Advanced functional materials with stimuli-responsive and/or adaptive behaviors are highly promising for several applications (e.g., sensing, piezoelectronics, mechanical actuation, filtration). Although dynamic behavior has been demonstrated in both two-dimensional (2D) and 3D frameworks, the successful prediction and bottom-up design of these properties remains a formidable challenge. This is because dynamic behavior necessitates a free-energy landscape which is traversable in a prescribed manner, permitting the population of various functional states. In Nature, the thermodynamics of proteins and biomolecular assemblies have been evolutionarily optimized to carry out their functions, and thus the design of new materials which rival these systems requires a detailed understanding of how to appropriately balance thermodynamic factors to achieve the desired conformational flexibility. Here we report the calculated free-energy landscape of a previously characterized dynamic 2D protein crystal, rationalize its profile in terms of the constituent enthalpic and entropic components, and harness this information to rationally perturb the landscape, giving rise to a predictable modulation of the crystal dynamics which is confirmed experimentally. By careful selection of the identity and placement of these mutations, we demonstrate that this modification simultaneously affords a novel degree of control over the conformational state of the lattice via reversible metal coordination interactions.