Chemistry Reference
In-Depth Information
Consequently, examination of the location of secondary structural elements,
distribution of polar and nonpolar amino acids, buried H-bond donors and acceptors,
interatomic distances and surface amino acids can help determine the degree of
reliability of a protein model. Chemical environment and its influence on the folding
reliability of the model can also be used to assess the protein model.
Although it may not be conclusive and could be sometimes misleading, the degree
of target-template sequence identity/similarity can be an indicator of the model's
quality. If there is major concern to model accurately the binding site, the fact that
the sequence identity is a global measure may mask the identity of the binding site
area. Proper refinement at backbone and side-chain level may offset low sequence
similarity in binding site areas. The performance of homology models in docking
does not always correlate well with sequence similarity. Despite not being a
sufficient condition, a correct sequence alignment is necessary to develop reliable
models. The final target-template pairwise alignment should be improved by
multiple sequence alignments.
The backbone structure of the template can be obtained from the crude model.
Nonetheless, even with state-of-the-art tools, large-scale backbone displacements
might be challenging. Mid-range and small displacements can be modeled by
molecular dynamics, Monte Carlo based techniques and normal mode analysis-
based procedures. However, this depends on the quality of the sampling and
energy functions used. Consequently single structures may not define well the
target-ligand complex opening space for receptor ensemble docking (RED) as a
valid approach to address protein flexibility in docking procedures.
Since diverse ligands may bind to the active site, binding side receptor flexibility
could be important. A good starting point can be obtained from backbone-
dependent side-chain rotamer libraries as well as information of known ligands.
Protein restraints obtained from previous receptor docking can be used. This is an
ensemble of ligand poses to produce models with diverse side-chain orientations
that is followed by local energy minimization and re-docking. Another approach
is to combine experimental knowledge of ligand binding with
in silico
modeling
of induced fit effects. In LSHM (ligand-steered homology model), known ligands
are used to shape and optimize the binding site through (flexible ligand)-(flexible
