Decoding biomolecular condensate dynamics: an energy landscape approach

Many eukaryotic proteins and RNAs contain low-complexity domains (LCDs) with a strong propensity for binding and driving phase separation into biomolecular condensates. Mutations in LCDs frequently disrupt condensate dynamics, resulting in pathological transitions to solid-like states. Understanding how the molecular sequence grammar of LCDs governs condensate dynamics is essential for uncovering their biological functions and the evolutionary forces that shape these sequences. To this end, we present an energy landscape framework that operates on a continuous 'stickiness' energy scale rather than relying on an explicit alphabet-based sequence. Sequences are characterized by Wasserstein distance relative to thoroughly shuffled or random counterparts. Armed with an energy landscape framework, map diagrams of material and dynamical properties governed by key energy landscape features modulated by the degree of complexity in LCD arrangements, including the periodicity and local disorder in LCDs. Highly periodic LCD patterns promote elasticity-dominated behavior, while random sequences exhibit viscosity-dominated properties. Our results reveal that minimum sticker periodicity is crucial for maintaining fluidity in condensates, thereby avoiding transitions to glassy or solid-like states. Moreover, we demonstrate that the energy landscape framework explains the recent experimental findings on prion domains and predicts systematic alterations in condensate viscoelasticity. Our work provides a unifying perspective on the sequence-encoded material properties whereby key features of energy landscapes are conserved while sequences are variable.