THE OF BIOLOGICAL CHEMISTRY 269, No. 40, Issue of pp ...11 DNA cleavage in a sequence-selective manner resulting in drug-specific cleavage intensity patterns in sequencing gels (3,

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 269, No. 40, Issue of October 7, pp. 25004-25009, 1994 Printed in U.S.A.

Unique Sequence Specificity of Topoisomerase I1 DNA Cleavage Stimulation and DNA Binding Mode of Streptonigrin*

(Received for publication, March 31, 1994, and in revised form, June 23, 1994)

Giovanni Capranicol, Manlio PalumboB, Stella Tinelli, and Franco Zunino From the Division of Experimental Oncology B, Zstituto Nazionale per lo Studio e la Cura dei lbmori, Via Venezian 1, 20133 Milan, Italy and the $Department of Pharmaceutical Sciences, University of Padova, 35100 Padova, Italy

Streptonigrin stimulated unique intensity patterns of topoisomerase 11-mediated DNA cleavage in agarose and sequencing gels with no similarity to those of doxorubi- cin, V”26,4’(9-acridinylamino)-methanesulfon-m-anisi- dide, genistein, and mitoxantrone. Surprisingly, a statis- tical analysis of 60 sites stimulated by streptonigrin in SV40 and pBR322 DNAs showed that the drug required the dinucleotide 5’-TA-3’ from 2- to 3-positions at the DNAcleavage site. Streptonigrin did not intercalate into the double helix; however, a positive value of the re- duced linear dichroism indicated that indeed the drug interacted with the DNA. An angle of 45” was found be- tween the major drug and local DNA axes, suggesting a minor groove binding mode. Moreover, a DNA winding assay showed that streptonigrin may tighten the helical twist of DNA, similar to the known minor groove binder distamycin. Drug competition for receptor site binding was then evaluated by drug combination in the cleavage reaction. DNA cleavage intensity patterns were altered only with the streptonigridmitoxantrone combination, suggesting that the two compounds may compete for ternary complex formation. The results indicate that streptonigrin may bind to the DNA in a manner similar to that of minor groove binders and that its pharma- cophore, possibly different from other topoisomerase I1 inhibitors, may be an important determinant of its unique sequence position specificity.

Mammalian DNA topoisomerase I1 is a nuclear enzyme that regulates DNA topology by a concerted DNA breaking-rejoining activity (1-4) and is the primary cellular target of antitumor agents, including intercalating drugs, etoposide, streptonigrin, and several other compounds (3-7). These drugs form a ternary complex, DNA-drug-enzyme, in which DNA strands are cleaved with the 5’ termini covalently linked to the protein. The drug action thus stimulates DNA cleavage in intact cells and in the presence of purified enzymes (3-6).

Antitumor drugs of different classes stimulate topoisomerase 11 DNA cleavage in a sequence-selective manner resulting in drug-specific cleavage intensity patterns in sequencing gels (3, 5, 6). Some knowledge of the structural determinants required for effective drug interactions has recently been gained when drug-stimulated cleavage sites of murine topoisomerase I1 were clustered depending on the drug used, and their sequences were analyzed by statistical means (8). These investigations

Ricerche “Progetto Finalizzato ACRO” (Rome, Italy) and by the Asso- * This work was partially supported by Consiglio Nazionale delle

ciazione Italiana per la Ricerca sul Cancro (Milan, Italy). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ To whom correspondence should be addressed. Tel.: 39-2-239-0203; Fax: 39-2-236-2692; E-mail: [email protected].

disclosed that specific nucleotides flanking the strand cut are required for drug stimulation of cleavage. Doxorubicin requires an adenine at least at one of the two 3’ termini (positions -1) of a cleavage site (81, whereas other drugs highly prefer different nucleotides immediately flanking the cleaved phosphodiester bond (9-12). Recently, these observations were further sup- ported by results of mutational analysis studies of topoisomer- ase I1 DNA cleavage in short DNA oligomers (13). Together, the data strongly suggest that drugs form ternary complexes inter- acting with nucleotides flanking the DNA cleavage site at the DNA-protein interface, resulting in inhibition of DNA religa- tion and cleavage stimulation (6, 8).

In a previous study, we showed that both m-AMSA‘ and bisantrene have an adenine requirement at position 1, and their molecular conformations revealed three-dimensional mo- tifs that are common to them but not to other drugs (12). There- fore, bisantrene and m-AMSA share common steric and elec- tronic features that may constitute a specific pharmacophore. We suggested that this pharmacophore may be an important determinant of the sequence position specificity of m-AMSA and bisantrene (12). The results also imply that antitumor topoisomerase I1 inhibitors may have partially distinct phar- macophores and receptor sites.

It was thus of great interest to extend the study to other chemically unrelated drugs in order to gain further information on drug pharmacophores effective against topoisomerase 11. In this work, we focused on the sequence specificity of streptoni- grin (Fig. l), which has recently been shown to stimulate to- poisomerase I1 DNA cleavage (7). The present results show that streptonigrin has a unique sequence specificity and that it may interact in the ternary complex in a manner different from other topoisomerase I1 inhibitors.

EXPERIMENTAL PROCEDURES Materials-Streptonigrin and m-AMSA were obtained from the Drug

Synthesis and Chemistry Branch, National Cancer Institute (Bethesda, MD), dissolved in 85% ethanol, 15% 1 M Tris-HCI, pH 7.9, and dimethyl sulfoxide, respectively, and stored at -20 “C. V”26 and doxorubicin were obtained from Bristol-Myers Italia (Latina, Italy) and Farmitalia- Carlo Erba (Milan, Italy), respectively. Genistein was purchased from Calbiochem and dissolved in dimethyl sulfoxide. Drugs were diluted in deionized water immediately before use. DNA topoisomerase I1 was purified from murine leukemia P388 cell nuclei by published procedures (14, 15) and was stored at -20 “C in storage buffer (20 mM KH,PO,, pH 7.0, 50% glycerol, 0.5 r n ~ phenylmethylsulfonyl fluoride, 0.1 mM EDTA, and 1 mM p-mercaptoethanol). Fractions containing only topoisomerase I activity were collected during the P11 phosphocellulose chromatogra- phy (14), and the enzyme was stored at -80 “C in storage buffer.

SequencingAnalysis ofDNA Cleavage Sites“SV40 and pBR322 DNA fragments were uniquely 5”end-labeled as described already (8, 111, separated by agarose gel electrophoresis, and purified by electroelution.

anesulfon-rn-anisidide; V”26, teniposide; LD, linear dichroism; LDr, The abbreviations used are: rn-AMSA, 4’(9-acridinylamino)-meth-

reduced linear dichroism.

25004

Streptonigrin-stimulated Topoisomerase 11 DNA Cleavage 25005

OCH, I

C T S C T S T.m I

C T S T'm

516 I

732,

727,

FIG. 1. Chemical structure of streptonigrin.

DNA cleavage reactions were performed in 20 p1 of 40 mM Tris-HCI, pH 7.5,80 mM KCl, 10 mM MgCl,, 0.5 mM dithiothreitol, 1 mMATP, 15 pg/ml bovine serum albumin with 106 units (14) of topoisomerase I1 and drugs a t 37 "C for 20 min. Reactions were stopped with SDS and proteinase K (1% and 0.1 mg/ml, respectively) and further incubated a t 42 "C for 45 min. DNAs were then ethanol-precipitated, resuspended in 2.5 p1 of 80% formamide, 10 mM NaOH, 1 mM EDTA, 0.1% dyes, heated a t 95 "C for 2 min, chilled on ice, and then loaded onto an 8% polyacrylamide denaturing gel. Gels were run a t 70 watts for 2 h. Autoradiograms of dried gels were on Amersham Hyperfilms.

Statistical Tests of Cleavage Site Sequences-The statistical tests have been described in detail previously (8,l l) . Briefly, they were (i) the 2 one sample test, used to determine the deviation from the expected base distribution a t each position of the aligned sequences and (ii) measurement of the probability ( p ) of deviation from expected base frequency; the opposite values of the logarithm of p , -log(p), are re- ported for each base a t each position around the cleavage site.

Linear Dichroism Determinations-Flow linear dichroism measure- ments were performed in 10 mM Tris-HC1, pH 7, 20 mM NaCl with a Jasco 5500 apparatus interfaced to an IBM personal computer. Four scans were accumulated for each sample. The spinning rate of the flow cell was 900 rpm. Isotropic absorption (Aim) was measured in a Perkin- Elmer Lambda 5 spectrophotometer. DNA base and streptonigrin con- centrations were 1 mM and 20 p, respectively. The reduced linear dichroism (LDr), defined as the ratio of LD to A,,, is related to the angle (a) between the chromophore electronic transition moment and the local DNA helix axis and to a DNA orientation factor (16). The a value can then be calculated using the LDr value of the DNA chromophores as an internal standard (16). The bases can be considered as essentially per- pendicular to the helix axis. Hence, according to Norden et al. (17), the a value can be calculated using the equation,

a = arccos(K9 - LDr(ligand)/3LDr(DNA))o.5 (Eq. 1)

with LDrtligand) and LDrtDNA) representing the LD values in the ligand (400 nm) and DNA (260 nm) regions, respectively.

DNA Winding Assay--Relaxed DNA was prepared by treatment of SV40 DNAwith an excess amount of topoisomerase I in 20 mM Tris-HC1, pH 7.5, 80 mM KCI, 10 mM MgCl,, 0.5 mM dithiothreitol, and 30 mg/ml bovine serum albumin at 37 "C for 3 h. After phenol extraction and ethanol precipitation, relaxed DNAwas resuspended in water. The DNA winding assay was in 20 pl of 20 mM Tris-HC1, pH 7.5,80 mM KCI, 10 mM MgCl,, 0.5 mM dithiothreitol, 30 mg/ml bovine serum albumin contain- ing 0.4 mg of relaxed DNA, excess amounts of topoisomerase I, and the indicated concentrations of test drugs. Reactions were for 1 h a t 37 "C and then stopped by SDS addition (1% final concentration). After phenol extraction and ethanol precipitation, DNA samples were resuspended in 20 p1 of 10 mM Tris-HCI, pH 8.0, l mM EDTA. DNA was then analyzed by two-dimensional gel electrophoresis according to Lee et al. (18). Briefly, DNA samples were loaded onto a 0.7% agarose gel and electro- phoresed a t 1.2 voltdcm in 90 mM Tris-HC1, 90 mM boric acid, 2.5 m~ EDTA, pH 7.8, for 15-17 h at 22 "C. Asecond-dimension electrophoresis, 90 O to the first dimension, was then conducted in the same buffer containing 0.03 mM ethidium bromide, a t 0.5-0.7 voltlcm for 17-19 h at 22 "C. The gel was finally stained with ethidium bromide, and photo- graphs were taken under U V illumination.

RESULTS

DNA Cleavage Intensity PatternsStreptonigrin (Fig. 1) stimulation of DNA cleavage mediated by murine DNA topoi-

8081

3481

poisomerase 11-mediated DNA cleavage. Plasmid pBR322 DNA FIG. 2. Sequence selectivity of streptonigrin stimulation of to-

fragments were "P-labeled a t Sal1 (left and middle panels) or BamHI (right panel) sites and incubated with the enzyme with or without 100 p~ streptonigrin. Lane C , control DNA, lane T, topoisomerase I1 only; lane S, with streptonigrin; lane T', topoisomerase I1 with drug solvent; lane m, molecular weight markers (purines). Numbers (genomic posi- tions in pBR322 DNA) and arrows indicate the nucleotides covalently linked to the enzyme a t the cleavage sites. The star indicates a con- taminating fragment.

somerase I1 was studied in selected regions of plasmid pBR322 and simian virus 40 DNAs (Figs. 2-4). The drug stimulated DNA cleavage with high selectivity, yielding unique cleavage intensity patterns completely distinct from those of V"26, doxorubicin, mitoxantrone, genistein (Figs. 3 and 41, and m- AMSA (not shown). Although many streptonigrin-stimulated sites were also stimulated by V"26 (Fig. 31, relative site in- tensities were totally different between the two drugs. There- fore, effective interactions of streptonigrin in the ternary com- plex take place at sites that are different from those of the other studied drugs (11, 12).

Without topoisomerase 11, streptonigrin did not cleave the DNA under our experimental conditions (Fig. 3). Cleavage stimulation could be seen a t drug concentrations as low as 10 m, and maximum activity was at 50-100 PM (Figs. 2 4 and data not shown). As expected for a weak DNA binder, cleavage suppression by streptonigrin was generally not observed in agreement with published data (7). The cleavage suppression at site 732 was likely caused by competition with the nearby site 727, highly stimulated by streptonigrin (Fig. 2).

Statistical Analysis of Site Sequences-Sixty streptonigrin- stimulated cleavage sites were collected, and their sequences were analyzed with the statistical methods described previ- ously (8, 11) (Fig. 5 and Table I). The results were completely different from those of other topoisomerase 11-trapping drugs (8, 9, 11, 12), since drug-preferred bases were not observed a t positions (-1 and +1) immediately flanking the strand cut.

Two major observations can be made. (i) Highly nonrandom base distributions were found at 2- and 3-positions. A thymine and an adenine were strongly preferred, whereas purines and pyrimidines were excluded at these positions, respectively (Fig.

25006 Streptonigrin-stimulated Topoisomerase 11 DNA Cleavage

C T s VM-26 DOX MTX -rn"rn-

Streptonigrin(pM) 100 - 10 100 - loo - loo - loo mw - . I_"- _R"-

2902

2069

2853

2829

k

t

t

k

bined with VM-26, doxorubicin, or mitoxantrone. An SV40 DNA FIG. 3. DNA cleavage intensity patterns of streptonigrin com-

fragment was 32P-labeled at the BclI site and incubated with streptoni- grin (S) at the indicated concentrations and/or V"26 (10 p ~ ) , doxoru- bicin (DOX, 1 PM), and mitoxantrone ( M T X , 1 PM). Lane C, DNA incu- bated with streptonigrin without enzyme; lane T, topoisomerase I1 only; lane m, molecular weight markers (purines). Numbers (genomic positions in SV40 DNA) and arrows indicate the nucleotides covalently linked to the enzyme at the cleavage sites. Stars indicate mitox- antrone-stimulated sites that are still stimulated by the streptonigrin/ mitoxantrone combination.

5 and Table I). These base requirements were further sup- ported by the observation that only 6 out of all 60 sites and 1 out of the strongest 21 sites lacked the thymine and the ade- nine a t positions 2 and 3, respectively. (ii) A striking result was the presence of random base distributions a t positions -1 and +1 (Fig. 5 and Table I), since these positions were always biased in previous studies. In particular, the enzyme has been shown to prefer pyrimidines and adenines at positions -1 and +1, respectively, in the absence of drugs (8, 9, 19).

Additional positions showed nonrandom base distributions, a thymine was excluded a t position 5, and a cytosine was pre- ferred at position 8 (Fig. 5 and Table I). These base require- ments, which correspond to adenine exclusion a t position -1 and guanine preference a t position -4 in the opposite strand, are sequence preferences of topoisomerase I1 itself as estab- lished by previous statistical analyses (8, 9, 19) and by a mu- tational study in short DNA oligonucleotides (13).

Effects of Drug Combinations-In an attempt to understand whether drugs may compete for the receptor site, we studied the effect of combining streptonigrin with V"26, m-AMSA, doxorubicin, genistein, or mitoxantrone in DNA cleavage ex-

m MTX GEN " - 100 - 100 Streptonigrin (IIM) 10 100

""mm"m **

i

h g f

0

d

C

b

a

1542 b

i o

""

iw" - "

L

.I _ _ I

bined with mitoxantrone or genistein. An SV40 DNA fragment was FIG. 4. DNA cleavage intensity patterns of streptonigrin com-

32P-labeled at the AccI site and incubated with streptonigrin a t the indicated concentrations andlor mitoxantrone ( " X , 1 PM) or genistein (GEN, 100 PM). Cleavage reactions were in 10 mM Tris-HCI, pH 6,lO mM MgCl,, 50 mM KCl, and 1 mM ATP. Lane C, control DNA, lane T, topoi- somerase I1 only; lune m, molecular weight markers (purines). On the left, a number indicates a genomic position in SV40 DNA; lowercase letters and lines indicate streptonigrin-stimulated cleavage sites.

periments (Figs. 3 and 4 and data not shown). In these experi- ments streptonigrin was always present at a saturating level (100 PM). Moreover, mitoxantrone and doxorubicin were at a concentration (1 m) showing the maximum cleavage activity (since a t higher concentrations cleavage suppression is promi- nent (11,20)), and genistein was also present at a high concen- tration. For V"26, m-AMSA, and doxorubicin, cleavage inten- sity patterns observed with drug combinations were the result of an addition of single drug effects (Fig. 3 and data not shown). Site 2829 was stimulated by streptonigrin and abolished by doxorubicin; their combination resulted in a cleavage level that was intermediate between those seen with the single drug (Fig. 3). Sites stimulated by both v"26 and streptonigrin were more intense in the presence of the two drugs (sites 2829,2853, and 2869) (Fig. 3).

However, cleavage intensity patterns of the mitoxantrone/ streptonigrin combination were not simply the sum of those stimulated by either drug alone (Figs. 3 and 4). Drug interfer- ence could be observed a t specific sites; several mitoxantrone- stimulated sites were not stimulated by the drug combination, whereas nearby sites (indicated by stars in Fig. 3) were still stimulated. At the same time, site 2902 (Fig. 3) and sites a, d, I, and m (Fig. 4) stimulated by streptonigrin were not observed when the drug was combined with mitoxantrone. It can be noted that other streptonigrin-stimulated sites were instead

Streptonigrin-stimulated Topoisomerase II DNA Cleavage 25007

90

80

70

60

50

N * 40

30

20

10

0

I

-25-20-15-10 -5 0 5 10 15 20 25 30

Position from the strand cut FIG. 5. Base preferences at the site of topoisomerase I1 DNA

cleavage stimulated by streptonigrin. Sixty sites were collected in the SV40 and pBR322 DNAs. 2 values indicate deviation from the expected distribution of base frequencies (forp = 0.05 and 0.01,J = 5.99 and 11.34, respectively). -log(p) values above and below the zero line indicate the probability of observing that deviation or more as either excess or deficiency, respectively, relative to the expected frequency of each base (A and T, 21.2%, G and C, 22.8%). On the abscissa axis, zero indicates the observed cleaved phosphodiester bond.

still stimulated (Figs. 3 and 4). Genistein decreased somewhat the stimulation effect of streptonigrin but without any site selectivity (Fig. 4). These results are thus consistent with drug receptor sites being at least partly overlapping for streptoni- grin and mitoxantrone, so that the binding of one drug to its receptor may compete with the binding of the other agent.

DNA Binding Mode of Streptonigrin-The DNA binding ac- tivity of streptonigrin has thus been examined by two inde- pendent tests: linear dichroism measurements and a DNA winding assay.

Attempts to determine precisely the affinity of streptonigrin for DNA were unsuccessful, since the binding constant is very low. We have overcome this limitation in the linear dichroism test by using a concentration of the drug 50-fold lower than that of DNA bases. The reduced linear dichroism values were -0.18 and +0.10 in the DNA and drug-DNA complex regions, respec-

tively. Since DNAintercalating agents show a negative value of the linear dichroism, the positive value detected for streptoni- grin in the ligand absorption region confirmed that the drug is a nonintercalating compound (7) and indicated that indeed streptonigrin interacts with DNA. Since the linear dichroism is related to the angle between the drug chromophore electronic transition moment and the local DNA helix axis (see "Experi- mental Procedures"), a value of 45" was found in the case of streptonigrin, consistent with a groove binding mode of the drug. A similar behavior has been observed for the well known groove binder (4',6-diamidino-2-phenylindole), for which this angle was found to be 43" (17).

We also examined whether streptonigrin was able to change the linking number of relaxed closed circular DNA molecules in the presence of DNA topoisomerase I, as analyzed with two- dimensional agarose gel electrophoresis (Fig. 6). The results indicate that streptonigrin (50 PM) produced a slight but con- sistent band shift toward the positively supercoiled topoiso- mers of SV40 DNA, indicating a winding activity. A marked winding activity was observed with distamycin (1 w), a known minor groove binder, whereas m-AMSA had a band-shifting activity toward the negatively supercoiled topoisomers (Fig. 6). In these experiments, similar results were obtained with either negatively supercoiled or relaxed topoisomers used as the ini- tial DNA substrate (not shown). These results indicate that streptonigrin binding to DNA, although weak, has an effect similar to distamycin and different from m-AMSA, indicating that the drug alters the DNA structure in a manner leading to a tightening of the helical twist.

DISCUSSION

The definition of the sequence specificity of drug stimulation of topoisomerase I1 DNA cleavage may be of great value in identifying the structural determinants of drug activity (6, 8, 11,121. The present results demonstrate that streptonigrin can trap topoisomerase 11-DNA cleavable complexes with a high and unique sequence specificity; the drug requires the dinucle- otide 5'-TA-3' at positions 2 and 3 to stimulate DNA cleavage. To our knowledge, streptonigrin is the first drug that shows primary base requirements at positions other than -1 and +l. The 2 and -log(p) values, used t o define preferred and ex- cluded bases, were almost as high as those found for a doxoru- bicin-stimulated site set (8) and thus are particularly signifi- cant since even sites weakly stimulated by streptonigrin were included (Fig. 5 and Table 1). In a way consistent with previous interpretations of this type of data (6,8,9), the present findings suggest that important interactions of streptonigrin in the ter- nary complex may occur with position 2 and 3 base pairs. This interpretation is further supported by recent findings in our laboratory. The simple isomerization of the amino group of the anthracycline molecule produces a marked alteration of drug cleavage intensity patterns and changes of drug-preferred bases close to the strand cut.2 These results cannot be easily explained by an indirect action of the drug at a distance from the cleavage, thus supporting the idea that the drug is placed at the cleavage site in the ternary complex interacting with the DNA and the enzyme.

Interestingly, streptonigrin has a slight but clear DNA wind- ing activity and shows a reduced linear dichroism in the pres- ence of the nucleic acid that has a positive value, suggesting that the drug does not intercalate into the DNA. Although the precise drug mode of DNA binding needs to be established conclusively, we show that the angle between the double helix axis and the drug molecule was about 45", a value close to that expected for a minor groove binder (17). Therefore, the receptor

G. Capranico and F. Zunino, unpublished data.

25008 Streptonigrin-stimulated Topoisomerase II DNA Cleavage TABLE I

Base frequencies at the site of streptonigrin-stimulated topoisomerase II DNA cleavage

sequences with the same proportions of sequences from the two DNAs, expected base frequencies were 27.2% for A and T and 22.8% for G and C. Sixty drug-stimulated sites were sequenced in SV40 and pBR322 DNAs; base frequencies are reported as percentages. In a collection of random

Base Position

-8 -7 -6 -5 -4 -3 -2 -1 +1 +2 +3 +4 +5 +6 +7 +8 +9 +10 +11 +12

A 25 23 22 22 33 38 18 25 30 13 67 18 30 23 18 17 27 22 28 30 C 20 20 20 28 23 13 27 30 22 18 8 22 30 20 18 43 22 25 28 17 G T

15 27 33 20 30 20 25 25 28 5 17 32 30 20 23 17 20 15 17 25 40 30 25 30 13 28 30 20 20 63 8 28 10 37 40 23 32 38 27 28

Base preferences" no T T A no T C no G no T noA no C

a Bases were considered preferred or excluded when the corresponding -log(p) values were equal or more than 2 (see Fig. 4). Only base preferences or exclusions are indicated.

20

1 0 1 ~ 1 R E Dista R Strep

S Top0 I mAMSA FIG. 6. DNA-winding ability of streptonigrin and distamycin.

?bp panels, relaxed closed circular SV40 DNA molecules were incubated with excess topoisomerase I without (R) or with 50 p~ streptonigrin (Strep) or 1 p~ distamycin (Dista). Bottom panels, negatively super- coiled DNA topoisomers ( S ) were incubated with excess topoisomerase I without (Top0 I) or with m-AMSA (50 p"). DNA topoisomers were analyzed by two-dimensional gel electrophoresis; arrows indicate direc- tions of first and second runs. See text for experimental details.

site in the enzyme-DNA complex might be peculiar for strep- tonigrin and in part different from those of other topoisomerase I1 inhibitors (6). Our present hypothesis is that streptonigrin may be placed externally to the DNA, possibly in the minor groove, interacting with the TA dinucleotide at positions 2 and 3 and interfering with the nearby DNA religation reaction cata- lyzed by the enzyme. It has recently been shown that minor groove binders can interfere with DNA topoisomerase I, stimu- lating enzyme-mediated DNAbreaks (21,221. Our present data indicate that such DNA binding agents may trap either topoi- somerase I or I1 cleavable complexes.

The streptonigrin requirements for the complementary bases T and A at positions 2 and 3, respectively, are symmetrically located relative to the two strand cuts. Since the -log(p) pa- rameter has a very similar high value for both T and A, this may suggest that drug interference occurs at both strand cuts in the ternary complex. In this case, either streptonigrin can bind to the DNA in two opposite orientations with similar abil- ity or an antiparallel drug dimer can form, similar to the case of distamycin A (231, thus affecting both strand cuts at a time.

It has been noted that the molecular structures of topoi- somerase 11-trapping drugs may be rationalized as a composite pharmacophore, since three domains have been identified that are not necessarily present in each individual compound (24). Indeed, some compounds are "pure intercalators" (ellipticine), some are not intercalators (etoposide, VM-261, and some have features of both intercalators and external binders (anthracy- clines, amsacrines, mitoxantrone, and actinomycin D). It has to

be emphasized that in the last type of compounds, generally considered intercalating agents, the external binder moiety of the molecule has been shown to constitute a crucial determi- nant of drug activity for anthracyclines, amsacrine, and mito- xantrone (20, 25-28). We have recently shown that the com- posite pharmacophore may be broken into at least two pharmacophores with more restrictions and requirements of spatial occupancy when considering compounds with the same sequence specificity (12). This idea would suggest the existence of different, although partially overlapping, receptor sites for drugs with diverse sequence position specificity.

The present results thus indicate that streptonigrin may be a prototype of still a different pharmacophore and that the region occupied by this drug in the ternary complex is distinct from those of other topoisomerase 11-trapping agents. Interest- ingly, drug interference in DNA cleavage stimulation has been observed with mitoxantrone but not with V"26, genistein, doxorubicin, or m-AMSA. Interestingly, only mitoxantrone has a base (guanine) preference at position 2, although secondary to those (cytosines and thymines) at position -1 (11). No other drug has been reported to have secondary base requirements at position 2 or 3 (6, 12). These observations might suggest that the mitoxantrone receptor site may overlap more largely than any other of the studied drugs with that of streptonigrin. In particular, a side chain of mitoxantrone might be placed in a DNA groove interacting with the guanine a t position 2.

Streptonigrin, a potent cytotoxic agent with a broad spec- trum of antitumor activity against murine models (291, was shown to inhibit leukocyte mitosis and cause extensive chro- mosomal breaks (30). Therefore, topoisomerase 11-mediated DNA cleavage may be stimulated by streptonigrin also in nu- clear chromatin. Since the drug has a unique sequence speci- ficity, it may be of help in evaluating the role of genomic het- erogeneity of topoisomerase I1 DNA cleavage in cell death mechanisms.

REFERENCES

2. Wang, J. C. (1985) Annu. Rev. Biochem. 54,665-697 1. Gellert, M. (1981)Annu. Reu. Biochem. 50, 879-910

3. Liu, L. F. (1989) Annu. Reu. Biochen. 58, 351-375 4. Osheroff, N. (1989) Pharmacol. & Ther. 41,223-241 5. Pommier, Y., and Kohn, K. W. (1989) in Deuelopments in Cancer Chemotherapy

(Glazer, R. I., ed) pp. 175-195, CRC Press, Boca Raton, FL 6. Capranico, G., and Zunino, F. (1992) Eur. J. Cancer 2 8 4 2055-2060 7. Yamashita, Y., Kawada, S., Fujii, N., and Nakano, H. (1990) Cancer Res. 60,

8. Capranico, G., Kohn, K. W., and Pommier, Y. (1990) Nucleic Acids Res. IS,

9. Pommier, Y., Capranico, G., Om, A,, and Kohn, K. W. (1991) Nucleic Acids Res.

10. Fosse, P., &ne, B., Le Bret, M., Paoletti, C., and Saucier, J. M. (1991) Nucleic

11. Capranico, G., De Isabella, P., Tinelli, S., Bigioni, M., and Zunino, F. (1993)

12. Capranico, G., Palumbo, M., Tinelli, S., Mabilia, M., Pozzan, A,, and Zunino, F.

5841-5844

6611-6619

19,5973-5980

Acids Res. 19,2861-2868

Biochemistry 32,3038-3046

Streptonigrin-stimulated Topoisomerase II DNA Cleavage 25009

13. Bigioni, M., Zunino, F., and Capranico, G. (1994) Nucleic Acids Res. 22,2274-

14. De Isabella, P., Capranico, G., Binaschi, M., Tinelli, S., and Zunino, F. (1990) 2281

15. Minford, J., Pommier, Y., Filipski, J., Kohn, K. W., Kemgan, D., Mattern, M., Mol. Pharmacol. 37, 11-16

16. Norden, B. (1978) Appl. Spectrosc. Reu. 14, 157-248 Michaels, S., Schwartz, R., and Zwelling, L.A. (1986)Biochemistry 25,s-16

17. Norden, B., Eriksson, M., Kim, S. K., Kubista, M., Lyng, R., and Akerman, B. (1990) in Molecular Basis of Specificity in Nucleic Acid-Drug Interactions (Pullman, B., and Jortner, J., eds) pp. 23-41, Kluwer Academic Publishers, Dordrecht, The Netherlands

18. Lee, C.-H., Mizusawa, H., and Kakefuda, T. (1981) Proc. Natl. Acad. Sei.

19. Spitzner, J. R., and Muller, M. T. (1988) Nucleic Acids Res. 16, 5533-5556 U. S. A. 18,2838-2842

20. Capranico, G., Zunino, F., Kohn, K. W., and Pommier, Y. (1990) Biochemistry

21. Chen, A. Y., Yu, C., Bodley, A,, Peng, L. F., and Lin, L. F. (1993) Cancer Res. 53,

(1994) J. M O L B ~ Z . 235, 1218-1230

29,562569

1332-1337

22. Chen, A. Y., Yu, C., Gatto, B., and Liu, L. F. (1993) Proc. Natl. Acad. Sci. U. S. A. 90,8131-8135

23. Pelton, J. G., and Wemmer, D. E. (1989) Proc. Natl. Acad. Sci. U. S. A. 86, 5723-5727

24. Macdonald, T. L., Lehnert, E. K., Loper, J. T., Chow, K.-C., and Ross, W. E. (1991) in DNA Topoisomerases in Cancer (Potmesil, M., and Kohn, K. W., eds) pp. 199-214, Oxford University Press, New York

25. Zwelling, L. A,, Mitchell, M. J., Satitpunwaycha, P., Mayes, J., Altschuler, E.,

26. Nelson, E. M., Tewey, K. M., and Liu, L. F. (1984) Proc. Natl. Acad. Sei. U. S. A. Hinds, M., and Baguley, B. C. (1992) Cancer Res. 52, 209-217

27. Rowe, T. C., Chen, G. L., Hsiang, Y. H., and Liu, L. F. (1986) Cancer Res. 46, 81, 1361-1365

28. De Isabella, P., Capranico, G., Palumbo, M., Sissi, C., Krapcho, A. P., and 2021-2026

29. Oleson, J. J., Calderella, L. A., Mjos, K. J., Reith, A. R., Thie, R. S., and Topin, Zunino, F. (1993) Mol. Pharmacol. 43, 715-721

30. Cohen, M. M., Shaw, M. W., and Craig, A. P. (1963) Proc. Natl. Acad. Sci. I. (1961) Antibiot. Chemothel: 11, 158-164

U. S. A. 50, 16-24