LAUSR

Linkers are ubiquitous in multidomain proteins. These linkers are integral to protein functions, and accumulating evidence suggests that the linkers' versatile roles are encoded in their sequences. However, a molecular picture of how amino acid differences in the linker influence protein function is still lacking. By using extensive Gaussian-accelerated MD coupled with dynamic network analysis, we reveal the molecular bases underlying the linker's role in Calmodulin (CaM), a highly conserved Ca2+-signaling hub in eukaryotes. Three CaM constructs comprising a wild-type linker, a flexible linker (four glycines at position D78-S81), and a rigid linker (four prolines at position D78-S81) were simulated. We show that the flexible linker resembles the wild type in allowing CaM to sample a large ensemble of conformations while the rigid linker confines the sampling. Our simulations recapture experimental observations that target binding enhances the Ca2+ affinity to CaM's EF-hand sites at the N-domain. However, only the wild-type linker can both correctly capture the Ca2+ binding order and maintain the α-helical structure of the domain. The other two constructs either bind Ca2+ in an incorrect order or exhibit unfolding of an N-domain helix. We demonstrate that the wild-type linker achieves these outcomes by transmitting interdomain dynamics efficiently. This was evidenced by stronger (anti)correlations among the linker residues, decoupling of the hydrogen bonds between A1-A15 and V35-E45, and structuring of the N-domain for Ca2+ binding. This decoupling was not evident for the other two constructs. Lastly, we show that the wild-type linker's optimal transmission stems from its thermodynamically favorable strain and solvation relative to the other two constructs. Our results show how the linker sequence tunes CaM function, suggesting possible mechanisms for changes in linker properties such as mutations or post-translational modifications to modulate protein/substrate binding.