The Interaction Between the Novel Intercalator Diethidium Cation and B-Form DNA: a Theoretical Study

Proline-rich peptides are known to adopt preferentially the extended polyproline II (PPII) helical conformation, which is involved in several protein-protein recognition events. By resorting to molecular modelling techniques, we wished to investigate the extent to which PPII helices could be used for the formation of isochelical peptide-DNA complexes leading to the selective recognition of the major groove of B-DNA. For that purpose, we have grafted to a cationic intercalator, 9-amino-acridine, an oligopeptide having the sequence: Pro- Arg-Pro-Pro-Arg-Pro-Pro-Arg-Pro-Pro-Asp-Pro-Pro. Each residue in the sequence was set in the D configuration, to prevent enzymatic hydrolysis, and each Arg residue was designed to target O6/N7 of a guanine base following the intercalation site. The Asp residue was designed to target a cytosine base, whilst simultaneously forming a bidentate complex with the Arg three residues upstream. Energy-minimization, using the JUMNA procedure, led to the following conclusions : 1) major groove binding is favoured over minor groove or exclusive binding to the phosphates by large energy differences, of over 50 and 90 kcal/mole, respectively: 2) the two best bound sequences are those having three successive guanine bases on the same DNA strand, immediately adjacent to the intercalation site. Sequence d(CGGGC G), encountered in the Primer Binding Site of the HIV retrovirus, thus ranks amongst the best-bound sequences; 3) replacement of an individual guanine amongst the three ones upstream of the intercalation site, by an adenine base, weakens by > 6 kcal/mole the binding energetics; 4) the conformational rigidity of the DNA-bound PPII helix should enable for a modulation of the base sequence selectivity, by appropriate replacements of the Arg and Asp residues. Thus sequence CGGCAAG, also encountered in the HIV genome, could be targeted by an oligopeptide having the sequence Pro-Arg-Pro-Pro-Asp-Pro-Pro-Asn-Pro-Pro-Asn-Pro-Pro-Arg-Ala.

Extensive molecular dynamics (MD) simulations have been used to investigate the intercalative binding of 9-hydroxyellipticine to the DNA oligonucleotide d(ATATATATATAT)2. Four independent simulations differing in the initial orientation of the drug at the intercalation site were carried out, and compared both with each other and a control simulation of the free DNA sequence. The structure of the latter was compared with structures obtained from x-ray crystallography and nmr spectroscopy, as well as the theoretically derived "alternating B-DNA" model [A. Klug et al. (1979), Journal of Molecular Biology, Vol. 131, p. 669]. The alternation of twist angles observed in experimental structures was reproduced in the simulation. All four independent simulations of the drug-DNA intercalation complex converged in placing the pyridine ring of the ellipticine chromophore in the major groove; in one case this involved a 180 degrees rotation of the drug at the intercalation site. At a more detailed level, the drug is seen to be capable of adopting several distinct orientations, each stable over a period of hundreds of pico-seconds. Despite the presence of several polar groups in the drug, however, no direct hydrogen bonding to the DNA occurs; instead, interactions between the methyl groups of the drug and the thymine bases at the intercalation site appear important in determining the orientational preferences of the drug. Comparison of the intercalation complexes with the free DNA sequence shows a degree of unwinding resulting from intercalation, in good agreement with experimental results, but spread over the three central base-pair steps, not confined to the intercalation site itself. Measurements of torsional rigidity indicate only a slight stiffening of the DNA restricted to the immediate site of intercalation. The structures obtained from the MD simulations were used to calculate theoretical CD spectra, with separate simulations giving very different results. This appears to indicate that given an accurate assignment of the main electronic transition dipole moment of the ellipticine chromophore, discrimination of the more realistic binding geometries may be possible. The relative merits of the various drug orientations observed in the simulations are discussed and a perpendicular orientation of the drug at the intercalation site is considered to be the most consistent with experimental data. While the simulations themselves represent a total of over 2 ns, however, the differences apparent between independent runs indicate that longer simulation times will be required before a complete, unequivocal view of DNA intercalation is obtained.

Flexibility of intercalation site geometries within a B-DNA helix was investigated in the twist-shift plane using energy minimization methods. The parameters optimized included sugar conformation, the glycosidic angles and phosphodiester torsion angles. Our calculations show several regions of energetically favorable intercalation geometries in the twist-shift plane. Modeling studies using interactive computer graphics and electrostatic potential surface compatibility provided initial hypotheses for the structures of the drug-DNA complexes. These hypotheses were supported and extended by energy minimizations of these complexes. Binding positions, conformational features and relative minimum binding energies of two anticancer drugs, mitoxantrone and bisantrene, were computed for intercalation complexes with DNA in the theoretically defined intercalation sites. Mitoxantrone intercalates DNA from the minor groove and the side chain OH or NH groups are involved in hydrogen bonds with the main chain phosphate groups of DNA, thereby cross-linking the complementary strands. The hydroxyl groups of mitoxantrone can also participate in hydrogen bonding with phosphate oxygens of another chain, thereby cross-linking DNA helices. Bisantrene intercalates DNA favorably from the major groove and the NH group of the dihydroimidazole ring can participate in hydrogen bonding with the phosphate oxygens of the backbone. These models are consistent with the physicochemical and electron microscopic studies of the interaction of mitoxantrone and bisantrene with DNA. Our results are now being used to guide the design of novel anticancer drugs that should interact with DNA in a manner similar to that proposed for our representative drugs.

Organometallic ruthenium(II)-arene (RA) antitumour compounds of the general type [Ru(II)(η6-arene)(X)2(pta)], RA-pta, and [Ru(II)(η6-arene)X(en)], RA-en, (X=leaving group; pta=1,3,5-triaza-7-phosphaadamantane, en=ethylenediamine) have been investigated computationally. The main focus has been i.e. the hydrolysis properties, ligand exchange reactions, binding to DNA and the analysis of the resulting structural perturbations. The goal of this thesis was to elucidate key steps in the reactivity of the above drugs that are not readily accessible by experimental techniques. The methodologies employed cover tailor made force fields in classical molecular dynamics (MD), density functional theory (DFT) MD simulations in implicit and explicit solvent as well as combined classical/quantum (QM/MM) MD simulations. For RA-pta compounds, binding energies between the metal centres and the surrounding ligands were calculated. The calculated energies rationalize the experimentally observed tendencies for arene loss, and show that the pta ligands are relatively strongly bound. Exchange of metal centre, methylation or protonation of the pta-ligand, or change of the arene result in significant differences in the metal-arene binding energies while leaving the metalphosphine bond strength essentially unchanged. Significantly lower binding energies and reduced hapticity are predicted for the exchange of arene by nucleobases. The latter show higher binding energies for nitrogen π-bonding than for p-bonding. No influence on the ruthenium-arene interaction was observed for RA-pta complexes bearing arenes with functionalized side chains. A combined DFT/continuum electrostatics approach has been used to estimate the protonation states and absolute pKa values of a series of RA-pta compounds and their hydrolysis products. Our results suggest that the selective in vivo activity towards cancer cells and the observed pH dependent DNA damage is due to a pH dependent activation/deactivation mechanism. In the context of rational drug design, it is hypothesised that analogues with fluorinated arene rings should result in improved selectivity towards hypoxic cancer cells. For RA-en complexes, we rationalized the chemoselectivity towards guanine. The calculated DFT binding energies for the three investigated nucleobases (G, A, C) decreases in the order G(N7) >> C(O2) ∼ C(N3) > A(N7) > G(O6) > OH2. The G(N7) complex is the most stable product due to a C6=O-HNen hydrogen bond while the corresponding C6-NH2-HNen interaction in adenine is repulsive. A study of the reaction of [(η6-benzene)Ru(en)(OH2)]2+ (1) towards the nucleobases G, A and C reveals a strong preference for a formation of an H-bonded cis vs. trans reactant adduct. Only guanine can form a thermally stable trans reactant adduct. The potential reaction sites of adenine and cytosine might loose their nucleophilicity by protonation while simultaneously the aqua leaving group of 1 is converted into an unreactive hydroxo ligand. For the reaction of 1 with guanine three different reaction pathways were identified, however, a "direct trans" reaction pathway is probably the most biologically relevant. Using classical and QM/MM MD we showed that both investigated ruthenium compounds, RA-pta and RA-en, can bind to the major groove of duplex DNA and are stable in this position. The DNA is highly flexible, adapts very fast and widens the major groove to accommodate the ruthenium complex. The local and global structural changes of the DNA (e.g. bending towards the major groove) observed for the RA-pta series are similar to those reported for cisplatin. A severe perturbation of the Watson-Crick base pairing adjacent to the binding site of RA-en was observed that can be linked to recent experimental results. Finally, a tailor made force field for RA-en was derived from our QM/MM trajectories using a force matching approach.

The interaction of daunomycin with B-DNA double helices of several methylated deoxynucleotides, d(C-G-m5C-G), d(m5C-G-C-G), d(C-G-m5C-G-C-G) and d(m5C-G-C-G-m5C-G) in solution was investigated by 1H-NMR spectroscopy at 500 MHz. At low temperature (t less than 20 degrees C for the tetramer and t less than 40 degrees C for the hexamers), several daunomycin-DNA complexes were observed in slow exchange with the drug-free DNA duplexes. The presence of daunomycin in a self-complementary double helix cancels the conformational symmetry of the two strands; the proton signals can split into several others owing to the difference between free and intercalated duplexes and to the many possible intercalation sites in a duplex (three for a tetramer, five for an hexamer). A model relating the chemical shifts of splitted proton signals to the various intercalated duplex conformations was given. The results show that one daunomycin molecule is associated with one duplex and that it can enter any intercalation site with equal probability; no side-effects were observed even for very short helices (of a tetramer). In the case of d(C-G-m5C-G) the association constant and the dissociation and association rates of the intercalated complex were evaluated.

A generalized procedure to generate nucleic acid structures is presented. In this procedure, the bases of a base pair are oriented first for characterization of particular DNA receptor sites. The resultant sites are then used in the study of specific molecule–DNA interactions. For example, intercalation sites, kinked DNA, and twisted and tilted bases are envisioned. Alterations of structures via anti → syn orientations of bases, as well as crankshaft motion about collinear bonds, provide additional conformations without disrupting the overall backbone structure. These approaches to the generation of nucleic acid structures are envisioned as required in studies of the intercalation phenomenon, minor adjustments of DNA to accommodate denaturation, binding of carcinogens to DNA, complex formation of transition metals with DNA, and antitumor agents as ligands. For these base‐pair and base orientations, backbone orientations are calculated by the AGNAS technique to yield physically meaningful conformations, namely, those conformations for which nonbonded contacts are favourable. A procedure is presented to generate dimer duplex units that are physically meaningful and to assemble these units into a polynucleotide duplex. Double helices that begin with B‐DNA, undergo a transition to one of the above‐mentioned receptor sites, and return to B‐DNA can be assembled from a catalog of dimer duplexes. Stereographic projections of the various receptor sites already being used to model binding to DNA are presented.

The conference 'From DNA-Inspired Physics to Physics-Inspired Biology' (1–5 June 2009, International Center for Theoretical Physics, Trieste, Italy) that myself and two former presidents of the American Biophysical Society—Wilma Olson (Rutgers University) and Adrian Parsegian (NIH), with the support of an ICTP team (Ralf Gebauer (Local Organizer) and Doreen Sauleek (Conference Secretary)), have organized was intended to establish stronger links between the biology and physics communities on the DNA front. The relationships between them were never easy. In 1997, Adrian published a paper in Physics Today ('Harness the Hubris') summarizing his thoughts about the main obstacles for a successful collaboration. The bottom line of that article was that physicists must seriously learn biology before exploring it and even having an interpreter, a friend or co-worker, who will be cooperating with you and translating the problems of biology into a physical language, may not be enough. He started his story with a joke about a physicist asking a biologist: 'I want to study the brain. Tell me something about it!' Biologist: 'First, the brain consists of two parts, and..' Physicist: 'Stop. You have told me too much.'

Folic acid binds DNA and RNA at different locations

Binding of the U1A protein to its RNA target U1 hairpin II has been extensively studied as a model for a high affinity RNA/protein interaction. However, the mechanism and kinetics by which this complex is formed remain largely unknown. Here we use real-time biomolecular interaction analysis to dissect the roles various protein and RNA structural elements play in the formation of the U1A.U1 hairpin II complex. We show that neutralization of positive charges on the protein or increasing the salt concentration slows the association rate, suggesting that electrostatic interactions play an important role in bringing RNA and protein together. In contrast, removal of hydrogen bonding or stacking interactions within the RNA/protein interface, or reducing the size of the RNA loop, dramatically destabilizes the complex, as seen by a strong increase in the dissociation rate. Our data support a binding mechanism consisting of a rapid initial association based on electrostatic interactions and a subsequent locking step based on close-range interactions that occur during the induced fit of RNA and protein. Remarkably, these two steps can be clearly distinguished using U1A mutants containing single amino acid substitutions. Our observations explain the extraordinary affinity of U1A for its target and may suggest a general mechanism for high affinity RNA/protein interactions.

Polyamines are known to be involved in cell growth regulation in breast cancer. To evaluate the efficacy of bis(ethyl)polyamine analogs for breast cancer therapy and to understand their mechanism of action we measured the effects of a series of polyamine analogs on cell growth, activities of enzymes involved in polyamine metabolism, intracellular polyamine levels, and the uptake of putrescine and spermidine using MCF-7 breast cancer cells. The IC50 values for cell growth inhibition of three of the compounds, N1,N12-bis(ethyl)spermine, N1,N11- bis(ethyl)norspermine, and N1,N14-bis(ethyl)homospermine, were in the range of 1-2 µM. Another group of three compounds showed antiproliferative activity at about 5 µM level. These compounds are also capable of suppressing colony formation in soft agar assay and inducing apoptosis of MCF-7 cells. The highly effective growth inhibitory agents altered the activity of polyamine biosynthetic and catabolic enzymes and down-regulated the transport of natural polyamines, although each compound produced a unique pattern of alterations in these parameters. HPLC analysis showed that cellular uptake of bis(ethyl)polyamines was highest for bis(ethyl)spermine. We also analyzed polyamine analog conformations and their binding to DNA minor or major grooves by molecular modelling and molecular dynamics simulations. Results of these analyses indicate that tetramine analogs fit well in the minor groove of DNA whereas, larger compounds extend out of the minor groove. Although major groove binding was also possible for the short tetramine analogs, this interaction led to a predominantly bent conformation. Our studies show growth inhibitory activities of several potentially important analogs on breast cancer cells and indicate that multiple sites are involved in the mechanism of action of these analogs. While the activity of an analog may depend on the sum of these different effects, molecular modelling studies indicate a correlation between antiproliferative activity and stable interactions of the analogs with major or minor grooves of DNA.Key words: polyamine analogs, breast cancer cells, apoptosis, molecular modelling.

Molecular modelling and chemical synthesis of a bisarginyl derivative of a tricationic porphyrin designed to target the major groove of DNA

The nature of water on the surface of a macromolecule is reflected in the temperature dependence of the heat effect, i.e., the heat capacity change, ΔCp, that accompanies its removal on forming a complex. The relationship between ΔCp and the nature of the surface dehydrated cannot be modeled for DNA by the use of small molecules, as previously done for proteins, since the contiguous surfaces of the grooves cannot be treated as the sum of small component molecules such as nucleotides. An alternative approach is used here in which ΔCp is measured for the formation of several protein/DNA complexes and the calculated contribution from protein dehydration subtracted to yield the heat capacity change attributable to dehydration of the DNA. The polar and apolar surface areas of the DNA dehydrated on complex formation were calculated from the known structures of the complexes, allowing heat capacity coefficients to be derived representing dehydration of unit surface area of polar and apolar surface in both grooves. Dehydration of apolar surfaces in both grooves is essentially identical and accompanied by a reduction in ΔCp by about 3 J K−1 mol−1 (Å2)−1, a value of somewhat greater magnitude than observed for proteins {ΔCp = − 1.79 J K−1 mol−1 (Å2)−1}. In contrast, dehydration of polar surfaces is very different in the two grooves: in the minor groove ΔCp increases by 2.7 J K−1 mol−1 (Å2)−1, but in the major groove, although ΔCp is also positive, it is low in value: + 0.4 J K−1 mol−1 (Å2)−1. Physical explanations for the magnitudes of ΔCp are discussed.

The DNA repair protein O(6)-alkylguanine alkyltransferase (AGT) is responsible for removing promutagenic alkyl lesions from exocyclic oxygens located in the major groove of DNA, i.e. the O(6) and O(4) positions of guanine and thymine. The protein carries out this repair reaction by transferring the alkyl group to an active site cysteine and in doing so directly repairs the premutagenic lesion in a reaction that inactivates the protein. In order to trap a covalent AGT-DNA complex, oligodeoxyribonucleotides containing the novel nucleoside N(1),O(6)-ethanoxanthosine ((e)X) have been prepared. The (e)X nucleoside was prepared by deamination of 3',5'-protected O(6)-hydroxyethyl-2'-deoxyguanosine followed by cyclization to produce 3',5'-protected N(1),O(6)-ethano-2'-deoxyxanthosine, which was converted to the nucleoside phosphoramidite and used in the preparation of oligodeoxyribonucleotides. Incubation of human AGT with a DNA duplex containing (e)X resulted in the formation of a covalent protein-DNA complex. Formation of this complex was dependent on both active human AGT and (e)X and could be prevented by chemical inactivation of the AGT with O(6)-benzylguanine. The crosslinking of AGT to DNA using (e)X occurs with high yield and the resulting complex appears to be well suited for further biochemical and biophysical characterization.

The indenoisoquinolines are a class of noncamptothecin topoisomerase I inhibitors that display significant cytotoxicity in human cancer cell cultures. They offer a number of potential advantages over the camptothecins, including greater chemical stability, formation of more persistent cleavage complexes, and induction of a unique pattern of DNA cleavage sites. Molecular modeling has suggested that substituents on the indenoisoquinoline lactam nitrogen would protrude out of the DNA duplex in the ternary cleavage complex through the major groove. This indicates that relatively large substituents in that location would be tolerated without compromising biological activity. As a strategy for increasing the potencies and potential therapeutic usefulness of the indenoisoquinolines, a series of compounds was synthesized containing polyamine side chains on the lactam nitrogen. The rationale for the synthesis of these compounds was that the positively charged ammonium cations would increase DNA affinity through electrostatic binding to the negatively charged DNA backbone, and the polyamines might also facilitate cellular uptake by utilization of polyamine transporters. The key step in the synthesis involved the condensation of Schiff bases, containing protected amine side chains, with substituted homophthalic anhydrides, to afford cis-3-aryl-4-carboxy-1-isoquinolones. These isoquinolones were then converted to indenoisoquinolines with thionyl chloride. Although monoamines were much more potent than the lead compound, no significant increase in potency was observed through incorporation of additional amino groups in the side chain. However, one of the monoamine analogues, which features a bis(2-hydroxyethyl)amino group in the side chain, proved to be one of the most cytotoxic indenoisoquinoline synthesized to date, with a GI50 mean-graph midpoint (MGM) of 0.07 microM in the NIH human cancer cell culture screen, and topoisomerase I inhibitory activity comparable to that of camptothecin.

An extensive conformational analysis has been carried out for two diastereoisomeric pairs of model estrogen quinone-derived DNA adducts, N6-(2-hydroxyestron-6(alpha,beta)-yl)-2'-deoxyadenosine (2-OHE1-6(alpha,beta)-N6-dA) and N2-(2-hydroxyestron-6(alpha,beta)-yl)-2'-deoxyguanosine (2-OHE1-6(alpha,beta)-N2-dG), in a B-DNA duplex and at a primer-template junction in a pol alpha family DNA polymerase. In vitro primer extension studies in pol alpha [Terashima, I., et al. (1998) Biochemistry 37, 13807-13815] have shown that the adenine adducts can incorporate dT, together with a small proportion of the incorrect base dC opposite the lesion, and they block less strongly than the guanine adducts. We have carried out conformational searches with energy minimization for four DNA duplexes containing 2-OHE1-6alpha-N6-dA, 2-OHE1-6beta-N6-dA, 2-OHE1-6alpha-N2-dG, or 2-OHE1-6beta-N2-dG. Our searches revealed that the four-ring nonplanar 2-hydroxyestrone (2-OHE1) moiety strongly prefers to reside in the major groove of the adenine adducts or the minor groove of the guanine adducts in a B-DNA duplex, with stereochemistry-dependent orientational differences in each case. No low energy conformations involving intercalation of the 2-OHE1 moiety were located in the searches. This stems from the largely nonplanar, nonaromatic nature of the 2-OHE1 ring system and implies that the proclivity for such bulky, nonplanar adducts to reside at the DNA helix exterior is a plausible conformational feature of other structurally similar estrogen quinone-derived DNA adducts, independent of base sequence context. In addition, the adenine adduct isomers, located in the major groove, manifest serious disturbance to the Watson-Crick base pairs at and near the lesion site, suggesting repair susceptibility. Possible structures of these adducts in a pol alpha family polymerase were also investigated through molecular modeling. The results rationalized the experimental in vitro primer extension studies. In addition, poor accommodation of the beta-stereoisomers within the polymerase was noted, suggesting that these stereoisomers would be more prone to cause blockage. Stereochemistry-dependent differences in adduct orientation could be expected to produce different biochemical effects, as has been observed in adducts derived from polycyclic aromatic hydrocarbons.