Facile entry to an efficient and practical enantioselective synthesis of a polycyclic cholesteryl ester transfer protein inhibitor

Article abstract of DOI:10.1021/ol501833g

An efficient enantioselective synthesis of the chiral polycyclic cholesteryl ester transfer protein (CETP) inhibitor 1 has been developed. The synthesis was rendered practical for large scale via the development of a modified Hantzsch-type reaction to prepare the sterically hindered pyridine ring, enantioselective hydrogenation of hindered ketone 6 utilizing novel BIBOP-amino-pyridine derived Ru complex, efficient ICl promoted lactone formation, and a BF₃ mediated hydrogenation process for diastereoselective lactol reduction. This efficient route was successfully scaled to produce multikilogram quantities of challenging CETP drug candidate 1.

Full text of DOI:10.1021/ol501833g

Letter
pubs.acs.org/OrgLett
Facile Entry to an Efficient and Practical Enantioselective Synthesis of
a Polycyclic Cholesteryl Ester Transfer Protein Inhibitor
Zhengxu S. Han,*,† Yibo Xu,† Daniel R. Fandrick,† Sonia Rodriguez,† Zhibin Li,† Bo Qu,†
Nina C. Gonnella,† Sanjit Sanyal,† Jonathan T. Reeves,† Shengli Ma,† Nelu Grinberg,† Nizar Haddad,†
Dhileep Krishnamurthy,† Jinhua J. Song,† Nathan K. Yee,† Waldemar Pfrengle,‡ Markus Ostermeier,‡
Jurgen Schnaubelt,‡ Zeno Leuter,‡ Sonja Steigmiller,‡ Jurgen Daubler,‡ Emanuel Stehle,‡
̈
̈
̈
Lukas Neumann,‡ Thomas Trieselmann,§ Patrick Tielmann,∥ Annette Buba,∥ Rainer Hamm,∥
Gunter Koch,∥ Svenja Renner,∥ Juan R. Dehli,∥ Florian Schmelcher,∥ Christian Stange,∥ Jurgen Mack,§
̈
Rainer Soyka,‡ and Chris H. Senanayake†
†Department of Chemical Development, Boehringer Ingelheim Pharmaceuticals Inc., 900 Old Ridgebury Road, P.O. Box 368,
Ridgefield, Connecticut 06877, United States
‡Departments of Chemical Development and §Medicinal Chemistry, Boehringer Ingelheim Pharma GmbH & Co. KG, Birkendorfer
Str. 65, 88397 Biberach/Riss, Germany
∥Global Department Process Development Chemicals, Boehringer Ingelheim Pharma GmbH & Co. KG, Binger Str. 173, 55216
Ingelheim am Rhein, Germany
S
* Supporting Information
ABSTRACT: An efficient enantioselective synthesis of the chiral polycyclic cholesteryl ester transfer protein (CETP) inhibitor 1
has been developed. The synthesis was rendered practical for large scale via the development of a modified Hantzsch-type
reaction to prepare the sterically hindered pyridine ring, enantioselective hydrogenation of hindered ketone 6 utilizing novel
BIBOP-amino-pyridine derived Ru complex, efficient ICl promoted lactone formation, and a BF3 mediated hydrogenation
process for diastereoselective lactol reduction. This efficient route was successfully scaled to produce multikilogram quantities of
challenging CETP drug candidate 1.
reduction, lactonization, and diastereoselective spirocycle
ompound 1 is a clinical candidate used as an inhibitor of
Ccholesteryl ester transfer protein (CETP) for the formation.
treatment of cardiovascular disorders.1,2 It is intended to
reduce the risk of atherosclerosis by improving the HDL/LDL
ratio.2 In order to support early clinical development, a scalable
process to produce 1 was required.
Our retrosynthetic strategy (Scheme 1) focused on the
synthesis of the advanced chiral intermediate 3. The
construction of the challenging chiral spirocycle could be
accomplished by a diastereoselective reduction of lactol 2
directed by the chiral alcohol, and 3 could be prepared either by
asymmetric reduction of 4 or from chiral intermediate 5. The
lactone in 3 or 4 could be accessible by an acid promoted
lactonization with the adjacent ester and olefin. Both 4 and 5
could be derived from the pyridine intermediate 6 which in
turn could be prepared by a four-component coupling reaction
(Hantzsch) of aldehyde 7, ketoester 8, dimedone 9, and an
appropriate ammonium salt.
The target compound 1 possesses a crowded tricyclic
pyridine core with an embedded chiral dihydrofuran moiety.
The discovery route for its synthesis suffers from not only low
yields and low stereoselectivity but also tedious steps for the
construction of the pyridine core.1c While the Hantzsch
reaction has been used extensively for the synthesis of pyridines
and a number of improved protocols that were reported
recently,3 its application to the synthesis of highly substituted
pyridine derivatives remains a challenge. To develop a cost
competitive process, rapid construction of the tricyclic pyridine
core was required. Herein, we describe an efficient
enantioselective synthesis of 1 by building the fully substituted
pyridine in one step followed by a catalytic asymmetric ketone
The initial focus was to develop an efficient route to install
the required carbon framework in a single step for the fully
Received: June 24, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ol501833g | Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
amount of enamine 12 required was due to the competitive
decomposition of 12 at high reaction temperatures. It was
determined that the optimal conditions included performing
the reaction at 90 °C with 2 equiv of enamine 12 in which
complete conversion was obtained in 4 h to furnish
dihydropyridine 13 in 87% isolated yield. Subsequent
aromatization of 13 with DDQ in NMP afforded the fully
substituted pyridine 6 in 93% isolated yield.
Scheme 1. Retrosynthetic Analysis
With the fully substituted pyridine core 6 in hand, the
synthesis of chiral intermediate 3 was investigated (Scheme 3).
Scheme 3. Synthesis of 3 via Lactone−Ketone Reduction
Scheme 2. Synthesis of 6 via Hantzsch-Type Reaction
An initial attempt was to form the spirocyclic γ-lactone 4 via an
intramolecular cyclization of either the ester 6 or acid 14 with
the alkene of the pyran ring. Nonhindered γ-lactone synthesis
from an olefin and acid has been reported although strong acids
are normally required to achieve high conversion.6 Acid 14 was
first investigated for the synthesis of 4, but as anticipated, its
conversion to 4 was found to be difficult. Among the many
acids surveyed,4 only neat H2SO4 or trifluoromethanesulfonic
acid (TfOH) could promote the cyclization. The reaction in
H2SO4 was sluggish, and an 80% conversion was obtained after
4 days at ambient temperature. Attempts to accelerate the
process with elevated temperatures generated many side
products. A clean reaction was observed in neat TfOH, and
the reaction was complete after 12 h at ambient temperature to
afford the desired lactone 4. However, presumably due to the
steric hindrance of pyridine 6, the synthesis of acid 14 by ester
hydrolysis was not a viable process, as the reaction needed to be
conducted in MeOH with LiOH at 120 °C in a pressure
reactor. A more direct protocol was later developed by the
cyclization of ester 6 with neat TfOH in 18 h at rt to afford 4 in
77% isolated yield.4 This effective transformation may account
for the TfOH promoted formation of the carbocationic species
of the alkene, followed by addition of the methoxy or carbonyl
oxygen atom, and the efficiently triflate induced demethylation
of the resultant oxonium species.6b
Due to the hindered environment around the ketone no
reduction was observed when 4 was subjected to asymmetric
hydrogenation or transfer hydrogenation conditions. Thus, we
explored borane reagents for the asymmetric reduction of 4.4,7
The optimal conditions were found using 2 equiv of a borane−
diethylaniline complex (BH3·Et2NPh) with 10 mol % of either
(R)-tert-leucinol or (1R,2S)-1-amino-2-indanol as a ligand in
THF. The reaction completed in 24 h at rt to provide chiral
alcohol 15 in an 80%−85% assay yield and 94:6 er. A higher
selectivity of 97:3 er was obtained when 15 mol % of (1R, 2S)-
substituted pyridine core 6 by a Hantzsch reaction (Scheme 2).
The required aldehyde 7 was prepared from the readily
available pyranone 10 in three steps and a 63% overall yield.4
Preliminary studies for the four-component coupling reactions
were unsuccessful for the parent example. When a mixture of 7,
8, 9, and NH4OAc was subjected to various reported
protocols,3 no desired product 13 was formed, but instead 11
was isolated as the major product. Consequently, the pyridine
synthesis via two-component coupling of 11 and 12 was
investigated.
Condensation of α,β-unsaturated aldehyde 7 with dimedone
9 was catalyzed by either 10 mol % of 2,6-lutidine or L-proline
to provide cyclic product 11 in 78% yield.4,5 Concurrently,
enamine 12 was prepared in 99% yield by reacting ketoester 8
with ammonium carbamate in MeOH.4 Unexpectedly, no
reaction was observed when 11 and 12 were subjected to the
reported reaction conditions in which EtOH, acetonitrile, or
1,2-dichloroethane were used as solvents in the presence or
absence of a Lewis acid additive at different temperatures.3a,b,4
Interestingly, moderate conversion was observed when a
mixture of tricycle 11 and enamine 12 in methyldiglyme was
heated to a higher reaction temperature of 140 °C. A higher
conversion (95%) to 13 was later identified once the reaction
was carried out in refluxing AcOH (100−110 °C). We believed
that the use of AcOH would drive the equilibrium toward the
active coupling intermediate 11′. Furthermore, it was found
that the yield was affected by the amount of enamine 12 and 4
equiv of 12 were required in order to consume 11. The large
B
dx.doi.org/10.1021/ol501833g | Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
1-amino-2-indanol was used. This protocol, however, provided
the product with 15%−20% of byproducts resulting from over-
reduction of the lactone. To obtain pure 15, two recrystalliza-
tions were needed to remove the byproducts, and the isolated
yield of 15 was a modest 55%. Treatment of alcohol 15 with
TBSOTf and 2,6-lutidine afforded the TBS ether 3 in 82%
yield. Use of TBSCl failed to provide the desired product.
Although the above process provided the desired chiral
product 3, it was not preferable for large scale synthesis. The
main drawbacks were the use of corrosive TfOH as solvent for
the lactone formation and the low overall isolated yield. To
overcome these issues, a more effective alternative was
developed by using asymmetric reduction of ketone 6 to install
the chiral alcohol first followed by lactone formation (Scheme
4). Similar to 4, the reduction of 6 by BH3·Et2NPh (1.6 equiv)
the subsquent debromination process generated about 28% side
products. After a survey of other reagents and conditions, it was
eventually found that treatment of 5 with ICl in DCM at rt
afforded the desired product 17b as a diastereomer in 95%
isolated yield, from which the lactone 3 was obtained in
quantitative yield by hydrogenative deiodination with the Pd/C
catalyst in MeOH.
Having successfully prepared the key tetracycle 3 we next
explored the construction of the chiral spirocycle 18 (Scheme
5). Addition of p-CF3−PhLi (ArLi) to 3 was first investigated.
Scheme 5. Synthesis of 1 by Lactol Reduction
Scheme 4. Synthesis of 3 via Ketone−Ester Reduction
The reaction was performed by reverse addition of 3 in THF to
an ArLi solution generated in situ by treatment of ArBr with n-
BuLi at −70 °C. The reaction was completed in 30 min after
warming to −35 °C to afford 2 in 95% yield. However, this
protocol was not feasible for large scale production due to the
limited thermal stability of the ArLi wherein up to 50% was
decomposed after 2 h at −70 °C. In order to obtain complete
conversion more than 3 equiv of the ArBr were required and
the reaction needed to be completed in less than 2 h.
Due to the stability issues with the aryllithium substrate, the
use of the corresponding Grignard reagent p-CF3−PhMgX was
investigated. The initial focus was to identify reaction
conditions for the preparation of the stable Grignard reagent.
Eventually, it was found that p-CF3−PhMgCl can be prepared
by treatment of ArBr with i-PrMgCl/LiCl and 2 equiv of
dioxane in THF.10 Although p-CF3−PhMgCl has been
previously reported as unstable,11 the in situ prepared Grignard
was found to decompose very slowly (<1%/h) at ambient
temperature. Therefore, by using 2 equiv of p-CF3−PhMgCl
the addition reaction reached full conversion in 4 h at 40 °C.
Attempts to isolate the addition product 2 directly were not
successful because it was an oil. Gratifyingly, the desired lactol 2
was easily isolated as a crystalline HCl salt in 96% yield and
high purity.
To complete the synthesis, a diastereoselective reduction of
lactol 2 with various reducing reagents promoted by Lewis
acids was first investigated. The main challenge was to identify
conditions to achieve a significant level of desired diaster-
eoselectivity, presumably induced by the already existing chiral
environment in 2. Initial results showed that the selectivity was
very sensitive to the reaction conditions, specifically, the
reagents, the Lewis acids, and the solvents used.12 Using
Et3SiH, a reagent commonly used for lactol reduction,13 with
TiCl4 in DCM an 83:17 dr of 18 was obtained favoring the
undesired diastereomer ((3S, 9S)/(3R, 9S)). With NaBH3CN
the selectivity was found to depend upon the solvent and Lewis
with (1R, 2S)-1-amino-2-indanol (15 mol %) as the ligand in
THF afforded the desired product 16 in 96.5:3.5 er. Unlike 4,
reduction of 6 gave a clean reaction to afford 16 in 93% isolated
yield and high purity.
To develop a greener and more efficient process, we
investigated the catalytic asymmetric hydrogenation of ketone
6. In contrast to lactone 4, high conversions were obtained
when 6 was subjected to different catalytic conditions with
commercially available RuCl2(diphosphine) (diamine) com-
plexes.8,9b The best result was found using 0.5 mol % of a
complex formed between RuCl2(tol-BINAP) with 2-amino-
methylpyridine (Ampy) in EtOH which afforded the chiral
alcohol 16 in quantitative yield and 98:2 er. Attempts to
decrease the catalyst loading resulted in low conversion. A more
efficient catalytic system was later identified by using our in-
house developed BIBOP ligand.9a Using the RuCl2(MeO−
BIBOP)−(Ampy) complex in isopropanol under 300 psi H2 at
20 °C, the reaction was complete in 20 h to afford 16 in 90%
isolated yield and >99:1 er. Furthermore, the yield and
stereoselectivity were not affected even by lowering the catalyst
loading to 0.01 mol %.
Treatment of 16 with TBSCl instead of TBSOTf in the
presence of t-BuOK afforded 5 in 91% isolated yield. The
lactone formation using the ester directly was studied. Since a
strong acid was not compatible with the substrate, the use of a
halogen-promoted lactonization was investigated. Using NIS or
I2 failed to deliver the desired product. Although NBS in
dioxane/water afforded the lactone 17a in 80% isolated yield,
C
dx.doi.org/10.1021/ol501833g | Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
acid employed.12 The best selectivity of 80:20 dr favoring the
desired diastereomer was obtained using TiCl4·(THF)2 in
THF. However, the reaction required 4 equiv of NaBH3CN
and a Lewis acid to achieve high conversion.
Gratifyingly, an improved selectivity of 85:15 dr was
observed with the hydrogenation in the presence of either
BF3·OEt2 or BF3·THF in DCM with 10 mol % wet Pd/C. The
selectivity was further increased to 89:11 dr when 3 mol % of
dry Pd/C was used, and the reaction was completed in 7.5 h at
40 °C to afford 18 in >99% conversion, which was used directly
in the next step.
A one-pot protocol for the hydrogenation of 2 and
deprotection of 18 was explored with the goal to avoid the
use of TBAF.14 Directly heating the reaction mixture after
hydrogenation provided low conversion and numerous
dehydration byproducts. A survey of other additives such as
acids and bases showed that p-TsOH·H2O gave a complete
reaction in 7.5 h at 35 °C to afford the crude desilylation
product 1 in 89:11 dr.
Recrystallization of different salts of 1 was studied in order to
upgrade its diastereomeric and enantiomeric purity. Recrystal-
lizations of the HCl, H3PO4, and HBr salts of 1, for example,
failed to enrich the product. An efficient protocol was finally
identified by the slow addition of H2SO4 to a solution of 1 in
THF, from which 1·H2SO4 was isolated in 75% overall yield in
three steps with enrichment of the er from 96:4 (from borane
reduction) to >99.5:0.5 and the dr from 89:11 to >99.5:0.5.
The final drug substance 1·HCl was prepared by treating the
free base 1 in acetone with 4 N HCl in dioxane to provide 1·
HCl in 92% yield and 99% purity. The stereochemistry of 1 was
unambiguously confirmed by a single crystal X-ray structure as
shown in Figure 1. This process was utilized to prepare 4.2 kg
of 1·HCl.
material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
■
*E-mail: steve.han@boehringer-ingelheim.com.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Scott Pennino and Keith McKellop (Boehringer
Ingelheim Pharm. Ridgefield, CT, USA) for HRMS analysis
and Keith Fandrick (Boehringer Ingelheim Pharm. Ridgefield,
CT, USA), Thomas Hampel (Boehringer Ingelheim Pharm.
Co., Ingelheim, Germany), and George Han (Johns Hopkins
University, Baltimore, MD, USA) for proofreading the
manuscript.
REFERENCES
■
(1) (a) Wagner, H.; Berta, D.; Fuchs, K.; Giovannini, R.; Hamprecht,
W.; Konetzki, I.; Streicher, R.; Trieselmann, T. WO2011/101424, Aug
25, 2011. (b) ibid, WO2012/110599, Aug 23, 2012. (c) Refer to
Scheme 1 in the Supporting Information (SI) for the discovery route
synthesis of 1.
(2) Mantlo, N. B.; Escribano, A. J. Med. Chem. 2014, 57, 1−17.
(3) (a) Evans, C. G.; Gestvicki, J. E. Org. Lett. 2009, 11, 2957−2959.
(b) Li, J.; He, P.; Yu, C. Tetrahedron 2012, 68, 4138−4144. (c) Saini,
A.; Kumar, S.; Sandhu, J. S. J. Sci. Ind. Res. 2008, 67, 95−111.
(d) Hsung, R. P.; Kurdyumov, A. V.; Sydorenko, N. Eur. J. Org. Chem.
2005, 25−44.
(4) See the SI for the synthesis of 7 and 11, the results under
different conditions for the Hantzsch reaction (Tables 1 and 2),
lactone formation (Tables 3 and 4), and ketone reduction (Table 5).
(5) Goswami, P.; Das, B. Tetrahedron Lett. 2009, 50, 897−900.
(6) (a) Gil, S.; Parra, M.; Rodriguez, P.; Segura, J. Mini-Rev. Org.
Chem. 2009, 6, 345−358. (b) Cermak, S. C.; Isbell, T. A. J. Am. Oil
Chem. Soc. 2000, 77, 243−248. (c) Munoz, M. P.; Lloyd-Jones, J. C.
̃
Eur. J. Org. Chem. 2009, 516−524.
(7) Brunin, T.; Cabou, J.; Bastin, S.; Brocard, J.; Pelinski, L.
Tetrahedron: Asymmetry 2002, 13, 1241−1243.
(8) (a) Ohkuma, T.; Hattori, T.; Ooka, H.; Inoue, T.; Noyori, R. Org.
Lett. 2004, 6, 2681−2683. (b) Noyori, R.; Ohkuma, O. Angew. Chem.,
Int. Ed. 2001, 113, 40−75 and references cited therein.
(9) (a) Tang, W.; Qu, B.; Capacci, A. G.; Rodriguez, S.; Wei, X.;
Haddad, N.; Narayanan, B.; Ma, S.; Grinberg, N.; Yee, N. K.;
Krishnamurthy, D.; Senanayake, C. H. Org. Lett. 2010, 12, 176−179.
(b) Rodriguez, S.; Qu, B.; Fandrick, K. R.; Bunono, F.; Haddad, N.;
Xu, Y.; Herbage, M. A.; Zeng, X.; Ma, S.; Grinberg, N.; Lee, H.; Han,
Z. S.; Yee, N. K.; Senanayake, C. H. Adv. Synth. Catal. 2014, 356, 301−
307.
Figure 1. Single X-ray crystal structure of 1.
In summary, an efficient enantioselective synthesis of the
CETP inhibitor 1 was developed. Its main features include the
development of an efficient Hantzsch-type reaction for the
construction of the fully substituted pyridine tetracyclic core, a
highly enantioselective ketone reduction via a newly developed
BIBOP−amino−pyridine derived Ru complex, a novel ICl
promoted lactonization, and BF3 mediated diastereoselective
lactol reduction. The asymmetric process was successfully
scaled to produce multikilogram quantities of 1 to support
preclinical drug development.
(10) Krasovskiy, A.; Straub, B. F.; Knochel, P. Angew. Chem., Int. Ed.
2006, 45, 159−162.
(11) Tang, W.; Sarvestani, M.; Wei, X.; Nummy, L. J.; Patel, N.;
Narayanan, B.; Byrne, D.; Lee, H.; Yee, N. K.; Senanayake, C. H. Org.
Process Res. Dev. 2009, 13, 1426−1430.
(12) See Table 6 in the SI for the results on lactol reduction under
different conditions.
(13) Fugedi, P. In The Organic Chemistry of Sugar; Levy, D. E.,
̈
Fugedi, P., Ed.; CRC Press: 2006; p 90.
̈
(14) Details are provided in the SI.
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental procedures, detailed reaction condition survey
results, and spectroscopic data for all new compounds. This
D
dx.doi.org/10.1021/ol501833g | Org. Lett. XXXX, XXX, XXX−XXX