180
determined by the relative affinities of the output domain interaction
and the magnitude of the input-induced thermodynamic change,
e.g., the strength of ligand binding [2]. Examples of output
domains that have been applied for these approaches include fluorescent or luminescent proteins domains forming FRET or BRET
pairs. To ensure efficient energy transfer between them in either
the on or off state, fluorescent domains can be used that have an
increased tendency to form intramolecular complexes, or their
close proximity can be promoted by so-called helper domains [3].
Our group used these self-associating fluorescent protein domains
to develop robust FRET sensor protein for a range of analytes,
including metal ions, bile acids, and antibodies [4–7]. In addition,
by replacing the self-associating fluorescent domains by an enzymeinhibitor pair with similar affinity, the optical readout of the antibody FRET sensor could be replaced by a more sensitive enzymatic
readout [8].
The modular organization of protein switches based on mutually exclusive domain interactions does not only allow exchange of
different output functions, but is also well suited to combine with
oligonucleotide-based biomolecular systems and networks. Most
approaches reported thus far to control enzyme activity by DNA are
based on the templated assembly of split enzyme fragments [9,
10]. Whereas split reporter enzymes have the advantage of providing a low background activity (at least when used intermolecularly),
split enzymes generally suffer from poor thermodynamic stability
and protein complementation is often not reversible. Therefore, our
group recently introduced an alternative approach that is based on
the reversible assembly of a complex between the reporter enzyme
TEM1-β-lactamase and its inhibitor domain BLIP [11–13]. This
design exploits the inherent difference in mechanical properties
between ssDNA and dsDNA to mechanically disrupt and spatially
separate the enzyme-inhibitor pair upon binding of an input oligonucleotide to the input-binding module. To facilitate straightforward exchange of input sequences, a noncovalent approach is
followed by conjugating both TEM1-β-lactamase and BLIP to individual, 21 nucleotide handle oligonucleotides. By appending antihandle sequences to each end of the synthetic ssDNA target
recognition sequence, a ternary complex is formed upon mixing of
the three protein switch components (Fig. 1). Formation of this ternary complex results in a large increase in effective concentration
and hereby induces an intramolecular interaction between the
enzyme-inhibitor pair. Hybridization of the complementary target
oligonucleotide to the input module results in the formation of a
rigid double-helix which mechanically disrupts the enzyme-inhibitor
interaction, hereby switching on enzyme activity.
Since this system is based on two well-folded and stable protein domains, it is more robust compared to the assembly of splitprotein fragments and readily reversible. Moreover, because of the
Wouter Engelen and Maarten Merkx
determined by the relative affinities of the output domain interaction
and the magnitude of the input-induced thermodynamic change,
e.g., the strength of ligand binding [2]. Examples of output
domains that have been applied for these approaches include fluorescent or luminescent proteins domains forming FRET or BRET
pairs. To ensure efficient energy transfer between them in either
the on or off state, fluorescent domains can be used that have an
increased tendency to form intramolecular complexes, or their
close proximity can be promoted by so-called helper domains [3].
Our group used these self-associating fluorescent protein domains
to develop robust FRET sensor protein for a range of analytes,
including metal ions, bile acids, and antibodies [4–7]. In addition,
by replacing the self-associating fluorescent domains by an enzymeinhibitor pair with similar affinity, the optical readout of the antibody FRET sensor could be replaced by a more sensitive enzymatic
readout [8].
The modular organization of protein switches based on mutually exclusive domain interactions does not only allow exchange of
different output functions, but is also well suited to combine with
oligonucleotide-based biomolecular systems and networks. Most
approaches reported thus far to control enzyme activity by DNA are
based on the templated assembly of split enzyme fragments [9,
10]. Whereas split reporter enzymes have the advantage of providing a low background activity (at least when used intermolecularly),
split enzymes generally suffer from poor thermodynamic stability
and protein complementation is often not reversible. Therefore, our
group recently introduced an alternative approach that is based on
the reversible assembly of a complex between the reporter enzyme
TEM1-β-lactamase and its inhibitor domain BLIP [11–13]. This
design exploits the inherent difference in mechanical properties
between ssDNA and dsDNA to mechanically disrupt and spatially
separate the enzyme-inhibitor pair upon binding of an input oligonucleotide to the input-binding module. To facilitate straightforward exchange of input sequences, a noncovalent approach is
followed by conjugating both TEM1-β-lactamase and BLIP to individual, 21 nucleotide handle oligonucleotides. By appending antihandle sequences to each end of the synthetic ssDNA target
recognition sequence, a ternary complex is formed upon mixing of
the three protein switch components (Fig. 1). Formation of this ternary complex results in a large increase in effective concentration
and hereby induces an intramolecular interaction between the
enzyme-inhibitor pair. Hybridization of the complementary target
oligonucleotide to the input module results in the formation of a
rigid double-helix which mechanically disrupts the enzyme-inhibitor
interaction, hereby switching on enzyme activity.
Since this system is based on two well-folded and stable protein domains, it is more robust compared to the assembly of splitprotein fragments and readily reversible. Moreover, because of the
Wouter Engelen and Maarten Merkx