Single-Molecule Studies on a FRET Biosensor: Lessons from a Comparison of Fluorescent Protein Equipped versus Dye-Labeled Species

Abstract

Bacterial periplasmic binding proteins (PBPs) undergo a pronounced ligand-induced conformational change which can be employed to monitor ligand concentrations. The most common strategy to take advantage of this conformational change for a biosensor design is to use a Förster resonance energy transfer (FRET) signal. This can be achieved by attaching either two fluorescent proteins (FPs) or two organic fluorescent dyes of different colors to the PBPs in order to obtain an optical readout signal which is closely related to the ligand concentration. In this study we compare a FP-equipped and a dye-labeled version of the glucose/galactose binding protein MglB at the single-molecule level. The comparison demonstrates that changes in the FRET signal upon glucose binding are more pronounced for the FP-equipped sensor construct as compared to the dye-labeled analog. Moreover, the FP-equipped sensor showed a strong increase of the FRET signal under crowding conditions whereas the dye-labeled sensor was not influenced by crowding. The choice of a labeling scheme should therefore be made depending on the application of a FRET-based sensor.

Keywords: Förster resonance energy transfer (FRET); biosensor; conformational change; fluorescent protein (FP); glucose sensor; hinge motion; ligand binding; single molecule studies.

Figure A1

Amino acid sequence of the MglB protein is given as used in this study. The two inserted cysteine residues at positions 42 and 137 are highlighted cyan and purple, respectively. The C-terminal His-tag is highlighted green.

Figure 1

(upper panels) Illustration of chosen dye labeling positions to design a FRET-based glucose sensor: 3D structures of ligand-free (left structure; 2FW0) and ligand-bound MglB (right structure; 2FVY) shown together with calculated AVs for donor (blue volumes, attached at position 137) and acceptor (red volumes, attached at position 42) dyes. Note the conformational change of the MglB structure induced by glucose (orange spheres) binding. (lower panels) For comparison, an equivalent illustration of both MglB structures is shown with donor FPs (blue, inserted at position 12) and acceptor FPs (red, inserted at the C-terminus). The relative orientations of the FPs are arbitrary and not based on experimental measures.

Figure 2

Comparison of smFRET histograms obtained from FP-equipped (left column) and dye-labeled (right column) species of MglB for four different conditions (from top to bottom): ligand-free state, ligand-bound state at the concentration near the K_d (for values see Figure 3), fully ligand-bound state and ligand-free state in presence of 10% (w/w) PEG 6,000 as a crowding agent. Data shown for the FP-equipped sample were taken from [22]. For the sample labeled with fluorescent dyes a vertical line corresponding to the position of the ligand-free state, is drawn for reference.

Figure 3

FRET parameter as a function of glucose concentration for the two kinds of MgIB-based sensors. For the FP-equipped constructs ensemble data were used (taken from [22]) to calculate a FRET ratio R = I_A/I_D (where I_A and I_D are the peak intensities of the acceptor and the donor emission, respectively) (left panel), while for the dye-labeled construct the smFRET statistical weights of the liganded population are reported (right panel).