Obtention of enantiomerically pure 5,5,5-trifluoro-l-isoleucine and 5,5,5-trifluoro-l-alloisoleucine

Article abstract of DOI:10.1016/j.jfluchem.2013.07.021

The fluorinated amino acids 5,5,5-trifluoro-l-isoleucine and 5,5,5-trifluoro-l-alloisoleucine constitute a pair of diastereoisomers with attractive properties as synthons for the obtention of biologically active molecules. However, difficulties in their synthesis and chiral resolution have restricted their use to a few examples of in vivo and in vitro applications. Here we described an alternative synthetic pathway for these highly valuable fluorinated amino acids and further resolution into enantiomerically pure samples by aminoacylase-catalyzed hydrolysis of their N-acetyl derivates. Resolution is achieved by means of distinctive preference toward substrate stereochemistry exhibited by acylase I from Aspergillus melleus compared to the analogous enzyme derived from porcine kidney. Straightforward access to enantiomerically pure samples for these fluorinated amino acids will definitely expand their relevance as building blocks for the production of new drugs and innovative biomaterials.

Full text of DOI:10.1016/j.jfluchem.2013.07.021

Journal of Fluorine Chemistry 156 (2013) 372–377
Contents lists available at ScienceDirect
Journal of Fluorine Chemistry
journal homepage: www.elsevier.com/locate/fluor
Short communication
Obtention of enantiomerically pure 5,5,5-trifluoro-
5,5,5-trifluoro- -alloisoleucine
L-isoleucine and
L
*
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 fluorinated amino acids 5,5,5-trifluoro-L-isoleucine and 5,5,5-trifluoro-L-alloisoleucine constitute a
Received 7 June 2013
pair of diastereoisomers with attractive properties as synthons for the obtention of biologically active
molecules. However, difficulties in their synthesis and chiral resolution have restricted their use to a few
examples of in vivo and in vitro applications. Here we described an alternative synthetic pathway for
these highly valuable fluorinated amino acids and further resolution into enantiomerically pure samples
by aminoacylase-catalyzed hydrolysis of their N-acetyl derivates. Resolution is achieved by means of
distinctive preference toward substrate stereochemistry exhibited by acylase I from Aspergillus melleus
compared to the analogous enzyme derived from porcine kidney. Straightforward access to
enantiomerically pure samples for these fluorinated amino acids will definitely expand their relevance
as building blocks for the production of new drugs and innovative biomaterials.
Received in revised form 19 July 2013
Accepted 27 July 2013
Available online 17 August 2013
Keywords:
5,5,5-Trifluoroisoleucine
5,5,5-Trifluoroalloisoleucine
2-Amino-5,5,5-trifluoro-3-
methylpentanoic acid
Acylase I
ß 2013 Elsevier B.V. All rights reserved.
Chiral resolution
1. Introduction
stereochemistry at all present chiral centers. For example,
introduction of 4-(R)-fluoro-L-proline into collagen can dramati-
In the last decades, fluorinated amino acids have found
numerous applications in several fields of structural and medicinal
biochemistry, such as synthons for therapeutic peptides, pharma-
ceuticals, agrochemicals, surface coatings, biological tracers, etc.
[1,2]. The introduction of fluorine has proved to dramatically
change the physical, chemical and biological properties of these
substances as a consequence of the high electronegativity and low
polarizability of the fluorine atom.
Highly fluorinated compounds also tend to segregate them-
selves from hydrophobic or hydrophilic phases and this phenom-
enon is referred to as the ‘‘fluorous effect’’ [3,4]. The fluorous effect
has been recently exploited by introducing highly fluorinated
amino acids into the hydrophobic core of proteins and peptides in
order to generate highly stable biopolymers with improved
resistance to high temperature or organic solvents denaturaliza-
tion [5,6]. Hexafluoroleucine, trifluoroleucine, hexafluorovaline
and trifluorovaline are some examples of fluorinated amino acids
carrying CF3 moieties that have been successfully employed for
this purpose, whether by in vivo or in vitro strategies [7].
cally improve its stability, while stereoisomer 4-(S)-fluoro-L-
proline actually results in a distinctly destabilization of the
collagen helix [8]. Furthermore the amino acid (2S,3R)-trifluorova-
line was efficiently activated by endogenous aminoacyl-tRNA
synthetases and incorporated into target proteins in vivo, while its
isomers (2S,3S) was completely inactive [9].
Other interesting applications for trifluorinated analogs of
isoleucine (1) and -alloisoleucine (2) (Fig. 1) were also described.
Incorporation of 5,5,5-trifluoro- -isoleucine (3) into murine inter-
L-
L
L
leukine-2 and a leucine zipper peptide derived from GCN4
transcriptional factor resulted in stable cognates with higher
resistance to thermal denaturalization compared to their wild-
type analogs [10,11]. Interestingly, both fluorinated proteins were
able to retain their DNA-binding properties. These studies showed
that 3 can be activated in vivo for incorporation into proteins
expressed in Escherichia coli. In these studies, 5,5,5-trifluoroiso-
leucine was employed as a
In addition, 5,5,5-trifluoro-
a building block for the asymmetric synthesis of a
D
,
L
-enantiomeric mixture.
-alloisoleucine (4) has been used as
-secretase
L
g
However, the outcome imprinted by the introduction of
fluorinated amino acids into biomaterials depends not only on
the number and position of the fluorine atoms but also on the
inhibitor, a promising agent for the treatment of Alzheimer’s
disease [12].
Three alternative methodologies for synthesizing racemic 2-
amino-5,5,5-trifluoro-3-methylpentanoic acid (5) have been
reported by Mu¨ller and slightly modified by Tirrell [10,13].
However, the most efficient approach requires preparative
electrolysis, which is not commonly available in modern organic
*
Corresponding author. Tel.: +49 30 314 23661; fax: +49 30 314 28279.
E-mail address: biava@chem.tu-berlin.de (H. Biava).
0022-1139/$ – see front matter ß 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jfluchem.2013.07.021

H. Biava, N. Budisa / Journal of Fluorine Chemistry 156 (2013) 372–377
373
Fig. 1. L-Isomers of 2-amino-3-methylpentanoic acid and their 5,5,5-trifluorinated analogs. The corresponding D-isomers can be obtained by inverting the configuration at C2
and C3 atoms.
chemistry laboratories. At the same time, resolution of the racemic
mixture into two diastereoisomeric pairs was achieved by means
of fractional recrystallization, assuming similar solubility proper-
ties as those known for 1 and 2 [14]. This method results
impractical on a small scale and usually leads to a pure sample for
applications on a small scale and normally produces a highly
complex mixture of several products besides the desired one.
In order to generate an easy and fast methodology for the
generation of 8 we followed the procedure shown in Scheme 1,
based on minor modifications from reported data [16].
the less soluble of the diastereoisomer pairs (5,5,5-trifluoro-
isoleucine) while only partially pure samples can be prepared for
D
,
L
-
Reaction between carboethoxymethylenetriphenyl-phosphor-
ane and 4,4,4-trifluoro-2-butanone in dry diethyl ether produces
the unsaturated ester 10 as mixture of E,Z isomers in 51% yield.
After purification by vacuum distillation, this product was
hydrogenated under low pressure to obtain the saturated ester
which was then hydrolyzed to obtain the free acid 8. Subsequent
treatment with thionyl chloride and bromine generated the
corresponding brominated acid 9 as a mixture of diastereoisomers
in 72% yield. Direct amination using an excess of concentrated
ammonia aqueous solution probed to be the best and easier
alternative to generate 5. This product was obtained in 40% yield
from 9 in a very straightforward manner when compared to the
former methodologies.
the more soluble constituent of the mixture (5,5,5-trifluoro-
D,L-
alloisoleucine).
Difficulties in their synthesis, isolation and the absence of a
methodology to access enantiomerically pure samples might be
the reasons for the restricted reported applications for these highly
valuable fluorinated amino acids hitherto.
Here we report a straightforward methodology for the synthesis
of 2-amino-5,5,5-trifluoro-3-methylpentanoic acid as a mixture of
diastereoisomers and first access to enantiomeric pure samples of
5,5,5-trifluoro-L-isoleucine and 5,5,5-trifluoro-L-alloisoleucine
through enzymatic resolution of their N-acetyl-derivates. This
methodology takes advantage of the distinction on substrate
configuration exhibited by commercial acylase I from Aspergillus
melleus compared to the equivalent enzyme from porcine kidney,
allowing resolution according to the stereochemistry at the C2 and
C3 atoms. Comparison with reported 4,4,4-trifluorovaline stereo-
isomers, which also bear two analogous chiral centers, allowed us
to unambiguously assign the absolute configuration in 3 and 4
despite the impossibility of obtaining appropriate samples for
crystallographic studies.
2.2. Resolution of 5,5,5-trifluoro-L-isoleucine and 5,5,5-trifluoro-L-
alloisoleucine
According to the published data [10,11,13], pure samples of
both pairs of diastereoisomers could be achieved by means of
fractional recrystallization. To our disappointment, this procedure
generated partially pure samples of 3 and 4, with contamination of
one pair of diastereoisomers with the other. We assayed more than
a few alternatives in order to improve this result, such as fractional
recrystallization of the N-acetyl derivates [14,17] and their
ammonium salts with various solvent mixtures [18], chro-
matographic purification of several derivates of 5 with protected
acid/amino functionalities [19–21], use of chiral selectors [22,23],
but all procedures probed to be equally inefficient. Nonetheless, it
would be desirable to have a simple method to access enantio-
merically pure samples of 3 and 4, which would result in possible
new applications for these amino acids.
Based on the reported difference in the rates of aminoacylase-
catalyzed hydrolysis of branched amino acids bearing two chiral
centers [24–26], we decided to evaluate the capability of acylase I
as biocatalyst for stereoselective deacetylation of N-acetyl
derivates of 5 employing enzymes from different sources. Fig. 2
shows the 1H NMR spectra of the products obtained after
treatment of N-acetyl-2-amino-5,5,5-trifluoro-3-methylpentanoic
acid (11) with acylase I from porcine kidney and acylase I from A.
melleus, in comparison to the mixture of diastereoisomers. It can be
easily noticed that acylase I from porcine kidney shows only a
partial substrate preference toward one of the diastereoisomers
while the enzyme from A. melleus deacetylates a single trifluori-
nated stereoisomer.
2. Results and discussion
2.1. Synthesis of 2-amino-5,5,5-trifluoro-3-methylpentanoic acid
Mu¨ller has reported three different procedures for the synthesis
of racemic 2-amino-5,5,5-trifluoro-3-methylpentanoic acid using
(a) 4,4,4-trifluoro-2-butanone (6), (b) 2-methyl-4,4,4-trifluoro-1-
butanol (7) or (c) 5,5,5-trifluoro-3-methylpentanoic acid (8) as
fluorinated synthons [13]. Methodologies a and b suffer from
several disadvantages such as poor yields, impossibility of
synthetic intermediates isolation and subsequent characterization,
employment of highly toxic compounds such as CrO3, etc. The most
efficient option c consists of bromination of 8 to obtain 2-bromo-
5,5,5-trifluoro-3-methylpentanoic acid (9) and successive amina-
tion to generate the desired fluorinated amino acid 5 in a racemic
form. The yield obtained was about 45%, according to the author.
Unfortunately, 8 is not commercially available and must be
synthesized. The procedure applied by Mu¨ller and Tirrell to obtain
this intermediate employs direct electrolysis of methallyl cyanide
and trifluoroacetic acid in presence of sodium acetate [10,13,15].
This procedure is decidedly inconvenient for most preparative

374
H. Biava, N. Budisa / Journal of Fluorine Chemistry 156 (2013) 372–377
Scheme 1. Synthesis of 2-amino-5,5,5-trifluoro-3-methylpentanoic acid. Reagents
and conditions: (a) Ph3PCHCOOEt, Et2O, reflux, 48 h; 51%; (b) H2, 10% Pd/C, MeOH,
overnight r.t.; (c) NaOH 1 M, reflux, 24 h, HCl cc., 80%; (d) SOCl2, 30 8C; (e) Br2, 70 8C,
overnight, then H2O, 72%; and (f) NH3 aq, 40 8C, 5 days, 55%;.
This observation prompted us to examine a sequential use of
these two biocatalysts in order to obtain pure stereoisomers of 5, as
shown in Scheme 2. N-acetyl-2-amino-5,5,5-trifluoro-3-methyl-
pentanoic acid, obtained from 5 following standard procedures,
was firstly treated with acylase I from A. melleus at pH = 7.5 for
24 h. After extracting the mixture with ethyl acetate, ion exchange
chromatography of the aqueous layer provided one stereoisomer
in 29% yield. This compound was assigned to (2S,3S)-2-amino-
Fig. 2. 1H NMR spectra in D2O showing the H signal for: (a) deacetylated samples
a
after treatment of 11 with Porcine kidney acylase I; (b) deacetylated samples after
treatment of 11 with Aspergillus melleus acylase I; (c) compound 5 (mixture of
diastereoisomers).
5,5,5-trifluoro-3-methylpentanoic acid or 5,5,5-trifluoro-
cine (3), according to 1H NMR spectra reported by Mu¨ller and
Tirrell for the 5,5,5-trifluoro- -isoleucine mixture [13] and the
irrefutable preference of acylases for the (2S) configuration [27].
After solvent evaporation from the organic layer, the remaining
residue was subsequently treated with porcine kidney acylase I at
pH = 7.5 for another 24 h. A similar workup provided a second
stereoisomer with 17% yield. This stereoisomer was assigned to
(2S,3R)-2-amino-5,5,5-trifluoro-3-methylpentanoic acid or 5,5,5-
L
-isoleu-
assignments, we employed this enzyme to deacetylate a sample of
racemic N-acetylated 4,4,4-trifluorovaline. This fluorinated amino
acid possesses two chiral centers in position 2 and 3, and the
isolation and characterization of the (2S,3S) and (2S,3R) forms have
been previously reported [20].
D,L
As expected, the presence of a smaller side chain reduces the
enzymatic ability of chiral discrimination at around C . Even so, 1H
b
NMR spectrum obtained after enzymatic deacetylation (Fig. 4)
shows a higher content of (2S,3R)-4,4,4-trifluorovaline, which
evidences A. melleus acylase I predilection in favor of this
stereochemical configuration.
trifluoro-L-alloisoleucine (4) according to the same line of
reasoning as in the case of 3.
The enantiomeric purity for compounds 3 and 4 was assessed
by1H NMR, 19F NMR and derivatization with (2R)-2-methoxy-2-
trifluoromethylphenylacetyl chloride (R-(ꢀ)-Mosher’s acid chlo-
ride, MTPA, see supplementary information) [28,29]. 1H NMR
spectra for compounds 3, 4 and 5 are shown in Fig. 3. All analytic
studies confirm the enantiomeric purity of the samples.
Unfortunately, we were unable to obtain suitable samples of 3
Newman projection through C2–C3 of (2S,3R)-4,4,4-trifluorova-
line and (2S,3S)-5,5,5-trifluoroisoleucine (Fig. 5) shows a similar
arrangement of substituents around C3 for both molecules,
confirming our stereochemical assignments for compounds 3
and 4. Note that the change in the stereochemical designation is a
consequence of Cahn–Ingold–Prelog priority rules and does not
stand for any alteration on the enzymatic specificity toward
substrate configuration.
and
4 for crystallographic characterization, despite several
conditions assayed. In order to access the stereochemical prefer-
ence of A. melleus acylase I and to confirm our stereochemical
Finally, the side products (2R,3R) and (2R,3S)-N-acetyl-2-
amino-5,5,5-trifluoro-3-methylpentanoic acid could be recycled
Scheme 2. Enzymatic resolution of 5,5,5-trifluoro-
L
-isoleucine and 5,5,5-trifluoro-
L
-alloisoleucine. Reagents and conditions: (a) Ac2O, NaOH, THF:H2O (1:1), r.t, 12 h, then HCl
cc., 88% yield; (b) Aspergillus melleus acelyase I, pH = 7.5, r.t., 24 h, 29% yield; (c) Porcine kidney acylase I, pH = 7.5, r.t., 24 h, 17% yield; (d) HCl 3 N, reflux, 12 h; and (e) AcOH,
benzaldehyde, 90 8C, 24 h.

H. Biava, N. Budisa / Journal of Fluorine Chemistry 156 (2013) 372–377
375
Fig. 4. 1H NMR spectrum of the main product after deacetylation of N-acetyl-4,4,4-
trifluorovaline with Aspergillus melleus acylase I. The spectrum shows the regions
Fig. 3. Comparison between 1H NMR spectra in D2O for: (a) compound 3, (b)
compound 4, and (c) compound 5.
corresponding to the H and CH3 groups.
b
Fig. 5. Preferred substrate absolute configuration exhibited by Aspergillus melleus acylase I as determinated by comparison between 5,5,5-trifluoroisoleucine and 4,4,4-
trifluorovaline deacetylation.
by deacetylation and racemization in the usual manner, to deliver
comparable to a bulky alkyl group [30], allows A. melleus acylase I
to fully discriminate between substrate configurations. The active
site of porcine kidney acylase I seems to be more flexible and both
(2S,3S) and (2S,3R) forms could be processed, even though the
reaction rate through the (2S,3R) form is slower.
racemic 5, as judged by 1H NMR and specific rotation.
3. Conclusions
We have applied a straightforward synthesis of 2-amino-5,5,5-
trifluoro-3-methylpentanoic acid and developed a new enzymatic
We believe that easy accessibility to these enantiomerically
pure fluorinated amino acids will open new possibilities in their
application as synthons for developing new therapeutic peptides
and pharmaceutics.
strategy to access enantiomeric pure samples of 5,5,5-trifluoro-
L-
isoleucine and 5,5,5-trifluoro- -alloisoleucine. Resolution was
L
achieved by means of complete discrimination toward (2S,3S)
configuration by A. melleus acylase I, whereas the analogous
enzyme from porcine kidney acylase I shows only a minor
preference toward this configuration. Successive treatment of 5
with both enzymes allowed access to both enantiomeric pure
forms, while the residual side products could be recycled.
The difference in the reactivity of both enzymes seems to be
related to the steric demand of the amino acid side chain. The
introduction of a CF3 moiety, whose size has been described as
4. Experimental
4.1. General
Reagents and chemicals were purchased from Sigma Aldrich,
Alfa Aesar or Fluorochem and used without further purification.
Solvents were used after single distillation. Acylase I from A.
melleus (E.C. 3.5.1.14, powder, >0.5 units/mg) and acylase I from

376
H. Biava, N. Budisa / Journal of Fluorine Chemistry 156 (2013) 372–377
porcine kidney (E.C. 3.5.1.14, grade II, salt-free, lyophilized
powder, 500–1.500 units/mg protein) was purchased from Sigma.
Ion exchange chromatography was performed on Dowex 50WX8
resin (200–400 mesh, Sigma) using distilled water and NH4OH 1 M
as eluants.
Nuclear magnetic resonance spectra were recorded on a Bruker
Avance 400 MHz or Bruker Avance III 500 MHz instrument using
standard deuterated solvents. 19F NMR was recorded using
2.11–2.00 (m, 1H), 1.14 (d, J = 6.4 Hz, 3H); 13C NMR (CDCl3,
125 MHz) 177.56, 126.88 (q, J = 275 Hz), 40.52, 39.50 (q,
J = 27 Hz), 24.95, 20.00. IR (liquid film, cmꢀ1):
2973 (m), 2943
(m), 2889 (w), 1708 (s), 1464 (s), 1416 (s), 1416 (s), 1391 (s), 1382
(s), 1276 (s), 1253 (s), 1214 (s), 1165 (s), 1127 (s), 1075 (s), 1038 (s).
MS-ESI: [m/z]ꢀ = 169.0480. Calculated for [C6H8F3O2]ꢀ = 169.0482.
d
y
4.4. Procedure for the synthesis of 2-bromo-5,5,5-trifluoro-3-
CF3CO2H (
d
= ꢀ76.50) for D2O or C6F6
(
d
= ꢀ164.9) for CDCl3 as
methylpentanoic acid (9)
internal standards. Room temperature specific rotation was
determinated on a Jasco P-2000 polarimeter in a 10 cm quartz
To a purged flask containing compound 8 (2.42 g, 14.3 mmol) at
30 8C, 3.44 g (29.0 mmol) of thionyl chloride was added dropwise.
The mixture was stirred until the gas evolution ceased and then the
flask was connected to a reflux condenser; whose top was attached
to a gas absorption trap. Bromine (3.43 g, 21.5 mmol) was added
slowly, in portions, and the mixture stirred at 70 8C overnight.
When the evolution of HBr stopped, the mixture was allowed to
cool at room temperature and then immersed into an ice bath.
10 ml of H2O was carefully added to the crude acid chloride. After
standing for 10 min, the mixture was extracted with ethyl ether
(3 ꢁ 100 ml), the organic layers combined, dried over Na2SO4 and
concentrated by rotary evaporation. The product was purified by
vacuum distillation (20 mbars, 120 8C) to give a colorless oil in 72%
yield (2.60 g, 10.4 mmol) as a 53:47 mixture of diastereoisomers
(determinated by 1H NMR). 1H NMR (mixture of diastereoisomers,
cell at
l = 589 nm (sodium D line). Analytical thin-layer chroma-
tography was performed using Merck silica gel Kieselgel 60 F254
(0.25 mm) plates and compound were made visible by KMnO4
basic solution, iodine or ninhydrin staining. Infrared spectra were
recorded on a Jasco FT/IR-4100 spectrometer with a resolution of
4 cmꢀ1. HPLC-MS was obtained on a LTQ Orbitrap XL using a
Syncronis C18 column (Thermo scientific), length 50 mm, ID 3 mm,
5
m
m. A 5–100% solvent 2 gradient in 12 min was employed for
elution. Solvent 1: H2O + 0.025% HCO2H. Solvent 2: MeOH + 0.025%
HCO2H. Flow rate: 1.3 ml minꢀ1
Gas chromatography was
.
performed on GC8000 Voyager instrument with a Chiral column
(Lipodex1 E).
4.2. Procedure for the synthesis of E,Z-5,5,5-trifluoro-3-methylpent-
2-enoic acid ethyl ester (10)
CDCl3, 400 MHz, ppm)
d
4.50–4.23 (two doublets, J1 = 6.4 Hz,
J1 = 4.3 Hz, 1H), 2.40–2.68 (m, 2H), 2.15 (m, 1H), 1.23 (m, 3H); 13C
NMR (mixture of diastereoisomers, CDCl3, 125 MHz, ppm)
0
To a round-bottom flask containing a suspension of carbothox-
ymethylenetriphenyl-phosphorane (6.60 g, 18.9 mmol) in anhy-
drous ether (20 ml) were added 4,4,4-trifluoro-2-butanone (6,
2.40 g, 18.9 mmol). The mixture was refluxed for 48 h. After
cooling at room temperature, the suspension was filtered and the
filtrate concentrated in vacuo to eliminate Et2O. The resulting
viscous oil was purified by vacuum distillation (20 mbar, 100 8C) to
afford 1.90 g (9.6 mmol, 51% yield, colorless oil) of a mixture of the
E,Z forms of 10 in a ratio E/Z = 60:40 as judged by NMR assignment.
NMR assignments for the E,Z isomers were made according to
previous reports on related compounds [31,32]. 1H NMR (CDCl3,
d
173.10, 172.80, 125.31 (q, J = 277 Hz), 125.07 (q, J = 278 Hz), 51.75,
51.33, 39.95 (q, J = 28 Hz), 37.34 (q, J = 28 Hz), 31.93 (m), 31.70 (m),
17.95, 16.00; IR (liquid film, cmꢀ1):
(w), 1716 (s), 1463 (s), 1437 (s), 1417 (s), 1394 (s), 1381 (s), 1274
(s), 1253 (s), 1168 (s), 1134 (s), 1032 (s). MS-ESI [m/z]ꢀ = 246.9590.
Calculated for [C6H7BrF3O2]ꢀ = 246.9587.
y 2980 (m), 2945 (m), 2889
4.5. Procedure for the synthesis of 2-amino-5,5,5-trifluoro-3-
methylpentanoic acid (5)
400 MHz, ppm): (E)
J = 10.6 Hz), 2.26 (s, 3H), 1.28 (m, 3H); (Z)
2H), 3.63 (q, 2H, J = 11.1 Hz), 2.0 (s, 3H), 1.28 (m, 3H); 13C NMR
(CDCl3, 101 MHz, ppm): (E) 165.19, 146.22 (m), 125.54 (q,
J = 278 Hz), 122.14, 60.11, 47.19 (q, J = 29 Hz); 19.82, 14.16; (Z)
165.69, 146.49 (m), 125.27 (q, J = 277 Hz), 122.46, 61.00, 44.34 (q,
J = 28 Hz), 25.12, 14.22; IR (liquid film, cmꢀ1):
2983 (s), 2942 (s),
d
5.84 (s, 1H), 4.17 (m, 2H), 2.89 (q, 2H,
Concentrated aqueous ammonia (10 ml) was carefully added to
9 (0.56 g, 2.26 mmol) and the mixture was heated at 40 8C and
stirred in a closed flask for 5 days. After that period, the excess of
ammonia was smoothly evaporated, the residue dissolved in HCl
1 M (final pH = 2–3) and washed with diethyl ether (2 ꢁ 50 ml).
The aqueous layer was lyophilized and purified by ion exchange
chromatography to generate 5 in 55% yield as a white powder
(0.23 g, 1.23 mmol) in a 55:45 (3:4) diastereoisomeric mixture as
determinated from 19F NMR. 1H NMR (mixture of diastereoi-
d
5.92 (s, 1H), 4.17 (m,
d
d
y
1716 (s), 1657 (s), 1447 (s), 1434 (s), 1371 (s), 1352 (s), 1337 (s),
1294 (s), 1273 (s), 1254 (s), 1220 (s), 1137 (s), 1102 (s), 1039 (s),
1016
(s).
GC-MS:
[m/z] = 196.3066.
Calculated
for
somers, D2O, 400 MHz, ppm)
d 3.65 (m, 1H); 2.50–2.24 (m, 2H),
[C8H11F3O2] = 196.0711.
2.13 (m, 1H), 0.95 (m, 3H); 13C NMR (mixture of diastereoisomers,
D2O, 125 MHz, ppm)
d
173.02, 172.44, 137–113 (m, 2ꢁ CF3); 58.42,
4.3. Procedure for the synthesis of 5,5,5-trifluoro-3-methylpentanoic
58.31, 35.99 (m, 2ꢁ CF3CH2-), 28.87, 28.83, 14.62, 14.26.19F NMR
acid (8)
(D2O, decoupled, 470 MHz, ppm)
d
ꢀ63.75 (s, 3F), ꢀ61.86 (s, 3F). IR
(cmꢀ1):
y IR 3117 (w), 3056 (w), 2987 (m), 2957 (m), 2705 (w),
Compound 10 (1.75 g, 8.9 mmol) was dissolved in 2.5 ml MeOH
and transferred into a flask containing 130 mg of 10% Pd/C
previously purged with Ar(g). The mixture were hydrogenated at
1 atm H2(g) pressure overnight. The catalyst was removed by
filtration and the filtrate was added to 50 ml NaOH 1 N. The
reaction was refluxed for 24 h. After cooling, the mixture was
extracted with diethyl ether (25 ml). The solution was acidified to
pH = 1 with concentrated HCl and extracted with ethyl ether
(3 ꢁ 100 ml). The organic layers were combined, the solvent
eliminated by rotary evaporation and the residue distilled under
vacuum (20 mbars, 80 8C) to produce 8 as a colorless oil (1.21 g,
2570 (br), 1594 (s), 1494 (vs), 1466 (w), 1443 (m), 1406 (s), 1383
(s), 1366 (m), 1334 (s), 1318 (s), 1289 (s), 1261 (w), 1249 (s), 1187
(s), 78 (w), 1164 (w), 1142 (s), 1112 (w), 1087 (s), 1052 (s). MS-ESI:
[m/z]+ = 186.0737. Calculated for [C6H11F3NO2]+ 186.0742.
4.6. Procedure for the synthesis of N-acetyl-2-amino-5,5,5-trifluoro-
3-methylpentanoic acid (11)
Amino acid 5 (0.22 g, 1.15 mmol) was dissolved in a 1:1 mixture
THF: H2O (5 ml) and NaOH (0.19 g, 4.6 mmol) were added with
stirring. The mixture was cooled at 0 8C and acetic anhydride
(0.17 ml, 1.73 mmol) was added dropwise. The mixture was stirred
for 12 h and then acidified with concentrated HCl to pH = 2.5. After
7.1 mmol) in 80% yield. 1H NMR (CDCl3, 500 MHz, ppm)
2.46 (m, 1H), 2.31–2.41 (m, 2H), 2.28–2.19 (m, 1H), 2.18 (s, 1H),
d 2.50–

H. Biava, N. Budisa / Journal of Fluorine Chemistry 156 (2013) 372–377
377
extracting with ethyl acetate (3 ꢁ 100 ml), the organic layers were
combined, dried over Na2SO4 and concentrated in vacuo. The
product was obtained in 88% yield as a white solid (0.25 g,
1.1 mmol). 1H NMR (mixture of diastereoisomers, D2O, 400 MHz,
removed by rotary evaporation and lyophilization. The product
obtained was dissolved in AcOH and 0.05 equiv. of benzaldehyde
was added. The mixture was heated at 90 8C with stirring for 24 h.
The solvent was removed under vacuum and the residue dissolved
in HCl 1 N, washed three times with ethyl acetate and water was
eliminated by lyophilization. A white solid (5) was obtained in
quantitative yield and judged to be racemic by 1H NMR analysis.
ppm)
3H), 2.06–2.04 (s, 3H), 1.05 (m, 3H). 13C NMR (mixture of
diastereoisomers, D2O, 101 MHz, ppm) 174.66, 174.38, 174.33,
d
4.58 (d, J1 = 4.2 Hz) 4.41 (d, J10 = 5.1 Hz) (1H), 2.60–2.10 (m,
d
174.12, 127.01 (q, J = 276 Hz), 126.93 (q, J = 276 Hz), 56.92, 55.98,
35.95 (q, J = 28 Hz), 35.81 (q, J = 28 Hz), 29.73 (m), 29.26 (m), 21.60
Acknowledgements
(m), 15.71, 14.21. 19F NMR (CDCl3, decoupled, 470 MHz, ppm):
d
ꢀ63.70 (s, 3F), ꢀ63.71 (s, 3F). IR (cmꢀ1):
y 3320 (b), 3075 (b), 2979
Dr. Biava is deeply thankful 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.
(m), 2945 (m), 2607 (w), 1717 (s), 1622 (s), 1539 (s), 1438 (m),
1252 (s), 1172 (s), 1122 (s), 1039 (s). MS-ESI: [m/z]+ = 228.0839.
Calculated for [C8H13F3NO3]+ 228.0848.
4.7. Procedure for the obtention of (2S,3S)-2-amino-5,5,5-trifluoro-3-
Appendix A. Supplementary data
methylpentanoic acid (5,5,5-trifluoroisoleucine, 3)
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.jfluchem.2013.
07.021.
Compound 11 (0.31 g, 1.36 mmol) was dissolved in 5–10 ml of
LiOH 1 N and the pH of the solution adjusted to pH = 7.5 with AcOH
10%. To this solution acylase I from A. melleus was added (100 mg)
and the solution stirred for 24 h maintaining the pH value between
7.5 and 8.0. After that period, HCl 1 N was added until reaching
pH = 1.5. The solution was heated at 60 8C for 10 min in presence of
activated charcoal. After filtering, the filtrate was extracted three
times with ethyl acetate (3 ꢁ 25 ml). The organic layers were
combined, dried with Na2SO4 and concentrated under vacuum. The
aqueous phase was purified by ion exchange chromatography to
obtain 3 as a white powder in 29% yield (74.3 mg, 0.40 mmol). 1H
References
[1] M. Cametti, B. Crousse, P. Metrangolo, R. Milani, G. Resnati, Chem. Soc. Rev. 41
(2012) 31–42.
[2] J.P. Begue´, D. Bonnet Delpon, Bioorganic and Medicinal Chemistry of Fluorine,
Wiley & Sons, New York, 2008.
[3] E.N. Marsh, Chem. Biol. 7 (2000) R153–R157.
[4] C. Ja¨ckel, B. Koksch, Eur. J. Org. Chem. (2005) 4483–4503.
[5] N.C. Yoder, K. Kumar, Chem. Soc. Rev. 31 (2002) 335–341.
[6] B.C. Buer, E.N. Marsh, Protein Sci. 21 (2012) 453–462.
[7] M. Salwiczek, E.K. Nyakatura, U.I. Gerling, S. Ye, B. Koksch, Chem. Soc. Rev. 41
(2012) 2135–2171.
[8] J.A. Hodges, R.T. Raines, J. Am. Chem. Soc. 125 (2003) 9262–9263.
[9] P. Wang, A. Fichera, K. Kumar, D.A. Tirrell, Angew. Chem. Int. Ed. 43 (2004)
3664–3666.
[10] P. Wang, Y. Tang, D.A. Tirrell, J. Am. Chem. Soc. 125 (2003) 6900–6906.
[11] S. Son, I.C. Tanrikulu, D.A. Tirrell, ChemBioChem 7 (2006) 1251–1257.
[12] Z. Wang, L. Resnick, Tetrahedron 64 (2008) 6440–6443.
[13] N. Mu¨ller, J. Flourine Chem. 36 (1987) 163–170.
[14] W.A. Hooper Huffman, A.W. Ingersoll, J. Am. Chem. Soc. 73 (1951) 3366–3369.
[15] N. Mu¨ller, J. Pharm. Sci. 75 (1986) 987–991.
NMR (D2O, 400 MHz, ppm)
2H), 2.28 (m, 1H), 0.95 (d, J = 6.4 Hz, 3H); 13C NMR (D2O, 101 mHz,
ppm) 172.43, 126.80 (q, J = 276 Hz); 58.34, 35.99 (q, J = 27 Hz),
28.81, 14.61. 19F NMR (D2O, decoupled, 470 MHz, ppm)
ꢀ61.86 (s,
3F). IR (cmꢀ1):
3032 (w), 2918 (m), 2708 (w), 2644 (br), 1733 (s),
d 3.78 (d, J = 3.1 Hz 1H), 2.54–2.41 (m,
d
d
y
1621 (m), 1585 (m), 1526 (s), 1431 (w), 1403 (s), 1391 (s), 1355 (m),
1323(s),1291(s),1261(w), 1255(s),1214(s), 78(w), 1122(w), 1087
(s), 1060 (w), 1043 (s). [a 28
] D = +15.28 (c 0.0915, 1 N HCl). MS-ESI:
[m/z]+ = 186.0729. Calculated for [C6H11F3NO2]+ 186.0742.
[16] C.A. Blum, X. Zheng, I. Martinez, B.L. Chenard, Preparation of haloalkyl-substituted
pyrimidinone derivatives, PCT Int. Appl. Patent WO 2008/024433 A2, published
28 February 2008.
[17] J.P. Greenstein, L. Levintow, C.G. Baker, J. White, J. Biol. Chem. 188 (1951) 647–
663.
4.8. Procedure for the obtention of (2S,3R)-2-amino-5,5,5-trifluoro-3-
methylpentanoic acid (5,5,5-trifluoroalloisoleucine, 4)
[18] T. Yajima, T. Horikawa, N. Takeda, E. Takemura, H. Hattori, Y. Shimazaki, T.
Shiraiwa, Tetrahedron: Asymmetry 19 (2008) 1285–1287.
[19] J.K. Montclare, S. Son, G.A. Clark, K. Kumar, D.A. Tirrell, ChemBioChem 10 (2009)
84–86.
[20] (a) X. Xing, A. Fichera, K. Kumar, J. Org. Chem. 67 (2002) 1722–1725;
(b) X. Xing, A. Fichera, K. Kumar, J. Org. Chem. 67 (2002) 8290.
[21] H. Biava, N. Budisa, Tetrahedron Lett. 54 (2013) 3662–3665.
[22] T. Shiraiwa, A. Ikawa, K. Sakaguchi, H. Korokawa, Bull. Chem. Soc. Jpn. 57 (1984)
2234–2237.
After solvent elimination under vacuum from the organic layer
obtained duringtheresolutionof3, thecrudematerialwastreatedin
a similar procedure but employing 50 mg of acylase I from porcine
kidney for 24 h at pH = 7.5. After purification of the aqueous layer by
ion exchange chromatography, compound 4 was obtained in 17%
yield as a white powder (0.045 g, 0.24 mmol). 1H NMR (D2O,
400 MHz, ppm)
2.36 (m, 1H), 2.34–2.19 (m, 1H), 1.08 (d, J = 7.2 Hz, 3H); 13C NMR
(D2O, 101 mHz, ppm) 173.26, 126.67 (q, J = 276 Hz); 58.39, 35.79
d 3.78 (d, J = 3.5 Hz 1H); 2.66–2.56 (m, 1H), 2.50–
[23] K.H. Kellner, A. Blasch, H. Chmiel, M. La¨mmerhofer, W. Lindner, Chirality 9 (1997)
268–273.
d
[24] S.S. Yuan, A.M. Ajami, J. Labelled Comp. Radiopharm. 22 (1985) 1309–1314.
[25] J.W. Banks, D. O‘Hagan, J. Flourine. Chem. 102 (2000) 235–238.
[26] M. Bakke, H. Ohta, U. Kazmaier, T. Sugai, Synthesis 9 (1999) 1671–1677.
[27] H.K. Chenault, J. Dahmer, G.M. Whitesides, J. Am. Chem. Soc. 111 (1989)
6354–6364.
[28] J. Dale, H. Mosher, J. Am. Chem. Soc. 95 (1973) 512–519.
[29] J.M. Seco, E. Quin˜oa´, R. Riguera, Chem. Rev. 104 (2004) 17–118.
[30] C. Ja¨ckel, M. Salwiczek, B. Koksch, Angew. Chem. Int. Ed. 45 (2006) 4198–4203.
[31] C.D. Poulter, D.M. Satterwhite, H.C. Rilling, J. Am. Chem. Soc. 98 (1976)
3376–3377.
(q, J = 28 Hz), 28.85, 14.20; 19F NMR (D2O, decoupled, 470 MHz,
ppm)
d
ꢀ63.75 (s, 3F). IR (cmꢀ1):
y 3073 (w), 2992 (m), 2969 (w),
2644 (br), 1624 (s), 1573 (s), 1520 (s), 1420 (m), 1397 (s), 1387 (m),
1356 (s), 1320 (s), 1287 (s), 1264 (w), 1167 (s), 1166 (s), 1146 (s),
1087 (s), 1093 (s), 1058 (s). [a 25
]
D = ꢀ91.78 (c 0.0780, 1 N HCl). MS-
ESI: [m/z]+ = 186.0730. Calculated for [C6H11F3NO2]+ 186.0742.
[32] D.L. Dull, I. Baxter, H.S. Mosher, J. Org. Chem. 32 (1967) 1622–1623.
4.9. Deacetylation and racemization of (2R,3R),(2R,3S)-N-acetyl-2-
amino-5,5,5-trifluoro-3-methylpentanoic acid
The organic residue after resolution of both 3 and 4 was
deacetylated by heating at reflux 12 h in HCl 3 N. The solvent was