J . Chem. Soc., Faraday Trans. I , 1984,80, 2439-2444 Effect of Alkyl Substituents on the Thermodynamics of the Self-association of Purine in Aqueous Solution BY HARRI LONNBERG,* JYRKI YLIKOSKI AND ANTTI VESALA Department of Chemistry and Biochemistry, University of Turku, SF-20500 Turku, Finland Received 12th September, 1983 The thermodynamic parameters for the self-association of several alkyl-substituted purines have been determined by applying the isodesmic model to the molar enthalpies of infinite dilution at different initial concentrations. Carbon-bonded alkyl groups have been shown to increase the tendency to association by making the association enthalpy more negative. In contrast, the increasing size of the N9-bonded alkyl groups appears to reduce the negative entropy contribution and thus stabilize the associated complexes.These findings are discussed in terms of dipole-induced-dipole interactions and hydrophobic bonding. Application of several experimental techniques, including o~mometry,~-~ calori- me try,8-12 ultrasonics, l3 9 l4 n.m. r . s p e c t r ~ s c o p y ~ ~ - ~ ~ and equilibrium sedimen t a t i ~ n , ~ ~ - ~ ~ has indicated that nucleic acid bases and nucleosides self-associate to a large extent in aqueous solutions. Usually the tendency to association is stronger with purines than with pyrimidines.l? 16$ 2 2 v 24 According to the widely adopted isodesmic model of Ts'O and c o ~ o r k e r s , ~ - ~ vertical stacking of several molecules occurs in a non-cooperative manner, although some evidence of both positive and negative cooperativity has also been reported.2v 26* 27 Undoubtedly the association goes far beyond the dimer stage.2q 4 9 26 Solvent water plays a decisive role in the stacking phenomenon.For example, addition of organic solvents to aqueous solutions results in marked destacking.ll9 1 5 7 2o Under non-aqueous conditions base stacking is not observed and base pairing by hydrogen bonding becomes the main interaction m e ~ h a n i s m . ~ ~ - ~ ~ The enthalpy, entropy and volume changes observed for the self-association in aqueous solutions are all negati~e.~-l~~ 1 7 - 1 9 9 2 3 7 2 5 9 3 2 9 33 This contrasts with the classical concept of hydrophobic interaction^,^^-^^ which are characterized by small or even positive enthalpy changes and positive entropy and volume changes.However, alkyl substi- tuents have been shown to enhance the association of purines, pyrimidines and their derivatives. 2-47 6 v 7 v l9 Accordingly, dipole-induced-dipole interactions have been suggested as being responsible for the while the hydrophobic contribution remains 39 The present paper reports our studies with alkyl-substituted purines. The equilibrium constants and the enthalpy and entropy changes for self-association have been calculated from the concentration dependences of the calorimetrically determined dilution enthalpies. The effects of the size and site of the alkyl groups on the thermodynamic quantities have been examined to elucidate the possible contribution of hydrophobic interactions in the base stacking. 24392440 THERMODYNAMICS OF THE SELF-ASSOCIATION OF PURINE EXPERIMENTAL MATERIALS Purine was a commercial product of Sigma Chemical Co.9-Methyl-, 6,9-dimethyl- and 8,9-dimethyl-purine were prepared and characterized as described 41 9-Ethylpurine was synthesized and separated from the corresponding N7 isomer using the method reported for 9-methylpurine40 with diethylsulphate as alkylating agent. The product was purified by sublimation at 130 "C at a pressure of mmHg. lH n.m.r. (ppm from TMS in CDCl,): 6 1.81 (CH3CH2-, t), 6 4.40 (CH3CH2-, q), 6 8.03 (H8, s), 6 8.92 (H2, s), 6 9.08 (H6, s). 13C n.m.r.: 6 152.4 (C2). Calculated for C,HsN4: C 56.74%, H 5.44% ; found: C 56.67%, H 5.56%. 9-Isopropylpurine was prepared using the alkylation method described for adenine.42 Purine (30 mmol) was converted into sodium purinide using sodium hydride (31.5 mmol) in DNF (300 cm3).Isopropyliodide (30 mmol) was added and the reaction mixture was agitated for 16 h at 30 "C in the dark. The N9 isomer was isolated chromatographically as described previo~sly.~~ The product was sublimed at 100 "C at a pressure of mmHg. lH n.m.r: 6 1.68 [(CH3),CH-, d], 6 4.97 [(CH,),CH-, m], 6 8.20 (HS, s), 6 8.73 (H2, s), 6 9.12 (H6, s). 13C n.m.r.: 6 22.5 [(CH,),CH-], 6 47.4 [(CH,),CH-1, 6 134.5 (C5), 6 143.1 (C6), 6 148.5 (C8), 6 151.1 (C4), 6 152.2 (C2). Calculated for CsHloN4: C 59.24%, H 6.21 % ; found: C 59.23%, H 6.22%. 6,8,9-Trimethylpurine was prepared as follows. 4,5-Diamino-6-methyl-2-thiopyrimidine (7 mmol), purchased from Sigma, was reduced to 4,5-diamino-6-methylpyrimidine with Raney nickel in boiling water (50cm3).The product was cyclized to 6,8-dimethylpurine by heating with acetic anhydride43 and methylated to 6,8,9-trimethylpurine as described for 9-methylp~rine.~~ The syrup obtained was crystallized from boiling n-hexane. lH n.m.r. : 6 2.67 (C8-CH3, s), 6 2.83 (C6-CH,, s), 6 3.80 (N9-CH,, s), 6 8.73 (H2, s). 13C n.m.r.: 6 14.1 6 153.4 (C8), 6 157.0 (C6). Calculated for C,H,,N,: C 59.23%, H 6.22% ; found: C 57.90%, H 6.22%. All the compounds prepared were assigned as N9 isomers on the basis of the 13C n.m.r. chemical shfts, using the method of Chenon et ~ 1 . ~ ~ The homogeneity of the products was checked by liquid chromatography with a TSK OSD5 column. Eluation (0.8 cm3 min-l) was performed with an acetic acid buffer (0.02 mol dm-3, buffer ratio 1 : 1) containing 20 vol.% acetonitrile. Distilled and degassed water was used for the calorimetric measurements. 6 15.3 (CH3CH2-, 6 38.9 (CH,CH-), 6 134.2 (C5), 6 145.0 (C6), 6 148.4 (C8), 6 151.3 (C4), (C8-CH3), 6 19.3 (C6-CH3), 6 28.7 (N9-CH3), 6 132.3 (C5), 6 151.4 (C2), 6 152.1 (C4), TREATMENT OF THE CALORIMETRIC DATA The dilution enthalpies for the alkyl-substituted purines were measured on a LKB 10700-2 batch microcalorimeter by mixing equal volumes (1.60 cm3) of aqueous purine solutions with water at 298.15 K. The reference cells were filled with equal volumes of water. The molar enthalpies of infinite dilution. AH,' d i l , were obtained by summing the molar enthalpy changes for successive dilutions from the concentration Ci to the concentration AHdil from the lowest experimentally accessible concentration to infinite dilution was estimated by linear extrapolation.The above method of obtaining the values of AH,,,il is based on the assumption that the solutes behave ideally, i.e. no volume changes take place on dilution. Previous data on the densities of aqueous solutions of 6-methylpurines lend support to this assumption. If it is further assumed that the association of monomeric purines proceeds to an infinite degree and that the association constants, K, and the enthalpy changes, AH*, for the successive steps are equal, the dependence of AHi,dil on Ci can then be expressed as AHe[ 1 + 2Ci K - (1 + 4Ci K)'] 2Ci K AHi,dil = Consequently, a two-parameter fitting procedure can be used to obtain K and AHe.Several earlier investigation^^-^. 45 show that the above assumptions are reasonable.H. LONNBERG, J. YLIKOSKI AND A. VESALA 2441 Table 1. Molar enthalpies of infinite dilution, AH,, dil, for some alkyl-substituted purines in aqueous solutions at 298.15 Ka AHi, dil/kJ mol-l 9- 9- 9- 8,9- 6,9- 6,8,9- c y 1 0 - 3 methyl- ethyl- isopropyl- dimethyl- dimethyl- trimethyl- mol dm-3 purine purine purine purine purine purine purine 1000 500 250 125 62.5 31.3 15.6 7.81 3.91 - 7.643 (- 7.680) -6.181 (- 6.159) -4.628 (-4.576) -3.150 ( - 3.1 12) - 1.888 (- 1.934) - 1.050 (- 1.1 10) - 0.530 (- 0.603) - - - - - 6.224 (- 6.243) - 5.090 (- 5.086) - 3.882 (- 3.853) -2.713 (- 2.678) - 1.679 (- 1.699) - 0.957 ( - 0.992) - 0.480 (- 0.545) - - - - - 5.538 (- 5.601) - 4.740 (-4.688) - 3.762 (- 3.675) -2.606 (- 2.657) - 1.739 (- 1.754) - 1.044 (- 1.061) -0.560 (- 0.597) - - - - - - - 3.956 ( - 4.006) - 3.247 ( - 3.1 82) - 2.372 (- 2.336) - 1.538 (- 1.568) - 0.944 (- 0.962) -0.500 (-0.548) - - - - - - - - - - -7.691 - 11.693 - (- 7.756) (- 11.769) - -6.336 - 10.145 - 12.792 (-6.292) (- 10.093) (- 12.820) -4.837 -8.221 - 10.996 (-4.741) (-8.165) (-10.968) -3.286 -6.176 -8.882 (-3.275) (-6.132) (-8.845) - 1.998 -4.206 -6.588 ( - 2.065) ( - 4.2 1 9) ( - 6.6 1 7) - 1.100 -2.589 -4.495 (- 1.201) (-2.652) (-4.534) - - 1.460 - 2.884 - (-1.537) (-2.837) - - - 1.620 - - (- 1.639) a Means of duplicate measurements.The values in parentheses are those calculated via eqn (1) using the values of K and A H e listed in table 2.RESULTS AND DISCUSSION Table 1 summarizes the molar enthalpies of infinite dilution for the alkyl purines investigated. Comparison with the values obtained by least-squares fitting indicates that eqn (1) satisfactorily describes the dependence of AHi. dil on Ci. The association constants, K, and the enthalpy changes, A@, which application of this isodesmic model yields are collected in table 2. Gill et aL9 have reported for purine the values of 2.9 & 0.2 kg mol-1 and - 15.5 f 0.8 kJ mol-1 for Kand AH*, respectively from calorimetry. The association constant of 2.80 k 0.06 dm3 mol-l obtained by equilibrium-sedimentation methodsz3 is of the same magnitude, while the concentration dependence of the molal osmotic coefficient has led to a lower value of 2.1 kg mol-l.l The osmometrically determined association constants for 9-methyl-, 9-ethyl- and 9-isopropyl-purine are also markedly lower than those obtained in the present Note that differences in the association constants measured by different experimental techniques are quite common.For example, the value for caffeine obtained calor- imetrically is 15 k 1 kg m ~ l - ' , ~ whereas n.m.r. spectr~scopy~~ and give 9.0 and 8.2 dm3 mol-l, respectively, at 298.15 K. With 6-dirnethylamino-9-@-~- ribofuranosyl) purine values of 22.2 kg mol-l, 34 dm3 mol-1 and 33.8 dm3 mol-1 have been observed at 298.15 K using osmometry,3 and equilibrium ~edimentation,~~ respectively. These differences are far greater than the experimental errors and possibly reflect the different sensitivities of the techniques to the basic2442 THERMODYNAMICS OF THE SELF-ASSOCIATION OF PURINE Table 2.Equilibrium constants and the enthalpy and entropy changes for the self-association of some alkyl-substituted purines in aqueous solutions at 298.15 Ka ~~ compound K/dm3 mol-l AHe/kJ mol-l A F / J K-l mol-l purine 3.1 fO.l - 13.42 f 0.20 - 35.6 f 1 .O 9-methylpurine 3.7 f 0.1 -10.43L0.12 - 24.1 f 0.7 9-e thylpurine 5.1 f0.3 -8.71 f0.16 - 15.7f 1.3 9-isoprop ylpurine 5.7 f 0.4 - 7.18 k 0.20 -9.6+ 1.7 8,9-dime thylpurine 7.1 f0.4 - 13.10&0.28 - 27.6 f 1.8 6,9-dimethylpurine 13.8L0.4 - 17.18k0.15 -35.8f 1.2 6,8,9-trimethylpurine 26.7 f 0.4 - 18.83 k 0.09 - 35.8 f 0.7 a Calculated from eqn (1) by least-squares fitting. assumptions of the isodesmic model.It appears reasonable to assume, however, that reliable conclusions concerning the influence of the solute structure on the thermodynamics of the association reaction can be drawn if data referring to a single experimental method are considered. The data in table 2 reveal that the associations of all the purines investigated are enthalpy driven and entropy opposed, as shown previously for a great variety of purine and pyrimidine derivati~es.~-l~$ 17-197 23 With the exception of unsubstituted purine, none of the compounds contains acidic hydrogen atoms. Accordingly, hydrogen bonding can be excluded as a possible explanation for the markedly different association abilities. It has been suggested previously38 that the dipole-induced-dipole interactions play a dominant role in the stacking of monomeric nucleic acid constituents.In other words, both the polarizing power of the polar bonds and the polarizability of the n-electron system would contribute to the stability of the associated complexes, the latter factor being more important with purine derivative^.^* 38 Part of the present results can be accounted for by this mode of interaction. The tendency to association of alkyl-substituted purines is markedly increased on going from 9-methylpurine to 8,9-dimethyl-, 6,9-dimethyl- and 6,8,9-trimethyl purine. The main reason for the stabilization of the stacks appears to be the increasingly negative enthalpy change. Part of this influence is cancelled out by the parallel changes in the entropy term, T A P . Since carbon-bonded alkyl groups are known to increase the polarizability of the purine bases, it seems possible that methyl substituents at C6 and C8 may enforce the dipole-induced-dipole interactions and thus make the association enthalpy more negative.For example, the polarizabili ties of purine and 6-methylpurine have been reported as 12.5 and 14.3 A3, re~pectively.~~ It remains obscure as to why the methyl group at C6 exerts a greater effect than the methyl group at C8. The suggested structures of the purine stacks3 offer a tentative explanation. Examination of the concentration dependences of the lH n.m.r. chemical shifts has led to the conclusion that the six-membered rings of purine nucleosides overlap in the associated complexes to a larger extent than the five-membered rings.3 Substituents in the overlapping part of the aromatic ring system may have a stronger influence on the interactions which cause the stacking phenomenon.The thermodynamic data referring to the association of 9-methyl-, 9-ethyl- and 9-isopropyl-purine are difficult to understand on the basis of the dipole-induced-dipole interactions. The ability to self-associate is again increased with increasing numberH. LONNBERG, J. YLIKOSKI AND A. VESALA 2443 of aliphatic carbon atoms, but with this series of compounds changes in the entropy term appear to be responsible for the stabilization of the associated species. Both the enthalpy and the entropy terms become less negative with increasing size of the N9 substituent, but changes in the latter are more marked.This finding is consistent with the classical concept of hydrophobic b ~ n d i n g . ~ ~ - ~ ~ Hydrophobic interactions between the N9-bonded alkyl groups may be expected to result in positive enthalpy and entropy contributions to the overall association enthalpy and entropy. On the basis of the present data these contributions appear to be 1-2 kJ mol-1 and 5-10 J K-l mol-1 for an additional methylene group. For comparison, AH* for the interaction between two methylene groups in long hydrocarbon chains has been suggested to be ca. 5 kJ mol-1 at 298.15 K and A S e is suggested to be ca. 20 J K-l m01-l.~~ Comparable evidence for the involvement of hydrophobic interactions in base stacking has been obtained from studies with alkyl-substituted uracils.’ Replacement of the N9 hydrogen atom of purine with a methyl group also exerts positive changes in both AH* and A P , the effect on AG* remaining relatively small.Analysing these influences is, however, difficult, since alkylation of the N9 site results in changes in the tautomeric composition, polarizing power, hydrogen-bonding ability and hydrophobic nature of the solute. In summary, the above discussion suggests that, besides dipole-induced-dipole interactions, hydrophobic bonding may contribute to the base stacking of purines. We thank the Academy of Finland, Council for the Natural Sciences, for financial support. P. 0. P. Ts’O, I. S. Melvin and A. C. Olson, J. Am. Chem. SOC., 1963,85, 1289. P. 0. P. Ts’O and S. I. Chan, J. Am. Chem. SOC., 1964,86,4176. A. D. Broom, M.P. Schweizer and P. 0. P. Ts’O, J. Am. Chem. SOC., 1967, 89, 3612. G. K. Helmkamp and N. S. Kondo, Biochim. Biophys. Acta, 1968, 157, 242. D. Porschke and F. Eggers, Eur. J. Biochem., 1972, 26, 490. E. Plesiewicz, E. Stepien, K. Bolewska and K. L. Wierzchowski, Nucleic Acids Res., 1976, 3, 1295. E. Plesiewicz, E. Stepien, K. Bolewska and K. L. Wierzchowski, Biophys. Chem., 1976, 4, 131. * P. R. Stoesser and S. J. Gill, J. Phys. Chem., 1967, 71, 564. S. J. Gill, M. Downing and G. F. Sheats, Biochemistry, 1967, 6, 272. lo E. L. Farquhar, M. Downing and S. J. Gill, Biochemistry, 1968, 7, 1224. l1 M. G. Marenchic and J. M. Sturtevant, J. Phys. Chem., 1973,77, 544. l2 A. Zielenkiewicz, E. Plesiewicz and K. L. Wiezchowski, Biophys. Chem., 1979, 10, 415. l3 M. P. Heyn, C.U. Nicola and G. Schwarz, J. Phys. Chem., 1977,81, 161 1. l4 P. Hemmes, A. A. Mayevski, V. A. Buckin and A. P. Sarvazyan, J. Phys. Chem., 1980, 84, 699. l5 S. I. Chan, M. P. Schweizer, P. 0. P. Ts’O and G. K. Helmkamp, J. Am. Chem. SOC., 1964,86,4182. l7 V. L. Antonovsky, A. S. Gukovskaja, G. V. Nekrasova, B. I. SukhorukovandI. I. Tchervin, Biochim. lCI W. Schimmak, H. Sapper and W. Lohman, Biophys. Struct. Mechanism, 1975, 1, 113. lCI W. Schimmak, H. Sapper and W. Lohman, Biophys. Struct. Mechanism, 1975, 1, 31 1. *O H. Sapper and W. Lohman, Biophys. Struct. Mechanism, 1978, 4, 327. *l D. M. Cheng, L. S. Kan, P. 0. P. Ts’O, C. Giessner-Prettre and B. Pullman, J. Am. Chem. SOC., 1980, 22 K. H. Scheller, F. Hofstetter, P. R. Mitchell, B. Prijs and H. Sigel, J.Am. Chem. SOC., 1981, 103, 247. 23 K. E. Van Holde and G. P. Rossetti, Biochemistry, 1967, 6, 2189. 24 T. N. Solie and J. A. Schellman, J. Mol. Biol., 1968, 33, 61. 26 R. Bretz, A. Lustig and G. Schwarz, Biophys. Chem., 1974, 1, 237. 26 F. Garland and S. D. Christian, J. Phys. Chem., 1975, 79, 1247. 27 J. Gajewska, A. Bierzynski, K. Bolewska, K. L. Wierzchowski, A. I. Petrov and B. I. Sikhorukov, 28 J. Pitha, R. N. Jones and P. Pithova, Can. J. Chem., 1966, 44, 1945. 29 Y. Kyogoko, R. C. Lord and A. Ritch, J. Am. Chem. SOC., 1967,89,496. M. P. Schweizer, S. I. Chan and P. 0. P. Ts’O, J. Am. Chem. SOC., 1965,87, 5241. Biophys. Acta, 1973, 331, 9. 102, 525. Biophys. Chem., 1982, 15, 191.2444 THERMODYNAMICS OF THE SELF-ASSOCIATION OF PURINE 30 K. A. Newmark and C. R. Cantor, J. Am. Chem. SOC., 1968,90, 5010. 31 G. G. Hammes and A. C. Park, J . Am. Chem. SOC., 1968,90,4151. 32 U. Gaarz and H. D. Liidemann, Ber. Bunsenges. Phys. Chem., 1976,80, 607. 33 D. D. Kasarda, Biochim. Biophys. Acta, 1970, 217, 535. 34 H. S. Frank and M. W. Evans, J . Phys. Chem., 1945, 13, 507. 35 W. Kauzmann, Adv. Protein Chem., 1959, 14, 1. 36 G. Nemethy and H. A. Scheraga, J . Chem. Phys., 1962, 36, 3401. 37 G. Nemethy and H. A. Scheraga, J. Phys. Chem., 1962, 66, 1773. 38 R. Lawaczeck and K. G. Wagner, Biopolymers, 1974, 13, 2003. 39 D. M. Crothers and D. I. Ratner, Biochemistry, 1968, 7, 1823. 40 J. Arpalahti and H. Lonnberg, Acta Chem. Scand., Ser. B, 1982, 36, 545. 41 J. Arpalahti and H. Lonnberg, Znorg. Chim. Acta, 1983, 78, 63. 42 M. Rasmussen and J. M. Hope, Aust. J. Chem., 1982, 35, 525. 43 A. Albert and D. J. Brown, J. Chem. SOC., 1954,2060. 44 M. T. Chenon, R. J. Pugmire, D. M. Grant, R. P. Panzica and L. B. Townsend, J . Am. Chem. SOC., 45 L. P. Vickers and G. K. Ackers, Arch. Biochem. Biophys., 1976, 174, 747. 46 B. Pullman, J. Chem. Phys., 1965,43, S233. 47 D. G. Oakenfull and D. E. Fenwick, Aust. J . Chem., 1977, 30, 741. 1975,97,4627. (PAPER 3/1601)