Development of a concise, asymmetric synthesis of a smoothened receptor (smo) inhibitor: Enzymatic transamination of a 4-piperidinone with dynamic kinetic resolution

Article abstract of DOI:10.1021/ol403630g

A concise, asymmetric synthesis of a smoothened receptor inhibitor (1) is described. The synthesis features an enzymatic transamination with concurrent dynamic kinetic resolution (DKR) of a 4-piperidone (4) to establish the two stereogenic centers required in a single step. This efficient reaction affords the desired anti amine (3) in >10:1 dr and >99% ee. The title compound is prepared in only five steps with 40% overall yield.

Full text of DOI:10.1021/ol403630g

Letter
pubs.acs.org/OrgLett
Development of a Concise, Asymmetric Synthesis of a Smoothened
Receptor (SMO) Inhibitor: Enzymatic Transamination of a
4‑Piperidinone with Dynamic Kinetic Resolution
Zhihui Peng,*,† John W. Wong,† Eric C. Hansen,† Angela L. A. Puchlopek-Dermenci,†
and Hugh J. Clarke‡
‡
†Chemical Research & Development, Analytical Research & Development, Pfizer Worldwide Research & Development, Eastern
Point Road, Groton, Connecticut 06340, United States
S
* Supporting Information
ABSTRACT: A concise, asymmetric synthesis of a smooth-
ened receptor inhibitor (1) is described. The synthesis features
an enzymatic transamination with concurrent dynamic kinetic
resolution (DKR) of a 4-piperidone (4) to establish the two
stereogenic centers required in a single step. This efficient
reaction affords the desired anti amine (3) in >10:1 dr and
>99% ee. The title compound is prepared in only five steps
with 40% overall yield.
Scheme 1. Retrosynthetic Analysis
he smoothened receptor (SMO), a component of the
hedgehog (Hh) signaling pathway, represents a novel
T
therapeutic target in a broad range of human cancers.1
Compound 1, a potent and orally bioavailable inhibitor of
SMO recently reported by Pfizer,2 has shown early signs of
efficacy and a good safety profile in phase 1 clinical trials
involving patients with various blood-related cancers, including
acute myeloid leukemia (AML).
To support the ongoing drug development program, a
practical and scalable synthesis of 1 was required. The key
synthetic challenge for the preparation of 1 was the effective
and practical establishment of the anti stereogenic centers on
the 2,4-positions of the piperidine ring. Previous syntheses2,3
employed (2R,4S)-4-hydroxypipecolic acid derivatives 2
(Scheme 1), which were not readily available on scale. The
syntheses were also tedious, and the overall yields were low.3
Furthermore, installation of the C-4 amino group required
three steps (alcohol activation, SN2 displacement with azide,
and reduction) and posed safety concerns in large scale
preparation. Over the past few decades, elegant methodologies
have been developed for the synthesis of chiral amines.4
However, reports of direct enantioselective reductive amination
of prochiral ketones in the absence of extra functional group
interconversion steps are rare.5 In this regard, enzymatic
transamination reactions provide an advantage over traditional
chemical synthesis.6 The increased availability of transaminases
and the advancement of enzyme engineering have made
enzymatic transamination an increasingly attractive and viable
manufacturing option for chiral amine synthesis in pharma-
ceuticals.7 We envisioned that an enzymatic transamination of
4-piperidone 4 would not only generate the C-4 chiral amino
group required in the penultimate 3 but also establish the anti
C-2 stereocenter at the same time through a dynamic kinetic
resolution (DKR) process.8 It should be noted that the facile
racemization and a crystallization-driven DKR of 2-aryl-4-
piperidones have been reported in the literature previously.9
To assemble 4-piperidone 4, we elected to explore the
nucleophilic addition of benzimidazole to a pyridinium salt,
which has proven to be a powerful method for the synthesis of
substituted piperidines.10 To the best of our knowledge,
heterocyclic carbon nucleophiles such as benzimidazole have
not been employed in this approach. We were pleased to find
that the N-Ts benzimidazole 511 was fully deprotonated with
LDA at −15 °C and underwent addition to 4-methoxy N-
methyl pyridinium triflate 612 to give the corresponding
dihydropyridine 7. Following mild acidic hydrolysis of the
Received: December 16, 2013
Published: January 22, 2014
© 2014 American Chemical Society
860
dx.doi.org/10.1021/ol403630g | Org. Lett. 2014, 16, 860−863

Organic Letters
Letter
enol ether the enone could be unmasked to furnish
dihydropyridone 8 (Scheme 2). A competing reaction pathway
HCl) cleanly afforded 4-piperidone 4 in 95% isolated yield by
direct filtration as a crystalline di-HCl salt hydrate.16
With 4-piperidone 4 in hand, the stage was set for the crucial
enzymatic transamination with DKR. To investigate the
feasibility of the racemization of 4, we obtained enantiomeri-
cally pure material by chiral SFC separation of racemic 4.
Complete racemization of a single enantiomer was found to
occur in less than 8 h at 40 °C in a 1/4 mixture of DMSO and
aqueous pH 10 buffer, the medium for the enzymatic
transamination. Although we did not have direct evidence of
the corresponding ring-opened intermediate 10, we suspected
that retro-aza-Michael/aza-Michael was the mechanism by
which the racemization of 4 occurred (Scheme 4).17
Scheme 2. Preparation of Dihydropyridone 8
Scheme 4. Racemization of 4-Piperidone 4
was found to be the deprotonation of the α-C−H of the
pyridinium salt13 by the lithiated benzimidazole and resulted in
regeneration of 10−20% of N-Ts benzimidazole 5. All attempts
to improve selectivity through examination of reaction
parameters (bases, temperature, concentration, addition rates,
etc.) were unfruitful. Nevertheless, the product could be
isolated in high purity (>98%) by crystallization from isopropyl
acetate in 75% yield.
Reduction of enone 8 to the corresponding ketone and
removal of the Ts group were required to prepare 4-piperidone
4. The conjugate reduction of 8 proved to be challenging due
to competitive over-reduction to the saturated alcohol.
Following experimentation, we found that LiAl(O-t-Bu)3H in
the presence of CuBr14 gave excellent selectivity with less than
2% of the over-reduced alcohol observed (Scheme 3).
Following the investigation of commercially available trans-
aminases, we were delighted to find several enzymes that
catalyzed the transamination of 4.18,19 Enzyme ATA-036 was
identified as catalyzing highly efficient transamination with
DKR to afford the desired (2R,4R)-amine 3 in high conversion
and excellent selectivities. Moreover, ATA-036 demonstrated
excellent thermo-stability and catalyzed the transamination of 4
in reactions with temperatures up to 75 °C. Unfortunately, 4-
piperidone 4 was found to be unstable at elevated temperatures
as shown by a control experiment, where in the absence of any
enzyme approximately 56% of 4 decomposed after 23 h at 60
°C. Performing reactions at less than 60 °C minimized
degradation of 4, but required longer reaction times compared
to reactions at higher temperatures. Consequently, a balance
was obtained for the transamination reaction through heating
the reaction at 50 °C for 50−60 h, generating amine 3 in 85%
assay yield with an anti/syn ratio of >10:1 and >99% ee
(Scheme 5).20
Scheme 3. Preparation of 4-Piperidone 4
Scheme 5. Transamination and DKR of 4-Piperidone 4
Optimized conditions were identified when a slight excess of
LiAl(O-t-Bu)3H (1.2 equiv) and only 5 mol % of CuBr were
employed.15 Piperidone 9 could be isolated in 90% yield and
with high purity (>99%) by crystallization from methanol−
H2O. Finally, the N-Ts protecting group of 9 was found to be
labile to both aqueous basic (NaOH) and acidic (HCl)
conditions. While the basic conditions generated over 10% of
the aldol dimer byproducts, the acidic conditions (THF, conc.
861
dx.doi.org/10.1021/ol403630g | Org. Lett. 2014, 16, 860−863

Organic Letters
Letter
The transamination product 3 was found to be highly soluble
in DMSO−H2O and was very difficult to extract with organic
solvents. Thus, we chose to employ an in situ N-Boc protection
of the resulting amine with (Boc)2O. After the transamination
was complete, the DMSO−H2O solution of 3 was treated with
excess (Boc)2O and Et3N to give a mixture of mono- and bis-
Boc compounds 14 and 15, which could be extracted with
EtOAc (Scheme 5). The crude mixture of N-Boc compounds in
EtOAc was then treated with excess tosylic acid for removal of
the Boc groups. Fortunately, the tritosylate salt of 3 was directly
crystallized from the reaction mixture and was isolated in 85%
yield and >99% purity with the undesired syn isomer observed
at <0.1%.
Scheme 7. Synthesis of 1 with Crude 3
With the key penultimate 3 prepared from the enzymatic
transamination, the final step of the synthesis of 1 involved
formation of the urea. Previously, the urea moiety was formed
by coupling of the amine with 4-cyanophenyl isocyanate. The
use of this isocyanate on scale, however, is undesirable due to
cost and stability issues. The corresponding N-carbamoylimi-
dazole reagent, which can be prepared from the desired aniline
and N,N-carbonyldiimidazole (CDI), has been utilized as an
efficient alternative to the isocyanate in urea formations.21
Indeed, we found that N-carbamoylimidazole 17 was easily
prepared by the reaction of 4-aminobenzonitrile 16 with a slight
excess of CDI (1.1 equiv) in toluene at ambient temperature,
and 17 was isolated via direct filtration as a 1:1 mixture with
imidazole in almost quantitative yield (Scheme 6). We also
amination reaction. Product 1 was formed in >90% assay yield
and was isolated similarly by crystallization from acetonitrile in
80% yield and >98% purity.25
In summary, an efficient enzymatic transamination and DKR
of 4-piperidone 4 enabled the development of a concise,
asymmetric synthesis of the SMO inhibitor 1. The entire
synthesis required only five linear steps and produced 1 in 40%
overall yield without the need for chromatography, thus,
rendering this synthesis suitable for large scale commercial
supply of 1.
Scheme 6. Preparation of N-Carbamoylimidazole 17 and
Synthesis of 1 with Isolated 3·(TsOH)3
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental procedures, compound characterization, and
1
copies of H and 13C NMR spectra. This material is available
free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
■
*E-mail: Zhihui.peng@pfizer.com.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
The authors would like to thank Pfizer colleagues Brian Jones
for HRMS analyses; Robert Singer and Frank Busch for helpful
■
́
discussions; and Nick Thomson and Stephane Caron for
management support.
discovered that the addition of only 1 mol % DBU yielded
shortened reaction times from over 12 h to 2−3 h.22,23 The
reaction of 17 (mixture with imidazole) with the tritosylate salt
of 3 proceeded rapidly in the presence of Et3N in THF to
furnish 1 in >95% assay yield. Following aqueous workup, 1
was isolated by crystallization from acetonitrile in 90% yield and
>99% purity.
Notably, we were delighted to find that N-carbamoyl-
imidazole 17 reacted with 3 directly in the aqueous medium
employed in the transamination reaction (Scheme 7), which
eliminated the need for isolation of 3 through the N-Boc/de-
Boc sequence.24 A slight excess of 17 (1.2 equiv) was required
to scavenge the remaining isopropylamine from the trans-
REFERENCES
■
(1) (a) Heretsch, P.; Tzagkaroulaki, L.; Giannis, A. Bioorg. Med.
Chem. 2010, 18, 6613−6624. (b) Rubin, L. L.; de Sauvage, F. Nat. Rev.
Drug Discovery 2006, 5, 1026−1033.
(2) Munchhof, M. J.; Li, Q.; Shavnya, A.; Borzillo, G. V.; Boyden, T.
L.; Jones, C. S.; LaGreca, S. D.; Martinez-Alsina, L.; Patel, N.; Pelletier,
K.; Reiter, L. A.; Robbins, M. D.; Tkalcevic, G. T. ACS Med. Chem.
Lett. 2012, 3, 106−111.
(3) The current scale-up route was enabled from the original
discovery route (ref 2), which produced tens of kilograms of 1 in 14
linear steps with an overall yield of only 4%. In this enabling route,
(2R, 4S)-4-hydroxypipecolic acid 2 was prepared with a modified
literature procedure: Gillard, J.; Abraham, A.; Anderson, P. C.;
862
dx.doi.org/10.1021/ol403630g | Org. Lett. 2014, 16, 860−863

Organic Letters
Letter
Beaulieu, P. L.; Bogri, T.; Bousquet, Y.; Grenier, L.; Guse, I.; Lavallee
́
,
(20) There were no significant differences in both diastereo- and
enantioselectivity with temperatures ranging from 40 to 70 °C.
(21) Rawling, T.; McDonagh, A. M.; Tattam, B.; Murray, M.
Tetrahedron 2012, 68, 6065−6070 and references cited therein.
P. J. Org. Chem. 1996, 61, 2226−2231.
(4) Nugent, T. C.; El-Shazly, M. Adv. Synth. Catal. 2010, 352, 753−
819.
́
(22) Larrivee-Aboussafy, C.; Jones, B. P.; Price, K. E.; Hardink, M. A.;
(5) Matsumura, K.; Zhang, X.; Hori, K.; Murayama, T.; Ohmiya, T.;
Shimizu, H.; Saito, T.; Sayo, N. Org. Process Res. Dev. 2011, 15, 1130−
1137.
McLaughlin, R. W.; Lillie, B. M.; Hawkins, J. M.; Vaidyanathan, R. Org.
Lett. 2010, 12, 324−327.
(23) The isolated solids consisting of 17 and imidazole showed no
sign of degradation after several months at ambient temperature.
(24) For urea formation with N-carbamoylimidzole in water, see:
Padiya, K. J.; Gavade, S.; Kardile, B.; Tiwari, M.; Bajare, S.; Mane, M.;
Gaware, V.; Varghese, S.; Harel, D.; Kurhade, S. Org. Lett. 2012, 14,
2814−2817.
(6) Koszelewski, D.; Tauber, K.; Faber, K.; Kroutil, W. Trends
Biotechnol. 2010, 28, 324−332.
(7) Savile, C. K.; Janey, J. M.; Mundorff, E. C.; Moore, J. C.; Tam, S.;
Jarvis, W. R.; Colbeck, J. C.; Krebber, A.; Fleitz, F. J.; Brands, J.;
Devine, P. N.; Huisman, G. W.; Hughes, G. J. Science 2010, 329, 305−
309.
(8) The transaminase DKR approach has been reported with the
chiral center α to the carbonyl groups; see: Chung, C. K.; Bulger, P.
G.; Kosjek, B.; Belyk, K. M.; Rivera, N.; Scott, M. E.; Humphrey, G. R.;
Limanto, J.; Bachert, D. C.; Emerson, K. M. Org. Process Res. Dev.
2014, 18, 215−227.
(25) A reslurry of the isolated solids was needed to purge the
undesired syn isomer to <1%.
(9) (a) Bravo, F.; Cimarosti, Z.; Tinazzi, F.; Smith, G. E.; Castoldi,
D.; Provera, S.; Westerduin, P. Org. Process Res. Dev. 2010, 14, 1162−
1168. (b) Lee, S. K.; Tambar, U. K.; Perl, N. R.; Leighton, J. L.
Tetrahedron 2010, 66, 4769−4774.
(10) Bull, J. A.; Mousseau, J. J.; Pelletier, G.; Charette, A. B. Chem.
Rev. 2012, 112, 2642−2713.
(11) Other protecting groups on the nitrogen of benzimidazole gave
a lower yield than the tosyl group.
(12) Other pyridinium salts such as iodide, tosylate, methyl sulfate,
and hexafluorophosphate were much less soluble in THF and gave
poor yields.
(13) The pKa of the α-C−H of the 4-methoxypyridinium salt was
measured to be ∼33; see: Wong, F. M.; Capule, C. C.; Chen, D. X.;
Gronert, S.; Wu, W. Org. Lett. 2008, 10, 2757−2760. And the D2O
quench of the reaction confirmed the deprotonation at the α-carbon of
the pyridinium salt.
(14) (a) Focken, T.; Charette, A. B. Org. Lett. 2006, 8, 2985−2988.
(b) Semmelhack, M. F.; Stauffer, R. D.; Yamashita, A. J. Org. Chem.
1977, 42, 3180−3188.
(15) Previously reported procedures required more than 1 equiv of
copper (I) salts; see ref 14.
(16) Alternatively, the free base of 4 could be isolated in 87% yield
with a slightly different procedure; see Supporting Information for
details.
(17) Indirect evidence for the racemization mechanism came from
the N-isopropyl impurity 13 that was formed in the transamination
reaction at 0.2−0.4% level. This impurity was generated by the
transamination of the N-isopropyl piperidone 12, which was produced
by a second retro-aza-Michael of 10 with the elimination of
methylamine followed by a double aza-Michael reaction with
isopropylamine, the amine donor for the transamination reaction:
(18) In the initial screening (30 °C, 19 h), eight (R)-selective
transaminases from Codexis gave >10% assay yields with various
diastereoselectivities (assay, anti/syn ratio): ATA-015 (15%, 1.7),
ATA-016 (21%, 1.1), ATA-024 (48%, 4.1), ATA-025 (41%, 10.3),
ATA-033 (42%, 9.1), ATA-034 (42%, 5.9), ATA-035 (35%, 10.2),
ATA-036 (61%, 8.3). Transaminase ATA-036 was selected for further
evaluation based on activity and diastereoselectivity, and subsequent
chiral analysis showed that reactions with this enzyme were highly
enantioselective and only produced the desired (2R,4R)-amine. Many
of the Codexis (S)-selective transaminases also showed activity to
produce the undesired (2S,4S)-amine.
(19) Both the free base and di-HCl salt of 4 were used in the
transamination reaction with similar results.
863
dx.doi.org/10.1021/ol403630g | Org. Lett. 2014, 16, 860−863