The crystal structure of the activin B:Fst288 complex and computational insights into the broad antagonistic activity and specificity of follistatin

Abstract

Members of the TGF-{beta} family regulate diverse biological processes, and their activity is tightly controlled by extracellular antagonists such as follistatin 288 (Fst288). Fst288 potently inhibits activins A and B, GDF8, and GDF11. Among these, ActB is the least structurally and functionally characterized, which limits a full understanding of Fst288s broad activity and its specificity for this TGF-{beta} subgroup. To address this, we solved the crystal structure of the ActB:Fst288 complex. As in other complexes, two Fst288 molecules surround ActB, occluding both receptor-binding sites. However, ActB engages Fst288 differently: its fingers bind most strongly the ND domain, while its acidic fingertips uniquely contact FSD3 disrupting interactions between the two Fst288 molecules and reducing the cooperativity seen in other complexes. Computational analysis revealed that, although ActB exhibits higher interaction enthalpy, entropic penalty lowers its overall affinity compared to ActA, consistent with experimentally measured Kd values. We therefore propose that broad ligand inhibition by Fst288 arises from variations in binding interactions and the utilization of cooperativity when direct contacts are insufficient for high affinity. In cellular contexts, antagonistic effectiveness is further impacted by interactions between complexes and the extracellular matrix, resulting in comparable in vitro IC50 values. Finally, sequence comparisons with non-binding TGF-{beta}s indicate that Fst288 specificity originates from ionic contacts with acidic ligand fingertips which likely initiate recognition, while interactions at the type I and II interfaces stabilize the complex. This study provides deeper insight into Fst288s regulation of activin signaling and paves the way for designing inhibitors with desired selectivity.