SUNY at Albany
June 19-23, 2001
Investigation of the energetics of C-HáááO hydrogen bonds in the DNA i-motif via the equilibrium between alternative intercalation topologies
The i-motif is formed by DNA sequences carrying a stretch of at least two cytidines. It involves the head-to-tail intercalation of two parallel-stranded duplexes, each spanned by hemi-protonated C¥CH+ base pairs (1). The structure includes two wide and two narrow grooves. The latter bring the phosphate groups in close proximity, thus enhancing electrostatic repulsion. This is surprising, because at first sight no structural element prohibits enlarging the narrow grooves by a relative rotation of the two duplexes (2). The narrowness of the groove also leads to proximity of sugar rings, with short distances between O4' on one strand and H1' on the anti-parallel strand across the narrow groove.
Berger et al. pointed out that the short inter-atomic distances are evidence for a systematic network of C-HáááO bonds, connecting the anti-parallel strands which form the rims of each narrow groove (2). They proposed that these bonds contribute to the stability of the i-motif and are the reason for the narrowness of the narrow groove, which is required for their formation.
The energetics of the C-HáááO bonds may be probed in the oligo-nucleotides d(Cn) which form i-motif tetramers with two intercalation topologies. In one, the four external residues of the i-motif are the residues at the 3' end of each strand (3'E topology), whereas in the other, they are those of the 5' end (5'E topology). Intercalation is maximum in both topologies.
In the 3'E topology, the n sugars are involved in C-HáááO bonds. For example, with n = 3, the bonds connect sugars 1, 2, 3 of one strand to (respectively) sugars 3, 2, 1 of the strand across the narrow groove. But in the 5'E topology, the sugar moiety of residue 1, at the 5' end, has no partner across the groove: for n = 3, the network now pairs off sugars 2, 3 of one strand with sugars 3, 2 of the other. With n - 1 instead of n hydrogen-bonded sugars, the 5'E topology is unfavored.
Another difference concerns the phosphate groups (the phosphate between sugars p and p +1 is indexed as p). In the 3'E topology, phosphates of anti-parallel strands face each other across the narrow grooves (for n = 3, phosphates 1, 2 face 2, 1, respectively) whereas they are shifted by one step (2 now facing 2) in the 5'E topology.This spreads out the phosphate charges by about 0.3 nm along the symmetry axis, and is therefore electrostatically favorable.The influence of this contribution can be reduced by increasing the salt concentration.
This description leads us to expect the following:
We studied the i-motif topology of d(Cn) with n limited to 3, 4 and 5, because of solubility problems. The temperature was 0¡C. The two topologies are easily observed in an NMR TOCSY spectrum. High salt favors the 3'E topology, in agreement with the first prediction. The concentration ratio of the two topologies in high salt varies only slightly with n, as expected from the second prediction. These observations support the model.
The concentration ratio in high salt is about 8, which corresponds to a free energy difference of 5.2 kJ/mol. Since the interactions described above amount to two to three C-HáááO bonds per sugar pair (2), and since the two topologies differ by two sugar pairs per i-motif tetramer, the free energy difference amounts at most to 1.3 kJ/mol per C-HáááO bond. This is much smaller than a typical estimate of 8 kJ/mol for such bonds (as compared to 4 kJ/mol for van der Waals associations, and 12 to 29 kJ/mol for N-HáááO bonds).
As in similar cases, the structural change from 3'E to 5'E topology does not result so much in a loss of H-bonds as in the substitution of internal H-bonds by H-bonds with water. Still, the small free energy derived from the measurements does not argue for a major contribution of the proposed C-HáááO bonds to the stability of the i-motif, or for an explanation of groove narrowness in terms of C-HáááO bonds.References and Footnotes
Jean-Louis Leroy, Karim Snoussi and Maurice Gu?ron
Groupe de Biophysique de l'UMR 7643 du CNRS, Ecole Polytechnique, 91128 Palaiseau, France.