Determination of the two-dimensional interaction rate constants of a cytokine receptor complex

Abstract

Ligand-receptor interactions within the plane of the plasma membrane play a pivotal role for transmembrane signaling. The biophysical principles of protein-protein interactions on lipid bilayers, though, have hardly been experimentally addressed. We have dissected the interactions involved in ternary complex formation by ligand-induced cross-linking of the subunits of the type I interferon (IFN) receptors ifnar1 and ifnar2 in vitro. The extracellular domains ifnar1-ectodomain (EC) and ifnar2-EC were tethered in an oriented manner on solid-supported lipid bilayers. The interactions of IFNalpha2 and several mutants, which exhibit different association and dissociation rate constants toward ifnar1-EC and ifnar2-EC, were monitored by simultaneous label-free detection and surface-sensitive fluorescence spectroscopy. Surface dissociation rate constants were determined by measuring ligand exchange kinetics, and by measuring receptor exchange on the surface by fluorescence resonance energy transfer. Strikingly, approximately three-times lower dissociation rate constants were observed for both receptor subunits compared to the dissociation in solution. Based on these directly determined surface-dissociation rate constants, the surface-association rate constants were assessed by probing ligand dissociation at different relative surface concentrations of the receptor subunits. In contrast to the interaction in solution, the association rate constants depended on the orientation of the receptor components. Furthermore, the large differences in association kinetics observed in solution were not detectable on the surface. Based on these results, the key roles of orientation and lateral diffusion on the kinetics of protein interactions in plane of the membrane are discussed.

FIGURE 1

Schematic of the dynamic equilibria of solution and surface interactions involved in the two-step formation and dissociation of the ternary IFN-receptor complex on a membrane (details in the text).

FIGURE 2

Relevance of the two possible dissociation pathways of the ternary complex. (A) Schematic of the experiments: ternary complex on fluid lipid membrane was formed by sequential tethering of (2 fmol/mm²) ifnar2-H10 (1) and (3 fmol/mm²) ifnar1-H10 (2) in stoichiometric amounts, followed by binding ^AF488IFNα2 to form the ternary complex (3). After the second injection of ^AF488IFNα2, additional ifnar1-H10 (top) or ifnar2-H10 (bottom) was rapidly tethered onto the membrane, and dissociation was monitored (4). (B) Course of a typical experiment as monitored by simultaneous TIRFS (top) and RIf (bottom) detection with addition loading of (4 fmol/mm²) ifnar1-H10. (C) Overlay of ligand dissociation curves with (red) and without (black) free ifnar1-H10 on the membrane. (D) Course of a typical experiment as monitored by simultaneous TIRFS (top) and RIf (bottom) detection with additional loading of (7 fmol/mm²) ifnar2-H10. (E) Overlay of ligand dissociation curves with (red) and without (black) free ifnar2-H10 on the membrane.

FIGURE 3

Monitoring two-dimensional dissociation kinetics by pulse-chasing the ternary complex. (A) Principle of surface-dissociation rate-constant determination as detected by FRET: The ternary complex on fluid lipid membrane is formed by sequential injection of ^AF488ifnar2-H10 (1), ifnar1-H10 (2), and ^AF568IFNα2 (3). Equilibrium is then perturbed by rapidly tethering an excess of nonlabeled ifnar2-H10 onto the membrane (4), which exchanges the labeled ifnar2-H10 in the ternary complex (5). (B) Course of a typical experiment monitoring donor fluorescence (green) and acceptor (red trace) fluorescence by TIRFS and the mass loading by RIf (black) (2 fmol/mm^{2 AF488}ifnar2-H10, 5 fmol/mm² ifnar1-H10, 16 fmol/mm² ifnar2-H10). (C) Comparison of the surface dissociation rates from donor (green) and acceptor (red) channels with the dissociation of ^AF568IFNα2 from ifnar2-H10 alone (blue). A control experiment carried out the same way, but with unlabeled ifnar2-H10 in 1 and with direct excitation of ^AF568IFNα2 confirmed negligible ligand dissociation from the surface (black). The residuals from monoexponential curve fits are shown in the bottom.

FIGURE 4

Determination of two-dimensional dissociation rate constants by ligand chasing. (A) Schematic of the assay: Ternary complex on fluid lipid membrane was formed by sequential injection of ifnar2-H10 (1), a large excess of ifnar1-H10 (2), and ^AF488IFNα2 (3). The excess of ifnar1 was then loaded with an unlabeled competitor (4), which binds ifnar1 with high affinity (IFNα2 HEQ) and exchanged the labeled ligand in the ternary complex (5). (B) Typical experiment carried out with the wild-type proteins as detected by TIRFS (green) and by RIf (black) (2 fmol/mm² ifnar2-H10, 20 fmol/mm² ifnar1-H10). After the second injection of ^AF488IFNα2, 2 μM ifnar2-tl was injected to eliminate rebinding. After the third injection of ^AF488IFNα2, 1 μM unlabeled IFNα2 HEQ was injected. (C) Overlay of the normalized ^AF488IFNα2 dissociation curves from panel B: spontaneous dissociation during washing with buffer (black) and with 2 μM ifnar2-tl (red), as well as dissociation while chasing with IFNα2 HEQ (green). Dissociation from ifnar2-H10 alone is shown for comparison (blue). The residuals from the curve fits are shown in the bottom. (D) Same experiment as in panel B carried out with ifnar2-H10 I47A (2 fmol/mm² ifnar2-H10 I47A, 20 fmol/mm² ifnar1-H10). (E) Overlay of the dissociation curves from panel D (same color-coding as in panel C) with comparison of ^AF488IFNα2 dissociation from ifnar2-H10 I47A alone (blue). The residuals from the curve fits are shown in the bottom.

FIGURE 5

Determination of two-dimensional rate constants for pathway 1. (A) Schematic of the assay: ternary complex on fluid lipid membrane was formed by sequential injection of ifnar1-H10 (1), excess ifnar2-H10 I47A (2), and ^AF488IFNα2 HEQ (3). Upon loading the excess binding sites of ifnar2-H10 with unlabeled IFNα2 (4), labeled IFNα2 in the ternary complex was exchanged (5). (B) Course of a typical experiment as detected by TIRFS (green) and RIf (black). During spontaneous ligand dissociation, 2 μM ifnar2-tl was maintained in the background to eliminate rebinding. After the second injection of ^AF488IFNα2 HEQ 1 μM unlabeled IFNα2-wild-type was injected (2.5 fmol/mm² ifnar1-H10, 50 fmol/mm² ifnar2-H10 I47A). (C) Overlay of the normalized dissociation curves: spontaneous dissociation from the ternary complex (red) and ligand exchange kinetics washing with 1 μM IFNα2 HEQ (green). Dissociation from Ifnar1-H10 alone is shown for comparison (blue). The residuals from the curve fits are shown in the bottom.

FIGURE 6

Population of the dissociation pathways under different conditions. Ligand dissociation was numerically simulated based on the experimentally determined two- and three-dimensional rate constants for the following species of ifnar2/IFNα2/ifnar1: wt/wt/wt (A); I47A/wt/wt (B); and wt/R144A/wt (C). In all cases, 2 fmol/mm² of both ifnar2 and ifnar1 were assumed to form ternary complex under three different condition: no excess of either of the receptor subunits (top panel), with an excess of 20 fmol/mm² ifnar1 (middle panel), and with an excess of 20 fmol/mm² ifnar2 (bottom panel).