Thomas R. Krugh

Evidence for sequence preferences in the intercalative binding of ethidium bromide to dinucleoside monophosphates

The formation of solution complexes of ethidium bromide with the self-complementary ribodinucleoside monophosphates, as well as ethidium complexes with mixtures of complementary and non-complementary ribodinucleoside monophosphates, has been monitored by circular dichroism, fluorescence, nuclear magnetic resonance, and visible spectroscopic techniques. Although ethidium bromide will form a complex with all of the dinucleoside monophosphates used in this study, it does exhibit a definite preference for binding to dinucleosides that have a pyrimidine (3′-5′) purine sequence when compared to their isomeric purine (3′-5′) pyrimidine sequence dinucleosides. The pyrimidine-purine sequence dinucleoside monophosphate complexes with ethidium exhibit fluorescence, circular dichroism, and visible spectra that closely resemble the spectra observed for the ethidium complexes with double-stranded nucleic acids. An analysis of the nuclear magnetic resonance spectra of the same ethidium-dinucleoside monophosphate solutions used for the optical experiments shows that complexation results in the formation of a miniature double-helical complex in which the phenanthridinium ring of ethidium bromide is intercalated between the adjacent base-pairs of a miniature double-helical complex. The formation of a double-helical complex is also indicated by the observation of the hydrogen-bonded guanine ring NH resonance in a complex of ethidium bromide with CpG dissolved in an aqueous solution. The combination of the optical spectroscopies and the nuclear magnetic resonance experiments on these model systems thus provides direct evidence for the intercalation model for the binding of ethidium bromide to double-stranded nucleic acids. The fact that ethidium preferentially interacts with dinucleoside monophosphate sequence isomers suggests that it binds to the various base sequences available as intercalation sites on DNA and RNA with significantly different binding constants.

Solution structure of a DNA duplex containing a replicable difluorotoluene-adenine pair.

A nonpolar aromatic nucleoside derivative based on 2,4-difluorotoluene (F), a non-hydrogen bonding shape analog of thymidine, was recently shown to be replicated against adenine with high efficiency and fidelity. This led to the suggestion that geometric matching, potentially even in the absence of hydrogen bonding between bases in a pair, may be sufficient to direct nucleotide selection during replication. We have examined the solution structure of the F–A pair in the context of a 12 base pair DNA duplex. We find that, despite the destabilization caused by this analog, the F–A pair very closely resembles that of a T·A pair in the same context. This lends support to the importance of shape matching in replication.

Drug-DNA interactions

Abstract Structures of drug-DNA oligomer complexes provide a molecular basis for the interpretation of sequence-selective binding as measured by DNA footprinting. Thermodynamic data, obtained from calorimetry, thermal denaturation, and equilibrium measurements are also an important aspect of drug-DNA complex formation. Although many ligand-DNA complexes exhibit a single conformation in solution, there are examples of systems undergoing exchange between two or more conformations in solution. A 2-aminofluorene-DNA oligomer duplex adopts two conformations, one exhibiting Watson-Crick pairing at the modified guanine, and a second conformation in which base pairing is disrupted and the fluorene moiety is stacked within the duplex. NMR-derived structures of oligonucleotides with actinomycin D, nogalomycin, bisintercalartors, lexitropsins and triplex binding ligands highlight recent advances in this area. The development of (pyridine-2-carboxamide-neotropsin)2-C6, a synthetic dimer that binds in the minor groove, is but one example of the synergistic interaction of NMR structural analysis, molecular modeling, and chemical synthesis.

Equilibrium binding of carcinogens and antitumor antibiotics to DNA: site selectivity, cooperativity, allosterism

The equilibrium binding of the carcinogens N-hydroxy-N-acetyl-2-amino-fluorene (HAAF) and 4-nitroquinoline-1-oxide (NQO) to phi X174RF DNA have been studied by phase partition techniques. Both molecules bind in a cooperative manner with only a few carcinogen molecules binding to each phi X174RF DNA molecule. The binding data for both HAAF and NQO fit a model in which two carcinogens cluster into a small number of sites--four sites for HAAF and twelve sites for NQO. Phase partition techniques were also used to study the binding of actinomycin D to both calf thymus DNA and poly (dG-dC) . poly (dG-dC) at much lower r values than had been previously reported. These data exhibit humped Scatchard plots which are indicative of cooperative binding; the overall shape of the Scatchard plots are consistent with a model for drug induced allosteric transitions in the DNA structure. The cooperativity in the actinomycin D binding to calf thymus DNA increases with decreasing sodium chloride concentration, suggesting a role for DNA flexibility in allosteric binding.

Proton Fourier transform nmr spectroscopy in H2O solutions on a JEOL PFT-100. The WEFT, SWEFT, and VASE pulse sequences

Abstract An inexpensive pulse generating circuit which adds a pulse to the pulse train generated by the JEOL DP-1 pulse programmer is described. This circuit allows for the generation of a 180°−τ−90° pulse sequence with a continuously variable interval (τ). The 180°−τ−90° pulse sequence has been used to record the 100-MHz proton Fourier transform NMR spectrum of 5 × 10−4 M pdC-dG in an H2O solution. The use of a 180°−τ1−180°−τ2−90° (SWEFT) pulse sequence for the simultaneous elimination of any two resonances is illustrated. A variable angle signal elimination (VASE) pulse sequence that allows for the use of a minimum interval between the pulses is discussed. An inexpensive circuit for the generation of a homogeneity spoiling pulse is also presented.

Large-scale purification of synthetic oligonucleotides and carcinogen-modified oligodeoxynucleotides on a reverse-phase polystyrene (PRP-1) column

A procedure is described for the large-scale purification of synthetic oligonucleotides using a polystyrene (PRP-1, Hamilton Co.) high-performance liquid chromatography (HPLC) column with a phosphate/methanol/acetonitrile solvent system. Pure oligonucleotides are obtained with a three-step procedure that involves only one column purification step. The dimethoxytrityl group is left on the oligomer for the HPLC purification. The use of the PRP-1 polystyrene column with a phosphate/methanol/acetonitrile solvent system provides excellent separation of the desired dimethoxytrityl-bearing oligonucleotide from failure sequences. The dimethoxytrityl group is removed by treatment with acetic acid and the oligonucleotide is desalted on a C-18 Sep-Pak cartridge. The oligodeoxynucleotides obtained are shown to be essentially pure by HPLC, polyacrylamide gel electrophoresis, and 500-MHzNMR spectroscopy. This procedure is especially useful for the large-scale purification of oligonucleotides required for NMR studies. The PRP-1 column and the phosphate/methanol/acetonitrile solvent system is useful for purifying modified oligonucleotides containing lipophilic groups such as the carcinogen 2-(acetylamino)fluorene.

Nuclear magnetic resonance studies of hydrogen bonded complexes of oligonucleotides in aqueous solution. I. pdG-dC and pdG-dT

The self-interaction of two deoxydinucleotides, pdG-dC and pdG-dT, was studied (in aqueous solution) by 100 MHz FT NMR spectroscopy. Concentration studies show that the self-complementary pdG-dC forms hydrogen bonded complexes. An analysis based on the concentration dependence of the chemical shift of the guanine amino protons strongly suggests that hydrogen bonded dimer formation occurs with a K for the dimerization equilibrium of 7.8 ± 0.7 M−1. On the other hand, the non-self-complementary pdG-dT does not give evidence of similar complex formation in the same concentration range which thus illustrates the importance of complementarity in the formation of hydrogen bonded complexes of deoxydinucleotides. Further impetus is therefore provided for the use of dinucleotides in modeling interactions common to nucleic acids.

Solute enhanced partition analysis-a novel method for measuring the binding of drugs to DNA.

Abstract A solute enhanced phase partition method is described for the measurement of the binding of drugs to nucleic acids. The inclusion of a solute which acts as a phase transfer reagent allows one to enhance the solubility of charged molecules in organic solvents by as much as three orders of magnitude. This development increases the utility of partition analysis as a general method for studying the interaction of small molecules with macromolecules. Solute enhanced partition analysis (SEPA) has been used to measure the DNA binding of positively charged drugs at very low levels of drug binding, where we have observed cooperative binding of daunorubicin to calf thymus DNA.

Single-cell partition analysis--a direct fluorescence technique for examining ligand-macromolecule interactions.

Single-cell partition analysis is described as a novel technique for examining ligand-macromolecule interactions. This procedure is a combination of the classical fluorescence titration technique and phase-partition techniques and allows three separate methods for calculating and comparing both free and bound drug concentrations. The value of this technique is demonstrated by the comparison of the binding properties of the potent antitumor antibiotic adriamycin and ethidium bromide to nucleic acids. Binding isotherms of both drugs were obtained at low r (concentration of bound drug per base pair) values, showing strikingly different results, thus allowing insight to be gained into the cooperative binding of these drugs to DNA.

Measurement of relaxation times by the adiabatic half passage (T1ϱ) technique

Abstract The use of the adiabatic half passage (T1ϱ) technique for the measurement of the relaxation times T1 and T2 for systems containing closely spaced resonances or large solvent peaks has been investigated. An analytic function for the time dependence of the dispersion mode signal employed in this technique has been developed and the conditions are given for the successful measurement of relaxation times in the presence of neighboring resonances. The modification of a JEOL-4H-100 spectrometer to measure nuclear magnetic resonance relaxation times is also described.