Proc. Not. Ad. Sei. USA Vol. 69, No. 9, pp. 2691-2695, September 1972 Initiation of DNA Synthesis: Requires RNA Synthesis Resistant to Rifampicin* Synthesis of 4x174 Replicative Form (dd gene product/dnaB gene product/Ml3 DNA/spermidine) RANDY SCHEKMAN, WILLIAM WICKNER, OLE WESTERGAARD, DOUGLAS BRUTLAG, KLAUS GEIDER, LEROY L. BERTSCE, AND ARTHUR KORNBERG Department of Biochemistry, Stanford University School of Medicine. Stanford, California 94305 Contti6uled by Arthur Kornberg, July 10, 1979 ABSTRACT Conversion of single-stranded DNA of phage 9x174 to the double-stranded replicative form in Escherichia coli uses enzymes essential for initiation and replication of the host chromosome. These enzymes can now be puri6ed by the assay that this phage system provides. The 9x174 conversion is distinct from that of M13. The reaction requires Werent host enzymes and is resistant to rifampicin and streptolydigin, inhibitors of RNA polymerase. However, RNA synthesis is essential for 9x174 DNA synthesis: the reaction is inhibited by low concentrations of actinomycin D, all four ribonucleoside triphosphates are required, and an average of one phospho- diester bond links DNA to RNA in the isolated double- stranded circles. Thus, we preeume that, as in the case of M13, synthesis of a short RNA chain primes the synthesis of a replicative form by DNA polymerase. Initiation of DNA synthesis by RNA priming is a mechanism of wide significance. Conversion of the viral single strand of M13 DNA to its double-stranded replicative form is inhibited by rifampicin (1). Studies with enzyme fractions have shown that RNA synthesis by RNA polymerase is required in this reaction (2). Our findings indicate that the RNA synthesis provides a primer to initiate DNA synthesis (1,2). How general is this RNA priming mechanism for DNA- strand initiation? Inhibition of other DNA synthetic events by rifampicin suggests the involvement of RNA priming in multiplication of the M13 replicative forms (1) and in syn- thesis of a colicinogenic factor (3). However, replication of +X174, a virus of base composition similar to that of M13, is not inhibited by rifampicin in vim (4) or in vitro (2). Does this absence of rifampicin inhibition imply a different mechanism of DNA-strand initiation or is there an RNA synthetic system that, unlike RNA polymerase, is resistant to rifampicin? We have found that Escherichia coli has distinctive enzymes for replication of +X and M13 DNA (2). We now report that the initiation of a +X DNA strand requires RNA synthesis, presumably in a primer role; the absence of rifampicin in- hibition points to an RNA synthetic system that behaves differently from RNA polymerase. A remarkable feature of the Abbreviations: +X, 4x174; SS, (phage) singlegtranded circular DNA; RF,'(phage) double-stranded DNA of circular replicative form; RF 11, (phage) RF with a discontinuity in at least one strand. * This is the third paper in a series on Initiation of DNA Syn- thesis; papers I and I1 in this series are refs. 1 and 2. enzymes for +X replication is that they include two or more of the gene products (dnaA and B) that are known to be re- quired for initiation and replication of the E. coli chromosome. The dna A and B proteins have been partially purified. MATERIALS AND METHODS Materials were from sources described (1). Streptolydigh (Upjohn Company, Lot no. 10799) was a gift from Dr. Walter Mangel. Actinomycin D was obtained from Merck. [a-atP]- deoxyribonucleoside triphosphates were synthesized by a modification of the procedure of Symons (5). Bacterial Strains. E. coli H560 (F+, thy-, endo I-, pol AI-) was provided by Dr. H. Hoffmann-Berling, H560 dnaB by Dr. F. Bonhoeffer, and CRT 4638 dnaA (F-, thy-, endo I-, pol AI-) by Dr. R. Wickner (strain originally from Dr. Y. Hirota). Temperature-resistant revertants were selected by spreading about l@ cells on tryptone-supplemenkd agar plates; clones that formed overnight at 42' were selected. Growth of Cella and Preparation of Extract. Cells were grown and collected as before (2), except that they were not chilled before harvest. Instead cells were suspended at room tem- perature (22-24") and quickly frozen in liquid nitrogen. Extracts and supernatant fractions were prepared aa before and used directly without addition of MgClt. Extracts from temperature-sensitive strains were prepared in the same way except that the 1-min warming step was between 25 and 30' rather than at 37'. The supernatant fraction (Fraction I) was prepared in polycarbonate tubes in the type 40 or Ti 60 rotor of the Beckman L265 B centrifuged at 100,000 to 200,000 X g for 30-60 min. Fraction I has about 20 mg of protein per ml. Assay of DNA Replication. The my measured incorpora- tion of a labeled nucleotide into an acid-insoluble fraction as before (2). The assay mixture (25 pl) contained 2.5 mM MgC12, 20 pM each of the four deoxyribonucleoside triphos- phates, 3 mM spermidine.HC1, 6% sucrose, 30 mM Tns. HCl (pH 7.5), 60 mM NaCl, 700 pM ATP, 100 pM each of CTP, GTP, and UTP, and 20 pM +X174 single-stranded DNA (nucleotide residues). The enzyme fraction (1-15 pl) was added last. [a-azP]deoxyribonucleoside triphosphates had a specific activity of 1 Ci/mmol (about 2 X 10' cpm/pmol). One unit of activity is defined as 1 pmol of deoxyribonucleo- tide incorporated per min at 30'. Fraction I has a specific activity of 5-10 units/mg protein. 2691 2692 Biochemistry: Schekman et ai. Proc. Nat. Acad. Sci. USA 69 (1972) TABLE 1. Polyamines stimulate conver8ion of +X SS to RF Polyamine concentration 1.7 mM 5.1 mM Deoxyribonucleotide incorporated (pmol) Spermidine 2.7 8.5 Cadaverine 0.6 1.0 Spermine 3.2 3.2 Putrescine 0.8 1.4 Diaminopropane 1.0 1.5 None 0.5 Standard assays (30 min at 30") were performed with 10 pl of Fraction I enzyme. RESULTS Soluble enzymes convert gX SS to RF Olivera and Bonhoeffer (6) showed that an E. coli lysate supported on a cellophane disc converted +X SS to RF:We developed a high-speed supernatant enzyme fraction capable of performing the conversion of an MI3 template to its RF (2). This crude fraction also proved capable, under compa- rable conditions (see below), of sustaining the +X SS to RF reaction (2). As with MI3 (2), the +X product sediments in a neutral sucrose velocity gradient as RF 11; of the input viral DNA (with 2 pM +X DNA) 80% was in the RF I1 product and the remainder appeared as single strands. The product was analyzed further by subjection of denatured RF I1 to velocity sedimentation in alkaline sucrose gradients or by equilibrium sedimentation in alkaline CsC1. The labeled product appeared as a full-length linear complementary strand, and the template appeared as an intact, circular, viral strand (see ref. 2 for conditions). Spermidine stimulates conversion of gX SS to RF In some experiments the rate of DNA synthesis on a +X single-stranded template was stimulated by spermidine 5- to 20-fold (Table 1). Although the stimulation was not always observed with different Fraction-I preparations, more puri- fied fractions showed an invariable and near absolute re- quirement for spermidine (data not shown). Other poly- amines were less effective. In the presence of spermidine, a rifampicin-sensitive reaction occurs at low concentrations of Fraction-I enzyme (1-5 pl in a 25p1 essay). The product of TABLE 2. Eflect of inhibitors of RNA synlhesis on conwsion of M13 and 4x174 SS to RF in vitro Single-strand template Inhibitor M13 *X yo of control Rifampicin, 5 pg/ml 10 80 Streptolydigin, 600 pg/ml 10 90 Actinomycin D, 5 pg/ml 7 11 Standard assays with 15 pl of Fraction I were for 10 min at 30'; 1bp1 samples were removed and precipitated. Control values in the rifampicin, streptolydigin, and actinomycin experi- ments were, respectively, as follows (in pmol): 25, 29, and 15 for M13; 25,31, and 15 for gX. Minutes Mutant dnaA protein is thermolabile in catalyzing the conversion of 4x174 SS to RF. CRT 4368 dnaA and a temperature-resistant revertant were grown at 30" in H-broth (2) to an optical density (595 nm) of 0.25. Fraction I (10 pl) was assayed except that spermidine waa omitted and 2 pM &X DNA WBS used. Parallel assays without gX DNA yielded values between 0 and 10% of those with gX DNA added; the values plotted are corrected for this background. FIG. 1. such a rifampicin-sensitive reaction is not RF 11, as judged by alkaline velocity sedimentation and by equilibrium sedi- mentation in neutral CsC1. Under such circumstances RNA polymerase may act nonspecifically to provide a primer for DNA synthesis. All assays presented here were done under conditions of rifampicin resistance. Replication proteins from E. coli function in 4X DNA synthesis E. co2i mutants that fail to duplicate their chromosomes at higher temperatures have been grouped in two classes: dnaA and C mutants that continue DNA synthesis at 42' but fail to initiate a new chromosome, and dnaB, D, E, F, and G mutants that stop DNA synthesis immediately at 42'. In extracts (Fraction I) of the thermosensitive dnaA and dnaB mutants there was little or no synthesis of @X DNA at 37", although the rates at 25' were at or near those of extracts of wild-type or temperature-resistant revertants (Figs. 1 and 2). When equal amounts of enzyme fractions from the dnaA and dnaB mutants were combined, +X DNA synthesis was at the wild-type level (Fig. 3). The rate of M13 DNA synthesis was the same in extracts of mutant cells as in those of wild-type cells whether measured at 25' or 37' (data not shown). dnnA protein participates in early phase of the reaction Since the dnaA mutation affects initiation of chromosome replication in E. coli, we tried to see whether the dnaA gene product acts at a stage before DNA synthesis. An aliquot of Fraction I from the dnaA mutant was incubated first with +X SS in the absence of deoxyribonucleoside triphosphates at either 25" or 37' for 8 min. At this point, the DNA precur- sors were added and synthesis was followed in both cases at Proc. Nat. Acad. Sci. USA 69 (1979) Initiation of DNA Synthesis 2693 I 25. b- 4- 2- - c, h - i' 8- *e- f. - B + z.6- 0 tamp-resistant - .- .- r: 2- 2? t 0 - - I 2:: 6 io Minutes FIG. 2. Mutant dnaB protein is thermolabile in catalyzing the conversion of 9x174 SS to RF. H560 dnaB and a temperature- resistant revertant were.grown at 30' in H-broth to an optical density (595 nm) of 0.5. Fraction I was assayed as in Fig. 1. 37". When incubations were performed at 37" throughout, no synthesis occurred; however preincubation at 25" enabled extensive DNA synthesis at 37" (Fig. 4). Thus, the dnaA gene product acts in the initiation phase of the reaction or is stabilized to a subsequent 37" exposure by an early interac- tion. Partial purification of the dnaA and dnaB proteins Complementation at 37" of Fraction I of the dnaB mutant provided a linear assay to determine amounts of the dnaB gene product in extracts of wild-type cells. A similar complementation assay for amounts of the dnaA gene product was feasible by use of Fraction I of the dnaA mutant. Comple- mentation reactions were rifampicin-resistant, and the product was RF 11. By fractionation procedures, to be re- ported in detail elsewhere, the dnaA and dnaB proteins have been partially purified. RNA synthesis is required for conversion of +X SS to RF Rifampicin, which prevents RNA polymerase from initiating synthesis (7), inhibits replication of M13 but not of t$X (2) (Table 2). The same result was found with streptolydigin, which prevents RNA polymerase from propagating a chain (8) (Table 2). From these results, one could conclude that RNA synthesis is not required for conversion of +X SS to RF or that RNA is synthesized by a system distinguishable from RNA polymerase. Involvement of RNA synthesis is indicated by three lines of evidence: (i) inihibition by low amounts of actinomycin D, (ii) requirement for all four ribonucleoside triphosphates, and (iii) covalent linkage of DNA to RNA in the product. 5 9) 0 6- dno A+ dna 8 extracts 9) T1 04- + - - z Q2 - dna B extract ,---- -----e- - - ~ dna A extract 0.- 8 0 10 20 Mlnutes 30 FIG. 3. Thermolabile dnaA and dnaB mutant extracts complement at the nonpermissive temperature. Fraction I from dnaA and dnaB mutant strains were assayed together (5 pl of each) or separately (10 pl of each) at 37' as in Figs. 1 and 2. Actinomycin D inhibits RNA synthesis by virtually all RNA polymerases (bacterial and animal) by intercalating into a duplex DNA template and binding to deoxyguanosine (9). Since t$X synthesis was inhibited by actinomycin D almost completely (as was M13) at concentrations that do not profoundly atTect DNA replication (Table 2), an RNA synthetic event is clearly indicated. Fraction I enzyme had a strong requirement for ATP for maximal t$X DNA synthesis. Although a dependence on added CTP, GTP, and UTP was not apparent in this frac- tion, it was clearly demonstrable with an enzyme fraction precipitated with ammonium sulfate (at 40% saturation). Preincubation ot 259 DNA synthesis at 3;. 0, 'Y -* L I I 1 IO 20 30 Minutes FIG. 4. dnaA protein is required for an early step of the 6x174 SS to RF reaction. Fraction I was prepared from CRT 4638 dnaA and assayed (10 pl enzyme in each assay) for +X template activity. Prior incubation denotes an 8-min incubation of the entire reaction mixture minus the four deoxyribonucleo- side triphosphates; the latter were added at 8 min. 2694 Biochemistry: Schekman et al. Proc. Nai. Acad. Sci. USA 69 (197.9) TABLE 3. Requiremats for ribonucleoside triphosphates in the conversion of gX SS + RF Deoxyribonucleotide incorporated (pmol) Ribonucleoside triphosphate additions -rifampicin +rifampicin 8.5 17.0 9.8 8.8 12.7 14.4 14.4 52.5 9.5 8.1 7.6 7.7 13.7 17.1 11.6 38.8 Fraction I (3 ml) was put onto a 0.6-ml DEAE-cellulose column (DE 52, H. Reeve Angel Inc., Clifton, N.J.) equilibrated with buffer A [50 mM Tris-HCl (pH 7.5)-10% sucrose-0.1 M NaCl]. Ammonium sulfate (0.72 g, 40% of saturation) was added to the DEAE pass-through and left for 30 min at 0'. The precipitate was mllected by centrifugation at 17,000 X g for 15 min. The drained pellet was redissolved in 2 ml of buffer A (with 0.2 mM dithiothreitol), and dialyzed for 5 hr with three changes of 200 ml of buffer A. After 8 hr at O", the fraction was used in standard assays (15 p1 of enzyme, 30-min incubation at 30°), except that only the indicated ribonucleoside triphos- phates were added (100 pM each). Assays were performed with or without rifampicin (10 pg/ml). * The value without gX added was 3.2 pmol. All four ribonucleoside triphosphates were required for max- imal synthesis of gX DNA; 2- to Sfold lower DNA synthesis with +X SS was observed when any one of the ribonucleoside triphosphates was omitted (Table 3). DNA is covalently linked to' RNA in gX and M13 SS conversions to RF A phosphodiester bridge between a deoxyribonucleotide and a ribonucleotide in the isolated RF product was reported earlier in the M13 reaction (2). The experiment entails alka- line hydrolysis of the RF I1 synthesized in the presence of four [a-a*P]deoxyribonucleoside triphosphates and isolation of the ribonucleotides from the alkaline digest. With the +X RF I1 products, as in the case of M13 (repeated in this study), about 1 mol of [a2P]ribonucleoside monophosphate was iso- lated for each mole of RF produced (Table 4). The mixtures of 2' and 3' ribonucleotides to which I2P was transferred con- tained largely AMP in the case of M13, but included sig- nificant amounts of all four ribonucleotides with AMP and GMP predominating in the case of +X. DISCUSSION DNA polymerases extend chains when provided with a DNA- primer strand and template but cannot initiate a chain. The unique capacity of RNA polymerases to initiate chains on DNA templates and the knowledge that DNA polymerase extends a primer, even when terminated mith a ribonu- cleotide, suggested to us that these two polymerases might act cooperatively in vivo for starting a DNA strand. This ex- pectation proved to be correct. In the initiation of a DNA strand on a phage M13 template, E. coli uses RNA poly- merase; the reaction is inhibited by rifampicin both in vivo and in vitro (1, 2). RNA priming of DNA-strand initiation on duplex DNA templates has also been inferred, from rifampicin sensitivity, with M13 RF replication (1) and colicinogenic factor El duplication (3). DNA synthesis that is unaffected by rifampicin implies a mechanism of strand initiations without RNA involvement or else RNA synthesis catalyzed by an enzyme system with properties different from RNA polymerase. The ongoing synthesis of an E. coli chromosome (10) and the replication of gX DNA (4) are such rifampicin-resistant processes. We chose to investigate the replication of gX DNA because we could analyze it with a soluble enzyme system (2). Our studies of th.is system now show that gX strands are initiated by an RNA primer synthesized either by an enzyme distinct from RNA polymerase, or some rifampicin-resistant modified form of the classic enzyme. The evidence is of three kinds: (i) action of specific inhibitors (Table 2), (ii) requirement for all four ribonucleoside triphosphates (Table 3), and (iii) covalent linkage of DNA to RNA (Table 4). The action of inhibitors requires extra comment. Actinomycin inhibits RNA polymerase action (bacterial and animal) by intercalation into a DNA helix and specific binding to deoxyguanosine residues (9). By contrast, rifam- picin inhibits the initiation of RNA chains, and streptolydigin their propagation, by binding the core polymerase directly. TABLE 4. The in vitro product contains RNA covahntly attached to DNA Labeled ribonucleotide alkaline released by M13 RF* gX RF* hydrolysis mol nucleotide/mol RF Exp. 1 Exp. 2 Exp.'l Exp. 2 AP 1.2 1.2 0.40 0.40 GP