Chem.-Bwl Interactwns, 85 (1992) 265-281
265
Elsewer Scientific Publishers Ireland Ltd
PHOTOCHEMICAL DNA MODIFICATIONS INDUCED BY 1,2-DIOXETANES
BERND EPE a, ELVIRA MULLER a, WALDEMAR ADAM b and CHANTU R SAHA-MOLLERb
alnst~tute of Pharmacology and Toxwology and blnst~tute of Organw Chemistry, Unwerszty of Wurzburg, D-8700 Wurzburg (Germany) (Received June 16th, 1992) (Rewmon received October 15th, 1992) (Accepted October 16th, 1992)
SUMMARY
1,2-Dioxetanes are efficient sources of triplet excited carbonyl compounds, into which they decompose on thermal or photochemical activation. In the presence of DNA, the decomposition of dioxetanes gives rise to DNA modifications, which have been studied by means of specific repair endonucleases. Cyclobutane pyrimidine dimers, which are generated by triplet-triplet energy transfer, were detected by a UV endonuclease; they made up between 2% and 30% of the total modifications recognized by a crude repair endonuclease preparation from Micrococcus luteus. For various 1,2-dioxetanes, the yield of pyrimidine dimers was proportional to their triplet excitation flux. DNA strand breaks, sites of base loss (AP sites; recognized by exonuclease III and endonuclease IV) and dihydropyrimidines (recognized by endonuclease III) were found to represent only a small fraction of the modifications. The majority of the modifications detected were recognized by formamidopyrimidine-DNA glycosylase (FPG protein) and represent 8-hydroxyguanine (7,8-dihydro-8-oxoguanine) residues or other yet not defined base modifications which are recognized by this enzyme. The modifications were generated in similar relative yields by thermal and photo-induced decomposition of the 1,2-dloxetanes and therefore emanate under both conditions from the excited carbonyl compounds. The formatmn of the FPG protein-sensitive modificatmns was efficiently quenched by azide anions; the Stern-Volmer quenching of these modifications was 150-fold more effective than that of the pyrimidine dimers. The relative amounts of the two types of modifications were strongly dependent on the structure of the 1,2dioxetanes and on the concentration of molecular oxygen. Singlet oxygen apCorrespondence to B Epe, Institute of Pharmacology and Toxicology, Umvermty of Wfirzburg, Versbaeher Str 9, D-8700 Wfirzburg, Germany
Abbrewatwns. AcTMD, acetoxymethyl-tmmethyl-l,2-dloxetane, HTMD, hydroxymethyltmmethyl1,2-dloxetane, NDP02 dmodlum salt of 1,4-etheno-2,3-benzodloxm-l,4-dlpropanom acid, PM2 DNA, bactemophage PM2 DNA, TMD, tetramethyl-l,2-dloxetane, TrMD, tmmethyl-l,2-dloxetane
0009-2797/92/$05 00 © 1992 Elsewer Scmntffic Pubhshers Ireland Ltd Pmnted and Published in Ireland
266
pears to be involved only to some extent in the generation of the FPG proteinsensitive base modifications as their yield was only moderately ( - 2-fold) mcreased in D20 as solvent. A mechamsm is suggested in whmh oxidized guanine is predominantly formed by a single-electron-transfer reactmn of the triplet excited carbonyl product derived from the 1,2-dmxetane, followed by unknown secondary oxidations, which involve molecular oxygen and/or undecomposed 1,2-dioxetane.
Key words: DNA damage -- 1,2-Dioxetanes -- Excited states -- Triplet ketones -- Singlet oxygen -- Photosensitization -- Pyrimidine dlmers -- Energy transfer -- Repair endonucleases -- Formam]dopyrimidme-DNA glycosylase -8-Hydroxyguanme -- Azide
INTRODUCTION
1,2-Dloxetanes are four-membered ring peroxides whmh decompose spontaneously (thermal activation) or under the influence of near-UV irradiation (photo-activation) into two carbonyl fragments, one of which is m an excited state [1] (Fig. 1). Such excited carbonyl compounds are of biologqcal interest since they may also be generated in cells under the influence of near-UV irradiation, by decomposition of lipid peroxides and in enzymatic oxidations [2-4]. Similar to other reactive oxygen species such as singlet oxygen (102) and oxygen radmals (superoxide, hydroxyl radicals) excited carbonyl compounds have been proposed to be involved in 'oxidative stress' [5]. The triplet energy of ketones such as acetone and acetophenone is h]gh enough to allow a triplet-triplet energy transfer to DNA, whmh results in DNA modifications identical with those generated by UV irradiation at 260 nm, in particular cyclobutane-type pyrimidine d]mers [6-8]. However, other types of DNA modifications can also be formed either directly by electron transfer reactions or cyclo-addition reactions of the tr]plet excited ketone (type I reactions) or indirectly, e.g. by an energy transfer to molecular oxygen, which results m the formation of highly reactive singlet oxygen (type II reaction) [9-11]. The
o-l' ,s 0--0
HaC'~R H3C CH3
A or near-UV
H3C"~CH3 (R)
+ O
R
H CH3 CH2OH CH2OAc
TrMD TMD HTMD AcTMD
HsC'J~ R (CH3) Fig. 1 1,2-I:hoxetanes used m this study and their fragmentation into two carbonyl products, one of which ]s in an electronically excited state
267
generation of triplet excited states from 1,2-dioxetanes should offer an opportunity to study their interactions with DNA. Incubation of isolated DNA with trimethyl-l,2-dioxetane (TrMD) has been shown to generate thymine dimers [12,13]. However, the spectrum of DNA modifications induced was later demonstrated to be not UV-like, i.e. unknown modifications other than pyrimidine dlmers prevail [14]. In cultured Syrian hamster embryo cells, hydroxymethyltrimethyl-l,2-dioxetane (HTMD) was shown to induce single strand breaks and cause the formation of mmronuclei and cell transformation, but no pyrimidine dimers could be detected [15]. HTMD induced SOS functions in E. coli strains [16]; however, this and several other 1,2dioxetanes proved to be non-mutagenic in Salmonella typh,mumum [17]. Photobinding of psoralen to DNA could be induced by thermal decomposition of HTMD [17,18]. Here, we describe an analysis of the spectrum of DNA modifications induced by TrMD, HTMD, tetramethyl-l,2-dioxetane (TMD) and acetoxymethyl-trimethyl-l,2-dioxetane (AcTMD) (Fig. 1) by means of specific repair endonucleases. MATERIALS AND METHODS
Chemicals The dioxetanes TrMD, TMD, HTMD and AcTMD were synthesized according to published procedures [19- 21]. The purity of the preparations was more than 95% in all cases, as determined by iodometric titration and by 1H-NMR spectroscopy with an internal standard. NDP02 (disodium salt of 1,4-etheno-2,3benzodioxin-l,4-dipropanom acid) was prepared as described by Nieuwmt et al. [22]. DNA from bacteriophage PM2 (PM2 DNA) was prepared according to Salditt et al. [23]. Ninety-five percent of the preparation consisted of the supercoiled form, as determined by the method described below. Enzymes Formamidopyrimidine-DNA glycosylase (FPG protein) was a gift from Dr. S. Boiteux, Villejuif. At the enzyme concentration applied, the incision of modffications induced in PM2 DNA by singlet oxygen was shown to be saturated [24]. Endonuclease III was provided by Dr. R.P. Cunningham, Albany. At the enzyme concentration applied, the incision of modifications reduced in PM2 DNA by osmium tetroxide (thymine glycols) [24] was shown to be saturated. Endonuclease IV was a gift from Dr. B. Demple, Boston. A crude cell extract from Micrococcus luteus (ATCC 4698) was obtained as described by Riazzudin [25]. A partially purified preparation of the UV endonuclease from M. luteus was obtained after streptomycin sulfate and ammonium sulfate precipitations by diethylaminoethyl (DEAE)-cellulose chromatography. This endonuclease was shown to fully recognize UV26°-generated pyrimidine dimers and H +-generated sites of base loss (AP sites) under the incubation conditions but have no activity against DNA modified by osmium tetroxide [24]. Exonuclease III was obtained from Boehringer Mannheim, FRG.
268
DNA mod,ficatwn Exposure of supercoiled PM2 DNA (10 #g/ml) to various 1,2-dioxetanes (up to 40 mM) was carried out in phosphate buffer (5 mM KH2POa, 50 mM NaC1, pH 7.4), which contained 20% acetone in the experiments shown in Fig. i and Table II. The modification was carried out either in the dark under thermal decomposition (1 h, 37°C or 55°C) or under UV-irradiation (360 nm, 5600 J/m 2, Osram HQV black light lamp) at 0°C for 4 rain (photo-induced decomposition). In some experiments H20 as solvent was replaced by D20; the final isotope purity was higher than 95%. In another set of experiments with photo-induced decomposition, argon or pure oxygen was bubbled through the solution by means of a glass capillary. Exposure of PM2 DNA to the endoperoxide NDPO2 (3.5 mM, 2 h in D20 buffer), to xanthine/xanthine oxidase/Fe(III)-EDTA (8 ~M xanthme) and to UV (254 nm) radiation was carried out as described earlier [14,24]. In all experiments, the DNA was precipitated by ethanol/sodium acetate and redissolved m BErbuffer (20 mM Tris-HC1 pH 7.5, 100 mM NaC1, 1 mM EDTA) for damage analysis.
DNA damage analys,s The assay makes use of the fact that supercoiled DNA is converted by either a single strand break (ssb) or the incision of a repair endonuclease into a relaxed (nicked) form which migrates separately from the supercoiled form in agarose gel electrophoresis. Quantification of both forms of DNA by fluorescence scanning allows the determination of the number of single strand breaks per PM2 molecule (10 t base pairs). If an incubation with repair endonucleases precedes the gel electrophoresis the number of single strand breaks plus endonucleasesensitive sites (ess) is obtmned from Eqn. 1 (see Ref. 14). ssb + ess = -ln[1.4"I/(1.4"I + II)] I: Fluorescence of supercoiled form II: Fluorescence of relaxed form
(1)
An aliquot of 0.3 t~g of the modified PM2 DNA in 20 td BE1 buffer was incubated for 30 min at 37°C with 10 t~l of one of the following repair endonuclease preparations: (i) exonuclease III, 300 U/ml in 20 mM Tris- HC1 (pH 8.0), 100 mM NaC1, 15 mM CaC12; (ii) a UV endonuclease preparation from M. luteus, 150 t~g/ml protein in BE15 buffer (BE15 buffer containing 15 mM EDTA); (iii) a crude extract from M. luteus, 0.3 mg/ml in BE15 buffer; (iv) FPG protein, 3 t~g/ml in BE15 buffer; (v) endonuclease III, 40 ng/ml m BE15 buffer and (vi) endonuclease IV, 10 U/ml in BEI~ buffer. The reactions were stopped by addition of 3 ul 10% sodium dodecyl sulfate and the DNA separated by agarose gel electrophoresis. After staining with ethidium bromide, the relative amounts of the supercoiled and the nicked form of PM2 DNA were determined by means of a fluorescence scanner (FTR20, Sigma Instruments, Berlin).
Activation parameters and quantum y~elds The chemiluminescence of 9,10-dibromoanthracene-2-sulfonate was used to
269
quantify the triplet states generated by thermal decomposition of 1,2-dioxetanes in water. Quantum yields were calculated from Stern-Volmer plots. The decomposition kinetics at various temperatures was followed as described earlier [26] for the determination of the rate constants and activation parameters. From the decomposition rate constants (Kd), the 1,2-dioxetane concentration (c) and triplet quantum yields (¢W) the triplet excitation flux (Ep T) was calculated according to Eqn. 2: Ep T = Kd- c . CT. NL
(2)
where NL is the Avogadro number. The triplet excitation flux represents the number of triplet excited states generated from the dioxetane per unit time per unit volume.
TrMD
HTMD
1.2
1,2
0,8
0,8
O~ 0.4
0,4
D. J~
&
0.0 YQ"
0,0
2,0
4,0
6.0
0.0
0.0
w e-
3
(g
2
o
to
.
.
.
.
.
1.0
.
.
.
.
2,0
AcTMD
"ID o
E < Z ¢:Z
20
40
0.0
1,0
2,0
dioxetane concentration [mM] Fig 2 DNA modifications reduced m PM2 DNA by incubation with 1,2-dloxetanes DNA single strand breaks (O) and sum of single strand breaks plus 0) rotes senmt]ve to a crude endonuclease preparatmn from M luteus (e), 00 the UV endonuclease from M luteus (4) and 011) exonuclease III from E cob (A)
270
RESULTS
DNA mod~ficatwns ~nduced by 1,2-dwxetanes are predominantly base modificatwns Supercoiled DNA from bacteriophage PM2 (104 base pairs) was exposed to various concentrations of 1,2-dioxetanes in phosphate buffer at 37°C and 55°C Subsequently, the DNA was analysed for the following types of modifications: (1) endonuclease-sensltive sites recogmzed by a crude protem preparation from M. luteus [25]; (fi) sites of base loss (AP sites) recogmzed by exonuclease III from E. coh [27]; (in) pyrimldme d:mers recognized, m addition to AP sites, by a specific UV endonuclease preparation from M luteus [25] and 0v) DNA single strand breaks. As shown in Fig. 2, all types of DNA modifications increased linearly with the 1,2-dioxetane concentration. From the best-fit slopes of these dose-dependent data, the numbers of modifications mduced per concentration umt were calculated (Table I). The data indmate that for all dioxetanes tested single strand breaks and AP sites represent only a minor fractmn of the total modificatmns
TABLE I SINGLE STRAND BREAKS AND ENDONUCLEASE-SENSITIVE SITES INDUCED BY VARIOUS 1,2-DIOXETANES IN PM2 DNA 1,2Dmxetane
Reactmn cond:tmns
Modificationsa
Time (mm)
Temp (°C)
M luteus b
TrMD HTMD TMD AcTMD
60 60 60 60
37 37 37 37
18 6 + 0 9 f 52+5 64+03 72+5
64 + 05 6 0 + 10 20+02 4 0 + 10
12 + 03 50+06 03+004 4 0 + 10
TrMD HTMD TMD AcTMD
60 60 60 60
55 55 55 55
630 + 780 + 190+ 400 +
114 + 4 101 + 4 24+2 78 + 4
34 100 90 73
TrMD
4 (dark) 4 ( + UV 36o)
02+03
TrMD
43 80 18 20
UV endo c
0
05+04
01+02
0
55+8
305+ 1
Exo III d
+ + + +
ssb e
3 4 1 4
10 + 0 2
0 40 + 0 15 040+02 010+001 090+04 7 + 09 13 + 1 20+03 19 + 2 (-03)+03 100+02
aNumber of modlficat:ons (+ S D ) per 106 base pmrs and mM calculated by hnear regression analysm from data points shown in Fig 2 bModfficat:ons detected by a crude repmr endonuclease preparation from M luteus CModlficatlons detected by a UV endonuclease preparation from M luteus dAp rotes detected by exonuclease III eSmgle strand breaks fA lower value reported earher [14] might have been due to metal 1on lmpumtms
271 TABLE II ACTIVATION AND EXCITATION PARAMETERS OF 1,2-DIOXETANES IN WATER AT 37°C a 1,2-Dloxetane
AG#b (kcal/mol)
tl/2 c (h)
oT d (%)
S T . 10-17 e (sp-1 1- 1)
TrMD TMD HTMD AcTMD
247±10 256±07 250±40 2 6 1 ± 13
65± 03 313± 09 3 8 7 ± 111 589± 30
5±1 4±1 5±1 10±1
9±2 4±1 2± 1 2± 1
apart of the data are taken from Ref 17. bFree enthalpms of activation CHalf-hfe dTrlplet excltlon yield eTrlplet excitation flux for 1 M dloxetane
generated. In the case of TrMD and TMD pyrimidine dimers (the number of which is obtained as the difference between sites sensitive to the UV endonuclease and exonuclease III) account at 37°C for approx. 27% and at 55°C for approx. 10% of the modifications recognized by the crude endonuclease preparation from M. luteus. For HTMD and AcTMD, however, the fraction of pyrimidine dimers is less than 2%. Unknown endonuclease-sensitive base modifications other than pyrimidine dimers prevail in all cases.
The number of pymmid,ne d~mers, but not of other base mod~ficatwns, ,s proportwnal to the triplet excitation flux To quantify the number of excited triplet states generated from 1,2-dioxetanes during incubation in aqueous solution, the activation and exotation parameters of TMD, TrMD, HTMD and AcTMD were determined (Table II). The half-hves and activation energies in water were found to be similar to the values reported earlier for toluene as solvent [21] while the triplet quantum yields turned out to be 3-10-fold lower. From the half-lives and quantum yields an excitation flux was calculated, which gives the number of triplet states generated per liter and second from a 1 M 1,2-dioxetane solution (Table II). The numbers of those DNA modifications which originate directly from the excited triplet carbonyl compounds are expected to be proportional to the corresponding excitation fluxes. Indeed, as shown in Fig. 3, a fair linear correlation is observed for pyrimidine dimers, but not for the other DNA base modifications detected by endonucleases from M. luteus. It has to be concluded that the latter modifications either do not originate from the excited triplet states at all or that their formation involves other (indirect) reaction mechanisms whmh are dependent on the particular structure of the dioxetanes. In the first case, a direct dark reaction between DNA and undecomposed dioxetanes might be considered as
272
Q;
0
~s
N
~'~
~o
~o I
~Z
0
..~
~-~
0
~o
o
&
~
o
o °
,
~
%'-
~
,.~
W
Q;
~o~ N
~j
e~
"Se=
INtU / dq 90L Jed SUOl;eOl;Ipotu VNO
~
273
most likely, since the decomposition products in their ground states, e.g. acetone in the case of TMD, are inert towards DNA under the conditions used (data not shown). The experiments described below were conducted to test this posslbihty.
The DNA modzfications ~nduced by 1,2-dwxetanes omg~nate from excited carbonyl compounds In addition to the thermal (spontaneous) decomposition, the cleavage of 1,2dioxetanes into excited carbonyl products can also be triggered by irradiation at wavelengths between 300 - 400 nm, at which 1,2-dloxetanes have a weak absorption [28]. This photo-induced decomposition can be carried out at low temperatures (0°C) at which dioxetanes do not decompose signfficantly m the dark Irradiation of TrMD on ice at 360 nm in the presence of PM2 DNA g~ves rise to much more DNA modifications than in the dark at the same temperature (Table I). As in the case of thermal decomposition at 37°C or 55°C, base modifications other than pynmldine dimers prevail. The DNA modificatmns mduced by irradiation at 0°C must originate d~rectly or indirectly from the excited carbonyl compounds generated by the fragmentatmn of the 1,2-dioxetanes, since (i) the lifetime of electromcally excited (undecomposed) 1,2-dioxetane is too short for a reaction with DNA [28] and (11)no or very few DNA modificatmns are formed under these conditions in the dark. Interestingly, independent of the method of 1,2-dioxetane decomposition, the fraction of pynmldme dlmers decreases with temperature. The photo-induced reaction generates a much higher percentage of singlet states compared to the thermal decomposition of 1,2-dmxetanes, which ymlds nearly exclusively excited triplet states [28]. Therefore, the result also indicates that the singlet or triplet nature of the excited states generated is not relevant for the DNA damage observed The endonuclease-sens,twe s, tes induced by 1,2-dwxetanes are largely base mod,ficatwns sensitive to FPG protein In order to obtain further information about those dioxetane-induced base modifications which are not pyrimldme dimers, two specffic repair endonucleases from E. coh were used to characterme the modifications, viz, endonuclease III and formamldopynmldme-DNA glycosylase (FPG protein). In addition, another AP endonuclease from E. cob, endonuclease IV, was mcluded in the assays. Endonuclease III has been shown to recognize, in addition to AP sites, various 5,6-dihydropyrimidine residues m DNA [29,30], which are characteristm base modifications induced by hydroxyl radmals [31,32]. FPG protein was originally described as a repair enzyme that recognizes formamldopyrimldines (ringopened purines) [33-35] and was later demonstrated to recognize 8-hydroxyguanme (7,8-dihydro-8-oxoguanme) as well [36,37]. Endonuclease IV probably differs from exonuclease III in the recogmtlon of AP s~tes oxidized m the sugar moiety [38]. The number of modifications recognized by the E coh endonucleases are shown m Fig. 4 for TrMD and HTMD in the form of DNA damage profiles Only a relatively small number of modifications is sensitive to endonuclease III or
274 ~. 3.0 .Q o T~ 1,--
m 2.0 (n ro em o
/ /
1.0
q~
"o o
E Z
0,0
L
TrMD
HTMD
endonuclease-sensltive
[] [] []
FPG protein endonuclease III UV endonuclease
NDPO 2
modifications
[ ] endonuclease IV [ ] exonuclease III
Xanthlns / UV254 X.oxidase/Fe III DNA strand breaks
•
single strand breaks
Fzg 4 DNA damage profiles single strand breaks and various endonucleasesensitive modifications observed in PM2 DNA after treatment with 0) TrMD (6 raM, 1 h, 37°C), (u) HTMD(2 5 mM, 1 h, 37°C), (111)NDPO2 m D20 buffer (3 5 mM, 2 h, 37°C), (Iv) xanthme (8/zM) m the presence of xanthme oxldase and Fe(III)-EDTA (30 mm, 20°C) and (v) UV (254 nm, 8 7 J/m 2, 0°C)
endonuclease IV. Therefore, dihydropyrim]dines and oxidlsed AP s]tes are minor lesions. The relative number of sites recognized by FPG protein, however, is high. The sum of FPG-sensitive sites and UV endonuclease-sensitwe sites is similar to the number of modifications recogmzed by the crude preparation from M. luteus. Therefore, most probably an endonuclease of M luteus analogous to FPG protein is responsible for the activity of the crude M. luteus protein preparation. The damage profiles induced by 1,2-dioxetanes can be compared with those induced by other agents for which the DNA damaging mechanism is better known (Fig. 4). DNA damage profiles induced by UV irradiation at 254 nm (direct excitation of DNA) and by xanthine m the presence of xanthine oxidase and Fe(III)EDTA (generation of hydroxyl radicals, see Ref. 14) are very different from those generated by the 1,2-dioxetanes. In the former case, UV endonucleasesensitive sites (pyrimidme dimers) are generated nearly exclusively, sites sensitive to the other enzymes each make up less than 1% of the total modifications detected. In the case of hydroxyl radicals, the damage profile is characterized by relatively high numbers of AP sites and single strand breaks. The damage profile induced by the endoperoxide NDP02, a chemmal source of singlet oxygen [39,40], however, resembles that of HTMD, which does not generate detectable
275 TABLE III EFFECT OF THE REPLACEMENT OF H20 AS SOLVENT BY D20 ON THE DNA DAMAGE INDUCED BY 1,2-DIOXETANES Mod~ficatmns senmtlve to
FPG protein UV endonuclease
Relative number of modifications (%)a,b HTMD c
TrMD c
199 ± 48 (8) nd
177 ± 25 (6) 113 ± 19 (8)
aObserved m D20 buffer compared to H20 buffer (defined as 100%) bValues represent mean ± S D , the number of determinations is given m parentheses CIncubatlons of the 1,2-dloxetanes with PM2 DNA were for 1 h at 37°C
pyrimldine dimers. Therefore, experiments were carried out to test the role of singlet oxygen in the generation of DNA damage by 1,2-dioxetanes.
S~nglet oxygen is only part~aUy involved in DNA damage by 1,2-dwxetanes The azide amon is known as an efficient, although not specific, quencher of singlet oxygen [41]. The influence of azide ions on the generation of DNA modifications by TrMD is shown in Fig. 5. Both the formation of UV endonuclease senmtive sites (mostly pynmidine dimers) and of the modifications sensitive to FPG protein is suppressed by azide anions, however with very different effectivity. The Stern-Volmer equation for dynamic quenching affords quenching constants of 64 ± 5 mM- 1 and 0.40 ± 0.02 mM- 1 for the FPG protein sensitive modifications and UV endonuclease-sensitive modifications, respectively. The life-time of singlet oxygen is approx. 10-fold greater in D20 than in H20 [42,43]. Therefore, DNA damage mediated by singlet oxygen is expected to increase in D20. In Table III, the effect of the solvent change on the generation of pyrimidine dimers and modifications sensitive to FPG protein is shown for TrMD and HTMD. Whereas the generation of pyrimidine dlmers is not significantly affected, that of FPG-sensitive sites is approximately 2-fold higher in D20. Control experiments indicate that 1,2-dioxetanes do not react significantly with azide ions under the conditions used (data not shown). To test for the participation of oxygen in the process of damage formation, the decomposition of TrMD and HTMD in the presence of DNA was carried out under argon gas, air and pure oxygen gas atmospheres. For practical reasons, the decomposition of the 1,2-dioxetanes in these experiments was induced by near-UV irradiation at 0°C (photo-activation). The results (Table IV) indicate that the generation of FPG-sensitive modffications is not significantly affected by the presence or absence of molecular oxygen. The generation of pyrimidine dimers, however, is lower in air than in argon gas and even lower in the presence of pure oxygen gas.
276 T A B L E IV E F F E C T OF M O L E C U L A R DIOXETANES Modifications sensitive to
OXYGEN
ON
Relative n u m b e r of modifications HTMD c
F P G protein UV endonuclease
THE
DNA
DAMAGE
INDUCED
BY
1,2-
(%)a,b TrMD c
air
oxygen
air
oxygen
97 :e 20 (6) n d
94 ± 18 (5) n d
90 =e 26 (6) 70 + 16 (7)
71 + 14 (6) 33 + 7 (5)
aObserved in air and pure o x y g e n c o m p a r e d to a r g o n (defined as 100%) bValues r e p r e s e n t m e a n + S D , t h e n u m b e r of d e t e r m i n a t i o n s is given m p a r e n t h e s e s CThe 1,2-dloxetanes were Irradiated m t h e presence of PM2 D N A with U V (360 nm) for 4 m m at 0°C
DISCUSSION
The results mdicate that 1,2-dioxetanes modify DNA both under conditions of thermal decomposition at 37°C or 55°C and Irradiation with near-UV (360 nm) at 0°C. At least under the latter conditions the DNA damage is initiated by 1,2dioxetane decomposition into excited carbonyl compounds and cannot be attributed to a direct oxidation or alkylation of the DNA by the 1,2-dioxetanes, whmh potentially can act both as one-electron acceptors [44] and as electrophiles [45]. The spectrum of DNA modifications is apparently the same under conditions of thermal and photo-induced decomposition (Table I). It consists predominantly of DNA base modifications. Only a minor part of the base modifications are pyrimidine dimers. These are well-known to be generated when acetone or acetophenone is Irradiated with near-UV in the presence of DNA [6-8] and result from energy transfer to DNA by the triplet excited carbonyl products. The generation of pyrimidine dimers is reduced in the presence of molecular oxygen (Table IV), probably because oxygen acts as a quencher for the excited carbonyl intermediates. Similarly, also the azide amon apparently can act as quencher of the excited carbonyl products, but only at relatively high concentrations (Fig. 5). A major fraction of the base modifications is recognized by FPG protein (Fig. 4). The only substrates of this repair endonuclease known so far are ringopened purmes (formamidopynmidines) and 8-hydroxyguanosme [33-37]; however, the recognition of other (guanine) modifications cannot be excluded. 8-Hydroxypurines and formamldopyrimidines are known to be generated by hydroxyl radicals by a mechanism that is fairly well understood [46]. However, in the case of hydroxyl radicals as oxidants the generation of the base modifica-
277 10 A 01 C
FPG prote~n UV endonuclease
100" oo~
O ,m
= o
",r-
80"
s
"O
_o 60'
265
Elsewer Scientific Publishers Ireland Ltd
PHOTOCHEMICAL DNA MODIFICATIONS INDUCED BY 1,2-DIOXETANES
BERND EPE a, ELVIRA MULLER a, WALDEMAR ADAM b and CHANTU R SAHA-MOLLERb
alnst~tute of Pharmacology and Toxwology and blnst~tute of Organw Chemistry, Unwerszty of Wurzburg, D-8700 Wurzburg (Germany) (Received June 16th, 1992) (Rewmon received October 15th, 1992) (Accepted October 16th, 1992)
SUMMARY
1,2-Dioxetanes are efficient sources of triplet excited carbonyl compounds, into which they decompose on thermal or photochemical activation. In the presence of DNA, the decomposition of dioxetanes gives rise to DNA modifications, which have been studied by means of specific repair endonucleases. Cyclobutane pyrimidine dimers, which are generated by triplet-triplet energy transfer, were detected by a UV endonuclease; they made up between 2% and 30% of the total modifications recognized by a crude repair endonuclease preparation from Micrococcus luteus. For various 1,2-dioxetanes, the yield of pyrimidine dimers was proportional to their triplet excitation flux. DNA strand breaks, sites of base loss (AP sites; recognized by exonuclease III and endonuclease IV) and dihydropyrimidines (recognized by endonuclease III) were found to represent only a small fraction of the modifications. The majority of the modifications detected were recognized by formamidopyrimidine-DNA glycosylase (FPG protein) and represent 8-hydroxyguanine (7,8-dihydro-8-oxoguanine) residues or other yet not defined base modifications which are recognized by this enzyme. The modifications were generated in similar relative yields by thermal and photo-induced decomposition of the 1,2-dloxetanes and therefore emanate under both conditions from the excited carbonyl compounds. The formatmn of the FPG protein-sensitive modificatmns was efficiently quenched by azide anions; the Stern-Volmer quenching of these modifications was 150-fold more effective than that of the pyrimidine dimers. The relative amounts of the two types of modifications were strongly dependent on the structure of the 1,2dioxetanes and on the concentration of molecular oxygen. Singlet oxygen apCorrespondence to B Epe, Institute of Pharmacology and Toxicology, Umvermty of Wfirzburg, Versbaeher Str 9, D-8700 Wfirzburg, Germany
Abbrewatwns. AcTMD, acetoxymethyl-tmmethyl-l,2-dloxetane, HTMD, hydroxymethyltmmethyl1,2-dloxetane, NDP02 dmodlum salt of 1,4-etheno-2,3-benzodloxm-l,4-dlpropanom acid, PM2 DNA, bactemophage PM2 DNA, TMD, tetramethyl-l,2-dloxetane, TrMD, tmmethyl-l,2-dloxetane
0009-2797/92/$05 00 © 1992 Elsewer Scmntffic Pubhshers Ireland Ltd Pmnted and Published in Ireland
266
pears to be involved only to some extent in the generation of the FPG proteinsensitive base modifications as their yield was only moderately ( - 2-fold) mcreased in D20 as solvent. A mechamsm is suggested in whmh oxidized guanine is predominantly formed by a single-electron-transfer reactmn of the triplet excited carbonyl product derived from the 1,2-dmxetane, followed by unknown secondary oxidations, which involve molecular oxygen and/or undecomposed 1,2-dioxetane.
Key words: DNA damage -- 1,2-Dioxetanes -- Excited states -- Triplet ketones -- Singlet oxygen -- Photosensitization -- Pyrimidine dlmers -- Energy transfer -- Repair endonucleases -- Formam]dopyrimidme-DNA glycosylase -8-Hydroxyguanme -- Azide
INTRODUCTION
1,2-Dloxetanes are four-membered ring peroxides whmh decompose spontaneously (thermal activation) or under the influence of near-UV irradiation (photo-activation) into two carbonyl fragments, one of which is m an excited state [1] (Fig. 1). Such excited carbonyl compounds are of biologqcal interest since they may also be generated in cells under the influence of near-UV irradiation, by decomposition of lipid peroxides and in enzymatic oxidations [2-4]. Similar to other reactive oxygen species such as singlet oxygen (102) and oxygen radmals (superoxide, hydroxyl radicals) excited carbonyl compounds have been proposed to be involved in 'oxidative stress' [5]. The triplet energy of ketones such as acetone and acetophenone is h]gh enough to allow a triplet-triplet energy transfer to DNA, whmh results in DNA modifications identical with those generated by UV irradiation at 260 nm, in particular cyclobutane-type pyrimidine d]mers [6-8]. However, other types of DNA modifications can also be formed either directly by electron transfer reactions or cyclo-addition reactions of the tr]plet excited ketone (type I reactions) or indirectly, e.g. by an energy transfer to molecular oxygen, which results m the formation of highly reactive singlet oxygen (type II reaction) [9-11]. The
o-l' ,s 0--0
HaC'~R H3C CH3
A or near-UV
H3C"~CH3 (R)
+ O
R
H CH3 CH2OH CH2OAc
TrMD TMD HTMD AcTMD
HsC'J~ R (CH3) Fig. 1 1,2-I:hoxetanes used m this study and their fragmentation into two carbonyl products, one of which ]s in an electronically excited state
267
generation of triplet excited states from 1,2-dioxetanes should offer an opportunity to study their interactions with DNA. Incubation of isolated DNA with trimethyl-l,2-dioxetane (TrMD) has been shown to generate thymine dimers [12,13]. However, the spectrum of DNA modifications induced was later demonstrated to be not UV-like, i.e. unknown modifications other than pyrimidine dlmers prevail [14]. In cultured Syrian hamster embryo cells, hydroxymethyltrimethyl-l,2-dioxetane (HTMD) was shown to induce single strand breaks and cause the formation of mmronuclei and cell transformation, but no pyrimidine dimers could be detected [15]. HTMD induced SOS functions in E. coli strains [16]; however, this and several other 1,2dioxetanes proved to be non-mutagenic in Salmonella typh,mumum [17]. Photobinding of psoralen to DNA could be induced by thermal decomposition of HTMD [17,18]. Here, we describe an analysis of the spectrum of DNA modifications induced by TrMD, HTMD, tetramethyl-l,2-dioxetane (TMD) and acetoxymethyl-trimethyl-l,2-dioxetane (AcTMD) (Fig. 1) by means of specific repair endonucleases. MATERIALS AND METHODS
Chemicals The dioxetanes TrMD, TMD, HTMD and AcTMD were synthesized according to published procedures [19- 21]. The purity of the preparations was more than 95% in all cases, as determined by iodometric titration and by 1H-NMR spectroscopy with an internal standard. NDP02 (disodium salt of 1,4-etheno-2,3benzodioxin-l,4-dipropanom acid) was prepared as described by Nieuwmt et al. [22]. DNA from bacteriophage PM2 (PM2 DNA) was prepared according to Salditt et al. [23]. Ninety-five percent of the preparation consisted of the supercoiled form, as determined by the method described below. Enzymes Formamidopyrimidine-DNA glycosylase (FPG protein) was a gift from Dr. S. Boiteux, Villejuif. At the enzyme concentration applied, the incision of modffications induced in PM2 DNA by singlet oxygen was shown to be saturated [24]. Endonuclease III was provided by Dr. R.P. Cunningham, Albany. At the enzyme concentration applied, the incision of modifications reduced in PM2 DNA by osmium tetroxide (thymine glycols) [24] was shown to be saturated. Endonuclease IV was a gift from Dr. B. Demple, Boston. A crude cell extract from Micrococcus luteus (ATCC 4698) was obtained as described by Riazzudin [25]. A partially purified preparation of the UV endonuclease from M. luteus was obtained after streptomycin sulfate and ammonium sulfate precipitations by diethylaminoethyl (DEAE)-cellulose chromatography. This endonuclease was shown to fully recognize UV26°-generated pyrimidine dimers and H +-generated sites of base loss (AP sites) under the incubation conditions but have no activity against DNA modified by osmium tetroxide [24]. Exonuclease III was obtained from Boehringer Mannheim, FRG.
268
DNA mod,ficatwn Exposure of supercoiled PM2 DNA (10 #g/ml) to various 1,2-dioxetanes (up to 40 mM) was carried out in phosphate buffer (5 mM KH2POa, 50 mM NaC1, pH 7.4), which contained 20% acetone in the experiments shown in Fig. i and Table II. The modification was carried out either in the dark under thermal decomposition (1 h, 37°C or 55°C) or under UV-irradiation (360 nm, 5600 J/m 2, Osram HQV black light lamp) at 0°C for 4 rain (photo-induced decomposition). In some experiments H20 as solvent was replaced by D20; the final isotope purity was higher than 95%. In another set of experiments with photo-induced decomposition, argon or pure oxygen was bubbled through the solution by means of a glass capillary. Exposure of PM2 DNA to the endoperoxide NDPO2 (3.5 mM, 2 h in D20 buffer), to xanthine/xanthine oxidase/Fe(III)-EDTA (8 ~M xanthme) and to UV (254 nm) radiation was carried out as described earlier [14,24]. In all experiments, the DNA was precipitated by ethanol/sodium acetate and redissolved m BErbuffer (20 mM Tris-HC1 pH 7.5, 100 mM NaC1, 1 mM EDTA) for damage analysis.
DNA damage analys,s The assay makes use of the fact that supercoiled DNA is converted by either a single strand break (ssb) or the incision of a repair endonuclease into a relaxed (nicked) form which migrates separately from the supercoiled form in agarose gel electrophoresis. Quantification of both forms of DNA by fluorescence scanning allows the determination of the number of single strand breaks per PM2 molecule (10 t base pairs). If an incubation with repair endonucleases precedes the gel electrophoresis the number of single strand breaks plus endonucleasesensitive sites (ess) is obtmned from Eqn. 1 (see Ref. 14). ssb + ess = -ln[1.4"I/(1.4"I + II)] I: Fluorescence of supercoiled form II: Fluorescence of relaxed form
(1)
An aliquot of 0.3 t~g of the modified PM2 DNA in 20 td BE1 buffer was incubated for 30 min at 37°C with 10 t~l of one of the following repair endonuclease preparations: (i) exonuclease III, 300 U/ml in 20 mM Tris- HC1 (pH 8.0), 100 mM NaC1, 15 mM CaC12; (ii) a UV endonuclease preparation from M. luteus, 150 t~g/ml protein in BE15 buffer (BE15 buffer containing 15 mM EDTA); (iii) a crude extract from M. luteus, 0.3 mg/ml in BE15 buffer; (iv) FPG protein, 3 t~g/ml in BE15 buffer; (v) endonuclease III, 40 ng/ml m BE15 buffer and (vi) endonuclease IV, 10 U/ml in BEI~ buffer. The reactions were stopped by addition of 3 ul 10% sodium dodecyl sulfate and the DNA separated by agarose gel electrophoresis. After staining with ethidium bromide, the relative amounts of the supercoiled and the nicked form of PM2 DNA were determined by means of a fluorescence scanner (FTR20, Sigma Instruments, Berlin).
Activation parameters and quantum y~elds The chemiluminescence of 9,10-dibromoanthracene-2-sulfonate was used to
269
quantify the triplet states generated by thermal decomposition of 1,2-dioxetanes in water. Quantum yields were calculated from Stern-Volmer plots. The decomposition kinetics at various temperatures was followed as described earlier [26] for the determination of the rate constants and activation parameters. From the decomposition rate constants (Kd), the 1,2-dioxetane concentration (c) and triplet quantum yields (¢W) the triplet excitation flux (Ep T) was calculated according to Eqn. 2: Ep T = Kd- c . CT. NL
(2)
where NL is the Avogadro number. The triplet excitation flux represents the number of triplet excited states generated from the dioxetane per unit time per unit volume.
TrMD
HTMD
1.2
1,2
0,8
0,8
O~ 0.4
0,4
D. J~
&
0.0 YQ"
0,0
2,0
4,0
6.0
0.0
0.0
w e-
3
(g
2
o
to
.
.
.
.
.
1.0
.
.
.
.
2,0
AcTMD
"ID o
E < Z ¢:Z
20
40
0.0
1,0
2,0
dioxetane concentration [mM] Fig 2 DNA modifications reduced m PM2 DNA by incubation with 1,2-dloxetanes DNA single strand breaks (O) and sum of single strand breaks plus 0) rotes senmt]ve to a crude endonuclease preparatmn from M luteus (e), 00 the UV endonuclease from M luteus (4) and 011) exonuclease III from E cob (A)
270
RESULTS
DNA mod~ficatwns ~nduced by 1,2-dwxetanes are predominantly base modificatwns Supercoiled DNA from bacteriophage PM2 (104 base pairs) was exposed to various concentrations of 1,2-dioxetanes in phosphate buffer at 37°C and 55°C Subsequently, the DNA was analysed for the following types of modifications: (1) endonuclease-sensltive sites recogmzed by a crude protem preparation from M. luteus [25]; (fi) sites of base loss (AP sites) recogmzed by exonuclease III from E. coh [27]; (in) pyrimldme d:mers recognized, m addition to AP sites, by a specific UV endonuclease preparation from M luteus [25] and 0v) DNA single strand breaks. As shown in Fig. 2, all types of DNA modifications increased linearly with the 1,2-dioxetane concentration. From the best-fit slopes of these dose-dependent data, the numbers of modifications mduced per concentration umt were calculated (Table I). The data indmate that for all dioxetanes tested single strand breaks and AP sites represent only a minor fractmn of the total modificatmns
TABLE I SINGLE STRAND BREAKS AND ENDONUCLEASE-SENSITIVE SITES INDUCED BY VARIOUS 1,2-DIOXETANES IN PM2 DNA 1,2Dmxetane
Reactmn cond:tmns
Modificationsa
Time (mm)
Temp (°C)
M luteus b
TrMD HTMD TMD AcTMD
60 60 60 60
37 37 37 37
18 6 + 0 9 f 52+5 64+03 72+5
64 + 05 6 0 + 10 20+02 4 0 + 10
12 + 03 50+06 03+004 4 0 + 10
TrMD HTMD TMD AcTMD
60 60 60 60
55 55 55 55
630 + 780 + 190+ 400 +
114 + 4 101 + 4 24+2 78 + 4
34 100 90 73
TrMD
4 (dark) 4 ( + UV 36o)
02+03
TrMD
43 80 18 20
UV endo c
0
05+04
01+02
0
55+8
305+ 1
Exo III d
+ + + +
ssb e
3 4 1 4
10 + 0 2
0 40 + 0 15 040+02 010+001 090+04 7 + 09 13 + 1 20+03 19 + 2 (-03)+03 100+02
aNumber of modlficat:ons (+ S D ) per 106 base pmrs and mM calculated by hnear regression analysm from data points shown in Fig 2 bModfficat:ons detected by a crude repmr endonuclease preparation from M luteus CModlficatlons detected by a UV endonuclease preparation from M luteus dAp rotes detected by exonuclease III eSmgle strand breaks fA lower value reported earher [14] might have been due to metal 1on lmpumtms
271 TABLE II ACTIVATION AND EXCITATION PARAMETERS OF 1,2-DIOXETANES IN WATER AT 37°C a 1,2-Dloxetane
AG#b (kcal/mol)
tl/2 c (h)
oT d (%)
S T . 10-17 e (sp-1 1- 1)
TrMD TMD HTMD AcTMD
247±10 256±07 250±40 2 6 1 ± 13
65± 03 313± 09 3 8 7 ± 111 589± 30
5±1 4±1 5±1 10±1
9±2 4±1 2± 1 2± 1
apart of the data are taken from Ref 17. bFree enthalpms of activation CHalf-hfe dTrlplet excltlon yield eTrlplet excitation flux for 1 M dloxetane
generated. In the case of TrMD and TMD pyrimidine dimers (the number of which is obtained as the difference between sites sensitive to the UV endonuclease and exonuclease III) account at 37°C for approx. 27% and at 55°C for approx. 10% of the modifications recognized by the crude endonuclease preparation from M. luteus. For HTMD and AcTMD, however, the fraction of pyrimidine dimers is less than 2%. Unknown endonuclease-sensitive base modifications other than pyrimidine dimers prevail in all cases.
The number of pymmid,ne d~mers, but not of other base mod~ficatwns, ,s proportwnal to the triplet excitation flux To quantify the number of excited triplet states generated from 1,2-dioxetanes during incubation in aqueous solution, the activation and exotation parameters of TMD, TrMD, HTMD and AcTMD were determined (Table II). The half-hves and activation energies in water were found to be similar to the values reported earlier for toluene as solvent [21] while the triplet quantum yields turned out to be 3-10-fold lower. From the half-lives and quantum yields an excitation flux was calculated, which gives the number of triplet states generated per liter and second from a 1 M 1,2-dioxetane solution (Table II). The numbers of those DNA modifications which originate directly from the excited triplet carbonyl compounds are expected to be proportional to the corresponding excitation fluxes. Indeed, as shown in Fig. 3, a fair linear correlation is observed for pyrimidine dimers, but not for the other DNA base modifications detected by endonucleases from M. luteus. It has to be concluded that the latter modifications either do not originate from the excited triplet states at all or that their formation involves other (indirect) reaction mechanisms whmh are dependent on the particular structure of the dioxetanes. In the first case, a direct dark reaction between DNA and undecomposed dioxetanes might be considered as
272
Q;
0
~s
N
~'~
~o
~o I
~Z
0
..~
~-~
0
~o
o
&
~
o
o °
,
~
%'-
~
,.~
W
Q;
~o~ N
~j
e~
"Se=
INtU / dq 90L Jed SUOl;eOl;Ipotu VNO
~
273
most likely, since the decomposition products in their ground states, e.g. acetone in the case of TMD, are inert towards DNA under the conditions used (data not shown). The experiments described below were conducted to test this posslbihty.
The DNA modzfications ~nduced by 1,2-dwxetanes omg~nate from excited carbonyl compounds In addition to the thermal (spontaneous) decomposition, the cleavage of 1,2dioxetanes into excited carbonyl products can also be triggered by irradiation at wavelengths between 300 - 400 nm, at which 1,2-dloxetanes have a weak absorption [28]. This photo-induced decomposition can be carried out at low temperatures (0°C) at which dioxetanes do not decompose signfficantly m the dark Irradiation of TrMD on ice at 360 nm in the presence of PM2 DNA g~ves rise to much more DNA modifications than in the dark at the same temperature (Table I). As in the case of thermal decomposition at 37°C or 55°C, base modifications other than pynmldine dimers prevail. The DNA modificatmns mduced by irradiation at 0°C must originate d~rectly or indirectly from the excited carbonyl compounds generated by the fragmentatmn of the 1,2-dioxetanes, since (i) the lifetime of electromcally excited (undecomposed) 1,2-dioxetane is too short for a reaction with DNA [28] and (11)no or very few DNA modificatmns are formed under these conditions in the dark. Interestingly, independent of the method of 1,2-dioxetane decomposition, the fraction of pynmldme dlmers decreases with temperature. The photo-induced reaction generates a much higher percentage of singlet states compared to the thermal decomposition of 1,2-dmxetanes, which ymlds nearly exclusively excited triplet states [28]. Therefore, the result also indicates that the singlet or triplet nature of the excited states generated is not relevant for the DNA damage observed The endonuclease-sens,twe s, tes induced by 1,2-dwxetanes are largely base mod,ficatwns sensitive to FPG protein In order to obtain further information about those dioxetane-induced base modifications which are not pyrimldme dimers, two specffic repair endonucleases from E. coh were used to characterme the modifications, viz, endonuclease III and formamldopynmldme-DNA glycosylase (FPG protein). In addition, another AP endonuclease from E. cob, endonuclease IV, was mcluded in the assays. Endonuclease III has been shown to recognize, in addition to AP sites, various 5,6-dihydropyrimidine residues m DNA [29,30], which are characteristm base modifications induced by hydroxyl radmals [31,32]. FPG protein was originally described as a repair enzyme that recognizes formamldopyrimldines (ringopened purines) [33-35] and was later demonstrated to recognize 8-hydroxyguanme (7,8-dihydro-8-oxoguanme) as well [36,37]. Endonuclease IV probably differs from exonuclease III in the recogmtlon of AP s~tes oxidized m the sugar moiety [38]. The number of modifications recognized by the E coh endonucleases are shown m Fig. 4 for TrMD and HTMD in the form of DNA damage profiles Only a relatively small number of modifications is sensitive to endonuclease III or
274 ~. 3.0 .Q o T~ 1,--
m 2.0 (n ro em o
/ /
1.0
q~
"o o
E Z
0,0
L
TrMD
HTMD
endonuclease-sensltive
[] [] []
FPG protein endonuclease III UV endonuclease
NDPO 2
modifications
[ ] endonuclease IV [ ] exonuclease III
Xanthlns / UV254 X.oxidase/Fe III DNA strand breaks
•
single strand breaks
Fzg 4 DNA damage profiles single strand breaks and various endonucleasesensitive modifications observed in PM2 DNA after treatment with 0) TrMD (6 raM, 1 h, 37°C), (u) HTMD(2 5 mM, 1 h, 37°C), (111)NDPO2 m D20 buffer (3 5 mM, 2 h, 37°C), (Iv) xanthme (8/zM) m the presence of xanthme oxldase and Fe(III)-EDTA (30 mm, 20°C) and (v) UV (254 nm, 8 7 J/m 2, 0°C)
endonuclease IV. Therefore, dihydropyrim]dines and oxidlsed AP s]tes are minor lesions. The relative number of sites recognized by FPG protein, however, is high. The sum of FPG-sensitive sites and UV endonuclease-sensitwe sites is similar to the number of modifications recogmzed by the crude preparation from M. luteus. Therefore, most probably an endonuclease of M luteus analogous to FPG protein is responsible for the activity of the crude M. luteus protein preparation. The damage profiles induced by 1,2-dioxetanes can be compared with those induced by other agents for which the DNA damaging mechanism is better known (Fig. 4). DNA damage profiles induced by UV irradiation at 254 nm (direct excitation of DNA) and by xanthine m the presence of xanthine oxidase and Fe(III)EDTA (generation of hydroxyl radicals, see Ref. 14) are very different from those generated by the 1,2-dioxetanes. In the former case, UV endonucleasesensitive sites (pyrimidme dimers) are generated nearly exclusively, sites sensitive to the other enzymes each make up less than 1% of the total modifications detected. In the case of hydroxyl radicals, the damage profile is characterized by relatively high numbers of AP sites and single strand breaks. The damage profile induced by the endoperoxide NDP02, a chemmal source of singlet oxygen [39,40], however, resembles that of HTMD, which does not generate detectable
275 TABLE III EFFECT OF THE REPLACEMENT OF H20 AS SOLVENT BY D20 ON THE DNA DAMAGE INDUCED BY 1,2-DIOXETANES Mod~ficatmns senmtlve to
FPG protein UV endonuclease
Relative number of modifications (%)a,b HTMD c
TrMD c
199 ± 48 (8) nd
177 ± 25 (6) 113 ± 19 (8)
aObserved m D20 buffer compared to H20 buffer (defined as 100%) bValues represent mean ± S D , the number of determinations is given m parentheses CIncubatlons of the 1,2-dloxetanes with PM2 DNA were for 1 h at 37°C
pyrimldine dimers. Therefore, experiments were carried out to test the role of singlet oxygen in the generation of DNA damage by 1,2-dioxetanes.
S~nglet oxygen is only part~aUy involved in DNA damage by 1,2-dwxetanes The azide amon is known as an efficient, although not specific, quencher of singlet oxygen [41]. The influence of azide ions on the generation of DNA modifications by TrMD is shown in Fig. 5. Both the formation of UV endonuclease senmtive sites (mostly pynmidine dimers) and of the modifications sensitive to FPG protein is suppressed by azide anions, however with very different effectivity. The Stern-Volmer equation for dynamic quenching affords quenching constants of 64 ± 5 mM- 1 and 0.40 ± 0.02 mM- 1 for the FPG protein sensitive modifications and UV endonuclease-sensitive modifications, respectively. The life-time of singlet oxygen is approx. 10-fold greater in D20 than in H20 [42,43]. Therefore, DNA damage mediated by singlet oxygen is expected to increase in D20. In Table III, the effect of the solvent change on the generation of pyrimidine dimers and modifications sensitive to FPG protein is shown for TrMD and HTMD. Whereas the generation of pyrimidine dlmers is not significantly affected, that of FPG-sensitive sites is approximately 2-fold higher in D20. Control experiments indicate that 1,2-dioxetanes do not react significantly with azide ions under the conditions used (data not shown). To test for the participation of oxygen in the process of damage formation, the decomposition of TrMD and HTMD in the presence of DNA was carried out under argon gas, air and pure oxygen gas atmospheres. For practical reasons, the decomposition of the 1,2-dioxetanes in these experiments was induced by near-UV irradiation at 0°C (photo-activation). The results (Table IV) indicate that the generation of FPG-sensitive modffications is not significantly affected by the presence or absence of molecular oxygen. The generation of pyrimidine dimers, however, is lower in air than in argon gas and even lower in the presence of pure oxygen gas.
276 T A B L E IV E F F E C T OF M O L E C U L A R DIOXETANES Modifications sensitive to
OXYGEN
ON
Relative n u m b e r of modifications HTMD c
F P G protein UV endonuclease
THE
DNA
DAMAGE
INDUCED
BY
1,2-
(%)a,b TrMD c
air
oxygen
air
oxygen
97 :e 20 (6) n d
94 ± 18 (5) n d
90 =e 26 (6) 70 + 16 (7)
71 + 14 (6) 33 + 7 (5)
aObserved in air and pure o x y g e n c o m p a r e d to a r g o n (defined as 100%) bValues r e p r e s e n t m e a n + S D , t h e n u m b e r of d e t e r m i n a t i o n s is given m p a r e n t h e s e s CThe 1,2-dloxetanes were Irradiated m t h e presence of PM2 D N A with U V (360 nm) for 4 m m at 0°C
DISCUSSION
The results mdicate that 1,2-dioxetanes modify DNA both under conditions of thermal decomposition at 37°C or 55°C and Irradiation with near-UV (360 nm) at 0°C. At least under the latter conditions the DNA damage is initiated by 1,2dioxetane decomposition into excited carbonyl compounds and cannot be attributed to a direct oxidation or alkylation of the DNA by the 1,2-dioxetanes, whmh potentially can act both as one-electron acceptors [44] and as electrophiles [45]. The spectrum of DNA modifications is apparently the same under conditions of thermal and photo-induced decomposition (Table I). It consists predominantly of DNA base modifications. Only a minor part of the base modifications are pyrimidine dimers. These are well-known to be generated when acetone or acetophenone is Irradiated with near-UV in the presence of DNA [6-8] and result from energy transfer to DNA by the triplet excited carbonyl products. The generation of pyrimidine dimers is reduced in the presence of molecular oxygen (Table IV), probably because oxygen acts as a quencher for the excited carbonyl intermediates. Similarly, also the azide amon apparently can act as quencher of the excited carbonyl products, but only at relatively high concentrations (Fig. 5). A major fraction of the base modifications is recognized by FPG protein (Fig. 4). The only substrates of this repair endonuclease known so far are ringopened purmes (formamidopynmidines) and 8-hydroxyguanosme [33-37]; however, the recognition of other (guanine) modifications cannot be excluded. 8-Hydroxypurines and formamldopyrimidines are known to be generated by hydroxyl radicals by a mechanism that is fairly well understood [46]. However, in the case of hydroxyl radicals as oxidants the generation of the base modifica-
277 10 A 01 C
FPG prote~n UV endonuclease
100" oo~
O ,m
= o
",r-
80"
s
"O
_o 60'
