DNA Repair in Human Cells Containing Photoadducts of 8-Methoxypsoralen Or Angelicin'

Home , Angelicin, Psoralen

[CANCER RESEARCH 40, 696-702, March 19801 0008-5472!80!0040-0000$02.00 DNA Repair in Human Cells Containing Photoadducts of 8-Methoxypsoralen or Angelicin'

James Kaye, Charles Allen Smith, and Philip C. Hanawalt

Department of Biological Sciences, Stanford University. Stanford, California 94305

ABSTRACT human subjects receiving photochemothenapy for psomiasis (31). Photoactivated 8-methoxypsoralen (8-MOP) has been Psomabensinwhich the 3 rings have a linear arrangement can proven to be clinically effective for a number of denmatobogical react with pynimidines at one (17, 18) or both (6, 8, 9) ends to conditions including lichen planus, mycosis fungoides, and form monofunctionabor bifunctional adducts (intenstrand cross psoniasis. 8-MOP forms two types of covalent photoproducts links), respectively. Psomabenssuchas angelicin, the rings of with DNA, monoadducts, and bifunctional adducts which cross which have an angular arrangement, are able to form only link the two DNA strands. Angelicin is a congener of 8-MOP monofunctionab adducts (1, 8) because the unneacted end is which forms only monoadducts. not properly oriented for reaction with a pynimidine in the other We have used the combined density and isotopic labeling strand of the native DNA helix. technique to study repair replication in cultured human fibro Since many agents which alter DNAstructure stimulate repair blasts treated with either of these compounds and exposed to responses in mammalian cells in vitro, we have examined the near-ultraviolet bight.In human dipboid fibroblasts (Wb-38), the repair replication response to psomalendamage and have at time course of repair replication for both compounds is similar. tempted to distinguish the effect of cross-links from that of Drug concentration and ultraviolet dose responses are also monoadducts by comparing the cellular response to 8-MOP similar for 8-MOP and angelicin. No repair replication was with that to angelicin. stimulated by either compound in xenodemmapigmentosum cells from Complementation Group A (XP12BE). These results MATERIALS AND METHODS suggest that repair replication in response to 8-MOP is primarily a response to monoadducts and that the enzymatic pathway Drugs. Crystalline angelicin was a generous gift from M. for this repair synthesis shares at least one step with the Ashwood-Smith, University of Victoria, British Columbia, Can pathway for repair of pynimidine dimens. ada. The sample was homogeneous by gas chromatognaphic Cross-link persistence in treated cells was assayed by use analysis, performed by I. Wilkinson, Stanford University Medi of the single-strand-specific S@nucleaseto digest DNA that did cabCenter. The crystals melted at 137â€”138Â°,in close agree not renature readily following heat denaturation. Partial removal ment with the reported value (14). 8-MOP was obtained from of cross-links was observed in normal, xenodenmapigmento Sigma Chemical Co. (St. Louis, Mo.). sum variant, and Fanconi's anemia fibmoblasts,but not in xer Cell Culture. Wl-38 normal human dipboid fibrobbasts, their oderma pigmentosum Group A cells. SV4O transformant VA-i 3, and fibroblasts from a patient with Fanconi's anemia (CAL 1196 from American Type Culture INTRODUCTION Collection, Rockvilbe,Md.) were grown at 37Â°in Eagle's basal medium supplemented with 10% fetal calf serum in a humidified The furocoumanins are a naturally occurring class of tnicyclic CO2atmosphere. XP cells (XP12BE, Complementation Group aromatic compounds which intercalate in DNA and form pho A, and XP4BE, variant) were grown in minimal essential me toadducts with the pynimidine bases when irradiated with UVA.2 dium with 20% fetal calf serum. T98G cells (30) obtained from Psoralen, the parent compound, was first isolated and purified G. Stein, University of Colorado, were grown in this medium from the seeds of an Indian plant, Psoralea corylifoila (13). with 5% fetal calf serum. Extracts from this and similar plants have been used for cen Drug Treatment and Irradiation. All solutions were prepared tunes together with sunlight for the treatment of certain skin and procedures carried out under yellow light (GE Buglight). 8- diseases (21). 8-MOP p.o. in combination with UVA exposure MOP or angebicinwas dissolved in 1 ml acetone; then PBS at forms the basis for the â€˜â€˜photochemothenapy',of psoniasis, 37Â°was added in at beast5-fold excess to give the desired currently under clinical trial in the United States (16, 20). The final concentration. The solutions were stirred for at beast30 properties, known effects, and uses of psonabenshave been mm in an uncovered beaker before use to allow evaporation of reviewed extensively (22, 26). Psonalens (plus UVA) are mu the acetone. Concentrations @50@tg/mIwereconsidered to be tagenic in bacteria (15) and carcinogenic in albino mice (11), saturated. When solutions were to be reused they were kept at and 8-MOP has recently been implicated as a carcinogen in â€”20Â°inthe dark, warmed to 37Â°,and stirred for at least 30 mm before use. Cells in 100-mm Petni dishes were washed with warm PBS ,ThisworkwassupportedbyGrantNP-i61fromtheAmericanCancerand then covered with 1 ml drug solution on PBS only and Society and by grants from NIH. 2 The abbreviations used are: UvA, near-ultraviolet light; 8-MOP, 8-methox placed in the 37Â°incubator for 5 mm, after which they were ypsoralen; XP, xeroderma pigmentosum; PBS, phosphate-buffered saline (0.14 irradiated under 2 parallel 15-watt black lights (GE Fl 5T8/ N NaCI:3 mMKCI:8 mMNa2HPO4:i5 [email protected],pH 7.5); dThd, deoxy BLB). 7.5 cm apart. Cells were irradiated through plastic Falcon thymidine; BrdUrd, bromodeoxyuridine. Received August 16, 1979; accepted November 26, 1979. 3003 Petnidish covers at a distance of 10 cm from the lamps.

696 CANCERRESEARCHVOL. 40

The output of these lamps was 20 J/sq m/sec as measured of linear 5 to 20% (w/v) sucrose gradients containing 0.1 N by a PUVA-meter (Herbert Waldmann, Schwenningen, Gem NaOH and 1 mMEDTA which were then centrifuged for 75 mm many). The spectral output measured in a Zeiss spectmopho at 30,000 rpm in an SW5O.i motorat 20Â°.Fractions were tometer was maximal at 360 nm, with less than 1% at lengths collected on paper strips and assayed for radioactivity as above 490 nm and below 300 nm. Following irradiation, the above. drug solution was removed, and the cells were washed with Digestion of DNA with Si Nuclease. DNA was diluted in a warm PBS before addition of medium on trypsin. Irradiation at 12- x 75-mmglass culture tube with single-strandedDNA in 254 nm utilized a Westinghouse IL 782-30 germicidal lamp. 10 mMTnis:i mMEDTAsothatthefinalconcentrationofsingle Semiconservative DNA Synthesis. Growing 32P-prelabeled stranded DNA was at least 20 @sg/ml,andthe final volume was VA13 cells were treated with drug and irradiated, then incu 160 @sb.Forty,zIof a 5X concentratedbuffer (0.9 M NaCI:0.3 bated in medium containing 2 @sCi/ml[3H]dThd(New England M sodium acetate, and 1 5 mM ZnCI2, pH 4) were added. S@ Nuclear Corp., Boston, Mass., 55 Ci/mmol) for 2 or 3 hn, nuclease (2 @sl)(SigmaChemical Co.; 100 units4tl) was then washed, trypsinized, and bysedin 0.5 ml 1 N NaOH. A portion added, the tubes were incubated in a 37Â°water bath for 30 of each bysatewas transferred to a plastic tube, and the DNA mm, then put in ice, and 10 ,zg calf thymus DNA were added. was precipitated with 10% CI3CCOOHand collected on Milbi DNA was precipitated with cold CI3CCOOH and collected on pore filters which were dried and assayed for radioactivity. glass fiber filters (Whatman GF/C, 2.4 cm) which were washed Repair Replication. Repair replication was measured as with 10 ml cold CI3CCOOHand then twice with 5 ml H2O. The described by Smith and Hanawalt (29). Confluent 32P-pmela filters were dried and assayed for radioactivity. The S1-resistant beled cells were incubated in medium containing I 0 zMBmdUmd fraction was calculated (except where specified otherwise) as and I @LMfluorodeoxyunidinefor 1 to 2 hn,then washed, treated the ratio of acid-precipitable radioactivity in denatured samples with psoralen or angebicin, and irradiated. The BndUmdand treated with S1to that in undenatured samples thus treated. S1 fluorodeoxyumidine-containing medium was then added back, treatment of these latter samples usually reduced the radioac with the addition of 5 @zCi/ml[3H]dThd(New England Nuclear tivity by about 10%. Corp., 55 Ci/mmol), and the cells were incubated at 37Â°.At sampling times, the cells were washed with PBS and bysedin RESULTS 0.5% sodium lauryl sulfate:i 0 mi,i Tnis:i 0 m@ EDTA. The lysates were incubated with proteinase K (100 @zg/ml;Beckman Inhibition of DNA Synthesis. We examined the effect on Instruments, Inc., Fullerton, Calif.) for 2 hr at 50Â°or overnight DNA synthesis in growing cells of angebicinand several 8-MOP at 37Â°and then passed rapidly through 25-gauge needles to concentrations with increasing doses of UVA irradiation (Chart shear the DNA. CsCI in 10 m@iTmis:1mt@iEDTAwas added to 1). No effect was seen with either drug or UVA irradiation @ form solutions with refractive indices of â€”1.401(p 1.718 g/ alone. However, considerable inhibition of DNA synthesis was ml), which were then centrifuged at 37,000 rpm for 40 hr at observed with low concentrations of 8-MOP and short expo 20Â°in a 5OTi rotor. Fractions were collected in plastic tubes, sure to UVA. As expected, the inhibition increased when either and aliquots were spotted on paper strips, which were washed drug concentration or dose of irradiation was increased. A in cold CI3CCOOH,ethanol, and acetone and then dried and saturated solution of angelicin also resulted in large inhibition assayed for radioactivity. Fractions comprising the bightpamen of DNA synthesis when followed by UVA. tab-densityDNA peak from each gradient were pooled, and 0.7 Cross-Link Formation. To verify that the 8-MOP was intro ml of 1 M K2HPO4,titrated to pH 12.5 with KOH, was added to ducing cross-links into the DNA of cells treated in vivo and to denature the DNA. CsCI and water were then added to make test for cross-link formation after angebicin treatment, we ex a solution of about 7 ml with final refractive index of 1.405. Centnifugation was carried out as before, and fractions were collected in plastic tubes containing 10 @sgcalfthymus DNA. DNA was precipitated with cold CI3CCOOH and collected on B@ z Millipore filters, which were dried and assayed for radioactivity. 2 80 Repair replication was calculated as the 3H:32Pratio in the parental DNA, or as 3H:,.tgDNA in cases in which the specific @ activity of DNA from untreated cultures was determined. 60 Purification of DNA. Cells were washed with PBS at 37Â°, 0 0 @ trypsinized, and collected by centrifugation. The pellet was 40 resuspended in 0.4 ml 10 mMTmis(pH 8):i m@EDTA, contain I ing proteinase K (125 @tg/mb).Afteraddition of 0.1 ml Sarkosyb C., LU 20 solution [5% (w/v) in 10 m@iTns:1 mi,i EDTA], the lysate was > incubated for 2 hr at 50Â°or overnight at 37Â°. DNA was I- 4 centrifuged through a CsCI step gradient consisting of a 3.5-mb layer of aqueous CsCl at a density of 1.4 g/mI in 10 m@Tnis: 1 mM EDTA on top of a CsCl solution at a density of 1.8 g/ml UVA, MINUTES in the same buffer. A 0.5-mb sample of lysate was layered on Chart 1. Inhibition of DNA synthesis in cells treated with 8-MOP or angelicin and Irradiated with UVA. vA-i 3 cells prelabeled with 32Pwere incubated with 1 top of this gradient, which was then centrifuged for 2 hr at (C) or 10 (0) itg 8-MOP per ml or 100 @gangelicin per ml (A) before irradiation 40,000 rpm in a SW5O.i rotor at 20Â°.Fractions were collected, for the time shown and then incubated for 2 hr in medium containing [3HIdThd and aliquots were assayed for radioactivity as above. (2 @Ci!mI).The3Hincorporated into acid-precipitable material was normalized to the 32Pactivity and is expressed as a percentage of the value for unirradiated Alkaline Sucrose Gradients. Samples were layered on top cells. 3Hincorporation in unirradiated cells was @25,000cpm,'point.

MARCH1980 697

amined the effects of these drugs on the sedimentation of DNA tion was observed in confluent human WI-38 fibrobbast cells from treated cells after direct bysis on alkaline sucrose gra treated with either 8-MOP onangelicin and irradiated with UVA. dients. Profiles from cells treated with UVA on 8-MOP alone No repair replication was detected in cells treated with satu were similar to those obtained from control cells and represent rated solutions of the drugs but not irradiated, or in cells number-average molecular weights in the range of 2 X 108. irradiated for up to 20 mm with UVA but not exposed to drug. However, DNAfrom cells treated with both a high concentration Equilibration between the drug in the buffer and that bound to of 8-MOP and moderate exposure to UVA sedimented much the DNA is achieved quickly. Cells incubated with 8-MOP (10 more rapidly (Chart 2a). When the centmifugation time was sg/ml) for 5, 15, on 30 mm prior to a 5-mm irradiation all shorten, the profile was a single sharp peak sedimenting much exhibited the same repair replication values. The rapid disso more rapidly than DNA from untreated cells. Although this ciation of drug from the DNA was shown by an experiment in profile undoubtedly represents anomalous sedimentation and which cells were incubated with 8-MOP (25 @sg/ml),andthen cannot be analyzed easily, the absence of DNA in that region 1 culturewas immediatelyirradiatedfor 2 mmwith UVAwhile of the gradient corresponding to about 2 x 108 dabtons mdi the drug-containing solution was removed from the other and cates that DNA molecules of that average size have been replaced with drug-free buffer. After 5 mm, this culture was altered in sedimentation properties by the 8-MOP treatment. In irradiated with UVA for 10 mm. Repair replication in the second contrast, treatment of cells with a saturated solution of angelicin culture was less than 20% of that in the first, even though the and an identical UVA exposure (Chart 2b) resulted in no change UVA dose was 5 times banger. in the sedimentation profile of DNA. These observations are The characteristics of repair replication after treatment with consistent with the known cross-linking ability of 8-MOP, which the 2 drugs were quite similar. At 25 @tg/ml,valuesfor repair effectively increases the molecular weight of the DNA under replication observed after UVA exposures up to 10 mm were denaturing conditions, while angebicin does not possess such comparable for the 2 drugs (Chart 3). (At this concentration, ability. the molamityof angebicin is 16% greater than that of 8-MOP.) Repair Replication in Normal Human Cells. Repair replica As expected, the repair replication was dependent upon UVA exposure. At a fixed UVA exposure of 5 mm, the drug dose response for the 2 drugs is also similar and nearly linear (Chart 4). Furthermore, the time course of repair replication in cells treated with 8-MOP or angelicin (25 @zg/ml)andexposed to UVA for 5 mm is similar for the 2 drugs (Chart 5). The absolute 2000 amount of nepaimreplication observed in 4 hr after treatment with drug (25 @sg/mI)and5 mm UVA exposure gives an mdi cation of the initial rate of repair synthesis (see Chart 5). The value obtained is equivalent to the repair replication observed 1000 in those cells following about 1.7 J/sq m of 254 nm light, and only about 20% of the value reached after saturating 254-nm doses. The possibility that unmeacted8-MOP might be inhibiting repair replication was eliminated by determining that repair I C,) replication following 254 nm UV was unaffected by the pres 0

0. 0

1000

500 z4 0 0

,.â€œ I C') 0 0 5 10 15 20 25 30 0. 0 FRACTION NUMBER Chart 2. Radioactivity profiles of alkaline sucrose gradients of DNA from VA 13 cells treated with 8-MOP or angelicin and irradiated with UVA.Cells prelabeled @ with [3HjdThd were incubated with 8-MOP (50 pgi'ml) (a) or angelicin (100 ml) (b), irradiated for 5 mm with UVA, washed with PBS, and incubated at 37Â°for 0 30 mm. After being harvested by trypsinization and counted, 2.5 x 1O@cells 0 5 10 were resuspended in 1 ml of 0. 1 N NaCI:10 [email protected],pH 7.5: 10 m@EDTA:bovine UVA, MINUTES serum albumin (1 mg!mI). A 40-id sample of this suspension was layered on a 250-gil lysis layer (1 M NaOH:0.1 N EDTA) which had been layered on top of a Chart 3. Repair replication in WI-38 cells treated with 8-MOP or angelicin as linear 5 to 20% alkaline sucrose gradient containing 0. 1 M NaOH:1 mM EDTA. a function of UVA irradiation. Cells were incubated with 8-MOP (25 @g!mI;116 The gradients were centrifuged at 20Â°in an SW5O.i rotor for 75 mm at 30,000 fLM)(â€¢)or angelicin (1 34 @tM)(A), irradiated for the times Indicated, and incubated rpm. Arrow, position of â€˜4C-T2phageDNA. also layered (as phage) in the same In the presence of [â€˜HJdThdfor4 hr. Repair replication was measured as gradients. described in â€˜â€˜MaterialsandMethods.â€•

698 CANCERRESEARCHVOL. 40

I I I I Irepair wasfrompatch (28). analysis @tg/mb)Aand Wl-38 cells treated with angelicin or 8-MOP (25 [3H]dThd0- 5 mm UVA, and incubated in medium containing 0.6 ,â€¢ - 100 the repair label. The small density shifts observed were 0. /. similar to those routinely observed in analyses of patch size C,) ,/. s, following exposure to 254 nm UV (27). Because of the slight selectivity of confluent WI-38 cells for dThd over BndUmd(29), 0. ./â€˜ ;; U we wished to repeat this experiment with [3H]BndUndas repair 0.4 - V I label. We chose another human cell line, T98G, which also C') / C,) / exhibits contact inhibition but which grows to a much higher 0. 50 U 7 - density. We exposed the cells to higher drug concentrations /A and increased the repair period to maximize the yield of 3H- I 0.2 -S/ >as labeled repair patches. The results (Chart 6) indicate that the / average size of the vast majority of repair patches initiated /A after either angelicin on 8-MOP treatment is similar to that / following 254 nm UV. The weight-average molecular weight of I the sonicated DNAwas roughly 200 nucleotides, as determined @@ 1@D 1@ 20 25.i@ by sedimentation in alkaline sucrose gradients. This leads to an estimate of about 25 nucleotides in the average patch. The DRUG CONCENTRATION (,@g/mt) evident skewing of the profiles to the dense side for the DNA Chart 4. Repair replication in WI-38 cells as a function of the concentration of 8-MOP or angelicin to which they were exposed prior to UVA irradiation. Cells were incubated with 8-MOP (â€¢)orangelicin (A) at the concentrations indicated and then irradiated with UVA for 5 mm. Repair incorporation was for 5 and 4 hr, respectively. The @PspecificactIvity for DNA was not available for the 8-MOP experiment, and repair is expressed per unit @P.

3 500

1' 2 I C')-@J

a. C') U I 500 C-) â€ã C-) I â€˜D â€Õ C,, â€Õ

0 10 15 500 TIME (HOURS) Chart 5. Time course of the repair replication in Wl-38 cells in response to the 2 drugs. Cells were incubated with 8-MOP (25 @g!ml)(â€¢)orangelicin (A) and UVA Irradiated for 5 mm. In the latter case, 32Pspecific activity for the DNA was not available. FRACTION NUMBER ence of 8-MOP (25 sg/mI) during the post-UV incubation Chart 6. Repair patch size distribution after exposure of cells to angelicin, 8- MOP, or 254 nm UV. ConfluentT986 cells prelabeledwith32Pwere exposedto period. saturated solutions of angelicin or 8-MOP for 5 mm, irradiated with UVA for 5 The average size of the DNA stretches synthesized by repair mm, and incubated in medium containing 10 [3HJBrdUrd(10 liCi!mI) for 20 hr. replication may be determined by reducing the molecular Parental density DNA from the cell lysates was Isolated from a neutral CsCI gradient and purified in a second gradient. It was then dialyzed, sonicated, and weight of the DNA to an average size of about 250 nucleotides centrifuged in 2.8-mi alkaline gradients as described 28). The gradients were by sonication prior to alkaline CsCI equilibrium sedimentation. fractionated directly into vials to which were added 1 ml H,O and 10 ml Omnifluor: Triton x-i 00 (2: 1) for radioactivity assay, the results of which are shown. The density of those DNA fragments carrying repair patches, Centrifugal field increases to the left. Bars, position of unsonicated hybrid-density labeled with 3H, and containing bromoumacibwill then be de DNA(recoveredfromtheneutralgradientof DNAfromcellsexposedto angelicin) tectably increased relative to that of the parental DNA, labeled included in a separate tube in the same run. Bottom, results of a similar experiment in which cells were exposed to UVA (10 J/sq m) and incubated for with 32P.The extent of the density shift and the independently 8 hr in mediumcontaining [3H]BrdUrd(5 @zCl/mI).Forthis alkaline gradient, larger determined size of the fragments allows calculation of the fractions were collected.

MARCH1980 699

Downloaded from cancerres.aacrjournals.org on October 2, 2021. © 1980 American Association for Cancer Research. J. Kaye et a!. from drug-treated cells amounts to only 10 to 12% of the Table 1 radioactivity. Whether this represents a very small number of SI nuclease resistance of linear SV4O DNA patches considerably banger than the avenage, or a small DNA samples were subjected to heat denaturation before incubation with or without 5, nuclease. amount of semiconsenvative synthesis, is not known. The pro file shown for cells exposed to 254 nm UV is from a separate (cpm)UVA 3H experiment, in which the molecular weight of the DNA was not nuclease8-MOPDrugTCA-precipitable + 5, nuclease â€”5, reduced as much, judging from the band width of the 32P- 7894 7124 labeled DNA component. 8-MOP â€” 116 7416 Repair Replication in XP Cells. Cells from patients with XP Angelicin + 1 58 5205 Angelicin+ â€” 5319Table 168 are defective to varying degrees in repair replication following UV irradiation or treatment with a variety of chemical camcino gens (12). Most strains in Complementation Group A, including 25, the strain used here (XP12BE) are almost completely defective DNA8-MOPTCA-precipitablenucleaseresistanceofWI-38cellular in this repair replication. Since it has been reported that XP (cpm)Heat 3H cells are able to remove cross-links produced by mitomycin C dena- + S nu- â€”S,nu (10), it appeared that XP cells might facilitate study of the clease+ UVA turation clease repair replication process, if any, associated with cross-link + 13.350 12,720 removal. However, these cells proved defective in repair mepli + + â€” 12,890 12,530 cation after treatment with either angelicin on psoraben at 25 + â€” + 180 10,840 + â€” â€” 11,410 11,130 zg/ml and a 5-mm UVA dose. The values obtained in a 4-hr â€” + + 270 14.260 posttreatment incubation period were less than 10% of the â€”+ + â€” 14,240 15,020 values obtained with Wl-38 cells treated similarly. Determination of lnterstrand Cross-Links in DNA. In order cross-linked; in theory, the continued reduction of molecular to estimate the frequency of cross-links produced in cells weight should beadto a smaller fraction of resistant molecules. treated with 8-MOP and UVA and to study cross-link removal, This was shown in a qualitative way by reducing the molecular we developed a method for cross-link estimation based upon weight of a cross-linked DNA sample through hydrodynamic both the rapid menatumabibityofcross-linked DNA and the spec shearing and varying degrees of sonication. Highly cross-linked ificity of Si nuclease for single-stranded DNA. DNA sheared through a needle was 90% resistant to S@diges Control experiments were conducted with radioactively Ia tion. Sonication of the same sample for 15, 20, and 30 sec at beled SV4O and bacteriophage A DNA. S@nuclease did not an arbitrary selling reduced the S@resistance to 67, 46, and degrade native DNA, and neither heat denaturation nor incu 20%, respectively. After 30 sec of sonication, 30% of the DNA bation in the absence of S1resulted in any bossof acid-precip became S@sensitive without heating, probably due to the itable DNA. However, after heat denatunation and incubation production of frayed ends in duplex molecules, or outright with S@,less than 5% of the DNA remained acid pmecipitable, denaturation. Similarly, in experiments with cellular DNA except in the case of closed circular SV4O DNA, which me sheared by passage through a 25-gauge needle, often a small mained 97% acid precipitable, indicating that little if any nicking percentage of the DNA was sensitive to S1 even when not of the closed circular strands occurred during the 30 sec at heated. DNA prepared from XP12BE cells which had been 100Â°.In this latter experiment, differentially labeled A DNA treated with 8-MOP (25 @g/ml)andexposed to 6 mm UVA was included in the sample (made by digesting form I DNA with sheared by 5 passages through a 25-gauge needle. This pro EcoRl nuclease), was 98% digested. For unit length linear cedumeyields double-strand moleculan weights of the order of SV4O DNA, a boiling time of 15 sec yielded the same extent of 10 x [email protected] incubations ranging from 20 to 60 mm, digestion as 30 sec or I mm, and 30 sec was then chosen as 40% of this DNA was rendered acid soluble, while 30% was the standard time. solubibized in a 10-mm digestion. (Our standard digestion in Unit length linear SV4O DNA was incuoated in saturated cubation was 30 mm.) solutions of 8-MOP or angelicin, exposed or not to 5 mm UVA, Cross-Unk Removal in Different Cell Types. In order to and examined for S@resistance by the above procedure. As evaluate the removal of 8-MOP cross-links in various human expected, DNA treated with angelicin and heat denatured was cell lines without having to make mathematical corrections for sensitive to S@whetherexposed to UVA or not, but DNA treated differences in the extent of shearing among different samples, with 8-MOP and exposed to UVA was completely resistant we used the following procedure. Cells were uniformly labeled (Table 1). Wl-38 cells were then treated with saturated sobu with either [3H]dThd or [â€˜4C]dThd,thengrown for at beast24 hn tions of 8-MOP, or exposed to 10 mm UVA, or both. Cells were in nonmadioactivemedium, and treated with 8-MOP and UVA bysed, and treated with proteinase K, and their DNA was as described above. Cells from 1 plate were then harvested isolated and purified on CsCI step gradients. Resistance to S@ and frozen at _200. An identically treated, differentially labeled after denatumation was observed only when the cells were plate was reincubated for the desired time; then, cells were subjected to both 8-MOP and UVA(Table 2). In this experiment, harvested, and the cell pellet was frozen. The 2 samples were the high-molecular-weight DNA recovered from the cells then thawed and combined, and the DNA was purified on CsCb treated with 8-MOP and exposed to UVA was completely step gradients, sheared, and analyzed for cross-links. resistant to Si digestion. Thus, every DNA fragment must have The above procedure was used to monitor changes in cross contained at least 1 cross-link. The determination of the fre linking in a number of cell strains (Chart 7). The amount of quency of cross-links requires a knowledge of the molecular rapidly renatumingDNA in Wl-38 and XP variant cells declined weight of DNA under conditions where not all molecules are with time. In separate experiments with Wl-38 and XP variant

700 CANCERRESEARCHVOL. 40

The photodependency of the damage provides a means of 30 limiting the time during which damage is produced and varying Lu the amount of damage produced at a constant drug concentma U) 4 tion. In addition, psoralen cross-links may offer an approach to Lu the study of damage recognition mechanisms and repair path U ways if, for example, mutant cells able to remove cross-links z but not monoadducts (on vice versa) could be found. It would U) be useful to understand the different biochemical effects of 0 20 cross-links and monoadducts in designing psorabentreatment of hyperproliferative skin diseases such as psoriasis. I- z We have demonstrated repair replication in Wl-38 human

U) cells in vitro following treatment with 8-MOP or angelicin and U) UVA. The process appears to be similar in several respects to Lu the repair replication response observed following exposure to z4 254 nm UV. It produces repair patches of the same size, is 0 10 absent in XP(A) but not in XP variant cells. The kinetics of U. repair is similar in that the rate of synthesis is initially high and 0 gradually declines. z We have no data on the number of monoadducts produced Lu U by angebicin for our drug and UVA doses on on the relative Lu number of monoadducts and cross-links produced by 8-MOP. 0. Previous studies (4, 8, 23, 24) have shown the relative photo 0 reactivity of 8-MOP to angelicin for both isolated native DNA and DNA in bacterial cells to be dependent on UVA dose and TIME (HOURS) to range from about 2 at low doses to 4 to 5 at higher doses. chart 7. Removalof cross-links by various cell strains as assayed by changes Estimates for the ratio of monoadducts to cross-links with in sensItIvItyto Si nuclease. Cells were exposed to 8-MOP (25 @ig/ml)and5 mm various psonalens range from 9:1 to 2.5:1 for 4,5',8-tnmethyl UVA.PreparationofDNAisdescribedinthetext. The acid-insolubleradioactivity remaining after treatment of the heat-denatured samples with 5, nuclease, psoralen (3, 5). With multiple reimmadiations,anupper limit for relative to the values obtained when the denaturation step was omitted, is shown. cross-links of 65% of bound 8-MOP has been reported (19). In Cell strains examined were normal diplold (WI-38, â€¢),xerodermacomplementa one experiment, we were abbeto triple the cross-linking fre tion Group A(XP1 2BE, A), xeroderma variant(XP4BE, Il), and Fanconi's anemia (CALl 196, 0). quency by reirradiating cells in the absence of 8-MOP following initialtreatment with 8-MOP at 25 @zg/mlanda 2-mm irradiation. cells, incubation for 24 hr did not markedly reduce the cross This suggests that in the other experiments reported in this link frequency below that at 12 hr. However, no reduction in paper the proportion of monoadducts to cross-links was meba cross-linking was observed in XPI 2BE cells, the Complemen tively high. tation Group A cells which also showed no repair replication The backof repair replication in XP(A) cells after treatment after treatment with 8-MOP. We were surprised to note consid with angebicin or 8-MOP and UVA suggests that the repair erable reduction of cross-linking in CALl 158, cells derived pathway for monoadducts produced by these agents shares at from a patient with Fanconi's anemia, since this genetic defect beastone enzymatic step with that for pyrimidine dimers and Is reported to involve a deficiency in cross-link repair (10, 25). other bulky adducts such as those produced by aflatoxin B1, In using this double-labeling procedure, we assumed the N-acetoxy-2-acetylammnofluonmne,and a number of other molecular weight distribution of DNA from cells that were agents. The absence of repair replication after 8-MOP treat allowed a repair period to be identical to that of DNA from cells ment also suggests that repair of psonabencross-links either before repair. However, reduction of the DNA molecular weight does not occur in these cells, is not associated with repair in the cells during repair might have occurred. We therefore replication as we define it, or is not detectable by our tech examined the molecular weight distribution of the DNA samples niques. Our inability to detect boss of cross-linking in these from Wl-38 and Fanconi's anemia cells by velocity sedimen cells supports the first hypothesis and suggests that XP(A) tation in neutral sucrose gradients. The profiles of the 2 iso cells backthe initial recognition step for both cross-links and topes In each sample were identical. Similar analysis of the monoadducts, a situation similar to that found in Escherichia resistant DNA remaining after denaturation-renaturation and S@ co!iuvrAmutantswhichdo not initiatecross-linkrepair(7). digestion showed that the larger DNA fragments in the original These results extend the finding of Baden et al. (2) that repair heterogeneous samples were more likely to contain cross-links of 8-MOP damage as determined by measurements of un than the smaller pieces, as expected if production of cross scheduled DNA synthesis is defective in XP cells. links and breakage of the DNA by shear were independent, A neport of Fujiwara et a!. (10) suggested that XP cells are random events. able to remove cross-links induced by treatment with mitomycin C but not similar monofunctionab adducts produced by decar DISCUSSION bamoybmitomycin C and that the reverse is true in Fanconi's anemia cells. A different XP(A) strain was used (XP205), and Psoralens represent a useful class of DNA-damaging agents a level of unscheduled synthesis was seen in these cells in research because they are highly specific for nucleic acids following mitomycin C treatment, which was equivalent to that and are relatively innocuous in the absence of photoactivation. observed in normal cells. The discrepency with our findings

MARCH1980 701

may be due to the ability of these cells to perform DNA repair benzofuran derivatives. VII. New synthesis of angelicin and 4-methyl ange or to the biochemical difference between adducts produced by 11cm.Bull.Chem. Soc. Jap., 35: 149â€”151,1962. 15. Matthews, M. M. Comparative study of lethal photosensitization of Sarcina mitomycin C (an alkylating agent) and psoralen photoadducts. luteaby 8-methoxypsoralenandby toluldineblue.J. Bacteriol.,85: 322â€” Our results in one Fanconi's anemia cell strain suggest that 328, 1963. 16. Melski, J. W., Tanenbaum, L., Parrish, J. A., Fitzpatrick, T. B., Bleich, H. L., these cells do not exhibit a generalized defect in cross-link and 28 participating investigators. Oral methoxsalen photochemotherapy removal. for the treatment of psorlasis: a cooperative clinical trial. J. Invest. Dermatol., 68: 328-335, 1977. 17. Musajo, L, Bordin, F., and Bevilacqua, R. Photoreactions at 3655 A linking ACKNOWLEDGMENTS the 3-4 double bond of furocoumarins with pyrimidmnebases. Photochem. Photoblol., 6: 927â€”931,1967. We appreciate the helpful discussions and critical reading of the manuscript 18. Musajo, L, Bordin, F., Caporale, C., Marciani, S., and Rigatti, G. Photo by Drs. Ann Ganesan and Vera Morhenn. We thank Rolando Arrabal for reliable reactions at 3655 A between pyrimidine bases and skin-photosensitizlng laboratory support and Maria Kent for manuscript preparation. furocoumanns. Photochem. Photoblol., 6: 71 1-71 9, 1967. 19. Ou, C., Tsai, C., Tapley, K., and Song, P. Photobindlng of 8-methoxypsoralen and 5,7-dlmethoxycoumarin to DNA and its effect on template activity. REFERENCES Biochemistry, 17: 1047-1053. 1978. 20. ParrIsh, J. A., Fitzpatrick, T. B., Tanenbaum, L, and Pathak, M. A. Photo 1. Ashwood-Smfth, M. J., and Grant, E. Conversion of psoralen DNA monoad chemotherapy of psoriasis with oral methoxsaien and Iongwave ultraviolet ducts in E. coli to interstrand DNA crosslinks by near UV light (320â€”360 light. N. EngI. J. Med., 291: 1207â€”1211, 1974. nm): inability of angelicin to form crosslinks in vivo. Experientia (Basel), 33: 21. Pathak, M. A., Kramer, D. M., and Fitzpatrick, T. B. Photoblologyand 384-386, 1977. photochemistry of furocoumarins (psoralens). In: T. B. Fitzpatrick, M. A. 2. Baden,H. P., Parrlngton,J. M., Delhanty,J. D. A., and Pathak, M. A. DNA Pathak, L C. Harber, M. Seiji, and A. Kukita. (eds.), Sunlight and Man, pp. synthesis in normal and xeroderma pigmentosum fibroblasts following treat 335-368. Tokyo: University of Tokyo Press, 1972 ment with 8-methoxypsoralen and long wave ultraviolet light. Blochim. 22. Rodighlero,G., and Dall'Acqua,F. Yearlyreview:biochemicaland medical Blophys. Acta, 262: 247â€”255,1972. aspects of psoralens. Photochem. Photobiol., 24: 647-653, 1976. 3. Ben-Hur,E., and Elkind,M. M. DNA cross-linkinginChinesehamstercells 23. Rodlghiero,G., Musajo,L, Dall'Acgua,F., Marciani,S., Caporale,G., and exposed to near-ultraviolet light in the presence of 4,5',8-trimethylpsoralen. Ciavatta, L Mechanism of skin photosensitization by furocoumarins. Pho Biochim. Biophys. Ada, 331: 181-193, 1973. toreactivity of various furocoumarins with native DNA and with ribosome 4. Bordin, F., Carlassare, F., Baccichetti, F., and Anselmo, L DNA repair and RNA. Biochim.Blophys.Acta, 2 17: 40â€”49,1970. recovery in Escherichia coil after psoralen and angelicin photosensitization. 24. Rodighlero,G., Musajo,L., Dall'Acqua,F., Marciani,S., Caporale,G., and Biochim. Biophys. Acta, 44 7: 249â€”259,1976. Clavatta, M. L. A comparison between the photoreactivity of some furocou 5. Cole, R. S. Psoralen monoadducts and lntersfrand cross-links in DNA. marins with native DNA and their skin-photosensitlzing activity. Experlentia, Blochim. Biophys. Acta, 254: 30â€”39,1971. 25:479-481, 1969. 6. Cole, R. S. Ught-induced cross-linking of DNA in the presence of a furocou 25. Sasaki, M. S., and Tonomura,A. A highsusceptibilityofFanconi'sanemia mann (psoralen). Studies with phage A, Escherichia coil, and mouse leuke to chromosome breakage by DNA cross-linking agents. Cancer Res., 33: mia cells. Biochim. Biophys. Acta, 2 17: 30â€”39,1970. 1829â€”1836,1973. 7. Cole, R. S., Levltan, D., and Sinden, R. P. Removal of psoralen lnterstrand 26. Scott,B. R., Pathak,M. A., and Mohn,G. R. Molecularandgeneticbasisof crosslinks from DNA of Escherichia co/i: mechanism and genetic control. J. furocoumarin reactions. Mutat. Res., 39: 29â€”74,1976. Molec. Biol., 103: 39-59, 1976. 27. Smith, C. A. Removal of T4 endonuclease â€˜/sensitivesites and repair 8. DaII'Acqua, F., Marciani, S., Ciavatta, L., and Rodighiero, G. Formation of replication in confluent human diploid fibroblasts. ln: P. C. Hanawalt, E. C. inter-strand cross-lmnkingsinthe photoreactions between furocoumarlns and Friedberg, and C. F. Fox (eds.), DNA Repair Mechanisms, pp. 31 1-314. DNA. Z. Naturforsch.,26b: 561-569, 1971. New York: Academic Press, Inc., 1978. 9. DaII'Acqua, F., Marciani, S., and Rodighiero, G. Interstrand cross-linkages 28. Smith, C. A., Cooper, P. K., and Hanawalt, P. C. Measurement of repair occurring In the photoreaction between psoralen and DNA. FEBS Left., 9: replication by equilibrium sedimentation. In: E. C. Frledberg and P. C. 121â€”123,1970. Hanawalt (eds.), Laboratory Procedures for DNA Repair Research. New 10. Fujiwara, Y., Tatsumi, M., and Sasaki, M. S. Crosslink repair in human cells York: Marcel Dekker, Inc., in press, 1960. and its possible defect in Fanconi's anemia cells. J. Molec. Biol., 113: 635â€” 29. Smith,C. A., and Hanawalt,P. C. Repair replicationin cufturednormaland 649, 1977. transformed human fibroblasts. Biochim. Biophys. Acta, 44 7: 121â€”132, 11. Griffin, A. C., Hakim, R. E., and Knox, J. The wave length effect upon 1976. erythemal and carcinogenic response in psoralen treated mice. J. Invest. 30. Stein, G. H. T98G: an anchorage-Independent human tumor cell line that Dermatol., 31: 289-295, 1958. exhibits stationary phase G1 arrest in vitro. J. Cell. Physiol., 99: 43-54, 12. Hanawalt, P. C., Cooper, P. K., Ganesan, A. K., and Smith, C. A. DNA repair 1979. in bacteria and mammaliancells. Annu. Rev. Biochem., 48:783â€”836,1979. 31. Stern, R. S., Thibodeau , L. A., Kleinerman, R. A., Parrish, J. A., Fitzpatrick, 13. Jois, H. S., Manjunath, B. L., and Venkatia Rao, S. Chemical examination of T. B., and 22 participatinginvestigators.Riskof cutaneouscarcinomaIn the seeds of Psoralea coryilfoila (Linn.). J. Indian Chem. Soc., 10: 41â€”46, patients treated with oral methoxsalen photochemotherapy for psorlasis. N. 1933. EngI. J. Med., 300: 809-81 3, 1979. 14. Kawase, Y., Nakayama, M., and Tamatsukuri, H. Synthetic studies on the

702 CANCERRESEARCHVOL. 40

James Kaye, Charles Allen Smith and Philip C. Hanawalt

Cancer Res 1980;40:696-702.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/40/3/696

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/40/3/696. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.