Biochimica et Biophysica Acta, 1090(1991)52-60 © 1991ElsevierSciencePublishersB.V.016%4781/91/$03.50 ADONIS 0167478191001898
Netropsin and bis-netropsin analogs as inhibitors of the catalytic activity of mammalian DNA topoisomerase II and topoisomerase cleavable complexes T.A. Beerman l, J.M. Woynarowski ~, R.D. Sigmund !, L.S. Gawron l, K.E. Rao 2 and J.W. Lown 2 t Department of Experimental Therapeutics, Roswell Park Cancer Institute, Buffalo, NY (U.S.A.) and 2 Department of Chemistry, University of Alberta, Edmonton (Canada)
(Received 29 March 1991)
Key words: Netropsin;DNA topoisomerase!I; (Mammalian) This study examined the ability of netropsin and relat~l minor groove binders to interfere with the actions of DNA topoisomerases 11 and !. We evaluated a series of netropsin dimers linked with flexible aliphatic chains of different lengths. These agents are potentially able to occupy longer stretches of DNA than the parental drug as a result of bidentate binding. Both netropsin and its dimers were found: (i) to inhibit the catalytic activity of isolated topoisomerase !! and (ii) to interfere with the stabilization of the cleavable complexes of topoisomerase H and I in nuclei. Dimers with linkers consisting of 0 - 4 and 6 - 9 methylene groups (n) were far more inhibitgry than netropsin against isolated enzyme and in the nuclear system. The compound with n = $ was less active than netropsin in both assays while the dimmerwith n = 10 inhibited only the isolated enzyme. The comparison of dimers with fixed Hnker length (n = 2) but varying number of N-methylpyrrole residues (from 1 to 3) revealed that the inhibitory properties were enhanced with increasing number of N-methylpyrroie units. For dimers with varying linker length, drug ability to inhibit catalytic activity of isolated topoisomerase 11 was positively correlated with calf thymus DNA association constants. In contrast, no such correlation existed in nuclei, However, the inhibitory effects in the nuclear system were correlated with the association constants for poly(dAdT). The results indicate that bidentate binding can significantly enhance anti-topoisomerase activity of netropsin related dimeric minor groove binders. However, other factors such as the length of the linker, the number of pyrrole moieties and the nature of the target (isolated enzyme/DNA versus chromatin in nuclei) also contribute to these activities.
Abbreviations: Hoechst 33258, 244-hydroxyphenyl)-5-[5-(4-methylpiperazin-l-yl)-benzimidazole-2-yl]benzimidazole;DAPI, 4',6-diamidino-phenylindole; VM-26, 4'-demethylcpipodophyllotoxin thionylidyneglucoside;m-AMSA,4'49-acridinylamino)methanesulfon-m-anisidide;kDNA, kinetoplastDNA from Crithidia fasciculata; SDS, sodiumdodecylsulfate; IC,~,,drug concentrationinhibitingby 90% decatenationof kDNAby isolated topoisomerasell; IC~. drug concentration inhibitingby 50% inductionof topoisomerase mediated DNA-protein crosslinks in nuclei; PBS, phosphate-buffered saline (0.14 M NaCI,2 mM KCI,6 mM Na2HPO4, I mM KH2PO4 (oH 7.2). Correspondence:T.A. Beerman,Departmentof ExperimentalTherapeutics, RoswellPark CancerInstitute,Buffalo,N.Y. 14263,U.S.A.
Introduction Some cytotoxic a n d / o r antivirai agents bind tightly to the minor groove in DNA without intercalating between DNA bases . For antibiotics distamycin and netropsin, such interaction widens locally the minor groove and induces a bend in the double helix while producing little or no long range distortion [2-5]. High selectivity of drug binding in the minor groove and, typically, preference for AT stretches, result from hydrogen bonding to DNA bases and Van der Waais contacts as well as electrostatic interactions [2-5].
53 These unique features of drug interactions with the minor groove offer the potential for designing agents reading defined DNA sequence ('lexitropsins', Ref. 6). Extensive synthetic programs are being carried out to obtain analogs of prototype minor groove binders with altered DNA sequence recognition and other DNA binding properties a n d / o r improved pharmacological properties [6-9]. One approach is to develop modified forms of the 'classical' minor groove binders. Recently, a series of netropsin dimers has been synthesized wherein two netropsin units are linked with polymethyiene chains of varying length . Since each netropsin subunit may bind to DNA, these dimers are, in principle, capable of bidentate interaction with DNA . Consequently, such dimers may occupy longer stretches on DNA than monomeric drugs and might also bind to recognition sequences that are interspersed with nonbinding regions. Several of these compounds showed significant cytotoxic and antiviral activity . Agents binding to the minor groove affect several DNA processing enzymes such as RNA and DNA polymerases and nucleases [12-14]. Our recent studies demonstrated that minor groove binders can inhibit DNA topoisomerases [15-18]. These enzymes control topology of cellular DNA by executing a nicking-closing reaction. Eukaroyotic topoisomerases ! and 11 are involved in several important cellular processes: DNA replication, transcription, chromosome segregation, and mitosis . Certain antitumor drugs convert topoisomerases to cellular poisons by stabilizing reaction intermediates (cleavable complexes) in whic:~ the enzyme is bound covalently to DNA . In contrast, minor groove binders impede enzyme action without stabilizing the cleavable complexes. We have shown that distamycin, Hocchst 33258 and DAPI inhibit the catalytic activity of isolated topoisomerase II and I [15,16]. In addition, these agents interfere with the stabilization of cleavable complexes by topoisomerase !I targeted drugs (VM-26 and m-AMSA) and a topoisomerase I inhibitor (camptothecin) [17,18]. Based on these findings, we proposed that the action of DNA topoisomerases involves the minor groove [15- 18,21]. Recent investigations confirmed that distamycin blocked the binding of topoisomerase I! and I to DNA [22,23]. Interestingly, anti-topoisomerase properties of distamycin, Hoechst 33258, and DAPI correlated with the size of their DNA binding sites [17,18]. Thus, it seemed reasonable that agents binding to longer stretches of DNA in the minor groove would be more potent inhibitors of DNA topoisomerases, in this study, we evaluated a series of dimeric netropsin analogs capable of bidentate binding (Fig. 1, 1-12). In addition to these linked dipyrrole oligopeptides, mono- and tripyrrole linked oligopeptides were studied (Fig. 1, 13, 14). Both netropsin and its dimers were found to inhibit the activity of isolated enzyme and topoisomerase-media-
ted effects in intact nuclei. The results indicate that bidentate binding can enhance anti-topoisomerase activity of netropsin-related dimeric minor groove binders. However, other factors such as the length of the linker, the number of pyrrole moieties, and the nature of the target (isolated enzyme/DNA versus chromatin in nuclei) are also important for these activities. Materials and Methods
Chemicals. Netropsin was obtained from Sigma (St. Louis, MO). Linked dimers (compounds 12-14, Fig. 1,) were synthesized as described previously . The stock solutions of minor groove binders were made in water and stored at - 20 ° C. VM-26 was a generous gift from Dr. W. Bradner at Bristol-Myers, Wallingford, CN, while m-AMSA was kindly provided by Warner Lambert Pharmaceutical, Ann Arbor, Ml. Camptothecin was obtained from the National Cancer Institute, Bethesda, MD. Stock solutions of VM-26, m-AMSA and camptothecin were made in dimethylsulfoxide and stored at - 20 ° C. Catalytic actiL'ity of topoisomerase !!. The assay measured the ability of topoisomerase 1I to decatenate highly catenated kDNA. The enzyme was isolated from L1210 leukemia cell as described elsewhere . This preparation, while not homogeneous, was devoid of topoisomerase i and endonuclease activities. Kinetoplast kDNA was isolated from Crithidia fasciculata according to England . Reaction mixtures contained 0.1/~g of kDNA, 1.5 U of topoisomerase II , 50 mM Tris HCI (pH 7.5), 50 mM KCI, 5 mM MgCl2, 0.1 mM EDTA, 0.5 mM dithiothreitol and 30 ~g/ml bovine serum albumin, all in 25/zl. Following 15 min incubation at 37°C, the reactions were terminated with 0.5% sodium dodecyl sulfate (SDS), 50 /~g/ml proteinase K and 10 mM EDTA. The samples were electrophoresed on 1% agarose, stained with ethidium bromide and photographed. The negatives were scanned on a computerized Molecular Dynamics laser densitometer. The results were expressed as concentration of minor groove binder inhibiting by 90% the release of decatcnated minicircles (IC~o) from the level seen in the absence of drugs. While gels shown in Figs. 2 and 4 illustrate typical results, the quantitation, such as in Fig. 3, was based on averaged data from two to four independent experiments. Stabi~zation of cleavable complexes of topoisomerase H and !. The assay determined topoisomerase-mediated DNA-protein crosslinks in isolated nuclei using the K+/SDS precipitation technique described previously  except that nuclei were isolated from [ '4C]thymidine labelled mouse fibrosarcoma cells (935.1 line). Briefly, cells were rinsed with phosphate-buffered saline (PBS), trypsinized, washed with PBS and finally
54 with cold ( 4 ° C) nuclei isolation buffer (2 mM KH2PO 4, 5 mM MgCI:, 150 mM NaCi, 1 mM EGTA (pH 6.9)). The cell pellet was resuspended in nuclei isolation buffer and adjusted to 0.3% (v/v) Triton X-100. After 20 min incubation at 4 ° C, samples were centrifuged at 300 x g for 13 min and nuclear pellet resuspended in the isolation buffer supplemented with 0.4 mM ATP at (0.3-0.5). 106 nuclei/ml. Aiiquots of nuclei suspension were incubated with VM-26 or camptothecin with and without minor groove binders as indicated in the legend to the figures. Following drug treatment, nuclei were lysed and DNA bound covalently to proteins was precipitated and its radioactivity determined . The results were expressed as percent of maximal effect (in the absence of minor groove binder) and represent mean values ( + S.E.) from two to four determinations carried out in duplicate. DNA binding. Drug-poi~(dAdT) binding constants were estimated by ethidium bromide displacement as described elsewhere . Binding data for calf thymus DNA have been published previously .
142N~.j~j Ns~.~¢,/~ O CH3 ~).'7 ~N
c~ H N
i41 n:2 (S) n=3 (6) n=4
(g) n,7 00) n=g (11) n=9
(7) n---5 (12) n=lO R,-(CH2)fR,
H ,, , N__ I~'N'~No~y~
Ol H N
Our recent studies showed that minor groove binders distamycin, Hoechst 33258 and DAPI interfere with the catalytic activity of DNA topoisomerases measured as relaxation of supercoiled DNA (topoisomerase 1 and topoisomerase ll) and decatenation of catenated networks of DNA from C. fasciculata (kDNA, topoisomerase 11). it seemed plausible that netropsin dimers consisting of two dipyrrole units capable of interaction with the minor groove would be more potent inhibitors of topoisomerase catalyzed reactions than the agents studied to date. In this investigation, we examined the effects of netropsin and related dimeric minor groove binders. The dimers differed primarily in the length of the aliphatic linker -(CH2) .- (Fig. 1). For n = 2, compounds with different numbers of pyrrole subunits were also examined. The decatenation of kDNA was employed as a basic assay for studying the effects of these agents on isolated topoisomerase I!. The examples of decatenation reactions analyzed by agarose gel electrophoresis are shown in Fig. 2. Untreated substrate consisting of highly catenated kDNA (lane 1) is unable to enter the gel. The action of the enzyme without a minor groove binder releases decatehated minicircles (lane 2). Addition of netropsin or the dimers resulted in a dose-dependent decrease in the amount of the released material. While this inhibitory effect was qualitatively similar for the agents studied, their relative activity varied significantly. For example, the compound with n ffi 0 produced complete inhibi-
NX[[" 10 CH3
Effects of netropsin dimers on catalytic actiuity of topoisomerase H
Fig. 1. Structures of netropsin (!), linked netropsin analogs (2-12) and dimers with 1 or 3 pyrrole/mer (13,14).
tion at concentration as low as 10 p M . In contrast, the compound with n = 5 was only marginally inhibitory at 100 /zM (Fig. 2). It is noteworthy that several minor groove binders at 10/zM tFig. 2) or at lower concentrations (data not shown) produced some stimulation of the decatenation reaction. This stimulation (ranging from 30% up to over 100%) was noted for r.,ost of the netropsin dimers as well as for netropsin at low drug levels (Fig. 2 and data not shown). Similar stimulation of topoisomerase catalytic activity has been observed previously at low concentrations of distamycin [15,22]. The inhibitory potency of the dimers was quantified as drug concentration inhibiting by 90% the release of decatenated DNA (IC90, Fig. 3). It is clearly seen that all the dimeric analogs (except for compound with n = 5) are more inhibitory than netropsin (IC9o = 200 pM). The most active compounds with n = 8, n - - 0 and n = 2 produced 1C90 equal to 7, 8 and 11 p M , respectively. The dimers with n = 3, 4, 6 and 7 with IC90 values ranging from 72 to 150 tzM were only slightly more potent than netropsin. Strikingly, the homolog with n --5 produced only about 50% inhibition at the highest concentration studied. Its IC90 (determined by extrapolation) amounted to 320 ~ M. Netropsin which contains two pyrrole units, is known to interact with four base pairs on DNA [2,3], while distamycin with three pyrroles binds to five base pairs as a result of increased hydrogen bonding [4,5]. Thus,
f 2 3 4 5 6 7 8 9fOil
Fig. 2. The effects of netropsin and selected netropsin dimers on the catalytic activity of isolated topoisomerase II in the decatenation assay, kDNA without (lane 1, top and bottom), and with Iopoisomerase II (lane 2, top and bottom), and with topoisomerase It and minor groove binders as indicated at concentrations of 10/,tM (lanes 3, 6, 9), 25 tiM (lanes 4, 7, 10), and I00/tM (lanes (5, 8. ID. Arrows indicate positions of electrophoresis origin (OR) and released monomeric circles (Mono).
variations in the n u m b e r of pyrrole moieties p e r binding unit could he expected to alter the size of D N A binding site of the dimeric derivatives. It was possible t h a t such c h a n g e s would affect topoisomerase inhibitory properties of the dimers. Fig. 4 shows the results o f the d e c a t e n a t i o n e x p e r i m e n t s for dimers with
Fig. 4. Inhibition of topoisomerase II activity by dimers with I, 2 m 3 pyrrole units/met, kDNA with topoisomerase !! and without drug (control, lane I), same as lane i with compound 13(I i~/rrole/mer) at 100 /xM (lane 2) and 200 ~tM (lane 3). same as lane I with compound 4 (2 pyrroles/mer) at 2 ttM (lane 4), 5 t~M (lane 5) l0 /tM (lane 6) and 25/tM (lane 7). ~me as lane I with compound t4 (3 pyrroles/mer) at 2 ~tM (lane 8), 5 p.M (lane 9) 10 ~tM (lane 10) and 25 /tM (lane ll). Arrows indicate positions of electrophoresis origin (OR) and released monomeric circles (Monomer).
n = 2 and 1, 2, or 3 pyrrole units/mer. The compound with a single pyrrole/me~ was practically inactive at c o n c e n t r a t i o n s as high as 200 p M (Fig. 4), O n the o t h e r hand, the c o m p o u n d with 3 p y r r o l e s / m e r (IC~0 = 5 ~ M ) was markedly more inhibitory t h a n the 2 p y r r o l e / m c r derivative ( I C a p - - 2 5 /~M). Again, some stimulation of the d e c a t e n a t i o n reaction was seen at low c o n c e n t r a t i o n s of the dimers with 1 and 2 pyrroles/mer.
Stabilization of the clear'able complexes of topoisomerase H and l
I.ImCER I . I m e ~ (a) Fig. 3. Concentrations of minor groove binders inhibiting Iopoisomerase !1 catalytic activity by 90% (ICq~) as a function of the linker length (n). The IC,~)values were determined using pooled inhibition data from 2-5 separate experiments.
M i n o r groove binders such as distamycin, Hoechst 33258 a n d D A P I were shown by us to interfere with t o p o i s o m e r a s e - m e d i a t e d D N A lesions induced by VM26 a n d m - A M S A . T h e s e lesions ( D N A - p r o t e i n crosslinks a n d D N A - b r e a k s ) reflect the stabilization of the cleavable complexes of topoisomerase II by VM-26 a n d m - A M S A , in this study, we employed the induction o f D N A - p r o t e i n crosslinks in nuclei by VM-26 to establish the effects of netropsin a n d its dimers. T h e use of nuclei offers the advantage o f a n e n d o g e n o u s topoisomerase acting o n intact c h r o m a t i n structures. Also, elimination o f cell m e m b r a n e barrier a n d cyto-
56 plasmic components removes the variable of drug transport. The effects of netropsin and its dimeric analogs on VM-26-induced crosslinks in nuclei from 935.1 cells are shown in Fig. 5 A, B and C. Both parental compound as well as dimers linked with 0-10 methylene groups Inhibited the stabilization of the cleavable complexes by VM-26. As with the isolated enz~ae, these minor groove binders markedly differed in their effectiveness. Clearly, netropsin was a less potent inhibitor than the dimers with 0 - 4 and 6 - 9 methylene groups. In contrast, dimers with n = 5 and n = 10 were less inhibitory than netropsin. The compound with n = 10 was the least active and produced less than 50% inhibition at
Drug Com:entrotion ( v M ]
Fig. 5. inhibition of topoisomerase |I mediated DNA-protein crosslinks induced in nuclei by Vial-26. Nuclei from [14C]tbymidineprelabelled mouse flbroblast cells (line 935.1) were incubated for 10 rain with 20/*M VM-26 and minor groove binders at varied concentrations given on abscissa, i00 percent value on the ordinate corresponded to the maximal crosslinldng induced by VM-26 in the
absence of minor groov~ binders and amounted to 28.5+ 1.3% of total nuclear DNA, (A) Netropsin (o), compound 2 ( n l 0 , o);, corapound $ ( n - l , A); compound 4 (n-2, e). (13)Compound $ (n-3. e);, compound 6 (n-4, a.); compound 7 (n=5, e); compound II(n- 6, o). (C) Compound9 (n 1 7, o); compound 10(n =8, o); compound !! (n ~ 9, ,,);, compound 12 (n ~ 10, O). The points represent mean values (:t:SE) from several (typically2-4) exper;ments.
140 1t0 .--,
40 40 20 m 0 1 | $ 4 6 I.Jlgam ~
? 8 (n)
Fig. 6. Concentrations of minor groove binders inhibiting by 50% topoisomerase I! mediated DNA-protein crosslinks in nuclei (1C~) as a function of the linker lenglh (n). The IC5o values were determined from the concentration dependence profiles shown in Fig. 5.
the highest concentration used (100 ~M). None of the dimers stabilized the cleavable complexes themselves, as the crosslinking observed in the absence of VM-26 was negligible (data not shown). Comparison of drug concentrations inhibiting VM-26 induced crosslinks by 50% (ICso) is given in Fig. 6. The dimers linked with 0 - 4 methylene groups produced the IC50 values from 10 to 14 p.M and were 5-6-times more active than netropsin (IC5o = G5 ~M). The dimers with n -- 6 - 9 were similarly active with the ICso values ranging from 8 to 27 ~M. By contrast, the compound with n = 5 was approximately one order of magnitude less inhibitory (IC50 = 120 ~ M ) than its immediate homologs. Likewise, the compound with n = 10 was practically inactive with an ICso value determined by extrapolation to be greater than 1000 t~M. Previously, we have found that distamycin, Hoechst 33258 and DAPI affect not only topoisomerase II but also topoisomerase 1 [17,18,21]. To see whether aetropsin analogs similarly interfere with the action of both enzymes, we examined the effect of selected dimers on topoisomerase l mediated cleavable complexes. This effect was measured as D N A - p r o t e i n crosslinks induced in nuclei by a topoisomerase ltargeted drug, camptothecin. The dimers with n = 4 or 6 were more potent inhibitors of topoisomerase I mediated crosslinks than netropsin (Fig. 7A). The homoiog with 5 methylene groups was again markedly less active. The IC~0 values produced in this assay by netropsin and the dimers with four, five and six methylene g~'oups (Fig. 7B) were very close to those obtained with topoisomerase II mediated crosslinks (Fig. 6).
Correlations between the effects on topoisomerase H and DNA binding properties of netropsin-related dimers Having shown that netropsin and related dimers interfere with topoisomerase-mediated reactions, we
57 addressed the question of whether these inhibitory properties are related to DNA binding parameters. Fig. 8,6, and 8B show the IC~0 values and IC50 values, respectively, plotted against association constants, K a, determined for calf thymus DNA. The property of netropsin and its dimeric congeners to inhibit the catalytic activity of topoisomerase II (the IC90 values) is correlated with the association constants. In contrast, the inhibition of the cleavable complexes and the association constants are not correlated (Fig. 8B). No correlation is obtained even after the omission of the most outlying compounds (n = 5 and n = 10) from the calculations. The association constants of minor groove binders are different for various DNA sequences [26,27]. Thus, the K , values for calf thymus DNA represent average affinities over all DNA binding sites. Since netropsin shows a marked preference for A-T regions, the K~
.,o t't ,o."'., ., "'" o/F-'" ol
' Y ~ ......... 4 i
dr to. ~ ' a ' " 4 I
I ~ (x 10")
Ko (x 10")
Fig. 8. Relationshipsbetweenthe abilityof minorgroove bindersto inhibittopoisomerase II mediatedeffects and their associationconstants (Ka) with c~lf thymus DNA (A, B) and pol~dAdT)(C, D). Panels A and C represent isc~latedenzymeinhibition(ICqo)versus Ka. Panels B and D 5ho~:inhibitionof DNA-proteincrmslinksin nuclei (ICso) versus the K,. Straight lines on panels A and D represent significantcorrelations(at P < 0.05). The line in Fig. 8A corresponds to equation Log 1C~o:-1.17"7"1os Ka+10.287. The correlation in Fig. 8D is reflected by equation: Log ICq0= - 3.714. log K, + 10.287. Panels B and C indicate lack of correlation with r-valuesof 0.164and 0.037,respectively.
IO 4O tO
B 14o 12o lOO o
1 t 8 4 S ~ 1 ~
F'~. 7. The effects of minor groove binders on topoisomerase ! mediated DNA-proteincrceslinksinducedin nucleiby 5 ~tM camptotbecin.(A) Concontrationdependence;netmpsin(o), compound6 (n ~ 4, o); compound 7 ( n - 5, • ) ; compound 8 (n ~ 6, [3). tO0
percent value on the ordinate correslmnded to the maximal cromiinkioginducedby camptotbecinin the absenceof minorgroove binders and amounted to 28.0:1:6.2% of total nuclear DNA. (13) Concentrationsof minor groove binders inhibitingby 50% DNAproteincrosslinks(ICso)as a functionof the linkerlensth(n).
values for the dimers were also determined with poly(dAdT). As expected, these values were, in general, several-fold higher than respective values for calf thymus DNA. The relationships between both sets of binding parameters and anti-topoisomerase activities, however, were quite different. Unlike association constants for calf thymus DNA, the K~ values for poly (dAdT) did not correlate with the IC90 v a l u ~ (Fig. 8C). In contrast, a correlation was found with the IC50 values (Fig. 8D). The correlations in Fig. 8A and D are not strong. They are, however, significant with correlation coefficients (r-values) amounting to 0.596 and 0.587, respectively. For 12 compounds (i.e., 10 degrees of freedom, d f = n - 2), the resulting t values (2.517 and 2.345, respectively) exceeded t,~,~a, = 2.28 (for P < 0.P3, d ] = 10). Thus, while the data for individual compounds oscillate, a relatively large number of derivatives inclw~ed in the studies enabled visualization of the trends e~pressed by these correlation. It should be emphasized that in contrast to Fig. 8A and 8D, the data for Fig. 8B and 8C show an almost ideal random relationship (with r-values of 0.164 and 0.03"7, respectively). Igscussiea Agents capable of bidentate interaction with the minor groove constitute a new class of DNA binding drugs. This study characterized for the first time the
58 effects of such drugs on topological enzymes. The results showed that dimeric analogs of netropsin (and also the parent drug) act as inhibitors of DNA topoisomerases. Both netropsin and its dimers were found to suppress: (i) the catalytic activity of isolated topoisomerase II and (ii) topoisomerase mediated DNA-protein crosslinks (topoisomerase cleavable complexes) in nuclei. The effects of netropsin and selected dimers on DNA-protein crosslinks mediated by topoisomerase II and I were qualitatively and quantitatively similar. This observation confirms that minor groove binders interfere with the action of both enzymes in a similar way [15-18]. Therefore, further interpretations based on the effects of netropsin and its dimers on topoisomerase II probably apply to topoisomerase I as well. Our previous studies on model minor groove binders, distamycin, Hoechst 33258 and DAPI indicated that these compounds inhibit DNA topoisomerases by competing with enzyme binding to DNA [15-18]. Netropsin and its analogs appear to inhibit topoisomerases II and I in the same way as these model agents. Netropsin binding to DNA, like distamycin, does not unwind or elongate the double helix [2-5]. Changes in DNA structure (widening of the minor groove and bending of the helix) occur only across the zone of attachment [2,3,5]. Thus, the results of this study suggest that DNA topoisomerases are unable to interact with DNA sites where the minor groove is occupied by netropsin or netropsin analog [15-18]. In addition, these findings lend further support to the idea that the action of DNA topoisomerases involves the minor groove. The nature of the stimulation of topoisomerase catalytic activity observed at low drug levels for several netropsin analogs remains unknown. Previously we and other authors have reported a similar stimulation for 0istamycin [15,22]. Also, distamycin was found to stimulate and inhibit transcription initiation by altering ;.he interaction of RNA pol~merase with DNA [12-13]. Topoisomerase stimulation can be related to the ability of distamycin and other minor groove binders to alter DNA bending [15,22]. Local change of DNA bent angle could enhance enzyme binding to DNA in the vicinity of drug attachment site. Increasing drug levels would eventually block available enzyme binding sites leading to the inhibition of enzyme activity. Monomeric netropsin is a relatively weak inhibitor or" topoisomerase I! catalytic activity and cleav~Ne complexes with IC90 and ICu) values of 200/tM a:~d 70 /~M, respectively. The respective values for a rela',~d drug, distamycin, amounted to 10 ~tM and 8 ~tM [15,17]. Relatively poor inhibitory properties of netropsin do not reflect any gross differences in DNA binding. While netropsin binds to 4 base pairs [2,3,5], an additional N-metbylpyrrole moiety enables distamyein to occupy five base pairs on DNA [4,5] Still, Hoechst 33258 which also occupies four base pairs
[28,29] was substantially more active [17,18] than netropsin. Also, distamycin, netropsin and Hoechst 33258 share common binding sites on DNA as found by footprinting analysis . However, differential interaction with specific sites seems possible as, for example, netropsin competed more efficiently than distamycin with HMG-1 protein binding . Increased size of DNA binding sites is the probable explanation for enhanced anti-topoisomerase activity of several dimers in comparison with netropsin. Each 'tooth' of a linked compound would be expected to protect a DNA stretch equal to that of monomeric netropsin if an appropriate recognition sequence is available. Consequently, the dimers should occupy twice as many base pairs as the monomer. If binding sequences for each unit are not contiguous, the dimers might span even longer tracts depending on the length of the linker and non-binchng interspersing sequences. Footprinting  confirmed that bidentate binding took place for netropsin-related dimers with n >_>_6. The footprinting experiments, however, used a DNA segment that contained only a limited number of strong binc:ing sites (stretches of four or more A or T) with no contiguous 'double' binding sequence. Thus, the compounds with n < 5 might be prevented from bidentate binding by too long interspersing non-binding regions. In contrast, natural DNA used in this study would be expected to contain a variety of binding sequences including longer stretches of ATs. Molecular modeling (Rao and Lown, unpublished data) suggested that drugs with n > 2 are inclined to acquire a conformation that is isohelical with DNA (required for tight fitting of a drug in the minor groove, Ref. 31). Thus, each of the binding units in these compounds has a potential to interact with DNA in a netropsin-like fashion. This is less likely for the dimers with shorter linkers. Still, derivatives with n = 0 or 1 may form a complex in which the second unit associates weakly with DNA rather than binding snugly in the minor groove. Such an 'operational' bidentate binding would also protect a longer stretch of DNA than the monomeric compound. The fact that even the dimer with n = 0 showed an enhanced activity supports this supposition. A surprising feature of anti-topoisomerase action of the linked ne*.ropsin dimers in the nuclear system as well as with the isolated enzyme was the relative inactivity of the derivative with n = 5. Thus, a seemingly minor difference of one methylene group in the linker for n = 4-6 could result in dramatic changes in the inhibitory potency. These differences cannot be explained only by variations in the strength of DNA interaction. Perhaps, in contrast to other homologs, only one unit of the compound with n - 5 interacts with DNA. Poor ,.'nhibitory activity of this compound may result from the combination of both its relatively
59 weak DNA binding and inability to form bidentate complexes. Noteworthy, a I' the dimers with n = 0-10 have identical molecule charge. Hence, the profound changes in antitopoisomerase activity cannot be attributed to a simple cation effect. Dimers with three N-methylpyrroles/mer should cover longer stretches than netropsin-type di-pyrrole/ mcr compounds. Accordingly, a tri-pyrrole dimer was more inhibitory than a di-pyrrole dimer with the same linker. In contrast, the dimer with one N-methyipyrrole/mer was virtually inactive. These observations emphasize the significance of the size of DNA binding sites for the interference with topoisomerases. An important question remains of how many base pairs should participate in drug binding for maximal inhibition? For comparison, the binding domain of Drosophila topoisomerase I1 comprises approx. 25 base pairs where the central 16 base pairs correspond to its cleavage consensus sequence . Similarly, eukaryotic topoisomerase II recognizes an 18 base consensus region, but protects a somewhat wider region against DNase 1 . A significant correlation exists between the ability of netropsin-related dimers to inhibit the catalytic activity of isolated enzyme and their association constants with calf thymus DNA. In nuclei, however, inh~ition of cleavable complex stabilization was not correlated with the DNA association constants. In addition, the dimer with n z 10 (inhibitory with the isolated enzyme) was completeb' inactive in nuclei. Thus, the strength of the binding to calf thymus DNA is indicative of antitopoisomerase effects with isolated enzyme but not in nuclei. The opposite is true when effects of netropsin dimers were related to association constants determined using poly(dAdT). A correlation was observed in nuclei but not with the isolated enzyme. While the explanation of these differences requires further studies, several possibilities exist. The different nature of the DNA target in the nuclear system (chromatin rather than naked DNA) may affect binding parameters such as number of binding sites or binding kinetics. Moreover, available drug a n d / o r en~,~te binding sites in nuclear chromatin may differ from those on isolated DNA. Consequently, the K a values for poly(dAdT) may be more relevant in nuclear systems than the average Ka values determined with isolated calf thymus DNA. On the other hand, interaction with the isolated DNA (at d r u g / D N A ratios as in the enzyme assay) may involve not only strong, purely A-T, binding sites but also weaker sites containing G or C [ii]. The suggestion that the binding of netropsin dimers to A-T may be important in nuclei is supported by other studies that implicated certain AT-rich sequences in the cellular function of topoisomerase !!. Nuclear matrix, where the enzyme is located, is at-
tached to very long AT domains (approx. 200 base pairs) termed scaffold- or matrix-associated regions (SAR or MAR, Refs. 34-36). T h e ~ regions bind preferentially and cooperatively topoi~merase Il  and contain cleavage sites for this enzyme . Consistent with our findings, topoisomerase II binding to these sequences can be prevented by distamycin  but not by chromomycin, an antibiotic that also bin:is to the minor groove but of GC rich DNA. Simila~'iy, bindir,g of SAR DNA to nuclear matrix can be blocked by distamycin . Thus, SAR (MAR) DNA is a likely target for anti-topoisomerase actions o¢ netropsin dimers in nuclei. In agreement with such a notion, another iexitropsin that requires G or C as central bases in its binding site (due to the presence of tfiazoles instead of pyrroles)  was virtually inactive in nuclei (our unpublished results). Overall, this investigation shows that the action of DNA topoisomerases can be manipulated by using AT-specific minor groove binders capable of bidentate interaction with DNA. Our resu"s indicate that the potential for bidentate binding can substantially enhance anti-topoisomerase activity of a minor groove binder. Studies arc now underway to design more potent inhibitors of topoisomerases based on the concept of polydentate interaction with the minor groove. Besides their pharmacological potential, such agents should b¢ useful as probes in the studies on DNA topoisomerases. Aekl~uts This study ~ a ; supported by grants (to T.A.B.) from the American Cancer Society, Grant CH-293, and the National Cancer Institute Grant CA-16056 and (to J.W.L.) from the Medicinal Research Council of Canada and the National Sciences and Engineering Research Council of Canada. References
1 Neidle. S.. Pearl. L.H. and Skelly,J.V. (1987) Biochcm. J. 243. 1-13. 2 Kopka,M.L., Yoon.C., Gtxxls¢ll,D.. Pjura.P. and Dickerson.E. (1985) ,I, Mol. Biol. 183,553-563. 3 Coil, M. Aymami.J.. Van der Mar¢l. G.A.. Van Boom. LH.. Rich, A. and Wang.A.HJ. 0989) Bio:hemisto'28. 310-320. 4 Coil,M.. Frederick.C.A.. Wang.A.H.J. and Rich.A. (1987)Proc. Natl. Acad. Sci. USA 84. 8385-8389. 5 Kopka.M.L.. Yoon.C, C_mmtsell,D.. Pjuta. P. and Dickerson.E. tl985) Proc. Natl. Acad.So. USA 82. 1376-1380. 6 [,own.J.W. (1988) Ami-CancerDrug Design3. ?25-40. 7 Denan. P. (1986)Science.23~ 464-471. 8 Arcamofle.F.. Lazzari. F_...Menozzi.M., Soranz~.C. and Verini. M.A. (1986) Anti-CancerDrug Design I. 235-244. 9 Dasgnpta.D.. ParracL P. and Sasisekharan.V. ti987) Biochemistl~. 26. 6 3 8 1 ~ . 10 Lown. J.W. Krowicki. IL. Baizarin/.J.. Ne~anan. R.A. and De Clercq. E. (1989)J. Med. Chem. 3" ".2368-2375.
60 11 Kissinger, K.L., Dabrowiak. J.C. and Lown, J.W. (1990) Chem. Res. Toxicol 30, 162-168. 12 Bruzik, J.P., Auble, D.T. and DeHaseth. P.L. (1987) Biochemistry 26, 950-956. 13 Levy, A., Weisman-Shomer, P. and Fry. M. (1989) Biochemistry 28, 7262-7267. 14 Shishido, K., Sakaguchi, R. and Nosoh, Y. (1984) Biochem. Biophys. Res. Commun. 124, 388-392. 15 Woynarowski, J.M., McHugh, M., Sigmund, R D . and Beerman, T.A. (1989) Mol. Pharmacol. 35, 177-182. 16 McHugh, M., Woynarowski, J.M., Sigmund, R.D. and Beerman, T.A. (1989) 3iochem. Pharmacol 38, 2323-2325. 17 Woynarowski, J.M., Sigmund, R.D. and Beerman, TA. (1989) Biochemistry 28, 3850-3855. 18 McHugh, M. Sigmund, R.D. and Beerman, T.A. (1990) Biochem. Pharmacol. 39, 707-714. 19 Wang, J.C. (1987) Biochim. Biophys. Acta 909, 1-9. 20 D'Arpa, P. and Liu, L.F. (1989) Biochim. Biophys. Acta. 989, 163-177. 21 Beerman. T.A., Woynarowski, J.M. and McHugh, M.M. (1991) in Topoisomerases in Cancer Chemotherapy (Kohn, K.W., ed.), Oxford University Press, ~,]ew York, in press. 22 Fesen, M. and Pommier, Y. (1989) J. Biol. Chem. 264, 1135411359. 23 Mortensen, UH., Stevnsner, T., Krogh, S., Olesen, K., Westergaard, O. and Bonven, B.J. (1990) Nucleic Acids Res. 18, 19831989. 24 Englund, P.T. (1978) Cell 14, 157-168.
25 Rao, K.E., Shea, R.G., Bathini, Y. and Lown, J.W. (1990) Anticancer Drug Des!gn 5, 3-20. 26 Zimmer, C., Marck. C.. Schneider, C. and Guschlbauer, W. (1979) Nucleic Acids Res. 6, 2831-2837. 27 Ward, B., Rehfuss, A., Goodisman, J. and Daorowiak, J. (1988) Biochemistry 27, 1198-1205. 28 De C.T. Carrondo, M.A.A.F., Coil, M., A:'mami, J., Wang, A.H.J., Van der Marel, G.A., Van Boom, J.H. and Rich, A. (1989) Biochemistry 28, 7849-7859. 29 Harshman, ICD. and Dervan, P.B. (1985) Nucleic Acids Res. 13, 31-112. 30 Wegner, M. and Grummt, F. (1990) Biochem. Biophys. Res. Commun. 166, 1110-1117. 31 Goodsell, D. and Dickerson, R.E. (1986) J. Med. Chem. 29, 727-733. 32 Lee, M.P., Sander, M. and Hsieh, T.-S. (1989) J. Biol. Chem. 264, 21779-21787. 33 Spitzner, J.R. and Muller, M.T. (1988) Nucleic Acids Res. 16, 5533-5556. 34 Adachi, Y., Kas, E. and Laemli, U.K. (1989) EMBO J. 8, 39974006. 35 Kits, E., lzaurralde, E. and Laemli, U.K. (1989) J. Mol. Biol. 210, 587-589. 36 Pommier, Y., Cockerill, P.N., Kohn, K.W. and Garrard, W.T., (1990) J. Virol. 64, 419-423. 37 Lee, M., Coulter, D.M., Pon, R.T., Krowicki, K. and 1.own, J.W. (1988) J. Biomol. Struct. Dyn. 5, 1059-1087.