Vol. 129, No. 3 Printed in U.S.A.
JOURNAL OF BACTERIOLOGY, Mar. 1977, p. 1415-1423 Copyright 0 1977 American Society for Microbiology
Deoxyribonucleic Acid Repair In Vitro by Extracts of Escherichia coli WARREN E. MASKER Biology Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37830
Received for publication 14 October 1976
Deoxyribonucleic acid (DNA) from bacteriophage T7 has been used to monitor the capacity of gently lysed extracts of Escherichia coli to perform repair resynthesis after ultraviolet (UV) irradiation. Purified DNA damaged by up to 100 J of UV radiation per m2 was treated with an endonuclease from Micrococcus luteus that introduces single-strand breaks in irradiated DNA. This DNA was then used as a substrate to study repair resynthesis by extracts ofE. coli. It was found that incubation with the extract and exogenous nucleoside triphosphates under suitable assay conditions resulted in removal of all pyrimidine dimers and restoration of the substrate DNA to its original molecular weight. Repair resynthesis, detected as nonconservative, UV-stimulated DNA synthesis, was directly proportional to the number of pyrimidine dimers introduced by radiation. The repair mode described here appears to require DNA polymerase I since it does not occur at the restrictive temperature in polA12 mutants, which contain a thermolabile polymerase. The addition of purified DNA polymerase I to extracts made from a polA mutant restores the ability to complete repair at the restrictive temperature. Living organisms rely upon deoxyribonucleic acid (DNA) repair mechanisms to maintain copies of their genetic information free from mutation due to the deleterious effects of radiation and foreign chemicals. In bacteria (and in other organisms as well) DNA repair can proceed by any of a number of alternate repair pathways (11, 37). The complexity of these pathways is indicated in part by the large number of genes that are involved in the repair process (6). The products of most of these genes have not yet been identified. Moreover, although enzymes with properties well suited for certain steps in DNA repair have been isolated (11), there is little assurance that the in vitro biochemical properties of these enzymes are not significantly altered from in vivo forms. Thus, although a wealth of information concerning DNA damage and repair has accumulated (33), relatively little is known about the relationships among various biochemical steps or about the factors that dictate which of the available repair pathways will be followed. Some information concerning details of the biochemical repair of DNA has been obtained through use of quasi in vitro systems such as toluene-treated or sucrose-plasmolyzed cells (1, 16, 17, 27, 28, 35). True in vitro systems have aided our understanding of the pyrimidine dimer excision process in extracts made from cells infected with bacteriophage T4 (9) and have been used
to study the excision of thymidine residues damaged by ionizing radiation (13). Also, it has been possible to achieve complete repair of ultraviolet (UV) radiation-damaged DNA in vitro by use of purified enzymes from Micrococcus luteus (12). As yet there has been no report of completion of dark repair of DNA in a true in vitro system with extracts of Escherichia coli, an organism that, because of extensive genetic investigation, is especially valuable for studies of DNA repair. Reported here are experiments on dimer excision and repair resynthesis performed in vitro by extracts of E. coli. An endonuclease from M. luteus was used to nick UV-irradiated DNA near the sites of pyrimidine dimers. This DNA was then used to monitor the repair capacity of extracts prepared from gently lysed E. coli. Incubation with these extracts resulted in removal of pyrimidine dimers, repair resynthesis, and ligation which restored the DNA to its original molecular weight.
MATERIALS AND METHODS Bacteria and bacteriophage. Bacterial strains used in this study included HMS146: thy recB21 (19); D10: thy end polAl (22); DRi10: a polA+ revertant of D110 (19); 011': thy sup+ (32); and MM387: thy polA12 recB21 (21). Bacteria were routinely grown in L broth (20) with shaking at 32 or 37°C. Bacteriophage T7 (from W. Studier) were grown as described by Studier (32) on strain 011'. 1415
J. BACTERIOL.
1416 MASKER DNA. DNA was purified from bacteriophage T7 as described by Richardson (25). DNAs from phage M13 and PM2 were gifts from S. Mitra and W. L. Carrier, respectively. All DNA concentrations are expressed as nucleotide phosphorous equivalents. Enzymes. An endonuclease from M. luteus that specifically attacks pyrimidine dimers formed after UV irradiation was generously provided by W. L. Carrier. The preparations used in this study were prepared by a modification of a previously described procedure (4). The enzyme preparation had a protein concentration of about 100 ,ug/ml and was approximately 1,000-fold purified (W. L. Carrier, personal communication). The enzyme was stored frozen in 10 mM tris(hydroxymethyl)aminomethane (Tris)-hydrochloride (pH 7.5)-l mM 2-mercaptoethanol-10% (vol/vol)ethylene glycol-bovine serum albumin (1 mg/ml). This preparation was routinely diluted 1:15 in an assay mixture containing 25 mM Tris-hydrochloride (pH 7.5) and 10 to 12 nmol of T7 DNA and incubated at 30°C for 15 min. Under these conditions the enzyme was in excess and could produce approximately one single-strand break per pyrimidine dimer in DNA irradiated with up to 70 J of UV per M2. DNA polymerase I from E. coli was a gift from S. Mitra. Lysozyme was purchased from Calbiochem. Other chemicals. Unlabeled nucleoside triphosphates and [methyl-3H]thymidine were purchased from Schwarz/Mann Bioresearch. [32P]deoxyadenosine 5'-triphosphate (dATP) was purchased from New England Nuclear Corp., nicotinamide adenine dinucleotide (NAD) from Calbiochem, and bromodeoxyuridine 5'-triphosphate (BrdUTP) from P-L Biochemicals. Preparation of extracts. Exracts of E. coli were prepared by a modification of the method described by Wickner et al. (36). E. coli were grown to an absorbance at 590 nm (A5,0) 1.7 (about 8 x 108 cells/ml) and then quickly chilled. The cells were collected by centrifugation, resuspended in 50 mM Tris-hydrochloride (pH 7.5)-10% sucrose at 100 times the original concentration, and then quickly frozen in liquid N2. The cells were thawed at room temperature and incubated with 0.1 M NaCl and 200 ,ug of lysozyme per ml for 45 min at 0°C and for 3 min at 37°C. The lysate was quickly chilled to 0°C and centrifuged at 39,000 rpm at 0°C for 30 min in a Spinco 40 rotor. The resulting supernatant had an A280 of approximately 35 and a ratio of A260 to A2.8 of about 2.0. For some experiments the extracts were dialyzed for 22 h at 0°C against 1,000 volumes of 50 mM Tris-hydrochloride (pH 7.5)-0.1 M NaCl-10% sucrose to reduce the level of endogenous nucleoside triphosphates. The dialyzed extracts were stable to freezing and storage under liquid nitrogen. UV irradiation. DNA to be used as a substrate was irradiated at 50 to 100 ,ul at a time at a concentration of 0.5 to 1.0 mM in a shallow-well microscope slide with constant stirring at room temperature. The depth of liquid was less than 1 mm. A pair of germicidal lamps with an incident dose of 1.04 J/m2 per s provided the UV source. Reaction conditions for restoration of DNA. The standard reaction mixture of 0.1 ml contained 30 mM Tris-hydrochloride (pH 7.5), 25 mM MgCl2, 80
,uM deoxyribonucleoside triphosphate (dNTP), 0.25 mM NAD, 12 mM 2-mercaptoethanol, 3 nmol of DNA, and 40 ,ul of extract prepared as described above. In early experiments 0.3 mM dNTPs and all four ribonucleoside triphosphates (rNTPs) at 0.3 mM were included in the reaction mix. Further experiments showed that the rNTPs were not required and that high concentrations of dNT.P's were unnecessary. The reactions were routinely incubated for 15 min at 30°C and then stopped by adding 10 i1. of 0.5 M ethylenediaminetetraacetic acid (EDTA) and chilling to 0°C or (for isopycnic gradient analysis) by adding 1.0 ml of ice-cold NET buffer (0.1 M NaCl-0.01 M EDTA-0.01 M Tris-hydrochloride [pH 8.0]). Under the conditions described here, DNA from approximately 10 phage particles is present per unit of protein released from one cell. Other methods. Zone sedimentation in alkaline sucrose, isopycnic gradient analysis, and determination of radioactivity have been described previously (16). Numbers of single-strand breaks in DNA sedimented through alkaline sucrose were calculated from the formula: 2(1/MW,,,, - 1IM,e,), where U, UV and Mw,, are the weight average molecular weights of the irradiated and unirradiated samples. In this study CsCl gradients were collected from a hole punctured in the bottom of the centrifuge tube, and fractions were precipitated with ice-cold 1 N HCl-0. 1 M inorganic pyrophosphate before being filtered onto Reeve Angel glass-fiber filters soaked in 0.1 M inorganic pyrophosphate. Amounts of thymine dimers were determined by two-dimensional paper chromatography as described by Carrier and Setlow (5). The total number of pyrimidine dimers was calculated from the number of thymine-containing dimers by use of values obtained by Setlow and Carrier (29).
=
RESULTS Removal of damaged sites in vitro. According to current models (11), the initial step in the excision repair process is recognition of the offending lesion and introduction of a singlestrand incision near the site of the damage. Subsequent to incision, the offending lesion with an adjacent stretch of nucleotides is excised, the DNA in the damaged region is resynthesized, and the new patch of DNA is joined to the contiguous parental DNA strand. To see whether the excision repair process could occur in vitro, we prepared extracts of E. coli, using recB mutants to avoid nuclease activity by exonuclease V. Linear duplex DNA from bacteriophage T7 provided a suitable substrate. Our initial efforts to detect incision of UV-irradiated T7 DNA incubated with extracts of E. coli were not successful, even in the presence of ATP (to promote incision) (35) and nicotinamide mononucleotide (to prevent ligation). In toluenetreated cells, incision breaks accumulate only at a small fraction of the UV-induced lesions
VOL. 129, 1977
DNA REPAIR IN VITRO
(28, 30). If incision is similarly limited in crude extracts, it is not likely that a small number of single-strand breaks could be detected by alkaline sucrose gradient analysis oflow-molecularweight DNA. Therefore, in an effort to bypass the limiting incision step and to make possible a study of excision and resynthesis in vitro, an endonuclease from M. lueus that is specific for irradiated DNA was used. It was necessary to examine the properties of the M. luteus damage-specific endonuclease under our assay conditions. The enzyme was tested on unirradiated DNA and found free from nonspecific endonuclease activity. Also, by measuring the amount of pyrimidine dimers in irradiated DNA incubated in our standard reaction conditions, we confirmed that the M. luteus enzyme could not by itself excise thymine dimers. To see whether the number of incisions introduced into UV-irradiated DNA by the enzyme was proportional to the number of pyrimidine dimers, the experiment shown in Fig. 1 was performed. DNA was irradiated, incubated with the M. luteus enzyme, and sedimented through alkaline sucrose. Figure 1A shows that under our experimental conditions the reciprocal of the M,, ofthe treated DNA was 1 ,16
A
0
2-
1.,
0
.80..44 z
4-
.
0 B
cw 0.30
w
0.20
z
a:
0.1
0
20
40 60 80 UV DOSE (J/m2)
100
FIG. 1. Comparison of thymine dimer content and enzyme-sensitive sites. Portions (50 pl) of 3H-labeled T7 DNA were irradiated with various fluences of UV. Portions of this DNA were incubated with the M. luteus damage-specific endonuclease, as described in the text. The DNA was sedimented through alkaline sucrose, and the M, values were determined. (A) 2/Mw as a function of UV fluence. The rest of the DNA was examined for thymine dimer content by paper chromatography. (B) Percentage of thymine as dimers shown as a function of UV dose.
1417
proportional to the UV fluence in the range 0 to 70 J/m2 and that the enzyme produced about 0.015 incisions/J per m2 per 106 molecular weight. The irradiated DNA was also examined for thymine dimer content. Figure 1B shows that about 0.0037% of the thymine present was converted to thymine-containing dimers by 1 J of UV per M2. This corresponds to about 0.019 pyrimidine dimers/J per m2 per 106 molecular weight (about 0.46 dimers/J per m2 per T7 genome of 25 x 106 daltons). Thus, in view of this observation, together with an earlier report by R. B. Setlow, W. L. Carrier, and J. Stewart (Biophys. J. 15:194a, Abstr. no. TH-PM-H2, 1975), we will assume that this enzyme produces one incision per pyrimidine dimer. To test the ability of extracts of E. coli to restore the molecular weight of damaged DNA, the experiment shown in Fig. 2 was performed. Purified UV-irradiated DNA was incubated with an excess of the damage-specific nuclease. A portion of the reaction mixture was withdrawn and examined on alkaline sucrose gradients along with unirradiated control DNA. The remainder of the endonuclease-treated DNA was incubated with extract prepared from E. coli and then subjected to sedimentation analysis with alkaline sucrose. As shown in Fig. 2, incubation with the endonuclease caused an accumulation of single-strand breaks in the irradiated DNA. A second incubation with an extract from E. coli restored the DNA to the same molecular weight as the unirradiated control. It was conceivable that the restoration of the DNA to its original molecular weight might be caused by ligation between juxtaposed ends without removal of the lesion. Therefore it was necessary to test whether any sites sensitive to the damage-specific endonuclease remained after completion of the treatment shown in Fig. 2. T7 DNA was irradiated with 40 J/m2 and subjected to the same treatment described in the legend to Fig. 2. As a control, a portion of the irradiated DNA was diluted into a reaction mixture without endonuclease or extract. Both mixtures were then subjected to extraction with phenol and dialyzed. A portion of the DNA was retained as a control, and the remainder was incubated with the damage-specific endonuclease from M. luteus. Figure 3A shows that DNA not treated with the in vitro repair system still contains the expected number of sites sensitive to the endonuclease. However, Fig. 3B shows that the DNA incised and then restored to the original molecular weight has no remaining endonuclease-sensitive sites. This result has been confirmed by use of paper chromatography to measure dimer content (Carrier and
1418 MASKER
J. BACTERIOL.
DNA. This experiment confirms that there is no enzymatic activity in the M. luteus enzyme preparation that is required for the dimer excision, repair resynthesis, or ligation steps needed to restore the incised DNA. Repair replication. The in vitro system described here is potentially valuable for measuring repair resynthesis since both the specific activity of the nucleoside triphosphates and the number of repaired sites are known. Therefore,
10
N
O
16
1.0
0.4 Q8 0.6 0.2 RELATIVE DISTANCE SEDIMENTED
0
FIG. 2. Alkaline sucrose sedimentation of DNA repaired in vitro. 3H-labeled DNA from bacteriophage T7 was irradiated with 40 Jof UV per Mi2. A 3nmol amount each of irradiated and unirradiated DNA was treated with a damage-specific endonuclease from M. luteus and then layered on alkaline sucrose gradients. Identical samples received a second incubation with an assay mixture containing extract from strain HMS146 before being layered on gradients and spun at 49,000 rpm for 120 min. Profiles of radioactivity recovered from these gradients are shown. (A) Treatment with M. luteus enzyme only; (B) treatment with enzyme followed by treatment with E. coli extract. Symbols: 0, -UV; A, +UV.
Masker, unpublished data). We conclude that all of the damage recognized by the M. luteus enzyme has been removed by incubation with the E. coli extract. To be sure that possible contaminants in the damage-specific endonuclease were not responsible for the restoration of incised DNA, the experiment shown in Fig. 4 was performed. Irradiated DNA was treated with the M. luteus endonuclease and then extracted with phenol to inactivate the enzyme. When this nicked DNA was incubated with an E. coli extract under standard assay conditions, the DNA was restored to nearly its original molecular weight. After incubation, EDTA was added to a separate portion of repaired DNA to a concentration sufficient to inhibit repair but allow incision by the M. luteus endonuclease. The DNA was incubated a second time with the damage-specific endonuclease before being sedimented through alkaline sucrose. As seen in Fig. 4 no endonuclease-sensitive sites remain in the repaired
0
1.0
0.4 0.8 0.6 0.2 RELATIVE DISTANCE SEDIMENTED
0
FIG. 3. Complete removal of enzyme-sensitive sites by E. coli extract. 3H-labeled T7 DNA was exposed to 40 JIm2. A portion of the DNA was diluted and retained as a control. The remainder was treated first with the damage-specific endonuclease from M. luteus and then with an extract prepared from E. coli strain HMS146. Both portions of DNA were extracted with phenol and dialyzed against 10 mM Tris-hydrochloride (pH 7.5)-10 mMNaCl-1 mMEDTA. A part of each sample was retained as a control; the remainder was treated a second time with the M. luteus endonuclease. The reaction mixtures were layered on alkaline sucrose gradients and spun as described in the legend to Fig. 2. Profiles of radioactivity recovered from these gradients are shown. (A) No treatment before extraction with phenol; (B) treated with endonuclease and extract. Symbols: 0, no second treatment with enzyme; A, incubated a second time with damage-specific endonuclease.
VOL. 129, 1977
DNA REPAIR IN VITRO
1419
0 w
a:
U
0 o 20-
0io0
dO z U r
0.
oL
1.0
0.6 0.4 0.2 0.8 RELATIVE DISTANCE SEDIMENTED
0
FIG. 4. DNA repaired after inactivation of the damage-specific endonuclease. "C-labeled T7 DNA was exposed to 40 J of UV per m2, treated with M. luteus damage-specific endonuclease, and extracted with phenol. After dialysis, a portion of the DNA was treated with an E. coli extract, as described in the text. A 50 mM concentration ofEDTA was added to a sample of the restored DNA, and this sample was treated a second time with the M. luteus enzyme before all the reaction mixes were sedimented through alkaline sucrose. Symbols: 0, unirradiated control; A, no treatment with extract; O, treated with complete reaction mixture; O, reincubated with damage-specific endonuclease and EDTA. we attempted to measure repair replication performed by extracts of E. coli. When experiments such as the one shown in Fig. 2 were performed with the four dNTPs absent from the reaction mixture, complete repair still occurred, presumably because of nucleoside triphosphates still present in the extract. However, when extracts were dialyzed extensively and then used in an experiment such as the one described in the legend to Fig. 2, the results shown in Fig. 5 were obtained. When DNA damaged by exposure to 100 J of UV per m2 was treated with the damage-specific endonuclease and reincubated in a complete reaction mixture including dialyzed extract and an 80 juM concentration of each dNTP, almost complete restoration of the DNA to its original molecular weight was achieved. However, when the dNTP's were not present, the single-strand breaks introduced by the endonuclease remained in the DNA. An extract similar to the one used in the previous experiment was dialyzed to remove dNTP's and then utilized to measure repair resynthesis. In this experiment BrdUTP was used in place of thymidine 5'-triphosphate and the dATP was labeled with 32P. Although it seemed unlikely that any semiconservative replication of the T7 DNA could occur in these extracts, CsCl gradient analysis was performed to be sure that no aberrant initiation was occurring as a result of UV exposure. The results
shown in Fig. 6 indicate that there was no density shift of 32P-labeled DNA relative to 3Hlabeled DNA caused by incorporation of BrdUTP. This is the result expected if only short patches of DNA were synthesized to perform the repair seen in Fig. 5. The amount of repair resynthesis was linear with a UV dose of up to at least 150 J/m2, as shown in the insert to Fig. 6. Based on the amount of [32P]dATP incorporated, our measured value of 4.6 dimers/T7 genome per 10 J of UV per m2, and the fact that 1 nmol of DNA is equivalent to 7.5 x 109 T7 genomes, we arrive at an estimate of 17 nucleosides incorporated per pyrimidine dimer removed. Because of possible residual DNA precursors in the extracts, this value may be an underestimate. Role of DNA polymerase I. In vitro repair in the absence of DNA polymerase I was examined by using a polA12 mutant temperature sensitive for this enzyme (21, 34). DNA from bacteriophage T7 was irradiated with 50 J of UV per m2 and incubated with the damagespecific endonuclease. Reincubation with an extract prepared from strain HMS146 (recB) caused restoration of the DNA to its original size, irrespective of whether the incubation was carried out at 30 or 4200 (data not shown). Figure 7 shows the results of reincubation with an extract prepared from strain MM387 (polA12 recB). When the reincubation was carried out at 30°C, DNA polymerase I was active
J. BACTERIOL.
1420 MASKER
0.4 0.6 RELATIVE DISTANCE SEDIMENTED
FIG. 5. Repair completed by dialyzed extract ofE. coli. 3H-labeled T7 DNA was exposed to 100 J of UV per m2. A portion of unirradiated DNA was retained as a control. The irradiated DNA was treated with the damage-specific endonuclease, and a portion was saved as a control. The remaining incised DNA was incubated with an extract from strain HMS146 that had been dialyzed at 0°C for 22 h against 10% sucrose-.1 M NaCl,50 mM Tris-hydrochloride (pH 7.5). In one sample the reaction mixture was complete; in the other sample all four dNTPs were missing. All samples were layered on alkaline sucrose gradients and spun at 49,000 rpm for 120 min. Profiles of radioactivity recovered from these gradients are shown. Symbols: 0, control, no UV; A, UV, no reincubation; V, reincubation in complete assay mixture; O, reincubation in assay mixture without dNTP's.
and the DNA was restored to its original molecular weight. However, when DNA polymerase I was inactivated by heating it to 42°C, singlestrand breaks remained in the DNA after reincubation. As a control, 2 U of purified DNA polymerase I was included in one reaction reincubated at 420C. As seen in Fig. 7, the purified enzyme substituted for the thermally inactivated enzyme to restore the damaged DNA. Identical results were obtained when the experiment was repeated with dialyzed extracts. Apparently DNA polymerase I play an important role in the completion of repair we observe in vitro, at least when incubation occurs at 420C. When 1.2 mM ATP was included in the reaction mixture in an effort to stimulate repair resynthesis by DNA polymerases II and III (16), the strain MM387 extract was still unable to restore the DNA to its original molecular weight when the incubation was carried out at 420C. The results reported here agree with previously reported observations (16, 30) on repair in toluene-treated cells. DISCUSSION The experiments presented above show that purified bacteriophage T7 DNA damaged by UV radiation and incised enzymatically at the sites of pyrimidine dimers can be repaired in
vitro by extracts prepared from E. coli. The criteria used to monitor repair include restoration of the DNA to its original molecular weight, removal of sites recognized by a damage-specific nuclease, and measurement of UVstimulated repair resynthesis. By examining the ability of UV-irradiated T7 DNA to be packaged in vitro (26) and to form viable phage particles, we have also obtained preliminary results which show that the procedure described above increases the biological activity of irriadiated DNA (Kuemmerle and Masker, unpublished data). In the present study an endonuclease from M. luteus was used to overcome incision deficiency in the extracts and to promote incision near the sites of pyrimidine dimers. This enzyme may be identical (or similar) to UV-specific endonucleases isolated and characterized by a number of workers (4, 15, 23). Previous studies with the enzyme used in this study have shown that only the dimer-containing DNA strand is nicked (4) and one nick per pyrimidine dimer is introduced (Setlow et al., Biophys. J. 15:194a). The present study confirms that this enzyme produces about one single-strand break per pyrimidine dimer, shows no activity toward unirradiated DNA, and cannot by itself excise thymine dimers. Thus, this enzyme proved suit-
VOL. 129, 1977
DNA REPAIR IN VITRO
1421
10 8642-
E 0L.
E
0
O
0.0
to
to
0
C)
0
I 8- 100
Q-
J/m2
N to
OI
to 0
1 60
4-
2S*
0
9S0
*.I
i.w
I
20
-jO
25
5 30 0 15 20 25 30 10 FRACTION NO. FIG. 6. Repair resynthesis in vitro. 3H-labeled T7 DNA was irradiated with 0, 50, 100, or 150 J of UV per m2 before being treated with the damage-specific endonuclease. Six-nanomole samples were incubated in 0.2 ml of standard reaction mixtures, except that BrdUTP was used in place of thymidine 5'-triphosphate and [32P]dATP was included at 480 cpm/pmol. After 15 min of incubation at 30°C with an extract from strain HMS146, the reactions were stopped and subjected to isopycnic CsCl gradient analysis. Profiles of radioactivity recovered from these gradients are shown. Symbols: 0, 3H; A, 32p. The insert shows picomoles of [32P]deoxyadenosine 5'-monophosphate (P]dAMP) incorporated into acid-insoluble material as a function of the UV dose. These values were corrected for small (
JOURNAL OF BACTERIOLOGY, Mar. 1977, p. 1415-1423 Copyright 0 1977 American Society for Microbiology
Deoxyribonucleic Acid Repair In Vitro by Extracts of Escherichia coli WARREN E. MASKER Biology Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37830
Received for publication 14 October 1976
Deoxyribonucleic acid (DNA) from bacteriophage T7 has been used to monitor the capacity of gently lysed extracts of Escherichia coli to perform repair resynthesis after ultraviolet (UV) irradiation. Purified DNA damaged by up to 100 J of UV radiation per m2 was treated with an endonuclease from Micrococcus luteus that introduces single-strand breaks in irradiated DNA. This DNA was then used as a substrate to study repair resynthesis by extracts ofE. coli. It was found that incubation with the extract and exogenous nucleoside triphosphates under suitable assay conditions resulted in removal of all pyrimidine dimers and restoration of the substrate DNA to its original molecular weight. Repair resynthesis, detected as nonconservative, UV-stimulated DNA synthesis, was directly proportional to the number of pyrimidine dimers introduced by radiation. The repair mode described here appears to require DNA polymerase I since it does not occur at the restrictive temperature in polA12 mutants, which contain a thermolabile polymerase. The addition of purified DNA polymerase I to extracts made from a polA mutant restores the ability to complete repair at the restrictive temperature. Living organisms rely upon deoxyribonucleic acid (DNA) repair mechanisms to maintain copies of their genetic information free from mutation due to the deleterious effects of radiation and foreign chemicals. In bacteria (and in other organisms as well) DNA repair can proceed by any of a number of alternate repair pathways (11, 37). The complexity of these pathways is indicated in part by the large number of genes that are involved in the repair process (6). The products of most of these genes have not yet been identified. Moreover, although enzymes with properties well suited for certain steps in DNA repair have been isolated (11), there is little assurance that the in vitro biochemical properties of these enzymes are not significantly altered from in vivo forms. Thus, although a wealth of information concerning DNA damage and repair has accumulated (33), relatively little is known about the relationships among various biochemical steps or about the factors that dictate which of the available repair pathways will be followed. Some information concerning details of the biochemical repair of DNA has been obtained through use of quasi in vitro systems such as toluene-treated or sucrose-plasmolyzed cells (1, 16, 17, 27, 28, 35). True in vitro systems have aided our understanding of the pyrimidine dimer excision process in extracts made from cells infected with bacteriophage T4 (9) and have been used
to study the excision of thymidine residues damaged by ionizing radiation (13). Also, it has been possible to achieve complete repair of ultraviolet (UV) radiation-damaged DNA in vitro by use of purified enzymes from Micrococcus luteus (12). As yet there has been no report of completion of dark repair of DNA in a true in vitro system with extracts of Escherichia coli, an organism that, because of extensive genetic investigation, is especially valuable for studies of DNA repair. Reported here are experiments on dimer excision and repair resynthesis performed in vitro by extracts of E. coli. An endonuclease from M. luteus was used to nick UV-irradiated DNA near the sites of pyrimidine dimers. This DNA was then used to monitor the repair capacity of extracts prepared from gently lysed E. coli. Incubation with these extracts resulted in removal of pyrimidine dimers, repair resynthesis, and ligation which restored the DNA to its original molecular weight.
MATERIALS AND METHODS Bacteria and bacteriophage. Bacterial strains used in this study included HMS146: thy recB21 (19); D10: thy end polAl (22); DRi10: a polA+ revertant of D110 (19); 011': thy sup+ (32); and MM387: thy polA12 recB21 (21). Bacteria were routinely grown in L broth (20) with shaking at 32 or 37°C. Bacteriophage T7 (from W. Studier) were grown as described by Studier (32) on strain 011'. 1415
J. BACTERIOL.
1416 MASKER DNA. DNA was purified from bacteriophage T7 as described by Richardson (25). DNAs from phage M13 and PM2 were gifts from S. Mitra and W. L. Carrier, respectively. All DNA concentrations are expressed as nucleotide phosphorous equivalents. Enzymes. An endonuclease from M. luteus that specifically attacks pyrimidine dimers formed after UV irradiation was generously provided by W. L. Carrier. The preparations used in this study were prepared by a modification of a previously described procedure (4). The enzyme preparation had a protein concentration of about 100 ,ug/ml and was approximately 1,000-fold purified (W. L. Carrier, personal communication). The enzyme was stored frozen in 10 mM tris(hydroxymethyl)aminomethane (Tris)-hydrochloride (pH 7.5)-l mM 2-mercaptoethanol-10% (vol/vol)ethylene glycol-bovine serum albumin (1 mg/ml). This preparation was routinely diluted 1:15 in an assay mixture containing 25 mM Tris-hydrochloride (pH 7.5) and 10 to 12 nmol of T7 DNA and incubated at 30°C for 15 min. Under these conditions the enzyme was in excess and could produce approximately one single-strand break per pyrimidine dimer in DNA irradiated with up to 70 J of UV per M2. DNA polymerase I from E. coli was a gift from S. Mitra. Lysozyme was purchased from Calbiochem. Other chemicals. Unlabeled nucleoside triphosphates and [methyl-3H]thymidine were purchased from Schwarz/Mann Bioresearch. [32P]deoxyadenosine 5'-triphosphate (dATP) was purchased from New England Nuclear Corp., nicotinamide adenine dinucleotide (NAD) from Calbiochem, and bromodeoxyuridine 5'-triphosphate (BrdUTP) from P-L Biochemicals. Preparation of extracts. Exracts of E. coli were prepared by a modification of the method described by Wickner et al. (36). E. coli were grown to an absorbance at 590 nm (A5,0) 1.7 (about 8 x 108 cells/ml) and then quickly chilled. The cells were collected by centrifugation, resuspended in 50 mM Tris-hydrochloride (pH 7.5)-10% sucrose at 100 times the original concentration, and then quickly frozen in liquid N2. The cells were thawed at room temperature and incubated with 0.1 M NaCl and 200 ,ug of lysozyme per ml for 45 min at 0°C and for 3 min at 37°C. The lysate was quickly chilled to 0°C and centrifuged at 39,000 rpm at 0°C for 30 min in a Spinco 40 rotor. The resulting supernatant had an A280 of approximately 35 and a ratio of A260 to A2.8 of about 2.0. For some experiments the extracts were dialyzed for 22 h at 0°C against 1,000 volumes of 50 mM Tris-hydrochloride (pH 7.5)-0.1 M NaCl-10% sucrose to reduce the level of endogenous nucleoside triphosphates. The dialyzed extracts were stable to freezing and storage under liquid nitrogen. UV irradiation. DNA to be used as a substrate was irradiated at 50 to 100 ,ul at a time at a concentration of 0.5 to 1.0 mM in a shallow-well microscope slide with constant stirring at room temperature. The depth of liquid was less than 1 mm. A pair of germicidal lamps with an incident dose of 1.04 J/m2 per s provided the UV source. Reaction conditions for restoration of DNA. The standard reaction mixture of 0.1 ml contained 30 mM Tris-hydrochloride (pH 7.5), 25 mM MgCl2, 80
,uM deoxyribonucleoside triphosphate (dNTP), 0.25 mM NAD, 12 mM 2-mercaptoethanol, 3 nmol of DNA, and 40 ,ul of extract prepared as described above. In early experiments 0.3 mM dNTPs and all four ribonucleoside triphosphates (rNTPs) at 0.3 mM were included in the reaction mix. Further experiments showed that the rNTPs were not required and that high concentrations of dNT.P's were unnecessary. The reactions were routinely incubated for 15 min at 30°C and then stopped by adding 10 i1. of 0.5 M ethylenediaminetetraacetic acid (EDTA) and chilling to 0°C or (for isopycnic gradient analysis) by adding 1.0 ml of ice-cold NET buffer (0.1 M NaCl-0.01 M EDTA-0.01 M Tris-hydrochloride [pH 8.0]). Under the conditions described here, DNA from approximately 10 phage particles is present per unit of protein released from one cell. Other methods. Zone sedimentation in alkaline sucrose, isopycnic gradient analysis, and determination of radioactivity have been described previously (16). Numbers of single-strand breaks in DNA sedimented through alkaline sucrose were calculated from the formula: 2(1/MW,,,, - 1IM,e,), where U, UV and Mw,, are the weight average molecular weights of the irradiated and unirradiated samples. In this study CsCl gradients were collected from a hole punctured in the bottom of the centrifuge tube, and fractions were precipitated with ice-cold 1 N HCl-0. 1 M inorganic pyrophosphate before being filtered onto Reeve Angel glass-fiber filters soaked in 0.1 M inorganic pyrophosphate. Amounts of thymine dimers were determined by two-dimensional paper chromatography as described by Carrier and Setlow (5). The total number of pyrimidine dimers was calculated from the number of thymine-containing dimers by use of values obtained by Setlow and Carrier (29).
=
RESULTS Removal of damaged sites in vitro. According to current models (11), the initial step in the excision repair process is recognition of the offending lesion and introduction of a singlestrand incision near the site of the damage. Subsequent to incision, the offending lesion with an adjacent stretch of nucleotides is excised, the DNA in the damaged region is resynthesized, and the new patch of DNA is joined to the contiguous parental DNA strand. To see whether the excision repair process could occur in vitro, we prepared extracts of E. coli, using recB mutants to avoid nuclease activity by exonuclease V. Linear duplex DNA from bacteriophage T7 provided a suitable substrate. Our initial efforts to detect incision of UV-irradiated T7 DNA incubated with extracts of E. coli were not successful, even in the presence of ATP (to promote incision) (35) and nicotinamide mononucleotide (to prevent ligation). In toluenetreated cells, incision breaks accumulate only at a small fraction of the UV-induced lesions
VOL. 129, 1977
DNA REPAIR IN VITRO
(28, 30). If incision is similarly limited in crude extracts, it is not likely that a small number of single-strand breaks could be detected by alkaline sucrose gradient analysis oflow-molecularweight DNA. Therefore, in an effort to bypass the limiting incision step and to make possible a study of excision and resynthesis in vitro, an endonuclease from M. lueus that is specific for irradiated DNA was used. It was necessary to examine the properties of the M. luteus damage-specific endonuclease under our assay conditions. The enzyme was tested on unirradiated DNA and found free from nonspecific endonuclease activity. Also, by measuring the amount of pyrimidine dimers in irradiated DNA incubated in our standard reaction conditions, we confirmed that the M. luteus enzyme could not by itself excise thymine dimers. To see whether the number of incisions introduced into UV-irradiated DNA by the enzyme was proportional to the number of pyrimidine dimers, the experiment shown in Fig. 1 was performed. DNA was irradiated, incubated with the M. luteus enzyme, and sedimented through alkaline sucrose. Figure 1A shows that under our experimental conditions the reciprocal of the M,, ofthe treated DNA was 1 ,16
A
0
2-
1.,
0
.80..44 z
4-
.
0 B
cw 0.30
w
0.20
z
a:
0.1
0
20
40 60 80 UV DOSE (J/m2)
100
FIG. 1. Comparison of thymine dimer content and enzyme-sensitive sites. Portions (50 pl) of 3H-labeled T7 DNA were irradiated with various fluences of UV. Portions of this DNA were incubated with the M. luteus damage-specific endonuclease, as described in the text. The DNA was sedimented through alkaline sucrose, and the M, values were determined. (A) 2/Mw as a function of UV fluence. The rest of the DNA was examined for thymine dimer content by paper chromatography. (B) Percentage of thymine as dimers shown as a function of UV dose.
1417
proportional to the UV fluence in the range 0 to 70 J/m2 and that the enzyme produced about 0.015 incisions/J per m2 per 106 molecular weight. The irradiated DNA was also examined for thymine dimer content. Figure 1B shows that about 0.0037% of the thymine present was converted to thymine-containing dimers by 1 J of UV per M2. This corresponds to about 0.019 pyrimidine dimers/J per m2 per 106 molecular weight (about 0.46 dimers/J per m2 per T7 genome of 25 x 106 daltons). Thus, in view of this observation, together with an earlier report by R. B. Setlow, W. L. Carrier, and J. Stewart (Biophys. J. 15:194a, Abstr. no. TH-PM-H2, 1975), we will assume that this enzyme produces one incision per pyrimidine dimer. To test the ability of extracts of E. coli to restore the molecular weight of damaged DNA, the experiment shown in Fig. 2 was performed. Purified UV-irradiated DNA was incubated with an excess of the damage-specific nuclease. A portion of the reaction mixture was withdrawn and examined on alkaline sucrose gradients along with unirradiated control DNA. The remainder of the endonuclease-treated DNA was incubated with extract prepared from E. coli and then subjected to sedimentation analysis with alkaline sucrose. As shown in Fig. 2, incubation with the endonuclease caused an accumulation of single-strand breaks in the irradiated DNA. A second incubation with an extract from E. coli restored the DNA to the same molecular weight as the unirradiated control. It was conceivable that the restoration of the DNA to its original molecular weight might be caused by ligation between juxtaposed ends without removal of the lesion. Therefore it was necessary to test whether any sites sensitive to the damage-specific endonuclease remained after completion of the treatment shown in Fig. 2. T7 DNA was irradiated with 40 J/m2 and subjected to the same treatment described in the legend to Fig. 2. As a control, a portion of the irradiated DNA was diluted into a reaction mixture without endonuclease or extract. Both mixtures were then subjected to extraction with phenol and dialyzed. A portion of the DNA was retained as a control, and the remainder was incubated with the damage-specific endonuclease from M. luteus. Figure 3A shows that DNA not treated with the in vitro repair system still contains the expected number of sites sensitive to the endonuclease. However, Fig. 3B shows that the DNA incised and then restored to the original molecular weight has no remaining endonuclease-sensitive sites. This result has been confirmed by use of paper chromatography to measure dimer content (Carrier and
1418 MASKER
J. BACTERIOL.
DNA. This experiment confirms that there is no enzymatic activity in the M. luteus enzyme preparation that is required for the dimer excision, repair resynthesis, or ligation steps needed to restore the incised DNA. Repair replication. The in vitro system described here is potentially valuable for measuring repair resynthesis since both the specific activity of the nucleoside triphosphates and the number of repaired sites are known. Therefore,
10
N
O
16
1.0
0.4 Q8 0.6 0.2 RELATIVE DISTANCE SEDIMENTED
0
FIG. 2. Alkaline sucrose sedimentation of DNA repaired in vitro. 3H-labeled DNA from bacteriophage T7 was irradiated with 40 Jof UV per Mi2. A 3nmol amount each of irradiated and unirradiated DNA was treated with a damage-specific endonuclease from M. luteus and then layered on alkaline sucrose gradients. Identical samples received a second incubation with an assay mixture containing extract from strain HMS146 before being layered on gradients and spun at 49,000 rpm for 120 min. Profiles of radioactivity recovered from these gradients are shown. (A) Treatment with M. luteus enzyme only; (B) treatment with enzyme followed by treatment with E. coli extract. Symbols: 0, -UV; A, +UV.
Masker, unpublished data). We conclude that all of the damage recognized by the M. luteus enzyme has been removed by incubation with the E. coli extract. To be sure that possible contaminants in the damage-specific endonuclease were not responsible for the restoration of incised DNA, the experiment shown in Fig. 4 was performed. Irradiated DNA was treated with the M. luteus endonuclease and then extracted with phenol to inactivate the enzyme. When this nicked DNA was incubated with an E. coli extract under standard assay conditions, the DNA was restored to nearly its original molecular weight. After incubation, EDTA was added to a separate portion of repaired DNA to a concentration sufficient to inhibit repair but allow incision by the M. luteus endonuclease. The DNA was incubated a second time with the damage-specific endonuclease before being sedimented through alkaline sucrose. As seen in Fig. 4 no endonuclease-sensitive sites remain in the repaired
0
1.0
0.4 0.8 0.6 0.2 RELATIVE DISTANCE SEDIMENTED
0
FIG. 3. Complete removal of enzyme-sensitive sites by E. coli extract. 3H-labeled T7 DNA was exposed to 40 JIm2. A portion of the DNA was diluted and retained as a control. The remainder was treated first with the damage-specific endonuclease from M. luteus and then with an extract prepared from E. coli strain HMS146. Both portions of DNA were extracted with phenol and dialyzed against 10 mM Tris-hydrochloride (pH 7.5)-10 mMNaCl-1 mMEDTA. A part of each sample was retained as a control; the remainder was treated a second time with the M. luteus endonuclease. The reaction mixtures were layered on alkaline sucrose gradients and spun as described in the legend to Fig. 2. Profiles of radioactivity recovered from these gradients are shown. (A) No treatment before extraction with phenol; (B) treated with endonuclease and extract. Symbols: 0, no second treatment with enzyme; A, incubated a second time with damage-specific endonuclease.
VOL. 129, 1977
DNA REPAIR IN VITRO
1419
0 w
a:
U
0 o 20-
0io0
dO z U r
0.
oL
1.0
0.6 0.4 0.2 0.8 RELATIVE DISTANCE SEDIMENTED
0
FIG. 4. DNA repaired after inactivation of the damage-specific endonuclease. "C-labeled T7 DNA was exposed to 40 J of UV per m2, treated with M. luteus damage-specific endonuclease, and extracted with phenol. After dialysis, a portion of the DNA was treated with an E. coli extract, as described in the text. A 50 mM concentration ofEDTA was added to a sample of the restored DNA, and this sample was treated a second time with the M. luteus enzyme before all the reaction mixes were sedimented through alkaline sucrose. Symbols: 0, unirradiated control; A, no treatment with extract; O, treated with complete reaction mixture; O, reincubated with damage-specific endonuclease and EDTA. we attempted to measure repair replication performed by extracts of E. coli. When experiments such as the one shown in Fig. 2 were performed with the four dNTPs absent from the reaction mixture, complete repair still occurred, presumably because of nucleoside triphosphates still present in the extract. However, when extracts were dialyzed extensively and then used in an experiment such as the one described in the legend to Fig. 2, the results shown in Fig. 5 were obtained. When DNA damaged by exposure to 100 J of UV per m2 was treated with the damage-specific endonuclease and reincubated in a complete reaction mixture including dialyzed extract and an 80 juM concentration of each dNTP, almost complete restoration of the DNA to its original molecular weight was achieved. However, when the dNTP's were not present, the single-strand breaks introduced by the endonuclease remained in the DNA. An extract similar to the one used in the previous experiment was dialyzed to remove dNTP's and then utilized to measure repair resynthesis. In this experiment BrdUTP was used in place of thymidine 5'-triphosphate and the dATP was labeled with 32P. Although it seemed unlikely that any semiconservative replication of the T7 DNA could occur in these extracts, CsCl gradient analysis was performed to be sure that no aberrant initiation was occurring as a result of UV exposure. The results
shown in Fig. 6 indicate that there was no density shift of 32P-labeled DNA relative to 3Hlabeled DNA caused by incorporation of BrdUTP. This is the result expected if only short patches of DNA were synthesized to perform the repair seen in Fig. 5. The amount of repair resynthesis was linear with a UV dose of up to at least 150 J/m2, as shown in the insert to Fig. 6. Based on the amount of [32P]dATP incorporated, our measured value of 4.6 dimers/T7 genome per 10 J of UV per m2, and the fact that 1 nmol of DNA is equivalent to 7.5 x 109 T7 genomes, we arrive at an estimate of 17 nucleosides incorporated per pyrimidine dimer removed. Because of possible residual DNA precursors in the extracts, this value may be an underestimate. Role of DNA polymerase I. In vitro repair in the absence of DNA polymerase I was examined by using a polA12 mutant temperature sensitive for this enzyme (21, 34). DNA from bacteriophage T7 was irradiated with 50 J of UV per m2 and incubated with the damagespecific endonuclease. Reincubation with an extract prepared from strain HMS146 (recB) caused restoration of the DNA to its original size, irrespective of whether the incubation was carried out at 30 or 4200 (data not shown). Figure 7 shows the results of reincubation with an extract prepared from strain MM387 (polA12 recB). When the reincubation was carried out at 30°C, DNA polymerase I was active
J. BACTERIOL.
1420 MASKER
0.4 0.6 RELATIVE DISTANCE SEDIMENTED
FIG. 5. Repair completed by dialyzed extract ofE. coli. 3H-labeled T7 DNA was exposed to 100 J of UV per m2. A portion of unirradiated DNA was retained as a control. The irradiated DNA was treated with the damage-specific endonuclease, and a portion was saved as a control. The remaining incised DNA was incubated with an extract from strain HMS146 that had been dialyzed at 0°C for 22 h against 10% sucrose-.1 M NaCl,50 mM Tris-hydrochloride (pH 7.5). In one sample the reaction mixture was complete; in the other sample all four dNTPs were missing. All samples were layered on alkaline sucrose gradients and spun at 49,000 rpm for 120 min. Profiles of radioactivity recovered from these gradients are shown. Symbols: 0, control, no UV; A, UV, no reincubation; V, reincubation in complete assay mixture; O, reincubation in assay mixture without dNTP's.
and the DNA was restored to its original molecular weight. However, when DNA polymerase I was inactivated by heating it to 42°C, singlestrand breaks remained in the DNA after reincubation. As a control, 2 U of purified DNA polymerase I was included in one reaction reincubated at 420C. As seen in Fig. 7, the purified enzyme substituted for the thermally inactivated enzyme to restore the damaged DNA. Identical results were obtained when the experiment was repeated with dialyzed extracts. Apparently DNA polymerase I play an important role in the completion of repair we observe in vitro, at least when incubation occurs at 420C. When 1.2 mM ATP was included in the reaction mixture in an effort to stimulate repair resynthesis by DNA polymerases II and III (16), the strain MM387 extract was still unable to restore the DNA to its original molecular weight when the incubation was carried out at 420C. The results reported here agree with previously reported observations (16, 30) on repair in toluene-treated cells. DISCUSSION The experiments presented above show that purified bacteriophage T7 DNA damaged by UV radiation and incised enzymatically at the sites of pyrimidine dimers can be repaired in
vitro by extracts prepared from E. coli. The criteria used to monitor repair include restoration of the DNA to its original molecular weight, removal of sites recognized by a damage-specific nuclease, and measurement of UVstimulated repair resynthesis. By examining the ability of UV-irradiated T7 DNA to be packaged in vitro (26) and to form viable phage particles, we have also obtained preliminary results which show that the procedure described above increases the biological activity of irriadiated DNA (Kuemmerle and Masker, unpublished data). In the present study an endonuclease from M. luteus was used to overcome incision deficiency in the extracts and to promote incision near the sites of pyrimidine dimers. This enzyme may be identical (or similar) to UV-specific endonucleases isolated and characterized by a number of workers (4, 15, 23). Previous studies with the enzyme used in this study have shown that only the dimer-containing DNA strand is nicked (4) and one nick per pyrimidine dimer is introduced (Setlow et al., Biophys. J. 15:194a). The present study confirms that this enzyme produces about one single-strand break per pyrimidine dimer, shows no activity toward unirradiated DNA, and cannot by itself excise thymine dimers. Thus, this enzyme proved suit-
VOL. 129, 1977
DNA REPAIR IN VITRO
1421
10 8642-
E 0L.
E
0
O
0.0
to
to
0
C)
0
I 8- 100
Q-
J/m2
N to
OI
to 0
1 60
4-
2S*
0
9S0
*.I
i.w
I
20
-jO
25
5 30 0 15 20 25 30 10 FRACTION NO. FIG. 6. Repair resynthesis in vitro. 3H-labeled T7 DNA was irradiated with 0, 50, 100, or 150 J of UV per m2 before being treated with the damage-specific endonuclease. Six-nanomole samples were incubated in 0.2 ml of standard reaction mixtures, except that BrdUTP was used in place of thymidine 5'-triphosphate and [32P]dATP was included at 480 cpm/pmol. After 15 min of incubation at 30°C with an extract from strain HMS146, the reactions were stopped and subjected to isopycnic CsCl gradient analysis. Profiles of radioactivity recovered from these gradients are shown. Symbols: 0, 3H; A, 32p. The insert shows picomoles of [32P]deoxyadenosine 5'-monophosphate (P]dAMP) incorporated into acid-insoluble material as a function of the UV dose. These values were corrected for small (
