323
complex strategies based on DNA recombination [5], in which
large DNA fragments (e.g., whole genes, specific protein domains,
or other DNA regulatory regions) are swapped among different
library variants. While these approaches are capable of generating
billions of unique mutant genes (or gene combinations), this
potential is much reduced by the necessary vector cloning and cell
transformation/transfection steps that follow. Both steps have limited efficiencies that significantly reduce the size of the mutant
library that can be generated. Furthermore, the introduction of
random point mutations throughout an entire gene inevitably
results in a large pool of inactive or unchanged mutants [6], which
can be seen as wasted potential. This problem has often been remediated at least in part by restricting the region of mutagenesis. In
this way, known structural features are avoided and the region of
interest is mutated more thoroughly. However, this solution comes
with bias, as catalytic sites are typically targeted because they are
often well defined and perceived as key to changing function [7,
8]. This neglects other possible functional changes, including network rewiring events which are rarely well characterized.
To ensure that network interactions are not excluded, it is thus
recommended that targeting specific gene regions be kept to a
minimum, especially in the absence of knowledge on the protein’s
interaction motifs. Instead, it is better to increase the efficiency of
cloning and transformation. As transformation efficiency is strongly
cell- and vector-dependent, we will not discuss it further. Instead,
we will focus on the mutagenesis and cloning steps. In our first
protocol, we will describe the construction of a library of random
point mutants through a combination of error-prone DNA polymerase, high-fidelity polymerase and type IIS restriction enzymes.
The latter enzymes, reviewed in [9], are notable for their ability to
cut DNA downstream of a recognition sequence. Though they are
useful for scar-free, one-directional cloning, perhaps their greatest
advantage comes from the possibility to design custom nonpalindromic overhangs. These can prevent self-ligation between
inserts (illustrated in Fig. 1) and can therefore dramatically increase
the efficiency of proper ligation.
Alternatively, one can avoid point mutagenesis altogether by
focusing instead on protein domains. Many proteins consist of
functional units, called domains, which can fold independently of
each other and accomplish self-contained functions. This modularity allows different domains to be joined into new proteins and is
thought to be an essential source of novelty in natural evolution
[10]. For instance, in the yeast mating pathway, protein kinases
and scaffolds involved in signal transduction are composed of one
or more interaction domains [11]. When these domains are manually “shuffled” with each other, new functional proteins are
generated and these can confer a novel response to the overall signaling pathway [11–13]. By promoting novel interactions between
Rewiring Signaling Networks
complex strategies based on DNA recombination [5], in which
large DNA fragments (e.g., whole genes, specific protein domains,
or other DNA regulatory regions) are swapped among different
library variants. While these approaches are capable of generating
billions of unique mutant genes (or gene combinations), this
potential is much reduced by the necessary vector cloning and cell
transformation/transfection steps that follow. Both steps have limited efficiencies that significantly reduce the size of the mutant
library that can be generated. Furthermore, the introduction of
random point mutations throughout an entire gene inevitably
results in a large pool of inactive or unchanged mutants [6], which
can be seen as wasted potential. This problem has often been remediated at least in part by restricting the region of mutagenesis. In
this way, known structural features are avoided and the region of
interest is mutated more thoroughly. However, this solution comes
with bias, as catalytic sites are typically targeted because they are
often well defined and perceived as key to changing function [7,
8]. This neglects other possible functional changes, including network rewiring events which are rarely well characterized.
To ensure that network interactions are not excluded, it is thus
recommended that targeting specific gene regions be kept to a
minimum, especially in the absence of knowledge on the protein’s
interaction motifs. Instead, it is better to increase the efficiency of
cloning and transformation. As transformation efficiency is strongly
cell- and vector-dependent, we will not discuss it further. Instead,
we will focus on the mutagenesis and cloning steps. In our first
protocol, we will describe the construction of a library of random
point mutants through a combination of error-prone DNA polymerase, high-fidelity polymerase and type IIS restriction enzymes.
The latter enzymes, reviewed in [9], are notable for their ability to
cut DNA downstream of a recognition sequence. Though they are
useful for scar-free, one-directional cloning, perhaps their greatest
advantage comes from the possibility to design custom nonpalindromic overhangs. These can prevent self-ligation between
inserts (illustrated in Fig. 1) and can therefore dramatically increase
the efficiency of proper ligation.
Alternatively, one can avoid point mutagenesis altogether by
focusing instead on protein domains. Many proteins consist of
functional units, called domains, which can fold independently of
each other and accomplish self-contained functions. This modularity allows different domains to be joined into new proteins and is
thought to be an essential source of novelty in natural evolution
[10]. For instance, in the yeast mating pathway, protein kinases
and scaffolds involved in signal transduction are composed of one
or more interaction domains [11]. When these domains are manually “shuffled” with each other, new functional proteins are
generated and these can confer a novel response to the overall signaling pathway [11–13]. By promoting novel interactions between
Rewiring Signaling Networks