Diffusion of glycophorin A in human erythrocytes

Abstract

Several lines of evidence suggest that glycophorin A (GPA) interacts with band 3 in human erythrocyte membranes including: i) the existence of an epitope shared between band 3 and GPA in the Wright b blood group antigen, ii) the fact that antibodies to GPA inhibit the diffusion of band 3, iii) the observation that expression of GPA facilitates trafficking of band 3 from the endoplasmic reticulum to the plasma membrane, and iv) the observation that GPA is diminished in band 3 null erythrocytes. Surprisingly, there is also evidence that GPA does not interact with band 3, including data showing that: i) band 3 diffusion increases upon erythrocyte deoxygenation whereas GPA diffusion does not, ii) band 3 diffusion is greatly restricted in erythrocytes containing the Southeast Asian Ovalocytosis mutation whereas GPA diffusion is not, and iii) most anti-GPA or anti-band 3 antibodies do not co-immunoprecipitate both proteins. To try to resolve these apparently conflicting observations, we have selectively labeled band 3 and GPA with fluorescent quantum dots in intact erythrocytes and followed their diffusion by single particle tracking. We report here that band 3 and GPA display somewhat similar macroscopic and microscopic diffusion coefficients in unmodified cells, however perturbations of band 3 diffusion do not cause perturbations of GPA diffusion. Taken together the collective data to date suggest that while weak interactions between GPA and band 3 undoubtedly exist, GPA and band 3 must have separate interactions in the membrane that control their lateral mobility.

Keywords: Band 3; Erythrocyte membrane structure; Glycophorin A; Single particle tracking; V(H)H.

Fig. 1. Recombinant biotinylated camel V_HH fragment IH4 specifically labels glycophorin A

(A) Immunoblot analysis of protein bands from whole human RBCs labeled with IH4 (left) or IH4-biotin (right), and then detected with murine anti-HA followed by anti-mouse-HRP (left) or anti-biotin-HRP (right). (B) Image of RBCs labeled first with IH4-biotin (upper panel) or IH4 (lower panel) and then detected with anti-biotin-FITC. (C) Titration of the antibody to estimate a concentration appropriate for single molecule tracking. Erythrocytes were incubated with the indicated concentrations of IH4-biotin, followed by anti-biotin-FITC and viewed under a confocal microscope.

Fig. 2. Recombinant monovalent camel antibody IH4 and IH4-biotin do not alter RBC deformability

(A) Representative deformability curves are shown for fresh RBCs treated with no antibody (solid), unlabeled IH4 (dashed), or the monoclonal anti-GPA antibody (R-10) known to increase erythrocyte rigidity (dotted). EI = elongation index (B) Comparison of the deformability of cells treated with or without IH4-biotin. To assure that IH4-biotin was indeed binding to the treated RBCs, binding was confirmed by agglutination with anti-biotin antibody after each sample was run.

Fig. 3. Recombinant camel IH4 antibody does not perturb band 3 mobility

Distributions of microscopic diffusion coefficients, macroscopic diffusion coefficients, and compartment sizes were measured in (A) untreated control cells, and (B) cells treated with IH4 at a concentration (4 μM) that is 4000-fold higher than the concentration used to measure glycophorin A diffusion.

Fig. 4. Comparison of glycophorin A and band 3 diffusion in unmodified (control) human erythrocytes

(A) GPA diffusion data collected on samples from multiple donors, and (B) band 3 diffusion data as previously published by Kodipilli et al. (21).

Fig. 5. Glycophorin A mobility is independent of the interaction of band 3 with the cytoskeleton

Distribution of GPA diffusion coefficients obtained from (A) control and (B) OV-treated red blood cells. Cells were incubated with 2 mM OV for 30-60 min, and GPA diffusion was monitored by single particle tracking. The distributions of GPA diffusion coefficients on untreated cells were measured the same day and on the same blood samples used for the controls.