Biology Reference
In-Depth Information
backbone coordinates.
5,6
An alternate approach to using molecular modeling to modify
an existing protein scaffold is the more challenging de novo design of a protein, whereby
amino acids and/or protein secondary structure elements are assembled (computationally)
to achieve the desired structure/function.
6
In the absence of structural information, a number of other experimental and molecular
modeling tools can be used to predict protein structure or structure
function
relationships. Homology modeling is commonly used to guide rational design of a
protein, based on existing structural information for a homologue.
7
Often experimental
results from studies of a protein
s homologues or mutants lead to successful rational
protein design (perhaps in combination with structural insights). For example, one may
compare homologous proteins having nearly identical sequences (or variants of a single
wild-type protein) and identify specific amino acid substitutions that are largely
responsible for differences in specific properties between the homologues. Techniques such
as alanine scanning mutagenesis,
8
in which individual amino acid positions are
substituted with alanine (chosen since it eliminates the side chain beyond the
'
carbon
and is least likely to change the protein conformation or impose electrostatic or steric
effects) are also commonly used to search for residues critical to a protein
β
'
s function and
can furnish insights for rational design.
Rational design often involves a single amino acid position targeted for site-directed
mutagenesis to another amino acid (e.g. to relax an enzyme
s substrate specificity
by increasing the volume of the binding pocket
9
). The more difficult task of identifying
multiple positions to mutate simultaneously to achieve improvements in function due
to coordinating point mutations is most commonly accomplished with the aid of molecular
modeling, but has also been demonstrated via visual structure analysis.
10
'
Another form of rational protein design involves the construction of site-specific protein
chimeras, fusions, and truncations. Common, simple examples are the addition of in-
frame tags (e.g. 6-histidines) and fluorescent protein fusions. More complex examples
involve the rearrangement, removal, or addition of specific protein domains or modules.
Here, a domain is defined as a structurally separable unit within a protein, while a
module is a functionally minimal unit that is transferable from one protein context to
another.
11
With sufficient structural and functional information about a domain or
module, one may be able to rationally combine their respective functions by careful, site-
directed insertions or fusions. However, even a seemingly straightforward protein
construction may not properly fold or retain function for unpredicted reasons. As
described below, the combinatorial approach of constructing chimeric libraries followed
by screening for active proteins can be more effective.
26
A growing area of rational protein design involves the incorporation of noncanonical
amino acids at specific sites in a protein sequence.
12,13
This allows the introduction of novel
chemical functionalities. While the techniques and efficiency of this technology continue
to evolve, the general methodology involves nonsense codon suppression with suppressor
tRNAs and their corresponding aminoacyl tRNA synthetases that have been engineered (using
directed evolution methods) to accept a desired noncanonical amino acid. While proteins
engineered to contain noncanonical amino acids have largely been developed for in vitro
protein applications and as molecular tools to probe biological functions, their use holds
promise for creating new biological functions and chemistries in synthetic biology.
Collectively, rational design has furnished many successfully engineered proteins.
Examples include proteins with improved thermal stability, altered substrate specificity,
proteins engineered to bind metals, and even enzymes with novel catalytic activity.
3,14,15
Despite such successes, in many cases it is overly difficult or impossible to predict a
mutagenesis strategy that will yield a desired protein property.
