This report describes initial characterization of the binding sites of ankyrin for spectrin and the anion exchanger using defined subfragments isolated from purified ankyrin domains. The spectrin-binding domain of ankyrin is comprised of two subdomains: an acidic, proline-rich region (pI = 4) involving the amino-terminal 80 residues from 828 to 908 and a basic region (pI = 8.8) that extends from 898 to 1386. The amino-terminal 70 amino acids of the spectrin-binding domain are critical for association with spectrin, since a subfragment missing this region is only 5% as active as the intact domain in displacing binding of spectrin to inside-out membrane vesicles, while deletion of the first 38 residues of the acidic domain results in a 10-fold reduction in activity. The anion exchanger-binding site is confined to an 89-kDa domain that was isolated and characterized as a globular molecule with approximately 30% alpha-helical configuration. A subfragment of the 89-kDa domain extending from residues 403 to 779 (or possibly 740) retains ability to associate with the anion exchanger. The 89-kDa domain is comprised of a series of tandem repeats of 33 amino acids that extend from residues 35 to 778 (Lux, S., John, K., and Bennett, V. (1990) Nature 344, 36-42). The activity of residues 403-779 demonstrates that the 33-amino acid repeats of the 89-kDa domain are responsible for association between ankyrin and the anion exchanger. The 33-amino acid repeating sequence of ankyrin represents an ancient motif also found in proteins of Drosophila, yeast, and Caenor habditis elegans. The finding that the 33-amino acid repeating sequence is involved in interaction with the anion exchanger implies that this motif may perform a role in molecular recognition in diverse proteins.