Biology Reference
In-Depth Information
DIRECTED EVOLUTION
Even if a protein structure is known,
site-directed mutations predicted by structural
and/or computational analyses to improve a function of interest are often deleterious or
do not provide the expected change in function due to unknown complexities of protein
structure and biochemistry. It is important to remember that properly folded and
functional proteins are typically only marginally stable relative to their unfolded or
non-functional state(s).
16
Even seemingly minor perturbations in sequence can lead to
significantly reduced stability or function. A simple demonstration of this is the
observation that a large majority of protein variants containing a single amino acid
substitution relative to the wild-type or parent sequence result in reduced or eliminated
parent fitness. One can therefore imagine a number of ways in which a mutation
predicted to improve binding toward a specific ligand results in a less stable protein,
even if the mutation does indeed improve binding (which may or may not be detected,
depending on the binding assay conditions).
'
rational
'
For these reasons an evolutionary approach to protein engineering is often preferred.
In general this entails generating genetic diversity in the form of a mutant library, followed
by screening the resulting protein variants for the improved properties of interest. The more
closely the screening conditions reflect the actual conditions in which the protein is to
have the desired functions, the more likely the resulting variant isolated will be useful.
Hence the directed evolution adage
(see below for examples
of screening assays used in directed evolution). For example, if stability is not included as a
fitness parameter in a screening assay, one might end up with a less stable but more active
variant. The required stability of course depends on the specific application.
'
you get what you screen for
'
Below we outline commonly used methods to generate libraries of protein variants. The
parent or wild-type protein may be a single sequence, and amino acid changes are simply
substitutions at various positions throughout that sequence. Alternately, for the case of
recombination, more than one gene sequence (corresponding to more than one protein
sequence) serves as the template for generating genetic (and hence protein) diversity,
and diversity is assessed by the number of crossovers between templates, as well as the
relative differences in amino acid sequence between variants of interest and their parents.
27
GENERATING SEQUENCE DIVERSITY
Random Mutagenesis
Random mutagenesis can be an attractive approach under a variety of circumstances in
which a protein function is to be improved by directed evolution. When little or nothing is
known about a protein
function relationships, this is the obvious
choice both to gain functional insights and to isolate improved variants. But even when
a great deal is known about the target protein, screening random mutations throughout
the sequence is still likely to yield improved variants when more targeted mutagenesis
techniques do not. A common strategy is to first generate and screen random mutation
libraries to identify potential
'
s structure or sequence
amino acid positions and regions of a protein,
which are then targeted for further mutagenesis. Random mutagenesis can also be applied
to specific gene segments, such as a gene region corresponding to one domain of a
multidomain protein.
'
hot spot
'
A variety of techniques have been developed to randomly introduce nucleotide substitutions
into gene segments. Early random mutagenesis methods involved exposing whole cells to
mutagenic conditions, such as UV exposure,
17
X-ray radiation,
18
or chemical mutagens.
19,20
'
strains have been constructed, for example
E. coli
XL1-red which is deficient in
three primary DNA repair pathways (
mutS
,
mutD
, and
mutT
).
21
Amplifying a plasmid via
growth in a mutator strain produces randomly mutagenized plasmid libraries. Since
Mutator
'
