Analysis of protein-ligand interactions by fluorescence polarization

Abstract

Quantification of the associations between biomolecules is required both to predict and understand the interactions that underpin all biological activity. Fluorescence polarization (FP) provides a nondisruptive means of measuring the association of a fluorescent ligand with a larger molecule. We describe an FP assay in which binding of fluorescein-labeled inositol 1,4,5-trisphosphate (IP(3)) to N-terminal fragments of IP(3) receptors can be characterized at different temperatures and in competition with other ligands. The assay allows the standard Gibbs free energy (ΔG°), enthalpy (ΔH°) and entropy (ΔS°) changes of ligand binding to be determined. The method is applicable to any purified ligand-binding site for which an appropriate fluorescent ligand is available. FP can be used to measure low-affinity interactions in real time without the use of radioactive materials, it is nondestructive and, with appropriate care, it can resolve ΔH° and ΔS°. The first part of the protocol, protein preparation, may take several weeks, whereas the FP measurements, once they have been optimized, would normally take 1-6 h.

Anistropy (A) increases linearly from A_D* to A_DR* as the fraction of bound D^* increases (eqtn 6).

The anisotropy (A) of entirely free D^* (A_D*) and of entirely bound D^* (A_DR*) define the maximal dynamic range of an FP assay, within which all experimental values lie. A_M, the measured A of the experimental incubation, includes contributions from free D^*, and D^* bound specifically to R and non-specifically to other components. A_I, measured uniquely for each experimental condition, is the A measured for each assay, but with all specifically bound D^* displaced by a saturating concentration of a competing ligand (I). A_NS (anisotropy due to non-specific binding) is calculated from A_I using eqtn 7.

Figure 1

Saturable binding of a ligand (D) to its target (R). (a) At equilibrium, increasing concentrations of D (expressed relative to its K_D) bind to R until all of the latter has associated to form DR. Note that when [D] = K_D, 50 % of R has bound to D. (b) Changes in Gibbs free energy are shown for mixtures of D, R and DR (each in their standard states). The standard Gibbs free energy change of the reaction (ΔG°, the difference in the molar Gibbs energies of the reactants and products) is shown and defines ΔG of the (hypothetical) reaction in which 1 mole of R and 1 mole of D combine to form 1 mole of DR (each in their standard states). Because reactions proceed spontaneously in the direction of reduced Gibbs free energy (i.e. yellow→blue, but not blue→yellow on the plot), the association of D and R proceeds to the minimal Gibbs free energy; this then defines the ratio [D][R]/[DR] at equilibrium (i.e. K_D).

Figure 2

Ligand binding analysed by fluorescence polarization. (a) If a fluorophore in which the absorbing and emitting dipoles are parallel is excited with plane-polarized light and remains immobile during its fluorescence lifetime, 60 % of the emitted light will remain polarized with respect to the excitation light (i.e. parallel). If the molecule tumbles during its fluorescence lifetime, less than 60 % of the emitted light will be polarized in the parallel plane and more will be detected in each of the perpendicular planes. (b) Typical layout of an FP apparatus used to read microplates. Excitation light (typically a xenon flash lamp) passes through excitation and polarization (horizontal) filters and is then reflected by a dichroic mirror to the sample. Emitted light passes through a dichroic mirror and is then split equally by a beam-splitter, which directs it to photomultiplier tubes (PMT) via horizontal (A) and vertical (B) polarization filters and then emission filters. Further details of equipment are provided in the Equipment requirements section of the Introduction. (c) A small fluorescent ligand (D^*) free in solution is excited with plane polarized light. Its rapid tumbling (due to its small molecular volume) causes emission of depolarized light and a low A (top). When D^* binds R, it increase its effective volume and therefore tumbles more slowly, causing more emitted light to remain polarized in the same plane as the excitation light; A therefore increases (bottom). This allows FP to be used to measure ligand binding to a larger macromolecule.

Figure 3

Structure of the IP₃ receptor and the ligands used. (a) Schematic representation of a single subunit of a tetrameric IP₃R1 showing key regions: the NT (residues 1-604) which comprises the IBC (residues 224-604) and SD (residues 1-223), and 6 transmembrane domains (TMD) (b) The crystal structure of the IBC with IP₃ bound (PDB, 1N4K) is shown, highlighting the 2-O-atom (arrow on the enlarged view) to which fluorescein 5-isothiocyanate is attached by a short linker to give FITC-IP₃. (c) Structures of IP₃, FITC-IP₃ and adenophostin A.

Figure 4

Experimental design flow chart. The protocol is divided into two parts: (i) protein preparation and (ii) anisotropy measurements and analysis.

Figure 5

Expression, purification and quantification of NT fragments of IP₃R. (a) Silver-stained gel and Western blot (WB) showing the NT (~67 kDa, arrows) and smaller contaminating proteins. Molecular weight markers are shown on the left. The relative intensities of the three major bands were similar in WB and after silver-staining, indicating that the smaller proteins probably correspond to C-terminally truncated fragments of the NT. These smaller products are unlikely to bind IP₃ (Refs.^,. (b) The Scatchard plot shows typical results from a saturation binding assay using ³H-IP₃, NT (30 ng per incubation) in TEM. From this plot, the density of binding sites (B_max, intercept on the x-axis) and affinity (−K_A, slope of the line) can be determined. Although the Scatchard plot conveniently presents binding results, non-linear curve-fitting to a Hill equation (Step 53) provides a more accurate means of determining B_max and K_A. Results, means ± SEM, n = 3, are reproduced from with permission from The American Society for Pharmacology and Experimental Therapeutics.

Figure 6

Typical 96-well plate layout for the FP saturation binding assay. All wells in columns A-H contain FITC-IP₃ (0.5 nM). Protein dilutions (1 -300 nM) are labelled R1-R15. Columns A-D include no further additions (allowing A_M to be determined); columns E-H also include a saturating concentration of IP₃ (10 μM, allowing A_NS to be determined). Columns I-L include only the serial dilutions of protein (to allow background fluorescence to be measured). Wells C8 and D8 include only FITC-IP₃ (to allow A_D* to be determined). Further details in Step 46A. The layout is typical, but it is advisable to vary the layout between experiments to avoid systematic errors in automated pipetting, etc.

Figure 7

Typical results from FP analysis of equilibrium ligand binding to N-terminal IP₃ receptor fragments (a) FP saturation binding assay at 4 °C using FITC-IP₃ (0.5 nM) and the indicated concentrations of NT (Step 53). (b) FP competition binding assay using FITC-IP₃ (0.5 nM), NT (80 nM) and the indicated concentrations of either IP₃ or adenophostin A at 4 °C (Step 54) (c) van’t Hoff plots for IP₃ and adenophostin A binding to the NT (Step 56). Results (a-c) are means ± S.E.M., n = 3. The same code applies to panels b-c. All FP analyses were performed in Ca²⁺-free CLM. Data reproduced from Ref. with permission from The American Society for Pharmacology and Experimental Therapeutics.