72
fluorophore in a distance-dependent manner, which results in
fluorescence of the acceptor fluorophore. FRET sensors can be
constructed by fusing a specific pair of fluorescent proteins to the
termini of a suitable recognition domain, which undergoes a conformational change when it binds the small molecule of interest.
This ligand-dependent conformational change alters the distance
between the fluorescent proteins, causing an observable change in
FRET efficiency, which can be observed as a change in the ratio of
the fluorescence intensities of the donor and acceptor fluorophores. The dynamic range of the sensor is the maximum change
in FRET efficiency between the unbound and ligand-bound states.
In addition to having a large dynamic range, an ideal sensor would
be specific for the target ligand and responsive over the physiological concentration range of the ligand; importantly, the sensor must
also be stable for extended periods under the experimental conditions (e.g., temperature and pH) required for the biological system
of interest. These properties of the sensor are largely determined
by the choice of recognition domain.
The solute-binding protein (SBP) superfamily is one set of recognition domains that is commonly used in FRET sensors for small
molecules [2–5]. SBPs exhibit high affinity and specificity toward a
diverse array of ligands, and they undergo a large conformational
change upon ligand binding, which can be transduced into an
optical signal in a FRET sensor construct. However, the low thermostability of existing SBPs can limit the utility of the resulting
sensors in biological environments, especially when destabilizing
modifications to the recognition domain, such as circular permutation or specificity-switching mutations, are necessary to improve
the dynamic range or specificity of the sensor [6].
Ancestral protein reconstruction (APR) is one method that has
been shown to produce consistently thermostable proteins, with
denaturation temperatures up to 40 °C greater than comparable
mesophilic proteins [7–11]. We have previously used this method
to reconstruct a thermostable SBP as a recognition domain for a
robust FRET sensor for l-arginine [6]. The main steps involved in
APR are: (1) collection and alignment of a sequence dataset representative of a protein family; (2) inference of a phylogenetic tree
describing the evolutionary relationships between the protein
sequences; (3) probabilistic reconstruction of the ancestral protein
sequences; (4) synthesis and cloning of genes encoding the ancestral proteins; and (5) expression, purification, and biophysical analysis of the ancestral proteins.
It has been argued that the high thermostability of reconstructed
ancestral proteins reflects the environment of the ancient, thermophilic organisms from which they originate [7, 8, 12]; thus, reconstruction of the ancestor of a sufficiently ancient protein family (for
instance, one that predates the divergence of the major bacterial
kingdoms) is a viable method for engineering a highly thermostable
Ben E. Clifton et al.
fluorophore in a distance-dependent manner, which results in
fluorescence of the acceptor fluorophore. FRET sensors can be
constructed by fusing a specific pair of fluorescent proteins to the
termini of a suitable recognition domain, which undergoes a conformational change when it binds the small molecule of interest.
This ligand-dependent conformational change alters the distance
between the fluorescent proteins, causing an observable change in
FRET efficiency, which can be observed as a change in the ratio of
the fluorescence intensities of the donor and acceptor fluorophores. The dynamic range of the sensor is the maximum change
in FRET efficiency between the unbound and ligand-bound states.
In addition to having a large dynamic range, an ideal sensor would
be specific for the target ligand and responsive over the physiological concentration range of the ligand; importantly, the sensor must
also be stable for extended periods under the experimental conditions (e.g., temperature and pH) required for the biological system
of interest. These properties of the sensor are largely determined
by the choice of recognition domain.
The solute-binding protein (SBP) superfamily is one set of recognition domains that is commonly used in FRET sensors for small
molecules [2–5]. SBPs exhibit high affinity and specificity toward a
diverse array of ligands, and they undergo a large conformational
change upon ligand binding, which can be transduced into an
optical signal in a FRET sensor construct. However, the low thermostability of existing SBPs can limit the utility of the resulting
sensors in biological environments, especially when destabilizing
modifications to the recognition domain, such as circular permutation or specificity-switching mutations, are necessary to improve
the dynamic range or specificity of the sensor [6].
Ancestral protein reconstruction (APR) is one method that has
been shown to produce consistently thermostable proteins, with
denaturation temperatures up to 40 °C greater than comparable
mesophilic proteins [7–11]. We have previously used this method
to reconstruct a thermostable SBP as a recognition domain for a
robust FRET sensor for l-arginine [6]. The main steps involved in
APR are: (1) collection and alignment of a sequence dataset representative of a protein family; (2) inference of a phylogenetic tree
describing the evolutionary relationships between the protein
sequences; (3) probabilistic reconstruction of the ancestral protein
sequences; (4) synthesis and cloning of genes encoding the ancestral proteins; and (5) expression, purification, and biophysical analysis of the ancestral proteins.
It has been argued that the high thermostability of reconstructed
ancestral proteins reflects the environment of the ancient, thermophilic organisms from which they originate [7, 8, 12]; thus, reconstruction of the ancestor of a sufficiently ancient protein family (for
instance, one that predates the divergence of the major bacterial
kingdoms) is a viable method for engineering a highly thermostable
Ben E. Clifton et al.