ø29 DNA Polymerase Residue Ser122, a Single-stranded DNA Ligand for 3′-5′ Exonucleolysis, Is Required to Interact with the Terminal Protein*

Three amino acid residues highly conserved in most proofreading DNA polymerases, a phenylalanine contained in the Exo II motif and a serine and a leucine belonging to the S/TLx2 h motif, were recently shown to be critical for 3′-5′ exonucleolysis by acting as single-stranded DNA ligands (de Vega, M., Lázaro, J.M., Salas, M. and Blanco, L. (1998) J. Mol. Biol. 279, 807–822). In this paper, site-directed mutants at these three residues were used to analyze their functional importance for the synthetic activities of ø29 DNA polymerase, an enzyme able to start linear ø29 DNA replication using a terminal protein (TP) as primer. Mutations introduced at Phe65, Ser122, and Leu123 residues of ø29 DNA polymerase severely affected the replication capacity of the enzyme. Three mutants, F65S, S122T, and S122N, were strongly affected in their capacity to interact with a DNA primer/template structure, suggesting a dual role during both polymerization and proofreading. Interestingly, mutant S122N was not able to maintain a stable interaction with the TP primer, thus impeding the firsts steps (initiation and transition) of ø29 DNA replication. The involvement of Ser122 in the consecutive binding of TP and DNA is compatible with the finding that the TP/DNA polymerase heterodimer was not able to use a DNA primer/template structure. Assuming a structural conservation among the eukaryotic-type DNA polymerases, a model for the interactions of ø29 DNA polymerase with both TP and DNA primers is presented.

The linear dsDNA 1 of Bacillus subtilis phage ø29 starts replication by a protein-priming mechanism (reviewed in Refs. 1 and 2) in which the viral DNA polymerase catalyzes the addition of dAMP to the hydroxyl group of Ser 232 of a free TP molecule. This event occurs opposite the second 3Ј nucleotide of the template, and then the initiation product (TP-dAMP) generated slides back one position to be paired to the terminal nucleotide (3). After sliding back, the same DNA polymerase molecule catalyzes the synthesis of a short elongation product (9 nucleotides) while it is still bound to the TP (transition step) (4). The incorporation of the next nucleotide forces dissociation of the TP/DNA polymerase heterodimer, starting normal DNA elongation initially coupled to strand displacement. Strand displacement synthesis proceeds from both DNA ends until the two replication forks meet. Then, the two partially replicated parental strands separate, and full-length DNA synthesis (19,285 base pairs in length) is completed.
On the other hand, three conserved amino acid motifs, named Exo I, II, and III, have been identified in the N-terminal portion of all prokaryotic and eukaryotic DNA-dependent DNA polymerases with a proofreading activity (11,18,19) that invariantly contain the five critical residues that were identified in Pol Ik acting as metal ligands and responsible for the 3Ј-5Ј exonuclease catalysis (20,21). The proposal of an evolutionarily conserved 3Ј-5Ј exonuclease active site (18) has been confirmed by 1) site-directed mutagenesis studies in some DNA polymerases (reviewed in Ref. 22) belonging to family A (Pol I-like) and family B (eukaryotic-type): Pol Ik (21,23), T7 DNA polymerase (24), ø29 DNA polymerase (Refs. 18, 25, and 26; reviewed in Ref. 5), T4 DNA polymerase (27)(28)(29)(30), herpes simplex virus DNA polymerase type 1 (31)(32)(33)(34)(35), E. coli DNA polymerase II (36,37), PRD1 DNA polymerase (38), Thermococcus littoralis DNA polymerase (39,40), and cellular DNA polymerases ␦ (41), ⑀ (19), and ␥ (42,43) from Saccharomyces cerevisiae, and family C (DNA polymerase III), consisting of ⑀ subunit of E. coli DNA polymerase III (44) and B. subtilis DNA polymerase III (45,46); and 2) the recent resolution of the crystal structure of a Nterminal fragment of T4 DNA polymerase (47) and that of phage RB69 DNA polymerase (48). Both structures served to identify a lysine as an additional active site residue (30,47,48). The functional importance of this residue was supported by site-directed mutagenesis analysis in ø29 DNA polymerase (49) together with its invariant presence in the eukaryotic DNA polymerase family (48,49). In addition to the residues involved in metal binding and catalysis at the 3Ј-5Ј exonuclease active site, other residues are structurally and functionally conserved at the exonuclease domain of most proofreading DNA polymerases. Among them, ø29 DNA polymerase residues Thr 15 and Asn 62 of the Exo I and II motifs, respectively, were shown to act as ssDNA ligands, having a critical role in the stabilization of the primer terminus at the 3Ј-5Ј exonuclease active site (50). Another highly conserved residue at the Exo II motif (Phe 65 in ø29 DNA polymerase) and two residues (Ser 122 and Leu 123 ) belonging to the newly identified S/TLx 2 h motif were shown to act also as ssDNA ligands for 3Ј-5Ј exonucleolysis (51), in agreement with the crystallographic data from T4 (47) and RB69 (48) DNA polymerases complexed with ssDNA.
To date, in the case of ø29 DNA polymerase, none of the mutations carried out at both metal and ssDNA ligands affected the synthetic activities of the enzyme (protein-primed initiation and DNA polymerization). However, as described in this paper, a detailed biochemical characterization of various mutants at the ssDNA ligands Phe 65 , Ser 122 , and Leu 123 of ø29 DNA polymerase demonstrated the involvement of these residues also in primer (both DNA and TP) binding during the synthetic reactions. Based on these data, and assuming conservation of the overall structure among eukaryotic-type DNA polymerases, a three-dimensional structure prediction for the ø29 DNA polymerase/TP interaction, modeled on the crystal structure of RB69 DNA polymerase (48), is proposed.
DNA Templates and Substrates-Oligonucleotides 15-mer (5Ј-GAT-CACAGTGAGTAC) and 21-mer (5Ј-TCTATTGTACTCACTGTGATC), which has a 5Ј-extension of six nucleotides in addition to the sequence complementary to the 15-mer, were supplied by Isogen. Oligonucleotide 15-mer was 5Ј-labeled with [␥-32 P]ATP and T4 polynucleotide kinase. 5Ј-Labeled 15-mer was hybridized to 21-mer in the presence of 0.2 M NaCl and 60 mM Tris-HCl, pH 7.5. ø29 DNA was obtained by proteinase K treatment of phage particles in the presence of SDS (56), phenol extraction, and ethanol precipitation. M13mp18 ssDNA was hybridized to the universal primer as described above, and the resulting molecule was used as a primer/template to analyze processive DNA polymerization coupled to strand displacement by ø29 DNA polymerase. Terminal protein-containing ø29 DNA (ø29 TP-DNA) was obtained as described (57).
Replication Assay (Protein-primed Initiation plus Elongation) with ø29 TP-DNA as Template-The incubation mixture contained, in 25 l, 50 mM Tris-HCl, pH 7.5; 10 mM MgCl 2 ; 20 mM ammonium sulfate; 1 mM dithiothreitol; 4% glycerol; 0.1 mg/ml BSA; 20 M each dCTP, dGTP, dTTP, and [␣-32 P]dATP (1 Ci); 0.5 g of ø29 TP-DNA; 125 ng of purified TP; and 5 ng of either wild-type or mutant ø29 DNA polymerase. After incubation for 10 min at 30°C, the reaction was stopped by adding 10 mM EDTA-0.1% SDS, and the samples were filtered through Sephadex G-50 spin columns. Relative activity was calculated from the Cerenkov radiation corresponding to the excluded volume. For size analysis, the labeled DNA was denatured by treatment with 0.7 M NaOH and subjected to electrophoresis in alkaline 0.7% agarose gels, as described (58). After electrophoresis, the position of unit-length ø29 DNA (19,285 bases) was detected by ethidium bromide staining, and then the gels were dried and autoradiographed. For the analysis of the transition products, 100 ng of DNA polymerase were used, providing the concentration of dNTP indicated in each case. The samples were analyzed in an SDS-12% polyacrylamide gel (360 ϫ 280 ϫ 0.5 mm) to obtain enough resolution to distinguish the TP bound to the first elongation products.
Replication of Primed M13 DNA-The incubation mixture contained, in 25 l, 50 mM Tris-HCl, pH 7.5; 10 mM MgCl 2 ; 1 mM dithiothreitol; 4% glycerol; 0.1 mg/ml BSA; 80 M each dCTP, dGTP, dTTP, and [␣-32 P]dATP (2.5 Ci); 0.25 g of primed M13mp8 ssDNA; and 25 ng of either wild-type or mutant ø29 DNA polymerase. After incubation for the indicated times at 30°C, the samples were processed, and the synthesized DNA was quantitated and analyzed as described above for the ø29 TP-DNA replication assay. After electrophoresis, unit-length M13mp8 ssDNA was detected by ethidium bromide staining, and then, gels were dried and autoradiographed.
TP-dAMP Formation (Protein-primed Initiation Assay)-The incubation mixture contained, in 25 l, 50 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 20 mM ammonium sulfate, 1 mM dithiothreitol, 4% glycerol, 0.1 mg/ml BSA, 0.2 M dATP ([␣-32 P]dATP) (2.5 Ci), 0.5 g of ø29 TP-DNA, 125 ng of purified TP, and 5 ng of either wild-type or mutant ø29 DNA polymerase and was incubated for 4 min at 30°C. In the case of the template-independent initiation assay, ø29 TP-DNA was omitted, 80 ng of either wild-type or mutant ø29 DNA polymerase were added, 1 mM MnCl 2 was used instead of MgCl 2 , and the incubation was maintained for 90 min at 30°C. The reactions were stopped by adding 10 mM EDTA and 0.1% SDS, filtered through Sephadex G-50 spin columns, and further analyzed by SDS-PAGE as described (57). Quantitation was done by densitometric analysis of the labeled band corresponding to the TP-dAMP complex, detected by autoradiography.
Interference Assay for TP Binding-Reactions were carried out as described for the template-independent initiation assay, but in the absence of template and using a limiting amount of TP and different proportions of a mixture of wild-type and mutant DNA polymerases. ø29 DNA polymerase mutant D249E (catalytically inactive but displaying a normal interaction with the TP) was used as a positive control for the interference assay, as described previously (15). The amounts of proteins used were as follows: 10 ng of TP, 25 ng of wild-type DNA polymerase, and either 25, 200, or 800 ng of each mutant derivative (D249E, F65S, or S122N). In all cases, the incubation was for 2 h at 30°C. After incubation, reactions were stopped and analyzed as indicated for the protein-primed initiation assay.
Analysis of the Interaction between TP and DNA Polymerase Mutants-The incubation mixture contained, in 150 l, 50 mM Tris-HCl, pH 7.5, 1 mM dithiothreitol, 0.1 mg/ml BSA, 20 mM ammonium sulfate, 1 g of TP, and 2 g of either wild-type or mutant DNA polymerase. After incubation for 30 min at 4°C, samples were loaded in 15-30% glycerol gradients (4 ml) in the presence of 50 mM Tris-HCl, pH 7.5, 20 mM ammonium sulfate, 0.2 M NaCl, 1 mM EDTA, and 7 mM 2-mercaptoethanol and centrifuged at 4°C for 24 h at 62,000 rpm in a Beckman SW65 rotor. Gradients were fractionated and subjected to SDS-PAGE in 12% polyacrylamide gels. After electrophoretical separation, the proteins in the gel were transferred to a polyvinylidene difluoride membrane (Millipore) during 120 min at 100 mA and incubated with polyclonal antibodies against ø29 DNA polymerase and TP. Detection was carried out using the ECL Western blotting system (Amersham Pharmacia Biotech).
DNA Gel Retardation Assay-A 5Ј-labeled 15-mer/21-mer hybrid molecule was used as DNA primer/template to analyze the interaction with either wild-type or mutant ø29 DNA polymerases. The incubation mixture, in a final volume of 20 l, contained 12 mM Tris-HCl, pH 7.5, 1 mM EDTA, 10 mM MgCl 2 , 20 mM ammonium sulfate, 0.1 mg/ml BSA, 0.18 ng of the 15-mer/21-mer molecule, and 5 ng of either wild-type or mutant ø29 DNA polymerase. To analyze the interaction of wild-type ø29 DNA polymerase/TP heterodimer with DNA primer/template and ssDNA, the heterodimer was preformed by incubating 40 ng of wildtype DNA polymerase with 20 ng of TP, in the presence of 12 mM Tris-HCl, pH 7.5, and 20 mM ammonium sulfate. After incubation for 30 min at 4°C, either the heterodimer, 40 ng of wild-type ø29 DNA polymerase, or 20 ng of TP were incubated with either 0.18 ng of the 15-mer/ 21-mer or with 0.075 ng of 15-mer molecule, under the conditions described above. When ssDNA was used as substrate, MgCl 2 was omitted to avoid its degradation. After incubation for 5 min at 4°C, the samples were subjected to electrophoresis in 4% (w/v) polyacrylamide gels (80:1, monomer:bis), containing 12 mM Tris acetate, pH 7.5, and 1 mM EDTA and run at 4°C in the same buffer at 8 V/cm, essentially as described (60). After autoradiography, ø29 DNA polymerase, TP, and DNA polymerase/TP complexed with ssDNA and dsDNA were detected as a mobility shift (retardation) in the migrating position of the labeled DNA. Quantitation of the binding capacity of the wild-type ø29 DNA polymerase versus its mutant derivatives was carried out by densitometry of the retarded band.
Polymerization of the Wild-type ø29 DNA Polymerase/TP Heterodimer on a Short Primer/Template Structure-The hybrid molecule 15-mer/21-mer can be used as a substrate for DNA-dependent DNA polymerization, because it contains a 6-nucleotide-long 5Ј-protruding end. The incubation mixture contained, in 12.5 l, 12 mM Tris-HCl, pH 7.5; 10 mM MgCl 2 ; 0.18 ng of 5Ј-labeled 15-mer/21-mer; 400 nM each of the four dNTPs; and either 40 ng of wild-type ø29 DNA polymerase; 20 ng of TP, or 60 ng of the DNA polymerase/TP complex (preformed as indicated above). After incubation for 2 min at 4°C (to guarantee the stabilization of the interaction between TP and DNA polymerase), the reaction was stopped by adding EDTA up to 10 mM. Samples were analyzed by 8 M urea-20% PAGE and autoradiography. Polymerization is detected as an increase in the size (15-mer) of the 5Ј-labeled 15-mer primer.
3Ј-5Ј Exonuclease Assay of the Wild-type ø29 DNA Polymerase/TP Heterodimer-The incubation mixture contained, in 12.5 l, 12 mM Tris-HCl, pH 7.5; 10 mM MgCl 2 ; and either 40 ng of the wild-type ø29 DNA polymerase, 20 ng of TP, or 60 ng of the DNA polymerase/TP heterodimer (preformed as indicated above). 0.075 ng of a 5Ј-labeled oligonucleotide (15-mer) was used as ssDNA substrate. Samples were incubated for 2 min at 4°C (to preserve the stabilization of the TP/DNA polymerase heterodimer) and quenched by adding 3 l of sequencing gel loading buffer. Reactions were analyzed by 8 M urea-20% PAGE and autoradiography.

The Exo II and (S/T)Lx 2 h Motifs of Protein-primed DNA
Polymerases- Fig. 1 shows a multiple alignment of two amino acid motifs, highly conserved in all DNA-dependent DNA polymerases endowed with a proofreading activity, but restricted here to the subgroup of eukaryotic-type DNA polymerases that use a protein-priming mechanism and to RB69 DNA polymerase, the crystal structure of which has been recently solved (48). The first block of amino acid similarity (originally called the Exo II motif; Ref. 18) can be defined by the consensus sequence hhxANx 2-3 (F/Y)Dx 2 Ahh. An aromatic residue (His in most cases) precedes the almost invariant Asn residue involved in ssDNA binding, both by crystallographic analysis (47,48,61) and site-directed mutagenesis studies (50). Similarly, the critical Asp residue of the Exo II motif is always preceded by a Phe or a Tyr residue. The second segment, defined by the consensus sequence (S/T)Lx 2 h (51), can be extended to hx 2 SLx 2 h among the group of TP-primed DNA polymerases. In this group, a Ser residue is present in 14 of 17 DNA polymerases, and Leu is invariantly present in all the sequences reported.
By site-directed mutagenesis in ø29 DNA polymerase, we have recently shown that the Phe residue of the Exo II motif, and the Ser and Leu residues of the hx 2 SLx 2 h motif are involved in ssDNA binding during 3Ј-5Ј exonucleolysis (51). Here we study the importance of these three latter residues for the other specific functions intrinsic to the mechanism used by ø29 DNA polymerase to replicate the viral genome. Six ø29 DNA polymerase mutant derivatives (F65Y, F65S, S122T, S122N, L123T, and L123A), overexpressed and purified as described (51), were analyzed using a variety of in vitro assays corresponding to the different stages of the TP-primed ø29 DNA replication mechanism.
Protein-primed TP-DNA Replication with the Mutant DNA Polymerases-ø29 DNA replication involves TP-primed initiation at both terminal origins, a special activity of ø29 DNA polymerase, that catalyzes the template-directed formation of a covalent complex between the viral TP and 5Ј-dAMP (initiation step) and the subsequent elongation (via strand displacement) of the initiation complex to produce full-length ø29 DNA (reviewed in Refs. 1 and 2).
Using a minimal replication system based on ø29 TP-DNA, ø29 DNA polymerase, and ø29 TP (62), the efficiency displayed by mutants F65S, S122T, and S122N was reduced 10-, 4-, and 4-fold, respectively, (see Table I) relative to that of the wildtype polymerase, although in all cases, elongation reached unit-length ø29 TP-DNA ( Fig. 2A). Indeed, when such an assay  74); Streptococcus pneumoniae phage Cp1-DNA polymerase (Cp1; GenBank TM data base accession number Z47794), and phage RB69 DNA polymerase (RB69) (48). Numbers between slashes indicate the amino acid position relative to the N terminus of each DNA polymerase. Numbers in parentheses indicate the length of the intervening amino acid sequence. The highly conserved Asn, Phe/Tyr, and Asp residues of the hhxANx 2-3 F/YDx 2 Ahh motif are shown in white letters over a black background. The conserved aromatic (A) and hydrophobic (h) residues of the motif are shown in boldface over a gray background; Ser/Thr and Leu residues of the hx 2 SLx 2 h motif are shown in white letters over a black background. The hydrophobic (h) residues of the motif are shown in boldface over a gray background. Residues studied in this paper are indicated with asterisks.
was carried out under more physiological conditions in which appropriate amounts of the four bacteriophage ø29 DNA replication proteins TP, DNA polymerase, ø29 dsDNA binding protein, and ø29 ssDNA binding protein are required to amplify limited amounts of ø29 TP-DNA molecules (59), essentially no amplification products were detected with these three mutant ø29 DNA polymerases ( Fig. 2B; see Table I).
Strand Displacement Capacity of Mutant ø29 DNA Polymerases-As mentioned above, ø29 DNA polymerase has to couple processive DNA synthesis to strand displacement to efficiently replicate ø29 TP-DNA. In order to analyze whether the defects in replicating ø29 TP-DNA that were exhibited by ø29 DNA polymerase mutants F65S, S122T, and S122N were specifically due to a defective strand displacement capacity, we performed a primed-M13 replication assay in which ø29 DNA polymerase starts polymerization from the 3Ј-OH group of a DNA oligonucleotide. The first replication round does not require strand displacement, but once it is completed, the polymerase encounters the 5Ј-end of the primer; the next rounds of replication (rolling circle-type) require, after this point, an active strand displacement. As shown in Fig. 3, the size of the replication products obtained with the mutant ø29 DNA polymerases (severalfold the length of the M13 DNA template) reveals that they are not specifically affected in strand displacement DNA synthesis. However, the global efficiency displayed by mutants F65S, S122N, and S122T was reduced 4-, 2.2-and 1.7-fold, respectively, with respect to that of the wild-type enzyme (see Table I). Such a defect could be pointing to a more general defect in the interaction with DNA (see below).
TP-dAMP Formation (Initiation of ø29 DNA Replication)-To start ø29 DNA replication, the DNA polymerase has to form an heterodimer with a free TP molecule to catalyze the template-directed insertion of dAMP onto the hydroxyl group of Ser 232 of the TP, i.e. the initiation reaction. Surprisingly, mutants F65S and S122N displayed a significantly reduced (3and 4-fold, respectively) TP-primed initiation capacity ( Fig. 4 and Table I). Because the initiation of ø29 DNA replication is a template-directed event (3), the initiation defect could be the consequence of a weak affinity for the template DNA. This possibility could be tested because ø29 DNA polymerase can catalyze the deoxynucleotidylation of TP in the absence of template (63). Under these conditions, the activity of the mutant derivatives of ø29 DNA polymerase was similar to or even lower than (in the case of mutant S122N) that corresponding to a templated TP-dAMP formation ( Fig. 4 and Table I). Thus, the low initiation capacity displayed by mutants points to a more specific defect, i.e. the interaction with the primer protein.
The ability of mutants F65S and S122N to interact with the TP was tested by using an interference assay (see under "Materials and Methods") in which wild-type and mutant polymerases compete for TP. As a control, ø29 DNA polymerase mutant D249E, which is catalytically inactive but has an intact capacity to interact with TP, was used (15). As expected (shown in Fig. 5), the inhibition profile obtained with this mutant paralleled the theoretical one. Contrarily, the wild-type enzyme was poorly competed by F65S and S122N mutant polymerases, probably reflecting a defective interaction with TP, and only mutant F65S was able to interfere the wild-type activity,  2. ø29 TP-DNA replication and amplification by point mutants of ø29 DNA polymerase. A, replication of ø29 TP-DNA. The assay was carried out as described under "Materials and Methods," in the presence of 5 ng of either wild-type or mutant ø29 DNA polymerase. After incubation for 10 min at 30°C, relative activity values were calculated (see Table I), and the length of the synthesized DNA was analyzed by alkaline agarose gel electrophoresis. The migration position of unit-length ø29 DNA is indicated. B, amplification of ø29 TP-DNA. The assay was carried out as described under "Materials and Methods," in the presence of 5 ng of ø29 TP-DNA, 5 ng of either wild-type or mutant ø29 DNA polymerases, 5 ng of TP, and 10 g of each ø29 dsDNA binding protein and ø29 ssDNA binding protein. After 90 min of incubation at 30°C, the reaction was stopped with 10 mM EDTA. The relative activity values were calculated (see Table I), and the length of the synthesized DNA was analyzed by alkaline agarose gel electrophoresis. but only using a high ratio of mutant versus wild-type enzyme (see Fig. 5).
To confirm these results, the interaction of these mutant polymerases with TP was directly analyzed by ultracentrifugation, as described under "Materials and Methods." Western blot analysis was used to identify the peaks corresponding to a TP/DNA polymerase heterodimer (97 kDa) and the free monomers of TP (31 kDa) and DNA polymerase (66 kDa). As can be observed in Fig. 6A, whereas the wild-type polymerase formed an heterodimer with TP, DNA polymerase mutant S122N and TP eluted separately as monomers (Fig. 6B), indicating a TP binding defect. Under these assay conditions, mutant F65S was able to form a complex with TP (not shown), suggesting that the interaction defect is not as drastic as that displayed by mutant S122N.

Affinity of Mutant ø29 DNA Polymerases for Primer/Template Structures-
The affinity for primer/template DNA molecules of wild-type and ø29 DNA polymerase mutants at residues Phe 65 , Ser 122 , and Leu 123 was directly studied using gel retardation assays, as described under "Materials and Methods." Under these conditions, the wild-type ø29 DNA polymerase produces a single retardation band using a labeled hybrid 15-mer/21-mer molecule (see Fig. 7) that has been interpreted as an enzyme-DNA complex competent for polymerization (16).
As also shown in Fig. 7, ø29 DNA polymerase mutant L123A had an affinity similar to that of the wild-type for primer/ template structures; mutants F65Y and L123T showed a reduced binding efficiency; and mutants F65S, S122T, and FIG. 3. Strand displacement coupled to M13 DNA replication by sitedirected mutants of ø29 DNA polymerase. Replication of primed M13 DNA was carried out as described under "Materials and Methods" using 80 M dNTPs and 25 ng of either wild-type or mutant ø29 DNA polymerases. After incubation for the indicated times at 30°C, relative activity values were calculated from dNMP incorporation (see Table I) and the length of the synthesized DNA was analyzed by alkaline 0.7% agarose gel electrophoresis and autoradiography. The position of full-length M13 DNA is shown at the right. Interference assay for TP binding. The assay of templateindependent formation of TP-dAMP complex by the wild-type DNA polymerase was performed in the presence of mutant polymerases F65S or S122N. An inactive DNA polymerase mutant, able to interact with TP, the D249E mutant (15), was used as a control of 100% competition that paralleled the theoretical profile. The TP-dAMP formed in the different competition conditions relative to that formed in the absence of competition (100%) is indicated. S122N were unable to produce any DNA retardation (see Table  I). Therefore, although Phe 65 , Ser 122 , and Leu 123 residues have been described as ssDNA ligands, stabilizing and orienting the primer terminus during 3Ј-5Ј exonucleolysis (51), it is apparent that mutations introduced at such positions also affect the stabilization of the DNA during polymerization.
The defective interaction displayed by mutants F65S and S122N with both TP and DNA primers could indicate that they occupy a common region of the ø29 DNA polymerase. To address this question, we analyzed whether the binding of TP is compatible with the binding of a primer terminus by the same DNA polymerase molecule. As shown by gel retardation assays (Fig. 8A, left panel), the capacity of the DNA polymerase and the TP to bind DNA is lost after formation of an heterodimer. In addition, the intrinsic DNA binding capacity of TP (52) is reduced after the formation of the heterodimer with ø29 DNA polymerase. In this assay, a significant portion of the labeled oligonucleotide remained single-stranded (see the legend to Fig. 8A). Because metal ions are included in this assay, the observed variation in the amount of this material is due to 3Ј-5Ј exonucleolysis by ø29 DNA polymerase. In agreement with these data, formation of a TP/DNA polymerase heterodimer prevented both exonucleolysis and polymerization on a DNA primer strand (Fig. 8A, right panel), even upon addition of a The assay was carried out as described under "Materials and Methods," using 5Ј-labeled 15mer/21-mer synthetic hybrid, in the presence of 5 ng of either wild-type or mutant ø29 DNA polymerases. Samples were analyzed by gel electrophoresis in the conditions described by Méndez et al. (16). The bands corresponding to free dsDNA (15/21 hybrid) and its complex with DNA polymerase (dsDNA/pol complex) are indicated. The band detected below free dsDNA (indicated with an arrow) corresponds to nonhybridized 15-mer oligonucleotide. high concentration of dNTPs (400 nM). These results support our hypothesis of a unique primer binding site to be nonsimultaneously occupied by TP or a DNA primer. We also analyzed whether binding of TP, which appears to require Ser 122 , interferes with the binding of ssDNA at the 3Ј-5Ј exonuclease active site of ø29 DNA polymerase. As shown in Fig.  8B (left panel), the retardation band obtained with the DNA polymerase alone, which has been interpreted as a stable interaction of ssDNA at the 3Ј-5Ј exonuclease active site (50), is not produced when a DNA polymerase/TP heterodimer is used. This suggests that TP binding occludes the access to the 3Ј-5Ј exonuclease binding cleft. In addition, when Mg 2ϩ ions were added to activate exonucleolysis, less than 25% of the ssDNA molecules were degraded by the heterodimer in comparison with the ø29 DNA polymerase alone.
Transition from Protein Priming to DNA Priming in ø29 DNA Replication-During in vitro replication of ø29 TP-DNA, a certain amount of nonelongated initiation complexes (TP-dAMP) and partially elongated abortive products (TP-(dNMP) 2-8 ) are produced as a consequence of a rate-limiting transition step from the initiation to the elongation mode (4). Such abortive products could be detected by using a 3Ј-5Ј exonuclease deficient ø29 DNA polymerase (4,25,26). It was interesting to study the transition step of mutants F65S and S122N, because both have a strongly reduced 3Ј-5Ј exonuclease activity (51) and a defective TP interaction (this paper). Thus, truncated elongation assays were carried out using a high amount of both wild-type or mutant DNA polymerase, to favor the initiation reaction, i.e. TP-dAMP formation (see under "Materials and Methods"). Providing dATP as the only nucleotide, the wild-type enzyme gave rise to TP-dAMP and TP-(dAMP) 2 products, as expected, taking into account its exonucleolytic activity reported to occur on the third nucleotide added to the TP (64), whereas mutant F65S was able to produce up to TP-(dNMP) 4 products (Fig. 9). With dATP, dGTP, and dTTP, elongation with the wild-type polymerase mainly occurred up to TP-(dNMP) 8 , which is the expected size when replication starts from the left origin (see sequence in Fig. 9). Under these conditions, mutant F65S carried out transition more efficiently than the wild-type enzyme. This apparent stimulation of the transition capacity of mutant F65S could be explained considering its strongly reduced 3Ј-5Ј exonuclease activity (51). However, when the four nucleotides were provided, to allow the DNA polymerase to fully replicate ø29 TP-DNA, mutant F65S still gave rise to abortive transition products up to TP-(dNMP) 8 not detected with the wild-type enzyme (Fig. 9). It has been FIG. 8. A, binding of TP to the DNA polymerase prevents binding and polymerization on DNA primer/template structures. The assay was carried out as described under "Materials and Methods," using 5Ј-labeled 15-mer/21-mer synthetic hybrid, in the presence of either 40 ng of wild-type ø29 DNA polymerase, 20 ng of TP, or 60 ng of TP/DNA polymerase complex. Samples were analyzed by gel electrophoresis in the conditions described by Méndez et al. (16). The bands corresponding to the free dsDNA and to the dsDNA complexed with either DNA polymerase or TP are indicated. The fastest migrating band correspond to a portion of nonhybridized labeled ssDNA. The right panel shows the inhibition of the polymerization capacity of ø29 DNA polymerase by addition of TP. The assay was carried out as described under "Materials and Methods," using the same conditions described above, incubating for 2 min at 4°C in the presence of 400 nM each of the four dNTPs. Samples were analyzed by 8 M urea-20% polyacrylamide gel electrophoresis and autoradiography. Polymerization is detected as an increase in the size (15-mer) of the 5Ј-labeled primer. B, binding of TP to the DNA polymerase prevents binding and 3Ј-5Ј exonucleolysis on ssDNA. The binding assay (left panel) was carried out under the same conditions described in A but using 5Ј-labeled 15-mer as ssDNA substrate. The bands corresponding to free ssDNA and to the DNA polymerase/ssDNA complex are indicated. 3Ј-5Ј exonucleolysis of ssDNA (right panel) was carried out as described under "Materials and Methods," by adding Mg 2ϩ ions and incubating for 2 min at 4°C. Samples were analyzed by 8 M urea-20% polyacrylamide gel electrophoresis and autoradiography. The position of different degradation intermediates of the 15-mer substrate is indicated. Abbreviations used are as follows: DNA pol, wildtype ø29 DNA polymerase; DNA pol/TP, wild-type ø29 DNA polymerase/TP heterodimer.

ø29 DNA Polymerase Ser 122 Mediates Interaction with TP
shown that during insertion of the first nine nucleotides, ø29 DNA polymerase remains bound to TP as an heterodimer (4), but both proteins dissociate once the next nucleotide has been incorporated. When this rate-limiting step has been completed, the stability of the DNA polymerase depends exclusively on the interactions with the DNA. Thus, it is tempting to speculate that the transition abortive bands appearing with the F65S mutant are produced as a consequence of its poor DNA binding capacity.
When either dATP alone or dATP, dGTP, and dTTP were provided, mutant S122N gave rise to a similar proportion of TP-(dAMP) 2 versus TP-dAMP products relative to the wild-type (Fig. 9), despite the fact that it has a very reduced 3Ј-5Ј exonuclease activity (51). Therefore, the polymerization advantage expected as a consequence of a decreased 3Ј-5Ј exonuclease activity in mutant S122N is probably counteracted by the defective interaction with TP displayed by this mutant that reduces the yield of all TP-(dNMP) n intermediates requiring a stable TP/DNA polymerase interaction. On the contrary, when the four nucleotides were provided, those TP-(dNMP) n intermediates that completed the transition stage, could be elon-gated via a facilitated TP/DNA polymerase dissociation, producing a good yield of fully replicated ø29 TP-DNA molecules. In agreement with that, and despite a reduction in its DNA binding capacity, mutant S122N did not produce the aborted transition products observed with mutant F65S. DISCUSSION ø29 DNA polymerase shares with other eukaryotic-type DNA polymerases (family B), Pol I-like DNA polymerases (family A), and Pol III DNA polymerases (family C) several regions of amino acid similarity at their 3Ј-5Ј exonuclease domain (11,12,18,19,49,51). Three of these regions, the so-called Exo I, II, and III motifs (18), contain the five catalytic amino acid residues acting as metal ligands, as they were originally identified by crystallographic studies of Pol Ik (20). The predictions of an evolutionarily conserved 3Ј-5Ј exonuclease active site has been confirmed with the resolution of the crystal structure of an N-terminal fragment of T4 DNA polymerase (47) and that of phage RB69 DNA polymerase complexed with ssDNA and divalent metal ions (48). Recently, a highly conserved aromatic (Phe or Tyr) residue, located at the Exo II motif, and a Ser and FIG. 9. Analysis of the transition products of ø29 DNA replication carried out by wild-type or mutant ø29 DNA polymerases. The assay was carried out as described under "Materials and Methods," using the different dNTPs at the indicated concentration. Samples were analyzed by high-resolution SDS-PAGE. The first 10 nucleotides of the left end of ø29 genome are depicted in the upper panel. The length of different transition products and the position corresponding to full-length TP-DNA are indicated at the right. In all cases, wild-type and mutant enzymes produced an extra abortive product corresponding to TP-(dNMP) 16 . This stop, previously described during replication of TP-DNA with exonuclease-deficient mutant polymerases (26), occurs after the transition event.
FIG. 10. Three-dimensional structure prediction for TP/DNA polymerase interaction. This prediction is based on the structure of bacteriophage RB69 DNA polymerase (18), the proofreading and polymerization domains of which are homologous to those of ø29 DNA polymerase. A, amino acid residues 1-102 are uncolored because they are not present in ø29 DNA polymerase. Residues 103-339, conforming the 3Ј-5Ј exonuclease domain, are shown in blue. The 5Ј-3Ј polymerization domain, formed by amino acids 340 -903, is depicted in green. In our model, the terminal protein (TP), in gray, which acts as an initiation primer for ø29 DNA polymerase, is predicted to bind the cleft formed by the thumb, palm and fingers. The polymerization catalytic core formed by RB69 DNA polymerase residues Asp 411 , Asp 621 , and Asp 623 (Asp 250 , Asp 456 , and Asp 458 of ø29 DNA polymerase) is shown as red spheres, whereas those residues involved in primer terminus stabilization at the polymerization active a Leu residue that form the S/TLx 2 h motif, located between the Exo II and Exo III motifs (51), were shown to be functionally important for a stable interaction with ssDNA during proofreading of DNA polymerization errors (51). In this study, we analyzed the importance of these three ø29 DNA polymerase residues for the synthetic activities, i.e. protein priming and DNA polymerization.
As was mentioned above, the first step of ø29 DNA replication requires the interaction between a DNA polymerase molecule and a free molecule of TP and the further recognition of the origins of replication, located at either DNA terminus, by the heterodimer. Initiation of replication occurs by the covalent linkage of dAMP to the hydroxyl group of Ser 232 of TP, in a reaction catalyzed by ø29 DNA polymerase (reviewed in Refs. 1 and 2) and directed by the second nucleotide at the 3Ј-end of the template. The first nucleotide is recovered by a sliding back mechanism (3), and the DNA polymerase remains bound to the TP until 9 dNMP residues have been incorporated (transition step) (4). Afterward, dissociation of DNA polymerase and TP takes place, and elongation occurs coupled to strand displacement, giving rise to fully replicated ø29 DNA molecules.
Interestingly, F65S mutant and those changes introduced at Ser 122 residue displayed a low efficiency in replicating and amplifying ø29 TP-DNA. Such a defect was not due to a defective strand displacement as deduced from their capacity to carry out rolling circle elongation using primed M13 DNA as template, strongly supporting the hypothesis that only those residues acting either directly or indirectly as metal ligands at the 3Ј-5Ј exonuclease active site are involved in strand displacement (25,26,49,50).
The analysis of the protein-primed initiation step revealed that mutant polymerases F65S and S122N were very affected in carrying out such an activity, both in the absence and in the presence of TP-DNA. This defect could be directly related with either a moderately (F65S mutant) or highly (S122N mutant) decreased capacity to form a stable TP/DNA polymerase heterodimer. Interestingly, mutants F65S, S122T, and S122N also displayed a poor capacity to interact with a DNA primer/template structure. The alteration of the capacity to interact with these two different primer structures (TP and DNA) as a consequence of the mutations introduced was apparent when the transition between protein and DNA priming, inherent to the ø29 DNA replication mechanism, was studied. Thus, mutant F65S was able to produce an efficient extension of the TP-dAMP product in the transition stage. However, its reduced capacity to interact with the newly created DNA primer leads to abortive transition products. Contrarily, the poor capacity to maintain the interaction with TP displayed by mutant S122N makes it difficult to reach the transition stage. At this point, a weakened TP/DNA polymerase interaction would facilitate entrance into the elongation stage. Little is known about the regions of the DNA polymerase that are making contacts with the TP. To date, only the conserved motif Tx 2 GR of ø29 DNA polymerase, the counterpart of which in Pol Ik (16) and RB69 (48) forms part of the palm subdomain at the polymerization domain, has been involved in interaction with TP (16) (see the following section). On the other hand, the fact that the Cterminal domain (polymerization domain) is unable to interact stably with the TP by itself, dropping up to 1000-fold its initiation capacity with respect to that of the complete enzyme (65), suggested that the N-terminal domain (3Ј-5Ј exonuclease domain) should contribute to TP binding. Here, we provide the first direct evidence involving a single residue of the N-terminal domain (Ser 122 of the hx 2 SLx 2 h motif) of ø29 DNA polymerase, in TP binding.
The interaction of ø29 DNA polymerase with a TP primer was modeled. Attempts to obtain ø29 DNA polymerase crystals adequate for x-ray diffraction analysis have not been successful so far. Therefore, our working models are based on extrapolation to the resolved three-dimensional structure of other DNA polymerases from family A (Pol Ik (66,67), Thermococus aquaticus DNA polymerase (68,69), Bacillus stearothermophilus DNA polymerase (70,71), and T7 DNA polymerase (72)) and from family B (RB69 DNA polymerase (48)). All of them showed a similar bimodular organization, with an N-terminal domain containing the 3Ј-5Ј exonuclease active site (or a vestige by loosing of the catalytic carboxylates, as in the case of T. aquaticus and B. sterarothermophilus DNA polymerases), and a C terminus containing the 5Ј-3Ј polymerization active site. However, when the polymerization domains of Pol Ik and RB69 DNA polymerase are overimposed, their exonuclease domains lie on opposite sides with respect to the polymerase active site (48), pointing to structural/functional differences between these two groups of DNA polymerases. As has been described above, ø29 DNA polymerase belongs to the eukaryotic-type group of DNA-dependent DNA polymerases (7), and therefore, its overall structure is likely similar to that of RB69 DNA polymerase.
Taking into account that ø29 DNA polymerase uses the same 5Ј-3Ј polymerization active site, located at its C-terminal domain, for the two synthetic activities (TP-primed initiation and DNA polymerization; reviewed in Ref. 5) and that the results presented in this paper demonstrate that it is unable to simultaneously bind both TP and DNA primers, suggesting a common binding site, an important question is how the structure of ø29 DNA polymerase is adapted to sequentially use both primers. Based on the structure of RB69 DNA polymerase (48), we propose that TP (represented in gray in Fig. 10) interacts with the DNA polymerase through its DNA primer cleft formed by the thumb, palm, and fingers subdomains (depicted in green). More than 40 residues of such a cleft have been shown to make a network of hydrogen bonds, ion pairs, and van der Waals contacts to stabilize the DNA/DNA polymerase interaction. Among them, the two residues (shown in Fig. 10A in dark blue spheres), belonging to the highly conserved motif Tx 2 GR (48), are of particular importance in the stabilization of the primer terminus at the 5Ј-3Ј polymerization active site. This assumption is based on site-directed mutagenesis studies of this motif of ø29 DNA polymerase (16) and crystallographic data showing that the arginine residue of the motif is involved in a direct interaction with the DNA primer terminus, both in T7 (72) and B. stearothermophilus (71) DNA polymerases. Interestingly, in the case of ø29 DNA polymerase, mutations at the Thr 434 and Arg 438 residues of motif Tx 2 GR caused a severe reduction in the capacity to interact not only with a DNA primer/template structure, but also with TP (16), as discussed above, which could be due to a defective interaction of the TP primer terminus (the hydroxyl group of Ser 232 in the case of ø29 TP) at the polymerization active site, strongly supporting the proposed location of the TP within the DNA polymerase structure. On site, Thr 588 and Gln 592 (Thr 434 and Arg 438 of ø29 DNA polymerase) are shown in dark blue. RB69 DNA polymerase residues Phe 221 , Ser 289 , and Leu 290 , homologous to the ø29 DNA polymerase residues Phe 65 , Ser 122 , and Leu 123 , studied in this paper, are shown in magenta, yellow, and orange, respectively. B, in this scheme, we propose that a unique primer binding site is initially occupied by the TP primer molecule during the initiation and transition stages of ø29 DNA replication. Afterward, during the elongation stage, the same cleft serves to hold the DNA primer/template structures, where residue Ser 122 seems to act as a DNA ligand during both DNA synthesis and proofreading of mismatched primer termini. The code color used is as described for A. the other hand, the positioning of the TP to reach the catalytic triad (formed by the aspartate residues of motifs Dx 2 SLYP and YGDTDS of eukaryotic-type DNA polymerases (11), depicted as red spheres in Fig. 10A, outlining the portion of the polymerization active site and their close vicinity to the Tx 2 GR motif), probably occludes the entrance to the ssDNA binding cleft (the length of which can bury the four 3Ј-terminal nucleotides of a ssDNA) (48) at the 3Ј-5Ј exonuclease domain (depicted in blue in Fig. 10). Such an occlusion would explain the incapacity of the DNA polymerase to bind a ssDNA 15-mer once it forms an heterodimer with a TP molecule. This fact could be physiologically relevant, because the heterodimer would prevent the exonucleolytic degradation of the 3Ј-ends of ø29 TP-DNA by ø29 DNA polymerase before replication starts.
The Phe residue of the Exo II motif (Phe 221 in RB69 DNA polymerase, represented in magenta spheres in Fig. 10) is located at the bottom of the 3Ј-5Ј exonuclease cleft and, therefore, far enough to preclude direct contacts with the primer structure (TP or DNA). The defective TP binding displayed by a mutant at the homologous residue (Phe 65 ) of ø29 DNA polymerase could be due to local alteration of the structure, although we cannot rule out the possibility that the structure of TP, or specific differences in the structure of TP-primed DNA polymerases, could favor such a contact. On the other hand, the fact that F65S and L123T mutants were affected in the interaction with a DNA primer/template (paralleling their low capacity to bind ssDNA; 51) could suggest that once the DNA polymerase is bound to a primer/template structure, the 3Јend of the former is first stabilized at the exonuclease active site by the ssDNA ligands, somehow mimicking the proofreading mode following misincorporation. Indeed, under gel retardation assay conditions, analysis of the retarded band revealed exonucleolytic degradation of the primer (17).
RB69 DNA polymerase residue Ser 289 (shown in yellow in Fig. 10), the counterpart of Ser 122 of ø29 DNA polymerase (this paper) is located 25 Å away from the polymerization active site, but facing the primer binding clef. This situation would allow this residue to make stable contacts with the TP during the initiation of ø29 DNA replication. In agreement with that, elimination of the hydroxyl group of Ser 122 (S122N mutant), largely reduced the capacity to interact with TP. As depicted in Fig. 10B, during the transition step, DNA polymerase contacts with TP are replaced by interactions with the nascent DNA to start normal elongation. Thus, the role of Ser 122 as a TP ligand would be transient, becoming a DNA ligand during the elongation stage. Moreover, the positioning of this residue just at the entry of the ssDNA binding cleft of the 3Ј-5Ј exonuclease domain allows it to be also involved in stabilization of the mismatched primer terminus at the 3Ј-5Ј exonuclease active site when corrections of elongation errors are required (51).