Synthesis of PPAR agonist via asymmetric hydrogenation of a cinnamic acid derivative and stereospecific displacement of (S)-2-chloropropionic acid

Article abstract of DOI:10.1021/ol050367e

(Chemical Equation Presented) The synthesis of the peroxime proliferator activated receptor (PPAR) α,γ-agonist (1) was accomplished with high enantio- and diastereoselectivity by employing an asymmetric hydrogenation strategy, of an α-alkoxy cinnamic acid

Full text of DOI:10.1021/ol050367e

ORGANIC
LETTERS
2005
Vol. 7, No. 10
1947-1950
Synthesis of PPAR Agonist via
Asymmetric Hydrogenation of a
Cinnamic Acid Derivative and
Stereospecific Displacement of
(S)-2-Chloropropionic Acid
Ioannis N. Houpis,*,† Lawrence E. Patterson,‡ Charles A. Alt,‡ John R. Rizzo,‡
Tony Y. Zhang,‡ and Michael Haurez†
Chemical Product Research and DeVelopment, Lilly DeVelopment Center, Rue
Granbonpre 11, B-1348, Mont. St. Guibert, Belgium, and Lilly Technology Center,
B-110, Indianapolis, Indiana 46285
houpis_ioannis@lilly.com
Received February 21, 2005
ABSTRACT
The synthesis of the peroxime proliferator activated receptor (PPAR)
r,γ-agonist (1) was accomplished with high enantio- and diastereoselectivity
by employing an asymmetric hydrogenation strategy, of an -alkoxy cinnamic acid derivative, to set the C-2 chiral center. A diastereospecific
r
SN2 displacement under mild basic conditions established the C-10 stereochemistry without any detectable racemization of the two epimerizable
chiral centers.
Peroxime proliferator activated receptor (PPAR) agonists
have attracted significant attention in the pharmaceutical
industry due to their potential usefulness in the treatment of
type 2 diabetes and dislepedimia,1 and several compounds
of this class are now in various stages of development.2 We
have recently described the synthesis of the PPAR agonist
1 via a sequence that utilized the starting material 2 (Scheme
1), which was readily available via several pathways from
our laboratory and recent literature precedence.3 Particular
features of this synthesis were the ethylation of 2 to give
the ethyl ether 3 via a variation of the Williamson ether
synthesis. Despite the fact that the chemistry performed well
on the kilogram scale, we were concerned about the
(2) (a) Haigh, D.; Birrell, H. C.; Cantello, B. C. C.; Eggleston, D. S.;
Haltiwanger, R. C.; Hindley, R. M.; Ramaswamy, A.; Stevens, N. C.
Tetrahedron: Asymmetry 1999, 10, 1353. (b) Liu, K. G.; Smith, J. S.;
Ayscue, A. H.; Henke, B. R.; Lambert, M. H.; Leesnitzer, L. M.; Plunket,
K. D.; Willson, T. M.; Sternbach, D. D. Bioorg. Med. Chem. J. 2001, 11,
2385. (c) Sauerberg, P.; Pettersson, I.; Jeppeses, L.; Bury, P. S.; Mogensen,
J. P.; Wassermann, K.; Brand, C. L.; Sturis, J.; Woeldike, H. F.; Fleckner,
J.; Andersen, A.-S. T.; Mortesen, S. B.; Svensson, L. A.; Rasmussen, H.
B.; Lehmann, S. V.; Polivka, Z.; Sindelar, K.; Panajotova, V.; Ynddal, L.;
Wulff, E. M. J. Med. Chem. 2002, 45, 789. (d) Deusen, H.-J.; Zundel, M.;
Valdois, M.; Lehmann, S. V.; Weil, V.; Hjort, C. M.; Ostergaard, P. R.;
Marcussen, E.; Ebdrup, S. Org. Process Res. DeV. 2003, 7, 82.
† Lilly Development Center.
‡ Lilly Technology Center.
(1) (a) Lehman, J. M.; Moore, L. B.; Smith-Oliver, T. A.; Wilkinson,
W. O.; Willson, T. M.; Kliewer, S. A. J. Biol. Chem. 1995, 270, 12953. (b)
Staels, B.; Dallongeville, J.; Auwerx, J.; Schoonjans, K.; Leitersdorf, E.;
Fruchart, J.-C. Circulation 1998, 41, 5020. (c) Oliver, M. F.; Heady, J. A.;
Morris, J. N.; Cooper, J. Lancet 1984, 2, 600.
10.1021/ol050367e CCC: $30.25
© 2005 American Chemical Society
Published on Web 04/15/2005

The key intermediate 11 was synthesized via the route
shown in Scheme 2. A mixture of the readily available
Scheme 1. Initial Approach to 5
Scheme 2. Optimized Synthesis of 5
robustness of this reaction, as some laboratory work indicated
significant variability in the retention of optical purity of the
existing C-2 center. Furthermore, the key coupling of 3 and
4, to produce intermediate 5, also posed particular challenges.
Namely, the reaction utilized pyrrophoric Na0 and, more
importantly, had to be performed at substantial dilution (30
mL/g) in order to avoid the formation of two impurities
resulting from single electron transfer reaction, 6 and 7.
Moreover, we were particularly concerned about the effect
of the Na0 quality, particularly its potassium content, on the
formation of 6 and 7. Therefore, for economic reasons and
due to the usefulness of the R-alkoxy, and particularly ethoxy
and methoxy, ester functionality in PPAR agonists and
coagonists as well as other biologically significant mol-
ecules,4 we sought a general solution to the introduction of
that functionality without the intermediacy of the alcohol 2
and the need for a difficult alkylation reaction.
benzaldehyde derivative 8 and ethoxy ethyl acetate 9 was
added to a cold (-40 °C) solution of potassium t-butoxide
in THF to produce the adduct, which was treated in situ with
trifluoroacetic anhydride to give 10. Compound 10 was not
isolated; instead, the reaction was warmed to ambient
temperature and treated with an additional amount (4 equiv)
of potassium t-butoxide to effect the elimination to the
corresponding R,â-unsaturated ester. This crude material was
hydrolyzed in MeOH-5 N NaOH to afford the desired acid,
11, in ca. 50-55% overall yield. It is worth noting that only
the (Z)-isomer was detected in the crude reaction mixture.
With the starting material in hand, we pursued the asym-
metric hydrogenation reaction.
Despite the explosion in the development of chiral ligands
for the asymmetric hydrogenation of a variety of olefin
derivatives, relatively few examples exist for the successful
reduction of R-alkoxy cinnamoyl derivatives.
More commonly, these enoate asymmetric reductions5
contain at the R-position an acetamido, acetyl, or other
functionality (e.g., MEM protecting group) that coordinates
the transition metal and thus provides an ordered transition
In this Letter, we would like to describe our synthesis of
intermediate 5 via an asymmetric hydrogenation of 11 and
subsequent coupling conditions with 4 to this key intermedi-
ate while preserving the epimerizable chiral centers and
avoiding the formation of the single electron transfer
byproducts.
(5) For recent reviews describing a plethora of ligand and metal systems,
see: (a) Noyori, R.; Kitamura, M.; Ohkuma, T. PNAS 2004, 101, 5356. (b)
Studer, M.; Blaser, H.-U.; Exner, C. AdV. Synth. Catal. 2003, 345, 45. (c)
Blaser, H.-U.; Malan, C.; Pugin, B.; Spindleer, F.; Steiner, H.; Studer, M.
AdV. Synth. Catal. 2003, 345, 103. (d) Crepy, K. V. L.; Imamoto, T. AdV.
Synth. Catal. 2003, 345, 79. (e) Sturm, T.; Weissensteiner, W.; Spindler,
F. AdV. Synth. Catal. 2003, 345, 160. (f) Genet, J.-P. Acc. Chem. Res. 2003,
36, 908.
(3) (a) Aikins, J.; Haurez, M.; Rizzo, J.; Van Hoeck, J. P.; Brione, W.;
Kestemont, J.-P.; Stevens, C.; Lemaire, X.; Forst, M.; Houpis, I. N. J. Org.
Chem., in press. (b) Liang, T.; Deng, L. J. Am. Chem. Soc. 2002, 124,
2870.
(4) Linderberg, M. T.; Moge, M.; Sivadasan, S. Org. Process Res. DeV.
2004, 8, 838. Lindenber, M. T.; Moge, M.; Sivadasan, S. Org. Process
Res. DeV. 2004, 8, 838.
1948
Org. Lett., Vol. 7, No. 10, 2005

BiNAP, BiPHEP, and DuPhos family of ligands that have
been successfully used in the reduction of MEM-protected
hydroxymethyl enoates in the multiton scale.6 Unfortunately,
in our case, these ligands afforded the product in low
conversion and low enantiomeric excess under a variety of
conditions. Furthermore, the ferrotane analogue 12 gave 81%
conversion and <20% ee.
As a consequence, a hydrogenation screen of over 250
catalysts and conditions was performed with our catalyst
libraries under our standard conditions.7 Figure 1 shows some
representative results of our screen.
Similarly to the Duphos family of ligands, Jopishos and
BiNAP8 derivatives (13 and 15-18 in Figure 1) gave either
low conversion or low enantiomeric excess under our
conditions. Gratifyingly, the Rh-based systems using the
aforementioned ligand motifs afforded conversions and
asymmetric induction between 70 and 80% ee. It was the
MandyPhos family of ligand that gave us the first truly useful
enantiomeric excess (ligands 19 and 20, 88% ee, Figure 1).
Table 1 shows a summary of our initial screen with
MandiPhos ligand 19.
Table 1. 11 w 21 with MandyPhos[(COD)2Rh]OTf 19 at 25
°C
solvent
pressure
conversion
% ee
TON
2800
TOF
140
MeOH
MeOH
EtOH
700 psi
200 psi
700 psi
700 psi
700 psi
700 psi
700 psi
100%
59%
88%
86%
83%
79%
71%
83%
79%
100%
100%
100%
100%
100%
iPrOH
dioxane
CH3CN
c-hexane
However, it was finally the Valphos ligand 22 (Scheme
3) that gave the best results upon optimization of the solvent
(MeOH), pressure (50 psi, Table 1 shows little effect of
pressure on ee), and additives (Et3N and MeONa). To date,
Scheme 3. Hydrogenation of 11 with Rh-Valphos
Figure 1. Results of the hydrogenation screen.
state that enhances enantioselectivity. To our knowledge,
R-alkoxy cinnamic acid derivatives have not been widely
described.
Our initial attempts to effect the desired transformation
utilized cationic Rh(I) and Ru(I) catalyst precursors and the
Org. Lett., Vol. 7, No. 10, 2005
1949

the finalized process involves the addition of a solution of
[(NDB)2Rh]BF4 and the ligand (1:1.1 ratio) to a mixture of
the acid and MeONa (10 mol %) in degassed MeOH. Under
these conditions, high pressures are not required as in the
case of the MandyPhos ligand. Therefore, the homogemeous
mixture was treated with H2 at 50 psi at ambient temperature
for 16 hours to afford the crude acid, which was transformed
to the product 21 in 78% yield.
Capillary electrophoresis and chiral HPLC methodologies
established the enantiomeric excess at ca. 92%. Although
our final process to the synthesis of 1 can easily accom-
modate such optical purity, an upgrade to ca. 98% ee can
be obtained by formation of the D-alaninol salt in ethanol
(94% yield). The latter allowed us to establish the absolute
stereochemistry of 21 by X-ray defraction analysis. It is worth
noting that in the absence of MeONa, only 33% of 21 was
produced, indicating that formation of the carboxylate is
essential for the success of this process. Release of the salt
and esterification with iPrOH and catalytic H2SO4, followed
by heterogeneous catalytic hydrogenation with 10% Pd/C
in EtOH, afforded 3 in nearly quantitative yield. No epimer-
ization was obserVed at the ciral center during the esteri-
fication.
With the latter in hand, we set out to identify conditions
that would effect the desired coupling reaction with complete
inversion at the (S)-2-chloropropionic (4) acid center while
avoiding epimerization of the starting materials or product
and the possibility for single election transfer.
A number of bases and solvents were evaluated in the
coupling of 3 and 4 in an automated fashion in a Chem-
Speed multireactor system. The bases used included K2CO3,
Cs2CO3, AcONa, NaOH (under phase-transfer catalysis
conditions), LDA KHMDS, NaHMDS, tBuOK, and tAmyl-
ONa, while solvents and solvent mixtures varied from THF,
toluene, and 2-methyl-THF to DMSO, methyl isobutyl
ketone, and CH3CN.9
Finally, we discovered that TMSONa in THF afforded
superior results to all the methods investigated, including
the original Na0 procedure (Scheme 2). Indeed deprotonation
of the phenol with 2 equiv of TMSONa at ambient temper-
ature in THF produced the phenoxide as a homogeneous
solution. Warming of the solution to 45 °C with slow (S)-
2-chloropropionic acid addition resulted in a thin slurry of
the sodium chloropropionate. The reaction proceeded to
completion overnight to give a good yield of the desired
compound with complete retention the optical purity at both
chiral centers.10 Solvent and counterion screens indicated
that the original conditions were optimal. For example,
TMSOK had the same reactivity in THF; however, it is more
expensive and less soluble in THF than the Na analogue.
Similarly, nonpolar and polar aprotic solvents gave lower
yields of the coupling product or an increased number of
impurities. From an operational standpoint, the reaction is
much more robust, as the use of the homogeneous base
avoided some of the issues observed in our previous
procedure such as coating of the Na metal by the precipitation
of Na-chloropropionate. It is worth noting that none of the
SET reaction products 6 or 7 were detected and the reaction
no longer exhibited any sensitivity in the concentration.
ReactIR experiments have established that the reaction does
not proceed via an R-lactone intermediate,11 while correlation
experiments have established the inversion of configuration
at the propionate center.
Diastereomeric upgrade and chemical purification of 5 can
be easily accomplished via the (R)-(+)-naphthylethylamine
salt in isopropyl acetate. That intermediate can be used to
conclude the synthesis of the final active pharmaceutical
ingredient, 1, as we have described previously.3
In conclusion, we have been able to complete the synthesis
of 1 through the direct access of the cinnamic acid derivative
21 via an asymmetric hydrogenation reaction. We have also
demonstrated the stereospecific SN2 displacement of (S)-2-
chloropropionic acid by phenoxide using TMSONa as the
base.
Despite these efforts, the results were disappointing,
ranging from no reaction with the lower pKa bases to
extensive hydrolysis (NaOH) to significant epimerization at
both C-2 and C-10 with the stronger bases.
Supporting Information Available: Experimental pro-
cedures to the final product 1, NMR spectra and chiral HPLC
spectra, X-ray data for the D-alaninol salt of 21, and proof
of the absolute configuration of 1. This material is available
free of charge via the Internet at http://pubs.acs.org.
OL050367E
(6) Bulliard, M.; Laboue, B.; Lastanet, J.; Roussiasse, S. Org. Process
Res. DeV. 2001, 5, 438.
(7) Screens were typically performed on 0.05 mmol of substrate in MeOH
at 200 psi with 2 mol % catalyst precursor and 2 mol % ligand, at ambient
temperature on an Endevor multireactor system
(8) BiNAP itself afforded 60 and 40% conversion with Rh(I) and Ru(I),
respectively. The product was racemic.
(10) Commercially available (S)-2-chloropropionic acid was determined
to be of 96% ee. Within the limit of detection of our HPLC method (ca.
0.1%), there was no erosion of the enantiomeric excess, indicating complete
inversion of stereochemistry at C-10. The C-2 center was similarly
unaffected.
(11) For formation and identification of R-lactones, see: Chapman, O.
L.; Wojtkowski, P. W.; Adam, W.; Rodriquez, O.; Rucktaeschel, R. J. Am.
Chem. Soc. 1972, 94, 1365. R-Lactones have also been postulated in
solvolysis reactions of R-bromo propionates: Grunwald, E.; Winstein, S.
J. Am. Chem. Soc. 1948, 70, 841.
(9) For examples of displacement of activated lactic acid or nonracemic
2-halo propionic acid derivatives, see: (a) Effenberger, F.; Burkard, U.;
Willfahrt, J. Angew. Chem. 1983, 95, 50. (b) Burkard, U.; Effenberger, F.
Chem. Ber. 1986, 119, 1594. (c) Sato, T.; Otera, J. J. Org. Chem. 1995,
60, 2627 and references therein. (d) Otera, J.; Nakazawa, K.; Sekoguchi,
K.; Orita, A. Tetrahedron 1997, 53, 13633. (e) Larcheveque, M.; Petit, Y.
Synthesis 1991, 162. (f) Nestler, H. J.; Hoerlein, G.; Handte, R.; Bieringer,
H.; Schwerdtle, F.; Langelueddeke, P.; Frisch, P. U.S. Patent US005712226A,
1998, and references therein.
1950
Org. Lett., Vol. 7, No. 10, 2005