A practical synthesis of (S)-oxybutynin

Article abstract of DOI:10.1016/S0040-4039(02)02135-4

(S)-Oxybutynin, an important drug acting as muscarinic receptor antagonist, was practically synthesized using catalytic enantioselective cyanosilylation of cyclohexyl phenyl ketone (6a) as a key step. Cyanohydrin 7a with 94% ee was obtained with 1 mol% of

Full text of DOI:10.1016/S0040-4039(02)02135-4

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 8647–8651
A practical synthesis of (S)-oxybutynin
Shuji Masumoto,a Masato Suzuki,a Motomu Kanaia,b and Masakatsu Shibasakia,*
aGraduate School of Pharmaceutical Sciences, The University of Tokyo, Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
bPRESTO, Japan Science and Technology Corporation (JST), Japan
Received 11 September 2002; revised 25 September 2002; accepted 27 September 2002
Abstract—(S)-Oxybutynin, an important drug acting as muscarinic receptor antagonist, was practically synthesized using catalytic
enantioselective cyanosilylation of cyclohexyl phenyl ketone (6a) as a key step. Cyanohydrin 7a with 94% ee was obtained with
1 mol% of Gd-5 catalyst. The key a-hydroxy carboxylic acid 8 was synthesized from 7a in an enantiomerically pure form without
column chromatography. Other cycloalkyl phenyl ketones, except cyclopentyl phenyl ketone (6e-H) gave products with high
enantioselectivity. The enantioselectivity of the reaction of 6e was dramatically improved using a-deuterium substituted 6e-D. The
dramatic deuterium effect provides an important insight into the competitive reaction pathway. © 2002 Elsevier Science Ltd. All
rights reserved.
1. Introduction
common feature of oxybutynin and related compounds
is the chiral tertiary a-hydroxy carbonyl moiety.
Although oxybutynin is currently prescribed in a
racemic form, the (S)-enantiomer is proposed to have
an improved therapeutic profile. Therefore, there is a
high demand for the development of an efficient enan-
tioselective synthetic route. Previously reported meth-
ods utilized diastereoselective reactions that require a
stoichiometric amount of chiral auxiliary or a chiral
starting material to construct the chiral stereocenter of
1.3 We describe the first practical catalytic enantioselec-
tive synthesis of the key synthetic intermediate 8, utiliz-
ing catalytic cyanosilylation of the ketone as a key step.
Oxybutynin (Ditropan, 1) is a widely utilized mus-
carinic receptor antagonist for the treatment of urinary
urgency, frequency, and incontinence.1 A number of
derivatives have been synthesized mainly by modifying
the ester (or the amide) part and the cycloalkyl part (for
examples, see Fig. 1).2 Some of these analogues have
improved M3-receptor subtype selectivity and thus the
side effects caused by antagonizing other subtypes, such
as M1- and M2-receptors are minimized. A structurally
2. Catalytic enantioselective synthesis of (S)-oxybutynin
We recently developed an (S)-selective catalytic
cyanosilylation of ketones4,5 using a chiral gadolinium
(Gd) complex prepared from Gd(OiPr)3 and the
D
-glu-
6
cose-derived chiral ligand 5 in a 1:2 ratio. We applied
this reaction to the synthesis of the chiral tertiary
a-hydroxy acid core of (S)-oxybutynin. Although the
reaction produces ketone cyanohydrins in high enan-
tioselectivity from a range of ketones, substrate 6a for
the oxybutynin synthesis is an extremely challenging
ketone. In general, to achieve high enantioselectivity in
a nucleophilic addition to carbonyl compounds, the
chiral catalyst needs to differentiate two lone pairs of
the carbonyl oxygen atom for the coordination to the
Lewis acid metal of the chiral catalyst. In the case of
ketone 6a, however, differences in the steric demand of
Figure 1. Oxybutynin and analogues.
Keywords: oxybutynin; muscarinic receptor antagonist; a-hydroxy
carboxylic acid; catalytic enantioselective cyanosilylation; ketones;
practical synthesis.
* Corresponding author. Tel.: +81-3-5841-4830; fax: +81-3-5684-5206;
e-mail: mshibasa@mol.f.u-tokyo.ac.jp
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)02135-4

8648
S. Masumoto et al. / Tetrahedron Letters 43 (2002) 8647–8651
of substrate ketone was added in one portion as a powder
at the last stage to a mixture of the catalyst and TMSCN
to start the reaction.10 Therefore, we first added the
substrate ketone to the catalyst, then cooled the reaction
temperature to −40°C, addition of the solvent, and slowly
added the TMSCN (over 15 min). Using this procedure
improved the results and the product was obtained with
94% ee (entry 4). The reaction and purification proce-
dures are practical and we performed the reaction on a
100 g-scale (see Section 5 for details). After the usual
aqueous work-up, the crude oil was purified through
short-pad silica gel column chromatography (Scheme 1).
Pure 7a was obtained with AcOEt/hexane (1/20) elution,
followed by MeOH/CHCl3 (1/15) elution to obtain the
ligand-containing fraction (a mixture of ligand 5 and the
silylated ligand). The pure ligand 5 was recovered in 98%
yield after acidic desilylation (1 M HCl aq. in THF) and
recrystallization.
these lone pairs are not sufficient because these lone pairs
are cis to sterically similar phenyl or cyclohexyl groups.7
Despite the expected difficulties, we were pleased to find
that the reaction proceeded at −60°C for 21 h in the
presence of 5 mol% of Gd-catalyst, giving the desired
(S)-7a in 96% yield with 95% ee (Table 1, entry 1, catalyst
concentration=0.075 M, ketone concentration=1.5
M).8,9 Lowering the catalyst loading to 1 mol%, however,
resulted in a sluggish reaction at −60°C, giving the
product in 39% with 64% ee (9 days, catalyst concentra-
tion=0.015 M, ketone concentration=1.5 M). Increas-
ing the reaction temperature to −40°C and using a higher
concentration, the reaction proceeded to completion to
give the product in 99% yield with 85% ee (entry 3,
catalyst concentration=0.075 M, ketone concentra-
tion=7.5 M). We hypothesized that the decrease in
enantioselectivity with the lower catalyst amount was due
in part to the initial heat generation when a large amount
Having established a practical method for the catalytic
enantioselective cyanosilylation of the ketone, the next
task was to convert the cyanide to the carboxylic acid.
The conversion was initially problematic, however, prob-
ably due to steric hindrance. Acidic hydrolysis or alco-
holysis gave ketone 6a predominantly through the
elimination of HCN. Basic hydrolysis after conversion to
the THP-protected cyanohydrin resulted in no reaction.
Therefore, we investigated the reduction–oxidation pro-
cedure (Scheme 1). Careful reduction of 7a with DIBAL-
H in toluene, desilylation with 4 N HCl aq. in THF,
followed by oxidation with NaClO2 gave the known
a-hydroxy carboxylic acid 8 in 80% overall yield (3 steps).
Chemically pure 8 was obtained through a base/acid
aqueous work-up without silica gel column chromatogra-
phy. Recrystallization from CH2Cl2/hexane gave an
enantiomerically pure intermediate for oxybutynin syn-
thesis.11
Table 1. Catalytic enantioselective cyanosilylation of 6a
Entry
Cat.
(x mol%)
Temp.
(°C)
Time
(h)
Yield
(%)a
ee
(%)b
1
2
5
5
1
1
−60
−40
−40
−40
21
1.5
50
40
96
99
99
95
91
85
94
3
3. Catalytic enantioselective cyanosilylation of other
aryl cycloalkyl ketones
4c
100
a Isolated yield.
b Determined by chiral HPLC analysis.
The unusually high enantiodifferentiation ability of the
catalyst on the apparently difficult substrate led us to
study the applicability of the catalytic enantioselective
c The ketone was added prior to TMSCN (see Section 5 for details).
In other entries, ketones were added at the last stage.
Scheme 1. Practical catalytic enantioselective synthesis of (S)-oxybutynin key intermediate.

S. Masumoto et al. / Tetrahedron Letters 43 (2002) 8647–8651
8649
cyanosilylation to other aryl cycloalkyl ketones. As
shown in Table 2, high to excellent enantioselectivity
was obtained except in the case of cyclopentyl phenyl
ketone (6e-H). Products obtained with high enantiose-
lectivity should be very useful for synthesizing new
chiral oxybutynin analogues.
The sharp contrast in the enantiomeric excess of 7e-H
(Table 2, entry 4) and products from other cycloalkyl
phenyl ketones is intriguing from a mechanistic point of
view. The results could not be explained from the
ground state structure difference between 6e-H and
other ketones, based on the comparison of the most
stable conformation of these substrates.12 The reactiv-
ity–enantioselectivity relationship of phenyl ketones
with no substituent on the aromatic rings (6a and
6d–g), indicated a general trend in that high enantiose-
lectivity was obtained when the reactions proceeded
smoothly (Table 1, entry 2 and Table 2, entries 3–6).
Thus, we initially hypothesized that the difference in
enantioselectivity might be due to the extent of the
microscopic reversibility of the reaction (Scheme 2,
hypothesis 1). As we clarified in the previous paper,4
the active nucleophile of this catalytic enantioselective
Scheme 2. Two hypotheses on the low reactivity and enan-
tioselectivity of cyclopentyl phenyl ketone.
Table 2. Catalytic enantioselective cyanosilylation of aryl
cycloalkyl ketones
reaction is a gadolinium cyanide of the chiral complex
containing Gd and 5 in a 2:3 ratio (Gd-5). Once the
gadolinium cyanide reacted with a ketone, the initial
product should be a gadolinium alkoxide 9 of the
corresponding ketone cyanohydrin, which would be
quite labile to the retro-reaction. If the equilibrium in
this first step is significantly faster than the second
silylation step, the microscopic reversibility of the enan-
tiodetermining step should diminish the enantioselectiv-
ity.13 Based on the hypothesis, we added a mixture of
free cyanohydrin 10 and TMSCN to the catalyst pre-
cursor 11, and observed the enantiomeric excess
changes. The silylation kinetics for each free cyanohy-
drins (10a, 10e-H, and 10g), however, was extremely
fast (<5 min for completion) even at −40°C,14 and
enantiomeric excess did not change in any cases starting
from enantiomerically enriched or racemic cyanohy-
drins (Scheme 2). Thus, hypothesis 1 was unequivocally
excluded.
Therefore, we next focused on the alternative hypothe-
sis in which the asymmetric catalyst might act as a
Brønsted base to deprotonate the starting ketones
(Scheme 2, hypothesis 2). This competitive pathway
might cause a ligand exchange to produce a less enan-
tioselective catalyst. Based on the hypothesis, we
planned to utilize a deuterium kinetic isotope effect to
prevent the undesired deprotonation.15 As we expected,
the reaction using 6e-D proceeded rapidly and the
product was obtained with 95% ee (Table 2, entry 5
versus entry 4). As far as we notice, this is the first
example of a dramatic advantage using an isotope
effect in catalytic enantioselective reactions. The results
also provide a very important insight into the reaction
a
The absolute configuration of the product was temporarily assigned
based on the analogy to 7a.
Isolated yield.
b
c Determined by chiral HPLC analysis.
d 85%-D.

8650
S. Masumoto et al. / Tetrahedron Letters 43 (2002) 8647–8651
mechanism, in which the Brønsted basicity of the Gd-5
can cause undesired decomposition of the highly enan-
tioselective catalytic species. Studies toward develop-
Japan Science and Technology Corporation (JST). We
thank Prof. E. M. Carreira in ETH for important
suggestions on the origin of low ee for 6e.
ment of
a
new enantioselective catalyst for
cyanosilylation of ketones that eliminates the Brønsted
basicity and demonstrates a further broad substrate
generality are currently under investigation.
References
1. Thompson, I. M.; Lauvetz, R. Urology 1976, 8, 452–
454.
4. Conclusion
2. For example, see: Carter, J. P.; Noronha-Blob, L.;
Audia, V. H.; Dupont, A. C.; McPherson, D. W.;
Natalie, K. J.; Rzeszotarski, W. J.; Spagnuolo, C. J.;
Waid, P. P.; Kaiser, C. J. Med. Chem. 1991, 34, 3965–
3974.
3. (a) Grover, P. T.; Bhongle, N. N.; Wald, S. A.;
Senanayake, C. H. J. Org. Chem. 2000, 65, 6283–6287;
(b) Mitsuya, M.; Ogino, Y.; Ohtake, N.; Mase, T. Tet-
rahedron 2000, 56, 9901–9907; (c) Senanayake, C. H.;
Fang, Q. K.; Grover, P.; Bakale, R. P.; Vandenbossche,
C. P.; Wald, S. A. Tetrahedron Lett. 1999, 40, 819–822.
4. Yabu, K.; Masumoto, S.; Yamasaki, S.; Hamashima,
Y.; Kanai, M.; Du, W.; Curran, D. P.; Shibasaki, M.
J. Am. Chem. Soc. 2001, 123, 9908–9909.
5. For other examples of catalytic enantioselective
cyanosilylation of ketones, see: (a) Tian, S.-K.; Deng,
L. J. Am. Chem. Soc. 2001, 123, 6195–6196; (b)
Belokon’, Y. N.; Green, B.; Ikonnikov, N. S.; North,
M.; Parsons, T.; Tararov, V. I. Tetrahedron 2001, 57,
771–779; (c) Deng, H.; Isler, M. P.; Snapper, M. L.;
Hoveyda, A. H. Angew. Chem., Int. Ed. 2002, 41,
1009–1012.
We developed a practical synthetic route of an impor-
tant pharmaceutical, (S)-oxybutynin. The chiral center
of the core tertiary a-hydroxy carboxylic acid was
constructed by the catalytic enantioselective cyanosilyl-
ation of ketones using 1 mol% of Gd-5 complex. These
procedures are suitable for large-scale synthesis with
minimal silica gel column chromatography purification.
We believe that the methodology described herein
should be practically useful for a process-type supply of
(S)-oxybutynin. Furthermore, a dramatic deuterium
isotope effect on the reaction kinetics and enantioselec-
tivity in the case of ketone 6e suggested that the ketone
deprotonation might become a possible competitive
pathway in some ketones.
5. Experimental
5.1. 100 g-scale cyanosilylation of 6a
Gd(OiPr)3 (5.31 mmol, 0.2 M stock solution in THF) in
THF (26.6 mL) was added to a solution of chiral ligand
5 (4.51 g, 10.6 mmol) in THF (106 mL) in an ice bath
and the mixture was stirred for 30 min at 45°C. After
cooling to room temperature, THF was evaporated and
the residue was dried for 6 h under vacuum (5 mmHg).
To this catalyst powder, ketone 6a (100 g, 0.531 mol)
was added. Propionitrile (71 mL) and TMSCN (85 mL,
0.637 mol) were successively added at −40°C, and the
mixture was stirred for 40 h at −40°C. H2O was added
to quench the reaction (caution: HCN is generated),
and the product and the ligand were extracted with
AcOEt. The combined organic layer was washed with
satd. NaCl aq. and dried over Na2SO4. Evaporation of
the solvent gave a crude oil, which was purified through
short pad SiO2 column chromatography (450 g: AcOEt/
hexane=1/20) to give pure 7a as a colorless oil (152 g,
100% yield). The enantiomeric excess of 7a was deter-
mined by chiral HPLC after conversion to 8 [DAICEL
CHIRALPAK AS, hexane/2-propanol/TFA=95/5/0.1,
1.0 mL/min, tR 7.3 (minor) and 10.6 min (major)]. The
ligand and the silylated ligand were eluted from the
column with CHCl3/MeOH=15/1, treated with HCl
aq. in THF, extracted, and purified by recrystallization
to recover pure ligand 5 in 98% yield.
6. Purchased from Kojundo Chemical Laboratory Co.,
Ltd. (Fax: +81-492-84-1351).
7. Corey et al. reported that electronic effect can deter-
mine the effective size of the carbonyl substituents:
Corey, E. J.; Helal, C. J. Tetrahedron Lett. 1995, 36,
9153–9156.
8. The absolute configuration of 7a was determined to be
(S) after conversion to carboxylic acid 8 and by a com-
parison of the optical rotation with the reported value
(Ref. 3a). This is the same enantioselectivity as with
other previously studied simple ketones such as ace-
tophenone, suggesting that the catalyst recognizes the
cyclohexyl group as a smaller group than the phenyl.
9. The (R)-selective titanium complex could not promote
the reaction from 6a at low temperature. The reaction
proceeded at room temperature and (R)-7a was
obtained in 96% yield with only 7% ee (36 h). For the
(R)-selective catalytic cyanosilylation of ketones, see:
Hamashima, Y.; Kanai, M.; Shibasaki, M. J. Am.
Chem. Soc. 2000, 122, 7412–7413.
10. This procedure was the representative one previously
(Ref. 4).
11. For the conversion of 8 to 1, see: Bakale, R. P.; Lopez,
J. L.; McConville, F. X.; Vandenbossche, C. P.;
Senanayake, C. H. US Pat. 6,140,529.
12. Molecular modeling studies indicated that 6e contains
similar conformation and electronic character with
other ketones.
Acknowledgements
13. A facile equilibrium followed by a kinetic resolution
through an enantioselective silylation might be another
possible pathway, however this possibility was also
Financial support was provided by RFTF of Japan
Society for the Promotion of Science and PRESTO of

S. Masumoto et al. / Tetrahedron Letters 43 (2002) 8647–8651
8651
15. For the use of an isotope effect in organic synthesis to
prevent the undesired deprotonation, see: (a) Dudley, G.
B.; Danishefsky, S. J.; Sukenick, G. Tetrahedron Lett. 2002,
43, 5605–5606; (b) Vedejs, E.; Little, J. J. Am. Chem. Soc.
2002, 124, 748–749.
excluded based on the experiments described in the text.
The related mechanism was proposed in Deng’s catalytic
enantioselective cyanosilylation of ketones (Ref. 5a).
14. In the absence of the catalyst precursor 11, the silylation
of the ketone cyanohydrins with TMSCN was very sluggish.