Factors Contributing to Aromatic Stacking in Water: Evaluation in the Context of DNA

Abstract

We report the use of thermodynamic measurements in a self-complementary DNA duplex (5'-dXCGCGCG)(2), where X is an unpaired natural or nonnatural deoxynucleoside, to study the forces that stabilize aqueous aromatic stacking in the context of DNA. Thermal denaturation experiments show that the core duplex (lacking X) is formed with a free energy (37 °C) of -8.1 kcal·mol(-1) in a pH 7.0 buffer containing 1 M Na(+). We studied the effects of adding single dangling nucleosides (X) where the aromatic "base" is adenine, guanine, thymine, cytosine, pyrrole, benzene, 4-methylindole, 5-nitroindole, trimethylbenzene, difluorotoluene, naphthalene, phenanthrene, and pyrene. Adding these dangling residues is found to stabilize the duplex by an additional -0.8 to -3.4 kcal·mol(-1). At 5 μM DNA concentration, T(m) values range from 41.7 °C (core sequence) to 64.1 °C (with dangling pyrene residues). For the four natural bases, the order of stacking ability is A > G ≥ T = C. The nonpolar analogues stack more strongly in general than the more polar natural bases. The stacking geometry was confirmed in two cases (X = adenine and pyrene) by 2-D NOESY experiments. Also studied is the effect of ethanol cosolvent on the stacking of natural bases and pyrene. Stacking abilities were compared to calculated values for hydrophobicity, dipole moment, polarizability, and surface area. In general, hydrophobic effects are found to be larger than other effects stabilizing stacking (electrostatic effects, dispersion forces); however, the natural DNA bases are found to be less dependent on hydrophobic effects than are the more nonpolar compounds. The results also point out strategies for the design nucleoside analogues that stack considerably more strongly than the natural bases; such compounds may be useful in stabilizing designed DNA structures and complexes.

Figure 1

The structures of natural and nonnatural deoxynucleosides in this study.

Figure 2

(A) Examples of thermal denaturation data for several of the dangling end sequences in this study. The dangling residues for each melting curve are given on the plot. (B) Van't Hoff plots for the same sequences as in A.

Figure 3

Relationships between stacking free energies (Table 2) and calculated physical properties of the DNA bases and aromatic analogues in this study (Table 1). (A) Hydrophobicity, as measured by log P for the methylated bases; (B) calculated polarizability; (C) estimated surface area of dangling residue excluded from solvent on stacking; (D) dipole moment of the methylated base. Data for natural DNA bases are given as circles, and other analogues, triangles. Abbreviations for the bases are as follows: A (adenine), B (benzene), C (cytosine), F (difluorotoluene), G (guanine), H (phenanthrene), I (5-nitroindole), M (4-methylindole), N (naphthalene), P (pyrrole), T (thymine), Y (pyrene), Z (trimethylbenzene).

Figure 4

Illustrations of possible 5′-end stacking geometries for the aromatic rings in this study. Models were built using canonical B-form geometry and placing the dangling residue at the 5′-end adjacent to a C–G pair. (A) Dangling pyrimidine (specifically thymine or difluorotoluene), (B) purine (specifically adenine or 4-methylindole), (C) benzene, (D) naphthalene, (E) phenanthrene, (F) pyrene. Stacked geometries were confirmed for adenine and pyrene by NMR experiments.

Figure 5

Imino proton spectra at 0 °C for (5′-dCGCGCG)₂ (bottom), (5′-dACGCGCG)₂ (center), and (5′-dYCGCGCG)₂ (top). Note the strong upfield shift of the imino proton in the terminal C-G pair caused by the adjacent stacking of the 5′-dangling bases A and Y.

Figure 6

The effect of added ethanol (as volume percent) on the T_m for five different dangling residues. The slopes of the lines are given in parentheses. Correlation coefficients for the lines range from 0.98 to >0.999.