Divalent cations and the electrostatic potential around DNA: Monte Carlo and Poisson-Boltzmann calculations.

The predictions of counterion condensation theory for divalent ions were tested by comparison with the results of Monte Carlo calculations on an all-atom model of DNA. Monovalent-divalent competition at the polyelectrolyte surface was investigated by varying the partial molar volume of divalent ions. To assess the viability of using Poisson-Boltzmann (PB) calculations for determining divalent ion concentrations at DNA surfaces, Monte Carlo (MC) calculations were compared with PB calculations using different models of the dielectric continuum. It was determined that, while standard PB calculations of divalent ion surface densities are about 25-30% below those predicted by MC techniques, and somewhat larger than errors previously determined for monovalent ions, errors due to the use of the mean-field approximation of PB theory are smaller than those arising from common assumptions regarding the dielectric continuum.

Proton equilibria in the minor groove of DNA.

Poisson-Boltzmann calculations by Pack and co-workers suggest the presence of regions of increased hydrogen ion density in the grooves of DNA. As an experimental test of this prediction, we have attached proton-sensitive probes, with variable linker lengths, to random-sequence DNA at G sites in the minor groove. The amino groups of beta-alanine, gamma-aminobutyric acid (GABA), and epsilon-aminocaproic acid have been coupled at pH 5, via a formaldehyde link, to the exocyclic amino group of guanine, utilizing a reaction that has been extensively investigated by Hanlon and co-workers. The resulting adducts at pH 5 retained duplex B form but exhibited typical circular dichroism (CD) changes previously shown to be correlated with the presence of a net positive charge in the minor groove. Increases in the solvent pH reversed the CD spectral changes in a manner suggesting deprotonation of the carboxylic acid group of the adduct. These data were used to calculate an apparent pK(a) for the COOH. The pK(a) was increased by 2.4 units for beta-alanine, by 1.7 units for GABA, and by 1.5 units for epsilon-amino caproic acid, relative to their values in the free amino acid. This agrees well with Poisson-Boltzmann calculations and the energy minimization of the structures of the adducts that place the carboxyl groups in acidic domains whose hydrogen ion density is approximately 2 orders of magnitude greater than that of bulk solvent.

Monte Carlo and Poisson-Boltzmann calculations of the fraction of counterions bound to DNA.

The counterion density and the condensation region around DNA have been examined as functions of both ion size and added-salt concentration using Metropolis Monte Carlo (MC) and Poisson-Boltzmann (PB) methods. Two different definitions of the "bound" and "free" components of the electrolyte ion atmosphere were used to compare these approaches. First, calculation of the ion density in different spatial regions around the polyelectrolyte molecule indicates, in agreement with previous work, that the PB equation does not predict an invariance of the surface concentration of counterions as electrolyte is added to the system. Further, the PB equation underestimates the counterion concentration at the DNA surface, compared to the MC results, the difference being greatest in the grooves, where ionic concentrations are highest. If counterions within a fixed radius of the helical axis are considered to be bound, then the fraction of polyelectrolyte charge neutralized by counterions would be predicted to increase as the bulk electrolyte concentration increases. A second categorization--one in which monovalent cations in regions where the average electrostatic potential is less than -kT are considered to be bound--provides an informative basis for comparison of MC and PB with each other and with counterion-condensation theory. By this criterion, PB calculations on the B form of DNA indicate that the amount of bound counterion charge per phosphate group is about .67 and is independent of salt concentration. A particularly provocative observation is that when this binding criterion is used, MC calculations quantitatively reproduce the bound fraction predicted by counterion-condensation theory for all-atom models of B-DNA and A-DNA as well as for charged cylinders of varying linear charge densities. For example, for B-DNA and A-DNA, the fractions of phosphate groups neutralized by 2 A hard sphere counterions are 0.768 and .817, respectively. For theoretical studies, the radius enclosing the region in which the electrostatic potential is calculated to be less than -kT is advocated as a more suitable binding or condensation radius than that enclosing the fraction of counterions given by (1 - epsilon-1). A comparison of radii calculated using both of these definitions is presented.

The effect of a variable dielectric coefficient and finite ion size on Poisson-Boltzmann calculations of DNA-electrolyte systems.

The results of variable dielectric coefficient Poisson-Boltzmann calculations of the counter-ion concentration in the vicinity of an all-atom model of the B-form of DNA are presented with an emphasis on the importance of spatial variations in the dielectric properties of the solvent, particularly at the macro-ion-solvent interface. Calculations of the distribution of hard-sphere electrolyte ions of various dimensions are reported. The presence of a dielectric boundary significantly increases the magnitude of the electrostatic potential with a concomitant increase in the accumulation of small counter-ions in the groove regions of DNA. Because ions with radii greater than 2 A have restricted access to the minor groove, the effect there is less significant than it is within the major groove. Changes in the dielectric coefficient for the electrolyte solution, allowing variation from 10 to 25, 40, 60, and 78.5 within the first 7.4 A of the surface of DNA, substantially increases the calculated surface concentration of counter-ions of all sizes. A lower dielectric coefficient near the macro-ion surface also tends to increase the counter-ion density in regions where the electrostatic potential is more negative than -kT. Regardless of the choice of dielectric coefficient, the number of ions in regions where the electrostatic potential is less than -kT remains the same for 0.153 M added 1-1 monovalent electrolyte as for the case without added salt. The strong dependence of the calculated distribution of counter-ion density on the choice of dielectric coefficients representing the solvent continuum suggests that care must be taken to properly characterize the physical system when studying electrostatic properties using these methods.

Acidic domains around nucleic acids.

The hydrogen ion concentration in the vicinity of DNA was mapped out within the Poisson-Boltzmann approximation. Experimental conditions were modeled by assuming Na-DNA to be solvated in a buffer solution containing 45 mM Tris and 3 mM Mg cations at pH 7.5. Three regions of high H+ concentration (greater than 10 microM) are predicted: one throughout the minor groove of DNA and two localized in the major groove near N7 of guanine and C5 of cytosine for a G.C base pair. These acidic domains correlate well with the observed covalent binding sites of benzo[a]pyrene epoxide (N2 of guanine) and of aflatoxin B1 epoxide (N7 of guanine), chemical carcinogens that presumably undergo acid catalysis to form highly reactive carbocations that ultimately bind to DNA. It is suggested that these regions of high H+ concentration may also be of concern in understanding interactions involving proteins and noncarcinogenic molecules with or near nucleic acids.

Counterion distribution around DNA: variation with conformation and sequence.

The three-dimensional Poisson-Boltzmann equation for the distribution of counterion charge density around double helical DNA has been solved by an iterative procedure. These computations have been performed for 0.01M monovalent salt solutions. A systematic study of the ten possible sequences found among adjacent nucleotide base pairs is presented for the A, B, and C conformations of DNA. In addition, calculations of the electrostatic stabilities of these conformations of DNA allows for comparison of the charge accumulation around each of the three conformers.

Quantum chemical calculations on the two-step mechanism of proflavin binding to DNA.

Quantum chemical calculations on the binding of proflavin to DNA lead to a model in which the outside binding to a phosphate group leads to an induced fit in the intercalation receptor site. The calculations suggest hydrogen bonding of the amine groups of the outside bound proflavin to the anionic oxygen of the backbone phosphate. The resulting partial neutralization facilitates the conformational transitions required for intercalation. The results are consistent with the observed preference of proflavin for dCpdG over dGpdC sequences and with the observed kinetics of the binding reaction.

The molecular structure and charge distribution of phorbol: parent compound of the tumor promoting phorbol diesters.

The geometry of phorbol, parent compound of the very highly active tumor promoting phorbol diesters, was totally optimized starting from the crystal structure of phorbol-5-bromofuroate-chloroform solvate (Petterson et al., J. Chem. Soc. B980, 1968). The full optimization of all atomic positions was done quantum chemically using the MINDO/3 potential energy surface and a modified Davidon-Fletcher-Powell search for local minima. The molecular structure and details of the electronic charge distribution are presented.

Quantum chemical studies of the metabolism of polycyclic aromatic amines and the stabilities and electrophilicities of their arylnitrenium ions in relation to their mutagenic/carcinogenic potencies.

Electronic parameters related to the cytochrome P450-catalyzed reactions of eight polycyclic aromatic amines have been calculated using all valence electron semiempirical molecular orbital methods. The reactions considered lead to the presumably carcinogenic arylnitrenium ions and to the competing hydroxylation and epoxidation products. The stabilities of the arylnitrenium ions relative to the N-hydroxylamines and their sulfate esters were also calculated, together with electrophilic reactivity parameters of the ions. The resulting parameters were used to predict major metabolites of the parent compounds and also to correlate with observed mutagenic activities of the four pairs of polycyclic aromatic amines studied. The major factor in determining mutagenic potencies of parent compounds appears to be the extent of N-hydroxylation and competing ring oxidations, as well as the electrophilic properties of the arylnitrenium ions.

Origins of the specificity in the intercalation of ethidium into nucleic acids. A theoretical analysis.

The sequence preferences observed in the intercalative binding of ethidium to dinucleoside phosphates have been theoretically examined. This specificity is for pyrimidine (3'--5') purine sequences as compared to their purine (3'--5') pyrimidine sequence isomers. It is shown that the stacking energies between the ethidium cation and the base pairs are fairly constant for all combinations of bases at the intercalation site. In contrast, the energy of unwinding the double helix to assume the geometry of the intercalation complex shows substantial sequence differences. Thus, the specificity observed is more readily explained in terms of these conformation energy changes than by preferential stacking interactions. These results imply that there may be a large class of intercalating molecules which exhibit similar pyrimidine (3'--5') purine sequence specificity.

Kinked DNA. Energetics and conditions favoring its formation.

Theoretical calculations demonstrate that the formation of kinks in DNA in the manner proposed by Crick and Klug produces a conformation which corresponds to a local energy minimum when the phosphate backbone is charged. This conformation becomes the global minimum under certain conditions which neutralize the anionic backbone oxygens. A sequence preference for kink formation is predicted. The possible role of kinking in experimentally observed phenomena and experimental means of demonstrating the existence of kinks are discussed.