5.1 QM Cluster Model
In this model, the ligand and binding site residues are extracted and treated using
electronic structure theory. Since the binding/catalytic site residues are mostly
dictating the binding energies with ligand and the rest of the residues only play
supportive role and are contributing to retain the structure of the enzyme, in particular the binding site conformation, this is reasonable approximation. Since not all
the amino acids of enzyme are included in the calculation, certain approximations
need to be applied. To avoid spurious charge accumulation in dangling bonds
which might alter the energetics of the whole protein–ligand systems, the cut bonds
are capped with hydrogens. Since the rigidity of the binding site was mostly stabilized by the rest of the protein, the free optimization of cluster might lead to
unrealistic distortions in the binding site geometry. So, certain terminal residues are
fixed in the space, only partial optimizations are performed, and the energies are
computed for these geometries. In certain cases, the QM cluster is placed in continuum solvent to mimic the protein-like environment and the dielectric constant for
the medium is chosen to be 4 [38]. There was the use of more than one quantum
mechanical theory in some cluster calculations. For example as in the case of 8‐Cl
TIBO bound to human immunodeficiency virus reverse transcriptase, authors
employed two-layer and three-layer ONIOM (in particular [MP2/6‐31G(d),
B3LYP/6‐31G(d,p) and PM3]) approach to estimate the interaction energy. The
residues closer to ligand are described using the high-level theory (like MP2) as
these contribute to total interaction energy dominantly, while the residues far away
from the ligand can be described using low-level theory as these contributions will
not be very significant [39]. There are not many studies which employ this
methodology to compute ligand binding energies or interaction energies with
receptor [38, 40–42]. However, for modelling a number of enzymatic reactions, this
method has been used successfully. In particular, the study on the enzymatic
reaction of acetylene hydratase to produce vinyl alcohol using two different
approaches, namely QM cluster model and QM/MM model, is worth recalling [43].
5.2 Hybrid QM/MM Approach
This approach combines the best of the two worlds, namely force-fields and electronic structure theory-based approach. Even though the receptor–ligand complex
system is too large in length scale, most of the time the region of relevance to us is
the ligand and certain residues that are in direct contact with the ligand. So, it is a
smarter idea to split the system into two regions and use the more accurate level of
theory (here, it is electronic structure theory) to describe the region of relevance and
to use a relatively less accurate but cheaper (here, it is force-field) approach to
describe the rest of the region. However, the harder part is the description of the
232
N. A. Murugan et al.
In this model, the ligand and binding site residues are extracted and treated using
electronic structure theory. Since the binding/catalytic site residues are mostly
dictating the binding energies with ligand and the rest of the residues only play
supportive role and are contributing to retain the structure of the enzyme, in particular the binding site conformation, this is reasonable approximation. Since not all
the amino acids of enzyme are included in the calculation, certain approximations
need to be applied. To avoid spurious charge accumulation in dangling bonds
which might alter the energetics of the whole protein–ligand systems, the cut bonds
are capped with hydrogens. Since the rigidity of the binding site was mostly stabilized by the rest of the protein, the free optimization of cluster might lead to
unrealistic distortions in the binding site geometry. So, certain terminal residues are
fixed in the space, only partial optimizations are performed, and the energies are
computed for these geometries. In certain cases, the QM cluster is placed in continuum solvent to mimic the protein-like environment and the dielectric constant for
the medium is chosen to be 4 [38]. There was the use of more than one quantum
mechanical theory in some cluster calculations. For example as in the case of 8‐Cl
TIBO bound to human immunodeficiency virus reverse transcriptase, authors
employed two-layer and three-layer ONIOM (in particular [MP2/6‐31G(d),
B3LYP/6‐31G(d,p) and PM3]) approach to estimate the interaction energy. The
residues closer to ligand are described using the high-level theory (like MP2) as
these contribute to total interaction energy dominantly, while the residues far away
from the ligand can be described using low-level theory as these contributions will
not be very significant [39]. There are not many studies which employ this
methodology to compute ligand binding energies or interaction energies with
receptor [38, 40–42]. However, for modelling a number of enzymatic reactions, this
method has been used successfully. In particular, the study on the enzymatic
reaction of acetylene hydratase to produce vinyl alcohol using two different
approaches, namely QM cluster model and QM/MM model, is worth recalling [43].
5.2 Hybrid QM/MM Approach
This approach combines the best of the two worlds, namely force-fields and electronic structure theory-based approach. Even though the receptor–ligand complex
system is too large in length scale, most of the time the region of relevance to us is
the ligand and certain residues that are in direct contact with the ligand. So, it is a
smarter idea to split the system into two regions and use the more accurate level of
theory (here, it is electronic structure theory) to describe the region of relevance and
to use a relatively less accurate but cheaper (here, it is force-field) approach to
describe the rest of the region. However, the harder part is the description of the
232
N. A. Murugan et al.