Novel and flexible entries into prostaglandins and analogues based on ring closing alkyne metathesis or alkyne cross metathesis

Article abstract of DOI:10.1021/ja003119g

The suitably functionalized cyclopentanone derivatives 12, 13, 19, and 37 serve as common precursors for the synthesis of various prostaglandins, prostaglandin-1,15-lactones, and unnatural analogues thereof. All of them contain a 2-butynyl entity which is

Full text of DOI:10.1021/ja003119g

J. Am. Chem. Soc. 2000, 122, 11799-11805
11799
Novel and Flexible Entries into Prostaglandins and Analogues Based
on Ring Closing Alkyne Metathesis or Alkyne Cross Metathesis
Alois Fu1rstner,* Karol Grela, Christian Mathes, and Christian W. Lehmann
Contribution from the Max-Planck-Institut fu¨r Kohlenforschung, D-45470 Mu¨lheim/Ruhr, Germany
ReceiVed August 21, 2000
Abstract: The suitably functionalized cyclopentanone derivatives 12, 13, 19, and 37 serve as common precursors
for the synthesis of various prostaglandins, prostaglandin-1,15-lactones, and unnatural analogues thereof. All
of them contain a 2-butynyl entity which is elaborated into the intact R side chain of the targets either via a
sequence comprising ring closing alkyne metathesis/Lindlar reduction or via alkyne cross metathesis (ACM)/
Lindlar reduction. These novel approaches are distinguished by (i) the ready accessibility of the required
cyclopentenone substrates via a three-component coupling reaction, (ii) the inherent flexibility which allows
one to make a series of analogues starting from these common platforms, (iii) a small number of steps, and
(iv) an excellent overall yield. The key alkyne metathesis reactions are efficiently catalyzed either by the
tungsten alkylidyne complex (t-BuO)3WtCCMe3 or, preferentially, by a catalyst formed in situ from Mo-
[N(t-Bu)(Ar)]3 and CH2Cl2, the reactivity of which can be fine-tuned by varying the Ar substituent on the
amido ligands. These organometallic tools exhibit a remarkable application profile, tolerate an array of polar
groups, rigorously distinguish between different π-electron systems, and catalyze the reactions under conditions
that are sufficiently mild to preserve even highly sensitive functionalities. The structures of the macrocyclic
prostaglandin lactone derivatives 22 and 32 were characterized by X-ray crystallography.
Introduction
Prostaglandins (PG’s) are signaling compounds of tremendous
importance in mammals and many other animals controlling
biological functions as diverse as smooth muscle contraction,
platelet aggregation, broncial and vascuolar constriction, chemo-
taxis, ion secretion (stomach acidity), reproduction, immune
responses, and neuro-endocrinal processes, just to mention the
most important effects.1 Due to their exceptional potency in vivo,
only minute amounts of PG’s are present in the organism; in
general, these secondary metabolites are biosynthesized de novo
from arachidonic acid as the common precursor following
chemical, immunological, or mechanical stimulation and are
metabolized immediately after eliciting the respective biological
response.1
In view of this well-established mode of action as local tissue
hormones that are formed “on demand” and “on site”, it was
most surprising to find that the marine nudibranch Tethys fimbria
produces fairly large amounts of prostaglandin-1,15-lactones
such as 1-3 and accumulates them in the cerata (dorsal
appendages) and ovotestis.2,3 This is the only example for an in
ViVo storage of PG’s reported in the literature. These molluscs
seem to have developed a rather sophisticated and economical
way of exploiting these unusual metabolites for more than one
purpose.2 Specifically, 1-3 are present in high concentrations
in the mucous secretion that is immediately released on
mechanical molestation. In view of their ichtyotoxic properties
(activity against the mosquito fish Gambusia affinis at concen-
trations of 1-10 µg/mL), the lactones themselves play an
important role in the chemical defense mechanism of T. fimbria
against potential predators. Moreover, hydrolytic enzymes
(1) See the following for leading and comprehensive treatises: (a)
Prostaglandins, Leucotrienes and Other Eicosanoids. From Biogenesis to
Clinical Applications; Marks, F., Fu¨rstenberger, G., Eds.; Wiley-VCH:
Weinheim, 1999. (b) Collins, P. W.; Djuric, S. W. Chem. ReV. 1993, 93,
1533. (c) Handbook of Eicosanoids: Prostaglandins and Related Lipids;
Willis, A. L., Ed.; CRC Press: Boca Raton, 1987. (d) Schro¨r, K.
Prostaglandine und Verwandte Verbindungen; Thieme: Stuttgart, 1984. (e)
Samuelsson, B. Angew. Chem. 1983, 95, 854; Angew. Chem., Int. Ed. Engl.
1983, 22, 805. (f) Bergstro¨m, S. Angew. Chem. 1983, 95, 865; Angew.
Chem., Int. Ed. Engl. 1983, 22, 858. (g) Vane, J. R. Angew. Chem., Int.
Ed. Engl. 1983, 22, 741. (h) Horton, E. W. Chem. Soc. ReV. 1975, 4, 589.
(i) Ramwell, P. W.; Shaw, J. E.; Corey, E. J.; Andersen, N. Nature 1969,
221, 1251.
(2) (a) Cimino, G.; Crispino, A.; Di Marzo, V.; Sodano, G.; Spinella,
A.; Villani, G. Experientia 1991, 47, 56. (b) Cimino, G.; Spinella, A.;
Sodano, G. Tetrahedron Lett. 1989, 30, 3589. (c) Cimino, G.; Crispino,
A.; Di Marzo, V.; Spinella, A.; Sodano, G. J. Org. Chem. 1991, 56, 2907.
(d) Marin, A.; Di Marzo, V.; Cimino, G. Marine Biol. 1991, 111, 353. (e)
Di Marzo, V.; Minardi, C.; Vardaro, R. R.; Mollo, E.; Cimino, G. Comp.
Biochem. Physiol. B: Biochem. Mol. Biol. 1992, 101, 99. (f) De Petrocellis,
L.; Di Marzo, V. Prostaglandins Leucotrienes Ess. Fatty Acids 1994, 51,
215. (g) Di Marzo, V.; Cimino, G.; Crispino, A.; Minardi, C.; Sodano, G.;
Spinella, A. Biochem. J. 1991, 273, 593.
(3) See the following for a comprehensive review on prostanoids of
marine origin: Gerwick, W. H. Chem. ReV. 1993, 93, 1807.
10.1021/ja003119g CCC: $19.00 © 2000 American Chemical Society
Published on Web 11/17/2000

11800 J. Am. Chem. Soc., Vol. 122, No. 48, 2000
Fu¨rstner et al.
readily convert them into the parent prostaglandins (e.g. 1 f
4) which are then used to control smooth muscle contractions
of the cerata extruding the defense secretion. Finally, fatty acid
esters of 3 likely serve as chemical release factors for ovulation
and therefore exert important functions in the reproduction of
these animals.2
Scheme 1
PG-1,15-lactones have also caught the attention of medicinal
chemists.4 The widespread use of prostaglandin esters as
prodrugs suggests that these intramolecular versions may serve
similar purposes. Since their polarities and shapes are consider-
ably different from those of the uncyclized hydroxy acids, they
may go unrecognized by normal PG processing enzymes;
depending on the different esterase activity in various tissues,
however, one may envisage a localized release of the parent
prostaglandin and thereby trigger a sustained onset of its specific
biological function.5
Intrigued by these diverse and promising physiological and
pharmacological properties, we have engaged in a program
aimed at the total synthesis of this class of natural products.6
Described below is a comprehensive summary of our results in
this field which led to the total synthesis of prostaglandin
E2-1,15-lactone (1), 15-epi-prostaglandin E2-1,15-lactone (22),
and the parent prostaglandin E2 (4) itself, as well as to a formal
total synthesis of prostaglandin A2-1,15-lactone (2). Although
1 can also be prepared from PGE2 4 by macrolactonization
reactions,7 we believe that the use of alkyne metathesis as a
fundamentally different and novel concept for the construction
of such target molecules offers significant advantages, including
a high overall efficiency, an excellent “economy of steps”,8 and
an inherent flexibility in structural terms. The latter aspect is
illustrated by the synthesis of several previously unknown PG
analogues differing from the natural product in the R side chain
which is particularly difficult to alter and manipulate by more
conventional approaches.
(E)- and (Z)-isomers, with the former dominating in most of
the recorded cases.9,10 This tendency to form product mixtures
constitutes an obvious drawback in target-oriented synthesis.
To circumwent this inherent problem, we have proposed an
indirect but stereoselectiVe approach to macrocyclic (Z)-alkenes
which consists of a ring-closing metathesis of diyne substrates
followed by Lindlar reduction of the resulting cycloalkyne
products (Scheme 1).11-14 Three different catalyst systems have
been used so far for this purpose, including (1) the tungsten
alkylidyne complex (t-BuO)3WtCCMe3 developed by Schrock,15
(2) a structurally unknown catalyst formed in situ from
Mo(CO)6 and p-chlorophenol (or related phenol additives),16
and (3) Mo[N(t-Bu)(Ar)]3 5 or 6 activated in situ by means of
CH2Cl2.13,17
Results and Discussion
Although all previous applications of these tools are very
promising,11-14,18 alkyne metathesis in general is still in its
infancy as compared with alkene metathesis and the envisaged
extension of this methodology to a total synthesis of 1 and
related targets therefore raises several new questions: Thus, it
Strategy. Ring closing metathesis of dienes (RCM) has
recently gained considerable importance as a tool for organic
synthesis.9,10 Among the few shortcomings that infringe upon
the superb overall application profile of RCM, the lack of control
over the configuration of the newly formed double bond is most
noteworthy if the reaction is applied to the macrocyclic series.
The products formed are usually obtained as mixtures of the
(11) Fu¨rstner, A.; Seidel, G. Angew. Chem. 1998, 110, 1758; Angew.
Chem., Int. Ed. Engl. 1998, 37, 1734.
(12) Fu¨rstner, A.; Guth, O.; Rumbo, A.; Seidel, G. J. Am. Chem. Soc.
1999, 121, 11108.
(13) Fu¨rstner, A.; Mathes, C.; Lehmann, C. W. J. Am. Chem. Soc. 1999,
121, 9453.
(14) (a) Fu¨rstner, A.; Rumbo, A. J. Org. Chem. 2000, 65, 2608. (b)
Fu¨rstner, A.; Seidel, G. J. Organomet. Chem. 2000, 606, 75. (c) Fu¨rstner,
A.; Dierkes, T. Org. Lett. 2000, 2, 2463.
(15) (a) Schrock, R. R.; Clark, D. N.; Sancho, J.; Wengrovius, J. H.;
Rocklage, S. M.; Pedersen, S. F. Organometallics 1982, 1, 1645. (b)
Freudenberger, J. H.; Schrock, R. R.; Churchill, M. R.; Rheingold, A. L.;
Ziller, J. W. Organometallics 1984, 3, 1563. (c) Listemann, M. L.; Schrock,
R. R. Organometallics 1985, 4, 74. (d) Schrock, R. R. Polyhedron 1995,
14, 3177.
(4) (a) Bundy, G. L.; Peterson, D. C.; Cornette, J. C.; Miller, W. L.;
Spilman, C. H.; Wilks, J. W. J. Med. Chem. 1983, 26, 1089. (b) Bundy, G.
L.; Morton, D. R.; Peterson, D. C.; Nishizawa, E. E.; Miller, W. L. J. Med.
Chem. 1983, 26, 790. (c) Spilman, C. H.; Beuving, D. C.; Forbes, A. D.;
Kimball, F. A. Prostaglandins 1977, 14, 477. (d) For related 1,20-lactones
see: Andersen, N. H.; Imamoto, S.; Subramanian, N. Prostaglandins 1981,
22, 831.
(5) 1 decreases gastric secretion in dogs by 90% when administered
intravenously at a dose of 100 µg/kg. Its antifertility activity is essentially
equal to that of parent PGE2 4 released by rapid enzyme catalyzed
hydrolysis. Therefore it has been concluded (ref 4) that PGE-lactones
themselves essentially lack intrinsic activity.
(6) For a preliminary communication see: Fu¨rstner, A.; Grela, K. Angew.
Chem. 2000, 112, 1292; Angew. Chem., Int. Ed. Engl. 2000, 39, 1234.
(7) (a) Corey, E. J.; Nicolaou, K. C.; Melvin, L. S. J. Am. Chem. Soc.
1975, 97, 653. (b) Narasaka, K.; Maruyama, K.; Mukaiyama, T. Chem.
Lett. 1978, 885. (c) See also ref 4.
(8) For a discussion of these and related strategic goals see: Fu¨rstner,
A. Synlett 1999, 1523.
(16) For early examples of scrambling or dimerization of simple alkynes
by structurally undefined catalyst systems see: (a) Mortreux, A.; Blanchard,
M. J. Chem. Soc., Chem. Commun. 1974, 786. (b) Villemin, D.; Cadiot, P.
Tetrahedron Lett. 1982, 23, 5139. (c) Kaneta, N.; Hirai, T.; Mori, M. Chem.
Lett. 1995, 627. (d) Kaneta, N.; Hikichi, K.; Asaka, S.-I.; Uemura, M.;
Mori, M. Chem. Lett. 1995, 1055. (e) Pschirer, N. G.; Bunz, U. H. F.
Tetrahedron Lett. 1999, 40, 2481.
(9) For the most recent review see: Fu¨rstner, A. Angew. Chem. 2000,
112, 3140; Angew. Chem., Int. Ed. Engl. 2000, 39, 3012.
(10) (a) Grubbs, R. H.; Chang, S. Tetrahedron 1998, 54, 4413. (b)
Fu¨rstner, A. Top. Catal. 1997, 4, 285. (c) Schuster, M.; Blechert, S. Angew.
Chem., Int. Ed. Engl. 1997, 36, 2036. (d) Fu¨rstner, A. Top. Organomet.
Chem. 1998, 1, 37. (e) Ivin, K. J.; Mol, J. C. Olefin Metathesis and
Methathesis Polymerization, 2nd ed.; Academic Press: New York, 1997.
(17) (a) See the following for an excellent review on the preparation of
5 and its reactions with small inorganic molecules: Cummins, C. C. Chem.
Commun. 1998, 1777. (b) Laplaza, C. E.; Odom, A. L.; Davis, W. M.;
Cummins, C. C.; Protasiewicz, J. D. J. Am. Chem. Soc. 1995, 117, 4999
and literature cited therein.
(18) For a short review on alkyne metathesis see: Bunz, U. H. F.;
Kloppenburg, L. Angew. Chem., Int. Ed. Engl. 1999, 38, 478.

NoVel and Flexible Entries into Prostaglandins
J. Am. Chem. Soc., Vol. 122, No. 48, 2000 11801
Scheme 2
Scheme 3
was not clear at the beginning of this investigation if the rather
electrophilic ketone group of 1 will survive alkyne metathesis
or if lengthy protecting group manipulations are required; this
aspect deserves particular consideration since tungsten alkyli-
dynes are known to be sufficiently nucleophilic to react, e.g.,
with acetone and related carbonyl compounds in a Wittig-type
process.19 The reactivity of the other alkyne metathesis catalysts
mentioned above toward ketones has not been studied at all.
Moreover, the aldol substructure of PGE2 is known to be quite
labile toward acid as well as base,20 thus rendering the envisaged
cyclization of the macrocyclic ring at the ∆5,6-bond (Scheme
2) a stringent test for the mildness of the method. Finally, it
should be noted that the catalyst has to distinguish between the
alkyne groups and the preexisting (E)-alkene in the cyclization
precursor to be useful in this preparative context.
Total Synthesis of (-)-Prostaglandin E2-1,15-Lactone.
Despite these issues, we have engaged in the total synthesis of
(-)-1 summarized in Scheme 3. “Three-component coupling”21
as the most elegant method for the construction of prostaglandin
skeletons efficiently provides the required cyclization precursor
15. Thus, commercially available propargyl alcohol 7 (ee 98%)
is protected as a triethylsilyl (TES) ether prior to conversion
into vinylstannane 8 under free radical conditions.22 Tin for
lithium exchange on exposure of compound 8 to n-BuLi at low
temperature followed by addition of Me2Zn affords a zincate
intermediate that undergoes smooth conjugate addition to the
known Michael acceptor 9 (ee 94%).23 Enolate 10 thus formed
is quenched with propargyl iodide 1124 to afford adduct 12.25
Selective deprotection of the TES group of the crude material
with dilute HOAc in aqueous THF delivers compound 13 in
diastereomerically and enantiomerically pure form in 74% yield
(19) (RO)3WtCCMe3 (R ) 2,6-diisopropylphenyl) reacts rapidly with
acetone, benzaldehyde, paraformaldehyde, ethyl formate, DMF, and aceto-
nitrile in Wittig-like reactions, cf.: Freudenberger, J. H.; Schrock, R. R.
Organometallics 1986, 5, 398.
over both steps (ee g99%).26a This excellent result reflects the
high reactivity of 11 as the electrophile in this three-component
assembly (vide infra). Subsequent esterification of the free -OH
group of 13 with 5-heptynoic acid 1424b in the presence of
diisopropyl carbodiimide and catalytic amounts of DMAP
readily provides diyne 15 and sets the stage for the crucial
macrocyclization via ring-closing alkyne metathesis.
Gratifyingly, this key transformation performs very well using
the catalyst formed in situ from Mo[N(t-Bu)(Ar)]3 (Ar ) 3,5-
dimethylphenyl) 5 and CH2Cl2 in toluene at 80 °C,13 affording
(20) (a) Outside of the pH range of ca. 5-8, E-type prostaglandins suffer
facile dehydration to give compounds of the A series which further isomerize
when exposed to base, cf.: Andersen, N. H. J. Lip. Res. 1969, 10, 320. (b)
This intrinsic lability consitutes a severe preparative constraint that has to
be taken into account in all routes to these targets. For a classical example
see Corey’s seminal total synthesis of PG’s, cf.: Corey, E. J.; Andersen,
N. H.; Carlson, R. M.; Paust, J.; Vedejs, E.; Vlattas, I.; Winter, R. E. K. J.
Am. Chem. Soc. 1968, 90, 3245. (c) Corey, E. J.; Vlattas, I.; Andersen, N.
H.; Harding K. J. Am. Chem. Soc. 1968, 90, 3247. (d) Corey, E. J.; Cheng,
X.-M. The Logic of Chemical Synthesis; Wiley: New York, 1989.
(21) See the following for pertinent reviews on “three-component
coupling”: (a) Noyori, R.; Suzuki, M. Angew. Chem. 1984, 96, 854; Angew.
Chem., Int. Ed. Engl. 1984, 23, 847. (b) Noyori, R.; Suzuki, M. Chemtracts-
Org. Chem. 1990, 3, 173. (c) Noyori, R. Asymmetric Catalysis in Organic
Synthesis; Wiley: New York, 1994; pp 298-322.
(22) Chen, S.-M. L.; Schaub, R. E.; Grudzinskas, C. V. J. Org. Chem.
1978, 43, 3450. This paper pretends that the hydrostannylation of the TES-
ether of alcohol 7 is regio- and diastereoselective according to 13C NMR
data. Careful analysis of the crude product by HPLC, however, shows that
the purity is only ca. 90%. This material can be used directly in the next
step, because the isomeric byproducts do not undergo productive 1,4-
addition, cf.: Johnson, C. R.; Penning, T. D. J. Am. Chem. Soc. 1988, 110,
4726.
(24) (a) Iodide 11 is conveniently prepared by treatment of commercially
available 2-butyn-1-ol with PPh3, I2, and imidazole according to a literature
procedure: Lange, G. L.; Gottardo, C. Synth. Commun. 1990, 20, 1473.
(b) Acid 14 is prepared from commercially available 3-pentyn-1-ol as
described in: Ansell, M. F.; Emmet, J. C.; Coombs, R. V. J. Chem. Soc. C
1968, 217.
(25) The three-component coupling was essentially carried out as
described in: Suzuki, M.; Morita, Y.; Koyano, H.; Koga, M.; Noyori, R.
Tetrahedron 1990, 46, 4809. For details see the Experimental Part in the
Supporting Information.
(23) (a) For the preparation of enone 9 see: Noyori, R.; Tomino, I.;
Yamada, M.; Nishizawa, M. J. Am. Chem. Soc. 1984, 106, 6717. (b) Forsyth,
C. J.; Clardy, J. J. Am. Chem. Soc. 1990, 112, 3497.
(26) (a) Determined by HPLC on a Chiracel OD-H column with
n-heptane/2-propanol as the mobile phase. (b) Determined by HPLC on a
Chiralpak AD column with n-heptane/2-propanol (95:5) as the mobile phase.

11802 J. Am. Chem. Soc., Vol. 122, No. 48, 2000
Fu¨rstner et al.
Table 1. Evaluation of Different Catalysts (10 mol % Each) for
the Cyclization of Diyne 19 to Cycloalkyne 20
Scheme 4
entry
catalyst
5/CH2Cl2
6/CH2Cl2
(t-BuO)3WtCCMe3
Mo(CO)6/p-ClC6H4OH
(1 equiv)
conditions
yield, %
1
2
3
4
toluene, 80 °C, 8 h
toluene, 80 °C, 8 h
toluene, 80 °C, 8 h
chlorobenzene,
81
87
65a
0b
130 °C, 24 h
a In addition, 4% of the substrate has been recovered by flash
chromatography. b Slow decomposition of the substrate is observed.
Table 2. Preparation of Truncated PG Analogues by Alkyne Cross
Metathesis (ACM) of Substrate 37 with Different Internal Alkynes
Using 5 (10 mol %) in CH2Cl2/Toluene as the Catalyst
Standard Lindlar hydrogenation of cycloalkyne 16 thus
formed followed by deprotection of the residual TBS ether with
aqueous HF in acetonitrile28 completes this total synthesis of
the marine natural product (-)-prostaglandin E2-1,15-lactone
1 which is obtained in respectable 28% overall yield. Its 600
MHz NMR spectra are in full agreement with the data reported
in the literature (for details see the Experimental Section).2 By
virtue of the known enzymatic hydrolysis of this compound into
prostaglandin E2 4 as well as by the known transformation of
1 into PGA2-1,15-lactone 2,4 formal total syntheses of these
interesting natural products have also been accomplished (vide
infra for an alternative approach to PGE2).
Finally it should be mentioned that the order of Lindlar
reduction and deprotection can also be inverted (Scheme 4).
Thus, cycloalkyne 16 can first be deprotected to afford free
alcohol 18 which constitutes an acetylenic congener of 1.
Biological studies of this and other analogues (see below)
prepared during this study are underway. Lindlar reduction then
completes a second route to 1. Given the usefulness of
derivatives with a triple bond at the 5,6-position in the R-chain
as common intermediates for the preparation of various natural
and unnatural PG’s, most notably of prostacyclin and derivatives
thereof,29 product 16 also opens many possibilities for further
transformations which are presently pursued in this labora-
tory.
Catalyst Optimization: 15-epi-Prostaglandin E2-1,15-
Lactone. Having in mind that the configuration at C-15 is a
critical determinant for PG-receptor affinity and specificity,1 we
have prepared the 15-epi-congener of 1, i.e., lactone 22 (Scheme
5), by following a route similar to the one outlined above.30
The crucial cyclization of diyne 19 to cycloalkyne 20 was also
taken as an opportunity to compare the performance of different
alkyne metathesis catalysts to gain a better understanding of
their specific application profiles in sensitive cases. The results
are compiled in Table 1 and deserve some comments.
a The rest consists of unreacted starting material that can be recovered
by FC.
Table 3. Compilation of Functional Groups Presently Known To
Be Compatible with Alkyne Metathesis Catalysts
Mo[N(t-Bu)(Ar)]3/CH2Cl2 (ref)
(t-BuO)3WtCCMe3 (ref)
ketone, alkyl chloride, nitrile,
alkene (this paper)
ester, tert-amide, ether, silyl ether,
acetal, thioether, pyridine (13)
nitro group, enoate, aldehyde, sulfone,
sulfonamide, glycoside (27)
ester, enoate, ketone, amide,
urethane, ether, alkene,
sulfone, silyl ether,
sulfonamide, acetal,
furan (11, 12, 14)
the desired cycloalkyne 16 in 68-73% isolated yield (4 runs).
The modified catalyst 6 in which the 3,5-dimethylphenyl group
on the amido ligands is replaced by a 4-fluorophenyl residue27
gives an even higher yield (77%). Careful chromatographic
inspection of the reaction mixture reveals that no racemization
takes place before or after ring closure and that the ee values
of the substrate (ee 99%)26a and of the product (ee 98%)26b are
also virtually identical.
This example highlights the excellent overall application
profile of this type of catalyst which is very reactive toward
alkynes but highly tolerant with an array of polar functional
groups (cf. Table 3). The transformation occurs under excep-
tionally mild conditions allowing one to preserve very labile
structural elements and guaranteeing the integrity of chiral
centers even if positioned R to a carbonyl group. Although the
precise mode of action of 5 still constitutes a subject of ongoing
investigations,27 it is remarkable that this species rigorously
distinguishes between different π-systems, since it selects
exclusively for triple bonds (reactive) whereas preexisting
double bonds in the substrate are inert.
In line with our expectations, the catalyst formed in situ from
5 and CH2Cl2 gave excellent results (entry 1).13 TLC control
shows a smooth and quantitative spot-to-spot conversion of the
substrate into the desired product, which is obtained in 81%
isolated yield. It is also interesting to note that the reactivity of
the catalyst can be tuned to some extend by changing the
electronic properties of its amido ligands. Thus, catalyst 6 in
which the Ar ) 3,5-dimethylphenyl substituents originally
(28) Newton, R. F.; Reynolds, D. P.; Webb, C. F.; Roberts, S. M. J.
Chem. Soc., Perkin Trans. 1 1981, 2055.
(29) A short summary of how to use acetylenic PG derivatives in further
transformations is given in ref 21c.
(30) Three-component coupling of racemic 8, 9, and 11 followed by
deprotection of the TES group provided a ca. 1:1 mixture of (()-13 and
(()-epi-13. These compounds can be easily separated by flash chromato-
graphy. Esterification of (()-epi-13 with 5-heptynoic acid afforded (()-
19, which was employed in this study. For details see the Supporting
Information.
(27) Mathes, C. Projected Ph.D. Thesis, University of Dortmund,
Germany.

NoVel and Flexible Entries into Prostaglandins
J. Am. Chem. Soc., Vol. 122, No. 48, 2000 11803
Scheme 5
Figure 1. ORTEP diagram of the molecular structure of compound
22. Anisotropic replacement parameter ellipsoids are drawn at 50%
probability; hydrogen atoms are omitted for clarity.
trans alkyl chain extends from the equatorial position of C9 to
the methyl terminus of C1. It is noteworthy that the ester group
is perpendicular to the main ring plane of the 13-membered
macrocycle. The torsion angles adjacent to the Z-alkene entity
C15-C16 are 80° and 98°, respectively, forcing the alkyl chains
to opposite faces of the double bond plane. For further details
see the Supporting Information.
Prostaglandin Lactone Analogues. The recent cloning of
several PG receptor subtypes33 has evoked considerably renewed
interest toward the development of subtype specific PG deriva-
tives with improved phamacological indices.34,35 It is well
established that the substitution pattern and stereochemical
display around the cyclopentane ring and, in particular, the
constitution of the side chains exert subtle effects on the PG
receptor affinity and specificity.1 Although prostaglandins in
general mediate a host of responses and some derivatives have
been successfully marketed as drugs for various clinical ap-
plications,1 much effort has recently centered on the identifica-
tion of PG derivatives relieving interocular pressure as caused
by glaucoma.36
employed are replaced by 4-fluorophenyl residues gave con-
sistently the best results (entry 2) (see also the previous and
the following sections of this paper). This finding confirms the
conclusion reached in an independent investigation that the
electrophilicity at Mo is a key parameter for adjusting the
reactivity of complexes of the general type Mo[N(tBu)(Ar)]3.27
Schrock’s tungsten catalyst (tBuO)3WtCCMe315 also turned
out to be compatible with the sensitive functional group pattern
of substrate 19 (entry 3). Particularly noteworthy is the fact that
the ketone groups in 19 and product 20 formed thereof are at
least kinetically inert toward this alkylidyne complex, although
closely related species are known to react rapidly with various
carbonyl derivatives.19 Unfortunately, however, the conversion
of 19 into 20 did not go to completion even on prolonged
reaction time employing high catalyst loadings. Difficulties in
separating the product from traces of unreacted diyne 19 (4%
have been recovered by flash chromatography) account for the
somewhat lower yield (65%) obtained with this otherwise quite
efficient catalyst.
In contrast to these successful initators, the“in situ” catalyst
system employing Mo(CO)6 and phenol additives is totally
unsuitable for applications to the prostaglandin series (entry 4).
Although this method deserves attention for its user-friendly
setup,16 the vigorous conditions required (130 °C in chloroben-
zene) preclude applications to elaborate and sensitive starting
materials such as 19. Therefore it must be concluded that the
scope of this “instant method” is likely restricted to rather robust
cases.12,31
In principle, “three-component coupling” as a triply conver-
gent method allows substantial alterations of all segments of a
PG.21 In practice, however, variations of the R-chain are known
to be problematic. Difficulties arise from undesirable equilibra-
tions and â-alkoxide elimination if the intermediate enolates
such as 10 must be trapped with poorly electrophilic R-chain
equivalents, some of which are rather unstable under reaction
conditions and must be used in fairly large excess.37
(32) (a) There are only a few other crystal structures of 13-membered
lactones known so far. The closest relatives to one reported here are some
brefeldin A derivatives [cf. b-d], also including a trans-disubstituted five-
membered ring. For a comparison of the structures of 22 and one of these
compounds see the Supporting Information. (b) Weber, H. P.; Hauser, D.;
Sigg, H. P. HelV. Chim. Acta 1971, 54, 2763. (c) Feldman, K. S.; Berven,
H. M.; Romanelli, A. L.; Parvez, M. J. Org. Chem. 1993, 58, 6851. (d)
Argade, A. B.; Haugwitz, R. D.; Devraj, R.; Kozlowski, J.; Fanwick, P. E.;
Cushman, M. J. Org. Chem. 1998, 63, 273.
(33) See the following for leading references and literature cited
therein: (a) Hirata, M.; Hayashi, Y.; Ushikubi, F.; Yokota, Y.; Kageyama,
R.; Nakanishi, S.; Narumiya, S. Nature 1991, 349, 617. (b) Coleman, R.
A.; Smith, W. L.; Narumiya, S. Pharmacol. ReV. 1994, 46, 205. (c) Cloning
of the gene encoding for the human PGE receptor: J. Biol. Chem. 1993,
268, 26767.
(34) For a pertinent example and a detailed discussion of the background
see: Suzuki, M.; Noyori, R.; Langstro¨m, B.; Watanabe, Y. Bull. Chem.
Soc. Jpn. 2000, 73, 1053.
(35) For studies along these lines employing combinatorial approaches
see: (a) Thompson, L. A.; Moore, F. L.; Moon, Y. C.; Ellman, J. A. J.
Org. Chem. 1998, 63, 2066. (b) Dragoli, D. R.; Thompson, L. A.; O’Brien,
J.; Ellman J. A. J. Comb. Chem. 1999, 1, 534. (c) Lee, K. J.; Angulo, A.;
Ghazal, P.; Janda, K. D. Org. Lett. 1999, 1, 1859.
Crystals of compound 22 suitable for X-ray analysis have
been obtained. This structure (Figure 1) is the first example of
a prostaglandin lactone to be described.32 Its 4-hydroxycyclo-
pentanone ring adopts an envelope conformation and an all-
(31) For a recent successful application in a cyclodimerization reaction
see: Pschirer, N. G.; Fu, W.; Adams, R. D.; Bunz, U. H. F. Chem. Commun.
2000, 87.
(36) Woodward, D. E. US Patent 5,462,968, Oct. 31, 1995.
(37) For a detailed discussion see: Gooding, O. W.; Beard, C. C.; Cooper,
G. F.; Jackson, D. Y. J. Org. Chem. 1993, 58, 3681.

11804 J. Am. Chem. Soc., Vol. 122, No. 48, 2000
Fu¨rstner et al.
Scheme 6
Figure 2. ORTEP diagram of the molecular structure of compound
32. Anisotropic replacement parameter ellipsoids are drawn at 50%
probability; hydrogen atoms are omitted for clarity.
Scheme 7
The inherent flexibility of our approach may help to overcome
this limitation. In this particular case, the electrophile used for
the three-component assembly process is the cheap and readily
available propargyl iodide 1124 which shows excellent reactivity.
We reasoned that product 13 thus obtained can be used as a
common platform for the synthesis of PG analogues simply by
esterification of the 15-OH group with various alkynoic acids.
Ring closure of the resulting diynes then installs a modified R
side chain; subsequent enzyme-mediated hydrolysis of the
lactones gives access to the free hydroxy acids, if desirable.
Furthermore, if PG-lactones are to be used as intramolecular
prodrugs enabling sustained release in the body, modifications
of the ester group and the size of the macrolactone ring should
allow a fine-tuning of the hydrolysis rate and hence the
bioavailability of the active component.4
To prove the viability of this concept, we have prepared two
PG-lactone analogues in addition to 15-epi-PGE2-1,15-lactone
22 and the acetylenic congeners 18 and 23, respectively. Thus,
esterification of alcohol 13 with 9-undecynoic acid in the
membered macrocyclic lactone is characterized by a planar (rms
deviation 0.04 Å) moiety extending from the 1,2-substituted
benzo ring to C18 on one side and to the tertiary carbon (C6)
on the other. The remaining portion of the ring can also be
approximated by a plane (rms 0.34 Å), forming an angle of
157° with the former. The conformation of this part of the
macrocycle is governed by the geometry of the annulated
4-hydroxycyclopentan-1-one ring, which adopts an envelope
conformation with C9 forming the flap. The alkyl chain between
C9 and C6 extending from an equatorial position of C9 includes
the E double bond C7-C8 and is all-trans. The other alkyl chain
between C13 and C18 contains the Z-alkene which is embedded
into the macrocycle with the adjacent torsion angles (C13-
C14-C15-C16 and C15-C16-C17-C18) of 125° and -141°,
respectively. A final gauche conformation C18-C19 completes
the macrocycle. The pentyl chain (all-trans except for the
i
presence of PrNdCdNiPr and DMAP catalyst delivers diyne
24 which cyclizes to afford cycloalkyne 25. Lindlar hydrogena-
tion of its triple bond followed by deprotection under standard
conditions affords the ring-expanded lactone 27 in excellent
overall yield (Scheme 6). Similarly, esterification of 13 with
the benzoic acid derivative 28, subsequent cyclization of diyne
29 by alkyne metathesis, and elaboration of the resulting
cycloalkyne 30 as described above readily provide the aromatic
ester analogue 32 (Scheme 7). In both cases, the crucial ring
closure has been carried out with the optimized molybdenum
amido catalyst formed in situ from complex Mo[N(tBu)-
(C6H4F)]3 6 and CH2Cl2 in toluene at 80 °C.
It was possible to obtain crystals of 32 suitable for X-ray
analysis (Figure 2). The molecular conformation of this 16-

NoVel and Flexible Entries into Prostaglandins
J. Am. Chem. Soc., Vol. 122, No. 48, 2000 11805
Scheme 8
Scheme 9
yield. Subsequent cross metathesis of this “truncated” prostanoid
with variously substituted internal alkynes catalyzed by
5/CH2Cl2 in toluene at 80 °C affords PG analogues 38-41 in
respectable yields (Table 2). Again, homodimerization of 37
does not interfere. These examples further illustrate the excellent
chemoselectivity of the catalyst which is fully compatible with
the alkyl chloride, acetal, nitrile, ester, ketone, and silyl ether
functions in the reaction partners.
Conclusions
Two novel concepts for the formation of natural and unnatural
prostaglandins have been described. The first one is based upon
a ring-closing alkyne metathesis reaction of suitable diyne
precursors, providing ready access to prostaglandin lactones.
These unusual natural products are the only “stock form” for
prostaglandins found in nature so far; they deserve attention on
their own right for their interesting biological properties and
constitute potential prodrugs for pharmaceutical applications.
The other approach makes strategic use of alkyne cross
metathesis (ACM) and represents the first application of this
type of transformation to natural product synthesis. Since both
entries into PG’s are inherently flexible and allow systematic
structural variations of the products, they may become useful
tools for lead optimization in a medicinal chemistry context.
The fact that alkyne metathesis allows the high-yielding
formation of very sensitive prostaglandin derivatives clearly
attests to the exceptional mildness of this emerging method-
ology.11-14 Furthermore, the (cyclo)alkynes thus obtained
provide a convenient solution for the yet unsolved problem of
conventional RCM regarding the formation of stereodefined
(cyclo)alkene products. The catalysts for alkyne metathesis
presently available, in particular the electronically tunable one
formed in situ from Mo[N(tBu)(Ar)]3 and CH2Cl2,13 are not only
efficient in terms of yield and reaction rate, but exhibit a
remarkable selectivity pattern which qualifies them beyond
doubt for even more demanding applications. To facilitate
retrosynthetic planning, Table 3 compiles those functional
groups that are presently known to be compatible with these
novel tools. Further applications of alkyne metathesis as well
as studies of the biological properties of some products described
in this paper are underway and will be disclosed soon.
terminating gauche conformation) points to the exo-face of the
macrocycle. For further details see the Supporting Information.
Alkyne Cross Metathesis (ACM). Synthesis of (-)-PGE2
and Further Prostaglandin Analogues. Compound 12, readily
available by an efficient three-component coupling reaction as
described above, opens yet another possibility to prepare intact
prostaglandins via alkyne cross metathesis (ACM). This type
of transformation has received very little attention so far and
was essentially confined to the preparation of simple acetylenic
derivatives by means of the rather forcing Mo(CO)6/phenol
methodology.16c,d To the best of our knowledge, no application
of ACM to the synthesis of natural products has been reported
to date.
As can be seen from Scheme 8, the recently introduced
catalyst system Mo[N(t-Bu)(Ar)]3 5/CH2Cl2 13 is very well suited
for this type of application as well. Specifically, reaction of
compound 12 bearing a butynyl entity as a handle for the
construction of an intact R-chain with a slight excess of the
symmetrical alkyne 33 in the presence of complex 5 and CH2-
Cl2 as the activating agent provides the desired cross metathesis
product 34 in 51% isolated yield. No trace of the homodimer
of 12 has been detected. The Lindlar reduction and subsequent
cleavage of the silyl protecting groups with aqueous HF in THF
proceed without incident, delivering PGE2-methyl ester 36 in
excellent yield. This compound is an established precursor for
the parent hydroxy acid 4.1,21
Acknowledgment. Generous financial support by the Deut-
sche Forschungsgemeinschaft (Leibniz award to A.F.) and the
Fonds der Chemischen Industrie is gratefully acknowledged.
K.G. thanks the Alexander-von-Humboldt foundation for a
fellowship.
To demonstrate the versatility of this approach even further,
we have prepared a small series of PG analogues differing from
the natural products in both side chains. Thus, reaction of n-BuLi
with Me2Zn and 1,4-addition of the resulting zincate38 to
cyclopentenone 9 followed by interception of the resulting
enolate with propargyl iodide 11 delivers product 37 in 73%
Supporting Information Available: Full Experimental
Section, compilation of the instrumentation used, details con-
cerning the X-ray structures of compounds 22 and 32, and copies
of the NMR spectra of all new compounds (PDF). This material
is available free of charge via the Internet at http://pubs.acs.org.
(38) For recent applications of zincates as Michael donors in total
synthesis see: (a) Fu¨rstner, A.; Weintritt, H. J. Am. Chem. Soc. 1998, 120,
2817. (b) Fu¨rstner, A.; Weintritt, H. J. Am. Chem. Soc. 1997, 119, 2944.
JA003119G