Operationally convenient asymmetric synthesis of (S)- and (R)-3-amino-4,4,4-trifluorobutanoic acid. Part II. Enantioselective biomimetic transamination of 4,4,4-trifluoro-3-oxo-N-[(R)-1-phenylethyl]butanamide

Article abstract of DOI:10.1016/j.jfluchem.2006.04.004

An asymmetric synthesis of (S)- and (R)-3-amino-4,4,4-trifluorobutanoic acid via DBU-catalyzed asymmetric 1,3-proton shift transfer reaction of (Z)-3-[(R)-1-phenylethylamino]-4,4,4-trifluoro-N-[(R)-1-phenylethyl]but-2-enamide has been developed. The intermediate Schiff base (S,R′)-9 was found to be relatively configurationally stable under the highly basic reaction conditions allowing preparation of the target amino acid in high chemical yield and enantioselectivity. This method was demonstrated to be practical for large (>25 g) scale synthesis of the target β-amino acid.

Full text of DOI:10.1016/j.jfluchem.2006.04.004

Journal of Fluorine Chemistry 127 (2006) 930–935
www.elsevier.com/locate/fluor
Operationally convenient asymmetric synthesis of (S)-
and (R)-3-amino-4,4,4-trifluorobutanoic acid
Part II. Enantioselective biomimetic transamination
of 4,4,4-trifluoro-3-oxo-N-[(R)-1-phenylethyl]butanamide
*
Vadim A. Soloshonok , Hironari Ohkura, Manabu Yasumoto
Department of Chemistry and Biochemistry, University of Oklahoma, Norman, OK 73019, United States
Received 23 March 2006; received in revised form 15 April 2006; accepted 17 April 2006
Available online 25 April 2006
Abstract
An asymmetric synthesis of (S)- and (R)-3-amino-4,4,4-trifluorobutanoic acid via DBU-catalyzed asymmetric 1,3-proton shift transfer reaction
of (Z)-3-[(R)-1-phenylethylamino]-4,4,4-trifluoro-N-[(R)-1-phenylethyl]but-2-enamide has been developed. The intermediate Schiff base (S,R0)-9
was found to be relatively configurationally stable under the highly basic reaction conditions allowing preparation of the target amino acid in high
chemical yield and enantioselectivity. This method was demonstrated to be practical for large (>25 g) scale synthesis of the target b-amino acid.
# 2006 Elsevier B.V. All rights reserved.
Keywords: 1,3-Proton shift reaction; Fluorine and compounds; Imines; Enamines; Operationally convenient conditions; Asymmetric synthesis; Biomimetic
reductive methodology
1. Introduction
extraction of Schiff base. However, on a lager scale the
extraction step is very complicated by the very viscous nature of
the DBU salt and results in relatively low (50–60%) recovery of
the product 4. Another issue with the current method is a partial
racemization of the intermediate product 4 under the reaction
conditions allowing preparation of 4 in about 80%ee. Though
final amino acid 6 can be obtained in high enantiomeric purity
after precipitation from MeOH, it still leads to reduced overall
yields. Therefore, to improve at least one of these issues we
decided to study an alternative approach using (Z)-3-[(R)-1-
phenylethylamino]-4,4,4-trifluoro-N-[(R)-1-phenylethyl]but-2-
enamide 7 (Scheme 2) as a starting compound for its
biomimetic transamination to target amino acid 6 [4].
In the preceding paper [1], we reported an operationally
convenient preparation of enantiomerically pure (S)- and (R)-3-
amino-4,4,4-trifluorobutanoic acid 6 (Scheme 1) via enantio-
selective biomimetic transamination [2] of isopropyl-4,4,4-
trifluoro-3-oxo-butanoate 1.
While this method is definitely convenient, and has been
already used by others [3], for preparation of gram quantities of
enantiomerically pure amino acid 6, we found that its
application on 5–10 g scale is rather complicated. The major
problem we encountered in scaling up the method is
problematic purification and separation of Schiff base 4 from
the excess of DBU. Thus, the current procedure for purification
of 4 involves conversion of DBU into its salt with acetic acid
followed by an extraction of product 4 with hexanes from the
resulting reaction mixture. On a small scale (up to 5 g) this
procedure is rather efficient and allows for virtually complete
2. Results and discussion
2.1. Enantioselective biomimetic transamination of
isopropyl-(Z)-3-((R)-1-phenylethylamino)-4,4,4-trifluoro-
N-((R)-1-phenylethyl)but-2-enamide
The starting enamine-amide 7 was chemoselectively
prepared in high chemical yields from commercially available
and inexpensive keto ester 1 and optically pure (R)-1-
DOI of original article: 10.1016/j.jfluchem.2006.04.003.
* Corresponding author. Tel.: +1 405 325 8279; fax: +1 405 325 6111.
E-mail address: vadim@ou.edu (V.A. Soloshonok).
0022-1139/$ – see front matter # 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jfluchem.2006.04.004

V.A. Soloshonok et al. / Journal of Fluorine Chemistry 127 (2006) 930–935
931
Scheme 1.
phenylethylamine, according to the previously published
procedure [5].
Considering the data obtained, we can make several general
conclusions. First, similar to what was observed in the
isomerization of enamine-ester 3 (Scheme 1) [4] the amount
of the base-catalyst had a noticeable effect on the rate of the
isomerization of enamine-amide 7 to Schiff base 9. However,
the isomerization of 7–9 occurred at relatively slower rates,
requiring about 24 h to reach a 90% conversion level even with
3.5 equiv. of the base (Table 1, entry 7). Second, the optimum
amount of the base might be about 3.5 equiv. as the application
of 4.0 and 5.0 equiv. had virtually no accelerating effect on the
reaction rate (entry 7 versus entries 8 and 9). Application of
acetonitrile as a solvent for this reaction (entry 10) resulted in
slower reaction rate, probably due to stabilization by the polar
solvent of the enamine form of 7 and its slow isomerization
to intermediate imine 8. The third and most impressive
conclusion, is that the stereochemical outcome of the
isomerization was only slightly influenced by the amount of
Considering structure of compound 7 one may raise a
concern regarding relatively high acidity of the N–H proton in
the amide moiety, which can be ionized under the basic
conditions and interfere with the desired isomerization reaction.
Therefore we were pleased to find that this structural feature
does not pose any problem and enamine-amide 7 can be cleanly
isomerized in the presence of DBU to Schiff base 9. As we did
in our previous study on the biomimetic transamination of
isopropyl-4,4,4-trifluoro-3-oxobutanoate 1 (Scheme 1) [4], we
decided to conduct a systematic investigation of the effect of the
base-catalyst amount on the rate and stereochemical outcome
of the isomerization of enamine-amide 7 to Schiff base 9. We
conducted a series of isomerizations without a solvent at 75 8C
gradually increasing the amount of the base starting from 0.5 to
5 equiv. The results are summarized in Table 1.
Scheme 2.

932
V.A. Soloshonok et al. / Journal of Fluorine Chemistry 127 (2006) 930–935
Table 1
Isomerization of (R,R0)-7 to (S,R0)-9 using gradually increasing amounts of
DBUa
Table 2
Racemization of (S,R0)-9 at 100 8C in the presence of 3.5 equiv. of DBU
Time (h)
%ee
Entry
Equivalent
Yield (%)
6 h
0
6
92.5
86.5
73.6
54.4
12 h
24 h
36 h
%ee
18
48
1
2
0.5
1.0
1.6
2.0
2.5
3.0
3.5
4.0
5.0
2.5b
6.3
15.0
26.4
34.8
39.0
39.8
47.4
46.2
53.3
8.4
9.0
23.4
48.0
57.4
63.2
63.8
73.3
64.0
71.2
14.9
16.5
44.6
75.2
81.2
87.8
87.5
91.8
90.7
93.1
34.1
27.1
59.9
86.9
91.6
95.9
96.5
98.3
97.4
98.1
44.4
95.8
95.5
94.0
93.8
93.5
93.6
93.0
92.8
92.5
71.0
3
4
5
withdrawing ester group would render the stereogenic C–H
more acidic and therefore less configurationally stable.
Nevertheless, considering the fact that in both Schiff bases 4
and 9 the stereogenic C–H is bearing three electron-
withdrawing substituents, two of which, the trifluoromethyl
group and Schiff base function, are very strong, their relatively
high configurational stability was truly remarkable. This
unexpected observation allows us to suggest for the first time a
novel dual role of a trifluoromethyl group attached to a
stereogenic carbon. Thus, on one hand, a trifluoromethyl group
definitely acidifies an adjacent C–H bond making the
corresponding stereogenic center less configurationally stable.
On the other hand, being a bulky [6] sphere with nine unshared
electron pairs it can act as a base/nucleophile-repelling
protecting group preventing from the corresponding racemi-
zation. This interesting, from our stand point, suggestion
should be studied in more detail.
6
7
8
9
10
a
Unless indicated, the reactions were conducted at 75 8C without a solvent.
The progress of the reactions was monitored by 19F NMR and GC analysis.
b
The reaction was conducted in acetonitrile.
the base. Thus, the highest value of enantioselectivity obtained
in the reaction catalyzed by 0.5 equiv. of the base (entry 1) was
95.8%ee, while in the reaction catalyzed by 4.0 equiv. the
enantioselectivity was only 3% lower (92.8%ee, entry 8). This
observation was in sharp contrast to the isomerization of
enamine-ester 3 (Scheme 1) where the difference in %ee
between the corresponding reaction was 15.4% [4].
These data indicated that product 9 is relatively config-
urationally stable in the presence of DBU at 75 8C. To
investigate this subject in more detail we studied racemization
rate of Schiff base 9 in the presence of 3.5 equiv. of DBU at
100 8C. The data obtained are summarized in Table 2.
Considering these very harsh reaction conditions the rate of
the racemization was not high. Thus, after 48 h the
enantiomeric purity of the Schiff base 9 was still high,
54.4%ee. It should be noted that under these conditions the rate
of chemical decomposition was higher than the racemization.
To rationalize the difference in the racemization rates between
Schiff bases 4 and 9, one may consider the difference in the
electron-withdrawing ability between the ester and amide
groups. Thus, in compound 4 (Scheme 1) the more electron-
2.2. Mechanistic considerations
The mechanistic rationale for the observed stereochemical
outcome might be similar to what we proposed for the
isomerization of enamine-ester (R)-3 to Schiff base (S)-4
(Scheme 1). In this regard we can suggest two transition states
(TSs) A and B (Fig. 1) leading, respectively, to the (S,R0)-
configured product 9 (major enantiomer) and the product 9 of
(R,R0) absolute configuration (minor enantiomer). In the TS A,
leading to the major enantiomer, trifluoromethyl and phenyl
groups play the role of stereocontrolling substituents regardless
the fact that the alkyl substituent bearing the amide function is
substantially bigger [1,6].
Fig. 1.

V.A. Soloshonok et al. / Journal of Fluorine Chemistry 127 (2006) 930–935
933
Scheme 3.
2.3. Determination of enantiomeric purity of products
isomerized product 9 and 3.5 equiv. of DBU (or any other,
depending on the conditions applied, see Table 1), was treated
with 1N HCl to selectively hydrolyze the Schiff base function in
compound 9. The resultant acetophenone was extracted in
ethereal phase while hydrochloride salts of amide 13 and DBU
were collected in the aqueous phase. Further separation of the
target product 14 and excess DBU was conveniently achieved
by neutralizing the salts in the aqueous phase with NaHCO3
followed by an extraction of free amine-amid 14 with ethyl
acetate. This separation procedure was reproduced on relatively
large scale (25 g) with overall (three steps) yield of 83%.
Finally, the amide 14 was hydrolyzed with concentrated
hydrochloric acid furnishing free amino acid 6 as shown in
Scheme 2.
For determination of enantiomeric purity of products 9
these compounds were hydrolyzed by concentrated hydro-
chloric acid to furnish a mixture of salts 10 and 5 (Scheme 3).
Without purification these salts were converted to the
corresponding amides 11 and 12 using standard [1] two-step
procedure indicated in Scheme 3. Enantiomeric purity of
compounds 11 and 12 was determined on crude mixtures [7]
using HPLC analysis of a chiral stationary phase. It is very
interesting to note that enantiomeric purity of the amide 11 was
found to be only 97.2%ee, while the producer of the
corresponding starting amine has claimed at least >99%ee.
Since some racemization is theoretically possible during the
reaction in the presence of DBU, this observation should be
considered as some sort of accusation, however, we are
currently are studying this issue in detail.
In conclusion, we found that DBU-catalyzed enantioselec-
tive biomimetic transamination of 4,4,4-trifluoro-3-oxo-N-
[(R)-1-phenylethyl]butanamide has at least two very significant
synthetic advantages over an alternative approach via
transamination of isopropyl-4,4,4-trifluoro-3-oxo-butanoate 1
[1]. These include, higher enantioselectivity and convenient
separation of the target products from excess DBU. On the
other hand the transamination of isopropyl-4,4,4-trifluoro-3-
oxo-butanoate 1 occurs at substantially higher reaction rates
and is convenient for preparation of gram amounts of the target
amino acid 6.
2.4. Removal of excess DBU and scaling-up issues
The general procedure for operationally convenient and
scalable isolation of the target amino acid 6 from acetophenone
and excess DBU is given in Scheme 4 (see also: Scheme 2).
Taking advantage of relative hydrolytic stability of the amide
functional group, the crude reaction mixture, containing
Scheme 4.

934
V.A. Soloshonok et al. / Journal of Fluorine Chemistry 127 (2006) 930–935
3. Experimental
HCl (6N) (150 mL) at r.t. and refluxing the resultant mixture at
120 8C for 12 h. Water was removed under reduced pressure to
provide the crude solid residue (the amino acid hydrochloride
salt and phenethylamine hydrochloride salt). Thus obtained
solid was dissolved in dried ethanol (40 mL) at r.t. and treated
with propylene oxide (20 mL). The resultant mixture was
stirred for 30 min at r.t. and evaporated under reduced pressure
to provide white solid residue which was washed with i-
propanol (40 mL). The crystalline material obtained was
filtered furnishing the target b-amino acid in 75% overall yield
and of 93%ee. The enantiomeric purity can be further increased
by crystallization of thus obtained 6 from MeOH [1].
3.1. General methods
Unless otherwise noted, all reagents and solvents were
obtained from commercial suppliers and used without further
purification. All the reactions were carried out under regular
atmosphere without any special caution to exclude air. Unless
indicated, 1H, 19F and 13C NMR spectra, were taken in CDCl3
solutions at 299.95, 282.24 and 75.42 MHz, respectively, on an
instrument in the University of Oklahoma NMR Spectroscopy
Laboratory. Chemical shifts refer to TMS (1H NMR and 13C
NMR) and fluorotrichloromethane (19F NMR) as the internal
standards. Yields refer to isolated yields of products of greater
3.5. Synthesis of the derivatives 12 and determination of
their enantiomeric purity
1
than 95% purity as estimated by H NMR spectrometry. All
new compounds were characterized by 1H, 19F and 13C NMR,
by mass spectrometry and/or elemental analysis.
Compound 9 (1.0023 g, 3.8511 mmol) was hydrolyzed with a
solution HCl (6N) (30 mL) in a reactor (25 mL) by refluxing at
120 8C for 12 h. Water was removed under reduced pressure to
providethecrudereactionmixturewhichwastreatedwiththionyl
chloride (4.5 mL) and ethanol (20 mL) and stirred for 12 h at r.t.
Excessofethanolwasremovedunderreducedpressuretoprovide
the crude reaction mixture which was treated with a solution of
triethylamine (2.5 mL) in CH2Cl2 (10 mL) at r.t., MgSO4 (0.5 g)
wasaddedandstirredfor5 min.Tothereactionmixtureobtained,
3,5-dinitrobenzoyl chloride (0.3043 g, 1.320 mmol) was added
andtheresultantmixturewasstirredfor 1 h. Thereactionmixture
was evaporated to dryness under reduced pressure, dissolved in
CHCl3 and passed through short silica-gel column [8] furnishing
derivative 12 suitable for HPLC analysis.
3.2. 3-((R)-1-phenylethylamino)-4,4,4-trifluoro-N-((R)-1-
phenylethyl)but-2-enamide (7) [5]
1H NMR d 1.51 (d, 3H, J = 6.6 Hz), 1.52 (d, 3H, J = 7.2 Hz),
4.66 (dq, 1H, J = 10.2, 6.6 Hz), 4.96 (s, 1H), 5.16 (q, 1H,
J = 7.2 Hz), 5.49 (d, J = 7.2 Hz), 7.21–7.36 (m, 10H), 9.12 (d,
1H, J = 10.2 Hz); 19F NMR d ꢀ66.64 (s); 13C NMR d 22.1,
25.1, 48.4, 53.8, 88.2 (q, J = 5 Hz), 120.4 (q, J = 277 Hz),
125.3, 126.0, 126.9, 127.3, 128.5, 128.6, 143.2, 144.5, 145.4 (q,
J = 31 Hz);
MS(TOF)
[M + Na],
[C20H21F3N2O + Na]: 385.1504, found: 385.1678.
Calcd.
for
3.3. (3S)-3-(1-phenylethylideneamino)-4,4,4-trifluoro-N-
((R)-1-phenylethyl)butanamide (9)
Column: SUMICHIRAL OA-4500; eluent: n-hexane/
dichloromethane/ethanol = 60/30/10; retention times: 8.4 for
(S) and 11.8 for (R).
1H NMR d 1.44 (d, 3H, J = 6.8 Hz), 2.19 (s, 3H), 2.68 (dd,
1H, J = 9.4, 13.9 Hz), 2.84 (dd, 1H, J = 3.1, 13.9 Hz), 4.72 (m,
1H), 5.04 (m, 1H), 5.96 (m, 1H), 7.00–7.14 (m, 5H), 7.34–7.46
(m, 3H), 7.69–7.71 (m, 2H); 19F NMR d 75.1 (d, J = 8.4 Hz);
13C NMR d 15.9, 21.6, 37.5, 48.9, 59.7 (q, J = 28 Hz), 125.3 (q,
J = 280 Hz), 125.9, 127.1, 127.1, 128.2, 128.5, 120.3, 140.1,
142.6, 167.9, 171.4. MS(TOF) [M + H]+, Calcd. for
[C20H21F3NO2 + H]+: 363.1684, found: 363.1887.
Acknowledgements
This work was supported by Department of Chemistry and
Biochemistry, University of Oklahoma. The authors gratefully
acknowledge generous financial support from Central Glass
Company (Tokyo, Japan) and Ajinomoto Company (Tokyo,
Japan).
3.4. Large scale procedure for preparation of (S)-3-amino-
4,4,4-trifluorobutanoic acid (6)
References
[1] V.A. Soloshonok, H. Ohkura, M. Yasumoto, J. Fluor. Chem. 127 (2006)
924–929.
To a reactor (250 mL), enamine-amide 7 (25.154 g,
69.39 mmol) and DBU (36.995 g, 243.0 mmol) were added
under air. The mixture was stirred for 36 h at 75 8C. The
resultant mixture was treated with HCl (1N) (150 mL) and Et2O
(300 mL) and stirred at r.t. for 2 h. An organic phase was
extracted with water (3 ꢁ 100 mL). The aqueous phase was
evaporated under reduced pressure to provide the crude solid
which was treated with saturated NaHCO3 (aq.) until pH 9. The
resultant mixture was extracted with AcOEt (3 ꢁ 150 mL). The
separated organic phase was dried and evaporated under
reduced pressure to furnish the product 14 (14.981 g, 83%).
Without purification compound 14 was hydrolyzed by adding
[2] For full papers on biomimetic transamination of fluorinated carbonyl
compounds, see:
(a) T. Ono, V.P. Kukhar, V.A. Soloshonok, J. Org. Chem. 61 (1996) 6563–
6569;
(b) V.A. Soloshonok, T. Ono, Tetrahedron 52 (1996) 14701–14712;
(c) V.A. Soloshonok, T. Ono, J. Org. Chem. 62 (1997) 3030–3031;
(d) V.A. Soloshonok, V.P. Kukhar, Tetrahedron 52 (1996) 6953–6964;
(e) V.A. Soloshonok, V.P. Kukhar, Tetrahedron 53 (1997) 8307–8314;
(f) V.A. Soloshonok, I.V. Soloshonok, V.P. Kukhar, V.K. Svedas, J. Org.
Chem. 63 (1998) 1878–1884.
[3] J. Jiang, H. Shah, R.J. DeVita, Org. Lett. 5 (2003) 4101–4103.
[4] Some details of this work were reported in the following short commu-
nication:

V.A. Soloshonok et al. / Journal of Fluorine Chemistry 127 (2006) 930–935
935
V.A. Soloshonok, T. Ono, I.V. Soloshonok, J. Org. Chem. 62 (1997) 7538–
7539.
(d) V.A. Soloshonok, A.D. Kacharov, D.V. Avilov, T. Hayashi, Tetrahedron
Lett. 37 (1996) 7845–7848;
[5] H. Ohkura, D.O. Berbasov, V.A. Soloshonok, Tetrahedron 59 (2003) 1647–
1656.
(e) V.A. Soloshonok, A.D. Kacharov, D.V. Avilov, K. Ishikawa, N. Naga-
shima, T. Hayashi, J. Org. Chem. 62 (1997) 3470–3479;
(f) M. Kanai, M. Yasumoto, Y. Kuriyama, K. Inomiya, Y. Katsuhara, K.
Higashiyama, A. Ishii, Org. Lett. 5 (2003) 1007–1010.
[7] Compounds 12 were found to undergo substantial enantiomer self-dispro-
portionation on regular, achiral silica-gel columns:
[6] For our discussions on quite controversial issue of a stereochemical bulk of
a trifluoromethyl group, see the following papers and references in them:
(a) V.A. Soloshonok, V.P. Kukhar, S.V. Galushko, N.Y. Svistunova, D.V.
Avilov, N.A. Kuzmina, N.I. Raevski, Y.T. Struchkov, A.P. Pisarevsky, Y.N.
Belokon, J. Chem. Soc. Perkin Trans. 1 (1993) 3143–3155;
(b) V.A. Soloshonok, T. Hayashi, K. Ishikawa, N. Nagashima, Tetrahedron
Lett. 35 (1994) 1055–1058;
(a) V.A. Soloshonok, Angew. Chem. In. Ed. Engl. 45 (2006) 766–769;
(b) V.A. Soloshonok, D.O. Berbasov, J. Fluor. Chem. 127 (2006) 597–
603.
(c) V.A. Soloshonok, D.V. Avilov, V.P. Kukhar, Tetrahedron 52 (1996)
12433–12442;
[8] We found that this procedure is safe in regard of the enantiomer self-
disproportionation effect.