A positively charged residue of φ29 DNA polymerase, highly conserved in DNA polymerases from families A and B, is involved in binding the incoming nucleotide

Abstract

Alignment of the protein sequence of DNA-dependent DNA polymerases has allowed the definition of a new motif, lying adjacent to motif B in the direction of the N-terminus and therefore named pre-motif B. Both motifs are located in the fingers subdomain, shown to rotate towards the active site to form a dNTP-binding pocket in several DNA polymerases in which a closed ternary complex pol:DNA:dNTP has been solved. The functional significance of pre-motif B has been studied by site-directed mutagenesis of φ29 DNA polymerase. The affinity for nucleotides of φ29 DNA polymerase mutant residues Ile364 and Lys371 was strongly affected in DNA- and terminal protein-primed reactions. Additionally, mutations in Ile364 affected the DNA-binding capacity of φ29 DNA polymerase. The results suggest that Lys371 of φ29 DNA polymerase, highly conserved among families A and B, interacts with the phosphate groups of the incoming nucleotide. On the other hand, the role of residue Ile364 seems to be structural, being important for both DNA and dNTP binding. Pre-motif B must therefore play an important role in binding the incoming nucleotide. Interestingly, the roles of Lys371 and Ile364 were also shown to be important in reactions without template, suggesting that φ29 DNA polymerase can achieve the closed conformation in the absence of a DNA template.

INTRODUCTION

φ29 DNA polymerase is a small 66 kDa monomeric enzyme belonging to the family of eukaryotic-type DNA replicases (1,2), also referred to as family B (3). In addition to 3′→5′ exonuclease and 5′→3′ polymerisation activities, common to most replicative DNA polymerases, this enzyme is able to use a protein as primer to initiate φ29 DNA replication (4). Initiation of replication (Fig. 1) occurs at both DNA ends and replication proceeds continuously without further priming events. Initiation is templated by the second nucleotide at the 3′ end of the DNA (5) and consists of the formation of a covalent phosphoester linkage between the hydroxyl group of Ser232 of the primer protein and 5′-dAMP (6) catalysed by φ29 DNA polymerase (7). The primer protein was named terminal protein (TP) since, as a consequence of this mechanism, a TP molecule remains covalently linked to each 5′ end of φ29 DNA, playing an important role as parental TP in the next replication round. ‘Sliding back’ of the initiation product TP–dAMP allows the second nucleotide of the template strand to again serve as template for the next step, TP–dAMP₂ formation (5). After some transition steps, in which the φ29 DNA polymerase remains complexed with TP and probably undergoes a structural change (8), the polymerase dissociates from TP and replicates each DNA strand while displacing the other. Due to its high processivity (>70 kb) and strand-displacement capacity, φ29 DNA polymerase does not need the help of any helicase or processivity factor to completely replicate its 19 kb linear natural template, TP–DNA (9,10). Interestingly, φ29 DNA polymerase is able to covalently link any of the four dNMP residues to TP in the absence of template DNA (11) (Fig. 1). A similar reaction during DNA-primed polymerisation is the addition of 1 nt to a blunt-ended DNA substrate (+1 addition) (12), a reaction that is also performed by φ29 DNA polymerase (J.A.Esteban, L.Blanco, A.Bernad and M.Salas, unpublished data). Extensive structure–function studies and biochemical characterisation of φ29 DNA polymerase (reviewed in 4; 10,13,14) has shown the various functions of this enzyme during protein-primed DNA replication. Site-directed mutational analysis of φ29 DNA polymerase allowed the demonstration of the existence of two independent domains (at the primary structure level), containing the synthetic and degradative activities of this enzyme (10; reviewed in 14). Thus, the N-terminal domain was shown to contain an evolutionarily conserved 3′→5′ exonuclease active site (2,15) as well as the strand-displacement activity (16). On the other hand, the C-terminal domain was shown to be involved in DNA polymerisation and protein-primed initiation. In agreement with these results, obtained by mutational analysis, the separately expressed C-terminal domain of φ29 DNA polymerase (44 kDa) had neither 3′→5′ exonuclease nor strand-displacement activity, but retained polymerisation and TP-primed initiation activities (17). A second degradative activity, pyrophosphorolysis, dependent on the presence of inorganic pyrophosphate and formally considered as polymerisation reversal, was shown to be located in the C-terminal domain (18).

Figure 1.

Enzymatic processes catalysed by φ29 DNA polymerase. (A) TP binding, as the formation of a stable complex between DNA polymerase and TP. (B) In the absence of template, TP–dNMP can be formed with any of the four dNTPs and is defined as TP deoxynucleotidylation. (C) In the presence of template, only TP–dAMP is preferentially formed, since it corresponds to the templated initiation reaction of φ29 DNA replication. (D) After initiation, elongation requires efficient transition, as well as strand displacement. The 3′→5′ exonuclease activity of φ29 DNA polymerase controls nucleotide insertion fidelity after initiation.

Taking into account the similarity of the open structure of RB69 DNA polymerase (19) with other crystal structures of family B DNA polymerases reported more recently (20–23), the RB69 enzyme is a good structural model for the family B DNA polymerases. Recently, the crystal structure of RB69 DNA polymerase as a closed complex has been published (24), being the only family B DNA polymerase crystallised as a closed ternary complex. As previously described for the DNA polymerases of family A (25,26), polymerase β (27) and reverse transcriptase families (28), the fingers subdomain of the RB69 DNA polymerase undergoes a substrate-induced conformational change upon binding of both DNA and dNTP, rotating towards the palm to form a dNTP-binding pocket. In agreement with this, the highly conserved Kx₃NSxYG motif, also named motif B (29), located in the fingers subdomain, was shown to be involved in both template/primer binding and dNTP selection. In φ29 DNA polymerase the role of the Kx₃NSxYG motif was studied by site-directed mutagenesis. Mutations in the invariant residue Lys383 led to a reduced dNTP-binding affinity (30). On the other hand, mutations in Gly391 and Phe393 drastically affected DNA binding, while mutations in Asn387 and Ser388 caused intermediate phenotypes (31). Tyr390 was proposed to be involved in checking proper base pairing of the incoming nucleotide (32,33).

In this study we define a new motif located adjacent to motif B, named pre-motif B, that is highly conserved among DNA-dependent DNA polymerases belonging to families A and B. We show that two of the residues of this motif (Ile364 and Lys371 of φ29 DNA polymerase) are very important for intrinsic binding of the incoming nucleotide during deoxynucleotidylation.

MATERIALS AND METHODS

Nucleotides and proteins

Unlabelled nucleotides were purchased from Pharmacia P-L Biochemicals. [α-³²P]dATP (3000 Ci/mmol) and [γ-³²P]ATP (5000 Ci/mmol) were obtained from Amersham International. Restriction endonucleases were from New England Biolabs. T4 polynucleotide kinase was from Boehringer Mannheim.

φ29 TP–DNA was isolated according to Peñalva and Salas (34). TP (M_r 30 918) was purified as described by Zaballos and Salas (35). The double-stranded DNA-binding protein p6 (M_r 11 873), obtained from Bacillus subtilis cells infected with phage φ29, was purified as described (36).

DNA templates and substrates

Oligonucleotides sp1 (5′-GATCACAGTGAGTAC), its complementary oligonucleotide sp1c (5′-GTACTCACTGTGATC) and sp1c+6 (5′-TCTATTGTACTCACTGTGATC), which has a 5′ extension of 6 nt in addition to the sequence complementary to sp1, were prepared with a DNA synthesiser from Applied Biosystems. The oligonucleotide sp1 was first purified by electrophoresis on 8 M urea–20% polyacrylamide gels and then 5′-labelled with [γ-³²P]ATP and T4 polynucleotide kinase. To analyse the polymerisation activity on a template/primer structure, 5′-labelled sp1 was hybridised to sp1c+6 in the presence of 0.2 M NaCl and 60 mM Tris–HCl, pH 7.5. The hybrid molecules sp1/sp1c (matched 15/15mer) and sp1/sp1c+6 (matched 15/21mer) were used as substrates for gel retardation and/or polymerisation and exonucleolysis experiments on a template/primer DNA.

Site-directed mutagenesis and expression of φ29 DNA polymerase mutants

The wild-type φ29 DNA polymerase gene, cloned into M13mp19 (M13mp19w21) (15), was used for site-directed mutagenesis, carried out essentially as described (37), using an oligonucleotide-directed in vitro mutagenesis kit from Amersham International. The fragments carrying the different mutations were subcloned in plasmid pT7-4w2 (13), which expresses φ29 DNA polymerase under the control of the T7 RNA polymerase-specific φ10 promoter (38). Presence of the desired mutation and absence of any other changes were confirmed by complete sequencing of each φ29 DNA polymerase mutant gene. Sequencing was carried out by the chain termination method, using Sequenase v.2.0 from US Biochemical Corp., and a set of synthetic oligonucleotides complementary to the φ29 DNA polymerase gene as sequencing primers. Expression of the mutant proteins was carried out in Escherichia coli strain BL21(DE3)pLysS, which contains the T7 RNA polymerase gene under control of the IPTG-inducible lacUV5 promoter, and a plasmid constitutively expressing T7 lysozyme (39,40). Purification of wild-type and mutant φ29 DNA polymerases (M_r 66 520) was done as described (13).

Buffers

The reaction buffer used in all assays contained 500 mM Tris–HCl, pH 7.5, 10 mM dithiothreitol, 40% glycerol and 1 mg/ml bovine serum albumin (BSA). The buffer used for dilution of the polymerase contained 25 mM Tris–HCl, pH 7.5, 120 mM NaCl, 1 mg/ml BSA and 50% glycerol.

DNA gel retardation assays

The interaction of either wild-type or mutant φ29 DNA polymerases with a template/primer structure was assayed using the 5′-labelled sp1/sp1c+6 (15/21mer) hybrid. The incubation mixture contained, in a final volume of 20 µl, 12 mM Tris–HCl, pH 7.5, 1 mM EDTA, 20 mM ammonium sulfate, 10 mM MgCl₂ or 1 mM MnCl₂, 0.1 mg/ml BSA, 1.2 nM 5′-labelled sp1/sp1c+6 and, as indicated, 1.2, 4.2, 8.4 or 16.8 nM either wild-type or mutant φ29 DNA polymerase (41). After incubation for 5 min at 4°C, the samples were subjected to electrophoresis in 4% (w/v) polyacrylamide gels (80:1 acrylamide:bisacrylamide) 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 (42). After autoradiography, enzyme–DNA stable interaction was detected as a shift in the position of the free labelled DNA and quantitated by densitometry of the autoradiographs corresponding to different experiments.

Polymerase/3′→5′ exonuclease coupled assay

The hybrid molecule sp1/sp1c+6 (15/21mer) contains a 6 nt long 5′-protruding end and, therefore, the primer strand can be used both as a substrate for 3′→5′ exonuclease activity and also for DNA-dependent DNA polymerisation. The reaction mixture contained, in a final volume of 12.5 µl, 1.25 µl reaction buffer, 10 mM MgCl₂ or 1 mM MnCl₂, 1.2 nM 5′-labelled hybrid molecule (sp1/sp1c+6), 16.8 nM either wild-type or mutant φ29 DNA polymerase and the indicated increasing concentrations of the four dNTPs. After incubation for 5 min at 25°C, the reaction was stopped by addition of 3 µl of sequencing loading buffer. Samples were analysed by 8 M urea–20% polyacrylamide gel electrophoresis and autoradiography. Polymerisation or 3′→5′ exonuclease were detected as an increase or decrease, respectively, in the size (15mer) of the 5′-labelled sp1 primer. Different dNTP concentrations are required for the wild-type φ29 DNA polymerase to polymerise the first nucleotide (from position 15 to 16), to effectively compete with the 3′→5′ exonuclease activity replicating until the last nucleotide (from position 16 to 20) and to replicate the last nucleotide of the template (from position 20 to 21). The exonucleolytic degradation value was obtained by calculating the number of catalytic events giving rise to each degradation product. From these data (obtained by densitometry of the autoradiographs), the catalytic efficiency (indicated in Table 1) of each mutant derivative, assayed under linear conditions of both time and enzyme amount, was calculated relative to wild-type φ29 DNA polymerase.

Table 1. Enzymatic activities of the mutant derivatives of φ29 DNA polymerase.

Activity assay		Substrate	φ29 DNA polymerase
			Wild-type	1364Q	K371T
DNA bindinga	Mg2+	sp1/sp1c+6	100	14	103
Exonucleasea	Mg2+	sp1/sp1c+6	100	18	90
Polymerase/exonuclease balanceb (µM)	Mg2+	sp1/sp1c+6	0.1	1	10
	Mn2+		0.025	0.025	0.1
KmdNTP (µM)	Mg2+	sp1/sp1c+6	0.3	12.1	59
	Mn2+		0.065	0.075	0.8
+1 additionc (µM)	Mg2+	sp1/sp1c	50	n.d.	n.d.
	Mn2+		5	100	n.d.
Pyrophosphorolysisa	Mg2+	sp1/sp1c+6	100	4	10

The different activity assays were carried out with the indicated substrates and metal activators as described in Materials and Methods. The numbers are the averages of several experiments.

^aNumbers indicate percentage activities obtained with respect to that of the wild-type DNA polymerase.

^bIn the polymerase/exonuclease coupled assay the dNTP concentration required, in µM, to effectively compete with the 3′→5′ exonuclease activity is indicated.

^cIn the +1 addition experiment the dNTP concentration required, in µM, to perform this non-templated polymerisation reaction is indicated. n.d. not detectable.

Measurement of nucleotide-binding affinity

Analysis of the Michaelis–Menten constant for nucleotide binding (K_m^dNTP) in DNA polymerisation [using the 15/21mer template/primer structure (sp1/sp1c+6) and MgCl₂ or MnCl₂ as metal activator] was carried out as described (33), with the following change: the assay was performed at 5°C, a temperature at which almost no exonuclease activity could be detected, to avoid exonucleolytic degradation of the template/primer structure. Therefore, no extra dCTP was added to the assay to prevent degradation of the DNA.

Non-templated +1 addition on a blunt-ended DNA substrate

The blunt-ended DNA substrate for this assay was the 5′-labelled hybrid molecule sp1/sp1c (15/15mer). The reactions were performed as described for the polymerase/exonuclease experiment, but using dNTP concentrations ranging between 5 and 500 µM. After incubation for 5 min at 25°C, the samples were analysed by 8 M urea–20% polyacrylamide gel electrophoresis. Non-templated +1 addition is seen as a band appearing at position 16.

Pyrophosphorolysis

Pyrophosphorolytic activity was measured on the template/primer substrate sp1/sp1c+6 (15/21mer). The reaction mixture was the same as for the polymerase/exonuclease experiment, but without dNTPs and using MgCl₂ as metal activator. Increasing concentrations of tetrasodium pyrophosphate (Sigma) were added (1–6 mM), as indicated. To reduce exonuclease activity on this substrate, the reactions were incubated for 10 min on ice. Exonuclease activity of the DNA polymerases was seen when no pyrophosphate was added, while the degradative activity seen with increasing concentrations of pyrophosphate corresponds to pyrophosphorolytic activity. Higher concentrations than 6 mM pyrophosphate inhibited the degradative activity of wild-type and mutant DNA polymerases.

Protein-primed initiation assay

The formation of TP–dAMP was performed either with or without TP–DNA as template. In both cases the reaction mixture contained, in a final volume of 25 µl, 2.5 µl reaction buffer, 20 mM ammonium sulfate, 0.1 µM [α-³²P]dATP (2 µCi), 83 nM purified TP and 16.8 nM either wild-type or mutant φ29 DNA polymerase. When indicated, 34 µM p6 was added to the reaction. Two conditions in the template-dependent assay (with 1.6 nM TP–DNA) were used: one contained 10 mM MgCl₂ as metal activator and the incubation time was 5 min at 30°C; the other contained 1 mM MnCl₂ and the incubation time was 1 min at 30°C. The assay without template was performed with 1 mM MnCl₂ as metal activator and incubation was for 4 h at 30°C. Reactions were stopped by addition of 10 mM EDTA and 0.1% SDS, filtration through Sephadex G-50 spin columns (in the presence of 0.1% SDS) and further analysis by SDS–PAGE as described (34). The TP–dAMP complex was detected by autoradiography and quantitated by densitometric analysis. To calculate the Michaelis–Menten constant for dATP binding (K_m and V_max) in the TP–DNA initiation reaction different dATP concentrations and incubation times were assayed.

TP–DNA replication

The incubation mixture was the same as described for the initiation reaction, but supplemented with 20 µM each dGTP, dCTP, dTTP and [α-³²P]dATP (2 µCi). After incubation at 30°C for times of between 5 and 60 min the reactions were stopped by addition of 10 mM EDTA and 0.1% SDS, and the samples were filtered through Sephadex G50 spin columns in the presence of 0.1% SDS. The Cerenkov radiation in the excluded volume was measured and used for quantitation. Elongation was studied by alkaline 0.7% agarose gel electrophoresis (43). The position of unit length φ29 DNA in the agarose gels was detected by ethidium bromide staining. Autoradiography of the dried gels revealed the transition step or elongation efficiency.

RESULTS

Definition of pre-motif B of DNA-dependent DNA polymerases

As shown in Figure 2, a new conserved amino acid sequence has been found in DNA-dependent DNA polymerases lying at a distance between 5 (φ29) and 70 amino acids (T4 and RB69; these two DNA polymerases here contain a divergent insertion of 55 amino acids; see ref. 19) from motif B in the direction of the N-terminus, and therefore named pre-motif B. In DNA polymerases that start replication by protein priming, this new motif is adjacent to the TPR-1 region, a conserved region only present in protein-primed DNA polymerases (44). The most conserved residue of pre-motif B is an invariant positively charged amino acid: Lys in the family B DNA polymerases (two exceptions with Arg) and His in the family A DNA polymerases (one exception). The corresponding residue of φ29 DNA polymerase, Lys371, has been changed by site-directed mutagenesis into the polar but uncharged amino acid Thr. Seven amino acids further upstream of this residue, two non-polar residues are conserved in nearly all DNA-dependent DNA polymerases. These two residues consist in most of the DNA polymerase sequences of a Val, Leu or Ile and an aromatic residue, Phe, Tyr or Trp, respectively. The order of the two kinds of residues is conserved in two groups: family A and protein-primed family B members. The cellular family B DNA polymerases do not contain an aromatic residue, but instead Leu, Ile or Val. One of these residues of φ29 DNA polymerase, Ile364, was subjected to site-directed mutagenesis introducing two non-conservative changes: one mutation into the polar uncharged amino acid Gln and another into the positively charged Arg. All these mutations in φ29 DNA polymerase were designed according to general suggestions for conservative substitutions (45) and secondary structure predictions (46,47).

Figure 2.

Alignment of pre-motif B of DNA-dependent DNA polymerases. Multiple alignment of the amino acid sequences near the Kx₃NSxYG motif (motif B) in the direction of the N-terminus of DNA-dependent DNA polymerases. DNA polymerase nomenclature and sequence references are compiled in Braithwaite and Ito (3), with the exception of pAL2 DNA polymerase codified by the linear plasmid pAL2 from Podospora anserina (54), Streptococcus pneumoniae phage Cp1 DNA polymerase (GenBank accession no. Z47794) and phage GA-1 DNA polymerase (55). Numbers between slashes indicate the amino acid position relative to the N-terminus of each DNA polymerase sequence. Invariant or highly conserved residues are indicated in white letters on a black background. Significant similarities are indicated by grey boxes. Identical residues aligned in more than half of the sequences are in bold. The consensus sequence of motif B for each subgroup of DNA polymerases is indicated, as well as its distance from pre-motif B. A multiple alignment of motif B of DNA-dependent DNA polymerases is shown in Saturno et al. (30). The residues of φ29 DNA polymerase studied here, Lys371 and Ile364 (indicated with arrows), and the previously studied residues of motif B (reviewed in 14;30) are indicated by an asterisk. The following conserved amino acids were considered: S and T; A and G; D and E; K, R and H; A, I, L, M, C and V; Y and F.

DNA-binding capacity of the φ29 DNA polymerase mutants

The interaction of the different mutant φ29 DNA polymerases with a template/primer structure was studied by gel retardation (described in Materials and Methods). In these assays the wild-type enzyme gives rise to a single retardation band (Fig. 3), which most likely corresponds to a protein–DNA interaction in which the primer terminus is stabilised at the polymerisation active site (41). A similar result was obtained with mutant DNA polymerase K371T (Fig. 3), suggesting that this mutation did not affect DNA-binding capacity. On the other hand, mutant DNA polymerase I364R was unable to stably bind the DNA substrate. Because of this drastic defect this mutant DNA polymerase had no polymerisation and initiation activities and is therefore not presented in the following experiments and in Table 1. Mutant polymerase I364Q showed a less drastic phenotype (Fig. 3): it was able to bind a template/primer structure, although less efficiently than the wild-type DNA polymerase (Table 1). The DNA-binding capacity of mutant DNA polymerases I364Q and K371T did not change in the presence of Mn²⁺ instead of Mg²⁺ as metal ion (not shown). The reduction in template/primer binding ability of mutant DNA polymerase I364Q is most probably the reason for the similar reduction in its exonuclease activity on the same template/primer substrate (Table 1). In agreement with this, the exonuclease activity of mutant DNA polymerase K371T was not affected (Table 1).

Figure 3.

DNA-binding capacity of the mutant φ29 DNA polymerases by gel retardation analysis. The 5′-labelled hybrid molecule sp1/sp1c+6 (15/21mer) was incubated with increasing amounts (nM) of the wild-type or mutant φ29 DNA polymerases, under the conditions described in Materials and Methods. After non-denaturating polyacrylamide gel electrophoresis, the mobilities of free DNA (sp1/sp1c6) and the polymerase–DNA complex were detected by autoradiography.

Polymerase/exonuclease balance on a template/primer structure

A change in the polymerase/exonuclease balance from exonucleolysis to polymerisation is obtained on a template/primer structure by increasing the dNTP concentration (see Materials and Methods). This polymerase/exonuclease balance not only depends on the presence of these two activities, but particularly on the relative affinity of the polymerisation active site for nucleotides and of both active sites for the primer strand. As shown in Figure 4, different dNTP concentrations are required for the wild-type φ29 DNA polymerase to (i) start replication and polymerise the first nucleotide (25 nM, from position 15 to 16); (ii) to effectively compete with the 3′→5′ exonuclease activity replicating until the penultimate nucleotide (100 nM, from position 16 to 20); (iii) to replicate the last nucleotide of the template/primer substrate (1 µM, from position 20 to 21). Mutant DNA polymerase I364Q needed a 10 times higher nucleotide concentration than the wild-type DNA polymerase for every step of replication of the same substrate in the presence of Mg²⁺ ions (Fig. 4 and Table 1). Mn²⁺ as metal activator is known to reduce the K_m for nucleotides during polymerisation (48). In agreement with this, no higher nucleotide concentration was needed with mutant DNA polymerase I364Q, in comparison to the wild-type DNA polymerase, to achieve effective polymerisation of the substrate in the presence of Mn²⁺ ions (Table 1). On the other hand, mutant DNA polymerase K371T required a higher nucleotide concentration than the wild-type DNA polymerase to effectively compete with its wild-type-like 3′→5′ exonuclease activity, replicating up to the fifth nucleotide (20mer) in the presence of Mg²⁺ ions at a nucleotide concentration of 10 µM (Fig. 4). Again, the use of Mn²⁺ as metal activator reduced the nucleotide requirement of this mutant DNA polymerase (Table 1). The latter polymerase was very inefficient in replicating the last nucleotide, since it required a nucleotide concentration of 200 µM even using Mn²⁺ (not shown). Mutations in both Lys371 and Ile364 did not affect insertion fidelity during polymerisation (not shown).

Figure 4.

DNA polymerase/exonuclease coupled assay. The assays were carried out as described in Materials and Methods, using the ³²P-labelled hybrid molecule sp1/sp1c+6 (15/21mer, sequence indicated) as template/primer DNA. After autoradiography of the 8 M urea–20% polyacrylamide gel, polymerisation or 3′→5′ exonuclease were detected as an increase or decrease, respectively, in the size (15mer) of the 5′-labelled sp1 primer. The positions of the non-elongated primer (15mer), elongated primer (16mer, 20mer and 21mer) and degraded primer (3mer) are shown. In lane c the DNA substrate was loaded alone as a control (15mer). The dNTP concentration required to compete with exonucleolysis, allowing efficient replication of the template, is shown in Table 1 for the wild-type and mutant φ29 DNA polymerases.

Determination of the K_m for nucleotides during polymerisation of the same template/primer substrate (see Materials and Methods) in the presence of Mg²⁺ as metal activator revealed a 40 times higher K_m for mutant DNA polymerase I364Q and a 200-fold higher K_m for mutant DNA polymerase K371T than for the wild-type enzyme (Table 1). These results are in agreement with those of the polymerase/exonuclease experiment that suggested that both mutant DNA polymerases had a strongly reduced affinity for dNTPs. Also in agreement with the polymerase/exonuclease experiments in the presence of Mn²⁺, the K_m for mutant DNA polymerase I364Q was similar to that of the wild-type DNA polymerase, while the affinity for dNTPs of mutant DNA polymerase K371T was still ∼10 times lower (Table 1).

Protein-primed initiation and replication activity

To study the effects of the mutations on protein-primed initiation activity the formation of TP–dAMP was first performed in the absence of a DNA template (11). In this assay the only interactions involved are those occuring between φ29 DNA polymerase and TP and the initiating nucleotide. As shown in Figure 5, wild-type φ29 DNA polymerase catalyses formation of TP–dAMP in the absence of TP–DNA, whereas mutant DNA polymerases I364Q and K371T were unable to perform this reaction (Fig. 5 and Table 2).

Figure 5.

TP-primed initiation activity of the φ29 DNA polymerase mutants. Template-dependent (TP–DNA) and template-independent (no template) formation of TP–dAMP complex was studied comparatively using the wild-type and mutant DNA polymerases. The reactions were carried out as indicated in Materials and Methods, using either 1 mM MnCl₂ or 10 mM MgCl₂ as metal activator and 34 µM protein p6 when indicated (+). The TP–dAMP complex is seen as a band after autoradiography of the SDS–PAGE gel. Mean activity values relative to the wild-type are given in Table 2.

Table 2. Initiation and polymerisation activity of the φ29 DNA polymerase mutants.

Activity assay		Substrate	φ29 DNA polymerase
			Wild-type	1364Q	K371T
TP–dAMP formation	Mg2+	TP–DNA, ± p6	100	6/3	<1/1
	Mn2+	TP–DNA, ± p6	100	15/33	1/1
	Mn2+	No template	100	1	<1
TP–dAMP formation, KmdATP (µM)	Mg2+	TP–DNA	5.4	13.5	n.d.
	Mn2+		1.1	1.5	10.4
TP–dAMP formation, Vmax (fmol/min)	Mg2+	TP–DNA	4.2	0.8	n.d.
	Mn2+		11.2	2.5	1.4
TP–DNA replication	Mg2+	TP–DNA, ± p6	100	1/1	<1
	Mn2+		100	22/30	1/1

The kinetic parameters of the initiation reaction were calculated as described in Materials and Methods. The values are the averages of several experiments. n.d., not detectable.

The interaction of both mutant DNA polymerases with TP was shown to be normal, since the mutant DNA polymerases (66 kDa) and TP (31 kDa) co-sedimented at a position corresponding to ∼95 kDa, indicating a TP–DNA polymerase heterodimer, after glycerol gradient centrifugation (48,49), and both mutant DNA polymerases were able to inhibit TP deoxynucleotidylation by the wild-type φ29 DNA polymerase at limiting TP concentration (44; not shown). Therefore, a defective interaction with dATP or catalysis itself must be the reason for their strongly affected non-templated TP deoxynucleotidylation activity. In comparison to template-independent TP deoxynucleotidylation, the template-directed TP–DNA initiation activity of wild-type φ29 DNA polymerase is much more effective. However, even with TP–DNA almost no initiation activity was detectable with mutant DNA polymerase K371T, even in the presence of Mn²⁺ and φ29 protein p6 (Fig. 5 and Table 2). Like Mn²⁺ (50), p6 also favours the initiation reaction due to a decrease in K_m for the initiating nucleotide, dATP (51). Determination of the K_m for the initiating nucleotide of mutant DNA polymerase K371T revealed a higher K_m and lower V_max than those of the wild-type DNA polymerase in the presence of Mn²⁺, being undetectable with Mg²⁺ (Table 2). Since almost no initiation activity was detectable with this mutant DNA polymerase, no further replication was observed (Table 2). On the other hand, TP–DNA initiation activity of mutant DNA polymerase I364Q was strongly reduced in the presence of Mg²⁺ ions, increasing in the presence of Mn²⁺ and p6 (Fig. 5 and Table 2). Similar values were obtained when replication of TP–DNA by this mutant DNA polymerase was assayed (Table 2), which was able to perform processive replication coupled to strand displacement and produce full-length φ29 DNA (not shown). Determination of the K_m for the initiating nucleotide for mutant DNA polymerase I364Q in the presence of Mg²⁺ showed a 2.5-fold higher value than that of the wild-type DNA polymerase, while the V_max was five times lower (Table 2). The K_m of the mutant DNA polymerase was reduced to a similar value to that of the wild-type DNA polymerase when Mn²⁺ was used as metal activator, while the V_max was still about five times lower (Table 2). Therefore, both mutant DNA polymerases were affected in their affinity for nucleotides during initiation and polymerisation.

Non-templated +1 addition activity on a blunt-ended DNA substrate

Deoxynucleotidylation of TP by φ29 DNA polymerase in the absence of template resembles +1 nt addition on blunt-ended double-stranded (ds)DNA, which has been described for several DNA polymerases (12; J.A.Esteban, L.Blanco, A.Bernad and M.Salas, unpublished data). Since both mutant DNA polymerases, I364Q and K371T, were affected in non-templated formation of TP–dAMP, the similar reaction under DNA-primed conditions was studied. As can be seen in Figure 6, the wild-type DNA polymerase is able to perform +1 addition at high dNTP concentrations (Table 1). On the other hand, polymerisation from position 15 to 16 could not be observed with either of the two mutant DNA polymerases. In comparison to the wild-type DNA polymerase the exonuclease activity of the mutant DNA polymerases on this substrate (Fig. 6, lane 0 µM dNTP) was similar to that on the template/primer DNA (Table 1), suggesting that their DNA-binding capacity was also similar for both substrates. When Mn²⁺ was used as metal activator the wild-type DNA polymerase required a lower dNTP concentration to perform +1 addition (Table 1). Under these conditions mutant DNA polymerase I364Q showed polymerisation to 16mer, while mutant DNA polymerase K371T was still unable to perform this reaction (Table 1).

Figure 6.

+1 addition on a blunt-ended DNA substrate. The template-independent addition of a nucleotide on the blunt-ended sp1/sp1c (15/15mer) substrate was performed as described in Materials and Methods using Mg²⁺ as metal activator. The positions of the non-elongated primer (15mer), elongated primer (16mer) and degraded primer (3mer) can be observed on this autoradiograph of an 8 M urea–20% polyacrylamide gel.

Pyrophosphorolytic activity of the φ29 DNA polymerase mutants

φ29 DNA polymerase has two 3′→5′ degradative activities: an exonuclease and an inorganic pyrophosphate-dependent degradative activity (18), pyrophosphorolysis, regarded as the reverse of the polymerisation reaction. As can be seen in Figure 7, addition of increasing pyrophosphate concentrations results in a degradative activity distinguishable from the exonuclease activity (the latter being detectable in the absence of pyrophosphate). Using conditions under which exonucleolysis was reduced by lowering the incubation temperature, both mutant DNA polymerases showed reduced pyrophosphorolytic activity, requiring high pyrophosphate concentrations to be detectable.

Figure 7.

Pyrophosphorolytic activity of the φ29 DNA polymerase mutants. Pyrophosphorolytic activity of the wild-type and mutant φ29 DNA polymerases on the template/primer substrate sp1/sp1c+6 (15/21mer) observed by addition of increasing concentrations of inorganic pyrophosphate (performed as described in Materials and Methods). To avoid exonuclease activity on the substrate the reaction was incubated on ice. In the autoradiograph of an 8 M urea–20% polyacrylamide gel, the positions of the non-elongated primer (15mer) and the degraded primers are indicated.

DISCUSSION

Motif B is highly conserved in all DNA-dependent DNA polymerases and, in agreement with this, the important functions of the amino acid residues of this motif are also highly conserved: dNTP selection and template/primer binding. Here we describe a new conserved region adjacent to motif B in the direction of the N-terminus, and therefore named pre-motif B. Our results show that the highly conserved Lys residue at position 371 in φ29 DNA polymerase is important for binding the incoming nucleotide. Mutation of this residue to Thr did not affect either the DNA- or the TP-binding capacity of the mutant DNA polymerase, indicating that this residue is not involved in binding the template/primer substrate. Conversely, this mutant DNA polymerase had a strongly decreased affinity for nucleotides in both the DNA-primed polymerisation and TP-primed initiation reactions. While the exonuclease activity of mutant DNA polymerase K371T was normal, its pyrophosphorolytic activity was strongly reduced. This indicates that residue Lys371 is probably involved in binding the β or γ phosphates of the incoming nucleotide. Interestingly, the role of this residue is also important in the non-templated TP deoxynucleotidylation and polymerisation reactions. When Lys371 is mutated to Thr no TP–dAMP formation was observed in the absence of template. Also, no +1 addition on a blunt-ended DNA substrate could be observed, using either Mg²⁺ or Mn²⁺ as metal activator. This supports a role of Lys371 as a dNTP ligand even when no templating nucleotide is positioned at the polymerisation active site. A similar role can be proposed for residue Ile364, since the phenotype of mutant DNA polymerase I364Q was similar to that of mutant DNA polymerase K371T: a decreased affinity for the incoming nucleotide in the polymerisation and initiation reactions (even in the absence of template) and a strongly affected pyrophosphorolytic activity. In contrast to mutant DNA polymerase K371T, mutant DNA polymerase I364Q was affected in binding the DNA substrate but not TP.

In the crystal structure of several DNA polymerases with dsDNA and dNTP it has been observed that in the closed conformation the fingers subdomain is rotated towards the palm to form a nucleotide-binding pocket [family A: Klentaq (26), T7 DNA polymerase (25), rat polymerase β (27) and HIV-1 reverse transcriptase (28)]. A similar movement of the fingers has been described recently for a family B DNA polymerase, that of phage RB69 (24). Our alignment of pre-motif B, located in the fingers subdomain, shows the correspondence of Lys371 of φ29 DNA polymerase to Lys486 of RB69, His639 of Klentaq DNA polymerase and His506 of T7 DNA polymerase (see Fig. 2). In the closed complex of Klentaq and T7 DNA polymerase this His residue of pre-motif B, located in helix N, has been shown to interact directly with the β-phosphate of the incoming nucleotide (25,26). Additionally, for these two family A DNA polymerases two more positively charged residues were shown to be involved in dNTP binding: the Arg and the Lys residues of motif B, located in helix O (Arg659 and Lys663 of Taq and Arg518 and Lys522 of T7 DNA polymerase). In the closed complexes of both DNA polymerases the Arg residues contact two oxygen atoms of the γ-phosphate and the Lys residues the α-phosphate of the incoming nucleotide. In the only family B DNA polymerase crystallised as a closed complex, RB69 DNA polymerase, residue Lys486, located in helix N, has been shown to interact with the γ-phosphate of the incoming nucleotide (24). The results obtained here for the corresponding residue of φ29 DNA polymerase, Lys371, are in agreement with the RB69 polymerase data. Additionally, in RB69 DNA polymerase two more positively charged residues appear to be involved in nucleotide binding (Fig. 8): Arg482 (pre-motif B) and Lys560 (motif B). Lys560, located in helix P, forms a hydrogen bond to the bridging oxygen between the α and β phosphates of the dNTP, while Arg482, located in helix N, interacts with the γ-phosphate of the incoming nucleotide.

Figure 8.

Residues of pre-motif B and motif B involved in dNTP binding. Crystal structure of the RB69 DNA polymerase active site in the open (light colours) and closed (dark colours) conformations (24) [crystallographic data from PDB1 IG9 (closed) and PDB1 H7 (open)]. The residues belonging to pre-motif B (Arg482, Lys486 and Val478) are coloured orange/red and that belonging to motif B (Lys560) is coloured blue. The incoming nucleotide is coloured green. The positions of the three catalytic aspartates (yellow) are indicated in the open and closed conformations as references.

The alignment of pre-motif B shows that the pair Phe363/Ile364 in φ29 DNA polymerase corresponds to Val478/Phe479 in RB69 DNA polymerase. In the crystal structure of RB69 DNA polymerase (open and closed) Val478 interacts with Arg482, the latter of which has been shown to interact with the γ-phosphate of the incoming nucleotide (Fig. 8). Therefore, Val478 seems to play a structural role, stabilising the position of one or more residues involved in dNTP binding. Taking into account the reduced affinity for nucleotides of φ29 DNA polymerase mutant I364Q, residue Ile364 may be the functional analogue of RB69 DNA polymerase residue Val478. We can therefore conclude that pre-motif B must play an important role in dNTP binding.

Patel et al. (52) have proposed that for Taq DNA polymerase I in the open pol:DNA:dNTP complex (low concentration of dNTP) the triphosphate moiety is coordinated by two catalytic aspartates (Asp610 and Asp785 in Taq) and the base portion of the dNTP stacks on top of the base pair forming the primer terminus, interacting with the conserved Tyr of motif B (Tyr671 in Taq). This Tyr residue (Tyr390 in φ29 DNA polymerase) occupies the same position as the first nucleotide of the 5′-template strand overhang (26,53), in the open conformation. During the transition between the open and closed complex the Tyr of motif B will move aside to allow the templating base to interact with the incoming nucleotide (25–27) and the Phe of motif B [nearly invariant in family A DNA polymerases and corresponding to an invariant Asn residue in family B DNA polymerases (Phe667 of Taq polymerase and Asn387 in φ29 DNA polymerase)] will form stacking interactions with the dNTP, probably facilitating the movement of motif B towards motif A (52). During non-templated reactions, such as those described in this paper using either DNA or protein as primer, the Tyr of motif B could be maintaining stacking interactions with the incoming nucleotide stable enough for the reaction to occur. Other dNTP ligands, such as Asn387 of φ29 DNA polymerase, could also contribute to this reaction (31). It can be predicted that the transition from an open to closed complex is also required for catalysis in the absence of an available template. The movement of the fingers subdomain in the transition between the open and closed complex during templated reactions brings the positively charged residues of pre-motif B and motif B closer to the active site (Fig. 8) and allows them to interact with the triphosphate moiety of the incoming nucleotide. Our results indicate that during non-templated reactions the positively charged residues of motif B and pre-motif B still interact with the triphosphate groups of the incoming nucleotide, as occurs in the closed conformation: the non-templated TP-primed initiation and DNA-primed +1 addition reactions were shown to be affected in φ29 DNA polymerases with mutations of two residues of pre-motif B involved in binding the incoming nucleotide, Lys371 and Ile364. Additionally, mutation of the invariant Lys residue of motif B (Lys383), shown to be directly involved in dNTP binding, resulted in mutant DNA polymerases that were unable to initiate in the absence of TP–DNA as template (30). This agrees with our proposal that non-templated reactions occur after the formation of a closed ‘ternary’ complex (pol:primer:dNTP). The fact that mutations of residues Lys371 and Ile364 of pre-motif B affect the pyrophosphorolytic activity of φ29 DNA polymerase suggests that during this reversal of polymerisation activity, DNA polymerase, pyrophosphate and the DNA substrate form a closed ‘ternary’ complex.

We can conclude that Lys371 of φ29 DNA polymerase, highly conserved among DNA polymerases of families A and B, is a direct dNTP ligand. On the other hand, residue Ile364, also localised in pre-motif B, seems to play a structural role, being important for dNTP and DNA binding. Therefore, pre-motif B must play an important role in binding the incoming nucleotide.