INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

Primases are essential enzymes in both prokaryotes and eukaryotes which synthesize primer RNA (pRNA) for initiation of DNA replication by DNA polymerases. The sole known activity of primase is to catalyze the synthesis of short RNA chains de novo on DNA templates (1).

Escherichia coli primase is responsible for priming of DNA replication in E. coli chromosome (2, 3), bacteriophage (4), X174 (5, 6), and G4 (7), and plasmids ColE1 (8) and pBR322 (9). In single-stranded phage G4, E. coli primase synthesizes pRNA from a special region termed as the origin of complementary DNA strand synthesis (oric) and requires an association with single-strand DNA-binding protein (SSB).¹ G4oric is a 140-nucleotide (nt) non-coding DNA sequence, containing three stem-loop structures (stem-loops I, II, and III) (10-12). The sequence for pRNA transcription begins at the unique 5'-CTG-3' trinucleotide (the pRNA initiation site) in the region 3'-flanking of stem-loop I (7, 13, 14). A specific structure of this priming complex has been reported (15). By contrast, in the general priming system primase associates with DnaB helicase and synthesizes pRNA chains from many sites on any single-stranded DNA (ssDNA) (16, 17).

In a study to investigate the effect of complementary oligodeoxynucleotides (referred to as oligonucleotides below) upon pRNA synthesis on G4oric, we have found that, in addition to synthesizing RNA de novo, primase catalyzed the synthesis of RNA chains from the 3'-hydroxyl terminus of oligonucleotides that hybridized to G4oric DNA. This primed RNA synthesis has been verified by size measurement, radioactive labeling, sequence mapping, and modification of the 3'-terminal group of oligonucleotide. The resulting DNA-RNA hybrid polynucleotide has been characterized by deoxyribonuclease I (DNase I) and ribonuclease H (RNase H) cleavages. Efficiency of the primed RNA synthesis on the oligonucleotide-primed template was close to the efficiency of pRNA synthesis de novo on a non-primed template. In the G4oric-specific priming system, the primed synthesis occurred within the pRNA transcribed DNA sequence; in the primase/DnaB/ssDNA general priming system, the primed synthesis did not require a specific DNA sequence and takes place with and without the 5'-CTG-3' sequence. This primed RNA can be used by E. coli DNA polymerase III holoenzyme as an effective primer for the synthesis of DNA chain. A similar priming reaction for the synthesis of RNA at the 3'-hydroxyl termini of DNA had been found in prokaryotic (18, 19) and eukaryotic (20) RNA polymerases. The significance of this finding to the observation of primase rebinding and synthesizing multimeric length RNA found in eukaryote (21, 22) is discussed.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

G4oric ssDNA, Oligonucleotides, and Proteins

The 278-nt G4oric ssDNA fragment (23) was generated by EcoRI cleavage of G4oric DNA from f1R199/G4oric circular ssDNA (f1R199 containing the 278-nt G4oric sequence (24)) with complementary oligonucleotides annealed to the EcoRI cleavage sites. The cleaved G4oric fragment was separated by polyacrylamide gel electrophoresis and eluted by soaking gel slices in 0.5 M NH₄OAc, 0.1% SDS, and 1 mM EDTA for several hours at room temperature. The eluted DNA was desalted with a Centricon-50 column (Amicon). The f1R199/G4oric and M13mp18 circular ssDNA was prepared as described in Ref. 23.

Oligonucleotides were chemically synthesized; some of them were modified at the 3' terminus with a phosphate group (DNAgency). All oligonucleotides were purified by gel electrophoresis, elution in H₂O, and passage through a Centricon-10 column. 5'-³²P-End labeling of oligonucleotides followed the method described in Ref. 25, using [-³²P]ATP (6000 Ci/mmol, NEN Life Science Products) and T4 polynucleotide kinase (U. S. Biochemical Corp.).

Primase was induced from the overexpressing plasmid pGNG1 (26) in E. coli BL21 cells (27). The protein was then purified by 40-50% NH₄(SO₄)₂ selective precipitation and FPLC Mono Q 5/5 column chromatography (Pharmacia Biotech Inc.) (28). SSB was purchased from U. S. Biochemical. dnaB protein was prepared as described in Ref. 29. DNase I (RNase-free) was purchased from Boehringer Mannheim and RNase H from U. S. Biochemical. DNA polymerase III holoenzyme was kindly supplied by Dr. Mike O'Donnell of the Rockefeller University.

Oligonucleotide-primed RNA Synthesis

The Primase/SSB/G4oric System-- Normal pRNA synthesis conditions (30) were used except for adding an annealing step prior to the synthesis reaction to generate a primed template. 0.2 pmol of 278-nt G4oric or f1R199/G4oric ssDNA template was mixed with 2.8 pmol of oligonucleotide (this ratio or higher yielded the maximum efficiency of primed synthesis) in 10 µl of 20 mM Tris-HCl, pH 7.5, the mixture was heated to 85 °C for 2 min then cooled slow to 35 °C. When using 5'-³²P-end labeled oligonucleotide for primed synthesis, the labeled oligonucleotide sample was heated to 67 °C for 15 min to inactivate T4 polynucleotide kinase and was used directly without further purification so that the concentration of oligonucleotide was not changed. The oligonucleotide-primed template was immediately added into a 50-µl synthesis reaction that was made up of pRNA synthesis buffer (20 mM Tris-HCl, pH 7.5, 8 mM dithiothreitol, 8 mM MgCl₂, and 4% sucrose), 4 µg of bovine serum albumin, 4.2 pmol of SSB protein (26 pmol for reaction using f1R199/G4oric), 10 pmol of primase, and different substrates. The complete synthesis reaction mixture contained 100 µM ATP, 20 µM each GTP, UTP, and CTP, and 20 µCi of [-³²P]GTP (3000 Ci/mmol). In the limited synthesis reaction, ddNTP was substituted for the rNTP. When 5'-³²P-end labeled oligonucleotides were used to label primed products, [-³²P]GTP was omitted from the reaction. Synthesis reactions were carried out at 30 °C for 20 min and stopped by adding EDTA to a final concentration of 20 mM. After precipitation with ethanol, the products were analyzed in 20% polyacrylamide (acrylamide:N',N'-methylene bisacrylamide, 19:1), 7 M urea gels.

The Primase/DnaB/ssDNA System-- The basic method used was as described in Ref. 16. Here, conditions were similar with the reaction using G4oric, except that M13mp18 ssDNA (0.2 pmol) was the DNA template and 2 µg of dnaB protein (7 pmol) was added instead of SSB. The efficiency of RNA synthesis was measured by exposure of gels to PhosphorImager^TM screen and quantitation using the ImageQuant^TM program (Molecular Dynamics). The poly(rA) size marker was prepared following the method described in Ref. 31; the ladder was standardized by comparison with the 9-nt pRNA synthesized by primase.

DNA Synthesis on the Primed RNA

The procedure of the RNA and DNA coupled syntheses followed was as described in Ref. 32, except for adding a step of annealing the template to the oligonucleotide before synthesis and separating primase and DNA polymerase reactions into successive processes. To prepare the primed DNA template with a primed RNA, a reaction mixture (20 µl) of 0.2 pmol of G4oric 278-nt ssDNA, 2.8 pmol of 17-nt oligonucleotide, 0.5 mM ATP, 0.1 mM CTP and UTP, 30 µM GTP, 30 µCi of [-³²P]GTP (3000 Ci/mmol), 80 µg/ml bovine serum albumin, 4.2 pmol of SSB, and 10 pmol of primase in pRNA synthesis buffer (see "Primed RNA Synthesis") was incubated at 30 °C for 20 min. DNA synthesis was immediately started by adding 50 µM each dGTP, dCTP, and TTP, 18 µM dATP plus 10 µCi of [-³⁵S]dATP (1,000 Ci/mmol, in the second half-life. Amersham), and 70 ng of pol III* plus 15 ng of -subunit. The DNA synthesis reactions (25 µl) were carried out at 30 °C for 10 min and stopped by adding 40 mM EDTA. Control reactions followed the same processes. DNA synthesis products were precipitated with ethanol once, then separated by electrophoresis in 20% polyacrylamide (acrylamide:N',N'-methylene bisacrylamide, 38:1), 7 M urea gel (20 × 40 cm), under 30 watts for 5 h. Two denaturing polyacrylamide gels were prepared. One for autoradiography was fixed with 12% methanol, 10% acetic acid, dried, then exposed to Kodak XAR-5 film without intensifying screen for 24 h. Another one for quantitation was exposed to a film without drying. The radioactive bands were excised from the gel and their radioactivity were measured in a liquid scintillation counter (Beckman, LS 6500).

Nuclease Digestion

DNase I and RNase H cleavages were performed immediately after RNA synthesis without denaturing treatment. 20 units of DNase I (RNase-free) was added directly to the 50-µl synthesis reaction and digestion was carried on for 40 min at 37 °C. For RNase H cleavage, the volume of synthesis sample was adjusted with RNase H buffer (20 mM Tris-HCl, pH 7.5, 20 mM KCl, 10 mM MgCl₂, 0.1 mM EDTA, and 0.1 mM dithiothreitol) to 100 µl. Samples were digested with 30 units of RNase H at 37 °C for 30 min and the digestion was stopped by adding 50 mM EDTA. The cleaved fragments were precipitated with ethanol once and then analyzed on a denaturing 20% polyacrylamide gel (16 × 15 cm). All control samples were passed through the same procedure.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

An Observation of Primed RNA Synthesis at the 3'-End of an Oligonucleotide Hybridized with G4oric ssDNA-- A 278-nt G4oric ssDNA fragment (23) was used in these experiments. The first oligonucleotide hybridized to G4oric was 20 nt in length and complementary to the 3'-sequence flanking the core stem-loop structures (Fig. 1A). The 3'-flanking region contains the 5'-CTG-3', the pRNA initiation site (7). pRNA synthesized by primase from the G4oric template annealed with the oligonucleotide was examined and compared with pRNA synthesized from unannealed templates including G4oric alone and mixed with an uncomplementary oligonucleotide. As seen in Fig. 1B, the majority of pRNA synthesized de novo from G4oric alone was a cluster of species from 24 to 28 nt in length, plus some smaller pRNA chains (lane 4). Synthesis from the G4oric annealed with the complementary oligonucleotide, however, mainly yielded a group of much larger products (approximately 38-42 nt, lane 5). It seemed unlikely that the large products were nascent pRNA because the maximum size of pRNA synthesized from G4oric by primase in the normal in vitro synthesis reaction we used is 29-nt (7). These products were not synthesized from the oligonucleotide itself (i.e. using the oligonucleotide as a template), as no products were observed when the oligonucleotide was used in the absence of G4oric (lane 7). Comparison of sizes of the large products with the 20-nt oligonucleotide (lane 2) and poly(rA) size marker (lane 3) showed that the large species were 19 to 23 ribonucleotides longer than the oligonucleotide. These sizes were equal to the lengths of RNA chains that were extended from the 3'-end of the oligonucleotide (being hybridized at site +5 on G4oric, see Fig. 1A) until the termination sites of pRNA synthesis (+24 to +28) on G4oric template. This result suggested that the large products were RNA chains extended from the 3'-end of the oligonucleotide by primase. A control 20-nt oligonucleotide (lane 1) of sequence unrelated to G4 sequence was not extended nor affected G4 pRNA synthesis with primase (lane 6).

View larger version (32K): [in this window] [in a new window]   Fig. 1.   pRNA synthesis by primase on G4oric with or without the complementary oligonucleotide. The procedure is described under "Experimental Procedures." A, the partial nucleotide sequence and the secondary structure of G4oric (10, 12) and the complementary oligonucleotides (20 nt). The DNA sequence for pRNA transcript is covered by a line, and the synthesis start sequence is indicated as +1. The ddCTP termination site G at +9 in the limited synthesis reaction is printed in bold. B, products from the complete RNA synthesis reaction with four rNTP and [-32P]GTP substrates were analyzed in a denaturing polyacrylamide gel (20 × 40 cm). Poly(rA) was used as a RNA size marker. The size of some nascent pRNA bands are marked and the larger products are given with two possible sizes (see the text). Lanes 1 and 2, the 5'-32P-end labeled non-G4 related oligonucleotide and the complementary oligonucleotide, respectively. Lane 3, the 5'-32P-end labeled poly(rA). Lanes 4-7 are synthesis products from different templates: lane 4, G4oric alone; lane 5, G4oric annealed with the complementary oligonucleotide; lane 6, G4oric annealed with the unrelated oligonucleotide; and lane 7, the complementary oligonucleotide alone. C, gel electrophoresis (16 × 15 cm) of the limited RNA synthesis reaction with ATP, GTP, [-32P]GTP, UTP, and ddCTP substrates. Templates are indicated above each lane. RNA synthesis products and the sizes are marked.

Verification of Primed Synthesis-- Two other oligonucleotides complementary to the G4oric 3'-sequence were then used to repeat the previous experiments. They were 17 and 23 nt long and had the same 3'-end as the 20-nt oligonucleotide. Under limited synthesis conditions with ddCTP terminator, both oligonucleotides generated a strong, large product which was equal in length to the original 17- or 23-nt oligonucleotide plus 4 ribonucleotides, in addition to small amount of the 9-nt de novo product (Fig. 2A). These results were consistent with the results obtained from the 20-nt oligonucleotide.

View larger version (39K): [in this window] [in a new window]   Fig. 2.   Primed RNA synthesis on the 17-nt or the 23-nt oligonucleotide annealed G4oric. A, the limited synthesis using ddCTP terminator as in Fig. 1C. The 5'-32P-end labeled 17- and 23-nt oligonucleotides are put on the left for comparison and the synthesis products are shown on the right in the same gel (because of much higher radioactivity in 32P-oligonucleotides, the left part received a much shorter exposure than the right part). The size of 9-nt nascent pRNA is marked. B, same limited synthesis using 5'-32P-end labeled oligonucleotides and nonradioactive rNTPs. Equal amounts of 32P-labeled oligonucleotide were used in oligonucleotide controls (left) and synthesis reactions (right). The arrow indicates the position of the 9-nt pRNA. C, mapping of primed synthesis with selective rNTP substrates and ddNTP terminators. Lanes 1 and 6, the 5'-32P-end labeled 17- and 23-nt oligonucleotides, respectively; lanes 2-4, synthesis products using the 17-nt oligonucleotide and lanes 7-9, same synthesis using the 23-nt oligonucleotide. Substrates used in these reactions were: lanes 2 and 7, ddGTP and [-32P]ATP; lanes 3 and 8, GTP plus [-32P]GTP and ddATP; and lanes 4 and 9, GTP plus [-32P]GTP, ATP, and ddCTP.

Requirement of the 3'-Hydroxyl Group-- All oligonucleotides used in the synthesis experiments had a 3'-hydroxyl terminus. If priming of RNA synthesis by primase requires a 3'-OH group in the primer, a modification to the 3'-terminal group should destroy this reaction. As expected, the 17- and 23-nt oligonucleotides modified with a 3'-terminal phosphate were unable to be extended for RNA synthesis (Fig. 3). However, like the reaction with unmodified oligonucleotides, a small amount of 9-nt de novo pRNA appeared in these reactions.

View larger version (46K): [in this window] [in a new window]   Fig. 3.   Primed synthesis of RNA on the modified 17- and 23-nt oligonucleotides with 3'-phosphate termini. All synthesis conditions were as in Fig. 1C.

Nucleases Digestion of the DNA-RNA Hybrid Polynucleotide-- If the primed synthesis product is a DNA-RNA hybrid polynucleotide chain, it should be cleaved either by deoxyribonuclease in its DNA region or by ribonuclease in its RNA region. DNase I that cleaves double-stranded DNA (35) and RNase H that only digests RNA in DNA-RNA duplexes (36) were employed to cleave the primed synthesis product when it remained hybridized with G4oric ssDNA template.

View larger version (80K): [in this window] [in a new window]   Fig. 4.   DNase I and RNase H cleavages of the DNA-RNA hybrid polynucleotide. The primed product was synthesized under ddCTP termination conditions from the 17-nt oligonucleotides annealed R199/G4oric ssDNA template. The procedure is described under "Experimental Procedures." Cleaved samples were separated by denaturing polyacrylamide gel electrophoresis. Components used in each reaction are indicated above each lane. The right part of the image was darkened so that the light bands were more visible.

Efficiency of Primed RNA Synthesis-- The efficiency of extending RNA chains on the oligonucleotide primer by primase appeared to be similar to the efficiency of pRNA synthesis de novo when estimated from the relative intensity of product bands in polyacrylamide gels. For example, in the complete synthesis reaction in Fig. 1B, the amounts of the two products judged from the density of the radioactive bands were similar; under limited synthesis conditions in Figs. 1C and 2A, the amounts of primed product seemed to decrease to about half of the nascent pRNA, which corresponded with 2 [-³²P]GMP molecules that would be incorporated in one extended RNA chain and 4 [-³²P]GMP incorporated into one de novo pRNA molecule (see Fig. 1A). Quantitation by PhosphorImager (Table I) of the two types of RNA yielded from increasing incubation times (1-30 min) showed that the efficiency of primed synthesis was always 80-90% of the efficiency of pRNA synthesis de novo (the 9-nt species). As there was a background of approximately 5-10% G4oric DNA molecules that were unbound by the oligonucleotide in these experiments (which was calculated from 9-nt pRNA synthesized in the priming reaction compared with that in the normal synthesis reaction), the efficiency of primed synthesis and de novo synthesis were very close. Both primed and de novo syntheses reached a plateau of the maximum after a 10-min reaction (Table I).

View this table: [in this window] [in a new window]   Table I Efficiency of primed RNA synthesis and de novo pRNA synthesis RNA was synthesized from templates of G4oric 278 nt alone or annealed by the 17-nt oligonucleotide under the limited conditions with ddCTP terminator and uniformly labeled with [-32P]GTP. RNA synthesis reactions were incubated from 1 to 30 min. The products were separated in a denaturing polyacrylamide gel. The value of each radioactive product was quantitated by a PhosphorImager. As 2 molecules of [-32P]GMP were incorporated into one primed RNA chain and 4 molecules of [-32P]GMP incorporated into one 9-nt nascent pRNA (see Fig. 1A), the values of primed products have been doubled in this table.

Primed Synthesis on G4oric Outside the pRNA Transcribed Sequence-- We have so far shown that the oligonucleotide-primed synthesis of RNA by primase on G4oric DNA took place immediately downstream of the pRNA initiation site. We had found, in fact, that only weak primed RNA synthesis occurred if oligonucleotides were hybridized outside of the DNA sequence for pRNA transcription on G4oric.

View larger version (56K): [in this window] [in a new window]   Fig. 5.   Primed RNA synthesis from the G4oric annealed with the 5' complementary oligonucleotides. A, the nucleotide sequence 5' flanking the core secondary structure in G4oric (10, 12) and the sequence of 5' complementary oligonucleotides. Arrows indicate termination sites of the primed synthesis by adding ddNTP in the experiment described below. B, gel electrophoresis of the primed products from limited conditions. Lane 1, 32P-labeled poly(rA). Lanes 2 and 6, 32P-labeled 20 and 15 nt oligonucleotides, respectively. The selective substrates used in each reaction were: lane 3, UTP, [-32P]UTP and ddGTP; lanes 4 and 5, UTP, [-32P]UTP, GTP, and ddATP; lane 7, GTP, [-32P]GTP plus ddATP; and lanes 8 and 9, GTP, [-32P]GTP, ATP, UTP, and ddCTP. The weak bands of the primed product in gel are marked by small circles, and the 9-nt nascent pRNA is indicated.

Primed Synthesis in the General Priming System-- The oligonucleotide-primed synthesis of RNA by primase was also examined in the primase/DnaB/ssDNA general priming system since this system contains many pRNA initiation sites (16, 17), which is contrary to the G4oric system. Although dnaB protein is a DNA helicase (37), its dominant function of unwinding of double-stranded DNA in the 5' to 3' direction did not impair priming RNA in the 3' to 5' direction by primase on the template-primer in our experiments. The lacZ DNA sequence in the M13mp18 ssDNA circle was used as a DNA template. As it was reported that 5'-CTG-3' is the primase-recognition DNA sequence (38), two complementary oligonucleotides (24-nt each) were utilized to test the priming reaction (Fig. 6A). One oligonucleotide hybridized to the DNA immediately upstream of a 5'-CTG-3' sequence ("CTG" oligonucleotide), another hybridized to a region without any 5'-CTG-3' sequence either inside or surrounding it (non-"CTG" oligonucleotide). Limited synthesis conditions were used so that we could reduce the numerous pRNA chains synthesized de novo from many other DNA regions. From a synthesis reaction containing the lacZ DNA annealed by the CTG oligonucleotide and only CTP plus ddATP, a single species was produced which was 2-nt longer than the oligonucleotide (Fig. 6B, lane 3). When the reaction was supplemented with CTP, ATP, and ddGTP, a second product, with one more additional ribonucleotide (i.e. 3-nt) in length, also appeared (lane 4). None of these products were generated if the oligonucleotide was omitted from reaction (lane 5). According to the DNA sequence, these lengths were equal to measurements predicted for the extension of RNA at the 3'-end of the oligonucleotide under the chain termination conditions used (see Fig. 6A). This result was confirmed by using the same oligonucleotide but modified with a 3'-terminal phosphate group. It resulted in a loss of extended products (see Fig. 6C, lanes 2 and 3).

View larger version (68K): [in this window] [in a new window]   Fig. 6.   Primed RNA synthesis in the primase/DnaB/M13 ssDNA priming system. The synthesis conditions are described under "Experimental Procedures." A, a partial nucleotide sequence of lacZ DNA in M13mp18 (39) and sequences of the two complementary oligonucleotides. Arrows indicate where the primed synthesis should be terminated by incorporation of a appropriate ddNTP substrate. B, gel electrophoresis of products synthesized from the M13mp18 ssDNA hybridized with the oligonucleotides in the presence of a selective ddNTP terminator. The substrates used were: lane 3, CTP, [-32P]CTP, and ddATP; lanes 4 and 5, CTP, [-32P]CTP, ATP, and ddGTP; lane 7, GTP, [-32P]GTP plus ddATP; and lanes 8 and 9, GTP, [-32P]GTP, ATP, and ddCTP. Lanes 2 and 6 were 32P-labeled oligonucleotides separately, and lane 1 is 32P-labeled poly(rA). C, synthesis reaction using oligonucleotides modified with a 3'-phosphate group. Nucleotide sequence of the two modified oligonucleotides are as the same as the oligonucleotides with the 3'-hydroxyl terminus used in panel B. The synthesis condition of lane 3 was the same as that of lane 5 in panel B, and lane 5 the same as that of lane 9 in panel B.

Synthesis of DNA from the Primed RNA Primer by DNA Polymerase III-- To investigate if the primed RNA, like de novo pRNA, can be utilized by DNA polymerase as a primer for DNA chain extension, DNA synthesis with the primed RNA by E. coli DNA polymerase III holoenzyme was assayed. To distinguish the three types of DNA chain that could be synthesized in this reaction (i.e. from de novo pRNA, the primed RNA, and the oligonucleotide), we used ³²P and ³⁵S to differentiate labeling of RNA and DNA regions in products and used the G4oric 278-nt fragment instead of viral circular DNA as a template in order to limit sizes of synthesized DNA within measurable lengths. To reduce competition between DNA polymerase III and primase using the annealed oligonucleotide for extension reaction, we delayed adding polymerase III until primase had fully synthesized RNA.

View larger version (46K): [in this window] [in a new window]   Fig. 7.   DNA synthesized from the primed RNA primer by DNA polymerase III. The procedure is described under "Experimental Procedures." DNA synthesis products were analyzed on a denaturing 20% polyacrylamide gel (20 × 40 cm). A, autoradiography of the film directly exposed to the dried gel. Lane 1, 32P-labeled pBR322-MspI double-stranded DNA fragments, as a size marker; lane 2, complete reaction; lane 3, reaction omitting the 17-nt oligonucleotide; and lane 4, reaction omitting primase. The three types of run-off DNA (see text) were indicated as I, II, and III. B, autoradiography of a second layer of the film that was placed on the top of the film shown in panel A, exposing to the same gel. Lanes 1'-4' represent same samples as lanes 1-4 in panel A. C, counts of run-off DNA products I, II, and III as indicated in panel A. The DNA bands were excised from another identical gel (wet) and the radioactive counts were measured by a liquid scintillation counter. The numbers of counts/min are minus a background and channel overlaps.

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

The data presented here demonstrates that primase can catalyze the synthesis of the RNA chain by covalent attachment of ribonucleotide to the 3'-hydroxyl terminus of an oligonucleotide hybridized to ssDNA. Primase therefore possesses two discrete activities to synthesize short RNA chains on DNA templates: (i) synthesis of RNA de novo and (ii) extension of RNA from the 3'-OH end of a pre-existing primer.

This reaction of synthesis of RNA chain from the 3'-OH terminus of DNA by primase has also been found in RNA polymerases. It was first reported in E. coli RNA polymerase using nicked poly(dAT) template (18, 19) and then found in wheat germ RNA polymerase II on restriction enzyme-cleaved simian virus 40 DNA (20). Later the extension reaction from oligonucleotide primers was used in a study of the conformation of the E. coli RNA polymerase-promoter-product complex (41, 42). As primase belongs to the family of DNA-dependent RNA polymerases, this activity of primed RNA synthesis may be a conserved property in RNA-synthesizing enzymes. Furthermore, considering that all known DNA polymerases synthesize DNA chains only by elongation from the 3'-OH terminus of primers, such primed synthesis is a common feature of the present nucleic acid polymerases.

Primase is a special RNA polymerase because it has very narrow specificity for DNA templates and can utilize either rNTP or dNTP as substrates (33, 34). Such similarity and distinction are also reflected in the primed synthesis activity. Both primase and RNA polymerase synthesize RNA on DNA template either de novo or from a free 3'-OH end of a DNA primer. RNA polymerase can also synthesize RNA on the 3'-OH of a RNA primer (43). This reaction has not been tested in primase in this work; however, it is expected that primase could use a RNA primer as well. Like RNA polymerase, primase displays both de novo and priming synthesis reactions simultaneously. For example, when an oligonucleotide primer was bound immediately upstream or partially covered the G4oric pRNA initiation sequences, primase catalyzed both nascent RNA and primed RNA chains simultaneously. The distinction between primase and RNA polymerase can be seen in the preference for primer. There is no difference in efficiency when primase catalyzes the synthesis of RNA from either 3'-deoxyribonucleotide or 3'-ribonucleotide of an oligonucleotide (data not shown). In contrast, primed synthesis by RNA polymerase from a 3'-ribonucleotide terminus of a primer was 5-10 times higher than synthesis from a 3'-deoxyribonucleotide terminus of a primer (41). This difference may be explained by the capability of primase to use both rNTP and dNTP substrates (33, 34), whereas RNA polymerase can only use rNTP.

We have shown in this report that the specificity of DNA sequence for the primed synthesis by primase is different between the specific G4oric and the general priming systems. This is in agreement with de novo synthesis by primase in the same systems. Such phenomenon may be explained by the structure of these different priming complexes. A specific structure of the primase/SSB/G4oric complex has been recently reported (15). The phasing and binding of SSB to the secondary structure of G4oric exposes the primase recognition sequences to the enzyme. Two primase molecules bind separately to G4oric on both of the 3' and 5' sides of the hairpin structures and synthesize pRNA at the initiation site from the 3'-region (15). Although one primase molecule binds to the 5'-region, there is no detectable nascent pRNA or only a trace of primed RNA synthesized by primase from that DNA region. In the primase/DnaB/ssDNA general priming system, a ternary conformation of DnaB/ssDNA was suggested (17). A physical interaction between primase and dnaB protein has been reported (44). These structures determine how primase interacts with DNA templates in each system for either de novo or the primed syntheses.

It has been known that eukaryotic primases synthesize RNA chains in monomer and multimer lengths. For example, the characteristic lengths are 12-14 nt and multimers for Drosophila primase (45), 8-12 nt and multimer for mouse (46) and yeast (47) primases. Primases mainly synthesize multimeric length pRNA when either isolated from the DNA primase-DNA polymerase complex or in the absence of dNTP substrates. Whereas they synthesize monomeric length pRNA when coupled with DNA polymerase or in the presence of dNTPs (48, 21, 49). For E. coli primase, a similar phenomenon has been found in several priming systems that the lengths of pRNA chains synthesized by primase were shortened to a unit size of 9-14 nt when primase associated with DNA polymerase III holoenzyme (31). Later, a study on yeast primase has revealed that dissociation and rebinding of primase from DNA template resulted in the synthesis of the multimeric length RNA after the synthesis of monomeric length RNA de novo (21, 22). However, this conclusion immediately raises a basic question: whether primases catalyze the extension of RNA chain from a pre-existing RNA on DNA template? To date, primases have been found to synthesize RNA primers de novo. The finding of the primed synthesis in E. coli primase would give a theoretical explanation for the observation of primase rebinding and synthesizing multi-modal length RNA. Although we have not examined the priming function of primase in eukaryotes, it is likely that eukaryotic primases possess this basic activity as the same as E. coli primase.