The acidic triad conserved in type IA DNA topoisomerases is required for binding of Mg(II) and subsequent conformational change.

The acidic residues Asp-111, Asp-113, and Glu-115 of Escherichia coli DNA topoisomerase I are located near the active site Tyr-319 and are conserved in type IA topoisomerase sequences with counterparts in type IIA DNA topoisomerases. Their exact functional roles in catalysis have not been clearly defined. Mutant enzymes with two or more of these residues converted to alanines were found to have >90% loss of activity in the relaxation assay with 6 mM Mg(II) present. Mg(II) concentrations (15-20 mM) inhibitory for the wild type enzyme are needed by these double mutants for maximal relaxation activity. The triple mutant D111A/D113A/E115A had no detectable relaxation activity. Mg(II) binding to wild type enzyme resulted in an altered conformation detectable by Glu-C proteolytic digestion. This conformational change was not observed for the triple mutant or for the double mutant D111A/D113A. Direct measurement of Mg(II) bound showed the loss of 1-2 Mg(II) ions for each enzyme molecule due to the mutations. These results demonstrate a functional role for these acidic residues in the binding of Mg(II) to induce the conformational change required for the relaxation of supercoiled DNA by the enzyme.

Escherichia coli DNA topoisomerase I is the best studied representative of the type IA DNA topoisomerases. This class of enzymes includes the bacterial and archeal DNA topoisomerase I and III, reverse gyrase, and yeast and mammalian topoisomerase III, with diverse roles in cellular functions (reviewed in Refs. 1 and 2). Mg(II) is required for the interconversion of DNA topological isomers catalyzed by these enzymes. Comparison of their polypeptide sequences showed that the conserved positions include the acidic residues Asp-111, Asp-113, and Glu-115 (3,4). When the crystal structure of the 67-kDa Nterminal transesterification domain of the enzyme was published, it was noted (5) that these three acidic residues in the active site are arranged similarly to the three acidic residues known to coordinate two divalent ions in Klenow fragment (6) that are required for the nucleotidyl transfer activity (7,8). These residues are found in domain I of the 67-kDa structure (5), which is similar to the BЈ domain of the Saccharomyces cerevisiae DNA topoisomerase II structure (9). There are cor-responding acidic residues that are conserved in type IIA DNA topoisomerases (9). Severe loss of DNA relaxation and cleavage activities resulted when one of these acidic triad residues in S. cerevisiae DNA topoisomerase II, Asp-530, was mutated (10). Another conserved glutamate at Glu-9 of E. coli DNA topoisomerase I and the aspartates motif DXD at Asp-111 and Asp-113 have been proposed to be conserved motifs in a catalytic domain named Toprim found in type IA and IIA topoisomerases, as well as a number of other nucleotidyl transferases and polynucleotide cleaving activities (11). However, results of site-directed mutagenesis in E. coli DNA topoisomerase I showed that conversion of a single one of these three conserved acidic residues to alanine did not abolish the relaxation activity (3,12). The exact role of Asp-530 in the catalytic mechanism of yeast topoisomerase II is also not entirely clear (10). Whether these three conserved acidic residues or the DXD motif participate in coordination of divalent ions by the enzyme-DNA complex during removal of negative supercoils remained to be established by biochemical analysis. It is possible that the loss of a single Mg(II) coordinating residue in the active site can be tolerated by the enzyme under reaction assay conditions with Ͼ2 mM Mg(II) present (3,12). Double and triple mutants lacking two or more of these acidic residues have now been purified for analysis of their enzymatic activities. The results presented here showed that relaxation activity for double mutants involving Asp-111, Asp-113, and Glu-115 requires higher concentrations of Mg(II) for maximal activity, supporting the hypothesis that the carboxylates in these three conserved acidic residues are involved in coordinating at least one of the two Mg(II) required for relaxation activity (13). We have also developed a proteolytic digestion assay using Glu-C endoproteinase to detect the Mg(II)-induced conformational change in E. coli DNA topoisomerase I.

EXPERIMENTAL PROCEDURES
Materials-The chemical reagents used were either ultrapure or ACS 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 contaminating metal ions. Plasmid DNA was purified by cesium chloride gradient centrifugation.
Mutagenesis-The QuikChange site-directed mutagenesis kit from Stratagene was used for construction of the mutants. The mutants were identified by DNA sequencing.
Enzyme Expression and Purification-Wild type topoisomerase I was expressed and purified as described (14). The double and triple mutants studied here were expressed in E. coli GP200 (gyrA(Nal r )gyrB225-⌬(topAcysB)204) (15) and purified with procedures similar to those used for the wild type enzyme.
Relaxation Activity Assay-Wild type and mutant enzymes were 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 the indicated concentration of MgCl 2 . 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 * This work was supported by Grant GM54226 from NIGMS, National Institutes of Health, United States 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  Intrinsic Tryptophan Fluorescence Measurements-Fluorescence measurements were performed with the Perkin-Elmer LS-5B luminescence spectrometer at room temperature. The spectral bandwidths were 5 and 10 nm, respectively for excitation (295 nm) and emission. Each data point reported was an average of three determinations. Enzyme was present at 1 M in 20 mM potassium phosphate, pH 7.4, 0.1 M KCl. All the measurements were corrected for the spectrum of the buffer solution used. To calculate the dissociation constant for Mg(II) binding, the data were analyzed as described previously (16).
Glu-C Proteolytic Digestion-Wild type or mutant topoisomerase I was dialyzed into 20 mM potassium phosphate, pH 7.5, 0.1 M KCl, 0.2 mM dithiothreitol with the indicated concentration of MgCl 2 present. Glu-C sequencing grade endoproteinase (from Roche Molecular Biochemicals) was added at a ratio of 1:50 (w/w) to 10 g of topoisomerase. Digestion was carried out at 37°C for the indicated length of time and stopped by the addition of equal volume of 2ϫ gel loading buffer for the Laemmli SDS gel (17). The samples were analyzed by electrophoresis in either a 15 or 5-20% gradient SDS-polyacrylamide gel followed by staining with Coomassie Blue. Comparison of intensities of the 14-kDa digestion product was carried out using the AlphaImager software.
Identification of the 14-kDa Glu-C Digestion Product-Mass spectroscopy was used to determine the exact molecular mass of the 14-kDa Glu-C digestion product. In-gel digestion with Lys-C endoproteinase and mass mapping were performed by Dr. V. A. Fried at the protein structure facility at New York Medical College. Mass spectroscopy was performed by matrix-assisted laser desorption ionization time-of-flight mass spectrometry (Kompact III, Kratos).

Expression and Purification of the Double and Triple Mutants-Expression of the alanine substitution mutants in E.
coli GP200 was confirmed by SDS gel electrophoresis of the soluble lysates followed by Coomassie Blue staining (data not shown). The enzymes were purified to Ͼ95% homogeneity ( Fig.  1) by the combination of phosphocellulose, hydroxyapatite, and single-stranded DNA agarose column chromatography.
Effect of the Double and Triple Mutations on Relaxation Activity with 6 mM Mg(II) Present-The purified double and triple mutants with the conserved acidic triad in the active site altered by alanine substitutions were diluted serially and assayed for relaxation activity in the presence of 6 mM Mg(II). The results (Fig. 2) showed that their relaxation activities were further diminished under this assay condition when compared with the activities reported for the single mutants (3,12). No activity was detected for the D111A/D113A double mutant and the D111A/D113A/E115A triple mutant. The D111A/E115A and D113A/E115A double mutants had Ͻ10% wild type activity.
The mutant with alanine substitution at Glu-9 has been shown to have Ͼ90% loss of relaxation activity (3,11). Double mutants with alanine substitutions at Glu-9 and one of the acidic triad had no detectable relaxation activity at 6 mM MgCl 2 (Fig. 2).
Double Mutants Involving the Acidic Triad Require Higher Mg(II) Concentration for Maximal Relaxation Activity-The effect of increasing Mg(II) concentration on the relaxation activity was compared between the wild type enzyme and the double and triple alanine substitution mutants. Maximal relaxation activity for the double mutants involving the acidic triad was observed at 15 mM MgCl 2 (Fig. 3). With wild type enzyme, such high concentration of MgCl 2 resulted in inhibition of relaxation activity. Previous studies showed that as little as 1.5-2.5 mM was sufficient for maximal activity for the wild type topoisomerase I (13,18). Relaxation activity could not be restored for the D111A/D113A/E115A triple mutant even at 20 mM MgCl 2 , and only partial relaxation of the input DNA can be achieved with the D111A/D113A mutant. The increased Mg(II) concentration had much less of an effect on the relaxation activity of the double mutants involving alanine substitution at Glu-9 and one of the acidic triad residues (Fig. 3). There is a slight enhancement of the relaxation activity of the E9A/D111A and E9A/D113A mutants at 15 mM MgCl 2 . The E9A/E115A mutant remained inactive.
To further compare the activity of the double mutants involving the acidic triad at the high Mg(II) concentration of 15 mM versus the wild type activity at 6 mM Mg(II), time course of the relaxation reaction was monitored. The results (Fig. 4) showed that 30 min were required for the DNA incubated with the double mutants to approach maximal relaxation, whereas the wild type enzyme reaction was nearly complete at 10 min.
Reduced Mg(II) Binding Stoichiometries for the Acidic Triad Double and Triple Mutants-We have previously shown that following incubation of the wild type topoisomerase I with 2 mM MgCl 2 , around 2 Mg(II) remains bound to each enzyme molecule after dialysis (13). Similar treatment of the acidic triad double and triple mutants followed by inductively coupled plasma analysis of the Mg(II) content (Table I) showed that their Mg(II) binding stoichiometries were reduced, in correlation with the relaxation activities observed. The D111A/D113A/ E115A mutant, which did not show any activity under all conditions tested, had a Mg(II) binding stoichiometry barely above the background. The D111A/E115A and D113A/E115A mutants retained binding of around 1 Mg(II) per enzyme molecule, whereas the double mutant D111A/D113A, with the more severely reduced activity, had significantly lower Mg(II) binding stoichiometry.

Mg(II) Binding Followed by Change in Fluorescence-
We have previously shown that Mg(II) binding result in a conformational change in topoisomerase I detectable by decreased tryptophan fluorescence (13). Analysis of the fluorescence data (Fig. 5A) as described previously (16)  Mg(II) Binding to Topoisomerase I Results in Change in Glu-C Digestion Pattern-Besides monitoring the change in tryptophan fluorescence, we have also developed a proteolytic assay using Glu-C digestion to detect the conformational change in topoisomerase I upon binding of Mg(II). As shown in Fig. 6, the presence of as little as 0.5 mM Mg(II) topoisomerase I resulted in accumulation of a 14-kDa proteolytic fragment not seen in the time course of digestion in the absence of Mg(II). The yield of this 14-kDa product was further enhanced when Mg(II) concentration was increased to 2 mM. The 14-kDa product was not seen when the D111A/D113A and D111A/D113A/E115A mutants were digested with Glu-C in the presence of 2 mM Mg(II) (Fig. 7). For the D111A/E115A and D113A/E115A mutants, the level of the 14-kDa product formed was about 3-fold lower than that from the wild type enzyme.
Identification of the 14-kDa Glu-C Digestion Product-The exact weight of the 14-kDa Glu-C digestion product was determined by mass spectroscopy to be 14,498. This is almost identical to the mass predicted for the Glu-C digestion product from residue 736 to 862 (14,509.5). The masses of the in-gel Lys-C digestion products of this 14 kDa fragment (Table II) confirmed this identification. DISCUSSION Type IA and type II DNA topoisomerases share similarity in their requirement for Mg(II) in the enzyme catalytic mechanism. When the sequences of these two classes of enzymes are compared, acidic residues corresponding to Glu-9, Asp-111, Asp-113, and Glu-115 of E. coli DNA topoisomerase I are found to be conserved (3,4). A role for Glu-9 in the breaking and rejoining of DNA strand has been proposed (3,12). However, conversion of one of the acidic triad Asp-111, Asp-113, or Glu-115 to alanine only resulted in partial loss of enzyme activity, with the mutation of Asp-111 having the least effect (3,12). The results described here with the double mutants provided support for the hypothesis that Asp-111, Asp-113, and Glu-115 bind a divalent ion during catalysis (5). Previous studies have shown a positive relationship between the relaxation activity of E. coli topoisomerase I and the concentration of Mg(II) from 0 to 10 mM, with near saturation of the metal requirement achieved at 2.5 mM (18). With double alanine substitutions at the acidic triad, the Mg(II) concentrations required for maximal relaxation activity are significantly higher, to a point that is inhibitory for the wild type enzyme activity. In the presence of DNA phosphates at the active site, the loss of a single Mg(II)   Double mutants involving Glu-9 and one of the acidic triads have no detectable relaxation activity at 6 mM Mg(II). Increased Mg(II) concentration had only a slight effect on the relaxation activity. This supports the interpretation that even though lower Mg(II) binding affinity may contribute to the decreased activity seen for the double mutants involving Glu-9, the major reason for activity deficiency is likely to be due to the effect of the Glu-9 mutation related to DNA cleavage/religation (3,12).
The Glu-C digestion patterns of all of the mutants examined here are quite similar in the absence of Mg(II). Therefore, the alanine substitutions did not lead to a severe change of enzyme folding that may account for the loss of activity. Addition of Mg(II) did not lead to the appearance of a 14-kDa fragment in the Glu-C digestion of the D111A/D113A and D111A/D113A/ E115A mutants This supports the role of the acidic triads in binding to Mg(II) and resulting in a conformational change in the enzyme.
The effect of double alanine substitutions at Asp-111 and Asp-113 was more severe in both the relaxation assay and the Glu-C digestion assay than the double alanine substitutions at one of these residues and Glu-115. In contrast, the effect of the single alanine substitution on relaxation activity was more severe at Glu-115 than at Asp-111 and Asp-113 (3). This indicates that Asp-111 and Asp-113 may have overlapping roles in Mg(II) binding, but elimination of both functional groups would greatly diminish Mg(II) binding. Mutation at Glu-115 may also influence another aspect of enzyme function, aside from binding of Mg(II) and the resulting conformational change.
The 14-kDa Glu-C digestion product that is preferentially formed in the presence of Mg(II) corresponds to a cleavage site after Glu-735. The region from amino acid 598 to the C terminus of the protein contains the three Zn(II) binding tetracysteine motifs (19), and another 14-kDa fragment that can be formed by tryptic cleavage of topoisomerase I at Arg-744 (20). The 14-kDa tryptic fragment has single-strand DNA binding activity on its own (20,21), but the affinity is greatly enhanced when it is linked to the Zn(II) binding motifs (22). The Zn(II) motifs needed for the relaxation of supercoiled DNA are thus part of a high affinity DNA binding domain. We denote this region from amino acid 598 to the C terminus as the ZD domain. The 67-kDa transesterification domain terminating at amino acid 597 can cleave DNA but has no relaxation activity (23). We postulate that in the absence of Mg(II), the Glu-C cleavage site at 735 is not accessible because of protein-protein interactions between the 67-kDa transesterification domain and the Zn(II) binding motifs. The last of the 12 Zn(II) coordinating cysteines can be found at position 736. In the presence of Mg(II), protein conformational change involving separation of the transesterification domain and ZD domain takes place to expose the Glu-C cleavage site at Glu-735, forming the observed 14-kDa product from residue 736 -862 that is resistant to further Glu-C digestion. Biochemical analysis of the role of the ZD domain in the relaxation reaction 1 is consistent with this interpretation.
The results presented here established a functional role for the acidic triad conserved in type IA and type IIA topoisomerases. The DXD motif residues Asp-111 and Asp-113 are clearly involved in binding of Mg(II) and the resulting enzyme conformational change. Besides additional structural information on the enzyme, identification of other amino acid residues needed for this Mg(II)-induced conformational change would be helpful for elucidation of the dynamic action of the topoisomerase activity.