Journal of Photochemistry and Photobiology, B: Biology, 7 (1990) 317--336

317

PHOTOSENSITIZATION OF D N A OF D E F I N E D SEQUENCE BY FUROCHROMONES, KHELLIN A N D VISNAGIN L. TRABALT,INI* Institut Curie-Biologie (Unitd de Recherche associde au CNRS 1292), 26 rue d'Ulm, F-75231 Paris Cddex 05 (France) P. MARTELLI and L. BOVALINI Institute of Biological C-Tle'mistry, University of Sienc~ Plan dei Mantellini 44, 1-53100 Siena (Italy) F. DALL'ACQUA Department of Pharmaceutical Sciv~es, Padova University, 1-35100 Padova (Italy) E. SAGE Institut Curie-Biologic (-Unitd de Recherche associ~e au CNR3 1292), 26 rue d'Ulm, I:'-75231 Paris Cddex 05 (France) (Received January 16, 1990; accepted April 4, 1990)

K e y w o r d s . Furochromones, furocoumarins, DNA adducts, photodynamic effect.

Summary The sequence specificity in the i n vi t ro DNA photobinding of khellin and visnagin, two naturally occurring furochromones p r o p o s e d for chemotherapy of vitiligo, waS investigated by using DNA sequencing methodology. The 3 ' - 5 ' exonuclease associated with the T4 DNA polymerase served as a tool for determining photoadducts distribution on DNA fragments of the /ac I gene of E s c h e r i c h i a coll. The photoadduct distribution of psoralen is also studied for comparison. Upon UVA irradiation, visnagin mainly forms monoadducts with thymine and to a lower extent with cytosine. Alternating (A-T), sequences are hot spots for visnagin photoaddition. This is a property shared with furocoumarins. T I T sites are also quite reactive to visnagin, as they are to methylated angelicins. In contrast, with psoralen derivatives, there is no preferential photobinding in 5'-TpA sites, and 5'-APT sites react as well. Furthermore, many sites such as T in the GC context, and C in any context, react, although weakly. The visnagin photoadduct distribution resembles very much the photoadduct distribution of methylated a~gelicins as tPresent address: Institute of Biological Chemist~, University of Siena, Plan dei Mantellini 44, 1-53100 Siena, Italy.

I011-1344/90/$3.50

© Elsevier Sequoia/Printed in The Netherlands

318 described by Miolo et al. The photoreaction of these two series of compounds is less sequence dependent than the photobinding of psoralen derivatives as described by Sage and Moustacchi and by Boyer et al. The sequence specificity in khellin-DNA photobinding is the same as for visnagin, even though it forms much fewer photoadducts. The absence of photo-oxidation of DNA after treatment with visnagin or khellin plus UVA suggests that furochromones do not present any photodynamic effect on DNA.

1. I n t r o d u c t i o n

Khellin and visnagin are naturally occuring furochromones extracted f r o m A m m i v i s n a g a L.(Lam.) [1]. Young green achenes o f A . v i s n a g a have the highest content in furochromones [2]. In ripe fruits, khellin and visnagin are principally localized in primary rib channels [3]. Among furochromones, kheUin has received particular attention since it is the major compound and it was used in the past as a coronary vasodilator. From the structural similarity of furochromones and furocoumarins, one may anticipate that furochromones would exhibit similar photochemical and phototherapeutic properties as furocoumarins. Indeed, khellin has been proven to be at least as effective as psoralen derivatives in the photochemotherapy of vitiligo [4-7]. The lack of long-term side effects and phototoxic skin erythema response makes kheUin a valuable alternative to the use of psoralen derivatives. Furthermore the low genotoxicity of khellin, in comparison with psoralen derivatives, is another advantage in the use of khellin in photochemotherapy, instead of bifunctional psoralens [6, 8, 9]. This encouraged the synthesis of new derivatives of furochromones as more therapeutically active and less toxic molecules [ 10]. As furocoumarins, khellin undergoes an intercalation complex with DNA in the dark; however, its DNA photobinding upon UVA irradiation is very low compared with psoralen derivatives [8, 11-13]. Interstrand DNA crosslinks induced by khellin and visnagin have been observed i n vitro at high UVA doses [11, 13, 14], but not in mammalian cells [6]. Khellin is also a poor type I (radicals) and type II (singlet oxygen) photodynamic sensitizer, whereas visnagin is better [15]. In respect with the use of kheUin in photochemotherapy, it is of interest to obtain more information on the mode of interaction of furochromones with DNA, since therapeutic activity of such compound has often been associated with their photoreaction with DNA. This paper mainly reports the sequence specificity in the photobinding of furochromones with DNA. Enzymatic assay coupled with DNA sequencing methodology serve as a tool to map at the nucleotide level and quantitative photolesions in DNA fragments of known sequence [ 16-18]. Potential photoreactions of furochromones with DNA, involving activated oxygen species, are also tested. With regards to the very low photoreactivity of khellin, most of the photochemical experiments are performed with visnagin, once we observed that khellin and visnagin

319

have the same sequence specificities. A comparison with psoralen is also reported.

2. M a t e r i a l s

and methods

2.1. Products Khellin and visnagin were commercial products from Fluka AG, Buchs (Switzerland) and Inverni and Della Beffa, Milan (Italy) respectively. Psoralen was from Sigma Chemical Co, St. Louis, MO (U.S,k.). The molecular structure of the tested compound is shown in Fig. 1. Restriction enzymes were from Boehringer (Mannheim). T4 DNA polymerase and exonuclease III ofEscherichia coli were from Bethesda Research Laboratory (U.S~.). 2.2. DNA substrates 5'-32P-end-labelled DNA fragments of interest were prepared as previously described [17]. The 49, 76 and 139 base pairs, derived from replicative form (RF) DNAs M13 mp8 /ac I 935 and 225-3, span the 1-49, 8 6 0 - 9 3 7 and 7 0 - 2 0 9 positions in the /ac I gene of E. coli respectively [16]. 2.3. Photoreaction of f u r o c h r o m o n e s w i t h DNA A standard reaction consisted of adding 2 ~1 of a khellin or visnagin solution at a concentration of 5 × 10 -8 M in DMSO to 18 p2 of a solution of labelled DNA in a buffer of 10 mM tris-hydrochloric acid (pH 7.5) and 1 mM ethylene diaminetetracetic acid. The mixture was kept in the dark for 30 min at room temperature and then irradiated at different UV doses using four HPW 125 Philips lamps mainly emitting at 365 nm (UVA) at a fluence of 20 J m - ~ s - 1. After irradiation, the unreacted furochromones were eliminated by two chloroform-isoamyl alcohol (19:1 v/v) extractions followed by ethanol precipitation. Psoralen was used as the reference compound and the photoreaction was done similarly. Untreated DNA and DNA incubated in the presence of furochromone without irradiation were prepared as control samples.

CH_

0

CH a

R R=-OCH 3

Khe||in

R= - H

Visnagin

Fig. 1. Molecular s t r u c t u r e o f t e s t e d c o m p o u n d s .

Psoralen

320

2.4. Alkali-labile lesions assay An aliquot of modified DNA was treated with 25 ]zl of an aqueous solution of 1 M piperidine at 90 ° C for 20 min. The lyophilization of piperidine was followed by three washes with 100 ~1 and two washes with 50 ~1 of water. Cleavage products were analysed on sequencing gels.

2.5. Digestion of modified DNA with T4 DNA polymerase (3'-5') exonuclease The modified DNA was dissolved in 20 ]~1 of 33 mM tris-acetate (pH 7.8), 10 mM magnesium acetate, 66 mM potassium acetate, 0.5 mM dithiothreitol and 0.1 mg ml-1 bovine serum albumin and digested with four units of T4 DNA polymerase ( 3 ' - 5 ' ) exonuclease for 2 h at 37 °C. The reaction was stopped by adding 1 ~g of carrier DNA. An aliquot of the reaction mixture was lyophylized. In the remaining part of the sample, the enzyme was eliminated by two chloroform-isoamyl alcohol extractions, followed by ethanol precipitation. The samples were then redissolved in water and photoreversion of potential cross-links or adducts was performed by irradiating with an UV dose of 6 kJ m -2 at 254 rim. This eliminates the effectof the presence of the furochromone on the migration of termination products.

2.6. Analysis o f sites o f photolesions by sequencing gel electrophoresis Both termination and chemical cleavage products were analysed on denaturing (7 M urea) polyacrylamide gel alongside the Maxam-Gilbert sequencing reaction products. The polyacrylamide concentration of the gels was 12% or 20% according to the length of the DNA fragments to be analysed. The termination products were visualized as bands on autoradiograms. The quantitative analysis of photolesions was performed by cutting out the corresponding bands from the gel and measuring their radioactivity by Cerenkov counting. Quantification of frequency of photoaddition slightly differs from that described by Sage and Moustacchi [17]. After the T4 DNA polymerase digestion, the reaction mixture was shared into two aliquots: one third of the mixture was directly lyophylized, whereas on the remaining two thirds was performed a chloroform-isoamyl alcohol extraction and successive ethanol precipitation. While extraction and precipitation were useful to free the termination products of enzymes and salts, the lyophylization avoided the loss of fully digested non-reacted molecules after ethanol precipitation and allowed the exact determination of the ratio of modified to unmodified DNA for each sample. Lyophylized and precipitated samples were loaded onto different gels. A short run of the gel where lyophylized samples were loaded allowed separation of unmodified molecules, which consequently were fully digested, from the adducted DNA molecules which were partially digested (termination at photoadducts). The total amount of photoadducts per DNA molecule was calculated according to a Polsson distribution. The fraction of radioactivity corresponding to DNA molecules with no photoadducts (fully

321 digested molecules) is the first term of the Poisson distribution. This gives the mean value of photoadducts per DNA molecule. The absence of salts in the samples and a long run of the gel where migrated precipitated DNA samples ensure a better resolution of bands, compared with the previous gel. This allows the assignment of sites of photoaddition by comparison with the Maxam--Gilbert sequencing ladder. Quantitation was also performed. The combination of results generated from both gels allow an exact quantification of frequency of photoaddition at each site.

2. 7. E x o n u c l e a s e III a s s a y The photoreaction of furochromones with supercoiled DNA (form I) was performed as described above. An aliquot of the modified DNA was digested in 50 ;zl of the buffer recommended by the manufacturer, with 2 units of E. coli exonuclease III for 30 min at 37 °C, while another aliquot was incubated in the buffer in the absence of enzyme. The same treatment was performed on untreated DNA and apurinic/apyrimidic (AP) DNA. Digested and undigested DNA samples were loaded on 0.8% agavose gel containing ethidium bromide. The presence of photolesions sensitive to exonuclease III is revealed by the conversion of supercoiled RF I DNA to circular opened RF II DNA.

3. R e s u l t s

3.1. S e q u e n c e specificity in D N A p h o t o b i n d i n g o f kheUin a n d v i s n a g i n Sites of photoaddition of khellin and visnagin were studied in 5'-endlabelled DNA fragments of known sequence, using DNA sequencing methodology. The 3 ' - 5 ' exonuclease activity associated with DNA polymerase of phage T4 is a good probe for mapping and quantifying the photoaddition of furochromones, as previously shown for furocoumavins [16-18]. This enzyme terminates its digestion of a DNA strand at an adducted base and is not blocked on the opposite strand. After enzymatic digestion of treated or untreated DNA fragments, the resulting termination products were analysed by polyacrylamide gel electrophoresis under denaturing conditions, together with the Maxam--Gilbert sequencing reaction products and control samples. An example of band pattern corresponding to exonuclease termination sites at photoadducts is given in Fig. 2 and reveals the sites of photobinding of visnagin and kheUin in a 139-base-pair DNA fragment. Lanes 1 and 2 exhibit a pattern of bands of different intensities, corresponding to termination sites at visnagin photoadducts. Visnagin photoreacts with DNA at many sites which ave scattered all along the DNA fragment. In lanes 4 and 5, which correspond to the photoreaction of khellin with DNA, only a few bands ave visible. They correspond to the strongest bands seen in lanes 1 and 2. This illustrates the poor photoreactivity of khellin with DNA compared with the photoreactivity of visnagin. The dark interaction of khellin or visnagin with DNA does not

322

1

2

3 4 5

G G A C CT

ACCA

TTT

TTAT T

TAT

/

Fig. 2. Qualitative analysis of the sites of khellin and vtsnagin photoaddttion in the 139obasepair DNA fragment. 5'-end-labelled DNA was incubated in the presence of 5 × 10 -4 M visnagin (lanes 1 and 2) or khellin (lanes 3-5) and UVA irradiated at doses of 72 kJ m -2 (lane 1), 144 kJ m -2 (lanes 2 and 4), 288 kJ m -2 (lane 5) or unirradiated (lane 3). All the samples were digested with T4 DNA polymerase ( 3 ' - 5 ' ) exonuclease. The termination products were resolved on 1296 polyacrylamide-urea 7' M gel, together with the Maxam-Gilbert sequencing reaction products. l e a d t o p r e m a t u r e t e r m i n a t i o n o f t h e e x o n u c l e a s e ( l a n e 3). T h e a b s e n c e o f m o l e c u l e s m i g r a t i n g as a d o u b l e - s t r a n d DNA f r a g m e n t d e m o n s t r a t e s that u n d e r o u r c o n d i t i o n s n e i t h e r v i s n a g i n n o r khellin form i n t e r s t r a n d DNA crosslinks. T a k i n g into a c c o u n t the low e x t e n t of khellin p h o t o a d d i t i o n to DNA

323

after exposure to UVA light, most of the experiments were performed with visnagin. The comparison of the termination sites with the Maxam--Gilbert sequencing ladder allows the assignment of sites of photoaddition. We consider a shift of 1-1.5 nucleotides in the migration of termination products due to a difference in the 3' terminus of digested and of chemically cleaved DNA fragments, as explained in ref. 16. A quantitative analysis of photolesions was performed as described in material and methods. This is illustrated in Fig. 3 which shows the distribution of visnagin photoadducts on both strands

o,

ll,.

II

,.,

s.GAATGTGAA AC CAGTAACGTTATACGATGTCGCAGAGTATGCCGGTGTCTC'~'A"TC,AGA CT TACACTT__.TG GTCA¥_TGCAATATGCTAC,AGC,GTC T CATA

TCT

I

0

I , I

|

5

GGC AAAP~CGC~CCACT T G G T C C G G T C G G T G ~

O~

'

'

T,TTC~CGCCC~/CrTC5

i

'

2L ~'~.g. 3. Distribution of visnagin p h o t o a d d u e t s in the 139-base-pair DNA fragment. The photoreaction was performed using a n excess of visnagin (standard protocol) a n d a n Irradiation dose of 144 kJ m -2 for b o t h strands. The total a m o u n t of adducts p e r DNA molecule is 0.4 for the 5'--3' s t r a n d a n d 0.3 for the 3 ' - 5 ' strand. The height of b a r s r e p r e s e n t s t h e frequency of visnagin photoaddltion at each reactive site. Single b a r a p p e a r s in c o r r e s p o n d e n c e to underlined site w h e n t h e assignment of b a n d s to specific 'Is was uncertain.

324 of the 139-base-pair DNA fragment. For each experiment, two UVA doses were systematically given. Maps of photoadduct distribution corresponding to the highest level of photobinding is generally reported. The following points are observed from Fig. 3: (1) AT-rich sequences are the strongest sites of photobinding of visnagin, i.e. TTATA, TATAA, TTATC, TAT and ATA; (2) visnagin photobinding on thymine is favoured in runs of pyrimidine--purine, i.e. the strongest site, 5'-TATA, in the lower strand; (3) the sequence context influences the photoreaction of visnagin, i.e. the difference in sensitivity of the two ATA sites in the lower strand; (4) runs of 3 Ts are rather strongly reactive, whereas runs of 4 Ts or 5 Ts are weak; (5) there is no obvious difference between the frequencies of photoaddition of 5'-APT sites and 5'TpA sites; (6) Ts surrounded by G or C are quite reactive; (7) photoaddition at Cs occurs at low frequencies and at AC or CA sites. No difference between the frequencies of total photoadduct and adduct distribution was observed when photoreaction was performed in a 10 mM sodium phosphate buffer (pH 7.4) (not shown). Since furochromones are supposed to intercalate in DNA [13], as furocoumarins do, it was of interest to study their photoreaction on supercoiled DNA. RF M13 mp8 /ac I 225-3, where the 139-base-pair DNA fragment is issued from, was photoreacted with visnagin and then the 5'-end-labelled DNA fragment is prepared as usual and digested with T4 DNA polymerase 3 ' - 5 ' exonuclease. For comparison 5'-end-labelled 139-base-pair DNA fragment 0inear DNA) which had already been purified was treated at the same drug-to-DNA ratio and the same UVA dose as supercoiled DNA. Figure 4(a) reveals that the same sites are reactive to visnagin in linear or supercoiled DN.~ A similar experiment performed with psoralen leads to the same conclusion (Fig. 403)) and demonstrates that it is a general feature for these families of compounds. If linear DNA (empty bars) seems more reactive at most of the sites than supercoiled DNA (full bars) does, it is evident that runs of Ts become more reactive in supercoiled DNA. This is particularly striking for psoralen which poorly binds to T r r sites in linear DNA. The photoreaction of furochromones with linear 139-base-pair DNA fragment was also performed at different drug-to-DNA ratios (Figs 5(a) and 503)). The distribution and frequency of photoaddition can be compared with those obtained under standard saturating conditions (full bars in Fig. 503), and Fig. 3). As expected, at a low visnagin-to-DNA ratio, photoadducts show up only at the strongest sites observed in Fig. 3. The same data are depicted for khellin in Fig. 503) (compare empty bars with full bars). Furthermore, as already observed in Fig. 2, khellin is much less reactive than visnagin. Indeed, a dose of 288 kJ m -2 was necessary to produce 0.04 kheUin photoadduct per DNA molecule, whereas ten times more photoadducts are formed when the same DNA is irradiated in the presence of visnagin at a UVA dose of 144 kJ m -2. Moreover, under standard saturating condition, kheUin photoadducts are only detected at the strong sites of photoreaction of visnagin. The weak sites would probably show up after longer exposure to UVA~ Figures 5(a) and 503) demonstrate that photobinding of khellin to

325 16

A

CGTTATACGATGTCGCAGAGTATGCCGGTGTCTCTTATCAGAC CG TCCCG/kCG"I"TTCTC,C -

-

-

-

-

-

3"

Fig. 4. Comparative analysis of the frequency of photoaddition of visnagin and psoralen at different sites on supercoiled DNA (full bars) and linear DNA (empty bars). 1 /~g of supercoiled DNA CRF M13 m p 8 /ac I 225.3) was irradiated in the p r e s e n c e of (a) 5 × 1 0 -4 M visnagin or (b) 5 × 1 0 -4 M psoralen at doses of 144 kJ m -~ and 72 kJ m -2 respectively. After the photoreaction, the 5'-end-labelled 139-base-pair DNA fragment was obtained from modified DNA using standard procedures. Linear DNA (5'.end-labelled 139-base-pair DNA fragment) was irradiated in the presence of 5 × 10 -4 M visnagin and psoralen at the doses of UVA light applied for the photoreaction with supercoiled DNA. W e deleted a part of the sequence corresponding to the arrow.

DNA exhibits the same s e que nc e specificity as the photobinding of visnagin, although khellin is m u c h less photoreactive.

3.2. Comparison with psoralen and other furocoumarins In order to com pa r e khellin and visnagin photoreactivity with furocoumatins, we used psoralen as reference c o m p o u n d (Fig. 5(c)). In addition, data will later be discussed in correlation with those obtained for angelicins [ 16 ] and others psoralen derivatives [ 18 ]. We observed a relevant difference in the extent of photoreactivity. In the p h o t o r e a c t i o n of psoralen with the 139-base-pair DNA fragment, 1.24 p h o to ad d u cts p er DNA molecule are f o r m e d after irradiating at a dose of 72 kJ m - e of UVA fight while, for visnagin, only 0.4 adducts per DNA molecule are f o r m ed after a dose of 144 kJ m -2 of UVA light. The co mp ari s on of Fig. 5(c) with Figs. 5(a) and 5(b) and Fig. 3 demonstrates a different distribution of phot oadduct s for khellin or visnagin and for psoralen. Although psoralen has a strong affinity for DNA, only a few sites show up, i.e. mainly alternated A-T-rich sequences, which are also

326 1.4,

A

0.~ 0 5.CGTTATA CGATGTCGCAGAGTATGCCGGTGTCTCTTATCAGACCGTTTCCCGACGTTTCTGC 3,

o

I

.,

=

J

rl

5.GTGAAACCAGTAACGTFATACGATGTGGCAGAGTATGCCGGTGTC'rCTTATCAGACGGTTTCCCGCGTTTCTGC3' 3 GACTT T GGT~T._TTGCAATATGCTACAGCGTGT CATACGGCCACAGAGAATAGTC TGGCJU~GGGCAGCAAAGACG5,

i. I'

'

II

I

3C

o

.I

I

I

,

s,CGTTATA C GATGTC GCAGAGTATGCCGGTGTC'rCTTATCAGAC CGTTTCCCG^ CI~rTTCTGC3.

Fig. 5. Analysis of photoreaction between the 139-base-pair DNA fragment and visnagin or khellin at different drug-to-DNA ratios, and comparison with psoralen. (a) A mixture of 5'end-labelled DNA plus 100 ng of sonicated calf thymus DNA was UVA irradiated at a dose of 144 kJ m -2 in the presence of visnagin at various concentrations. The ratios visnagin to DNA were 1:10 (full bars), 1:5 (lined bars) and l : l (empty bars). The number of visnagin photoadducts per DNA molecule was 0.02, 0.03 and 0.06 corresponding to photoreaction at the ratios 1:10, 1:5 and 1:1 respectively. (b) DNA was irradiated in the presence of khellin in excess (full bars) or at a khellin-to-DNA ratio of 1:1 (empty bars), at UVA doses of 288 kJ m -2 (upper strand) and 216 kJ m -2 (lower strand). The number of khellin photoadducts per DNA molecule was 0.02, 0.04 and 0.02 corresponding respectively to photoreaction in the upper strand at a ratio of 1:1 and saturation, and in the lower strand at saturation. (c) 5'-end-labelled DNA was treated with 5 × 10 -4 M psoralen and irradiated at UVA dose of 72 kJ m -z, leading to 1.2 photoadducts per DNA molecule.

the strongest sites for visnagin photobinding. The photobinding of psoralen is v e r y s e q u e n c e s p e c i f i c . T h i s f e a t u r e h a s a l r e a d y b e e n o b s e r v e d b y u s a n d is a c h a r a c t e r i s t i c o f t h e p s o r a l e n s e r i e s . It h a s t o b e n o t e d t h a t , a t s i t e 5 ' TATA, t h e l o w e x t e n t o f p s o r a l e n p h o t o a d d i t i o n o b s e r v e d o n t h e s t r a n d s h o w n i n Fig. 5 ( c ) is l i k e l y t o r e s u l t f r o m f o r m a t i o n o f m o s t p h o t o a d d u e t s o n t h e c o m p l e m e n t a r y s e q u e n c e , a s o b s e r v e d f o r a n g e l i c i n s [ l 9]. T h e t w o T T T s i t e s are much less reactive to psoralen than to khellin or visnagin.

327

Other D N A sequences have been analysed for their reactivity towards these derivatives.Figure 6 is a comparison of visnagin (fullbars) and psoralen (empty bars) photoadduct distribution in the 49-base-pair (A) and 76-basepair (B) D N A fragment. The same features as above are observed: ( 1 ) m u c h fewer photoadducts are induced by visnagin than by psoralen; (2) m a n y more sites react to visnagin than to psoralen; (3) photobinding of psoralen is m u c h more sequence specific, and only the A-T-rich sites are strongly reactive; (4) no preferential photobinding of visnagin at 5'-TpA sites vs. 5'A p T sites is observed, while a low extent of psoralen photoadducts are detected at 5'-APT, according to what was previously observed with derivatives of the psoralen series [17, 18]. Psoralen does not photoreact either at C or at T positions in a GC context, in contrast with visnagin. In conclusion, these observations lead us to conclude that visnagin and khellin are very poorly photoreactive, compared with furocoumarins; their photobinding to DNA is less sequence specific than the photoaddition of psoralen derivatives. Although alternated A-T sequences are the hot spot of photoaddition, as for furocoumarins, furochromones photoreact with runs

2.8

0

-A

1, I,,

GGACACCATCGAA~GCAAAACCTTTCGcGGT--A~GGCATGATAGCG 3, 37

,,,,,

,

°sGAACGGCGGGATATAACATGAGCTGTCTTCGGTATCGT-CGTATCCCACq~ATATCCGCACCAACGC 3~ Fig. 6. Comparative analysis of photoaddition of visnagin (fullbars) and psoralen (empty bars) in the (a) 49- and Co) 76-base-pair D N A fragments. The photoreaction was performed under standard conditions. The irradiation doses were 144 kJ m -2 for the isn gin and 72 kJ m -~ for the psoralen. After the T4 D N A polymerase (3'-5') exonuclease digestion,the termination products were resolved on 2 0 % polyacrylamide-urea 7 M gel. The amount of vlsnagin photoadducts was 0.47 per 49-base-pair D N A fragment and 0.71 per 76-base-pair D N A fragment. The amount of psoralen photoadducts was 0.35 per 49-base-palr D N A molecule, and 1.75 per 76-base-pair D N A molecule.

328 of three T and with T sites in a GC context. They do not exhibit any preference for 5'-TpA vs. 5'-APT.

3.3. P h o t o d y n a m i c effect o f kheUin a n d v i s n a g i n Beside covalent binding to DNA, certain furocoumarins m ay photoinduce a n o t h e r type o f DNA damage involving oxygen, i.e. photo-oxidation of guanine [ 2 0 - 2 2 ]. The capacity of various furocoumarins to generate activated oxygen species have been determined [ 2 3 - 2 7 ] . Since visnagln and khellin at a lower extent have b ee n r e p o r t e d as pr oducer s of singlet oxygen and superoxide radicals [15], it was of interest to test their ability to induce photo-oxidation in DNA. Two different a p p r o a c h e s were used, as described by Sage et al. [20] for 3-carbethoxypsoralen (3-CPs). Since some of the oxidation products in DNA have b een demonstrated to be sensitive to hot alkali, a chemical cleavage by hot piperidine and an analysis of the cleavage products on sequencing gel were used to test the potential photodynamic effect of furochromones. The 139-base-pair 5'-end-labelled DNA fragment was left untreated or incubated in the presence of f u r o c h r o m o n e s without further irradiation or UVA irradiated at doses which pr oduc e phot oa d duct s (as in Figs. 3, 5(a) and 5Co)) and e x p o s e d to 1 M piperidine at 90 °C for 20 min. We did not detect any cleavage at G positions, nor at other positions, in DNA treated with f u ro ch r o mo n es plus UVA. No difference was observed between treated samples and control samples (untreated or unirradiated) (autoradiograms not shown). Since tris(hydroxymethyl)aminomethane present in the irradiation mixture is a good radical scavenger and can form adduct during photo-oxidation of DNA, the same experiments were p e r f o r m e d in a 10 mM sodium phosphat e buffer. Again, we did not observe any cleavage above the background. The AP-endonuclease activity of exonuclease III of E. coli has been demonstrated to recognize photo-oxidation induced by 3-CPs and cleave 3CPs-treated DNA at G positions [20]. We used the conversion of supercoiled DNA (RF I DNA) to circular o p e n e d DNA (RF II DNA) to detect an eventual cleavage by E. coli exonuclease III after treatment of supercoiled DNA with f u r o c h r o m o n e plus UVA. Psoralen, known as a p r o d u c e r of singlet oxygen, was used as a positive control [23]. DNA-containing apurinic sites served as a test for the enzyme. The results are given in Table 1. Photoreaction of DNA with visnagin does not p r o d u c e sensitive sites to exonuclease III, since the ratio of RF II to RF I does not vary significantly; this also applies to khellin (not shown). On the contrary, this ratio increases after irradiation of RF I DNA at a UVA dose of 144 kJ m -~ in the presence of psoralen. The psoralen tr ea t m e nt by itseff induces singie-strand breaks after incubation at 37 °C. The ratio of RF II to RF I is slightly increased after digestion of the same DNA with exonuclease III. Cleavage by the enzyme probably occurs at photo-oxidized G p r o d u c e d by singiet oxygen. Assuming that there is a Poisson distribution of damages on the different DNA molecules, one can estimate that 0.6 single-strand breaks are p r o d u c e d after exposure of RF I DNA to psoralen plus 144 kJ m -2 of UVA light and subsequent digestion

329

=

C~ ~O

b.-

0

u

•

•

o

C~

0

C~ C3

0ul

q) O

C

C~

et

C

× ~3

r,..

et

C

OC

U

c~

O0

0

I:1

0 N

c I

330 with exonuclease HI. No cleavage was observed at lower UVA doses. For comparison, the same enzymatic treatment of RF I DNA irradiated in the p res en ce of 3-CPs at a dose of 36 kJ m -2 or in the presence of acetylpsoralen at a UVA dose of 216 kJ m -2 leads to 0.9 and 0.5 single-strand breaks per DNA molecule respectively. The relative importance of the photodynamic effect of the tested c o m p o u n d s on DNA is 3-CPs > psoralen > acetylpsoralen >> visnagin, khellin The photodynamic effect of visnagin on DNA is irrelevant, at least under our reaction conditions.

4. D i s c u s s i o n In this paper, we used an enzymatic a p p r o a c h coupled with DNA sequencing meth odol ogy to study the sequence specificity in the i n v i t r o photobinding o f furochromones, visnagin and khellin to DNA. The substrates were three DNA fragments of length 49, 76 and 139 base pairs. They are parts o f the /ac I gene of E. coli, a target gene often chosen to determine mutational specificity of drugs, including furocoumarins [ 2 9 - 3 2 ] . The 139base-pair DNA fragment is of particular interest since it contains most of the mutation sites induced by mutagens [33]. The tool which served to map photoadducts and determine the frequency of photoaddition at each site was the 3 ' - 5 ' exonuclease activity associated to the DNA polymerase of phase T4. It already served to study the photobinding of psoralen derivatives and angelicins [ 1 6 - 1 8 , 34]. This enzymatic activity has been proven to be useful to study the distribution of bulky adducts on DNA precisely [35, 36]. Other exonucleases have been tested to map psoralen photoadducts, i.e. exonuclease HI of E. coli, 5 ' - 3 ' exonuclease from bacteriophage, or nuclease Bal 31 [ 37- 39] . These enzymes may exhibit either non-specific stops or bypass of m onoadduct s [38, 40--42]. The nuclease Bal 31 bypasses monoa dduc t s and is blocked only by interstrand cross-links [43]. The Uvr ABC endonuclease which is in charge of excision and repair of psoralen photoadducts in E. coli, could be used to map phot oadduct s [44, 45]. However this enzymatic complex is not commercially available. In previous work, we have shown that at least for psoralen phot oadduct s the difficulties raised by the use of other enzymes are not encount ered with the 3 ' - 5 ' exonuclease of T4 DNA polymerase [17]. Khellin and visnagin are found to form a low extent of photoadducts, c o m p a r e d with psoralen (this study) and other furocoumarins [16]. KheUin even p r o d u ces at least ten times less photoadducts than visnagin does. A very low extent of i n v i t r o and i n v i v o photobinding of khellin to DNA has been r ep o r ted [13]. This is in line with its low affinity for DNA and low rate constant [13, 46]. A steric hindrance due to the two m e t h o x y groups at positions C5 and C8 may explain this low photoreaction of khellin.

331

Under our experimental conditions, no formation of interstrand DNA cross-links was observed for visnagin or for khellin as in ref. 6. Several studies have reported the photoinduction of cross-links by furochromones, although far less efficiently than by bifunctional psoralen derivatives [11, 13, 14]. It has to be noted that in most of these studies cross-links were barely detectable at a UVA dose as high as 520 kJ m -2. The fact that khellin, and visnagin to a lower extent, are poor sensitizers probably explains their low genotoxicity. Furthermore khellin and visnagin behave as monofunctional agents. Indeed, it is admitted that furocoumarins bioadducts are more genotoxic than monoadducts and more effective for inducting recombination [11, 47]. It has been reported that khellin is less genotoxic in yeast than bifunctional 5-methoxypsoralen (5-MOP) [6] and that the relative order of toxicity in bacteria, fungi and Chinese hamster ovary cells is 8-methoxypsoralen (8-MOP) > visnagin > khellin [8]. A similar order in ant[viral potency has been observed [9]. Visnagin is shown to photoreact mainly with thymine, although photobinding at cytosine is also observed at a low extent (Figs. 3 and 6). Owing to their chemical nature, furochromones are expected to photoreact with DNA as do psoralen derivatives. Indeed, these compounds form an intercalation complex with DNA in the dark [12]. As for furocoumarins, the main adduct is a c / s - s y n furan-side thymine monoadduct, whereas a pyrone-side monoadduct may also be present [8, 13]. We demonstrate that khellin and visnagin exhibit a preferential photobinding to alternated (A-T)~ sites. Run of 3 Ts are also quite reactive. This is a feature shared with psoralen derivatives [17, 18] and angelicins [ 16]. These types of sequence were defined as "strong sites" of furocoumarin photobinding. This implies that furochromones follow the same general rule for photoaddition as furocoumarins. In fact, the few sites which react with visnagin at low visnagin-to-DNA ratio or with khellin are the only sites which react with psoralen and psoralen derivatives (Fig. 5) [17, 18]. The frequencies of photoadducts at these sites slightly differ between the two series of compounds. The extent of photoreaction at a potential site is greatly influenced by the flanking sequence. The local DNA conformation conferred by an alternance of pyrimidine-purine favors photoreaction of furochromones. This is illustrated in Fig. 3 by the difference between the reactivities of the two ATA sites which are in a different context. The role of DNA conformation may also explain the difference between reactivity of the site T I ' r c c c towards visnagin and the reactivity towards psoralen in supercoiled and linear DNA (Fig. 4). Nevertheless, some substantial differences between the distribution of visnagin and psoralen photoadducts exist, as is apparent in Fig. 6. Many minor sites of visnagin photoaddition show up. An even more striking difference is that visnagin photoreacts in 5'-APT, as well as in 5'-TpA, if not better, whereas there is a strong preference of psoralen derivatives for 5'-TpA sites (Fig. 6) [17, 48].

332 The visnagin photoadduct distribution on the 49-, 76- and 139-basepair DNA fragments is very close to the TMA photoadduct distribution. The absence of preferential photoreaction at 5'-TpA vs. 5'-APT, the photoreaction at T surrounded by G or C, the photoreaction at C in any context, the high photobinding at run of 3 Ts and the low photoaifanity at Tn ( n = 4 , 5) are also the characteristics of the photoreaction of methylated angelicins. The main difference between the two families of compounds is that the site 2 T r ( c c c ) of the 139-base-pair DNA fragment is no longer one of the strongest as it is for TMA. The similarity of the sequence specificity in the photobinding of furochromones and methylated angelicins may seem surprising since these molecules have different shapes (linear vs. angular), different alfmities for DNA and form very different extents of photoadducts. Furochromones, as methylated angelicins, are less sequence specific than psoralen derivatives. Much less sites are reactive to psoralen derivatives (Figs. 5 and 6 [17, 18]). These last compounds show a preferential photoreaction in 5'-TpA, while 5'-APT sites barely react. Runs of 3 Ts are also less reactive to psoralen derivatives than to methylated angelicins and furochromones. Meanwhile the geometry of the molecules of furochromones is closer to the geometry of psoralen than to that of angelicins. Owing to the presence of oxygene at position 4 and other substituants, the intercalated furochromones may have the same overlapping and the same interaction with adjacent bases as angelicins do. A stronger stacking interaction of psoralen and adjacent bases in the intercalation complex may prevent the photoreaction with T or C at certain sequences (by lack of a sufficient overlapping of functional groups, for example). It is remarkable that at all the strong sites of photoaddition of visnagin and kheIlin, i.e. A-T-rich sites, which are also the hot spot of photoaddition of psoralen derivatives, mutations can be recovered in /ac I gene after treatment o f E . coli with angelicin or 8-MOP plus UVA [30, 32]. The hottest point of mutation for angelicin plus UVA is observed at the site TTAT/AATA which is one of the strongest sites for visnagin photoaddition in the 139base-pair DNA fragment. Although visnagin (and khellin to a lower extent) was reported to generate activated oxygen species upon irradiation [15], we did not observe any DNA photosensitization involving oxygen by these compounds. Indeed, we did not detect any photo-oxidized guanine in DNA, which would be sensitive to hot alkali and E . coli exonuclease III as was reported for 3-CPs [16]. In contrast, upon irradiation with high UVA doses, psoralen produces DNA lesions sensitive to exonuclease HI. The capacity to generate singlet oxygen by various furocoumarins correlates well with the oxygen effect observed on DNA. The quantum yield of singlet oxygen formation for 3-CPs is by far the highest, while it is lower for psoralen, and very low for 8-MOP, 5-MOP and angelicin [23, 25-27]. No photosensitized oxidation of 2'-deoxyguanosine has been observed for 8-MOP or angelicin, whereas photo-oxidation of 2'-deoxyguanosine by 5-MOP mainly involves the radical [49, 50]. The percentage of photodynamic degradation of 2'-deoxyguanosine via singlet oxygen has been

333 r e p o r t e d to be slightly smaller for visnagin than for psoralen, while it is nonexistent for khellin [15]. Indeed, the photo-oxidation of DNA by 3-CPs is quite important unde r conditions that pr od uce photoaddition, whereas it is lower for psoralen and undetectable for visnagin, khellin, 5-MOP, 8-MOP, angelicin and TMA (this work) [16, 20]. The discrepancy bet w een the results obtained with visnagin of 2'-deoxyguanosine and DNA is only apparent since G oxidation p h o t o i n d u c e d in DNA by psoralen is only detected at the highest dose of UVA irradiation (see Table 1). Furthermore, in the experiment p e r f o r m e d to detect any photodynamic effect on DNA, the conditions are those which p r o d u c e photoaddition, and which could be relevant to biological events. Another possibility is that singlet oxygen which is p r o d u c e d at different extents b y all these c o m p o u n d s may be mainly deactivated by the solvent [51]. Visnagin is a good p r o d u c e r of superoxide radical and kheUin a m oderat e one. Nevertheless, superoxide radical does not directly damage DNA. The lack of p h o to to xi c erythemal r es pons e when treating vitiligo patients with khellin m a y b e related to the absence of photodynamic effect of this compound. Our work confirms the p o o r photosensitization of DNA by visnagin and khellin. The extent of photoaddition is low c o m p a r e d with most furocoumarins, and photo-oxidation of DNA involving oxyge n is not observed. These properties of visnagin and khellin are of particular interest since this confers a low toxicity to these compounds. Furthermore, bot h c o m p o u n d s have b e e n r e p o r t e d to be as efficient as psoralen in the phototherapeutic treatment of vitiligo [4--7]. These observations suggest that the mechanism of action of f u r o c h r o m o n e s in the t r eat m e nt of vitiligo may differ from that of furocoumatins. Acknowledgments Thanks are due to Dr. Ethel Moustacchi who host ed Lorenza Trabalzini in her laboratory for this investigation. This work was s uppor t ed by grants f r om Centre National de la Recherche Scientffique, Institut National de la Sant~ et de la Recherche Mddicale (Grant 852017), Ligue Nationale Fran~aise contre le Cancer, Association p o u r la Recherche sur le Cancer (Grant 6381) and the EEC (Grant BIO-151-F). References 1 P. Martelli, L. Bovalini, S. Ferri and G. G. Franchi, Rapid separation and quantitative determination of khellin and vlsnagin in A m m i visnaga (L.)Lam. fruits by high-performance liquid chromatography, J. Chromatogr., 301 (1984) 297-302. 2 G. G. Franchi, L. Bovalini, P. Martelli, S. Ferri and E. Sbardellati, High performance liquid chromatography analysis of the furanochromones khellin and visnagin in various organs of A m m i visnaga (L.)Lam. at different developmental stages, J. Ethnopharmacol., 14 (1985) 203-212. 3 G. G. Franchi, S. Ferri, L. Bovalini and P. Martelli, A m m i visnaga (L.)Lam.: Occurrence of khellin and visnagin in primary rib channels and endosperm, and emptiness of vittae, revealed by UV microscopy, Int. J. Crude Drug Res., 25 (1987) 137-144.

334 4 A. Abdel-Fattah, M. N. Aboul-Enein, G. M. Wassel and B. S. EI-Menshawi, An approach to the treatment of vitiligo by khellin, Dermatologica, 165 (1982) 136-140. 5 H. HSnigsmann and B. Ortel, Khellin photochemotherapy of vitiligo, Photodermato/ogy, 2 (1985) 193-194. 6 P. Morli~re, H. HSnigsmann, D. Averbeck, M. Dardalhon, G. Hfippe, B. Ortel, R. Santus and L. Dubertret, Phototherapeutic, photobiologlc and photoseusitizing properties of khellin, J. Invest. Derma~L, 90 (1988) 720-724. 7 B. Ortel, A. Tanew and H. HSnigsmann, Treatment of vitiligo with khellin and UVA, J. Am. Acad~ DermatoL, 18 (1988) 693--701. 8 B. F. Abeysekera, Z. Abramowski and G. H. N. Towers, Genotoxicity of the natural furochromones, khellin and vianagin, and the identification of a khellin-tikvn~e photoadduct, Photovhe'm. PhotobioL, 38 (1983) 311-315. 9 J. B. Hudson, E. A. Graham, G. C. Chan and G. H. N. Towers, Differential effects of photoactive furanyl compounds on virus functions, Photochem. PhotobioL, 42 (1985) 523--528. 10 S. Simoni, M. Fimlani, P. Martelli, L. Bovalini, D. Vedaldi, L. Andreassi and F. Dall'Acqua, Evaluation of new photosensitizers by means of candida test, Med. BIOL Environ., 17 (1989), 823-829. 11 E. Cassuto, N. Gross, E. Bardwell and P. Howard-Flanders, Genetic effects of photoadducts and photocross-links in the DNA of phage lambda exposed to 360 nm light and tri-methylpsoralen or khellin, Biochim. Biophys. Acta, 475 (1977) 589-600. 12 N. Niccolai, L. Bovalird and P. Martelli, The mechanisms of interaction between furanochromones and DNA. A heteronuclear overhauser effect study on the khellin-thymidine model system, Biophys. Chem., 24 (1986) 217-220. 13 D. Vedaldi, S. Caffieri, F. Dall'Acqua, L. Andreassi, L. Bovaltui and P. Martelli, Khellin, a naturally occurring furochromone, used for the photochemotherapy of skin diseases: mechanism of action, Farmaco, Ed~ Sci., 43 (1988) 333--346. 14 M. Altamirano-Dimas, J. B. Hudson, Z. Abramowski and G. H. N. Towers, The photoinduction of DNA cross-links by the furanochromone visnagin, Photobioche~ Photobiophys., 10 (1985) 121-130. 15 P. Martelli, L. Bovalini, S. Ferri, G. G. Franchi and M. Barl, Active oxygen forms in photoreaction between DNA and furanochromones khellin and visnagin, FEBS Lett., 189 (1985) 255-257. 16 G. Miolo, F. Dall'Acqna, E. Moustacchi and E. Sage, Monofunctional angular furocoumarins: sequence specificity in DNA photobinding of 6,4,4'-trimethylangelicin and other angelicins, Photochem. PhotobioL, 50 (1989) 75-84. 17 E. Sage and E. Moustacchi, Sequence context effects on 8-methoxypsoralen photobinding to defined DNA fragments, Biochemistry, 25 (1987) 3307--3314. 18 V. Boyer, E. Moustacchi and E. Sage, Sequence specificity in photoreaction of various psoralen derivatives with DNA= Role In biological activity, Biochemistry, 27 (1988) 3011-3018. 19 G. Miolo and E. Sage, unpublished data, 1988. 20 E. Sage, T. Le Doan, V. Boyer, D. E. Helland, L. Kittler, C. H~l~ne and E. Moustacchi, Oxidative DNA damage photo-induced by 3-carbethoxypsoralen and other furocoumartns: mechanisms of photooxidation and recognition by repair enzymes, J. MoL BioL, 209 (1989) 297-314. 21 J. Cadet, C. Decarroz, S. Y. Wang and W. R. Midden, M e c ~ and products of photosensitized degradation of nucleic acids and related model compounds, Isr. J. ~ , 23 (1983) 420--429. 22 J. Cadet, C. Deccaroz, L. Voituriez, F. Gaboriau and P. Vigny, Sensitized photoreactlons of purine and pyrimidine 2'-deoxyribonucleosides by 8-methoxypsoralen and 3-carbethoxypsoralen, in W. Bors, M. Saran and D. Talt (eds.) Oxygen Radicals in Chemistry and Bio/ogy, Walter de Gruyter, Berlin, 1984, pp. 485--490. 23 D. Vedaldi, F. Dall'Acqua, A. Gennaro and G. Rodighiero, Photosensitized effects of furocoumarins: the possible role of singlet oxygen, Z. Naturforsch., 38c (1983) 866-869.

335 24 M. A. Pathak and P. C. Joshi, Production of active oxygen species (~02 and 02-) by psoralens and ultraviolet radiation (320-400 rim), Biochim. Biophys. Acta, 798 (1984) 115--126. 25 D. Averbeck, L. Dubertret, M. Craw, T. G . Truscott, F. Dall'Acqua, P. Rodighlero, D. Vedaldi and E. J. Land, Photophysical, photochemical and photobiological studies of 4'methylangelicins, potential agents for photochemotherapy, Farmaco, Ed. Sci., 39 (1984) 57-69. 26 D. Averbeck, G. Innocenti, D. Vedaldi and F. Dali'Acqua, Photobiological effects of methylpsoralens in haploid yeast: the role of oxygen, Med. Biol. Environ., 14 (1986) 17-25. 27 C. N. Knox, E. J. I.and and T. G. Trnscott, Singiet oxygen generation by furocoumarin triplet states. I: Linear furocoumarins (psoralens), Photoche~ Photobiol., 43 (1986) 359-363. 28 L. Trabalzini and E. Sage, unpublished data, 1989. 29 J. H. Miller, Carcinogens induce targeted mutations in Escherichia coil, Cell, 31 (1982) 5-7. 30 S. S. Miller and E. Eisenstadt, Analysis of supressible base substitution mutations induced by angelicin (isopsoralen) in the /ac I gene of Escherichia coli: implications for the mechanism of SOS-mutagenesis, J. Bacteriol., 169 (1987) 2724-2729. 31 F. Yatagai and B. W. Glickman, Mutagenesis by 8-methoxypsoralen plus near-UV treatment: analysis of specificity in the /ac I gene of Escherichia coli, Murat. Res., 153 (1986) 209-224. 32 F. Yatagai, M. J. Horsfall and B. W. Glickman, Defect in excision repair alters the mutational specificity of PUVA treatment in the /ac I gene of Esche'richia coli~ J. Mol. Biol., 194 (1987) 601-607. 33 A. J. E. Gordon, P. A. Burns, D. F. Fix, F. Yatagal, F. L. Allen, M. J. Horsfall, J. A. Halliday, J. Gray, C. Bernelot-Moens and B. W. Glickman, Missense mutation in the /ac I gene of Escherichia coli: interferences on the structure of the repressor protein, J. Mol. Biol., 200 (1988) 239--251. 34 E. Sage, V. Boycr and E. Moustacchi, Sequence specificity in psoralen-DNA photobinding, Biochem. Pharmacol~, 37 (1988) 1829-1830. 35 R. P. P. Fuchs, DNA binding spectrum of the carcinogen N-acetoxy-N-2-acetylaminofluorene significantly differs from the mutation spectrum, J. Mol. Biol., 177 (1984) 173--180. 36 P. W. Doetsch, G. L. Chart and W. A. Haseltine, T4 DNA polymerase ( 3 ' - 5 ' ) exonuclease, an enzyme for the detection and quantitation of stable DNA lesions: the ultraviolet light example, Nucleic Acids Res., 13 (1985) 3285-3304. 37 W. P. Zhen, O. Buchardt, H. Nielsen and P. E. Nielsen, Site specificity of psoralen-DNA intel~h-,tnd cross-linking determined by nuclease Bal 31 digestion, Biochemistry, 25 (1986) 6598-6603. 38 E. A. Ostrander, R. A. Karty and L. M. Hallick, High resolution psoralen mapping reveals an altered DNA helical structure in the SV40 regulatory region, Nucleic Acids Res., 16 (1988) 213--227. 39 T. J. Kochel and R. R. Sinden, Hyperreactivity of B--Z junctions to 4,5',8-trtmethylpsoralen photobinding assayed by an exonuclease HI. Photoreversal mapping procedure, J. Mol. Biol., 205 (1989) 91--102. 40 B. Royer-Pokora, L. K. Gordon and W. A. Haseltine, Use of exonuclease HI to determine the site of stable lesions in defined sequences of DNA: the cyclobutane pyrimidine dimer and cis and trans dichlorodiammine platinum H examples, Nuc/e/c Ac/ds Res., 9 (1981) 4595--4609. 41 P. W. Doetsch, personal communication, 1988. 42 E. Sage, unpublished, 1986. 43 C. Gardin and E. Sage, unpublished, 1985. 44 A. Sancar, IC A. Franklin, G. Sancar and M. S. Tang, Repair of psoralen and acetylaminofluorene DNA adducts by ABC excinuclease, J. Mol. BioL, 184 (1985) 725--734. 45 A. T. Yeung, B. K. Jones, M. Capraro and T. Chu, The repair of psoralen monoadducts by the Escherichia coli Uvr ABC endonuclease, Nucleic Acids Res., 15 (1987) 49574971.

336 46 P. C. Beaumont, E. J. Land, S. Nataratnam, B. J. Parsons and G. O. Phillips, A pulse radiolysis study of the complexing of furocoumarins with DNA and proteins, Biochi~n~ Biophys. Acta, 608 (1980) 182-189. 47 D. Averbeck and S. Averbeck, Genotoxic effects of mono- and bifunctional furocounmrins in yeast involvement of DNA Photoadditions and oxygen dependent reactions, in R. V. Bensasson, G. Jori, E. J. Land and T. G. Truscott (eds.), Primary Photo.~rrocesses in Biology and Medicine, Plenum, New York, 1985, pp. 295-300. 48 H. Gamper, J. Piette and J. E. Hearst, Efficient formation of a cross-linkable HMT monoadduct at the Kpn I recognition site, Photochem. PhotobioL, 40 (1984) 29-34. 49 J. Cadet, M. Berger, C. Decarroz, J. R. Wagner, J. E. Van Lier, Y. M. Ginot and P. Vlgny, Photosensitized reactions of DNA and related compounds, Biochimie, 58 (1986) 813-834. 50 J. Cadet, personal communication, 1988. 51 P. C. C. Lee and M. A. J. Rodgers, Laser flash photokinetic studies of rose bengal sensitized photodynamic interactions of nucleotides and DNA, Photoc/urm. PhotobioL, 45 (1987) 79-86.