Kinetic aspects involved in the simultaneous enzymatic synthesis of (S)-3-fluoroalanine and (R)-3-fluorolactic acid

Article abstract of DOI:10.1016/j.jfluchem.2003.08.009

Desaminative oxidation of rac -3-fluoroalanine (rac-1) catalyzed by l-alanine dehydrogenase (l-ALADH) was studied by means of initial rate experiments both in the presence and in the absence of a product. Estimates of kinetic parameters of the reaction were obtained. The reaction mechanism is the sequential ordered BiTer mechanism with inhibition by excess of (S)-3-fluoroalanine. This kinetic study was the basis for the development of a couple enzymatic system for the simultaneous synthesis of (S)-3-fluoroalanine (1a) and (R)-3-fluorolactic acid (3) with l-ALADH and l-lactate dehydrogenase using rac-1 and NAD⁺. Analysis of isolated products revealed 1a in 60% yield and 86% ee and 3 in 80% yield and over 99% ee. Compounds 1a and 3 represent chiral building blocks for the synthesis of several products with pharmacological activity. The presence of the fluorine atom in the substrate causes a better interaction of it in the active site of the enzyme.

Full text of DOI:10.1016/j.jfluchem.2003.08.009

Journal of Fluorine Chemistry 124 (2003) 219–227
Kinetic aspects involved in the simultaneous enzymatic synthesis
of (S)-3-fluoroalanine and (R)-3-fluorolactic acid
Luciana P.B. Gonc¸alves, O.A.C. Antunes1, Gerson F. Pinto, Enrique G. Oestreicher*
´ ´
Instituto de Quımica, UFRJ, CT Bloco A Labs 641 and 536 B, Cidade Universitaria, Rio de Janeiro, RJ 21945-970, Brazil
Received 5 August 2003; received in revised form 5 August 2003; accepted 20 August 2003
Abstract
Desaminative oxidation of rac-3-fluoroalanine (rac-1) catalyzed by L-alanine dehydrogenase (L-ALADH) was studied by means of initial
rate experiments both in the presence and in the absence of a product. Estimates of kinetic parameters of the reaction were obtained. The
reaction mechanism is the sequential ordered BiTer mechanism with inhibition by excess of (S)-3-fluoroalanine. This kinetic study was the
basis for the development of a couple enzymatic system for the simultaneous synthesis of (S)-3-fluoroalanine (1a) and (R)-3-fluorolactic acid
(3) with L-ALADH and L-lactate dehydrogenase using rac-1 and NADþ. Analysis of isolated products revealed 1a in 60% yield and 86% ee
and 3 in 80% yield and over 99% ee. Compounds 1a and 3 represent chiral building blocks for the synthesis of several products with
pharmacological activity. The presence of the fluorine atom in the substrate causes a better interaction of it in the active site of the enzyme.
# 2003 Elsevier B.V. All rights reserved.
Keywords: L-Alanine dehydrogenase; (S)-3-Fluoroalanine; (R)-3-Fluorolactic acid; Enantiospecific desaminative oxidation; Enantioselective reduction
1. Introduction
However, the substrate specificity and the enantioselec-
tivity of these enzymes are very variable.
Oxidoreductions, especially the reduction of carbonyl
compounds, the aminative reduction of a-ketoacids and
the oxidative deamination of rac-a-amino acids catalyzed
by stereoselective NAD(P)þ-dependent dehydrogenases, are
very important organic reactions in asymmetric synthesis
leading to the production of valuable chiral building blocks.
The use of isolated enzymes to catalyze these reactions has
greatly increased in the last years [1–12].
For instance, alcohol dehydrogenases are enzymes that,
although having a broad substrate specificity show variable
enantioselectivity [10–14]. On the other hand, L-lactic dehy-
drogenase from rabbit muscle (L-LDH) and L-alanine dehy-
drogenase (L-ALADH) have proved to be highly
enantioselective, but with a rather narrow range of substrate
specificity, restricted to relative small modifications on the
b-carbon atom of pyruvate [5,15–17].
Fluorinated chiral a-hydroxyacids, such as (R)-3-fluoro-
lactic acid, are considered to be highly versatile chiral
building blocks in asymmetric synthesis for the production
of several compounds of pharmacological interest, such as,
b-adrenergic blocking agents of the aryloxypropanolamines
type, such as, (S)-propranolol and (S)-moprolol and of other
products of pharmaceutical interest such (S)-3-hydroxypyr-
rolidine-2-one which in its open form, (S)-4-amino-2-hydro-
xybutanoic acid, is considered to be one of the most potent
known inhibitors of the neurotransmitter GABA, also show-
ing anticancer activity [5,18].
On the other hand, (S)-3-fluoroalanine has been described
as an antibiotic of wide spectra [19] via irreversible inacti-
vation of bacterial alanine racemase [20], an enzyme
involved in the biosynthesis of the cell wall. This compound
also acts as an inhibitor of serine palmitoyl transferase [21]
Abbreviations: L-ALADH, Bacillus subtilis L-alanine dehydrogenase;
L-LDH, rabbit muscle L-lactate dehydrogenase; A, NADþ; B, (S)-alanine
or (R)-3-fluoroalanine; P, NH4þ; Q, pyruvate or 3-fluoropyruvate; R,
NADH; BSA, bovine serum albumine; E, free form of the enzyme; Kia,
dissociation constant of L-ALADH-NADþ complex; Ka, Michaelis
constant of NADþ; Kb, Michaelis constant of (S)-alanine or (R)-3-
fluoroalanine (3-fluoro-L-alanine); KM, apparent value of the Michaelis
constant of NADþ; Kii, apparent value of the dissociation constant of L-
ALADH-NADþ-inhibitor complex; Kis, apparent value of L-ALADH-
inhibitor complex; v0, initial velocity; V0, maximum velocity; Vapp
,
apparent value of the maximum velocity; ee, enantiomeric excess; HPLC,
high resolution liquid chromatography; IR, infrared spectrometry
* Corresponding author. Tel.: þ55-21-25627366;
fax: þ55-21-2562-7266.
E-mail addresses: octavio@iq.ufrj.br (O.A.C. Antunes),
enrique@iq.ufrj.br (E.G. Oestreicher).
1 Co-corresponding author. Tel.: þ55-21-2562-7248;
fax: þ55-21-2562-7559.
0022-1139/$ – see front matter # 2003 Elsevier B.V. All rights reserved.
doi:10.1016/j.jfluchem.2003.08.009

220
L.P.B. Gonc¸alves et al. / Journal of Fluorine Chemistry 124 (2003) 219–227
and in addition, it is a potential precursor of fluoroamine
compounds [22,23].
muscle (L-(þ)-lactate:NADþ oxidoreductase, EC 1.1.1.27),
solution in 50% glycerol containing 10 mM potassium
phosphate buffer, pH 7.5, L-glutamate dehydrogenase (L-
GLUDH) from bovine liver type III (L-glutamate:NAD(P)þ
oxidoreductase, EC 1.4.1.3), lyophilized, NADþ (grade
IIIC), NADH (grade III), bovine serum albumine (fraction
V), sodium pyruvate, sodium 3-fluoropyruvate, rac-alanine
and Dowex 50W-X8 cation exchange resin were obtained
from Sigma. a-Ketoglutarate, NaBH4 and the Diazald1 kit
(for CH2N2 generation) were purchased from Aldrich Che-
mical. All other chemicals were of analytical grade and
obtained from Merck, Darmstad.
The utilization of enzymes as catalyst for the production
of chiral building blocks is widely reported in the literature
[7,8]. However, the use of dehydrogenases in preparative
applications requires effective processes for coenzyme
regeneration since these cofactors are too expensive to be
used as stoichiometric reagents. Several of these processes
have already been reported by our group [4–6].
In addition, the use of fluorinated blocks in the synthesis
of pharmaceuticals has been receiving considerable atten-
tion in the past years [24].
(R)-3-Fluorolactic acid was obtained enzymatically in our
laboratory in 80% overall yield and ee > 99% in a system
involving the enantioselective reduction of 3-fluoropyruvate
catalyzed by L-LDH and NADH which was in situ regen-
erated by the oxidation of cis-1,2-bis(hydroxymethyl)cyclo-
hexane catalyzed by HLADH [5,6]. (R)-3-Fluoroalanine was
obtained by Oshima et al. [12] with L-ALADH from Bacillus
sphaericus by reductive amination of 3-fluoropyruvate in the
presence of NADH. The reduced form of the coenzyme was
in situ regenerated by the oxidation of ammonium formate
catalyzed by yeast formate dehydrogenase. No enantiose-
lective process for the production of (S)-3-fluoroalanine has
been described before our preliminary communication [4].
The only process described in the literature for the synthesis
of (S)-3-fluoroalanine is the photofluorination of (R)-alanine
with CF3-OF with a conversion of 50% [19].
An enzymatic system able to produce simultaneously (S)-
3-fluoroalanine via kinetic resolution of rac-3-fluoroalanine
by deaminative oxidation catalyzed by L-ALADH and (R)-3-
fluorolactic acid by enantioselective reduction of 3-fluoropyr-
uvate generated in the former reaction catalyzed by L-LDH, is
reported in the present article. In order to improve the
performance of this coupled enzymatic system since the
deaminative oxidation reaction, the rate limiting step of the
whole process, is thermodynamically unfavorable and strong
inhibition phenomena by excess of substrate and/or by the
reaction products certainly are present, a kinetic study of this
reaction was undertaken. This study involved initial velocity
measurements in the absence of products and the analysis of
the inhibition by the products of the reaction. From this study
reaction conditions for coupled enzymatic system were found
that allowed the simultaneous production of (S)-fluoroalanine
in 60% yield and 86% ee and (R)-3-fluorolactic acid in 80%
yield and over 99% ee. The presence of fluorine allowed a
stronger interaction between the enzyme and the substrate.
2.2. Assay of L-ALADH activity and estimation of
kinetic parameters
L-ALADH activity (assayed in the direction of rac-3-
alanine deaminative oxidation) was measured by following
NADH absorption at 340 nm at pH 7.9 and 25 8C. All assays
were carried out in quartz cuvettes with a 1 cm light path
using a Beckman DU 70 spectrophotometer equipped with a
dot-matrix printer. The temperature of the cell holder was
kept at 25 8C by forced circulating water.
Reaction mixtures contained in a total volume of
1 ml:50 mM sodium phosphate buffer, pH 7.9 with varying
concentrations of NADþ and rac-3-fluoroalanine and for
product inhibition studies, contained also four different fixed
concentrations of 3-fluoropyruvate or NH4Cl, that are indi-
cated in the legends of Figs. 2 and 3, respectively. NADþ, rac-
3-fluoroalanineandproductinhibitorssolutionsweremadeup
immediately before use in the same buffer and maintained in
an ice bath. After addition of the substrates or of the substrates
and one inhibitor, the reaction mixtures were incubated for
5 min in the spectrophotometer cell holder before the addition
of the enzyme solution. Reaction was started by adding 20 ml
of an enzyme solution containing in a volume of
1.0 ml:130 mg (protein basis) corresponding to 0.45 IU of a
desalinized L-ALADH enzyme preparation (obtained by
ultra-filtration with Centricon-10 concentrators (AMICON),
through five consecutive centrifugations at 5 8C for 1 h and
2000 Â g and reconstitution of the original volume with
sodium phosphate buffer 10 mM, pH 7.7 and 50% (v/v) of
glycerol. The slopes of the recording lines were kept close to
458 by varying the absorbance full scale and/or the time full
scale of the spectrophotometer. The printer curves obtained
were extrapolated to the time of enzyme addition and the
tangents of the curves at this time were taken as initial
velocities. The initial velocity of the reaction was calculated
in terms of mM of NADH produced per minute, using a molar
2. Materials and methods
absorption coefficient for NADH of 6220 MÀ1 cmÀ1
.
2.1. Materials
2.3. Kinetic data processing
Bacillus subtilis L-alanine dehydrogenase (L-alanine:
NADþ oxidoreductase, EC 1.4.1.1), suspension in 2.4 M
(NH4)2SO4, pH 7.0, L-lactate dehydrogenase from rabbit
The values (symbols) appearing in the figures are the
mean experimentally determined initial velocities (tripli-
cates), and were used in determining the kinetic parameters.

L.P.B. Gonc¸alves et al. / Journal of Fluorine Chemistry 124 (2003) 219–227
221
Estimates of parameters and of their asymptotic standard
errors were obtained by fitting the appropriate rate equations
to data using a nonlinear least-squares computer program,
developed in our laboratory and specifically devised for
steady-state studies of enzymes kinetics [25]. Before the
data for a whole experiment were analyzed with the aid of
the computer program, the linearity of the individual lines
was checked graphically by plotting the reciprocal velocities
against the reciprocal of the varied substrate concentration at
various concentrations of the fixed changing ligand. The
same procedure was extended for the effect of the fixed
changing ligand on intercepts and/or slopes of the Line-
Weaver–Burk plots obtained. Based on these preliminary
plots, the proper pattern was selected for each experiment
and the appropriate equations (see below) were fitted to the
mean data points.
trically at 340 nm until total consumption of the substrate.
The concentration of NH4þ was calculated from the differ-
ence of NADH absorption (before and after) the addition of
L-GLUDH by using the molar absorption coefficient of
NADH. This method gave results that are in agreement with
those obtained by chiral HPLC in Section 2.6.
2.5. Isolation of target products
Isolation and characterization of products formed was
accomplished by thermal treatment of the reaction mixture
(100 8C for 15 min) in order to inactivate the enzymes. After
centrifugation to remove the coagulated proteins, the super-
natant fluid was acidified with 3 M HCl until pH 2.0 and
submitted to cation exchange column chromatography on
Dowex 50W-X8 (8 cm  2:5 cm). The column was initially
eluted with glass bidistilled water in order to recover 3-
fluorolactic acid and then with 1 M NH4OH to isolate 3-
fluoroalanine. The fractions containing the former product
were acidified until pH 1.7 with 1 M HCl, saturated with
NaCl and continuously extracted with Et2O for 72 h to
eliminate the coenzyme. The organic phase, containing 3-
fluorolactic acid, was methylated with CH2N2/Et2O accord-
ing to Blank et al. [27]. The resulting product, 3-fluorolactic
acid methyl ester was obtained with an overall yield of 80%.
The ninhydrin positive fractions that were eluted with
NH4OH containing 3-fluoroalanine wereþconcentrated in
rotatory evaporator at 40 8C until all NH4 was eliminated
and vacuum dried in Speed Vac (Savant). The resulting
product was re-crystallized from iso-propanol/H2O as fol-
lows. The dried material was re-suspended in glass bi-
distilled water until complete dissolution, 30 mg of active
charcoal were added and the suspension was maintained
under continuous stirring for 15 min. The sample was
filtered and the operation repeated until a colorless solution
was obtained that was heated at 60 8C with the dropwise
addition of iso-propanol/H2O mixture (1:3, v/v) at 60 8C.
The sample was kept at room temperature for 30 min and
then for 3 h in an ice bath. After initiation of the crystal-
lization process, drops of iso-propanol were added and the
crystals formed were collected by filtration under vacuum
and were washed successively with iso-propanol/H2O 90%
(v/v), iso-propanol and hexane. Isolated yield was 60%.
It will be assumed now, and justified in Section 3 that the
reaction mechanism is the sequential ordered BiTer [26].
Then the initial velocity equation in the absence of products
is:
V0AB
v0 ¼
(1)
KiaKb þ KbA þ KaB þ AB
For the inhibition studies, the rate equations used were:-
linear noncompetitive inhibition,
Vapp
KMð1 þ I=KisÞ þ Sð1 þ I=KiiÞ
linear uncompetitive inhibition,
Vapp
KM þ Sð1 þ I=KiiÞ
S
v0 ¼
(2)
(3)
S
v0 ¼
2.4. Laboratory preparative enzymatic production of (S)-3-
fluoroalanine and (R)-3-fluorlactate
The reaction was performed in a jacketed batch reactor of
120 ml, continuously stirred. Reaction medium contained in
a total volume of 50 ml:0.05 M sodium phosphate buffer, pH
7.9, 20 mM rac-3-fluoroalanine, 373 IU of L-ALADH,
4250 IU of L-LDH and 0.05% (w/v) BSA. The reaction
was started by adding 0.25 mM NADþ. The reaction mix-
ture was maintained under continuous magnetic stirring and
the temperature was kept at 25 8C by circulating water
through the jacket of the reactor with a thermocirculating
bath.
2.6. Characterization of (R)-3-fluorolactic acid methyl
ester and (S)-3-fluroalanine
The extent of reaction was followed by removing at
different times aliquots of 10–50 ml of the reaction mixture
that were immediately diluted with 0.5 ml of 0.05 M sodium
phosphate buffer, pH 7.9 and heated at 100 8C for 5 min iþn
order to inactivate the enzymes. The concentration of NH4
present in these samples was determined by ‘‘end-point’’
assays in the presence of 0.2 mM NADH, 0.5 mM of
a-ketoglutarate and 1 mg of L-GLUDH in a total volume
of 2.0 ml at 25 8C. The decrease of NADH absorption after
the addition of the enzyme was followed spectrophotome-
Both target products were analyzed and characterized by:
IR spectrometry with a Nicolet Magna-IR 760 spectrophot-
ometer; 1H NMR spectroscopy with a 200 MHz Varian
(Gemini) equipment; polarimetry, carried out with a Jasco
DIP-370 digital polarimeter; Chiral HPLC performed by
using an ISCO 2350 chromatograph equipped with a
Nucleosil Chiral-11 column (Macherey-Nagel, ET 250/8/
4) by using 1 mM CuSO4, pH 4.6 as eluent at a flow rate of
1 ml/min, 23 8C and UV detection at 235 nm.

222
L.P.B. Gonc¸alves et al. / Journal of Fluorine Chemistry 124 (2003) 219–227
2.7. Determination of enantiomeric excess
pounds of pharmacological interest [4] that in addition
presents a strong antibiotic activity [19].
The enantiomeric excesses of both target products were
determined by chiral HPLC as described in Section 2.6, rac-
3-fluorolactic acid methyl ester and rac-3-fluoroalanine
were used as standards.
Rac-3-fluorolactic acid methyl ester was prepared as
previously described by us in [5]. Rac-3-fluoroalanine
was obtained by reductive amination of 3-fluoropyruvate
(3.49 mmol) with NaBH4 (5.5 eq.) in the presence of
NH4OH (68 mmol) as described by Dolling et al. [28].
The product was isolated, purified, crystallized and identi-
fied as described in Sections 2.5 and 2.6.
Fig. 1 shows the fitting of Eq. (1) to the initial velocity
data obtained in the absence of products. In Fig. 1A rac-3-
fluoroalanine was the variable substrate at seven fixed con-
centrations of NADþ. Since an intersecting pattern of
straight lines was obtained, this was the first indication that
the kinetic mechanism is most probably of the sequential
type involving ternary or central complexes, not necessarily
kinetically important. This experiment also revealed that
high concentrations of the varied substrate have inhibitory
activities. In order to confirm these experimental observa-
tions, the concentration of NADþ was varied keeping the
concentration of 3-fluoroalanine fixed at various levels. As
shown in Fig. 1B the same pattern of intersecting lines was
obtained, however, when concentrations of 3-fluoroalanine
higher than 4.0 mM were used, parallels lines indicative of
an uncompetitive inhibition by excess of 3-fluoroalanine,
were obtained (Fig. 1B).
2.8. Determination of protein concentration
Protein concentration was determined by using a method
suitable for the detection of low protein contents [29].
Bovine serum albumin was used as standard.
If the experimental points showing inhibition (Fig. 1A and
B) by excess of substrate are eliminated and Eq. (1) is fitted
to the remaining data, estimates of the parameters of this rate
equation are obtained. These estimates together with the
corresponding values determined by Grimshaw and Cleland
[30] for the oxidative deamination of L-alanine are displayed
in Table 1. On comparing the values of the kinetic para-
meters obtained for the oxidative deamination of 3-fluor-
oalanine with those for L-alanine determined by Grimshaw
and Cleland [30] (Table 1), it can be seen that the values
estimates for Kia are not significantly different and thus, Kia
does not depends on the nature of the other substrate (3-
fluoroalanine and L-alanine, respectively), suggesting that,
in fact, this parameter is the dissociation constant of the L-
ALADH-NADþ complex, and that for the oxidative deami-
nation reaction of 3-fluoroalanine the order of substrate
combination with the enzyme is the same as for the corre-
sponding reaction of L-alanine catalyzed by L-ALADH [30].
On the other hand, markedly in the case of Kb, the value
obtained for 3-fluoroalanine was significantly lower than the
3. Results and discussion
Grimshaw and Cleland [30] determined that the kinetic
mechanism for the oxidative deamination reaction of L-
alanine catalyzed by L-ALADH from B. subtilis in the
presence of NADþ was the sequential ordered BiTer kinetic
model, with NADþ binding to the free form of the enzyme
followed by the binding of L-alanine. The first product to be
released was NH4þ, followed by pyruvate and NADH was
the last one (Scheme 1).
Considering that L-ALADH from B. subtilis shows an
absolute stereoespecificity for the L-enantiomer of alanine
[30] and that compounds harboring an halogen substituent
on the b-atom carbon of alanine have not been before used to
determine the kinetic mechanism of the oxidative deamina-
tion reaction of this enzyme, we decided to investigate the
kinetic mechanism of the oxidative deamination reaction of
rac-3-fluoroalanine catalyzed by L-ALADH. Also, the
kinetic resolution of rac-3-fluoroalanine by oxidative dea-
mination of the R-enantiomer of this compound will produce
(S)-3-fluoroalanine, an important chiral building block in
asymmetric synthesis for the production of several com-
Table 1
Estimates of kinetic parameters derived from initial velocity studies on the
oxidative deamination of (R)-3-fluoroalanine catalyzed by L-ALADH
Parameter
Value Æ S:E:a
Value Æ S:E:b
95% Confidence
limitsc
d
V0 (mM/min)
H C
3
CO H
CO H
2
H C
3
2
–
0.032 Æ 0.002
0.394 Æ 0.036
2.499 Æ 0.233
0.266 Æ 0.350
0.029–0.036
0.321–0.468
2.023–2.975
0.194–0.338
Kia (mM)
Kb (mM)
Ka (mM)
0.33 Æ 0.07
4.40 Æ 0.70
0.22 Æ 0.02
+
NH
4
(P)
O
NADH
(R)
+
NH
NAD
(A)
2
(B)
(Q)
a Data from [26] for the oxidative deamination of (S)-alanine (L-
alanine).
b Asymptotic standard error of parameter. Estimates of parameters
obtained by nonlinear regression analysis of experimental data depicted in
Fig. 1.
E
EA
EQR
ER
E
EAB
(EPQR)
oxidative deamination reaction of (S)-alanine catalyzed by L-ALADH.
c Calculated according to Student’s t-test adapted for nonlinear
Scheme 1. Sequential ordered steady-state kinetic mechanism for the
regression as described by Metzler [33].
d Value not supplied in [29].

L.P.B. Gonc¸alves et al. / Journal of Fluorine Chemistry 124 (2003) 219–227
223
(A)
400
300
200
100
0
-0,4
0
0,4
0,8
1,2
1/fluoro-L-alanine (mM)-1
(B)
400
300
200
100
0
-3,5
0
3,5
7
1/NAD+ (mM)-1
Fig. 1. Effect of 3-fluoroalanine and NADþ on the initial velocity of the reaction of oxidative deamination of (R)-3-fluoroalanine catalyzed by L-ALADH.
(A) Rac-3-fluoroalanine was the variable substrate (its concentration is expressed as (R)-3-fluoroalanine considering 95% purity of the synthesized racemate).
The concentrations of NADþ were: (^) 0.15 mM; (&) 0.2 mM; (~) 0.25 mM; (*) 0.3 mM; ( ) 0.4 mM; (*) 0.6 mM, and (þ) 1.0 mM. (B) NADþ was the
variable substrate. The concentrations of (R)-3-fluoroalanine (considering 95% purity of the synthesized racemate) were: (&) 0.9 mM; (&) 1.35 mM; (~)
1.8 mM; (*) 2.25 mM; (~) 3.15 mM; (*) 4.5 mM; and (*) 9.0 mM. Other experimental condition are given in Sections 2.1 and 2.2.
value for the limiting Michaelis constant of the natural
substrate of the enzyme (P > 0:05 calculated from the
Student’s t-test). This result may indicate that the presence
of the fluorine atom in the substrate causes a better inter-
action of it in the active site of the enzyme.
opyruvate produced a linear uncompetitive inhibition pat-
tern (Fig. 3). Estimates of parameters of Eqs. (2) and (3) and
of their asymptotic standard errors are displayed in Table 2.
The fact that NH4þ behaves as a noncompetitive inhibitor in
relation to NADþ (Fig. 2), that 3-fluoropyruvate is an
uncompetitive inhibitor (Fig. 3) and that NADH is a com-
petitive inhibitor in relation to NADþ (results not show),
strongly suggests that the desaminative oxidation of 3-
fluoro-D-alanine catalyzed by B. subtilis L-ALADH in the
presence of NADþ follows a sequential ordered BiTer
kinetic mechanism [26], with NADþ being the first sub-
strate to bind to the enzyme followed by the binding of 3-
fluoro-D-alanine. Moreover, since NADH produced linear
In order to confirm the substrates binding sequence with
the enzyme as well as to obtain some information concern-
ing the order of release of products from the central com-
plexes, a study of product inhibition was carried out. The
results of these experiments by using NADþ as the varied
substrate are illustrated in Figs. 2 and 3. When NH4þ was
assayed as inhibitor a linear noncompetitive inhibition
pattern was obtained (Fig. 2) on the other hand, 3-fluor-

224
L.P.B. Gonc¸alves et al. / Journal of Fluorine Chemistry 124 (2003) 219–227
400
300
200
100
0
1/NAD (m M )-1
-3,5
0
3,5
7
Fig. 2. Inhibition of the reaction of oxidative deamination of (R)-3-fluoroalanine catalyzed by L-ALADH by NH4þ. NADþ was the variable substrate and
þ
(R)-3-fluoroalanine concentration was kept constant at 2.25 mM. Concentrations of NH4 were: (^) zero; (&) 2.0 mM; (~) 3.5 mM; and (*) 5.0 mM.
Other experimental conditions are given in Sections 2.1 and 2.2.
competitive inhibition in relation to NADþ, the reduced
coenzyme must be the last product to be release in thþe
reaction. The mixed type inhibition produced by NH4
(Fig. 2) and the linear uncompetitive inhibition found for
3-fluoropyruvate (Fig. 3), strongly suggest that NH4þ is the
first product to be release in the reaction, followed by the
release of 3-fluoropyruvate. Considering this order of pro-
ducts release, the uncompetitive inhibition by rac-3-fluor-
oalanine in relation to NADþ (Fig. 1B) can be attributed to
the formation of a L-ALADH-NADH-3-fluoro-L-alanine
‘‘dead-end’’ complex. The kinetic mechanism determined
with this series of experiments for the deaminative oxida-
tion of rac-3-fluoroalnine catalyzed by L-ALADH is shown
in Scheme 2.
Grinshaw and Cleland [30] described that D-alanine
behaved as a linear competitive inhibitor with respect to
L-alanine. This inhibition implies in the competition of both
enantiomers of alanine for the L-ALADH-NADþ complex
300
200
100
0
1/NAD (mM )-1
0
3
,5
7
Fig. 3. Inhibition of the reaction of oxidative deamination catalyzed by L-ALADH by 3-fluoropyruvate. NADþ was the variable substrate and (R)-3-
fluoroalanine concentration was kept at 2.25 mM. Concentrations of 3-fluoropyruvate were: (^) zero; (&) 0.02 mM; (~) 0.04 mM; and (*) 0.06 mM.
Other experimental conditions are given in Sections 2.2 and 2.3.

L.P.B. Gonc¸alves et al. / Journal of Fluorine Chemistry 124 (2003) 219–227
225
Table 2
Estimates of kinetic parameters derived from product inhibition studies on
CO H
2
the oxidative deamination of (R)-3-fluoroalanine catalyzed by L-ALADH
F
Parameter
Estimate Æ S:E:a (product inhibitor)
NH
2
+
NH
4
H
2
O
3-Fluoropyruvatec
1a
þb
CO H
2
NH4
F
Vapp (mM/min)
KM (mM)
Kii (mM)
0.017 Æ 0.002
0.016 Æ 0.002
0.364 Æ 0.011
0.063 Æ 0.022
–
CO H
2
CO H
2
NH
F
2
0.338 Æ 0.013
5.750 Æ 0.355
10.826 Æ 0.458
F
L-ALADH
rac-1
2
O
NH
2
Kis (mM)
+
NAD
NADH
1b
a Asymptotic standard error of parameter.
b Estimates of parameters obtained by fitting Eq. (2) to experimental
CO H
2
data depicted in Fig. 2 by nonlinear regression analysis.
c Estimates of parameters obtained by fitting Eq. (3) to experimental
data depicted in Fig. 3 by nonlinear regression analysis.
F
L-LDH
OH
3
(Scheme 1). Since the experiments described in this work
were performed with racemic 3-fluoralanine, the interaction
of 3-fluoro-D-alanine with the above mentioned complex can
be discarded, otherwise, instead of the linear uncompetitive
inhibition by excess of rac-3-fluoroalanine observed
(Fig. 1B), a linear mixed type inhibition should be evident.
This latter type of inhibition would arise by formation of
both L-ALADH-NADþ-3-fluoro-D-alanine and L-ALADH-
NADH-3-fluoro-L-alanine dead-end complexes. The absence
of the former complex is indicative of a much stronger
interaction of 3-fluor-L-alanine with the L-ALADH-NADþ
than L-alanine as suggested by the value of Kb found for the
fluorinated substrate (Table 1).
As stated earlier in the present work, (R)-3-fluorolactic
acid methyl ester and (S)-3-fluoroalanine can be considered
as valuable chiral building blocks for the synthesis of several
enantiomeric pure products with therapeutically activity.
(R)-3-Fluorolactic acid methyl ester has been previously
obtained in our laboratory [5] in 80% overall yield and
ee > 99% by enantiospecific reduction of 3-fluoropyruvate
catalyzed by rabbit muscle L-lactate dehydrogenase. NADH
needed in this reaction was used in catalytic concentration
by utilizing a NADH in situ regeneration system consisting
of the oxidation of cis-1,2-bis(hydroxymethyl)cyclohexane
to the chiral lactone (þ)-(1R,6S)-cis-8-oxabicyclo[4.3.0]
nonan-7-one catalyzed by horse liver alcohol dehydrogenase
Scheme 3. Coupled enzymatic redox system for the simultaneous
production of (S)-3-fluoroalanine (1a) and (R)-3-fluorolactic acid (3)
using rac-3-fluoroalanine (rac-1) as substrate. (R)-3-Fluoroalanine (1b); 3-
fluoropyruvic acid (2).
[5]. (S)-3-Fluoroalanine, on the other hand, has never been
before obtained enzymatically. In 1970, Kollonitsch et al.
[19] obtained this compound by photofluorination of
D-alanine with fluoroxytrifluoromethane with yield of 50%.
Since the kinetic properties of L-ALADH acting on rac-3-
fluoroalanine as well as those of L-LDH catalyzing the
enantioselective reduction of 3-fluoropyruvate [5,6] were
studied by our group, we envisaged an enzymatic system
suitable for the simultaneous production of (S)-3-fluoroala-
nine (1a) and (R)-3-fluorolactic acid (3). As shown in
Scheme 3 this enzymatic system involves the deaminative
oxidation of the R-enantiomer (1b) of rac-1 to produce
NH4þ, 3-fluoropyruvate (2) and NADH. The latter two
products of the main reaction are used by L-LDH to produce
(R)-3-fluorolactic acid (3) and NADþ, then if the velocity of
this reaction is not rate limiting of the whole system, just
NH4þ and (R)-3-fluoralanine will accumulate in the reaction
medium and NADþ will be continuously recycled until all
(1b) has been consumed. Moreover, since NH4þ is a rather
weak inhibitor of L-ALADH (Table 2) and the concentra-
tions of NADH and of (2) will be almost zero in this
condition, the extent of conversion must be high and (1a)
must be obtained with a high ee and a yield of approximately
50% (total conversion of 1b). This coupled redox system is
then able to produce simultaneously two valuable chiral
blocks, (1a) by kinetic resolution of rac-3-fluoroalanine
catalyzed by L-ALADH and (3) by enantioselective reduc-
tion of (2) catalyzed by L-LDH.
CO H
CO H
2
2
F
F
+
NH
4
(P)
NADH
(R)
+
NH
O
NAD
(A)
2
(B)
(Q)
The time–course curve for the deaminative oxidation of
(1b) catalyzed by L-ALADH with the concomitant enantio-
selective reduction of (2) catalyzed by L-LDH that also acts
as NADþ in situ regeneration system in a semi-preparative
scale of approximately 150 mg (20 mM rac-3-fluoroalanine)
is shown in Fig. 4. The progress of the reaction indicates that
after 24 h of reaction, under the experimental conditions
described in Section 2.4, 93% of (1b) were consumed,
E
E
EQR
ER
E
EAB
( )
EPQR
CO H
2
F
NH
ERB
2
Scheme 2. Sequential ordered BiTer kinetic mechanism for the oxidative
deamination reaction of (R)-3-fluoroalanine catalyzed by L-ALADH.

226
L.P.B. Gonc¸alves et al. / Journal of Fluorine Chemistry 124 (2003) 219–227
100
80
60
40
20
0
0
10
20
30
40
50
Time (h)
Fig. 4. Time–course curve of the coupled enzymatic system for the
simultaneous production of (S)-3-fluoroalanine and (R)-3-fluorolactic acid
in a laboratory preparative scale (150 mg). Tþhe progress of reaction was
followed by enzymatic determination of NH4 concentration at the times
indicated in the figure as described in Section 2.4 (*) or by chiral HPLC
determination of (S)-3-fluoroalanine as described in Section 2.6 (&).
suggesting that at the end of the reaction, theoretically
(considering that L-ALADH is absolutely enantiospecific
for (1b) as it has been shown for the natural substrate by
Grimshaw and Cleland [30]), the reaction medium must
contain 9.3 mM of (1a) with an ee of 86% and the same
concentration of enantiopure (3), respectively.
Fig. 5. Chiral HPLC analysis of 3-fluoroalanine. (A) Rac-3-fluoroalanine
obtained by aminative reduction of 3-fluoropyruvate as described in
Section 2.7. (B) Product obtained enzymatically as described in Section
2.4. The numbers appearing in the figure represent the retention times
(minutes) of the peaks. Other experimental details are given in Section 2.6.
At the end of the reaction, after thermal inactivation of
enzymes, both target products (1a) and (3) were isolated and
purified as described in Section 2.5. Compound 1a was
Target product (3) that remained stable in the reaction
mixture was recovered in 80% yield and was analyzed
before methylation by chiral HPLC as described for (1a).
Just one peak with retention time of 2.84 min was obtained
(results not shown). From these data an ee higher than 99%
was determined for (3). This product was methylated as
described previously [5] and analyzed by polarimetry, IR,
elemental analysis (C, H, N) and 1H NMR. The methyl ester
25
isolated in 60% yield, ½aꢀD ¼ À8:7 (c ¼ 0:8, 1 M HCl).
IR revealed the following bands of axial deformation:
3428 cmÀ1 (NH3þ); 3061, 2987 and 2923 cmÀ1 (aliphatic
C–H); 1615 cmÀ1 (C¼O of carboxylate); 1356 cmÀ1 (C–N);
1022 cmÀ1 (C–F). 1H NMR showed the following chemical
shifts: d 4.02 (ddd) JH;F ¼ 29:5 Hz, JH;H ¼ 4:5 and 3.1 Hz; d
4.81 (ddd) JH;F ¼ 47:4, JH;H ¼ 4:5 and 3.0 Hz.
25
Fig. 5 shows the analysis of 3-fluoroalanine by chiral
HPLC. Fig 5A shows that when rac-3-fluoroalanine was
analyzed two peaks of the same area were obtained with
retention times of 4.71 and 5.73 min, respectively. When the
product obtained enzymatically was analyzed (Fig. 5B), the
same two peaks were found. The one eliciting a retention
time of 4.71 min representing 7% of the relative area
corresponded to (R)-3-fluoroalanine that was not consumed
by L-ALADH. The other one, with a retention time of
5.73 min, and relative area of 93%, corresponded to (S)-
3-fluoroalanine. From these data an ee of 86% was deter-
mined for (S)-3-fluoroalanine. This ee determined experi-
mentally is very similar to that calculated with the data of
Fig. 4 thus confirming the absolute stereospecificity of L-
ALADH for (R)-3-fluoroalanine.
of (3) showed ½aꢀD ¼ À4 (c ¼ 1, EtOH) and elemental
analysis, IR and 1H NMR data that are fully consistent with
our previously reported results [5]. The absolute configura-
tion of (3) was considered to be (R) based on the fact that
L-LDH presents an absolute enantioselectivity and it is
considered to be a reagent to determine the absolute con-
figuration of the a-C atom of compounds that are substrates
of this enzyme [31,32].
4. Concluding remarks
This paper describes an efficient enzymatic process to
produce simultaneously in the same reaction medium two
important chiral blocks, enantiopure (R)-3-fluorolactic acid

L.P.B. Gonc¸alves et al. / Journal of Fluorine Chemistry 124 (2003) 219–227
227
[9] M.J. Kim, G.M. Whitesides, J. Am. Chem. Soc. 110 (1988) 2959–
2964.
methyl ester a 1,2,3-trisubstituted three-carbon compound
with several applications in asymmetric organic synthesis
and enantioenriched (S)-3-fluoroalanine. This latter com-
pound as stated earlier, in addition to synthetic applications
shows an interesting activity as wide spectra antibiotic.
Enzymatic systems as the one reported in the present paper
can be easily set up, allowing the preparation of numerous
optically active products with biological or pharmacological
activity. It has been observed that the presence of the fluorine
atom in the substrate causes a better interaction of it in the
active site of the enzyme.
[10] C.W. Bradshaw, W. Hummel, C.H. Wong, J. Org. Chem. 57 (1992)
1532–1536.
[11] W. Hummel, M.R. Kula, Eur. J. Biochem. 184 (1989) 1–13.
[12] T. Ohshima, C. Wandey, D. Conrad, Biotechnol. Bioeng. 34 (1989)
394–397.
[13] E. Keinan, K.K. Seth, R. Lamed, Ann. NYAcad. Sci. 50 (1987) 130–
149.
[14] C.V. Bradshaw, H. Fu, C.I. Shen, C.-H. Wong, J. Org. Chem. 57
(1992) 1526–1532.
[15] B.L. Hirchbein, G.M. Whitesides, J. Am. Chem. Soc. 104 (1982)
4458–4460.
[16] A.J. Pratt, Chemistry in Britain, March 1989, pp. 282–286.
[17] H.U. Bergmeyer, M. Grabl, H.-H. Walter, in: J. Bergmeyer, M. Grabt
(Eds.), Methods of Enzymatic Analysis, vol. II, third ed., VCH,
Weinheim, 1988, pp. 138–139.
Acknowledgements
[18] J.M. Bentley, H.J. Wadsworth, C.L. Willis, J. Chem. Soc. Chem.
Commun. (1995) 231–232.
Financial support from CNPq, FUJB/UFRJ, CAPES,
FAPERJ, FINEP and PRONEX is acknowledged. We also
´
thank the ‘‘Nucleo de Pesquisas de Produtos Naturais-
UFRJ’’ for running NMR analysis.
[19] J. Kollonitsch, L. Barash, F.M. Kahan, H. Kropp, Nature 243 (1973)
346–347.
[20] N. Esaki, C.T. Walsh, Biochemistry 25 (1986) 3261–3267.
[21] K. Sandvig, O. Garred, A. van Helvoort, G. van Meer, B. van Deurs,
Mol. Biol. Cell. 7 (1996) 1391–1404.
[22] C.L. Schengrund, P. Kovac, Carbohydr. Res. 319 (1999) 24–28.
[23] C.A. Hutton, Org. Lett. 1 (1999) 295–297.
[24] M.B. Arcuri, S.J. Sabino, O.A.C. Antunes, E.G. Oestreicher, J.
Fluorine Chem. 121 (2003) 55–56.
References
[1] M.B. Arcuri, S.J. Sabino, O.A.C. Antunes, E.G. Oestreicher, J. Mol.
Catal. B: Enzym. 21 (2003) 107–111.
[25] G.F. Pinto, E.G. Oestreicher, Comput. Biol. Med. 18 (1988) 135–
144.
[2] M.B. Arcuri, S.J. Sabino, O.A.C. Antunes, E.G. Oestreicher, Catal.
Lett. 79 (2002) 1–4.
[26] W.W. Cleland, Biochim. Biophys. Acta 67 (1963) 104–137.
[27] I. Blank, J. Mager, E.D. Bergman, J. Chem. Soc. (1955) 2190–2193.
[28] U.H. Dolling, A.W. Douglas, E.J.J. Grabowsky, E.F. Schoenewaldt,
P. Sohar, M. Sletzinger, J. Org. Chem. 43 (1978) 1634–1640.
[29] E.F. Hartree, Anal. Biochem. 48 (1972) 422–427.
[30] C.E. Grimshaw, W.W. Cleland, Biochemistry 20 (1981) 5650–
5655.
[3] M.B. Arcuri, S.J. Sabino, O.A.C. Antunes, G.F. Pinto, E.G.
Oestreicher, Amino Acids 19 (2000) 477–482.
[4] L.P.B. Gonc¸alves, O.A.C. Antunes, G.F. Pinto, E.G. Oestreicher,
Tetrahedron: Asymmetry 11 (2000) 1465–1468.
[5] L.P.B. Gonc¸alves, O.A.C. Antunes, G.F. Pinto, E.G. Oestreicher, J.
Mol. Catal. B: Enzym. 4 (1998) 67–76.
[6] F.M. Bastos, T.K. Franc¸a, G.D.C. Machado, G.F. Pinto, E.G.
Oestreicher, L.M.C. Paiva, J. Mol. Catal. B: Enzym. 19 (2002)
459–465.
[31] J.R. Matos, M.B. Smith, C.-H. Wong, Bioorg. Chem. 13 (1985) 121–
130.
[32] F. Noll, in: J. Bergmeyer, M. Grabt (Eds.), Methods of Enzymatic
Analysis, vol. VI, third ed., VCH, Weinhein, 1988, pp. 582–588.
[33] C.M. Metzler, in: L. Endrenyi (Ed.), Kinetic Data Analysis: Design
and Analysis of Enzyme and Pharmacokinetic Experiments, Plenum
Press, New York, NY, 1981, pp. 25–37.
[7] C.H. Wong, G.M. Whitesides, Enzymes in Synthetic Organic
Chemistry, Elsevier, Oxford, 1994.
[8] J.M. Fang, C.H. Lin, C.W. Brashaw, C.-H. Wong, J. Chem. Soc.
Perkin Trans. 1 (1995) 967–978.