200
This greatly facilitates their subsequent purification, in vitro characterization, and engineering by means of rational design or highthroughput screening approaches.
The second consideration concerns structural features of the
protease: Ideally, the N- and C-termini of the protease are located
in close proximity of the active site. In this way, an AI-domain is
displayed at a high effective concentration such that it can effectively compete with the substrate and thus shut down the activity
of the protease transducer in the basal state. Conversely, its activity
can be strongly induced after the AI-domain is irreversibly cleaved
off by an activating protease or reversibly dislodged from the active
site through a binding event. The exact mechanism of activation
ultimately depends on the molecular recognition elements in the
linker connecting the AI-domain with the protease transducer.
Similar considerations concern the choice of ligand-binding receptors that serve as molecular recognition elements to detect the target analyte. Binding receptors can either recognize two distinct
epitopes with separate binding sites on the same target analyte or
feature an allosteric receptor that undergoes a significant conformational change upon binding the target analyte. In case of the
latter, binding receptors that undergo cooperative binding interactions upon associating with their target analyte are preferred as this
overcomes potential problems with multivalent interactions and
the subsequent formation of oligomeric complexes. For instance,
we exploited a binding receptor based on artificially engineered
affinity clamps consisting of a circularly permutated PDZ domain
attached to an enhancer domain that recognizes the PDZ domain
exclusively in its ligand bound form and thus forms a sandwich
complex upon binding its cognate ligand [6, 7].
From a protein engineer’s perspective, it is highly desirable to
recombine autoinhibited protease modules with distinct families of
binding receptors that display similar structural and biophysical
characteristics, yet diverse ligand specificities. In this way, the protein engineering process is rendered as generally applicable as possible. In particular, the orientation of the N- and C-termini is
critical if a distinct class of binding receptors is to be repeatedly
recombined with an autoinhibited protease module with little
empirical optimization to yield synthetic protease sensors with
diverse ligand specificities. Ideally, a distinct class of binding
receptors can be sourced from naturally occurring receptor families. Alternatively, binding receptors can be engineered using standardized work flows as for example recombinant antibody-like
binders that can be derived using a variety of display technologies.
AI-domains are engineered using a combination of rational,
structure- guided protein engineering in combination with highthroughput screening. Notably, for protease-based signal transducers,
the proteolytic cleavage product provides a good molecular lead to
1.3 Choice
of Binding Receptors
as Molecular
Recognition Elements
1.4 General
Considerations
on the Construction
of AI-Domains
Viktor Stein and Kirill Alexandrov
This greatly facilitates their subsequent purification, in vitro characterization, and engineering by means of rational design or highthroughput screening approaches.
The second consideration concerns structural features of the
protease: Ideally, the N- and C-termini of the protease are located
in close proximity of the active site. In this way, an AI-domain is
displayed at a high effective concentration such that it can effectively compete with the substrate and thus shut down the activity
of the protease transducer in the basal state. Conversely, its activity
can be strongly induced after the AI-domain is irreversibly cleaved
off by an activating protease or reversibly dislodged from the active
site through a binding event. The exact mechanism of activation
ultimately depends on the molecular recognition elements in the
linker connecting the AI-domain with the protease transducer.
Similar considerations concern the choice of ligand-binding receptors that serve as molecular recognition elements to detect the target analyte. Binding receptors can either recognize two distinct
epitopes with separate binding sites on the same target analyte or
feature an allosteric receptor that undergoes a significant conformational change upon binding the target analyte. In case of the
latter, binding receptors that undergo cooperative binding interactions upon associating with their target analyte are preferred as this
overcomes potential problems with multivalent interactions and
the subsequent formation of oligomeric complexes. For instance,
we exploited a binding receptor based on artificially engineered
affinity clamps consisting of a circularly permutated PDZ domain
attached to an enhancer domain that recognizes the PDZ domain
exclusively in its ligand bound form and thus forms a sandwich
complex upon binding its cognate ligand [6, 7].
From a protein engineer’s perspective, it is highly desirable to
recombine autoinhibited protease modules with distinct families of
binding receptors that display similar structural and biophysical
characteristics, yet diverse ligand specificities. In this way, the protein engineering process is rendered as generally applicable as possible. In particular, the orientation of the N- and C-termini is
critical if a distinct class of binding receptors is to be repeatedly
recombined with an autoinhibited protease module with little
empirical optimization to yield synthetic protease sensors with
diverse ligand specificities. Ideally, a distinct class of binding
receptors can be sourced from naturally occurring receptor families. Alternatively, binding receptors can be engineered using standardized work flows as for example recombinant antibody-like
binders that can be derived using a variety of display technologies.
AI-domains are engineered using a combination of rational,
structure- guided protein engineering in combination with highthroughput screening. Notably, for protease-based signal transducers,
the proteolytic cleavage product provides a good molecular lead to
1.3 Choice
of Binding Receptors
as Molecular
Recognition Elements
1.4 General
Considerations
on the Construction
of AI-Domains
Viktor Stein and Kirill Alexandrov