Beyond Intra-Base-Pair Hydrogen Bonds in DNA Structures: A Comprehensive Analysis

Jürgen Sühnel (jsuehnel@imb-jena.de), Klaus Lindauer (klaus@imb-jena.de)
Institut für Molekulare Biotechnologie, Postfach 100813, D-07708 Jena / Germany

This is the HTML version of the poster. The Figures are available in GIF and Postscript format. Therefore you need separate viewers for displaying GIF and Postscript files. In addition, the molecular structures are displayed in the VRML format, which enables you to interact with them and which again requires that you have available a VRML viewer. A short introduction to the VRML format as well as a collection of more than 600 molecular VRML images can be found in the Virtual Reality Division of the Image Library of Biological Macromolecules at the IMB Jena. See also the MGMS poster Image Library of Biological Macromolecules - Virtual Reality Modeling at Work.
In addition to the HTML poster you can check out a version of the poster in PDF format. This requires that you have installed the Acrobat reader, which you can get for free from Adobe. The PDF poster contains links which are indicated by the pointer changing to a pointing finger. The VRML links work only correctly if you have specified your favorite VRML viewer in the Edit/Preferences/WWW Browser Application menu of the Acrobat Reader.

Introduction

Hydrogen bonding in Watson-Crick pairs is of utmost importance for the structure of DNA. From the increasing number of structural studies on DNA we have learned, however, that this standard view of hydrogen bonding in DNA has to be supplemented by hydrogen bonding interactions of the non-Watson-Crick type, usually called mismatches or non-canonical pairings. In studies on the tracts of DNA dodecamers Nelson et al. (Nature 1987, 330, 221) and Coll et al. (PNAS 1987, 84, 8385) have identified for the first time cross-strand diagonal H-bonds. They represent a third type of H-bonds in DNA structures because they connect different base-pairs.These inter-base-pair-hydrogen bonds have been claimed to contribute to a particular stability and rigidity of AT tracts in DNA. In order to gain deeper insight into the importance of this type of H-bonds for DNA we have performed a systematic search for non-intra-base-pair hydrogen bonds in a set of 202 DNA structures from the Protein Data Bank.

Method

The DNA structure set was scanned for hydrogen bonds using our program HBexplore. It is a new tool for the comprehensive analysis of hydrogen bonding patterns in biological macromolecules (Lindauer, Bendic, Sühnel, CABIOS, in press). Here you can download a preprint. HBexplore identifies hydrogen bonds according to geometrical criteria.
The following criteria were used:

aHA < 2.5 Å, dDA < 3.9 Å, aDHA > 90 deg, aDAA1 > 90 deg, aHAA1 > 90 deg,

where d and a stand for distance and angle, D and A are the H-bond donor and acceptor atoms and A1 is the bonded neighbor of A.

Results

The results shown include distributions of geometrical H-bond parameters for all H-bonds identified and for the non-intra-base-pair H-bonds alone, other statistical data on non-intra-base-pair H-bonds and a few individual motifs.

Distributions of geometrical H-bond parameters

The total number of H-bonds identified in the set of 202 DNA structures is 5776. Figures 1 and 2 show the corresponding distribution of the H...A distances and of the D-H...A angles. To the best of our knowledge, this for the first time that a distribution of geometrical hydrogen bond parameters for a set of DNA structures which can be assumed to be representative for all DNA structures currently known is presented. We have identified 82 non-intra-base-pair H-bonds, which corresponds to a fraction of about 1.4% as compared to the total number of H-bonds. The Figures 3 and 4 exhibit the corresponding distributions for the non-intra-base-pair H-bonds alone.

Fig. 1. Distribution of H...A distances for all H-bonds (GIF, Postscript)
Fig. 2. Distribution D-H...A angles for all H-bonds (GIF, Postscript)
Fig. 3. Distribution of H...A distances for non-intra-base-pair H-bonds (GIF, Postscript)
Fig. 4. Distribution of D-H...A angles for non-intra-base-pair H-bonds (GIF, Postscript)

H-bond motifs with non-intra-base-pair H-bonds

In Figure 5 a few general motifs with non-intra-base-pair hydrogen bonds are displayed.

Fig. 5. General motifs including non-intra-base-pair H-bonds (GIF, Postscript)

Figures 6 to 9 show various motif types with non-intra-base-pair H-bonds identified in the DNA structures analyzed.

Fig. 6. Base-base inter-strand H-bond motif in an A-DNA (PDB code: 1dn6) (GIF, PostScript, VRML)
Fig. 7. Base-backbone intra-strand H-bond motif from a telomeric DNA (PDB code: 200d) (GIF, Postscript, VRML)
Fig. 8. Base-base intra-strand H-bond motif in a DNA bisintercalator complex (PDB code: 108d) (GIF, Postscript, VRML)
Fig. 9. Motif with a hydrogen which is involved in two non-intra-base-pair H-bonds in a telomeric DNA (PDB code: 186d) (GIF, Postscript, VRML)

Statistical Information

A complete list of all structures analyzed is given in Table I and a database of all non-intra-base-pair hydrogen bonds identified is provided in Table II together with their geometrical parameters and the H-bond motif they are involved in. Finally, Table III contains additional statistical information on the non-intra-base-pair H-bonds.

Summary

A set of 202 DNA structures was scanned for non-intra-base-pair H-bonds. We could identify 82 H-bonds of this type in 44 structures. This corresponds to a fraction of about 1.4% as compared to the total number of H-bonds (5776).

Non-intra-base-pair H-bonds were found in A-DNA, B-DNA, DNA-drug complexes, DNA-protein complexes and unusual DNA structures, like tetraplexes, but not in Z-DNA. In DNA-protein complexes the fraction of structures with non-intra-base-pair H-bonds is significantly higher than for the other DNA classes. Less then 50% of the non-intra-base-pair H-bonds are three-center bonds and about 30% are of the base-backbone type. The preferred nucleotide pair is AT, even though all other nucleotide combinations, except for AC, occur as well. This is also correct for the GC interaction in spite of its assumed rigidity due to the three H-bonds. On the other hand, there is a large number of AT sequence ranges which do not show non-intra-base-pair H-bonds at all. Moreover, the great majority of known structures has central AT tracts, which means that the results are biased towards the AT case. Therefore, the assumption that non-intra-base-pair H-bonds in AT tracts are responsible for their particular rigidity is not clearly substantiated.

When increasing the cutoff value for the HA distance from 2.5 to 3Å the fraction of non-Z-DNA structures with non-intra-base-pair H-bonds increases from 24% in the first case to 75% in the second one.

Finally the distributions of HA distances and DHA angles for all H-bonds in the structure set are reported. These distributions have maxima at HA distances between 1.8 and 1.9 Å and at DHA angles between 160 and 170 °.