Role of DNA Polymerases in Excision Repair in Escherichia coli*

The excision repair process following ultraviolet irradiation has been fractionated into its individual steps in tolu- ene-treated cells. Incision can be examined in vitro independently of other reactions by omission of the deoxyribonucleoside triphosphates which prohibits repair synthesis and causes incisions to accumulate. Incision requires ATP, continues from 10 to 15 min at 37”, and is specific for the excision repair pathway in toluene-treated cells. The exci- sion of pyrimidine dimers in strains containing DNA polymerase I is rapid when all components are present and results in 30 to 40% excision in the first 5 min. When the deoxyribonucleoside triphosphates are omitted, the excision rate, but not extent, is much reduced. This pattern of excision is comparable to that observed in intact cells deficient in DNA polymerase I. Neither DNA polymerase II nor III appears to influence the rate of dimer removal. The requirement for repair synthesis in excision repair has been evalu-ated by the addition of the deoxyribonucleoside triphosphates subsequent to incision accumulation, thus allowing repair synthesis and ligation to return the DNA to its original size. The reformation of the DNA to high molecular weight is rapid and nearly complete by 2


Role of DNA Polymerases in Excision
The excision repair process following ultraviolet irradiation has been fractionated into its individual steps in toluene-treated cells. Incision can be examined in vitro independently of other reactions by omission of the deoxyribonucleoside triphosphates which prohibits repair synthesis and causes incisions to accumulate.
Incision requires ATP, continues from 10 to 15 min at 37", and is specific for the excision repair pathway in toluene-treated cells. The excision of pyrimidine dimers in strains containing DNA polymerase I is rapid when all components are present and results in 30 to 40% excision in the first 5 min. When the deoxyribonucleoside triphosphates are omitted, the excision rate, but not extent, is much reduced. This pattern of excision is comparable to that observed in intact cells deficient in DNA polymerase I. Neither DNA polymerase II nor III appears to influence the rate of dimer removal. The requirement for repair synthesis in excision repair has been evaluated by the addition of the deoxyribonucleoside triphosphates subsequent to incision accumulation, thus allowing repair synthesis and ligation to return the DNA to its original size. The reformation of the DNA to high molecular weight is rapid and nearly complete by 2 min in cells containing DNA polymerase I. The reformation is slower and less complete in the absence of DNA polymerase I. This slower reformation is apparently catalyzed by DNA polymerase III as synthesis is observed in a mutant lacking both DNA polymerase I and II.

Ultraviolet
(UV) radiation causes the formation of dimers between adjacent pyrimidines in DNA (1). Excision repair is one mechanism whereby this damage is removed and corrected in the DNA (2,3). This repair pathway is characterized genetically in Escherichia coli by the uvr mutants (4, 5 the minimum, the pathway must have individual steps of incision near the site of the UV damage, excision of the damage, insertion of correct nucleotides by repair replication, and ligation of these new nucleotides to the old strand. We have attempted to manipulate this process in order to gain a clearer understanding of these individual steps. We have utilized toluene-treated cells for this study (6-10). This system has been useful in demonstrating the involvement of various DNA polymerases in repair (6, 10, 111, the requirement for ATP during repair synthesis (6, ll-13), and the requirement for ATP in UV-specific incision (7,14).
Incision is difficult to study in live cells as an independent event because the nicks are rapidly repaired. However, incision can be followed independently of subsequent repair events in vitro by omitting the deoxyribonucleoside triphosphates (dNTPs)' to prohibit repair replication.
Since repair replication is blocked, the incisions can not be repaired and thus accumulate during the postirradiation incubation. Incision can be studied in toluene-treated cells by following changes in the size of the DNA on alkaline sucrose gradients. Methods that block ligase such as NMN addition (15) or the use of a ligase mutant (16) also allow incision study (17,18). The advantage of allowing incision accumulation by dNTP starvation rather than the introduction of inhibitors is that it is not necessary to remove the inhibitor for the study of events subsequent to incision.
We have examined the influence of repair synthesis on the excision and reformation of high molecular weight DNA following incision accumulation.
This was done by utilizing cells deficient in the different polymerases and by manipulation of the dNTPs. Dimer excision requires ATP (9) and proceeds in the absence of repair synthesis (-dNTP), but at a rate simulating polk excision (9). The reformation experiments indicate that the repair synthesis mediated by DNA polymerase I leads to efficient ligation and a rapid reformation of intact DNA strands. Reformation in the absence of DNA polymerase I is slow. In strains lacking both DNA polymerase I and II, reformation is observed as a result of repair synthesis by DNA polymerase III. These data suggest that there are two alternatives for repair synthesis and reformation following incision.
Cells were grown in L-broth (10 g of tryptone, 5 g of yeast extract, 5 g of sodium chloride, and 1 liter of water) supplemented with 10 hg/ml of thymine and the DNA was labeled with tritiated thymidine (2 Z&i/ml, 6 Ci/mmol). Cells which were to be used for dimer analysis received a second supplement of tritiated thymidine after 30 min of growth.
Toluene Treatment -Cells were toluene treated as previously described (22). Fifty-milliliter cultures were harvested at a cell density of 0.5 to 1 x lo0 cells/ml and resuspended in 0.05 M potassium phosphate buffer (pH 7.4). The cells were exposed to 1% toluene at room temperature for 5 to 10 min with slow stirring. After centrifugation the cells were resuspended at a concentration of 5 x lOa to 10'" cells/ml in potassium phosphate buffer. Zrradiatiorr and Assay for Repair Synthesis -The toluene-treated cells were exposed to radiation emitted largely at 254 nm from an unfiltered GE G8T5 lamp at a distance of 100 cm. Radiant energy was 6 J/m*/min as measured by an ultraviolet intensity meter (Ultraviolet Products, Inc.) and resulted in the expected pyrimidine dimer production (9 Our results indicate two distinguishable rates of excision (Fig. 3). The faster rate requires synthesis by DNA polymerase I (Refs. 9 and 31 and Fig. 2). In poZBstrains deficient in DNA polymerase II, the patterns of excision (polA+ or polA-) are identical with those in polB+ strains. These data indicate that DNA polymerase II is not critical for either fast or slow dimer excision. To examine the influence of DNA polymerase III, it was necessary to use conditional mutants as the polC gene function is required for  4. Influence of DNA polymerase I on reformation of high molecular weight DNA after incision accumulation. Prelabeled toluene-treated cells were exposed to 30 J/m2 UV dose and incubated for 10 min as in Fig. 1 in the presence of ATP but in the absence of dNTPs. The dNTPs were then added and the reactions stopped at various times and analyzed as in Fig. 1 repair synthesis to proceed and if the appropriate configuration exists in the repaired DNA, then ligation will result in the reformation of high molecular weight DNA. The influence of DNA polymerase I on repair synthesis is examined in Fig.  4. The data in the left panel for the polA+ strain demonstrate that after the incision period, the addition of dNTPs results in rapid reformation.
Longer incubation results in a profile which is close to that obtained from unirradiated cells carried through the same protocol. In thepoZA-cells, there is a much slower and less complete reformation with only a small proportion of the DNA achieving a molecular weight similar to the unirradiated control. It is important to keep in mind that it is the parental DNA which is followed in this experiment and not newly incorporated repair synthesis. The amount of synthesis during the postirradiation incubations is insufficient to account for the production of high molecular weight material de novo; rather, the labeled parental strands must be joined as a result of the repair synthesis and ligation.
The participation of DNA polymerase II in excision repair was examined in polB-cells (Fig. 5). In a polB-strain with an active DNA polymerase I, the reformation is fast and The results with HMS85, deficient in DNA polymerase II but containing DNA polymerase I, are to the Zefi and HMS83, deficient in both DNA polymerase I and II, to the right. Experiment was as in Fig. 4. 0, 10 min without dNTPs; 0, dNTPs added for 2 min; A, dNTPs added for 20 min; A, unirradiated cells 10 min without dNTPs followed by 20 min with dNTPs.
complete as in other polA+ strains (left panel). This indicates that DNA polymerase II is not required for fast reformation. This is consistent with the lack of UV sensitivity observed for this strain (21). In a strain deficient in both DNA polymerases I and II, a slow and less complete reformation is observed, as in Fig. 4. These data suggest that DNA polymerase III may be responsible for the low level of reformation that is seen in the absence of DNA polymerase I. DISCUSSION We have fractionated the multistep process of excision repair into individual reactions (Fig. 6). The incision reaction is a consequence of the UV irradiation and requires ATP but neither dNTPs nor repair synthesis. Incisions accumulate at 37" for approximately 15 min in vitro. This time period is the same as that seen for excision of pyrimidine dimers (9) and incorporation during repair synthesis (6). In irradiated poZAcells, there is a slow rate of strand breakage in the absence of ATP (Fig. 1) which is also present in unirradiated control cells. The lower stability of the DNA in poZA-cells may account for an indication that incision was ATP-independent since DNA polymerase I-deficient cells were used in an earlier study (17). We have demonstrated that the low level of repair synthesis seen in the absence of ATP does not vary appreciably with UV dose (32).
The number of breaks introduced into the DNA is less than the number of pyrimidine dimers produced by the given doses. We observe incisions at 10 to 15% of the dimer sites. This value is less than the 30% observed by Ganesan (33) for in vitro incision with T4 UV endonuclease in Brij-treated cells but about the same as values observed by Waldstein et al. (14). We observe approximately 30% excision in vivo under the same conditions. The excision of pyrimidine dimers from irradiated DNA does not require either dNTPs or repair synthesis. However, the excision rate in cells that contain DNA polymerase I is stimulated by simultaneous repair synthesis. The 5' + 3' exonuclease activity of DNA polymerase I is also stimulated by simultaneous polymerization (29,30). We have also shown a requirement for the 5' + 3' exonuclease function of DNA polymerase I for fast excision in the presence of dNTPs (9). These data taken together suggest the fast excision in the presence of simultaneous repair synthesis is catalyzed by DNA polymerase I, probably by a nick translation mechanism. DNA polymerase I is suited to participate in an efficient repair as it has been shown in vitro to be capable of dimer excision (29, 34), and it can carry out nick translation synthesis in which both the polymerizing and 5' + 3' exonucleolytic functions are simultaneously active (35,36). The absence of excision stimulation by dNTPs in poZA -strains might indicate that removal of dimers and polymerization are uncoupled, since these cells contain nearly normal levels of the 5' -+ 3' exonuclease activity of DNA polymerase I.
We reported earlier that toluene-treated wild type cells maintain their DNA in a high molecular weight form following irradiation when incubated in a complete reaction mixture (37). For this to occur following incision, the steps of excision and repair synthesis that require a nick in the DNA must be completed rapidly so that ligation will limit the number of breaks present in the DNA at any one time. That is, the incisions are repaired and resealed before they can accumulate. DNA polymerase I-mediated repair, resulting in breaks remaining in the DNA for the shortest time, is consistent with the observation that a small number of nucleotides are inserted into the damage site in poZA+ cells (38) and with low levels of synthesis in z&-o (32). The rapid reformation and small patch size in poZA + cells indicate an effective interaction of polynucleotide ligase in the process.
The data presented here indicate that DNA polymerase I is required for efficient reformation following incision. It is evident from a comparison of the rate of reformation with that of incision accumulation that the steps of repair synthesis and ligation are not rate-limiting for the complete process in poZA + cells. The rapid reformation would account for the lack of any appreciable drop in the size of the DNA in poZA+ cells in complete reaction mixtures although it is conceivable that ligase is simply resealing incisions before excision. This is unlikely, however, since excision proceeds under these conditions. DNA polymerase I is not the only DNA polymerase which participates in excision repair synthesis. Even though poZAstrains are UV sensitive, they are not as sensitive as uurAor uvrB-mutants which suggests that there is another DNA polymerase capable of acting during excision repair. DNA polymerase III has been shown to participate in viuo in the absence of DNA polymerase I (39, 40). DNA polymerase III has also been shown to be responsible for postirradiation repair incorporation (6) which is dependent on the uurA gene function (8). In this report, we have shown that in poZA-or poZA-, poZB-strains there is some, albeit poor, reformation of intact DNA strands.
DNA polymerase III has been reported to possess a 5' + 3' exonuciease activity that is capable of excising dimers (41). In cells containing a conditionalpolc mutation, under restrictive conditions, the excision rate is unaffected. This result indicates that the excision of dimers and repair synthesis are not coupled in this case, suggesting separate enzymes for excision and resynthesis. This is consistent with the lack of stimulation of excision by dNTPs in poZA-cells. At this point, it is not evident what nuclease is responsible for dimer excision in poZA -strains.