Site-directed Mutagenesis of Conserved Aspartates, Glutamates and Arginines in the Active Site Region of Escherichia coli DNA Topoisomerase I*

To catalyze relaxation of supercoiled DNA, DNA topoisomerases form a covalent enzyme-DNA intermediate via nucleophilic attack of a tyrosine hydroxyl group on the DNA phosphodiester backbone bond during the step of DNA cleavage. Strand passage then takes place to change the linking number. This is followed by DNA religation during which the displaced DNA hydroxyl group attacks the phosphotyrosine linkage to reform the DNA phosphodiester bond. Mg(II) is required for the relaxation activity of type IA and type II DNA topoisomerases. A number of conserved amino acids with acidic and basic side chains are present near Tyr-319 in the active site of the crystal structure of the 67-kDa N-terminal fragment of Escherichia coli DNA topoisomerase I. Their roles in enzyme catalysis were investigated by site-directed mutation to alanine. Mutation of Arg-136 abolished all the enzyme relaxation activity even though DNA cleavage activity was retained. The Glu-9, Asp-111, Asp-113, Glu-115, and Arg-321 mutants had partial loss of relaxation activity in vitro. All the mutants failed to complement chromosomal topA mutation inE. coli AS17 at 42 °C, possibly accounting for the conservation of these residues in evolution.

DNA topoisomerases (for review, see Refs. 1-7) catalyze the interconversion of different DNA topological isomers by first forming a covalent enzyme-DNA intermediate via nucleophilic attack of a tyrosine hydroxyl on the DNA phosphodiester linkage. After strand passage through the break, religation involving nucleophilic attack of the displaced DNA hydroxyl group on the phosphotyrosine linkage takes place. Type IA and type II DNA topoisomerases are linked to the 5Ј-phosphoryl end of the cleaved DNA while type IB DNA topoisomerases are linked to the 3Ј-phosphoryl end. Mg(II) is required for the relaxation activities of both type IA and type II DNA topoisomerases but not for the type IB enzymes. The detailed catalytic mechanism of DNA cleavage and religation by topoisomerases remains to be elucidated. The mechanism of the type IA and type II topoisomerase may share similarities with other enzymes that also require Mg(II) for nucleotidyl transfer activity.
Tyr-319 of Escherichia coli DNA topoisomerase I is the catalytic residue that provides the hydroxyl group for forming the covalent intermediate with DNA. The three-dimensional struc-ture of the 67-kDa N-terminal domain of this enzyme has been determined by x-ray crystallography (8). In this structure, Tyr-319 is present in the interface between domains I and III. It has been pointed out (8) that the spatial arrangement of the three acidic residues Asp-111, Asp-113, and Glu-115 in the active site region is similar to the acidic residues that coordinate two divalent cations in the exonuclease catalytic site of Klenow fragment (9). However, the structure observed has to undergo additional conformational changes before there is sufficient space in the active site region for DNA and possibly Mg(II) to bind. A number of residues found in the active site, including Glu-9, Asp-111, Asp-113, and Glu-115, and Arg-321, are strictly conserved among type IA DNA topoisomerase sequences (Fig.  1). This evolutionary conservation extends from the archaeal (Methanococcus jannaschii) topoisomerase I to human topoisomerase III. There is a lower degree of conservation for Arg-136 which is also near the active site tyrosine. The acidic and basic functional groups on these residues may be important for the enzyme catalytic function. For DNA polymerases, two divalent metal ions coordinated by carboxylates are essential for the enzyme activity (10 -14). One or two metal ion mechanisms utilizing carboxylates for coordination have been proposed for the EcoRI and EcoRV restriction endonucleases (15,16). Certain bacterial transposases and retroviral integrases have a conserved DDE motif that has been proposed to be involved directly in catalysis (17)(18)(19). In a mechanism proposed for the yeast site-specific recombinases Flp and R, invariant arginines either facilitate the phosphoryl transfer by stabilizing the leaving group or the partial charge on the phosphorus in the transition state during the phosphodiester cleavage or exchange reactions (20). These recombinases do not require divalent ions for their activity. Their mechanisms may also share some similarities with topoisomerases. It should be noted that Mg(II) is not absolutely required for the DNA cleavage activity of E. coli DNA topoisomerase I.
To investigate the roles of the carboxylates and arginine residues in the active site region of E. coli topoisomerase I, they were altered by site-directed mutagenesis to alanines, abolishing the acidic or basic functional groups. The mutant enzymes were expressed and purified. Different enzymatic assays were carried out to determine how the mutation affected the interaction of the topoisomerase enzyme with DNA and/or Mg(II).

EXPERIMENTAL PROCEDURES
Materials-All chemical reagents used were ultrapure or Baker analyzed reagent grade. Solutions were prepared with water first deionized with the Barnstead Nanopure system and then passed over a Bio-Rad Chelex 100 resin (100 -200 mesh sodium form) to remove any remaining contaminating metal ions. Tubes, spectrophotometric cells, and glassware for metal ion-sensitive experiments were first washed with 10 mM EDTA and then rinsed extensively with metal-free water * This work was supported by Grant GM54226 from NIGMS, National Institutes of Health, Department of Health and Human Services. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. before use. Plasmid DNA was purified by cesium chloride centrifugation.
Enzyme Expression and Purification-Wild-type E. coli DNA topoisomerase I enzyme was expressed from E. coli MV1190 cells transformed with plasmid pJW312. The R136A and R321A mutants were expressed in E. coli AS17 (topA am PLL1(Tc r supD ts ), from R. E. Depew, Northeastern Ohio University). The E9A, D111A, D113A, and E115A mutants were expressed in E. coli GP200 (gyrA(Nal r ) gyrB225⌬ (topAcysB)204) (22). Enzyme purification was carried out using the previously described procedures (23) with an additional hydroxyapatite chromatography step between the phosphocellulose and singlestranded DNA-agarose columns.
Relaxation Activity Assay-Wild-type and mutant enzymes were serially diluted and assayed for relaxation activity in 20 l with 0.5 g of negatively supercoiled plasmid DNA, 10 mM Tris-HCl, pH 8.0, 50 mM NaCl, 0.1 mg/ml gelatin, and 6 mM MgCl 2 unless indicated otherwise. Incubation was at 37°C for 30 min. The reactions were stopped by the addition of 5 l of 50% glycerol, 50 mM EDTA, and 0.5% (v/v) bromphenol blue. After electrophoresis in a 0.7% agarose gel with TAE buffer (40 mM Tris acetate, pH 8.1, 2 mM EDTA), the DNA was stained with ethidium bromide and photographed over UV light.
Cleavage of 5Ј-End-labeled Single-stranded DNA-Plasmid pT7-1 DNA (from U. S. Biochemical Corp.) was cleaved with EcoRI and then labeled at the 5Ј end with [␥-32 P]ATP and T4 polynucleotide kinase. The labeled DNA was denatured to single strand and incubated with the wild-type and mutant enzymes, and then sodium hydroxide was added to trap the covalent complex and cleavage of the DNA (24). After electrophoresis in a 6% DNA sequencing gel, the 5Ј-end-labeled DNA cleavage products were visualized by autoradiography.
Covalent Complex with Oligonucleotide Substrate-The singlestranded oligonucleotide 5Ј-CAATGCGCT-3Ј, containing the sequence of a strong cleavage site (24), was labeled at the 3Ј end with [␣-32 P]dATP and terminal deoxynucleotidyl transferase. After incubation of 0.15 g of the labeled oligonucleotide with 0.4 g of the enzyme in 10 mM Tris-HCl, pH 7.6, 1 mM EDTA, 20 mM potassium phosphate at 37°C for 5 min, the reaction was stopped with the addition of 1% SDS. The enzyme was separated from the non-covalently bound oligonucleotides by electrophoresis in a 10% SDS-polyacrylamide gel. The labeled covalent complex was visualized by autoradiography of the dried gel.
tion was carried out using the Molecular Dynamics PhosphorImager.
Intrinsic Tryptophan Fluorescence Measurements-Fluorescence measurements were performed with the Perkin-Elmer LS-5B spectrometer with excitation at either 280 or 295 nm at either 42°C or room temperature (ϳ25°C). The spectral bandwidths were 3 and 10 nm, respectively, for excitation and emission. The wild-type or mutant topoisomerase I was present at 0.1 mg/ml in 20 mM potassium phosphate, pH 7.5, 0.1 M KCl, 0.2 mM dithiothreitol. All the measurements were corrected for the spectrum of the buffer used.
Equilibrium Dialysis to Determine Mg(II) Binding Stoichiometry-1 ml of topoisomerase I (0.4 mg/ml) was dialyzed against 400 ml of buffer (20 mM potassium phosphate, pH 7.5, 0.1 M KCl, 0.2 mM dithiothreitol, and 400 M MgCl 2 ) at room temperature for 7 h. The enzyme and dialysis buffer samples were submitted to Quantitative Technologies, Inc., NJ, for Mg(II) content analysis using the inductively coupled plasma method.

RESULTS
Expression and Purification of the Topoisomerase I Mutants-After site-directed mutagenesis of the plasmid pJW312 to produce the desired alanine substitutions, expression of the mutant proteins in E. coli strain AS17 was examined by SDSgel electrophoresis of the total soluble extract followed by Coomassie Blue staining of the gel. The expression level of the Arg-136 mutant was comparable with that of the wild-type topoisomerase I while expression of the Arg-321 mutant was detectable but lower than the wild type. Bands corresponding to the Glu-9, Asp-111, Asp-113, and Glu-115 mutants could not be identified in AS17 extracts (data not shown). Expression of these three mutant enzymes were then carried out in E. coli strain GP200. Detectable levels of expression of these mutant enzymes allowed the purification of the mutant enzymes to homogeneity (Fig. 2).
Effect of the Mutations on Enzyme Activities-Purified mutant enzymes were diluted serially and compared with the wild-type topoisomerase I for relaxation of negatively supercoiled plasmid DNA in relaxation buffer containing 6 mM MgCl 2 (Fig. 3A). The results showed that the Arg-136 mutant was totally inactive in the relaxation assay. No relaxation activity was observable even with 400 ng of the enzyme, a 50-fold excess over the amount needed to observe relaxation by the wild-type enzyme. The other mutants had varying degrees of loss of relaxation activity. Examination of the effect of dilutions and time course of relaxation (Fig. 3B) showed that the Glu-9 mutant had the greatest reduction in catalytic activity (Ͼ90% reduction). The Glu-115, Asp-113, and Arg-321 mutants had about 80 -90% reduction in activity. The Asp-111 mutant was closest to the wild-type enzyme in activity (Ͻ50% reduction).
The E. coli strain AS17 does not grow at 42°C because of the  temperature sensitivity of the suppressor for the chromosomal topA am mutation. The pJW312 plasmid carrying the wild-type topoisomerase I gene can complement this chromosomal mutation (21). Each one of the active site mutations tested here was found to abolish this in vivo complementation even though they showed varying degrees of loss of in vitro relaxation (Table I).
The inability of the Glu-9, Glu-111, Asp-113, and Glu-115 mutants to complement efficiently in E. coli AS17 may be due to their low level of expression in this E. coli strain. To evaluate the in vivo activities of these mutants, the thermosensitivities of GP200 expressing these mutants were examined (Table II). Loss of topA activity in E. coli can lead to a lower rate of survival when the temperature was raised to 52°C (reviewed in Ref. 25). The rate of survival of E. coli GP200 was increased ϳ50-fold when plasmid encoded wild-type topA activity was present (Table II). The survival rates of GP200 transformed with plasmid encoding the Glu-9, Glu-111, Asp-113, and Glu-115 mutants correlated with their in vitro activity, with the Glu-111 mutant conferring close to wild-type thermoresistance and the Glu-9 mutant conferring the least amount of thermoresistance.
The Arg-136 Mutant Could Cleave DNA and Form the Covalent Complex-The cleavage activities of the mutant enzymes were examined using 5Ј-end-labeled single-stranded DNA. Even though the Arg-136 mutant was totally inactive in the relaxation assay, the amount of cleaved DNA formed was comparable with that from the wild-type enzyme (Fig. 4). For the other mutants, the amounts of cleavage products observed were decreased, with the Glu-9 mutant having the lowest cleavage activity, so that the reduction of relaxation activity seen with these other mutants might be due to the decrease in the amount of cleaved complex formed by these mutants. The Asp-111 mutant had the same DNA cleavage efficiency as the wild-type enzyme.
A 3Ј-end-labeled oligonucleotide substrate was used to form a covalent complex with the enzyme. At 5 min after the addition of the enzyme to the oligonucleotides, the amounts of the covalent complex observed for the wild-type and the Arg-136 mutant were identical (Fig. 5). The amount of covalent complex formed by the Asp-111 mutant was also close to that of the wild type while the other mutant enzymes gave lower levels of  This experiment also demonstrated that the Arg-136 mutation affected a step in the enzyme relaxation mechanism that took place after DNA cleavage. This could be the strand passage or the DNA religation step. The inter-molecular strand transfer activity of the Arg-136 mutant was compared with the wild type enzyme using the covalent complex formed with this 3Ј-end-labeled oligonucleotide substrate and HindIII-digested DNA as acceptor molecules (26). The result from one experiment is shown in Fig. 6. Data from several experiments indicated that the inter-molecular religation activity is about the same for the wild-type and the Arg-136 mutants. The efficiency of this reaction is low for both the wild-type and Arg-136 mutant enzymes. It may or may not reflect the relative activity of the intra-molecular religation that takes place during relaxation of supercoiled DNA.
Effect of the Glu-9, Asp-113, Glu-115, and R321 Mutations on Non-covalent DNA Binding-The decreased amount of cleaved complex formed by some of these mutants could be due to the effect of the mutation on non-covalent binding to DNA. The gel mobility shift assay was used to evaluate the non-covalent topoisomerase-DNA complex formation. The results (Fig. 7) showed that the Arg-321 and Glu-9 mutants had about 50% of the DNA binding activity as that of the wild-type enzyme. The Glu-115 mutant had slightly less binding activity (40%) while the Asp-113 mutant had the lowest DNA binding activity (25%).
The Glu-9 and Glu-115 Mutations Altered the Mg(II) Binding-The mutant enzymes were tested for Mg(II) binding by equilibrium dialysis against buffer containing 0.4 mM MgCl 2 . The Mg(II) binding stoichiometry was determined by inductively coupled plasma analysis (Table III). At this Mg(II) concentration, each molecule of the wild-type enzyme has been found to bind around 2 Mg(II) (27). A higher Mg(II) binding stoichiometry was observed for the Arg-136 and Arg-321 mutants. The replacement of the positively charged arginine residues by an alanine might make available an additional Mg(II) binding site in the enzyme, but this higher Mg(II) binding did not compensate for the mutation in the enzyme. The Asp-111 mutant had only a slight reduction in the amount of bound Mg(II) when compared with the wild-type enzyme. The Glu-9 and Glu-115 mutants had the greatest reduction in the binding of Mg(II).
Change in Topoisomerase I Fluorescence from Mutation-The fluorescence spectra of the wild-type and mutant topoisomerase I mutants were first compared at room temperature ( Fig. 8 and Table IV). The Glu-9 and Glu-115 mutants were found to have an ϳ30% drop in maximal fluorescence intensity, indicating a change in the protein conformation influencing the environment of the tryptophan residues in the enzyme. The decreases in fluorescence intensities were to a lesser extent for the Arg-136 and Arg-321 mutants (around 20%) while the Asp-111 and Asp-113 mutants had an ϳ25% drop in maximal fluorescence intensity. The fluorescence measurements were repeated at 42°C. The wild-type enzyme fluorescence was not affected significantly by the temperature shift. In contrast, the fluorescence intensity of the Arg-321 mutant decreased by more than 40%, indicating lower stability. The Glu-9 mutant had a much smaller decrease in fluorescence intensity at the higher temperature (Ͻ14%) while the fluorescence intensities of the other mutants did not change significantly.

DISCUSSIONS
The site-directed mutagenesis study described here aimed at elucidating the function of several strictly conserved acidic or basic amino acid residues found at the proximity of the active site nucleophile Tyr-319. For Arg-136, mutation to alanine abolished relaxation activity totally. Mutations of the other residues produced enzymes with reduced but observable in vitro activities. Therefore, the strict conservation in evolution did not necessarily correlate with absolute requirement of the residue for in vitro activity. Nevertheless, all the mutants tested failed to complement E. coli AS17 for growth at 42°C. This might at least partly be due to effect of the mutations on the stability and thus expression level of the enzyme in E. coli AS17. The Arg-136 mutant was the only one among the active site mutants tested here found to have an expression level in E. coli AS17 comparable with that of the wild-type enzyme. The  7. Effect of the mutations on the formation of non-covalent complex as assayed by the gel mobility shift assay using a 5-end-labeled 36-mer.
other mutant proteins were accumulated at a reduced or unobservable level. In contrast, mutations in the Zn(II) binding domain resulted in mutant proteins that were accumulated at higher levels than the wild-type enzyme (27). In the E. coli strain GP200 where the Glu-9, Asp-111, Asp-113, and Glu-115 mutants were stably expressed, these mutants were able to confer some degree of thermoresistance to compensate for the absence of chromosomal topA activity.
In the crystal structure of the 67-kDa N-terminal fragment, Glu-9, Asp-111, Asp-113, Glu-115, and Arg-321 are part of an extensive network of hydrogen bonds and salt bridges that are responsible for the domain-domain interactions (8). Mutation of these residues to alanines might abolish some of these interactions and destabilize the protein structure. This potential effect of the mutation on the enzyme structure and stability is consistent with the protein fluorescence data shown in Table  IV, with the structure of Arg-321 mutant particularly sensitive to the increase in temperature.
Mg(II) is required for the relaxation of DNA by E. coli DNA topoisomerase I. It is not required for the cleavage of singlestranded DNA although the rate of cleavage of small oligonucleotide substrates can be enhanced by the presence of Mg(II) (28,29). Mutations of Glu-9 and Glu-115 reduced the Mg(II) binding stoichiometry significantly after equilibrium dialysis against buffer containing 0.4 mM MgCl 2 . This might have resulted from one of the multiple coordination sites for Mg(II) in the enzyme being removed or replaced by a ligand of lower affinity to Mg(II). The effects of the mutations of Glu-9, Asp-111, Asp-113, and Glu-115 to alanine on the relaxation activity were less severe than those observed for mutations of carboxylates proposed to be Mg(II) coordination sites involved in nucleotidyl transfer in other enzyme systems (11,14,30,31). It is possible that for E. coli DNA topoisomerase I, water or a DNA phosphate could substitute for the mutated carboxylate in Mg(II) coordination in the relaxation reaction to provide partial enzymatic activity, especially at Mg(II) concentration significantly above the minimal concentrations needed for relaxation activity. The effect of mutation of Asp-111 on the enzyme activity and Mg(II) binding was significantly smaller than mutations of the other carboxylates, so it is unlikely that Asp-111 is a catalyticly important residue.
The Glu-9, Asp-113, Glu-115, and Arg-321 mutants also had reduced DNA binding activity. This could be due to the participation of the residue in direct protein-DNA interaction which is a plausible role for Arg-321. Alternatively, the reduced DNA binding could be due to the effect of the mutation on the protein folding. It also cannot be ruled out that the lower Mg(II) binding stoichiometries observed for the Glu-9 and Glu-115 mutants might be due to the effect of the mutations on the protein structure and not due to loss of a Mg(II) coordination site. A Mg(II) binding site distinct from the catalytic center has been proposed for the EcoRV restriction endonuclease (32). E. coli DNA topoisomerase I may also have a second Mg(II) binding site away from the active site region that is required for relaxation activity because of its effect on the protein conformational changes that take place during the relaxation reaction cycle (28). The carboxylates at the active site may be conserved for their roles in protein structure instead of catalytic functions. The magnitude of the effect of a single mutation of these catalytically non-essential residues on the overall protein stability may depend on the E. coli strain background, but nevertheless be of sufficient significance to account for the evolutionary conservation.
The steps of DNA cleavage and religation may or may not involve the same catalytic residues in the activation of nucleophiles as well as the stabilization of transition states and leaving groups. Presently there is a lack of data to address this question, and it remains unclear what these catalytic residues may be. Among the mutants tested, the Glu-9 mutant had the greatest loss of DNA cleavage activity with only a modest reduction in non-covalent DNA binding. Besides a possible role in binding Mg(II), this residue might be involved in the catalytic step of DNA cleavage. In contrast, the Arg-136 mutant had normal DNA cleavage activity but no relaxation activity. It might be needed in the DNA strand passage step since it could perform the inter-molecular rejoining of DNA.
There is also a lack of information on the orientation of the DNA substrate when it binds to the active site. Additional biochemical data and/or structural information of the enzyme-  a The values represent the average of three sets of measurements and are normalized relative to the maximum emission of wild-type topoisomerase I with the excitation wavelength of 280 nm at 25°C. DNA or enzyme-Mg(II) complexes would be needed to arrive at a more detailed mechanism of catalysis by the enzyme.