Báo cáo Y học: Synthesis of phosphoenol pyruvate (PEP) analogues and evaluation as inhibitors of PEP-utilizing enzymes pot

Synthesis of phospho
enol
pyruvate (PEP) analogues and evaluation
as inhibitors of PEP-utilizing enzymes
Luis Fernando Garcı
´
a-Alles and Bernhard Erni
Departement fu
¨
r Chemie und Biochemie, Universita
¨
t Bern, Switzerland
The synthesis of 10 new phosphoenolpyruvate (PEP)
analogues with modifications in the phosphate and the
carboxylate function is described. Included are two potential
irreversible inhibitors of PEP-utilizing enzymes. One incor-
porates a reactive chloromethylphosphonate function
replacing the phosphate group of PEP. The second contains
a chloromethyl group substituting for the carboxylate
function of PEP. An improved procedure for the prepar-
ation of the known (Z)- and (E)-3-chloro-PEP is also given.
The isomers were obtained as a 4 : 1 mixture, resolved by
anion-exchange chromatography after the last reaction step.
The stereochemistry of the two isomers was unequivocally
assigned from the
3
J
H-C
coupling constants between the
carboxylate carbons and the vinyl protons.
All of these and other known PEP-analogues were tested

as reversible and irreversible inhibitors of Mg
2+
-andMn
2+
-
activated PEP-utilizing enzymes: enzyme I of the phos-
phoenolpyruvate:sugar phosphotransferase system (PTS),
pyruvate kinase, PEP carboxylase and enolase. Without
exception, the most potent inhibitors were those with sub-
stitution of a vinyl proton. Modification of the phosphate
and the carboxylate groups resulted in less effective com-
pounds. Enzyme I was the least tolerant to such modifica-
tions. Among the carboxylate-modified analogues, only
those replaced by a negatively charged group inhibited
pyruvate kinase and enolase. Remarkably, the activity of
PEP carboxylase was stimulated by derivatives with neutral
groups at this position in the presence of Mg
2+
, but not with
Mn
2+
. For the irreversible inhibition of these enzymes,
(Z)-3-Cl-PEP was found to be a very fast-acting and efficient
suicide inhibitor of enzyme I (t
1/2
¼ 0.7 min).
Keywords: phosphoenolpyruvate analogues; chemical
synthesis; inhibition; irreversible inhibitor; PEP-utilizing
enzymes.
Phosphoenolpyruvate (PEP) is a small and highly

functionalized molecule that plays a central role in metabo-
lism. It is not only important because of its high phosphate
group-transfer potential (DG ¼ )61.9 kJÆmol
)1
), but also
because it is a versatile C
3-
synthon in C–C, C–P and C–O
bond-formation reactions [1]. Representative examples of
the first function are the synthesis of ATP catalysed by
pyruvate kinase, and the transport with concomitant
phosphorylation of carbohydrates across the bacterial
membrane, mediated by the PEP:sugar phosphotransferase
system (PTS) [2]. Examples of the second function are the
fixation of CO
2
in plants (mediated by PEP carboxylase) [3],
the generation of natural phosphonates (PEP mutase) [4],
the first step in peptidoglycan cell-wall biosynthesis (cata-
lysed by UDP-GlcNAc enolpyruvyl transferase) and the
biosynthesis of aromatic amino acids (3-deoxy-
D
-arabino-
heptulosonate-7-phosphate synthase and 5-enolpyruvyl-
shikimate-3-phosphate synthase) [1].
Because of its pivotal role in metabolism, PEP has been
the subject of extensive chemical modification. Most of the
pseudosubstrates or competitive inhibitors discovered so far
differed from PEP by the presence of substitutions distal to
the phosphate group (position C-3, similar to compounds

1b–e, Scheme 1) [5–7]. Some of these compounds turned out
to be crucial in mechanistic studies of PEP-utilizing
enzymes, for instance in the establishment of the stereo-
chemical course of enzymatic processes mediated by
enzyme I of the PTS [8], UDP-GlcNAc enolpyruvyl trans-
ferase and 5-enolpyruvyl-shikimate-3-phosphate synthase
[9], 3-deoxy-
D
-arabino-heptulosonate-7-phosphate synthase
[10], pyruvate kinase [11], KDOP synthase [12], enolase [13],
PEP carboxykinase [14], and PEP carboxylase [15].
A representative example is the study of UDP-GlcNAc
enolpyruvyl transferase, and 5-enolpyruvyl-shikimate-
3-phosphate synthase, with (Z)-F-PEP (1b), an analogue
that allowed the isolation and characterization of stable
fluoro analogues of the otherwise unstable tetrahedral
intermediate of the normal reaction [1].
The carboxylic and the phosphate functionalities of PEP
have been modified less frequently. Several studies indicated
that both groups might be essential to establish the correct
substrate–active site contacts in pyruvate kinase and enolase
[6]. Important exceptions are phosphoenolthiopyruvate [16],
thiophosphoenolpyruvate [17], and the remarkable case of
sulfoenolpyruvate (3a) a substrate that transfers its sulfuryl
group to ADP in the presence of pyruvate kinase [18].
This paper presents the results obtained with 17 PEP
analogues (Scheme 1) as inhibitors of the reactions
catalysed by the enzyme I of the PTS, pyruvate kinase,
Correspondence to L. F. Garcı
´

a Alles. Departement fu
¨
r Chemie
und Biochemie. Universita
¨
t Bern. Freiestrasse 3. CH-3012 Bern,
Switzerland.
Fax: + 41 31/631 48 87, Tel.: + 41 31/631 37 92,
E-mail:
Abbreviations: PEP, phosphoenolpyruvate; PTS, phosphoenolpyru-
vate:sugar phosphotransferase system; FC, flash chromatography;
HRMS, high resolution mass spectrometry.
(Received 20 February 2002, revised 23 April 2002,
accepted 15 May 2002)
Eur. J. Biochem. 269, 3226–3236 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.02995.x
PEP carboxylase, and enolase. Compounds 2a–e, 2g–i and
3c are new, and for that reason their synthesis and
characterization is also reported, as well as an improved
method for the preparation of the 3-Cl-PEP analogues 1c
and 1d. The last two isomers and the chlorinated analogues
2g and 3c are candidates for the irreversible inactivation of
PEP-utilizing enzymes.
MATERIALS AND METHODS
Enzyme I, and the rest of components from the PTS were
expressed and purified as previously described [19]. Pyruvate
kinase from rabbit muscle (2000 UÆmL
)1
),
L
-lactate dehy-

drogenase from rabbit muscle (550 UÆmg
)1
), and glucose-
6-phosphate dehydrogenase from yeast (350 UÆmg
)1
)were
from Boehringer Mannheim. Malate dehydrogenase from
porcine heart (2700 UÆmL
)1
) and PEP carboxylase from
maize (50 UÆmL
)1
) were from Fluka. Enolase from bakers
yeast (500 UÆmL
)1
) was purchased from Sigma. NADP
(sodium salt), pyruvic acid,
D
-2-phosphoglyceric acid
(sodium salt), bromoacetyl bromide (4d), 1,3-dichloro-2-
propanone (4e), bromo and chloro trimethylsilane, trimethyl
phosphite, methyl
D
,
L
-lactate, chloromethylphosphonic acid
dichloride and potassium thioacetate were from Fluka.
4-Chloroacetoacetic acid methyl ester (4a) and 1-acetoxy-3-
chloroacetone (4b) were from TCI America. Ethyl pyruvate,
3-bromo-1,1,1-trifluoroacetone (4c) and dimethyl chloro-

phosphate from Aldrich. PEP (cyclohexylammonium salt)
and NADH (disodium salt) were from Sigma. Solvents were
usually of the highest purity commercially available. Ben-
zene was dried by continuous refluxing over and distillation
from sodium. Fluka silica gel 60/230-400 mesh was used in
column chromatography purification. Ion-exchange chro-
matography was carried out using Dowex 50W X-8 (50–100
mesh) from Fluka and Sephadex DEAE A-25 column from
Pharmacia. Deuterated solvents were purchased from
Armar AG (D
2
O, CD
3
OD) and Fluka (CDCl
3
).
Characterization of the PEP-analogues 1–3, butyrate is
given in Table 1. (Z)-3-F-PEP (1b)and(Z)-phosphoenol-
butyrate [1e (Z)-3-Me-PEP] were a generous gift of
R. L. Somerville (Department of Biochemistry, Purdue
University, West Lafayette, IN, USA). They were contam-
inated with around 6% of their E-isomers, as judged
from their
1
H-NMR spectra. Phospho-
D
,
L
-lactic acid (1f)
was obtained via condensation of methyl-

D
,
L
-lactate and
dimethyl chlorophosphate [20], followed by phosphate ester
demethylation using trimethylsilyl bromide (step 2,see
below), and hydrolysis of the carboxylic methyl ester at
pH ¼ 12.0 (step 3). Published methods were also followed
for the synthesis of sulfoenolpyruvate (3a)[18],and
a-(dihydroxyphosphonylmethyl) acrylic acid (3b)[6].
1
H- and
13
C-NMR spectra were recorded at 300.1 and
75.47 MHz, respectively. Spectra in D
2
O were calibrated
against sodium 3-(trimethyl)propane-1-sulfonate (external
standard).
31
P-NMR at 81 MHz were calibrated against a
85% phosphoric acid external standard (d ¼ 0.00 p.p.m.).
Due to the pH dependence of phosphate and phosphonate
chemical shifts, phosphorus data reported for the final
products were acquired in double-distilled H
2
Oat
pH ¼ 7.1–7.4. Iodide (m/z 126.9045) and taurocholate
(m/z 514.2839) were used as internal standards in negative
mode high-resolution ESI-MS measurements.

Ethyl 3,3-dichloropyruvate (5a) was prepared by stirring
a mixture of fresh ethyl pyruvate (2.9 g, 25 mmol), sulfuryl
chloride (4.1 mL, 50 mmol), and p-toluenesulfonic acid
dihydrate (0.24 g, 1.25 mmol) at 70 °C. Extra sulfuryl
chloride (50 mmol) was added after 4 and 8 h of reaction.
The reaction was continued for a total of 24 h. Excess
sulfuryl chloride was removed by distillation and water
(10 mL) was added. The reaction mixture was extracted
with diethyl ether (3 · 15 mL), the organic layer dried over
anhydrous magnesium sulfate, and the solvent evacuated.
Silica gel flash chromatography (FC) (hexanes/ethyl acetate
Scheme 1.
Table 1.
1
H-NMR spectral data (D
2
O, noninterchangeable signals) of
analogues 1–3. Products as cyclohexylammonium salts, except 1c, 1d,
2b (triethylammonium), 3a (potassium salt) and 3b (acid form). Signals
due to cyclohexylamonium: d ¼ 3.11 p.p.m. (m, 1H), 1.94 (m, 2H),
1.76 (m, 2H), 1.62 (m, 1H), 1.30 (m, 5H) and triethylammonium:
d ¼ 3.17 p.p.m. (q, 6 H), 1.05 (t, 9H).
Product
d (p.p.m.) (multiplicity, J in Hz)
Vinyl protons R
1
,R
2
,R
3

1b 7.62 (dd, 73.5, 2.9) –
1c 6.72 (d, 1.1) –
1d 6.17 (d, 1.1) –
1e 6.32 (dq, 7.4, 2.2) 1.75
(dd, 7.4, 2.9, 3H)
1f 4.38 (p, 7.0), 1.34 (d, 7.0, 3H) –
2a 4.76 (d, 1.1), 4.42 (d, 0.7) 3.24 (s, 2H)
3.73 (s, 3H)
2b 4.77 (dd, 2.2, 1.8), 4.56 (t, 1.8) 3.20 (s, br, 2H)
2c 4.84 (s, br), 4.57 (s, br) 4.52 (s, br, 2H)
2.14 (s, 3H)
2d 4.64 (s, br), 4.44 (s, br) 3.91 (s, br, 2H)
2e 5.16–5.12 (m, 2H) –
2f 5.23–5.08 (m, 2H) –
2g 4.82 (s, br)
a
, 4.68 (s, br) 4.09 (s, br, 2H)
2h 4.66 (s, br), 4.49 (s, br) 3.57 (s, 2H)
2.35 (s, 3H)
2i
b
4.61 (s, br), 4.48 (s, br) 3.16 (s, br, 2H)
3a 5.84 (d, 2.2), 5.58 (dd, 2.2) –
3b 6.29 (d, 5.7), 5.84 (d, 5.7) 2.86 (d, 22.3, 2H)
3c 5.54 (dd, 1.8, 3.7), 5.23
(dd, 4.1, 2.2)
3.57 (m, 2H)
a
Partially overlapped with water signal.
b

5% of the oxidized form
also present: 4.56 (s), 3.41 (s). These signals disappear upon addi-
tion of dithiothreitol.
Ó FEBS 2002 Inhibitors of phosphoenolpyruvate-utilizing enzymes (Eur. J. Biochem. 269) 3227
7 : 3, v/v) furnished ethyl dichloropyruvate: 1.8 g, 40%.
1
H-NMR (CDCl
3
) d:5.97(1H,s,CHCl
2
), 4.38 (2H, q,
J ¼ 7.0 Hz, CH
2
O),1.36(3H,t,J ¼ 7.0 Hz, CH
3
).
Synthesis of the enolphosphates 1c,d and 2a-i
Step (1): Perkow reaction. Ten millimoles of ethyl
dichloropyruvate (5a)orthea-haloketones (4a–e)was
added dropwise to a flask containing 10 mmol (1.18 mL) of
trimethyl phosphite (20 mmol for the preparation of 6d)at
0–10 °C. Violent bubbling took place in some cases. After
addition, the ice-bath was removed and the reaction was
allowed to proceed at the temperature indicated below until
31
P-NMR indicated the complete disappearance of trimeth-
yl phosphite (typically overnight). Small amounts of
trimethyl phosphite were eliminated in vacuo (0.1 mbar) at
room temperature. Details: 6a: reaction at room tempera-
ture, purification by FC (hexanes/ethyl acetate, 2 : 3, v/v),

1.8 g, 80%. 6b: reaction at room temperature, FC (hexanes/
ethyl acetate, 1 : 1, v/v), 1.45 g, 65%. 6c: prepared following
Cherbuliez et al. instructions [21], replacing the triethyl
phosphite for trimethyl phosphite, 2.2 g, 100%. 6d:pre-
pared from bromoacetyl bromide (4d,R
2
¼ Br, X ¼ Br),
1 h reaction at 60 °C, 2.54 g, 100% [22]. 6e: 2.0 g, 100%
[23]. 7a:reactionat70°C for 1 h. Purified by FC (hexanes/
ethyl acetate, 1 : 1, v/v): 1.5 g, 58% yield, 4 : 1 mixture of
the Z-andE-isomers.
Preparation of the enolphosphate dimethyl ester 6f. A
mixture of 6e (1 g, 5 mmol) and potassium thioacetate
(5.7 g, 5 mmol) was stirred into 5 mL of dimethylforma-
mide, at room temperature. The reaction mixture was
sonicated periodically. After 3 h, 10 mL of diethyl ether
were added, and the resulting suspension was passed
through a small path of silica gel, with hexanes/ethyl
acetate (1 : 1, v/v) as eluent. The product fractions were
collected and the solvent removed.
1
H-NMR revealed the
presence of around 20% starting material, together with
the desired product. To drive the reaction to completion
the mixture was subjected to two more reaction cycles,
adding consecutively 2 and 1 mmol of potassium thioac-
etate, until all starting material 6e had disappeared. 6f:
0.53 g, 45%.
Step (2): Removal of phosphate ester groups in 6a–f and
7a. The simple and mild demethylation procedure des-

cribed by McKenna et al. was employed [24]. Trimethyl-
silyl bromide (2 mmol, 0.27 mL) was slowly added to a
flask containing 1 mmol of compound 6a–f or 7a,kept
under argon at 0–4 °C. 4 mmol of trimethylsilyl bromide
were used in the reaction with compound 6d. The mixture
was stirred for 1 h and then for an additional 1 h at room
temperature. After evaporation of excess trimethylsilyl
bromide at high vacuum, 2 mmol of cyclohexylamine in
15 mL of methanol/ether (1 : 5, v/v) were added. The
white solid was collected by filtration and washed with
3 · 8mL of ether. 2a: dicyclohexylammonium salt,
0.28 g, 70%. 2c: dicyclohexylammonium salt, 0.35 g,
89%. 2e: dicyclohexylammonium salt, 0.38 g, 97%. 2f:
tricyclohexylammonium salt, 0.58 g, 61% [25,26]. 2g:
dicyclohexylammonium salt, 0.33 g, 89%. 2h:dic-
yclohexylammonium salt, 0.25 g, 62%. 8a: dicyclohexyl-
ammonium salt, 0.30 g, 71%.
Step (3): Hydrolysis of carboxylic acid ester
groups. Compounds 2b, 2d and 2i were prepared from
2a, 2c and 2h, respectively. Compound 2b was obtained by
addition of 5 molar equvalents of KOH (1
M
) to the residue
obtained after evaporation of excess trimethylsilyl bromide
in the previous step. Hydrolysis was allowed to proceed for
3–4 min. The aqueous solution was passed through a
DowexWX-8column(H
+
-form) and the acidic fractions
were pooled and neutralized with 2 mmol of cyclohexyl-

amine. The product was further purified by anion-exchange
chromatography, following the procedure described for the
separation of (Z)- and (E)-3-Cl-PEP (see below). It was
detected after the first chromatography step at 220 nm.
Fifty-milliliter fractions were collected and lyophilized after
the second chromatography, giving the tristriethylammoni-
um salt of 2b: 0.1 g, 20% yield.
With compounds 2d and 2i, the cyclohexylammonium
cations of 2c or 2h (1 mmol in 2–3 mL of deionized water)
were first exchanged against Na
+
by loading on a Dowex
XW-8 column (Na
+
form). The sodium salts were eluted
with 3 · 5 mL deionized water and adjusted to pH 12.0–
12.5 with 1
M
KOH. Around 3–4 mmol of KOH were
usually added before reaction completion (1–2 h). The
whole reaction volume was passed through the Doxew
XW-8 column (4 °C, H-form), the eluate neutralized with
cyclohexylamine (2 mmol, 0.23 mL) and then lyophilized.
2d: dicyclohexylammonium salt, 0.30 g, 86%. 2i:
dicyclohexylammonium salt, 0.28 g, 75%.
(Z)- and (E)-3-chlorophosphoenolpyruvate (1c,d). A
portion (1.2 g; 2.8 mmol) of the 4 : 1 mixture of isomers
8a was hydrolysed similarly to compounds 2c and 2h.The
solution was kept at pH 12.5 for 5 h and then neutralized
with 1

M
HCl (final pH value ¼ 6.0). The two isomers were
separated following the procedure of Poyner et al. with
modifications [27]. The mixture was diluted with 300 mL
deionized water and slowly loaded at 4 °C to a Sephadex
DEAE A-25 column (30 g, Cl
–
form), which was then eluted
with a KCl gradient (2 mLÆmin
)1
, 10 mL per fraction,
0.15
M
to 0.35
M
in 475 min). The compounds were detected
at 254 nm. Product 1c started to elute at 0.19
M
,whereas1d
appeared at 0.27
M
KCl. The corresponding fractions were
pooled and diluted three times with deionized water. They
were loaded on a second Sephadex DEAE A-25 column
(HCO
3
–
form) and eluted with 2 mLÆmin
)1
tryethylammo-

nium bicarbonate (0.2
M
to 1
M
in 475 min). The fractions
containing the product were pooled and lyophilized.
Analytical HPLC (DEAE-60-7, Macherey–Nagel, condi-
tions in legend to Fig. 1) revealed that the isolated
products were more than 99% pure. 1c: ditriethylam-
monium salt, 0.47 g, 42% [28]. 1d: ditriethylammonium salt,
0.11 g, 10%.
Chloromethylphosphonate 3c
Trimethylsilyl 2-trimethylsilyloxypropenoate (9)waspre-
pared as previously described [22]. Chloromethylphosphon-
ic acid dichloride (10 mmol, 1 mL) in 20 mL of dry benzene
was added dropwise to a flask containing 2.3 g (10 mmol)
of 9 at 50 °C. The reaction mixture was refluxed for 4 h.
Benzene was removed under vacuum, and the unstable
cyclic acylphosphate 10 was Kugelrohr distilled at around
130 °C (0.1 mbar): 0.4 g, 22%,
1
H-NMR (CDCl
3
), d:5.82
3228 L. F. Garcı
´
a-Alles and B. Erni (Eur. J. Biochem. 269) Ó FEBS 2002
(1H, dd, J ¼ 3.7, 1.8 Hz, CH
2
¼ C), 5.54 (1H, d,

J ¼ 3.7 Hz, CH
2
¼ C), 4.09 (2H, d, J ¼ 10.7 Hz, CH
2
Cl);
31
P-NMR (CDCl
3
) d: +26.2. The product 3c was obtained
after addition of 10 to 5 mL of ice-cold H
2
Oand
neutralization with 0.65 mL of cyclohexylamine. The solu-
tion was lyophilized and the product recovered by filtration
after triturating with 25 mL MeOH/ether, 1 : 4, v/v. 3c:
dicyclohexylammonium salt, 0.61 g, 15%.
Stability of PEP analogues 1–3
Most of PEP-derivatives 1–3 were stable over months
when stored as 250 m
M
solutions at pH ¼ 7.0–7.3 and
)20 °C. However, compounds 2b, 2d and 2i decomposed
under these conditions. Periodical inspection by
31
P-NMR revealed a continuous increase of the inorganic
phosphate signal (+1.96 p.p.m. at pH 7.1 in double-
distilled H
2
O).
Competitive inhibition enzyme assays

Unless otherwise indicated, all experiments were performed
at 30 °C, in 96-well microtitre plates. Progress curves were
recorded and the initial rates were calculated as the maximal
slope of the absorption curve obtained. IC
50
values were
measured using 0.1 m
M
PEP (0.1 m
MD
-2-phosphoglyceric
acid in the case of enolase) and in the presence of 0–5 m
M
inhibitor at the enzyme and metal concentrations indicated
below.
Enzyme I activity was measured by coupling the
formation of glucose-6-phosphate to its oxidation to
6-phosphoglucono-d-lactone. This process is catalysed by
D
-glucose-6-phosphate dehydrogenase and produces
NADPH, which can be monitored at 340 nm. The reaction
conditions were as described (150 lL per well) [19]:
0.02 l
M
enzyme I, 1 l
M
HPr, 20 l
M
IIA
Glc

,1lLof
membrane extract, 1 m
MD
-glucose, 0.1 units
D
-glucose-
6-phosphate dehydrogenase, 1 m
M
NADP
+
,50m
M
Hepes
pH ¼ 7.5, 2.5 m
M
dithiothreitol and 2.5 m
M
NaF. Either
5m
M
MgCl
2
or 1 m
M
MnCl
2
were also present.
Pyruvate kinase activity was determined in a coupled
assay with
L

-lactate dehydrogenase. The initial rates of
formation of pyruvic acid released from PEP were monit-
ored by the decrease of absorption at 340 nm due to NADH
consumption, as described previously [5,29]. The assays
were carried out in the presence of 0.015 UÆmL
)1
of
pyruvate kinase and 5 m
M
MgCl
2
or 0.05 UÆmL
)1
of
pyruvate kinase and 1 m
M
MnCl
2
.
PEP carboxylase activity was determined in a coupled
assay with malate dehydrogenase, as described previously
[30]. The rate of formation of oxalacetic acid was
calculated from the rate of disappearance of NADH.
Studies were conducted in the presence of 0.3 UÆmL
)1
of
PEP carboxylase and either 5 m
M
MgCl
2

or 1 m
M
MnCl
2
.
Enolase inhibition by 1b–d (0–200 l
M
), 1f and 2f (with
Mn
2+
) was directly monitored as the increase of absorption
at 235 nm due to the formation of the conjugated C–C
double bond of PEP from
D
-2-phosphoglyceric acid [29].
Reversible inhibition with the rest of compounds was
assayed by coupling PEP formation with NADH consump-
tioninthepresenceofpyruvatekinaseand
L
-lactate
dehydrogenase [27]. The experiments were carried out in the
presence of 0.04 UÆmL
)1
of enolase and 5 m
M
MgCl
2
or
0.15 UÆmL
)1

of enolase and 2 m
M
MnCl
2
.
Enzyme inactivation experiments
The time-dependent inactivation assays were carried out
under turnover conditions. The enzymes (5 l
M
enzyme I,
3UÆmL
)1
pyruvate kinase, 1.1 UÆmL
)1
PEP carboxylase or
2UÆmL
)1
enolase) were preincubated for 10 min at 30 °C
in the presence of enough of the rest of components to
maintain multiple turnovers (as indicated above). MgCl
2
(5 m
M
) was present during the incubation (also 0.5 m
M
MnCl
2
with PEP carboxylase and enolase) together with
0.5 m
M

of 1c,d,or5m
M
of 2g, 3a and 3c. Aliquots
(15–20 lL) were withdrawn at time intervals and diluted in
cold quenching buffer (285–130 lL) containing 1 m
M
PEP
or
D
-2-phosphoglyceric acid in the case of enolase. The
residual enzymatic activity was determined under the
conditions of the IC
50
assays, after addition of the enzyme
to a fresh mixture of the rest of components, 1 m
M
PEP or
D
-2-phosphoglyceric acid, and 5 m
M
MgCl
2
.
RESULTS
Preparation of the PEP-analogues 2a-i
The synthesis has been based in the Perkow reaction
(Scheme 2) [31]. The commercially available a-haloketones
4a–e were reacted with trimethyl phosphite, giving the
enolphosphate dimethyl esters 6a–e,inmostcasesin
quantitative yields. The thioester 6f was prepared from the

1-chloromethyl-vinyl derivative 6e, by nucleophilic displace-
ment with potassium thioacetate. Subsequent replacement
of the phosphate methyl ester for trimethylsilyl groups, by
treating with trimethylsilyl bromide [24], and final
methanolysis furnished 2a–i. These compounds were
purified by precipitating their cyclohexylammonium salts.
All attempts to synthesize 2b and 2d from the haloketones
4f (R
2
¼ CH
2
CO
2
H, X ¼ Cl) and 4g (R
2
¼ CH
2
OH,
X ¼ Cl), obtained after enzymatic hydrolysis of 4a and
4b, were unsuccessful. [Note that 4a (5 mmol) was hydro-
lysed with the lipase B from Candida antarctica (0.5 g) after
2hat37°C in water-saturated
t
BuOMe (50 mL). White
needles of 4-chloro-3-oxo-butyric acid (4f)formed(62%
yield) after removal of the enzyme by filtration, evaporation
of the solvent and recrystallization from hexanes/MeOH,
4 : 1, v/v. 4b was hydrolyzed under the same conditions in
the presence of LypozymeÒ. 1-Chloro-3-hydroxy-propan-
2-one (4g)wasobtainedin74%yieldafterFCwith

hexanes/AcOEt, 3 : 2, v/v]. The free carboxylate and
hydroxyl groups probably promote nucleophilic displace-
ments on the postulated phosphonium intermediate of the
Perkow reaction [31], thereby precluding the elimination of
methyl chloride. This course of the reaction is indicated by
the isolation of product 6a (R
2
¼ CH
2
CO
2
Me) from the
reaction between 4f and trimethyl phosphite. Therefore 2b
and 2d,aswellas2i were prepared by alkaline hydrolysis of
the esters 2a, 2c and 2h, respectively. However, 2a was
stable to hydrolysis at pH 12 and the reaction had to be
carried out under more harsh conditions (1
M
KOH). As a
consequence, small amounts of side-products were formed,
as shown by
1
H-NMR, and 2b had to be purified by anion-
exchange chromatography.
Ó FEBS 2002 Inhibitors of phosphoenolpyruvate-utilizing enzymes (Eur. J. Biochem. 269) 3229
Synthesis of potential irreversible inhibitors
Only a few irreversible inhibitors of PEP-utilizing enzymes
are described in the literature. Two examples are the
antibiotic fosfomycin [(1R,2S)-1,2-epoxypropylphosphonic
acid], which targets UDP-GlcNAc enolpyruvyl transferase

[32], and (Z)-3-bromo-phosphoenolpyruvate [(Z)-Br-PEP],
employed as a mechanism-based inhibitor of pyruvate
kinase [33], pyruvate phosphate dikinase [5], and PEP
carboxylase [30]. We present here the preparation of four
candidates for the irreversible inhibition of PEP-utilizing
enzymes.
The enolphosphates 1c,d,and2g can be considered as
potential suicide inhibitors. They are nonreactive molecules
but are transformed by enzyme-catalysed dephosphorylation
into enolates, which in turn by protonation/tautomerization
are converted to 3-chloropyruvic acid (in the case of 1c,d)or
chloroacetone (from 2g). These a-halocarbonyl compounds
can then react with nucleophilic amino-acid residues [34,35].
Because they will be generated in the active site of the protein,
the probability of labelling catalytically relevant residues,
therefore inactivating the enzyme, is increased.
These compounds were also synthesized via the Perkow
reaction (Scheme 2). Ethyl 3,3-dichloropyruvate (5a)or
commercially available 1,3-dichloroacetone (4e) were reac-
ted with trimethyl phosphite. The dimethyl enolphosphates
7a and 6e were obtained in excellent yields. The preparation
of 6e by this route had been reported previously [23]. On the
other hand, the synthesis of the isomeric mixture 7a
resembles the procedure proposed by Liu et al.forthe
preparation of pure (Z)-3-chlorophosphoenolpyruvate (1c)
from 3,3-dichloropyruvic acid (5b,R
2
¼ CO
2
H, X ¼ Cl)

[28]. We instead decided to use the ethyl ester 5a, because it
afforded a 1 : 4 mixture of the (E)- and (Z)-isomers 7a,
therefore allowing the simultaneous preparation of the two
isomers 1c and 1d. Besides, in our hands, the compound 1c
obtained following the described procedure was contamin-
ated with around 5% PEP, which could not be removed.
This contamination probably derives from the presence of
small amounts of 3-chloropyruvic acid mixed with the
3,3-dichloropyruvic acid prepared following the reported
procedure.
The derivative 2g and the ethyl esters 8a (Z/E mixture)
were obtained after treatment of 6e and 7a with trimeth-
ylsilyl bromide and methanolysis. Finally, the ethyl ester
group of 8a was hydrolysed under basic conditions, and the
Z-andE-isomers were separated by anion-exchange
chromatography. Compounds 1c and 1d could be obtained
in this way at the same time and in higher than 99% purity
(Fig. 1A,B).
The analogue 3c carries a chloromethylphosphonate
group instead of the phosphate present in PEP. This
functionality can react with nucleophiles located in the
active-site of PEP-utilizing proteins. 3c might be particularly
suited to label residues which are transiently phosphorylated
in the course of the catalytic cycle, for instance, the active-
site histidines of enzyme I of the PTS [2], phosphoenolpyru-
vate synthase, and pyruvate phosphate dikinase, or the
presumed active-site aspartic acid residue of phos-
phoenolpyruvate mutase [36].
The synthesis of the chloromethylphosphonate 3c was
accomplishedasdepictedinScheme3.Thestrategywas

Scheme 2.
Scheme 3.
Fig. 1. Stereochemical assignment of (Z)-3-Cl-PEP (left) and (E)-3-Cl-
PEP (right). (A,B) Analytical anion-exchange HPLC of purified 1c
and 1d. Chromatographic analysis was carried out in a DEAE-60-7
column [1 mLÆmin
)1
,20m
M
KH
2
PO
4
,pH¼ 6.0, KCl (0 m
M
for
2minto360m
M
in 16 min)]. The effluent was monitored at 240 nm.
Retention times for each isomer are indicated. (C,D)
1
H-decoupled
13
C-NMR spectra of 1c and 1d (ditriethylammonium salts) in CD
3
OD.
Only the carboxylate carbon region is shown. (E,F)
13
C-NMR spectra
in CD

3
OD without decoupling.
3230 L. F. Garcı
´
a-Alles and B. Erni (Eur. J. Biochem. 269) Ó FEBS 2002
based on the formation of the mixed cyclic anhydride 10,
which was expected to be readily hydrolysable to furnish 3c.
Similar five-membered ring phosphates are known to be
exceptionally susceptible to nucleophilic attacks [37], and
have been used as strong phosphorylating agents. An
analogous cyclic acyl phosphate is probably formed during
the intramolecular carboxylate-catalysed hydrolysis of PEP
phosphate esters [38], and was also proposed as an
explanation to the
18
O distribution pattern observed when
PEPisheatedinacidicH
18
2
O[39].
Compound 10 was prepared by a method used for the
preparation of similar structures [40]. The moisture-sensitive
trimethylsilyl 2-trimethylsilyloxypropenoate (9) prepared
from pyruvic acid [22], was reacted with chloromethylphos-
phonic acid dichloride.
31
P-NMR of the crude reaction
mixture revealed three major signals appearing upfield of the
chloromethylphosphonic acid signal. The cyclic acyl phos-
phate 10 could be isolated by distillation and was partially

characterized by
1
H- and
31
P-NMR, in spite of its instability.
As expected, hydrolysis of 10 produced compound 3c.
Characterization of PEP analogues
Compounds 1–3 were characterized by
1
H (see Table 1),
13
C- and
31
P-NMR and mass spectrometry. The stereo-
chemistry of compounds 1c and 1d could not be established
by comparison with published information, which disagree
in this respect. The major product of the Perkow reaction
between 3,3-dichloropyruvic acid and trimethyl phosphite
was first reported by Liu et al.tobetheZ-isomer [28].
However, Poyner et al. indicated that the dominant product
of the same reaction was the E-isomer [27]. In view of this
contradiction the NMR coupling constant (
3
J
HC
)between
the carboxylic carbon atom and the vinyl proton (HOO
13
C-
C ¼ C-

1
H) for each isomer has now been measured. It is
known that, without exception, the coupling constant
between two nuclei substituted directly on the carbons of
a carbon–carbon double bond is stronger when they are in
the trans rather than the cis orientation [41].
3
J
HC
was
determined in two steps: (a) the phosphorus-carbon coup-
ling constant (HOO
13
C-C-O-
31
P) was measured in the
13
C-
NMR
1
H spin decoupled spectrum of the pure isomer
(Fig. 1C,D); and (b) the additional coupling, due to the
vinyl proton, was obtained from the coupled spectra
(Fig. 1E,F). The compound with the strongest
1
H-
13
C
coupling constant (
3

J
HC
¼ 7.2 Hz,
3
J
PC
¼ 5.7 Hz) was
assigned as the E-isomer and the compound with the
weakest coupling (
3
J
HC
¼ 1.6 Hz,
3
J
PC
¼ 1.5 Hz) was
assigned as the Z-isomer. Therefore (E)-Cl-PEP is the
compound presenting the vinyl proton signal (6.17 p.p.m.,
D
2
O, pH ¼ 7.0) upfield from the vinyl signal of the
Z-isomer (6.72 p.p.m., D
2
O, pH ¼ 7.1), in agreement with
Liu et al.[28].
PEP analogues as reversible inhibitors of PEP-utilizing
enzymes
The compounds presented in Scheme 1 and phospho-
D

,
L
-lactic acid (1f) were screened as inhibitors of (a) enzyme I
of the PTS from E. coli; (b) rabbit muscle pyruvate kinase;
(c) maize PEP carboxylase; and (d) yeast enolase. Because
the selection of the metal cofactor required by many PEP-
dependent enzymes has been reported to influence the
inhibition results in some cases, the assays have been
performed with Mg
2+
-andMn
2+
-activated enzymes.
Derivatives 1b, 1c, 1e, 1f, 2f, 3a and 3b were used previously
to study some of these enzymes. They have been included in
the present work, for comparison, and to complete the data
for the four enzymes. However, due to the number of assays
to be performed IC
50
values were calculated. The results are
presented in Table 2. The inhibition type and the value of
the inhibition constant (K
i
) have been determined only in
some representative cases.
Inhibition of enzyme I. The bacterial PTS catalyses uptake
with simultaneous phosphorylation of the carbohydrates [2].
The PTS is a group transfer pathway: a phosphoryl group
derived from PEP is transferred sequentially along a series of
proteins to the sugar molecule. Enzyme I is the protein at the

top of this system. It transfers the phosphoryl group from
PEP to a phosphocarrier protein, HPr. (Z)-phosphoenol-
butyrate (1e) is the only analogue that has been used to
study this enzyme. It was employed to establish the
stereochemistry of protonation of the released enolate [8].
Before measuring inhibition, compounds 1–3 were
checked as phosphoryl donors to enzyme I in a glucose
phosphotransferase assay. Glucose 6-phosphate formation
occurred with compounds 1b–e, but with one to three orders
of magnitude lower catalytic rates than with PEP (data not
shown). All compounds were then tested as competitive
inhibitors with respect to PEP. As shown in Table 2, only
the Z-isomers of 3-F-PEP (1b)and3-Cl-PEP(1c)weakly
inhibit enzyme I. Inhibition by these compounds is com-
petitive and the same K
i
¼ 0.4 m
M
was calculated for the
two compounds [Fig. 2A, only shown for (Z)-3-F-PEP].
Inhibition by analogues with a hydroxymethylene (2d)ora
phosphonate group (2f) instead of the carboxylate of PEP,
and by
D
,
L
-phospholactic acid (1f) was weak. In contrast to
the rest of enzymes shown in Table 2, no significant
differences were observed when changing the metal present.
Interestingly, compound 3b, in which only difference to PEP

is the replacement of the phosphate-bridging oxygen by a
CH
2
moiety, is completely inactive, suggesting that this
oxygen participates in hydrogen bonds with the enzyme or
in the coordination to the metal cofactor. Similar results
have been obtained with compound 3b as inhibitor of PEP
mutase [36], and pyruvate kinase [6].
Inhibition of pyruvate kinase. This enzyme catalyses the
regeneration of ATP from ADP and PEP, in the last step of
glycolysis. Due to its physiological relevance, pyruvate
kinase is one of the best studied enzymes and many PEP
analogues have been used with it [5–7,16,18,25,28,29,33,42].
Compounds 1–3 were tested as inhibitors of the reaction
betweenADPandPEPcatalysedbypyruvatekinase.
Pyruvic acid is one of the products of this reaction.
Therefore, activity was measured by coupling the formation
of pyruvate with its NADH-dependent reduction to
L
-lactate, a process catalysed by
L
-lactate dehydrogenase.
Compounds 1b,c potently inhibit phosphotransfer from
PEP to ADP, in accordance with their published inhibition
constants (K
i
): 57 n
M
for 1b [5], and 39 n
M

for 1c [28].
Strikingly, however, E-Cl-PEP 1d is 2400-fold less inhibitory
than its Z-isomer 1c.
In a general sense, modification of the phosphate group
or the carboxylate function is counterproductive for binding
Ó FEBS 2002 Inhibitors of phosphoenolpyruvate-utilizing enzymes (Eur. J. Biochem. 269) 3231
to pyruvate kinase (Table 2). Nevertheless, a remarkable
dependence of inhibition on the metal employed is observed.
Compound 2f had been described previously as not
interacting with Mg
2+
-activated pyruvate kinase [25]. We
have found, however, that this compound becomes a strong
inhibitor when Mn
2+
is present. Under such conditions, 2f
inhibits competitively pyruvate kinase with a K
i
¼ 80 l
M
(Fig. 2B). This observation might be of relevance, for
instance, in efforts intended to cocrystallise pyruvate kinase
with a nonreactive PEP analogue and ADP. Similar metal
dependence is observed with the racemic mixture of
compounds 1f and to a lesser extent with sulfoenolpyruvate,
3a, in agreement with published data [18].
Inhibition of PEP carboxylase. This enzyme catalyses the
addition of bicarbonate to PEP to produce oxalacetate and
inorganic phosphate [3]. PEP carboxylase is widely distri-
buted in plants. It is particularly important in C

4
plants,
where it concentrates CO
2
before it enters the Calvin cycle.
Inhibition was studied by measuring the rate of oxalac-
etate formation from PEP in the presence of increasing
concentrations of compounds 1–3. Activity was detected in
a coupled assay with NADH/malate dehydrogenase. All
compounds were first checked as pseudosubstrates, in order
to verify incompatibilities with the inhibition studies. The
activity detected with the known substrates of PEP
carboxylase 1b and 1c was very low [15,28,30]. With the
rest of compounds no activity could be detected, indicating
that either they are not substrates of PEP carboxylase, or the
products formed are not substrates of malate dehydroge-
nase. PEP carboxylase inhibition was then measured. Again
the most potent inhibitors with respect to PEP were those
modified at C-3. Measured IC
50
values are well correlated
with the reported K
i
:85l
M
for 1b [30], 63 l
M
for 1c [28],
and 18 l
M

for 1e [43]. The results also highlight a common
feature of PEP-utilizing enzymes: the preference for Z-over
E-isomers (compare IC
50
values measured for 1c and 1d).
In presence of Mg
2+
most of the PEP-derivatives with
modifications of the carboxylic position do not inhibit and
instead stimulate PEP carboxylase activity. The effect is
more pronounced in the presence of the trifluoromethyl and
chloromethyl analogues 2e and 2g. Besides, only com-
pounds presenting neutral groups instead of the carboxylate
function of PEP are stimulatory. In contrast, none of the
compounds that present negatively charged groups at this
position, namely 2b or 2f, induce an increase in the activity.
It is important to point out that the inhibition studies have
been carried out at pH 7.5. Under similar conditions the
compound 2f has previously been described to be a
competitive inhibitor, with a K
i
of 2.2 m
M
[25].
Similarly to pyruvate kinase, metal ion plays an import-
ant role in PEP carboxylase inhibition. Several compounds,
e.g. phospholactate (1f), become inhibitors in the presence
of Mn
2+
. The inhibition constant for the

L
-isomer of
Table 2. Half-inhibitory concentrations (IC
50
) and half inactivation times (t
50
) of PEP-utilizing enzymes with analogues 1–3. IC
50
values (given in m
M
)
were obtained using 0.1 m
M
PEP, in the presence of 5 m
M
MgCl
2
or 1 m
M
MnCl
2
. t
50
values (given in min) were measured at 30 °C with 0.5 m
M
1c,d or 5 m
M
2g, 3a and 3c, in the presence of 5 m
M
MgCl

2
. K
m
values for PEP (m
M
) are indicated; these values were taken from the indicated
references. NM, not measured; ND, no time dependent inactivation detected.
Comp
Enzyme I PKase PEPCase Enolase
IC
50
IC
50
IC
50
IC
50
b
Mg
2+
Mn
2+
t
50
Mg
2+
Mn
2+
t
50

Mg
2+
Mn
2+
t
50
a
Mg
2+
Mn
2+
t
50
a
PEP 0.2 [52] – 0.03 [5] 0.02 [5] 0.8 [28] 0.3 [28] 0.05 [5] 0.08 [5]
Modified in vinyl region
1b 0.8 NM 1 · 10
)4
NM 0.07 NM 3 · 10
-4
NM
1c 0.9 NM 0.7 5 · 10
)5
NM ND 0.02 NM ND 0.25 NM ND
1d >10 NM  60 0.12 NM ND 0.4 NM ND 0.25 NM ND
1e >10 NM 0.15 NM 0.04 NM 0.04 NM
1f >10 5  7 0.16 0.5 0.08 1.9 5
Modification of the carboxylic group
2a >10 >10 >10 >10 25%
c

>10 >10 >10
2b >10 >10 >10 >10 >10 0.9 1.2  8
2c >10 >10 >10 >10 >10 >10 >10 >10
2d >10  6 >10 >10 35%
c
2 >10 >10
2e >10 >10 >10 >10 55%
c
>10 >10 >10
2f  8>10  9 0.03 5 3 1.4 0.3
2g >10 >10 ND >10 >10 ND 60%
c
28%
c
240
d
>10 >10 ND
2h >10 >10 >10 >10 >10 0.8 >10 >10
2i >10 >10 >10  9 >10 1.2 >10 >10
Modified in the phosphate position
3a >10 >10 ND 5 1.2 >10 >10 ND >10 >10
3b >10 >10  6 >10 0.2 0.1 0.3 9 · 10
)3
3c >10 >10 ND >10 >10 ND  9 3 ND >10 >10 ND
a
Incubation also in the presence of 0.5 m
M
MnCl
2
.

b
Using 0.1 m
MD
-2-phosphoglyceric acid, in the presence of 5 m
M
MgCl
2
or 2 m
M
MnCl
2
.
c
Enhancement of the activity observed upon addition of the compound. The percentage of increase of activity achieved with 5 m
M
of compound is indicated.
d
In presence of 5 m
M
MgCl
2
. Extrapolated from inhibition observed after 2 h incubation.
3232 L. F. Garcı
´
a-Alles and B. Erni (Eur. J. Biochem. 269) Ó FEBS 2002
phospholactate was reported to shift from 100 to 1 l
M
when
Mg
2+

was replaced by Mn
2+
[18]. Five compounds, among
those modified at the carboxylic position, inhibit moderately
Mn
2+
-activated PEP carboxylase. In presence of this metal,
inhibition by compounds 2b (not shown), 2h (not shown)
and 3c (Fig. 2C) are competitive with K
i
values of 1.1, 1.2
and 3.9 m
M
, respectively. Thus, PEP carboxylase is the only
enzyme shown in Table 2 that is able to interact with the
chloromethyl phosphonate 3c, arguing in favour of its
structural tolerance. A second effect related to the metal
present is that only the compound 2g is now able to
stimulate the activity in the presence of Mn
2+
.
Inhibition of enolase. Enolase is a glycolytic enzyme that
catalyses the reversible elimination of water from
2-phospho-
D
-glycerate to form PEP. The reaction proceeds
with anti stereochemistry [44], via a carbanion (enolate)
intermediate generated by abstraction of the C-2 proton of
D
-2-phosphoglyceric acid [27]. The reaction is nearly

isoenergetic [45]. Several PEP-analogues have been
employed in the study of enolase [5–7,16,46,47]. Two
compounds (Z)-3-F-PEP (1b)anda-(dihydroxyphosphinyl-
methyl)acrylate (3b) function as alternative substrates [6]. In
the case of (Z)-3-F-PEP the product of the reaction is the
enol of tartronate semialdehyde phosphate, a potent
reversible inhibitor of the enzyme. This compound is
formed after OH
–
attack at C-3, followed of F
–
elimination.
Enolase also catalyses the formation of tartronate semial-
dehyde phosphate from (Z)-3-Cl-PEP (1c) [27]. Neverthe-
less, these compounds display a far lower catalytic efficiency
than PEP, and for that reason it is still possible to study
them as reversible inhibitors.
Inhibition of enolase was followed in two ways. In most
of cases the formation of PEP from
D
-2-phosphoglyceric
acid was coupled to the pyruvate kinase/
L
-lactate dehy-
drogenase assay. Obviously, this methodology was not
applied with compounds 1b–e (also 1f and 2f when Mn
2+
was present) as they are good inhibitors of pyruvate kinase.
In these cases the formation of PEP was directly monitored
at 235 nm, in the presence of variable amounts of the

PEP-analogues.
Compound 1b strongly inhibited enolase. Other com-
pounds, such as 1c–e and 3b were good competitors
compared to
D
-2-phosphoglyceric acid. Unlike the other
enzymes assayed, enolase did not discriminate between the
(Z)- and the (E)-3-Cl-PEP isomers. Phospholactate 1f,the
carboxymethyl analogue 2b and the phosphonate 2f
displayed moderate inhibitory potencies. The last analogue
has been described to competitively inhibit enolase, with a
K
i
of 2.2 m
M
[25]. The same type of inhibition and similar
inhibition constant value was measured for 2b in the
presence of both Mg
2+
(K
i
¼ 2.0 m
M
, not shown) and
Mn
2+
(K
i
¼ 2.2 m
M

, Fig. 2D). Preserving the negative
charge of the carboxylate group of PEP seems to be essential
for recognition by enolase. It is likely that such a negative
charge is required for binding to one of the two divalent ions
present in the active site of enolase [48]. Finally, in the case
of enolase the metal selected did not affect the inhibition
results as markedly. Only with the compound 3b a strong
enhancement of inhibition was observed in the presence of
Mn
2+
. Under such conditions this analog noncompetitively
inhibits enolase with a K
i
of 6 l
M
(not shown).
Enzyme inactivation studies
To screen for irreversible/suicide inhibition, the target
enzymes were first incubated at 30 °C under turnover
conditions with the PEP analogues 1c,d, 2g and 3c in the
presence of Mg
2+
, and were then assayed for residual
catalytic activity with their natural substrates. Inactivation
of PEP carboxylase and enolase was also studied in the
presence of Mn
2+
. Enzyme I and PEP carboxylase were
also treated with 3a, as this compound might transfer the
sulfuryl group to a catalytic residue, thereby blocking the

enzyme.
Incubation of enzyme I with (Z)-3-Cl-PEP (1c)resulted
in a fast time-dependent inactivation (Fig. 3). The time to
half-inactivate the enzyme (t
1/2
) was 0.7 min. Multiple
turnover conditions were required, highlighting the suicidal
nature of this inhibitor. The E-isomer promoted a much
slower irreversible inhibition (t
1/2
¼ 60 min). No inactiva-
tion was detected with the rest of derivatives (Table 2).
Details on the enzyme I/1c interaction have been recently
reported [19].
Fig. 2. Lineweaver–Burk plots of inhibition of PEP-utilizing enzymes in
the presence of PEP analogs 1–3. (A) Inhibition of Mg
2+
-activated
enzyme I by 0 m
M
(squares), 0.12 m
M
(circles), 0.36 m
M
(triangles)
and 1.08 m
M
(stars) (Z)-3-F-PEP (1b). (B) Inhibition of Mn
2+
-acti-

vated pyruvate kinase by 0 l
M
(squares), 11 l
M
(circles), 33 l
M
(tri-
angles) and 100 l
M
(stars) compound 2f. (C) Inhibition of Mn
2+
-
activated PEP carboxylase in the presence of 0 m
M
(squares), 1 m
M
(circles), 3 m
M
(triangles) and 9 m
M
(stars) compound 3c. (D) Inhi-
bition of Mn
2+
-activated enolase by 0 m
M
(squares), 0.33 m
M
(cir-
cles), 1 m
M

(triangles) and 3 m
M
(stars) analog 2b. PEP concentrations
(
D
-2-phosphoglyceric acid with enolase) were varied between 0 and
2m
M
in all cases except for (B) (0–0.5 m
M
). Other conditions were as
indicated under Materials and methods for the calculation of IC
50
values. The same data, plotted in the Michaelis–Menten form were
used to derive the K
app
m
for PEP or
D
-2-phosphoglyceric acid at each
inhibitor concentration. K
i
values were then calculated by linear
regression of K
app
m
values vs. inhibitor concentration [I], according to
the equation K
app
m

¼ K
m
(1 + [I]/K
i
).
Ó FEBS 2002 Inhibitors of phosphoenolpyruvate-utilizing enzymes (Eur. J. Biochem. 269) 3233
Pyruvate kinase was not irreversibly inhibited by any of
the PEP analogues 1c,d, 2g or the chloromethylphospho-
nate 3c. Incubations were prolonged for up to 2 h at 30 °C
without significant effect. Slow inactivation of PEP
carboxylase was induced by compound 2g (25% of activity
loss after 2 h). Incubation with the 3-Cl-PEP isomers 1c and
1d for the same time inactivated PEP carboxylase by less
than 10%. It was therefore not possible to reproduce the
results obtained when the enzyme was incubated at 25 °C
with the analogue 1c inthepresenceofMn
2+
(reported
t
1/2
¼ 5 h) [28]. In the case of enolase no inactivation was
observed with compound 3c. Compounds 1c,d and 2g were
also tested, in spite of the fact that they were not expected to
behave as mechanism-based inhibitors, as the enzyme does
not catalyse dephosphorylation reactions. They were not
inhibitory.
DISCUSSION
The data presented in this work, in combination with
multiple studies presented previously with these and other
PEP-utilizing enzymes, indicate that the most active

analogues of PEP are those differing by substitutions at
C-3 (1b–e). Modifications of the phosphate and the
carboxylate groups resulted in inactive compounds, in
general terms. These two anionic centres probably contrib-
ute to the chelation of the metal required for binding to
these enzymes; disruption of one contact may suffice to
abolish binding. This is the most likely explanation of the
fact that, in a general sense, the best inhibitors among
compounds 2a–i were those preserving a negative charge in
the modified position, namely compounds 2b and 2f.The
preference of these enzymes for the Z- (like 1c)overthe
E-steroisomer (1d) is also a common feature. Only enolase
did not discern between the two isomers.
According to the data presented in Table 2, enzyme I of
the PTS imposes the most stringent geometrical restrictions
to its substrate. This protein is thought to be active only as a
dimer, and its dimerization is induced by the PEP molecule
[49]. Therefore, this process might become an alternative
checkpoint for substrate binding, and reduce the chances of
finding good inhibitors for this enzyme. In fact, it might be
possible that some of the screened compounds bind to the
dimer without displaying inhibition because they cannot
compete with PEP during the dimerization step. Note that
the inhibition assays were carried out at low enzyme I
concentrations, where protein association plays an import-
ant role. Further experiments will be carried out under
conditions where enzyme I is predominantly a dimer in
order to clarify this possibility.
Pyruvate kinase is an intensively studied enzyme. Most of
the data presented in this work for this enzyme corroborates

previous results. However, some observations are of special
interest. For instance, the remarkable difference observed
between the Z-andE-isomers of 3-Cl-PEP as inhibitors: the
Z-isomer is three to four orders of magnitude stronger than
the E-isomer. A much smaller difference has been observed
between the two isomers of phosphoenolbutyrate, a com-
pound presenting a more voluminous substitution than
chlorine: 7.1 l
M
K
i
for (Z)-phosphoenolbutyrate (1e)vs.
49.5 l
M
for its E-isomer [29]. Consequently, the data
measured with the 3-Cl-PEP isomers cannot be justified
on the basis of steric arguments. Other electronic factors
must contribute differently with each isomer, to establish the
interactions with the enzyme.
From the enzymes studied in this work, PEP carboxylase
has been found to be the most tolerant to modifications on
the structure of PEP. In concrete, it is interesting to call the
attention to the results obtained with 2b, 2h and 2i.The
carboxymethyl, acetylsufanylmethyl and mercaptomethyl
functions of these analogues are either considerably bulkier
than the carboxylate group of PEP or electronically very
different. Therefore they are not expected to occupy the
pocket that PEP carboxylase uses for the carboxylate group
of PEP. A bicarbonate binding pocket is also present in this
enzyme. Consequently, these derivatives might exert inhibi-

tion by adopting an alternative orientation in the active site,
similar to that shown in Fig. 4. A second possibility is that
the modified group is embedded into a hydrophobic pocket
Fig. 4. Alternative binding mode of inhibitors to PEP carboxylase. (A)
Schematic representation of PEP and bicarbonate bound into the
active site. (B) Suggested alternative binding mode for compounds 2b
(shown), 2h and 2i.
Fig. 3. Irreversible inhibition of enzyme I. Inactivation 30 °C induced
by 0.5 m
M
(Z)-3-Cl-PEP (h), 0.5 m
M
(E)-3-Cl-PEP (s), 5 m
M
2g (n),
5m
M
3a ()), 5 m
M
3c (I) and no inhibitor (j). The incubation mix-
ture contained 5 m
M
MgCl
2
,1m
MD
-glucose and catalytic amounts of
the rest of components of the PTS necessary to maintain multiple
turnovers.
3234 L. F. Garcı

´
a-Alles and B. Erni (Eur. J. Biochem. 269) Ó FEBS 2002
that is known to be present in close proximity to the
methylene group of PEP [50].
Interestingly, some of the PEP derivatives modified at the
carboxylic position did not inhibit but instead stimulated
the activity of PEP carboxylase, especially in the presence of
Mg
2+
. This effect might be due to binding of these
compounds to the glucose-6-phosphate allosteric site of
the enzyme, similarly to what has been observed with
fosfomycin by Mu´ jica-Jime
´
nez et al.[51].Inthatstudy,the
metal-free form of fosfomycin was proposed to compete
with free PEP for the enzyme’s allosteric site. Similarly,
complex formation with Mg
2+
is probably precluded in the
stimulatory compounds 2a,d,e,g, because neutral chemical
functions are replacing the coordinating carboxylate group
of PEP in these molecules. In agreement with this propo-
sition, the compounds presenting negatively charged groups
in that position, 2b and 2f, did not enhance PEP carboxylase
activity.
ACKNOWLEDGEMENTS
We are indebted to the Swiss National Science Foundation (grant
31–45838.95) and the Secretarı
´

a de Estado de Educacio
´
nyUniversi-
dades (Spain) for financial support. Special thanks to Prof Ronald L.
Somerville (Purdue University) for his kind donation of (Z)-3-F-PEP
and (Z)-3-Me-PEP. We also thank Dr Eloy Arenas Bernal for assisting
us with the HPLC work, and Prof Peter Bigler (University of Bern) for
his helpful advice with NMR experiments.
REFERENCES
1. Walsh, C.T., Benson, T.E., Kim, D.H. & Lees, W.J. (1996) The
versatility of phosphoenolpyruvateanditsvinyletherproductsin
biosynthesis. Chem. Biol. 3, 83–91.
2. Postma, P.W., Lengeler, J.W. & Jacobson, G.R. (1993) phos-
phoenolpyruvate:carbohydrate phosphotransferase systems of
bacteria. Microbiol. Rev. 57, 543–594.
3. O’Leary, M.H. (1982) Phosphoenolpyruvate carboxylase: an
enzymologist’s view. Annu. Rev. Plant Physiol. 33, 297–315.
4. Seidel, H.M., Freeman, S., Seto, H. & Knowles, J.R. (1988)
Phosphonate biosynthesis: isolation of the enzyme responsible for
the formation of a carbon–phosphorus bond. Nature 335, 457–
458.
5. Duffy, T.H. & Nowak, T. (1984) Stereoselectivity of interaction of
phosphoenolpyruvate analogues with various phosphoenol-
pyruvate-utilizing enzymes. Biochemistry 23, 661–670.
6. Stubbe, J.A. & Kenyon, G.L. (1972) Analogs of phosphoe-
nolpyruvate. Substrate specificities of enolase and pyruvate kinase
from rabbit muscle. Biochemistry 11, 338–345.
7. Wirsching, P. & O’Leary, M.H. (1985) (E)-3-Cyanophos-
phoenolpyruvate, a new inhibitor of phosphoenolpyruvate-
dependent enzymes. Biochemistry 24, 7602–7606.

8. Hoving, H., Nowak, T. & Robillard, G.T. (1983) Escherichia coli
phosphoenolpyruvate-dependent phosphotransferase system:
stereospecificity of proton transfer in the phosphorylation of
enzyme I from (Z)-phosphoenolbutyrate. Biochemistry 22,
2832–2838.
9. Kim, D.H., Tucker-Kellogg, G.W., Lees, W.J. & Walsh, C.T.
(1996) Analysis of fluoromethyl group chirality establishes a
common stereochemical course for the enolpyruvyl transfers cat-
alyzed by EPSP synthase and UDP-GlcNAc enolpyruvyl trans-
ferase. Biochemistry 35, 5435–5440.
10. Sundaram, A.K. & Woodard, R.W. (2000) Probing the Stereo-
chemistry of E. coli 3-deoxy-
D
-arabino-heptulosonate 7-phosphate
synthase (phenylalanine-sensitive)-catalyzed synthesis of KDO
8-P analogues. J. Org. Chem. 65, 5891–5897.
11. Adlersberg, M., Dayan, J., Bondinell, W.E. & Sprinson, D.B.
(1977) Stereochemical studies of the pyruvate kinase reaction with
(Z)- and (E)-phosphoenol-alpha-ketobutyrate. Biochemistry 16,
4382–4387.
12. Dotson, G.D., Nanjappan, P., Reily, M.D. & Woodard, R.W.
(1993) Stereochemistry of 3-deoxyoctulosonate 8-phosphate syn-
thase. Biochemistry 32, 12392–12397.
13.Reed,G.H.,Poyner,R.R.,Larsen,T.M.,Wedekind,J.E.&
Rayment, I. (1996) Structural and mechanistic studies of enolase.
Curr. Opin. Struct. Biol. 6, 736–743.
14. Hwang, S.H. & Nowak, T. (1986) Stereochemistry of phosphoe-
nolpyruvate carboxylation catalyzed by phosphoenolpyruvate
carboxykinase. Biochemistry 25, 5590–5595.
15. Janc, J.W., Urbauer, J.L., O’Leary, M.H. & Cleland, W.W. (1992)

Mechanistic studies of phosphoenolpyruvate carboxylase from
Zea mays with (Z)- and (E)-3-fluorophosphoenolpyruvate as
substrates. Biochemistry 31, 6432–6440.
16. Sikkema, K.D. & O’Leary, M.H. (1988) Synthesis and study of
phosphoenolthiopyruvate. Biochemistry 27, 1342–1347.
17. Hansen, D.E. & Knowles, J.R. (1982) The stereochemical course
at phosphorus of the reaction catalyzed by phosphoenolpyruvate
carboxylase. J. Biol. Chem. 257, 14795–14798.
18. Peliska, J.A. & O’Leary, M.H. (1989) Sulfuryl transfer catalyzed
by pyruvate kinase. Biochemistry 28, 1604–1611.
19.Garcia-Alles,L.F.,Flukiger,K.,Hewel,J.,Gutknecht,R.,
Siebold, C., Schurch, S. & Erni, B. (2002) Mechanism-based
inhibition of enzyme I of the Escherichia coli phosphotransferase
system: Cys502 is an essential residue. J. Biol. Chem. 277,
6934–6942.
20. Nishiyama, K. & Inouye, Y. (1982) Stereochemistry of 1,3-elim-
inative cyclopropanation. Agric. Biol. Chem. 46, 1027–1034.
21. Cherbuliez, E., Weber, G. & Rabinowitz, J. (1965) Formation and
transformation of esters. LXI. Reaction of monochlorinated and
monobrominated trifluoroacetylacetic esters or monochlorinated
and monobrominated trifluoroacetone with ethyl phosphite. Helv.
Chim. Acta 48, 1423–1426.
22. Sekine,M.,Futatsugi,T.,Yamada,K.&Hata,T.(1982)Silyl
phosphites. Part 20. A facile synthesis of phosphoenolpyruvate
and its analog utilizing in situ generated trimethylsilyl bromide. J.
Chem. Soc. Perkin Trans. 1, 2509–2513.
23. Herzig, C. & Gasteiger, J. (1982) Reaction of 2-chlorooxiranes
with phosphites and phosphines: a new route to beta-carbonyl-
phosphonic esters and phosphonium salts. Chem. Ber. 115, 601–
614.

24. McKenna, C.E., Higa, M.T., Cheung, N.H. & McKenna, M.C.
(1977) The facile dealkylation of phosphonic acid dialkyl esters by
bromotrimethylsilane. Tetrahedron Lett. 155–158.
25. Bearne, S.L. & Kluger, R. (1992) Phosphoenol acetylpho-
sphonates: substrate analogs as inhibitors of phosphoenolpyruvate
enzymes. Bioorg. Chem. 20, 135–147.
26. Benenson, Y., Belakhov, V. & Baasov, T. (1996) 1-(Dihydroxy-
phosphynyl) vinyl phosphate: the phosphonate analog of phos-
phoenolpyruvate is a pH-dependent substrate of Kdo8P synthase.
Bioorg. Med. Chem. Lett. 6, 2901–2906.
27. Poyner, R.R., Laughlin, L.T., Sowa, G.A. & Reed, G.H. (1996)
Toward identification of acid/base catalysts in the active site of
enolase: comparison of the properties of K345A, E168Q, and
E211Q variants. Biochemistry 35, 1692–1699.
28. Liu, J., Peliska, J.A. & O’Leary, M.H. (1990) Synthesis and study
of (Z)-3-chlorophosphoenolpyruvate. Arch. Biochem. Biophys.
277, 143–148.
29. Duffy, T.H., Saz, H.J. & Nowak, T. (1982) Stereospecificity of
(E)- and (Z)-phosphoenol-alpha-ketobutyrate with chicken liver
phosphoenolpyruvate carboxykinase and related phospho-
enolpyruvate-utilizing enzymes. Biochemistry 21, 132–139.
30. Dı
´
az, E., O’Laughlin, J.T. & O’Leary, M.H. (1988) Reaction of
phosphoenolpyruvate carboxylase with (Z)-3-bromophos-
Ó FEBS 2002 Inhibitors of phosphoenolpyruvate-utilizing enzymes (Eur. J. Biochem. 269) 3235
phoenolpyruvate and (Z)-3-fluorophosphoenolpyruvate. Bio-
chemistry 27, 1336–1341.
31. Lichtenthaler, F.W. (1961) Chemistry and properties of enol
phosphates. Chem. Rev. 61, 607–649.

32. Marquardt, J.L., Brown, E.D., Lane, W.S., Haley, T.M.,
Ichikawa, Y., Wong, C.H. & Walsh, C.T. (1994) Kinetics, stoi-
chiometry, and identification of the reactive thiolate in the
inactivation of UDP-GlcNAc enolpyruvoyl transferase by the
antibiotic fosfomycin. Biochemistry 33, 10646–10651.
33. Blumberg, K. & Stubbe, J. (1975) Chemical specificity of pyruvate
kinase from yeast. Biochim. Biophys. Acta 384, 120–126.
34. Craik, C.S., Roczniak, S., Largman, C. & Rutter, W.J. (1987) The
catalytic role of the active site aspartic acid in serine proteases.
Science 237, 909–913.
35. McCray, J.W. & Weil, R. (1982) Inactivation of interferons:
halomethyl ketone derivatives of phenylalanine as affinity labels.
Proc. Natl Acad. Sci. USA 79, 4829–4833.
36. Seidel, H.M. & Knowles, J.R. (1994) Interaction of inhibitors with
phosphoenolpyruvate mutase: implications for the reaction
mechanism and the nature of the active site. Biochemistry 33,
5641–5646.
37. Westheimer, F.H. (1968) Pseudo-rotation in the hydrolysis of
phosphate esters. Acc. Chem. Res. 1, 70–78.
38. Clark, V.M. & Kirby, A.J. (1963) Neighboring group participa-
tion in phosphate ester hydrolysis. J. Am. Chem. Soc. 85, 3705–
3706.
39. O’Neal Jr, C.C., Bild, G.S. & Smith, L.T. (1983) Facile oxygen
exchanges of phosphoenolpyruvate and preparation of
[
18
O]phosphoenolpyruvate. Biochemistry 22, 611–617.
40. Schwarz, R. & Ugi, I. (1981) Synthesis of high-reactivity cyclic
enediol phosphates and cyclic acyl phosphates by direct phos-
phorylation ring closure. Angew. Chem. 93, 836–838.

41. Bovey, F.A. (1969) Nuclear Magnetic Resonance Spectroscopy.
Academic Press, New York, NY.
42. Wirsching, P. & O’Leary, M.H. (1988) 1-Carboxyallenyl phos-
phate, an allenic analogue of phosphoenolpyruvate. Biochemistry
27, 1355–1360.
43. Gonzalez, D.H. & Andreo, C.S. (1986) Phosphoenolpyruvate
carboxylase from maize leaves. Studies using beta-methylated
phosphoenolpyruvate analogs as inhibitors and substrates. Z.
Naturforsch. C. Biosci. 41, 1004–1010.
44. Cohn, M., Pearson, J.E., O’Connell, E.L. & Rose, I.A. (1970)
Nuclear magnetic resonance assignment of the vinyl hydrogens of
phosphoenolpyruvate. Stereochemistry of the enolase reaction. J.
Am. Chem. Soc. 92, 4095–4098.
45. Burbaum, J.J. & Knowles, J.R. (1989) Internal thermodynamics
of enzymes determined by equilibrium quench: values of Kint for
enolase and creatine kinase. Biochemistry 28, 9306–9317.
46. Stubbe, J. & Abeles, R.H. (1980) Mechanism of action of enolase:
effect of the beta-hydroxy group on the rate of dissociation of the
alpha-carbon-hydrogen bond. Biochemistry 19, 5505–5512.
47. Appelbaum, J. & Stubbe, J.A. (1975) Enolase catalyzed beta,
gamma-alpha,beta isomerization of 2-phospho-3-butenoic acid to
(Z)-phosphoenol-alpha-ketobutyrate. Biochemistry 14, 3908–3912.
48. Larsen, T.M., Wedekind, J.E., Rayment, I. & Reed, G.H. (1996)
A carboxylate oxygen of the substrate bridges the magnesium ions
at the active site of enolase: structure of the yeast enzyme com-
plexed with the equilibrium mixture of 2-phosphoglycerate and
phosphoenolpyruvate at 1.8 A
˚
resolution. Biochemistry 35, 4349–
4358.

49. Chauvin, F., Brand, L. & Roseman, S. (1994) Sugar transport
by the bacterial phosphotransferase system. Characterization
of the Escherichia coli enzyme I monomer/dimer transition
kinetics by fluorescence anisotropy. J. Biol. Chem. 269, 20270–
20274.
50. Matsumura, H., Terada, M., Shirakata, S., Inoue, T., Yoshinaga,
T., Izui, K. & Kai, Y. (1999) Plausible phosphoenolpyruvate
binding site revealed by 2.6 A
˚
structure of Mn
2+
-bound phos-
phoenolpyruvate carboxylase from Escherichia coli. FEBS Lett.
458, 93–96.
51. Mujica-Jimenez, C., Castellanos-Martinez, A. & Munoz-Clares,
R.A. (1998) Studies of the allosteric properties of maize leaf
phosphoenolpyruvate carboxylase with the phosphoenolpyruvate
analog phosphomycin as activator. Biochim. Biophys. Acta 1386,
132–144.
52. Waygood, E.B. & Steeves, T. (1980) Enzyme I of the phospho-
enolpyruvate: sugar phosphotransferase system of Escherichia coli.
Purification to homogeneity and some properties. Can. J. Bio-
chem. 58, 40–48.
SUPPLEMENTARY MATERIAL
The following material is available from ck
well-science.com/products/journals/suppmat/EJB/EJB2995/
EJB2995sm.htm
Table S1.
31
P- and

13
C-NMR and mass spectrometry data
of analogues 1-3.
Table S2.
1
H- and
31
P-NMR data of derivatives 6a–f, 7a
and 8a.
3236 L. F. Garcı
´
a-Alles and B. Erni (Eur. J. Biochem. 269) Ó FEBS 2002