Biocatalytic synthesis of (2S)-5,5,5-trifluoroleucine and improved resolution into (2S,4S) and (2S,4R) diastereoisomers

Article abstract of DOI:10.1016/j.tetlet.2013.04.128

The introduction of fluorinated amino acid carrying a CF₃ moiety in therapeutical agents and protein engineering requires the accessibility of highly pure chiral samples in order to correctly understand the effect of fluorination on bioactivity. Here we report an easy enzymatic approach for the synthesis of (2S)-5,5,5-trifluoroleucine and its subsequent resolution into its (2S,4S) and (2S,4R) diastereoisomers. This strategy stands for a sustainable synthesis and enhanced resolution by means of alternative protecting groups for the amino/acid functionalities than that previously reported.

Full text of DOI:10.1016/j.tetlet.2013.04.128

Tetrahedron Letters 54 (2013) 3662–3665
Contents lists available at SciVerse ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Biocatalytic synthesis of (2S)-5,5,5-trifluoroleucine and improved resolution
into (2S,4S) and (2S,4R) diastereoisomers
⇑
Hernan Biava , Nediljko Budisa
Department of Biocatalysis, Institute of Chemistry, Technical University Berlin, Strasse des 17. Juni 135, 10623 Berlin, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
The introduction of fluorinated amino acid carrying a CF3 moiety in therapeutical agents and protein
engineering requires the accessibility of highly pure chiral samples in order to correctly understand
the effect of fluorination on bioactivity. Here we report an easy enzymatic approach for the synthesis
of (2S)-5,5,5-trifluoroleucine and its subsequent resolution into its (2S,4S) and (2S,4R) diastereoisomers.
This strategy stands for a sustainable synthesis and enhanced resolution by means of alternative protect-
ing groups for the amino/acid functionalities than that previously reported.
Received 25 March 2013
Revised 29 April 2013
Accepted 30 April 2013
Available online 9 May 2013
Keywords:
Ó 2013 Elsevier Ltd. All rights reserved.
Fluorinated amino acid
5,5,5-Trifluoroleucine
Chiral resolution
The introduction of amino acid analogs carrying a CF3 moiety
into the structure of therapeutical agents, proteins, and peptides
has been extensively applied as a tool for improving their activity
and stability.1,2 The enhanced pharmacological properties are gen-
erally explained in terms of an increased hydrophobicity of the tri-
fluoromethyl group when compared to the methyl group, which
most probably improves diffusion through cell membranes and
resistance to hydrolysis by endogenous proteases.
Highly fluorinated amino acids have been additionally exploited
in rational protein design, by introducing CF3 substituents into the
hydrophobic core of proteins and peptides.3,4 In principle, this
modification would lead to an enhanced stability of the folded
form of the biopolymer in presence of organic solvents or high
temperature conditions (so-called ‘fluorous effect’).5,6 These are
both greatly desired properties for biocatalytical purposes.
Due to their attractive applications, the synthesis of fluorinated
amino acids varying the number and position of the CF3 substitu-
ent has retained a significant role in the scientific community for
several decades.7 With the purpose of correctly pondering the out-
come of fluorination into bioactivity, it is imperative to control the
stereochemistry at the substituted carbon atoms. The amino acid
(2S)-5,5,5-trifluoroleucine (TFLeu, 1), which can exist as a mixture
of two diastereoisomers 1a and 1b (Fig. 1), is a typical example of
the many efforts devoted to stereospecific synthesis or preparative
resolution from commercial racemic sources.8–14
making their application unfeasible for bigger scale purposes.
Therefore, chromatographic resolution after derivatization of com-
mercial racemic mixtures is still the most practical method for pre-
paring pure samples for 1a and 1b isomers.
Here we report an easy chemoenzymatic approach for the syn-
thesis of (2S)-TFLeu followed by its subsequent resolution into the
(2S,4S) and (2S,4R) diastereoisomers. This synthetic strategy pro-
vides ready access to the S stereochemistry at the C
olution of N-acetyl derivates by acylases and subsequent necessity
for R-C
isomers recycling.15 The introduction of alternative pro-
a avoiding res-
a
tecting groups for the amino/acid functionalities allows a more
efficient separation by flash column chromatography on silica gel
than that previously reported.12,13
The synthesis of TFLeu on a milligram scale using Alcaligenes
faecalis whole cells has been previously reported.11 This procedure
makes use of an
a-keto ester and L-glutamic acid as amino donor
for a transamination reaction. Another biocatalytic reductive ami-
nation for the synthesis of (S)-5,5,5,50,50,50-hexafluoroleucine and
(S)-5,5,50,50-tetrafluoroleucine has been accomplished at gram
Even though a number of synthetic pathways for the stereoiso-
mers of TFLeu have been published, these methodologies are com-
plex, require multiple steps, or employ expensive chiral auxiliaries,
Figure 1. Chemical structure of (2S)-5,5,5-trifluoroleucine (1), (2S,4S)-5,5,5-triflu-
oroleucine (1a), and (2S,4R)-5,5,5-trifluoroleucine (1b). The asterisk in structure 1
indicates unresolved stereochemistry.
⇑
Corresponding author. Tel.: +49 30 314 23661; fax: +49 30 314 28279.
E-mail address: biava@chem.tu-berlin.de (H. Biava).
0040-4039/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.tetlet.2013.04.128

H. Biava, N. Budisa / Tetrahedron Letters 54 (2013) 3662–3665
3663
scale, employing phenylalanine dehydrogenase, NHþ4 and NADH as
reducing agent.16,17 The latter method requires fewer steps for the
obtention of the -keto ester intermediates and therefore we
a
examined the possibility to extend this methodology for the syn-
thesis of 1.
For this purpose, 1,1,1-trifluoroacetone was reacted with the
ylide 3 in a Wittig reaction to give the unsaturated pyruvate ester
4 (Scheme 1).18 No contamination with side-products was detected
after vacuum distillation, whether the reaction was performed in
THF, toluene, or benzene, contrary to the observation made by Chiu
and Chen.16 As previously reported,19 the reaction is highly diaste-
reoselective toward the E isomer (E:Z >95:5 judged from 1H NMR
spectra). The stereochemistry of the product was unambiguously
assigned as E by 1H{19F} NOE experiments (Fig. S1, Supplementary
data).20,21
Taking advantage of the high diasteroselectivty observed for the
synthesis of 4, we envisioned the likelihood for enantioselective
hydrogenation of the double bond in order to control the
stereochemistry at the second chiral center at this step of the
synthesis. For this purpose, we inspected (S,S0,R,R0)-TangPhos
(cyclooctadiene)rhodium(I)TBF22–24 and [((4S,5S)-Cy2-Ubaph-
ox)Ir(COD)]BARF25–27 as possible chiral catalysts (Fig. S2, Supple-
mentary data). Both metal complexes have been successfully
employed for the enantioselective hydrogenation of related sub-
strates, with good ee and acceptable yields. However, despite the
variety of conditions assayed, such as various solvents, catalyst load-
ing, temperature, reaction time, H2 pressure, etc., all experiments re-
sulted in a racemic mixture of the desired product plus a variable
amount of the over-hydrogenated alcohol, as judged by chiral
CG–MS and NMR analysis of the reaction products. The formation
of the alcohol as by-product seems to be attributable to the presence
of multiple electron-withdrawing groups and the conjugated car-
bonyl, which makes the double bond in 4 particularly electron defi-
cient. We decided then to solve this problem in a further step by
resolution of the final diastereoisomeric mixture.
Figure 2. 1H NMR in the d = 2.8–1.7 ppm window, showing diastereotopic hydro-
gen atoms for compounds (a) 1; (b) 1a; and (c) 1b.
it can be seen on the 1H NMR spectra for the resulting product after
purification by ion-exchange chromatography (Fig. 2a).29 The spec-
tral pattern corresponds to 1:1 mixture of the two isomers of (2S)-
trifluoroleucine. This result evidences once again the enzymatic
versatility of phenylalanine dehydrogenase to act on a wide range
of substrates, with different sizes and polarities. In order to make
the synthesis viable in terms of regeneration of the expensive
NADH, we employed in addition formate dehydrogenase (EC
232.844.2) to regenerate NADH, with concomitant oxidation of
the formate in the buffer.
When confined to a dialysis membrane, both enzymes could be
reused several times, without no evident loss in the activity or the
enantioselectivity, making this synthesis an excellent alternative to
the methodologies published by Matsumura et al.11
The enantioselectivity of the enzymatic reductive amination
was assessed by derivatization of new synthesized (2S)-5,5,5-tri-
fluoroleucine with (2R)-2-methoxy-2-trifluoromethylphenylacetyl
chloride (R-(ꢀ)-Mosher’s acid chloride, MTPA).30–32 The 1H NMR
of the derivatized sample is shown in Figure 3. Only two peaks
which correspond to OCH3 groups of the amide derived from dias-
teroisomers (2S,4S) and (2S,4R) plus unreacted MTPA were de-
tected, which confirms the enantioselectivity of the enzymatic
reaction.
A non-enantioselective hydrogenation of substrate 4 under low
H2 pressure and Pd/C as catalyst gave the corresponding reduced
a
-keto ester 5 in 82% yield.28 Interestingly, TiCl3(aq) did not perform
this reaction successfully, albeit it has been employed by Chiu for
the reduction of the hexafluoro analog.16 This result supports
author’s statement that the number and position of fluorine atom
strongly affects reactivity for these types of molecules.
In a further step, we studied the ability of the enzyme phenyl-
alanine dehydrogenase (EC 1.4.1.20) to act on substrate 5 which
has a marked polarity difference when compared to the substrates
assessed by Chiu (Scheme 1). After hydrolysis in the presence of
Na2CO3 to generate the free salt, an enzymatic amino reduction
catalyzed by phenylalanine dehydrogenase at pH 8.5 was per-
formed, in the presence of NADH as reducing agent. Interestingly,
both enantiomers of 5 were evenly processed by the enzyme, as
As previously stated, two related methodologies have been re-
ported for the resolution of all four isomers of 5,5,5-triflouroleu-
cine.12,13 These procedures involved conversion of the racemic
commercial samples into the N-Boc-5,5,5-trifluoroleucine t-butyl
Scheme 1. Biocatalytic synthesis of (2S)-5,5,5-trifluoroleucine. Yields correspond to individual steps.

3664
H. Biava, N. Budisa / Tetrahedron Letters 54 (2013) 3662–3665
practical point of view, the absence of any chromophores on the
TFLeu derivates makes TLC detection rather tricky, and contamina-
tion after fraction collection during flash chromatography were
usual for us when applying these methodologies.
Consequently, an alternative procedure was developed in order
to improve yield and TLC detection (Scheme 2). Initially (2S)-5,5,
5-trifluoroleucine was quantitatively converted into its N-Boc
protected derivate,33 which then reacted with benzyl chloride in
the presence of cesium carbonate, to give the N-Boc-5,5,5-triflu-
oroleucine benzyl esters (7) in 78% yield. These derivates were
easily separated into two diastereoisomers on silica gel using
n-hexane/t-butyl methyl ether (5:1) as the eluent and TLC detec-
tion by UV absorption or staining with phosphomolybdic acid.34
The time required for performing this chromatographic column
was at least three times less than that required for N-Boc-5,5,5-tri-
fluoroleucine t-butyl esters or N-Boc-5,5,5-trifluoroleucine methyl
esters. The higher difference of Rf values for the diasteroisomers 7a
and 7b permitted to make use of a more polar solvent mixture,
reducing elution time and without compromising on the purity
of the final products.
Removal of the protecting group by standard procedures gave
the hydrochloric salts of (2S,4S)-5,5,5,-trifluoroleucine (1a) and
(2S,4R)-5,5,5,-trifluoroleucine (1b) without racemization (Fig. 2b
and c).35 The optical purity of the samples was asserted by 1H
NMR, 19F NMR and specific rotation. Our total yield after the reso-
lution and the protecting group removal was around 78%, which is
higher than those previously reported in Refs.12,13
In brief, we have successfully extended an enzymatic procedure
with phenylalanine dehydrogenase for the enantioselective syn-
thesis of (2S)-5,5,5-trifluoroleucine in a simple, fast, and sustain-
able methodology, despite the high dipolar moment of the
pyruvate substrate. The use of alternative protecting groups, differ-
ent from those previously reported, allowed us to further resolve
the mixture into the (2S,4S) and (2S,4R) diastereoisomers by
means of flash chromatography with improved fraction detection
and higher yield.
Figure 3. 1H NMR in the d = 3.6-3.4 ppm window, showing the OMe signals for
(2S)-5,5,5-TFLeu after derivatization with MTPA.
esters or N-Boc-5,5,5-trifluoroleucine methyl esters. These mix-
tures could be both subsequently separated into two diastereoiso-
meric pairs by flash column chromatography. Following removal of
the protecting groups and enzymatic deacylation of the N-acetyl
derivates with porcine kidney acylase I, all four diastereoisomers
could be obtained in optically pure form.
Nevertheless, these procedures have certain disadvantages,
such us low yield after derivatization and removal of protecting
groups, use of highly toxic reagents such as pyridinium dichro-
mate, and generation of (2R)-N-acetyl amino acid as side-products
that have to be racemized for reutilization. Furthermore, from a
Acknowledgments
The authors are kindly grateful to Alexander von Humboldt
Foundation for financial support and postdoctoral fellowship. We
also would like to thank Technical University Berlin for the use
of facilities and equipment.
Supplementary data
Supplementary data associated (1H{19F} NMR spectra for com-
pound 4, chemical structure of metal catalysts and experimental
conditions) with this article can be found, in the online version,
at http://dx.doi.org/10.1016/j.tetlet.2013.04.128.
References and notes
1. Giese, C.; Lepthien, S.; Metzner, L.; Brandsch, M.; Budisa, N.; Lilie, H.
ChemMedChem 2008, 3, 1449–1456.
2. Begué, J.-P.; Bonnet Delpon, D. Bioorganic and Medicinal Chemistry of Fluorine;
Wiley & Sons: New York, 1993.
3. Merkel, L.; Budisa, N. Org. Biomol. Chem. 2012, 10, 7241–7261.
4. Bauer, B.; Marsch, E. Protein Sci. 2012, 21, 453–462.
5. Marsh, E. Chem. Biol. 2000, 7, R153–R157.
6. Jäckel, C.; Koksch, B. Eur. J. Org. Chem. 2005, 2005, 4483–4503.
7. Qiu, X.-L.; Qing, F.-L. Eur. J. Org. Chem. 2011, 3261–3278.
8. Weinges, K.; Kromm, E. Liebigs Ann. Chem. 1985, 90–102.
9. Kitazume, T.; Lin, J.; Yamazaki, T. Tetrahedron: Asymmetry 1991, 2, 235–237.
10. Vlasacova, V.; Tolman, V.; Zivny, K. J. Chromatogr. 1993, 639, 273–279.
11. Matsumura, Y.; Urushihara, M.; Tanaka, H.; Uchida, K.; Yasuda, A. Chem. Lett.
1993, 1255–1256.
12. Xing, X.; Filcera, A.; Kumar, K. J. Org. Chem. 2002, 67, 1722–1725.
13. Montclare, J.; Son, S.; Clark, G.; Kumar, K.; Tirrell, D. ChemBioChem 2009, 10,
84–86.
Scheme 2. Functional group protection for 1 and resolution into 1a and 1b. Yields
correspond to individual steps (otherwise noted).

H. Biava, N. Budisa / Tetrahedron Letters 54 (2013) 3662–3665
3665
14. Pigze, J.; Quach, T.; Molinski, T. J. Org. Chem. 2009, 74, 5510–5515.
15. Chenault, H.; Dahmer, J.; Whitesides, G. J. Am. Chem. Soc. 1989, 111, 6354–6364.
16. Chiu, H.-P.; Chen, R. Org. Lett. 2007, 9, 5517–5520.
17. Chiu, H.-P.; Suzuki, I.; Gullickson, D.; Ahmad, R.; Kokona, B.; Fairmen, R.; Cheg,
R. J. Am. Chem. Soc. 2006, 128, 15556–15557.
3.77 (m, 1H) 2.58 (m, 1H) 2.16 (m, 1H) 1.86 (m, 1H), 1.17 (d, J = 7.0 Hz, 3H). 13
C
NMR (D2O, 100 MHz): d 173.99, 173.77, 127.96 (q, J = 278 Hz), 127.89 (q,
J = 277 Hz), 52.40, 52.24, 34.31 (q, J = 26.3 Hz), 34.38 (q, J = 27.07 Hz) 30.90,
30.88, 14.03, 12.15 ½a 2D5 +14.9° (c 0.0261, 1 N HCl) HRMS analysis: [MH]+:
ꢁ
186.0742, observed: 186.0732.
18. Procedure for the synthesis of 5,5,5-trifluoro-4-methyl-2-oxo-pent-3-enoic acid
ethyl ester (4): 1,1,1-Trifluoroacetone (9 mmol, 1.35 mL) was added to a flask
containing ethyl (triphenylphosphoranylidene) pyruvate (6 mmol, 2.26 g)
dissolved in 13.5 mL benzene at 45 °C in an oil bath. The reaction was stirred
at 45 °C for 48 h. The solvent was then carefully evaporated under reduced
pressure. The residue was purified by vacuum distillation with a dry ice–
acetone bath for collecting the product as a light yellow oil (1.03 g, 3.9 mmol,
65% yield). 1H NMR (400 MHz, CDCl3): d 7.34 (m, 1H) 4.37 (q, J = 7.2 Hz, 2H)
2.28 (d, J = 1.5 Hz, 3H) 1.39 (t, J = 7.2 Hz, 3H); 13C NMR (100.6 MHz, CDCl3) d
182.79, 160.80, 145.86 (m), 121.63 (q, J = 275 Hz), 121.87 (m), 63.20, 14.11,
13.34; IR (liquid film, KBr): 2990m, 1735s, 1712s, 1449s, 1374s, 1299s, 1267s,
1130s, 1079s, 1012s; HRMS analysis calculated [MH]+: 211.0582, observed:
211.0576.
30. Dale, J.; Dull, D.; Mosher, H. J. Org. Chem. 1969, 34, 2543–2549.
31. Dale, J.; Mosher, H. J. Am. Chem. Soc. 1973, 95, 512–519.
32. Seco, J.; Quiñoá, E.; Riguera, R. Chem. Rev. 2004, 104, 17–117.
33. Bilgicer, B.; Fichera, A.; Kumar, K. J. Am. Chem. Soc. 2001, 123, 4393–4399.
34. Procedure for the synthesis of (2S)-N-Boc-5,5,5-trifluoroleucine benzyl ester (7)
and resolution into 7a and 7b: To an ice-cooled solution of N-Boc-(2S)-5,5,5-
trifluoroleucine (0.47 g, 1.65 mmol) in anhydrous DMF (10 mL) was added
Cs2CO3 (0.65 g, 2 mmol) and the mixture stirred at 0 °C during 1 h. Benzyl
bromide (0.25 mL, 2 mmol) was added dropwise to the reaction mixture and
maintained under stirring at 0 °C for another 30 min. The reaction was allowed
to reach room temperature overnight and then poured in to water (10 mL), and
extracted with MTBE (3 ꢂ 15 mL). The organic layers were combined, washed
with Brine and dried over Na2SO4. Rotary evaporation of the solvent gave a
yellowish oil that was purified by flash chromatography using hexane: MTBE
(5:1) as eluent. Three fractions were collected: 237 mg of (2S,4S)-N-Boc-5,5,5-
trifluoroleucine benzyl ester (7a) (38%), 161 mg of (2S,4R)-N-Boc-5,5,5-
trifluoroleucine benzyl ester (7b) (26%), and 90 mg of the mixture of 7a and
7b (14%).The total yield was 78%.
19. Dull, D.; Bacter, I.; Mosher, H. J. Org. Chem. 1967, 32, 1622–1623.
20. Kondratov, I.; Gerus, I.; Furmanova, M.; Vdovenko, S.; Kukhar, V. Tetrahedron
2007, 63, 7245–7246.
21. Fioravanti, S.; Colantoni, D.; Pellacani, L.; Tardella, P. J. Org. Chem. 2005,
70(3296), 329.
22. Tang, W.; Zhang, X. Org. Lett. 2002, 4, 4159–4161.
23. Tang, W.; Liu, D.; Zhang, X. Org. Lett. 2003, 5, 205–207.
24. Dai, Q.; Yang, W.; Zhang, X. Org. Lett. 2005, 7, 5343–5345.
(2S,4S)-N-Boc-5,5,5-trifluoroleucine benzyl ester (7a) Rf: Hexane/MTBE (5:1)
0.52; 1H NMR (CDCl3, 500 MHz) d 7.36 (m, 5H.) 5.20 (m, 2H) 5.10 (m, 1H, NH)
4.41 (m, 1H) 2.30 (m, 1H) 2.18 (m, 1H), 1.67 (m, 2H), 1.44 (s, 9H) 1.11 (d, 3H,
J = 6.4 Hz); 13C NMR (CDCl3, 100 MHz) d 171.79, 154.99, 134,98, 126.47 (q,
25. Ranu, B.; Dutta, J.; Guchhait, S. Org. Lett. 2001, 3, 2603–2605.
26. Lu, W. J.; Chen, Y.-W.; Hou, X.-L. Angew. Chem. 2008, 47, 10133–10136.
27. Church, T.; Andersson, P. Coord. Chem. Rev. 2008, 252, 513–531.
28. Procedure for the synthesis of 5,5,5-trifluoro-4-methyl-2-oxo-pentanoic acid ethyl
ester (5): Compound 4 (1.60 g, 7.61 mmol) was dissolved in 20 mL of THF and
transferred into a flask containing 10% Pd/C (0.40 g), previously purged with
argon. The mixture was hydrogenated at low H2 pressure for 4 h (TLC control).
The metal catalyst was removed by filtration and the filtrate was evaporated
under reduced pressure to obtain 5 as colorless oil (1.48 g, 6.19 mmol, 82%
yield). 1H NMR (CDCl3, 500 MHz): d 4.34 (q, J = 4.2 Hz, 2H), 3.21–2.83 (m, 3H),
1.38 (t, J = 7.2 Hz, 3H)1.16 (d, J = 6.4 Hz, 3H); 13C NMR (CDCl3, 100 MHz): d
191.15, 160.45, 127.83 (q, J = 275 Hz), 63.05, 39.63 (q, J = 2.7 Hz), 33.65 (q,
J = 27.8 Hz) 33.38, 14.10; IR (liquid film, KBr): 2987 m, 1732b, 1467s, 1389s,
1301s, 1260s, 1170s, 1114s, 1067s, 1014s; HRMS analysis calculated [MH]+:
213.0739, observed: 213.0731.
2
JCF = 279 Hz) 128.68, 128.65, 128.30, 67.38, 51.61; 35.96 (q, JCF = 26.9 Hz);
32.91 (m); 28.22, 12.96; 19F NMR (CDCl3, decoupled, 470 MHz): d ꢀ73.10 (s,
3F); IR (liquid film, KBr): 3356m, 2981s, 2933s, 1740s, 1715s,, 1499s, 1367s,
1265s, 1255s,1164s, 1132s, 1025s; ½ ꢁ ꢀ13.8° (c 0.0168, CHCl3); HRMS
a 2D5
analysis calculated [M+H]+: 376.1726, observed: 376.1736.
(2S,4R)-N-Boc-5,5,5-trifluoroleucine benzyl ester (7b):Rf: hexane/MTBE (5:1)
0.32; 1H NMR (CDCl3 500 MHz) d 7.36 (m, 5H.) 5.19 (m, 2H) 5.08 (m, 1H)
4.39 (m, 1H) 2.29 (m, 1H) 2.16 (m, 1H) 1.65 (m, 1H) 1.43 (s, 9H) 1.11 (d, 3H,
J = 6.9 Hz); 13C NMR (CDCl3, 100 MHz)
d 171.9, 155.2, 135.1, 126.93 (q,
2
J
CF = 278 Hz) 128.8, 128.7,128.5, 67.6, 51.7, 35.0 (q, JCF = 25.5 Hz) 33.8, 28.4,
12.1; 19F NMR (CDCl3, decoupled, 470 MHz): d ꢀ73.68 (s, 3F); IR (liquid film,
KBr) 3346m, 2982s, 2935s, 1743s, 1713s, 1499s, 1265s, 1255s,1159s, 1126s,
1021s;
½
a 2D5
ꢁ
ꢀ1.6° (c 0.0251, CHCl3); HRMS analysis calculated [M+H]+:
376.1726, observed: 376.1736.
29. Procedure for the synthesis of (2S)-2-amino-5,5,5-trifluoro-4-methyl-pentanoic
acid (1): Compound 5 (1.48 g, 7.0 mmol) was dissolved in 10% w/v Na2CO3(aq)
(7 mL) and stirred at room temperature overnight to give the pyruvate salt.
This solution was lyophilized and added to 1.5 M HCOONH4(aq) (20 mL) with
NADH (51.3 mg, 0.07 mmol), and the pH was adjusted to 8.5 with NH4OH 1 M
at room temperature. Then 20 units of formate dehydrogenase and 50 units of
35. Procedure for the synthesis: (2S,4S)-5,5,5-trifluoroleucine hydrochloride (1a) and
(2S,4R)-5,5,5-trifluoroleucine hydrochloride (1b): Compound 7a (142 mg,
0.38 mmol) was dissolved in EtOH and transferred into a flask containing
10% Pd/C (50 mg) and previously purged with Ar(g). The mixture was stirred
under low H2 pressure until the reaction was completed. The solvent was
eliminated under vacuum and the residue was treated with 4 mL of 4 M HCl in
dioxane at 0 °C under Ar(g) for 30 min. Vacuum removal of the solvent gave the
hydrochloride of 1a as a white powder (56 mg, 72% yield). 1H NMR (D2O,
400 MHz) d 3.97 (m, 1H) 2.60 (m, 1H) 2.27 (m, 1H) 2.83 (m,1H) 1.18 (d,
L
-phenylalanine dehydrogenase were dissolved into a capped dialysis
membrane (CelluSepÒ T3) with a small portion of the reaction mixture. The
dialysis membrane was then immersed into the solution with continuous
stirring. The pH was monitored and adjusted to 8.5 with NH4OH 1 M during
reaction. After 48 h, the solution was lyophilized and the solid crude was
purified by ion exchange chromatography (Dowex WX50-200) using water
followed by 1 M NH4OH as eluents. Fractions with the desired product were
combined and lyophilized to give 1 as light white powder (0.58 g, 3.11 mmol,
45% yield). The capped dialysis membrane could be reused several times
without apparent loss in activity and selectivity. 1H NMR (CDCl3, 500 MHz): d
J = 7.0 Hz, 3H); 19F NMR (D2O, decoupled, 470 MHz): d ꢀ73.23 (s, 3F); ½a 2D5
ꢁ
ꢀ3.6° (c 0.0267, 1N HCl); [MꢀH]ꢀ: 184.0591, observed: 184.0587. Compound
7b (73 mg, 0.20 mmol) was treated in a similar way to give the hydrochloride
of 1b as a white powder (38 mg, 85% yield). (1H NMR (D2O, 400 MHz) d 3.97
(m, 1H) 2.54 (m, 1H) 2.15 (m, 1H) 2.06 (m, 1H) 1.18 (d, J = 6.7 Hz, 3H); 19F NMR
(D2O, decoupled, 470 MHz): d ꢀ75.90 (s, 3F); ½a D25
+18.9° (c 0.0251, 1N HCl);
ꢁ
[MꢀH]ꢀ: 184.0591, observed: 184.0586.