Enantioselective synthesis of (1R,2S,4S)-7-oxabicyclo[2.2.1]heptan-2-exo-carboxylic acid

Article abstract of DOI:10.1016/j.tetasy.2009.12.030

A scalable enantioselective access to (1R,2S,4S)-7-oxabicyclo[2.2.1]heptan-2-exo-carboxylic acid 7, a key precursor in the synthesis of A2a receptor antagonist 1 by means of an enzymatic resolution of the respective butyl ester with lipase A from Candida antarctica, is described.

Full text of DOI:10.1016/j.tetasy.2009.12.030

Tetrahedron: Asymmetry 21 (2010) 159–161
Contents lists available at ScienceDirect
Tetrahedron: Asymmetry
journal homepage: www.elsevier.com/locate/tetasy
Enantioselective synthesis of (1R,2S,4S)-7-oxabicyclo[2.2.1]heptan-
2-exo-carboxylic acid
Beat Wirz , Paul Spurr , Christophe Pfleger c
a,*
b
a
Chemical Synthesis, Biocatalysis, Hoffmann-La Roche Ltd, Pharma Research Basel, 4070 Basel, Switzerland
Chemical Synthesis and Process Research, Hoffmann-La Roche Ltd, Pharma Research Basel, 4070 Basel, Switzerland
Chemical Synthesis, Kilolab, Hoffmann-La Roche Ltd, Pharma Research Basel, 4070 Basel, Switzerland
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 8 December 2009
Accepted 21 December 2009
A scalable enantioselective access to (1R,2S,4S)-7-oxabicyclo[2.2.1]heptan-2-exo-carboxylic acid 7, a key
precursor in the synthesis of A2a receptor antagonist 1 by means of an enzymatic resolution of the
respective butyl ester with lipase A from Candida antarctica, is described.
Ó 2010 Elsevier Ltd. All rights reserved.
1
. Introduction
Adenosine is a key modulator of many neurophysiological pro-
of them really revealed sufficient practical promise in terms of
yield, cost or selectivity to warrant further thorough investigation.
Instead, under relatively tight time constraints, an existing pub-
2
cesses and the high concentration of A2a-receptors present in the
dopamine-rich regions of the brain are intimately involved in the
control of emotion, reward and pleasure. Benzothiazole derivative
lished, non-patented route to compound 12 as the racemate was
chosen and adapted accordingly. Essentially it comprised the
uncatalyzed cycloaddition of furan to methyl acrylate, affording a
6:1 endo-/exo-mixture of rac-ester adducts which were separated
by chromatography after hydrogenation. The remaining steps par-
alleled those shown in Scheme 1. Since the cycloaddition reactions
were unselective and the isomer separation was impractical, a
more manageable synthesis was essential for the preparation of
enantiopure 12, especially for scale-up purposes. The Diels–Alder
reaction was significantly ameliorated by implementing instead a
1
was under evaluation as a A2a receptor antagonist candidate for
1
the treatment of major depression, and in view of clinical studies,
a scalable synthesis had to be developed. Among several alterna-
tive approaches evaluated, only the convergent assembly from
two key components, the bicyclic methylamine 12 and aminoben-
zothiazole 2 (via an activated carbamate, cf. Scheme 1) turned out
to be viable on a larger scale.
2
ZnCl -mediated cycloaddition (0.3 equiv, rt/16 h) of furan with
acrylonitrile, generating the bicyclic eneether 3 as a 1:1 mixture
H
3
O
of endo- and exo-isomers which after hydrogenation to the satu-
O
O
(S)
3a
NHMe.HCl
rated nitrile isomers 4 were hydrolyzed under alkaline conditions
N
S
(
(
aq KOH/EtOH, 75 °C/2 h) exclusively to the racemic exo-acid 5
95% from 3) thus solving the initial selectivity issue. Enantioselec-
H
NH
H
N
1
2
tive enzymatic hydrolysis of the respective racemic exo-butylester
6b gave rise to the (S)-acid 7 which was re-esterified to the ethyl
ester 8 required for successful conversion in the subsequent reac-
tion sequence. In a one-pot, three-step process, successive hydrazi-
O
1
O
. Results and discussion
Several potential routes to building block 12 based on various
nolysis, treatment of the acylhydrazide 9 with NaNO /aq HCl and
Curtius rearrangement in EtOH of the acylazide 10 intermediate
2
2
created in situ produced the carbamate 11. Reduction with LAH
furnished the bicyclic methylamine 12, isolated as the HCl-salt in
4
an overall yield of 15% (Scheme 1).
cycloaddition reactions followed by enzymatic or classical resolu-
tion steps were considered and partially evaluated. However, none
2
.1. Evaluation of chemoenzymatic routes
In the course of route evaluation, the potential of an array of
enzyme-catalyzed resolution reactions was cursorily screened
*
Corresponding author. Tel.: +41 61 688 3390; fax: +41 61 688 1673.
E-mail address: beat.wirz@roche.com (B. Wirz).
(see below); only the kinetic resolution of the racemic exo-ester
0
957-4166/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2009.12.030

1
60
B. Wirz et al. / Tetrahedron: Asymmetry 21 (2010) 159–161
R =
1
. (COCl) / cat. DMF
tol, RT/1h
O
O
O
2
O
a: ethyl
b: butyl
c: pentyl
d: hexyl
e: octyl
CN
H / Pd-C
2
KOH
COOH
H
COOR
H
O
CN
CN
cat. ZnCl2
EtOAc
RT/24h
~100%
EtOH
75°/1.5h
~95%
2. nBuOH, RT/1h
RT/16h
~95%
~90%
3
4
5
rac-6
Lipase Cand. ant.
(
rac~1:1 endo/exo)
(rac~1:1 endo/exo)
(rac exo)
aq. NaOH- ~35% (>98% ee)
NaH PO
2
4
(
pH 7) RT/72h
O
O
(
O
O
CON3
NaNO / aq. HCl
CONHNH2
H
N H .H O
COOEt
cat. H SO / EtOH
COOH
H
2
2
4
2
2
4
CH Cl , 0°/3h
80°/16h
80°/16h; ~90%
2
2
H
H
(S)-10
(S)-7
S)-9
(S)-8
CH Cl -EtOH
2
2
~80% (3 steps)
4
0°/48h
OMe
PhOCOCl / pyr /
CH Cl2
N
2
1
. LAH / THF
O
RT/3h then iPr NEt
40°/16h; ~85%
2
NH2
O
NHCOOEt
H
NHMe
HCl
6
0°/1.5h
S
.
2
. aq. HCl
EtOAc
2
H
~75%
(
S)-12
(
S)-11
~8 steps, ~ 15% o.a.
O
1
Scheme 1. Chemoenzymatic route to A2a receptor antagonist 1.
6
delivered results with practical relevance (see next section). Res-
of which is Chirazyme L-5 (Roche Diagnostics) or the bulk enzyme
Novocor AD L (Novozymes)—exhibited moderate selectivity (E
ꢀ35). The longer chain alkyl esters 6c–e exhibited E-values >90
at rt, nevertheless the reaction had to be stopped before reaching
50% conversion in order to obtain the acid in >97% ee which was
mandatory owing to the modest enantiomeric enrichment attain-
able in the subsequent steps.
olution of the structurally related esters 13 derived from the unsat-
urated compound exo-7-oxabicyclo[2.2.1]hept-5-ene-2-carboxylic
acid has been described in the literature. However, both the
methyl ester (using Candida rugosa ) and the ethyl ester (using li-
pase from Candida antarctica, form B5b) were resolved only with
moderate selectivity (E <15), insufficient for large-scale prepara-
tion of the formed (S)-configured acid in high enantiomeric excess.
In addition, problems due to concomitant retro Diels–Alder reac-
tion were observed during workup of the product.5b
5
5a
Employing the pentyl ester 6c at a technically acceptable sub-
strate concentration of 5%, acid 7 was obtained in ꢁ98% ee at
ꢁ41% conversion (E P180) using the liquid enzyme formulation
Novocor AD L (s/e 5) in 0.1 M NaCl and 4 mM sodium phosphate
O
4
buffer pH 7.0 at 28 °C. At higher substrate concentrations the
COOR
selectivity deteriorated significantly. Since on a kg-scale the prep-
aration of the pentyl ester suffered from thermal stress in the re-
moval of excess pentanol, the butyl ester 6b was favoured.
Optimization of reaction parameters for 6b revealed the positive
effect of elevated phosphate ion concentration (100 mM; pH 7.0)
on the reaction rate (ꢁ1.4-fold acceleration). This in turn permitted
the reaction temperature to be lowered to 10 °C, improving selec-
tivity. Finally, at a butyl ester concentration of 5%, the product was
1
3
A more direct enantioselective access to acid 7 was attempted
by the stereoselective action of nitrilases on nitrile 4 prepared with
an improved exo/endo-ratio of 9:1 when employing 1 equiv of
, 50 °C, 72 h in the Diels–Alder reaction. However, screening
of microbial strains containing nitrilase activity as well as a panel
of commercially available nitrilases afforded only modest selectiv-
ities (E <14). Likewise, no appreciable reaction was observed in
the attempted hydroxylaminolysis of the pentyl ester 6c with li-
pase A and B from C. antarctica in various solvents, as well as in
the enzymatic formation of hydroxamic acid from acid 5.
6
ZnCl
2
9
obtained in 98% ee at 40% conversion (E P180). Since the enzyme
7
did not act upon the endo-ester, present in only a very small
amount in the substrate, the endo-form was easily removed to-
gether with the retained ester and the fairly water-soluble product
acid 7 was isolated by repeated extraction with organic solvent at
acidic pH. Acid 7 was obtained in 98% ee and 36% yield and could
be enantiomerically enriched to >99.5% ee by means of digestion
2
.2. Enantioselective hydrolysis of bicyclic ester 6
9
from MTBE-heptane 1:2 (32% yield). The procedure was success-
In an extensive enzyme screening for the enantioselective
fully scaled-up to a multi-kg-scale.
hydrolysis of ethyl ester 6a, none of the enzymes showed apprecia-
ble preference for retention of the (1R,2S,4S)-configurated ester
which would have provided a short and robust route to ester 8.
Therefore, the pathway had to proceed via the formed (1R,2S,4S)-
acid 7 which required a high selectivity (E >100).8 With the butyl
ester 6b only lipase from C. antarctica form A—a commercial source
3. Conclusion
The present enzymatic resolution procedure of 6b enabled a
scalable enantioselective synthesis of the enantiopure bicyclic
moiety 12 (eight steps, ꢁ15% o.a.) with technical potential for

B. Wirz et al. / Tetrahedron: Asymmetry 21 (2010) 159–161
161
the large scale access to the A2a receptor antagonist 1 devoid of
chromatographic purifications, extreme conditions and toxic/exo-
tic reagents.
4. (a) Rumi, L.; Pfleger, C.; Spurr, P.; Klinkhammer, U.; Bannwarth, W. Org. Process
Res. Dev. 2009, 13, 747–750; (b) Spurr, P.; Wirz, B. Patent Application
US2008154043.
5.
(a) Schueller, C. M.; Manning, D. D.; Kiessling, L. L. Tetrahedron Lett. 1996, 37,
Route assessment including several enzyme screenings indi-
cated the enzymatic kinetic resolution of racemic 7-oxa-bicy-
clo[2.2.1]heptan-2-exo-carboxylic acid ester 6 as the favoured
option and work on a larger scale indicated the butyl ester 6b as
the substrate of choice. The screening of various chemical and
physical reaction parameters revealed an enhanced reaction rate
at elevated phosphate buffer concentration. The increased reaction
rate in turn allowed a reduction of the reaction temperature,
enhancing enzyme selectivity. The enzymatic procedure estab-
lished was scaled-up to a multi-kg level.
8853–8856; (b) Abrecht, S.; Karpf, M.; Trussardi, R.; Wirz, B. European Patent EP
1
127 872 B1, 2001.
6
.
Abrecht, S.; Cordon Federspiel, M.; Estermann, H.; Fischer, R.; Karpf, M.; Mair,
H.-J.; Oberhauser, T.; Rimmler, G.; Trussardi, R.; Zutter, U. Chimia 2007, 61, 93–
99.
Personal communication Steven Paul Hanlon and Hans Iding, Hoffmann-La
Roche Ltd.
7
8
9
.
.
.
Chen, C.-S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J. J. Am. Chem. Soc. 1982, 104,
7294–7299.
Butyl ester 6b (100 g, 504 mmol, 99%) was emulsified in 2 L 100 mM sodium
phosphate buffer pH 7. The pH was adjusted to 7.1 with 5.7 ml aq NaOH (32%)
and the temperature lowered to 10 °C. Hydrolysis was started with 15 ml C.
antarctica A lipase Novocor AD L (Novozymes, Bagsvaerd, Denmark) and the pH
was maintained at 7.0 by the controlled addition of 1 M NaOH solution under
vigorous stirring. After a consumption of 200.3 ml (40% conversion; 72 h; 97.9%
ee) the reaction was stopped by the addition of 350 ml dichloromethane,
followed by 40 g filter aid dicalite. The mixture was stirred for 0.3 h, filtered and
extracted with 1.2 L and 0.5 L MTBE. The organic phases were washed separately
with 100 ml deionized water and the combined aqueous phases acidified with
Acknowledgements
Thanks go to R. Agra, K.-J. Gutmann, P. Stocker, R. Santillo, P.
Schmitt, S. Wang, S. Corpataux, S. Frrokaj, C. Jenny, F. Koch, F. Mon-
tavon, S. Polat, M. Pozar, J. Reimann, S. Weiss, K. Wetzel, M. Christ,
F. Gantz, D. Hediger, S. Hildbrand, H. Stahr, F. Tixeront and H. Wet-
zel for the technical support and to A. Stämpfli, S. Brogly, C. Tour-
noux, A. Becht, R. Ineichen and O. Zemp for the analytical
assistance.
55 ml hydrochloric acid (25%) to pH 2. Then 40 g of Dicalite were added, and
after stirring for 0.3 h and filtering, the filtrate was saturated with 800 g NaCl
and extracted four times with 0.8 L MTBE. The combined extracts were dried
over Na
2 4
SO , filtered and evaporated at 50 °C/440-280 mbar, yielding 26 g (36%)
of a white solid (GLC purity: 97.5%, ee: 98.3%, [
a
]
D
= ꢂ28.8 (c 1.00, EtOH). For
further purification, a main portion of the crude material (24.6 g) was suspended
in 25 ml MTBE and heated to 50 °C. The partial solution attained was treated
after 0.2 h with 50 ml heptane at the same temperature. The white suspension
was cooled to rt and stirred for 3 h. The product was filtered, washed three times
with 2.5 ml MTBE/heptane 1:2 and dried for 4 h at 50 °C/30 mbar providing
References
2
1.8 g 7 (89% recovery, 32% step yield based on assay for entire amount) as a
+
1.
2.
3.
Flohr, A.; Jakob-Roetne, R.; Norcross, R.; Riemer, C. US2004235915.
Nelson, W.; Allen, D. J. Heterocycl. Chem. 1972, 9, 561–568.
(a) Sundermann, B.; Scharf, H. Synlett 1996, 703–704; (b) Garrigues, B.; Oussaid,
A. J. Organomet. Chem. 1999, 585, 253–255; (c) Fraile, J.; Garcia, J.; Gomez, M.; de
la Hoz, A.; Mayoral, J.; Moreno, A.; Prieto, P.; Salvatella, L.; Garcia, E. Eur. J. Org.
Chem. 2001, 2891–2899.
white crystalline solid. Analysis: 99.6% GLC; 99.6% ee; EI-MS: 143.2 (M+H ),
142.2 (M); IR (nujol): 2923, 1720, 1230, 1192 (COOH), 1140 (C–O–C) cm
NMR (CDCl , 400 MHz): d (ppm) 10.6 (s br, 1H, COOH), 4.86 (d, J = 5.1 Hz, 1H, H-
1), 4.69 (t, J = 5.0 Hz, 1H, H-4), 2.67 (dd, J = 9.1, 4.6 Hz, 1H, H-2), 2.06–2.22 (m,
ꢂ1
1
; H
3
0
00
0
0
00
00
1H, H-3 ), 1.69–1.85 (m, 3H, H-3 , H-5 , H-6 ), 1.43–1.60 (m, 2H, H-5 , H-6 ).