Chemo- and enantioselective routes to chiral fluorinated hydroxyketones using ketoreductases

Article abstract of DOI:10.1021/ol701810v

Chiral fluorinated hydroxyketones were synthesized with excellent ee (>98%) and yield by a chemo- and stereoselective reduction of prochiral methyl/trifluoromethyl diketones using commercially available ketoreductase enzymes. By using p- and m-trifluoroac

Full text of DOI:10.1021/ol701810v

ORGANIC
LETTERS
2007
Vol. 9, No. 24
4951-4954
Chemo- and Enantioselective Routes to
Chiral Fluorinated Hydroxyketones
Using Ketoreductases
Brendan T. Grau,* Paul N. Devine, Lisa N. DiMichele, and Birgit Kosjek
Department of Process Research, Merck Research Laboratories, Merck and Co., Inc.,
PO Box 2000, Rahway, New Jersey, 07065
brendan_grau@merck.com
Received July 27, 2007
ABSTRACT
Chiral fluorinated hydroxyketones were synthesized with excellent ee (>98%) and yield by a chemo- and stereoselective reduction of prochiral
methyl/trifluoromethyl diketones using commercially available ketoreductase enzymes. By using p- and m-trifluoroacetyl substituted
acetophenones, we demonstrate that ketoreductases can selectively differentiate between methyl and trifluoromethyl ketones within the same
molecule. As a result, useful catalysts were identified that eliminated the need for costly and time-consuming protection/deprotection of the
ketone moiety, enabling a more convergent synthesis of hydroxyketones. Further, a route to chiral methyl hydroxyketones is provided where
an enzyme selectively reduces the unactivated ketone.
Over the past decade, there has been a dramatic increase in
the demand for fluorinated intermediates1 as chemists rou-
tinely incorporate fluorine into molecules in order to modify
the electronic properties, binding affinities, and bioavail-
abilities of compounds.2 As Swinson indicates, fluorines are
prevalent in antidepressant, antibacterial, antiinflammatory,
and cholesterol-lowering drugs. Therefore, the ability to
produce optically pure halogenated intermediates is of great
value to the pharmaceutical industry in order to develop
robust, cost-efficient syntheses of complex molecules. One
potentially useful intermediate for synthesis that is not readily
available is a fluorinated chiral hydroxyketone. An elegant
and efficient method to access these compounds is utilizing
a chemo- and stereoselective reduction of a prochiral dike-
tone. This route will allow access to a chiral hydroxyketone
without the need for protecting groups and additional steps
simply by using a commercially available catalyst.
In this paper, we describe the highly chemo- and enantio-
selective preparation of chiral fluorinated hydroxyketones
employing reduction of a p- and m-diketone, 1 and 2
respectively, catalyzed by commercially available ketore-
(1) Thayer, A. M. Chem. Eng. News 2006, 84, 15-24.
(2) Swinson, J. PharmaChem 2007, 6, 4, 38-41; Swinson, J. Pharma-
Chem 2005, 4, 6, 26-30.
10.1021/ol701810v CCC: $37.00
© 2007 American Chemical Society
Published on Web 10/31/2007

a broad substrate range including substituted acetophe-
nones,10-13 as well as 2,2,2-trifluoroacetophenone.14-16 Fin-
ally, the catalysts that we screened are synthetically useful
as they are commercially available17 as isolated enzymes and
have been used to deliver multikilograms of optically pure
chiral alcohol.13
Scheme 1. Chemoselective Reductions of Methyl/
Trifluoromethyl Diketones To Produce Hydroxyketonesa
The methyl/trifluoromethyl diketones 1, 2, and 3 were
synthesized by modifying the literature procedure.3 Ketals
were formed from the corresponding bromo-acetophenones
using ethylene glycol and were catalyzed by p-toluene
sulfonic acid. This step was followed by a metal/halogen
exchange using n-hexyl lithium and quench with methyl
trifluoroacetate. Deprotection with 1 N HCl and crystalliza-
tion afforded the trifluoroacetyl-substituted acetophenones
1, 2, and 3 as solids with 60-80% isolated yield. The three
diketones 1, 2, and 3 were screened against a library
consisting of 72 commercially available enzymes in a 96-
well format.17 The o-diketone was isolated as a cyclized
dihydrate 3a (Scheme 1) and found to be inert in all chemical
and enzymatic reactions. The reactions were monitored by
an achiral GC assay to determine the product ratios and
conversions. The ee value of hydroxyketones was determined
by chiral SFC for 4 and 5 and by chiral GC for 7 and 8.
The majority of the ketoreductase library was active toward
the p-diketone 1 and produced the trifluoromethyl hydroxy-
ketone 4 in excess. The enzymes listed in Table 1 showed
the highest chemoselectivity toward the p-diketone. In fact,
49 of the enzymes had a mixture of trifluoromethyl hydroxy-
ketone 4 and bis-alcohol 10 when sampled.
a 3a: The o-diketone was isolated as a cyclized dihydrate and
was inert in all reactions.
ductase enzymes (Scheme 1). This new method of producing
optically pure hydroxyketones eliminates the need to protect
prochiral diketones and allows a more efficient and conver-
gent synthesis. Importantly, we demonstrate a route to (S)-
methyl hydroxyketones (7, 8) which are difficult compounds
to synthesize due to the relative ease of a chemical reduction
of a trifluoromethyl ketone located within five bonds of the
methyl carbonyl.
The chemoselective reduction (Scheme 1) of these com-
pounds is reported in the literature,3 but this reaction has
limited application in the synthesis of pharmaceutical inter-
mediates as it produces racemic hydroxyketones. Synthesiz-
ing chiral trifluoromethyl hydroxyketones by a chemoselec-
tive asymmetric hydrogenation of fluorinated â-diketones
was demonstrated using Pt/Al2O3 and chiral modifiers.4
However, this metal-catalyzed transformation is limited to
reduction of the activated ketone in the â-diketones exam-
ined.
The aim of our research was to demonstrate that keto
reductases can chemo- and stereoselectively differentiate
between methyl and trifluoromethyl ketones within the
same molecule by using p- and m-trifluoroacetyl-substituted
acetophenones. Enzymes are an excellent choice as catalysts
to mediate this chemoselective diketone reduction as the
elaborate chiral environment should be able to differentiate
between the ketones due to the steric and electronic differ-
ences of the carbonyls. However, prior to the results
presented here, there are no examples in the literature of an
enzymatic chemoselective reduction of methyl/trifluoro-
methyl diketones, although there are a number of examples
of enzymes catalyzing the regio- and enantioselective reduc-
tions of substrates such as R-diketones5-7 and â-diketones.7-9
Additionally, these biocatalysts have been shown to have
The results of screening the p-diketone 1 toward the library
demonstrates that by using isolated enzymes one can chemo-
selectively access both enantiomers of the trifluoromethyl
hydroxyketone 4 in high enantiomeric excess. In fact, there
(6) Mahmoodi, N. O.; Mohammadi, H. G. Monatsh. Chem. 2003, 134,
1283-1288.
(7) Edegger, K.; Stampfer, W.; Seisser, B; Faber, K; Mayer, S. F.;
Oehrlein, R.; Hafner, A.; Kroutil, W. Eur. J. Org. Chem. 2006, 1904-
1909.
(8) Muller, M.; Wolberg, M.; Shubert, T.; Hummel, W. AdV. Biochem.
Eng. Biotechnol. 2005, 92, 261-287.
(9) Wolberg, M.; Hummel, W.; Wandrey, C.; Muller, M. Angew. Chem.,
Int. Ed. 2000, 39, 23, 4306-4308.
(10) Zhu, D.; Rios, B. D.; Rozzell, J. D.; Hua, L. Tetrahedron:
Asymmetry 2005, 16, 1541-1546.
(11) Groger, H.; Hummel, W.; Rollman, C.; Chamouleau, F.; Husken,
H.; Werner, H.; Wunderlich, C.; Abokitse, K.; Drauz, K.; Buchholz, S.
Tetrahedron 2004, 60, 633-640.
(12) Rosen, T. C.; Feldmann, R.; Dunkelmann, P.; Daubmann, T.
Tetrahedron Lett. 2006, 47, 4803-4806.
(13) Pollard, D.; Truppo, M.; Pollard, J.; Chen, C.; Moore, J. Tetrahe-
dron: Asymmetry 2006, 17, 554-559.
(14) Yamazaki, Y.; Kobayashi, H. Tetrahedron: Asymmetry 1993, 6,
1287-1294.
(15) Bradshaw, C. W.; Hummel, W.; Wong, C. H. J. Org. Chem. 1992,
57, 1532-1536.
(16) Caron, D.; Coughlan, A. P.; Simard, M.; Bernier, J.; Piche, Y.;
Chenevert, R. Biotechnol. Lett. 2005, 27, 713-716.
(17) Isolated enzymes were used as lyophilized solids or liquid prepara-
tions. KRED-XXX enzymes, alcohol dehydrogenase from Rhodococcus
erythropolis (ADH-RE), alcohol dehydrogenase from Candida parapsilosis
(ADH-CP), glucose dehydrogenase-103 (GDH-103), NAD+, NADH, and
NADPH were purchased from Biocatalytics, Inc. (Pasedena, Ca, U.S.A.).
ADH CDX010, ADH CDX013, alcohol dehydrogenase from Lactobacillus
breVis (ADH-LB), and alcohol dehydrogenase from Thermoanaerobacter
sp. (ADH T) were purchased from Julich Chiral Solutions, Inc (Julich,
Germany). All chemicals used were certified as reagent grade and purchased
from Sigma-Aldrich and Fisher Scientific. Screening conditions are listed
in Tables 1 and 2 under footnote “a”.
(3) Sasaki, S.; Yamauchi, T.; Kubo, H.; Kanai, M.; Ishii, A.; Higash-
iyama, K. Tetrahedron Lett. 2005, 46, 1497-1500.
(4) Hess, R.; Diezi, S.; Mallat, T.; Baiker. A. Tetrahedron: Asymmetry
2004, 15, 251-257.
(5) Boutante, P.; Mousset, G.; Veschambre, H. New J. Chem. 1998, 247-
251.
4952
Org. Lett., Vol. 9, No. 24, 2007

Table 1. Enzyme-Catalyzed Reduction of p-Diketone 1
4 yield
(%)b
7 yield
(%)b
10 yield
(%)b
ee
(%)c,d
entry
enzyme
1
2
3
4
5
6
7
8
KRED-112
KRED-129
KRED-131
KRED-A1i
KRED-A1n
KRED-A1x
CDX-013
KRED-111
KRED-B1d
KRED-B1e
KRED-113
CDX-010
GDH-103
ADH-RE
ADH-T
ADH-CP
KRED-A1p
100
100
100
100
100
100
92
92
90
88
86
71
72
47
27
0
0
0
0
0
0
0
0
0
0
0
0
0
9
0
10
0
100
87
0
0
0
0
0
0
8
8
10
12
14
20
28
43
73
0
>99 (S)
>99 (R)
>99 (R)
>99 (S)
>99 (R)
>99 (R)
>99 (S)
97 (S)
>99 (S)
>99 (S)
>99 (S)
>99 (R)
13 (R)e
>99 (R)f
>99 (R)
98 (S)
9
10
11
12
13
14
15
16
17
Figure 1. ADH-RE-catalyzed chemoselective oxidation of p-bis
alcohol 10. At 16 h timepoint, 1 equiv of NADH was added to
drive the reaction to completion due to instability of the cofactor
under the reaction conditions.18
conditions18 and the initial reduction rates of the carbonyls,
our hypothesis is that the oxidized cofactor (NAD+) accum-
ulates within the system and the enzyme chemoselectively
oxidizes the methyl alcohol of 10 to yield highly enantiopure
4. This oxidation was observed as a result of our screening
conditions employing excess cofactor. In contrast, addition
of an enzyme that recycles cofactor to the reaction to
regenerate NADH yields bis-alcohol 10 as the sole product.
A similar regioselective oxidation of diols was presented
by Edegger, et al.7 as a route to provide enantiopure
hydroxyketones. In order to prove that we were observing a
chemoselective oxidation, we ran reactions using ADH-RE
and charged racemic bis-alcohol (10) and excess NAD+,
which yielded 21% of 4 and no methyl hydroxyketone 7.
Therefore, the chemoselective oxidation of similar bis-
alcohols using enzymes such as ADH-RE may allow a route
to produce chiral hydroxyketones when combined with a
system to recycle the oxidized cofactor.19
Identifying catalysts that chemoselectively reduce the
methyl ketone proved to be more difficult. Yet as shown in
Table 1 (entries 16 and 17), the (S)-enantiomer of the methyl
hydroxyketone 7 can be synthesized. Two enzymes, the
alcohol dehydrogenase from Candida parapsilosis (ADH-
CP) and KRED-A1p demonstrated chemoselectivity toward
the reduction of the methyl carbonyl producing the methyl
hydroxyketone 7 with low levels of bis-alcohol 10 and none
of the other regioisomer 4. This result was confirmed at larger
scale with a 94% yield, 98% ee, and less than 1% of 10.
The ability of the enzyme to successfully mediate this
chemoselective reduction may be due to a combination of
factors including a pseudoprotection of the trifluoromethyl
carbonyl as a stable hydrate under aqueous conditions
(observed by 1H NMR), an electronic/hydrophobic effect of
the fluorines within the active site, and a steric difference in
13
98 (S)
a Reaction conditions: 0.4 g/L enzyme, 2.5 g/L diketone, 3 equiv of
NAD(P)H, 5% v/v DMF, in 100 mM potassium phosphate buffer (pH 7) at
30 °C. b Analyzed using achiral GC assay. c Analyzed using chiral SFC
for 4 and chiral GC assays for 7. d Absolute configuration for 4 and 7 was
determined by using the modified Mosher method21 and chiral hydroxyke-
tones produced from the catalysts listed in entries 1 and 16. e Entry 13 is
listed although it shows moderate chemoselectivity and low enantioselec-
tivity due to relevance of GDH-103 as a cofactor recycling enzyme and its
activity toward 1. f ee and absolute configuration is listed for trifluoromethyl
hydroxyketone 4.
are a number of enzymes whereby one can choose to produce
either of the enantiomers with high yield. With perfect
chemoselectivity, KRED-112 or KRED-A1i (Table 1, entries
1 and 4) can be used to synthesize the (S)-trifluoromethyl
hydroxyketone 4, while KRED-129, KRED-131, KRED-
A1n, and KRED-A1x (Table 1, entries 2, 3, 5, and 6) produce
the (R)-enantiomer 4.
To confirm the results of the screen and isolate the chiral
hydroxyketones, reactions were run at the 100-mg scale.
These experiments indicate that 4 is synthesized by two
routes. One route, demonstrated by KRED-112 (Table 1,
entry 1), shows a perfectly chemoselective reduction of the
trifluoromethyl carbonyl, producing the (S)-enantiomer of 4
cleanly with less than 1% bis-alcohol 10 and no production
of the other regioisomer 7. The second route to 4 is through
a mechanism of chemoselective oxidation with the bis-
alcohol 10 acting as an internal cofactor recycling reagent.
This route (Figure 1) is demonstrated using alcohol dehy-
drogenase from Rhodococcus erythropolis (ADH-RE, Table
1, entry 14) which rapidly reduces methyl carbonyl of 1 to
produce 7. Gradually, compounds 4 and 10 begin to
accumulate in the system. At 16 h, additional NADH was
charged to drive the reaction forward and results in an
increase of 7 and 10 with a decrease of 4. At the end of the
reaction, the sole products are 4 and 10 in a 1:1 mixture.
Considering the instability of NADH under the reaction
(18) Wu, J. T.; Wu, L. H.; Knight, J. A. Clin. Chem. 1986, 3, 314-319.
(19) Hummel, W.; Riebel, B. Biotechnol. Lett. 2003, 25, 51-54.
Org. Lett., Vol. 9, No. 24, 2007
4953

the relative size of a trifluoromethyl group as compared to
the methyl analogue.20
ketone 5 can be synthesized in high ee (>99%) by a number
of catalysts, such as KRED-A1i and KRED-112 (Table 2,
entries 1 and 6) while the (R)-enantiomer was achieved using
ADH-RE (Table 2, entry 14). A time course analysis of the
reactions using 100 mg of 2 and the catalysts, KRED-112
(Table 2, entry 6) and KRED-A1y (Table 2, entry 2) revealed
excellent chemoselectivity to 5 with <1% bis-alcohol 11
when harvested at complete consumption of the diketone 2.
ADH-CP (Table 2, entry 17) produces the (S)-enantiomer
of the m-methyl hydroxyketone 8 in moderate ee and
chemoselectivity. A switch from the p-diketone to the
m-diketone showed reduction of stereoselectivity and chemose-
lectivity for this catalyst.
The m-diketone 2 results indicate that both enantiomers
of the m-trifluoromethyl hydroxyketone 5 can be synthesized
with high ee and good chemoselectivity. The enzymes listed
in Table 2 show the highest chemoselectivity toward the
Table 2. Enzyme-Catalyzed Reduction of m-Diketone 2
In fact, there were a number of enzymes that showed
differences in chemo- and stereoselectivity against the p- and
m-diketones 1 and 2 under the same conditions. The
substrate-dependent variations in enzyme selectivity are likely
due to the role that steric and electronic effects play within
the active site of these catalysts. Therefore, there is value in
screening a moderate sampling of distinct enzymes22 to
ensure that one identifies the best catalyst for a particular
substrate.
The use of readily available commercial ketoreductases
has been demonstrated to be very effective for reducing the
p- and m-diketones 1 and 2 with high ee and minimal
formation of side products. We have successfully demon-
strated the direct access to both enantiomers of the trifluo-
romethyl hydroxyketones 4 and 5. Additionally, we have
provided a route to the (S)-enantiomers of methyl hydroxy-
ketones 7 and 8.
5 yield
8 yield
11 yield
ee
entry
enzyme
(%)b
(%)b
(%)b
(%)c,d
1
2
3
4
5
6
7
8
KRED-A1i
KRED-A1y
KRED-B1e
KRED-124
KRED-B1d
KRED-112
KRED-113
KRED-115
KRED-129
GDH-103
KRED-A1q
KRED-126
ADH-LB
89
88
88
87
87
86
86
86
86
85
84
83
83
82
78
76
4
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
76
11
12
12
13
13
14
14
14
14
15
16
17
17
18
22
24
20
>99 (S)
76 (R)
>99 (S)
>99 (S)
>99 (S)
>99 (S)
>99 (S)
96 (S)
9
63 (R)
10
11
12
13
14
15
16
17
15 (S)e
>99 (S)
98 (S)
>99 (S)
>99 (R)
>99 (S)
42 (R)
ADH-RE
CDX-013
KRED-130
ADH-CP
90 (S)f
Note Added after ASAP Publication. Compound 1 was
incorrect in Table 1 in the version published ASAP October
31, 2007; the corrected version was published ASAP
November 2, 2007.
a Reaction conditions: 0.4 g/L enzyme, 2.5 g/L diketone, 3 equiv of
NAD(P)H, 5% v/v DMF, in 100 mM potassium phosphate buffer (pH 7) at
30 °C. b Analyzed using achiral GC assay. c Analyzed using chiral SFC
for 5 and chiral GC assays for 8. d Absolute configuration for 5 and 8 was
determined by using the modified Mosher method21 and chiral hydroxyke-
tones produced from the catalysts listed in entries 6 and 17. e Entry 13 is
listed although it shows low enantioselectivity due to relevance of GDH-
103 as a cofactor recycling enzyme and its activity toward 2. f Absolute
configuration is listed for methyl hydroxyketone 8.
Supporting Information Available: Experimental pro-
cedures for synthesizing the substrates and products as well
as analytical methods and spectroscopic data for all com-
pounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
m-diketone under the screening conditions. Sixty-eight of
the catalysts (94%) had a mixture of m-trifluoromethyl
hydroxyketone 5 and bis-alcohol 11. Overall, the chemose-
lectivity of the catalysts under the conditions employed was
lower for the m-diketone 2 as compared to the p-diketone 1.
Again, the (S)-enantiomer of the m-trifluoromethyl hydroxy-
OL701810V
(20) Timperley, C. M.; White, W. E. J. Fluorine Chem. 2003, 123, 65-
70.
(21) Xiao, L.; Yamazaki, T.; Kitazume, T.; Yonezawa, T.; Sakamoto,
Y.; Nogawa, K. J. Fluorine Chem. 1997, 84, 19-23.
(22) Homann, M. J.; Vail, R. B.; Previte, E.; Tamarez, M.; Morgan, B.;
Dodds, D. R.; Zaks, A. Tetrahedron. 2004, 60, 789-797.
4954
Org. Lett., Vol. 9, No. 24, 2007