Multiscale simulation of monoamine oxidase catalyzed decomposition of phenylethylamine analogs

Author’s Accepted Manuscript Multiscale simulation of monoamine oxidase catalyzed decomposition of phenylethylamine analogs Gabriel Oanca, Jernej Stare, Robert Vianello, Janez Mavri www.elsevier.com/locate/ejphar

PII: DOI: Reference:

S0014-2999(17)30392-8 http://dx.doi.org/10.1016/j.ejphar.2017.05.061 EJP71251

To appear in: European Journal of Pharmacology Received date: 21 October 2016 Revised date: 31 March 2017 Accepted date: 31 May 2017 Cite this article as: Gabriel Oanca, Jernej Stare, Robert Vianello and Janez Mavri, Multiscale simulation of monoamine oxidase catalyzed decomposition of phenylethylamine analogs, European Journal of Pharmacology, http://dx.doi.org/10.1016/j.ejphar.2017.05.061 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Multiscale simulation of monoamine oxidase catalyzed decomposition of phenylethylamine analogs Gabriel Oanca,a,b Jernej Stare,a Robert Vianello,c and Janez Mavria* a

Laboratory of Biocomputing and Bioinformatics, National Institute of Chemistry, Ljubljana, Slovenia. Faculty of Physics, Alexandru Ioan Cuza University of Iasi, Iasi, Romania. c Computational Organic Chemistry and Biochemistry Group, Ruđer Bošković Institute, Zagreb, Croatia. * Corresponding author E-mail: [email protected] b

Abstract Phenylethylamine (PEA) is an endogenous amphetamine and its levels are increased by physical activity. As with other biogenic monoamines, it is decomposed by monoamine oxidase (MAO) enzymes. The chemical mechanism of MAO, and flavoenzymes in general, is a subject of heated debate. We have previously shown that the rate-limiting step of MAO catalysis involves a hydride transfer from the substrate methylene group vicinal to the amino group to the N5 atom of the lumiflavin co-factor moiety. By using multiscale simulation on the Empirical Valence Bond (EVB) level, we studied the chemical reactivity of the monoamine oxidase B catalyzed decomposition of PEA and its two derivatives: p-chloro--methylphenylamine (p-CMP) and p-methoxy-methylphenethylamine (p-MMP). We calculated activation free energies of 17.1 kcal/mol (PEA), 18.4 kcal/mol (p-MMP) and 20.0 kcal/mol (p-CMP), which are in excellent agreement with the experimental values of 16.7 kcal/mol for PEA and 18.3 kcal/mol for p-MMP, while the experimental value for p-CMP is not available. This gives strong support to the validity of our hydride transfer mechanism for both MAO A and B isoforms. The results are discussed in the context of the interplay between MAO point mutations and neuropsychiatric disorders. Keywords monoamine oxidase; molecular simulation; QM/MM methodology; Empirical Valence Bond; neurotransmitters; phenylethylamine

1. Introduction Monoamine oxidases (MAOs) are enzymes found on the mitochondrial membrane of cells (Edmondson et al., 2009; Youdim et al., 2006). The reaction they catalyze is the oxidative deamination of biogenic and dietary monoamines (Abad et al., 2013; Poberžnik et al., in print; Repič et al., 2014b; Vianello et al., 2012; Zapata-Torres et al., 2015) such as dopamine, serotonin, histamine, noradrenaline and phenylethylamine. MAOs exist in two isoforms, MAO A which mainly decomposes serotonin and MAO B which predominantly metabolizes dopamine and phenylethylamine (Shih et al., 1999). The human A and B isoforms are quite similar, sharing 70% of amino-acid sequence and the same lumiflavin cofactor (Bach et al., 1988). By calculating the pKa values of ionizable residues and comparing the corresponding active site geometries, we have shown that MAO A and MAO B have very similar preorganized electrostatics and that it is very likely that both isoforms catalyze reactions by the same chemical mechanism (Repič et al., 2014a), although this has been questioned in the literature (Orru et al., 2013).

The general reaction of flavin amine oxidases, including MAO, can be divided into two half-reactions. In the reductive half-reaction, the equivalent of two hydrogen atoms are transferred from the substrate to the flavin, thus reducing it to FADH2, while the oxidative half-reaction involves the oxidation of the reduced flavin back to FAD by molecular oxygen, producing hydrogen peroxide. The reductive half-reaction represents the rate-limiting step in both MAO A and B isoforms, although the weakly oxygen concentration dependent rate constant for the oxidative half-reaction in MAO B indicates that the barrier associated with this part of the catalytic turnover is not significantly lower in MAO B (Reid McDonald et al., 2010). Oxidation of an amine substrate necessarily involves the removal of two protons and two electrons as the carbon–nitrogen single bond is converted to a double bond. Taking into consideration the three-dimensional structures of the two MAO isoforms (Binda et al., 2002; De Colibus et al., 2005; Son et al., 2008) in conjunction with the spectroscopic and kinetic data, three mechanisms for the catalytic reaction have been proposed: (a) a direct hydride transfer mechanism, (b) a radical transfer mechanism, and (c) a polar nucleophilic mechanism. Studies on deuterated substrate analogues have demonstrated that the rate-limiting step is the cleavage of a carbon-hydrogen bond vicinal to the amino group (Klinman and Matthews, 1985) and, hence, the catalytic proposals differ in the nature of the hydrogen being transferred, namely a hydride (H–) in (a), a hydrogen atom (H•) in (b), and a proton (H+) in (c) (Vianello et al., 2016; Wang and Edmondson, 2011), commonly by the flavin N5 atom. We have clearly demonstrated that the rate limiting step involves a hydride transfer from the methylene group vicinal to the amino group that is picked up by the N5 atom on the flavin adenine dinucleotide (FAD) cofactor (Vianello et al., 2012). Other new computational studies (Abad et al., 2013; Akyuz and Erdem, 2013; Atalay and Erdem, 2013; Zapata-Torres et al., 2015; Zenn et al., 2015) and in vitro experiments (Ralph et al., 2007) support the proposed hydride transfer mechanism. Phenylethylamine (PEA) is more a neuromodulator than an excitatory neurotransmitter and increased levels are associated with schizophrenia (Wolf and Mosnaim, 1983). PEA is synthesized from L-phenylalanine by the aromatic amino acid decarboxylase (Berry, 2007; Broadley, 2010), although it is rapidly metabolized. In neurons, it is estimated that PEA has a half-life of about 30 seconds (Berry, 2004; Broadley, 2010). It remains a challenge to calculate the rate constant for PEA decomposition catalyzed by MAO B using a multiscale molecular simulation. In addition to calculating the activation free energy that is analytically connected with the reaction rate, multiscale simulation yields insight into the electronic effects of the enzyme environment (e.g. point mutations, protonation states) on the rate constant. Substituted PEA have different reactivities and analysis of their MAO B catalyzed decomposition was studied experimentally in order to gain insight into the mechanism by using the methods of physical organic chemistry (Li et al., 2006). In a very recent study Fierro and co-workers computationally studied a series of p-substituted phenylethylamines by using a cluster model of the MAO enzyme at the quantum mechanical level (Fierro et al., 2016). They compared the activation energies with the available experimental values and, overall, a good agreement was found. The focus of this article is to examine the chemical reactivity of three phenylethylamine derivatives using a multiscale QM/MM method on the Empirical Valence Bond level, by considering the full dimensionality of the enzyme with extensive thermal averaging. The obtained results will provide further support in favor of the hydride transfer mechanism being the same in both A and B MAO isoforms.

2. Computational Methods 2.1. Quantum chemical calculations The geometry of all three substrates (PEA, p-MMP and p-CMP) was initially optimized by employing quantum chemical calculations in Gaussian09 (Frisch et al., 2009) using the HF/6-31G(d) level of theory. Structures of all three phenylethylamine derivatives and the lumiflavin co-factor moiety are shown in Scheme 1. Optimized geometries were used to calculate the initial set of atomic charges using the MerzKollman scheme at the same level of theory.

PEA

p-MMP

p-CMP

FAD

Scheme 1 Structures of phenylethylamine (PEA) and its two derivatives: p-chloro--methylphenylamine (p-CMP) and p-methoxy--methylphenethylamine (p-MMP), together with the lumiflavin moiety of the FAD co-factor.

A flavin molecule was added to each substrate and the corresponding systems were subjected to the transition state search in accordance with our hydride transfer mechanism (Vianello et al., 2012). Once we located the matching TS structures, the geometries of the reactants and the products were found by applying a scan along the reaction coordinate corresponding to the imaginary eigenvector using the IRC procedure. In this way, the activation and reaction energies were obtained. All calculations were performed at the M062X/6–31+G(d,p) level of theory. In the IRC procedure the Hessian was recomputed every ten steps. For practical reasons associated with the buildup of the classical model, the reference reaction free energy was taken as the difference between the energy at the reactants’ point and the transient intermediate point on the products’ reaction coordinate path, as we have done before. (Oanca et al., 2016; Poberžnik et al., in print, Repič et al., 2014b). The energy profiles obtained in this way were used for the calibration of the subsequent QM/MM EVB simulations.

2.2. EVB Parameterization of the Reference Reaction in the Gas Phase Program package Q version 5 was used for all EVB calculations (Marelius et al., 1998). Considering the lack of solution-phase experimental data, we parameterized our EVB Hamiltonian to reproduce reliable M062X/6-31+G(d,p) energetics in vacuo (∆G‡gas and ∆Ggas). Thus, any gas phase barrier and reaction free energy will be the same, regardless of the simulation run, all the differences being reflected in the EVB coupling term and gas shift parameters, H12 and a22. The substrate was subsequently

moved to the MAO B active site using the same parameter set, which is a valid approximation due to the demonstrated phase-independence of the EVB off-diagonal (Hij) coupling term (Hong et al., 2006). This approach allows us to explore the effect of changing the intrinsic gas phase environment to a more complex environment such as the MAO B active site. Application of a gas phase reference reaction is an established approach (Mo and Gao, 2000). In order to calculate the gas phase profile with Q, we followed the same FEP simulation protocol as in the enzyme (see section 2.3). Since the equilibration in the gas phase is faster than in the enzyme, we ran the equilibration for 100 ps at 300 K. Special attention was given to the flavin-substrate orientation, which was restrained to the reference positions corresponding to the geometry from the enzyme by applying the harmonic potential to all non-hydrogen atoms with a force constant of 0.5 kcal/mol/Å 2. The applied values for the coupling term H12 and the gas shift a22 are shown in Table 1. Table 1 Empirical Valence Bond Parameters for phenylethylamine (PEA) and its two derivatives: pchloro--methylphenylamine (p-CMP) and p-methoxy--methylphenethylamine (p-MMP). The values for the coupling term H12 and the gas shift a22 are given in kcal/mol. Please note that the parameters reproduce the gas phase barriers and energies of reactions.

substrate

PEA p-MMP p-CMP

H12 84.36 65.20 72.12

a22 109.05 109.25 115.15

2.3. EVB Simulation of the Reaction in Enzyme The optimized structures of the substrates were manually docked into the enzyme active site using UCSF Chimera software (Pettersen et al., 2004). The initial structures, along with the topologies, were generated by the qprep program from the Q package. The systems were hydrated, each of them with about 1800 water molecules (1829, 1795 and 1784 water molecules for PEA, p-MMP and p-CMP, respectively). The entire system was encapsulated inside a sphere with a radius of 30 Å centered at the reactive N5 atom of the flavin moiety. A few enzyme atoms outside the sphere were restrained to their original positions with a harmonic force constant of 200 kcal/mol/Å2. The flavin moiety and the substrate molecules were treated as reactive atoms within the framework of the EVB approach. The applied protocol and the van der Waals parameters were the same as in our previous work concerning PEA and MAO A point mutations (Oanca et al., 2016).

Molecular dynamics simulation runs were performed using the qdyn program from the Q package in conjunction with the OPLS–AA force field. Before simulating the reaction event, each system was carefully equilibrated. We started at a low temperature of 1 K applying position restraints for the entire system with a harmonic force constant of 200 kcal/mol/Å2 for 5 ps. Then we gradually increased the temperature and released the position restraints by performing eight additional consecutive simulations of 10–30 ps in length. The force constants for the position restraints were decreased to 0.5 kcal/mol/Å2 and the temperature was increased to 300 K. At the end we performed a final equilibration of 1 ns, at a temperature of 300 K with the position restraints imposed only on all non-hydrogen atoms in the EVB region, with a force constant of 0.5 kcal/mol/Å2. In the final equilibration step we also imposed a distance restraint with a harmonic force constant of 10 kcal/mol/Å2 on the equilibrium distance of 3 Å between the –carbon of the PEA moiety and the N5 atom of the flavin, and a harmonic restraint of 10 kcal/mol/Å2 on the equilibrium distance of 2 Å between the reactive substrate –hydrogen atom and the flavin N5 atom. In this way, we obtained the equilibrated structure of the enzyme-substrate complex with the reactive conformation for the enzymatic reaction. The relaxed system was mutated from a reactant diabatic state to a product diabatic state in 51 mapping frames, each consisting of a molecular dynamics simulation of 10 ps, using a 1 fs time step, for a total simulation time of 510 ps. The EVB simulation of the reactive trajectory was performed 10 times, starting from different initial configurations, giving rise to ten independent free energy profiles. The differences in calculated activation free energies represent a measure of convergence. The abovementioned starting configurations were provided by running 10 short simulations of 50 ps, starting from the equilibrated procedure described above in this section, the only difference being the random number seed for the initial velocities. A snapshot from the equilibrated structure of MAO B with the substrate PEA is shown in Figure 2.

All three substrates react only in the neutral (unionized) form, which is consistent with several experimental results that convincingly showed that the neutral form is required for the reaction (Dunn et al., 2008; Jones et al., 2007; Tan and Ramsay, 1993). This agrees well with our previous results where we have shown that the reaction with monocationic substrates has a very high barrier (Vianello et al., 2012). Nevertheless, under physiological conditions (pH = 7.4) all three substrates are protonated monocations. Therefore, it is essential to estimate and include in the calculations the free energy required to deprotonate the amino group to attain the reactive systems. We assumed that in the MAO B active site the substrate pKa value does not change significantly relative to the experimental value in aqueous solution. This assumption proved to be valid for both MAO A and MAO B active sites (Borštnar et al., 2012; Repič et al., 2014a). In order to calculate the free energy for the abovementioned substrate deprotonation we used the following formula: (

)

(

)

We assumed a physiological pH value of 7.4 and used the measured experimental pKa values for PEA of 9.64 (Jones et al., 2007). An experimental value for

p-MMP was not available so we took the value for its -methylated p-MMP analog with a pKa value of 9.53 (Tucker et al., 1993). For p-CMP, the experimental pKa value was also not available and we took the same experimental value as for p-MMP. The obtained free energies of deprotonation were added to the calculated free energies of activation in order to have the appropriate comparison with the experimental values. The obtained corrections for deprotonation are 3.08 kcal/mol for PEA and 2.93 kcal /mol for both p-MMP and p-CMP. Please note that the calculated values for the activation free energy for the enzyme reaction shown in Table 2 are corrected for the free energies of deprotonation, while, in the gas phase, the amino group is neutral and the deprotonation correction is not necessary. The experimental G‡ values were taken from (Fierro et al., 2016; Li et al., 2006). 3. Results and Discussion The calculated activation energies and the reaction free energies are collected in Table 2, while the corresponding reaction profiles are shown in Figure 3, taking p-CMP as an illustrative example. Experimental values for the activation free energies were calculated from the experimental rate constants by assuming the validity of the transition state theory. The activation energy relates to the rate constant by the following equation:

where kB is the Boltzmann constant, h is the Planck constant, T is the temperature in K, and G‡ represents the activation energy. The gas phase reactions for all three substrates are highly endergonic with activation barriers exceeding 35 kcal/mol. The results show that in the gas phase the reaction equilibrium will be strongly shifted in the direction of the reactants as all three reaction free energies are highly endergonic and cluster around 27 kcal/mol. This high reaction free energy in the gas phase is due to the transient intermediate point considered in place of the product state (see 2.1), ultimately leading to a covalently bound substrate on the flavin, as is the case in the gas phase DFT calculations. Inclusion of the enzyme environment significantly lowers the barriers and makes the hydride abstraction thermodynamically feasible. A comparison between the experimental and calculated activation energies for the PEA and p-MMP shows a small difference of 0.4 kcal/mol for PEA, which is within the standard deviation, while almost perfect agreement is reached for p-MMP. This gives strong support in favor of our hydride transfer mechanism and suggests that, in the absence of experimental results, the value calculated for p-CMP is accurate as well. On the other hand, this may indicate that the calculated value reported by Fierro and co-workers of ΔG‡ = 35.0 kcal/mol for p-CMP (Fierro et al., 2016) is very likely overestimated. The standard deviations of the calculated activation free energies are between 0.7 and 1 kcal/mol, which serve as a measure of the uncertainty.

Table 2 EVB QM/MM activation free energies (ΔG‡) and reaction free energies (ΔG°) for phenylethylamine (PEA) and its two derivatives: p-chloro--methylphenylamine (p-CMP) and pmethoxy--methylphenethylamine (p-MMP).

substrate G‡exp PEA 16.7 p-MMP 18.3 p-CMP NA

G‡enz 17.1 ± 0.9 18.4 ± 1.0 20.0 ± 0.7

G0enz 1.3 ± 1.4 -8.3 ± 1.7 -1.1 ± 0.6

G‡gas 36.6 ± 0.7 35.5 ± 0.4 36.0 ± 0.4

G0gas 27.7 ± 0.9 26.9 ± 0.6 27.6 ± 0.7

The obtained results show that both p-MMP and p-CMP are worse MAO B substrates than the unsubstituted PEA, as they are associated with larger activation free energies. The highest ΔG‡ value of 20.0 kcal/mol was obtained for p-CMP, in line with the experiments by Kinemuchi et al. (Kinemuchi et al., 1988) and Kim et al. (Kim et al., 1991), who both demonstrated that p-CMP was a highly selective MAO B inhibitor. Interestingly, when its p-Cl group is replaced by the methoxy group, as in p-MMP, the barrier drops down to 18.4 kcal/mol indicating a change from the reversible inhibitor to the MAO B substrate, yet the barrier is still higher than for PEA. This seems to indicate that the steric and electronic features of the investigated para-substituents in p-MMP and p-CMP create unfavorable interactions within the active site, leading to increased barriers. This aspect even outperforms a likely favorable effect of the electron-donating β-methyl groups, which are expected to stabilize the formed carbocationic transition state during the hydride abstraction reaction. Overall, it follows that the nature of the para-substituent determines the behavior of these compounds with MAO B, which is an important insight for the development of novel and more efficient MAO B inhibitors to be clinically employed to treat neurodegeneration.

4. Conclusions We calculated the activation free energies for phenylethylamine and two of its derivatives, and very good agreement with the experimental data was found. Obviously, the combined presence of the β-methyl group and para-substituents of similar volume with different electronic properties on the aromatic ring determines the reactivity of these compounds. When calculating reaction profiles, we performed intensive conformational sampling and we feel that our free energy calculations are well converged. While experimental data for a larger set of compounds would most likely enrich this discussion, a good agreement with the experimental activation free energies is an additional proof of the validity of the hydride transfer mechanism. Future activities will be directed towards understanding the relationships between MAO point mutations and biogenic amines levels, which is of prime importance for genomic medicine. Acknowledgements This work benefited from the Erasmus program scholarship (G. O.). R.V. gratefully acknowledges the European Commission for an individual FP7 Marie Curie Career Integration Grant (contract number PCIG12–GA–2012–334493). J.S. and J.M. would like to thank the Slovenian Research Agency for financial support within the framework of the Program Group P1-0012. Part of this work was supported by the COST Action CM1103. We would like to thank Ms. Charlotte C. W. Taft for careful reading of the manuscript and linguistic corrections.

References Abad, E., Zenn, R.K., Kastner, J., 2013. Reaction Mechanism of Monoamine Oxidase from QM/MM Calculations. Journal of Physical Chemistry B 117, 1423814246. doi: http://dx.doi.org/10.1021/jp4061522 Akyuz, M.A., Erdem, S.S., 2013. Computational modeling of the direct hydride transfer mechanism for the MAO catalyzed oxidation of phenethylamine and benzylamine: ONIOM (QM/QM) calculations. Journal of Neural Transmission 120, 937-945. doi: http://dx.doi.org/10.1007/s00702-0131027-8 Atalay, V.E., Erdem, S.S., 2013. A comparative computational investigation on the proton and hydride transfer mechanisms of monoamine oxidase using model molecules. Computational Biology and Chemistry 47, 181-191. doi: http://dx.doi.org/10.1016/j.compbiolchem.2013.08.007 Bach, A.W.J., Lan, N.C., Johnson, D.L., Abell, C.W., Bembenek, M.E., Kwan, S.W., Seeburg, P.H., Shih, J.C., 1988. Cdna Cloning of Human-Liver Monoamine Oxidase-a and Oxidase-B - Molecular-Basis of Differences in EnzymaticProperties. Proceedings of the National Academy of Sciences of the United States of America 85, 4934-4938. doi: http://dx.doi.org/10.1073/pnas.85.13.4934 Berry, M., 2007. The Potential of Trace Amines and Their Receptors for Treating Neurological and Psychiatric Diseases. Reviews on Recent Clinical Trials 2, 3-19. doi: http://dx.doi.org/10.2174/157488707779318107 Berry,

M.D., 2004. Mammalian central nervous system trace amines. Pharmacologic amphetamines, physiologic neuromodulators. Journal of Neurochemistry 90, 257-271. doi: http://dx.doi.org/10.1111/j.14714159.2004.02501.x

Binda, C., Newton-Vinson, P., Hubalek, F., Edmondson, D.E., Mattevi, A., 2002. Structure of human monoamine oxidase B, a drug target for the treatment of neurological disorders. Nature Structural Biology 9, 22-26. doi: http://dx.doi.org/10.1038/nsb732 Borštnar, R., Repič, M., Kamerlin, S.C.L., Vianello, R., Mavri, J., 2012. Computational Study of the pK(a) Values of Potential Catalytic Residues in the Active Site of Monoamine Oxidase B. Journal of Chemical Theory and Computation 8, 3864-3870. doi: http://dx.doi.org/10.1021/ct300119u Broadley, K.J., 2010. The vascular effects of trace amines and amphetamines. Pharmacology & Therapeutics 125, 363-375. doi: http://dx.doi.org/10.1016/j.pharmthera.2009.11.005 De Colibus, L., Li, M., Binda, C., Lustig, A., Edmondson, D.E., Mattevi, A., 2005. Threedimensional structure of human monoamine oxidase A (MAO A): Relation to

the structures of rat MAO A and human MAO B. Proceedings of the National Academy of Sciences of the United States of America 102, 12684-12689. doi: http://dx.doi.org/10.1073/pnas.0505975102 Dunn, R.V., Marshall, K.R., Munro, A.W., Scrutton, N.S., 2008. The pH dependence of kinetic isotope effects in monoamine oxidase A indicates stabilization of the neutral amine in the enzyme-substrate complex. Febs J 275, 3850-3858. doi: http://dx.doi.org/10.1111/j.1742-4658.2008.06532.x Edmondson, D.E., Binda, C., Wang, J., Upadhyay, A.K., Mattevi, A., 2009. Molecular and Mechanistic Properties of the Membrane-Bound Mitochondrial Monoamine Oxidases. Biochemistry 48, 4220-4230. doi: http://dx.doi.org/10.1021/bi900413g Fierro, A., Edmondson, D.E., Celis-Barros, C., Rebolledo-Fuentes, M., Zapata-Torres, G., 2016. Why p-OMe- and p-Cl-beta-Methylphenethylamines Display Distinct Activities uponMAO-B Binding. Plos One 11. doi: http://dx.doi.org/10.1371/journal.pone.0154989 Gaussian 09. Revision A.02, 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., Nakatsuji, H., Caricato, M., Li, X., Hratchian, H.P., Izmaylov, A.F., Bloino, J., Zheng, G., Sonnenberg, J.L., Hada, M., Ehara, M., Toyota, K., Fukuda, R., Hasegawa, J., Ishida, M., Nakajima, T., Honda, Y., Kitao, O., Nakai, H., Vreven, T., Montgomery Jr., J.A., Peralta, J.E., Ogliaro, F., Bearpark, M.J., Heyd, J., Brothers, E.N., Kudin, K.N., Staroverov, V.N., Kobayashi, R., Normand, J., Raghavachari, K., Rendell, A.P., Burant, J.C., Iyengar, S.S., Tomasi, J., Cossi, M., Rega, N., Millam, N.J., Klene, M., Knox, J.E., Cross, J.B., Bakken, V., Adamo, C., Jaramillo, J., Gomperts, R., Stratmann, R.E., Yazyev, O., Austin, A.J., Cammi, R., Pomelli, C., Ochterski, J.W., Martin, R.L., Morokuma, K., Zakrzewski, V.G., Voth, G.A., Salvador, P., Dannenberg, J.J., Dapprich, S., Daniels, A.D., Farkas, Ö., Foresman, J.B., Ortiz, J.V., Cioslowski, J., Fox, D.J., Gaussian, Inc., Wallingford, CT, USA, 2009. Hong, G.Y., Rosta, E., Warshel, A., 2006. Using the constrained DFT approach in generating diabatic surfaces and off diagonal empirical valence bond terms for modeling reactions in condensed phases. Journal of Physical Chemistry B 110, 19570-19574. doi: http://dx.doi.org/10.1021/jp0625199 Jones, T.Z.E., Balsa, D., Unzeta, M., Ramsay, R.R., 2007. Variations in activity and inhibition with pH: the protonated amine is the substrate for monoamine oxidase, but uncharged inhibitors bind better. Journal of Neural Transmission 114, 707-712. doi: http://dx.doi.org/10.1007/s00702-0070675-y Kim, S.K., Toyoshima, Y., Arai, Y., Kinemuchi, H., Tadano, T., Oyama, K., Satoh, N., Kisara, K., 1991. Inhibition of Monoamine-Oxidase by 2 Substrate-Analogs, with Different Preferences for 5-Hydroxytryptamine Neurons.

Neuropharmacology 30, 329-335. doi: http://dx.doi.org/10.1016/00283908(91)90057-I Kinemuchi, H., Arai, Y., Toyoshima, Y., Tadano, T., Kisara, K., 1988. Studies on 5Fluoro-Alpha-Methyltryptamine and Para-Chloro-BetaMethylphenethylamine - Determination of the Mao-a or Mao-B SelectiveInhibition Invitro. Jpn J Pharmacol 46, 197-199. doi: http://dx.doi.org/10.1254/jjp.46.197 Klinman, J.P., Matthews, R.G., 1985. Calculation of Substrate Dissociation-Constants from Steady-State Isotope Effects in Enzyme-Catalyzed Reactions. Journal of the American Chemical Society 107, 1058-1060. doi: http://dx.doi.org/10.1021/ja00290a052 Li, M., Binda, C., Mattevi, A., Edmondson, D.E., 2006. Functional role of the "aromatic cage" in human monoamine oxidase B: Structures and catalytic properties of Tyr435 mutant proteins. Biochemistry-Us 45, 4775-4784. doi: http://dx.doi.org/10.1021/bi051847g Marelius, J., Kolmodin, K., Feierberg, I., Aqvist, J., 1998. Q: A molecular dynamics program for free energy calculations and empirical valence bond simulations in biomolecular systems. Journal of Molecular Graphics & Modelling 16, 213-225. doi: http://dx.doi.org/10.1016/S10933263(99)00012-1 Mo, Y.R., Gao, J.L., 2000. Ab initio QM/MM simulations with a molecular orbitalvalence bond (MOVB) method: Application to an S(N)2 reaction in water. Journal of Computational Chemistry 21, 1458-1469. doi: http://dx.doi.org/10.1002/1096-987X(200012)21:16%3C1458::AIDJCC4%3E3.0.CO;2-2 Oanca, G., Purg, M., Mavri, J., Shih, J.C., Stare, J., 2016. Insights into enzyme point mutation effect by molecular simulation: phenylethylamine oxidation catalyzed by monoamine oxidase A. Physical Chemistry Chemical Physics 18, 13346-13356. doi: http://dx.doi.org/10.1039/c6cp00098c Orru, R., Aldeco, M., Edmondson, D.E., 2013. Do MAO A and MAO B utilize the same mechanism for the C-H bond cleavage step in catalysis? Evidence suggesting differing mechanisms. Journal of Neural Transmission 120, 847-851. doi: http://dx.doi.org/10.1007/s00702-013-0991-3 Poberžnik, M., Purg, M., Repič, M., Mavri, J., Vianello, R., in print. Empirical valence bond simulations of the hydride transfer step in the monoamine oxidase A catalyzed metabolism of noradrenaline. J. Phys. Chem. B. doi: http://pubs.acs.org/doi/abs/10.1021/acs.jpcb.6b09011 Pettersen, E.F., Goddard, T.D., Huang, C.C., Couch, G.S., Greenblatt, D.M., Meng, E.C., Ferrin, T.E., 2004. UCSF chimera - A visualization system for exploratory

research and analysis. Journal of Computational Chemistry 25, 1605-1612. doi: http://dx.doi.org/10.1002/jcc.20084 Ralph, E.C., Hirschi, J.S., Anderson, M.A., Cleland, W.W., Singleton, D.A., Fitzpatrick, P.F., 2007. Insights into the mechanism of flavoprotein-catalyzed amine oxidation from nitrogen isotope effects on the reaction of Nmethyltryptophan oxidase. Biochemistry-Us 46, 7655-7664. doi: http://dx.doi.org/10.1021/bi700482h Reid McDonald, G., Olivieri, A., Ramsay, R. R., Holt, A., 2010. On the Formation and Nature of the Imidazoline I2 Binding Site on Human Monoamine Oxidase-B. Pharmaceutical Research 62, 475-488. doi: 10.1016/j.phrs.2010.09.001 Repič, M., Purg, M., Vianello, R., Mavri, J., 2014a. Examining Electrostatic Preorganization in Monoamine Oxidases A and B by Structural Comparison and pK(a) Calculations. Journal of Physical Chemistry B 118, 4326-4332. doi: http://dx.doi.org/10.1021/jp500795p Repič, M., Vianello, R., Purg, M., Duarte, F., Bauer, P., Kamerlin, S.C.L., Mavri, J., 2014b. Empirical valence bond simulations of the hydride transfer step in the monoamine oxidase B catalyzed metabolism of dopamine. ProteinsStructure Function and Bioinformatics 82, 3347-3355. doi: http://dx.doi.org/10.1002/prot.24690 Shih, J.C., Chen, K., Ridd, M.J., 1999. Monoamine oxidase: From genes to behavior. Annual Review of Neuroscience 22, 197-217. doi: http://dx.doi.org/10.1146/annurev.neuro.22.1.197 Son, S.Y., Ma, A., Kondou, Y., Yoshimura, M., Yamashita, E., Tsukihara, T., 2008. Structure of human monoamine oxidase A at 2.2-angstrom resolution: The control of opening the entry for substrates/inhibitors. Proceedings of the National Academy of Sciences of the United States of America 105, 57395744. doi: http://dx.doi.org/10.1073/pnas.0710626105 Tan, A.K., Ramsay, R.R., 1993. Substrate-Specific Enhancement of the Oxidative Half-Reaction of Monoamine-Oxidase. Biochemistry 32, 2137-2143. doi: http://dx.doi.org/10.1021/bi00060a003 Tucker, S.A., Amszi, V.L., Acree, W.E., 1993. Studying Acid-Base Equilibria in 2Phase Solvent Media - Analyzing Results Using Aqueous Surfactant Solutions with Organic-Solvents in the Titration of Analytes with Low Solubilities and Low Dissociation-Constants. Journal of Chemical Education 70, 80-82. doi: http://dx.doi.org/10.1021/ed070p80 Vianello, R., Domene, C., Mavri, J., 2016. The Use of Multiscale Molecular Simulations in Understanding a Relationship between the Structure and Function of Biological Systems of the Brain: The Application to Monoamine Oxidase Enzymes. Front Neurosci 10. doi: http://dx.doi.org/10.3389/fnins.2016.00327

Vianello, R., Repič, M., Mavri, J., 2012. How are Biogenic Amines Metabolized by Monoamine Oxidases? European Journal of Organic Chemistry 2012, 70577065. doi: http://dx.doi.org/10.1002/ejoc.201201122 Wang, J., Edmondson, D.E., 2011. H-2 Kinetic Isotope Effects and pH Dependence of Catalysis as Mechanistic Probes of Rat Monoamine Oxidase A: Comparisons with the Human Enzyme. Biochemistry 50, 7710-7717. doi: http://dx.doi.org/10.1021/bi200951z Wolf, M.E., Mosnaim, A.D., 1983. Phenylethylamine in Neuropsychiatric Disorders. General Pharmacology 14, 385-390. doi: http://dx.doi.org/10.1016/03063623(83)90020-4 Youdim, M.B.H., Edmondson, D., Tipton, K.F., 2006. The therapeutic potential of monoamine oxidase inhibitors. Nature Reviews Neuroscience 7, 295-309. doi: http://dx.doi.org/10.1038/nrn1883 Zapata-Torres, G., Fierro, A., Barriga-Gonzalez, G., Salgado, J.C., Celis-Barros, C., 2015. Revealing Monoamine Oxidase B Catalytic Mechanisms by Means of the Quantum Chemical Cluster Approach. Journal of Chemical Information and Modeling 55, 1349-1360. doi: http://dx.doi.org/10.1021/acs.jcim.5b00140 Zenn, R.K., Abad, E., Kastner, J., 2015. Influence of the Environment on the Oxidative Deamination of p-Substituted Benzylamines in Monoamine Oxidase. Journal of Physical Chemistry B 119, 3678-3686. doi: http://dx.doi.org/10.1021/jp512470a

Figure legends Figure 1 Structure of the hydrated MAO B enzyme with the PEA substrate used in the simulation. The sphere with the radius of 30 Å is shown in cyan blue. The substrate and the flavin moiety atoms are shown explicitly.

Figure 2 A snapshot from the equilibrated structure of MAO B with the PEA substrate corresponding to the reactants. The substrate, flavin cofactor and nearby tyrosine residues forming the aromatic cage are shown in atomic resolution.

Figure 3 Free energy profiles for the reaction involving the p-chloro--methylphenylamine (p-CMP) substrate in the gas phase (above) and the MAO B enzyme (below). For computational details see text.