1975 1977 Purines. Part I. Kinetics of Interaction of Nickel(ii) with some Purine Bases and Nucleosides By Anne Casper and G. Victor Fazakerley,' Department of Inorganic Chemistry University of Cape Town, Rondebosch 7700 S. Africa The interaction of NiT1 with adenine 4-aminopyrazolo[3,4-d]pyrimidine (app) adenosine hypoxanthine and inosine have been studied by stopped-flow techniques. Adenine and app show a complex reaction pattern with both the neutral (HL) and protonated (H,L+) species attacking the nickel ion with subsequent deprotonation, the forward rate constants for adenine being 336 and 1.44 x 1 O3 I mol-l s-1 respectively. The preference for the protonated ligand may be explained in terms of the neutral ligand hydrogen bonding with a water molecule in the metal-ion co-ordination sphere poorly aligning the ligand for attack on Ni'I.Hypoxanthine and inosine show no pH dependence in the region pH 2-6 and give rates within the range of ' normal ' substitution. RELATIVELY few kinetic studies have been reported on the interaction of purines nucleosides and nucleotides l-* with first-row transition-metal ions despite the im-portance of such studies in understanding the role of these ions in biological systems. The nickel(@-adenine system was studied by Boivin and Zador who observed a pH-independent reaction and by Kustin4 who postu-lated a pH-dependent reaction involving only the neutral ligand which deprotonates after attack. It has been suggested that linkage isomerism occurs giving rise to observed non-exponential traces because this behaviour was absent with theophylline (1,3-dihydro-2,6-dioxopurine).Our own results suggest that a bad fit of the data is due to the presence of a more complex reaction scheme than previously considered. EXPERIMENTAL Adenine 4-aminopyrazolo[3,4-d]pyrimidine (app) hypo-xanthine and inosine were obtained from Aldrich Chemi-cal Co. Nickel(I1) perchlorate was prepared from AnalaR Ni[SO,] and Na[HCO,] and treating the Ni[HCO,] produced with AnalaR HClO,. Excess of Ni[HCO,] was filtered off and Ni[ClO,] crystallized. Solutions of Ni[C10,] were standardized by adding a known concentration of ethylene-diaminetetra-acetate (edta) and back titrating with standard Mg[SO,] using Eriochrome Black T as indicator. Dissociation constants were determined by potentiometric titration of acidified ligand solutions using standardized carbonste-free NaOH (GU 0 .1 ~ ) as titrant.? The solutions were 0 . 4 ~ in Na[ClO,]. Measurements were made at 283 K on an Orion 801 digital pH meter. A Beckman 39000 glass electrode was used with a Wilhelm-type bridge reference electrode. Hydrogen-ion activities were divided by 0.683 a to obtain [H+]. To calculate OH- concentrations it was necessary to evaluate K under the conditions of the experi-ments. This was achieved by titrating NaOH against HClO a t I = 0.4~ (Na[ClO,]) and taking points in the region pH 10-11.5; K was found to be 1.22 x Reactions were followed at 280 nm for adenine and ino-sine 290nm for app 270nm for hypoxanthine and 310nm for adenosine on a Durrum D-110 stopped-flow spectropho-tometer at 283 An acetate buffer was used (5 x G.Boivin and M. Zador Bull. Soc. chim. France 1971 12, G. Boivin and M. Zador Canad. J . Chem. 1972 50 3117. G. Boivin and M. Zador Canad. J . Chem. 1973,51 3322. R. L. Karpel K. Kustin and M. A. Wolff J . Phys. Chem., I<. Kustin and M. A. Wolff J.C.S. Dalton 1973 1031. mo12 l-,. 0.5 K. t IM = 1 mol dm-3. 4279. 1971 75 799. ~O-,M) and solutions were made up to a constant ionic strength of 0 . 4 ~ with Na[C104]. Ligand concentrations were 5 x 10-4~ in the reaction mixture. Although acetate ions act as ligands towards NiI1,lO the reaction is too fast to be observed by stopped-flow techniques and the formation constant is relatively small. Test runs in the presence and absence of acetate established that the experimental rate constants were not affected by its presence.Reactions were run under pseudo-first-order conditions with NiII con-centrations in 10-100-fold excess t o ensure that only the 1 1 species was formed. Oscilloscope traces yielded excel-lent first-order rate constants linear for a t least 90% com-pletion of reaction. Values of bobs. used in evaluating further parameters were an average of 3-5 individual determina-tions. RESULTS Protolytic Dissociation Constants.-Both pKBl and pKa, of adenine and app were determined. Values of pK, and pK, of hypoxanthine and inosine were too low and too high respectively to warrant consideration but the PKa, values were determined. The results are listed in Table 1 with available literature values.Adenine aPP The reactions were first order with respect to metal-ion concentration linear plots being obtained of hobs. against [Ni2+]. These plots did not pass through the origin show-ing that the kinetics of complex formation obey the revers-ible mixed second- and first-order rate equation (1) where (1) k f and hd are the observed formation and dissociation rate constants respectively. Adenine and app behaved unusually towards NiII in that, when the Ni2+ concentration was held constant and the pH 6 H. Sternlicht D. E. Jones and K. Kustin J . Amev. Chem. Soc 1968 90 7110. 7 G. G. Hammes and S. A. Levison Biochemistry 1964 8, 1504. 8 R. S. Brundage R. L. Karpel K. Kustin and J. Weisel, Biochem. Biophys. Acta 1972 267 268.H. S. Harned and B. B. Owen ' The Physical Chemistry sf Electrolyte Solutions,' 2nd edn. Reinhold Publishing Co, New York 1950 p. 567. -d [L]~/dt = kf !Ni2+] [LIT - h d [Product] lo R. G. Wilkens Accounts Chem. Res. 1970 3 408 1978 J.C.S. Dalton varied kobs. increased with increasing hydrogen-ion concen-tration. The pH effect indicates that hydrogen ions are released on complex formation and/or that more than one ligand species is reactive towards the metal the fastest TABLE 1 Protolytic dissociation constants Conditions (e,rc, Ligand PK, PK* I/M) Ref. Adenine 4.26 & 0.03 9.90 f 0.02 10 0.4 a 4.22 9.80 25 0.05 11 4.25 9.83 20 b 4.18 9.70 25 0.05 12 "PP 4.57 f 0.03 11.0 & 0.1 10 0.4 a H ypoxanthine 8.76 & 0.04 10 0.4 a 8.83 25 0.05 11 1.98 8.94 20 C Inosine 8.68 f 0.04 10 0.4 a ca.1.5 8.82 20 C b A. Albert and E. P. Serjeant Bzochenz. J., C A. Albert Biocheun. J. 1953 54 646. a This work. 1960 76 621. reacting species being that which increases in concentration as the pH is lowered. I n the pH range over which this study was conducted (6.5-4.2 for adenine and 6.2-4.6 for app) this species is the protonated form of the ligand, H,L+. The coulombic repulsion produced by two positively charged species interacting causes the extent of reaction to decrease considerably decreasing the total optical absorb-ance change and resulting in unfavourable signal to noise ratios; hence any measurements made at pH values below the pKal values of the ligands have a high uncertainty. Nevertheless although it was not possible to study the reac-tions in a region where H,L+ is the dominant species a clear pattern emerged making the following mechanism in which all three ligand species are postulated as reacting to give identical products the most likely.k , k2 ks k i H,L+ + Ni2+ + [NiL]+ + 2H' (2) HL + Ni2+ [NiL]+ + H" (3) L- + Ni2+ [NiL]+ (4) k , k0 In the pH range studied the concentration of L- is The Since (kl[H+l+ k3Ka1) + MH+J2+ k4CH+I +k6 negligible; hence the k path need not be considered. rate equation resulting from this mechanism is (5). [Ni2f] ha1 + [H+l kobs. = ( 5 ) it is likely that k2[H+]2<k4[H+] or k6 this equation reduces t o (6). Plots of against [Ni2+]/(Ka1 + [H']) [Ni2+] kobs. = should be linear with gradient (kl[H+] + k,Kal) and inter-cept (k4[H+] + &)* These constant-pH studies were con-ducted for both ligands and families of straight lines were obtained each family consisting of four lines corresponding t o four pH values.Values of the gradients and intercepts l1 H. Reinert Abh. Deut. Akad. Wiss. Berlin KZ. Med. 1964, 373. f k3Ka1) + k,[H+] + (6) Ka1 + [H+l (Table 2) were then plotted against [H+] to yield k and k , and k and K respectively (Table 3). TABLE 2 Gradients and intercepts obtained by plotting kobs. against [Ni2+]/(Ka1 + [H']) Ligand 105[H+] /M Gradient/s-' Intercept/s-l Adenine 0.583 0.0268 & 0.002 26.8 & 0.3 1.46 0.0390 & 0.002 27.4 -J= 0.4 2.93 0.0608 f 0.003 28.4 & 0.6 4.63 0.0842 0.006 29.5 * 1.1 0.158 0.0147 & 0.005 79.5 f 1.5 82.3 If 1.8 0.583 0.0480 f 0.006 1.46 0.1132 0.007 87.0 & 1.1 2.32 0.1838 f 0.014 92.8 f 2.8 "PP The good fit of the data to the rate equation adds confi-dence to the proposed mechanism.A further test was applied more rigorous in that it encompassed a wider pH range. This consisted of rearranging equation(6) to give a function of kobs. linear in [H+] when the Ni2+ concentration is held constant [equation (7)]. The value of k obtained kobs. ([H+] + Ka1) - k4[H+12 = [H+]([Ni2+]k1 + kdKa1 + (7) kc) + [Ni2+Ik&ai + h6Kai from the [Ni2+]-dependence studies was used to calculate the left-hand side of equation (7) corresponding to each [H+] value. This was then plotted against [H+j. As seen in Figures 1 and 2 the points fall on a straight line as antici-pated.The gradient of the line is ([Ni2+]k1 f k4Ka1 + k6) and the intercept ([Ni2+]k3Kal + k6Ka1). Table 4 shows the experimental gradients and intercepts together with those derived from the data in Table 3. The agreement both for the adenine and app systems is very satisfactory. 901 1 1 5 - L o 0 1 2 3 4 5 105[H+]/M FIGURE 1 Determination of k and k for adenine (0) and aPP (0) Hypoxanthine and Inosine .-Hypoxanthine and inosine were also found to obey the reversible mixed second- and first-order rate equation in their reactions with NiII. How-ever unlike adenine and app no change in kobs. was dis-cerned on varying [H+] over three pH units. This indicates that only one ligand species is reactive towards the metal; this is assumed to be the neutral species (H,L) which is predominant in the range studied.The mechanism proposed is as in (8) for which the rate kl k l H,L + Ni2+ + [Ni(H2L)I2' (8) 13 T. R. Harkins and H. Freiser J . Amer. Clzem. Soc. 1958,80, 1132 1975 1979 TABLE 3 Summary of results obtained on NiT1-adenine and -app studies Equilibrium (a) Adenine ki Si2+ + HL-. [Ni(HL)I2+ kd t Conditions Rate constants Ref. pH 4.2 1 = 1.224 (Na[ClO,]) (zl 8 "C (lij 31 oc pH 4.9-6 1 = 0 . 1 ~ (K[NO,]) 25 "C kf = 300 1 mol-l s-l kd = lo7 1 mo1-l s-l 4 pH 4.2-6.5 I = 0.491 (Na[ClO,]) 10 "C k = (1.44 &- 0.08) x lo3 1 mol-l s-l This work. k = pH 4.6-6.2 I = 0 . 4 ~ (Na[ClO,]) 10 "C k = k = k = TABLE 4 k = 336 f. 30 1 mol-l s-l k = (6.64 & 1.5) x lo4 1 mol-l s-l 26.44 & 0.2 s-l (7.67 & 0.4) x lo3 1 mol-'s-' This work 96.6 & 50 1 mol-' SO k = (6.09 & 78.5 & 1.3 s-l 0.6) x lo5 1 mol-l s-l Comparison of observed gradient and intercept obtained on plotting {Kob,.([Hf] +- Kal) - hJH+l2] against [H+] and those predicted using equation 7 and Figures 1 and 2 Gradientls-1 lo3 Intercept$ r n o l - ' ~ - ~ h r \ P > 1.68 & 0.04 Adenine 52.0 f 2.0 47.8 & 2.0 1.7 -J= 0.1 aPP (2.23 f.0.1) x lo2 (2.05 & 0.12) x lo2 2.3 f. 0.15 2.15 & 0.05 Ligand obs. calc." obs . ca1c.b a Gradient = [Ni2+]kl f k4Ka1 + k,. In order to determine K and h, Kobs. equation (9) applies. Kobs. = k,[Ni2+] + K (9) was measured as a function of nietal-ion concentration. Linear plots were obtained and from the gradients and n 30 t / 495 - 85 I I I 175 0 1 2 3 4 .5 105[H+]/M FIGURE 2 Determination of k4 and k for adenine (0) and aPP (0) intercepts respectively the calculated rate constants are : hypoxanthine K = (1.90 & 0.1) x lo3 1 mol-l s-l K = 32.1 f 2.0 s-1; inosine K = (1.12 0.10) x 103 1 mol-1 s-l k = 59.5 f 1.5 s - ~ . DISCUSSION Adenine and @+.-The mechanism postulated here differs from those proposed previously l p 4 where only the l3 E. Sletten Acta Cryst. 1969 B25 1480. l4 E. Sletten Acta Cryst. 1970 B26 1609. l5 D. M. L. Goodgame and K. A. Price Nature 1966 220 783. Intercept = [Ni2+]k,Ka1 + k,Kal. neutral ligand HL was the reactive species. The results are summarized in Table 3. Although the experimental conditions vary widely the previous results can be ques-tioned.Boivin and Zador found no evidence for reac-tion at pH values greater than 5 where the neutral species predominates and yet propose this as the reacting species.l Kustin4 quotes a rather high experimental deviation of &25% on relaxation data because of ligand instability. However under the conditions of our reactions the ligand spectrum remains unchanged over several days. The poor fit arises because the protonated ligand species also reacts with NiII followed by deprotonation of the inter-mediate species. It is interesting to note the preference of NiII for H2L+ rather than HL. The reaction with HL is unusually slow being about an order of magnitude less than that for ' normal ' substitution.This is not a tem-perature effect as Kustin found anomalously slow be-haviour a t room temperature. We suggest that the neutral ligand initially forms a hydrogen bond from N(l) to a water molecule in the metal's inner co-ordination sphere "(1) is involved in hydrogen bonding between adenine and thymine in the Watson-Crick model of base pairing]. Metal binding at N(l) seems unlikely in adenine; none of the studies in solution or in the solid state have indicated an N(1)-metal co-ordinate bond.11-19 This misorientation caused l6 R. Weiss and H. Venner 2. Physiol. Chem. 1963 333 169. l7 E. Sletten Chem. Comm. 1971 558. l8 M. M. Taqui Kahn and C. R. Krishnamoorthy J . Inorg. lS L. Srinivasan and M. R. Taylor Chem. Comm. 1970 1668. Nuclear Chem.1971 33 1417 J.C.S. Dalton by hydrogen bonding slows the reaction. If N(1) is protonated20 in H,L+ hydrogen bonding cannot occur and ' normal ' substitution is observed The appNiII system is rationalized in the same way. The differences in k, and k for the two ligands results from N(1) in app being more basic resulting in a stronger hydrogen-bonding effect. The binding site on adenine and app is proposed to be N(9). This may explain why the NiII-adenosine reac-tion was too fast to be measured on the stopped-flow apparatus as N(9) is blocked and anothei mechanism must be operative. The formation rate constant for this reaction is reported to be about five times greater than for adenine. Most crystal-structure studies show N(9) and N(3) as binding sites but this may not be so in solution.I t seems unlikely that a chelate structure involving N(7) and the amino-group occurs because of the large value of k,. This should be much smaller if an SCS (sterically controlled substitution) mechanism were operative. The steric strain involved in distorting the orbitals would almost certainly slow down the reaction. Hypoxanthine and Inosine.-Both ligands gave forma-tion rate constants (k,) with NiII within the range of ' normal ' substitution with a neutral ligand. attempted to study the hypoxanthine-NiII system but observed non-exponential relaxation traces and suggested that this might be due to linkage isomerism. The excel-lent exponential traces observed in this study indicate that if linkage isomers do coexist the various binding sites on the ligand must have either near-identical affinities for the metal ion or their affinities are so different only one effect can be observed. We thank the University of Cape Town and the Council for Scientific and Industrial Research for the award of research grants and study bursaries. Kustin [4/2091 Received 8th October 19741 2o R. M. Izatt J. J. Christensen and J. H. Rytting Clzem. Rev. 1971 71 439