Biochemical characterization of
Bacillus subtilis
type II isopentenyl
diphosphate isomerase, and phylogenetic distribution of isoprenoid
biosynthesis pathways
Ralf Laupitz
1
, Stefan Hecht
1
, Sabine Amslinger
1
, Ferdinand Zepeck
1
, Johannes Kaiser
1
, Gerald Richter
1
,
Nicholas Schramek
1
, Stefan Steinbacher
2
, Robert Huber
3
, Duilio Arigoni
4
, Adelbert Bacher
1
,
Wolfgang Eisenreich
1

subtilis chromosome extending from bp 184 997–186 043
with similarity to the idi-2 gene of Streptomyces sp. CL190
specifying type II isopentenyl diphosphate isomerase was
expressed in a recombinant Escherichia coli strain. The
recombinant protein with a subunit mass of 39 kDa was
purified to apparent homogeneity by column chromatog-
raphy. The protein was shown to catalyse the conversion of
dimethylallyl diphosphate into isopentenyl diphosphate and
vice versa at rates of 0.23 and 0.63 lmolÆmg
)1
Æmin
)1
,
respectively, as diagnosed by
1
H spectroscopy. FMN and
divalent cations are required for catalytic activity; the highest
rates were found with Ca
2+
. NADPH is required under
aerobic but not under anaerobic assay conditions. The
enzyme is related to a widespread family of (S)-a–hydroxy-
acid oxidizing enzymes including flavocytochrome b
2
and
L
-lactate dehydrogenase and was shown to catalyse the
formation of [2,3-
13
C

oids, hopanoids), predator repulsion and pollinator or mate
attraction [1].
Despite their enormous structural and functional com-
plexity, all terpenoids are assembled from two simple
precursors, isopentenyl diphosphate (IPP) and dimethylallyl
diphosphate (DMAPP) (Fig. 1). The biosynthesis of these
universal terpene precursors via the mevalonate pathway
has been studied in considerable detail in yeast and animals.
These classical studies established the formation of IPP
from three acetate moieties via mevalonate (reviewed in
[2–5]). IPP is then converted into DMAPP by an isopentenyl
diphosphate isomerase which is essential in all organisms
using the mevalonate pathway (reviewed in [6,7]).
The elucidation of the mevalonate pathway culminated in
the development of the statin type drugs which inhibit
3-hydroxy-3-methylglutaryl-CoA reductase and reduce car-
diovascular morbidity and mortality by reduction of blood
cholesterol levels and probably also by down-regulation
of inflammatory processes [8,9]. Certain statins such as
LipitorÒ and ZocorÒ are record holders with regard to
current drug sales.
A second isoprenoid biosynthesis pathway starting with
1-deoxy-
D
-xylulose 5-phosphate has been discovered in the
last decade (reviewed in [10–14]). The linear carbohydrate
precursor is transformed into a branched polyol derivative,
2C-methyl-
D
-erythritol 4-phosphate [15] which is further

Whereas type I isomerases only require divalent cations
for catalytic activity, the type II isomerase of Streptomyces
sp. CL190 has been reported to require FMN and NADPH
as well as divalent metals [28]. The structure of a type II IPP
isomerase from Bacillus subtilis has been elucidated by
X-ray crystallography [29]. This paper reports on the
biochemical properties of the recombinant enzyme from
B. subtilis. Phylogenetic patterns of IPP isomerases in the
archaeal and eubacterial kingdoms with respect to the two
IPP/DMAPP biosynthesis pathways were analysed by
bioinformatic methods.
Experimental procedures
Materials
IPP, DMAPP and [3,4,5-
13
C
3
]DMAPP were prepared
by published procedures [30,31]. [U-
13
C
3
]acetone and
[2,3-
13
C
2
]pyruvate were obtained from Isotec (Miamisburg,
OH, USA). Restriction enzymes were purchased from New
England Biolabs (Frankfurt, Germany). Oligonucleotides

kanamycin (50 mgÆL
)1
). Cultures were incubated at 37 °C
with shaking. At an optical density of 0.7 (600 nm),
isopropyl thio-b-
D
-galactoside was added to a final concen-
tration of 2 m
M
, and the culture was incubated for 5 h. The
cells were harvested by centrifugation, washed with 0.9%
(w/v) sodium chloride, and stored at )20 °C under anaer-
obic conditions.
The following steps were carried out under anaerobic
conditions. Frozen cell mass (4 g) was thawed in 38 mL of
100 m
M
Tris hydrochloride, pH 8.0, containing 0.5
M
sodium chloride and 20 m
M
imidazole hydrochloride. The
suspension was passed through a French press and was then
centrifuged. To the supernatant (60 mL), 40 mL of water
were added, and the mixture was applied to a column
of Ni-chelating Sepharose FF (column volume, 11 mL;
Amersham Pharmacia Biotech) which had been equili-
brated with 100 m
M
Tris hydrochloride, pH 8.0, containing

100 m
M
Tris hydrochloride, pH 8.0, 10 m
M
MgCl
2
,10l
M
FMN, 2 m
M
sodium acetate, 10.8 m
M
DMAPP or IPP, and
protein. The mixtures were incubated at 37 °C under
Fig. 1. Biosynthesis of IPP and DMAPP.
Ó FEBS 2004 B. subtilis type II IPP isomerase (Eur. J. Biochem. 271) 2659
anaerobic conditions. The reaction was terminated by the
addition of EDTA to a final concentration of 26 m
M
.After
the addition of D
2
O to a final concentration of 10% (v/v),
the samples were analysed by NMR spectroscopy.
Assay of lactate dehydrogenase activity
Assay mixtures containing 100 m
M
Tris hydrochloride,
pH 8.0, 17 m
M

Alto, CA) equipped with ultraviolet and interference
optics. Experiments were performed with double sector
cells equipped with aluminum centerpieces and sapphire
windows. Partial specific volumes and buffer densities
were estimated according to published procedures [35].
Samples contained 100 m
M
Tris hydrochloride, pH 8.0.
Mass spectrometry
Mass spectra were recorded with a Biflex III MALDI-TOF
mass spectrometer from Bruker Instruments, Karlsruhe,
Germany. Samples contained 25 m
M
Tris hydrochloride,
pH 8.0, 33% (v/v) CH
3
CN, saturated a-cyanohydroxycin-
namic acid, 0.1% (v/v) trifluoracetic acid and 0.7 mg of IPP
isomerase per mL.
Bioinformatics
Similarity searches in the GenBank database of completed
and unfinished prokaryotic genomes (among them not yet
specifically assigned genomes) (.
gov) were performed with the programs
BLASTP
and
TBLASTN
using the gapped
BLAST
and

idi-2
gene from
B. subtilis
An open reading frame (Acc. no. P50740) extending from
bp position 184 997–186 043 on the B. subtilis chromosome
with similarity to the idi-2 gene of Streptomyces sp. CL190
[28] (37% identical amino acid residues; Fig. 2A) was
amplified by PCR and was cloned into the plasmid pQE30
affording the recombinant plasmid pQEidi2 (see Experi-
mental procedures). An E. coli strain carrying this plasmid
produced copious amounts of a 39 kDa protein as judged
by SDS electrophoresis (Fig. 3).
The recombinant protein was purified by affinity chro-
matography on Nickel-chelating sepharose and appeared
homogeneous as judged by SDS/PAGE (Fig. 3). Partial
N-terminal Edman degradation afforded the amino acid
sequence MRGSHHHHHHGSVTRAE in agreement with
the sequence of the recombinant gene. MALDI-TOF mass
spectrometry showed a relative mass of 38 463 Da in good
agreement with the calculated mass of 38 455 Da (data not
shown).
Hydrodynamic studies on Idi-2 protein of
B. subtilis
X-ray structure analysis of the B. subtilis enzyme in the
presence of FMN indicated a D
4
symmetric homooctamer
structure with a relative mass of 309 kDa [29]. In order to
check for the potential influence of substrates and cofactors
on the quaternary structure of the enzyme, we performed

H NMR assignments of DMAPP shown in
Table 1 were confirmed by two-dimensional NOESY
experiments indicating strong NOE interactions between
the methyl signal at 1.79 p.p.m. (E-methyl group) and the
signal at 5.47 p.p.m. (methine group). It should be noted
that some confusion with respect to these assignments reigns
in the literature. Whereas the correct
1
H NMR assignments
are given in the text of [41], reversed assignments of the
2660 R. Laupitz et al. (Eur. J. Biochem. 271) Ó FEBS 2004
methyl signals are given in footnote 26 of that paper and in
[28]. When the enzyme was incubated with IPP as substrate
under aerobic conditions in the presence of NADPH and
Fig. 2. Amino acid residues essential for
functionality of Idi-2 protein. (A) Amino acid
sequence comparison of Idi-2 proteins.
Sequences included in this analysis were
B. subtilis Idi-2 protein and Idi-2 proteins
from major human pathogens. Residues
absolutely conserved in all Idi-2 amino acid
sequences available in the GenBank database
are labelled by open triangles. Residues
involved in FMN binding (as found in the
crystallographic structure of the B. subtilis
protein, see below) are shown by filled
triangles. (B) Stereo representation of the
FMN-binding site of B. subtilis Idi-2 protein.
The disordered regions between Met256 and
Phe263 and Tyr211 and Arg226, respectively,

4 4.88 111.3 71, 3
5 1.80 21.4 41, 3
Dimethylallyl diphosphate
1 4.49 6.6 6.6
2 5.48 7.1
3 139.7 41, 42
4(E-methyl) 1.79 24.7 42, 4
5(Z-methyl) 1.75 17.0 41, 4
a
Referenced to external trimethylsilylpropane sulfonate;
b
observed
with [3,4,5-
13
C
3
]DMAPP and [2,3-
13
C
3
]IPP.
Ó FEBS 2004 B. subtilis type II IPP isomerase (Eur. J. Biochem. 271) 2661
FMN, we observed the appearance of the signals of both
methyl groups and of the methine group of the enzymat-
ically formed DMAPP (Fig. 4A). Concomitantly, the
signals of IPP were progressively diminished. Using acetate
as an internal standard, the signal integrals afforded the
concentrations of IPP and DMAPP as a function of time
(Fig. 5).
Figure 4B illustrates the reverse reaction, i.e. the conver-

13
C
3
]DMAPP as
Fig. 4.
1
H-NMR assay of type II IPP isomerase from B. subtilis.
A, part of the
1
H NMR spectrum of the reaction mixture (lower lane)
obtained from IPP (
1
H NMR signals, see upper lane) by the catalytic
action of Idi-2 protein under aerobic conditions. B, part of the
1
H
NMR spectrum of the reaction mixture (lower lane) obtained from
DMAPP (
1
H NMR signals, upper lane) by the catalytic action of Idi-2
protein under aerobic conditions. Assay mixtures contained 100 m
M
Tris hydrochloride, pH 8.0, 10 m
M
MgCl
2
,1m
M
dithiothreitol,
2.5 m

M
sodium acetate, and 10.8 m
M
IPP or DMAPP.
j, formation of DMAPP from IPP; s, formation of IPP from
DMAPP.
Table 2. Catalytic rates of Idi-2 protein under different conditions.
Reaction mixtures contained MgCl
2
and were prepared as described
under Experimental procedures.
Procedure/condition
Specific activity
(lmolÆmin
)1
Æmg
)1
)
Conversion of IPP into DMAPP
Aerobic
a
0.63 ± 0.042
b
Anaerobic 0.62 ± 0.037
b
Conversion of DMAPP into IPP
Aerobic
a
0.23 ± 0.007
b

mixtures contained NADH.
2662 R. Laupitz et al. (Eur. J. Biochem. 271) Ó FEBS 2004
substrate. The decrease of the
13
C-coupled signals of the
Z-andE-methyl groups resonating at 17.0 and
24.7 p.p.m., respectively, as well as that of the quaternary
carbon atom resonating at 139.7 p.p.m. was accompanied
by the appearance of three new
13
C-coupled signals at
143.4, 111.3 and 21.4 p.p.m. assigned as the carbon atoms
3, 4, and 5 of IPP, respectively (cf. Table 1 and Fig. 6).
Within the limits of experimental accuracy the catalytic
rates determined with this assay were the same as those
described above.
Whereas NADPH or NADH was required for catalytic
activity under aerobic conditions, the reaction could
proceed without NADPH under anaerobic conditions using
enzyme which had been purified under anaerobic condi-
tions. FMN, however, was required under aerobic as well as
under anaerobic conditions. The reaction rates were similar
under aerobic and anaerobic conditions (Table 2). Photo-
metric analysis gave no evidence for reduction of FMN in
aerobic or anaerobic assay mixtures (data not shown). The
recombinant enzyme has an absolute requirement for a
divalent metal ion for catalytic activity; the highest rates
were found with Ca
2+
(Table 3). A different order for the

Fig. 2) in the direct neighborhood of the FMN binding site
are also absolutely conserved (Fig. 2B).
Type II isomerases are restricted to the archaeal and
eubacterial kingdoms. With the exception of Halobacterium
sp. NRC-1, Mycobacterium marinum and Photorhabdus
luminescens featuring both a putative idi-1 and a putative
idi-2 gene, the distribution of type I and type II isomerases in
the prokaryotic kingdom appears to be mutually exclusive.
Genes specifying type II isomerases were found in 19 of 20
(95%) archaebacterial species. Nanoarchaeum equitans is
devoid of IPP biosynthesis as well as of idi genes. In the
group of 263 eubacterial genomes studied, 35 (13%) carry
an idi-1 gene,72(27%)carryanidi-2 gene, 2 carry both an
idi-1 and idi-2 gene, and 154 (59%) appear to be devoid of
IPP isomerases.
Phylogenetic analyses of 102 type II isomerases were
performed as described under Experimental procedures.
The final consensus phylogenetic tree (majority rule) shows
the major phylogenetic grouping of 76 type II isomerases in
the archaeal and eubacterial kingdoms as illustrated in
Fig. 7. Bacillales and Lactobacillales form a cluster which is
separated from other lineages with statistical relevance
(bootstrap value: 100%). Some actinobacteria (Streptomy-
ces sp. CL190, Kitasatospora griseola and Actinoplanes sp.
A40644) group also within this cluster. The separation of
the archaeabacterial from the eubacterial kingdom was not
found to be statistically relevant (bootstrap values < 50%).
In the eubacterial kingdom, Cyanobacteria with the
exception of Crocosphaera watsonii and Synechocystis sp.
PCC6803 (which group together with the sulfur bacterium

,1m
M
dithiothreitol, 2.5 m
M
NADPH, 10 l
M
FMN, and 5.2 m
M
[3,4,5-
13
C
3
]DMAPP.
Table 3. Activation of Idi-2 protein by divalent cations. Reaction mix-
tures were prepared as described under Experimental Procedures.
Metal ions were added to a final concentration of 10 m
M
.
Metal ion Relative Activity (%)
Ca
2+
100
Mg
2+
65
Mn
2+
17
Zn
2+

genes (Dichelobacter nodosus, Legionella pneumophila and
P. luminescens,allthreec-proteobacteria; and the Spiro-
chete Borrelia burgdorferi). Interestingly, the idi-2 gene of
P. luminescens, whose genome specifies the enzymes of the
deoxyxylulose phosphate pathway together with both the
Fig. 7. Consensus cladogram of Idi-2 proteins
from various microorganisms. The simplified
tree (majority rule) was deduced by Neighbor-
joining analysis based on the alignment of the
amino acid sequences of 76 Idi-2 proteins.
Gaps were removed from the alignment, and
the total number of positions taking into
account was 320. The numbers at the nodes
are the statistical confidence estimates
computed by the bootstrap procedure. Only
groups with bootstrap probablity values
>50% were retained. The bar represents 0.137
PAM distance.
2664 R. Laupitz et al. (Eur. J. Biochem. 271) Ó FEBS 2004
idi-1 as well as the idi-2 genes, is interrupted by a
counterclockwise located transposase gene effectively
knocking it out.
Type I isomerases are found preferably in the Actino-
bacteria group (Corynebacterium sp., Mycobacterium sp.
and Streptomyces sp.), but also in the Bacteroidetes group
(Cytophaga hutchinsonii),andsomeinthea-subgroup
(Rhodobacter sphaeroides and Silicibacter pomeroyi)andin
the c-subgroup (Coxiella burnetii, Erwinia carotovora,
Escherichia sp., Klebsiella pneumoniae, P. luminescens,
Salmonella and Shigella sp.) of the proteobacteria group.

and animals. On the other hand, all completely sequenced
eukaryotic genomes comprise putative orthologs of the type
I isomerase. Highest degrees of similarity were found to
Idi-1 proteins of the c-proteobacteria A. vinelandii and
P. luminescens and to Idi-1 protein of the Bacteroid
C. hutchinsonii (expect values 3e-22, 5e-19 and 8e-20,
respectively).
Paralogs of Idi-2 proteins
Database searches conducted with the
BLASTP
program
retrieved a considerable number of proteins with
substantially lower similarity (Expect value >1e-10) to the
B. subtilis Idi-2 protein which was used as search motif.
Notably, type II isomerase shows weak, but significant
similarity with a family of FMN dependent (S)-a-hydroxy-
acid dehydrogenases (pfam database accession no.
PF01070) including flavocytochrome b
2
from yeast
[EC 1.1.2.3 (FCb2)] [42], long chain hydroxyacid oxidase
from mammals [43], glycolate oxidase from spinach
[EC 1.1.3.15 (GOX)] [44],
L
-lactate dehydrogenases from
bacteria (EC 1.1.1.27) [45] (S)-mandelate dehydrogenase
from P. putida (MD) [46] and inosine 5¢-monophosphate
dehydrogenases (EC 1.1.1.205) [47] over the entire length of
their respective sequences (Fig. 9A) (expect values 0.002,
1.1, 0.083, 2e-06, 0.47 and 0.001, respectively).

Thus, under structural aspects, Idi-2 appears as a fairly
distant relative of the FMN-dependent a-hydroxyacid-
oxidizing enzymes. However, the sequence similarity in
conjunction with the TIM barrel fold and the conserved
FMN binding site leave no doubt about the evolutionary
relatedness of Idi-2 with the dehydrogenase superfamily.
Fig. 8. Distribution of isoprenoid biosynthesis pathways and IPP
isomerases in 283 completed and unfinished prokaryotic genomes. MEV,
mevalonate pathway genes; DXP, deoxyxylulose pathway genes.
Ó FEBS 2004 B. subtilis type II IPP isomerase (Eur. J. Biochem. 271) 2665
L
-Lactate dehydrogenase activity of Idi-2 protein
from
B. subtilis
Partial sequence similarity of the Idi-2 gene and the
paralogous lldD gene specifying
L
-lactate dehydrogenase
together with similarities in the TIM-barrel fold and
FMN binding site of the two respective proteins promp-
ted a search for the presence of a residual redox activity
in type II isomerase. In order to obtain maximum
sensitivity in combination with maximum selectivity, we
used [2,3-
13
C
2
]pyruvate as substrate. Using NADH as
cofactor, we observed the formation of [2,3-
13

genase was added based on a sequence
alignment to glycolate oxidase. Secondary
structures and sequence numbers refer to gox
(top lines) and idi (bottom lines), respectively.
The eight ba modules of the TIM-barrel
domainarecoloredyellowandnamedS1toS8
and H1 to H8, respectively. Active site residues
of the a-hydroxyacid dehydrogenase family are
shown in red, the characteristic
NHG[GA]RQL-motif is boxed (note that it is
not conserved in IDI proteins). FMN-binding
residues are coloured blue and residues con-
served in IDI proteins located close to FMN in
green. The similarity between the a-hydroxy-
acid dehydrogenase and Idi-2 proteins is most
pronounced in the C-terminal half which har-
bours the standard phosphate binding site
(SPB). B, stereo-view of the superposition of
Idi-2 (N-terminal extension in yellow, TIM-
barrel in grey, C-terminal extension in green)
and glycolate dehydrogenase (N-terminal
extensionindarkblue,TIM-barrelinlight
blue, C-terminal extension in purple). Secon-
dary structure elements of the TIM-barrel
superimpose very well, especially for modules
b7/a7andb8/a8 which harbor the standard
phosphate binding site (SPB). Deviations are
found for the N- and C-terminal extensions.
FMNisdepictedinorange(Idi-2)andgreen
(GOX). In addition, the GOX structure dis-

Streptomyces enzyme. However, when the enzyme is
purified and assayed under anaerobic conditions, NADPH
is not required. The catalytic activities observed with IPP as
substrate are similar under aerobic and anaerobic condi-
tions, in the range of 0.6 lmolÆmg
)1
Æmin
)1
. No evidence was
obtained for redox cycling of FMN. The amino acid
residues involved in the FMN binding site are absolutely
conserved throughout a large number of orthologs
(Figs 2,9). This suggests an essential role for FMN despite
the low affinity of the enzyme for that cofactor and the
apparent absence of a redox process as part of the catalytic
cycle. The substrate binding site of the type II enzyme
remains veiled. However, a patch of absolutely conserved
amino acid residues comprising the polar amino acid
residues H147, N149, Q152 and E153 in close proximity
to the FMN binding site suggests that the substrates could
be bound in close proximity to the isoalloxazine moiety. In
the absence of direct evidence for the involvement of a redox
process it is tempting to postulate that the cofactor might act
as a dipole stabilizing a cationic intermediate or transition
state of the reaction. A similar role has been postulated for
tryptophan 121 in the case of E. coli Idi-1 [51].
During the preparation of this manuscript three groups
reported the catalytic properties of recombinant type II IPP
isomerases from B. subtilis [52]andthetwoArchaea
Methanothermobacter thermoautothrophicus [53] and Sulf-

lateral gene transfer similar to that reported for 3-hydroxy-
3-methylglutaryl coenzyme A reductase [55,56]. It is inter-
esting in this context that both types of isomerases are found
in the Actinobacteria group. The anomalous positions for
some eubacterial species (e.g. Cyanobacteria and Actino-
bacteria) observed here (cf. Figure 7) may be explained by a
loss of evolutionary constraints due to nonessential func-
tions of Idi-2 proteins in bacteria using the deoxyxylulose
phosphate pathway.
With regard to the complex distribution of the two
different terpenoid pathways and of the two different
isomerase types in the eubacterial kingdom, it is relevant to
emphasize that certain highly pathogenic Gram-positive
cocci including Enterococcus and Staphylococcus species use
type II isomerases in conjunction with the mevalonate
pathway which has an absolute requirement for isomeriza-
tion of IPP in order to generate DMAPP. Hence, the type II
isomerase is an essential enzyme in this group of human
pathogens.
Enterococci and Staphylococci have a dramatic history of
resistance development against virtually all currently avail-
able antibiotics. Most notably, many strains are multidrug
resistant, and the rapidly spreading resistance against
vancomycin and methicillin constitutes a life-threatening
problem in affected patients [57]. Clearly, there is an urgent
requirement for novel therapeutic strategies directed at
these microorganisms. As the human type I IPP isomerase
and the type II isomerase of the microorganisms mentioned
have no detectable similarity, it should be possible to
develop inhibitors for the bacterial enzyme which have little

Ó FEBS 2004 B. subtilis type II IPP isomerase (Eur. J. Biochem. 271) 2667
7. Koyama, T. & Ogura, K. (1999) Isopentenyl diphosphate, iso-
merase and prenyltransferases. In Comprehensive Natural Product
Chemistry (Cane, D., ed.), Vol. 2, pp. 69–96. Pergamon, Oxford.
8. Slater, E.E. & MacDonald, J.S. (1988) Mechanism of action and
biological profile of HMG CoA reductase inhibitors. A new
therapeutic alternative. Drugs 36, 72–82.
9. Stancu, C. & Sima, A. (2001) Statins: mechanism of action and
effects. J. Cell Mol. Med. 5, 378–387.
10. Schwarz, M. & Arigoni, D. (1999) Gingkolide biosynthesis. In
Comprehensive Natural Product Chemistry (Cane, D., ed.), Vol. 2,
pp. 367–399. Pergamon, Oxford.
11. Rohmer, M. (1999) A mevalonate-independent route to isopen-
tenyl diphosphate. In Comprehensive Natural Product Chemistry
(Cane, D., ed.), Vol. 2, pp. 45–68. Pergamon, Oxford.
12. 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.
13. Eisenreich, W., Rohdich, F. & Bacher, A. (2001) Deoxyxylulose
phosphate pathway to terpenoids. Trends Plant Sci. 6,78–84.
14. Rohdich, F., Hecht, S., Bacher, A. & Eisenreich, W. (2003)
Deoxyxylulose phosphate pathway of isoprenoid biosynthesis.
Discovery and function of ispDEFGH genes and their cognate
enzymes. Pure Appl. Chem. 75, 393–405.
15. Takahashi, S., Kuzuyama, T., Watanabe, H. & Seto, H. (1998) A
1-deoxy-
D
-xylulose 5-phosphate reductoisomerase catalyzing the
formation of 2-C-methyl-

-erythritol 2,4-cyclodiphosphate.
Proc.NatlAcad.Sci.USA97, 2486–2490.
19. 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.NatlAcad.Sci.USA98, 14837–14842.
20. Kollas, A.K., Duin, E.C., Eberl, M., Altincicek, B., Hintz, M.,
Reichenberg, A., Henschker, D., Henne, A., Steinbrecher, I.,
Ostrovsky, D.N., Hedderich, R., Beck, F., Jomaa, H. & Wiesner,
J. (2002) Functional characterization of GcpE, an essential
enzyme of the non-mevalonate pathway of isoprenoid biosynth-
esis. FEBS Lett. 532, 432–436.
21. Wolff, M., Seemann, M., Grosdemange-Billiard, C., Tritsch, D.,
Campos, N., Rodriguez-Concepcion, M., Boronat, A. & Rohmer,
M. (2002) Isoprenoid biosynthesis via the methylerythritol phos-
phate pathway. (E)-4-Hydroxy-3-methylbut-2-enyl diphosphate:
chemical synthesis and formation from methylerythritol cyclo-
diphosphate by a cell-free system from Escherichia coli.
Tetrahedron Lett. 43, 2555–2559.
22. Rohdich, F., Hecht, S., Ga
¨
rtner, K., Adam, P., Krieger, C.,
Amslinger,S.,Arigoni,D.,Bacher,A.&Eisenreich,W.(2002)
Studies on the nonmevalonate terpene biosynthetic pathway:
metabolic role of IspH (LytB) protein. Proc. Natl Acad. Sci. USA
99, 1158–1163.
23. Adam, P., Hecht, S., Eisenreich, W., Kaiser, J., Gra
¨
wert, T.,
Arigoni, D., Bacher, A. & Rohdich, F. (2002) Biosynthesis

H. (2001) An unusual isopentenyl diphosphate isomerase found in
the mevalonate pathway gene cluster from Streptomyces sp. strain
CL190. Proc. Natl Acad. Sci. USA 98, 932–937.
29. Steinbacher, S., Kaiser, J., Gerhardt, S., Eisenreich, W., Huber,
R., Bacher, A. & Rohdich, F. (2003) Crystal structure of the type
II isopentenyl diphosphate: dimethylallyl diphosphate isomerase
from Bacillus subtilis. J. Mol. Biol. 329, 973–982.
30. Davisson, V., Jo, Woodside, A.B., Neal, T.R., Stremler, K.E.,
Mu
¨
hlbacher, M. & Poulter, C.D. (1986) Phosphorylation of iso-
prenoid alcohols. J. Org. Chem. 51, 4768–4779.
31. Christensen, D.J. & Poulter, C.D. (1994) Enzymatic synthesis of
isotopically labeled isoprenoid diphosphates. Bioorg.Med.Chem.
2, 631–637.
32. Bullock, W.O., Fernandez, J.M. & Short, J.M. (1987) XL1-Blue: a
high efficiency plasmid transforming recA Escherichia coli strain
with beta-galactosidase selection. Biotechniques 5, 376–379.
33. Zamenhof, P.J. & Villarejo, M. (1972) Constructions and prop-
erties of Escherichia coli strains exhibiting a-complementation of
b-galactosidase fragments in vivo. J. Bacteriol. 110, 171–178.
34. Sanger, F., Nicklen, S. & Coulson, A.R. (1977) DNA sequencing
with chain-terminating inhibitors. Proc. Acad. Natl Sci. USA 74,
5463–5468.
35. Laue, T.M., Shah, B.D., Ridgeway, T.M. & Pelletier, S.L. (1992)
Computer-aided interpretation of analytical sedimentation data
for proteins. In Analytical Ultracentrifugation in Biochemistry
and Polymer Science (Harding, S.E., Rowe, A.J. & Horton,
J.C., eds), pp. 90–125. Royal Society of Chemistry, Cambridge,
UK.

Molecular cloning and nucleotide sequence of cDNA encoding rat
kidney long-chain
L
-2-hydroxyacid oxidase. Expression of the
catalytically active recombinant protein as a chimaera. Eur. J.
Biochem. 214,17–25.
44. Lindqvist, Y. & Branden, C.I., (1989) The active site of spinach
glycolate oxidase. J. Biol. Chem. 264, 3624–3628.
45. Garvie, E.I. (1980) Bacterial lactate dehydrogenases. Microbiol.
Rev. 44, 106–139.
46. Sukumar, N., Xu, Y., Gatti, D.L., Mitra, B. & Mathews, F.S.
(2001) Structure of an active soluble mutant of the membrane-
associated (S)-mandelate dehydrogenase. Biochemistry 40, 9870–
9878.
47. Whitby, F.G., Luecke, H., Kuhn, P., Somoza, J.R., Huete-Perez,
J.A., Phillips, J.D., Fletterick, R.J. & Wang, C.C. (1997) Crystal
structure of Tritrichomonas foetus inosine-5¢-monophosphate
dehydrogenase and the enzyme-product complex. Biochemistry 36,
10666–10674.
48. Stenberg, K. & Lindqvist, Y. (1997) Three-dimensional structures
of glycolate oxidase with bound active-site inhibitors. Protein Sci.
6, 1009–1015.
49. Tegoni, M. & Cambillau, C. (1994) The 2.6-A refined structure of
the Escherichia coli recombinant Saccharomyces cerevisiae flavo-
cytochrome b
2
-sulfite complex. Protein Sci. 3, 303–313.
50. Sukumar, N., Dewanti, A.R., Mitra, B. & Mathews, F.S. (2004)
High resolution structures of an oxidized and reduced flavopro-
tein: The water switch in a soluble form of (S)-mandelate dehy-

58. Kuzmic, P. (1996) Program DYNAFIT for the analysis of enzyme
kinetic data: application to HIV proteinase. Anal. Biochem. 237,
260–273.
Supplementary material
The following material is available from ck
wellpublishing.com/products/journals/suppmat/EJB/EJB4194/
EJB4194sm.htm
Table S1. Isoprenoid biosynthesis in completed and unfin-
ished prokaryotic genomes.
Ó FEBS 2004 B. subtilis type II IPP isomerase (Eur. J. Biochem. 271) 2669