The Transducer Domain Is Important for Clamp Operation in Human DNA Topoisomerase IIα*

DNA topoisomerase II is a multidomain homodimeric enzyme that changes DNA topology by coupling ATP hydrolysis to the transport of one DNA helix through a transient double-stranded break in another. The process requires dramatic conformational changes including closure of an ATP-operated clamp, which is comprised of two N-terminal domains from each protomer. The most N-terminal domain contains the ATP-binding site and is directly involved in clamp closure, undergoing dimerization upon ATP binding. The second domain, the transducer domain, forms the walls of the N-terminal clamp and connects the clamp to the enzyme core. Although structurally conserved, it is unclear whether the transducer domain is involved in clamp mechanism. We have purified and characterized a human topoisomerase IIα enzyme with a two-amino acid insertion at position 408 in the transducer domain. The enzyme retains both ATPase and DNA cleavage activities. However, the insertion, which is situated far from the N-terminal dimerization area, severely disrupts the function of the N-terminal clamp. The clamp-deficient enzyme is catalytically inactive and lacks most aspects of interdomain communication. Surprisingly, it seems to have retained the intersubunit communication, allowing it to bind ATP cooperatively in the presence of DNA. The results show that even distal parts of the transducer domain are important for the dynamics of the N-terminal clamp and furthermore indicate that stable clamp closure is not required for cooperative binding of ATP.

DNA topoisomerase II is a multifunctional and highly complex enzyme that is able to change the topological conformation of DNA in response to different physiological alterations (1)(2)(3). Topoisomerase II enzymes mediate topological changes by introducing a transient break in one DNA duplex, the G segment, while another duplex, the T segment, is coordinately transported through the gated DNA. During the process, ATP is required to drive the enzyme through a series of dramatic conformational changes dependent on both interdomain and intersubunit communication (4 -8).
Eukaryotic topoisomerase II is a homodimeric enzyme consisting of three distinct regions. The N-terminal and core regions, containing the two catalytic entities, are highly conserved among eukaryotic organisms and also share homology with DNA gyrase, which represents the bacterial DNA topoisomerase II counterpart. The active site for ATP hydrolysis is encompassed in the N-terminal region, whereas that of DNA cleavage and ligation is located in the central region. The C-terminal part shows no sequence conservation and is dispensable for catalytic activity in vitro (9 -11).
The N-terminal region of topoisomerase II forms an ATPoperated clamp that closes upon ATP binding, allowing trapping of the T segment (6,8,12). The crystal structures of the yeast and bacterial N-terminal topoisomerase II fragments reveal that the dimeric N-terminal clamp contains two domains in each protomer (8,12). The most N-terminal domain holds the ATP-binding site that dimerizes upon nucleotide binding (8,12,13). This domain shares homology with the GHKL-type ATPases including topoVI-B (14) and MutL (15), for which a similar clamp operation has been observed. The second domain, called the transducer domain, bridges the N-terminal ATPase domain and the core region. Based on the yeast topoisomerase II structure, the transducer domain comprises the walls in a 6-Å-wide hole formed upon clamp closure, and it has been suggested to push the T segment through the DNA gate (12). The domain furthermore contains a loop that extends into the ATP-binding pocket and harbors a highly conserved lysine that contacts the ␥-phosphate of the bound nucleotide (8,12,13). Based on structural and biochemical analyses of the loop region, communication between the ATP-binding GHKL domain and the central domain of the enzyme responsible for DNA cleavage/ligation has been suggested to go through the transducer domain (13,16). Upon nucleotide binding and closure of the N-terminal clamp, the transducer domain undergoes a large domain rotation, probably facilitating opening of the DNA gate in the G segment bound by the central domain (14). Recent studies surprisingly revealed that the transducer domain also bears structural homology to MutL and the archea topoVI-B, and similar conformational effects of nucleotide binding have been reported for the transducer domain of these GHKL-ATPases, indicating that the structure and motion of this domain play a conserved role in the clamp mechanism (14).
According to the available structures of the yeast topoisomerase II core, it is a heart-shaped dimer comprised of two domains, AЈ and BЈ, showing homology to the subunits of DNA gyrase, gyrase A and gyrase B, respectively (17). The AЈ domain contains the active site tyrosine covalently attached to the 5Ј end of the DNA during cleavage (18) and encompasses a large cavity holding the transported DNA after passage through the G segment (17). The BЈ domain constitutes the interface be-tween the transducer domain and the AЈ domain (17). It is essential for DNA cleavage (17,19) and has a proposed role in the separation of the cleaved DNA ends moving the AЈ domain by undergoing large rearrangements in response to signals from the ATPase domain (5).
Minimization of energy usage in DNA topoisomerase II requires a tight coupling between ATP consumption and DNA strand passage (20). Several observations have indicated that this is obtained through an extensive interdomain and intersubunit communication facilitating an appropriate temporal order of catalytic steps and conformational changes. Thus, binding of ATP to the N-terminal domain stimulates cleavage of the G segment bound by the core region, and DNA binding increases the turnover rate of ATP hydrolysis (4,7). Furthermore, ATP binding occurs cooperatively only in the presence of a DNA substrate, meaning that DNA is required for intersubunit communication between the two ATP-binding sites in topoisomerase II (20). Binding of ATP per se induces clamp closure and triggers enzyme catalysis (6,21), whereas hydrolysis of ATP is believed to accelerate the DNA strand passage event as well as enzyme resetting (7,22).
In an earlier study, we have demonstrated that deletion of amino acids 351-408 located in the C-terminal part of the transducer domain of human topoisomerase II␣ disturbs interdomain communication (4). To further dissect the role of the transducer domain, we have characterized a human topoisomerase II␣ enzyme with two amino acids, leucine and glutamic acid, inserted at position 408. The insertion leaves both catalytic domains active, but strand passage is abolished. The clamp function is severely disrupted, even though the insertion is situated far from the dimerization domain, demonstrating the importance of the transducer domain for the clamp mechanism. The clamp-deficient enzyme is disturbed in domain communication, because the ATPase activity of the mutant enzyme lacks the normal response to DNA, and the DNA cleavage activity is unaffected by nucleotide binding. However, to our surprise, the DNA-mediated cooperativity in ATP binding seems to persist in 408i, indicating that stable clamp closure is not a prerequisite for intersubunit communication.

EXPERIMENTAL PROCEDURES
Yeast Strains and Plasmids-The Saccharomyces cerevisiae strains BJ201 (Mat␣ ura3 trp1 pep4::HIS3 prb1 can1 top2::TRP1) and JEL1⌬Top1 (kindly provided by J. C. Wang) were used for complementation and overexpression of topoisomerase II constructs, respectively. Plasmid pBY105 contains the yeast TPI promotor inserted into the polylinker region of the LEU2/ARS-CEN plasmid pRS315, which was used as the backbone for pHT212, pHT300, and pHT408i, carrying the wild-type human TOP2␣ cDNA with a C-terminal hexahistidine tail, the untagged human TOP2␣ cDNA, and the human TOP2␣ cDNA with an insertion of two amino acids (LE) at position 408, respectively. Modified versions of YEpWOB6 were used for overexpression of the hexahistidine-tagged human topoisomerase II␣ and 408i enzymes.
Construction of Plasmids-The construction of pHT300 and pHT408i (pHT407-2) was described by Jensen et al. (10). For overexpression of the hexahistidine version of the human topoisomerase II␣ wild-type and 408i enzymes, the topoisomerase II␣ cDNA of YEpWOB6 was first modified with a hexahistidine tail at the C terminus. For this purpose, a C-terminal fragment of topoisomerase II␣ fused to a hexahistidine tail was generated via polymerase chain reaction using pHT212 as template. The 3Ј primer was designed with a stretch of 54 overhanging nucleotides containing the hexahistidine tail and three restriction sites for cloning. The annealing segment of the primer was the reverse complementary sequence of the human c-Myc epitope (5Ј-TCC CCC CGG GGC GGC CGC CTC GAG CTA ATG ATG GTG GTG ATG GTG GCT CCC ACG GTT CAA GTC TTC TTC AGA GAT CAA C-3Ј). The 5Ј primer sequence was identical to nucleotides 2973-2991 of the human topoisomerase II␣ cDNA sequence (5Ј-GAG AGA GTT GGA CTA CAC-3Ј). The generated polymerase chain reaction fragment was used to replace the corresponding fragment of human topoisomerase II␣ in YEpWOB6 employing BlnI and XmaI as 5Ј and 3Ј cloning sites, respectively.
Yeast Transformation and Complementation-Yeast cells were transformed using a modified version of the LiAc method of Ito et al. (23). To test the ability of 408i to complement the lack of endogenous topoisomerase II in BJ201, the LEU2-based pHT408i construct was transformed into BJ201, and the cells were transferred to media plates containing 5Јfluoro-orotic acid (1 mg/ml) to select against the URA3 plasmid carrying the Schizosaccharomyces pombe top2 ϩ gene (10). pHT300 was used as a positive control.
Human Topoisomerase II␣ Induction, Overexpression, and Purification-The recombinant human topoisomerase II enzymes were overexpressed in yeast strain JEL1⌬Top1 by the addition of galactose to glucose-free medium (18). Yeast cells were extracted with 2 volumes of extraction buffer (50 mM Tris-HCl, pH 7.8, 1 M NaCl, 1 mM phenylmethylsulfonyl fluoride) and 1 volume of acid-washed glass beads (425-600 m; Sigma). Further preparation of yeast extracts was done according to the procedure of Jensen et al. (10). The initial purification step using a 6-ml Ni 2ϩ -nitriloacetic acid-agarose column was as described previously by Biersack et al. (24). For further purification, the fractions pooled from the Ni 2ϩ column were loaded onto a 1-ml heparin-Sepharose column (Amersham Biosciences), and elution was performed by a 15-ml linear NaCl gradient ranging from 400 to 1500 mM. The peak fractions were further applied to a phosphocellulose column (P11 cellulose phosphate; Whatman) for concentration of the enzyme. Elution was performed in 750 mM NaCl, 0.5 mM dithiothreitol, 0.1 mM EDTA, 40% glycerol, 50 mM Tris-HCl, pH 7.7. Fractions containing topoisomerase II enzyme were pooled and stored in liquid nitrogen for later use. Homogeneity of the topoisomerase II preparations was determined by SDS-polyacrylamide gel analysis after staining with Coomassie Blue dye.
Topoisomerase II-mediated DNA Relaxation-DNA relaxation was performed by incubating 2 or 10 nM topoisomerase II and 6.5 nM supercoiled pUC19 DNA in 50 mM Tris-HCl, pH 8, 140 mM KCl, 1 mM EDTA, 8 mM MgCl 2 supplemented with 1 mM ATP. The reactions were incubated at 37°C and stopped after 20 min by the addition of SDS and EDTA to a final concentration of 0.1% and 10 mM, respectively. The samples were subjected to electrophoresis in 1% agarose gels in TBE buffer (100 mM Tris borate, pH 8.3, 2 mM EDTA). DNA was stained with 1 g/ml ethidium bromide and visualized by UV light.
Clamp Closing Assay-For clamp closing experiments, 10 nM topoisomerase II was preincubated with 6.5 nM negatively supercoiled pUC19 DNA for 5 min in a total volume of 20 l of 50 mM Tris-HCl, pH 8, 140 mM KCl, 1 mM EDTA, 8 mM MgCl 2 . After preincubation, AMPPNP 1 (Roche Applied Science) was added to a final concentration of 1 mM, and the reactions were incubated for an additional 5 min. The reactions were next stopped by the addition of either NaCl or SDS to final concentrations of 800 mM and 1%, respectively. The sample volume was increased to 70 l by adding 50 l of an 800 mM NaCl solution. To trap enzyme-DNA catenanes, phenol extraction was performed by adding 1 volume of phenol. The samples were vortexed and centrifuged at 13,000 rpm in an Eppendorf centrifuge for 10 min. The water phase was removed, ethanol-precipitated, and dissolved in 10 l of TE buffer (10 mM Tris-HCl, pH 7.4, 1 mM EDTA) for gel analysis. The combined phenol phase and phenol/water interphase was washed one time in high salt washing buffer (2 M NaCl, 50 mM Tris-HCl, pH 7.9, and 5 mM MgCl 2 ) and one time in 0.6 M LiCl. Upon removal of the water phase after the last wash, the remaining material was ethanol-precipitated and dissolved in 10 l of TE buffer containing 1 mg/ml proteinase K. The samples were next subjected to electrophoresis in a 1% agarose gel in TBE buffer containing 1 g/ml ethidium bromide.
Clamp Closing Analyzed by Analytical CsCl Density Gradient Ultracentrifugation-The technique used is essentially that described previously by Morris et al. (25). Reaction mixtures of 30 l containing 150 nM topoisomerase II, 250 nM relaxed pBR322 (topoGEN), or linearized pUC19 plasmid DNA in sample buffer (50 mM Tris-HCl, pH 8, 60 mM NaCl, 1 mM EDTA, 8 mM MgCl 2 ) were incubated for 5 min at 30°C. AMPPNP was added to a final concentration of 1 mM, and incubation was continued for 10 min. were incubated for 10 min at 37°C and loaded onto a nitrocellulose filter using a slot blot apparatus (PR 648 Slot Blot Manifold; Amersham Biosciences). The filter was washed in a buffer containing 50 mM Tris-HCl, pH 7.9, and 800 mM KCl. Loading of samples were performed at a rate of 1 ml/min. A Molecular Imager (Bio-Rad) was used for quantification. For measure of protein retention, a similar experiment was performed in the absence of [␥-35 S]ATP, but in this case the filter was subjected to Western blot analysis with the pentahistidine antibody (Qiagen) recognizing the His-tagged topoisomerase II enzyme.
Topoisomerase II-mediated Hydrolysis of ATP-The ATPase assay was based on the method of Osheroff et al. (26). The reactions contained 100 nM topoisomerase II in the presence or absence of 20 nM negatively supercoiled pUC19 plasmid DNA. The reactions were carried out in 20 l of 50 mM Tris-HCl, pH 8, 140 mM KCl, 1 mM EDTA, 8 mM MgCl 2 containing a final concentration of 1 mM cold ATP and 2 Ci (0.033 M) of [␥-32 P]ATP (3000 Ci/mmol; Amersham Biosciences). The mixtures were incubated at 37°C, and 2.5-l aliquots were spotted onto thin layer cellulose plates impregnated with polyethylenimine (Merck). Chromatography was performed using freshly made 0.4 M NH 4 HCO 3 . Levels of free phosphate were quantified using a Molecular Imager (Bio-Rad).
Oligonucleotides-DNA oligonucleotides were obtained from DNA Technology Corp. and purified by preparative polyacrylamide gel electrophoresis as described by Andersen et al. (27). The 28-mer used as the bottom strand in the suicide substrate was modified at the 3Ј end by a phosphate group to inhibit ligation to this end.
Topoisomerase II-mediated Cleavage of Suicide Substrates-Hybridization and labeling of the synthetic oligonucleotides were done according to the procedures described by Andersen et al. (27). For topoisomerase II-mediated DNA cleavage, 20 nM topoisomerase II was incubated with 0.05 pmol of labeled substrate in 50 l of 10 mM Tris-HCl, pH 7.0, 2.5 mM MgCl 2 , 2.5 mM CaCl 2, 20 mM NaCl, and 0.1 mM EDTA at 37°C, and the reactions were stopped by the addition of SDS to 1%. Covalent topoisomerase II-DNA cleavage complexes were recovered from a phenol/water interphase according to Gocke et al. (28). The complexes were subsequently ethanol-precipitated and treated with proteinase K (500 g/ml) for 3 h at 42°C. One volume of gel loading buffer (50% formamide, 0.05% bromphenol blue, 0.03% xylene cyanole, 5 mM EDTA) was added, and the material was subjected to electrophoresis in a 12% denaturing polyacrylamide gel.
Topoisomerase II-mediated Ligation-A topoisomerase II-mediated suicide cleavage reaction was performed as described above. After incubation at 37°C for 45 min, the cleavage reaction was stopped by the addition of NaCl to 0.4 M, thereby preventing further cleavage during the ligation reaction. Ligation was initiated by the addition of 50 pmol of 45-mer ligation substrate. After 15 min of further incubation, the reaction was stopped by addition of SDS to 1%. The samples were ethanol-precipitated, proteinase K-digested, and analyzed by electrophoresis in a 12% denaturing polyacrylamide gel.
Topoisomerase II-mediated Cleavage of Circular DNA-Topoisomerase II-mediated DNA cleavage was performed by incubating 100 nM of topoisomerase II and 6.5 nM negatively supercoiled pUC19 DNA in a total volume of 20 l in 50 mM Tris-HCl, pH 8, 140 mM KCl, 1 mM EDTA, 8 mM MgCl 2 . The samples were incubated at 37°C for 7 min, and the cleavage products were trapped by the addition of SDS to 1%. After proteinase K digestion (0.8 mg/ml), the samples were subjected to electrophoresis in 1% agarose gels in TAE buffer (40 nM Tris acetate, pH 8.3, 2 mM EDTA). DNA was visualized by ethidium bromide staining (1 g/ml). When reactions were carried out in the presence of ATP or AMPPNP, the concentration of the nucleotide was 1 mM.

Human Topoisomerase II␣ Containing a Two-amino Acid Insertion at Position 408 Is Unable to Complement Growth in a
Yeast top2 Deletion Strain-To study the role of the transducer domain in DNA topoisomerase II, we have characterized a human topoisomerase II␣ enzyme (408i) having an insertion of two amino acids at position 408 in the transducer domain (Fig.  1A, left panel). Based on alignment to the homologous part of yeast topoisomerase II, the insertion is located in the lower The ATP-binding subdomain is shown in white, the core region is in dark gray, and the variable C-terminal region is in black. The transducer domain, which is shown in light gray, bridges the two catalytic entities and is thought to encompass the walls of the clamp potentially interacting with the T segment. Y indicates the position of the active site tyrosine. Right panel, the structure of the N-terminal fragment of yeast topoisomerase II (12) showing the location of position 397 (spheres, indicated with arrows) corresponding to position 408 in human topoisomerase II␣. The figure is prepared with PYMOL (34). B, a single-copy ARS/CEN plasmid carrying either the 408i (pHT408i) or the wild-type human topoisomerase II␣ (pHT300) cDNA behind a TPI promotor was transformed into the yeast strain BJ201. As a control, the cells were transformed with the empty vector pRS315. In BJ201, the chromosomal TOP2 gene has been disrupted by insertion of the structural TRP1 gene, whereas the essential topoisomerase II activity is provided by the S. pombe top2 ϩ gene carried on a single copy URA3-based plasmid. After transformation, the cells were grown on medium containing 5Ј-fluoro-orotic acid to counterselect against the URA3 plasmid. C, purification of wild-type and 408i topoisomerase II␣. The enzymes were purified after overexpression in yeast through three different steps involving Ni 2ϩ -nitriloacetic acid-agarose, heparin-Sepharose, and phosphocellulose column chromatography. The homogeneity of the enzyme preparations was determined in a 10% SDS-polyacrylamide gel stained with Coomassie Blue. Protein size markers are indicated to the left of the gel. Bovine serum albumin (BSA) has been loaded as a marker for protein concentration. wt, wild-type. part of the transducer domain, as indicated in the structure of the N-terminal yeast fragment (Fig. 1A, right panel). The insertion has been presented earlier in a linker scanning analysis of human topoisomerase II␣ (10), where it was found to be among the few that abolished the ability of the human enzyme to sustain growth in an S. cerevisiae top2 deletion strain. This conclusion was confirmed by the complementation analysis shown in Fig. 1B. For studies of the in vitro capabilities of 408i, the enzyme fused to a hexahistidine tail at the C-terminal end was overexpressed in a top1 null strain and purified to homogeneity as seen from the Coomassie-stained gel in Fig. 1C.
Purified 408i Is Unable to Perform Relaxation of Supercoiled Plasmid DNA-To investigate whether the inability of 408i to complement is due to a defect in the DNA strand passage activity of the enzyme, a DNA relaxation assay was performed, and the catalytic activity of the mutant enzyme was compared with that of the wild-type enzyme. Although 10 nM of the wild-type enzyme relaxed all the supercoiled DNA within 20 min, the insertion mutant showed no sign of relaxation in these molar ranges (Fig. 2). Thus, consistent with the lack of in vivo complementation, these results reveal that insertion of two amino acids at position 408 is detrimental to the overall enzyme activity. Either the mutant enzyme fails to perform one or more of the steps involved in strand passage and catalytic turnover, or alternatively, the two-amino acid insertion disrupts correct folding of the protein.
Insertion at Position 408 Severely Disturbs the Function of the N-terminal Clamp-Preceding strand passage, topoisomerase II binds ATP and changes into the closed clamp conformation. To test whether the enzyme, bearing an insertion in the N-terminal transducer domain, has retained this important function of the N terminus, a clamp closing assay was performed (modified from Bjergbaek et al. (4)). The enzyme was preincubated with supercoiled plasmid DNA before the addition of the nonhydrolyzable ATP analog, AMPPNP. With the wild-type enzyme this leads to the formation of salt stable topoisomerase II-DNA interlinked complexes, which can be collected from a phenol-water interphase (Fig. 3A). The mutant enzyme, however, was unable to trap a detectable amount of supercoiled DNA under the present experimental conditions. Thus, insertion of two amino acids at position 408 disturbs stable clamp closure. None of the enzymes allowed trapping of DNA upon the addition of SDS, which disrupts the interlink between enzyme and DNA or in the absence of AMPPNP, where stable clamp closing does not occur.
To perform a more detailed investigation of the clamp function of 408i, the enzyme was subjected to analytical CsCl density gradient ultracentrifugation. In this assay, the clamp closing reaction can be studied in the presence of higher concentrations of DNA and enzyme, and it is thus possible to

FIG. 3. Insertion at position 408 disturbs the N-terminal clamp.
A, to investigate clamp closure, enzyme and DNA were preincubated for 5 min before the addition of AMPPNP. After 5 min of further incubation, the reactions were stopped by the addition of either NaCl or SDS to 800 mM and 1%, respectively, and the samples were phenol-extracted. The water phase was removed and ethanol-precipitated, whereas the remaining phenol interphase was washed in high salt before the material was ethanol-precipitated and proteinase K-digested. The samples were subjected to electrophoresis in a 1% agarose gel containing ethidium bromide. Enzyme, nucleotide, and stop solution used in the samples are indicated above the gel. i and w denote interphase and water phase, respectively. The position of supercoiled (SC) and catenated (Cat) circular plasmid DNA are indicated. B, For analysis of clamp closure by analytical ultracentrifugation, enzyme, DNA and AMPPNP were incubated as described under "Experimental Procedures." The samples were loaded on a CsCl gradient and run for 40 h. Absorbance traces at 260 nm of the CsCl density gradients are shown for each reaction. The major peak at the bottom of the gradient corresponds to DNA without any protein bound, whereas enzyme bound to one plasmid is represented by the peak in the upper part of the gradient. Enzyme bound to catenated plasmids sediments as a minor peak in the middle of the gradient. The small peak observed in the control experiment with linear DNA is due to a contaminating fraction of nonlinearized plasmid DNA in the linear DNA preparation. wt, wild-type. detect even very low levels of clamp closure. Although this assay revealed the formation of a small amount of the salt stable closed clamp complex with 408i in the presence of AMP-PNP (Fig. 3B), it still confirmed that the mutant enzyme is severely disrupted in its ability to close the clamp and/or to keep the clamp stably closed at high concentrations of salt.
408i Retains Normal ATP Binding Capabilities-To investigate whether the clamp closing defect of 408i is caused by an impairment of ATP binding, we tested the enzyme for its ability to bind a nonhydrolyzable, radioactive ATP analog using a filter binding assay (Fig. 4, upper panel). The results are depicted graphically in the lower panel of Fig. 4, where levels of ATP binding are given in arbitrary units. In the absence of DNA, the ATP binding capacity of 408i was similar to that of the wild-type enzyme (compare the third and sixth columns), showing that insertion at position 408 does not disrupt folding of the ATP-binding pocket. Thus, the inability of 408i to perform clamp closure does not result from a lack of ATP binding.
ATP binding to the wild-type enzyme occurs cooperatively in the presence of DNA (7). Because cooperativity depends on intersubunit communication, we wanted to investigate whether this effect of DNA was maintained in the clampdeficient 408i enzyme. Cooperative ATP binding can be revealed experimentally if relative small concentrations of an unlabeled ATP analog stimulate binding of the labeled ATP analog rather than compete with it. As seen in Fig. 4, the addition of 20 M cold AMPPNP surprisingly increases binding of [␥-35 S]ATP to 408i in the presence of DNA, as is the case with the wild-type enzyme. Under the given experimental condi-tions, a 2-3-fold stimulation is obtained with the wild-type enzyme, whereas a 2-fold stimulation is obtained with 408i (Fig. 4, lower panel, compare fourth and fifth columns and  seventh and eighth columns, respectively). At higher concentrations of cold AMPPNP, binding of [␥-35 S]ATP is competed by AMPPNP (data not shown). Thus, 408i seems to have retained its ability to sense the presence of DNA, responding with cooperative ATP binding despite its lack of clamp closing activity. Western blot analysis was performed to rule out that the observed differences in enzyme binding of [␥-35 S]ATP were caused by an AMPPNP-or DNA-mediated change in the retention efficiency of topoisomerase II on the nitrocellulose filter (data not shown).
408i Has ATPase Activity, but This Activity Is Not Stimulated by DNA-To investigate whether the N-terminal ATPase domain has retained its intrinsic catalytic ability, we tested the ATPase activity of 408i. ATPase experiments were performed in the presence of 1 mM ATP, and thin layer chromatography followed by Molecular Imager analysis was used to measure the level of released P i (Fig. 5). The result shows that 408i is able to hydrolyze ATP, although its ATPase activity is lower than that of the wild-type enzyme (Fig. 5A). A comparison of 408i to that of an ATPase-deficient topoisomerase II␣ enzyme (G164I) (7), which was purified in the same way, confirmed that the ATPase activity of 408i is not caused by contaminating phosphatases. ATPase activity is therefore an intrinsic capability of 408i, demonstrating that the N-terminal domain of the mutant enzyme folds into a catalytically active entity.
Earlier studies on topoisomerase II ATPase activity have shown that the catalytic turnover of hydrolysis is stimulated by DNA (25,26). To test whether an inability to undergo the conformational changes necessary for stable clamp closure disturbs this stimulatory effect of DNA, the ATPase activity was measured also in the presence of DNA (Fig. 5B). Under the conditions used in the present study, the wild-type ATPase activity displayed a 2-fold stimulation in the presence of DNA. However, no stimulation was obtained with the mutant enzyme, implying that at least under saturating concentrations of ATP, the ATPase activity is unaffected by DNA in 408i.
The 408i Enzyme Withholds both DNA Cleavage and Ligation Abilities-To determine whether the insertion in 408i influences the activities of the central domain, we analyzed the ability of the enzyme to perform DNA cleavage and ligation. For this purpose, advantage was taken of the topoisomerase II suicide system, which allows a separation of the two halfreactions as schematically illustrated in Fig. 6A (upper panel). The suicide substrate consists of a 16-base-long 5Ј-recessed top strand with only three nucleotides 5Ј to the cleavage position and a 28-base-long bottom strand. Use of the suicide substrate has been demonstrated to cause an uncoupling of the DNA cleavage and ligation half-reactions because of the release of the trinucleotide 5Ј to the cleavage position on the top strand upon topoisomerase II-mediated cleavage (27). The substrate was labeled at the 3Ј end of the recessed top strand and incubated with either the wild-type or the 408i enzyme. Cleavage complexes were isolated from a phenol/water interphase and analyzed by gel electrophoresis. As seen from the gel in Fig. 6A  (lower panel), the mutant enzyme retains the ability to cleave the suicide DNA substrate with a cleavage level similar to that obtained with the wild-type enzyme.
The topoisomerase II cleavage complex generated upon cleavage of a suicide substrate is kinetically competent (27) and able to perform ligation if a suitable ligation substrate is added to the cleavage mixture as schematically illustrated in Fig. 6B  (left panel). To investigate whether the central domain of 408i also withholds ligation activity, topoisomerase II-DNA cleav- age complexes were prepared for the ligation assay as described under "Experimental Procedures." After an increase in salt concentration to inhibit further cleavage, ligation was initiated by addition of a 45-mer DNA oligonucleotide able to hybridize to the bottom strand of the cleaved substrate. As for cleavage, the level of ligation obtained with the mutant enzyme is very similar to that of the wild-type enzyme (Fig. 6B,  right panel).
To further verify that the cleavage/ligation domain of 408i retains a normal DNA interaction, the sequence specificity of the mutant and wild-type enzymes were compared using a 4330-bp pBR322 fragment labeled only at one 3Ј end as a DNA cleavage substrate. No difference was observed in the specificity of the two enzymes (data not shown), indicating that insertion at position 408 does not interfere with the normal DNA sequence preference of topoisomerase II. The result further substantiates that the central domain of 408i has retained its normal properties.
The Cleavage/Ligation Equilibrium of 408i Is Not Affected by ATP or AMPPNP-In addition to the stimulatory effect of DNA on topoisomerase II-catalyzed ATP binding and hydrolysis, the communication between the N-terminal and central regions is also manifested during topoisomerase II-mediated DNA cleavage, where ATP or an ATP analog shifts the DNA cleavage/ ligation equilibrium toward cleavage (21,29).
To further probe the communication abilities of the mutant enzyme, topoisomerase II-mediated cleavage of supercoiled plasmid DNA was performed in the absence or presence of ATP or the ATP analog, AMPPNP (Fig. 7, upper panel). Although more cleavage was obtained with the mutant enzyme relative to the wild-type enzyme in the absence of nucleotide (Fig. 7, lower panel), only the wild-type enzyme was stimulated by AMPPNP. This shows that 408i is unable to transmit the effect of nucleotide binding to the core region and thus indicates that the insertion directly or through inhibition of clamp closure abolishes the transmission event. In the presence of ATP, the wild-type enzyme relaxed the DNA. Even with these high enzyme concentrations, ATP only allowed 408i to cleave the DNA, in support of its inability to relax supercoiled DNA.

DISCUSSION
The ability of topoisomerase II to change the topological conformation of DNA is based on a highly controlled series of conformational changes that propagate from the N-terminal region throughout the entire enzyme. One large and very important switch occurs upon binding of ATP, which causes the N-terminal arms to dimerize and thereby closes the clamp. Because binding of a nonhydrolyzable ATP analog has been demonstrated to sustain one strand passage reaction (26,30), ATP binding per se is sufficient to allow strand passage, whereas hydrolysis accelerates the reaction and resets the enzyme for a new catalytic cycle (22,31).
In the present work, we have characterized a mutant human topoisomerase II␣ enzyme that has a two-amino acid insertion at position 408 in the distal C-terminal part of the transducer domain. The enzyme has a severe defect in clamp closure but nevertheless can bind and hydrolyze ATP and is able to perform DNA cleavage and ligation. Our results show that the insertion causes a quite specific disturbance of the N-terminal clamp function. It results in an enzyme that fails to relax supercoiled DNA and therefore is unable to sustain mitotic growth in a yeast top2 null strain.
Taken together, our results strongly suggest that closure of the N-terminal clamp is the most important consequence of ATP binding and critical for the strand passage reaction. This is in agreement with results obtained from studies of a human topoisomerase II␣ heterodimer, which was still able to relax DNA, although disabled in ATP binding in one subunit (7). The enzyme was disturbed in some aspects of domain communication but nevertheless was able to perform clamp closure, supportive of a strong correlation between clamp closure and DNA strand passage.
Surprisingly, we find that 408i, which is unable to make a stable N-terminal dimerization, still seems to respond to DNA with cooperative ATP binding. This indicates that certain types of intersubunit communication can be mediated either via dimer contacts in the core region or through less stable interactions in the N-terminal dimerization domains, which might still occur in 408i, when DNA is bound to the enzyme.
The observations that DNA does not stimulate ATP hydrolysis in 408i but still seems to mediate cooperative ATP binding suggest that DNA normally influences the ATPase domain through two separate mechanisms, only one of which is retained in 408i. The stimulatory role of DNA on ATP hydrolysis, which is lacking in 408i, might thus rely on a fully coordinated clamp function and/or DNA strand passage. In support of two stimulatory mechanisms of DNA, studies of a cleavage-deficient topoisomerase II showed that this enzyme had a normal indicates radioactive labeling of the substrate. Lower panel, for suicidal cleavage, the wild-type enzyme or 408i was incubated with the 3Ј end-labeled substrate, and cleavage was stopped by the addition of SDS to 1%. Topoisomerase II-DNA cleavage complexes were recovered from a phenol/water interphase. The complexes were subsequently ethanolprecipitated and treated with proteinase K, before they were subjected to electrophoresis in a 12% polyacrylamide gel. C, DNA control. M, DNA size marker increasing in steps of two bases. The substrate remaining in the interphase after phenol extraction and the cleavage product, for which migration is retarded with ϳ1 base because of residual undigested protein, are indicated by S and Cl, respectively. *, cleavage products with a larger protein fragment covalently linked because of incomplete proteinase K digestion. The experiments were carried out in duplicate. B, left panel, schematic illustration of the topoisomerase II-mediated ligation reaction showing ligation of a 45-mer to the cleaved DNA in the suicide cleavage complex. The positions of the topoisomerase II-mediated DNA cleavage are indicated by arrows. The asterisk represents radioactive labeling. B, right panel, topoisomerase II-mediated ligation was performed as described under "Experimental Procedures" in the absence or presence (ϩ) of the 45-mer ligation substrate. The samples were withdrawn after 15 min of ligation, and the products were analyzed in a 12% denaturing polyacrylamide gel and visualized on a Molecular Imager (Bio-Rad). The enzyme used in the reaction is indicated above the lanes. C, DNA control. M1 and M2 indicate DNA size markers increasing in steps of two or 10 bases, respectively. Cl, cleavage product. S, cleavage substrate. L, ligation product. wt, wild-type. effect of DNA on ATP binding but a very reduced effect on ATP turnover (25).
In addition to the inability of 408i to increase its rate of ATP hydrolysis upon the addition of DNA, the clamp-disabled 408i enzyme does not show the normal increase in DNA cleavage when either ATP or an ATP analog is present. 408i thus has a disruption in the two-way communication, which normally exists between the N-terminal and central enzyme regions. Our findings correlate with the results obtained from studies of a topoisomerase II␣ enzyme bearing a deletion of amino acids 351-408 (4). In addition to severe interdomain communication defects, the deletion enzyme had a malfunctioning N-terminal clamp. Our results confirm that the conformational changes involved in clamp closure are important for interdomain communication.
Sequence alignment with the recently crystallized yeast topoisomerase II fragment suggests that position 408 is situated at the outer surface in the distal end of the transducer domain and probably is in close contact with the BЈ domain. It is thus located far from the ATP-binding site and the N-terminal dimerization area, both of which have well established roles in the mechanism of clamp closure (8,12). No obvious role for the transducer domain in the clamp mechanism has been suggested, yet comparison of several GHKL-ATPases surprisingly revealed that the structural similarity was not confined to the ATPase domain, but in some cases included the transducer domain as well (14,15). The observed disturbance of clamp function by insertion of two amino acids at position 408 demonstrates an important role of the transducer domain for the dynamics of the N-terminal clamp consistent with the structural conservation of the domain.
Interdomain communication is critical for topoisomerase II to go through the conformational changes required for catalysis including clamp closure (4,5,8,30). This complex process involves global rearrangements, and therefore it is difficult to identify residues critical for the process. However, the transducer domain harbors an important loop, in which a highly conserved lysine (Lys 378 ) forms a hydrogen bond with the ␥-phosphate of the bound nucleotide (8). Both structural and biochemical studies suggest that this loop is sensing the nucleotide-bound state of the enzyme and helps direct the motions of the ATP-binding domain relative to the transducer domain (13,14,16). According to the alignment, position 408 in human topoisomerase II␣ is connected to this loop by a stretch of amino acids that directly traverses the transducer domain. The fact that position 408 is located at the interface of the BЈ domain and the transducer domain thus indicates the existence of a structural communication pathway reaching from the ATPase domain to the DNA cleavage/ligation domain through the area around position 408. The observed clamp deficiency resulting from insertion at position 408 thus suggests that disturbance of communication between the transducer domain and the BЈ domain is fatal for clamp function, maybe because of restrictions in the movements of the N-terminal arms that still remain in 408i after nucleotide binding.
The fact that 408i displays a significantly higher cleavage level on supercoiled plasmid DNA in the absence of nucleotide infers that the cleavage region in 408i does not fall into quite the same conformation as in the wild-type enzyme under these conditions. This observation leads to the hypothesis that the transducer domain in the absence of ATP normally induces a specific conformation of the BЈ domain through the area around position 408, which results in an autoinhibition of the cleavage activity. Upon nucleotide binding, the inhibition is released, allowing a DNA gate to be formed. This interpretation is consistent with the crystal structure of the yeast core topoisomerase II that spontaneously falls into an open conformation in the absence of DNA and nucleotide (5,32). In line with this, hypercleavage has also been reported for a catalytically active human topoisomerase II␣ enzyme bearing a point mutation at position 437 (33), also located at the interface of the BЈ domain and the transducer domain.
In conclusion, the insertion of two amino acids at position 408 in the transducer domain of human topoisomerase II␣ severely disturbs the function of the N-terminal clamp. The clamp-deficient enzyme is catalytically inactive and lacks most aspects of interdomain communication. The results show that even distal parts of the transducer domain are important for the dynamics of the N-terminal clamp. Furthermore, the study strongly indicates that closure of the N-terminal clamp is a crucial prerequisite for strand passage.