Search tips
Search criteria 


Logo of narLink to Publisher's site
Nucleic Acids Res. 2010 November; 38(20): 7167–7178.
Published online 2010 June 30. doi:  10.1093/nar/gkq588
PMCID: PMC2978363

Class-specific restrictions define primase interactions with DNA template and replicative helicase


Bacterial primase is stimulated by replicative helicase to produce RNA primers that are essential for DNA replication. To identify mechanisms regulating primase activity, we characterized primase initiation specificity and interactions with the replicative helicase for gram-positive Firmicutes (Staphylococcus, Bacillus and Geobacillus) and gram-negative Proteobacteria (Escherichia, Yersinia and Pseudomonas). Contributions of the primase zinc-binding domain, RNA polymerase domain and helicase-binding domain on de novo primer synthesis were determined using mutated, truncated, chimeric and wild-type primases. Key residues in the β4 strand of the primase zinc-binding domain defined class-associated trinucleotide recognition and substitution of these amino acids transferred specificity across classes. A change in template recognition provided functional evidence for interaction in trans between the zinc-binding domain and RNA polymerase domain of two separate primases. Helicase binding to the primase C-terminal helicase-binding domain modulated RNA primer length in a species-specific manner and productive interactions paralleled genetic relatedness. Results demonstrated that primase template specificity is conserved within a bacterial class, whereas the primase–helicase interaction has co-evolved within each species.


DNA replication is required by all autonomous organisms to duplicate and transmit their genetic material. During this process, the replicative helicase unwinds the chromosomal double helix while primase catalyzes the de novo synthesis of a template-directed RNA polymer at least once on the leading strand template and numerous times on the lagging strand template (1,2). Primase, encoded by dnaG in bacteria, provides RNA primers with a free 3′-hydroxyl group required by all known DNA polymerases. Extension of the RNA primers by the DNA polymerase holoenzyme results in the formation of Okazaki fragments of approximately 1000 to 2000 bases in bacteria (3).

The modular structure of bacterial DnaG primase includes an N-terminal zinc-binding domain (ZBD) that generally contains three cysteines and one histidine that coordinate the binding of a zinc ion (4,5). The central RNA polymerase domain (RPD) contains the active site where metal-dependent polymerization occurs (6). The C-terminal helicase-binding domain (HBD) of DnaG directly interacts with helicase (7–9). Fluorescence resonance energy transfer and in vitro activity assays indicated that two truncated primases can interact, suggesting that the extended ZBD from one primase binds to the DNA template within the RPD catalytic center of the other truncated protein (10).

Primases from both bacteriophages and bacteria initiate de novo primer synthesis on defined trinucleotides (11). The Firmicute primases from the Bacilli class, specifically Staphylococcus aureus and Geobacillus stearothermophilus (formerly known as Bacillus stearothermophilus), predominantly prime 5′-d(CTA) with moderate initiation on 5′-d(TTA) (12,13). DnaG from the γ-Proteobacteria Escherichia coli catalyzes primer synthesis on 5′-d(CTG) (14,15) and the Aquificae member Aquifex aeolicus initiates preferentially on 5′-d(CCC) (16).

Evidence to date suggests that the bacterial replicative helicase stimulates primase activity by acting as a platform for primase, thereby increasing the local concentration of the ssDNA template (9,10,17). Stoichiometric analyses of this critical protein–protein interaction under saturating conditions indicate that three primases are present on each helicase hexamer (9,18). The helicase N-terminal domain directly interacts with the primase HBD (9,19–21), while the C-terminal domain of helicase is responsible for the hexameric quaternary structure and for hydrolyzing ATP (22,23). The functional interaction between the HBD and helicase N-terminal domain appears to be specific, since DnaG from E. coli and S. aureus are not stimulated by the heterologous helicase (12). Further, E. coli helicase is unable to productively interact with B. anthracis primase (24).

Important questions remain regarding the mechanistic process by which bacterial primases recognize specific trinucleotides, including whether this attribute is conserved within bacterial classes. Moreover, the evolutionary and functional relationship between the essential primase-helicase interaction has not been defined, nor is it known whether orthologs of these proteins from phylogenetic siblings can productively interact. To begin to address these questions and better understand the species-specific differences in primase and helicase, we functionally characterized these enzymes from multiple species within the divergent gram-positive Firmicutes and gram-negative Proteobacteria. Our findings suggest that core sequences in the primase ZBD predict template specificity within a bacterial class and that the primase HBD–helicase interaction has co-evolved despite differing environmental niches for mesophilic and thermophilic bacteria.


Strains and nucleic acids

Bacterial strains were obtained from ATCC (B. anthracis Sterne and Pseudomonas aeruginosa PA01) or from the Armed Forces Institute of Pathology following the requirements of the Select Agent Program as outlined in the APHIS/CDC Form 2 (Yersinia pestis CO92) or as previously published (S. aureus, E. coli and G. stearothermophilus) (12,25,26). All manipulations of viable Y. pestis material were performed by authorized individuals within BSL-3 laboratories verified for select agent work using laboratory biosafety criteria as required. The E. coli dnaG mutant CR34/308 was obtained from the Coli Genetic Stock Center at Yale University. RNA standards were synthesized by Invitrogen and deoxyribonucleotide templates and PCR primers were generated by Integrated DNA Technologies.

Cloning and sequence analysis of primases and helicases

Primase and replicative helicase from G. stearothermophilus were cloned as previously described (25,26). Cloning strategies used for the construction of the truncated, chimeric and wild-type primases, as well as replicative helicases are described in the Supplementary Data and the PCR primers are listed in Supplementary Table S3. Nucleotide sequencing was performed by the UNMC Molecular Biology Core Facility. Sequences were aligned using the BESTFIT and PILEUP programs in the Wisconsin Package computation software (Accelrys Inc.) and/or ClustalW2.

Expression and purification of primases and helicases

Geobacillus stearothermophilus primase and replicative helicase expression and purification were performed as previously described (25,26). Truncated, chimeric, mutated and wild-type primases, as well as replicative helicases were expressed in E. coli BL21(DE3) and purified by nickel affinity chromatography as suggested by the manufacturers (Novagen EMD Biosciences, GE Healthcare). Enzyme purity and concentration were determined by 10% (w/v) SDS–PAGE analysis and spectrophotometry using the appropriate ε at 280 nm (Supplementary Table S4) and determined to be greater than 98% pure.

Site-directed mutagenesis of primase

Site-directed mutagenesis of the S. aureus primase ZBD was carried out using the QuikChange Multi Site-directed Mutagenesis Kit (Stratagene) with the appropriate oligonucleotides according to the manufacturer’s recommendations. The presence of desired mutations and the absence of spurious mutations were confirmed by DNA sequencing.

RNA primer synthesis assay and thermally denaturing HPLC analysis

Reaction conditions and mixtures for mesophilic primases were used as previously optimized for S. aureus and E. coli with 400 µM of each NTP (12,27). Assay conditions for optimum G. stearothermophilus primase activity were determined using 1 mM of each NTP. All RNA primer synthesis reactions were performed in 100 µl nuclease-free water containing 50 mM HEPES, 100 mM potassium glutamate, 10 mM DTT, 10 mM magnesium acetate, and 2 µM of the designated template. Priming reactions contained 3.6 µM primase, unless otherwise noted, and were incubated for 1 h at the optimum temperature for primer synthesis. Reactions containing DnaG from mesophilic bacteria were incubated at 30°C and from thermophilic G. stearothermophilus at 50°C or 37°C in the absence or presence of helicase, respectively. Samples were processed as previously described using HPLC under thermally denaturing conditions at 80°C (16). Retention times of the primers were correlated to the retention times of the appropriate RNA standard (Supplementary Table S5) to confirm composition and length. RNA primers were quantified by using the ssDNA template as an internal control to normalize for system variability. All reactions were performed in triplicate, a minimum of three independent experiments with at least two different protein preparations.

Phylogenetic analyses

Protein sequences were aligned using the default settings of ClustalW2. Each multiple sequence alignment of the full-length proteins or their domains was analyzed using the SEQBOOT program in PHYLIP 3.69 package to generate 100 random trees per data set. The PRODIST algorithm was then utilized to generate distances for each new random bootstrap tree and the FITCH program was used to minimize the distances. CONSENSE created a majority-rule consensus tree with bootstrap values. The final unrooted trees were created using the DRAWTREE program.

In vivo complementation assay

The E. coli CR34/308 RPD mutations P207L and Q252L in the dnaG allele were confirmed by PCR amplification and subsequent DNA sequencing of the cloned amplicon. To test complementation, E. coli dnaG mutant CR34/308 was transformed with the designated expression plasmid by electroporation. Equal amounts of the transformation mixture were plated for growth on LB agar plates supplemented with 0.36% glucose and 0.01% thiamine and assessed after overnight incubation at either the permissive temperature of 30°C or the nonpermissive temperature of 42°C. Untransformed and transformed cells were plated on both supplemented LB agar plates with or without kanamycin to confirm selection for those cells containing the kanamycin-resistant expression plasmid and incubated at both temperatures.


Trinucleotide initiation specificity for primase is class-specific

Based on earlier studies demonstrating that DnaG primase from gram-positive S. aureus and G. stearothermophilus initiated RNA primer synthesis from d(CTA) and d(TTA) and that the initiation trinucleotide for gram-negative E. coli was d(CTG), we hypothesized that trinucleotide initiation specificity would be restricted at the bacterial class level. To test this question, DnaG from the Bacilli Firmicute B. anthracis and the γ-Proteobacteria Y. pestis and P. aeruginosa were first evaluated in our template screening assay that tests initiation on all 64 possible trinucleotides. RNA polymer synthesis and the resulting lengths of these products were assessed to predict possible initiation trinucleotides. DNA templates with the sequence 5′-CAGA(CA)5 XYZ (CA)3, where XYZ is the trinucleotide of interest, were then used in the primer synthesis assay to confirm initiation specificity (Figure 1A). The RNA products were analyzed by thermally denaturing high-performance liquid chromatography (HPLC), a methodology that provides more precise qualitative and consistent quantitative data than traditional gel electrophoresis with radiolabeled NTPs (28).

Figure 1.
Trinucleotide initiation specificity for primases from B. anthracis, Y. pestis and P. aeruginosa. (A) Schematic of primase activity assay using the 23-mer trinucleotide-specific template where XYZ is the trinucleotide of interest (upper sequence) and ...

As shown in Figure 1B, the gram-positive B. anthracis primase preferentially initiated from d(CTA) and to a slightly lesser degree on d(TTA), similar to the primases from S. aureus and G. stearothermophilus. For both the d(CTA) and d(TTA) templates, the major primer product synthesized by B. anthracis DnaG was a full-length 16-mer, with the former RNA polymer eluting earlier, consistent with its more hydrophilic composition. The primases from gram-negative Y. pestis and P. aeruginosa preferentially primed on d(CTG) and to a slightly reduced extent on d(CTA) (Figure 1C and D). Although E. coli DnaG was known to initiate on d(CTG), d(CTA) was also recognized (Figures 2A and 3A; Supplementary Figure S1), concurring with in vivo analyses in which the majority of RNA primers began with the sequence 5′-pppAG (29). Therefore, d(CTA) serves as a common initiation trinucleotide for DnaG from both the Firmicutes and Proteobacteria, representing a possible ancestral link with the presumably antecedent Firmicute clade. These data demonstrate that trinucleotide recognition by primase within these bacterial classes is conserved.

Figure 2.
Contribution of the primase HBD on initiation specificity. (A–C) Comparison of primer synthesis by wild-type DnaG from S. aureus (Sa) and E. coli (Ec) to chimeric primases in which the HBD was exchanged. Primases were tested with the 23-mer template ...

CTP titration experiments using the 23-mer CTA-specific template containing a single guanine at the antepenultimate position (Figure 1A) were performed to determine the initial nucleotide copied. The absence of CTP resulted in the predominant synthesis of a 13-mer RNA product, while increasing the CTP concentration decreased synthesis of the 13-mer with a concurrent increase in the full-length 16-mer primer (Supplementary Figure S2). A minor amount of RNA polymers longer than 13 ribonucleotides were also observed in the absence of CTP due to nucleotide misinsertion, a fidelity aspect shared by the primases from S. aureus and E. coli (27,30). These results indicate that DnaG from B. anthracis, Y. pestis and P. aeruginosa initiate from the middle nucleotide in the recognition trinucleotide, analogous to other fully characterized bacterial primases.

Primase ZBD and RPD interactions determine initiation of RNA primer synthesis

Experiments using the E. coli, A. aeolicus and T7 systems showed that the ZBD and RPD were sufficient to catalyze template-directed oligoribonucleotide synthesis (10,16,31). However, the potential influence of the HBD on primase specificity has not been examined. We found that the S. aureus and E. coli primases without their HBD produced RNA primers that were qualitatively and quantitatively similar to their wild-type equivalent, with the exception of truncated E. coli primase on d(CTG) (Supplementary Figure S1). HBD-swapped S. aureus and E. coli primases retained trinucleotide specificity of the respective ZBD+RPD; however, differences in primer synthesis efficiency were observed (Figure 2A–C). Collectively, these data demonstrate that neither the HBD of S. aureus nor E. coli primases regulates class-specific trinucleotide recognition.

The exchange of the ZBD between a Firmicute and Proteobacteria was predicted to transfer template specificity. To examine whether the ZBD was solely responsible for recognition of the initiation site, we tested the activity of a chimera that contained the ZBD of S. aureus DnaG and the RPD+HBD of E. coli primase. No priming occurred on templates containing the preferred Proteobacterial trinucleotide d(CTG); however, initiation did occur on d(CTA) and to a considerably lesser degree on d(TTA) (Figure 3A). A minor amount of primers that were longer than the expected 16-mer were synthesized on templates containing either d(CTA) or d(TTA), indicating that non-specific initiation occurred 3′ to these trinucleotides. This additional priming reflected the activity observed for wild-type E. coli DnaG and the enzymes containing the E. coli RPD, differing from the high specificity observed for wild-type S. aureus primase (Figures 2 and 3A; Supplementary Figure S1). Several studies have determined that relatively rigid polymerase active sites exhibit higher specificity than flexible catalytic centers (32–34). Therefore, the promiscuous priming observed for E. coli DnaG may be due to a slightly more flexible RPD active site relative to other bacterial primases.

Figure 3.
Key residues in the primase ZBD determine initiation specificity. (A) Denaturing HPLC analyses of RNA primers synthesized by the ZBD-swapped chimeric primase compared to wild-type primase from S. aureus (Sa) or E. coli (Ec) on templates containing d(CTA), ...

Overall, the negligible initiation on d(TTA) and the lack of primer synthesis on d(CTG) demonstrated that neither the presence of the S. aureus ZBD nor the E. coli RPD+HBD are independently sufficient to confer class-associated trinucleotide initiation specificity. These findings indicate that a heterologous ZBD and RPD must interact in a precise manner for specific and significant initiation. In addition, template recognition must be determined by either critical amino acids in both the cognate ZBD and RPD or alternatively, key residues are located in only one of these domains.

Key residues in the ZBD β4 strand confer trinucleotide specificity

The ZBD crystal structure from G. stearothermophilus revealed a number of conserved basic and hydrophobic residues on the outer or exposed surface of the β sheet (4). We hypothesized that residues in the ZBD β sheet, conserved within members of the Firmicutes or Proteobacteria, and spatially located such that they potentially interact with the ssDNA template, contribute to class-associated trinucleotide specificity. To test this possibility, a sequence alignment of the ZBD from those primases with known initiation specificity was created using an algorithm that clustered the primases into their respective bacterial class. This alignment identified residues in the β sheet that were conserved between members of either the Firmicutes or Proteobacteria (Figure 3B). To assess the contribution of these residues and several other prospective amino acids on template specificity, S. aureus DnaG mutants were created by converting the ZBD residue to the corresponding amino acid in Proteobacterial primases (R32K, L37C, I56F, C57Y, E28K and D42H).

All S. aureus ZBD mutants were initially tested with the two screening templates that together comprise all 64 possible trinucleotides to evaluate enzyme activity and to determine if a new specificity was created. Template specificity was then confirmed using a 23-mer trinucleotide-specific template in the RNA priming assay (Figure 1A). All of the single-residue ZBD mutants exhibited the highest activity on the d(CTA) template, indicating that none of these amino acids alone were capable of changing the initiation specificity (Figure 3C). Double and triple substitutions in the ZBD of S. aureus DnaG were also evaluated (I56F/C57Y, E28K/D42H, R32K/I35H, R32K/G65H and E28K/D42H/C57Y). Of these ZBD mutants, only the S. aureus I56F/C57Y mutant gained function on d(CTG) (Figure 3C). These findings confirm that these two adjacent amino acids in the ZBD β4 strand are the dominant residues conferring specificity, and that trinucleotide recognition and initiation can be transferred across bacterial classes by providing the appropriate residues in the same β sheet location.

Class-associated trinucleotide specificity regulates a trans interaction between the ZBD and RPD of chimeric primases

Structural and activity analyses by Berger and associates demonstrated that the ZBD of one truncated primase can interact with the RPD active site of a second truncated primase (10). Class-associated trinucleotide specificity and the construction of two chimeras with different activities provided for a direct examination of this phenomenon for full-length bacterial primases. The chimeric enzyme containing the S. aureus ZBD and E. coli RPD+HBD was capable of primer synthesis on templates containing d(CTA), but had little or no activity on the preferred Proteobacterial trinucleotide d(CTG) (Figures 3A and 4A). The second chimera comprised the E. coli ZBD and the S. aureus RPD+HBD had no intrinsic activity on d(CTG) (Figure 4B) or any template (not shown). Mixing studies were performed with the expectation that if the E. coli ZBD conferred template recognition and initiation in trans, then de novo primer synthesis should occur on d(CTG). As shown in Figure 4C and D, the addition of the inactive E. coli ZBD chimeric primase to the reaction containing the chimera with the S. aureus ZBD and the E. coli RPD+HBD resulted in the production of primers on d(CTG). Doubling the molar concentration of the chimeric primases resulted in a 3-fold increase in primer synthesis. The most plausible explanation for the gain in function on d(CTG) or change in template specificity was that the E. coli ZBD on the inactive chimera was functionally interacting in trans with the E. coli RPD on the other full-length chimeric primase.

Figure 4.
Full-length chimeric primases interact in trans on class-associated trinucleotides. (A and B) Representative chromatograms comparing RNA primer production on the Gamma-proteobacterial initiation trinucleotide d(CTG) by the chimeric primase comprising ...

A second phenomenon occurred that had been previously observed with E. coli DnaG in that helicase stimulation resulted in shorter primers (35). In the mixed chimera activity assays, the major RNA product was a 14-mer rather than a full-length 16-mer (Figure 4C and D), similar to E. coli primer lengths in vivo (29). This result further substantiates the model that the intersubunit interaction restricts primer length and suggests an important mechanism that contributes to the regulation of primase processivity even in the absence of helicase.

Productive helicase stimulation of primase parallels genetic relatedness

Before testing the cross-reactivity of DnaG primases and replicative helicases within the same bacterial class, the primary sequence relatedness of each full-length protein and their associated domains were established for three Firmicutes and three Proteobacteria (Figure 5A–G; Supplementary Tables S1 and S2). The phylogenies of these full-length enzymes indicated there was a much stronger divergence between the Firmicute and Proteobacteria sequences for primase compared to replicative helicase (Figure 5A and B). The observed phyla split for DnaG is consistent with their template initiation specificities. Within each clade, the primases from E. coli and Y. pestis were the most similar followed by the primases from G. stearothermophilus and B. anthracis (Figure 5A; Supplementary Table S1). These findings suggested an evolutionarily conserved structural relationship that may translate into interchangeable functional activity. Based on these findings, we hypothesized that only primases and helicases that share considerable sequence homology within the interacting domains would cross-stimulate primer synthesis.

Figure 5.
Cross-stimulation of primase activity by replicative helicases. (A–G) Unrooted phylogenetic trees based on sequence identities for DnaG primase and replicative helicase and their associated domains from the Bacilli Firmicutes S. aureus (Sa), ...

To assess the evolutionary relationship of the domains that comprise DnaG and replicative helicase, the sequences for their associated domains were compared. Domain boundaries were based on published limited tryptic digestion studies and the resolved crystal structure for each domain. The phyla splits and branch lengths for the three primase domains were very similar to each other and to the full-length protein, despite their wide range of sequence similarities (Figure 5C–E; Supplementary Table S1). The ZBD had the greatest sequence identity and the HBD had the least. The strong phyla split for all three primase domains suggested that swapping domains between the two phyla would probably not result in a transfer of functionality. As was observed with full-length primase and the primase domain phylogenies (Figure 5A and C-E), the two helicase domain phylogenetic trees for the E. coli and Y. pestis clade branched the closest, followed by the G. stearothermophilus and B. anthracis clade (Figure 5F and G). Together these observations predicted that if heterologous stimulation was feasible, the highest level of functional cross-reactivity would occur between primase and helicase from these two pairs of bacteria.

Mixing experiments demonstrated that helicase from the Firmicutes B. anthracis and G. stearothermophilus cross-stimulated the heterologous primase for enhanced primer synthesis (Figure 5H). Similarly, DnaG and helicase from the Proteobacteria E. coli and Y. pestis productively interacted (Figure 5I). Interestingly, Y. pestis helicase stimulated E. coli primase to synthesize full-length primers rather than shorter RNA products, in contrast to the native E. coli DnaG–helicase interaction (Figure 5I). However, neither S. aureus nor P. aeruginosa was capable of productive interactions with other members of the same bacterial class (data not shown). Together these results indicate that genetic relatedness predicts the outcome of structure–function activity for helicase-stimulated primase activity.

Both E. coli and S. aureus helicase stimulation of DnaG results in a quantitative increase in the amount of primers produced, but RNA polymer lengths are reduced in E. coli (12,35). In this study, B. anthracis, G. stearothermophilus, Y. pestis and P. aeruginosa helicases stimulated their cognate primase to synthesize predominately full-length primers (Figure 5H and I; data not shown). Therefore, the E. coli helicase–DnaG interaction differs from other Proteobacteria and Firmicutes in terms of primer lengths produced. To investigate whether this effect was transferrable, both the S. aureus and E. coli HBD-swapped chimeric primases were tested with their cognate HBD-associated helicase. S. aureus helicase stimulated a modest increase in the production of predominantly full-length RNA primers by the chimeric primase containing the E. coli ZBD and RPD with the S. aureus HBD (Supplementary Figure S3A). In contrast, the chimera with the S. aureus ZBD and RPD and E. coli HBD synthesized shorter primers in the presence of E. coli helicase (Supplementary Figure S3B). These data demonstrate that the interplay between helicase and the cognate HBD modulates primer synthesis in a manner, mirroring the species-specific wild-type primase–helicase interaction.

In vivo complementation of an E. coli strain with a temperature-sensitive DnaG mutation

To determine whether RPD mutations could be rescued in vivo, we employed genetic complementation assays using the E. coli CR34/308 mutant carrying the thermosensitive dnaG allele. CR34/308 is a ‘quick-stop mutant’ that terminates DNA synthesis at the restrictive temperature of 42°C (36). This E. coli dnaG mutant contains two substitutions in the region that encodes the RPD, specifically P207L and Q252L. These conditional-lethal mutations in the catalytic core of this essential gene product allow for DNA replication to proceed at the permissive temperature of 30°C, but not at 42°C. Transformation of the empty plasmid did not rescue the mutant strain, whereas introduction of plasmid with E. coli dnaG permitted robust cell growth at the nonpermissive temperature of 42°C (Table 1). Moreover, transformation of the E. coli dnaG mutant with a plasmid expressing an N-terminally histidine-tagged E. coli DnaG rescued growth at 42°C, indicating that covalent linkage of the N-terminal tag does not affect the in vivo function of primase.

Table 1.
In vivo complementation of an E. coli strain with a temperature-sensitive DnaG mutationa

When genetic complementation assays were performed with plasmids expressing Y. pestis, P. aeruginosa or S. aureus primases, only expression of Y. pestis DnaG rescued the E. coli dnaG mutant strain at the restrictive temperature (Table 1). This in vivo result concurred with the in vitro experiments in which mutually productive cross-species stimulation of primer synthesis was obtained between primase and helicase from E. coli and Y. pestis (Figure 5I). These data indicate that primary sequence similarities directly correlate with the ability of primase and helicase from two different species to productively interact and function in vivo.


Class-restricted trinucleotide initiation specificity of bacterial primase

The demonstration of primase initiation on d(CTA) and d(TTA) for Firmicutes and d(CTG) and d(CTA) for Proteobacteria supports the general contention that members of the same phylogenetic class share trinucleotide template specificity. All three of these recognition trinucleotides code for leucine, an amino acid often determined to be among the most abundant in bacterial proteins. Given the high chromosomal skew for coding sequences, a high abundance of these three codons on the lagging strand template would provide a greater amount of initiation templates where primase acts. In each of these initiation sequence templates, the first two nucleotides are pyrimidines and the third is a purine. Initiation from these sequence motifs ensures that the first two primer NTPs are purines, bases known to stack strongly, thereby enhancing complex formation and subsequent dinucleotide condensation.

In the bacteriophage T7 primase, both Asp31 and His33 were determined to be essential for template specificity (37). Mutation of His33 to Ala altered synthesis at sequences containing a cryptic cytosine in the T7 primase recognition trinucleotide d(GTC) to d(GTA) and d(GTG). We determined that substitution of both Ile56 and Cys57 in the ZBD β4 strand of S. aureus DnaG to the Proteobacterial ZBD residues Phe and Tyr was required to switch specificity from d(CTA) to d(CTG). Similar to the findings in the T7 system, the preference for the cryptic third base in the initiation trinucleotide was altered. The spatial location of these two ZBD residues is on the surface of a β sheet proposed to bind ssDNA (Figure 6A) (4) and are structurally equivalent to Asp31 and His33 in the T7 primase. These key S. aureus and T7 ZBD residues are near the third residue (Cys) involved in zinc binding, indicating a conserved architecture for bacterial and bacteriophage primases and a possible communication between the zinc ion and nucleic acids.

Figure 6.
Regulation of primer synthesis by critical DnaG primase and replicative helicase residues. (A) Comparison of the S. aureus (Sa) and E. coli (Ec) primase ZBD models based on consensus structures obtained for A. aeolicus and G. stearothermophilus (PDB codes ...

Implications for trans interaction between the ZBD and RPD of primases

Productive crosstalk between primase subunits has been demonstrated in both the bacteriophage T4 and T7 by complementing catalytically inactive, full-length mutant primase with a truncated primase containing only the RPD (38,39). In bacterial systems, structural analyses suggested that primase without the HBD alternates between two different conformations: a compact form and an extended form (10). Our findings support the emerging paradigm for bacteria that the extended conformation of full-length DnaG is active and that the trans mode of interplay between subunits provides a regulatory mechanism by which site-specific primer initiation and elongation is restricted.

Even though each zinc finger domain in a classical zinc finger transcription factor recognizes and binds three nucleotides, three or more zinc finger domains are required for higher specificity and binding affinity to the 8 or 9 base pair recognition sequence. In the proposed model (Figure 6B), bacterial primase has a single zinc finger for sequence-specific recognition, binding and initiation that interacts in trans with the RPD on a separate primase. An exquisite level of specificity is achieved since the ZBD/RPD interaction can discriminate between a single nucleotide. Since hexameric helicase binds several primases, the close proximity of DnaG would further enhance ZBD and RPD intersubunit interactions.

Species specificity of primase-helicase interactions

Structural differences in the 5-bundle helix within the HBD of primase from the Firmicute G. stearothermophilus and the Proteobacteria E. coli have been determined (9,20,21). These differing structures, which must affect their orientation and flexibility, provide an explanation for the experimental absence of cross-stimulation between primases and helicases from these bacterial classes. Here we show that cross-stimulation between primase and helicase was restricted to closely related members of the same bacterial class, namely E. coli and Y. pestis, as well as B. anthracis and G. stearothermophilus. These results support a close ancestral relationship between these bacteria, despite their adaptation to diverse environments. This is particularly evident from the ability of the mesophilic B. anthracis helicase to stimulate thermophilic G. stearothermophilus primase and vice versa.

A comparison of the primase HBD and helicase interface from G. stearothermophilus with a B. anthracis model provided insight into residues that may be essential for the productive cross-species interaction (Figure 6C–E). Although a number of residues are conserved in the HBD–helicase interface for members of the same or both bacterial classes, several amino acids are unique to B. anthracis and G. stearothermophilus, as well as E. coli and Y. pestis. Studies investigating the contribution of these species-specific and the conserved amino acids within this critical interface should illuminate their impact on this essential protein–protein interaction.

E. coli helicase stimulated E. coli DnaG to synthesize shorter rather than full-length primers, unlike the other productive bacterial helicase–primase interactions. A recent study suggested that differences in electrostatic potential surface charge on the N-terminal domain of E. coli helicase compared to another gram-negative bacteria may influence the binding stability to DnaG (40). This comparatively low negative charge distribution on the E. coli helicase N-terminal domain may contribute to the observed functional differences. However, in vivo complementation of the E. coli dnaG RPD mutant by Y. pestis DnaG and in vitro cross-stimulation between the primases and helicases from E. coli and Y. pestis demonstrates a degree of plasticity.

The productive interaction between DnaG and the replicative helicase hexamer has emerged as a critical element in the coordination of leading and lagging strand DNA synthesis (41). Our analyses of the interactions between primase and ssDNA and primase and helicase have revealed a series of properties that influence either class- or species-specific function. These key functional similarities and differences are controlled by the modular components of primase and have important evolutionary implications for the regulation of DNA replication. Since both primase and helicase are essential for cell propagation and survival, only those mutations that are either neutral or enhance functional activity will persist. The conserved mechanism by which primase recognizes specific sequence motifs and initiates de novo primer synthesis indicates that few mutations are tolerated in the primer-synthesizing domains. Additional studies on the molecular mechanisms that regulate primase activity and helicase–primase interactions will improve our understanding of the evolution of species-specific functions.


Supplementary Data are available at NAR Online.


The Defense Advanced Research Program Agency (Award W911NF0510275 to S.H.H.); Biotechnology and Biological Sciences Research Council (Award E004717/1 to P.S.). Funding for open access charge: The Department of Pathology and Microbiology, University of Nebraska Medical Center.

Conflict of interest statement. None declared.

Supplementary Material

Supplementary Data:


We thank Amanda Bartling and Matthew Shortridge for assistance with data collection, and Oluwatoyin Asojo for technical help with PyMOL software.


1. Kornberg A, Baker T. DNA Replication. 2nd edn. Freeman: San Francisco; 1992.
2. Okazaki R, Okazaki T, Sakabe K, Sugimoto K, Sugino A. Mechanism of DNA chain growth: I. Possible discontinuity and unusual secondary structure of newly synthesized chains. Proc. Natl Acad. Sci. USA. 1968;59:598–605. [PubMed]
3. Wu CA, Zechner EL, Reems JA, McHenry CS, Marians KJ. Coordinated leading- and lagging-strand synthesis at the Escherichia coli DNA replication fork: V. Primase action regulates the cycle of Okazaki fragment synthesis. J. Biol. Chem. 1992;267:4074–4083. [PubMed]
4. Pan H, Wigley DB. Structure of the zinc-binding domain of Bacillus stearothermophilus DNA primase. Structure. 2000;8:231–239. [PubMed]
5. Griep MA, Lokey ER. The role of zinc and the reactivity of cysteines in Escherichia coli primase. Biochemistry. 1996;35:8260–8267. [PubMed]
6. Godson GN, Schoenich J, Sun W, Mustaev AA. Identification of the magnesium ion binding site in the catalytic center of Escherichia coli primase by iron cleavage. Biochemistry. 2000;39:332–339. [PubMed]
7. Bird LE, Pan H, Soultanas P, Wigley DB. Mapping protein–protein interactions within a stable complex of DNA primase and DnaB helicase from Bacillus stearothermophilus. Biochemistry. 2000;39:171–182. [PMC free article] [PubMed]
8. Chang P, Marians KJ. Identification of a region of Escherichia coli DnaB required for functional interaction with DnaG at the replication fork. J. Biol. Chem. 2000;275:26187–26195. [PubMed]
9. Bailey S, Eliason WK, Steitz TA. Structure of hexameric DnaB helicase and its complex with a domain of DnaG primase. Science. 2007;318:459–463. [PubMed]
10. Corn JE, Pease PJ, Hura GL, Berger JM. Crosstalk between primase subunits can act to regulate primer synthesis in trans. Mol. Cell. 2005;20:391–401. [PubMed]
11. Frick DN, Richardson CC. DNA primases. Ann. Rev. Biochem. 2001;70:39–80. [PubMed]
12. Koepsell SA, Larson MA, Griep MA, Hinrichs SH. Staphylococcus aureus helicase but not Escherichia coli helicase stimulates S. aureus primase activity and maintains initiation specificity. J. Bacteriol. 2006;188:4673–4680. [PMC free article] [PubMed]
13. Thirlway J, Soultanas P. In the Bacillus stearothermophilus DnaB–DnaG complex, the activities of the two proteins are modulated by distinct but overlapping networks of residues. J. Bacteriol. 2006;188:1534–1539. [PMC free article] [PubMed]
14. Yoda K, Okazaki T. Specificity of recognition sequence for Escherichia coli primase. Mol. Gen. Genet. 1991;227:1–8. [PubMed]
15. Swart JR, Griep MA. Primase from Escherichia coli primes single-stranded templates in the absence of single-stranded DNA-binding protein or other auxiliary proteins. Template sequence requirements based on the bacteriophage G4 complementary strand origin and Okazaki fragment initiation sites. J. Biol. Chem. 1993;268:12970–12976. [PubMed]
16. Larson MA, Bressani R, Sayood K, Corn JE, Berger JM, Griep MA, Hinrichs SH. Hyperthermophilic Aquifex aeolicus initiates primer synthesis on a limited set of trinucleotides comprised of cytosines and guanines. Nucleic Acids Res. 2008;36:5260–5269. [PMC free article] [PubMed]
17. Soultanas P. The bacterial helicase–primase interaction: a common structural/functional module. Structure. 2005;13:839–844. [PMC free article] [PubMed]
18. Mitkova AV, Khopde SM, Biswas SB. Mechanism and stoichiometry of interaction of DnaG primase with DnaB helicase of Escherichia coli in RNA primer synthesis. J. Biol. Chem. 2003;278:52253–52261. [PubMed]
19. Tougu K, Peng H, Marians KJ. Identification of a domain of Escherichia coli primase required for functional interaction with the DnaB helicase at the replication fork. J. Biol. Chem. 1994;269:4675–4682. [PubMed]
20. Syson K, Thirlway J, Hounslow AM, Soultanas P, Waltho JP. Solution structure of the helicase-interaction domain of the primase DnaG: a model for helicase activation. Structure. 2005;13:609–616. [PMC free article] [PubMed]
21. Oakley AJ, Loscha KV, Schaeffer PM, Liepinsh E, Pintacuda G, Wilce MC, Otting G, Dixon NE. Crystal and solution structures of the helicase-binding domain of Escherichia coli primase. J. Biol. Chem. 2005;280:11495–11504. [PubMed]
22. Biswas EE, Biswas SB. Mechanism of DnaB helicase of Escherichia coli: structural domains involved in ATP hydrolysis, DNA binding, and oligomerization. Biochemistry. 1999;38:10919–10928. [PubMed]
23. Nakayama N, Arai N, Kaziro Y, Arai K. Structural and functional studies of the dnaB protein using limited proteolysis. Characterization of domains for DNA-dependent ATP hydrolysis and for protein association in the primosome. J. Biol. Chem. 1984;259:88–96. [PubMed]
24. Biswas SB, Wydra E, Biswas EE. Mechanisms of DNA binding and Regulation of Bacillus anthracis DNA primase. Biochemistry. 2009;48:7373–7382. [PMC free article] [PubMed]
25. Bird LE, Wigley DB. The Bacillus stearothermophilus replicative helicase: cloning, overexpression and activity. Biochim. Biophys. Acta. 1999;1444:424–428. [PubMed]
26. Pan H, Bird LE, Wigley DB. Cloning, expression, and purification of Bacillus stearothermophilus DNA primase and crystallization of the zinc-binding domain. Biochim. Biophys. Acta. 1999;1444:429–433. [PubMed]
27. Swart JR, Griep MA. Primer synthesis kinetics by Escherichia coli primase on single-stranded DNA templates. Biochemistry. 1995;34:16097–16106. [PubMed]
28. Koepsell S, Bastola D, Hinrichs SH, Griep MA. Thermally denaturing high-performance liquid chromatography analysis of primase activity. Anal. Biochem. 2004;332:330–336. [PubMed]
29. Kitani T, Yoda K, Ogawa T, Okazaki T. Evidence that discontinuous DNA replication in Escherichia coli is primed by approximately 10 to 12 residues of RNA starting with a purine. J. Mol. Biol. 1985;184:45–52. [PubMed]
30. Koepsell SA, Larson MA, Frey CA, Hinrichs SH, Griep MA. Staphylococcus aureus primase has higher initiation specificity, interacts with single-stranded DNA stronger, but is less stimulated by its helicase than Escherichia coli primase. Mol. Microbiol. 2008;68:1570–1582. [PubMed]
31. Frick DN, Baradaran K, Richardson CC. An N-terminal fragment of the gene 4 helicase/primase of bacteriophage T7 retains primase activity in the absence of helicase activity. Proc. Natl Acad. Sci. USA. 1998;95:7957–7962. [PubMed]
32. Skopalik J, Anzenbacher P, Otyepka M. Flexibility of human cytochromes P450: molecular dynamics reveals differences between CYPs 3A4, 2C9, and 2A6, which correlate with their substrate preferences. J. Phys. Chem. B. 2008;112:8165–8173. [PubMed]
33. Thorpe IF, Brooks CL., 3rd Molecular evolution of affinity and flexibility in the immune system. Proc. Natl Acad. Sci. USA. 2007;104:8821–8826. [PubMed]
34. Tokuriki N, Tawfik DS. Protein dynamism and evolvability. Science. 2009;324:203–207. [PubMed]
35. Johnson SK, Bhattacharyya S, Griep MA. DnaB helicase stimulates primer synthesis activity on short oligonucleotide templates. Biochemistry. 2000;39:736–744. [PubMed]
36. Grompe M, Versalovic J, Koeuth T, Lupski JR. Mutations in the Escherichia coli dnaG gene suggest coupling between DNA replication and chromosome partitioning. J. Bacteriol. 1991;173:1268–1278. [PMC free article] [PubMed]
37. Kusakabe T, Hine AV, Hyberts SG, Richardson CC. The Cys4 zinc finger of bacteriophage T7 primase in sequence-specific single-stranded DNA recognition. Proc. Natl Acad. Sci. USA. 1999;96:4295–4300. [PubMed]
38. Yang J, Xi J, Zhuang Z, Benkovic SJ. The oligomeric T4 primase is the functional form during replication. J. Biol. Chem. 2005;280:25416–25423. [PubMed]
39. Lee SJ, Richardson CC. Interaction of adjacent primase domains within the hexameric gene 4 helicase-primase of bacteriophage T7. Proc. Natl Acad. Sci. USA. 2002;99:12703–12708. [PubMed]
40. Kashav T, Nitharwal R, Abdulrehman SA, Gabdoulkhakov A, Saenger W, Dhar SK, Gourinath S. Three-dimensional structure of N-terminal domain of DnaB helicase and helicase-primase interactions in Helicobacter pylori. PloS ONE. 2009;4:e7515. [PMC free article] [PubMed]
41. Tanner NA, Hamdan SM, Jergic S, Loscha KV, Schaeffer PM, Dixon NE, van Oijen AM. Single-molecule studies of fork dynamics in Escherichia coli DNA replication. Nat. Struct. Mol. Biol. 2008;15:170–176. [PMC free article] [PubMed]

Articles from Nucleic Acids Research are provided here courtesy of Oxford University Press