Measuring the energetics of membrane protein dimerization in mammalian membranes

Abstract

Thus far, quantitative studies of lateral protein interactions in membranes have been restricted peptides or simplified protein constructs in lipid vesicles or bacterial membranes. Here we show how free energies of membrane protein dimerization can be measured in mammalian plasma membrane-derived vesicles. The measurements, performed in single vesicles, utilize the quantitative imaging FRET (QI-FRET) method. The experiments are described in a step-by-step protocol. The protein characterized is the transmembrane domain of glycophorin A, the most extensively studied membrane protein, known to form homodimers in hydrophobic environments. The results suggest that molecular crowding in cellular membranes has a dramatic effect on the strength of membrane protein interactions.

Figure 1

Overview of measurements of GpA dimerization energy in plasma membrane-derived vesicles.

Figure 2

Flowchart describing the steps involved in the QI-FRET method. The parameters determined in Phase I are used to calculate the free energy of GpA dimerization in Phase II.

Figure 3

One vesicle loaded with GpA_EYFP and GpA_mCherry. Intensities per unit membrane area were obtained by integrating the Gaussian intensity profiles across the membrane.

Figure 4

(A) FRET data and proximity contribution for GpA in CHO plasma membrane-derived vesicles. Each data point represents a single vesicle, for which E, C_A and C_D are determined using the QI-FRET method. Data analysis suggests that the scatter is largely due to random noise in image acquisition (see text). (B) Calculated dimeric fraction times Ẽ. Ẽ is the FRET efficiency in a GpA dimer with a donor and an acceptor. (C) Fit of the data to the dimerization model while varying the two unknowns, Ẽ and the dimerization constant K_D.

Figure 5

Analysis of uncertainties in calculated dimeric fractions due to uncertainties in image acquisition. 3 independent images of 10 vesicles were taken while re-focusing for each image. (A) shows the averages of the calculated membrane intensities $I_D^m$, $I_{\mathit{\text{FRET}}}^m$ and $I_A^m$for each vesicle, along with the standard deviations. (B) shows the averages of the dimeric fractions, calculated independently from each image set, along with their standard deviations.

Figure 6

Differences in calculated dimeric fractions, when using two methods to determine the membrane intensities $I_D^m$, $I_{\mathit{\text{FRET}}}^m$ and $I_A^m$for each vesicle analyzed in Figure 4, with and without filtering the random noise in the fluorescence images. The calculated differences due to noise are similar to the observed differences between vesicles in Figure 4, suggesting that random noise in the fluorescence images leads to the observed scatter in Figure 4.