ABSTRACT
The antitumour antibiotics bleomycin and actinomycin are commonly used therapeutically in combination. One causes metal ion- and oxygen-dependent oxidative damage to DNA, while the other acts at the level of DNA via intercalation of its phenoxazone chromophore and probable inhibition of topoisomerases. Both drugs bind and/or cleave DNA primarily at guanine-containing sequences, which could lead to mutual interference. Using three different restriction fragments we show that binding of actinomycin to DNA causes major alterations in the sequence specificity of bleomycin-Fe-mediated cleavage, including the appearance of new cleavage sites and the suppression of others. The subtle sequence-dependence of the interference is illustrated by the different effects of actinomycin on DNA cleavage by the deglycobleomycin-Fe complex. Actinomycin sharply decreases the extent of cleavage at GpC sites by both bleomycin and deglycobleomycin whereas cleavage at GpT sites is much less affected, while novel cleavage sites are generated at GpA, ApT and, to a lesser extent, TpT steps. A dramatic increase in bleomycin-Fe cutting at GpA is barely detectable with deglycobleomycin-Fe, confirming that the carbohydrate moiety of bleomycin is important for DNA recognition. The results contribute to a better understanding of how two individually well-characterized small molecules interact simultaneously with specific sequences in DNA and as such assist clarification of the principles governing drug-DNA recognition.
Bleomycin and actinomycin (Fig. 1 ) are naturally occurring antitumour antibiotics produced by different strains of Streptomyces. After more than three decades of practice they have an established place in the clinic for the treatment of many forms of cancer. Actinomycin is used primarily to treat paediatric solid tumours including Wilms' tumour, Ewing's sarcoma and embryonal rhabdomyosarcoma, where it may be curative, whereas the principal therapeutic use of bleomycin is in the treatment of testicular carcinomas (1 ). They are also used together in combination chemotherapy for treating patients with gynaecologic cancers, disseminated melanoma and Kaposi's sarcoma (2 ,3 ). Regimens associating bleomycin and actinomycin plus an alkylating agent and/or a microtubule inhibitor have proved successful for the treatment of advanced germ-cell tumours (4 ,5 ). Combinations with vinblastine (VAB-6 regimen) are used routinely to treat malignant germ-cell tumours of the ovary and testicular cancers (6 ,7 ). Although both bleomycin and actinomycin are poorly selectively toxic (they are potent inhibitors of nucleic acid synthesis in practically all types of cells), they preferentially kill aerobic cell populations of solid tumours whereas many other DNA-binding drugs like daunomycin and mitomycin C are preferentially toxic to cells under hypoxic conditions (8 ).
It is well recognized that nucleic acids are the main target of both bleomycin and actinomycin. Binding of actinomycin to DNA involves insertion of the phenoxazone chromophore between consecutive base pairs coupled with fitting of its two bulky peptide rings into the minor groove of the double helix (9 ,10 ). Actinomycin displays a strong preference for intercalating at GpC-containing sites (11 -14 ) and exhibits the rare property of inhibiting both type I and type II topoisomerases (15 ,16 ). Bleomycin is a very potent cleaver of RNA as well as DNA in the presence of transition metal ions such as iron and copper. The degradation involves oxygenation of the deoxyribose moiety at C4'-H. The cutting reaction occurs in the minor groove of DNA predominantly at Gp
Ammonium persulphate, tris base, acrylamide, bis-acrylamide, ultrapure urea, boric acid, tetramethylethylenediamine and dimethyl sulphate were from BDH. Formic acid, piperidine, hydrazine and formamide were from Aldrich. Photographic requisites were from Kodak. Bromophenol blue and xylene cyanol were from Serva. All other chemicals were analytical grade reagents, and all solutions were prepared using multiply deionized, filtered water from a Milli-Q water purification system (Millipore). Restriction endonucleases EcoRI, AvaI and PvuII (Boehringer Mannheim, Germany) were used according to the supplier's recommended protocol in the activity buffer provided. Alkaline phosphatase, AMV reverse transcriptase and T4 polynucleotide kinase were from Pharmacia. [[alpha]-32P]dATP and [[gamma]-32P]ATP (6000 Ci/mmol) were purchased from New England Nuclear. Unlabelled ATP and dATP (ultrapure grade) were purchased from Pharmacia.
The initial stock solution was made by shaking 2 mg of actinomycin D (Sigma Chemical Co.) in 5 ml of water at 4oC for ~2 h and checking the concentration spectrophotometrically applying a molar extinction coefficient of 25 300 M-1 * cm-1 at 425 nm. A sample of pure bleomycin A2 was obtained from Roger-Bellon Laboratories (Neuilly-sur-Seine, France). The gulose- mannose moiety of bleomycin A2 was specifically cleaved by HF solvolysis according to a published protocol (20 ). Deglycobleomycin A2 was purified by HPLC and the integrity of the deglycosylated antibiotic was checked by 1H-NMR and HR-FAB mass spectrometry. Bleomycin A2 and deglycobleomycin were first dissolved in water to give a 1 mM stock solution and subsequent dilutions of the drugs were made with 10 mM Tris, 10 mM NaCl buffer, adjusted to pH 7.0. The initial dilutions (to 250 [mu]M) were apportioned into aliquots and stored frozen at -20oC.
Plasmids pUC12, pBS and pKMp27 were isolated from Escherichia coli by a standard sodium dodecyl sulphate-sodium hydroxide lysis procedure and purified by banding twice in CsCl-ethidium bromide gradients. Ethidium was removed by several isopropanol extractions followed by exhaustive dialysis against Tris-EDTA buffer. Purified plasmids were then precipitated and resuspended in appropriate buffer prior to digestion by the restriction enzymes. Three restriction fragments were used to investigate antibiotic-mediated cleavage of DNA: (i) a 178 bp EcoRI-PvuII fragment from pUC12, (ii) a 265 bp fragment cut out of plasmid pBS with EcoRI and PvuII, and (iii) the 160 bp tyrT(A93) fragment cut out of plasmid pKMp27. In each case the plasmid was double digested with EcoRI plus PvuII or AvaI for 3 h at 37oC then incubated with AMV reverse transcriptase in the presence of [[alpha]-32P]dATP to label specifically the 3'-end at the EcoRI site. The singly end-labelled DNA fragment was then purified by preparative electrophoresis on a non-denaturing 6.5% 1.5 mm polyacrylamide gel (200 V for 2 h in TBE buffer: 89 mM Tris base, 89 mM boric acid, 2.5 mM Na2 EDTA, pH 8.3). After a few seconds of autoradiography to locate the DNA, the band was excised from the gel, minced with a blade and extracted overnight in 500 mM ammonium acetate, 10 mM magnesium acetate. The purified DNA was then precipitated twice with 70% ethanol prior to resuspension in 10 mM Tris, 10 mM NaCl buffer (pH 7.0).
In a typical experiment, a solution of actinomycin was mixed with the 3'-radiolabelled DNA (~1 nM) in a final volume of 6 [mu]l in 10 mM Tris-HCl buffer pH 7.0 containing 10 mM NaCl and incubated for 1 h at 37oC. After equilibration, a freshly prepared bleomycin-FeII complex (4 [mu]l) was added to each tube. The bleomycin-FeII complex consisted of 2 [mu]l of a 5 [mu]M solution of bleomycin A2 mixed with 2 [mu]l of 5 [mu]M Fe(NH4)2(SO4)2[middot]6H2O just prior to the experiment. After incubation for 1 min at room temperature the cleavage reaction was stopped by freezing. Samples were lyophilised, resuspended in 50 [mu]l of water and lyophilised again. With deglycobleomycin, the digestion time was extended to 2 min to compensate for the slightly lower efficiency of DNA cleavage by the deglycobleomycin-FeII complex compared with the bleomycin-FeII complex. The cleavage products were resuspended in 4 [mu]l of formamide-dye solution and resolved on a denaturing polyacrylamide gel as described below.
DNA cleavage products were resolved by electrophoresis under denaturing conditions on a 0.3 mm 8% polyacrylamide gel containing 8 M urea in TBE buffer. After ~2.5 h at 60 W (1600 V; BRL sequencer model S2), gels were soaked in 10% acetic acid for 15 min, transferred to Whatman 3MM paper, dried under vacuum at 80oC, and examined by autoradiography using either a phosphorimager or X-ray films (Fuji R-X) exposed at -70oC with an intensifying screen usually for 24 h. For quantitative analysis, a Molecular Dynamics 425E PhosphorImager was used to collect data from storage screens exposed to the dried gels overnight at room temperature. Base line-corrected scans were analyzed by integrating all the densities between two selected boundaries using ImageQuant version 3.3 software. Each resolved band was assigned to a particular bond within the DNA fragment by comparison with its position relative to sequencing standards (G or G+A tracks).
Cleavage experiments with the FeII complexes of bleomycin and deglycobleomycin were performed using three different DNA fragments to provide an assessment of the effect of actinomycin with respect to a wide variety of potential binding/cutting sites. The 265mer and 178mer EcoRI-PvuII restriction fragments from the plasmids pBS and pUC12 respectively were examined together with the 160 bp fragment from E.coli containing the tyrT promoter sequence that has been used previously to probe the sequence-specific cleavage of DNA by bleomycin (21 ,22 ). All three fragments have been employed previously in footprinting experiments to determine the location of actinomycin binding sites (12 ,14 ,23 ,24 ). After pre-equilibration of each DNA fragment with actinomycin, the bleomycin or its deglycosylated analogue was allowed to react with the actinomycin-DNA complex. The sites of cleavage were visualised on a sequencing gel. Typical autoradiograms of a bleomycin cleavage assay on each fragment are shown in Figure 2 . It is immediately apparent that actinomycin modifies significantly the cutting by both bleomycin and deglycobleomycin. At some sites the cleavage intensity is markedly enhanced whereas at other sites the cleavage diminishes or is completely suppressed. At low concentrations of actinomycin (5-10 [mu]M) new cleavage sites can appear, such as at position 28 on the tyrT fragment and position 55 on the 265mer. Both the decreases and increases in cleavage rate are dose-dependent. Clear changes in cleavage intensity and/or location can be seen with equimolar concentrations of actinomycin and bleomycin (5 [mu]M). The changes become substantially more marked as the actinomycin concentration is raised to 50 [mu]M.
Before considering the effects of actinomycin, it must be recorded that the present results corroborate our previous findings that the gulose-mannose moiety of bleomycin A2 plays a significant role in the recognition of preferred nucleotide sequences (19 ). The data in Figure 4 confirm that, although each drug cleaves DNA primarily at Gp
The effects of actinomycin on bleomycin-induced cleavage can easily be explained. Since actinomycin exhibits a marked preference for binding to GpC-containing sites such as TGCA and CGCA (26 ) it is not surprising that it powerfully inhibits cleavage at GpC steps by bleomycin. According to the literature, the affinity of actinomycin for sites containing GpC must be at least one order of magnitude higher than that of bleomycin. Indeed, Bailey et al. (27 ) measured an affinity constant of 4.5 * 106 M-1 and 2.5 * 106 M-1 for actinomycin D bound to oligonucleotides containing -TGCA- and -CGCA- sequences respectively, whereas Chien et al. (28 ) found the affinity of bleomycin for DNA to be 1.2 * 105 M-1. The weak effect of actinomycin on bleomycin cleavage at GpT sites could also be anticipated on the basis of reports that actinomycin binds poorly to oligonucleotides devoid of GpC but containing GpT steps (29 ). However, although sequences surrounding GpT dinucleotides generally do not furnish strong binding sites for actinomycin, it has been observed that certain sequences such as CGTCGACG (29 ) as well as GTTTG and, to a lesser extent, GTTG (27 ) can bind the antibiotic with high affinity. These data are entirely compatible with the present results because we observe that weak cutting at the TGG
The actinomycin-induced enhancement of cleavage at GpA and ApT sites by bleomycin must surely result from structural distortions of the double helix due to intercalation of actinomycin molecules at preferred sites nearby (12 ). Crystallographic studies on actinomycin complexes provide compelling evidence of minor groove changes (9 ). It has previously been shown that DNA structure can influence sequence-specific cleavage by bleomycin (30 ). In fact the response of bleomycin to actinomycin binding is reminiscent of that of DNAase I which experiences a large increase in cutting efficiency at AT sequences adjacent to actinomycin-bound GpC sites (31 ). Recently, Patel and co-workers have shown that upon binding of actinomycin to the oligonucleotide (AAAGCTTT)2, the double helix develops a pronounced kink and is fully unwound at the central GpC site, thereby inducing an opening and widening of the minor groove at adjacent sequences (10 ). Such actinomycin-induced variation in the width of the minor groove could easily account for the observed alteration of bleomycin cutting since we have recently shown that the minor groove width appears to be an important determinant for sequence recognition by bleomycin (22 ). Another factor which may contribute to the modification of bleomycin specificity is DNA bending. Intercalation of actinomycin into DNA causes the double helix to bend towards the major groove (10 ) and the interaction of bleomycin with a bent DNA fragment (such as kinetoplast DNA) results in significant cleavage at GpA steps (32 ). Accordingly it is reasonable to conclude that the enhanced cleavage at GpA sites in the presence of actinomycin is a manifestation of actinomycin-induced bending of the helix. The fact that with deglycobleomycin the enhanced cleavage at GpA sites is not as pronounced as with bleomycin suggests that the drug lacking the gulose-mannose moiety may be less sensitive to DNA structural changes than the intact antibiotic. This again is consistent with the belief that the carbohydrate domain of bleomycin contributes to the recognition of DNA. In summary, the results indicate that actinomycin affects the cleavage of DNA by bleomycin both directly via masking some of the preferred cleavage sites (competition between the two antibiotics for GpC-containing recognition sites) and indirectly as a result of the structural consequences introduced by intercalation of the actinomycin chromophore into the helix.
It has previously been reported that DNA-binding drugs can interfere with phosphodiester backbone cleavage by bleomycin. Mono- and bis-intercalating drugs like ethidium bromide and echinomycin respectively, alter the degradative activity of bleomycin (33 ,34 ). The minor groove binder distamycin also inhibits bleomycin-induced cleavage at GpA and GpT sites. Both distamycin and ethidium reduce the cutting efficiency of the bleomycin-Fe complex but, unlike actinomycin, they do not provoke a redistribution of cleavage sites (33 ). On the other hand, platination of DNA with the widely-used antitumour agent cis-dichlorodiammineplatinumII causes major alterations in the sequence-specificity of bleomycin-Fe-mediated cleavage (35 ). Using a platinated DNA oligomer of defined structure, Gold et al. (36 ) demonstrated that platination generates a novel preferred sequence for bleomycin-Fe cleavage (decreased cleavage at a GpC site and increased cutting at a TpT step) as a result of platinum- induced distortion of the helical structure. Clearly our results are consistent with several studies suggesting that conformational alterations in DNA can lead to novel patterns of bleomycin- mediated DNA cleavage.
The marked effect of actinomycin in modifying DNA cleavage by bleomycin is reminiscent of a known effect of the anthracycline antibiotic daunomycin. Using a photosensitive analogue of actinomycin (the 7-azido derivative) it was recently shown that daunomycin alters the binding of actinomycin such that the antibiotic is displaced from its primary GpC sites on to secondary sites in the DNA (37 ). As in the present case, these findings raise practical implications since actinomycin and daunomycin have frequently been used in combination chemotherapy. The successful use of drug regimens containing bleomycin and actinomycin may be due, at least in part, to the modification of the range of DNA sequences cleaved by bleomycin under such conditions. There is evidently a need to extend investigations on drug-nucleic acid complexes to probe simultaneous interactions of two or more drugs with DNA. Evidence gleaned about synergistic interactions between antibiotics and DNA may be of profound benefit for formulating classes of antibiotics to be used in combination chemotherapy. Moreover, the rational design of new sequence-targeted compounds depends crucially on our understanding of the mechanisms by which active antibiotics bind to and recognise defined sequences in DNA. The results reported here contribute to a better understanding of how two individually well-characterized small molecules interact simultaneously with specific sequences, and as such assist clarification of the principles governing drug- DNA recognition.
The authors thank Dean Gentle for his expert technical assistance. This work was supported by grants (to M.J.W.) from the Cancer Research Campaign, the Wellcome Trust, the Association for International Cancer Research and the Sir Halley Stewart Trust; (to C.B.) from the Association pour la Recherche sur le Cancer (ARC 6932).
*To whom correspondence should be addressed. Tel: +33 320 16 92 18; Fax: +33 320 16 92 29; Email: bailly@lille.inserm.fr
REFERENCES
Return