A Homogeneous Type I1 DNA Topoisomerase from HeLa Cell Nuclei”

Using kinetoplast DNA networks as a substrate in a decatenation assay, we have purified to apparent homogeneity a type I1 DNA topoisomerase from HeLa cell nuclei. The most pure preparations contain a single polypeptide of 172,000 daltons as determined by sodium dodecyl sulfate-gel electrophoresis. The molecular weight of the native protein, based on sedimentation and gel filtration analyses, is estimated to be 309,000. These results suggest that the enzyme is a dimer of 172,000-dalton subunits. The enzyme is a type I1 topo- isomerase as demonstrated by its ability to change the linking number of DNA circles in steps of two and to decatenate or unknot covalently closed DNA circles. No gyrase activity is detectable. ATP is required for the relaxation, decatenation, and unknotting of DNA, and a DNA-dependent ATPase activity is present in the most pure fractions. ATP is hydrolyzed to ADP in this reaction. This enzyme is very similar in its catalytic properties to T4 DNA topoisomerase (Liu, L. F., Liu, C. C., and Alberts, B. M. (1979) Nature 281,456-461). Topoisomerases are enzymes which introduce transient breaks in the DNA backbone and thereby participate in a number of genetic processes. Some reactions catalyzed by these enzymes which depend breakage and reunion mechanism are: supercoiling or relaxation of closed circular duplex DNA; catenation and decatenation DNA circles; knotting and viral

Topoisomerases are enzymes which introduce transient breaks in the DNA backbone and thereby participate in a number of genetic processes. Some reactions catalyzed by these enzymes which depend on the breakage and reunion mechanism are: 1) supercoiling or relaxation of closed circular duplex DNA; 2) catenation and decatenation of DNA circles; 3) knotting and unknotting of DNA circles; 4) viral integration or joining together of DNA strands to form novel sequences; and 5) complete renaturation of single-stranded circles of complementary sequence (for reviews see Refs. [1][2][3][4]. Topoisomerases have been found in both procaryotic and eucaryotic cells; they are present in the virion of vaccinia virus (5), and they are induced by several bacteriophages (6)(7)(8).
Topoisomerases have recently been divided into two classes which are distinguished by their reaction mechanisms (12,13).
The type I enzymes transiently break one strand of the helix, permitting the linking number to change in steps of one. Omega protein (14) from Escherichia coli and the eucaryotic nicking-closing enzymes (15) are examples of type I enzymes. Known type I enzymes do not require energy input for their reactions. The energy of the phosphodiester bond is conserved after strand breakage by the covalent attachment of the * This research was supported by National Institutes of Health Grants GM-27608-13 and GM-27731. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

+ Supported by Public Health Predoctoral Fellowship Training
Grant 5T01-GM-00184. enzyme to one end of the DNA, allowing the DNA backbone to be reformed after topoisomerization (1, 2).
The type I1 topoisomerases catalyze the topological passing of two double-stranded DNA segments, presumably by introducing a transient enzyme-bridged double strand break on one of the crossing DNA segments (12, [16][17][18]. This mechanism results in a change in linking number in steps of two (12, [18][19][20]. A consequence of this mechanism is that type I1 enzymes can catenate and decatenate covalently closed circles as well as relax supercoils. Gyrase (21), which can also induce negative supercoiling, is the best characterized example of this class of enzymes (for reviews see Refs. 3 and 4). T4 topoisomerase, another well studied type I1 topoisomerase, has not been shown to supercoil DNA, but can catalyze relaxation, knotting, and catenation reactions (12). Based on genetic studies, the T4 enzyme is involved in DNA synthesis (reviewed in Ref. 7 ) . Several eucaryotic enzymes, similar to T4 topoisomerase, have been identified by their ability to catenate and decatenate DNA (18, 22).
Because of the likely importance of type I1 topoisomerases in eucaryotic DNA replication, we are studying the topoisomerases present in HeLa cells. To search for a type I1 enzyme we have used as a substrate the mitochondrial DNA (kinetoplast DNA) of Crithidia fasciculata, an insect trypanosomatid. This DNA is in the form of networks consisting primarily of about 5000 covalently closed minicircles (2.5 kilobases) which are topologically interlocked (23). Type I1 topoisomerases, such as T4 topoisomerase and Micrococcus luteus DNA gyrase, decatenate this DNA to form individual minicircles which are easily detectable by gel electrophoresis (24). Type I enzymes cannot decatenate these networks, making a type I1 enzyme easy to detect even in the presence of a large amount of type I enzyme. Using this assay, we have purified to apparent homogeneity a new HeLa topoisomerase activity, which we call HeLa topoisomerase 11. It is a type I1 topoisomerase. We have characterized its physical and enzymatic properties.

Cell Growth and Nuclei Isolation-HeLa cells (S-3), a gift of Dr.
Bernard Moss (National Institutes of Health), were grown at 37 "C in minimal essential medium for suspension cultures supplemented with 5% horse serum, 100 units/ml of penicillin, 0.1 mg/ml of streptomycin sulfate (all from Grand Island Biological Co.). Fresh cells (15 liters of culture; 5-6 X IO5 cells/ml) were processed for each purification. The culture was chilled to 4 "C, and the nuclei isolation was carried out at 0-4 "C. The cells (about 20 g, wet weight) were centrifuged (3000 X g for 10 min), washed twice with 100 ml of 0.15 M NaCI, 10 mM sodium phosphate, pH 7.5, and then resuspended in 100 ml of extraction buffer (5 mM potassium phosphate, pH 7.0,2 mM MgClz, 1 mM PMSF' (added as a 100 mM solution in isopropyl alcohol), 1 mM mercaptoethanol, 0.5 mM dithiothreitol, 0.1 mM EDTA). After swelling at 0 "c for 30 min, the cells were broken by Dounce homogenization (about 20 strokes with a loose pestle). Cell disruption was monitored by ' The abbreviations used are: PMSF, phenylmethylsulfonyl fluoride; NaDodS04, sodium dodecyl sulfate; PEG, polyethylene glycol. phase microscopy. Nuclei from the broken cells were collected by centrifugation (200 X g for 10 min) and washed once with 50 ml of nuclei wash buffer (extraction buffer without MgCh and EDTA).
Topoisomerase Assays-The standard topoisomerase I1 reaction mixture (20 pl) contained 50 mM Tris-HC1, pH 7.9, 120 mM KC1, 10 mM MgC12,0.5 mM dithiothreitol, 0.5 mM EDTA, 30 p g / d of bovine serum albumin, 0.5 mM ATP, 20 p g / d of DNA. Kinetoplast DNA was used for the decatenation reactions, and plasmid pBR322 for the relaxation reactions. For the unknotting reactions, knotted P4 DNA or knotted pBR322 was used in the standard reaction. After incubation at 30 "C for 30 min, the reactions were terminated by the addition of 5 rJ of 5% (w/v) NaDodSOr, 25% (w/v) Ficoll, 0.25 m g / d of bromphenol blue. Reactions were loaded on 1% horizontal agarose gels and electrophoresed in 90 mM Tris-borate, pH 8.3,2.5 mM EDTA. Gels were electrophoresed for 3 h at 6 V/cm, stained with 10 pg/ml of ethidium bromide, destained in water, and photographed under UV illumination. One unit of activity is the amount of topoisomerase needed to decatenate about 50% of the kinetoplast circles under these conditions. The assay of topoisomerase I was described previously 6% PEG. Addition of PEG to the diluent causes a stimulation of (25). Enzymes were diluted for assay in reaction mixture containing activity of approximately &fold.
ATPase Assays-The reaction conditions were identical with those used in the topoisomerase assay (with pBR322 DNA), except that the ATP concentration was reduced to 230 PM and it was labeled with 3H (6.5 Ci/mol). The reactions were stopped by the addition of 3 pl of solution containing ATP, ADP, and AMP (3 mM each), and by spotting onto polyethyleneimine-cellulose strips. After elution with 0.5 M LiCI, 2 M formic acid, the reference nucleotides were located under UV illumination. Radioactivity was assayed by scintillation counting.
Other Methods"NaDodSO4 gels were run according to the methods of Laemmli (26). Protein concentration was determined by the method of Bradford (27). Topoisomers of a unique linking number were isolated as described previously (12). Sedimentation coefficient was determined according to Martin and Ames (28). Centrifugation was in an SW 50.1 rotor for 6 h at 4 "C at 50,000 rpm on a 5-20% sucrose gradient. The gradient was buffered with 0.2 M potassium phosphate, pH 7.0, and contained 1 mM PMSF.
DNA-Plasmid DNA was purified by phenol deproteinization of a cleared lysate followed by CsCl/ethidium bromide equilibrium centrifugation (7). Kinetoplast DNA (form I, containing covalently closed circles) was purified by differentia1 centrifugation and CsCl/propidium diiodide equilibrium centrifugation (23). P4 DNA (from mature phage) and P4 knotted DNA (from phage heads) were isolated by phenol extraction after CsCl equilibrium centrifugation of a lysate of Escherichia coli C-2048 (P2) infected with P4 del Knotted pBR322 was prepared by treatment of pBR322 DNA with T4 topoisomerase (7).
Other Materials-Bio-Gel A-1.5m and hydroxylapatite (Bio-Gel HTP) were from Bio-Rad, phosphocellulose (P-11) was from Whatman, and DNA cellulose was prepared by the method of Alberts and Herrick (29). HeLa topoisomerase I (the 100, OOO form) (25) and T4 was from New England Nuclear and [a-"PIATP was from ICN. topoisomerase (7) were purified as described previously.
["]ATP PMSF was from Eastman Kodak, PEG 6000 was from Baker, and polyethyleneimine plates were from Brinkmann. All other chemicals were of highest grade commercially available.

Purification
A summary of the purification of HeLa topoisomerase I1 is presented in Table I. All steps of the purification were done at 0-4 "C, a n d the purification was completed in less than 3 days.
1. Preparation of the PEG Supernatant-Nuclei from 20 g of fresh cells were resuspended in 100 m l of nuclei wash buffer, and EDTA was added to a final concentration of 4 mM. After 15 m i n a t 0 "C, the nuclei were lysed by the slow addition, with stirring, of 100 ml of 2 M NaCl, 50 mM Tris-HC1, pH 7.5, 10 mM mercaptoethanol, 1 mM PMSF. After another 15 min at 0 "C, 100 ml of 18% (w/v) PEG, 1 M NaCl, 50 mM Tris-HC1, pH 7.5, 10 mM mercaptoethanol, 1 mM PMSF were slowly added with constant stirring. This solution was incubated for 40 min at 0 "C with occasional stirriig and then centrifuged at 12,000 X g for 30 min. The supernatant, fraction I (375 ml), was then used for further purification.
The topoisomerase I1 activity elutes slightly ahead of t h e topoisomerase I. The two enzymes were pooled as indicated in Fig. 1A. The topoisomerase II pool is fraction 11. Further purification of topoisomerase I has been described previously (25); further purification of topoisomerase I1 is described below.

Gel
Filtration on %io-Gel A-1.5m"Fraction IV was loaded directly on a Bio-Gel A-1.5m column (1.5 x 85 cm) equilibrated with 0.2 M potassium phosphate, pH 7.0, in solution A, and the column was washed with the same buffer. Topoisomerase I1 activity eluted at 68 ml in a symmetrical peak (fraction V). Column fractions were also assayed by NaDodS0,-gel electrophoresis. The enzyme activity correlated with a single Coomassie blue-stained polypeptide of i' vl, = 172,000, and no other polypeptides were detectable in the active fractions (Fig. 2). Since this polypeptide copurified with An additional 30% of loaded activity trailed into the topoisomerase I pool, To reduce topoisomerase I contamination, this activity was discarded.
'Estimate of protein concentration was based on intensity of Unpublished experiments in collaboration with R. Calendar.
Coomassie blue-stained band on an NaDodSOI gel. the topoisomerase activity throughout the purification, and since it cosedimented with it on a sucrose gradient (see below), we assume that this protein is responsible for the activity. Therefore, fraction V is virtually homogeneous.

Stability of the Enzyme
In preliminary purifications of topoisomerase I1 we used frozen cells and found that the stability of all fractions, even in the presence of PMSF, was very poor. Total loss of activity in all fractions was observed after several days a t 4 "C. We believe that proteolysis makes a large contribution to the instability, as on NaDodS04 gels a number of smaller molecular weight bands are generated in fractions I11 and IV during storage at 4 "C. By using fresh (never frozen) cells we found that the stability of the enzyme increased dramatically. Fraction IV is stable for 1 month at 4 "C and then loses activity with a half-time of approximately 1 month. At -20 "C, the enzyme is stable for a t least 3 months. Fractions 1-111 have a half-life of approximately 3 weeks a t 4 "C. The tremendous loss of activity during the gel filtration step may not be due to instability. We believe it is due to adsorption of the enzyme to the Bio-Gel resin, as recoveries improved if the column was used several times.

Calculation of Molecular Weight
As shown in Fig. 2, the subunit molecular weight of topoisomerase I1 is 172,000. We determined the native molecular weight from its hydrodynamic properties. A portion of fraction I11 (0.1 ml containing 3700 units) was mixed with catalase, aldolase, thyroglobulin, and ferritin (200 pg each) as markers and centrifuged on a 5-20% sucrose gradient. The sedimentation coefficient was determined by comparing the sedimentation rate of topoisomerase (measured by decatenation assay) with the sedimentation rate of the markers. The sedimentation coefficient is 9.2 s. In a separate experiment, the decatenation activity sedimented together with a polypeptide of M, = 172,000, and no other detectable polypeptide cosedimented with activity (data not shown). The Stokes radius was determined by comparing the elution position on a Bio-Gel A-1.5m column of the topoisomerase with those of the same markers used for sedimentation. T o calculate the Stokes radius, (-log K.v)t2 was plotted against elution volume for the markers and a straight line was obtained (30). From the elution volume of topoisomerase 11, its Stokes radius was found to be 78 A. The molecular weight of the native protein was calculated from its sedimentation coefficient and Stokes radius by the method of Siege1 and Monty (30) to be 309,000. Since the subunit M, = 172,000, the active enzyme appears to be a dimer. Fig. 3 shows examples of the reactions catalyzed by HeLa topoisomerase 11. Lanes a-c show the effect of the enzyme on negatively supercoiled pBR322 DNA. No relaxation is detectable in the absence of ATP (lane c ) , but relaxation is extensive in the presence of ATP (lane 6). The enzyme will relax positive supercoils as well. If the pBR322 DNA in the standard assay is positively supertwisted by addition of ethidium bromide (2 pg/ml), this DNA is readily relaxed by topoisomerase I1 (data not shown). after electrophoresis on an NaDodSO.,-5% polyacrylamide gel. Portions (one-fourth of the total volume) of each indicated fraction of the Bio-Gel column were precipitated by addition of one-third volume of 50% trichloroacetic acid, 2 mg/ml of deoxycholate. After centrifugation in a Beckman microfuge, the protein was resuspended in sample buffer (10% glycerol, 5% /l-mercaptoethanoI, 3% NaDodSO.,), and titrated to pH 6.8 with 1 M Tris base. The samples were boiled for 2 min before loading on the gel. Molecular weights were determined by comparison with standards. E shows the amount of pBR322 relaxation activity present in each fraction. Relaxation was assayed on 1 pI of each fraction as described under "Materials and Methods." In the presence of the DNA-condensing agents spermidine or histone H1, the topoisomerase can catenate plasmid DNA (Fig. 3). Lune g shows plasmid DNA (untreated). When treated with topoisomerase I in the presence of spermidine (lanes h and]), no catenation is observed. When treated with topoisomerase I1 in the presence of spermidine, the DNA forms catenanes, each consisting of many interlocked circles which do not enter the gel (lane i ) . Catenation is also observed if histone H1 (8 pg/ml) is used as condensing agent.

Reactions Catalyzed by Topoisomerase 11
HeLa topoisomerase I1 can also unknot topologically knotted P4 DNA (lanes l and n, Fig. 3). pBR322 DNA which had been knotted by treatment with T4 topoisomerase I1 is also unknotted by HeLa topoisomerase I1 (data not shown).

Requirements for Activity
The components of the standard reaction mixture, described under "Materials and Methods," are present at optimal concentrations. If Mg' or ATP is omitted, no activity is detected. Mn", at an optimum concentration of 1 mM, can substitute for Mg", but activity is about 50% of that obtained with Mg", dATP can fully substitute for ATP; however, no reaction is observed if UTP, GTP, or CTP is used as the nucleotide cofactor. The activity is reduced to about 50% if dithiothreitol is omitted. Using Tris-HC1 buffers, the enzyme exhibited a broad pH optimum from about 7.5 to 9 with about 50% activity at pH 7.3. Addition of adenosine 5'-O-(thiotriphosphate) (50 p~) to the standard reaction mixture inhibits the reaction by about 50%. Since adenosine 5'-0-(thiotriphosphate) is probably poorly hydrolyzed, this result suggests ATP hydrolysis may be required for activity. In agreement with this possibility is the presence of an ATPase activity in the most pure preparation of topoisomerase. ATP is hydrolyzed to ADP in a DNA-dependent reaction (Table 11).

Inhibitors
We have studied the effect of several procaryotic topoisom- TARLE I1 ATPase activity of purified topoisomerase II Assays were carried out under the standard conditions described under "Materials and Methods" except that the ATP was labeled with "H, the DNA was pBR322. and the incubation was for 60 min. Chromatography of the reaction products on polyethyleneimine-cellulose showed that the ATP was hvdrolvzed to ADP. erase inhibitors on HeLa topoisomerase 11. In the standard assay, 50% inhibition is obtained with 100 pg/ml of oxolinic acid, 500 pg/ml of nalidixic acid, and 200 pg/ml of novobiocin. These concentrations are significantly higher than those re-quired to inhibit E. coli DNA gyrase by 50% (10 pg/ml for oxolinic acid (31), 200 pg/ml for nalidixic acid (31), and 1 pg/ ml for novobiocin (32)).

HeLa Topoisomerase 11 Is a Type 11 Topoisomerase
Since this enzyme can decatenate and catenate covalently closed circular DNA, it is presumably a type I1 enzyme. To provide further evidence for this classification we treated a topoisomer of a unique linking number, n, with HeLa topoisomerase I or HeLa topoisomerase 11. As shown by gel electrophoresis, the topoisomerase I1 products are topoisomers with linking numbers n and n-2 (Fig. 4, lane c ) . Linking number changes in steps of two are characteristic of a type I1 topoisomerase (12, 13). In contrast, the HeLa topoisomerase I products differ in linking number by steps of one (lanes b  and d, Fig. 4).

DISCUSSION
We have purified a type I1 topoisomerase from HeLa nuclei to apparent homogeneity, using an assay which depends on the ability of type I1 enzymes to decatenate covalently closed circular DNA. This assay has allowed us to detect and quantitate type I1 activity in crude extracts without interference from type I topoisomerases. Besides its ability to decatenate DNA circles, we have demonstrated that the enzyme can relax positive and negative supercoils and unknot topologically knotted DNA circles. No gyrase activity is detectable. The enzyme requires ATP for activity, and consistent with this requirement is the presence of a DNA-dependent ATPase. The most pure preparation contains a single polypeptide of M, = 172,000 as measured by NaDodS04-gel electrophoresis. This polypeptide correlates well with activity throughout the purification and in sedimentation in a sucrose gradient. The native molecular weight, calculated from hydrodynamic properties, is 309,000, suggesting that a dimer of M, = 172,000 subunits is the active species.
This enzyme shares several properties with other type I1 topoisomerases. The T4 topoisomerase, which has similar catalytic properties, probably has three subunits with M, = 63,000, 52,000, and 15,000 (7). No data on molecular weight of the native T4 enzyme is available. Type I1 topoisomerases, which probably are similar to HeLa topoisomerase I1 and T4 topoisomerase, have been detected in Drosophila embryos (18) and Xenopus germinal vesicles (22). These enzymes differ from the prokaryotic DNA gyrase in that they are unable to induce negative supercoiling (7, 18,22). Nevertheless, all the type I1 enzymes may be related in their reaction mechanisms. The mechanism involves the transient introduction of an enzyme-bridged double-stranded break and the passage of an intact DNA helix through the break. The ends of the broken DNA helix cannot rotate freely relative to one another during this event. This mechanism accounts for the observed change in linking number in steps of two (12, 17-19).
There is a clear need for energy input (ATP hydrolysis) in the supercoiling reaction catalyzed by gyrase. However, the requirement for ATP hydrolysis by other type I1 enzymes, which have not been shown to supercoil DNA, is less obvious and remains a matter of speculation. The ATPase activity in T4 topoisomerase is stoichiometrically related to strand passage (7). The HeLa topoisomerase I1 ATPase may also be related to strand passage, as we calculate that roughly the same number of molecules of ATP are hydrolyzed as strands passed. The HeLa and T4 enzymes may actually have some energy-requiring functions (e.g. gyrase activity) in uiuo. Alternatively ATP may be required simply in the strand passage mechanism (3,4, 7). One difference between gyrase and other type I1 enzymes may be that gyrase can bind the DNA exclusively in a positive supercoil (33). The other type I1 enzymes may not be able to wrap the DNA to generate a positively supercoiled domain, or may require factors such as a specific DNA sequence or other proteins to orient directionally on the DNA.
Type I1 topoisomerases may function in a number of important genetic processes. In replication they may play a role in initiation, elongation, and termination. In initiation they may serve to separate the strands at the replication origin in an origin-specific gyrase reaction (7). In elongation, they may relax positive supertwists which are generated by unwinding of the template strands. In termination, they may assist segregation of daughter molecules by passing one helix through another (34). Topoisomerases are thought to be important in transcription; for example, supercoiling may increase the activity of certain promoters (3). They may also facilitate assembly of DNA into its condensed chromosomal state. The ability of type I1 topoisomerases to pass one helix through another could be important in generating and maintaining this complex organization. Further studies will reveal the precise role played by HeLa topoisomerase I1 in these and other processes.