Challenges in Modelling Metalloenzymes
519
the molybdenum enzymes as well as structural distortion of metal sites in many
other enzymes; for more examples, see recent review by Stanek et al. [79] As a
representative example here, we take the case of molybdenum proteins. The strained
entatic state was observed especially for DMSO reductase family of molybdenum
enzymes that harbour a bis-molybdopterin cofactor (Moco). Such cofactor comprises
Mo ion in its centre, two pterins, each providing two dithiolene ligands to Mo, a
protein residue of serine, cysteine, aspartate or selenocysteine, depending on the
enzyme, as well as one oxo ligand, the latter being sensitive to the Mo-oxidation
state. The structurally extended pterin ligands interact with protein surrounding via
multiple hydrogen bonds and ionic interactions, and as a result, the strain is introduced
to the structure of the Mo complex. This effect can be structurally quantified by
dihedral angle between planes of pterin ligands, referred to as an angle φ, similar
to the Rây–Dutt twist used in studies of arsenate oxidase [80], or by S 1 -S 2 -S 3 -S 4
dihedral, which was used in studies on ethylbenzene dehydrogenase [81, 82].
The oxidized (Mo
VI ) Moco is six-coordinate and contains the oxo (O) ligand.
Despite the presence of six ligands, in most of the enzymes the Mo coordination
does not attain a (distorted) octahedral geometry, which is typical for its model
complexes [83, 84], but rather a distorted trigonal prism, where each trigonal face
is formed by two sulphurs from opposite dithiolenes and one oxygen atom [85].
The S 1 -S 2 -S 3 -S 4 dihedral in oxidized enzymes attains values in the range from
−20° to −30° [86–88]. On the other hand, the reduced (Mo
IV ) Moco is usually fivecoordinate (the prior O ligand is already transferred to the reagent) and attains a shape
of tetragonal pyramid with sulphur atoms of pterins forming its square base with relatively small twist. Alternatively, the product, which is loosely bound to Mo(IV) by
the dative bond, can serve as the sixth ligand. Such geometries were confirmed experimentally by X-ray crystallography results for: (a) the structure of non-active resting
state of EBDH [89], in which the reduced Moco features S 1 -S 2 -S 3 -S 4 dihedral of
−4.86°, (b) the reinterpreted structure of the pentacoordinate Moco of formate dehydrogenase H from E. coli (FDH) with S 1 -S 2 -S 3 -S 4 dihedral of −6.23° [90], as well
as (c) reduced perchlorate reductase (PChR) with the substrate analogue bound to
Moco (hexacoordinate, PDB: 5CHC) or without the substrate (pentacoordinate, PDB:
4YDD), where S 1 -S 2 -S 3 -S 4 dihedrals are in the range from −6° to −9° [91].
The relaxed nature of the almost planar arrangement of pterins in the reduced
molybdenum enzymes is confirmed by geometry optimization in vacuum, which
yields almost identical conformation as that observed in the EBDH crystal structure
(S 1 -S 2 -S 3 -S 4 dihedral of −5.3°).
The strain imposed by the protein on Mo coordination geometry is usually modelled by freezing the position of the fringe hydrogen atoms that are saturating bonds
at places where the cofactor was truncated. However, in contrast to the amino acid
residues, it is much more difficult to select truncation and constrain sites that will
allow both to simulate the real flexibility of the cofactor on the one hand and to
properly model its electronic characteristics on the other. For example, constraining
H atoms (or sometimes carbon atoms of CH 3 groups introduced instead of H) of the
− S-CH=CH-S
− gives a very rigid structure, and as a result, the modelled reaction
pathway may be biased towards oxidized or reduced geometry of Moco, depending
Précédent

- 528/540

Suivant