BIOCHEMICAL
Vol. 188, No. 2, 1992 October 30, 1992
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS Pages 584-589
C-l’ HYDROGEN ABSTRACIION OF DEOKYRIBOSE IN DNA STRAND SCISSION BY DYNEMICIN A Takashi Shiraki, Motonari Uesugi, and Yukio Sugiura* Institute for Chemical Research, Kyoto University, Uji, Kyoto 611, Japan Received
July
6,
1992
Summary: Dynemicin A, which is a hybrid antitumor antibiotic containing anthraquinone and enediyne cores, abstracts the C-l’ hydrogen of DNA deoxyribose and then the damaged DNA leads to strand breaks with the formation of S- and 3’-phosphate termini. The lesions of C-4 hydrogen also occur at 3’ side of G-C base pairs (i. e., 5’-CT and 5’-G&, leading to 5’phosphate and 3’-phosphoglycolate termini or 4’-hydroxylated abasic sites. The C-l’ hydrogen abstraction by dynemicin A is distinct from the preferential C-5’ hydrogen abstraction of calicheamicin and neocarxmostatin. 0 1992 Academic Press, Inc.
Enediyne antitumor antibiotics, such as esperamicin, calicheamicin, and neocarzinostatin, are actively studied.1 A curmnt question of these antibiotics is to determine which hydrogen of the deoxyribose, such as C-5’, C-4, and C-l’, is abstracted by the putative carbon-centered diradical species. As to calicheamicin, one radical of the diradical abstracts one of the C-5’ hydrogens and another radical seems to attack the C-4’ hydrogen.2*3*4 As to neocarxinostatin, the C-5’ hydrogen abstraction of thymidine and adenosine residues causes single stranded breaks.5 One of the neocarzinostatin-induced double-stranded lesions consists of an abasic site at the C residue and a strand break at the 1 residue in the sequence AGcGa.
This lesion results
from C-l’ attack of the C residue with the formation of 2deoxyribonolactone and C-5’ attack of the 1 residue with the formation of thymidine 5’-aldehyde.6 The other bistranded lesion consists of a strand break and an abasic lesion resulting from C-4’ hydrogen abstraction at the 1 residue of the GT step and a strand break resulting from C-5’ hydrogen abstraction at the T residue of the CT step in the sequence AGPAa.7 Dynemicin A, recently isolated from Micromonosporu chersina, is a novel antitumor antibiotic consisting of enediyne and anthraquinone components (Figure 1). The novel antibiotic shows a *To whom correspondence should be addressed. 0006-291X192 Copyright All rights
$4.00
0 1992 by Academic Press, Inc. of reproduction in any form reserved.
584
Vol.
188,
No.
2, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
COOH OCHB
OH
0
OH
Figure 1. The chemicalstructureof dynemicin A.
potent antitumor activity in vivo and in vitro. 8vg The antineoplastic action of dynemicin A has been suggested to be associated with the intercalation and cleavage of DNA by the antibiotic. As to the action mechanism of dynemicin A, we indicated that (1) DNA strand scission by dynemicin A is significantly enhanced by the presence of NADPH, thiol compounds, or visible light, (2) the antibiotic interacts with the minor groove of DNA duplex, and (3) intercalation of the anthraquinone core into the DNA followed by attack of the phenylene diradical formed from the enediyne corn is most likely.1Ql1912 As for dynemicin A, no studies on hydrogen abstraction from DNA have been reported. Computer modeling studies on dynemicin A-DNA complex have predicted the 5’-hydrogen abstraction from one of the diradical of dynemicin A. 13~14 Here, we present the experimental evidences that the C-l’ hydrogen of deoxyribose is abstracted in dynemicin A-induced DNA strand scission. Materials and Methods m Purified dynemicin A was kindly supplied by Bristol-Myers Squibb Research Institute, Tokyo, Japan. Peplomycin, one of the bleomycin analogues, was supplied by Nippon Kayaku. Plasmid pBR322 DNA was isolated from Escherichiu coli C 600, and restriction endonucleases Bum HI, Hhu I, Sal I, and Sph I were obtained from Takara Shuzo (Kyoto, Japan). Micrococcal nuclease was purchased from Sigma, and T4 polynucleotide kinase and bacterial alkaline phosphatase (I??.cofi C 75) were obtained from Takara Shuzo (Kyoto, Japan). All other,chemicals used wem of commercial reagent grade. of 5 -m . . the t&y&&. To confirm S-phosphate termini at the cleavage sites in dynemicin A-induced DNA strand scission, the S-phosphatase reaction was carried out. The 3’-end labeled pBR322 (Bum HI-Hha I fragment) was cleaved by dynemicin A in the presence of NADPH, and then treated with 4 units of bacterial alkaline phosphatase (BAP) (E. coli C 75) at 65 ‘C for 1 hr. As the S-hydroxyl terminus marker, the DNA fragment digested by micrococcal nuclease (MNase) was prepared. 15 These DNA fragments were electrophomsed by sequencing gels. l6 . rmlru To determine 3’-phosphate termini at the cutting sites in dynemicin A-induced DNA cleavage, the 3’-phosphatase reaction was carried out. l7 The 5’-end labeled pBR322 (Sal I-Hhu I fragment) was cleaved by dynemicin A-NADPH system, and then treated with 6 units of T4 polynucleotide kinase (T4 PNK) at 37 OC for 45 min in the 585
Vol.
188,
No.
2,
1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
absence of ATP. As the 3’-hydroxyl terminus marker, the DNA fragment digested by DNase I was prepared. , DNA fragments cleaved by Detection of3 -ohosDho~lvcolare the cleave peplomycin-Fe (PeM-Fe) complex possess 3’-phosphoglycolate ends as a result of C-4 hydrogen abstraction. l8 Dynemicin A-cleaved DNA fragments were compared with those of peplomycin-Fe complex by using the S-end . . labeled Sal I-Hhu I fragment. , Detecti sf 4 -hyitu&&&astc u&x Partially, C-4’ radical of deoxyribose is known to lead to the formation of 4’-hydroxylated abasic sites instead of strand breaks.l9*20 Therefore, the detection also makes a diagnosis of the C-4’ hydrogen abstraction. When 4’-hydroxylated abasic sites are treated with hydrazine, the DNA strands ate breaked with the production of 3’pyridazinyhnethyl termini. The 5’-end labeled Sal I-Hha I fragment was cleaved by dynemicin A in the presence of NADPH, and then treated with 0.1 M hydrazine-HCl (pH 8.0) at 90 “C for 5 min. After ethanol precipitation, the DNA fragments were electrophoresed by 15 % polyacrylamide sequencing gels. Results and Discussion Production of 5’-phosphate and 3’-phosphoglycolate tetmini results from the C-4’ hydrogen abstraction as shown in bleomycin-Fe system. 18 The 5’-aldehyde and 3’-phosphate ends are formed by the C-5’ hydrogen abstraction which is demonstrated in neocarzinostatin.5 On the other hand, the 5’-phosphate and 3’-phosphate ends are considered to be final products in the Cl’ hydrogen abstraction as indicated by 1, 10-phenanthroline-Cu.21~22
Figure 2 shows an
analysis for 5’-terminal structures of dynemicin A-cleaved DNA by using the 3’-end labeled DNA fragments. As shown in lane 4, all the cleaved fragments possess the same mobility as the Maxam-Gilbert sequencing markers (lanes 2 and 3), indicating that these sites have 5’-phosphate termini.16 Furthermore, the 5’-phosphatase reaction converted all the dynemicin A-cleaved DNA bands into slower migrating fragments (lane 5) that corn&rate with micrococcal nucleasedigested DNA fragments, that is, 5’-hydroxyl ended fragments (lane 6). Thus, all the dynemicin A-cleaved DNA fragments seem to have 5’-phosphate termini. The mobility of dynemicin Acleaved DNA was not altered by base treatment, showing that no 5’-nucleoside aldehyde termini arc produced by dynemicin A. The structures of 3’-termini at the cleaved sites are somewhat complex. Figure 3 shows an analysis for 3’-terminal structures of dynemicin A-cleaved DNA by using 5’-end labeled DNA fragments. Most of the cleaved fragments (lane 4) possess the same mobility as tire MaxamGilbert sequencing markers (lanes 2 and 3), that is, these sites have 3’-phosphate termini.16 To confirm this point, the 3’-phosphatase reaction was carried out. The 3’-phosphatase reaction converted dynemicin A-cleaved fragments into slower migrating ones (lane 5) that comigrated with DNase I-digested fragments (lane 6), that is, 3’-hydroxyl ended fragments. Therefore, most of the cleaved sites possess 3’-phosphate termini. Similar 5’- and 3’-phosphate productions have
Vol.
188,
No.
2,
1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
/
02 Figure 2. Analysis of S-terminal structures in dynemicin A-induced DNA cleavage. The 3’-end labeled DNA was cleaved by dynemicin A (20 pM) with NADPH (1 mM) at 37 “C for 18 hr (lane 4), followed by S-phosphatase reaction using bacterial alkaline phosphatase (BAP) (lane 5). Lanes 2 and 3 show the Maxam-Gilbert sequencing reactions for C+T and G+A, respectively, and lane 1 indicates intact DNA. Lane 6 shows micrococcal nuclease (MNase)digested DNA as the 5’-hydroxyl marker. Figure 3. Analysis of 3’-terminal structures in dynemicin A-induced DNA cleavage. The S-end labeled DNA was cleaved by dynemicin A (20 l.tM) with NADPH (1 mM) at 37 ‘C for 18 hr (lane 4), followed by 3’-phosphatase reaction using T4 polynucleotide kinase (T4 PNK) (lane 5). Lanes 2 and 3 show the Maxam-Gilbert sequencing reactions for C+T and G+A, respectively, and lane 1 indicates intact DNA. Lane 6 shows DNase I-digested DNA as the 3’hydroxyl marker.
been reported in the DNA cleavage by 1, IO-phenanthroline-copper. The 5’- and 3’-phosphate, free bases, and 5-methylene-2-SKfuranone
(5-MF) are known to form as a result of the C-l’
hydrogen abstraction.21922 Dynemicin A clearly produced 5’- and 3’-phosphates and free bases (data not shown), although 5-MF was not detected. The C-l’ hydrogen abstraction is the most likely for explanation of dynemicin A-induced DNA damage (Figure 6). Since calicheamicin and neocarzinostatin are believed to abstract preferentially
C-S hydrogen,2731475 dynemicin A
appears to have a umque cleaving mode among the enediyne antitumor antibiotics. As shown in Figure 4, certain cleaved sites such as 5’-CT- and 5’-GA, split into doublet (lanes 2 and 3). The bands corn&rated with the Maxam-Gilbert sequencing markers have 3’phosphate termini, and then the other bands (indicated by arrows), which are slightly ahead of the 3’-phosphate-ended fragments, run identically with the peplomycin-Fe-cleaved fragments (lane 4). The complex of peplomycin with iron is known to cleave DNA by the C-4’ hydrogen 587
Vol.
188, No. 2, 1992
BIOCHEMICAL
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
.-P N m P
04
123
1234
Figure 4. Comparison of 3’-terminal structures produced by dynemicin A and peplomycin-Fe. The S-end labeled DNA was cleaved by dynemicin A (20 pM) under visible light radiation at 20 ‘C for 1 hr (lane 2) or in the presence of methyl thioglycolate (1 mM) at 37 ‘C for 18 hr (lane 3). Lane 1 indicates the Maxam-Gilbert sequencing reaction for G+A. Lane 4 shows DNA cleavage pattern by peplomycin-Fe (PeM-Fe) as the 3’-phosphoglycolate marker. Figure 5. Detection of 4’-hydroxylated abasic sites. The S-end labeled DNA was cleaved by dynemicin A (20 PM) under visible light radiation (lane 2), followed by hydrazine (0.1 M, pH 8.0) treatment at 90 ‘C for 5 min (lane 3). Lane 1 shows the Maxam-Gilbert sequencing reaction for G+A.
abstraction and to form 3’-phosphoglycolate termini. 18 Therefore, dynemicin A seems to attack the C-4’ hydrogen at these cleavage sites. The C-4’ radicals of deoxyribose are known to lead in part to the formation of 4’hydroxylated abasic sites instead of strand scission. 19920 Figure 5 shows an analysis for 4’hydroxylated abasic sites in dynemicin A-cleaved DNA. As indicated in lane 3, hydrazinetreatment produced several new bands migrating about a half base slower than the MaxamGilbert sequencing markers. These new bands (indicated by arrows) are presumed to be 3’pyridazinylmethyl
terminal fragments reported in the studies on bleomycin. Consistent with
spontaneous formation of 3’-phosphoglycolate
ends, 4’-hydroxylated
abasic sites are also
formed at 5’-CT_and S-GA. Thus, these sites undergo the C-4’ hydrogen abstraction in addition to the C-l’ lesion (Figure 6). It should be noted that both C-l’ and C-4’ hydrogens lie in the minor groove of DNA. This is consistent with the minor groove interaction of dynemicin A reported
previously.lO
Most of the C-4’ lesions occur at 3’ side of G*C base pain, and hence 2-
amino group of guanine may partially inhibit accessof dynemicin A to deep position in the minor 588
Vol.
188, No. 2, 1992
BIOCHEMICAL
AND BIOPHYSICAL
-
0
0
RESEARCH COMMUNICATIONS
3phosphu!&mtate
? 0:p.o
0
&
o=;-cr
-w
OR s-phqhate o:*.cr OR 0
0 O:P-0 c
B’pyridazinylmelhyl temeus
b c-hydmxyhted
abask
sile
Figure 6. Proposed schemefor the DNA cleavage by dynemicin A via C-l’ (upper) or C-4 (lower) hydrogen abstraction.
groove of B-form DNA. Indeed, we found previously that modification of the guauine 2-amino group with authtarnyciu greatly inhibits DNA cleavage by dynemicin A.10 References 1. Nicolaou, K. C., andDai, W.-M. (1991) Angew. Chem. Int. Ed. Engl. 30. 1387-1416. 2. Zein, N., Sinha. A. M., McGahren, W. J., and Ellestad, G. A. (1988) Science 240, 1198-1201. 3. Zein, N.. Poncin, M., Nilakantan. R., and Ellestad, G. A. (1989) Science 244, 697-699. 4. De Voss. J. J., Townsend, C. A., Ding, W.-D., Morton, Cl. 0.. Ellestad, G. A., Zein, N., Tabor, A. B., and Schreiber, S. L. (1990) J. Am. Chem. Sot. 112.9669-9670. 5. Goldberg, I. H. (1987) Free Radical Biol. Med. 3,41-54. 6. Kappen, L. S., and Goldberg, I. H. (1989) Biochemistry 28, 1027-1032. 7. Dedon, P. C.. and Goldberg, I. H. (1990) J. Biol. Chem. 265, 14713-14716. 8. Konishi, M.. Ohkuma, H., Matsumoto. K., Tsuno. T., Kamei, H., Miyaki, T., Oki, T., and Kawaguchi, H. (1989) J. Antibiot. 42, 1449-1452. 9. Konishi, M., Ohkuma, H.. Tsuno, T., Oki, T., VanDuyne, G. D., and Clardy, J. (1990) J. Am. Chem. Sot. 112,3715-3716. 10. Sugiura, Y., Shiraki, T., Konishi, M., and Oki, T. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 3831-3835. 11. Shiraki, T., and Sugiura, Y. (1990) Biochemistry 29,9795-9798. 12. Sugiura, Y., Arakawa, T., Uesugi, M., Shiraki, T., Ohkuma. H., and Konishi, M. (1991) Biochemistry 30.2989-2992. 13. Langley, D. R., Doyle, T. W., and Beveridge, D. L. (1991) J. Am. Chem. Sot. 113,4395-4403. 14. Wender, P. A., Kelly, R. C., Beckham, S.. and Miller, B. L. (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 8835-8839. 15. Richardson, C. C., and Komberg. A. (1964) J. Biol. Chem. 239, 242-250. 16. Maxam, A. M., and Gilbert, W. (1980) Methods Enzymol. 65,499-560. 17. Cameron, V., and Uhlenbeck, 0. C.. (1977) Biochemistry 16,5120-5126. 18. Stubbe, J., and Kozarich, J. W. (1987) Chem. Rev. 87, 1107-1136. 19. Sugiyama, H., Xu, C.. Murugesan, N., Hecht, S. M.. van der Marel. G. A., and van Boom, J. H. (1988) Biochemistry 27.58-67. 20. Sugiyama, H., Kawabata, H., Fujiwara, T., Dannoue, Y., and Saito, I. (1990) J. Am. Chem. Sot. 112,5252-5257. 21. Kuwabara, M., Yoon, C.. Goyne. T., Thederahn. T.. and Sigman, D. S. (1986) Biochemistry 25,7401-7408. 22. Goyne. T. E., and Sigman, D. S. (1.987) J. Am. Chem. Sot. 109,2846-2848. 589
Vol. 188, No. 2, 1992 October 30, 1992
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS Pages 584-589
C-l’ HYDROGEN ABSTRACIION OF DEOKYRIBOSE IN DNA STRAND SCISSION BY DYNEMICIN A Takashi Shiraki, Motonari Uesugi, and Yukio Sugiura* Institute for Chemical Research, Kyoto University, Uji, Kyoto 611, Japan Received
July
6,
1992
Summary: Dynemicin A, which is a hybrid antitumor antibiotic containing anthraquinone and enediyne cores, abstracts the C-l’ hydrogen of DNA deoxyribose and then the damaged DNA leads to strand breaks with the formation of S- and 3’-phosphate termini. The lesions of C-4 hydrogen also occur at 3’ side of G-C base pairs (i. e., 5’-CT and 5’-G&, leading to 5’phosphate and 3’-phosphoglycolate termini or 4’-hydroxylated abasic sites. The C-l’ hydrogen abstraction by dynemicin A is distinct from the preferential C-5’ hydrogen abstraction of calicheamicin and neocarxmostatin. 0 1992 Academic Press, Inc.
Enediyne antitumor antibiotics, such as esperamicin, calicheamicin, and neocarzinostatin, are actively studied.1 A curmnt question of these antibiotics is to determine which hydrogen of the deoxyribose, such as C-5’, C-4, and C-l’, is abstracted by the putative carbon-centered diradical species. As to calicheamicin, one radical of the diradical abstracts one of the C-5’ hydrogens and another radical seems to attack the C-4’ hydrogen.2*3*4 As to neocarxinostatin, the C-5’ hydrogen abstraction of thymidine and adenosine residues causes single stranded breaks.5 One of the neocarzinostatin-induced double-stranded lesions consists of an abasic site at the C residue and a strand break at the 1 residue in the sequence AGcGa.
This lesion results
from C-l’ attack of the C residue with the formation of 2deoxyribonolactone and C-5’ attack of the 1 residue with the formation of thymidine 5’-aldehyde.6 The other bistranded lesion consists of a strand break and an abasic lesion resulting from C-4’ hydrogen abstraction at the 1 residue of the GT step and a strand break resulting from C-5’ hydrogen abstraction at the T residue of the CT step in the sequence AGPAa.7 Dynemicin A, recently isolated from Micromonosporu chersina, is a novel antitumor antibiotic consisting of enediyne and anthraquinone components (Figure 1). The novel antibiotic shows a *To whom correspondence should be addressed. 0006-291X192 Copyright All rights
$4.00
0 1992 by Academic Press, Inc. of reproduction in any form reserved.
584
Vol.
188,
No.
2, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
COOH OCHB
OH
0
OH
Figure 1. The chemicalstructureof dynemicin A.
potent antitumor activity in vivo and in vitro. 8vg The antineoplastic action of dynemicin A has been suggested to be associated with the intercalation and cleavage of DNA by the antibiotic. As to the action mechanism of dynemicin A, we indicated that (1) DNA strand scission by dynemicin A is significantly enhanced by the presence of NADPH, thiol compounds, or visible light, (2) the antibiotic interacts with the minor groove of DNA duplex, and (3) intercalation of the anthraquinone core into the DNA followed by attack of the phenylene diradical formed from the enediyne corn is most likely.1Ql1912 As for dynemicin A, no studies on hydrogen abstraction from DNA have been reported. Computer modeling studies on dynemicin A-DNA complex have predicted the 5’-hydrogen abstraction from one of the diradical of dynemicin A. 13~14 Here, we present the experimental evidences that the C-l’ hydrogen of deoxyribose is abstracted in dynemicin A-induced DNA strand scission. Materials and Methods m Purified dynemicin A was kindly supplied by Bristol-Myers Squibb Research Institute, Tokyo, Japan. Peplomycin, one of the bleomycin analogues, was supplied by Nippon Kayaku. Plasmid pBR322 DNA was isolated from Escherichiu coli C 600, and restriction endonucleases Bum HI, Hhu I, Sal I, and Sph I were obtained from Takara Shuzo (Kyoto, Japan). Micrococcal nuclease was purchased from Sigma, and T4 polynucleotide kinase and bacterial alkaline phosphatase (I??.cofi C 75) were obtained from Takara Shuzo (Kyoto, Japan). All other,chemicals used wem of commercial reagent grade. of 5 -m . . the t&y&&. To confirm S-phosphate termini at the cleavage sites in dynemicin A-induced DNA strand scission, the S-phosphatase reaction was carried out. The 3’-end labeled pBR322 (Bum HI-Hha I fragment) was cleaved by dynemicin A in the presence of NADPH, and then treated with 4 units of bacterial alkaline phosphatase (BAP) (E. coli C 75) at 65 ‘C for 1 hr. As the S-hydroxyl terminus marker, the DNA fragment digested by micrococcal nuclease (MNase) was prepared. 15 These DNA fragments were electrophomsed by sequencing gels. l6 . rmlru To determine 3’-phosphate termini at the cutting sites in dynemicin A-induced DNA cleavage, the 3’-phosphatase reaction was carried out. l7 The 5’-end labeled pBR322 (Sal I-Hhu I fragment) was cleaved by dynemicin A-NADPH system, and then treated with 6 units of T4 polynucleotide kinase (T4 PNK) at 37 OC for 45 min in the 585
Vol.
188,
No.
2,
1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
absence of ATP. As the 3’-hydroxyl terminus marker, the DNA fragment digested by DNase I was prepared. , DNA fragments cleaved by Detection of3 -ohosDho~lvcolare the cleave peplomycin-Fe (PeM-Fe) complex possess 3’-phosphoglycolate ends as a result of C-4 hydrogen abstraction. l8 Dynemicin A-cleaved DNA fragments were compared with those of peplomycin-Fe complex by using the S-end . . labeled Sal I-Hhu I fragment. , Detecti sf 4 -hyitu&&&astc u&x Partially, C-4’ radical of deoxyribose is known to lead to the formation of 4’-hydroxylated abasic sites instead of strand breaks.l9*20 Therefore, the detection also makes a diagnosis of the C-4’ hydrogen abstraction. When 4’-hydroxylated abasic sites are treated with hydrazine, the DNA strands ate breaked with the production of 3’pyridazinyhnethyl termini. The 5’-end labeled Sal I-Hha I fragment was cleaved by dynemicin A in the presence of NADPH, and then treated with 0.1 M hydrazine-HCl (pH 8.0) at 90 “C for 5 min. After ethanol precipitation, the DNA fragments were electrophoresed by 15 % polyacrylamide sequencing gels. Results and Discussion Production of 5’-phosphate and 3’-phosphoglycolate tetmini results from the C-4’ hydrogen abstraction as shown in bleomycin-Fe system. 18 The 5’-aldehyde and 3’-phosphate ends are formed by the C-5’ hydrogen abstraction which is demonstrated in neocarzinostatin.5 On the other hand, the 5’-phosphate and 3’-phosphate ends are considered to be final products in the Cl’ hydrogen abstraction as indicated by 1, 10-phenanthroline-Cu.21~22
Figure 2 shows an
analysis for 5’-terminal structures of dynemicin A-cleaved DNA by using the 3’-end labeled DNA fragments. As shown in lane 4, all the cleaved fragments possess the same mobility as the Maxam-Gilbert sequencing markers (lanes 2 and 3), indicating that these sites have 5’-phosphate termini.16 Furthermore, the 5’-phosphatase reaction converted all the dynemicin A-cleaved DNA bands into slower migrating fragments (lane 5) that corn&rate with micrococcal nucleasedigested DNA fragments, that is, 5’-hydroxyl ended fragments (lane 6). Thus, all the dynemicin A-cleaved DNA fragments seem to have 5’-phosphate termini. The mobility of dynemicin Acleaved DNA was not altered by base treatment, showing that no 5’-nucleoside aldehyde termini arc produced by dynemicin A. The structures of 3’-termini at the cleaved sites are somewhat complex. Figure 3 shows an analysis for 3’-terminal structures of dynemicin A-cleaved DNA by using 5’-end labeled DNA fragments. Most of the cleaved fragments (lane 4) possess the same mobility as tire MaxamGilbert sequencing markers (lanes 2 and 3), that is, these sites have 3’-phosphate termini.16 To confirm this point, the 3’-phosphatase reaction was carried out. The 3’-phosphatase reaction converted dynemicin A-cleaved fragments into slower migrating ones (lane 5) that comigrated with DNase I-digested fragments (lane 6), that is, 3’-hydroxyl ended fragments. Therefore, most of the cleaved sites possess 3’-phosphate termini. Similar 5’- and 3’-phosphate productions have
Vol.
188,
No.
2,
1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
/
02 Figure 2. Analysis of S-terminal structures in dynemicin A-induced DNA cleavage. The 3’-end labeled DNA was cleaved by dynemicin A (20 pM) with NADPH (1 mM) at 37 “C for 18 hr (lane 4), followed by S-phosphatase reaction using bacterial alkaline phosphatase (BAP) (lane 5). Lanes 2 and 3 show the Maxam-Gilbert sequencing reactions for C+T and G+A, respectively, and lane 1 indicates intact DNA. Lane 6 shows micrococcal nuclease (MNase)digested DNA as the 5’-hydroxyl marker. Figure 3. Analysis of 3’-terminal structures in dynemicin A-induced DNA cleavage. The S-end labeled DNA was cleaved by dynemicin A (20 l.tM) with NADPH (1 mM) at 37 ‘C for 18 hr (lane 4), followed by 3’-phosphatase reaction using T4 polynucleotide kinase (T4 PNK) (lane 5). Lanes 2 and 3 show the Maxam-Gilbert sequencing reactions for C+T and G+A, respectively, and lane 1 indicates intact DNA. Lane 6 shows DNase I-digested DNA as the 3’hydroxyl marker.
been reported in the DNA cleavage by 1, IO-phenanthroline-copper. The 5’- and 3’-phosphate, free bases, and 5-methylene-2-SKfuranone
(5-MF) are known to form as a result of the C-l’
hydrogen abstraction.21922 Dynemicin A clearly produced 5’- and 3’-phosphates and free bases (data not shown), although 5-MF was not detected. The C-l’ hydrogen abstraction is the most likely for explanation of dynemicin A-induced DNA damage (Figure 6). Since calicheamicin and neocarzinostatin are believed to abstract preferentially
C-S hydrogen,2731475 dynemicin A
appears to have a umque cleaving mode among the enediyne antitumor antibiotics. As shown in Figure 4, certain cleaved sites such as 5’-CT- and 5’-GA, split into doublet (lanes 2 and 3). The bands corn&rated with the Maxam-Gilbert sequencing markers have 3’phosphate termini, and then the other bands (indicated by arrows), which are slightly ahead of the 3’-phosphate-ended fragments, run identically with the peplomycin-Fe-cleaved fragments (lane 4). The complex of peplomycin with iron is known to cleave DNA by the C-4’ hydrogen 587
Vol.
188, No. 2, 1992
BIOCHEMICAL
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
.-P N m P
04
123
1234
Figure 4. Comparison of 3’-terminal structures produced by dynemicin A and peplomycin-Fe. The S-end labeled DNA was cleaved by dynemicin A (20 pM) under visible light radiation at 20 ‘C for 1 hr (lane 2) or in the presence of methyl thioglycolate (1 mM) at 37 ‘C for 18 hr (lane 3). Lane 1 indicates the Maxam-Gilbert sequencing reaction for G+A. Lane 4 shows DNA cleavage pattern by peplomycin-Fe (PeM-Fe) as the 3’-phosphoglycolate marker. Figure 5. Detection of 4’-hydroxylated abasic sites. The S-end labeled DNA was cleaved by dynemicin A (20 PM) under visible light radiation (lane 2), followed by hydrazine (0.1 M, pH 8.0) treatment at 90 ‘C for 5 min (lane 3). Lane 1 shows the Maxam-Gilbert sequencing reaction for G+A.
abstraction and to form 3’-phosphoglycolate termini. 18 Therefore, dynemicin A seems to attack the C-4’ hydrogen at these cleavage sites. The C-4’ radicals of deoxyribose are known to lead in part to the formation of 4’hydroxylated abasic sites instead of strand scission. 19920 Figure 5 shows an analysis for 4’hydroxylated abasic sites in dynemicin A-cleaved DNA. As indicated in lane 3, hydrazinetreatment produced several new bands migrating about a half base slower than the MaxamGilbert sequencing markers. These new bands (indicated by arrows) are presumed to be 3’pyridazinylmethyl
terminal fragments reported in the studies on bleomycin. Consistent with
spontaneous formation of 3’-phosphoglycolate
ends, 4’-hydroxylated
abasic sites are also
formed at 5’-CT_and S-GA. Thus, these sites undergo the C-4’ hydrogen abstraction in addition to the C-l’ lesion (Figure 6). It should be noted that both C-l’ and C-4’ hydrogens lie in the minor groove of DNA. This is consistent with the minor groove interaction of dynemicin A reported
previously.lO
Most of the C-4’ lesions occur at 3’ side of G*C base pain, and hence 2-
amino group of guanine may partially inhibit accessof dynemicin A to deep position in the minor 588
Vol.
188, No. 2, 1992
BIOCHEMICAL
AND BIOPHYSICAL
-
0
0
RESEARCH COMMUNICATIONS
3phosphu!&mtate
? 0:p.o
0
&
o=;-cr
-w
OR s-phqhate o:*.cr OR 0
0 O:P-0 c
B’pyridazinylmelhyl temeus
b c-hydmxyhted
abask
sile
Figure 6. Proposed schemefor the DNA cleavage by dynemicin A via C-l’ (upper) or C-4 (lower) hydrogen abstraction.
groove of B-form DNA. Indeed, we found previously that modification of the guauine 2-amino group with authtarnyciu greatly inhibits DNA cleavage by dynemicin A.10 References 1. Nicolaou, K. C., andDai, W.-M. (1991) Angew. Chem. Int. Ed. Engl. 30. 1387-1416. 2. Zein, N., Sinha. A. M., McGahren, W. J., and Ellestad, G. A. (1988) Science 240, 1198-1201. 3. Zein, N.. Poncin, M., Nilakantan. R., and Ellestad, G. A. (1989) Science 244, 697-699. 4. De Voss. J. J., Townsend, C. A., Ding, W.-D., Morton, Cl. 0.. Ellestad, G. A., Zein, N., Tabor, A. B., and Schreiber, S. L. (1990) J. Am. Chem. Sot. 112.9669-9670. 5. Goldberg, I. H. (1987) Free Radical Biol. Med. 3,41-54. 6. Kappen, L. S., and Goldberg, I. H. (1989) Biochemistry 28, 1027-1032. 7. Dedon, P. C.. and Goldberg, I. H. (1990) J. Biol. Chem. 265, 14713-14716. 8. Konishi, M.. Ohkuma, H., Matsumoto. K., Tsuno. T., Kamei, H., Miyaki, T., Oki, T., and Kawaguchi, H. (1989) J. Antibiot. 42, 1449-1452. 9. Konishi, M., Ohkuma, H.. Tsuno, T., Oki, T., VanDuyne, G. D., and Clardy, J. (1990) J. Am. Chem. Sot. 112,3715-3716. 10. Sugiura, Y., Shiraki, T., Konishi, M., and Oki, T. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 3831-3835. 11. Shiraki, T., and Sugiura, Y. (1990) Biochemistry 29,9795-9798. 12. Sugiura, Y., Arakawa, T., Uesugi, M., Shiraki, T., Ohkuma. H., and Konishi, M. (1991) Biochemistry 30.2989-2992. 13. Langley, D. R., Doyle, T. W., and Beveridge, D. L. (1991) J. Am. Chem. Sot. 113,4395-4403. 14. Wender, P. A., Kelly, R. C., Beckham, S.. and Miller, B. L. (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 8835-8839. 15. Richardson, C. C., and Komberg. A. (1964) J. Biol. Chem. 239, 242-250. 16. Maxam, A. M., and Gilbert, W. (1980) Methods Enzymol. 65,499-560. 17. Cameron, V., and Uhlenbeck, 0. C.. (1977) Biochemistry 16,5120-5126. 18. Stubbe, J., and Kozarich, J. W. (1987) Chem. Rev. 87, 1107-1136. 19. Sugiyama, H., Xu, C.. Murugesan, N., Hecht, S. M.. van der Marel. G. A., and van Boom, J. H. (1988) Biochemistry 27.58-67. 20. Sugiyama, H., Kawabata, H., Fujiwara, T., Dannoue, Y., and Saito, I. (1990) J. Am. Chem. Sot. 112,5252-5257. 21. Kuwabara, M., Yoon, C.. Goyne. T., Thederahn. T.. and Sigman, D. S. (1986) Biochemistry 25,7401-7408. 22. Goyne. T. E., and Sigman, D. S. (1.987) J. Am. Chem. Sot. 109,2846-2848. 589
