Purification of a DNA polymerase-DNA primase complex from calf thymus glands.

An immunoabsorbent column, prepared by covalently linking mouse monoclonal anti-calf thymus DNA polymerase-alpha to Protein A-Sepharose, was used as the primary purification step for rapid isolation of DNA polymerase-alpha from calf thymus-gland extracts. In a 4-step procedure consisting of the removal of nucleic acids by protamine sulfate precipitation, chromatography on the immunoabsorbent column, desalting on Sephadex G-50, and removal of bovine immunoglobulins on Protein A-Sepharose, DNA polymerase-alpha activity was purified about 5000-fold from the crude extract with greater than 40% recovery of total enzyme activity. The antibody column-purified DNA polymerase-alpha fraction contains a DNA primase activity that is efficient in replication of single-stranded DNA and poly(dT) when rNTPs are included in the replication reactions. Synthesis by calf thymus DNA polymerase-primase is totally dependent on added template. Complete replication of circular single-stranded phage DNA is achieved with polymerase-primase producing a nicked circular DNA containing oligoribonucleotide primer in the final product. Primers synthesized with single-stranded phage DNA as template were up to 10 nucleotides long when dNTPs were omitted from the reaction and 8 or less nucleotides long when dNTPs were present. Primers synthesized using poly(dT) consisted of three populations when dATP was absent from the reaction, averaging 20 nucleotides, 10 nucleotides, and 3-4 nucleotides. The 20-nucleotide population was not found when dATP was included in the reaction.


From the Department of Biochemistry, Uniformed Services University of the Health Sciences, Bethesda, Maryland 20814
An immunoabsorbent column, prepared by covalently linking mouse monoclonal anti-calf thymus DNA polymerase-a to Protein A-Sepharose, was used as the primary purification step for rapid isolation of DNA polymerase-a from calf thymus-gland extracts. In a 4step procedure consisting of the removal of nucleic acids by protamine sulfate precipitation, chromatography on the immunoabsorbent column, desalting on Sephadex G-50, and removal of bovine immunoglobulins on Protein A-Sepharose, DNA polymerase-a activity was purified about 5000-fold from the crude extract with greater than 40% recovery of total enzyme activity.
The antibody column-purified DNA polymerase-a fraction contains a DNA primase activity that is efficient in replication of single-stranded DNA and poly(dT) when rNTPs are included in the replication reactions. Synthesis by calf thymus DNA polymeraseprimase is totally dependent on added template. Complete replication of circular single-stranded phage DNA is achieved with polymerase-primase producing a nicked circular DNA containing oligoribonucleotide primer in the final product. Primers synthesized with single-stranded phage DNA as template were up to 10 nucleotides long when dNTPs were omitted from the reaction and 8 or less nucleotides long when dNTPs were present. Primers synthesized using poly(dT) consisted of three populations when dATP was absent from the reaction, averaging 20 nucleotides, 10 nucleotides, and 3-4 nucleotides. The 20-nucleotide population was not found when dATP was included in the reaction.
Purification of DNA polymerase-a from eukaryotic sources by conventional procedures requires lengthy processing and is often hindered by problems with proteolysis (1)(2)(3)(4)(5)(6). With the development of monoclonal antibodies to human DNA polymerase-a (7) and immunoaffinity chromatography, rapid purification schemes have been developed for DNA polymerase-a from human KB cells (8) and from calf thymus glands (9).
The immunoaffinity-purified human KB cell DNA polymerase-a was found to contain a tightly bound DNA primase and is capable of replicating single-stranded circular phage DNA and poly(dT) when rNTPs are supplied in the replication reactions (8,10). Purification of the calf thymus DNA polymerase-a, utilizing a monoclonal antibody to human KB *This research was supported in part by Grant CA23365 from National Cancer Institute, Department of Health and Human Services. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. cell enzyme (9) produced a high-molecular-weight enzyme complex, but the purification scheme used resulted in 1% activity recovery.
In the present work, we describe a procedure for rapid purification of DNA polymerase-a from calf thymus glands on a reusable immunoabsorbent column with excellent recovery of enzyme activity. The purified enzyme contains DNA primase activity that is closely coupled to the deoxynucleotide polymerization process. We believe that analysis of several forms of DNA polymerase-primase complex will be necessary to reveal the components of the eukaryotic replication process.

Materials
Unlabeled rNTPs' and dNTPs were purchased from Sigma.  RF I DNA were prepared as described by Kunkel and Loeb (11).
Herpes simplex type I DNA was provided by Dr. William Ruyechan of Uniformed Services University of the Health Sciences. Calf thymus DNA and RNase-free bovine pancreatic DNase I were from Worthington. Poly(dT), chain length of 3000, was synthesized with terminal deoxynucleotidyltransferase (12). Oligodeoxyadenylates were prepared as previously described (13). Protein A-Sepharose was from Pharmacia. Proteins for molecular weight standards were from Bio-Rad. Other chemicals are reagent grade.
Mouse monoclonal antibodies to calf thymus DNA polymerase-ol were prepared by a procedure essentially as described for the panel of mouse monoclonal antibodies to human terminal transferase (14). Four different monoclonal antibodies (all are IgG, subtype) were used for experimental purification of calf thymus DNA polymerase-a.
Mouse monoclonal antibody to Protein A-Sepharose was coupled to Protein A-Sepharose with dimethyl suberimidate by mixing mouse IgG, (purified from hybridoma culture supernatant on a Protein A-Sepharose column) with Protein A-Sepharose in 0.2 M triethylamine HC1 buffer a t pH 8.7 at a ratio of 2 mg of IgGJml of Protein A-Sepharose. After mixing for 1 h at room temperature, the suspension was centrifuged in a clinical centrifuge, and the amount of IgG remaining in the supernatant was determined by Am. Binding was usually greater than 80%. After resuspension in the same solution, solid dimethyl suberimidate was added with mixing to 3 mg/ml. Mixing was allowed to continue for 4 h at room temperature. The gel suspension was then poured into a column and washed with at least 10 volumes each of phosphate-buffered saline (pH 7.4), 0.2 M NaC1, 3.5 M MgC12, 0.2 M NaCl, phosphate-buffered saline containing 0.05% sodium azide, and then stored at 4 "C until use. The efficiency of the coupling reaction was determined by monitoring the Am of the  Crude extract was made from 440 g of frozen calf thymus glands. bAbout 30% of DNA Dolvmerase activitv Dresent in the protamine sulfate supernatant did not bind to the mouse monoclonal antibody (Mab) column. " " . effluent from the column and was greater than 90% for the immunoabsorbent resins used in this study.

Methods
Enzyme Assays-DNA polymerase activity was measured in a 250pl reaction containing 0.04 M potassium phosphate buffer at pH 7.0, 0.2 mg/ml of activated calf thymus DNA, 1 mM dithiothreitol, 8 mM MgCI2, 0.1 mM each of dATP, dCTP, dGTP, and [CH3-3H]dTTP (100 counts/min/pmol), and 100 gg/ml BSA. Incubation was carried out a t 35 "C, and aliquots (30 to 50 gl) were removed a t various times and washed up on glass-fiber filters for acid-insoluble products. One unit of enzyme is defined as the amount that catalyzes the incorporation of 1 nmol of total dNTP into acid-insoluble material in 60 min a t 35 "C.
DNA primase activity was measured in a 250-p1 reaction containing 0.05 M Tris-HCI buffer at pH 8.0, 0.1 mM poly(dT), 1 mM ditniothreitol, 0.2 mM [CY-~'P]ATP at 1000 to 5000 cpm/pmol, 8 mM MgClz, and 100 pg/ml BSA. lncubation was carried out a t 35 "C, and aliquots (30 to 50 p l ) were removed and processed on DE-81 paper (Whatman) and counted under double-isotope conditions in a liquid scintillation counter (15). One unit of DNA primase activity is defined as the amount that catalyzes the incorporation of 1 nmol of ATP into products in 60 min a t 35 "C.
When single-stranded circular phage DNAs were used, typical reactions (200 to 400 p l ) contained 0.04 M potassium phosphate a t pH 7.0, 8 mM MgCI', 1 mM dithiothreitol, 0.1 mM each of dCTP, dGTP, dTTP, and [8-3H]dATP at 300 to 400 cpm/pmol, 0.2 mM each UTP, CTP, GTP, and [n-"P]ATP or [y3*P]ATP at 1000 to 5000 cpm/pmol, 10 pg/ml DNA, and 100 pg/ml BSA. When [(u-~'P]~ATP was used in the reaction the concentration of ATP was 1 mM. Products formed in the reaction were processed on DE-81 paper as described above. Purification of DNA Polymerase-a from Calf Thymus Glands-All operations were carried out a t 4 "C. Frozen calf thymus glands (440 g) were homogenized in a Waring blender in 1760 ml of 0.04 M potassium phosphate at pH 7.4,0.04 M NaCI, 1 mM phenymethylsulfonyl fluoride, and 1% dimethyl sulfoxide (Buffer A). A cloudy supernatant solution was obtained after centrifugation of the homogenate for 15 min a t 4500 rpm in an H4000 rotor in a Sorvall3B centrifuge. After filtering through 4 layers of cheesecloth, 71 ml of 3% protamine sulfate solution were added with stirring. Stirring was continued for 15 min, and then the suspension was allowed to settle for 15 min. The supernatant solution obtained after centrifugation a t 8500 rpm for 30 min in a GS-3 rotor in a Sorvall 5B centrifuge still showed slight turbidity and was clarified by addition of 10% Nonidet P-40 to a final concentration of 0.1%.
The clarified protamine sulfate supernatant (1720 ml) was loaded directly onto a 5-ml mouse monoclonal antibody column containing 10 mg of mouse IgG, to calf thymus DNA polymerase-n a t a flow rate of 4 ml/min. After the loading was complete, the column was washed with 250 ml of Buffer A, 50 ml of 1 M NaCl in 0.05 M Tris-HCI buffer a t pH 8.0, followed by 50 ml of 3 M NaCl in 0.05 M Tris-C1 buffer a t pH 8.0. Protein remaining on the column was eluted with 3.2 M MgC12 buffered with 0.05 M Tris-HCI at pH 8.0. Greater than 95% of the DNA polymerase and DNA primase activity adsorbed were found by direct assay of the 3.2 M MgC1, elute. This fraction was desalted on a Sephadex G-50 column (2.3 X 60 cm) equilibrated with 0.05 M NaCl in 0.05 M Tris-HCI at pH 8.0 and 10% glycerol. Bovine IgG, present in calf thymus extracts and bound to the unsaturated Protein A sites on the Sepharose, was removed by passing the Sephadex G-50 fraction SDS-polyacrylamide gel as described under "Methods." Peptide bands were visualized with Coomassie Blue staining. Protein standards were myosin, phosphorylase b, BSA, ovalbumin, and carbonic anhydrase. through a 5-ml Protein A-Sepharose column equilibrated with 0.05 M NaCl in 0.05 M Tris-HCI at pH 8.0 and 10% glycerol. After the Protein A-Sepharose treatment, the fraction was dialyzed against 25 mM potassium phosphate at pH 7.4, 0.2 mM dithiothreitol, and 50% glycerol. Activities were found to be stable during storage for 2 months a t -20 "C.
Gel Electrophoresis-Peptides present in the purified calf thymus DNA polymerase-a fraction were analyzed on SDS-polyacrylamide gels prepared as described by Laemmli (16) using a 6% stacking gel and a 10% separating gel.
Agarose gel electrophoresis of DNA samples was carried out on 1% gels a t neutral pH as previously described (17). After visualization of DNA bands by staining with ethidium bromide, the gel was dried onto a sheet of DE-81 paper and autoradiographed on Kodak X-Omat AR film a t -70 'C with a DuPont Cronex Lightning Plus intensifying screen. Restriction fragments of DNA samples were analyzed on 6% polyacrylamide gels run in 40 mM Tris acetate buffer at pH 8.0 containing 2 mM EDTA. RNA primers synthesized by the calf thymus enzyme was analyzed on urea-polyacrylamide sequencing gels containing 20% acrylamide, 7.0 M urea, 0.09 M Tris borate at pH 8.3 and 2.5 mM EDTA. Prior to electrophoresis, nucleic acid samples were brought to 49% formamide, 5 mM EDTA, 0.05% bromphenol blue, and 0.05% xylene cyano1 and boiled for 5 min. Electrophoresis was carried out on a sequencing gel apparatus at 3000 V for 2.5 to 3 h.

RESULTS
Purification of Calf Thymus DNA Polymerase-a-A summary of the calf thymus DNA polymerase-a purification procedure using a monoclonal antibody column is presented in Table I. This procedure for preparation of DNA polymerase-a from calf thymus glands outlined under "Experimental Procedures" has several advantages. First, the entire procedure takes 2 days. Second, the enzyme isolated is fully active. Third, recovery of the total enzyme activity is good. Fourth, the capacity of the antibody column is high, since up to 3 mg of fully active mouse monoclonal antibody can be covalently coupled to each ml of Protein A-Sepharose. Finally, the procedure does not destroy the immunoabsorbent resin, and the column can be reused.
In the purification of DNA polymerase-a from calf thymus glands an additional step must be included to remove endogenous IgG in the thymus extracts. This simply requires passing the fraction through a Protein A-Sepharose column. A similar procedure, originally developed for purification of human terminal deoxynucleotidyltransferase from human lymphoblastoid cells, provides rapid preparation of homoge-neous terminal transferase preparations.* The choice of an appropriate monoclonal antibody for enzyme purification is an important factor. We have tested four mouse monoclonal antibodies to calf thymus DNA polymerase-a for enzyme purification. All four antibody columns were prepared with 2 mg of IgG/ml of Protein A-Sepharose, and the capacities for binding enzyme varied more than 3fold. Approximately 15 to 30% of the activity found in the clarified crude extract does not bind to the column. The immunoabsorbent column used in the study was made from mouse monoclonal antibody 42. With this column, about 30% of the activity present in the protamine sulfate supernatant did not bind to the column, and the unbound fraction remained constant for up to 1600 ml of thymus extract. The unbound fraction increased to 40% during the loading of additional 500 ml of thymus extract, and this amount increased to 60% upon further loading of 1000 ml of thymus extract. When the immunoabsorbent resin was made with mouse monoclonal antibody 17, the amount of enzyme activity found in the flowthrough fraction began at 15% and increased to about 30% after 3 liters of thymus extract had passed through the column. On the other hand, greater than 90% of the activity loaded onto the column is recovered in the MgC12 F. J. Bollum and L. M. S. Chang, unpublished results. eluate when the column was made with antibody 42 and only about 35% when antibody 17 was used. SDS-polyacrylamide gel analysis of the peptides as well as protein measurements on the fraction eluted from antibody 17 column showed good recovery of protein, but inactivation of enzyme activity appears to have taken place.
The differences in binding capacity and recovery of enzyme activity from the immunoabsorbent column must be related to the affinity of antibody as well as the specific nature of the antibody-antigen interaction. All four mouse monoclonal antibodies tested in this study are high-affinity antibodies as indicated by the stability of antigen-antibody interaction in 3 M NaCl, yet the total recoveries of enzyme and binding capacities varied. In order to develop an efficient process for protein purification purposes, it may be important to compare a number of different monoclonal antibodies.
Concentrated MgC1, solutions appear to provide effective but mild antigen-antibody dissociation. When similar procedures are used in the purification of low-molecular-weight terminal transferase from immunoabsorbent columns, activity recovery was near 100% when eluted with unbuffered 1 M MgC1, solution. With calf thymus DNA polymerase-a, the recoveries of enzyme activity were 2 and 0% when unbuffered MgC1, solution was used for antibody 42 column and antibody 17 column, respectively. Enzyme activity recoveries were greatly improved when the MgC& solution was buffered with Tris-HC1 at pH 8.0. The concentration of MgC& as well as the pH of the solution to be used to elute protein from the column should be determined for effective operation of this kind of immunoabsorbent columns.
The use of MgCl, for dissociation of the antigen and anti-body complexes on the immunoabsorbent resin made by covalent coupling of antibody to Protein A-Sepharose also allows indefinite reuse of the resin. We have used a mouse monoclonal antibody coupled to a Protein A-Sepharose column for over 50 purifications of calf thymus terminal transferase. The column still retains greater than 50% of the original binding capacity. Table I shows a summary of purification of calf thymus DNA polymerase-a on the antibody 42 column. The overall recovery of enzyme activity exceeds 40%. The clarification of the crude extract appears to be important for successful purification of the enzyme. Crude extracts were carefully titrated with protamine sulfate to obtain maximum removal of nucleic acids and minimum loss of enzyme activity. For a 20% homogenate of calf thymus glands, we found 0.12% protamine sulfate provided the best results. The small amount of turbidity remaining after the protamine sulfate treatment was eliminated by addition of a small amount of nonionic detergent Nonidet P-40. The major loss in enzyme activity occurs during loading of the protamine sulfate supernatant onto the antibody column since about 30% of the activity is found in the flowthrough. Over 90% of enzyme activity bound to the immunoabsorbent resin is recovered in the MgCl, eluate. A small amount of activity, less than 0.5% of DNA polymerase and DNA primase activities, was found in the 1 M NaCl wash. Lane A in Fig. 1 shows the SDS peptides present in the purified calf thymus DNA polymerase fraction.  found in both fractions and the 1 M NaCl fraction is not active, we believe that the unique peptides (185, 160, and 68 kDa) must be responsible for the DNA polymerase and DNA primase activities. Earlier experiments suggest that the 160-kDa peptide is DNA polymerase?
The lower-molecularweight peptides, certainly tightly bound either directly to the column resin or to bound proteins, may represent degraded enzyme peptides or modulating peptides in the enzyme complex.
The DNA polymerase-a fraction isolated from the immunoabsorbent column contains DNA primase activity that can be measured directly by incorporation of [a-32P]ATP in a poly(dT) primed reaction, showing a final specific activity of 747 units/mg in the complex (Table I). The presence of DNA primase in immunoabsorbent-purified DNA polymerase fraction suggests that these enzymes are present as a complex in the crude extract (8,18). Estimation of the molecular weight of the complex would be helpful in understanding peptide stoichiometry.
An unacceptable degree of aggregation of the protein in the purified DNA polymerase-primase fraction was detected by sucrose gradient analysis of the final purified fraction. Addition of NaCl to 0.5 M did not disaggregate these fractions. In L. M. S. Chang and P. Plevani  Tris-HCI a t pH 7.5, 0.5 mM dithiothreitol, 100 pg/ml BSA, and 5 mM MgC12. HaeIII digestion was terminated by addition of SDS and EDTA to 0.5% and 20 mM, respectively. Half of each of the HaeIII digests was separated on a 6% polyacrylamide gel, and the DNA fragments were visualized by ethidium bromide staining (positions of HaeIII fragments are marked on the right of autoradiogram) and autoradiography as described under "Methods." order to obtain some information on the native molecular weight of the enzyme complex, the fraction was first treated with 1.5 M NaCl and then dialyzed into 0.5 M NaCl before analysis on a sucrose gradient made up in 0.5 M NaCl. The enzyme activity profile on the sucrose gradient is presented in Fig. 2. DNA polymerase activity, assayed with activated DNA, sedimented as a double peak a t 7.4 and 9.3 S corresponding to molecular weights of about 158,000 and 223,000, respectively. DNA primase activity, assayed directly by [a-'"PIATP incorporation in poly(dT) replication or 4x174 DNA replication, is found exclusively in the faster sedimenting species. These results suggest that the immunoabsorbent column-purified DNA polymerase-primase fraction consists of free DNA polymerase together with a DNA polymerase-primase complex. Free native DNA polymerase is about 160 kDa, and the polymerase-primase complex appear to be 223 kDa. One possible explanation is that the calf thymus DNA polymerase-primase complex we have detected consists of the 160and the 68-kDa peptides. The lower-molecular-weight peptides do not appear to be specific to the complex, and the 185-kDa peptide is present in trace amount only. We would prefer to postpone speculation about the relation of these peptides to the complex structure.
Properties of Calf Thymus DNA Polymerase-DNA Primase-Synthesis catalyzed by the purified calf thymus enzyme was completely dependent on added template. With circular single-stranded phage DNA as template, maximum synthesis is dependent on the presence of all 4 rNTPs, although omission of pyrimidine ribonucleoside triphosphates decreased synthesis by only 50% (Table 11). ATP is required for the replication of poly(dT). Addition of 100 units/ml of E. coli DNA polymerase I stimulated the reaction rate only about 2-fold. In contrast, poly(dT) replication catalyzed by the yeast DNA polymerase-primase complex was found to be stimulated 7-to %fold by added E. coli DNA polymerase I (18). The results obtained with the calf thymus enzyme suggest that the polymerase and primase activities are more tightly coupled than the yeast enzyme complex.
RNA primers synthesized in the replication reaction can be demonstrated by direct incorporation of ribonucleotides into reaction products. Fig. 3 shows the time courses of poly(dT) and 4x174 DNA replication. When dNTPs were left out of the reaction mixtures the incorporation of [a-"'PIAMP in the poly(dT) reaction was linear until greater than 50% of avail-able template was utilized. We have utilized this property of the calf thymus enzyme to measured primase activity directly (Table I). When dATP was included in the replication reaction, incorporation of [a-"'PIAMP is depressed about 4-fold. The rate of incorporation of AMP in 4x174 DNA replication is less than 1% of that found in poly(dT) replication, and the net incorporation of AMP is also depressed 2-to %fold when dNTPs were present in the reaction. Similar experiments on 6x174 DNA replication using [y-:"P]ATP show that less than 4 molecules of [y-:"P]ATP are incorporated per DNA molecule after complete replication (data not shown).
The calf thymus DNA polymerase-primase has a very high affinity for template DNA. The incorporation curves in Fig.  4 demonstrate that the enzyme is nearly saturated at 1 pg/ml of 4x174 DNA. After 60 min of incubation, replication of 0.16 and 0.8 pg/ml of phage DNA were 75 and 81%, respectively (Fig. 4). At higher concentrations of DNA replication continued until near completion. Template saturation measurements estimate the K , of the calf thymus DNA polymerase-primase for fd DNA to be 0.36 pg/ml (data not shown).
Although the calf enzyme complex replicates heat-denatured calf thymus DNA (data not shown) and single-stranded circular phage DNA efficiently, the enzyme complex carried out only limited synthesis on double-stranded DNAs. When native calf thymus, 4x174 RF I, SV40, and herpes simplex type I DNAs were used as templates, an initial synthesis (amounting to less than 2% replication) was observed, but the reactions did not progress (data not shown). This initial synthesis may represent replication of single-stranded regions in these DNA samples.

Products of Synthesis Catalyzed by the Calf Thymus DNA Polymerase-Primase-Incorporation
experiments demonstrate that replication of 4x174 DNA by calf thymus polymerase-primase proceeds to completion, suggesting that the entire phage DNA molecule might be replicated by the enzyme complex. Fig. 5 shows the time course of appearance of HaeIII fragments in the products of 4x174 DNA replication. When replication was greater than 50% (lanes D and E ) , all of the HaeIII restriction fragments were detected in the product, proving that template molecules are fully replicated.
The principal product of 4x174 DNA replication catalyzed by the calf polymerase-primase complex is nicked circular DNA (Fig. 6A). Single-stranded circular phage DNA is quantitatively converted to a double-stranded form as illustrated by ethidium bromide staining of the products of the replication reaction after separation by electrophoresis on 1% agarose. In addition to the major population of nicked circular form, a minor population of double-stranded linear molecules is also present. The double-stranded linear molecules are probably produced from single-stranded linear DNA molecules present in the 4x174 DNA used as template. The RNA primers associated with the products when the replication reaction was carried out in the presence of [m-"'P]ATP (lanes a and a ' ) and in the presence of [y-" 'PIATP (lanes b and b') are shown in the autoradiograms in Fig. 6B. Because of the small amount of [y-:"P]ATP incorporation in 4x174 DNA replication products, we thought that the initiation sites on 4x174 DNA might be specific sequences. When the [y-'"P] ATP labeled DNA products were cleaved with H u e 1 1 1 and analyzed by gel electrophoresis, radioactivity was found to be associated with all fragments (data not shown) suggesting no site-specific initiation.
We attempted to analyze the size of the RNA primers synthesized in the replication reaction after isolation of the reaction products by phenol treatment followed by ethanol precipitation. The major problem encountered in this analysis is the poor recovery of the RNA products. This is probably due to solubility of short oligonucleotides in ethanol. In order to examine the total populations of RNA primers synthesized in the reaction, we carried out the DNase I digestion directly on the reaction mixtures at the end of the replication reaction and analyzed [a-""PIAMP-labeled products on a DNA-sequencing gel. Reaction products of poly(dT) and 4x174 DNA replications, with and without DNase I treatment, were loaded on the DNA-sequencing gel twice at 30 min apart, and the autoradiogram of the gel is presented in Fig. 7. The RNA products in poly(dT) reactions containing only ATP exhibit three populations, averaging about 20 nucleotides, 10 nucleotides, and 3 to 4 nucleotides (lanes A and A'). Treatment of these RNA products with DNase I did not significantly alter the size distribution (lanes B and B'). The RNA products synthesized in poly(dT) replication when dATP was present also showed three size distributions, averaging about 16 nucleotides, 3-4 nucleotides, and polymer form (lanes C and (2'). When these poly(dT) replication products were digested with DNase, the oligoriboadenylates produced belong to two size distributions, 9-10 nucleotides and 3-4 nucleotides (lanes D -+ "-c -+ "+ -+ -+ " -+ To the remainder of each sample, SDS and EDTA were added to 0.5% and 10 mM, respectively. The reaction mixtures were then diluted with an equal volume of 98% formamide, 10 mM EDTA, 0.1% bromphenol blue, and 0.1% xylene cyanol. After further dilution with 49% formamide, 5 mM EDTA, 0.05% bromphenol blue, and 0.05% xylene cyanol, the samples were boiled for 5 min, and 2 pl of each diluted sample were loaded on the urea-  A and A', B and B', C and C', D and D', E and E', F and F and D'). These results suggest that in the absence of dATP, the primase synthesizes primers in multiples of 10 nucleotides as well as smaller species. In the presence of dATP in poly(dT) replication, primers synthesized by the primase and utilized by DNA polymerase have an average size of 9-10 nucleotides. The 16-nucleotide population found in the untreatedpoly(dT) replication products (lanes C and C') may represent primer with a few deoxyadenylate residues. The results and interpretation of the primer sizes in poly(dT)-primed reactions obtained in this study are consistent with those observed with the mouse primase and the human KB cell DNA polymeraseprimase (10, 19).
Discrete RNA primers detected in 6x174 DNA reactions are 10 nucleotides or shorter when dNTPs were omitted from the reaction mixture (Fig. 7, lanes E and E ' ) and 8 nucleotides or shorter when dNTPs were included (lanes G and G'). No discrete larger species were detected in 9x174 DNA reactions. Treatment of 6x174 reaction products with DNase I did not affect the size distribution of RNA primers (lanes F, H, F' and H') suggesting the absence of deoxynucleotides in the primers synthesized by the calf enzyme complex.

DISCUSSION
Calf thymus was the original source of material for largescale purification of eukaryotic DNA polymerase-a (20), leading to the demonstration of the high-molecular weight (21), oligoribonucleotide initiation (22), and absence of associated exonucleases (23). Traditional methods of purification (1,4,24) demonstrated multiple species of DNA polymerase-cy, and interpretation of the basis of the heterogeneity observed has not been possible.
The existence of DNA primase, providing oligoribonucleotide initiation for DNA synthesis, has now been demonstrated in several eukaryotic systems (8,18,19,(25)(26)(27)(28) following the lead established by genetic and biochemical studies on E. coli DNA polymerase I11 (29). The DNA primase presumably occurs as some sort of a complex with DNA polymerase-a, but genetic analysis is not easily available in eukaryotic systems.
Use of monoclonal antibody purification now provides the possibility for more extensive analysis of the proteins participating in the DNA replication process in complex cells. The multipeptide complex we have isolated in the current work is of high specific activity and can be obtained in large quantity and excellent yield. The complex contains three unique peptides and three or more "associated" lower-molecular-weight species. We assign the 160-kDa peptide as the catalytic subunit for deoxynucleotide polymerization based on previous studies demonstrating the association of DNA polymerase activity with this peptide by renaturation in situ after electrophoresis on polyacrylamide gels in the presence of SDS. 3 We associate the 68-kDa peptide with DNA primase. The last association is rather tentative, and the functions of other proteins remain to be determined.
The rather simple peptide pattern obtained suggests that we have not isolated the complete DNA replication complex. Replication of 4x174 DNA by the complex is ribonucleotide dependent, showing rapid initiation and closely coupled DNA synthesis of full 4x174 RF I1 molecules. Double-stranded DNA molecules are not replicated to any appreciable extent. Additional protein components and origins of proper DNA replication are obviously missing in our current studies.
Our studies used several monoclonal antibodies to demonstrate that the yield of active complex obtained by immu-noaffinity purification is a function of the monoclonal antibody used. This suggests that further studies should explore a variety of monoclonal antibodies for yield, peptide structures, and enzyme activities present. The DNA polymerase-DNA primase complex obtained in this study should be a good source of antigen for generating new antibody panels.
The monoclonal antibody columns used in this study should also permit examination of cell-extraction procedures with the intent of devising conditions that will allow isolation of complexes more approaching the true biological state. Complexes of this kind should provide a clearer picture of peptide stoichiometry. Through use of immunoaffinity purification, the analysis of DNA replication in eukaryotic systems at the molecular level now appears to be resolvable.