A FRET-Based Biosensor for the Src N-Terminal Regulatory Element

Abstract

In signaling proteins, intrinsically disordered regions often represent regulatory elements, which are sensitive to environmental effects, ligand binding, and post-translational modifications. The conformational space sampled by disordered regions can be affected by environmental stimuli and these changes trigger, vis a vis effector domain, downstream processes. The disordered nature of these regulatory elements enables signal integration and graded responses but prevents the application of classical approaches for drug screening based on the existence of a fixed three-dimensional structure. We have designed a genetically encodable biosensor for the N-terminal regulatory element of the c-Src kinase, the first discovered protooncogene and lead representative of the Src family of kinases. The biosensor is formed by two fluorescent proteins forming a FRET pair fused at the two extremes of a construct including the SH4, unique and SH3 domains of Src. An internal control is provided by an engineered proteolytic site allowing the generation of an identical mixture of the disconnected fluorophores. We show FRET variations induced by ligand binding. The biosensor has been used for a high-throughput screening of a library of 1669 compounds with seven hits confirmed by NMR.

Keywords: NMR; c-Src; fluorescence; fuzzy complexes; high-throughput screening; intrinsically disordered proteins (IDP).

Figure 1

(A) Domain architecture and cartoon representation of CLOBY. (B) Fluorescence emission spectra of intact CLOBY (blue) and after treatment with TEV protease (red). Excitation wavelength was 506 nm and protein concentration was 1 µM.

Figure 2

NMR chemical shift perturbation of CLOBY (A,B) and TEV-treated CLOBY (C,D) with respect to the isolated SNRE region of Src without the fluorescent proteins. Chemical shift differences were measured at 278 K for the disordered regions (A,C) and at 298 K for the SH3 domains (B,D). Red lines represent five standard deviations from the mean of the lowest 10% CSD. SAXS data of CLOBY are presented in panel (E). The main figure is the intensity profile, while the inserts represent a normalized Kratky plot (upper right) and a Guinier plot (lower left). Panel (F) presents the R_g distribution of the EOM-selected subensembles that best reproduce experimental SAXS data (blue) compared with the R_g distribution of the initial pool (in black).

Figure 3

Concentration dependence of the normalized fluorescence of CLOBY (A) and TEV-treated CLOBY (B). The comparison of CLOBY and CLOBY-TEV presented in panel (C) highlights that the intermolecular FRET in the two samples is the same, and that intramolecular FRET dominates the fluorescence spectra at the 1 μM concentration used for screening experiments. The hydrodynamic radius measured by dynamic light scattering (D) at much higher concentrations (up to 275 μM for intact CLOBY) is also concentration dependent and provides an estimated dimer dissociation constant of 0.7 mM, confirming that self-association is negligeable at the working concentrations used in this study.

Figure 4

(A) CLOBY FRET is decreased in the presence of the VLS12 peptide, known to disturb the fuzzy complex of the SNRE. (B) Statistics of the first screening results based on fluorescence changes at 530 nm. Green and red bars represent the number of compounds increasing or decreasing the fluorescence by more than 3 standard deviations (indicated by **), respectively, only in intact CLOBY. The blue bar corresponds to colored compounds that could not be analyzed in the high-throughput assay. Orange bar corresponds to compounds with an effect (positive or negative) between one and two standard deviations (indicated by *).

Figure 5

NMR validation of FRET results. Representative results of the three classes of hits defined by comparing fluorescence data and NMR perturbations. In each panel the fluorescence of CLOBY and CLOBY-TEV (at 1 μM concentration), in the presence or in the absence of excess drug (100 μM) are shown in the left. Chemical shift perturbations caused by the drug in the isolated SNRE, i.e., in the absence of the fluorescent proteins, are shown in the right. The effect on the disordered region (at 278 K) is shown on top and that that on the SH3 domain (at 298 K) are shown below. (A): GW4064; (B): Eltrombopag; (C): Flavoxate.