444
M. G. Quesne and S. P. de Visser
3 Developments in QM Cluster Calculations
This section presents an overview of the latest developments in QM cluster model
techniques. Several brief case studies are used that highlight specific breakthrough
in this field.
3.1 Co-factor Free Dioxygenase Reaction Mechanism
Our first example concerns one of the most abundant groups of metalloenzymes,
namely the dioxygenases that generally use transition metals to catalyze the activation of molecular oxygen in otherwise spin forbidden reactions [39–42]. However, a
small subgroup of this superfamily is able to direct this triplet to singlet conversion
without the incorporation of a co-factor [43]. One of the earliest examples of these
types of catalysts was (1H)-3-hydroxy-4-oxoquinaldine 2,4-dioxygenase (HOD),
which catalyzes oxygen-dependent N-heteroaromatic ring cleavage, in the (1H)-3
hydroxy-4-oxoquinaldine (QND) substrate [44]. Two recent papers used combined
experimental DFT cluster models and QM/MM techniques to explore how this cofactor-free enzyme is able to replicate metalloenzyme functionality without incorporating a metal or flavin co-factor [45, 46].
Cluster models of various sizes incorporated the same active site residues in both
studies (Fig. 2). The smaller model (shown in the blue circle) includes the substrate
and truncated models of His 251 , Asp 126 , Ser 101 and Trp 160 as well as the backbone of
Trp 36 . The larger model retains all these atoms and also incorporates the imidazole
groups of His 38 , His 100 and His 102 as well as the three water molecules that were
deemed to be within interaction distance of several QM residues. Importantly, the
O 2 molecule was only included in the latter study [45]; while the former used some
QM/MM calculations to validate the cluster model-derived reaction mechanism [46].
The first paper focused on the initial reaction step of HOD catalysis and explored
the effect of amino acid substitution using both experimental, cluster model and
QM/MM techniques [46]. This combined approach highlights the importance of
the His/Asp dyad; whereby, it was determined that His 251 was not basic enough to
catalyze substrate deprotonation alone. Site-specific mutations produced the variants
D126A and H251A that targeted each residue in this dyad separately. These mutants
were then replicated theoretically by removing a carboxylate and an imidazole groups
from the wild-type (WT) cluster models. These cluster models utilized the hybrid
DFT functional UB3LYP [5, 47] in combination with the basis set 6−31G(3d,p), as
implemented in the Gaussian software package [36]. Additionally, QM/MM models
were constructed and optimized with a combination of the CHARMM27 force field
[32] and UB3LYP, as implemented in DL_POLY [48] and TURBOMOLE [49],
respectively. The protocol used here is identical to that discussed in more depth in
Sect. 4; therefore, here we will only state that the protein databank (PDB) file 2WJ4,
a solvent sphere of 35Å and the basis set SV(P) was used in this study.
M. G. Quesne and S. P. de Visser
3 Developments in QM Cluster Calculations
This section presents an overview of the latest developments in QM cluster model
techniques. Several brief case studies are used that highlight specific breakthrough
in this field.
3.1 Co-factor Free Dioxygenase Reaction Mechanism
Our first example concerns one of the most abundant groups of metalloenzymes,
namely the dioxygenases that generally use transition metals to catalyze the activation of molecular oxygen in otherwise spin forbidden reactions [39–42]. However, a
small subgroup of this superfamily is able to direct this triplet to singlet conversion
without the incorporation of a co-factor [43]. One of the earliest examples of these
types of catalysts was (1H)-3-hydroxy-4-oxoquinaldine 2,4-dioxygenase (HOD),
which catalyzes oxygen-dependent N-heteroaromatic ring cleavage, in the (1H)-3
hydroxy-4-oxoquinaldine (QND) substrate [44]. Two recent papers used combined
experimental DFT cluster models and QM/MM techniques to explore how this cofactor-free enzyme is able to replicate metalloenzyme functionality without incorporating a metal or flavin co-factor [45, 46].
Cluster models of various sizes incorporated the same active site residues in both
studies (Fig. 2). The smaller model (shown in the blue circle) includes the substrate
and truncated models of His 251 , Asp 126 , Ser 101 and Trp 160 as well as the backbone of
Trp 36 . The larger model retains all these atoms and also incorporates the imidazole
groups of His 38 , His 100 and His 102 as well as the three water molecules that were
deemed to be within interaction distance of several QM residues. Importantly, the
O 2 molecule was only included in the latter study [45]; while the former used some
QM/MM calculations to validate the cluster model-derived reaction mechanism [46].
The first paper focused on the initial reaction step of HOD catalysis and explored
the effect of amino acid substitution using both experimental, cluster model and
QM/MM techniques [46]. This combined approach highlights the importance of
the His/Asp dyad; whereby, it was determined that His 251 was not basic enough to
catalyze substrate deprotonation alone. Site-specific mutations produced the variants
D126A and H251A that targeted each residue in this dyad separately. These mutants
were then replicated theoretically by removing a carboxylate and an imidazole groups
from the wild-type (WT) cluster models. These cluster models utilized the hybrid
DFT functional UB3LYP [5, 47] in combination with the basis set 6−31G(3d,p), as
implemented in the Gaussian software package [36]. Additionally, QM/MM models
were constructed and optimized with a combination of the CHARMM27 force field
[32] and UB3LYP, as implemented in DL_POLY [48] and TURBOMOLE [49],
respectively. The protocol used here is identical to that discussed in more depth in
Sect. 4; therefore, here we will only state that the protein databank (PDB) file 2WJ4,
a solvent sphere of 35Å and the basis set SV(P) was used in this study.