Sterol biosynthesis by a prokaryote: First in vitro identification of the genes encoding squalene epoxidase and lanosterol synthase from Methylococcus capsulatus

Article abstract of DOI:10.1271/bbb.70331

Sterol biosynthesis by prokaryotic organisms is very rare. Squalene epoxidase and lanosterol synthase are prerequisite to cyclic sterol biosynthesis. These two enzymes, from the methanotrophic bacterium Methylococcus capsulatus, were functionally expressed in Escherichia coli. Structural analyses of the enzymatic products indicated that the reactions proceeded in a complete regio- and stereospecific fashion to afford (3S)-2,3-oxidosqualene from squalene and lanosterol from (3S)-2,3-oxidosqualene, in full accordance with those of eukaryotes. However, our result obtained with the putative lanosterol synthase was inconsistent with a previous report that the prokaryote accepts both (3R)-and (3S)-2,3-oxidosqualenes to afford 3-epi-lanosterol and lanosterol, respectively. This is the first report demonstrating the existence of the genes encoding squalene epoxidase and lanosterol synthase in prokaryotes by establishing the enzyme activities. The evolutionary aspect of prokaryotic squalene epoxidase and lanosterol synthase is discussed.

Full text of DOI:10.1271/bbb.70331

Biosci. Biotechnol. Biochem., 71 (10), 2543–2550, 2007
Sterol Biosynthesis by a Prokaryote: First in Vitro Identification
of the Genes Encoding Squalene Epoxidase and Lanosterol
Synthase from Methylococcus capsulatus
1
1
1;2
Chiaki NAKANO, Akihiro MOTEGI, Tsutomu SATO,
3
1;2;y
Masayuki ONODERA, and Tsutomu HOSHINO
1
Graduate School of Science and Technology, Niigata University, Ikarashi, Niigata 950-2181, Japan
Department of Applied Biological Chemistry, Faculty of Agriculture, Niigata University,
2
Ikarashi, Niigata 950-2181, Japan
3
Department of Applied Chemistry and Biotechnology, Niigata Institute of Technology,
Kashiwazaki 945-1195, Japan
Received May 29, 2007; Accepted July 11, 2007; Online Publication, October 7, 2007
[doi:10.1271/bbb.70331]
Sterol biosynthesis by prokaryotic organisms is very
oxidosqualene, even though the epoxide is employed as
the common substrate. The polycyclization cascade of
(3S)-2,3-oxidosqualene is one of the most complicated
biochemical reactions catalyzed by a single protein. It
leads to the formation of new carbon-carbon bonds with
rare. Squalene epoxidase and lanosterol synthase are
prerequisite to cyclic sterol biosynthesis. These two
enzymes, from the methanotrophic bacterium Methyl-
ococcus capsulatus, were functionally expressed in
Escherichia coli. Structural analyses of the enzymatic
products indicated that the reactions proceeded in a
complete regio- and stereospecific fashion to afford (3S)-
1
,2)
accurate regio- and stereochemical specificities. Tri-
terpenoid cyclases (OSCs) show a high similarity among
a variety of biological species, and recent progress in
1
,2)
2
,3-oxidosqualene from squalene and lanosterol from
3S)-2,3-oxidosqualene, in full accordance with those of
functional analysis of the active sites is remarkable.
(
(3S)-2,3-Oxidosqualene is produced from squalene by
the enzymatic action of squalene epoxidase (SE). Both
SE and OSC are prerequisite to the production of cyclic
sterols in eukaryotes, as shown in Scheme 1, and thus
the two enzymes are essential to eukaryotic species. On
the other hand, bacterial species have no steroid, thus
differing from eukaryotes. However, in 1971, a com-
paratively large amount of steroids was found excep-
tionally in the methanotrophic bacterium Methylococcus
eukaryotes. However, our result obtained with the
putative lanosterol synthase was inconsistent with a
previous report that the prokaryote accepts both (3R)-
and (3S)-2,3-oxidosqualenes to afford 3-epi-lanosterol
and lanosterol, respectively. This is the first report
demonstrating the existence of the genes encoding
squalene epoxidase and lanosterol synthase in prokar-
yotes by establishing the enzyme activities. The evolu-
tionary aspect of prokaryotic squalene epoxidase and
lanosterol synthase is discussed.
3
)
capsulatus (Bath), the chemical structures of which
were later determined to be 4,4-dimethyl-5ꢀ-cholesta-
8
(14),24-dien-3ꢁ-ol,
4,4-dimethyl-5ꢀ-cholesta-8(14)-
4
)
Key words: squalene; oxidosqualene; lanosterol; hopene;
Methylococcus capsulatus
en-3ꢁ-ol and their 4ꢀ-methyl derivatives. There are
some indications that steroids occur at very low levels in
prokaryotes, but it is still doubtful whether steroids are
biosynthesized by prokaryotic organisms. Instead of
sterol, bacteria usually produce pentacyclic triterpenes,
so-called hopanoids, as building blocks in their cell
membranes, produced from squalene by the action of
squalene-hopene cyclase (SHC) (Scheme 1). Very
recently, steroid precursors and/or steroids, viz., lano-
sterol, cycloartenol and their metabolites, were further
isolated from the following prokaryotic cells: Myxo-
5
)
Sterols are ubiquitous components of eukaryotes that
play a key role in controlling fluidity and flexibility of
their cell membranes. Lanosterol and cycloartenol are
steroid precursors; the former is found in vertebrates and
fungi, and the latter in photosynthetic plants. The cyclic
sterols or triterpenes are produced from (3S)-2,3-
oxidosqualene by enzymatic reactions with 2,3-oxidos-
qualene cyclases (OSCs). The cyclase enzymes give rise
to the structural diversity of cyclic triterpene scaffolds
during the polycyclization process of acyclic (3S)-2,3-
6
)
7
)
bacterium species of Stigmatella aurantiaca and
8
)
9)
Nannocystis excedens, Gemmata obscuriglobus, and
y
To whom correspondence should be addressed. Fax: +81-25-262-6854; E-mail: hoshitsu@agr.niigata-u.ac.jp
Abbreviations: SE, squalene epoxidase; OSC, 2,3-oxidosqualene cyclase; SHC, squalene-hopene cyclase

2544
C. NAKANO et al.
H
H
H
H
H
14
Phytosteroids
Steroids
3
4
Lanosterol
HO
Parkeol
OSC
HO
Cycloartenol
HO
OSC-CAS
OSC-LS
H
H
H
O /SE
2
SHC
3
3
3
HO
O
Squalene
3β-Hydroxyhopene
(
3S)-2,3-Oxidosqualene
SHC
H
H
H
H
H
H
SHC
3
3
HO
O
Hopene
(3R)-2,3-Oxidosqualene
3α-Hydroxyhopene
Scheme 1. Biochemical Reactions Leading to Sterols, Which Are Catalyzed by Squalene Epoxidase (SE) and Oxidosqualene Cyclases (OSCs).
Lanosterol synthase (OSC-LS) and cycloartenol synthase (OSC-CAS) are found in eukaryotes and are produced by the polycyclization
reaction of (3S)-2,3-oxidosqualene. On the other hand, prokaryotes usually produce hopene by direct polycyclization reaction of squalene, which
is mediated by squalene-hopene cyclase (SHC). Prokaryotic SHC also catalyzes the reactions of artificial substrates (3R)- and (3S)-2,3-
oxidosqualenes into 3ꢀ- and 3ꢁ-hydroxyhopenes, respectively.18) Parkeol and lanosterol are found in the cells of prokaryotic G. obscuriglobus.9)
Methylosphaera hansonii.10) To establish the occurrence
of steroids in prokaryotes, identification of the genes
responsible for the production of sterol and/or steroids
is indispensable as well as findings of steroid compo-
nents in the cells. In 1998, Tippel et al. reported the
cloning of the shc gene from M. capsulatus, but they did
were recorded with a JEOL SX 102 or a Jms-Q1000GC
K9 mass spectrometers operated at 70 eV. GC analyses
were done on a Shimadzu GC-8A apparatus with a flame
ionization detector. A non-polar DB-1 capillary column
(0:53 mm ꢀ 30 m, J & W Scientific, CA, USA) was
used, and the GC conditions were as follows: injection
1
1)
ꢁ
ꢁ
not detect an osc gene in a hybridization experiment.
temperature, 290 C; column temperature, 270 C; flow
2
Thus the cloning and the characterization of the bacterial
genes responsible for steroid biosynthesis is still
unaccomplished work. The genome project for this
rate of N carrier, 1.0 kg/cm . HPLC was carried out
2
with a Hitachi L-1700 (pump) and an L-7405 (UV
detector), the HPLC peaks having been monitored at
1
2)
ꢁ
strain was completed in 2004. A comparison of the
amino acid alignment with those of known SEs and
OSCs from eukaryotes allowed us to deduce the
presence of SE and OSC in M. capsulatus, but no in
vitro evidence has been given hitherto. Here we provide
proof that sterol pathway is indeed involved in this
bacterium; the enzyme activities of the SE and OSC
were demonstrated using heterologously expressed
enzymes in E. coli. These two enzymatic reactions
proceeded in a regio- and stereospecific fashion identical
to those from eukaryotic species, implying that the two
bacterial genes may have been acquired from ancient
eukaryotes through lateral gene transfer.
210 nm. Specific rotation values were obtained at 25 C
with a Horiba SEPA-300 instrument.
Growth conditions. M. capsulatus (ATCC 33009) was
ꢁ
grown at 30 C under an atmosphere of methane and air
1
3)
(3:7) in a 3.8-liter gas-flask, where 0.7-liter of the
following medium was contained: 1-liter culture me-
dium, adjusted to pH 6.8, contained NaNO3 (2.0 g),
.
.
MgSO4 7H2O (0.2 g), KCl (0.04 g), CaCl2 7H2O
.
.
(0.02 g), NaH2PO4 12H2O (0.53 g), Na2HPO4 2H2O
.
.
(0.12 g), FeSO4 7H2O (1.0 mg), CuSO4 5H2O (5 mg),
H3BO3 (10 mg), MnSO4 4H2O (9.3 mg), ZnSO4 7H2O
(70 mg), and MoO (10 mg).
.
.
3
Materials and Methods
Cloning of the se and the osc from M. capsulatus and
construction of the expression plasmid. Chromosomal
DNA was isolated according to the procedure of Tippel
Analytical method. H-NMR and 13C-NMR spectra
1
1
1)
were measured in a C6D6 or CDCl3 solution with a
et al. The putative se and osc were amplified with
þ
1
Bruker DMX 600 instrument. Chemical shifts in H- and
1
KOD polymerase (Toyobo, Tokyo) from the chromo-
somal DNA with the following primers: for the putative
3
C-NMR spectra (ppm) were relative to 7.28 and
28.0 ppm for the solvent peak of C D , and to 7.26 and
7.0 ppm for the CDCl solution. GC–MS (EI) spectra
0
1
7
se, 5 -CGGGTTACATATGATGAGTTCGATTGAAT-
0 0
6
6
TGGG-3 (sense primer) and 5 -CGCGGATCCTCA-
3

Sterol Biosynthesis by Prokaryotes
2545
0
TTTCGCCCGCAGCGCC-3 (anti-sense primer), NdeI
and BamHI restriction sites being underlined: for the
azole and 0.1% of Triton X-100. The extracts obtained
were subjected to centrifugation at 10,000 g for 20 min.
The resulting supernatants were loaded onto a Ni
agarose column (Qiagen, Hilden, Germany) equilibrated
with a buffer (20 mM of Tris–HCl, pH 7.9, 300 mM of
NaCl, 10 mM of imidazole and 0.1% Triton X-100). SE
with the His tag was eluted with 250 mM imidazole
including 0.1% Triton X-100. The purity was checked
by SDS–PAGE containing 10% acrylamide, and esti-
mated to be approximately > 95%. The purified enzyme
(0.5 mg) was used and incubated with squalene (490 mM)
under the following conditions: NADPH-P450 reductase
140 nM, NADPH 1 mM, FAD 1 mM, Triton X-100 0.5%,
50 mM Tris–HCl (pH 7.5), total volume 5 ml, incubated
0
putative osc, 5 -CCGGGATCCCATGAAGCATCTGC-
0
0
TTTCGC-3 (sense primer), 5 -ATCGGTCGACTCTG-
0
CGATAACCCGTGTCAT-3 (anti-sense primer), Bam-
HI and SalI sites being underlined.
A polymerase chain reaction (PCR) was performed
with these primers and the chromosomal DNA as a
template. In a 50 ml PCR reaction, 300 ng of chromoso-
mal DNA, 0.5 mM of each primer, and 0.2 mM of each
þ
dNTPs in 1 ꢀ KOD buffer (Toyobo) were amplified
þ
ꢁ
with 1U KOD polymerase for 30 cycles (98 C, 1 min;
5
ꢁ ꢁ
0 C, 1 min; 68 C, 7 min). The resulting PCR product
of the putative se (1,350 bp) was digested with NdeI and
BamHI, and then ligated into the NdeI-BamHI large
fragment of pET16b to obtain the expression plasmid.
The PCR-amplified osc gene (2,013 bp) was ligated to
pET-22b using the BamHI and SalI sites. The BamHI-
SalI fragment in the osc gene was transferred from
pET22b into pET32b, because expression and solubility
were somewhat improved by pET32b. These constructs
were transformed into E. coli BL 21 (DE3). The
integrity of the constructs was confirmed by sequencing.
A DNA sequencing reaction was carried out by dideoxy
chain termination method with a Thermo Sequenaseꢀ
Primer Cycle Sequencing Kit (GE Healthcare, Buck-
inghamshire, UK). The automated sequencing was
performed on LIC-4000 (LI-COR).
ꢁ
at 25 C for 20 h.
Enzyme assay of the putative OSC. The culture
conditions of the E. coli (BL21) encoding the putative
OSC were the same as those harboring the putative SE
described above, including IPTG concentration and
culture temperature. To the collected E. coli pellet
(100 ml culture) was added 10 ml of Tris–HCl buffer
solution (100 mM), which was then subjected to ultra-
sonication. The cell-free homogenates thus prepared
were used as the enzyme source. The enzyme reaction
contained a racemic mixture of (3R,S)-2,3-oxidosqua-
lene (470 mM), Triton X-100 (0.5%), MgCl (6.4 mM),
2
mercaptoethanol (8 mM), Tris–HCl (89 mM), and 4 ml of
the cell-free extract in a 5 ml total volume. Incubation
ꢁ
Enzyme assays of the putative SE and incubation
conditions with the cell-free homogenates and with the
purified SE. E. coli BL 21(DE3) encoding the putative
was done for 20 h at 25 C, and quenched by adding
KOH/MeOH. Extraction of lipophilic materials and
removal of the excess of Triton X-100 were carried out
in a way similar to that for SE.
ꢁ
SE was grown at 37 C in 100 ml of an LB medium with
ampicillin (5 mg/100 ml). Expression of the recombi-
nant protein was induced by adding 1 mM of IPTG, when
the optical density at 600 nm reached about 0.5–0.6.
Incubation of 2,3-oxidosqualene with A. acidocaldar-
ius SHC. The plasmid (pET3a) harboring shc was
ꢁ
14)
Cultivation was continued for an additional 5 h at 25 C.
transformed into E. coli BL21(DE3). The cells were
ꢁ
To the pellets collected after the centrifugation was
added 10 ml of 20 mM Tris–HCl buffer solution (pH 7.4)
containing 0.1% Triton X-100, and then this was
subjected to ultrasonication to disrupt cells. The super-
natant obtained by centrifugation was subjected to
enzymatic assay. The reaction mixture contained squa-
lene (490 mM), the homogenous cell-free extract (4 ml),
NADPH (1 mM), FAD (1 mM), NADPH-cytochrome
P450 reductase (140 nM), Triton X-100 (0.5%), and
grown at 30 C (OD ¼ 0:6) in Luria-Bertani (LB)
medium containing 50 mg/liter ampicillin, and then
induced for 8 h with isopropyl-ꢁ-D-thiogalactopyrano-
side (IPTG, final concentration, 0.2 mM). The E. coli,
grown in a 100 ml LB medium, was collected after
centrifugation. To the collected pellets was added 10 ml
of 50 mM Tris–HCl buffer solution (pH 8.0) containing
1.0% Triton X-100, and then the suspension was
subjected to ultrasonication and centrifugation to obtain
the supernatant as the enzyme source. One mg of 2,3-
oxidosqualene, obtained by incubating squalene with the
putative SE of M. capsulatus, was emulsified with the
Tris–HCl in a total volume of 5 ml of 20 mM Tris–HCl
ꢁ
(pH 7.4). After incubation at 25 C for 20 h, the enzyme
reaction was terminated by adding 6 ml of 15% KOH/
ꢁ
MeOH and heating at 75 C for 30 min, and then the
aq. solution of Triton X-100 (20 mg) in H O (0.1 ml). To
2
lipophilic materials were extracted with hexane. After
removing the excess of Triton X-100 by passing the
hexane extract through a short SiO2 column with
hexane:EtOAc (100:20), the eluted fraction was sub-
jected to GC analysis. Purification of the SE was carried
out as follows: The harvested cells were subjected to
ultrasonication on ice in a buffer containing 20 mM of
Tris–HCl (pH 7.9), 300 mM of NaCl, 10 mM of imid-
the solution were added 3.0 ml of 0.1 M citrate buffer
(pH 6.0) and 2.0 ml of the cell-free homogenates (SHC),
ꢁ
which was then incubated at 45 C for 12 h. The
enzymatic reaction was quenched by adding 6 ml of
ꢁ
15% KOH/MeOH and heating it at 70 C. The lipophilic
materials were extracted with hexane (4 ml ꢀ 3). Triton
X-100 included in the hexane-extract was removed by
a short SiO2 column chromatography eluting with

2546
C. NAKANO et al.
A SE
N-terminal hydrophobic region
5
30
1
120
51
GXGXXG motif
DG motif
GD motif
1
1
24
49
164
178
279
207
178
192
293
221
284
303
401
334
314
36
333
431
364
126
139
70
57
B OSC
VXDCTA motif
QW motif
QW motif
QW motif
QW motif
373
381
336
481
453
374
378
386
341
486
458
379
1
8
14
23
46
55
61
70
12
27
99
80
17
114
94
149
126
63
164
141
78
32
QW motif
483
490
441
591
562
470
498
531
546
589
596
547
702
672
576
604
611
562
717
687
591
505
456
606
577
485
538
489
640
614
518
553
504
655
629
533
C M. capsulatus
D G. obscuriglobus
6
50 amino acid
454 amino acid
449 amino acid
670 amino acid
OSC
SE
SE
OSC
AGT
GTA
AGT
GTA
AGTA
AGTA
GTA
AGT
GTCTGCCGT
Fig. 1. Amino Acid Alignment of the Putative Squalene Epoxidases (SE) and Oxidosqualene Cyclases (OSC).
The characteristic sequences of DGXXXXXR and GD motifs in SEs (A) and VXDCTA and QW motifs in OSCs (B), are illustrated. The
conserved and similar residues are highlighted. The blank boxes in (A) show the membrane-spanning region, which was suggested by SOSUI
program (http://bp.nuap.nagoya-u.ac.jp/sosui/sosui submit.html). LS, CAS, and SHC indicate lanosterol synthase, cycloartenol synthase, and
hopene synthase, respectively. Mc, M. capsulatus (accession no. SE: AAU91076, LS: AAU91075); Go, G. obscuriglobus (gnl TIGR 214688
contig: 4355: G. obscuriglobus UQM2246); Sa, Stigmatella aurantiaca (CAS: CAD39196, the se gene has not yet been sequenced); Hs, Homo
sapiens (SE: Q14534, LS: P48449); At, Arabidopsis thaliana (SE: NP 568033, CAS: P38605); Aa, A. acidocaldarius (SHC: BAA25185). McSE
showed identities of 25.2% and 20.4% to GoSE and AtSE, respectively. McLS showed the identities of 44.8%, 23.5% and 35.6% to GoLS, AaSHC
and SaCAS, respectively. (C) Arrangement of the putative OSC and SE found in M. capsulatus, where the translational coupling was found through
four base pairs (LS: MCA2873, start 3068808 – end 3066796, SE: MCA2872, start 3066799 – end 3065450, DDBJ database), and (D) that of the
putative SE and OSC in G. obscuriglobus, both genes of which were also contiguous at a distance of nine base pairs (gnl TIGR 214688 contig:
4355: G. obscuriglobus UQM2246, which includes both SE and OSC; Institute for Genomic Research (TIGR) database; http://www.tigr.org/.
hexane:EtOAc (100:20), and the eluted solution was
subjected to GC analysis.
synthase from A. thaliana, respectively. These eukary-
otic SEs and OSCs were identified by in vitro assay or
gene disruption. Figure 1 shows the amino acid align-
ments of SEs and OSCs derived from various biological
sources, in which the characteristic sequences only are
illustrated. The putative SE had three characteristic
motifs of the FAD-binding domains (Fig. 1A): the
dinucleotide-binding motif of GXGXXG, DG, and
GD motifs, all of which are highly conserved among
eukaryotic SEs. These motifs are also found in the
putative SE, but the GXGXXG motif was replaced by
GXSXXG sequence. The membrane-spanning region,
frequently found in those from yeasts and vertebrates,
was lacking in the putative SE. Generally, eukaryotic
OSCs have two characteristic alignments, called QW
Results and Discussions
The genes of MCA2872 and MCA2873 (DDBJ
database, accession nos. AAU91076 and AAU91075)
were assumed to be SE and OSC, respectively, by
analyses of NCBI BLAST (http://www.ncbi.nlm.nih.
gov/BLAST) and CLUSTALW (http://clustalw.genome.
jp). The putative SE had identities of 16.1% and 19.9%
to Homo sapience and Sacchromyces cerevisiae, re-
spectively. The putative OSC showed 33.0%, 27.9% and
1
5)
28.1% identities to the lanosterol synthases from
H. sapience and S. cerevisiae and to the cycloartenol

Sterol Biosynthesis by Prokaryotes
2547
A
B
A
1
2
2
B
Retention time (min)
Fig. 3. Gas Chromatograms of Hexane-Extracts from Incubation
Mixtures of 2,3-Oxidosqualenes with the SHC from A. acidocal-
darius.
Retention time (min)
Triton X-100 included in the hexane extract was removed by
passing a short SiO2 column eluting with a mixed solvent of hexane/
EtOAc (100:20). Peaks 1 and 2 show 3ꢀ- and 3ꢁ-hydroxyhopene,
respectively. 3ꢀ- and 3ꢁ-Hydroxyhopenes were selectively produced
Fig. 2. GC Traces of Hexane-Extract Obtained by Incubating
Squalene with Cell-Free Homogenates from E. coli Having Putative
SE (B) and Lacking This Gene (A).
Peaks 1 and 2 show squalene and (3S)-2,3-oxidosqualene, re-
spectively.
18)
from (3R)- and (3S)-2,3-oxidosqualene, respectively (Scheme 1).
The substrates used in this enzymatic reaction were as follows: (A) a
chemically synthesized racemic mixture of (3R,S)-2,3-oxidosqua-
lenes; (B) 2,3-oxidosqualene prepared from incubation with the
putative M. capsulatus SE.
motif and VXDCTA motif (Fig. 1B); the role of the
1
6)
former is to reinforce protein architecture, while that
of the latter is to initiate the polycyclization reaction by
1
7)
proton attack on the epoxide ring. The putative OSC
had five QW motifs and a VSDCAA sequence corre-
sponding to VXDCTA motif.
oxidosqualenes to give 3ꢀ- and 3ꢁ-hydroxyhopene,
1
8)
respectively.
As shown in Fig. 3B, no detectable
amount of 3ꢀ-hydroxyhopene was produced, clearly
demonstrating that no (3R)-oxidosqualene was contami-
nated in the incubation mixture. Thus this bacterial SE
produces only the (3S)-form with complete regio- and
stereospecificities in a way similar to eukaryotic SEs.
SDS–PAGE showed that almost all of the expressed
proteins formed an inclusion body, but 4–5% of the
protein was soluble (data not shown). The SE, purified
by Ni agarose affinity column chromatography followed
by SDS–PAGE, showed enrichment of a major protein
at approximately 50 kDa comparable with the predicted
molecular weight of the His10-fused se gene product
(50.8 kDa), and it had enzyme activity under the in-
cubation conditions described in ‘‘Materials and Meth-
ods,’’ affording 2,3-oxidosqualene, which was confirmed
on SiO2–TLC (Rf ¼ 0:4, hexane:EtOAc = 100:5) and
by GC–MS analysis.
Figure 2 shows the GC chromatogram of the hexane-
extract from incubating squalene with the cell-free
extract from E. coli encoding the putative SE. GC–MS
analysis showed that the new peak was 2,3-oxidosqua-
lene, the EI-MS of which was superimposable on that of
1
7)
chemically synthesized 2,3-oxidosqualene:
(
m=z 69
þ
100%), 81 (85), 135 (37), 153 (45), and 426 (5, M ).
Large-scale incubation afforded 12.3 mg after complete
purification with HPLC (hexane:2-PrOH = 100:0.5),
1
13
the H- and C NMR spectra of which were identical
1
7,18) 1
to those of the synthetic one:
H NMR (600 MHz,
C6D6), ꢂH (ppm): 1.22 (3H, s), 1.27 (3H, s), 1.66 (3H, s),
1
2
.68 (3H, s), 1.72 (6H, s), 1.72 (3H, s), 1.79 (3H, s),
.15–2.34 (20H, m), 2.67 (1H, t, J ¼ 6:2 Hz), 5.36 (5H,
20
m), and the specific optical rotation, ½ꢀꢂ
ꢃ1:47 (c
D
20
0
.14, EtOH), was identical to the published value ½ꢀꢂ
19)
D
ꢃ1:4 (c 1.0, EtOH). These data indicate that this
epoxidation reaction proceeds with complete regio- and
stereospecificities in the same way as that of eukaryotic
epoxidases, that is, oxygenation occurs at the terminal
double bond to afford 3S-epoxide selectively. To
establish the stereochemical destiny during this enzy-
matic reaction, oxidosqualene thus obtained was incu-
bated with SHC from A. acidocaldarius, which cata-
lyzes the polycyclization reaction of (3R)- and (3S)-2,3-
The putative osc gene product expressed by the
pET22b construct almost formed an inclusion body in
the E. coli expression system. Replacement of the
BamHI-SalI fragment in the osc gene with pET32b
afforded somewhat improved solubility. This may have
occurred due to enhancement of correct folding of the
thioredoxin-fused protein. The E. coli BL21(DE3) har-
boring pET32b-OSC was grown and the putative OSC
was expressed. SDS–PAGE of the cell-free homoge-

2548
C. NAKANO et al.
A
B
A
1
1
426
2
B
426
Retention time (min)
Retention time (min)
Fig. 4. GC Profiles of the Hexane-Extract Obtained by Incubating a
Racemic Mixture of (3R,S)-2,3-Oxidosqualene with Cell-Free
Extracts, Which Were Obtained from Grown E. coli Having the
Putative OSC (B) and That Lacking the OSC (A).
The Triton X-100 included in the reaction mixture was removed
with a short SiO2 column eluting with hexane:EtOAc (100:20).
Peaks 1 and 2 show (3R,S)-2,3-oxidosqualene and lanosterol,
respectively.
Fig. 5. EI-MS of Authentic Lanosterol (A) and the Enzymatic
Product (B) Obtained by Incubating a Racemic Mixture of (3R,S)-
2
,3-Oxidosqualene with the Putative OSC.
nates showed enrichment of 90 kDa, equivalent to the
molecular weigh of the predicted thioredoxin-fused OSC
(
94 kDa). To the cell-free homogenates was added
careful experiments to obtain 3-epi-lanosterol from the
incubation mixture of a racemic mixture of (3R,S)-2,3-
oxidosqualenes failed, while lanosterol was successfully
isolated. No other homologous osc was found even after
a careful inspection of the genome of this microorgan-
ism. At present, we are unable to account for this
stereochemical inconsistency between the previous
report and our result. There is no report that the
enzyme activities of OSCs was detected with proteins
expressed by E. coli, and the studies of the function of
racemic (3R, S)-2,3-oxidosqualene, and then incubated
as described in ‘‘Materials and Methods.’’ Figure 4
shows the GC trace of the lipophilic fraction from the
incubation mixture. A new peak was found in the
incubation mixture with the cell-free extract of the
E. coli encoding the putative OSC (Fig. 4B), but no
peak (other than 2,3-oxidosqualene substrate) was
observed from that lacking the gene (Fig. 4A). GC–
MS (EI) showed the following ions characteristic of
lanosterol: m=z 69 (70%), 81 (20), 393 (46), 411 (100),
1
3)
OSCs have been carried out using only lanosterol
20,21)
þ
and 426 (M , 65) (Fig. 5). The enzymatic product was
isolated by SiO2 column chromatography eluting with
hexane:EtOAc (100:5). The characteristic signals in the
1
synthase-deficient (erg7) S. cerevisiae.
This is due
to the serious formation of inclusion body in E. coli and
to the failure of solubilization. Thus, this is the first
report that the enzyme activity of OSC was successfully
detected using the protein expressed in E. coli.
H NMR spectrum (600 MHz, CDCl ) were as follows:
3
ꢂH (ppm), 0.69 (3H, s), 0.81 (3H, s), 0.87 (3H, s), 0.91
(
3H, d, J ¼ 6:5 Hz), 0.98 (3H, s), 1.00 (3H, s), 1.60 (3H,
s), 1.68 (3H, s), 3.24 (1H, dd, J ¼ 11:7, 4.4 Hz, H-3),
.10 (1H, t, J ¼ 7:0 Hz, H-24). These data, including
The biochemical results described here unambigu-
ously verify that SE and OSC from M. capsulatus have
the same features as those from eukaryotes with regard
to substrate and product specificities; epoxidation of
squalene proceeds with complete regio- and stereo-
specificities to afford (3S)-2,3-oxidosqualene, and the
polycyclization reaction by the OSC is active only as to
(3S)-2,3-oxidosqualene, but inert as to the (3R)-form.
These findings suggest that a couple of se and osc genes
from M. capsulatus may have been acquired from
eukaryotes through lateral gene transfer.
From the viewpoint of molecular evolution, it is
intriguing to address the question of how prokaryotes
acquired the eukaryote-type se and osc genes from
M. capsulatus. Lateral gene transfer from higher-order
extant eukaryotes into bacteria is unlikely, and thus the
5
other proton signals, were in full accordance with those
of authentic lanosterol. The apparent coupling constants
of H-3 (dd, J ¼ 11:7, 4.4 Hz) definitively indicate that
H-3 was arranged in an ꢀ-configuration, i.e., ꢁ-orienta-
tion for 3-OH, but no proton signal expected from ꢁ-
arranged H-3 (br s) was detected. Rohmer et al. reported
that (3S)- and (3R)-2,3-oxidosqualenes were converted
into lanosterol (3ꢁ-OH) and 3-epi-lanosterol (3ꢀ-OH),
respectively,1 by incubation with cell-free homoge-
nates from grown cells of M. capsulatus, and they
suggested that broad substrate specificity that can accept
both (3R)- and (3S)-oxidosqualenes is characteristic of
SHCs and OSCs from prokaryotes.13) However, our
3)

Sterol Biosynthesis by Prokaryotes
2549
gene transfer may have occurred between bacteria and
In conclusion, we describe here the first in vitro
evidence for the occurrence of SE and OSC in
M. capsulatus by establishing the enzyme activities.
The regio- and stereospecificities accompanied by the
enzymatic action were completely identical to those of
extant eukaryotes. The expressed OSC gave lanosterol
selectively from (3S)-2,3-oxidosqualene, but no produc-
tion of 3-epi-lanosterol, which is inconsistent with a
9
)
ancient eukaryotes, as suggested by Person et al. In
fact, cholestane and its 28- to 30-carbon analogs have
been found in molecular fossils, giving persuasive
evidence that eukaryotes existed 500 million to 1 billion
years ago.2 G. obscuriglobus also accumulates sterols
2)
9
)
of lanosterol and parkeol in the cells. The open reading
frames (ORFs) of the putative se and osc genes involved
in M. capsulatus are contiguously organized through
translational coupling (Fig. 1C), and those of G. ob-
scuriglobus also are positioned in a franking region
1
3)
previous paper. This new finding sheds light on the
evolutionary aspect of the steroid biosynthetic genes in
prokaryotes.
(
Fig. 1D). In contrast to the prokaryotes, the two genes
involved in extant eukaryotes are located on distinct
chromosomes; e.g., human OSC and SE are on chro-
mosomes 21 and 8, respectively, and S. cerevisiae OSC
and SE are on chromosomes 8 and 7, respectively. As
described above, it is likely that bacteria and early
eukaryotes exchanged the biosynthetic pathways, sug-
gesting that ancient unicellular eukaryotes may have had
contiguous ORFs of SE and OSC. This would have
facilitated gene transfer of the intact sterol pathway
between bacteria and ancient eukaryotes, before the two
genes were separated and rearranged on the different
chromosomes during evolution. The structures of steroid
Acknowledgments
This work was made possible by the financial support
(nos. 16208012 and 18380001) provided by the Ministry
of Education, Culture, Sports, Science and Technology
of Japan.
References
1
)
Hoshino, T., and Sato, T., Squalene-hopene cyclase:
catalytic mechanism and substrate recognition. Chem.
Commun., 291–301 (2002).
4
)
components found in M. capsulatus indicate that
lanosterol is further metabolized with aid of 14ꢀ-
demethylase (cytochrome P-450), followed by oxidation
of the 4ꢀ-methyl group. The 14ꢀ-demethylase of
2) Yoder, R. A., and Johnston, J. N., A case study in
biomimetic total synthesis: polyolefin carbocyclizations
to terpenes and steroids. Chem. Rev., 105, 4730–4756
(
2005).
2
3)
3) Bird, C. W., Lynch, J. M., Pirt, F. J., Reid, W. W.,
Brooks, C. J. W., and Middleditch, B. S., Steroids and
squalene in Methylococcus capsulatus growth on meth-
ane. Nature, 230, 473–474 (1971).
M. capsulatus was recently investigated, but none of
the biosynthetic genes that follow after the 14ꢀ-
demethylase is yet known. There is a phylogenetic
implication that bacterial genes of sterol 14ꢀ-demethy-
4
)
Bouvier, P., Rohmer, M., Benveniste, P., and Ourisson,
8ð14Þ
2
4)
lase (CYP51) may have been acquired from plants.
G., ꢀ -Steroids in bacterium Methylococcus capsu-
latus. Biochem. J., 159, 267–271 (1976).
The information regarding the sterol biosynthetic genes
of prokaryotes is extremely limited. The following
prokaryotic genes have been sequenced: both se and osc
from M. capsulatus and G. obscuriglobus, but only osc
from S. aurantiaca (the se has not yet been sequenced)
5) Volkman, J. K., Sterols in microorganisms. Appl.
Microbiol. Biotechnol., 60, 495–506 (2003).
6) Poralla, K., Cycloartenol and other triterpene cyclase. In
‘‘Comprehensive Natural Products Chemistry’’ Vol. 2,
eds. Barton, D., Nakanishi, K., and Meth-Cohn, O.,
Pergamon/Elesevier, Amsterdam, pp. 299–319 (1999).
Bode, H. B., Zeggei, B., Silakowski, B., Wenzei, S. C.,
Reichenbach, H., and M u¨ ller, R., Steroid biosynthesis in
prokaryotes: identification of myxobacterial steroids and
cloning of the first bacterial 2,3-oxidosqualene cyclase
from the myxobacterium Stigmatella aurantiaca. Mol.
Microbiol., 47, 471–481 (2003).
(
Fig. 1). However, the putative genes of the latter two
prokaryotic species have not yet been characterized by
confirmation of their enzymatic activities in vitro.
Further identification of steroid biosynthetic genes,
including other bacterial species, is necessary for a
better understanding of unusual steroid production. As
described above, bacteria usually use pentacyclic
hopene as a reinforcer of their cell membranes. G. ob-
scuriglobus is also presumed to have shc gene, as found
in M. capsulatus, that is, the shc gene of G. obscur-
iglobus UQM 2246 was surmised from TIGR database
7
)
8) Kohl, W., Glo, A., and Reichenbach, H., Steroids from
the Myxobacterium Nannocystis exedens. J. Gen. Micro-
biol., 129, 1629–1635 (1983).
9
)
Pearson, A., Budin, M., and Brocks, J. J., Phylogenetic
and biochemical evidence for sterol synthesis in the
bacterium Gemmata obscuriglobus. Proc. Natl. Acad.
USA, 100, 15352–15357 (2003).
(
http://www.tigr.org/, gnl TIGR 214688 contig:4535).
The co-existence of hopanoid triterpene and steroids
may further reinforce the cell membrane. Otherwise,
there may be a difference in physiological function
between steroids and hopanoids. Further biochemical
studies will be required to address the question of why
M. capsulatus and G. obscuriglobus possess both SHC
and OSC. To understand steroid production by bacteria,
many unresolved issues remain.
1
0) Shouten, S., Bowman, J. P., Rijpstra, W. I. C., and
Damst e´ , J. S. S., Sterols in a psychrophilic methano-
troph, Methylosphaera hansonii. FEMS Microbiol. Lett.,
1
86, 193–195 (2000).
1
1) Tippelt, A., Jahnke, L., and Poralla, K., Squalene-hopene
cyclase from Methylococcus capsulatus (Bath): a bacte-
rium producing hopanoids and steroids. Biochem. Bio-

2550
C. NAKANO et al.
phys. Acta, 1391, 223–232 (1998).
2) Ward, N., Larsen, O., Sakwa, J., Bruseth, L., Khouri, H.,
Biosci. Biotechnol. Biochem., 63, 2189–2198 (1999).
1
18) Hoshino, T., Nakano, S., Kondo, T., Sato, T., and
Miyoshi, A., Squalene-hopene cyclase: final deprotona-
tion reaction, conformational analysis for the cyclization
of (3R,S)-2,3-oxidosqualene and further evidence for the
requirement of an isopropylidene moiety for initiation of
the polycyclization cascade and for the formation of the
5-membered E-ring. Org. Biomol. Chem., 2, 1456–1470
(2004).
Durkin, A. S., Dimitrov, G., Jiang, L., Scanlan, D., Kang,
K. H., Lewis, M., Nelson, K. E., Methe, B., Wu, M.,
Heidelberg, J. F., Paulsen, I. T., Fouts, D., Ravel, J.,
Tettelin, H., Ren, Q., Read, T., DeBoy, R. T., Seshadri,
R., Salzberg, S. L., Jensen, H. B., Birkeland, N. K.,
Nelson, W. C., Dodson, R. J., Grindhaug, S. H., Holt, I.,
Eidhammer, I., Jonasen, I., Vanaken, S., Utterback, T.,
Feldblyum, T. V., Fraser, C. M., Lillehaug, J. R., and
Eisen, J. A., Genomic insights into methanotrophy: the
complete genome sequence of Methylococcus capsulatus
19) Yamada, Y., Seo, C. W., and Okada, H., Asymmetric
synthesis of (S)-(ꢃ)-squalene-2,3-epoxide. Agric. Biol.
Chem., 45, 1741–1742 (1981).
(
Bath). PLoS Biol., 2, 1616–1628 (2004).
20) Herrera, J. B., Bartel, B., Wilson, W. K., and Matsuda, S.
P., Cloning and characterization of the Arabidopsis
thaliana lupeol synthase gene. Phytochemistry, 49,
1905–1911 (1998).
21) Zhang, H., Shibuya, M., Yokota, S., and Ebizuka, Y.,
Oxidosqualene cyclases from cell suspension cultures of
Betula platyphylla var. japonica: molecular evolution of
oxidosqualene cyclases in higher plants. Biol. Pharm.
Bull., 26, 642–650 (2003).
22) Brocks, J. J., Logan, G. A., Buick, R., and Summons, R.
E., Archean molecular fossils and the early rise of
eukaryotes. Science, 285, 1033–1036 (1999).
23) Jackson, C. J., Lamb, D. C., Marezylo, T. H., Warrilow,
A. G. S., Manning, N. J., Lowe, D. J., Kelly, D. E., and
Kelly, S. L., A novel sterol 14ꢀ-demethylase/ferredoxin
fusion protein (MCCYP51FX) from Methylococcus
capsulatus represents a new class of the cytochrome
P450 superfamily. J. Biol. Chem., 277, 46959–46965
(2002).
1
3) Rohmer, M., Bouvier, P., and Ourisson, G., Non-specific
lanosterol and hopanoid biosynthesis by a cell-free
system from the bacterium Methylococcus capsulatus.
Eur. J. Biochem., 112, 557–560 (1980).
4) Sato, T., and Hoshino, T., Overexpression of squalene-
hopene cyclase by the pET Vector in Escherichia coli
and first identification of tryptophan and aspartic acid
residues inside the QW motif as active sites. Biosci.
Biotechnol. Biochem., 62, 407–411 (1998).
5) Abe, I., Abe, T., Lou, W., Masuoka, T., and Noguchi, H.,
Site-directed mutagenesis of conserved aromatic resi-
dues in rat squalene epoxidase. Biochem. Biophys. Res.
Commun., 352, 259–263 (2007).
6) Sato, T., and Hoshino, T., Kinetic studies on the function
of all the conserved tryptophans involved inside and
outside the QW motifs of squalene-hopene cyclase:
stabilizing effect of the protein structure against thermal
denaturation. Biosci. Biotechnol. Biochem., 63, 1171–
1
1
1
1
180 (1999).
24) Rezen, T., Debeljak, N., Kordis, D., and Rozman, D.,
New aspects on lanosterol 14ꢀ-demethylase and cyto-
chrome P450 evolution: lanosterol/cycloartenol diversi-
fication and lateral transfer. J. Mol. Evol., 59, 51–58
(2004).
1
7) Sato, T., and Hoshino, T., Functional analysis of the
DXDDTA motif in squalene-hopene cyclase by site-
directed mutagenesis experiments: initiation site of the
polycyclization reaction and stabilization site of the
carbocation intermediate of the initially cyclized A-ring.