Regulation of Catalysis by the Smallpox Virus Topoisomerase*

The poxvirus type IB topoisomerases catalyze relaxation of supercoiled DNA by cleaving and rejoining DNA strands via a pathway involving a covalent phosphotyrosine intermediate. Recently we determined structures of the smallpox virus topoisomerase bound to DNA in covalent and non-covalent DNA complexes using x-ray crystallography. Here we analyzed the effects of twenty-two amino acid substitutions on the topoisomerase activity in vitro in assays of DNA relaxation, single cycle cleavage, and equilibrium cleavage-religation. Alanine substitutions at 14 positions impaired topoisomerase function, marking a channel of functionally important contacts along the protein-DNA interface. Unexpectedly, alanine substitutions at two positions (D168A and E124A) accelerated the forward rate of cleavage. These findings and further analysis indicate that Asp168 is a key regulator of the active site that maintains an optimal balance among the DNA cleavage, religation, and product release steps. Finally, we report that high level expression of the D168A topoisomerase in Escherichia coli, but not other alanine-substituted enzymes, prevented cell growth. These findings help elucidate the amino acid side chains involved in DNA binding and catalysis and provide guidance for designing topoisomerase poisons for use as smallpox antivirals.

Topoisomerases are the targets of several classes of drugs that are important in anti-cancer and anti-microbial chemotherapy. These "topoisomerase poisons" act by trapping the covalent protein-DNA complex, thereby stabilizing toxic DNA breaks. Accumulation of DNA breaks then poisons cancer cells or infecting microbes. The poxvirus topoisomerases are thus attractive drug targets for poxvirus-specific topoisomerase poisons (7)(8)(9)(10)(11).
The elucidation by x-ray crystallography of two structures of smallpox virus topoisomerase bound to its DNA recognition site has allowed enzyme activation to be understood in structural terms (22). The poxvirus topoisomerases are two-domain proteins (22)(23)(24)(25) that bind to the 5Ј-(T/C)CCTT-3Ј site as monomers by forming C-shaped clamps around the DNA (22,23,26). Comparison of the vaccinia amino and carboxyl domain structures determined in the absence of DNA (23,25) to the smallpox-topoisomerase/DNA complex (22) allowed the conformational changes that mediate activation to be visualized. The N-terminal domain binds to DNA essentially as a rigid body, but the C-terminal catalytic domain undergoes a considerable structural reorganization upon binding (22). Multiple protein side chains along the DNA interface make contacts that are positioned to stabilize the active conformation. Particularly notable is the formation of ␣-helix 5. This segment of the protein is disordered in the absence of DNA (23), but upon binding to the favored 5Ј-(T/C)CCTT-3Ј site, forms an ␣-helix that inserts into the DNA major groove (22). Amino acid side chains projecting from helix 5 make specific contacts to base pairs ϩ3 to ϩ6 (for numbering of the DNA base pairs in the recognition site see Fig. 1B). Formation of helix 5 allows nearby segments of the amino acid chain to bind DNA and deliver one of the catalytic residues (Arg 130 ) to the vicinity of the scissile phosphate, thereby activating covalent chemistry. Other notable topoisomerase-DNA interactions involve: 1) ␤-strand 5, which inserts into the DNA major groove and makes sequence-specific interactions to base-pairs ϩ2 to ϩ4, 2) Arg 206 , which makes an upstream contact to a base pair edge at ϩ9, and 3) the active site residues (Arg 130 , Lys 167 , Arg 223 , His 265 , Tyr 274 ), which make a network of contacts involving residue ϩ1 and the scissile phosphate (22).
The topoisomerase/DNA complex structures, together with extensive previous work, have led to a detailed picture of the catalytic chemistry. The reaction is catalyzed by a combination of acid-base catalysis, in which the attacking nucleophile is deprotonated and the leaving group protonated, and electrostatic stabilization, in which the developing negative charge at the scissile phosphate is stabilized by basic side chains at the active site (Refs. 4, 6, 15, 27 and references therein). A study of the topoisomerase enzyme of Leishmania bound to DNA and containing vanadate, a transition state analog, at the position of the scissile phosphate has yielded a view of the probable transition state (28), confirming and extending the models developed in the poxvirus and human topoisomerase systems.
Here we present a study of amino acid side chains of the smallpox topoisomerase that were suggested by the structural analysis to have a functional role. We focused particularly on amino-acids at the protein-DNA interface that were not thoroughly studied in previous work in the vaccinia system. The effects of amino acid substitutions were analyzed in vitro in assays for full DNA relaxation or the individual steps in the topoisomerase reaction cycle. This work revealed a network of new contacts important for enzyme function, and unexpectedly also disclosed amino acid substitutions that increased the apparent rate of cleavage of suicide substrates in vitro.

EXPERIMENTAL PROCEDURES
Site-directed Mutagenesis-Mutations were introduced into the variola double cysteine mutant topoisomerase gene by using the QuikChange mutagenesis kit (Stratagene, La Jolla, CA). Plasmid pET29a double cysteine mutant VAR-TOPO was the template for the mutagenesis reaction. All mutations were confirmed by dideoxy DNA sequencing method.
Topoisomerase Expression and Purification-pET29based plasmids containing mutant topoisomerase genes were transformed into Escherichia coli BL21(DE3) Codon-Plus RIL (Stratagene, La Jolla, CA) and were induced by adding IPTG. Double cysteine and mutant topoisomerases were purified from soluble bacterial lysates using SP-Sepharose column chromatography. The protein concentrations of the SP-Sepharose preparations were determined by using the dye binding method (Bio-Rad) with bovine serum albumin as the standard.
DNA Relaxation Assay-Reaction mixtures (10 l) containing 20 mM Tris-HCl, pH 8.0, 100 mM NaCl, 1 g of pUC19 plasmid DNA and 4.6 ng of double cysteine or mutant topoisomerases (4-fold molar excess of plasmid molecules over topoisomerase molecules) were incubated at 25°C for appropriate time. The reactions were stopped by the addition of a solution containing SDS (2% final concentration), glycerol, and bromphenol blue. Samples were analyzed by electrophoresis through a 1.0% horizontal agarose gel in TAE buffer (40 mM Tris acetate and 1 mM EDTA). The gels were stained in 0.5 g/ml ethidium bromide solution and photographed. The density of bands was measured by scanning the photograph using a STORM Phospho-Imaging Analyzer. All assays in this study were repeated at least twice to estimate error.
Forward Cleavage Assays-A 16-mer CCCTT-containing DNA oligonucleotide was 5Ј end-labeled by enzymatic phosphorylation in the presence of [␥-32 P]ATP and T4 polynucleotide kinase and annealed to a complementary 18-mer bottom strand. Cleavage reaction mixtures containing (per 20 l) 20 mM Tris-HCl, pH 8.0, 100 mM NaCl, 8 pmol of 16-mer/18-mer DNA, and topoisomerase were incubated at 25°C. Covalent complexes were denatured by addition of SDS to 2%. The samples were electrophoresed through a 12% polyacrylamide gel

Selection of Amino Acids for
Analysis-The amino acid substitutions studied in this work are listed in Table 1. All mutants were generated in a gene encoding the smallpox virus topoisomerase that also contained the substitutions C100S and C211S (termed "CSCS" below). These substitutions were required for formation of diffraction quality crystals of the smallpox topoisomerase with DNA (22), and so were included here to allow direct comparison with structural data. The C211S substitution diminished the activity of the enzyme slightly (Table 1), which had the advantage of slowing the reaction rates and simplifying some of the kinetic studies.
The DNA-contacting residues ( Fig. 1B) were typically substituted with alanine to allow the effects of selective truncation of the side chain to be studied. Residue His 39 projects off the N-terminal domain and makes a contact to the T at position Ϫ1. Gln 69 projects off ␤-5 of the N-terminal domain, and makes a sequence-specific contact to the major groove at position ϩ2. Arg 80 projects from ␣-helix 3, which connects the N-and C-terminal domains, and contacts the Thy at ϩ1. Lys 133 and Lys 135 extend from helix 5 and make major groove contacts at positions ϩ5 and ϩ6. Arg 206 makes the most upstream contact, extending to the ϩ9 position. Tyr 209 contacts ϩ6 and Thr 266 is positioned to contact Ϫ1 (though this part of the DNA is missing in the structures).
The active site residues (Arg 130 , Lys 167 , Arg 223 , His 265 , and Tyr 274 ) have all been studied previously in the vaccinia topoisomerase background (4,6,15,(17)(18)(19)(20). Here we substituted these residues with alanine, as in previous work. Catalytic site residues Arg 130 , Lys 167 , and Arg 223 were also substituted with methionine in order to maintain similar hydrophobic contacts along the length of the side chain, while selectively eliminating the charged functional group, a type of substitution not previously studied. The Tyr 274 catalytic tyrosine was substituted with Phe to remove the hydroxyl group that serves as the nucleophile in the transesterification reaction mediating initial cleavage, a type of substitution also studied previously (16,29).
Residue Asp 168 was implicated as potentially important because it lies on the outer rim of the active site, positioned to interact with the 5Ј-hydroxyl group at the scissile phosphate, and well positioned to participate in the religation reaction. In the structures, the main chain NH of Asp 168 contacts the DNA phosphate on the 3Ј-side of ϩ2. The Asp 168 side chain is available to contact the cleaved DNA strand in the vicinity of Ϫ1 or Ϫ2, but this portion of the DNA is missing in the available structures so detailed interactions are uncertain. An initial study of a D168A topoisomerase derivative indicated that this side chain was important for DNA relaxation (22), but did not specify the step involved.
Residue Glu 124 is an unusual buried acidic residue. The side chain is in a generally hydrophobic environment in the protein interior, but it does have the potential to exchange protons through a series of relays to the active site. Possibly the relay system involves an imidic acid tautomer of the peptide backbone (22). Previous studies of the effects of E124A on DNA relaxation in the vaccinia (21) and variola enzymes (22) showed no strong reduction in rate.
Topoisomerase derivatives containing each of these substitutions were analyzed for their effects on DNA relaxation, the forward rate of DNA cleavage, and the position of the equilibrium between cleavage and religation. In what follows, mutant derivatives are referred to as "CSCS" followed by specifying the substitution (e.g. CSCS D168A).
Effects of Amino Acid Substitutions on DNA Relaxation-DNA relaxation was analyzed using supercoiled bacterial plasmids as substrates. Following incubation with topoisomerase protein, supercoiled and relaxed topoisomers were separated by electrophoresis and quantified. In these reactions, 0.13 pmol of topoisomerase were added to 0.57 pmol of plasmid, so that relaxation of the substrate required that each protein molecule carry out several cycles of DNA relaxation, product release, and rebinding to new DNAs. Sample reactions are shown in Fig. 2A for the CSCS reference enzyme and CSCS K167M. Results are quantified in Fig. 2B and summarized in Table 1. For some of the mutants, relaxation half-times have been previously reported (denoted by the asterisks in Table 1 and Ref. 22).
Substitutions in the previously identified active site residues (R130A, K167A, R223A, H265A, and Y274F) severely impaired relaxation, paralleling previous studies in the vaccinia system (4,6,15,(17)(18)(19)(20)22). The CSCS topoisomerase derivatives containing methionine substitutions R130M, K167M, and R223M were also severely reduced in activity, implicating the terminal basic charge of the side chain as important for catalysis, and arguing against the idea that creation of a cavity in the active site by the alanine substitution was the main deficit. CSCS D168A was also severely impaired. The CSCS E124A and CSCS E124Q  Effects of Amino Acid Substitutions on Cleavage of Suicide Substrates-The forward rate of DNA cleavage (k cl ) can be monitored in isolation using "suicide substrates" (Fig. 3A). In these reactions, formation of the phosphotyrosine intermediate generates a short oligonucleotide that is released after cleavage, trapping the covalent complex. Sample time course studies comparing the CSCS enzyme and CSCS D168A and CSCS R206A are shown in Fig. 3B. To analyze the forward rate of cleavage, it is important to carry out reactions in the presence of saturating enzyme, so titration reactions were carried out initially to identify saturating conditions (data not shown). Examples of quantitation of time course reactions are shown in Fig. 3, C and D, and rates are summarized in Table 1.
Topoisomerase derivatives with substitutions in the N-terminal domain and connecting ␣-three helix (H39A, Q69A, and R80A) were reduced in rate 10 -100-fold. Some of the catalytic domain contact residues were also strongly affected (CSCS with K133A, R206A, Y209A, and T266A), while CSCS K135A was nearly equal to CSCS. All of the catalytic site residues (CSCS with R130A K167A, R223A, H265A, and Y274F) were severely reduced in activity as expected from previous work.
Residue Arg 206 was the only side chain contacting position ϩ9 in the structure, allowing the basis of sequence specific recognition at this DNA position to be studied in more detail ( Fig.  3D and Table 2). We found that substituting the ϩ9 Gua residue with each of the other three bases reduced the forward rate of cleavage by the CSCS enzyme about 10 -50-fold, indicating that this DNA position is recognized in a sequence specific fashion by the topoisomerase. When the Arg 206 side chain was truncated by alanine substitution, the reaction rate with the standard DNA substrate was reduced about 50-fold. Thus disrupting the ϩ9 contact with either a DNA mutation or the R206A side chain truncation had quantitatively similar effects. However, when the CSCS R206A enzyme was tested on DNA substrates with substitutions at ϩ9, no further reduction in rate was observed. Changes at ϩ10 or ϩ8 had little or no effect on the CSCS enzyme (Ͻ3-fold), and CSCS R206A was similarly unaffected. Both the CSCS topoisomerase and CSCS R206A were strongly reduced in activity by an Ade substitution at position ϩ2 within the 5Ј-CCCTT-3Ј pentamer. This indicates that truncating the Arg 206 side chain resulted in a loss of ability to discriminate the base pair at position ϩ9, but this was not a general loss of discrimination because CSCS R206A responded similarly to the CSCS topoisomerase to changes at positions ϩ10, ϩ8, and ϩ2. Thus the functional analysis indicates that the Arg 206 contact is the basis of sequence discrimination at position ϩ9.
Surprisingly, alanine substitutions at both Glu 124 and Asp 168 substantially increased the measured k cl , a mutant phenotype  not previously reported despite the wealth of mutagenesis studies for the related vaccinia enzyme. For CSCS D168A, the apparent forward rate of cleavage was increased by more than 10-fold. The E124A substitution increased the forward cleavage rate 5-fold. Curiously, the E124Q substitution slightly slowed the cleavage rate. We discuss the possible explanations for these observations below.
Effects of Mutants Assayed on Equilibrium Cleavage Substrates-The position of the equilibrium defined by the rates of covalent complex formation and religation (k cl ) can be measured on equilibrium oligonucleotide substrates. These substrates contain the 5Ј-CCCTT-3Ј core pentamer in the center of a 36-mer oligonucleotide. After transesterification to form the 3Ј-phosphotyrosine intermediate, all the strands remain annealed because of the longer DNA lengths (Fig. 4A). This is in contrast to the suicide substrate, where the free portion of the cleaved strand is released (Fig. 3A). Consequently, in the equilibrium substrate, the 5Ј-hydroxyl of the cleaved strand is available for religation. The cleavage and religation reactions are quite fast on these substrates, achieving equilibrium in most cases within 4 s (data not shown). Thus the fraction of the DNA in the covalent complex at times Ͼ4 s reports the position of the cleavage-religation equilibrium. In addition, the religation rates (k rel ) can be calculated from the measured forward rates of cleavage (k cl ) and the position of the cleavage religation equilibrium (K cl ) using the relationship K cl ϭ k cl /k rel . These values are shown in the right column in Table 1.
The amounts of enzyme required for saturation were first determined by titration as shown in Fig. 4, B and C. Table 1 summarizes the fraction of DNA molecules in the covalent complex at equilibrium for all the topoisomerase derivatives tested. For the CSCS reference enzyme, the fraction of substrate molecules in the covalent complex at equilibrium is 0.13. For CSCS with several of the active site substitutions, the equilibrium position was not reached by Ͼ24 h (R130A, K167A, and R223A), or no covalent complex formation was detected (CSCS Y274F). CSCS H265A reached equilibrium after 4 min. At equilibrium, the percentage of substrate in the covalent complex was 0.35, consistent with a more severe defect for the CSCS H265A enzyme in religation than in the forward cleavage step, as suggested in previous work in the vaccinia system (17,20).
For several of the enzymes with substitutions in DNA contact residues, the position of the equilibrium K cl was changed, so that at the steady state less of the DNA was in the covalent complex (Table 1). This is consistent with the idea that these protein-DNA contacts are more important for activating the initial cleavage reaction than for religation. Relatively strong effects were seen for CSCS with Q69A, R80A, K133A, R206A, Y209A, and T266A. Weaker but possibly significant reductions in ratio were seen at CSCS with K135A and H39A.
For CSCS D168A, the fraction of substrate molecules in the covalent complex was increased to 0.36. The Asp 168 side chain is positioned on the side of the active site near His 265 and close to the inferred position of the 5Ј-OH group that participates in religation. Thus Asp 168 is implicated by this data in religation as well.
For the Glu 124 substitutions, the two changes studied had different effects. CSCS E124Q was slightly altered, so that only about half as much of the substrate was in the covalent complex at equilibrium (0.06). For CSCS E124A, there was no significant change compared with the CSCS reference enzyme.
Effects of 5 mM MgCl 2 on the Activity of CSCS, CSCS D168A, and CSCS E124A-A possible mechanism for the reduction in relaxation rate with CSCS D168A is that it may be defective in product release. In the relaxation assay, the plasmid DNA substrate is in excess, so each topoisomerase molecule must undergo several cycles of DNA cleavage, religation, and product release to complete relaxation of the plasmid population. Studies of the vaccinia topoisomerase have shown that relaxation reactions carried out under the conditions in Fig. 2 are limited at the product release step. However, addition of 5 mM MgCl 2 accelerates release without affecting the cleavage rate k cl (30). We thus compared reactions containing 5 mM MgCl 2 to reactions containing 2.5 mM EDTA (Fig. 5).
For the CSCS enzyme, we found that the time to 50% relaxation was reduced from 12 min in EDTA to 0.75 min in the presence of MgCl 2 , a reduction of 16-fold. For CSCS D168A, the half-time was unmeasurable in 2.5 mM EDTA because a portion of the substrate became relaxed quickly but no further relaxation took place even at long incubation times. Addition of 5 mM MgCl 2 allowed complete relaxation of the substrate, with a half-time of 2.4 min, representing an increase of at least 600fold. For CSCS E124A, the half-time went from 9.3 min in 2.5 mM EDTA to 1 min in 5 mM MgCl 2 , roughly paralleling the acceleration seen with the CSCS reference enzyme. Thus the time necessary to relax half the substrate was reduced for CSCS, CSCS D168A, and CSCS E124A in 5 mM MgCl 2 , and the change was comparatively greater for CSCS D168A. This supported the hypothesis that product release is limiting, and that the barrier is comparatively greater with CSCS D168A.
Effects of NaCl Concentration on the Activity of CSCS, CSCS D168A, and CSCS E124A-We next investigated whether CSCS D168A binds DNA more tightly than the CSCS reference enzyme. Previous studies of the vaccinia topoisomerase have shown that DNA binding affinity can be probed in the single cycle cleavage assay by titration of NaCl in reaction mixtures. The increased salt concentration diminishes DNA affinity by ionic screening without affecting k cl (30), so that altered DNA affinity results in slower covalent complex formation.
Comparison of covalent complex formation for CSCS, CSCS D168A, and CSCS E124A in the presence of different NaCl concentrations showed that CSCS D168A was less sensitive to increased NaCl concentrations (Fig. 6). To carry out this study, reaction times were adjusted for each enzyme to maximize the sensitivity of detecting small differences. Titration of NaCl revealed that all enyzmes showed a progressive reduction in rate above 250 mM NaCl, but the decline was faster for CSCS and CSCS E124A (Fig. 6, A-C, results are plotted in Fig. 6D). Reaction rates were measured at 250 mM NaCl, and found to be reduced (compared with 100 mM NaCl) by 43-fold for CSCS, 10.5-fold for CSCS E124A, and 1.9-fold for CSCS D168A (Fig.  6E). These data support the idea that CSCS D168A is considerably increased in DNA binding affinity compared with CSCS, while CSCS E124A may be slightly increased.
Toxicity of the D168A Topoisomerase in E. coli-We also observed that the CSCS D168A was more toxic to E. coli cells during overexpression than the unmodified CSCS enzyme (Fig.  7). To document this, E. coli cells transformed with topoisomerase expression vectors were streaked on plates lacking IPTG, 3 and so not inducing topoisomerase expression, or plates containing 400 M IPTG, which induce topoisomerase expression to high levels. After overnight incubation at 37°C, no growth was seen for cells containing the D168A mutant in the presence of 400 M IPTG. Growth was seen, however, in the presence of CSCS D168A in the absence of IPTG. Cells expressing the CSCS topoisomerase, or containing the empty expression vector, grew in both the presence and absence of 400 M IPTG. Expression of the CSCS did have some effect on cells, since there was less growth than for cells containing the empty expression vector, and after prolonged storage of plates many of the transformed cells died (data not shown). For the case of CSCS E124A and CSCS E124Q, there was no reduction of growth of E. coli under inducing conditions (data not shown). Active site substitutions also abrogated the toxicity (data not shown). Comparison of cells expressing the CSCS topoisomerase with cells expressing the fully wild-type enzyme showed that the wild type was slightly more toxic (data not shown), consistent with the CSCS substitutions reducing the enzyme activity slightly. We did not attempt to study D168A or E124A in the fully wild-type background (i.e. Cys at 100 and 211) because of the expected increase in toxicity. In summary, the most notable finding was that truncating the Asp 168 side chain by the alanine substitution resulted in increased toxicity of topoisomerase expression in E. coli. 3 The abbreviation used is: IPTG, isopropyl-1-thio-␤-D-galactopyranoside.

DISCUSSION
The phenotypes of mutant topoisomerase derivatives analyzed in this study are summarized on the topoisomerase/DNA complex structure in Fig. 8 together with mutants previously analyzed in the vaccinia topoisomerase system (6, 12, 15, 17-20, 29, 31-33)(summarized in supplemental Table S1). Positions of substitutions that reduced the forward rate of cleavage k cl are shown in red in Fig. 8A. Substitutions that increased k cl are shown in blue in Fig. 8B. Amino acid side chains important in the forward cleavage reaction line the protein/DNA interface, extending from DNA residues ϩ9 to at least Ϫ1. These amino acids include the catalytic site residues and amino acids within the central cavity of the enzyme. Most of the residues implicated as making sequence specific contacts from analysis of the x-ray structures had detectable phenotypes when substituted with alanine. Studies of alanine substitutions of Arg 206 , Asp 168 , and Glu 124 were particularly informative and are discussed further below.
The importance of the contact between Arg 206 and Gua at ϩ9 could be documented by a "loss-ofdiscrimination" experiment. Substituting the Arg 206 side chain with alanine resulted in a reduction in the forward rate of cleavage. Substitutions at position ϩ9 in the DNA similarly diminished cleavage by the CSCS topoisomerase. However, when CSCS R206A was tested in the presence of substitutions at ϩ9, no further reduction in rate was found. Both the CSCS enzyme and CSCS R206A remained capable of discriminating base pair substitutions at position ϩ2. Thus the selective loss of discrimination with CSCS R206A at the ϩ9 base pair implicates the R206 contact as the basis of specific recognition.
CSCS D168A and CSCS E124A showed a novel phenotype, which was an acceleration of the apparent forward rate of cleavage on the suicide substrate. Both Asp 168 and Glu 124 lie at the margins of the active site. Asp 168 lies near the protein surface close to the inferred binding site for the 5Ј-oxygen that acts as the nucleophile for religation. Glu 124 is buried in the protein interior. Both of these residues are strictly conserved in the poxvirus topoisomerases, though not universally conserved in the other type IB enzymes. The conservation suggests that these side chains carry out important functions in the poxvirus enzymes, but that they are probably not strictly required for catalysis.
Studies of CSCS D168A were particularly informative. CSCS D168A showed a reduction in the rate of DNA relaxation under conditions requiring multiple cycles of DNA relaxation and product release. Titration of NaCl in forward cleavage assays indicated that CSCS D168A has an increased affinity for DNA. Accelerating product release in the relaxation assay by addition of 5 mM MgCl 2 resulted in a dramatic increase in rate for CSCS D168A, consistent with the idea that tighter DNA binding led to impaired product release. One simple explanation for the increased DNA affinity would be that the acidic D168 side chain normally engages in an electrostatic repulsion with the negatively charged phosphates in the substrate DNA, so that remov-  DECEMBER 8, 2006 • VOLUME 281 • NUMBER 49 ing the side chain increased the affinity. The D168A substitution could have increased the forward rate of cleavage by several means. Single cycle cleavage assays were carried out in the presence of saturating topoisomerase, so tighter DNA binding alone is not likely to be resposible for the faster rate. The D168A substitution might increase binding in a conformation required for cleavage, or act directly at the chemical step to accelerate catalysis. Assays using the equlibrium cleavage substrate revealed a greater proportion of covalent complexes at steady state for CSCS D168A than for wild-type CSCS, indicating that the forward rate of cleavage was increased more than the religation rate (13-fold for k cl versus 4.6-fold for k rel ). In the structure, the Asp 168 side chain lies near the expected position of the 5Ј-hydroxyl group that is the nucleophile for religation, close to the His 265 side chain that is also implicated in religation by data in Table 1 and previously published studies (6,12,15,17,18,20,32). The Asp 168 side chain may normally help remove a proton from the 5Ј-hydroxyl group that is the nucleophile for religation, and conversely the D168A substitution may make the environment more acidic, thereby promoting cleavage. More broadly, the topoisomerase must maintain a delicate balance between cleavage, allowing release of DNA supercoils, and religation, which restores DNA continuity but prevents further supercoil release (15). This work implicates Asp 168 as a key element of the active site that modulates and optimizes DNA affinity and the cleavage and religation efficiencies.

Catalysis by the Smallpox Virus Topoisomerase
The role of the Glu 124 side chain is less clear. Substitution with alanine accelerated the rate of cleavage on the suicide substrate 5-fold, but the E124Q substitution had no effect. For CSCS E124A, the position of the cleavage-religation equilbrium was unchanged, implying that the religation rate was accelerated as much as cleavage. CSCS E124A showed slightly reduced sensitivity to 250 mM NaCl, indicating possible tigher DNA binding, though this was not pronounced enough to inhibit religation by impairing product release. Possibly the E124A substitution enhanced binding in a conformation promoting cleavage or increased the catalytic rate, though it would require further experiments to clarify the mechanism.
A goal of studies of the smallpox topoisomerase has been to develop topoisomerase poisons that might be used as therapy for poxvirus infection. The finding that CSCS D168A is toxic in E. coli may be helpful in this regard. The D168 and H265 side chains are near one another, specifying a region of the enzyme surface important for religation. Potentially small molecules might be identified that bind to this region and so interfere with religation and thus could serve as poisons of the poxvirus topoisomerases. The smallpox topoisomerase/DNA complex x-ray structures (22) provide templates for possible design of antiviral molecules based on this idea.