Comparison of DNA replication and repair enzymology using permeabilized baby hamster kidney cells.

The roles of DNA polymerases in replication and repair were examined in lysolecithin-permeabilized baby hamster kidney cells. Lysolecithin treatment yields cell preparations that are capable of normal replicative DNA synthesis at in uiuo rates, when supplied with deoxyribonucleoside triphosphates (dNTPs) (Miller et al. (1978) Biochemistry 17, 1073-1090; Miller et al. (1979) Exp. Cell Res., 120, 421-425). The K’,,, of dNTPs for replication is 50 pM. Evidence presented here shows that permeabilized Go cells carry out DNA repair synthesis after exposure to bleomycin. The K’, of dNTPs for repair is 160 FM. The sensitivity of replication and repair in permeabilized cells to known inhibitors of DNA polymerases (Y and /I in cell-free extracts was investigated. Replication synthesis was strongly inhibited by cytosine arabinoside triphosphate, high KCl, and N-ethylmaleimide, as is DNA polymerase (Y in extracts. Repair synthesis was relatively insensitive to these agents, as is DNA polymerase p activity in extracts. Thus replication and repair synthesis can be clearly distinguished by these inhibitors. These differences in inhibition, K’,, and cell cycle occurrence indicate that DNA polymerase LY is the major replication polymerase, whereas DNA polymerase /I is the major repair polymerase. Ribonucleoside &phosphates serve well as precursors for DNA replication and repair in lysolecithinpermeabilized cells, indicating that ribonucleotide reductase is highly active. Hydroxyurea inhibits replication and ribonucleotide reductase. But hydroxyurea does not inhibit repair, except partially at very low concentrations of substrates. This result suggests that noninhibition of repair by hydroxyurea has a quantitative basis-the very low requirement of dNTPs for repair.

The roles of DNA polymerases in replication and repair were examined in lysolecithin-permeabilized baby hamster kidney cells. Lysolecithin treatment yields cell preparations that are capable of normal replicative DNA synthesis at in uiuo rates, when supplied with deoxyribonucleoside triphosphates (dNTPs)  Biochemistry 17, 1073-1090; Miller et al. (1979) Exp. Cell Res., 120, 421-425).
The K',,, of dNTPs for replication is 50 pM. Evidence presented here shows that permeabilized Go cells carry out DNA repair synthesis after exposure to bleomycin. The K', of dNTPs for repair is 160 FM.
The sensitivity of replication and repair in permeabilized cells to known inhibitors of DNA polymerases (Y and /I in cell-free extracts was investigated. Replication synthesis was strongly inhibited by cytosine arabinoside triphosphate, high KCl, and N-ethylmaleimide, as is DNA polymerase (Y in extracts. Repair synthesis was relatively insensitive to these agents, as is DNA polymerase p activity in extracts. Thus replication and repair synthesis can be clearly distinguished by these inhibitors.
These differences in inhibition, K',, and cell cycle occurrence indicate that DNA polymerase LY is the major replication polymerase, whereas DNA polymerase /I is the major repair polymerase.
Ribonucleoside &phosphates serve well as precursors for DNA replication and repair in lysolecithinpermeabilized cells, indicating that ribonucleotide reductase is highly active. Hydroxyurea inhibits replication and ribonucleotide reductase. But hydroxyurea does not inhibit repair, except partially at very low concentrations of substrates. This result suggests that noninhibition of repair by hydroxyurea has a quantitative basis-the very low requirement of dNTPs for repair.
Animal cells contain at least three DNA polymerases: (Y, ,L?, and y. The (Y and /? enzymes have been highly purified, and the y enzyme has been partially purified (1). The mitochondrial DNA polymerase is probably a form of the y enzyme (2). The three DNA polymerases can be distinguished from one another by their sensitivity to sulfhydryl group blockers and salt concentration, molecular weight, and template preferences (1,3).
The functions of the DNA polymerases have not been defined clearly. Which polymerases act in replication and which act in repair synthesis is suggested from the correlation of activity with cell cycle phase. Several laboratories report that the predominant activity in proliferating cells is DNA polymerase (Y, whereas the major activity in quiescent cells is DNA polymerase p (4-S). Bertazzoni et al. (9) found that UVirradiated lymphocytes have increased DNA repair and higher levels of DNA polymerase p than proliferating lymphocytes. These observations indirectly suggest that the (Y enzyme is important in replication and that the /3 enzyme is involved in repair. DNA polymerase y is a minor enzyme. We have examined roles of the DNA polymerases using lysolecithin'-permeabilized BHK cells. This technique yields a cell preparation which can carry out normal, replicative DNA synthesis at i:-~:lvo Yates when supplied with deoxyribonucleoside triphosphates (10, 11). We show here that repair synthesis also can be measured in cells permeabilized with lysolecithin.
The different sensitivities of replication and repair synthesis to various inhibitors indicate that DNA polymerase (Y plays a rate-limiting role in replication, whereas DNA polymerase ,8 plays a major role in repair. The role of hydroxyurea in inhibiting DNA replication but not repair is also considered. Roles of Mammalian DNA Polymerases in Replication and Repair 6905 mM 4-(2.hydroxyethyl)-1-piperazineethanesulfonic acid (pH 7.4), 50 mM sucrose, 80 mu KCl, 4 mM MgC12, 7.5 mu potassium phosphate (pH 7.4), and 1.0 mM CaClz (Solution A). Lysolecithin in 2 ml of Solution A was added to the cells for 2 min at O"C, at a concentration dependent on: 1) the density of the culture, 2) the degree of permeability desired, and 3) the cell type. These factors have been discussed in detail previously (10,11 were washed five times with cold 5% trichloroacetic acid, then dissolved in 1.0 ml of 0.1 N NaOH for 30 min at 37"C, neutralized with 0.1 ml of 1.0 N HCl, and the entire sample was counted in a liquid scintillation counter with 10 ml of Biofluor.

Replication
of DNA in Permeable Cells--We have previously shown that DNA synthesis in exponential cultures of lysolecithin-permeabilized cells proceeds in a normal replicative fashion; i.e., it is semiconservative, discontinuous, cell cycle-dependent, and occurs at in vivo rates (10, 11). This in situ system was further utilized to examine the DNA polymerase activities in BHK cells. The K', of dNTPs was determined by varying the concentration of all four dNTPs simultaneously over a wide range and measuring incorporation of ["HITTP into DNA of permeabilized exponential cells. Incorporation from 30 s to 20 min of incubation using 0.25 IIIM dNTPs showed that extrapolating to zero time yields very close to zero incorporation and that incorporation is linear (10). The K', for dNTPs was calculated from 15-min incubations.
From the specific activity of the assay mix, we calculated the picomoles of dNMPs incorporated into DNA/ unit of time. A Lineweaver-Burk plot of a typical experiment is presented in Fig. 1 cells grown in suspension (10) and with BHK cells made permeable using a high NaCl concentration (13). We next examined the sensitivity of DNA replication to inhibitors that affect mammalian DNA polymerases in extracts differently.
The cells were permeabilized and then incubated in the DNA synthesis assay solution containing different concentrations of KC1 or N-ethylmaleimide (1). Fig.  2A shows that replication activity is inhibited >80% at concentrations of KC1 above 0.3 M. NaCl at 0.08 M inhibited replication by 50% (not shown). Fig. 2B shows that N-ethylmaleimide concentrations above 0.3 mM inhibit DNA replication by >90%. We have previously reported that 20 pM cytosine arabinoside triphosphate inhibited DNA synthesis by 90%, as did 1.0 IIIM hydroxyurea when rNDPs were supplied as precursors to permeable cells (10, 13). The preferential inhibition of DNA polymerase (Y by Ara-CTP is well documented (see ref. 14 for a review).
DNA Repair in Permeable Cells-To measure repair synthesis, BHK cells were first arrested in Go by placing them in medium containing 0.2% calf serum for 48 h. This greatly reduced replication synthesis which otherwise would obscure the much lower repair activity. The low serum medium was replaced with medium containing 10% calf serum for 1 to 2 h prior to permeabilization to ensure that the cells were as physiologically healthy as possible. Flow microfluorimetry showed virtually none of these cells in S phase. The undamaged cells incorporated very little ["HITTP-approximately 0.1% of the replication rate (Fig. 3). DNA was damaged by exposing these cultures to x-rays or UV light prior to permeabilization.
Other cultures were permeabilized and incubated in the DNA synthesis assay solution containing bleomycin. Bleomycin is a basic glycopeptide that binds to DNA and causes nicking and strand scission. DNA repair follows any of these treatments (see ref. 15 and 16 for reviews).
Permeable BHK cells were shown to carry out repair synthesis after x-ray treatment. BHK cells arrested in Go were  3. Unscheduled DNA synthesis in permeable BHK cells. BHK cells were plated in 60-mm culture dishes at 5 x lo" cells/dish. After 24 h the cells were put in Dulbecco's medium containing 0.2% calf serum for an additional 48 to 60 h. A, Go cells were permeabilized and exposed to the DNA synthesis solution containing bleomycin at the indicated doses. M, 0.05 mM dNTPs; M, 0.01 mM dNTPs. B, GO cells were exposed to the indicated dose of x-rays 30 min prior to permeabilization. The DNA synthesis solution contained 0.05 mM dNTPs.
All incubations were for 20 min at 37°C. 20 pCi/ml[JH]TTP was used in each experiment.
exposed to x-rays 30 min prior to permeabilization with lysolecithin. There was a 1.5 to 2-fold increase in [3H]TTP incorporation relative to unirradiated permeable Go cells (Fig.  3B).
Irradiation with UV (1200 pW/cm' for up to 5 min) did not increase E3H]TTP uptake appreciably.
Bleomycin produced an increase in ['H]TTP incorporation, which was dose-dependent up to 10 pg/ml (Fig. 3A). Maximum repair after bleomycin treatment (Fig. 3) occurred at about 1% of the replication rate in an exponential culture (lo), about a lo-fold increase in ["HITTP incorporation relative to unirradiated permeable Go cells. Synthesis was linear for at least 15 min. In contrast to replication, omitting GTP, CTP, and UTP from the DNA synthesis assay mixture did not affect the rate or amount of DNA synthesis observed after bleomytin treatment, so these rNTPs were omitted from subsequent DNA repair experiments.
The DNA synthesized after bleomycin treatment of Chinese hamster ovary cells was analyzed on benzoylated naphthoylated DEAE-cellulose columns using the method of Scudiero et al. (17). Double-stranded (or native) DNA will not bind to the column and washes through with 1 M NaCl. Singlestranded DNA will bind tightly to the column and does not elute with 1 M NaCl. [3H]TTP incorporated into repaired DNA is found in double stranded DNA and elutes with 1 M NaCl, whereas [3H]TTP incorporated into replicating DNA is single-stranded and remains bound to the column during a 1 M NaCl wash. This method has been shown to yield results identical to the BUdR density-labeling procedure of Hanawalt and Cooper (18). Following this procedure, we found that 85% of the increase in radioactivity after bleomycin treatment was eluted with 1 M NaCl, indicating it is in repaired DNA. BHK cells were arrested i n GO and permeabilized as described under "Experimental Procedures." They were incubated with 20 pg/ ml of bleomycin and with the indicated concentrations of either dNTPs or rNDPs plus dTTP (20 pCi of "H/ml). After 20 min the reaction was stopped, and the cells were processed for scintillation counting as described under "Experimental Procedures." The amount of incorporation by permeabilized GO cells not incubated with bleomycin was subtracted from each point. where the subscripts denote to which member of the concentration pair the reaction velocity (u) and substrate concentration (S) refer. The K', value obtained for all four dNTPs was 0.170 mM + 0.05 mM. This is substantially higher than the 0.05 mM K', found in replication synthesis (Fig. 1). To further distinguish this unscheduled DNA synthesis from replication activity, the sensitivity of bleomycin-induced synthesis to KC1 and N-ethylmaleimide was examined. Bleomycin-induced DNA synthesis was much less sensitive than replication to both high KC1 ( Fig. 2A) and N-ethylmaleimide (Fig. 2B). Furthermore, 20 pM Ara-CTP, a concentration which inhibits replication in permeable cells by 90% (10, 13), reduced bleomycin-induced synthesis only 20%. Eliminating ATP and phosphoenolpyruvate from the assay mixture inhibited bleomycin-induced synthesis by 10 to 30% in different experiments; the absence of these energy sources reduced replication synthesis by >80% (10, 13). The rNDPs can support repair synthesis in permeable cells. Using ADP, GDP, and CDP with ["H]TTP as the DNA synthesis precursors consistently gave 80 to 90% as much repair synthesis as the same concentrations of all four dNTPs (Table I). Thus reductase activity appears to be in great excess relative to repair polymerase activity. The presence of hydroxyurea (up to 5 XnM) did not inhibit repair synthesis when dNTPs were used at any concentration, nor until the exogenously supplied concentration of rNDPs dropped below 5 pM. Even the combination of very low (0.5 pM) rNDPs and 5 mM HU only reduced repair synthesis about one-third relative to the rate with dNTPs.

DISCUSSION
It is well documented that DNA replication and repair in animal cells are very different processes (19,20), and our data are consistent with these observations. A major purpose of the experiments reported here was to assign functions to the DNA Roles of Mammalian DNA Polymerases in Replication and Repair Yes polymerases cx and /3. Many properties of these enzymes have been well studied in cell extracts, and their activities with respect to cell cycle phase have been examined using synchronized intact cells (see Ref. 1). To bridge the gap between enzymes and their roles in intact cells, we have used lysolecithin-permeabilized BHK cells. This subcellular yet highly physiological cell preparation carries out normal, replicative DNA synthesis (10, 11) as well as repair synthesis (Fig. 3).
Bleomycin-induced ["HITTP incorporation was shown to be predominantly (>80%) repair by the following criteria, which distinguish it from replication: 1) It is unscheduled synthesis; i.e. it occurs in non-S phase cells (Fig. 3). 2) [3H]-TTP incorporation reached a plateau at higher bleomycin concentrations; saturation at high levels of DNA damage is characteristic of repair (17,21). 3) It is not dependent on ATP. 4) It elutes from BND-cellulose columns in the manner expected of repaired DNA.
A small amount (10 to 20%) of the bleomycin-induced ["H]TTP incorporation does not elute in the 1 M NaCl repair fraction. This could be due to a background of replication carried out by the few cells not arrested in Go or to the induction of a small amount of replication synthesis by bleomycin in Go cells. The presence of a small amount of replication synthesis is consistent with the 10 to 30% inhibition of bleomycin-induced ["HITTP incorporation produced by Nethylmaleimide, KCl, and Ara-CTP. We consider the remaining incorporation to be repair synthesis. Table II compares the characteristics of both types of DNA synthesis to the properties of DNA polymerases (Y and /?, as studied in extracts by other investigators. Each of the differences between replication and repair is characteristically different between DNA polymerases (Y and p, respectively. DNA replication involves many proteins and enzymes besides a polymerase. Although it is possible that one or more of these molecules is sensitive to salt, N-ethylmaleimide, or Ara-CTP, it is unlikely that another component of the replication (or repair) process would have the same sensitivity to all three inhibitors as DNA polymerase (Y (or, if so, that none of the components of the repair system have these sensitivities). These data provide strong evidence that the (Y enzyme is the polymerase for replication synthesis, whereas the p enzyme is the polymerase for repair synthesis. The simplest explanation for the apparently rate-limiting role of these polymerases is that they catalyze elongation, the polymerization of the bulk of the DNA precursors. An alternative explanation that we cannot rule out is that these polymerases are involved in an initiation or priming step which limits the subsequent rate. We did not observe repair synthesis in permeable Go cells after UV irradiation, probably because rodent cells are generally thought to be weak in UV repair, and repair of the resulting thymidine dimers usually requires 48 to 72 h for complete excision (21).
The primary action of HU is to inhibit ribonucleotide reductase (13), although HU also has been reported to inhibit a gap-filling step of SV40 DNA synthesis (22). HU is frequently used to inhibit replication preferentially while repair is measured (23). Insensitivity of repair to HU is generally attributed to the far smaller requirement for dNTPs, as compared to replication, so that preexisting pools of dNTPs might be adequate. Another possibility that we test here is that inhibition of ribonucleotide reductase by HU might be so incomplete under the usual conditions that enough dNTPs would be produced to permit repair. If this is so, HU should inhibit DNA synthesis most strongly when the reductase catalyzes the reaction at a low rate, e.g. at low substrate (rNDP) concentrations.
In contrast to strong inhibition by 1 mM HU of DNA replication in permeable cells, even when high concentrations (500 ,uM) of rNDPs were supplied (10, 13), 5 ITIM HU did not inhibit repair synthesis until the concentrations of rNDPs were below 5 pM (Table I). Even at 0.5 pM rNDPs, inhibition was only about 30%. At fist glance, the higher R,, of repair would seem to make it more sensitive to HU than replication. However, the repair of DNA is a low rate multistep process in which the high activity of ribonucleotide reductase appears to be far from rate-limiting.
If so, then a nearly complete inhibition of reductase activity would be required before any possible effect on repair could be observed. Even under conditions of low substrate and high HU concentrations, no definitive evidence was found for the inhibition of repair by HU.
Use of permeable cells to study enzyme kinetics assumes that 1) exogenous substrate has free access to the enzyme, and 2) there are no intracellular pools of substrates available to the enzyme. In these experiments, cells were treated with high doses of lysolecithin, resulting in "highly permeable" cell preparations (11). This should reduce intracellular pools (e.g. of dNTPs) and also ensure maximum accessibility of exogenous compounds to the enzyme of interest. Even so, complete elimination of enzyme or intracellular pool compartmentalization is not certain. For instance, the lack of dependence of bleomycin-induced repair on ATP is characteristic of "shortpatch" repair observed in prokaryotes and eukaryotes (21). We cannot rule out the possibility that permeable cells still contain a protected pool of ATP, or that salvage pathways are still active and are supplying ATP for repair synthesis.
The results presented here have examined enzymological aspects of DNA replication and repair in lysolecithin-permeabilized BHK cells. Since a wide variety of animal cells can be rapidly and easily permeabilized (11,13), the technique should be useful for examining many cellular processes, especially for relating results obtained with enzymes and cell extracts to phenomena observed in vivo.
Acknowledgments-We thank David Schneider for assistance in preparing the manuscript for publication.