Knowledge-based prediction of DNA hydration using hydrated dinucleotides as building blocks

Abstract

Water plays an important role in stabilizing the structure of DNA and mediating its interactions. Here, the hydration of DNA was analyzed in terms of dinucleotide fragments from an ensemble of 2727 nonredundant DNA chains containing 41 853 dinucleotides and 316 265 associated first-shell water molecules. The dinucleotides were classified into categories based on their 16 sequences and the previously determined structural classes known as nucleotide conformers (NtCs). The construction of hydrated dinucleotide building blocks allowed dinucleotide hydration to be calculated as the probability of water density distributions. Peaks in the water densities, known as hydration sites (HSs), uncovered the interplay between base and sugar-phosphate hydration in the context of sequence and structure. To demonstrate the predictive power of hydrated DNA building blocks, they were then used to predict hydration in an independent set of crystal and NMR structures. In ten tested crystal structures, the positions of predicted HSs and experimental waters were in good agreement (more than 40% were within 0.5 Å) and correctly reproduced the known features of DNA hydration, for example the `spine of hydration' in B-DNA. Therefore, it is proposed that hydrated building blocks can be used to predict DNA hydration in structures solved by NMR and cryo-EM, thus providing a guide to the interpretation of experimental data and computer models. The data for the hydrated building blocks and the predictions are available for browsing and visualization at the website https://watlas.datmos.org/watna/.

Keywords: DNA hydration; WatNA; dinucleotide fragments; knowledge-based prediction; water.

Figure 1

The procedure for determining DNA hydrated building blocks and their application for hydration prediction. (a) Selected structures from the analyzed data set of crystal structures from the PDB, containing 2727 nonredundant DNA chains (gray cartoon) together with the reported crystal water molecules (red balls). (b) Top: example of an extracted dinucleotide with water molecules closer than d _assoc = 4.0 Å. Middle: hydrated building block defined as the reference dinucleotide with the water distribution obtained by transferring all water molecules associated with the dinucleotides of the particular NtC/sequence combination. Bottom: 3D maps of water probability density calculated from the water distribution by Fourier averaging (Schneider et al., 1998 ▸). Water probability density maps were calculated separately for waters closer than d _calc = 3.4 Å to the base (blue mesh) and to the sugar-phosphate atoms (yellow mesh) of the reference dinucleotide. Peaks in the hydration densities, shown as yellow and cyan spheres, are interpreted as hydration sites (HSs). (c) Water probability density maps calculated for PDB entry 6l75 (Jhan et al., 2021 ▸) by overlapping the relevant hydrated building blocks over the crystal DNA structure and performing Fourier averaging.

Figure 2

Water densities in uncomplexed DNA (light colors) and protein–DNA complexes (darker colors); the HSs are depicted as small balls. The water distributions associated with bases and backbone are shown in blue and yellow, respectively. Water densities around dinucleotides of (a) BB00/CG and (b) BB00/AA NtC/sequence combinations are shown.

Figure 3

The hydration patterns of selected sequences of highly populated NtC classes. Water probability densities are displayed as a mesh and HSs as small balls. Base-related water density and HSs are displayed in blue and those for the backbone in yellow. These and other hydration patterns can be visualized using WatNA. (a) Typical hydration of bases in the BB00 class shown for the CG sequence. Pyrimidines have one HS in each groove; purines have two in the major groove and one in the minor groove. (b) HSs of the phosphate charged O atoms OP1 and OP2 in all 16 dinucleotide sequences of BB00, the NtC class best representing the BI form. Spheres in different shades of yellow represent HSs in different sequences; water densities are only displayed in matching colors for four of the sequences for clarity. The alignment shows three HSs for both OP1 and OP2 atoms. The most prominent HS in all sequences is the HS hydrating the OP2 atom. Waters constituting this HS are stabilized by van der Waals contact with the pyrimidine C6 or purine C8 base atom (base and sugar atoms are not shown for clarity). (c) Both NtC classes representing the BI form, BB00 and BB01, have similar structures and their hydration is also very similar. Data are for the AC sequence. (d) The BII form, NtC class BB07, has a unique HS in the minor groove. The HS is stabilized by contacts to N3(R)/O2(Y) of the first base and deoxyribose O4′ atoms of both sugar rings. The BII hydration is shown here for the CA sequence. (e) The hydration of the A form represented here by NtC class AA00 in the CG sequence. The most prominent difference relative to the hydration pattern of the BI form is the phosphate hydration, with two large water densities connected to the phosphate OP2 atom. One of these densities overlaps with the base hydration and water would fluctuate between the positions of the HSs highlighted by red arrows. (f) Hydration of the NtC class BBS1 occurring in the GG sequence, mostly in quadruplex structures. (g) The hydration pattern of the CG step of the Z form that is described by NtC class ZZ1S. (h) The hydration pattern of the GC step of the Z form, described by NtC class ZZS1. The hydration pattern of the left-handed DNA is very different from that in both right-handed forms: water densities are mostly located between the base planes rather than in the planes, and base and phosphate hydration merge in the minor rather than the major groove.

Figure 4

Histograms showing the percentages of predicted HSs located at a given distance interval from any of the crystallographically observed water molecules in the ten analyzed crystal structures listed in Table 2 ▸. Percentages for DNA base hydration are shown in blue and those for the sugar-phosphate backbone are in yellow. The graphs were created using matplotlib (Hunter, 2007 ▸).

Figure 5

Prediction of the hydration structure in the minor groove of B-like DNA duplexes. (a) The predicted hydration reproduces the first-shell portion of the spine of hydration in the AT-rich region of PDB entry 7cby (Dai et al., 2020 ▸) between the T12T13 step of chain A base-paired to the A4A5 step of chain B. (b) Double string of hydration reproduced in PDB entry 7cby (Dai et al., 2020 ▸). Crystal water positions are shown as red spheres and predicted hydration sites for the bases and sugar-phosphate backbone are shown as cyan and yellow spheres, respectively; the corresponding hydration probability densities are shown as cyan and yellow mesh. Figures are clipped for clarity. The predicted hydration densities for PDB entry 7cby and nine others can be viewed interactively using the WatNA web application.

Figure 6

The predicted water probabilities and HSs correctly reproduce differences between phosphate hydration in the A and B forms. (a) Phosphate hydration in the A-DNA structure with PDB code 6l75 (Jhan et al., 2021 ▸). The phosphates are bridged by water molecules in accordance with the ‘economy of hydration’ principle. (b) Phosphate groups hydrated separately in the B-DNA structure with PDB code 7m7n (Gregory et al., 2021 ▸).

Figure 7

Screenshot of the A-B-Z tab of the WatNA application. The left column contains the interactive data table, the middle column contains the visualization panel with the Mol* viewer (Sehnal et al., 2021 ▸) and the right column contains the Measurements panel (collapsed) and the Controls panel. Hydration of the BB00 CG and TA dinucleotides is visualized. The reference dinucleotide structures are shown in a ball-and-stick representation and HS positions are shown as colored spheres; the hydration probability distribution is contoured at a selected pseudo-occupancy level using a colored mesh.