A molecular handoff between bacteriophage T7 DNA primase and T7 DNA polymerase initiates DNA synthesis.

The T7 DNA primase synthesizes tetraribonucleotides that prime DNA synthesis by T7 DNA polymerase but only on the condition that the primase stabilizes the primed DNA template in the polymerase active site. We used NMR experiments and alanine scanning mutagenesis to identify residues in the zinc binding domain of T7 primase that engage the primed DNA template to initiate DNA synthesis by T7 DNA polymerase. These residues cover one face of the zinc binding domain and include a number of aromatic amino acids that are conserved in bacteriophage primases. The phage T7 single-stranded DNA-binding protein gp2.5 specifically interfered with the utilization of tetraribonucleotide primers by interacting with T7 DNA polymerase and preventing a productive interaction with the primed template. We propose that the opposing effects of gp2.5 and T7 primase on the initiation of DNA synthesis reflect a sequence of mutually exclusive interactions that occur during the recycling of the polymerase on the lagging strand of the replication fork.

The T7 DNA primase synthesizes tetraribonucleotides that prime DNA synthesis by T7 DNA polymerase but only on the condition that the primase stabilizes the primed DNA template in the polymerase active site. We used NMR experiments and alanine scanning mutagenesis to identify residues in the zinc binding domain of T7 primase that engage the primed DNA template to initiate DNA synthesis by T7 DNA polymerase. These residues cover one face of the zinc binding domain and include a number of aromatic amino acids that are conserved in bacteriophage primases. The phage T7 singlestranded DNA-binding protein gp2.5 specifically interfered with the utilization of tetraribonucleotide primers by interacting with T7 DNA polymerase and preventing a productive interaction with the primed template. We propose that the opposing effects of gp2.5 and T7 primase on the initiation of DNA synthesis reflect a sequence of mutually exclusive interactions that occur during the recycling of the polymerase on the lagging strand of the replication fork.
DNA synthesis is initiated on the lagging strand of the replication fork by a site-specific RNA polymerase known as a DNA primase (1). The primase synthesizes a short RNA primer that is extended by DNA polymerase to create an Okazaki fragment (2,3). During replication, interactions between the primase and other replicative proteins such as DNA polymerase, DNA helicase, and single-stranded DNA (ssDNA) 1 -binding protein are required to initiate DNA synthesis. It has been difficult to identify these essential interactions because they occur transiently and involve weak interactions between multiple partners.
We have been studying the physical and functional interactions between proteins of the comparatively simple bacteriophage T7 replication system. Phage T7 replicates DNA using four proteins, T7 gene 4 primase-helicase protein (gp4), T7 gene 2.5 ssDNA-binding protein (gp2.5), and T7 DNA polymerase (gp5) complexed with its processivity factor, Escherichia coli thioredoxin (4). The reactions catalyzed by the individual proteins are coordinated during replication by the assembly of a multiprotein complex that moves with the replication fork (5,6). During replication, the primase-helicase directly contacts both the DNA polymerase and gp2.5 (7)(8)(9)(10). A C-terminal acidic segment of the primase-helicase is required for its interaction with the DNA polymerase (10). An interaction between gp2.5 and T7 DNA polymerase is required for the coordinated synthesis of both leading and lagging strands of the replication fork (5,6). A C-terminal 21-residue region of gp2.5 is essential not only for the interactions with the DNA polymerase and the primase-helicase but also for dimerization of gp2.5 (9). Although many pairwise interactions between T7 replication proteins have been identified, the overall physical arrangement of subunits and their stoichiometries in the replication complex are unknown.
The T7 primase (residues 1-255) consists of two domains located in the N-terminal half of the primase-helicase protein.
The zinc binding domain (ZBD; residues 1-59) and the RNA polymerase domain (RPD; residues 60 -255) comprise the active primase (15,23), and they are tethered to the helicase by a flexible linker (16). T7 primase synthesizes tetraribonucleotide primers at specific priming sites with a consensus sequence 5Ј-(G/T)(G/T)GTC-3Ј that templates the synthesis of pppACC(C/A) and pppACAC (20,28). The conserved 3Ј-C of the priming site is required for primer synthesis, but it is not copied by the primase (29). Previous biochemical and mutational experiments have shown that the ZBD recognizes the minimal priming site (5Ј-GTC-3Ј) and that the RPD contains the active site for polymerization of ribonucleotides (30 -34). In addition to the full-length primase-helicase, a 56-kDa isoform lacking the ZBD is produced from an in-frame translation initiation site within T7 gene 4 during phage growth (35). The 56-kDa isoform of the primase-helicase has helicase activity but no primer synthesis activity (30,36).
A second important activity of T7 primase is to promote the utilization of oligoribonucleotide primers by T7 DNA polymerase (37). The primase remains associated with its tetraribo-nucleotide product on the DNA template until DNA polymerase initiates synthesis (38). In this role the primase can prevent dissociation of these short primers from the template and stabilize the primed template in the active site of DNA polymerase (3). The zinc binding domain alone of T7 primase stimulates primer utilization by T7 DNA polymerase (15). The proteinprotein interactions mediating the transfer of a primed template to DNA polymerase are starting to be elucidated in several different replication systems. During replication in E. coli, the primed DNA template is transferred from the DnaG primase to DNA polymerase III by the combined action of the ␤ clamp loader and the ssDNA-binding protein (SSB) (39). In contrast, T7 primase-helicase directly transfers a newly synthesized primer to T7 DNA polymerase and stimulates primer extension by the polymerase through an intimate physical interaction of the primase and polymerase (29,30,38).
We proposed a mechanistic model for primer synthesis and transfer to DNA polymerase that is based on a crystal structure of T7 primase along with NMR and biochemical data (15) (Fig.  1). T7 primase has a bipartite structure consisting of the ZBD and RPD separated by an extended linker. These domains do not interact with each other in the absence of substrates (open conformation of the primase) (15). During primer synthesis, the primase undergoes a conformational change in the linker region that brings the ZBD and RPD in contact with the template and nucleotide substrates (closed conformation). After a primer is synthesized, the RPD releases the primed DNA template, and the ZBD alone delivers the primed template to DNA polymerase. The recent crystal structure of the ring-shaped T7 primase-helicase is consistent with this mechanism of primer synthesis (16). The structure reveals the outward flexure of the primase domain that makes room between two neighboring primase subunits where the RNA polymerase active sites are located. This flexible architecture would allow the ZBD to engage the template along with the RNA polymerase active site.
The small size of the ZBD (59 residues) raises questions about how it binds to the primed template and engages DNA polymerase to stimulate primer extension activity. In this study, we focused on the interactions between the ZBD and primed template that support the synthesis of oligoribonucleotide primers and their utilization by T7 DNA polymerase. NMR binding experiments and mutagenic studies of the ZBD revealed that the same surface of the ZBD remains in contact with the primed template during primer synthesis and extension by the polymerase, consistent with the important role of the ZBD for both steps of the initiation of DNA synthesis. The utilization of primers was inhibited by the T7 ssDNA-binding protein (gp2.5), which interacts strongly with the polymerase and may chaperone the polymerase to the site of primer synthesis during each round of Okazaki fragment synthesis.
Protein Expression and Purification-The expression and purification of the ZBD and RPD of the T7 primase have been described previously (15). An expression plasmid for a C-terminally His 6 affinitytagged ZBD was constructed by cloning the NdeI-HindIII fragment of the wild-type ZBD plasmid between the NcoI and HindIII sites of pET28a (Novagen). Site-directed mutations of the C-terminally His 6tagged ZBD were constructed by inserting PCR fragments amplified from plasmids of mutant gene 4 proteins (32) into the NcoI-HindIII site of pET28a or by using the QuikChange site-directed mutagenesis kit (Stratagene) with the His 6 -tagged ZBD plasmid. The primers for sitedirected mutagenesis were 27 nucleotides in length and contained the sequence 5Ј-GCG-3Ј at the center encoding the alanine mutation. The full list of primers is found in Supplemental Table 1. The His 6 -tagged ZBD proteins were expressed in the E. coli strain BL21(DE3) by the addition of 1 mM isopropyl-1-thio-␤-D-galactopyranoside and incubation for 3 h at 37°C. The cells were pelleted; resuspended in lysis buffer containing 25 mM potassium phosphate buffer (pH 7.4), 500 mM KCl, and 0.1 mM PMSF; and then homogenized by the microfluidizer Emul-siFlex-C5 (Avestin). The lysate was clarified by centrifugation at 40,000 ϫ g followed by filtration of the supernatant (0.45-m pore). The protein solution was then loaded on a HiTrap nickel-chelating column (5 ml, Amersham Biosciences), and the column was washed with 60 ml of buffer A containing 25 mM phosphate buffer (pH 7.4), 500 mM KCl, 0.1 mM PMSF, and 60 mM imidazole. The bound proteins were eluted by stepwise increases in imidazole from 80 to 160 mM in 20 mM increments. The His 6 -tagged ZBDs eluted at about 120 mM imidazole, and the eluted fractions were combined and concentrated using a Centricon-3 ultrafiltration device (Amicon). The concentrated ZBD proteins were dialyzed against 20 mM Hepes (pH 7.5), 5 mM DTT, 0.1 mM PMSF, 200 mM NaCl, and 30% glycerol and then stored at Ϫ20°C. Some of the mutant ZBDs were less soluble in this buffer, and the sodium chloride concentration was increased up to 400 mM to maintain solubility.
Circular Dichroism Spectroscopy-CD measurements were performed using an AVIV model 62DS spectropolarimeter with a thermostated cuvette of 1-mm path length. All proteins were prepared at the concentration of 10 M in 300 l of sample buffer containing 20 mM potassium phosphate (pH 7.5) and 100 mM NaCl. Five recorded wavelength scans (250 -195 nm) were collected at 10°C with a signal-averaging time of 3 s and averaged together before subtracting the background spectrum of a buffer blank. The CD spectra were analyzed with the K2d program 2 to estimate the fraction of ␣-helical and ␤-strand secondary structures (43).
NMR Spectroscopy and Resonance Assignments-Two-dimensional 1 H-15 N HSQC spectra were collected from 500-l samples containing 90 M 15 N-labeled ZBD in 200 mM Hepes (pH 7.5), 100 mM NaCl, 20 mM DTT, 1 mM EDTA, 0.1 mM PMSF, and 20 mM MgCl 2 in 90% H 2 O, 10% D 2 O at 25°C using a 400-MHz Unityplus, a 500-MHz UnityInova, or a 750-MHz UnityInova Varian spectrometer. Some spectra were measured with a Cold Probe. The DNA template (5Ј-GGGTCAA-3Ј) and synthetic RNA primer (5Ј-ACCC-3Ј) were lyophilized in separate tubes. The spectrum of the ZBD alone was measured, and then the lyophilized DNA template was resuspended with the buffered protein solution to yield a final concentration of 0.4 mM ssDNA. The spectrum of ZBD-ssDNA mixture was recorded, and then the lyophilized RNA primer was dissolved in the protein-DNA solution (0.4 mM primer final concentration), and the NMR spectrum of the ZBD-ssDNA-RNA was recorded.
Primer Synthesis and Extension Assays-Primer synthesis assays were carried out as described before (38) with the following modifications. The reaction mixture (10 l) contained 1 mM each of ATP and CTP, 0.2 Ci/l [␥-32 P]ATP, 50 M 26-mer DNA template (5Ј-CAGT-GACGGGTCGTTTATCGTCGGCA-3Ј), 50 M RPD, and 10 M ZBD or ZBD mutants in reaction buffer A consisting of 40 mM Tris-HCl (pH 7.5), 10 mM MgCl 2 , 10 mM DTT, and 50 mM potassium glutamate. The reactions were incubated at 37°C for 1 h and then quenched with an equal volume of 98% formamide, 0.1% bromphenol blue, and 20 mM EDTA. The products were separated by electrophoresis in a 25% acrylamide gel containing 3 M urea and visualized by autoradiography. The relative activities of the mutant ZBDs were compared after dividing the product intensities of primer extension reaction by that of a control reaction using the wild-type ZBD.
Primer extension assays were carried out in 10-l reactions containing 50 M 26-mer DNA template (5Ј-CAGTGACGGGTCGTTTATC-GTCGGCA-3Ј), 50 M 5Ј-[ 32 P]ACCC, and 0.3 mM each of dATP, dTTP, dCTP, and dGTP along with 1 M T7 DNA polymerase in reaction buffer A containing the indicated concentrations of the ZBD, RPD, or T7 primase. The reactions were incubated for 1 h at room temperature, and the products were separated in 25% acrylamide gels containing 3 M urea and then visualized by autoradiography. The intensities of the fully extended primer strands were subtracted from the background extension activity of the DNA polymerase without primase. The relative enhancements of primer extension of each primase were calculated by dividing the intensities with that of the background extension. For the primer extension assays of the ZBD mutants, a fixed concentration (10 M) of the ZBD or ZBD mutants was used in the same reaction as above. The relative activities of the mutant ZBDs for stimulating primer extension by T7 DNA polymerase were normalized to the activity of the wild-type ZBD.
To measure inhibition by gp2.5, the primer extension reactions (10 l) contained 50 M 26-mer DNA template (5Ј-CAGTGACGGGTCGT-TTATCGTCGGCA-3Ј), 50 M 5Ј-[ 32 P]ACCC, and 0.3 mM each of dATP, dTTP, dCTP, and dGTP along with 1 M T7 DNA polymerase; 50 M ZBD, 10 M T7 primase, or 10 M primase-helicase protein; and the specified concentrations of gp2.5 (either full-length gp2.5 or the Cterminally truncated mutant gp2.5⌬26) (9) in reaction buffer A. The mixtures were incubated for 1 h at room temperature, and then the products were separated by electrophoresis in a 25% acrylamide gel containing 3 M urea and visualized by autoradiography. The primer extension activities in the presence of different concentrations of gp2.5 were normalized to the activity of a reaction without gp2.5 by taking the ratio of the product intensities.
Isolation of the Priming Complex-The isolation of the protein-DNA complexes containing T7 DNA polymerase and T7 primase were carried out in reactions (10 l) containing 20 M 26-mer biotinylated DNA template (5Ј-CAGTGACGGGTCGTTTATCGTCGGCA-biotin-3Ј), 3 mM each of ATP and CTP, 0.9 mM each of dGTP and ddTTP, 20 M T7 DNA polymerase, and 20 M T7 primase in the reaction buffer. After incubation for 1 h at room temperature, dCTP (10 mM) and Tween 20 (0.2%) were added to the reactions followed by another incubation on ice for 10 min. Streptavidin-agarose beads (Sigma) were washed in a spin filter (Nanosep MF, Pall Filtron) with an equilibrium buffer containing 20 mM Tris-HCl (pH 7.5) and 0.2% Tween 20, and the reaction mixtures were mixed with the washed beads followed by incubation for 10 min on ice. For reactions inhibited by gp2.5, 20 M gp2.5 was added at this time. The beads were washed three times with 100 l of wash buffer at 4°C in the spin filter. The wash buffer consisted of the reaction buffer plus 0.2% Tween 20, 200 mM NaCl, and 5% glycerol. The washed avidin beads were resuspended in 20 l of SDS-PAGE sample buffer and centrifuged to elute the bound proteins. The eluted proteins were separated by SDS-PAGE and visualized by staining with Coomassie Blue R-250.
For the complex containing the ZBD and the DNA polymerase, a primer strand was synthesized separately in synthesis buffer conta-ining 100 M 26-mer biotinylated DNA template (5Ј-CAGTGACGGGT-CGTTTATCGTCGGCA-biotin-3), 1 mM each of ATP and CTP, 0.3 mM each of dGTP and ddTTP, 100 nM T7 DNA polymerase, and 100 nM T7 primase in the reaction buffer. The reaction was incubated for 2 h at room temperature. The binding mixture (10 l) contained 2 l of the primer synthesis reaction (the final concentration of template was 20 M), 20 M T7 DNA polymerase, 100 M ZBD, 10 mM dCTP, and 0.2% Tween 20. The mixture was incubated on ice for 10 min. The binding and elution processes were the same as described above.
DNA Binding Assay-The binding of an ssDNA template to the primase was monitored by equilibrium fluorescence anisotropy of a 5Ј-labeled DNA (the 6-FAM label, see description under "Materials") recorded with excitation and emission wavelengths of 495 and 525 nm, respectively, in a Photon Technology International spectrofluorometer (Lawrenceville, NJ). Reactions contained 20 mM Tris-HCl (pH 7.5), 10 mM MgCl 2 , 1 mM DTT, 50 mM NaCl, 50 nM 6-FAM-labeled DNA template (17-mer), and the indicated concentrations of either the ZBD, RPD, or T7 primase. Anisotropy measurements were taken in the L-format using binding buffer as blank and correcting for the instrument G-factor as described previously (47).

RESULTS
The Conformation of the T7 Primase Changes during Primer Synthesis-We previously used x-ray crystallographic and NMR methods to show that the T7 primase undergoes a large conformational change when it synthesizes a primer (15). The ZBD and the RPD are separated by a flexible linker, and both domains move independently in solution. The two-dimensional 1 H-15 N HSQC spectrum of T7 primase and that of the isolated ZBD show identical resonances for residues within the ZBD, indicating that the ZBD tumbles independently of the RPD in the absence of DNA and nucleotide substrates (15). We refer to this structure as an "open conformation" of the primase (Fig. 1).
In the presence of an oligoribonucleotide-primed DNA template, the resonances of residues within the ZBD and RPD change, indicating that both domains contact the ssDNA template and/or the synthesized primer. We conclude from these experiments that the conformation of the primase changes from an open inactive conformation to a catalytically active, closed conformation with both ZBD and RPD engaging the primed DNA template. Here we identified residues within the ZBD that contact the primer-template heteroduplex and are required for primer synthesis and the efficient utilization of primers by T7 DNA polymerase Resonance Shifts within the ZBD upon Binding to a Primed DNA Template-After a primer is synthesized by T7 primase, the ZBD alone can stimulate primer extension by T7 DNA polymerase (15) (Figs. 1 and 2). This result suggested that following primer synthesis the RPD is released from the primed DNA template and the ZBD alone participates in the primer extension reaction with T7 DNA polymerase. The interaction of the ZBD with the primed template was confirmed by monitoring changes in the HSQC spectra of the ZBD after FIG. 1. A mechanistic model for primer synthesis and the initiation of DNA synthesis by bacteriophage T7. i, T7 primase consists of a ZBD and RPD that are connected together by a flexible linker. The ZBD is essential for the recognition of priming sites on a DNA template, and the RPD contains the active site for primer synthesis. However, these domains do not interact in the absence of substrates. ii, during primer synthesis, both the ZBD and RPD bind to the template, inducing a closed conformation of the primase. iii, after primer synthesis, the primase opens up again, and only the ZBD remains bound to the primed template. iv, the ZBD delivers the primed DNA template to T7 DNA polymerase to stimulate primer extension.

T7 Primase-Polymerase Complex
addition of a template annealed to a tetraribonucleotide primer or addition of the ssDNA template alone. Many of the resonances within the ZBD shifted position or exhibited diminished intensity after the addition of a primed DNA template (Fig. 3,  compare a and b). However, the addition of ssDNA alone had no effect (Fig. 3c) indicating that the isolated ZBD fails to stably interact with the template alone (Fig. 4). This result was unexpected since residues within the ZBD influence the DNA binding preference of the primase during primer synthesis (30 -32). We therefore further examined the DNA binding activities of the intact primase and its constituent domains using the fluorescence anisotropy assay described below. The NMR resonances of the ZBD that respond to complexation with the primed template are in good agreement with a set of resonances in the spectrum of the intact T7 primase (15) (Fig. 6).
The NMR experiments indicate that the ZBD binds to the primed template in a similar manner alone or when tethered to the RPD of T7 primase. The flexible linker that connects the ZBD and RPD of T7 primase apparently places few constraints on the binding of the ZBD.
DNA Binding Activities of Primase Domains-The T7 primase binds to a ssDNA template during primer synthesis with an apparent dissociation constant of 10 M (48). We examined the DNA binding activities of the ZBD and RPD of the primase to determine which domain contributes most strongly to interactions with the DNA template. In agreement with the NMR experiments described above, the isolated ZBD of T7 primase did not stably bind to the priming sequence of 5Ј-GTC-3Ј of the DNA template (Fig. 4). However, the RPD bound to the ssDNA template nearly as efficiently as the intact primase. The ssDNA binding activity of the RPD is nonspecific: the RPD binds efficiently to an oligo lacking the priming site (data not shown). We conclude that the RPD is likely to initiate the interaction of the primase with DNA before the ZBD engages the substrates on the priming site.
Mapping the DNA Binding Surface of the ZBD-To understand how the ZBD contacts the primed template, the NMR resonances were assigned to residues within the ZBD using a three-dimensional 15 N-edited TOCSY-HSQC spectrum, a three-dimensional 15 N-edited NOESY-HSQC spectrum, and six three-dimensional triple resonance spectra, HNCA, HN-(CO)CA, HNCO, HN(CA)CO, CBCA(CO)NH, and CBCANH (Fig. 5). Almost all of the residues within the ZBD were assigned except for five N-terminal residues that are disordered in the crystal structure of T7 primase (15). These N-terminal residues are apparently unstructured even in the complex with the primed DNA template.
We previously identified 31 resonances in the 1 H-15 N HSQC spectrum of T7 primase that were perturbed upon binding to a primed DNA template (15) (Fig. 6). The central region of the primase HSQC spectrum is crowded with many unresolved peaks that are probably derived from the RPD. We therefore focused on identifying the residues within the isolated ZBD that are perturbed in the complex with a primed DNA template. Nineteen residues were affected by complexation with the primed template (Figs. 3 and 5). Two groups of residues were identified: those with resonances that changed position in the two-dimensional spectrum and those with diminished intensity (Figs. 5 and 7). The affected residues cluster in three regions of the amino acid sequence (Fig. 7), and they are distributed across one entire surface of the ZBD (Fig. 6). The binding surface of the ZBD faces the active site of the primase in the open conformation of T7 primase that was crystallized (15).
Two charged residues within the ZBD, Asp-31 and His-33, are important determinants of specific binding to the priming site 5Ј-GTC-3Ј (32). However, the resonances for these residues did not change when the ZBD was mixed with the ssDNA template or with a primer-template heteroduplex (Figs. 3 and 6). The resonances of Asp-31 and His-33 were not resolved in the spectrum of T7 primase and are presumably hidden in the crowded central region of the spectrum (Fig. 5) (15). Although these residues contribute to the catalytic specificity of primer synthesis at 5Ј-GTC-3Ј priming sites (Fig. 8), they apparently do not directly contact the template or the primer in our experiments (Fig. 9).
Mutational Studies of the ZBD-Previous mutational studies of T7 primase have identified residues within the ZBD that contribute to the catalytic activity of primer synthesis (32). The ZBD of T7 primase also stimulates the utilization of a primed DNA template by T7 DNA polymerase (15). NMR experiments have not succeeded in the identification of residues within the ZBD that contact T7 DNA polymerase. We therefore used alanine scanning mutagenesis to identify residues within the ZBD of T7 primase that are required for the stimulation of DNA synthesis by T7 DNA polymerase in a reaction primed by the tetraribonucleotide 5Ј-pACCC-3Ј that was annealed to an oligonucleotide template. Twenty-two residues were chosen for substitution with an alanine. The mutated residues included highly conserved residues in the bacteriophage primases and those of T7 primase that interact with the primer-template heteroduplex (Fig. 7). The mutant ZBD proteins were overexpressed in E. coli and purified before measuring their activities in primer synthesis and primer extension assays (Fig. 8). Four highly conserved cysteine residues within the ZBD coordinate a zinc atom that is undoubtedly important for protein folding (15). These cysteines were not mutated because the bound metal is required for the interaction of the primase with the DNA template (31). We confirmed by CD spectroscopy that the alanine scanning mutants of the ZBD are folded and yield a CD spectrum similar to that of the wild-type ZBD. One exception was the F29A ZBD, which showed a slight deviation that might be indicative of a conformational change (Supplemental Fig. 1). Because the mutations were introduced into the ZBD alone, primer synthesis activity was assayed using a mixture of the mutant ZBD and the wild-type RPD that had been overexpressed and purified separately. The two separate domains of T7 primase complement one another for primer synthesis activity (15).
The mutant ZBDs were classified into three groups based on their primer synthesis activities: wild-type activity, signifi-FIG. 2. The ZBD, but not the RPD, of T7 primase stimulates primer utilization by T7 DNA polymerase. The DNA synthesis activity of T7 DNA polymerase was measured by the extension of a synthetic 5Ј-[ 32 P]ACCC primer annealed to a 26-mer DNA template in reactions containing T7 primase (residues 1-255), the ZBD, or the RPD. The relative amounts of DNA synthesis are expressed as the -fold enhancement over the background rate measured in the absence of the primase. The ZBD of T7 primase stimulates primer utilization nearly as well as the intact primase, whereas the RPD lacks this activity.

T7 Primase-Polymerase Complex
cantly decreased activity, and no activity (Figs. 7 and 8). The S27A, E47A, and D48A mutants retained wild-type levels of primer synthesis activity. The residues Glu-47 and Asp-48 are disordered in the crystal structure of T7 primase (15), and it is likely that they are located away from the DNA binding surface (Fig. 9). In contrast, Ser-27 is located at the center of the surface that interacts with the DNA. The HSQC resonance peak for Ser-27 was only slightly diminished when the ZBD was bound to DNA, and the analogous peak in the HSQC spectrum of the intact T7 primase is unaffected by DNA binding (15). The mutant proteins N19A and K41A showed intermediate levels of primer synthesis activity (Fig. 8). Asn-19 and Lys-41 are located on the back side of the ZBD near the bound zinc atom and away from the proposed DNA binding surface, but they could indirectly influence DNA binding activity by affecting the conformation of the ZBD.
Seventeen of the 22 alanine mutants that were constructed had no detectable primer synthesis activity (Fig. 8). Most of these essential residues populate the surface of the ZBD that faces the primase active site in the crystal structure (15) (Fig.  9). NMR experiments indicated these residues were directly affected by binding to the primer-template duplex (Fig. 7). Although they lacked detectable primer synthesis activity, the CD spectra of these alanine mutants closely resembled the wild-type spectrum, indicating that the mutant proteins are not grossly misfolded. It is remarkable that nearly all of the residues on this surface of the ZBD are required for primer synthesis activity (Fig. 9).
The ZBD of T7 primase also promotes the efficient utilization of tetraribonucleotide primers by T7 DNA polymerase (17,37).
The mutant ZBDs were therefore tested for their ability to stimulate the primer extension activity of T7 DNA polymerase in an in vitro assay. In general, the activities of the mutant ZBDs in the primer extension assay mirrored their activities in primer synthesis with different mutants showing levels of activity ranging from wild-type activity to no activity (Fig. 8). However, several of the mutant proteins (S23A, D24A, N26A, F29A, and W42A) retained significant activity in the primer extension assay, although they were inactive in primer synthesis (Fig. 8). Ser-23, Asp-24, and Asn-26 are located in the connection between ␤-strands 2 and 3 of the ZBD, and these residues are less exposed on the surface that binds to DNA (Fig.  9). Phe-29 and Trp-42 are located at the base of the ZBD near the connection to the RPD. Although these residues flank His-33, an essential residue for both primer synthesis and primer extension, Phe-29 and Trp-42 are dispensable for primer utilization by T7 DNA polymerase.
Gene 2.5 Protein Inhibits Oligoribonucleotide Utilization by T7 DNA Polymerase-The SSB encoded by gp2.5 generally   FIG. 3. The HSQC spectrum of the ZBD changes dramatically upon binding to a DNA template annealed to the tetraribonucleotide 5-ACCC-3. Two-dimensional 1 H-15 N HSQC spectra of the ZBD are shown in the absence of DNA (a), in the presence of a primed DNA template (b), and in the presence of the DNA template only (c) (see "Experimental Procedures" for details). Upon binding to the primed template, many resonances (circled) are diminished or disappear altogether, indicating these residues are perturbed by binding interactions. The ZBD does not stably bind to the DNA template without the primer strand, and no change in the HSQC spectrum is observed (compare a and c). 4. The RPD, not the ZBD, of T7 primase binds to a ssDNA template. DNA binding assays were carried out in the reactions containing a 3Ј-labeled ssDNA template containing the priming site. The anisotropic fluorescent emission (r) of the 6-FAM-labeled DNA was measured at equilibrium in the presence of the indicated concentrations of T7 primase, the ZBD, or the RPD. The RPD bound ssDNA with an affinity similar to that of the intact primase, whereas the ZBD did not bind.
FIG. 5. HSQC resonance assignments for the ZBD of T7 primase. The assignments of the ZBD residues were carried out with six three-dimensional triple resonance spectra and two three-dimensional 15 N-edited TOCSY-HSQC and NOESY-HSQC spectra. The resonances were assigned for the backbone amides of 52 of 57 non-proline residues in the ZBD. The N-terminal five residues could not be assigned, and they are disordered in the crystal structure (15). Side chain resonances are designated with "S" after the residue numbers.
stimulates DNA replication (49 -51). gp2.5 physically interacts with T7 DNA polymerase (8), and this protein-protein interaction is required to couple DNA synthesis of the leading strand of the replication fork to synthesis of the lagging strand (5,6). The E. coli SSB protein is required for the initiation of DNA synthesis (1), and it stimulates the loading of E. coli DNA primase (DnaG) onto DNA (52). E. coli SSB also participates in the transfer of newly synthesized primers to E. coli DNA polymerase III (39). We therefore examined how the T7 gp2.5 affects primer utilization by T7 DNA polymerase. Primer extension reactions were carried out with a 10-fold molar excess of the ssDNA template and limiting amounts of gp2.5, T7 primase, and T7 DNA polymerase. In the absence of the primase, the initiation of DNA synthesis from the oligoribonucleotide primer 5Ј-pACCC-3Ј was very inefficient (Fig. 10a, lane  2). Addition of the full-length T7 primase-helicase, the primase fragment, or the ZBD to the primer extension reaction greatly stimulated primer utilization and DNA synthesis by T7 DNA polymerase (Fig. 10a, lanes 5, 8, and 11). Surprisingly the addition of gp2.5 specifically inhibited DNA synthesis initiated from a tetraribonucleotide primer either in the presence or in the absence of T7 primase (Fig. 10a, lanes 3, 6, 9, and 12). DNA synthesis that was primed by a longer DNA oligonucleotide was unaffected by the addition of gp2.5 (Fig. 10a, lanes 14 -16).
The inhibitory effect of gp2.5 does not result from gp2.5 binding to ssDNA and sequestering the template for DNA synthesis. Strong inhibition was observed with substoichiometric concentrations of gp2.5 in reactions with a 10-fold molar excess of the ssDNA template and oligoribonucleotide primer. A C-terminally truncated gp2.5 mutant (gp2.5⌬26) with decreased affinity for T7 DNA polymerase (9) had only a slight effect on primer utilization by T7 DNA polymerase (Fig. 10a,  lanes 4, 7, 10, and 13). gp2.5⌬26 binds to ssDNA with 40-fold higher affinity than wild-type gp2.5 (9,53), yet the mutant protein interfered weakly with primer utilization by T7 DNA polymerase alone, and gp2.5⌬26 did not inhibit DNA synthesis in the presence of T7 primase. These results indicate the interaction of T7 DNA polymerase with gp2.5 specifically interferes with DNA synthesis that is initiated from an oligoribonucleotide primer.
The opposite effects of gp2.5 and T7 primase on primer utilization by T7 DNA polymerase suggested that these pro-teins might compete for binding to T7 DNA polymerase. To test this idea, the concentration of each protein was varied in the presence of a fixed amount of the other protein. In the presence of 10 M T7 primase, gp2.5 was a potent inhibitor of primer utilization by T7 DNA polymerase (Fig. 10b) with 50% inhibition occurring in the presence of about 1 M gp2.5 protein.
Strong inhibition by gp2.5 was evident even when the primase (10 M) and the DNA template (50 M) were present in vast excess over gp2.5 (Fig. 10b). The gp2.5⌬26 protein inhibited the reaction with roughly 10-fold lower potency, consistent with the low affinity of gp2.5⌬26 for T7 DNA polymerase (9). In the absence of gp2.5, increasing concentrations of T7 primase stimulated primer utilization (Fig. 10c). However, the addition of more T7 primase did not overcome the inhibitory effect of gp2.5. Thus, the inhibition of oligoribonucleotide utilization by gp2.5 does not result from a simple competition between T7 primase and gp2.5 for binding to DNA polymerase.
A Physical Interaction between T7 DNA Polymerase and the ZBD-It has been shown previously that T7 primase-helicase remains associated with T7 DNA polymerase for several cycles of nucleotide insertion during DNA synthesis primed by a tetraribonucleotide (38). A monomeric mutant of T7 primasehelicase was pulled down along with T7 DNA polymerase on an immobilized DNA template after primer extension was blocked by the incorporation of a 2Ј,3Ј-dideoxynucleotide. In the absence of the primase-helicase, the polymerase binds weakly to the DNA template annealed to a tetraribonucleotide primer. The polymerase-DNA complex with the nascent primer is stabilized by the interaction with T7 primase. Once the growing primer strand is extended beyond a critical length, the polymerase no longer requires the primase to assist with DNA synthesis (38). We used a pull-down assay to examine the physical interaction of T7 DNA polymerase and T7 primase, or the ZBD, with an immobilized DNA template (Fig. 11). In the complete reaction (lane 1), DNA synthesis primed by an oligoribonucleotide was terminated by the insertion of a 2Ј,3Ј-dideoxynucleotide. Both the primase and T7 DNA polymerase remained bound to the immobilized template in the presence of a high concentration of the incoming nucleotide (dCTP) specified by the template sequence. If any of the nucleotide or protein components was omitted from the reaction, neither protein was pulled down with the immobilized DNA (Fig. 11, lanes 2-5).
The ZBD alone could stabilize the primer extension complex on an immobilized DNA (Fig. 11, lane 8) albeit less efficiently than T7 primase. In the complete reaction with the ZBD (Fig.  11, lane 8) the amount of T7 DNA polymerase remaining in the complex was about one-half the amount in the complex formed with T7 primase (Fig. 11, lane 1 versus lane 8). Although the ZBD co-migrated with a small amount of avidin released from the beads, the ZBD in the primer extension complex stained more intensely than the corresponding avidin band from a control reaction (lane 9). We conclude that the ZBD alone supports the formation of the priming complex even though the complex was less stable than the corresponding complex with T7 primase. It is notable that the ZBD had very low affinity for a primed template in the absence of T7 DNA polymerase (Fig.  11), suggesting that productive binding to the nascent primertemplate heteroduplex requires the participation of both proteins. Once assembled, this primer extension complex is resistant to disruption by gp2.5 (not shown).

DISCUSSION
The tetraribonucleotides synthesized by T7 primase are not effective primers for DNA synthesis by T7 DNA polymerase. The primase protein stimulates primer utilization by T7 DNA polymerase (37,54), forming a complex with the primed template that is subsequently engaged by T7 DNA polymerase to FIG. 6. Identification of the surface residues of the ZBD that interact with a primed DNA template. The resonances that were perturbed upon binding to a primed DNA template are shown on the corresponding residues of the ZBD structure, which is shown in stereo. A bound zinc atom is depicted as a green ball. NMR experiments were carried out with T7 primase (15) or the ZBD (Fig. 3). The residues within the ZBD showing changes in response to binding are labeled with blue circles for NMR experiments using T7 primase and with gray circles for experiments using the ZBD. The labels denote resonances that disappeared (filled circles) and those that shifted position in the HSQC spectrum (half-filled circles) in the presence of the primed DNA template.

T7 Primase-Polymerase Complex
initiate DNA synthesis (38). The small (59-residue) N-terminal ZBD of T7 primase promotes efficient utilization of oligoribonucleotides by the polymerase (15), raising questions about how the ZBD contacts the primed template and secures it in the polymerase active site. In this study, we showed that the ZBD binds specifically, albeit weakly, to the primed DNA template through interactions with a large number of aromatic and polar residues decorating one surface of the ZBD (Fig. 6). The primase and DNA polymerase bound cooperatively to the primed template with both enzymes required to stabilize the protein-DNA complex (Fig. 11). The utilization of tetraribonucleotide primers for DNA synthesis was specifically blocked by the T7 single-stranded DNA-binding protein gp2.5, raising the possibility that the interaction of gp2.5 with T7 DNA polymerase could regulate the initiation of Okazaki fragment synthesis on the lagging strand of the replication fork.
Two-dimensional 1 H- 15 N HSQC experiments showed that the isolated ZBD binds to a primed DNA template in the same manner as the ZBD within the intact primase (Figs. 3 and 6). The corresponding resonance peaks in the HSQC spectra of the ZBD and T7 primase were perturbed similarly upon binding to a primed DNA template. The NMR resonances that changed upon binding were assigned (Fig. 5), and most correspond to a patch of highly conserved residues (Phe-11, His-14, Phe-29, Asp-31, Phe-35, and Tyr-37) located on one surface the ZBD (Fig. 6). Alanine scanning mutagenesis confirmed the importance of residues on this surface of the ZBD for both primer synthesis and primer utilization activities (Figs. 7 and 8). The abundance of aromatic residues on the DNA binding surface of the ZBD is notable as these residues could stack against the nucleobases of the primer or template strand. These observations support and extend the previous model in which this surface of the ZBD engages the DNA during primer synthesis (Supplemental Fig. 2) (15,59). The interaction of the ZBD with substrates requires the primase to change conformation from the open conformation seen in the crystal structure of T7 primase to a closed conformation in which the ZBD and RPD can both engage substrates for primer synthesis (Fig. 1).
The ZBD bound specifically to a primer-template heteroduplex, but it failed to interact with the ssDNA template alone (Figs. 3 and 4). In contrast, the RPD had ssDNA binding activity that was comparable to that of the intact primase (Fig.  4). The RPD may initiate the interaction of the primase with a DNA template before the ZBD is brought into position for FIG. 7. Conserved residues of ZBD are important for binding to the primed template and mediating primer synthesis and extension activities during replication. The sequences of the ZBDs of bacteriophage primases were aligned using the program ClustalW (58), and the secondary structure of the T7 ZBD is shown above the sequence. The sequences shown are: BPT7, bacteriophage T7; BPT3, bacteriophage T3; BPYe0312, bacteriophage Ye0312; PPGH1, Pseudomonas spp. phage GH-1; PPKT2400, Pseudomonas putida (strain KT2440); CPP60, cyanophage P60; RPSI01, roseophage SI01. The conserved residues are highlighted by color. Highly conserved residues are colored red or green, and strongly conserved residues are colored dark or light gray. The results of the resonance shifts and alanine scanning mutagenesis are summarized below the sequences using the same labeling scheme as in Figs. 6 and 9. The biochemical activities of the mutant ZBDs are also summarized (see Fig. 8). Primer synthesis activity (red circles) and stimulation of primer utilization by T7 DNA polymerase (green circles) are shown. Filled circles denote a complete loss of detectable activity, half-filled circles indicate mutants with significantly decreased activity, and open circles denote a wild-type level of activity.

FIG. 8. Primer synthesis and extension activities of mutant
ZBDs. The activities of mutant ZBDs containing alanine substitutions at the indicated positions were measured for primer synthesis and for the stimulation of primer utilization by T7 DNA polymerase. The mutant ZBDs were overexpressed in E. coli and purified, and their activities were assayed in reactions containing the purified wild-type RPD of T7 primase (see "Experimental Procedures" for details). These results are summarized in Fig. 9. WT, wild type. Filled circles indicate a complete loss of activity resulting from substitution of alanine; half-filled circles denote a significant decrease in activity, and open circles denote a wild-type level of activity of alanine scanning mutagenesis of the ZBD are mapped onto a stereograph of the ZBD structure (see Fig. 8). The residues affected by alanine substitutions are labeled with red circles denoting primer synthesis activity and green circles denoting stimulation of primer utilization by T7 DNA. primer synthesis. Two residues located in the ZBD of the primase (Asp-31 and His-33) are determinants of primer synthesis initiated at 5Ј-GTC-3Ј priming sites (32). Alanine substitutions involving Asp-31 or His-33 strongly interfered with primer synthesis activity and the stimulation of primer utilization (Figs. 8 and 9). However, our NMR experiments with the ZBD failed to detect an interaction of these residues with DNA ( Figs.  6 and 7). The resonances for Asp-31 and His-33 are unresolved in the HSQC spectrum of T7 primase (15), so we cannot determine whether or not these residues interact with DNA in the more stable complex with T7 primase.
The T7 primase and T7 DNA polymerase collaborate to initiate DNA synthesis, forming a multiprotein complex that could be isolated on an immobilized DNA template (Fig. 11). The T7 single-stranded DNA-binding protein gp2.5 specifically interfered with DNA synthesis that was initiated from an oligoribonucleotide (Fig. 10). Longer oligonucleotide primers were efficiently extended by T7 DNA polymerase without assistance from the primase, and this reaction was unaffected by gp2.5 (Fig. 10a, lanes 14 -16). Several lines of evidence suggest that gp2.5 inhibits primer utilization by binding to T7 DNA polymerase rather than by sequestering the DNA template or dislodging the oligoribonucleotide from the template. First, gp2.5 strongly inhibited primer utilization in the presence of a vast excess of ssDNA template and tetraribonucleotide primer. Second, a mutant gp2.5 protein (gp2.5⌬26) with decreased binding affinity for T7 DNA polymerase was a less potent inhibitor of primer utilization even though gp2.5⌬26 binds to ssDNA 40fold more tightly than wild-type gp2.5 (53). Although gp2.5 binds to T7 primase-helicase (9), it is unlikely this interaction involves the region comprising the primase. An interaction between gp2.5 and T7 primase has not been reported, and an N-terminal truncation of the primase-helicase protein that removes the ZBD of the primase does not affect the interaction with gp2.5 (9). This evidence suggests that gp2.5 inhibits primer utilization by T7 DNA polymerase through its interaction with T7 DNA polymerase instead of binding to the primase or the primed DNA template.
The inhibitory effect of gp2.5 on the utilization of oligoribonucleotide primers is surprising because it has been shown previously that gp2.5 increases the initiation of RNA-primed DNA synthesis catalyzed by T7 primase-helicase and T7 DNA polymerase (50). However, there are two significant differences between our experiments and those in the previous report. The first difference is the length of the template used for DNA synthesis. We used a short oligonucleotide (26-mer) instead of M13 ssDNA (50). gp2.5 stimulates DNA synthesis by T7 DNA polymerase on large ssDNA templates (49,55) probably by preventing secondary structure that interferes with templated DNA synthesis. In addition, stoichiometric amounts of gp2.5 or E. coli SSB facilitate strand displacement synthesis and thereby increase the extent of DNA synthesis. We observed no effect of gp2.5 on DNA synthesis catalyzed by T7 DNA polymerase using an oligonucleotide template (Fig. 10a, lanes 14 -16), and there is no effect of gp2.5 on the efficiency of primer synthesis by T7 primase on an oligo template (data not shown). The other difference between our experiments and the previous report (50) is the amount of gp2.5 protein used in comparison to the concentration of ssDNA binding sites in the reaction. gp2.5 protein stimulates RNA-primed DNA synthesis when it is present in molar excess over the ssDNA binding sites (based on ϳ7 nucleotides bound per monomer) (56). In contrast, we observed that primer utilization was inhibited by substoichiometric amounts of gp2.5 mixed with a 10-fold or greater excess of template and tetraribonucleotide primer. It is also worth noting that gp2.5 does not bind efficiently to oligos less than 30 nucleotides in length (57). These results highlight two different functions of gp2.5, promoting DNA synthesis through its DNA binding activity and selectively blocking DNA synthesis initiated from an oligoribonucleotide primer.
What is the biological significance of the inhibitory interaction of gp2.5 with T7 DNA polymerase? The physical interaction of gp2.5 and the DNA polymerase could serve two functions during replication: to retain the lagging strand polymerase in the replication complex during polymerase recycling and to regulate the initiation of Okazaki fragment synthesis. gp2.5 is an essential T7 replication factor that coordinates the synthesis of the leading strand of the replication fork with DNA synthesis on the lagging strand (5, 6). gp2.5 coats the ssDNA template on the lagging strand (5,6) and is therefore well positioned to participate in a handoff of the polymerase from the end of one Okazaki fragment to the initiation site for the next. The inhibitory interaction of gp2.5 with T7 DNA The utilization of tetraribonucleotide primers is strongly inhibited by the addition of the wild-type (W) gp2.5 ssDNA-binding protein (lanes 3, 6, 9, and 12). DNA synthesis primed by a 14-mer oligonucleotide is unaffected by the addition of gp2.5 (lanes 14 and 15). A C-terminally truncated mutant of the gp2.5 protein (gp2.5⌬26 (⌬)) that binds tightly to DNA but interacts weakly with T7 DNA polymerase is less effective at inhibiting primer utilization by the polymerase (lanes 4, 7, 10, and 13). b, the primer extension activity of T7 DNA polymerase was monitored in the presence of different concentrations of gp2.5 and fixed amounts of T7 primase (10 M) and DNA polymerase (1 M). The relative enhancements of primer extension activity, in comparison to the background activity measured in the absence of primase, are plotted for reactions containing different concentrations of wild-type gp2.5 (closed circles) or the gp2.5⌬26 mutant (open circles). c, the primer extension activity of T7 DNA polymerase was monitored in the presence of different concentrations of T7 primase and a fixed concentration of the DNA polymerase (1 M) in the absence or presence of gp2.5 or gp2.5⌬26 (1.75 M). The relative enhancement of primer utilization by the polymerase is plotted for reactions without gp2.5 (diamonds) or with wild-type gp2.5 (filled circles) or gp2.5⌬26 (open circles). WT, wild type. polymerase may correspond to an intermediate step during the recycling of the polymerase from one Okazaki fragment to the next. The polymerase in this intermediary complex is prevented from initiating DNA synthesis until it is transferred to the priming site where the primase initiates the next round of DNA synthesis. What triggers the release of the polymerase to allow productive interactions with the primed template? The proposed handoff of DNA polymerase from gp2.5 to T7 primase might be facilitated by protein interactions that are not present in the limited reactions studied here. Studies using a reconstituted T7 replication system (5,6) can address this question. FIG. 11. Formation of a stable priming complex between T7 DNA polymerase and T7 primase. The priming complex that is formed when the primase delivers a primed DNA template to the DNA polymerase was isolated by a pull-down experiment using an immobilized ssDNA template attached to avidin-agarose beads through a 3Ј-biotin group. The priming complex between T7 DNA polymerase and T7 primase was captured in the complete reaction (lane 1), whereas when either component was left out from the reaction, the proteins did not form the stable complex (lanes 2-5). The purified proteins are shown in lane 6 (T7 DNA polymerase) and lane 7 (T7 primase), respectively. In the same way, the priming complex between T7 DNA polymerase and the ZBD was captured (lane 8). Although the ZBD co-migrates with streptavidin that came off of the agarose beads, lane 8 has a more intense signal than lane 9 at the ZBD/avidin position, indicating that the ZBD is captured by the immobilized template together with the DNA polymerase. pol, polymerase.