A cell cycle-associated pathway for repair of DNA-protein crosslinks in mammalian cells

75

Mutation Research, 183 (1987) 75-87 DNA Repair Reports Elsevier MTR06199

A cell cycle-associated pathway for repair of DNA-protein crosslinks in mammalian cells R. Gantt Laboratory of Cellular and Molecular Biology, National Cancer Institute, Building 37, 2D07, Bethesda, MD 20892 (U.S.A.) (Received 19 December 1985) (Revision received 28 August 1986) (Accepted 8 September 1986)

Keywords: Cell-cycle associated pathway; DNA-protein crosslinks, repair; trans-Platinum(II)diamminedichloride; Nucleotide excision pathway; DNA replication.

Summary Bulky adducts to DNA including DNA-protein crosslinks formed with trans-platinum(II)diamminedichloride are repaired largely by the nucleotide excision pathway in mammalian cells. The discovery in this laboratory that cells deficient in nucleotide excision repair, i.e., SV40-virus transformed SV-XP20S cells, can efficiently repair DNA-protein crosslinks implicates a second pathway. In this report, details concerning this pathway are presented. DNA-protein crosslinks induced with 20/~M trans-platinum were assayed by the membrane alkaline elution procedure of Kohn. DNA replication was measured by CsC1 gradient separation of newly synthesized DNA that had incorporated 5-bromodeoxyuridine. The following results indicate that this new repair pathway is associated with cell cycling: (A) Whereas rapidly proliferating human cells deficient in excision repair (SV40 transformed XP20S, group A) are proficient in repair of DNA-protein crosslinks, the more slowly growing untransformed parent line is deficient but can complete repair after prolonged periods of 4-6 days, the approximate doubling time of the cell population. (B) Either "used" culture medium or cycloheximide (1 /~g/ml) inhibits cell proliferation, protein synthesis, DNA replication and crosslink repair. In the presence of increasing concentrations of cycloheximide (0.01-5 #g/ml) the percent of DNA replication decreases and is essentially equivalent to the percent of crosslink repair. The following results indicate that this new repair pathway, though associated with cell cycling, is independent of DNA replication per se. (C) The rates of DNA-protein crosslink repair and DNA replication are essentially the same in mouse L1210 cells rapidly proliferating in 20% serum supplement; however, to slower proliferation rates in 1% serum rate of crosslink repair is slower but differs from that of DNA replication. (D) In the presence of aphidicolin (10 #g/ml) cells can repair DNA-protein crosslinks in virtually the complete absence of DNA replication, though the rate is slower in both nucleotide excision-proficient and Correspondence: Dr. Raymond Gantt, Laboratory of Cellular and Molecular Biology, National Cancer Institute, Building 37, 2DO7, Bethesda, MD 20892 (U.S.A.). 0167-8817/87/$03.50 © 1987 Elsevier Science Publishers B.V. (Biomedical Division)

76 -deficient cells. Thus, DNA replication is not essential for repair of DNA-protein crosslinks. (E) Comparison of the kinetics of replication and DNA-protein crosslink repair of pulse-labeled DNA indicates that, in the absence of metabolic inhibitors, repair of the crosslinks is independent of DNA replication per se and, therefore, DNA recombination events are not involved in this repair process. We conclude, therefore, that the new repair pathway is not coupled with DNA replication but is linked with cell cycling.

Since the discovery in 1962 that ultraviolet light causes DNA-protein crosslinks in bacterial (Alexander and Moroson, 1962) and mammalian (Smith, 1962) cells, various other agents including X-rays (Yamamoto, 1973), visible fluorescent light (Gantt et al., 1978, 1979), and numerous chemical carcinogens (Fornace and Little, 1979; Thomas, 1975) have been found to produce these lesions. When induced with trans-platinum(II)diamminedichloride (trans-platinum), DNA-protein crosslinks are reported to increase sister-chromatid exchanges (Bradley et al., 1979), transform 3T3 and 10T1/2 mouse cells (Fornace and Little, 1980) and to be repaired by the nucleotide excision pathway (Fornace and Seres, 1982). Recently, we found that a rapidly dividing line of xeroderma pigmentosum group A (XP-A) cells, line XP20S, transformed by simian virus 40 (SV40) repaired trans-platinum-induced DNA-protein crosslinks as competently as normal human foreskin fibroblasts (Gantt et al., 1984). These results were surprising in view of the known deficiency of XP-A cells in nucleotide excision repair and the report by Fornace and Seres (1982) that another XP-A cell line, XP12BE, is deficient in crosslink repair. We also observed that the slow-growing untransformed XP20S cells could complete repair of crosslinks only after prolonged periods of 4-6 days. If the SV40 virus does not directly introduce information for nucleotide excision repair, these findings suggest that DNA replication or cell cycling is an important factor in the repair of these lesions, since the rapid rate of cell cycling distinguishes the virus-transformed from untransformed cells. Conceivably, DNA replication per se could facilitate accessibility of repair enzymes to damaged sites in the chromatin (Cleaver, 1978); Mortelmans et al. (1976) have suggested, for example, that the repair deficiency in XP-A cells may result from an inability of the repair enzymes

to reach the damaged DNA in the chromatin. This idea has been both supported (Kano and Fujiwara, 1983) and contested (Helland et al., 1984). Alternatively, the SV40 virus may indirectly activate a function normally present in the cell. This alternative implicates a second pathway for repair of DNA-protein crosslinks in addition to the nucleotide excision mechanism as we presently understand it. Here we present evidence from studies with mouse and human cells that this second pathway for the repair of trans-platinuminduced DNA-protein crosslinks is not dependent on DNA replication but is linked with cell cycling. Materials and methods

Cell culture Suspension cultures of mouse leukemia cells, L1210 (a gift from Dr. Kurt Kohn, NCI), were carried without agitation in plastic T-25 flasks (Falcon) in RPMI 1630 medium (GIBCO, Grand Island, NY) supplemented with 20% fetal bovine serum (Flow Laboratories, McLean, VA), 100 International Units/ml penicillin and 100 /~g/ml streptomycin. At 3-4-day intervals the cultures were split 1 to 3 or 4 and grown at 37°C in an air atmosphere. At bi-monthly intervals stock cells were grown in antibiotic-free medium for 3 weeks and when tested for mycoplasma were found negative (Flow Laboratories) by direct and indirect tests. For alkaline elution experiments, cells which had grown to about 2 x 106/ml were adjusted to 5 × 105 cells/ml and incubated with either [2-14C]thymidine (0.4 /xCi/ml, 50 mCi/mmole) or [Me-3H]thymidine (0.8 /~Ci/ml, 160 mCi/mmole), overnight or for 72 h. After labeling, the cells were resuspended in thymidinefree medium (6 x 104 cells/ml), treated if appropriate, and 5 ml of cell suspension pipetted

77 into plastic T-25 flasks. When feasible, control cultures labeled with one isotope were mixed with treated cultures labeled with the other isotope and co-cultivated during the repair period; this procedure and the use of batch techniques for labeling DNA, treatment of cells with trans-platinum, used medium, aphidicolin, etc., and maintenance of 37 o C, significantly increases reproducibility. Numerous reversed isotope experiments show that an isotope effect does not influence the DNA elution rates under these conditions. Monolayer cultures of SV40-transformed XP2OS cells (Institute for Medical Research, Camden, N J, repository No. GM4312A) were carfled in T-25 flasks in Dulbecco's modified Eagle medium (MA Bioproducts, Walkersville, MD) containing 100/~g garamycin/ml (Schering-Plough Corp., Kenilworth, NJ) and supplemented with 10% fetal bovine serum (Flow Laboratories). The cells were subcultured 1:5 weekly and grown at 37°C with 10% CO2:air gas phase. Normal human foreskin fibroblasts were initiated and cultured as previously described (Reigner et al., 1976). For comparison of.DNA elution patterns, confluent T-25 cultures were harvested with 0.1% crystalline trypsin (Worthington Biochemical Corp., Freehold, N J) in 0.02% EDTA and planted in T-75 flasks. The planting medium was replaced with fresh medium 48 h later, and cells labeled with radioactive thymidine as above. After an additional 72 h the cells were harvested and replicate cultures prepared in thymidine-free medium at 3 × 105 cells per T-25 flasks.

Trans-platinum treatment A pool of L1210 cells, 6 × 104 cells per ml of complete medium were treated with 20/~M transplatinum (Vega Biochemicals, Tucson, AZ) by dilution of a 1 mM stock solution prepared in complete medium 30 min earlier at 37°C. Immediately after treatment, 5 ml of the suspension was pipetted into the appropriate number of T-25 flasks. The monolayer cultures of human fibroblasts were treated 24 h after planting by replacing the medium with complete medium containing 20 #M trans-platinum as above. The drug was not removed from the cells. The 20/~M concentration was chosen because it produces sufficient DNA-protein crosslinks for measurement but has

little or no toxic effect on L1210 mouse cells as reported by Zwelling et al. (1979) and confirmed in our laboratory for the cell types used. Cells and culture medium were shielded from light of < 500 nm wavelength.

Assay of DNA-protein crosslinks The membrane alkaline elution procedure of Kohn (Kohn et al., 1976, 1981) with minor modifications was used to measure crosslinks. Briefly, 14C- and 3H-prelabeled, trans-platinum-treated and untreated control cells were mixed placed on a polyvinyl chloride filter (BSWP02500, 2 /tm pore, Millipore, Bedford, MA) lysed with 5 ml of 20 mM sodium EDTA (pH 10.2) containing 0.3% sarkosyl (wt./vol.) and 2 M NaC1 followed by a burst of air forced thru the membrane with a syringe to fragment the DNA (Gantt et al., 1979). The DNA was then eluted at a flow rate of 0.33 ml/min with 20 mM EDTA-tetrapropylammonium hydroxide (free acid form of EDTA titrated to pH 12.2 with tetrapropylammonium hydroxide). Fractions were collected every 15 min for 3-4 h. Glacial acetic acid (0.1 ml) was added to each fraction and the lac and 3H were counted as a gel in 10 ml Ready-Solv MP (Beckman Instruments, Fullerton, CA) at 61% and 34% efficiency, respectively, in a Beckman LS250 scintillation counter. Nitrocellulose filters (Millipore, HAWP 02500, 0.45 /~m pore) were used in place of polyvinyl chloride filters for certain experiments as indicated. The lysis and elution fluid remained the same but the elution rate was changed to 0.05 ml/min and 1.5-ml fractions were collected and counted as a single phase in 10 ml Ready-Solve MP at efficiencies of 65% for 14C and 40% for 3H. The alkaline elution of DNA through nitrocellulose filters is not the same as through polyvinyl chloride filters. The most obvious difference is the convex elution profile through nitrocellulose versus the concave shape through polyvinyl chloride filters. However, the conditions chosen results in DNA elution differences between trans-platinum treated and untreated cells that are indistinguishable from those following elution through polyvinyl chloride filters. The elution rate of DNA from treated cells lysed on nitrocellulose filters in buffer containing proteinase K (0.4 mg/ml) is restored to that of un-

78 treated cells as it is with polyvinyl chloride filters. The radioactivity remaining on the filters was solubilized with dilute HC1 (0.5 M, 70°C for 30 min) and counted as described (Kohn et al., 1974). The cells and DNA were shielded from light after harvesting the cultures and during the elution procedure. The percent D N A - p r o t e i n crosslinks repaired was calculated as follows: Each elution filter of an experimental series contained a control and experimental D N A elution profile. To compensate for elution rate variations from individual filters in an experiment, the difference between control and experimental elution profiles was calculated at the lowest "percent DNA remaining" value (the Y coordinate) common to all control profiles in the experiment. The difference calculated from the first elution profile after trans-platinum treatment, when crosslinks were maximum, was set equal to 100%. The difference for the following time points was then expressed as a percent of this value. This procedure, thus, compares all elution profiles of an experiment at a point where the cell D N A of the controls have eluted to the same extent. The time of ehition varied at most by 30 min. The results of several filters were presented on one set of coordinates by first plotting the elution data obtained from the filter with the transplatinum treated and untreated control cells at zero time. Then the results from the other filters in the same series were plotted by taking the difference between their experimental and internal control profiles and adding this difference, point by point, to the untreated control profile of the zero time filter. No adjustments for filter variation were made.

Procedure for pulse-label experiments Stock cultures of L1210 mouse cells or XP12BE human cells were split (1 : 4) 24 h before labeling and then two T-150 plastic flasks were inoculated with 100 ml each of complete medium at 37°C containing 6 × 104 cells per ml. The flasks were continuously agitated on a rocker and all manipulations including centrifugation, medium change, and sampling were carried out at 37°C. Radioactive thymidine was added, x4C in one flask and 3H in the other, and 20/~M trans-platinum was added

to one of the flasks shortly thereafter in the case of a simultaneous labeling and treatment experiment. 1 h (or 2 h for XP12BE) later the radioactive medium was removed and the cells combined and suspended in a total of 200 ml of warm m e d i u m containing b r o m o d e o x y u r i d i n e (25 /~g/ml). In the case of delayed treatment, transplatinum was added to one flask 5 h (or 12 h for XP12BE) after the pulse label; the medium was changed on both the treated and untreated control. At appropriate times after trans-platinum treatment, 5-10 ml samples were removed and analyzed by a modified alkaline elution procedure ( G a n t t e t al., 1985) for D N A - p r o t e i n crosslinks, and an additional 5-10 ml sample was analyzed for DNA replication as described below.

Measurement of DNA replication To measure percent DNA replication and crosslink repair on the same cell suspension the culture volume (6 × 104 cells/ml) was doubled. Immediately after trans-platinum addition, 5bromodeoxyuridine (25/~g/ml) was added for the repair period. At harvest, half the cell pool was sued for measurement of crosslink repair and the other half was centrifuged at 600 × g for 10 min and the supernatant decanted. Then, 0.5 ml of 0.1 M Tris-HC1, pH 7.5, containing 0.3% sarkosyl was added and the sample stored overnight at - 2 0 ° C or until the end of the experiment. The samples were then thawed, heated in a water bath at 60 o C for 30 min, cooled and suspended in 5 ml of CsC1 (final density of 1.721 g / c m 3) in 0.1 M Tris-HC1, pH 7.5. 5 ml was then pipeted into a Polyallomer (Beckman Instruments) tube and the samples centrifuged in a SW50.1 rotor at 35000 rpm at 25 o C for 54-72 h. After centrifugation the tubes were tapped from the bottom, 10 drops collected per fraction, refractive index determined on every third tube, and all 24-28 fractions counted in 10 ml Ready-Solv MP + 1 ml water. Percent replication was calculated by summing the radioactivity in the heavier DNA peak, dividing by the radioactivity contained in both the heavy and light peaks and multiplying by 100. Incubation of cells with 5-bromodeoxyuridine does not effect the DNA ehition profiles under the conditions used.

79

Chemicals Aphidicolin (NSC No. 234714) was provided by the Natural Products Branch, NCI, Bethesda, MD. Sarkosyl (sodium lauryl sarcosinate) was purchased from Ciba-Geigy, Greensboro, NC. 5Bromodeoxyuridine and cycloheximide were purchased from Sigma Chemical Co., St. Louis, MO, tetrapropylammonium hydroxide (10% in water) from Eastman Kodak, Rochester, NY, and proteinase K from EM Biochemicals, Darmstadt (Germany).

Results In experiments on the crosslinks effects of both

cis- and trans-platinum, Zwelling et al. (1979) found that the trans-isomer formed largely DNA-protein crosslinks, whereas cis-platinum formed predominantly DNA-DNA crosslinks. They showed that at various concentrations of trans-platinum the relative retention of DNA from treated L1210 cells gradually approached that of untreated cells with time. Relative retention of DNA from cells treated with 20 #M trans-platinum was completely restored to the retention of untreated cells within 24 h. In this elution assay, the higher the molecular weight of the DNA or the more DNA with protein adducts, the greater the relative retention of DNA on the filter or the slower the elution rate. When the assay is carried out with proteinase K added to the lysis buffer, the elution rate of treated cells is immediately restored to that of the untreated controls. Using the air burst technique instead of X-irradiation to fragment the DNA, we corroborated the results of ZweUing et al. (1979) on cells treated with 20 #M and 30 #M trans-platinum and interpreted the findings to indicate that the lesions are DNA-protein crosslinks which can be repaired within 24 h. Fig. 1A shows the effect of trans-platinum treatment on the elution rate of DNA from mouse L1210 cells immediately after treatment and the restoration of the elution rate of the treated cells 24 h later indicating repair of the DNA-protein crosslinks; further when proteinase K is included in the lysis buffer the elution rates of DNA from treated and untreated cells are essentially identical (data not shown). Incidently, caffeine (200 #g/ml), which can have deleterious effects on DNA repair

processes, does not appear to affect the repair of DNA-protein crosslinks as shown in Fig. 1A. Additional experiments with caffeine concentrations up to 400 #g per ml also did not show repair inhibition though slight toxicity was observed (data not shown). Fig. 1B shows that trans-platinum treatment of L1210 cells results in a linear dose response up to 40 #M trans-platinum in the DNA-protein crosslink assay. This dose response is comparable to that obtained by Zwelling et al. (1979). However, it should be noted that crosslink yields are significantly dependent on the freshness of the medium at lower trans-platinum concentrations, presumably because of medium sulfhydryl group oxidation. In asynchronous cells populations, repair of DNA adducts usually starts rapidly and gradually slows down similar to a first order process. To examine the kinetics of DNA-protein crosslink repair, experiments were designed to measure on the same cell pool the percent crosslinks repaired and percent DNA replicated during the repair period. Fig. 2A shows that the rate of DNA-protein crosslink repair is essentially identical to the replication rate of DNA in the cell population having a doubling time of 12-15 h in 20% serum supplement. Moreover, the kinetics look essentially linear after a brief lag and repair goes to completion in about 15 h. This result suggests that crosslink repair and replication of DNA and/or cell cycling may be coupled. If coupling does occur, then increasing the doubling time of the cell population should reduce the rate of repair. Cells adapted to growth in 1% serum supplement have a population doubling time of 30 h. Repair of DNA-protein crosslinks in these slower growing cells took twice as long (30 h, Fig. 2B) as repair in the more rapidly growing cells (15 h, Fig. 2A). However, DNA replication took even longer in 15[ serum. After 54 h the DNA in trans-platinum treated cells was only 75% replicated and though the rate of replication was linear it was only half the DNA-protein crosslink repair rate. This result would suggest that repair need not be coupled to DNA replication, but interpretation is difficult because 20 #M trans-platinum is significantly more toxic to cells cultured in 1% rather than 20% serum supplement. DNA replication and cell cycling can be in-

80 I

I

I

I

I

1

95 I

0

85

0

l

0')

75 o rr

O

D

I

i

I

1

0

0

0

0

0

4)

<

i)

z

g

I

40

I=

E

E3

0

65

20 -

a

m

A 55

• I

I 6

0

I

I 12

I

0

I • 18

I

i

20

0

40

klM tr_ans-platinum

Hours

Fig. 1. Alkaline elution rates of D N A from L1210 cells treated with 20 # M trans-platinumwith and without 200 /xg/ml caffeine during the 24-h repair period and a tram-platinumdose response curve. Panel A: Cells at 6 × 1 0 4 / m l in complete medium, prelabeled with radioactive thymidine, were treated as a batch with trans-platinumand pipetted (5 ml) into T-25 flasks. At "zero" time, which is 4 h after treatment begins because crosslinks are then at a maximum, and again after 24 h, control cells and trans-platinumtreated cells were mixed, deposited on a polyvinyl chloride filter, lysed, and the D N A eluted at p H 12.1. trans-Platinum-treated( O ) and untreated (O) cells at 0 times; trans-platinum-treatedcells with caffeine (A) and without caffeine (zx) at 24 h. Panel B: Dose response of alkaline elution rates of D N A from L1210 cells 4 h after trans-platinumtreatment. 2 experiments each containing 3 elution profiles were averaged for the individual points; the bars represent the range of the values. Calculation of the Y coordinate, "Difference in % D N A Remaining" on the filter, is described in Materials and Methods.

I

100

I

I

I

'

~ //'¢

I

I(30

i

i

i

I

i

o/

,

o

/

/ !

// /

80

80

60

6(3

i//I

_66 l1/1///G/ 40

40

QcZ UO

/" Q /d / / ~

20

0 01 I

0

]

3

I

1

6

I

I

L

9 HOURS

I

12

t

I

15

J

I

18

I

I

I

I

10

20

30

40

i I50

I

HOURS

Fig. 22. Comparison of the rates of D N A replication and repair of D N A - p r o t e i n crosslinks by L1210 cells. Panel A: 20% serum supplement. Prelabeled cells were treated in batch at 6 X 104 cells/ml with trans-platinumand 5-bromodeoxyuridine (25 # g / m l ) and pipetted (5 ml) into T-25 flasks; all operations were carried out at 37°C. At appropriate times untreated control cells were mixed with the treated cells and half the mixture was assayed for D N A - p r o t e i n crosslinks by the m e m b r a n e alkaline elution technique and the other half was assayed for genome replication by equilibrium density centrifugation in CsCI. Calculations are described in Materials and methods. Panel B: 1% serum supplement. After 2 weeks adaptation of cells to growth in medium with 1% serum, 6 x 104 cells/ml were trans-platinumtreated as a batch in this medium from a 1.0 m M stock solution prepared, as usual, 30 rain previously in medium containing 20% serum supplement. All measurements and manipulations were identical to the experiments performed in Panal A. (O), % D N A - p r o t e i n crosslink repair; ([3), % D N A of the cell population that has replicated.

81

hibited by suspending proliferating cells in "conditioned" or used medium. If either of these processes is required for repair of DNA-protein crosslinks, then used medium should inhibit crosslink repair because cell proliferation is retarded under these conditions. Fig. 3A shows the alkaline elution profiles of cells treated and untreated as usual with tram-platinum, then resuspended in used medium and assayed for DNA-protein crosslinks 24 h later. It can be seen that used medium substantially inhibits repair of the DNA-protein crosslinks. However, we found that in addition to inhibition of cell proliferation, protein synthesis was significantly decreased after cells were suspended in used medium: 30% of normally proliferating cells soon after suspension and 23% of normal in cells 24 h after suspension in used medium (data not shown). Since the used suspension medium supports at least an additional doubling of the cell population, metabolic adaptation of the cells and not the lack of available energy or nutrients is probably responsible for repair inhibition. The protein synthesis inhibitor, cycloheximide, 100

'

°6O

also indirectly interferes with cell cycling and DNA replication. When L1210 cells were treated with 1 /~g cycloheximide per ml during the 24-h repair period, DNA-protein crosslink repair was again substantially inhibited (Fig. 3B), to an extent similar to that obtained with conditioned medium. At a concentration of 1 #g/ml, cycloheximide effectively blocks protein synthesis in L1210 cells ( < 10% of normal). These results suggest that either extent of protein synthesis or of DNA replication may correlate with the extent of DNA-protein crosslink repair. Table 1 shows the results of experiments in which inhibition of crosslink repair, inhibition of DNA replication, and inhibition of protein synthesis are compared in the presence of various concentrations of cycloheximide during the repair period. At 0.01 #g/ml, protein synthesis is inhibited 22-35% but there is no inhibition of either crosslink repair or DNA replication. At each of the subsequent cycloheximide concentrations there is an approximate equivalence between inhibition of crosslink repair and DNA replication, but not with inhibition of protein synthesis. These results suggest that DNA replication or

,

oooo

90 100

t

I

I

I

I

80 •o

c

"E E nr-

o

o

0

0

0

8

70

,¢ 0

0

0

A

•

A

60

Z El

0

z~

9t

60

z~

~t

40

A

•

I 50

] 0

|

]

6

[ 12

Hours

30 I 11~

0

[I

B I

I

I

6

I

12

I

dll JK.

18

Hours

Fig. 3. DNA elution rates from trans-platinum-treated L1210 cells cultured for 24 h in used medium or cycloheximide. Panel A: Repair in used medium. All cells were treated in fresh medium with trans-platinum. Used medium was prepared by centrifugation of cultures grown to a density of 106 cells/ml; a density of 2.0-2.5 × 106 cells/ml is attained when a culture is permitted to plateau. (O), (O), trans-platinum treated and untreated cells, respectively, at zero time; 24 h repair period of trans-platinum treated cells transferred to used medium (A) or fresh medium (zx). Panel B: Cycloheximide treatment. L1210 cell were treated with tram-platinum and cultured for 24 h in the presence or absence of cycloheximide. Cycloheximide (1 ~ g / m l ) was added at the beginning of trans-platinum treatment. (O), (O), trans-platinum treated and untreated cells, respectively, at zero time; tram-platinum treated cells with (A) or without (zx) cycloheximide for 24 h.

82 TABLE 1

i n h i b i t e d with aphidicolin. At a c o n c e n t r a t i o n of 10 F g / m l , aphidicolin inhibits t h y m i d i n e incorpor a t i o n more than 95% within 30 m i n a n d n o D N A h e a v y s t r a n d d u e to i n c o r p o r a t i o n of 5b r o m o d e o x y u r i d i n e is detectable b y CsC1 b a n d i n g after a 24-h i n c u b a t i o n . Repair was 68% a n d 97% complete after 19 h a n d 40 h, respectively (data n o t shown). This shows that m o u s e L1210 cells c a n repair D N A - p r o t e i n crosslinks in virtually the complete absence of D N A replication; though the rate is significantly slower, repair goes to completion. These results indicate that D N A replication is n o t essential for repair of D N A - p r o t e i n crosslinks. It has b e e n reported ( F o r n a c e a n d Seres, 1982) that D N A - p r o t e i n crosslinks in h u m a n cells are repaired b y the nucleotide excision pathway, a n d that XP12BE, group A, cells (unrelated to XP20S) are deficient in repairing these lesions. However, as n o t e d earlier, we f o u n d that XP20S, group A cells, t h o u g h clearly deficient in repair of D N A - p r o t e i n crosslinks, can eventually effect complete repair, a n d in addition, that the SV40-

CYCLOHEXIMIDE INHIBITION OF DNA-PROTEIN CROSSLINK REPAIR, DNA REPLICATION, AND PROTEIN SYNTHESIS IN L1210 CELLS Cycloheximide Percent inhibition a (/~g/ml) Crosslink Replication repair (%) of DNA (%)

Protein sysnthesis(%)

0.01 0.10 1.0 5.0

22-35 70-72 91-93 96-97

0 20- 40 70- 89 93-100

0 34-38 67-88 90-97

The values for crosslink repair and DNA replication were determined 24 h after 20 ~M trans-platinum treatment Cycloheximide was added at the beginning of treatment. Crosslinks were measured by alkaline elution of the same cell pool that had incorporated 5-bromodeoxyuridine for DNA replication measurements as described in the legend to Fig. 3. Inhibition of protein synthesis was determined from 14C-leucineincorporation after 6 h treatment with cycloheximide. a Data are presented as the range of at least 3 Expts. cell cycling is c o u p l e d in some way with D N A - p r o t e i n crosslink repair. T o test the effect of halting D N A synthesis o n repair of D N A - p r o t e i n crosslinks, replication was

,00 F . . . . . . . . . . . . . 9O

70

==80

!

'E

'~

,

#

J

60 55

' 0

4

8

Fraction

~

12

J 30

0

B

~

2

~,

'

6 Fraction

8

10

,

Fig. 4. Elution profiles of DNA from transformed XP fibroblasts and normal cells repairing DNA-protein crosslinks in the presence of aphidicolin. Panel A: SV40-XP20S, treated with 20 /~M trans-platinum and then cultured with aphidicolin (10 /tg/ml). trans-Platinum-treated (O) and untreated (O) cells with zero incubation time for repair, trans-Plafinum-treated cells cultured in the presence of aphidicolin for 48 h (~) and for 72 h (~). Panel B: Normal human foreskin fibroblasts were treated with 20 ~M trans-platinum and then cultured with aphidicolin (10 /~g/ml). trans-Plafinum-treated ([2) and untreated (O) cells with zero incubation time for repair, trans-Platinum-treated cells cultured in the presence of aphidicolin for 24 h (xT) and for 48 h (~,). Nitrocellulose filters were used for DNA elution instead of polyvinyl chloride filters (see Materials and Methods).

83

transformed derivative line is as repair-proficient as normal human foreskin fibroblasts (Gantt et al., 1984). These results were interpreted to mean that either the process of DNA replication compensates for the XP-A genetic defect, perhaps by making the lesion accessible (Mortelmans et al., 1976), or that repair is carried out by another pathway. If the process of DNA replication makes the lesion accessible to repair enzymes, then inhibition of DNA synthesis should result in concomitant inhibition of DNA-protein crosslink repair. To test the effect of inhibiting DNA replication on the repair of DNA-protein crosslinks in SV40-transformed XP20S, group A cells, cells were treated with 20/~M trans-platinum and aphidicolin was added shortly thereafter. Fig. 4A shows that nearly one-half the DNA-protein crosslinks were repaired after 48 h and by 72 h virtually all the crosslinks were repaired. In an identical experiment with normal human foreskin fibroblasts (Fig. 4B) which complete DNA-protein crosslink repair in 24-36 h, as do the SV40-transformed XP20S cells (data not shown) (Gantt et al., 1984), repair was complete at about 48 h in the presence of aphidicolin (Fig. 4B). Thus, aphidicolin retards repair of these trans-platinum-induced lesions in both nucleotide excision-proficient and -deficient cells, but it appears to retard the repair in the SV40-transformed XP20S cells significantly more than in the normal human foreskin fibroblast cells. This difference in crosslink removal in the presence of aphidicolin may result from the proficient nucleotide excision repair in the normal fibroblast cells. Snyder and Regan (1981) reported that aphidicolin addition to cells during repair of UV light damage to DNA results in a single-strand break at the site of damage. This phenomenon could account for the apparent repair of DNA-protein crosslinks in the presence of aphidicolin because production of single-strand breaks would increase the DNA elution rate without crosslink repair. To test this possibility trans-platinum treated and untreated control cells were incubated with aphidicolin (10 btg/ml) for 24 and 48 h and then assayed for repair of DNA-protein crosslinks with proteinase K (0.5 mg/ml) in the lysis buffer. The DNA elution profiles of the trans-platinum treated

and untreated cells were identical indicating that repair of the lesion had taken place in the presence of aphidicolin. If repair had not taken place and single-strand breaks were present at the site of the lesion, the DNA from the treated cells would have eluted faster than the untreated cells after proteinase K digestion. Because of the specificity limitations of the metabolic inhibitors aphidicolin and cycloheximide, and the possibility of unexpected interactions, another approach was used to investigate the relationship between DNA-protein crosslink repair and DNA replication. Rather than labeling the cellular DNA 16-36 h, a short pulse label (1 h) was given to mark all cells in S phase. DNA-protein crosslinks were induced either simultaneously with the pulse label or after a 5 h delay. Bromodeoxyuridine was introduced after removal of the trans-platinum so that DNA-protein crosslinks and DNA replication could be measured on the same cell pool at various intervals. This approach therefore, compares the kinetics of DNA-protein crosslink repair with time of DNA replication in the absence of metabolic inhibitors. When crosslinks were induced during the pulse label as shown in Fig. 5, repair proceeded linearly to completion and was about 90% com,

8G

i

~

ec

e¢ll~ 20

O~ 0

2

6

10

14

Hours

Fig. 5. Kinetics of DNA-protein crosslink repair and DNA replication in L1210 cells. Repair and replication rates are compared in a simultaneous experiment (open and filled circles) in which DNA-protein crosslinks are formed while DNA is pulse labeled for I h and in a delayed experiment (open and filled squares) in which DNA-protein crosslinks are formed 5 h after DNA is pulse-labeled. Percent repair of DNA-protein crosslinks for the simultaneous (O) and delayed tram-platinum treatment (D). Percent replication of the pulse-labeled DNA for the simultaneous (e) and delayed treatment (11).

84

plete before D N A replication began. In the experiment in which crosslinks were induced 5 h after pulse labeling there was a 3 - 4 h delay before D N A - p r o t e i n crosslink repair began and it then proceeded linearly at about the same rate as repair of crosslinks induced during the pulse label. However, DNA replication began somewhat before crosslink repair and proceeded substantially faster. At completion of D N A replication 55-60% of the D N A - p r o t e i n crosslinks had been repaired; at the time of 50% D N A replication, D N A - p r o t e i n crosslink repair was about 1% completed. These results clearly show that D N A - p r o t e i n crosslink repair in L1210 cells is not coupled with DNA replication per se. The results of similar experiments with cells from patients who have the hereditary disease xeroderma pigmentosum, group A, are consistent with the conclusion above, that D N A replication per se is not coupled with repair of transplatinum-induced DNA-protein crosslinks (Fig. 6). D N A replication was comparable to that of L1210 cells if allowance is made for the 26 h doubling time compared to the 12 h doubling time of L1210 cells. However, there was a striking difference in repair kinetics between the two cell lines. In contrast to the linear kinetics of L1210 cells, the XP12BE cells repair 40% to 60% of the D N A - p r o tein crosslinks and plateau for 10-12 h. This period of near zero repair is followed by an increase in rate to near the rate of the early time points. Another XP group A cell line, SV40-transformed XP20S, was chosen to repeat the comparison of simultaneous formation and delayed formation of D N A - p r o t e i n crosslinks relative to pulse labeling of the DNA. Unlike mouse L1210 and XP12BE cells which grow as suspension cultures, the SV40-transformed XP20S cells require a surface substrate for attachment. Experiments using replicate cultures of either the transformed or untransformed XP20S and normal foreskin fibroblasts resulted in poor reproducibility from flask to flask measuring radioactivity incorporated during pulse labels, percent D N A replicated, and percent repair of D N A - p r o t e i n crosslinks. To circumvent this problem, the SV40-transformed cells were detached from the surface and pulse-labeled, treated with trans-platinum and maintained in a

8O A

[]o o c~

40

20

o

6

v

12

18

24

30

36

Hours

Fig. 6. Kinetics of DNA-protein crosslink repair and DNA replication in XP12BE cells• Repair and replication rates are compared in a simultaneous experiment (open and filled circles) in which D N A - p r o t e i n crosslinks are formed while DNA is pulse-labeled for 2 h and in a delayed experiment (open and filled squares) in which DNA-protein crosslinks are formed 12 h after the DNA is pulse-labeled. Percent repair of D N A - p r o tein crosslinks after simultaneous (©) and after delayed (n) trans-platinum treatment. Percent replication of pulse labeled DNA for simultaneous (O) and for delayed (m) trans-platinum treatment of the cells. The open circles are the average of 2 Expts, and the error bars represent the range.

spinner culture to prevent re-attachment of cells. Samples were taken at intervals and assayed for D N A replication and D N A - p r o t e i n crosslink repair as for the L1210 and XP12BE suspension cultures. It was found that there was no replication of the labeled DNA but the D N A - p r o t e i n crosslinks formed during the pulse label were repaired rapidly to a plateau of about 85% within 16-20 h (data not shown). There was also a small lag suggesting a sigmoid curve. The repair kinetics of transformed XP20S cells do not show the twophased repair kinetics seen with the XP12BE cells; rather, they resemble those of the mouse L1210 cells. Attempts to measure replication and repair kinetics on cells treated with trans-platinum 12 h after the pulse label were unsuccessful; percent D N A replication, when measurable, was low and most of the radioactivity passed through the membrane filter during lysis of the cells indicating D N A fragmentation and cell death. Discussion

The fact that XP-A cells transformed by SV40 virus are as proficient as normal cells in crosslink repair, though still deficient in repair of UV lightinduced damage, provides direct evidence (Gantt

85 et al., 1984) that another pathway, in addition to nucleotide excision, can repair trans-platinum-induced DNA-protein crosslinks. This report presents additional results indicating that both mouse and human cells utilize this second pathway associated with cell cycling. Evidence that repair of these DNA-protein crosslinks is associated with cell cycling consists of the following: (A) Rapidly proliferating L1210 cells repair DNA-protein crosslinks at a rate indistinguishable from the rate of genome replication, i.e., cell doubling (Fig. 2A). (B) Reduction of the proliferation rate reduced the repair rate, though not to the level of genome replication (Fig. 2B) possibly because the trans-platinum is somewhat toxic in 1% serum supplement or, alternatively, excision repair makes a measurable contribution to the total repair in L1210 mouse cells. (C) Used medium and cycloheximide, both of which inhibit cell cycling, inhibit DNA-protein crosslink repair (Fig. 3). Furthermore, cycloheximide at various concentrations inhibits DNA synthesis, i.e., cell cycling, to different extents and the extent of inhibition is approximately equal to that of crosslink repair inhibition (Table 1). (D) Untransformed XP-A cells, though deficient in DNA-protein crosslink repair, completely repair these lesions after prolonged time periods of 4-6 days, the approximate doubling time of the cell population; further, virus-transformed derivate XP-A cells which proliferate rapidly are proficient at DNA-protein crosslink repair (Gantt et al., 1984) In spite of an apparent association of enhanced DNA repair in general, with DNA replication (Cleaver, 1978), the two processes can be completely dissociated in the case of trans-platinuminduced DNA-protein crosslinks. Not only does repair go to completion while replication is inhibited > 95% by aphidicolin (Fig. 4), repair of pulse-labeled DNA can be shown to be independent of replication per se in 3 cell lines, mouse L1210 and two XP group A lines, in the absence of any DNA synthesis inhibitors (Figs. 5,6). Thus, the repair mechanism, at least for these DNA-protein crosslinks, is unlikely to be a DNA recombination event. Also, the idea that removal of the various chromatin-associated proteins during DNA replication of XP-A cells, per se, pro-

vides the necessary access for repair of transplatinum-induced DNA-protein crosslinks seems incorrect. The XP12BE cells (Fig. 6) repair the DNA-protein crosslinks formed with pulselabeled DNA with similar kinetics whether the crosslinks are formed soon after DNA replication or 12 h later. Replication and repair are similarly dissociated in L1210 mouse cells (Fig. 5) and SV40-transformed XP20S cells though the kinetics differ in each case. To date, no inducible repair system has been unequivocally demonstrated in mammalian cells. It is not clear whether the enhancement of repair observed in various cell lines and animal tissues is due to induction of repair enzymes not ordinarily present in the cell or to the activation or increased synthesis of enzymes that are constitutive (for a review, Schendel, 1981). Recent reports implicate DNA repair activities that vary with the cell cycle in complex ways. For example, in growth arrested 10T1/2 mouse cells (Smith et al., 1981) N-3-methyladenine is lost rapidly during G1 and S phases following N-methyl-N'-nitro-N-nitrosoguanidine treatment, while N-7- and O-6-methylguanine are removed during the G1 phase but not during the S phase. In the small mammal, Microtus agrestis, UV-dependent incision changes rapidly thru the cell cycle after release from quiescence (Downes et al., 1982) and maximum incision potential occurs in late G1 phase. After growth arrestment, when normal human fibroblasts of line WI38 are treated with UV light or chemical carcinogens, DNA replication is preceded by enhancement of the nucleotide excision pathway followed by enhanced activity of base excision (Gupta and Sirover, 1981). Although enhanced activities are generally considered to be increases of constitutive proteins, they may include induction of new protein in some cases. The above considerations suggest a mechanism which seems to accommodate the results to date. Since extent of genome replication is approximately equivalent to extent of crosslink repair in the presence of various concentrations of cycloheximide, and since proliferation but not DNA replication is closely associated with repair, a possible explanation is that a cycle-dependent synthesis of protein(s) is required for this second pathway.

86 T h e second p a t h w a y p o s t u l a t e d here to be cell c y c l e - d e p e n d e n t for D N A - p r o t e i n crosslink rep a i r w o u l d not b e i n d u c i b l e in the sense of bacterial SOS repair in that d e l i b e r a t e D N A d a m a g e is not required. T h e p a t h w a y is p r e s e n t in S phase cells b e f o r e the crosslink lesions are i n t r o d u c e d b y t r a n s - p l a t i n u m t r e a t m e n t , b u t the cells n o t in S p h a s e at the time of t r e a t m e n t p r e s u m a b l y m u s t cycle to activate the p a t h w a y . W e infer this from the e x p e r i m e n t a l results of a d d i n g c y c l o h e x i m i d e o r a p h i d i c o l i n to the culture m e d i u m d u r i n g the r e p a i r p e r i o d as discussed below. A c c o r d i n g to this m e c h a n i s m , the results with c y c l o h e x i m i d e (i.e., that extent of g e n o m e replic a t i o n is a p p r o x i m a t e l y equivalent to extent of crosslink repair) a n d the results from the a p h i dicolin a n d p u l s e - l a b e l i n g e x p e r i m e n t s could be e x p l a i n e d as follows: cells that enter S p h a s e b e f o r e c y c l o h e x i m i d e or a p h i d i c o l i n a d d i t i o n w o u l d have synthesized the required protein(s) a n d thus c o m p l e t e crosslink repair. Those cells that have n o t reached S p h a s e when c y c l o h e x i m i d e is a d d e d w o u l d be u n a b l e to synthesize the p r o tein(s) a n d could n o t utilize this repair p a t h w a y . O n the o t h e r hand, those cells that are not in S p h a s e when a p h i d i c o l i n is a d d e d would c o n t i n u e to cycle p r e s u m a b l y to the G 1 - S p h a s e b o r d e r , have synthesized the protein(s), stop cycling but c o m p l e t e repair o f the lesions. T h e site of essential p r o t e i n synthesis for this p a t h w a y is p l a c e d near the G 1 - S p h a s e b o u n d a r y because in the presence of c y c l o h e x i m i d e the extent o f r e p a i r a p p r o x i m a t e s the extent of replication. R e p a i r w o u l d be e x p e c t e d to exceed replication to the extent that the time of p r o t e i n synthesis takes place n e a r e r the S - G 2 p h a s e b o u n d a r y . T h e second p a t h w a y w o u l d b e a c t i v a t e d irrespective of the presence of del i b e r a t e D N A d a m a g e ; it w o u l d represent one of the cell functions p r e p a r i n g for cell division, a n d it w o u l d serve to m o n i t o r the D N A before replic a t i o n to r e m o v e a n y crosslink lesions. T h e r e are several possible ways for cells to r e m o v e p r o t e i n crosslinks o t h e r than nucleotide excision. F o r example, a base with a d d u c t could be directly r e m o v e d b y a glycosylase; if the p r o tein a d d u c t is linked to deoxyribose, the nucleoside could be r e m o v e d b y p h o s p h o e s t e r a s e activity. Alternatively, the b u l k of the p r o t e i n could be cleaved from the b a s e or sugar b y a p r o t e a s e a n d

the p l a t i n u m - c o n t a i n i n g residue be r e p a i r e d b y the usual pathways. A n o t h e r possibility a n a l o g o u s to r e m o v a l of a d d u c t s b y O - 6 - m e t h y l g u a n i n e transferase is that p r o t e i n a d d u c t s m a y be removed b y a specific p r o t e i n acceptor(s) which then leave the b a s e intact in the D N A strand.

References Alexander, P., and H. Moroson (1962) Crosslinking of DNA to protein following ultraviolet irradiation of different cells, Nature (London), 194, 882-883. Bradley, M.O., I.C. Hsu and C.C. Harris (1979) Relationship between sister chromatid exchange and mutagenicity, toxicity, and DNA damage, Nature (London), 282, 318-320. Cleaver, J.E. (1978) DNA repair and its coupling to DNA replication in eukaryotic cells, Biochim. Biophys. Acta, 516, 489-516. Downes, C.S., R.T. Johnson and A.R.S Collins (1982) Cell cycle-dependent regulation of excision repair of UV damage, Exp. Cell Res., 142, 47-56. Fornace Jr., A.J., and J.B. Little (1979) DNA-protein crosslinking by chemical carcinogens in mammalian cells, Cancer Res., 39, 704-710. Fornace Jr., A.J., and J.B. Little (1980) Malignant transformation by the DNA-protein crosslinking agent trans-Pt(II)diamminedichloride, Carcinogenesis 1,989-994. Fornace Jr., A.J., and D.S. Seres (1982) Repair of transPt(II)diamminedichloride DNA protein crosslinks in normal and exicision-deficient human cells, Mutation Res., 94, 277-284. Gantt, R., R. Parshad, R.A.G. Ewig, K.K. Sanford, G.M. Jones, R.E. Tarone and K.W. Kohn (1978) Fluorescent light-induced DNA crosslinkage and chromatid breaks in mouse cells in culture, Proc. Natl. Acad. Sci. (U.S.A.), 75. 3809-3812. Gantt, R., G.M. Jones, E.V. Stephens, A.E. Baeck and K.K. Sanford (1979) Visible light-induced DNA crosslinks in cultured mouse and human cells, Biochim. Biophys. Acta, 565, 231-240. Gantt, R., W.G. Taylor, R.F. Camalier and E.V. Stephens (1984) Repair of DNA-protein crosslinks in an excisiondeficient human cell line and its SV40-transformed derivative, Cancer Res., 44, 1809-1812. Gantt, R., E.V. Stephens and S.R. Davis (1985) Measurement of DNA-protein crosslinks in mammalian cells without X-irradiation, Anal. Biochem., 149, 365-368. Gupta, P.K., and M.A. Sirover (1981) Cell cycle regulation of DNA repair in normal and repair deficient human cells, Chem.-Biol. Interactions, 36, 19-31. Helland, D., R. Kleppe, J.R. Lillehaug and K. Kleppe (1984) Xeroderma pigmentosum: in vitro complementation of DNA repair endonuclease, Carcinogenesis, 5, 833-836. Kano, Y., and Y. Fujiwara (1983) Defective thymine dimer excision from xeroderma pigmentosum chromatin and its characteristic catalysis by cell-free extracts, Carcinogenesis, 4, 1419-1424.

87 Kohn, K.W., C.A. Friedman, R.A.G. Ewig and Z.M. Igbal (1974) DNA chain growth during replication of asynchronous L1210 cells, Alkaline elution of large DNA segments from cells lysed on filters, Biochemistry, 13, 4134-4139. Kohn, K.W., L.C. Erickson, R.A.G. Ewig and C.A. Friedman (1976) Fractionation of DNA from mammalian cells by alkaline elution, Biochemistry, 15, 4629-4637. Kohn, K.W., R.A.G. Ewig, L.C. Erickson and L.A. Zwelling (1981) Measurement of strand breaks and crosslinks by alkaline elution, in: E. Friedberg and P. Hanawalt (Eds.), DNA Repair: a Laboratory Manual of Research Procedures, Vol. 1, Part B, Marcel Dekker, New York, pp. 379-401. Mortelmans, K., E.C. Friedberg, H. Slor, G. Thomas and J.E. Cleaver (1976) Defective thymine dimer excision by cell-free extracts of xeroderma pigmentosum cells, Proc. Natl. Acad. Sci. (U.S.A.), 73, 2757-2761. Reigner, D.A., T. McMichael, J.C. Berno and G.E. Milo (1976) Processing of human tissue to establish primary cultures in vitro, Tissue Culture Assoc. Manual, 2, 273-275. Schendel, P.F. (1981) Inducible repair systems and their implications for toxicology, CRC Crit. Rev. Toxicol., 8, 311-362. Smith, G.J., J.W. Grisham and D.G. Kaufman (1981) Cycledependent removal of certain methylated bases from DNA

of 10T1/2 cells treated with N-methyl-N'-nitro-N-nitrosoguanidine, Cancer Res., 41, 1373-1378. Smith, K.C. (1962) Dose dependent decrease in extractability of DNA from bacteria following irradiation with ultraviolet light or with visible light plus dye, Biochem. Biophys. Res. Commun., 8, 157-163. Snyder, R.D., and J.D. Regan (1981) Aphidicolin inhibits repair of DNA in UV-irradiated human fibroblasts, Biochem. Biophys. Res. Commun., 99, 1088-1094. Thomas, J.O. (1975) Chemically-induced DNA-protein crosslinks, in: K.C. Smith (Ed.), Aging, Carcinogenesis, and Radiation Biology, Plenum, New York, pp. 193-205. Yamamoto, O. (1973) Radiation-induced binding of nucleic acid constituents with protein constituents and with each other, Int. J. Radiat. Phys. Chem., 5, 213-229. Zwelling, L.A., T. Anderson and K.W. Kohn (1979a) DNAprotein and DNA interstrand crosslinking by cis- and trans-platinum(II)diamminedichloride in L1210 mouse leukemia cells and relation to cytotoxicity, Cancer Res., 39, 365-369. Zwelling, L.A., J. Filipski and K.W. Kohn (1979b) Effect of thiourea on survival and DNA crosslink formation in cells treated with platinum(II) complexes, L-phenylalanine mustard, and bis(2-chloroethyl)methylamine, Cancer Res., 39, 4989-4995.