A preprint modeling study reports two different simulated patterns in the stiffness of short single-stranded DNA homopolymers under polyvalent ions. For two 30-nucleotide strands, fitted persistence length—a measure of chain stiffness—decreased as magnesium or calcium concentration increased. In the spermine simulations, it did not change beyond 0.5 millimolar across the tested range of 0.1 to 20 millimolar.
The spermine result is a model-based prediction from a preprint. The authors say no clear theory explains it yet and that the predicted concentration dependence needs experimental testing.
A narrow comparison of two sequences
The modeled strands were dT30, made entirely from thymine, and dA30, made entirely from adenine. Each was examined with magnesium, calcium or spermine, with 20 millimolar sodium chloride present in the stated simulations. The comparison therefore covered two short, uniform polymers under specified ion conditions rather than mixed DNA sequences.
The researchers used an explicit-ion, coarse-grained nucleic-acid model called TIS. They estimated persistence length by fitting a mathematical distribution for a semiflexible chain to the simulated distribution of end-to-end distances, the span from one end of a strand to the other.
The fitted bare persistence length was about 0.4 nanometres for dT30 and 1.1 nanometres for dA30. The analysis described dT30 as random-coil-like and dA30 as more helical.
Magnesium and calcium followed a common pattern
For both magnesium and calcium, simulated persistence length decreased monotonically as ion concentration increased. The fitted values also showed a linear dependence on κ−1, the model’s term for electrostatic screening used in the analysis.
The study summarizes the divalent-ion pattern as lp ∼ κ−1 ∼ c−0.5, with a reported concentration exponent of −0.5. It notes that this scaling agrees with flexible-polyelectrolyte theory.
The charge analysis showed its largest change at lower magnesium concentrations. In 20 millimolar sodium chloride without added magnesium, the renormalized phosphate charge was close to −0.9. At magnesium concentrations up to 1 millimolar, the analysis reported about a 25% charge reduction; up to 20 millimolar, it reported about 13% further reduction, with the change leveling off near 50%.
At the highest concentrations considered for dT30, the charge accounting attributed about 47% of charge neutralization to magnesium and about 1% to sodium. The calculation counted a counterion as condensed using a model-defined distance increment of 0.15 nanometres.
The spermine pattern leveled off
In the simulations, persistence length did not change beyond 0.5 millimolar spermine in either sequence across the 0.1-to-20-millimolar range. It was larger for dA30 than for dT30 at every tested concentration.
The renormalized charge also reached a plateau near 0.5 millimolar spermine and remained lower than under magnesium or calcium. The paper describes this as weaker charge screening by spermine than by the divalent ions.
A separate count showed another plateau: the number of spermine molecules in direct contact with phosphate groups stayed nearly constant at about three to six across the full 0.1-to-20-millimolar range. The authors offer a qualitative interpretation in which spermine’s anisotropic structure and volume exclusion may limit access to neighboring phosphate sites and restrict simultaneous binding.
That explanation remains qualitative. The authors acknowledge that no clear theory explains the spermine result and present the concentration dependence of persistence length as a prediction awaiting experimental testing.
What the simulations leave open
The evidence comes from two modeled homopolymers, dT30 and dA30, under specified ionic and simulation conditions. Mixed sequences, other strand lengths and other structures were not tested, so the reported pattern cannot yet be assumed to apply to longer or biologically complex single-stranded DNA.
The supplied analysis does not report the number of independent trajectories or replicates, sampling convergence, confidence intervals, standard errors or error bars. Exact concentration-specific persistence-length values and complete fit uncertainties are also not provided in the extracted text.
The coarse-grained calculations are model-based evidence rather than independent experimental validation. The study does not experimentally establish spermine’s concentration dependence or the proposed steric-saturation explanation, and it provides no evidence for causal effects in humans or other organisms.
The analysis leaves open experimental measurements of persistence length against spermine concentration, tests of whether steric saturation is general across polyamine–single-stranded-DNA interactions, and development of a theory incorporating spermine’s anisotropic structure, volume exclusion and binding entropy.
A prediction still awaiting a laboratory test
The document is arXiv:2608.25046v1, dated 25 August 2026, and is identified in the supplied metadata as a preprint with no journal or DOI listed. The acknowledgments name NSF grant CHE 2320256 and Welch Foundation grant F-0019 as research support.
The paper states that supporting data are available from the corresponding author upon reasonable request.
Paper data and sources
Original title: Persistence length of short homopolymeric single-stranded DNA sequences in polyvalent cations
Authors: Balaka Mondal, D. Thirumalai
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-25
DOI: Not available
Original paper · Full text