Intermediate-Resolution Modeling of Dynamic DNAs and Their Phase Separation

Abstract

DNA is a fundamental biomolecule in eukaryotic cells, playing central roles in processes ranging from genome organization and transcription to innate immune signaling. Recent studies have revealed that DNA can undergo protein-free phase separation in the presence of divalent cations, yet the underlying molecular mechanisms, including the interplay of base stacking, base pairing, electrostatics, and ion interactions, remain poorly understood. Here, we introduce an intermediate-resolution model for condensates of DNAs (iConDNA) that can capture key local and long-range structural features of dynamic DNAs and simulate their spontaneous phase transitions. By introducing explicit base stacking and pairing interactions, the iConDNA model not only reproduces major conformational properties of DNA homopolymers but also folds DNA hairpins and duplexes and captures their thermodynamic properties. With an effective model of explicit Mg2+, iConDNA successfully captures the temperature and magnesium concentration dependence of DNA properties. Together, these features enable iConDNA to qualitatively recapitulate homotypic DNA phase separation, providing a suitable tool to study DNA homotypic phase separation in biological and engineering applications.