DNA Repair Reactions by Purified HeLa DNA Polymerases and Exonucleases*

Transcription

1 THE JOURNAL OF BIOLOGICAL CHEMISTRY 1988 by The American Society for Biochemistry and Molecular Biology, Inc. Vol. 263, No. 25, Issue of September 5, pp ,1988 Printed in U.S.A. DNA Repair Reactions by Purified HeLa DNA Polymerases and Exonucleases* HAkan RandahlS, George C. Elliott, and Stuart Linn From the Department of Biochemistry, University of California, Berkeley, California 9472 (Received for publication, March 3,1988) PM2 duplex DNA substrates containing small gaps were utilized to study DNA repair reactions of extensively purified HeLa DNase V (a bidirectional double strand DNA exonuclease) and DNA polymerases 8, y (mitochondrial and extramitochondrial), and a holoenzyme, and 6 as a function of ionic strength. At 5 mm NaCl, DNase V carried out extensive exonucleolytic degradation, and &polymerase exhibited strand displacement synthesis. However, at 15 mm NaCl, the DNase appeared only to remove damaged nucleotides from DNA termini while B-polymerase catalyzed only gap-filling synthesis. When present in equimolar amounts, B-polymerase and DNase V (which can be isolated as a 1: 1 complex) catalyzed more degradation than synthesis at 5 mm NaCl; however, at 15 mm NaCl a coupled very limited nick translation reaction ensued. At physiological ionic strength DNA polymerase a holoenzyme was not active upon these substrates. In 15 mm KC1 it could fill small gaps and carry out limited nick translation with undamaged DNA, but it could not create a ligatable substrate from UV-irradi- ated DNA incised with T4 UV endonuclease. Mitochondrial DNA polymerase y was more active at 15 mm NaCl than at lower ionic strengths. It readily filled small gaps but was only marginally capable of stranddisplacement synthesis. The extramitochondrial form of y-polymerase, conversely, was less sensitive to ionic strength; it too easily filled small gaps but was not effective in catalyzing strand displacement synthesis. Finally, DNA polymerase 6 was able to fill gaps of several to 2 nucleotides in.5 M NaCl, but at higher NaCl concentrations there was little activity. DNA polymerases 6 did not demonstrate strand displacement synthesis. Therefore, at physiological ionic strength, it appears that either DNA polymerase B or extramitochondrial DNA polymerase y might aid in short patch DNA repair of nuclear (or transfecting) DNAs, whereas mitochondrial y-polymerase might fill small gaps in mitochondrial DNA. A previous report from this laboratory (1) demonstrated that HeLa DNA could fill small gaps in DNA and then catalyze limited strand displacement synthesis. Moreover, HeLa DNase V, which alone could extensively degrade duplex DNA exonucleolytically, could be coupled with * This work was supported by Grant GM3415 from the National Institutes of Health and by Fellowship 1F32 GM1668 (to G. C. E.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. j Present address: Dept. of Tumor Biology, Karolinska Institute, Box 64, S-141 Stockholm, Sweden DNA to catalyze a nick translation reaction. This latter process required a judicious choice of levels of the two enzymes, however, in order to suppress strand displace- ment synthesis by the polymerase or uncontrolled degradation by the DNase. In addition, the process could be terminated only by the presence of DNA ligase. Studies of the DNA polymerase 8-DNase V coupled reaction using damaged and incised SV4 minichromosomes in place of purified DNA showed a much more regulated reaction; there was little degradation of the chromatin DNA beyond removal of the damage, and synthesis was limited to filling of the excision gap (2). Surprisingly, these latter reactions appeared not to be regulated by the more complex nature of the chromatin substrate but rather by the more physiological ionic strength which was being utilized to keep the chromatin structure undisrupted. The studies in this paper were aimed at learning how these two enzymes might be regulated with ionic strength such that DNase V acts only as an excision exonuclease, DNA does not catalyze strand displacement synthesis, and the two enzymes catalyze a coupled and limited excision/ resynthesis reaction. Similar studies were also undertaken with highly purified preparations from HeLa cells of DNA polymerase a holoenzyme, mitochondrial DNA polymerase y, extramitochondrial DNA polymerase y, and DNA polymerase 6. EXPERIMENTAL PROCEDURES Materials-Unlabeled deoxyribonucleoside triphosphates were from Sigma, and [(Y-~'P]~TI'P, [3H]dTTP, and [met/~yl-~h]thymidine were from Amersham Corp. PM2 DNA (>go% Form I, 7,-1, cpm/nmol) was isolated from phage grown on a thymidine auxotroph or a wild type of strain of Alterornonus espejiana Ba131 (3, 4). The DNA was UV-irradiated at "C with 254-nm light in 1 mm Tris- HCl, ph 7.5,2% glycerol at a fluence of 2 J/mz/s for 3 s. T4 DNA ligase was obtained from Boehringer Mannheim; one unit is the amount required to change 1 nmol of "P from pyrophosphate into Norit-adsorbable material in 2 min at 37 "C(5). T4 UV endonuclease (fraction IV) (6) and Neurospora crassa endonuclease (fraction IX) (7) were prepared as described. Endonuclease activity was measured using a nitrocellulose filter procedure which retains nicked but not supercoiled PM2 [3H]DNA (4). One unit of endonuclease introduces 1 pmol of nicks into Form I DNA/min at 37 'C when the substrate is saturating. HeLa DNA polymerases and y were prepared as described by Krauss and Linn (8) with further purification by chromatography upon hydroxylapatite and DNA cellulose.' HeLa DNA polymerase a holoenzyme was prepared as described by Vishwanatha et al. (9). HeLa DNA polymerase 6 was the hydroxylapatite fraction described by Nishida et al. (1) and was free of other DNA polymerases. Incision of DNA-Incision by T4 UV endonuclease was carried out in 25 mm Tris-HC1, ph 7.5, 2 mm NaC1, 8 ng/ml bovine serum albumin,.3 mm UV-irradiated PM2 DNA (calculated as mononu- ' H. Randahl and S. Linn, unpublished data. Further purification steps based on procedures described by Mosbaugh and Linn (1) are to be published in detail.

2 cleotide), and T4 UV endonuclease. It was assumed that incisions occur randomly among DNA molecules according to a Poisson distribution (4). Incision by Neurospora crassa endonuclease was carried out in 5 mm Tris-HC1, ph 7.5, 1 mm MgC12,.3 mm PM2 DNA (calculated as mononucleotide). It should be noted that the endonuclease cleaves Form I DNA only once under these conditions (ll), so it is assumed that all Form I1 molecules generated have one, and only one, site of incision. Preparation of DNA Containing Larger Gaps-Incised PM2 [3H] DNA was incubated at 37 C with HeLa DNase V in 5 mm Tris- HC1, ph 8.2, mm dithiothreitol,.1 mm EDTA,.5 mg/ml bovine serum albumin, 1 mm NaC1, and.1 mm PM2 DNA for up to 12 min so as to produce the desired mean gap lengths. To determine the amount of 3H label released by the exonuclease, aliquots of about 1 nmol of DNAwere removed, chilled, and mixed with 2 p1 of a solution containing 1.8 mg/ml calf thymus DNA and 25 mg/ml bovine serum albumin. To precipitate the DNA,.2 ml of 1% trichloroacetic acid was added and, after 3 min at C, the samples were centrifuged for 5 min and.2 ml of the supernatant fluid was counted by liquid scintillation. These assays were performed in triplicate and the mean gap length was determined after taking into account the number of nicks present as determined by the endonuclease assay. DNA Synthesis Reactions--Reaction mixture (.1 ml) containing 5 mm Tris-HC1, ph 8.2, 7.5 mm MgCl,, 2 mm dithiothreitol,.1 mm EDTA,.5 mg/ml bovine serum albumin,.5 mm datp, dctp, dgtp, and [32P]dTTP (1-2 cpm/pmol), gapped PM2 [3H] DNA, and the indicated NaCl or KC1 concentration. After incubation at 37 C for the indicated time, reactions were terminated by addition of.2 ml of.1 M sodium pyrophosphate and.7 ml of 1% trichloroacetic acid, collected on Whatman GF/C filters, and washed 2 times with 3 ml of 1 M HCl,.1 M sodium pyrophosphate. Filters were dried and counted by liquid scintillation. For samples that were to be incubated with T4 DNA ligase, polymerase reactions were terminated by incubation at 7 C for 3 min, and then the samples were made.6 mm in ATP, 1 mm in MgCl2, and.5 unit of T4 DNA ligase was added per nmol of DNA. The samples were incubated for another 3 min at 37 C, and then the reaction was terminated by the addition of EDTA to 4 mm. Alkaline Sucrose Gradient Sedimentation-To samples from ligase reactions were added.5 volume of.3 M Tris-HC1, ph 8.,.25 M HeLa DNA Polymerases C NaCl (mm) Time (min) 1 2 Time (min) FIG. 1. Effect of NaCl concentration upon activities of HeLa DNA polymerase fl and DNase V. A, PM2 [3H]DNA was treated with N. crassa endonuclease to.65 nicks/genome, and then 5.5-nmol aliquots of the DNA were assayed with.12 unit at the NaCl concentrations indicated. Blanks contained polymerase but were immediately quenched by stopping buffer. Incubation of the NaOH,.9 M NaCl,.5 M EDTA, and then the samples were samples was for 6 min at 37 C, and then the reactions were termilayered onto a 5-25% linear sucrose gradient in the above buffer. nated by addition of stopping buffer and the samples filtered as After centrifugation in an SW 5.1 rotor at 43, rpm for 9 min at described under Experimental Procedures. B, PM2 [3H]DNA was 2 C, fractions of.18 ml were collected from the bottom of the tube, UV-irradiated at 2 J/m2/s for 4 s and incised by T4 UV endonuclease and each fraction was adjusted to.5 mg/ml calf thymus DNA and to yield an average of one nick/genome, and then aliquots of 45 nmol 7% trichloroacetic acid. Precipitates were collected on Whatman GF/ were treated with.7 unit of DNase V at various NaCl concentrations C filters and washed 2 times with 3 ml of 1 M HC1,.1 M sodium as described under Experimental Procedures. Duplicate samples pyrophosphate, and then the filters were dried and counted by liquid were taken at the times indicated and monitored for nucleotides scintillation. released. H,.5 M NaCI; *,.1 M NaCI; D,.15 M NaCl. C, PM2 [3H]DNA(27 nmol) was UV-irradiated at a dose rate of 2 J/m2/s RESULTS Effect of Ionic Strength on the Activities of DNA Polymerase for 4 s and incised by T4 UV endonuclease to an average of 1.5 nicks/genome. The DNA was divided into 6 aliquots and added to three pairs of reaction mixtures with the indicated NaCl concentra- P and DNase V-When Form I DNA is treated with N. crassa tion. The polymerase and nuclease reaction mixtures contained [32P] endonuclease, Form I1 molecules are produced, each of which dttp or unlabeled dttp, respectively, in addition to unlabeled catp, dctp, and dgtp. DNA polymerase (8.4 units) and DNase appears to contain a gap of one or two molecules. Since these V (5 units) were mixed and incubated on ice for 1 min before gaps contain 3 -OH primer termini, they provide suitable addition to each sample. Incubation was at 37 C and, at the indicated substrates for studying gap filling and strand displacement times, aliquots were taken in duplicate and the extents of reaction reactions, and such molecules were therefore utilized to study determined as described under Experimental Procedures. A, A,.5 the activity of DNA polymerase p at different concentrations M NaCI; e,,o.lo M NaC1; W, Q.15 M NaCl. of NaCl (Fig. IA). At.5 M NaC1, the polymerase catalyzed strand displace- bose-!%phosphate formed by reaction) at the ment at a linear rate for at least 6 min, incorporating terminus and a pyrimidine dimer nucleotide at the 5 dtmp residues (roughly 12 total nucleotides)/gap. Increased terminus (12). salt concentrations suppressed the strand displacement activity, and at.15 M NaC1, only a total of 1-2 nucleotides could be incorporated per gap. Hence at this physiological ionic Y Y B4 strength, P-polymerase appears only to catalyze gap filling. To study the effect of ionic strength upon DNase V, PM2 DNA was UV-irradiated and then treated with phase T4 UV...P$ pj pjp$ P... endonuclease. The treatment resulted in a population with This substrate was then exposed to DNase V at several NaCl approximately 1 nick/molecule which contained a baseless concentrations (Fig. 1B). With 5 mm NaCl in the reaction unsaturated sugar (presumably 2,3-didehydro-2,3-dideoxyri- mixture, degradation was extensive; 33 dtmp residues were 1 4 Y

3 1223 HeLa DNA Polymerases released in 2 h, equivalent to more than 1 total nucleotides released. A sharp decline in activity was observed for increased NaCl concentrations, however, and with.15 M NaCl, only 2 thymine residues were released as a limit of the reaction. Within the limits of detectability of this protocol, this level is consistent with the removal only of the pyrimidine dimer nucleotide (and possibly the baseless sugar), but no undamaged nucleotides from the DNA. Hence, at physiological ionic strength, DNase V appears only to catalyze excision of damage from DNA termini. DNase V and DNA polymerase P can be isolated as an equimolar complex (l), suggesting that these two enzymes might function jointly to effect excision and repair synthesis in uiuo. In previous studies, such a reaction could be elicited in uncontrolled reactions in vitro, but no evidence was ob- tained for the enzymes acting in a regulated, complexed manner (1). Since ionic strength had such a profound effect upon the enzymes individually (Fig. 1, A and B), the action of the two enzymes together was studied upon the incised UVirradiated substrate at several NaCl concentrations (Fig. IC). At.5 M NaC1, when the enzymes were added at what were estimated to be equimolar amounts, the DNase V excised nucleotides at a higher rate than P-polymerase incorporated them. However, at.1 or.15 M NaCl, the two enzymes seemed to work in concert, excising and incorporating equal numbers of nucleotides in an apparently coupled manner. Furthermore, beyond simple gap filling or excision of damage as is seen for the enzymes individually, a slow nick translation reaction ensued in which each enzyme appeared to enable the other to go beyond the limited reaction. The nick translation reaction was further studied by the subsequent addition of DNA ligase in order to test the extent of coupling of the polymerase and nuclease reactions (Fig. 2). In this case, PM2 DNA was treated with N. crassa DNase so that all molecules were gapped, and then these molecules were incubated with the P-polymerase and DNase V. After heating to 7 "C, these were finally treated with T4 DNA ligase. Sixty percent of the PM2 DNA molecules was sealed by the ligase, and these sealed molecules contained.9 [32P]dTMP residues (roughly 3.5 total nucleotides) incorporated by the polymerase. The unsealed molecules contained roughly.6 [32P]dTMP residues each. By these criteria, the polymerase and DNase reactions are coupled to within 1 or 2 nucleotides in their joint action with the nuclease, possibly leading the polymerase reaction. Action of Mitochondrial DNA Polymerase y-dna polymerase y is reported to have maximal activity at M KC1 (13). To determine the effect of salt with a gapped substrate, nicks were introduced into UV-irradiated PM2 [3H] DNA with the T4 UV endonuclease, and then these nicks were extended by DNase V in low salt so as to form gaps of roughly 58 nucleotides. This DNA was used as substrate for y-polymerase at various NaCl concentrations (Fig. 3A). As opposed to DNA polymerase 6, the y-polymerase was more active with increasing NaCl concentration, though the differences were not so great over the range tested. To investigate whether DNA polymerase y could utilize nicks or small gaps, UV-irradiated DNA was treated with T4 UV endonuclease and then preincubated with DNase V in.15 M NaCl so as to excise only the pyrimidine dimer. At.15 M NaCl the DNA polymerase y filled the small gaps slowly and about 3.5 [32P]dTMP residues were inserted per gap after 18 min, possibly indicating limited strand displacement synthesis (Fig. 3B, squares). Similar observations were made with DNA containing small gaps created by the N. crassa nuclease. loo zoo FIG. 2. Analysis by alkaline sucrose gradient sedimentation of DNA polymerase B plus DNase V reaction products after ligation. PM2 [3H]DNA was treated with N. crmsa endonuclease to produce molecules containing 1 gap, and DNA synthesis reaction mixtures were prepared containing 24 nmol of this DNA, 3.1 units of &polymerase, 1.3 units of DNase V, and [32P]dTTP. One sample was heated immediately for 3 min at 7 "C, made.6 mm in ATP, and then 1 units of T4 DNA ligase were added and the mixture was incubated for 3 min at 37 "C. This sample did not incorporate significant amounts of [32P]dTTP but was 56% Form I. (T4 DNA ligase is able to seal one-nucleotide gaps (l).) Two other samples were incubated with the polymerase for 2 h at 37 'C and then heated for 3 min at 7 "C and made.6 mm in ATP. To one was added 1 units of T4 DNA ligase (El) and to the other was added buffer (+), and incubation was continued for 3 min at 37 "C. All samples were then applied to 5-25% alkaline sucrose gradients and treated as described under "Experimental Procedures." The specific activities of the [3H] DNA and [32P]dTTP were 97 and 425 cpm/pmol, respectively. When DNase V was present with the DNA polymerase y and the substrate nicked by the UV endonuclease, there was a lag before any polymerization occurred (Fig. 3B, open triangles). This lag was shorter at lower concentrations of NaCl in which the DNase V was more active. As opposed to the combination of DNA polymerase P and DNase V, which, when added together, appeared to stimulate one another (Fig. lc), DNase V and y-polymerase do not appear to stimulate one another (Fig. 3B, triangles). At.15 M NaCl there was no difference in apparent activity of DNase V whether y-polymerase was present or not, and the polymerase activity lagged behind the nuclease. This difference was more pronounced at lower salt concentrations, since DNase V has increased activity whereas y-polymerase has decreased activity in lower salt. In a final experiment, PM2 [3H]DNA was treated with N. crassa nuclease and then with DNA polymerase y. After inactivating the polymerase, the samples were treated with T4 DNA ligase and analyzed by alkaline sucrose sedimenta-

4 3 A HeLa DNA Polymerases , I Time [minl FIG. 3. Utilization of gaps in PM2 [%]DNA by mitochondrial DNA polymerase y. PM2 [3H]DNA was irradiated with UV light at 2 J/mZ/s for 4 a and incised by T4 UV endonuclease to an average of.7 nicks/genome. A, the nicked DNA was treated with DNase V to form gaps of roughly 58 nucleotides and then exposed to DNA polymerase y in.5 M NaCl (A),.1 M NaCl (+), and.15 M NaCl (El). All measurements were in duplicate and utilized.3 unit of polymerase/assay. B, incised UV-irradiated PM2 [3H]DNA was treated with DNase V in.15 M NaCl to excise only the pyrimidine dimer and then added to a polymerase reaction mixture containing.15 M NaCl and.3 unit of y-polymerase/sample (El). Incised UVirradiated PM2 [3H]DNA was added to reaction mixtures containing.15 M NaC1,.3 unit of y-polymerase, and.5 unit DNase V/sample containing unlabeled (A) or 32P-labeled dttp (A) and monitored for nuclease (A) or polymerase (A) activity. Allpoints represent duplicate assays. The specific activities of the [3H]DNA and [32P]dTTP were 97 and 636 cpm/pmol, respectively. tion. Initially, all of the PM2 DNA was nicked by the nuclease and after polymerization and ligation, 41% of the DNA migrated as Form I (Fig. 4). The mean of incorporation of ["PI dtmp in the Form I DNA was.3 [32P]dTMP/molecule, as expected from the filling of gaps of one or two nucleotides. The DNA in the Form I1 region had roughly the same degree of incorporation, except at the trailing edge where 32P may have been incorporated into some DNA fragments. It therefore appears that the DNA polymerase y can fill small gaps though it is not apparent why the ligation efficiencywas somewhat low. Action of the Extramitochondrial DNA Polymerase y-the filling of small gaps in DNA treated with the N. crassa nuclease by this polymerase does not vary much between.5 and.15 M NaCl (Fig. 5). About.8 dtmp residues were incorporated per gap at the initial rate. The time course of the reaction (Fig. 5) also shows a subsequent slow phase in which several more nucleotides are incorporated over a 3-h period, presumably by strand displacement synthesis. When the gapped DNA was treated with the polymerase, then with T4 DNA ligase, and finally analyzed by alkaline sucrose gradient sedimentation, 46% of the DNA sedimented as Form I (Fig. 6). The peak fraction of Form I DNA contained.7 [32P]dTMP residues/molecule, as expected for filling gaps of roughly 2-3 nucleotides. The DNA in the Form I1 region had roughly the same degree of incorporation, except at the trailing edge where 32P again may have been incorporated into some DNA fragments. In summary, the activity of the extra-.o FIG. 4. Analysis by alkaline sucrose gradient sedimentation of mitochondrial DNA polymerase y reaction products after ligation. PM2 [3H]DNA was treated with N. crassa endonuclease followed by 1 unit of DNA polymerase y per 45-nmol aliquot. The remaining protocol was as in Fig. 2. El, with DNA ligase; +, without ligase. The specific activities of the [3H]DNA and [32P]dTTP were 97 and 625 cpm/pmol, respectively. Time [minl FIG. 5. Utilization of PM2 DNA treated with N. crassa nuclease by the extramitochondrial DNA polymerase 7. N. crassa endonuclease was utilized to nick 125 nmol of PM2 [3H]DNA to.8 scissions/genome. The DNA was divided in three aliquots and added to a reaction mixture containing the indicated NaCl concentrations. DNA polymerase (2 units) was added to each aliquot, and the reactions were incubated at 37 "C. At the indicated times, duplicate.1-ml aliquots were monitored for nucleotide incorporation. W,.5 M NaC1; +,.1 M NaC1; El,.15 M NaCl. The specific activity of the [32P]dTTP was 428 cpm/pmol. mitochondrial DNA polymerase y does not differ significantly from that for the mitochondrial form of the enzyme. Action of DNA Polymerase &-DNA polymerase 6 is an a- like polymerase that has an intrinsic 3' to 5' exonuclease activity. When tested on DNA containing gaps of 2 nucleotides, the enzyme was very sensitive to ionic strength; complete gap filling was approached only at low concentrations of NaCl (Fig. 7A).

5 HeLa DNA Polymerases 8 A I B ( FIG. 6. Analysis by alkaline sucrose gradient sedimentation of extramitochondrial DNA polymerase y reaction product after ligation. The experiment was performed as in Fig. 2, except that 4.95 units of extramitochondrial DNA polymerase y were used. El, with DNA ligase; e, without ligase. The specific activities of the [3H]DNA and [32P]dTTP were 97 and 425 cpm/pmol, respectively. PM2 [3H]DNA was also incised with N. crassa endonuclease and used as substrate for &polymerase (Fig. 7B). At 1-15 mm NaCl, incorporation was barely detectable and then only after prolonged incubation. At.5 M NaC1, somewhat less than 1 dtmp was incorporated per nick. Thus the enzyme appeared to fill small gaps only in low salt concentrations and did not appear to catalyze strand displacement synthesis. DNA nicked by N. crassa endonuclease was treated with DNA polymerase 6 and T4 DNA ligase, and the products were analyzed by alkaline sucrose gradient sedimentation. When the DNA polymerase 6 reaction was carried out in.15 M NaCl, very little of Form I DNA was found, presumably due to a lack of activity of the &polymerase on this substrate. With.5 M NaCl present in the polymerase reaction, about 37% of the DNA sedimented as Form I (Fig. 8), and this DNA contained roughly 1 [32P]dTMP residue/molecule. Thus, DNA polymerase 6 can fill small gaps only at hypotonic concentrations of NaC1. Action of DNA Polymerase a Holoenzyme-In a previous study with purified DNA polymerase a catalytic subunit (14), it was found that gaps of up to 65 nucleotides could be reduced to about 15-nucleotides in length, following which the DNA became refractory to further synthesis. To test whether accessory factors could confer upon this polymerase the ability to totally fill gaps, a large holoenzyme complex of a-polym- erase was prepared according to the procedure of Vishwanatha et al. (9). This complex contained and exonuclease activities as well as other peptides as described by those authors. At physiological ionic strength, the holoenzyme has little 1 Time [min] FIG. 7. Utilization of gaps in PM2 [%]DNA by DNA polymerase 8. A, PM2 [3H]DNA was treated with UV light at a dose rate of 2 J/m2/s for 4 s and incised by T4 UV endonuclease to 1.3 nicks/ genome. The nicked DNA was treated with DNase V at 5 mm to form gaps averaging 2 nucleotides. The DNA (11.5 nmol) was then used as a substrate for 1.8 units of DNA polymerase 6 in.6-ml reactions containing.5 M NaCl (A),.1 M NaCl (e), or.15 M NaCl (El). At the indicated times, duplicate.1-ml aliquots were assayed for nucleotides incorporated. B, PM2 [3H]DNA was incised by N. crmsa endonuclease to.53 gaps/genome, and nucleotide in- corporation was measured as above. The specific activity of the 32P[dTTP] was 139 cpm/pmol. activity. However, in 15 mmkc1, the enzyme could convert PM2 DNA treated with N. crassa endonuclease to a form which could be sealed by DNA ligase (Fig. 9). In this case, 1 total nucleotides were incorporated per sealed nick in Form I DNA, and roughly 8 nucleotides/nick were found in the unsealed Form I1 DNA. This extent of incorporation implies that the exonuclease activity in the holoenzyme preparation might be able to serve in a nick translation type of reaction with the polymerase. In fact, some nucleotides are removed from Form I1 DNA by the holoenzyme preparation. This nuclease does not appear to have a DNA excision repair capability, however, since UV-irradiated DNA treated with T4 endonuclease was not a substrate for the holoenzyme polymerase activity (and was not sealable by DNA ligase after polymerase treatment). As a final caveat, in the presence of the four dntps, the holoenzyme preparation could seal the filled gaps initially made by the N. crassa endonuclease without the addition of exogeneous DNA ligase (Fig. 9). (DNA treated with UV light, T4 UV endonuclease, and holoenzyme was not a substrate for this endogenous DNA ligase.) Whether this ligase is part of the holoenzyme complex, or indeed, whether the capability to fill small gaps is conferred by a peptide of the holoenzyme complex, awaits further characterization of that complex, especially at very low ionic strengths. What can be concluded is that some factor or factors exists in HeLa cells which, at

6 HeLa DNA Polymerases OOC 4C 2c 5 I ( 8OC t 6C 4C 2c C FIG. 8. Analysis by alkaline sucrose gradient sedimentation of DNA polymerase 6 reaction product after ligation. The experiment was performed as in Fig. 2, except that 1 units of DNA polymerase 6 were used. D, with DNA ligase; 4, without DNA ligase. The specific activities of [3H]DNA and [3ZP]dTTP were 97 and 547 cpm/pmol, respectively. least at hypotonic salt concentrations, allow(s) a-polymerase to fill small gaps. DISCUSSION Our laboratory has pursued several studies directed at methodically studying DNA excision repair with well characterized DNA or chromatin substrates and highly purified nucleases and polymerases from HeLa cells (1,2,14). This report emphasizes the unanticipated degree of importance for controlling such a simple parameter as ionic strength and points out the profound qualitative effects alteration of such a parameter might have. The most impressive observations were with DNA polymerase p and DNase V. At isotonic.15 M NaC1, these enzymes acted in a controlled physiologically sensible manner, either independently or in a concerted nick translation reaction. In hypotonic conditions, where these enzymes are most active and are normally characterized, they act without apparent control in a nonphysiological manner. This report examines the activities of all of the known HeLa DNA polymerases in gap-filling clearly is the most proficient at filling gaps, and at physiological ionic strength its strand displacement reaction is retarded. The y-polymerases also appear to be capable of filling small gaps in.15 M NaC1, though not in lower salt concentrations. DNA polymerase LY catalytic subunit can add nucleotides to large gaps but leaves a gap of roughly 15 nucleotides as a final product (14). However, a-polymerase holoenzyme, prepared as described by Vishwanatha et al. (9), is capable of forming ligatable substrates from small gaps, but only at very low salt concentrations. DNA polymerase d also FIG. 9. Analysis by alkaline sucrose gradient sedimentation of DNA polymerase a holoenzyme reaction product after ligation. The experiment was performed as in Fig. 2 with 1 nmol of PM2 DNA, 2 units of DNA polymerase, and T4., DNA ligase (incubated for 6 min) where indicated. Only 75% of the PM2 DNA was nicked after treatment with N. crassa nuclease. D, DNA nicked with N. crassa nuclease and treated with polymerase and T4 DNA ligase; 4, nicked DNA treated with T4 DNA ligase only; un-nicked DNA treated with T4 DNA ligase only. The specific activities of the [3H] DNA and [32P]dTTP were 344 and 31 cpm/pmol, respectively. The 32P]dTTP] appeared to be contaminated with some acid-insoluble material which remained near the top of the gradient. appears to be able to fill gaps, but not efficiently and only at hypotonic NaCl concentrations. Hence, as in many other respects, &polymerase resembles, but is not identical to, a- polymerase. How might these observations relate to the functioning of these enzymes in uiuo? Under isotonic conditions, DNA polymerase p or extramitochondrial y-polymerase, but not a- or &polymerases might aid in gap-filling DNA repair processes of nuclear or transfecting DNAs. Moreover, small gaps in mitochondrial DNA might be filled by the mitochondrial y- polymerase.* Uncontrolled strand displacement synthesis, which might induce recombination or duplication events, would appear most likely to be due to,&polymerase. Gap filling in the absence of accessory fidelity factors by 8- or possibly y-polymerase might contribute to the high rate of mutagenesis observed with transfecting DNAs in mammalian cells (15). This information, coupled with fidelity studies, might therefore help to understand the roles of the various DNA polymerases and their accessory factors in both normal and abnormal DNA metabolic events in mammalian cells. Ackmwkdgrnents-We thank Dr. Craig Nishida for providing 6- polymerase, Roberta Johnson for her expert help with tissue culture, and Professor Earl Baril for his patient advice for purifying a- polymerase holoenzyme. REFERENCES 1. Mosbaugh, D. W., and Linn, S. (1983) J. Biol. Chern. 258, While long-patch nucleotide excision DNA repair appears not to occur in mitochondria, our laboratory has detected and purified distinct mitochondrial forms of DNA glycosylases and AP endonucleases which might elicit short-patch base excision DNA repair (A. E. Tomkinson, T. Bonk, N. Bartfeld, and S. Linn, unpublished data).

7 12234 HeLa DNA Polymerases 2. Evans, D. H., and Linn, S. (1984) J. Biol. Chem. 259, Vishwanatha, J. K., Coughlin, S. A., Wesolowski-Owen,M., and 1259 Baril, E. F. (1986) J. Biol. Chem. 261, Espejo, R. T., and Canelo, E. S. (1968) Virology 34, Nishida, c., Reinhard, p., and Linn, s. (1988) J. Bwl. Chem. 4. Kuhnlein, U., Penhoet, E. E., and Linn, S. (1976) Proc. Natl. 263,51-51 Acad. Sci. U. S. A. 73, Kato, A. C., Bartok, K., Fraser, M. J., and Denhardt, D. T. (1973) 5. Weiss, B., Jacquemin-Sablon, A., Live, T. R., Fareed, G-C., and Biochim. Bwphys. Acta 38, Richardson, C. C. (1968) J. Biol. Chem. 243, Kim, J., and Linn, S. (1988) Nucleic Acids Res. 16, Knopf, K.-W., Yamada, M., and Weissbach, A. (1976) Biochem- 6. Friedberg, E. C., and King, J. J. (1971) J. Bacteriol. 16: 5- istry 15, Mosbaugh, D. W., and Linn, S. (1984) J. Biol. Chem. 269, Linn, S., and Lehman, I. R. (1965) J. Biol. Chem. 24, Razzaque, A., Chakrabarti, S., Joffee, S., and Seidman, M. (1984) 8. Krauss, S. W., and Linn, S. (1988) J. Cell. Physiol. 126, Mol. Cell. Biol. 4,