In the Simian Virus 40 In Vitro Replication System, Start Site Selection by the Polymerase α-Primase Complex Is Not Significantly Altered by Changes in the Concentration of Ribonucleotides

Abstract

The simian virus 40 (SV40) in vitro replication system was previously used to demonstrate that the human polymerase (Pol) α-primase complex preferentially initiates DNA synthesis at pyrimidine-rich trinucleotide sequences. However, it has been reported that under certain conditions, nucleoside triphosphate (NTP) concentrations play a critical role in determining where eukaryotic primase initiates synthesis. Therefore, we have examined whether increased NTP concentrations alter the template locations at which SV40 replication is initiated. Our studies demonstrate that elevated ribonucleotide concentrations do not significantly alter which template sequences serve as initiation sites. Of considerable interest, the sequences that serve as initiation sites in the SV40 system are similar to those that serve as initiation sites for prokaryotic primases. It is also demonstrated that regardless of the concentration of ribonucleotides present in the reactions, DNA synthesis initiated outside of the core origin. These studies provide additional evidence that the Pol α-primase complex can initiate DNA synthesis only after a considerable amount of single-stranded DNA is generated.

Studies employing the simian virus 40 (SV40) in vitro replication system have been used to establish much of what is known about the enzymology of eukaryotic DNA replication (10, 40, 81). This system has also been used to study the reaction mechanisms that take place during particular stages in the replication process; for instance, during the initiation of DNA replication. SV40 replication is initiated when T antigen (T-ag), the single viral protein necessary for replication, site specifically binds to the viral origin (reviewed in references 8, 11, and 31). Upon binding, T-ag assembles into a double hexamer (21, 23, 51, 61, 77) that is able to function as a helicase (22, 34, 66, 68, 84). Owing to its helicase activity, T-ag is capable of catalyzing origin-specific unwinding, provided replication protein A (RPA) and topoisomerase I are present in the reaction (15, 22, 27, 88).

Recent insights into initiation of DNA replication at the SV40 origin include images of T-ag double hexamers assembled on the viral core origin (77). Additional studies have helped to define the minimal core origin sequences necessary for T-ag double hexamer formation (41, 42, 67). Further insights into T-ag's interactions with the central region of the core origin, site II, were provided by the solution structure of the T-ag origin binding domain (T-ag-obd_131–260) (49). These studies established that a pair of loops is utilized to make sequence-specific contacts with site II (49); a similar surface is used by the DNA-binding domain of papillomavirus to bind to DNA (29). Recent experiments have also helped to unravel how T-ag's interactions with the core origin are controlled by the cell cycle machinery (5; reference 83 and references therein).

The SV40 replication system has also been used to investigate the formation of nascent DNA in the vicinity of the SV40 origin following T-ag catalyzed DNA unwinding (reviewed in references 10, 11, 79). Unwinding of the SV40 origin enables the four subunit Polymerase (Pol) α-primase complex (81) to gain access to the newly formed single-stranded DNA and to initiate DNA synthesis (reviewed in reference 11). It was previously demonstrated that the Pol α-primase complex initiates DNA replication by synthesizing a small RNA/DNA hybrid that has been termed primer-RNA/DNA (16, 56–58). Primer-RNA/DNA is formed exclusively from lagging-strand DNA templates, initially in the vicinity of the SV40 origin and, at later stages of synthesis, at progressively distal locations (13, 25, 50). Once synthesized, primer-RNA/DNA molecules are extended by a second PCNA-dependent polymerase to form full-length Okazaki fragments (16, 56–58, 78).

To provide additional insights into nascent DNA formation by the Pol α-primase complex, primer-extension techniques were used to establish the template locations at which primer-RNA/DNA is formed (13, 14, 17). It was concluded that a significant feature of the initiation signals for the Pol α-primase complex is the trinucleotide 3′-NTT-5′ and, to a lesser extent, the dinucleotide 3′-NT-5′ (where N encodes the nucleotide encoding the 5′ end of the primer) (17). However, whether these sequences will serve as recognition sites for primase under conditions other than those present in the standard in vitro SV40 replication assays has yet to be examined. For example, the ribonucleotide concentrations used in standard SV40 replication assays are 200 μM GTP, UTP, and CTP and 4 mM ATP (48, 70, 86); however, it has been proposed that the actual ribonucleotide concentrations present in cells are higher than the standard ribonucleotide concentrations (43). Moreover, using the purified Pol α-primase complex isolated from calf thymus and the elevated nucleoside triphosphate (NTP) levels suggested to reflect in vivo concentrations (e.g., 1 mM CTP, 1 mM UTP, 2 mM GTP, and 4 mM ATP), it was reported that primer synthesis occurs at the first potential site to which primase binds (43). These observations raise the question of whether the previously described 3′-NTT-5′ and 3′-NT-5′ initiation sites will be selected in SV40 replication reactions conducted in the presence of the alleged in vivo concentrations of ribonucleotides. To address this issue, the sites utilized by the Pol α-primase complex during primer-RNA/DNA synthesis in HeLa cell crude extracts have been analyzed in the presence of the elevated ribonucleotide concentrations.

It was also previously reported that the human Pol α-primase complex does not, in general, initiate replication within the SV40 core origin (13, 14, 16). However, as with individual start site selection, it is not clear if this observation will be reproducible under a range of reaction conditions. Therefore, we have repeated the primer-extension experiments over the core origin using primer-RNA/DNA formed in the presence of the elevated ribonucleotide concentrations. The results from these studies provide additional evidence that the Pol α-primase complex does not initiate DNA synthesis within the SV40 core origin.

MATERIALS AND METHODS

Commercial enzymes and reagents.

Restriction endonuclease SphI was from Gibco-BRL Life Technologies. HindIII, RNasin and calf intestinal alkaline phosphatase (CIP) were from Promega. Sequenase version 2.0 was from U.S. Biochemicals. Phage lambda cut with HindIII was from New England Biolabs.

Preparation of SV40 T-ag, HeLa cell crude extracts, and DNA stocks.

SV40 T-ag was produced using a baculovirus expression vector containing the T-ag-encoding SV40 A gene (59) and isolated by immunoaffinity techniques using the PAb 419 monoclonal antibody as previously described (26, 65, 86). Purified T-ag was dialyzed against T-ag storage buffer (20 mM Tris-HCl [pH 8.0], 50 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride, 0.2 μg of leupeptin per ml, 0.2 μg of antipain per ml, and 10% glycerol) and frozen at −70°C until use. HeLa cell crude extracts were prepared according to the method of Wobbe et al. (86) and dialyzed overnight against storage buffer (20 mM HEPES [pH 7.5], 0.1 mM EDTA, 50 mM NaCl, 1 mM dithiothreitiol, 0.1 mM phenylmethylsulfonyl fluoride, and 10% glycerol) prior to freezing at −70°C.

Plasmid pSV01ΔEP, a 2,796-bp plasmid having the SV40 core origin containing an EcoRII fragment cloned into the EcoRI site of pBR322 (86) was isolated using standard techniques (60) and stored in TE buffer (10 mM Tris [pH 8.0], 1 mM EDTA). To provide single-stranded templates for primer extensions and sequencing reactions, pSV01ΔEP was cleaved with PstI and cloned into M13mp19 (17). A clone containing single-stranded DNA complementary to SV40 early mRNA was isolated and termed M13SV01ΔEP-Pst#2. Single-stranded DNA from clone M13SV01ΔEP-Pst#2 was obtained using routine methods (60).

Pulse- and continuous-labeling reactions.

Pulse-labeling reactions (120 μl), were performed as described previously (12, 15, 17, 25). Reaction mixtures contained 2 μg of T-ag, 0.5 mM dithiothreitol, 40 mM creatine phosphate (di-Tris salt [pH7.7]); 2.8 μg of creatine phosphokinase, 1.5 μg of SV40 origin-containing pSV01ΔEP, 60 μl of HeLa cell extract (∼17 mg/ml), and various amounts of MgCl₂ (see below). The concentration of ribonucleotides used in one set of pulse reactions were those found in the standard SV40 in vitro replication reactions (200 μM each for CTP, UTP, and GTP and 4 mM ATP) (48, 70, 86). The second set of pulse reactions contained the high ribonucleotide levels suggested to reflect the in vivo concentrations (1 mM CTP, 1 mM UTP, 2 mM GTP, and 4 mM ATP) (43). The elevated ATP concentration (4 mM) is required to support T-ag assembly and subsequent unwinding events (9, 21, 24). For the reactions containing the high ribonucleotide levels, the MgCl₂ concentrations were increased so that they were 5 mM in excess of the total NTP concentration (i.e., 13 mM). Moreover, the MgCl₂ concentration in the standard reactions was increased from 7 to 9.6 mM; as a result, the MgCl₂ concentration in the standard reactions was also 5 mM in excess of the total NTP concentration (i.e., 4.6 mM). All pulse mixes contained the same concentration of dNTPs: (dATP, dGTP, and dTTP [final concentration, 100 μM each]) and [α-³²P]dCTP (final concentration, 3 μM [∼250 cpm/fmol]). Pulse reactions were terminated after 10 s by adding 12 μl of stop mixture (EDTA, N-lauroylsarcosine [pH 7.7], and proteinase K; final concentrations of 14 mM, 0.25 mg/ml, and 0.45 mg/ml, respectively). Aliquots (10 μl) were withdrawn to monitor incorporated label via trichloroacetic acid precipitations. Finally, the continuous-labeling reactions (13) are identical to the pulse reactions described above except that the nucleotides in the pulse mix are added at the same time as T-ag.

Primer extension reactions.

To isolate primer-RNA/DNA, nascent DNA formed during 10-s pulse reactions or the continuous-labeling reactions were pooled (in both instances, 9 samples were used in the “standard-ribonucleotide” reactions and 17 were used in the “high-ribonucleotide” reactions). The samples (∼1,000,000 cpm; ∼12 pmol) were ethanol precipitated, washed with 80% ethanol, and dried. The pellets were resuspended in 10 μl of TE and 20 μl of formamide loading buffer (60), boiled for 4 min, and loaded on 10% polyacrylamide gels containing 8 M urea. Primer-RNA/DNA was isolated (∼10,000 cpm; ∼0.125 pmol) as described elsewhere (12, 25).

RNA primers were removed from one primer-RNA/DNA aliquot with alkali; phosphate residues were removed from a second aliquot with CIP (17). Treated primer-RNA/DNA molecules were hybridized to single-stranded M13SV01ΔEP-Pst#2 DNA. The primer extension reactions were performed exactly as described elsewhere (17). Restriction endonuclease digestions were conducted for 2 h in 40-μl reactions with 20 U of enzyme according to the manufacturer's recommendations; 40 U (1 μl) of rRNasin (Promega) was added to each reaction to inhibit RNase activity.

Gel electrophoresis and PhosphorImager analysis.

Sequencing reactions (62), used as size markers for the analysis of the primer extension reactions, were conducted employing a kit purchased from U.S. Biochemicals (dITP labeling mix). Sequencing reactions were primed with a 21-nucleotide (nt) oligonucleotide (5′-TCATGAGCGGATACATATTTG-3′), purchased from Oligos, Etc., Inc., hybridized to M13SV01ΔEP-Pst#2 (17). Primer extension reactions (∼1,000 cpm/lane) and sequencing reactions (∼25,000 cpm/lane of [³⁵S]dATP) were loaded onto 8% polyacrylamide gels containing 8 M urea; the gels were electrophoresed and processed as described earlier (17). Quantitation of the primer extension products was performed on a Molecular Dynamics PhosphorImager.

RESULTS

There is presently some controversy regarding the ribonucleotide concentrations present in eukaryotic cells. Studies by Hauschka (37) indicate that the actual physiological concentration of ribonucleotides in eukaryotic cells is considerably higher than the concentration of ribonucleotides currently used in the SV40 in vitro system (see Materials and Methods). In contrast, recent characterization of the nucleotide concentrations present in eukaryotic cells (74) indicates that the ribonucleotide concentrations used in SV40 in vitro replication reactions are actually within the physiological range. In view of the uncertainty associated with these studies and the observations made by Kirk et al. (43), we elected to investigate the consequence of increasing the ribonucleotide concentration on various aspects of the initiation of nascent DNA synthesis in the SV40 in vitro replication system.

Establishing whether increased ribonucleotide concentrations influence primer-RNA/DNA formation and the positions at which primer-RNA/DNA is synthesized.

In an initial set of reactions, we determined whether elevated ribonucleotide concentrations alter, either quantitatively or quantitatively, the formation of primer-RNA/DNA (Fig. 1). Primer-RNA/DNA formed in the presence of high ribonucleotides (lane 3) had essentially the same overall size distribution as that formed under standard ribonucleotide concentrations (lane 2). However, closer inspection of lane 3 demonstrates that the upper (∼34 nt) primer-RNA/DNA species is reduced in the presence of elevated ribonucleotide concentrations, a result observed in several additional experiments (data not shown). To quantitate these experiments, the primer-RNA/DNA species formed in this and two identical experiments was removed and counted in a scintillation counter. These analyses revealed that the amount of primer-RNA/DNA formed in the presence of high ribonucleotide concentrations was approximately one-third of that formed under standard ribonucleotide concentrations (for quantitation, see the legend to Fig. 1). The molecular basis for the drop in primer-RNA/DNA synthesis at elevated ribonucleotide concentrations is not understood. Nevertheless, these studies demonstrate that SV40 replication reactions, conducted in vitro in the presence of elevated ribonucleotide concentrations, have a slightly modified form of primer-RNA/DNA that is synthesized at reduced levels relative to primer-RNA/DNA formed in the presence of standard ribonucleotide concentrations.

FIG. 1.

Primer-RNA/DNA formed during a 10-s pulse in the presence of standard and high ribonucleotide concentrations. Following pulse-labeling (see Materials and Methods), the samples were ethanol precipitated, and the pellets were resuspended in 15 μl of TE buffer. The samples were then boiled for 4 min in a 90% formamide loading buffer (60) and applied to a 10% denaturing polyacrylamide gel. Lanes 2 and 3 show the results of pulse reactions conducted at standard and high ribonucleotide concentrations, respectively. The location of primer-RNA/DNA, containing two distributions centered around ∼24 and ∼34 nt, is indicated. Size markers (sizes shown in nucleotides) used during gel electrophoresis (lane 1) were from an MspI digest of pBR322, labeled with kinase by standard methods (60). Finally, upon isolation of primer-RNA/DNA, the average number of counts per minute in three separate 120 μl “standard-ribonucleotide” reactions was 1,829; in the “high-ribonucleotide” pulse reactions, the corresponding number was 599. Thus, relative to the standard-ribonucleotide reactions there was an approximately threefold decrease in primer-RNA/DNA synthesis in the high-ribonucleotide pulse reactions.

Previous studies indicated that the first nucleotide to bind primase helps to stabilize the Pol α-primase complex at a given start site (64; reviewed in reference 2). It follows that an increase in the concentration of a particular ribonucleotide might stabilize the Pol α-primase complex at different template sequences and thereby alter the initiation profile. Indeed, it has been reported that at elevated ribonucleotide concentrations, the Pol α-primase complex purified from calf thymus initiates synthesis in a sequence-independent manner (43). Therefore, we next addressed whether elevated ribonucleotide concentrations had similar effects on start site selection in the SV40 replication system.

Primer-RNA/DNA was formed in pulse reactions conducted for 10 s (see Materials and Methods). To isolate primer-RNA/DNA formed under standard ribonucleotide concentrations, nascent DNA from nine pulse reactions was pooled. However, since primer-RNA/DNA formation is reduced when reactions are conducted under high ribonucleotide concentrations (Fig. 1), it was necessary to pool the nascent DNA from 17 pulse reactions. The pooled samples were then loaded onto a 10% denaturing polyacrylamide gel that was subjected to electrophoresis (data not shown), and primer-RNA/DNA was purified as described earlier (12, 25). Primer extension techniques, depicted in Fig. 2, were then used to map the template locations used to synthesize primer-RNA/DNA molecules (12, 14, 17).

FIG. 2.

Map of the SV40 origin region and depiction of the primer extension reactions used to map primer-RNA/DNA start sites. Single-stranded M13 DNA is depicted by the thick lines at the ends of the figure, while DNA derived from pSV01ΔEP is symbolized by the thinner line. Primer-DNA is symbolized by the small rectangle at the end of the dashed line; primer-RNA is depicted by the circle. Rectangles not associated with a circle represent primer-DNA molecules formed as a result of alkali treatment. Primer extension products, resulting from Sequenase 2-catalyzed elongation of primer-RNA/DNA molecules hybridized to M13SV01ΔEP-Pst#2, are depicted by the dashed arrows. The primer extension products were cleaved at the indicated restriction endonuclease sites. Nucleotide positions in plasmid pSV01ΔEP are indicated; numbers in parentheses correspond to the SV40 numbering system (73).

Results of primer extension reactions, conducted with primer-RNA/DNA formed under standard ribonucleotide concentrations, are presented in Fig. 3 (lanes 1 to 4). As previously demonstrated (14, 17), these experiments permit mapping of initiation sites on the “late side” of the SV40 origin. As a control for possible Sequenase pause sites, aliquots of the nonrestriction endonuclease cleaved primer-extension products are displayed in Fig. 3 (lanes 1 and 2). The primer extension reaction displayed in lane 1 was performed with primer-RNA/DNA that had been treated with CIP; those in lane 2 were performed with primer-RNA/DNA that had been pretreated with alkali. Additional aliquots of the primer extension reactions were treated with SphI (lanes 3 and 4). It is apparent that the previously described set of start sites, L4 to L9 (13, 14, 17), were utilized under this set of reaction conditions. Similar primer-extension experiments were conducted with primer-RNA/DNA molecules formed under high ribonucleotide concentrations (Fig. 3, lanes 5 to 8). As a control, aliquots of the non-restriction-endonuclease-cleaved primer extension products are displayed in lanes 7 and 8. The primer extension reaction displayed in lane 7 was performed with primer-RNA/DNA that had been treated with CIP; those in lane 8 were performed with primer-RNA/DNA that had been pretreated with alkali. Additional aliquots of the primer extension reactions were treated with SphI (lanes 5 and 6). A comparison of lanes 3 and 4 with lanes 5 and 6 demonstrates that the band patterns are very similar. Thus, in pulse reactions, the previously described start sites L4 to L9 are used under both concentrations of ribonucleotides. It is also apparent from Fig. 3, lanes 4 and 6, that alkali treatment of primer-RNA/DNA formed under either standard or high ribonucleotide concentrations reduced the size of the primer extension products to the same extent. Thus, molecules formed under both sets of conditions have RNA primers of similar length.

FIG. 3.

Example of primer extension reactions used to map initiation sites on the late side of the SV40 origin using both standard in vitro ribonucleotide concentrations (lanes 1, 2, 3, and 4) and high ribonucleotide concentrations (lanes 5, 6, 7, and 8). Aliquots of the primer extension products (∼1000 cpm), formed by using M13SV01ΔEP-Pst#2 and primer RNA-DNA (either CIP treated [lanes 1 and 7] or alkali treated [lanes 2 and 8]) were loaded onto an 8% denaturing polyacrylamide gel. Additional aliquots of the primer extension products were cleaved with SphI at pSV01ΔEP position 2518 (lanes 3, 4, 5, and 6). Prior to conducting the primer extension reactions, primer-RNA/DNA was either CIP treated (lanes 3 and 5) or alkali treated (lanes 4 and 6). Individual primer-RNA and primer-DNA start sites (L stands for late) are indicated; as previously reported (14, 17), the start sites are numbered sequentially from the first strong site, L1. The sequencing ladders (see Materials and Methods) used as size markers are indicated by the letters A, C, G, and T.

The data in Fig. 3 were quantitated using a PhosphorImager, a plot of lane 3 is presented in Fig. 4A, and a plot of lane 5 is presented in Fig. 4B. These studies confirm what is suggested by visual inspection of Fig. 3 (lanes 3 and 5), namely, that the same start sites are used to form primer-RNA/DNA under both standard and elevated ribonucleotide concentrations. Collectively, these studies indicate that over a range of ribonucleotide concentrations, the Pol α-primase complex initiates DNA synthesis at the same distinct positions.

FIG. 4.

Results from PhosphorImager analyses of the primer extension reactions. (A) Results of a PhosphorImager scan, measured in PhosphorImager counts, of Fig. 3 (lane 3). Individual primer-RNA start sites, selected during DNA synthesis in the presence of standard ribonucleotide concentrations, are indicated. (B) Results of a PhosphorImager scan, measured in PhosphorImager counts, of Fig. 3 (lane 5). Individual primer-RNA start sites, selected in the presence of high ribonucleotide concentrations, are indicated.

Establishing whether increased ribonucleotide concentrations affect initiation site selection in the vicinity of the SV40 origin.

We previously reported that primer-RNA/DNA start sites are suppressed over the SV40 core origin (13). In contrast, in vivo-based mapping studies indicated that start sites for DNA synthesis are situated within the core origin and are used to initiate leading-strand DNA synthesis (38). Moreover, recent studies of the yeast ARS1 and human lamin B2 origins of replication indicate that DNA synthesis initiates within eukaryotic origins of replication (1, 7). Therefore, to further characterize initiation events, we elected to determine if our previous in vitro mapping studies could be duplicated in the presence of elevated levels of ribonucleotides. Since it is agreed that initiation events do not take place in the core origin on the strand encoding late mRNA (13, 38), these studies were limited to an analysis of initiation events on the template encoding early mRNA, the lower strand in Fig. 9.

FIG. 9.

Map of the nucleotide positions at which primer-RNA/DNA was initiated in the vicinity of the SV40 origin in the presence of standard and high ribonucleotide concentrations. The locations of part of one of the SV40 enhancers (Δ enhancer), the 21-bp repeats, the core origin and the T-ag binding site I are indicated. Nucleotide positions in plasmid PSV01ΔEP are also indicated; where present, numbers in parentheses are those used to number SV40 (73). Sequences encoding the 5′ termini of primer-RNA are indicated by boldface capital letters, and those encoding the 5′ termini of primer-DNA are indicated by boldface lowercase letters. Primer-RNA molecules are symbolized by arrows; the dotted arrows represent the relatively weak initiation events taking place over the 21-bp repeats and core origin in the pulse reactions (Fig. 5 and 6) (13). To aid in establishing the distribution of pyrimidine residues on the template for early mRNA synthesis, the pyrimidine residues are reprinted on the separate, lower, lines. Also indicated is DD2, the oligonucleotide used to synthesize the sequencing ladders used as size markers. Minor differences with previous mapping studies include the La DNA start site being at 2642 rather than 2641 (Fig. 5) and the failure to detect the previously described (13) faint start site Lb (Fig. 6 and 8). In addition, compared to previous studies (13, 17), the L1 initiation site was utilized at relatively low levels (Fig. 6 and 8). Finally, in the presence of the elevated ribonucleotide concentrations, start sites Lx1 and Lx2 were detected (Fig. 7 and 8).

Figure 5 presents the results of primer extension experiments conducted with primer-RNA/DNA, formed using standard and elevated ribonucleotide concentrations, and M13SV01ΔEP-Pst#2 (complementary to SV40 early mRNA). Upon cleavage with HindIII, these experiments enabled mapping of initiation sites from the pSV01ΔEP HindIII site at position 2718 through the SV40 core origin, 21-bp repeats, and enhancer sequences. As previously reported (13), start sites for primer-RNA/DNA formed under standard ribonucleotide concentrations were, in general, absent over the core origin (lane 3). Start sites for primer-RNA/DNA were detected over the 21-bp repeats (Lc to Li). However, stronger start sites were more apparent beyond this region. As in previous studies, alkali treatment (lane 4) reduced the size of the primer extension products by approximately 9 nt. The experiments in lanes 5 and 6 permitted mapping of start sites for primer-RNA/DNA formed in the presence of elevated ribonucleotide concentrations. It is apparent that the start sites for primer-RNA/DNA formed in the presence of elevated ribonucleotide concentrations (lanes 5 and 6) are very similar to those formed in the presence of standard ribonucleotide concentrations (lanes 3 and 4). In both instances, there is very little indication of initiation events occurring within the core origin. As controls for “sequenase 2” pause sites, aliquots of the non-restriction-endonuclease-cleaved primer extension products, formed under standard and high ribonucleotide concentrations, are displayed in lanes 1 and 2 and lanes 7 and 8, respectively.

FIG. 5.

Representative primer extension reactions used to map initiation sites in the vicinity of the core origin, using primer-RNA/DNA formed under standard (lanes 1 to 4) and high (lanes 5 to 8) ribonucleotide concentrations. As controls for Sequenase pause sites, aliquots of the primer extension products, formed by using M13SV01ΔEP-Pst#2 and primer RNA-DNA (either CIP treated [lanes 1 and 7] or alkali treated [lanes 2 and 8]) were loaded onto an 8% denaturing polyacrylamide gel. Additional aliquots of the primer extension products were cleaved with HindIII at pSV01ΔEP position 2718 (lanes 3, 4, 5, and 6). Prior to conducting the primer extension reactions, primer-RNA/DNA was either CIP treated (lanes 3 and 5) or alkali treated (lanes 4 and 6). Individual primer-RNA and primer-DNA start sites (L stands for late) are indicated to the right of the figure, while the map to the left of the figure indicates the pSV01ΔEP regions covered by these primer extension reactions. The sequencing ladders used as size markers (see Materials and Methods) are indicated by the letters A, C, G, and T.

PhosphorImager-based quantitation of Fig. 5, lanes 3 and 5, are presented in Fig. 6A and B. Inspection of these plots confirms that, on the template for early mRNA synthesis, little or no primer-RNA/DNA is formed within the core origin. Furthermore, regardless of the ribonucleotide concentrations in the pulse mix, the start sites over the 21-bp repeats are relatively weak; stronger start sites are detected at template locations situated beyond the 21-bp repeats.

FIG. 6.

PhosphorImager analyses of primer extension reactions presented in Fig. 5. (A) Results of a PhosphorImager scan, measured in PhosphorImager counts, of the standard-ribonucleotide primer extension reactions presented in Fig. 5 (lane 3). (B) Results of a PhosphorImager scan of the high-ribonucleotide primer extension reactions presented in Fig. 5 (lane 5). Individual primer-RNA start sites, measured in PhosphorImager counts, are indicated. Maps of the regions in the vicinity of the core origin covered by the primer extension reactions are presented below the scans.

Repeating the primer extension reactions with primer-RNA/DNA formed in continuous-labeling reactions.

As previously noted, the first nucleotide to bind to primase helps to stabilize the primase-template complex (64). The second nucleotide to bind primase becomes the nucleotide present at the 5′ end of the primer, while the initially bound nucleotide becomes the second nucleotide in the RNA primer. These binding events are followed by the formation of a full-length RNA primer (20, 45, 64). Based on these studies, and results from our previous primer-RNA/DNA mapping experiments, we proposed that in the SV40 replication system an ATP molecule initially binds to primase opposite the 3′ proximal thymidylate of an 3′-NTT-5′ or, less frequently, 3′-NT-5′ sequence. We also proposed that the second nucleotide binds primase opposite the apparently random nucleotide at position N (17). In light of these studies and the fact that our pulse reactions are incubated for 15 min in the presence of 4 mM ATP, one could argue that primase molecules are prebound to the template prior to the introduction of the remaining ribonucleotides in the pulse mix. To remove this possibility, we repeated the origin proximal mapping studies using primer-RNA/DNA molecules formed in continuous-labeling reactions (see Materials and Methods). In these experiments, the pulse mix is added to the reactions at the same time as the T-ag. Therefore, it cannot be argued that, owing to the binding of an ATP molecule, primase is selectively stabilized at a given initiation site.

Results from representative primer extension reactions conducted with primer-RNA/DNA formed during continuous-labeling reactions, in the presence of standard and high ribonucleotide concentrations, are presented in Fig. 7. The primer extension products (see Materials and Methods) were cleaved at the HindIII site at position 2718, thereby enabling mapping of initiation sites through the SV40 core origin, 21-bp repeats, and enhancer sequences. It is apparent from lanes 1 and 2 that start sites for primer-RNA/DNA formed in continuous-labeling reactions, in the presence of standard ribonucleotide concentrations are, in general, absent over the core origin. Inspection of lane 1 also reveals that strong start sites for primer-RNA, relative to those in Fig. 5, are present over the 21-bp repeats (Lc to Li). As in previous studies, alkali treatment (lane 2) reduced the size of the primer extension products by approximately 9 nt. The experiments in lanes 3 and 4 permit mapping of start sites for primer-RNA/DNA formed in continuous-labeling reactions in the presence of the high ribonucleotide concentrations. Comparison of lanes 3 and 4 with lanes 1 and 2 demonstrates that basically the same origin proximal start sites are utilized in the presence of standard and high concentrations of ribonucleotides.

FIG. 7.

Representative primer extension reactions used to map initiation sites in the vicinity of the core origin, using primer-RNA/DNA molecules formed in continuous-labeling reactions. Reactions performed with primer-RNA/DNA molecules formed using standard ribonucleotide concentrations are presented in lanes 1 and 2; those formed under high ribonucleotide concentrations are presented in lanes 3 and 4. Aliquots of the primer extension products were cleaved with HindIII at pSV01ΔEP position 2718 (lanes 1 to 4). Prior to conducting the primer extension reactions, primer-RNA/DNA was either CIP treated (lanes 1 and 3) or alkali treated (lanes 2 and 4). As in Fig. 5, individual primer-RNA and primer-DNA start sites (L stands for late) are indicated to the right of the figure, while the map to the left of the figure depicts the pSV01ΔEP regions covered by these primer extension reactions. As in previous examples, the sequencing ladders used as size markers (see Materials and Methods) are indicated by the letters A, C, G, and T.

The data in Fig. 7 were quantitated using a PhosphorImager; a plot of lane 1 is presented in Fig. 8A, and a plot of lane 3 is presented in Fig. 8B. In general, these studies confirm what is suggested by visual inspection of Fig. 7, namely, that in continuous-labeling reactions the same origin proximal start sites are used to form primer-RNA/DNA under both standard and high ribonucleotide concentrations. However, certain minor bands (e.g., L×1 and L×2) were detected in the reactions containing the high ribonucleotide concentrations, but they were not detected in the standard reactions. Thus, the presence of elevated ribonucleotides may slightly extend the repertoire of template sequences used by the Pol α-primase complex as initiation sites. However, it is apparent that even in the presence of elevated ribonucleotides, the previously described set of initiation sites is preferentially used.

FIG. 8.

PhosphorImager analyses of the primer extension reactions presented in Fig. 7. (A) Results of a PhosphorImager scan of the standard-ribonucleotide primer extension reactions presented in Fig. 7 (lane 1). (B) Results of a PhosphorImager scan of the high-ribonucleotide primer extension reactions presented in Fig. 7 (lane 3). Individual primer-RNA start sites, utilized in the continuous-labeling reactions, are labeled. Maps of the regions in the vicinity of the core origin covered by the primer extension reactions are presented below the scans; the y axis presents the PhosphorImager counts of the bands in a given lane.

The primer extension products shown in Fig. 3, 5, and 7 were run next to sequencing ladders that served as size markers; this enabled the identification of those template sequences that serve as initiation sites for primer-RNA/DNA synthesis (Fig. 9). With the exception of L×1 and L×2, the average distance between primer-RNA/DNA start sites over the 21-bp repeats, in either the pulse or continuous-labeling experiments, was ∼10 nt; a similar phasing of primer-RNA/DNA start sites is not observed at the more-distal template locations. It was also previously noted that the start sites over the 21-bp repeats do not, as a rule, contain the 3′-NTT-5′ sequence. Rather, many of the origin proximal start sites (Lc-Li) contain purines at position 1 and a thymidylate at position 2 (3′-PuT-5′). A similar sequence, 3′-NT-5′, is used to support relatively weak initiation events at more distal template locations (17). Finally, comparison of Fig. 5 and 6 with Fig. 7 and 8 demonstrates that the origin proximal start sites are preferentially utilized in the continuous-labeling reactions. We suspect that this result may be related to the extent of unwinding that takes place in the two reactions. For instance, it has been reported that the priming sites nearest a given replication fork are preferentially used (36). Walter and Newport discuss additional evidence that the DNA Pol α-primase complex specifically binds to forks formed during DNA unwinding (80). Therefore, a slightly greater degree of unwinding in the incubation requiring pulse reactions would shift the initiation events to more distal sites. In contrast, relatively limited unwinding in the continuous-labeling reactions would result in a greater utilization of origin proximal sites.

DISCUSSION

Previous studies have demonstrated that eukaryotic primase preferentially initiates within pyrimidine-rich sequences (3, 32, 35, 40, 44, 71, 85, 89). In order for primase to initiate DNA synthesis, a stable complex must form between primase and the pyrimidine-rich initiation sites. Recent experiments indicate that stable binding of two NTPs to primase is a prerequisite for stable complex formation between primase and single-stranded DNA templates (43, 64). Related studies demonstrated that fluctuations in ribonucleotide pools can influence exactly which template sequences are selected as initiation sites (43).

To further address how the Pol α-primase complex initiates DNA synthesis in eukaryotic cells, we have been characterizing the formation of primer-RNA/DNA. We previously reported that template sequences containing 3′-NTT-5′ and, to a lesser extent 3′-NT-5′, sequences serve as preferred initiation sites for synthesis of primer-RNA/DNA (13, 14, 17). The present studies demonstrate that the same 3′-NTT-5′ and 3′-NT-5′ start sites are used during 10-s pulse reactions, even in the presence of elevated levels of ribonucleotides. That this result reflects the preloading of the Pol α-primase complex at selected start sites prior to pulse-labeling is unlikely. This conclusion is based on equilibrium thermodynamic arguments that suggest that a significant percentage of the Pol α-primase complexes will be unbound just prior to pulse-labeling. Therefore, upon introduction of the pulse mixes, distinct initiation sites should be present if the ribonucleotide concentrations were significantly influencing start site selection. Furthermore, we have demonstrated that basically the same origin proximal start sites are utilized in continuous-labeling reactions in the presence of standard and elevated ribonucleotide concentrations. Thus, in the SV40 replication system, the start site specificity of the Pol α-primase complex is not lost at elevated ribonucleotide concentrations; however, consistent with studies by Kirk et al. (43), it may be extended to include additional template sequences (e.g., 3′-NC-5′).

It is of interest to consider why our SV40-based assays have detected a greater sequence specificity for primase-catalyzed initiation events than was detected in previous studies. One advantage of this system is that we are characterizing primer-RNA/DNA, molecules whose RNA moieties are intact and can be directly analyzed. Similar studies cannot be undertaken with full-length Okazaki fragments given the extensive processing of the 5′ ends during maturation (reviewed in reference 4). Moreover, unlike assays conducted with the purified Pol α-primase complex, SV40 replication reactions are performed in HeLa cell crude extracts that contain additional factors [e.g., topoisomerases, RPA, etc.] that may influence start site selection. Consistent with this proposal, it was previously reported that in the presence of the single-stranded DNA-binding protein from Escherichia coli, the specificity of priming by the calf thymus Pol α-primase complex was increased (35). Furthermore, the templates for lagging-strand DNA synthesis flanking the SV40 origin are rich in thymidylate residues and have a relatively low abundance of deoxycytidylate residues (Fig. 9). Thus, preferential utilization of the 3′-NTT-5′ and 3′-NT-5′ initiation signals may reflect, in part, the sequence composition of the DNA in the vicinity of the SV40 origin of replication. While these issues need further examination, our studies provide additional evidence that the pyrimidine rich 3′-NTT-5′ and 3′-NT-5′ sequences are preferentially used by the Pol α-primase complex during initiation of SV40 DNA synthesis.

That a pyrimidine-rich trinucleotide serves as the preferred recognition signal for the Pol α-primase complex in the SV40 in vitro replication system is of considerable interest when one considers the initiation signals utilized by prokaryotic primases. The recognition signals for T7 gene 4 protein (54), the T4 gene 61 protein (19, 39), and the E. coli DnaG (72) are 3′-CTG-5′, 3′-T(C/T)G-5′, and 3′-GTC-5′, respectively. Moreover, mutant forms of DnaG have been described that recognize the more general sequence 3′-PuPyPy-5′ (90). Thus, in both size and sequence composition, there is considerable overlap between prokaryotic initiation sites and those used to initiate SV40 replication. These observations provide additional evidence that basic features of replication have been conserved between prokaryotic and eukaryotic organisms (69).

In previous primer extension studies, we also investigated where the Pol α-primase complex initiates DNA synthesis in the vicinity of the core origin (13). Based on these studies, it was concluded that initiation events are suppressed over the core origin. Relatively strong start sites for initiation of DNA replication were detected at a considerable distance (i.e., ∼70 nt) beyond the core origin. Indeed, we previously suggested that extensive DNA unwinding is a prerequisite for the generation of adequate single-stranded DNA for use as a substrate by the Pol α-primase complex (14). However, as with studies designed to map the individual start sites, it is possible that experimental conditions, such as changes in ribonucleotide concentrations, may influence where DNA synthesis events initiate relative to the core origin. Nevertheless, the present pulse-and continuous-labeling studies demonstrate that even in the presence of the elevated ribonucleotide concentrations, DNA synthesis is, in general, initiated outside of the SV40 core origin.

Recent studies of the replication origin of E. coli, oriC, demonstrated that two hexamers of Dna B must unwind 65 nt or more before the primase can function (30). Based on these studies, it was also concluded that helicase activity is required to produce a considerable amount of single-stranded DNA prior to forming a primer by the E. coli primase, DnaG. In light of the results obtained in the SV40 and E. coli systems, it will be interesting to determine if helicases in higher eukaryotic organisms (e.g., the MCM complex [46, 91]) must also unwind a considerable amount of DNA before the Pol α-primase complex initiates primer synthesis. However, in vivo-based studies support the opposite conclusion namely, that DNA synthesis takes place within the SV40 (38), yeast ARS1 (7), and CHO dihydrofolate reductase (18) origins of replication. Related studies of the yeast ARS1 and the human lamin B2 origins indicate that the oppositely moving leading strands initiate at the same positions on the two complementary helices (1, 6). Thus, additional experiments are needed to determine whether a general property of origins of replication is that they undergo extensive unwinding prior to serving as templates for initiation by the Pol α-primase complex.

The studies described in the preceding sections provide additional insight into the initiation of SV40 DNA replication. Upon forming a double hexamer on the SV40 core origin, T-ag unwinds the origin and recruits additional factors necessary for initiation (e.g., Topo I [33] and HSSB [RPA] [53; reviewed in reference 11]) and the Pol α-primase complex (28, 52, 63, 75; reference 82 and references therein). Once a 3′-NT-5′ or 3′-NTT-5′ initiation signal is encountered by the Pol α-primase complex, synthesis of a primer-RNA/DNA molecule is initiated. Previous studies indicate that RPA(HSSB) (87), particularly the 32-kDa subunit, is also required for the synthesis of primer-RNA/DNA (47, 50) and for coordinating a Pol switch with Pol δ (76, 92). Based on observations made in prokaryotic (30) and human systems (92), it is proposed that the synthesis of the origin proximal primer-RNA/DNA molecules is a prerequisite for the recruitment of RFC, PCNA, and Pol δ and the subsequent initiation of leading-strand synthesis. Okazaki fragment formation is initiated by synthesis of primer-RNA/DNA at more distal NTT sites (reviewed in reference 11). Subsequent steps in the initiation process, such as the processing of the RNA primers, have been reviewed (4, 11, 79). The extent to which observations based on the SV40 model system apply to initiation events at other eukaryotic replication origins remains to be determined. However, one additional reason for continued interest in RNA primers is that their synthesis has been recently shown to be the activator of the checkpoint that prevents entrance into mitosis until the S phase is completed (55).