Biology Reference
In-Depth Information
Optimization Methods for
De Novo Protein Design
C. A. Floudas & H. K. Fung
The de novo peptide and protein design, first suggested almost two
decades ago, begins with a postulated or known flexible protein three-
dimensional structure and aims at identifying amino acid sequence(s)
compatible with this structure. Initially, the problem was denoted as
the “inverse folding problem'' [1,2] since protein design has intimate
links to the well-known protein folding problem [3]. In contrast to the
characteristic of protein folding to associate a given protein sequence
with its own unique shape, the inverse folding problem exhibits high
levels of degeneracy, that is, a large number of sequences will be com-
patible with a given protein structure, although the sequences will vary
with respect to properties such as activity and stability.
In silico protein design allows for the screening of overwhelmingly
large sectors of sequence space, with this sequence diversity subse-
quently leading to the possibility of a much broader range of properties
and degrees of functionality among the selected sequences. Allowing
for all 20 possible amino acids at each position of a small 50-residue
protein results in 20
50
combinations, or more than 10
65
possible sequences.
From this large number of sequences, the computational sequence
selection process aims at selecting those sequences that will be com-
patible with a given structure using efficient optimization of energy
functions that model the molecular interactions.
In an effort to make the difficult nature of the energy modeling and
combinatorial optimization manageable, the first attempts at computa-
tional protein design focused only on a subset of core residues and
explored steric van der Waals-based energy functions through exhaustive
searches for compatible sequences [4,5]. Over time, the models have
evolved to incorporate improved rotamer libraries in combination with
detailed energy models and interaction potentials. Although the
consideration of packing effects on structural specificity is sometimes
sufficient, as shown through the design of compatible structures using
backbone-dependent rotamer libraries with only van der Waals energy
evaluations for a subset of hydrophobic residues [6,7], there has been
42
Search WWH ::
Custom Search