Synthetic Template Specificity of a Deoxyribonucleic Acid Polymerase from Regenerating Rat Liver*

The template response of a purified DNA polymerase from regenerating rat liver has been studied. This enzyme does not copy a polyribonucleotide template; however, the ribo strand of a synthetic RNA-DNA hybrid can be used as an initiator for the synthesis of the complementary deoxyribo strand. This enzyme also has a marked preference for at least partially single stranded polydeoxyribonucleotide templates. The copolymer poly[d(A-T) .d(T-A)] is multiplied 4-or 8-fold. The single stranded poly(dA) acts as an efficient template for the regenerating rat liver DNA polymerase, provided short complementary oligodeoxyribonucleotide initiators are added. added initiator. of the polydeoxyribonucleotide in an amount of to the amount The


SUMMARY
The template response of a purified DNA polymerase from regenerating rat liver has been studied. This enzyme does not copy a polyribonucleotide template; however, the ribo strand of a synthetic RNA-DNA hybrid can be used as an initiator for the synthesis of the complementary deoxyribo strand.
This enzyme also has a marked preference for at least partially single stranded polydeoxyribonucleotide templates.
The single stranded poly(dA) acts as an efficient template for the regenerating rat liver DNA polymerase, provided short complementary oligodeoxyribonucleotide initiators are added.
Poly(dC) is used as a tempIate without added initiator.
Synthesis of the complementary polydeoxyribonucleotide continues in both cases until an amount of polymer equal to the amount of initial template is produced.
The enzyme does not copy poly(dG), either in a single stranded form in the presence of oligo (dC) or in a duplex form such as poly[(dC) .(dG)].
The specificities detected with synthetic polymers show that this DNA-dependent DNA polymerase is distinguishable from RNA-dependent DNA polymerases.
We previously reported the characterization of DNA polymerases present in normal rat liver and hcpatomas, with special emphasis on their template specificities (I, 2). A DNA polymerase present in regenerating rat liver has been purified and optimal conditions (pH, Mg++, and I<+ ion requirements, saturation concentrations of substrates and templat,cs) for its action have been determined.
The present paper deals with the use of synthetic templates by the partially purified DNA polymerase from regenerating rat liver in order to clarify the relat,ionship between the enzyme and the illiti~~Oor-tenlplatel molecule.
* This work was supported by a Research Grant from the Ligue Nationale Frangaise cont,re lc Cancer and from the Commissariat b 1'Energie Atomique.
1 The term initiator is used to designate t.he chain containing a free 3'OH, physically complexed with the template.
The replication of the template st,nrts from the 3'OH end of the initiator (3). Similar experiments have been reported with I)?r'A polymerase from calf thymus (4-8).
The major emphasis of the present study is a definition of criteria for the dist,inction between cellular and viral I>?~A polymerases (9, 10). These criteria are of great importance for studying cellular transformation by oncogenic RNA viruses since an RNA-dependent DNA polymerasp, a hybrid-dependent DNA polymerase, and a DNA-dependent l)iYh polymerase have been shown to be present in the virions (1 I). Such an approach has been recently pursued by Ualtimore (12) and Spiegelman (13).
Cavalieri (19) first demonstrated that Escherichia coli DSA polymcrase I can use RNA or DNA-RNA hybrids as templates. These resulbs have been recently confirmed at least for solne steps of purification of E. coli polymerase 1 (12,13,20). In contrast, t,wo purified synthetic RNA-dependent DNA polymera.ses can be clearly distinguished from the DNA-dependent DNA polymerases of HeLa cells (21,22). The purified DKA polymerase studied here is strictly DNA-dependent and has been successfully separated from a hybrid-dependent activity present in crude extracts of regenerating rat liver.

MATERIALS AND METHODS
Enzyme Preparation-Two-third part,inl hepntectomies were performed on 20 or 25 male WAG rats weighing 200 to 300 g by removal of the left lateral and median lobes. Forty hours after the partial hepatectomies the rats were killed by decapitation and the regenerating livers rapidly removed :d placed in ice. All subsequent fractionation steps were pert'ormrd at O-4". Homogenates were prepared by disrupting the liver cells with a Potter homogenizer in 3 volumes of Buffer h cont:Lining 0.25 M sucrose (Buffer A:50 mM Tris-HCI, pH 7.5, 1 mM MgCl,, 6 mM KCI, 1 mnI 2-mercaptoethanol).
The whole homogenate was centrifuged at 105,000 X g for 2 hours in a preparative Spinco ultracentrifuge.
The postmicrosomal supernataut. (Fraction I) was separated from the sediment and from the fatty overlayer with a syringe. 132 with continuous mixing 150 ml of an ammonium sulfate saturated solution of Buffer A. After 20 min, this suspension was centrifuged at 5000 rpm for 30 min. The c1ea.r supernat,ant layer was removed and adjusted to 40% saturation of ammonium sulfate in the same way as described above. Sft,er 20 min, the precipitate was sedimented at 5000 rpm for 30 min, and the pellet was dissolved in 50 ml of Buffer A containing 20% (w/v) glycerol.
Then the solution wa,s dialyzed against the same buffer for 18 hours (Fraction II).
At 0.08 Y NaCl, a first peak of DNA polymerase w-as eluted; this fraction still shows an appreciable level of activity in the absence of one to three deoxyribonucleoside triphosphates with DNA as template; moreover, the copying of the ribo strand of a poly[(dT) . (A)] hybrid can be catalyzed by this fraction.
At 0.15 M ,"aCl, a DNA polymerase was eluted, which requires the presence of all four deoxyribonucleoside triphosphates and a denatured DNA as template (Fraction III).
At 0.25 M NaCl and 0.6 M NaGI, inactive proteins and nucleic acids were eluted, respectively.
After dialysis, Fraction III (50 to 60 ml) was adsorbed onto an hydroxylapatite column (1.5 x 25) which had been equilibrated with 0.1 M phosphate buffer pH 7.5, 6 mM KCl, 1 mM MgCIs, 1 mM 2-mercaptoethanol and 20% glycerol (w/v). The column was washed with the same buffer and adsorbed proteins were eluted with 100 ml of a linear potassium phosphate gradient (0.1 to 0.5 M). DNA polymerase activity was eluted at 0.20 M potassium phosphate.
In the presence of poly[d (A-T) .d(T-A)] and denatured DNA, the optimum Mg++ concentration is 2 to 5 m&r, the optimum Kf concentration, 2.5 mM, and the optimum nucleoside triphosphate concentration is 0.2 my each. The activity is proport.ional to the enzyme concentration and under the assay conditions used ribonucleoside triphosphates cannot replace deoxyribonucleoside triphosphates as substrat.es. As judged by polyacrylamide gel electrophoresis, Fraction IV is not yet homogeneous; the DNA polymerase activity is cluted in the exclusion volume of a Sephadex G-200 column.
Endo-and esonuclease activities have been assayed by the incubation of labeled DNA wit.h the enzyme preparation under the conditions of the DNA polymerase reaction with or without the four deoxyribonucleoside triphosphates, or at various pH values. Sephadex gel filtration was used to monitor the course of esonucleolytic and endonucleolytic hydrolysis of DNA and to fractionate the products of digestion accord-ing to the method of Birnboim (23). In all cases less than 2 y0 of the initial radioactivity was eluted in the region of mononucleotides, while a defined peak was eluted in the region of undegraded DNA4. Under these conditions a "nickase" activity cannot be detected and, in fact, t.he prcincubation of denatured DNA with Fraction IV increases by 2-or 3-fold the priming activity of this DNA.
For the present study two different preparations of Fraction IV, having specific activities of 100 and 150 units per mg of protein in the preseuce of poly[d (A-T) sd(T--4) 1, were used. One unit of DNA polymerase is defined as the amount of activity required to convert 1 nmole of total nueleotide into an acid in soluble product in 1 hour under the standard conditions specified below (Table I).
Templates Incubations were carried out at 38", aliquots were withdrawn at various time intervals, and the acid-insoluble fraction was precipitated on Whatman GF/C glass filters in 5% perchloric acid and 2y0 sodium pyrophosphate. Filters were rinsed with perchloric acid, with ethanol, and then dried. The radioactivity retained on filters was measured in a Packard or Intertechnique scintillation counter.
With these counters, efficiency of aH determination was 22% and 36%, respectively.

Detection of Hybrid-dependent
Activity-Tn crude extracts of regenerating rat liver an activity catalyzing the copying of the  In contrast, poly-[(dC) . (dG)] and poly[(dC) .(G)] polymers in R conditions were used as effective templates (Fig. 1). In R and D conditions, NaCl, nucleotide and template concentrations, as well as pH, were different.
A further study of these factors revealed pH to be the most significant parameter.
As shown in Fig. 2, two optimal pH values were found: pH 7 for the use of deoxyribonucleotide polymers containing A and T as a t.emplate, and pH 7.6 for the replication of poly(dC to the same extent and the yield of the reaction rapidly exceeded 100 %.2 Among the single stranded polydeoxyribonucleotides, only poly(dC) was active as a template.
A very small amount of oligo(dG) present in the poly(dC) preparation could possibly act as an initiator although the purity of the poly(dC) preparation was confirmed by the following criteria: (a) spectral base properties and base analysis showed the -presence of a single nucleotide; (b) in the presence of E. coli RNA polymerase only one ribonucleoside triphosphate was polymerized, i.e. poly(dC) could stimulate the polymerization of only GTP. This control was made by Biopolymers Inc. This point will be discussed later.
Addition of short complementary oligodeoxyribonucleotides to poly(dA) or poly(dT) initiated the replication of these single stranded templates.
Only an initiator that could pair with the template to form standard base pairs was active; oligo(dC) was not effective.
In all cases, the poly(dG) sequences of the polymers so far studied were not used as templates.
In the presence of the double stranded homopolymer poly[(dC) . (dG)] the incorporation of dGMP was considerably greater than that of dCMP.
In contrast to poly(dA) or poly(dT), poly(dG) would not act as template for dCMP polymerization even in the presence of oligo(dC).
This fact is not due to inhibitory compounds in the reaction mixture (e.g. deoxyribonucleoside diphosphates or radiation products present in the [3H]dCTP preparation).
Indeed, when a calf thymus DNA was used as template, the incorporation of dCiMP reflected the base composition of this template.
Furthermore, the slight incorporation of the homologous nucleotide in the presence of a single stranded polymer shows that a terminal deoxynucleotidyl transferase activity (24) does 2 (Yield = total mononucleotides incorporated template mononucleotides x 100) 2. Influence of pH on template efficiency of various polynucleotides.
Samples were incubated for 60 min at 38" at the pH indicated.
In 40 mM Tris-

d(A-T) .d(T-A)] with dATP and ["H]dTTP.
Act,iviLy was expressed in nanomoles of labeled deoxyribonucleotide incorporated by 1 ml of enzyme in 1 hour.
not account for the differences in template properties of these polymers.
A comparison of the template capabilities of the ribo strand and the deoxyribo strand of the heteroduplexes poly[(d'l') . (A)] and poly[(dC) .(G)] leads to the conclusion that the enzyme specifically copies a deoxyribonucleotide template. Poly(A) and poly(G) strands of the hybrid polymers are not replicated by the enzyme, although they can be used as initiators.
Addition of oligo(dT) to poly(h) did not stimulate the replication of ~01~ (A).
Template Activity of Deoxyribonucleotide Polymers. Yield of Reaction-Under the standard conditions described above, and with concentrations of poly[d (A-T) .d(T-A)] ranging from 5 to 22 PM, more than a doubling of the template occurred.
The extensive synthesis (Fig. 3) observed in the presence of low concentrations of template implies the utilization of the newly synthesized product as template by the DNA polymerase or a slippage mechanism with hairpin-like extensions from a main chain as postulated by Kornberg (25) for E. coli DNA polymerase I.
As shown in Fig. 4, dAMP and TMP were incorporated to approximately the same extent, resulting in a total synthesis of 14 nmoles of acid-insoluble product for 5.5 nmoles of template added. This synthesis is obtained i n 3 hours, after which no radioactivity is rendered acid-soluble during further incubation.

Comparison of the kinetics of [aH]dGMP incorporation
in the presence of various synthetic polymers having dC sequences shows that poly(dC) was the most effective template. The polymerase reaction proceeded until the template poly(dC) was duulicated and then stopped or continued at a very low rate   . 4). After 10 hours the polymerase reaction is only increased by the addition of template. I'oly(dC) was used as a template without added initiator.
Addition of oligo(dG) increased the initial rate of synthesis without modifying the yield of the reaction.
A decrease of the tot.al synthesis was noticeable only when the ratio oligo(dG):poly(dC) exceeded 1, while doubling the poly(dC) template concont.ration always doubled the yield of the reaction (Figs. 4 and 5). 3 A different situation was observed in the case of poly(dA). Poly(dA) alone is a poor template as a double stranded poly-[(dA) . (dT)], but addition of oligo(d'l') great.ly stimulated the incorporation of [3H]TMP. As shown in Fig. 6, t,he rate of synthesis increased in proportion t.o the amount of oligomer added until the ratio oligo(dT):poly(d:Y) rcachrd 1. Under these conditions, replication of the templntc was complete and the reaction ceases at 100%. When the oligo(dT):poly(dA) ratio was greater than 1, the rcnction st.opped prematurely. Within the range of oligo(dT) concentration studied, no appreciable t.erminal addition was observed.
a As suggested by Baltimore (12), variations around the 100% value for the yield of the reaction can be attributed to errors in the determination of the concentjration of the templates. The incubation was essentially t,he same as in the standard assay system, except for the total volume (1 ml). Aliquots of 25 ~1 were withdrawn at indicated times and assayed as described under "Materials and Methods." Each point represents the average of duplicate experiments. Activity was expressed in nanomoles of labeled deoxyribonucleotide incorporated in the presence of 100 fig of enzyme and 5.5 nmoles of each template.
The late self-initiation described by l3ollum (6) for calf thymus DNA polymerase was not obtained on poly(dA) with DNA polymerase from regenerating rat liver.
The critical factor for the replication of this polymer appears to be the concentration of free 3'OH available to start the reaction, as demonstrated by Baltimore (12)  factor does not appear to be important for poly(dC), and compa.rison of Figs. 5 and 6 clearly shows the difference between the two types of template.
Using synthetic polymers, we have shown the absolute requirement of regenerating rat liver DNA polymerase for a deoxyribonucleotide template and have defined specific properties of this enzyme. In contrast to E. coli DNA polymerase I, the Fraction IV of regenerating rat liver DNA polymerase did not copy a ribonucleotide template; however, this enzyme could use the ribo strand of an RNA-DNA hybrid as an initiator for the copying of the complementary deoxyribo strand.
The two enzymes differ, moreover, in their sensitivity to sulfhydryl group reagents.
In contrast to E. co& DNA polymerase I, this DNAdependent DNA polymerase from regenerating rat liver is inhibited by p-chloromercuribenzoate and N-ethylmaleimide (27) like E. coli DNA polymerases II (28) and III (29).
In other respects, such as the capacity to use poly[d(A-T) .

dQ-A)]
as a template and to achieve a net synthesis with this template, the present enzyme resembIes E. coli DNA polymerase I and the DNA polymerase recently purified from normal and regenerating rat liver (30).
The capacity to catalyze the reaction with partially single stranded DNA and the requirement for a hydrogen bonded initiator for synthesis on a single stranded template poly(dA) or poly(dT) recalls the properties of calf thymus DNA polymerase (7) and I<. coli DNA polymerase II (28). The enzyme studied in this paper differs from the rat liver polymerase purified by Berger et al. (30) which cannot utilize denatured DNA as a template.
The high rate of synthesis on poly(dC) is also a characteristic property of the present enzyme. Spiegelman et aZ. (14) have previously noted that homopolymer poly(dC) is a very good template for different DNA polymcrases.
In our experiments, this is probably due to the partially ordered structure of poly(dC) in the incubation medium (pHm4 = 7.5) (31,32), which may be conducive to a self-initiation.
We have not, however, ruled out the possibility that small amounts of oligo(dG), acting as an initiator, are present either in the enzyme preparation or in the poly(dC) preparation.
In the case of poly(dG), it is probable that this polymer has a very stable secondary structure which inhibits its action as a template for l)NA polymcrase (32,33).
A study of the products synthesized in the presence of known templates is rendered possibIe by the duplication of these templates. The different behavior of the synthetic polymers used could be related to their physical characteristics or to biological properties specific to A-T or G-C sequences (e.g. differences in the affinity of the enzyme for these sequences, involvement of these sequences in binding, initiation or termination processes) (34). Further investigation with highly purified regenerating rat liver DNA polymerase and with natural DNA should give us better understanding of the enzymatic specificities exhibited in the presence of synthetic polymers.
We are in the process of studying the in vitro replication of the two components of Cancer pagurus DNA: on the one hand poly d(AT), a duplex DNA which is predominantly an alternating copolymer of A and T with about 3 mole% of GC base pairs, on the other hand, main component with 45 mole% of GC base pairs. 4 pH, denotes the midpoint of t.he pH-induced helix-random coil transition taken at the pH where the increase in absorbance is 5070 complete (31).
The results with homopolymers demonstrate that DNA-dependent DNA polymerase and RNA-dependent DNA polymerases are distinguishable.
The use of homopolymer templates with oligodeoxyribonucleotide initiators provides an effective system for testing unknown enzymes for their template preference.
The ribo strand-dependent activity detected with poly[(dT) b (A)] in crude extracts prepared from regenerating rat liver is separated from the DNA-dependent activity in the course of the DEAE-cellulose chromatography (see "Materials and hiethods").
This fact suggests that these activities correspond to different enzymes or at least to different parts of a complex structure which may be concerned with the regulation of replication and transcription in viva. Recently Chang and Bollum (3, 35, 36) have found several forms of DNA polymerase in eukaryotic cells (rat liver included).
The low molecular weight species may behave, notably with poly[(dT) .(A)], as reverse transcriptases but do not use oligoribonucleotide initiators. The high molecular weight forms are the major activity in rapidly growing cells; these forms do not copy a polyribonucleotide template but use oligoribonucleotide initiators. Both of the molecular weight forms of DNA polymerase from eukaryotic cells are inhibited by antibody directed against the high molecular weight DNA polymerase from calf thymus gland. This observation suggests a common subunit or polypeptide sequences between various forms of DNA polymerase in t,he cell (37). The properties of DNA polymerase from regenerating rat liver, studied here, are similar to the properties of the high molecular weight species; in fact, this enzyme is eluted in the exclusion volume of a Sephadex G-200 column (see "Materials and Methods").
We are in the process of purifying the hybriddependent DNA polymerase of regenerating rat liver to compare this enzyme with the DNA-dependent DNA polymerase of the same tissue and with the reverse transcriptase of oncogenic RNA viruses.