Biology Reference
In-Depth Information
“Focused” strategy.
The focused approach to protein engineer-
ing overcomes many of the limitations of the other two approaches.
The structural approach, which is based on global QSAR knowledge,
is inefficient because of the shortage of such knowledge. The func-
tional approach, which is based on
in vitro
directed evolution of pro-
teins, bypasses rational design but has limited capacity to build
proteins to functional specifications. Although still at its inception and
limited in application, the focused approach promises significant
progress in protein engineering due to its capacity for direct genera-
tion of the QSAR knowledge required for the rational design of pro-
teins with predetermined properties. Whereas the structural strategy
reveals gaps in knowledge and the functional approach “breeds”
desired properties without building new knowledge, the focused
strategy breeds new knowledge that is focused on obtaining new pro-
tein structures with desired properties.
At the core of the focused approach is computer-assisted mathe-
matical modeling of the association between the structural and func-
tional properties of proteins. The mathematical model can be built
using a rather limited data set of the protein structural variants dis-
playing variations in its activity. Although the size of the data set
affects the accuracy of the model, the model can be applied as a guide
for the virtual exploration of sequence space in search of protein
structures with improved or new properties. The new structures
found through this virtual process can be constructed and tested
experimentally in the laboratory. If the desired properties are not
matched by the predicted proteins, indicating inaccurate modeling,
the new data obtained from the evaluation can be used to adjust the
model and the mathematical modeling process. The virtual search and
experimental testing can be repeated until satisfactory results are
obtained. Thus, the focused strategy is a cyclic process designed to
accelerate the development or evolution of knowledge by use of a
mathematical model. The major advantage of this approach is its use
of a virtual process instead of experimentation to evaluate structural
variants for a desired function. Although the functional approach can
also screen vast protein sequence spaces using experimental
approaches, it relies on preexisting protein libraries and is impractical
