Highly enantioselective synthesis of fluorinated β-amino ketones via asymmetric organocatalytic Mannich reactions: A case study of unusual reversal of regioselectivity

Article abstract of DOI:10.1055/s-0030-1259513

The highly enantioselective direct Mannich protocol employing fluoroacetone, p-anisidine, and aldehydes catalyzed by 4-siloxyproline was developed, allowing efficient access for pharmaceutically important fluorinated β-amino ketones. On the basis of DFT c

Full text of DOI:10.1055/s-0030-1259513

CLUSTER
477
Highly Enantioselective Synthesis of Fluorinated b-Amino Ketones via
Asymmetric Organocatalytic Mannich Reactions: A Case Study of Unusual
Reversal of Regioselectivity
E
nantioselective Synth
i
e
sis of
nFluorinated
b
-Ami
L
no
K
etones u, Yunpeng Lu, Puay Kah Amy Tan, Qiu Ying Lau, Guofu Zhong*
Division of Chemistry and Biological Chemistry, School of Physical and Mathematical Sciences, Nanyang Technological University,
21 Nanyang Link, Singapore 637371, Singapore
Fax +6567911961; E-mail: guofu@ntu.edu.sg
Received 1 December 2010
ful application of a wide array of ketone nucleophiles in
the organocatalytic direct asymmetric Mannich reac-
tions,8 however, the employment of fluoroacetone re-
mains elusive.9 Herein, we show that fluoroactone is a
Abstract: The highly enantioselective direct Mannich protocol em-
ploying fluoroacetone, p-anisidine, and aldehydes catalyzed by 4-
siloxyproline was developed, allowing efficient access for pharma-
ceutically important fluorinated b-amino ketones. On the basis of
DFT calculations and mechanistic probes, the origin of an unusual powerful donor in the direct Mannich reaction with unac-
tivated aldimines catalyzed by 4-siloxyproline,10 giving
reversal of regioselectivity was unraveled.
fluorinated b-amino ketones with excellent enantiosele-
tivities (Scheme 1).
Key words: organocatalysis, Mannich reaction, DFT calculation,
fluoroacetone, b-amino ketones
From the outset, we realized that due to the similar size of
fluorine atom compared to hydrogen, the overall regiose-
lectivity may well exhibit a complex dependence on sev-
eral competing factors, such as initial preferences, relative
rates, and reversibilities of individual steps, which all
have the potential to exert an influence (vide infra). On the
basis of our initial DFT calculations (Scheme 2),11 it was
found that in sharp contrast to its isoelectronic counter-
part, namely hydroxyacetone, fluoroacetone-derived
enamine, as well as the transition state (TS) have little
preference in the gas phase (which can be generalized to
nonpolar solvents), enamine I (which leads to linear prod-
uct) is only slightly more favored by 0.23 kcal/mol, while
its corresponding TS I exhibits a higher energy barrier of
0.54 kcal/mol, thus rendering the Mannich reaction with
almost no regioselectivity. However, with DMSO as sol-
vent, the energy of enamine I is 0.98 kcal/mol lower than
that of enamine II, and TS I is more favored by 0.80 kcal/
mol, with net effect of increased selectivity towards linear
product. This unusual reversal in regioselectivity was in
Since the year of 2000, elegant studies on the proline-cat-
alyzed direct asymmetric Mannich reactions by List1 and
Barbas III et al.2 have unveiled a new field of opportuni-
ties in the synthesis of b-amino carbonyl compounds.3
These Mannich products are of great value due to the fact
that the b-amino carbonyl compounds are ubiquitous drug
intermediates and building blocks of other biologically
active compounds.3 Meanwhile, organofluorine com-
pounds have gained an increased attention from synthetic
community as well as pharmaceutical industry.4 Due to
fluorine’s strong electron-withdrawing nature, the re-
placement of the hydroxy or hydrogen atoms in drug mol-
ecules with fluorine atoms can bring about drastic changes
to both the physical and chemical properties of the mole-
cules, in particular, their pharmacological activity.5
We reported the utility of commercially available fluoro-
acetone in the organocatalytic asymmetric direct aldol re-
actions,6 which paves the way to subsequent extensive
investigation and mechanistic study.7 Despite the success-
O
O
O
NHPMP
R
O
NHPMP
R
NH2
4 (20 mol%)
+
+
+
solvent
r.t.
H
R
MeO
F
F
F
1
2
3
5
6
TBSO
5/6 = 6:1 to >10:1
53 to 81% yield
94 to >99% ee
CO2H
N
CO2H
N
H
H
4a
4b
Scheme 1 Direct organocatalytic asymmetric Mannich reactions of fluoroacetone, p-anisidine, and aldehydes
SYNLETT 2011, No. 4, pp 0477–0480
x
x.
x
x.
2
0
1
1
Advanced online publication: 27.01.2011
DOI: 10.1055/s-0030-1259513; Art ID: Y02210ST
© Georg Thieme Verlag Stuttgart · New York

478
M. Lu et al.
CLUSTER
enamine I
enamine II
O
O
N
H
N
H
branched
product
linear
product
O
O
PMP
F
PMP
N
N
F
Ph
Ph
TS I
TS II
Geometries
Energy + ZPE correction Energy +ZPE correction
in gas phase (a.u.)
in DMSO solvent (a.u.)
enamine I
enamine II
TS I
–616.708625
–616.721175
–616.708246
–616.719612
–1287.404806
–1287.405674
–1287.436601
–1287.435327
TS II
Scheme 2 DFT calculations of possible enamine derived from fluoroacetone and L-proline (enamine I and enamine II) and lowest energy
transition state (TS I and TS II) (1 a.u. = 627.5095 kcal/mol)
Table 1 Optimization of Reaction Conditionsa
stark contrast to most of previous reports on other ketone
donors.12
Entry
Time
(h)
Solvent
l/bb
Yield of 5a ee of 5a
(%)c
(%)d
Encouraged by these results, we set out to test our pro-
posed organocatalytic three-component Mannich reaction
between fluoroacetone (1), p-anisidine (2), and butanal
(3a), using L-proline (4a) as catalyst (Table 1). Indeed, the
reaction proceeds well at room temperature in DMSO,
yielding fluorinated b-amino ketone as 4:1 regioisomers
in favor of linear product (entry 1). Although other polar
aprotic solvents display similar regioselectivities (entries
2, 3), switching the solvent to less polar MeCN, THF, or
CH2Cl2, the linear-to-branch ratio (l/b) drops significantly
to almost 1:1 (entries 4–6). Similar observations were ob-
tained when the reaction was conducted in water or under
solvent-free conditions (entries 7, 8).
1
2
5
7
DMSO
DMF
NMP
MeCN
THF
4:1
55
80
4.5:1
4:1
53
83
3
12
24
24
24
24
24
18
3
48
85
4
1.2:1
1.1:1
1:1
<10
<10
<10
<10
<10
50
n.d.
n.d.
n.d.
n.d.
n.d.
77
5
6
CH2Cl2
H2O
7
1:1
8
neat
1:1
During the optimization study, we found the initial regio-
selectivity was extremely high (l/b > 20:1, by crude 19F
NMR analysis), while prolonged reaction time or higher
reaction temperature resulted in lower l/b ratio (entry 9,
10), possibly due to instability of linear product (vide in-
fra).13 Given the inferior performance of 4a at lower tem-
perature (entry 11), we screened more catalysts and
finally arrived at the optimized conditions (entry 12).
9
DMSO
DMSO
DMF
DMF
2.5:1
2:1
10e
11f
12f,g
38
75
48
60
5:1
31
90
10:1
70
94
a General reaction conditions: fluoroacetone (1, 0.60 mmol), p-anisi-
dine (2, 0.55 mmol), butanal (3a, 0.50 mmol), L-proline (4a, 20
mol%), solvent (2.0 mL), r.t.
With the optimal conditions in hand, we examined the
substrate scope, and the results are displayed in Table 2.
The protocol tolerates a wide spectrum of R groups, in-
cluding linear alkyls (entries 1, 2), branched hydrocarbons
(entry 3), as well as aromatics with various electronic and
steric properties (entries 4–10). It was found that the more
sterically demanding R groups gave better l/b ratios and
ee, at the expense of reactivity. Remarkably, in marked
contrast to previous proline-catalyzed direct Mannich re-
actions, electron-donating and heteroaromatic R groups
all provide desired fluorinated b-amino ketones with ex-
cellent stereocontrol (entries 9, 10).
b Determined by 1H NMR of crude reaction mixture.
c Yields of 5 after flash column chromatography on silica gel.
d Determined by HPLC on chiral stationary phase (Daicel chiralpak
AS-H).
e Reaction was carried out at 35 °C.
f Reaction was carried out at 4 °C.
g Reaction was carried out with 3 equiv of 1, 3 equiv of H2O, 20 mol%
of 4b.
retro-Michael reaction of initially formed linear Mannich
product 5f (pathway B), while the branched counterpart 6f
remained intact. This readily explains the erosion of l/b ra-
tios observed in the cases with prolonged reaction time
and higher temperature (vide supra). We reasoned that the
addition of water may facilitate retro-aldol condensation
process of the enone, which regenerate the substrates 1
and 3 (pathway C).
Although water was not essential in the case of aliphatic
aldehydes,14 we isolated enone 7f as a major byproduct
during investigation on aromatic aldehydes without the
addition of water (Scheme 3). Control experiments re-
vealed that secondary and tertiary amine may catalyze the
Synlett 2011, No. 4, 477–480 © Thieme Stuttgart · New York

CLUSTER
Enantioselective Synthesis of Fluorinated b-Amino Ketones
479
R2
R3
N
H
B: retro-Michael
O
NHPMP
R1
O
NHPMP
R1
O
NH2
R1
F
F
F
6f (R1 = 4-O2NC6H4)
MeO
5b (R1 = n-Bu)
7f
2
R1 = 4-O2NC6H4
isolated
5f (R1 = 4-O2NC6H4)
H2O
O
O
H
R1
R1
3
1
A: Mannich
C: retro-aldol
F
8f
O
F
not observed
Scheme 3 Plausible decomposition pathway
Table 2 Substrate Scopesa
Representative Procedure
To a solution of catalyst 4b (24.5 mg, 0.10 mmol, 20 mol%) and p-
anisidine (2, 67.7 mg, 0.55 mmol) in anhyd DMF (2 mL, 0.25 M)
with H2O (25 mL, 1.4 mmol) at 4 °C, fluoroacetone (1, 100 mL, 1.5
mmol) was added in one portion, followed by butanal (3a, 45 mL,
0.5 mmol), the stirring was continued at 4 °C until full consumption
of the imine as indicated by crude 1H NMR (ca. 24 h). The reaction
was quenched with half-sat. NH4Cl solution, and the organic mate-
rials were extracted with EtOAc three times. The combined organic
layers were washed with brine, dried on MgSO4, and concentrated
in vacuo. A small portion of the crude reaction mixture was taken to
determine the l/b ratio by 1H NMR, the rest were purified by flash
column chromatography on silica gel (EtOAc–hexane = 10:90 to
15:85) to afford (R)-1-fluoro-4-[(4-methoxyphenyl)amino]heptan-
2-one (5a) as a colorless oil (88 mg, 70% yield) in 94% ee.
Entry
3 R
l/bb
10:1
Yield (%)c ee (%)d
1
2
3b n-Bu
73
71
68
65
70
72
81
68
53
65
94
3c Ph(CH2)2
3d i-Bu
>10:1
>10:1
7:1
95
3
98.5
97
4
3e Ph
5
3f 4-O2NC6H4
3g 4-ClC6H4
3h 1-Naphth
3i 2-FC6H4
3j 4-MeOC6H4
3k 2-thienyl
8:1
99
6
8:1
98.5
7
>10:1
9:1
>99
1H NMR (300 MHz, CDCl3): d = 6.77 (m, 2 H), 6.58 (m, 2 H), 4.74
(d, 2 H, J = 47.7 Hz), 3.83 (m, 1 H), 3.70 (s, 3 H), 2.80 (br, 1 H),
2.74 (dd, 1 H, J = 6.6, 16.7 Hz), 2.64 (dd, 1 H, J = 6.6, 16.7 Hz),
1.59–1.26 (m, 4 H), 0.89 (t, 3 H, J = 7.1 Hz). 13C NMR (75 MHz,
CDCl3): d = 206.5 (d, J = 19.4 Hz), 152.5, 141.1, 115.2, 115.0, 85.2
(d, J = 184.1 Hz), 55.8, 50.7, 43.0, 37.7, 19.4, 13.9. 19F NMR (376
MHz, CDCl3): d = –226.1 (t, J = 47 Hz). ESI-HRMS: m/z calcd for
8
98
94
97
9
6:1
10
7:1
a General reaction conditions: fluoroacetone (1, 1.50 mmol), p-anisi-
dine (2, 0.55 mmol), aldehyde 3 (0.50 mmol), 4-siloxyproline (4b, 20
mol%), H2O (25 mL), DMF (2.0 mL), 4 °C, 36–96 h.
b Determined by 1H NMR of crude reaction mixture.
c Yields of 5 after flash column chromatography on silica gel.
d Determined by HPLC on chiral stationary phase (Daicel chiralpak
AS-H, AD-H).
C14H21FNO2 [M + H+]: 254.1556; found: 254.1552. [a]D –6.9 (c
22
0.75, CH2Cl2, 94% ee). HPLC (Daicel chiralpak AD-H, 2-PrOH–
hexane = 15:85, flow rate 1.0 mL/min, detection UV at 254 nm):
tR = 8.9 min (major), 11.2 min (minor).
Acknowledgment
In conclusion, we have demonstrated the direct asymmet-
ric Mannich reactions employing fluoroacetone as donor
Research support from the Ministry of Education in Singapore
(ARC12/07, No. T206B3225) and Nanyang Technological Univer-
with unactivated aldimines as acceptors, allowing highly sity (URC, RG53/07) are gratefully acknowledged. Y.L. thanks the
Division of Chemistry and Biological Chemistry at Nanyang Tech-
nological University for providing the computational resources.
enantioselective synthesis of fluorinated b-amino carbon-
yl compounds. The unusual reversal of regioselectivity
was rationalized on the basis of the DFT calculations and
control experiments. We hope our mechanistic insight
into the fundamental processes and interactions will lead
to the development of more active and selective catalytic
systems, by showing several interesting features that point
the way towards approaches that may well prove success-
ful.
References and Notes
(1) (a) List, B. J. Am. Chem. Soc. 2000, 122, 9336. (b) List, B.;
Pojarliev, P.; Biller, W. T.; Martin, H. J. Am. Chem. Soc.
2002, 124, 827.
(2) (a) Notz, W.; Sakthivel, K.; Bui, T.; Zhong, G.; Barbas,
C. F. III. Tetrahedron Lett. 2001, 42, 199. (b) Cordova, A.;
Notz, W.; Zhong, G.; Betancort, J. M.; Barbas, C. F. III.
J. Am. Chem. Soc. 2002, 124, 1842. (c) Cordova, A.;
Synlett 2011, No. 4, 477–480 © Thieme Stuttgart · New York

480
M. Lu et al.
CLUSTER
(10) Hayashi, Y.; Yamaguchi, J.; Hibino, K.; Sumiya, T.;
Watanabe, S.; Tanaka, F.; Notz, W.; Barbas, C. F. III. J. Am.
Chem. Soc. 2002, 124, 1866. (d) Zhang, H.; Mifsud, M.;
Tanaka, F.; Barbas, C. F. III. J. Am. Chem. Soc. 2006, 128,
9630.
Urushima, T.; Shoji, M.; Hashizume, D.; Koshino, H. Adv.
Synth. Catal. 2004, 346, 1435.
(11) DFT calculations were carried out with the Gaussian 09
package (Frisch M. J. et al. Gaussian 03, Revision D.01,
Gaussian, Inc.: Wallingford, CT, 2004). The geometries for
two enamines and two possible transition states which
involved hydrogen bonding, are fully optimized by M06
method using 6-31G+(d,p) basis set. The geometries have
been confirmed to be equilibrium geometries or transition
states by the harmonic frequencies calculations at the same
level of theory. The energy values in the gas phase are
further calculated with M06/6-311++G(d,p) method based
on the optimal geometries and corrected with the zero-point
energy calculated from M06/6-31G+(d,p) method. Solvent
effect from DMSO is mimicked by the PCM model and
single point energy calculation is done with M06/6-
311++G(d,p) with the gas phase geometry.
(3) For recent reviews, see: (a) Verkade, J. M. M.; van Hemert,
L. J. C.; Quaedflieg, P. J. L. M.; Rutjes, F. P. J. T. Chem. Soc.
Rev. 2008, 37, 29. (b) Ting, A.; Schaus, S. E. Eur. J. Org.
Chem. 2007, 5797.
(4) (a) Müller, K.; Faeh, C.; Diederich, F. Science 2007, 317,
1881. (b) Gerebtzoff, G.; Li-Blatter, X.; Fischer, H.;
Frentzel, A.; Seelig, A. ChemBioChem 2004, 5, 676.
(c) Jeschke, P. ChemBioChem 2004, 5, 570.
(5) (a) Asymmetric Fluoroorganic Chemistry: Synthesis,
Applications and Future Directions; Ramachandran, P. V.,
Ed.; American Chemical Society: Washington DC, 2000.
(b) Biomedical Frontiers of Fluorine Chemistry; Ojima, I.;
McCarthy, J. R.; Welch, J. T., Eds.; American Chemical
Society: Washington DC, 1996. (c) Welch, J. T.;
Eswarakrishnan, S. In Fluorine in Bioorganic Chemistry;
John Wiley and Sons: New York, 1990. (d) Resnati, G.
Tetrahedron 1993, 49, 9385.
(12) Normally, the Mannich reaction will proceed at the more
substituted site, see ref. 8a. Reversal of regioselectivity was
observed previously, see ref. 8b.
(6) Zhong, G.; Fan, J.; Barbas, C. F. III. Tetrahedron Lett. 2004,
45, 5681.
(7) (a) Chen, X.-H.; Luo, S.-W.; Tang, Z.; Cun, L.-F.; Mi,
A.-Q.; Jiang, Y.-Z.; Gong, L.-Z. Chem. Eur. J. 2007, 13,
689. (b) Xu, X.-Y.; Wang, Y.-Z.; Cun, L.-F.; Gong, L.-Z.
Tetrahedron Asymmetry 2007, 18, 237.
(8) For extensive substrate scope of ketones, see: (a) Notz, W.;
Watanabe, S.-I.; Chowdari, N. S.; Zhong, G.; Betancort,
J. M.; Tanaka, F.; Barbas, C. F. III. Adv. Synth. Catal. 2004,
346, 1131. (b) Chowdari, N. S.; Ahmad, M.; Albertshofer,
K.; Tanaka, F.; Barbas, C. F. III. Org. Lett. 2006, 8, 2839.
(9) Two examples of Mannich reaction between fluoroacetone
with preformed PMP-imine derived from activated ethyl
glyoxalate were reported with low ee (61% for ref. 2b) of
14%, see: Cobb, A. J. A.; Shaw, D. M.; Longbottom, D. A.;
Gold, J. B.; Ley, S. V. Org. Biomol. Chem. 2005, 3, 84.
(13) We found the linear product was rather sensitive in the
presence of acid, and will slowly undergo dimerization upon
prolonged storage, thus all aldehydes used in this study were
distilled or recrystallized prior to usage, and the products
were stored after reduction to corresponding alcohol. For
reduction, see: Hayashi, Y.; Tsuboi, W.; Shoji, M.; Suzuki,
N. J. Am. Chem. Soc. 2003, 125, 11208.
(14) The addition of H2O will slightly improve the ee, for similar
observations, see: (a) Enders, D.; Grondal, C.; Vrettou, M.
Synthesis 2006, 2155. For our previous study on the role of
H2O in the transition state, see: (b) Lu, M.; Zhu, D.; Lu, Y.;
Hou, Y.; Tan, B.; Zhong, G. Angew. Chem. Int. Ed. 2008, 47,
10187. (c) Zhu, D.; Lu, M.; Chua, P. J.; Tan, B.; Wang, F.;
Yang, X.; Zhong, G. Org. Lett. 2008, 10, 4585. (d) Tan, B.;
Zeng, X.; Lu, Y.; Chua, P. J.; Zhong, G. Org. Lett. 2009, 11,
1927.
Synlett 2011, No. 4, 477–480 © Thieme Stuttgart · New York

Copyright of Synlett is the property of Georg Thieme Verlag Stuttgart and its content may not be copied or
emailed to multiple sites or posted to a listserv without the copyright holder's express written permission.
However, users may print, download, or email articles for individual use.