Enantioselective hydrogenation of β-ketoesters using a MeO-PEG-supported Biphep ligand under atmospheric pressure: A practical synthesis of (S)-fluoxetine

Article abstract of DOI:10.1055/s-2006-949650

The preparation of a novel chiral 2,2′-bis(MeO-PEG-supported)-6, 6′-bis(diphenylphosphanyl)biphenyl (MeO-PEG-Biphep) ligand is described. The derived ruthenium complex catalyzes the hydrogenation of β-ketoesters in up to 99% yield and 99% ee under atmospheric pressure. The accelerating effects exerted by the PEG linkage are dramatic when compared to the unsupported analogue, MeO-Biphep-RuBr₂. Furthermore, the catalyst can be recovered easily and the recycled catalysts were shown to maintain their efficiency in two consecutive runs, albeit with declining activity. One of the products, (S)-ethyl-3-hydroxy-3-phenylpropanoate, is useful in the preparation of (S)-fluoxetine. Georg Thieme Verlag Stuttgart.

Full text of DOI:10.1055/s-2006-949650

LETTER
2395
Enantioselective Hydrogenation of b-Ketoesters Using a MeO-PEG-Supported
Biphep Ligand under Atmospheric Pressure: A Practical Synthesis of
(S)-Fluoxetine
E
nantioselective
i
Hydro
t
genati
i
on of
b
n
-Ketoester
s
g Chai, Huansheng Chen, Zhiming Li, Quanrui Wang,* Fenggang Tao
Department of Chemistry, Fudan University, 200433 Shanghai, P. R. of China
Fax +86(21)65641740; E-mail: qrwang@fudan.edu.cn
Received 27 March 2006
the expensive chiral catalyst. An attempt to surmount this
Abstract: The preparation of a novel chiral 2,2¢-bis(MeO-PEG-
supported)-6,6¢-bis(diphenylphosphanyl)biphenyl (MeO-PEG-Bi-
phep) ligand is described. The derived ruthenium complex catalyzes
the hydrogenation of b-ketoesters in up to 99% yield and 99% ee
problem has been to ‘heterogenize’ a homogeneous cata-
lyst by anchoring the catalyst onto an insoluble polymer
support.9 Unfortunately, it is difficult to characterize these
under atmospheric pressure. The accelerating effects exerted by the insoluble polymer-supported catalysts and, crucially, they
PEG linkage are dramatic when compared to the unsupported ana-
often lead to a reduction in enantioselectivity due to the
logue, MeO-Biphep-RuBr2. Furthermore, the catalyst can be recov-
restrictions of the polymer matrix, which results in limited
ered easily and the recycled catalysts were shown to maintain their
mobility and accessibility of the active sites. To circum-
efficiency in two consecutive runs, albeit with declining activity.
vent these limitations, another different type of polymeric
One of the products, (S)-ethyl-3-hydroxy-3-phenylpropanoate, is
chiral catalyst has been developed using a soluble poly-
mer support and has gained increasing popularity.10 This
type of catalyst is soluble in organic solvents, thereby
allowing the reaction to proceed homogeneously, a facet
that often improves activity and selectivity. The ligand
can be isolated simply by precipitation on addition of a
suitable solvent, followed by filtration. Thus, this ap-
proach may exploit the advantages of both heterogeneous
and homogeneous catalysts. Herein we wish to report the
preparation of a novel soluble chiral MeO-PEG-supported
ligand and its use in Ru(II)-mediated enantioselective
hydrogenation of several b-ketoesters.11
useful in the preparation of (S)-fluoxetine.
Key words: asymmetric synthesis, supported catalysis, hydrogena-
tion, biphenyl ligands, fluoxetine
Enantioselective reduction of prochiral ketones allows an
extraordinarily efficient synthesis of optically active sec-
ondary alcohols, which are valuable synthetic intermedi-
ates.1 In this regard, several important antidepressants,
such as fluoxetine, tomoxetine, nisoxetine, and duloxetine
having a 3-aryloxy-3-arylpropylamine motif can be con-
structed by using the respective b-amino alcohol as a key
precursor. In particular, fluoxetine (ProzacTM, Eli Lilly
Co.) has been widely used as an antidepressant that shows
little effect on noradrenergic or dopaminergic systems.2
Although fluoxetine hydrochloride is currently used ther-
apeutically as the racemate, related studies demonstrated
that the two enantiomers have quite different pharmaco-
logical potencies and metabolites.3,4 As a consequence,
there has been an increased effort to devise an enantiose-
lective synthesis of the individual enantiomers of fluoxet-
ine.5 In all these syntheses, the major strategy hinges on
the production of suitable chiral alcohol intermediates.
The axially dissymmetric diphosphine (R)-MeO-Biphep
(1), which was introduced by Joseph Foricher in 1991,
was prepared according to the reported method using 3-
bromoanisole as the starting material.12 Demethylation of
1 was effected with boron tribromide in dichloromethane
solution to provide (R)-HO-Biphep (2) in 85% yield.13
Compound 2 was anchored to the PEG linkage by reaction
with MeO-PEG-OMs (Mn = 2100) in the presence of
Cs2CO3 in DMF to furnish MeO-PEG-(R)-Biphep (3) in
up to 96% yield (1H NMR).14 According to a reported pro-
cedure,8 the ruthenium catalyst 4 was conveniently pre-
pared in situ by reaction of ligand 3 and the bis(2-
methylallyl)cycloocta-1,5-diene ruthenium(II) complex
in the presence of HBr in acetone (Scheme 1).15
Asymmetric hydrogenation of prochiral ketones catalyzed
by chiral ruthenium–atropisomeric diphosphine complex-
es has proved to be one of the most fundamental strate-
gies.6,7 Genêt has disclosed that b-ketoesters could be
hydrogenated under atmospheric pressure with the homo-
geneous catalyst generated in situ from (6,6¢-dimethoxy-
biphenyl-2,2¢-diyl)bis(diphenylphosphine) (MeO-Biphep)
in combination with (COD)Ru[CH2(CH3)2]2 in the pres-
ence of HBr.8 However, these homogeneous catalytic sys-
tems suffered from a difficulty in separating and recycling
Poly(ethylene glycol) monomethyl ether (MeO-PEG) was
the solid support of choice because it offers the following
advantages. First, MeO-PEG is an inexpensive and com-
mercially obtainable material, which can be readily con-
verted to the methanesulfonate form (MeO-PEG-OMs) on
treatment with methanesulfonyl chloride. Secondly, it has
a broad solubility profile in organic solvents (it is soluble
in methanol, tetrahydrofuran, dichloromethane, toluene,
N,N-dimethylformamide, acetonitrile and water but insol-
uble in diethyl ether, tert-butyl methyl ether, cold isopro-
pyl alcohol and cold ethanol).16 This allows an easy
SYNLETT 2006, No. 15, pp 2395–2398
1
8
.0
9
.2
0
0
6
Advanced online publication: 08.09.2006
DOI: 10.1055/s-2006-949650; Art ID: W06406ST
© Georg Thieme Verlag Stuttgart · New York

2396
L. Chai et al.
LETTER
purification of the product and the possibility to recycle 5a using (R)-MeO-Biphep-RuBr2, 18 hours were required
the catalyst on addition of a suitable solvent.
under comparable conditions (Table 2, entry 3). This
demonstrated a dramatically increased activity of (R)-4.
To further test the effects of the PEG backbone in the cat-
alyst (R)-4 on activity, we carried out the reaction using
similar conditions with (R)-MeO-Biphep-RuBr2, but we
added two equivalents of poly(ethylene glycol) monome-
thyl ether or its mesylate to the reaction mixture. Over
three hours, the conversion rates of 5a were only 46% and
20%, respectively (Table 2, entries 4 and 5). Thus, the
function of the PEG motif in enhancing the activity of the
supported catalyst seems not to be only as a result of phase
transfer catalysis. Presumably, the attachment of the cata-
lyst to the PEG polymer hampers dimer formation of the
catalyst, thereby making the catalyst more efficient.
The 31P NMR spectra of the MeO-PEG supported MeO-
Biphep ligand 3 was similar to those of the monomeric
systems 1 and 2 (Table 1). Only one 31P NMR peak was
observed for 3, suggesting that both hydroxyl groups in 2
had been etherified with the polymer backbone. It was cal-
culated that the immobilized ligand 3 has a loading of 0.21
mmol ligand/g support.
Table 1 The 31P NMR Spectra of the Supported Ligand and its Par-
ent Systems
Ligand
MeO–Biphep HO–Biphep MeO–PEG–Biphep
(1)
(2)
(3)
31P NMR (ppm)
–14.00
–13.08
–14.05
(R)-4
O
O
OH
O
R2
R2
R1
O
R1
O
H2 (1 bar)
6
5
a: R1 = Ph, R2 = Et
b: R1 = Me, R2 = Et
c: R1 = Et, R2 = Me
d: R1 = i-Pr, R2 = Me
BBr3, CH2Cl2
MeO
MeO
PPh2
PPh2
HO
HO
PPh2
PPh2
–78 °C to r.t.,
overnight
e: R1 = 4-MeOC6H4, R2 = Me
85%
(R)-1
(R)-2
Scheme 2
MeO-PEG-OMs (MW = 2100)
Cs2CO3, DMF, 65 °C, 18 h
One of the advantages of polymer-supported catalysts is
their easy recovery and reuse. Thus, the longevity of the
supported catalyst (R)-4 was verified using 5a as the sub-
strate. At the end of the reaction, the catalyst was precipi-
tated by adding cold Et2O (0 °C) and recovered by
filtration. In two consecutive reactions, the catalyst af-
forded conversion rates of 90% (in 36 hours) and 46% (in
60 hours), both with an ee of 95% (Table 2, entries 6 and
7). The practically colorless solutions indicate that little or
no catalyst leached during the reaction and work-up (ICP
analysis found no leached ruthenium in the ether filtrate
down to the limit of detection). Unexpectedly, the recov-
ered catalyst gave a slightly better enantioselectivity but
was less active, plausibly due to degradation of the PEG-
MeOPEG O
MeOPEG O
PPh2
PPh2
96%
(R)-3
(R)-4
(COD)Ru(2-methylallyl)2,
HBr, MeOH, acetone
Ph
Ph
Ph
MeOPEG O
P
RuBr2
r.t., 30 min
MeOPEG
O
P
Ph
Scheme 1 The synthesis of (R)-MeO-PEG-supported Biphep-
RuBr2 (R)-4
As a model reaction to test the efficacy of the supported containing ligand. The asymmetric hydrogenation was
catalyst, enantioselective hydrogenation of ethyl-3-oxo-3- then extended to other b-ketoester substrates, and an inter-
phenylpropanoate (5a) catalyzed by the in situ generated esting set of results was achieved. As can be seen from
complex (R)-4 was performed under atmospheric pressure Table 2, the immobilized catalyst (R)-4 exhibited excel-
at 55 °C for 3 hours (Scheme 2). Ethanol is known to be lent activities and enantioselectivities for the asymmetric
an appropriate solvent in terms of providing high enantio- hydrogenation of a variety of aliphatic b-ketoesters, in-
selectivity in hydrogenation reactions7 and since the sup- cluding ethyl acetoacetate (5b), methyl propionylacetate
ported catalyst (R)-4 is soluble in hot ethanol, most of the (5c) and methyl isobutyrylacetate (5d) under an atmo-
reactions were carried out in ethanol (unless otherwise spheric pressure of H2 (Table 2, entries 8–14). For exam-
stated). The results are summarized in Table 2. We found ple, with ethyl acetoacetate, complete conversion was
that, in the presence of 2 mol% of the catalyst (R)-4 under observed at 55 °C in two hours, with 98% ee (Table 2, en-
an atmospheric pressure of H2, hydrogenation of 5a in eth- try 8). After the first run, the recovered catalyst showed
anol provided (S)-ethyl-3-hydroxy-3-phenylpropionate good activity and enantioselectivity to afford 93% of iso-
(6a) in 98% isolated yield and 92% ee in three hours lated product 6b with an excellent ee of 99% (Table 2, en-
(Table 2, entry 1). The parent catalyst (R)-MeO-Biphep- try 9). As in the case of 5a, the accelerating effect of the
RuBr2 gave only 21% conversion of 5a under similar con- PEG-linkage in the catalyst (R)-4 was dramatic as using
ditions, although the ee was slightly higher at 94% (R)-MeO-Biphep-RuBr2 required 24 hours to achieve an
(Table 2, entry 2). In order to achieve 92% conversion of almost complete conversion with a 99% ee (Table 2, entry
Synlett 2006, No. 15, 2395–2398 © Thieme Stuttgart · New York

LETTER
Enantioselective Hydrogenation of b-Ketoesters
2397
Table 2 Asymmetric Hydrogenation of b-Ketoesters at Atmospheric Pressure17
Entry
1
Substrate
5a
Cat.b
Temp (°C)
Time (h)c
Yield (%)d
98
ee (%)e
(R)-4
55
55
55
55
55
60
60
55
60
55
50
55
55
58
35
3
3
92
94
94
2
5a
(R)-MeO-Biphep-RuBr2
21
3
5a
(R)-MeO-Biphep-RuBr2
18
3
92
4
5a
(R)-MeO-Biphep-RuBr2 + MeO-PEG-OH
46
5
5a
(R)-MeO-Biphep-RuBr2 + MeO-PEG-OMs
3
20
6
5a
(R)-4 (2nd use)
(R)-4 (3rd use)
(R)-4
36
60
2
90
95
95
7
5a
46f
100f
93
8
5b
98
9
5b
(R)-4 (2nd use)
(R)-MeO-Biphep-RuBr2
(R)-4h
36
24
2.5
30
3
99
10
11g
12g
13g
14g
15g
5b
98
99
5c
97
>99
99
5c
(R)-4 (2nd use)h
(R)-4
87
5d
96
>99.5
99
5d
(R)-4 (2nd use)
(R)-4
32
3.5
83
5e
52
85
a The hydrogenation reactions were carried out in ethanol (0.1 g cat./1 mL ethanol) unless otherwise stated.
b 2 mol% of catalyst.
c Not optimized.
d Isolated yield.
e Determined by HPLC on a Daciel Chiralcel OD column or chiral GC. The configuration of the products was determined by comparison of
optical-rotation data with literature data.
f GC conversion.
g Using methanol as solvent.
h 1 mol% was used.
10). Another experiment revealed that lowering the cata- conditions as described for 5a. However, we performed
lyst loading could give comparable results, as in the case this reaction at 35 °C in order to achieve better enantiose-
of the asymmetric reduction of methyl propionylacetate lectivity, and we used methanol as the solvent to avoid the
(5c). With the use of 1 mol% of catalyst (R)-4, the reaction etherification. The reaction of 5e with 2 mol% of the cat-
proceeded smoothly in methanol giving 97% isolated alyst (R)-4 proceeded somewhat sluggishly as compared
yield with nearly exclusive enantioselectivity after two with 5a, affording a 52% yield of the corresponding opti-
and a half hours (Table 2, entry 11). The good perfor- cally active 3-hydroxy ester 6e with an 85% ee, demon-
mance of the recovered, immobilized catalyst was further strating a decrease in enantioselectivity.
verified by the asymmetric hydrogenation of 5c and 5d
(Table 2, entries 12 and 14). Finally, the reaction of meth-
The reduced product 6a could be readily transformed to
the amide 7 in 70% yield on treatment with a 40% aqueous
yl-3-(4-methoxybenzoyl)acetate (5e) containing an elec-
solution of methylamine. Reduction of 7 with LiAlH4 fur-
tron-donating group was conducted under similar
nished the b-amino alcohol 8 in 90% yield. According to
OH
O
OH
MeNH2 (40% aq)
THF, 70 °C
OH
O
LiAlH4–THF
0–25 °C
Me
Me
OEt
N
N
H
H
90%
7
6
70%
8
CF3
4-CF3C6H4Cl, NaH, DMAc
90 °C; then (CO2H)2, Et2O
O
Me
(CO2H)2
N
H
90%
9
Scheme 3 Synthesis of (S)-fluoxetine oxalate 9
Synlett 2006, No. 15, 2395–2398 © Thieme Stuttgart · New York

2398
L. Chai et al.
LETTER
Longbottom, D. A.; Nesi, M.; Scott, J. S.; Storer, R. I.;
Taylor, S. J. J. Chem. Soc., Perkin Trans. 1 2000, 3815.
(10) (a) Toy, P. H.; Tanda, K. D. Acc. Chem. Res. 2000, 33, 546.
(b) Wentworth, P. Jr.; Janda, K. D. Chem. Commun. 1999,
1917. (c) Pu, L. Chem. Eur. J. 1999, 5, 2227.
a known procedure,5d compound 8 was converted to (S)-
fluoxetine, which was isolated as the oxalate salt 9 in 90%
yield with a 94% ee (Scheme 3).
In summary, a MeO-PEG-supported Biphep-RuBr2 cata-
lyst [(R)-4] was synthesized. This catalyst is easy to pre-
pare and practical to use. A series of b-ketoesters have
been successfully reduced in a few hours by using 1–2
mol% of the catalyst. Generally, the reaction rates exceed-
ed those obtained with the non-tethered parent catalyst.
The catalyst could be recovered and reused, yet the yield
and rate dropped considerably in subsequent reactions. An
interesting aspect of these reactions is that the catalyst is
active under hydrogen at atmospheric pressure, negating
the use of an autoclave. This renders the method opera-
tionally convenient.
(d) Bergbreiter, D. E. Catal. Today 1998, 42, 389.
(e) Deng, G.-J.; Li, G.-R.; Zhu, L.-Y.; Zhou, H.-F.; He, Y.-
M.; Fan, Q.-H.; Shuai, Z.-G. J. Mol. Catal. A: Chem. 2006,
244, 118.
(11) This work was first communicated at the 15th International
Symposium on Fine Chemistry and Functional Polymers
held at Shanghai during October 2005.
(12) Schmid, R.; Foricher, J.; Cereghetti, M.; Schonhoizer, P.
Helv. Chim. Acta 1991, 74, 370.
(13) Zhang, Z.-G.; Qian, H.; Longmire, J.; Zhang, X. M. J. Org.
Chem. 2000, 65, 6223.
(14) Preparation of (R)-3: MeO-PEG-OMs (2.6 g, 1.24 mmol),
(6,6¢-dihydroxybiphenyl-2,2¢-diyl)bis(diphenylphosphine)
(2) (0.34 g, 0.62 mmol) and Cs2CO3 (0.8 g, 2.44 mmol) were
added to rigorously degassed DMF (20 mL). The resulting
stirred solution was heated to 65 °C for 18 h. Most of the
DMF was removed under reduced pressure. The resulting
mixture was cooled to 0 °C, H2O (20 mL) and HCl (2 M,
2 mL) were added carefully and the mixture was extracted
with CH2Cl2 (2 × 20 mL). The CH2Cl2 extracts were
combined and washed with brine (20 mL), dried over
MgSO4 and concentrated (to ~ 4 mL). Et2O (400 mL) was
added to the solution and the mixture was stirred under 0 °C
for 0.5 h. The resulting precipitate was isolated by filtration
and washed with cold i-PrOH (30 mL) and Et2O (100 mL),
to give the supported ligand (R)-3 as an off-white solid (2.6
g, 92%). 1H NMR (500 MHz, DMSO-d6): d = 7.12–7.31 (m,
22 H), 6.82 (d, J = 8.4 Hz, 2 H), 6.57 (d, J = 7.0 Hz, 2 H),
3.2–3.8 (polyethylene glycol peaks, ~380 H). 31P NMR (400
MHz, CDCl3): d = –14.05 (s).
(15) Catalyst Preparation: To a mixture of diphosphine ligand
3 (60 mg, 0.0126 mmol) and bis(2-methylallyl)cycloocta-
1,5-diene ruthenium (II) complex (4 mg, 0.0126 mmol) in
anhydrous degassed acetone (1.5 mL) was added 0.18 M
methanolic HBr (0.14 mL, 0.025 mmol). The amber mixture
was stirred at room temperature for 0.5 h and the solvent
removed thoroughly in vacuo to leave the active catalyst,
which was used immediately as a hydrogenation catalyst.
(16) Fan, Q.-H.; Deng, G.-J.; Lin, C.-C.; Chan, A. S. C.
Tetrahedron: Asymmetry 2001, 12, 1241.
(17) Typical Hydrogenation Procedure: The appropriate
ketone (0.63 mmol) was dissolved in degassed EtOH (1 mL)
and the solution was canulated into a Schlenk tube and
degassed by 3 cycles of vacuum/argon. The mixture was
added to the in situ generated catalyst (2 mol%) in a glass
vessel and placed under argon. The argon atmosphere was
replaced with H2 (1 atm) and the mixture was heated for the
period specified in Table 2. After the reaction was complete,
the mixture was cooled to 0 °C and the residue was treated
with cooled Et2O (30 mL). The precipitated polymeric
catalyst was collected by filtration for reuse in the next run.
Both the yield and the ee value of the alcohol were
determined from the filtrate.
Acknowledgment
We are grateful to the Municipal Government of Shanghai for fi-
nancial support (grant no. 03DZ19209).
References and Notes
(1) (a) Kitamura, M.; Tokunaga, M.; Ohkuma, T.; Noyori, R.
Org. Synth. 1993, 71, 1. (b) Mikami, K.; Yusa, Y.;
Korenaga, T. Org. Lett. 2002, 4, 1643. (c) Noyori, R.;
Ohkuma, T.; Kitamura, M.; Takaya, H.; Sayo, N.;
Kumobayashi, H.; Akutagawa, S. J. Am. Chem. Soc. 1987,
109, 5856.
(2) Robertson, D. W.; Krushinski, J. H.; Fuller, R. W.; Leander,
J. D. J. Med. Chem. 1988, 31, 1412.
(3) (a) Chénevert, R.; Fortier, G. Chem. Lett. 1991, 1603.
(b) Chénevert, R.; Fortier, G.; Rhlid, R. B. Tetrahedron
1992, 48, 6769.
(4) (a) Sakuraba, S.; Achiwa, K. Chem. Pharm. Bull. 1995, 43,
748. (b) Lily Co. Drugs Future 1996, 21, 83.
(5) Representative synthesis of optically active fluoxetine:
(a) Srebnik, M.; Ramachandran, P. V.; Brown, H. C. J. Org.
Chem. 1988, 53, 2916. (b) Corey, E. J.; Reichard, G. A.
Tetrahedron Lett. 1989, 30, 5207. (c) Sakuraba, S.; Achiwa,
K. Synlett 1991, 689. (d) Kumar, A.; Ner, D. H.; Dike, S. Y.
Tetrahedron Lett. 1991, 32, 1901. (e) Devocelle, M.;
Agbossou, F.; Mortreux, A. Synlett 1997, 1306. (f) Wang,
G.-Y.; Liu, X.-S.; Zhao, G. Tetrahedron: Asymmetry 2005,
16, 1873.
(6) Huang, H.-L.; Liu, L.-T.; Chen, S.-F.; Ku, H. Tetrahedron:
Asymmetry 1998, 9, 1637.
(7) Ratovelomanana-Vidal, V.; Girard, C.; Touati, R.;
Tranchier, J. P.; Ben Hassine, B.; Genêt, J. P. Adv. Synth.
Catal. 2003, 345, 261.
(8) Genêt, J. P.; Pinel, C.; Ratovelomanana-Vidal, V.; Mallart,
S.; Pfister, X.; Caño De Andrade, M. C.; Laffitte, J. A.
Tetrahedron: Asymmetry 1994, 5, 665.
(9) For excellent reviews, see (a) Benaglia, M.; Puglisi, A.;
Cozzi, F. Chem. Rev. 2003, 103, 3401. (b) Ley, S. V.;
Baxendale, I. R.; Bream, R. N.; Jackson, P. S.; Leach, A. G.;
Synlett 2006, No. 15, 2395–2398 © Thieme Stuttgart · New York