Identification and mutational analysis of Mg2+ binding site in EcoP15I DNA methyltransferase: involvement in target base eversion.

EcoP15I DNA methyltransferase catalyzes the transfer of the methyl group of S-adenosyl-l-methionine to the N6 position of the second adenine within the double-stranded DNA sequence 5'-CAGCAG-3'. To achieve catalysis, the enzyme requires a magnesium ion. Binding of magnesium to the enzyme induces significant conformational changes as monitored by circular dichroism spectroscopy. EcoP15I DNA methyltransferase was rapidly inactivated by micromolar concentrations of ferrous sulfate in the presence of ascorbate at pH 8.0. The inactivated enzyme was cleaved into two fragments with molecular masses of 36 and 35 kDa. Using this affinity cleavage assay, we have located the magnesium binding-like motif to amino acids 355-377 of EcoP15I DNA methyltransferase. Sequence homology comparisons between EcoP15I DNA methyltransferase and other restriction endonucleases allowed us to identify a PD(X)n(D/E)XK-like sequence as the putative magnesium ion binding site. Point mutations generated in this region were analyzed for their role in methyltransferase activity, metal coordination, and substrate binding. Although the mutant methyltransferases bind DNA and S-adenosyl-l-methionine as well as the wild-type enzyme does, they are inactive primarily because of their inability to flip the target base. Collectively, these data are consistent with the fact that acidic amino acid residues of the region 355-377 in EcoP15I DNA methyltransferase are important for the critical positioning of magnesium ions for catalysis. This is the first example of metal-dependent function of a DNA methyltransferase. These findings provide impetus for exploring the role(s) of metal ions in the structure and function of DNA methyltransferases.

sequence homology and type of methylation catalyzed (2,3). Structure-guided sequence comparison analysis revealed that all three classes of MTase are closely related to one another (4,5). EcoP15I DNA methyltransferase (M.EcoP15I) adds a methyl group to the second adenine in the recognition sequence 5Ј-CAGCAG-3Ј in the presence of AdoMet (6). It is an N 6adenine MTase, and like all N 6 -adenine MTases, M.EcoP15I contains two highly conserved sequences, FXGXG (motif I) at position 444 -448 and DPPY (motif IV) at position 125-128 (7). Although mutations in motif I completely abolished AdoMet binding but left target DNA recognition unaltered, mutations in motif IV resulted in loss of enzyme activity, but AdoMet and DNA binding were not affected (8). By employing chemical modification using thiol-directed agents and site-directed mutagenesis, it was demonstrated that cysteine 344 in M.EcoP15I was necessary for enzyme activity and played an essential role in DNA binding (9). We showed that when M.EcoP15I binds to its recognition sequence, both the adenine bases in the recognition site appear to be structurally distorted supporting the proposed base flipping mechanism for this enzyme (10).
Structural and mutational analysis of several type II restriction endonucleases reveal the sequence PD(X) n (D/E)XK motif as a catalytic/Mg 2ϩ binding signature motif (11). The charged residues from the catalytic triad are implicated in metal ionmediated catalysis (12). An inspection of amino acid sequences of DNA methyltransferases did not reveal any characteristic metal binding motif. However, in EcoP15I DNA methyltransferase, a PD(X) n (D/E)XK-like motif is present in which the partially conserved proline is replaced by methionine (7).
In this investigation, we report the identification and characterization of a magnesium binding motif in M.EcoP15I. We generated and characterized a number of mutations in the acidic residues in this region and demonstrate that several of these residues play a role in metal binding and catalysis. Our data allow us to conclude that methylation by the enzyme indeed requires magnesium and that magnesium binding to the PD(X) n (D/E)XK-like motif participates in base flipping. More importantly, our results have implications in understanding the role of metal ions in protein structure, function, and evolution of metal-independent methyltransferases.

General Recombinant Techniques
Restriction enzymes, Klenow fragment of DNA polymerase, T4 DNA ligase, and T4 polynucleotide kinase were purchased and used according to the manufacturer's recommendations. Digestions with type II restriction enzymes, ligations, transformations, and DNA electrophoresis were performed as described previously (14). Plasmid DNA pUC18 or pGEM3Zf(Ϫ) was prepared as described previously (14).

Oligonucleotides and Radiolabeling
Basic procedures of labeling of oligonucleotides were done as described previously (15). All oligonucleotides showed a purity of Ͼ95%. Concentrations of oligonucleotides were determined by UV absorbance at 260 nm using the sum of the extinction coefficients of the individual bases. The oligonucleotides used in the mutagenesis reactions are listed in Table I.

Site-directed Mutagenesis
Site-directed mutagenesis was performed to replace acidic amino acid residues at positions 355, 358, 363, 370, 372, 377, to Gly or Ala, Ala, Leu, Ala, Leu, Ala, and 599, 601, 603, 617 to Ala, respectively, using suitable primers A to K. The sequence of primer A was designed to change aspartic acid at position 355 to glycine and create a RsaI restriction site. Single-stranded DNA template containing uracil residues was prepared from E. coli strain CJ236 that harbored pGEM3Zf(Ϫ)M.EcoP15I. Primer A was hybridized to this singlestranded DNA, and oligonucleotide-directed mutagenesis was performed as described previously (16). The resultant plasmid was termed pGEM3Zf(Ϫ)M.EcoP15I-D355G. The mutants were identified by digesting the plasmid DNA with RsaI. Similarly, substitutions at other positions were made using appropriate primers (see Table I) whose sequences were designed such that a restriction site was created or eliminated (see Table I). In all cases, the mutants were identified by digesting the plasmid with the appropriate restriction enzyme. The resultant plasmids after mutagenesis are listed in Table I. Double and triple mutants were constructed using appropriate single-stranded DNA and primers (Table II). All mutants were confirmed by digesting the plasmid DNA with the appropriate restriction enzyme. DNA fragments containing the individual mutations were released from the respective pGEM3Zf(Ϫ) constructs using EcoRI and HindIII restriction enzymes. These fragments were swapped separately into the pUC18 vector containing the wild-type M.EcoP15I gene. The resultant plasmids, for instance pD355G, pD355A, pD370A, pE372L, pD601A, pD358A/E377A, and pD355A/D358A/E377A, were used for expression and purification of mutant EcoP15I-D355A, D355G, D370A, E372L, D601A, D358A/E377A, and D355A/D358A/E377A, respectively.

Overexpression and Purification of Mutant M.EcoP15Is
Wild-type and mutant M.EcoP15Is were purified according to the method described in Ref. 13 to near homogeneity. Peak fractions from the heparin-Sepharose column containing the enzyme were pooled and concentrated by using Amicon ultrafiltration unit with 30-kDa cutoff membrane. The purity of the wild-type and mutant enzymes was judged as being greater than 99% on SDS-PAGE (17). Protein concentration was estimated by the method of Bradford (18).
Inactivation and Cleavage of M.EcoP15I by Fe 2ϩ /Ascorbate 2 M M.EcoP15I was incubated with a freshly prepared solution of 0 -500 M ferrous sulfate in 20 mM Tris-HCl buffer, pH 8.0. The solution was incubated for 15 min on ice. Sodium ascorbate (final concentration 10 mM) was then added. Aliquots from the reaction mixture were assayed for residual enzyme activity at various times and were also analyzed by electrophoresis on 10% polyacrylamide gels containing 0.1% SDS, to detect the presence of fragments generated by cleavage of the enzyme. The concentration of reactants and conditions for inactivation and cleavage of enzyme were maintained the same in all experiments, unless specified otherwise. Protection against inactivation was determined by incubating the enzyme with various metal ions before the addition of iron and ascorbate.

Metal Ion Elimination
10 M M.EcoP15I was dialyzed against 100 -200 volumes of 10 mM EDTA and 20 mM Tris-HCl, pH 8.0, for 3 h at 4°C. The dialysis was then repeated for 1 h against fresh dialysis buffer. The protein solution was passed through Chelex 100 resin prepared as described by the vendor to remove any metal contamination. All buffers were analyzed by atomic absorption spectroscopy to confirm that there were no contaminating metals. For removal of the bound metal, a sample of M.EcoP15I was dialyzed sequentially versus 50 mM EDTA and then versus metal-free 20 mM Tris-HCl, pH 8.0. Metals were introduced into the apoenzyme by dialyzing a 10 M protein sample against a 1 mM solution of the selected metal ions in 20 mM Tris-HCl buffer, pH 8.0, followed by two changes of 100 volumes for removing excess metal.

Atomic Absorption Spectroscopy
All metal analyses were performed by atomic absorption spectroscopy. Each analysis was complemented by appropriate control experiments. The levels of Mg 2ϩ were determined in a 10 M solution of M.EcoP15I formulated in 20 mM Tris-HCl buffer, pH 8.0. The standard error for these determinations was Ϯ 5%. Total Mg 2ϩ concentration was measured on a Hitachi, Zeeman atomic absorption spectrophotometer. The system was calibrated using metal calibration standard solutions in the range of 0 -2 ppm. The enzyme metal concentrations were about 1 ppm. Metal determinations were performed at two dilutions, each in duplicate.

Assays for Methylation Activity
In Vitro M.EcoP15I Activity-MTase activity was monitored by incorporation of tritiated methyl groups into pUC18 DNA, and the specific activity of the enzyme was measured as described previously (8).
Sensitivity to Restriction Endonuclease-The sensitivity of the plasmid DNA containing the wild-type or mutant EcoP15I gene to restriction digestion by EcoP15I restriction enzyme (R.EcoP15I) was used to assay qualitatively for methylase activity. Plasmid DNA was isolated using the alkaline lysis method (14) and subjected to R.EcoP15I digestion as described previously (9).

Electrophoretic Mobility Shift Analysis
Binding reactions were carried out in 50 mM Tris-HCl, pH 7.5, 20 mM MgCl 2 , 7 mM 2-mercaptoethanol using duplexes I ( Table I). The mobility shift assay was performed as described previously (15). Protein-DNA complexes were visualized by exposing the dried gels to x-ray films for 12-16 h at Ϫ70°C using film cassettes containing intensifying screen.

Steady-state Fluorescence Measurements
Fluorescence emission spectra and the fluorescence intensities from titrations of M.EcoP15I were measured in a Shimadzu UV-160 recording spectrophotometer. The emission spectra were recorded over a wavelength of 300 -400 nm with an excitation wavelength of 290 nm. Slit widths of 10 nm for excitation and 10 nm for emission were used. Titrations of M.EcoP15I were carried out from 0 to 20 mM MgCl 2 or MnCl 2 or CaCl 2 . M.EcoP15I was incubated for 5 min at room temperature before the beginning of the titration. Each spectrum recorded was an average of three scans. All fluorescence measurements were made using 1-ml quartz cuvettes, with a 1-cm path length at 25°C. The concentration of the protein used was 3 M. The sample solution was mixed well and spectra recorded. At least 2-5 min time was given for stabilization of the reading. Appropriate corrections were made for dilution of the protein sample upon addition of metal ions as well as for the inner filter effect of protein and metal ions.
All samples were incubated until the equilibrium was established under the particular set of conditions before measuring the steady-state fluorescence intensities. All fluorescence emission spectra and fluores-cence intensities from titrations were corrected for protein tryptophan fluorescence by subtraction of control spectra and control titrations. In addition, fluorescence data were corrected for variable background emission of the solutions. Fluorescence intensities of all duplex oligonucleotides containing 2-AP showed a linear dependence on their concentrations. Each spectrum recorded was an average of three scans.
Steady-state fluorescence emission spectra of the 2-AP-containing oligonucleotides were measured as described previously (10). Sequences of the oligonucleotides L and M used to measure steady-state fluorescence emission spectra are given in Table I

Circular Dichroism Measurements
CD spectra were recorded on a Jasco J-500A spectropolarimeter. All experiments were done at 25°C in 20 mM potassium phosphate buffer, pH 7.0. The protein solutions were incubated on ice for 5 min in 1-mm path length quartz cells in a final volume of 400 l, and the CD spectrum was recorded. The observed ellipticities were converted to mean residue ellipticity [ MRE ], using the following equation, where obs is the observed ellipticity in degrees, mrm is the mean residue molecular mass based on a molecular mass of 150 kDa, c is the protein concentration in mg/ml, and l is the path length of the cell in cm. The protein concentration used was 5 M. The ellipticity of magnesium in this spectra was near zero. Each experimental spectrum represents the best fit at least three determinations.

Limited Proteolysis
Limited proteolysis of M.EcoP15I was performed as described previously (15). Briefly, 2 M M.EcoP15I in 20 mM Tris-HCl pH 8.0 buffer was incubated with 1% (w/w) Staphylococcus V8 protease (in 0.5 M sodium phosphate buffer, pH 7.5) at room temperature for 20 min. The digestion was stopped by adding SDS-sample buffer followed by boiling at 90°C for 5 min. The samples were analyzed on 10% polyacrylamide gels containing 0.1% SDS.

Requirement of Magnesium for DNA Methylation by
M.EcoP15I-We investigated the effect of various metal ions on the methylation activity of M.EcoP15I, to determine the specificity of metal requirement. 2 M M.EcoP15I was incubated with varying amounts of Mg 2ϩ , Mn 2ϩ , or Ca 2ϩ and the catalytic activity was determined (Fig. 1). The results of the methylation activity carried out with Mg 2ϩ , Mn 2ϩ , and Ca 2ϩ ions clearly indicate that only in the presence of increasing amounts of Mg 2ϩ was there a corresponding increase in methylation activity ( Fig. 1). Maximum activity was seen after the addition of 6 mM MgCl 2 . Further additions of up to 25 mM of MgCl 2 had no further effect on enzymatic activity.
The metal ion concentration-dependent enzymatic activity assay illustrated that Mn 2ϩ and Ca 2ϩ do not facilitate catalytic reaction at any concentration, and therefore the requirement for metal ions is specific to Mg 2ϩ (Fig. 1). The methylation activity seen in the presence of magnesium is recognition sitespecific at the EcoP15I. This was demonstrated by the inability a For details, see Table I. of M.EcoP15I-modified DNA to be cleaved by R.EcoP15I in the presence of ATP (data not shown). The stimulating influence of a magnesium ion on methylation activity may be associated with the enhanced ability of the enzyme to form a more stable complex with DNA. The lack of any activity in the presence of manganese or calcium could be attributed to the fact that the enzyme does not bind DNA in the presence of these two metal ions. However, earlier work from our laboratory clearly showed that M.EcoP15I was able to bind specifically DNA containing EcoP15I recognition sequence in the presence of manganese and calcium (15). The results presented above suggest that M.EcoP15I probably harbors a metal binding center important for methylation. Earlier, Hadi et al. (19) reported that, although R.EcoP15I comprising the restriction (Res) and modification (Mod) subunits could methylate DNA in the absence of magnesium, EcoP15I modification subunit, which acts as a DNA MTase, methylated DNA only in the presence of Mg 2ϩ . More recently, Marks et al. (20) reported that the methylation reaction by AhdI methyltransferase is dependent in the presence of Mg 2ϩ and inhibited by EDTA. Although all studied DNA MTases carry out methylation reactions in the absence of any metal ion, it is worth mentioning here that there are a few cases where the methylation reaction is stimulated in the presence of metal ions. It has been reported that the methylation reaction catalyzed by EcoB modification methylase (21), BcgI (22), MmeI (23), AloI (24) restriction enzymes and Eco57I modification methylase (25) are stimulated by metal ions. Inactivation of M.EcoP15I by Ferrous Sulfate-We intended to identify the metal binding region of M.EcoP15I using an affinity cleavage technique. Fenton chemistry uses transition metals to oxidize catalytically the peptide linkage that is in close proximity to the amino acid residues involved in metal ligation. Fe 2ϩ is used as the redox active transition metal (26). It is expected that Fe 2ϩ would bind to the methyltransferase at the Mg 2ϩ binding site.
In the presence of ascorbate, ferrous sulfate caused a rapid time-dependent inactivation of M.EcoP15I at pH 8.0 and 4°C with a half-time of about 5 min. As a control experiment, under the same conditions but without Fe 2ϩ and ascorbate added, the enzyme was found to be completely stable during the experimental period (Fig. 2). Replacing ferrous sulfate with manganese sulfate or removing ascorbate from the incubation mixture did not lead to any inactivation (data not shown). Inclusion of 10 mM magnesium chloride, which was 25 times molar in excess over ferrous sulfate, in an enzyme solution, completely protected the enzyme from Fe 2ϩ -induced inactivation (Fig. 2). Addition of 10 mM MgCl 2 to a partially inactivated enzyme solution prevented further inactivation of the remaining enzyme activity (Fig. 2).
SDS-PAGE Pattern of the Fe 2ϩ -inactivated M.EcoP15I-Successful application of Fe 2ϩ -mediated cleavage depends on specific interactions between the metal ion and M.EcoP15I protein. This approach has been used successfully in identifying the metal binding sites of XPF-ERCC 1, a structure-specific DNA repair endonuclease (27), V(D)J recombinase (28), TaqI endonuclease (29), and in a number of other type II restriction enzymes (30). We first determined the concentration of FeSO 4 required for cleavage to be 380 M (Fig. 3A) and the optimal time required for the cleavage of M.EcoP15I to be 30 min (Fig.  3B). Extensively inactivated M.EcoP15I was subjected to PAGE under reducing conditions in the presence of SDS to examine the possible oxidative cleavage of the polypeptide chain. The native enzyme has a subunit of 73 kDa (15). The inactivated enzyme was clearly cleaved into two fragments having molecular masses of 36 and 35 kDa (Fig. 3, A and B). It must be mentioned here that the putative magnesium binding site, MDX 18 ELK motif lies almost right in the middle of the protein sequence (Fig. 4). Taking this into consideration, the cleavage pattern obtained above is consistent. The 36-and the 35-kDa polypeptides represent the N-terminal half and C-terminal half of the protein. On inspection of the amino acid sequence of the N-terminal peptide, it is apparent that it contains a dipeptide, aspartate-proline, which is part of catalytic motif, DPPY (7). Formic acid is known to cleave the peptide bond specifically between the two amino acids and should yield two fragments of sizes 14 and 22 kDa. An SDS-PAGE analysis of the products of digestion of Fe 2ϩ -cleaved M.EcoP15I with 70% formic acid generated 14-, 22-, and 35-kDa products (Fig.  3C). This result clearly indicates that the original two bands of 36 and 35 kDa are products of the Fenton reaction. Inclusion of magnesium inhibited the cleavage of the protein, whereas Mn 2ϩ or Ca 2ϩ did not inhibit cleavage (Fig. 3B). The effect of the addition of the radical scavenger glycerol to the Fe 2ϩ / ascorbate-induced cleavage of M.EcoP15I was examined. As shown in Fig. 3D, glycerol and EDTA inhibited cleavage of protein and instead afforded protection from cleavage.
Identification of Presumptive Mg 2ϩ Binding Motif in M.EcoP15I by Sequence Alignment-As mentioned earlier, structural and mutational information obtained with several type II restriction endonucleases revealed a PDX 10 -30 (D/E)XK motif as a magnesium binding signature motif (11,31). Se-quence alignment of the M.EcoP15I with this motif led to the identification of the following sequence MDRLLSEEKIIFGD-DENKIIELK (MDX 18 ELK). Here, the amino acid proline is replaced by methionine. This motif is located in the middle of the protein sequence between positions 355 and 377 (Fig. 4).
A DXDXD motif has been identified previously in several EcoP15I or mutant methyltransferase D355G as described above. Fenton cleavage products were further treated with 70% formic acid as described previously (15). The Fenton-formic acid cleavage products were visualized on 10% polyacrylamide gels containing 0.1% SDS. D, effect of metal ions, radical scavengers on cleavage of M.EcoP15I. M.EcoP15I was incubated with or without 10 mM MgCl 2 or MnCl 2 or CaCl 2 or glycerol or EDTA for 5 min at 4°C. 380 M FeSO 4 was added, and the reaction mixture was incubated further at 4°C for 15 min. After 15 min, the sample mixture was incubated at room temperature for 30 min in the presence of 10 mM sodium ascorbate. The samples were visualized on 0.1% SDS and 10% polyacrylamide gel. MWM denotes molecular mass markers. The arrowheads indicate the cleavage products. calcium-binding proteins including calmodulin (32). Both structural and mutational studies of several calcium-binding proteins have clearly revealed that the acidic amino acid residues in this motif are known to coordinate calcium and thus play a very important role in catalysis. It is also known that in many calcium-binding proteins, the DXDXD motif can bind metals like magnesium, in addition to calcium (33).
A close look at the M.EcoP15I amino acid sequence revealed a DXDXD motif in the C-terminal end between positions 599 and 603 (7) (Fig. 4). It was therefore of interest to know whether the conserved acidic amino acid residues in either the MD(X) 18 ELK or DXDXD motifs are responsible for coordinating magnesium and, if so, to assign a possible role for these amino acids in catalysis. As a first step in this direction, we have focused on the conserved acidic amino acid residues likely to be critical for metal binding and enzyme activity and used site-directed mutagenesis to make mutant proteins with defined substitutions at these positions. Activities of Mutant M.EcoP15I-To assess the modification phenotype of the M.EcoP15I mutants, we carried out in vitro restriction assay. Plasmid DNA from cells expressing the wildtype or mutant MTases were isolated, and their sensitivity to digestion by the cognate R.EcoP15I was determined. Plasmids harboring an active methylase would result in methylation of all M.EcoP15I sites in vivo and thereby be resistant to subsequent in vitro cleavage by R.EcoP15I. It is clear from Fig. 5A that the plasmid DNAs carrying the mutant M.EcoP15I-D355G, D370A, and E372L were not protected from cleavage by R.EcoP15I, indicating these to be inactive methylases. Plasmid DNAs carrying the mutant M.EcoP15I-D355A, D358A, E363L, and E377A were protected from cleavage by R.EcoP15I, indicating these methylases to be functionally active (Fig. 5A). Similarly, double mutant MTases containing the following combinations D355A/D358A, D355A/E363L, D355A/E377A, D358A/E363L were all active methyltransferases. Interestingly, only the plasmid DNA carrying the mutant D358A/ E377A was not protected from R.EcoP15I cleavage (Fig. 5B). Also, as shown in the Fig. 5B, among the triple mutants, only plasmid DNAs carrying D355A/D358A/E377A and D358A/ E363L/E377A were not protected from R.EcoP15I cleavage, suggesting that these were inactive methylases. It should be noted that single substitution at positions 358, 363, and 377 produced mutant MTases, which were all catalytically active. By making site-directed mutations at these positions and analyzing the ability of mutant proteins to promote DNA methylation, it is clear that any plasmid DNA that encoded mutant mod genes in which the aspartic acid at positions 355, 358, 370, and glutamic acid at position 372 and E377 were changed singly or in combination, they were susceptible to cleavage by R.EcoP15I. These results also clearly point out the importance of these acidic amino acid residues for methylation activity.
The aspartic acid residues at positions 599, 601, and 603 and the glutamic acid residue at position 617 were replaced by alanine using the site-directed mutagenesis protocol as described under "Experimental Procedures." Double mutants D599A/D601A, D599A/D603A, D599A/E617A, D601A/D603A, D601A/E617A, and D603A/E617A were also constructed as described above (Fig. 5C). Plasmid DNAs from cells expressing the above single (Fig. 5C) or double mutant enzymes were all resistant to R.EcoP15I digestion, suggesting that all of these mutant methyltransferases were active and that the amino acids at these positions do not have any significant role in methylation activity.
Ca 2ϩ versus Mg 2ϩ -Mutational analysis of the DXDXD region clearly indicated that the aspartic acid residues in the C-terminal region of M.EcoP15I did not have a role in methylation activity. It is noteworthy to mention that several Ca 2ϩbinding proteins usually contain a mixture of calcium and magnesium binding sites, and there is pronounced magnesium antagonism (34). To determine whether M.EcoP15I can interact with Ca 2ϩ despite its failure to support DNA methylation, the rate of methylation was measured in buffer containing both MgCl 2 and CaCl 2 . First, M.EcoP15I was preincubated with 10 mM CaCl 2 for 10 min at 4°C, and varying amounts of magnesium chloride (0 -25 mM) were added. Methylation activity was performed as described earlier. The result of such a preincubation experiment clearly showed that enzyme activity increased as magnesium concentration increased, and maximal activity was observed when the magnesium concentration was 6 mM (Fig. 6). Clearly, preincubation with Ca 2ϩ had no effect on methylation activity. Next, the enzyme was preincubated with 10 mM MgCl 2 and the methylation activity was determined as a function of increasing amounts of calcium chloride (0 -25 mM). As can be seen from Fig. 6, the methylation activity neither increased nor decreased, indicating that Ca 2ϩ had no significant effect on the magnesium-dependent methylation activity catalyzed by M.EcoP15I. The simplest interpretation of these observations is that M.EcoP15I does not bind calcium, and therefore magnesium stimulated methylation activity is not inhibited in the presence of calcium. Calcium often acts as a nonreactive analog of Mg 2ϩ . It has been used extensively as a Mg 2ϩ surrogate. With orthodox type II restriction endonucleases, Ca 2ϩ is observed to increase the affinity for cognate DNA sequences by factors of up to 10 Ϫ4 (35). It has been shown recently that only in the presence of Ca 2ϩ , the dimeric FokI complex is formed (36). Similarly, it was demonstrated that (MboII) 2 -(DNA) 2 species is formed only in the presence of Ca 2ϩ (37). In this context, it must be pointed out that the "final" proof for a PD(X) n (D/E)XK motif involved in catalysis requires a crystal structure analysis of an enzyme-substrate complex in conjunction with mutational analysis and mechanistic studies.
Characterization of Mutant M.EcoP15I-Several of the mutant methyltransferases (D355G, D355A, D358A, D370A, E372L, E377A, D601A, D358A/E377A, and D355A/D358A/ E377A) were purified as described under "Experimental Procedures." The purified enzymes were analyzed, using SDS-PAGE and Western blotting, for alterations in the electrophoretic mobilities, and no apparent changes were detected vis-à -vis the wild-type protein (data not shown). Methylation activity of the wild-type and mutant MTases was also determined by measuring their ability to transfer the 3 H-labeled methyl group from AdoMet to pUC18 DNA. In vitro methylation activity measurements clearly showed that purified preparations of D355A and D358A were as active as the wild-type enzyme, whereas D355G, D370A, and E372L were inactive (data not shown). Increasing the amounts of these enzymes (up to 10 g) in standard assay conditions led to no enhancement in methylation activity, clearly suggesting that these residues do play crucial role in catalysis. It is interesting to note that when aspartic acid at position 355 was replaced by alanine (D355A), the methylation activity was found to be similar to that of wild-type M.EcoP15I, whereas replacing the aspartic with glycine resulted in complete loss of activity. We surmise that D355G does not interact with metal ions, as does D355A substitution. The double mutant D358A/E377A and the triple mutant D355A/D358A/E377A enzymes showed negligible methyl- ation activity, consistent with results described earlier (data not shown).
Atomic Absorption Spectroscopic Analysis of Wild-type and Mutant M.EcoP15Is-To detect metal ions in wild-type or mutant M.EcoP15I, atomic absorption spectroscopy was employed. We first showed that purified wild-type or mutant D355G, D370A, and E372L M.EcoP15I were metal-free. We next tested whether the wild-type or mutant M.EcoP15I binds to metal ions by providing exogenous metal ions. Toward this, wild-type M.EcoP15I and mutant enzymes (10 M) were incubated in the presence of 1 mM MgCl 2 or CaCl 2 and dialyzed extensively. Although the wild-type enzyme contained 1 mol of Mg 2ϩ /mol of protein, D355G, D370A, and E372L mutant enzymes showed no magnesium binding. The addition of 3-30 eq of Ca 2ϩ to M.EcoP15I followed by extensive washing with metal-free 20 mM Tris-HCl buffer, pH 8.0, at 4°C did not contain any bound calcium ions. None of the mutant MTases bound calcium. These observations quite convincingly demonstrate that substitutions at positions 355, 370, and 372 with glycine, alanine, or leucine do indeed interfere with Mg 2ϩ binding. The inactivity of D355G, D370A, and E372L is primarily the result of an inability of metal binding needed for catalysis. Even though this study focuses on ionizable Mg 2ϩ ligands, it is possible that there are other contributors to metal ion affinity. Water-mediated contacts are very common and also cannot be ruled out (38). We tested two other mutant methylases M.EcoP15I-D355A and M.EcoP15I-D601A for magnesium binding by atomic absorption spectroscopy. As mentioned earlier, the D355A mutant methyltransferase is catalytically active. D601A mutant methyltransferase, which harbors a mutation in the DXDXD region of M.EcoP15I, is also catalytically active. Both of these catalytically active mutant MTases are able to bind magnesium (1 mol of Mg 2ϩ /mol of protein) but not calcium.
Fe 2ϩ /Ascorbate treatment of D355G Mutant of M.EcoP15I-M.EcoP15I was inactivated and specifically cleaved (Figs. 2 and 3) when incubated with ferrous sulfate and ascorbate. Aspartic acid at position 355 was identified as one of the acidic amino acids, which probably coordinates a magnesium ion. Mutating the aspartic acid residue at position 355 of M.EcoP15I to glycine inactivates the enzyme (Fig. 5A, compare  fifth and sixth lanes). The D355G mutant of M.EcoP15I was subjected to Fe 2ϩ /ascorbate treatment. Fig. 3C shows the SDS-PAGE pattern of the wild-type, wild-type following treatment with Fe 2ϩ and ascorbate, M.EcoP15I with D355 mutation, and the mutant M.EcoP15I following treatment with Fe 2ϩ and ascorbate, respectively. Distinct cleavage products are observed for wild-type M.EcoP15I as described earlier. However, no specific cleavage products were observed for the D355G mutant of M.EcoP15I. This result is consistent with the observation that the aspartic acid residue at position 355 of M.EcoP15I is involved in coordination of the divalent metal.
CD Analysis of Wild-type and Mutant M.EcoP15Is-The loss of catalytic activity of the D355G, D370A, and E372L variants could be the result of an altered protein conformation rather than the absence of a functional group that participates in catalysis. To check for this possibility, CD spectroscopy, which is extremely sensitive to protein secondary structure, was used. Spectra were collected from 200 to 250 nm at 25°C for wildtype and the mutant variants.
When wild-type M.EcoP15I was titrated with increasing concentrations of magnesium, the CD spectrum changed significantly (Fig. 7A), indicating that the secondary structure of protein did alter in the presence of magnesium. No such changes were observed when magnesium was replaced by manganese or calcium (data not shown). More importantly, in the absence of any metal ion, there was no apparent difference in the CD spectra of the wild-type and D355G (Fig. 7B), D370A (Fig. 7C) and E372L (data not shown) mutant MTases. Thus, the substitutions of any of the three amino acid residues did not affect the overall secondary structure of M.EcoP15I per se. It is also evident that the CD spectra of the mutant proteins did not alter significantly in the presence of magnesium, substantiating the earlier observations that these mutant proteins are not capable of binding magnesium. Because we detected no major conformational changes from substitutions, it is likely that the changes observed resulted from a direct effect of the substitution on the active site rather than from some more generalized change in MTase structure.
Fluorescence Spectroscopy-Fluorescence of tryptophan residues is highly dependent on their local environment and enzyme conformation. Titration of M.EcoP15I with increasing concentrations of MgCl 2 was carried out while monitoring the fluorescence emission spectrum of the enzyme with excitation at 295 nm. At this excitation wavelength, the fluorescence signal should be caused by the six tryptophan residues in M.EcoP15I. An increase in the overall fluorescence emission intensity was observed with increasing concentrations of magnesium, reaching a maximum at 20 mM (Fig. 8A). The effect of magnesium chloride on the fluorescence properties of M.EcoP15I was totally reversed by the addition of 10 mM EDTA (data not shown). The addition of 15 mM MnCl 2 or CaCl 2 had no effect on the protein fluorescence emission (data not shown). This indicated that the changes in fluorescence observed with the addition of Mg 2ϩ was not caused by a change in the ionic strength. Thus, we demonstrate here that significant changes in enzyme structure could be observed in the presence of 20 mM MgCl 2 , affecting the tryptophan emission intensity 340 nm. This suggests that magnesium binding may allow the enzyme to assume a conformation necessary for DNA methylation to be catalyzed. There were no significant changes in fluorescence properties of D355G (Fig. 8B), D370A (Fig. 8C), and E372L (data not shown) EcoP15I mutant enzymes with the addition of magnesium chloride. We hypothesize that M.EcoP15I exists in at least two conformations, an open conformation in the absence of magnesium and closed or more compact conformation upon the addition of magnesium ions.
Limited Proteolysis-Staphylococcus V8 protease has been used to generate limited peptide fragments of protein. V8 protease cleaves on the carboxyl side of glutamic acid and is predicted to generate more than 50 fragments on complete digestion of M.EcoP15I. To determine structural changes, if any, we subjected wild-type M.EcoP15I and mutant proteins to limited V8 proteolytic digestion and analyzed the digestion products by electrophoresis. A V8 protease (1%) sensitivity assay was performed to monitor conformational changes upon binding of MgCl 2 . In the absence of MgCl 2 , the digestion of V8 protease in wild-type M.EcoP15I results in accumulation of three clear bands of sizes approximately within molecular mass ranges of 50, 32, and 30 kDa (Fig. 9A). On the other hand, with digestion in the presence of the Mg 2ϩ divalent cation, there was full protection against degradation, implying that Mg 2ϩ stabilizes the enzyme against proteolysis (Fig. 9A). Similarly, a V8 protease experiment was done in the presence of Ca 2ϩ and Mn 2ϩ metal ions. The digestion pattern was found to be similar to that seen in the absence of the metal ion (Fig. 9A). All three mutant proteins D355G, D370A, and E372L were proteolyzed to the same extent in the absence or presence of magnesium, indicating that the activity of V8 protease is unaffected by the addition of divalent cation (Fig. 9B). These results are consistent with our CD and fluorescence spectroscopic analysis in that the binding of Mg 2ϩ to M.EcoP15I brings about conformational changes.
DNA Binding Properties of Mutant M.EcoP15Is-The basis for the requirement of Mg 2ϩ for methylation activity needs to be elucidated to understand fully the enzyme mechanism. There are several possible roles for Mg 2ϩ in the methylation of DNA by M.EcoP15I. Because the mutant MTases were not functional in DNA methylation, one of the possible reasons could be the loss of DNA or AdoMet binding ability. To test this hypothesis, DNA binding properties of mutant MTases were investigated using an electrophoretic mobility shift assay. Duplex I, which is a 31-mer oligonucleotide containing the EcoP15I recognition sequence 5Ј-CAGCAG-3Ј, was used in an electrophoretic mobility shift assay to monitor specific binding. The D355G, D370A, and E372L mutant MTases (2 M) bind to 100 nM DNA almost with the same affinity as wild-type enzyme, thus demonstrating that the catalytic inactivity is not caused by their inability to bind DNA.
Photo Cross-linking of Mutant M.EcoP15Is-UV cross-linking experiments were carried out in an attempt to correlate the lack of MTase activity with an inability to bind AdoMet. Exposure of wild-type M.EcoP15I and M.EcoP15I-D355A, D355G, D370A, E372L, and D601A MTases to UV light in the presence of [methyl-3 H]AdoMet resulted in strong photolabeling with the same yield as the wild-type enzyme (data not shown). It is therefore very unlikely that the amino acid substitutions had affected the ability of these mutants to cross-link AdoMet with a similar, if not the same, conformation as the wild-type M.EcoP15I enzyme. Fluorescence Changes Induced by Wild-type and Mutant M.EcoP15Is upon Binding to 2-AP-containing Duplexes-Base flipping is involved in the catalytic mechanism of various enzymes interacting with DNA (39). Base flipping generally pulls the base out of the DNA helix and results in large changes in the environment of the flipped target. As reported earlier (10), M.EcoP15I binds to its recognition sequence 5Ј-CAGCAG-3Ј, and in the process of methylating the second adenine residue, there are drastic conformational changes seen within the recognition sequence. Using a fluorescence-based assay, we showed that the target base is indeed in a different environment when the enzyme binds. Most likely, the target base is flipped out (10). Potassium permanganate footprinting also lent support to base flipping by M.EcoP15I (10). So far, there are no reports regarding DNA MTases using metal ions in the base flipping process. Hence, we extended our studies to test the effect of metal ions on base flipping, if any. We used DNA containing the EcoP15I recognition sequence CAGCAG, but the second adenine (target base) was replaced by a fluorescent adenine analog, 2-AP. This probe was used to monitor temporal changes in 2-AP fluorescence in the absence or presence of magnesium ions. It must be mentioned that M.EcoP15I shows significant enhancements in 2-AP fluorescence only in the presence of AdoMet or S-adenosyl-L-homocysteine or sinefungin (10). As shown in the Fig. 10, M.EcoP15I enzyme does not show significant enhancement in 2-AP fluorescence in the absence of magnesium. To test further whether 2-AP unstacking is linked directly with magnesium concentrations, 2-AP fluorescence change over a range of MgCl 2 concentrations from 1 to 20 mM was studied. Throughout this range, the fluorescence increases with the MgCl 2 concentration (Fig. 10, curves a-d). Magnesium alone, when added to duplex DNA containing 2-AP, had no effect on the fluorescence intensity (data not shown). Encouraged by these results, we attempted to monitor the effect of magnesium ions on 2-AP fluorescence by mutant M.EcoP15Is. To our surprise, the mutant enzymes D355G (Fig. 11A), D370A (Fig. 11B), and D358A/E377A (Fig. 11C) showed no enhancement in 2-AP fluorescence upon addition of Mg 2ϩ . The simplest interpretation of these results is that the acidic amino acid residues in conjunction with Mg 2ϩ coordinated to these amino acids do stabilize the extra helical adenine base. To the best of our knowledge, this is the first evidence that metal ion binding can provide a driving force to stabilize base stacking interactions. The D601A mutant enzyme, which is catalytically active, upon addition of Mg 2ϩ showed almost the same 2-AP fluorescence as the wild-type M.EcoP15I enzyme (Fig. 11D). We had shown earlier (10) that on binding DNA, which contains 2-AP at the first adenine base in the recognition sequence, an increase in fluorescence was observed. The mutant enzymes, D355G, D370A, and D358A/E377A and E372L did not show any increase in fluorescence upon binding DNA (data not shown). These observations are consistent with the findings reported above that the interaction of acidic amino acids with magnesium ions is required for stabilization of the extra helical adenine base, which is probably flipped out during the methylation reaction. Collectively, these findings are thus key to understanding how catalytically inactive M.EcoP15I mutants are unable to position reaction components so that they are poised for specific chemistry. Our studies demonstrate that Mg 2ϩ is an essential metal ion in promoting M.EcoP15I methylation activity.
In view of similar mechanism of methylation by DNA MTases, it is interesting to note that methylation by some MTases either requires or is stimulated by Mg 2ϩ (21)(22)(23)(24)(25). This study clearly demonstrates that Mg 2ϩ is required for base flipping by M.EcoP15I and suggests that this may be true for other MTases that show metal ion dependence. Several of the MTases where such an effect is seen belong to types I, IIB, IIG, IIS, or type III R-M systems (40). This is in contrast to the DNA MTases belonging to the orthodox type II R-M enzymes, which do not show any requirement of metal ion for methylating DNA. Considering that orthodox type II R-M enzymes are more efficient and probably arose later in evolution than the other types of R-M systems (41), one can possibly speculate that Mg 2ϩ -independent MTases evolved from Mg 2ϩ -dependent MTases. However, detailed structural comparisons of the active sites of Mg 2ϩ -dependent and -independent MTases could provide insights into understanding how side chains of amino acids can compensate for the absence of Mg 2ϩ . Interestingly, the recent identification of a metal-independent restriction endonuclease lends support to our understanding of how amino acids compensate for metal ions at the active site (42). Based on such analysis, we have initiated studies to design an Mg 2ϩindependent M.EcoP15I.