Human erythrocyte net sugar transport is hypothesized to be rate-limited by reduced cytosolic diffusion of sugars and/or by reversible sugar association with intracellular macromolecules [Naftalin, R. J., Smith, P. M., & Roselaar, S. E. (1985) Biochim. Biophys. Acta 820, 235-249]. The present study examines these hypotheses. Protein-mediated 3-O-methylglucose uptake at 4 degrees C by human erythrocytes and by resealed, hypotonically lysed erythrocytes (ghosts) is inhibited by increasing solvent viscosity. Protein-mediated transport and transbilayer diffusion of the slowly transported substrate 6-NBD glucosamine are unaffected by increasing solvent viscosity. These findings suggest that protein-mediated 3-O-methylglucose transport is diffusion-limited in erythrocytes. More detailed analyses of red cell 3-O-methylglucose uptake (at 4 degrees C and at limiting extracellular sugar levels) reveal that net influx is a biexponential process characterized by rapid filling of a small compartment (C-1 = 29 +/- 6% total cell volume; k(1) = 7.4 +/- 1.7 min(-1)) and slow filling of a larger compartment (C-2 = 71 +/- 6% total cell volume; k(2) = 0.56 +/- 0.11 min(-1)). Erythrocyte D-glucose net uptake at 4 degrees C is also a biphasic process. Transmembrane sugar leakage is a monoexponential process indicating that multicomponent, protein-mediated uptake does not result from sugar uptake by two cell populations of differing cellular volume. Sugar exit at limiting 3-O-methylglucose concentrations is described by single exponential kinetics. This demonstrates that multicomponent sugar uptake does not result from influx into two populations of cells with widely different sugar transporter content. We conclude that biexponential sugar uptake results from slow (relative to transport) exchange of sugars between serial, intracellular sugar compartments. Biexponential sugar uptake is observed under equilibrium exchange conditions (intracellular sugar concentration = extracellular sugar concentration) but only at 3-O-methylglucose concentrations of less than 1 mM. Above this sugar concentration, exchange uptake is a monoexponential process. Because diffusion rates are independent of diffusant concentration, this suggests that multicomponent uptake results from high-affinity sugar binding within the cell. The concentration of cytosolic binding sites (30 mu M, K-d(app) = 400 mu M) was estimated from the equilibrium cellular 3-O-methylglucose space. Biexponential net 3-O-methylglucose uptake is also observed in human erythrocyte ghosts, in control human K562 cells, and in K562 cells induced to synthesize hemoglobin by prolonged exposure to hemin. This demonstrates that neither membrane-bound nor free cytosolic hemoglobin forms the sugar-binding complex. alpha-Toxin-permeabilized cells fill rapidly (within 5 s) with 3-O-methylglucose and L-glucose (a nontransported sugar), indicating that the glucose binding compartment does not extend across the entire intracellular margin of the plasma membrane. Rather, it must be restricted to domains of locally high-glucose transporter density. Immunofluorescence microscopy of erythrocytes indicates that GLUT1 is not distributed uniformly across the cell surface, while the anion transporter shows a uniform cell surface distribution. Red cell hexokinase land GLUT1 appear not to colocalize in hypotonically lysed erythrocytes. The kinetics of sugar uptake and exit are quantitatively mimicked by a model in which newly imported sugars enter the bulk intracellular water only following interaction with an intracellular, sugar-binding complex. We conclude that steady state sugar transport assays in human erythrocytes measure two processes: rapid sugar translocation across the bilayer and slow sugar release into bulk cytosol. The conclusions of previous steady state analyses which assume net transport reflects only sugar translocation may require reconsideration.
