Remarkably Strong T-Shaped Interactions between Aromatic Amino Acids and Adenine: Their Increase upon Nucleobase Methylation and a Comparison to Stacking

Occurrence and classification of T-shaped interactions between nucleobases in RNA structures

Understanding the frequency and structural context of discrete noncovalent interactions between nucleotides is of pivotal significance in establishing the rules that govern RNA structure and dynamics. Although T-shaped contacts (i.e., perpendicular stacking contacts) between aromatic amino acids and nucleobases at the nucleic acid-protein interface have recently garnered attention, the analogous contacts within the nucleic acid structures have not been discussed. In this work, we have developed an automated method for identifying and unambiguously classifying T-shaped interactions between nucleobases. Using this method, we identified a total of 3261 instances of T-shaped (perpendicular stacking) contacts between two nucleobases in an array of RNA structures from an up-to-date data set of ≤3.5 Å resolution crystal structures deposited in the Protein Data Bank.

Quantum machine learning corrects classical forcefields: Stretching DNA base pairs in explicit solvent

In order to improve the accuracy of molecular dynamics simulations, classical forcefields are supplemented with a kernel-based machine learning method trained on quantum-mechanical fragment energies. As an example application, a potential-energy surface is generalized for a small DNA duplex, taking into account explicit solvation and long-range electron exchange-correlation effects. A long-standing problem in molecular science is that experimental studies of the structural and thermodynamic behavior of DNA under tension are not well confirmed by simulation; study of the potential energy vs extension taking into account a novel correction shows that leading classical DNA models have excessive stiffness with respect to stretching. This discrepancy is found to be common across multiple forcefields. The quantum correction is in qualitative agreement with the experimental thermodynamics for larger DNA double helices, providing a candidate explanation for the general and long-standing discrepancy between single molecule stretching experiments and classical calculations of DNA stretching. The new dataset of quantum calculations should facilitate multiple types of nucleic acid simulation, and the associated Kernel Modified Molecular Dynamics method (KMMD) is applicable to biomolecular simulations in general. KMMD is made available as part of the AMBER22 simulation software.

Collective dynamics of molecular rotors in periodic mesoporous organosilica: a combined solid-state 2H-NMR and molecular dynamics simulation study

Molecular rotors offer a platform to realize controlled dynamics and modulate the functions of solids. The motional mechanisms in arrays of rotors have not been explored in depth. Crystal-like porous organosilicas, comprising p-phenylene rotators pivoted onto a siloxane scaffold, were modelled using molecular dynamics (MD) simulations. Long simulations, on a microsecond scale, allowed to follow the reorientation statistics of rotor collections and single out group configurations and frequency distributions as a function of temperature. The motions observed in the MD simulations support a multiple-site model for rotor reorientations. Computed motional frequencies revealed a complex rotatory phenomenon combining an ultra-fast libration motion (oscillation up to 30°) with a slow and fast 180° flip reorientation. Adopting a multiple-site model provides a more accurate simulation of the ²H-NMR spectra and a rationalization of their temperature dependence. In particular, rotators endowed with distinct rates could be explained by the presence of slower rings locked in a T-shaped conformation.

Impacts of Ca2+ cation and temperature on bovine α-lactalbumin secondary structures and foamability - Insights from computational molecular dynamics

We used computational molecular dynamics (MD) to assess molecular conformations of apo- and holo-forms (respectively without and with Ca²⁺) of bovine α-lactalbumin (α-La) at different temperatures, and to correlate them with the protein's foaming properties. At 4 °C and 25 °C no major protein conformation changes occurred. At 75 °C, lots of changes were evidenced: the Ca²⁺ depletion triggered the complete loss of h2b, h3c helices and S1, S2 and S3 β-sheets, and partial losses of H1, H2 and H3 α-helices. The absence of Ca²⁺ in apo-α-La and its leaving from holo-α-La triggered electrostatic repulsion among Asp82, Asp84 and Asp87, leading to the formation of a hydrophobic cluster involving Phe9, Phe31, Ile1, Va42, Ile55, Phe80 and Leu81. These conformational changes were related to an interfacial tension decrease and to a foaming capacity increase, for both apo-α-La and holo-α-La. This study exemplifies how powerful MD is as a tool to provide a better understanding of the molecular origins of food proteins' techno-functionalities.

Quantum Mechanics Helps Uncover Atypical Recognition Features in the Flavin Mononucleotide Riboswitch

Estimating the binding energies of small molecules to RNA could help uncover their molecular recognition characteristics and be used to rationally design RNA-targeting chemical probes. Here, we leveraged the ability of the fragment molecular orbital (FMO) method to provide detailed pairwise energetic information to examine the interactions between the aptamer domain of the flavin mononucleotide (FMN)-responsive riboswitch and small-molecule ligands. After developing an efficient protocol for executing high-level FMO calculations on RNA-ligand complexes, we applied our protocol to nine FMN-aptamer-ligand complexes. We then used the results to identify "hot-spots" within the aptamer and decomposed pairwise interactions between the hot-spot residues and the ligands. Interestingly, we found that several of these hot-spot residues interact with the ligands via atypical CH···O hydrogen bonds and anion-π contacts, as well as (face-to-edge) T-shaped π-π interactions. We envision that our results should pave the way for the wider and more prominent use of FMO calculations to study structure-energy relationships in diverse RNA-ligand systems, which in turn may provide a basis for dissecting the molecular recognition characteristics of RNAs.

Probing the Nature of Noncovalent Interactions in Dimers of Linear Tyrosine-Based Dipeptides

Tyrosine-based dipeptides self-assemble to form higher order structures. To gain insights into the nature of intermolecular interactions contributing to the early stages of the self-assembly of aromatic dipeptides, we study the dimers of linear dityrosine (YY) and tryptophan-tyrosine (WY) using quantum-chemical methods with dispersion corrections and universal solvation model based on density in combination with energy decomposition and natural bond orbital (NBO) analyses. We find that hydrogen bonding is a dominant stabilizing force. The lowest energy structure for the linear YY dimer is characterized by Ocarboxyl···H(O)tyr. In contrast, the lowest energy dimer of linear WY is stabilized by Ocarboxyl···H(N)trp and πtyr···πtyr. The solvent plays a critical role as it can change the strength and nature of interactions. The lowest energy for linear WY dimer in acetone is stabilized by Ocarboxyl···H(O)tyr, πtrp···H(C), and πtrp···H(N). The ΔG of dimerization and stabilization energies of solvated dipeptides reveal that the dipeptide systems are more stable in the solvent phase than in gas phase. NBO confirms increased magnitudes for donor-acceptor interaction for the solvated dipeptides.

Regioselective synthesis of spiroxindolopyrrolidine: a one step cycloaddition reaction twists inherent optical and fluorescence property of ferrocene–anthracene dyad

We report the regioselective synthesis of spiroxindole using a 1,3-dipolar cycloaddition reaction (1,3-DC). Our experiments show that cycloaddition can alter the inherent optical and fluorescence properties of the ferrocene–anthraced dyad system.

DNA-protein π-interactions in nature: abundance, structure, composition and strength of contacts between aromatic amino acids and DNA nucleobases or deoxyribose sugar

Four hundred twenty-eight high-resolution DNA-protein complexes were chosen for a bioinformatics study. Although 164 crystal structures (38% of those searched) contained no interactions, 574 discrete π-contacts between the aromatic amino acids and the DNA nucleobases or deoxyribose were identified using strict criteria, including visual inspection. The abundance and structure of the interactions were determined by unequivocally classifying the contacts as either π-π stacking, π-π T-shaped or sugar-π contacts. Three hundred forty-four nucleobase-amino acid π-π contacts (60% of all interactions identified) were identified in 175 of the crystal structures searched. Unprecedented in the literature, 230 DNA-protein sugar-π contacts (40% of all interactions identified) were identified in 137 crystal structures, which involve C-H···π and/or lone-pair···π interactions, contain any amino acid and can be classified according to sugar atoms involved. Both π-π and sugar-π interactions display a range of relative monomer orientations and therefore interaction energies (up to -50 (-70) kJ mol(-1) for neutral (charged) interactions as determined using quantum chemical calculations). In general, DNA-protein π-interactions are more prevalent than perhaps currently accepted and the role of such interactions in many biological processes may yet to be uncovered.

Significant strength of charged DNA-protein π-π interactions: a preliminary study of cytosine

The present work characterized the preferred gas-phase structure and optimum interaction energy of both parallel stacked and perpendicular T-shaped dimers between cytosine (C), as a representative nucleobase, and aspartic/glutamic acid (DE), aspartate/glutamate (DE(-)) or arginine (R(+)), using detailed M06-2X/6-31+G(d,p) potential energy surface scans as a function of the relative monomer orientation. Through comparison to previous literature on the π-π interactions between the DNA nucleobases and the aromatic amino acid residues, this work will allow for comparisons between DNA-protein interactions involving aromatic and acyclic R-side chains, as well as comparisons of the relative geometric dependence and magnitude of π-π (C:DE), πcation-π (C:R(+)), and πanion-π (C:DE(-)) interactions. Our results show that the preferred relative monomer orientation is highly dependent on the monomer composition and charge, and is dictated by electrostatic-driven interactions. More importantly, for the first time, we report that the π-π interactions between cytosine and (neutral) aspartic/glutamic acid are up to approximately -40 kJ mol(-1), while the πcation-π or πanion-π interactions between cytosine and arginine or aspartate/glutamate are up to approximately -90 and -99 kJ mol(-1), respectively. An extensive investigation of the effects of the computational methodology implemented, including comparisons to detailed CCSD(T)/CBS potential energy surfaces and interaction energies, supports the use of M06-2X, as well as ωB97X-D, to study DNA-protein π-π interactions of varying composition and charge. Most importantly, the CCSD(T)/CBS results verify the strong nature of these DNA-protein π-π interactions, as well as the unique nature of the πcation-π and πanion-π counterparts. Therefore, our results emphasize that a wide variety of different types of noncovalent interactions between both cyclic and acyclic π-containing components can significantly contribute to the stability of DNA-protein complexes and likely play a larger role in biology than currently accepted.

yDNA versus yyDNA pyrimidines: computational analysis of the effects of unidirectional ring expansion on the preferred sugar-base orientation, hydrogen-bonding interactions and stacking abilities

The properties of natural, y- and yy-pyrimidines are compared using computational (B3LYP, MP2) methods. Ring expansion upon incorporation of benzene or naphthalene into the natural pyrimidines affects the preferred orientation of the base about the glycosidic bond in the corresponding nucleoside to a similar extent. Specifically, although the natural pyrimidines preferentially adopt the anti orientation with respect to the 2'-deoxyribose moiety, the expanded analogues will likely display (anti/syn) conformational heterogeneity, which may lead to alternate hydrogen-bonding modes in double-stranded duplexes. Nevertheless, the A:T Watson-Crick hydrogen-bond strengths do not significantly change upon base expansion, while the G:C interaction energy is slightly strengthened upon incorporation of either expanded pyrimidine. The largest effect of base expansion occurs in the stacking energies. Specifically, the maximum (most negative) stacking energies in isolated dimers formed by aligning the nucleobase centers of mass can be increased up to 45% by inclusion of a single y-pyrimidine and up to 55% by consideration of a yy-pyrimidine. Similar increases in the stacking interactions are found when a simplified duplex model composed of two stacked (hydrogen-bonded) base pairs is considered, where both the intrastrand and interstrand stacking interactions can be increased and the effects are more pronounced for the yy-pyrimidines. Moreover, the total stability (sum of all hydrogen-bonding and stacking interactions) is greater for duplexes containing expanded yy-pyrimidines compared to y-pyrimidines, which is mainly due to enhanced stacking interactions. Thus, our calculations suggest that multiple unidirectional increases in the size of the nucleobase spacer can continuously enhance the stability of expanded duplexes.

Combined effects of π-π stacking and hydrogen bonding on the (N1) acidity of uracil and hydrolysis of 2'-deoxyuridine

M06-2X/6-31+G(d,p) is used to study the simultaneous effects of π-π stacking interactions with phenylalanine (modeled as benzene) and hydrogen bonding with small molecules (HF, H(2)O, and NH(3)) on the N1 acidity of uracil and the hydrolytic deglycosylation of 2'-deoxyuridine (dU) (facilitated by fully (OH(-)) or partially (HCOO(-)···H(2)O) activated water). When phenylalanine is complexed with isolated uracil, the proton affinity of all acceptor sites significantly increases (by up to 28 kJ mol(-1)), while the N1 acidity slightly decreases (by ~6 kJ mol(-1)). When small molecules are hydrogen bound to uracil, addition of the phenylalanine ring can increase or decrease the acidity of uracil depending on the number and nature (acidity) of the molecules bound. Furthermore, a strong correlation between the effects of π-π stacking on the acidity of U and the dU deglycosylation reaction energetics is found, where the hydrolysis barrier can increase or decrease depending on the nature and number of small molecules bound, the nucleophile considered (which dictates the negative charge on U in the transition state), and the polarity of the (bulk) environment. These findings emphasize that the catalytic (or anticatalytic) role of the active-site aromatic amino acid residues is highly dependent on the situation under consideration. In the case of uracil-DNA glycosylase (UNG), which catalyzes the hydrolytic excision of uracil from DNA, the type of discrete hydrogen-bonding interactions with U, the nature of the nucleophile, and the anticipated weak, nonpolar environment in the active site suggest that phenylalanine will be slightly anticatalytic in the chemical step, and therefore experimentally observed contributions to catalysis may entirely result from associated structural changes that occur prior to deglycosylation.

Modeling the chemical step utilized by human alkyladenine DNA glycosylase: a concerted mechanism AIDS in selectively excising damaged purines

Human alkyladenine DNA glycosylase (AAG) initiates the repair of a wide variety of (neutral or cationic) alkylated and deaminated purines by flipping damaged nucleotides out of the DNA helix and catalyzing the hydrolytic N-glycosidic bond cleavage. Unfortunately, the limited number of studies on the catalytic pathway has left many unanswered questions about the hydrolysis mechanism. Therefore, detailed ONIOM(M06-2X/6-31G(d):AMBER) reaction potential energy surface scans are used to gain the first atomistic perspective of the repair pathway used by AAG. The lowest barrier for neutral 1,N(6)-ethenoadenine (εA) and cationic N(3)-methyladenine (3MeA) excision corresponds to a concerted (A(N)D(N)) mechanism, where our calculated ΔG(‡) = 87.3 kJ mol(-1) for εA cleavage is consistent with recent kinetic data. The use of a concerted mechanism supports previous speculations that AAG uses a nonspecific strategy to excise both neutral (εA) and cationic (3MeA) lesions. We find that AAG uses nonspecific active site DNA-protein π-π interactions to catalyze the removal of inherently more difficult to excise neutral lesions, and strongly bind to cationic lesions, which comes at the expense of raising the excision barrier for cationic substrates. Although proton transfer from the recently proposed general acid (protein-bound water) to neutral substrates does not occur, hydrogen-bond donation lowers the catalytic barrier, which clarifies the role of a general acid in the excision of neutral lesions. Finally, our work shows that the natural base adenine (A) is further inserted into the AAG active site than the damaged substrates, which results in the loss of a hydrogen bond with Y127 and misaligns the general base (E125) and water nucleophile to lead to poor nucleophile activation. Therefore, our work proposes how AAG discriminates against the natural purines in the chemical step and may also explain why some damaged pyrimidines are bound but are not excised by this enzyme.

Evaluating how discrete water molecules affect protein-DNA π-π and π(+)-π stacking and T-shaped interactions: the case of histidine-adenine dimers

Changes in the magnitude of (M06-2X/6-31+G(d,p)) π-π stacking and T-shaped (nucleobase-edge and amino acid-edge) interactions between (neutral or protonated) histidine (His) and adenine (A) dimers upon microsolvation with up to four discrete water molecules were determined. A variety of histidine-water interactions were considered including conventional (N-H···O, N···H-O, C-H···O) hydrogen bonding and nonconventional (X-H···π (neutral His) or lone-pair···π (protonated His)) contacts. Overall, the effects of discrete His-H(2)O interactions on the neutral histidine-adenine π-π contacts are negligible (<3 kJ mol(-1) or 15%) regardless of the type of water binding, the number of water molecules bound, or the His-A dimer (stacked or (amino acid- or nucleobase-edge) T-shaped) configuration. This suggests that previously reported gas-phase binding strengths for a variety of neutral amino acid-nucleobase dimers are likely relevant for a wide variety of (microsolvated) environments. In contrast, the presence of water decreases the histidine-adenine π(+)-π interaction by up to 15 kJ mol(-1) (or 30%) for all water binding modes and orientations, as well as different stacked and T-shaped His(+)-A dimers. Regardless of the larger effect of discrete histidine-water interactions on the magnitude of the π(+)-π compared with π-π interactions, the π(+)-π binding strengths remain substantially larger than the corresponding π-π contacts. These findings emphasize the distinct nature of π(+)-π and π-π interactions and suggest that π(+)-π contacts can provide significant stabilization in biological systems relative to π-π contacts under many different environmental conditions.

Aromatic rings in chemical and biological recognition: energetics and structures

This review describes a multidimensional treatment of molecular recognition phenomena involving aromatic rings in chemical and biological systems. It summarizes new results reported since the appearance of an earlier review in 2003 in host-guest chemistry, biological affinity assays and biostructural analysis, data base mining in the Cambridge Structural Database (CSD) and the Protein Data Bank (PDB), and advanced computational studies. Topics addressed are arene-arene, perfluoroarene-arene, S⋅⋅⋅aromatic, cation-π, and anion-π interactions, as well as hydrogen bonding to π systems. The generated knowledge benefits, in particular, structure-based hit-to-lead development and lead optimization both in the pharmaceutical and in the crop protection industry. It equally facilitates the development of new advanced materials and supramolecular systems, and should inspire further utilization of interactions with aromatic rings to control the stereochemical outcome of synthetic transformations.

Effects of the biological backbone on stacking interactions at DNA-protein interfaces: the interplay between the backbone···π and π···π components

The (gas-phase) MP2/6-31G*(0.25) π···π stacking interactions between the five natural bases and the aromatic amino acids calculated using (truncated) monomers composed of conjugated rings and/or (extended) monomers containing the biological backbone (either the protein backbone or deoxyribose sugar) were previously compared. Although preliminary energetic results indicated that the protein backbone strengthens, while the deoxyribose sugar either strengthens or weakens, the interaction calculated using truncated models, the reasons for these effects were unknown. The present work explains these observations by dissecting the interaction energy of the extended complexes into individual backbone···π and π···π components. Our calculations reveal that the total interaction energy of the extended complex can be predicted as a sum of the backbone···π and π···π components, which indicates that the biological backbone does not significantly affect the ring system through π-polarization. Instead, we find that the backbone can indirectly affect the magnitude of the π···π contribution by changing the relative ring orientations in extended dimers compared with truncated dimers. Furthermore, the strengths of the individual backbone···π contributions are determined to be significant (up to 18 kJ mol(-1)). Therefore, the origin of the energetic change upon model extension is found to result from a balance between an additional (attractive) backbone···π component and differences in the strength of the π···π interaction. In addition, to understand the effects of the biological backbone on the stacking interactions at DNA-protein interfaces in nature, we analyzed the stacking interactions found in select DNA-protein crystal structures, and verified that an additive approach can be used to examine the strength of these interactions in biological complexes. Interestingly, although the presence of attractive backbone···π contacts is qualitatively confirmed using the quantum theory of atoms in molecules (QTAIM), QTAIM electron density analysis is unable to quantitatively predict the additive relationship of these interactions. Most importantly, this work reveals that both the backbone···π and π···π components must be carefully considered to accurately determine the overall stability of DNA-protein assemblies.

The assessment of density functionals for DNA–protein stacked and T-shaped complexes

The present work uses 129 nucleobase – amino acid CCSD(T)/CBS stacking and T-shaped interaction energies as reference data to test the ability of various density functionals with double-zeta quality basis sets, as well as some semi-empirical and molecular mechanics methods, to accurately describe noncovalent DNA–protein π–π and π+–π interactions. The goal of this work is to identify methods that can be used in hybrid approaches (QM/MM, ONIOM) for large-scale modeling of enzymatic systems involving active-site (substrate) π–π contacts. Our results indicate that AMBER is a more appropriate choice for the lower-level method in hybrid techniques than popular semi-empirical methods (AM1, PM3), and suggest that AMBER accurately describes the π–π interactions found throughout DNA–protein complexes. The M06–2X and PBE-D density functionals were found to provide very promising descriptions of the 129 nucleobase – amino acid interaction energies, which suggests that these may be the most suitable methods for describing high-level regions. Therefore, M06–2X and PBE-D with both the 6–31G(d) and 6–31+G(d,p) basis sets were further examined through potential-energy surface scans to better understand how these techniques describe DNA–protein π–π interactions in both minimum and nonminimum regions of the potential-energy surfaces, which is critical information when modeling enzymatic reaction pathways. Our results suggest that studies of stacked nucleobase – amino acid systems should implement the PBE-D/6–31+G(d,p) method. However, if T-shaped contacts are involved and (or) smaller basis sets must be considered due to limitations in computational resources, then M06–2X/6–31G(d) provides an overall excellent description of both nucleobase – amino acid stacking and T-shaped interactions for a range of DNA–protein π–π and π+–π interactions.

A preliminary investigation of the additivity of pi-pi or pi+-pi stacking and T-shaped interactions between natural or damaged DNA nucleobases and histidine

Previous computational studies have examined pi-pi and pi(+)-pi stacking and T-shaped interactions in nucleobase-amino acid dimers, yet it is important to investigate how additional amino acids affect these interactions since simultaneous contacts often appear in nature. Therefore, this paper investigates the geometries and binding strengths of amino acid-nucleobase-amino acid trimers, which are compared to the corresponding nucleobase-amino acid dimer interactions. We concentrate on systems containing the natural nucleobase adenine or its (cationic) damaged counterpart, 3-methyladenine, and the aromatic amino acid histidine, in both the neutral and protonated forms. This choice of molecules provides information about pi-pi and pi(+)-pi stacking and T-shaped interactions in asymmetric, biologically relevant systems. We determined that both stacked and T-shaped interactions, as well as both pi-pi and pi(+)-pi interactions, exhibit geometric additivity. To investigate the energetic additivity in our trimers, the synergy (E(syn)) and the additivity (E(add)) energy were examined. E(add) reveals that it is important to consider the interaction between the two amino acids when examining the additivity of nucleobase-amino acid interactions. Additionally, E(syn) and E(add) indicate that pi(+)-pi interactions are quite different from pi-pi interactions. The magnitude of E(add) is generally less than 2 kJ mol(-1), which suggests that these interactions are additive. However, the interaction energy analysis does not provide information about the individual interactions in the trimers. Therefore, the quantum theory of atoms in molecules (QTAIM) was implemented. We find inconsistent conclusions from our QTAIM analysis and interaction energy evaluation. However, the magnitudes of the differences between the dimer and trimer critical point properties are extremely small and therefore may not be able to yield conclusive descriptions of differences (if any) between the dimer and trimer interactions. We hypothesize that, due to the limited number of investigations of this type, it is currently unclear how QTAIM can improve our understanding of pi-pi and pi(+)-pi dimers and trimers. Therefore, future work must systematically alter the pi-pi or pi(+)-pi system to definitively determine how the geometry, symmetry, and system size alter the QTAIM analysis, which can then be used to understand biologically relevant complexes.

Effects of the biological backbone on DNA-protein stacking interactions

The pi-pi stacking (face-to-face) interactions between the five natural DNA or RNA nucleobases and the four aromatic amino acids were compared using three different types of dimers: (1) a truncated nucleoside (nucleobase) stacked with a truncated amino acid; (2) a truncated nucleoside (nucleobase) stacked with an extended amino acid; and (3) a nucleoside (extended nucleobase) stacked with a truncated amino acid. Systematic (MP2/6-31G*(0.25)) potential energy surface scans reveal important information about the effects of the deoxyribose sugar and protein backbone on the structure and binding energy between truncated nucleobase and amino acid models that are typically implemented in the literature. Most notably, electrostatic and steric interactions arising from the bulkiness of the biological backbones can change the preferred relative orientations of DNA and protein pi-systems. More importantly, the protein backbone can strengthen the stacking energy (by up to 10 kJ mol(-1)), while the deoxyribose moiety can strengthen or weaken the stacking interaction depending on the positioning of the amino acid relative to the sugar residue. These effects are likely due to additional interactions between the amino acid or nucleobase ring and the backbone in the extended monomer rather than significant changes in the properties of the biological pi-systems upon model extension. Since the present work reveals that all calculated DNA-protein stacking interactions are significant and approach the strength of other noncovalent interactions between biomolecules, both pi-pi and backbone-pi interactions must be considered when attempting to gain a complete picture of DNA-protein binding.

A quantum chemical study of repair of O6-methylguanine to guanine by tyrosine: evaluation of the winged helix-turn-helix model

The winged helix-turn-helix model for the repair of O6-MeG to guanine involving the reaction of O6-MeG with a tyrosine residue of the protein O6-alkylguanine-DNA alkyltransferase (AGT) was examined by studying the reaction mechanism and barrier energies. Molecular geometries of the species and complexes involved in the reaction, i.e. the reactant, intermediate and product complexes as well as transition states, were optimized employing density functional theory in gas phase. It was followed by single point energy calculations using density functional theory along with a higher basis set and second order M(phi)ller-Plesset perturbation theory (MP2) along with two different basis sets in gas phase and aqueous media. For the solvation calculations in aqueous media, the integral equation formalism of the polarizable continuum model (IEF-PCM) was employed. Vibrational frequency analysis was performed for each optimized structure and genuineness of transition states was ensured by visualizing the vibrational modes. It is found that tyrosine can repair O6-MeG to guanine by a two-step reaction. The present results have been compared with those obtained considering the helix-turn-helix model where the repair reaction primarily involves cysteine and occurs in a single-step. It is concluded that the repair through tyrosine envisaged in the winged helix-turn-helix model would be less efficient than that through cysteine envisaged in the helix-turn-helix model.