Alternative esters in the synthesis of ZD0947

Article abstract of DOI:10.1016/j.tetlet.2005.03.057

Alternatives to the original iso-borneol and allyl esters were investigated in the synthesis of ZD0947 (1). Homologous allylic esters were prepared in higher yields and were more readily purified than for the existing route, but offered no further benefits later in the synthesis. However, the PNB-substituted ester series provided crystalline intermediates, higher yields and simplified isolations through-out, and utilised a heterogeneous hydrogenation, thus reducing the residual catalyst levels.

Full text of DOI:10.1016/j.tetlet.2005.03.057

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 3179–3181
Alternative esters in the synthesis of ZD0947
Jonathan D. Moseley*
AstraZeneca, Process Research and Development, Avlon Works, Hallen, Bristol, BS10 7ZE, UK
Received 17 February 2005; accepted 10 March 2005
Abstract—Alternatives to the original iso-borneol and allyl esters were investigated in the synthesis of ZD0947 (1). Homologous
allylic esters were prepared in higher yields and were more readily purified than for the existing route, but offered no further benefits
later in the synthesis. However, the PNB-substituted ester series provided crystalline intermediates, higher yields and simplified
isolations through-out, and utilised a heterogeneous hydrogenation, thus reducing the residual catalyst levels.
Ó 2005 Elsevier Ltd. All rights reserved.
ZD0947 (1) is a potassium channel opener in clinical
development for the treatment of urinary urge inconti-
nence (Ôoveractive bladderÕ).1 The pharmacological pro-
file of such a compound should have beneficial effects on
an overactive bladder but with reduced cardiovascular
side effects, and the side effects associated with existing
anti-muscarinic agents. The initial medicinal chemistry
synthesis, in which the desired (S)-enantiomer of 1 was
isolated after a lengthy classical resolution has been dis-
closed.2 An asymmetric synthesis, which significantly
improved on this synthesis has also been reported.3 At
the same time as that work, an investigation into alter-
native esters was conducted to improve the early stages
of the racemic synthesis and avoid the use of a homo-
genous catalyst. This investigation is the subject of this
letter.
acid-labile iso-borneol ester had been chosen to avoid
decarboxylation under alkaline conditions, but on
scale-up, considerable decarboxylation to racemic
ZD0947 (6) was observed under the forcing acid condi-
tions. A change to the allyl ester 2c was made; 4c could
then be cleaved with WilkinsonÕs catalyst to avoid the
decarboxylation. This was a major improvement and
supplied ZD0947 for pre-clinical trials. However, it
proved difficult to remove the Rh catalyst residues to
the low levels required for toxicity testing (i.e.,
<5 ppm).
Preparation of allyl ester 2c also presented some difficul-
ties. Trans-esterification of ethyl TFA acetate (2a) with
allyl alcohol was achieved by distilling off ethanol, but
the yield was moderate (ꢀ35%) and the distillation
was inefficient due to the similar boiling points of the
two alcohols, which required several cycles of distil-
lation and recharging of allyl alcohol.6 Lastly, the key
Hantzsch reaction yielded a moderate 59% after an
involved work-up procedure.
The original medicinal chemistry synthesis is shown in
Scheme 1. The key step is the unsymmetrical Hantzsch
reaction4 between the benzaldehyde, cyclohexa-1,3-
dione, ammonia and acetoacetate 2b (R = iso-borneol)
to give the Hantzsch product 3b. Dehydration of this
unusually stable tetrahydropyridine gave dihydropyr-
idine 4b, and cleavage of the ester gave the desired race-
mic acid 5, from which 1 could be isolated after
resolution with (S)-(a)-methylbenzylamine.
Since the ester was discarded during the synthesis, we
felt that these factors could be improved by the use of
alternatives to the allyl and iso-borneol esters. The alco-
hol component needed to be commercially available and
readily cleaved to avoid decarboxylation. Higher boiling
allylic alcohol homologues (2d–g) offered the prospect of
achieving both of these aims, whilst the benzyl esters
(2h–j) offered the additional advantage of cleavage by
heterogeneous catalysis (and other methods in the case
of 2i and 2j)7 (Scheme 2). Both sets of alternatives con-
served methodology with the existing synthesis for the
final steps.
Although this synthetic sequence was straightforward
and short, a number of issues were apparent.5 The
Keywords: Hantzsch; ZD0947.
*
Tel.: +44 117 938 5601; fax: +44 117 938 5081; e-mail:
jonathan.moseley@astrazeneca.com
0040-4039/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2005.03.057

3180
J. D. Moseley / Tetrahedron Letters 46 (2005) 3179–3181
CN
O
O
O
O
CN
O
O
O
b
c
OR
CF3
a
+
+
OEt
F3C
OR
F3C
O
N
H
3b-j
OH
CHO
2a (R = ethyl)
2a
2b (R = isoborneal)
2c (R = allyl)
CN
CN
O
CN
O
CN
O
O
O
O
c (or f)
d, e
OR
CF3
OH
CF3
+
N
H
CF3
N
H
N
H
N
CF3
H
1 (ZD0947)
4b-j
5
6
Scheme 1. Reagents and conditions: (a) alcohol, distillation (76%); (b) NH4OAc, EtOH, reflux (59%); (c) pTSA, AcOH, 100 °C (54%); (d) (S)-(a)-
methylbenzylamine, toluene/n-BuOH (26%); (e) NMP, heat (85%); (f) n-BuOAc/EtOH/H2O (4:2:1), AcOH, (Ph3P)3RhCl, 70 °C (yields a–c given for
iso-borneol derivatives).
2d, R =
2h (X = H), R =
O
O
2e, R =
2f, R =
2g, R =
2i (X = OMe), R =
2j (X = NO2), R =
OR
F3C
X
2d-j
Ph
Scheme 2.
Esters 2e–j were readily prepared by distilling off ethanol
in toluene.8 Yields were significantly higher (50–80%)
than for 2c but processes were not optimised. The analo-
gous Hantzsch products 3e–j were prepared in each case,
and dehydrated to the analogous dihydropyridines 4e–
j.9 During the dehydration, both the isoprenyl (3e, 4e)
and geranyl (3f, 4f) derivatives showed signs of cleavage
to 6. A screen of acids and solvents suggested that pTSA
or concd H2SO4 in toluene or n-butanol were the most
promising combinations to limit this. The cinnamyl
derivative (3g, 4g) proved to be more stable but the
Hantzsch reaction yield was poor. Consequently, the
WilkinsonÕs catalyst ÔdeprotectionÕ was not tested in
any of these cases.
The acid-labile PMB-substituted series (2/3/4i) proved to
be too sensitive to the acidic dehydration conditions and
4i could not be isolated. However, the PNB-substituted
series (2/3/4j) gave stable crystalline compounds. Initial
formation of the Hantzsch product 3j appeared low
(22%), due to large quantities of the intermediate pyran
7 being formed instead (50%). However, addition of
NMP as a co-solvent (17% v/v with EtOH) converted
7 in situ to the desired 3j, and a simple aqueous
drown-out gave a 78% overall yield. Conversion to 4j
proceeded as expected and heterogeneous hydrogen-
ation of the PNB group under standard or transfer
hydrogenation10 conditions gave 5 in near quantitative
yield.11
In conclusion, formation of the esters 2e–j proceeded in
higher yields and were more readily purified than for 2c,
but the allyl homologues offered no further advantages
compared to the allyl or iso-borneol groups. However,
the PNB-substituted series (2/3/4j) provided crystalline
intermediates, higher yields and simplified isolations
through-out, and utilised a heterogeneous hydrogen-
ation, thus reducing the residual catalyst levels.
CN
O
O
O
NO2
CF3
OH
X
7 (X = O); 3j (X = NH)
It was hoped that the benzyl substituents would confer
greater crystallinity on the Hantzsch and dihydropyr-
idine analogues, but the benzyl series (2/3/4h) proved to
be no better than the existing allyl series in this respect.
Acknowledgements
The author thanks Claire Ancell for additional labora-
tory preparations.

J. D. Moseley / Tetrahedron Letters 46 (2005) 3179–3181
3181
7. Hudlicky, M. Reductions in Organic Chemistry; Ellis
Horwood: Chichester, 1984.
References and notes
8. The hindered compound 2d could not easily be prepared
due to the similar b.p. of its alcohol with ethanol.
1. Abdel-Karim, A. M.; Bialecki, R. A.; Elhilali, M. M.
J. Urol. 2002, 168, 837–842.
2. Ohnmacht, C. J.; Trainor, D. A.; Forst, J. M.; Stein, M.
M.; Harris, R. J. US Patent 5,622,964.
1
9. All new compounds gave satisfactory H and 13C NMR
and mass spec. data, and purities were assayed by HPLC
(or GC in the case of the acetoacetates 2b–j) and by
comparison to the H NMR spectra (see also Ref. 11).
3. Ashworth, I.; Hopes, P.; Levin, D.; Patel, I.; Salloo, S.
Tetrahedron Lett. 2002, 43, 4931–4933.
1
10. Ram, S.; Ehrenkaufer, R. E. Synthesis 1988, 91.
11. For the PNB-substituted series, dehydration was achieved
with pTSA in toluene at reflux (82%), standard hydrogen-
ation with 10% Pd/C and H2 in methanol (95%), or
transfer hydrogenation with 10% Pd/C and HCO2NH4 in
methanol (80%). Full experimental details and analyses of
these and other key compounds are described in: Jones,
A.; Moseley, J.; Patel, I.; Snape, E.; Young, M. World
Patent WO 2003062203, 2003.
4. For a recent review including the Hantzsch reaction, see
the following and references cited therein: Simon, C.;
Constantieux, T.; Rodriguez, J. Eur. J. Org. Chem. 2004,
4957–4980.
5. The loss of half of the material as the (R)-enantiomer was
addressed by the asymmetric route (see Ref. 3).
6. This process was later improved by the availability of large
scale short-path distillation equipment, and subsequently
by an alternative synthesis.