Effect of Hydrogen Bonding and Partial ... - ACS Publications

Feb 2, 2018 - Research School of Chemistry, Australian National University, Canberra, Australian Capital Territory 2600, Australia. J. Phys. Chem. A ,...
0 downloads 0 Views 1MB Size
Article Cite This: J. Phys. Chem. A XXXX, XXX, XXX−XXX

pubs.acs.org/JPCA

Effect of Hydrogen Bonding and Partial Deprotonation on the Oxidation of Peptides Published as part of The Journal of Physical Chemistry virtual special issue “Manuel Yáñez and Otilia Mó Festschrift”. Bun Chan,*,†,¶ Christopher J. Easton,‡ and Leo Radom*,¶ †

Graduate School of Engineering, Nagasaki University, Bunkyo 1-14, Nagasaki 852-8521, Japan School of Chemistry, University of Sydney, Sydney, New South Wales 2006, Australia ‡ Research School of Chemistry, Australian National University, Canberra, Australian Capital Territory 2600, Australia ¶

S Supporting Information *

ABSTRACT: In a recent computational study, we found that hydrogen bonding/partial deprotonation facilitates subsequent electron transfer from amides to HO•. We have now analyzed these and related reactions with a glycine derivative as a model peptide, investigating not only reaction energies but also barriers for the individual steps. We find that partial deprotonation not only assists subsequent electron transfer (a sequential proton-loss electron-transfer (SPLET)-type reaction pathway) but also promotes sequential hydrogenatom transfer (HAT, in a sequential proton-loss hydrogenatom-transfer (SPLHAT)-type process), both being potential alternatives to direct HAT as routes for peptide oxidation. Each of these alternative pathways is calculated to have energy requirements that make them accessible and competitive. These oxidative processes may produce α-carbon-centered peptide radicals that, through deprotonation, are readily oxidized to the corresponding imines. We have also examined the possibility of competing reactions of amino acid side chains by comparing reactions of the glycine model with those of an analogous valine derivative. We find that, while the side chains of amino acids are more reactive toward direct HAT, a preceding partial deprotonation instead continues to favor the SPLET- and SPLHAT-type reactions, leading to the production of α-carbon-centered peptide radicals. Taken together, these processes have broad implications that impact many aspects of the science and utility of peptides.



INTRODUCTION The reactivity of peptides toward oxidation is important in diverse areas of biology, chemistry, and materials science. For example, oxidative processes are extensively involved in enzyme catalysis of peptide substrates and the synthesis of peptide secondary metabolites,1−3 as well as being closely associated with peptide damage resulting from oxidative stress.4 Conversely, the resistance of peptides toward oxidation is key to their stability and use in oxidative environments.5 This has led to extensive continuing research, including a number of computational studies by others6 and by us,7−11 aimed at a better understanding of the underlying chemistry. Very recently,11 we reported the substantial effect of hydrogen bonding and associated partial deprotonation on the oxidative susceptibility of amides. Given the relevance to peptides with their ubiquitous hydrogen bonds, this phenomenon is potentially of great significance. Indeed, our results suggest that such hydrogen-bonding interactions can even alter the course of peptide oxidation. Because of these potential ramifications, in the present study, we comprehensively explore and develop our preliminary findings. In particular, (i) we have analyzed a model glycine derivative instead of a simple amide; © XXXX American Chemical Society

(ii) we have examined additional reaction pathways and investigated not only reaction energies but also barriers for the individual steps, which is an important consideration because a number of closely related processes have been found to be kinetically controlled/contrathermodynamic;5,7−9 and, (iii) finally, we have also studied a valine derivative as a representative of peptide amino acid residues with a side chain because the behavior of glycine compared with such residues has sometimes been found to be anomalous.12−14



COMPUTATIONAL DETAILS Wave function and density functional theory calculations were carried out with Gaussian 09.15 Geometries were optimized at the M05-2X/6-31+G(d) level16 with inclusion of the effect of continuum solvation using the SMD model.17 The SMD solvation model has been optimized for several theoretical procedures, among which M05-2X/6-31G(d) is one of the best Received: November 30, 2017 Revised: January 16, 2018

A

DOI: 10.1021/acs.jpca.7b11797 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

[MeC(O)NH−CH2−C(O)NHMe, 6] instead of the amide 1, in order to better model the reactivity of an amino acid residue within a peptide. In addition, we have calculated not only reaction energies but also barriers for the various reaction pathways. Both SPLET-type reaction at either of the amide groups in 6 and direct HAT can produce the α-carbon-centered radical 7. The reaction schemes for these processes are displayed in Figure 2. The energetics for direct HAT by HO• from the α-carbon of 6 to produce 7 establishes a baseline for comparison. The calculated free energy (ΔG) for this reaction is −153.6 kJ mol−1, compared with −117.4 kJ mol−1 (ΔH = −112.6 kJ mol−1 in Figure 1) in our simpler model 1. The higher exergonicity reflects the generation of the captodatively stabilized glycyl radical 7. This reaction needs to overcome a barrier of 40.2 kJ mol−1 in order to proceed. In comparison, the top reaction pathway of Figure 2 represents the SPLET-type reaction of the N-terminal amido group of 6 to give the radical 7. The process begins with the formation of a hydrogen-bonded complex (8) between HO− and the amide N−H. This reaction is unfavorable entropically and is slightly endergonic (by 22.8 kJ mol−1). Further migration of the proton to produce the alternative complex 9 is mildly exergonic (−7.6 kJ mol−1), with a very small barrier (1.8 kJ mol−1). Of course, at an ambient temperature of 298 K, such a small barrier would mean that 8 and 9 effectively correspond to a single “low-barrier hydrogen-bonded” intermediate. This would also be applicable to intermediates 14 and 15, to be discussed shortly. The proper representation of such species is in itself a subject of considerable interest.25 Nonetheless, the main focus of the present study is on steps subsequent to the formation of these initial hydrogen-bonded complexes, and as a result, the exact nature of the hydrogen bond is unlikely to affect the discussion that follows. Electron transfer from the complex 9 to an HO• radical to yield the radical 10 is quite favorable, with a reaction free energy of −38.7 kJ mol−1 and a free energy barrier of 15.1 kJ mol−1. Subsequent exergonic dissociation of water (−29.0 kJ mol−1) gives the N-centered radical 11. From this point, generation of the α-carbon-centered radical 7 could occur in a multistep tautomerization. The first step is deprotonation from the α-carbon of 11 to give the radical anion 12, which is associated with a reaction free energy of −114.7 kJ mol−1 and a barrier of 25.8 kJ mol−1. The production of 12 is followed by a slightly endergonic (two-step) reprotonation at nitrogen (with an overall reaction energy of 13.6 kJ mol−1) to give 7. The barrier for the final dissociation of HO− is +20.4 kJ mol−1. The bottom reaction pathway of Figure 2 corresponds to the SPLET-type reaction of the C-terminal amido group of 6. In this case, when hydrogen bonding and the associated partial deprotonation at N−H produces the complexes 14 and 15, electron transfer to give 16 is also facilitated. Following dissociation of water from 16, deprotonation from the α-carbon of the nitrogen-centered radical 17 to give 18 has a substantial barrier of 62.6 kJ mol−1. For an amidyl radical in an extended peptide chain, deprotonation from the carbon adjacent to either the amide carbonyl, which is analogous to the conversion of 17 to 18, or nitrogen, which corresponds to the reaction of 11 to give 12, is conceivable. The relatively much lower barrier for the deprotonation of 11 to give 12 (of 25.8 kJ mol−1) indicates that the latter is more likely, as might have been expected. New Pathways That May Contribute to Peptide Oxidation. The calculations identify yet another pathway for

performing methods.17 We use this method in the present study but with supplementation of the 6-31G(d) basis set with standard diffuse functions [6-31+G(d)] in order to better describe anionic species. Improved gas-phase single-point energies were obtained using the high-level G4(MP2)-6X composite procedure.18 Zero-point vibrational energies (ZPVEs) and thermal corrections for 298 K enthalpies (ΔH298) and entropies (S298) were obtained using M05-2X/631+G(d) harmonic vibrational frequencies scaled by 0.9631 (ZPVE), 0.9504 (ΔH298), or 0.9445 (S298).19 The total energy of each species comprises the gas-phase G4(MP2)-6X electronic energy and SMD[M05-2X/6-31+G(d)] solvation energy, ZPVE, ΔH298, and S298 components. Barriers for electron-transfer processes were estimated with the standard Marcus equation20 using (structural) reorganization energies between the initial and the electron-transferred species. We have recently applied such an approach to the ionization process of superbases.21 Unless otherwise noted, relative energies in the text refer to aqueous condensed-phase free energies at 298 K in kJ mol−1.



RESULTS AND DISCUSSION Background. In our previous study,11 we employed Nmethylacetamide (1) as a model amide. HO• is the most likely hydrogen-abstracting species in reactions involving hydrogen peroxide,9 and one possible pathway for oxidation of 1 to give 2 is by direct hydrogen-atom transfer (HAT)22,23 to HO• (Figure 1). Additionally, we found that when the amide group in 1 is

Figure 1. Possible pathways from N-methylacetamide (1) to the radical 2 and the associated reaction enthalpies [G4(MP2)-6X + SMD aqueous solvation, ΔH, 298 K, kJ mol−1].

partially deprotonated as a result of N−H hydrogen bonding with a strong base (HO−), electron transfer to HO• from the resulting anion 3 proceeds with favorable thermodynamics (−48.4 kJ mol−1) to produce the radical 4. Subsequent dissociation of water yields 5, from which a tautomeric rearrangement involving the N-methyl group then produces the carbon-centered radical 2. The reaction of 1 to give 4 via 3 has similarities to the process termed SPLET (sequential proton loss electron transfer)6,24 in that partial deprotonation of 1 affords the anion 3, facilitating the subsequent electron transfer to produce 4. It is potentially significant as providing an alternative to HAT as a way of generating the radical 2. Analysis of the Effect of Partial Deprotonation on Oxidation of a Glycine Derivative. In the present study, we have improved and expanded on our previous theoretical investigation,11 first through analysis of the glycine derivative B

DOI: 10.1021/acs.jpca.7b11797 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

Figure 2. Reactions of the glycine derivative [MeC(O)NH−CH2−C(O)NHMe, 6] with HO• under neutral and basic conditions and the associated reaction free energies and barriers (in parentheses) (G4(MP2)-6X + SMD aqueous solvation, ΔG, 298 K, kJ mol−1).

the conversion of 6 to 7. The radical anion 13 can be generated by partial deprotonation-assisted HAT from 9 to HO• (9 → 13), which has a barrier of 19.4 kJ mol−1 (versus 40.2 kJ mol−1 without the preceding partial deprotonation) and a reaction free energy of −162.1 kJ mol−1. This process is analogous to SPLHAT (sequential proton loss HAT).6,26 The corresponding reaction of the C-terminal amido group, involving partial deprotonation-assisted HAT from 15 to produce 19, has a barrier of 38.2 kJ mol−1. As with the SPLET-type reactions discussed above, for a partially deprotonated amide in an extended peptide chain, HAT from the carbon adjacent to either the amide carbonyl or nitrogen is conceivable. The larger barrier for 15 → 19, compared with that of 9 → 13, indicates

that the HAT from the carbon adjacent to the amide nitrogen is preferred. α-Carbon-centered peptide radicals undergo a variety of subsequent reactions, including the formation of imines, the hydrolysis of which has been associated with the backbonecleavage of proteins exposed to oxidative stress4 R′C(O)NH−CR•−C(O)NHR″ → R′C(O)NCR −C(O)NHR″ → R′C(O)NH 2 + OCR−C(O)NHR″

In this regard, we note that the radical anion 12, generated as an intermediate in the deprotonation-assisted pathway from 6 to 7, is readily oxidized to 20. The ionization energy of 12 is only 318.8 kJ mol−1; therefore, only a mild oxidant (such as C

DOI: 10.1021/acs.jpca.7b11797 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

known to occur from the side chain of α-substituted amino acids, under kinetic control due to polar effects, despite the much greater thermodynamic stability of the corresponding captodative α-centered radicals.5,7−9 Therefore, we have also investigated reactions of the valine derivative MeC(O)NH− CH(i-Pr)−C(O)NHMe (23) as a model of an α-substituted amino acid peptide residue. The reaction energies and barriers for direct HAT by HO• from the α- and β-positions of 23 and for the SPLHAT-type partial deprotonation-assisted HAT reactions at those centers are displayed in Figure 4, where they are compared with the corresponding data for the glycine derivative 6. The barriers and free energies for the direct HATs from the α- and β-positions of the valine derivative 23 (23 → 24 and 23 → 25) are 35.8 and 26.8, and −147.6 and −137.2 kJ mol−1, respectively. In comparison, the corresponding values for the barrier and free energy of α-HAT from the glycine derivative 6 (6 → 7) are 40.2 and −153.6 kJ mol−1. Thus, while α-HAT from the valine 23 is somewhat less exergonic than that for the glycine 6, in a contrathermodynamic sense, the barrier is slightly smaller for the valine 23. On the basis of similar observations with closely related systems,3,12,13,27 the lower exergonicity for reaction of the valine 23 is attributable to steric interactions preventing the valyl radical 24 from adopting the conformation that would allow maximum overlap of the πorbitals of the MeC(O)NH− and −C(O)NHMe groups with the formally singly-occupied orbital at the radical center. Of the three direct HATs, abstraction from the β-position of the valine 23 to give 25 has the lowest barrier, despite being the least

molecular oxygen) is required. Hydrolysis of 20 to produce 21 plus 22 is also thermodynamically favorable, with a free energy of −26.9 kJ mol−1 (see Figure 3).

Figure 3. Oxidation of the radical anion 12 followed by hydrolysis and the associated ionization showing reaction free energies (G4(MP2)-6X + SMD aqueous solvation, ΔG, 298 K, kJ mol−1).

Valine versus Glycine. Our calculations discussed above show that partial deprotonation assists not only the oxidation of 9 to 10 and then 11 but also HAT from 9 by HO• to produce 13 and then 7 (Figure 2). Thus, they suggest not just one but two alternatives to direct HAT for peptide oxidation, which are both energetically feasible, based on both their free energies and barriers. Glycine is the only α-amino acid that does not contain a side chain (α-substituent). The reactions of the other amino acids are more complex and sometimes show unusual reactivity. For example, HAT from the α-position of derivatives of αsubstituted amino acids to give tertiary captodatively stabilized radicals is less favorable rather than more favorable when compared with reaction of the corresponding glycine derivatives to give the secondary α-carbon-centered radicals.12,13 Most relevant to this investigation, HAT by HO• is

Figure 4. Reactions of (A) the glycine derivative 6 and (B) the valine derivative 23 with HO• under neutral and basic conditions and the associated reaction free energies and barriers (in parentheses) [G4(MP2)-6X + SMD aqueous solvation, 298 K, kJ mol−1]. D

DOI: 10.1021/acs.jpca.7b11797 J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A



exergonic. This is in accord with previous experimental and theoretical findings, noted above, that HO• displays preferential HAT from the side chains of α-substituted amino acid derivatives, with the reactions being kinetically controlled and the regioselectivity being determined by polar effects and electrostatic interactions in the reaction transition structures.5,7−9,28,29 Considering now the SPLHAT-type reactions, hydrogen bonding and partial deprotonation at N−H in 23 to give the complex 26 leads to an increase in the barrier and exergonicity of subsequent HAT from the β-position to produce 28, with an exergonicity of −158.0 vs −137.3 kJ mol−1 and a barrier of +34.1 vs +26.8 kJ mol−1. By contrast, hydrogen bonding and partial deprotonation at N−H in 23 markedly facilitates αabstraction, with the barrier being lowered from 35.8 (23 → 24) to 16.1 kJ mol−1 (26 → 27). The magnitude of this reduction is very similar to that calculated for the glycine derivative 6, where formation of the complex 9 decreases the barrier from 40.2 for 6 → 7 to 19.4 kJ mol−1 for 9 → 12. The overall effect is that, whereas abstraction from the β-position of the valine 23 has the lowest barrier of the three direct HATs, the sequential partial deprotonation of 23 to give 24, followed by HAT, results in abstraction from the α-position being preferred. That is, it changes the regioselectivity of reaction with HO•.

AUTHOR INFORMATION

ORCID

Bun Chan: 0000-0002-0082-5497 Christopher J. Easton: 0000-0002-2263-1204 Leo Radom: 0000-0001-8249-1314 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We gratefully acknowledge research funding from the Japan Society for the Promotion of Science (JSPS) (Grant Number 16H07074001) and the Australian Research Council (Discovery Grant DP150101425) and generous grants of computer time from the RIKEN Advanced Center for Computing and Communication (ACCC), Japan, the Institute for Molecular Science (IMS), Japan, and the National Computational Infrastructure (NCI) National Facility, Australia.



REFERENCES

(1) Shisler, K. A.; Broderick, J. B. Glycyl Radical Activating Enzymes: Structure, Mechanism, and Substrate Interactions. Arch. Biochem. Biophys. 2014, 546, 64−71. (2) Fluehe, L.; Marahiel, M. A. Radical S-Adenosylmethionine Enzyme Catalyzed Thioether Bond Formation in Sactipeptide Biosynthesis. Curr. Opin. Chem. Biol. 2013, 17, 605−612. (3) Easton, C. J. Free-Radical Reactions in the Synthesis of α-Amino Acids and Derivatives. Chem. Rev. 1997, 97, 53−82. (4) Davies, M. J. Protein Oxidation and Peroxidation. Biochem. J. 2016, 473, 805−825. (5) Watts, Z. I.; Easton, C. J. Peculiar Stability of Amino Acids and Peptides from a Radical Perspective. J. Am. Chem. Soc. 2009, 131, 11323−11325. (6) For a comprehensive review, see: Galano, A. Free Radicals Induced Oxidative Stress at a Molecular Level: The Current Status, Challenges and Perspectives of Computational Chemistry Based Protocols. J. Mex. Chem. Soc. 2015, 59, 231−262. (7) O’Reilly, R. J.; Chan, B.; Taylor, M. S.; Ivanic, S.; Bacskay, G. B.; Easton, C. J.; Radom, L. Hydrogen Abstraction by Chlorine Atom From Amino Acids: Remarkable Influence of Polar Effects on Regioselectivity. J. Am. Chem. Soc. 2011, 133, 16553−16559. (8) Chan, B.; O’Reilly, R. J.; Easton, C. J.; Radom, L. Reactivities of Amino Acid Derivatives Toward Hydrogen Abstraction by Cl• and OH•. J. Org. Chem. 2012, 77, 9807−9812. (9) Amos, R. I. J.; Chan, B.; Easton, C. J.; Radom, L. Hydrogen-Atom Abstraction from a Model Amino Acid: Dependence on the Attacking Radical. J. Phys. Chem. B 2015, 119, 783−788. (10) Chan, B.; Karton, A.; Easton, C. J.; Radom, L. α-Hydrogen Abstraction by •OH and •SH Radicals from Amino Acids and Their Peptide Derivatives. J. Chem. Theory Comput. 2016, 12, 1606−1613. (11) Chan, B.; Moran, D.; Easton, C. J.; Radom, L. Impact of Hydrogen Bonding on the Susceptibility of Peptides to Oxidation. Chem. - Asian J. 2017, 12, 1485−1489. (12) Easton, C. J.; Hay, M. P. Preferential Reactivity of Glycine Residues in Free Radical Reactions of Amino Acid Derivatives. J. Chem. Soc., Chem. Commun. 1986, 55−57. (13) Burgess, V. A.; Easton, C. J.; Hay, M. P. Selective Reaction of Glycine Residues in Hydrogen Atom Transfer from Amino Acid Derivatives. J. Am. Chem. Soc. 1989, 111, 1047−1052. (14) Salamone, M.; Basili, F.; Bietti, M. Reactivity and Selectivity Patterns in Hydrogen Atom Transfer from Amino Acid C−H Bonds to the Cumyloxyl Radical: Polar Effects as a Rationale for the Preferential Reaction at Proline Residues. J. Org. Chem. 2015, 80, 3643−3650. (15) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,



CONCLUDING REMARKS Our computational quantum chemistry study establishes that hydrogen bonding/partial deprotonation-facilitated electron transfer, in an SPLET-type reaction, provides an alternative pathway to direct HAT for peptide oxidation and the formation of α-carbon-centered peptide radicals. In addition, we find that partial deprotonation facilitates subsequent HAT in an SPLHAT-type manner, as yet another potentially competitive process. Both alternatives are calculated to have free energies and barriers that make them energetically viable. We also find that, through deprotonation, the α-centered radicals are readily oxidized to the corresponding imines. Comparing reactions of the glycine derivative 6 with those of the valine 23, as representative of an amino acid residue with a side chain, the valyl side chain is found to be more reactive toward direct HAT, whereas partial deprotonation changes the regioselectivity, such that the SPLET- and SPLHAT-type processes favor backbone radical formation. Analogous processes are likely to be important in biology, for example, within localized environments at the active sites of enzymes involved in oxidative peptide metabolism and, more broadly, in the reactions of peptides and proteins exposed to oxidative stress. They also have potential utility in chemical synthesis and should be considered in the context of any application of peptides in materials science.



Article

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpca.7b11797. G4(MP2)-6X vibrationless energies, zero-point vibrational energies, thermal corrections for 298 K enthalpies, 298 K entropies, and M05-2X solvation energies for each species (Table S1), optimized geometries (Table S2), and free-energy values corresponding to the enthalpies in Figure 1 (Figure S1) (PDF) E

DOI: 10.1021/acs.jpca.7b11797 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A B.; Petersson, G. A.; et al. Gaussian 09, revision C.03; Gaussian, Inc.: Wallingford, CT, 2009. (16) Zhao, Y.; Schultz, N. E.; Truhlar, D. G. Design of Density Functionals by Combining the Method of Constraint Satisfaction with Parametrization for Thermochemistry, Thermochemical Kinetics, and Noncovalent Interactions. J. Chem. Theory Comput. 2006, 2, 364−382. (17) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. Universal Solvation Model Based on Solute Electron Density and a Continuum Model of the Solvent Defined by the Bulk Dielectric Constant and Atomic Surface Tensions. J. Phys. Chem. B 2009, 113, 6378−6396. (18) Chan, B.; Deng, J.; Radom, L. G4(MP2)-6X: A Cost-Effective Improvement to G4(MP2). J. Chem. Theory Comput. 2011, 7, 112− 120. (19) Merrick, J. P.; Moran, D.; Radom, L. An Evaluation of Harmonic Vibrational Frequency Scale Factors. J. Phys. Chem. A 2007, 111, 11683−11700. (20) Marcus, R. A. Chemical and Electrochemical Electron-Transfer Theory. Annu. Rev. Phys. Chem. 1964, 15, 155−196. (21) Poad, B. L. J.; Reed, N. D.; Hansen, C. S.; Trevitt, A. J.; Blanksby, S. J.; Mackay, E. G.; Sherburn, M. S.; Chan, B.; Radom, L. Preparation of an Ion with the Highest Calculated Proton Affinity: ortho-Diethynylbenzene Dianion. Chem. Sci. 2016, 7, 6245−6250. (22) Lai, W.; Li, C.; Chen, H.; Shaik, S. Hydrogen-Abstraction Reactivity Patterns from A to Y: The Valence Bond Way. Angew. Chem., Int. Ed. 2012, 51, 5556−5578. (23) Mayer, J. M. Understanding Hydrogen Atom Transfer: From Bond Strengths to Marcus Theory. Acc. Chem. Res. 2011, 44, 36−46. (24) Litwinienko, G.; Ingold, K. U. Solvent Effects on the Rates and Mechanisms of Reaction of Phenols with Free Radicals. Acc. Chem. Res. 2007, 40, 222−230. (25) Gilli, G.; Gilli, P. Towards an Unified Hydrogen-Bond Theory. J. Mol. Struct. 2000, 552, 1−15. (26) Estévez, L.; Otero, N.; Mosquera, R. A. A Computational Study on the Acidity Dependence of Radical-Scavenging Mechanisms of Anthocyanidins. J. Phys. Chem. B 2010, 114, 9706−9712. (27) Croft, A. K.; Easton, C. J.; Radom, L. Design of RadicalResistant Amino Acid Residues: A Combined Theoretical and Experimental Investigation. J. Am. Chem. Soc. 2003, 125, 4119−4124. (28) Roberts, B. P. Polarity-Reversal Catalysis of Hydrogen-Atom Abstraction Reactions: Concepts and Applications in Organic Chemistry. Chem. Soc. Rev. 1999, 28, 25−35. (29) Chan, B.; Easton, C. J.; Radom, L. Outcome-Changing Effect of Polarity Reversal in Hydrogen-Atom-Abstraction Reactions. J. Phys. Chem. A 2015, 119, 3843−3847.

F

DOI: 10.1021/acs.jpca.7b11797 J. Phys. Chem. A XXXX, XXX, XXX−XXX