Enumeration of oligomerization states of membrane proteins in living cells by homo-FRET spectroscopy and microscopy: theory and application

Abstract

Protein-protein interactions play a pivotal role in biological signaling networks. It is highly desirable to perform experiments that can directly assess the oligomerization state and degree of oligomerization of biological macromolecules in their native environment. Homo-FRET depends on the inverse sixth power of separation between interacting like fluorophores on the nanometer scale and is therefore sensitive to protein oligomerization. Homo-FRET is normally detected by steady-state or time-resolved fluorescence anisotropy measurements. Here we show by theory and simulation that an examination of the extent of homotransfer as measured by steady-state fluorescence anisotropy as a function of fluorophore labeling (or photodepletion) gives valuable information on the oligomerization state of self-associating proteins. We examine random distributions of monomers, dilute solutions of oligomers, and concentrated solutions of oligomers. The theory is applied to literature data on band 3 protein dimers in membranes, GPI-linked protein trimers in "rafts," and clustered GFP-tagged epidermal growth factor receptors in cell membranes to illustrate the general utility and applicability of our analytical approach.

FIGURE 1

Simulations of the anisotropy as a function of labeling (f) for discrete oligomers (containing N-subunits) under the situation that energy migration depolarizes the fluorescence to an extent that the anisotropy is decreased to r_m/N, where N is the number of labeled monomers per oligomer. N-values in order from top to bottom: N = 1, 2, 3, … , 8. Simulations generated using Eq. 6 in text, with r₁ = 0.4.

FIGURE 2

Comparison of anisotropy as a function of fluorophore labeling for homogeneous oligomers with N subunits (dotted lines) and a Poisson distribution of oligomers with mean oligomerization number of N (solid lines). N-values in order from top to bottom: N = 1, 2, 3, and 4. Simulations generated with Eqs. 6 and 7 in text using the model in Fig. 1.

FIGURE 3

The effect of monomer on the anisotropy enhancement curves. Monomer-tetramer equilibrium with increasing fraction of fluorescence caused by monomer: 75% monomer (circles), 50% monomer (squares), 25% monomer (triangles). Simulations generated with Eq. 8 from text.

FIGURE 4

The effect of 2D concentration on the anisotropy versus label plot. Monomers at 1 million molecules on the surface of a cell with surface area 1000 μm² (β = 0.075) (circles). Dimers at a concentration of 1 million dimers on a cell with surface area 1000 μm² (with r_o = 0.4, r_∞ = 0.2, w = 1 ns, τ = 4 ns) in the presence (β = 0.075) (triangles) and absence (β = 0) (squares) of concentration depolarization.

FIGURE 5

Three-dimensional concentration depolarization of dimers. Plot of the gradient for the r versus f curve at different f values when β = 0 (squares), 0.001 (circles), 0.01 (apex-up triangles), and 0.1 (apex-down triangles), and r_o = 0.4, r_∞ = 0.2, w = 1 ns, and τ = 4 ns.

FIGURE 6

Plot of the average steady-state anisotropy as a function of the degree of labeling for the eosin-labeled band 3 in purified dimers (apex-up triangles) and in cells/membranes (apex-down triangles) (data from Fig. 1 of Blackman et al. (16)). Simulations of the corresponding anisotropy plots are denoted by the solid lines. The dimer simulation was generated using the parameters obtained from dynamic depolarization measurements of eosin-labeled band 3 in cross-linked dimers (with r_o = 0.37, r_∞ = 0.31, w = 1 ns, τ = 3 ns) and assuming no concentration depolarization between dimers. The membrane ghost simulations used identical parameters to the dimers except β = 0.06.

FIGURE 7

Plot of the average anisotropy as a function of the degree of photobleaching of GPI-linked proteins (data from Fig. 3 of Yeow et al. (13)). Lines indicate fits to dimer model (dotted lines) and trimer model (solid lines).