Tài liệu Báo cáo khoa học: Isoprenoid biosynthesis via the methylerythritol phosphate pathway Mechanistic investigations of the 1-deoxy-D-xylulose 5-phosphate reductoisomerase ppt

Eur. J. Biochem. 269, 4446–4457 (2002) Ó FEBS 2002

doi:10.1046/j.1432-1033.2002.03150.x

Isoprenoid biosynthesis via the methylerythritol phosphate pathway
Mechanistic investigations of the 1-deoxy-D-xylulose 5-phosphate reductoisomerase
Jean-Francois Hoeffler, Denis Tritsch, Catherine Grosdemange-Billiard and Michel Rohmer
¸
Universite´ Louis Pasteur/CNRS, Institut Le Bel, Strasbourg, France

The 1-deoxyxylulose 5-phosphate reductoisomerase (DXR,
EC 1.1.1.267) catalyzes the conversion of 1-deoxy-D-xylulose
5-phosphate (DXP) into 2-C-methyl-D-erythritol 4-phosphate (MEP). This transformation is a two-step process
involving a rearrangement of DXP into the putative intermediate 2-C-methyl-D-erythrose 4-phosphate followed by a
NADPH-dependent reduction of the latter aldehyde. By
using [1-13C]DXP as a substrate, the rearrangement of DXP
into [5-13C]2-C-methyl-D-erythrose 4-phosphate was shown
to be NADPH dependent, although it does not involve
a reduction step. The putative aldehyde intermediate,
obtained by chemical synthesis, was converted into MEP by

the DXR in the presence of NADPH and into DXP in the
presence of NADP+, indicating the reversibility of the
reaction catalyzed by the DXR. This reversibility was confirmed by the conversion of MEP into DXP in the presence
of NADP+. The equilibrium was, however, largely displaced in favour of the formation of MEP. The reduction
step required the presence of a divalent cation such as Mg2+
or Mn2+.

Many bacteria, the unicellular green algae and the chloroplasts from phototrophic organisms synthesize their
isoprenoids via the mevalonate-independent 2-C-methylD-erythritol phosphate 5 (MEP) pathway (Fig. 1) [1–3].
The initial step of this route is the formation of 1-deoxyD-xylulose 5-phosphate 3 (DXP) by the condensation of

(hydroxyethyl)thiamine resulting from pyruvate 1 decarboxylation on glyceraldehyde 3-phosphate 2 catalyzed by
the thiamine diphosphate-dependent DXP synthase (DXS)
[4–6]. The second enzyme of this biosynthetic pathway,
the DXP reductoisomerase (DXR), catalyzes the transformation of DXP into MEP 5 in two steps. DXR is a
class B dehydrogenase [7,8]. The corresponding gene has
now been cloned from Escherichia coli [9], Zymomonas
mobilis [10], Mentha x piperita [11], Arabidopsis thaliana
[12], Synechocystis sp. [7], Streptomyces coelicolor [13] and
Pseudomonas aeruginosa [14]. In the postulated mechanism
of the reaction catalyzed by the DXR, DXP 3 is first
rearranged into 2-C-methyl-erythrose-4-phosphate 4 [15],
which is subsequently reduced by NADPH to yield MEP
5. The latter aldehyde intermediate 4 was, however, never

characterized, neither directly, nor indirectly. It is apparently not released from the enzyme active site during the
catalysis [16,17]. Three reactions are successively performed on the MEP framework, yielding three additional
intermediates of the MEP pathway: conversion of MEP 5
into 4-diphosphocytidyl-2-C-methyl-D-erythritol 6 [18,19],
phosphorylation of the C-2 hydroxyl group of 6 yielding 7
[20,21] and conversion of 7 into 2-C-methyl-D-erythritol
2,4-cyclodiphosphate 8 [22,23]. The two last steps of the
pathway were identified by a combination of genetic and
biochemical methods. An E. coli strain engineered for the
utilization of exogenous mevalonate accumulated tritiumlabelled 2-C-methyl-D-erythritol 2,4-cyclodiphosphate upon
incubation of [1-3H]-2-C-methyl-D-erythritol and after
disruption of the gcpE gene, suggesting that 2-C-methylD-erythritol 2,4-cyclodiphosphate 8 is the substrate of the
GcpE protein [24]. Incubation of [3-14C]-2-C-methylD-erythritol 2,4-cyclodiphosphate 8 with a crude cell-free
system from an E. coli strain overexpressing gcpE resulted
in the formation of 4-hydroxy-3-methylbut-2-enyl diphosphate 9 [25,26]. Deletion of the lytB gene in a similarly
engineered E. coli strain, resulted in the accumulation of

the same diol diphosphate 9 [27]. In addition, feeding
with uniformly labelled [U-13C5]-1-deoxy-D-xylulose E. coli
strains overexpressing the gene of the xylulose kinase
(responsible for the phosphorylation of free 1-deoxy-Dxylulose) as well as of all genes of the enzymes downstream of gcpE or lytB resulted in the accumulation of
uniformly labelled 4-hydroxy-3-methylbut-2-enyl diphosphate 9 or of isopentenyl diphosphate (IPP) 10 and
dimethylallyl diphosphate 11, respectively [28,29]. The
nature of the cofactors required for the conversion of
2-C-methyl-D-erythritol 2,4-cyclodiphosphate 8 into IPP 10
and dimethylallyl diphosphate 11 is still a matter of
investigation (Fig. 1).
This paper focuses on the two intriguing consecutive
steps catalyzed by the DXR from E. coli. Recently,

´
Correspondence to M. Rohmer, Universite Louis Pasteur/CNRS,
Institut Le Bel, 4 rue Blaise Pascal, 67070 Strasbourg Cedex, France.
Fax: +33 3 90241345, E-mail:
Abbreviations: AHIR, acetohydroxy acid isomeroreductase;
H2-NADPH, dihydro-NADPH; DXP, 1-deoxy-D-xylulose
5-phosphate; DXR, 1-deoxy-D-xylulose 5-phosphate reductoisomerase; DXS, 1-deoxy-D-xylulose 5-phosphate synthase;
H-DXR, His-tagged DXR; IPP, isopentenyl diphosphate;
MEP, 2-C-methyl-D-erythritol 4-phosphate.
Enzymes: acetohydroxy acid isomeroreductase (EC 1.1.1.86),
1-deoxy-D-xylulose 5-phosphate reductoisomerase (EC 1.1.1.267),
1-deoxy-D-xylulose 5-phosphate synthase (EC 4.1.3.7),
NADP-dependent alcohol dehydrogenase (EC 1.1.1.2).
(Received 12 June 2002, accepted 24 July 2002)

Keywords: isoprenoid, 2-C-methyl-D-erythritol 4-phosphate,
1-deoxyxylulose 5-phosphate reductoisomerase, 2-C-methylD-erythrose4-phosphate.

Ó FEBS 2002

Deoxyxylulose phosphate reductoisomerase (Eur. J. Biochem. 269) 4447

Fig. 1. 2-C-Methyl-D-erythritol 4-phosphate pathway for isoprenoid biosynthesis.

experiments performed using different combinations of
substrate, inhibitor and NADPH have been reported [17].
They suggested that NADPH binds before the normal
substrate DXP 3 or an inhibitor such as fosmidomycin
and were consistent with an ordered mechanism. DXR,
like all enzymes of the MEP pathway, is a potential target
for inhibitors acting as antibacterial or antiparasitic drugs
or as herbicides. Fosmidomycin, an inhibitor of the latter
enzyme [30], has been shown to be active against the
parasite responsible for malaria [31]. Knowledge of the
intimate mechanism of the reaction catalyzed by DXR is
required for the design of such inhibitors. The following
questions were thus addressed: (1) What is the role of
methylerythrose phosphate in the conversion of DXP into
MEP? (2) Is the cofactor NADPH required for the sole
isomerization of DXP into methylerythrose phosphate? (3)
Is the reaction catalyzed by DXR reversible? (4) What
kind of mechanism is involved in the rearrangement
leading to the branched MEP carbon skeleton? An a-ketol
rearrangement or a retroaldolization/aldolization would
both afford the same reaction product.

MATERIALS AND METHODS
General methods
Unlabelled DXP was prepared either enzymatically [32],
or chemically (Hoeffler et al., unpublished results). [1-13C]
DXP and [2-13C]DXP were prepared enzymatically from
glyceraldehyde phosphate and from [3-13C]pyruvate

(Isotec, Miamisburg, OH, USA) or from [2-13C]pyruvate
(Isotec, Saclay, France), respectively. MEP was obtained
either by chemical synthesis [33], or by enzymatic synthesis
[9]. [5-13C]MEP or [2-13C]MEP were prepared enzymatically from [1-13C]DXP or [2-13C]DXP, respectively. [2-13C]
Glycerol was purchased from Euriso-top (Saclay,
France). Unless otherwise indicated, substrates, coenzymes and enzymes were from Sigma. Dihydro-NADPH
(H2-NADPH) was synthesized from NADPH by catalytic
hydrogenation as previously described [34]. The concentration of H2-NADPH was determined from the absorbance at
263 nm (e ¼ 18 500 M)1Ỉcm)1). All enzymatic DXR assays
were recorded on an Uvikon 933 spectrophotometer
(Kontron Instruments) by following the variation of the
NADPH concentration. Glycoaldehyde 2-phosphate 12
was synthesized from glycerol 3-phosphate by treatment
with sodium metaperiodate and purified by anion exchange
chromatography. D-Erythrose 4-phosphate was purchased
from Fluka.
All nonaqueous reactions were run in dry solvents under
an argon atmosphere. ÔDried and concentratedÕ refers to the
removal of residual amounts of water with anhydrous
Na2SO4 followed by evaporation of the solvent on a rotary
evaporator. Flash chromatography [35] (Merck silica gel,
40–63 lm) and TLC (Merck 1.05553) were performed using
the same solvent system. TLC plates were developed by

heating up to 100 °C after spraying with an ethanol solution
of p-anisaldehyde (2.5%), sulfuric acid (3.5%) and acetic
acid (1.6%) or with an ethanol solution of phosphomolybdic
acid reagent (10% w/v). NMR spectra were recorded on a


4448 J.-F. Hoeffler et al. (Eur. J. Biochem. 269)

Ó FEBS 2002

Bruker AC200 spectrometer at 200 MHz for 1H-NMR and
50 MHz for 13C-NMR, on a Bruker AC300 at 300 MHz for
1
H-NMR, at 75 MHz for 13C-NMR and at 121.5 MHz for
31
P-NMR and also on Bruker ARX 500 at 500 MHz
for 1H-NMR, 125 MHz for 13C-NMR and 202 MHz for
31
P-NMR. 31P NMR spectra were calibrated against an
external H3PO4 standard (d ¼ 0.00 p.p.m.). NMR experiments were carried out in CDCl3 or D2O using as internal
standard CHCl3 (d ¼ 7.26 p.p.m.), DHO (d ¼ 4.56 p.p.m.)
for 1H-NMR and 13CDCl3 (d ¼ 77.03 p.p.m.) for
13
C-NMR. Negative mode electrospray mass spectrometry
was performed on a Hewlett Packard 1100MS spectrometer
using acetonitrile/water (1 : 1) as solvent. GC-MS by
chemical ionization was performed on a Finnigan-MAT
TSQ 700 spectrometer with a 70 eV ionization energy using
i-butane as gas.

(20 mL) at )5 °C, was added tert-butyldimethylsilyl triflate
(3.6 mL, 15.5 mmol, 1.5 equivalents). After 1 h at )5 °C,
the reaction was quenched with water (15 mL), diluted with
diethyl ether (45 mL), washed with 1 M HCl (15 mL), 5%
aqueous NaHCO3 (15 mL) and brine (15 mL). The organic
layer was dried and concentrated. The residue was purified
by flash chromatography to afford 16 as colourless oil
(2.2 g, 97%, Rf ¼ 0.48, hexane/ethyl acetate, 80 : 20). 1HNMR (300 MHz, CDCl3): d ¼ 0.14 (3H, s); 0.16 (3H, s);
0.90 (9H, s); 2.21 (1H, dddd, J ¼ 12.6 Hz, J ¼ 9.1 Hz,
J ¼ 9.1 Hz, J ¼ 8.6 Hz, 3-Ha); 2.46 (1H, dddd,
J ¼ 12.6 Hz, J ¼ 9.1 Hz, J ¼ 6.4 Hz, J ¼ 3.4 Hz, 3-Hb);
4.18 (1H, ddd, J ¼ 9.1 Hz, J ¼ 9.1 Hz, J ¼ 6.4 Hz; 4-Ha);
4.38 (2H, m, 4-Hb and 2-H). 13C-NMR (75 MHz, CDCl3):
d ¼ )4.90 (CH3); 17.86 (quaternary C); 25.57 (CH3); 38.12
(C-3); 68.08 (C-2); 76.08 (C-4); 175.63 (C-1).

Synthesis of 2-C-methyl-D-erythrose 4-phosphate 4

Synthesis of (3S )-3-hydroxypentan-2-one
5-phosphate 14

(3S)-3-(tert-Butyldimethylsiloxy)pentan-2-one 5-dibenzylphosphate 17. Methyl lithium (1.6 M solution in diethyl
ether, 5.9 mL, 9.5 mmol, 1.1 equivalents) was added
dropwise to a stirred solution cooled to )78 °C of silyl
ether 16 (1.9 g, 8.5 mmol, 1 equivalent) in tetrahydofuran
(40 mL). After stirring at )78 °C for 3 h, the reaction was
quenched by addition of water (20 mL) and diluted with
diethyl ether (40 mL). The organic layer was separated,
and the aqueous layer was extracted with diethyl ether
(3 · 40 mL). The combined organic extracts were dried and

concentrated in vacuo to give crude (3S)-3-(tert-butyldimethylsiloxy)-2-methyltetrahydrofuran-2-ol (1.8 g, 90%) as
a colourless oil corresponding to the mixture of diastereomers at C-2, which was used for the next step without
further purification.
Dibenzylphosphorochloridate [37] (1.6 g, 5.4 mmol, 1.5
equivalents) was added under stirring to a solution of
(3S)-3-(tert-butyldimethylsiloxy)-2-methyltetrahydrofuran-2-ol (1.0 g, 4.3 mmol, 1 equivalent) in pyridine (20 mL)
at 0 °C. The reaction was stirred at room temperature for
2 h. After quenching by addition of water (2 mL), the
solvents were removed under vacuum by azeotropic distillation with toluene. Flash chromatography gave 17 as
colourless oil (340 mg, 16%, Rf ¼ 0.35, hexane/ethyl acetate, 60 : 40). 1H-NMR (300 MHz, CDCl3): d ¼ 0.03 (3H,
s); 0.05 (3H, s); 0.90 (9H, s); 1.91 (1H, m, 4-Ha); 1.92 (1H, m,
4-Hb); 2.13 (3H, s, 1-H); 4.12 (3H, m, 3-H, 5-H); 5.02 (4H,
m, CH2Ph); 7.28–7.35 (10H, m). 13C-NMR (75 MHz,
CDCl3): d ¼ )5.18 (CH3); )4.99 (CH3); 18.03 (quaternary
C); 25.30 (C-1); 25.66 (3 · CH3); 35.04 (d, J ¼ 6.6 Hz,
C-4); 63.41 (d, J ¼ 4.9 Hz, C-5); 69.28 (d, J ¼ 4.9 Hz,
2 · CH2); 75.06 (C-3); 127.94, 128.53, 135.74 and 135.90
(aromatic C); 210.97 (C-2). 31P-NMR (121.5 MHz, CDCl3):
d ¼ )3.46.

(S)-2-Hydroxy-c-butyrolactone tert-butyldimethylsilylether 16. To a solution of (S)-2-hydroxy-c-butyrolactone
15 (1.1 g, 10.3 mmol, 1 equivalent) and 2,6-lutidine
(3.0 mL, 25.8 mmol, 2.5 equivalents) in dichloromethane

(3S)-3-Hydroxypentan-2-one 5-dibenzylphosphate 18. To
a stirred solution of the silyl ether 17 (200 mg, 0.41 mmol, 1
equivalent) in tetrahydofuran (5 mL) was added tetrabutylammonium fluoride (160 mg, 0.49 mmol, 1.2 equivalents).

2-O-Benzyl-2-C-methyl-D-erythrose 4-dibenzylphosphate 13
[33] was hydrogenated (200 mg, 0.41 mmol) over 10%

Pd/C (20 mg) in methanol (10 mL) for 30 min at room
temperature and atmospheric pressure (Fig. 2). The mixture
was filtered, and the filtrate diluted in water (15 mL),
concentrated in order to remove the methanol and treated
with a 1 M NaOH solution to reach pH 7.5, yielding a
mixture of 2-C-methyl-D-erythrose 4-phosphate 4 and its
dimethylacetal. This mixture was treated with an ion
exchange resin (Dowex 50 W-X4, H+ form) in water
(10 mL) at 37 °C for 2 h [36]. After filtration, the pH of the
filtrate was adjusted to 7.5 with a 1 M NaOH solution. For
NMR and mass spectra analyses, an aliquot of the solution
was lyophilized to dryness to afford the sodium salt of
2-C-methyl-D-erythrose 4-phosphate 4. This aldehyde cannot be stored pure or in concentrated solution and was
accordingly kept in water solution. The aldehyde concentration was determined using the DXR assay. 1H-NMR
(500 MHz, D2O, 1 : 2 mixture of the aldehyde and of the
hydrate): d ¼ 0.93 (2H, s, CH3, hydrate); 1.10 (1H, s, CH3,
aldehyde); 3.62 (2H, m, 2 · 4-H); 3.76 (1H, m, 3-H); 4.74
(0.7H, s, 1-H, hydrate); 9.41 (0.3H, s, 1-H, aldehyde). 13CNMR (125 MHz, D2O): d ¼ 16.34 (CH3); 18.25 (CH3);
63.65 (d, J ¼ 4.8 Hz, C-4); 65.12 (d, J ¼ 4.8 Hz, C-4);
73.78 (d, J ¼ 7.2 Hz, C-3); 74.11 (d, J ¼ 7.2 Hz, C-3);
75.27 (C-2); 79.83 (C-2); 91.78 (C-1, hydrate); 205.72 (C-1,
aldehyde). 31P-NMR (202 MHz, D2O): d ¼ 5.2 and 5.0.
Electrospray MS: m/z ¼ 213 (M-H, molecular ion of the
2-C-methyl-D-erythrose 4-phosphate mono-anion).

Fig. 2. Synthesis of 2-C-methyl-D-erythrose
4-phosphate 4. (i) H2, Pd/C in methanol;
(ii) DOWEX 50 W-X4Ò, H+, in water at
40 °C.

Ó FEBS 2002

Deoxyxylulose phosphate reductoisomerase (Eur. J. Biochem. 269) 4449

The mixture was stirred at room temperature for 30 min
and evaporated to dryness, and the residue was purified by
flash chromatography to afford 18 as a colourless oil
(135 mg, 88%, Rf ¼ 0.41, ethyl acetate/hexane, 90 : 10).
1
H-NMR (300 MHz, CDCl3): d ¼ 1.77 (1H, m, 4-Ha); 2.14
(1H, m, 4-Hb); 2.15 (3H, s, 1-H); 3.74 (1H, s, OH); 4.17 (3H,
m, 3-H, 5-H); 4.97 (1H, d, J ¼ 11.8 Hz, CH2Ph), 5.01 (1H,
d, J ¼ 11.6 Hz, CH2Ph), 5.04 (1H, d, J ¼ 11.8 Hz,
CH2Ph), 5.08 (1H, d, J ¼ 11.6 Hz, CH2Ph); 7.34 (10H,
m). 13C-NMR (75 MHz, CDCl3): d ¼ 25.11 (C-1); 33.79 (d,
J ¼ 6.6 Hz, C-4); 63.56 (d, J ¼ 6.6 Hz, C-5); 69.38 (d,
J ¼ 4.9 Hz, 2 · CH2); 73.20 (C-3); 127.97, 128.56, 135.64
and 135.77 (aromatic C); 209.40 (C-2). 31P-NMR
(121.5 MHz, CDCl3): d ¼ )3.35.
(3S)-3-Hydroxypentan-2-one 5-phosphate 14. (3S)-3Hydroxypentan-2-one 5-dibenzylphosphate 18 (35 mg,
0.01 mmol) was hydrogenated over 10% Pd/C (4 mg) in
ethanol (2 mL) for 2 h at room temperature and atmospheric pressure. The mixture was filtered, and the filtrate
concentrated. The residue was dissolved in water (1 mL),
and the pH adjusted to 7.5 with a 1 M NaOH solution. The
mixture was lyophilized to give the sodium salt of 18
(20 mg, 98%). 1H-NMR (200 MHz, D2O): d ¼ 1.60
(1H, m, 4-Ha); 1.93 (1H, m, 4-Hb); 2.01 (3H, s, 1-H); 3.69
(2H, ddd, J4a,5 ¼ J4b,5 ¼ J5,P ¼ 6.2 Hz, 5-H); 4.23 (1H,
dd, J ¼ 8.9 Hz, J ¼ 3.7 Hz, 3-H). 13C-NMR (75 MHz,

D2O): d ¼ 25.30 (C-1); 33.45 (C-4); 60.53 (C-5); 73.95 (C-3);
215.15 (C-2). 31P-NMR (121.5 MHz, D2O): d ¼ 4.25.
Synthesis of (4S)-4-hydroxypentan-2-one 5-phosphate 19
(R)-(tert-Butyldiphenylsiloxymethyl)oxirane 21. To a
stirred solution of (R)-glycidol 20 (2.7 g, 36.0 mmol, 1
equivalent) and imidazole (3.0 g, 44 mmol, 1.2 equivalents)
in dry dichloromethane (30 mL) was added at 0 °C tertbutylchlorodiphenylsilane (10.5 mL, 40 mmol, 1.1 equivalents). After 1 h at room temperature, the reaction mixture
was poured into water (30 mL), and the organic layer was
separated. The aqueous layer was extracted three times with
dichloromethane (250 mL). The combined extracts were
dried, filtered, concentrated and purified by flash chromatography to give 21 as a colourless oil (10.4 g, 92%,
Rf ¼ 0.21, hexane/ethyl acetate, 95 : 5). 1H-NMR
(200 MHz, CDCl3): d ¼ 1.09 (9H, s); 2.63 (1H, dd,
J ¼ 5.2 Hz, J ¼ 2.5 Hz, 3-Ha); 2.76 (1H, dd, J ¼ 5.2 Hz,
J ¼ 4.2 Hz, 3-Hb); 3.15 (1H, m, 2-H); 3.73 (1H, dd,
J ¼ 11.8 Hz, J ¼ 4.7 Hz, 1-Ha); 3.88 (1H, dd,
J ¼ 11.8 Hz, J ¼ 3.2 Hz, 1-Hb); 7.36–7.49 (6H, m); 7.69–
7.75 (4H, m). 13C-NMR (50 MHz, CDCl3): d ¼ 19.27
(quaternary C); 26.78 (3 · CH3); 44.45 (C-3); 52.25 (C-2);
64.34 (C-1); 127.74, 129.77, 133.31 and 135.64 (aromatic C).
(2S)-4-Methylpent-4-ene-1,2-diol 22. Into a flask equipped with a mechanical stirrer, an addition funnel and
containing anhydrous copper iodide (110 mg, 0.58 mmol,
0.1 equivalents) was added tetrahydofuran (20 mL). After
cooling at )30 °C, isoproprenylmagnesium bromide (0.5 M
in tetrahydofuran, 58 mL, 28.8 mmol, 5 equivalents) was
drop wise added. The temperature never exceeded )30 °C.
After stirring for 30 min at )30 °C (R)-(tert-butyldiphenylsiloxymethyl)oxirane 21 (1.8 g, 5.8 mmol, 1 equivalent) in
tetrahydofuran (10 mL) was slowly added, maintaining the

temperature at )30 °C. After stirring for 1 h at )30 °C, the

reaction was quenched by addition of a saturated NH4Cl
solution and warmed up to room temperature. The reaction
was filtered through a sintered glass funnel containing celite
and the tetrahydofuran was removed under reduce pressure.
The filtrate was diluted with diethyl ether (50 mL) and the
organic layer was washed with water (20 mL) and brine
(20 mL), dried and concentrated in vacuo. Purification by
flash chromatography afforded (2S)-1-tert-butyldiphenylsiloxy-2-hydroxypent-4-ene as a colourless oil (1.9 g, 98%,
Rf ¼ 0.28, ethyl acetate/hexane, 10 : 90).
To a stirred solution of the former silyl ether (1.9 g,
5.7 mmol, 1 equivalent) in tetrahydofuran (50 mL) was
added tetrabutylammonium fluoride (2.0 g, 6.2 mmol, 1.1
equivalents). The mixture was stirred at room temperature
for 3 h and evaporated to dryness. The residue was purified
by flash chromatography to afford 22 as colourless oil
(610 mg, 91%, Rf ¼ 0.33, ethyl acetate). 1H-NMR
(200 MHz, CDCl3): d ¼ 1.74 (3H, s, 4-CH3); 2.14 (2H, m,
2 · 3-H); 3.17 (2H, -OH); 3.42 (1H, dd, J ¼ 11.3 Hz,
J ¼ 7.1 Hz, 1-Ha); 3.63 (1H, dd, J ¼ 11.3 Hz, J ¼ 3.0 Hz,
1-Hb); 3.84 (1H, m); 4.77 (1H, m, 2-H); 4.83 (1H, m, 4-H).
13
C NMR (50 MHz, CDCl3): d ¼ 22.42 (4-CH3); 41.69
(C-3); 66.44 (C-1); 69.69 (C-2); 113.35 (C-3); 141.93
(quaternary C-4).
(2S)-2-Hydroxypent-4-en-1-ol dibenzylphosphate 23.
Dibenzylphosphorochloridate [36] (1.6 g, 5.4 mmol, 1.2
equivalents) was added under stirring to a solution of (2S)4-methylpent-4-ene-1,2-diol 22 (520 mg, 4.5 mmol, 1 equivalent) in pyridine (10 mL) at )40 °C during a period of
20 min. The reaction was stirred at )40 °C for 2 h, quenched
by addition of water (2 mL), and the solvents were removed
under vacuum by azeotropic distillation with toluene. Flash

chromatography gave 23 as a colourless oil (920 mg, 54%)
(Rf ¼ 0.24, ethyl acetate/hexane, 65 : 35). 1H-NMR
(200 MHz, CDCl3): d ¼ 1.72 (3H, s, 4-CH3); 2.13 (2H, m,
3-H); 2.69 (1H, -OH); 3.93 (3H, m, 1-H, 2-H); 4.75 (1H, m,
5-Ha); 4.83 (1H, m, 5-Hb); 5.07 (4H, m, 2 · -CH2Ph); 7.35
(10H, m). 13C-NMR (50 MHz, CDCl3): d ¼ 22.39
(4-CH3); 41.19 (C-3); 68.09 (d, J ¼ 6.1 Hz, C-2); 69.50 (d,
J ¼ 5.1 Hz, 2 · CH2); 71.49 (d, J ¼ 5.8 Hz, C-1); 113.64
(C-5); 127.97, 128.58, 135.60 and 135.74 (aromatic C); 141.33
(C-4). 31P-NMR (121.5 MHz, CDCl3): d ¼ )2.55.
(4S)-4-Hydroxypentan-2-one 5-dibenzylphosphate 24. To
a biphasic solution of (2S)-2-hydroxypent-4-en-1-ol dibenzylphosphate 23 (110 mg, 0.29 mmol, 1 equivalent) and
NaIO4 (263 mg, 1.2 mmol, 4.2 equivalents) in a mixture of
acetonitrile/carbon tertrachloride/H2O (2 : 2 : 3, 7 mL)
was added ruthenium trichloride (7 mg, 0.03 mmol, 0.1
equivalents) [38]. After vigorous stirring for 15 min at room
temperature, water (10 mL) and dichloromethane (10 mL)
were added, and the two phases were separated. The upper
aqueous phase was extracted four times with dichloromethane (4 · 25 mL). The combined organic extracts were
dried and concentrated. The residue was purified by flash
chromatography to afford 24 as a colourless oil (88 mg,
80%, Rf ¼ 0.41, ethyl acetate/hexane, 80 : 20). 1H-NMR
(200 MHz, CDCl3): d ¼ 2.13 (3H, s, 1-H); 2.55 (2H, m,
3-H); 3.40 (1H, broad s, -OH); 3.94 (2H, m, 5-H); 4.18 (1H,
m, 4-H); 4.99 (1H, d, J ¼ 11.6 Hz, CH2Ph), 5.03 (1H, d,
J ¼ 11.8 Hz, CH2Ph), 5.05 (1H, d, J ¼ 11.6 Hz, CH2Ph),


4450 J.-F. Hoeffler et al. (Eur. J. Biochem. 269)

5.09 (1H, d, J ¼ 11.8 Hz, CH2Ph); 7.34 (10H, m); 13CNMR (50 MHz, CDCl3): d ¼ 30.70 (C-1); 45.73 (C-3);
66.45 (d, J ¼ 6.5 Hz, C-4); 69.55 (d, J ¼ 5.0 Hz, 2 · CH2);
70.41 (d, J ¼ 6.2 Hz, C-5); 127.84, 128.01, 128.61, 135.57
and 135.68 (aromatic C); 207.82 (C-2). 31P-NMR
(121.5 MHz, CDCl3): d ¼ )2.73.
(4S)-4-Hydroxypentan-2-one 5-phosphate 19. (4S)-4Hydroxypentan-2-one 5-dibenzylphosphate 24 (55 mg,
0.15 mmol) was hydrogenated over 10% Pd/C (6 mg) in
ethanol (2 mL) for 2 h at room temperature and atmospheric pressure. The mixture was filtered, and the filtrate
and evaporated to dryness. The residue was dissolved in
water (1 mL), and the pH adjusted to 7.5 with a 1 M NaOH
solution. The mixture was lyophilized to give the sodium
salt of 19 (35 mg, 99%). 1H-NMR (500 MHz, D2O):
d ¼ 2.06 (3H, s); 2.55 (1H, dd, J ¼ 16.8 Hz, J ¼ 8.9 Hz,
3-Ha); 2.61 (1H, dd, J ¼ 16.8 Hz, J ¼ 4.1 Hz, 3-Hb); 3.51
(1H, m, 5-Ha); 3.57 (1H, m, 5-Hb); 4.07 (1H, m, 4-H).
13
C-NMR (75 MHz, D2O): d ¼ 29.66 (CH3); 45.64 (CH2);
46.02 (d, J ¼ 19.4 Hz, CH2); 66.95 (d, J ¼ 12.1 Hz, CH);
213.59 (CO). 31P-NMR (121.5 MHz, D2O): d ¼ 4.10.
Purification of His-tagged deoxyxylulose 5-phosphate
reductoisomerase
The coding region for the dxr gene from E. coli was cloned
into the pRSET vector (Invitrogen) between the BglII and
HindIII restriction sites. This vector contains a DNA
sequence encoding for six histidine residues. Plasmid
pRSET-DXR was introduced into E. coli strain
BL21(DE3)pLysE. After induction of enzyme expression
by addition of IPTG (0.4 mM) at mid-log phase (OD600,
0.7) at 37 °C, the culture was incubated for additional 3 h
at the same temperature. Cells (from 3 · 500 mL cultures)

were harvested by centrifugation and washed with water.
They were resuspended in a 50 mM Tris/HCl, 250 mM
NaCl, 5 mM 2-mercaptoethanol pH 8 buffer (10 mL) and
disrupted by sonication (8 · 30 s pulses at 40-W output,
duty cycle 50%) with cooling in an ice bath. The cell-free
system was centrifuged at 18 000 g for 30 min at 4 °C in a
Sigma 3K30 centrifuge. The crude cell extract was applied
on a column of Ni-nitrilotriacetic acid agarose (Qiagen,
0.8 · 2 cm) equilibrated with the same buffer. The column
was first washed with the same buffer, and then with the
buffer containing imidazole (5 mM). The enzyme was eluted
by applying a linear gradient of imidazole (5–120 mM) in the
same buffer (2 · 30 mL). Fractions containing His-tagged
DXR (H-DXR) were pooled and concentrated by ultrafiltration on a Centricon 30 unit (Millipore). The enzymatic
solution was dialysed against 50 mM Tris/HCl, 100 mM
NaCl, dithiothreitol (2 mM), pH 8 buffer by several
concentration/dilution steps using Centricon 30 units. The
concentration of protein was determined using the method
of Bradford [39].
H-DXR enzymatic activity
The enzymatic activity was determined routinely at 37 °C in
a 50 mM Tris/HCl, 1 mM MnCl2, 2 mM dithiothreitol
pH 7.5 buffer containing 0.15 mM NADPH and 0.5 mM
DXP. H-DXR was added to have an absorbance decrease
of about 0.1 min)1. The rate was measured by following the

Ó FEBS 2002

decrease of the absorbance at 340 nm due to the formation
of NADP+ from NADPH.

To compare the kinetic parameters (Km and V) of DXP 3
and 2-C-methyl-D-erythrose 4-phosphate 4, assays were
carried out in a 50 mM triethanolamine/HCl, 1 mM MnCl2
(or 3 mM MgCl2), 2 mM dithiothreitol, pH 7.7, at a fixed
concentration of NADPH (0.15 mM). The concentration of
DXP varied from 31 to 310 lM, while the concentration of
2-C-methyl-D-erythrose 4-phosphate 4 varied from 93 to
620 lM. The concentrations of the stock solutions of
substrate were determined enzymatically using the
H-DXR. The enzyme (4.3 lg) was added lastly in order
to initiate the reaction.
D-Erythrose 4-phosphate was tested as the substrate of
H-DXR at concentrations up to 1 mM and H-DXR concentrations up to 13 lgỈmL)1. The influence of D-erythrose
4-phosphate (1 mM) on the activity of H-DXR was checked
with DXP (96 lM) as the substrate. The kinetic parameters
(Km and V) in the reverse reaction were determined at a
fixed concentration of NADP+ (0.15 mM). The assays were
performed at 37 °C in a 50 mM Tris/HCl, 1 mM MnCl2,
2 mM dithiothreitol pH 7.5 buffer. The concentration of
MEP 5 varied from 75 to 375 lM. The concentration of the
stock solution of MEP 5 was determined by titration of the
phosphate according to the method of Leloir & Cardini [40].
The enzyme (4 lg) was added lastly in order to initiate the
reaction.
Reduction of 2-C-methyl-D-erythrose 4-phosphate 4 to
2-C-methyl-D-erythritol 4-phosphate 5 by H-DXR
To show that the reduction of 2-C-methyl-D-erythrose
4-phosphate 4 really gives 2-C-methyl-D-erythritol 4-phosphate 5, the aldehyde 4 (10 mg) was treated overnight with
DXR (1.2 mg) in the presence of NADPH (0.5 mM) in a
triethanolamine/HCl, 1 mM MnCl2, 2 mM dithiothreitol

pH 7.7 buffer at 37 °C (4 mL final volume). NADPH was
regenerated using the isopropanol/alcohol dehydrogenase
system from Thermoanaerobium brockii [41]. After hydrolysis of the phosphate esters with alkaline phosphatase
(bovine intestinal mucosa, Sigma, 0.5 mg) for 4 h at 37 °C,
the medium was lyophilized, and the residue was acetylated
overnight with a mixture of acetic anhydride and pyridine
(0.2 mL, 1 : 1 v/v). After evaporation of the reagents, the
residue was analysed by TLC. Methylerythritol triacetate
was isolated and identified by 1H-NMR (Rf ¼ 0.41, ethyl
acetate/hexane, 50 : 50). 1H NMR (200 MHz, CDCl3):
d ¼ 1.24 (3H, s, CH3); 2.04 (3H, s, CH3COO); 2.09
(3H, s, CH3COO); 2.11 (3H, s, CH3COO); 2.49 (1H, s,
OH); 3.89 (1H, d, J1a,1b ¼ 11.6 Hz, 1 Ha); 4.15 (1H, d,
J1a,1b ¼ 11.6 Hz, 1-Hb); 4.16 (1H, dd, J4a,4b ¼ 12.1 Hz,
J3,4a ¼ 8.1 Hz, 4-Ha); 4.56 (1H, dd, J4a,4b ¼ 12.1 Hz,
J3,4b ¼ 2.7 Hz, 4-Hb); 5.18 (1H, dd, J3,4a ¼ 8.1 Hz,
J3,4b ¼ 2.7 Hz, 3-H); 13C-NMR (50 MHz, CDCl3):
d ¼ 19.80 (CH3); 20.62 (CH3); 20.71 (CH3); 62.67 (CH2);
68.02 (CH2); 71.95 (quaternary C, C-2); 72.54 (CH, C-3);
169.99 (CO); 170.85 (2 · CO).
Isomerization of 2-C-methyl-D-erythrose 4-phosphate 4
into DXP 3 by H-DXR
2-C-Methyl-D-erythrose 4-phosphate 4 (10 mg) was treated
overnight with H-DXR (1.2 mg) in the presence of NADP+


Ó FEBS 2002

Deoxyxylulose phosphate reductoisomerase (Eur. J. Biochem. 269) 4451

(0.5 mM) in a 5 mM triethanolamine/HCl, 1 mM MnCl2,
2 mM dithiothreitol pH 7.7 buffer at 37 °C (4 mL, final
volume). The carbohydrate phosphates were identified after
dephosphorylation and acetylation by the usual method.
The acetylated crude residue was analysed by GCMS, and
the analytical data compared with those of a synthetic
reference of deoxyxylulose triacetate.
Reversibility of the formation of DXP 3 from MEP 5
by H-DXR
MEP 5 (10 mg) was treated overnight with H-DXR
(1.2 mg) in the presence of NADP+ (0.5 mM) in a 5 mM
triethanolamine/HCl, 3 mM MgCl2, 2 mM dithiothreitol
pH 7.7 buffer at 37 °C (4 mL final volume). NADP+ was
regenerated using the acetone/alcohol dehydrogenase from
Thermoanaerobium brockii [40]. After dephosphorylation
using an alkaline phosphatase, the mixture was lyophilized
and acetylated. Deoxyxylulose triacetate was isolated by
TLC and identified by 1H-NMR [6].
In other experiments, [5-13C]MEP or [2-13C]MEP (8 mM)
was treated overnight with H-DXR (1.1 mg) in the presence
of NADP+ (3 mM) in a 50 mM NH4HCO3, 3 mM MgCl2
and 2 mM dithiothreitol buffer at 37 °C (0.5 mL, final
volume). NADP+ was regenerated using the acetone/
alcohol dehydrogenase from Thermoanaerobium brockii.
The reaction was directly performed in a NMR tube and
monitored by 13C-NMR (50 MHz) using [2-13C]glycerol as
internal reference (d ¼ 71.3 p.p.m.).
Determination of the apparent equilibrium constant
of the DXR reaction
The assays were performed in a 50 mM Tris/HCl pH 7.5

buffer containing 1 mM MnCl2 and 2 mM dithiothreitol at
37 °C. H-DXR (12 lg) was incubated in the presence of
0.116 mM MEP and NADP+ at different concentrations
(0.088–0.352 mM) or at fixed concentration of NADP+
(0.176 mM) with MEP at different concentrations (0.058–
1.16 mM). The reactions were followed at 340 nm until the
absorbance reached a plateau. The concentration of
produced NADPH was determined from the absorbance
(e ¼ 6220 M)1 cm)1, kmax ¼ 340 nm). The influence of
DXP 3 (0.51–0.153 mM final concentration) or NADPH
(6.9–20.4 lM final concentration) on the concentration of
NADPH formed during the incubation of the enzyme with
NADP+ (0.176 mM) and MEP 5 (0.116 mM) was determined by the same UV absorption method.
13

C-NMR study of the rearrangement reaction of H-DXR

The reactions were directly performed in NMR tubes
(5 mm diameter) in a 50 mM NH4HCO3 buffer containing
3 mM MgCl2 and 2 mM dithiothreitol at 37 °C. The
[1-13C]DXP concentration was 12.5 mM. H-DXR (100 lg)
was added to initiate the enzymatic reaction. The influence
of 0.5 mM NADP+, 0.5 mM ATP-ribose and 0.5 mM H2NADPH was tested. The activity of the enzyme was
demonstrated by adding NADPH (0.3 mM) and its regenerating system, isopropanol/alcohol dehydrogenase from
Thermoanaerobium brockii. The reaction medium (620 lL
final volume) contained D2O (100 lL) and [2-13C]glycerol
(1 mg) as an internal reference (d ¼ 71.3 p.p.m.). 13C-

NMR spectra were recorded after 4 h incubation. The 13C
chemical shifts of the possible metabolites resulting from the

retro-aldol cleavage of DXP are hydroxyacetone 25 and
glycoaldehyde phosphate 12. The 13C shifts of hydroxyacetone 25 (0.7 M) and glycoaldehyde phosphate 12 (0.3 M)
were determined in the same medium. Hydroxyacetone 25:
13
C-NMR (50 MHz, 50 mM NH4HCO3): d ¼ 24.0 (C-3,
CH3), 66.7 (C-1, CH2OH), 211.0 (C-2, CO). Glycoaldehyde
2-phosphate 12: 13C-NMR (50 MHz, 50 mM NH4HCO3):
d ¼ 66.0 (d, C-2, J ¼ 3.3 Hz, CH2OP), 88.4 [d, C-1,
J ¼ 6.6 Hz, CH(OH)2].
Kinetic studies of (3S )-3-hydroxypentan-2-one
5-phosphate 14 and (4S )-4-hydroxypentan-2-one
5-phosphate 19 with H-DXR
H-DXR was incubated with (3R)-3-hydroxypentan-2-one
5-phosphate 14 (0.5 mM) or (4S)-4-hydroxypentan-2-one
5-phosphate 19 (0.5 mM) and NADPH (0.15 mM) in a
50 mM Tris/HCl, 1 mM MnCl2, 2 mM dithiothreitol pH 7.5
buffer. The reaction was followed at 340 nm to observe the
formation of NADP+. The inhibition of the enzymatic
activity of DXR by (3S)-3-hydroxypentan-2-one 5-phosphate 14 and (4S)-4-hydroxypentan-2-one 5-phosphate 19
was studied by determining the influence of the two
compounds [0.8–2.4 mM for (3S)-3-hydroxypentan-2-one
5-phosphate 14, 0.022–0.110 mM for (4S)-4-hydroxypentan-2-one 5-phosphate 19] on the enzymatic rate. The
concentration of DXP 3 varied between 75 and 510 lM.
DXR (4 lg) was added last to initiate the reaction.

RESULTS AND DISCUSSION
2-C-Methyl-D-erythrose 4-phosphate 4 as intermediate
in the DXR-catalyzed reaction
2-C-Methyl-D-erythrose 4-phosphate 4 was postulated as an
intermediate in the first step of the reaction catalyzed by

DXR. It results from an a-ketol rearrangement of DXP
and, after reduction, yields MEP 5. From the analogy of the
latter reaction sequence with that involved in the formation
of the carbon skeleton of amino acids with a branched sidechains, aldehyde 4 was expected to be only a transient
intermediate not released from the enzyme active site, much
like 3-hydroxy-3-methyl-2-oxobutyrate resulting from rearrangement of 2-acetolactate by acetohydroxy acid isomeroreductase (AHIR; EC 1.1.1.86) in the biosynthesis of
branched-chain amino acids [42].
For testing its possible role, unlabelled 2-C-methylD-erythrose 4-phosphate 4 was synthesized by an adaptation
of our former synthesis of MEP 5 (Fig. 2) [33]. 2-O-Benzyl2-C-methyl-D-erythrose 4-dibenzylphosphate 13 was
obtained as previously described in six steps from commercially available 1,2-O-isopropylidene-a-D-xylofuranose 26.
Hydrogenolysis of the benzyl groups in methanol yielded
the 2-C-methyl-D-erythrose 4-phosphate dimethylacetal,
which upon hydrolysis with an acidic DowexÒ resin
afforded 2-C-methyl-D-erythrose 4-phosphate 4. The putative intermediate was tested as substrate of H-DXR assays
and was utilized as reference material for the detection of
this aldehyde in the DXR enzyme tests.
Preliminary kinetic studies were performed in order to
determine the optimal conditions for the DXR-catalyzed


Ó FEBS 2002

4452 J.-F. Hoeffler et al. (Eur. J. Biochem. 269)

enzymatic reaction. Assays were performed in a triethanolamine/HCl buffer instead of the usual Tris/HCl buffer [9]
because Tris is known to react with aldehydes [43]. DXR
requires divalent cations such as Mn2+, Co2+ or Mg2+ for
its catalytic activity [9,16]. For our enzyme system, maximal rates were found for 1 mM Mn2+ and 3 mM Mg2+
concentrations, indicating that the enzyme has more affinity
for Mn2+ cations than for Mg2+ cations. The usual concentration of NADPH is 0.3 mM [9]. However, a 0.15 mM

concentration was chosen for NADPH because higher
concentrations resulted in lower rates. When 2-C-methylD-erythrose 4-phosphate 4 was incubated with H-DXR in the
presence of NADPH, consumption of the cofactor was
shown by the decrease in the absorption at 340 nm,
suggesting that the enzyme reduced the aldehyde 4. For the
identification of MEP, the expected reaction product, the
reaction mixture obtained after the reduction of 2-C-methylD-erythrose 4-phosphate 4 with NADPH was dephosphorylated using an alkaline phosphatase and freeze-dried. The
crude residue was acetylated, and TLC allowed the isolation
of methylerythritol triacetate, which was identified by NMR
by comparison with a synthetic reference sample [44]. This
confirmed that 2-C-methyl-D-erythrose 4-phosphate 4 was
effectively reduced to MEP 5 by H-DXR in the presence of
NADPH. Under these reaction conditions, deoxyxylulose
triacetate could not be isolated after TLC. The reverse
reaction, isomerization of 2-C-methyl-D-erythrose 4-phosphate 4 into DXP 3 did not take place significantly in the
presence of NADPH. The formation of DXP 3 from
methylerythrose 4 by H-DXR was, however, observed by
incubating the aldehyde in the presence of NADP+. As
methylerythrose and DX, as well as the diacetate of
methylerythrose and the triacetate of DX, have the same
Rf, the presence of DX triacetate was checked by GC and
GCMS. The retention times of the detected products were
compared with those of synthetic 1-deoxy-D-xylulose triacetate. In contrast with the almost quantitative formation of
MEP 5 from aldehyde 4, the formation of DXP 3 was very
low (% 7% yield as shown by GC detection of DX
triacetate). Furthermore, GC-MS (chemical ionization with
i-butane) of the acetylated crude reaction mixture showed a
peak with the retention time of deoxyxylulose triacetate and
characterized by a pseudo molecular ion at (M + H)+
(m/z ¼ 261) and by an ion corresponding to the loss of acetic

acid from the deoxyxylulose triacetate (m/z ¼ 201). This
confirmed the presence of small amounts of 1-deoxyD-xylulose triacetate.
The Km values measured for methylerythrose phosphate
(294 lM in the presence of 1 mM MnCl2 and 158 lM in the

presence of 3 mM MgCl2) for the E. coli H-DXR were
significantly higher than those found for DXP (73 lM for
1 mM MnCl2 and 97 lM for MgCl2) and also depended on
the nature of the divalent cation. Despite several reproducible measurements, for unknown reasons the Km values we
determined for DXP 3 in the presence of MnCl2 (1 mM)
differed significantly from those found in the literature for
the same enzyme from E. coli (Km ¼ 250 lM) [45] or from
Z. mobilis (Km ¼ 300 lM) [10]. However, the Km values for
DXP (Km ¼ 97 lM) when MgCl2 was used were similar to
those published for the purified E. coli enzyme wild-type
(Km ¼ 99 lM) [45] and for S. coelicolor DXP reductoisomerase (Km ¼ 60 lM) [13]. The results obtained with an
enzyme bearing a His-tag, like most those of the literature
concerning His-tagged proteins, may not be directly extended to the native enzyme. As the amino-terminal part of
DXR is involved in the binding of the cofactor [46], the Histag, which is localized at the N-terminal end, may influence
the enzymatic activity of H-DXR.
As for AHIR [42], the reduction step required the
presence of a divalent metal cation, which may be involved
in the binding of the aldehyde and/or the cofactor to the
enzyme. Whether such a metal cation is also required for the
isomerization remains to be shown. With DXP 3 as a
substrate, no significant difference of the kinetic constants
was observed at optimal concentrations of Mn2+ (1 mM)
and Mg2+ (3 mM) (Table 1). The influence of the nature of
the divalent cation was, however, more pronounced for
methylerythrose phosphate 4. In the presence of Mn2+, the

binding of the aldehyde 4 was less efficient than in the
presence of Mg2+. Indeed, although the chemical and
biochemical behaviour of Mn2+ resembles that of Mg2+,
˚
˚
Mn2+ (0.75 A) is somewhat larger than Mg2+ (0.65 A). In
2+
addition, Mn
binds more readily to a site containing
nitrogen in addition to oxygen than Mg2+, which prefers
oxygen only [47]. These peculiar properties of the two metal
cations may influence the binding of methylerythrose
phosphate 4 to the active site, and thus explain the different
Km values for the aldehyde 4. The higher rate of reduction
observed with Mn2+ could be due to a faster release of
MEP, the reaction product.
Interestingly, the methyl group of methylerythrose phosphate 4 is essential for the binding to the enzyme. In our
reaction conditions, D-erythrose 4-phosphate was neither a
substrate, nor an inhibitor of the H-DXR (data not shown).
It was recently reported that D-erythrose 4-phosphate is a
poor substrate of DXR [17]. According to our results,
methylerythrose phosphate apparently has a good affinity
with the enzyme, at least as compared with that of DXP 3

Table 1. Determination of the kinetic parameters (Km and V) of DXP 3 and 2-C-methyl-D-erythrose 4-phosphate 4. Assays were carried out in a
50 mM triethanolamine/HCl, 1 mM MnCl2 (or 3 mM MgCl2), 2 mM dithiothreitol pH 7.7 buffer at a fixed concentration of NADPH (0.15 mM).
The concentration of DXP varied from 31 to 310 lM while the concentration of 2-C-methyl-D-erythrose 4-phosphate 4 varied from 93 to 620 lM.
The concentrations of the stock solutions of substrate were determined enzymatically using the H-DXR. The enzyme (4.3 lg) was added last in
order to initiate the reaction.
DXP 3

2-C-Methyl-D-erythrose 4-phosphate 4

1 mM MnCl2
Km (lM)
V (lmolỈmin)1Ỉmg DXR)1)

3 mM MgCl2

1 mM MnCl2

3 mM MgCl2

73
10.5

97
10.5

294
20.6

158
11.9


Ó FEBS 2002

Deoxyxylulose phosphate reductoisomerase (Eur. J. Biochem. 269) 4453

(Table 1). As erythrose 4-phosphate has a very weak affinity
with the DXR, it appears that the methyl group at C-2 must
play a crucial role for the binding of the substrate to the
enzyme active site.
DXR and AHIR, an enzyme involved in the biosynthesis of branched-chain amino acids [42], catalyze similar
reactions. The latter converts 2-acetolactate or 2-aceto-2hydroxybutyrate into 2,3-dihydroxy-3-isovalerate or 2,3dihydroxy-3-methylvalerate. This reaction proceeds in two
steps: an isomerization, consisting of an alkyl migration, is
followed by an NADPH-dependent reduction of the oxo
group to give the final product. In the reactions catalyzed by
the two enzymes, the ketol-acid and the DXP isomeroreductase, the formation of the expected intermediates,
3-hydroxy-3-methyl-2-oxo-butyrate or methylerythrose
phosphate, respectively, has never been shown. For AHIR,
it was suggested that the intermediate may be tightly bound
to the enzyme or that the reduction takes place during the
alkyl transfer so that the intermediate is never really formed
[42]. In none of our assays could the formation of
methylerythrose phosphate 4 be detected. Assays designed
to dissociate the transposition step from the reduction when
using DXP 3 as substrate were performed for a tentative
direct identification of methylerythrose phosphate. In the
absence of cofactor, no isomerization was observed. Accordingly, the simultaneous presence of the divalent cation and
of the cofactor might be required for the subsequent fixation
of DXP 3 or methylerythrose phosphate 4 [17]. DihydroNADPH, an NADPH analogue [34] which is not a reducing
cofactor, was expected to bind to the enzyme much like the
natural coenzyme [17]. Inhibition of the reaction by
dihydro-NADPH would suggest that this analogue was
bound to the active site of the enzyme. It was, however,
impossible to induce the isomerization of DXP 3 into
methyl erythrose phosphate 4.
The reduction step seems to represent the driving force to

perform the rearrangement. The fact that the postulated
oxo intermediate in both isomeroreductase-catalyzed reactions are substrates for the reduction step with higher Km
than those of the normal substrates suggests that the first
proposition, their tight binding to the enzyme, is rather
improbable. The reduction step might be necessary in order
that the isomerization takes place.
Reversibility of the DXR-catalyzed reaction: formation
of DXP 3 from MEP 5
In order to verify that the H-DXR is capable of catalyzing
the reverse reaction, the enzyme was incubated with MEP 5
and NADP+. A first series of experiments was performed
with 13C-labelled MEP. When H-DXR was incubated in the
presence of [5-13C]MEP or [2-13C]MEP and NADP+, a
decrease of the C-5 (d ¼ 17.7 p.p.m.) or of the C-2
(d ¼ 73.2 p.p.m.) signals from MEP was observed, accompanied by a concomitant appearance and following increase
of new signals corresponding to C-1 (d ¼ 25.5 p.p.m.) and
C-2 (d ¼ 212.4 p.p.m.) from [1-13C]DXP or [2-13C]DXP,
respectively. A second experiment was performed with
unlabelled MEP. The increase of the absorbance at 340 nm,
due to the formation of NADPH, suggested that MEP 5
was at least oxidized to 2-C-methyl-D-erythrose 4-phosphate
4, the supposed intermediate of the reaction. To show that
the enzyme is capable of performing the two steps of the

reverse process, converting MEP 5 into DXP 3, i.e.
oxidation of MEP into 4 and rearrangement of 4 into
DXP 3, MEP 5 was treated with H-DXR in the presence
of NADP+ and a regeneration system of the coenzyme to
favour the reaction in the direction of the formation of DXP
3. After dephosphorylation and acetylation of the reaction

mixture, 1-deoxy-D-xylulose triacetate and 2-C-methylD-erythritol triacetate were isolated by TLC and identified
by 1H-NMR, proving that H-DXR catalyzed the reverse
transformation of MEP 5 into DXP 3, including not only the
oxidation of MEP 5 to 2-C-methyl-D-erythrose 4-phosphate
4, but also the rearrangement of the latter aldehyde into
DXP 3. Methylerythrose diacetate, which coelutes with DX
triacetate, was not observed.
When the reverse reaction with MEP 5 (115 lM) and
NADP+ (175 lM) was followed during several minutes
the absorbance increase stopped completely when about
13–14% of the MEP 5 was transformed, corresponding to
the production of some NADPH (15.6 lM). The addition of
more H-DXR to the reaction medium did not induce any
additional absorbance increase. Ceasing NADPH formation was not due to the inactivation of the enzyme, but
rather to the fact that the equilibrium had been reached.
This low production of NADPH suggests that the reaction
equilibrium is largely in favour of the production of MEP 5
from DXP 3. In order to confirm this hypothesis, the
influence of the concentrations of NADP+ and MEP 5, the
substrates of the reverse reaction, and of NADPH and DXP
3, the products of the reaction, on the total amount of
NADPH produced were analysed.
The apparent equilibrium constant
Kẳ

ẵDXPeq ẵNADPHeq ẵHỵ
ẵMEPeq ẵNADPỵ eq

where [DXP], [MEP], [NADPH] and [NADP+] represent
the concentrations of the different compounds at the

equilibrium, was calculated in each case and found to be
approximately the same (average value 4.6 ± 0.5 · 10)10 M
at 37 °C). Attempts to determine the Km of MEP 5
showed that the Kms of MEP 5 (116 lM) and DXP 3
(76 lM) had similar values. The V of the reverse reaction
(3.5 mMỈmin)1Ỉmg protein)1) was about 60% of that of
the formation of MEP 5 from DXP 3 (5.6 mMỈmin)1Ỉmg
protein)1).
Even the very transitory existence of methylerythrose
phosphate 4 is no longer an assumption, as this aldehyde 4
is the substrate of the DXP reductoisomerase with a good
affinity. It is converted into MEP 5, as well as into DXP 3.
Furthermore, our NMR data afforded direct evidence for
the reversibility of the reaction catalyzed by the DXR by the
conversion of MEP 5 into DXP 3.
Rearrangement of DXP 3 to 2-C-methyl-D-erythrose4-phosphate 4: the role of NADPH
The conversion of DXP 3 into MEP 5 by the DXR requires
the presence of NADPH as cofactor. The first step of this
conversion, i.e. the rearrangement of DXP 3 into methylerythrose phosphate 4, is, however, formally fully independent of this cofactor, as this rearrangement only
corresponds to an isomerization. In order to try to shed
light on the possible role of methylerythrose phosphate 4,


Ó FEBS 2002

4454 J.-F. Hoeffler et al. (Eur. J. Biochem. 269)

the DXR-catalyzed reaction was followed by 13C-NMR
using a 13C-labelled substrate (99% isotope abundance) in
order to improve the sensitivity of the detection method.

[1-13C]DXP was incubated in the presence of NADPH and
of a NADPH regenerating system. The enzyme preparation
was active. The decrease of the intensity of the C-1 signal of
[1-13C]DXP (d ¼ 25.5 p.p.m.) was accompanied by the
concomitant increase of the C-5 signal of [5-13C]MEP
(d ¼ 17.7 p.p.m.). In a second experiment, NADPH was
omitted in order to check whether the presence of the
reducing cofactor is essential in the transposition step. When
[1-13C]DXP was incubated alone in the absence of
NADPH, no additional 13C signal was observed, and
especially no signal corresponding to the C-5 methyl group
of methylerythrose phosphate 4 (d ¼ 16.34 p.p.m., aldehyde form; d ¼ 18.25 p.p.m. hydrate form), indicating that
no reaction had occurred, at least within the limits of the
13
C-NMR detection. The presence of the native coenzyme
seems essential for the rearrangement, although it is not
formally required. In the presence of NADPH analogues,
such as NADP+, dihydro-NADPH or ATP-ribose, DXP
remained intact and no conversion into methylerythrose
phosphate was observed.
In conclusion, direct evidence for the formation of 2-Cmethyl-D-erythrose 4-phosphate 4 was not obtained in this
enzymatic reaction, and no isomerization took place in the
absence of NADPH. These results were consistent with
those of previous observations reported in the literature: the
binding of NADPH first only allowed the binding of the
DXP 3 [17].

Formation of 2-C-methyl-D-erythrose 4-phosphate
4 from DXP: an a-ketol rearrangement or a
retro-aldolization reaction?

Incubation of [4,5-13C2]glucose into triterpenoids of the
hopane series from Methylobacterium fujisawaense demonstrated for the first time the presence of a rearrangement in
the MEP pathway [48]. By analogy with the biosynthesis of
the amino acids with branched side-chains, the conversion
of DXP 3 into MEP 5 was considered as an a-ketol
rearrangement (Fig. 3). This mechanism involves the
deprotonation of the hydroxyl group at C-3 of DXP 3,
followed by the migration of the phosphate-bearing C2
subunit to afford methylerythrose phosphate 4. Examples
of reactions utilizing a retroaldolization/aldolization type
mechanism in the place of an a-ketol rearrangement are also
found in the literature, e.g. the reaction catalyzed by the
ribulose 5-phosphate 4-epimerase [49,50]. In addition, such
aldolization reactions often require the presence of a
divalent cation [50], much like the reaction catalyzed by
DXR. For the formation of MEP 5 from DXP 3, such an
alternative mechanism would involve the deprotonation of
the C-4 hydroxyl group of DXP 3, followed by the cleavage
of the carbon–carbon bond between the carbon atoms C-3
and C-4 to give the enolate of hydroxyacetone 25 and
glycoaldehyde phosphate 12 (Fig. 3). Recombination of the
two resulting moieties, rearranged via an aldolization by
formation of a novel carbon–carbon bond between the
carbon atoms derived from C-2 and C-4 of DXP 3, would
give 2-C-methyl-D-erythrose 4-phosphate 4. In order to try

Fig. 3. Conversion of DXP into MEP by the DXR: a-ketol rearrangement vs. retro-aldolization/aldolization.


Ó FEBS 2002

Deoxyxylulose phosphate reductoisomerase (Eur. J. Biochem. 269) 4455

to get more insight into the mechanism of the DXRcatalyzed reaction, the DXP analogue 14 was synthesized
and analysed for its behaviour towards the DXR. On
the one hand, if DXP analogue 14 is transformed to the
(2R)-2-hydroxy-2-methylbutanol 4-phosphate 27, the reaction is most probably an a-ketol rearrangement. On the
other hand, the retro-aldolization would imply the intermediary formation of glycoaldehyde phosphate 12 and of the
enol of hydroxyacetone 25 (Fig. 3).
Formation of 2-C-methyl-D-erythrose 4-phosphate 4 from
DXP: an a-ketol rearrangement? The possibility that
DXR catalyzes the formation of 2-C-methyl-D-erythrose
4-phosphate 4 from DXP 3 via an a-ketol rearrangement
(Fig. 3) was checked by testing (3S)-3-hydroxypentan-2-one
5-phosphate 14 as potential substrate of H-DXR. This
compound was synthesized from the commercially available
(S)-2-hydroxy-c-butyrolactone 15, which has the required
configuration for the C-2 asymmetric carbon of 14 (Fig. 4).
Protection of the secondary alcohol of 15 afforded the silyl
ether 16 in 97% yield. Addition of methyl lithium gave a
lactol, which opened under standard phosphorylation
condition to the enantiomerically pure ketone 17, but with
low yield [37]. Deprotection of the silyl ether 17 with fluoride
salts (TBAF), followed by hydrogenolysis of the benzyl
groups, afforded (3R)-3-hydroxypentan-2-one 5-phosphate
14. Incubation of the DXP analogue 14 with H-DXR in the
presence of NADPH did not induce any decrease of the
absorbance at 340 nm. The DXP analogue 14 was not a
substrate of DXR. It was, however, recognized by DXR
and reversibly inhibited the enzyme as a mixed-type

inhibitor (Ki ¼ 120 lM). In conclusion, no information
was obtained on the reaction mechanism of the DXRcatalyzed reaction, but the crucial role of the C-4 hydroxy
group of DXP 3 was pointed out in the isomerization steps.
Furthermore, as the DXP analogue 14 inhibited the
DXR, its isomer, (4S)-4-hydroxypentan-2-one 5-phosphate
19, was also synthesized and tested on H-DXR. The
synthesis started with the preparation of the silyl ether 21
from commercially available (R)-glycidol 20 (Fig. 4). Epoxide opening of 21 with isoproprenylmagnesium bromide in
the presence of CuI, followed by the deprotection of the silyl
ether, gave diol 22 in excellent yield. Selective phosphorylation of the primary alcohol of 22 with dibenzylphosphate
chloride [37] at low temperature followed by oxidative
cleavage [38] of the double bond yielded the protected
ketone 23 in 39% yield over two steps. Finally the benzyl
groups were quantitatively removed by hydrogenolysis in
the presence of a catalytic amount of palladium on activated
carbon to yield (4S)-4-hydroxypentan-2-one 5-phosphate 19
(Fig. 4). Analogue 19 also inhibited the DXR as a mixed
type inhibitor (Ki ¼ 800 lM).
Formation of 2-C-methyl-D-erythrose 4-phosphate 4 from
DXP: a retro-aldolization mechanism? In the case of an
alternative retro-aldolization mechanism, hydroxyacetone
25 and glycoaldehyde phosphate 12 are the two intermediates leading to the formation of methylerythrose phosphate
4 (Fig. 3). Detection of hydroxyacetone 25 was attempted
by incubation of [1-13C]DXP and following the reaction by
13
C-NMR spectroscopy. No signal corresponding to the
C-3 of hydroxyacetone 25 (d ¼ 24.0 p.p.m.), which was
expected to be labelled in the case of hydroxyacetone

Fig. 4. Synthesis of (3S)-3-hydroxypentan-2-one 5-phosphate 14 and

(4S)-4-hydroxypentan-2-one 5-phosphate 19. (A) Synthesis of (3R)-3hydroxypentan-2-one 5-phosphate 14. (i) TBDMSOTf, 2,6-lutidine,
dichloromethane, 97%; (ii) CH3Li, tetrahydofuran, )78 °C, 67%;
(iii) (BnO)2POCl, pyridine, 16%; (iv) Bu4NF, tetrahydofuran, 88%;
(v) H2, Pd/C, EtOH, 100%. (B) Synthesis of (4S)-4-hydroxypentan-2one 5-phosphate 19. (i) TBDPSCl, imidazole, dichloromethane, 92%;
(ii) CH2 ¼ CHMgBr, CuI, tetrahydofuran; (iii) Bu4NF, tetrahydofuran, 89% from two steps; (iv) (BnO)2POCl, pyridine, )40 °C, 54%;
(v) RuCl3, NaIO4, CH3CN, CCl4, H2O, 72%; (vi) H2, Pd/C, EtOH,
100%.

formation, was observed next to those of C-1 of DXP
(d ¼ 25.5 p.p.m.) and C-5 of MEP (d ¼ 17.7 p.p.m.). In
addition, the influence of hydroxyacetone 25 and glycoaldehyde phosphate 12, the two intermediates in a retroaldolization/aldolization mechanism (Fig. 3), on the activity
of DXR was checked. When the enzyme was incubated in
the presence of the two compounds at concentrations up to
1 mM, and NADPH, no decrease of the absorbance at
340 nm was observed, indicating that no MEP 5 was
produced. In addition, the two compounds, either alone or
together at concentrations of up to 1 mM, did not inhibit the
production of MEP 5 from DXP 3. They do not seem to be
recognized by DXR.
These negative results did not enable us to retain or
exclude either one or the other mechanism. The absence of
NADPH consumption during the incubation of the enzyme


Ó FEBS 2002

4456 J.-F. Hoeffler et al. (Eur. J. Biochem. 269)

with analogue 14 might be in favour of a retro-aldolization
but this feature is better explained by the absence of

formation of a productive complex between the analogue
and the divalent ions. Both DXP analogues, 14 and 19, each
lacking a hydroxyl group either at C-4 or at C-3, respectively, inhibited the DXR. These hydroxyl groups are
apparently essential in the rearrangement reaction. It is
very likely that the hydroxyl groups are sites for the
chelation of a divalent cation such as Mg2+ or Mn2+, most
probably acting as Lewis acids facilitating the rearrangement reaction. Furthermore the sequential mechanism
followed by the enzyme has been confirmed: NADPH
and the divalent cation bind first and precede the binding of
the substrate [17]. There is some evidence that this sequence
implies a conformational change of the enzyme to allow the
reaction as well for the isomerization step as for the
reduction, as methylerythrose phosphate 4 is a substrate of
DXR. More detailed analysis of the mechanism of the
DXR-catalyzed reaction will be possible as the crystal
structure of this enzyme is now being determined [46,51].

ACKNOWLEDGEMENTS
We thank Mr J.-D. Sauer for all NMR measurements, Ms. E. Mastio
for recording the mass spectra and O. Besumbes and Dr S. Imperial
(University of Barcelona, Spain) for the construction of the E. coli
strain overexpressing the DXR. This investigation was supported by a
grant to MR from the ÔInstitut Universitaire de FranceÕ.

9.

10.

11.

12.

13.

14.

15.

REFERENCES
1. Rohmer, M. (1999) The discovery of a mevalonate-independent
pathway for isoprenoid biosynthesis in bacteria, algae and higher
plants. Nat. Prod. Rep. 16, 565–574.
2. Lichtenthaler, H.K., Rohmer, M. & Schwender, J. (1997) Two
independent biochemical pathways for isopentenyl diphosphate
and isoprenoid biosynthesis in higher plants. J. Physiol. Plant. 101,
643–652.
3. Eisenreich, W., Schwarz, M., Cartayrade, A., Arigoni, D., Zenk,
M.H. & Bacher, A. (1998) The deoxyxylulose phosphate pathway
of terpenoid biosynthesis in plants and microorganisms. Chem.
Biol. 5, R221–R233.
4. Sprenger, G.A., Schorken, U., Wiegert, T., Grollessss, S., Graaf,
ă
A.A., Taylor, S.V., Begley, T.P., Bringer-Meyer, S. & Sahm, H.
(1997) Identification of a thiamin-dependent synthase in
Escherichia coli required for the formation of the 1-deoxy-Dxylulose 5-phosphate precursor to isoprenoids, thiamin, and pyridoxol. Proc. Natl Acad. Sci. USA 94, 12857–12862.
5. Lange, B.M., Wildung, M.R., McCaskill, D. & Croteau, R. (1998)
A family of transketolases that directs isoprenoid biosynthesis via
a mevalonate-independent pathway. Proc. Natl Acad. Sci. USA
95, 2100–2104.
6. Lois, L.M., Campos, N., Rosa-Putra, S., Danielsen, K., Rohmer,

M. & Boronat, A. (1998) Cloning and characterization of a gene
from Escherichia coli encoding a transketolase-like enzyme that
catalyzes the synthesis of 1-D-deoxyxylulose 5-phosphate, a common precursor for isoprenoid, thiamin, and pyridoxol biosynthesis. Proc. Natl Acad. Sci. USA 95, 2105–2110.
7. Proteau, P.J., Woo, Y.-H., Williamson, R.T. & Phaosiri, C. (1999)
Stereochemistry of the reduction step mediated by recombinant
1-deoxy-D-xylulose 5-phosphate isomeroreductase. Org. Lett. 1,
921–923.
8. Radykewicz, T., Rohdich, F., Wungsintaweekul, J., Herz, S., Kis,
K., Eisenreich, W., Bacher, A., Zenk, M.H. & Arigoni, D. (2000)
Biosynthesis of terpenoids: 1-deoxy-D-xylulose-5-phosphate

16.

17.

18.

19.

20.

21.

22.

reductoisomerase from Escherichia coli is a class B dehydrogenase.
FEBS Lett. 465, 157–160.
Kuzuyama, T., Takahashi, S., Watanabe, H. & Seto, H. (1998)
Direct formation of 2-C-methyl-D-erythritol 4-phosphate from
1-deoxy-D-xylulose 5-phosphate by 1-deoxy-D-xylulose 5-phosphate reductoisomerase, a new enzyme in the non-mevalonate

pathway to isopentenyl diphosphate. Tetrahedron Lett. 39, 4509–
4512.
Grolle, S., Bringer-Meyer, S. & Sahm, H. (2000) Isolation of the
dxr gene of Zymomonas mobilis and characterization of the
1-deoxy-D-xylulose 5-phosphate reductoisomerase. FEMS
Microbiol. Lett. 191, 131–137.
Lange, B.M. & Croteau, R. (1999) Isoprenoid biosynthesis via a
mevalonate-independent pathway in plants: cloning and heterologous expression of 1-deoxy-D-xylulose-5-phosphate reductoisomerase from peppermint. Arch. Biochem. Biophys. 365, 170174.
Schwender, J., Muller, C., Zeidler, J. & Lichtenthaler, H.K. (1999)
ă
Cloning and heterologous expression of a cDNA encoding
1-deoxy-D-xylulose-5-phosphate reductoisomerase of Arabidopsis
thaliana. FEBS Lett. 455, 140–144.
Cane, D.E., Chow, C., Lillo, A. & Kang, I. (2001) Molecular
cloning, expression and characterization of the first three genes in
the mevalonate-independent isoprenoid pathway in Streptomyces
coelicolor. Bioorg. Med. Chem. 9, 1467–1477.
Altincicek, B., Hintz, M., Sanderbrand, S., Wiesner, J., Beck, E. &
Jomaa, H. (2000) Tools for discovery of inhibitors of the 1-deoxyD-xylulose 5-phosphate (DXP) synthase and DXP reductoisomerase: an approach with enzymes from the pathogenic
bacterium Pseudomonas aeruginosa. FEMS Microbiol. Lett. 190,
329–333.
Rohmer, M., Seemann, M., Horbach, S., Bringer-Meyer, S. &
Sahm, H. (1993) Glyceraldehyde 3-phosphate and pyruvate as
precursors of isoprenic units in an alternative non-mevalonate
pathway for terpenoid biosynthesis. J. Am. Chem. Soc. 118, 2556–
2566.
Takahashi, S., Kuzuyama, T., Watanabe, W. & Seto, H. (1998) A
1-deoxy-D-xylulose 5-phosphate reductoisomerase catalyzing the
formation of 2-C-methyl-D-erythritol 4-phosphate in an alternative nonmevalonate pathway for terpenoid biosynthesis. Proc.
Natl Acad. Sci. USA 95, 9879–9884.

Koppisch, A.T., Fox, D.T., Blagg, B.S.J. & Poulter, C.D. (2002)
E. coli MEP synthase: steady-state kinetic analysis and substrate
binding. Biochemistry 41, 236–243.
Rohdich, F., Wungsintaweekul, J., Fellermeier, M., Sagner, S.,
Herz, S., Kis, K., Eisenreich, W., Bacher, A. & Zenk, M.H. (1999)
Cytidine 5¢-triphosphate-dependent biosynthesis of isoprenoids:
ygbP protein of Escherichia coli catalyzes the formation of 4diphosphocytidyl-2-C-methylerythritol. Proc. Natl Acad. Sci.
USA 96, 11758–11763.
Kuzuyama, T., Takagi, M., Kaneda, K., Dairi, T. & Seto, H.
(2000) Formation of 4-(cytidine 5¢-diphospho)-2-C-methyl-D-erythritol from 2-C-methyl-D-erythritol 4-phosphate by 2-C-methylD-erythritol 4-phosphate cytidyltransferase, a new enzyme in the
nonmevalonate pathway. Tetrahedron Lett. 41, 703706.
Luttgen, H., Rohdich, F., Herz, S., Wungsintaweekul, J., Hecht,
ă
S., Schuhr, C.A., Fellermeier, M., Sagner, S., Zenk, M.H., Bacher,
A. & Eisenreich, W. (2000) Biosynthesis of terpenoids: ychB
protein of Escherichia coli phosphorylates the 2-hydroxy group of
4-diphosphocytidyl-2C-methyl-D-erythritol. Proc. Natl Acad. Sci.
USA 97, 1062–1067.
Kuzuyama, T., Takagi, M., Kaneda, K., Watanabe, H., Dairi, T.
& Seto, H. (2000) Studies on the nonmevalonate pathway: conversion of 4-(cytidine 5¢-diphospho)-2C-methyl-D-erythritol to its
2-phospho derivative by 4-(cytidine 5¢-diphospho)-2-C-methylD-erythritol kinase. Tetrahedron Lett. 41, 2925–2928.
Herz, S., Wungsintaweekul, J., Schuhr, C.A., Hecht, S., Luttgen,
ă
H., Sagner, S., Fellermeier, M., Eisenreich, W., Zenk, M.H.,


Ó FEBS 2002

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

Deoxyxylulose phosphate reductoisomerase (Eur. J. Biochem. 269) 4457

Bacher, A. & Rohdich, F. (2000) Biosynthesis of terpenoids:
ygbB protein converts 4-diphosphocytidyl-2C-methyl-D-erythritol
2-phosphate to 2C-methyl-D-erythritol 2,4-cyclodiphosphate.
Proc. Natl Acad. Sci. USA 97, 2486–2490.

Takagi, M., Kuzuyama, T., Kaneda, K., Watanabe, H., Dairi, T.
& Seto, H. (2000) Studies on the nonmevalonate pathway:
formation of 2-C-methyl-D-erythritol 2,4-cyclodiphosphate from
2-phospho-4-(cytidine
5¢-diphospho)-2-C-methyl-D-erythritol.
Tetrahedron Lett. 41, 3395–3398.
´
Seemann, M., Campos, N., Rodrı´ guez-Concepcion, M., Hoeffler,
J.-F., Grosdemange-Billiard, C., Boronat, A. & Rohmer, M.
(2002) Isoprenoid biosynthesis via the methylerythritol phosphate
pathway: accumulation of 2-C-methyl-D-erythritol 2,4-cyclodiphosphate in gcpE deficient mutant of Escherichia coli. Tetrahedron Lett. 43, 775–778.
´
´
Seemann, M., Campos, N., Rodrı´ guez-Concepcion, M., Ibanez,
E., Duvold, T., Tritsch, D., Boronat, A. & Rohmer, M. (2002)
Isoprenoid biosynthesis in Escherichia coli via the methylerythritol
phosphate pathway: enzymatic conversion of methylerythritol
cyclodiphosphate into a phosphorylated derivative of (E)-2methylbut-2-ene-1,4-diol. Tetrahedron Lett. 43, 1413–1415.
Wolff, M., Seemann, M., Grosdemange-Billiard, C., Tritsch, D.,
´
Campos, N., Rodrı´ guez-Concepcion, M., Boronat, A. &
Rohmer, M. (2002) Isoprenoid biosynthesis via the methylerythritol phosphate pathway. (E)-4-Hydroxy-3-methylbut-2-enyl
diphosphate: chemical synthesis and formation from methylerythritol cyclodiphosphate by cell-free system from Escherichia coli.
Tetrahedron Lett. 43, 2555–2559.
Hintz, M., Reichenberg, A., Altincicek, B., Bahr, U., Gschwind,
R.M., Kollas, A.-K., Beck, E., Wiesner, J., Eberl, M. & Jomaa, H.
(2001) Identification of (E)-4-hydroxy-3-methylbut-2-enyl pyrophosphate as a major activator for human cd T cells in Escherichia
coli. FEBS Lett. 509, 317–322.
Hecht, S., Eisenreich, W., Adam, P., Amslinger, S., Kis, K.,
Bacher, A., Arigoni, D. & Rohdich, F. (2001) Studies on the

nonmevalonate pathway to terpenes: the role of the GcpE (IspG)
protein. Proc. Natl Acad. Sci. USA 98, 14837–14842.
Rohdich, F., Hecht, S., Krieger, C., Amslinger, S., Arigoni, D.,
Bacher, A. & Eisenreich, W. (2002) Studies on the nonmevalonate
pathway to terpenes: metabolic role of IspH (LytB) protein. Proc.
Natl Acad. Sci. USA 99, 1158–1163.
Kuzuyama, T., Shimizu, T., Takahashi, S. & Seto, H. (1998)
Fosmidomycin, a specific inhibitor of 1-deoxy-D-xylulose
5-phosphate reductoisomerase in the nonmevalonate pathway for
terpenoid biosynthesis. Tetrahedron Lett. 39, 7913–7916.
Jomaa, H., Wiesner, J., Sanderbrand, S., Altincicek, B.,
Weidemeyer, C., Hintz, M., Turbachova, I., Eberl, M., Zeidler, J.,
ă
Lichtenthaler, H.K., Soldati, D. & Beck, E. (1999) Inhibitors of
the nonmevalonate pathway of isoprenoid biosynthesis as antimalarial drugs. Science 285, 1573–1576.
Taylor, S.V., Vu, L.D. & Begley, T.P. (1998) Chemical and
enzymatic synthesis of 1-deoxy-D-xylulose-5-phosphate. J. Org.
Chem. 63, 2375–2377.
Hoeffler, J.-F., Pale-Grosdemange, C. & Rohmer, M. (2000)
Chemical synthesis of enantiopure 2-C-methyl-D-erythritol 4phosphate, the key intermediate in the mevalonate-independent
pathway for isoprenoid biosynthesis. Tetrahedron 56, 1485–1489.
Biellmann, J.-F. & Jung, M.J. (1971) Mechanism of the alcohol
dehydrogenases from yeast and horse liver. Eur. J. Biochem. 19,
130–134.
Still, W.C., Khan, M. & Mitra, A. (1978) Rapid chromatographic
technique for preparative separations with moderate resolution.
J. Org. Chem. 43, 2923–2925.

36. Bischofberger, N., Waldmann, H., Saito, T., Simon, E.S., Lees,
W., Bednarski, M.D. & Whitesides, G.M. (1988) Synthesis of

analogues of 1,3-dihydroxyacetone phosphate and glyceraldehyde
3-phosphate for use in studies of fructose-1,6-diphosphate aldolase. J. Org. Chem. 53, 3457–3465.
37. Hall, R.H. & Khorana, H.G. (1954) Nucleoside polyphosphates.
II. A synthesis of uridine-5¢-di- and triphosphate. J. Am. Chem.
Soc. 76, 5056–5060.
38. Carlsen, P.H.J., Katsuki, T., Martin, V.S. & Sharpless, K.B.
(1981) A greatly improved procedure for ruthenium tetraoxide
catalyzed oxidations of organic compounds. J. Org. Chem. 46,
3936–3938.
39. Bradford, M.M. (1976) A rapid and sensitive method for the
quantitation of microgram quantities of protein utilizing the
principle of protein-dye binding. Anal. Biochem. 72, 248–254.
40. Leloir, L.F. & Cardini, C.E. (1957) Characterisation of phosphorus compounds by acid lability. Methods Enzymol. 3, 840–
850.
41. Keinan, E., Hafeli, E.K., Seth, K.K. & Lamed, R. (1986) Thermostable enzymes in organic synthesis. 2. Asymmetric reduction of
ketones with alcohol dehydrogenase from Thermoanaerobium
brockii. J. Am. Chem. Soc. 108, 162–169.
42. Dumas, R., Biou, V., Halgand, F., Douce, R. & Duggleby, R.G.
(2001) Enzymology, structure, and dynamics of acetohydroxy acid
isomeroreductase. Acc. Chem. Res. 34, 399–408.
43. Bubb, W.A., Berthon, H.A. & Kuchel, P.W. (1995) Tris buffer
reactivity with low-molecular-weight aldehydes: NMR characterization of the reactions of glyceraldehyde 3-phosphate. Bioorg.
Chem. 23, 119–130.
44. Duvold, T., Calı´ , P., Bravo, J.M. & Rohmer, M. (1997)
Incorporation of 2-C-methyl-D-erythritol, a putative isoprenoid
precursor in the mevalonate-independent pathway, into ubiquinone and menaquinone of Escherichia coli. Tetrahedron Lett. 35,
6181–6184.
45. Kuzuyama, T., Takahashi, S., Takagi, M. & Seto, H. (2000)
Characterization of 1-deoxy-D-xylulose 5-phosphate reductoisomerase, an enzyme involved in isopentenyl diphosphate biosynthesis, and identification of its catalytic amino acid residues.
J. Biol. Chem. 275, 19928–19932.

46. Reuter, K., Sanderbrand, S., Jomaa, H., Wiesner, J., Steinbrecher,
I., Beck, E., Hintz, M., Klebe, G. & Stubbs, M.T. (2002) Crystal
structure of 1-deoxy-D-xylulose-5-phosphate reductoisomerase, a
crucial enzyme in the non-mevalonate pathway of isoprenoid
biosynthesis. J. Biol. Chem. 277, 5378–5384.
47. Bock, C.W., Kaufman-Katz, A., Markham, G.D. & Glusker, J.P.
(1999) Manganese as a replacement for magnesium and zinc:
functional comparison of the divalent ions. J. Am. Chem. Soc. 121,
7360–7372.
48. Rohmer, M., Knani, M., Simonin, P., Sutter, B. & Sahm, H.
(1993) Isoprenoid biosynthesis in bacteria: a novel pathway for the
early steps leading to isopentenyl diphosphate. Biochem. J. 295,
517–524.
49. Johnson, A.E. & Tanner, M.E. (1998) Epimerization via carbon–
carbon bond cleavage. L-Ribulose-5-phosphate 4-epimerase as a
masked class II aldolase. Biochemistry 37, 5746–5754.
50. Lee, L.V., Vu, M.V. & Cleland, W.W. (2000) 13C and deuterium
isotope effects suggest an aldol cleavage mechanism for L-ribulose5-phosphate 4-epimerase. Biochemistry 39, 4808–4820.
51. Yajima, S., Nonaka, T., Kuzuyama, T., Seto, H. & Ohsawa, K.
(2002) Crystal structure of 1-deoxy-D-xylulose 5-phosphate
reductoisomerase complexed with cofactors: implications of a
flexible loop movement upon substrate binding. J. Biochem. 131,
313–317.