assessing the sequence specificity in the binding of co(iii) to dna via a thermodynamic approach

Michael HicksGeorge Wharton III Assessing the SequenceDaniel H. Huchital

Specificity in the Binding ofW. Rorer Murphy, Jr.Richard D. Sheardy Co(III ) to DNA via a

Department of Chemistry,Seton Hall University, Thermodynamic Approach

400 South Orange Avenue,South Orange,

NJ 07079-2694

Received 19 September 1996;accepted 17 April 1997

Abstract: The interaction specificities of Co(III) with DNA were investigated via consider-ation of thermodynamic characteristics of the duplex to single strand transition for DNAoligomers incubated in the presence of [Co(NH3)5(OH2)](ClO4)3 . It has previously beendemonstrated that incubation of the DNA oligomer [(5medC-dG)4]2 with this cobalt complexleads to coordination of the cobalt center to the DNA, presumably at N7 of guanine bases [D.C. Calderone, E. J. Mantilla, M. Hicks, D. H. Huchital, W. R. Murphy, Jr. and R. D. Sheardy,(1995) Biochemistry 34, 13841] . In this report, DNA oligomers of different sequence wereincubated with [Co(NH3)5(OH2)](ClO4)3 via protocols developed previously and the treatedoligomers were subjected to thermal denaturation for comparison to the untreated oligomers.The DNA oligomers were designed in order to investigate the sequence specificity, if any, inthe reaction of the cobalt complex with DNA. The values of Tm , DH

£H , and Dn ( the differentialion binding term) obtained from the thermal denaturations were used to assess the sequencespecificity of the interaction. For all oligomers, treated or untreated, Tm and DH

£H vary linearlywith log [Na/] and hence the value of Dn is a function of the Na/ concentration. The resultsindicate no significant reaction between the cobalt complex and oligomers possessing isolated-GA- or -CG- sites; however, the thermodynamic characteristics of DNA oligomers pos-sessing either an isolated -GG- site or an isolated -GC- site were altered by the treatment.Atomic absorption studies of the treated oligomers demonstrate that only the DNA oligomerspossessing isolated -GG- or -GC- sites bind cobalt. Hence, the changes in the thermodynamic

properties of these oligomers are a result of cobalt binding with a remarkable sequence specificity.q 1997 John Wiley & Sons, Inc. Biopoly 42: 549–559, 1997

Keywords: thermodynamic characterization; optical melting studies; sequence specificity;binding of Co(III) to DNA; designed oligomers

Correspondence to: Richard D. SheardyContract grant sponsor: National Institutes of Health (NIH)

and Bristol-Meyers-Squibb/Research Corporation (BMS); con-tract grant number: GM51069-01 (NIH) and HS0316 (BMS)q 1997 John Wiley & Sons, Inc. CCC 0006-3525/97/050549-11

549

8K30 5503/ 8K30$$5503 08-18-97 11:27:19 bpa W: Biopolymers

550 Hicks et al.

plementary DNA oligomers referred to as the COTARINTRODUCTION(cobalt target) series:

The development of new molecules that bind COTAR1: 5*-ATTAAT-CTTAAG-ATTAAT-3 *strongly and selectively to nucleic acids expands the

COTAR2: 5*-ATTAAT-GGATCC-ATTAAT-3 *range of potential pharmaceutical agents whose modeof bioactivity is through interaction with DNA or COTAR3: 5*-ATTAAT-AAGCTT-ATTAAT-3 *RNA. Ideal candidates for such reagents are transi-

COTAR4: 5*-ATTAAT-TTCGAA-ATTAAT-3 *tion metal complexes due to the wide range of struc-tures and reactivities found in these molecules.1 The

The design constraints limited the number of G bases:relatively large size of these complexes allows theCOTAR1 has a single G flanked by two A bases; CO-sampling of a significant portion of the nucleic acidTAR2 has a GG site; COTAR3 has a GC site and CO-

surface, increasing the likelihood of substantial speci-TAR4 has a CG site ( the sequence variations are indi-

ficity. The positive charge of many readily prepared cated in bold) . Due to the high AT content, these oligo-metal complexes can potentially increase the strength mers are all 18-mers in order to ensure duplex stabilityof complex–nucleic acid interaction due to electro- at 377C and 50 mM Na/ , our incubation conditions. Instatic effects. We have developed protocols for re- addition, each oligomer possesses at least two differentacting a Co(III) complex bearing a labile ligand with restriction endonuclease recognition sites: COTAR1

has a A f l II site (C/TTAAG); COTAR2 has a BamH Ia DNA oligomer to yield a robust cobalt–nucleicsite (G/GATCC); COTAR3 has a Hind III site (A/acid adduct.2 For these preliminary studies, we syn-AGCTT); COTAR4 has a Bstb I site (TT/CGAA); andthesized an eight-base self-complimentary oligomerall have an Asn I site (AT/TAAT). Note that the unique(medC-dG)4, designated as Z8. Z8 was incubatedsites possess at least one G base.in the presence of [Co(NH3)5(OH2)](ClO4)3 at var-

DNA oligomers were synthesized with an ABI 380Bious DNA/cobalt ratios followed by exhaustive so-

DNA Synthesizer (Applied Biosystems, Foster City, CA)dium exchange dialysis to remove any unreacted co- and purified as previously described by reverse phasebalt complex. Atomic absorption spectroscopic anal- high performance liquid chromatography (HPLC).2 Afterysis of the treated oligomers indicated a gradual the second HPLC purification, the oligomers were sub-uptake of cobalt reaching a limiting value of one jected to exhaustive dialysis vs water and then lyophilizedcobalt per four base pairs. Thermal denaturation, UV to dryness. DNA purity was confirmed via analytical

HPLC and gel electrophoresis.visible absorption, and CD spectroscopic studies ofthe cobalt–nucleic acid adduct indicated that its con-formational properties were dramatically altered Preparation of Cobalt Modified DNAfrom the untreated oligomer. The results were inter- Oligomerspreted in terms of coordination of the cobalt center

Lyophilized DNA oligomers were dissolved in standardto N7 of the guanine bases of the DNA with concomi-phosphate buffer (5 mM phosphate, pH 7.0, 50 mM Na/)tant loss of the aquo ligand.to give a final concentration of 40–75 mM ( in base pairs) .This interaction is quite different from the equi-The [Co(NH3)5(OH2)](ClO4)3 complex was then addedlibrium binding of [Co(NH3)6]3/ to DNA oligo-to give final concentrations ranging from 100 to 400 mM.mers which proceeds primarily via hydrogen bond-Each sample was incubated at 377C for 48 h followed bying.2,3–5 The binding of Co(III) through loss of aexhaustive dialysis vs 200 mM NaCl then water. The

labile ligand to a DNA oligomer possessing G bases aqueous samples were then lyophilized to dryness. Priorshould not be surprising since Co(II) has been to thermal denaturation, the samples were reconstitutedshown to bind to N7 of G bases in other synthetic in standard phosphate buffer with NaCl added to obtainDNA oligomers.6,7 Here, we report studies aimed at various concentrations of Na/ , heated to 807C for 2 min,assessing the sequence specificity of this reaction. and then slowly reannealed by cooling to room tempera-

ture.

Optical Melting StudiesMATERIALS AND METHODSTwo types of studies were carried out. In the first setof studies, the DNA oligomers were treated at variousDNA Oligomer Design and Synthesisconcentrations of [Co(NH3)5(OH2)](ClO4)3 as de-scribed above. After dialysis and lyophilization, the sam-In order to investigate the sequence specificity in the

reaction of [Co(NH3)5(OH2)](ClO4)3 with DNA, we de- ples were reconstituted in standard buffer with NaCladded to give a final concentration of Na/ of 115 mM.signed and synthesized a preliminary series of self-com-


Assessing Sequence Specificity 551

In the second set of experiments, the DNA oligomers Thus the variation in Tm , and hence DG , withwere treated with a fixed concentration of 200 mM log[Na/] is attributed to the entropic dependence[Co(NH3)5(OH2)](ClO4)3 , dialyzed, lyophilized, and on the concentration of sodium ions.12 However,reconstituted in standard phosphate buffer with NaCl recent studies on DNA dumbbell molecules haveadded to give final concentrations of Na/ of 53, 150, or shown that DH may in fact be mildly dependent330 mM. For these experiments, the concentration of

upon the sodium ion concentration.15 It can beDNA was 5.4 1 1005M in base pairs.shown, in this particular case, that the differentialFor each thermal denaturation experiment, the ab-ion binding terms can be defined by Eq. (2) .sorbance at 260 nm was monitored as the temperature

was ramped from 20 to 857C at a rate of 0.37C/min. Thedata were transferred to a personal computer for analysis. Dn Å (DH /2.303RT 2

m)(dTm /d log[Na/])The melting temperatures reported here, Tm , were ob-

/ (2.303RT )01(1 0 T /Tm) (2)tained from the inflection points from the first derivativeof the A260 vs T plots. 1 (dDH /d log[Na/])

Here, Dn is the differential ion binding term andGraphite Furnace Atomic Absorptionrepresents the number of sodium ions/duplex re-(AA) Studiesleased from the polymer as it undergoes the duplex

In order to determine the presence of any cobalt bound to single strand transition. The question naturallyto the DNA after the incubation/dialysis treatment, AA

arises about the suitability of analyzing thermody-spectra of the cobalt treated oligomers were recorded withnamic data obtained for simple linear oligomers ina Polarized Zeeman Spectrometer Z-8270 from Hitachilight of the polyelectrolyte theory due to differentialusing a platform graphite tube with Argon purge. Thebinding of sodium ions at the ends of the oligomertemperature program involved drying from 80 to 1407C,relative to the interior.13,14 However, since all theashing at 2007C for a 20 s hold, ramping to 8007C for a

30 s hold, and then atomizing at 27007C and reading the oligomers under study here are the same length andabsorbance at 240.7 nm. Readings were performed in the same terminal bases, the end effects should betriplicate using a standard SSC-300 Hitachi Autosampler the same for all oligomers. The polyelectrolyte the-and compared to an external standard curve at 20, 50, ory also states that the number of monovalent coun-and 100 ppb. terions bound to a lattice of evenly spaced charges

is relatively independent of the bulk ionic strengthfor a polymeric DNA.9–13 However, for oligomers,

THEORY the differential thermodynamic ion association termvaries linearly with 1/N , where N is the number ofnucleotides.12 Thus, when discussing the release ofThe optical melting curves can be analyzed via Eq.bound sodium ions from oligomers, one must spec-(1) to obtain the enthalpy of the duplex to singleify the bulk concentration of sodium ions for thestrand transition via the two-state model.8 Equationparticular calculation.(1) is appropriate for self-complementary oligomers

where DH£H is the van’t Hoff enthalpy change, Tm

is the inflection point of the a vs T plot and ais the fraction of single strands. By plotting the RESULTS AND DISCUSSIONdifferential melting curve (i.e., dA /dT vs T ) , onecan directly obtain Tm and da /dT vida infra from The initial set of experiments were carried out tothe x and y coordinates, respectively, of the peak determine how treatment of the COTAR oligomersmaximum. with varying concentrations of the cobalt complex,

via established protocols, would affect the Tm ofDH

£H Å 6RT 2m(da /dT )TÅTm

(1) each oligomer. The results, plotted as DTm vs con-centration of cobalt complex in the incubation and

According to counterion condensation theory,9–14 shown in Figure 1, indicate that the Tm values forthere is a linear relationship between Tm and COTAR1 and COTAR4 show little difference fromlog[Na/] (at concentrations of Na/ less than 300 the untreated oligomer regardless of the incubationmM) in the thermal denaturation of polymeric concentration of the cobalt complex. However, CO-DNA. In addition, it is often assumed that the en- TAR2 shows a dramatic decrease in its Tm and CO-

TAR3 displays a slight increase in its Tm with in-thalpy of the transition, DH , and DH /RT 2m are both

independent of the concentration of sodium ions. creasing [Co(NH3)5(OH2)](ClO4)3 in the incuba-


552 Hicks et al.

reported in Table II. One of the major sources oferror in these determination arises from the scatterin the da /dT values (and hence DH

£H values) ob-tained.

Examination of the plots in Figures 2–6 indicatea linear relationship between Tm and log[Na/] , asexpected, for all oligomers. The slopes and y inter-cepts varies among the different oligomers. Com-paring the data of a treated oligomer to its untreatedparent indicates little variation, within experimentalerror, in the values of Tm and dTm /d log[Na/] forCOTAR1 and for COTAR4. However, there aredramatic differences in Tm and dTm /d log[Na/] fortreated COTAR2 and COTAR3 relative to their un-treated parents. Treatment of COTAR2 with the co-balt complex indicates a more pronounced depen-dence of Tm on log[Na/] , while treatment of CO-TAR3 results in a decreased dependence of Tm onFIGURE 1 Variation of DTm as a function of the incu-log[Na/] . Consistency with these observationsbation concentration of [Co(NH3)5(OH2)](ClO4)3 : CO-arises from consideration of the DH

£H values forTAR1 (circles) , COTAR2 (squares) , COTAR3 (trian-cobalt treated and untreated oligomers. COTAR1gles) , and COTAR4 (upside down triangles) . Meltingand 4 show no perturbation of their respective ther-profiles were obtained in standard buffer at 115 mM Na/

and Tm values obtained at the midpoint of the duplex to modynamic properties under all conditions exam-single strand transition.8 DTm Å Tm ,t 0 Tm ,u , where Tm ,t ined. However, treatment of COTAR2 with the co-and Tm ,u are the melting temperature of treated and un- balt complex leads to an enthalpic destabilizationtreated DNA oligomer respectively. For these experi- relative to the untreated parent while treatment ofments, the temperature was ramped from 20 to 857C at COTAR3 results in enthalpic stabilization relative0.37C/min, the transition monitored at 260 nm and to the untreated parent.[DNA] was 5.4 1 1005M in base pairs.

One of the most notable observations is the milddependence of DH

£H on log[Na/] for these oligo-

tion. It is important to note here that all melts werereversible, indicating that any alterations were per-manent.

It has been argued that DTm data should not beused as a substitute for thermodynamic characteriza-tions when analyzing the interactions of small mole-cules with DNA.16 Hence, thermal denaturationstudies were then performed on untreated andtreated COTAR oligomer samples in standard bufferwith 53, 150, or 330 mM total Na/ concentration.Figure 2 shows a typical set of differential meltingprofiles at the three different Na/ concentrations.The slight asymmetry in the derivative plots mayindicate slight deviations from true two state behav-ior. However, the effect of any slight deviation fromtwo state behavior should be minimal for the follow-

FIGURE 2 Typical differential melting curve (i.e.,ing analysis. The values of Tm and DH£H [via Eq.

dA260 nm/dT vs T ) for untreated COTAR4. For these ex-(1)] were obtained for each oligomer at each Na/ periments, the temperature was ramped from 20 to 857Cion concentration. Plots of Tm vs log[Na/] and at 0.37C/min, the transition monitored at 260 nm andDH

£H vs log[Na/] resulted in linear fits (Figures [DNA] was 5.4 1 1005M in base pairs. The DNA oligo-3–6) of high correlation (r 2ú 0.992). The primary mer was prepared in standard phosphate buffer and NaCldata used to construct these plots are given in Table was added to give final concentrations of 53 (solid line) ,

150 (dotted line) , or 330 (dash-dot line) mM total Na/ .I and the slopes and y intercepts for these plots are



the charge density of a polymeric random coil andhas a reported value of 0.71,12,14 and Np is the num-ber of phosphates in the DNA segment under con-sideration (for a linear 18-mer, Np Å 34). For thiscalculation using Eq. (3) , Np Å 34, CD Å 0.88, andCC Å 0.71, giving Dn Å 5.78. The value of 5.78noted above thus represents the upper limiting valuefor Dn . 14,15 Inspection of the values listed in TableIII indicate that Dn increases with increasing [Na/]but are lower than the limiting value as expectedfrom end effects.14

Comparison of the Dn values of a treated oligo-mer with its untreated parent indicates slight differ-ences (within experimental error) for COTAR1 andCOTAR4 (differences of about 0.02 on average), amoderate difference for COTAR3 (increasing from00.12 to 00.05 with increasing [Na/]) and a sig-nificant difference for COTAR2 (decreasing from0.26 to 0.13 with increasing [Na/]) . The negativeDn for COTAR3 suggests that the treated oligomerreleases more sodium ions upon denaturation.

FIGURE 3 Plots of Tm vs log[Na/] and DH£H vs

log[Na/] for the thermal denaturation of untreated andtreated COTAR1: the solid points are for the untreatedoligomer, the open points are for the treated oligomer.The lines are the results of the linear regression analysisof the data. Experimental conditions for these denatur-ations are the same as for the data presented in Figure 2.

mers. Further, there are differences in the values ofdDH

£H /d log[Na/] among the various oligomers.These differences are most likely due to subtle dif-ferences in base composition and sequence. Hence,the enthalpy dependence on Na/ concentration mustbe included in the evaluation of the differential ionbinding term Dn via Eq. (2) .

Table III lists the values of DH£H /RT 2

m and Dncalculated from the experimentally determinedvalues of Tm , DH

£H , dTm /d log[Na/], and dDH£H/d

log[Na/] listed in Tables I and II. Both DH£H/RT2

m

and Dn are mildly dependent on log[Na/] . Ac-cording to Eq. (3) , the number of sodium ions re-leased from an eighteen base pair segment embed-

FIGURE 4 Plots of Tm vs log[Na/] and DH£H vsded in a polynucleotide should be 5.78.14,17

log[Na/] for the thermal denaturation of untreated andtreated COTAR2: the solid points are for the untreated

Dn Å Np(Cd 0 Cc) (3)oligomer; the open points are for the treated oligomer.The lines are the results of the linear regression analysis

In Eq. (3) , Cd is the charge density of a polymeric of the data. Experimental conditions for these denatur-ations are the same as for the data presented in Figure 2.duplex and has a reported value of 0.88,12,14 Cc is


554 Hicks et al.

clude that each site binds a cobalt since there aretwo -GG- sites per duplex. It is surprising, however,that less cobalt was detected bound to COTAR3after 48 hours of incubation. However, after an addi-tional 12 h of incubation, COTAR3 also binds 1.9{ 0.1 cobalts per duplex. The AA data confirmsthat cobalt is present for two of the four oligomers.This suggests a selectivity, given by the fact thatCOTAR2 binds more cobalt than COTAR3 in thesame time period. Since COTAR2 binds substan-tially more cobalt, it is not surprising that the Dndata is different from COTAR3, which may havemore unreacted sites. We have previously shownthat Z8 binds ca. one cobalt for every four basepairs.2 COTAR3 possesses one -GC- site while Z8possesses multiple -GC- sites/duplex. Thus, the re-activity of COTAR3 to bind cobalt may be influ-enced by the sequences flanking the reactive sites.We are currently investigating this question andhow it relates to reaction time.


log[Na/] for the thermal denaturation of untreated andtreated COTAR3: the solid points are for the untreatedoligomer; the open points are for the treated oligomer.The lines are the results of the linear regression analysisof the data. Experimental conditions for these denatur-ations are the same as for the data presented in Figure 2.

The thermal denaturation data presented in Ta-bles I and III demonstrate that treatment of theCOTAR2 and COTAR3 oligomers with [Co(NH3)5

(OH2)](ClO4)3 alters their duplex to single strandequilibria. In addition, the thermal melting profilesare reversible indicating that the alterations are es-sentially permanent. These results suggest that co-balt is tightly bound to the DNA. Further, the ap-parent absence of any alterations in the confor-mational properties of COTAR1 and COTAR4suggests that cobalt is not bound to the DNA. Thus,DNA oligomers that were treated with 200 mM[Co(NH3)5(OH2)](ClO4)3 for 48 h, followed bydialysis and lyophilization were also assayed for thepresence of tightly bound cobalt (III) by AA. The


results shown in Table IV are consistent. As can log[Na/] for the thermal denaturation of untreated andbe seen, no cobalt was detected for COTAR1 and treated COTAR1: the solid points are for the untreatedCOTAR4, while COTAR2 binds 1.9 { 0.1 cobalt / oligomer; the open points are for the treated oligomer.duplex and COTAR3 binds 0.6 { 0.1 cobalt /du- The lines are the results of the linear regression analysisplex. of the data. Experimental conditions for these denatur-

ations are the same as for the data presented in Figure 2.In the case of COTAR2, it is tempting to con-



Table I Melting Parameters for Treated and Untreated COTAR DNA Oligomersa

Untreated Treated

DHvH DHvH

Oligomer [Na/] (mM) Tm (K) (kcal/mol) Tm (K) (kcal/mol)

COTAR1 53 323.8 74.7 323.9 76.4150 327.8 81.5 328.1 83.3330 330.6 88.4 330.8 88.3

COTAR2 53 322.0 109.9 320.7 82.6150 325.8 114.2 325.3 87.5330 329.4 117.1 329.2 91.1

COTAR3 53 312.7 76.1 314.2 92.8150 318.0 82.2 318.1 99.3330 321.6 87.5 321.9 103.9

COTAR4 53 316.5 87.0 317.0 87.3150 321.2 90.6 321.5 90.9330 324.5 93.9 325.0 93.8

a The reported Tm values are {0.37C and the DHvH values are {3–6%.

There have been many studies on the interactions been used to induce the B to Z transition in syntheticDNA oligomers possessing such repeats.19–21 The sta-of simple metal complexes with both oligomeric

and polymeric DNA as well as RNA.18–34 The types bilization of Z-DNA by [Co(NH3)6]3/ has been at-tributed to specific hydrogen bonds between three ofof reversible interactions observed are based on

electrostatic, hydrogen bonding, and hydrophobic the cobalt ammine groups with guanine and phosphateacceptor sites on the DNA helix.3effects. A number of small metal complexes with

appropriate ligands have also been shown to interact This reagent has also been shown to induce anunusual non-Z-like structure in a DNA oligomerwith DNA via coordination of the nitrogenous bases

to the metal via loss of a labile ligand from the containing a (dC-dG)4 segment22 and to induce theB to A transition in oligomers of sequence dCC-metal. Both reversible and irreversible interactions

display sequence specificity in the binding. For ex- CCGGGG.23 Braunlin’s group has also observeddifferences in the binding modes of [Co(NH3)6]3/ample, it is well known that cobalt(III) hexammine,

[Co(NH3)6]3/ , induces the B to Z transition in to a wide variety of DNA oligomers.24 This reagentbinds to polymeric native DNA with high affinity4,25DNA polymers possessing long runs of alternating

purine–pyrimidine dinucleotide subunits such as dC- and with a preference to GC-rich DNAs5 and cancondense DNA into secondary structures.26 All ofdG and 5medC-dG.18 This cobalt complex has also

Table II Line Parameters for Tm vs log[Na/] and DHvH vs log[Na/] Plotsa

Tm vs log[Na/] DHvH vs log[Na/]

Oligomer Slope y Intercept Slope y Intercept

COTAR1 8.58 334.8 17.1 96.3COTAR1 treated 8.72 335.1 14.99 95.6COTAR2 9.27 333.7 9.09 121.6COTAR2 treated 10.67 334.2 10.71 96.3COTAR3 11.23 327.1 14.3 94.2COTAR3 treated 9.64 326.4 14.0 110.7COTAR4 10.08 329.4 8.65 97.9COTAR4 treated 10.06 329.8 8.17 97.7

a The values listed were generated by linear regression analysis of the data presented in Figures 3–6.


556 Hicks et al.

Table III DHvH/RT2m and Dn Values for Treated and Untreated COTAR DNA Oligomersa

Untreated Treated

Oligomer [Na/] (mM) DHvH/RT 2m Dn DHvH/RT 2

m Dn

COTAR1 53 0.358 1.85 0.366 1.82150 0.317 2.07 0.389 2.06330 0.407 2.26 0.406 2.20

COTAR2 53 0.533 2.38 0.404 2.12150 0.541 2.49 0.416 2.28330 0.543 2.56 0.423 2.43

COTAR3 53 0.392 1.99 0.473 2.11150 0.409 2.24 0.494 2.32330 0.426 2.43 0.505 2.48

COTAR4 53 0.437 2.05 0.437 2.04150 0.442 2.14 0.443 2.14330 0.449 2.39 0.447 2.22

a The errors for the DHvH/RT 2m are typically 3–6% and the error in Dn values are {0.004–0.006.

these interactions are reversible and indicate se- plexes such as cisplatin react by an associative (a-quence specificity in the binding, most likely to the type) mechanism involving the formation of a fivemost basic position, G-N7, in the major groove. coordinate intermediate. Co(III) ammines such as

Coordination of metal complexes to DNA has also [Co(NH3)5(OH2)](ClO4)3 typically follow a dis-been observed. For example, [Ru(NH3)5(OH2)]2/ sociative (d-type) intimate mechanism, where thehas been shown to bind to calf thymus DNA pri- breaking of the bond between the metal and themarily at N7 of guanine bases.27 Cisplatin, cis- leaving group is the key contribution to the activa-Pt(NH3)2Cl2 , also prefers sites rich in G bases since tion energy. The transition state results from a de-the point of covalent attachment is also to the N7 crease in coordination number (from 6 to 5) in theof guanine.28–30 In this particular case, both in- purely d-type mechanism, but the dissociative inter-terstrand and intrastrand cross-linking have been ob- change or Id mechanism involves the initial forma-served with intrastrand cross-linking predominant tion of an outer sphere complex; the expanded coor-at 5*-GG-3 * sites.31–33 Hopkins et al. have demon- dination does not involve such a transition state. Ionstrated that the interstrand cross-linking occurs very pairing between the complex and the entering groupefficiently at 5 *-GC-3* sites.30 It has also been sug- will increase KOS and the likelihood of an Id intimategested that cis-[Ru(NH3)2Cl2]/2 behaves in a man- mechanism. This would most likely be the case inner similar to cisplatin.34

the modification of nucleic acids with cationic co-The substitutional mechanisms by which Co(III) balt complexes.

and Pt(II) react are quite different.35–38 Pt(II) com- The coordination of the complex to the DNA,ultimately depends on the ease of formation of theouter sphere complex with the DNA and its over-

Table IV AA Analysis of Treated COTAR DNA all affect on the reaction chemistry. The complexOligomera

pKa allows for 15% [Co(NH3)5OH2]3/ and 85%[Co(NH3)5OH]2/ at the reaction pH of 7.0. BasedOligomer ron the relative ion pairing strengths, the electrostati-cally driven outer sphere complexation equilibriumCOTAR1 NDlies ca. 2.7 times more in favor of the outer sphereCOTAR2 1.9 { 0.1complex for the trivalent species over the divalentCOTAR3 0.60 { 0.1

COTAR4 ND species.35,36 The acid–base equilibrium of the com-plex must be maintained as the trivalent aquo spe-

a The value of r is defined as the binding ratio and equals to cies binds more readily to the DNA. In addition,number of cobalt(III) atoms bound per DNA duplex. The entrythe t1/2 for the rate of aquation at the reaction tem-of ND indicates that no cobalt was detected for values at or

below 0.10 mg/mL. perature, 377C, is 16 h less than the estimated rate of



PROPOSED SCHEME 1

anation for the hydroxo.* The reaction rate observed pend upon the mode of interaction of the Co(III)with the DNA. Thus, the changes in Tm and DH

£Hfrom conformational changes to the DNA via CDat l Å 287 nm indicate that the reaction has a t1/2 with ionic strength can be used to assess the se-

quence specificity of the binding reaction.of 13–15 h.39 A possible explanation is that theinitial outer sphere coordination chemistry for The variations in the oligomers considered here

arise from the position and number of G bases in a[Co(NH3)5OH2]3/ coupled to the faster rate ofaquation favors the refurbishment of the aquo spe- six base pair segment in the middle of the sequence.

The results presented here indicate a sequence spec-cies over the hydroxo (Proposed Scheme 1). Thehydroxo species does, however, offset this reaction ificity in the reaction between [Co(NH3)5(OH2)]

(ClO4)3 and these oligomers. The oligomer pos-for the smaller percentage that coordinates to theDNA. Regardless of the speciation, both will ligand sessing a single G base flanked by two A bases

(COTAR1) and the oligomer possessing a CG sitesubstitute within the 48 h reaction time as deter-mined by AA for COTAR2 and COTAR3 as it did (COTAR4) are apparently unaffected by treatment

with the cobalt complex suggesting that Co(III)for Z8 in the past.2

We have used thermodynamic characterization does not react with either oligomer to any significantextent. However, the sequence that contains twoof the thermal denaturation of DNA oligomers to

study the base sequence preference of the reaction adjacent G bases (COTAR2) apparently bindsCo(III) and is quite sensitive to its presence. Theof [Co(NH3)5(OH2)](ClO4)3 with specific DNA

oligomers. These studies indicated whether reaction resultant cobalt–nucleic acid adduct has a signifi-cantly different charge density and enthalpic stabil-with the complex occurred and how the DNA re-

sponded thermodynamically to the resultant pertur- ity. The sequence possessing a GC (COTAR3) siteis also altered by treatment with [Co(NH3)5(OH2)]-bation. Certainly the coordination of cobalt(III) to

the DNA substrate would dramatically alter the (ClO4)3 resulting in an enthalpic stabilization. Al-though the results obtained here for COTAR3 withcharge density of the oligomer. In light of the coun-

terion condensation theory, this alteration should a single GC site/strand cannot be directly comparedto Z8, which possesses 3–4 GC sites/strand as de-thus change the number of sodium counterions

bound to either the duplex, the single strand, or termined by AA,2 both sets of results point toCo(III) reacting with DNA oligomers with GCboth, resulting in a change in Dn . In addition, the

presence of any positively charged, covalently sites. The changes in the Tm values for COTAR2and COTAR3 are similar to what would be observedbound species, such as Co(III) , to a DNA oligomer

alters the enthalpic contribution to the duplex stabil- for intrastrand and interstrand cross-linking by cis-ity.40 The alteration may, for example, lead to stabi- platin, respectively.41 It has been demonstrated thatlization arising from the resultant favorable electro- cisplatin preferentially cross-links DNA in an intra-static interaction or destabilization arising from he- strand fashion at GG sites and, to a lesser extent,lix distortion at the point of coordination. The most at GA sites31–33 and cross-links in an interstrand fash-likely scenario is a combination of both effects. The ion at GC sites.30,42 It is tempting to speculate cross-magnitude and sign of the changes in Dn and DH

£H linking at GG or GC sites by [Co(NH3)5(OH2)]-and their dependence on Na/ concentration will de- (ClO4)3 , a process that could mandate the loss of

an ammine ligand by the entropically driven chelateeffect.35 Further studies are necessary in order to* Via application of the Arrhenius equation using data pub-

lished in Ref. 32 for the rate of aquation and anation at 257C. delineate the interaction specificities and reaction


558 Hicks et al.

2. Calderone, D. C., Mantilla, E. J., Hicks, M., Huchi-molecularity of this simple cobalt (III) complex andtal, D. H., Murphy, W. R., Jr. & Sheardy, R. D.analogues with DNA.(1995) Biochemistry 34, 13841.

3. Gessner, R. V., Quigley, G. J., Wang, A. H.-J., vander Marel, G. A., van Boom, J. H. & Rich, A. (1985)CONCLUSIONSBiochemistry 24, 237.

The incubation of [Co(NH3)5(OH2)](ClO4)3 with 4. Braunlin, W. H., Anderson, C. F. & Record, M. T.,Jr. (1987) Biochemistry 7724.four oligomers for 48 h indicated that of the four,

5. Braunlin, W. H. & Xu, Q. (1992) Biopolymers 32,COTAR1 (5 *-(-CTTAAG-)-3 *] and COTAR41703.[5 *-(-TTCGAA-)-3 *] were not thermodynamically

6. Marzilli, L. G. (1977) Prog. Inorg. Chem. 23, 255.altered by cobalt treatment. COTAR2 [5*-(-GGA-7. Gao, T.-G., Sriram, M. & Wang, A. H.-J. (1993)TCC-)-3 *] and COTAR3 [5*-(-AAGCTT-)-3 *] ,

Nucleic Acids Res. 21, 4093.on the other hand, did show notable changes in8. Marky, L. A. & Breslauer, K. J. (1987) Biopolymersthermodynamics through changes in counterion

26, 1601.condensation release (Dn) and duplex–single 9. Manning, G. S. (1972) Biopolymers 11, 937.strand equilibria (Tm and DH

£H) . These differences 10. Manning, G. S. (1978) Quart. Rev. Biophys. 11, 179.are believed to be the result of the selective coordi- 11. Record, M. T., Jr. (1975) Biopolymers 14, 2137.nation of the complex to COTAR2, and to a lesser 12. Record, M. T., Jr., Anderson, C. F. & Lohman, T. M.degree to COTAR3 within the reaction time investi- (1978) Quart. Rev. Biophys. 11, 102.gated. Although the precise nature of the adduct is 13. Record, M. T., Jr., Mazur, S. J., Melancon, P., Roe,unclear at this time, atomic absorption data confirm J. H., Shaner, S. L. & Unger, L. (1981) Ann. Rev.

Biochem. 30, 997.that the complex is irreversibly bound to the two14. Olmsted, M. C., Anderson, C. F. & Record, M. T.,oligomers following treatment and exhaustive so-

Jr. (1991) Biopolymers 31, 1593.dium exchange dialysis.15. Paner, T. M., Riccelli, P. V., Owczarzy, R. & Be-The thermodynamic stability of the cobalt treated

night, A. S. (1996) Biopolymers 39, 779.COTAR2 relative to COTAR3 implies that the ad-16. Pilch, D. S., Plum, G. E. & Breslauer, K. J. (1995)duct that forms upon reannealing is sequence depen-

Curr. Opin. Struct. Biol. 5, 334.dent. This sequence dependency we feel is a direct17. Record, M. T., Jr. & Lohman, T. M. (1978) Biopoly-

result of the -GG- vs -GC- modifications in the mers 17, 159.center of COTAR2 and COTAR3, respectively. Un- 18. Behe, M. & Felsenfeld, G. (1981) Proc. Natl. Acad.like other commonly investigated trivalent cations Sci. USA 78, 1619.such as [Co(NH3)6]3/ , the species that result from 19. Winkle, S. A. & Sheardy, R. D. (1990) Biochemistry[Co(NH3)5(OH2)](ClO4)3 , [Co(NH3)5(OH2)]3/ , 29, 6514.and [Co(NH3)5(OH)]2/ , both form outer sphere 20. Winkle, S. A., Aloyo, M. C., Morales, N., Zambrano,

T. Y. & Sheardy, R. D. (1991) Biochemistry 30,complexes that ultimately lead to the covalent at-10601.tachment of the complex. The most likely position

21. Lu, M., Kallenbach, N. R. & Sheardy, R. D. (1992)of attachment is the most basic position of the majorBiochemistry 31, 4712.groove, the G-N7. These results are significant since

22. Winkle, S. A., Aloyo, M. C., Lee-Chee, T., Morales,an alteration of the binding environment also affectsN., Zambrano, T. Y. & Sheardy, R. D. (1992) J.the recognition and reactivity of enzymes such asBiomol. Struct. Dynam. 10, 389.restriction endonucleases. These studies are pres-

23. Xu, Q., Shoemaker, R. K. & Braunlin, W. H. (1993)ently under investigation.

Biophys. J. 65, 1039.24. Xu, Q., Jampani, S. R. B. & Braunlin, W. H. (1993)This work was supported by the National Institutes of

Biochemistry 32, 11754.Health (GM51069-01) and Bristol-Meyers-Squibb/Re-25. Plum, G. E. & Bloomfield, V. A. (1988) Biopoly-search Corporation (Grant HS0316). The authors also

mers 27, 1045.thank Merck & Co. for the use of their Atomic Absorption26. Thomas, T. J. & Bloomfield, V. A. (1983) Biopoly-Spectrometer and for supporting Michael Hicks during

mers 22, 1097.his leave of absence.27. Clarke, M. J., Jansen, B., Marx, K. A. & Kruger, R.

(1986) Inorg. Chim. Acta.28. Murray, V., Motyka, H., England, P. R., Wickham,REFERENCES

G., Lee, H. H., Denny, W. A. & McFadyen, W. D.(1992) Biochemistry 31, 11812.1. Tullius, T. D. (1989) in Metal-DNA Chemistry, Tul-

lius, T. D., Ed., American Chemical Society, Wash- 29. Brabek, V., Reedjil, J. & Leng, M. (1992) Biochem-istry 31, 12307.ington, DC.



30. Hopkins, P. B., Millard, J. T., Woo, J., Weidner, 36. Wilkins, R. G. (1970) Acc. Chem. Res. 3, 408.37. Ghosh, M. C., Bhattacharya, P. & Banerjee, P.M. F., Kirchner, J. J., Sigurdsson, S. T. & Raucher,

S. (1992) Tetrahedron 47, 2475. (1988) Coord. Chem. Rev. 91, 1.38. Cotton, F. A. & Wikinson, G. (1988) Advanced Inor-31. Bruhn, S. L., Toney, J. H. & Lippard, S. J. (1990)

Prog. Inorg. Chem. 38, 477. ganic Chemistry, 5th ed., Wiley Interscience, NewYork, p. 732.32. Pinto, A. L. & Lippard, S. J. (1985) Biochem. Bio-

phys. Acta 780, 167. 39. Hicks, M. H., Xiao, B., Pantano, T., Sheardy, R. D.,Huchital, D. H. & Murphy, W. R., Jr., manuscript in33. Rahmouni, A. & Leng, M. (1987) Biochemistry 26,

7229. preparation.40. Rentzeperis, D., Marky, L. A., Dwyer, T. J., Geiers-34. Clarke, M. J. (1980) in Inorganic Chemistry in Biol-

ogy and Medicine, Miller, A. E., Ed., American tanger, B. H., Pelton, J. G. & Wemmer, D. E. (1995)Biochemistry 34, 2937.Chemical Society, Washington, DC, p. 157.

35. Douglas, B. E., McDaniel, D. H. & Alexander, J. J. 41. Macquet, J.-P. & Butour, J.-L. (1978) Biochimie 60,901.(1994) in Concepts and Models of Inorganic

Chemistry, 3rd ed., John Wiley & Sons, New York, 42. Malinge, J.-M., Perez, C. & Leng, M. (1994) NucleicAcids Res. 22, 3834.p. 493.


assessing the sequence specificity in the binding of co(iii) to dna via a thermodynamic approach

Documents