Purification of Topoisomerase I1 from Amsacrine-resistant P388 Leukemia Cells EVIDENCE FOR TWO FORMS OF THE ENZYME*

Topoisomerase I1 was purified from an amsacrine- resistant mutant of P388 leukemia. A procedure has been developed which allows the rapid purification of nearly homogeneous enzyme in quantities sufficient for enzyme studies or production of specific antisera. The purified topoisomerase I1 migrated on sodium dodecyl sulfate-polyacrylamide gel electrophoresis as two bands with apparent molecular masses of 180 (pl80) and 170 kDa (p170); both proteins unknotted P4 DNA in an ATP-dependent manner and displayed amsacrine-stimulated covalent attachment to DNA. Staphylococcus V8 protease cleavage patterns of p170 and pl80 showed distinct differences. Specific poly- clonal antibodies to either p170 or p180 recognized very selectively the form of the enzyme used to generate the antibodies. Immunoblotting with these specific antibodies showed that both p180 and p170 were present in cells lysed immediately in boiling sodium dodecyl sulfate. Comparison of the purified topoisomerase I1 from amsacrine-resistant P388 with that from amsa-crine-sensitive P388 demonstrated that each cell type contained both pl80 and p170;

Topoisomerase I1 was purified from an amsacrineresistant mutant of P388 leukemia. A procedure has been developed which allows the rapid purification of nearly homogeneous enzyme in quantities sufficient for enzyme studies or production of specific antisera. The purified topoisomerase I1 migrated on sodium dodecyl sulfate-polyacrylamide gel electrophoresis as two bands with apparent molecular masses of 180 (pl80) and 170 kDa (p170); both proteins unknotted P4 DNA in an ATP-dependent manner and displayed amsacrine-stimulated covalent attachment to DNA. Staphylococcus V8 protease cleavage patterns of p170 and pl80 showed distinct differences. Specific polyclonal antibodies to either p170 or p180 recognized very selectively the form of the enzyme used to generate the antibodies. Immunoblotting with these specific antibodies showed that both p180 and p170 were present in cells lysed immediately in boiling sodium dodecyl sulfate. Comparison o f the purified topoisomerase I1 from amsacrine-resistant P388 with that from amsacrine-sensitive P388 demonstrated that each cell type contained both p l 8 0 and p170; however, the relative amounts of the two proteins were consistently different in the two cell types. The data strongly suggest that p170 is not a proteolytic fragment of pl80. Thus, P388 cells appear to contain two distinct forms of topoisomerase 11.
Topoisomerases are enzymes which control the topological state of DNA (for reviews, see Refs. [1][2][3]. Type I1 topoisomerases catalyze DNA strand passage through transient double strand breaks in the DNA; this ability to change the linking number of DNA allows the enzymes to mediate DNA interconversions such as supercoiling (demonstrated for prokaryotic only) and relaxation of supercoiling, catenation and decatenation, and knotting and unknotting (1)(2)(3). Topoisomerase I1 appears to be a component of the nuclear scaffold (4, 51, perhaps located near the base of chromatin loops (6,7). Indirect evidence suggests the enzyme may be associated with gene enhancer regions (8). These enzymes have been implicated in a number of critical cellular processes including DNA * This work was supported in part by Grant CA40884 from the National Cancer Institute. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom reprint requests should be addressed.
An important feature of the topoisomerase 11-mediated DNA strand passage reaction is the covalent attachment of the enzyme to the DNA via a tyrosine-DNA phosphodiester linkage; under denaturing conditions the enzyme can be trapped in this form, resulting in topoisomerase-associated DNA strand breakage (12). Of particular interest is the observation that several structurally diverse classes of antitumor agents such as anthracyclines, anthracenediones, epipodophyllotoxins, and ellipticines greatly increase the number of topoisomerase 11-associated strand breaks in DNA (13)(14)(15). This has led to the suggestion that topoisomerase I1 is a common target of these drugs (14). Consistent with this idea has been the finding that several cell lines which have developed resistance to one class of topoisomerase 11-active compounds are cross-resistant to other topoisomerase 11-active compounds (16)(17)(18)(19). For example, a murine leukemia cell line which has developed resistance to amsacrine (P388/A20)' is cross-resistant to teniposide, bisantrene, and doxorubicin, but not to the topoisomerase I-active drug, camptothecin; resistance to these drugs could not be explained by differential uptake (20). Furthermore, the amsacrine-resistant cell line contained 2-3-fold less topoisomerase I1 activity than its parental line, as well as reduced topoisomerase I1 immunoreactivity (20). These data suggest that resistance to amsacrine in P388/A20 may be due to changes in the expression or chemosensitivity of topoisomerase 11.
To test the hypothesis that resistance to amsacrine in P388j A20 is due to an alteration in topoisomerase I1 and to permit detailed study of the enzyme, topoisomerase I1 has been purified from P388/A20 cells. A purification scheme is described which is both rapid and reproducible and provides quantities of highly purified topoisomerase I1 sufficient for enzyme studies and development of topoisomerase-specific antisera. Topoisomerase I1 from amsacrine-resistant P388 cells was purified as two distinct polypeptides of 180 and 170 kDa. Both polypeptides had topoisomerase I1 activity. Differences in the antigenicity and proteolytic cleavage patterns of the two forms of the enzyme strongly suggest that the 170-kDa enzyme is not simply a proteolytic fragment of the 180-tease was from Worthington. Nitrocellulose (0.2-pm pore) was from Schleicher and Schuell. CHAPS and IODO-GEN were obtained from Pierce Chemical Co.
Purification of Topoisomerase 11-All steps were performed at 0-4 "C. Chromatography was done on a Pharmacia Biotechnology, Inc. fast protein liquid chromatography system composed of two P-500 pumps and an LCC-500 controller.
Topoisomerase Extraction-On day 7 postinoculation, the ascites tumors were removed from syngeneic mice and placed on ice. The cells were washed three times in 4-5 volumes of PBS containing 1 mM PMSF, 1 mM benzamidine, 1 pg/ml soybean trypsin inhibitor and centrifuged for 5 min at 1000 X g in a Beckman J-6M centrifuge. The cells were resuspended in 4 volumes (1.5-2.0 X 10s cells/ml) of 5 mM potassium phosphate, pH 7.0,2 mM MgClZ, 0.1 mM EDTA, 1 mM PMSF, 1 mM benzamidine, 10 pg/ml soybean trypsin inhibitor, 50 pg/ml leupeptin, 10 mM 2-mercaptoethanol and stirred slowly with a magnetic stirring bar at 4 "C for 15 min. Nuclei were prepared by lysis of the cells with 10 strokes from a chilled Dounce homogenizer. The extent of cell lysis was monitored by microscopy. The lysate was centrifuged for 10 min at 1000 X g. The nuclei were washed twice with 5 volumes of 1 mM potassium phosphate, pH 6.5, 5 mM MgClz, 1 mM EGTA, 10% glycerol, 100 mM NaCl, 1 mM PMSF, 1 mM benzamidine, 10 pg/ml soybean trypsin inhibitor, 50 pg/ml leupeptin, and 10 mM 2-mercaptoethanol by centrifuging at 1000 X g for 5 min.
The washed nuclei were resuspended to 2 X 108/ml in extraction buffer (5 mM potassium phosphate, pH 7.0, 2 mM MgC12, 0.1 mM EDTA, 1 mM PMSF, 1 mM benzamidine, 10 pg/ml soybean trypsin inhibitor, 50 pg/ml leupeptin, 10 mM 2-mercaptoethanol, 10% glycerol). NaCl (5 M) was added slowly to give a final concentration of 0.35 M, and topoisomerase I1 was extracted for 60 min at 4 "C with constant stirring. The extract was centrifuged at 3,000 rpm for 15 min in a Beckman J-6M; the supernatant was further centrifuged at 25,000 X g in a Beckman J2-21M centrifuge for 15 min. The final supernatant was the 0.35 M NaCl extract.
Ultrogel Hydroxylapatite-A 1.6 X 21-cm column of Ultrogel hydroxylapatite was prewashed with four gradients of 0.2 M potassium phosphate, pH 7.0, 10% glycerol, 10 mM 2-mercaptoethanol to 1.0 M potassium phosphate, pH 7.0, 10% glycerol, 10 mM 2-mercaptoethanol, followedby equilibration in the 0.2 M potassium phosphate buffer. The 0.35 M NaCl extract was loaded onto the column at 1 ml/min, either through a 50-ml superloop or directly through pump A (without the in-line filter). The column was then washed with the 0.2 M potassium phosphate buffer until the Am returned to base line.
Topoisomerase I1 was eluted with a 100-rnin linear gradient of 0.2-1.0 M potassium phosphate buffers at 1.25 ml/min. Four-min fractions were collected (5 ml). Fractions were assayed for P4 unknotting activity (described below) and active fractions pooled.
Mono Q or Fast Flow Q-Either a Mono Q (0.5 X 5 cm) or Fast Flow Q column (1.0 X 3.5 cm) was washed extensively according to manufacturer's instructions (ix. 2 N NaOH, 2 M NaCl, 50% acetic acid) followed by four gradients of 0-1 M NaCl in Q buffer (20 mM Tris-HC1, pH 7.5, 10% glycerol, 10 mM 2-mercaptoethanol, 0.5 mM EDTA). The column was then equilibrated in 0.2 M NaCl in Q buffer. tite were diluted with Q buffer to a final PO, concentration below 100 Immediately before loading, the pooled fractions from hydroxylapa-mM and loaded onto the column from a superloop at 0.5-1 ml/min. (Alternatively, the pooled hydroxylapatite fractions were loaded directly through pump B and diluted with the appropriate amount of Q buffer through pump A.) After loading, the column was washed with 5-10 column volumes of 0.2 M NaCl in Q buffer. Topoisomerase I1 was eluted with a 30-40-min linear gradient of 0.2-0.6 M NaCl in Q buffer at 0.5 ml/min. One-min fractions were collected (0.5 ml). stopped by the addition of 2 p1 of stop solution (5% SDS, 50 mM EDTA, 50% glycerol, 0.5 mg/ml bromphenol blue). The reaction mixture was electrophoresed on a 0.7% agarose gel in TPE buffer (80 mM Tris phosphate, 2 mM EDTA); the gel was then stained with ethidium bromide and photographed with Polaroid type 55 film. Photographic negatives were scanned with a Beckman DUdB spectrophotometer to quantify the amount of unknotted P4 DNA (23).
SDS-PAGE-Samples were electrophoresed on a Bio-Rad Protean I1 (160 X 160 X 0.75 mm) according to the method of Laemmli (24). The gels (7%) were run at 6 mA/gel for 14-16 h and stained with silver according to the method of Merril et al. (25).
Topoisomerase II Antibodies-Purified topoisomerase I1 was electrophoresed on SDS-polyacrylamide gels to separate the 180-kDa band from the 170-kDa band (approximately 25 pg of each band). After staining of the gel with Coomassie Blue, the individual bands were excised from the gel and placed into PBS. The gel slices were crushed with a Dounce homogenizer, an equal volume of Freund's complete adjuvant was added, and the samples were emulsified by brief sonication. New Zealand White rabbits were injected intradermally or subcutaneously with the antigen. Booster injections (approximately 10 pg/band SDS-PAGE separated topoisomerase) were given after 3-5 weeks and every 3-4 weeks thereafter. Animals were bled from the central ear artery 7-13 days after each boost. Sera were stored at -70 "C. Western Transfer and Immunoblotting-Following SDS-PAGE, gels were electrophoretically transferred (Hoefer Transfor, TE 52) to nitrocellulose at 400 mA for 4-6 h (26). Blots were blocked with 0.5% gelatin in PBS and probed with anti-topoisomerase I1 antibodies appropriately diluted into WBWB (25 mM Tris-HC1, pH 7.5, 150 mM NaCl, 5 mM EDTA, 0.1% gelatin, 0.25% Triton X-100) containing 1% normal goat serum; blots were incubated overnight at room temperature. Following three 5-min washes in WBWB, the blots were incubated in a 1:lOO dilution of AuroProbe BL GAR, if insufficient staining intensity was observed, the probes were silver-enhanced with IntenSE according to manufacturer's instructions. Alternatively, imjugated goat anti-rabbit IgG (Protoblot, Promega Biotec) according munoreactive material was detected with alkaline phosphatase-conto manufacturer's instructions. Staining of total protein on nitrocellulose blots was done either with AuroDye according to manufacturer's instructions or with India ink (27).
Inhibition of P4 Unkmtting Activity by Anti-pl70 Antibody-Anti-p170 serum or preimmune serum was precipitated with 50% ammonium sulfate; the IgG precipitate was redissolved in PBS and desalted p180 were first diluted into 10 mM Tris-HC1, pH 7.5,1% CHAPS, 10 over G-50. Fractions from Mono S containing predominantly p170 or mM 2-mercaptoethanol and then added to an equal volume of either preimmune IgG or various dilutions of anti-pl70 I& in preimmune IgG. After 3 h on ice, a 5-p1 aliquot of this was assayed for P4 unknotting activity as described above.
Affinity Purification of Antibody-For affinity purification of an-tibodies, the serum was first precipitated with 50% ammonium sulfate and desalted over G-50. The IgG fraction was diluted in WBWB and used to probe a nitrocellulose blot of purified topoisomerase I1 as described above. The topoisomerase band was cut from the nitrocellulose sheet, and the antibodies were extracted as described by Smith and Fisher (28). The extracted antibodies were diluted in WBWB/ 1% normal goat serum and stored at 4 "C with 0.025% sodium azide.
Covalent Transfer of "P from DNA to Protein-Covalent transfer of 32P from DNA to p170 and p180 was performed according to Rowe et al. (29,30). Briefly, a 3.1-kilobase BamHI fragment of plasmid pGFc5a (generously provided by Dr. Maxwell Lee, Duke University) was labeled with [32P]dCTP by the random priming technique (Prime Time, IBI Technologies). To this labeled DNA (2.5 ng) was added 1 pg of purified topoisomerase I1 (Mono S fractions pooled to contain approximately equal amounts of p170 and p180) in 50 mM Tris-HC1, pH 7.5, 100 mM KC1, 10 mM MgClz, 0.5 mM EDTA, 1 mM ATP, 30 pg/ml bovine serum albumin, either in the absence or presence of 40 pg/ml amsacrine. After 10 min at 37 "C, the reaction was stopped by the addition of SDS to 1%. An equal volume of 40 mM Tris, pH 8.0, 25 mM CaCl, was added, and the DNA was digested with 4 units of Ba131 nuclease (Bethesda Research Laboratories) for 60 min at 30 "C.
The sample was then boiled for 10 min in an equal volume of SDS-PAGE sample buffer and electrophoresed as described above. Following silver staining, the gel was dried and exposed to Kodak XAR-5 film for 18 h with an intensifying screen.
Proteolytic Fragmentation of Topoisomerase ZI-Comparison of the proteolytic cleavage patterns of the 180-and 170-kDa proteins was done by the method of Cleveland et al. (31). Briefly, either purified topoisomerase I1 or radioiodinated purified topoisomerase I1 was electrophoresed on a 7% SDS-polyacrylamide gel and the bands of interest located by Coomassie Blue staining or autoradiography of the wet gel. The bands were carefully excised from the gel and loaded into the wells of a 7-23% gradient SDS-polyacrylamide gel. The gel slices were overlaid with 20% glycerol in Cleveland buffer (31), followed by the appropriate amount of Staphylococcus V8 protease in 10% glycerol in Cleveland buffer. The samples were electrophoresed at 6 mA/gel for 14-16 h and either stained with silver (25) or dried and exposed to Kodak XAR-5 film for 2-5 days at -70 "C with an intensifying screen.
Protein Determination-Protein concentrations were measured by the method of Bradford (34) with lysozyme as a standard.

RESULTS
The first step in the purification of topoisomerase TI was a selective extraction of the enzyme from isolated P388/A20 nuclei with 0.35 M NaCl (35). This procedure extracts approximately 90% of the topoisomerase I1 from the isolated nuclei as judged either by the ability to re-extract topoisomerase I1 with more harsh conditions (e.g. 1 M NaCl or 1% CHAPS) or by immunoblot analysis of the extracted nuclei (data not shown). Selective extraction with 0.35 M NaCl also removed about 80% of the total cellular protein ( Table I).
Protease inhibitors were included in each step of the nuclear isolation and 0.35 M NaCl extraction; each inhibitor was freshly prepared and added to the buffers just prior to use.
The 0.35 M NaCl extract was immediately loaded onto a column of Ultrogel hydroxylapatite which had been equilibrated with 0.2 M potassium phosphate. Greater than 90% of the protein which was loaded onto the column either passed through without binding or was eluted with a 0.2 M potassium phosphate wash (data not shown). Bound protein was eluted from the column with a gradient of 0.2-1.0 M potassium phosphate; topoisomerase I1 activity began to elute at 0.55-0.6 M potassium phosphate and formed a broad trailing peak. About 60% of the topoisomerase I1 activity loaded onto the hydroxylapatite column was recovered (Table I).
Active fractions from hydroxylapatite were pooled and applied to either Mono Q or Fast Flow Q. Mono Q gave the best separation of topoisomerase I1 but had a greater tendency to build up pressure during sample loading. Because of the high ionic strength of the sample after hydroxylapatite, the potassium phosphate concentration had to be reduced before loading. This was done by diluting the sample in 20 mM Tris-HC1, pH 7.5, 10% glycerol, 10 mM 2-mercaptoethanol (Q buffer) immediately before loading. Topoisomerase I1 activity eluted as a single peak at about 0.4 M NaCl and was associated with the appearance of two bands on SDS-PAGE at 180 (p180) and 170 kDa (p170) (Fig. 1, panels A-C). Recovery of topoisomerase I1 activity from Mono Q or Fast Flow Q was usually 25-50% but occasionally dropped below 10% (Table I). Sudden losses of topoisomerase I1 activity at the Q column step occurred more frequently when P388/A20 was used as the source of the enzyme than when P388/WT20 was used (data not shown). This likely represents enzyme denaturation (e.g. due to enzyme dilution) and may reflect a greater lability of the enzyme from the amsacrine-resistant cells. However, it may also reflect loss of the enzyme or perhaps even removal of an unknown stimulatory factor. The final step in the purification was chromatography of the pooled Q fractions over Mono S, which separated the remaining proteins to give a nearly homogeneous preparation (Fig. 2 A ) . Topoisomerase I1 binds to both Mono Q (anion exchange) and Mono S (cation exchange) at pH 7.5, requiring about 400 mM NaCl for elution from either column; proteins which co-purified with the enzyme over Q showed different elution profiles on Mono S. Depending upon the Q fractions that were pooled, minor contaminating proteins were sometimes observed in the purified preparations (e.g. Fig. 2A  A duplicate gel to p a n e l B was transferred to nitrocellulose and probed with anti-pl7O antiserum (1:lOOO dilution); the immunoreactive material was visualized with AuroProbe BL GAR. approximately 225 ng. In p a n e l B, a duplicate gel to p a n e l A was blotted onto nitrocellulose and probed with affinity-purified anti-pl80 antibodies. 100-fold above the 0.35 M NaCl extract (Table I); this repre-allow the two proteins to be assayed for topoisomerase I1 sents about a 500-fold purification of the enzyme from whole activity independently (compare fractions 10 and 11 with cells (assuming a 90% extraction of the enzyme). The rela-fractions 15 and 16). This type of analysis strongly suggests tively low fold purification may reflect loss of activity of the that both p170 and p180 possessed topoisomerase I1 activity. enzyme during purification, as the enzyme activity was found Further evidence that both proteins have topoisomerase I1 to decrease a t 4 "C with a half-life of only several days, activities is shown in Fig. 4, where specific polyclonal antiespecially when the enzyme was in dilute solutions. The bodies to p170 (described below) were used to inhibit P4 topoisomerase I1 activity of the purified preparation could be unknotting activity. Addition of anti-pl70 antibodies to a maintained for several months, however, when stored at -20 "C in 50% glycerol.
Topoisomerase I1 activity correlated very closely with the appearance of p170 and p180 (Fig. 3). The SDS-polyacrylamide gel shown in Fig. 2A was densitometrically scanned to give the relative amounts of p170 and p180 in each fraction (Fig. 3B, left panel), and each fraction was assayed for P4 unknotting activity (Fig. 3A). Topoisomerase I1 activity did not correlate with either p170 or p180 alone (Fig. 3B, left  panel) but was very well correlated with the sum of the two sample containing primarily p180 (approximately 20% p170 as determined by densitometric scanning) resulted in only a slight inhibition of P4 unknotting activity (Fig. 4, A and B). When added to a sample containing only p170, however, the same amount of anti-pl7O antibodies resulted in almost complete inhibition of P4 unknotting activity (Fig. 4, A and B). Inhibition of P4 unknotting activity with specific anti-pl70 antibodies demonstrated that p170 possessed topoisomerase I1 activity. The inability to significantly inhibit activity in a sample containing primarily p180 clearly indicated that a proteins (Fig. 3B, right panel). An advantage of Mono S was second topoisomerase I1 activity was present. Although it that it provided sufficient separation of p170 from p180 to appeared likely that the second topoisomerase I1 activity resided in p180, this could not be determined immunologically, since anti-pl80 had no effect on P4 unknotting activity of either p170 or p180 (data not shown).
To conclusively demonstrate that both p170 and p180 possessed topoisomerase I1 activity, an experiment was done which takes advantage of the ability to trap an intermediate in the topoisomerase reaction in which the enzyme is covalently attached to DNA. As Rowe et al. (29,30) previously demonstrated, denaturation of the complex formed between unlabeled topoisomerase I1 and 32P-labeled DNA results in covalent incorporation of 32P label into the topoisomerase 11; by digesting the bulk of the DNA with Ba131 nuclease followed by SDS-PAGE and autoradiography, it can be determined which proteins in the sample mediate the DNA cleavage reaction. As shown in the left lane of Fig. 5, when a sample of purified topoisomerase I1 containing p170 and p180 was subjected to this type of analysis, both proteins incorporated the 32P label. In the absence of added topoisomerase 11, no 32P bands were seen (data not shown). The addition of amsacrine Lane F shows total protein staining with India ink (27).
to the reaction caused an increase in the amount of 32P incorporated into the proteins (Fig. 5, right hne), as expected for topoisomerase I1 (30). Thus, it is clear that both p170 and p180 have topoisomerase I1 activity.
To determine the relationship between the two forms of topoisomerase 11, two types of studies were undertaken. The amounts of p17O and p180 were electrophoresed on a 7% SDSpolyacrylamide gel and located by Coomassie Blue staining; alternatively, radioiodinated topoisomerase I1 was included in the sample, and p170 and p180 were located by autoradiography of the wet gel. The individual bands were excised and P388/WT20 and P388/A20 cells were grown in RPMI medium containing 20% fetal calf serum, 10 mM 2-mercaptoethanol in 75-cm2 flasks. Logarithmically growing cells were washed twice by centrifugation a t 1000 X g for 5 min in PBS containing 10 pg/ml soybean trypsin inhibitor, 1 mM benzamidine, 1 mM PMSF, 2 mM EDTA, 50 pg/ml leupeptin. The washed cells were immediately lysed in electrophoresis sample buffer (containing 2% SDS) in a boiling water bath. Equal amounts of protein (150 pg/lane) were loaded onto a 7% SDSpolyacrylamide gel; electrophoresis and transfer to nitrocellulose were as described under "Experimental Procedures." The nitrocellulose blots were probed with either ( A ) anti-pl70 antisera (1:250 dilution) or ( B ) affinity-purified anti-pl80 antibodies. Detection of the immunoreactive bands was done with alkaline phosphatase-conjugated secondary antibody.
proteolyzed in a second SDS-polyacrylamide gel according to the method of Cleveland et al. (31). As can be seen in Fig. 6, the proteolytic cleavage patterns of the two proteins were significantly different. Clearly, peptide fragments were produced by V8 protease digestion of p170 that were not generated by p180. Furthermore, intermediate fragments generated by low concentrations of V8 protease were also different between the two. Protein p170 was proteolyzed more readily by V8, but increasing amounts of V8 protease (up to 2.5 pg, data not shown) still produced many peptide fragments from each protein that were not formed from the other at any V8 amount. A low level of protein degradation was always observed in the absence of protease under the incubation conditions used (Fig. 6A, lanes 2 and 3); the cause of this proteolysis is not known. However, even this spontaneous proteolysis of the two proteins yielded distinctly different fragments.
The second type of comparison of p170 and p180 was done by analysis of their antigenic similarity. Specific polyclonal antibodies to either p170 or p180 were generated in rabbits and tested for their ability to cross-react with the different proteins. As described above (Fig. 4), antibodies to p170 were able to inhibit the catalytic activity of p17O but had very little, if any, effect upon p180 catalytic activity. As can be seen in Fig. 7A, lane B, antibodies raised against p170 were highly selective for that protein even when equal amounts of purified p170 and p180 were probed. Antisera against p180 was of lower titer but showed similar selectivity (data not shown); after affinity purification of the p180 antibody, it showed no opoisomerase II 16745 cross-reactivity with p170 (Fig. 7A, lane C). The same specificity was observed when a 0.35 M NaCl extract of P388/A20 nuclei was probed (Fig. 7B), indicating that the antigenic differences observed in the purified preparations were not an artifact of the purification.
The results suggested that two forms of topoisomerase I1 were present in the P388/A20 cell line. Since this cell type was selected for its resistance to amsacrine, it was of interest to determine whether the P388/WT20 cells also possessed the p170 and p180 forms of topoisomerase 11. Fig. 8 shows topoisomerase I1 purified from both cell types. Peak fractions from Mono S chromatography (fractions 10-15) were pooled and the samples run on SDS-PAGE and stained with silver. Comparison of the samples shows that each cell type contained both p170 and p180, though the ratio of the two bands was different. In each preparation, the relative amount of p170 was greater in the P388/WT20 cells. This ratio difference was also observed by immunoblotting of cell lysates that had been prepared by immediate lysis of the cells in boiling SDS (Fig. 9). This clearly demonstrated that p170 and p180 were present in the cells (and were not an artifact of extraction or purification) and indicated that the increased amount of p170 found in purified topoisomerase I1 preparations from P388/WT20 cells was not due to selective enrichment of p17O during purification. In addition, p170 and p180 antisera showed the same selectivity in the P388/WT20 as was demonstrated in P388/A20 (Fig. 9). Thus, while the relative amounts of two forms of topoisomerase I1 in the amsacrinesensitive and -resistant P388 cells were different, it is clear that each cell type contained both forms.

DISCUSSION
Topoisomerase I1 was purified from P388/A20 cells to near homogeneity by a rapid highly reproducible purification scheme. The procedure requires less than three days, reducing not only the amount of time devoted to enzyme purification, but also minimizing the loss of activity of this labile enzyme. From approximately 2 X 10" P388 cells, one can typically obtain several hundred micrograms of purified topoisomerase 11. The availability of this quantity of purified enzyme has made it possible to generate antibodies which are highly specific for topoisomerase 11.
The enzyme from P388/A20 cells was purified as two bands of 180 and 170 kDa on SDS-PAGE. Both forms of the enzyme unknotted P4 DNA in an ATP-dependent manner, an activity that can be accomplished only by a type I1 topoisomerase (22). Each form of the enzyme displayed amsacrine-stimulated covalent attachment to DNA. It appears highly unlikely that p170 is simply a proteolytic fragment of p180 for the following reasons: 1) fingerprinting analysis of the proteins by V8 protease revealed significantly different patterns; 2) specific polyclonal antibodies to the two proteins were highly selective for the protein used to generate the antibodies, both in inhibition of catalytic activity and in immunoblotting; 3) the ratio of the two proteins was consistent within a given cell type; 4) multiple protease inhibitors were used; 5) no conversion of p180 to p170 was ever observed, even after prolonged storage; 6) no other evidence of proteolysis was present, such as lower molecular weight forms; 7) the ratio of the two proteins was characteristic of a particular cell type; 8) both bands were present in cell lysates (in cells immediately lysed in boiling SDS).
P388 cells were chosen as the source of the enzyme for this study because of the availability of clones from both wild type P388 and an amsacrine-resistant mutant which may contain an altered topoisomerase I1 (20). The ability to purify the enzyme from both amsacrine-resistant and -sensitive cells will allow comparison of the enzymatic properties of each and may lead to an understanding of the mechanism of amsacrineinduced topoisomerase I1 DNA cleavage and the development of resistance to this drug. Initial studies of the two forms of topoisomerase I1 have indicated that p170 is more sensitive than p180 to inhibition by amsacrine.' Furthermore, preliminary evidence with specific antibodies to p170 and p180 demonstrates a correlation between the amount of p170 in a cell type and the cytotoxic potency of amsacrine.' Thus, the relative levels of the two forms of topoisomerase I1 in a cell may be an important determinant of amsacrine sensitivity.
It is important to emphasize that the appearance of two forms of topoisomerase I1 is not limited to the amsacrineresistant P388 cells. As shown in Figs. 8 and 9, the amsacrinesensitive P388 cells also possess both enzymes. Specific antibodies to p180 recognize the same band in P388/WT20 as in P388/A20; antibodies to p170 behave similarly. Furthermore, preliminary immunoblotting experiments with specific antibodies to p170 or p180 have shown both forms of the enzyme to be present in lysates of other cell types as well, including human cells.' Thus, it does not appear that one form of the enzyme is simply a mutant that developed during acquisition of amsacrine resistance.
Previous studies have also reported the purification of topoisomerase 11, but no evidence was presented for different forms of the enzyme that was not due to proteolysis. Miller et al. (36) found only a 172-kDa topoisomerase from HeLa cells, and a 160-kDa topoisomerase I1 was isolated from Plasmodium berghei (37). Topoisomerase I1 from yeast was purified as a "polydisperse" band of 150 kDa, indicating to the authors that some proteolysis was occurring (38). In calf thymus, bands of 125 and 140 kDa were found, though these clearly represented proteolytic products of a 180-kDa form found by immunoblotting of cell lysates (39). Tryptic maps of the two proteins showed no significant differences (39). A more recent purification of the enzyme from calf thymus found protein bands of 175 and 150 kDa (40); while no comparison of the two bands was made, the authors did state that antibodies raised against the 175-kDa protein crossreacted with the 150-kDa protein. Purification of the enzyme from Drosophila has resulted in the appearance of multiple bands on SDS-PAGE ranging from 172 to 132 kDa (41-43). As the authors point out, the most likely explanation for these multiple bands is proteolysis, since higher molecular weight species could be converted to the lower molecular weight species (41, 43), proteolytic cleavage patterns of the bands were very similar (41), and antibodies to the peptides showed cross-reactivity to the entire cluster of peptide bands (40,41). The reason that different forms of topoisomerase I1 have not been previously detected is not obvious. However, numerous differences exist between the present study and previous ones, including topoisomerase 11 extraction conditions, protease inhibitors, types of chromatography, and speed of the purification. Additionally, the type of cell used in the purification may be important. Preliminary results indicate that the amounts of p170 and p180 can be markedly different in different cell types'; perhaps the cell types used in previous purifications contained so little of one form of topoisomerase I1 that it escaped detection.
It cannot be determined from the data whether the two forms of topoisomerase I1 are due to separate genes (or different splicing of one gene) or to post-translational modification of a single gene product. Several potential modifications of the enzyme have been demonstrated in vitro, including phos-* F. Drake, unpublished observations. phorylation (44-46) and ADP-ribosylation (47), but these have not been found in uiuo. A gene for topoisomerase I1 has been cloned from yeast (48) and Drosophila (49). In yeast it was shown by gene disruption to be a single copy essential gene (48). This finding does not exclude the possibility that more than one gene for topoisomerase I1 exists in mammalian cells, however. For example, yeast has a single copy metallothionein gene, while two copies of the gene are present in the mammalian genome (50). A topoisomerase I1 gene from mammalian cells has not as yet been isolated, so no data are available to exclude multiple genes. Additional studies will be required to determine the relationship of the two forms of topoisomerase 11. In light of the importance of topoisomerase I1 to cellular function, it is interesting to speculate what the roles of two forms of the enzyme may be. For example, the enzymes may have a differential subcellular distribution, such as in their association with the chromatin scaffold. Perhaps one of these enzymes associates with particular subsets of chromatin, such as regions which are actively transcribing or replicating. The quantity of topoisomerase I1 has been shown to be cell cycledependent (51); it may be that only one form of the enzyme displays this cell cycle dependence. The two enzymes could also differ in their enzymatic properties, such as cofactor requirements or substrate specificities. Previous studies of topoisomerase I1 have suggested that the enzyme is a homodimer (1-3); perhaps p170 and p180 may interact as a heterodimer under certain conditions. The ability to purify the enzymes in active form, as well as the availability of specific antibodies to the two proteins, will allow these questions to be addressed.