DNA polymerase delta mediates excision repair in growing cells damaged with ultraviolet radiation DARELJ.
HUNTING' AND
BONNIEJ. GOWANS
Medical Research Council Group in the Radiation Sciences, Department of Nuclear Medicine and Radiobiology, University of Sherbrooke, Sherbrooke, Quebec, Canada J1H 5N4 AND
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by CONCORDIA UNIV on 11/11/14 For personal use only.
STEVEN L.
DRESLER~
Department of Pathology, Washington University School of Medicine, 660 South Euclid Avenue, St Louis, MO 63110. U.S.A. Received September 30, 1990 HUNTING,D. J., GOWANS, B. J., and DRESLER,S. L. 1991. DNA polymerase delta mediates excision repair in growing cells damaged with ultraviolet radiation. Biochem. Cell Biol. 69: 303-308. In confluent, stationary phase cells, an aphidicolin-sensitive DNA polymerase mediates UV-induced excision repair, but the situation in growing cells is still controversial. The sensitivity of repair synthesis to aphidicolin, an inhibitor of DNA polymerases alpha and delta, was determined in growth phase and confluent normal human fibroblasts (AG1518) using several techniques. Repair synthesis in confluent cells was always inhibited by aphidicolin, no matter which measurement technique was used. However, the inhibition of repair synthesis in growth-phase cells by aphidicolin was only detectable when techniques unaffected by changes in nucleotide metabolism were used. We conclude that UVinduced repair synthesis in growing cells is actually aphidicolin sensitive, but that this inhibition can be obscured by changes in nucleotide metabolism. Employing butylphenyl-deoxyguanosine triphosphate, a potent inhibitor of polymerase alpha and a weak inhibitor of delta, we have obtained evidence that polymerase delta is responsible for repair synthesis in growth-phase cells following UV irradiation. Key words: DNA polymerase, excision repair, ultraviolet radiation.
D. J., GOWANS, B. J., et DRESLER,S. L. 1991. DNA polymerase delta mediates excision repair in growing HUNTING, cells damaged with ultraviolet radiation. Biochem. Cell Biol. 69 : 303-308. Dans les cellules confluentes en phase stationnaire, une DNA polymerase sensible A l'aphidicoline permet la rkparation d'une excision induite par UV, mais la situation est encore controverske dans les cellules en croissance. Utilisant plusieurs techniques, nous avons dktermine la sensibilitk de la synthtse rkparatrice A l'aphidicoline, un inhibiteur des DNA polymkrases alpha et delta dans les fibroblastes (AG1518) humains normaux confluents et dans les fibroblastes en phase ' de croissance. La synthbe rkparatrice dans les cellules confluentes est toujours inhibke par l'aphidicoline, quelle que soit la technique de mesure utilide. Cependant, l'inhibition de la synthtse rkparatrice par l'aphidicoline dans les cellules en phase de croissance n'est dktectable que si nous utilisons des techniques non affectkes par des changements dans le mktabolisme des nuclkotides. Nous concluons que la synthbe rkparatrice induite par l'ultraviolet dans les cellules en croissance est vkritablement sensible a I'aphidicoline, mais que cette inhibition peut h r e occultke par des changements dans le mktabolisme des nuclkotides. Utilisant le butylphknyl-dkoxyguanosine triphosphate, un puissant inhibiteur de la polymerase alpha et un faible inhibiteur de la polymkrase delta, nous avons obtenu la preuve que la polymkrase delta est responsable de la synthtse rkparatrice dans les cellules en phase de croissance a p r b irradiation A l'ultraviolet. Mots cl6s : DNA polymtrase, rkparation d'une excision, radiations ultraviolettes. [Traduit par la rkdaction]
Introduction Numerous studies over the last several years have provided convincing evidence that the DNA polymerase responsible for the bulk of repair synthesis in stationary phase cells damaged by UV is sensitive to aphidicolin and arabinocytosine but not to dideoxythymidine (Hanaoka et al. 1979; Berger et al. 1979; Ciarrocchi et al. 1979; Snyder and Regan 1981; Tyrrell et al. 1985 and references therein). These data are consistent with the involvement of either DNA polymerase alpha or delta, but not beta in the UV repair process. Recent studies have implicated delta as the UV repair
polymerase in stationary phase cells (Dresler and Frattini 1986; Dresler and Kimbro 1987; Dresler et al. 1988b). The question of whether an aphidicolin-sensitive or -resistant polymerase mediates repair synthesis in growing cells is still controversial. Several studies have reported no effect of aphidicolin on repair synthesis in growing mammalian cells as measured by radioactive precursor incorporation (i.e., the BrdUrd density shift method and (or) unscheduled DNA synthesis) (Pedrali-Noy and Spadari 1980; Hardt et al. 1981; Giulotto and Mondello 1981; Bohr and Kober 1982; Downes and Collins 1982; Mullinger et al. 1983). When strand break accumulation and dimer removal were measured, Snyder ABBREVIATIONS: BrdUrd, 5-bromo-2'-deox~uridine;dThd, and Regan (1982) found that UV-induced repair in growthymidine; ~ A T P ,dCTP, dGTP, dTTP, and BrdUTP, the ing human fibroblasts was resistant to aphidicolin, while 5-triphosphates of 2'-deox~adenosine, 2'-deoxycytidine. 2'Collins (1983) found that both growing human fibroblasts deoxyguanosine, thymidine, and 5-bromo-2' -deoxyuridine, respectively; BuPh-dGTP, ~~-(~n-but~l~hen~l)-2'-deox~~uanosine-5'and HeLa were sensitive. 'OmWe have used several experimental approaches triphosphate; PBS, phosphate-buffered saline; EDTA, ethylpare the sensitivity of excision repair to aphidicolin in growth enediaminetetraacetic acid; rad, radiation absorbed dose. phase with that in growth-arrested normal human fibro' ~ u t h o rto whom correspondence should be addressed. '~eceased February 12, 1989. blasts. Our results are consistent with the hypothesis that
Printed in Caoada / Imprim6 au Canada
BIOCHEM. CELL BIOL. VOL. 69, 1991
304
excision repair synthesis in both confluent and growing human fibroblasts is mediated by an aphidicolin-sensitive polymerase. The apparent resistance of repair synthesis in growth-phase cells to aphidicolin results from changes in nucleotide metabolism. Further, by employing BuPh-dGTP, a potent inhibitor of polymerase alpha and a weak inhibitor of polymerase delta (Byrnes 1985), we have obtained evidence that polymerase delta is responsible for repair synthesis in UV-irradiated growth-phase cells.
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by CONCORDIA UNIV on 11/11/14 For personal use only.
Materials and methods Cell culture
Human diploid fibroblasts (AG1518; Institute for Medical Research) were grown in monolayers, labeled with [I4c]dThd (Amersham, 50-60 mCi/mmol; 1 Ci = 37 GBq) during log-phase growth and grown to confluence. In studies of growing cells, the fibroblasts were subcultured at a dilution of 1:3. incubated 24 h, labeled 24 h with 10 nCi/mL [I4C]dThD,then incubated 24 h in medium without radioactive precursor. Such cultures had reached 50-60% confluence and contained many mitotic cells. For autoradiography, cells were grown in plastic tissue culture plates containing ethanol-washed glass microscope slides, and were not labeled with [14c]dThD. Cells are damaged with UV radiation from a G15T8 germicidal lamp, as described previously (Hunting and Gowans 1988). Measurement of repair synthesis in intact cells using autoradiography
Cells, grown on glass microscope slides, were damaged with UV radiation and incubated for 2 h in medium containing 5 pCi/mL of [ 3 ~ ] d ~with h ~or, without aphidicolin. The slides were then washed with PBS in medium containing 400 pM dThD for 30 min at 37°C and again washed with PBS. The cells were fixed in 70% ethanol, dried, dipped in NBT-emulsion (Kodak; diluted 1:l with water), dried, and exposed for 2 or 7 days at 4OC. Slides were developed in D-19 (Kodak) and counter stained with Harris hematoxylin (Sheehan and Hrapchak 1980). Silver grains overlying 150-250 nuclei were counted for each sample. Repair synthesis was calculated as the difference between mean numbers of grains in non-S phase cells in damaged and undamaged samples. Measurement of repair synthesis in intact cells by BrdUrd density shift Cells were incubated for 3 h with 50 pM BrdUrd and for 30 min
with aphidicolin prior to irradiation. Following irradiation, cells were incubated in the same medium, with the further addition of [ 3 ~ ] d ~(40 h dpCi/mL), for 2 h. The cells were then harvested; the DNA was isolated and dissolved in alkaline CsCl(0.1 M KOH, 1.79 gm/mL CsCl) and centrifuged for ca. 18 h at 50 000 x g in a Beckman VTi 65.2 rotor. As a result of the large amount of replicative synthesis in growing cells, the replicative and parental DNA were not cleanly separated and therefore the peak of parental density DNA was recentrifuged under identical conditions. Measurement of repair synthesis in permeable cells
Intact cells were incubated with 50 pM BrdUrd for 30 rnin; aphidicolin, when used, was added 30 min prior to irradiation. Cells were either irradiated before or after permeabilization. When intact cells were irradiated, they were either harvested immediately or incubated at 37°C before harvesting and permeabilization. Cells were permeabilized and incubated as described previously (Dresler et al. 1982). Measurement of DNA strand breaks
A modified version of the alkaline elution method of Kohn et al. (1976) was used, as described previously (Hunting and Gowans 1988). Standard curves were constructed by irradiation of fibroblasts with Y-rays from a 6 0 ~ source o (gamma cell 200,
Atomic Energy of Canada) so that data could be expressed as rad equivalents. Results and discussion Although it is well established that aphidicolin, an inhibitor of DNA polymerases alpha and delta but not beta, inhibits DNA repair synthesis in growth-arrested cells following UV irradiation, the extent to which aphidicolin inhibits repair synthesis in growing cells may depend on the methodology employed (see Introduction). We have therefore examined the assumptions and results of several experimental approaches in an attempt to determine the involvement of polymerases alpha and (or) delta in excision repair in growing cells.
Autoradiography The advantage of determining DNA repair synthesis by autoradiography is that one can easily differentiate between repair and replicative synthesis even in growth-phase cultures, without the need for drugs such as hydroxyurea or BrdUrd. The main assumption of this technique is that thymidine metabolism is identical in control and experimental samples. Aphidicolin (65 pM) reduced incorporation of [ 3 ~ ] d ~ into h d the nuclear DNA of non-S phase cells to 19% of control in confluent cultures following UV irradiation. However, aphidicolin caused a slight stimulation of UV-induced t 3 ~ ] d ~ incorporation, hd to 110% of control, into non-S phase cells in growing cultures. The presence of S phase cells in cultures of both growing and confluent cells, albeit at very different frequencies, served as a positive control. Aphidicolin completely inhibited DNA replicative synthesis in both types of cultures, thus ruling out inactivation of aphidicolin by growing cells. Bromodeoxyuridine density shift Incorporation of the density label, BrdUrd, together with a radioactive label, [ 3 ~ ] d ~ hinto d , replicating DNA followed by isopycnic centrifugation in alkaline CsCl separates the denser replicative DNA from the parental DNA, thus allowing repair synthesis in the parental DNA to be quantitated. Aphidicolin, at both 6.5 and 65 pM, was a potent inhibitor of repair synthesis in confluent cells (34 and 18% of control, respectively); however, in growing cells it had no effect at 6.5 pM and actually stimulated repair synthesis at 65 pM to 138% of control. The low and high concentrations of aphidicolin inhibited replicative synthesis in the irradiated cells by 92 and 96070, respectively, indicating that the lack of potency of aphidicolin in growth-phase cells is specific to repair synthesis. The apparent lack of effect or slight stimulation in growing cells, and potent inhibition in confluent cells, of repair synthesis by aphidicolin as determined by autoradiography and the BrdUrd density shift method are in agreement with results from several laboratories (Pedrali-Noy and Spadari 1980; Hardt et al. 1981; Guillotto and Mondello 1981; Snyder and Regan, 1982; Bohr and Kgber 1982; Mullinger et al. 1983). DNA strand break accumulation During the excision repair process, DNA strand breaks are created by the incision step and removed by the ligation step, following synthesis of the repair patch. Given the rapid rate of incision following UV irradiation (see the review by Collins and Johnson 1984) and the short half-life of
,
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by CONCORDIA UNIV on 11/11/14 For personal use only.
F r a c t i o n Number
FIG. 1. DNA strand break accumulation induced by aphidicolin in confluent and growth-phase human fibroblasts. Exponentially growing cells (A) or confluent cells (B) were incubated with or without 6.5 pM aphidicolin or 10 rnM hydroxyurea plus 6.5 pM aphidicolin for 30 min and were then damaged with 5 J/m2 UV radiation. Following a further 15 rnin incubation in the presence or absence of drugs, the cells were harvested and DNA strand break frequencies were determined by alkaline elution (Methods and materials). Open symbols, 0 J/m2: 0,no drugs; 0 , aphidicolin; A, hydroxyurea plus aphidicolin. Closed symbols, 5 J/m2: l , no drugs; W , aphidicolin; A , hydroxyurea plus aphidicolin. Broken lines, 1000 rad Y radiation standards.
FIG. 2. Inhibition by aphidicolin of UV-induced DNA excision repair in permeable growth-phase human fibroblasts. Exponentially growing AG1518 fibroblasts, prelabeled with [14c]dT'hd and preincubated for 30 min in medium containing 50 pM BrdUrd, were made permeable (Materials and methods). Permeable cells were irradiated with 100 J/m2 and incubated for 15 min at 37°C with reaction mix containing 3 pM each of dATP, BrdUTP, dGTP, and [ C Y - ~ ~ P I ~ Cand T P ,either (A) no inhibitor or (B) 100 pM aphidicolin. DNA was isolated and analyzed by isopycnic centrifugation on alkaline CsCl gradients (Materials and methods). The gradients were fractionated and 14c( 0 ) and 3 2 (A) ~ radioactivity were measured. Net repair synthesis, determined b taking the difference between the specific dCMP incorporation ( P/'~c) in the parental DNA peaks of corresponding irradiated and unirradiated samples (repair synthesis), was 0.50 in the absence of inhibitor and 0.008 (2% of control) in the presence of 100 pM aphidicolin.
incomplete repair patches (3-10 rnin) (Erixon and Ahnstrom 1979; Hunting et al. 1985; Hunting and Gowans 1988), any treatment which results in an inhibition of ligation will cause a rapid accumulation of strand breaks. In addition to its sensitivity, the use of DNA strand break accumulation to measure the inhibition of excision repair has the further advantage of being unaffected by changes in nucleotide metabolism. As shown in Fig. 1, aphidicolin and aphidicolin plus hydroxyurea had little effect on strand break frequencies in the DNA of undamaged cells, whether confluent or growing. Note that strand breaks created in newly replicated DNA would not be detected, since this DNA would not be radioactively labeled. A 15-min incubation at 37OC following damage induced few breaks in the DNA of growth-phase cells in the absence of inhibitors (ca. 100 rad equivalents), while that of confluent cells contained ca. 650 rad equivalents, consistent with the burst of strand breaks previously noted in UV-irradiated confluent cells (Erixon and
Ahnstrom 1979; Hunting and Gowans 1988). A 15-min incubation of damaged cells with aphidicolin resulted in the accumulation of ca. 900 rad equivalents of breaks in growing cells and greater than 2000 rad equivalents in confluent cells. Thus, the additional breaks caused by aphidicolin were 800 and 1350 rad equivalents in growing and confluent cells, respectively. Hydroxyurea, an inhibitor of deoxyribonucleotide synthesis, has been found to enhance the inhibitory effect of aphidicolin on excision repair (Hunting et al. 1985; Dresler et al. 1988b), probably by lowering deoxyribonucleoside triphosphate concentrations. In the present experiment, hydroxyurea had little effect on the potency of aphidicolin in confluent cells, but did increase strand breaks by 30%, to 1200 rad equivalents, in growthphase cells. The rapid accumulation of strand breaks in both confluent and growing cells indicates that repair synthesis must be inhibited in both cell types by aphidicolin.
I 1.0 0.8
0.6 3~
0.4
02
FRACTION RETAINED
L
306
BIOCHEM. CELL BIOL. VOL. 69, 1991
1 0 04.
75.-
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by CONCORDIA UNIV on 11/11/14 For personal use only.
so--
irradiated (Fig. 2B) and unirradiated cells (data not shown) and reduced repair synthesis to 0% of control (Fig. 2B). This extent of inhibition is identical to that seen with permeable confluent cells under the same experimental conditions (see Fig. 3B). Thus, when deoxyribonucleotides are supplied exogenously, aphidicolin is a potent inhibitor of repair synthesis in both growing and confluent cells.
DNA repair synthesis in permeable cellsfollowing treatment of damaged intact cells with aphidicolin 25A In the experimental protocol described above, cells were permeabilized, then damaged, and therefore, both the incision and polymerization steps took place in the permeable 0 0 * cells. However, the permeable cell system can also be used 0.1 1 10 100 to estimate the relative number of incomplete repair sites [Aphidicolin] (W) accumulated in intact cells in the presence of inhibitors of repair synthesis (Snyder 1986). In the present experiment, intact cells were incubated with BrdUrd to maximize labeling of replicative DNA (see above), damaged with 12 ~ / UV m ~ radiation, incubated in the presence or absence of aphidicolin, then harvested, permeabilized and incubated in a reaction mix containing I ~ H I ~ C T BrdUTP, P, dGTP, dATP, and ATP. In both growth-phase and confluent cells, aphidicolin caused a rapid accumulation of incomplete repair sites in the intact cells, as shown by the increase in repair synthesis in the permeable cells (fivefold for growing and eightfold for confluent cells). The amount of this stimulation of repair synthesis was greatest in confluent cells in agreement with the DNA strand break data (Fig. l), and is consistent with a higher rate of incision in confluent cells. DNA polymerases alpha and delta DNA polymerases alpha and delta are both aphidicolin FIG. 3. Inhibition of UV-induced DNA repair synthesis and sensitive and have other similar properties as well as cerpurified DNA polymerases alpha and delta by aphidicolin and tain important differences, including the fact that delta has butylphenyl-dGTP in permeable growth-phase and confluent a 3 ' -5 ' exonuclease activity (Crute et al. 1986). BuPh-dGTP human fibroblasts. Cells were grown in glass roller bottles, is a potent inhibitor of polymerase alpha and a weak prelabeled with [I4c]d~hd,incubated for 30 min with 5 0 pM inhibitor of polymerase delta ( B y e s 1985; Lee et al. 1985); BrdUrd, harvested, made permeable, washed, irradiated with for example, 50% inhibition of purified polymerases alpha 100 ~ / m 'UV, incubated for 15 min with repair synthesis reaction and delta required 25 nM and 40 FM BuPh-dGTP, respecmix containing (A) 3 pM dATP, 3 pM dGTP, 3 pM BrdUTP, 3 pM [ 3 2 ~ ] dand ~ indicated ~ ~ , concentration of aphidicolin, or (B) tively (Dresler et al. 1988a). We therefore determined the 3 pM dATP, 3 pM dCTP, 3 pM BrdUTP, 0.3 pM [cY-~'P]~GTP, sensitivity of repair synthesis in permeable growth-phase cells and the indicated concentration of BuPhdGTP. DNA was isolated to BuPh-dGTP. As shown in Fig. 3B, the dose curve for and analyzed by isopycnic centrifugation in alkaline CsCI. Gra- inhibition of repair synthesis in permeable growth-phase cells dients were fractionated and radioactivity was determined. Repair synthesis is expressed as a percentage of repair synthesis in samples by BuPh-dGTP is virtually identical to that obtained previously with permeable confluent cells under identical condiincubated without inhibitor. Polymerase delta was isolated from tions (Dresler et al. 1988b). However, repair in both types calf thymus as described (Crute et al. 1986). Polymerase activity of cells is ca. 1/3000 as sensitive to BuPh-dGTP as is purified was assayed using a poly(dA)-oligo(dT) template-primer as described (Crute et al.). In Fig. 3B the symbols are as follows: 0 , DNA polymerase alpha. Figure 3A shows the dose curve purified pol alpha taken from Dresler et al. (1988b); ,purified for inhibition of repair synthesis in permeable growth-phase pol delta; A , repair synthesis in permeable growth-phase cells; A, cells by aphidicolin that is identical to that obtained previrepair synthesis in permeable confluent cells assayed under iden- ously with confluent cells (Dresler 1984). tical conditions, taken from Dresler et al. (1988b). Although several studies have reported approximately 80% inhibition of repair synthesis by aphidicolin and have proposed that polymerase beta is responsible for the other DNA repair synthesis in permeable cells 20% of UV-induced repair synthesis, our results (Fig. 3) The presence of substantial numbers of S phase cells in demonstrate that aphidicolin can completely inhibit repair the permeable growth-phase cultures necessitated the use of synthesis at concentrations that would not inhibit density labeling and isopycnic centrifugation to quantitate polymerase beta. We therefore conclude that polymerase repair synthesis (Fig. 2). Intact cells were incubated with delta is responsible for all excision repair synthesis after 50 pM BrdUrd for 30 min prior to permeabilization to maxdamage by UV radiation in growing human fibroblasts. imize labeling of replicative DNA and thus give clean separaThree hypotheses have been proposed to explain the tions of replicative and repaired DNA in the isopycnic denapparent lack of inhibition by aphidicolin of repair synthesis sity gradients (Materials and methods). Aphidicolin in growing cells, as measured by radioactive precursor incor(100 pM) completely inhibited replicative synthesis in both
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by CONCORDIA UNIV on 11/11/14 For personal use only.
NOTES
3W
tion and characterization of two new high molecular weight forms of DNA polymerase 6. Biochemistry, 25: 26-36. DOWNES,C.S., and COLLINS,A.R.S. 1982. Effects of DNA replication inhibitors on UV excision repair in synchronised human cells. Nucleic Acids Res. 10: 5357-5368. DRESLER,S.L. 1984. Comparative enzymology of ultravioletinduced DNA repair synthesis and semiconservative DNA replication in permeable diploid human fibroblats. J. Biol. Chem. 259: 13 947 - 13 952. DRESLER, S.L., and FRATTINI, M.G. 1986. DNA replication and UV-induced DNA repair synthesis in human fibroblasts are much less sensitive than DNA polymerase a! to inhibition by butylphenyl-deoxyguanosinetriphosphate. Nucleic Acids Res. 14: 7093-7102. DRESLER, S.L., and KIMBRO, K.S. 1987. 2',3'-Dideoxythymidine 5 ' -triphosphate inhibition of DNA replication and ultravioletinduced DNA repair synthesis in human cells: evidence for involvement of DNA polymerase delta. Biochemistry, 26: 2664-2668. DRESLER,S.L., ROBERTS,J.D., and LIEBERMAN, M.W. 1982. Characterization of deoxyribonucleic acid repair synthesis in permeable human fibroblasts. Biochemistry, 21: 2557-2564. DRESLER,S.L., ERATTINI,M.G., and ROBINSON-HILL, R.M. 1988a. In situ enzymology of DNA replication and ultravioletinduced DNA repair synthesis in permeable human cells. Biochemistry, 27: 7247-7254. DRESLER,S.L., GOWANS,B.J., ROBINSON-HILL, R.M., and HUNTING,D.J. 1988b. Involvement of DNA polymerase 6 in DNA repair synthesis in human fibroblasts at late times after ultraviolet irradiation. Biochemistry, 27: 6379-6383. ERIXON,K., and AHNSTROM, G. 1979. Single-strand breaks in DNA during repair of UV-induced damage in normal human and xeroderma pigmentosum cells as determined by alkaline DNA unwinding and hydroxylapatite chromatography. Mutat. Res. 59: 257-271. GIULOTTO, E., and MONDELLO, C. 1981. Aphidicolin does not inhibit the repair synthesis of mitotic chromosomes. Biochem. Biophys. Res. Commun. 99: 1287-1294. HANAOKA,F.. KATO, H., IKEGAMI,S., OHASHI,M., and YAMADA, M. 1979. Aphidicolin does inhibit repair replication in HeLa cells. Biochem. Biophys. Res. Commun. 87: 575-580. Acknowledgements HARDT,N., PEDRALY-NOY, G., FOCHER,F., and SPARADI,S. 1981. Aphidicolin does not inhibit DNA repair synthesis in This manuscript is dedicated t o the memory of Dr. Steven ultraviolet-irradiated HeLa cells. Biochem. J. 199: 453-455. L. Dresler. This work was supported by the Medical HUNTING,D., and GOWANS,B.J. 1988. Inhibition of repair Research Council of Canada. D.J. Hunting is a scholar of patch ligation by an inhibitor of Poly(ADP-ribose) synthesis in the Fonds d e la recherche en santk d u Quebec. normal human fibroblasts damaged with ultraviolet radiation. Mol. Pharmacol. 33: 358-362. BERGER, N.A., KUROHARA, K.K., PETZOLD, S.J., and SIKORSKI, HUNTING,D. J., DRESLER,S.L., and LIEBERMAN, M.W. 1985. G.W. 1979. Aphidicoli inhibits eukaryotic DNA repair. ImplicaMultiple conformational states of repair particles in chromatin tions for involvement of DNA polymerase a in both processes. during DNA excision repair. Biochemistry, 24: 3219-3225. Biochem. Biophys. Res. Commun. 89: 218-225. KOHN,K. W., ERICKSON, L.C., EWIG,R.A.G., and FRIEDMAN, BYRNES,J.J. 1985. Differential inhibitors of DNA polymerases C.A. 1976. Fractionation of DNA from mammalian cells by alpha and delta. Biochem. Biophys. Res. Commun. 132: alkaline elution. Biochemistry, 15: 4629-4637. 628-634. LEE, M.Y., TOOMEY, N.L., and WRIGHT,G.E. 1985. DifferenBOHR,V., and KOBER,L. 1982. The effect of aphidicolin on DNA tial inhibition of human placental DNA polymerases 6 and a by repair in resting and mitogen stimulated human lymphocytes. Biochem. Biophys. Res. Commun. 108: 797-803. BuPdGTP and BuAdATP. Nucleic Acids Res. 13: 8623-8630. CIARROCCHI, G., GRIEBROK JOSE,J., and LINN,S. 1979. Further MULLINGER, A.M., COLLINS, A.R.S., and JOHNSON, R.T. 1983. Cell growth state determines susceptibility of repair DNA syncharacterization of a cell-free system for measuring replicative thesis to inhibition by hydroxyurea and 1-P-D-arabinoand repair DNA synthesis with cultured human fibroblasts and furanosylcytosine. Carcinogenesis (London), 4: 1039-1043. evidence for the involvement of DNA polymerase a in DNA NICANDER, B.. and REICHARD, P. 1985. Evidence for the involverepair. Nucleic Acids Res. 7: 1205-1219. ment of substrate cycles in the regulation of deoxyribonucleoside COLLINS, A. 1983. DNA repair in ultraviolet-irradiated HeLa cells triphosphate pools in 3T6 cells. J. Biol. Chem. 260: 9216-9222. is disrupted by aphidicolin: the inhibition of repair need not imply PEDRALI-NOY, G., and SPARADI,S. 1980. Aphidicolin allows a the absence of repair synthesis. Biochim. Biophys. Acta, 741: 341-347. rapid and simple evaluation of DNA-repair synthesis in damaged human cells. Mutat. Res. 70: 389-394. COLLINS,A.R.S., and JOHNSON,R.T. 1984. The inhibition of DNA repair. Adv. Radiat. Biol. 11: 71-113. SHEEHAN, D.C., and HRAPCHAK, B.B. 1980. Theory and practice of histotechnology. C.V. Mosby Co., St-Louis. pp. 142-143. CRUTE,J.J., WAHL,A.F., and BAMBARA, R.A. 1986. Purifica-
poration: growing cells use a n aphidicolin-insensitive D N A polymerase (presumably polymerase beta) to perform repair synthesis (Pedrali-Noy and Spadari 1980; Hardt et al. 1981; Giulotto and Mondello 1981; Bohr and K#ber 1982); some characteristic of growing cells blocks the inhibitory effects of aphidicolin o n repair synthesis (e.g., high d C T P concentrations; Snyder and Regan 1982); or, aphidicolin actually inhibits repair synthesis in growing cells, but this inhibition is masked by compensatory changes i n nucleotide metabolism o r repair patch length (e.g., increased phosphorylation of [ 3 ~ ] d ~ hDownes d; and Collins 1982; Mullinger et al. 1983). All three hypotheses are consistent with the autoradiographic and BrdUrd density shift data from the present study. However, only the third hypothesis is consistent with our results: in both confluent and growing cells, aphidicolin inhibits repair synthesis in permeable cells, causes strand break accumulation in UV-irradiated intact cells, and stimulates repair synthesis in permeable cells following incubation of UV-irradiated intact cells with aphidicolin. T o mask the inhibition of repair synthesis by aphidicolin would require an increase in the specific activity of the d T T P pool. This could result from either increased phosphorylation of [ 3 ~ ] d ~ ho rd reduced synthesis of d T T P via ribonucleotide reductase and thymidylate synthetase. The latter possibility is supported by the observation that aphidicolin substantially inhibited the in situ activities of both ribonucleotide reductase and thymidylate synthetase in growing 3T3 cells (Nicander and Reichard 1985). The fact that the inhibition of repair synthesis by aphidicolin is masked only in growing cells may result from the much higher levels of deoxyribonucleotide synthesis in growing than in confluent cells. This possibility is consistent with the observed synergism between hydroxyurea, an inhibitor of deoxyribonucleotide synthesis, and aphidicolin in growing but not in confluent cells (Fig. 1; Snyder and Regan 1982).
308
BIOCHEM. CELL BIOL. VOL. 69, 1991
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by CONCORDIA UNIV on 11/11/14 For personal use only.
SNYDER,R.D. 1986. Evaluation of putative inhibitors of DNA excision repair in cultured human cells by the rapid nick translation assay. Mntat. Res. 173: 279-286. SNYDER, R.D., and REGAN, J.D. 1981. Aphidicolin inhibits repair of DNA in UV-irradiated human fibroblasts. Biochem. Biophys. Res. Commun. 99: 1088-1094. -1982. Differential responses of log and stationary phase
human fibrobla& to inhibition of DNA repair by aphidicolin. Biochim. Biophys. Acta, 697: 229-234. TYRRELL, R.M., KEY=, S.M.. AMAUDRUZ, F., and PIDOUX, M. 1985. Excision repair in U.V. (254 nm) damaged non-dividing human skin fibroblasts: a major biological role for DNA polymerase alpha. Int. J. Radiat. Biol. 48: 723-735.
HUNTING' AND
BONNIEJ. GOWANS
Medical Research Council Group in the Radiation Sciences, Department of Nuclear Medicine and Radiobiology, University of Sherbrooke, Sherbrooke, Quebec, Canada J1H 5N4 AND
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by CONCORDIA UNIV on 11/11/14 For personal use only.
STEVEN L.
DRESLER~
Department of Pathology, Washington University School of Medicine, 660 South Euclid Avenue, St Louis, MO 63110. U.S.A. Received September 30, 1990 HUNTING,D. J., GOWANS, B. J., and DRESLER,S. L. 1991. DNA polymerase delta mediates excision repair in growing cells damaged with ultraviolet radiation. Biochem. Cell Biol. 69: 303-308. In confluent, stationary phase cells, an aphidicolin-sensitive DNA polymerase mediates UV-induced excision repair, but the situation in growing cells is still controversial. The sensitivity of repair synthesis to aphidicolin, an inhibitor of DNA polymerases alpha and delta, was determined in growth phase and confluent normal human fibroblasts (AG1518) using several techniques. Repair synthesis in confluent cells was always inhibited by aphidicolin, no matter which measurement technique was used. However, the inhibition of repair synthesis in growth-phase cells by aphidicolin was only detectable when techniques unaffected by changes in nucleotide metabolism were used. We conclude that UVinduced repair synthesis in growing cells is actually aphidicolin sensitive, but that this inhibition can be obscured by changes in nucleotide metabolism. Employing butylphenyl-deoxyguanosine triphosphate, a potent inhibitor of polymerase alpha and a weak inhibitor of delta, we have obtained evidence that polymerase delta is responsible for repair synthesis in growth-phase cells following UV irradiation. Key words: DNA polymerase, excision repair, ultraviolet radiation.
D. J., GOWANS, B. J., et DRESLER,S. L. 1991. DNA polymerase delta mediates excision repair in growing HUNTING, cells damaged with ultraviolet radiation. Biochem. Cell Biol. 69 : 303-308. Dans les cellules confluentes en phase stationnaire, une DNA polymerase sensible A l'aphidicoline permet la rkparation d'une excision induite par UV, mais la situation est encore controverske dans les cellules en croissance. Utilisant plusieurs techniques, nous avons dktermine la sensibilitk de la synthtse rkparatrice A l'aphidicoline, un inhibiteur des DNA polymkrases alpha et delta dans les fibroblastes (AG1518) humains normaux confluents et dans les fibroblastes en phase ' de croissance. La synthbe rkparatrice dans les cellules confluentes est toujours inhibke par l'aphidicoline, quelle que soit la technique de mesure utilide. Cependant, l'inhibition de la synthtse rkparatrice par l'aphidicoline dans les cellules en phase de croissance n'est dktectable que si nous utilisons des techniques non affectkes par des changements dans le mktabolisme des nuclkotides. Nous concluons que la synthbe rkparatrice induite par l'ultraviolet dans les cellules en croissance est vkritablement sensible a I'aphidicoline, mais que cette inhibition peut h r e occultke par des changements dans le mktabolisme des nuclkotides. Utilisant le butylphknyl-dkoxyguanosine triphosphate, un puissant inhibiteur de la polymerase alpha et un faible inhibiteur de la polymkrase delta, nous avons obtenu la preuve que la polymkrase delta est responsable de la synthtse rkparatrice dans les cellules en phase de croissance a p r b irradiation A l'ultraviolet. Mots cl6s : DNA polymtrase, rkparation d'une excision, radiations ultraviolettes. [Traduit par la rkdaction]
Introduction Numerous studies over the last several years have provided convincing evidence that the DNA polymerase responsible for the bulk of repair synthesis in stationary phase cells damaged by UV is sensitive to aphidicolin and arabinocytosine but not to dideoxythymidine (Hanaoka et al. 1979; Berger et al. 1979; Ciarrocchi et al. 1979; Snyder and Regan 1981; Tyrrell et al. 1985 and references therein). These data are consistent with the involvement of either DNA polymerase alpha or delta, but not beta in the UV repair process. Recent studies have implicated delta as the UV repair
polymerase in stationary phase cells (Dresler and Frattini 1986; Dresler and Kimbro 1987; Dresler et al. 1988b). The question of whether an aphidicolin-sensitive or -resistant polymerase mediates repair synthesis in growing cells is still controversial. Several studies have reported no effect of aphidicolin on repair synthesis in growing mammalian cells as measured by radioactive precursor incorporation (i.e., the BrdUrd density shift method and (or) unscheduled DNA synthesis) (Pedrali-Noy and Spadari 1980; Hardt et al. 1981; Giulotto and Mondello 1981; Bohr and Kober 1982; Downes and Collins 1982; Mullinger et al. 1983). When strand break accumulation and dimer removal were measured, Snyder ABBREVIATIONS: BrdUrd, 5-bromo-2'-deox~uridine;dThd, and Regan (1982) found that UV-induced repair in growthymidine; ~ A T P ,dCTP, dGTP, dTTP, and BrdUTP, the ing human fibroblasts was resistant to aphidicolin, while 5-triphosphates of 2'-deox~adenosine, 2'-deoxycytidine. 2'Collins (1983) found that both growing human fibroblasts deoxyguanosine, thymidine, and 5-bromo-2' -deoxyuridine, respectively; BuPh-dGTP, ~~-(~n-but~l~hen~l)-2'-deox~~uanosine-5'and HeLa were sensitive. 'OmWe have used several experimental approaches triphosphate; PBS, phosphate-buffered saline; EDTA, ethylpare the sensitivity of excision repair to aphidicolin in growth enediaminetetraacetic acid; rad, radiation absorbed dose. phase with that in growth-arrested normal human fibro' ~ u t h o rto whom correspondence should be addressed. '~eceased February 12, 1989. blasts. Our results are consistent with the hypothesis that
Printed in Caoada / Imprim6 au Canada
BIOCHEM. CELL BIOL. VOL. 69, 1991
304
excision repair synthesis in both confluent and growing human fibroblasts is mediated by an aphidicolin-sensitive polymerase. The apparent resistance of repair synthesis in growth-phase cells to aphidicolin results from changes in nucleotide metabolism. Further, by employing BuPh-dGTP, a potent inhibitor of polymerase alpha and a weak inhibitor of polymerase delta (Byrnes 1985), we have obtained evidence that polymerase delta is responsible for repair synthesis in UV-irradiated growth-phase cells.
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by CONCORDIA UNIV on 11/11/14 For personal use only.
Materials and methods Cell culture
Human diploid fibroblasts (AG1518; Institute for Medical Research) were grown in monolayers, labeled with [I4c]dThd (Amersham, 50-60 mCi/mmol; 1 Ci = 37 GBq) during log-phase growth and grown to confluence. In studies of growing cells, the fibroblasts were subcultured at a dilution of 1:3. incubated 24 h, labeled 24 h with 10 nCi/mL [I4C]dThD,then incubated 24 h in medium without radioactive precursor. Such cultures had reached 50-60% confluence and contained many mitotic cells. For autoradiography, cells were grown in plastic tissue culture plates containing ethanol-washed glass microscope slides, and were not labeled with [14c]dThD. Cells are damaged with UV radiation from a G15T8 germicidal lamp, as described previously (Hunting and Gowans 1988). Measurement of repair synthesis in intact cells using autoradiography
Cells, grown on glass microscope slides, were damaged with UV radiation and incubated for 2 h in medium containing 5 pCi/mL of [ 3 ~ ] d ~with h ~or, without aphidicolin. The slides were then washed with PBS in medium containing 400 pM dThD for 30 min at 37°C and again washed with PBS. The cells were fixed in 70% ethanol, dried, dipped in NBT-emulsion (Kodak; diluted 1:l with water), dried, and exposed for 2 or 7 days at 4OC. Slides were developed in D-19 (Kodak) and counter stained with Harris hematoxylin (Sheehan and Hrapchak 1980). Silver grains overlying 150-250 nuclei were counted for each sample. Repair synthesis was calculated as the difference between mean numbers of grains in non-S phase cells in damaged and undamaged samples. Measurement of repair synthesis in intact cells by BrdUrd density shift Cells were incubated for 3 h with 50 pM BrdUrd and for 30 min
with aphidicolin prior to irradiation. Following irradiation, cells were incubated in the same medium, with the further addition of [ 3 ~ ] d ~(40 h dpCi/mL), for 2 h. The cells were then harvested; the DNA was isolated and dissolved in alkaline CsCl(0.1 M KOH, 1.79 gm/mL CsCl) and centrifuged for ca. 18 h at 50 000 x g in a Beckman VTi 65.2 rotor. As a result of the large amount of replicative synthesis in growing cells, the replicative and parental DNA were not cleanly separated and therefore the peak of parental density DNA was recentrifuged under identical conditions. Measurement of repair synthesis in permeable cells
Intact cells were incubated with 50 pM BrdUrd for 30 rnin; aphidicolin, when used, was added 30 min prior to irradiation. Cells were either irradiated before or after permeabilization. When intact cells were irradiated, they were either harvested immediately or incubated at 37°C before harvesting and permeabilization. Cells were permeabilized and incubated as described previously (Dresler et al. 1982). Measurement of DNA strand breaks
A modified version of the alkaline elution method of Kohn et al. (1976) was used, as described previously (Hunting and Gowans 1988). Standard curves were constructed by irradiation of fibroblasts with Y-rays from a 6 0 ~ source o (gamma cell 200,
Atomic Energy of Canada) so that data could be expressed as rad equivalents. Results and discussion Although it is well established that aphidicolin, an inhibitor of DNA polymerases alpha and delta but not beta, inhibits DNA repair synthesis in growth-arrested cells following UV irradiation, the extent to which aphidicolin inhibits repair synthesis in growing cells may depend on the methodology employed (see Introduction). We have therefore examined the assumptions and results of several experimental approaches in an attempt to determine the involvement of polymerases alpha and (or) delta in excision repair in growing cells.
Autoradiography The advantage of determining DNA repair synthesis by autoradiography is that one can easily differentiate between repair and replicative synthesis even in growth-phase cultures, without the need for drugs such as hydroxyurea or BrdUrd. The main assumption of this technique is that thymidine metabolism is identical in control and experimental samples. Aphidicolin (65 pM) reduced incorporation of [ 3 ~ ] d ~ into h d the nuclear DNA of non-S phase cells to 19% of control in confluent cultures following UV irradiation. However, aphidicolin caused a slight stimulation of UV-induced t 3 ~ ] d ~ incorporation, hd to 110% of control, into non-S phase cells in growing cultures. The presence of S phase cells in cultures of both growing and confluent cells, albeit at very different frequencies, served as a positive control. Aphidicolin completely inhibited DNA replicative synthesis in both types of cultures, thus ruling out inactivation of aphidicolin by growing cells. Bromodeoxyuridine density shift Incorporation of the density label, BrdUrd, together with a radioactive label, [ 3 ~ ] d ~ hinto d , replicating DNA followed by isopycnic centrifugation in alkaline CsCl separates the denser replicative DNA from the parental DNA, thus allowing repair synthesis in the parental DNA to be quantitated. Aphidicolin, at both 6.5 and 65 pM, was a potent inhibitor of repair synthesis in confluent cells (34 and 18% of control, respectively); however, in growing cells it had no effect at 6.5 pM and actually stimulated repair synthesis at 65 pM to 138% of control. The low and high concentrations of aphidicolin inhibited replicative synthesis in the irradiated cells by 92 and 96070, respectively, indicating that the lack of potency of aphidicolin in growth-phase cells is specific to repair synthesis. The apparent lack of effect or slight stimulation in growing cells, and potent inhibition in confluent cells, of repair synthesis by aphidicolin as determined by autoradiography and the BrdUrd density shift method are in agreement with results from several laboratories (Pedrali-Noy and Spadari 1980; Hardt et al. 1981; Guillotto and Mondello 1981; Snyder and Regan, 1982; Bohr and Kgber 1982; Mullinger et al. 1983). DNA strand break accumulation During the excision repair process, DNA strand breaks are created by the incision step and removed by the ligation step, following synthesis of the repair patch. Given the rapid rate of incision following UV irradiation (see the review by Collins and Johnson 1984) and the short half-life of
,
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by CONCORDIA UNIV on 11/11/14 For personal use only.
F r a c t i o n Number
FIG. 1. DNA strand break accumulation induced by aphidicolin in confluent and growth-phase human fibroblasts. Exponentially growing cells (A) or confluent cells (B) were incubated with or without 6.5 pM aphidicolin or 10 rnM hydroxyurea plus 6.5 pM aphidicolin for 30 min and were then damaged with 5 J/m2 UV radiation. Following a further 15 rnin incubation in the presence or absence of drugs, the cells were harvested and DNA strand break frequencies were determined by alkaline elution (Methods and materials). Open symbols, 0 J/m2: 0,no drugs; 0 , aphidicolin; A, hydroxyurea plus aphidicolin. Closed symbols, 5 J/m2: l , no drugs; W , aphidicolin; A , hydroxyurea plus aphidicolin. Broken lines, 1000 rad Y radiation standards.
FIG. 2. Inhibition by aphidicolin of UV-induced DNA excision repair in permeable growth-phase human fibroblasts. Exponentially growing AG1518 fibroblasts, prelabeled with [14c]dT'hd and preincubated for 30 min in medium containing 50 pM BrdUrd, were made permeable (Materials and methods). Permeable cells were irradiated with 100 J/m2 and incubated for 15 min at 37°C with reaction mix containing 3 pM each of dATP, BrdUTP, dGTP, and [ C Y - ~ ~ P I ~ Cand T P ,either (A) no inhibitor or (B) 100 pM aphidicolin. DNA was isolated and analyzed by isopycnic centrifugation on alkaline CsCl gradients (Materials and methods). The gradients were fractionated and 14c( 0 ) and 3 2 (A) ~ radioactivity were measured. Net repair synthesis, determined b taking the difference between the specific dCMP incorporation ( P/'~c) in the parental DNA peaks of corresponding irradiated and unirradiated samples (repair synthesis), was 0.50 in the absence of inhibitor and 0.008 (2% of control) in the presence of 100 pM aphidicolin.
incomplete repair patches (3-10 rnin) (Erixon and Ahnstrom 1979; Hunting et al. 1985; Hunting and Gowans 1988), any treatment which results in an inhibition of ligation will cause a rapid accumulation of strand breaks. In addition to its sensitivity, the use of DNA strand break accumulation to measure the inhibition of excision repair has the further advantage of being unaffected by changes in nucleotide metabolism. As shown in Fig. 1, aphidicolin and aphidicolin plus hydroxyurea had little effect on strand break frequencies in the DNA of undamaged cells, whether confluent or growing. Note that strand breaks created in newly replicated DNA would not be detected, since this DNA would not be radioactively labeled. A 15-min incubation at 37OC following damage induced few breaks in the DNA of growth-phase cells in the absence of inhibitors (ca. 100 rad equivalents), while that of confluent cells contained ca. 650 rad equivalents, consistent with the burst of strand breaks previously noted in UV-irradiated confluent cells (Erixon and
Ahnstrom 1979; Hunting and Gowans 1988). A 15-min incubation of damaged cells with aphidicolin resulted in the accumulation of ca. 900 rad equivalents of breaks in growing cells and greater than 2000 rad equivalents in confluent cells. Thus, the additional breaks caused by aphidicolin were 800 and 1350 rad equivalents in growing and confluent cells, respectively. Hydroxyurea, an inhibitor of deoxyribonucleotide synthesis, has been found to enhance the inhibitory effect of aphidicolin on excision repair (Hunting et al. 1985; Dresler et al. 1988b), probably by lowering deoxyribonucleoside triphosphate concentrations. In the present experiment, hydroxyurea had little effect on the potency of aphidicolin in confluent cells, but did increase strand breaks by 30%, to 1200 rad equivalents, in growthphase cells. The rapid accumulation of strand breaks in both confluent and growing cells indicates that repair synthesis must be inhibited in both cell types by aphidicolin.
I 1.0 0.8
0.6 3~
0.4
02
FRACTION RETAINED
L
306
BIOCHEM. CELL BIOL. VOL. 69, 1991
1 0 04.
75.-
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by CONCORDIA UNIV on 11/11/14 For personal use only.
so--
irradiated (Fig. 2B) and unirradiated cells (data not shown) and reduced repair synthesis to 0% of control (Fig. 2B). This extent of inhibition is identical to that seen with permeable confluent cells under the same experimental conditions (see Fig. 3B). Thus, when deoxyribonucleotides are supplied exogenously, aphidicolin is a potent inhibitor of repair synthesis in both growing and confluent cells.
DNA repair synthesis in permeable cellsfollowing treatment of damaged intact cells with aphidicolin 25A In the experimental protocol described above, cells were permeabilized, then damaged, and therefore, both the incision and polymerization steps took place in the permeable 0 0 * cells. However, the permeable cell system can also be used 0.1 1 10 100 to estimate the relative number of incomplete repair sites [Aphidicolin] (W) accumulated in intact cells in the presence of inhibitors of repair synthesis (Snyder 1986). In the present experiment, intact cells were incubated with BrdUrd to maximize labeling of replicative DNA (see above), damaged with 12 ~ / UV m ~ radiation, incubated in the presence or absence of aphidicolin, then harvested, permeabilized and incubated in a reaction mix containing I ~ H I ~ C T BrdUTP, P, dGTP, dATP, and ATP. In both growth-phase and confluent cells, aphidicolin caused a rapid accumulation of incomplete repair sites in the intact cells, as shown by the increase in repair synthesis in the permeable cells (fivefold for growing and eightfold for confluent cells). The amount of this stimulation of repair synthesis was greatest in confluent cells in agreement with the DNA strand break data (Fig. l), and is consistent with a higher rate of incision in confluent cells. DNA polymerases alpha and delta DNA polymerases alpha and delta are both aphidicolin FIG. 3. Inhibition of UV-induced DNA repair synthesis and sensitive and have other similar properties as well as cerpurified DNA polymerases alpha and delta by aphidicolin and tain important differences, including the fact that delta has butylphenyl-dGTP in permeable growth-phase and confluent a 3 ' -5 ' exonuclease activity (Crute et al. 1986). BuPh-dGTP human fibroblasts. Cells were grown in glass roller bottles, is a potent inhibitor of polymerase alpha and a weak prelabeled with [I4c]d~hd,incubated for 30 min with 5 0 pM inhibitor of polymerase delta ( B y e s 1985; Lee et al. 1985); BrdUrd, harvested, made permeable, washed, irradiated with for example, 50% inhibition of purified polymerases alpha 100 ~ / m 'UV, incubated for 15 min with repair synthesis reaction and delta required 25 nM and 40 FM BuPh-dGTP, respecmix containing (A) 3 pM dATP, 3 pM dGTP, 3 pM BrdUTP, 3 pM [ 3 2 ~ ] dand ~ indicated ~ ~ , concentration of aphidicolin, or (B) tively (Dresler et al. 1988a). We therefore determined the 3 pM dATP, 3 pM dCTP, 3 pM BrdUTP, 0.3 pM [cY-~'P]~GTP, sensitivity of repair synthesis in permeable growth-phase cells and the indicated concentration of BuPhdGTP. DNA was isolated to BuPh-dGTP. As shown in Fig. 3B, the dose curve for and analyzed by isopycnic centrifugation in alkaline CsCI. Gra- inhibition of repair synthesis in permeable growth-phase cells dients were fractionated and radioactivity was determined. Repair synthesis is expressed as a percentage of repair synthesis in samples by BuPh-dGTP is virtually identical to that obtained previously with permeable confluent cells under identical condiincubated without inhibitor. Polymerase delta was isolated from tions (Dresler et al. 1988b). However, repair in both types calf thymus as described (Crute et al. 1986). Polymerase activity of cells is ca. 1/3000 as sensitive to BuPh-dGTP as is purified was assayed using a poly(dA)-oligo(dT) template-primer as described (Crute et al.). In Fig. 3B the symbols are as follows: 0 , DNA polymerase alpha. Figure 3A shows the dose curve purified pol alpha taken from Dresler et al. (1988b); ,purified for inhibition of repair synthesis in permeable growth-phase pol delta; A , repair synthesis in permeable growth-phase cells; A, cells by aphidicolin that is identical to that obtained previrepair synthesis in permeable confluent cells assayed under iden- ously with confluent cells (Dresler 1984). tical conditions, taken from Dresler et al. (1988b). Although several studies have reported approximately 80% inhibition of repair synthesis by aphidicolin and have proposed that polymerase beta is responsible for the other DNA repair synthesis in permeable cells 20% of UV-induced repair synthesis, our results (Fig. 3) The presence of substantial numbers of S phase cells in demonstrate that aphidicolin can completely inhibit repair the permeable growth-phase cultures necessitated the use of synthesis at concentrations that would not inhibit density labeling and isopycnic centrifugation to quantitate polymerase beta. We therefore conclude that polymerase repair synthesis (Fig. 2). Intact cells were incubated with delta is responsible for all excision repair synthesis after 50 pM BrdUrd for 30 min prior to permeabilization to maxdamage by UV radiation in growing human fibroblasts. imize labeling of replicative DNA and thus give clean separaThree hypotheses have been proposed to explain the tions of replicative and repaired DNA in the isopycnic denapparent lack of inhibition by aphidicolin of repair synthesis sity gradients (Materials and methods). Aphidicolin in growing cells, as measured by radioactive precursor incor(100 pM) completely inhibited replicative synthesis in both
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by CONCORDIA UNIV on 11/11/14 For personal use only.
NOTES
3W
tion and characterization of two new high molecular weight forms of DNA polymerase 6. Biochemistry, 25: 26-36. DOWNES,C.S., and COLLINS,A.R.S. 1982. Effects of DNA replication inhibitors on UV excision repair in synchronised human cells. Nucleic Acids Res. 10: 5357-5368. DRESLER,S.L. 1984. Comparative enzymology of ultravioletinduced DNA repair synthesis and semiconservative DNA replication in permeable diploid human fibroblats. J. Biol. Chem. 259: 13 947 - 13 952. DRESLER, S.L., and FRATTINI, M.G. 1986. DNA replication and UV-induced DNA repair synthesis in human fibroblasts are much less sensitive than DNA polymerase a! to inhibition by butylphenyl-deoxyguanosinetriphosphate. Nucleic Acids Res. 14: 7093-7102. DRESLER, S.L., and KIMBRO, K.S. 1987. 2',3'-Dideoxythymidine 5 ' -triphosphate inhibition of DNA replication and ultravioletinduced DNA repair synthesis in human cells: evidence for involvement of DNA polymerase delta. Biochemistry, 26: 2664-2668. DRESLER,S.L., ROBERTS,J.D., and LIEBERMAN, M.W. 1982. Characterization of deoxyribonucleic acid repair synthesis in permeable human fibroblasts. Biochemistry, 21: 2557-2564. DRESLER,S.L., ERATTINI,M.G., and ROBINSON-HILL, R.M. 1988a. In situ enzymology of DNA replication and ultravioletinduced DNA repair synthesis in permeable human cells. Biochemistry, 27: 7247-7254. DRESLER,S.L., GOWANS,B.J., ROBINSON-HILL, R.M., and HUNTING,D.J. 1988b. Involvement of DNA polymerase 6 in DNA repair synthesis in human fibroblasts at late times after ultraviolet irradiation. Biochemistry, 27: 6379-6383. ERIXON,K., and AHNSTROM, G. 1979. Single-strand breaks in DNA during repair of UV-induced damage in normal human and xeroderma pigmentosum cells as determined by alkaline DNA unwinding and hydroxylapatite chromatography. Mutat. Res. 59: 257-271. GIULOTTO, E., and MONDELLO, C. 1981. Aphidicolin does not inhibit the repair synthesis of mitotic chromosomes. Biochem. Biophys. Res. Commun. 99: 1287-1294. HANAOKA,F.. KATO, H., IKEGAMI,S., OHASHI,M., and YAMADA, M. 1979. Aphidicolin does inhibit repair replication in HeLa cells. Biochem. Biophys. Res. Commun. 87: 575-580. Acknowledgements HARDT,N., PEDRALY-NOY, G., FOCHER,F., and SPARADI,S. 1981. Aphidicolin does not inhibit DNA repair synthesis in This manuscript is dedicated t o the memory of Dr. Steven ultraviolet-irradiated HeLa cells. Biochem. J. 199: 453-455. L. Dresler. This work was supported by the Medical HUNTING,D., and GOWANS,B.J. 1988. Inhibition of repair Research Council of Canada. D.J. Hunting is a scholar of patch ligation by an inhibitor of Poly(ADP-ribose) synthesis in the Fonds d e la recherche en santk d u Quebec. normal human fibroblasts damaged with ultraviolet radiation. Mol. Pharmacol. 33: 358-362. BERGER, N.A., KUROHARA, K.K., PETZOLD, S.J., and SIKORSKI, HUNTING,D. J., DRESLER,S.L., and LIEBERMAN, M.W. 1985. G.W. 1979. Aphidicoli inhibits eukaryotic DNA repair. ImplicaMultiple conformational states of repair particles in chromatin tions for involvement of DNA polymerase a in both processes. during DNA excision repair. Biochemistry, 24: 3219-3225. Biochem. Biophys. Res. Commun. 89: 218-225. KOHN,K. W., ERICKSON, L.C., EWIG,R.A.G., and FRIEDMAN, BYRNES,J.J. 1985. Differential inhibitors of DNA polymerases C.A. 1976. Fractionation of DNA from mammalian cells by alpha and delta. Biochem. Biophys. Res. Commun. 132: alkaline elution. Biochemistry, 15: 4629-4637. 628-634. LEE, M.Y., TOOMEY, N.L., and WRIGHT,G.E. 1985. DifferenBOHR,V., and KOBER,L. 1982. The effect of aphidicolin on DNA tial inhibition of human placental DNA polymerases 6 and a by repair in resting and mitogen stimulated human lymphocytes. Biochem. Biophys. Res. Commun. 108: 797-803. BuPdGTP and BuAdATP. Nucleic Acids Res. 13: 8623-8630. CIARROCCHI, G., GRIEBROK JOSE,J., and LINN,S. 1979. Further MULLINGER, A.M., COLLINS, A.R.S., and JOHNSON, R.T. 1983. Cell growth state determines susceptibility of repair DNA syncharacterization of a cell-free system for measuring replicative thesis to inhibition by hydroxyurea and 1-P-D-arabinoand repair DNA synthesis with cultured human fibroblasts and furanosylcytosine. Carcinogenesis (London), 4: 1039-1043. evidence for the involvement of DNA polymerase a in DNA NICANDER, B.. and REICHARD, P. 1985. Evidence for the involverepair. Nucleic Acids Res. 7: 1205-1219. ment of substrate cycles in the regulation of deoxyribonucleoside COLLINS, A. 1983. DNA repair in ultraviolet-irradiated HeLa cells triphosphate pools in 3T6 cells. J. Biol. Chem. 260: 9216-9222. is disrupted by aphidicolin: the inhibition of repair need not imply PEDRALI-NOY, G., and SPARADI,S. 1980. Aphidicolin allows a the absence of repair synthesis. Biochim. Biophys. Acta, 741: 341-347. rapid and simple evaluation of DNA-repair synthesis in damaged human cells. Mutat. Res. 70: 389-394. COLLINS,A.R.S., and JOHNSON,R.T. 1984. The inhibition of DNA repair. Adv. Radiat. Biol. 11: 71-113. SHEEHAN, D.C., and HRAPCHAK, B.B. 1980. Theory and practice of histotechnology. C.V. Mosby Co., St-Louis. pp. 142-143. CRUTE,J.J., WAHL,A.F., and BAMBARA, R.A. 1986. Purifica-
poration: growing cells use a n aphidicolin-insensitive D N A polymerase (presumably polymerase beta) to perform repair synthesis (Pedrali-Noy and Spadari 1980; Hardt et al. 1981; Giulotto and Mondello 1981; Bohr and K#ber 1982); some characteristic of growing cells blocks the inhibitory effects of aphidicolin o n repair synthesis (e.g., high d C T P concentrations; Snyder and Regan 1982); or, aphidicolin actually inhibits repair synthesis in growing cells, but this inhibition is masked by compensatory changes i n nucleotide metabolism o r repair patch length (e.g., increased phosphorylation of [ 3 ~ ] d ~ hDownes d; and Collins 1982; Mullinger et al. 1983). All three hypotheses are consistent with the autoradiographic and BrdUrd density shift data from the present study. However, only the third hypothesis is consistent with our results: in both confluent and growing cells, aphidicolin inhibits repair synthesis in permeable cells, causes strand break accumulation in UV-irradiated intact cells, and stimulates repair synthesis in permeable cells following incubation of UV-irradiated intact cells with aphidicolin. T o mask the inhibition of repair synthesis by aphidicolin would require an increase in the specific activity of the d T T P pool. This could result from either increased phosphorylation of [ 3 ~ ] d ~ ho rd reduced synthesis of d T T P via ribonucleotide reductase and thymidylate synthetase. The latter possibility is supported by the observation that aphidicolin substantially inhibited the in situ activities of both ribonucleotide reductase and thymidylate synthetase in growing 3T3 cells (Nicander and Reichard 1985). The fact that the inhibition of repair synthesis by aphidicolin is masked only in growing cells may result from the much higher levels of deoxyribonucleotide synthesis in growing than in confluent cells. This possibility is consistent with the observed synergism between hydroxyurea, an inhibitor of deoxyribonucleotide synthesis, and aphidicolin in growing but not in confluent cells (Fig. 1; Snyder and Regan 1982).
308
BIOCHEM. CELL BIOL. VOL. 69, 1991
Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by CONCORDIA UNIV on 11/11/14 For personal use only.
SNYDER,R.D. 1986. Evaluation of putative inhibitors of DNA excision repair in cultured human cells by the rapid nick translation assay. Mntat. Res. 173: 279-286. SNYDER, R.D., and REGAN, J.D. 1981. Aphidicolin inhibits repair of DNA in UV-irradiated human fibroblasts. Biochem. Biophys. Res. Commun. 99: 1088-1094. -1982. Differential responses of log and stationary phase
human fibrobla& to inhibition of DNA repair by aphidicolin. Biochim. Biophys. Acta, 697: 229-234. TYRRELL, R.M., KEY=, S.M.. AMAUDRUZ, F., and PIDOUX, M. 1985. Excision repair in U.V. (254 nm) damaged non-dividing human skin fibroblasts: a major biological role for DNA polymerase alpha. Int. J. Radiat. Biol. 48: 723-735.
