A comparison of associated enzyme activities in various deoxyribonucleic acid polymerases.

Abstract A comparison of activities associated with Escherichia coli polymerase I, calf thymus 6 to 8 S DNA polymerase and calf thymus 3.4 S DNA polymerase demonstrates that neither exonuclease activity nor pyrophosphorolysis is essential for polymerization. The 3.4 S enzyme has no demonstrable 3'- to 5'-exonuclease activity and cannot carry out pyrophosphate exchange. Under these conditions phosphodiester bond formation is kinetically irreversible.

polymerase I, calf thymus 6 to 8 S DNA polymerase and calf thymus 3.4 S DNA polymerase demonstrates that neither exonuclease activity nor pyrophosphorolysis is essential for polymerization.
The 3.4 S enzyme has no demonstrable 3'-to 5'-exonuclease activity and cannot carry out pyrophosphate exchange.
Under these conditions phosphodiester bond formation is kinetically irreversible.
The over-all reaction for "replicative" polymerization of deoxynucleoside triphosphates into DNA is commonly written : dNTP + initiated template Enzyme \ dNilIP-initiated template + PPi M++ and implies a pyrophosphate exchange. The substitution of OH-for PPi in this formulation produces exonuclease activity (I). The relationship of the polgmerase function to the pyrophosphate exchange and hydrolytic functions in Escherichia coli polymerase I has been studied extensively (l-5).
The high molecular weight species of DNA polymerase from calf thymus also has a pyrophosphate exchange (6), but exonuclease activities are apparently absent.
This report demonstrates that the low molecular weight DNA polymerase (also from calf thymus glandj exhibits neither detectable pyrophosphate exchange nor esonuclease activities.
These results indicate that pyrophosphorolysis and hydrolysis are not necessary consequences of polymerization function of DNA polymerases and point to the 3.4 S DNA polymerase as a favorable subject for mechanism studies.
The difference between the two species of mammalian DNA polymerases in "associated" enzymes activities is also of special interest, since they are immunologically related (7).
* This research was supported by United States Pttblic Health Service Research Grant CA 08487 from the National Cancer Institute.
Enq?les-Calf thymus low molecular weight I)Nh polymerase was prepared from calf thymus chromatin using slight modifications of the procedure described for rabbit bone marrow low molecular weight DNA polymerasc (8). The enzyme was further purified by affinity chromatography on denatured DSX cellulose (9). The enzyme preparation obtained is essentially homogeneous and the purification procedure is reported in a separate publication (10). The high molecular weight D?;A polymerase was prepared from the soluble extract of calf thymus gland, essentially as previously described (11). A homogeneous preparation of E. co2i DNA polgmerase I (12) was generously supplied by Dr. L. -4. Loch, Institute for Cancer Research, Fox Chase, Philadelphia, Pa. Chemicals-Deoxynucleoside triphosphatcs (dNTl's) ,I polydeoxynucleotides, polydeosynucleotides containing a radioactive 3'.terminal nucleotide, oligodeosynucleotides and DNase I-treated calf thymus DNA were prepared as described earlier (8,13). Polydeoxynucleotide-oligodeoxynucleotide complexes and polydecxynucleotide-polydeosynucleotide complexes were prepared by physical mixing of the appropriate materials in 0.01 M KC1 followed by heating to 60" and slow cooling.
The use of initiated polydeoxynucleotide templates has been described in earlier work (13,14 Abbreviations for homopolymers bearing subscripts refer to the chain length. lIefinite sllbscript numbers refer to the chain length of oligon!lcleotides. A bar over a definite subscript ntlmber denotes average chain length. Ammediol is used as a trivial name for 2-amino-2-methyl-1,3-propandiol. PP i is used to abbreviate pyrophosphate ion, and OH-is used to abbreviate hydroxyl ion. except that 0.05 M ammediol buffer at pH 8.6 or Tris-Cl at pH 7.6 was used as buffer. The E. coli polymerase I reactions were carried out as described for calf thymus 6 to 8 S polymerase except that 8 mM MgClz was used. Polymerase activity was measured by the increase of acid-insoluble material on GF/C discs (15). One unit of DNA polymerase is defined as 1 nmole of dTMP incorporated into acid-insoluble material per hour. Specific activity of the enzyme is defined as units of enzyme activity per mg of protein.
The init,iated template system used to determine the specific activity of various DNA polymerases is d(pA)s.d(pT)lt.
The specific activities of enzymes used with this template system are 167,000 for E. coli polymerase I, 5,000 for calf thymus 6 to 8 S enzyme, and 250,000 for calf thymus 3.4 S enzyme.
Pyrophosphate an aliquot of the reaction mixture was added to ice-cold 5% trichloroacetic acid, nucleotides were absorbed onto activated charcoal, the charcoal was washed, and the nucleotides were then eluted from the charcoal with 3 M acetic acid in pyridine.2 Quantitative elution of dNTP was obtained by this method.
The amount of [32P]dNTP was det.ermined by counting Cerenkov radiation of the acetic acidpyridine solution directly in the liquid scintillation counter. Pyrophosphate exchange reactions with calf thymus 3.4 S polymerase were carried out under identical conditions or using 0.05 M Tris-Cl (pH 7.6) or 0.05 M ammediol buffer (pH 8.6) as buffer.
E. coli polymerase I pyrophosphate exchange reactions were carried out under conditions similar to those used for the 6 to 8 S calf thymus polymerase and for the 3.4 S calf thymus polymerase.
The polymerization activity in pyrophosphateinhibited reactions was assayed under the usual exchange conditions except that nonradioactive sodium pyrophosphate and [3H]dTTP were used, and the progress of the reaction was followed by measuring acid-insoluble material. Pyrophosphate exchange reactions for the various enzymes were also carried out with initiated homopolymer templates under conditions described for the polymerization reactions in the presence of various amounts Of [32P]PPi.
Controls for pyrophosphate exchange reactions were carried out in the absence of divalent cation, DNase I-treated calf thymus DNA, or dNTPs.
The lack of pyrophosphatase activity in the calf thymus DNA polymerase preparations was ascertained by assaying radioactive orthophosphate production from pyrophosphate by the method of Martin and Doty (16). The Cerenkov radiation of the isobutyl alcohol-benzene solution of phosphomolybdate was counted directly in the liquid scintillation counter.
Pyrophosphorolysis Reactions-Pyrophosphorolysis reactions of the various DNA polymerases were measured in conjunction with pyrophosphate exchange reactions when dNTPs were omitted from the reaction mixtures.
An alternative method, similar to that described by Brutlag and Kornberg (5), for measuring pyrophosphorolysis activity was also used in this study. in the presence of 2 mM sodium pyrophosphate, 8 mrvr MgC12, 1 mM 2-mercaptoethanol, 100 pg per ml of bovine serum albumin, and 0.02 M potassium phosphate at pH 7.2. Progress of the reaction was followed by taking aliquots of the reaction mixture at various times, adding an equal volume of a marker solution containing dNTP, dNDP, and dNMP (dTTP, dTDP, and dTMP when the matched template is used, and dCTP, dCDP, and dCMP when the mismatched template is used) in 0.1 M EDTA, and separating the entire mixtures on Whatman No. 1 paper in isobutyric acid-NH3-H20.
The chromatograms were dried, ultraviolet absorbing areas were cut out, and the radioactivity was counted in the liquid scintillation counter. Pyrophosphorolysis reactions using the matched and the mismatched templates were also carried out using the calf thymus 3.4 S polymerase except that the reactions were carried out in 0.05 M Tris-Cl at pH 7.6, 0.5 rnnt MnC12, 1 mM 2-mercaptoethanol, 100 pg per ml of bovine serum albumin, and 0.2 mM sodium pyrophosphate.
Y-to 5'-Exonuclease Reactions-The 3'-to 5'-exonuclease reactions of the various DNA polymerases were carried out using the single-stranded 3' terminus-labeled polydeoxynucleo- in the presence or absence of [14C]dTTP. The reaction conditions were the same as described for polymerization reactions.
Reaction progress was followed by measuring the decrease of acid-insoluble material using the glass fiber disc method.
When [i4C]dTTP was present, radioactivities were determined by double isotope counting procedures.
Since the 3'-to 5'.exonuclease activity in the calf thymus enzyme preparations was absent or at a very low level, these enzyme preparations were tested for the exonuclease activity at various PI-IS, in MgC12 or MnC12, and over a range of protein concentrations.
Deoxynucleoside Triphosphate Degradation Reactions-The dNTP degradation reactions of the calf thymus enzymes were carried out under conditions described for polymerization except that 0.5 mM [14C]dTTP (5 times the template nucleotide concentration; specific activity, 8000 cpm per nmole) was used, and the reactions were carried to completion (one complete round of replication of the available template nucleotide). Analysis of the products of the reaction with added markers (dTTP, dTDP, and dTMP) was carried out on Whatman No. 1 paper strips developed with isobutyric acid-NHs-H20 and liquid scintillation counting as above.
5'-to S'-Exonuclease Reaclions-Degradation from the 5' terminus was measured by using d ( Progress of reactions was followed by measuring the decrease of acidinsoluble radioactive material.

Polymerization-All
three enzymes used in this study seem to have similar requirements for polymerization. The essential reactants are the template: initiator system, divalent cation, and dNTPs.
The enzymes act by extending the initiator chain by phosphodiester bond formation with complementary nucleotides Although almost 700,000 cpm were exchanged by E. coli polymerase I in a 10.min reaction, no charcoal-absorbable count significantly above the background was found for calf thymus 3.4 S polymerase after 1 hour of reaction.
Considerable experimentation has been done to verify and evaluate the significance of the negative results obtained with the 3.4 S polymerase.
In addition to the 30-fold increase in enzyme concentration and loo-fold increase in specific activity of the [32P]pyrophosphate in the reaction noted above, complete pyrophosphate concentration curves, studies at pH 7.2, 7.6, and 8.6, and exchange with synthetic templates in Mg++ and Mn+f were carried out. All attempts to observe an exchange reaction using the 3.4 S polymerase were negative.
The 3.4 S enzyme was tested directly for pyrophosphatase and dNTP degrading activities and none were found.
We conclude that the 3.4 S DNA polymerase does not catalyze a pyrophosphate exchange.
Pyrophosphorolysis-The ability of polymerase I from E. coli  (Table II) as part of the controls in the pyrophosphate exchange experiments described in Table I. No pyrophosphorolysis was found with 3.4 S polymerase, although both polymerase I from E. coli (Table II) and calf thymus 6 to 8 S polymerase were active (1, 3, 6).
Using template-initiator systems containing a/-labeled termini, we also examined the nature of the products of degradation in the presence of pyrophosphate.
The results listed in Table III  The 3.4 S polymerase produced no hydrolytic or pyrophosphorolytic products.
Although the 3'-to 5'-nuclease activity was easily demonstrated with polymerase I, no significant level of this activity could be detected with the 6 to 8 S and 3.4 S calf thymus DNA polymerases. The 3'-to 5'.exonuclease activity described as "proofreading exonuclease" in E. coli polymerase I proceeds from the 3' terminus toward the 5' end (5). The results obtained with E. coli polymerase I and calf thymus DNA polymerases are exhibited in Fig. 1. In contrast to the easily detectable level of activity observed with polymerase I, t,he 3.4 S calf thymus polymerase had no proofreading exonuclease,3 and the activity seen with the 6 to 8 S enzyme was really too low to be interpreted as an important "associated" activity.
In the pyrophosphorolysis ex- periment (Table III), where enzyme concentrations were higher sively studied for Td DNA polymerase (17). This reaction is and when the products were analyzed, calf thymus 6 to 8 S en-assumed to be a combination of hydrolytic and polymerization zyme produced no hydrolytic product ([3H]dTMP) when a reactions. When the calf thymus enzymes were tested for dNTP matched template was used and produced a very low level of degradation, none was observed, as might be expected for enhydrolytic product ([3H]dCMP) when the template with a mis-zymes not having 3'-to 5'-exonuclease activity. A typical matched base was used. result for calf thymus 3.4 S enzyme is shown in Fig. 2

. The
Deoxynucleoside Triphosphafe Degradation-The observation trace amount of [14C]dTMP present (note that the figure is on a that dNTPs are converted to dNMPs in a template-dependent 4 cycle log scale) in the substrate was unchanged during the 160reaction was first described by Deutscher and Kornberg for E. min incubation.
Similar results were obtained with the 6 to 8 S coli polymerase I (3). The same reaction has also been extenpolymerase. To examine the nature of PPi inhibition of calf thymus DNA polymerases, we studied the effects of PPi on the polymerization rates of the calf thymus enzymes at various concentrations of dTTP.
These results are shown in Fig. 3. At 0.215 M PPi or less inhibition of the 6 to 8 S polymerase was noncompetitive with respect to dTTP (Fig. 3A). At higher concentrations of PPi (above 0.286 mu) the inhibition was mixed and eventually became competitive at concentrations of 0.358 mM and 0.429 mM. The 3.4 S DNA polymerase was inhibited by PPi at concentrations of PPi less than 0.15 mM and the inhibition was competitive with respect to dNTP (Fig. 3B). Since the optimum Mn++ ion concentration (0.25 to 0.5 mu) was about the same for both calf thymus enzymes under our reaction conditions, the competitive inhibition of the 3.4 S enzyme by such low pyrophosphate concentrations could not be due to depletion of Mn++ ions (present at 0.6 mM). The inhibition of the 6 to 8 S calf thymus enzyme at such concentrations (less than 0.215 mM PPi) appears to be noncompetitive. The data clearly indicate that at higher concentrations of PPi (above 0.215 mM) the kinetics of inhibition becomes competitive between PPi and dNTP.
Since PPi is competing with dNTP for Mn++ ions only at higher concentrations, the competitive inhibition seen with the 3.4 S enzyme at less than 0.15 mrvr PPi cannot be attributed simply to competition for Mn++ ions. Equilibrium dialysis studies with E. co& polymerase I showed that' pyrophosphate does not compete with dNTP (18) and kinetic studies indicated t,hat PPi competes with OH-in the FIG. 2. Degradation of deoxynucleoside triphosphate by calf thymus 3.4 S polymerase.
The conditions for the reaction and the analyses of the products were as described under "Materials and Methods" with 126 units of polymerase present. Frame A shows the distribution of products after 1 min of incubation.

Frame B
is after 40 min of incubation and Frame C is after 160 min of incubation. The numbers 1, 2, 3, and 4 designate the posit,ions of polymer product, dTTP, dTDP, and dTMP, respectively. pyrophosphate exchange reaction catalyzed by E. coli polymerase I (3). The results of our experiments, together with those reported for polymerase I, allow the simple conclusion that the action of PPi on calf thymus 3.4 S enzyme is different from its action on polymerase I (3, 18) and calf thymus 6 to 8 S enzyme. The pyrophosphate exchange activity levels of these enzymes also suggest that the action of PPi is different for each enzyme.
It is also interesting to note that orthophosphate is an extremely effective inhibitor for the calf thymus 3.4 S enzyme.
E. coli polymerase I and the calf thymus 6 to 8 S enzyme are not affected at comparable concentrations of Pi. The effects of Pi on E. coli polymerase I, calf thymus 6 to 8 S DNA polymerase, and calf thymus 3.4 S polymerase are compared in Table IV.
Our unpublished results show that potassium phosphate at neutral pH is the most suitable buffer for general use in calf thymus 6 to 8 S enzyme reactions. Although Tris-Cl, N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (Hepes), and cacodylate buffers at comparable pH can be used in poly dA replication, they are somewhat inhibitory with DNA, poly dC, and poly dT templates. Table IV shows that calf thymus 6 to 8 S enzyme activity is not inhibited by Pi at pH 7.0 in the presence of 25 mM Pi. When Tris-Cl at pH 7.6 was used as buffer, the calf thymus 6 to 8 S enzyme was more sensitive to Pi, yet E. coli polymerase I was not inhibited by Pi under the same conditions. As would be expected, the effect of Pi in a DNA polymerase reaction is a complex function of the divalent cation, the type of buffer used, and the pH of the reaction.
The strong inhibition of the 3.4 S enzyme probably cannot be explained by the competition for R/In++ between dNTP and Pi and the result amplifies the differences that may be found in these DNA polymerases.

DISCUSSION
The enzymatic properties and associated enzyme activities of E. coli polymerase I have been studied extensively (l-5, 16, 17 frequently been used as a model for all other DNA-synthesizing enzymes. With the exception of the 5'-to 3'-exonuclease function of polymerase I the associated enzyme activities appear to be present in all prokaryotic DNA polymerases studied (19). The highly purified enzyme preparations we obtained from calf thymus gland have allowed us to carry out a comparison of the enzyme activities of these mammalian DNA polymerases with E. coli polymerase I. The results clearly indicate that the mammalian DNA polymerases we have examined differ from E. coli polymerase I both in the kinds and the levels of associated enzyme activities present.
A summary of the results of this comparison is shown in Table V.
The absence of the 5'-and 3'-exonuclease function of DNA polymerase has been demonstrated in a number of prokaryotic and eukaryotic enzymes (19). Results presented in this study as well as our unpublished data on several purified mammalian DNA polymerases showed the absence of this activity. This particular exonuclease activity appears to be present only in certain prokaryotic DNA polymerases. Both E. coli polymerase I and the calf thymus 6 to 8 S enzyme carry out pyrophosphate exchange and pyrophosphorolysis reactions.
In contrast to E. coli polymerase I the 6 to 8 S enzyme did not exhibit a significant level of 3'-to 5'-exonuclease activity with matched templates nor did it degrade dNTP.
Since it was possible to demonstrate the presence of a trace level of 3'-to 5'-exonuclease activity in the 6 to 8 S calf thymus enzyme when a mismatched template was used, a final interpretation of this finding must be postponed.
The 6 to 8 S enzyme preparation used in this study is not homogeneous and it may be possible   On the other hand, it may be possible that the conformation of the template-enzyme complex determines the magnitude of the 3'to 5'-exonuclease function.
We have noted, for example, that the calf thymus 6 to 8 S enzyme does seem to stabilize the interaction between the initiator chain and the template strand4 (20) and this stabilizing effect of the enzyme may be a reflection of conformation near the active center of the enzyme.
In any event, it should be emphasized that the ratio of exonucleolytic activity to polymerizing activity observed with the 6 to 8 S enzyme is several orders of magnitude lower than that of E. coli polymerase I. The absence of all associated activities in the calf thymus 3.4 S enzyme is the most unusual finding in this investigation.
Although it does seem obvious that pyrophosphorolysis is merely reversal of the polymerization reaction of DNA polymerase (5, 6), the negative findings with the 3.4 S polymerase suggest that this may not be generally true. The different effects of PPi on the various DNA polymerases also suggest that all features of the reaction pathways of polymerization and pyrophosphorolysis may not be identical in these enzymes. The competition between PPi and dNTP on the calf thymus 3.4 S enzyme and the lack of a pyrophosphorolysis reaction suggest that binding of PPi at the dNTP site on the enzyme surface does not allow pyrophosphorolysis to take place. The noncompetitive effect of PPi with dNTP on calf thymus 6 to 8 S enzyme suggests the possibility of a secondary site on this enzyme that does bind PPi. The presence of a PPi-binding site around the catalytic center might effect a higher concentration of PPi there and may also be required for the apparent reversal of the polymerization reaction (pyrophosphate exchange and pyrophosphorolysis) .
If the enzymes studied are involved in DNA metabolism in living cells then it is of interest to consider possible biological implications of associated activities. For example, the 3'-to 5'.exonuclease found associated with E. coli polymerase I has been called a proofreading exonuclease (5). This term implies that base pairing errors occurring during polymerization are efficiently removed before further polymerization takes place. This type of activity has also been demonstrated in Td-induced polymerase and the fact that this polymerase does produce some errors may be deduced from the observation that noncomplementary dNTPs are converted to dNMPs in homopolymerdirected reactions (17). A relation between 3'-to 5'-exonuclease activity and "antimutator" activity has also been observed in the Tq system (5, 21). It will be of interest to examine the mechanisms that control the error frequency in biological sysplate and initiator, d(pT)s can be used effectively by polymerase I to initiate poly(dA) replication.
terns that contain DNA polymerases devoid of prcofreading exonuclease activity.
The absence of pyrophosphate exchange and pyrophosphorolysis associated with the 3.4 S enzyme is somewhat discomforting to chemical instincts and conveys no obvious biological advantage.
At the present moment we can only assume that there must be some beneficial effect of the apparent irreversibility of DNA synthesis catalyzed by this chromatin-bound species of DNA polymerase.