A novel convergent synthesis of the antiglaucoma PGF2α analogue bimatoprost

Article abstract of DOI:10.1002/chir.22123

The 17-phenyl PGF_2α analogue bimatoprost (10a) is the most efficacious ocular hypotensive agent currently available for the treatment of glaucoma or ocular hypertension. A novel convergent synthesis of 13,14-en-15-ol prostamideF_2α an

Full text of DOI:10.1002/chir.22123

CHIRALITY 25:170–179 (2013)
A Novel Convergent Synthesis of the Antiglaucoma PGF2a
Analogue Bimatoprost
1
1
1
1
2
2,3
*
IWONA DAMS, MICHAŁ CHODYŃSKI, MAŁGORZATA KRUPA, ANITA PIETRASZEK, MARTA ZEZULA, PIOTR CMOCH,
MONIKA KOSIŃSKA,2 AND ANDRZEJ KUTNER1
1R&D Chemistry Department, Pharmaceutical Research Institute, Rydygiera 8, 01-793 Warsaw, Poland
2R&D Analytical Chemistry Department, Pharmaceutical Research Institute, Rydygiera 8, 01-793 Warsaw, Poland
3Institute of Organic Chemistry, Polish Academy of Sciences, Kasprzaka 44/52, 01-224 Warsaw, Poland
ABSTRACT
The 17-phenyl PGF2a analogue bimatoprost (10a) is the most efficacious ocular
hypotensive agent currently available for the treatment of glaucoma or ocular hypertension. A
novel convergent synthesis of 13,14-en-15-ol prostamideF2a analogues was developed employing
Julia–Lythgoe olefination of the structurally advanced phenylsulfone (+)-(5Z)-15 with an enantio-
merically pure aldehyde o-chain synthon (–)-(S)-16a. Subsequent hydrolysis of protecting
groups and final amidation of the diol 26a yielded bimatoprost (10a). The main advantage of
the current strategy is the preparation of high-purity bimatoprost (10a). The novel convergent
strategy allows the synthesis of a whole series of 13,14-en-15-ol prostamideF2a analogues with
the desired C-15 asymmetric center configuration from a common and structurally advanced
prostaglandin intermediate (+)-(5Z)-15. The preparation and identification of two synthetic impu-
rities, 15-epi isomer (10b) of bimatoprost and a new prostaglandin related amide (+)-(5Z)-18, are
also described. Chirality 25:170–179, 2013. © 2013 Wiley Periodicals, Inc.
KEY WORDS: bimatoprost; prostaglandins; prostamides; Corey lactone; nucleophilic addition
INTRODUCTION
side chain (o-chain) via a Horner–Wadsworth–Emmons
(HWE) condensation of the Corey aldehyde with a suitable
ketophosphonate. The HWE reaction is generally affected
by some drawbacks, such as easy epimerization of labile
stereogenic centers12 and, depending on the base used for
deprotonation, formation of additional byproducts.13 Subse-
quent reduction of the o-chain 13,14-en-15-one to a 13,14-en-
15-ol, addition of the a-chain by Wittig olefination leading to
a cis-5,6-alkene, and some further simple transformations com-
prise a sequence of reactions needed for the Corey synthesis
of trans-13,14-en-15-ol PGF2a analogues such as travoprost (8)
and bimatoprost (10a).
The tedious nature of prostaglandin purification and separa-
tion of diastereoisomers necessitates a highly diastereoselec-
tive industrial preparation of PGF2a analogues used as active
pharmaceutical ingredients. The multiple chiral centers pres-
ent in prostaglandin molecules continue to create practical
challenges, and existing synthetic routes to PGF2a analogues
by the Corey strategy, in spite of considerable progress made
in this field, are not completely free of contaminating diaster-
eoisomers. Besides the a-chain 5,6-trans diastereoisomer
typically formed from the Wittig reaction, particular difficulty
The pharmacological management of glaucoma and ocular
hypertension has significantly changed over the last 18 years
with the introduction of new prostaglandin analogues/prosta-
mides (Fig. 1).1 Unoprostone isopropyl ester (4, ResculaW, Ciba
Vision, Duluth, GA, USA), latanoprost (6, XalatanW, Pfizer, New
York, NY, USA), travoprost (8, TravatanW, Alcon, Fort Worth,
TX, USA), bimatoprost (10a, LumiganW, Allergan, Irvine, CA,
USA) and tafluprost (12, TaflotanW, Santen Oy, Finland) are
the active ingredients of topical hypotensive medications
approved by the U.S. Food and Drug Administration and the
European Medicines Agency for use as first-line agents for
lowering intraocular pressure in patients with glaucoma
or ocular hypertension. The ability of PGF2a analogues to
effectively reduce intraocular pressure with once-per-day
dosing, their ocular tolerability comparable to timolol, and the
general lack of systemic adverse effects have made them the
mainstay of pharmacological therapy for glaucoma and ocular
hypertension all over the world.
Synthesis of diverse prostaglandin analogues over the past
decades has attracted the attention of many organic chemists,
as attested by the number of papers dealing with this subject.
So far, however, there is no convergent methodology that
would allow effective and highly diastereoselective prepara-
tion of a whole series of prostaglandin analogues from a com-
mon and structurally advanced prostaglandin intermediate.
Most synthetic procedures used to attain PGF2a (1) and its
analogues 2 ꢀ 12 employ one of the known variants of the
Corey method, in which lower and upper side chains are
sequentially attached in a specific order to a derivative of
the commercially available (–)-Corey aldehyde/lactone.2–11
The Corey strategy involves first the installation of the lower
Additional Supporting Information may be found in the online version of this
article.
Contract grant sponsor: EU European Regional Development Fund, Innova-
tive Economy Operational Programme; Contract grant number: UDA-
POIG.01.03.01-14-068/08-00.
*Correspondence to: Iwona Dams, R&D Chemistry Department, Pharmaceuti-
cal Research Institute, Rydygiera 8, 01-793 Warsaw, Poland. E-mail: i.dams@
ifarm.eu
Received for publication 8 August 2012; Accepted 28 September 2012
DOI: 10.1002/chir.22123
Published online 5 February 2013 in Wiley Online Library
(wileyonlinelibrary.com).
© 2013 Wiley Periodicals, Inc.

SYNTHESIS OF THERAPEUTICALLY USEFUL PGF2 ANALOGUES
171
Fig. 1. Chemical structure of PGF2a (1) and synthetic prostanoids 2–12 used in the therapy of glaucoma.
has been associated with maintaining the purity at the o-chain
C-15 carbinol asymmetric center, which typically is formed by
a diastereoselective reduction. The known methods for such
reduction require expensive reagents, often low-temperature
conditions, and, in practice, do not lead to a completely
stereoselective formation of the desired prostaglandin diaste-
reoisomer.2–11,14 The 15-epi isomers of diverse PGF2a analo-
gues, including (15S)-latanoprost, (15S)-travoprost, and
(15R)-bimatoprost (10b), have been reported to form even
in diastereoselective reduction of the o-chain 15-keto func-
tion in amounts exceeding at least several percent.2-11,14
The physicochemical properties of 15R/15S PGF2a diastereoi-
somers are usually very similar, so it is not a straightforward
task to discern between them by NMR spectroscopy, optical
rotation, or even analytical reverse-phase high-performance
liquid chromatography (HPLC). Furthermore, the application
of preparative scale HPLC to remove even small amounts
of undesired 15-epi isomer is usually very costly and far
from trivial, whereas 5,6-trans diastereoisomer is much easier
to separate.
A growing demand for hypotensive PGF2a analogues stimu-
lates the need for the development of a novel convergent strat-
egy for prostaglandin synthesis that would also overcome
disadvantages of existing synthetic methods. The Corey
lactone still provides the most facile and practical access to
PGF2a analogues, yet the difficulties with 15-epi isomers need
to be overcome. In 2007, we reported preparation of the struc-
turally advanced prostaglandin phenylsulfone (+)-(5Z)-15
(Scheme 1) with the attached a-chain possessing carboxyl
group protected by 4-methyl-2,6,7-trioxabicyclo[2.2.2]octane
(methyl-OBO).15 The o-chain elongation of the phenylsulfone
15 by SN2 alkylation with enantiomerically pure alkyl halides
allowed the synthesis of 13,14-dihydro-15-ol PGF2a analogues
with the desired C-15 asymmetric center configuration, such
as latanoprost (6). Importantly, we envisaged that the new
prostaglandin key intermediate 15 could find the application
in convergent syntheses of diverse active prostaglandins and
related compounds, depending on the structure of o-chain
and the sequence of chemical reactions needed for the o-chain
elongation. Since two other commercially available 13,14-en-
15-ol PGF2a drugs, travoprost (8) and bimatoprost (10a),
contain the trans-13,14-double bond in the o-chain, a new
approach for the highly diastereoselective construction of the
o-chain allylic alcohol moiety with the desired C-15 asymmetric
center configuration needed to be developed.
Bimatoprost (LumiganW) is a commonly prescribed syn-
thetic prostamide that continues to hold a key position in
the antiglaucoma drug market with global sales of over US
$849 million in 2011. Evidence suggest that bimatoprost
reduces intraocular pressure more effectively than any other
single medication available for glaucoma management.11,16
Additionally, bimatoprost is the active ingredient in LatisseW
(0.03% solution, Allergan), a new prescription medicine to
treat hypotrichosis of eyelashes. As a result of growing
demand for prostaglandin antiglaucoma drugs in the pharma-
ceutical market, we have developed a novel convergent
synthesis of bimatoprost (10a) from the structurally advanced
prostaglandin phenylsulfone 15 and an enantiomerically pure
a-hydroxy protected aldehyde (–)-(S)-16a (Scheme 2). This
article describes the first application of the prostaglandin
intermediate 15 to the highly diastereoselective synthesis of
trans-13,14-en-15-ol prostamide F2a analogues with the desired
C-15 asymmetric center configuration, based on the reverse
order of side chain attachment to the Corey lactone via Wittig
Scheme 1. Strategy for convergent synthesis of PGF2a analogues via prostaglandin phenylsulfone 15, based on the reverse order of side chain attachment to
the derivative of Corey lactone (13).
Chirality DOI 10.1002/chir

172
DAMS ET AL.
Scheme 2. A novel strategy for convergent synthesis of bimatoprost (10a) from the phenylsulfone 15 and aldehyde 16a.
liquid chromatograph coupled with an Applied Biosystems Qtrap 3200 mass
spectrometer equipped with a Chiralpak OD-3R (3 mm, 150 ꢂ 4.6 mm),
Kinetex XB-C18 (2.6 mm, 150 ꢂ 4.6 mm), Poroshell 120 EC-C8 (2.7 mm,
150 ꢂ 4.6 mm), and Xterra RP18 (3.5 mm, 150 ꢂ 4.6 mm) column. The
HPLC systems were qualified and HPLC methods were validated.
and Julia–Lythgoe olefinations. The preparation and identifica-
tion of two synthetic impurities, 15-epi isomer ((+)-(15R)-10b)
of bimatoprost and a new prostaglandin related amide (+)-
(5Z)-18 (Fig. 2), are also described.
EXPERIMENTAL
(–)-(S)-2-Hydroxy-4-phenylbutyl pivalate (22a)
Reagents
Trimethylacetyl chloride (1.57 ml, 12.634 mmol) was added to a stirred
solution of diol (–)-(S)-20a (2.0 g, 12.032 mmol) in a mixture of CH2Cl2
and pyridine (1:1, 50 ml) at 0 ꢁC under an argon atmosphere. After stirring
at 0 ꢁC for 1 h and at room temperature for 1 h, the reaction was quenched
with crushed ice (25 g) and the whole was portioned between AcOEt
(25 ml) and 10% aqueous HCl (25 ml). The resulting layers were sepa-
rated and the aqueous phase was extracted with AcOEt (3 ꢂ 25 ml). The
combined organic extracts were washed successively with H2O
(150 ml), saturated aqueous NaHCO3 (150 ml), and brine (200 ml), and
dried over anhydrous Na2SO4. Filtration and evaporation in vacuo
furnished the crude ester (3.27 g), which was purified by flash column
chromatography over silica gel (hexanes/AcOEt 4:1) to afford the
The chemical reagents were purchased from Sigma-Aldrich and Iris
Biotech Gmbh chemical companies. 1- < (Z)-6-{(1R,2R,3R,5S)-2-[(Phenyl-
sulfonyl)methyl]-3,5-bis(triethylsilyloxy)cyclopentyl}hex-4-enyl> -4-methyl-
2,6,7-trioxabicyclo{2.2.2]octane15 ((+)-(5Z)-15, a mixture of 5Z/5E isomers
in the ratio of 91.55%:8.45%, 83.1% de) and ((–)-S)-4-phenyl-1,2-butanediol15
((–)-(S)-20a, 99.6% ee) were manufactured by the Pharmaceutical
Research Institute (Warsaw, Poland). The HPLC standards, bimatoprost
(10a) and its diastereoisomers 10b–10c, were purchased from
Cayman Chemicals.
General Procedures
Analytical thin-layer chromatography (TLC) was performed on silica
gel DC-Alufolien Kieselgel 60 F254 (Merck) with mixtures of hexanes,
ethyl acetate, dichloromethane, methanol, and 2-propanol in various
ratios as developing systems. Compounds were detected by spraying
the plates with 1% Ce(SO4)2/2%H3[P(Mo3O10)4] in 10% H2SO4 followed
by heating to 120 ꢁC. Preparative column chromatography was carried
out on silica gel (Kieselgel 60, 40–63 mm, 230–400 mesh, Merck) with
mixtures of ethyl acetate, dichloromethane, methanol, and 2-propanol in
varying ratios as eluents. Flash chromatography was run with a positive
air pressure generated by handheld, pear-shaped rubber pumps. 1H
NMR spectra were recorded in CDCl3 solutions on a Varian VNMRS-
600 (600 MHz) spectrometer with TMS as internal standard. IR spectra
were taken for liquid films or KBr pellets on a Nicolet Imapct 410 FT-IR
spectrophotometer. HRMS spectra were recorded on an AMD 604
Inectra Gmbh spectrometer or a Mariner PE Biosystem ESI-TOF
spectrometer. Melting points were determined from DSC termograms
performed on a Mettler Toledo DSC822E differential scanning calorimeter.
Optical rotations were measured with a Perkin Elmer 341 automatic
polarimeter in acetone, EtOH, CH2Cl2, or CHCl3 solutions as the
solvents with percent concentrations. Analytical HPLC were performed
on a Waters Alliance 2695 system equipped with a 2998 PDA detector,
using the following columns: Chiracel OD OD00CE-EL068 (10 mm,
250 ꢂ 4.6mm), Chiracel OD-H (5mm, 250 ꢂ 4.6 mm), Chiralpak OD-3R
(3 mm, 150 ꢂ 4.6mm), Kinetex XB-C18 (2.6mm, 150 ꢂ 4.6 mm), Poroshell
120 EC-C8 (2.7 mm, 150 ꢂ 4.6 mm), and AS-3R (3mm, 150 ꢂ 4.6 mm). The
LC-MS were recorded on a Shimadzu LC-2010A HT high-performance
20
pivalate (–)-(S)-22a (2.83 g, 93.9% yield, 99.64% ee) as a white solid. ½aꢃD
= –19.97 (c 1.0, EtOH). mp 44.55–50.24 ꢁC, peak 47.02 ꢁC, heating rate
10.00 ꢁC/min. FT-IR (KBr) n (cmꢀ1): 3542, 3400, 3082, 3027, 2971, 2949,
2916, 1707, 1478, 1457, 1367, 1280, 1175, 1070, 1036, 917, 744, 700. 1H
NMR (600 MHz, CDCl3) d (ppm): 1.22 (s, 9H, –C(CH3)3), 1.80 (m, 2H,
CH2-3), 2.16 (s, 1H, –OH), 2.71 (m, 1H, one of the CH2-4 group), 2.82
(m, 1H, one of the CH2-4 group), 3.84 (m, 1H, CH-2), 4.02 (dd, J = 6.7
and 11.4 Hz, 1H, one of the CH2-1 group), 4.12 (dd, J = 3.4 and 11.4 Hz,
1H, one of the CH2-1 group), 7.18 (m, 1H, aromatic H-4), 7.20 (m, 2H,
aromatic H-2 and H-6), 7.28 (m, 2H, aromatic H-3 and H-5). 13C NMR
(150 MHz, CDCl3) d (ppm): 27.2 (3C, –C(CH3)3), 31.6 (C-4), 35.0 (C-3),
38.9 (–C(CH3)3), 68.5 (C-1), 69.3 (C-2), 126.0 (aromatic C-4), 128.4 (2C,
aromatic C-2 and C-6), 128.5 (2C, aromatic C-3 and C-5), 141.6 (aromatic
C-1), 178.8 ((CH3)3C-C(O)O–). HRMS (ESI): calcd. for C15H22O3Na
[M + Na]+ 273.14612; found 273.1451. HPLC: Chiracel OD-H, 5 mm,
250 ꢂ 4.6 mm column, hexanes/ethanol/2-propanol 100:1:1 (v/v/v),
0.7 ml/min, Rt = 22.779 min. (99.82% yield of 22a), Rt = 34.108 (0.18% yield
of 22b), 99.64% ee.
(+)-(S)-2-(tert-Butyldimethylsilyloxy)-4-phenylbutyl
pivalate (23a)
tert-Butyldimethylsilyl chloride (1.96 g, 12.607 mmol) was added in one
portion to a stirred solution of alcohol (–)-(S)-22a (2.63 g, 10.506 mmol)
and imidazole (1.16 g, 16.81 mmol) in anhydrous DMF (25 ml) at 0 ꢁC
under an argon atmosphere. The reaction was allowed to proceed for
Fig. 2. Structures of the key synthetic impurities 10b, 10c and 18 of bimatoprost.
Chirality DOI 10.1002/chir

SYNTHESIS OF THERAPEUTICALLY USEFUL PGF2 ANALOGUES
173
18 h at room temperature and then quenched with crushed ice (25 g). The
resulting mixture was portioned between hexanes (25 ml) and H2O
(50 ml). The aqueous layer was extracted with hexanes (3 ꢂ 25 ml). The
combined organic extracts were washed successively with H2O (100 ml)
and brine (150 ml) and dried over Na2SO4. Filtration and evaporation in
vacuo furnished the crude product (3.97 g) as a light yellow oil, which
was purified by flash column chromatography (silica gel, from 20:1 to
10:1 hexanes/AcOEt) to give TBDMS ether (+)-(S)-23a (3.64 g, 95.1%
at room temperature for 30 min. The resulting layers were separated and
the aqueous phase was extracted with CH2Cl2 (3 ꢂ 25 ml). The com-
bined extracts were washed with brine (3 ꢂ 150 ml) and dried over
Na2SO4. Filtration and evaporation in vacuo furnished the crude product
(2.34 g), which was purified by flash column chromatography (silica gel,
hexanes/AcOEt 10:1) to afford the aldehyde (–)-(S)-16a (2.22 g, 94.2%)
20
as a colorless oil. ½aꢃD = –19.77o (c 1.0, CHCl3). FT-IR (thin film) n (cmꢀ1):
3063, 3028, 2954, 2929, 2799, 2857, 1736, 1497, 1472, 1463, 1361, 1255,
1116, 1006, 973, 838, 778, 747, 699, 670. 1H NMR (600 MHz, CDCl3) d
(ppm): 0.086 (s, 3H, CH3-Si), 0.088 (s, 3H, CH3-Si), 0.95 (s, 9H, (CH3)3C-Si),
1.96 (m, 2H, CH2-3), 2.71 (m, 2H, CH2-4), 4.02 (ddd, J = 1.5, 5.3 and 6.8 Hz,
1H, CH-2), 7.18 (m, 2H, aromatic H-2 and H-6), 7.19 (m, 1H, aromatic
H-4), 7.28 (m, 2H, aromatic H-3 and H-5), 9.59 (d, J = 1.5 Hz, 1H, –CHO).
13C NMR (150 MHz, CDCl3,) d (ppm): –4.6 (CH3-Si), –4.9 (CH3-Si), 18.2
((CH3)3C-Si), 25.8 (3C, (CH3)3C-Si), 30.8 (C-4), 34.5 (C-3), 77.1 (C-2),
126.1 (aromatic C-4), 128.4 (2C, aromatic C-2 and C-6), 128.5 (2C, aromatic
C-3 and C-5), 141.2 (aromatic C-1), 204.1 (–CHO).
20
yield) as a colorless oil. ½aꢃD = +3.98 (c 1.0, EtOH). FT-IR (thin film) n
(cmꢀ1): 3063, 3027, 2956, 2930, 2857, 1732, 1496, 1472, 1462, 1397,
1362, 1283, 1256, 1155, 1120, 1064, 1004, 836, 776, 699. 1H NMR
(600 MHz, CDCl3) d (ppm): 0.09 (s, 6H, (CH3)2Si), 0.91 (s, 9H, (CH3)3C-
Si), 1.21 (s, 9H, (CH3)3C-C), 1.82 (m, 2H, CH2-3), 2.64 (ddd, J = 5.8, 10.8
and 13.8 Hz, 1H, one of the CH2-4 group), 2.72 (ddd, J = 5.8, 10.8 and
13.8 Hz, 1H, one of the CH2-4 group), 3.91 (m, 1H, CH-2), 4.02 (m, 2H,
CH2-3), 7.18 (m, 2H, aromatic H-2 and H-6), 7.19 (m, 1H, aromatic H-4),
7.28 (m, 2H, aromatic H-3 and H-5). 13C NMR (150 MHz, CDCl3) d
(ppm): –4.7 (CH3-Si), –4.5 (CH3-Si), 18.0 ((CH3)3C-Si), 25.8 (3C, (CH3)3C-
Si), 27.2 (3C, (CH3)3C-C), 31.3 (C-4), 36.5 (C-3), 38.8 ((CH3)3C-C),
67.7 (C-1), 69.6 (C-2), 125.8 (aromatic C-4), 128.3 (2C, aromatic C-2 and
C-6), 128.4 (2C, aromatic C-3 and C-5), 142.2 (aromatic C-1), 178.4
((CH3)3CC(O)O–) ppm. HRMS (ESI): calc. for C21H36O3NaSi [M + Na]+
387.2326; found 387.2326.
Method B. Sulfur trioxide pyridine complex (9.381g, 57.759 mmol) was
added over 15min to a stirred solution of alcohol (–)-(S)-24a (5.4 g,
19.253 mmol) and Et3N (16.35ml, 115.518 mmol) in anhydrous DMSO
(90 ml) under an argon atmosphere. After stirring for 1 h, the mixture was
diluted with CH2Cl2 (180 ml) and poured into saturated aqueous NH4Cl
(180 ml) at 0 ꢁC. The resulting layers were separated and the aqueous phase
was extracted with CH2Cl2 (3 ꢂ 50ml). The combined organic extracts were
washed successively with H2O (150 ml) and brine (150 ml) and dried over
anhydrous Na2SO4. Filtration and evaporation in vacuo furnished the crude
product (5.34 g), which was purified by flash column chromatography
over silica gel (hexanes/AcOEt 10:1) to afford the aldehyde (–)-(S)-
(–)-(S)-2-(tert-Butyldimethylsilyloxy)-4-phenylbutan-1-
ol (24a)
Diisobutylaluminum hydride in toluene (1.0 M, 24 ml, 24.0 mmol) was
added dropwise over 15 min to a stirred solution of pivalate (+)-(S)-23a
(3.45 g, 9.463 mmol) in anhydrous CH2Cl2 (50 ml) at –78 ꢁC under an
argon atmosphere. The resulting mixture was allowed to warm to –20ꢁC
for a 30-min period and stirred at this temperature for another 2 h. TLC
analysis (hexanes/AcOEt 8:1) indicated disappearance of the starting
pivalate (+)-(S)-23a. The clear colorless solution was recooled to –
78 ꢁC and the excess of DIBAL was quenched by addition of MeOH
(15 ml) dropwise. On warming to 0 ꢁC, 10% aqueous potassium sodium
tartrate (100 ml) was added and the mixture was stirred vigorously at
room temperature for 2 h. The resulting layers were separated and the
aqueous layer was extracted with CH2Cl2 (3 ꢂ 25 ml). The combined
extracts were washed with water (100 ml) and brine (150 ml) and dried
over anhydrous Na2SO4. Filtration and evaporation in vacuo furnished
the crude product (2.89 g), which was purified by flash column chroma-
tography (silica gel, 3–9% hexanes/AcOEt) to afford the primary alcohol
20
16a (4.503 g, 84.0%) as a colorless oil. ½aꢃD = –19.24 (c 1.0, CHCl3).
The characterization data from IR and NMR spectra were identical in
all aspects with those of (–)-(S)-16a obtained according to Method A.
1-{(Z)-6-[(1R,2R,3R,5S)-2-[(1R/1S,2R/2S,3S)-3-(tert-
Butyldimethylsilyloxy)-5-phenyl-1-(phenylsulfonyl)pent-1-yl]-
3,5-bis(triethylsilyloxy)cyclopentyl]hex-4-enyl}-4-methyl-2,6,
7-trioxabicyclo[2.2.2] octane (17a)
n-BuLi (4.50 ml, 7.20 mmol, 1.6 M in hexanes) was added dropwise to a
solution of diisopropylamine (0.65 ml, 7.755 mmol) in anhydrous THF
(10 ml) at –78 ꢁC under an argon atmosphere. After 15 min at –78 ꢁC,
the phenylsulfone 15 (3.85 g, 5.539 mmol) in anhydrous THF (5 ml)
was added dropwise with vigorous stirring. The reaction mixture was
allowed to warm to –40 ꢁC and the crude aldehyde (–)-(S)-16a (2.0 g,
7.20 mmol) in anhydrous THF (5 ml) was added dropwise. After being
stirred at –40 ꢁC for 1 h, TLC analysis (hexanes/AcOEt 5:1) indicated
disappearance of the starting phenylsulfone 15. The cold reaction was
quenched with saturated aqueous NaCl solution (50 ml). The resulting
layers were separated and the aqueous phase was extracted with AcOEt
(3ꢂ 25ml). The combined organic extracts were washed with brine
(150 ml), dried over Na2SO4, filtered, and concentrated in vacuo. The crude
product was carefully dried in vacuo to yield the crude mixture of epimeric
hydroxy sulfones 17a (5.98g) as a light yellow oil. The hydroxy sulfones
17a were directly used for the next step without further purification.
(–)-(S)-24a (2.51 g, 94.6% yield, 99.68% ee) as a colorless oil.½aꢃ2D0 = –12.69
(c 1.0, EtOH). FTIR (thin film) n (cmꢀ1): 3420, 3063, 3026, 2953, 2929, 2857,
1469, 1496, 1472, 1462, 1388, 1361, 1255, 1113, 1045, 988, 836, 776, 698,
1
665 cm–1. H NMR (600MHz, CDCl3) d (ppm): 0.08 (s, 3H, CH3-Si), 0.09
(s, 3H, CH3-Si), 0.91 (s, 9H, (CH3)3C-Si), 1.83 (m, 2H, CH2-3), 1.85 (m, 1H,
–OH), 2.64 (m, 2H, CH2-4), 3.52 (dd, J = 5.2 and 11.1Hz, 1H, one of the
CH2-1 group), 3.60 (dd, J = 3.8 and 11.1 Hz, 1H, one of the CH2-1 group),
3.79 (m, 1H, CH-2), 7.18 (m, 2H, aromatic H-2 and H-6), 7.19 (m, 1H,
aromatic H-4), 7.28 (m, 2H, aromatic H-3 and H-5). 13C NMR (150 MHz,
CDCl3) d (ppm): –4.5 (CH3-Si), –4.5 (CH3-Si), 18.1 ((CH3)3C-Si), 25.9
(3C, (CH3)3C-Si), 31.7 (C-4), 35.7 (C-3), 66.1 (C-1), 72.4 (C-2), 125.8
(aromatic C-4), 128.3 (2C, aromatic C-2 and C-6), 128.4 (2C, aromatic C-3
and C-5), 142.0 (aromatic C-1). HRMS (ESI): calc. for C16H28O2NaSi
[M + Na]+ 303.17508; found 303.1750. HPLC: Chiracel OD-H, 5mm,
250 ꢂ 4.6mm column, hexanes/ethanol/methanol 98:1.5:0.5 (v/v/v),
1.0ml/min, Rt = 8.607min. (99.84% yield of 24a), Rt = 10.846 (0.16% yield of
24b), 99.68% ee.
1-{(Z)-6-[(1R,2R,3R,5S)-2-[(S,E)-3-(tert-
Butyldimethylsilyloxy)-5-phenyl-1-penten-1-yl]-3,5-bis-
(triethylsilyloxy)cyclopentyl]hex-4-enyl}-4-methyl-2,6,7-
trioxabicyclo[2.2.2] octane (25a)
The crude mixture of diastereoisomeric hydroxy sulfones 17a (5.98 g)
was dissolved in anhydrous MeOH (100 ml) and Na2HPO4 (6.29 g,
44.312 mmol) was added in one portion. On cooling to 0 ꢁC under an
argon atmosphere, sodium amalgam (20%, 7.64 g, 66.468 mmol Na) was
added portionwise over 1 h with vigorous stirring. The cooling bath was
removed and the resulting suspension was stirred for 5 h at room temper-
ature until the disappearance of the starting hydroxy sulfones 17a (TLC
hexanes/AcOEt 5:1). The solution was decanted from the remaining
amalgam and concentrated under reduced pressure. The residue was
diluted with water (100 ml) and AcOEt (50 ml). The resulting layers were
separated and the aqueous phase was extracted with AcOEt (3 ꢂ 25 ml).
(–)-(S)-2-(tert-Butyldimethylsilyloxy)-4-phenylbutanal (16a)
Method A. Dess-Martin periodinane (4.45 g, 10.182 mmol) was added
portionwise to a cold (0 ꢁC) suspension of alcohol (–)-(S)-24a (2.38 g,
8.485 mmol) and dry NaHCO3 (2.14 g, 25.455 mmol) in anhydrous CH2Cl2
(50 ml). After being stirred for 1 h at room temperature, TLC analysis
(hexanes/AcOEt 10:1) indicated disappearance of the starting alcohol (–)-
(S)-24a. Saturated aqueous NaHCO3 (100ml) and Na2SO3 (7.49g,
59.395 mmol) were then added simultaneously and the mixture was stirred
Chirality DOI 10.1002/chir

174
DAMS ET AL.
The combined organic extracts were washed with brine (150 ml) and
dried over anhydrous Na2SO4. Filtration and evaporation in vacuo
furnished the crude product 25a (4.85 g) as a light yellow oil. The crude
olefine 25a was directly used for the next step without further purification.
5–10% CH3OH/AcOEt to afford the title compound 10a (1.48 g, 87.2%
yield, 10a:10b:(5E,15S)-10c = 91.50%:0.12%:8.38%) as a pale yellow oil.
The oil was treated with tert-butyl methyl ether (20 ml); the resulting pre-
cipitate was filtered off and dried under reduced pressure to give bimato-
prost 10a. This sample was further purified by recrystallization from a
mixture of ethyl acetate/tert-butyl methyl ether to give a pharmaceutical
2,2-Bis(hydroxymethyl)propyl (Z)-7-{(1R,2R,3R,5S)-3,5-
dihydroxy-2-[(S,E)-3-hydroxy-5-phenylpent-1-enyl]-
cyclopentyl}hept-5-enoate (26a)
20
grade bimatoprost 10a (1.25 g, 73.7% yield) as a white solid. ½aꢃD = +39.07
20
(c 1.0, CH2Cl2). (lit.10a ½aꢃD = +32.7 (c 0.33, CH2Cl2)). mp 65.70–72.70ꢁC,
peak 69.52 ꢁC, heating rate 10.00 ꢁC/min (lit.10a mp 67–68ꢁC). FT-IR
(KBr) n (cmꢀ1): 3420, 3327, 3084, 3011, 2914, 2865, 2933, 1620, 1546,
1496, 1456, 1372, 1317, 1290, 1249, 1151, 1097, 1055, 1027, 976, 920,
698. 1H NMR (600 MHz, CDCl3) d (ppm): 1.10 (t, J = 7.2 Hz, 3H, –
CH2CH3), 1.46 (m, 1H, CH-8), 1.62 (m, 1H, one of the CH2-3 group),
1.68 (m, 1H, one of the CH2-3 group), 1.74 (m, 1H, one of the CH2-10
group), 1.78 (m, 1H, one of the CH2-16 group), 1.90 (m, 1H, one of the
CH2-16 group), 2.02–2.06 (m, 2H, one of the CH2-4 group and one of
CH2-7 group), 2.11–2.15 (m, 3H, CH2-2 and one of the CH2-4 group), 2.21
(m, 1H, one of the CH2-10 group), 2.29 (m, 1H, one of the CH2-7 group),
2.34 (m, 1H, CH-12), 2.67 (m, 2H, CH2-17), 3.22 (m, 2H, –CH2CH3), 3.55
(s, 3H, three –OH), 3.91 (m, 1H, CH-11), 4.08 (m, 1H, CH-15), 4.12 (m,
1H, CH-9), 5.34 (m, 1H, CH-5), 5.41 (m, 1H, CH-6), 5.47 (dd, J = 9.0 and
15.3 Hz, 1H, CH-13), 5.59 (dd, J = 7.3 Hz and 15.3 Hz, 1H, CH-14), 5.98
(t, J = 5.1 Hz, 1H, >NH), 7.17 (m, 1H, aromatic H-4), 7.18 (m, 2H, aromatic
H-2 and H-6), 7.26 (m, 2H, aromatic H-3 and H-5). 13C NMR (150MHz,
CDCl3) d (ppm): 14.8 (–CH2CH3), 25.4 (C-7), 25.6 (C-3), 26.7 (C-4), 31.9
(C-17), 34.4 (–CH2CH3), 35.8 (C-2), 38.8 (C-16), 42.9 (C-10), 50.2 (C-8),
55.5 (C-12), 72.3 (C-15), 72.3 (C-9), 77.7 (C-11), 125.8 (aromatic C-4), 128.4
(2C, aromatic C-3 and C-5), 128.4 (2C, aromatic C-2 and C-6), 129.2 (C-6),
129.7 (C-5), 133.2 (C-13), 135.1 (C-14), 142.0 (aromatic C-1), 173.4 (C-1).
TBAF (16.62 ml, 16.620 mmol, 1.0 M in THF) was added dropwise to a
solution of the crude silyl protected prostaglandin analogue 25a (4.85 g)
in anhydrous THF (15 ml). The resulting brown solution was stirred at
65–70 ꢁC under an argon atmosphere for 1 h, whereupon THF was evapo-
rated under reduced pressure. The viscous residue was then diluted with
10% solution of citric acid (25 ml) to hydrolyze 4-methyl-OBO carboxyl
masking group. After being stirred for 15 min, the product was salted
out with sodium chloride, separated and dried in vacuo to give the crude
pentaol 26a (3.76 g). Purification by silica gel flash chromatography with
gradient elution from 1% to 10% methanol/ethyl acetate afforded the pros-
taglandin analogue 26a (2.39 g, 87.9% yield from phenylsulfone 15,
26a:26b:(5E,15S)-isomer = 91.42%:0.17%:8.41%) as a light yellow viscous
20
oil. ½aꢃD = +29.58 (c 1.0, CHCl3). FT-IR (thin film) n (cmꢀ1): 3373, 3024,
2932, 2837, 1732, 1603, 1496, 1454, 1374, 1246, 1172, 1047, 972, 923,
748, 701, 609. 1H NMR (600 MHz, CDCl3) d (ppm): 0.84 (s, 3H, –CH3),
1.48 (m, 1H, cyclopentane CH-1), 1.67–1.70 (m, 3H, a-chain CH2-3 and
one of the cyclopentane CH2-4 group), 1.80 (m, 1H, one of the o-chain
CH2-4 group), 1.90 (m, 1H, one of the o-chain CH2-4 group), 2.06–2.09
(m, 2H, one of the a-chain CH2-7 group and one of the a-chain CH2-4
group), 2.12–2.26 (m, 3H, one of the a-chain CH2-4 group, one of the
a-chain CH2-7 group and one of the cyclopentane CH2-4 group), 2.32
(m, 1H, cyclopentane CH-2), 2.34 (m, 2H, a-chain CH2-2), 2.67 (m, 2H,
o-chain CH2-5), 3.52 (m, 4H, two –CH2OH), 3.89 (m, 1H, cyclopentane
CH-3), 4.06 (m, 1H, o-chain CH-3), 4.08 (m, 2H, a-chain CH2-1), 4.11 (m,
1H, cyclopentane CH-5), 5.34 (m, 1H, a-chain CH-5), 5.42 (m, 1H, a-chain
CH-6), 5.45 (dd, 1H, J = 9.0 Hz and 15.29 Hz, o-chain CH-1), 5.58
(dd, J = 7.5 Hz and 15.2Hz, o-chain CH-2), 7.17 (m, 1H, aromatic H-4), 7.19
(m, 2H, aromatic H-2 and H-6), 7.26 (m, 2H, aromatic H-3 and H-5). 13C
NMR (150 MHz, CDCl3) d (ppm): 16.8 (–CH3), 24.7 (a-chain C-3), 25.6
(a-chain C-7), 26.5 (a-chain C-4), 31.8 (o-chain C-5), 33.5 (a-chain C-2),
38.7 (o-chain C-4), 40.5 (–C(CH3)(CH2OH)2), 42.8 (cyclopentane C-4),
49.5 (cyclopentane C-1), 55.4 (cyclopentane C-2), 66.5 (–CH2C(CH3)
(CH2OH)2), 66.6 (2C, two –CH2OH), 72.4 (o-chain C-3), 72.5 (cyclopentane
C-5), 77.5 (cyclopentane C-3), 125.8 (aromatic C-4), 128.4 (2C, aromatic C-3
and C-5), 128.5 (2C, aromatic C-2 and C-6), 129.3 (a-chain C-6), 129.4
(a-chain C-5), 133.3 (o-chain C-1), 135.2 (o-chain C-2), 141.9 (aromatic C-
1), 174.6 (C= O). HRMS (ESI): calc. for C28H42O7Na [M + Na]+ 513.28228;
found 513.2837. Chiral HPLC: Chiralpak OD-3R, 3mm, 150ꢂ 4.6 mm
column, H2O/TEA (1000/1, adjusted to pH = 4.5 with H3PO4 (phase A)/
CH3CN (phase B) with gradient elution 80–10%, 1.0 ml/min, Rt = 25.01 min.
(91.42% yield of (+)-(15S)-26a), Rt = 26.91min. (8.41% yield of (5E,15S)-
isomer) Rt = 27.68 min. (0.17% yield of (+)-(15R)-26b). LC-MS (ESI):
Chiralpak OD-3R, 3 mm, 150 ꢂ 4.6 mm column, H2O/TEA (1000/1, ad-
justed to pH = 4.5 with H3PO4 (phase A)/CH3CN (phase B) with gradient
elution 80–10%, 1.0 ml/min, Rt = 26.55 min. (m/z = 490.2 [M + H]+ for (+)-
(15S)-26a), Rt = 28.34 min. (m/z = 490.2 [M + H]+ for (5E,15S)-isomer),
Rt = 30.03 min. (m/z = 490.2 [M + H]+ for (+)-(15R)-26b).
HRMS (ESI): calc. for
C
25H37NO4Na [M + Na]+ 438.26148; found
438.2632. HPLC: Kinetex XB-C18, 2.7mm, 150 ꢂ 4.6 mm column, H2O/
CH3CN (8:2, phase A)/ H2O/CH3CN (1:1, phase B) with gradient elution
90–70%, 1.0ml/min, Rt = 21.44min. (0.12% yield of (+)-(15R)-10b), Rt = 21.85
min. (8.38% yield of (5E,15S)-10c), Rt = 22.56min. (91.50% yield of (+)-(15S)-
10a). LC-MS (ESI): Kinetex XB-C18, 2.7mm, 150 ꢂ 4.6mm column, (600 ml
NH3 ꢄ H2O: 500 ml CH3COOH: 1 dm3 H2O): CH3CN (8:2, phase A)/
(600 ml NH3 ꢄ H2O: 500 ml CH3COOH: 1 dm3 H2O): CH3CN (8:1, phase
B) with gradient elution 100–75%, 1.0 ml/min, Rt = 21.79 min. (m/
z = 416.3 [M + H]+ for (+)-(15R)-10b), Rt = 22.33 min. (m/z = 416.3
[M + H]+ for (5E,15S)-10c), Rt = 22.91 min. (m/z = 416.3 [M + H]+ for
(+)-(15S)-10a).
RESULTS AND DISCUSSION
The accessibility of the stable prostaglandin phenylsulfone
15 (Scheme 1) drew our attention to Julia–Lythgoe olefina-
tion17, which via reductive elimination of b-hydroxy sulfones
could enable construction of the allylic alcohol moiety of
various prostaglandins. The unique stereochemical features of
the Julia–Lythgoe olefination should allow us to preserve the
relative cis/trans stereochemistry of side chains in the phenyl-
sulfone 15 and give rise to stereodefined trans-13,14-double
bond (Fig. 1) required in the o-chain. It is well documented that
sodium amalgam reduction of aliphatic a-hydroxy sulfones
furnishes olefins having the trans configuration of the o-chain
13,14-double bond.17b Additionally, the appropriately functiona-
lized aldehydes with steric encumbrance should afford trans
alkenes as the only products of Julia reaction.18
(8R,9S,11R,15S)-9,11,15-Trihydroxy-17-phenyl-18,19,20-
trinor-5Z,13E-prostadienoic acid ethylamide (10a)
EtNH2 (20 ml, 251.553 mmol, 70% in water) was added in one portion to
the prostaglandin analogue 26a (2.0 g, 4.076 mmol). The resulting solu-
tion was stirred for 60 h to disappearance of the starting material 26a
(TLC, CH2Cl2/CH3OH 9:1). The excess of EtNH2 was then removed by
evaporation under reduced pressure and the aqueous residue was diluted
with AcOEt (25 ml). The resulting layers were separated and the aqueous
phase was extracted with AcOEt (3 ꢂ 25 ml). The combined extracts were
washed with brine (150 ml) and dried over anhydrous Na2SO4. Filtration
and evaporation in vacuo furnished the crude product 10a (1.93 g), which
was purified by silica gel flash chromatography with gradient elution from
In the chemical structure of bimatoprost (10a) the carbon
17 position in the o-chain is substituted by a phenyl ring. We
envisaged that combining the sulfone 15 (Scheme 2) with the
a-hydroxy protected aldehyde 16a should give the mixture of
hydroxy sulfones 17a, which in a few simple steps could be
transformed to bimatoprost 10a. Since the first step of the
Julia–Lythgoe protocol is a nucleophilic addition of a-sulfonyl
carbanion to aldehyde, the aldehyde synthon should possess
the stereochemical arrangement corresponding to the “15S”
Chirality DOI 10.1002/chir

SYNTHESIS OF THERAPEUTICALLY USEFUL PGF2 ANALOGUES
175
stereochemistry in the final prostaglandin analogue. Critical
to this disconnection, the enantiomeric purity of the aldehyde
o-chain synthon 16a should be very high, thus a priori estab-
lishing a very high diastereoisomeric excess of the final prod-
uct 10a. Importantly, the aldehyde o-synthon 16a with a
well-chosen protecting group should not be capable of epi-
merization under basic reaction conditions.
source of chirality. Thus, the diol 20a was synthesized in
high enantiomeric purity (99.60% ee) from the aldehyde 19,
which is commercially available or can be easily prepared
from an important chiral template 1,2:5,6-di-O-isopropylidene-
D-mannitol.15 The enantiomeric impurity (+)-(R)-20b was
synthesized by Mitsunobu21 esterification of the diol 20a with
the excess of p-nitrobenzoic acid to the corresponding bis-
(4-nitrobenzoate) 21a (Scheme 3), followed by subsequent
hydrolysis of diester 21a under mild basic conditions to
the diol (+)-(R)-20b (99.50% ee). The course of all reactions
and the enantiomeric purity of products were controlled by
TLC and HPLC. The stereochemical assignment of the com-
pounds 20b and 21a was confirmed by the values of optical
rotation.22 Their structures were also proved by NMR and IR
spectra.
Selecting a protecting group strategy for hydroxyl groups in
the diol 20a that would minimize side reactions was critical in
the design of the synthetic approach to the new aldehyde 16a.
Initially, we thought that selective monotosylation23 of the
primary hydroxyl group of 20a followed by silylation of the
secondary hydroxyl group and Kornblum oxidation24 of tosy-
late would be the best way to synthesize the aldehyde 16a.
However, monotosylation of the diol 20a was not selective
enough, yielding 20% of the unwanted secondary monotosylate.
The Kornblum oxidation of the a-silyloxy protected monotosy-
late also failed, probably due to the fast decomposition of the
final product 16a. After exploring various possibilities, tri-
methyl acetyl chloride was found to be the best protecting
reagent. Selective esterification of the primary hydroxyl
group of 20a with pivaloyl chloride afforded the new pivalo-
ate a-hydroxy ester (–)-(S)-22a as the main product of this
reaction (93.9% yield, 99.64% ee). A silicon-based protecting
group strategy at C-2 was envisaged to be sufficiently stable
to the reaction sequence yet easily removed under mild basic
conditions in the final steps of the synthesis. Our later
attempts to olefinate the phenylsulfone 15 with various a-
silyloxy protected aldehydes 16a using the Julia–Lythgoe
protocol showed that the tert-butyldimethylsilyl protecting group
provides better 15S/15R stereoselectivity for o-chain elongation
than triethylsilyl protection. Thus, the secondary hydroxyl group
of 22a was protected by silylation with tert-butyldimethylsilyl
Synthesis of the Aldehyde v-Chain Synthon (–)-(S)-16a and
Its Enantiomeric Impurity (+)-(R)-16b
To the best of our knowledge, the enantiomerically pure alde-
hydes (–)-(S)-16a and (+)-(R)-16b (Scheme 3) have never
been obtained in optically active form. In 1997, Uguen reported
the preparation of racemic 2-(tert-butyldimethylsilyloxy)-4-
phenylbutanal (16) based on DIBAL-H reduction of a-tert-
butyldimethylsilyloxy protected cyanohydrin obtained from
3-phenylpropanal by Kiliani–Fischer synthesis.19 Another
low-yield strategy for the racemic 16 involved oxidative cleav-
age of the double bond of 4-(tert-butyldimethylsilyloxy)-6-
phenylhex-1-ene with a mixture of OsO4, N-methylmorpholine
N-oxide, and NaIO4.20 Similar conditions applied to the racemic
3-(tert-butyldimethylsilyloxy)-5-phenylpentanal caused uncom-
mon degradation to one carbon lower homologue 16 with only
58% yield.20 Unfortunately, none of these methods allows the
preparation of the optically active aldehyde o-chain synthon
(–)-(S)-16a. The novel aldehyde synthon 16a used for the
construction of the o-chain of bimatoprost (10a) contains a
stereogenic center at the carbon C-2 corresponding to C-15 in
the target molecule, with chirality corresponding to the “15S”
configuration in the final product 10a. In all areas of pharma-
ceutical research it is most essential to investigate pure and
well-defined compounds. Thus, the aldehyde 16a should
possess a very high enantiomeric excess, preferably above
99.8% ee, as individual impurity content below 0.1% facilitates
the registration process of active pharmaceutical ingredients.
We envisaged that the aldehyde o-chain synthon 16a or its
enantiomeric impurity 16b could be prepared in four-step
syntheses from optically active diol (–)-(S)-20a or (+)-(R)-
20a (Scheme 3). The desired enantiomeric purity of the
aldehyde 16a was achieved by employing the known (+)-
(R)-2,2-dimethyl-1,3-dioxolane-4-carboxaldehyde (19) as a
Scheme 3. Synthesis of the aldehyde o-chain synthon (–)-(S)-16a and its enantiomer (+)-(R)-16b. Reagents and conditions: (1) p-NO2-PhCOOH, PPh3, DIAD,
toluene, 0 ꢁC, then 2 h at r.t., 88.7% yield; (2) LiOH . H2O, MeOH, 5 h, 97.4% yield (99.50% ee); (3) PivCl, Py, CH2Cl2, 0 ꢁC for 1 h, then 1 h at r.t., 93.9% yield of (–)-(S)-
22a (99.64% ee); (4) TBDMSCl, ImH, DMF, 0 ꢁC for 15 min, then 18 h at r.t., 95.1% yield of (+)-(S)-23a; (5) DIBAL-H, CH2Cl2, –78 ꢁC for 20 min, 2 h at r.t., 94.6%
yield of (–)-(S)-24a (99.68% ee); (6) Method A: DMP, NaHCO3, CH2Cl2, 0 ꢁC, then 1 h at r.t., 94.2% yield of (–)-(S)-16a. Method B: SO3. Py, TEA, DMSO, 1 h,
84.0% yield of (–)-(S)-16a.
Chirality DOI 10.1002/chir

176
DAMS ET AL.
chloride to give the silyl ether (+)-(S)-23a with 95.1% yield. Sub-
sequent deprotection of the C-1 pivaloate ester with DIBAL-H
provided the desired alcohol (–)-(S)-24a in 84.5% yield over
three steps from the diol 20a. The optical purity of the alcohol
24a was determined as 99.68% ee on the basis of HPLC using
a chiral stationary phase. The enantiomeric impurity (+)-(R)-
24b (99.51% ee) was synthesized by the same strategy as illus-
trated in Scheme 3, but starting from the chiral diol (+)-(R)-20b
(99.50% ee).
(+)-(15S)-10a according to the convergent strategy illustrated
in Scheme 4. The main problems accompanying the o-chain
elongation of phenylsulfone 15 to the prostaglandin pentahy-
droxy ester (+)-(15S)-26a were moderate reactivity of the alde-
hyde 16a and the lability of triethylsilyl protecting groups in
phenylsulfone 15. Considering the chemical structure of the
o-chain synthon 16a, phenyl- and tert-butyldimethylsililoxy-
substituents with electron-withdrawing properties should
activate the aldehyde 16a in nucleophilic addition reactions.
On the other hand, the bulky a-situated tert-butyldimethylsilyl
group creates steric hindrance reducing the ability of the
phenylsulfone 15 addition to the aldehyde 16a. The influ-
ence of steric encumbrance caused by a-silyloxy substituent
was clearly seen in our later attempts to olefinate the sulfone
15 with a bulky, a-tert-butyldiphenylsilyl protected aldehyde
16a (~60% conversion). Thus, the main problem accompa-
nying the synthesis of the prostaglandin precursor 17a was
the choice of an optimal base capable of efficient and highly
diastereoselective o-chain elongation of phenylsulfone 15
with sterically hindered aldehyde 16a. Based on the known
behavior of alkyl and aryl sulfones,17,27 we studied the effects
of various bases (LHMDS, KHMDS, n-BuLi, t-BuOK, LDA)
to determine the optimal conditions for high-yield and highly
diastereoselective synthesis of pentahydroxy ester 26a. Of all
the bases screened, freshly prepared LDA was found to be
the best base providing almost full conversion (~95%) of phenyl-
sulfone 15 to the mixture of diastereoisomeric hydroxy
sulfones 17a. The choice of 4-methyl-2,6,7-trioxabicyclo[2.2.2]-
octane (methyl-OBO) carboxyl masking group in phenylsul-
fone 15 was justified by its high stability under basic Julia–
Lythgoe reaction conditions. However, the limited stability of
the methyl-OBO protecting group on silica gel caused by the ef-
fect of lithium salts present in the crude reaction product
resulted in significant losses of 17a during column chromatog-
raphy. Thus, the crude mixture of b-hydroxy sulfones 17a was
subjected to reductive elimination with a fresh 20% sodium
amalgam to afford the methyl-OBO protected prostaglandin
precursor 25a. The phosphate-buffered conditions of Trost28
were used to ensure the optimal pH conditions and prevent
deprotection of the cyclopentane hydroxy groups. Similar to
the methyl-OBO protected b-hydroxy sulfones 17a, the prosta-
glandin intermediate 25a was not purified by silica gel column
The main problem accompanying the synthesis of the enan-
tiomerically pure aldehyde 16a was its high instability under
oxidation conditions. The Parikh–Doering oxidation25 (SO3 ꢄ
Py, DMSO, Et3N) of the alcohol 24a resulted in the aldehyde
16a formation with satisfactory yield (84.0%). However, after
exploring various possibilities, we have found that oxidation
of the alcohol 24a with Dess–Martin periodinane26 affords
the aldehyde 16a with the best yield and the highest chemi-
cal purity. The use of Dess–Martin periodinane allowed us to
avoid some of the difficulties encountered in using other
methods, such as adverse products, long reaction times, diffi-
cult workup procedures, or the need to apply a large excess of
the oxidizing agent. Most importantly, Dess–Martin periodi-
nane is well known not to racemize chiral centers of optically
active compounds. Thus, the enantiomeric purity of the alde-
hyde o-chain synthon 16a should be the same as the starting
alcohol 24a. a-Hydroxy protected aldehydes are usually read-
ily available and stable compounds. However, their stability
is often insufficient for identification purposes. It was not
surprising then, that we were unable to evaluate its enantio-
meric and chemical purity by HPLC on various stationary
phases. The enantiomeric aldehyde (+)-(R)-16b was synthe-
sized by the same strategy as illustrated in Scheme 3, but using
the chiral alcohol (+)-(R)-24b as the starting material. The pres-
ence of the aldehyde functional group in compounds 16a and
16b was confirmed by IR and NMR spectra.
Synthesis of Bimatoprost (+)-(15S)-10a and Its (+)-(15R)-
10b-epi Isomeric Impurity
With the novel aldehyde o-chain synthon (–)-(S)-16a and
the prostaglandin phenylsulfone (+)-(5Z)-1515 (a mixture
of 5Z/5E isomers in the ratio of 91.55%:8.45%, 83.1% de) in
hand, we continued the synthesis of the target compound
Scheme 4. Synthesis of bimatoprost (10a) and its (15R)-(+)-epi isomer 10b from phenylsulfone 15 and aldehyde 16a or 16b. Reagents and conditions used:
(1) LDA, THF, –78 ꢁC, 1 h; (2) 20% Na/Hg, Na2HPO4, MeOH, 0 ꢁC for 30 min, then 5 h at r.t.; (3) n-Bu4NF, 65 ꢀ 70 ꢁC, 1 h; (4) HOOCC(OH)(CH2COOH)2, H2O,
15 min, 87.9% yield of 26a from 15; (5) NH2Et, H2O, 60 h, 87.2% yield of 10a.
Chirality DOI 10.1002/chir

SYNTHESIS OF THERAPEUTICALLY USEFUL PGF2 ANALOGUES
177
chromatography. Significantly, there was no need for purifica-
tion of the prostaglandin intermediates 17a and 25a, which
was a major advantage for any future scale-up. The triethylsilyl
and tert-butyldimethylsilyl protecting groups in prostaglandin
precursor 25a were all removed under the action of tetrabuty-
lammonium fluoride (TBAF) solution, immediately followed
by hydrolytic opening of the methyl-OBO masking group with
citric acid solution to the stable 2,2-bis(hydroxymethyl)propyl
ester 26a. The epimeric ester 26b was synthesized by the
same synthetic strategy as illustrated in Scheme 4, but using
the chiral aldehyde 16b as the starting material. The 15S/
15R stereoselectivity of o-chain elongation was first evaluated
at this step. Thus, both compounds 26a and 26b were purified
by silica gel flash chromatography and identified by spectro-
scopic and chromatographic methods. The HPLC and LC-MS
recorded for the ester 26a confirmed the presence of the
(5E,15S)-isomer at the level of 8.41%, which was comparable
with the content of (5E)-isomer in the starting phenylsulfone
(5Z)-15. Fortunately, this kind of contamination leads to
the (5E,15S)-10c diastereoisomer of bimatoprost in the final
steps of the synthesis, which is easily removable from the final
product 10a by recrystallization or preparative HPLC. More
importantly, the content of the epimeric impurity (15R)-26b
in the prostaglandin precursor 26a was scarcely more than a
trace (0.17%). Considering the high enantiomeric purity of the
alcohol 24a (24a:24b, 99.84%:0.16%), the chromatographic
data for the ester 26a (26a:26b, 99.81%:0.19%) confirmed
that the bulky tert-bytyldimethylsilyl group made a very good
protection for the aldehyde a-proton preventing the o-chain
synthon 16a from racemization. Following the final steps of
the synthesis illustrated in Scheme 4, the ester contamination
26b leads to the (15R)-10b epimer of bimatoprost, only a trace
amount of which in the final product 10a is readily separable by
recrystallization or preparative HPLC. Since the olefinic diaster-
eoisomers 26 proved to be hardly separable at this juncture, we
found that better results could be obtained when separation
was performed on the final and less polar product 10a.
(15R)-10b, far more difficult to remove, in the final product
10a was scarcely more than a trace (0.12%). The physicochem-
ical properties of 10a and 10b are very similar, so it is not a
straightforward task to discern between these two compounds
by NMR spectroscopy, optical rotation or even by analytical
1
HPLC. The H and 13C NMR chemical shifts determined for
epimers 10a and 10b are almost identical, especially in peak
multiplicity and chemical shifts given by o-chain protons. The
two protons of the o-chain CH2-16 group in the bimatoprost
(10a) are seen as two multiplets at 1.78 and 1.90 ppm, whereas
in the 15-epi isomer they give one multiplet at 1.84 ppm. The
H-13 and H-14 protons of the o-chain trans-double bond in
the bimatoprost (10a) are seen as two doublet of doublets
at 5.47 and 5.59ppm; as well in 15-epi isomer 10b they give
two doublet of doublets at 5.22 and 5.61ppm. The 1H and 13C
NMR chemical shifts and peak multiplicity determined for the
o-chain H-15 proton and C-15 carbon in the bimatoprost 10a
(4.08 and 72.25 ppm) are highly comparable with those of the
15-epi isomer 10b (4.10 and 71.51 ppm). Considering very
similar physicochemical properties of epimers 10a and 10b,
recrystallization or preparative HPLC can only be applied to
remove trace amounts of (15R)-10b diastereoisomer. Thus,
the crude prostamide was further recrystallized from the
mixture of tert-butyl methyl ether/ethyl acetate to give a
pharmaceutical grade bimatoprost 10a in 73.7% yield from
the pentahydroxy ester 26a.
Synthesis of the Novel Prostaglandin-Related Amide (+)-(5Z)-
18, the New Synthetic Impurity of Bimatoprost (+)-(15S)-10a
The Julia–Lythgoe olefination entails one main drawback in
the step of formation of the carbon–carbon bond with a-sulfonyl
carbanion and carbonyl compounds. Since the hydrogen on the
carbon bearing the sulfonyl group is highly acidic, in some
cases the addition reaction of a-sulfonyl carbanion to carbonyl
compounds is reversible.29 The Julia–Lythgoe olefination of
the phenylsulfone 15 with the aldehyde 16a occurred under
mild conditions and gave good yields of stereodefined trans-
13,14 alkene 26a. The reaction stopped at ~95% conversion of
phenylsulfone 15 to b-hydroxy sulfones 17a despite using
the aldehyde 16a in excess under highly anhydrous and air-
free conditions. The equilibrium was strongly shifted toward
the products 17a; however, up to 5% of phenylsulfone 15
was left unreacted and subsequently went into side reactions
leading to the novel prostaglandin-related amide (+)-(5Z)-18
(Scheme 5). Finally, the crude bimatoprost (10a) was contam-
inated with the new impurity 18, which was readily separable
by silica gel flash chromatography.
The amide 18 was synthesized and identified to establish the
full profile of impurities accompanying our novel synthesis of
the active pharmaceutical ingredient 10a. Thus, phenylsulfone
15 (83.1% de) was subjected to the reductive desulfonylation
followed by deprotection of triethylsilyl groups with TBAF solu-
tion and subsequent hydrolytic opening of the methyl-OBO
protecting group with citric acid solution to afford the stable 2,2-
bis(hydroxymethyl)propyl ester (+)-(5Z)-28 (90.7% yield from
15). Amidation of the ester 28 with ethylamine aqueous solu-
tion gave the prostaglandin-related amide 18. The LC-MS con-
firmed the presence of (5E)-18 isomer in the amount compara-
ble with that of (5E)-15 isomer in the starting phenylsulfone
15. Similarly to the bimatoprost (10a) and its a-chain 5,6-trans
diastereoisomer 10c, the allylic carbons C-4 and C-7 of (5E)-18
isomer are seen in the 13C NMR spectrum as two singlets at a
lower field (31.87 and 31.96 ppm) compared with the same
Chirality DOI 10.1002/chir
Most synthetic procedures used to attain bimatoprost
(10a) proceed via bimatoprost acid (9a) and its esterification
to the corresponding isopropyl ester, subsequently followed
by amidation with ethylamine.10,14e One of the main advan-
tages of our novel convergent strategy is direct amidation of
the complex pentahydroxy ester 26a, without the need of
bimatoprost acid (9a) and its simple alkyl esters preparation.
Thus, the ester 26a was successfully amidated to bimatoprost
10a with ethylamine aqueous solution. The diastereoisomeric
impurity (+)-(15R)-10b of bimatoprost was synthesized by the
same synthetic strategy as illustrated in Scheme 4, but using
the 2,2-bis(hydroxymethyl)propyl ester 26b as the starting
material. The crude bimatoprost (10a) and its (15R)-epi isomer
(10b) were purified by silica gel flash chromatography and
identified by spectroscopic and chromatographic methods.
The HPLC and LC-MS recorded for the bimatoprost (10a)
confirmed the presence of the (5E,15S)-isomer (10c) (8.38%)
formed from the (5E)-isomer present in the starting phenylsul-
fone (5Z)-15. The NMR spectra of bimatoprost 10a and its 5,6-
trans diastereoisomer 10c are hardly discerned. However, the
a-chain allylic carbons C-4 and C-7 of 10c are seen in the 13
C
NMR spectrum as two singlets at a lower field (31.93 and
31.16 ppm) compared with the same carbons of 10a (26.70
and 25.38 ppm), therefore the C-4 and C-7 carbons of 10c
are more effectively deshielded by the trans-double bond of
the a-chain. More importantly, the content of epimeric impurity

178
DAMS ET AL.
Scheme 5. Synthesis of the new impurity (+)-(5Z)-18 of bimatoprost. Reagents and conditions used: (1) Na/Hg, Na2HPO4, MeOH, 0 ꢁC, 30 min, then 3 h at r.t.;
(2) n-Bu4NF, 1 h; (3) HOOCC(OH)(CH2COOH)2, H2O, 15 min, 90.7% yield of 28 from phenylsulfone 15; (4) NH2Et, MeOH, 72 h, 87.8% yield of 18.
Toris CB, Gabelt BT, Kaufman PL. Update on the mechanism of action
of topical prostaglandins for intraocular pressure reduction. Surv Ophthal-
mol 2008;53:S107 ꢀ 120; (f) Stjernschantz JW. From PGF2a-isopropyl ester
carbons of (5Z)-18 isomer (26.36 and 26.61 ppm); therefore the
C-4 and C-7 carbons of (5E)-18 are more effectively deshielded
by the trans-double bond. The (5Z)-18 diastereoisomer can be
easily purified from its (5E)-18 diastereoisomer by preparative
HPLC on silica gel stationary phases.
to latanoprost: a review of the development of Xalatan. Invest Ophthalmol
Vis Sci 2001;42:1134 ꢀ 1145; (g) Brooks TC, Burlingame M, Burk M,
Tortorice K, Dong D, Glassman PA, Lynch JA, Good CB, Rickman HS,
Claborn M, Appleman D, Jones W. A prostaglandin analogue for the
treatment of glaucoma. Formulary 2009;44:322 ꢀ 328; (h) Waugh J, Jarvis
B. Travoprost. Drugs Aging 2002;19:465 ꢀ 471; (i) Burk RM, Woodward
DF. Bimatoprost, a novel efficacious hypotensive drug now recognized
as a member of a new class of agents called prostamides. Drug Dev Res
2007;68:147 ꢀ 155; (j) Sharif NA, Klimko P. Update and commentary on
the pro-drug bimatoprost and a putative “prostamide receptor”. Expert
Rev Ophthalmol 2009;45:477 ꢀ 489; (k) Easthope SE, Perry CM. Topical
bimatoprost. A review of its use in open-angle glaucoma and ocular
hypertension. Drugs Aging 2002;19:231 ꢀ 248; (l) Woodward DF, Phelps
RL, Krauss AH-P, Weber A, Short B, Chen J, Liang Y, Wheeler LA.
CONCLUSIONS
The 17-phenyl PGF2a analogue bimatoprost (10a) has
potent topical ocular activity. A novel convergent synthesis
of 13,14-en-15-ol prostamideF2a analogues with the desired
C-15 asymmetric center configuration was developed employ-
ing Julia–Lythgoe olefination of the structurally advanced
phenylsulfone (+)-(5Z)-15 with an enantiomerically pure alde-
hyde o-chain synthon (–)-(S)-16a. Subsequent hydrolysis of
protecting groups and final amidation of the diol 26a yielded
bimatoprost (10a). The novel aldehyde o-chain synthon (–)-
(S)-16a used for the construction of the o-chain of bimato-
prost (10a) was envisaged to contain a stereogenic center
at the carbon C-2 corresponding to C-15 in the target mole-
cule, with chirality corresponding to the “15S” configuration
in the final product 10a. This ensured the absence of an
appreciable quantity of the undesired (+)-(15R)-10b diaste-
reoisomer in the final prostaglandin analogue 10a. The main
advantage of the new method is the preparation of high-purity
bimatoprost (10a). The novel convergent strategy allows the
synthesis of a whole series of trans-13,14-en-15-ol prostami-
deF2 analogues from a common and structurally advanced
prostaglandin intermediate (+)-(5Z)-15.
Bimatoprost:
a novel antiglaucoma agent. Cardiovas Drug Rev
2004;22:103 ꢀ 120; (m) Sit AJ, Asrani S. Effects of medications and
surgery on intraocular pressure fluctuation. Surv Ophthalmol 2008;53:S45ꢀ
S55; (n) Cantor LB. Bimatoprost: a member of a new class of agents, the
prostamides, for glaucoma management. Exp Opin Invest Drugs
2001;10:721 ꢀ 731; (o) Perry CM, McGavin JK, Culy CR, Ibbotson T. Latano-
prost. An update of its use in glaucoma and ocular hypertension. Drugs
Aging 2003;20:597ꢀ630.
2. For a review of Corey synthesis of PGF2a, see: (a) Corey EJ, Weinshenker
NM, Schaaf TK, Huber W. Stereo-controlled synthesis of prostaglandins
F2a and E2 (dl). J Am Chem Soc 1969;91:5675–5677; (b) Corey EJ, Schaaf
TK, Huber W, Koelliker U, Weinshenker NM. Total synthesis of prosta-
glandins F2a and E2 as the naturally occurring forms. J Am Chem Soc
1970;92:397–398; (c) Corey EJ, Noyori R, Schaaf TK. Total synthesis of
prostaglandins F1a, E1, F2a, and E2 (natural forms) from a common synthetic
intermediate. J Am Chem Soc 1970;92:2586–2587; (d) Corey EJ. The logic of
chemical synthesis: multistep synthesis of complex carbogenic molecules.
Angew Chem Int Ed Engl 1991;30:455–465; (e) Corey EJ, Cheng X-M. The
logic of chemical synthesis. New York: Wiley; 1989. 255–256.
3. Resul B, Stjernschantz J, No K, Liljebris C, Selén G, Astin M, Karlsson M,
Bito LZ. Phenyl-substituted prostaglandins: potent and selective antiglau-
coma agents. J Med Chem 1993;36:243–248.
4. Selliah R, Dantanarayana A, Haggard K, Egan J, Do EU, May JA. Synthesis
of [phenyl-2-3H]-travoprost: isopropyl ester prodrug of a selective prostaglan-
din FP receptor agonist. J Labelled Cpd Radiopharm 2001;44:173–183.
ACKNOWLEDGMENTS
Financial support was provided by the European Union’s
European Regional Development Fund (ERDF) under the Inno-
vative Economy Operational Programme (UDA-POIG.01.03.01-
14-068/08-00).
5. Collins PW, Djuric SW. Synthesis of therapeutically useful prostaglandin
and prostacyclin analogs. Chem Rev 1993;93:1533–1564.
6. Liljebris C, Selén G, Resul B, Stjernschantz J, Hacksell U. Derivatives
of 17-phenyl-18,19,20-trinorprostaglandin F2a isopropyl ester: potential
antiglaucoma agents. J Med Chem 1995;38:289–304.
7. Klimko P, Hellberg M, McLaughlin M, Sharif N, Severns B, Williams G,
Haggard K, Liao J. 15-Fluoro prostaglandin FP agonists: a new class of
topical ocular hypotensives. Bioorg Med Chem 2004;12:3451–3469.
8. Matsumura Y, Mori N, Nakano T, Sasakura H, Matsugi T, Hara H, Morizawa
Y. Synthesis of the highly potent prostanoid FP receptor agonist, AFP-168: a
novel 15-deoxy-15,15-difluoroprostaglandin F2a derivative. Tetrahedron Lett
2004;45:1527–1529.
LITERATURE CITED
1. (a) Ishida A, Odani-Kawabata N, Shimazaki A, Hara H. Prostanoids in the
therapy of glaucoma. Cardiovas Drug Rev 2006;24:1 ꢀ 10; (b) Bean GW,
Camras CB. Commercially available prostaglandin analogs for the reduc-
tion of intraocular pressure: similarities and differences. Surv Ophthalmol
2008;53:S69 ꢀ 84; (c) Eisenberg DL, Toris CB, Camras CB. Bimatoprost
and travoprost: a review of recent studies of two new antiglaucoma drugs.
Surv Ophthalmol 2002;47:S105 ꢀ 115; (d) Alm A Prostaglandin derivatives
as ocular hypotensive agents. Prog Retin Eye Res 1998;17:291 ꢀ 312; (e)
Chirality DOI 10.1002/chir

SYNTHESIS OF THERAPEUTICALLY USEFUL PGF2 ANALOGUES
179
9. Bowler J, Crossley NS. Cyclopentane derivatives. GB 1974;1:350,971.
Bimatoprost versus latanoprost in lowering intraocular pressure in glaucoma
and ocular hypertension: results from parallel-group comparison trials. Adv
Ther 2004;21:247ꢀ 262; (g) Vand der Valk R, Webers CA, Schouten JS,
Zeegers MP, Hendrikse F, Prins MH. Intraocular pressure-lowering effects
of all commonly used glaucoma drugs. Ophthalmol. 2005;112:1177ꢀ1185;
(h) Holmstrom S, Buchholz P, Walt J, Wickstrøm J, Aagren M, Curr Med
Res Opin 2005;21:1875ꢀ 1883; (i) Cantor LB, Hoop J, Morgan L, WuDunn
D, Catoira Y, Intraocular pressure-lowering efficacy of bimatoprost 0.03%
and travoprost 0.04% in patients with glaucoma or ocular hypertension. Br
J Ophthalmol 2006;90:1370ꢀ 1373; (j) Denis P, Lafuma A, Khoshnood B,
Mimaud V, Berdeaux G. A meta-analysis of topical prostaglandin analogues
intraocular pressure lowering in glaucoma therapy. Curr Med Res Opin
2007;23:601ꢀ608.
10. (a) Gutman A, Nisnevich M, Etinger I, Zaltzman L, Yudovich B, Pertsikov
B. Tishin B. Process for the preparation of prostaglandin derivatives. US
2007;7:166,730; (b) Woodward DF, Andres SW, Burk RM, Garst ME.
Non-acidic cyclopentane heptanoic acid, 2-cycloalkyl or arylalkyl deriva-
tives as therapeutic agents. WO 94/06433, 31.03.1994.
11. Henegar KE. Process and intermediates to prepare latanoprost. US2003/
0187071 A1, 2.10.2003.
12. Hwang S-W, Adiyaman M, Khanapure SP, Rokach J. Total synthesis of
12-epi-PGF2a. Tetrahedron Lett 1996;37:779ꢀ782.
13. Kim S, Bellone S, Maxey KM, Powell WS, Lee G-J, Rokach J. Synthesis of
15R-PGD2: a potential DP2 receptor agonist. Bioorg Med Chem Lett
2005;15:1873ꢀ1876.
17. (a) Julia M, Paris J-M. Syntheses a l’aide de sulfones V(+)—methode de synth-
ese generale de doubles liaisone. Tetrahedron Lett 1973;14:4833 ꢀ 4836; (b)
Kocieński PJ, Lythgoe B, Ruston S. Scope and stereochemistry of an olefin
14. (a) Corey EJ, Albonico SM, Koelliker U, Schaaf TK, Varma RK. New
reagents for stereoselective carbonyl reduction. An improved synthetic
route to the primary prostaglandins. J Am Chem Soc 1971;93:1491–1493;
(b) Corey EJ, Becker KB, Varma RK. Efficient generation of the 15S
configuration of prostaglandin synthesis. Attractive interactions in
synthesis from b-hydroxysulphones.
J Chem Soc Perkin Trans 1
1978;829 ꢀ 834; (c) Kocieński PJ, Lythgoe B, Waterhouse I. The influence
of chain-branching on the steric outcome of some olefin-forming reactions.
J Chem Soc Perkin Trans 1 1980;1045ꢀ 1050; (d) Kocieński PJ. Recent
sulphone-based olefination reactions. Phosphorus and Sulphur and the
Related Elements 1985;24:97ꢀ 127; (e) Blakemore PR, Cole WJ, Kocieński
PJ, Morley A. A stereoselective synthesis of trans-1,2-disubstituted alkenes
based on the condensation of aldehydes with metallated 1-phenyl-1H-tetra-
zol-5-yl sulfones. Synlett 1998;1:26ꢀ28.
stereochemical control of carbonyl reduction.
J Am Chem Soc
1972;94:8616–8618; (c) Corey EJ, Bakshi RK, Shibata S, Chen C-P, Singh
VK. A stable and easily prepared catalyst for the enantioselective reduc-
tion of ketones. Applications to multistep syntheses. J Am Chem Soc
1987;109:7925–7926; (d) Aswathanarayanappa C, Bheemappa E, Bodke
YD. Diastereoselective reduction of the enone intermediate of travoprost.
Org Process Res Dev 2011;15:1085–1087; (e) Zanoni G, D’Alfonso A,
Porta A, Feliciani L, Nolan SP, Vidari G. The Meyer-Schuster rearrange-
ment: a new synthetic strategy for prostaglandins and their drug analogs,
bimatoprost and latanoprost. Tetrahedron 2010;66:7472–7478; (f) Corey
EJ. New enantioselective routes to biologically interesting compounds. Pure
18. Blakemore PR, Ho DKH, Nap WM. Ethyl (benzothiol-2-ylsulfonyl)acetate:
a new reagent for the stereoselective synthesis of a,b-unsaturated esters
from aldehydes. Org Biomol Chem 2005;3:1365ꢀ1368.
19. Adjé N, Vogeleisen F, Uguen D. Improved conditions for the Kiliani-
Fischer synthesis. Tetrahedron Lett 1996;37:5893ꢀ5896.
ꢀ
Appl Chem 1990;62:1209ꢀ 1216; (g) Obadalová I, Pilarcík T, Slavíková M,
20. Belotti D, Andreatta G, Pradaux F, BouzBouz S, Cossy J. Degradation of alde-
ꢀ
Hájícek J. Synthesis of latanoprost diastereoisomers. Chirality 2005;17:
hydes to one carbon lower homologs. Tetrahedron Lett 2003;14:3613ꢀ3615.
S109 ꢀ S113; (h) Hutton J, Senior M, Wright NCA. The reduction of pros-
taglandin intermediates using alumina hydride reagents. Synth Commun
1979;9:799 ꢀ 808; (i) Hutton J Di-isobornyloxyaluminium isopropoxide—
21. Mitsunobu O. The use of diethyl azodicarboxylate and triphenylpho-
sphine in synthesis and transformation of natural products. Synthesis
1981;1:1ꢀ28.
a
new highly selective Meerwein-Ponndorf-Verley reagent. Synth
Commun 1979;6:483ꢀ 486; (j) Node M, Nishide K, Shigeta Y, Shiraki H,
Obata K. A novel tandem Michael addition/Meerweinꢀ Ponndorf ꢀ Verley
reduction: asymmetric reduction of acyclic a,b-unsaturated ketones using a
chiral mercapto alcohol. J Am Chem Soc 2000;122:1927ꢀ 1936; (k) Noyori
R, Tomino I, Yamada M, Nishizawa M. Asymmetric synthesis via axially
dissymmetric molecules. 7. Synthetic applications of the enantioselective
reduction by binaphthol-modified lithium aluminum hydride reagents J
Am Chem Soc 1984;106:6717ꢀ6725.
22. Wang F-D, Yue J-M. A total synthesis of (+)- and (ꢀ)-dihydrokavin with
a sonochemical blaise reaction as the key step. Eur J Org Chem
2005;12:2575ꢀ2579.
23. Martinelli MJ, Nayyar NK, Moher ED, Dhokte UP, Pawlak JM, Vaidyanathan
R. Dibutyltin oxide catalyzed selective sulfonylation of a-chelatable primary
alcohols. Org Lett 1999;1:447ꢀ450.
24. (a) Kornblum N, Jones WJ, Anderson GJ. A new and selective method
of oxidation. The conversion of alkyl halides and alkyl tosylates to alde-
hydes. J Am Chem Soc 1959;81:4113 ꢀ 4144; (b) Kornblum N, Powers
JW, Anderson GJ, Jones WJ, Larson HO, Levand O, Weaver WM. A new
and selective method of oxidation. J Am Chem Soc 1957;79:6562.
15. (a) Martynow JG, Jówik J, Szelejewski W, Achmatowicz O, Kutner A,
Winiewski K, Winiarski J, Zegrocka-Stendel O, Gołbiewski P. A new
synthetic approach to high-purity (15R)-latanoprost. Eur J Org Chem
2007;4:689–703.
25. Parikh JR, Doering WE. Sulfur trioxide in the oxidation of alcohols by
16. (a) Higginbotham EJ, Schuman JS, Goldberg I, Gross RL, VanDenburgh
AM, Chen K, Whitcup SM. One-year, randomized study comparing
bimatoprost and timolol in glaucoma and ocular hypertension. Arch
Ophthalmol 2002;120:1286 ꢀ 1293; (b) Cohen JS, Gross RL, Cheetham
JK, VanDenburgh AM, Bernstein P, Whitcup SM. Two-year double-
masked comparison of bimatoprost with timolol in patients with glaucoma
or ocular hypertension. Surv Ophthalmol 2004;49:S45ꢀ S52; (c) Williams
RD, Cohen JS, Gross RL, Liu CC, Safyan E, Batoosingh AL. Long-term effi-
cacy and safety of bimatoprost for intraocular pressure lowering in glaucoma
or ocular hypertension: year 4. Br J Ophthalmol 2008;92:1387 ꢀ 1392; (d)
Aptel F, Cucherat M, Denis P. Efficacy and tolerability of prostaglandin
dimethyl sulfoxide. J Am Chem Soc 1967;89:5505ꢀ5507.
26. Dess DB, Martin JC. Readily accessible 12-I-51 oxidant for the conversion
of primary and secondary alcohols to aldehydes and ketones. J Org Chem
1983;48:4155ꢀ4156.
27. (a) Trost BM. Chemical chameleons. Organosulfones as synthetic build-
ing blocks. Bull Chem Soc Jpn 1988;61:107 ꢀ 124; (b) Simpkins NS.
The chemistry of vinyl sulphones. Tetrahedron 1990;46:6951 ꢀ 6984; (c)
Baldwin IR, Whitby RJ. Total synthesis of (ꢅ)-acetoxyodontoschismenol
using zirconium chemistry. Chem Commun 2003;2786ꢀ2786.
28. Trost BM, Arndt HC, Strege PE, Verhoeven TR. Desulfonylation of aryl
analogs:
a meta-analysis of randomized controlled clinical trials. J
alkyl sulfones. Tetrahedron Lett 1976;17:3477ꢀ3478.
Glaucoma 2008;17:667 ꢀ 673; (e) Noecker RS, Dirks MS, Choplin NT,
Bernstein P, Batoosingh AL, Whitcup SM. A six-month randomized
clinical trial comparing the intraocular pressure-lowering efficacy of bimato-
prost and latanoprost in patients with ocular hypertension or glaucoma. Am J
Ophthalmol 2003;135:55 ꢀ 63; (f) Simmons ST, Dirks MS, Noecker RJ.
29. (a) Marko IE, Murphy F, Dolan S. Efficient preparation of trisubstituted
alkenes using the Julia-Lythgoe olefination of ketones. On the key-role
of SmI2 in the reductive elimination step. Tetrahedron Lett 1996;37:2089
ꢀ 2092; (b) Hart DJ, Wu W-L. Observations regarding the first step of the
Julia-Lythgoe olefin synthesis. Tetrahedron Lett 1996;37:5283ꢀ5286.
Chirality DOI 10.1002/chir