Molecular determinants for DNA minor groove recognition: design of a bis-guanidinium derivative of ethidium that is highly selective for AT-rich DNA sequences

The phenanthridinium dye ethidium bromide is a prototypical DNA intercalating agent. For decades, this anti-trypanosomal agent has been known to intercalate into nucleic acids, with little preference for particular sequences. Only polydA-polydT tracts are relatively refractory to ethidium intercalation. In an effort to tune the sequence selectivity of known DNA binding agents, we report here the synthesis and detailed characterization of the mode of binding to DNA of a novel ethidium derivative possessing two guanidinium groups at positions 3 and 8. This compound, DB950, binds to DNA much more tightly than ethidium and exhibits distinct DNA-dependent absorption and fluorescence properties. The study of the mode of binding to DNA by means of circular and electric linear dichroism revealed that, unlike ethidium, DB950 forms minor groove complexes with AT sequences. Accurate quantification of binding affinities by surface plasmon resonance using A(n)T(n) hairpin oligomer indicated that the interaction of DB950 is over 10-50 times stronger than that of ethidium and comparable to that of the known minor groove binder furamidine. DB950 interacts weakly with GC sites by intercalation. DNase I footprinting experiments performed with different DNA fragments established that DB950 presents a pronounced selectivity for AT-rich sites, identical with that of furamidine. The replacement of the amino groups of ethidium with guanidinium groups has resulted in a marked gain of both affinity and sequence selectivity. DB950 provides protection against DNase I cleavage at AT-containing sites which frequently correspond to regions of enhanced cleavage in the presence of ethidium. Although DB950 maintains a planar phenanthridinium chromophore, the compound no longer intercalates at AT sites. The guanidinium groups of DB950, just like the amidinium group of furamidine (DB75), are the critical determinants for recognition of AT binding sites in DNA. The chemical modulation of the ethidium exocyclic amines is a profitable option to tune the nucleic acid recognition properties of phenylphenanthridinium dyes.     

Two human colon tumor cell lines with similar nuclease sensitivities have different ethidium bromide binding characteristics

Chromatin from two human colon adenocarcinoma cell lines (HT-29 and LoVo) showed similar digestion kinetics when sensitivities to DNase I and micrococcal nuclease were examined. Chromatin conformations were probed by examining the binding of ethidium bromide. A Scatchard plot revealed that both chromatins bound the same amount of ethidium bromide per mole of DNA, but the DNA from LoVo cells was more accessible to the intercalator. The results indicate that differences in chromatin conformation are not necessarily accompanied by different nuclease sensitivities.     

Site-directed mutagenesis of the -10 region of the lacUV5 promoter. Introduction of dA4.dT4 tract suppresses open complex formation

Homopolymeric dAn.dTn sequences, where n is 4 or greater, have special properties leading to increased duplex stability and DNA bending. The lacUV5 promoter was used to examine the functional consequences of changing the -10 TATAAT consensus sequence to the sequence TAAAAT. The transversion mutation at the underlined site was accomplished with site-directed mutagenesis using translation termination as the selection procedure. For free DNA, structural differences at the 5' and 3' junction regions of the dA4.dT4 tract can be readily detected by DNase I digestion. However, site binding by Escherichia coli RNA polymerase appeared unaltered by the TAAAAT sequence since identical DNase I footprints were obtained for the lacUV5 and mutant promoters. Binding competition studies under different ionic strengths revealed a significant reduction in mutant promoter open complex formation relative to the lacUV5 promoter. Mutant promoter open complexes also dissociated faster and to a greater extent than the corresponding lacUV5 promoter open complexes when challenged with heparin or a combination of heparin and increased KCl concentration. Consequently, mutant promoter open complexes appear less stable than lacUV5 promoter open complexes.     

Effects of the antitumor antibiotic mithramycin on the structure of repetitive DNA regions adjacent to its GC-rich binding site.

Regions of An.Tn, (GA)n.(TC)n, and (GT)n.(AC)n have been cloned into the SmaI (CCC/GGG) site of plasmid pUC19. HindIII-EcoRI restriction fragments containing these inserts have been used as substrates for footprinting experiments using DNase I, DNase II, and micrococcal nuclease as probes. These present good mithramycin binding sites (GGG) flanking repetitive regions to which the drug does not bind. In each case, mithramycin footprints are observed at the CCC/GGG sites, which are not affected by the nature of the surrounding sequences. Some weaker binding is detected at TCGA and ACCA sites and at regions of alternating GA. No binding is found to regions of alternating GT. An.Tn inserts (n = 23 or 69) are normally resistant to cleavage by all these probes; in the presence of mithramycin, a dramatic increase in DNase I cleavage is observed throughout the entire insert and is indicative of an alteration in DNA structure. Similar changes are seen with DNase II and micrococcal nuclease. These changes cannot be explained by invoking changes in the ratio of free substrate to cleavage agent. In contrast, cleavage of (GA)n.(CT)n and (GT)n.(AC)n inserts is not affected by drug binding. The results are consistent with a model in which mithramycin causes dramatic changes in the width of the DNA minor groove, generating a structure which has some properties of A-DNA, and suggest that this can be propagated into surrounding DNA regions in a sequence-dependent manner. The structural alterations with An.Tn are highly cooperative and can be transmitted over at least three turns of the DNA helix.     

Bifunctional intercalator [N-MeCys3,N-MeCys7]TANDEM binds to the dinucleotide TpA.

The binding of [N-MeCys3,N-MeCys7]TANDEM has been examined by DNase I footprinting and diethyl pyrocarbonate modification of several synthetic DNA fragments containing AT-rich regions. DNase I footprinting reveals that at low concentrations the ligand binds preferentially to the center of (AT)n regions. A fragment containing the tetranucleotide AATT was unaffected by the ligand. Diethyl pyrocarbonate modification of several fragments containing blocks of (AT)n revealed a pattern in which alternate adenines were rendered more reactive in the presence of the ligand. These reactive adenines were staggered across the two DNA strands in the 3'-direction, consistent with ligand binding to the dinucleotide TpA. In sequences of the type (TAA)n.(TTA)n, binding of [N-MeCys3,N-MeCys7]TANDEM resulted in strong modification of the second adenine in the sequence TAA, i.e., the base on the 3'-side of the ligand binding site. Data for binding to (AT)n are best explained by suggesting that the adenines sandwiched between the quinoxaline chromophores are rendered most reactive to diethyl pyrocarbonate.     

New dsDNA binding unnatural oligopeptides with pyrimidine selectivity

Solid phase peptide library screening followed by extension of a lead recognition element for binding to a dsDNA sequence (NF binding site of IL6) using solution phase screening, delivered a new DNA binding peptide, Ac-Arg-Ual-Sar-Chi-Chi-Tal-Arg-CONH2. In the present research, the contribution of the different amino acid side chains to the binding strength of the peptide to dsDNA was investigated using an ethidium bromide displacement test. Based on these results, the lead structure was optimized by deconvolution. Eight new unnatural amino acids were evaluated at two positions of the heptapeptide replacing the Ual-Sar fragment. The strongest dsDNA binding was observed using ([(3-chlorophenyl)methyl]amino)acetic acid (Cbg) and beta-cyclohexyl-l-alanine (Cha) respectively, at those two positions. A 10-fold increase in affinity compared to the Ual-Sar sequence was obtained. Further enhancement of dsDNA binding was obtained with hybrid molecules linking the newly developed peptide fragment to an acridine derivative with a flexible spacer. This resulted in ligands with affinities in the microM range for the dsDNA target (K(d) of 2.1 x 10(-6) M). DNase I footprinting with the newly developed oligopeptide motifs showed the presence of a pronounced pyrimidine specificity, while conjugation to an intercalator seems to redirect the interaction to mixed sequences. This way, new unnatural oligopeptide motifs and hybrid molecules have been developed endowed with different sequence selectivities. The results demonstrate that the unnatural peptide library approach combined with subsequent modification of selected amino acid positions, is very suited for the discovery of novel sequence-specific dsDNA binding ligands.     

The DNA-binding properties of polyomavirus large T antigen are altered by ATP and other nucleotides.

We have examined the influence of ATP on the DNA-binding properties of polyomavirus large T antigen (Py TAg). Utilizing nitrocellulose filter binding, DNase I footprinting, and gel mobility shift assays, we observed that ATP increased Py TAg binding to DNA fragments containing either all Py TAg-binding sites (whole origin) or those sites within (core origin) or adjacent to (early) the origin of replication. Even nonspecific binding to DNA fragments lacking Py TAg-binding sites was increased somewhat by ATP. Binding to the core origin was increased to a greater extent than binding to other DNA fragments tested. Gel band mobility shift assays revealed that ATP increased the production of core origin-specific Py TAg-DNA complexes of high molecular weight. ATP stimulation depended on the presence of MgCl2. Other nucleotides and nonhydrolyzable ATP analogs also increased Py TAg binding to the core origin but to various degrees: ATP, dATP, 5'-adenylyl imidodiphosphate (AMPPNP) greater than 5'-adenylyl methylenediphosphate (AMPPCP) greater than dCTP greater than UTP greater than TTP. GTP and dGTP did not increase DNA binding by Py TAg. The rates of association and disassociation of Py TAg with all the DNA fragments were altered by the presence of ATP. DNase I footprinting showed that ATP extensively extended the region protected within the core origin and also produced a distinctive DNase I-hypersensitive site on the late strand at nucleotides 5255 to 5262 (TTACTATG).     

Solution structural studies of the Ag(I)-DNA complex.

We report equilibrium dialysis and electric dichroism studies of the two strong complexes (I and II) of silver ion with DNA. Cooperative conversion of DNA to the stronger type I complex results in a 9% length decrease, and a structure in which intercalated ethidium is perpendicular to the helix axis. Upon addition of more Ag+ to form the type II complex, the DNA length reverts to its original value and bound ethidium once again becomes tilted from the plane perpendicular to the helix axis. In both type I and type II Ag (I) - DNA complexes, ethidium binding is mildly cooperative. We interpret the results in terms of a sequence of silver-induced cooperative switches of DNA from its B-form structure with propeller twisted base pairs to a structure with flat base pairs in the type I complex, and back again to propellered base pairs in the type II complex.     

Deoxyribonuclease II from the Icelandic scallop (Chlamys islandica): isolation and partial characterization

A deoxyribonuclease (DNase) was isolated from viscera of the cold-adapted marine bivalve Icelandic scallop. The 42 kDa DNase was shown to be a single polypeptide which catalyses DNA hydrolysis in the absence of divalent cations. The isolated enzyme showed maximal activity at pH 6 and no activity above pH 7.2 against native DNA. The scallop DNase was slightly more susceptible to heat denaturation than porcine DNase II and makes double-strand breaks in circular DNA substrate as the porcine enzyme. The N-terminal sequence of the scallop DNase was shown to be closely similar to DNase II (EC 3.1.22.1) proteins from other organisms. The scallop DNase is in addition to plancitoxin I from A. planci, the only DNase II enzyme isolated from marine invertebrates.     

Detection of deoxyribonuclease I and II activities in Japanese quail oocytes

Birds exhibit physiological polyspermy, i.e. numerous spermatozoa enter the germinal disc of an oocyte and form pronuclei during fertilisation. However, only one of them unites with the female pronucleus to form a zygote nucleus; the supernumerary spermatozoal nuclei degenerate at the early cleavage stages. To establish a factor responsible for spermatozoal degeneration, the presence of DNase activity was studied in vitro in extracts of Japanese quail oocytes using lambda DNA/HindIII as a substrate. The experimental conditions were designed to reveal the presence of either DNase I or DNase II activities, separately. Degradation of the substrate DNA was evaluated by electrophoresis on agarose gels stained with ethidium bromide. High activities of DNase I and DNase II were found in the germinal discs of the largest vitellogenic oocytes. DNase I activity was estimated to be about 3 x 10(-3) Kunitz units and DNase II about 4 x 10(-2) Kunitz units per germinal disc. DNase I activity in an oocyte seems to increase during oogenesis since DNA degradation by the extracts from the germinal discs of the largest vitellogenic oocytes was much higher than by those from previtellogenic and small vitellogenic oocytes. The presence of high DNase I and II activities in the largest vitellogenic oocytes would point to their role in degradation of DNA from supernumerary spermatozoa entering the ovum during polyspermic fertilisation in birds. The enzymes could be a factor, or one of the factors, in the late block to polyspermy in the cytoplasm of avian eggs. It is suggested here that the DNase activities might also be responsible for poor efficiency in obtaining transgenic birds by microinjection of exogenous DNA into the fertilised chick ovum.     

Mode of DNA binding by SopA protein of Escherichia coli F plasmid

The binding of SopA to the promoter region of its own gene, in which four copies of SopA's recognition sequence, 5'-CTTTGC-3', are arrayed asymmetrically, was examined in vitro. Titration using electrophoretic mobility shift assay showed that the stoichiometry of SopA protomers to the promoter-region DNA is 4 and that the binding is highly co-operative. The co-operativity was corroborated by EMSA and DNase I footprinting for a number of mutant DNA fragments in which 5'-CTTTGC-3' was changed to 5'-CTTACG-3'. EMSA in the style of circular permutation showed that SopA bends DNA. Mutation at either outermost binding site had a different effect on DNA bending by SopA, reflecting the asymmetry in the arrangement of the binding sites, for which the results of DNase I footprinting were in agreement. Gel filtration chromatography and analytical ultracentrifugation of free SopA showed that the protein can exist as a monomer and oligomers in the absence of ATP. Hence, the results indicate that the co-operativity in SopA's DNA binding is based on its intrinsic protein-protein interaction modified by DNA interaction.     

Nuclease sensitivity of estradiol-charged estrogen receptor binding sites in nuclei isolated from normal and neoplastic rat mammary tissues

The interaction of partially purified calf uterine estradiol-charged estrogen receptor ([3H]ER) with rat nuclei was studied in vitro. We previously observed a significantly greater number of [3H]ER binding sites (at saturation) in nuclei of R3230AC mammary tumors from intact vs ovariectomized (ovex) rats with no difference in the affinity of [3H]ER binding for these nuclei. We now report on the nuclease sensitivity of [3H]ER binding sites in nuclei from these tumors and from normal rat tissues. Digestion of tumor nuclei with deoxyribonuclease I (DNase I) prior to incubation with [3H]ER in vitro resulted in a progressive loss of [3H]ER binding capacity, which was not accompanied by alterations in the affinity of [3H]ER for the nuclei (Kd = 1-3 nM). A significantly lower concentration (P less than 0.005) of DNase I eliminated 50% of the [3H]ER binding sites in nuclei of tumors from intact hosts (8 unit.min/ml) compared to tumors from ovex hosts (22 unit.min/ml). These results indicate that DNA regions capable of binding ER are more susceptible to DNase I digestion in tumors from intact rats than those from ovex hosts, suggesting that the endogenous hormonal milieu is responsible, at least in part, for maintenance of nuclease-sensitive DNA conformations in this hormone-responsive mammary tumor. The amount of DNase I required to eliminate 50% of [3H]ER binding to nuclei from lactating mammary gland, liver, and kidney ranged from 14 to 56 unit.min/ml. Therefore, accessibility of [3H]ER binding sites to nuclease digestion in normal rat tissue is generally less than that of R3230AC tumors.     

Effects of alkylation by dimethyl sulfate, nitrogen mustard, and mitomycin C on DNA structure as studied by the ethidium binding assay.

The extent of alkylation of DNA by dimethyl sulfate, nitrogen mustard, and the antibiotic mitomycin C is related to the resulting decrease in the fluorescence of intercalated ethidium. The fluorescence losses due to the first two types of reagents show a marked pH dependence, with greater losses of fluorescence being observed at alkaline pH values. At pH 11.6 the fluorescence shows a slow recovery, so that with low levels of methylation (4% deoxyguanosine residues modified) one observes complete return of fluorescence. We postulate that these phenomena are due to conversion of 7-methyldeoxyguanosine to the zwitterionic form, and partial denaturation of the DNA duplex with loss of ethidium binding sites. Hydroxide-ion-catalyzed imidazole ring opening, and the removal of the positive charge permits reannealing with concomitant return of the ethidium intercalation sites. This conclusion is substantiated by enzymatic hydrolysis of 14C-labelled methylated DNA and identifiions of the ethidium assay. The distinctly different behavior of mitomycin C confirms previous conclusions that its alkylation, preferentially on guanine, does not take part at the N-7 position.     

Circular dichroism studies of ethidium bromide binding to the isolated nucleolus.

Circular dichroism in the 300-360 nm region and fluorescence induced by intercaltating binding of ethidum bromide to both DNA and RNA components were studied in isolated HeLa nucleoli. Both DNA and RNA compoents contribute to the induced dichroic elliticity. Digestion of nucleoli by RNase or DNase shows that most of the induced ellipticity comes from the DNA component. In nucleoli with an RNA/DNA = 0.8/1.0 the RNA component gives only 20% of the total ellipticity when measured at an ethidium bromide/DNA = 0.25. Spectro-fluorometric titration shows that ethidium bromide intercalates mostly into DNA in nucleoli. Both circular dichroism and fluorescence studies indicate that both DNA and RNA components in isolated nucleoli are less accessible to intercalating binding by ethidium bromide when compared to purified nucleolar DNA, DNA in chromatin or purified ribosomal RNA. Circular dichroic measurements of intercalating binding of ethidium bromide to to nucleoli may be used to study changes in nucleoli under different physiological or pathological conditions.     

Methidiumpropyl-EDTA.Fe(II) and DNase I footprinting report different small molecule binding site sizes on DNA.

DNase I and MPE.Fe (II) footprinting both employ partial cleavage of ligand-protected DNA restriction fragments and Maxam-Gilbert sequencing gel methods of analysis. One method utilizes the enzyme, DNase I, as the DNA cleaving agent while the other employs the synthetic molecule, methidium-propyl-EDTA (MPE). For actinomycin D, chromomycin A3 and distamycin A, DNase I footprinting reports larger binding site sizes than MPE.Fe (II). DNase I footprinting appears more sensitive for weakly bound sites. MPE.Fe (II) footprinting appears more accurate in determining the actual size and location of the binding sites for small molecules on DNA, especially in cases where several small molecules are closely spaced on the DNA. MPE.Fe (II) and DNase I report the same sequence and binding site size for lac repressor protein on operator DNA.     

High-resolution assessment of protein DNA binding affinity and selectivity utilizing a fluorescent intercalator displacement (FID) assay

Protein titration displacement of ethidium bromide bound to hairpin deoxyoligonucleotides containing any sequence of interest provides a well-defined titration curve (measuring the loss of fluorescence derived from the DNA bound ethidium bromide) that provides both absolute binding constants (K(a)) and stoichiometry of binding. This use of a fluorescent intercalator displacement (FID) assay for establishing protein DNA binding affinity and selectivity is demonstrated with the examination of the LEF-1 HMG domain binding to hairpin deoxyoligonucleotides containing its commonly accepted consensus sequence 5'-CTTTGWW (W=A or T) and those modified (5'-CTNTGWW) to examine sequences implicated in early studies (5'-CTNTG). The effectiveness of the FID assay coupled with its technically non-demanding experimental use makes it an attractive alternative or complement to selection screening, footprinting or affinity cleavage, and electrophoretic mobility shift assays for detecting, characterizing, and quantitating protein DNA binding affinity and selectivity.     

Translocation of the pre-synaptic complex formed upon DNA uptake by Streptococcus sanguis and its inhibition by ethidium bromide.

Summary Donor DNA in its initially bound, singlestranded form exists in a chromosomally-unassociated complex where it is resistant to exogenous DNase I but sensitive to micrococcal nuclease. Most of the complexes are readily recuperable from the supernatant of recipients converted into spheroplasts. Subsequent to formation of this superficially located complex, donor DNA progressively associates with the recipient chromosome into which it is eventually integrated. Treatment of recipients with ethidium bromide at various times after initial DNA binding almost immediately halts translocation of whatever donor material is not yet synapsed with the chromosome. On the other hand, donor DNA that has already synapsed experiences no difficulty in becoming genetically integrated. Some degradation occurs to DNA that fails to undergo translocation as a result of ethidium bromide treatment, the acid-soluble products appearing in the culture medium. DNA in untranslocated complexes surviving treatment is not appreciably different in single-strand length from that in untreated complexes. When these surviving complexes are isolated from a cell lysate, the contained DNA can be shown by spectrofluorometry to have bound the drug.