Slow-binding inhibition of 2-keto-3-deoxy-6-phosphogluconate (KDPG) aldolase

Article abstract of DOI:10.1016/j.bmc.2004.03.039

2-Keto-3-deoxy-6-phosphogluconate (KDPG) aldolase is a key enzyme in the Entner-Doudoroff pathway of bacteria. It catalyzes the reversible production of KDPG from pyruvate and D-glyceraldehyde 3-phosphate through a class I Schiff base mechanism. On the basis of aldolase mechanistic pathway, various pyruvate analogues bearing β-diketo structures were designed and synthesized as potential inhibitors. Their capacity to inhibit aldolase catalyzed reaction by forming stabilized iminium ion or conjugated enamine were investigated by enzymatic kinetics and UV-vis difference spectroscopy. Depending of the substituent R (methyl or aromatic ring), a competitive or a slow-binding inhibition takes place. These results were examined on the basis of the three-dimensional structure of the enzyme.

Full text of DOI:10.1016/j.bmc.2004.03.039

Bioorganic & Medicinal Chemistry 12 (2004) 2965–2972
Slow-binding inhibition of 2-keto-3-deoxy-6-phosphogluconate
(KDPG) aldolase
a
Remi Braga, Laurence Hecquet and Casimir Blonski
b
a,
*
ꢀ
aUniversitꢀe Paul Sabatier, LSPCMIB, UMR 5068, Groupe de Chimie Organique Biologique, B^at IIR1, 118 Route de Narbonne,
31062 Toulouse Cedex 4, France
bUniversitꢀe Blaise Pascal, Laboratoire SEESIB, UMR 6504, 24, avenue des Landais, Campus des Cꢀezeaux, 63177
Aubiꢁere Cedex, France
Received 28 October 2003; accepted 16 March 2004
Available online 21 April 2004
Abstract—2-Keto-3-deoxy-6-phosphogluconate (KDPG) aldolase is a key enzyme in the Entner–Doudoroff pathway of bacteria. It
catalyzes the reversible production of KDPG from pyruvate and D-glyceraldehyde 3-phosphate through a class I Schiff base
mechanism. On the basis of aldolase mechanistic pathway, various pyruvate analogues bearing b-diketo structures were designed
and synthesized as potential inhibitors. Their capacity to inhibit aldolase catalyzed reaction by forming stabilized iminium ion or
conjugated enamine were investigated by enzymatic kinetics and UV–vis difference spectroscopy. Depending of the substituent R
(methyl or aromatic ring), a competitive or a slow-binding inhibition takes place. These results were examined on the basis of the
three-dimensional structure of the enzyme.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
It has been shown that the accumulation of KDPG in
bacteria is correlated with an immediate and significant
decrease in growth.8 This suggests that KDPG aldolase
can be considered as a target for the development of new
bacteriostatic or bactericidal drugs. With the exception
of bromopyruvic acid9 and fluorodinitrobenzene,4;10 few
inhibitors of KDPG aldolase have been developed to
date, despite interest in such compounds. On the basis of
the aldolase mechanistic pathway (Scheme 1), we con-
sidered the possibility of obtaining slow-binding inhi-
bitions of this enzyme through the stabilization of
reaction intermediates (iminium ion or enamine) within
the enzyme by the use of pyruvate analogues bearing a
b-dicarbonyl structure.11 The expected reaction
where occurs the formation of iminium ion and
conjugated enamine is shown in Scheme 2. The stabil-
ization of these intermediates is possible through
hydrogen bonding and equilibrium between several
structures.
Aldolases are subject of continuous interest owing to their
central place in catalyzing carbon–carbon bond forma-
tion (or cleavage) in living organisms. 2-Keto-3-deoxy-6-
phosphogluconate (KDPG) aldolase (E.C. 4.1.2.14) is a
key enzyme in the Entner–Doudoroff pathway of bacte-
ria.1 It catalyzes the reversible production of KDPG from
pyruvate and
D-glyceraldehyde 3-phosphate (D-GAP)
through a class I Schiff base mechanism.2;3 The aldol
condensation proceeds by several ordered steps (Scheme
1): iminium ion (or Schiff base) formation between the
carbonyl of pyruvate and the e-amino groupof essential
Lys residue (Lys 133 for the enzyme from E. coli);4 (ii)
enamine formation after proton abstraction at C3 in the
iminium ion; (iii) enamine reaction with the carbonyl of
D
-GAP to form a new C–C bond and a second Schiff base;
and (iv) hydrolysis of the latter iminium ion leading to
KDPG with a S-configuration at C4 and free enzyme.
This enzyme displays considerable substrate tolerance
and appears as an useful catalyst for stereocontrolled
carbon–carbon bond formation between pyruvate and a
range of unnatural aldehydes.5–7
In the present paper, we describe the synthesis of
seven pyruvate analogues bearing such a structure
(Scheme 3) and the mechanisms by which these
compounds inhibit KDPG aldolase from E. coli. The
mode of interaction of these compounds with aldolase
was examined by enzyme kinetics and UV–vis difference
spectroscopy.
* Corresponding author. Tel.: +335-61-55-64-86; fax: +335-61-55-60-
11; e-mail: blonski@cict.fr
0968-0896/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.03.039

2966
R. Braga et al. / Bioorg. Med. Chem. 12 (2004) 2965–2972
Lys 133
NH2
H2O
H2O
H
NH
NH
H
NH
HO
HO
H
HO
O
CH3
OH
H
HO
H
O
O
O
CH3
pyruvic acid
HB
D-GAP
O
B
Glu 45
D-GAP
NH2
NH
OH
O
NH
H
H
H2O
H2O
H
HO
HOOC
HOOC
O
O
P
O
P
OH
OH
H
O
P
O
OH
OH
KDPG
Scheme 1. Reaction catalyzed by KDPG aldolase and mechanistic pathway.
NH2
H
H
H
H
H
O
O
H2O
H2O
N
O
H
H
N
O
N
O
HO
HO
HO
HO
R
R
R
R
H
H
O
O
O
O
B
Scheme 2. Expected mechanistic pathway for KDPG aldolase inhibition by synthesized compounds.
O
R1
O
O
R1
O
OEt
R2
R3
OEt
R2
R3
i
+
EtO
CH3
O
O
2a-f
ii
a: R1=R2=R3=H
b: R1=CF3, R2=R3=H
c: R2=CF3, R1=R3=H
d: R3=CF3, R1=R2=H
e: R2=OCH3, R1=R3=H
f: R3=OCH3, R1=R2=H
R1
O
O
R2
R3
OH
O
O
O
3a-f
OH
H3C
O
1
Scheme 3. Synthetic scheme for synthesized compounds. (i) NaH; (ii) NaOH, H2SO4. Surrounded is indicated the structure of compound 1.
2. Results
ketone with diethyl oxalate in the presence of sodium
hydride in refluxing toluene.12 Diketoesters 2a–f are then
converted to the expected pyruvic acid analogues 3a–f
(Scheme 3). Acetylpyruvic acid 1 is obtained starting
from the commercially available ethyl ester derivative in
the same way.13
2.1. Synthesis
The b-diketo structure of compounds 2a–f is generated
by Claisen condensation of the appropriate methyl-

R. Braga et al. / Bioorg. Med. Chem. 12 (2004) 2965–2972
2967
2.2. Enzymatic inhibition studies
16
14
12
10
8
2.2.1. Competitive inhibition of aldolase from E. coli by
acetylpyruvic acid 1. KDPG aldolase activity was com-
petitively inhibited in the presence of compound 1, with
Ki value of 250 ꢀ 20 lM. The Michaelis constant (Km)
determined for KDPG as substrate was 60 ꢀ 3 lM (not
shown).
6
2.2.2. Time-dependent inhibitions of aldolase by com-
pounds 3a–f. Incubation of aldolase with compound 3a–f
resulted in a first-order rate loss of enzyme activity (as
an example, Fig. 1 shows the results obtained with
compound 3a). The enzymatic activity is slowly but fully
restored by the addition of saturating concentration
(10 mM) of substrate pyruvate to the enzyme–inhibitor
complex solutions (data not shown). This result is con-
sistent with an inhibition occurring at the active site of
the enzyme.
4
2
0
0.00
-0.03
-0.02
-0.01
0.01
1/[I] (µM-1)
0.02
0.03
0.04
0.05
Figure 2. Compound 3a as slow-binding aldolase inhibitor. Residual
activity values at equilibrium were obtained from Figure 1. Data
points are experimental, and the line represents the best fit of the data
to Eq. 1. The Kiꢂ value determined was 70 ꢀ 10 lM.
The time-dependent reversible inhibition observed (Fig.
1) and reversion by addition of excess of substrate
suggest that compounds 3a–f behave as slow-binding
inhibitors of aldolase.14 The analysis of the extent of
inhibition (as an example, Fig. 2 shows the result ob-
tained with compound 3a) are consistent with a slow
equilibrium formation between the enzyme, the inhibitor
and the enzyme–inhibitor complex (see Eq. 1 in the
Experimental section). This allows to determine
the inhibition constant values, which are indicated in
Table 1.
Table 1. Dissociation constant values (Kiꢂ) describing the interaction
of compounds 3a–f with aldolase
Compounds
Kiꢂ (lM)
3a
3b
3c
3d
3e
3f
70 ꢀ 20
40 ꢀ 10
30 ꢀ 5
380 ꢀ 50
75 ꢀ 10
180 ꢀ 50
2.3. UV–vis difference spectroscopy
spectra characterized by the presence of isosbestic
points, which indicates that there was no accumulation
of intermediates nor side reactions. As an example,
Figure 3 shows spectra obtained with compound 3a.
2.3.1. Interactions of compounds 3a–f with aldolase. The
interactions of aldolase (10 lM subunits) with com-
pounds 3a–f (25–500 lM) resulted in UV–vis difference
The reactions of compound 3a–f with the lysine residue
of aldolase was modelled through the reaction of 3a–f
(250 lM) with aminocaproic acid (50 mM).15;16 The
0.18
0.16
0.14
0.12
25µM
0.10
0.08
50µM
0.06
100µM
0.04
200µM
500µM
0.02
0.00
0
10
20
40
50
time (mi3n0)
Figure 1. Time-dependent inhibition of aldolase by compound 3a.
Native recombinant aldolase (0.2 mg mLꢁ1) was incubated in the
presence of compound 3a at the indicated concentration in TEA buffer
(1 mL final volume). Aliquots were analyzed at various times for
enzymatic activity (see the Experimental section). In control performed
without inhibitor, no inhibition was observed.
Figure 3. Interaction of aldolase with compound 3a. UV–vis difference
spectra of 10 lM aldolase subunits and 100 lM of compound 3a in
TEA buffer (1 mL final volume, pH 7.6) with time (interval of 2 min).
Spectra are characterized by maximum at 390 nm, minimum at 318 nm
and the presence of isosbestic points around 270 and 350 nm.

2968
R. Braga et al. / Bioorg. Med. Chem. 12 (2004) 2965–2972
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
3. Discussion
As a first remark, we underline the simple and efficient
route that the Claisen reaction represents for the intro-
duction of the b-diketo frame present in all inhibitors
used in this work.
Second, the results obtained with compound 1, which
behaves as a competitive inhibitor, indicate that this
frame is recognized by the enzyme: the extension of the
carbon chain of two more carbon atoms does not pre-
vent the recognition by the enzyme, although there is a
loss of affinity from the substrate to compound 1 by a
factor of 4.
0
100
200
300
400
Then, the replacement of the methyl group in 1 by an
aromatic substituent in the series 3a–f changes the
inhibition, which becomes time dependent. The analysis
of the interactions between the enzyme and this set of
compounds is made by two independent approaches:
[I] (µM)
Figure 4. Determination of kinetic parameters for aldolase inhibition
by compound 3a. Data are experimental (provided by Fig. 3) and the
line represent the best fit of the data to Eq. 2. The k1 and k values
ꢁ1
determined were (2.61 ꢀ 0.03) · 103 Mꢁ1 minꢁ1
and 0.109 ꢀ
0.009 minꢁ1, respectively. This allowed us to calculate Kiꢂ value, which
was 70 ꢀ 10 lM.
• by inhibition kinetics: results shown in Figure 2, the
reversion of inhibition by adding an excess of sub-
strate, both evidence a slow-binding process, the
kinetic parameters of which can be determined. As
a matter of fact, compounds 3b and 3c exhibit Kiꢂ
values, clearly lower than Km;
same pattern of UV–vis difference spectra as for the
reactions with aldolase was obtained (not shown).
• by following in UV–vis difference spectroscopy the
formation of the complex between an inhibitor of
the 3a–f series and the enzyme; the significant absor-
bance and the parallel with the reference reaction
observed with aminocaproic acid indicate the forma-
tion of a stabilized intermediate, compatible with the
proposed Scheme 2 and in agreement with the
observed time-dependent kinetics. Moreover, the fair
agreement obtained by the two independent methods
between the two sets of Kiꢂ values, indicates that the
enzyme inactivation and the formation of the com-
plex are correlated. This point is also confirmed by
Difference absorbance spectra such as shown in Figure 1
were consistent with a first-order process, which corre-
lated with the loss of enzyme activity. Observed first-
order rate constants present no saturation kinetics, a
result consistent with the slow-binding scheme presented
in the Experimental section and with the enzymatic
inhibition studies. The kinetic parameter values associ-
ated with the EIꢂ complex formation were obtained by
fitting the data to Eq. 2 (as an example, Fig. 4 shows the
result obtained with compound 3a).
The resulting values are reported in Table 2 and the
dissociation constant values (Kiꢂ) calculated from these
results are in good agreement with those of Table 1,
determined by an other approach.
the good agreement on the k values, obtained in
ꢁ1
the one hand by inhibition kinetics and on the other
by the chase pulse experiment where the inhibitor is
displaced from the complex by adding large amounts
of substrate. As this complex has a significant absor-
bance, it is likely that not only the Schiff base forma-
tion but also the proton abstraction step occur within
the enzyme for the substrate analogues. This point
has to be confirmed by other methods, particularly
site directed mutagenesis. This will be done when pos-
sible residues acting as proton abstractor are identi-
fied from the X-ray structure, presently in progress,
Dissociation of the enzyme-3a–f complexes followed by
UV–vis difference spectroscopy was observed at satu-
rating concentration (10 mM) of substrate pyruvate (not
shown). The pyruvate chase fully restored enzyme
activity and the displacement of inhibitor from the
enzymatic complex was characterized by first-order rate
constants (k values) in good agreement with those
ꢁ1
determined by the other method (Table 2).
Table 2. Kinetic parameter values associated with aldolase inhibition by compounds 3a–f, determined by UV–vis difference spectroscopy technique
Kinetic parameters
Compounds
k1 ꢃ 103 (Mꢁ1 minꢁ1
)
k
(minꢁ1
)
Kiꢂ (lM)
ꢁ1
3a
3b
3c
3d
3e
3f
2.61 ꢀ 0.03
3.90 ꢀ 0.05
1.20 ꢀ 0.01
0.46 ꢀ 0.07
1.20 ꢀ 0.02
0.43 ꢀ 0.01
0.109 ꢀ 0.009
0.023 ꢀ 0.002
0.075 ꢀ 0.005
0.207 ꢀ 0.003
0.160 ꢀ 0.070
0.250 ꢀ 0.003
(0.123 ꢀ 0.025)a
(0.062 ꢀ 0.011)
(0.035 ꢀ 0.005)
(0.137 ꢀ 0.014)
(0.117 ꢀ 0.006)
(0.106 ꢀ 0.010)
70 ꢀ 10
10 ꢀ 1
60 ꢀ 10
450 ꢀ 10
130 ꢀ 60
580 ꢀ 20
a Values in parentheses were determined by chase pulse experiments (see text).

R. Braga et al. / Bioorg. Med. Chem. 12 (2004) 2965–2972
2969
of crystals of inhibitor 3b and KDPG aldolase from
E. coli. For now, we underlined the interest of the
UV spectroscopic method, which leads to reliable
results and is of particular interest to investigate com-
plexes formed between effectors and enzymes even
after they lost their catalytic activity, for instance
mutants.
measured by means of triose-phosphate isomerase/
glycerol phosphate dehydrogenase as previously
reported.21 The protein concentration was estimated
using the molecular absorbance at k ¼ 280 nm (e ¼
18; 260 L molꢁ1 cmꢁ1, calculated by Schepartz Lab Bio-
polymer Calculator22). The subunit concentration was
determined assuming a molecular weight of 67,000 for
the trimeric aldolase.
The next point is related to the substituent effect ob-
served with inhibitors of the 3a–f series: it can be noticed
from Table 2 that the best Kiꢂ values are obtained with
inhibitors bearing an electron withdrawing substituent
at position ortho or meta (3b and 3c), whereas the
reverse situation is observed with the electron releasing
4.3. Kinetic parameters and equilibrium constants values
determination
The data were subjected to a linear least square fit to the
appropriate equation using the Origin program.
grouppresent in
3e and 3f. To account for this, we
considered the X-ray structure obtained with KDPG
aldolase from E. coli where the substrate pyruvate has
been trapped in the carbinolamine form.17 This structure
shows that an aromatic residue, namely phenyl-alanine
135 points towards the C3 carbon (or methyl group) in
pyruvate, position where the group Aryl–C(O) is linked.
It is therefore conceivable that p–p interactions occur
between Phe-135 and the aryl groupin the inhibitor, the
interaction being stronger between an electron rich and
an electron depleted aromatic rings than between two
rings of similar densities, as observed for instance in
Meisenheimer complexes.18 Along these lines, it can be
understood that CF3 as substituent on the ring leads to
stronger inhibition than compounds 3e or 3f bearing the
4.4. Competitive inhibition study
The dissociation constant of the enzyme–inhibitor
complex was determined by double reciprocal plots of
initial velocities of aldolase (0.33 lg mLꢁ1) for KDPG
and compound 1 concentrations in the range of 100–500
and 0–500 lM, respectively.
4.5. Inhibition study
Aldolase (0.2 mg mLꢁ1 in 0.1 mL TEA buffer) was
incubated in the presence of the compound under study
(0.025–0.5 mM). The enzymatic activity was assayed as a
function of time with 10 lL aliquots. Control experi-
ments were run without inhibitor and all measurements
were made in duplicate.
electron donating groupCH O, the phenyl ring in Phe-
3
135 being electron rich. Again, a better understanding of
this situation will be obtained when the X-ray structure
of the complex of E. coli KDPG aldolase-3b will be
solved. This may allow also to introduce other substit-
uents on the chain, possibly contributing to an improved
affinity. For now, we present the access to a new class of
KDPG aldolase inhibitors and a strategy to study and
improve them.
4.6. Equations
In this study, slow-binding inhibition14 involves a slow
equilibrium formation between the enzyme E, the
inhibitor I and the enzyme–inhibitor complex EIꢂ as
shown below:
4. Experimental
4.1. Enzymes and reagents
k1
k
k1
ꢁ1
E þ I ꢀ EIꢂ Ki ¼
Pyruvic acid sodium salt, glycerol phosphate dehydro-
genase, triose-phosphate isomerase and lactic dehydro-
genase were purchased from Boehringer Mannheim. All
other chemicals were purchased from Aldrich and were
used without extra purification. KDPG aldolase from
E. coli was kindly provided to us by Pr Sygusch (Mon-
treal University). Substrate KDPG was prepared as
described previously.19
k
ꢁ1
Kiꢂ is the dissociation constant for the EIꢂ complex.
Characterization of slow-binding inhibition: in the
absence of substrate, the following two equations were
used for the determination of the kinetic parameter
values:
1
1
Kiꢂ
¼
þ
ð1Þ
ð2Þ
4.2. Assay methods
Vm ꢁ Vi Vm Vm½IŠ
Aldolase activity (25 units mgꢁ1 at 25 ꢁC) was measured
by means of lactic dehydrogenase method in 1 mL of
0.1 M triethanolamine hydrochloride buffer (pH 7.6 and
ionic strength 0.15), using KDPG (1 mM) as substrate as
described elsewhere,20 except for compound 1, which
was found to inhibit lactic dehydrogenase; aldolase
activity in the presence of this compound was then
kobs ¼ kꢁ1 þ k1½IŠ
Vm and Vi represent, respectively, the residual enzymatic
activity in the absence and in the presence of the
inhibitor; the apparent first-order rate constants (kobs
)
describing the formation of EIꢂ parallel the increasing
concentration of inhibitor ([I]).

2970
R. Braga et al. / Bioorg. Med. Chem. 12 (2004) 2965–2972
4.7. UV–vis difference spectroscopy
25 mL of water. The organic layer was dried over so-
dium sulfate and evaporated in vacuo. The remaining
product was purified by silica chromatography (CHCl3)
to yield 2a as a yellow oil (0.9 g, 25%). 1H NMR
Absorbance spectra were measured using a Cary 1E
Varian spectrophotometer at constant 25 ꢁC. The same
TEA buffer as for enzymatic kinetic studies was utilized.
Absorbance spectra were measured using a previously
described method.15;16 Absorption spectra were scanned
between 250 and 450 nm and recorded as a function of
time. Measurements were initiated by addition of the
compound under study at various final concentrations
to TEA buffer solutions containing a fixed concentration
of aldolase (10 lM subunit). The measured absorption
spectra of the enzyme complex recorded at timed
intervals were corrected for absorption by buffer, com-
pound under study and the enzyme alone. For the model
study, absorption spectra were recorded at timed inter-
vals and corrected for absorption by buffer, compounds
and aminocaproic acid. Observed apparent first-order
rate constants, kobs were obtained for each assay by fit-
ting the time-dependent absorption data against a first-
order kinetic equation using the Origin program. The
dissociation constant (Kiꢂ) and the rate constants k1 and
3
(250 MHz, CDCl3): 7.95 (d, 2H, JH–H ¼ 8 Hz), 7.56–
7.36 (m, 3H), 7.01 (s, 1H), 4.30 (q, 2H,
3
3JH–H ¼ 7 Hz),1.34 (t, 3H, JH–H ¼ 7 Hz). 13C NMR
(63 MHz, CDCl3): 190.7, 169.7, 139.5, 134.8, 133.7,
128.8, 127.8, 97.9, 62.5, 14.1.
4.8.3. Ethyl 2,4-dioxo-4-[2-(trifluoromethyl)phenyl]but-
anoate (2b). The compound was prepared by following
the same procedure as 2a, using (2-trifluoromethyl)ace-
tophenone (1 g, 5.3 mmol), diethyloxalate (1.1 mL,
8.0 mmol) and sodium hydride (0.3 g, 10.6 mmol). The
remaining product was purified by silica chromatogra-
phy (toluene/AcOEt 7:3–1:1) to yield 2b as an orange oil
1
(0.4 g, 23%). 19F NMR (188 MHz, CDCl3): 17.39. H
3
NMR (250 MHz, CDCl3): 7.61 (d, 1H, JH–H ¼ 7 Hz),
7.38 (s, 2H), 7.25 (m, 1H), 6.21 (s, 1H), 4.12 (q, 2H,
3
3JH–H ¼ 7 Hz), 1.21 (t, 3H, JH–H ¼ 7 Hz). 13C NMR
k
(corresponding to the formation and dissociation of
(63 MHz, CDCl3): 192.3, 169.4, 166.9, 142.0, 125.4 (q,
3JC–F ¼ 250 Hz), 130.4, 129.8, 127.4, 126.2, 99.0, 62.0,
13.8.
ꢁ1
the slow-binding aldolase–inhibitor complex, respec-
tively), are derived from analysis of apparent rate con-
stants using Eq. 2.
The first-order rate constant k was also derived inde-
4.8.4. Ethyl 2,4-dioxo-4-[3-(trifluoromethyl)phenyl]but-
anoate (2c). The same procedure as for 2b was fol-
lowed, using (3-trifluoromethyl)acetophenone as
starting material. The remaining residue was purified by
silica chromatography (toluene to toluene/AcOEt 8:2) to
yield 2c as an orange oil (0.2 g, 12%). 19F NMR
(188 MHz, CDCl3): 12.63. 1H NMR (250 MHz, CDCl3):
ꢁ1
pendently from absorbance data corresponding to the
displacement of the inhibitor under study from the
enzymic complex by 10 mM pyruvate.
4.8. Synthesis
3
8.17 (s, 1H), 8.11 (d, 1H, JH–H ¼ 8 Hz), 7.80 (d, 1H,
3
NMR spectra were recorded in CDCl3, CD3OD, ace-
tone-d6 or DMSO-d6 on Bruker AC200 and AC250
spectrometers. All chemical shifts (d) are reported in
parts per million (ppm) with respect to TMS for 1H and
13C spectra and trifluoroacetic acid for 19F spectra as
internal standard.
3JH–H ¼ 8 Hz), 7.60 (t, 1H, JH–H ¼ 8 Hz), 7.04 (s, 1H),
3
3
4.36 (q, 2H, JH–H ¼ 7 Hz), 1.37 (t, 3H, JH–H ¼ 7 Hz).
13C NMR (63 MHz, CDCl3): 188.7, 170.7, 161.8, 131.5
3
(q, JC–F ¼ 250 Hz), 130.8, 129.9, 127.6, 124.3, 121.3,
120.2, 97.7, 62.7, 13.9.
4.8.5. Ethyl 2,4-dioxo-4-[4-(trifluoromethyl)phenyl]but-
anoate (2d). Reagents were used according to the same
procedure as 2a, using (4-trifluoromethyl)acetophenone
(0.5 g, 2.7 mmol), diethyloxalate (0.6 mL, 4.0 mmol) and
sodium hydride (0.2 g, 5.3 mmol). The product was
purified by silica chromatography (CHCl3 to CHCl3/
4.8.1. Acetylpyruvic acid (1). This compound was pre-
pared as described previously,13 starting form the ethyl
ester derivative.
1H NMR (250 MHz, CDCl3): 6.46 (s, 1H), 2.28 (s, 3H).
13C NMR (63 MHz, CDCl3): 199.2, 167.6, 164.4, 101.8,
27.1.
CH3OH 9:1) to yield 2d as an orange oil (0.3 g, 37%). 19
F
NMR (188 MHz, CDCl3): 12.4. 1H NMR (250 MHz,
3
CDCl3): 8.09 and 7.72 (d, 4H, JH–H ¼ 8 Hz), 7.06
3
The esters 2a–f were prepared by Claisen condensation
of diethyloxalate with the corresponding methyl ketone
as previously reported.23
(s, 1H), 4.38 (q, 2H, JH–H ¼ 7 Hz), 1.26 (t, 3H,
3JH–H ¼ 7 Hz). 13C NMR (63 MHz, CDCl3): 188.5,
3
171.4, 162.9, 137.8, 132.5 (q, JC–F ¼ 32 Hz), 128.4,
125.9, 122.1, 97.9, 45.8, 14.1.
4.8.2. Ethyl 2,4-dioxo-4-phenylbutanoate (2a). To a
mixture of acetophenone (2 g, 16.6 mmol) and diethy-
loxalate (3.4 mL, 25.0 mmol) in 50 mL of anhydrous
toluene was added sodium hydride (0.8 g, 33.3 mmol).
The resulting mixture was stirred at 80 ꢁC for 2 h; the
solvent was evaporated under vacuo, the crude mixture
placed in 50 mL of ethyl acetate and washed once with
4.8.6. Ethyl 4-(3-methoxyphenyl)-2,4-dioxobutanoate
(2e). (3-Methoxy)acetophenone (1 g, 6.7 mmol), diethyl-
oxalate (2.0 mL, 10.0 mmol) and sodium hydride (0.3 g,
13.4 mmol) were utilized as described above. The crude
product was purified by silica chromatography (CH2Cl2/
n-C6H12 8:2) to yield 2e as a colourless oil (0.8 g, 49%).

R. Braga et al. / Bioorg. Med. Chem. 12 (2004) 2965–2972
2971
3
1H NMR (250 MHz, acetone-d6): 7.38 (t, 2H,
131.5 (q, JC–F ¼ 250 Hz), 130.9, 127.8, 126.1, 124.7,
121.3, 96.4. Anal. Calcd for C11H7O4F3: C, 50.78; H,
2.71. Found: C, 50.38; H, 2.59. Mass spectrometry
(FAB): 261 [MþH]þ.
3
3JH–H ¼ 8 Hz), 7.27 (t, 1H, JH–H ¼ 8 Hz), 7.01 (d, 1H,
3
3JH–H ¼ 8 Hz), 6.90 (s, 1H), 4.24 (q, 2H, JH–H ¼ 7 Hz),
3
3.71 (s, 3H), 1.28 (t, 3H, JH–H ¼ 7 Hz). 13C NMR
(63 MHz, acetone-d6): 190.6, 169.2, 162.0, 159.9, 155.2,
152.1, 149.5, 142.3, 112.2, 98.0, 62.4, 55.3, 14.0.
4.8.11. 2,4-Dioxo-4-[4-(trifluoromethyl)phenyl]butanoic
acid (3d). By following the same procedure as for 3b,
compound 2d (0.29 g, 1.0 mmol) yielded the expected
compound 3d as a white powder (0.11 g, 44%). 19F
NMR (188 MHz, DMSO-d6): 18.6. 1H NMR (250 MHz,
4.8.7. Ethyl 4-(4-methoxyphenyl)-2,4-dioxobutanoate
(2f). The same procedure as for 2a was followed, using
(4-methoxy)acetophenone (1 g, 6.7 mmol), diethyloxa-
late (2.0 mL, 10.0 mmol) and sodium hydride (0.3 g,
13.4 mmol). The product was purified by silica chro-
matography (diethyl ether/petroleum ether 1:1) to yield
3
DMSO-d6): 8.23 (d, 2H, JH–H ¼ 8 Hz), 7.90 (d, 2H,
3JH–H ¼ 8 Hz), 7.11 (s, 1H), 2.49 (m, 2H). 13C NMR
(63 MHz, DMSO-d6): 187.8, 171.2, 162.9, 138.0, 132.5
(q, 3JC–F ¼ 32 Hz), 130.0, 125.8, 121.4, 98.2. Anal. Calcd
for C11H7O4F3,H2O: C, 47.49; H, 3.26. Found: C, 47.12;
H, 3.08. Mass spectrometry (FAB): 261 [MþH]þ.
1
2f as a colourless oil (0.4 g, 25%). H NMR (250 MHz,
3
CDCl3): 7.82 (d, 2H, JH–H ¼ 9 Hz), 6.83 (d, 2H,
3
3JH–H ¼ 9 Hz), 6.88 (s, 1H), 4.27 (q, 2H, JH–H ¼ 7 Hz),
3
3.75 (s, 3H), 1.29 (t, 3H, JH–H ¼ 7 Hz). 13C NMR
(63 MHz, CDCl3): 190.2, 167.9, 164.3, 162.3, 158.2,
154.3, 152.9, 150.1, 114.1, 97.6, 62.4, 55.5, 14.0.
4.8.12. 4-(3-Methoxyphenyl)-2,4-dioxobutanoic acid (3e).
This compound was prepared from 2e (0.40 g, 1.6 mmol)
by following the above procedure to yield the expected
1
4.8.8. 2,4-Dioxo-4-phenylbutanoic acid (3a). To a solu-
tion of 2a (0.12 g, 0.51 mmol) in 5 mL of water was ad-
ded 0.25 mL of sodium hydroxide 4 N, and the mixture
was vigorously stirred at room temperature for 2 min.
The solution was then acidified with H2SO4 6 N to pH 3,
the crude product was extracted six times with 10 mL
diethyl ether; the organic layers were pooled and the
solvent evaporated in vacuo. The remaining product
was crystallized (using CCl4 as solvent) to yield 3a as a
yellow powder (0.10 g, 67%). 1H NMR (250 MHz,
compound 3e as a pink powder (0.20 g, 57%). H NMR
(250 MHz, CD3OD): 7.68 (d, 1H, 3JH–H ¼ 8 Hz), 7.58 (s,
3
1H), 7.49 (t, 1H, JH–H ¼ 8 Hz), 7.26 (d, 1H,
3JH–H ¼ 8 Hz), 7.14 (s, 1H), 3.92 (s, 3H). 13C NMR
(63 MHz, CD3OD): 191.3, 163.1, 161.0, 130.9, 130.8,
121.1, 120.9, 120.7, 120.0, 98.5, 55.8. Anal. Calcd for
C11H10O5: C, 59.46; H, 4.54. Found: C, 59.28; H, 4.34.
Mass spectrometry (FAB): 223 [MþH]þ.
3
CDCl3): 8.02 (d, 2H, JH–H ¼ 8 Hz), 7.67–7.49 (m, 3H),
4.8.13. 4-(4-Methoxyphenyl)-2,4-dioxobutanoic acid (3f).
This compound was prepared from 2f (0.40 g, 1.6 mmol)
by following the above procedure to yield 3f as a red
7.18 (s, 1H). 13C NMR (63 MHz, CDCl3): 187.3, 173.3,
163.5, 133.5, 134.1, 129.0, 127.9, 95.9. Anal. Calcd for
C10H8O4: C, 62.50; H, 4.20. Found: C, 62.35; H, 4.15.
Mass spectrometry (FAB): 193 [MþH]þ.
1
powder (0.22 g, 61%). H NMR (250 MHz, acetone-d6):
3
3
8.11 (d, 2H, JH–H ¼ 8 Hz), 7.13 (d, 2H, JH–H ¼ 8 Hz),
7.11 (s, 1H), 3.93 (s, 3H). 13C NMR (63 MHz, acetone-
d6): 191.2, 165.4, 163.4, 131.1, 130.6, 125.4, 115.1, 97.9,
56.0. Anal. Calcd for C11H10O5: C, 59.46; H, 4.54.
Found: C, 59.44; H, 4.54. Mass spectrometry (FAB):
223 [MþH]þ.
4.8.9.
2,4-Dioxo-4-[2-(trifluoromethyl)phenyl]butanoic
acid (3b). This compound was prepared from 2b
(0.34 g, 1.2 mmol) by following the same procedure as
for 3a using sodium hydroxide pellets (0.06 g, 2.4 mmol),
except that the mixture was vigorously stirred at room
temperature for half an hour. The crude product was
then crystallized (CCl4) to yield 3b as a white powder
(0.11 g, 36%). 19F NMR (188 MHz, DMSO-d6): 23.43.
Acknowledgements
We are thankful to J. Sygusch for providing KDPG
aldolase. We also acknowledge Y. Boublik, M. Bardet
and M. Sancelme for technical assistance.
1H NMR (250 MHz, DMSO-d6): 7.87–7.69 (m, 4H),
13
6.47 (s, 1H).
C NMR (63 MHz, DMSO-d6): 185.4,
3
163.3, 137.2, 132.6, 126.5 (q, JC–F ¼ 250 Hz), 130.0,
129.2, 128.9, 128.7, 102.5. Anal. Calcd for C11H7O4F3:
C, 50.78; H, 2.71. Found: C, 50.35; H, 2.61. Mass
spectrometry (FAB): 261 [MþH]þ.
References and notes
1. Conway, T. FEMS Microbiol. Rev. 1992, 9, 1.
2. Meloche, H. P.; Wood, W. A. J. Biol. Chem. 1964, 239,
3511.
3. Grazi, E.; Meloche, H. P.; Martinez, G.; Wood, W. A.;
Horecker, B. L. Biochem. Biophys. Res. Commun. 1963,
10, 4.
4. Ingram, J. M.; Wood, W. A. J. Biol. Chem. 1965, 240,
4146.
5. Allen, S. T.; Heintzelman, G. R.; Toone, E. J. J. Org.
Chem. 1992, 57, 426.
4.8.10. 2,4-Dioxo-4-[3-(trifluoromethyl)phenyl]butanoic
acid (3c). By following the same procedure as for 3b,
compound 2c (0.14 g, 0.5 mmol) yielded the expected
compound 3c as a white powder (0.05 g, 39%). 19F NMR
(188 MHz, CDCl3): 12.65. 1H NMR (250 MHz, CDCl3):
3
8.25 (s, 1H), 8.17 (d, 1H, JH–H ¼ 8 Hz), 7.88 (d, 1H,
3
3JH–H ¼ 8 Hz), 7.67 (t, 1H, JH–H ¼ 8 Hz), 7.20 (s, 1H).
13C NMR (63 MHz, CDCl3): 186.3, 173.7, 162.4, 134.4,

2972
R. Braga et al. / Bioorg. Med. Chem. 12 (2004) 2965–2972
6. Machajewski, T. D.; Wong, C. H. Angew. Chem., Int. Ed.
2000, 39, 1352.
7. Shelton, M. C.; Cotterill, I. C.; Novak, S. T. A.;
Poonawala, R. M.; Sudarshan, S.; Toone, E. J. J. Am.
Chem. Soc. 1996, 118, 2117.
8. Fuhrman, L. K.; Wanken, A.; Nickerson, K. W.; Conway,
T. FEMS Microbiol. Lett. 1998, 159, 261.
9. Meloche, H. P. Biochemistry 1967, 6, 2273.
10. Barran, L. R.; Wood, W. A. J. Biol. Chem. 1971, 246,
4028.
11. Gefflaut, T.; Blonski, C.; Perie, J. Bioorg. Med. Chem.
1996, 4, 2043.
12. Marvel, C. S.; Dreger, E. E.. In: Org. Synth. Coll.; Wiley:
New York, 1941; Vol. I, p238.
13. Lehninger, A. L.; Witzemann, E. J. J. Am. Chem. Soc.
1942, 64, 874.
14. Morrison, J. F.; Walsh, C. T. Adv. Enzyme Relat. Areas
Mol. Biol. 1987, 61, 201.
15. Santi, D. V.; Ouyang, T. M.; Tan, A. K.; Gregory, D. H.;
Scanlan, T.; Carreras, C. W. Biochemistry 1993, 32, 11819.
16. Blonski, C.; De Moissac, D.; Perie, J.; Sygusch, J.
Biochem. J. 1997, 323, 71.
17. Allard, J.; Grochulski, P.; Sygusch, J. Proc. Natl. Acad.
Sci. U.S.A. 2001, 98, 3679.
18. Jones, J. B.; Sih, C. J.; Perlman, D. In Applications of
Biochemical Systems in Organic Chemistry––Part II; Wiley:
New York, 1976; Vol. X.
19. O’Connell, E. L.; Meloche, H. P. Methods Enzymol. 1982,
89, 98.
20. Meloche, H. P.; Ingram, J. M.; Wood, W. A. Methods
Enzymol. 1967, 9, 520.
21. Rose, I. A.; O’Connell, E. L.; Mehler, A. H. J. Biol. Chem.
1965, 240, 1758.
22. Schepartz Lab Biopolymer Calculator, found in http://
paris.chem.yale.edu/extinct.html.
23. Pei, Y.; Wickham, B. O. S. Tetrahedron Lett. 1993, 34, 7509.