Reconstitution of enzymatic activity by the association of the cap and catalytic domains of human topoisomerase I.

When human topoisomerase I binds DNA, two opposing lobes in the enzyme, the cap region (amino acid, residues 175-433) and the catalytic domain (Deltacap, residues 433 to the COOH terminus) clamp tightly around the DNA helix to form the precleavage complex. Although Deltacap contains all of the residues known to be important for catalysis and binds DNA with an affinity similar to that of the intact enzyme, this fragment lacks catalytic activity. However, a mixture of Deltacap and topo31 (residues 175-433) reconstitutes enzymatic activity as measured by plasmid DNA relaxation and suicide cleavage assays. Although the formation of an active complex between topo31 and Deltacap is too unstable to be detected by pull-down experiments even in the presence of DNA, the association of topo31 with Deltacap persists and is detectable after the complex catalyzes the covalent attachment of the DNA to Deltacap by suicide cleavage. Removal of topo31 from Deltacap-DNA after suicide cleavage reveals that, unlike the cleavage reaction, religation does not require the cap region of the protein. These results suggest that activation of the catalytic domain of the enzyme for cleavage requires both DNA binding and the presence of the cap region of the protein.

Eukaryotic type I topoisomerases promote the relaxation of supercoiled DNA by nicking and rejoining one of the strands of the DNA. These enzymes are important for many biological processes including DNA replication, transcription, and recombination (1,2). Eukaryotic topoisomerase I, the poxviral topoisomerases, and some bacterial topoisomerases belong to the type IB subfamily of topoisomerases (1,3). The type IB subfamily members bind to double-stranded DNA and can relax either positive or negative supercoils in the absence of energy cofactors or divalent cations. DNA cleavage is initiated by the nucleophilic attack of the O-4 atom of the active site tyrosine on the scissile phosphate with the resultant covalent attachment of the enzyme to the 3Ј end of the broken strand. Rotation of the DNA at the site of the break relaxes any supercoiling tension followed by religation of the DNA and release of the enzyme (2,4).
Human topoisomerase I is a member of the type IB subfamily and is composed of 765 amino acids (91 kDa). Sequence comparisons and limited proteolytic analyses in combination with crystallographic studies of the enzyme define four major do-mains: an NH 2 -terminal domain (Met 1 -Gly 214 ), a core domain (Ile 215 -Ala 635 ), a linker domain (Pro 636 -Lys 712 ), and a COOHterminal domain (Gln 713 -Phe 765 ) (Fig. 1A) (5)(6)(7). The highly charged NH 2 -terminal domain is dispensable for enzymatic activity in vitro (5) and contains nuclear targeting signals and binding sites for other proteins, such as nucleolin and SV40 large T antigen (8 -10). Topo70 is a truncated form of human topoisomerase I with a molecular mass of 70 kDa that lacks residues 1-174 of the NH 2 -terminal domain but retains full enzyme activity in vitro (11,12). The core domain is highly conserved and contains all of the residues directly implicated in catalysis except the active site Tyr 723 . The COOH-terminal domain is also highly conserved and contains the active site Tyr 723 . Separately purified COOH-terminal and core domains can interact with each other and reconstitute topoisomerase I activity in vitro (5,13). The linker region that connects the core domain to the COOH-terminal domain is not conserved and is dispensable for activity in vitro, although without the linker the enzyme has a reduced processivity (12).
In the crystal structure of the human topoisomerase I complexed with DNA, the protein clamps around the DNA with most of the protein-DNA contacts involving the core and COOH-terminal domains (6,14). The core domain can by further divided into three subdomains: 1) core subdomain I (residues 215-232, 320 -433), 2) core subdomain II (residues 233-319), and 3) core subdomain III (residues 434 -635). Core subdomains I and II form the top lobe or "cap" of the enzyme and cover the top of the DNA as the structure is usually oriented (6,14). Core subdomain II does not come in contact with the DNA in the structure, but the folding of core subdomain II is similar to that of the homeodomain region found in a family of DNA-binding proteins (6,14). Two long positively charged ␣ helices (␣ 6 from core subdomain I and ␣ 5 from core subdomain II) form a "V"-shaped structure on the front end of the cap that may contact the DNA during the rotation process (6,14). Core subdomain III and the COOH-terminal domain form the bottom lobe of the protein. This region of the protein is homologous to the catalytic domain of the site-specific recombinases that include HP1 integrase, integrase, and Cre recombinase, and also to the catalytic domain of vaccinia topoisomerase (6,15). The bottom lobe of the protein is attached to the cap region through a long ␣ helix extending upwards from core subdomain III on one side of the bound DNA and by a salt bridge on the other side of the DNA formed by amino acid side chains extending from a pair of loops in core subdomains I and III (6,14).
To dissect the function of the core domain and explore the functional relationship between the cap and the core subdomain III, we have studied the properties of three fragments of human topoisomerase I either alone or in pairwise combinations (Fig. 1A). Topo31 (residues 175-433), a 31-kDa fragment that consists of the cap region of the protein, binds DNA well, whereas topo17 (residues 175-320), a 17-kDa fragment that mainly comprises core subdomain II (the homeodomain-like region), cannot bind DNA. The third fragment corresponds to the catalytic domain of the enzyme and contains core subdomain III, the linker region, and the COOH-terminal domain (residues 433-765), and the fragment is referred to as ⌬cap. Although ⌬cap contains all of the elements required for catalysis and can bind DNA, it is catalytically inactive. However, when topo31 and ⌬cap are combined, they reconstitute enzymatic activity.

EXPERIMENTAL PROCEDURES
Generation of Truncation Mutants-All of the truncation mutants were generated using standard PCR methodology. To generate topo17, pGEX-topo70 DNA (13) was used as the PCR template, in combination with a plus-sense primer that starts at position 852 of the topoisomerase I cDNA sequence (16) and a minus-sense primer that contained two stop codons followed by an AvrII restriction site immediately after the codon for residue 320 of the protein. The PCR products were purified and digested with NdeI and AvrII and ligated to pGEX-topo70-WT DNA that had been cleaved with the same two restriction enzymes. To generate topo31, we used the same plus-sense primer and a minus primer that contains two stop codons followed by an AvrII restriction site after the codon for residue 433 of the protein. The PCR fragment was purified and digested with SphI and AvrII and ligated to the pGEX-topo70-WT that had been cut with the same two restriction enzymes. The ⌬cap mutant was generated using a plus-sense primer containing a BamH1 site followed by an AUG codon that annealed upstream of residue 433 and a minus-sense primer that annealed downstream of position 2392 of the topoisomerase I cDNA sequence. The PCR fragment was purified and digested with BamH1 and NheI and ligated to pFASTBAC1 topo70 (17) that had been cut with the same two restriction enzymes. All of the mutations were confirmed by dideoxy sequencing.
Suicide Cleavage Reactions-The duplex oligonucleotide suicide cleavage substrate CL14/CP25 was labeled and annealed as described previously (18). The suicide cleavage reactions were carried out by incubating 2 g of topo17, topo31, or ⌬cap alone or in the indicated combinations in 20 l of reaction buffer with 5 ng of the suicide cleavage substrate at 23°C for 3 h. Topo70 (0.5 g) was used as a positive control 1 The abbreviations used are: GST, glutathione S-transferase; topo70, NH 2 -terminal truncation of human topoisomerase I missing first 174 amino acids; topo58, COOH-terminal truncation of topo70 missing last 106 amino acids; topo17, human topoisomerase I from residues 175-320; topo31, human topoisomerase I from residues 175-433; ⌬cap, NH 2 -terminal truncation of human topoisomerase I beginning at residue 433; DTT, dithiothreitol. for suicide cleavage. 5 l of 5ϫ SDS loading buffer (5% SDS, 20% glycerol, 100 mM Tris-HCl, pH 8.0, 5% 2-mercaptoethanol, 0.12% bromphenol blue) was added to quench the reactions. The samples were boiled for 5 min and analyzed by 10% SDS-PAGE. The gel was stained with Coomassie Blue to visualize the protein bands and dried before exposure to film to detect the radiolabeled proteins.
Religation Kinetics-Covalent complexes used as substrates for the religation reaction were generated by suicide cleavage as described previously (18). Suicide cleavage was carried out in 100 l of reaction buffer containing 20 nM labeled suicide substrate and 0.5 M topo70 for 1 h or 2.5 M topo31 plus 2.5 M ⌬cap for 6 h at 23°C. The suicide cleavage reaction with topo70 was stopped by adding KCl to a final concentration of 0.5 M to prevent further cleavage during religation. High salt inactivation of the reaction with the reconstituted enzyme was unnecessary owing to the slow cleavage rate, and furthermore, high salt was found to dissociate the topo31-⌬cap complex (see Fig. 7). The reactions were transferred to 37°C and preincubated for 2 min. Religation was initiated by the addition of a 300-fold molar excess of the 11-mer religation acceptor oligonucleotide (R11) that is complementary to the region downstream of the cleavage site (18). Aliquots of 10 l were removed at different time points (5 s, 15 s, 30 s, 2 min, 5 min, and 60 min), and the reactions were stopped by the addition of an equal volume of 1% SDS. Religation was complete by the 60-min time point in both cases. Samples were ethanol-precipitated and dissolved in 10 l of 1 mg/ml trypsin and digested at 37°C for 1 h to remove all but a short topoisomerase-derived peptide from the covalent complexes. The samples were analyzed by electrophoresis in a 20% sequencing gel. The religation product migrates as a 23-mer and is well resolved from the oligonucleotide-peptide covalent complex that migrates slower than the uncleaved oligonucleotide. The percentage of religation at each time point was quantified using a phosphorimager and the ImageQuant software.
Gel Shift Assay-The labeling and annealing of the 25-mer duplex oligonucleotide, CL25/CP25, has been described previously (18). The DNA binding assay was carried out by incubating the labeled CL25/ CP25 DNA (0.5 nM) with aliquots of 2-fold serial dilutions of the indicated protein in 10 l of reaction buffer. For topo70, topo58, topo31, and ⌬cap, the protein concentrations used in the assay ranged from 1 M to 8.7 nM. For topo17 and the lysozyme control, the concentrations extended from 10 M to 87 nM. The reactions were incubated at 23°C for 15 min before the addition of 2.5 l of 50% glycerol, followed by analysis on a 6% native polyacrylamide gel at 4°C. The running buffer contained 25 mM Tris-HCl, pH 8.5, and 162 mM glycine. Due to the high pI values for the topo70 protein and the truncation fragments used here (Ͼ9.0), free protein and protein-DNA complexes migrated to the cathode and therefore only the free oligonucleotides entered the gel. The amount of unbound oligonucleotide in the gel was quantified using a phosphorimager and the ImageQuant software. The dissociation constant (K d ) was estimated from the protein concentration at which one-half of the total duplex oligonucleotide was bound to the protein (19).
Association of GST-topo31 with ⌬cap-DNA-The reaction was carried out by incubating GST-topo31 (ϳ2.5 M on beads) with an equal molar concentration of ⌬cap in the presence of 20 nM labeled suicide substrate in 45 l of reaction buffer. The reaction was rotated at 23°C for 2.5 h to permit suicide cleavage, and then the reaction was divided into three equal portions. The first 15-l portion was quenched by the addition of 5 l of 5ϫ SDS loading buffer and used as the control for the total amount of DNA-⌬cap generated in the reaction by suicide cleavage. The second 15-l portion was centrifuged at 10,000 rpm for 2 min to remove the beads, and the supernatant was added to 5 l of 5ϫ SDS loading buffer. The beads were washed once with 50 l of reaction buffer and suspended in 20 l of 1ϫ SDS loading buffer, and both the supernatant and the beads were analyzed by SDS-PAGE to test for the association of ⌬cap-DNA with GST-topo31. The third 15-l portion was centrifuged at 10,000 rpm for 2 min, and the beads were washed once with 50 l of 1ϫ reaction buffer, then suspended in 20 l of 300 mM KCl, 10 mM Tris-HCl, pH 7.4, 1 mM EDTA, 1 mM DTT, and rotated at 23°C for 5 min to elute the ⌬cap-DNA. The beads with the bound GST-topo31 were removed by centrifugation at 10,000 rpm for 2 min, and the supernatant containing the ⌬cap-DNA was adjusted to 10 mM Tris-HCl, pH 7.4, 1 mM EDTA, 1 mM DTT, and 100 mM KCl. Fresh glutathione-Sepharose 4B beads were added to the eluted ⌬cap-DNA and the mixture was rotated at 23°C for 2.5 h to test whether ⌬cap-DNA associates nonspecifically with the beads.
To test the effects of salt on the GST-topo31 interaction with ⌬cap-DNA, bead-bound covalent complexes were prepared as described previously and incubated for 5 min in 10 mM Tris-HCl, pH 7.4, 1 mM EDTA, 1 mM DTT, and KCl concentrations ranging from 100 to 400 mM. The beads were collected by centrifugation and resuspended in 15 l of 1ϫ SDS loading buffer and analyzed by SDS-PAGE. 5 l of 5ϫ SDS loading buffer was added to each supernatant and similarly analyzed. The gel was stained with Coomassie Blue and dried before exposure to film to compare the amount of labeled DNA present in the beads with the amount in the supernatant at the different salt concentrations.
Religation by ⌬cap in Absence of Topo31-To generate the substrate for religation, GST-topo31 bound to beads (ϳ2.5 M) was mixed with an equal molar amount of ⌬cap, and the mixture was incubated with 20 nM of labeled suicide substrate in reaction buffer in a total volume of 100 l at 23°C for 6 h to allow suicide cleavage to occur. The beads were collected by centrifugation, suspended in 100 l of 300 mM KCl and 10 mM Tris-HCl, pH 7.4, 1 mM EDTA, 1 mM DTT, and rotated at 23°C for 5 min to elute the ⌬cap-DNA from the beads. The beads were removed by centrifugation at 10,000 rpm for 2 min. To eliminate all traces of GST-topo31, 5 l of pre-equilibrated glutathione-Sepharose 4B beads were added to the supernatant and rotated at 23°C for 30 min before removal of the beads by centrifugation. This step was repeated three times. The supernatant that contains the ⌬cap-DNA was divided into four 25-l aliquots prior to preparing the samples for the religation assay. 50 l of 10 mM Tris-HCl, pH 7.4, 1 mM EDTA, 1 mM DTT was added to two of the aliquots to make 75-l reactions with a final KCl concentration of 100 mM. 3 l of topo31 (0.7 g/l) was added to the first 75-l reaction (ϩtopo31), and an equal volume of storage buffer was added to the second 75-l reaction (Ϫtopo31). The same procedure was used to prepare two additional reactions for religation except the KCl concentration was adjusted to 800 mM. All the reactions were incubated at 37°C for 2 min before religation was initiated by the addition of 5 l of the R11 religation oligonucleotide (final concentration, 2 M). A 10-l aliquot was removed prior to the addition of R11 for the zero time point analyses. 10-l aliquots were removed at the indicated times and quenched by adding an equal volume of 1% SDS. All the reactions were ethanol-precipitated and digested with 10 l of 1 mg/ml of trypsin at 37°C for 1 h. The samples were analyzed on a 20% sequencing gel and subjected to phosphorimager analysis using ImageQuant software.

Expression and Purification of Topoisomerase I Fragments-
The various fragments of human topoisomerase I employed in this study are shown schematically in Fig. 1, panel A. Both topo17 (residues 175-320) and topo31 have the same NH 2 terminus as topo70, but topo17 extends to the end of core subdomain II, whereas topo31 (residues 175-433) includes all of the cap region of the protein. The ⌬cap fragment corresponding to the catalytic domain begins where topo31 ends and extends through to the COOH terminus of the protein. Fusion constructs of topo17 and topo31 containing NH 2 -terminal GST were expressed and purified from E. coli. The fusion proteins were either used directly or the GST portion was removed by Factor Xa prior to use. Topo58 (residues 175-659) and the ⌬cap fragment (residues 433-765) were purified from recombinant baculovirus-infected insect cells. Analysis of the purified proteins by SDS-PAGE (Fig. 1, panel B) showed that all four fragments were essentially homogenous.
The Reconstitution of Relaxation Activity by a Mixture of Topo31 and ⌬cap-Although the ⌬cap fragment contains all of the amino acids that constitute the active site of the enzyme, no plasmid DNA relaxation activity was detectable in the purified fragment (Fig. 2, lane 9). As expected, no relaxation activity was associated with topo31 alone (Fig. 2, lanes 2-8). However, relaxation activity could be reconstituted by the addition of topo31 to ⌬cap. Under these conditions (1-h incubation), activity was first detectable at approximately a 2:1 molar ratio of topo31 to ⌬cap, and essentially complete relaxation was achieved at a molar ratio of 4 (Fig. 2, lanes 13 and 14,  respectively).
The Suicide Cleavage Activity of the Reconstituted Enzyme-A 5Ј end-labeled suicide substrate that contained a 14-bp duplex with an 11 base 5Ј-tail (Fig. 3, top of panel B) was used to test the cleavage activity of the topoisomerase I fragments alone or in pairwise combinations. Upon cleavage and formation of the covalent complex with the 5Ј end-labeled DNA, the AG dinucleotide at the 3Ј end of the scissile strand is released, preventing religation. Suicide cleavage resulted in the formation of a labeled oligonucleotide-protein species that could be detected by SDS-PAGE analysis. The results showed that none of the protein fragments alone had cleavage activity and that combining topo17 with ⌬cap did not reconstitute cleavage activity (Fig. 3, lanes 2-5). However, the combination of topo31 with ⌬cap yielded a labeled protein band that migrated slightly above the ⌬cap protein band on the Coomassie Blue-stained SDS-polyacrylamide gel (Fig. 3, panels A and B, lane 6) and that corresponded in size to a ⌬cap-DNA covalent complex. The amount of cleavage observed for the combination of topo31 with ⌬cap was less than that observed with a smaller amount of topo70 (Fig. 3, lane 7). These results confirmed that topo31 and ⌬cap can reconstitute topoisomerase I cleavage activity. When a suicide cleavage time course combining topo31 and ⌬cap was performed, cleavage reached a plateau after ϳ20 h of incubation under these conditions (data not shown).
Single Turnover Religation Activity-Religation was studied under single turnover conditions by assaying the ability of the covalent intermediate to attach a 5Ј-hydroxyl-terminated 11mer to the cleaved oligonucleotide (12-mer) to form a 23-mer product (18,20). The first step of the reaction was carried out by incubating the suicide cleavage substrate described previously with topo70 or with the reconstituted topo31-⌬cap complex to generate the corresponding covalent complexes. The religation reactions were initiated by the addition of the 11-mer acceptor oligonucleotide to the reaction mixtures. The religation assay for topo70 was carried out at 0.5 M KCl to promote dissociation of the topoisomerase after strand closure and to prevent recleavage of the product. Religation by the reconstituted enzyme had to be carried out at 100 mM KCl because higher salt concentrations dissociated the topo31-⌬cap complex (see below). Under these conditions the cleavage rate by the reconstituted activity was too slow to interfere with the religation measurement. The samples were treated with trypsin to remove all but a short trypsin-resistant peptide from the topoisomerase I-DNA covalent complexes prior to analysis in a sequencing gel (Fig. 4, panel A). The percentage of religated product formed at each time point was plotted for topo70 and the topo31-⌬cap mixture (Fig. 4, panel B). The results indicated that the religation kinetics for the reconstituted topo31-⌬cap complex and topo70 are very similar and that topo31 and ⌬cap can fully reconstitute the religation activity of topoisomerase I. DNA Binding as Measured by a Native Gel Shift Assay-⌬cap contains all of the critical residues involved in catalysis yet it lacks enzymatic activity. This lack of catalytic activity could result from a reduced affinity of ⌬cap for DNA. To test this possibility a native gel mobility shift assay was used to measure the DNA binding properties of the various topoisomerase-derived fragments. Similar to topo70, topo31, topo17, and ⌬cap are positively charged, and because a covalently bound oligonucleotide only partially neutralizes the positive charge, the protein-DNA complexes fail to enter the native gel. Under these conditions, K d is equal to the protein concentration at which the amount of unbound oligonucleotide observed in the gel has been reduced by a factor of 2 (19). Lysozyme has a similar pI value and was therefore used as a negative control for DNA binding by topo31, topo17, and ⌬cap. The binding assays showed that the affinity of the topo17 protein for the DNA was about the same as that of the lysozyme control, indicating that the binding is relatively nonspecific (K d of ϳ5 ϫ 10 Ϫ6 M) (Fig. 5, inset). Topo31 bound DNA with a K d of ϳ4 ϫ 10 Ϫ7 M, whereas ⌬cap bound the substrate DNA with a higher affinity (ϳ1 ϫ 10 Ϫ7 M), which is only 2-fold lower than that of topo70 (ϳ5 ϫ 10 Ϫ8 M). These results showed that topo31 binds DNA with a somewhat reduced affinity compared with topo70 and that the absence of activity for ⌬cap is not due to a failure to bind DNA. GST-topo31 Remains Associated with ⌬cap After Suicide Cleavage-Topo31 and ⌬cap together can reconstitute complete topoisomerase I activity, indicating that topo31 interacts with and activates ⌬cap either before or after the addition of DNA. However GST-topo31 bound on the glutathione-Sepharose 4B beads failed to pull down a detectable quantity of ⌬cap after incubating the two proteins together either in the presence or absence of DNA (data not shown). This result could indicate either that the interaction between ⌬cap and GST-topo31 is too weak to form a stable complex even in the presence of DNA, or that only a small fraction of the proteins interact to form the complex.
An alternative and more sensitive approach to detect the existence of the complex is to ask whether radioactively labeled ⌬cap-DNA and GST-topo31 remain associated after suicide cleavage. GST-topo31 bound to glutathione-Sepharose 4B beads was incubated with ⌬cap and labeled suicide substrate to form the covalent complex. The beads were collected by centrifugation to eliminate uncleaved substrate DNA and free ⌬cap. The bead-associated material was analyzed by SDS-PAGE (Fig. 6). The results showed that most of the radiolabeled ⌬cap-DNA covalent complex was associated with the beads (96%) (Fig. 6B, lanes 1 and 2), indicating that ⌬cap-DNA forms a complex with GST-topo31. However, it should be noted that the amount of ⌬cap containing covalently attached DNA was too small to be detected in the Coomassie Blue-stained gel (Fig.  6A, lane 1). The control analysis with isolated ⌬cap-DNA showed that the covalent complex did not associate nonspecifically with the glutathione-Sepharose 4B beads (Fig. 6B, lanes 4  and 5), indicating that GST-topo31 was responsible for mediating the association of ⌬cap-DNA with the beads. These results demonstrated that GST-topo31 activates ⌬cap and remains associated with ⌬cap after suicide cleavage.
Salt Effect on the Association of GST-topo31 and ⌬cap-DNA-To evaluate the stability of the ⌬cap-DNA/GST-topo31 complex, we tested the sensitivity of the complex to salt. The covalent ⌬cap-DNA complex that bound to the GST-topo31 on glutathione-Sepharose 4B beads was generated as described under "Experimental Procedures." The beads were divided into several aliquots and washed with buffers at different KCl concentrations (Fig. 7). Whereas only a trace amount of ⌬cap-DNA (2.9%) was released from the beads at 100 mM KCl, most of the labeled material was released into the supernatant (89%) at 300 mM KCl (Fig. 7, panels B and C, lanes 2-4). Over the range of KCl concentrations tested, the association of GST-topo31 with the beads was stable (Fig. 7, panel A). Washing with KCl concentrations of 400 mM or higher did not significantly increase the amount of label released from the beads (Fig. 7, panels B and C, lane 5 and data not shown). These results indicated that the association of ⌬cap-DNA with GST-topo31 was disrupted between 200 and 300 mM KCl.
Religation by ⌬cap Alone-The salt sensitivity of the GST-topo31/⌬cap-DNA interaction facilitated the separation of the ⌬cap-DNA covalent complex from GST-topo31. We purified the radiolabeled ⌬cap-DNA complex away from GST-topo31 in the presence of high salt and then carried out a religation reaction to test whether the covalent ⌬cap-DNA complex remains active and competent for religation. The religation reactions were carried out at 100 and 800 mM KCl in the presence or absence of added topo31 (Fig. 8). The results at 100 mM KCl showed that ⌬cap could carry out religation in the absence of added topo31, but the religation rate by ⌬cap-DNA in presence of added topo31 was at least 5-fold faster than the religation rate in the absence of topo31 (Fig. 8, panel B). Because topo31 can interact with ⌬cap at 100 mM KCl, we cannot exclude the possibility that the small amount of religation observed at this salt concentration in the absence of added topo31 was due to a trace amount of contaminating GST-topo31 undetected by Western blot analysis (data not shown). However, at 800 mM KCl contaminating GST-topo31 should be unable to bind ⌬cap and consequently any observed religation should result solely from the activity of ⌬cap. As shown in Fig. 8, religation by ⌬cap readily occurred at 800 mM KCl in the absence of topo31. The lack of stimulation by exogenously added topo31 established that the observed religation was not mediated by contaminating GST-topo31. Thus the result showed that ⌬cap-DNA could carry out the religation reaction alone, although the religation rate at 800 mM KCl was much slower than the rate at 100 mM KCl in the absence of topo31. DISCUSSION A functional analysis of fragments of human topoisomerase I shows that the cap region (topo31) of human topoisomerase I, which lacks catalytically important residues, can bind DNA although with a reduced affinity when compared with topo70. In an earlier study, a fragment corresponding to the same region of the Saccharomyces cerevisiae topoisomerase I was similarly shown to bind DNA (7). However, core subdomain II (topo17) does not bind DNA by itself despite displaying a structural homology to the homeodomain region found in a family of DNA-binding proteins (6). The ⌬cap fragment corresponds to the catalytic domain of the enzyme and contains all of the residues known to be directly involved in catalysis yet it is still enzymatically inactive. This lack of activity is not due to a defect in DNA binding because its affinity for DNA is only 2-fold lower than that of topo70. The absence of activity for ⌬cap is particularly surprising in view of the structural similarity of this portion of human topoisomerase I with that of the catalytic domains of integrase and vaccinia topoisomerase, both of which retain some catalytic activity as the isolated domains (14,15,21,22).
The catalytically active domain of vaccinia topoisomerase that lacks the NH 2 -terminal region has been shown to bind DNA at specific sites with a lower affinity than the full-length enzyme and to exhibit a reduced catalytic activity. The missing NH 2 -terminal domain contributes to DNA binding because point mutations at Tyr 70 and Tyr 72 in the NH 2 -terminal domain that have been shown to contact the substrate near the scissile phosphate in the major groove cause a similar effect on the activity as removing the NH 2 -terminal domain (23). When the crystal structures are compared, most of the active site residues of vaccinia topoisomerase spatially superimpose very well on the corresponding residues in human topoisomerase I except for the active site tyrosine (Tyr 274 ), which is displaced away from the active site pocket and is not in a position to directly attack the scissile phosphate (15,22,24). This observation suggests that a precleavage conformational change in the catalytic domain is necessary to establish the correct position of the active site tyrosine for nucleophilic attack on the DNA. It has been suggested that DNA contacts by the NH 2terminal domain of the enzyme are involved in facilitating this precleavage conformational change (15,23). In the crystal FIG. 5. DNA binding assays. The gel shift assay was carried out as described under "Experimental Procedures." The percentage of unbound duplex oligonucleotide present in the gel was quantified using the phosphorimager and plotted against the protein concentration. Because lysozyme and topo17 have a lower affinity for DNA, the protein concentrations used in the gel shift assays were 10-fold higher than those used for topo70 (‚), topo58 (ϫ), topo31 (᭜), and ⌬cap (Ⅺ). The binding profiles for topo17 (OE) and lysozyme (E) are shown in the inset.  structure of the integrase catalytic domain, the active site tyrosine is located on a flexible segment and is similarly distant from the other catalytic residues (21,22,25), again suggesting that a conformational change must precede cleavage. In both of these cases, the protein fragments that contain the displaced active site tyrosines retain some enzymatic activity, suggesting DNA binding alone is sufficient to induce the precleavage conformational change required for catalysis.
A possible explanation for the lack of activity of the isolated catalytic domain of human topoisomerase I (⌬cap) is that, similar to vaccinia topoisomerase and integrase, the architecture of the active site is not properly assembled for catalysis. More- over, by analogy with the vaccinia topoisomerase and the tyrosine recombinases, it may be that the nucleophilic Tyr 723 in the isolated catalytic domain is displaced from the proper position for cleavage. Because the isolated catalytic domain is inactive, we propose that DNA binding is not sufficient in this case to induce the precleavage conformational change that assembles the active site for catalysis, and instead the conformational change depends also on the presence of the cap region. A crystal structure of human topoisomerase I in the absence of DNA is required to test this conjecture. Nonetheless a 68-kDa recombinant human topoisomerase I has been shown by Raman and CD spectroscopy to undergo a conformational change after DNA binding and cleavage in solution (26). This transformation is mostly localized to the core and the COOH-terminal domains, and primarily reflects the relative movement of domains upon DNA binding. Whether the observed effects are related to a precleavage conformational change that assembles the proper active site remains to be determined.
Our results show that in human topoisomerase I, the cap region can bind DNA and is required to activate ⌬cap for cleavage. However, the cap region is not required for the religation reaction, and apparently once ⌬cap becomes covalently attached to the DNA after suicide cleavage, it is competent to carry out religation in the absence of the cap. What most clearly distinguishes religation from cleavage is the nature of the nucleophile. If indeed the cap is required to reorient the nucleophilic Tyr 723 for the cleavage reaction, then it might not be required for the religation reaction where the nucleophilic 5Ј-hydroxyl is held in place by virtue of the base pairing between the religation oligonucleotide and the downstream nonscissile strand. Alternatively, if ⌬cap is inactive for reasons other than the positioning of Tyr 723 , then the simple presence of the covalently bound DNA appears sufficient to maintain the appropriate activated conformation for religation.
The activation of ⌬cap by topo31 implies that there is an interaction between the two protein fragments at least when they are bound together on the DNA. Given that the only noncovalent contact between the cap and the remainder of the protein in the crystal structure involves a salt bridge between Lys 369 and Glu 497 within the "lips" region of the protein (6), it is not surprising that no interaction was detectable using a Coomassie Blue-stained SDS-polyacrylamide gel analysis in the absence of DNA. Although no interaction was seen in the presence of DNA using the same assay, the use of a much more sensitive assay involving cleavage of a radiolabeled suicide substrate did permit detection of a stable interaction. Our results do not allow us to address whether detection in this case was simply due to the increased sensitivity of the assay or whether the interaction between topo31 and the ⌬cap is stabilized by the formation of the covalent complex.
In the ternary topo31-⌬cap-DNA complex, the most likely region for communication between the cap and the catalytic domains is within the lip region where the two lobes of the intact protein interact with each other. Two lines of evidence support the notion that interactions within this region could be transmitted to the active site region of the enzyme or to the bound substrate DNA. First, two residues located within the lips region, the side chain of His 367 and a main chain nitrogen of Arg 364 , both contact a phosphate located in the nonscissile strand one nucleotide away from the cleavage site (6). Second, it has been shown that the camptothecin-resistant mutation G363V in the lips region of human topoisomerase I can suppress the lethal phenotype of a T718A mutant that, by itself, mimics camptothecin treatment by stabilizing the covalent intermediate (27). Molecular modeling studies further substantiate that structural changes within the lips region of the protein could influence the architecture of the active site of the enzyme and therefore potentially have an effect on the chemistry of catalysis (27).