Quantitative analysis of multi-protein interactions using FRET: application to the SUMO pathway

Abstract

Protein-protein binding and signaling pathways are important fields of biomedical science. Here we report simple optical methods for the determination of the equilibrium binding constant K(d) of protein-protein interactions as well as quantitative studies of biochemical cascades. The techniques are based on steady-state and time-resolved fluorescence resonance energy transfer (FRET) between ECFP and Venus-YFP fused to proteins of the SUMO family. Using FRET has several advantages over conventional free-solution techniques such as isothermal titration calorimetry (ITC): Concentrations are determined accurately by absorbance, highly sensitive binding signals enable the analysis of small quantities, and assays are compatible with multi-well plate format. Most importantly, our FRET-based techniques enable us to measure the effect of other molecules on the binding of two proteins of interest, which is not straightforward with other approaches. These assays provide powerful tools for the study of competitive biochemical cascades and the extent to which drug candidates modify protein interactions.

Figure 1.

(A) Schematic diagrams of the 6-His-TEV fluorescent protein constructs, FRET signals arising between tags of bound proteins (excitation at 400 nm leads to emission at 530 nm) and the nonbinding case where no FRET takes place (excitation at 400 nm leads to emission at 475 nm). (B) Normalized absorption spectra (dashed lines) and emission spectra (solid lines, excitation 400 nm) of cyan fluorescent protein (CFP) and yellow fluorescent protein (YFP). (C) Fluorescence emission spectra (excitation 400 nm) demonstrating FRET between 1 μM CFP–SUMO1 and 1 μM YFP–Ubc9 fusion proteins (solid black line), emission of the individual components at this concentration (dashed lines) and the sum of their signals (gray solid line).

Figure 2.

Steady-state FRET binding assay. (A) Fluorescence emission spectra (excitation 400 nm) recorded during the titration of YFP–Ubc9 into 1.1 μM CFP–SUMO1. YFP–Ubc9 was added in increments up to 8 μM. The solid black line is the pre-titration emission. (B) Fluorescence emission at 530 nm of the FRET binding assay: YFP–Ubc9 added to CFP–SUMO1 (solid black diamonds, data from spectra shown in A); control for nonspecific interactions, YFP–Ubc9 added to 1.1 μM CFP (gray squares); measurement of dilution, buffer added to CFP–SUMO1 (control 1, gray triangles); measurement of direct excitation, YFP–Ubc9 added to buffer (control 2, black triangles). The sum of controls 1 and 2 is displayed as a separate series (white triangles).

Figure 3.

Steady-state FRET binding assay: data analysis. (A) FRET data as a function of total YFP–Ubc9 concentration, following the subtraction of controls accounting for dilution (control 1) and direct excitation (control 2). (B) Bound protein determined from the FRET data as a function of free YFP–Ubc9 concentration, fitted by a hyperbola with B _max = 1.1 μM and K_d = 0.59 ± 0.09 μM.

Figure 4.

Time-resolved FRET binding assay. (A) Selected time-resolved fluorescence traces (excitation 397 nm, emission 475 nm) of solutions with various amounts of YFP–Ubc9 added to CFP–SUMO1, resulting in ratios of 1:0 (pre-titration, black), 1:1 (gray), 1:2 (dark gray), and 1:3 (light gray). The instrumental response is recorded by means of a scattering solution (dashed gray). (B) Fluorescence lifetimes of solutions of CFP-SUMO1 and varying YFP-Ubc9 concentrations. (C) Bound protein versus free YFP–Ubc9 calculated from the CFP lifetime at 475 nm.

Figure 5.

Multi-protein binding assay. (A) Emission spectra (excitation 400 nm) of a three-protein binding titration, beginning with 1 μM CFP–SUMO1 and 1 μM YFP–Ubc9 solution adding in untagged RanBP2(l) in increments up to 3 μM as indicated. The arrow indicates the decrease in FRET signal at 530 nm. Dilution during the titration cancels the increase in the 475 nm peak due to decreasing energy transfer. (B) Inhibition curve displaying percent of initial CFP–SUMO1 + YFP–Ubc9 complex split as a function of RanBP2(l) added. Data are derived from the 530-nm peak decrease in A and fitted with a single exponential.

Figure 6.

(A) FRET signal monitored as the 530 nm/480 nm ratio for three titrations beginning with an equimolar solution of 1 μM CFP–SUMO1 and YFP–Ubc9 and adding SUMO1 (circles), RanBP2(s) (squares), and RanBP2(l) (diamonds). (B) FRET signal monitored as the 530 nm/480 nm ratio for a titration beginning with an equimolar solution of 1 μM CFP–RanBP2(1) and YFP–Ubc9 and adding SUMO1.