J. Mol. Biol.
(1991)
217,
53-62
Drosophila Topoisomerase II-DNA Interactions by DNA Structure Michael Tao-shih
are Affected
T. HowardI, Maxwell P. Lee’ Hsieh2 and Jack D. Griffithlf-
1Lineberger Cancer Research Center Chapel Hill, NC 27514, U.S.A. 2Department of Biochemistry Duke University Medical Center Durham, NC 27’710, U.S.A. (Received
23 May
1990; accepted 5 September
1990)
The binding of purified Drosophila topoisomerase II to the highly bent DNA segments from the SV40 terminus of replication and C. fasciculata kinetoplast minicircle DNA (kDNA) was examined using electron microscopy (EM). The probability of finding topoisomerase II positioned at or near the bent SV40 terminus and Crithidia fasciculata kDNA was two- and threefold higher, respectively, than along the unbent pBR325 DNA into which the elements had been cloned. Closer examination demonstrated that the enzyme bound preferentially to the junction between the bent and non-bent sequences. Using gel electrophoresis, a cluster of strong sodium dodecyl sulfate-induced topoisomerase II cleavage sites was mapped to the SV40 terminus DNA, and two weak cleavage sites to the C. fasciculata kDNA. As determined by EM, Drosophila topoisomerase II foreshortened the apparent length of DNA by only 15 base-pairs when bound, arguing that it does not wrap DNA around itself. When bound to pBR325 containing the C. fasciculata kDNA and the XV40 terminus, topoisomerase II often produced DNA loops. The size distribution was that predicted from the known probability of any two points along linear DNA colliding. In vitro mapping of topoisomerase II on DNA whose ends were blocked by avidin protein revealed that binding is enhanced at sites located near a blocked end as compared to a free end. These observations may contribute towards establishing a framework for understanding topoisomerase II-DNA interactions.
1. Introduction
In vitro, the addition
of sodium dodecyl sulfate to complexes of topoisomerase II and DNA in a double strand DNA (dsDNA) break (Liu et al., 1983; Sander & Hsieh, 1983). Mapping of such topoisomerase II cleavage sites has shown that some sites are more efficiently cleaved than others. A consensus cleavage sequence for Drosophila (Sander & Hsieh, 1985) and chicken erythrocyte (Spitzner & Muller, 1988) topoisomerase II has been determined by sequencing in vitro cleavage sites. The degree of correlation between a DNA sequence and the cleavage consensus does not always accurately predict the intensity of cleavage by topoisomerase II. This suggests that DNA sequence may only be one of several parameters that determine the (SDSI) results
The topological state of DNA has important implications for many aspects of DNA metabolism including transcription, recombination, chromosome segregation, repair and replication. The topoisomerase enzymes can profoundly affect these processes by altering DNA topology (Wang, 1985). Type II topoisomerases break both strands of the double helix and catalyze a strand passage and religation event (Maxwell & Gellert, 1986; Hsieh, 1990). The unique ability of topoisomerase II to pass DNA strands makes it essential for the separation of newly replicated DNAs (DiNardo et al., 1984; Holm et al., 1985; Yang et al., 1987; Uemura et al., 1987).
t Author addressed.
to whom
all correspondence
should
$ Abbreviations used: SDS, sodium dodecyl dsDNA, double strand DNA; bp, base-pair(s); kinetoplast DNA; EM, electron microscopy.
be
53 0022-2S36/91/010053-10
$03.00/O
0
1991
Academic
sulfate;
kDNA;
Press
Limited
54
,U. T. Howard
strength of topoisomerase II cleavage. Cleavage of DNA can be enhanced by the addition of anti-tumor drugs such as the acridines, anthracylines, elliptitines and epipodophyllotoxins (for reviews, see Glisson & Ross, 1987; Liu, 1989). In viwo, it appears that topoisomerase 11 is the target for these drugs (Yang et al., 1985a,b; Rowe et al., 1986), which act by trapping a putative intermediate in topoisomerase II-mediated DNA strand passage. This topoisomerase II-DNA complex stabilized by drugs is termed the “cleavable complex”. It has been suggested that t,he collision of replication enzymes with the cleavable complex generates a DNA lesion which is lethal to actively replicating cells. Topoisomerase 11 cleavage and DNA footprinting have been used extensively to predict sites of topoisomerase 11 activity. In most cases, strong binding correlates with strong cleavage (Liu & Wang, 1978a,b; Klevan & Wang, 1980; Fisher et al., 1981; Morrison & Cozzarelli, 1981; Lee et al., 1989a,b). Although one site of topoisomerase 11 binding has been identified that is not a strong cleavage site (Kirkegaard & Wang, 1981). To understand further the parameters that determine DNA binding, cleavage, strand passage and religation by topoisomerase II, its will be necessary to charact,erize the physical properties of the enzyme-DNA complex. Some proteins interact with DNA by wrapping DNA about themselves. The histone proteins preferentially assemble at sites of DNA bending (Drew & Travers, 1985; Satchwell et al., 1986; Hsieh & Griffith, 1988): due to the decreased energy required to coil an already bent DNA around the histone core. Kleven & Wang (1980) have shown that DNA gyrase will form a “gyrasome” on DNA with 145 base-pairs (bp) of DNA wrapped about the enzyme. If other topoisomerase 11 molecules also wrap DNA it would be expected that segments of DNA containing severe sequence-directed bends might act as preferential binding sites for these enzymes. Alternatively, DNA bending might influence the positioning of topoisomerase II in an indirect manner by creating DNA loops. A 223 bp DNA segment from Crithidia fasciculata kinetoplast DNA (BDNA) contains a severe sequence-directed bend, with the individual molecules forming almost perfect circles (Griffith et aE., 1986). When this segment of bent DNA is inserted within non-bent DPL’A, loops are formed bringing two segments of DNA into close proximity. If topoisomerase 11 prefers t,o bind DNA at, sites where two DNAs come in contact then DNA bending could create a strong topoisomerase II binding site at the base of such a loop. Two topoisomerase II enzymes have been visualized by electron microscopy (EM) bound to DNA: Escherichia c&I gyrase (Moore et al.: 1983) and phage T4 topoisomerase (Moore, 1982; Kreuzer & Huang, 1983); both enzymes were seen to locate at the base of DNS loops. Here we have employed EM to further characterize the topoisomera,se II-DNA complex, and to directly visualize the position of Drosophila topoisomerase II molecules on DNA containing sequence-
et al. directed bends. Also; phoresis to map strong within these DNAs.
we have used gel electrctSDS-induced cleavage sit,es
2. Materials and Methods (a) Proteins and DNA Plasmid DNAs were isolated by the alkaline iysis method (Maniatis et al.; 1982). Restriction enzymes were purchased from Pl’ew England Biological laborat,ories (XEBL) or Biological Research Laboratories and used according to manufacturers’ specifications. Restriction digested DEA was end-labelled by a fill-in reaction using T4 DNA polymerase (PI’EBL) and biotinylated deoxycytosine triphosphate purchased from EPU’ZO Diagnostics. Inc.: followed by incubation with avidin (C. Bortner & J. Griffith, unpublished results). Topoisomerase II was purified as described (Hsieh. 1983) with modifications (Lee et al., 1989a). (b) Topoisomerase of 10 50 10
II binding
reaction8
Binding reactions were carried out in a volume of 20 ~1 10 mivf-Tris. NC1 (pH 7.9), 50 m&l-KCi, @I m&r-EDTA. miv-MgCl,. 1.25 m;M-ATP. with 10 ~M-D?U’A and n;ll-topoisomerase II. The reactions were incubated for min at 30°C. cc) Cleavage reaction
The plasmid pJGCl/svt (Hsieh & Griffith: 1988) !inearized by treatment with H1:ndTIT was ineubat.ed with topoisomerave II under binding conditions for 10 min at 30°C. SDS was then added to 1 y0 (w/v) and incubation continued for 2 min. The DNA-protein complexes were digested with 50 ,ug of proteinase K/ml for 1 h at 45°C. and electrophoresed on I “;o (W/V) agarose gels. Mock cleavage reactions were identical except that EDTA was added to 25 m&r in place of 1 v/0 SIX. DN.4 was transferred to a Zeta-Probe membrane using the alkaline transfer method and probed as described by Reed B Mann (1985). Autoradiographs were obtained by exposure of Koda,k X-Omat AR film.
The t,opoisomerase IT-DSA complexes were fixed bag adding glutaraldehyde Tao 1 O/b, for 5 min at 30°C. The samples were then chrornatographed over 2-ml columns of Sepharose 4B and prepared for EM as described (Griffith 6t Christiansen, 1958). including rotary shadowca,st,ing with tungsten. Samples were examined in a Philips 400 EM TLG. The lengt,hs of molecules were measured from images on electron micrographs using a Summagraphics digitizer coupled to an IBM PC-AT comput’er programmed with software developed by Dr Richard R’ubin.
(a) To;r,~isomerase il pref~renbially binds clear t/be bent D-VA4 elements fro’mb C. fasciculata Line~o$~~t minicircles and the SY40 terwhus of replimtim The plasmid pJGCll/svt (Fig. l(a)) consists of a 223 bp fragment from C. fasciculata kDNA and a 1216 bp fragment, from the simian virus 40 (SF’4O) terminus region cloned into separate sites in
Drosophila
Topoisomerase
II-DNA
I 0;;
interactions
ShI&vun
(a)
(b)
SV40 3204
bp
PstI
55
terminal
DNA
T
P.tI
1988
bp
Avail
Figure 1. Restriction map of pJGCl/svt, pPK20l/cat and the SV40 terminus of replication. (a) pJGCl/svt contains 223 bp of bent kDNA from C. fasciculata and 1216 bp of bent SV40 DNA cloned into pBR325 (Hsieh & Griffith, 1988). (b) pPK2Ol/cat contains the bent 223 bp kDXA cloned into theBamH1 site of pSP65 (Kitchin et al., 1986). (c) The 1216 br) fragment of SV40 from base-pair 1988.3204, containing the terminus of replication and a sequence directed bend (&iek& Griffith, 1988). pBR325 (Hsieh & Griffith, 1988). These segments have been shown to have strong sequence-directed curvature. The 223 bp kDNA forms a tight loop (Griffith et al., 1986) and the 1216 bp fragment is bent through nearly 180” (Hsieh & Griffith, 1988) as seen by EM. The pJGCl/svt DNA was linearized with SphI and XalI, and end-tagged with a biotinavidin complex at the Sal1 site (see Materials and Methods). Purified DrosophiZa topoisomerase II was incubated with end-labeled pJGCl/svt DNA in a buffer containing ATP and Mg’+, fixed with glutaraldehyde, and prepared for EM (Fig. 2). Topoisomerase II appeared bound to the DNA at a single site or at two sites simultaneously. When topoisomerase II bound two sites simultaneously the intervening DNA formed a loop and these loops showed a broad spectrum of sizes. Approximately 100 molecules were measured from the untagged end of the DNA to each topoisomerase II molecule and the results displayed (Fig. 3). At this level of resolution, the enzyme appears to have an affinity for two large areas of DNA spanning the C. fasciculata kDNA and SV40 terminal sequences. It was calculated that topoisomerase II bound the C. fasciculata kDNA and SV40 terminus DNA three- and twofold more often, respectively, than the surrounding pBR325 sequences. To establish the pattern of topoisomerase II binding more precisely in the vicinity of the bent DNA, a PvuII-SphI-digested DNA fragment from pPK201/ cat (Kitchen et al.; 1986; Fig. l(b)) containing the bent C. fasciculata kDNA; and an AvaII-PstI fragment of the SV40 terminus region (Fig. 1 (c)), were isolated and uniquely end-tagged with biotin-avidin at the PvuII and PstI sites, respectively, incubated with topoisomerase II under binding conditions, fixed with glutaraldehyde and prepared for EM (Fig. 4(a) and (b)). Topoisomerase II binding was mapped by measuring DNA molecules from the untagged end to each topoisomerase II particle. The results revealed that for both DNAs there was a
strong localization of the enzyme to the junctions of straight and bent DNA (Fig. 5(a) and (b)). It was noted that 20% of the C. fasciculata kDNA molecules appeared looped with topoisomerase II particles bound where the strands crossed and 87% of these loops contained the highly bent kDNA.
(b) SDS-induced cleavage of DNA by topoisomerase II is more prevulent in the XV40 terminus than in the bent C. fasciculata kDNA Purified Drosophila topoisomerase II was incubated with linear pJGC/svt DNA, as described above, and cleavage induced by the addition of SDS to 1 yc. The DNA fragments were then treated with Proteinase K, electrophoresed on a 1 y. (w/v) a,garose gel and transferred to a nylon membrane (see Materials and Methods). Strong cleavage sites on pJGCl/svt DNA were mapped by probing the membrane with a HindIII-SaZI-digested fragment from pJGCl/svt that overlaps a unique end of pJGCl/svt DNA linearized with HindIII. A number of sites were identified, and the strength of cleavage was determined by comparing band intensities. A cluster of strong cleavage sites mapped to the SV40 terminus region and two weak cleavage sites to the C. fasciculata kDNA sequences (Fig. 6). Comparison of topoisomerase II binding and cleavage sites within the C. fasciculata and SV40 DNAs revealed that strong binding at the C. fasciculata kDNA site did not correspond to strong cleavage, but that the SV40 terminus region, which contains a very strong cluster of cleavage sites, did show a corresponding increase in binding (compare Figs 3 and 6). Sequence analysis of the SV40 terminus region shows two sites with 86% homology to the Drosophila topoisomerase II cleavage consensus located at’ position 2297 and 2549. The C. fasciculata kDNA did not contain significant sequence homology to the consensus.
Al. T. Howard
56
et
al.
Figure 2. Visualization of Drosophila topoisomerase II bound to DK4. A 1350 bp &“$I-&x/I fragment of pJG:Cljsvt DSA uniquely end-labeled at the S&I site with a biotin-avidin conjugate was incubated with Drosophila topoisomerase II. The proten-DNA complexes were fixed and prepared for EM by mounting onto thin carbon films and rotar? shadowcasting with tungsten (see Materials and IMethods). Shown in reverse contrast. The bar represent!s 0.1 pm.
(c) Drosophila,
topoisomerase around
/I itself
does
not
wrap
DNA
Drosophila topoisomerase IT wa,s incubated with DNA containing the BamHT-&&I HSP8787 intergenie heatshock region (Lee et al.; 19896) under binding reaction condit’ions, fixed with glutaralde-
hyde and prepared for EM (see Xat’eriais and Methods). Measurements were made of the HSP87A7 intergenie heat shock DNA, which contains four regions enriched in topoisomerase II cleavage sites (Lee et al.: 19896). with and Without Drosophila topoisomerase II bound. The results (Fig. 7) demonstrate only 50 L& (1 A=O.l nm) 01
Drosophila
Topoisomerase
223
II-DNA
interactions
57
bp kDNA
40 Percentage
60 distance
Figure 3. Positioning of topoisomerase II on pJGCl/svt DNA. The position of 205 topoisomerase II particles on the 7350 bp pJGCl/svt DPU’A was determined by measuring molecules on the electron micrographs, prepared as described in the legend to Fig. 2, from the &&I site to each bound protein. The distance along pJGCl/svt DNA is given in percentage distance from the &&I site towards the Sal1 site. The position of the bent 223 bp C. fax&data kDNA and the bent 1216 bp SV40 DPU’A from the terminus of replication (svt) are indicated.
about 15 bp foreshortening of the DNA when bound by a single topoisomerase II molecule. This agrees with footprinting data which demonstrate only 20 to 30 bp of protection by Drosophila topoisomerase II (Lee et aZ., 19893).
(d) Drosophila
topoisomerase
II
loops DNA
The binding of Drosophila topoisomerase II to pJGCl/svt DNA produced loops of a broad spectrum of sizes (Fig. 2). Topoisomerase II-associated DNA loops were also observed when binding was carried out in the presence of adenosine 5’-0-(3thiotriphosphate) or in the absence of any nucleotide co-factor, demonstrating that the formation of loops is not dependent on the binding or hydrolysis of ATP. These loops varied greatly in size from 200 to 2000 bp with the modal number of loops measuring about 500 bp (Fig. 8). The lack of dependence of loop size on ATP hydrolysis argues against a model in which the loops are generated by the translocation of the DNA through the enzyme complex from a single site. An alternative model is one in which the enzyme binds to a single site on DNA and then captures a second distant DNA site through a collision event. If this model is correct, then the distribution of loop sizes should reflect the probability of two sites on DNA coming in contact with each other by collision. Clearly if two sites are very close, their contact will be excluded by the stiffness of the DKA and if they are very far apart, their contact will also be unlikely. The optimal separation along DNA for achieving contact through a diffusion-driven process appears to be about 500 bp as revealed by cyclization studies
(Shore et al., 1981; Shore & Baldwin, 1983), and the size distribution of DNA loops associated with topoisomerase II closely agrees with this cyclization probability (Fig. 8). Loops of DNA associated with topoisomerase II were also observed when linear pJGCl/svt was incubated with topoisomerase II under binding conditions, and prepared for EM by a method employing freeze-drying and no chemical fixation (Thresher & Griffith, 1990), excluding the possibility that looping is an artifact of fixation. An identical distribution of loop sizes was observed when pBR325 was bound by topoisomerase II (data not shown).
(e) Topoisomerase avidin-bound
II
preferentially ends
localizes of
near
DNA
Topoisomerase II was found more frequently at sites located near an end blocked by a biotin-avidin conjugate than at sites near a free unblocked end (Fig. 9). To further investigate this, a DNA fragment containing the BamHI-XphI HSP87A7 intergenie heatshock region was bound by avidin at either end to biotinylated deoxycytosine (see Materials and Methods) The DNA was then incubated with Drosophila topoisomerase II, fixed with glutaraldehyde and prepared for EM. When the avidin complex was located at the BamHI end, binding of topoisomerase II at the cleavage site nearest the avidin was approximately eightfold higher than when the BamHI end was unblocked. Likewise, if the avidin was located at the XphI end, binding was approximately eightfold higher at the site nearest the SphI end than when the same end was unblocked.
58
M. T. Howard
et al.
Figure 4. Visualization of Drosophila t,opoisomerase II bound to C. jascicuhta iaDX8 and SV40 t,ermintiti i3S.A (a) pPK2Ol/cat DNA was digested with SphI and PvuII. t,he 890 bp fragment produced was labeled with biotic and avidin at the Pl;uII end, incubated with topoisomerase II and mounted for EM as in Fig. 2. (b) SV40 DK’A was digested with AvaT1 and P&I, and the resulting 1191 bp DNA was labeled with biotin and aridin at the P&I end. bound by topoisomerase II and prepared for EM as described in the legend to Fig. 2. Shown in reverse contrast,. The bar represents 0.1 pm.
4. Discussion In viz~o, it is likelv that’ a number of factors determine the DNA binding, cleavage and catalytic sites for topoisomerase IT. These factors might include DNA sequence, structure and the presence of other DNA binding proteins. The pattern of druginduced cleavage of DT\‘A by topoisomerase TI is different in z&o than t,hat observed in vitro, with some cleavage sites being enhanced and others decreased in intensity (Yang et ccl., 1985a,b; Udvardy et al.; 1986). Here, we have investigated one unusual DNA structure, sequence-directed bending, for its effect on topoisomerase II binding and cleavage of DNA. The in vitro binding of Drosophila topoisomerase TT to the SV40 terminus of replication and C. fasciculnta kDNA as shown by EM revealed that, these bent DNAs have a higher
affinity for the enzyme than surrounding vector sequences. SDS-induced cieavage of DXA by topoisomerase IT demonstrated that the C. ,jaksascictda~ta cleaved, whereas the bent kDNA wa.s not efficiently SV40 terminus DNA was found to contain a, cluster of strong cleava,ge sites. The relationship between sites of topoisomera,se II binding, cleavage and catalysis is not well understood, A DNA gyrase binding site that’ is not a cleavage site has been mapped by Kirkegaard 8 Wang (1981). This demonstrates tha,t topoisomerase II binding sites do not necessarily define sites of and the cleavage. In the case of the SC40 terminus C. fasciculata kDNAs the presence of a strong cleavage site may require the conjunction of 2; strong binding sit’e, as defined by the structure of DNA4; and a site of strong homology to t,hc topoisomerase TT consensus cleavage Lbosophila
TopoisomeraseII-DNA
Drosophila
interactions
1
59
Bentsvt DNA
Bent kDNA
40 Percentage
60 distance
1
li
h-L 80
Percentage
diston ce (bl
(0)
Figure 5. Positioning of Drosophila topoisomerase II on the bent C. fascicuZata kDNA and SV40 terminus of replication. (a) The position of topoisomerase II on the 890 bp fragment from pPK20l/cat containing the bent kDPjA was determined by measuring molecules on electron micrographs such as those shown in Fig. 4(a) from the SphI site to each bound topoisomerase labeled as described in Distances are represented bent DNA is represented
II. (b) Positioning of topoisomerase II on the 1191 legend to Fig. 4(b); was measured from the AvaII as percentage distance from the unlabeled end towards by the horizontal line above the bars in (a) and (b).
the
topoisomerase II-DNA interactions will hopefully clarify the relationship between binding, cleavage
sequence. The C. fasciculata kDNA does not contain regions of sequence homology to the consensus and demonstrates only weak cleavage. The SV40 terminus DNA, however, contains two sequences 86% homology to the Drosophila topoisowith merase IT cleavage consensus sequence and is cleaved efficiently by topoisomerase II. This
and catalysis.
Several proteins, including histones and DNA gyrase, have been shown to wrap DNA around a protein core even though the bending of DNA in such a manner is energetically unfavorable. The presence of sequence-directed DNA bending will lower the energy required to bend and hold DNA in such a conformation, and therefore could create a favored binding site for DNA-wrapping proteins. To date, there is no evidence that DNA is wrapped
observation is of particular interest in the light of the findings of Fields-Berry & DePamphilis (1989), which suggest that the SV4Q terminus of replication can inhibit the decatenation activity of topoisomerase II in wivo. Further investigation into
Lane
:
t
2
3
4
bp SV40 terminal DPjA, isolated and end to each bound topoisomerase II. the labeled end. The position of the
5
6
7
Figure 6. Topoisomerase II cleavage mapping of C. fasciculata kDNA and SV40 DXA from the terminus of replication. Topoisomerase II cleavage of HindIII-digested pJGC/svt was induced by addition of SDS to in vitro topoisomerase II binding reactions. A Southern blot of cleavage fragments was hybridized with a radioactive probe that overlaps a unique end of pJGCl/svt. Lane 1: BamHI-digested pJGCl/svt; lane 2, EcoRI-digested pJGCl/svt; lane 3, HindIIT-digested pJGCl/svt mock (EDTA added to 25 rnM in place of SDS) topoisomerase II cleavage reaction; lanes 4, 5 and 6, topoisomerase II cleavage of HindIII-digested pJGCl/svt at a DNA to protein molar ratio of 1 : 100, 1 : 50 and 1 : 25, respectively; lane 7, HindIII-digested 1 size marker. The center of the kDEA and SV40 terminus bend are represented by arrows.
60
M. T. Howard
et al. 30
20 “n -8 6 2 E ; IO
ib)
0
Id00 Loop
2600 size
i bp)
Figure 8. Comparison of Droso#la topoisomerase IIassociated DNA loop size with the Shore & Baldwin (1983) probability of cyclization. The number of loops of a measured length is represented by the open boxes (0) and the J-factor for probability of cyclization of a linear DlUA of a certain size with compiementary overhanging ends is represented hy the filled boxes (m).
0.12
0.14 Length
0.16 (pm)
Figure 7. Determination of the linear foreshortening of DKA by the binding of Drosophila topoisomerase II. of DSA containing the (a) Length histogram RamHI-&‘@ fragment from the heat shock intergenic region of HSP87A7 (Lee et al., 19896) bound by a single topoisomerase II molecule. (b) Length histogram of the same DNA free of protein. Samples in (a) and (b) were prepared for EM as described in the legend of Fig. 2, and lengths were measured from molecules on electron micrographs.
around eukaryotic topoisomerase II molecules. DBase I footprinting of the Drosophila topoisomerase II revealed a protected region of 20 to 30 bp with no indication of hypersensitive sites at a 10 bp interval (Lee et aZ., 1988, 1989b). In this study, the lengths of DNA bound by Drosophila topoisomerase II and protein-free DNA have been compared to determine if the DNA is wrapped around the enzyme. These results and the EM observations demonstrated that Drosophila topoisomerase II does not wrap DNA about itself and consequently would not be affected by DKA bending in the way proposed for nucleosome-like interactions. Proteins that bind a single DNA at two sites simultaneously must bend DNA to bring these sites into close juxtaposition. Curvature of the intervening DSA would facilitate this process. The C. fascicuZata bent kDNA formed topoisomerase II-associa,ted DNA loops in which 8’i oh of the molecules contained the bent DNA within the loop. The
binding of topoisomerase II to the junction of straight and bent DNA from C. fasciculata kDT\‘A may reflect t,he propensity of t,opoisomerase II to bind at the base of a DI\;A loop where these two regions will be in close proximity due to the bent, nature of the DNA. Topoisomerase II-associar5ed loops were observed within the SV40 terminus DNA much less frequently than within the C. fasciculata kDNA. A probable explanation for this difference may come from the observation that t,he 6’. fasciculata kDNA bends through nearly 360” (Griffith et al., 1986) whereas the SV40 terminus DNA appears to bend only about 180” (Hsieh & Griffith, 1988). If binding DNA at two separate siLes simultaneously increases the stability of t,he t’opoisomerase II-DNA complex, then the lower affinity of the SV40 terminus in compa,rison to the C. fasciculata kDNA for topoisomerase II may be explained by the probability of looping DSA at each site. An alternative explanation for the increased binding of topoisomerase II at, the junctions of straight and bent DNA resides in the evidence that suggests that topoisomerase II can move along DSA (Osheroff. 1986). We have observed that @rosophila topoisomerase II was more frequently found near an avidin-bound end of linear DNA t&n near a free end. This may reflect linear diffusion of the enzyme along DNA and t’he inability of the enzyme to be released from an end of DXA blocked by avidin. This interpretation must be approa,ched cautiously as a non-specific affinity of topoisomerase IT for avidin could also explain this finding. If such
TopoisomeraseII-DNA
Drosophila
interactions
61
We thank John Benson for helpful discussion. This work was supported by grants from the National Institutes of Health (GM31819, GM42342) to J.D.G., and (GM29006) to T.H., and the American Cancer Society (NP-583) to J.D.G. References
0
20
40 Percentage
60 distance
80
100
Figure 9. Enhanced localization of Drosophila topoisomerase II near avidin-bound ends of DNA. (a) DNA containing the BamHI-SphI HSP87A7 intergenic heatshock region (Lee et al., 1989b) was labelled at the BamHI end with biotin and avidin, incubated with topoisomerase II and prepared for EM as described in the legend to Fig. 2. The position of the topoisomerase II molecules was measured from the BamHI end of the DNA. (b) The same DNA fragment was labeled at the #phi end, reacted with topoisomerase II and the position of topoisomerase II molecules was determined as described for (a).
a passive diffusion model is correct,, enhanced binding of topoisomerase II could occur at, or near, sites where movement of topoisomerase II along DNA would be inhibited. These blocks to movement might include unusual DNA conformations or other proteins bound along the DNA. The primary recognition sequence has been determined for many DNA-binding proteins. Little is known about how the local DNA structure and other proteins nearby may influence recognition of these sequences by specific DNA-binding proteins. The local environment is of particular interest when the recognition sequence is not strictly defined but is part of a larger group of sequencesrelated to the consensus. Here, we have demonstrated the DNA-topoisomerase II interaction can be affected by DNA structure (bending), and have proposed that DNA looping and diffusion of topoisomerase II along DNA are characteristics of this interaction that may be involved in determining sites of topoisomerase II binding, cleavage and catalysis.
DiNardo, S., Voelkel, K. & Sternglanz, R. (1984). Proe. Nat. Acad. Sci., U.S.A. 81, 2616-2620. Drew, H. R. & Travers, A. A. (1985). J. Mol. BioZ. 186, 773-790. Fields-Berry, S. C. & DePamphilis, M. L. (1989). NucZ. Acids Res. 17, 3261-3273. Fisher, L. M., Mizuuchi, K., O’Dea, M. H., Ohmori, H. & Gellert, M. (1981). Proc. Nut. Acad. Xci., U.S.A. 78, 41654169. Glisson, B. S. & Ross, W. E. (1987). Pharmac. Ther. 32, 89-106. Griffith, J. D. 85 Christiansen, G. (1978). Annu. Rev. Biophys. Bioeng. 7, 19-35. Griffith, J., Bleyman, M., Rauch, C. A.; Kitchen; P. A., Englund, P. T. (1986). Cell, 46, 117-724. Holm, C., Goto, T., Wang, J. C. & Botstein, D. (1985). Cell, 41, 553-563. Hsieh, T. (1983). Methods Enzymol. 100, 161-170. Hsieh, T. (1990). In Biological Effects of DNA Supercoiling (Cozzarelli, N. R. & Wang, J. C., eds), Cold Spring Harbor Laboratory Press, New York, in the press. Hsieh, C. & Griffith, J. D. (1988). Cell, 52, 535-544. Kirkegaard, K. & Wang, J. C. (1981). CeZZ,23, 721-729. Kitchin, P. A., Klein, V. A., Ryan, K. A., Gann, K. L.; Rauch, C. A., Kang, D. S., Wells, R. D. & Englund, P. T. (1986). J. Biol. Chem. 261, 11302%11309. Klevan, L. & Wang, J. C. (1980). Biochemistry, 19, 5229-5234. Kreuzer, K. N. & Huang, W. M. (1983). In Bacteriophage T4 (Mathews, C. K., Kutter, E. M., Mosig, G. & Berget, P. B., eds), pp. 90-96, Am. Sot. Microbial., Washington, DC. Lee, M. P., Sander, M. & Hsieh, T. (1988). J. Cell Biol. 107; 402a. Lee, M. P., Sander, M. & Hsieh, T. (1989a). J. Biol. Chem. 264, 13510-13518. Lee, M. P., Sander, M. & Hsieh, T. (1989b). J. BioZ. Chem. 264, 21779-21787. Liu, L. F. (1989). Annu. Rev. Biochem. 58, 351-375. Liu, L. F. & Wang, J. C. (1978a). Proc. Nat. Acad. Xci., U.S.A. 74, 2098-2102. Liu, L. F. & Wang, J. C. (19876). Cell, 15, 979-984. Liu, L. F., Rowe, T. C., Yang, L., Tewey, K. M. 8: Chen, G. L. (1983). J. BioZ. Chem. 258, 15365-15370. Maniatis, T., Fritsch, E. F. & Sambrook; J. (1982). Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Maxwell, A. & Gellert, M. (1986). Adwan. Protein Chem. 38, 69-107. Moore, C. L. (1982). Ph.D. thesis, University of North Carolina at Chapel Hill. Moore, C. L., Klevan, L., Wang, J. C. & Griffith, J. D. (1983). J. BioZ. Chem. 258, 46124617. Morrison, A. & Cozzarelli, N. R. (1981). Proc. Nat. Acad. Sci., U.S.A. 78, 1416-1420. Osheroff, N. (1986). J. BioZ. Chem. 261, 9944-9950. Reed, K. C. & Mann, D. A. (1985). NucZ. Acids Res. 13, 1207-7221. Rowe, T. C., Chen, G. L., Hsiang, Y. & Liu, L. F. (1986). Cancer Res. 46, 2021-2026.
M. T. Howard
62
Sander,
M. &I Hsieh,
T. (1983).
J. Biol.
Chem.
258,
8421-8428.
Sander,
M. & Hsieh, T. (1985).
Nucl.
Acids
Res. 13,
1057-1072.
Satchwell, S. C., Drew, H. R. & Travers, A. A. (1986). J. Mol. Riol. 191, 659-675. Shore, D. & Baldwin, R. L. (1983). J. Mol. Riol. 170, 957-98 1. Shore, D., Langowski, J. & Baldwin, R. L. (1981). Proc. Nat. Acad. Sci., U.S.A. 78, 48334837. Spitzner, J. R. & Muller, M. T. (1988). Nucl. Acids Res. 16, 5533-5556. Thresher, R. J. & Griffith, J. D. (1990). Proc. Nat. Acad. Sci.,
U.S.A.
et al.
Udvardy, A., Schedl, I’., Sander, MM. & Hsieh. 7’. (1986). J. Mol. Biol. 191, 231-246. Uemura, T.; Ohkura; H.; Adachi, Y., Plorino, K., Shiozaki, K. & Yanagida, M. (1987). Cell, 50. 917-925. Wang, J. C. (1985). Annu. flev. Biochem. 54, 665-697. Yang, L., Rowe, T. C. & Liu, L. F. (1985a). Cancer I?es. 45, 5872-5876.
Yang, L., Rowe, T. C., Nelson: E. M, & Liu, L. F. (19856). Cell, 41, 127-132. Yang, L., Weld. M. S., Li, J. J., Kelly, T. J. & Liu, L. F. (1987). Proc. Nat. Acad. Ski., U.X.A. 84, 950-954.
87, 5056-5060.
Edited
by P. van Hippel
(1991)
217,
53-62
Drosophila Topoisomerase II-DNA Interactions by DNA Structure Michael Tao-shih
are Affected
T. HowardI, Maxwell P. Lee’ Hsieh2 and Jack D. Griffithlf-
1Lineberger Cancer Research Center Chapel Hill, NC 27514, U.S.A. 2Department of Biochemistry Duke University Medical Center Durham, NC 27’710, U.S.A. (Received
23 May
1990; accepted 5 September
1990)
The binding of purified Drosophila topoisomerase II to the highly bent DNA segments from the SV40 terminus of replication and C. fasciculata kinetoplast minicircle DNA (kDNA) was examined using electron microscopy (EM). The probability of finding topoisomerase II positioned at or near the bent SV40 terminus and Crithidia fasciculata kDNA was two- and threefold higher, respectively, than along the unbent pBR325 DNA into which the elements had been cloned. Closer examination demonstrated that the enzyme bound preferentially to the junction between the bent and non-bent sequences. Using gel electrophoresis, a cluster of strong sodium dodecyl sulfate-induced topoisomerase II cleavage sites was mapped to the SV40 terminus DNA, and two weak cleavage sites to the C. fasciculata kDNA. As determined by EM, Drosophila topoisomerase II foreshortened the apparent length of DNA by only 15 base-pairs when bound, arguing that it does not wrap DNA around itself. When bound to pBR325 containing the C. fasciculata kDNA and the XV40 terminus, topoisomerase II often produced DNA loops. The size distribution was that predicted from the known probability of any two points along linear DNA colliding. In vitro mapping of topoisomerase II on DNA whose ends were blocked by avidin protein revealed that binding is enhanced at sites located near a blocked end as compared to a free end. These observations may contribute towards establishing a framework for understanding topoisomerase II-DNA interactions.
1. Introduction
In vitro, the addition
of sodium dodecyl sulfate to complexes of topoisomerase II and DNA in a double strand DNA (dsDNA) break (Liu et al., 1983; Sander & Hsieh, 1983). Mapping of such topoisomerase II cleavage sites has shown that some sites are more efficiently cleaved than others. A consensus cleavage sequence for Drosophila (Sander & Hsieh, 1985) and chicken erythrocyte (Spitzner & Muller, 1988) topoisomerase II has been determined by sequencing in vitro cleavage sites. The degree of correlation between a DNA sequence and the cleavage consensus does not always accurately predict the intensity of cleavage by topoisomerase II. This suggests that DNA sequence may only be one of several parameters that determine the (SDSI) results
The topological state of DNA has important implications for many aspects of DNA metabolism including transcription, recombination, chromosome segregation, repair and replication. The topoisomerase enzymes can profoundly affect these processes by altering DNA topology (Wang, 1985). Type II topoisomerases break both strands of the double helix and catalyze a strand passage and religation event (Maxwell & Gellert, 1986; Hsieh, 1990). The unique ability of topoisomerase II to pass DNA strands makes it essential for the separation of newly replicated DNAs (DiNardo et al., 1984; Holm et al., 1985; Yang et al., 1987; Uemura et al., 1987).
t Author addressed.
to whom
all correspondence
should
$ Abbreviations used: SDS, sodium dodecyl dsDNA, double strand DNA; bp, base-pair(s); kinetoplast DNA; EM, electron microscopy.
be
53 0022-2S36/91/010053-10
$03.00/O
0
1991
Academic
sulfate;
kDNA;
Press
Limited
54
,U. T. Howard
strength of topoisomerase II cleavage. Cleavage of DNA can be enhanced by the addition of anti-tumor drugs such as the acridines, anthracylines, elliptitines and epipodophyllotoxins (for reviews, see Glisson & Ross, 1987; Liu, 1989). In viwo, it appears that topoisomerase 11 is the target for these drugs (Yang et al., 1985a,b; Rowe et al., 1986), which act by trapping a putative intermediate in topoisomerase II-mediated DNA strand passage. This topoisomerase II-DNA complex stabilized by drugs is termed the “cleavable complex”. It has been suggested that t,he collision of replication enzymes with the cleavable complex generates a DNA lesion which is lethal to actively replicating cells. Topoisomerase 11 cleavage and DNA footprinting have been used extensively to predict sites of topoisomerase 11 activity. In most cases, strong binding correlates with strong cleavage (Liu & Wang, 1978a,b; Klevan & Wang, 1980; Fisher et al., 1981; Morrison & Cozzarelli, 1981; Lee et al., 1989a,b). Although one site of topoisomerase 11 binding has been identified that is not a strong cleavage site (Kirkegaard & Wang, 1981). To understand further the parameters that determine DNA binding, cleavage, strand passage and religation by topoisomerase II, its will be necessary to charact,erize the physical properties of the enzyme-DNA complex. Some proteins interact with DNA by wrapping DNA about themselves. The histone proteins preferentially assemble at sites of DNA bending (Drew & Travers, 1985; Satchwell et al., 1986; Hsieh & Griffith, 1988): due to the decreased energy required to coil an already bent DNA around the histone core. Kleven & Wang (1980) have shown that DNA gyrase will form a “gyrasome” on DNA with 145 base-pairs (bp) of DNA wrapped about the enzyme. If other topoisomerase 11 molecules also wrap DNA it would be expected that segments of DNA containing severe sequence-directed bends might act as preferential binding sites for these enzymes. Alternatively, DNA bending might influence the positioning of topoisomerase II in an indirect manner by creating DNA loops. A 223 bp DNA segment from Crithidia fasciculata kinetoplast DNA (BDNA) contains a severe sequence-directed bend, with the individual molecules forming almost perfect circles (Griffith et aE., 1986). When this segment of bent DNA is inserted within non-bent DPL’A, loops are formed bringing two segments of DNA into close proximity. If topoisomerase 11 prefers t,o bind DNA at, sites where two DNAs come in contact then DNA bending could create a strong topoisomerase II binding site at the base of such a loop. Two topoisomerase II enzymes have been visualized by electron microscopy (EM) bound to DNA: Escherichia c&I gyrase (Moore et al.: 1983) and phage T4 topoisomerase (Moore, 1982; Kreuzer & Huang, 1983); both enzymes were seen to locate at the base of DNS loops. Here we have employed EM to further characterize the topoisomera,se II-DNA complex, and to directly visualize the position of Drosophila topoisomerase II molecules on DNA containing sequence-
et al. directed bends. Also; phoresis to map strong within these DNAs.
we have used gel electrctSDS-induced cleavage sit,es
2. Materials and Methods (a) Proteins and DNA Plasmid DNAs were isolated by the alkaline iysis method (Maniatis et al.; 1982). Restriction enzymes were purchased from Pl’ew England Biological laborat,ories (XEBL) or Biological Research Laboratories and used according to manufacturers’ specifications. Restriction digested DEA was end-labelled by a fill-in reaction using T4 DNA polymerase (PI’EBL) and biotinylated deoxycytosine triphosphate purchased from EPU’ZO Diagnostics. Inc.: followed by incubation with avidin (C. Bortner & J. Griffith, unpublished results). Topoisomerase II was purified as described (Hsieh. 1983) with modifications (Lee et al., 1989a). (b) Topoisomerase of 10 50 10
II binding
reaction8
Binding reactions were carried out in a volume of 20 ~1 10 mivf-Tris. NC1 (pH 7.9), 50 m&l-KCi, @I m&r-EDTA. miv-MgCl,. 1.25 m;M-ATP. with 10 ~M-D?U’A and n;ll-topoisomerase II. The reactions were incubated for min at 30°C. cc) Cleavage reaction
The plasmid pJGCl/svt (Hsieh & Griffith: 1988) !inearized by treatment with H1:ndTIT was ineubat.ed with topoisomerave II under binding conditions for 10 min at 30°C. SDS was then added to 1 y0 (w/v) and incubation continued for 2 min. The DNA-protein complexes were digested with 50 ,ug of proteinase K/ml for 1 h at 45°C. and electrophoresed on I “;o (W/V) agarose gels. Mock cleavage reactions were identical except that EDTA was added to 25 m&r in place of 1 v/0 SIX. DN.4 was transferred to a Zeta-Probe membrane using the alkaline transfer method and probed as described by Reed B Mann (1985). Autoradiographs were obtained by exposure of Koda,k X-Omat AR film.
The t,opoisomerase IT-DSA complexes were fixed bag adding glutaraldehyde Tao 1 O/b, for 5 min at 30°C. The samples were then chrornatographed over 2-ml columns of Sepharose 4B and prepared for EM as described (Griffith 6t Christiansen, 1958). including rotary shadowca,st,ing with tungsten. Samples were examined in a Philips 400 EM TLG. The lengt,hs of molecules were measured from images on electron micrographs using a Summagraphics digitizer coupled to an IBM PC-AT comput’er programmed with software developed by Dr Richard R’ubin.
(a) To;r,~isomerase il pref~renbially binds clear t/be bent D-VA4 elements fro’mb C. fasciculata Line~o$~~t minicircles and the SY40 terwhus of replimtim The plasmid pJGCll/svt (Fig. l(a)) consists of a 223 bp fragment from C. fasciculata kDNA and a 1216 bp fragment, from the simian virus 40 (SF’4O) terminus region cloned into separate sites in
Drosophila
Topoisomerase
II-DNA
I 0;;
interactions
ShI&vun
(a)
(b)
SV40 3204
bp
PstI
55
terminal
DNA
T
P.tI
1988
bp
Avail
Figure 1. Restriction map of pJGCl/svt, pPK20l/cat and the SV40 terminus of replication. (a) pJGCl/svt contains 223 bp of bent kDNA from C. fasciculata and 1216 bp of bent SV40 DNA cloned into pBR325 (Hsieh & Griffith, 1988). (b) pPK2Ol/cat contains the bent 223 bp kDXA cloned into theBamH1 site of pSP65 (Kitchin et al., 1986). (c) The 1216 br) fragment of SV40 from base-pair 1988.3204, containing the terminus of replication and a sequence directed bend (&iek& Griffith, 1988). pBR325 (Hsieh & Griffith, 1988). These segments have been shown to have strong sequence-directed curvature. The 223 bp kDNA forms a tight loop (Griffith et al., 1986) and the 1216 bp fragment is bent through nearly 180” (Hsieh & Griffith, 1988) as seen by EM. The pJGCl/svt DNA was linearized with SphI and XalI, and end-tagged with a biotinavidin complex at the Sal1 site (see Materials and Methods). Purified DrosophiZa topoisomerase II was incubated with end-labeled pJGCl/svt DNA in a buffer containing ATP and Mg’+, fixed with glutaraldehyde, and prepared for EM (Fig. 2). Topoisomerase II appeared bound to the DNA at a single site or at two sites simultaneously. When topoisomerase II bound two sites simultaneously the intervening DNA formed a loop and these loops showed a broad spectrum of sizes. Approximately 100 molecules were measured from the untagged end of the DNA to each topoisomerase II molecule and the results displayed (Fig. 3). At this level of resolution, the enzyme appears to have an affinity for two large areas of DNA spanning the C. fasciculata kDNA and SV40 terminal sequences. It was calculated that topoisomerase II bound the C. fasciculata kDNA and SV40 terminus DNA three- and twofold more often, respectively, than the surrounding pBR325 sequences. To establish the pattern of topoisomerase II binding more precisely in the vicinity of the bent DNA, a PvuII-SphI-digested DNA fragment from pPK201/ cat (Kitchen et al.; 1986; Fig. l(b)) containing the bent C. fasciculata kDNA; and an AvaII-PstI fragment of the SV40 terminus region (Fig. 1 (c)), were isolated and uniquely end-tagged with biotin-avidin at the PvuII and PstI sites, respectively, incubated with topoisomerase II under binding conditions, fixed with glutaraldehyde and prepared for EM (Fig. 4(a) and (b)). Topoisomerase II binding was mapped by measuring DNA molecules from the untagged end to each topoisomerase II particle. The results revealed that for both DNAs there was a
strong localization of the enzyme to the junctions of straight and bent DNA (Fig. 5(a) and (b)). It was noted that 20% of the C. fasciculata kDNA molecules appeared looped with topoisomerase II particles bound where the strands crossed and 87% of these loops contained the highly bent kDNA.
(b) SDS-induced cleavage of DNA by topoisomerase II is more prevulent in the XV40 terminus than in the bent C. fasciculata kDNA Purified Drosophila topoisomerase II was incubated with linear pJGC/svt DNA, as described above, and cleavage induced by the addition of SDS to 1 yc. The DNA fragments were then treated with Proteinase K, electrophoresed on a 1 y. (w/v) a,garose gel and transferred to a nylon membrane (see Materials and Methods). Strong cleavage sites on pJGCl/svt DNA were mapped by probing the membrane with a HindIII-SaZI-digested fragment from pJGCl/svt that overlaps a unique end of pJGCl/svt DNA linearized with HindIII. A number of sites were identified, and the strength of cleavage was determined by comparing band intensities. A cluster of strong cleavage sites mapped to the SV40 terminus region and two weak cleavage sites to the C. fasciculata kDNA sequences (Fig. 6). Comparison of topoisomerase II binding and cleavage sites within the C. fasciculata and SV40 DNAs revealed that strong binding at the C. fasciculata kDNA site did not correspond to strong cleavage, but that the SV40 terminus region, which contains a very strong cluster of cleavage sites, did show a corresponding increase in binding (compare Figs 3 and 6). Sequence analysis of the SV40 terminus region shows two sites with 86% homology to the Drosophila topoisomerase II cleavage consensus located at’ position 2297 and 2549. The C. fasciculata kDNA did not contain significant sequence homology to the consensus.
Al. T. Howard
56
et
al.
Figure 2. Visualization of Drosophila topoisomerase II bound to DK4. A 1350 bp &“$I-&x/I fragment of pJG:Cljsvt DSA uniquely end-labeled at the S&I site with a biotin-avidin conjugate was incubated with Drosophila topoisomerase II. The proten-DNA complexes were fixed and prepared for EM by mounting onto thin carbon films and rotar? shadowcasting with tungsten (see Materials and IMethods). Shown in reverse contrast. The bar represent!s 0.1 pm.
(c) Drosophila,
topoisomerase around
/I itself
does
not
wrap
DNA
Drosophila topoisomerase IT wa,s incubated with DNA containing the BamHT-&&I HSP8787 intergenie heatshock region (Lee et al.; 19896) under binding reaction condit’ions, fixed with glutaralde-
hyde and prepared for EM (see Xat’eriais and Methods). Measurements were made of the HSP87A7 intergenie heat shock DNA, which contains four regions enriched in topoisomerase II cleavage sites (Lee et al.: 19896). with and Without Drosophila topoisomerase II bound. The results (Fig. 7) demonstrate only 50 L& (1 A=O.l nm) 01
Drosophila
Topoisomerase
223
II-DNA
interactions
57
bp kDNA
40 Percentage
60 distance
Figure 3. Positioning of topoisomerase II on pJGCl/svt DNA. The position of 205 topoisomerase II particles on the 7350 bp pJGCl/svt DPU’A was determined by measuring molecules on the electron micrographs, prepared as described in the legend to Fig. 2, from the &&I site to each bound protein. The distance along pJGCl/svt DNA is given in percentage distance from the &&I site towards the Sal1 site. The position of the bent 223 bp C. fax&data kDNA and the bent 1216 bp SV40 DPU’A from the terminus of replication (svt) are indicated.
about 15 bp foreshortening of the DNA when bound by a single topoisomerase II molecule. This agrees with footprinting data which demonstrate only 20 to 30 bp of protection by Drosophila topoisomerase II (Lee et aZ., 19893).
(d) Drosophila
topoisomerase
II
loops DNA
The binding of Drosophila topoisomerase II to pJGCl/svt DNA produced loops of a broad spectrum of sizes (Fig. 2). Topoisomerase II-associated DNA loops were also observed when binding was carried out in the presence of adenosine 5’-0-(3thiotriphosphate) or in the absence of any nucleotide co-factor, demonstrating that the formation of loops is not dependent on the binding or hydrolysis of ATP. These loops varied greatly in size from 200 to 2000 bp with the modal number of loops measuring about 500 bp (Fig. 8). The lack of dependence of loop size on ATP hydrolysis argues against a model in which the loops are generated by the translocation of the DNA through the enzyme complex from a single site. An alternative model is one in which the enzyme binds to a single site on DNA and then captures a second distant DNA site through a collision event. If this model is correct, then the distribution of loop sizes should reflect the probability of two sites on DNA coming in contact with each other by collision. Clearly if two sites are very close, their contact will be excluded by the stiffness of the DKA and if they are very far apart, their contact will also be unlikely. The optimal separation along DNA for achieving contact through a diffusion-driven process appears to be about 500 bp as revealed by cyclization studies
(Shore et al., 1981; Shore & Baldwin, 1983), and the size distribution of DNA loops associated with topoisomerase II closely agrees with this cyclization probability (Fig. 8). Loops of DNA associated with topoisomerase II were also observed when linear pJGCl/svt was incubated with topoisomerase II under binding conditions, and prepared for EM by a method employing freeze-drying and no chemical fixation (Thresher & Griffith, 1990), excluding the possibility that looping is an artifact of fixation. An identical distribution of loop sizes was observed when pBR325 was bound by topoisomerase II (data not shown).
(e) Topoisomerase avidin-bound
II
preferentially ends
localizes of
near
DNA
Topoisomerase II was found more frequently at sites located near an end blocked by a biotin-avidin conjugate than at sites near a free unblocked end (Fig. 9). To further investigate this, a DNA fragment containing the BamHI-XphI HSP87A7 intergenie heatshock region was bound by avidin at either end to biotinylated deoxycytosine (see Materials and Methods) The DNA was then incubated with Drosophila topoisomerase II, fixed with glutaraldehyde and prepared for EM. When the avidin complex was located at the BamHI end, binding of topoisomerase II at the cleavage site nearest the avidin was approximately eightfold higher than when the BamHI end was unblocked. Likewise, if the avidin was located at the XphI end, binding was approximately eightfold higher at the site nearest the SphI end than when the same end was unblocked.
58
M. T. Howard
et al.
Figure 4. Visualization of Drosophila t,opoisomerase II bound to C. jascicuhta iaDX8 and SV40 t,ermintiti i3S.A (a) pPK2Ol/cat DNA was digested with SphI and PvuII. t,he 890 bp fragment produced was labeled with biotic and avidin at the Pl;uII end, incubated with topoisomerase II and mounted for EM as in Fig. 2. (b) SV40 DK’A was digested with AvaT1 and P&I, and the resulting 1191 bp DNA was labeled with biotin and aridin at the P&I end. bound by topoisomerase II and prepared for EM as described in the legend to Fig. 2. Shown in reverse contrast,. The bar represents 0.1 pm.
4. Discussion In viz~o, it is likelv that’ a number of factors determine the DNA binding, cleavage and catalytic sites for topoisomerase IT. These factors might include DNA sequence, structure and the presence of other DNA binding proteins. The pattern of druginduced cleavage of DT\‘A by topoisomerase TI is different in z&o than t,hat observed in vitro, with some cleavage sites being enhanced and others decreased in intensity (Yang et ccl., 1985a,b; Udvardy et al.; 1986). Here, we have investigated one unusual DNA structure, sequence-directed bending, for its effect on topoisomerase II binding and cleavage of DNA. The in vitro binding of Drosophila topoisomerase TT to the SV40 terminus of replication and C. fasciculnta kDNA as shown by EM revealed that, these bent DNAs have a higher
affinity for the enzyme than surrounding vector sequences. SDS-induced cieavage of DXA by topoisomerase IT demonstrated that the C. ,jaksascictda~ta cleaved, whereas the bent kDNA wa.s not efficiently SV40 terminus DNA was found to contain a, cluster of strong cleava,ge sites. The relationship between sites of topoisomera,se II binding, cleavage and catalysis is not well understood, A DNA gyrase binding site that’ is not a cleavage site has been mapped by Kirkegaard 8 Wang (1981). This demonstrates tha,t topoisomerase II binding sites do not necessarily define sites of and the cleavage. In the case of the SC40 terminus C. fasciculata kDNAs the presence of a strong cleavage site may require the conjunction of 2; strong binding sit’e, as defined by the structure of DNA4; and a site of strong homology to t,hc topoisomerase TT consensus cleavage Lbosophila
TopoisomeraseII-DNA
Drosophila
interactions
1
59
Bentsvt DNA
Bent kDNA
40 Percentage
60 distance
1
li
h-L 80
Percentage
diston ce (bl
(0)
Figure 5. Positioning of Drosophila topoisomerase II on the bent C. fascicuZata kDNA and SV40 terminus of replication. (a) The position of topoisomerase II on the 890 bp fragment from pPK20l/cat containing the bent kDPjA was determined by measuring molecules on electron micrographs such as those shown in Fig. 4(a) from the SphI site to each bound topoisomerase labeled as described in Distances are represented bent DNA is represented
II. (b) Positioning of topoisomerase II on the 1191 legend to Fig. 4(b); was measured from the AvaII as percentage distance from the unlabeled end towards by the horizontal line above the bars in (a) and (b).
the
topoisomerase II-DNA interactions will hopefully clarify the relationship between binding, cleavage
sequence. The C. fasciculata kDNA does not contain regions of sequence homology to the consensus and demonstrates only weak cleavage. The SV40 terminus DNA, however, contains two sequences 86% homology to the Drosophila topoisowith merase IT cleavage consensus sequence and is cleaved efficiently by topoisomerase II. This
and catalysis.
Several proteins, including histones and DNA gyrase, have been shown to wrap DNA around a protein core even though the bending of DNA in such a manner is energetically unfavorable. The presence of sequence-directed DNA bending will lower the energy required to bend and hold DNA in such a conformation, and therefore could create a favored binding site for DNA-wrapping proteins. To date, there is no evidence that DNA is wrapped
observation is of particular interest in the light of the findings of Fields-Berry & DePamphilis (1989), which suggest that the SV4Q terminus of replication can inhibit the decatenation activity of topoisomerase II in wivo. Further investigation into
Lane
:
t
2
3
4
bp SV40 terminal DPjA, isolated and end to each bound topoisomerase II. the labeled end. The position of the
5
6
7
Figure 6. Topoisomerase II cleavage mapping of C. fasciculata kDNA and SV40 DXA from the terminus of replication. Topoisomerase II cleavage of HindIII-digested pJGC/svt was induced by addition of SDS to in vitro topoisomerase II binding reactions. A Southern blot of cleavage fragments was hybridized with a radioactive probe that overlaps a unique end of pJGCl/svt. Lane 1: BamHI-digested pJGCl/svt; lane 2, EcoRI-digested pJGCl/svt; lane 3, HindIIT-digested pJGCl/svt mock (EDTA added to 25 rnM in place of SDS) topoisomerase II cleavage reaction; lanes 4, 5 and 6, topoisomerase II cleavage of HindIII-digested pJGCl/svt at a DNA to protein molar ratio of 1 : 100, 1 : 50 and 1 : 25, respectively; lane 7, HindIII-digested 1 size marker. The center of the kDEA and SV40 terminus bend are represented by arrows.
60
M. T. Howard
et al. 30
20 “n -8 6 2 E ; IO
ib)
0
Id00 Loop
2600 size
i bp)
Figure 8. Comparison of Droso#la topoisomerase IIassociated DNA loop size with the Shore & Baldwin (1983) probability of cyclization. The number of loops of a measured length is represented by the open boxes (0) and the J-factor for probability of cyclization of a linear DlUA of a certain size with compiementary overhanging ends is represented hy the filled boxes (m).
0.12
0.14 Length
0.16 (pm)
Figure 7. Determination of the linear foreshortening of DKA by the binding of Drosophila topoisomerase II. of DSA containing the (a) Length histogram RamHI-&‘@ fragment from the heat shock intergenic region of HSP87A7 (Lee et al., 19896) bound by a single topoisomerase II molecule. (b) Length histogram of the same DNA free of protein. Samples in (a) and (b) were prepared for EM as described in the legend of Fig. 2, and lengths were measured from molecules on electron micrographs.
around eukaryotic topoisomerase II molecules. DBase I footprinting of the Drosophila topoisomerase II revealed a protected region of 20 to 30 bp with no indication of hypersensitive sites at a 10 bp interval (Lee et aZ., 1988, 1989b). In this study, the lengths of DNA bound by Drosophila topoisomerase II and protein-free DNA have been compared to determine if the DNA is wrapped around the enzyme. These results and the EM observations demonstrated that Drosophila topoisomerase II does not wrap DNA about itself and consequently would not be affected by DKA bending in the way proposed for nucleosome-like interactions. Proteins that bind a single DNA at two sites simultaneously must bend DNA to bring these sites into close juxtaposition. Curvature of the intervening DSA would facilitate this process. The C. fascicuZata bent kDNA formed topoisomerase II-associa,ted DNA loops in which 8’i oh of the molecules contained the bent DNA within the loop. The
binding of topoisomerase II to the junction of straight and bent DNA from C. fasciculata kDT\‘A may reflect t,he propensity of t,opoisomerase II to bind at the base of a DI\;A loop where these two regions will be in close proximity due to the bent, nature of the DNA. Topoisomerase II-associar5ed loops were observed within the SV40 terminus DNA much less frequently than within the C. fasciculata kDNA. A probable explanation for this difference may come from the observation that t,he 6’. fasciculata kDNA bends through nearly 360” (Griffith et al., 1986) whereas the SV40 terminus DNA appears to bend only about 180” (Hsieh & Griffith, 1988). If binding DNA at two separate siLes simultaneously increases the stability of t,he t’opoisomerase II-DNA complex, then the lower affinity of the SV40 terminus in compa,rison to the C. fasciculata kDNA for topoisomerase II may be explained by the probability of looping DSA at each site. An alternative explanation for the increased binding of topoisomerase II at, the junctions of straight and bent DNA resides in the evidence that suggests that topoisomerase II can move along DSA (Osheroff. 1986). We have observed that @rosophila topoisomerase II was more frequently found near an avidin-bound end of linear DNA t&n near a free end. This may reflect linear diffusion of the enzyme along DNA and t’he inability of the enzyme to be released from an end of DXA blocked by avidin. This interpretation must be approa,ched cautiously as a non-specific affinity of topoisomerase IT for avidin could also explain this finding. If such
TopoisomeraseII-DNA
Drosophila
interactions
61
We thank John Benson for helpful discussion. This work was supported by grants from the National Institutes of Health (GM31819, GM42342) to J.D.G., and (GM29006) to T.H., and the American Cancer Society (NP-583) to J.D.G. References
0
20
40 Percentage
60 distance
80
100
Figure 9. Enhanced localization of Drosophila topoisomerase II near avidin-bound ends of DNA. (a) DNA containing the BamHI-SphI HSP87A7 intergenic heatshock region (Lee et al., 1989b) was labelled at the BamHI end with biotin and avidin, incubated with topoisomerase II and prepared for EM as described in the legend to Fig. 2. The position of the topoisomerase II molecules was measured from the BamHI end of the DNA. (b) The same DNA fragment was labeled at the #phi end, reacted with topoisomerase II and the position of topoisomerase II molecules was determined as described for (a).
a passive diffusion model is correct,, enhanced binding of topoisomerase II could occur at, or near, sites where movement of topoisomerase II along DNA would be inhibited. These blocks to movement might include unusual DNA conformations or other proteins bound along the DNA. The primary recognition sequence has been determined for many DNA-binding proteins. Little is known about how the local DNA structure and other proteins nearby may influence recognition of these sequences by specific DNA-binding proteins. The local environment is of particular interest when the recognition sequence is not strictly defined but is part of a larger group of sequencesrelated to the consensus. Here, we have demonstrated the DNA-topoisomerase II interaction can be affected by DNA structure (bending), and have proposed that DNA looping and diffusion of topoisomerase II along DNA are characteristics of this interaction that may be involved in determining sites of topoisomerase II binding, cleavage and catalysis.
DiNardo, S., Voelkel, K. & Sternglanz, R. (1984). Proe. Nat. Acad. Sci., U.S.A. 81, 2616-2620. Drew, H. R. & Travers, A. A. (1985). J. Mol. BioZ. 186, 773-790. Fields-Berry, S. C. & DePamphilis, M. L. (1989). NucZ. Acids Res. 17, 3261-3273. Fisher, L. M., Mizuuchi, K., O’Dea, M. H., Ohmori, H. & Gellert, M. (1981). Proc. Nut. Acad. Xci., U.S.A. 78, 41654169. Glisson, B. S. & Ross, W. E. (1987). Pharmac. Ther. 32, 89-106. Griffith, J. D. 85 Christiansen, G. (1978). Annu. Rev. Biophys. Bioeng. 7, 19-35. Griffith, J., Bleyman, M., Rauch, C. A.; Kitchen; P. A., Englund, P. T. (1986). Cell, 46, 117-724. Holm, C., Goto, T., Wang, J. C. & Botstein, D. (1985). Cell, 41, 553-563. Hsieh, T. (1983). Methods Enzymol. 100, 161-170. Hsieh, T. (1990). In Biological Effects of DNA Supercoiling (Cozzarelli, N. R. & Wang, J. C., eds), Cold Spring Harbor Laboratory Press, New York, in the press. Hsieh, C. & Griffith, J. D. (1988). Cell, 52, 535-544. Kirkegaard, K. & Wang, J. C. (1981). CeZZ,23, 721-729. Kitchin, P. A., Klein, V. A., Ryan, K. A., Gann, K. L.; Rauch, C. A., Kang, D. S., Wells, R. D. & Englund, P. T. (1986). J. Biol. Chem. 261, 11302%11309. Klevan, L. & Wang, J. C. (1980). Biochemistry, 19, 5229-5234. Kreuzer, K. N. & Huang, W. M. (1983). In Bacteriophage T4 (Mathews, C. K., Kutter, E. M., Mosig, G. & Berget, P. B., eds), pp. 90-96, Am. Sot. Microbial., Washington, DC. Lee, M. P., Sander, M. & Hsieh, T. (1988). J. Cell Biol. 107; 402a. Lee, M. P., Sander, M. & Hsieh, T. (1989a). J. Biol. Chem. 264, 13510-13518. Lee, M. P., Sander, M. & Hsieh, T. (1989b). J. BioZ. Chem. 264, 21779-21787. Liu, L. F. (1989). Annu. Rev. Biochem. 58, 351-375. Liu, L. F. & Wang, J. C. (1978a). Proc. Nat. Acad. Xci., U.S.A. 74, 2098-2102. Liu, L. F. & Wang, J. C. (19876). Cell, 15, 979-984. Liu, L. F., Rowe, T. C., Yang, L., Tewey, K. M. 8: Chen, G. L. (1983). J. BioZ. Chem. 258, 15365-15370. Maniatis, T., Fritsch, E. F. & Sambrook; J. (1982). Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Maxwell, A. & Gellert, M. (1986). Adwan. Protein Chem. 38, 69-107. Moore, C. L. (1982). Ph.D. thesis, University of North Carolina at Chapel Hill. Moore, C. L., Klevan, L., Wang, J. C. & Griffith, J. D. (1983). J. BioZ. Chem. 258, 46124617. Morrison, A. & Cozzarelli, N. R. (1981). Proc. Nat. Acad. Sci., U.S.A. 78, 1416-1420. Osheroff, N. (1986). J. BioZ. Chem. 261, 9944-9950. Reed, K. C. & Mann, D. A. (1985). NucZ. Acids Res. 13, 1207-7221. Rowe, T. C., Chen, G. L., Hsiang, Y. & Liu, L. F. (1986). Cancer Res. 46, 2021-2026.
M. T. Howard
62
Sander,
M. &I Hsieh,
T. (1983).
J. Biol.
Chem.
258,
8421-8428.
Sander,
M. & Hsieh, T. (1985).
Nucl.
Acids
Res. 13,
1057-1072.
Satchwell, S. C., Drew, H. R. & Travers, A. A. (1986). J. Mol. Riol. 191, 659-675. Shore, D. & Baldwin, R. L. (1983). J. Mol. Riol. 170, 957-98 1. Shore, D., Langowski, J. & Baldwin, R. L. (1981). Proc. Nat. Acad. Sci., U.S.A. 78, 48334837. Spitzner, J. R. & Muller, M. T. (1988). Nucl. Acids Res. 16, 5533-5556. Thresher, R. J. & Griffith, J. D. (1990). Proc. Nat. Acad. Sci.,
U.S.A.
et al.
Udvardy, A., Schedl, I’., Sander, MM. & Hsieh. 7’. (1986). J. Mol. Biol. 191, 231-246. Uemura, T.; Ohkura; H.; Adachi, Y., Plorino, K., Shiozaki, K. & Yanagida, M. (1987). Cell, 50. 917-925. Wang, J. C. (1985). Annu. flev. Biochem. 54, 665-697. Yang, L., Rowe, T. C. & Liu, L. F. (1985a). Cancer I?es. 45, 5872-5876.
Yang, L., Rowe, T. C., Nelson: E. M, & Liu, L. F. (19856). Cell, 41, 127-132. Yang, L., Weld. M. S., Li, J. J., Kelly, T. J. & Liu, L. F. (1987). Proc. Nat. Acad. Ski., U.X.A. 84, 950-954.
87, 5056-5060.
Edited
by P. van Hippel
