Site-specific Cleavage of a DNA Hairpin by Topoisomerase I1 DNA SECONDARY STRUCTURE AS A DETERMINANT OF ENZYME RECOGNITION/CLEAVAGE*

To further define the nucleic acid determinants that govern the recognition of DNA by topoisomerase 11, the ability of the enzyme to cleave a 51-base oligonucleotide that contained a centrally located 19-base hairpin was characterized. Topoisomerase I1 cleaved the 51-mer in a site-specific fashion, within the hairpin, one nucleotide from the 3’-base of the stem. Protein denaturants were not required to trap cleavage products. Although the sequence of the oligonucleotide influenced levels of en- zyme-mediated DNA scission, it did not affect the spatial location of cleavage. DNA scission required a double- strandedkingle-stranded junction at the 3’-base of the hairpin and a tail (either single- or double-stranded) at least 8 bases in length on the 5’-side. Cleavage was not observed when base-pairing within the oligonucleotide was eliminated or when the hairpin was extended to produce a completely double-stranded substrate. Fi-nally, the enzyme displayed a size constraint for both the stem and loop structures of the hairpin. These results indicate that topoisomerase 11 is capable of recognizing secondary structure within nucleic acids and identifies the first secondary structure-specific DNA recognition/ cleavage site for the type I1 enzyme. 70 "C 10 min, and cooling to 25 "C over a period of 2 h. Annealing was confirmed by electrophoresis on nondenaturing 14% polyacryl- amide gels. The standard cleavage assay contained 100 I" topoisomerase I1 and 50 n~ oligonucleotide substrate in a total of 20 pl of 10 nm Tris-HC1, pH 7.9, 50 nm NaCl, 50 nm KCl, 0.1 nm EDTA, 2.5% glycerol (v/v), and 7.5 m~ MgC12. Reactions were incubated at 30 "C for 30 min and stopped by the addition of 1 pl of 250 nm EDTA followed by 2 pl of 10% SDS. Proteinase K (2 pl of a 1 mdml solution) was added, and topoisomerase I1 was digested at 45 "C for 30 min. Loading buffer (20 pl) was added and reactions were heated to 95 "C for 2 min. DNA cleavage products were resolved by electrophoresis at 500 V for 4 h in 14% polyacrylamide, 7 M urea gels in 100 nm Tris borate, pH 8.3, 2 nm EDTA and were visualized by autoradiography with Kodak XAR film and a Du Pont Lightning Plus screen. DNA cleavage was quantitated by scanning autoradiograms with an E-C Apparatus model EC910 scanning densitom- eter in conjunction with Hoefer GS-370 Data System Software. Alter-natively, DNA cleavage products were quantitated using a Molecular Dynamics PhosphorImager system. In some cases, DNA bands were excised and quantitated using a Beckman LS-7500 liquid scintillation counter. reactions utilizing a 3"radiolabeled oligo- nucleotide were camed out as described above. Reactions were stopped by the addition of 1 pl of 250 m~ EDTA followed by 2 pl of 10%


Site-specific Cleavage of a DNA Hairpin by Topoisomerase I1
DNA SECONDARY STRUCTURE AS A DETERMINANT OF ENZYME RECOGNITION/CLEAVAGE* (Received for publication, August 26, 1993, and in revised form, December 5, 1993) Stacie J. Froelich-AmmonS, Kevin C. Gale  To further define the nucleic acid determinants that govern the recognition of DNA by topoisomerase 11, the ability of the enzyme to cleave a 51-base oligonucleotide that contained a centrally located 19-base hairpin was characterized. Topoisomerase I1 cleaved the 51-mer in a site-specific fashion, within the hairpin, one nucleotide from the 3'-base of the stem. Protein denaturants were not required to trap cleavage products. Although the sequence of the oligonucleotide influenced levels of enzyme-mediated DNA scission, it did not affect the spatial location of cleavage. DNA scission required a doublestrandedkingle-stranded junction at the 3'-base of the hairpin and a tail (either singleor double-stranded) at least 8 bases in length on the 5'-side. Cleavage was not observed when base-pairing within the oligonucleotide was eliminated or when the hairpin was extended to produce a completely double-stranded substrate. Finally, the enzyme displayed a size constraint for both the stem and loop structures of the hairpin. These results indicate that topoisomerase 11 is capable of recognizing secondary structure within nucleic acids and identifies the first secondary structure-specific DNA recognition/ cleavage site for the type I1 enzyme.
Thus far, two models for topoisomerase II-mediated nucleic acid recombination have been proposed. They are the subunit exchange (37,55,56) and the intermolecular ligation (57,58) models, respectively. The subunit exchange model requires that two independent enzyme homodimers simultaneously introduce double-stranded breaks in the backbones of separate DNA helices (37,55,56). Recombination is mediated by the exchange of subunits between the two homodimers followed by ligation to generate new DNA species. In this model, the 3'-hydroxyl of the cleaved DNA (which is not covalently attached to the enzyme) never dissociates from the active site of topoisomerase 11.
In contrast to the above, the intermolecular ligation model requires the action of only a single homodimer of the type I1 enzyme (57,58). In this latter model, one of the 3"hydroxyl DNA termini generated by topoisomerase II-mediated cleavage dissociates from the active site of the enzyme and is replaced by a separate nucleic acid terminus. Recombination is mediated by the nucleophilic attack of the invading 3"hydroxyl on the enzyme-bound 5"phosphate terminus and subsequent ligation of the two DNAmolecules. The dissociation event that is critical to this model is rare when double-stranded nucleic acid substrates are cleaved by topoisomerase I1 (55,58). However, dissociation occurs frequently when molecules with doublestrandedsingle-stranded junctions (59), DNA duplexes containing nicks,' or single-stranded DNA molecules with the potential to form secondary structures (57,58) are utilized by the enzyme.
The fact that the DNA structures described above exist in vivo (60, 61) and are often located at sites of nucleic acid recombination (62-65) suggests that the intermolecular ligation model for topoisomerase II-mediated DNA recombination may have physiological relevance. Therefore, in order to detail the mechanism by which this potential recombination pathway occurs, interactions between the type I1 enzyme and a DNA hairpin-containing oligonucleotide were characterized. Results indicate that topoisomerase I1 cleaves the DNA hairpin in a sitespecific fashion and that the 3'-DNA terminus generated by scission is able to diffuse from the active site of the enzyme. Furthermore, the secondary structure rather than the sequence of the oligonucleotide is the dominant feature recognized by the type I1 enzyme. The substrate characterized in the present study represents the first example of a DNA recognitiodcleavage site for topoisomerase I1 that is dictated K. C. Gale  [Y-~~PIATP (3000 Cilmmol), l~c-~~PlddATP (3000 Cilmmol), and I~c-~~Sld-dATP (1300 Ci/mmol) were from Amersham; T4 polynucleotide kinase was from DuPont NEN SDS and proteinase K were from E. Merck Biochemicals; Tris and urea were from Sigma; and EcoLume liquid scintillant was from ICN.
Preparation of Oligonucleotide Substrates-Oligonucleotides were synthesized on an Applied Biosystems DNA Synthesizer. The DNA was dissolved in water, mixed with loading buffer (85% formamide, 50 nm Tris borate, pH 8.3, 0.05% bromphenol blue, 0.05% xylene cyano1 FF), and resolved by electrophoresis in 8% polyacrylamide/7 M urea gels in 100 nm Tris borate, pH 8.3, 2 n m EDTA. Samples were excised from gels, extracted overnight in 500 nm NH,OAc, 10 nm MgOAc2, and 1 nm EDTA, and purified by chromatography on SepPak C-18 (Millipore) columns as described by the manufacturer. 30-min incubation at 37 "C, radiolabeled samples were purified and resuspended as described above and diluted with nonlabeled oligonucleotide to a specific activity of -2 x lo5 cpdpmol. For experiments in which oligonucleotides were labeled on their 3"hydroxyl termini following topoisomerase 11-mediated DNA cleavage, [~~~P l d d A T p and terminal deoxynucleotidyltransferase were added to cleavage reaction mixtures and were allowed to incubate for an additional 1 h at 30 "C.
lbpoisomerase II-mediated DNA Cleauage of Oligonucleotides-Prior to DNA cleavage assays, hairpin-containing oligonucleotides were incubated at room temperature for 30 min. Base-pairing within the hairpin was confirmed by CD optical spectroscopy. When appropriate, oligonucleotides were annealed by mixing equimolar amounts, heating to 70 "C for 10 min, and cooling to 25 "C over a period of 2 h. Annealing was confirmed by electrophoresis on nondenaturing 14% polyacrylamide gels.
The standard cleavage assay contained 100 I " topoisomerase I1 and 50 n~ oligonucleotide substrate in a total of 20 pl of 10 nm Tris-HC1, pH 7.9, 50 nm NaCl, 50 nm KCl, 0.1 nm EDTA, 2.5% glycerol (v/v), and 7.5 m~ MgC12. Reactions were incubated at 30 "C for 30 min and stopped by the addition of 1 pl of 250 nm EDTA followed by 2 pl of 10% SDS. Proteinase K (2 pl of a 1 mdml solution) was added, and topoisomerase I1 was digested at 45 "C for 30 min. Loading buffer (20 pl) was added and reactions were heated to 95 "C for 2 min. DNA cleavage products were resolved by electrophoresis at 500 V for 4 h in 14% polyacrylamide, 7 M urea gels in 100 nm Tris borate, pH 8.3, 2 nm EDTA and were visualized by autoradiography with Kodak XAR film and a Du Pont Lightning Plus screen. DNA cleavage was quantitated by scanning autoradiograms with an E-C Apparatus model EC910 scanning densitometer in conjunction with Hoefer GS-370 Data System Software. Alternatively, DNA cleavage products were quantitated using a Molecular Dynamics PhosphorImager system. In some cases, DNA bands were excised and quantitated using a Beckman LS-7500 liquid scintillation counter.

Covalent Attachment of lbpoisomerase II to Oligonucleotide Cleavage
Products-DNA cleavage reactions utilizing a 3"radiolabeled oligonucleotide were camed out as described above. Reactions were stopped by the addition of 1 pl of 250 m~ EDTA followed by 2 pl of 10% SDS. DNA. The position of topoisomerase 11-mediated DNA cleavage (be-Samples were mixed with 20 pl of 250 nm Tris-HC1, pH 6.8, 8% SDS, 20% 2-mercaptoethanol, 40% glycerol (v/v), 0.004% bromphenol blue, and were subjected to electrophoresis in a 2% agarose (MCB) gel in 100 nm Tris borate, pH 8.3, 2 nm EDTA, and 0.1% SDS. Following electrophoresis, topoisomerase I1 was visualized by staining with 0.05% Coomassie Brilliant Blue R-250 in 25% isopropyl alcohol, 10% acetic acid. The gel was destained by diffusion in 25% isopropyl alcohol, 10% acetic acid and was dried by partial vacuum. Topoisomerase I1 that was covalently attached to radiolabeled DNA cleavage products was visualized by autoradiography as described in the preceding section. Alternatively, DNA cleavage reactions treated with or without proteinase K were monitored by electrophoresis on a denaturing 14% polyacrylamide gel. DNA cleavage products were visualized by autoradiography.

RESULTS
Previous studies demonstrated that the type I1 topoisomerase from bacteriophage T4 (67) or D. melanogaster (57,581 was able to cleave w 1 7 4 (+)-strand DNA and that cleavage intermediates could be trapped even in the absence of protein denaturants. Although the viral DNA is single-stranded in nature, it has considerable potential for forming secondary structures. Since strong sites of topoisomerase 11-mediated DNA cleavage in the w 1 7 4 viral strand often map in the vicinity of sequences that form hairpins (67),2 it is likely that the enzyme has the capability to recognize specific structures in nucleic acids. Therefore, to characterize interactions between topoisomerase I1 and regions of DNA secondary structure, the ability of the Drosophila enzyme to cleave a hairpin-containing oligonucleotide was assessed.
The DNA substrate employed in this study is shown in Fig. 1.
The oligonucleotide is 51 bases long and contains an 8-base pair stem, a 3-base single-stranded loop, and 16-base singlestranded tails on both the 5'-and the 3'-ends of the hairpin. The sequence of the oligonucleotide represents bases 3941-3991 of w 1 7 4 (+)-strand DNA (68h3 This region of the viral DNA is thought to play an important role in transcription termination (69). Furthermore, sites of topoisomerase 11-mediated DNA cleavage (as determined by high resolution mapping) appear to be located near this sequence (67).
Cleavage of the Hairpin-containing Oligonucleotide by Topoisomerase ZZ-Topoisomerase I1 cleaved the hairpin-containing 51-mer in a site-specific fashion (Fig. 2). When the DNA substrate was labeled with [32Plphosphate on its 5'-terminus, a unique 34-mer was generated following incubation with the enzyme. The length of the reaction product was determined by comparison with oligonucleotide standards derived from the 5"terminus of the 51-mer (Fig. 2 ) and was confirmed by COmigration with the predicted 34-mer (not shown). As denoted by the arrow in Under the conditions employed, -8% of the initial substrate was cleaved following a 30-min incubation with the enzyme.
To further characterize cleavage of the 51-mer by topoisomerase 11, alterations were made in the reaction conditions. For comparison, relative levels of DNA cleavage were set to 1.0 for complete 30-min assays. As seen in Table I, cleavage required the presence of topoisomerase I1 and a divalent cation. Although less efficient than magnesium, calcium (which supports cleavage of double-stranded substrates (70)) also supported scission of the 51-mer. Cleavage of the hairpin did not require ATP and was diminished when topoisomerase I1 was incubated with the nonhydrolyzable ATP analog, APP(NH)P,4 prior to the addition of oligonucleotide. This latter finding is consistent with the fact that nonhydrolyzable ATP analogs induce a conformational change that impedes DNA substrates from entering the active site of the enzyme (71,72 (57,58,67). This is due to the fact that the 3'-hydroxyl termini generated by cleavage of the viral DNA strand can dissociate from the active site of topoisomerase I1 and uncouple the cleavageh-eligation equilibrium of the enzyme (57). The DNA cleavage reactions shown in Fig. 2 and Table I were terminated by the addition of EDTA. This finding strongly suggests that the 3'-hydroxyl of the 34mer product can dissociate from topoisomerase I1 upon cleavage. This hypothesis was confirmed by several experiments. First, no reversal of DNA cleavage was observed when reactions were treated with salt (500 mM NaCI) prior to EDTA (Table I). Second, levels of the 34-mer cleavage product increased with time (Fig. 3). This is in marked contrast to the time independent cleavage that occurs with double-stranded DNA substrates (25,57). Finally, the 3"terminus of the 34-mer The abbreviation used is: APP(NH)P, adenyl-5"yl P,y-imidodiphosphate.  Terminated with salt' Terminated with SDSr Terminated with salt, then SDSE Unless noted otherwise, all reactions were carried out as described under "Experimental Procedures" and were terminated by the addition of EDTA.
Relative levels of cleavage in complete reaction mixtures were set to 1.0. Data are the averages of three independent experiments. Standard deviations are indicated.
A concentration of 5 mM CaClz was employed in place of MgC12. A concentration of 1 mM APP(NH)P was employed. Sodium chloride (500 mM final concentration) was added, and reactions were incubated at 30 "C for 5 min prior to the addition of EDTA.
f SDS (1% final concentration) was added to reactions prior to EDTA. Sodium chloride (500 m~ final concentration) was added and reactions were incubated at 30 "C for 5 min. then SDS (1% final concentration) was added prior to EDTA. Data are plotted as the percentage of initial DNA substrate. could be labeled by the addition of terminal deoxynucleotidyltransferase and [ ( Y -~~P I~~A T P (Fig. 3, inset ). Following either a 30or 60-min cleavage assay, -75% of the 34-mer generated by scission of the hairpin-containing oligonucleotide was labeled. No 3'4abeling of the 34-mer was observed when topoisomerase I1 was absent from reaction mixtures.

I
Although a protein denaturant was not required to trap DNA cleavage products, higher levels (-2-fold) of 34-mer were observed in reactions that were terminated by the addition of SDS (Table I). Thus, release of the newly generated DNA 3'hydroxyl appears to be less efficient than the cleavage event. It is likely that topoisomerase I1 undergoes more than one round of cleavage/religation with the 51-mer before the 34-mer cleavage product dissociates from the active site of the enzyme. This suggestion is supported by the fact that treatment of reactions with salt prior to SDS reduced levels of DNA cleavage to those observed in reactions terminated with EDTA (Table I). A hallmark of topoisomerase 11-mediated cleavage of doublestranded DNA is the covalent linkage of the enzyme to the 5"DNAtermini generated by the scission event (24-28). Asimilar covalent enzyme-DNA linkage was formed during cleavage of the hairpin-containing oligonucleotide. When 3"radiolabeled 51-mer was utilized as a substrate, the corresponding labeled 17-mer cleavage product did not migrate into polyacrylamide gels unless topoisomerase I1 was digested with proteinase K prior to electrophoresis (not shown). Furthermore, as determined by the co-migration of topoisomerase I1 with radiolabeled DNA in a denaturing protein gel (Fig. 4, lane 2 ) , the 17-mer became covalently attached to the enzyme. The covalent attachment of DNA cleavage products was not observed in the absence of topoisomerase 11-mediated DNA scission (lane 1 ) or when 5"radiolabeled 51-mer was used as the substrate for cleavage (not shown).
Effects of Nucleotide Sequence on Cleavage of the Hairpincontaining OligonucleotideSites of topoisomerase II-mediated cleavage in double-stranded DNA are dictated by the nucleotide sequence of the substrate (24, 73-77). In many cases, substitution of even a single base within the nucleic acid recognition sequence of the enzyme abolishes cleavage (77). To examine the effects of primary structure on the ability of topoisomerase I1 to cleave the hairpin-containing oligonucleotide, a series of 51-mers with altered DNA sequences was synthesized. It should be noted that the secondary structure of the parent oligonucleotide was maintained throughout this series. 'The first oligonucleotide examined was one in which the se uence of the entire 51-mer was inverted (Le. the sequence that read 5' + 3' was constructed 3' + 5'). As seen in Figs. 5 ar.d 6, topoisomerase 11-mediated cleavage of the inverted oligcnucleotide was -6-fold lower than that observed with the parent substrate. Remarkably, however, the spatial location of DNA cleavage (i.e. one nucleotide from the 3'-base of the hairpin and 34 bases from the 5"terminus) was identical with that obtained with the parent 51-mer (Fig. 5). No cleavage was observed at any other sequence position.
To further assess the effects of DNA sequence on the cleavage of the oligonucleotide by topoisomerase 11, three additional 51mers were constructed; the first inverted the 2 bases directly flanking the cleavage site, the second inverted the hairpin (i.e. s t e d o o p ) sequence, and the third replaced the single-stranded tails with their complementary sequences (Fig. 6). As above, the spatial location of cleavage for all three oligonucleotides was identical with that obtained in the parent 51-mer. However, the primary structure of the oligonucleotides influenced that of the inverted 51-mer. Finally, the sequence of the singlestranded tails had little effect on topoisomerase 11-mediated DNA cleavage.
These results indicate that the DNA sequence of the 51-mer influences levels of topoisomerase 11-mediated DNA scission, with alterations in the vicinity of the cleavage site having the greatest effect. However, since DNA sequence did not affect the spatial location of cleavage, this suggests that the site specificity of the enzyme for this hairpin is dictated by the secondary, rather than the primary structure of the oligonucleotide.
Requirement for a Double-stranded /Single-stranded DNA Junction in the OligonucleotideSince the site of cleavage by topoisomerase I1 is one base from the double-strandedkinglestranded DNA junction on the 3'-side of the hairpin, the requirement for this junction in enzyme recognitiodscission was characterized (Fig. 7). When base-pairing within the hairpin was eliminated by converting the bases in the 5'-half of the stem to their complements, cleavage of the oligonucleotide was abolished. An identical result was obtained when the hairpincontaining substrate was replaced by two oligonucleotides that made up the 5'and 3"halves of the 51-mer but were not annealed. These results indicate that base pairing is essential for topoisomerase 11-mediated cleavage of the parent oligonucleotide. Finally, no cleavage was observed when the doublestrandedsingle-stranded DNA junction was eliminated by converting the sequence of the 3'-tail to the complement of the 5"tail. A similar result was obtained even when cleavage reactions with this double-stranded substrate were terminated by the addition of SDS (not shown). Thus, topoisomerase I1 does not recognize the 51-mer when the nucleic acid region of enzyme-mediated DNA cleavage is entirely double-stranded. The above findings strongly suggest that topoisomerase I1 requires a double-strandedsingle-stranded junction for cleavage of the hairpin-containing oligonucleotide.
Recognition of the StemlLoop Region of the Oligonucleotide-To further characterize the influence of DNA secondary structure on the cleavage of the 51-mer by topoisomerase 11, the length of the stem and loop regions of the hairpin were altered. The hairpin in the parent 51-mer contains a n 8-base pair stem and a 3-base loop.
Increasing the length of the double-stranded DNA stem had a dramatic effect on the efficiency of enzyme-mediated scission (Fig. 8). Adding even 2 base pairs to the top of the stem decreased cleavage by >go%. Moreover, adding 4 base pairs nearly eliminated cleavage. Decreasing the length of the stem by removing the 2 base pairs adjacent to the loop had a lesser effect on DNA scission. When this oligonucleotide with a 6-base pair stem was employed, the level of DNA cleavage was -70% that of the parent hairpin. It is not clear whether this latter effect is due to a n altered interaction of the oligonucleotide with the enzyme or to a diminished stability of the shortened stem.
These results suggest that topoisomerase I1 has a size constraint for its recognition of the hairpin. If this suggestion is correct, increasing the size of the loop could also adversely affect the ability of the enzyme to cleave the oligonucleotide. As shown in Fig. 7, this was the case. When the 3-base loop of the parent 51-mer was replaced by a 6-or 9-base loop, levels of cleavage dropped by -84% or -96%, respectively. Thus, it appears that topoisomerase I1 recognizes not only the doublestrandeasingle-stranded junction of the oligonucleotide, but the size of the hairpin as well.
Requirement for the Single-stranded DNA Tails of the Oligonucleotide-Footprinting data indicate that topoisomer-". . . ase I1 interacts with approximately 10-15 bases on either side of its cleavage site in double-stranded DNA (78,79). Therefore, the importance of the single-stranded DNA tails that flank the hairpin was determined.
The length of the single-stranded 5'-tail was critical for topoisomerase 11-mediated cleavage (Fig. 9). Decreasing the length of the 5'-tail from 16 to 12 bases diminished DNA scission by S O % . Furthermore, DNA cleavage was abolished when the length of the tail was reduced to 6 bases. It does not appear that the enzyme requires more than 16 bases on the 5'-side of the hairpin, as increasing the size of the tail to 19 bases had no effect on DNA cleavage. In contrast to results with the 5'-tail, the length of the 3'-tail was inconsequential (Fig. 9). Even an oligonucleotide with a 2-base 3'-tail was cleaved with an efficiency that was comparable to that of the parent 51-mer.
Finally, the requirement for the DNA tails to be singlestranded was examined (Fig. 10). When the parent oligonucleotide was converted from a hairpin to a cruciform by annealing it with a 51-mer hairpin-containing oligonucleotide that contained tails complementary to those of the parent substrate, no DNA cleavage was observed. This finding suggests that the single-stranded nature of at least one of the tails is important for topoisomerase 11-mediated DNA scission. Consequently,  -tail (3'-DS fl2bp)). All annealmg oligonucleotides were constructed to yield blunt ends at the respective termini of the parent substrate. Data represent the averages of 2 independent experiments. Standard errors are indicated. The relative level of DNA cleavage for the parent 51-mer was set to 1.0. complementary 15-base oligonucleotides were annealed to either the 5'-or 3'-tail. These 15-mers were constructed to yield blunt ends at the termini of the parent substrate. Converting the 5'-tail to a double-stranded region had only a small (<20%) effect on cleavage. However, conversion of the 3'-tail to a double-stranded region eliminated cleavage of the hairpin. When the 15-mer was replaced by a 12-base complementary oligonucleotide (that annealed to bases 40-511, the level of DNA scission was restored to that of the parent 51-mer. Therefore, it appears that topoisomerase I1 requires the DNA flanking the 3'-side of the hairpin to be single-stranded for at least 2 4 nucleotides. DISCUSSION Considerable circumstantial evidence suggests that topoisomerase I1 mediates recombinational events within the cell (37). In vitro, nucleic acid molecules with the potential to form secondary structures are efficient substrates for topoisomerase 11-mediated DNA cleavage (57,58,67) and intermolecular ligation (57,58). Considering that such DNAs are often hot spots for illegitimate recombination within the genetic material (62-651, and that hairpins are intermediates in the recombination of V(DM regions (67), interactions between topoisomerase I1 and a hairpin-containing DNA oligonucleotide were examined.
Topoisomerase I1 cleaved the hairpin in a site-specific fashion, one nucleotide from the 3'-base of the stem. Furthermore, the 3"hydroxyl generated by DNA cleavage was able to dissociate from the active site of the enzyme. Therefore, DNA hairpins are potential substrates for topoisomerase 11-mediated illegitimate recombination.
Little is known about the features that govern the intrinsic site specificity of topoisomerase I1 for its nucleic acid target. Although the topological state of DNA(80), the presence ofATP or nonhydrolyzable ATP analogs (24, 81,82), and the substitution of other divalent cations for magnesium (28, 70) all influence the efficiency of enzyme-mediated cleavage, none of these affects the site of nucleic acid scission. All currently available evidence indicates that topoisomerase I1 binding (78,79) and cleavage (24,73-77) are directed by the primary structure of its DNA substrate. However, while several consensus nucleic acid sequences for cleavage have been proposed, they bear little resemblance to one another (24, 73-77). Thus, the precise se-quence determinants that dictate site specificity still remain equivocal.
TO date, all DNA cleavage sites identified for topoisomerase I1 are double-stranded in nature (21)(22)(23). Although sites of DNA cleavage have been described in substrates that contain double-strandedsingle-stranded junctions (59), topoisomerase I1 also cuts these sites when they are located in completely double-stranded regions (24, 59, 83).
The DNA hairpin described in the present study represents a unique class of recognition sites for topoisomerase 11-mediated cleavage. The enzyme does not utilize this site when it is embodied in purely single-stranded or double-stranded DNA. Furthermore, in marked contrast to sites of double-stranded DNA scission, site specificity within the hairpin is not influenced by nucleic acid sequence. Finally, in addition to its requirement for a double-strandedlsingle-stranded DNA junction, the enzyme also recognizes the size of both the stem and the loop structures. Thus, topoisomerase I1 must be interacting with regions of the hairpin that are distal to the spatial location of cleavage.
In conclusion, the present study demonstrates that topoisomerase I1 is capable of recognizing secondary structure within nucleic acids and identifies the first secondary structure-specific DNA recognitiodcleavage site for the type I1 enzyme.