3.2. Enantioselective Syntheses
Enantioselective synthesis of the (
S)-enantiomer of oxybutynin has attracted much interest and deserves a detailed attention, even if oxybutynin is commercially available for medical use as a racemic mixture only. Most efforts have been made during the first decade of this century and were dedicated to the obtention of the (
S)-enantiomer of 2-cyclohexyl-2-hydroxy-2-phenylacetic acid 9. The simplest way to isolate it was to treat, as described in the patent of Bakale
et al [
23], the racemic (100 g) with L-tyrosine methyl ester in order to afford the expected (
S)-oxybutynin in 42 % yield.
(
S)-Oxybutynin hydrochloride ((
S)-4.HCl) [
14]. M.p. 117-118 °C.
1H NMR (400 MHz, DMSO-d6): δ 0.9 to 1.1 (m, 4H), 1.1 to 1.2 (m, 7H), 1.3 (m, 1H), 1.4 (m, 1H), 1.6 (m, 2H), 1.7 (m, 1H), 2.9 (d, 4H), 4.1 (s, 2H), 4.8 (s, 2H), 5.7 (s, 1H), 7.2 (m, 1H), 7.3 (m, 2H), 7.5 (m, 2H), 11.4 (s, 1H) ppm.
13C NMR (100 MHz, DMSO-d6): δ 8.9, 25.2, 25.1, 25.8, 25.9, 25.9, 40.5, 45.7, 46.8, 52.4, 75.4, 80.9, 83.8, 125.7, 127.2, 127.9,141.1, 173.5 ppm.
Besides separation of diastereoisomers, more sophisticated methods have been described to prepare (S)-9 and they are summarized hereafter. Evidently, those protocols can be adapted to afford the (R)-enantiomer of 9.
For example, Senanayake
et al. [
24] obtained (
S)-2-cyclohexyl-2-hydroxy-2-phenylacetic acid in enantiomeric excesses higher than 98 % by forcing a Grignard reagent to preferentially attack one face rather than the other face of 2-oxo-2-phenylacetic acid or 2-oxo-cyclohexylacetic acid derivatives. To achieve that goal, the authors used bulky chiral auxiliaries based on substituted
cis-1-amino-2-indanol moieties (1-
para-tolylsulfonyl group and acetonide). One representative case is depicted in scheme 5. Thus, 2-oxo-2-phenylacetic acid (5) was converted into its acyl chloride and then reacted with
cis-(1
S,2
R)-2-
para-tolylsulfonamidoindanol to afford the ester 18. Subsequent Grignard reaction and hydrolysis of the ester bond yielded the targeted (
S)-2-cyclohexyl-2-hydroxy-2-phenylacetic acid (
S)-9. Other structurally-related chiral intermediates (19-21) evaluated in the study are represented in
Scheme 4. Best yields were obtained from the esters 18 and 19 bearing the
N-
para-tolylsulfonamidoindanyl group.
Scheme 5.
Preparation of (
S)-2-cyclohexyl-2-hydroxy-2-phenylacetic acid (
S)-9 following [
24]. Reagent(s); catalyst; solvent(s); yield(s). (i) (COCl)
2; HCON(CH
3)
2; CH
2Cl
2; 95 %. (ii)
Cis-(1
S,2
R)-2-
para-tolylsulfonamidoindanol; (C
2H
5)
3N; tetrahydrofuran; yield not mentioned. (iii) Bromocyclohexane, Mg; ZnCl
2; tetrahydrofuran; 55%; (iv) NaOHaq; CH
3OH; > 95 %.
Scheme 5.
Preparation of (
S)-2-cyclohexyl-2-hydroxy-2-phenylacetic acid (
S)-9 following [
24]. Reagent(s); catalyst; solvent(s); yield(s). (i) (COCl)
2; HCON(CH
3)
2; CH
2Cl
2; 95 %. (ii)
Cis-(1
S,2
R)-2-
para-tolylsulfonamidoindanol; (C
2H
5)
3N; tetrahydrofuran; yield not mentioned. (iii) Bromocyclohexane, Mg; ZnCl
2; tetrahydrofuran; 55%; (iv) NaOHaq; CH
3OH; > 95 %.
Later, Chattopadhyay
et al. [
25] described another stereocontrolled Grignard reaction (
Scheme 6) by addition of cyclohexylmagnesium bromide on the phenyl ketone 23, obtained from (
R)-2,3-
O-cyclohexylydene-D-glyceraldehyde 22. Hydrolysis of the acetal 24 in fairly good yield required a preliminary benzylation of the alcohol function, thus lengthening the sequence giving access,
via the aldehyde 25, to the expected
S-9. An acceptable overall yield of 25 % was obtained at the end of that 8-step procedure.
In another approach, that could be performed at a 100 g scale (
Scheme 7), trimethylacetaldehyde (pivaldehyde) was acetalized with (
S)-2-hydroxy-2-phenylacetic acid (26; (
S)- mandelic acid). The resulting dioxolone (27) was deprotonated and stereoselectively coupled with cyclohexanone at – 78 °C. Dehydration of the so-formed alcohol (28) followed by hydrolysis and hydrogenation (or the inverse sequence) yielded (
S)-9 in excellent enantiomeric excess (> 99.9 %) and an overall yield of 66% [
26].
Catalytic enantioselective cyanosilylation of cyclohexyl phenyl ketone 29 (
Scheme 8) with a chiral gadolinium complex was the first step of the procedure reported by Shibasaki
et al. [
27]. Subsequent reduction, desilylation, and oxidation afforded the expected (
S)-9. Performed from 100 g of 29, (
S)-9 was isolated in 80 % yield with an enantiomeric excess higher than 99.5 %. To be complete, let us mention that, at the mmol scale, enzymatic resolution of racemic cyanohydrins structurally related to 30 has been the subject of a study of Gotor
et al. [
28].
Alternatively (
Scheme 9), the ketone 29 was introduced in a Wittig reaction with methylenetriphenylphosphorane yielding the alkene 31. Asymmetric dihydroxylation with osmium tetraoxide under Sharpless conditions gave 32. Then, a Swern reaction oxidized the diol into the corresponding aldehyde 25, which could be further oxidized to the acid (
S)-9 with an overall yield of 45 % [
29]. Notice that the aldehyde 25 has also been obtained from 3-[(
Z)-2-phenyl-2-cyclohexylvinyl]oxazolidin-2-one 33 (scheme 9) as proposed by Gourdet and Lam [
30].
Following Maruoka
et al. [
31], (
S)-9 could be prepared through a
L-proline-catalyzed asymmetric aldol reaction between cyclohexanone (in 10-fold excess) and ethyl 2-oxo-2-phenylacetate (34,
Scheme 10). That reaction yielded the ester 35 in good yield (79 %), good diastereoselectivity (dr = > 20:1) and good enantiomeric excess (96 %). However, obtention of the pure corresponding acid (
S)-9 was not straightforward so that the authors had to design a tedious 5-step sequence starting from 32. The overall yield, calculated on 34, fell to 40 %.
In the work of Trost
et al. [
32], the initial precursor of (
S)-9 was the commercially available cyclohex-2-en-1-ol 36 (
Scheme 11), which was converted, in basic medium by treatment with carbon dioxide and then 2-bromoacetophenone, into the ketocarbonate 37. Deprotonation of 37 and protection of the enol by
tert-butyldimethylsilyl chloride was accompanied by an intramolecular rearrangement affording 38. In the subsequent step, an internal allylic alkylation involving a chiral palladium catalyst gave the aldehyde 39. Reduction of the cyclohexenyl ring, oxidation and deprotection afforded the expected (
S)-acid 9 with an enantiomeric excess higher than 99 % but an overall yield of 22 %.
Having in hands (
S)-2-cyclohexyl-2-hydroxy-2-phenylacetic acid 9 (or the (
R) enantiomer), optically active oxybutynin (hydrochloride) could be synthesized using one of the procedures described in
Scheme 3 and
Scheme 4. The activated (
S)-acid 11 has been involved, with the butynyl alcohol 13, in the multigram preparation (40 g) of (
S)-oxybutynin hydrochloride [
14].
3.3. Resolution of the Racemic
Although high-performance thin-layer chromatography has been cited [
33], enantiomers of oxybutynin have been generally separated by high-performance liquid chromatography. The silica-based support of the columns was covalently bonded to ovomucoid [
34] or, more often, coated with polysaccharides such as amylose-tris(3,5-dimethylphenylcarbamate) [
35,
36] or cellulose-tris(4-methylbenzoate) [
37]. Also reported and noteworthy, the use of β-cyclodextrin derivatives, essentially hydroxypropyl-β-cyclodextrin, as chiral selectors for the separation by high-performance liquid chromatography [
38] as well as by electrophoresis [
39], liquid-liquid reactive extraction [
40], and recycling high-speed counter-current chromatography [
41].
Astonishingly, preparation of diastereoisomers was seldom described. One example could be found in the patents of Molnar and Johnston [
18,
19]. The inventors treated the racemic (86.1 g), under its free base form, with D-malic acid in 2-propanol and isolated the D-malate salt of (
R)-oxybutynin in 41 % yield. In the most recent patent [
19], the same inventors claimed that “eleven other chiral acids were tested for production of (
R)-oxybutynin salt from racemic oxybutynin: L-tartric acid, D-tartric acid, L-(+)-lactic acid, D-glucuronic acid,
D-gluconic acid,
L-malic acid, (1
R,3
S)-(+)-camphoric acid, (
S)-(+)-mandelic acid, (1
R)-(-)-10-camphorsulfonic acid, L-pyroglutamic acid, and D-(-)-quinic acid. None of the other eleven chiral acids were successful in chiral resolution to produce an (
R)-oxybutynin salt from racemic oxybutynin.”