Rhodium-catalyzed and zinc(II)-triflate-promoted asymmetric hydrogenation of tetrasubstituted α,β-unsaturated ketones

Article abstract of DOI:10.1021/ol203264a

The asymmetric hydrogenation of tetrasubstituted α,β-unsaturated ketones has been accomplished using an in situ formed rhodium-Josiphos catalyst. The reaction is enhanced by addition of catalytic zinc(II) triflate, which significantly improves turnover fr

Full text of DOI:10.1021/ol203264a

ORGANIC
LETTERS
2012
Vol. 14, No. 4
1038–1041
Rhodium-Catalyzed and Zinc(II)-Triflate-
Promoted Asymmetric Hydrogenation of
Tetrasubstituted r,β-Unsaturated
Ketones
Joel R. Calvin,† Michael O. Frederick,† Dana L. T. Laird,‡ Jacob R. Remacle,†,§ and
Scott A. May*,†
Chemical Product Research and Development, and Discovery Chemistry Research and
Technology, Eli Lilly and Company, Indianapolis, Indiana 46285, United States
may_scott_a@lilly.com
Received December 20, 2011
ABSTRACT
The asymmetric hydrogenation of tetrasubstituted R,β-unsaturated ketones has been accomplished using an in situ formed rhodium-Josiphos
catalyst. The reaction is enhanced by addition of catalytic zinc(II) triflate, which significantly improves turnover frequency while suppressing
epimerization of the products.
Homogeneous asymmetric hydrogenation of olefins is a
powerful tool for the synthesis of enantiopure chiral
materials.1 Numerous examples of this general method-
ology exist in the synthesis of molecules of pharmaceutical
importance.2 The selective reduction of R,β-unsaturated
ketones is an appealing approach to the synthesis of
optically active ketone products.3 The products of such
transformations may contain one or two new stereocen-
ters depending upon the degree of substitution in the start-
ing material. Unlike more widely studied systems such as
R,β-unsaturated acids and esters, enones require an addi-
tional degree of chemoselectivity such that the carbonyl
functionality is not reduced in either the starting substrate
or product.4 Several research groups have independently
reported the asymmetric reduction of trisubstituted en-
ones with iridium5ꢀ7 and ruthenium-based8 catalysts.
† Chemical Product Research and Development, Eli Lilly and Company
‡ Discovery Chemistry Research and Technology, Eli Lilly and Company
§ Dow Corning Corporation, 3901 S. Saginaw Rd., Midland, Michigan
48640, United States
(4) For examples of chemo- and stereoselective reductions of enones
to provide allylic alcohols, see: (a) Ohkuma, T.; Koizumi, M.; Doucet,
H.; Pham, T.; Kozawa, M.; Murata, K.; Katayama, E.; Yokozawa, T.;
Ikariya, T.; Noyori, R. J. Am. Chem. Soc. 1998, 120, 13529. (b) Xie, J.-H.;
Liu, X.-Y.; Xie, J.-B.; Wang, L.-X.; Zhou, Q.-L. Angew. Chem., Int. Ed.
2011, 50, 7329.
(5) (a) Lu, S.-M.;Bolm,C. Chem.;Eur. J. 2008, 14, 7513. (b) Lu, S.-M.;
Bolm, C. Angew. Chem., Int. Ed. 2008, 120, 8920.
(6) Lu, W.-J.; Chen, Y.-W.; Hou, X.-L. Angew. Chem., Int. Ed. 2008,
47, 10133.
(7) Rageot, D.; Woodmansee, D. H.; Pugin, B.; Pfaltz, A. Angew.
Chem., Int. Ed. 2011, 50, 9598.
(8) (a) Massonneau, V.; Maux, P. L.; Simonneaux, G. Tetrahedron
Lett. 1986, 27, 5497. (b) Massonneau, V.; Maux, P. L.; Simonneaux,
G. J. Organomet. Chem. 1987, 327, 269.
(1) A classic example of asymmetric alkene reduction is the synthesis
of L-DOPA: Knowles, W. S.; Sabacky, M. J. J. Chem. Soc., Chem.
Commun. 1968, 1445. Also see Ager, D. J. Enantioselective Alkene
Hydrogenation: Introduction and Historic Overview. In Handbook of
Homogeneous Hydrogenation; de Vries, J. G., Elsevier, C., Eds.; Wiley-
VCH: Weinheim, Germany, 2007; Vol 2, pp 745ꢀ772.
(2) See: Asymmetric Catalysis on Industrial Scale, 2nd ed.; Blaser, H.-U.,
Federsel, H.-J., Eds.; Wiley-VCH: Weinheim, Germany, 2006; Vol. 2.
(3) (a) Moritani, Y.; Appella, D. H.; Jurkauskas, V.; Buchwald, S. L.
J. Am. Chem. Soc. 2000, 122, 6797. (b) Lipshutz, B. H.; Servesko, J. M.;
Petersen, T. B.; Papa, P. P.; Lover, A. A. Org. Lett. 2004, 6, 1273. (c)
Ouellet, S. G.; Walji, A.; Macmillan, D. W. C. Acc. Chem. Res. 2007, 40,
1327.
r
10.1021/ol203264a
Published on Web 01/30/2012
2012 American Chemical Society

However, examples of asymmetric reductions of the cor-
responding tetrasubstituted enones are rare.9ꢀ11 In con-
junction with the synthesis of ERβ agonist LY500307, we
required the stereoselective reduction of enone 1 to satu-
rated ketone 2 (Scheme 1). Attempts to reduce 1 under
several known conditions failed, which highlighted the
need to develop methodology that expanded the scope of
this reaction type. We report herein the development of a
highly chemo- and enantioselective rhodium-catalyzed
asymmetric reduction of tetrasubstituted enones. A key
feature of this method is the use of catalytic zinc triflate,
which significantly improves turnover frequency and sup-
presses epimerization in the R-position of the products.
The mechanistic landscape for rhodium-catalyzed asym-
metric hydrogenation is dominated by studies associated
with N-acyldehydroamino acids and derivatives thereof.16
Applying this previous work, a catalytic cycle for enone1 is
shown in Scheme 2. Key dihydride complex IV can be
formed from complex III or solvate dihydride II. Evidence
supporting either the solvate dihydride (II)17 or unsatu-
rated solvate (I)18 has not been studied for enone systems.
We speculated that the combination of severe steric envi-
ronments of both the ligandꢀmetal complex and the
substrate were contributing factors to slow substrate metal
coordination to form complex III or IV.
Scheme 2. Plausible Catalytic Cycle for Enone Hydrogenation
Scheme 1. Synthesis of LY500307 via Asymmetric Hydrogena-
tion of Enone 1
Catalyst screening with enone 1 led to the identification
of a rhodium complex with ligand 3 (Figure 1),12 which
provided high stereoselectivity (>95% ee) and conversion
(>95%).13 High throughput screening14 led to the identi-
fication of the rhodium complex of ligand 4, which was
also highly stereoselective (95% ee) but superior to 3 with
regard to chemoselectivity and stability across a wider range
of reaction conditions. Efforts to further develop both leads
were disappointing, however, as catalyst loadings were too
high for practical development.15
A screen of additives revealed that several Lewis acids
dramatically improved the rate of the reaction (Table 1).
For example, 5 mol % Zn(OTf)2 afforded full conversion
at a catalyst loading of S/C 2000, while the control reaction
did not proceed. Furthermore, whenzinc triflate wasadded
to a control experiment after 16 h (1ꢀ2% conversion), the
reaction proceeded at an identical rate to that of a standard
reaction where zinc triflate was present at the outset
(Figure 2). This demonstrated that the catalyst was stable
over long reaction times, but turnover frequencywaslow.19
We observed that Lewis acids derived from superacids
such as triflic acid or tetrafluoroboric acid led to in-
creased turnover, while others such as zinc chloride and
(14) This supplemental high throughput screening was conducted by
Dr. Felix Spindler and Dr. Erhard Bappert at Solvias AG, Basel,
Switzerland. This screen was designed to further examine the ferrocene-
based scaffold that had been previously identified.
Figure 1. Optimized ligands from screening.
(15) Maximum S/C was 100 for 3 and 200 for 4. While this may be of
some utility for small scale material production, these loadings are easily
an order of magnitude too high for practical process development.
(16) Dehydroaminoacids such as (Z)-β-amidocinnamate (MAC)
have a different substrateꢀmetal binding relationship than that of
R,β-unsaturated ketones. This difference is manifested in a metallocycle
V, which is bound to the β-carbon instead of the R-carbon as in MAC.
(17) Gridnev, I. D.; Imamoto, T.; Hoge, G.; Kouchi, M.; Takahashi, H.
J. Am. Chem. Soc. 2008, 130, 2560.
(18) There is still considerable question concerning the sequence of
events involved in the formation of complex IV. For leading references,
see: (a) Chan, A. S. C.; Halpern, J. J. Am. Chem. Soc. 1980, 102, 838. (b)
Gridnev, I. D.; Imamoto, T. Acc. Chem. Res. 2004, 37, 633.
(9) Asymmetric hydrogenation of unfunctionalized tetrasubstituted
alkenes is known; see: Roseblade, S. J.; Pfaltz, A. Acc. Chem. Res. 2007,
40, 1402.
(10) Chemoselective reduction of tetrasubstituted enones has been
reported using Rh/Al. See: Yamaguchi, M.; Nitta, A.; Reddy, R. S.;
Hirama, M. Synlett 1997, 177.
(11) An example of ruthenium-catalyzed asymmetric hydrogenation
of endocylic tetrasubstituted enones is the synthesis of Jasmone odor-
ants. See: Saudan, L. A. Acc. Chem. Res. 2007, 40, 1309 and references
therein. Similarly, Rh-catalyzed asymmetric reduction of trisubstituted
endocyclic enones has also been reported. See: Ohshima, T.; Tadaoka,
H.; Hori, K.; Sayo, N.; Mashima, K. Chem.;Eur. J. 2008, 14, 2060.
(12) Togni, A.; Breutel, C.; Schnyder, A.; Spindler, F.; Landert, H.;
Tijani, A. J. Am. Chem. Soc. 1994, 116, 4062.
(19) At S/C 2000, the control experiment without Zn(OTf)2 pro-
gressed to ∼2% after 18 h, while the experiment with Zn(OTf)2 was
complete in 8 h. This represents a >100ꢁ increase in TOF .
(13) See the Supporting Information for the full details of this screen.
Org. Lett., Vol. 14, No. 4, 2012
1039

Other acids such as TFA, MSA, and nitric acid all failed to
promote the reaction as well. When 1 equiv of zinc triflate
wasused(Table1, entry 7), no epimerization wasobserved,
which suggested that triflic acid was not the source of rate
acceleration.22 To further isolate the effect of zinc triflate,
we examined the reaction rates between a standard reac-
tion and one where 10 mol % 2,6-di-tert-butylpyridine was
added as an acid scavenger.23 After 8 h, the conversion was
96% with the acid scavenger, just slightly slower than the
rate of the control reaction (>99%). This data also sug-
gested that both zinc triflate and triflic acid catalyze this
reaction independently. As such, any mechanistic rationale
must account for both possibilities.
The reduction of 1 could be accelerated further by in-
creasing temperature24 and pressure.25 In both cases, these
changes had almost no effect on enantioselectivity.26 Other
solvents were examined,27 but the mixed system of metha-
nol and ethyl acetate was superior from reaction rate,
solubility, impurity profile, and environmental consid-
erations.28 Reduction of des-keto 1 failed, which showed
the necessity of the directing group.29 Experiments with D2
showed equal 96%30 incorporation at both the R and β
positions of the enone, which is consistent with the pathway
in Scheme 2. This result eliminated the possibility of alkene
migration prior to reduction. When optimized conditions31
were applied to substrate 1(10 g scale), ketone 2was obtained
in 90% isolated yield (>99.8% ee32) after crystallization.33
A possible mechanistic rationale for this rate accelera-
tion is shown in Scheme 3. Mixed ketal (VII) could form in
Table 1. Study of Key Additivesa
entry
additive
equiv
conv (%)b
5 (%)b
ee (%)c
1
2
3
4
5
6
7
8
9
none
1
100
12
0
40
0
Sc(OTf)3
ZnCl2
0.20
0.15
0.25
0.30
0.05
1.0
95
94
Zn(OAc)2
Zn(BF4)2
Zn(OTf)2
Zn(OTf)2
TfOH
0
0
100
100
100
100
17
0
95
95
95
95
96
0
0
0.01
0.001
6
TfOH
0
a All reactions were conducted on 100 mg of 1 in 1.1 mL of 30%
MeOH/EtOAc. b Peak area % by reverse phase HPLC at 220 nm shown.
c Reported on the basis of normal phase HPLC analysis with a chiral
column (AD-H Chiral Pak) at 220 nm.
(23) This could include residual TfOH in Zn(OTf)2 or TfOH gener-
ated from reaction with residual water or methanol.
(24) A screen of temperature was conducted with ligand 5 (S/C 2000
and 4000) and 5 mol % Zn(OTf)2. Reaction profiles were clean at
temperatures up to 90 °C, but above 90 °C, the reaction profiles showed
increased impurities, notably epimerization in the R-position. A slight
erosion of ee was also observed (94 vs 95% ee).
(25) A significant pressure effect was noted, where reactions con-
ducted at S/C 4000 were easily driven to completion under 1000 psig of
hydrogen.
(26) The lack of pressure effect on ee may suggest that the solvate
dihydride is an important intermediate in the catalytic cycle.
(27) 1,2-DCE (with precatalyst formation in methanol) is an excellent
solvent system in terms of rate. However, we see increased epimerization
of product, and the use of this solvent is discouraged because of the
environmental impact and toxicological liabilities. Other mixed solvent
systems, MeOH/toluene and MeOH/THF, resulted in slower reaction
rates.
Figure 2. Overlay of a standard reaction with Zn(OTf)2 with a
reaction where Zn(OTf)2 was added after 16 h.
zinc acetate did not provide similar results. This prompted
us to determine if Brønsted acids were contributing to this
acceleration. As shown in Table 1, triflic acid20 is a potent
cocatalyst (<0.1 mol %);21 however, these reactions were
prone to epimerization of product to form trans-ketone 5.
(28) Alcohols are commonly used in asymmetric hydrogenations;
however, compounds 1 and 2 are poorly soluble in alcohols, which
prevents their use as the sole solvent. Ethyl acetate is added to maintain a
homogeneous solution, but EtOAc is not a viable solvent alone and
slows the reaction down at higher concentrations.
(29) The allylic alcohol derived from 1 was also made but was
unstable under reaction conditions.
(30) The remaining 4% was equal H incorporation in the same
positions. We believe this is due to exchange from the solvent catalyzed
by zinc triflate. Reactions at high catalyst loading without zinc triflate
showed 100% deuterium insertion.
(31) S/C 4000, 0.1 equiv of Zn(OTf)2, 1000 psig H2, 30% MeOH/
EtOAc, 70 °C, 18 h.
(32) Crude ee was 94ꢀ95%, and upgrade to >99.5% was obtained
by crystallization. Complete details are available in the experimental
section.
(33) See: May, S. A.; Johnson, M. D.; Braden, T. M.; Calvin, J. R.;
Haeberle, B. D.; Jines, A. R.; Miller, R. D.; Rener, G. A.; Richey, R. N.;
Schmid, C. R.; Vaid, R. K.; Yu. H. Org. Process Res. Dev. (submitted) for
a full account of the scale up to 140 kg (88% yield) using a continuous process
in a plug flow reactor.
(20) Tetrafluoroboric acid was also examined, and similar results
were observed. However, HBF4 led to slightly lower ee and has the
additional liability of generating HF. On the basis of this, we focused the
work on triflic acid and salts thereof.
(21) For an example of cooperative metalꢀBrønsted acid catalysis,
see: Li, C.; Villa-Marcos, B.; Xiao, J. J. Am. Chem. Soc. 2009, 131, 6967.
(22) If TfOH was generated via Zn(OTf)2 þ MeOH, increasing the
Zn(OTf)2 concentration by 20ꢁ should result in higher TfOH concen-
tration. This, in turn, should result in an increased amount of product
epimerization. We would also expect the hydrolysis rate of zinc triflate to
be negligible on the basis of literature precedent. See: Corey, E. J.;
Shimoji, K. Tetrahedron Lett. 1983, 24, 169.
1040
Org. Lett., Vol. 14, No. 4, 2012

the presence of either Brønsted or Lewis acids, resulting in a
more electron rich alkene as well as a better directing group
in the hydroxyl group of the mixed ketal. The result may be
an increase in the rate of substrate ligand complex forma-
tion to deliver III (Scheme 2). Additional work is underway
to further examine the mechanism of this reaction.
Figure 3. Influence of olefin geometry on stereoinduction.
Scheme 3. Possible Catalytic Function of Lewis or Bronsted
Acid
As the substitution about the ketone changes from phenyl
to alkyl (Table 2, entries 1ꢀ4), higher catalyst loading was
required, and there was slight erosion of ee as R2 decreased
in size. Varying substitution of the aryl ring in the β-
position (Table 2, entries 5ꢀ10) showed that steric bulk
does influence enantioselectivity.
Interestingly, in the case of acylic substrates 8 and 10, the
alkene geometry influenced the stereochemical course of
the reaction (Figure 3). Z-alkene 8 proceeded to the
expected (2R,3S) product 9, while the E-alkene 10 deliv-
ered opposite facial selectivity, the (2R,3R) product 11
under identical conditions.
Table 2. Expanded Scope of Rh-Catalyzed Asymmetric Re-
duction
In conclusion, a chemo- and enantioselective Rh-
catalyzed reduction of tetrasubstituted R,β-unsaturated
ketones has been developed. Catalyst turnover frequency
is enhanced by the addition of catalytic zinc triflate, which
allows for high S/C while suppressing epimerization in the
R-positionof the products. Furtherefforts are underway to
expand the scope34,35 and understand the mechanism of
this transformation.
entrya
R1
R2
Ph
ee (%)b
yield (%)c,d
1 (6a)
2 (6b)
3 (6c)
4 (6d)
5 (6e)
6 (6f)
7 (6g)
8 (6h)
9 (6i)
10 (6j)
2,6-(BnO)2C6H3
2,6-(BnO)2C6H3
2,6-(BnO)2C6H3
2,6-(BnO)2C6H3
C6H5
95
90
92
96
80
77
82
95
95
94
90
45e
91e
95e
97
94
94
98
95
86
Me
Et
Acknowledgment. We wish to thank Professors Clark
Landis, William Roush, Peter Wipf, Eric Carreira, and Dr.
Joseph Martinelli for helpful discussions regarding this
work. We also thank Jonas Y. Buser and David Foley for
NMR work relating to the D2 reactions. We also wish to
acknowledge ChaoYe, You Chen, Ling Xia Jiang, Hui Liu,
Anping Ren, Xiaoyang Wang, Weifeng Xi, Baitao Gao,
Hui Lin, and Zhihua Liu for preparation and characteriza-
tion of acrylic acids used in the preparation of substrates 6
in Table 2.36
iPr
H
4-CH3OC6H4
3-CH3OC6H4
2-CH3OC6H4
2-EtOC6H4
4-OMe
3-OMe
2-OMe
2-OEt
2-OiPr
2-i-PrOC6H4
a All reactions were conducted on 200 mg of substrate. b ee is reported
on crude reaction products. c Yields represented are isolated after
purification. d Poor conversion was observed without Zn(OTf)2 in all
cases under otherwise identical conditions. e S/C 100 used in these cases
to ensure reaction completion in 18 h.
Supporting Information Available. Detailed experimen-
tal procedures and spectral data for all compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
Examination of alternative substrates established that
the methodology has reasonably broad scope (Table 2).
(35) Efforts on other tetrasubstituted substrates (tetrasubstituted N-
acyl derivatives, for example) will be reported in due course.
(36) Address: Shanghai PharmExplorer Co., Ltd, No. 998 Halei Road,
Zhangjiang Hi-Tech Park, Pudong New Area, Shanghai, 201203, China.
(34) The addition of zinc triflate to ruthenium-based catalysts does
improve turnover as well, but the effect is not as dramatic as in the
rhodium case. For example, the reduction of 1 with a Ru-(4) complex
(S/C 50) fails to give any conversion without zinc triflate, while the
reaction with 0.2 equiv proceeds to 19% conversion with 60% ee.
Results on related systems will be reported in due course.
The authors declare no competing financial interest.
Org. Lett., Vol. 14, No. 4, 2012
1041