Mechanism of cleavage of apurinic sites by 9-aminoellipticine.

THEJOURNAL

OF BIOLOGICAL CHEMISTRY 0 1989 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 264, No . 24, Issue of August 25, pp. 14172-1417S,1989 Printed in U.S.A.

Mechanism of Cleavage of Apurinic Sites by 9-Aminoellipticine* (Received for publication, March 8, 1989)

Jean-Remi Bertrandl, Jean-Jacques Vasseurg, Alain GouyetteS, Bernard Raynerg, Jean-Louis Imbachg, ClaudePaolettil, and Claude MalvyST From the SLaboratoire de Biochimie-Enzymologie,UA 147 Centre National de la Rechmche Scientifique, U 140 Institut National de la Sante et de la Recherche Medicale,Institut Gustaue Roussy, 94805 VillejuifCider, France and SLaboratoire de Chimie Bio-Organique, UA 488 Centre National de la Recherche Scientifique, Uniuersite des Sciences et Techniques du Languedoc, Place Eugene-Bataillon, 34060 Montpellier Cedex, France

We have studied the kinetics of breakage of apurinic (AP) sitesby the intercalating agent 9-aminoellipticine using fluorimetric methods with single(ss)- and double (ds)-stranded apurinic DNA. In order to understand the chemical process, high performance liquid chromatography was used to follow the reaction kinetics with the apurinic oligonucleotide model T(AP)T. The unstable intermediate, which is responsible for the 8elimination step, is a Schiff base resulting from the interaction of the amino group of the aromatic amine with the aldehyde function of the deoxyribose moiety (AP site). Fluorescence occurs simultaneously with the breakage of both ss and ds DNA and of the oligonucleotide and arises from the formation of a conjugated double bond onthe Schiff base through the B-elimination reaction. In optimal conditions, the second order rate constant for the fluorescence build up is 15 x lo3 min“ M” for ds DNA and 0.105 X lo3 min” M“ for T(AP)T. The ability of 9-aminoellipticine to induce fluorescence and breakage of ss DNA and T(AP)T shows that intercalation is not essential for this reaction to occur. Nevertheless, the greater rate constant with DNA suggests that stacking is an important parameter for the reaction of the aromatic amine with the AP site.

and phenylhydrazineare also able to cleave DNA at A P sites through a /3-elimination process. Although Livingston (11) proposed that the formation of a Schiff base leads to a pelimination breakage, the exact mechanism was not firmly established. We have previously shownthat the intercalating agent 9-aminoellipticine (9-AE, Fig. 1) cleaves DNA at the AP sites witha high efficiency. Whereas 30 PM Lys-Trp-Lys the AP sites in PM2 DNA, only1p~ is required to break all 9-AE is required under the same conditions (12). We have shown that, while 9-AE breaks AP sites in pBR322 DNA, it has no action on apyrimidinic sites (13). As a consequence of its interaction with apurinic sites, 9AE has been shown to potentiatethe cytotoxic activity of an alkylating agent, dimethyl sulfate, in E. coli (14). It is therefore a candidate for a specific inhibition of DNA repair at apurinic sites. Its binding to AP sites is followed by a timedependent increaseof fluorescence (excitation: 322 nm, emission: 547 nm (15)).The aim of this work is: (i) to determine the exact nature of the fluorophore; (ii) to ascertain whether a Schiff base formation with 9-AE is responsible for the 0elimination breakage; (iii) to find out whether stable covalent and a p u r i n i c D N A adducts canbe formed between this amine and (iv) to determine the compounds formed after mixing apurinic DNA and 9-AE. The different stepsare described in this paper, but for the sake of clarity the final model which can be proposed with the oligonucleotide T(AP)T is indicated in Fig. 7.

Apurinic (AP)l sites inDNA are considered to be common intermediates in mutagenesis (1).AP endonucleases constiEXPERIMENTALPROCEDURES tute the first step the for repair of these lesions and have been described in both prokaryotes(2, 3) and eukaryotes (4). GenMaterials on the 3’-side of the erally, these enzymes cleave DNA either ChemicalCompounds-9-AEwas synthesized as previously deAP site to produce 3’-OH-deoxyribose and 5’-phosphomono- scribed (16, 17). The purity was checked by HPLC (Waters Associor o n t h e 5‘-side to produce 3”OH ester termini (type I) ates) and found to be 95% as determined by UV spectrophotometry 11) at 313 nm (Cia pBondapakOreverse phase column with the following nucleotide and deoxyribose5’-phosphatetermini(type eluent system: 70% methanol, 30% 0.1 M ammonium acetate, pH 5 ) . (for review, see Ref. 5). However, a recent report by Bailly and Verly (6) indicates that Escherichia coli endonuclease I11 T(AP)T 1 was prepared according to a published procedure (18). Nucleic Acids-Calf thymus DNA was purchased from Boehringer acts as a @elimination catalyst and produces 2‘,3’-unsatuMannheim. It was purified by two phenol extractions and sonicated 5’-phosphomonoestertermini(6). Poly- for 30 s at 40 watts with a Branson Sonic sonifier (Lower Co.). The rated deoxyribose and amines (7), the tripeptide Lys-Trp-Lys (8, 9), acridine deriv- DNA concentration is always indicated as the concentration of nuatives (lo), and aldehyde reagents(11) such as semicarbazide cleotides. Apurinic DNA-Calf thymus DNA was alkylated at purines with * This investigation was supported by a grantfrom the Association dimethyl sulfate and thenheat-depurinated as follows: a 130 p M DNA pour la Recherche sur le Cancer (Villejuif)for the project: “Reactivite solution was incubated for 1 h at 37 “C with 30.7 p~ dimethyl sulfate des Sites Apuriniques de I’ADN.” The costs of publication of this (Sigma) in 200 mM sodium cacodylate, 130 mM sodium perchlorate, article were defrayed in part by the payment of page charges. This 1 mM EDTA, at pH 7. In order to create one A P site per 25 bases, article must therefore be hereby marked “aduertisement” in accord- depurination was performed by heating the methylated DNA for 6 h at 55 “C in 25 mM sodium acetate, pH 5 (15, 19, 20). The amount of ance with 18 U.S.C. Section 1734 solely to indicate this fact. li To whom correspondence and reprint requests should be ad- AP sites was estimated according to the twofollowing described results (15, 20). (i) The number of methylated bases is proportional dressed. The abbreviations used are: AP, apurinic; 9-AE, 9-aminoellipti- to dimethyl sulfate concentration and is therefore easily calculated cine; HPLC, high performance liquid chromatography; pdT, 5”phos- for 30.7 p~ dimethyl sulfate. (ii) In the above described conditions, all alkylated bases are removed from DNA. The number of apurinic phate thymidine.

14172

Mechanism of Cleavage of Apurinic DNA by 9-AE %N&

CHJ

9-A

E

FIG. 1. Structure of the aromatic amine 9-AE. sites is therefore equal to the number of methylated bases. High Performance Liquid Chromatographic Analyses-These were carried out on a Radial-Pak CIS (10 pm) cartridge in a Waters RCM 100 module. A Waters U6K injector, two 6000 A pumps, a M720 solventprogrammer, and a M730 microprocessor-controlled data system were used. UV spectra were recorded with a Pye-Unicam PU 4021 multichannel detector and a Pye-Unicam PU 4850 video chromatography control center. Proton Nuclear Magnetic Resonance ('H NMR) Spectra-These spectra were recorded a t 72 "C with a Bruker WB UM 360 spectrophotometer. NMR samples were made up in dimethyl sulfoxide-d6. Fast Atom Bombardment (FAB) Mass Spectra-These spectra were recorded in positive mode or in negative mode on a double focusing mass spectrophotometer VG 70-250 fitted with a FAB gun (Ion Tech, United Kingdom). Glycerol was used as thematrix. Fluorescence Measurements-These measurements were performed with a Kontron SFM 23/B spectrofluorimeter coupled with a LKB 2210 recorder. Thermostated 1-ml quartz cuvettes (Hellma) were used. Methods Fluorescence-The DNA was dissolved in 10 mM sodium phos. phate, 1 mM EDTA, pH 7.4, at a final concentration of 65 p ~ The excitation wavelength was set at 322 nm and emission at 547 nm. Kinetic data were entered on a Minc computer (Digital), and a fit was performed with a Marquart algorithm. This allowed the determination of the kinetic parameters. T(AP)T was dissolved either in 10 mM sodium acetate, 1 mM EDTA, pH 4, or in 10 mM sodium phosphate, 1 mM EDTA, pH 7.4. Fluorescence spectra and kinetic measurements (excitation:322 nm, emission: 547 nrn) were performed at pH4 and 7.4. Measurement of DNA Cleavage-65 pM apurinic calf thymus DNA was incubated with 6.6 PM 9-AE at 25 "C in 10 mM sodium phosphate, 1 mM EDTA, pH 7.4. The reaction was stopped at various times by treating 500-pl aliquots with 6 M sodium borohydride (final concentration 0.35 M ) for 10 min. Neutralization was then performed with 250 p1 of 1 M HCI, and DNA was precipitated overnight in ethanol at -70 "C. After centrifugation, pellets were dissolved in 3.5 M urea, 15% saccharose, 15 mM EDTA, 0.45 M Trishorate, pH 8, then boiled for 10 min and cooled in ice. Electrophoresis was then performed in a 7 M urea, 5-20% polyacrylamide gel for 2 h at 300 V with a BioRad protein apparatus. After staining with ethidium bromide, a direct scanning of fluorescence was performedwitha RFT Transidyne scanner (21). Data were collected on an online M21 Olivetti computer using thePclab software (DATA TRANSLATION) for analogic digital conversion. Phage 3x174 DNA digested by HaeIII was used as a molecular weight marker. Strand breakage was determined from analysis of the DNA length pattern by the scanning of each slot. At any migration distance, the fluorescence intensity of ethidium bromide is proportional to the concentration of DNA according to the equation: IF(d) = Q X C where IF(d) = fluorescence intensity at migration distance d, Q = proportionality factor, and C = DNA concentration. Fragment size in bases S is given by the equation: logS=-axd+b where d = migration distance, and a,b = constants. a and b are determined using the standard curve obtained with 9x174. The DNA strand concentration (SC) a t a distance d can therefore be expressed as: SC(d) =

Wd)

Q x (exp(-a

X

d

+ b))

Integration of SC from the start to thebottom of the gel leads to a determination of the total strandconcentration which is the concentration of breaks.

14173

Kinetics of p d T Elimination as Function of 9-AE Concentration during Reaction with Compound 1-The rate of @-eliminationof pdT was investigated at pH 5. Reactions of 1 (0.1 mM) with 9-AE (0.5, 0.6, and 0.8 mM) were performed in amethanol, 100 mM acetate buffer (32, v/v) mixture at 25 "C. At various times, aliquots (30 p I ) were withdrawn and analyzed by HPLC. Elution was performed at a flow rate of 3 mlmin" with alineargradient of 0 to 8.5% of acetonitrile in 0.1 M ammonium acetate, pH 5.9, in 20 min. The apparent rate constantsof k', and k'? (Fig. 5) were calculated from the half-time value of pdT elimination (k'' (9-AE) = In 2/t,,+) and from the tmeX and Cmaxvalues of compound 2.

p d T Elimination as Function of p H during Reaction of I with 9AE-The extent of breakage a t 37 "C of T(AP)T (0.1 mM) by 9-AE (0.1 mM) at 37 "C was investigated at 5 pH values (1 mM EDTA, 10 mM sodium dihydrogen phosphate adjusted to pH values ranging from 3 to 7 with phosphoric acid or sodium hydroxide). Reduction of the Schiff Base Formed between 9-AE and TfAP)T9.8 X mmol) was added to a 9-AE(diacetate form, 3.7 mg, mg, 9.8 X mmol) and sodium solution of 1 (Na' form, 7.7 X 10" mmol) in 10 mM sodium cyanoborohydride (61.6mg,9.8 phosphate (pH 5, 98 ml; 1 mM with respect to EDTA), 1 mM EDTA, pH 5, and the mixture was kept at 37 "C for 1 h. After evaporation of the solvent mixture, compound 3 (Fig. 11)was purified by reverse phasechromatographyon an Amicon Cs column using stepwise increments of methanol (0 to 100%) in water, followed by reverse phase HPLC. A linear gradient of 0-25% acetonitrile, 0.1 M ammonium acetate, pH 5.9, was applied for 20 min at a 3 ml min-' flow rate. The appropriate fractions were combined, evaporated, dissolved in water, and applied on a Sep-Pak c18 cartridge (washing withwater and then elution with a 50:50 v/v methanol/water mixture) to give compound 3 as its ammonium salt (6.8 mg, yield 68%; HPLC purity better than 98%) afterlyophilization. The identity of this compound was established by mass spectrometry and NMR spectroscopy analysis. Fitting of Fluorescence Kinetics Obtained with T(AP)T-Fluorescence kinetics parameters were determined by a computer analysis on an Olivetti M21 computer. Kinetic experiments performed with DNA suggested that fluorescence is caused by the accumulation of a product occurring shortly after DNA cleavage. HPLC experiments indicate that the reaction proceeds further aftercleavage of T(AP)T. Thereaction was therefore modeled according to a 2-step system: kl

kz

A+B-+AB+X A = 9-AE, B = AP sites, AB is the fluorophore, X = uncharacterized unstable compound (see below). AB = b X (exp(-kl X t ) (exp(-kz x t)), where b is a proportionality factor.

-

RESULTS

The Reactionbetween 9-AE and AP Calf Thymus DNA Triggers a Time-dependent Fluorescence-The formation of a Schiff base between amines and apurinic DNA has been suggested by Livingston (11).It hasalso been shown that the introduction of a conjugated C=N bond on the benzene ring opposite to the pyrimidine ring of the ellipticine molecule enhances its fluorescence properties inDNA (22). 9-AE is not fluorescent in DNA. We therefore looked for an increase of fluorescence when mixing apurinic DNA and 9-AE as indirect evidence for Schiff base formation. Fig. 2 shows that when calf thymus DNA (65 p M ) , carrying one AP site per 25 bases, is mixed with 9-AE (from 0.8 pM to 13 pM), an asymptotic build up of fluorescence is observed with a plateau dependent on the concentration of 9-AE. This plateau increases from 0.8 p~ to 6.6 pM and thendecreases. The reaction follows pseudois first order kinetics above 6 PM 9-AE, and the rate constant 15 X lo3min" M" at 25 "c(result not shown).We previously described (12) that 9-AE induces breakage at apurinic sites. We have verified (results not shown) that the number of

Mechanism of Cleavage of Apurinic DNA by 9-AE

14174

w 15

W

I

0 0

//

1

1

1

I

10

20

30

40

TIME

0

10

(rnin.)

20

30

40

TIME (min.)

FIG. 2. Fluorescence is time-dependent when apurinic calf thymus DNA is mixed with 9-AE (excitation: 322 nm; emission: 547 nm). The observed kinetic curve is dependent on 9-AE concentration. In all experiments, calf thymus DNA with one AP site per 25 bases is used at 6.5 X M bases in the following buffer: 10 mM phosphate, 1 mM EDTA, pH 7.4, as described under “Methods.” The experimental curve is fitted either with a biexponential model (graph A ) or with a monoexponential model (graph B).

breaks increases with concentration when the concentration of 9-AE is higher than 6.6 p ~Therefore, . the decrease of the fluorescence plateau is not due to a lower number of breaks, but more likely to quenching by a high dose of chromophore (23). We therefore investigated the relationship between fluorescence and breakage at 6.6 /*M in order to ascertain the role of the fluorophore in thecleavage reaction. Comparison of Fluorescence and Cleavage on ss and ds DNA-After various times (from0 to 1h) of contact between 6.6 PM 9-AE and 65 p M apurinic DNA (1 site per 50 bases), aliquots were analyzed on an ethidium bromide-stained denaturing gel (5-20% acrylamide, 7 M urea). After gel scanning, the extentof cleavage was calculated from the analysis of the migration distances (see “Methods”). The number of breaks was at a plateau between 40 and 60 min incubation time at 25 “e.

We cannot ascertain that the calculated number of breaks has an absolute meaning. The value at 60 min was therefore arbitrarily set to 100%.However, cleavage at theplateau is in

the same range as the number of apurinic sites induced by dimethyl sulfate (1per 50 bases). Fig. 3 shows that when the same calculation is made on the fluorescence data, both kinetics of fluorescence and breakage are quite similar. These results, suggest that the chemical events responsible for fluorescence and for breakage occur within a very short time span. They could therefore be identical or closely related. When apurinic ss calf thymus DNA was used, identical results were obtained for both fluorescence and cleavage (results not shown), indicating that intercalation is not required for interaction of 9-AE with AP sitesin DNA. According to Livingston’s hypothesis, the reaction between 9AE and theapurinic site would produce a Schiff base, then a @-eliminationprocess would induce strand breakage. In order to trapa possible Schiff base, experiments were performed in the presence of300 mM sodium cyanoborohydride. This reagent is known to selectively reduce Schiff bases (24,25). We observed a ZOO-fold increase in fluorescence intensity with a shift toward low wavelengths (Xmaxex293 nm, , , , . ,X 435 nm) when compared to the fluorescence obtainedwithout the reducing agent at pH 7.4 ,,X(, 322 nm, Amax 547 nm (15)). Meanwhile, no cleavage of apurinic DNAwas detected (results not shown). This result is consistent with the occurrence of a Schiff base in DNA prior to cleavage at apurinic sites. Neither fluorophore nor cleaved apurinic sites canbe better characterized using apurinic DNA. This study has therefore been extended to theapurinic oligonucleotide T(AP)T. Evidence for an Intermediate Schiff Base between 9-AE and T(AP)T-Experiments were performed with T(AP)T (0.1 mM) and various concentrations of 9-AE in a methanol, 0.1 M acetate buffer at pH 5 (2:3, v/v). They indicate the disappearance of 1 according to a pseudo-first order kinetics and the concomitant formation of the 2‘,3’-unsaturated aldehyde 2 and pdT (Fig. 4). A compound X which could result from 1,4 addition of 9-AE to 2, as previously found with 3-aminocarbazole (26), was detected by HPLC, but itwas too unstable to allow further characterization. The following values were found for the apparent rate constants(Fig. 5): k ’ l , 9.7 X

0 2

a -s z w

a% > w

5% A

W

u a

0

a 3

2 2

o u O .

2

00

20

40

60

T I M E (min)

FIG. 3. Comparison of 9-AE-induced apurinic DNA cleavage and fluorescence (excitation: 322 nm; emission: 647 nm) as a function of time. 6.5 X lo-’ M double-stranded apurinic DNA (1 site for 50 bases) was mixed with 6.6 X M 9-AE. Strand cleavage was determined from gels, as described under “Methods.” Data are expressed as a percentage of maximum effect.

@dT

j

*@df

2

FIG. 4. Observed intermediates duringthe reaction between 9-AE and T(AP)T. P indicates 3’,5’-phosphodiester linkage.


14175

cyanoborohydride at pH 5. We then observed at the same time the disappearance of T(AP)T and the formation of a new compound 3 without pdT elimination. This derivative was purified, and its structure was established by NMR and mass spectroscopy as theproduct of the reduction of a Schiff base between the aromatic amine and the aldehyde function in the 1’ position of deoxyribose (Fig. 7, compound 3).Positive and negative ion FAB mass spectra of the ammonium salt of 3 (Fig. 8A) showed MH’ at m/z 988 and (M - HI- at m/z 986, respectively. Furthermore,fragments at m/z 245 (Ar’) and m/z 742 (M - Ar)- were observed, respectively, in positive and negative ion FAB mass spectra, establishingthat the aminoellipticine (ArNH2)was linked through its N-9atom to the oligonucleotide moiety. Fig. 8B shows the ‘H NMR spectrum of 3 in the region 5.5-10 ppm. The assignment of each signal was established by means of selective decoupling experiments taking intoaccount the downfield shift of protons TIME Imin) ortho to N atom (H-1 andH-3) in pyridine moiety of pyridoFIG. 5. Kinetics of the reaction between T(AP)T (0.1 mM) carbazole heterocycles (28, 29). Compound 3 is a fluorescent and 9-AE (0.5 mM). Reaction of 1 (0.1 mM) with 9-AE (0.5 mM) was performed as described under “Experimental Procedures.” In the derivative (Fig. 9A, excitation 305 nm, emission 497 nm) of first step, the unsaturated aldehyde 2 is formed upon 5”phosphate 9-AE which itself is not fluorescent. We have therefore investhymidine (pdT) (%elimination. Then, addition of 9-AE to 2 results tigated the action of 9-AE on T(AP)T by fluorescence. in compound 3.This last product was too unstable to be quantified. Fluorescence Study of the Reaction between 9-AE and The rate constantswere calculated according to mechanism (Fig. 4). T(AP)T-The appearance of 3 was followed by fluorescence Theoretical curves (lines) ( k f 1= 9.7 min-’, k’2 = 9.2 rnin”) were in good agreement with experimental concentration values of after mixing 10 PM 9-AE and 10 p~ T(AP)T in the presence of 350mM sodium cyanoborohydride. These experimentswere compounds 1,2,and pdT. performed in a pH 4 buffer (10 mM acetate, 1 mM EDTA) which has been shown above to be the optimal pH for cleavage of T(AP)T. As expected, we observed saturable kinetics of fluorescence using the excitation and emission wavelengths of 3. The fluorescence spectra of isolated 3 and of the compounds formed after reaction at pH 4 between 9-AE and T(AP)T in the presence of cyanoborohydride are almost identical (Fig. 9). Fig. 10 shows that when 0.1 mM T(AP)T was mixed with 0.5 mM 9-AE (pH 4,lOmM acetate, 1 mM EDTA) a biphasic fluorescence curve was recorded, between 0 and 20 min at 25 “C, using the parameters described above for apurinic DNA (excitation: 322 nm, emission: 547 nm). At pH 7, whatever the concentrations of 9-AE and T(AP)T were, no fluorescence was observed. A two-step reaction (see fitting of fluorescence under “Methods”) can account for the fluorescence at pH4 in aqueous buffer. The rate constantkl, for the pH of Reactlon M l x t u r e build up of fluorescence, was found to be 105 min” M” and FIG. 6. pH dependence of the reaction between 9-AE and the rate constant k2, for the decrease of fluorescence, to be T(AP)T. Reaction between 1 (0.1 mM) and 9-AE (0.1 mM) was 1.53 min”.

performed as described under “Experimental Procedures.”The extent of pdT elimination was determined by HPLC analysis of an aliquot (30 pl) after 5 min of reaction.

min” at 0.5 mM 9-AE; k’*, 9.2 X rnin”. The value of k’’ appears to be linearly related (result not shown) to the concentration of 9-AE in excess, which indicates a second order reaction (kl = 19 min” ”l). We also studied the p-elimination breakage of T(AP)T by 9-AE in aqueous buffer from pH 3 to 7. Fig. 6 shows that this breakage is maximum at about pH 4. In opposition to the initial simplified model of Fig. 4, the second order reaction for 9-AE-induced cleavage of T(AP)T indicates that 9-AE is directly involved in the chemical mechanism. This result and the above described influence of pH are consistent with the occurrence of an intermediate Schiff base (27). A short lifetime of this Schiff base can account for the absence of its detection by HPLC. Therefore, the limiting would be the formation of step in the cleavage reaction (JZ’~) the Schiff base. In order to ascertain the occurrence of this derivative, we performed experimentsin the presence of 10 mM sodium

DISCUSSION

9-AE displays two unusual features when mixed with apurinic DNA: a breakage at low dose (12) and the development of a specific fluorescence (15). In this article, we have compared the action of 9-AE on apurinic DNA and on the apurinic oligonucleotide T(AP)T. We have shown in both cases the occurrence of breakage and fluorescence. However, fluorescence measurements hadto be performed at pH 4 with T(AP)T. The occurrence of a Schiff base is the most likely hypothesis to account for the 9-AE-induced breakage of DNA at apurinic sites. Hydrazine derivatives have indeed been shown to form hydrazones with apurinic sites in DNA (ll), andwe recently studied this reaction with our T(AP)T model (30). Furthermore, ethidium bromide has been shown to form a Schiff base with apurinic RNA (31). However, two results do not support this hypothesis for 9-AE: 1) derivatives of ellipticine, which are deprived of exocyclic amines, are also able, although with lower efficiency, to break DNA at apurinicsites (12); 2) intermediate compounds detected by HPLC in the reaction

14176


FIG. 7. Summarized chemical pathway of the reaction between 9apurinic theAE and oligonucleotide

'"zNw CM3

''@P*'_dT@p+a; c

T(AP)T. Pindicates 3',5'-phosphodiester linkage.

t

M

/

/

"@-H=NQ~

~

I

@ *T

cn3

~

A

~~''J$'NH H

O

6

CH3

0 HO'

' FAB' MH' mlz

980

FAB- IM-HI- mlz 986

FIG. 8. Mass and NMR data of 9-AE derivative 3. A, main ions observed from mass spectrometry. B, 'H NMR spectrum (6-10 ppm) at 360 MHz.

OH

.

B

n

TH 6

-H

N

181

' 171

)Ifti10 TH6

W 0 13)

to1

H

CH 3

-

10

9.5

9

of 9-AE with T(AP)T have been identified on the basis of their UV spectra; no Schiff base has been detected. This suggests that either no Schiff base occurs during that reaction or that it could not be detected by HPLC due to its low concentration and/or high reactivity in a next reaction step. This last hypothesis is the right one. Indeed, experiments performed in the presence of the reducing agent sodium cyanoborohydride indicate that a new fluorescent compound is accumulating during the reaction and cleavage is inhibited. This compound was isolated and appears to be a reduced Schiff base (this property can be used to synthesize oligonucleotides covalently linked at their 3' terminus to an intercalating agent) (32). The formation of a Schiff base is therefore a mandatory step before the occurrence of a nick in T(AP)T. Experiments performed with sodium cyanoborohydride also demonstrate the similarity of the reaction of

8.5

8

s ( PPm)

7.5

7

I

6.5

I . .

6

T(AP)T and apurinic DNA with 9-AE. In both cases, we observed a great increase of fluorescence and a suppression of cleavage. Fluorescence spectra of T(AP)T or DNA with 9AE and cyanoborohydride are very similar (Figs. 9B and 11). They are also similar to thespectrum of the purified reduced Schiff base (Fig. 9A). In absence of sodium cyanoborohydride, breakage and fluorescence are therefore probably caused by the formation of a Schiff base between 9-AE and deoxyribose in the AP site. The chemical process which leads tothe @-elimination breakage is therefore identical with the scheme proposed by Livingston (11).According to our results, one might suggest that intrinsic intensities of fluorescence of the Schiff base before cleavage, its reduced form, and the a,@-unsaturated Schiff base are similar, although spectramay vary. However, because of the short life of the Schiff base before cleavage,


14177

c N

200

300

400

500

600

700

WAVELENGTH (nm)

TIME

(min.)

FIG.10. Fluorescence (excitation 322 nm, emission 547 nm) is time-dependent when 0.1 mM 9-AE is mixed with 0.5 mM T(AP)T in 10 mM sodium acetate, 1 mM EDTA, pH 4. Graph A: the observed kinetic curve indicates three different phases. Graph B: computer modeling of the first two phases: -, observed fluorescence; - - -, computer modeling. Calculations are described under “Methods.” Kinetic constants are k , 105 min” M”, k, 1.53 min”.

1

306

497

WAVELENGTH (nm) FIG.9. Emission and excitation spectra of sodium cyanoborohydride-treated solutions of 9-AE and T(AP)Tsites. Graph A: 3.3 PM HPLC-purified compound 3 in 0.1 M sodium acetate, 1 mM EDTA, pH 4. Graph B: 0.1 mM T(AP)T was mixed with 0.1 mM 9-AE for 1 h in 0.35 M CNBH3Na, 0.1 M sodium acetate, 1 mM EDTA, pH 4.

the fluorescence observed at pH 7.4 on DNA is probably caused by the a,@-unsaturatedSchiff base. In thepresence of cyanoborohydride, no DNA cleavageoccurs, the reduced form of the Schiff base before cleavage accumulates, and fluorescence increases. One can expect the spectrum of the a,@unsaturated Schiff base to be different from the spectrum of the Schiff base prior to cleavage. This may explain the fact that thefluorescence spectrum of AP DNA in the presence of 9-AE is different if the AP DNA is first reduced with cyanoborohydride. We have not succeeded in isolating this Schiff base with T(AP)T, although a new compound is detected in HPLC analysis when cyanoborohydride is added after breakage. We have nevertheless demonstrated with phenylhydrazine that the @-eliminationprocess gives the corresponding a,@-unsaturatedhydrazone as an intermediate of reaction (30). In thiscase, the relative stability of this hydrazone could be due to theconjugation of the a,@-unsaturated imino system with the nitrogen atom doublet of phenylhydrazine (27). Ac-

W A V E L € N G T H (nm)

FIG.11. Emission and excitation spectra of sodium cyanoborohydride-treated solutions of 9-AE and apurinic DNA. 6.5 X 1O“j M apurinic DNA (1 site per 25 bases) was mixed with 6.6 X P M 9-AE in: 0.35 M CNBH3Na, 0.1 M sodium acetate, 1 m M EDTA, pH 4.

cordingly, one can assume that the supposed Schiff base during reaction of T(AP)T with 9-AE is not stable enough and leads to otherderivatives. This is consistent with the bell shape kinetic curve obtained for fluorescence (Fig. 10). One can suggest that the low stability of the fluorophore with T(AP)T, compared to apurinic DNA, could be caused by differences in stacking interactions. The secondary increase of fluorescence with higher incubation times confirms the evolution of the reaction. A t this point, 9-AE induces a precipitation of the mixture when working with the apurinic oligonucleotide.A characterizationof the products is therefore not technically possible. The only complete reaction that can be followed is the one between T(AP)T and3-aminocarbazole (26). According to the structure of the final adduct isolated with 3-aminocarbazole, one might imagine a dissociation of the a,P-unsaturatedSchiff base followed by 1,4-addition of 9AE to the 2‘,3’-unsaturated deoxyribose in the apurinic site.

14178


This, via unstableintermediates, would lead to the 9-AE analogue of the final adduct observed with 3-aminocarbazole (26). The action of 9-AE at low doses is due either to its high affinity binding for DNA (33) or to its preferential reactivity with the AP site itself. This reactivity could be related to structural properties of 9-AE and to its ability to form a Schiff base prior to p-elimination. We have shown that the rate constantfor breakage is two orders of magnitude higher with apurinic ss DNA than with the oligonucleotide T(AP)T. Published resultssuggest that 9AE does not need to first bind with the unmodified part of DNAbefore reacting with AP sites (12). This higher rate constant could therefore be caused by a different structure of AP sites in single-stranded (ss) DNA and in the less stacked T(AP)T. The AP site structure could indeed influence the occurrence of a Schiff base. In conclusion, we have shown that the properties of the DNA binding agent, 9-AE, toward apurinic sites (cleavage and specific fluorescence) are caused by the formation of a Schiff base at 1’-position of deoxyribose. This could be used to detect apurinic sitesin cells treated by ionizing radiations or anti-tumor drugs. According tothe results obtained with the related compound, 3-aminocarbazole (26), one might envisage that the Schiff base leads to a final adduct on the 3‘ position of deoxyribose. 9-AE, or related compounds, could be potential candidates for inhibition of DNA repair (14) through chemical reaction with apurinic sites. Acknowledgment-We thank Dr. Annette Larsen(Villejuif) for her help in the writing of the manuscript. REFERENCES 1. Loeb, L. A. (1985) Cell 40, 483-484 2. Shaaper, R. M., and Loeb, L. M. (1981) Proc. Natl. Acad. Sci.

U. S. A. 78,1773-1777 3. Kunkel, T. A. (1984) Proc. Natl. Acad. Sci. U. S. A. 281, 14941498 4. Gentil, A., Margot, A., and Sarasin, A. (1984) Mutat. Res. 129, 141-147 5. Lindahl, T.(1982) Annu. Reu. Biochem. 51, 61-87 6. Bailly, V., and Verly, W . G. (1987) Biochem. J . 242, 565-572 7. Lindahl, T.,and Andersson, A. (1972) Biochemistry 11, 36183623

8. Pierre, J., and Laval, J. (1981) J. Biol. Chem. 256, 10217-10220 9. Behmoaras, T., Toulm6, J. J., and HClene, C. (1981) Nature 292, 858-859 10. Constant, J. F., O’Connor, T. R., Lhomme, J., and Laval, J. (1988) Nucleic Acids Res. 1 6 , 2691-2703 11. Livingston, D. C. (1964) Biochim. Biophys. Acta87,538-540 12. Malvy, C., Prevost, P., Gansser, C.,Viel,C., and Paoletti, C. (1986) Chem.-Biol. Interact. 67,41-53 13. Malvy, C., and Bertrand, J. R. (1986) FEES Lett. 208, 155-157 14. Malvy, C., Safraoui, H., and Bertrand, J. R. (1988) Anti-Cancer Drug Des. 2, 361-370 15. Bertrand, J. R., Malvy, C., and Paoletti, C. (1987) Biochem. Biophys. Res. Commun. 143, 768-774 16. Dat-Xuong, N., Adeline, M. T., Lecointe, P., and Janot, M. M. (1975) C. R. Hebd. Seances Acad. Sci. Ser. C Chim. 2 8 1 , 623626 17. Gansser, C., Viel, C., Malvy, C., and Cros, D. (1980) I1 Farmaco 11,887-895 18. Vasseur, J. J., Rayner, B., and Imbach, J. L. (1986) Biochem. Biophys. Res. Commun. 134,1204-1208 19. Behmoaras, T.,Toulme, J. J., and Helene, C. (1981) Proc. Natl. Acad. Sci. U. S. A. 78, 926-930 20. Uhlenhopp, E. L., and Krasna, A.I. (1971) Biochemistry 10, 3290-3295 21. Malvy, C. (1984) Anal. Biochem. 143, 158-162 22. Banoun, H., Le Bret, M., and Auclair, C. (1985) Biochemistry 2 4 , 701-707 23. Udenfriends, S. (1962) in Fluorescence Assayin Biology and Medicine, pp. 13-37, Academic Press, Orlando, FL 24. Borch, R. F., Bernstein, M. D., and Dupont Durst, H. (1971) J. Am. Chem. SOC.93,2897-2904 25. Lane, C. F. (1975) Synthesis 135-146 26. Vasseur, J. J., Rayner, B., Imbach, J. L., Verma, S., McCloskey, J. A., Lee, M., Chang, D.-K., and Lown, J. W. (1987) J. Org. Chem. 52,4994-4998 27. Sollenberger, P. Y.,and Martin, R. B. (1968) in The Chemistry of the Amino Group (Patai, S., ed) pp. 349-406, Wiley Interscience, London 28. Van der Donckt, E., Martin, R. H., and Geerts-Evrard, F. (1964) Tetrahedron 20,1495-1503 29. Perche, J. C., Saint-Ruf, G., and Buu-Hai, N. P. (1972) J. Chem. SOC. PerkinI260-262 30. Vasseur, J. J., Rayner,B., and Imbach, J. L. (1988)J. Heterocyclic Chem. 25,389-392 31. Nishikawa, K., Adams, B. L., and Hecht, S. M. (1982) J. Am. Chem. SOC.104,326-328 32. Vasseur, J.-J., Gauthier, C., Rayner, B., Paoletti, J., and Imbach, J. L. (1988) Biochem. Biophys. Res. Commun. 152,56-61 33. Le Pecq, J. B., Dat-Xuong, N., Gosse, C., and Paoletti, C. (1974) Proc. Natl. Acad. Sci. U. S. A. 71,5078-5082