Protamine, an arginine-rich disordered protein, replaces histones to compact DNA in the sperm nucleus. Whereas folded domains can bind DNA both nonspecifically and in a site-specific manner, whether disordered proteins exhibit site specificity in DNA binding is an open question. Here, we used all-atom molecular dynamics (MD) simulations to characterize protamine binding to double-stranded DNA. In conventional MD simulations, 8 copies of protamine (33 residues with 21 arginines) quickly wrap around a 35-bp DNA. In addition to salt bridges with DNA phosphate groups, Arg forms hydrogen bonds with nucleobases, in the form of either wedges (involving bases from both strands) or clamps (one or more bases on the same strand). Base interactions favor GC over AT and the major groove over minor groove. At the major groove, Arg strongly prefers guanine, as both its O6 and N7 atoms can act as hydrogen bond acceptors based on chemical and geometrical considerations. At the minor groove, Arg shows a modest preference for AT over GC. Free-energy simulations confirm these observations. For experimental validation, we designed DNA sequences spanning 0% to 40% GC contents. Protamine-DNA mixtures form amorphous aggregates, and the aggregation propensity increases with increasing GC content. This work provides a mechanistic framework for understanding how DNA sequence composition, in particular GC content, regulates DNA condensation by protamine.
Dhiman et al. (Sun,) studied this question.