Search

A detailed description of the interaction between Fe(II).bleomycin A2 and the Dickerson-Drew dodecamer d(CGCGAATTCGCG)2 is presented. The reaction between bleomycin and this substrate leads to DNA cleavage at two major sites, adenosine5 and cytidine11, and two minor sites, cytidine3 and thymidine8. The pattern and relative intensities of cleavage at these sites was not entirely consistent with what would be predicted based on the preference of the drug for cleavage at the pyrimidines of 5'-GC-3' and 5'-GT-3' sites. Insight into the origins of the apparent alteration of selectivity was provided by examination of the structure of the duplex which had been determined by X-ray crystallography. This indicated that the C4' hydrogens of the two nucleotides located at the strongest cleavage sites, C11 on one strand and A5 on the other, were oriented toward each other in the minor groove. Two-dimensional NMR measurements and molecular dynamics modeling indicated that a metalloBLM could bind to the duplex in an orientation that positioned the metal center roughly equally close to each of these hydrogen atoms. On the basis of this observation, it was proposed that these two residues represented a double-stranded BLM cleavage site. This hypothesis was tested through the study of the BLM-mediated cleavage of the related decamer duplex, d(CGCGAATTCG).d(CGAATTCGCG), as well as the hairpin sequence d(CGCGAATTCGIIIITTTTCCCCCGAATTCGCG). By the use of the hairpin oligonucleotide 32P-labeled alternately at the 5' and 3'-ends, unequivocal evidence was obtained for BLM-mediated double-strand cleavage. Quantitative analysis of the proportion of damage involving double-strand cleavage was effected by the use of the hairpin substrate; for damage initiated at the predominant cleavage site (cytidine31, analogous to cytidine11 in the dodecanucleotide), it is estimated that 43% of all damage leads to double-stranded lesions. The exceptional efficiency of double-strand cleavage observed in this system must reflect the spatial proximity and orientation of the two sugar H's whose abstraction is required to produce double-stranded lesions.

[structure: see text]. Tallysomycins are glycopeptide antibiotics that were first isolated from fermentation broths of Streptoalloteichus hindustanus. They are structurally related to the bleomycins but contain an additional talose sugar attached via a unique glycosylcarbinolamide linkage. Herein we report the synthesis of a key tallysomycin intermediate that incorporates the glycosylcarbinolamide moiety unique to the tallysomycins.

[structure: see text]. Several conformationally rigid analogues of the methylvalerate subunit contained within the linker domain of the antitumor antibiotic bleomycin have been prepared. These compounds have been protected in a fashion suitable for the solid-phase synthesis of bleomycin. Bleomycin congeners containing these analogues should facilitate a more detailed understanding of the nature of the conformational bend that the methylvalerate moiety is thought to impart to the natural product.

The syntheses of several novel halogenated bithiazoles structurally related to the bithiazole moiety of bleomycin A(5) are described. Also described is the ability of these compounds to mediate photoactivated DNA cleavage. Chlorinated bithiazole analogues were shown to be much more active than an analogous brominated derivative. DNA strand scission activity was strictly light dependent and was accompanied by dechlorination of the bithiazole nucleus, apparently in a stoichiometric fashion. Inhibition of DNA cleavage in the presence of DMSO, as well as photoaddition to 1-octene by both brominated and chlorinated bithiazole derivatives, suggest strongly that the initial step in photoactivated DNA cleavage involves homolysis of the thiazole carbon-halogen bond. The chlorinated bithiazoles were found to mediate sequence selective cleavage of a (32)P-end labeled DNA, although the selectivity observed was not the same as that of bleomycin itself. The implications of this observation are discussed.

The DNA cleavage properties of metallobleomycins conjugated to three solid supports were investigated using plasmid DNA, relaxed covalently closed circular DNA, and linear duplex DNA as substrates. Cleavage of pBR322 and pSP64 plasmid DNAs by Fe(II).BLM A(5)-CPG-C(2) was observed with efficiencies not dissimilar to that obtained using free Fe(II).BLM A(5). Similar results were observed following Fe(II).BLM A(5)-CPG-C(2)-mediated cleavage of a relaxed plasmid, a substrate that lacks ends or negative supercoiling capable of facilitating strand separation. BLMs covalently tethered to solid supports, including Fe(II).BLM A(5)-Sepharose 4B, Fe(II).BLM A(5)-CPG-C(6), and Fe(II).BLM A(5)-CPG-C(2), cleaved a 5'-(32)P end labeled linear DNA duplex with a sequence selectivity identical to that of free Fe(II).BLM A(5); cleavage predominated at 5'-G(82)T(83)-3' and 5'-G(84)T(85)-3'. To verify that these results could also be obtained using other metallobleomycins, supercoiled plasmid DNA and a linear DNA duplex were employed as substrates for Co(III).BLM A(5)-CPG-C(2). Free green Co(III).BLM A(5) was only about 2-fold more efficient than green Co(III).BLM A(5)-CPG-C(2) in effecting DNA cleavage. A similar result was obtained using Cu(II).BLM A(5)-CPG-C(2) + dithiothreitol. In addition, the conjugated Co.BLM A(5) and Cu.BLM A(5) cleaved the linear duplex DNA with a sequence selectivity identical to that of the respective free metalloBLMs. Interestingly, when supercoiled plasmid DNA was used as a substrate, conjugated Fe.BLM A(5) and Co.BLM A(5) were both found to produce Form III DNA in addition to Form II DNA. The formation of Form III DNA by conjugated Fe.BLM A(5) was assessed quantitatively. When corrected for differences in the intrinsic efficiencies of DNA cleavage by conjugated vs free BLMs, conjugated Fe.BLM A(5) was found to produce Form III DNA to about the same extent as the respective free Fe.BLM A(5), arguing that this conjugated BLM can also effect double-strand cleavage of DNA. Although previous evidence supporting DNA intercalation by some metallobleomycins is convincing, the present evidence indicates that threading intercalation is not a requirement for DNA cleavage by Fe(II).BLM A(5), Co(III).BLM A(5), or Cu(I).BLM A(5).

[structure in text] To explore the possibility of modifying bleomycin in a fashion that could alter its physiological distribution in a therapeutic setting, a new analogue of bleomycin has been prepared. This analogue is intended to target the asialoglycoprotein receptor on liver cells. Critically, despite the large C-substituent, the bleomycin conjugate was found to degrade DNA in the same fashion as bleomycin A(5) itself, and with only modestly decreased efficiency.

The solid-phase syntheses of two deglycobleomycin A(5) analogues were achieved using a commercially available polystyrene resin containing triphenylmethyl-linked spermidine. The final products were deblocked and released from the resin, analyzed, and purified by C(18) reversed phase HPLC and characterized by high-field (1)H NMR spectroscopy and mass spectrometry. The purified products relaxed supercoiled plasmid DNA in a concentration-dependent fashion and to the same extent as authentic material derived from natural BLM A(5).

Bleomycin was used as the reference compound in a COMPARE analysis to identify extracts in the National Cancer Institute's Natural Products Repository exhibiting cytotoxicity profiles similar to this antitumor agent. One of the extracts so identified was a CH(2)Cl(2)-methanol extract prepared from Gymnosporia trigyna, which effected relaxation of supercoiled pSP64 DNA in the presence of Cu(2+). Bioassay-guided fractionation using DNA strand-scission activity as an end point resulted in the isolation of four active principles. These were identified as syringaldehyde (1), (-)-syringaresinol (2), (+)-catechin (3), and (+)-epicatechin (4). Compounds 1 and 2 represent a new type of DNA strand-scission agent.

The bleomycin group antitumor antibiotics have long been of interest as a consequence of their efficacy in the treatment of certain tumors, not to mention their unique structures and properties in mediating dioxygen activation and sequence selective degradation of DNA. At a chemical level, the structure originally assigned to bleomycin was subsequently reassigned and the new structure has been confirmed by total synthesis. Through the elaboration of structurally modified bleomycin congeners and fragments, synthetic efforts have also facilitated an understanding of the contribution of individual structural domains in bleomycin to sequence selective DNA binding and cleavage, and have also provided insights into the nature of the chemical processes by which DNA degradation takes place. Within the last several years, it has also become apparent that bleomycin can mediate the oxidative degradation of all major classes of cellular RNAs; it seems entirely plausible that RNA may also represent an important locus of action for this class of antitumor agent. In parallel with ongoing synthetic and mechanistic efforts using classical methods, the study of bleomycins attached to solid supports has been shown to provide important mechanistic insights, and the actual elaboration of modified bleomycins by solid phase synthesis constitutes a logical extension of such efforts.

The chemistry of RNA degradation by Fe.bleomycin was studied using two RNA substrates that are modified efficiently at a small number of sites by the antitumor antibiotic. Cleavage of tRNAHis precursor transcript by Fe(II).BLM A2 was shown to require O2; cleavage was also observed when the same substrate was treated with Fe(III).BLM A2 + H2O2. Consistent with earlier observations made for DNA, the extent of tRNAHis precursor cleavage was greater for Fe(II).BLM A5 than for Fe(II).BLM A2; the least cleavage was obtained using Fe(II).BLM demethyl A2. By the use of 32P end labeled tRNAHis precursor transcript that was also 3H labeled within the uracil moieties, it was shown that release of uracil was nearly stoichiometric with tRNA strand scission by Fe(II).BLM A2. Nonetheless, treatment of the tRNAHis with hydrazine following BLM-mediated cleavage indicated formation of a new product that must have derived from a BLM-induced lesion. Also employed for characterization of BLM cleavage of RNA were the octanucleotides CGCTAGCG, C3-ribo-CGCTAGCG and C3-ara-CGCTAGCG. Analysis of the products of cleavage indicates that Fe.BLM is capable of mediating cleavage by abstraction of a H atom either from C-4' H or c-1' H of the chimeric oligonucleotides.

Two crystallographically defined tRNAs, yeast tRNAAsp and tRNAPhe, were used as substrates for oxidative cleavage by Fe.bleomycin to facilitate definition at high resolution of the structural elements in RNAs conducive to bleomycin binding and cleavage. Yeast tRNAAsp underwent cleavage at G45 and U66; yeast tRNAPhe was cleaved at four sites, namely G19, A31, U52 and A66. Only two of these six sites involved oxidative cleavage of a 5'-G.Pyr-3' sequence, but three sites were at the junction between single- and double-stranded regions of the RNA, consistent with a binding model in which the bithiazole + C-terminal substituent of bleomycin bind to minor groove structures on the RNA. Also studied were four tRNA transcripts believed on the basis of biochemical and chemical mapping experiments to share structural elements in common with the mature tRNAs. Cleavage of these tRNAs by Fe.bleomycin gave patterns of cleavage very different from each other and than those of the mature tRNAs. This observation suggests strongly that Fe.bleomycin cannot be used for chemical mapping in the same fashion as more classical reagents, such as Pb2+ or dimethyl sulfate. However, the great sensitivity of Fe.bleomycin to changes in nucleic acid structure argues that those species which do show similar patterns of cleavage must be very close in structure.

The specific cleavage of a DNA triple helix by FeII.bleomycin (BLM) is demonstrated. Triplex-specific cleavage was observed on both strands of the 32-base pair (bp) duplex at the duplex-triplex junctions. Strand scission products and alkali labile lesions were both formed. The strongest BLM cleavage site was located at the 5'-duplex-triplex junction, which is also the preferred triplex binding site of intercalating agents [Collier, D. A., Mergny, J.-L., Thuong, N. T., & Hélène, C. (1991) Nucleic Acids Res. 19, 4219-4224]. The preference of BLM for the 5'-junction does not appear to derive from selective intercalative binding at this site. This is supported by the observation that phleomycin, which contains a thiazolinylthiazole moiety rather than a planar bithiazole ring system, exhibited the same selectivity of triplex cleavage as BLM. Cleavage of the triple helix by FeII.BLM was unaffected by concentrations of Mg2+ up to 5 mM, suggesting possible therapeutic applications of this novel DNA target. Molecular-modeling calculations of the triplex region suggested that dramatic variations in minor groove width and depth occur at the duplex-triplex junctions, particularly at the 5'-junction. Moreover, the minor groove at these sites was calculated to be somewhat shallower and wider than the minor groove of B-DNA. These results suggest that the preference of BLM for the duplex-triplex junctions derives from selective recognition of minor groove shape at these sites and thus reflects conformation-selective, rather than sequence-selective, DNA recognition by FeII.BLM.

Bleomycin A2 (BLM) was found to mediate sequence specific hydrolysis of tRNA(Phe) in the absence of added metal ions. BLM A2 promoted phosphodiester bond hydrolysis 3' to the pyrimidine residue at all resolved Py-Pu sites not involving modified bases, as demonstrated by high-resolution electrophoretic analysis of 5'- and 3'-32P-end-labeled substrates. The reaction proceeded with surprising facility, approaching efficiency that of oxidative strand scission mediated by the FeII.BLM A2 complex. By the use of a number of BLM congeners, as well as a study of the time, temperature, and salt dependence of the hydrolysis, it was shown that in many respects the hydrolytic reaction parallels the oxidative degradation of RNA and DNA mediated by metallobleomycins. Thus, in contrast to the well-characterized oxidative degradation of DNA and RNA by bleomycin studied for two decades, the present report documents the ability of certain metal-free bleomycins to mediate RNA hydrolysis.

The effect of iron(II) bleomycin on a DNA-RNA heteroduplex was investigated using a substrate formed by reverse transcription of Escherichia coli 5S ribosomal RNA. Both strands of the heteroduplex were cleaved by FeII.BLM A2 at comparable concentrations; complete digestion of both strands was observed using 5 microM FeII.BLM A2. The DNA strand of the heteroduplex was cleaved predominantly at 5'-G-pyr-3' sites; the sites of cleavage of the DNA strand were a subset of those observed for the corresponding DNA strand of a DNA duplex of identical sequence. The sites of cleavage of the RNA strand of the heteroduplex involved both purines and pyrimidines and were found to be different than the sites of cleavage of the 5S rRNA alone, demonstrating that cleavage of the former must actually have involved heteroduplex recognition by FeII.BLM A2. Both the DNA and RNA strands of the heteroduplex were cleaved by FeII.BLM A2 in the presence of physiological concentrations of Mg2+, consistent with the possibility that DNA-RNA heteroduplexes may be therapeutically relevant targets for bleomycin.

A recent study of Fe(II).bleomycin-mediated DNA strand scission suggested that the metal binding domain of the drug is also the primary determinant of the observed sequence selectivity of strand scission (Carter, B. J., Murty, V. S., Reddy, K. S., Wang, S.-N., and Hecht, S. M. (1990) J. Biol. Chem. 265, 4193-4196). Although it is well established that the bithiazole moiety+C-terminal substituent of bleomycin are required for DNA binding, the role of the bithiazole in sequence-selective DNA recognition remains unclear. To determine whether the bithiazole moiety exhibits an intrinsic DNA binding selectivity, three synthetic EDTA-conjugated bithiazole derivatives were used to mediate DNA cleavage in the presence of Fe2+ and dithiothreitol. Incubation of these Fe(II).EDTA-bithiazoles in the presence of a 5'-32P end-labeled DNA duplex resulted in strand scission at every position to essentially the same extent. The relative cleavage efficiencies among the bithiazoles were a strong function of their ionic state. These findings imply that the bithiazoles can bind to many sites on the DNA; they support a model of bleomycin-DNA interaction in which the bithiazole moiety+C-terminal substituent are required only for DNA binding, whereas the metal binding domain is responsible for metal ion coordination and oxygen activation as well as being the primary determinant of sequence-selective DNA cleavage.

Previously, Fe.bleomycin (BLM) has been shown to mediate RNA cleavage in a fashion more highly selective than that of DNA. Because RNAs often assume secondary and tertiary structures not commonly encountered with DNAs, it was not clear whether the greater selectivity of RNA cleavage was a consequence of differences in the mononucleotide constituents of RNA and DNA, or of the three-dimensional structures of the individual substrates. Accordingly, we prepared a "tDNA" identical in sequence with Bacillus subtilis tRNA(His) precursor, the latter of which is known to be a good substrate for Fe(II).BLM A2 and which undergoes oxidative cleavage predominantly at U35. Remarkably, the tDNA underwent cleavage predominantly at T35. At higher concentrations of Fe(II).BLM A2, the tDNA was extensively degraded, while the tRNA(His) precursor was not. Competition experiments suggested that this was not due to more efficient binding of Fe.BLM to the tDNA; in fact the tRNA precursor appeared to be bound more efficiently. The lesser cleavage of the tRNA(His) may be due to limitations in the facility of chemical transformation following Fe.BLM binding, or else to the formation of RNA lesions that do not lead directly to RNA strand scission.

The ability of iron(II).bleomycin to mediate RNA degradation was further characterized. At micromolar concentrations, FeII.BLM was shown to effect cleavage of Escherichia coli tRNA(1His) and a Schizosaccharomyces pombe amber suppressor tRNA construct in an efficient fashion. In contrast, E. coli tRNA(Cys) and yeast mitochondrial tRNA(Asp) and tRNA(fMet) precursors were not substrates for FeII.BLM. Also shown to be a good substrate for cleavage by FeII.BLM was yeast 5S ribosomal RNA. Since HIV-1 reverse transcriptase mRNA has previously been shown to be degraded by Fe.BLM (Carter et al., 1990a), members of the three major classes of RNA have now been shown to undergo Fe.BLM-mediated strand scission. For each of the substrate RNAs, cleavage occurred at sites unique to that substrate. Although RNA cleavage occurred at numerous sequences, 5'-G-pyr-3' sites were prominent. Likewise, while cleavage was noted in regions anticipated to be double-stranded, as well as in single-stranded regions, a disproportionate number of cleavages were noted at the junction between single- and double-stranded regions. As found in earlier studies, RNA cleavage was much more selective than DNA cleavage. Further, when RNA cleavage was carried out in the presence of reagents such as Mg2+, spermidine, and NaCl, the selectivity of cleavage was further enhanced. The highly selective and efficient cleavage of a number of RNA molecules reinforces our earlier suggestion that RNA may constitute a therapeutically relevant target for bleomycin.

Bleomycin is an antitumor agent whose activity has long been thought to derive from its ability to degrade DNA. Recent findings suggest that cellular RNA may be a therapeutically relevant locus. At micromolar concentrations, Fe(II)-bleomycin readily cleaved a Bacillus subtilis tRNAHis precursor in a highly selective fashion, but Escherichia coli tRNA(Tyr) precursor was largely unaffected even under more forcing conditions. Other substrates included an RNA transcript encoding a large segment of the reverse transcriptase from human immunodeficiency virus 1. RNA cleavage was oxidative, approximately 10-fold more selective than DNA cleavage, and largely unaffected by nonsubstrate RNAs. RNA sequence analysis suggested recognition of RNA tertiary structure, rather than recognition of specific sequences; subsets of nucleotides at the junction of single- and double-stranded regions were especially susceptible to cleavage. The ready accessibility of cellular RNAs to xenobiotic agents, the high selectivity of bleomycin action on RNAs, and the paucity of mechanisms for RNA repair suggest that RNA may be a therapeutically relevant target for bleomycin.

Previous studies of Fe-bleomycin-mediated DNA cleavage have established that the bithiazole moiety + C-terminal substituent of bleomycin are required for DNA binding, while the metal binding domain is responsible for O2 activation. Although recent studies have indicated that the metal binding domain also participates in DNA unwinding, and in determining the sequence and strand selectivity of DNA cleavage, no study has defined the structural domain that bears primary responsibility for the observed pattern of bleomycin-mediated DNA degradation. Presently, by the use of four synthetic analogs of bleomycin demethyl A2 having the functional domains connected by rigid spacers of varying lengths, the source of DNA cleavage specificity has been determined. When the four analogs cleaved 242- and 127-base pair 5'-32P-end-labeled DNA restriction fragments containing isolated Fe-bleomycin cleavage sites, all four produced cleavage at the same preferred sites. Because the (oligo)glycine spacers altered the distance between the domains by as much as 14 A, the identical cleavage patterns argue that the primary determinant of sequence specificity for these analogs is the metal binding domain.

Bleomycin is hypothesized to cause cell growth inhibition and cell death via DNA cleavage. We have attempted to determine if net DNA cleavage is directly related to growth inhibition by measuring whether both parameters vary in parallel. Of primary importance to these studies was use of several bleomycin congeners. We have shown that these congeners vary in their abilities both to inhibit cell growth and to cause DNA damage. Bleomycin B2, tallysomycin, and phleomycin were the most potent growth inhibitors, and bleomycin B2 caused the most DNA damage. N-Acetylbleomycin A2 was inactive in both assays. The net amount of DNA damage measured at two levels of growth inhibition was compared for each congener and was found to vary widely among the congeners. Similarly, the degree of growth inhibition at a given level of submaximal DNA damage was found to vary widely when individual congeners were compared to each other. Hence, growth inhibition and net DNA damage due to bleomycin are not directly correlated with each other when individual congeners are compared to each other.

The role of the bithiazole moiety of bleomycin in the interaction of the antibiotic with DNA has been studied by the use of synthetic bithiazole derivatives. The DNA affinity of individual C-terminal (bithiazole) analogues of bleomycin was measured in terms of the ability of these species to block the binding of bleomycin to DNA, as judged by diminution of the DNA degradation that attends bleomycin binding. DNA degradation was monitored both by release of [3H]thymine from radiolabeled PM-2 DNA and by alteration of bleomycin-treated DNA oligomers of defined sequence derived from Escherichia coli plasmid pLJ3. It was found that the affinity of the bithiazole derivatives for DNA depended on the presence of the bithiazole moiety itself but more importantly on the number and spacing of positively charged groups; 2'-(2-aminoethyl)-2,4'-bithiazole-4-[3-[(4-aminobutyl) amino]propyl]carboxamide (14), having three positively charged groups at neutral pH, was a reasonably effective inhibitor of DNA degradation by bleomycin. Consistent with the importance of the spacing of the positively charged groups, tetrapeptide S (12) was found to be significantly less inhibitory toward DNA degradation by bleomycin than tripeptide S, in spite of their equal number of positively charged groups and the greater structural similarity of the former to bleomycin A2. Bleomycin is known to cleave DNA perferentially at certain sequences. It was shown that the inhibitors employed in this study diminished DNA cleavage proportionately at each cleavage site; no alteration was observed in the specificity of cleavage. A number of the bithiazole analogues employed as inhibitors of bleomycin-mediated DNA degradation were also utilized in fluorescence quenching experiments with calf thymus DNA. Consistent with the belief that these species inhibit bleomycin degradation by competitive binding to the DNA substrate, the best inhibitors exhibited the greatest fluorescence quenching upon admixture of DNA.

In the presence of NADPH and O2, NADPH-cytochrome P-450 reductase was found to activate Fe(III)-bleomycin A2 for DNA strand scission. Consistent with observations made previously when cccDNA was incubated in the presence of bleomycin and Fe(II) + O2 or Fe(III) + C6H5IO, degradation of DNA by NADPH-cytochrome P-450 reductase activated Fe(III)-bleomycin A2 produced both single- and double-strand nicks with concomitant formation of malondialdehyde (precursors). Cu(II)-bleomycin A2 also produced nicks in SV40 DNA following activation with NADPH-cytochrome P-450 reductase, but these were not accompanied by the formation of malondialdehyde (precursors). These findings confirm the activity of copper bleomycin in DNA strand scission and indicate that it degrades DNA in a fashion that differs mechanistically from that of iron bleomycin. The present findings also-establish the most facile pathways for enzymatic activation of Fe(III)-bleomycin and Cu(II)-bleomycin, provide data concerning the nature of the activated metallobleomycins, and extend the analogy between the chemistry of cytochrome P-450 and bleomycin.

DNA strand scission by bleomycin in the presence of Cu and Fe was further characterized. It was found that DNA degradation occurred readily upon admixture of Cu(I) or Cu(II) + dithiothreitol + bleomycin, but only where the order of addition precluded initial formation of Cu(II)--bleomycin or where sufficient time was permitted for reduction of the formed Cu(II)--bleomycin to Cu(I)--bleomycin. DNA strand scission mediated by Cu + dithiothreitol + bleomycin was inhibited by the copper-selective agent bathocuproine when the experiment was carried out under conditions consistent with Cu chelation by bathocuproine on the time scale of the experiment. Remarkably, it was found that the extent of DNA degradation obtained with bleomycin in the presence of Fe and Cu was greater than that obtained with either metal ion alone. A comparison of the sequence selectivity of bleomycin in the presence of Cu and Fe using 32P-end-labeled DNA duplexes as substrates revealed significant differences in sites of DNA cleavage and in the extent of cleavage at sites shared in common. For deglycoblemycin and decarbamoylbleomycin, whose metal ligation is believed to differ from that of bleomycin itself, it was found that the relative extents of DNA cleavage in the presence of Cu were not in the same order as those obtained in the presence of Fe. The bleomycin-mediated oxygenation products derived from cis-stilbene were found to differ in type and amount in the presence of added Cu vs. added Fe. Interestingly, while product formation from cis-stilbene was decreased when excess Fe was added to a reaction mixture containing 1:1 Fe(III) and bleomycin, the extent of product formation was enhanced almost 4-fold in reactions that contained 5:1, as compared to 1:1, Cu and bleomycin. The results of these experiments are entirely consistent with the work of Sugiura [Sugiura, Y. (1979) Biochem. Biophys. Res. Commun. 90, 375-383], who first demonstrated the generation of reactive oxygen species upon admixture of O2 and Cu(I)--bleomycin.

Three DNA duplexes differing substantially in sequence were derived from pBR322 plasmid DNA and supercoiled SV40 DNA by digestion with appropriate restriction endonucleases. Following treatment with the restriction methylase HhaI (recognition sequence: GCGC) or HhaI and HpaII (CCGG), the unmethylated and methylated DNAs were compared as substrates for the antitumor agent bleomycin. Bleomycin-mediated strand scission was shown to diminish substantially at a number of sites in proximity to the methylated cytidine moieties, especially where multiple sites had been methylated within a DNA segment of limited size. Detailed analysis of the DNA substrates revealed that both strands of DNA within a methylated region became more refractory to cleavage by bleomycin and that the protective effect could extend as many as 14 base pairs in proximity to the 5-methylcytidine moieties. Among the methylated DNA segments that became more resistant to bleomycin cleavage was a HpaII site of SV40 DNA, methylation of which has previously been shown to diminish the synthesis of the major late viral capsid protein following microinjection into Xenopus laevis oocytes. Study of the cleavage reaction at varying salt levels suggested that diminished bleomycin strand scission may be due, at least in part, to local conformational changes of the DNA to Z form (or other non-B-form structures). The results are generally consistent with the hypothesis that one mechanism for the expression of selective therapeutic action by certain DNA damaging agents could involve the recognition of specific methylation patterns.

Copper(I)-bleomycin [Cu(I) X BLM] was characterized in detail by 13C and 1H NMR. Unequivocal chemical shift assignments for Cu(I) X BLM and Cu(I) X BLM X CO were made by two-dimensional 1H-13C correlated spectroscopy and by utilizing the observation that Cu(I) X BLM was in rapid equilibrium with Cu(I) and metal-free bleomycin, such that individual resonances in the spectra of BLM and Cu(I) X BLM could be correlated. The binding of Cu(I) by bleomycin involves the beta-aminoalaninamide and pyrimidinyl moieties, and possibly the imidazole, but not N alpha of beta-hydroxyhistidine. Although no DNA strand scission by Cu(II) X BLM could be demonstrated in the absence of dithiothreitol, in the presence of this reducing agent substantial degradation of [3H]DNA was observed, as was strand scission of cccDNA. DNA degradation by Cu(I) X BLM was shown not to depend on contaminating Fe(II) and not to result in the formation of thymine propenal; the probable reason(s) for the lack of observed DNA degradation in earlier studies employing Cu(II) X BLM and dithiothreitol was (were) also identified. DNA strand scission was also noted under anaerobic conditions when Cu(II) X BLM and iodosobenzene were employed. If it is assumed that the mechanism of DNA degradation in this case is the same as that under aerobic conditions (i.e., with Cu(I) X BLM + O2 in the presence of dithiothreitol), then Cu X BLM must be capable of functioning as a monooxygenase in its degradation of DNA.

The sites of cleavage of DNA by bleomycin A2, bleomycin B2, phleomycin, tallysomycin A, and Blenoxane (Bristol-Meyers) in reactions containing equimolar Fe2+ and atmospheric oxygen were analyzed by gel electrophoresis of 32P end labeled DNA fragments. Bleomycin A2 and bleomycin B2 reactions cleaved DNA at all sites with a frequency equal to that of Blenoxane. At high concentrations of bleomycin the site specificity of cleavage was unchanged. Bleomycin cleavage sites and phleomycin cleavage sites are a subset of sites cleaved in reactions containing tallysomycin A. The nature of 5' and 3' termini induced by bleomycin cleavage was investigated. Electrophoresis of bleomycin-induced fragments after alkaline phosphatase or polynucleotide kinase treatment indicated that 5' termini are phosphoryl groups but 3' termini are not simple phosphoryl groups. Analysis of bleomycin cleavage of single-stranded DNA substrate showed that cleavage occurs only in regions of potentially double-stranded looped-back sequences. Possible mechanisms for determination of bleomycin cleavage sequence specificity are discussed.

The bleomycins (BLMs) are a structurally related group of antitumor antibiotics used clinically for the treatment of certain malignancies. The mechanism of action of the BLM is believed to involve DNA strand scission, a process that requires O2 and an appropriate metal ion; the therapeutically relevant metal is probably iron or copper. DNA strand scission by activated Fe X BLM involves oxygenation C-4' of deoxyribose and leads to two sets of products. One set results from scission of the C-3'--C-4' bond of deoxyribose, with concomitant cleavage of the DNA chain. The other set of products consists of free bases and an alkali-labile lesion, the latter of which leads to DNA chain cleavage on subsequent treatment with base. The structures of all of these degradation products have now been established by direct comparison with authentic synthetic samples. Also studied was the activation of BLM with (mono)oxygen surrogates such as iodosobenzene. The chemistry of the activated BLM so formed was remarkably similar to that of activated cytochrome P-450 and structurally related metalloporphyrins, which suggests a mechanistic analogy between the two. Remarkably, both Fe X BLM and Cu X BLM were also shown to be activated by NADPH cytochrome P-450 reductase in a transformation that was dependent on metal ion, O2 and NADPH.

Cu(I) and Cu(II) form stable 1:1 complexes with bleomycin (BLM). The affinity of both metals for the drug is greater than that of Fe(II). Cu(I) . BLM A2 binds to calf thymus DNA with about the same affinity as Fe(II) . BLM, as judged by DNA-induced fluorescence quenching of the bithiazole moiety of BLM. Based on 1H NMR and potentiometric titration data, the Cu(I) complexes of BLM are shown to have geometries very different than those of other BLM . metal(II) complexes studied thus far. As Cu(I) . BLM is oxidation-reduction active, its geometry is of importance in defining the structural requirements for BLM activity.

Two sets of products are formed from DNA upon treatment with Fe(II).bleomycin + O2. One set, which is believed to derive from a C-4' hydroperoxy derivative of the DNA deoxyribose moiety, includes the four possible base propenals, as well as DNA oligomers having deoxynucleoside 3'-(phosphoro-2"-O-glycolates) at their 3'-termini. The other set of products consists of free bases and alkali-labile lesions, the latter of which had not previously been characterized structurally. By use of the self-complementary dodecanucleotide d(CGCTTTAAAGCG) having a site modified by Fe-bleomycin three nucleotides from the 5'-end, it has been possible to characterize the alkali-labile product as a C-4' hydroxyapurinic acid. When the bleomycin-treated dodecanucleotide was treated with agents that effected decomposition of the alkali-labile lesion, products of the form CpGpx were obtained, and these proved useful for structural characterization of the alkali-labile lesion. Treatment with alkali produced CpGpx, where x was 2,4-dihydroxycyclopentenone. Alternatively, treatment with hydrazine provided a pyridazine derivative, and aqueous alkylamines led to formation of CpGp itself. The structures of all dinucleotides produced from the alkali-labile lesion were verified by direct comparison with authentic synthetic samples.

Bleomycin-mediated DNA strand scission has been shown to be diminished at certain sequences in proximity to 5-methylcytidines. We have investigated the molecular basis of this observed diminution using selective bleomycin (BLM) modifications at the C-terminus. Of the four different bleomycin congeners investigated, only bleomycin A2 and bleomycin BAPP were substantially affected by cytidine methylation. We have also examined the effect of other DNA modifications on bleomycin-mediated strand scission. Methylation at the N6 position of adenosine resulted in diminution of DNA cleavage by all four bleomycin congeners. The presence of bulky 5-(glucosyloxy)methyl groups in the major groove of T4 DNA had little effect on the efficiency of DNA strand scission mediated by bleomycin A2 or B2, suggesting the absence of important steric interactions between Fe(II).BLM and DNA in the major groove. In contrast, DNA cleavage mediated by bleomycin congeners was very sensitive to a major DNA conformational change, the B----Z transition. Salt and MgCl2 titrations of the DNA copolymers poly(dG-dC).poly(dG-dC) and poly(dG-MedC).poly(dG-MedC) demonstrated that bleomycin A2 and B2 did not cleave Z-DNA efficiently. In addition, circular dichroism titrations of these copolymers revealed that both bleomycin congeners increased the cation concentration necessary to induce the B----Z transition, implying that bleomycin preferentially binds to and stabilizes B-form DNA. These results are consistent with a model in which cytidine methylation at appropriate sequences of DNA is sufficient to induce subtle conformational changes that render the helix unreceptive to cleavage by some bleomycin congeners.

The unwinding of plasmid DNA by bleomycin A2 (BLM A2) was investigated by use of two-dimensional gel electrophoresis. It was found that Cu2+ ions greatly facilitated the unwinding of topoisomers of plasmid DNA by BLM A2 at concentrations where cupric ions alone had no effect on DNA supercoiling. The concentration of BLM A2 required for observable unwinding was reduced at least 100-fold in the presence of equimolar Cu2+. A plot of [Cu2+] vs extent of DNA unwinding in the presence of 10(-4) M BLM A2 gave a curve consistent with the action of cupric ions on BLM in an allosteric fashion, possibly rearranging the drug into a conformation that facilitates DNA unwinding. The participation of the metal center in enhancing DNA unwinding via direct ionic interaction with one or more negatively charged groups on the DNA duplex also seems possible. Further analysis of the structural factors required for BLM-mediated DNA unwinding was carried out with Cu2+ + BLM demethyl A2, the latter of which differs from BLM A2 only in that it lacks a methyl group, and associated positive charge, at the C-terminus. Cu(II).BLM demethyl A2 was found to be much less effective than Cu(II).BLM A2 as a DNA unwinding agent, emphasizing the strong dependence of this process on the presence of positively charged groups within the BLM molecule. These findings constitute the first direct evidence that the metal center of BLM can participate in DNA interaction, as well as in the previously recognized role of oxygen binding and activation.

Certain properties of the bleomycin analogs deglycobleomycin A2 and decarbamoylbleomycin A2 have been characterized. In common with bleomycin A2, both deglycobleomycin A2 and decarbamoylbleomycin A2 were found to mediate DNA degradation in the presence of Fe(II) + O2. Both analogs were found to have essentially the same sequence selectivity for DNA strand scission as bleomycin A2 when a 5'-[23P]-end-labeled linear duplex DNA derived from SV40 DNA was employed as a substrate. Product analysis for the three analogs was carried out by assay for malondialdehyde (precursors) after digestion of calf thymus DNA, and also by hplc analysis of the digestion products formed from the dodecanucleotide d(CGCTTTAAAGCG). All three Fe(II) X bleomycin A2 analogs produced the same products, albeit not in the same relative amounts.

Growth inhibition and cell killing by bleomycin are believed to be related to the ability of this antibiotic to cleave chromosomal DNA. Because bleomycin has an intracellular site of action, its ability to cross biological membranes must be critical to its overall effectiveness as an antitumor agent. The local anesthetic dibucaine acts to enhance membrane fluidity; therefore, the reported ability of this local anesthetic to modulate bleomycin effects on KB cells was investigated. Cells were treated with various bleomycin congeners in the presence or absence of dibucaine for 24 h. Dibucaine enhanced the inhibition of cell growth mediated by bleomycin A2, demethylbleomycin A2, bleomycin B2, and isobleomycin A2. N-Acetylbleomycin A2 did not inhibit cell growth in the absence of dibucaine, but it was inhibitory in the presence of dibucaine. Cells treated simultaneously for analysis of DNA breakage on alkaline sucrose gradients revealed that breakage was also enhanced in the presence of dibucaine. The degree of enhancement varied with dose and bleomycin congener. N-Acetylbleomycin A2 did not induce DNA breakage in either the absence or the presence of dibucaine. While growth inhibition and net DNA breakage correlated reasonably well in the absence of dibucaine for each bleomycin analogue tested, proportionality was lost in the presence of dibucaine, and very little DNA breakage was present when growth inhibition was complete. These observations imply that, at least in the presence of dibucaine, bleomycin may mediate growth inhibition at some locus in addition to chromosomal DNA and, also, that a given net amount of bleomycin analogue induced DNA damage per se does not produce a specific degree of growth inhibition.

Both Fe(III) and Cu(II) complexes of bleomycin (BLM), but not N-acetyl BLM . Fe(III), mediated the transfer of oxygen from iodosobenzene to organic substrates. In analogy with results obtained using certain cytochrome P-450 analogs, cis-stilbene was converted cleanly to the respective oxide, while no more than traces of trans-stilbene oxide were formed from trans-stilbene under identical conditions. The possible relevance of these observations to the degradation of DNA by bleomycin was also studied. In both the presence and absence of O2, BLM . Cu(II) . C6H5IO effected DNA degradation, as judged by the release of [3H]thymine from radiolabeled Escherichia coli DNA. These findings provide a valuable new assay system for the study of bleomycin analogs and suggest the possibility that bleomycin may function as an "oxygen transferase" in its degradation of DNA in situ.

Proton NMR studies at 360 MHz establish the binary Fe(II)-bleomycin complex to be paramagnetic with a spectrum covering 70 ppm. Addition of carbon monoxide generates a stable, diamagnetic Fe(II)-bleomycin-CO complex that is a putative structural analog of the "active" Fe(II)-bleomycin-O2 complex. The following six groups have been determined to be coordinated to the Fe(II) ion from analysis of the highly resolved 1H NMR spectra of this complex: CO, the primary and secondary amine nitrogens of the beta-aminoalanine moiety, the carbamoyl moiety on the 3-position of mannose, the pyrimidine N-1, and the imidazole N-1. The Fe(II)-bleomycin-CO complex binds to DNA, as shown by fluorescence quenching experiments, but Fe(II)-bleomycin-CO does not mediate thymine release. These results necessitate a major revision in the current model for metal coordination to bleomycin.

By the use of DNA, copolymers of defined nucleotide composition, and a synthetic dodecanucleotide having putative bleomycin cleavage sites in proximity to the 5'- and 3'-termini, the products formed concomitant with DNA strand scission have been isolated and subjected to structural identification and quantitation via direct comparison with authentic synthetic samples. The products of DNA strand scission by Fe(II)-bleomycin include oligonucleotides having each of the four possible nucleoside 3'-(phosphoro-2''-O-glycolates) at their 3'-termini, as well as the four possible base propenals. At least for 3-(adenin-9'-yl)propenal and 3-(thymin-1'-yl)propenal, the products formed were exclusively of the trans configuration.

We have studied the DNA- and metal-binding properties of a bleomycin A2 derivative in which the alpha-amino group of the beta-aminoalanine moiety has been N-acetylated. The modified antibiotic has been shown to be without activity in mediating the in vitro release of [3H]thymine from Pm-2 DNA. Fluorescence experiments indicate that the degree of quenching by DNA of the bithiazole fluorescence is unaffected by N-acetylation of bleomycin. Furthermore, 1H NMR experiments demonstrate that N-acetylation does not alter the stoichiometry of metal binding. The Fe(II)--Ac-bleomycin A2 complex, however, has been found to be stable in the presence of both O2 and CO, and thus inactivation appears to be accounted for by the loss of the ability to bind and/or reduce O2. Comparison of the 1H NMR spectra of the Fe(II)--bleomycin and Fe(II)--Ac-bleomycin A2 complexes indicates that either a drastic reorganization of the ligands with respect to the central iron atom has occurred or that an altered spin state is stabilized. These experiments establish that the ability of bleomycin to cause DNA damage is sensitive to even minor structural alterations within the antibiotic.

In analogy with bleomycin, deglyco-bleomycin B2 has been found to form a stable, diamagnetic complex with Fe(II) and CO. Although the stoichiometry of this complex appeared to be the same as that formed with bleomycin, the geometry of the deglyco-bleomycin complex was fundamentally different, especially as regards orientation of the beta-aminoalanine moiety. In the presence of Fe(II) and O2, deglyco-bleomycin A2 and deglyco-bleomycin B2 were found to release [3H]thymine from radiolabeled PM-2 DNA; when employed at limiting concentrations, deglyco-bleomycin A2 and B2 gave about half as much [3H]thymine release as the respective bleomycins. In view of the spectral evidence (Burger, R. M., Horwitz, S. B., Peisach, J., and Wittenberg, J. B. (1979) J. Biol. Chem. 254, 12299-12302) that Fe(II) . bleomycin . CO has the same geometry as the complex formed by initial association of bleomycin, Fe(II), and O2, the accumulated data suggest strongly that all metal complexes of bleomycin (derivatives) capable of DNA degradation need not have the same geometry.