ORIGINAL ARTICLE

Johan Kjellstrom Æ Sofi K.C. Elmroth

Kinetics and mechanism for platination of thione-containingnucleotides and oligonucleotides: evaluation of the salt dependence

Received: 28 November 2001 /Accepted: 23 May 2002 / Published online: 14 August 2002� SBIC 2002

Abstract Reactions of cis-[PtCl(NH3)(CyNH2)(OH2)]+

(Cy=cyclohexyl) with thione-containing single-strandedoligonucleotides d(T8XT8) and d(XT16) (X=s6I or s4U)and the mononucleotides 4-thiouridine (s4UMP) and 6-mercaptoinosine (s6IMP) have been studied in aqueoussolution at pH 4.1. The reaction kinetics was followedusing HPLC methodology as a function of ionic strengthin the interval 5.0 mM £ I £ 300 mM. A two-fold ki-netic preference for reaction with the s6I moiety over s4Uis observed in both monomeric and oligomeric systems.The rate for adduct formation with the oligonucleotidesd(T8XT8) and d(XT16) decreases with increasing ionicstrength of the medium. The effect is most pronouncedfor adduct formation with the middle positions, e.g. ford(T8

s6IT8): k2,app=370 M–1 s–1 and 13 M–1 s–1 at I=5.0and 300 mM, respectively, and slightly less pronouncedfor adduct formation at the end positions, e.g. ford(s6IT16): k2,app=130 M–1 s–1 and 11 M–1 s–1 at I=5.0and 300 mM, respectively. Analysis of the salt depen-dence using the Brønsted-Debye-Huckel relationshipshows that the reactions with the monomers are welldescribed as an interaction between a monovalent cationand a monovalent anion. In contrast, a similar analysisof the oligonucleotide reactions indicates influence frompolyelectrolyte effects. The results support a mechanismin which pre-association on the DNA surface precedesadduct formation, regardless of the exact location of thefinal binding site.Electronic supplementary material to this paper, com-prising Figs. S1–S6 and Tables S1–S3, can be obtained

by using the Springer Link server located at http://dx.doi.org/10.1007/s00775-002-0384-9.

Keywords Platinum Æ DNA Æ Kinetics Æ6-Mercaptoinosine Æ 4-Thiouridine

Introduction

Coordination of square-planar platinum(II) complexessuch as cisplatin (cis-[PtCl2(NH3)2]) and cis-[PtCl2(NH3)(CyNH2)] (JM118) (Cy=cyclohexyl) tonative DNA preferentially takes place at purine-rich re-gions [1, 2, 3, 4]. The initial step of the reaction, typicallythe formation of a monofunctional guanine-N7 adduct,is likely an effect mainly governed by electrostatic in-teractions [5, 6, 7]. In contrast, the subsequent formationof intra- and inter-strand bifunctional adducts relies on acombination of favourable steric and electrostatic in-teractions during formation of the second covalent bond,giving rise to a molecule with strong geometrical con-straints [8, 9, 10, 11]. The driving force for formation ofthe common d(GpG) adduct of cisplatin is large enoughto compensate for the energy required to convert thereacting, B-type helical DNA to a bent, A/B-type DNAhybrid [8, 9]. The structural change is important for thebiological processing of DNA in cancer cells, since rec-ognition of the bent DNA structure around the plati-nation site likely diverts HMG-type proteins from theircrucial function in the rapidly growing cell [10, 11, 12,13]. However, current lack of potent alternatives to cis-platin illustrates the need for a better understanding ofhow the delicate balance between electrostatic and stericinteractions may be optimized for construction of apharmaceutically active drug.

A focus for on-going work in our laboratory is toincrease the knowledge concerning the influence of weakinteractions on the rate and mechanism for adduct for-mation between ribonucleic acid surfaces and chargedmetal complexes. Previous studies by us have shown thatthe kinetics for adduct formation reactions with donor

J Biol Inorg Chem (2003) 8: 38–44DOI 10.1007/s00775-002-0384-9

Electronic supplementary material to this paper, comprisingFigs. S1–S6 and Tables S1–S3, can be obtained by using theSpringer Link server located at http://dx.doi.org/10.1007/s00775-002-0384-9

J. Kjellstrom Æ S.K.C. Elmroth (&)Inorganic Chemistry, Chemical Center,Lund University, P.O. Box 124, 221 00 Lund, SwedenE-mail: sofi.elmroth@inorg.lu.seTel.: +46-46-2228106Fax: +46-46-2224439

groups located on the charged phosphodiester backboneare sensitive towards factors such as (1) the size of thereacting DNA oligomer [7, 14], (2) the exact location ofthe binding site [15, 16] and (3) the choice and concen-tration of counterions used in the supporting electrolyte[7, 15, 16]. The kinetics data could be fitted to a reactionmodel in which pre-association of the platinum complexprecedes the adduct formation reaction. The salt de-pendence suggested the magnitude of the electrostaticinteractions to be dependent mainly on the size of theoligomer, and thus rather insensitive towards the exactlocation of the binding site. Further, the reduced reac-tion rates and magnitude of the electrostatic interactionsobserved after a change of Na+ for Mg2+ in the sup-porting electrolyte supports a mechanism where thecharged platinum complex replaces a cation in thecondensation layer surrounding the oligomers. The in-fluence on kinetics from exchange of surface bound bulkcations for the platinum complex seems reasonable,considering the close proximity of the electrostaticallypreferred interaction sites along the phosphodiesterbackbone and the reactive phosphorothioate group [5, 6,17, 18, 19]. The present study has been designed to ob-tain information on how the kinetics is influenced by atranslocation of the reactive group from the chargedbackbone to the non-charged bases. The adduct for-mation reaction with the DNA surface was studied byuse of the metal complex cis-[PtCl(NH3)(CyN-H2)(OH2)]

+ and single-stranded, 17-mer poly d(T)oligonucleotides containing a single thione moiety,i.e. 4-thiouridine or 6-mercaptoinosine. The platinumcomplex is considered to be the active metabolite ofthe orally administered anticancer active compoundcis,trans,cis-[PtCl2(OC(O)Me)2(NH3)(CyNH2)] (JM216)[20, 21, 22]. Both thiones have functions in vivo: the 4-thiouridie moiety, d(s4U), is naturally occurring at theeighth position of Escherichia coli tRNA [23], and 6-mercaptoinosine, d(s6I), can be used as a protein inhib-itor during treatment of acute leukemia [24, 25, 26, 27].The presence of the soft sulfur donor in these bases fa-cilitates the analysis by providing binding sites that arekinetically preferred over the common DNA bases [28,29, 30]. In addition, their location is similar to that ofguanine-N7 with respect to the phosphodiester backbone[31]. The present results give support for a common re-action model for formation of adducts on the DNAsurface, regardless of the exact location of the bindingsite, involving electrostatically driven pre-association of

the reacting cation on the surface prior to the rate-determining formation of the covalent bond.

Materials and methods

Supplies and reagents

Buffer solutions were prepared from potassium hydrogen phthalate(KHC8H4O4, Acros), and the pH was adjusted by addition ofHClO4 (Merck) [32]. The phthalate buffer was diluted 100 times forthe kinetics experiments ([K+]=[C8H4O4

–]=0.5 mM in all exper-iments). Sodium perchlorate (NaClO4, Merck) was added to adjustthe ionic strength. The pH was determined by use of a Methrom744 pH meter. Stock solutions were kept at room temperature.

The compound cis-[PtCl2(NH3)2] was purchased (Sigma) andcis-[PtCl2(NH3)(CyNH2)] (1) was synthesized according to the lit-erature [33]. The monochloro complex cis-[PtCl(NH3)(CyNH2)X]+/0 (X=DMF or NO3

–) was prepared by addition of0.98 equiv of AgNO3 (Baker) to 1 dissolved in DMF (Lab-Scan).The reaction mixture was allowed to vortex in the dark for about20 h, and the AgCl precipitate was removed by centrifugation.Stock solutions of approximately 35 mM cis-[PtCl(NH3)(CyNH2)X]+/0 (X=DMF or NO3

–) were stored in the dark at+8 �C. Addition of cis-[PtCl(NH3)(CyNH2)X]+/0 (X=DMF orNO3

–) to an aqueous solution results in rapid and quantitativeconversion to the corresponding aqua complex cis-[PtCl(NH3)(CyNH2)(OH2)]

+ (2). Previous studies have shown thatthe Cl– ligand trans to the coordinated cyclohexylamine moiety isthe most labile one [34], and the corresponding aqua complex isshown in Scheme 1. However, the other Cl– ligand is also labile,giving rise to a mixture of aqua complexes with similar reactivity inthe solution.

The monomers s4UMP, s6IMP, d(GMP) and d(s6I) (Sigma) andthe oligonucleotide d(T8GT8) (Scandinavian Gene Synthesis) wereused as received without further purification. The sulfur-containingoligonucleotides d(T8

s4UT8), d(T8s6IT8), d(s4UT16) and d(s4IT16)

were synthesized as described earlier [35, 36, 37]. Stock solutions ofmonomers and oligomers were kept frozen at –80 �C. Concentra-tions were determined spectrophotometrically using calculatedextinction coefficients [38]: �260(T17)=140,900 M–1 cm–1, �332=23,000 M–1 cm–1 and �322=23,100 M–1 cm–1 for s4U and s6I, re-spectively [39, 40]. Spectra were recorded at 25 �C and ambientpressure using a Milton Roy 3000 diode-array or a Cary 300 BioSpectrophotometer and thermostatted 1.00 cm Quartz Suprasilcells.

1H NMR spectra

Proton NMR spectra were recorded on a Varian Unity 300 spec-trometer working at 299.79 MHz. The inversion-recovery tech-nique was used to determine the longitudinal relaxation time, T1,for the HDO protons. The deuterated solvents D2O (Acros) andDMF-d7 (Cambridge Isotope Laboratories) were used as received.Internal HDO and DMF signals were used as references. The re-actions were studied under second-order conditions with typical

Scheme 1

39

reactant concentrations of 5–10 mM. Samples were prepared bydissolution of the appropriate d(GMP), s6IMP or s4UMP in a5 mm NMR tube (Glaser, Basel) and adding an appropriateamount of D2O. Reactions were initiated by addition of dissolved 1or cis-[PtCl2(NH3)2] in DMF-d7 to the mononucleotide solution.

Circular dichroism spectroscopy

Spectra of d(T8GT8) as a function of increasing [Na+] were col-lected from 350 to 245 nm using a Jasco J-500A spectropolarimeterat ambient temperature. An aqueous stock solution with[d(T8GT8)]=4.9 lM was prepared in phthalate buffer (pH 4) with[Na+, K+]=1.0 mM. For measurements, 2.0 mL of the stock so-lution were placed in the cylindrical quartz cuvette (5 mm pathlength, Hellma). The concentration of Na+ was stepwise increasedfrom 1.0 mM to 300 mM by removal of a small aliquot of oli-gonucleotide solution in the cuvette and replacing it by an equalamount of 1.0 M NaClO4. The resulting decrease in oligonucleo-tide concentration was taken into account for calculation of themolar ellipticity (D�). The rotation was observed in millidegrees (H)and was converted to D� by use of the relations MH=100H/CL andD�=MH/3298, where C and L denote the molar concentration ofd(T8GT8) and the path length in mm, respectively [41, 42].

HPLC measurements

The chromatograms were obtained by use of a LaChrom chro-matograph system (Merck-Hitachi, working under MicrosoftWindows 3.51) with a D-7000 interface and a D-7400 UV/visspectrometer at a constant temperature of 25.0±0.2 �C (L-7350oven and cooling module, Merck-Hitachi). The separation condi-tions were optimized on a reverse phase Protein & Peptide C18column (300·4.6 mm i.d., 10 lm particle diameter, Vydac) equip-ped with a guard. Solutions of 0.10 M ammonium acetate (Merck),A, adjusted to pH 6.0 with acetic acid (Merck), and a 1:1 mixtureof A and acetonitrile (Lab-Scan), B, were used as eluents. Lineargradients in which the ratio A:B changed from 84:16 to 70:30 over aperiod of 22 min, or a constant mixture of 100:0 during 6.5 minand then a linear gradient to 87:13 over a period of 10 min, weregenerated by a low-pressure gradient system at a constant flow of1 mL/min for separation of the platinated oligo- and mononucle-otides, respectively. The chromatograms were evaluated by use ofan on-line HPLC System Manager Software. The time-dependentchanges of peak areas for reactants and corresponding productswere used for the kinetics evaluation.

Kinetics

Reactions of cis-[PtCl(NH3)(CyNH2)(OH2)]+ (2) with s4UMP,

s6IMP, d(T8s4UT8), d(T8

s6IT8), d(s4UT16) and d(s6IT16) were stud-

ied in dilute phthalate buffer at pH 4.1±0.1. At this pH, the aquaform of the platinum complex dominates over the less reactivehydroxo form, and the sulfur-containing bases are considered to bein the thione form [25, 34, 43]. An approximately 40-fold excess ofplatinum complex over thiones was used to ensure pseudo-first-order reaction conditions. The platination reactions were studied asa function of ionic strength in the interval 5.0 mM £ I £ 300 mMby a variation of the concentration of NaClO4 in the supportingelectrolyte (4.5 mM £ [Na+] £ 0.30 M). The kinetics reactionswere initiated by addition of a small volume of 2 in DMF to athermostatted solution of buffered mono- or oligonucleotide solu-tion. Sample aliquots were withdrawn at appropriate time intervalsand immediately quenched by dilution in buffer. The samples werestored in liquid nitrogen at –196 �C and injected on the HPLCdirectly after thawing. Observed pseudo-first-order rate constantswere determined by a fit of a single exponential to the normalizedtime-dependent peak areas corresponding to unreacted mono- oroligonucleotide. Apparent second-order rate constants were ob-tained according to k2,app=kobs/CPt. Errors correspond to onestandard deviation.

Results

CD spectroscopy

The CD spectrum of d(T8GT8) in phthalate buffer and[Na+]=10 mM exhibits a positive band at 277 nm anda negative band at 252 nm, both features characteristicof B-form DNA (see Fig. S1) [44]. Successive addition ofNaClO4 to a final concentration of 300 mM indicatesthat the B-form structure is maintained in the whole saltinterval studied, since both the D� values and the posi-tion of the negative or positive bands remain constant(cf. inset in Fig. S1).

1H NMR spectra

Specific platination of the thione moiety was verified bya comparison of the 1H NMR spectrum obtained afterreaction of the s4U and s6I moieties with that of a similarbase carrying an exocyclic oxygen atom at the C4 or C6

position, i.e. T and G, respectively. The spectrum ofd(TMP) was studied as a function of time during 11 hafter addition of 1 to the reaction mixture, without anynotable change of the spectral characteristics (results notshown). Similarly, reaction of cis-[PtCl2(NH3)2] withd(GMP) results in only minor changes of the spectra:after 1.5 h reaction time the peak originating from C8-Hin unplatinated d(GMP) (d=8.27 ppm) has decreasedby only approximately 5% and a small product peakappears at d=8.76 ppm (results not shown). In contrast,addition of 1 to d(s4UMP) gives rise to the appearance ofnew peaks in the aromatic region (d=7.94, 7.17 and7.11 ppm, 3JH-H=7.7, 6.4 and 6.4 Hz, respectively), andthe doublets from the C6 and C5 protons decrease as afunction of time (d=8.03 and 6.68 ppm, 3JH-H=7.71and 7.68 Hz, respectively) (see Fig. 1). Similarly, thereaction of cis-[PtCl2(NH3)2] with d(s6I) results in sig-nificant changes of the spectrum, with the signals fromthe C2-H and C8-H protons (d=8.59 and 8.44 ppm,respectively) decreasing to approximately 35% of theirinitial height by 1.5 h after mixing of the reactants. Inaddition, two major (d=9.00 and 8.69 ppm) and threeminor (d=8.95, 8,82, and 8.52 ppm, Fig. S2) productpeaks appear in the aromatic region.

UV/vis spectroscopy

Platination of the thione group in the oligonucleotideswas confirmed by a study of the absorption spectrum asa function of time around the local absorption maxi-mum of d(s4U) and d(s6I) at k=332 or 322 nm, respec-tively. Representative spectral changes obtained afteraddition of 2 to d(T8

s6IT8) are shown in Fig. 2. As canbe seen, the decrease of absorbance originatingfrom unreacted thione takes place in parallel with anincrease of absorbance at k=350 nm. The presence of

40

an isosbestic point indicates a simple reaction with twoabsorbing components, in agreement with a direct re-action between the thione moiety and the platinumcomplex. Control experiments were performed to in-vestigate the stability of the thione unit in the absence ofmetal reagent. No spectral changes were observed dur-ing the time interval used for the kinetics studies (datanot shown).

Kinetics

The kinetics of the reaction of 2 with s4UMP, s6IMP,d(T8

s6IT8), d(T8s4UT8), d(

s6IT16) or d(s4UT16) were in-vestigated at pH 4.1±0.1 using HPLC methodology.The time course for the change of the integrated area ofthe reacting DNA fragment was found to obey first-order kinetics, both when evaluated at 260 nm and atthe thione-specific maximum. The direct proportionalitybetween the excess platinum concentration and the ob-served pseudo-first-order rate constant was verified bycontrol studies of the concentration dependence for re-action of 2 with s4UMP and d(s6IT16) (Fig. S3; data inTable S1). A selection of apparent second-order rateconstants is given in Table 1 (complete data in Ta-ble S2). The rate constants for adduct formation withthe s6IMP monomer are approximately two times largercompared with those obtained for reaction with s4UMPin the salt interval studied. The rate constant for adductformation in the interval 5.0 mM<I<0.10 M can becharacterized by the average k2,app=2.2±0.3 and4.6±0.5 M–1 s–1 for s4UMP and s6IMP, respectively. Asystematic decrease of the reactivity is observed as afunction of increasing salt concentration, however, and

analysis of the salt dependence by use of the Brønsted-Debye-Huckel equation1 [45] (corresponding plots aregiven in Fig. S4) results in negative slopes, as expectedfor a reaction between a cation and an anion, with val-ues close to the expected one of –1 for both monomers(Table S3).

Incorporation of the s4U and s6I moieties into the oli-gonucleotides d(T8XT8) and d(XT16) results in signifi-cantlymore rapid reactionswith these targets (see Table 1and Fig. S5). The kinetic preference for reaction with thes6I moiety is maintained in the DNA environment. For a

Fig. 2 Spectra at 1.5 min intervals for the reaction of 2 withd(T8

s6IT8) at 25.0 �C, pH 4.1 and [salt]=35 mM; [Pt(II)]=8.00·10–5 M, [DNA]=8.00·10–6 M

Fig. 11H NMR spectrum as a function of time for the reaction of

cis-[PtCl2(NH3)2] with d(s4UMP) at 25.0 �C; [Pt(II)]=7.05·10–3 Mand [s4UMP]=7.01·10–3 M

Table 1 Selected second-order rate constants for the reaction of 2with various types of ribonucleic acid fragments with Na+ as thepredominant cation

Ribonucleic acid k2,app (M–1 s–1) Ref

[Cation] (mM)

1.0 or 5.0 35 >100

s4UMP 2.5±0.3b 2.1±0.2 1.9±0.3 This works6IMP 5.1±0.6b 4.4±0.3 3.8±0.5 This workd(Tp(S)T) 1.0±0.5a 0.4±0.1 0.5±0.1 [16]d(s4UT16) 55±10b 17±10 8.1±1.0 This workd(s6IT16) 130±10b 41±8 23±4 This workd(Tp(S)T15) 30±10a 19±3 1.4±0.7 [16]d(T8

s4UT8) 150±20b 29±6 14±3 This workd(T8

s6IT8) 370±90b 79±10 38±6 This workd(T8p(S)T8) 200±40a 74±4 6±2 [16]

a[Na+]=1.0 mMb[Na+]=5.0 mM

1logk2,app=logk0+2·AZAZBI1/2/(1+I1/2)

41

given thione, adduct formation with the middle positionis kinetically preferred over reaction with the end posi-tion. The kinetic preference increases with decreasingionic strength, resulting in an approximately three-foldlarger rate constant for adduct formation with d(T8XT8)compared with d(XT16) at the lowest salt concentrationstudied, I=5.0 mM. The salt dependence for the reac-tions with the oligonucleotides is illustrated in Fig. 3 bythe reaction of d(T8

s6IT8) with 2 (complete data in Ta-ble 2, Table S3 and Fig. S6). Analysis of the data ac-cording to the Brønsted-Debye-Huckel equation(Fig. 3A) results in values of the product zAzB similar towhat has been determined previously for interactionswith phosphorothioates: compare, for example, zAzB=–4.3±0.7 and –4.8±0.5 for d(T8

s4UT8) and d(T8s6IT8),

respectively, with zAzB=–4.9±0.5 for d(T8p(S)T8) [16].Interestingly, the present data exhibit a systematic devi-ation from linear dependence, as would be expected for areaction influenced by polyelectrolyte effects [46] (videinfra).

Discussion

Monomers

Platination of the thione-containing bases s4UMP ands6IMP typically gave rise to reactions with half-lives inthe range 10 min £ t1/2 £ 30 min. As expected for thesegood nucleophiles, the second-order rate constants foradduct formation are significantly larger compared withthose found for adduct formation with N7 in purines [29,30], while exhibiting a reactivity only slightly above thatof phosphorothioates (see Table 1). Of the two thionesstudied here, s6IMP displays the higher reactivity, inagreement with previous kinetics studies for interactionwith Pd(II) [47] and Au(III) [48] centres. The higherreactivity of s6IMP might partly be explained by themore favourable conditions that can be expected for theinitial electrostatic interaction of the positively chargedplatinum complex with the negative potential around the

reactive sulfur atom. In addition, approach to the s6Imoiety is likely to take place without competition fromother groups in its vicinity. In contrast, the negativepotential around the exocyclic oxygen in s4U mightcompete with that around the sulfur atom in s4U (seeScheme 1). The ionic strength dependence clearly indi-cates that the monophosphate group influences theinteraction between 2 and the non-charged thione group,with a decreasing reactivity as a function of increasingionic strength. The common slope of about –0.7, ratherthan the expected one of –1.0, is likely a result ofthe relatively large distance between the location of thecharged phosphodiester group and the adduct site onthe bases.

Oligonucleotides

The use of the non-charged thione group as the targetfor interaction of 2 with DNA models gives rise to ki-netics characteristics similar to those already reportedfor interactions with the charged phosphodiester back-bone [16]. More specifically, both location of the targetwithin the DNA fragment and the salt concentration ofthe reaction medium have a profound influence on re-activity. Further, previous studies have shown that therate for adduct formation is strongly dependent on thepH of the solution in the interval 5 £ pH £ 7 [7]. The pHdependence is mainly a result of the pH-dependent dis-tribution between the reactive aqua complex and itscorresponding base, the unreactive hydroxo complex ofthe type cis-[PtCl(OH)LL¢], around the pKa for the co-ordinated aqua ligand [34]. The low pH chosen for thestudy, ca. pH 4, thus simplifies the interpretation of thekinetics data by restricting the reaction to the directreaction of the aqua complex with the target group onthe DNA. In addition, it minimizes the kinetics influencefrom a possible pH gradient which may arise due to ahigher concentration of oxonium ions on the oligonu-cleotide surface compared that of the bulk solution[49, 50, 51] (vide infra). The observed reactivity for the

Fig. 3 Evaluation of the saltdependence for reaction of 2with d(T8

s6IT8) according to Athe point-charge model (theBrønsted-Debye-Huckel equa-tion) and B the polyelectrolytemodel (Eq. 1)

42

corresponding reactions at pH 7 can be estimated to beat least one order of magnitude lower compared to theones determined in the present study [7].

The results obtained from the CD spectra suggestthat the time-averaged conformation is of the B-DNAtype. However, recent calculations have indicated thatthese short DNAs might exhibit a tendency for forma-tion of bent structures [19]. It might thus be possible thatthe location-dependent kinetics is a combined result of(1) a varying tendency for cation accumulation along theoligomer and (2) the formation of a transient, bentstructure facilitating interaction with the middle posi-tions [19, 52, 53, 54, 55, 56, 57]. Comparison of presentdata with those obtained for interactions with thephosphorothioate group shows clearly that binding tothe non-charged bases is less influenced by a variation oflocation of the binding site in the DNA environment.For example, at [salt]=1 mM the relative reactivity ofs6I in the monomer, d(s6IT16) and d(T8

s6IT8) is 1:25:70whereas that of reaction with the phosphorothioategroup is 1:30:200. A similar comparison of the rateconstants obtained at [salt]=35 mM results in the sameconclusion (see Table 1). These results nicely agree withthose reported recently for reaction with G-N7 and thussupport the previous assumption of the distance fromthe phosphodiester backbone as a parameter of influenceon the adduct formation process during reactions ofcations with the DNA surface [58].

The variation of k2,app with the salt concentrationwas used in an attempt to evaluate whether the adductformation reaction with the non-charged thiones can bedescribed by the Brønsted-Debye-Huckel relationship,i.e. assuming point-charge interactions [45], or is influ-enced by polyelectrolyte-like properties of the oligomers.The Brønsted-Debye-Huckel relationship predicts a lin-ear relation between logk2,app and I1/2/(1+I1/2), whereasinfluence from polyelectrolyte behaviour suggests a lin-ear relationship between logk2,app and log[Na+] ac-cording to Eq. 1:

log k2;app¼ log k0 � nW log Naþ½ � ð1Þ

where n denotes the number of ionic interactions (heren=1), and Y the fraction of condensed counterions perphosphate on the oligomer surface [46]. As can be seenby comparison of Fig. 3A and B, treatment of data ac-cording to Eq. 1 displays the better fit. The resulting

slopes, and corresponding values of Y, summarized inTable 2, are slightly larger for platination of the middleposition in d(T8

s4UT8) and d(T8s6IT8) with Y=0.74 and

Y=0.80, respectively, compared with platination of theend positions in d(s4UT16) and d(s6IT16), which can becharacterized by a common slope of 0.60±0.01. Im-portantly, these values agree nicely with the calculatedvalues for single-stranded DNA and are significantlylarger than the value expected for a polyelectrolyte-in-fluenced reaction in the absence of a pre-equilibrium(Y=0.12) [5]. Thus, present data together with previousstudies by us give strong support for a common reactionmechanism involving pre-accumulation of the metalcomplex on the DNA prior to rate-determining adductformation, an assumption that has further support byrecent studies by Burgess and co-workers [59].

Conclusions

We have shown that reactions between cationic inter-mediates of anticancer-active platinum complexes andDNA are accelerated by the presence of a chargedphosphodiester backbone also when the reactive groupis a non-charged base. The increased reactivity in theDNA environment, compared with that obtained foradduct formation with the corresponding mononucleo-tide, is significant already in small oligomeric systems.Analysis of the salt dependence according to theBrønsted-Debye-Huckel relationship shows that thevariation in reactivity can be accounted for by assuminga net charge of these 17-mer oligomers close to –4. Thereactions are best described by a mechanism assumingpolyelctrolyte-like behaviour of the oligonucleotides,involving pre-association of the cationically chargedmetal complex on the DNA surface prior to the rate-determining adduct formation step.

Acknowledgements We are grateful to Professor Robert S. Coleman(Department of Chemistry and Comprehensive Cancer Center, TheOhio State University, USA) for providing us with the thio-modifiedoligonucleotides and preliminary studies by Dr. Anna Ericson.We thank the department of Organic Chemsitry 1 at Lund Uni-versity for access to the CD spectrophotometer. Financial supportfrom the Swedish Cancer Society (3308-B00-07XCC, S.K.C.E.) isgratefully acknowledged.

References

1. Reedijk J (1999) Chem Rev 99:2499–25102. Wong E, Giandomenico CM (1999) Chem Rev 99:2451–24663. Yarema KJ, Lippard SJ, Essigmann JM (1995) Nucleic Acids

Res 23:4066–40724. Bradley LJN, Yarema KJ, Lippard SJ, Essigmann JM (1993)

Biochemistry 32:982–9885. Manning GS (1978) Q Rev Biophys II 2:179–2466. Pullman A, Pullman B (1981) Q Rev Biophys 14:289–3807. Elmroth SKC, Lippard SJ (1995) Inorg Chem 34:5234–52438. Takahara P, Rosenzweig AC, Frederick CA, Lippard SJ (1995)

Nature 377:649–6529. Takahara P, Frederick CA, Lippard SJ (1996) J Am Chem Soc

118:12309–12321

Table 2 Slopes (nY) and intercepts (logk0) after analysis of the saltdependence according to Eq. 1a

Oligonucleotide Slopes (nY) Intercepts (logk0)

d(T8s6IT8) 0.80±0.02 0.72±0.04

d(T8s4UT8) 0.74±0.05 0.44±0.07

d(s6IT16) 0.60±0.01 0.75±0.02d(s4UT16) 0.60±0.03 0.36±0.05Double-stranded DNA 0.88b –Single-stranded DNA 0.56b –

aBuffered with 0.5 mM KHC8H4O4 and pH adjusted with HClO4bCalculated values [5]

43

10. Jamieson ER, Lippard SJ (1999) Chem Rev 99:2467–249811. Cohen SM, Lippard SJ (2001) Prog Nucleic Acid Res 67:93–

13012. Toney JH, Donahue BA, Kellett PJ, Bruhn SL, Essigmann JM,

Lippard SJ (1989) Proc Natl Acad Sci USA 86:8328–833213. Hughes EN, Engelsberg BN, Billings PC (1992) J Biol Chem

267:13520–1352714. Elmroth SKC, Lippard SJ (1994) J Am Chem Soc 116:3633–

363415. Kjellstrom J, Elmroth SKC (1997) Chem Commun 1701–170216. Kjellstrom J, Elmroth SKC (1999) Inorg Chem 38:6193–619917. Anderson CF, Record MTJ (1990) Annu Rev Biophys Chem

19:423–46518. Record MTJ, Zhang W, Anderson CF (1998) Adv Protein

Chem 51:281–35319. Martınez JM, Elmroth SKC, Kloo L (2001) J Am Chem Soc

123:12279–1228920. Kelland LR, Murrer BA, Abel G, Giandomenico CM, Mistry

P, Harrap KR (1992) Cancer Res 52:822–82821. Kelland LK (2000) Expert Opin Invest Drugs 9:1372–138222. Poon GK, Raynaud FI, Mistry P, Odell DE, Kelland LR,

Harrap KR, Bernard CFJ, Murrer BA (1995) J Chromatogr A712:61–66

23. Lipsett MN (1965) J Biol Chem 240:3975–397824. Montgomery JA (1995) In: Foye WO (ed) Cancer chemother-

apeutic agents. American Chemical Society, Washington, pp58–66

25. Dubler E (1996) In: Siegel A, Siegel H (eds) Metal ions inbiological systems, vol 32. Dekker, New York, pp 301–338

26. Elion GB (1989) Science 244:41–4727. Waters TR, Swann PF (1997) Biochemistry 36:2501–250628. Arpalahti J (1999) Perspect Bioinorg Chem 4:165–20829. Arpalahti J, Lippert B (1990) Inorg Chem 29:104–11030. Arpalahti J, Mikola M, Mauristo S (1993) Inorg Chem

32:3327–333231. Xu Y (2000) Prog Nat Sci 10:401–41332. Lide D (ed) (1994) CRC handbook of chemistry and physics.

CRC Press, Boca Raton, Fla33. Giandomenico CM, Abrams MJ, Murrer BA, Vollano JF,

Rheinheimer MI, Wyer SB, Bossard GE, Higgins JDI (1995)Inorg Chem 34:1015–1021

34. Barton SJ, Barnham KJ, Habtemariam A, Sue RE, Sadler PJ(1998) Inorg Chim Acta 273:8–13

35. Coleman RS, Siedlecki JM (1992) J Am Chem Soc 114:9229–9230

36. Coleman RS, Kesicki EA (1994) J Am Chem Soc 116:11636–11642

37. Coleman RS, Arthur JC, McCary JL (1997) Tetrahedron53:11191–11202

38. http://paris.chem.yale.edu/extinct.html39. Igarashi-Yamamoto N, Tajiri A, Hatano M, Shibuya S, Ueda

T (1981) Biochim Biophys Acta 656:1–1540. Hampton A, Maguire MH (1961) J Am Chem Soc 83:150–15741. Cantor CR, Schimmel PR (1980) Biophysical chemsitry. Part

II: techniques for the study of biological structure and function.Freeman, San Francisco, p 412

42. Patil SD, Rhodes DG (2000) Nucleic Acids Res 28:2439–244543. Psoda A, Kazimierczuk Z, Shugar D (1974) J Am Chem Soc

96:6832–683944. Rodger A, Norden B (1997) Circular dichroism and linear

dichroism. Oxford University Press, New York45. Espenson JH (1995) Chemical kinetics and reaction mecha-

nisms. McGraw-Hill, Singapore, pp 203–21046. Lohman TM, DeHaset PL, Record MTJ (1978) Biophys Chem

8:281–29447. Ericson A, McCary JL, Coleman RS, Elmroth SKC (1998)

J Am Chem Soc 120:12680–1268148. Ericson A, Iljina Y, McCary JL, Coleman RS, Elmroth SKC

(2000) Inorg Chim Acta 297:56–6349. Lamm G, Pack GR (1990) Proc Natl Acad Sci USA 87:9033–

903650. Hanlon S, Wong L, Pack GR (1996) Biophys J 72:291–30051. Pack GR, Wong L (1996) Chem Phys 204:279–28852. Fenley MO, Manning GS, Olson WK (1990) Biopolymers

30:1191–120353. Olmsted MC, Anderson CF, Record MTJ (1989) Proc Natl

Acad Sci USA 86:7766–777054. Olmsted MC, Bond JP, Anderson CF, Record MTJ (1995)

Biophys J 68:634–64755. Stein VM, Bond JP, Capp MW, Anderson CF, Record MTJ

(1995) Biophys J 68:1063–107256. Zhang W, Bond JP, Anderson CF, Lohman TM, Record MTJ

(1996) Proc Natl Acad Sci USA 93:2511–251657. Zhang W, Ni H, Capp MW, Anderson CF, Lohman TM,

Record MTJ (1999) Biophys J 76:1008–101758. Sykfont A, Ericson A, Elmroth SKC (2001) Chem Commun

1190–119159. Wang Y, Farrell N, Burgess JD (2001) J Am Chem Soc

123:5576–5577

44