DNA-independent deoxynucleotidylation of the phi 29 terminal protein by the phi 29 DNA polymerase.

In this paper, we show that the phi 29 DNA polymerase, in the absence of DNA, is able to catalyze the formation of a covalent complex between the phi 29 terminal protein (TP) and 5'-dAMP. Like the reaction in the presence of phi 29 DNA, TP.dAMP complex formation is strongly dependent on activating Mn2+ ions and on the efficient formation of a TP/DNA polymerase heterodimer. The nature of the TP-dAMP linkage was shown to be identical (a O-5'-deoxyadenylyl-L-serine bond) to that found covalently linking TP to the DNA of bacteriophage phi 29, indicating that this DNA-independent reaction actually mimics that occurring as the initiation step of phi 29 DNA replication. Furthermore, as in normal TP-primed initiation on the phi 29 DNA template, this novel reaction showed the same specificity for TP Ser232 as the OH donor and the involvement of the YCDTD amino acid motif, highly conserved in alpha-like DNA polymerases. However, unlike the reaction in the presence of phi 29 DNA, the DNA-independent deoxynucleotidylation of TP by the phi 29 DNA polymerase did not show dATP specificity, being possible to obtain any of the four TP.dNMP complexes with a similar yield. This lack of specificity together with the poor efficiency of this reaction at low deoxynucleoside triphosphate (dNTP) concentration reflect a weak, but similar stability of the four dNTPs at the phi 29 DNA polymerase dNTP-binding site. Thus, the presence of a director DNA would mainly contribute to stabilizing a complementary nucleotide, giving base specificity to the protein-primed initiation reaction. According to all these data, the novel DNA polymerase reaction described in this paper could be considered as a "non-DNA-instructed" protein-primed deoxynucleotidylation.

In this paper, we show that the $29 DNA polymerase, in the absence of DNA, is able to catalyze the formation of a covalent complex between the $29 terminal protein (TP) and 6'-dAMP. Like the reaction in the presence of $29 DNA, TP-dAMP complex formation is strongly dependent on activating Mn2+ ions and on the efficient formation of a TP/DNA polymerase heterodimer. The nature of the TP-dAMP linkage was shown to be identical (a 0-5'-deoxyadenylyl-L-serine bond) to that found covalently linking TP to the DNA of bacteriophage $29, indicating that this DNA-independent reaction actually mimics that occurring as the initiation step of $29 DNA replication. Furthermore, as in normal TP-primed initiation on the $29 DNA template, this novel reaction showed the same specificity for TP SerZs2 as the OH donor and the involvement of the YCDTD amino acid motif, highly conserved in a-like DNA polymerases. However, unlike the reaction in the presence of $29 DNA, the DNA-independent deoxynucleotidylation of TP by the $29 DNA polymerase did not show dATP specificity, being possible to obtain any of the four TP-dNMP complexes with a similar yield. This lack of specificity together with the poor efficiency of this reaction at low deoxynucleoside triphosphate (dNTP) concentration reflect a weak, but similar stability of the four dNTPs at the $29 DNA polymerase dNTP-binding site. Thus, the presence of a director DNA would mainly contribute to stabilizing a complementary nucleotide, giving base specificity to the protein-primed initiation reaction. According to all these data, the novel DNA polymerase reaction described in this paper could be considered as a "non-DNA-instructed" protein-primed deoxynucleotidylation.
One of the initiation mechanisms for linear DNA replication is based on the use of specific proteins as primers. A stable and specific interaction of these proteins with a specialized DNA polymerase allows synthesis to be directed by the 3"terminal base of the linear DNA template strand. Most likely, as a result of this mechanism, several groups of viruses and linear plasmids have terminal proteins covalently linked to the B'-ends of their linear genomes. Since the first proposal * This work was supported by Research Grant 5R01 GM27242-12 from the National Institutes of Health, Grant PB87-0323 from the Direcci6n General de Investigacibn Cientifica y TBcnica, and an instutional grant from the Fundaci6n Ram6n Areces. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisernent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ Postdoctoral Fellow from the Spanish Research Council.
ll To whom reprint requests should be addressed.  (3), with a specific viral protein of 31 kDa covalently linked at the two 5'-ends through a phosphoester bond between the OH group of Ser232 and dAMP, the 5"terminal nucleotide at both DNA ends (4). By using a highly purified 429 DNA replication system, it has been demonstrated that a free molecule of terminal protein (TP)' becomes covalently bound to 5'-dAMP in a reaction catalyzed by the 429 DNA polymerase in the presence of the 429 DNA-TP template and 5'-dATP; subsequently, the TP dAMP complexes formed at both DNA ends are elongated by the 429 DNA polymerase in a highly processive way by a strand-displacement mechanism until completion of both DNA strands (5, 6). By site-directed mutagenesis in the 429 DNA polymerase, we have demonstrated that critical residues and amino acid motifs forming the polymerization active site are also critical for the protein-primed initiation activity. These structure-function studies allowed us to conclude that the 429 DNA polymerase is structured in two separate domains, with the N-terminal portion containing the 3' + 5' exonuclease activity, whereas both synthetic activities, i.e. protein-primed initiation and DNA polymerization, are located in the C-terminal domain (7-9).
In this paper, we show that the 429 DNA polymerase, in the absence of DNA, is able to covalently link any of the four dNMPs to 429 TP. The specificity of this reaction, in terms of nucleotide selection together with the effect of site-directed mutants in both TP and DNA polymerase, will be also discussed in the context of a structural model for the interactions occurring during TP-primed initiation of linear DNA replication.
Formation of TP-Nucleotide Covalent Complex-The standard incubation mixture contained, in 25 pl, 50 mM Tris-HC1, pH 7.5, 1 mM The abbreviations used are: TP, terminal protein; dNMPs and dNTPs, deoxynucleoside mono-and triphosphates, respectively; SDS, sodium dodecyl sulfate. Protein Deoxynucleotidylation dithiothreitol, 5% glycerol, 1 mM MnC12, 0.25 p~ [cY-~'P]~ATP, and 125 ng of both 629 TP and either wild-type or exonuclease-deficient 629 DNA polymerase, either in the presence or absence of $29 DNA-TP as template. When indicated, particular components were omitted. After incubation for the indicated time a t 30 "C (different temperatures are otherwise indicated), the reaction was stopped by adding EDTA to 10 mM and SDS to 0.1%, filtered through Sephadex G-50 spun columns in the presence of 0.1% SDS, and analyzed by SDSpolyacrylamide gel electrophoresis and autoradiography. Quantitation was done by excising from the gel the radioactive band corresponding to the TP-nucleotide complex and counting the Cerenkov radiation or by densitometry of the autoradiographs. To study both the affinity and specificity for different nucleotide substrates, the reaction was carried out essentially as described, but using the indicated concentrations of any of the four [cY-~'P]~NTPs or particular [a-32P]NTPs. To further study the specificity of the reaction in the absence of DNA, different mutants in both TP and $29 DNA polymerase (described above) were used, in the amounts indicated, instead of the corresponding wild-type proteins.
Identification of the Amino Acid Involved in Formation of TP-Nucleotide Linkage-TP-nucleotide complex formation was assayed as described for the standard assay using exonuclease-deficient 629 DNA polymerase and 0.5 p~ [cY-~'P]~ATP (5 pCi). After overnight incubation at 30 "C, the unreacted [cY-~'P]~ATP was removed by gel filtration on Sephadex G-50 spun columns, and the excluded material was subjected to SDS-polyacrylamide gel electrophoresis. The radioactive band running a t the TP position was cut out; eluted overnight at 37 "C with Tris-HC1, pH 7.5, 0.1% SDS; precipitated with 4 volumes of acetone; and resuspended in water. A sample was treated with 5.8 N, HCl for 2 h a t 110 "C in a sealed capillary tube, and the hydrolysis products were subjected to two-dimensional thin-layer analysis on cellulose plates (10 X 10 cm; Merck). In the first dimension, electrophoresis was performed in 2.5% formic acid, 7.8% acetic acid, pH 2.0, at 600 V. In the second dimension, chromatography in isopropyl alcohol/HCI/H*O (7015:15) was developed. Bromphenol blue was used as visual marker; and 0-phospho-L-serine, O-phospho-L-threonine, and 0-phospho-L-tyrosine (40 pg each; Sigma), detected by ninhydrin staining, were used as phosphoamino acid markers.
Identification of Nucleotide Moiety Linked to TP-A sample of the "P-labeled TP-nucleotide complex obtained as described above was subjected to alkaline hydrolysis in 0.5 M piperidine for 3 h a t 37 "C, followed by vacuum removal of the piperidine. Degradation products were identified by thin-layer chromatography on polyethyleneiminecellulose plates as described previously (15). The chromatogram was developed in lithium formate, pH 3.0. dADP, dAMP and Pi, obtained by partial digestion of [(Y-~'P]~ATP with alkaline phosphatase, were used as internal markers. In addition, unlabeled 5'-dAMP was run in parallel, and its position was determined by UV absorbance. Cosedimentatwn of TP.dAMP Complex and 629 DNA Polymerase-TP-nucleotide complex formation was assayed as described for the standard assay, except that 0.5 p~ [(u-~'P]~ATP (20 pCi), 750 ng of TP, and 600 ng of wild-type $29 DNA polymerase were used in a final volume of 100 pl. After overnight incubation at 25 "C, samples were layered on top of linear 15-30% glycerol gradients (5 ml) in buffer containing 50 mM Tris-HC1 pH 7.5,44 mM NaCl, and 20 mM (NH&SO. and centrifuged for 28 h at 290,000 X g at 0 "C. Fractions of -0.2 ml were collected from the bottom, and TP and 629 DNA polymerase were detected by radioimmunoassay as described (10). The position corresponding to the TP-dAMP complex was determined by SDS-polyacrylamide gel electrophoretic analysis of the different fractions, followed by autoradiography.

DNA-independent Deoxyadenylylatwn of 429 TP by 429
DNA Polymerase-When highly purified 429 TP and exonuclease-deficient 429 DNA polymerase were incubated with were <0.01% ( l a n e a), ~0.01% ( l a n e b), ~0 . 0 1 % ( l a n e c), and 5.1% ( l a n e d). 100% activity corresponds to 0.15 fmol of 32P-labeled TPnucleotide complex. B, the assay was carried out essentially as described for A, but in the absence or presence of 629 DNA-TP (0.5 pg) as template. After incubation for the indicated times at 30 "C, the samples were analyzed, and the reaction was quantitated as described for A. C, the assay was carried out as described for B either for 15 min in the absence or for 1.5 min in the presence of 629 DNA-TP (0.5 pg) as template at the indicated temperatures. Analysis and quantitation, expressed as a percentage of the maximum value obtained, were as described for A. 100% activity corresponds to either 0.96 fmol (DNA-independent reaction) or 0.125 pmol (DNA-dependent reaction) of 32P-labeled TP.dAMP complex. c). The requirement of this reaction for divalent metal ions showed a specificity similar to that observed for the initiation of 429 DNA replication: with Mn2+ as the best activator; other active metals were Co2+ and M e , with an activation (relative to Mn2+) of 5% for Co2+ and 1% for Mg2+ (data not shown). Furthermore, the labeling of the TP band was strongly inhibited either by anti-TP or anti-429 DNA polymerase sera (data not shown). NH: ions were shown to be critical for the stable formation of an equimolar stoichiometric complex between TP and DNA polymerase as the first step in the initiation of 429 DNA replication (16). In agreement with this, the 32P-labeled TP band obtained in the absence of DNA was drastically decreased when (NH4)2S04 was omitted (lane d ) , indicating that the efficiency of this DNA-independent reaction also requires a stable TP/429 DNA polymerase 'Esteban, J. A., Bernad, A., Salas, M., and Blanco, L. (1992) Biochemistry, in press. heterodimer. Therefore, this reaction can be defined as a DNA-independent deoxyadenylylation of TP by 429 DNA polymerase. Fig. 1B shows the time kinetics of the DNA-independent reaction compared to that carried out in the presence of 429 DNA-TP. The reaction was maintained roughly linear for at least 17 h at 30 "C. Considering this interval, the estimated rate of DNA-independent catalysis under the standard assay conditions described under "Materials and Methods'' is 0.02 fmol of TP-nucleotide complex formed in 1 min at 30 'C, which is -250-300-fold lower than that in the presence of 429 DNA-TP. As shown in Fig. IC, the protein-primed initiation of 429 DNA replication has an optimum temperature requirement of -30-37 "C, a condition that has been normally used in the standard assays; however, the DNA-independent reaction had an optimum of -15 "C, with the activity at higher temperature being decreased to -45 and 25% at 30 and 37 "C, respectively. These results, together with the low temperature optimum for protein-primed initiation when single-stranded 429 DNA ori sequences were used as template3 suggest that the differences in optimum temperature shown in Fig. 1C reflect the existence of a thermodynamically unfavored step during protein-primed initiation of the linear duplex 429 DNA, namely the opening of the 429 DNA replication origins.
Characterization of Linkage as Phsphserine Bond between TP and 5'-dAMP: Specificity of Se?32 as OH Donor-The novel nature of the DNA-independent DNA polymerase-catalyzed reaction described in this paper led us to characterize the product to determine whether or not it was similar to that obtained in the presence of DNA, the actual protein-primed initiation event. Thus, acid hydrolysis of the 32P-labeled TP band obtained in the absence of DNA yielded O-phosphoserine ( Fig. 2A), as was the case for both the 429 DNA-TP template (17) and the TP. dAMP complex formed as initiation of 429 DNA-TP replication (15); treatment of the 32P-labeled T P band with 0.5 M piperidine for 2 h at 37 "C, conditions that hydrolyze 429 DNA-TP linkage (17), released a radioactive product that co-migrated with 5'-dAMP when subjected to thin-layer chromatography on polyethyleneiminecellulose (Fig. 2B). As a control, no radioactivity at the position of 5'-dAMP appeared when the sample was not treated with piperidine. Therefore, these results allow us to characterize the product as 5'-dAMP covalently bound to TP through a phosphoserine linkage, i.e. a 0-5'-deoxyadenylyl-L-serine bond, identical to that found covalently linking TP to the DNA of bacteriophage 429. T P SerZ3' was shown to be the only serine (18 serine residues in a total of 266 amino acids) involved in the linkage to 429 DNA (4). Recent results have shown that a change of TP SerZ3' to threonine completely abolishes its priming capacity (12), i.e. the ability to be used as primer for initiation of 429 DNA-TP replication. Furthermore, a change of TP SerZ3* to cysteine had a reduced (0.7%), but detectable activity (13). As shown in Fig. 3A, when these two TP mutants (with the linking amino acid (SerZ3') changed to threonine (ThrZ3') or cysteine (CysZ3')) were assayed for DNA-independent deoxyadenylylation by 429 DNA polymerase, only the CysZ3' mutant was slightly active (4%), whereas the ThrZ3' mutant was essentially inactive (<0.05%). As a control, no labeled bands moving at the position of the TP-dAMP complex were obtained in the absence of 429 DNA polymerase (Fig. 3A). Therefore, as in the TP-primed initiation of 429 DNA-TP replication, this DNA-independent reaction shows the same specificity for TP SerZ3' as the OH donor. L. Blanco at 110 "C, and the hydrolysis products were subjected to two-dimensional thin-layer analysis as described under "Materials and Methods." Electrophoresis was run in the first dimension, and chromatography was run in the second one. The hydrolysis product of 32Plabeled TP was detected by autoradiography. 0-Phospho-L-serine, 0phospho-L-threonine, and 0-phospho-L-tyrosine, used as internal markers, were detected by ninhydrin staining. B, characterization of the nucleotide moiety linked to TP. A sample of 32P-labeled TP obtained in uitro under the conditions described for A was applied to a polyethyleneimine-cellulose plate before (-Pip) or after (+Pip) incubation with 0.5 M piperidine for 3 h at 37 "C. Chromatography was performed with 0.15 M lithium formate, pH 3.0. As internal marker, [cY-~*P]~ATP, untreated (MI) or treated (M2) with phosphatase, was used. In all cases, unlabeled 5'-dAMP, detected by UV absorbance, was used as an internal standard.
Involvement of YCDTD Motif of 429 DNA Polymerase-Based on the high conservation of this sequence among alike DNA polymerases (reviewed in Refs. 7 and 9; see also Ref. 18) and on site-directed mutagenesis studies on the YCDTD sequence of the 429 DNA polymerase (9, 14), we have proposed that this motif forms part of a critical domain involved not only in the elongation but also in the proteinprimed initiation activity of the 429 DNA polymerase. Thus, the latter activity was essentially abolished by the single mutations T457 to P and D458 to G and severely affected by the D456G mutation, although none of these mutations affected the 3' + 5' exonuclease activity and the ability of the enzyme to interact with T P (9). As shown in Fig. 3B, the 429 DNA polymerase mutants T457P and D458G were also inactive in the DNA-independent deoxyadenylylation of TP, whereas the activities of mutants Y454F and C455G were similar (93 and 30%, respectively) to those obtained in the presence of 429 DNA-TP (107 and 45%, respectively) (9). Therefore, comparison of the results obtained (this paper and Ref. 9) indicates that the effect of these mutations is independent of the presence of DNA, as it would be expected from the hypothesis (9, 19) that this motif is particularly involved in metal binding associated with the dNTP site. Interestingly, as also shown in Fig. 3B, the D456 to G mutation did not severely affect this DNA-independent reaction (39% lower than the wild type) in comparison with the data obtained from the analysis of the initiation of 429 DNA-TP replication (90% lower than the wild type) (9). A plausible explanation of this difference could be that a slight change in the metalcoordinated orientation of the incoming nucleotide due to the D45fi to G mutation may have a drastic effect when this nucleotide is constrained to base-pair with a DNA template.
TPedAMP Complex Formed in Absence of DNA Remains Attached to 429 DNA Polymerase-As has been demonstrated, the protein-priming mechanism of initiation at the ends of the linear 429 DNA molecule results in the formation of an initiation complex between TP and 5'-dAMP, which provides the 3'-OH group needed for elongation. 429 DNA polymerase is the only polymerase involved in both the initiation and elongation steps of 429 DNA replication; and therefore, dissociation of the TP/429 DNA polymerase heterodimer is likely to occur to replace the protein-protein interactions required for initiation by the protein-DNA interactions required for the elongation of the newly created DNA primer. This step, similar to a normal translocation step during processive DNA polymerization, was shown to be rate-limiting in the absence of viral protein p6, a 429 origin-binding protein probably involved in opening the replication origins (20,21). Therefore, it is likely that this specific step, named "transition" between protein-primed initiation and elongation (20), requires a conformational change in the 429 DNA polymerase to adapt the different nature of both protein and DNA primers. The possibility of forming a TP -dAMP complex in the absence of DNA raised the question of whether this complex would remain attached to 429 DNA polymerase or, on the contrary, as a result of the reaction, both proteins would dissociate. Under the assay conditions described under "Materials and Methods," -50% of 429 TP (31 kDa) interacted with 429 DNA polymerase (66 kDa) to form a 1:l complex (-90 kDa) as detected by glycerol gradient sedimentation (Fig. 4); and the remaining T P sedimented at the position of monomer (-31 kDa). As also shown in Fig. 4, the newly synthesized ~~'P-labeled TP cosedimented in the position corresponding to the TP/429 DNA polymerase heterodimer (-90 kDa), indicating the formation of a stable ternary complex (TP. dAMP/429 DNA polymerase). This result supports the hypothesis of a strong protein-protein interaction as the main factor blocking transition to elongation (20). Thus, the different steps of the DNA-independent deoxyadenylylation of TP by 429 DNA polymerase can be schematized as follows, .dATP A TP.dAMP/pol+ PPi a = formation of the TP/429 DNA polymerase (pol) heterodimer, b = binding of dATP to the 429 DNA polymerase dNTP-binding site, and c = deoxyadenylyl transfer reaction coupled to pyrophosphate release.
Substrate Specificity of "Non-DNA-instructed" Reaction-As has been reported, the protein-primed initiation of 429 DNA replication results in the formation of a T P .dAMP covalent complex (15). As shown in Fig. 5A, the reaction in the presence of the 429 DNA-TP template showed a strong specificity for the formation of TP-dAMP (-400-1000-fold more efficient than the formation of the other T P .dNMP complexes). This specificity can be explained either by: (i) a requirement for template instruction (T is the first 3'-base at both 429 DNA ends), as in a normal DNA-dependent DNA polymerase-catalyzed reaction; (ii) a special DNA-independent selection for dATP (in this case, the special characteristics of the serine-nucleotide linkage could require a strict specificity for dATP or involve a particular dATP-binding site for this reaction). The DNA-independent formation of the TPnucleotide complex described in this paper allowed us to distinguish between these two possibilities by comparing the efficiency of formation of the four possible TP dNMP complexes in the absence of the 429 DNA-TP template. Fig. 5A shows that this reaction had a similar efficiency in the formation of the four TP-dNMP complexes, with the exception of T P . dCMP, whose formation significantly varied with different [a-"PIdCTP batches used. Therefore, these results indicate that the specificity for dATP is only provided by the presence of the 429 DNA-TP template. No reaction was obtained when [cx-~*P]ATP or [cx-~*P]UTP was used as nucleotide substrate for both DNA-dependent and DNA-independent reactions (data not shown). Therefore, although a direct role of TP in substrate selection cannot be excluded, these results suggest a DNA-independent discrimination at the 429 DNA polymerase substrate-binding site, which is also able to distinguish between the ribo-and deoxynucleotide moieties.
As shown in Fig. 5B, when the dATP concentration (0.25 PM) used in the standard assay conditions was increased up to 250 p~, the rate of catalysis of the DNA-independent reaction was linearly increased; on the contrary, in the presence of 429 DNA-TP, the reaction reached a value close to the maximum rate at a relatively low (25 PM) dATP concentration. These results can be explained considering the strong differences in affinity for dATP at the 429 DNA polymerase active site, with the absence of DNA as a condition of minimum stability. Furthermore, the fact that the reaction obtained at a high dATP concentration gave similar values in either the presence or absence of DNA indicates that the proper orientation of the dNTP substrate is not as strictly template-dependent as the affinity and selection for the correct nucleotide.

DISCUSSION
There are several facts in support of the idea that the first step in the initiation of 429 DNA replication is the formation of a TP/DNA polymerase heterodimer. 1) Both genes are contiguous in the genetic map, probably being transcribed into a polycistronic mRNA (22,23); 2) a TP. DNA polymerase complex can be isolated from 429-infected cells (24); and 3) the two proteins interact in the absence of 429 DNA-TP with a 1:1 stoichiometry (16). As shown in this paper, the fact that the 429 DNA polymerase is able, in the absence of DNA, to catalyze the addition of a nucleotide to T P in a reaction very similar to that occurring as the initiation step of 429 DNA replication strongly emphasizes the functional importance of the TP-DNA polymerase interaction. However, the fact that this DNA-independent reaction can occur with any of the four dNTPs and that it is not very efficient at a low dNTP concentration fits with the general mechanism proposed by Kornberg (25) for DNA-dependent DNA polymerization onto DNA or RNA primers in which the presence of a director DNA would be necessary to stabilize a "complementary nucleotide," resulting in an increase of its resident time at the dNTP-binding site. If this model also applies to TP-primed initiation, a precise interaction with the terminal 3'-base of the template must be required to guarantee the proper selection of the initiating nucleotide, giving base specificity to the TP-primed initiation reaction.
To initiate synthesis, all known DNA polymerases require a free 3'-OH group, usually provided by a DNA or RNA primer. The synthesis and/or location of these primers is normally based on base complementarity with part of the template strand. In this way, primer and template are in close enough vicinity to be simultaneously recognized by the DNA polymerase to initiate template-directed DNA synthesis. The results shown in this paper agree with a model for protein priming in which both requirements (primer and template) are independently recognized in two differentiated steps. 1) The primer (TP) is initially recognized by the 429 DNA polymerase to form a functionally active heterodimer; and 2) once the T P is loaded in the 429 DNA polymerase molecule, recognition of the template strand at the replication origins should take place. Subsequently, after formation of a T P primer-429 DNA polymerase-template ternary complex, efficient binding and selection of the complementary nucleotide can occur. The analysis of N-terminal deletion mutants in TP (26) and the fact that single-stranded DNA can be efficiently replicated by a protein-primed initiation mechanism3 support the idea that binding of TP to the 3'-end of the template strand is needed to precisely locate the TP/DNA polymerase heterodimer at both replication origins. According to this functional model, an important question is how the structure of the 429 DNA polymerase is adapted to use both a protein (for initiation) and DNA (during processive elongation) as primers. Recently, we have found significant similarities among the two main families of DNA-dependent DNA polymerases (a-like and polymerase I-like), suggesting the existence of an evolutionary conserved Klenow fragment-like core in either protein-primed or DNA-and RNA-primed DNA polymerases (8,27). According to the three-dimensional structure of the Klenow fragment, the cleft, which is involved in binding the double-stranded region (containing the DNA or RNA primer strand) of the replicating DNA molecule, is proposed to be the TP-binding site. In agreement with a structural equivalence between TP and the double-stranded region of a template-primer DNA molecule, it has been re- The novel DNA polymerase-mediated reaction described here not only provides data on the structural and functional relevance of the TP-DNA polymerase interaction, but, as in part exemplified in this paper, also constitutes an invaluable tool to analyze and define: 1) TP critical residues involved in the priming function and TP domains involved in binding the DNA template and/or the 429 DNA polymerase and 2) 429 DNA polymerase critical residues directly affecting catalysis, being possible to define and discriminate between residues directly involved in nucleotide binding versus those contributing to (via DNA binding) or affecting nucleotide selection. Thus, as shown in this paper, the effect of single changes in the TP linking amino acid (SerZ3') was independent of the presence of the 429 DNA and parental TP. This independence together with the fact that the mutations introduced (SerZ3' + Thr, S e P -Cys) do not affect the overall interaction with $29 DNA polymerase (12, 13) suggest that they are directly affecting the active site for protein priming; in the case of the ThrZ3' mutant, the OH group is probably in an incorrect orientation with respect to the 429 DNA polymerase dNTP-binding site, whereas the low efficiency of the CysZ3' mutant could be due to differences in both orientation and reactivity of the substituting SH group.