On the cyclization mechanism of squalene: A ring expansion process of the five-membered D-ring intermediate

Article abstract of DOI:10.1039/a806948d

Site-directed mutagenesis experiments with W169F, W169H and W489F for the squalene-hopene cyclase, and the formation of 10 possessing the five-membered D-ring and a tetrahydrofuran moiety as the enzyme product of the analogue 8 with a hydroxy group, strongly suggest that a ring expansion reaction from the five- to the six-membered ring is responsible for the D-ring formation of hopene.

Full text of DOI:10.1039/a806948d

On the cyclization mechanism of squalene: a ring expansion process of the
five-membered D-ring intermediate
a
b
ab
Tsutomu Sato, Takamasa Abe and Tsutomu Hoshino*
a
Graduate School of Science and Technology, and
b
Department of Applied Biological Chemistry, Faculty of Agriculture, Niigata University, Ikarashi, Niigata
50-2181, Japan. E-mail: hoshitsu @agr.niigata-u.ac.jp
9
Received (in Cambridge, UK) 7th September 1998, Accepted 26th October 1998
Site-directed mutagenesis experiments with W169F, W169H
and W489F for the squalene-hopene cyclase, and the
formation of 10 possessing the five-membered D-ring and a
tetrahydrofuran moiety as the enzyme product of the
analogue 8 with a hydroxy group, strongly suggest that a
ring expansion reaction from the five- to the six-membered
ring is responsible for the D-ring formation of hopene.
W489F). The wild-type has a catalytic optimum at 60 °C and pH
3,5
6.0. A large scale incubation of 1 (150 mg) for 16 h with a
cell-free extract from a 6 l culture of W169F afforded 2, 3 and
4 (an oil) (100.5, 10 and 6 mg, respectively) after the separation
2
with a SiO column (hexane–EtOAc). The incubation of other
mutants conducted with the same quantities of 1 and cell-free
extracts as for W169F gave the following isolated yields: for
mutant W169H: 2 (73.5 mg), 3 (6.3 mg) and 4 (33.0 mg), and for
mutant W489F: 2 (36.8 mg), 3 (3.3 mg) and 4 (9.8 mg). Detailed
The cyclization of squalene 1 into pentacyclic triterpenes, hop-
6
22(29)-ene 2 and hopan-22-ol 3, is an outstanding reaction from
2D NMR analyses revealed that 4 had a dammarene skeleton
the point of view of both stereo- and regio-chemical specificity,
with the formation of five new carbon–carbon bonds and nine
with an exomethylene group, but the 17-side chain had an a-
orientation (17-epi-dammarene). No other product was detected
in the reaction mixtures except for recovered 1.
1
new chiral centers. Oxidosqualene also undergoes the poly-
1
olefin cyclization analogous to squalene. Recent progress on
Concomitant production of 4 together with the two normal
products 2 and 3 indicates that a common intermediate 5 is
being produced during the polyolefin cyclization process
(Scheme 2). Formation of the dammarene cation 5 having a
6/6/6/5-fused ring system is thermodynamically favored by
Markovnikov control. Proton elimination from the methyl
group would give 4 [path (b)], while the ring expansion process
from a five- to six-membered D-ring would give the hopanyl
C22-cation 7 via a disfavored anti-Markovnikov C-17 cation 6
[path (a)], the latter being formed after the ring expansion of 5
has been processed. The ring expansion competes with the
deprotonation. Steric factors also favor the formation of 5; given
that the cyclization reaction proceeds by adopting a pre-chair
conformation 1a for the D-ring construction, greater repulsion
would occur due to the 1, 3-diaxial arrangement between the
two methyls at the 15- and 19-positions (squalene numbering),
thus resulting in the less hindered conformation 1b.
the two cyclases of squalene and oxidosqualene has spurred
mechanistic studies of the polycyclization reactions. Squalene-
hopene cyclase (SHC) is believed to fold the linear molecule 1
into an all pre-chair conformation 1a inside the enzyme cavity,
leading to 2 and 3 through the generation of a series of
2
carbocation intermediates (Scheme 1). Scheme 1 also shows
that the C- and D-rings are formed by anti-Markovnikov
closures. Site-directed mutagenesis experiments of SHC re-
vealed that both D-376 and D-377 were crucial for the
catalysis.3 Recently, an X-ray analysis of Alicyclobacillus
4
acidocaldarius SHC has been reported. We have independ-
ently succeeded in an overexpression of the SHC and reported
the first identification of the tryptophan residues 169 and 489 as
components of the active sites; substitution of these tryptophans
with valine and leucine by point mutations resulted in complete
5
loss of the enzyme activity. Here, we report that mutants of
W169F, W169H and W489F produce the normal cyclization
products 2 and 3 together with an abnormal tetracyclic product
A similar ring expansion has been proposed for the C-ring
formation in lanosterol biosynthesis, based on the trapping of
4
consisting of a 6/6/6/5-fused ring system, the formation of 4
being in agreement with the Markovnikov rule. This finding
leads us to propose that a ring expansion reaction is involved in
the D-ring formation of 2 and 3.
Cell-free homogenates of the mutants prepared by site-
directed mutagenesis were incubated with 1 at optimal tem-
peratures (45 °C for W169F and W169H, and 53 °C for
Enz-AH+
1b
Markovnikov Closure
Ha
H
17
b
D
C
20
19
21H
H
H
Enz-AH+
15
5
1a
(a)D-Ring Expansion
(b) Deprotonation
Longer life time of 5
1
7
E
H
H
H
13
H
17
1
9
18
11
C
D
29
12
20
A
B
10
2
1
9
H
8
15
6
anti-Markovnikov
Cation
14
16
21
6
3
4
5
7
2
30
H
28
4
OH
7
3
=
Anti-Markovnikov addition
2
+
3
Scheme 1
Scheme 2
Chem. Commun., 1998, 2617–2618
2617

that the W169 would bind to 1 rather than stabilizing the
carbocation, while W489 may exhibit both binding and cation
stabilization, and also suggest that the higher electron density of
the p-electrons, the greater affinity to 1. The looser binding of
the phenylalanine or histidine residues to 1, near the D-ring, in
the mutant SHCs would lead to the longer lifetime of 5, as
inferred from the thermodynamic and steric preferences.
Compared to W169F, W169H significantly increased the
amount of 4 5.5-fold; the histidine residue may abstract a proton
from the 21-methyl [path (b)], indicating that the position of
W169 in the cavity may possibly be close to the 21-methyl of 5,
but further evidence is required to confirm this.
OH
OH
Enz-AH+
Enz-AH+
8
a
8
b
H
H
O
O
9
10
Scheme 3
Notes and references
1 I. Abe, M. Rohmer and G. D. Prestwich, Chem. Rev., 1993, 93, 2189.
the five-membered C-ring intermediate (a 6/6/5-fused ring)
from incubation experiments with substrate analogues. Incuba-
tion of the squalene analogue 8 (C27-OH), prepared via
7
2
3
G. Ourisson, M. Rohmer and K. Poralla, Annu. Rev. Microbiol., 1987,
1, 301.
C. Feil, R. Sussmuth, G. Jung and K. Poralla, Eur. J. Biochem., 1996,
42, 51.
4
treatment of H
reduction with LiAlH
5
IO
6
with 2,3-oxidosqualene followed by
, with the wild-type SHC afforded 106
4
2
almost quantitatively (Scheme 3). Compound 9 and other
products were not detected. Formation of 10 strongly supported
the suggestion that the cyclization reaction proceeded via the
prefolded 8b (like 1b), but not through 8a (like 1a), and also
gave unequivocal evidence for the involvement of a discrete
metastable C-20 carbocation intermediate like 5 prior to the ring
expansion and further cyclization; the hydroxy group would
have attacked the tertiary C-20 cation thus produced due to its
highly nucleophilic nature, resulting in the formation of a
tetrahydrofuran ring in 10. A dammarene cation similar to 5 was
postulated for the cyclization mechanism of 2,3-dihydrosqua-
4
5
K. U. Wendt, K. Poralla and G. E. Schulz, Science, 1997, 277, 1811.
T. Sato, Y. Kanai and T. Hoshino, Biosci. Biotechnol. Biochem., 1998,
6
2, 407.
6 All the HRMS (EI) and NMR (H-H COSY 45, HOHAHA, NOESY,
DEPT, HMQC and HMBC) spectra were consistent with the proposed
structures of 4 and 10.
7
E. J. Corey and H. Cheng, Tetrahedron Lett., 1996, 37, 2709; E. J.
Corey, S. C. Virgil, H. Cheng, C. H. Baker, S. P. T. Matsuda, V. Singh
and S. Sarshar, J. Am. Chem. Soc., 1995, 117, 11819; T. Hoshino and Y.
Sakai, Chem. Commun., 1998, 1591.
I. Abe and M. Rohmer, J. Chem. Soc., Perkin Trans. 1, 1994, 783.
I. Abe, T. Dang, Y. F. Zheng, B. A. Madden, C. Fei, K. Poralla and
G. D. Prestwich, J. Am. Chem. Soc., 1997, 119, 11333.
8
9
8
9
lene and 29-methylidene-2,3-oxidosqualene by SHC.
Since the mutants of W169V and W489L were completely
10 K. Poralla, Bioorg. Med. Chem. Lett., 1994, 4, 285; I. Abe and G. D.
Prestwich, Proc. Natl. Acad. Sci. U.S.A., 1995, 92, 9274; D. A.
Dougherty, Science, 1996, 271, 163.
1 The mutations gave a lowering of the optimal temperature, but no
change with pH. Reactions with 5 mg of purified SHC were conducted
at 30 °C and pH 6.0 for 1 h; thermal denaturation of the SHCs was not
5
inactive, it appears that the tight binding to 1 comes from the
aromatic ring residue, not from the hydrophobic aliphatic
residues of SHC. To date, cation–p interactions induced by
aromatic moieties, resulting in the carbocation stabilization,
have been proposed for the catalysis and/or acceleration of the
1
found. The kinetic values of K
Lineweaver–Burk plots as follows: K
maxs: 0.09, 0.078, 0.045 and 0.017 nmol min mg , respectively, for
the wild-type, W169F, W169H and W489F.
m
and Vmax were determined from
10
polycyclization reaction. Kinetic values for the mutants were
m
s: 16.7, 277, 280 and 92 m ; and
M
compared with that of the wild-type.11 For the mutant W169F,
21
21
V
K
m
increased 17-fold, but Vmax remained unchanged. On the
other hand, for the mutant W489F, K increased 5.5-fold, but
max was only 14% of the wild type. These kinetic results imply
m
V
Communication 8/06948D
2618
Chem. Commun., 1998, 2617–2618