Understanding How Transient Interactions Shape the Affinity of NF-κB Transcription Factor for DNA via Molecular Dynamics Simulations and Machine Learning

Nuclear factor kappa B (NF-κB) transcription factors play a pivotal role in gene regulation by binding to κB DNA elements with a variable central region surrounded by conserved flanking sequences. The RelA homodimer, a widely expressed NF-κB TF, shows differing affinities based on the central DNA base pair, with A or T conferring tenfold higher affinity compared to G or C. However, crystal structures offer limited insights into the underlying mechanisms for these affinity differences. Our study utilized extensive molecular dynamics (MD) simulations and two machine learning models to identify two key residues, R187 and R124, in RelA and elucidate their distinct dynamic behaviors. R187 influences base contacts by interacting with major groove-bound residues through differential transient states, while R124 enhances interactions by favoring transient insertion into the minor groove in A-κB sequences over G-κB. These residues exhibit selective interactions that orchestrate a complex interplay among DNA-interacting residues, validated by mutagenesis studies. Our discoveries shed light on the intricate mechanisms governing NF-κB's selective binding to κB DNA, underscoring the importance of MD simulations in unraveling the dynamic aspects of TF-DNA interactions.