An ONIOM and MD Investigation of Possible Monofunctional Activity of

Article pubs.acs.org/JPCB

An ONIOM and MD Investigation of Possible Monofunctional Activity of Human 8‑Oxoguanine−DNA Glycosylase (hOgg1) Jennifer L. Kellie, Katie A. Wilson, and Stacey D. Wetmore* Department of Chemistry and Biochemistry, University of Lethbridge, 4401 University Drive West, Lethbridge, Alberta, Canada, T1K 3M4 S Supporting Information *

ABSTRACT: Since the formation of 8-oxoguanine (OG) is one of the most common DNA-damaging events, cells have evolved efficient repair processes to avoid the mutagenic effects associated with this lesion, including base excision repair (BER) initiated by hOgg1. In the present work, three distinct mechanisms for deglycosylation catalyzed by hOgg1 that represent monofunctional activity were characterized using a combination of molecular dynamics (MD) simulations on the full DNA−enzyme complex and ONIOM calculations on a truncated DNA−protein model. The initial lysine activation step common to all pathways involves proton transfer from (cationic) K249 to (anionic) C253 and subsequent active-site rearrangement to align key amino acids and/or water for the next reaction step. In the first mechanism, K249 initiates deglycosylation as the nucleophile and the resulting DNA−protein cross-link is hydrolyzed to generate an abasic site. In the remaining two mechanisms, an active-site water molecule is the nucleophile, which is activated by either K249 or D268. These latter mechanisms are supported by MD simulations that reveal an abundance of water in the active site that could function as the nucleophile. Our ONIOM model suggests that the most likely mechanism involves water nucleophile activation by K249, which allows the active-site aspartate (D268) to electrostatically stabilize the charge buildup on the sugar residue throughout the entire reaction pathway. This newly conjectured mechanism is consistent with the proposed activity of other monofunctional glycosylases. In addition to providing the first atomic level evidence for a monofunctional hOgg1 catalytic pathway, the mechanistic details revealed in the present work can be used to direct future large-scale reaction modeling on the entire DNA− protein complex, which can be coupled with experimental kinetic data to afford a reliable comparison of the potential mono- and bifunctional activity of this crucial enzyme.

■

following exposure to ionizing radiation.9−11 It has been estimated that OG is formed 103 times per (normal) cell per day,12 and OG formation has been associated with cancer13−16 and neurodegenerative diseases.9,17,18 The high mutagenicity of OG lesions has been linked to G:C → T:A transversion mutations, which commonly result from a DNA polymerase misreading the Hoogsteen face of OG as thymine during replication.19 Therefore, it is critical that cells efficiently invoke defense mechanisms, including BER, to combat the accumulation of OG. In humans, BER of OG in DNA is initiated by 8oxoguanine−DNA glycosylase (hOgg1), which has shown catalytic activity for both the deglycosylation and β-lyase steps.20,21 The proposed hOgg1 bifunctional mechanism of action involves four phases (Figure 1): (a) deglycosylation of the OG nucleotide by an active-site lysine (K249), (b) rearrangement to a ring-opened Schiff-base intermediate, (c) β-elimination of the 3′-phosphate, and (d) hydrolysis of the DNA−protein cross-link to regenerate the enzyme and release

INTRODUCTION Since cellular DNA is damaged on a regular basis,1 multiple repair pathways exist in humans to maintain our genetic information.2 The base excision repair (BER) pathway is responsible for correcting nonbulky DNA nucleobase damage.3 BER is initiated by a DNA glycosylase that identifies the damaged base and catalytically cleaves the sugar−nucleobase (N-glycosidic) bond.4−7 Specifically, the glycosidic bond is either hydrolyzed by a monofunctional glycosylase, or the damaged base is displaced by an active-site lysine or terminal proline residue by a bifunctional glycosylase. Regardless of the glycosylase activity mode, base removal results in an abasic (apurinic or apyrimidinic, AP) site, which is further processed through β-elimination of the 3′-phosphate and δ-elimination of the 5′-phosphate that is facilitated by either a bifunctional glycosylase or an AP endonuclease. Subsequently, a DNA polymerase and lyase complete the repair process by inserting the complementary canonical nucleobase according to Watson−Crick pairing. One of the most prevalent forms of nonbulky DNA damage is 7,8-dihydro-8-oxoguanine (8-oxoguanine, OG),8 which arises following attack of DNA by reactive oxygen species (ROS) generated during the normal metabolism of oxygen or © 2015 American Chemical Society

Received: April 28, 2015 Revised: May 26, 2015 Published: May 27, 2015 8013

DOI: 10.1021/acs.jpcb.5b04051 J. Phys. Chem. B 2015, 119, 8013−8023

Article

The Journal of Physical Chemistry B

to β-elimination. This suggests that it may be feasible to kinetically trap the deglycosylated intermediate, and a possible monofunctional mechanism may involve hydrolysis of the cross-linked deglycosylation product.38 Unfortunately, since kinetic isotope effect experiments on monofunctional hOgg1 activity that parallel studies on other glycosylases (such as MutY39 and hUNG240) have yet to be performed, there is no experimental evidence indicating a preference for one of these pathways. Although it is difficult to definitively compare possible monofunctional mechanisms using traditional experimental techniques, computational studies have provided essential information about the molecular details of the deglycosylation reactions facilitated by several other DNA glycosylases.41−45 However, most computational work to date on hOgg1 has focused on the bifunctional mechanism of action, investigating SN1,46−49 SN2,38,46,47,49−51 AN + DN,52 and σ-bond substitution49 mechanisms of OG deglycosylation facilitated by K249 as the nucleophile. Indeed, only a single study has examined the hydrolysis of the OG glycosidic bond by a (fully or partially activated) water molecule (OH− and HCOO−···H2O model nucleophiles, respectively).50 Nevertheless, the computational model employed in this hydrolysis study was too small to afford information about the relative importance of different monofunctional pathways. Thus, the present work uses a combination of molecular dynamics (MD) and quantum mechanical (QM) calculations to further explore the likelihood of monofunctional hOgg1 activity and characterize the most probable monofunctional mechanisms of action. MD simulations, complemented by an ONIOM conformational search, provide currently missing information on the charge states and conformational landscape of key active-site residues, as well as the availability of water in the active site to facilitate nucleotide deglycosylation. Subsequently, truncated DNA−protein (ONIOM) models are used to characterize three distinct reaction pathways. Together, this work provides the first atomic level evidence for a monofunctional hOgg1 mechanism of action. The mechanistic intricacies revealed in the present work provide crucial missing information to inspire future large-scale enzymatic modeling that can use experimental kinetic data to reliably compare both the possible mono- and bifunctional activity modes for hOgg1.

Figure 1. General proposed mechanism for the bifunctional activity of hOgg1, including the (A) deglycosylation, (B) ring-opening, (C) βelimination, and (D) enzyme regeneration steps.

the product.22,23 In addition to determining the catalytic role of K249,24−28 X-ray crystal structures of several hOgg1 mutants at various stages of the BER pathway, as well as kinetic data on hOgg1 mutants, have led to proposals that D268 is a key catalytic residue for several reaction steps,28−31 likely being responsible for electrostatically stabilizing (charged) Schiff base intermediates (Figure 1). Despite experimental evidence classifying hOgg1 as a bifunctional glycosylase, the β-lyase activity is known to be significantly slower than the glycosylase activity.26,32−35 Furthermore, there is debate in the literature regarding the relevance of hOgg1 β-lyase activity in vivo. Specifically, experimental evidence suggests that the E. coli AP endonuclease APE1 can dissociate hOgg1 prior to β-elimination of the AP site,26,33 while human AP endonuclease 1 (HAP1) increases hOgg1 glycosylase activity and uncouples lyase activity.32,36,37 This suggests that hOgg1 may exhibit monofunctional activity that is not limited by slow lyase activity due to subsequent processing by an AP endonuclease. However, to the best of our knowledge, no literature to date provides atomic level information about the possible mechanism for the monofunctional activity of hOgg1. Since the sugar−phosphate backbone near the damaged nucleoside is typically solvent exposed when bound to a DNA glycosylase, an hOgg1 monofunctional mechanism similar to that invoked by other monofunctional glycosylases can be envisioned in which a water molecule transfers a proton to an active-site residue (general base) and acts as the nucleophile for the deglycosylation step.4−7 This potential pathway would avoid the (likely slow) enzyme-regeneration step following deglycosylation assisted by K249. Alternatively, our group proposed a different monofunctional mechanism based on calculations that investigated the bifunctional mechanism of action of hOgg1 using a truncated DFT model.38 Specifically, the most stable point on the reaction surface corresponding to the bifunctional hOgg1 mechanism of action was found to be the DNA−protein cross-linked intermediate that occurs prior

■

COMPUTATIONAL DETAILS Molecular Dynamics Simulations. A crystal structure (PDB ID: 3KTU) of hOgg1 cocrystallized with DNA containing the 2′-fluoro-dOG inhibitor (FdOG, Figure 2),

Figure 2. Active site of hOgg1 bound to 2′-fluoro-dOG (FdOG, PDB ID: 3KTU). 8014


Article


the system is −3 and includes 1130 atoms (559 heavy (nonhydrogen) atoms). The ONIOM methodology was applied by dividing the computational model into three distinct regions: (1) unconstrained high-level (DFT) region, (2) unconstrained low-level (semiempirical) region, and (3) constrained low-level (semiempirical) region. The unconstrained region of the model includes the high-level (DFT) region and a portion of the lowlevel (semiempirical) region (Figure 3). Specifically, the DFT

which mimics the reactant complex, was used as a starting point for MD simulations. Initially, PROPKA 3.153−56 and chemical intuition were used to assign the protonation states of the amino acid R-groups. Since H270 has been modeled as either cationic or neutral in different computational studies,57,58 both protonation states were examined to afford the comparison that is currently missing in the literature. The effects of proton transfer from K249 to C253 were also examined by considering charged separated (K249+−C253−) and charged neutral (K249−C253) complexes (with the preferred (cationic) H270 protonation state). The parameters for 8-oxoguanine were assigned as previously reported,43 while all amino acid residues and nucleotides were assigned AMBER ff99SB59 parameters. Missing heavy atoms in D81, K82, Q125, K126, and Q325, and all hydrogen atoms were added using the LEaP module of AMBER 11.60 All DNA−protein complexes were charge neutralized by adding (22 or 23) Na+ ions and solvated in an 8.0 Å TIP3P water box using the LEaP module of AMBER 11. MD simulations of the hOgg1:DNA complex were carried out using the SANDER module of AMBER 1160 or 12,61 with a 2 fs time step, and the periodic boundary condition implemented throughout. Minimization of the solvent and ions was conducted using a 2092 kJ mol−1 Å−2 constraint on the DNA and protein for 500 steps of steepest decent minimization and 500 steps of conjugate gradient minimization. Subsequently, 1000 steps of unrestrained steepest decent and 1500 steps of unrestrained conjugate gradient minimization were performed on the entire system. The system was then heated at constant volume from 0 to 300 K over 20 ps with a 41.84 kJ mol−1 Å−2 restraint on the DNA and protein. Finally, the entire system was simulated without restraints for 20 ns at 300 K and 1 atm. All reported simulations have a stable root-mean-square deviation (RMSD) over the production run (Figure S1, Supporting Information). Trajectory analysis was completed using the ptraj module of AMBER 11. Solvent distributions are shown at a 1σ contour and were generated on the basis of a 30 Å box centered around OG using a 0.5 Å grid. Quantum Mechanical Calculations. Model Generation. Initial heavy atom coordinates were taken from the hOgg1:FdOG crystal structure (PDB ID: 3KTU). This choice is justified on the similarity of the crystal structure and MD representative structure from our preferred (cationic H270, neutral K249, and neutral C253) model (i.e., the RMSD based on an overlay of the heavy (non-hydrogen) atoms is 1.094 Å; Figure S2c, Supporting Information). Residues with heavy atoms within 10 Å of the substrate (FdOG) were included in the model. Additional residues were included to bridge small gaps (1−2 amino acids) in the backbone. Residues with Rgroups directed away from the body of the resulting model were truncated to alanine. A calcium ion near the 3′-phosphate of FdOG was included with a coordinated water molecule (HOH55C), and two new water molecules were added to complete the coordination sphere. A full list of the model contents can be found in Table S1, Supporting Information. Cleaved bonds were capped with a hydrogen atom in the same orientation as the original atom. The remaining hydrogen atoms were added by hand to ensure optimal hydrogenbonding arrangements. Prior to the conformational search and reaction surface calculations, the locations of the hydrogen atoms were optimized using PM6. The K249−C253 interaction was initially modeled as charge separated, D268 and D322 were anionic, and R324 was cationic. H270 was modeled as cationic due to the results from our MD simulations. The total charge of

Figure 3. Schematic (upper) of the high-level (black) and unconstrained low-level (gray) regions of the ONIOM model. 3D model (lower) of the high-level region of the hOgg1 reactant with unconstrained low-level atoms in green.

region contains the FdOG nucleoside, 3′-phosphate, the peptide bond between G42 and Q43 (including Cα of both residues), the R-groups of S147, K249, C253, D268, H270, Q315, and F319, and three water molecules (HOH109C, HOH55C, and HOH1A). The high-level region includes 121 atoms (63 heavy atoms) and has a charge of −1. The unconstrained portion of the low-level region includes the 5′phosphate, C5′ of DG24 (attached to the 3′-phosphate), the Rgroups of F144 and M257, the Ca2+, and the two coordinated water molecules (a total of 39 atoms, 18 heavy atoms). The remainder of the model was constrained in the low-level region. Conformational Search. In a previous study of MutY activity,43 an ONIOM conformational search of the enzyme active site provided supporting information to MD simulations regarding the conformational flexibility of active-site residues. Therefore, a similar search of the conformational space of the hOgg1 active site was carried out in the present work. The orientations of residues near the OG nucleobase, namely, F144, K249, C253, D268, H270, Q315, and F319, were included in the search, and each side chain was individually considered. 8015


Article


Figure 4. Three possible hOgg1 monofunctional mechanisms (K249−Nuc, K249−GB, and D268−GB) characterized in the present work.

K249 and C253 were modeled as neutral for the search due to evidence from MD, which was further supported by subsequent ONIOM calculations. First, the dihedral angles defining the conformation of the R-group under investigation were modified in 30° increments for all residues except K249 and Q315, which were rotated in 60° increments. PM6 single-point energy calculations were carried out on each conformation generated. Subsequently, conformers within 100 kJ mol−1 of the most stable geometry for each amino acid were optimized using our ONIOM scheme. Reaction Surfaces. For each potential hOgg1 monofunctional pathway characterized in the present work (Figure 4), the reactants, intermediates, and products were fully optimized using ONIOM. However, calculations of the force constants required to aid full transition state optimizations with the entire ONIOM model were prohibitively expensive. Therefore, reaction potential energy surfaces (PES) near the transition state were used to refine transition state geometries. This approach has been successfully used to characterize the mechanism of action of MutY,43 and data on related glycosylases (AAG, hUNG2) for which fully optimized TS could be obtained verify the accuracy of this approach.41,45 An initial guess for each transition state was generated by hand to connect the two adjacent stationary points, and then, the reaction coordinates were altered to refine the portion of the surface around the TS. The region of the reaction surface associated with each transition state was described by a different set of coordinates that reflect the reaction pathway under consideration (selected bond lengths and dihedral angles are provided in Table S2, Supporting Information). For each transition state, the two coordinates were systematically altered (the step-sizes implemented are provided in Table S2, Supporting Information), and held fixed during the ONIOM

optimizations to generate the associated region of the reaction surface. In ONIOM optimizations, the high-level region was treated with the M06-2X density functional method in conjunction with the 6-31G(d) basis set, and the low-level region was treated with the PM6 semiempirical method. Previous work justifies the use of M06-2X with a small basis set for modeling DNA repair enzymes.41,62 For all minima and transition states, the DFT region was extracted and frequency calculations were performed with M06-2X/6-31G(d) to verify the nature of the stationary point, and estimate the zero-point vibrational (ZPV) correction and the thermal correction to the Gibbs energy. In addition, it was ensured that any imaginary frequencies corresponded to the truncation points (for both minima and TS) and the reaction coordinate (for TS). This approach of estimating the ZPV correction has been successfully applied to a variety of other enzymatic reactions modeled with ONIOM.63−65 High-quality single-point energies were obtained with ONIOM(M06-2X/6-311+G(2df,2p):PM6). All QM calculations were carried out using Gaussian 09 (revisions A.02 and C.01).66,67

■

RESULTS AND DISCUSSION Charge State and Conformation of H270. As mentioned in the Computational Details section, MD simulations were initially used to investigate the effects of cationic versus neutral H270 on the active-site geometry, since H270 has been modeled as both cationic and neutral in previous computational studies of hOgg1.47,51 When H270 is modeled as neutral, substantial distortion occurs to the active site compared to the crystal structure geometry, which primarily occurs in the vicinity of H270, including D268, M271, and F319 (RMSD based on the side chains of active-site residues of 3.192 Å; Figure S2A, Supporting Information). Indeed, H270 adopts a 8016


Article

The Journal of Physical Chemistry B different orientation in the MD simulation than in the crystal structure (∠(CαCβCγNδ) = −66.5 versus 71.5° for the 3KTU crystal structure versus MD simulation, respectively; Table S3, Supporting Information). The MD predicted orientation prevents a hydrogen-bonding interaction between neutral H270 and OG (Table S4, Supporting Information). Specifically, H270 and OG become significantly separated (by 8.656 Å) compared to the crystal structure distance (3.193 Å) when H270 is neutral (Table S5, Supporting Information). Although most DNA binding proteins use several lysine and/or arginine residues to interact with the sugar−phosphate backbone at the lesion site,68 H270 is the only hOgg1 residue positioned to hydrogen bond with the backbone, and therefore, the loss of this H270−DNA interaction would likely affect enzyme binding and/or function.28 In contrast, the crystal structure orientation of the active site is better maintained when H270 is cationic (RMSD based on the side chain of the active-site residues of 1.120 Å; Figure S2B, Supporting Information). Indeed, cationic H270 maintains a hydrogen bond with the OG phosphate moiety for up to 86.0% of the MD simulation depending on the K249 and C253 protonation state (Table S4, Supporting Information). Interestingly, the distortion of the active site observed in the neutral H270 simulation is consistent with previous simulations on the apoenzyme,58 which attributed the distortion to the R46Q, R131Q, and R154H mutations (although an alternative H270 protonation state was not examined). Together, this data suggests that H270 is likely protonated in the active site, and therefore was modeled as cationic throughout the remainder of this work. Charge States and Conformations of C253 and K249. Previous computational investigations of the hOgg1 bifunctional mechanism of action have employed a cationically charged K249 residue.46,47,49,52 Indeed, there is support that OG substrate binding involves a K249+−C253− contact.69 Specifically, the dipole between the two residues is antialigned with the local dipole formed between O8 and N7H of OG (Figure 2). This dipole−dipole interaction has been proposed to assist the differentiation between undamaged guanine and OG.69 However, it has been proposed that K249 is activated during substrate binding,38 and mutational studies suggest that the active site may change the pKa of K249 to afford a neutral residue.38 When coupled with debates in the literature regarding the charge and potential K249 nucleophile activation mode,38,39,49 these studies suggest that the protonation states of K249 to C253 must be more carefully considered prior to modeling hOgg1 (bi- or monofunctional) glycosylase activity. To explore the protonation states of K249 and C253, the dynamics of the active site was initially compared when the salt bridge is charge separated (K249+−C253−) versus neutral (K249−C253). There is an increased amount of active-site water in simulations with K249+−C253− rather than K249− C253 (Figure 5 and Table S6, Supporting Information). The increased water could prevent direct interactions between the substrate and active-site residues that are necessary for catalysis. Interestingly, a hydrogen bond between SγC253 and a water molecule occurs with >50% occupancy for K249+−C253− (Table S6, Supporting Information). Interestingly, this hydrogen bond with SγC253 occurs with multiple water molecules, which consecutively form hydrogen bonds to result in >50% occupancy of this type of hydrogen-bonding interaction. These water molecules consist of both water that was originally present in the crystal structure and water that was added prior to simulation when the system was solvated. In contrast, no

Figure 5. Solvent density around OG in MD simulations with (a) K249+−C253− and (b) K249−C253.

hydrogen bonds occur between C253 and water for K249− C253 (Table S6, Supporting Information), which suggests the negative charge on C253 must be stabilized in some way (i.e., through hydrogen bonding or proton transfer). C253 adopts an orientation consistent with the crystal structure in both charged states (Table S3, Supporting Information). Although three K249 conformations are adopted in both simulations, the relative frequency of each conformation is dependent on the K249−C253 protonation state (Table S3, Supporting Information). When K249 and C253 are neutral, the most common orientation breaks the K249−C253 hydrogen-bonding interaction and positions K249 by the sugar moiety of OG, which is a good location for K249 to act as the nucleophile. Interestingly, this catalytically favorable orientation is adopted with a significantly lower frequency when K249 is cationic (10%) than neutral (51%; Table S3, Supporting Information). Furthermore, although the K249+−C253− salt bridge is present for ∼100% of the MD simulation, K249 moves away from C1′ and occupies a larger range of distances with respect to the substrate when the residues are charge separated (Table S5, Supporting Information). Thus, a combination of less active-site water and the preferable active-site conformation when C253 and K249 are neutral suggests the salt bridge is likely not stable after binding or, in other words, proton transfer likely occurs upon binding. 8017


Article


Table 1. Reaction Free Energies (ΔG, kJ mol−1) for the Three hOgg1 Monofunctional Mechanisms Investigated in the Present Worka,b

To further analyze the binding arrangement of K249 and C253, an ONIOM conformational search was considered for the charge neutral pair (Table S7, Supporting Information). Three unique binding conformations were identified for C253 (Table S8, Supporting Information), and each invokes different noncovalent interactions. The most stable orientation contains an S−H···O hydrogen bond to the K249 backbone carbonyl. This orientation is consistent with the preferred MD structure in which K249 is repositioned in the active site upon proton transfer to C253 and SγC253 hydrogen bonds to OK249 with a significant occupancy (68.9%, Table S4, Supporting Information). The two other ONIOM geometries invoke a lone-pair−π or S−H···π interaction with F144. A large number of conformations were isolated for K249 and Q315 in the ONIOM conformational search (Table S8, Supporting Information), many of which are (albeit slightly) more stable than the crystal structure geometry (Table S7, Supporting Information). Interestingly, some of the alternate structures place K249 in a better position for nucleophilic attack (d(C1′··· Nε) < 4 Å), which was also observed in the MD simulations. Most importantly, since the conformational search was carried out with both K249 and C253 in a neutral state (i.e., after Lys activation), our results suggest that the K249−C253 interaction is not stable after proton transfer. Since it appears that proton transfer within the K249+− C253− salt bridge may provide a more catalytically competent active-site conformation, the barrier for this proton transfer was considered (TS1−All, Figures 4 and 6). The proton transfer

RC1 TS1 RC2 TS2 RC3 TS3 IC1 PC

K249−Nuc

D268−GB

K249−GB

0 −2.3 −58.5 −7.0 −6.2 213.6 −28.1

0 −2.3 −58.5 NCc −47.3 210.5

0 −2.3 −58.5 NCc 1.2 174.4

123.9

17.1

a

See Figure 4 for the definitions of the mechanisms and stationary points. bONIOM(M06-2X/6-311+G(2df,2p):PM6) single-point energies including unscaled thermal corrections to the Gibbs energies. c Could not be characterized.

with our MD results and those from the ONIOM conformational search, indicating that this interaction can break following proton transfer. Together, our combined MD and ONIOM approach suggests the first step in the (mono- or bifunctional) mechanism of action of hOgg1 is likely proton transfer from K249 to C253, which activates K249 and can facilitate subsequent K249 conformational rearrangement for the next reaction (deglycosylation) step(s). Lysine Facilitated Deglycosylation and Hydrolysis of the DNA−Protein Cross-Link. As discussed above, MD and ONIOM conformational searches suggest that K249 can be positioned in the active site in a way to facilitate the first step of the base excision repair process through nucleophilic attack at C1′ of the substrate, which corresponds to the first deglycosylation pathway examined in the present study (denoted as K249−Nuc, Figure 4). This pathway involves two phases: (1) SN2 deglycosylation of OG facilitated by K249 and (2) hydrolysis of the DNA−protein cross-link to yield the AP-site product. Before deglycosylation can occur, K249 must undergo a conformational change that breaks the hydrogen bond with C253 and reorients Nε below the deoxyribose plane. This rearrangement from RC2−All to RC3−KNuc costs 52.3 kJ mol−1 (Table 1 and Figure S4, Supporting Information). Most of the rearrangement energy is associated with breaking the K249−C253 hydrogen bond, which has an estimated contribution of 51.5 kJ mol−1 (modeled by rotating the C253 Cβ−Oγ bond to remove the interaction). However, this step is also associated with other active-site motion, including C253, the 3′-phosphate, and Ca2+, which is supported by the MD simulations. After rearrangement, the transition state occurs with glycosidic bond and nucleophilic distances of 2.350 and 2.300 Å, respectively (Figure 7). Pyramidalization of N7 occurs in the transition state as the negative charge accumulating on OG is delocalized to O8 and N7 (TS3−KNuc, Figure 7). Our ONIOM model predicts the deglycosylation barrier to be ΔG = 213.6 kJ mol−1 (relative to RC1), while the resulting intermediate is exothermic (ΔG = −28.1 kJ mol−1) due to proton transfer from the cross-link to N9 of OG, and restoration of the D268−N2 hydrogen bond that breaks in the TS. The stability of the crosslinked intermediate is consistent with the results from our previous small model study.38 Since the deglycosylated intermediate is more stable than RC1, it is possible that the DNA−protein cross-link persists for some time before proceeding to the ring-opened step (bifunctional activity) or

Figure 6. Refined stationary points for the lysine activation step, including hydrogen-bond distances (Å) and angles (deg, parentheses). ONIOM(M06-2X/6-311+G(2df,2p):PM6) Mulliken charges for NH7 and O8 shown (green). For clarity, only the OG, K249, and C253 residues are included.

step is barrierless within our ONIOM model (Figure S3, Supporting Information), and the resulting activated reactant (RC2) is 58.5 kJ mol−1 more stable than RC1 (Table 1 and Figure S4, Supporting Information). Thus, although the salt bridge may be stable in the free enzyme (i.e., when the active site is solvent accessible), the environment shifts to be less polar after the substrate is bound, which activates the essential K249 residue. As anticipated, the neutral hydrogen bond is substantially longer than the charged interaction (by 0.192 Å, Figure 6), which decreases the strength of the contact and reduces the barrier to K249 rearrangement. This is consistent 8018


Article


along the DNA backbone by O3′OG and O5′OG (Figure 5), which is consistent with the location of some crystallographic water. More importantly, 11 different water molecules form 19 hydrogen bonds at the O3′OG position (Table S6, Supporting Information). Although these water molecules are transient (14 of the hydrogen bonds are occupied for less than 20% of the simulation and the highest occupancy is 36%; Table S6, Supporting Information), a water molecule that can act as the nucleophile is present for the majority of the MD simulation. In addition to water, a general base (GB) is required to activate the nucleophile for monofunctional glycosylase activity.70 Upon close examination of the active site, two residues may function in this capacity, and the ability of each residue to facilitate an alternative hOgg1 monofunctional mechanism is considered in the subsequent sections. Glycosidic Bond Hydrolysis Facilitated by D268. A commonly accepted role for an active-site aspartate in monofunctional glycosylases is activation of the nucleophile.5,6 Like many monofunctional glycosylases, hOgg1 contains an active-site aspartate (D268) that may be able to catalyze monofunctional activity. Although this would require the helix cap between D268 and the backbone of H270 and M271 to break, our ONIOM conformational search supports the possibility of such changes to the D268 conformation. Specifically, as previously observed for MutY,43 two conformations of D268 were identified (Tables S7 and S8, Supporting Information). One structure maintains the capping interaction, while the other breaks the hydrogen bond with the backbone. Interestingly, the cost of breaking the helix cap is much smaller for hOgg1 (ΔE = 12.2 kJ mol−1) than previously reported for MutY (ΔE = 51.6 kJ mol−1),43 since the D268− backbone hydrogen bond is replaced with a D268−OG interaction (involving the exocyclic amino group) in the hOgg1 active site. As previously demonstrated for MutY,43 this provides support for the conformational flexibility of D268, which may translate to D268 having the ability to act as the general base in the monofunctional activity of hOgg1. Therefore, a possible mechanism for this pathway is considered within our ONIOM scheme (D268−GB, Figure 4). Although the transition state (TS2) corresponding to D268 rearrangement could not be characterized in the D268−GB pathway, the barrier must be at least 11.2 kJ mol−1 (RC2 to RC3 energy difference; Table 1 and Figure S4, Supporting Information). In the reactant corresponding to the deglycosylation step (RC3−DGB, Figure 8), the water nucleophile interacts with D268, which assists activation of the nucleophile and orients the water for attack on the anomeric sugar carbon. K249 stabilizes the OG nucleobase in the reactant and transition state through a hydrogen bond to N3 (d(Nε−H··· N3) = 2.082−2.099 Å, Figure 8). The deglycosylation transition state occurs at d(C1′···OWat) = 2.000 Å, with a hydrolysis barrier of ΔG⧧ = 210.5 kJ mol−1 (relative to RC1). Proton transfer from the water molecule leads to an endothermic product (ΔG = 123.9 kJ mol−1 for PC−DGB, Figure 8). Although our ONIOM model suggests the barrier for hydrolysis of the K249−substrate cross-link is large, a small energetic difference is predicted between OG deglycosylation facilitated by direct attack of K249 versus hydrolysis involving D268. Therefore, it is interesting to further explore whether an alternative active-site residue can behave as the general base to activate the water nucleophile. Glycosidic Bond Hydrolysis Facilitated by K249. The results in the present work for the hydrolysis of OG facilitated

Figure 7. Refined stationary points for the K249−Nuc deglycosylation step (top), including hydrogen-bond distances (Å) and angles (deg, parentheses). For clarity, only the OG, K249, and D268 residues are included. ONIOM(M06-2X/6-31G(d):PM6) reaction potential energy surface (ΔE relative to RC1−All, kJ mol−1) near the transition state (bottom), in which each contour corresponds to 0.30 kJ mol−1.

being hydrolyzed (monofunctional activity). However, preliminary calculations on the hydrolysis of the DNA−protein crosslink (TS4−KNuc, Figure 4) indicate that the barrier is >300 kJ mol−1 (Figure S3, Supporting Information). Although the large barrier could be due to our ONIOM model (for example, the truncation and related constraints imposed on the low-level region may prohibit a required active-site rearrangement and/ or OG release), another monofunctional glycosylation route may be more feasible. Availability of a Water Nucleophile for Glycosidic Bond Hydrolysis. On the basis of the mechanism used by other monofunctional glycosylases,4−7 it is reasonable to propose a water molecule may function as the nucleophile in the hOgg1 monofunctional pathway, which would require a water molecule to be in close proximity to C1′ of OG. Similar ONIOM studies on other monofunctional glycosylases indicate that the water nucleophile typically falls within 4.0 Å of C1′ of the substrate in the reactant complex.41,43,45 Although there is never more than one water molecule within 3.6 Å of C1′OG during MD simulations on hOgg1, there is a water molecule within 3.6 Å of C1′OG for 72% of the simulation. The position of the water is also relevant for a hydrolysis mechanism. Specifically, hOgg1 active-site water close to OG mainly resides 8019


Article


characterized but costs at least 59.7 kJ mol−1 (RC2 to RC3 energy difference, Table 1 and Figure S4, Supporting Information). RC3−KGB contains a hydrogen bond between the water nucleophile and D268 that helps orient the nucleophile, and K249 is positioned such that the lone pair on Nε is below the sugar plane and ready to accept a proton from the water molecule (Figure 9). The slightly dissociative,

Figure 8. Refined stationary points for the D268−GB deglycosylation step (top), including hydrogen-bond distances (Å) and angles (deg, parentheses). For clarity, only the OG, K249, and D268 residues, as well as the water nucleophile, are included. ONIOM(M06-2X/631G(d):PM6) reaction potential energy surface near the transition state (ΔE relative to RC1−All, kJ mol−1; bottom), in which each contour corresponds to 0.30 kJ mol−1.

by D268 parallel a study from our group on the monofunctional activity of human uracil DNA glycosylase (hUNG2),45 which excises uracil from DNA. Specifically, although it was experimentally proposed that D145 activates the water nucleophile,71 the preferred pathway identified using a similar computational approach as employed in the current work (truncated ONIOM model) involves water activation by a neighboring histidine (H148), which permits D145 to provide maximum electrostatic stabilization of the charge buildup on the sugar moiety throughout the entire reaction. For hOgg1, representative structures can be identified from MD simulations with a water molecule located between the deoxyribose moiety and K249. Therefore, a possible reaction pathway was considered in which K249 activates the water nucleophile (K249−GB, Figure 4), which will determine whether this residue can play an important role in the monofunctional mechanism of action of hOgg1. The transition state for reorganization of the active site to place the water nucleophile and K249 general base in a position to initiate the deglycosylation reaction could not be

Figure 9. Refined stationary points for the K249−GB deglycosylation step (top), including hydrogen-bond distances (Å) and angles (deg, parentheses). For clarity, only the OG, K249, and D268 residues, as well as the water nucleophile, are included. ONIOM(M06-2X/631G(d):PM6) reaction potential energy surface (ΔE relative to RC1− All, kJ mol−1) near the transition state (bottom), in which each contour corresponds to 0.25 kJ mol−1.

but still synchronous, deglycosylation transition state occurs with a glycosidic bond length of 2.750 Å and a nucleophile distance of 2.250 Å (TS3−KGB, Figure 9). In the transition state, an interaction between the water nucleophile and D268 is broken, and partial proton transfer to K249 has occurred. The associated barrier is 174.4 kJ mol−1 (relative to RC1). Immediately following the transition state, a proton on the water nucleophile is transferred to N9 of the OG anion in association with motion that places N9 of the nucleobase below the plane of the sugar. This occurs spontaneously, which leads 8020


Article

The Journal of Physical Chemistry B to a discontinuity in the upper right corner of the reaction PES near the transition state (Figure 9). The final product is destabilized with respect to the overall reactant (ΔG = 17.1 kJ mol−1). A similar proton transfer was observed in small model studies of OG deglycosylation (albeit from cationic Lys rather than a water nucleophile as in the current mechanism).46,49 In the product complex, D268 interacts with both the new 1′hydroxyl and N2 of OG. Although the OG deglycosylation barrier predicted by our truncated ONIOM model for each of the three pathways is greater than that anticipated for an enzyme-catalyzed reaction, our results suggest that the mechanism involving water activation by K249 may be the most likely route through which hOgg1 can exhibit monofunctional activity. This mechanism points to the important role of K249 regardless of the (mono- or bifunctional) activity mode of the enzyme. This role is consistent with experimental studies that show mutation of K249 eliminates enzymatic activity.24−26 Furthermore, despite previous computational studies on the hOgg1 mechanism of action neglecting to include D268 in the model,46,47,52 D268 also plays a critical role in the monofunctional mechanism characterized in the present work. Specifically, although an active-site aspartate has been implicated to be the general base in the mechanism of action of other monofunctional glycosylases,5,6 and a computational study has confirmed this role for MutY,43 proton transfer from the water nucleophile to another residue in the active site allows D268 to most effectively stabilize the formation of the cationic charge on the sugar moiety throughout the reaction pathway. Indeed, the final hOgg1 monofunctional mechanism is consistent with that characterized using a similar computational approach for (monofunctional) hUNG2,45 which suggests activation of the water nucleophile by a neighboring histidine is more likely than by an active-site aspartate. This proposed role is consistent with mutational studies that found enzymatic activity is eliminated when D268 is replaced with a neutral residue (alanine or asparagine) but maintained when replaced with a charged residue (glutamate).28−31 Nevertheless, the exact hOgg1 monofunctional activity mode must be verified using largescale models of the entire DNA−protein complex, which will permit accurate comparisons to experimentally measured kinetic data.

to activate the nucleophile. For the first time, our ONIOM model suggests that the most likely mechanism of monofunctional hOgg1 action uses a water nucleophile and K249 acts as the general base that shuttles a proton from the nucleophile to the OG− leaving group. Nevertheless, D268 is also essential for this pathway, likely participating in the deglycosylation reaction step. Although this newly conjectured mechanism is consistent with the proposed activity of other monofunctional glycosylases, the relative importance of the three mechanisms characterized for the first time in the present work using truncated ONIOM models must be confirmed using large-scale models of the entire DNA−protein complex. These large-scale models will allow for greater flexibility of the active site, as well as permit frequency calculations on the entire model and full transition state optimizations. Nevertheless, this study provides the first atomic level details of possible monofunctional pathways, and therefore represents a critical step toward understanding different (mono- and bifunctional) activity modes of this essential repair enzyme.

CONCLUSIONS Due to evidence that hOgg1 behaves as a monofunctional glycosylase (i.e., no lyase activity) in vivo, the present work uses a combined MD and ONIOM approach to characterize possible mechanisms for the monofunctional activity of hOgg1. Three distinct pathways were identified, which all begin with proton transfer from K249 to C253 and a subsequent active-site rearrangement that orients K249 and/ or D268 for the next reaction step. In the first mechanism, K249 acts as the nucleophile in the deglycosylation step, which parallels the first step in the proposed bifunctional hOgg1 mechanism of action. After depurination, the resulting DNA− protein cross-link is hydrolyzed to yield an AP site and regenerate the enzyme. Although a previous DFT study from our group suggests this pathway would divert the stable DNA− protein cross-linked intermediate from β-elimination of the backbone, our ONIOM model predicts the hydrolysis of the cross-link to be energetically prohibited. Nevertheless, MD simulations support a mechanism that invokes a water molecule as the nucleophile and either K249 or D268 as the general base

ACKNOWLEDGMENTS We thank the Natural Sciences and Engineering Research Council of Canada (NSERC), the Canadian Research Chair Program, and the Canadian Foundation for Innovation (CFI) for financial support. J.L.K. acknowledges NSERC (CGS-D) and the University of Lethbridge, while K.A.W. acknowledges NSERC (USRA), for student scholarships. Computational resources from the Upscale and Robust Abacus for Chemistry in Lethbridge (URACIL) and those provided by Westgrid and Compute/Calcul Canada are greatly appreciated.

■

ASSOCIATED CONTENT

* Supporting Information S

ONIOM model description (Table S1); ONIOM reaction coordinates (Table S2); key dihedral angles from MD simulations (Table S3); active-site hydrogen bonding from MD (Table S4); key distances from MD simulations (Table S5); active-site water hydrogen bonding from MD (Table S6); ONIOM conformational search data (Tables S7 and S8); RMSD from MD simulations (Figure S1); crystal structure− MD overlays (Figure S2); PES for K249 activation (Figure S3); full citation for refs 43, 44, and 47; and ONIOM energetics and coordinates for all stationary points. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcb.5b04051.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: (403) 329-2323. Fax: (403) 329-2057. Notes

The authors declare no competing financial interest.

■

■

■

REFERENCES

(1) Hoeijmakers, J. H. J. DNA Damage, Aging, and Cancer. New Engl. J. Med. 2009, 361, 1475−1485. (2) Jackson, S. P.; Bartek, J. The DNA-Damage Response in Human Biology and Disease. Nature (London, U. K.) 2009, 461, 1071−1078. (3) Zharkov, D. O. Base Excision DNA Repair. Cell. Mol. Life Sci. 2008, 65, 1544−1565. (4) Schermerhorn, K. M.; Delaney, S. A Chemical and Kinetic Perspective on Base Excision Repair of DNA. Acc. Chem. Res. 2014, 47, 1238−1246. 8021


Article


F319, Three Amino Acids of the 8-Oxoguanine-Binding Pocket. Nucleic Acids Res. 2004, 32, 570−578. (26) Dalhus, B.; Forsbring, M.; Helle, I. H.; Vik, E. S.; Forstrøm, R. J.; Backe, P. H.; Alseth, I.; Bjørås, M. Separation-of-Function Mutants Unravel the Dual-Reaction Mode of Human 8-Oxoguanine DNA Glycosylase. Structure (Cambridge, MA, U. S.) 2011, 19, 117−127. (27) Kuznetsova, A. A.; Kuznetsov, N. A.; Ishchenko, A. A.; Saparbaev, M. K.; Fedorova, O. S. Step-by-Step Mechanism of DNA Damage Recognition by Human 8-Oxoguanine DNA Glycosylase. Biochim. Biophys. Acta 2014, 1840, 387−395. (28) Koval, V.; Knorre, D.; Fedorova, O. Structural Features of the Interaction between Human 8-Oxoguanine DNA Glycosylase Hogg1 and DNA. Acta Naturae 2014, 6, 52. (29) Norman, D. P. G.; Chung, S. J.; Verdine, G. L. Structural and Biochemical Exploration of a Critical Amino Acid in Human 8Oxoguanine Glycosylase. Biochemistry 2003, 42, 1564−1572. (30) Krokan, H. E.; Standal, R.; Slupphaug, G. DNA Glycosylases in the Base Excision Repair of DNA. Biochem. J. 1997, 325, 1−16. (31) Rosenquist, T. A.; Zharkov, D. O.; Grollman, A. P. Cloning and Characterization of a Mammalian 8-Oxoguanine DNA Glycosylase. Proc. Natl. Acad. Sci. U. S. A. 1997, 94, 7429−7434. (32) Zharkov, D. O.; Rosenquist, T. A.; Gerchman, S. E.; Grollman, A. P. Substrate Specificity and Reaction Mechanism of Murine 8Oxoguanine-DNA Glycosylase. J. Biol. Chem. 2000, 275, 28607− 28617. (33) Hill, J. W.; Hazra, T. K.; Izumi, T.; Mitra, S. Stimulation of Human 8-Oxoguanine-DNA Glycosylase by AP-Endonuclease: Potential Coordination of the Initial Steps in Base Excision Repair. Nucleic Acids Res. 2001, 29, 430−438. (34) Kuznetsov, N. A.; Koval, V. V.; Zharkov, D. O.; Nevinsky, G. A.; Douglas, K. T.; Fedorova, O. S. Kinetics of Substrate Recognition and Cleavage by Human 8-Oxoguanine-DNA Glycosylase. Nucleic Acids Res. 2005, 33, 3919−3931. (35) Krishnamurthy, N.; Haraguchi, K.; Greenberg, M. M.; David, S. S. Efficient Removal of Formamidopyrimidines by 8-Oxoguanine Glycosylases. Biochemistry 2008, 47, 1043−1050. (36) Vidal, A. E.; Hickson, I. D.; Boiteux, S.; Radicella, J. P. Mechanism of Stimulation of the DNA Glycosylase Activity of Hogg1 by the Major Human AP Endonuclease: Bypass of the AP Lyase Activity Step. Nucleic Acids Res. 2001, 29, 1285−1292. (37) Morland, I.; Luna, L.; Gustad, E.; Seeberg, E.; Bjørås, M. Product Inhibition and Magnesium Modulate the Dual Reaction Mode of Hogg1. DNA Repair 2005, 4, 381−387. (38) Kellie, J. L.; Wetmore, S. D. Mechanistic and Conformational Flexibility of the Covalent Linkage Formed During β-Lyase Activity on an AP-Site: Application to Hogg1. J. Phys. Chem. B 2012, 116, 10786− 10797. (39) McCann, J. A. B.; Berti, P. J. Transition-State Analysis of the DNA Repair Enzyme MutY. J. Am. Chem. Soc. 2008, 130, 5789−5797. (40) Werner, R. M.; Stivers, J. T. Kinetic Isotope Effect Studies of the Reaction Catalyzed by Uracil DNA Glycosylase: Evidence for an Oxocarbenium Ion-Uracil Anion Intermediate. Biochemistry 2000, 39, 14054−14064. (41) Rutledge, L. R.; Wetmore, S. D. Modeling the Chemical Step Utilized by Human Alkyladenine DNA Glycosylase: A Concerted Mechanism Aids in Selectively Excising Damaged Purines. J. Am. Chem. Soc. 2011, 133, 16258−16269. (42) Brunk, E.; Arey, J. S.; Rothlisberger, U. Role of Environment for Catalysis of the DNA Repair Enzyme MutY. J. Am. Chem. Soc. 2012, 134, 8608−8616. (43) Kellie, J. L.; Wilson, K. A.; Wetmore, S. D. Standard Role for a Conserved Aspartate or More Direct Involvement in Deglycosylation? An ONIOM and MD Investigation of Adenine−DNA Glycosylase. Biochemistry 2013, 52, 8753−8765. (44) Alexandrova, A. N. Promiscuous DNA Alkyladenine Glycosylase Dramatically Favors a Bound Lesion over Undamaged Adenine. Biophys. Chem. 2010, 152, 118−127.

(5) Stivers, J. T.; Jiang, Y. L. A Mechanistic Perspective on the Chemistry of DNA Repair Glycosylases. Chem. Rev. 2003, 103, 2729− 2759. (6) Berti, P. J.; McCann, J. A. B. Toward a Detailed Understanding of Base Excision Repair Enzymes: Transition State and Mechanistic Analyses of N-Glycoside Hydrolysis and N-Glycoside Transfer. Chem. Rev. 2006, 106, 506−555. (7) Scott, T. L.; Rangaswamy, S.; Wicker, C. A.; Izumi, T. Repair of Oxidative DNA Damage and Cancer: Recent Progress in DNA Base Excision Repair. Antioxid. Redox Signaling 2014, 20, 708−726. (8) Kamiya, H. Mutagenic Potentials of Damaged Nucleic Acids Produced by Reactive Oxygen/Nitrogen Species: Approaches Using Synthetic Oligonucleotides and Nucleotides. Nucleic Acids Res. 2003, 31, 517−531. (9) Halliwell, B. Oxidative Stress and Neurodegeneration: Where Are We Now? J. Neurochem. 2006, 97, 1634−1658. (10) Cadet, J.; Douki, T.; Ravanat, J.-L. Oxidatively Generated Base Damage to Cellular DNA. Free Radicals Biol. Med. 2010, 49, 9−21. (11) Cadet, J.; Carell, T.; Cellai, L.; Chatgilialoglu, C.; Gimisis, T.; Miranda, M.; O’Neill, P.; Ravanat, J. L.; Robert, M. DNA Damage and Radical Reactions: Mechanistic Aspects, Formation in Cells and Repair Studies. Chimia 2008, 62, 742−749. (12) Loft, S.; Danielsen, P. H.; Mikkelsen, L.; Risom, L.; Forchhammer, L.; M?ller, P. Biomarkers of Oxidative Damage to DNA and Repair. Biochem. Soc. Trans. 2008, 36, 1071−1076. (13) Tsuzuki, T.; Nakatsu, Y.; Nakabeppu, Y. Significance of ErrorAvoiding Mechanisms for Oxidative DNA Damage in Carcinogenesis. Cancer Sci. 2007, 98, 465−470. (14) Paz-Elizur, T.; Sevilya, Z.; Leitner-Dagan, Y.; Elinger, D.; Roisman, L. C.; Livneh, Z. DNA Repair of Oxidative DNA Damage in Human Carcinogenesis: Potential Application for Cancer Risk Assessment and Prevention. Cancer Lett. 2008, 266, 60−72. (15) Kryston, T. B.; Georgiev, A. B.; Pissis, P.; Georgakilas, A. G. Role of Oxidative Stress and DNA Damage in Human Carcinogenesis. Mutat. Res., Fundam. Mol. Mech. Mutagen. 2011, 711, 193−201. (16) Ziech, D.; Franco, R.; Pappa, A.; Panayiotidis, M. I. Reactive Oxygen Species (ROS) − Induced Genetic and Epigenetic Alterations in Human Carcinogenesis. Mutat. Res., Fundam. Mol. Mech. Mutagen. 2011, 711, 167−173. (17) Evans, M. D.; Dizdaroglu, M.; Cooke, M. S. Oxidative DNA Damage and Disease: Induction, Repair and Significance. Mutat. Res. 2004, 567, 1−61. (18) Tanrikulu, S.; Dogru-Abbasoglu, S.; Ozderya, A.; Ademoglu, E.; Karadag, B.; Erbil, Y.; Uysal, M. The 8-Oxoguanine DNA NGlycosylase 1 (hOgg1) Ser326Cys Variant Affects the Susceptibility to Graves’ Disease. Cell Biochem. Funct. 2011, 29, 244−248. (19) Markkanen, E.; Hubscher, U.; van Loon, B. Regulation of Oxidative DNA Damage Repair: The Adenine: 8-Oxo-Guanine Problem. Cell Cycle 2012, 11, 1070−1075. (20) Faucher, F.; Doublie, S.; Jia, Z. C. 8-Oxoguanine DNA Glycosylases: One Lesion, Three Subfamilies. Int. J. Mol. Sci. 2012, 13, 6711−6729. (21) Rogacheva, M. V.; Kuznetsova, S. A. Repair of 8-Oxoguanine in DNA. The Mechanisms of Enzymatic Catalysis. Russ. Chem. Rev. 2008, 77, 765−788. (22) Banerjee, A.; Verdine, G. L. A Nucleobase Lesion Remodels the Interaction of Its Normal Neighbor in a DNA Glycosylase Complex. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 15020−15025. (23) Fromme, J. C.; Bruner, S. D.; Yang, W.; Karplus, M.; Verdine, G. L. Product-Assisted Catalysis in Base-Excision DNA Repair. Nat. Struct. Biol. 2003, 10, 204−211. (24) Nash, H. M.; Lu, R. Z.; Lane, W. S.; Verdine, G. L. The Critical Active-Site Amine of the Human 8-Oxoguanine DNA Glycosylase, hOgg1: Direct Identification, Ablation and Chemical Reconstitution. Chem. Biol. 1997, 4, 693−702. (25) van der Kemp, P. A.; Charbonnier, J. B.; Audebert, M.; Boiteux, S. Catalytic and DNA-Binding Properties of the Human Ogg1 DNA N-Glycosylase/Ap Lyase: Biochemical Exploration of H270, Q315 and 8022


Article


Oniom Study and Principal Component Analysis. J. Mol. Model. 2012, 18, 2227−2240. (64) Ke, Z.; Abe, S.; Ueno, T.; Morokuma, K. Catalytic Mechanism in Artificial Metalloenzyme: Qm/Mm Study of Phenylacetylene Polymerization by Rhodium Complex Encapsulated in Apo-Ferritin. J. Am. Chem. Soc. 2012, 134, 15418−15429. (65) Chung, L. W.; Sameera, W. M. C.; Ramozzi, R.; Page, A. J.; Hatanaka, M.; Petrova, G. P.; Harris, T. V.; Li, X.; Ke, Z.; Liu, F.; et al. The ONIOM Method and Its Applications. Chem. Rev. 2015, DOI: 10.1021/cr5004419. (66) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. Gaussian 09, revision A.02; Gaussian, Inc.: Wallingford, CT, 2009. (67) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. Gaussian 09, revision C.01; Gaussian, Inc.: Wallingford, CT, 2010. (68) Voet, D.; Voet, J. G. Biochemistry, 3rd ed.; John Wiley & Sons: New Jersey, 2004. (69) Banerjee, A.; Yang, W.; Karplus, M.; Verdine, G. L. Structure of a Repair Enzyme Interrogating Undamaged DNA Elucidates Recognition of Damaged DNA. Nature (London, U. K.) 2005, 434, 612−618. (70) Wolfenden, R. Degrees of Difficulty of Water-Consuming Reactions in the Absence of Enzymes. Chem. Rev. 2006, 106, 3379− 3396. (71) Mol, C. D.; Arvai, A. S.; Slupphaug, G.; Kavli, B.; Alseth, I.; Krokan, H. E.; Tainer, J. A. Crystal Structure and Mutational Analysis of Human Uracil-DNA Glycosylase: Structural Basis of Specificity and Catalysis. Cell (Cambridge, MA, U. S.) 1995, 80, 869−878.

(45) Przybylski, J. L.; Wetmore, S. D. A Qm/Qm Investigation of the hUNG2 Reaction Surface: The Untold Tale of a Catalytic Residue. Biochemistry 2011, 50, 4218−4227. (46) Schyman, P.; Danielsson, J.; Pinak, M.; Laaksonen, A. Theoretical Study of the Human DNA Repair Protein hOgg1 Activity. J. Phys. Chem. A 2005, 109, 1713−1719. (47) Calvaresi, M.; Bottoni, A.; Garavelli, M. Computational Clues for a New Mechanism in the Glycosylase Activity of the Human DNA Repair Protein hOgg1. A Generalized Paradigm for Purine-Repairing Systems? J. Phys. Chem. B 2007, 111, 6557−6570. (48) Zheng, Y.; Xue, Y.; Yan, S. G. The Effects of Oxidation and Protonation on the N-Glycosidic Bond Stability of 8-Oxo-2′Deoxyguanosine: Dft Study. THEOCHEM 2008, 860, 52−57. (49) Š ebera, J.; Trantírek, L.; Tanaka, Y.; Sychrovský, V. Pyramidalization of the Glycosidic Nitrogen Provides the Way for Efficient Cleavage of the N-Glycosidic Bond of 8-Oxog with the hOgg1 DNA Repair Protein. J. Phys. Chem. B 2012, 116, 12535− 12544. (50) Shim, E. J.; Przybylski, J. L.; Wetmore, S. D. Effects of Nucleophile, Oxidative Damage, and Nucleobase Orientation on the Glycosidic Bond Cleavage in Deoxyguanosine. J. Phys. Chem. B 2010, 114, 2319−2326. (51) Šebera, J.; Trantírek, L.; Tanaka, Y.; Nencka, R.; Fukal, J.; Sychrovský, V. The Activation of N-Glycosidic Bond Cleavage Performed by Base-Excision Repair Enzyme Hogg1; Theoretical Study of the Role of Lys 249 Residue in Activation of G, Oxog and Fapyg. RSC Adv. 2014, 4, 44043−44051. (52) Osakabe, T.; Fujii, Y.; Hata, M.; Tsuda, M.; Neya, S.; Hoshino, T. Quantum Chemical Study on Base Excision Mechanism of 8Oxoguanine DNA Glycosylase: Substrate-Assisted Catalysis of the NGlycosidic Linkage Cleavage Reaction. Chem-Bio Inf. J. 2004, 4, 73− 92. (53) Bas, D. C.; Rogers, D. M.; Jensen, J. H. Very Fast Prediction and Rationalization of pK(a) Values for Protein-Ligand Complexes. Proteins: Struct., Funct., Bioinf. 2008, 73, 765−783. (54) Li, H.; Robertson, A. D.; Jensen, J. H. Very Fast Empirical Prediction and Rationalization of Protein pK(a) Values. Proteins: Struct., Funct., Bioinf. 2005, 61, 7040−721. (55) Olsson, M. H. M.; Sondergaard, C. R.; Rostkowski, M.; Jensen, J. H. Propka3: Consistent Treatment of Internal and Surface Residues in Empirical pK(a) Predictions. J. Chem. Theory Comput. 2011, 7, 525−537. (56) Sondergaard, C. R.; Olsson, M. H. M.; Rostkowski, M.; Jensen, J. H. Improved Treatment of Ligands and Coupling Effects in Empirical Calculation and Rationalization of pK(a) Values. J. Chem. Theory Comput. 2011, 7, 2284−2295. (57) Pinak, M. Impact of Radiation Damage on DNA Determined by Computational Simulation. Jpn. J. Health Phys. 2004, 39, 35−41. (58) Anderson, P. C.; Daggett, V. The R46Q, R131Q and R154H Polymorphs of Human DNA Glycosylase/β-Lyase hOgg1 Severely Distort the Active Site and DNA Recognition Site but Do Not Cause Unfolding. J. Am. Chem. Soc. 2009, 131, 9506−9515. (59) Hornak, V.; Abel, R.; Okur, A.; Strockbine, B.; Roitberg, A.; Simmerling, C. Comparison of Multiple Amber Force Fields and Development of Improved Protein Backbone Parameters. Proteins: Struct., Funct., Bioinf. 2006, 65, 712−725. (60) Case, D. A.; Darden, T. A.; Cheatham, T. E., III; Simmerling, C. L.; Wang, J.; Duke, R. E.; Luo, R.; Crowley, M.; Walker, R. C.; Zhang, W.; et al. Amber 11; University of California: San Francisco, CA, 2010. (61) Case, D. A.; Darden, T. A.; Cheatham, T. E., III; Simmerling, C. L.; Wang, J.; Duke, R. E.; Luo, R.; Walker, R. C.; Zhang, W.; Merz, K. M.; et al. Amber 12; University of California: San Francisco, CA, 2012. (62) Kellie, J. L.; Wetmore, S. D. Selecting DFT Methods for Use in Optimizations of Enzyme Active Sites: Applications to Oniom Treatments of DNA Glycosylases. Can. J. Chem. 2013, 91, 559−572. (63) Treesuwan, W.; Hirao, H.; Morokuma, K.; Hannongbua, S. Characteristic Vibration Patterns of Odor Compounds from BreadBaking Volatiles Upon Protein Binding: Density Functional and 8023


An ONIOM and MD Investigation of Possible Monofunctional Activity of

Recommend Documents