Synthesis of the potent antiglaucoma agent, travoprost

Article abstract of DOI:10.1021/op010097p

A commercial synthesis of the antiglaucoma agent, travoprost 2, is described. A total of 22 synthetic steps are required to provide the single enantiomer prostanoid, with the longest linear sequence being 16 steps from 3-hydroxybenzotrifluoride. The route is based upon a cuprate-mediated coupling of the single enantiomer vinyl iodide 13 and the tricyclic ketone 5, of high stereochemical purity, to yield the single isomer bicyclic ketone 15. A Baeyer - Villiger oxidation provides the lactone 16 as a crystalline solid, thus limiting the need for chromatographic purification. DIBAL-H reduction, Wittig reaction, esterification, and silyl group deprotection complete the synthesis of travoprost.

Full text of DOI:10.1021/op010097p

Organic Process Research & Development 2002, 6, 138−145
Synthesis of the Potent Antiglaucoma Agent, Travoprost
Lee T. Boulton, Dean Brick, Martin E. Fox, Mark Jackson, Ian C. Lennon,* Raymond McCague, Nicholas Parkin,
Darren Rhodes, and Graham Ruecroft
Chirotech Technology Limited, Cambridge Science Park, Milton Road, Cambridge, CB4 0WG, UK
Abstract:
A commercial synthesis of the antiglaucoma agent, travoprost
2, is described. A total of 22 synthetic steps are required to
provide the single enantiomer prostanoid, with the longest linear
sequence being 16 steps from 3-hydroxybenzotrifluoride. The
route is based upon a cuprate-mediated coupling of the single
Figure 1.
enantiomer vinyl iodide 13 and the tricyclic ketone 5, of high
stereochemical purity, to yield the single isomer bicyclic ketone
15. A Baeyer-Villiger oxidation provides the lactone 16 as a
crystalline solid, thus limiting the need for chromatographic
purification. DIBAL-H reduction, Wittig reaction, esterification,
and silyl group deprotection complete the synthesis of travo-
prost.
Figure 2.
Introduction
The prostaglandin analogue 16-[3-(trifluoromethyl)phen-
oxy]-17,18,19,20-tetranor PGF2R 1 and its ester derivatives,
in particular the isopropyl ester travoprost 2, are potent drugs
for the treatment of glaucoma and ocular hypertension
(Figure 1).1 Additionally, certain phenyl-substituted prosta-
glandin analogues are known to be potent and selective anti-
glaucoma agents.2 Recently, novel analogues in which a
cyclohexane unit3 has replaced the cyclopentane core and a
series containing a boronate4 in the R-side chain have been
synthesized. The racemic form of 1, fluprostenol, was devel-
oped by ICI for use as a veterinary contraceptive agent.5 The
ICI synthesis commenced from the Corey lactone aldehyde
3,6 with the R- and ω-side chains being built up by Horner-
Emmons/Wittig alkenylation reactions.
The biological profile and synthesis of 2 and its analogues
have been described by Klimko et al.1 Their synthetic strategy
was based on the Corey lactone route wherein the cyclo-
pentane ring embedded in a lactone intermediate of type 3
has relative stereochemistry correctly defined across four
chiral centres (Figure 2). This approach has two major
limitations (a) the reduction of the keto function in the ω-side
chain of intermediate 4 using (-)-B-chlorodiisopinocam-
pheylborane gave an insufficiently high diastereoisomeric
excess, requiring the removal of the unwanted 15-(S) isomer
by a difficult chromatographic separation, and (b) this route
requires the purchase or synthesis of multikilogram quantities
of the Corey lactone aldehyde 3. Hence, an alternative route
more amenable to the manufacture of kilogram quantities
of travoprost 2 was sought.
Elegant three-component coupling methodologies have
been developed for the synthesis of prostaglandins.7 An
alternative three-component, two-step variation has also been
devised by Johnson.8 Gooding describes the application of
a three-component coupling to unnatural prostaglandins with
(E)-CHdCHCH(OH)CH2OPh as the ω-side chain, intro-
duced by conjugate addition of a stannane-derived cuprate.9
Other novel routes to prostaglandin analogues based on
alkyne metathesis10 and nickel-catalysed cyclisation11 to form
the cyclopentane core have appeared recently. From our
experience in this area over the past six years we felt that
none of these methods were suitable for a scaleable,
commercial synthesis of travoprost 2. In particular, none of
these methods provided the rigorous control of all the
stereocentres that we sought to satisfy the regulatory authori-
ties. Therefore, an alternative synthetic strategy was required.
* Author for correspondence. E-mail: ilennon@dow.com.
(1) Klimko, P. G.; Bishop, J.; Desantis Jr., L.; Sallee, V. L. EP 0639563, 1995;
Chem. Abstr. 1995, 122, 290579.
(2) Resul, B.; Stjernschantz, J.; No, K.; Liljebris, C.; Sele´n, G.; Astin, M.;
Karlsson, M.; Bito, L. Z. J. Med. Chem. 1993, 36, 243.
(3) Klimko, P. G.; Davis, T. L.; Griffin, B. W.; Sharif, N. A. J. Med. Chem.
2000, 43, 3400.
(7) For some examples, see: Noyori, R.; Suzuki, M. Angew. Chem., Int. Ed.
Engl. 1984, 23, 847; Morita, Y.; Suzuki, M.; Noyori, R. J. Org. Chem.
1989, 54, 1785; Suzuki, M.; Morita, Y.; Koyano, H.; Koga, M.; Noyori, R.
Tetrahedron 1990, 46, 4809; Noyori, R.; Suzuki, M. Chemtracts: Org.
Chem. 1990, 3, 173.
(4) Feng, Z.; Hellberg, M. Tetrahedron Lett. 2000, 41, 5813.
(5) a) Mallion, K. B. U.S. Patent 3,935,240, 1976; Chem. Abstr. 1974, 81,
120097. (b) Bowler, J.; Crossley, N. S. U.S. Patent 4,321,275, 1982; Chem.
Abstr. 1977, 86, 43256.
(6) Corey, E. J.; Cheng, X.-M. The Logic of Chemical Synthesis; Wiley: New
York, 1989; pp 250-266 and references therein.
(8) Johnson, C. R.; Braun, M. P. J. Am. Chem. Soc. 1993, 115, 11014.
(9) Gooding, O. W.; Beard, C. C.; Cooper, G. F.; Jackson, D. Y. J. Org. Chem.
1993, 58, 3681.
(10) Fu¨rstner, A.; Grela, K.; Mathes, C.; Lehmann, C. W. J. Am. Chem. Soc.
2000, 122, 11799.
(11) Sato, Y.; Takimoto, M.; Mori, M. Chem. Pharm. Bull. 2000, 48, 1753.
138
•
Vol. 6, No. 2, 2002 / Organic Process Research & Development
10.1021/op010097p CCC: $22.00 © 2002 American Chemical Society
Published on Web 02/07/2002

Scheme 1
Figure 3.
Results and Discussion
We were attracted to a highly convergent approach to
prostaglandin analogues involving addition of a cuprate
reagent, incorporating the entire ω-side chain, to tricyclo-
[3.2.0.02,7]heptanone 5 developed by Roberts and Newton.12
This route required three synthons: the commercially avail-
able Wittig reagent (4-carboxybutyl)triphenylphosphonium
bromide, the silyl-protected bromohydrin 6 to provide the
cyclopentane core, and the vinyl iodide 13 to give access to
the ω-side chain (Figure 3). Thus, 6 and 13 had to be
prepared with high stereochemical purity.
The single enantiomer, bicycloheptenone (-)-7 (99% ee),
was liberated after treatment of the salt with aqueous sodium
carbonate solution. This was carried out prior to the
bromohydrin formation to minimise handling of the volatile
bicycloheptenone, which had also shown a tendency to
degrade at ambient temperature over short periods. Treatment
of (-)-7 with 1,3-dibromo-5,5-dimethylhydantoin in acetone/
water gave the crude bromohydrin with excellent diastereo-
isomeric and regioisomeric control (<10% minor isomers).
The rationale behind the highly selective bromohydrin
formation in this system has been fully detailed.16 The
bromohydrin contained small amounts of hydantoin impuri-
ties that could not be readily removed by crystallisation, and
consequently, the material was used crude. Silyl-protected
6 was readily recrystallised to provide analytically pure
product in 38% yield from the salt 8. The silyl-protected
bromohydrin 6 was converted into the tricycle 5 as required.
Avoidance of awkward late-stage reduction to establish
the required configuration of the C-15-OH functionality
provides the major advantage of this route. The stereochem-
istry at C-15 is controlled by using a single enantiomer side
chain derived from the acetylene 9. The acetylene was
derived from the racemic alcohol 1017 via a bioresolution
sequence (Scheme 2).18
The tricyclic ketone 5 was prepared from the TBDMS-
protected bromohydrin 6, which was in turn derived from
bicyclo[3.2.0]hept-2-en-6-one 7 (Scheme 1).13 The literature
synthesis of 7 was readily scaled up to kilogram levels,
except that we did not batch distill the intermediate
dichlorobicyclo[3.2.0]hept-2-en-6-one, or 7.14 DSC analysis
showed that the distillation of these materials could be
hazardous principally because of decomposition of the
residues. When heated at 10 °C/min in sealed aluminium
pans, the crude dichlorobicycloheptenone decomposed exo-
thermically from about 160 °C, the distilled material
displayed two endotherms (probably due to evaporation), and
the distillation residue decomposed exothermically from
about 90 °C. The DSC results also showed that the distillation
of 7 left an unstable residue, which degraded exothermically
upon heating. Both of these materials were therefore purified
by a short path wipe film distillation for any quantities above
50 g. The resolution of racemic bicyclo[3.2.0]hept-2-en-6-
one 7 by forming diastereomeric salts of its R-hydroxysul-
fonic acid derivative with (+)-R-methylbenzylamine 8 and
separation by crystallisation has been described.15 To develop
a synthesis capable of producing multihundred-gram to
kilogram quantities of travoprost 2 we needed to operate this
resolution at pilot-plant scale. The process developed at 5-L
scale in the laboratory was subsequently transferred to 200-L
vessels without any major process changes.
Enzymatic acylation of the racemic alcohol 10, using
Chirazyme L9, afforded the ester (R)-11 and the alcohol (S)-
10 (Scheme 2). This mixture could be separated, after which
basic hydrolysis of the butyrate ester (R)-11 gave the desired
(12) Lee, T. V.; Roberts, S. M.; Dimsdale, M. J.; Newton, R. F.; Rainey, D. K.;
Webb, C. F.,J. Chem. Soc., Perkin Trans. 1 1978, 1176.
(13) Grieco, P. A. J. Org. Chem. 1972, 37, 2363.
(16) Grudzinski, Z.; Roberts, S. M. J. Chem. Soc., Perkin Trans. 1 1975, 1767.
(17) Tolstikov, G. A.; Miftakhov, M. S.; Danilova, N. A.; Galin, F. Z. Z. J.
Org. Chem. USSR (Engl. Transl), 1983, 19, 1624-1631.
(14) The distillation of 7,7-dichlorobicylco[3.2.0]hept-2-en-6-one has been
described. Minns, R. A. Organic Syntheses; Wiley & Sons: New York,
Collect. Vol. 6, 1988, p 1037.
(18) Full details of this process are published elsewhere. (a) Fox, M. E.; Parratt,
J. S. U.S. Patent 6,214,611; Chem. Abstr. 2000, 133, 309067. (b) Fox, M.
E.; Jackson, M.; Lennon, I. C.; McCague, R.; Parratt, J. S. AdV. Synth. Catal.
2002, 344. In press.
(15) Wallis, C. J. EP 0074856, 1983; Chem. Abstr. 1984, 99, 139627.
Vol. 6, No. 2, 2002 / Organic Process Research & Development
•
139

Scheme 2
Scheme 3
alcohol (R)-10 in >98% ee. In this case the maximum yield
of 10 is 50%, with typical yields being 38-45%. To increase
the throughput of material from this resolution we developed
a mesylation and inversion process.18 Reaction of the mixture
of ester (R)-11 and alcohol (S)-10 with methanesulfonyl
chloride gave a mesylate 12/ester (R)-11 mixture. Treatment
of this mixture with butyric acid and triethylamine at 110
°C afforded the desired butyrate ester (R)-11 with ap-
proximately 90% ee. Enzymatic hydrolysis of the ester, using
Chirazyme L2, gave the alcohol (R)-10 with >99% ee,
containing some residual (S)-ester. The acetylene 9 was
obtained in 58-65% yields from the racemate 10 and 99.5%
ee after protection of the alcohol with TBDMSCl, basic
hydrolysis of the residual ester and hemiphthalate formation.
The final purification involved a base wash to remove the
hemiphthalate and a simple filtration through a pad of silica
gel.
The vinyl iodide 13 was prepared from the acetylene 9
via a hydrozirconation-iodination sequence (Scheme 3).19
In this procedure the reactive zirconium reagent, iBuZrCp2-
Cl,19 is formed in situ from readily available zirconocene
dichloride and tert-butylmagnesium chloride, thus avoiding
the need for the more expensive Schwartz’s reagent (Cp2-
Zr(H)Cl). Vinyl iodide 13 was metalated with tert-butyl-
lithium and then treated with lithium 2-thienylcyanocuprate.
In our hands the use of n-butyllithium for the metalation
failed to give satisfactory results. Fresh preparation of lithium
2-thienylcyanocuprate, by lithiation of thiophene with n-butyl-
lithium followed by treatment with copper(I) cyanide, gave
the most reproducible results. Reaction of alkenylcuprate 14
with tricycle 5 formed the bicyclic ketone 15. The tricycle
could be isolated as a low-melting, white solid on a small
scale (<10 g), but on larger scale extensive decomposition
of the tricycle was observed upon drying. Therefore, tricycle
5 was used without purification as a concentrated toluene
solution. The crude ketone 15 was semipurified by passing
through a silica gel pad to provide 15 in greater than 90%
chemical purity and 64% yield, and an analytically pure
sample was obtained by silica gel chromatography.
The most problematic process step was the Baeyer-
Villiger oxidation of the ketone 15. The use of peracetic acid
and sodium acetate in acetic acid resulted in a 3:1 mixture
of regioisomeric lactones, 16 and 17 isolated as an oil
(Scheme 4). This ratio is consistent with the Baeyer-Villiger
(19) Swanson, D. R.; Nguyen, T.; Noda, Y.; Negishi, E. J. Org. Chem. 1991,
56, 2590.
oxidation of similar systems reported in the literature.20
A
140
•
Vol. 6, No. 2, 2002 / Organic Process Research & Development

Scheme 4
Figure 4. ORTEP diagram of the molecular structure of
lactone 16. Hydrogen atoms are omitted for clarity.
Scheme 5
variety of oxidation systems were studied including m-
chloroperbenzoic acid and various metal-catalysed reac-
tions.21 In all cases the ratio of regioisomers was similar to
that outlined above. Extensive epoxidation of the ω-side
chain olefin was also observed with most of these reagents.
Using a tosyl group in place of the TBDMS group on the
oxabicyclooctanone ring and carrying out the Baeyer-
Villiger oxidation with peracetic acid, led to a 10:1 mixture
of regioisomers, in favour of the desired isomer. Unfortu-
nately, the tosylate intermediate was not suitable for the
synthesis of travoprost 2.
Conveniently, the minor and unwanted regioisomer 17
was selectively hydrolysed by treatment with aqueous sodium
hydroxide in acetonitrile (Scheme 4).22 The unreacted lactone
16 was separated from the hydroxy acid 18 by crystallisation
to provide the desired product in 38% yield. Purification of
the mother liquors by passing through a silica gel pad and
crystallisation, led to a further crop. An overall yield of 47%
of analytically pure, crystalline lactone 16 was obtained.23
Another key advantage of our route to travoprost 2 is the
crystallinity of lactone 16. This allows a high-purity, late-
stage intermediate to be isolated without the need for careful
chromatography. As the lactone was crystalline, we con-
firmed its structure and relative configuration by X-ray
crystallography (Figure 4).24
19 was achieved using diisobutylaluminium hydride at
-70 °C. Wittig reaction with the ylide generated from (4-
carboxybutyl)triphenylphosphonium bromide and potassium
tert-butoxide followed by esterification using DBU and
2-iodopropane afforded the bis-silyl-protected prostanoids
20a and 20b. Carrying out the Wittig reaction at 0 °C limited
the amount of trans isomer to 3% (whereas reaction at room
temperature typically gave 5% trans isomer). The mixture
of silyl compounds was also observed due to migration of
the protecting group on the cyclopentane ring.22,25 Depro-
tection using aqueous hydrochloric acid in 2-propanol
(24) Crystal data for C30H47F3O5Si2: colourless, monoclinic, P21, a ) 6.803(10)
Å, b ) 32.53(4) Å, c ) 15.27(4) Å, â ) 90.24(10)°, V ) 3380(10) Å3,
Z ) 4, dcalc ) 1.181 g/cm3, F(000) ) 1288, µ (Mo KR) ) 0.155 mm-1
,
θ range 1.33-20.81° (Mo KR), 123831 measured reflections on a Rigaku
RAXIS diffractometer, 6797 independent reflections [R(int) ) 0.052],
the structure determination using direct methods (SHELXS86), R(F) )
0.0625, Rw ) 0.1703 for 5791 reflections with I > 2σ(I), GOF )
1.070. Crystallographic data for structure (16) reported in this paper
have been deposited with the Cambridge Crystallographic Data Centre
as deposition No. CCDC-170249. Copies of the data can be obtained
on application to the CCDC, 12 Union Road, Cambridge CB2 1EZ, UK
[e-mail: deposit@ccdc.cam.ac.uk, http://www.ccdc.cam.ac.uk].
Lactone 16 was converted to the target prostanoid 2 using
conventional processes (Scheme 5). Reduction to the lactol
(20) Coleman, M. J.; Crookes, D. L.; Hill, M. L.; Singh, H.; Marshall, D. R.;
Wallis, C. J. Org. Process Res. DeV. 1997, 1, 20.
(21) Gottlich, R.; Yamakoshi, K.; Sasai, H.; Shibasaki, M. Synlett 1997, 971.
(22) Newton, R. F.; Roberts, S. M. Tetrahedron 1980, 36, 2163.
(23) Jackson, P. M.; Lennon, I. C. U.S. Patent 6,294,679; Chem. Abstr. 2000,
133, 309792.
(25) For a related example of silyl migration see: Torisawa, Y.; Shibasaki, M.;
Ikegami, S. Chem Pharm. Bull. 1983, 31, 2607.
Vol. 6, No. 2, 2002 / Organic Process Research & Development
•
141

completed the synthesis of travoprost 2. This material was
isolated with good chemical purity, but was not of high
enough purity for a bulk active. A simple purification to
remove minor diastereoisomers and minor impurities was
developed using a Biotage Prep 75-L chromatography unit.
This provided travoprost 2 with excellent chemical purity,
suitable for toxicology and clinical trials. This route has been
developed further, with only minor modifications, to provide
commercial quantities of travoprost 2.
To conclude, we have demonstrated that the tricycle
rearrangement route12 to prostanoids can be developed to
provide a commercially viable route to the antiglaucoma
agent, travoprost 2, with rigorous control of all five stereo-
centres. The longest linear sequence was 16 steps from
3-hydroxybenzotrifluoride, and yields in the range of 4.0-
6.9% were achieved. In total, 22 synthetic steps were
required, and a total of 450 g of travoprost was synthesised
for clinical trials. We are now using this route for the
manufacture of travoprost 2 to provide low-kilogram quanti-
ties of the active pharmaceutical ingredient.
0-5 °C and held at this temperature overnight (17 h) and
then collected by filtration. The product was washed with
2-propanol (1.8 kg). The wet cake [7.00 kg, estimated dry
weight 4.62 kg, 24.7%; liberated (-)-bicycloheptenone, 89%
ee] was charged to the vessel with 2-propanol (11.6 kg) and
heated to 67 ( 5 °C to give a solution. The mixture was
cooled to 0-5 °C and left to stir overnight. The product was
collected by filtration and washed with 2-propanol (1.6 kg).
The 2-propanol wet cake [5.45 kg, estimated dry weight 4.34
kg, 23.2%; liberated (-)-bicycloheptenone, 97% ee] was
charged to the vessel with 2-propanol (10.9 kg) and heated
to 65 ( 5 °C, at which point the solids dissolved. The
solution was cooled to 0-5 °C and stirred overnight. The
resulting slurry was collected by filtration. The wet cake was
washed with 2-propanol (1.2 kg) and collected. The wet cake
was dried to constant weight to give the final product 8 (4.25
kg, 22.7%), mp 127 °C (lit.15 86-88 °C); [R]D23 -14.8 (c 5,
water).
1H NMR (200 MHz, d6-DMSO) δ 7.55-7.35 (5H, m,
ArH), 5.72 (2H, m), 4.40 (1H, q, J ) 8.0), 3.25 (1H, m),
2.92-2.62 (3H, m), 2.25 (1H, dd, J ) 16.0, 10.0), 1.51 (1H,
m), and 1.48 (3H, d, J ) 8.0); Liberated (-)-bicyclo-
heptenone 99% ee, [GC, Chirasil Dex CB, 90 °C, hold for
10 min, ramp to 200 °C at 30 °C/min, retention time: 5.52
min (-); 6.09 min (+)].
1(R)-2-exo-Bromo-3-endo-tert-butyldimethylsilyloxy-
bicyclo[3.2.0]heptan-6-one (6). Demineralized water (54
kg), Na2CO3 (10.26 kg, 96.8 mol), and (-)-6-hydroxybicyclo-
[3.2.0]hept-2-en-6-sulfonic acid (+)-R-methylbenzylamine
(1:1) salt 8 (9.00 kg, 28.9 mol) were charged to a 200-L
nitrogen-purged vessel. The mixture was agitated. CH2Cl2
(12.0 kg) was charged to the reactor, and the mixture was
stirred for 15-20 min. The lower CH2Cl2 layer was
separated, and the aqueous phase was extracted again with
CH2Cl2 (12.0 kg). The combined CH2Cl2 extracts were
charged to the vessel, and the vessel contents were adjusted
to pH 1 by the controlled addition of hydrochloric acid (7.2%,
16.2 kg) over a 70-min period, maintaining the temperature
in the range 8-16 °C. The CH2Cl2 layer was separated and
washed with demineralised water (6 × 9.0 kg) and then with
brine (21 kg). The CH2Cl2 solution was dried (MgSO4) and
filtered, and the solids were washed with CH2Cl2 (4.2 kg).
The resulting solution was collected and concentrated by
rotary evaporation to give an oil (3.1 kg, quantitative).
The oil (3.1 kg) was dissolved in acetone (10.7 kg) and
charged to a vessel, and demineralised water (2.6 kg) was
added. 1,3-Dibromo-5,5-dimethylhydantoin (4.14 kg, 14.5
mol) was charged to the reactor over 60-80 min, maintaining
the temperature below 10 °C. The temperature was adjusted
to 15-20 °C over 2 h, and the reaction mixture was held at
this temperature overnight. The reaction mixture was quenched
with 10% aqueous Na2S2O5 solution (13.5 kg), maintaining
the temperature of the mixture below 30 °C. The mixture
was extracted with CH2Cl2 (3 × 12.0 kg). The CH2Cl2
extracts were washed with brine (22 kg) and dried (MgSO4,
1.6 kg). The mixture was filtered, washed with CH2Cl2 (24
kg), and concentrated on a rotary evaporator to give the crude
bromohydrin (4.75 kg, ∼88% purity ) 4.18 kg actual, 62%),
Experimental Section
General Procedures. Melting points were determined on
an Electrothermal capillary apparatus and are uncorrected.
1H NMR spectra were recorded at 200 MHz (Bruker AM200)
or 400 MHz (Bruker DPX 400). 13C NMR spectra were
recorded at 100 MHz. Chemical shifts (δ) are quoted in ppm,
and coupling constants (J) are given in Hz. Optical rotations
were determined using a Perkin-Elmer 341 polarimeter and
are given in 10-1 deg cm2 g-1. IR spectra were recorded
using a Perkin-Elmer 1600 series FTIR. Mass spectra were
recorded using a Navigator LC/MS system, or a Hewlett-
Packard GC/MS. Analytical thin-layer chromatography was
performed on Merck silica gel precoated plates and visualised
using ceric ammonium molybdate or potassium permanga-
nate solution.
(-)-6-Hydroxybicyclo[3.2.0]hept-2-en-6-sulfonic Acid
(+)-r-Methylbenzylamine (1:1) Salt (8). (+)-R-Methyl-
benzylamine (7.30 kg, 60.2 mol), CH2Cl2 (86.3 kg) and
demineralised water (1.2 kg) were charged to a 200-L
nitrogen-purged vessel. The mixture was agitated, and sulfur
dioxide gas (17.4 kg, 272 mol) was charged to the vessel
with stirring in the temperature range 18-23 °C. (()-
Bicyclo[3.2.0]hept-2-en-6-one 7 (6.50 kg, 60.1 mol) was
charged to the vessel, followed by CH2Cl2 (5.2 kg),
maintaining the temperature of the vessel contents below 25
°C during this addition. The reaction mixture was stirred at
18-23 °C until crystallisation was observed. The contents
of the vessel were cooled to between 0 and 5 °C, and the
mixture was stirred overnight within this temperature range.
The resulting solid was collected by filtration and washed
with CH2Cl2 (5.2 kg). The solid was washed once more with
CH2Cl2 (5.2 kg). The damp cake [9.75 kg, estimated dry
weight 8.1 kg, 43.3%; liberated (-)-bicycloheptenone was
75% ee] was charged back to the vessel, and 2-propanol (20.3
kg) was added. The mixture was heated to 67 °C. The hazy
solution was circulated through an in-line 5 µm filter to give
a clear solution. The circulation lines were flushed clear of
product using 2-propanol (1.9 kg). The batch was cooled to
142
•
Vol. 6, No. 2, 2002 / Organic Process Research & Development

which was then charged to a vessel with DMF (11.0 kg)
and imidazole (2.4 kg, 35.2 mol). The mixture was stirred
and cooled to 0-5 °C. tert-Butyldimethylsilyl chloride (3.5
kg, 23.4 mol) was charged in portions, maintaining the
temperature below 5 °C. The reaction mixture was stirred
overnight at 20 ( 2 °C. A further portion of tert-butyldi-
methylsilyl chloride (200 g, 1.34 mol) was charged, and the
mixture was stirred for 2 h. The reaction mixture was
quenched with demineralised water (36 kg) and extracted
with heptane (12.3 kg, then 6.2 kg). The combined heptane
extracts were washed with water (9.0 kg) and brine (9.0 kg).
The heptane solution was dried (MgSO4, 1.8 kg), and the
slurry was filtered through a silica gel pad (0.84 kg). The
resulting heptane liquors were concentrated under reduced
pressure to give an oil (4.1 kg). The oil was charged to a
vessel along with heptane (12.3 kg) and stirred to form a
solution. The resulting solution was cooled to -27 ( 2 °C
over 1 h and left to stir at -27 ( 2 °C for 1 h. The crystalline
product was collected by filtration. Heptane (0.7 kg, pre-
cooled to -20 ( 2 °C) was charged to the vessel and then
transferred to the filter cake. The solid was dried to constant
weight to give 6 (3.53 kg, over two crops, 38%), retention
time 26.3 min (GC, 25QC2/BP5, 60 °C for 5 min, ramp to
300 °C at 10 °C/min, hold for 25 min), [R]2D3 -0.7 (c 1.0,
CH2Cl2); mp 58-62 °C, 1H NMR (200 MHz, CDCl3) δ 4.51
(1H, d, J ) 3.5), 4.21 (1H, s), 3.78 (1H, m), 3.33-3.13 (3H,
m), 2.48 (1H, m), 2.20 (1H, d, J ) 14), 0.85 (9H, s), 0.08
(3H, s), and 0.06 (3H, s).
3(S)-endo-tert-Butyldimethylsilyloxytricyclo[3.2.0.02,7]-
heptan-6-one (5). KOtBu (114.2 g, 1.02 mol) was suspended
in toluene (2 L) and cooled to -15 °C. The suspension was
stirred under a nitrogen atmosphere, and a solution of the
bromoketone 6 (250 g, 0.783 mol) in dry toluene (400 mL)
was added over 1 h. The internal temperature was maintained
at -10 to -20 °C. The mixture was stirred for 1 h before
being warmed to room temperature, activated carbon (75 g)
was added, and the mixture was stirred for 5 min. The
mixture was filtered through Celite, and the cake was washed
with toluene (2.5 L). The filtrates were concentrated under
reduced pressure, at 20 °C, to an approximately 700-mL
volume. This solution was used directly in the next step.
(E)-1-Iodo-4-(m-trifluoromethylphenoxy)-3(R)-tert-butyl-
dimethylsilyloxy-1-butene (13). A dry 5-L three-necked
flask was purged with nitrogen, and bis(cyclopentadienyl)-
zirconium dichloride (459 g, 1.57 mol) and toluene (2 L)
were added. The vessel was covered with aluminium foil to
exclude light, evacuated, and purged with nitrogen. tert-
Butylmagnesium chloride (2 M in ether, 785 mL, 1.57 mol)
was added over 30 min (an exotherm from 20 to 28 °C was
observed). The mixture was heated at 50 °C for 1 h (during
which time isobutylene evolution was observed). Alkyne 918
(450 g, 1.31 mol) in toluene (500 mL) was added, and
heating was continued at 50 °C for 5 h. The reaction mixture
was cooled to -40 °C (dry ice/acetone), and iodine (497 g,
1.96 mol) in THF (600 mL) was added over 35 min. The
mixture was warmed to room temperature over 1 h. Aqueous
Na2S2O5 solution (1 M, 2 L) and heptane (3 L) were added
(a bright-yellow dense precipitate was formed). The mixture
was filtered through a No. 3 filter paper, and the filter cake
was washed with heptane (1 L). The organic layer was
separated, and the aqueous phase was extracted with heptane
(1 L). The combined organic phases were washed with
aqueous Na2S2O5 solution (1 M, 3 L), saturated aqueous
NaHCO3 solution (2 L) (Note: effervescence; use care when
adding bicarbonate solution), and brine (2 L). The organic
phase was dried (MgSO4), filtered, and concentrated under
reduced pressure to afford a brown oil. The residue was
passed through a pad of activated aluminium oxide (Neutral
Brockmann 1, 150 mesh, 750 g), eluting with heptane (6
L). The solvent was concentrated under reduced pressure,
the residue was dissolved in heptane (1 L) and filtered
through a No. 3 filter paper. The solvent was evaporated to
provide the iodide 13 as a dark brown liquid (441 g, 72%),
[R]2D3 -15.5 (c 0.96, CH2Cl2); IR νmax (film) 1609, 1592,
1
1329, and 1130 cm-1; H NMR (200 MHz, CDCl3) δ 7.40
(1H, t, J ) 8), 7.24 (1H, m), 7.11-7.04 (2H, m), 6.68 (1H,
dd, J ) 14, 5), 6.50 (1H, dd, J ) 14, 1), 4.51 (1H, m), 3.91
(2H, d, J ) 6), 0.92 (9H, s), and 0.11 (6H, s); m/z (GC/MS,
t
EI) 415 (M - Bu, 15), 288 (45), and 219 (100).
7-anti[4-(m-Trifluoromethylphenoxy)-3(R)-tert-butyldi-
methylsilyloxy-1(E)-butenyl]-5(S)-endo-tert-butyldimeth-
ylsilyloxybicyclo[2.2.1]heptan-2-one (15). A dry 10-L
flange flask, fitted with an overhead stirrer, temperature
probe, nitrogen inlet, and pressure-equalised dropping funnel,
was purged with nitrogen and cooled to -70 °C. A solution
of tert-butyllithium (1.7 M in pentane, 1013 mL, 1.72 mol)
was added. The solution was recooled to -70 °C, and a
solution of the vinyl iodide 13 (436 g, 0.923 mol) in diethyl
ether (1.3 L) was added over 90 min, maintaining the internal
temperature below -60 °C.
In the meantime, thiophene (75.2 mL, 0.94 mol) was
placed in a dry 2-L three-necked flask, under a nitrogen
atmosphere. THF (600 mL) was added, and the solution was
cooled to -30 °C. n-Butyllithium (2.5 M in hexanes, 376
mL, 0.94 mol) was added over 20 min. The solution was
stirred for 20 min at -20 °C, and then the resulting yellow
solution was added to a suspension of copper(I) cyanide
(84.15 g, 0.94 mol) in THF (800 mL) at -20 °C over 15
min. The resulting dark-brown solution was recooled to -10
°C and stirred for 20 min.
The freshly prepared lithium 2-thienylcyanocuprate solu-
tion, at -10 °C, was added to the vinyllithium solution at
-70 °C over 20-30 min. The resulting solution was stirred
for 30 min at -70 °C. The tricycle solution 5 (approximately
187 g in 600 mL toluene, + 100 mL THF added) was cooled
to -70 °C and added to the cuprate solution 14 at -70 °C
over 20 min. The mixture was stirred at -70 °C for 1 h, the
cooling bath was removed from the reaction vessel, and
saturated aqueous NH4Cl solution (3 L) was added. The
mixture was allowed to warm to room temperature with
stirring (the aqueous layer became deep blue in colour, and
a yellow/green precipitate formed). The mixture was filtered
through a No. 3 filter paper, and the filter cake was washed
with MTBE (1 L). The organic layer was separated, and the
aqueous layer was extracted with MTBE (1 L). The
combined organic layers were washed with brine (2 L), dried
Vol. 6, No. 2, 2002 / Organic Process Research & Development
•
143

(MgSO4), and decolourised with activated carbon. After 20
min the solution was filtered, the cake was washed with
MTBE (2.5 L), and the filtrate was evaporated under reduced
pressure.
1733 cm-1; 1H NMR (400 MHz, CDCl3) δ 7.40 (1H, t, J )
8), 7.22 (1H, d, J ) 8), 7.10 (1H, s), 7.05 (1H, d, J ) 8),
5.70 (2H, m), 4.52 (3H, m), 3.87 (2H, d, J ) 6), 3.16 (1 H,
d, J ) 18), 3.00 (1H, d, J ) 6), 2.56 (1H, dd, J ) 18, 6),
2.46 (1H, m), 2.39 (1H, m), 1.88 (1H, dt, J ) 16, 3), 0.95
(9H, s), 0.89 (9H, s), 0.10 (6H, s), and 0.04 (6H, s); 13C
NMR (100 MHz, CDCl3) δ (170.39, 158.75, 132.43, 131.92
(q, J ) 32), 130.05, 127.87, 123.91 (q, J ) 270), 118.06,
117.68, 111.11, 82.20, 72.26, 71.64, 71.10, 48.94, 42.48,
40.40, 33.05, 25.75, 18.30, 18.04, -4.64, -4.76, -4.88 and
The residue was taken up in heptane and passed through
a plug of silica (1.5 kg), eluting with 2% EtOAc/heptane to
10% EtOAc/heptane to provide the pure ketone 15 as a
yellow solid (293 g, 64%), mp 64-72 °C; [R]2D0 +35.9 (c
1.05, CH2Cl2); (Found: C, 61.50; H, 8.04. C30H47F3O4Si2
1
requires C, 61.61; H, 8.10); IR νmax (Nujol) 1738 cm-1; H
t
-5.10; m/z (GC/MS, EI) 543 (M - Bu, 7), 219 (43), and
73 (100).
NMR (400 MHz, CDCl3) δ 7.38 (1H, t, J ) 8), 7.25 (1H, d,
J ) 8), 7.13 (1H, s), 7.07 (1H, d, J ) 8), 5.86 (1H, dd, J )
16, 8), 5.73 (1H, dd, J ) 16, 6), 4.55 (2H, m), 3.90 (2H, d,
J ) 7), 2.80 (1H, m), 2.77 (1H, d, J ) 18), 2.57 (2H, m),
2.45 (1H, m), 2.05 (1H, dd, J ) 18, 4), 1.35 (1H, m), 0.95
(9H, s), 0.90 (9H, s), 0.15 (3H, s), 0.14 (3H, s), and 0.05
(6H, s); 13C NMR (100 MHz, CDCl3) δ 216.0, 158.8, 132.0,
131.5 (q, J ) 32), 129.9, 127.8, 123.9 (q, J ) 270), 118.0,
117.5, 111.1, 72.3, 71.3, 69.9, 54.4, 50.2, 46.2, 38.8, 33.4,
25.8, 18.3, 18.0, -4.7, -4.7, -4.9, and -4.9; m/z (GC/MS,
8-anti[4-(m-Trifluoromethylphenoxy)-3(R)-tert-butyldi-
methylsilyl-1(E)-butenyl]-6(S)-endo-tert-butyldimethyl-
silyloxy-2-oxabicyclo[3.2.1]octan-3-ol (19). The lactone 16
(180 g, 300 mmol) was dissolved in dry toluene (2 L), and
the solution was cooled to -70 °C under nitrogen. DIBAL-H
(1.5 M in toluene, 300 mL) was added dropwise over 1.5 h,
and the solution was stirred for 3 h. The reaction was
quenched with methanol (100 mL) at -70 °C. Dilute aqueous
H2SO4 (2 N, 500 mL) was added, maintaining the temper-
ature below -30 °C. The cold bath was removed, and after
warming to ambient temperature, the mixture was extracted
with MTBE (3 × 500 mL). The organic phases were washed
with aqueous H2SO4 (2 N, 500 mL), water (500 mL), and
brine (500 mL) and dried (MgSO4). The mixture was filtered,
and solvent was evaporated to give the lactol 19 (181 g,
100%) as a colourless oil, IR νmax (film) 3428 and1724 cm-1;
1H NMR (200 MHz, CDCl3) δ 9.78 (1H, s), 7.40 (1H, t, J
) 8), 7.20-7.05 (3H, m), 5.65 (2H, m), 4.50 (1H, m), 4.30
(1H, m), 3.95-3.86 (1H, m), 3.88 (2H, d, J ) 6.5), 2.80
(1H, dd, J ) 19, 10), 2.45-2.00 (4H, m), 1.78 (1H, m),
1.60 (1H, br), 0.90 (18H, s), 0.06 (6H, s), 0.04 (3H, s), and
0.00 (3H, s).
(5Z,13E)-(9S,11R,15R)-11-Hydroxy-9,15-bis-(tert-butyldi-
methylsilyloxy)-16-(m-trifluoromethylphenoxy)-17,18,19,20-
tetranor-5,13-prostadienoic Acid, Isopropyl Ester (20a)
and 9-Hydroxy-11,15-bis-(tert-butyldimethylsilyloxy)-
compound (20b). KOtBu (1 M in THF, 1.8 L) was added
to a stirred solution of (4-carboxybutyl)triphenylphosphonium
bromide (402 g, 0.907 mol) in dry THF (3 L) under nitrogen
whilst maintaining the temperature below 10 °C. After 30
min the mixture was cooled to 0 °C, and a solution of the
lactol 19 (181 g, 0.30 mol) in dry THF (2.5 L) was added
over 30-45 min, while maintaining the temperature below
2 °C. The mixture was stirred for 1.5 h. TLC (EtOAc/pentane
1:4) showed complete reaction of the lactol. The reaction
was quenched with saturated NH4Cl (1 L), and the mixture
was extracted with EtOAc (1 L, then 2 × 500 mL). The
organic phases were washed with water (1 L) and brine (1
L) and dried (Na2SO4, 250 g). The solvent was evaporated,
and the crude prostadienoic acid was dissolved in acetone
(2 L). DBU (280 g, 1.88 mol) was added. After 5 min,
isopropyl iodide (306 g, 1.88 mol) was added, and the
solution was stirred at room temperature for 20 h. TLC
(EtOAc/pentane 1:4) showed complete consumption of the
acid. Acetone was removed by rotary evaporation, and the
residue was dissolved in EtOAc (3.5 L). The organic solution
t
EI) 527 (M - Bu, 6), 409 (11), 219 (19), and 73 (100).
8-anti[4-(m-Trifluoromethylphenoxy)-3(R)-tert-butyldi-
methylsilyl-1(E)-butenyl]-6(S)-endo-tert-butyldimethyl-
silyloxy-2-oxabicyclo[3.2.1]octan-3-one (16). Ketone 15
(362.6 g, 0.62 mol) and NaOAc (170 g, 2.07 mol) were
dissolved in glacial acetic acid (1.7 L). The reaction vessel
was placed in a water bath at 20 °C, and peracetic acid (40%
in dilute acetic acid, 176.7 mL, 0.93 mol) was added over a
period of 20 min. The solution was stirred at room temper-
ature for 3 h. More peracetic acid (30 mL) was added, and
the solution was stirred for a further 2 h. The reaction mixture
was poured onto water (2.5 L), and the products were
extracted into MTBE (2 × 750 mL, and 500 mL). The
combined organic extracts were washed with water (2 L).
The aqueous phase was back-extracted with MTBE (500
mL). The combined organic extracts were neutralised with
saturated Na2CO3 solution (500 mL), and water (2 L) was
added to aid phase separation. The organic phase was washed
with water (1 L) and brine (1 L), dried (MgSO4), and
evaporated under reduced pressure to give a yellow oil (363.8
g). The crude product (consisting of a 3:1 mixture of
regioisomers) was dissolved in acetonitrile (1 L) at room
temperature. Aqueous NaOH solution (1 M, 300 mL) was
added, and the solution was stirred at room temperature for
2 h. Water (1 L) was added, and the product was extracted
into MTBE (3 × 500 mL). The combined organic phases
were washed with brine (500 mL), dried (MgSO4), filtered,
and evaporated under reduced pressure. The residue was
recrystallised from heptane (1 L) at -70 °C (cold bath
temperature), filtered, and washed with cold ( -70 °C, cold
bath) heptane (2 × 200 mL). The solid was dried to give
the lactone 16 as a white solid (141.2 g, 38%). The mother
liquors were filtered through silica gel (1 kg), eluting with
20% CH2Cl2/heptane to remove baseline material. Recrys-
tallisation from cold ( -70 °C, cold bath) heptane (400 mL)
gave a second batch of lactone (34 g, 9%), mp 77-78 °C;
[R]2D0 -10.9 (c 1.05, CH2Cl2); (Found: C, 59.91; H, 7.90.
C30H47F3O5Si2 requires C, 59.97; H, 7.88); IR νmax (Nujol)
144
•
Vol. 6, No. 2, 2002 / Organic Process Research & Development

was washed with saturated KH2PO4 (3.5 L). The aqueous
layer was washed with EtOAc (1 L then 0.5 L). The
combined organic phases were washed with brine (800 mL),
dried with (Na2SO4, 210 g), and filtered. The solvent was
removed under reduced pressure, and the residue was
dissolved in 1:6 EtOAc/heptane (500 mL). The solution was
passed down a flash column of silica (1.44 kg) in a sintered
funnel. The product was eluted with 1:6 EtOAc/heptane (13
L). The solvent was evaporated to give the title compounds
(20a+b) (209 g, 95%) as a colourless oil (Found: C, 62.76;
H, 8.69. C38H63F3O6Si2 requires C, 62.60; H, 8.71); IR νmax
(film) 3521 and 1731 cm-1; 1H NMR (400 MHz, CDCl3) δ
7.37 (1H, t, J ) 8), 7.18 (1H, d, J ) 8), 7.10 (1H, s), 7.05
(1H, d, J ) 8), 5.60 (2H, m), 5.40 (2H, m), 5.00 (1H, heptet,
J ) 6), 4.52 (1H, q, J ) 6), 4.25-4.05 (2H, m), 3.90 (2H,
m), 2.55 (1H, d, J ) 10), 2.40-1.95 (7H, m), 1.80 (1H, d,
J ) 15), 1.65 (3H, m), 1.50 (1H, m), 1.20 (6H, d, J ) 6),
0.90 (9H, s), 0.85 (9H, s), 0.10 (9H, s), and 0.05 (3H, s);
m/z (CI) 729 (MH+, 20), 597 (31), 435 (65), 303 (100), and
(84).
with brine (1 L), dried (Na2SO4), filtered, and evaporated.
The residue was dissolved in EtOAc (200 mL). The solution
(250 g) was chromatographed (100% EtOAc) in four portions
on the Biotage 75 medium-pressure silica chromatography
system. All relevant fractions were combined and concen-
trated to give the title compound 2 (97 g, 70%) as a colourless
oil, [R]2D0 +14.6 (c 1.0, CH2Cl2); IR νmax (film) 3374 and
1727 cm-1; 1H NMR (400 MHz, CDCl3) δ 7.39 (1H, t, J )
8), 7.22 (1H, d, J ) 8), 7.15 (1H, s), 7.08 (1H, d, J ) 8),
5.70 (2H, m), 5.40 (2H, m), 4.98 (1H, heptet, J ) 6.5), 4.52
(1H, m), 4.18 (1H, m), 3.97 (3H, m), 3.25 (2H, br s), 2.60
(1H, br s), 2.38 (1H, m), 2.30-1.96 (7H, m), 1.76 (1H, dd,
J ) 16, 4), 1.65 (2H, quintet, J ) 7), 1.55 (1H, m), and
1.20 (6H, d, J ) 6); 13C NMR (100 MHz, CDCl3) δ 173.57,
158.67, 135.45, 131.87 (q, J ) 32), 130.02, 129.85, 129.75,
128.93, 123.89 (q, J ) 270), 118.06, 117.82, 111.48, 77.77,
72.70, 71.99, 70.86, 67.72, 55.82, 50.24, 42.84, 34.00, 26.60,
25.48, 24.83, and 21.81; m/z (CI) 501 (MH+, 21), 321 (34),
303 (44), and 249 (100).
(5Z,13E)-(9S,11R,15R)-9,11,15-Trihydroxy-16-(m-trifluoro-
methylphenoxy-17,18,19,20-tetranor-5,13-prostadienoic
Acid, Isopropyl Ester (2). The silyl-protected compound
(20a+b) (202 g, 277 mmol) was dissolved in 2-propanol
(2.2 L). Aqueous HCl (2 N, 750 mL) was added over
30 min. The solution was stirred at room temperature for
2 h. TLC (100% EtOAc) showed complete deprotection
had been achieved. Saturated aqueous NaHCO3 solution
(2 L) was added, and the mixture was extracted with EtOAc
(1 L, then 2 × 0.4 L). The combined extracts were washed
Acknowledgment
We thank Dr. Madeleine Helliwell, Department of Chem-
istry, University of Manchester for determination of the X-ray
crystal structure. We are grateful to Catherine Rippe´ and Sean
Savage for the development of analytical methods, and Dr.
John O’Rourke for DSC analysis.
Received for review October 15, 2001.
OP010097P
Vol. 6, No. 2, 2002 / Organic Process Research & Development
•
145