Proc. Natl. Acad. Sci. USAVol. 83, pp. 6317-6321, September 1986Biochemistry

Reaction of nucleic acids and cis-diamminedichloroplatinum(II) inthe presence of intercalating agents

[ethidium bromide/proflavine/acridine/trans-diamminedichloroplatinum(ll)]

JEAN-MARC MALINGE AND MARC LENGCentre de Biophysique Moldculaire, Centre National de la Recherche Scientifique, 1A, Avenue de la Recherche Scientifique, 45071 Orleans, Cedex 2, France

Communicated by Gary Felsenfeld, May 9, 1986

ABSTRACT The reaction of cis-diamminedichloroplati-num(II) and several synthetic or natural double-strandedpolydeoxyribonudeotides has been carried out in the presenceof such intercalating agents as ethidium bromide, proflavine,and acridine. After incubation of the reaction mixtures at 37Cfor 24 hr, some ethidium or proflavine, but no acridine,molecules are tightly bound to nucleic acids. Tight binding isdefined by resistance to extraction with butanol, assayed byfiltration at acid pH or by thin-layer chromatography at basicpH. In the ternary complexes, there is about one tightly boundethidium (or proflavine) per platinum residue. At 370C, but notat 40C, tightly bound ethidium exchanges with free ethidium,whereas platinum residues do not exchange. The binding andthe release of tightly bound ethidium are very slow (severalhours). It is suggested that in the ternary complexes, nucleicacid-cis-Pt(NH3)2-intercalating agent, a bidentate adduct (gua-nine-ethidium or -proflavine)-cis-Pt(NH3)2, is formed. Notightly bound ethidium or proflavine is found when cis-diam-minedichloroplatinum(II) is replaced by trans-diamminedi-chloroplatinum(II). Competition experiments between cis-di-amminedichloroplatinum(II), poly(dG-dC), and poly(dG)--poly(dC) or poly(dA-dT) show that the presence of ethidiumbromide, proflavine, or acridine interferes with the distribu-tion of platinum between the polynucleotides. These resultsmight help to explain the synergism for drugs used in combi-nation with cis-diamminedichloroplatinum(ll) and in the de-sign of new chemotherapeutic agents.

Numerous works have been devoted to the study of thebinding of the antitumor drug cis-diamminedichloroplati-num(II) [cis-Pt(NH3)2Cl2] to DNA. cis-Pt(NH3)2CI2 is clini-cally active against many different neoplasms and it is oftenused in combination with other anticancer drugs, such asdoxorubicin (adriamycin) or bleomycin. Several studies havereported enhanced effects for combination drug therapy,including synergy (1).The mechanism of action of cis-Pt(NH3)2Cl2 is not yet

known, but it is often thought that its antitumor activity isrelated to its binding to DNA (2-5). In the reaction ofcis-Pt(NH3)2C12 and DNA, guanine residues are the preferredbinding sites but adenine and cytosine residues also react,which leads to the formation of various types of adducts(6-11). On the other hand, for a given DNA, the nature of theadducts depends upon DNA conformation. In the reaction ofcis-Pt(NH3)2CI2 and poly(dG-m5dC) in the Z conformation, amonodentate adduct is formed. In the reaction of cis-Pt(NH3)2CI2 and poly(dG-m5dC) in the B conformation, abidentate adduct is formed (12).The conformations of B-DNA and Z-DNA are very differ-

ent (13, 14). To know whether smaller conformationalchanges of DNA could also play a role in the nature of the

adducts, we have carried out the platination of nucleic acidsin the presence of intercalating compounds. It has beenalready reported that new cis-Pt(NH3)2Cl2 binding sites weredetected by means of exonuclease III when DNA wasincubated with the intercalating agent ethidium bromidebefore the addition of cis-Pt(NH3)2Cl2 (15, 16). We show hereby competition experiments between various nucleic acidsthat the preferential binding of cis-Pt(NH3)2Cl2 to (dG)n (dC)nsequences is abolished when the reaction is carried out in thepresence of ethidium bromide, proflavine, and acridine. Inaddition, we show that the stability of the ternary complexes[nucleic acid-cis-Pt(NH3)2-dye] strongly depends upon thenature of the dye. In the ternary complexes, ethidium andproflavine are tightly bound to the nucleic acids. They cannotbe removed from the nucleic acids, as judged by severalassays, such as extraction with organic solvent, filter assayat acid pH, or thin-layer chromatography (TLC) at basic pH.By these three assays, acridine is completely removed. It issuggested that cis-Pt(NH3)2 cross-links guanine residues andethidium or proflavine but not acridine.

MATERIALS AND METHODSCalf thymus DNA and poly(dG)-poly(dC) (purchased fromBoehringer Mannheim), poly(dG-dC), poly(dA-dC)-poly(dG-dT), and poly(dA-dT) (purchased from Pharmacia) weretreated twice with phenol and then precipitated with ethanol.Stock solutions of nucleic acids were made in 10 mMNaCl04/1 mM phosphate buffer, pH 7.5. Unlabeled ethidiumbromide (Sigma) and [6-14C]ethidium bromide (18.3 mCi/mmol; 1 Ci = 37 GBq) (Centre Energie Atomique, France),acridine (Fluka), and 3,6-diaminoacridine (proflavine, Sigma)were used without any further purification. cis-Pt(NH3)2Cl2and trans-Pt(NH3)2Cl2 were kindly provided by J. L. Butour.The synthesis of poly(d[8-14C]G-dC) was performed as de-scribed (12).The reaction with cis-Pt(NH3)2Cl2, the nucleic acids, and

the intercalating agents was carried out as follows. Anappropriate volume of the intercalating agent was first addedto the nucleic acid solution (0.5 mM). After 5 min, a freshlymade solution of cis-Pt(NH3)2Cl2 was added and the reactionmixture was incubated at 37°C. The molar ratio of boundethidium per nucleotide was determined by the followingthree assays. (i) Filter assay: the nucleic acid of the reactionmixture was precipitated on a microfiber glass filter(Whatman) by a 5% solution of trichloroacetic acid (Merck).The filter was washed and then dried, and the radioactivity of[14C]ethidium bound to the nucleic acid was counted with anLKB 1216 counter. (ii) TLC was run on plates of Silica gel 60F 254 (Merck) with the solvent system isopropanol/ammo-nia/water, 6:3:1 (vol/vol). The radioactivity of the differentspots was then determined. (iii) Butanol extraction: the saltconcentration of the reaction mixture was adjusted to 0.2 M

Abbreviations: cis- and trans-Pt(NH3)2C12, cis- and trans-diam-minedichloroplatinum(II); rb, molar ratio drug/nucleotide.

6317

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

May

4, 2

020

6318 Biochemistry: Malinge and Leng

NaClO4. The mixture was extracted four times with water-saturated 1-butanol and then dialyzed against 0.2 MNaClO4/1 mM phosphate buffer, pH 7.5, for 2 hr in the cold.The dye concentration was determined either by radioactiv-ity or by absorption. The absorption spectra were recordedon a Kontron Uvikon 810 spectrophotometer. It has beenassumed that the extinction coefficients of the dyes interca-lated in nucleic acids and in the ternary complexes, nucleicacid-cis-Pt(NH3)2-dye, were the same. The Pt contents of thesamples were determined with an atomic absorption spec-trophotometer as described (12).

Competition Experiments. Poly(d[8-14C]G-dC), variousdouble-stranded polydeoxyribonucleotides, and cis-Pt-(NH3)2C12 were incubated at 370C for 24 hr in the absence orin the presence of intercalating agents. The reaction mixtureswere extracted with butanol as described above and thenincubated for 75 hr at 370C. From time to time, the solutionswere extracted with butanol. After dialysis of the solutions,the amount of Pt bound to poly(d[8-14C]G-dC) was deter-mined by TLC after enzymatic hydrolysis (DNase I, nucleaseS1, and alkaline phosphatase) (12-17).

RESULTSReaction in the Presence of Ethidium Bromide. It is known

that cis-Pt(NH3)2Cl2 binds to poly(dG-dC) in the B confor-mation. The main adduct in this platinated polymer, namedpoly(dG-dC)-cis-Pt(NH3)2, arises from a cross-link betweenthe N-7 of two guanine residues (6, 17). The stability of thecomplex and the nature of the adducts are quite differentwhen platination ofpoly(dG-dC) is carried out in the presenceof ethidium bromide.Formation of the Ternary Complexes. To a solution of

poly(dG-dC) was first added ethidium bromide; cis-Pt(NH3)2C12 was added 5 min later and then the reactionmixture was incubated at 37°C. After a few hours of incuba-tion, some ethidium was tightly bound to poly(dG-dC).Tightly bound ethidium is defined as ethidium that cannot beremoved from poly(dG-dC), as judged by extraction withbutanol, by filter assay at acid pH, or by TLC at basic pH.Even in more drastic conditions, the molar ratio rb (tightlybound ethidium per nucleotide) was unchanged, as deter-mined by the filter assay on the reaction mixture or on thereaction mixture first diluted 10 times with 2 M NaClO4 andthen extracted 3 times with butanol. As shown in Fig. 1, rbincreased as a function of the time of incubation and after 16hr remained constant over a long period of time. Within theexperimental error, rb was the same when the salt concen-tration of the reaction mixture was 5 or 50 mM NaClO4.Similar results were obtained when poly(dG-dC) was re-placed by poly(dG)-poly(dC), poly(dA-dC)poly(dG-dT), ornatural DNA (not shown). If cis-Pt(NH3)2Cl2 was replaced bytrans-Pt(NH3)2Cl2, all of the added ethidium bromide wasremoved (Fig. 1).

After incubation of ethidium bromide and poly(dG-dC) orpoly(dG-dC)-cis-Pt(NH3)2 at 37°C for 16 hr, no tightly boundethidium was found, as judged by the three assays. More-over, in the absence of poly(dG-dC), no reaction betweenethidium bromide and cis-Pt(NH3)2Cl2 was detected by TLCor by absorption.As shown in Fig. 2, for a given quantity of cis-Pt(NH3)2Cl2,

the amount of tightly bound ethidium depends upon theamount of added ethidium bromide (rb was determined after20 hr of incubation ofthe reaction mixture by filter assay). Onthe other hand, the amount of Pt bound to poly(dG-dC,) wasalmost the same in the absence or in the presence of ethidiumbromide. Similar results were obtained when poly(dG-dC)was replaced by poly(dG)-poly(dC), poly(dA-dC)-poly(dG-dT), or natural DNA. After incubation of the reactionmixtures at 37°C for 20 hr, rb (ethidium) and rb (Pt) were

Time, hr

FIG. 1. Kinetics of binding of ethidium bromide (EtdBr) topoly(dG-dC) in the presence of cis-Pt(NH3)2Cl2 (e) or trans-Pt(NH3)2CI2 (i). Solvent, 5 mM NaClO4/lmM phosphate buffer, pH7.5; temperature, 37°C. The input molar ratios cis-Pt(NH3)2Cl2,trans-Pt(NH3)2C12, and ethidium bromide per nucleotide = 0.08,0.08, and 0.30, respectively. The amount of tightly bound ethidiumper nucleotide, rb, was determined by filter assay.

almost the same in all of these complexes for input ratiosequal to 0.3 (ethidium bromide) and <0.1 [cis-Pt(NH3)2Cl2],respectively.

Stability of the Ternary Complex. The ternary complexpoly(dG-dC)-cis-Pt(NH3)2-ethidium obtained after incuba-tion of poly(dG-dC), cis-Pt(NH3)2Cl2, and ethidium bromidereaction mixture at 37°C for 20 hr was freed of non-tightlybound ethidium by butanol extraction in the cold and itsstability was studied under various experimental conditions.At 4°C, rb (tightly bound ethidium) remained almost constantover a long period of time (Fig. 3). At 37°C, rb decreased andthen remained constant. If, at various times, the solution wasextracted with butanol, rb became <0.01 after about 100 hr ofincubation. On the other hand, the amount of Pt bound topoly(dG-dC) remained constant as a function of time (Fig. 3).In the presence of 100 mM thiourea, the ternary complex waseven less stable. After 1 hr of incubation at 37°C, almost allthe ethidium molecules can be removed.Ethidium Exchange. The tightly bound ethidium in the

ternary complex poly(dG-dC)-cis-Pt(NH3)2-ethidium can ex-

0.08 _=_ .0o80~~~~~

0.04 0.04

0.06 0.12 0.18 0.24 0.30EtdBr/nucleotide

FIG. 2. rb, molar ratio of tightly bound ethidium (o) or boundcis-Pt(NH3)2 (A) per nucleotide residue of poly(dG-dC) as a functionof the input molar ratio ethidium bromide (EtdBr) per nucleotideresidue. The input molar ratio, cis-Pt(NH3)2C12 per nucleotide, =0.08. The reaction mixture was incubated at 37°C for 24 hr. Solvent,5mM NaClO4/1 mM phosphate buffer, pH 7.5. The values of rb weredetermined by filter assay.

Proc. Natl. Acad. Sci. USA 83 (1986)

Dow

nloa

ded

by g

uest

on

May

4, 2

020

Proc. Natl. Acad. Sci. USA 83 (1986) 6319

e 0.035 - 0.035

0 50 100 150Time, hr

FIG. 3. Kinetics of release of tightly bound ethidium to poly(dG-dC) at 4°C (C) and at 370C (o, +, o). The squares (a) are relative tothe amount of bound cis-Pt(NH3)2. Solvent, 10 mM NaCI04/1 mMphosphate buffer, pH 7.5. The solutions were extracted at varioustimes with cold butanol (o) or were not extracted (e). +, Unlabeledethidium bromide was added [input ratio ethidium bromide(EtdBr)/nucleotide = 0.3]. The values of rb were determined by filterassay and by visible absorption.

change with non-tightly bound ethidium bromide, as suggest-ed by the results shown in Fig. 1. The exchange wasdemonstrated by the following experiment. Poly(dG-dC),[14C]ethidium bromide, and cis-Pt(NH3)2Cl2 were incubatedat 37°C for 20 hr. The non-tightly bound [14C]ethidium wasremoved by extraction with butanol in the cold and then wasreplaced by unlabeled ethidium bromide. At 37°C, theamount of tightly bound [14C]ethidium decreased (Fig. 3) butthe total amount of tightly bound labeled and unlabeledethidium remained constant, as determined by absorptionafter butanol extraction. It took several hours for the totalexchange.

Pt Exchange. There was no exchange of cis-Pt(NH3)2bound to the ternary complex. After 20 hr of incubation ofpoly(d['4C]G-dC), ethidium bromide, and cis-Pt(NH3)2Cl2(the input molar ratios cis-Pt(NH3)2Cl2/nucleotide andethidium/nucleotide = 0.08 and 0.30, respectively), a 2-foldexcess of poly(dG-dC) and ethidium bromide (input ratioethidium/nucleotide = 0.30) was added and then the solutionwas incubated at 37°C for 24 hr. The amount of cis-Pt(NH3)2bound to poly(d['4C]G-dC) (determined by enzymatic hydrol-ysis of the polymer after removal of the tightly boundethidium) before and after the addition of poly(dG-dC) wasthe same.

Characterization of Platinated Poly(dG-dC). Some proper-ties of poly(dG-dC)-cis-Pt(NH3)2 and of platinated poly(dG-dC) after complete removal of the tightly bound ethidiumfrom poly(dG-dC)-cis-Pt(NH3)2-ethidium complex have beencompared. Both platinated samples behaved similarly, asjudged by circular dichroism [the circular dichroism spectraof poly(dG-dC) and poly(dG-dC)-cis-Pt(NH3)2 (rb = 0.05) aredifferent (17-20)], by the binding to antibodies to Z-DNA[poly(dG-dC)-cis-Pt(NH3)2 is well recognized by the antibod-ies to Z-DNA (20)], and by the nature of the adducts(determined after enzymatic hydrolysis), mainly (deoxygua-nosine)2-cis-Pt(NH3)2 (6, 16) (results not shown).

Reaction in the Presence of Proflavine. Only a few experi-ments have been performed with proflavine. In the firstapproximation, proflavine behaves like ethidium bromide.After 20 hr of incubation of proflavine, cis-Pt(NH3)2Cl2 andpoly(dG-dC) or poly(dG)-poly(dC) or poly(dA-dC)poly(dG-dT) or natural DNA, some proflavine molecules were tightlybound to the nucleic acid. All of the proflavine was removedwhen there was no cis-Pt(NH3)2Cl2. At 37°C, the ternary

complex nucleic acid-cis-Pt(NH3)2-proflavine was relativelyunstable and behaved like the ternary complex nucleicacid-cis-Pt(NH3)2-ethidium. There was no tightly boundproflavine if cis-Pt(NH3)2Cl2 was replaced by trans-Pt(NH3)2CI2. Finally, in the absence of nucleic acid, noreaction of cis-Pt(NH3)2C12 and proflavine was detected byabsorption or TLC.

Reaction in the Presence of Acridine. After incubation ofpoly(dG-dC) or natural DNA with acridine and cis-Pt(NH3)2C12 at 370C for 20 hr (same experimental conditionsas in Fig. 1) followed by extraction with butanol and dialysis,there was no tightly bound acridine but some Pt was boundto the nucleic acid. Thus, acridine and proflavine behavequite differently.

Several experiments were performed with platinatedpoly(dG-dC) after removal of acridine by extraction withbutanol and dialysis in the cold. In an initial experiment, theplatinated poly(dG-dC) was incubated at 370C for 20 hr andthen ethidium bromide or proflavine were added. No tightlybound dye was found. In a second experiment, the platinatedpoly(dG-dC) was mixed with proflavine or ethidium bromideat 40C (input ratio dye/nucleotide = 0.3) and then incubatedat 37TC. The amount of tightly bound dye increased as afunction of time and then remained constant. A plateau wasreached only after 7 hr (not shown).

Acridine stabilizes an intermediate form in the ternarycomplex poly(dG-dC)-cis-Pt(NH3)2-ethidium. This complexwas incubated at 370C in the presence of acridine (input molarratio acridine/nucleotide = 0.3). As shown in Fig. 4, rb(tightly bound ethidium) decreased as a function of time.After 20 hr of incubation, acridine was extracted with coldbutanol, ethidium bromide was added (input molar ratio =0.3), and the solution was incubated at 37°C. The value of rbincreased and, after several hours of incubation, reached avalue slightly smaller than that at the beginning of theexperiment.

Preferential Binding of cis-Pt(NH3)2CI2. Competition exper-iments between cis-Pt(NH3)2Cl2 and several double-strandedpolydeoxynucleotides have been performed. In all of theseexperiments poly(d['4C]G-dC) was taken as reference. Thesame amount of cis-Pt(NH3)2Cl2 was added on the one handto poly(d[14C]G-dC) and on the other hand to an equimolarmixture of poly(d['4C]G-dC) and poly(dG)-poly(dC) orpoly(dA-dT) in the absence or in the presence, respectively,

0.06

e 0.03 -

8 16 24 32 40Time, hr

FIG. 4. Kinetics of release of tightly bound ethidium to poly(dG-dC) in the presence of acridine. The complex was incubated at 370Cin the presence of acridine (input ratio acridine/nucleotide = 0.3).After 20 hr, the solution was extracted with cold butanol and then[14C]ethidium bromide was added [input ratio ethidium bromide(EtdBr)/nucleotide = 0.25]. Solvent, 5 mM NaClO4/1 mM phos-phate buffer, pH 7.5. The filter assay was used for determination ofrb.

Biochemistry: Malinge and Leng

Dow

nloa

ded

by g

uest

on

May

4, 2

020

6320 Biochemistry: Malinge and Leng

of dye (ethidium bromide, proflavine, or acridine). Theamount of cis-Pt(NH3)2C12 bound to poly(d["4C]G-dC) was

determined by TLC after enzymatic hydrolysis of thepolynucleotide, the tightly bound dye having been firstremoved by incubation at 37°C for 80 hr. The results are

summarized in Table 1 (the percentages are known within±5%). In the absence of dye, cis-Pt(NH3)2C12 bound prefer-entially to poly(dG)-poly(dC) as compared to poly(dG-dC)and preferentially to poly(dG-dC) as compared to poly(dA-dT). In the presence of dye, cis-Pt(NH3)2C12 bound equallywell to poly(dG-dC) and to poly(dG)-poly(dC). In the com-

petition experiments with poly(dA-dT) and ethidium bro-mide, all of the cis-Pt(NH3)2C12 bound to poly(dG-dC).

DISCUSSIONThe reaction of cis-Pt(NH3)2CI2 with nucleic acids proceedsthrough loss of chloride ions to form hydrolysis products andthe limiting step is the extent of hydrolysis (1-5). In the invitro reaction with nucleic acids the most frequent lesions are

a cross-link between two bases and only a small proportionof monofunctional adducts has been detected, mainly forshort incubation times (6-11). In platinated DNA, the lesionsprevent the binding of ethidium bromide (21).

After incubation of poly(dG-dC), cis-Pt(NH3)2C12, andethidium bromide for about 16 hr at 37°C, some ethidiummolecules form a very stable complex with poly(dG-dC) andcannot be removed by extraction with butanol, by filter assayat acid pH, by TLC at basic pH, or even by a combination ofextraction with butanol and filter assay. The molar ratio rb

(tightly bound ethidium per nucleotide) depends upon theethidium bromide and cis-Pt(NH3)2C12 per nucleotide inputratios. For ratios = 0.3 and 0.08, respectively, almost all ofthe added cis-Pt(NH3)2CI2 is bound to poly(dG-dC), andbound cis-Pt(NH3)2 and tightly bound ethidium molaramounts are about the same.

After incubation of ethidium bromide and poly(dG-dC) or

poly(dG-dC)-cis-Pt(NH3)2, all of the dye can be removed bythe three assays. Thus, the ternary complexes formed in thereaction of cis-Pt(NH3)2CI2 and poly(dG-dC) in the presenceof ethidium bromide are different from those formed betweenplatinated or unplatinated poly(dG-dC) and ethidium bro-mide. Similar conclusions were obtained from the study ofternary complexes in which poly(dG-dC) was replaced bypoly(dG)-poly(dC), poly(dA-dC)-poly(dG-dT), and naturalDNA. There was no tightly bound ethidium when cis-Pt(NH3)2CI2 was replaced by trans-Pt(NH3)2Cl2.The stability of the ternary complex poly(dG-dC)-cis-

Pt(NH3)2-ethidium (all of the non-tightly bound ethidiumhaving been removed by extraction with cold butanol) hasbeen studied at two temperatures. At 4°C, only a smallamount of ethidium is released with time. At 37°C, almost allthe tightly bound ethidium is released after about 100 hr,

whereas the amount of bound Pt remains constant. Aftercomplete loss of the tightly bound ethidium, the platinatedpoly(dG-dC) has all of the characteristics of poly(dG-dC)-cis-Pt(NH3)2, as judged by circular dichroism, by the bindingto the antibodies to Z-DNA, and by the nature of the adducts.The ternary complex poly(dG-dC)-cis-Pt(NH3)2-ethidium

is even less stable in the presence ofthiourea, which is knownto have a high affinity for Pt (22). Previous results have shownthat the presence of ethidium during the platination ofDNAmakes easier the removal of Pt by KCN (16).

Tightly bound ethidium in the ternary complex poly(dG-dC)-cis-Pt(NH3)2-[14C]ethidium can exchange with freeethidium. On the other hand, there is no exchange of thebound Pt even in the presence of an excess of poly(dG-dC).These results can be summarized as follows:

Poly(dG-dC) + cis-Pt(NH3)2C12 -poly(dG-dC)-cis-Pt(NH3)2.

Poly(dG-dC) + cis-Pt(NH3)2C12 + ethidium -poly(dG-dC)-cis-Pt(NH3)2-ethidium.

Poly(dG-dC)-cis-Pt(NH3)2-ethidium -

poly(dG-dC)-cis-Pt*(NH3)2 + ethidium.

Poly(dG-dC)-cis-Pt*(NH3)2 -7

poly(dG-dC)-cis-Pt(NH3)2.

[1]

[2]

[3]

[4]

In poly(dG-dC)-cis-Pt(NH3)2 prepared according to reac-tion 1, the main adduct arises from a cross-link between twoguanine residues (6, 17). In poly(dG-dC)-cis-Pt(NH3)2-ethidium (reaction 2), ethidium is tightly bound and cis-Pt(NH3)2 is not bound to the bases as in poly(dG-dC)-cis-Pt(NH3)2. This is suggested by the reaction with thiourea andbecause there is no tightly bound ethidium when the dye isadded to an already formed poly(dG-dC)-cis-Pt(NH3)2. More-over, in poly(dG-dC)-cis-Pt(NH3)2-ethidium there is aboutone ethidium per platinum. Tightly bound ethidium can

exchange with non-tightly bound ethidium (reaction 3). Theexchange is slow (several hours) and the intermediate com-

pound is poly(dG-dC)-cis-Pt*(NH3)2. Incubation of poly(dG-dC)-cis-Pt*(NH3)2 in the absence of ethidium leads topoly(dG-dC)-cis-Pt(NH3)2 (reaction 4).Some experiments have been performed in the presence of

proflavine and acridine. Proflavine behaves like ethidiumbromide. By contrast, after incubation of acridine, cis-Pt(NH3)2C12, and poly(dG-dC) (or natural DNA) reactionmixture at 37°C for 16 hr, no tightly bound acridine has beendetected, although cis-Pt(NH3)2C12 reacts with the nucleicacid. After removal of the dye and further incubation, theplatinated poly(dG-dC) behaves as poly(dG-dC)-cis-Pt(NH3)2. If the incubation is carried out in the presence ofethidium bromide, some ethidium is tightly bound. Thus,

Table 1. Competition experiments between cis-Pt(NH3)2Cl2, poly(dG-dC), and poly(dG)-poly(dC)or poly(dA-dT) in the absence or in the presence of ethidium bromide, proflavine, oracridine, respectively

% bound Pt

To poly(dG) poly(dC)To poly(dG-dC) or poly(dA-dT)

Poly(dG-dC), poly(dG)-poly(dC) 23 77Poly(dG-dC), poly(dG)-poly(dC), acridine 42 58Poly(dG-dC), poly(dG)-poly(dC), proflavine 43 57Poly(dG-dC), poly(dG)poly(dC), ethidium bromide 60 40Poly(dG-dC), poly(dA-dT) 80 20Poly(dG-dC), poly(dA-dT), ethidium bromide 100 0

In all experiments, the input molar ratio cis-Pt(NH3)2Cl2 per nucleotide = 0.08 and the input molarratio ethidium bromide, proflavine, or acridine per nucleotide = 0.30.

Proc. Natl. Acad. Sci. USA 83 (1986)

Dow

nloa

ded

by g

uest

on

May

4, 2

020

Biochemistry: Malinge and Leng

acridine can stabilize an intermediate form of platinatedpoly(dG-dC) that is able to bind ethidium tightly. This is alsoshown by the study of poly(dG-dC)-cis-Pt(NH3)2-ethidium inthe presence of acridine. At 370C, rb (tightly bound ethidium)decreases as a function oftime. After removal of acridine andof released ethidium by butanol extraction in the cold andthen addition of ethidium bromide, one again gets a ternarycomplex poly(dG-dC)-cis-Pt(NH3)2-ethidium in which theamount of tightly bound ethidium is almost equal to that at thebeginning of the experiment. These results argue stronglythat this intermediate form and poly(dG-dC)-cis-Pt*(NH3)2(reaction 3) are identical.At this point, it is necessary to discuss how cis-Pt(NI3)2C12

binds to poly(dG-dC) in the presence of the dyes. In thepresence of ethidium or proflavine, it can be excluded thatcis-Pt(NH3)2 cross-links two guanine residues, as alreadydiscussed. On the other hand, after complete removal of thetightly bound ethidium, the platinated poly(dG-dC) behaveslike poly(dG-dC)-cis-Pt(NH3)2. Moreover, the Pt residues donot exchange in the ternary complexes. These results suggeststrongly that in the presence of the dyes, one function of Ptis bound to N-7 ofguanine residues. In the ternary complexespoly(dG-dC)-cis-Pt(NH3)2-ethidium or -proflavine the molaramounts of tightly bound dye and Pt are almost equal. Thissuggests that Pt is involved in the binding site of the tightlybound dye. These platinated sites interact quite differentlywith the two related molecules acridine and proflavine. Theassociation and dissociation rates of ethidium (or proflavine)and platinated binding sites are very slow. It is proposed thata bidentate adduct (guanine-ethidium or -proflavine) isformed and that the slow step of the reaction is the bindingof Pt-guanine adduct to the dye. In this reaction, the nucleicacid can be considered as a matrix achieving a favorableorientation ofthe reactants. In the absence of nucleic acid, noreaction between cis-Pt(NH3)2Cl2 and the dye has beendetected (this work; refs. 15, 16).

Several results strongly suggest that cis-Pt(NH3)2Cl2 bindstoDNA in a sequence-dependent manner (7, 9, 17, 23-25). Bymeans of exonuclease III, new cis-Pt(NH3)2Cl2 binding siteshave been detected by platination treatment of DNA in thepresence of ethidium (15, 16). By competition experimentsbetween various double-stranded polynucleotides and cis-Pt(NH3)2CI2, in the presence of ethidium bromide, profla-vine, or acridine, we show that the dyes interfere with thepreferential binding of cis-Pt(NH3)2Cl2 to (dG)"-(dC)" se-quences. For example, in the absence of ethidium bromide,more cis-Pt(NH3)2C12 binds to poly(dG)-poly(dC) than topoly(dG-dC) and more binds to poly(dG-dC) than to poly(dA-dT). In the presence of ethidium bromide about the sameamount of Pt is now bound to poly(dG-dC) and topoly(dG)-poly(dC). In the competition experiment betweenpoly(dG-dC) and poly(dA-dT), there is no Pt bound topoly(dA-dT). It is possible that in vivo intercalating agentsredirect cis-Pt(NH3)2Cl2, which might increase the therapeu-tic efficiency of cis-Pt(NH3)2C12.

In conclusion, in the reaction of cis-Pt(NH3)2Cl2 andnucleic acids, intercalating dyes such as ethidium bromide,proflavine, and acridine prevent cis-Pt(NH3)2Cl2 from bind-ing preferentially to (dG),-(dC). sequences. The stability ofthe ternary complexes nucleic acid-cis-Pt(NH3)2-dye de-

Proc. Natl. Acad. Sci. USA 83 (1986) 6321

pends strongly upon the nature of the dye. More generally,since many compounds intercalate into DNA, one expectslarge differences not only in the distribution of bound Pt butalso in the interactions between bound Pt and the intercalatedcompound as a function of the chemical nature of theintercalating compound, the base sequence, and the confor-mation of DNA. The study of the ternary complexes mighthelp to explain the synergism for drugs used in combinationwith cis-Pt(NH3)2Cl2 and in the design of new chemothera-peutic agents.

We are indebted to Dr. B. Malfoy and to Dr. J. Ramstein for helpfuldiscussion. We thank Dr. J. L. Butour and Dr. A. Gouyette foranalysis of platinated samples.

1. Prestayko, A. W., Crooke, S. T. & Carter, S. K., eds. (1980)Cisplatin: Current Status and New Developments (Academic,New York).

2. Roberts, J. J. & Thomson, A. J. (1979) Proc. Nucleic AcidRes. Mol. Biol. 22, 71-133.

3. Lippard, S. J. (1982) Science (Washington, DC) 218,1075-1082.

4. Macquet, J. P., Butour, J. L., Johnson, N. P., Razaka, H.,Salles, B., Vieussens, C. & Wright, M. (1984) in PlatinumCoordination Complexes in Cancer Chemotherapy, eds.Hacker, M. P., Douple, E. B. & Krakoff, I. H. (Nijhoff,Boston), pp. 29-38.

5. Pinto, A. L. & Lippard, S. J. (1985) Biochim. Biophys. Acta780, 167-180.

6. Eastman, A. (1983) Biochemistry 22, 3927-3933.7. Fichtinger-Schepman, A. M. J., van der Veer, J. L., den

Hartog, J. H. J., Lohman, P. H. M. & Reedijk, J. (1985)Biochemistry 24, 707-713.

8. Eastman, A. (1985) Biochemistry 24, 5027-5032.9. Royer-Pokora, B., Gordon, L. K. & Haseltine, W. A. (1981)

Nucleic Acids Res. 9, 4595-4609.10. Pinto, A. L. & Lippard, S. J. (1985) Proc. Natl. Acad. Sci.

USA 82, 4616-4619.11. Johnson, N. P., Mazard, A. M., Escalier, J. & Macquet, J. P.

(1985) J. Am. Chem. Soc. 107, 6376-6380.12. Malinge, J. M. & Leng, M. (1984) EMBO J. 3, 1273-1279.13. Rich, A., Nordheim, A. & Wang, A. H. (1984) Annu. Rev.

Biochem. 53, 791-846.14. Leng, M. (1985) Biochim. Biophys. Acta 825, 339-344.15. Tullius, T. D. & Lippard, S. J. (1982) Proc. NatI. Acad. Sci.

USA 79, 3489-3492.16. Merkel, C. M. & Lippard, S. J. (1982) Cold Spring Harbor

Symp. Quant. Biol. 47, 355-360.17. Rahmouni, A., Malinge, J. M., Schwartz, A. & Leng, M.

(1985) J. Biomol. Struct. Dyn. 3, 363-375.18. Malfoy, B., Hartmann, B. & Leng, M. (1981) Nucleic Acids

Res. 9, 5659-5669.19. Ushay, H. M., Santella, R. M., Caradonna, J. P., Grunberger,

D. & Lippard, S. J. (1982) Nucleic Acids Res. 10, 3573-3588.20. Malinge, J. M., Ptak, M. & Leng, M. (1984) Nucleic Acids

Res. 12, 5767-5778.21. Macquet, J. P. & Butour, J. L. (1978) Biochimie 60, 901-914.22. Micetich, K., Zwelling, L. A. & Kohn, K. W. (1983) Cancer

Res. 43, 3609-3613.23. Stone, P. J., Kelman, A. D. & Sinex, F. M. (1974) Nature

(London) 251, 736-737.24. Stone, P. J., Kelman, A. D., Sinex, F. M., Bhargava, M. M.

& Halvorson, H. 0. (1976) J. Mol. Biol. 104, 793-801.25. Cohen, G. L., Ledner, J. A., Bauer, W. R., Ushay, H. M.,

Caravana, C. & Lippard, S. J. (1980) J. Am. Chem. Soc. 102,2487-2488.

Dow

nloa

ded

by g

uest

on

May

4, 2

020