Journal of Photochemistry and Photobiology, B: Biology, 3 (1989) 33 - 52 33

UVA-INDUCED BINDING OF 8METHOXYPSORALEN TO CELLS OF SACCHAROMYCES CERE VISIAE: SEPARATION AND CHARACTERIZATION OF DNA PHOTOADDUCTS

MICHAEL BANKMANN and MARTIN BRENDEL+

Znstitut fiir Mikrobiologie der J. W. Goethe-Universitiit, Theodor-Stern-Kai 7, Haus 75, 6000 Frankfurt/Main (F.R.G.)

(Received April 8,1988;accepted July 14,1988)

Keywords. Psoralen, hydroxylapatite chromatography, gel chromatography, DNA adducts, molecular dosimetry, Saccharomyces cereuisiae.

Summary

We present methods for the determination of UVA-induced binding of 8methoxypsoralen (&MOP) to nucleic acids and protein and for a quantita- tive assay of radioactively labelled &MOP plus UVA induced DNA photo- products in the yeast Saccharomyces cereuisiae. For the dose range up to 60 kJ m-*, with a wild-type survival of 1% or higher, binding to DNA is loo-fold and to RNA lo- to 20-fold more efficient than that to protein. Between 20% and 65% of the 8-MOP binds to macromolecules that are neither nucleic acids nor protein. The number of DNA-bound 8-MOP mole- cules for the haploid genome rises from 14 (unirradiated control) to 349 at the highest UVA exposure dose (60 kJ m-*). Gel chromatography reveals three types of DNA thymine photoproduct, the pyrone-side monoadducts, the furan-side monoadducts and the diadducts. Among these, pyrone-side monoadducts always constitute the smallest fraction, regardless of whether the treatment is with in vitro or in uiuo 8-MOP plus UVA.

A bbreuiations

HA: hydroxylapatite; HMT: 4’-(hydroxymethyl)-4,5’,8-trimethylpsoralen; ICL: interstrand cross-link; 8-MOP: 8-methoxypsoralen; UVA: ultraviolet light 320 - 400 nm.

1. Introduction

8-Methoxypsoralen (8-MOP) belongs to a group of naturally occurring bifunctional furocoumarins whose antimitotic potential is used in the photo-

‘Author to whom correspondence should be addressed.

loll-1344/89/$3.50 0 Elsevier Sequoia/Printed in The Netherlands

34

therapy of certain skin diseases [ 1, 21. Its applicability as a therapeutic agent, however, is limited by its proven mutagenicity, recombinagenicity, carcinogenicity and teratogenicity [ 21. The photobiological activity of furocoumarins resides in their ability to bind covalently to DNA [3, 41, RNA [ 51 and protein [6] after activation by UVA irradiation. Reaction with DNA is thought to be the most important contribution to the biological effectiveness. Three steps take place in the interaction of &MOP with DNA: (a) formation of a DNA-ES-MOP complex in the dark by intercalation of the planar molecules between two base pairs of DNA, (b) UVA-induced cyclo- addition to the 5,6 double bond of thymine, either with the pyrone-side or the furan+ide of &MOP, to yield the pyrone-side or furan-side monoadduct respectively and (c) further photoactivation of furan-side monoadducts to form furan-side pyrone-side diadducts (interstrand cross-links; ICL). 8- MOP+UVA treatment of DNA, therefore, will produce three photoadducts, two monoadducts and one diadduct. Both monoadducts and diadducts in DNA may be genotoxic [7] but splitdose experiments in which mono- adducts were photoactivated to form diadducts (ICL) in the absence of further intercalated psoralen molecules clearly showed ICL to have a greater genotoxic potential [8,9]. These experiments, however, give no information on the number of remaining monoadducts since pyrone-side monoadducts and noncross-linkable furan-side monoadducts were not measured.

While characterization of photoactivated psoralen-thymine adducts [ 31 and isolation and characterization of photoadducts of in vitro psoralen-DNA interaction [lo, 111 as well as the chemical structure of photoadducts of other monofunctional and bifunctional furocoumarins have been described (for review see ref. 12), the quantitative assay of cellular DNA photopro- ducts, owing to their low number, has only been reported for mammalian cells [13]. Recently, a highly sensitive immunological assay (detection of photoproducts in femtomole concentrations) for in uiuo 8-MOP+UVA- induced DNA adducts has been introduced, which is based on monoclonal antibodies raised against 8-MOP-DNA photoproducts [ 141.

For yeast there are reports on the total number of psoralen-DNA photoadducts [15, 161; however, there are no data on the quantitative distribution of the different types of adduct after 8-MOP+UVA treatment in uiuo. In this communication we introduce a method that (a) yields com- parative data on the UVA-activated binding of 8-MOP to DNA, RNA and protein and (b) will allow a quantitative assay of all 8-MOP-DNA photo- adducts produced in uiuo for the yeast Saccharomyces cereuisiae.

2. Materials and methods

2.1. Strains The following haploid strains of Saccharomyces cereuisiae were used:

BM 7126-9A MATa RAD lysl-1 ade2-1 his438 leu2-3 ura rho* F51 MA Ta rude-1 6 tmpl - 1 Of8 rho* MB1114-5D/A MA Ta RAD metl-1 his5-2 1~~1-1 turn2 rho*

35

Strain F51 is an efficient utilizer of exogenous dTMP as described recently

1171.

2.2. Media and growth conditions Complete medium (YEPD) was prepared according to Brendel and

Haynes [18], medium N and 4N (supplemented with 2 X lo-’ mol 1-i dTMP) have been described previously [ 191.

YEPD medium was inoculated with (0.5 - 1) X 10’ cells ml-l from a pre-grown culture (18 h, 30 “C) and shaken vigorously for 15 - 20 h at 30 “C. Cells were harvested at (2 - 3) X lo7 cells ml-’ (65% - 85% budding cells) and washed three times with phosphate buffer (0.067 M, pH 7).

2.3. Mu tagens &MOP was obtained from Sigma. Stock solutions (330 c(g ml-‘) in

ethanol-water (3:7, v/v) were prepared and stored at room temperature, shielded from light. Under these conditions g-MOP remained stable for several weeks.

8-[Methoxyl-3H] methoxypsoralen in ethanolic solution (spec. act. 2.59 TBq mmol-‘, cont. 185 MBq ml-‘) was a special sample purchased from Amersham and was stored at -20 “C. The purity of the radiolabelled substance was checked by thin layer chromatography.

2.4. In vivo labelling of DNA, RNA and protein DNA specific labelling of strain F51 was done according to Fleer and

Brendel [20]. For in vivo labelling of cellular proteins [3H]dTMP was substituted by either 0.185 MBq ml-’ [ 14C] leucine, 0.185 MBq ml-’ [ 14C] - lysine, 0.185 MBq ml-’ [ 14C]histidine (all purchased from Amersham) or all three together. In contrast, specific RNA labelling was achieved by sup- plementing either 0.74 MBq ml-l [3H]uracil or 0.37 MBq ml-’ [ 14C]adenine (both from Amersham) to medium N (w/o uracil and adenine). Cells from strain MB1114-5D/A were inoculated (lo6 cells ml-‘) and grown in a gy- rator-y water bath shaker for 48 h at 30 “C. About 96% of incorporated radiolabelled bases were found in RNA as tested by CsCl density gradient centrifugation [ 201.

2.5. SMOP+UVA treatment of yeast cells Yeast cells harvested from log-phase were adjusted to a cell density of

10’ cells ml-’ and the suspension was cooled down to 4 “C. For each ex- perimental point, 13 ~1 8-MOP stock solution and 11.1 MBq [3H]8-MOP were added to samples of 1.3 ml (= 1.3 X log cells) to give a final 8-MOP concentration of 18.4 PM. The cell samples were kept in the dark for 20 min at 4 “C.

Since 8-MOP is rapidly metabolized at high cell densities, a process which can be influenced by both incubation time and final ethanol concen-

36

tration (4.9%) (from ref. 21 and own observations), both parameters were held constant with great accuracy. Irradiation of the cells was performed with an Osram high pressure mercury lamp (HBO 200 Ll) under constant agitation. The light was focussed on a quartz cuvette 10 mm thick, and the irradiated area was 1.3 cm2. To cut out UV light of wavelength below 360 nm a colourglass filter (WG 360, Schott) was placed in front of the quartz cuvette. The dose rate was 15 kJ mm2 min-’ as determined by actinometry

w31. Survival of treated and untreated cells was assayed as described pre-

viously [ 231.

2.6. 8-MOP+UVA treatment of calf thymus DNA To a solution of 1.44 mg calf thymus DNA (Sigma, type I) in 2 ml

buffer (10 mM Tris-HCl, 1 mM sodium EDTA, pH 8), 0.2 ml &MOP stock solution and 3.7 MBq [3H]8-MOP were added and kept in darkness for 20 min at 4 “C. Irradiation was performed without agitation of the DNA solution. Unreacted or photodegraded &MOP molecules were removed by chloroform extraction (3 X 5 ml).

2.7. Isolation of 8-MOP+UVA-modified DNA, RNA and protein The irradiated cells were washed three times with 5 ml ice-cold phos-

phate buffer (Heraeus Christ centrifuge), resuspended in 1.5 ml homogeniza- tion buffer (12 mM sodium phosphate, 1% w/v sodium lauryl sarcosin, 10 mM sodium EDTA, pH 6.8) filled in a pre-cooled (-80 “C!) Eaton press [24] and were homogenized after 1 - 2 h of freezing at -80 “C. The efficiency of cell disruption was always greater than 95%. The thawed homogenate (10 min, 60 “C) was washed twice (15 min, 13 000 rev min-‘, Eppendorf centrifuge), and the combined supernatants were ultrasonicated with a Branson Sonifier B-12 (10 s, 40 scale-parts) and desalted by using pre-packed PD-10 columns (Pharmacia). These extracts were subjected to hydroxylapatite (HA) chromatography.

Columns made from 5 ml plastic syringes with a polyethylene-sintered plug were filled with 500 mg HA gel (1:l w/w mixture of Bio-rad HTP grade and DNA grade) and mounted into an aluminium block heated to 60 “C. After suspending the gel in elution buffer (EB, 0.0125 M sodium phosphate) the column was loaded with the celI extract. The elution procedure was as follows: 4 X 5 ml 0.0125 M EB; 4 X 5 ml 0.1 M EB; 4 X 5 ml 0.2 M EB; 4 X 5 ml 0.33 M EB.

All fractions were collected, chilled in ice and assayed for acid-insoluble radioactivity by adding equal amounts of ice-cold 10% w/v trichloracetic acid. After 30 min at 0 “C the samples were collected on membrane filters (Millipore, pore size 0.45 pm), washed three times with 5 ml ice-cold water and 5 ml ethanol and dried, and the radioactivity assayed in a dioxan- methanol-based scintillator. The HA gel was dried (80 “C, 15 h) and radio- activity was assayed in a toluene-based scintillator. Dpm values were calcu-

37

lated by determination of the quench factor of each sample using an external standard (Beckman LS 330 scintillation counter). When separation of the DNA photoadducts was intended, the DNA fraction (0.33 M EB eluate) was dialysed against distilled water (2 X 30 min and then overnight, 4 “C) and thereafter concentrated by Aquacide III (Calbiochem). The final volume was approximately 1 ml.

2.8. Calculation of the isolation efficiency of DNA, RNA and protein Both HA elution conditions and determination of the DNA, RNA and

protein recovery were worked out by using in uiuo labelled DNA, RNA and protein standards which had been isolated from &MOP+UVA-treated yeast cells (15.3 PM, 30 kJ mv2; isolation procedure, see above). By HA chroma- tography, maximum separation of these biopolymers was achieved using sodium phosphate as elution buffer (EB) at 60 “C. RNA was eluted at an ionic strength of 0.2 M EB, whereas DNA was eluted with 0.33 M EB. Most of the protein remained on the HA gel or was eluted with 0.0125 M and 0.1 M EB respectively. The elution efficiency was calculated by summation of both the TCA-precipitable radioactivity of all fractions and the remaining radioactivity on the HA gel (=total radioactivity). The radioactivity of each fraction was expressed as a percentage of the total.

Taking into account the efficiency of cell disruption, losses after desalting procedure (approx. 10%) and the elution efficiency of HA chroma- tography, recovery of cellular DNA, RNA and protein was calculated as 71%, 73% and 80% respectively. Enzymatic digestion of protein and RNA was performed by adding 50 pg proteinase K to the crude cell extract and 50 E.cg pancreatic ribonuclease (pre-treated, 10 min 100 “C) to the desalted cell extract of 1.3 X lo9 cells respectively and incubation at 60 “C for 1 h. After this treatment more than 97% of RNA and protein became TCA soluble.

2.9. Separation of DNA photoadducts [ 3H]8-MOP-modified DNA was hydrolysed by adjusting the DNA

solution to 0.4 N HCl and heating to 75 “C for 4 h. After this treatment less than 2% of the DNA remained TCA-precipitable. After cooling down to room temperature the hydrolysis mixture was adjusted to pH 2.2 and lay- ered on a Sephadex LH-20 column (Pharmacia; 1.5 X 90 cm). Elution was performed with a buffer (pH 2.2; 10 mM KH2P04)/methanol gradient at a flow rate of 10 ml h-‘. 2 ml fractions were collected and the radioactivity determined in a dioxan-based scintillator.

2.10. Isolation of the DNA photoadducts To isolate the DNA photoproducts, an additional purification step was

introduced prior to gel chromatography. The DNA hydrolysate was applied to a preconditioned CrsSep Pak cartridge (Waters), and was first eluted with distilled water (10 ml) and then with ethanol (5 ml). The ethanolic fraction

38

was evaporated to dryness, redissolved in buffer (pH 2.2, 10 mM KH,P04) and subjected to gel chromatography. After this procedure, peak I was quantitatively removed whereas peaks II - IV seemed to be unaffected. The combined fractions of the various peaks were freeze dried, redissolved in buffer (pH 2.2,lO mM KH,PO,) and stored at -20 “C.

Thin layer chromatography of the photoadducts was performed on PEI sheets (Macherey and Nagel CEL 300).

2.11. Photoreversal of the photoproducts Small aliquots of the isolated photoproducts (25 - 50 ~1) were spotted

onto plastic petri dishes and UV-irradiated by an Osram HNSl2 source (intensity maximum at 254 nm) with an output of 1.5 J m-* s-l.

All manipulations with &MOP+UVA-treated material were carried out under yellow light.

3. Results

3.1. Separation of DNA, RNA and protein via HA chromatography When using radioactively labelled 8-MOP to study UVA-induced

binding of the chemical to cells of the yeast Saccharomyces cerevisiae, the cellular macromolecules must be separated extremely carefully, since after photoactivation the drug is known to react with and bind to nucleic acids and protein [ 2, 5, 61. Yeast cells were radioactively labelled in their DNA, RNA and protein respectively, and after 8-MOP+UVA treatment (15.3 PM; 30 kJ m-*) these macromolecules were liberated from the cells and their separation determined via HA chromatography. Optimal separation of DNA, RNA and protein was achieved by increasing stepwise the molarity of the elution buffer from 0.1 M to 0.2 M and finally to 0.33 M (Table 1). Over 86% of extracted RNA and 81% of extracted DNA eluted with 0.2 M and 0.33 M phosphate buffer respectively, whereas binding of protein to HA was less specific in that a major fraction of cellular protein was not eluted (depending on the quality of [‘4C]amino acid labelling, between 46% and 82%); another protein fraction eluted at very low molarity (Table 1, columns 1, 3, 5 and 6). With this elution protocol we found the following cross- contamination amongst the macromolecules: 5% RNA and 2% protein in DNA; 5% DNA and 4% protein in RNA; 13% DNA and 9% RNA in protein.

Digestion by RNAse and proteinase K led to the total elimination of the respective macromolecules; however, application of RNAse, although pre- conditioned at 100 “C for 10 min, also -decreased the fraction of DNA molecules (Table 1, columns 2,4 and 7).

HA chromatography of a yeast cell homogenate treated by r3H]8- MOP+UVA showed that more than half of the cell-bound radioactivity remained on the column whereas the fractions of labelled DNA and RNA were rather small (Table 1, column 8). The data in Table 1 are representa- tive for a batch of experiments all of which revealed principally the same distribution of macromolecule-bound radioactivity (RNA fraction, 5% - 19%;

39

TABLE 1

Fractions of DNA, RNA and protein separated by HA chromatography (a)

Sodium phosphate DNA (mol 1-l)

(b) (c)

RNA Protein [ 3H]8-MOP (e)

(b) (c) (b)* (b)** (d) (b) (c) (d)

0.0125 0.1

1.1 1.8 4.1 0.2 12.2 48.0 0.1 14.8 16.0 8.2

0.2 5.5 4.2 86.5 1.8 3.9 4.6 0.2 8.2 2.2 8.0 0.33 81.0 68.3 4.8 0.3 2.0 1.1 0.4 15.7 10.1 13.1 HA 12.4 11.0 4.6 0.4 81.9 46.2 1.5 61.3 57.0 55.0

Ix 100.0 85.3 100.0 2.7 100.0 100.0 2.2 100.0 85.3 84.3

(a) Percentage of high-molecular bound radioactivity layered on HA column. (b) Untreated homogenate. (c) Homogenate treated with pancreatic ribonuclease. (d) Homogenate treated with proteinase K. (e) Homogenate from a rho+ strain (MB1114-5D/A); 3.7 MBq [3H]8-MOP, 15.3 m 8-MOP, 30 kJ rnez. *Protein labelled with [ r4C]lysine. **Protein labelled with [ 14C]lysine, [ 14C]leucine and [ 14C]histidine

DNA fraction, 5% - 16%; protein fraction, 65% - 90%). These variations were also found when using other rho+ yeast strains and may reveal variations in protein and RNA contents of the log cells.

Addition of RNAse to a homogenate of [3H]8-MOP+UVA-treated yeast cells led to the complete digestion of RNA, as proven by isopycnic centri- fugation (data not shown); radioactivity found in the RNA fraction after HA chromatography of the same homogenate is most probably due to protein contamination (Table 1, column 9). Again, RNAse digestion leads to partial loss of radioactive molecules in the DNA fraction.

After digestion of the homogenate with proteinase K most of the bound [ 3H]8-MOP remained on the HA column (Table 1, column 10); the efficiency of protein digestion, however, differed between experiments (10% - 40%). The proteinase-K-resistant “protein fraction” therefore varied between 0.4 and 0.65. Since proteins bind a large amount of the [3H]8-MOP, the fractions of RNA and of DNA are overestimated by 4% and 2% protein contamination respectively. Therefore when determining the UVA-induced binding of &MOP to RNA and DNA we must correct by this protein frac- tion. Cross-contamination of DNA with RNA and vice versa is negligible owing to the small fraction of bound radioactivity; for the same reason we do not have to correct the protein fractions (0.0125 and 0.1 M elution, and the fraction remaining on the HA gel) for RNA and DNA contamina- tion.

3.2. UVA-induced 8-MOP binding to cellular components Results of separation of DNA from RNA and protein after extraction

from yeast cells that had been treated with different doses of UVA (surviving

40

fraction from 100% to 1%) are shown in Table 2. There was substantial binding of &MOP (last column) without UVA irradiation; after an exposure dose of 60 kJ mm2 total binding to cellular macromolecules increased lo- fold. Adduct formation with DNA and RNA rose about 20-fold whereas protein and protein-associated macromolecules (proteinase K resistant), while constituting the largest &MOP binding fraction, increased their share by approximately a factor of 10. Proteinase-K-resistant protein constituted about 30% and was lower than in the experiments reported above. DNA-

TABLE 2

UVA-induced [3H]8-MOP binding to cellular components (a) of yeast strain BM 7126-9A rho0

UVA Survival DNA (kJ rne2) (%)

RNA Profein

(b) (c)

2 (dpm X 103)

0 100 0.9 (1.9) 0.5 (1.1) 30.9 (66.0) 14.5 (31.0) 46.8 7.5 44.8 4.6 (3.5) 2.1 (1.6) 98.2 (74.9) 26.2 (20.0) 131.1

15 18.2 6.2 (3.3) 4.7 (2.5) 122.6 (65.1) 54.8 (29.1) 188.3 30 3.1 18.0 (5.2) 5.5 (1.6) 218.2 (62.9) 105.2 (30.3) 346.9 60 1.4 22.1 (4.6) 10.4 (2.2) 319.8 (66.4) 128.8 (26.8) 481.1

Numbers in parentheses refer to the percentage of total radioactivity. The RNA and DNA fractions were corrected for protein contamination: 4% or 2% of the radioactivity of the protein fraction (column b) was subtracted from the RNA or the DNA fraction respec- tively. (a) Measured as dpm (X 1 03) bound. (b) (dpm HA gel + dpm of 0.0125 M and 0.1 M wash fractions) -(c). (c) dpm HA gel + dpm of 0.0125 M and 0.1 M wash fractions after treatment of homo- genate with proteinase K.

TABLE 3

Number of 8-MOP molecules bound to DNA, RNA and protein

UVA Photoadducts bound 8-MOP-molecules bounda

(kJ m-*) Cella Genomeb RNA ProteinC c(g DNA pg RNA mprotein (X 10-s)

0 658 14 9 433 2.9 0.04 0.33 7.5 1848 72 33 1376 14.5 0.13 1.06

15 2659 98 74 1719 19.6 0.30 1.32 30 4902 284 87 3058 57.0 0.35 2.35 60 6675 349 164 4468 69.8 0.66 3.44

Total recovery of cellular DNA, RNA and protein, see Section 2. ‘Calculated on the basis of radioactivity bound to macromolecules. bCalculated for 2.0 X lo9 treated genomes per experimental point. ‘Referred to column b in Table 2. dDNA, RNA and protein content of a single haploid cell was calculated as 5 X lo-l4 g [25], 2.5 X lo-l2 g [25] and 1.3 X lo-” g [26] respectively.

41

bound &MOP with the highest specificity, about a loo-fold over RNA and about 10 - 20-fold over protein (Table 3).

3.3. Quantitative measurement of DNA photoproducts Figure 1 compares the elution profiles of photoadducts after acid

hydrolysis of in vitro %MOP+UVA-treated calf thymus and yeast DNA. We did not apply the hitherto favoured enzymatic hydrolysis [lo, 11, 13, 271 since the mild alkaline conditions intrinsic to this method might have led to decomposition of the pyrone-side thymine adducts [28, 291. We obtained five peaks of radioactively labelled material in the elution profile after photoactivated binding of radiolabelled &MOP to calf thymus DNA (Fig. l(a)), whereas only three peaks appeared in the corresponding experiment in which yeast DNA specifically radiolabelled in its thymine residues reacted with non-labelled 8-MOP+UVA (Fig. l(b)). In both experiments peaks II, III and IV appeared in the same fractions. Peak V in Fig. l(a) showed that

2LO

5 - x 180

E ,"

120

60

a

6

elut~on volume I ml I

Fig. 1. Gel chromatography (Sephadex LH-20) of acid-hydrolysed DNA treated with 8-MOP in vitro. Mobile phase: buffer (10 mM KH2P04, pH 2.2)/methanol gradient. -1, void volume; ......., methanol content (%); Peaks I - V, see text. (a) Calf thymus DNA treated with [3H]8-MOP+UVA; l +, 270 kJ rnA2; O--O, 810 kJ m-‘. (b) [3H]dTMP- labelled yeast DNA (from 5 X 10’ cells) treated with &MOP (200 pg): Pre-chromato- graphed via a C1s SepPak cartridge; a--0, without irradiation; o-0, 450 kJ mp2.

42

unbound &MOP (identical fraction of eluate as free 8-MOP; also appearing in non-UVA-treated controls) is still present after threefold extraction with chloroform; no such peak was found in the experiment where unlabelled &MOP was applied. Peak I appeared in unirradiated as well as in UVA- irradiated S-MOP-treated DNA and also in &MOP+UVA-treated RNA and protein (data not shown), which makes it an unlikely candidate for an UVA-induced 8-MOP-thymine photoproduct. This left peaks II, III and IV as putative photoproducts of the 8-MOP-DNA reaction. Since peaks II - IV were obtained after UVA photoactivation using either radioactive 8-MOP or radioactively (thymine) labelled DNA, all three photoproducts must con- stitute adducts formed by 8-MOP and thymine.

3.4. Identification of DNA photoproducts Lacking direct evidence for the structure of molecules eluting in peaks

II - IV we devised experiments to elucidate indirectly the nature of these substances. Three major photoproducts of 8-MOP+UVA treatment of DNA are known, two of them monoadducts and one a diadduct [2]. When UVA irradiating an initially 8-MOP+UVA-treated DNA from which all unbound 8-MOP has been removed one can induce diadducts from furan-side mono- adducts [ 11,271. If two of the peaks II, III and IV shown in Fig. 1 represent the monoadduct and diadduct respectively, we should obtain an increase in one with a concomitant decrease in another of the peaks in such a “UVA chase” experiment. Table 4 summarizes the results of UVA-dose-dependent production of 8-MOP-thymine photoproducts and the outcome of a UVA-chase of 8-MOP-treated DNA. While the UVAdosedependent relative decrease of photoproduct in peak III already indicated that this is the furan- side monoadduct [27], additional support came from the UVA chase (last two lines of Table 4). Irradiation in the absence of free 8-MOP significantly decreased material in peak III whereas peaks II and IV showed an increase. Upon comparing the absolute counts (of radioactivity) we find that most of the radioactivity has shifted from peak III to peak II, with a small increase in peak IV. While this shift could be seen in both of the additional UVA chase experiments, the distribution of radioactivity in the void volume and in fraction I did not follow a consistent pattern (small increase, small de- crease, and unchanged). We may thus state that in our gel filtration analysis, peak II contains the diadduct, peak III the furan-side monoadduct and peak IV the pyrone-side monoadduct of the UVA-induced 8-MOP reaction with the thymine residues of DNA.

Additional proof for the nature of the 8-MOP+UVA-induced DNA photoproducts was derived from their photochemical reaction after irradia- tion with UVZ54nm. For psoralens, opening of the cyclobutane ring was generally expected so that from the diadduct both types of monoadduct as well as free 8-MOP and thymine should be generated [ll, 301 whereas the monoadducts should split in 8-MOP and thymidine (thymine) [28,31 - 331. This is exactly what we found after UV irradiation of material in peaks II and III (Fig. 2): peak II yielded UV-cleavage products eluting like those in

43

TABLE 4

Quantitative determination of &MOP-DNA photoadducts induced after in vitro [3H]S- MOP+UVA treatment of calf thymus DNA

UVA (kJ rn-?)

Vo Peak Iz

Z zz zzz IV (dpm x 103)

0

30

90

270

270’

810

810b

30+ oc

30 + 450c

%

%

%

%

%

%

1 117 0.8 99.2

51 159 18.3 56.9

145 254 19.6 34.4

279 393 296 248 61 1277 21.9 30.8 23.2(48.9) 19.4(41.0) 4.8(10.1)

17.3 82.1 0.1 0.4 0.1

453 820 804 445 148 2670 17.0 30.7 30.1(57.5) 16.7(31.9) 5.5(10.6)

222 1 13.0 0.05

31 96 12.8 39.9

861 471 146 1701 50.6(58.2) 27.7(31.9) 8.6(9.9)

33 13.6(28.7) ;:.3(65.5)

51 19 21.2(62.1) 8.3(23.5)

7 243 2.9(5.5)

41 107 17.8 46.5

12 230 5.2(14.4)

0 0 0 118

18 51 1 280 6.3(25.3) 18.2(72.3) 0.3(1.3)

155 175 9 738 21.0(45.6) 23.8(51.7) 2.5(2.7)

Numbers in parentheses refer to the percentage of combined radioactivity of peaks II - IV. ‘Without acid hydrolysis. bPrechromatographed via a Sep-Pak Cis cartridge. ‘Pulse-chase experiment. Vo, radioactivity eluted with the void volume.

peaks III and IV (the presumed monoadducts) and peak V (8-MOP), while UV irradiation of peak III yielded only peak V material, i.e. radioactive &MOP; being unlabelled, the small amount of thymine that must also appear by cleavage of the furan-side monoadduct could not be detected by photometric means. The photoproduct in peak IV (presumed to be pyrone- side monoadduct) when in aqueous solution apparently consisted of two fractions with different stabilities. After UV irradiation we detected not only &MOP but several other cleavage products that could not be found after UV irradiation of S-MOP alone (Fig. 2).

3.5. Purity of 8-MOP+UVA-induced DNA photoproducts To check the purity of material eluting in peaks II - IV, we isolated the

material and further analysed it by thin layer chromatography (Figs. 3(a) -

44

75

60

peak IV

10

0 0 40 60 120 160 200 240 260

elution volume (ml) Fig. 2. Gel chromatography (Sephadex LH-20) of isolated DNA photoproducts eluted in peak II, peak III and peak IV : O*, without irradiation; n, 2.7 kJ rn-’ UVz% W

60

Fig. 3. Thin layer chromatography (PEI sheets) of the isolated DNA photoproducts. Mobile phase: 80:20 (v/v) methanol-water. (a) Peak II, (b) peak III, (c) peak IV, (d) peak II, in buffer (KH~POQ, pH 2.2); adjusted to 1 M NaCl; (e) m, adjusted to I N NaOH (1 h at room temperature; l +, adjusted to 1 N HCl(1 h at 80 “C). Before chroma- tography, the solutions were adjusted to pH 2.2 (with either concentrated HCl or IO N NaOH respectively).

3(c)). Furan-side monoadduct (material in peak III) migrated in a single condensed band (Fig. 3(b)) while the photoproducts from peaks II and IV yielded two separate bands each (Figs. 3(a) and 3(c)). Pyrone-side mono- adducts and diadducts are known to open the pyrone ring to yield the respective carbonic acid in aqueous solution [ll, 28 - 321. This process is reversible, and while the open pyrone ring predominates in alkaline solution the closed la&one ring is reproduced on heating in acid solution [28, 291.

III

60 60 120 140 160 180 200 220 elutiCm VOlWllC [IId]

45

Alkaline treatment of the photoproduct of peak II (1 N NaOH, 1 h at room temperature) eliminated the fast migrating band while acid treatment (1 N HCl, 1 h at 80 “C) eliminated the slowly migrating band in thin layer chroma- tography (Figs. 3(d) and 3(e)). Essentially the same changes were found for photoproduct from peak IV (putative pyrone-side monoadduct; data not shown). The photoproduct from peak III did not show any changes in migration when subjected to thin layer chromatography after alkali or acid treatments as would be expected from a furan-side monoadduct.

3.6. Applicability of developed methods in vivo The same types of photoproduct were found in the DNA of yeast cells

after treatment with radioactively labelled SMOP+UVA (Fig. 4). The relative peak height of photoadducts eluting in peaks II, III and IV again depended on the exposure dose of UVA, higher doses leading to a larger fraction of material in peak II (diadduct) at the expense of peak III material (furan-side monoaddud). The pyrone-side monoadduct clearly was a minor- ity photoproduct in uiuo as even an exposure dose of SMOP+UVA leading to only 1% survival in haploid wild-type strain BM7126-9A still only yielded a relative fraction of approximately 1%.

Fig. 4. Gel chromatography (Sephadex LH-20) of acid-hydrolysed yeast DNA isolated from [3H]8-MOP+UVA-treated yeast cells (BM7126-9ArhoO). 4, void volume; for calcu- lation of % dpm, fractions of the void volume were not taken into account. m, 7.5 kJ m-*; W, 60 kJ m-*.

The fraction of radioactivity of the void volume (maximum 50%; not shown in Fig. 4) was rather large when compared with the same fraction in the gel elution of in vitro [3H]8-MOP+UVA-treated calf thymus DNA; this reveals protein contamination of the isolated yeast DNA.

Discussion

We have shown (I) that it is possible to separate quantitatively DNA, RNA and protein via HA chromatography, (II) that 8-MOP binds to all three above-mentioned macromolecules after UVA activation and (III) that

46

&MOP-DNA photoadducts produced in vivo or in vitro may be separated via gel chromatography. The separation of double-stranded DNA and single- stranded RNA is due to their different adsorption to HA [34] which is a temperature-dependent process [35]. Thus, our DNA and RNA elution at 60 “C proceeds faster than that described at 4 “C [36,37].

Proteins adhere tightly to HA, and native proteins bind stronger than denatured ones [38]. Labelling of proteins with radioactive amino acids reveals a great inhomogeneity among protein molecules as judged by the different elution profiles of protein labelled by one or a mixture of three amino acids (Table 1). Nevertheless, we find about 94% of all proteins either at low molarity of elution buffer (0.0125 M and 0.1 M) or remaining on the HA column. In contrast to our method, the elution of highly purified DNA (90%yield; 0.1% RNA and protein) on HA normally requires urea denaturation of RNA and protein [39,40]. For yeast, however, this method led to a lower DNA yield and a higher RNA contamination [41]. Also, there is no separation of RNA and protein with this method. Therefore our method has the advantage of separating the three macromolecules, hence allowing simultaneous measurement of &MOP binding to all three species.

Binding of [3H]8-MOP occurs in vivo to all three macromolecules, albeit with different efficiencies. Without UVA activation there is very little binding to the nucleic acids but significant reaction with protein. Concurrent with UVA irradiation, binding to RNA and DNA increases by a factor of 20 - 25, whereas binding to protein is only elevated by a factor of 10 (Table 3). Binding of 8MOP to protein in the dark may not be via covalent linkage, as shown for the reaction of furocoumarins with serum albumin [42, 431. After UVA activation two modes of binding to proteins have been described for 8-MOP: (a) indirectly, via photo-oxidation steps [6, 441 or (b) directly, via opening of either 3,4 or 4’,5’ double bonds [45]. At present the nature of the molecules that bind 8-MOP and remain on HA after di- gestion with proteinase K is unclear. Since in the controls neither [3H]8- MOP nor proteinase-Kdigested protein (labelled by three radioactive amino acids) remained on the HA gel, we offer four other possible explanations: (I) There might be a preference for 8-MOP to proteins not digested by proteinase K; this is unlikely, as even combined digestion with pronase and proteinase K did not substantially change the fraction of non-eluting 8-MOP. (II) Alternatively there might be normal digestion of protein marked with 8-MOP; the resulting polypeptides, however, stay on the HA gel because they are linked [46, 471. This possibility was excluded by an experiment where [3H]8-MOP was reacted with [ 14C]protein (labelled by three amino acids) and then digested by proteinase K: regardless of the exposure dose to UVA, and hence to the severity of 8-MOP-protein interaction, the [14C]radio- activity (showing residual protein or polypeptide binding) remaining on HA was always around 3% (data not shown), making strong polypeptide binding unlikely. (III) Free 8-MOP might adhere to the HA gel in the presence of cellular material. This is not realistic since neither a larger elution volume nor a higher molarity of the buffer (0.5 M) nor additional washing with 10 ml

ethanol led to any reduction of bound &MOP. (IV) Finally one can imagine &MOP bound to different sorts of molecule e.g. to lipids. This lipid-&MOP interaction has been described by several authors [ 48 - 501 and could explain why proteinase digestion had no effect. Some support for lipids (or mem- branes) as binding molecules for 8-MOP comes from our data, where elution of 8-MOP+UVA-reacted macromolecules stemming from a rho+ strain is compared with those extracted from a rho0 strain (Tables 1 and 2): clearly, homogenates from a rho+ strain (large share of mitochondrial membranes) have a larger protein fraction resistant to proteinase K (40% - 65%; Table 1; unpublished data) than those of rho0 strains (20% - 30%; Table 2). However, to what extent mitochondrial membranes and their proteins contribute to the proteinase-K-resistant fraction of yeast protein is not yet known, and this problem needs further investigation. Nevertheless, binding studies in mammalian cell lines also revealed that 8-MOP adheres, apart from the nucleus, to cytoplasmic structures [ 51, 521, a process interpreted as binding to proteins or membranes, or both [ 511.

The number of DNA-bound 8-MOP molecules for a haploid yeast genome rose linearly with the UVA dose from 14 (u&radiated control) to 349 (60 kJ mp2; 1.4% survival). 2 X lo9 genomes were treated per aliquot, and the genome number corrected for budding cells. Our 8-MOP binding data are considerably lower than the values reported by Averbeck [15]. Starting from 186 DNA-bound 8-MOP molecules (control) he found 5633 DNA-bound 8-MOP molecules per diploid genome (survival 1% at 25.2 kJ mv2). While the rate of increase is of similar magnitude (25-fold us. 30-fold), the absolute number of DNA photoadducts does reflect the different me- thods employed for their assay. Averbeck TCA-precipitated whole cells on glass fibre filters, followed by alkaline hydrolysis of RNA and enzymatic digestion of protein. The remaining filter-bound radioactivity (of 8-MOP) was attributed to photoadducts of DNA. Own experiments yielded much higher bound 8-MOP radioactivity when applying this method (data not shown), and one explanation for this might be non-specific 8-MOP binding to different types of macromolecule. If we add to our data those unspe- cifically bound 8-MOP molecules (column c in Table 2) we obtain 217 “adducts” (control) us. 2043 for the highest UVA exposure dose (haploid genomes), values comparable with those reported by Averbeck [15]. Also, use of rho+ strains can lead to an overestimation of photoadducts in nuclear DNA since it is known that mtDNA, although only constituting 15% - 25% of total DNA [53], has higher binding of 8-MOP by up to a factor of 10 [ 541. Most of the above arguments and the fact that we can verify the same low values for one type of 8-MOP-DNA adduct, the interstrand cross-link, by another method [55], lends credibility to our claim that UVA-induced 8-MOP adduct formation in yeast DNA is about one order of magnitude lower than estimated thus far. Our data show that 8-MOP binding to DNA is about loo-fold that of binding to RNA, whereas published data indicate a 6 - 8-fold higher binding to DNA [16]. Again, this might be due to the analytical standard employed.

48

Acid hydrolysis of [ 3H]8-MOP+UVA-treated calf thymus DNA yielded five radioactive reaction products after gel chromatography (Fig. l(a)). Peak V contained unbound 8-MOP and peaks II, III and IV thymine photoproducts as verified by the photoactivated reaction of g-MOP with [3H]dTMP-labelled yeast DNA (Fig. l(b)). Material in peak I has not been identified but it does not constitute an &MOP-thymine photoproduct since this peak also appears in experiments in which there was no UVA irradiation, where DNA was not hydrolysed and where RNA or protein was treated with [3H]8-MOP+UVA in vitro. Most likely, it is some sort of chemically altered form of &MOP which loosely adheres to macromolecules and dissociates rapidly during gel chromatography. A strong indication for this comes from experiments in which [3H]8-MOP was subjected to acid hydrolysis with or without prior UVA-irradiation; both procedures led to the formation of a “minority product” whose elution coincided with that of the peak I material (data not shown).

The material of peaks II and III was identified as diadduct and furan- side monoadduct respectively, by the UVA “pulse chase experiment” (Table 4). Only furan-side monoadducts absorb UVA320nm and may be activated to form diadducts [32, 561. Pulse chase experiments by Tessman et al. [27] showed that apart from the formation of diadducts from furan-side mono- adducts there was also an increase in pyrone-side monoadducts, apparently by direct photoisomerization; this is in accordance with our data in Table 4. Further analysis by thin layer chromatography revealed that material from peaks II and IV consisted of two components each (Figs. 3(a) and 3(c)). Their interchange, depending on alkaline or acidic treatments, makes them good candidates for the open and closed ring structure described for the pyrone-side monoadducts and diadducts [ 11, 28, 29, 31, 321. Broadening of peak IV (Figs. 1 and 2) may be attributed to the simultaneous presence of molecules with either ring structure. Irradiation by UV2sGnm of the pyrone-side monoadducts which already are unstable in aqueous solution [ 28,311 leads to further degradation products (Fig. 2, bottom).

HMT-DNA photoproducts (monoadducts and diadducts) which have an open pyrone ring at the pyrone-side are unable to undergo photoreversion at the pyrone side [30, 321. Diadducts with an open pyrone ring can there- fore only yield monoadducts with an open pyrone ring, whereas open pyrone ring monoadducts will not undergo photoreversion. UV254nm irradia- tion of material from peak IV, however, led to decay of both components (Fig. 2). This might be explained as follows: both ring-type moieties form an equilibrium, which at a pH of 2.2 favours the closed lactone-type of pyrone- side monoadduct (Fig. 2). After UV-induced splitting of the closed-ring monoadducts (into &MOP and thymine and other products) they are re- moved from the equilibrium, resulting in the conversion to the closed-ring- type of open-ring-type monoadducts.

We do not expect any pyrone-ring cleavage under pH 2.2 elution condi- tions [28,31] ; UV irradiation of the diadduct, however, yields both types of pyrone-side monoadduct (Fig. 2). We can thus speculate that cleavage of the

49

pyrone ring, either from the diadduct or the pyrone-side monoadduct, may already occur during UVA irradiation and thus also during in uiuo treatment. Since the open ring form might lead to a stronger distortion of the DNA double helix than the other photoproducts, such a putative “minority product” of S-MOP+UVA might well be of biological significance [ 311. Our results clearly show that the furan-side monoadduct and the diadduct are the major DNA lesions after S-MOP+UVA; pyrone-side monoadducts constitute a minority photoproduct. This would be expected from the results of other in vitro experiments [ 10, 11, 271. A direct comparison of the relative amounts of photoproducts, however, would be of little significance since their formation, especially that of fur-an-side monoadduct and that of the diadduct, very much depends on the wavelength of the UVA applied [27, 571. However, the distribution of DNA photoadducts after S-MOP+UVA treatment of yeast DNA in vitro and in uiuo or of calf thymus DNA in vitro is similar (Figs. 1 and 4); certainly the pyroneaide monoadducts constitute the minority product while the formation of diadducts increases with the UVA exposure dose (Fig. 4). The same was found by Zolan working with a monkey cell line and using the psoralen HMT [ 131. A detailed in uiuo study on the formation of S-MOP+UVA-induced DNA lesions and their biological consequences for yeast will be published shortly [ 551.

The biological consequences of UVA-induced S-MOP binding to RNA and protein should not be neglected. In vitro studies of psoralen binding to enzymes revealed many enzymes to be inactivated [46, 581; in the case of polymerase I of E. coli even mutagenicity was postulated [59]. However, taking into account the relatively low S-MOP binding capacity of proteins on the one hand and the marked negative effects on survival of DNA repair mutants on the other hand, we are inclined to play down the biological effects of S-MOP protein interaction.

Acknowledgments

We thank Ms. M. Marcovic for drawing the figures and the Willkomm- Stiftung for purchasing the labelled S-MOP. The data are from the doctoral thesis of M.B., who received a fellowship from the Graduiertenfijrderung of the Hessian State.

References

1 B. R. Scott, M. A. Pathak and G. R. Mohn, Molecular and genetic basis of furo- coumarin reactions, Mutat. Res., 39 (1976) 29 - 74.

2 E. Ben-Hur and P. S. Song, The photochemistry and photobiology of furocoumarins (psoralens), Adu. Radiat. Biol., 11 (1984) 131 - 171.

3 L. Musajo and G. Rodighiero, Studies on the photo-C-cycloaddition reactions be- tween skin photosensitizing furocoumarins and nucleic acids, Photochem. Photobiol., 11 (1970) 27 - 35.

50

4 F. Dall’Acqua, S. Marciani and G. Rodighiero, Interstrand cross-linkages occurring by the photoreaction between psoralen and DNA, FEBS Lett., 9 (1970) 121 - 123.

5 G. D. Cimino, H. B. Gamper, S. T. Isaacs and J. E. Hearst, Psoralens as photoreactive probes of nucleic acid structure and function: organic chemistry, photochemistry, and biochemistry, Ann. Rev. Biochem., 54 (1985) 1151 - 1193.

6 K. Yoshikawa, N. Mori, S. Sakakibara, N. Mizuno and P. S. Song, Photoconjugation of 8-methoxypsoralen with proteins, Photochem. Photobiol., 29 (1979) 1127 - 1133.

7 D. Averbeck, Photobiology of furocoumarins. In C. Helene, M. Charlier, Th. Montenay-Garestier and G. Laustriat (eds.), Trends in Photobiology, Plenum, New York, pp. 295 - 308.

8 N. Babudri, B. Pani, S. Venturini, M. Tamaro, C. Monti-Bragadin and F. Bordin, Mutation induction and killing of V79 Chinese hamster cells by 8-methoxypsoralen, Mutat. Res., 91 (1981) 391 - 394.

9 R. Chanet, C. Cassier and E. Moustacchi, Genetic control of the bypass of mono- adducts and of the repair of cross-links photoinduced by 8-methoxypsoralen in yeast, Mutat. Res., 145 (1985) 145 - 155.

10 D. Kanne, K. Straub, H. Rapoport and J. E. Hearst, Psoralen-deoxyribonucleic acid photoreaction. Characterization of the monoaddition products from 8-methoxy- psoralen, Biochemistry, 21 (1982) 861 - 871.

11 D. Kanne, K. Straub, J. E. Hearst and H. Rapoport, Isolation and characterization of pyrimidine-psoralen-pyrimidine photodiadducts from DNA, J. Am. Chem. Sot., 104 (1982) 6754 - 6764.

12 P. Vigny, F. Gaboriau, L. Voituriez and J. Cadet, Chemical structure of psoralen- nucleic acid photoadducts, Biochimie, 67 (1985) 317 - 325.

13 M. E. Zolan, C. A. Smith and P. C. Hanawalt, Formation and repair of furocoumarin adducts in (Y deoxyribonucleic-acid and bulk deoxyribonucleic acid of monkey cells, Biochemistry, 23 (1984) 63 - 69.

14 R. M. Santella, N. Dharmaraja, F. P. Gasparro and R. P. Edelson, Monoclonal anti- bodies to DNA modified by 8-methoxypsoralen and ultraviolet A light, Nulceic Acids Res., 13 (1985) 2533 - 2544.

15 D. Averbeck, Relationship between lesions photoinduced by mono- and bi-functional furocoumarins in DNA and genotoxic effects in diploid yeast, Mutat. Res., 58 (1985) 217 - 233.

16 D. Averbeck, E. Moustacchi and E. Bisagni, Biological effects and repair of damage photoinduced by a derivative of psoralen substituted at the 3,4 reaction site. Photo- reactivity of this compound and lethal effect in yeast, Biochim. Biophys. Acta, 518 (1978) 464 - 481.

17 E. Bender and M. Brendel, Effects of excess thymidylate on thymidylate low- requiring strains of Saccharomyces cerevisiae: high mutagenicity and absence of DNA strand breaks, Mutat. Res., 197 (1988) 59 - 66.

18 M. Brendel and R. H. Haynes, Interactions among genes controling sensitivity to radiation and alkylation in yeast, Mol. Gen. Genet., 125 (1973) 197 - 216.

19 W. W. Faith, M. Brendel, W. Laskowski and E. Lehmann-Brauns, Economizing DNA- specific labelling by deoxythymidine monophosphate in Saccharomyces cerevisiae, Mol. Gen. Genet., 132 (1974) 335 - 345.

20 R. Fleer and M. Brendel, Formation and fate of cross-links induced by polyfunctional anticancer drugs in yeast, Mol. Gen. Genet., 176 (1979) 41 - 52.

21 P. Prognon, J. Blais, P. Vigny, D. Averbeck, S. Averbeck and A. Gond, The meta- bolism of 8-methoxypsoralen by Saccharomyces cerevisiae. Evidence for an inducing effect of ethanol, I1 Farmaco, Ed. SC., 39 (1984) 739 - 751.

22 C. G. Hatchard and C. A. Parker, A new sensitive chemical actinometer. II. Potassium ferrioxalate as a standard chemical actinometer, Proc. R. Sot., London, Ser. A, 235 (1956) 518 - 536.

23 A. Ruhland and M. Brendel, Mutagenesis by cytostatic alkylating agents in yeast strains of differing repair capacities, Genetics, 92 (1979) 83 - 97.

51

24 N. R. Eaton, New press for disruption of microorganisms, J. Bacterial., 83 (1962) 1359 - 1360.

25 J. C. Mounolou, The properties of composition of yeast nucleic acids. In A. H. Rose and J. S. Harrison (eds.), The Yeasts, Vol. 2, Academic Press, New York, 1971, pp. 309 - 333.

26 R. F. Fowell, Sporulation and hybridization. In A. H. Rose and J. S. Harrison (eds.), The Yeasts, Vol. 1, Academic Press, New York, 1969, pp. 303 - 383.

27 J. W. Tessman, S. T. Isaacs and J. E. Hearst, Photochemistry of the furan-side 8- methoxypsoralen-thymidine monoadduct inside the DNA helix. Conversion to diadduct and to pyroneside-monoadduct, Biochemistry, 24 (1985) 1669 - 1676.

28 J. Cadet, L. Voituriez, J. Ulrich, P. C. Joshi and S. Y. Wang, Isolation and charac- terization of the mono-heterodimers of 8methoxypsoralen and thymidine involving the pyrone moiety, Photobiochem. Photobiophys., 8 (1984) 35 - 49.

29 B. S. Hahn, P. C. Joshi, L. S. Kan and S. Y. Wang, Heterodimers of psoralen and thymine derivatives: properties, structure and stereochemistry, Photobiochem. Photobiophys., 3 (1981) 113 - 124.

30 G. D. Cimino, Y. Shi and J. E. Hearst, Wavelength dependence for the photoreversal of a psoralen-DNA crosslink, Biochemistry, 25 (1986) 3013 - 3020.

31 P. C. Joshi, S. Y. Wang, W. R. Midden, L. Voituriez and J. Cadet, Heterodimers of 8-methoxypsoralen and thymine, Photobiochem. Photobiophys., 8 (1984) 51 - 60.

32 Y. Shi and J. E. Hearst, Wavelength dependence for the photoreactions of DNA- psoralen monoadducts. I. Photoreversal of monoadducts, Biochemistry, 25 (1987) 5895 - 5902.

33 S. C. Shim and Y. Z. Kim, Photoreaction of 8-methoxypsoralen with thymidine, Photochem. Photobiol., 38 (1983) 265 - 271.

34 G. Bernardi, Hydroxyapatite chromatography. In S. P. Colowick and N. 0. Kaplan (eds.), Methods in Enzymology, Vol. 21, Academic Press, New York, 1971, pp. 95 - 137.

35 J. Kalmakoff and C. C. Payne, A simple method for the separation of single-stranded and double-stranded RNA on hydroxyapatite, Anal. Biochem., 55 (1973) 26 - 33.

36 G. Bernardi, F. Carnevali, A. Nicolaieff, G. Piperno and G. Tecce, Separation and characterization of a satellite DNA from a yeast cytoplasmatic “petite” mutant, J. Mol. Biol., 37 (1968) 493 - 505.

37 G. Bernardi, Chromatography of nucleic acids on hydroxyapatite III. Chromato- graphy of RNA and polyribonucleotides, Biochim. Biophys. Acta, 174 (1969) 449 - 457.

38 G. Bernardi and T. Kawasaki. Chromatography of polypeptides and proteins on hydroxyapatite columns, Biochim. Biophys. Acta, 160 (1968) 301 - 310.

39 G. G. Markov and J. G. Ivanov, Hydroxyapatite column chromatography in proce- dures for isolation of purified DNA, Anal. Biochem., 59 (1974) 555 - 563.

40 P. I. Adriaenssens, C. J. Bixler and M. W. Anderson, Isolation and quantitation of DNA-bound benzo(a)pyrene metabolites: comparison of hydroxyapatite and precipi- tation procedures, Anal. Biochem., 123 (1982) 162 - 169.

41 M. Kircher and M. Brendel, DNA alkylation by mustard-gas in yeast strains of dif- ferent repair capacity, Chem. Biol. Interact., 44 (1983) 27 - 39.

42 F. M. Veronese, R. Bevilacqua, 0. Schiavon and G. Rodighiero, The binding of 8-methoxypsoralen by human serum albumin, Zl Farmaco, Ed. Sci., 33 (1978) 667 - 673.

43 F. M. Veronese, R. Bevilacqua, 0. Schiavon and G. Rodighiero, Drug-protein inter- action: plasma protein binding of furocoumarins, Zl Farmaco, Ed. Sci., 34 (1979) 716 - 725.

44 F. M. Veronese, 0. Schiavon, R. Bevilacqua, F. Bordin and G. Rodighiero, The effect of psoralens and angelicins on proteins in the presence of UV-A irradiation, Photochem. Photobiol., 34 (1981) 351 - 354.

45 S. Lerman, J. Megaw and I. Willis, The photoreactions of 8-methoxypsoralen with tryptophan and lens proteins, Photochem. Photobiol., 31 (1980) 235 - 242.

52

46 0. Schiavon, R. Simonic, S. Ronchi, R. Bevilacqua and F. M. Veronese, The modifi- cation of ribonuclease-A by near ultraviolet irradiation in the presence of psoralen, Photochem. Photobiol., 39 (1984) 25 - 30.

47 0. Schiavon and F. M. Veronese, Extensive crosslinking between subunits of oligo- merit proteins induced by furocoumarins plus UV-A irradiation, Photochem. Photo- biol., 43 (1986) 243 - 246.

48 L. Kittler, W. R. Midden and S. Y. Wang, Interactions of furocoumarins with subunits of cell constituents. Photoreaction of fatty acids and aromatic amino acids with trimethylpsoralen (TMP) and 8methoxypsoralen (8-MOP), Stud. Biophys., 114 (1986) 139 - 148.

49 A. Y. Potapenko, L. N. Bezdetnaya, E. P. Lysenko, V. L. Sukhorukov, A. N. Remisov and Y. A. Vladimirov, Mechanisms of furocoumarin-sensitized damage to biological membranes, Stud. Biophys., 114 (1986) 159 - 170.

50 K. Kittler and G. Lober, Photoreactions of furocoumarins with membrane consti- tuents. Results with fatty acids and artificial bilayers, Stud. Biophys., 101 (1984) 69 -72.

51 J. D. Laskin, E. Lee, E. J. Yurkow, D. L. Laskin and M. A. Gallo, A possible mecha- nism of psoralen phototoxicity not involving direct interaction with DNA, Proc. Natl. Acad. Sci. USA, 82 (1985) 6158 - 6162.

52 M. Sasaki, F. Meguro, E. Lumazawa, H. Fujita, H. Kakishima and T. Sakata, Evidence for uptake of 8-methoxypsoralen and 5-methoxypsoralen by cellular nuclei, Mutat. Res., 197 (1988) 51 - 58.

53 R. Hall, P. Nagley and A. Linnane, Biogenesis of mitochondria, Mol. Gen. Genet., 145 (1976) 169 - 175.

54 N. Magaiia-Schwencke, J. A. P. Henriques, R. Chanet and E. Moustacchi, The fate of 8-methoxypsoralen photo-induced crosslinks in nuclear and mitochondrial yeast DNA; comparison of wild-type and repair-deficient strains, Proc. Natl. Acad. Sci. USA, 79 (1982) 1722 - 1726.

55 M. Bankmann and M. Brendel, Molecular dosimetry of 8-MOP+UVA induced DNA- photoadducts in Saccharomyces cerevisiae: correlation of lesions number with geno- toxic potential, J. Photochem. Photobiol., B: Biol., to be published.

56 Y. Shi and J. E. Hearst, Wavelength dependence for the photoreactions of DNA- psoralen monoadducts. II. Photocross-linking of monoadducts, Biochemistry, 26 (1987) 3792 - 3798.

57 P. K. Chatterjee and C. R. Cantor, Photochemical production of psoralen-DNA monoadducts capable of subsequent photo-crosslinking, Nucleic Acids Res., 5 (1978) 3619 - 3633.

58 F. M. Veronese, 0. Schiavon, R. Bevilacqua, F. Bordin and G. Rodighiero, Photo- inactivation of enzymes by linear and angular furocoumarins, Photochem. Photobiol., 36 (1982) 25 - 30.

59 M. Granger and C. Helene, Photoaddition of 8-methoxypsoralen to E. coli DNA polymerase I. Role of psoralen photoadducts in the photosensitized alterations of Poll enzymatic activities, Photochem. Photobiol., 38 (1983) 563 - 568.