• Small Ankyrin-1 and Obscurin as a Model for Ankyrin Binding Motifs

      Willis, Chris; Bloch, Robert J. (2011)
      Small ankyrin-1 binds with high affinity to two regions within obscurin A. This interaction provides a molecular link between the sarcoplasmic reticulum and myofibrils in striated muscle. Here, I show that four hydrophobic residues within the hydrophobic "hotspot" of the ankyrin-like repeats of sAnk1 are involved in binding obscurin. Alanine scanning mutagenesis of each of the four residues inhibits binding to the high affinity binding site, Obsc6316-6345, whereas two of the mutations had no effect on binding to the lower affinity site, Obsc6231-6260. Mutagenesis identified three central residues within Obsc6316-6345 that are critical for binding sAnk1, but no single residue within Obsc6231-6260 that is essential. Instead, only a triple mutant of neighboring residues of Obsc6231-6260 decreases binding. This suggests the positional importance of the hydrophobic residues within the central region of Obsc6316-6345 that have a direct role in docking to sAnk1, in contrast to a more general, hydrophobic contribution of the hydrophobic residues of Obsc6231-6260. Modeling of sAnk1 and Obsc6316-6345 are consistent with the idea that specific hydrophobic residues interact in the docked complex. Using the identified hydrophobic and previously reported charged amino acids of obscurin involved in binding, I identify two classes of ankyrin binding motifs (ABMs) The first, type 1, contains ~35% alpha-helix, between two less structured regions. The second, type 2, contains~17% alpha-helix at its amino-terminal, flanked by a disordered carboxy-terminal sequence. Using a custom matrix, I classify several known ABMs as type 1 or 2 by structural comparison and sequence alignment within the ABM-specific matrix. I use this matrix to predict ABMs in proteins that bind ankyrins. This method identified a type 1 ABM in histone deacetylase-4 and in retinoblastoma protein that bind their ligands, RFXAnk and gankyrin, respectively. I confirm this binding with in vitro assays and the ~30% helical content of retinoblastoma by circular dichroism. The method also identified a type 2 ABM in the retinoblastoma protein that binds gankyrin in vitro and has ~15% helical content. I conclude that my methods can be used to predict ABMs from primary sequence which will be applicable to the prediction of other protein-protein interactions.