Functional analyses of Tyr420 and Leu607 of Alicyclobacillus acidocaldarius squalene-hopene cyclase. Neoachillapentaene, a novel triterpene with the 1,5,6-trimethylcyclohexene moiety produced through folding of the constrained boat structure

Article abstract of DOI:10.1271/bbb.66.1660

The functions of Tyr420 and Leu607 were analyzed by constructing various site-directed mutants. The mutation at position 420 into Ala and Gly gave bicyclic α-and γ-polypodatetraene in significant amounts, but with a trace amount of tricyclic malabaricatriene. The kinetic data for and the product distribution of the Y420F mutant indicate that the major function of Tyr420 is to stabilize the 6/6-fused bicyclic cation. Mutation experiments on Leu607 demonstrate that the appropriate steric bulk size at position 607 is required to strongly bind with the product-like conformation formed during the polycyclization process. Introduction of the bulkiest Trp residue into 420 or 607 led to the production of a novel monocyclic triterpene having the (5R,6R)-1,5,6-trimethylcyclohexene ring, named neoachillapentaene, indicating that the enzymatic cyclization proceeded via a constrained boat structure. Folding of the squalene molecule into a boat conformation by squalene cyclase has not been reported before.

Full text of DOI:10.1271/bbb.66.1660

This article was downloaded by: [University of Regina]
On: 17 November 2014, At: 23:47
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer
House, 37-41 Mortimer Street, London W1T 3JH, UK
Bioscience, Biotechnology, and Biochemistry
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/tbbb20
Functional Analyses of Tyr420 and Leu607 of
Alicyclobacillus acidocaldarius Squalene-Hopene
Cyclase. Neoachillapentaene, a Novel Triterpene
with the 1,5,6-Trimethylcyclohexene Moiety…
Tsutomu SATOa, Shigehiro SASAHARAa, Toshiyuki YAMAKAMIa & Tsutomu HOSHINOa
a Department of Applied Biological Chemistry, Faculty of Agriculture, Niigata University
Ikarashi, Niigata 950-2181, Japan
Published online: 22 May 2014.
To cite this article: Tsutomu SATO, Shigehiro SASAHARA, Toshiyuki YAMAKAMI & Tsutomu HOSHINO (2002) Functional
Analyses of Tyr420 and Leu607 of Alicyclobacillus acidocaldarius Squalene-Hopene Cyclase. Neoachillapentaene, a Novel
Triterpene with the 1,5,6-Trimethylcyclohexene Moiety…, Bioscience, Biotechnology, and Biochemistry, 66:8, 1660-1670,
DOI: 10.1271/bbb.66.1660
To link to this article: http://dx.doi.org/10.1271/bbb.66.1660
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of
the Content. Any opinions and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied
upon and should be independently verified with primary sources of information. Taylor and Francis shall
not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other
liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or
arising out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions

Biosci. Biotechnol. Biochem., 66 (8), 1660–1670, 2002
Functional Analyses of Tyr420 and Leu607 of Alicyclobacillus acidocaldarius
Squalene-Hopene Cyclase. Neoachillapentaene, a Novel Triterpene
with the 1,5,6-Trimethylcyclohexene Moiety Produced through
Folding of the Constrained Boat Structure
†
Tsutomu SATO, Shigehiro SASAHARA, Toshiyuki YAMAKAMI, and Tsutomu HOSHINO
Department of Applied Biological Chemistry, Faculty of Agriculture, Niigata University, Ikarashi,
Niigata 950-2181, Japan
Received January 10, 2002; Accepted March 27, 2002
mutants and substrate analogs,1,2) we have estab-
The functions of Tyr420 and Leu607 were analyzed
by constructing various site-directed mutants. The mu-
tation at position 420 into Ala and Gly gave bicyclic
and -polypodatetraene in signiˆcant amounts, but with
lished that this polycyclization reaction consists of
eight reaction steps (Scheme 1): (1) 1st cyclization to
form A-ring 4 by proton attack on the terminal dou-
ble bond,5) (2) 2nd ring closure to give the B-ring (6
a-
g
a trace amount of tricyclic malabaricatriene. The kinetic
data for and the product distribution of the Y420F
mutant indicate that the major function of Tyr420 is to
W
6-fused A B ring system 5),6) (3) 3rd cyclization to
W
yield the 5-membered C-ring (6 6 5-fused A B C-
WW
W W
stabilize the 6 6-fused bicyclic cation. Mutation experi-
tricyclic ring system 6) by Markovnikov closure, (4)
W
ments on Leu607 demonstrate that the appropriate ster-
ic bulk size at position 607 is required to strongly bind
with the product-like conformation formed during the
polycyclization process. Introduction of the bulkiest
Trp residue into 420 or 607 led to the production of a
which then undergoes ring expansion to form the 6-
novel monocyclic triterpene having the (5
trimethylcyclohexene ring, named neoachillapentaene,
indicating that the enzymatic cyclization proceeded via
R,6R)-1,5,6-
a
constrained boat structure. Folding of the squalene
molecule into a boat conformation by squalene cyclase
has not been reported before.
Key words: squalene; hopene; terpene cyclase; un-
natural natural product; Alicyclobacillus
acidocaldarius
The cyclization of squalene 1 into pentacyclic
hopene 2 and hopanol 3 is one of the most complicat-
ed biochemical reactions (Scheme 1)1–4) which is cata-
lyzed by squalene-hopene cyclases (SHCs) [EC
5.4.99-] from prokaryotic species. The polycycliza-
tion reaction proceeds under precise enzymatic con-
trol to form ˆve new rings and nine new chiral cen-
ters. This polycyclization mechanism is analogous to
that mediated by eukaryotic oxidosqualene cyclases
(OSCs) which give lanosterol and numerous plant
Scheme 1. Polycyclization Reaction of Squalene 1, Leading to
Hopene 2 and Hopanol 3.
This polycyclization reaction comprises eight reaction steps
which has been demonstrated by isolating the abortive cycliza-
tion products formed by numerous site-directed mutants and by
trapping the carbocation intermediates by using the substrate
analogs.1,2)
triterpenes from (3S
)-2,3-oxidosqualene.2–4) In the
past several years, there have been remarkable ad-
vances with the catalytic mechanism of Alicyclobacil-
lus acidocaldarius SHC.1,2) By using site-directed
†
To whom correspondence should be addressed. Fax: +81-25-262-6854; E-mail: hoshitsu
＠
agr. niigata-u.ac.jp
Abbreviations: SHC, squalene-hopene cyclase; OSC, oxidosqualene cyclase; LDAO,
N
,
N
-dimethyldodecylamine-N-oxide; GC, gas chro-
matography; TLC, thin-layer chromatography; CD, circular dichroism

Tyr420 and Leu607 of Squalene Cyclase
1661
membered C-ring (6 6 6-fused tricyclic ring system
data were presented. In order to infer the precise
catalytic function, numerous mutants have to be con-
structed, in which Tyr420 and Leu607 are replaced by
other diŠerently categorized amino acids with diŠer-
WW
7),7) (5) 5th cyclization to give the thermodynamically
favored 5-membered D-ring (6 6 6 5-fused A B C
WWW
W W W
D ring system 8, 17-epi-dammarenyl cation),8) (6) fol-
lowed by the second ring enlargement process to
ent electronic and or bulk size environments, and the
W
form the 6-membered D-ring (6 6 6 6-fused A B
kinetic data of the mutants have also to be evaluated
in detail. We planned to prepare many mutants tar-
geted at positions 420 and 607 in order to provide
deeper insight into the functions of Tyr420 and
Leu607.
WWW
W W
C D-ring system, prohopanyl cation 9),9,10) (7) the
W
last ring closure process to construct the 6 6 6 6 5-
WWWW
fused A B C D E-ring system (10, hopanyl ca-
W W W W
tion),10) and (8) the ˆnal deprotonation reaction to in-
troduce the double bond between C-22 and C-29. The
polycyclization reaction is characterized by the fold-
ing of 1 into an all pre-chair conformation during the
multiple reaction processes. We have demonstrated
We present here deˆnitive evidence that the major
function of Tyr420 is to stabilize bicyclic cation inter-
mediate 5 formed during the polycyclization cascade.
Furthermore, we describe that the steric bulk size at
position 607 is critical to the optimal folding of 1
leading to 5 having a chair form for the B-ring; the
most appropriate bulk size of Leu gives rise to perfect
contact around the B-ring formation site. These
inferences are drawn from the kinetic data of
numerous site-directed mutants targeted at Tyr420
and Leu607 and from the structures of the abortive
cyclization products. Replacement by the bulkier
that cation-p interaction between p-electrons of aro-
matic amino acids and the transient carbocations
greatly contributed to the acceleration of the polycy-
clization reaction. The replacement of Phe365,6)
Phe6017) and Phe60510) by Tyr greatly enhanced the
reaction velocity due to the elevated
p-electron den-
sity of Tyr. In addition, a few tyrosine residues were
shown to intensify the cation-p interaction by being
placed at correct positions in the reaction cavity.11)
The steric bulk size of active site residues is likely to
be responsible for stereochemical control during the
polycyclization reaction; the substitution of Ile with
smaller Ala or Gly at position 261 aŠorded false in-
amino acids at the two positions, such as Y420
and L607 F or W, gave an unnatural natural triter-
pene having the (5 ,6 )-1,5,6-trimethylcyclohexene
ªW
ª
R
R
ring. This carbocyclic triterpene skeleton has never
been reported before, despite more than 90 diŠerent
triterpene skeletons being known at the present time.
termediates having 13a-H in the 6 6 5-fused tricyclic
WW
and 17
a
-H in the 6 6 6 5-fused tetracyclic cation,9)
WWW
the conˆgurations of which are opposite to those of
true intermediates 6 (13b-H) and 8 (17b-H) (Scheme
Materials and Methods
1
1). However, the question has remained unresolved
as to how stereochemical speciˆcity is attained along
the polycyclization pathway; in other words, why
only single stereoisomer 2 is produced, despite
stereoisomers of 29 being possible. Matsuda and co-
workers have recently reported several examples, in
which the steric bulk size of active sites involved in
OSCs in‰uenced the polycyclization pathway of ox-
idosqualene.12–14) To answer the question about the
stereochemical speciˆcity, a number of mutants hav-
ing diŠerent bulk size should be constructed and eval-
uated.
Analytical method. H-NMR and 13C-NMR spec-
tra were measured in a C6D6 solution by Bruker DPX
400 and DMX 600 instruments. The chemical shifts
1
of H- and 13C-NMR spectra (ppm) are relative to
7.28 and 128.0 ppm for the solvent peak. MS spectra
were recorded with a Jeol SX 102 mass spectrometer.
GC analyses were carried out by a Shimadzu GC-8A
×
instrument with a DB-1 capillary column (0.53 mm
30 m; injection temperature, 290 C; column
9
9
temperature, 270 C; and ‰ow rate of N2 carrier,
1.0 kg cm2). Speciˆc rotation values were measured
W
at 25
9
C with a Horiba SEPA-300 instrument.
The two residues of Tyr420 and Leu607 are not
highly conserved in all known SHCs. The Tyr420 and
Leu607 residues of A. acidocaldarius SHC are
replaced by Phe and Ile, respectively, in some other
bacterial SHCs.15) However, X-ray information on
the SHC-inhibitor complex shows that these residues
General methods. Details of the experimental pro-
tocol have been reported in the previous papers5,11)
for the overexpression systems with the pET vector in
E. coli BL21 (DE3), DNA sequence analyses, prepa-
ration of cell-free homogenates and enzyme puriˆca-
tion methods.
are located close to the inhibitor of
N, N-dimethyl-
dodecylamine-
N
-oxide (LDAO),16,17) implying that
the two residues may be responsible for catalysis.
Indeed, Poralla and co-workers have reported that
the Tyr420Ala mutant produced three abortive
Site-directed mutagenesis. All site-directed muta-
genesis experiments were performed as described in
the previous papers.5,11) The following primers were
used:
cyclization products consisting of two bicycles,
and -polypodatetraene, and one tricycle, malabari-
catriene,18,19) and that the Leu607Lys mutant gave
bicyclic
-polypodatetraene.20) However, no kinetic
a-
g
Y420G, 5
CGCCCCAAC]-3
Y420A, 5 -pd[GCTCGTGTTGTCCACGTC
?
-pd[CTCGTGTTGTCCACGTC
G
CCGG-
?
(„SalI)
g
?
G
GCG-

1662
T. SATO et al.
×
n-hexane (200 ml 3) from each
GCGCCCCAACC]-3
Y420D, 5 -pd[CTCGTGTTGTCCACGTCGT
CGCCCCAAC]-3 („SalI)
Y420H, 5 -pd[GTTGTCGACGTCGT
CAACCG]-3 („BbeI)
Y420F, 5 -pd[CTCGTGTTGTCCACGTCCC
CGCCCCAAC]-3 („SalI)
Y420W, 5 -pd[CTCGTGTTGTCCACGTC
CGCCCCAAC]-3 („SalI)
L607A, -pd[CATGGTGTAGCC
ATCCCCGGGGAAGCCCG]-3 (+SmaI)
L607I, 5 -pd[GGTGTAGCCGA GTAGAAATCC-
CCGGGGAAGCCCG]-3 (+SmaI)
L607F, -pd[CATGGTGTAGCCGA
ATCCCCGGGGAAGCCCG]-3 (+SmaI)
L607W, 5 -pd[CATGGTGTAGCCCCAGTAGAA-
ATCCCCGGGGAAGCCCG]-3 (+SmaI)
The bold letters designate the altered bases, and
the italic letters show the target mutations. The un-
derlined letters show the silent mutations for easy
screening of the desired mutants by a restriction frag-
ment analysis. The created or deleted restriction sites
are shown in parentheses. To ascertain that the
desired mutation had been carried out, the entire
region of all the inserted DNAs was sequenced.
?
(„Sal I)
was extracted with
incubated mixture, and then subjected to SiO2
column chromatography by eluting with -hexane,
giving 1, 2, 11 and 13–16 in a pure state. A mixed sol-
vent of -hexane and EtOAc (100:2) was used to elute
3 and 12 due to the high polarity. The complete
puriˆcation of 12 was achieved by HPLC [ -hex-
?
C
GG-
AGCGCCC-
GG-
CCGG-
?
n
?
G
?
n
?
A
?
n
＝
?
G
ane:isopropanol 100:0.1]. The following yields
were obtained: 1 (3.6 mg), 2 (94.3 mg), 3 (11.3 mg),
14 (9.1 mg), 15 (21.2 mg) and 16 (1.4 mg) for Y420A;
1 (3.5 mg), 2 (0.3 mg), 11 (3.2 mg), 13 (26.7 mg), 14
(6.0 mg) and 15 (3.3 mg) for Y420W; 1 (12.5 mg), 2
(0.6 mg), 11 (4.3 mg), 12 (0.5 mg), 13 (39.1 mg), 14
(6.8 mg) and 15 (4.9 mg) for L607F; and 1 (0.5 mg),
11 (7.3 mg), 12 (0.4 mg), 13 (53.7 mg) and 14
(8.0 mg) for L607W.
?
5?
GGCGTAGAA-
?
?
T
?
5?
AGTAGAA-
?
?
?
The structures of 11–16 were determined by
spectral analyses with EIMS and NMR, including
COSY 45, HOHAHA, NOESY, DEPT, HMQC and
HMBC. Products 13, 14 and 16 were 3-deox-
yachilleol,
g
-polypodatetraene and 17(
E )-13a(H )-
malabarica-14(27),17(
E
),21-triene, respectively, the
complete NMR assignments of which have been
reported before.6,7,11) The spectroscopic data (NMR
assignment, EI-MS and speciˆc rotation) for 11, 12
and 15, which had not previously been isolated in our
experiments, are described below.
Thermal stability and kinetic analysis. All the mu-
tated SHCs were homogeneously puriˆed according
Compound 11 (oil, neoachilla-3,9(
21-pentaene cf IUPAC name: 1,5(
methyl-6-[3,8,12,16-tetramethyl-heptadeca-3(
E
),13(
E
),17(
E ),
to the protocol described in the previous paper.5,11)
A
M
(
R
),6(
R
)-tri-
mixed solution, which was composed of 0.5 m
squalene, 0.2 Triton X-100 and 5 g of the
homogeneously puriˆed enzyme in a sodium citrate
buŠer (60 m , pH 6.0), was prepared in a ˆnal
E
),
1
z
m
7(
E
),11(
E
),15(
E
)-tetraenyl]-cyclohexene). The H-
and 13C-NMR data measured at 400 MHz are shown
M
in Table 1. EIMS fragments m z (
z
): 410 (16) [M+],
W
volume of 5 ml for the enzyme reactions. Incubation
was done for 60 min at diŠerent temperatures (30, 35,
395 (1) [M+„Me], 341 (9), 273 (1), 205 (12), 137
(26), 123 (100) [trimethylcyclohexene ring], 69 (63).
40, 45, 50, 55, 60, 65, or 70
9
C) to examine the ther-
HREIMS for C30 H50: calcd., 410.3913; found,
25
＝
mal stability of the cyclases. To estimate the kinetic
parameters, incubation was conducted for 60 min at
410.3888. [
a
]
„4.6 (
c
0.27, CHCl3).
-hydoroxyachilla-9(
),21-tetraene). 600 MHz, dH (ppm): 5.53
6.4 Hz, H-10), 5.46 (1H, bt, 6.1 Hz,
6.1 Hz, H-17), 5.37 (1H, t,
7.0 Hz, H-21), 2.48 (1H, m, H-8), 2.32 (4H, m,
D
Compound 12 (oil,
13( ),17(
(1H, bt,
H-13), 5.42 (1H, t,
6
a
E ),
45
9
C before adding 15
z
methanolic KOH (6 ml) to
E
E
＝
terminate the enzyme reaction. The lipophilic
J
J
＝
＝
J
products (2 and 3) and starting material (1) which
＝
J
had remained unchanged were extracted with
n
-
hexane (5 ml
×
4) from each reaction mixture, and ˆ-
H-16 and H-20), 2.31 (5H, m, H-8, H-11 and H-12),
2.22 (4H, m, H-15 and H-19), 1.84 (3H, s, H-26),
1.83 (1H, m, H-7), 1.80 (3H, s, H-29), 1.75 (3H, s,
H-28 or H-27), 1.74 (3H, s, H-27 or H-28), 1.72 (1H,
m, H-1), 1.69 (3H, s, H-30), 1.56 (1H, m, H-7), 1.50
(1H, m, H-2), 1.41 (1H, m, H-2), 1.35 (1H, m, H-3),
nal product 2 was quantiˆed by a GC analysis with a
DB-1 capillary column (30 m in length). The kinetic
value of
Burk plots.
Km and kcat were estimated from Lineweaver-
＝
Isolation and structural determination of products
from mutated SHCs. A cell-free homogenates from
the cultured E. coli cells (6 liters), in which the mutat-
ed SHCs of Y420A, Y420W, L607F and L607W were
expressed, was prepared as described in the previous
papers.5,11) Squalene 1 (150 mg, 50 mg, 75 mg and
75 mg, respectively, for Y420A, Y420W, L607F and
L607W) was separately incubated with the
homogenate at the optimal temperature for 16 h, and
the reaction mixture was lyophilized. The product
1.29 (1H, ddd,
J
3.7, 12.6, 12.6 Hz, H-1), 1.20 (1H,
＝
J
ddd,
1.15 (1H, t,
4.2, 12.7, 12.7Hz, H-3), 1.20 (3H, s, H-25),
＝
J
4.5 Hz, H-5), 1.05 (3H, s, H-23), 0.87
(3H, s, H-24). dC (ppm): 136.5 (C-9), 135.21 (C-14),
134.94 (C-18), 131.07 (C-22), 124.95 (C-21), 124.88
(C-17 or C-13), 124.84 (C-13 or C-17), 124.7 (C-10),
73.5 (C-6), 56.85 (C-5), 43.95 (C-1), 43.39 (C-8),
41.76 (C-3), 40.22 (C-15 or C-19), 40.19 (C-19 or
C-15), 35.59 (C-4), 32.93 (C-23), 28.78 (C-11 or
C-12), 28.75 (C-11 or C-12), 27.22 (C-16 or C-20),

Tyr420 and Leu607 of Squalene Cyclase
1663
27.13 (C-20 or C-16), 25.81 (C-29), 25.34 (C-7), 23.5
(C-25), 21.59 (C-24), 20.84 (C-2), 17.71 (C-30), 16.33
(C-26), 16.19 (C-28 or C-27), 16.09 (C-27 or C-28).
The relative stereochemistry was conˆrmed by the
NOE cross peak of Me-23 with H-5 and that of
with those of authentic samples, which had previous-
ly been isolated by us, and quantiˆed by a GC analy-
sis.
Results
Me-24 with Me-25. EIMS fragments m z
(z):
W
428 (0.7) [M+], 410 (21) [M+„H2O], 395 (3)
[M+„H2O„Me], 341 (8), 273 (3), 205 (30), 137 (73)
[allylic cleavage between C-7 and C-8], 69 (100).
HREIMS for C30H50 M+„H2O): found, 410.3905.
To provide insight into the functions of Tyr420
and Leu607, the following ten mutants were con-
structed: Y420G, Y420A, Y420D, Y420H, Y420F,
Y420W, L607A, L607I, L607F and L607W. Tyr is an
aromatic amino acid having a phenolic hydroxyl
group. To examine the role of the hydroxyl group
25
D
23
D
[
a
]
+0.5 (
c
0.0034, CHCl3), cf
[
a
]
+2.5 (c 0.6,
CHCl3).21)
Compound 15 (oil,
dH (ppm): 5.46 (1H, bt,
bt, 7 Hz, H-17), 5.37 (1H, bt,
5.09 (1H, bs, H-26), 4.84 (1H, bs, H-26), 2.51(1H,
a
-polypodatetraene). 600 MHz,
and the aromatic
residue, the Tyr was mutated into Phe and Trp. To
diminish the role of the aromatic -electrons, the
p-electron system of the Tyr
＝
J
7.1 Hz, H-13), 5.44 (1H,
6.9 Hz, H-21),
J
＝
J
＝
p
Y420G and Y420A mutants were prepared. The
mutation experiments for the Tyr into Asp and His
were carried out to evaluate how the acidic function
＝
12.8, 4.7, 2.4 Hz, H-7), 2.43 (1H, m, H-12),
ddd,
J
¿
2.3 (4H, m, H-15, H-16, H-19 and H-20), 2.16
＝
(1H, m, H-12), 2.10 (1H, ddd,
J
12.8, 12.8, 4.7 Hz,
of Asp (or anion of the carboxylate) and the p-elec-
H-7), 1.82 (1H, m, H-1), 1.81 (3H, s, H-29), 1.78
(1H, m, H-9), 1.77 (3H, s, H-27), 1.75 (1H, m, H-6),
1.74 (3H, s, H-28), 1.69 (3H, s, H-30), 1.69 (1H, m,
H-11), 1.64 (1H, m, H-2), 1.62 (1H, m, H-11), 1.55
tron density of His (or the basic function of proton-
abstraction) would aŠect the polycyclization cascade.
To get better knowledge about the steric bulk size at
position 607, Ala with a smaller bulk size and Ile with
an equivalent size were introduced. The Phe and Trp
mutants were also constructed to evaluate the eŠect
of the aromatic ring on the polycyclization reaction.
＝
12.9 Hz, H-3), 1.42
(1H, m, H-2), 1.48 (1H, bd,
J
(1H, dddd,
J
＝
12.9, 12.9, 12.9, 4.2 Hz, H-6), 1.27
＝
J
(1H, ddd,
12.9, 12.9, 3.8 Hz, H-3), 1.12 (1H, dd,
＝
＝
13.0, 13.0,
J
12.6, 2.5 Hz, H-5), 1.08 (1H, ddd,
J
3.6 Hz, H-1), 0.967 (3H, s, H-23), 0.921 (3H, s,
H-24), 0.865 (3H, s, H-25). dC (ppm): 148.85 (C-8),
135.0 (C-14), 134.97 (C-18), 131.08 (C-22), 125.73
(C-13), 124.94 (C-21), 124.84 (C-17), 106.65 (C-26),
56.46 (C-9), 55.66 (C-5), 42.43 (C-3), 40.24 (C-15 or
C-19), 40.23 (C-15 or C-19), 39.82 (C-10), 39.31
(C-1), 38.73 (C-7), 33.71 (C-4), 33.65 (C-23), 27.38
(C-12, C-16 or C-20), 27.25 (C-12, C-16 or C-20),
27.15 (C-12, C-16 or C-20), 25.82 (C-29), 24.77
(C-6), 24.23 (C-11), 21.89 (C-24), 19.75 (C-2), 17.72
(C-30), 16.17 (C-28), 16.12 (C-27), 15.61 (C-25). The
relative stereochemistry was conˆrmed by the NOE
cross peaks among H-5, H-9 and Me-23. EIMS frag-
Enzymic products formed by the site-directed
mutants
We have previously shown that point mutants,
altered at active sites, aŠorded prematurely cyclized
products. Based upon the product distribution, the
placement of various active residues inside the central
active cavity has been proposed.1,2) With a large
amount of the cell-free homogenate from the grown
mutants, 1 was incubated for 16 h at the optimal tem-
perature and the enzymic products formed by each
mutant were analyzed by GC. As one example, the
product distribution pattern by the L607F mutant is
shown in Fig. 1. Five abortive cyclization compounds
(11–15) were produced by the L607F mutant (Figs. 1
and 2). All the enzymic products were puriˆed by
SiO2 column chromatography, and their structures
determined by detailed NMR analyses including 2D
NMR (COSY 45, HOHAHA, NOESY, HMQC and
HMBC), these beeing conˆrmed by the EI-MS frag-
ment patterns. Products 13 and 14 were 3-deox-
ments m z (
z
): 410 (26) [M+], 395 (18) [M+„CH3],
W
341 (3), 273 (8), 205 (7), 204 (4), 191 (12) [allylic
cleavage between C9 and C11], 137 (51), and 69
25
(100). HREIMS for C30H50: found, 410.3921. [a]
D
+14.2 (c 0.046, CHCl3), cf+27.4 (c
0.4, CHCl3).22)
Product distribution pattern. One mg of 1 was
separately incubated at 45 C for 16 h with cell-free
9
yachilleol and
which had previously been isolated by us.5,6) Product
15 was conˆrmed to be -polypodatetraene by the de-
g-polypodatetraene, respectively,
extracts from the mutants of Y420G, Y420D,
Y420H, Y420F, and L607A and L607I (Table 3),
which had been grown in a 50-ml LB medium. The
hexane extract was obtained from the reaction mix-
ture. The detergent was removed by passing through
a
tailed NMR analyses (see the Materials and Methods
section). The production of 11 and 12 has not been
reported before. The structural determination of 11
and 12 is described later. Prolonged incubation with
the Y420G and Y420A mutants aŠorded three
products: bicyclic 14 and 15 and tricyclic 16. Com-
＝
a short SiO2 column [n-hexane:EtOAc 100:20],
before the extract was subjected to a GC analysis to
examine the product distribution pattern and the
product amount. The enzymatic products were iden-
tiˆed by comparing the GC-MS fragment patterns
pound 16 was determined to be (17
E )-13a(H )-
malabarica-14(27),17,21-triene by the NMR and EI-

1664
T. SATO et al.
Fig. 2. Structures of the Abortive Cyclization Products Isolated
by Some Mutated SHCs Constructed in This Paper.
We propose neoachillapentaene as the name for unnatural
natural triterpene 11. Products 12, 13, 14, 15 and 16 were known
6-
-polypodatetraene,
malabarica-14(27),17,21-triene, respectively.
a
-hydroxy-9(
E
),13(
E
),17(
E
),21-tetraene, 3-deoxyachilleol,
g
a
-polypodatetraene, and 17(
E
)-13a(H )-
Fig. 1. Product Distribution Pattern of the Leu607Phe Mutant
by a GC Analysis.
Five abortive cyclization products (11–15 ) having mono- or
bicyclic skeletons were found in addition to 2. One mg of 1 was
incubated with a cell-free extract in a 200-ml culture under cata-
lytically optimal conditions (559C and pH 6.0), whereby all em-
ployed 1 was consumed. Triton X-100 was removed by passing
the hexane extract of the reaction mixture through a short SiO2
column with a mixture of hexane: EtOAc (100:20) as the eluent
(see the Materials and Methods section). The amount of the mu-
tated SHC used, equivalent to 2 mg of pure Leu607Phe SHC,
was signiˆcantly larger than that in Fig. 4 and Table 3 by 400-
fold and 4-fold, respectively.
MS analyses, and had previously been isolated from
mutants F601A and I261A.7,9) All the abortive cycli-
zation products discussed in this paper are listed in
Fig. 2. The abortive cyclization products were not
found in any detectable amount when the kinetic
data were estimated by incubating 1 with small
Fig. 3. Selected HMBC, NOEs and EI-MS Fragment Ion (m z
W
123, base peak) Data for the Structural Determination of 11.
No NOE was found between Me-24 and Me-25, or between
H-6 and H-7. The numbering on the left side is according to the
neoachillapentaene skeleton, while that of the 1,5,6-trimethylcy-
clohexene ring (the right side) is to IUPAC nomenclature.
amounts of the SHCs (5 mg) and for a short incuba-
tion time (60 min) (see Table 2 and Fig. 4), but were
detected when using large amounts of the SHCs and
by incubating for a prolonged period of incubation
time (16–20 h) (see Table 3 and Fig. 1).
1.01 (Me-24) and dH 1.79 (bs, Me-23) with dC 139.7
(s, C-4) indicate that one allyl methyl group (Me-23)
and one tertiary methyl group (Me-24) were attached
to the monocyclic ring having a double bond (Fig. 3).
In addition, the protons of both Me-25 and Me-24
were correlated with C-5 (s, dC 40.7) in the HMBC
spectrum, conˆrming the position of Me-25. The
double bond position was further conˆrmed by the
Structural determination of 11
Substrate 1 had eight allyl methyl groups and six
1
double bonds in the molecule. In the H-NMR spec-
trum of 11, six allyl methyl protons were found at dH
1.80, 1.79, 1.77, 1.740, 1.736 and 1.69, suggesting
that four double bonds remained unchanged without
participating in the cyclization reaction, inferring
that 11 was a monocyclic compound. The other two
methyl protons appeared in a higher ˆeld of dH 0.98
HMBC cross peaks of both dH 1.52 (H-1) and dH 2.03
(H-2) with dC 124.5 (d, C-3). These data deˆnitively
demonstrate the presence of the 1,5,6-trimethylcyclo-
hexene moiety in 11 (IUPAC numbering). The
trimethylcyclohexene ring was further supported by
observing m z 123 as a base peak in the EI-MS spec-
W
＝
(d,
J
6.8 Hz, Me-25) and
d
H 1.01 (s, Me-24), sugges-
trum (Fig. 3). The relative stereochemistry of the
1,5,6-trimethylcyclohexene ring was determined by
the NOESY spectrum. Strong NOEs of Me-24 with
H-6 and of Me-25 with H-7 revealed the same geo-
ting that the secondary and tertiary methyl groups
was on the cyclized monocyclic ring. The clear
HMBC cross peaks of the two methyl protons of dH

Tyr420 and Leu607 of Squalene Cyclase
1665
Table 1. 1H- and 13C-NMR Data for Compound 11 (neoachillapentaene) in C6D6
The position numbering is according to Fig. 2.
Position
1H
13C
Position
1H
13C
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
1.52 (1H, m); 1.59 (1H, m)
2.03 (1H, m); 2.09 (1H, m)
5.56 (1H, bm)
—
27.4 (t)
25.9 (t)
124.5 (d)
139.7 (s)
40.7 (s)
33.5 (d)
35.6 (t)
34.7 (t)
136.2 (s)
124.2 (d)
28.7 (t)
28.7 (t)
124.8 (d)
135.2 (s)
40.2 (t)
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
2.32 (2H, m)
5.46 (1H, m)a
—
2.23 (2H, m)
2.32 (2H, m)
5.37 (1H, bt,
—
27.1 (t)b
124.8 (d)
134.9 (s)
40.2 (t)
—
27.2 (t)b
124.9 (d)
131.1 (s)
19.4 (q)
21.2 (q)
16.0 (q)
16.4 (q)
16.2 (q)c
16.1 (q)c
25.8 (q)
17.7 (q)
1.88 (1H, m)
J 6.8 Hz)
＝
1.70 (1H, m); 1.67 (1H, m)
1.93 (1H, m); 2.14 (1H, m)
—
5.46 (1H, m)
2.28 (2H, m)
1.79 (3H, bs)
1.01 (3H, s)
0.98 (3H, d,
1.77 (3H, s)
1.74 (3H, s)c
1.736 (3H, s)c
1.80 (3H, s)
1.69 (3H, s)
J 6.8 Hz)
＝
2.28 (2H, m)
5.42 (1H, bt,
—
J
＝
7.0 Hz)a
2.23 (2H, m)
a,b,c
The assignments may be interchangeable.
metry between Me-24 and H-6, and between Me-25
and H-7, but no NOE was observed between Me-24
and Me-25, or between H-6 and H-7. The selected
NOE data are shown in Fig. 3. Thus, the relative
stereochemistry of the 5- and 6-positions in the 1,5,6-
trimethylcyclohexene ring was determined to be all
ature by using the homogeneously puriˆed enzymes.
The Phe mutant had an optimal temperature of
9
65 C, higher than that of the native SHC (Fig. 4A),
suggesting that reinforcement of the protein structure
had acted against thermal denaturation. However,
the speciˆc enzyme activity was relatively less than
that of the wild-type, strongly indicating that the Tyr
residue of the wild-type had acted to stabilize the
transient carbocation intermediate. Table 2 shows
the kinetic data, which were estimated from the
amount of 2 produced. Despite the Phe mutant hav-
ing enhanced a‹nity to 1 (the smaller Km by 1.75-
R
-conˆguration. The results of the detailed NMR
analyses are consistent with the proposed structure
for 11 shown in Fig. 2. The complete NMR assign-
ments are shown in Table 1. Compound 11 is a novel
triterpene, which we name neoachillapentaene for
unnatural natural product 11.
fold), the reaction velocity (kcat) was signiˆcantly
Structural determination of 12
decreased by 3.6-fold, in a comparison with the wild-
type. This ˆnding unequivocally indicates that the
free electron pair of the phenolic hydroxyl group
played a crucial function for stabilizing the transient
cationic intermediate produced during the polycycli-
zation reaction. However, the Trp mutant showed
little activity, despite an enhanced cyclization rate
Compound 12 had high polarity, i.e., no move-
ment on SiO2 TLC, but other products 11 and 13–16
showed Rf values in the range 0.45–0.75 for hexane.
The observation of the fragment ion M+„H2O (m z
W
410) also suggested the involvement of a hydroxyl
group in 12, which was conˆrmed by ˆnding the sig-
nal at
d
C 73.5 (s, C-6), a diagnostic signal for an alco-
being expected owing to the increased p-electron den-
holic carbon, in the 13C-NMR spectrum. Five allyl
methyl groups and four double bonds were left un-
changed without any responsibility for the cycliza-
tion reaction, indicating the monocyclic ring for 12.
The position of the hydroxyl group was conˆrmed by
the HMBC cross peak of Me-25 with C-6. The
relative stereochemistry was determined from the
NOESY spectrum: clear NOEs of Me-24 and Me-25,
and of Me-23 with H-5, indicating a chair structure
for the monocyclic ring. All the NMR data support
sity. This loss of the enzyme activity would be at-
tributable to the signiˆcantly decreased binding with
1, which resulted from the increased steric bulk size.
We have previously indicated that an appropriate
bulk size is required for eŠective cation-p interaction;
the F365Y mutant had a greater reaction velocity
than the F365W mutants.6) The His mutant had
higher activity, when compared to the Ala and the
Asp mutants (Table 2). This may provide additional
evidence for the crucial role of aromatic p-electrons
12 being 6
a
-hydroxyachilla-9(
E
),13(
E
),17(
E
),21-
at position 420, since His moiety not only abstracts a
proton as a base, but also stabilizes the transient in-
tetraene, which has previously been isolated from the
21)
fern, Polypodiodes formosana
.
termediate through cation-p interaction.
As already described, prolonged incubation with a
large amount of Y420G- and Y420A-mutated SHC
aŠorded two bicycles (14 and 15) and one tricycle (16)
as abortive cyclization products, as shown in Table 3.
Functional analysis of Tyr420
Figure 4 shows the speciˆc enzyme activities for the
formation of 2 plotted against the incubation temper-

1666
T. SATO et al.
Table 2. Kinetic Data and Optimal Temperatures for the Wild-type and Mutated SHCs
The kinetic values of Km and kcat were determined from Lineweaver-Burk plots by using 5
The enzyme activities were assayed by estimating the amount of 2 produced after incubating after 45
m
g each of the homogeneously puriˆed SHCs.
C for 60 min, where no thermal denatu-
9
ration occurred. In the cases of the three mutants of Y420W, L607F and L607W, the production of all the enzymic products (2, 3 and abor-
tive cyclization products 11–16) was too small to estimate the kinetic parameters.
SHC
Km
(
m
M
)
kcat (min„1
)
kcat Km
Relative activity (
100.0
z
)
Optimal temperature (
60
9
C)
W
Wild-type
1.62
×
10
51.8
3.19
Y420G
Y420A
Y420D
Y420H
Y420F
Y420W
2.56
9.09
×
×
102
10
1.21
13.2
15.1
46.6
14.4
ND
4.73
1.45
×
×
10„3
0.1
4.5
50
55
55
50
65
55
10„1
„1
×
×
8.22 10
1.83 10
5.7
6.89
×
10
6.76
×
1.56
ND
10„1
21.2
48.8
ND
×
0.926 10
ND
„1
„1
×
×
L607A
L607I
L607F
L607W
8.78 10
57.8
59.9
ND
ND
6.58 10
20.6
16.3
ND
ND
55
55
55
55
×
×
11.5 10
5.21 10
ND
ND
ND
ND
ND: No enzymic product including the abortive cyclization products (11–16) was found.
The Y420F and Y420D mutants aŠorded only bicy-
clic 14 and 15. The Y420H gave mono-13, and bicy-
clic 14 and 15. It is noteworthy that the Trp mutant
gave signiˆcant amounts of monocyclic 11 and 13,
together with relatively small amounts of bicycles 14
and 15 and a trace of 2. Except for the Trp mutant,
all the mutants aŠorded a signiˆcant amount of 2.
The ratio of the product distribution is summarized
in Table 3. All the mutants other than the Trp mutant
produced large amounts of bicyclic compounds,
indicating that Tyr420 was situated close to the C-8
cation of 5 during the polycyclization reaction and
thus could stabilize cationic intermediate 5 through
the phenolic hydroxyl group. The signiˆcantly bulky
Trp residue, which was incorporated in the mutated
SHC, may have interfered with the 2nd cyclization
reaction, thus resulting in a high production ratio of
monocyclic 13 to bicyclic 14 and 15.
Functional analysis of Leu607
As shown in Fig. 4B, the mutations of Leu into Ala
and Ile dramatically decreased the speciˆc activity for
the production of 2. This could be explained in terms
of the increased Km values, but not of the decreased
k
cat values (Table 2); that is, the bindings with 1 were
dramatically decreased, but the cyclization velocities
were nearly the same. The Ala mutant had sig-
niˆcantly lower binding to 1 by 5.4 fold, indicating
that the enzyme perturbation had occurred, possibly
due to the markedly lower bulk size of Ala. However,
it is surprising that the mutation of Leu into Ile gave
rise to a markedly increased Km value by 7.1 fold,
despite the steric bulk size of Ile being almost equiva-
lent to that of Leu. The structural diŠerence is just a
branching point: iso-butyl for Leu and sec-butyl for
Ile. The remarkable diŠerence in Km value between
the Leu and the Ile mutants strongly indicates that
the isobutyl residue of Leu 607 more precisely ˆtted
to the folded product-like conformation of 1, when
Fig. 4. Speciˆc Enzyme Activities of the Mutated and Wild-type
SHCs Plotted against the Incubation Temperatures, These
Being Estimated from the Production of Hopene 2.
One mg of 1 was incubated with 5
puriˆed SHC for 60 min at pH 6.0.
(A) Mutated SHCs targeted at Y420:
$
mg of a homogeneously

wild-type,

Y420G,
#
Y420A,

Y420D,

Y420H,

Y420F, Y420W (B) Mu-
$
tated SHCs targeted at L607:
L607F, L607W

wild-type,

L607A, L607I,

×

Tyr420 and Leu607 of Squalene Cyclase
1667
Table 3. Product Distribution (
z
) of the Wild-type and Mutated SHCs Estimated by GC Analyses
One mg of 1 was incubated with a large amount of the cell-free homogenate from 50 ml of cultured cells, equivalent to 0.5 mg of each pure
SHCs, under the conditions of pH 6.0, 45 C and 16 h. Under these conditions, the conversion ratio of 1 to all the enzymic products (2, 3 and
abortive cyclization products 11–16), was as follows: 95.3 for Y420G, 17.5 for Y420W, 26.5 for L607F, 20.8 for L607W and 100
9
z
z
z
z
z
for the other mutants, while the wild-type SHC consumed all of 1 in a short incubation time of 60 min. The small conversion ratios of
Y420W, L607F and L607W indicate that the activity for the formation of all the enzyme products was decreased and that some of 1 was left
unchanged.
Monocyclic
Bicyclic
Tricyclic
Hopene
Hopanol
Total yield of
SHC
2
3
11
12
13
14
15
16
11–16
Wild-type
84.1
15.9
—
—
—
—
—
—
—
Y420G
Y420A
Y420D
Y420H
Y420F
Y420W
47.9
62.7
63.4
63.7
75.1
0.4
9.3
11.7
12.3
12.1
15.0
—
—
—
—
—
—
9.6
—
—
—
—
—
—
—
—
—
9.1
—
69.9
11.3
6.9
7.6
1.5
0.9
28.1
16.8
16.7
13.6
9.0
3.4
1.9
—
—
—
42.8
25.6
24.3
24.2
9.9
12.2
4.9
—
99.6
L607A
L607I
L607F
L607W
84.8
84.1
1.6
15.2
14.3
—
—
—
9.0
10.1
—
—
—
45.3
74.6
—
—
21.1
10.4
—
—
17.9
—
—
—
—
—
—
1.6
98.1
100
1.6
4.8
4.9
—
—
compared to the sec-butyl group of Ile. The enzyme
activities of the L607F and L607W mutants for the
formation of 2 were almost quenched (Fig. 4B and
Table 2), suggesting that this position could not work
to stabilize the intermediary carbocation through
Discussion
The formation mechanisms for products 12–16 are
explained as follows. Monocyclic 12 and 13 could
have been formed from monocyclic cation 4;
deprotonation at the 25-position gave 13, while 12
was produced as a result of nucleophilic attack of
water molecule to the cation 4. Deprotonation reac-
tions at the 7- and 26-positions of bicyclic intermedi-
ate 5 aŠorded 14 and 15, respectively. Compound 16
could have been produced by the deprotonation of
cation-
p interaction. The signiˆcantly decreased
speciˆc activity, i.e. the initial velocity (Fig. 4 and
Table 2) should be explained in terms of the bulkiness
of Phe and Trp. The bulky substituents may have dis-
turbed the access of 1 to the protonation site (the
DXDDTA motif5)) which is responsible for initiating
the polycyclization reaction. However, the prolonged
incubation of the L607F and L607W mutants aŠord-
ed signiˆcant amounts of mono- and bicyclic com-
pounds 11–15 without completing the polycyclization
to ˆnal product 2 (Table 3). No tricyclic product such
as 16 was found. This distribution pattern for the
abortive cyclization products suggests that L607 was
located around the B-ring formation site. The L607F
mutant gave signiˆcant amounts of bicyclic products
14 and 15, but the L607W mutant aŠorded monocy-
clic 11–13 in larger amounts than the L607F mutant
did. This observation indicates that the larger the
steric bulk size, the higher the production of monocy-
clic compounds. Therefore, Leu607 may possibly be
positioned around the B-ring formation site, the bulk
size at this position being critical to the optimal fold-
ing of 1 leading to 5. A comparison of the product
distribution pattern of the Tyr420 mutants with that
of the Leu607 mutants suggests that the location sites
of the two residues in the interior cavity were close to
each other. The position of Leu607 may possibly be
closer to the A-ring formation site compared to that
of Tyr420, because the L607W mutant gave higher
production of monocycles than the Y420 W mutant
(Table 3).
Me-27 from the 6 6 5-fused tricyclic cation, whose
stereochemistry at C-13 was opposite to that of true
intermediate 6.
WW
The kinetic results for the Tyr420 mutants
(Table 2) clearly indicate that the function was to
stabilize the transient carbocation formed during the
polycyclization reaction. Figure 4(A) and Table 2
show that the higher the electron density, the faster
the reaction rates. The p-electron density of Tyr (the
wild type) is higher than that of Phe (the Y420F
mutant), while the steric bulk size of Tyr (the wild-
type) is somewhat larger than that of Phe (the Phe
mutant), leading to enzyme perturbation. This may
have caused the higher thermal stability of the wild-
type (Fig. 4 and Table 2); the optimal temperature
for the wild-type having the Tyr residue (60
lower than that for the Phe mutant (65 C). This in-
9C) was
9
ference is consistent with our previous report that the
replacement of F365 and F605 by Tyr enhanced the
reaction velocity, but the optimal temperatures was
decreased.6,10) It is thus evident that the function of
Tyr420 was for stabilizing the cationic intermediate.
The decreased electron density of the Phe mutant,
compared to that of the wild-type (Tyr residue),
could aŠord bicyclic 14 and 15, in spite of the in-

1668
T. SATO et al.
the LDAO inhibitor, thus the planar skeleton of the
three Phe residues could stabilize the cationic inter-
mediates through a cation-
interaction.6,7,10) On the
p
other hand, the hydroxyl group of the Tyr420 residue
was arranged to point toward the inhibitor inside the
interior cavity.16,17) The unshared electron pair of the
hydroxyl group, whose electron density was further
enhanced by the extended resonance through the
phenyl p-electron ring system, may have been placed
in the proximity to the C-8 cation of 5, thus resulting
in greatly enhanced stabilization of 5. Alternatively,
the negative charge of the phenolate of Tyr420 may
have worked for cation stabilization. It was recently
reported that a variety of the abortive cyclization
products, including bicyclic 14 and 15, tetra- and
pentacyclic skeletons, are accumulated in nearly
equivalent amounts in Zymomonas mobilis cells.23)
Z. mobilis SHC has Phe residue at the position corre-
sponding to Tyr420 of A. acidocaldarius SHC.15) The
formation of bicyclic 14 and 15 by Z. mobilis SHC
may possibly be attributable to the poorer electron
density of Phe than of Tyr, as demonstrated here
(Tables 2 and 3). The production of energetically un-
favorable product 15 was higher than that of favora-
ble 14 (a deprotonation product according to Zait-
sev's rule), possibly suggesting that a basic amino
acid for abstracting a proton from Me-26 of 5 may
have been located near the Tyr420 residue. The Gly
and the Ala mutants having smaller bulk size
aŠorded tricyclic 16 (a Markovnikov adduct). The
stereochemistry of 13
WW
a
-
H
is opposite to that of a true
) 6 (see
6 6 5-fused tricyclic intermediate (13b-H
Scheme 1),7,9) suggesting that erroneous folding of 1
may have occurred during the polycyclization proc-
ess. The signiˆcantly altered enzyme structure, which
was veriˆed by the markedly increased Km values
(15.8-fold for the Gly mutant and 5.6-fold for the
Ala mutant), may have led to this erroneous folding.
This ˆnding may suggest that an appropriate steric
bulk size is required at this position to complete the
normal polycyclization. However, this function is
minimal, because production of 16 is too low to
allocate the role of stereochemical control to this
Fig. 5. (A) Natural Sesqui- and Diterpenes Containing 1,5,6-
Trimethylcyclohexene Ring.
A chair conformation was adopted for the A-ring formation
to yield the stereochemistry of 5R,6S in the ring of 17–19. The
hydride, methyl shifts and proton elimination reactions proceed-
ed in an anti-parallel concerted manner. (B) Boat Form to Give
the Stereochemistry of 5R,6R in the Cyclohexene Ring of 11.
creased binding to 1 (Table 2). In addition, the loss of
-electrons by introducing an aliphatic amino acid
p
halted the polycyclization at the bicyclic ring stage,
leading to the high production of 14 and 15 (Table 3).
These ˆndings strongly indicated that the Tyr420
residue was oriented near the C-8 cation of 5 in the
reaction cavity. We have previously shown that
residues Phe3656) and Tyr60911) also worked to stabi-
lize the C-8 cation in 5. The amounts of bicyclic com-
pounds produced by the F365A, Y609A and Y420A
position (only 2–3z of all the enzymic products,
Table 3). The Y420H mutant gave monocyclic 13,
having an exocyclic methylene group, suggesting that
the His residue could also have had a basic function
responsible for proton elimination from Me-25 of 4,
in addition to the cation-stabilizing function (a close
kcat value to that of the wild-type, Table 2).
Of particular interest is the fact that the Y420W
and L607W mutants both produced a novel product
11 having the 1,5,6-trimethylcyclohexene ring. The
CD spectra of the mutants were indistinguishable
from that of the wild-type (data not shown), suggest-
ing that there was little alteration of the protein struc-
ture by the mutations. A few sesqui- and diterpenes
having the trimethylcyclohexene ring have been
mutants were ca. 96
indicating that stabilization of the C-8 cation may
have been attained mainly by the -electrons of
z, 50z and 24z, respectively,
p
Phe365, but further enhanced by those of Tyr609 and
Tyr420. The results of the X-ray analysis show that
the faces of the expanded
p-electron system of the
F365, F601 and F605 residues were in the direction of

Tyr420 and Leu607 of Squalene Cyclase
1669
found in the liverwort, Ptychanthus striatus24) and in
the Okinawan sea sponge (Fig. 5).25) However, the
stereochemistry at the 6-position of the trimethylcy-
clohexene ring (IUPAC numbering) in 11 is opposite
to that of striatene 17 and striatol 18 (sesquiterpenes)
and agelasidine B 19 (diterpene). Compounds 17–19
could be produced via an organized chair structure
20, which underwent the sequential 1, 2-shift reac-
tions of hydride and methyl group with subsequent
deprotonation (Fig. 4(A)), in anti-parallel concerted
manner. On the other hand, 11 could have been
formed via boat structure of 21 as shown in Fig. 5(B),
taking into consideration the stereochemistry
tained for the polycyclization reaction. The exact
steric bulk size of some or numerous active site
residues may possibly guide the single stereoisomer,
despite numerous isomers being possible. It has re-
cently been shown that the mutation of Ile into Leu
in cycloartenol synthase resulted in the production of
lanosterol and parkeol besides cycloartenol, in spite
of a little diŠerence in the bulk size.14) This report
also validates the importance of the correct bulk size
for directing the structural diversity of natural triter-
penes.
In conclusion, this is the ˆrst report of a boat
structure being constructed during the cyclization
process by altering the enzyme structure, despite
SHC always adopting an all pre-chair conformation
for the construction of 2 and 3. The introduction of a
diŠerent steric bulk size at a given position(s) as the
protocol for site-directed mutagenesis will be promis-
ing in future for creating unnatural natural products.
(5R,6R) of the trimethylcyclohexene ring (IUPAC
numbering). We have most recently reported that a
constrained boat form was also constructed from the
cyclization of the 10-ethylated-2, 3-oxidosqualene by
lanosterol synthase, leading to the monocyclic triter-
pene having the 2,3,4-trimethylcyclohexanone ring;26)
the bulky ethyl group enforced a boat conformation
upon the substrate analog, once cyclization had start-
ed in the enzyme cavity. The large steric bulk size of
the Trp residue of the Y420W and L607W mutants
would have signiˆcantly interfered with the access of
1 to the protonation site (the DXDDTA motif5)),
resulting in a remarkably decreased initial velocity
(Fig. 4 and Table 2). However, once cyclization had
started during the prolonged incubation, 1 would
have been constrained to fold with a boat structure
21; the large bulk size of the replaced Trp residue
would have disturbed the formation of the normal
chair conformation. This notion is in good agree-
ment with the fact that the bulky ethyl group (incre-
ment of the C1 unit to 1) at the 10-position imposed a
boat structure to give the trimethylcyclohexanone
ring. However, the Trp mutant could still have
adopted a chair form for the A-ring formation (12
and 13); the amounts of 12 and 13 produced from the
chair structure of 4 were higher than that of 11 der-
ived from boat intermediate 21 (the yield of 11 was
Acknowledgments
We are grateful for the ˆnancial supports (Nos.
13660150 and 13306009) provided by the Ministry of
Education, Science and Culture of Japan. We are
also indebted for the ˆnancial support from the
Fujisawa Foundation which was given to T. H.
References
1) Hoshino, T., Insight into the active sites of squalene
cyclase and the cyclization mechanism (in Japanese).
Bioscience & Industry, 59, 167–170 (2001).
2) Hoshino, T., and Sato, T., Squalene-hopene cyclase:
catalytic mechanism and substrate recognition. J.
Chem. Soc. Chem. Commun., 291–301 (2002).
3) Abe, I., Rohmer, M., and Prestwich, G. D., En-
zymatic cyclization of squalene and oxidosqualene to
sterols and triterpenes. Chem. Rev., 93, 2189–2206
(1993).
4) Wendt, K. U., Schulz, G. E., Corey, E. J., and Liu,
D. R., Enzyme mechanisms for polycyclic triterpene
formation. Angew. Chem. Int. Ed., 39, 2812–2833
(2000).
5) 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. Biosci. Biotechnol. Biochem., 63, 2189–2198
(1999).
only ca. 10
z
for all the enzymic products of the Trp
mutants). Intermediate 4 could have undergone sub-
sequent cyclization to give bicyclic 14 and 15, but
their amounts were relatively decreased with increas-
ing amounts of monocyclic 12 and 13 when the bulki-
ness was increased (Table 3, compare L607W with
L607F). Thus, the bulky Trp residue acted to inter-
fere with the 2nd cyclization process (4
ª5).
It is noticeable that a slight diŠerence in the steric
bulk size at position 607 (compare Leu with Ile) had a
remarkable in‰uence on the substrate recognition
6) Hoshino, T., and Sato, T., Functional analysis of
phenylalanine 365 in hopene synthase, a conserved
amino acid in the families of squalene and oxidosqua-
(
K
m), although the product distribution and the reac-
lene cyclases. J. Chem. Soc. Chem. Commun.
2005–2006 (1999).
,
tion velocity ( cat) were almost the same as those of
k
the wild-type (Tables 2 and 3). Substitution with the
larger bulk size of Phe or Trp gave rise to a boat
structure during the cyclization process. These ˆnd-
ings seem to be informative in answering the question
of why regio- and stereochemical control was at-
7) Hoshino, T., Kouda, M., Abe, T., and Ohashi, S.,
New cyclization mechanism for squalene: a ring ex-
pansion step for the ˆve-membered C-ring intermedi-
ate in hopene biosynthesis. Biosci. Biotechnol.

1670
T. SATO et al.
Biochem., 63, 2038–2041 (1999).
cyclase at 2.0 Å resolution. J. Mol. Biol., 286,
175–187 (1999).
18) Merkofer, T., Pale-Grosdemange, C., Wendt, K. U.,
Rohmer, M., and Poralla, K., Altered product pat-
terns of a squalene-hopene cyclase by mutagenesis of
active site residues. Tetrahedron Lett., 40, 2121–2124
(1999).
19) Pale-Grosdemange, C., Merkofer, T., Rohmer, M.,
and Poralla, K., Production of bicyclic and tricyclic
triterpenes by mutated squalene-hopene cyclase.
Tetrahedron Lett., 40, 6009–6012 (1999).
8) Sato, T., Abe, T., and Hoshino, T., On the cycliza-
tion mechanism of squalene: a ring expansion process
of the ˆve-membered D-ring intermediate. J. Chem.
Soc. Chem. Commun., 2617–2618 (1998).
9) Hoshino, T., Abe, T., and Kouda, M., Unnatural
natural triterpenes produced by altering isoleucine
into alanine at position 261 in hopene synthase and
the importance of having the appropriate bulk size at
this position for directing the stereochemical destiny
during the polycyclization cascade. J. Chem. Soc.
Chem. Commun., 441–442 (2000).
10) Hoshino, T., Kouda, M., Abe, T., and Sato, T.,
Functional analysis of Phe605, a conserved aromatic
amino acid in squalene-hopene cyclase. J. Chem. Soc.
Chem. Commun., 1485–1486 (2000).
11) Sato, T., and Hoshino, T., Catalytic function of the
residues of phenylalanine and tyrosine conserved in
squalene-hopene cyclases. Biosci. Biotechnol.
Biochem., 65, 2233–2242 (2001).
12) Hart, E. A., Hua, L., Darr, L. B., Wilson, W. K.,
Pang, J., and Matsuda, S. P. T. Directed evolution to
investigate steric control of enzymatic oxidosqualene
cyclization. An isoleucine-to-valine mutation in
cycloartenol synthase allows lanosterol and parkeol
biosynthesis. J. Am. Chem. Soc., 122, 9887–9888
(1999).
20) Schmitz, S., Fu
ä ll, C., Glaser, T., Albert, K.,
and Poralla, K., Speciˆc production of
g-
polypodatetraene or 17-isodammara-20(21), 24-diene
by squalene-hopene cyclase mutant. Tetrahedron
Lett., 42, 883–885 (2001).
21) Arai, Y., Hirohara, M., Ageta, H., and Hsuä , H. Y.,
Fern constituents: two new triterpenoid alcohols
with mono- and bi-cyclic skeletons, isolated from
Polypodiodes formosana. Tetrahedron Lett., 33,
1325–1328 (1992).
22) Shiojima, K., Arai, Y., Masuda, K., Kamada, T.,
and Ageta, H., Fern constituents: polypodatetraenes,
novel bicyclic triterpenoids, isolated from Polypodia-
ceous and Aspidiaceous plants. Tetrahedron Lett.
,
24, 5733–5736 (1983).
23) Douka, E., Koukko, A.-I., Drainas, C., Grosdeman-
ge-Billiard, C., and Rohmer, M., Structural diversity
of the triterpenic hydrocarbons from the bacterium
Zymomonas mobilis: the signature of defective squa-
13) Joubert, B. M., Hua, L., and Matsuda, S. P. T.,
Steric bulk at position 454 in Saccharomyces
cerevisiae lanosterol synthase in‰uences B-ring for-
mation but not deprotonation. Org. Lett., 2, 339–341
(2000).
lene cyclization by the squalene hopene cyclase.
W
FEMS Microbiol. Lett., 199, 247–251 (2001).
24) Takeda, R., Naoki, H., Iwashita, T., Mizukawa, K.,
Hirose, Y., Ishida, T., and Inoue, M., Sesquiterpe-
noid constituents of the liverwort, Ptychanthus stria-
14) Matsuda, S. P. T., Darr, L. B. Hart, E. A., Herrera,
J. B. R., McCann, K. E., Meyer, M. M., Pang, J.,
and Schepmann, H. G., Steric bulk at cycloartenol
synthase. Position 481 in‰uences cyclization and
deprotonation. Org. Lett., 2, 2261–2263 (2000).
15) Tippelt, A., Jahnke, L., and Poralla, K., Squalene-
hopene cyclase from Methylococcus capsulatus
tus (Lehm. et Lindenb. Nees). Bull. Chem. Soc. Jpn.
56, 1125–1132 (1983).
,
25) Nakamura, H., Wu, H., Kobayashi, J., Kobayashi,
M., Ohizumi, Y., and Hirata, Y., Agelasidines.
Novel hypotaurocyamine derivatives from the
Okinawan sea sponge Agelas nakamurai Hoshino. J.
Org. Chem., 50, 2494–2497 (1985).
(Bath):
a bacterium producing hopanoids and
steroids. Biochim. Biophys. Acta, 1391, 223–232
(1998).
16) Wendt, K. U., Poralla, K., and Schulz, G. E., Struc-
ture and function of a squalene cyclase. Science, 277,
1811–1815 (1997).
17) Wendt, K. U., Lenhart, A., and Schulz, G. E., The
structure of the membrane protein squalene-hopene
26) Hoshino, T., and Sakai, Y., Enzymic products of the
2,3-oxidosqualene analog having an ethyl residue at
10-position. First trapping of the trimethylcyclohexa-
none ring by lanosterol synthase. Tetrahedron Lett.
42, 7319–7323 (2001).
,