Biosynthesis of terpenoids: Efficient multistep biotransformation procedures affording isotope-labeled 2C-methyl-D-erythritol 4-phosphate using recombinant 2C-methyl-D-erythritol 4-phosphate Synthase

Article abstract of DOI:10.1021/jo015890v

This paper describes the recombinant expression of the ispC gene of Escherichia coli specifying 2C-methyl-D-erythritol 4-phosphate synthase in a modified form that can be purified efficiently by metal-chelating chromatography. The enzyme was used for the preparation of isotope-labeled 2C-methyl-D-erythritol 4-phosphate employing isotope-labeled glucose and pyruvate as starting materials. The simple one-pot methods described afford numerous isotopomers of 2C-methyl-D-erythritol 4-phosphate carrying ³H, ¹³C, or ¹⁴C from commercially available precursors. The overall yield based on the respective isotope-labeled starting material is approximately 50%.

Full text of DOI:10.1021/jo015890v

7770
J . Org. Chem. 2001, 66, 7770-7775
Biosyn th esis of Ter p en oid s: Efficien t Mu ltistep Biotr a n sfor m a tion
P r oced u r es Affor d in g Isotop e-La beled 2C-Meth yl-D-er yth r itol
4-P h osp h a te Usin g Recom bin a n t 2C-Meth yl-D-er yth r itol
4-P h osp h a te Syn th a se
Stefan Hecht, J uraithip Wungsintaweekul, Felix Rohdich,* Klaus Kis, Tanja Radykewicz,
Christoph A. Schuhr, Wolfgang Eisenreich, Gerald Richter, and Adelbert Bacher
Lehrstuhl fu¨r Organische Chemie und Biochemie, Technische Universita¨t Mu¨nchen, Lichtenbergstrasse 4,
D-85747 Garching, Germany
felix.rohdich@ch.tum.de
Received J uly 4, 2001
This paper describes the recombinant expression of the ispC gene of Escherichia coli specifying
2C-methyl-D-erythritol 4-phosphate synthase in a modified form that can be purified efficiently
by metal-chelating chromatography. The enzyme was used for the preparation of isotope-labeled
2C-methyl-D-erythritol 4-phosphate employing isotope-labeled glucose and pyruvate as starting
materials. The simple one-pot methods described afford numerous isotopomers of 2C-methyl-D-
erythritol 4-phosphate carrying 3H, 13C, or 14C from commercially available precursors. The overall
yield based on the respective isotope-labeled starting material is approximately 50%.
In tr od u ction
chemical synthesis. Moreover, biotransformation technol-
ogy enables the introduction of isotope labels at high yield
and with rigorous regiochemical and stereochemical
control.
The biosynthesis of structurally complex natural prod-
ucts in living systems has several unique features in
comparison to chemical synthesis in the laboratory or
production plant. Chemical synthesis is usually organized
in a stepwise fashion, and individual reaction steps are
plagued by the formation of undesired side products that
must be removed after every step. In contrast, biosyn-
thetic processes occur in very complex cellular environ-
ments comprising thousands of different macromolecular
and low-molecular-weight components. Nevertheless,
individual synthetic processes are characterized by quan-
titative yields (in other words, without undesired side
products) over numerous reaction steps in conjunction
with virtually perfect stereocontrol. The main reasons for
the superior performance of cellular biochemistry com-
pared to conventional synthetic organic chemistry are (i)
the use of virtually perfect catalysts (i.e., enzymes), (ii)
the recycling of reagents (i.e., cofactors), and (iii) the
efficient management of Gibbs free energy gradients.
In principle, it appears possible to implement these
favorable process properties in extracellular synthetic
procedures. A wide variety of enzymes are commercially
available or can be obtained in high yield using recom-
binant DNA technology. Cofactors can be recycled in situ
using appropriate auxiliary enzymes. In such a biomi-
metic setting, the central feature is then the control of
free energy gradients in the absence of cellular energy
production facilities. Sufficiently large free energy gra-
dients can be provided in vitro by enzymatic coupling of
the desired biotransformation to exergonic auxiliary
reactions.
As a practical example, we report procedures for the
one-pot preparation of the branched polyol, 2C-methyl-
D-erythritol 4-phosphate, in an isotope-labeled form. The
compound is the first committed intermediate in the
non-mevalonate pathway of terpenoid biosynthesis,1-7
which supplies the building blocks for the vast majority
of the approximately 30 000 terpenoids known to date
(Figure 1).
The methods reported in this paper afford a large
variety of different isotopomers of the target compound
in high yield by simple and rapid one-pot reactions using
commercially available isotope-labeled glucose and/or
pyruvate samples as starting materials.
Resu lts
The ispC gene8 specifying 2C-methyl-D-erythritol 4-phos-
phate synthase, also named 1-deoxy-D-xylulose 5-phos-
phate reductoisomerase,9 was amplified by PCR from
chromosomal Escherichia coli DNA and was placed under
the control of a T5 promoter and lac operator in the
(1) Takahashi, S.; Kuzuyama, T.; Watanabe, H.; Seto, H. Proc. Natl.
Acad. Sci. U.S.A. 1998, 95, 9879-9884.
(2) Lange, B. M.; Croteau, R. Arch. Biochem. Biophys. 1999, 365,
170-174.
(3) Radykewicz, T.; et al. FEBS Lett. 2000, 465, 157-160.
(4) Proteau, P. J .; Woo, Y.-H.; Williamson, R. T.; Phaosiri, C. Org.
Lett. 1999, 1, 921-923.
(5) J omaa, H.; et al. Science 1999, 285, 1573-1576.
(6) Kuzuyama, T.; Takahashi, S.; Takagi, M.; Seto, H. J . Biol. Chem.
2000, 275, 19928-19932.
The implementation of biotransformation processes for
synthetic organic chemistry is particularly worthwhile
for compounds that are not conveniently accessible by
(7) Arigoni, D.; Eisenreich, W.; Latzel, C.; Sagner, S.; Radykewicz,
T.; Zenk, M. H.; Bacher, A. Proc. Natl. Acad. Sci. U.S.A. 1997, 94,
10600-10605.
(8) Meganathan, R. In Vitamins and Hormones; Litwack, G., Ed.;
Academic Press: San Diego, 2001; Vol. 61, pp 199-201.
(9) Koppisch, A. T.; Blagg, B. S. J .; Poulter, D. T. Org. Lett. 2000, 2,
215-217.
* To whom correspondence should be addressed. Phone: +49-89-
289-13364. Fax: +49-89-289-13363.
10.1021/jo015890v CCC: $20.00 © 2001 American Chemical Society
Published on Web 10/13/2001

Biosynthesis of Terpenoids
J . Org. Chem., Vol. 66, No. 23, 2001 7771
En zym e P r op er ties. The recombinant enzyme sedi-
ments in the analytical ultracentrifuge at an apparent
velocity of 5.0 S (20 °C). Analysis of the sedimentation
data using the program ULTRASCAN indicated a slight
asymmetry of the moving boundary, which suggested
concentration-dependent aggregation.10,11 A van Holde-
Weischet analysis of the data showed a crossover effect
that is characteristic for nonideal systems with a con-
centration dependency of the sedimentation coefficient.10-12
The data are compatible with a homotetrameric protein
that is subject to aggregation under the experimental
conditions used.
In agreement with earlier studies, the enzyme requires
divalent metal ions for catalytic activity.1-6 Manganese
and magnesium stimulate the activity to a similar degree
at concentrations above 5 mM. At concentrations below
1 mM, manganese was more efficient than magnesium.
2C-Methyl-D-erythritol 4-phosphate synthase requires
NADPH as a reducing agent.1-6 About 1% of the activity
was found when NADPH was replaced by NADH (0.5
mM).
The enzyme activity can be monitored photometri-
cally.1 The KM values determined with Mn2+ were 171
and 25 µM for 1-deoxy-D-xylulose 5-phosphate and NAD-
PH, respectively; vmax had a value of 18 µmol min-1 mg-1
corresponding to a turnover number of 13 s-1 per subunit.
Syn th etic P r oced u r es. The recombinant protein was
used for the preparation of 2C-methyl-D-erythritol 4-phos-
phate isotopomers by rapid one-pot methods specifically
designed for the efficient incorporation of 13C or of
radioisotopes (3H, 14C) as labels for biosynthetic and
biological studies.
Five in vitro biotransformation strategies were opti-
mized for different isotope labeling patterns. Procedures
1-4, designed for introducing 13C or 14C, use D-glyceral-
dehyde 3-phosphate prepared from dihydroxyacetone
phosphate or glucose as the starting material. Procedure
5 enables the incorporation of 3H into the 1R position of
4 using 1-deoxy-D-xylulose 5-phosphate (3) as the starting
material.
F igu r e 1. Non-mevalonate pathway of isoprenoid biosynthe-
sis. Dxs, 1-deoxy-D-xylulose 5-phosphate synthase; IspC, 2C-
methyl-D-erythritol 4-phosphate synthase; IspD, 4-diphospho-
cytidyl-2C-methyl-D-erythritol synthase; IspE, 4-diphospho-
cytidyl-2C-methyl-D-erythritol kinase; IspF, 2C-methyl-D-
erythritol 2,4-cyclodiphosphate synthase.
Ta ble 1. Ba cter ia l Str a in s a n d P la sm id s Used in Th is
Stu d y
strain or
plasmid
genotype or
relevant characteristic
reference or
source
E. coli K-12
Ca r bon La bels in P osition s 2 a n d /or 2′ (P r oce-
d u r es 1 a n d 2). Carbon atoms 2 and 2′ of 4 are derived
from C-2 and C-3, respectively, of pyruvate (1; Figure 2).
Hence, 13C label can be introduced into one or both of
these positions (designated by the letters d and e in
Figure 2) using appropriately 13C-labeled pyruvate. More
specifically, [2-13C1]-, [3-13C1]-, and [2,3-13C2]pyruvate
afford [2-13C1]-, [2′-13C1]-, and [2,2′-13C2]4, respectively
(procedure 1). Moreover, 14C-labeled 4 can be prepared
starting with 14C-labeled pyruvate (procedure 2).
The unlabeled glyceraldehyde 3-phosphate required as
the starting material can be obtained by hydrolysis of 7
followed by treatment with commercially available triose
phosphate isomerase. NADPH is recycled using glucose
dehydrogenase with glucose as the substrate.
M15 [pREP4]
Lac, ara, gal, mtl, recA+,
uvr+, [pREP4, lacI, kanr]
RecA1, endA1, gyrA96,
thi-1, hsdR17, supE44,
relA1, lac, [F′, proAB,
lacIqZ∆M15, Tn10 (tetr)]
23
24
XL1-Blue
plasmids
pQE30
high-copy N-terminal
His-tag vector
Qiagen
pQEYAEMECO expression construct for
this study
the ispC gene of E. coli
expression plasmid pQE30 (Table 1). The recombinant
gene specified the N-terminal sequence motif MRGSH-
HHHHHGS followed by amino acid residue 1 of the E.
coli open reading frame.
Recombinant E. coli cells carrying that plasmid ex-
pressed the ispC gene in very high yield. 2C-Methyl-D-
erythritol 4-phosphate synthase represented approxi-
mately 50% of the total protein in the recombinant E.
coli cells after induction with isopropyl thio-D-galacto-
pyranoside (IPTG).
The reaction sequence starting with 7 can be conve-
niently carried out as a one-pot reaction involving four
enzymes. The reaction progress can be monitored in real
time by 13C NMR analysis. Due to the specific 13C
labeling, the NMR analysis selectively detects the reac-
The recombinant protein could be purified very ef-
ficiently by affinity chromatography on Ni2+-chelating
Sepharose affording homogeneous protein. The apparent
subunit molecular mass was 43 kDa as shown by SDS-
PAGE in agreement with the calculated mass of 43.3 kDa.
(10) Demeler, B.; Saber, H.; Hansen, J . C. Biophys. J . 1997, 72, 397-
407.
(11) Hansen, J . C.; Lebowitz, J .; Demeler, B. Biochemistry 1994, 33,
13155-13163.
(12) Van Holde, K. E.; Weischet, W. O. Biopolymers 1978, 17, 1387-
1403.

7772 J . Org. Chem., Vol. 66, No. 23, 2001
Hecht et al.
Ta ble 2. NMR Da ta of 2C-Meth yl-D-er yth r itol 4-P h osp h a te Isotop om er s (10% D2O)
chemical shifts (ppm) coupling constants (Hz)
high-intensity 13C signalsd
position
1H
13C
31P
C1
C2
11.7a
3.9,a 39.6b
C3
C4
C2′
P
[2-13C]-4
[3,4,5-13C3]-4
[U-13C5]-4
1
2
2′
3
4
3.50
66.2
74.1
18.4
73.6
64.5
143.0a
s
s
s
m
d
m
m
39.6b
1.04
3.64
3.77
4.01
127.0a
135.4a
42.5b
6.5c
4.4c
dd
dd
42.5b 141.3a
146.2a
P
7.0
s
m
m
a
d
1H-13C coupling constants. b 13C-13C coupling constants. c 13C-31P coupling constants. s, singlet; d, doublet; dd, double doublet;
ddd, double double doublet; m, multiplet.
tants (pyruvate, intermediates, and product), whereas the
unlabeled components such as cofactors, buffer compo-
nents, and proteins do not contribute appreciably to the
NMR spectrum. The product, 4, can be purified by
chromatography on ECTEOLA cellulose. The enzyme-
mediated biotransformation is essentially quantitative,
but the workup procedure used is only capable of recov-
ering about 50% of the product in highly purified form.
Ca r bon La bels in P osition s 1, 2, 2′, 3, a n d /or 4
(P r oced u r e 3). 2C-Methyl-D-erythritol 4-phosphate with
carbon labels in positions 1, 2, 2′, 3, and/or 4 can be
obtained from 13C-labeled glucose in conjunction with 13C-
labeled pyruvate (Figure 2). The fate of different precur-
sor atoms is indicated in Figure 2 by lowercase letters
(a-e). For any desired labeling pattern, the appropriately
labeled starting materials can be selected easily on the
basis of Figure 2. The reaction is performed in two
sequential steps without isolation of intermediate 3.
NADPH is recycled in situ using glucose dehydrogenase.
To avoid isotopic dilution by the glucose used as the
substrate for NADPH regeneration, the glycolytic en-
zymes (A and C-F) and 1-deoxy-D-xylulose 5-phosphate
synthase (G) operative in the first reaction step must be
deactivated by heat treatment after the first reaction
phase. As an example, the preparation of [U-13C5]4 is
reported in detail in the Experimental Section.
Ca r bon La bels in P osition s 1, 3, a n d /or 4 (P r oce-
d u r e 4). 2C-Methyl-D-erythritol 4-phosphate with carbon
labels in position 1, 3, or 4 can be obtained from
appropriately labeled glucose as the starting material.
ATP consumed in the glycolysis sequence can be regener-
ated using phosphoenolpyruvate and pyruvate kinase
(Figure 2). Pyruvate generated in situ from phospho-
enolpyruvate in the ATP-recycling step can serve as
endogeneously generated substrate for 1-deoxy-D-xylulose
5-phosphate synthase.
The reaction sequence is performed in two sequential
steps without isolation of intermediates. Initially, labeled
glucose (9) is converted into 1-deoxy-D-xylulose 5-phos-
phate (3). Only 2 equiv of phosphoenolpyruvate, just
sufficient for the complete conversion of labeled 9 into 2,
should be used in this step. In the subsequent reduction
and isomerization step, a recycling system comprising
unlabeled glucose and glucose dehydrogenase is used for
NADPH regeneration. If phosphoenolpyruvate (from the
first step) is still present at this reaction stage, unlabeled
glucose will enter the glycolytic pathway and thus will
cause an isotopic dilution in the product. The product 4
is stable under the reaction conditions. 13C NMR spectra
recorded at several stages of the one-pot reaction are
shown in Figure 3, which indicates the near-quantitative
conversion of 13C-labeled 9 into the final product, [1,3,4-
13C3]4.
Tr itiu m in P osition 1 (P r oced u r e 5). Tritium can
be introduced into position 1 of 4 by reduction of 3 with
tritium-labeled NADPH, which can be obtained in situ
using glucose dehydrogenase and [1-3H]D-glucose as the
tritium source. As shown recently, the hydrogen of
NADPH is introduced stereospecifically into the 1R
position of 4.3,4 Any 13C isotopomer of 3 can be converted
into the corresponding 2C-methyl-D-erythritol 4-phos-
phate isotopomer containing tritium in position 1.
Discu ssion
2C-Methyl-D-erythritol 4-phosphate synthase of E. coli
had been reported earlier to have a specific activity of
0.5 s-1 with Co2+ as the cofactor.6 The N-terminally
modified enzyme described in this study has a ca. 26-
fold higher catalytic activity with Mg2+ or Mn2+ as the
cofactor. Co2+ was a poor activator of enzyme activity
with our enzyme. The different enzyme properties might
be related to the N-terminal modification.
Isotope-labeled precursors have played a central role
in the exploration of the non-mevalonate pathway of
terpenoid biosynthesis and have the potential to contrib-
ute still further to the elucidation of the missing reaction
steps and their mechanisms.7,13-17
Due to the versatility of our synthetic procedures, a
wide variety of different isotopomers of 2C-methyl-D-
erythritol 4-phosphate can be prepared from commer-
cially available isotope-labeled glucose and pyruvate as
starting materials. Our one-pot procedures make use of
cofactor-recycling systems, which facilitates workup and
product purification. The procedures can be scaled up
for the preparation of gram amounts of the desired
isotopomer. Furthermore, for the preparation of 4 labeled
with 14C or 3H at high specific activity, the procedures
can be scaled down linearly to very small volumes.
The yields of the enzyme-catalyzed reaction sequences
are almost quantitative as shown by the series of NMR
spectra in Figure 3 (although significant losses occur
during workup and purification). The driving forces of
the reaction are (i) the release of carbon dioxide in the
(13) Broers, S. T. J . Ph.D. Thesis, Eidgeno¨ssische Technische
Hochschule, Zu¨rich, Switzerland, 1994.
(14) Schwarz, M. K. Ph.D. Thesis, Eidgeno¨ssische Technische Hoch-
schule, Zu¨rich, Switzerland, 1994.
(15) Herz, S.; Wungsintaweekul, J .; Schuhr, C. A.; Hecht, S.;
Lu¨ttgen, H.; Sagner, S.; Fellermeier, M.; Eisenreich, W.; Zenk, M. H.;
Bacher, A.; Rohdich, F. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 2486-
2490.
(16) Lu¨ttgen, H.; Rohdich, F.; Herz, S.; Wungsintaweekul, J .; Hecht,
S.; Schuhr, C. A.; Fellermeier, M.; Sagner, S.; Zenk, M. H.; Bacher,
A.; Eisenreich, W. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 1062-1067.
(17) Rohdich, F.; Wungsintaweekul, J .; Fellermeier, M.; Sagner, S.;
Herz, S.; Kis, K.; Eisenreich, W.; Bacher, A.; Zenk, M. H. Proc. Natl.
Acad. Sci. U.S.A. 2000, 96, 11758-11763.

Biosynthesis of Terpenoids
J . Org. Chem., Vol. 66, No. 23, 2001 7773
F igu r e 3. Real-time 13C NMR analysis of a reaction mixture
(procedure 4). The detailed reaction conditions are given in
the Experimental Section. 1H-Decoupled 13C NMR spectra
were measured at the beginning of the reaction (A), after
conversion of precursors to [3,4,5-13C3]3 at 12 h after the start
of the reaction (B), and at the end of the reaction sequence
(C). Color codes of signals refer to the respective reaction
component indicated.
formation of 3, (ii) the recycling of NADPH, and (iii)
under certain conditions (procedure 4), the consumption
of the pyruvate generated in the ATP regeneration
reaction by the 1-deoxy-D-xylulose 5-phosphate synthase
reaction. J ointly, these exergonic partial reactions afford
a Gibbs free energy gradient that is large enough to drive
the main reaction virtually to completion.
By judicious selection of the auxiliary reactions, it is
also possible to streamline the composition of the final
reaction mixture with regard to product purification. In
the reaction sequences described above, the only side
products formed in stoichiometric amounts are gluconate,
pyruvate, and bicarbonate. Moreover, ethanol is formed
stoichiometrically in procedure 1, and 2 equiv of ADP is
formed in procedure 3 where the recycling of ADP would
result in the dilution of the isotope-labeled pyruvate used
as the substrate.
It is obvious that all reaction sequences described in
this article can be performed with nonlabeled precursors.
Procedures 1 and 4 are best suited for the preparation
of unlabeled 4.
We have previously reported the chemical synthesis
of 4 that was comprised of 14 reaction steps with an
overall yield of 9%.18 The method had been designed to
allow the incorporation of tritium into position 1 of 4.
F igu r e 2. Synthesis of 13C-labeled 2C-methyl-D-erythritol
4-phosphates. A, hexokinase; B, pyruvate kinase; C, glucose
6-phosphate isomerase; D, fructose 6-phosphate kinase; E,
aldolase; F, triose phosphate isomerase; G, 1-deoxy-D-xylulose
5-phosphate synthase; H, 2C-methyl-D-erythritol 4-phosphate
synthase; I, glucose dehydrogenase; the transfer of 13C or 14C
labels from starting material to product is indicated by
lowercase letters (a-e).
(18) Kis, K.; Wungsintaweekul, J .; Eisenreich, W.; Zenk, M. H.;
Bacher, A. J . Org. Chem. 2000, 65, 587-592.

7774 J . Org. Chem., Vol. 66, No. 23, 2001
Hecht et al.
The resulting [1-3H]4 played a crucial role in the iden-
tification of the ispD gene of the non-mevalonate path-
way.15 Two more synthetic procedures for the preparation
of 4 have been published so far that allow the introduc-
tion of stable isotope labels.9,19 However, in general,
chemical syntheses of multiply labeled compounds need
specific strategies and numerous reaction and isolation
steps in contrast to the universality of our enzymatic
approach.
In comparison to chemical syntheses, the procedures
in this paper have the combined advantages of (i) a high
yield, (ii) short reaction times, (iii) access to a virtually
unlimited variety of 13C- or 14C-isotopomers from com-
mercially available precursors, (iv) virtually perfect ste-
reocontrol, (v) the simplicity of the one-pot reaction
conditions in aqueous solution, (vi) the possibility of real-
time monitoring of the reaction progress by NMR, (vi)
the use of commercially available enzymes for all but two
reaction steps, and (vii) the exclusive use of starting
materials that are commercially available.
It should be mentioned that these favorable aspects
of “long shot” in vitro biotransformation are not in any
way limited to the biosynthesis of terpenoids. Similar
synthetic strategies could be used for a wide variety
of natural products in an isotope-labeled or unlabeled
form. The rapidly growing body of genomic data from a
wide variety of organisms, the advanced state of re-
combinant DNA technology, and the specific design of
recombinant enzymes for ease of purification provide a
steadily improving framework for the generalization of
this approach.
enzymes. The ligation mixture was electroporated into E. coli
XL1-Blue and M15 [pREP4] cells affording the recombinant
strains XL1-pQEYAEMECO and M15-pQEYAEMECO, re-
spectively.
Sequ en ce Deter m in a tion . DNA sequencing was per-
formed by the automated dideoxynucleotide method.21 N-
Terminal peptide sequences were obtained by the pulsed-liquid
mode.
P u r ifica tion of Recom bin a n t 2C-Meth yl-D-er yth r itol
4-P h osp h a te Syn th a se fr om E. coli. Recombinant E. coli
M15-pQEYAEMECO cells (2 g) were suspended in 25 mL of
buffer A (100 mM Tris hydrochloride (pH 8.0), 0.5 M sodium
chloride) containing 20 mM imidazole hydrochloride, 25 mg
of lysozyme, and 2.5 mg of DNase I. The mixture was
incubated at 37 °C for 30 min, cooled on ice, and subjected to
ultrasonic treatment. The suspension was centrifuged at
15 000 rpm for 30 min. The supernatant was loaded on a Ni2+
-
chelating Sepharose column (2 × 8 cm) at a flow rate of 3 mL/
min, which had been equilibrated with 20 mM imidazole in
buffer A. The column was developed with a linear gradient of
20-500 mM imidazole in buffer A (total volume ) 300 mL).
Fractions were combined and concentrated by ultrafiltration.
Assa y of 2C-Met h yl-D-er yt h r it ol 4-P h osp h a t e Syn -
th a se. 2C-Methyl-D-erythritol 4-phosphate synthase was as-
sayed according to Takahashi et al.1
An a lytica l Ultr a cen tr ifu ga tion . The concentrated protein
from the purification step described above was loaded on top
of a Superdex 75 HR 26/60 column, which had been equili-
brated with 100 mM Tris hydrochloride (pH 8.0) containing 1
mM dithiothreitol, 0.02% sodium azide, and 50 mM sodium
chloride at a flow rate of 3 mL/min. 2C-Methyl-D-erythritol
4-phosphate synthase was eluted to a volume of 132 mL. The
peak fraction with an absorbance of about 1 OD at 280 nm
was used as a sample for analytical ultracentrifugation.
Hydrodynamic studies were performed with an analytical
ultracentrifuge Optima XL-1 (Beckman Coulter) equipped with
UV and interference optics. Experiments were performed with
double-sector cells equipped with aluminum centerpieces and
sapphire windows. Partial specific volumes and buffer densi-
ties were estimated according to published procedures.22
P r ep a r a tion of [2-13C1]2C-Meth yl-D-er yth r itol 4-P h os-
p h a te (P r oced u r e 1). A solution containing 440 mg (660
µmol) of dimeric dihydroxyacetone phosphate acetal (7) in
10 mL of water was added to a suspension of Dowex 50 WX8
(5 mL, H+ form). The mixture was incubated at 65 °C for 2 h.
The solid was filtered off and washed with water. The
combined solution was lyophilized. Dihydroxyacetone phos-
phate glass (242 mg, containing about 92 mg (0.33 mmol) of
8) was dissolved in 2.5 mL of water, and the pH was adjusted
to 8.0 by the addition of 8 M sodium hydroxide. A solution
(1.5 mL) containing 450 mM Tris hydrochloride (pH 8.0), 30
mM magnesium chloride, 8.5 mg (20 µmol) of thiamine
diphosphate, 1.5 mg (10 µmol) of dithiothreitol, and 363 mg
(330 µmol) of [2-13C1]pyruvate (1, sodium salt) was added. The
pH was adjusted to 8.0, and 88 units (15 µg) of triose phosphate
isomerase and 9.6 units (3.0 mg) of 1-deoxy-D-xylulose 5-phos-
phate synthase were added. The mixture was incubated at 37
°C for 11 h. NADPH (317 mg, 350 µmol) and 2C-methyl-D-
erythritol 4-phosphate synthase (1 mg) were added, and
incubation was continued at 37 °C for 2 h. The product was
purified by chromatography on ECTEOLA cellulose as de-
scribed below, yielding 43.6 mg (201 µmol, 61%).
Exp er im en ta l Section
Ma ter ia ls. The following materials were obtained from the
sources indicated in parentheses: 13C-labeled compounds
(Isotec, Miamisburg, OH); [1-3H]glucose (Amersham Pharma-
cia Biotech, Freiburg, Germany); sodium [2-14C]pyruvate
(NEN, Boston, MA); oligonucleotides (MWG Biotech, Ebers-
berg, Germany); T4 ligase (Gibco-BRL, Eggenstein, Germany);
Sepharose Q FF, Superdex 75 HR 26/60, and restriction
enzymes (Amersham Pharmacia Biotech); chemicals and
enzymes (Sigma Chemicals, Deisenhofen, Germany); DNase
I (Roche Diagnostics, Mannheim, Germany); Taq polymerase
and IPTG (Eurogentec, Seraing, Belgium); RNase A and
Nucleosil 10SB (16 × 250 mm) (Macherey & Nagel, Du¨ren,
Germany); ECTEOLA 23 cellulose (Fluka, Deisenhofen, Ger-
many); TLC plates (Merck, Darmstadt, Germany); and Ni2+
-
chelating Sepharose and Dowex ion exchangers (Amersham
Pharmacia, Braunschweig, Germany). 1-Deoxy-D-xylulose
5-phosphate was prepared as described elsewhere.20
P r ot ein s. The preparation of recombinant 1-deoxy-D-
xylulose 5-phosphate synthase from Bacillus subtilis has been
reported elsewhere.20
Micr oor ga n ism s a n d P la sm id s. Bacterial strains and
plasmids used in this study are summarized in Table 1.
Con str u ction of a n Exp r ession P la sm id . The ispC gene
of E. coli coding for 2C-methyl-D-erythritol 4-phosphate syn-
thase (GenBank accession no. AE000126) was amplified from
base pair position 9887 to 11 083 by PCR using the oligo-
nucleotides ECOYAEM1 (5′-ggaggatccatgaagcaactcacc-3′) and
ECOYAEM2 (5′-gcgcgactctctgcagccgg-3′) as primers and chro-
mosomal E. coli DNA as the template. The amplificate was
digested with the restriction endonucleases BamHI and PstI.
The fragment was ligated into the expression plasmid vector
pQE30, which had been digested with the same restriction
P u r ifica tion of 2C-Meth yl-D-er yth r itol 4-P h osp h a te
(Gen er a l P r oced u r e). Crude reaction mixtures obtained as
described above were lyophilized. The crude product was
(21) Sanger, F.; Nicklen, S.; Coulson, A. R. Proc. Acad. Natl. Sci.
U.S.A. 1977, 74, 5463-5468.
(22) Lane, T. M.; Shah, B. D.; Rigeway, R. M.; Pelltier, S. L. In
Analytical Ultracentrifugation in Biochemistry and Protein Science;
Harding, S. E., Rowe, A. J ., Horton, J . C., Eds.; Royal Society of
Chemistry: Cambridge, 1992; pp 90-125.
(23) Zahmenhof, P. J .; Villarejo, M. J . Bacteriol. 1972, 110, 171-
178.
(24) Bullock, W. O.; Fernandez, J . M.; Short, J . M. BioTechniques
1987, 5, 376-379.
(19) Fontana, A. J . Org. Chem. 2001, 66, 2506-2508.
(20) Hecht, S.; Kis, K.; Eisenreich, W.; Amslinger, S.; Wungsin-
taweekul, J .; Herz, S.; Rohdich, F.; Bacher, A. J . Org. Chem. 2001,
66, 3948-3952.

Biosynthesis of Terpenoids
J . Org. Chem., Vol. 66, No. 23, 2001 7775
dissolved in water (1 g/100 mL). The solution was placed on a
column of ECTEOLA 23 cellulose (5 × 30 cm), which was
developed with a linear gradient of 0 to 0.3 M triethylammo-
nium acetate (total volume ) 4 L) at a flow rate of 1.5 mL/
min. Fractions (10 mL) were collected and analyzed by TLC
using Silica NH-Sil plates, which were developed with a
mixture of 2-propanol/ethyl acetate/water (6:1:3, v/v). The
product was identified as a brown spot after staining with a
mixture of vanillin/sulfuric acid (10 g/L). Fractions were
combined and lyophilized. The chemical purity of 2C-methyl-
D-erythritol 4-phosphate (triethylammonium salt) was ap-
Nucleosil 10 SB (16 × 250 mm) using 0.5 M formic acid as the
eluent at a flow rate of 13 mL/min. The effluent was monitored
refractometrically. The product was eluted at 14.5 min. Frac-
tions were combined and lyophilized, yielding 86 mg (0.89
mmol, 50%). The chemical purity of 2C-methyl-D-erythritol
4-phosphate was > 97%, as estimated by 1H NMR spectros-
copy.
P r ep a r a t ion of [1,3,4-13C3]2C-Met h yl-D-er yt h r it ol 4-
P h osp h a te (P r oced u r e 4). A solution containing 150 mM
Tris hydrochloride (pH 8.0), 10 mM magnesium chloride, 1.0
g (5.4 mol) of [U-13C6]glucose, 0.23 g (1.5 mmol) of dithiothrei-
tol, 0.3 g (0.7 mmol) of thiamine diphosphate, 0.1 g (0.2 mmol)
of ATP (disodium salt), and 2.2 g (11 mmol) of potassium
phosphoenolpyruvate in a total volume of 300 mL was adjusted
to pH 8.0 by the addition of 8 M sodium hydroxide. Hexokinase
(1500 units, 10 mg), 300 units (400 µg) of phosphoglucose
isomerase, 105 units (1 mg) of fructose 6-phosphate kinase,
610 units (100 µg) of triose phosphate isomerase, 57 units (10
µg) of aldolase, 400 units (2.8 mg) of pyruvate kinase, and 8
units (3 mg) of 1-deoxy-D-xylulose 5-phosphate synthase were
added. The solution was incubated at 37 °C for 24 h. Subse-
quently, 2.0 g (11 mmol) of glucose, 200 mg (200 µmol) of
NADP+, 100 units (1.8 mg) of glucose dehydrogenase, and 39
units (5.7 mg) of 2C-methyl-D-erythritol 4-phosphate synthase
were added, and incubation at 37 °C was continued for 18 h.
The product was purified by the general method described
above, yielding 1.0 g (4.6 mmol, 41%).
1
proximately 90%, as estimated by H NMR spectroscopy.
P r ep a r a tion of [2-14C]2C-Meth yl-D-er yth r itol 4-P h os-
p h a te (P r oced u r e 2). A solution containing 260 mM Tris
hydrochloride (pH 8.0), 10 mM magnesium chloride, 80 µg
(0.50 µmol) of dithiothreitol, 0.4 mg (0.9 µmol) of thiamine
diphosphate, 1.0 mg (4.8 µmol) of dihydroxyacetone phosphate
(lithium salt), 0.5 mg (4.8 µmol) of [2-14C]pyruvate (sodium salt,
10.4 µCi/µmol), and 250 units (30 µg) of triose phosphate
isomerase in a total volume of 450 µL was incubated at 37 °C
for 15 min. A solution of 0.5 units (0.2 mg) of 1-deoxy-D-
xylulose 5-phosphate synthase from B. subtilis in 50 µL of 100
mM Tris hydrochloride (pH 8.0) was added, and the mixture
was incubated for 2 h. Subsequently, a solution containing 100
mM Tris hydrochloride (pH 8.0), 10 mM magnesium chloride,
1.8 mg of NADPH (2.0 µmol), and 0.1 mg (0.9 units) of
recombinant 2C-methyl-D-erythritol 4-phosphate synthase in
a
total volume of 200 µL was added. The mixture was
P r ep a r a t ion of (1R)-[1-3H ]2C-Met h yl-D-er yt h r it ol 4-
P h osp h a te (P r oced u r e 5). A solution containing 100 mM
Tris hydrochloride (pH 8.0), 10 mM magnesium chloride, 0.16
mg (0.20 µmol) of NADP+, 1 µg (6 nmol) of [1-3H1]D-glucose
(120 µCi), 5.0 µg (23 nmol) of 1-deoxy-D-xylulose 5-phosphate,
2 units (36 µg) of glucose dehydrogenase, and 0.2 units (29
µg) of 2C-methyl-D-erythritol 4-phosphate synthase in a total
volume of 301 µL was incubated at 37 °C for 1 h. The mixture
was applied to a column of Dowex 1X8 (0.5 × 5.5 cm, formate
form), which was washed with 6 mL of water followed by
10 mL of 1 M formic acid. The effluent was lyophilized
affording 68 µCi of (1R)-[1-3H]2C-methyl-D-erythritol 4-phos-
phate (radiochemical yield ) 57%).
incubated at 37 °C for 30 min. Aliquots (10 µL) were applied
to an HPLC column of Nucleosil 10SB (4.6 × 250 mm) that
was developed with 10 mM ammonium formate in 0.2 M formic
acid at a flow rate of 1 mL/min. The effluent was monitored
by a solid-state scintillation counter. [2-14C]2C-Methyl-D-
erythritol 4-phosphate was eluted at 16 mL. Fractions were
combined and lyophilized affording 20 µCi of [2-14C]2C-methyl-
D-erythritol 4-phosphate (radiochemical yield ) 40%).
P r ep a r a tion of [U-13C5]2C-Meth yl-D-er yth r itol 4-P h os-
p h a te (P r oced u r e 3). A reaction mixture containing 150 mM
Tris hydrochloride (pH 8.0), 10 mM magnesium chloride, 44
mg (0.10 mmol) of thiamine diphosphate, 1.02 g (1.79 mmol)
of ATP, 0.17 g (0.89 mmol) of [U-13C6]glucose, 200 mg (1.79
mmol) of [2,3-13C2]pyruvate, 410 units (0.05 mg) of triose
phosphate isomerase, 360 units (3.6 mg) of hexokinase, 50
units (0.07 mg) of phosphoglucose isomerase, 20 units (3.2 µg)
of fructose 6-phosphate kinase, 35 units (6.0 µg) of aldolase,
and 2 units (0.6 mg) of recombinant 1-deoxy-D-xylulose 5-phos-
phate synthase from B. subtilis in a total volume of 58 mL
was incubated at 37 °C overnight, while the pH was held at a
constant value of 8.0 by the addition of 1 M sodium hydroxide.
The reaction was stopped by adding 3 mL of 2 M hydrochloric
acid. The precipitate was removed by centrifugation, and the
pH of the solution was readjusted to 8.0. A solution (72 mL)
containing 0.97 g (5.4 mmol) of glucose, 100 mg (0.1 mmol) of
NADP+, 10 units (1.4 mg) of 2C-methyl-D-erythritol 4-phos-
phate synthase, and 120 units (2.2 mg) of glucose dehydroge-
nase was added. The mixture was incubated at 37 °C over-
night. Aliquots (5 mL) were applied to an HPLC column of
NMR. 1H NMR and 1H-decoupled 13C NMR spectra were
1
recorded at 500.1 and 125.6 MHz for H and 13C, respectively.
The chemical shifts were referenced to external trimethyl-
silylpropane sulfonate. 31P NMR spectra were recorded at a
frequency of 101.3 MHz. 31P chemical shifts were referenced
to external 85% H3PO4. 2C-Methyl-D-erythritol 4-phosphate
spectra were recorded in 90% H2O/10% D2O (pH 8.0, uncor-
rected glass electrode reading).
Ack n ow led gm en t . This work was supported by
the Deutsche Forschungsgemeinschaft, the Fonds der
Chemischen Industrie, and the Hans-Fischer-Gesell-
schaft. We thank Katrin Ga¨rtner and Ingrid Obersteiner
for their skillful technical assistance.
J O015890V