Download Access of the substrate to the active site of squalene and

1202
Biochemical Society Transactions (2005) Volume 33, part 5
Access of the substrate to the active site of
squalene and oxidosqualene cyclases:
comparative inhibition, site-directed mutagenesis
and homology-modelling studies
S. Oliaro-Bosso*, T. Schulz-Gasch†, S. Taramino*, M. Scaldaferri*, F. Viola* and G. Balliano*1
*Dipartimento di Scienza e Tecnologia del Farmaco, Facoltà di Farmacia, Università degli Studi di Torino, Via Pietro Giuria 9, I-10125 Turin, Italy,
and †F. Hoffmann-La Roche Ltd., Molecular Design, PRBD-CS 92/2.10D, CH-4010 Basel, Switzerland
Abstract
Substrate access to the active-site cavity of squalene-hopene cyclase from Alicyclobacillus acidocaldarious
and lanosterol synthase [OSC (oxidosqualene cyclase)] from Saccharomyces cerevisiae was studied by
an inhibition, mutagenesis and homology-modelling approach. Crystal structure and homology modelling
indicate that both enzymes possess a narrow constriction that separates an entrance lipophilic channel
from the active-site cavity. The role of the constriction as a mobile gate that permits substrate passage
was investigated by experiments in which critically located Cys residues, either present in native protein or
inserted by site-directed mutagenesis, were labelled with specifically designed thiol-reacting molecules.
Some amino acid residues of the yeast enzyme, selected on the basis of sequence alignment and a
homology model, were individually replaced by residues bearing side chains of different lengths, charges or
hydrophobicities. In some of these mutants, substitution severely reduced enzymatic activity and thermal
stability. Homology modelling revealed that in these mutants some critical stabilizing interactions could no
longer occur. The possible critical role of entrance channel and constriction in specific substrate recognition
by eukaryotic OSC is discussed.
In cholesterol and ergosterol biosynthesis, the most significant structural alteration occurring along the pathway,
that is generation of the steroid nucleus after assembly of
the triterpene backbone, is brought about by lanosterol
synthase. Lanosterol synthase belongs to the large family
of OSCs (oxidosqualene cyclases), eukaryotic enzymes that
catalyse the cyclization of 2,3-oxidosqualene into different
cyclic compounds: lanosterol alone in non-photosynthetic
organisms (fungi and mammals), cycloartenol, precursor of
phytosterols and other tetra- and pentacyclic compounds in
plants [1].
Prokaryotes possess an enzyme similar to OSCs: SHC
(squalene-hopene cyclase) that converts squalene into hopene
or diplopterol, pentacyclic precursors of hopanoids [2].
Interestingly, while all the known OSCs only accept 2,3oxidosqualene as substrate, SHC not only accepts its physiological substrate squalene, but also the eukaryotes’ substrate,
2,3-oxidosqualene [3].
Squalene and OSCs catalyse among the most fascinating
and complex monoenzymatic reactions (Figure 1). During
the process, which is primed by protonation of the substrate,
four or five rings are formed, several chiral centres are created,
hydride and methyl groups are 1,2 shifted and a proton is
Key words: enzyme model, homology modelling, mutagenesis, oxidosqualene cyclase (OSC),
squalene-hopene cyclase (SHC), thiol inhibitor.
Abbreviations used: OSC, oxidosqualene cyclase; SHC, squalene-hopene cyclase.
1
To whom correspondence should be addressed (email [email protected]).
C 2005
Biochemical Society
extracted (or a hydroxyl added, as in diplopterol formation)
and the cyclic product is released [2]. During the past
50 years, an impressive mass of data has been collected on
these enzymes; investigations began with pure chemistry
and later used inhibitors, mutagenesis studies and homology
modelling [2,4,5]. As for other enzymes however, key clues
to understanding the structure–function relationship came
from solving the crystal structure, which for this class of
lipophilic proteins occurred quite recently. In 1997, the
structure of SHC from Alicyclobacillus acidocaldarious was
solved at medium resolution [6] and a few years later to 2 Å
(1 Å = 10−10 m or 0.1 nm) resolution [7]. More recently, the
same protein was co-crystallized with the close squalene
analogue 2-azasqualene and the complex was analysed by
X-ray diffraction to 2.13 Å [8]. Finally, in 2004, human lanosterol synthase was crystallized and its structure solved by
a Hoffmann–La Roche group [9]. Crystallographic analysis
revealed the presence of a cavity in these proteins, shaped
so as to accept the substrate in a prefolded product-like
conformation (Figure 2). Both SHC and OSC are monotopic
membrane proteins, since they are shaped so as to submerge
into the non-polar part of the phospholipid bilayer from one
side of the membrane without protruding through it [7–9].
Structural analysis has also shown that the protruding
part of the protein contains a lipophilic channel leading to
the active-site cavity. Cavity and channel are separated by a
narrow constriction formed by four amino acid residues that
Seventh Yeast Lipid Conference
Figure 1 Reactions catalysed by (A) prokaryotic SHC and (B) lanosterol synthase
Reproduced with permission from Milla, P., Lenhart, A., Grosa, G., Viola, F., Weihofen, W.A., Schulz, G.E. and Balliano, G.
c Blackwell
(2002) Third-modifying inhibitors for understanding squalene cyclase function. Eur. J. Biochem. 269, 2108–2116 Publishing Ltd.
Figure 2 Surface representation of SHC sliced in the middle of the
molecule
A detergent molecule N,N-dimethyldodecylamine-N-oxide that has
been found in a crystal structure of SHC [11] is shown by the
ball-and-stick model. The catalytic acid Asp376 and the residues forming
the channel constriction are indicated. Reproduced with permission from
Milla, P., Lenhart, A., Grosa, G., Viola, F., Weihofen, W.A., Schulz, G.E.
and Balliano, G. (2002) Third-modifying inhibitors for understanding
c Blackwell
squalene cyclase function. Eur. J. Biochem. 269, 2108–2116 Publishing Ltd.
appear to block it (Figure 2). Two residues of the constriction,
however, belong to a loop that appears to be sufficiently
mobile to permit passage of the substrate in its extended
conformation [7].
When the hypothesis that the substrate entered through
the channel constriction of SHC was first raised, we thought
we had a suitable tool to verify it: one of the four amino acid
residues forming the narrow constriction is a Cys (Cys435 ,
SHC numbering), an amino acid residue that is easy to
modify and we had previously used some thiol reagents as
inhibitors of yeast OSC [10]. A reagent could be used as a
sort of obstruction device at the constriction of SHC: the
entrance gate hypothesis would be confirmed, if an inhibitory
effect occurred. The experiments gave the expected results,
especially significant because we studied not only a wildtype recombinant protein, which possesses five Cys residues
and no disulphide bridges, but also a pair of mutants obtained
by site-directed mutagenesis in the laboratory of Schulz [11]:
a single mutant lacking the Cys at the channel constriction
and a quadruple mutant bearing only the Cys at the channel
constriction. Thiol-reacting agents proved more effective on
the quadruple than on the single mutant [11].
SHC has long been the only template structure for building
homology models that is able to give important mechanistic
insights into oxidosqualene cyclization by eukaryotic OSCs
[12]. It was obvious, however, that SHC was only of limited
use as a model for unravelling the OSC reaction, because of
differences between SHC and OSC in substrate, high-energy
intermediates and products.
The appearance of the SHC structure triggered the hunt
for the first eukaryotic cyclase (OSC) structure and at the end
of 2004, the news that all OSC searchers were awaiting was
announced: the structure of human lanosterol synthase had
been solved [9]. The outstanding achievement obtained with
human OSC should be considered both as the conclusion of
50 years of history, since it permits the mechanism of the first
enzymatic oxidosqualene cyclization to be explored at the
molecular level, and as the beginning of new research projects
based on predicting the structures of all OSCs cloned thus
far, by homology modelling.
Concerning the question of substrate access to the active
site, the structure of human OSC confirmed the existence
of a lipophilic entrance channel and a constriction also in
eukaryotic OSC and allowed single residues or sections of
the peptide backbone, which can alter their conformation
or change their arrangement to permit substrate passage, to
be identified [9]. Although the architecture of the lipophilic
channel was clear from the crystal structure, its role in
C 2005
Biochemical Society
1203
1204
Biochemical Society Transactions (2005) Volume 33, part 5
Figure 3 Alignment of supposed OSCs channel constriction sequences with SHC
Alignments were achieved using the MultAlin program (http://prodes.toulouse.inra.fr/multalin/multalin.html).
substrate–enzyme interaction is still far from being fully
understood. One would wonder, for instance, how specific
the channelling of the substrate to the active site through the
lipophilic channel and constriction actually is. Is the higher
substrate specificity of OSCs (which transform only oxidosqualene) compared with SHCs (which transform both
squalene and oxidosqualene) [3] due to the ability of the
lipophilic channel of OSCs to reject the squalene nonepoxide before entrance into the active-site cavity, or does
the rejection occur after entrance? Several findings suggest
that selection occurs outside the active site. For instance, if an
extended conformation is required for the substrate to pass
through the narrow constriction [7], it might be thought that
the substrate passes the constriction and enters the active-site
cavity with its epoxide edge turned towards the protonating
apparatus located at the top of the active-site cavity. In OSC,
the active-site cavity appears to be more spatially restricted
than that in SHC [13]; in this case, it might not be easy for
the substrate to turn within the active site. If this were the
case, the channelling of properly oriented substrate would
undoubtedly be a task for the lipophilic channel and channel
constriction. Furthermore, metabolic reasons suggest that
selection of substrate preferably occurs outside the active site,
to reduce possible competitive inhibition of OSC by the close
substrate analogue squalene.
We approached the problem of the role of the constriction
region and lipophilic channel in OSCs through inhibition
experiments, site-directed mutagenesis and homology modelling. We planned initially to prepare mutants of lanosterol
synthase from Saccharomyces cerevisiae bearing Cys at
the channel constriction in order to repeat the labelling
experiments with thiol reagents, as we had previously done
with SHC [11]. Thus a suitable control mutant had to be
prepared. All eukaryotic OSCs possess a critical Cys residue
inside the active site, belonging to the highly conserved
catalytic motif DCTAE (455–459) [14,15]. We thus needed
a control protein lacking the active-site Cys residue, to avoid
possible competition for thiol-reacting agents between this
residue and the Cys residues that we planned to insert at the
constriction. Starting from this control mutant, we prepared
channel constriction mutants. Positions where Cys residues
were inserted were selected on the basis of both sequence
alignment (Figure 3) and homology modelling. Mutants were
treated with thiol-reacting agents of different shapes and sizes
and their possible obstructing effect was evaluated to verify
the role of the constriction as a gate for substrate passage. As
C 2005
Biochemical Society
expected, Cys-bearing mutants proved to be more susceptible
to thiol-reacting reagents than the wild-type.
One of the Cys-bearing mutants proved to be less active
and less stable than the control enzyme, thus indicating
that the site-directed mutagenesis substitution had touched
a single critical position of the protein [16]. The homology
model revealed that a section of the entrance constriction
bearing the critical position is crucially linked via hydrogen
bonds to some distant (in sequence) parts of the enzyme.
Additional substitution at this position further confirmed
its critical role and smoothed the way for a more extensive
mutagenesis study.
The results obtained with both mutagenesis and inhibition
experiments suggest that the entrance channel and the channel
constriction could play a critical role in specific substrate
recognition by OSC. Much biochemical investigation will
be required to confirm the proposed role: novel mutants
will have to be tested for their catalytic activity and thermal
stability and important information may also be obtained
by determining the mutants’ susceptibility to known sitedirected inhibitors such as 2-azasqualene [8] and RO 48-8071
[17,18], and by studying the possible competitive effect of
squalene. A challenge raised by the use of human OSC as
a template should be the comparison, through homologymodelling studies, of the three-dimensional structures of all
the cloned OSCs. Comparison of different entrance channels
could clarify the molecular basis of their common ability to
specifically recognize the substrate.
This work was supported by the Ministero dell’Istruzione, Università
e Ricerca (MIUR) (ex 60%). Thanks are due to Professor G. Grosa
(Università del Piemonte Orientale ‘Amedeo Avogadro’, Novara,
Italy) for supplying thiol-reacting agents and to Professor G.E. Schulz
(Universität Freiburg, Germany) for providing the original picture of
Figure 2 [11].
References
1 Nes, W.D. (1990) Recent Adv. Phytochem. 24, 283–327
2 Wendt, K.U., Schulz, G.E., Corey, E.J. and Liu, D.R. (2000)
Angew. Chem. Int. 39, 2812–2833
3 Abe, I. and Rohmer, M. (1994) J. Chem. Soc. Perkin Trans. 1, 783–791
4 Woodward, R.B. and Bloch, K. (1953) J. Am. Chem. Soc. 75, 2023–2024
5 Abe, I., Rohmer, M. and Prestwich, G.D. (1993) Chem. Rev. 90,
2189–2206
6 Wendt, K.U., Poralla, K. and Schulz, G.E. (1997) Science 277, 1811–1815
7 Wendt, K.U., Lenhart, A. and Schulz, G.E. (1999) J. Mol. Biol. 286,
175–187
Seventh Yeast Lipid Conference
8 Reinert, D.J., Balliano, G. and Schulz, G.E. (2004) Chem. Biol. 11, 121–126
9 Thoma, R., Schulz-Gasch, T., D’Arcy, B., Benz, J., Aebi, J., Dehmlow, H.,
Hennig, M., Stihle, M. and Ruf, A. (2004) Nature (London) 432, 118–122
10 Balliano, G., Grosa, G., Milla, P., Viola, F. and Cattel, L. (1993) Lipids 28,
903–906
11 Milla, P., Lenhart, A., Grosa, G., Viola, F., Weihofen, W.A., Schulz, G.E. and
Balliano, G. (2002) Eur. J. Biochem. 269, 2108–2116
12 Schulz-Gasch, T. and Sthal, M. (2003) J. Comput. Chem. 24, 741–753
13 Dehmlow, H., Aebi, J.D., Jolidon, S., Ji, J.H., von der Marck, E.M., Himber, J.
and Morand, O. (2003) J. Med. Chem. 46, 3354–3370
14 Corey, E.J., Matsuda, S.P.T. and Bartel, B. (1994) Proc. Natl. Acad.
Sci. U.S.A. 91, 2211–2215
15 Shi, Z., Buntel, C. and Griffin, J.H. (1994) Proc. Natl. Acad. Sci. U.S.A. 91,
7370–7374
16 Oliaro-Bosso, S., Schulz-Gasch, T., Balliano, G. and Viola, F. (2005)
ChemBioChem, in the press
17 Morand, O.H., Aebi, J.D., Dehmlow, H., Ji, J.H., Gains, N., Langsfeld, H. and
Himber, J. (1997) J. Lipid Res. 38, 373–390
18 Lenhart, A., Reinert, D.J., Aebi, J.D., Dehmlow, H., Morand, O.H. and
Schulz, G.E. (2003) J. Med. Chem. 46, 2083–2092
Received 1 June 2005
C 2005
Biochemical Society
1205