Background

Molecular binding in liquids is fundamental to many processes in chemistry, soft matter, and biology, including viral attachment, transcription factor binding, enzymatic reactions, self-assembly, and drug interactions. Traditionally, binding dynamics have been modeled as memoryless, meaning the probability of remaining bound decays exponentially with time and can be described by a dissociation rate constant. But this view neglects rebinding, even though many biological and soft-matter functions depend not on one strong binding event, but on repeated weak binding and rebinding. Because fast nanoscale interactions are hard to measure experimentally, it has remained unclear how often rebinding occurs and how much it shapes overall binding behavior. The paper argues that existing models cannot distinguish a single long binding event from many short rebindings, creating a need for a new framework.

Question

Whether liquid molecules exhibit a genuine binding memory effect, rather than purely memoryless stochastic binding, and if so, whether that memory follows general physical rules. How this memory depends on system dimensionality, binding affinity, viscosity, crowding, fractal geometry, and environmental heterogeneity. 

Main findings

Binding memory is universal in liquids and is often expressed as a power-law decay in the BAF rather than an exponential decay. In homogeneous systems, the scaling exponent depends mainly on dimensionality and diffusion properties, while factors such as viscosity, binding strength, crowding, and cutoff distance primarily affect the amplitude of the memory signal rather than its exponent. The study further shows that rebinding can make a major contribution to effective binding time, so ignoring it can strongly underestimate how long molecules remain functionally associated. In more complex heterogeneous environments, such as fractal spaces, chromatin-like binding landscapes, and phase-separated systems, additional emergent memory behaviors arise that cannot be fully captured by traditional diffusion-based theory. The authors also extend the concept to biological settings, including chromatin-associated motion and lipid bilayers, and experimentally detect strong binding memory on both artificial and live-cell membranes. Overall, the paper proposes binding memory as a new physical principle and as a possible probe of local material properties in complex biological systems.

Key methods used

The study combines three major approaches. First, it uses large-scale all-atom molecular dynamics simulations of hydrogen-bonding liquids such as water and ethanol to directly compute BAFs across different dimensions. Second, it uses Langevin dynamics simulations to systematically vary physical parameters like viscosity, binding strength, crowding, confinement, fractal geometry, and heterogeneous binding landscapes, and to test how these factors influence memory. Third, it develops a scaling theory linking BAF power-law exponents to dimensionality and anomalous diffusion. To validate the theory experimentally, the authors use real-time high-resolution single-particle tracking microscopy, specifically 3D-SMARTER, with microsecond temporal and nanometer spatial precision, to measure adsorption and membrane-associated binding dynamics in vitro and in live cells. This simulation-theory-experiment integration is a major strength of the paper.

Research article link: Binding memory of liquid molecules | Nature Communications