Use of drug-specific antibodies to identify ethidium adducts produced in Trypanosoma brucei by photoaffinity labeling.

A photoreactive azido analog of the trypanocide ethidium bromide, 3-amino-8-azido-5-ethyl-6-phenylphenanthridinium chloride, attached covalently to calf thymus DNA (CT DNA) by photoaffinity labeling, was used to generate antibodies for the drug analog. The specificity of the antiserum was tested using enzyme-linked immunoadsorbant assays (ELISA) against immobilized antigen (photoaffinity labeled DNA) and by both the avidin-biotin peroxidase reaction and indirect immunofluorescence performed on smears of drug treated trypanosomes. The reaction of the antiserum with the covalently bound drug adduct was diminished effectively by prior incubation with an excess of ethidium monoazide, ethidium diazide, and ethidium bromide, and to a lesser extent by the DNA-ethidium complex, the diazide-DNA or RNA adduct, and the monoazide-RNA adduct. DNA which had been photoaffinity labeled with either the propidium or the acridine moiety did not react. The antiserum recognition of DNA photoaffinity labeled with ethidium monoazide was based on the substituted phenanthridinium ring system of the parent ethidium, as evidenced by competition binding studies involving the free monoazido analog (EA1), the diazido analog (EA2), and the parent compound, ethidium bromide (EB). This approach and the sensitivity it provides should prove useful for identifying the distribution and fate of covalently bound drugs resulting from antiparasitic drug treatment, and for studying their roles in antiparasitic action.

Photodynamic mutagenic action of acridine compounds on yeast Saccharomyces cerevisiae.

The photodynamically produced mutagenicity and toxicity of 8 acridine compounds were compared in Saccharomyces cerevisiae under resting and growing conditions. Without irradiation none of the acridines induced respiratory-deficient ('petite') colonies, indicative of mitochondrial DNA damage, in resting cells; and only acriflavine and proflavine induced 'petites' in growing cells. Also, without irradiation none of the acridines were significantly toxic or mutagenic for nuclear DNA under resting or growing conditions. However, with irradiation, acriflavine, proflavine, acridine yellow and rivanol became effective 'petite'-inducing mutagens and highly toxic for resting cells, while acriflavine, proflavine, and acridine orange became effective nuclear mutagens for resting cells. Acridine, quinacrine and 9-aminoacridine were not at all biologically effective with irradiation for resting cells. The results presented here indicate that singlet oxygen is generated by a photodynamic mechanism when acriflavine is irradiated, and further, that acridine, quinacrine and 9-aminoacridine are ineffective photosensitizers, because they are incapable of generating singlet oxygen with irradiation.

Subcellular localization of photoreactive ethidium analogs in Trypanosoma brucei by fluorescence microscopy.

To identify the in vivo targets of the trypanocide, ethidium bromide, the fluorescent staining of T. brucei was examined for a series of ethidium analogs using fluorescence microscopy. Determination of the biological targets for most drugs is limited by the reversible nature of their interactions. To overcome this limitation, photoaffinity (azido) analogs of ethidium, which are capable of covalent attachment with photoactivation, were used to identify the ethidium binding sites within the parasites. Two of these compounds, when covalently attached, demonstrated an enhancement of fluorescent staining and were selective for the kinetoplast at low drug concentrations. These compounds were also those found previously to have the highest trypanocidal activity. Propidium, a phenanthridinium analog identical to ethidium except for a larger, more ionic substitution at R5, showed more nonspecific binding as determined by its general staining of the cytoplasm.

Ethidium binding sites on plasmid DNA determined by photoaffinity labeling.

Photoaffinity labeling of pBR322 with ethidium monoazide (8-azido-3-amino-5-ethyl-6-phenylphenanthridinium chloride) was used to provide evidence for the sequence specificity of ethidium binding to native DNA. DNA-drug interactions were examined at concentrations of eight covalently bound ethidium drugs per molecule of pBR322 (4363 base pairs). Restriction enzyme cutting was blocked by the covalent binding of a drug molecule at (or near) the enzyme recognition sequence. This phenomenon was observed with all restriction enzymes tested and was not limited to specific regions of the pBR322 molecule. Double-digestion experiments indicated that a drug molecule may bind 2 to 3 base pairs outside the recognition sequence and still block restriction enzyme digestion. Intact plasmid was treated with [3H]ethidium monoazide and digested with restriction enzymes. The amount of covalently-linked ethidium analog was quantitated for different restriction fragments and the G-C content of each fragment was determined from the DNA sequence. In approximately half of the fragments the drug appeared to preferentially bind at a G-C base pair. However, no preference for specific sequences such as 5'-C-G-3' was detected, as had been suggested by previous modeling studies with ethidium bromide. The other fragments were located in specific map regions of the plasmid and did not bind drug with a strict dependence on GC content suggesting that binding specificity may depend on more than one structural feature of the DNA.

Identification of an acridine photoaffinity probe for trypanocidal action.

Twenty-four acridine derivatives were screened for trypanocidal activity in Trypanosoma brucei in order to determine which structural features of the acridine molecule confer maximal antiparasitic activity. The synthesis of several new azidoacridine derivatives are also reported as well as an assessment of their value as possible photoaffinity probes for the study of acridine trypanocidal action. The most effective and selective acridine trypanocides, with and without irradiation, were the 3-amino-10-methylacridinium salt derivatives. With brief irradiation, one azidoacridine, 3-amino-6-azido-10-methylacridinium chloride, showed considerable trypanocidal activity at very limiting drug concentrations (10(-7)M) and warrants consideration as a possible photoaffinity probe.

Petite induction in Saccharomyces cerevisiae by ethidium analogs. Action on mitochondrial genome.

Petite induction of ethidium analogs was examined in both resting and growing yeast cells. All of the analogs used in these experiments were active in dividing cells of Saccharomyces cerevisiae; only the parent ethidium bromide was mutagenic under resting conditions. Incorporation of adenine into mitochondrial DNA appeared to be prevented completely by ethidium and partially inhibited by other analogs. Treatment of growing cells with analogs affected fragmentation of pre-existing DNA as seen by the loss of a mitochondrial antibiotic resistance marker. The rates of elimination of the marker were different; ethidium generated greater loss than the monoamino analogs (3-amino and 8-amino-); and the deaminated analog was least effective. However, in resting yeast the marker was partially eliminated only with treatment of the parent ethidium. The degradation of the mitochondrial DNA by exposure to ethidium compounds was confirmed by agarose gel electrophoresis. Electrophoretic patterns of the mitochondrial DNA treated with each of the analogs under growing conditions and only with ethidium under resting conditions showed degradation of the mitochondrial DNA.

Production of frameshift mutations in Salmonella by phenanthridinium derivatives: enzymatic activation and photoaffinity labeling.

The effect of metabolic activation on the mutagenic potential of some phenanthridinium compounds was examined in Salmonella typhimurium strains TA1538 and TA1978 . All of the compounds tested were mutagenic in TA1538, a DNA excision-repair-deficient strain, when metabolizing enzymes were included in the assay. Reversions were not detected when these compounds were examined under the same conditions in TA1978 , the isogenic strain of TA1538 proficient in DNA repair. The mutagenic activity of an azido analog of propidium iodide was also examined using photoactivation and enzymatic activation, and with both conditions, reversions were observed in TA1538 but not in TA1978 . Furthermore, the ranking of mutagenic activity of propidium azide relative to ethidium azide analogs was comparable for both types of activation. The evidence from several studies suggests that the structural requirements for mutagenic activity for this series of phenanthridinium compounds appear to be the same whether mutagenesis is induced via photoactivation or metabolic activation. The interaction with DNA resulting in covalent alteration of the DNA is implicated as the mutagenic mechanism whether the active species is generated by metabolic- or photo-activation.

Induction of cytoplasmically inherited respiration-deficient ('petite') mutants by photodynamic action of acridine compounds.

All acridines used (acriflavine, proflavine, acridine orange and 3-azido-10-methylacridinium chloride) produced killing in yeast cells when activated with visible light. Acriflavine, proflavine and 3-azido-10-methylacridinium chloride, but not acridine orange, produced petite and sectored colonies. Both cell killing and petite induction by light activation of acriflavine resulted apparently from photodynamic action mediated by singlet oxygen (1O2) since the effect were prevented by either sodium azide or anaerobiosis. The biological effects of 3-azido-10-methylacridinium chloride, which was developed as a potential photoaffinity probe for studying the binding and biological effects of acridines, appeared to be due to a photodynamic action analogous to that of acriflavine. Sodium azide or anaerobiosis prevented the light-activated effects of 3-azido-10-methylacridinium chloride despite the fact that the initial chemical breakdown of the azido derivative induced by light was not affected. Cells suspended in D2O demonstrated an enhanced response to 3-azido-10-methylacridinium chloride with irradiation. These results indicate that singlet oxygen mediates the light-activated biological effects of both acriflavine and 3-azido-10-methylacridinium chloride.

Structure-function characterization of phenanthridinium compounds as mutagens in Salmonella.

The mutagenic activity of some phenanthridinium compounds was examined by using Salmonella typhimurium strain TA98. Microsomal enzyme activation of the compounds was necessary for the detection of frameshift mutagenesis. Amino and/or azido functions at both R3 and R8 were structural requisites for significant mutagenic activity, although mutagenicity was severely reduced for the diazido analog. When an azido or amino group was substituted by hydrogen at either R3 or R8, mutagenic activities were minimal, and the deaminated compound (3,8-dihydro derivative) was not mutagenic. Propidium was only slightly more mutagenic than the monoamino and monoazido analogs. Relationships between mutagenic activity and chemical structure of phenanthridinium compounds are discussed.

Demonstration of specific high affinity binding sites in plasmid DNA by photoaffinity labeling with an ethidium analog.

We have used photoaffinity labeling of pBR322 DNA with 8-azido-3-amino-5-ethyl-6-phenylphenanthridinium chloride to demonstrate high affinity ethidium-binding sites. Plasmid equilibrated with as little as 1 drug/DNA molecule was photoactivated, freed of uncomplexed drug by ethanol precipitation, and subjected to restriction analysis. There was highly specific, rather than random, blockage of HhaI sites (d(GCGC)) at low drug concentrations. Furthermore, the same 7 new digestion fragments were generated at drug to nucleotide ratios ranging from 1:100 to 1:8000. All the new DNA fragments had chain lengths greater than the largest HhaI fragment (393 base pairs). At higher ligand concentrations closely approximating those needed for equilibrium binding studies, detection of the high affinity sites was greatly masked. Drug binding to HhaI restriction fragments which had been prepared prior to the action of drug did not induce new bands. Furthermore, the larger DNA fragments from drug-labeled plasmid were resistant to HhaI digestion over a wide range of enzyme concentrations. These findings suggest that ligand binding can be highly selective even between sites which have the same tetranucleotide sequence. Therefore, selective drug binding must be dictated not only by local base sequence preference, but also by other long range parameters.

Histochemical applications of two phenanthridinium compounds.

The fluorescent compounds ethidium monoazide and ethidium bromide were found to react intensely with nucleic acids of fixed, paraffin embedded tissues of rat and mouse. For routine staining, 10(-5) M solutions of ethidium bromide and its monoazide analogue were virtually identical in their reactions. Fresh frozen sections of the tissues reacted in the same manner as fixed, paraffin embedded samples. Fluorescence of DNA and RNA in rat pancreas could be selectively abolished by taking advantage of the greater sensitivity of RNA to acid hydrolysis. Hydrolysis in aqueous solutions (1 N HCl at 55-60 C) abolished RNA fluorescence in 5 min, whereas 20 min or longer were required to destroy DNA fluorescence. DNA fluorescence was selectively abolished by 3 hr in 0.1 N HCl in anhydrous methanol while the RNA remained unaffected. Rat pancreas stained with the 10(-5) M ethidium compounds below pH 5.0 showed reduced RNA fluorescence, but the DNA continued to fluoresce brightly at pH 0.6. Reducing the pH of the staining solution to pH 1.0, therefore, was an additional method of selectively abolishing RNA fluorescence. Ethidium solutions in 5.0 M NaCl at pH 5.0 had little effect on DNA or RNA fluorescence. This new method of examining nucleic acids in fixed tissue samples opens new approaches to the histochemistry of these substances. The method also offers new possibilities for the study of mutagenic drug-DNA interactions.

Antitrypanosomal action enhanced by photoaffinity labeling with ethidium analogs.

The trypanocidal activity of photoreactive azido analogs of ethidium was tested to determine the suitability of using such compounds as in vivo probes to study the mechanism of the antitrypanosomal activity of ethidium. Eight ethidium analogs, including three nonphotoreactive compounds, were tested for their ability to kill T. brucei both with and without photolytic activation. Two analogs tested, the monoamino-monoazido isomers, showed greater that 100-fold enhancement of trypanocidal activity following photolytic activation in situ. Without photolytic activation, only the nonphotoreactive monoamino precursor analogs showed activity greater than the parent ethidium compound. The availability of suitable ethidium analogs which can be covalently attached by in situ photoactivation provides a new approach for studying the mechanism by which ethidium exerts its trypanocidal activity.

Comparison of petite induction in yeast by acridines, ethidium and their photoaffinity probes.

The production of petite mutations by different acridine analogs was studied in Saccharomyces cerevisiae. Compounds with amino substituents at the 2 and 3 positions of the acridine nucleus and methylation at position 10 were effective for petite induction in growing cells but not in resting cells, while those with chloro, nitro and methoxy substituents were not effective in either resting or growing cells. Photosensitive azido derivatives of the acridines were tested to evaluate the role of covalent drug attachment for mutagenesis in resting cells. Photolysis of resting cells with 9-axido, 3-azido-6-amino-, 9-azido-10-methyl-, or 3-azido-6-amino-10-methyl-acridine was highly toxic. 3-Azido-6-amino-acridine, and especially 3-azido-10-methyl-, and 3-azido-6-amino-10-methyl-acridine, were effective petite inducers in resting cells. Thus, the photosensitive (azido) group at position 9 produced only cell killing while the azido group at position 3 and/or 6 led to effective petite induction in resting cells. In contrast, petite induction was observed only for growing cells, for dark control experiments with these compounds or with the monoazide precursor compounds.

Azido analogs of acridine: photoaffinity probes for frameshift mutagenesis in Salmonella typhimurium.

In order to identify a photoaffinity probe for 9-aminoacridine frameshift mutagenesis, 20 azido analogs of acridine were synthesized and tested in Ames' Salmonella tester strains, TA1535, TA1537, TA1538 and their corresponding excision-repair-coefficient strains TA1975, TA1977 and TA1978, to determine their mutagenicity and toxicity relative to 9-aminoacridine. The substituent-mutagenicity patterns observed for these compounds agree very well with those obtained previously for non-azidoacridines. The results presented here show that the 2-azido-analog of 9-aminoacridine demonstrates biological activity similar to 9-aminoacridine prior to photolytic activation. With light activation, however, the 9-amino-2-azido derivative becomes more effective at producing frameshift mutations characteristic of 9-aminoacridine. Furthermore, this photolytic enhancement of mutagenesis appears to be due to the repairable lesion suggesting that covalent attachment of the drug occurs.

Acridine structure correlated with mutagenic activity in Salmonella.

The structural basis for direct mutagenicity of acridines was studied by testing 50 different analogs in the Ames Salmonella tester strains without the addition of mammalian activating enzymes. These experiments showed that the single most effective substituent for frameshift mutagenesis in strain TA1537 is an amino group at the "9" position, while an amino group at either the "3" or "1" position is less effective. Other substitutions at the "9" position demonstrate decreased frameshifting activity compared to 9-aminoacridine. Furthermore, all substituents in combination with the amino group of 9-aminoacridine also decrease frameshifting activity, except for the addition of another amino group at the "1" position or a methyl at the ring nitrogen. Nitro substituents at the "1" and "3" positions enhance 9-aminoacridine toxocity. All nitro substituents decrease typical acridine-frameshift mutagenesis for strain TA1537, but they induce mutagenic activity either in the other type of frameshift strain, TA1538, or in the base-pair substitution strain TA1535. These studies have provided important structure-function relationships for acridine mutagenicity and toxicity in Salmonella. Consequently, this biological system has provided a sensitive means for determining the structural requirements for mutagenic mechanisms.