The effects of tautomerization and protonation on the adenine–cytosine mismatches: a density functional theory study

In the present work, we demonstrate the results of a theoretical study concerned with the question how tautomerization and protonation of adenine affect the various properties of adenine–cytosine mismatches. The calculations, in gas phase and in water, are performed at B3LYP/6-311++G(d,p) level. In gas phase, it is observed that any tautomeric form of investigated mismatches is more stabilized when adenine is protonated. As for the neutral mismatches, the mismatches containing amino form of cytosine and imino form of protonated adenine are more stable. The role of aromaticity on the stability of tautomeric forms of mismatches is investigated by NICS(1)ZZ index. The stability of mispairs decreases by going from gas phase to water. It can be explained using dipole moment parameter. The influence of hydrogen bonds on the stability of mismatches is examined by atoms in molecules and natural bond orbital analyses. In addition to geometrical parameters and binding energies, the study of the topological properties of electron charge density aids in better understanding of these mispairs.


Introduction
Prototropic tautomerism of nucleotide bases is believed to induce the formation of rare tautomers of DNA bases (Colominas, Luque, & Orozco, 1996;Schreiber & González, 2007;Šponer & Hobza, 2003). In some cases, the complexes of rare tautomers are more stable than ground state one. For example, Samijlenko et al. indicated that the three rare adenine tautomers generate more stable complexes with CH 3 COO − and Na + ions in comparison with the ground state case (Samijlenko, Krechkivska, Kosach, & Hovorun, 2004). The rare tautomeric forms can also cause mispairing during replication and, if it escapes repair, eventually may lead to a point mutation (Lippert & Gupta, 2009;Topal & Fresco, 1976). Furthermore, the mispairing can be incorporated into the DNA strand without disturbing the geometrical and energetic demands of the DNA double helix on the base-pair formation (Guckian, Krugh, & Kool, 2000;Harris et al., 2003). Das and Lyngdoh have proposed that the configuration of a solitary base pair is a major factor to predict whether it would occur at the wobble position or not (Das & Lyngdoh, 2014). Moreover, Watson and Crick explained that tautomerization alters the hydrogen bonding partners and therefore could enable mismatches to assume the structure of canonical base pairs (Watson & Crick, 1953). For instance, Wang et al. have observed adenine-cytosine mismatch using high-resolution X-ray crystallographic analysis of a DNA polymerase that catalyzes replication in crystal (Wang, Hellinga, & Beese, 2011). In adenine-cytosine mispair, if adenine (A) tautomerizes to its imino form (from hereon designated A * ), the latter will pair with cytosine (C) instead of thymine (T). This behavior is also observed for imino form of cytosine (denoted as C * ). On the other words, the C * incorrectly pairs with A. This mismatch was theoretically characterized by Danilov, Anisimov, Kurita, and Hovorun (2005) and Fonseca Guerra, Bickelhaupt, Saha, and Wang (2006). Danilov et al. indicated that the formation energy of A·C * is more favorable than A * ·C. They also studied the nature of their hydrogen bonding based on the energy decomposition analysis of Morokuma-Kitaura and the reduced-variational-space methods to elucidate the possibility of the tautomerization of DNA base pairs (Danilov et al., 2005). Also, Fonseca Guerra et al. confirmed that A * ·C mispair is a suitable candidate for the incorporation into DNA (Fonseca Guerra et al., 2006). Hovorun and Brovarets' have indicated that exactly the A·C * base mispair is an active player of the point mutational events and is effectively dissociated by the replication machinery into the A and C * monomers in contrast to the A * ·C base mispair, playing the mediated role of a provider of the A·C * base mispair in DNA that is synthesized (Brovarets' & Hovorun, 2015c). In order to understand the origin of the spontaneous transitions, Hovorun and Brovarets' have also proposed the new physicochemical mechanism on the basis of tautomerization, induced by the interaction of the DNA polymerase recognition center with the canonic nucleotide bases, of the pyrimidine bases in wobble base pairs A·C and G·T which transition into the pairs A·C * and G * ·T accordingly (Brovarets' & Hovorun, 2009). In other work, the tautomeric transition between wobble A·C(w) mismatch and Watson-Cricklike A·C * (WC) base mispair has theoretically studied by proceeding non-dissociatively via the sequential proton transfer between the bases through the planar, highly stable, and zwitterionic TS A+·C-A·C(w) ↔ A·C*(WC) transition state. Here, the authors have suggested that biologically significant A·C(w) → A·C * (WC) tautomerisation is kinetically controlled pathway for the formation of the enzymatically competent Watson-Crick-like A·C * (WC) DNA base mispair in the essentially hydrophobic recognition pocket of the high-fidelity DNA polymerase responsible for the occurrence of spontaneous point AC/CA incorporation errors during DNA biosynthesis (Brovarets' & Hovorun, 2015a). Recently, the interesting practical applications of non-Watson-Crick base pairs have also been investigated. Theoretical calculations predict A-C, A-G, and C-C mispairs as ideal to be incorporated into DNA sequences for developing a molecular switch. The experimental proof further substantiates the conceivability of an A-C mispair-based pH switch. In addition, it has been observed that pH-dependent molecular switches can be constructed harnessing the stability of a protonated A-C mispair (Jissy & Datta, 2014).
Protonation of nucleic acid bases plays a significant role in many biochemical (i.e., enzymatic reactions, stabilization of triplex structures) and mutagenic processes (Sinden, 1994;Wilcox, Ahluwalia, & Bevilacqua, 2011). The protonation/deprotonation of base in any canonical nucleoside significantly perturbs DNA-like conformations. Nevertheless, the ionization mechanism cannot explain entirely the nature of the spontaneous transitions (Brovarets', Zhurakivsky, & Hovorun, 2010). Halder et al. have benchmarked the conventional formalisms for modeling the process of protonation and concluded that considering water as a proton donor might provide a physicochemically relevant picture of the relative order of protonation propensity of different sites of the nucleobases. Also, while the availability of stabilization possibilities determines the feasibility of occurrence of protonated bases, their occurrence context and specific functional roles are important factors determining their occurrence propensities (Halder, Halder, Bhattacharyya, & Mitra, 2014). Specifically, the protonation of adenine, adenosine, and adenine-containing nucleotides have been studied extensively in both gas and condensed phases by theoretical and experimental means (Russo, Toscano, Grand, & Jolibois, 1998;van Zundert et al., 2011). For example, Marian et al. produced protonated adenine by electrospray ionization (Marian, Nolting, & Weinkauf, 2005). In the other work, protonated adenine was produced after the ionization of cold neutral adenine dimers (Cheong et al., 2011). Also, Rajabi et al. investigated the structure of the protonated adenine dimer by infrared multiple photon dissociation spectroscopy and electronic structure calculations (Rajabi, Theel, Gillis, Beran, & Fridgen, 2009).
Base mispairing may also be brought about by a charged nucleobase. Such processes could in fact circumvent the necessity of tautomerization (Goodman, 1995). The present work deals with the mispairs involving adenine nucleobase, in its neutral and protonated forms, and cytosine. Here, we focus on the mispairs which can exist in a cognate base-pair conformation in the gas phase and in water. Protonated adenine comes in various tautomeric forms. This complexity arises from the fact that the two stable neutral adenine tautomers can be protonated at different sites. On the basis of experimental data and theoretical calculations, the four lowest ones in energy as well as their imino forms are shown in Scheme 1. All the other tautomers of the protonated adenine have considerably higher energies and are not studied here. Also, 1H,9H-adeninium (1H-9H-A (+) ) is not considered in this manuscript whereas it leads to wobble base paring.

Computational details
All calculations have been implemented in the Gaussian 09 suite of programs (Frisch et al., 2009) at the spin-restricted level. The density functional theory (DFT) method where electron correlation is taken into account by means of non-local exchange and correlation functional is emerging as a cost-effective alternative to the time consuming (Hertwig & Koch, 1995;Leulliot, Ghomi, Scalmani, & Berthier, 1999;Shukla & Leszczynski, 2000). Here, the geometries were optimized at the B3LYP/6-311++G(d,p) level of theory. The B3LYP/6-311++G(d,p) has been successfully applied on similar systems recently studied and have been verified to give accurate normal mode frequencies, characteristics of intra-and intermolecular hydrogen bonds and geometries (Brovarets', Zhurakivsky, & Hovorun, 2013a;Brovarets', Zhurakivsky, & Hovorun, 2013b;Matta, 2010;Ponomareva, Yurenko, Zhurakivsky, Mourik, & Hovorun, 2014;Wiberg, 2004;Yurenko, Zhurakivsky, Samijlenko, & Hovorun, 2011). Solute-solvent interactions can have dramatic effects on molecular energies, structures, and properties (Erikson, 2001;Leszczynski, 1995). If it were necessary to consider each solvent molecule as a separate molecule, the computational cost of modeling a solvent-mediated chemical reaction would grow prohibitively high. Modeling the solvent as a polarizable continuum, rather than individual molecules, makes ab initio computation feasible. In many cases, solvation effects can be computed very effectively in the framework of continuum solvation models (Dolney et al., 2000;Tomasi & Persico, 1994). In these models, the bulk of the solvent is represented as a structureless polarizable medium characterized mainly by its dielectric constant. Even when specific interactions require the introduction of some solvent molecules, strongly bound to the solute, the continuum picture is still very useful and often necessary (Cossi, Rega, Scalmani, & Barone, 2003). To examine the effects of water solvent on the stability of adenine-cytosine mispairs, the polarizable continuum model (PCM) (Barone & Cossi, 1998) was performed at the B3LYP/6-311++G(d,p) level. The PCM is one of the most reliable continuum solvation procedures and is commonly used in computational chemistry to model solvation effects (Kannappan, Suganthi, & Sathyanarayanamoorthi, 2014;Tomasi, Mennucci, & Cammi, 2005). Such a success is mainly because of the continuous improvements, both in terms of computational efficiency and generality, made by all the people involved in the PCM project. The result of these efforts is that nowadays, PCM, with all its different variants, is the default choice in many computational codes to couple a quantum-mechanical description of a molecular system with a continuum description of the environment (Mennucci, 2012). The symmetrical constraint (Cs) was considered during optimization. It has been previously shown that the effect of constraining the nucleobase geometry to Cs symmetry on the bond energy of natural Watson-Crick pairs (which are close to Cs symmetric) is in most cases very small or zero (Fonseca Guerra, Bickelhaupt, Snijders, & Baerends, 2000;Paragi et al., 2008). The nature of the stationary points was confirmed by frequency calculations at the same level of theory. All the structures are found to be at local minima on the potential energy surface. To obtain credible complexation energies in gas phase and in water, the single-point energies of the species were further refined at the MP2/6-311++G(d,p)//B3LYP/6-311++G(d,p) level of theory. The aromaticity of adenine and cytosine tautomers was also investigated by nucleus-independent chemical shift (NICS) calculations. It is defined as the negative value of the absolute shielding computed at a ring center or at some other interesting points of a system. Rings with large negative NICS values are considered aromatic. As shown by Lazzeretti and Aihara (Aihara, 2002;Lazzereti, 2000;Lazzeretti, 2004), NICS values at the geometrical center of the ring (NICS(0)) contain important spurious contributions from the in-plane tensor components that are not related to aromaticity. NICS(1) (1Å above/below the plane of the ring) essentially reflects π-effects and it is a better indicator of the ring current than the value at the center, because at this point, the effects of the local σ-bonding contributions are diminished (Corminboeuf, Heine, Seifert, Schleyer, & Weber, 2004;von Ragué Schleyer et al., 2001). The outof-plane component of the NICS(1), NICS(1) ZZ , correctly reflects the π-electron effects and probably is a better descriptor of aromaticity (Ebrahimi, Habibi, Masoodi, & Gholipour, 2009). For determining NICS (1) ZZ values, the NMR calculations were performed at the B3LYP/6-311++G(d,p) level using GIAO formalism (Wolinski, Hinton, & Pulay, 1990).
A typical adenine-cytosine mismatch is given in Scheme 2(b). It is observed that adenine-cytosine mismatch can exist in two forms: A * …C and A…C * . In order to compare the stability of amino and imino forms of adenine and cytosine in gas phase and in water, the cohesive energy or the binding energy (BE) per atom for any tautomer was calculated according to the following formula (Rohrer, 2001;Roohi & Bagheri, 2013): where a, b, d, e, and f are the number of C, N, H, O, and H + species, respectively. E C , E N , E H , E O , and E H + are the ground state total energies of C, N, H, O, and H + , respectively, and E is the total energy of the optimized tautomer. As shown from Table 1, the energetic order of neutral and protonated forms of adenine in gas phase and in water is as following: It is also found that the BE value of C is higher than that of C * . These findings indicate that the protonated forms of adenine are more stable than neutral ones. Also, the amino form is more possible than imino case for any species. Here, the aromaticity of six-membered ring of any species is examined at both amino and imino forms. With reference to the values of NICS(1) ZZ , it is concluded that the amino form is higher in aromaticity than the imino case (see Table 1). Thus, the more stability of amino form may be attributed to aromaticity factor. It is observed that the BE value increases by going from gas phase to water. As can be seen from Table 1, the dipole moment of adenine and cytosine tautomers in water is greater than that in gas phase. Whereas water is a polar solvent, the more stability of tautomers in water may be ascribed to hydration effects.

Theoretical study of adenine-cytosine mismatches
Here, the complexation energy has been calculated using equation where E complex and E mon are optimized energies of adenine-cytosine mismatch and each individual component monomer, respectively. The results in Table 2 indicate that the ΔE values are negative. In other words, the complexation energies show an upward trend as is observed for the A * …C, A…C * , A (+)* …C, and A (+) …C * mismatches. In gas phase, any tautomeric form of adenine-cytosine mispair is more stabilized when adenine is protonated. It is also found that the ΔE value of A * …C/ A (+)* …C complexes is more negative than that of A… C * /A (+) …C * ones, respectively, indicating that A * …C and A (+)* …C mismatches are more stable. The aromaticity of six-membered ring of adenine and cytosine changes during complexation (see Table 3). A meaningful relationship can be found between the stability of adenine-cytosine mismatches and aromaticity changes (ΔNICS(1) ZZ ). From aromaticity point of view, it seems that the imino forms of adenine and cytosine are more suitable for mismatch formation. Whereas the ΔNICS(1) ZZ in adenine (at both A and A * forms) is greater than that in cytosine, it is expected that the A * … C/A (+)* …C to be more possible than A…C * /A (+) …C *  (1) (3) mispairs, respectively. This result is confirmed by the energetic order of investigated complexes. Compared to gas phase, the absolute values of ΔE (|ΔE|) decrease in water. Here, the role of dipole moment (μ) is investigated. It is observed that the investigated mismatches have lower dipole moments than the total dipole moments of the monomers. The stronger interaction of polar solvent (water) with monomers leads to a decrease in stability of adenine-cytosine mismatches.
In addition, we have calculated the interaction (ΔE int ) and deformation (ΔE def ) energies of the considered systems (see Table 2). The ΔE int values were evaluated as the difference between the complexation energy and the sum of the energies of the separated monomers, with the same geometries as they have in the complex (i.e., frozen geometries). The difference between the ΔE and the ΔE int is the deformation energy of the monomers. While the value of ΔE int is negative and makes a positive contribution to the ΔE, this behavior is reversed for ΔE def . For any investigated system, the ΔE int and ΔE def values in A * …C/A (+)* …C are, respectively, greater than those in A…C * /A (+) …C * . It is observed that the contribution of ΔE int is dominant in all complexes. Also, the difference between ΔE int and ΔE def is amplified in the presence of protonated adenine. Considering the difference between ΔE int and ΔE def , it is expected that the A * …C and A (+)* …C mismatches to be more stable in neutral and positively charged adenine-cytosine mispairs. This result is in accord with the order of ΔE values.
The adenine-cytosine mispair is stabilized by three hydrogen bonds (H-bonds). As shown in Scheme 2(b), they are denoted by HB1, HB2, and HB3. In the following, the H-bonds are examined using geometrical, topological, and energetic parameters. The bond length of A…H (A is proton acceptor) is often treated as a rough measure of the strength of H-bond (Chęcińska & Grabowski, 2006). For studied mismatches, the bond lengths of H-bonds are gathered in Table S1 (see supplementary materials). It can be seen that the length of HB1 in A * …C is shorter than that in A…C * . An opposite order is observed for mismatches containing protonated adenine. These observations can be explained on the basis of atomic charges of N(1) and H(1) atoms. The natural charges (the nuclear charge minus the summed natural populations of natural atomic orbitals on the atom) on the H and N atoms (q H , q N ) obtained using NBO calculations at the B3LYP/6-311++G(d,p) level of theory are given in Table S2 (see supplementary materials). The N(1) and H(1) atoms involved in HB1 have negative and positive charges, respectively. The increase in the absolute value of q N(1) (|q N(1) |) and q H(1) leads to stronger HB1. The protonation of adenine leads to increasing q H(1) and decreasing |q N(1) | in its amino and imino forms, respectively. Thus, the HB1 is, respectively, strengthened/weakened in A (+) …C * /A (+)* …C forms. Compared to A…C * , it is also found that the length of HB2 in A * …C is shortened. This difference is intensified in the presence of protonated adenine. In A * …C mismatch, the H(2) and N(2) atoms involved in HB2 correspond to adenine and cytosine, respectively. This contribution is reversed for A…C * mispair. The increase  in q H(2) and |q N(2) | leads obviously to strengthen HB2. In gas phase, the q H(2) in C * and |q N(2) | in C are, respectively greater than those in A * and A. In comparison with q H(2) values in C * and A * , the difference in |q N(2) | values of A and C is further revealed. Hence, it is expected that the HB2 in A * …C to be stronger than that in A…C * . In water, the values of q H(2) in A * and |q N(2) | in C are both higher than those in C * and A, respectively. The protonation of adenine is accompanied by increasing q H(2) in A (+)* and decreasing |q N(2) | in A (+) . Thus, HB2 in A (+)* …C is obviously stronger than that in A (+) …C * . The HB3 includes H and O atoms of adenine and cytosine, respectively. It is observed that the length of HB3 in A * …C is shorter than that in A…C * . This behavior can be ascribed to natural atomic charges on O (3) and H(3) atoms. As shown in Table S2, the q H(3) and |q O(3) | values in A * and C are, respectively, higher than those in A and C * . The protonation of adenine intensifies this difference. Topological criteria were also proposed to detect the existence of H-bond (Grabowski, Sokalski, & Leszczynski, 2005;Popelier, 2000). The bond critical points (BCPs) of the N…H and O…H interactions were found and the features of them were analyzed since it is well known that characteristics of BCPs, such as the electron densities (ρ HB ), their laplacians (∇ 2 ρ HB ), and the energetic properties (H HB ) of BCPs, allow us to categorize interactions, and these topological parameters are also treated as measures of H-bond strength (Domagała & Grabowski, 2005;Gálvez, Gómez, & Pacios, 2003). The values of ρ HB , ∇ 2 ρ HB , and H HB at the BCPs were evaluated by the means of AIM approach at the B3LYP/6-311++G(d,p) level of theory (see supplementary materials, Table S3). A typical molecular graph is shown in Figure 1. In general, the ρ HB value can be a useful parameter for describing the strength of H-bonds (Grabowski, 2006). The order of ρ HB1 values is identical in gas phase and in water. In neutral complexes, the value of ρ HB1 in A * …C is higher than that in A…C * . An opposite behavior is observed in the presence of protonated adenine. Rozas et al. have introduced a classification of H-bonds according to their strength. Weak H-bonds show both ∇ 2 ρ HB and H HB values positive; for medium H-bonds, ∇ 2 ρ HB > 0 and H HB < 0, and also for strong H-bonds, the ∇ 2 ρ HB as well as H HB are negative (Rozas, Alkorta, & Elguero, 2000). In gas phase, ∇ 2 ρ HB1 and H HB1 values in A…C * form are positive, indicating that HB1 considered as weak Hbonds while A * …C form is characterized by the positive ∇ 2 ρ HB1 and negative H HB1 showing that HB1 may be classified as medium H-bonds. This order is reversed for positively charged complexes. In water, it is also observed that the HB1 in all neutral complexes can be considered as medium H-bonds. The ρ HB2 values in gas phase and in water indicate that HB2 in A * …C form is stronger than that in A…C * . This difference is intensified in the presence of protonated adenine. According to Rozas classification, HB2 in all neutral complexes is of medium type. The presence of protonated adenine increases ρ HB2 in A (+)* …C form and decreases it in A (+) …C * case so that HB2 in A (+)* …C and A (+) …C * forms can be classified as medium and weak H-bonds, respectively. In any investigated system, the ρ HB2 decreases by going from gas phase to water. Considering ρ HB3 value, it is concluded that HB3 in A * …C form is stronger than that in A…C * case. This difference is intensified by protonation of adenine. On the basis of ∇ 2 ρ HB3 and H HB3 values, HB3 can be characterized as weak H-bonds in all complexes. Compared to gas phase, the value of ρ HB3 in A * …C form increases in water while a reverse behavior is observed for A…C * . Also, ρ HB3 values in both forms of A (+)* …C and A (+) …C * decrease by going from gas phase to water. The AIM results are confirmed by H-bond lengths.

HB3 HB3
9H-A * …C 9H-A …C * Figure 1. The molecular graphs for 9H-adenine-cytosine mismatch obtained using AIM analysis. Small red spheres, small yellow spheres, and lines represent BCPs, ring critical points, and bond paths, respectively. E Iog; HB ¼ 0:33 Á ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Dm À 40 p where E Iog , HB is a hydrogen bond energy in kcal mol −1 , Δν is the magnitude of the red shift (relative to the free molecule) in cm −1 of the stretching mode of H-bonded AH groups in an AH…B complex (A and B = N, O). The partial deuteration was applied to minimize the effect of mechanical resonances .
The Iogansen's formula has frequently been applied for studying various base pairs Brovarets', Yurenko, & Hovorun, 2015;Brovarets', Zhurakivsky, & Hovorun, 2015). For example, Brovarets' et al. have studied Hyp * ·Hyp base pair at B3LYP/6-311++G(d,p) level of theory, where Hyp and Hyp * are keto and enol tautomers of the hypoxanthine, respectively. They estimated the energy of NH…N H-bond in Hyp * .Hyp base pair using Iogansen's formula. On the basis of their results, the energy of NH…N H-bond in gas phase is obtained equal to 29.16 kJ mol −1 (Brovarets', Zhurakivsky & Hovorun, 2013c). In other work, the Iogansen's formula has been used to examine NH…N H-bond in DNA base mispair containing amino and imino tautomers of cytosine. The results obtained at B3LYP/6-311++G(d,p) level in gas phase showed that the energies of upper and lower NH…N H-bonds correspond to 27.87 and 27.07 kJ mol −1 . In the present work, the results in Table 4 indicate that the ranges of E HB1 and E HB2 , in gas phase, are from 19.87 to 35.93 and 11.92 to 40.46 kJ mol −1 , respectively. In water, these ranges correspond to 21.25−28.90 and 15.06−34.00 kJ mol −1 .
Nevertheless, the strength of CH…O H-bond cannot be examined using Iogansen's formula. Thus, the energies of the intermolecular H-bonds were also evaluated by the empirical Espinosa-Molins-Lecomte (EML) formula (Espinosa, Molins, & Lecomte, 1998) Tables 4 and S3. Nikolaienko et al. indicated that the EML formula usually overestimates the energy of the classical H-bonds (Nikolaienko, Bulavin, & Hovorun, 2012). Whereas the evaluation of energy of CH…O H-bond is only possible using EML formula, however, the correlations between E EML,HB values and energetic, geometrical and topological properties are considered in the following sections.
As observed from  (Brovarets', Yurenko & Hovorun, 2014). The results in   Figure 2(a), a direct relationship can be found between ∑E EML,HB and ΔE values.
To gain more insight into the influences of tautomerization and protonation on H-bonds, the NBO analysis has been performed on the investigated complexes. The NBO calculations show that the most important donoracceptor interaction in HB1 and HB2 is Lp N → σ * N-H while it is Lp O → σ * C-H in HB3 (see Scheme 2(b)). The energy values of these interactions (E (2) HB ) are gathered in Table S4 (see supplementary materials). For neutral complexes in gas phase, the value of E (2) HB1 in A * …C form is greater than that in A…C * case. An opposite behavior is observed for positively charged mispairs. With the exception of 7H-A…C * and 7H-A * …C forms, the order of E (2) HB1 values in water is similar to gas phase. In water, the value of E (2) HB1 in 7H-A…C * is greater than that in 7H-A * …C. The difference of E (2) HB1 values in tautomeric forms of any mismatch is reduced in water. It may be attributed to solvent effects on donor-acceptor interactions. In comparison with any mismatch in gas phase, the Lp N(1) → σ * N-H(1) interaction in A…C * /A * …C is reinforced/weakened in water so that E (2) HB1 value in 7H-A…C * would be even greater than that in 7H-A * …C. The value of E (2) HB2 in A * …C is also higher than that in A…C * . This difference is intensified by the protonation of adenine. Compared to gas phase, the Lp N(2) → σ * N-H(2) interactions in A * …C/A…C * mismatches become stronger/weaker in water, while this interaction in both forms of A (+)* …C and A (+) …C * is generally weakened by going from gas phase to water. Similar to E (2) HB2 , the value of E (2) HB3 in A * …C is greater than that in A…C * . This difference is enhanced in the presence of protonated adenine. As going from gas phase to water, Lp O(3) → σ * C-H(3) interaction is weakened/strengthened in A…C * /A * …C forms while it is weakened in both forms of A (+) …C * and A (+)* …C. For any mismatch, the total E (2) HB (∑E (2) HB ) is calculated. In agreement with |ΔE| and |∑E EML,HB |, the value of ∑E (2) HB in A * …C/A (+)* …C is higher than that in A…C * /A (+) …C * . In comparison with gas phase, it is observed that ∑E (2) HB decreases in water. A direct relationship can be seen between ∑E (2) HB and ΔE values in Figure 2(b).

Conclusions
In this manuscript, the effects of tautomerization and protonation of adenine on the various properties of adenine-cytosine mismatches have been investigated. Here, we examined the mispairs which can exist in a cognate base-pair conformation in the gas phase and in water. At first, the neutral and protonated adenine and also neutral cytosine tautomers were studied. With reference to the values of NICS(1) ZZ , it was concluded that the amino forms of adenine and cytosine is higher in aromaticity than the imino case. Thus, the more stability of amino form may be attributed to aromaticity factor. It was observed that the stability of species increases by going from gas phase to water. The dipole moment of adenine and cytosine tautomers in water is greater than that in gas phase. Whereas water is a polar solvent, the more stability of tautomers in water may be ascribed to hydration effects.
The results show that the any tautomeric form of adenine-cytosine mispair in gas phase is more stabilized when adenine is protonated. The ΔE value of A * …C/A (+)* …C complexes is, respectively, more negative than that of A…C * /A (+) …C * ones, indicating that A * …C and A (+)* …C mismatches are more stable. From aromaticity point of view, it seems that the imino forms of adenine and cytosine are more suitable for mismatch formation. Whereas the changes of NICS(1) ZZ in adenine (at both A and A * forms) are greater than that in cytosine, it is expected that the A * …C and A (+)* …C to be more possible than their corresponding mispairs. This result is confirmed by the energetic order of investigated complexes. Compared to gas phase, the |ΔE| values decrease in water. It is observed that the investigated mismatches have lower dipole moments than the total dipole moments of the monomers. The stronger interaction of polar solvent (water) with monomers leads to decreasing stability of adenine-cytosine mismatches. For any investigated mispair, the interaction and deformation energies were calculated. These values in A * …C/A (+)* …C are, respectively, greater than those in A…C * /A (+) …C * . It is observed that the contribution of interaction energy is dominant in all complexes.
The adenine-cytosine mispair is stabilized by three H-bonds. The nature and strength of H-bonds were characterized by AIM analysis. The energies of the intermolecular H-bonds were evaluated by the empirical Iogansen's and EML formulas. Also, the most important donor-acceptor interactions were studied using NBO analysis. The excellent correlations were found between the complexation energy and the results of AIM and NBO analyses.

Supplementary material
The supplementary material for this paper is available online at http://dx.doi.10.1080/07391102.2015.1072734.

Disclosure statement
No potential conflict of interest was reported by the authors.