93 1
Biochemistry 1987, 26, 93 1-942
Copper-Dependent Cleavage of DNA by Bleomycid Guy M . Ehrenfeld,*-*Joshua B. Shipley,* David C. Heimbrook,s Hiroshi Sugiyama,$ Eric C . Long,$ Jacques H . van Boom," Gijs A. van der Marel," Norman J. Oppenheimer,I and Sidney M . Hecht*,tgg Departments of Chemistry and Biology, University of Virginia, Charlottesville, Virginia 22901, Chemical Research and Development, Smith Kline & French Laboratories, Swedeland, Pennsylvania 19479, Department of Organic Chemistry, University of Leiden, Leiden, The Netherlands, and Department of Pharmaceutical Chemistry, University of California, San Francisco, California 94143 Received July 8, 1986; Revised Manuscript Received September 24, 1986
Fe was further characterized. It was found that D N A degradation occurred readily upon admixture of Cu(1) or Cu(I1) dithiothreitol bleomycin, but only where the order of addition precluded initial formation of Cu(I1)-bleomycin or where sufficient time was permitted for reduction of the formed Cu(I1)-bleomycin to Cu(1)-bleomycin. D N A strand scission mediated by C u 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 D N A 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 C u and Fe using 32P-end-labeledD N A duplexes as substrates revealed significant differences in sites of D N A cleavage and in the extent of cleavage at sites shared in common. For deglycobleomycin and decarbamoylbleomycin, whose metal ligation is believed to differ from that of bleomycin itself, it was found that the relative extents of D N A cleavage in the presence of C u 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:l Fe(II1) and bleomycin, the extent of product formation was enhanced almost 4-fold in reactions that contained 5: 1, as compared to 1:1, C u and bleomycin. The results of these experiments are entirely consistent with the work of Sugiura [Sugiura, Y. (1979) Biochem. Biophys. Res. Commun. 90, 375-3831, who first demonstrated the generation of reactive oxygen species upon admixture of O2 and Cu(1)-bleomycin. ABSTRACT: D N A strand scission by bleomycin in the presence of C u and
+
+
+
+
B l e o m y c i n ( B L M ) ~is a glycopeptide-derived antineoplastic agent capable of effecting DNA strand scission (Hecht, 1979). In addition to a source of oxygen, bleomycin-mediated DNA degradation has been shown to require a metal cofactor (Sausville et al., 1978a,b). A description of the structure and behavior of metallobleomycin complexes is central to an understanding of the way in which bleomycin behaves as a DNA-interactive molecule and antineoplastic agent. While it has been shown that the iron (Sausville et al., 1978a), cobalt (Sugiura, 1980; Chang & Meares, 1982, 1984; Albertini & Garnier-Suillerot, 1982b), manganese (Ehrenfeld et al., 1984; Burger et al., 1984; Suzuki et al., 1985a), and vanadyl (Kuwahara et al., 1985) complexes of bleomycin are capable of mediating DNA degradation, considerable confusion has been evident regarding the properties of copper-bleomycin. A number of interesting observations have been made concerning Cu-BLM, and these provide the motivation for its careful characterization. These have included the findings that BLM bound Cu(I1) with a greater affinity than other physiologically relevant metals (Sugiura et al., 1979), that the administration of metal-free BLM was followed by Cu(I1) 'This work was supported at the University of Virginia by NIH Research Grants CA27603 and CA38544. *Address correspondence to this author at the Department of Chemistry, University of Virginia. 'University of Virginia. $Smith Kline & French Laboratories. University of Leiden. University of California.
0006-2960/87/0426-0931$01.50/0
binding in vivo (Kano et al., 1973), and that Cu(II).BLM has potentially useful cytotoxic properties (Ishizuka et al., 1967; Crooke & Bradner, 1977; Takahashi et al., 1977; Umezawa et al., 1968; Nunn & Lunec, 1978; Rao et al., 1980). A number of structures have been suggested for Cu(II).BLM on the basis of physicochemical studies (Iitaka et al., 1978; Takita et al., 1978; Bereman & Winkler, 1980; Solaiman et al., 1980; Dabrowiak, 1980; Antholine et al., 1984); the structure of Cu(I).BLM has been shown to differ significantly from that of other metallobleomycin complexes (Oppenheimer et al., 1981; Ehrenfeld et al., 1985). Cu(II).BLM can be reduced to Cu(I).BLM, although the facility of this process is unclear and probably dependent on the specific conditions employed (Takahashi et al., 1977; Antholine et al., 1982; Freedman et al., 1982; Kilkuskie et al., 1984). On the other hand, Cu(1)-BLM is oxidized readily to Cu(II).BLM (Oppenheimer et al., 1981), reportedly with the generation of reactive oxygen radicals (Sugiura, 1979). Since Cu-BLM is also known to bind to DNA (Povirk et al., 1981; Oppenheimer et al., 1981), it has seemed logical to study the ability of the putative oxygen radicals to degrade DNA in analogy with the transformation mediated by other metallobleomycins. Although initial investigation of CuqBLM in the presence of 2-mercaptoethanol suggested that the complex could not degrade phage DNA (Shirakawa et al., 1971) and that Cu(I1)
'
Abbreviations: BLM, bleomycin; DTT, dithiothreitol; EDTA, ethylenediaminetetracetic acid; Tris, tris(hydroxymethyl)aminomethane.
0 1987 American Chemical Society
932
BIOCHEMISTRY
could inhibit the degradation of DNA mediated by Fe(I1)BLM (Umezawa et al., 1970; Ishida & Takahashi, 1975; Sausville et al., 1978b), studies in these laboratories subsequently demonstrated Cu(II).BLM-mediated DNA strand scission in an C6H510-dependent reaction that required no added reducing agent (Murugesan et al., 1982; Ehrenfeld et al., 1985). Iodosobenzene-activated Cu(II).BLM was also shown to transfer oxygen to olefinic substrates, producing oxygenated substrates qualitatively and quantitatively different than those obtained by using Fe(III).BLM + C6H510(Murugesan & Hecht, 1985). Further studies suggested that CwBLM, in the presence of oxygen and an external reducing agent such as dithiothreitol or sodium dithionite, could also effect DNA strand scission (Ehrenfeld et al., 1985). In contrast to these reports, Suzuki et al. (1985b) have recently asserted that Cu-BLM has no significant DNA cleavage activity; the observed activity in earlier reports was attributed to adventitious Fe. In an effort to reconcile these apparent differences in experimental observations, and to explore further the factors that led to earlier reports of Cu. BLM inactivity, we have reinvestigated the behavior of Cu. BLM. Reported herein are conditions optimal for the activation of Cu(I).BLM for DNA degradation, the demonstration of bathocuproine sensitivity for the Cu-BLM-mediated process, and an analysis of the effects on DNA degradation noted for deferoxamine. Also described for the first time is the potentiation of DNA strand scission by bleomycin when both Cu and Fe were present, relative to that obtained with either metal ion alone. Characterization of DNA strand scission by CumBLM also included an analysis of qualitative and quantitative differences in strand selectivity of cleavage of 32Pend-labeled DNA duplexes, differences in the timing of metallobleomycin activation, and the effect of metal ligand alterations on the ability of single congeners of bleomycin to mediate DNA strand scission in the presence of Cu and Fe. Additionally, this report outlines clear differences in the nature and amounts of products formed from Cu-BLM and Fe-BLM and the intriguing finding that product formation was not maximal at a 1:l ratio of Cu and BLM. EXPERIMENTAL PROCEDURES Materials. Blenoxane was obtained from Bristol Laboratories through the courtesy of Dr. William Bradner; it was fractionated as described (Chien et al., 1977; Oppenheimer et al., 1979). Deglycobleomycin A2 was obtained by partial hydrolysis of bleomycin A, (Muraoka et al., 1981; Oppenheimer et al., 1982) or by synthesis (Aoyagi et al., 1982). Decarbamoylbleomycin A, was obtained by partial hydrolysis of bleomycin A, by modification of the published procedure (Naganawa et al., 1977); the structure of the product was verified by total synthesis. Iso-BLM A, was obtained by isomerization of BLM A,, as described (Nakayama et al., 1973). (See Chart I.) T4 polynucleotide kinase, DNA polymerase I (Klenow fragment), bacterial alkaline phosphatase, SV40 DNA, dithiothreitol, and agarose were obtained from Bethesda Research Laboratories. Deferoxamine mesylate was obtained from Ciba-Geigy Pharmaceuticals. Bathocuproinedisulfonate and calf thymus DNA were purchased from Sigma Chemical Co. EcoRI and HaeIII restriction endonucleases were purchased from New England Biolabs; TaqI and RsaI restriction endonucleases were from Boehringer Mannheim Biochemicals. Chelex- 100 resin was purchased from Bio-Rad. Copper(1) chloride (99.99% pure), Fe"(NH4)2(S04)2, and Cu( 11) chloride dihydrate were obtained from Aldrich Chemicals. The purity of these metals was determined by atomic absorption; contaminating Fe was 0.28% in Cu-
EHRENFELD ET AL.
Chart I
"
OH
&NH2
oqxNH2 bleomycinA2
H
o
L" ti
oqHs"z deglycobleowin A2
0
decarbarncylbleornycinA2
NH,
is0 bleomycin A?
CI2.2H2Oand 0.13% in CuCI. Bleomycin-Mediated Cleavage of DNA in the Presence of Cu(Zr) and Dithiothreitol. Reaction mixtures (50-pL total volume) contained 50 mM sodium cacodylate buffer, pH 7.0, 100 pM NaEDTA, 500 ng of SV40 DNA (>95% form I), 10 pM bleomycin B2, 40 pM DTT, 30 pM CuCl,, and, where present, either 100 MMdeferoxamine mesylate or 300 pM bathocuproinedisulfonate. All reaction solutions were prepared immediately prior to use with water that had been purified by passage through a Millipore-MilliQ apparatus. Further removal of adventitious metals was effected by batch treatment of solutions with portions of Chelex-100 resin (Ehrenfeld et
VOL. 26, NO. 3, 1987
CLEAVAGE O F D N A BY BLEOMYCIN
933
was carried out in 100 mM Tris-borate buffer, pH 8.4, conal., 1985). Each reaction was carried out for 10 min at room taining 100 mM EDTA at 1000 V for 3 h. Autoradiography temperature. The order of addition of reaction components was carried out at -80 OC with an intensifying screen (Kodak was as indicated in individual figure legends. Reactants were X-Omat AR film). added from 10-fold concentrated stock solutions; aliquots were added to the DNA-buffer solution or premixed at 10-s inDeglyco-BLM-Mediated Cleauage of 32P-End-Labeled tervals in the stated order prior to addition to the DNA-buffer DNA in the Presence of Fe(II) or Cu(II) + Dithiothreitol. solution. In those cases in which certain reactants were preReaction mixtures (30-pL total volume) contained 50 mM mixed separately, the reactions were initiated by combination sodium cacodylate, pH 7.0, sonicated calf thymus DNA (1 5 of the preformed solutions. Following each reaction 5 pL of pM DNA nucleotide concentration), 5’-32P-end-labeledDNA 75% glycerol was added to each reaction mixture, and the ( lo4 cpm), 20, 10, or 5 pM deglyco-BLM A2, and 1 equiv resulting solution was applied to a 1.2% agarose gel containing of Fe(I1) or Cu(I1) 1 mM DTT. The reaction mixtures were 1 pg/mL ethidium bromide. Horizontal gel electrophoresis initiated as described above for decarbamoyl-BLM A2, was carried out (80 V; 3 h) in 40 mM Tris-HC1 buffer, pH maintained at 25 “ C for 2 min, and analyzed for site-specific 7.8, containing 10 mM NaOAc and 5 mM NaEDTA. The cleavage as described above. gels were then visualized (UV light box) and photographed Time-Dependent Activation of Bleomycin with Fe(II) + (Kodak no. 9 filter; Polaroid type 55 P / N film). DTT or C u ( I 0 DTT. To reaction mixtures (30-pL total Preparation of DNA Restriction Fragments. SV40 DNA volume) containing 20 mM solution cacodylate, pH 7.0, 40 was digested with EcoRI and 3’-end-labeled with [ C Y - ~ ~ P I ~ A T PmM NaC1, 1 mM DTT, sonicated calf thymus DNA (1 5 pM by use of the Klenow fragment of DNA polymerase I (Sanger DNA nucleotide concentration), and 5’-32P-end-labeledDNA & Coulson, 1975). The 3’-end-labeled DNA was then sub(1.25 X lo4 cpm) was added Cu(II).BLM B, or Fe(II).BLM jected to digestion with HaeIII. The resulting solution was Bz such that the final metal ion concentration was 10 pM Cu subjected to electrophoresis on a 5% nondenaturing polyor 5 pM Fe. acrylamide gel. Two bands (275 and 477 base pairs, reThe metallobleomycins were preformed by admixture of spectively) were isolated by electroelution. Fe(NH4)2(S04)zor CuC12 with 1.2 equiv of bleomycin B2. The 5’-end-labeled DNA employed was prepared by diThe reactions were then initiated by admixture of these pregesting SV40 form I DNA with TuqI, followed by treatment formed metallobleomycins to the reaction mixture. The rewith bacterial alkaline phosphatase. The resulting DNA was action mixture was incubated at 25 OC, and aliquots were labeled with [y-3ZP]ATPas described (Maxam & Gilbert, removed at predetermined time intervals, quenched by ethanol 1980) and then treated with RsaI to yield 136- and 570precipitation of the DNA, and analyzed for site-specific DNA base-pair fragments, which were isolated from a 5% nondecleavage on polyacrylamide gels as described above. Control naturing polyacrylamide gel by electroelution. reactions were performed in the absence of Fe(I1) or Cu(I1); the ability of the ethanol precipitation to quench bleomycinA second set of restriction fragments was prepared from mediated DNA cleavage was also verified. SV40 form I DNA by digestion with restriction endonuclease BclI, followed by dephosphorylation of the restricted DNA Cleavage of d(CGCTTTAAAGCG)by Metallobleomycins. with calf intestine alkaline phosphatase (Maniatis et al., 1982). Reaction mixtures (50-pL total volume) contained 50 mM The resulting linear duplex DNA was 5’-3zP-end-labeledas sodium cacodylate, pH 7.0, d(CGCTTTAAAGCG) (200 pM described above and then treated with EcoRII, which resulted final nucleotide concentration), 10-20 pM BLM A2, Fe(I1) in the generation of a 127-base-pair fragment and a 242or Cu(I1) at the indicated concentrations, 1 mM DTT, and base-pair fragment. These DNA fragments were then purified deferoxamine mesylate where present. Each reaction mixture by polyacrylamide electrophoresis on a 5% gel and isolated was initiated by addition of a solution containing Fe(I1) or as described (Maxam & Gilbert, 1980). Cu(II), DTT, and deferoxamine as indicated; this solution was preincubated at 37 “ C for 5 min prior to admixture to the Decarbamoyl-BLM and Iso-BLM-Mediated Cleavage of buffered solution containing d(CGCTTTAAAGCG). The 32P-End-LabeledDNA in the Presence of Fe(II) or Cu(IZ) + reaction mixture was incubated at 37 “ C for 15 min and then Dithiothreitol. Reaction mixtures (50-pL total volume) analyzed promptly by HPLC on a Rainin C I SMicrosorb (0.46 contained 50 mM sodium cacodylate, pH 7.0, sonicated calf X 10 cm) column. The samples were eluted with 0.2 M thymus DNA (50 pM DNA nucleotide concentration), 3’ammonium formate buffer at a flow rate of 1.5 mL/min 32P-end-labeledDNA (-lo4 cpm), 25, 10, or 5 pM decar(monitored by A254). The amounts of cytosine and transbamoyl-BLM A2 or iso-BLM A2, and 2 equiv (relative to 3-(cytosin- 1-y1)propenal formed were quantified as described decarbamoyl-BLM) of Fe(I1) or Cu(I1) 1 mM DTT. (Sugiyama et al., 1985a,b). Control reactions employed bleomycin at final concentrations of 10 and 5 pM. The reaction mixtures were demetalized as Analysis of DNA Cleavage Gels. The autoradiograms were described above. scanned with an LKB 2202 laser densitometer interfaced to a Hewlett-Packard 3390 integrator. The data are reported Metal ion-DTT complexes were formed first and then as fractional cleavage, normalized over the scanned segment, combined with solutions containing the BLM congener of with a correction factor subtracted to compensate for backinterest. This combined solution was then added to the ground absorbance of the film. buffered DNA-containing solution to initiate the reaction. Spectrophotometric Determination oj- Cu(I1)-Bleomycin Reactions were maintained at 25 OC for 30 min and then in the Presence of Bathocuproine. To a I-mL aerobic solution treated successively with 5 pL of 3.5 M NaOAc and 2 volumes containing 50 mM sodium cacodylate buffer, pH 7.0, 80 pM of cold ethanol. The precipitated DNA was isolated by cenCuC12,and 100 pM bleomycin A, was added 20 pL of 2 mM trifugation, washed with cold 80% ethanol, and dried. The bathocuproine. The combined solution was maintained at resulting pellet was resuspended in a loading buffer consisting room temperature in a cuvette for 15 min, during which time of 100 mM Tris-borate, pH 8.4, and 50% formamide 0.15% the amount of Cu(1I)-bleomycin was quantified by the abbromophenol blue. The DNA was heat-denatured (90 OC, 2 sorption of this complex at 600 nm = 0.1 mM-’ cm-’) min) and applied to a (0.4 X 35 X 20 cm) 8% polyacrylamide (Freedman et al., 1982). gel (1 :20 cross-linked) containing 50% urea. Electrophoresis
-
+
+
+
+
EHRENFELD ET AL.
FIGURE 1: Effect of activation method on SV40 DNA strand scission by IO rM BLM B, in the presence of copper and dithiothreitol. Reactions were carried out as described under Experimental Procedures. The order of addition of each reagent IO the DNA solution is as indicated. and brackets indicate premixing of certain reactants prior to their (simultaneous)addition to the DNA solution: lane I. DTT, Cu(ll). B L M lane 2. DTT. BLM, Cu(ll); lane 3. BLM. Cu(ll), DW, lane 4. BLM. DTT, Cu(ll); lane 5. Cu(ll). BLM. DTT, lane 6. [Cu(ll). BLM]. D T . lane 7. [Cu(ll), BLM, DTT]: lane 8. (Cu(ll). DTT. BLMI; lane 9. [DTT. BLM]; lane IO. [Cu(ll). ELM]: lane 11. [Cu(ll). DTT]; lane 12, BLM; lane 13. DNA alone.
Specrrophoromerric Determination of Cu(l)-Balhocuproine in the Presence of Bleomycin. To a degassed solution ( I-mL total volume) containing 50 mM sodium cacodylate, pH 7.0, 80 p M CuCI. and 200 p M bathocuproine were added two degassed IO-pL aliquots of a degassed IO mM bleomycin A, solution; the addition of aliquots was separated by 3 min. Following equilibration a t r m m temperature in a Thunberg cuvette, the amount of Cu(l)-bathocuproine was quantified by measuring the absorption of this complex at 480 nm (e1, = 13.5 mM-' cm-l) (Sanchez-Rasero. 1981). Spectrophotometric Determination of Cu(1l)-Bleomycin in the Presence of Deferoxamine. To a I-mL solution containing 50 mM sodium cacodylate buffer, pH 7.0, 100 p M CuCI,, and 100 p M deferoxamine were added four degassed 2.5-pL aliquots of IO mM bleomycin A, a t 3-min intervals. Following each addition, the mixture was allowed to equilibrate a t room temperature in a cuvette, and the absorption of Cu(11)-bleomycin was measured a t 600 nm. Indirect Specrrophoromerric Determinarion of Cu(l)Bleomycin in the Presence of Deferoxamine. To a degassed I-mL solution in a Thunberg cuvette containing 50 mM sodium cacodylate, pH 7.0.80 p M CuCI, and 100 JIM bathocuproine was added IO pL of a degassed IO mM deferoxamine solution. Following equilibration a t room temperature, the amount of Cu(1)-bathocuproine present was determined by measuring the absorbance a t 480 nM. Comparison of the relative effects of deferoxamine and bleomycin on the concentration of the Cu(l)-bathocuproine complex afforded an indirect measure of the relative affinities of these ligands for Cu(1). RESULTS
DNA Cleavage by Bleomycin in the Presence of Cu(1l) and Dithiothreitol. As shown in Figure I , the extent of DNA strand scission by BLM in the presence of CuCI, and dithiothreitol varied significantly as a function of the order of addition of reagents. While some cleavage was obtained in each incubation mixture containing BLM and DTT, presumably reflecting the presence of adventitious metal ions, maximal cleavage was obtained when CuCI, was premixed with dithiothreitol and then with BLM prior to their simultaneous addition lo the buffer DNA solution (Figure I . lane 8). Control experiments indicated that this extent of cleavage could be obtained only if BLM, CuCI,, and DTT were present (cf. lanes 9-13). Of particular interest was the observation that alternative protocols for BLM activation using the same reagents were less productive. For example, the sequential addition of BLM. CuCI,, and DTT to the buffered DNA solution gave no cleavage beyond that obtained in the absence of added CuCI, (cf. lanes 3 and 9). In each case, lack of CuCI,-dependent BLM-mediated DNA strand scission was
flOURE 2: Effect of bathocuproine on BLM B, mediated dcgradation ofSV40 DNA in the prcscncc of copper and dithiothrcitol. Rcautions were carried out as described under Experimental Procedures. The components listed were premixed in the order given prior to their (simultaneous) addition to the buffered solution containing SV40 DNA: lane I. Cu(ll). DTT. bathocuproinc;lane 2, Cu(ll), bathocuproine. DTT lane 3. Cu(ll). bathocuproine: lane 4, Cu(ll). DTT, bathocuproine, BLM; lane 5. Cu(ll). bathocuproine, DTT. BLM; lane 6. Cu(ll). DTT. BLM; lane 7. DTT, BLM; lane 8. BLM: lane 9, Cu(ll). DTT, lane IO. DNA alone.
Table I: Spctroscopic Assay of Relative Binding Constants of Blwrnycin. Deferoxamine. and Bathocuproinc for Cu(l) and Cu(ll)' metal complex apparent I
II
Klq
(W)
Cu(ll)-bleomycin A2 Cu(ll)-bathoeuproine
