The static and dynamic structural heterogeneities of B-DNA: extending Calladine-Dickerson rules

Abstract

We present a multi-laboratory effort to describe the structural and dynamical properties of duplex B-DNA under physiological conditions. By processing a large amount of atomistic molecular dynamics simulations, we determine the sequence-dependent structural properties of DNA as expressed in the equilibrium distribution of its stochastic dynamics. Our analysis includes a study of first and second moments of the equilibrium distribution, which can be accurately captured by a harmonic model, but with nonlocal sequence-dependence. We characterize the sequence-dependent choreography of backbone and base movements modulating the non-Gaussian or anharmonic effects manifested in the higher moments of the dynamics of the duplex when sampling the equilibrium distribution. Contrary to prior assumptions, such anharmonic deformations are not rare in DNA and can play a significant role in determining DNA conformation within complexes. Polymorphisms in helical geometries are particularly prevalent for certain tetranucleotide sequence contexts and are always coupled to a complex network of coordinated changes in the backbone. The analysis of our simulations, which contain instances of all tetranucleotide sequences, allow us to extend Calladine-Dickerson rules used for decades to interpret the average geometry of DNA, leading to a set of rules with quantitative predictive power that encompass nonlocal sequence-dependence and anharmonic fluctuations.

Figure 1.

Symmetric Kullback–Leibler divergence per degree of freedom between Gaussian distributions, which is a combined measure of differences in values of first and second moments, for each of the thirteen oligomers in the miniABC training library, but for cgDNA model parameter sets fitted to different MD simulation protocols (see Materials and Methods and Supplementary Methods).

Figure 2.

Average values of intra-basepair helical coordinates of the central base-pair (x-axis) in all possible 64 trinucleotide sequence contexts (y-axis). Results obtained from the miniABC_BSC1-K simulations. The global averages (white) are over all sequence contexts and standard deviations reflect the variation among trinucleotide contexts. The blue squares mean that a specific base-pair has an average value above the global average plus one standard deviation, while red squares mean an average value below the global average minus one standard deviation.

Figure 3.

Average values of inter-basepair, or step, helical coordinates for the central junction (x-axis) in all possible 256 tetranucleotide contexts (y-axis). Results obtained from the miniABC_BSC1-K simulations. Tetranucleotides classified as bimodal (half-square) are polymorphic (i.e. they sample two clear conformational substates). The global averages (white), exhibited on the legend at the right of each squared-plot, were computed from the weighted-averages obtained through BIC (see Materials and Methods and Supplementary Methods), and while the standard deviations reflect the variation among the tetranucleotide sequences that share the same central basepair step. The blue squares mean that a specific step has an average value above the global average plus one standard deviation, while red squares mean an average value below the global average minus one standard deviation.

Figure 4.

Sequence dependence of BII backbone conformations comparing K⁺ and Na⁺. (A) miniABC_BSC1-K BII percentages. (B) miniABC_BSC1-Na BII percentages. (C) Correlation between the percentage of BII (%BII, horizontal axis) and the occurrence of formation of the C–H···O H-bonds (%HB, vertical axis) at the central base step of each of the 256 possible tetranucleotide sequences, colour-coded according to the base type of the central base step.

Figure 5.

Schema of the polymorphic, or multi-well, landscape exhibited by B-DNA at the tetranucleotide level expressed in the purine (R)/pyrimidine (Y) alphabet, for which only 10 distinct combinations exist, but which still distinguish all possible behaviors. The only helical coordinates that exhibited multi-modality are shift, slide and twist, and each junction in the figure is marked with which coordinates can be multi-modal in it. There is a very high correlation between the occurrence of multi-modality and the formation of a noncanonical hydrogen bond in either the same or a neighboring junction, along with its associated BI/BII backbone transition (see text).

Figure 6.

Comparison between experimental structures (X-ray and NMR) determined for the Drew−Dickerson Dodecamer (9), and conformations predicted by using the MD datasets produced herein. The six intra-basepair parameters were predicted (red lines) and compared with all the experimental structures (gray lines) and Calladine–Dickerson prediction for Roll and Twist (cyan line). Vertical dashed lines represent predicted binormal/bimodal steps, while vertical dotted lines represent steps with clear multi-peaked distributions although not bimodal according to Helguerro (see Materials and Methods). In the last row, predicted BI% (yellow and green) were compared with ³¹P-NMR gold-standard measurements and multi-microsecond long MD simulation of the same sequence using PARMBSC1 force field (9). NMR1 stands for the work by Schwieters et al. (74), and NMR2 from Tian et al. (68). 3D superposition showing the capability of the miniABC library and our set of rules to predict two conformational substates observed experimentally for DDD at the highly polymorphic GCGA tetranucleotide (NMR structures with PDB id: 171D, and 1GIP). The RMSD was measured after aligning the backbones between the experimental structures and the two substates captured by the miniABC library.