Structure-guided Mutational Analysis of the Nucleotidyltransferase Domain of Escherichia coli NAD+-dependent DNA Ligase (LigA)*

NAD+-dependent DNA ligase (LigA) is essential for bacterial growth and a potential target for antimicrobial drug discovery. Here we queried the role of 14 conserved amino acids of Escherichia coli LigA by alanine scanning and thereby identified five new residues within the nucleotidyltransferase domain as being essential for LigA function in vitro and in vivo. Structure activity relationships were determined by conservative mutagenesis for the Glu-173, Arg-200, Arg-208, and Arg-277 side chains, as well as four other essential side chains that had been identified previously (Lys-115, Asp-117, Asp-285, and Lys-314). In addition, we identified Lys-290 as important for LigA activity. Reference to the structure of Enterococcus faecalis LigA allowed us to discriminate three classes of essential/important side chains that: (i) contact NAD+ directly (Lys-115, Glu-173, Lys-290, and Lys-314); (ii) comprise the interface between the NMN-binding domain (domain Ia) and the nucleotidyltransferase domain or comprise part of a nick-binding site on the surface of the nucleotidyltransferase domain (Arg-200 and Arg-208); or (iii) stabilize the active site fold of the nucleotidyltransferase domain (Arg-277). Analysis of mutational effects on the isolated ligase adenylylation and phosphodiester formation reactions revealed different functions for essential side chains at different steps of the DNA ligase pathway, consistent with the proposal that the active site is serially remodeled as the reaction proceeds.

DNA ligases are grouped into two families, ATP-dependent ligases and NAD ϩ -dependent ligases, according to their nucleotide substrate requirement. DNA ligases seal 5Ј-phosphate and 3Ј-hydroxyl termini by means of three nucleotidyl transfer reactions (1). In the first step, attack on the ␣ phosphorus of ATP or NAD ϩ by ligase results in the release of pyrophosphate or NMN and the formation of a covalent ligase-adenylate intermediate. In the second step, the AMP is transferred to the 5Ј end of the 5Ј phosphate-terminated DNA strand to form a DNA-adenylate intermediate. In the third step, ligase catalyzes attack by the 3Ј-OH of the nick on DNA-adenylate to join the two polynucleotides and release AMP.
The ATP-dependent DNA ligases are found in bacteria, Archaea, and eukarya, whereas the NAD ϩ -dependent enzymes are present only in bacteria and entomopox viruses (2,3). All known bacterial species encode at least one highly conserved NAD ϩ -dependent DNA ligase (LigA), 1 which is essential for growth, even when the bacterium encodes additional NAD ϩdependent or ATP-dependent ligase enzymes (4,5). Therefore, LigA presents an attractive target for broad spectrum antimicrobial therapy predicated on blocking the reaction of LigA with NAD ϩ (6).
Crystal structures of a LigA apoenzyme and a covalent LigAadenylate intermediate (7,8) revealed that the AMP-binding pocket of bacterial NAD ϩ -dependent ligases is located within a nucleotidyltransferase domain shared with ATP-dependent DNA ligases, certain ATP-dependent RNA ligases, and GTPdependent mRNA capping enzymes (9 -13). The polynucleotide ligases and capping enzymes comprise a covalent nucleotidyltransferase superfamily, members of which catalyze transfer of a nucleoside monophosphate to the 5Ј phosphorylated end of a polynucleotide via a covalent enzyme-lysyl-nucleoside monophosphate intermediate. Five signature motifs of the nucleotidyltransferase domain (motifs I, III, IIIa, IV, and V in Fig. 1) form the nucleotide-binding pocket (14). Extensive mutational analyses of ATP-dependent DNA ligases, RNA ligases, and RNA capping enzymes have identified essential catalytic side chains within these motifs and begun to clarify their distinctive contributions at different steps along their respective reaction pathways (15)(16)(17)(18)(19). In contrast, mutational analyses of NAD ϩdependent ligases are less advanced, at least with respect to the nucleotidyltransferase domain.
Previous studies of NAD ϩ -dependent ligases have ascribed functions to individual domains of the LigA protein. All LigA enzymes consist of a central "ligase module" composed of an adenylyltransferase domain and an OB fold domain, which is flanked by a N-terminal domain (domain Ia) and several Cterminal modules, including a tetracysteine zinc-binding motif, a helix-hairpin-helix domain, and the BRCT domain (7,8). An N-terminal fragment composed of the Ia and nucleotidyltransferase domains retains full ligase adenylylation activity with NAD ϩ but is unable to catalyze phosphodiester formation at a standard 5Ј-PO 4 nick or at a pre-adenylated nick (3, 20 -22). An instructive finding was that deletion of domain Ia abolished the reaction of LigA with NAD ϩ to form ligase-adenylate but had no effect on phosphodiester bond formation at a pre-adenylated nick (3,23); these results implicated domain Ia in recognition of the NAD ϩ substrate. Mutations of five individual amino acids within domain Ia either reduced or abolished sealing of a 5Ј-PO 4 nick and adenylate transfer from NAD ϩ , without affecting ligation of pre-formed DNA-adenylate. It was suggested that these five side chains comprise a binding site for the NMN moiety of NAD ϩ (3, 23), a prediction borne out by the recently reported structure of the N-terminal domain of LigA bound to NMN (24).
Mutational analysis of the nucleotidyltransferase domain of LigA (25,26) has highlighted essential roles for five amino acids: the motif I lysine (Lys-115 in Escherichia coli LigA), which is the site of covalent attachment of AMP to the enzyme (27); the motif I aspartate (Asp-117 in EcoLigA); the motif IV aspartate (Asp-285 in EcoLigA); and the motif V lysine (Lys-314 in EcoLigA). The essential residues above are denoted by "͉" over the sequence alignment in Fig. 1. Several residues, found by alanine scanning to be nonessential for LigA function, are indicated by "ϩ" above the sequence alignment in Fig. 1.
Here we performed a new alanine scan of 14 residues of the nucleotidyl transferase domain of EcoLigA indicated by "?" in Fig. 1. The choice of positions to mutate was guided by side chain conservation and, in some cases, proximity to the covalently bound AMP in the crystal structure of the Thermus filiformis LigA-AMP adduct. Specific residues were targeted to test models of metal-dependent catalysis proposed by Lee et al. (8) on the basis of the LigA-AMP structure. We report the identification of five new essential residues in EcoLigA, clarify structure-activity relationships at these and other essential positions by conservative mutagenesis, and analyze mutational effects on individual steps of the ligation pathway. We discuss our findings in light of the recently reported structure of Enterococcus faecalis LigA bound to NAD ϩ (24).

EXPERIMENTAL PROCEDURES
Ligase Mutants-Missense mutations were introduced into the pET-EcoLIG expression plasmid using the PCR-based two-stage overlap extension method as described previously (23). The entire ligA gene was sequenced in every case to confirm the desired mutation and exclude the acquisition of unwanted changes during PCR amplification and cloning. The expression plasmids were transformed into E. coli BL21(DE3). Mutant and wild-type ligases were purified from the soluble lysates of isopropyl 1-thio-␤-D-galactopyranoside-induced BL21(DE3) cells by nickel-agarose chromatography, as described previously (23). The protein concentrations were determined using the Bio-Rad dye reagent with bovine serum albumin as a standard.
Nick Ligation-Reaction mixtures (20 l) containing 50 mM Tris-HCl (pH 7.5), 10 mM (NH 4 ) 2 SO 4 , 5 mM dithiothreitol, 5 mM MgCl 2 , 20 M NAD ϩ , 1 pmol of 5Ј 32 P-labeled nicked duplex DNA substrate, and aliquots of serial 2-fold dilutions of wild-type or mutant ligases were incubated at 22°C for 10 min. The products were resolved by electrophoresis through a 15-cm 18% polyacrylamide gel containing 7 M urea in TBE (90 mM Tris borate, 2.5 mM EDTA). The extents of ligation were determined by scanning the gel with a Fujix BAS2500 imager. The specific activities of wild-type and mutant ligases were determined from the slopes of the titration curves in the linear range of enzyme dependence. The activities of the mutant ligases were normalized to the wildtype specific activity (defined as 100%).
Ligase Adenylylation-Reaction mixtures (20 l) containing 50 mM Tris-HCl (pH 7.5), 5 mM dithiothreitol, 5 mM MgCl 2 , 1 M [ 32 P]adenylate-NAD (purchased from PerkinElmer Life Sciences), and 8 pmol of wild-type or mutant LigA were incubated at 22°C for 10 min. The reactions were quenched with SDS, and the products were analyzed by SDS-PAGE. The ligase-[ 32 P]AMP adduct was visualized by autoradiography and quantified by scanning the gel with a Fujix imager.
Ligation at a Pre-adenylated Nick-The nicked DNA-adenylate substrate is shown in Fig. 4. The 5Ј adenylated 32 P-labeled 18-mer strand (AppDNA) was prepared using Mycobacterium tuberculosis LigC (5) as follows. Ten reaction mixtures (100 l each) containing 50 mM Tris-HCl (pH 7.5), 10 mM MgCl 2 , 5 mM dithiothreitol, 1 mM ATP, 50 pmol of a singly nicked three-piece duplex DNA substrate containing a 5Ј 32 Plabeled 18-mer strand at the nick, and 10 g of MtuLigC were incubated at 37°C for 30 min. The reactions were halted by adding EDTA to a concentration of 25 mM. The samples were pooled and then ethanolprecipitated. The pelleted material was resuspended in formamide, heated at 95°C for 5 min, and then subjected to electrophoresis through an 18% polyacrylamide gel containing 7 M urea in TBE. The 32 P-labeled AppDNA strand, which was separated clearly from the 18-mer input pDNA strand, was located by autoradiography of the wet gel and then eluted from an excised gel slice. The 18-mer AppDNA was ethanolprecipitated and resuspended in 10 mM Tris-HCl (pH 8.0) and 1 mM EDTA. The nicked DNA-adenylate substrate was formed by annealing the gel-purified AppDNA strand to the complementary 36-mer and a 3Ј-OH acceptor strand at a molar ratio of 1:4:4.
AppDNA ligation reaction mixtures (100 l) containing 50 mM Tris-HCl (pH 7.5), 5 mM dithiothreitol, 5 mM MgCl 2 , 600 fmol nicked DNAadenylate substrate, and 6 pmol of wild-type or mutant EcoLigA proteins were incubated at 22°C. The sealing reactions were initiated by adding ligase. Aliquots (10 l) were withdrawn at the times specified in Fig. 4 and quenched immediately with EDTA and formamide. The products were resolved by denaturing PAGE, and the extents of ligation were determined by scanning the gel with a Fujix imager.
Mutational Effects on LigA Function in Vivo in Yeast-NdeI-BamHI restriction fragments containing LigA mutant alleles were excised from their respective pET-based LigA plasmids and inserted into pYX1-His (CEN TRP1). In this vector, expression of ligA is under the control of the yeast TPI1 promoter. The pYX1-LIGA plasmids were transformed into a yeast cdc9⌬ strain bearing a CEN URA3 CDC9 plasmid. Trp ϩ isolates were streaked on agar plates containing 0.75 mg/ml of 5-fluoroorotic acid (FOA). Lethal mutations (Ϫgrowth) were those that formed no FOA-resistant colonies at 18, 25, 30, or 37°C. The viable cdc9⌬ ligA yeast strains were tested for growth on YPD agar at 18, 25, 30, and 37°C. "ϩϩϩ" indicates colony size indistinguishable from a cdc9⌬ strain expressing wild-type LigA.

Alanine Scanning Identifies LigA Residues Essential for Nick
Joining in Vitro-Single alanine substitutions were introduced at 14 conserved positions in the nucleotidyltransferase domain of EcoLigA (Fig. 1). His 10 -tagged G118A, L119A, D138A, E143A, G172A, E173A, N198A, R200A, R208A, R277A, D283A, G286A, V288A, and K290A mutants were produced in E. coli and purified by nickel-agarose chromatography in parallel with wild-type EcoLigA ( Fig. 2A). The extent of ligation of singly nicked 3Ј-OH/5Ј-PO 4 DNA by wild-type EcoLigA and each mutant was gauged as a function of the input enzyme, and the specific activities were normalized to the wild-type value (defined as 100%). Our operational definition of an essential amino acid was one at which alanine substitution reduced nick-sealing activity to Յ10% of the wild-type level. By this criterion, Gly-118, Glu-173, Arg-200, Arg-208, and Arg-277 were deemed essential for ligation (Fig. 2B).
Eight residues were judged to be nonessential for ligation in vitro (Fig. 2B). The L119A change flanking motif I had no impact on nick-joining activity. This leucine side chain was initially reported to line the AMP-binding pocket of TfiLigA and to make multiple van der Waals contacts with the ribose sugar of the covalently bound adenylate (8). It has since been determined that the amino acid sequence used to build the TfiLigA structure contained numerous errors (including incorrect side chains, deleted amino acids, and extra amino acids in the nucleotidyltransferase domain), which the authors have recently rectified. In the revised version of the TfiLigA structure (Protein Data Bank 1V9P), the leucine corresponding to Leu-119 of EcoLigA makes no contact with the bound adenylate. (Henceforth, all references to the TfiLigA structure will be to the revised version.) The equivalent leucine side chain makes no contacts with NAD ϩ in the EfaLigA structure (24).
The D138A change caused a modest reduction of ligase activity to 37% of wild-type LigA. The conserved Asp side chain is located on the surface of TfiLigA ϳ15 Å from the bound AMP. The E143A mutation had little effect on ligation (52% of wildtype activity); this conserved glutamate is located on the surface of TfiLigA ϳ17 Å away from the bound adenylate. The modest decrement in activity elicited by the G172A change in motif III (36% of wild-type nick joining) might reflect a subtle effect on the position of the vicinal Glu-173 side chain, which lines the AMP-binding pocket and is essential for EcoLigA catalysis. The N198A change reduced nick-joining activity to one-fourth of wild-type LigA. This conserved Asn side chain is located ϳ9 Å away from the AMP phosphate in TfiLigA-adenylate and is 7-8 Å away from the NMN moiety of NAD ϩ in the EfaLigA structure.
The D283A mutation had little impact on ligase activity (57% of wild-type), a finding that vitiates the model of Lee et al. (8), which proposes that this carboxylate side chain is required to coordinate an essential divalent cation cofactor during the step of phosphodiester formation. The G286A mutation in nucleotidyl transferase motif IV lowers nick-joining activity to 14% of wild-type; we suspect that the loss of activity reflects steric effects of the alanine methyl group on the conformation of the motif IV beta strand, especially the position of the vicinal Asp285 side chain, which is essential for EcoLigA activity (26). The V288A change in motif IV was benign (74% of wild-type activity). This conserved aliphatic side chain makes van der Waals contacts with the adenine base of AMP in TfiLigA and the adenine of NAD ϩ in EfaLigA (24). We conclude that such contacts are not important for EcoLigA function. The K290A change reduced nick-joining activity to 13% of the wild-type level. This conserved lysine side chain lines the adenosine-binding pocket of TfiLigA and makes van der Waals contacts with the N1 and C2 atoms of the adenine base. Although Lys-290 does not meet our criterion for essentiality, it probably does play a role in substrate binding (see "Discussion").

Effects of Alanine Mutations on EcoLigA Function in Vivo in
Yeast-A deletion of the essential Saccharomyces cerevisiae ATP-dependent DNA ligase Cdc9 can be complemented by expression of EcoLigA (26). Viability of the yeast cdc9⌬ strain is contingent on maintenance of an extrachromosomal CDC9 gene on a CEN URA3 plasmid. Hence, cdc9⌬ cells cannot grow on medium containing 5-FOA (a drug that selects against the URA3 CDC9 plasmid), but they can grow on 5-FOA if the cells have been transformed with a CEN TRP1 plasmid expressing wild-type EcoLigA under the control of the constitutive yeast TPI1 promoter. Here we tested, by the plasmid shuffle assay (26), whether the EcoLigA-Ala mutants were functional in yeast. We found that mutations of the five residues defined as essential for EcoLigA function in vitro (Gly-118, Glu-173, Arg-200, Arg-208, Arg-277) were all lethal in vivo, i.e. they were unable to support growth of cdc9⌬ on 5-FOA at either 18, 25, 30, or 37°C (Fig. 2B). Mutations of seven of the residues defined as nonessential for nick-joining activity in vitro were functional in cdc9⌬ complementation in vivo. Only the least  Nick-joining activity of the LigA-Ala mutants was gauged by enzyme titration as described under "Experimental Procedures" and expressed as the percent of the wild-type value. LigA complementation of a yeast cdc9⌬ strain was tested by plasmid shuffle (26). Growth was scored as described under "Experimental Procedures." active of the nonessential mutants, K290A, with 13% of wildtype activity in vitro, was unable to complement growth of cdc9⌬. The viable cdc9⌬ ligA-Ala strains L119A, D138A, E143A, G172A, N198A, D283A, and V288A grew as well as the wild-type cdc9⌬ ligA strain on YPD agar medium at 18, 25, 30, and 37°C, as gauged by colony size (not shown). The G286A strain grew as well as a wild-type ligA strain at 25 and 30°C but formed smaller colonies at 37°C (not shown).
Structure-Activity Relationships at Essential Residues of EcoLigA-The present results, together with previous mutational data (26), highlight eight individual non-glycine amino acids within the nucleotidyltransferase domain that are essential for EcoLigA activity in vitro: Lys-115, Asp-117, Glu-173, Arg-200, Arg-208, Arg-277, Asp-285, and Lys-314. To better evaluate the contributions of these residues to the DNA ligase reaction, we tested the effects of conservative substitutions whereby arginine was replaced by lysine and glutamine, glutamate by glutamine and aspartate, aspartate by asparagine and glutamate, and lysine by arginine and glutamine. Sixteen conservative EcoLigA mutants were produced in E. coli and purified from soluble bacterial extracts by nickel-agarose chromatography. The LigA polypeptide was the predominant species detected by SDS-PAGE, and the extents of purification were comparable for mutant and wild-type EcoLigA (Fig. 3A).
The nick-joining and ligase adenylylation activities of the mutants were determined and normalized to the wild-type value (Fig. 3B). Conservative replacement of the motif I Lys-115 by either arginine or glutamine abolished nick-joining activity and the reaction with [ 32 P]NAD ϩ to form the ligaseadenylate intermediate. These results highlight a strict requirement for lysine as the active site nucleophile. Replacing Asp-117 with asparagine abolished nick-joining activity (Ͻ0.1% of wild-type), and a glutamate substitution resulted in a minimal restoration of function (to 3% of wild-type). These data establish the requirement for a carboxylate residue at position 117, and we infer that there is a tight steric constraint that precludes accommodation of the extra methylene group of a glutamate. Previous studies showed that replacing the motif I Asp with alanine does not prevent the ligase adenylylation reaction with NAD ϩ (25). Here we find that the conservative D117N and D117E mutants retained 35 and 53% of wild-type adenylyltransferase activity, respectively. These findings point to an essential role of the conserved motif I aspartate during either step 2 or step 3 of the ligase pathway (25).
LigA nick-joining activity was abolished when the motif III Glu-173 side chain was replaced with either aspartate or glutamine. Thus, activity requires a carboxylate and a minimum distance from the main chain to the carboxylate that is met by glutamate but not aspartate. The E173D and E173Q mutants retained 13 and 54% of wild-type adenylyltransferase activity, respectively, signifying that an amide group suffices for the reaction of LigA with NAD ϩ to form ligase-adenylate, and the stringent requirement for a carboxylate is probably exerted at a downstream step of the ligation pathway.
Conservative replacement of Arg-200 and Arg-208 with either lysine or glutamine failed to restore nick-joining activity. Changing Arg-277 to glutamine reduced nick joining to 1.8% of the wild-type level, similar to the R277A mutant, and there was only marginal restoration of activity with a lysine substitution (to 5.6% of wild-type). Thus a positive charge is not sufficient at positions 200, 208, or 277; rather, we posit that these side chains engage in bidentate hydrogen bonding or ionic contacts that are sustained uniquely by arginine. Arg-208 is apparently essential for the reaction of LigA with NAD ϩ , insofar as the adenylyltransferase activity of the R208Q and R208K proteins was 0.1-2% of the wild-type activity (Fig. 3B). In contrast, Arg-200 plays a more complex role whereby adenylyltransferase activity is partially restored by a lysine (14% of wild-type) but not a glutamine (1% activity); but overall ligation requires an arginine. The effects of the R277Q and R277K changes on ligase adenylylation closely paralleled their impact on overall ligation.
Replacing the essential motif IV Asp-285 side chain with asparagine abolished nick-joining activity (Ͻ0.1% of wild-type), and introducing a glutamate had minimal salutary effect (4% of wild-type). Thus, a carboxylate is required at position 285, and the extra methylene group of a glutamate side chain is not easily accommodated. The D285E and D285N mutants retained 6 and 23% of wild-type adenylyltransferase activity, respectively, signifying that, although an amide group at the proper distance from the main chain partially restores reactivity of LigA with NAD ϩ , a carboxylate is required at a downstream step.
Conservative substitution of the motif V Lys-314 side chain with glutamine abolished nick-joining activity and reduced adenylyltransferase to 3% of wild-type. Introducing an arginine increased ligation and adenylyltransferase activities to 5 and 13% of wild-type, respectively.
Conservative Mutational Effects on EcoLigA Function in Vivo in Yeast-The conservative mutants were tested by plasmid shuffle for cdc9⌬ complementation (Fig. 3B). The results underscore the theme that LigA mutations that abolish or severely depress nick-joining activity in vitro are lethal in vivo in yeast. The R227K mutant, which was the most active of the proteins in this collection (with 5.6% of wild-type ligase function), was able to sustain yeast growth on YPD medium at 25 and 30°C, but not at 37°C (scored as ts in Fig. 3B). The classical ligA-ts7 mutant strain of E. coli has a similar low residual level of strand-joining activity at the permissive temperature (28).

Effects of Lys-115, Glu-173, Arg-200, Arg-208, and Arg-277 Mutations on the Isolated
Step of Phosphodiester Formation at a Pre-adenylated Nick-The third step of the ligation pathway entails attack of the 3Ј-OH of the nick on the 5Ј-PO 4 of the DNA-adenylate to form a phosphodiester and release AMP. We assayed step 3 of the ligation reaction using a pre-adenylated nicked DNA substrate labeled with 32 P at the 5Ј-PO 4 of the DNA-adenylate strand (Fig. 4). Reaction of wild-type EcoLigA with the nicked DNA-adenylate in the presence of magnesium without added NAD ϩ resulted in strand closure, evinced by formation of a radiolabeled 36-mer product. A kinetic analysis of the sealing of nicked DNA-adenylate by wild-type ligase and mutants of residues Lys-115, Glu-173, Arg-200, Arg-208, and Arg-277 is presented in Fig. 4. As reported previously (23), the wild-type step 3 reaction of wild-type LigA attained an end point in 5 min (Fig. 4A). Conservative changes in the motif I lysine had starkly disparate effects on phosphodiester formation whereby K115Q was unreactive with the nicked DNAadenylate, whereas K115R was just as active as wild-type LigA (Fig. 4A). This result shows that a positive charge at this motif I position is necessary and sufficient for step 3 catalysis and underscores how the motif I lysine plays different roles at different steps of the ligation pathway.
Mutations of Glu-173 in motif III have instructive effects on sealing at a pre-adenylated nick. Although alanine and glutamine substitutions slow the initial rate to 3 and 2% of the wild-type, respectively, the aspartate substitution elicits a gain of step 3 function to 14% of the wild-type rate (Fig. 4B). The hierarchy of conservative effects on step 3 activity, whereby E174D is more active than E173Q, is opposite to that on step 1 ligase adenylylation, where E173Q is more active than E173D (Fig. 3B). Thus, the motif III side chain acts differently at different steps along the reaction pathway, with an amide functional group sufficing for step 1 but not step 3 and a carboxylate being required for step 3.
Changing Arg-208 to either alanine or lysine abolished overall ligation and precluded the ligase adenylylation step (Fig.  3B); thus it is remarkable that these same mutations had relatively little impact on the kinetics of phosphodiester formation at a pre-adenylated nick (Fig. 4C). The rate of step 3 by R208A and R208K was about one-half that of wild-type LigA, signifying that Arg-208 is not required for catalysis of the sealing step. Nonetheless, the R208Q change was apparently deleterious to step 3 catalysis, insofar as the step 3 rate of R208Q was ϳ7-fold slower than that of either R208A or R2308K (Fig. 4C).
Mutations R277A, R277Q, and R277K slowed the rate of the isolated sealing step to 20, 10, and 4% of the wild-type LigA, respectively (Fig. 4D). Although the lysine change had a greater impact on step 3 than did glutamine, the opposite was seen for the nick ligation and ligase adenylylation reactions, for which the R277K mutant was more active than R277Q. All of the Arg-200 mutations strongly suppressed sealing at a pre-adenylated nick (Fig. 4E). The step 3 rates of the R200K, R200Q, and R200A proteins were 7, 3, and 2% of the wild-type, respectively. Thus, although a lysine could partially substitute for arginine in ligase-adenylylation, an arginine was strictly required for step 3 and overall nick joining. DISCUSSION The LigA reaction is initiated by the attack of the motif I lysine on the NAD ϩ substrate to form ligase-adenylate. The reaction likely proceeds through a pentacoordinate phosphorane transition state in which the attacking lysine nucleophile is apical to the NMN leaving group. The structure of E. faecalis ligase revealed that coordination of the NMN moiety of NAD ϩ is achieved by the closure of domain Ia over the AMPbinding pocket of the nucleotidyltransferase domain. Breaking of the ␣-␤ phosphoanhydride bond of NAD ϩ upon enzymeadenylate formation releases the tether of domain Ia to the nucleotidyltransferase domain and allows LigA to adopt an open conformation in which the AMP phosphate is exposed on the now free surface of the nucleotidyltransferase domain, to which the DNA nick must bind for catalysis of steps 2 and 3. The nucleotidyltransferase domain of LigA is not able to bind DNA by itself; rather the DNA-binding step and catalysis of phosphodiester formation require contribution from the flanking C-terminal domain modules (3, 20 -22, 25).
Here we performed an alanine scan of 14 conserved amino acids of the nucleotidyltransferase domain of EcoLigA and thereby identified five essential residues (Gly-118, Glu-173, Arg-200, Arg-208, and Arg-277). Structure-activity relationships were determined for Glu-173, Arg-200, Arg-208, and Arg-277, as well as four other essential side chains (Lys-115, Asp-117, Asp-285, and Lys-314). In addition, we identified Lys-290 as important for LigA activity, although not essential by our cut-off criterion. Reference to the EfaLigA crystal structure (24) allows us to discriminate three classes of essential/important side chains: (i) those that contact NAD ϩ directly (e.g. Lys-115, Glu-173, Lys-290, and Lys-314), (ii) those that either comprise the interface between domain Ia and the nucleotidyltransferase domain in the closed conformation or comprise part of the DNA docking site on the surface of the nucleotidyltransferase domain in the open conformation (e.g. Arg-200 and Arg-208), and (iii) those that stabilize the fold of the nucleotidyltransferase domain (e.g. Arg-277). Although not directly implicated as such from available LigA structures, we suggest that Asp-285 and perhaps Asp-117 comprise a metal-binding site. The several classes of essential residues are considered individually below.
NAD ϩ -binding Residues-The role of the motif I lysine nucleophile in the ligase adenylylation step is established biochemically and structurally (8,25) (Fig. 5). Here we show that arginine is unable to replace lysine in forming the ligase-AMP intermediate of EcoLigA; similar findings were reported previously for Thermus thermophilus LigA (25). We presume that the higher pK a of arginine hinders the necessary deprotonation of the nucleophilic nitrogen atom during catalysis of step 1. Yet, arginine is fully able to perform the essential role of the motif I lysine during the step of phosphodiester bond formation by EcoLigA. A role for the motif I lysine in step 3 of the NAD ϩdependent ligase reaction had not been demonstrated previously. Our results suggest that the positively charged side chain of the motif I lysine provides an essential contact with the AMP leaving group during the attack of the 3Ј-OH of the nick on the DNA 5Ј-phosphate of the AppDNA intermediate.
The essential motif III glutamate coordinates the adenosine ribose O2Ј atom in the NAD ϩ -bound substrate complex of Efa-LigA (Fig. 5, left panel). An amide functional group is apparently capable of coordinating the ribose during the ligase adenylylation reaction, insofar as the E173Q mutant retained about half of wild-type activity in ligase-AMP formation. The ribose contact to the motif III glutamate undergoes remodeling during progression through the ligation pathway in synchrony with a syn-anti conformational switch of the adenosine nucleoside (Fig. 5). Adenosine is in the syn conformation in the closed complex of EfaLigA bound to NAD ϩ , but it adopts an anti conformation in the open state of the TfiLigA-adenylate covalent intermediate (Fig. 5, right panel). The rotation of the ribose sugar around the glycosidic bond disrupts the contact of the motif III glutamate with the ribose hydroxyl. The fact that the motif III glutamate is required for catalysis of step 3 suggests that the ribose undergoes yet another conformational switch after catalysis step 2, such that a contact is restored between the glutamate and the adenosine of AppDNA. This putative step 3 contact is apparently fulfilled only by a carboxylate functional group, insofar as E173D is more active in phosphodiester formation than is E173Q. It is important to note that changes in nucleoside conformation from syn to anti after step 1 catalysis, and domain reopening accompanied by remodeling of the motif III glutamate contact to the adenosine ribose, are also seen in ATP-dependent DNA ligases (11). The motif III glutamate is essential for step 3 catalysis by ATP-dependent DNA ligase, and it is reported that Asp can function in lieu of Glu during step 3, although not during step 1 or the composite nick-joining reaction of an ATP-dependent DNA ligase (15).
Lys-290 is a conserved residue immediately flanking motif IV of the bacterial LigA enzymes (Fig. 1). Loss of this side chain reduced EcoLigA activity in vitro to 13% of wild-type and precluded LigA function in vivo in the yeast cdc9⌬ complementation assay. The EfaLigA structure reveals that this lysine side chain donates a hydrogen bond to the N1 atom of the adenine of NAD ϩ and also forms a bifurcated salt bridge to the carboxylate oxygens of a conserved glutamate (Glu-113 in EcoLigA) located two residues upstream of the motif I lysine (24). The E113A mutant of EcoLigA retained 40% of wild-type nickjoining activity in vitro and was fully active in cdc9⌬ complementation (26); the more severe effects of the K290A mutation of EcoLigA imply that Lys-290 is not merely functioning through Glu-113 but rather that the contact of Lys-290 to the adenine base is important per se for ligase activity.