Stereoinversion of chiral carbon centers via oxidation by protonated hypochlorous acid

Jon Babi, University of Toronto, Canada

Yupeng (Kevin) Liu, University of Toronto, Canada

Natalie J. Galant, University of Toronto, Canada

Anita Rágyanszki, University of Toronto, Canada
University of Miskolc, Hungary

Imre G. Csizmadia University of Toronto, Canada
University of Miskolc, Hungary

Abstract

Chiral carbon centers are present in a wide array of biological molecules, and the loss or inversion of their stereochemistry leads to abnormal structure and function. One way that carbon centers reverse their stereochemistry is through racemization reactions by oxidative or reductive species. Protonated hypochlorous acid (ClOH2 +), formed spontaneously in vivo as a consequence of reactive oxidative species, can oxidize and abstract a hydride from its chiral carbon center, which produces a planar prochiral carbocation. This planar prochiral carbocation is susceptible to hydrogenation on either of its faces, allowing for inversion of its stereochemistry.

This study investigated the thermodynamics of the racemization of carbon stereocenter analogues methane, propane, and alanine, by hydride abstraction with ClOH2 + and hydrogenation from another stereocenter. Using ab initio calculations in Gaussian09 at the B3LYP/6-31G(d) level of theory, it was determined that the reaction is exergonic and characterized by a favourable energy of activation in vacuum. Racemization completed using an additional stereocenter as a hydride donor resulted in the exergonic formation of another prochiral carbon center in addition to the racemized stereocenter. This implied that biologically active chiral carbon centers may be spontaneously oxidized by ClOH2 + to form a reactive prochiral carbon, which may react with other monomers to form aberrant racemization products. This racemization of carbon stereocenters disrupts the structure and function of biological molecules, contributing to neurodegeneration, cardiovascular disease, and several age-related diseases.

Introduction

A vast number of biological molecules, including nearly all amino acids and proteins, are stereospecific. Their specificity is crucial for normal functioning, and racemization may result in protein dysfunction that results in a host of human diseases [1]. The majority of amino acids in the human body are L-amino acids, with a small percentage of D-amino acids (D-AAs). While these D-AAs have selectively important physiological roles, their presence in high concentration can lead to disease in the central nervous system [2], vascular system [3], and musculoskeletal system [4]. Although degradative pathways for D-AAs do exist, they have not been extensively studied. If the rate of D-AA production is too high, these degradative processes may not be able to maintain normal levels of them [5]. Therefore, it is vitally important that the body maintain a low rate of D-AA formation.

In the past decade, it has become increasingly clear that natural oxidative processes important for biological functioning also contribute to an increase in D-AAs [6]. Generation of superoxide (O2 -) occurs as a by-product of many of these processes, which may be reduced to form a variety of reactive oxygen species (ROS) [7]. Superoxide spontaneously reacts with water to produce hydrogen peroxide (H2O2), which may be enzymatically converted to hypochlorous acid (HOCl) [8].

Cl^– + H₂O₂ → HOCl + HO⁺

Cellular compartments specialized for degrading biomolecules, such as lysosomes and autophagosomes, have an acidic environment where the pH drops below 5.5 [9]. This allows localized and nearby chlorine species to be protonated and form protonated hypochlorous acid (ClOH2 +) [10]. Although it is known that there are enzymatic pathways for the synthesis of D-AAs, their spontaneous formation by oxidation and racemization of L-AAs by ROS such as ClOH2 + has not been studied [5]. It is known that HOCl may oxidize certain amino acid groups, but not the backbone of a peptide [11]. This study proposes that a more reactive form, ClOH2 +, may be potent enough to oxidize the alpha carbon of an amino acid, therefore catalyzing the formation of a prochiral carbocation intermediate and its racemization.

Computational chemistry approaches can be used to evaluate the viability of oxidative racemization of amino acids. Starting with simple carbon models such as methane and propane, then more complex protein residue structures such as alanine and alanine diamide, reactions involving ClOH2 + and the proton-alpha carbon moiety (H-Cα) of the models can be studied, and the thermodynamic values associated with hydride abstraction reactions can be determined. Furthermore, these approaches can be extended to evaluate a mechanism for the propagation of this racemization reaction involving a hydride transfer between two different alpha carbons.

Investigating a mechanism for the spontaneous ROS-mediated racemization of an amino acid model can illuminate a potential process through which D-AAs are generated in the cell. Although they are traditionally thought to arise from natural biosynthetic pathways, this study suggests that there may be a considerable amount of D-AAs derived from the racemization of L-amino acids by reaction with ROS. Further work on evaluating the viability of the propagation of this racemization by hydride transfer between the planar carbocation intermediate and other Cα to amplify the total amount of racemized alpha carbons in a chain-like reaction can prove useful as well. Analysis of the thermodynamics of these reactions will aid in evaluating the viability of ROS-promoted racemization.

Materials and methods

Thermodynamic analysis of total (rovibrational + electrical) energy (ΔE_total), enthalpy (ΔH), and Gibbs free energy (ΔG) of the reactants, transition states, intermediates, and products for the reactions in Figure 1 were determined through ab initio calculations using density function theory (DFT) at the B3LYP level of theory [12, 13] with a 6-31G(d) basis set [14] with an ultrafine grid integration. While the transition states of the reactions required the building of a unique Z-matrix to ensure appropriately- defined bond and angle parameters, the reactants and products were designed using GaussView 5 software [15] and the calculations were conducted using Gaussian09 software [16].

The reactants and products were optimized at a standard 297K in vacuo, which were then used to perform a frequency calculation in the same conditions to yield the thermodynamic values ΔE, ΔH, and ΔG. These values were then converted from Hartrees to kJ/mol, where 1 Hartree = 2625.5 kJ/mol in standard conditions. Using the optimized values of the bond lengths, bond angles, and dihedral angles from the calculations conducted on the reactants, the unique Z-matrices were designed to preserve these values as the initial points of the transition states and the intrinsic reaction coordinate (IRC) calculations. These were calculated with an initial force constant calculation, which was optimized to a transition state minimum.

The reactants and products were optimized at a standard 297K in vacuo, which were then used to perform a frequency calculation in the same conditions to yield the thermodynamic values ΔE, ΔH, and ΔG. These values were then converted from Hartrees to kJ/mol, where 1 Hartree = 2625.5 kJ/mol in standard conditions. Using the optimized values of the bond lengths, bond angles, and dihedral angles from the calculations conducted on the reactants, the unique Z-matrices were designed to preserve these values as the initial points of the transition states and the intrinsic reaction coordinate (IRC) calculations. These were calculated with an initial force constant calculation, which was optimized to a transition state minimum.

Results

The calculations outlined in the previous section yielded values of ΔE_total, ΔH, and ΔG for the reactants, transition states, intermediates, and products in (Figure 1). From these values, various vital thermodynamic values such as forward and reverse activation energy (ΔG^‡_f, ΔG^‡_r) and overall change in Gibbs free energy (ΔG_reaction) were calculated. In interpreting the viability of the reactions from a thermodynamic standpoint, ΔG^‡_f and ΔG_reaction are key indicators, and are shown in Figure 2 and Figure 3.

As a general measure of the relative stability of a given chemical species, the evolution of the Gibbs free energy as a function of the reaction coordinate plays a key role in understanding the oxidation reactions. These relationships are shown in Figure 4, Figure 5, and Figure 6.

Discussion

Thermodynamics and reaction conditions
The thermodynamic values of the oxidation of H-C_α moieties by ClOH₂⁺ in this exploratory study offer insight into the viability of spontaneously generated D-AAs from oxidation of L-amino acids. The calculations for every reaction showed negative values of ΔG_reaction, indicating that the reactions occurred spontaneously. The products of the reactions were species classically known to be exceptionally stable, as most of the reactions yielded HCl and H₂O as side products, which contributed to the stability of the products compared to reactants. The reactions yielded three product molecules from two reactant molecules, which corresponded to an overall larger change in entropy and further contributed to a more negative (and therefore, more favourable) change in ΔG_reaction.
The forward activation energies (ΔG^‡_f) varied from 16.996kJ/ mol to 50.759kJ/mol. There are multiple reasons for the variation in ΔG^‡_f between similar species, such as the oxidation of methane compared to propane. These differences are largely due to electronic effects, which refer to the ability of the structure to stabilize theorized transition states and final products. In the transition state, the H-C_α moiety is weakened and a positive charge is formed on the C_α, which develops into a formal positive charge in the final product. The two -CH₃ groups present in propane are excellent cationic charge stabilizers when compared to -H groups, due to the multiple σ-bonds present in the group. These bonds can stabilize carbocation charge by hyperconjugation, allowing for the dispersion of the charge over their own surface.
When comparing alanine and alanine diamide, potential steric effects can come into play on top of the positive contribution from the electron donating groups’ electronic effects, as the large amide groups on the molecule represent major steric bulk. Given that residues exist in vivo as parts of much larger peptides, steric bulk will be an increasingly important consideration for future study. If these values of ΔG^‡_f can be supplied by in vivo systems, oxidative molecules including ROS such as ClOH₂⁺ may abstract a hydride from the C_α of an amino acid, break the H-C_α moiety, and form a prochiral carbocation intermediate. This planar prochiral carbon intermediate may be protonated on either side to form the L- or D-AA, representing an important mechanism for the generation of D-AA using ROS produced during cellular stress may to mediate the stereoinversion.
Given the exploratory nature of the study, the in silico conditions were not selected to match in vivo conditions due to computational limitations associated with simulations that mimic solvation in a variety of media. The differences in the in vacuo conditions used when compared to the variety of solvents, free ions, cellular interactions (including the ionic strength in the cell), and metal cofactors available in biological systems imply that the results of the study may be widely affected by these conditions [17]. Metal cofactors and high ionic stress can contribute to an increased possibility of potentially stabilizing electronic interactions between charged ions and transition states or final carbocationic products. Exploring these effects, along with the potential steric limitations presented by highly complex and well-organized protein structures that could limit the access of ClOH₂⁺ to C_α sites, will prove vital in evaluating the potential viability of the reaction in vivo.
Biological implications
D-AAs are elevated in a variety of pathological conditions, most notably in the central nervous system [2], vascular system [3], and musculoskeletal disease [4]. In the human aorta, D-aspartate accumulates in elastin, but not in other structural proteins, due to the spontaneous racemization of L-aspartate [3]. Collagen and other peptides do not display this D-isomer accumulation due to their rapid turnover [3]. In CNS disorders, there may be aggregations of peptides that are not recycled or degraded similar to elastin in the aorta. The b-amyloid (Ab) peptide affiliated with Alzheimer disease forms toxic insoluble aggregates in the brain when b-amyloid degrading proteases are disrupted [18].
Not only are these insoluble peptides more prone to accumulating D-AAs due to their low turnover [3], the elevated D-serine also overstimulates excitatory neural NMDA receptors, leading to cellular exhaustion and death [19]. Racemized forms of amino acids also accumulate in the musculoskeletal system and their elevated concentration is an indicator of disease [4]. Not only are D-AAs relevant in disease, but their concentration may also be used to determine the age of a protein and tissue [4]. In the aorta, the first-order rate constant of this racemization was calculated to be 1.14(x10)3 [3], but this constant may vary in different environments. In all, this evidence suggests that there is a spontaneous mechanism for endogenous amino acid racemization which may cause build-up of D-AAs in proteins not rapidly being turned over and ultimately contribute to disease.
Looking forward
Our findings are an initial evaluation on the possibility of amino acid racemization by ROS, but further study on this mechanism may provide a more accurate understanding. The next step is to re-evaluate the energetics of this reaction using a more rigorous computational theory, since B3LYP/DFT may not yield accurate calculations for long-distance, non-covalent van der Waals interactions in the transition state. In addition, this reaction should be verified under in vivo conditions and with the consideration of larger protein structures, including possible metal cofactors, to ensure that the reaction is favourable. However, we predict that the increased ionic strength of the environment will accelerate the reaction. In vivo research should also focus on quantifying, with a high degree of accuracy, the generation of ClOH₂⁺ and its lifespan in tissues.

Conclusion

All racemization reactions between ClOH₂⁺ and alpha carbon models were found to be spontaneous and exergonic in vacuum at standard conditions. The planar carbocation intermediate propagated this racemization by a hydride transfer mechanism. This suggests that there may be a similar mechanism leading to spontaneous generation of D-AAs from L-amino acids in vivo, which is a topic for further research. This racemization of amino acids in peptides leads to disruption of protein structure, enzymatic function, and metabolic pathways when these peptides are not degraded. Interestingly, an increase in D-AA concentration is also implicated in many diseases, and the potential application of computational chemistry in evaluating the viability of reactions in given conditions is a topic with much promise.

Acknowledgements

We would like to acknowledge our appreciation to the Department of Chemistry and the Research Opportunity Program at the University of Toronto for the opportunity to conduct this study. We would like to thank Professor Csizmadia for his guidance, Natalie Galant for her overarching support of the entire group, and Anita Rágyanszki for her consistent support and assistance in times of great stress. We would also like to thank Nidaa Rasheed and Fiser Béla for their support, guidance, and technical assistance throughout the study, and every member of the Csizmadia group for fostering an environment that encouraged exploration and discovery across two continents.

References

1. Friedman M. Formation, nutritional value, and safety of D-amino acids. Adv Exp Med Biol.1991;289:447-81.
2. Chervyakov A, Gulyaeva N, Zakharova M. D-amino acids in normal ageing and pathogenesisof neurodegenerative diseases. Neurochem J. 2011;5(2):100-14.
3. Powell J, Vine N, Crossma M. On the accumulation of D-aspartate in elastin and other proteinsof the ageing aorta. Atherosclerosis. 1992;97(3):201-8.
4. McCudden C, Kraus V. Biochemistry of amino acid racemization and clinical application tomusculoskeletal disease. Clin Biochem. 2006;39(12):1112-30.
5. Ohide H, Miyoshi Y, Maruyama R, Hamase K, Konno R. D-Amino acid metabolism in mammals:biosynthesis degradation and analytical aspects of the metabolic study. J Chromatogr BAnalyt Technol Biomed Life Sci. 2011;879(29):3162-68.
6. Kuge K, Kitamura K, Nakaoji K, Hamada K, Fujii N, Saito T, et al. Oxidative stress inducesthe formation of D-aspartyl residues in the elastin mimic peptides. Chem Biodivers.2010;7(6):1408-12.
7. Scheiber M, Chandel N. ROS function in redox signaling and oxidative stress. Curr Biol.2014;24(10):R453062.
8. Parker H, Dragunow M, Hampton M, Kettle A, Winterboum CC. Requirements for NADPHoxidase and myeloperoxidase in neutrophil extracellular trap formation differ depending onthe stimulus. J Leukoc Biol. 2012;92(4):841-49.
9. Mindell, J. Lysosomal acidification mechanisms. Annu Rev Physiol. 2012;74:69-86.
10. Fengel D, Wegener G. Wood: chemistry, ultrastructure, reactions. Walter de Gruyler & Co.1983;1:284-285.
11. Pullar J, Vissers M, Winterbourn C. Critical review living with a killer: the effects of hypochlorousacid on mammalian cells. IUBMB Life. 2000;50:259-66.
12. Becke, AD. Density-functional thermochemistry. III. The role of exact exchange. J ChemPhys. 1993;98:5648-5652.
13. Lee C, Yang W, Parr RG. Development of the Colle-Salvetti correlation energy formula intoa functional of the electron density. Phys Rev. 1988;37:785-789.
14. Ditchfield R, Hehre WJ, Pople JA. Self-consistent molecular-orbital methods. IX. An extendedGaussian-type bases for molecular-orbital studies of organic molecules. J Chem Phys.1971;54:724-28.
15. Dennington R, Keith T, Millam J. GaussView, Version 5. Semichem Inc., Shawnee MissionKS. 2009.
16. Frisch M, et al. Gaussian 09 Revision E.01. DJ Gaussian Inc., Wallingford CT. 2009.
17. Haraldsdottir HS, Thiele I, Fleming RMT. Quantitative assignment of reaction directionalityin a multicompartmental human metabolic reconstruction. Biophysical J. 2012;1:1705.
18. Moro M, Collins M, Cappellini E. Alzheimer’s disease and amyloid beta-peptide depositionin the brain: matter of “aging”? Biochem Soc Trans. 2010;28(2):539-44.
19. Shleper M, Kartvelishvity E, Wolosker H. D-serine is the dominant endogenous coagonistfor NMDA receptor neurotoxicity in organotypic hippocampal slices. J Neurosci.2005;25(41):9413-17.Este nihil ius dis reiunt.