THEO CHEM ELSEVIER Journal of Molecular Structure (Theochem)466(1999)49-58 Structure and properties of hydroxyl radical modified nucleic acid components:tautomerism and miscoding properties of 5-hydroxycytosine Piotr Cysewski Department of Clinical Biochemistry.University School of Medical Sciences in Bydgose-.Karlowica 24.85-092 Bydgos-c Poland Received 17 February 1997:accepted 11 August1998 Abstract vapour state the keto- mino isomer is most sta However,in the presence of a solvent field the most preferred is the keto-amino isomer.Thus,there is significant influence of the environment polarity on the tautomers succession.Such a behaviour is related to discrepancies in the dipole moments between the two most stable tautomers of 5-OH-C.The keto-imino form is less polar and has a dipole moment one order less then the keto-amino structure.The pairing potential of 5-OH-C was estimated on the basis of SCF calculations for the two most favourable tautomers.Both structures are able to form stable dimers with guanine,cytosine and to a lesser extent with lead to miscod However,the form pair led to the conclusion that the resulting stabilisation energy is higher for pairs with guanine and cytosine but slightly lower for pairs with thymine and adenine.The presence of one of these odd pairs in the DNA may be responsible for CG=GC and CGAT transversion or CG=TA transitions.These facts are in good agreement with in vivo and in vitro experimental observations.In the light of our results the source of such mutations may be related not only to the miscoding potential of dominant tautomer,but also to other,less stable tautomers of 5-OH-C.1999 Elsevier Science B.V.All rights reserved. Keywords:Ab initio,Tautomerism;Solvation,Mispairing.5-hydroxycytosine 1.Introduction mass spectrometry with the selected-ion-monitoring) technique developed by group [1]. Free radical-induced damage to DNA in vivo is Exp sure of pyrimidin of DNA to ionising radia implicated to play a role in carcinogenesis.Evidence tion under aerobic conditions or oxidising agents exists that DNA damage by endogenous free radicals results in attack on the 5,6 double bond of the pyri- occurs in vivo,and there is a steady-state level of free midine ring or on the exocyclic 5-methyl group.The radical-modified bases in cellular DNA.Elucidation primary product of oxidation of the 5,6 double bond of of the chemical nature of such DNA lesions at biolo- eytosine yields cytosine glycol,which decomposes to gically significant quantities is usually done by the aid xycytosine,5-hydroxyuracil and uracil glycol of a sensitive GC-MS SIM(gas chromatography- [2-5].The presence of stable 5-hydroxycytosine 0166-1280/99/-see front matter1999 Elsevier Science B.V.All rights reserved. Pm:S0166-1280(98)00337-6
Structure and properties of hydroxyl radical modified nucleic acid components: tautomerism and miscoding properties of 5-hydroxycytosine Piotr Cysewski Department of Clinical Biochemistry, University School of Medical Sciences in Bydgoszcz, Karłowicza 24, 85-092 Bydgoszcz, Poland Received 17 February 1997; accepted 11 August 1998 Abstract The SCF ab initio quantum chemistry calculations of the 5-hydroxycytosine (5-OH-C) tautomer and its pairs with standard nucleic acid bases were performed. In the vapour state the keto–imino isomer is most stable among all possible H1 tautomers. However, in the presence of a solvent field the most preferred is the keto–amino isomer. Thus, there is significant influence of the environment polarity on the tautomers succession. Such a behaviour is related to discrepancies in the dipole moments between the two most stable tautomers of 5-OH-C. The keto–imino form is less polar and has a dipole moment one order less then the keto–amino structure. The pairing potential of 5-OH–C was estimated on the basis of SCF calculations for the two most favourable tautomers. Both structures are able to form stable dimers with guanine, cytosine and to a lesser extent with thymine and adenine. The first does not lead to miscoding abilities. However, the formation of all other dimers may be responsible for observed in vivo miscoding properties of this DNA lesion. A comparison of the stabilisation energies to AT pair led to the conclusion that the resulting stabilisation energy is higher for pairs with guanine and cytosine but slightly lower for pairs with thymine and adenine. The presence of one of these odd pairs in the DNA may be responsible for CG ) GC and CG ) AT transversion or CG ) TA transitions. These facts are in good agreement with in vivo and in vitro experimental observations. In the light of our results the source of such mutations may be related not only to the miscoding potential of dominant tautomer, but also to other, less stable tautomers of 5-OH–C. q 1999 Elsevier Science B.V. All rights reserved. Keywords: Ab initio; Tautomerism; Solvation; Mispairing; 5-hydroxycytosine 1. Introduction Free radical-induced damage to DNA in vivo is implicated to play a role in carcinogenesis. Evidence exists that DNA damage by endogenous free radicals occurs in vivo, and there is a steady-state level of free radical-modified bases in cellular DNA. Elucidation of the chemical nature of such DNA lesions at biologically significant quantities is usually done by the aid of a sensitive GC-MS SIM (gas chromatography– mass spectrometry with the selected-ion-monitoring) technique developed by group [1]. Exposure of pyrimidines of DNA to ionising radiation under aerobic conditions or oxidising agents results in attack on the 5,6 double bond of the pyrimidine ring or on the exocyclic 5-methyl group. The primary product of oxidation of the 5,6 double bond of cytosine yields cytosine glycol, which decomposes to 5-hydroxycytosine, 5-hydroxyuracil and uracil glycol [2–5]. The presence of stable 5-hydroxycytosine Journal of Molecular Structure (Theochem) 466 (1999) 49–58 0166-1280/99/$ - see front matter q 1999 Elsevier Science B.V. All rights reserved. PII: S0166-1280(98)00337-6
50 P.Cysewski/Journal of Molecuar Structure (Theochem)466(199)49-58 cancerous s fairl Ho e 5-hydro T181 and s-hydrowuracil for sation of side context-dependent mispairing in vitro [10,11]. especially amino substituents.The right non-pla Purmal et al.studied the miscoding properties of2 geometry of these groups is predicted only if the deoxy-5-hydroxycytidine (5-OHdC). They have polarisation functions are supported.The relative shown that 5-OHdCTP can replace dCTP,and to a stability of different tautomers is usually predicted by the calculation of the ice between 公 e) fragmen of the this correspom onu n nu .5-OHdc in the h ndent In o dG was the redominant nucleotide incornorated can hope however that with this a ach errot opposite 5-OHdC with dA incorporation also are of the same order for all compounds of the same observed.However,in another sequence context,dC ype and will cancel when relative energies are adngloceosne In this paper the full gra I tran wa pe h cytosine the biold e of 5- d MD to. tical inv the final now concerng properties of this cytosine derivative road range of standard Gaussian tyr The aim of this paper is to describe the tautomeric and basis sets were used:starting from 3-21G,though 6 miscoding properties of 5-OH-C in vapour and esti- 31G,6-311G,6-31G**and ending on 6-311G** mate the influence of the solvent field on the tautomer Additionally the importance of solvation effect on ism of this cytosine derivative. the tautomer stabilisation was estimate For thi eason th CM)[22]wa to sim 2.Methods In thi The tautomerism of standard and modified nuclei and is only the source of the filed related to acid bases was described on different levels [12-20] its dielectric constant the calculations were restricted Starting from semiempirical Hamiltonian [12] to only the two most favourable tautomers,for which through density functional theory (DFT)[13,14],the the geometry was optimised in the presence of the Hartree-Fock self consistence field [15,16]and solvent field ending on the post SCF techniques with correlation The misco n into a on The supposed res ons were impos alitat dard 9 as well as free radical.modified 2 211 The di atio n the nucleic acid base tautomerism.The DFT a 321G level was followed by single point eneray esti- which retains the simplicity of the one mation on the basis of the 6-31G**level,in vacuum approximation,is able to estimate most static and and in water solution.Again the SCI-PCM model [22] dynamic contributions of electron correlation was used for estimating the solvation effect.Addition However,the correspondence between results based ally the basis set superposition error (BSSE)wa and S( was not culated on the acid bases. he
product in chromatin of various human cancerous tissues is fairly well documented [2–10]. The major oxidative products of cytosine, 5-hydroxycytosine and 5-hydroxyuracil, exhibit sequence context-dependent mispairing in vitro [10, 11]. Purmal et al. studied the miscoding properties of 2 0- deoxy-5-hydroxycytidine (5-OHdC). They have shown that 5-OHdCTP can replace dCTP, and to a much lesser extent dTTP, as a substrate for Escherichia coli DNA polymerase I Klenow fragment (exonuclease free). The specificity of such nucleotide incorporation opposite 5-OHdC in the template was sequence context dependent. In one sequence context, dG was the predominant nucleotide incorporated opposite 5-OHdC with dA incorporation also observed. However, in another sequence context, dC was the predominant base incorporated opposite 5- OHdC. These data suggest that the 5-hydroxycytosine has the promutagenic potential leading to C ! T transitions and C ! G transversions. Despite the biological significance of 5-hydroxycytosine no theoretical investigations were presented till now concerning properties of this cytosine derivative. The aim of this paper is to describe the tautomeric and miscoding properties of 5-OH–C in vapour and estimate the influence of the solvent field on the tautomerism of this cytosine derivative. 2. Methods The tautomerism of standard and modified nucleic acid bases was described on different levels [12–20]. Starting from semiempirical Hamiltonian [12], through density functional theory (DFT) [13,14], the Hartree–Fock self consistence field [15,16] and ending on the post SCF techniques with correlation effects taken into account [17]. The semiempirical methods are supposed to be accurate enough for a qualitative description of standard [18,19], as well as free radical, modified [20,21] nucleic acid base tautomerism. The DFT approach, which retains the simplicity of the one-particle approximation, is able to estimate most static and dynamic contributions of electron correlation. However, the correspondence between results based on DFT and SCF methods was not stated till now for modified nucleic acid bases. The SCF method provides a reliable tool for prediction of the molecular properties. However, as it was emphasised in the literature [18], the polarisation functions are crucial for correct geometry optimisation of side groups, especially amino substituents. The right non-planar geometry of these groups is predicted only if the polarisation functions are supported. The relative stability of different tautomers is usually predicted by the calculation of the difference between total energy of the tautomers. Since this corresponds to the difference of two large and nearly equal numbers, the errors involved in the energy estimations may have significant influence on the final result. One can hope, however, that with this approach errors are of the same order for all compounds of the same type and will cancel when relative energies are considered. In this paper the full gradient geometry optimisation was performed for the six most preferred H1 tautomers of 5-OH-cytosine. The HF SCF method was used to estimate the optimal geometry and MP2 approximations were applied for finding the final energy. The broad range of standard Gaussian type basis sets were used: starting from 3-21G, though 6- 31G, 6-311G, 6-31G** and ending on 6-311G**. Additionally the importance of solvation effect on the tautomers’ stabilisation was estimated. For this reason the self-consistent isodensity polarised continuum model (SCI-PCM) [22] was chosen to simulate the water, methanol, acetone and cyclohexane solutions. In this model the solvent is not present explicitly and is only the source of the continuum filed related to its dielectric constant. The calculations were restricted to only the two most favourable tautomers, for which the geometry was optimised in the presence of the solvent field. The miscoding properties of the two most stable tautomers of 5-OH-C were studied on the basis of the HF SCF method. No restrictions were imposed on the pairs’ geometry during the optimisation procedure. The full gradient minimisation performed on the 3-21G level was followed by single point energy estimation on the basis of the 6-31G** level, in vacuum and in water solution. Again the SCI-PCM model [22] was used for estimating the solvation effect. Additionally the basis set superposition error (BSSE) was calculated on the basis of the counterpoise method (CP) proposed by Boys and Bernardi [23]. 50 P. Cysewski / Journal of Molecular Structure (Theochem) 466 (1999) 49–58
P.Cysewski/Journal of Molecular Smructure (Theochem)466(1999)49-58 深 ndicate the potential rotation of side groups.Three classes The gradient convergence criterion for all sce enol-keto and/or imino-amino tautomerisation.The geometry calculations was set to o0005 The calcula- possible structures of this cvtosine derivative are tions were performed on the basis of Gaussian [24] schematically presented in Fig.1.The three classes and Gamess [25]programs. comprise all possible tautomers having a hydrogen atom attached to NI nitrogen.Only such a tautomeric 3.Results and discussion os nyar ding on 3.1.Tautomeric properties of 5-hydroxycytosine Corm In to des cribe the this side group the energy changes related to OHo The 5-hydroxycytosine may potentially undergo rotation were estimated.The relative energies with 20.0 16.0 120 -lla1 .装.b1 女一13a 2.0 180 -420 -60 120 180 The value of the H5-05-C5-C4 dihedral angle Fig.2.The calculated changes in the total e of che ion of the of Hs-O:-C-Ca dihedral angle.The singl poimtcaleulationscrefmd ed for eac lue of th angle on the
The gradient convergence criterion for all SCF geometry calculations was set to 0.0005. The calculations were performed on the basis of Gaussian [24] and Gamess [25] programs. 3. Results and discussion 3.1. Tautomeric properties of 5-hydroxycytosine The 5-hydroxycytosine may potentially undergo enol–keto and/or imino–amino tautomerisation. The possible structures of this cytosine derivative are schematically presented in Fig. 1. The three classes comprise all possible tautomers having a hydrogen atom attached to N1 nitrogen. Only such a tautomeric form may appear in the nucleosides and DNA. The O5HO5 hydroxyl group is freely rotable and may adopt different orientations, depending on the tautomeric form. In order to describe the potential behaviour of this side group the energy changes related to O5HO5 rotation were estimated. The relative energies with P. Cysewski / Journal of Molecular Structure (Theochem) 466 (1999) 49–58 51 Fig. 1. The structures of all possible H1 tautomers of 5-hydroxy-cytosine. The arrows indicate the potential rotation of side groups. Three classes of structures correspond to the keto–amino (I), keto–imino (II) and enol–imino (III) tautomers. Fig. 2. The calculated changes in the total energy of chosen tautomers as a function of the value of H5–O5–C5–C4 dihedral angle. The single point calculations were performed for each value of the torsion angle on the basis of the HF SCF (6-13G**) method. The other geometrical parameters were taken from full gradient optimisation of a given tautomer. The reference point corresponds to the most stable tautomer IIa1 in vacuum
52 P.Cysewski/Journal of Molecular Structure (Theochem)466(199)49-58 mand the fomation ofal hd sults of the full gradient metr Symbol H(keal/mol) μ(D comprises the PM3 derived heat of formation for all 55.d -54. possible H tautomers of 5-OH-C.The results allow us to omit from further analysis all structures belonging -48 036161 to class III.The values of their heat of formation are 2 uch higher compared tolected and analysed in detail analysed in de ted in Table 2 It esulting energes there PM3 and SCE mers I1 and 12 were the most stable.but the HF SCF and MP2 calculations of the isolated 5-OH-C mole- cule led to the conclusion that tautomer llal is char acterised by the energy tor all basis h timal Il stru able.Thus.l and out set for all intermediate structures.From Fig 2 it is evident that two rotamers are to be considered for optimised geometry of tautomers Ilal and Il are amino tautomers and those imino-ones,which have presented in Fig.3.The tautomer Ilal is flat with HN41 hydrogen atom in the b position (see Fig coplanar H and Hos atoms.In contrast,the tautomer equal in the energy,whic has th Hos hydrogen ator by the full gradient optimis plane with lue of on angle 4-H only the amin group.Thesed e pyra the Table 2 Energies 4648513 4.8499 464.855 164.8379 61.849 464.8502 467 -467 167.350 .46761 -467. ,467.479 -467.3 -467.4 -46902048 -A 00 _A600188 MP246-311G*
respect to the optimal I1 structure are presented in Fig. 2. The presented plots were obtained as a result of the single point energy evaluation in the 6-21G** basis set for all intermediate structures. From Fig. 2, it is evident that two rotamers are to be considered for amino tautomers and those imino-ones, which have HN41 hydrogen atom in the ‘‘b’’ position (see Fig. 1). These isomers are not equal in the energy, which was confirmed by the full gradient optimisation. On the contrary, the imino tautomers, having the HN41 hydrogen atom in the ‘‘a’’ position, are able to form only one rotamer due to the interactions between N4 and HO5 atoms and the formation of an internal hydrogen bond. The results of the full gradient geometry optimisations are collected in Tables 1 and 2. The first one comprises the PM3 derived heat of formation for all possible H1 tautomers of 5-OH-C. The results allow us to omit from further analysis all structures belonging to class III. The values of their heat of formation are much higher compared to classes I and II. The rest of the six tautomers were selected and analysed in detail by HF SCF method. The resulting energies are presented in Table 2. It is interesting to note that there is a discrepancy between PM3 and SCF predictions. In the semiempirical approximation, the tautomers I1 and I2 were the most stable, but the HF SCF and MP2 calculations of the isolated 5-OH-C molecule led to the conclusion that tautomer IIa1 is characterised by the lowest energy for all basis sets. The next tautomers, I1 and I2, are about 3 kcal/mol less stable. Thus, in light of the ab initio calculations, one may conclude that in the vacuum the C5 modified cytosine may exist as a mixture of at least two tautomers: keto–imino, IIa1, and keto–amino, I1 or I2. The optimised geometry of tautomers IIa1 and I1 are presented in Fig. 3. The tautomer IIa1 is flat with coplanar H41 and HO5 atoms. In contrast, the tautomer I1 has the HO5 hydrogen atom oriented above the ring plane with an almost perpendicular bond HO5–O5. The value of the improper torsion angle: H41–N4–H42– C4 2 170.48 indicates the pyramidal character of the amino group. These differences in the geometry of 52 P. Cysewski / Journal of Molecular Structure (Theochem) 466 (1999) 49–58 Table 1 Results of the semiempirical PM3 geometry optimisation of all studied tautomers of 5-OH-cytosine Symbol Hf (kcal/mol) m (D) I1 2 55.0 4.8 I2 2 54.6 5.9 IIa1 2 52.5 1.0 IIa2 2 48.7 2.3 IIb1 2 53.1 4.6 IIb2 2 52.0 3.1 III13a 2 45.2 2.6 III13b 2 38.8 3.1 III14a 2 39.2 3.8 III14b 2 32.4 4.7 III23a 2 41.1 4.0 III23b 2 42.3 3.1 III24a 2 35.2 5.5 III24b 2 39.5 4.8 Table 2 Results of the ab initio gradient geometry optimisation of the most stable tautomers of 5-OH-cytosine. The superscripts denote the full gradient optimisation (1), or single point calculations on the basis of the 6-311G geometry (2), respectively. The energies are expressed in Hartrees and dipole moments in Debay Energies I1 I2 IIa1 IIa2 IIb1 IIb2 3-21G(1) 2 464.85144 2 464.84999 2 464.85572 2 464.83793 2 464.84936 2 464.85020 6-31G(1) 2 467.25137 2 467.25093 2 467.25573 2 467.23793 2 467.24994 2 467.25000 6-311G(1) 2 467.36381 2 467.36310 2 467.36757 2 467.35008 2 467.36186 2 467.36205 MP2(6-311G)(2) 2 468.33306 2 468.33263 2 468.33732 2 468.32062 2 468.33126 2 468.33063 6-31G**(1) 2 467.47951 2 467.47774 2 467.48333 2 467.346888 2 467.47721 2 467.47836 6-311G**(2) 2 467.58420 2 467.58252 2 467.58835 2 467.57422 2 467.58250 2 467.58350 MP2(6-311G**)(2) 2 469.02048 2 469.02033 2 469.02552 2 469.01099 2 469.01883 2 469.01847 Dipole moments MP2(6-311G)(2) 7.576 8.425 0.382 3.619 6.702 5.478 MP2(6-311G**)(2) 7.253 7.936 0.715 3.481 6.138 5.062
P.Cvsewski Journal of Molecular Structure (Theochem)466 (1999)49-58 3 H1-N4H2.C4■.1704 H 1193 H05C5-C4=86 113 、24 1216 1357 180 121 tauto both taut stabili- 3.2.Solvation of 5-hydroxyeytosine the favourable isomer in the vacuum need not be the most preferred in the more polar environmental Usually the vacuum and non-polar solvents stabilise structures of low polarity but the increase of the envir Table 3 pola ealculated dipole ed in th initio calculations.The tautomer al is characterised 20.70)and cyelohexane (202) by about one order lower dipole moment than the second stable tautomer I1.In such a situation the influ- Solvent I ence of the solvent field seems to be very important basis of the 6-311G -467280 -PC t stable tautomer Wate -4672883 water,acetone ar -467.26536 -46726510 The 6-311G* e would 4676125 4676064 pect,the inversion of the relative stability of the -467.59466 -467.59509 studied tautomer was observed for polar and non polar solvents.The more polar the solvent,the more
both tautomers will have consequences on the stabilisation of the systems containing 5-OH-C. Both pairing and stacking will be strongly dependent on the tautomeric form. Besides, the solvation effect may also be tautomer dependent. 3.2. Solvation of 5-hydroxycytosine The solvation effect may play an important role in the stabilisation of a particular tautomer. The most favourable isomer in the vacuum need not be the most preferred in the more polar environmental. Usually the vacuum and non-polar solvents stabilise structures of low polarity but the increase of the environment polarity results in the stabilisation of polar structures. This is the case, for example, for 2-OHadenine [21]. Table 2 comprises the calculated dipole moments of the studied tautomer estimated on the basis of ab initio calculations. The tautomer IIa1 is characterised by about one order lower dipole moment than the second stable tautomer I1. In such a situation the influence of the solvent field seems to be very important. The solvent effect was estimated on the basis of the SCI-PCI technique for the two most stable tautomers in three distinct solvents: water, acetone and cyclohexane. The resulting energies are given in Table 3 and additionally plotted in Fig. 4 as curves related to the energy level of tautomer I1 in water. As one would expect, the inversion of the relative stability of the studied tautomer was observed for polar and nonpolar solvents. The more polar the solvent, the more P. Cysewski / Journal of Molecular Structure (Theochem) 466 (1999) 49–58 53 Fig. 3. The optimised geometry parameters of the two most stable tautomers of 5-OH-cytosine from HF SCF (6-31G**) optimisation. The values of the bond lengths (in italics) are given in angstroms and the values of the bond angles are expressed in degrees. Table 3 The solvent effect was estimated on the basis of the self-consistent isodensity polarised continuum model [22]. The full gradient geometry optimisation (1) or single point energy calculations related to previously obtained geometry (2) were performed in the presence of the solvent field. Three solvents were considered: water (1 78.54), acetone (1 20.70) and cyclohexane (1 2.02). Values of the dielectric constant are given in parentheses Solvent I1 IIa1 6-311G(1) Water 2 467.2883 2 467.28040 Acetone 2 467.28651 2 467.27889 Cyclohexane 2 467.26536 2 467.26510 6-311G**(2) Water 2 467.61255 2 467.60641 Acetone 2 467.61079 2 467.60530 Cyclohexane 2 467.59466 2 467.59509