5.6 Conclusions

The mesoscopic model (that uses the NN model) with the UO energies qualitatively describes the FDC of DNA unzipping experiments. At a quantitative level, the UO energies overestimate the mean unzipping force, especially at low salt concentration.

In order to reach a quantitative agreement between the experiments and the mesoscopic model, it has to be completed with: 1) an accurate model for the ssDNA elasticity, 2) an empiric model to correct the experimental drift and 3) a specific contribution for the end loop.

The NNBP energies can be fit to match the mesoscopic model with the experimental FDC. A Monte Carlo fit of the data gives rapid and accurate estimations of the NNBP energies. The fit is robust and provides acceptable confidence intervals for the energies.

The extension of the experiments and the Monte Carlo fit to several molecules and salt concentrations provides valuable information about the salt dependence of the NNBP energies. A homogeneous salt correction does not describe the results found. Instead, a sequence-specific salt correction offers a more trustworthy rule.

The heterogeneous salt correction might be justified by two different explanations: 1) the differences in solvation between the specific nucleotides and monovalent ions at different salt concentrations or; 2) the differences in the elastic response of the bases of the ssDNA at different salt concentrations. Specific experiments should be carried out to discern between the two possibilities.

The unzipping experiments do not provide the enthalpies and the entropies of formation of the NNBP elements. However, the unzipping and melting experiments can be combined to obtain an estimation of these thermodynamic magnitudes.

Although melting and unzipping experiments are based on disruption processes triggered by different external agents (temperature and force respectively), the agreement between the thermodynamic magnitudes is remarkable. Both types of experiments are correctly described by the NN model.

The thermodynamic magnitudes obtained with the unzipping experiments not only predict the melting temperatures well but also improve the melting temperature prediction as compared to the UO prediction for oligos longer than 15 bp. The misprediction of melting temperatures for short oligos might be due to underestimated boundary effects.

The unzipping experiments provide an alternative determination of the NNBP parameters in which the folding/unfolding transition does not need to be two-state. Besides, instead of several short oligos of different sequence, one long molecule is sufficient to infer the NNBP energies.

The main limitations of the method exposed are the strong dependence of the results on the elastic response of the ssDNA and the impossibility to determine the bimolecular initiation factors.

The unzipping approach can be extended to extract free energies, entropies and enthalpies in DNA and RNA structures under different solvent and salt conditions. The method can be also applied to extract free energies of other structural motifs in DNAs (e.g., sequence dependent loops, bulges, mismatches, junctions). The enthalpies and the entropies can be directly extracted from unzipping experiments performed at different temperatures. The force methods make possible to extract free energies in conditions not accessible to bulk methods (such as the dsRNA that hydrolyzes in the presence of magnesium in melting experiments). Another extension of the method can be used to obtain the binding free energies of DNAs and RNAs bound to proteins, where the proteins denaturalize below the dissociation melting transition. Finally, this method could be also useful in cases where molecular aggregation and other collective effects in bulk preclude accurate free energy measurements.

This methodology can serve as a basis to search for long-range context effects (e.g., 2nd-nearest and 3rd-nearest interactions) in DNA. This might be especially interesting in those regions along the sequence where the NN model seems to fail.

Summing up, this chapter establishes a novel methodology to obtain thermodynamic information from single-molecule experiments.

JM Huguet 2014-02-12