Mapping of the Glycolytic Enzyme Binding Sites on Human Erythrocyte Membrane Band 3.

Glycolytic enzymes (GEs) including aldolase, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), phosphofructokinase (PFK), pyruvate kinase (PK) and lactate dehydrogenase (LDH) are known to associate with the inner surface of the human erythrocyte membrane. Previous studies have suggested that the N-terminus of the cytoplasmic domain of band 3 (cdb3) constitutes the likely binding site for aldolase, GAPDH and PFK, however, the membrane docking sites for PK and LDH have not been elucidated. In this study, we demonstrate that PK and LDH exhibit no affinity for band 3, regardless of whether the association is measured by co-immunoprecipitation assay, binding competition studies, or catalytic inhibition analyses. We further find that the binding sites for GAPDH, aldolase and PFK on band 3 are distinct but partially overlapping, as evidenced by the fact that: 1) deletion of residues 1–11 of cdb3 eliminates the binding of aldolase, but not PFK or GAPDH, 2) fusion of thioredoxin (Trx) to the N-terminus of residues cdb3 blocks aldolase binding, but not the association of GAPDH or PFK, 3) deletion of sequences 1–50, 1–40, 1–31, or 1–23 of cdb3 blocks cdb3 association with all three GEs, whereas deletion of residues 12-23 only abrogates aldolase binding (while reducing the affinity of PFK and GAPDH), 4) the presence of both sequences, 6–DDYED-10 and 19-EEYED-23, are necessary for cdb3 association with aldolase, whereas the presence of either sequence alone (especially 19-EEYED-23) is sufficient to maintain association with GAPDH and PFK, 5) mutation of all of the acidic residues in the above two sequences to their corresponding amides (E→Q and D→N) results in loss of affinity for all GEs. Because i) kidney cdb3 (which lacks residues 1-65 of cdb3) shows no affinity for any of the GEs, ii) residues 1–55 of cdb3 show near normal affinity for aldolase, GAPDH and PFK, and iii) GAPDH, aldolase, and PFK all compete with each other for cdb3, we conclude that the binding sites for aldolase, GAPDH and PFK are all located within the first 23 residues of cdb3, with the docking site for aldolase likely residing somewhat more N-terminal than the binding sites for GAPDH and PFK. Finally, because each band 3 monomer contains three homologous sequences (6-DDYED-10, 19-EEYED-23 and 902-DEYDE-906) that are found in other proteins that bind GEs (e.g. actin, β-tubulin, troponin T), and since the first two of these sequences were shown to be required for cdb3 binding, the question naturally arose whether different GEs might also associate with residues 902-DEYDE-906 at the extreme C-terminus of band 3. Similar binding studies demonstrate that such an interaction does not occur. Thus, our data show that GAPDH, aldolase, and PFK all bind near the N-terminus of band 3 and that PK and LDH must dock somewhere else on the red cell membrane.