Serological analysis of human deoxyribonucleic acid polymerases. Preparation and properties of antiserum to deoxyribonucleic acid polymerase I from human lymphoid cells.

The preparation and properties of an antiserum to human DNA polymerase I (6 to 8 S) are described. Care was taken in the purification of the antigen to remove certain other DNA polymerases found in human cells. An incubation of antigen and antiserum lasting about 48 hours is necessary to achieve maximal inhibition. About 1 mug of the antipolymerase immunoglobulin G, prepared in rats, neutralizes 60% of the activity present in 54 ng of the enzyme. Tritrations varying both antiserum and enzyme demonstrate clear regions of antigen and antibody excess. Inhibition of enzyme activity is about the same whether the templateprimer is (dA)n-(dT)12-18, or partially digested DNA. An assay was developed which measures the remaining activity in the supernatant after precipitation of enzyme-antibody complexes with goat anti-rat immunoglobulin G. In this assay, 2.2 mug of the antipolymerase immunoglobulin G quantitatively bind 33 ng of DNA polymerase I. With use of the direct neutralization assay and the immuno-precipitation test, we found little, if any, antigenic relationship between DNA polymerase I and DNA polymerase II (3.4 S). Similarly, little, if any, relationship was found to the DNA polymerases from five RNA tumor viruses. The activities of RNA-directed DNA polymerases from the blood leukocytes of two patients with acute myelogenous leukemia and from the placentas of rhesus monkeys were not inhibited in neutralization assays which were shortened because these enzymes were thermolabile. In identically shortened neutralization assays, the antipolymerase immunoglobulin G neutralized up to 76% of the activity of DNA polymerase I. In addition to its utility in distinguishing cellular DNA polymerases, the rat antiserum should be useful reagent for testing of novel DNA polymerases isolated in small quantities from human tumors for contamination with DNA polymerase I. This enzyme is present in abundance in proliferating tissue and often confuses the biochemical characterization of these novel enzymes.

R. GRAHAM  An assay was developed which measures the remaining activity in the supernatant after precipitation of enzyme-antibody complexes with goat anti-rat immunoglobulin G. In this assay, 2.2 pg of the antipolymerase immunoglobulin G quantitatively bind 33 ng of DNA polymerase I. With use of the direct neutralization assay and the immunoprecipitation test, we found little, if any, antigenic relationship between DNA polymerase I and DNA polymerase II (3.4 S). Similarly, little, if any, relationship was found to the DNA polymerases from five RNA tumor viruses. The activities of RNA-directed DNA polymerases from the blood leukocytes of two patients with acute myelogenous leukemia and from the placentas of rhesus monkeys were not inhibited in neutralization assays which were shortened because these enzymes were thermolabile.
In identically shortened neutralization assays, the antipolymerase immunoglobulin G neutralized up to 76% of the activity of DNA polymerase I. In addition to its utility in distinguishing cellular DNA polymerases, the rat antiserum should be a useful reagent for testing of novel DNA polymerases isolated in small quantities from human tumors for contamination with DNA polymerase I. This enzyme is present in abundance in proliferating tissue and often confuses the biochemical characterization of these novel enzymes.
Two major DNA polymerases have been isolated, partially or completely purified, and characterized from proliferating mammalian cells (l-7).
DNA synthesizing activities also have been isolated from mammalian cell cytoplasm (8)(9)(10) and mitochondria (11)(12)(13)(14) which, although less adequately characterized, are probably distinct from the two major polymerases. Another DNA polymerase, which is biochemically (15,17) and immunologically (16,17) closely related to the RNA-directed DNA polymerases from type C RNA tumor viruses derived from primate tumors, has been isolated from human leukemic blood leukocytes. As part of an effort to distinguish or define relationships among each of these enzymes, we are preparing antisera to highly purified, biochemically distinguishable DNA polymerases from human cells. This approach is especially helpful in the absence of complete purification of these enzymes, which is often difficult because available quarltities of tissue are small and some of the enzymes are labile.
In this paper, we describe the preparation, characteristics, and specificity of an antiserum to DiC'A polymerase I1 from human lymphoid cells. The purified antiserum inhibits other cellular DNA polymerases and DNA polymcrases from a variety of RNA tumor viruses only at concentrations at least 50 times higher than those required to inhibit DNA polymerase I. EXPERIMENTAL  cellulose, which did not occur in previous work (5). This loss may be due in part to difficulties encountered in scaling up the procedures; e.g., the large volume of Fraction 2 required 3 days to process on phosphocellulose. The final step of purification, chromatography on DEAE-cellulose, is illustrated in Fig. 1A. Chromatography of DNA polymerase 1 (Fraction 5) on DNAcellulose is shown in Fig. IB. The major peak of activity (using    Table  I), arranged in order from left to right, performed in 7.5% polyacrylamide gels ("Experimental Procedure"). Gels were 9 cm long; electrophoresis was for 7 hours at 7.8 ma per tube. The amounts of protein applied to each gel were 66, 52, 50, 31, and 33 pg, respectively, for Fractions 1 through 5.
ing sodium dodecyl sulfate is shown in Fig. 2 The major band of protein in Fraction 5 has a mobility corresponding to a molecular weight of 89,000. Although this band may not correspond to a component of DNA polymerase I, a major band of identical mobility has been observed in highly purified fractions of DNA polymerase I isolated from normal blood lymphocytes.4

Properties of DNA Polymerases
I and II-DNA polymerase I used for antigen possesses properties that clearly distinguish this enzyme from other cellular DNA polymerases. The enzyme sediments in sucrose gradients at s20,W = 6.5, and elutes at 1.15 times the void volume on gel filtration through Eephadex G-200. The enzyme binds strongly to DEAE-cellulose at pH 7.0, suggesting that DNA polymerase I is an acidic protein (isoelectric point < 7.0). This enzyme is fully inhibited by 0.25 mM Nethylmaleimide and has a Ki of 2.7 PM for the dCTP analogue I-fi-n-arabinofuranosylcytosine 5'-triphosphate (32).
DNA polymerase II purified from RPM1 1788 cells sediments at ~20,~ = 3.6 and elutes at about 2 times the void volume on Sephadex G-200. This enzyme binds strongly to phosphocellulose and does not bind to DEAE-cellulose at pH 7.0, suggesting that this enzyme is a basic protein, as is DNA polymeiase II purified from blood lymphocytes (isoelectric point 9.4) (5). DNA polymerase II is not inhibited by 0.25 mM N-ethylmaleimide and has a Kc of 12.8 PM for I-P-n-arabinofuranosylcytosine 5'-triphosphate (32). The wide separation of DNA polymerase I from polymerase II achieved during each of the last three steps of purification should ensure that polymerase I (Fraction 5) is free of polymerase II. The final step should remove any remaining traces of polymerase II, since, under the conditions used, this enzyme appears quantitatively in the unbound fraction. Chromatography of DK A polymerase I (Fraction 5) on DNA-cellulose was performed to determine the extent of contamination by DNA polymerase III, since this procedure resolves these two enzymes (25,33). Although the recovery of DNA polymerase I was low, DNA polymerase III was undetectable at a sensitivity of assay adequate to detect 1% of the activity of DNA polymerase I (Fig.  1B). In direct assays of DNA polymerase I (Fraction 5), under conditions optimized to measure DNA polymerase III (S), a specific activity of 1 unit per pg was detected. This is 0.27% of the specific activity measured in Assay System A. DNA polymerase I (Fraction 5) contains no detectable terminal deoxynucleotidyltransferae activity.
Preparalion of Antiserum-A prolonged course of immunization was necessary before enzyme neutralizing activity was obtained; five immunizations of 470 pg protein per injection were required. Subsequently, the titer of neutralizing antibodies was maintained at an approximately constant level by triweekly immunizations with DNA polymerase I for an additional four immunizations. A total of approximately 30 ml of immune serum was recovered from the animal.
Properties of Neutralization Reaction-The kinetic data of neutralization of DNA polymerase I are shown in Fig. 3. In this experiment, small quantities of IgG (1.0 pg) and enzyme (65 ng) were used. It is noteworthy that under these conditions, a prolonged preincubation (48 to 72 hours) of polymerase and immune IgG was required to achieve maximal inhibition of enzyme activity. At this low concentration, nonimmune IgG did not consistently influence enzyme activity; amounts greater than 5 pg per reaction mixture stimulated enzyme activity. For this reason, in comparative neutralization assays (Figs. 6 and 7), total rat IgG was kept constant, whereas immune IgG was varied. Incubation for longer than 12 hours caused a partial loss of enzyme activity even in the absence of immune IgG (Fig. 3). This 10~s was minimized by including purified bovine serum albumin in the reaction mixture. The albumin also minimized but did not completely abolish the stimulation of enzyme activity by nonimmune IgG.
If larger amounts of immune IgG (>25 pg) were used, partial inhibition of DNA polymerase I occurred without any preliminary incubation of antigen and IgG. Under these condit,ions inhibition was independent of the time of incubation up to 1 hour, or of the temperature of incubation up to 37". Assessment of the effect of longer incubation periods at 25" or 37" was impossible because of thermolability of the polymerase. Therefore, in the routine neutralization reaction, a period of 48 to 72 hours of incubation of antigen and IgG at 4' was adopted as the best compromise between comnletion of immune inhibition and avoidance of thermal loss of enzyme activity.
A titration of increasing amounts of immune IgG against two fixed amounts of DNA polymerase I is shown in Fig. 4. The enzyme was incubated with IgG for 72 hours prior to determination of residual enzyme activity, which is expressed as a per cent of the activity in the presence of the same amount of nonimmune TgG. The activity of 16.3 ng of enzyme was reduced to 40% of control by 1.2 I.cg of immune IgG; approximately 16 times as much immune IgG was required to reduce the activity of 1630 ng of enzyme to this level. By increasing the immune IgG to 19 pg, more than 950/, of the activity of the smaller amount of enzyme was inhibited.
A titration of increasing amounts of DNA polymerase I against a fixed amount of immune IgG (3.8 pg) is shown in Fig. 5. Above enzyme inputs of 100 ng, progressively less inhibition was observed, until at 1630 ng of enzyme only 15% inhibition occurred. This experiment clearly defines a region of "antigen excess" in the titration.
In experiments not shown, immunological inhibition of DNA polymerase I was about the same whether the template-primer was activated DNA, (dA),. (dT) 12-18, or (dC), . (dG)n-rs. Inhibition was also independent of concentration of template-primer over a range of 10 to 40 pg per ml for activated DNA and from 2.5 to 20 pg per ml for (dA)n .(dT)rz-1s. These ranges extend from limiting to saturating concentrations of template-primer under Assay Conditions A and B, respectively.

Double Antibody Immunoprecipitation
Assay-In these experiments, nonimmune or anti-DNA polymerase I IgG was mixed   6 and 7). All of the activity of this polymerase is precipitable in the double antibody assay (Table II). Therefore, under these Procedure" and in the legend to Fig. 6, containing < 2.5, <5, < 12.5, <20, and 13.5 ng of protein of the above enzymes, respectively.
After incubation, DNA polymerase activity was assayed as described under "Experimental Procedure." O-O, control titration of DNA polymerase I as described in Fig. 6, performed simultaneously.
conditions, 2.2 pg of immune IgG quantitatively bound 33 ng of DNA polymerase I.
SpeciJicity of Antiserum-In the neutralization assay, (Figs. 6 and 7), 1.0 pg of immune IgG neutralized 56% of the activity of DNA polymerase I. DNA polymerase 11 from RPMI 1788 cells or from blood lymphocytes was not inhibited by this quantity of immune IgG (Fig. B), nor were the DNA polymerases from five RNA tumor viruses (Fig. 7). Therefore, at low concentra-tions of immune IgG, this antiserum distinguishes DNA polymerase I from these other polymerases.
At much higher inputs of immune IgG, some of the other polymerases were partially inhibited.
DNA polymerase II from RPM1 1788 cells was not inhibited by immune IgG at 64 pg per reaction mixture; in a control reaction, this amount of IgG inhibited 99% of the activity of DNA polymerase I (Fig. 6). In the same experiment, 64 pg of immune IgG inhibited 33% of the activity of DNA polymerase II from blood lymphocytes (Fig. 6). This quantity of immune IgG inhibited none of the activity of the DNA polymerase from the Mason-Pfizer virus and a minority of the activities of the polymerases from murine leukemia virus, simian sarcoma virus, and gibbon ape lymphosarcoma virus (Fig. 7). Only the polymerase from avian myeloblastosis virus was inhibited more than 50y0 (Fig. 7).
In the double antibody immunoprecipitation assay, none of the activity of DNA polymerase II from RPhlI 1788 cells or of the RNA-directed DNA polymerases from murine leukemia virus or gibbon ape lymphosarcoma virus was precipitable. About one-fourth of the activity of DNA polymerase 11 from blood lymphocytes and one-third of the activities of the RNA-directed DNA polymerases from avian myeloblastosis and simian sarcoma viruses were precipitable (Table II).

DISCUSSION
The purification of antigen was designed to prepare DNA polymerase I as free as possible from contamination by other human DNA polymerases, specifically DNA polymerase II (l-7) and DNA polymerase III (a-10, 25, 33). Therefore, chromatography on DEAE-cellulose at pH 7.0 was added to the purification scheme that we have previously described (5). This step effectively removed remaining traces of DNA polymerase II. Moreover, previous reports suggested that DEAE-cellulose chromatography separates DNA polymerase III from DNA polymerase I (8,9). This procedure produced DNA polymerase I (Fraction 5) which was contaminated by less than 0.3% DNA polymerase II15. The purified enzyme (Fraction 5) was also free of terminal transferase activity.
Thus, although Fraction 5 was not purified to homogeneity (Fig. 2), it was substantially free of other DNA polymerase activities found in human cells. This degree of purity is required for the production of analytically useful antiserum, i.e. useful in distinguishing or relating various human DNA polymerases and determining any possible relationships to the DNA polymerases of RNA tumor viruses.
The properties of DNA polymerase I are similar to those previously described by a number of workers (l-6, 25, 32).
An unusual feature of the neutralization reaction is the requirement for prolonged incubation of antigen and immune IgG in order to reach maxirnal inhibition (Fig. 3). This feature differs from the kinetics of immune inhibition of RNA-directed DNA polymerases described by others (21,23). Chang and Rollum (39) obtained an antiserum in rabbits to calf thymus DNA polymerase I. These workers incubated the antigen-antibody mixture overnight prior to enzyme assay, but detailed neutralization kinetics was not described.
Prolonged kinetics is a well known characteristic of immunoprecipitation reactions (40), including the precipitation of enzymes (41). In our experience, prolonged incubations were particularly necessary for maximum L We found (25,33) that chromatoaraphv on DNA-cellulose provided a better separation of DNA polymerases I and III than did DEAE-cellulose chromatonraDhv. DNA-cellulose chromatography was avoided in the pFepLrlt,ion of DNA polymerase I, however, because the yields were low. enzyme neutralization when the concentrations of DNA polymerase I and immune IgG were low (Fig. 3). Demonstration of typical features of immunological titrations, e.g. regions of antigen and antibody excess (Figs. 4 and 5), required prolonged incubations to achieve maximal inhibition.
If incubations were brief (less than 1 hour), misleading results were obtained. For example, the end point of the immune IgG was greatly underestimated.
Also, in titrations of fixed immune IgG against increasing concentration of enzyme, a flat curve was obtained in the neutralization assay rather than the concave upward curve shown in Fig. 5. Attempts to increase the rate of enzyme inhibition by increasing the temperature of incubation were unsuccessful due to the thermolability of DNA polymerase I. In comparative neutralization tests, the assays were designed to contain very small amounts of DNA polymerases.
In this way, the detection of cross-reactions was as sensitive as possible since the greatest extent of antibody excess was achieved.
These conditions required the use of sensitive assays for DNA polymerases.
However, certain types of cross-reaction will not be detected in these neutralization assays. For example, classes of immune IgG might bind to heterologous DNA polymerases in ways which do not affect or even stimulate these enzymes. This type of binding might occur at regions distant from the active sites of the polymerases.
For this reason, a method was needed of assaying directly for antigen binding, and a double antibody immunoprecipitation test was developed. This assay measures the fraction of DNA polymerase activity not bound by immune IgG and does not depend solely on direct inactivation of enzyme activity by immune IgG.
W7ith use of both the neutralization and immunoprecipitation assays, we tested the specificity of anti-DNA polymerase I IgG. In neither of these assays did this JgG inhibit or bind to DNA polymerase II from RPM1 1788 cells, even at concentrations sufficient to inactivate or bind all of DNA polymerase I ( Fig. 6 and Table II).
On the other hand, a small fraction of DNA polymerase II from phytohemagglutinin-stimulated blood lymphocytes was bound (Table II) and inactivated (Fig. 6). It is possible that this enzyme was contaminated with a small amount of DNA polymerase I, since the technique of purification of DNA polymerase II from blood lymphocytes was less rigorous than that used with RPM1 1788 cells. At least 60 times as much immune IgG was required to inhibit DNA polymerase II to the same extent as DNA polymerase I (Fig. 6). This is a much greater difference than reported by Chang and Zlollum for an antiserum to calf thymus DNA polymerase I produced in rabbits (39). At approximately the same titers, this rabbit antiserum equally inhibited DNA polymerases I and II from several mammalian species (39). This rabbit antiserum conceivably recognizes antigenic site(s) on the enzymes that are different from the site(s) which our rat antiserum recognizes.
DNA polymerase I and II may contain polypeptide regions in common which our antiserum does not recognize. Neutralization and immunoprecipitation tests in the reciprocal systems (anti-DNA polymerase II against DNA polymerase I) should provide more information on this important point.
We reported elsewhere that anti-DNA polymerase I IgG does not inhibit human DNA polymerase III (33). In those experiments, the incubation of antigens and IgG was brief (10 min). Further studies using more prolonged conditions of incubation are now in progress in order to define a possible serologic relationship between DNA polymerases I and III.
Several DNA polymerases isolated from RNA tumor viruses were inhibited or bound to a small extent by rat anti-DNA polymerase I (Fig. 7 and Table II).
Surprisingly, the RNA-directed DNA polymerase of avian myeloblastosis virus was inhibited ( Fig. 7) or bound (Table II) more than any other viral polymerase tested. In general, however, there were large differences in the concentration of immune IgG required to produce inhibitions of the viral enzymes equal to that of the homologous enzyme; for example, 54 times more immune IgG in the case of the polymerase from avian myeloblastosis virus (Fig. 7). These titrations are essentially the reciprocal of the experiments performed by others @O- 22), and confirm in general that the antigenie relationships are distant between cellular DNA polymerase I and viral RNA-directed DNA polymersses. This specificity is a useful property of our rat anti-DNA polymerase I IgG, which should distinguish other polymerases from DNA polymerase I and detect contamination of RNA-directed DNA polymerases by DNA polymerase I when such enzymes are isolated from human tumor cells.
We have reported the isolation of RNA-directed DNA polymerases from the leukemic blood cells of several patients (15)(16)(17). These enzymes were inactivated by antisera to the DNA polymerases of simian sarcoma and gibbon ape lymphosarcoma viruses (16,17). Two such enzymes were not inhibited by the rat anti-DNA polymerase I IgG. Neither did this IgG inhibit an RNA-directed DNA polymerase isolated from the placentas of rhesus monkeys (42). The activities of both the human leukemic and rhesus placental reverse transcriptases declined during prolonged incubations at 4" to such an extent that incubations of polymerase and immune IgG for less than 1 hour were necessary. However, under these circumstances, large amounts of immune IgG (up to 34 pg) were used so that up to 76% of the activity of DNA polymerase I was neutralized in control assays.