"Cytoplasmic" deoxyribonucleic acid polymerase. Structure and properties of the highly purified enzyme from human KB cells.

The freshly prepared crude cytoplasmic fraction of aqueously extracted KB cells contains a single major species of DNA polymerase activity (DNA polymerase C) that sediments homogeneously in low ionic strength sucrose gradients with a peak at 10.8 S. The enzyme activity from frozen crude extracts sediments heterogeneously under these conditions with peaks at 8.4 and 10 S. In 0.45 M salt-containing gradients all of the polymerase activity is recovered as a single 6.4 S species. When purified to a specific activity of 7,300, DNA polymerase C sediments in low ionic strength gradients as a single species of 6.5 S. From combined sedimentation and gel filtration analysis, we estimate the molecular weight of the active protomeric species of the polymerase to be about 170,000. Under no conditions of ionic strength does the enzyme disaggregate to active species smaller than 6.4 to 6.5 S. Sodium dodecyl sulfate-polyacrylamide gel analysis of the most highly purified enzyme fractions reveals two major protein bands of 87,000 and 175,000 daltons, respectively. These data suggest that DNA polymerase C contains an 87,000-dalton component and permit the interpretation that the active protomer of Mr equal 170,000 may be a dimer. The purified enzyme shows maximal activity with gapped duplex DNA and has an absolute requirement for 3'-hydroxyl termini. It utilizes initiated polydeoxynucleotide templates poorly and initiated polyribonucleotide templates not at all. Although the polymerase is inhibited by PPi it has only minimal ability to promote PPi exchange (0.8% of the polymerase activity). The purified enzyme is free of endonuclease and exonuclease activities (less than or equal to 0.003% of the polymerase activity) and demonstrates no primer-template-dependent conversion of substrate dNTP to free dNMP during the polymerization reaction. Finally, DNA polymerase C does not excise misparied primer termini from a synthetic homopolymer primer-template but can utilize such termini as initiation sites, although at a very slow rate.

W. DAVID  In recent years it has been recognized that eukaryotic cells contain at least four classes of DNA polymerase activity that can * These studies were supported by Grant NP-96B from the American Cancer Society and Grants CA-14835 and AI-08806 from the National Institutes of Health.
be distinguished to varying degrees by their apparent subcellular distribution and by physical, enzymatic, and immunological properties.
These generally accepted classes include a DNA polymerase species that is associated with mitochondria (l-4); a physically heterogeneous (6 to 12 S) high molecular weight activity, usually derived from the cytoplasmic fraction of aqueously extracted cells (5-16), which we have designated KB cell DNA polymerase C (17, 18); a homogeneously sedimenting (3 to 3.5 S) low molecular weight polymerase, obtained from purified nuclei (7, 9-11, 13, 14, 19-22), which we have designated KB cell DNA polymerase Nl (17,18,23) ; and an activity that has been identified both in nuclear and cytoplasmic fractions and that shows a marked predilection for the replication of oligodeoxynucleotideprimed homopolymeric ribonucleotide templates, particularly of the structure oligo(dT) .poly(A) (R-DNA polymerase) (24)(25)(26)(27)(28)(29). Although the precise roles played by these enzymes in in viva DNA replication remain to be established, particular interest has focused on the 6 to 12 S "cytoplasmic" polymerase for several reasons. First, this species constitutes the majority of the total DNA polymerase activity that is ordinarily recovered from aqueously prepared crude extracts of growing cells. Second, the activity of this species, in contrast to the 3 to 3.5 S nuclear polymerase, is responsive to cell proliferation, increasing significantly when quiescent cells are stimulated to divide (30, 31) and falling sharply as cell division ceases-(32,33).
Third, a number of reports have described in purified nuclear fractions a high molecular weight DNA polymerase activity (9, 17, 18) whose presence and quantity appear to be closely correlated with active cell multiplication (13,21,34). Our physical studies reported here and recently published immunological evidence (35) suggest that this large nuclear polymerase and the "cytoplasmic" enzyme are very similar, if not identical. Fourth, although the "cytoplasmic" enzyme was the first eukaryotic DNA polymerase to be identified (5), it has thus far proved particularly difficult to purify. Thus the structure of the enzyme, the significance of the physical heterogeneity that it exhibits in crude or partially purified preparations (16,(35)(36)(37), its proposed association with a cytoplasmic smooth membrane fraction (10, 21), and its putative relationship to other eukaryotic DNA polymerases, particularly to the 3.5 S nuclear enzyme (3%41), remain matters of current controversy.
In previous publications we have described the isolation and partial characterization of three DNA polymerases from freshly harvested KB cells (17,18)   Under optimum homopolymer assay conditions, the specific activity was generally 57, of that observed with activated DNA.
Sucrose Gradient Centrifugation-Linear 5 to 207, (w/w) sucrose gradients were prepared as before (17). The gradients were buffered with 0.01 M Tris-HCl, pH 7.9, and contained 1 mM each of pmercaptoethanol and EDTA with or without NaCl at 0.45 or 2.0 M. Enzyme solution was loaded onto the top of the gradient, and sedimentation was carried out for 17 hours at 40,000 rpm at 5" in the SW 40.1 rotor in a Spinco L-265B centrifuge.
At the end of the centrifugation, fractions of equal volume were collected by pumping from the bottom of the centrifuge tube at 1 ml/min with a Technicon pump. Sephadex G-200 Gel Filtration-Gel filtration was carried out essentially as before (17. 18). Specific conditions for each filtration run are given in the figure legends.
Vertical Slab Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis-Gel electrophoresis was carried out in a vertical slab gel apparatus (17) with a 23-ml linear gradient of 5 to 12% or 5 to 15Yo polyacrylamide gel concentration (bisacrylamide to monomer weight ratio, 0.8/30) used as a seDarating ael. ElectroDhoresis was initiated and maintained at 30 ma and 30" until the front-running buffer ions had migrated into the separating gel, and then the current was increased to 60 ma. The gel slab was fixed and stained according to Weber and Osborn (44). Quantitation of protein content of stained bands was performed by measuring absorption at 650 nm with a Transidyne RFT densitometer (Transidyne General Corp., Ann Arbor, Mich.), and employing the linear portion of a calibration curve derived from the reference protein, human r-globulin.
Samples were prepared for gel electrophoresis as follows.
(a) For sodium dodecyl sulfate treatment, aliquots of enzyme fraction were dialyzed for 17 hours at room temperature against 1 mM Tris-HCl, pH 6.8, 0.0270 sodium dodecyl sulfate and then lyophilized to dryness.
The dry pellet was-redissolved in >fc of the original sample volume of 10% glycerol in Hz0 and then boiled for 5 min in a 100" water bath. (b) For NaCl treatment, enzyme fractions were first dialyzed against the desired molarity of NaCl for at least 2 hours and then dialvzed extensivelv against several changes of the Tris-sodium dodecyl sulfate buff& and processed as described above. (c) Reduction and alkylation of enzyme fractions with p-mercaptoethanol and iodoacetamide were performed by a modification of a standard procedure (45). Protein (10 to 50 mg/ml) was equilibrated with Hi for 2 to 3 hours. For each 1% of protein, dithiothreitol was added to 7 mM, and the solution was incubated for 1 hour at 37". For each 1% of protein, iodoacetamide was then added to 20 mM. The protein solution was incubated with the alkylating agent for 1 hour at 0" and then prepared for sodium dodecyl sulfate treatment as above.

Assay of Exonuclease
Activities-The assay for 3'-+5'-exonuclease activity was carried out exactly as before (23) KPi. The peak fractions were pooled and concentrated by dialysis against solid sucrose at 4" (Fraction V).

Second Phosphocellulose Column Chromatography
The concentrated enzyme fractions were adjusted to 0.15 M KPi, pH 7.2, and loaded on a phosphocellulose column (2.5 x 7.5 cm). The column was washed with 3 bed volumes of the same buffer and was then developed with a 50.ml linear gradient of 0.15 M to 0.31 M KPi, pH 7.2. The peak fractions of enzyme activity were pooled, concentrated by dialysis against solid sucrose, and stored over liquid nitrogen (Fraction VI). After 6 months of storage, about 40% of Fraction VI activity was lost, but longer storage for over 1 year did not result in any further loss of activity.
In each of the three steps of column chromatography in this purification protocol only a single major peak of DNA polymerase activity is detected by the standard assay with activated DNA primer-template. Although the practice of pooling peak tubes at each step might lead to the discarding of minor polyrnerase species, we have no evidence to suggest the presence of additional significant species of D-DNA polymerase activity3 in our Fraction III preparations.

Sephadex G-200 Gel Filtration
When Fraction VI was filtered on a column (1.5 x 20 cm) of Sephadex G-200 that had been equilibrated with 0.01 M Tris-HCl, pH 7.9, 1 mM each of EDTA and &mercaptoethanol, and 10% glycerol, about 707, of the loaded enzyme activity was recovered in a single peak that eluted close to the void volume (V,/V, = 1.06) (Fraction VII).
Of note was the fact that Fraction VII contained only 5 to 10% of the total Coomassie blue staining material that, was present in Fraction VI, which might indicate a specific activity of Fraction VII enzyme of >lOO,OOO. However, since Coomassie blue is not a specific stain for protein on sodium dodecyl sulfate gels (48,49), it is not yet clear to what extent this step of purification removes protein, phospholipid, glycolipid, or some combination of these. Because of the instability of Fraction VII, all of the studies of the catalytic properties of polymerase C that are presented here were carried out with Fraction VI enzyme.

Distribution
of Polymerase C during Subcellular Fractionation Baril et al. (10,21) have reported that the cytoplasmic DNA polymerase from rat liver is associated with a subcellular fraction that is rich in smooth cndoplasmic reticulum. We attempted to confirm this observation by fractionating freshly harvested KB cells according to the protocol of Baril et al. (10). Although the KB cell cytoplasmic DNA polymerase activity was indeed associated with a pellet fraction (corresponding to fraction P4 of Baril et al. (10)) containing smooth endoplasmic reticulum, the polymerase activity could be clearly separated from the membrane-rich fraction by subsequent sedimentation through a discontinuous sucrose gradient. As shown in Fig. 1, DNA polymerase C activity is found predominantly in a region of the gradient that contains no recognizable structures other than occasional dense material that resembles glycogen.
1. Distribution of cytoplasmic DNA polymerase activity during subcellular fractionation of KB cells. The figure shows the recovery of DNA polymerase C activity from a discontinuous sucrose gradient (A) and representative fields of Gradient Fractions 2 (B), 3 (C), and 4 (D) as observed by electron microscopy. Freshly harvested KB cells were suspended in 0.025 M Hepes buffer, pH 7.5, 1 mM EDTA, 2 mM MgC12, and 1 mM P-mercaptoethanol and were broken by Dounce homogenization (17). The resulting crude extract was adjusted to 0.025 M Hepes, pH 7.5, 3 mM MgC12,0.33 mM EDTA, 1 mM fl-mercaptoethanol, 20 mM KCl, and 0.2 M sucrose. Nuclei and mitochondria were removed by centrifugation at 10,000 X g for 15 min, and microsomes were pelleted by centrifugation at 100,000 X g for 90 min. The supernatant was then centrifuged at 78,000 X g for 16 hours to obtain a pellet fraction that is essentially equivalent to the P4 fraction of Barilet al. (10). 68% of the initial DNA polymerase activity in the crude extract was recovered in this step, 47y0 in the pellet and 21% in the supernatant.
The pellet was carefully suspended in 0.025 M Hepes, pH 7.  Fig. 2 and Table II. In sucrose density gradients of low ionic strength, the polymerase activity in frozen crude cytoplasmic extracts exhibited the usual heterogeneous sedimentation profile with peaks at about 8 S and 10 S ( Fig. 2A). In contrast, in the same gradients, the polymerase activity in freshly prepared cytoplasmic extracts sedimented as a relatively homogeneous single species with a peak at 10.8 S (Fig. 2B). With either preparation of Fraction I enzyme, sedimentation in gradients containing 0.45 M NaCl yielded a single sharp peak of polymerase activity at 6.4 S (Fig. 2C), and if the 6.4 S species was dialyzed against 0.01 M Tris-HCl, pH 7.9, and subsequently sedimented through a hypotonic sucrose gradient, the heterogeneous activity profile was regenerated, with the reappearance of peaks at about 8 S and 10 S (Fig. 20). In all of these experi-7049 ments the recovery of loaded enzyme activity was always better than 90% (Table II).
The sedimentation of Fraction IV enzyme was examined after preparation either by the standard procedure (Table I) or by a modified protocol from which the step of acid precipitation (Step II) had been omitted.
Sedimentation of the modified Fraction IV enzyme produced an activity profile (Fig. 2E) very similar to that observed with frozen crude extract, while sedimentation of the standard Fraction IV preparation (Fig. 2E) showed a substantial shift of the activity profile toward predominance of the 8 S component.
This shift is not due to the acid precipitation step itself, however, since acid precipitation of freshly prepared Fraction I enzyme does not significantly alter the sedimentation profile illustrated in Fig. 2B. As observed with the crude extracts, centrifugation of Fraction IV enzyme in gradients containing 0.45 M NaCl resulted in quantitative conversion of the enzyme activity to the 6.4 S species (Table II). Finally, sedimentation of Fraction VI enzyme in gradients with or without NaCl (Fig. 28') revealed only a single peak of enzyme activity at 6.5 S, although different preparations of Fraction VI occasionally showed some skewing of the activity profile toward higher S values in gradients lacking NaCl.
Analysis by Gel Filtration-The results of Sephadex G-200 gel filbration analysis of DNA polymerase C are presented in Fig. 3. At low ionic strength, Fraction II enzyme from the freshly prepared cytoplasmic extract filtered as a large aggregate in a sharp peak coincident with the void volume of the column (Fig. 3A). In the presence of 0.45 M NaCl, the same enzyme fraction eluted sharply just ahead of the bovine y-globulin marker (Fig. 3B). Filtration of Fraction VI enzyme in the absence of salt yielded a somewhat heterogeneous peak of activity between the void volume and the position of bovine y-globulin (Fig. 3A), while in the presence of 0.45 M NaCl, the elution profile was identical to that of Fraction II illustrated in Fig. 3B. The results from gel filtration agree with those from sedimentation analysis in that they provide no evidence for the presence of more than a single principal species of DNA polymerase activity in the cytoplasmic preparation. ?vloreover, under no condition of gel filtration, including the presence of 2 M NaCl or 100 mM P-mercaptoethanol (data not shown), have we ever observed conversion of polymerase C to active species of the size of DNrl polymerase Nl. Estimation of the apparent M, of Fraction VI enzyme by a standard plot (Fig. 3C) yielded values of about 2255,000 (no salt) and 190,000 (0.45 M salt), respectively. These values are considerably larger than those predicted from sedimentation COefficients (Table II) and demonstrate the anomalous behavior of this enzyme on gel filtration that has been noted with crude enzyme fractions from calf and rat by others (15,16,36). To obtain a possibly more valid approximation of the size of the Fraction VI polymerase activity we estimated Stokes radius values from our gel filtration data (56, 57) (Fig. 30) and used these, together with the S values (Table II)   we consistently observed the presence of one or both of two discrete bands, which we have designated Bands A (Mr of 175,000) and B (M, of Si,OOO), respectively.
Sodium dodecyl sulfate-polyacrylamide gel analysis of Fraction VII enzyme indicated that 90 to 957, of the total Coomassie blue stainable material in Fraction VI was removed by gel filtration, but there was persistence of protein Bands A and B. To determine whether these bands might be related to the polymerase C protein, two portions of Fraction VI enzyme were separately filtered through Sephadex G-200, and the elution profile of polymerase activity was determined. In one case, each fraction in the activity peak was individually reduced and alkylated and then subjected to sodium dodecyl sulfate gel electrophoresis. By this procedure, both Bands A and B were detected, and they represented >60% of the total Coomassie blue staining material in the gel (Fig. 4A). The protein concentration of Band A plus Band B closely followed the profile of polymerase C activity (Fig.  4B). In the second case, each fraction across the peak of activity was separately treated with 2 M NaCl and then analyzed by sodium dodecyl sulfate gel electrophoresis.
As seen in Fig. 4C, the gel pattern revealed a single predominant component, Band B, the concentration of which across the peak was closely coincident with the profile of polymerase C activity (Fig. 40). Thus by this technique, Band B, with M, of 87,000, is the major detectable protein in the peak of the polymerase activity that elutes at the void volume of the G-200 column in a position consistent with a size of about 200,000 daltons. Within the limits of protein quantitation by measurement of Coomassie blue stain intensity, the total amount of protein recovered in Band B of Fraction VII, after NaCl treatment, was about equal to that recovered in Bands A plus B of Fraction VII after reduction and alkylation. Some Comments Regarding KB Cell DNA Polymerase NI--We have previously (17,18) reported the presence of a high molecular weight DNA polymerase activity in purified nuclear fractions from KB cells. This activity, DNA polymerase N2, could be clearly distinguished from DNA polymerase Xl and seemed also to differ in certain enzymatic properties from DNA polymerase C. However, polymerase N2 had not been highly purified, and thus its identification as a separate polymerase species remained uncertain. We have examined some of the physical properties of a preparation of polymerase N2 that had been purified through the step of DEAE-cellulose column chromatography (17, 18). When this fraction was centrifuged through sucrose gradients in the absence or presence of salt, sedimentation profiles identical with those illustrated in Fig. 2 (Fig. 5A) but shows a peak at pH 9.25 with Mg2+ as divalent action. In contrast, with (dT)%-poly(dA) as primer-template and Mn2+ as divalent cation, the pH profile of the enzyme peaks sharply at pH 7.6.  respectively.
Kinetics of Polymerization-Under standard conditions with activated DNA primer-template (Fig. 50), dNMP incorporation is brisk and reasonably linear for 120 min but) ceases soon thereafter. The termination of the reaction is due to enzyme instability, since the addition of fresh enzyme at 120 min (Fig. 50)  Under these conditions the E. coli polymerase generated acid-soluble [32P]dTMP at a rate equal to 5~~ of its rate of polymerization with activated DNA. By contrast,, 5'+3'-exonuclease activity could be excluded from polymerase C to a level of < 1O-4% of its polymerizing activity.

Y-G-Exonuclease
Activity-DNA polymerase C had no detectable 3'-5-exonuclease activity (<0.002% of the polymerizing activity) when tested with a double-stranded DNA substrate that was very highly labeled at every 3' terminus (23) ( Table  III).
Under these same assay conditions, the 3'45'.exonuclease activity of E. coli polymerase I was readily detected at a ratio of nuclease activity to polymerase activity of 0.94%. When polymerase C was challenged with a 3'-terminally labeled singlestranded DNA substrate, or with a 3'.terminally mismatched homopolymer substrate, a trace level of exonuclease activity was observed that was close to the limit of resolution of the assay (Table III).
In contrast, when the same mismatched homopolymer was exposed to E. coli polymerase I, complete hydrolysis of the mispaired dCMP residues to the theoretical limit predicted from the known degree of enzyme saturation occurred within 2.5 min. We believe that the trace of 3' -+5'-exonuclease activity that is present in polymerase C Fraction VI cannot be considered a significant polymerase-associated activity and is most likely due to a minor contaminant. and 10% giyceroi.'Polymerase activity (0---0) was-recovered in a single peak (V,/V, = 1.06) (only those fractions that contained enzyme activity are shown in B and D); recovery of loaded activity was 45% in B, 7170 in D. A and B, each of the fractions in the activity peak was separately reduced and alkylated and then analyzed by sodium dodecyl sulfate gel electrophoresis.
The gel lane corresponding to the peak fraction of enzyme activity in Panel B is shown in Panel A; Bands A and B comprise 62yo of the total Coomassie blue staining material in the gel. of Band A plus Band B in each fraction is represented by the closed circles. (The details of protein quantitation are described under "Methods").
C and D, each of the fractions in the activity peak was separately treated with 2 M NaCl and then analgzed by sodium dodecyl sulfate gel electrophoresis. The gel lane corresponding to the peak fraction-of enzymeactivity in Panel D is shown in Panel C: Band B comDrises 6497, of the total Coomassie blue staining material in the gel: The concentration of protein in Band B in each enzyme fraction is represented by the closed circles. The standard marker proteins are: 1, bovine yglobulin; 2, conalbumin; 3, ovalbumin; 4, chymotrypsinogen; 6, cytochrome C.
polymerase Nl. In contrast, with DNA polymerase C we have been unable to detect this activity under any conditions of polymerization that were tested (Table IV). When, the incorrect substrate, dCTP, was tested with the primer-template (dT)k. poly(dA), there was neither incorporation nor turnover of dCTP. This result is identical with that we previously reported (23)  At the times indicated, 250 ~1 aliquots were removed and assayed for the incorporation of [3H]dTMP. At 120 min (1 ) 1.5 units of fresh enzyme were added to a 0.5 ml portion of the reaction mixture.
Since the fraction of total dCMP residues that was at the 3'. terminal position was 0.7 (23), the data indicate that by 120 min of reaction, about 10% of the mismatched termini had been utilized for polymerization.

DISCUSSION
This paper presents a detailed study of the structure and enzymatic properties of DNA polymerase C (17, IS), the KB cell counterpart of the high molecular weight polymerase activity that has been identified as the principal DNA polymerase species in a wide range of eukaryotic sources that include yeast (65), sea urchin (34), tetrahymena (66), chick embryo (14,61), and diverse mammalian cells and tissues (5-7, 9-13, 15-18, 37). Although the "cytoplasmic" DNA polymerase was the first eukaryotic DNA polymerase to be isolated (5, 6), its molecular structure has not yet been established.
The heterodisperse sedimentation profile that crude polymerase preparations exhibit in low ionic strength density gradients has proved perplexing and has led to uncertainty as to whether the "cytoplasmic" polymerase might be comprised of more than a single enzyme species (16,31,36), as well as to claims that the large polymerase is an aggregate form of the 3.5 S nuclear polymerase (39-41). The DNA polymerase activity in the freshly prepared crude cytoplasmic fraction of aqueously extracted KB cells sediments in low ionic strength sucrose density gradients as a single discrete species of 10.8 S and is not associated with smooth membranes (10,21) or with any other recognizable subcellular structure. The 10.8 S activity is quantitatively converted to a 6.4 S protomeric species in the presence of 0.45 M NaCl, and this conversion is at least partially reversible. The shape of the sedimentation profile varies to some extent with the degree of enzyme purification, and the most highly purified preparation, polymerase Fraction VI, sediments identically as a homogeneous 6.5 S activity in the presence or absence of salt. From combined sedimentation and gel filtration analyses (56) we estimate the molecular weight of Fraction VI enzyme activity to be between 170,000 and 200,000. This estimate of the size of the active polymerase C protomer is in good agreement with those reported for the comparable polymerase activity from human lymphocytes (ll), rat liver (15), and chick embryo (14).
Although DNA polymerase C has not yet been purified to homogeneity, sodium dodecyl sulfate-polyacrylamide gel analyses of Fraction VII suggest that protein Bands A and 13, of 175,000 and 87,000 daltons, respectively, may be major components of the enzyme, and that Band A may be a dimer of Band B. Although not conclusive, these findings permit the interpretation that the active protomer of polymerase C is a 175,000-dalton dimer that is composed of two enzymatically inactive 87,000dalton monomers.
We note that Smith et al. (67) have observed a major band of protein of 89,000 daltons in a partially purified (specific activity, 359) preparation of the "cytoplasmic" DNA polymerase from cultured human lymphoblastoid cells. The sedimentation and gel filtration data indicate that polymerase C has a pronounced tendency to associate at low ionic strength wit.h itself or with other unidentified constituents of similar size (at least as defined by S value), and that our purification scheme results in the removal or alteration of factors that are required for this interaction.
Although such aggregation has been encountered in most studies of "cytoplasmic" DNA polymerase activities from whatever source, we do not yet understand the phenomenon, nor do we know whether it may be of biological significance or is merely adventitious. Our findings do strongly suggest, however, that conclusions regarding multiplicity of "cytoplasmic" polymerase species (16, 36)) interconversions between polymerase species (39-41), and hypothetical schemes of polymerase regulation (37, 40) that are based largely on alterations or differences in sedimentation and gel filtration profiles should be received with caution.
It is important to emphasize that by no method of analysis of DNA polymerase C that we have employed, neither by sedimentation nor gel filtration at salt concentrations up to 2 M, nor by sodium dodecyl sulfate gel electrophoresis, have we ever observed active enzyme species of 3 to 3.5 S or protein bands of 43,000 to 45,000 daltons that might correspond to DNA polymerase NI (23). We underscore this point because there are several reports in the literature (38-41) that suggest that these two Nl itself forms large aggregates that can be dissociated to the 3.5 S monomer in the presence of salt. The polymerase Nl aggregate retains the characteristic enzymatic properties of the monomer and can thus be unambiguously distinguished from the polymerase C protomer or aggregate forms by appropriate tests (17). However, given the strong propensity of both of these classes of DNA polymerase to participate in ionic strengt.h-dependent aggregation reactions, and considering that crude cell and tissue extracts are usually prepared in hypotonic buffers that favor such interactions, it is apparent that attempts to discriminate these classes of activity in impure preparations on the basis of size criteria can be erroneous.G L We had previously reported that DNA polymerase C used exonuclease III-treated DNA primer-template about 45% as well as activated DNA primer-template (17,18). We have subsequently found that the relative degree of utilization varies profoundly with different preparations of the former, probably reflecting differences in the length of the gaps that are introduced and consequently the actual primer-template conformation that the polymerase encounters.
B Dr. N. B. Hecht has just sent us a preprint of a paper in press in Biochim. Biophys. Acta in which he has clarified his earlier observations regarding the interconvertibility of the large and small DNA polymerases.
He has now obtained evidence that the cytoplasmic fraction of murine cells contains two species of large DNA polymerase, only one of which can be converted by salt to a 3.5 S form. Hecht thus concludes that the salt-dissociable species is an aggregate form of the small polymerase, while the non-dissociable 7 to 10 S species represents the "cytoplasmic" DNA polymerase.
In striking contrast to the apparent structural complexity of DNA polymerase C, the enzymatic capabilities of the highly purified Fraction VI protomer are remarkably simple, an assessment that is in good agreement with that reached by Chang and Bollum (8) from their study of the comparable calf thymus enzyme. The general conclusion is now emerging (8, 23) that the eukaryotic DNA polymerases generally lack most or all of the associated properties that are so characteristic of their counterparts in prokaryotes (68). From the data presented here and in our earlier report (17), we may summarize the catalytic properties of polymerase C as follows. (u) With an appropriate primertemplate, which may be either gapped DNA or oligodeoxynucleotide-or oligoribonucleotide-initiated polydeoxynucleotide template, polymerase C will incorporate correctly paired dNTPs 011 3'.hydroxyl primer termini with the stoichiometric cleavage of PPi. However, this enzyme is totally unable to copy an initiated polyribonucleotide template.
(b) The enzyme carries out a detectable, but minimal, PPi exchange reaction at a very slow rate that is just below the limit of detection of our earlier study of polymerase Nl (23). Thus it remains possible that the latter polymerase might in fact possess a PPi exchange capability of this very low magnitude.
(c) Polymerase C is devoid of endonuclease and exonuclease activities.
(d) The enzyme cannot perform the dNTP turnover reaction, a capability that is prominently displayed by our homogeneous preparations of DNA polymerase NI (3). (e) Like polymerase Nl, polymerase C not only fails to excise mispaired primer termini, but it can utilize such termini as functional initiation sites, albeit at only about 10% of the rate observed with the nuclear enzyme. This lack of stringency with regard to the conformation of the 3'-primer terminus is not a peculiarity of the DNA polymerases derived from malignant or transformed cells since the same behavior has been clearly established for the small nuclear polymerase from calf thymus (8,69). It thus appears that for both of the two well defined classes of D-DNA polymerases, replication fidelity can be accomplished only by means of appropriate dNTP selection. If either polymerase encounters or introduces a mispaired dNMP, it possesses no mechanism for error correction and, at least in the in vitro situation, will replicate through it. In the last few years studies in E. coli have sharply defined the obligatory participation of a large number of purified proteins and as yet uncharacterized gene products in meaningful, in vitro DNA replication (68, see Chapter 7). The use of more sophisticated primer-templates than activated DNA or synthetic homopolymers with eukaryotic DNA polymerases, in an effort to identify other, non-polymerase components of a "replication complex, " is a clear direction for future research in this area. This approach is perhaps made even more compelling by virtue of the very limited catalytic repertoire of those highly purified eukaryotic polymerases that have been characterized t.o date.