Differential responses of log and stationary phase human fibroblasts to inhibition of DNA repair by aphidicolin

Biochimica et Biophysica Acta, 697 (1982) 229-234

229

Elsevier Biomedical Press BBA 91060

DIFFERENTIAL RESPONSES OF LOG AND STATIONARY PHASE HUMAN FIBROBLASTS TO INHIBITION OF DNA REPAIR BY APHIDICOLIN R O N A L D D. S N Y D E R * and JAMES D. R E G A N

Umcersi O" of Tennessee-Oak Ridge Graduate School of Biomedical Sciences and Biology Division, P.O. Box Y, Oak Ridge National Laborato W, Oak Ridge, TN 37830 (U.S.A.) (Received October 19th, 1981 )

Key words: DNA repair," Aphidicofin; (Human fibroblast)

The tetracyclic diterpenoid, aphidicolin, is an effective inhibitor of DNA repair in human cells following ultraviolet irradiation. This inhibition is very efficient in confluent resting cells but not in rapidly cycling cells as measured by (1) analysis of DNA single-strand breaks by alkaline sucrose sedimentation, (2) chromatographic analysis of pyrimidine-dimer removal, and (3) repair replication using CsCI density centrifugation. The inhibition is reversed by deoxycytidine or thymidine but not by deoxyadenosine or deoxyguanosine during the repair period. The data suggest that differences in deoxynucleoside triphosphate pools between cycling and confluent resting cells determine the different efficacies of the agent in these two situations.

Introduction It is generally agreed that DNA polymerase a is the replicative enzyme in mammalian cells (Refs. 1-4, for review). The role of DNA polymerase fl is less well understood, but it has been implicated to be involved in DNA repair processes by several investigators [2,5,6]. One obvious approach to further elucidate the role(s) of DNA polymerase a in DNA metabolic processes is through the use of specific inhibitors of the enzyme. Aphidicolin has been shown to be an inhibitor of DNA polymerase a but not of 13 or "/polymerases from mammalian cells (Refs. 7, 11, 12 for review) and it was therefore of interest to determine the effects of this agent on DNA repair. We had earlier reported [13] * Current address: Dept. of Biochemistry, School of Hygiene and Public Health; Johns Hopkins University; Baltimore, M D 21205. U.S.A. By acceptance of this article, the publisher or recipient acknowledges the U.S. Government's right to retain a nonexclusive, royalty-free license in and to any copyright covering the article.

that aphidicolin does, in fact, inhibit repair in human cells following ultraviolet irradiation. This finding is consistent with some reports [14-16] but in apparent contradiction to others [17-19]. We have now determined that part of the reason for the above inconsistency could arise from the fact that the inhibition of repair by aphidicolin is very efficient in confluent non-cycling cells, but not in log phase cells. We report here the use of three different repair assays to substantiate this point, and on the reversal of this inhibition by deoxynucleosides.

Materials and Methods

Cell culturing and labeling. Normal human skin fibroblasts were grown in modified Eagle's medium supplemented with 15% fetal calf serum and nonessential amino acids (Gibco). The cells were maintained at 37°C in humidity- and carbon dioxide-controlled incubators. All experiments were conducted by inoculating approx. 5.0.104 cells into 60 mm plastic tissue culture dishes, allowing

230

48 h for growth, then labeling for 24 h with 1.9 C i / m m o l [3H]thymidine at 4t~Ci/ml (Amersham Corporation) in modified Eagle's medium with 10% calf serum. Alkaline sucrose sedimentation. Appropriately treated cells were scraped from dishes in cold saline/& 12% EDTA and 50/~1 cells were lysed for 1 h at room temperature in 200 /~1 1M N a O H overlayed on 5-20%, 4 ml alkaline sucrose gradients containing 2 M NaCI. Centrifugation was carried out in a Beckman SW56 rotor for 150 min at 30000 rpm (130000g). Gradients were fractionated by the filter strip method [20] and counted for radioactivity. Molecular weight determinations of the peak D N A fractions were made by computer. Dimer chromatography. Determinations of the pyrimidine dimer content of D N A were made by the two-dimensional paper chromatography technique as described previously [21]. CsCI repair replication. Repair replication in ultraviolet-irradiated cells was assayed by isopycnic CsC1 density gradient centrifugation and [3H]bromodeoxyuridine labeling [22]. Log phase or confluent cells growing in 50 mm flasks were labeled with 515 Ci,/mol [laC]thymidine at 0.25 ffCi/ml for I h, and then washed with medium and allowed to grow for approx. 24 h. 2 h prior to ultraviolet irradiation, cells were incubated in medium containing 3 # g / m l BrdUrd/0.25 ~ g / m l FUrd. Cells were irradiated with 20 J / m 2 254 nm light and at appropriate times after treatment were incubated for 3 h in medium containing 12 C i / m mol [3H]BrdUrd (Amersham C o r p o r a t i o n ) , / 3 / ~ g / m l non-radioactive BrdUrd/0.25 ~ g / m l FUrd. At the end of this labeling period, the cells were harvested and the DNA was isolated and centrifuged in alkaline CsCI gradients for 40-45 h at 37000 rpm in an SW40 rotor. DNA banding at normal density was pooled and rebanded until all apparent contributions from semi-conservative synthesis had been removed. The remaining 3Hlabeled DNA at normal density reflected repair replication and its specific activity (cpm//xg DNA) was calculated. Results

We have shown that aphidicolin is capable of inducing breaks in ultraviolet-irradiated cellular

~

8

i. I

x0

~

001

.

.

.

.

,

.... OI

10

, .~.

"~ ~00

1000

APHIDICOLIN CONCN /~.g,'ml)

Fig. I. Aphidicolin dose response for production of DNA single-strand breaks in log phase and confluent cells. Media was removed from dishes and cells were ultraviolet-irradiated as monolayers with 20 J / m 2 with a G.E. germicidal lamp with peak emission at 254 nm. Immediately after irradiation, cells were treated with various concentrations of aphidicolin. Following 3 h incubation at 37°C, cells were collected and subjected to alkaline sucrose sedimentation for DNA molecular weight determinations. & - A , confluent cells: • • . log phase cells.

D N A [13]. We extend those observations in Fig. 1 to indicate that the number of such strand breaks observed increases in a linear fashion as the concentration of aphidicolin increases. Moreover, there is a dramatic difference between the responses of cycling and non-cycling cells at a given concentration of aphidicolin. For example, whereas a treatment of 6/~g/ml aphidicolin yields only 0.4 breaks/10 s daltons DNA in cycling cells, this same treatment yields over 6 b r e a k s / 1 0 ~ daltons D N A in confluent resting cells. This treatment causes no breaks in non-irradiated control DNA [131. To study the effect of the addition of deoxynucleosides on this strand-break accumulation, ultraviolet-irradiated log phase or confluent cells were treated with a dose of aphidicolin that led to approx. 1 DNA strand break/10 8 daltons DNA and then incubated for 3 h in the presence of various concentrations of deoxynucleosides. Fig. 2 shows that production of strand breaks in both log and confluent cells by aphidicolin is reversible by deoxycytidine and that about a 4-fold lower concentration of deoxycytidine is needed in confluent rather than in log phase cells to achieve maximum reversal of the aphidicolin effect. It also appears that this reversal is less effective at higher deoxycytidine concentrations. In contrast, thymidine

231 I

8o!

=# i /\ ~=~o~ I X /

--c~cc

om

oz

0

2

4 79 15 4C COFJCrl (Fm)

30

dT CONCN(pm)

_Z

DEOXYNUGLEOSIDE CONCN(~rn)

Fig. 2. Effect of deoxycytidine on the accumulation of aphid•colin-induced DNA single-strand breaks. Cells were irradiated with 20 J / m 2 and were then incubated with 20 p,g/ml aphid•colin (log cells) or with 0.25 p,g/ml aphid•colin (confluent cells) for 3 h at 37°C in the presence of various concentrations of deoxycytidine (dC). Cells were then sedimented in alkaline sucrose and the molecular weights of the DNA were determined, m .... A, confluent cells; • 0 , log cells. DNA strand breaks in the absence of deoxycytidine were 1.5 and 2.8 for log and confluent cells, respectively. (Left-hand figure.) Fig. 3. Effect of thymidine on the accumulation of aphid•colin-induced DNA single-strand breaks. Cells were treated as described in the legend to Fig. 2 except that various concentrations of thymidine (dT) were added during the 3 h incubation. /', A, confluent cells: • O, log cells. (Middle-figure.) Fig. 4. Effect of deoxyadenosine and deoxyguanosine on the accumulation of aphid•colin-induced DNA single-strand breaks. Cells were treated as described in the legend to Fig. 2 except that deoxyadenosine (A A) and deoxyguanosine ( • • ) were added at various concentrations. (Right-hand figure.)

also seems to reverse the aphid•colin effect (Fig. 3) and does so in a direct dose-dependent fashion. Concentrations of thymidine up to 500 #M are no more effective than 30 t~M and do not reverse the deoxynucleoside effect as seen with deoxycytidine (data not shown). Fig. 4 shows that both deoxyguanosine and deoxyadenosine are ineffective

at reversing the effect of aphid•colin when tested from 0.25 to 500 #M. (Data shown only for confluent cells.) The apparent inhibition of repair by aphid•colin as detected by the accumulation of DNA strand breaks can be verified by direct analysis of loss of pyrimidine dimers from the DNA [ 13]. Fig. 5 shows

4O > 302 o

o o X4

8

,~

,~

~o

"

6o

APHIDICOLIN CONCN ( M g / m l )

Fig. 5. Removal of pyrimidine dimers from human cellular DNA in the presence of aphid•colin. Cells were labeled with [3H]thymidine and irradiated as described in the legend to Fig. I, The cells were then collected after 12 h incubation at 37°C in the presence of aphid•colin at various concentrations. Following two-dimensional paper chromatography, the percent dimers removed was calculated as described previously [21]. ZX ZX, confluent cells; • • , log phase cells. 20 J / m 2 ultraviolet irradiation produces 0.09% of total thymidine in dimer form.

-)

90

0

4

12 20 3 0 0 dC CONCN (p.m)

5

15 50 200 crr CONCN (ffm)

Fig. 6. effect of deoxycytidine (dC) and thymidine (dT) on the inhibition of dimer removal by aphid•colin in confluent cells. Cells were irradiated with 20 J / m 2 as described in the legend to Fig. 1 and were incubated for 12 h in the presence of aphid•colin (0.25 p g / m l , final concentration). A. • • , aphid•cofin plus dooxycytidine; O O, deoxycytidine alone. B. • 0 , aphid•colin plus thymidine; O O, thymidine alone.

232 TABLE I R E P A I R SYNTHESIS IN A P H I D I C O L I N - T R E A T E D LOG PHASE A N D C O N F L U E N T H U M A N FIBROBLAS fS Cells were irradiated With 20 J / m 2 ultraviolet Light as in the legend to Fig. I and I)NA repair replication wa~ monitored o~er the first 3 h following treatment by CsCI density centrifugation as described in Materials and Methods. n c c×pcrimcnt not conducted n.d. - not determined. Treatment

Ultraviolet Ultraviolet plus 6 / z g / m l aphidicolin Ultraviolet plus 20 # g / m l aphidicolin Ultraviolet plus 6 / ~ g / m l aphidicolin Ultraviolet plus 4 ~ M deoxycytidine No ultraviolet; 2 0 / ~ g / m l aphidicolin No ultraviolet; 4/~M deoxycytidine

Log

Confluent

cpm//~g D N A

~ Inhibition

cpm/t~g D N A

c~ Inhibition

550 460 410 n.c. n.c. n.c. n.c.

0 16 25 -

5(}0 290 n.c. 512 512

0 42

that the removal of dimers is greatly inhibited by low concentrations of aphidicolin in confluent resting cells, but that this inhibition is much less effective in log phase cells as would be predicted from the differential dose responses for strandbreak accumulation. Moreover, the effects of deoxycytidine at various concentrations in reversing aphidicolin-induced D N A strand breaks are also seen in the dimer analysis presented in Fig. 6A. Lower concentrations of deoxycytidine effectively reverse the aphidicolin block of dimer removal, but higher concentrations are less effective. A very different situation exists, however, in the case of thymidine. While dThy reduces the number of aphidicolin breaks (Fig. 3), this is not accompanied by a reversal of dimer removal inhibition (Fig. 6B) as is seen with deoxycytidine. In fact, thymidine itself blocks dimer removal. It thus seems likely that the observed reduction in aphidicolin breaks by thymidine is due to a block in the initiation of the repair itself leading to less total sites undergoing repair rather than to an interaction between aphidicolin and thymidine. Elsewhere we will examine the effect of thymidine on DNA repair in more detail (unpublished data). Table I shows the effect of aphidicolin on D N A repair replication in the first 3 h following ultraviolet irradiation of log and confluent cells. Again, it is clear that the inhibition is more efficient in non-cycling than in log-phase cells, and that deoxycytidine reverses this inhibition.

0 0

n.d. -

n.d.

-

Discussion

Previous studies examining the effect of aphidicolin on repair of ultraviolet-irradiated DNA in human cells have yielded conflicting results. In the work presented here, we show that these incons,istencies could arise in part from the use of cell ~ultures with different proportions of cycling cells. We have not determined the percentage of cycling cells in log and confluent cultures. It is obvious, however, that whereas a significant proportion of t h e cells in log cultures are cycling, a very small proportion are cycling in the confluent cultures since these latter cultures are used after being at confluency for at least a week. Consistent with the notion that the cycling state of the cells may play a crucial role in the response to aphidicolin is the observation that mitotic cells show no inhibition [19], but non-stimulated human lymphocytes are inhibited in ultraviolet repair by aphidicolin [15]. Also, inhibition of repair has been reported in non-confluent cultures treated with hydroxyurea prior to and during the repair period [14]. This suggests that pool sizes may be a determining factor for the efficacy of aphidicolin, as hydroxyurea is generally considered to be an agent that inhibits ribonucleotide reductase, thus lowering pools of deoxynucleoside triphosphates. We observe a significant enhancement of inhibition by aphidicolin in log phase cells and only a slight enhancement in confluent cells by hydroxy-

233

urea pretreatment (data not shown). In the present communication we show that the effects of aphidicolin may be partially reversed by the addition of deoxycytidine and thymidine, but not by deoxyadenosine or deoxyguanosine. Apparently, the thymidine effect is due to a general inhibition of repair resulting in a lowering of the number of sites undergoing repair rather than an interaction with the aphidicolin and is not dealt with further here. Deoxycytidine, however, seems to work in a manner similar to its reversal of the aphidicolin block of replicative DNA synthesis [18,23,24]. Thus, the reversal of both repair and replication blocks by aphidicolin proceeds solely through deoxycytidine antagonism and the case for a common DNA polymerase (a) action in both processes is strengthened. The rather curious decrease in effectiveness of deoxycytidine at higher concentrations is not completely understood. It could presumably occur by a net lowering of dCTP pools by feedback inhibition at higher deoxycytidine levels [25]. Further supporting the notion that dNTP pool sizes may dictate the efficiency of repair inhibition by aphidicolin is the discovery of a mutant mouse cell line resistant to aphidicolin and with an altered ribonucleotide reductase [26,27]. The effect of this mutation is to increase the pools of dCTP and dATP. An alternative explanation for the different responses of log and confluent cells to aphidicolin might center around the DNA polymerase levels. In log phase cells, polymerase a levels are very high, whereas in confluent cells this activity constitutes only about 5% of the total polymerase activity. One could suppose that there is too much a polymerase in log cells to be effectively inhibited by aphidicolin. Although this explanation seems somewhat unlikely, it is supported by the reports of mouse and Drosophila cell mutants that are resistant to aphidicolin apparently because of their increased DNA polymerase a activity [28,29]. It is thus unclear which, if either, of the two alternatives might be involved in producing the differential responses of log and confluent cells to aphidicolin. Nevertheless, the specificity of aphidicolin for DNA polymerase a and the similarity of the deoxynucleoside effect in reversing both repair and replication argue strongly that DNA

polymerase a is involved in DNA repair in human fibroblasts.

Acknowledgements The authors wish to acknowledge the helpful advice of Ben Van Houten, and Drs. R. Ley and W.E. Cohn for the critical reading of the .manuscript. We also wish to thank Dr. A.H. Todd, Pharmaceutical Division, Imperial Chemical Industries Ltd., Macclesfield, U.K., for the generous gift of aphidicolin. Research was sponsored jointly by the National Cancer Institute under Interagency agreement No. 40-5-63, and the Office of Health and Environmental Research, U.S. Department of Energy, under contract W-7405-eng-26 with the Union Carbide Corporation. R.D.S. is a postdoctoral investigator supported by training grant No. T32-CA-09336 from the National Cancer Institute.

References 1 Bollum, F.Y. (1975) Progr. Nucl. Acids Res. Mol. Biol. 15, 109-144 2 Hubscher, U., Kuenzle, C.C. and Spadari, S. (1979) Proc. Natl. Acad. Sci. U.S.A. 76, 2316-2320 3 Weissbach, A. (1977) Annu. Rev. Biochem. 46, 25-47 4 Falaschi, A. and Spadari, S. (1978) in DNA synthesis: present and future (Molineaux, J. and Kohiyama, M. eds.), pp. 487-515, Plenum Press, New York 5 Bertazzoni, U., Stefanini, M., Pedrali-Noy, G., Giulotto, E., Nuzzo, F., Falaschi, A. and Spadari, S. (1976) Proc. Natl. Acad. Sci. U.S.A. 73, 785-789 6 Waser, J., Hubscher, U., Kuenzle, C.C. and Spadari, S. (1979) Eur. J. Biochem. 97, 361-368 7 Ikegami, S., Taguchi, T., Ohashi, M., Ogura, M., Nagano, H. and Mano, Y. (1978) Nature (London) 275, 458-460 80hashi, M., Taguchi, T. and lkegami, S. (1978) Biochem. Biophys. Res. Commun. 82, 1084-1090 9 Dalziel, W., Hesp, B., Stevenson, K. and Jarvis, J.A.J. (1973) J. Chem. Soc. Perkins Trans. 1, 2841-2852 10 Pedrali-Noy, G. and Spadari, S. (1979) Biochem. Biophys. Res. Commun. 88, 1194-2002 11 Wist, E. and Prydz, H. (1979) Nucl. Acids Res. 6, 1583-1590 12 Huberman, J.A. (1981) Cell 23,647-648 13 Snyder, R.D. and Regan, J.D. (1981) Biochem. Biophys. Res. Commun. 99, 1088-1094 14 Hanaoka, F., Kato, H., Ikegami, S., Ohashi, M. and Yamada, M. (1979) Biochem. Biophys. Res. Commun. 87, 575-580 15 Berger, N., Kurohara, K., Petzold, S. and Sikorski, G. (1979) Biochem. Biophys. Res. Commun. 89, 218-225 16 Ciarrocchi, G., Jose, J.G. and Linn, S. (1979) Nucl. Acids Res. 7, 1205-1219

234 17 Pedrali-No~¢, G. and Spadari, S. (1980) Mutat. Res. 70, 389-394 18 Seki, S., Oda, T. and Ohashi, M. (1980) Biochim. Biophys. Acta 610, 413-420 19 Giulotto, E. and Mondello, C. (1981) Biochem. Biophys. Res. Commun. 99, 1287-1294 20 Carrier, W.L. and Setlow, R.B. (1971) Anal. Biochem. 43,427-432 21 Carrier, W.L. and Setlow, R.B. (1971) Methods Enzymol. 21,230-237 22 Cleaver, J.E. (1975) Methods Cancer Res. X I, 123-165 23 Pedrali-Noy, G. and Spadari, S. (1980) J. Virol. 36,457-464 24 Ogura, M., Suzuki-Hori, C. Nagano, H., Mano, Y. and

Ikegami, S. (1979) Eur. J. Biochem. 07, 6()3-607 25 de Saint Vincent, B.R:, Dechamps, M. and Buttin, G. (1980) J. Biol. Chem. 255, 162 167 26 Ayusawa, D., Iwata, K., Kozu, T., Ikegami, S. and Seno, Y. (1979) Biochem. Biophys. Res. Commun. 91,946 954 27 Ayusawa, D., Iwata, K. and Seno, T. (1981) Somat. Cell Genet. 7, 27-42 28 Nishimura, M., Yasuda, H., Ikegami, S., Ohashi, M. and Yamada, M. (1979) Biochem. Biophys. Res. Commun. 91, 939-945 29 Sugino, A. and Nakayama, K. (1980) Proc. Natl. Acad. Sci. U.S.A. 77, 7049-7053