 (d), 15.4 (d), 18.1 (u), 18.2 (u), 18.4 (u), 22.1
(u), 25.9 (d), 26.0 (d), 26.1 (u), 29.2 (u), 32.5 (u), 36.0 (u), 37.8 (d),
38.3 (u), 39.9 (d), 42.6 (u), 44.5 (d), 56.1 (d), 63.2 (u), 76.2 (u), 76.3
(d), 78.1 (u), 78.4 (d), 121.7 (d), 131.6 (d), 132.8 (d), 141.9 (u); IR
(CHCl3): ν 2933 (s), 2858 (s), 1667 (w), 1467 (m), 1384 (m), 1254
(s), 1107 (s), 1005 (w), 974 (w), 938 (w), 938 (w), 909 (w), 839 (s)
cm-1; MS (EI, 70 eV) m/z (relative intensity, %) 689 (3), 688 (M+, 5),
633 (25), 632 (55), 631 (100), 609 (16), 608 (36), 607 (71), 557 (16),
556 (26), 500 (26), 499 (57), 476 (21), 475 (46), 450 (12), 449 (30),
425 (21), 367 (12), 293 (28), 251 (10), 225 (17), 211 (11), 185 (15),
183 (27), 172 (10), 171 (66), 159 (14), 147 (29), 145 (12), 133 (19),
115 (11), 105 (11), 91 (14); HRMS calcd for C40H76O3Si3+-C4H9
631.439808, found 631.439840.
Acknowledgment. Financial support of this work by the
Deutsche Forschungsgemeinschaft (Collaborative Research
Center “Asymmetric Synthesis with Chemical and Biological
Methods”) is gratefully acknowledged. We thank Dr. J. Runsink
for NMR spectroscopic investigations, Prof. Dr. C. Bolm and
H. G. Do¨teberg, Gru¨nenthal, Aachen, for supplying use with
preparative Chiralcel AD and OD HPLC columns, Dr. H. Dahl,
Schering AG, Berlin, for a most generous gift of cis-tetrahy-
dropentalene-2,5(1H,3H)-dione, Dr. M. Lerm for a sample of
24, Prof. Dr. A. Fu¨rstner for the NMR data of 16, Professor K.
A. Jørgensen and M. Marigo for copies of NMR spectra of R,R-
4,5-diphenylimidazolidine, B. Sommer and S. Engels for
technical assistance, and Degussa AG, Du¨sseldorf, for D-
phenylalanine.
Supporting Information Available: Schemes S1-S4, Table
S1, general information, experimental procedures and spectro-
scopic data for 16S-2, Z-16S-2, 16S-3, 5, 7, 8, 17, 18, rac-18,
19, rac-19, 23, 26, 27, 30, 31, 32, 34, 35, 36, 37, 38, 40, 43,
acetate of 45, 51, 52, 53, and copies of the NMR spectra of
16S-2, Z-16S-2, 16S-3, 5, E-6, Z-6, 7, 8, 26, 27, 30, 31, 32, 34,
35, 36, 38, 40, 43, 44, 45, 52, and 53. This material is available
free of charge via the Internet at http://pubs.acs.org.
(+)-(E)-5-((3aS,4R,5R,6aS)-5-(tert-Butyldimethylsilyloxy)-4-
((3S,4S,E)-3-(tert-butyldimethylsilyloxy)-4-methyl-oct-1-en-6-ynyl)-
hexahydropentalen-2(1H)-ylidene)-pentan-1-ol (45). To a solution
of the silyl ether 44 (118 mg, 0.171 mmol) in hexanes (18 mL) was
added neutral alumina (with 3% of H2O, 70-230 mesh) (11 g). The
mixture was stirred at ambient temperature for 24 h, then filtered, and
the alumina washed with hexanes (20 mL) and the organic phase was
JA0558037
9
17920 J. AM. CHEM. SOC. VOL. 127, NO. 50, 2005