In Silico design of hydroxylated coumarins and thermodynamic investigation of their free radical scavenging mechanism

Coumarin and its derivatives have been found to exhibit antioxidant activities. However, a detailed description of the thermodynamics of free radical scavenge by these compounds has not received adequate attention. In order to explore the reaction energetics of the free radical scavenging mechanism of newly designed coumarin compounds, a theoretical study, based on density functional theory (DFT) using the Becke’s three-parameter Lee-Yang-Parr hybrid functional (B3LYP) in connection with the 6-311G* basis set, was employed. Hydrogen atom transfer (HAT), single electron transfer followed by proton transfer (SET-PT) and sequential proton loss electron transfer (SPLET) mechanisms were investigated by thermodynamic studies in the gas phase and aqueous solution. Various reaction enthalpies such as bond dissociation enthalpy (BDE), adiabatic ionization potential (AIP), proton dissociation enthalpy (PDE), proton affinity (PA), electron transfer enthalpy (ETE) and Gibbs free energy changes that characterize these mechanisms were calculated. The high solvation enthalpies of the electron and proton resulted in lower values of AIP and PA respectively in aqueous solution than in the gas phase. The reaction Gibbs free energy results indicate that hydroxylated coumarin molecules are efficient free radical scavengers by HAT and SPLET mechanisms in the gas phase and aqueous solution. The SET-PT mechanism was thermodynamically prohibited in the gas phase, but feasible in aqueous solution. Also these compounds were observed to be more reactive towards HOO. than CH3OO . radicals.


INTRODUCTION
The coumarins belong to the benzopyrone family, which are abundantly distributed among the plant genus Calophyllum. They occur in the fruits, leaves, stems and roots (Jain Joshi, 2012). They represent a class of active compounds associated with a wide spectrum of important biological activities such as anti-inflammatory Arora et al., 2014), antiviral (Shen et al., 2018), antitumor Salem et al., 2016;An et al., 2018), antibacterial (Khan et al., 2004), antifungal (Khan et al., 2004;Forezi et al., 2018), anticancer (Morsy et al., 2017) and anticoagulant (Lei et al., 2015) activities. Also, the antioxidant activities of coumarin and its derivatives have been investigated widely Arora et al., 2014;Salem et al., 2016;Pe´rez-Cruz et al., 2018;Onar and Vardar, 2018;Xia et al., 2018). Antioxidants are compounds that scavenge free radicals by employing various mechanisms in their mode of action in the human system. Three major mechanisms of free radical scavenging have been recognised. They include, hydrogen atom transfer (HAT), single electron transfer followed by proton transfer (SET-PT) and sequential proton loss electron transfer (SPLET) (Medina et al., 2014;Ngo et al., 2016;Bayat and Fattahi, 2017). The trend in the implication of oxidative damage in many pathology cases (e.g. atherosclerosis, hypertension, diabetes, and cardiovascular diseases) has triggered interest in the design and synthesis of new antioxidants with potent free radical scavenging properties. However, the mechanism of free radical scavenge by these compounds needs to be analysed.
In this research, the earlier developed quantitative structure activity relationship (QSAR), model for coumarin antioxidants (Alisi et al., 2018), was employed in the design of a new set of coumarin compounds by ligand based virtual screening technique. Virtual screening by the method of QSAR is highly beneficial in the prediction and identification of new biologically active compounds. (Melagraki et al., 2009;Asadollahi et al., 2011). This technique has emerged as a reliable, quick and costeffective technique for the discovery of lead compounds. The QSAR model was also employed in predicting the antioxidant activities of the designed compounds. The applicability domain of these compounds was accessed by the leverage approach. Three of the compounds with best antioxidant activities were subjected to thermodynamic studies in order to generate the various reaction enthalpies that influence their antioxidant mechanisms. The frontier orbital energies of the molecules and the spin density distribution of their radicals were computed. Three reaction mechanisms namely, HAT, SET-PT and SPLET, were considered. The various reaction enthalpies namely, the electron transfer enthalpy (ETE), the adiabatic ionization potential (AIP), the bond dissociation enthalpy (BDE), the proton dissociation enthalpy (PDE) and the proton affinity (PA), that characterize the various steps in these mechanisms were computed.
A crucial factor that influences the radical scavenging mechanism is the characteristics of the free radicals scavenged (Amic´ et al., 2017;Xie et al., 2014). The HOO⋅ and CH3-OO⋅ are examples of peroxyl radicals (ROO⋅), which also includes methyl peroxyl radical ( CH3-OO⋅), vinyl peroxyl radical (CH2=CH-OO⋅), allyl peroxyl radical ( CH2=CH-CH2-OO⋅), etc. The ROO⋅ are formed within living systems and have been proposed as free radicals of biological relevance which need to be efficiently scavenged in order to retard oxidative stress (Galano and Alvarez-Idaboy, 2013). This is attributed to their moderate halflives, which are an efficient requirement for interception by phenolic compounds. They may mimic lipid peroxyl radicals (LOO⋅), which are abundantly formed in biological systems. The thermodynamically preferred mechanism of free radical scavenging in the present research was decided by calculating the Gibbs free energy change of the reaction (Δ r G), for scavenging the two important peroxyl radicals, the hydroperoxyl radical (HOO⋅) and the methyl peroxyl radical (CH 3 OO⋅) in the gas phase and aqueous solution.

In Silico design of new compounds
With the aid of the developed QSAR model for coumarin antioxidants (Alisi et al., 2018), a new set of coumarin antioxidants were designed by ligand based virtual screening. This was achieved by insertion, deletion and substitution of different substitutes on the template molecule as dictated by the results of the coumarin antioxidant model (Melagraki et al., 2009;Asadollahi et al., 2011;Mitra et al., 2011;Alisi et al., 2019). Compound M45 (Fig. 1) listed in Table 1 of (Alisi et al., 2018), was observed to possess the best antioxidant activity ( pIC 50 = 5.769). Based on this reason, it was chosen as the template molecule in the present research. The chemical structures of the compounds were drawn with the aid of the ChemDraw program (Li et al., 2004). Also, minimization and further optimization of the molecular structures were accomplished at the density functional theory (DFT), level using the Becke's threeparameter Lee-Yang-Parr hybrid functional (B3LYP), and 6-311G* basis set without symmetry constraints with the aid of Spartan 14 program (Shao et al., 2006). Through this computation, quantum chemical descriptors were generated. The optimized structures were converted into Sdf files and submitted for the generation of molecular descriptors using the PADEL program package version 2.20 (Yap, 2011). The effects of the structural modifications on the antioxidant activity of the resulting compounds were investigated using the coumarin antioxidant model. Subsequently, the applicability domains of the designed coumarins were assessed by the leverage approach.

Calculation of antioxidant descriptors
The evaluation of the preferred mechanism of free radical scavenge was accomplished by the computation of various antioxidant descriptors in the gas phase and aqueous solution as presented below: The homolytic bond dissociation enthalpy (BDE), was calculated under standard conditions of 1 atm. and 298.15 K using Eq. 1. BDE is the standard enthalpy change at a given temperature when a particular chemical bond is broken under standard conditions (Ruscic, 2015). The BDE value determines the stability of the corresponding hydroxyl group. When the BDE value is low, the stability of the corresponding O-H bond is low, and thus can easily be broken (Sun and Jin, 2013). This gives rise to better antioxidant activity for the considered molecule.
(1) Adiabatic ionization potential (AIP), values represents the ability of the antioxidant to transfer electrons to the free radical. It describes the process of electron donation by the antioxidant. The lower the AIP value for a given molecule, the easier is the ability to transfer electrons. Molecules with lower AIP values have higher tendency to undergo ionization. Such molecules have been observed to possess stronger antioxidant activities (Işın, 2016). The AIP values were estimated according to Eq. (2); (2) The proton dissociation enthalpy (PDE), describes the ability of the molecules to donate protons. The computation of PDE values was accomplished using Eq. (3). Antioxidants with lower PDE values are more susceptible to proton abstraction (Mikulski et al., 2014).
(3) Proton affinity (PA) is the molar enthalpy change at 298.15 K. lower PA values result in higher antioxidant activity of molecules. This parameter was calculated using Eq. (4); (4) The electron transfer enthalpy (ETE), values were calculated using Eq. (5). The lower the ETE value for a given molecule, the more active is the resulting phenoxide anion. In this study, the total enthalpies of the species were obtained as the sum of total electronic energy, zero-point energy and the translational, rotational and vibrational contributions to the total enthalpy as presented in Eq. 6. The RT (PV-work) term was added to convert the energy to enthalpy (Najafi et al., 2011). (6) Where, are the translational, rotational, and vibrational contributions to the enthalpy respectively.  (Nenadis and Tsimidou, 2012;Bartmess. 1994;Bizarro et al., 1999;Rimarcik et al., 2010). Geometry optimization of all molecular structures in the gas phase was performed at the DFT/B3LYP/6-311G* level of theory. Water (ε=78.39) is the physiological medium of human living cells. Consequently, the computation of the solvation effect of water on the antioxidant activity was accomplished using the self-consistent reaction field (SCRF) method with a polarized continuum model (PCM) Bayat and Fattahi, 2017) at the DFT/B3LYP/6-31G* level.

Investigation of the thermodynamically favoured mechanism
The reaction Gibbs free energy (∆ r G), is employed to determine the thermodynamically favoured mechanism (Amic´ et al., 2017;Zheng et al., 2017). In this research, the changes in Gibbs free energy between the reactants and products in the gas phase and aqueous solution for the studied mechanisms of the reactions between the antioxidants and the peroxyl radicals (HOO⋅ and CH 3 -OO.) were estimated.
The reaction between a free radical and an antioxidant is said to be thermodynamically favourable if the reaction Gibbs free energy is exergonic (Eq. 7).
The Gibbs free energy change for the HAT mechanism is given by (Eq. 8).
The Gibbs free energy change for the SET-PT mechanism is given by (Eq. 9) and (Eq. 10). While that of the SPLET mechanism is given by (Eq. 11) and (Eq. 12). = Gibbs free energy of anion.
In the above computations, the values of and were employed as the Gibbs free energy of the electron (e − ) and proton (H + ) respectively in the gas phase. However, in aqueous solution, and were used as the Gibbs free energy for the electron and proton respectively (Tissandier et al., 1998;Hwang and Chung, 2005).

Analysis of ligand based virtual screening
A total of twenty coumarin antioxidant derivatives with improved antioxidant activities were successfully designed from the template molecule ( Figure 1) and their plC 50 results and leverage values calculated and presented in Table 1. All the designed molecules recorded improved antioxidant activity when compared with the template molecule (MCM 45). Also the results of applicability domain investigation for the developed coumarin model gave a leverage threshold, h* value of 0.4286 (Alisi et al., 2018).  The calculated leverage results for the coumarin molecules revealed the presence of two structural outliers, which are outside the applicability domain of the developed coumarin antioxidant model. These include MCM 001 and MCM 018 molecules. The remaining molecules were found within the applicability domain of the coumarin model with MCM 002, MCM 008, and MCM 015 having the best antioxidant activities. The molecular structures of these compounds are presented in Figure 2.
It was observed that the molecule MCM 015, which has the best antioxidant activity among the designed molecules, contains two phenol groups which are hydrogen donor groups attached at positions 3 and 13, and two chlorine atoms which are electronegative groups attached at positions 2 and 14 on rings B and A, respectively. The two phenol groups are fused together to give a total of five rings for this molecule. This unique structure gave rise to the exceptionally good antioxidant activity of this molecule as reflected in its pIC 50 value ( Table 1).

Analysis of the HAT mechanism
The calculated BDE results at the various OH positions for MCM 002, MCM 008 and MCM 015 in the gas phase and aqueous solution are presented in Table 2. For MCM 002, the BDE at the 2-OH position in vacuum has the highest value of 317.72 kJ/mol in comparison to the other positions which have values of 276.11 kJ/mol, 287.37 kJ/mol and 274.08 kJ/mol at the 5-OH, 7-OH and 8-OH positions respectively. The high BDE value of MCM 002 at the 2-OH position could be attributed to the existence of an intramolecular hydrogen bond between the 2-OH and the adjacent ketone group in the parent molecule. Observe that hydrogen abstraction from 2-OH also requires breaking the intramolecular hydrogen bond. The existence of a hydrogen bond also accounts for the high BDE of MCM 008 at the 8-OH position and MCM 015 at the 16-OH position.
The lowest BDE for MCM 002 occurred at the 8-OH with a value of 274.08 kJ/mol (Table 2). This is closely followed by that at the 5-OH position with a value of 276.11 kJ/mol. These results confirm that HAT from the 8-OH or 5-OH occurs more readily than from the 7-OH or 2-OH positions. Therefore, the primary site of free radical attack on the MCM 002 molecule, based on the HAT mechanism, has almost equal chances at either the 8-OH or 5-OH positions. Similarly, MCM 015 with lowest values of BDE at the 9-OH (253.79 kJ/mol) position, has this position as the most probable site for free radical attack based on the HAT mechanism. For MCM 008, four possible sites for HAT exist with their associated BDE's namely; the 5-OH (282.63 kJ/mol), 6-OH (280.57 kJ/mol), 7-OH (288.94 kJ/ mol) and 8-OH (289.97 kJ/mol). These positions have very close BDE with the 6-OH having the lowest value, while the 8-OH has the highest value. This implies that there are very close chances of HAT occurring in any of the four -OH positions with the 6-OH position having the greatest chance. A similar sequence with lower BDE values was observed in aqueous solution. The results of BDE for the three compounds at all the possible sites of free radical attack are lower than that of phenol (327.55 kJ/mol) which is usually considered as a reference compound. This shows that these compounds are better free radical scavengers than phenol through HAT. From the results presented in Table 2, the BDE values in water are lower than the corresponding values in water. For instance, the BDE value of MCM 015 at the 9-OH site in vacuum is 253.79 kJ/mol. While the value in water is 222.62 kJ/mol. This suggests that HAT mechanism is more preferred in aqueous solution than in the gas phase.

Analysis of spin density distribution
To explain the differences in the BDE results, the spin density distribution of MCM 002, MCM 008 and MCM 015 radicals were calculated and presented in Table 2. The spin density distribution is a useful parameter employed in rationalising the stability of the antioxidant radical (Xue et al., 2014;Wang et al., 2015;Zheng et al., 2017). The lower the spin density on the radical, the more delocalised is the spin density, the easier the radical is formed, the higher the stability of the radical, the lower the BDE, the higher the antioxidant activity of the compound. As presented in Table  2, results of the spin density on the oxygen atom of the different OH radicals indicate that the stabilization of these radicals for MCM 002 increases in the order MCM 002 2-OH ˂ MCM 002 7-OH ˂ MCM 002 5-OH ˂ MCM 002 8-OH . Similarly for MCM 008 radicals, we have MCM 008 8-OH ˂ MCM 008 7-OH ˂ MCM 008 5-OH ˂ MCM 008 6-OH. Also, the increasing order for MCM 015 radicals is: MCM 015 16-OH ˂ MCM 015 15-OH ˂ MCM 015 7-OH ˂ MCM 015 9-OH. This is in very good agreement with the order in which their BDE decrease. Further, the density is more delocalised for radicals issued from MCM 002 8-OH, MCM 008 6-OH and MCM 015 9-OH for the three considered antioxidant compounds respectively. These are the sites with the lowest BDE value, and subsequently the preferred site of free radical scavenge in each of these molecules.

Analysis of SET-PT mechanism
The results of adiabatic ionization potential (AIP) and proton dissociation enthalpy (PDE) for the studied molecules and their radicals in vacuum and water are presented in Table 2. The first step of the SET-PT mechanism is characterized by AIP. Among the various sites that are subject to electron abstraction for the MCM 002 molecule in vacuum, the 2-OH site is observed to possess the highest electron donation ability to the free radical, since it has the lowest AIP value of 447.43 KJ/mol. This value is also lower than that of phenol (572.78 kJ/ mol) computed at the same level of theory. This implies that electron abstraction at this site is more favoured than that of phenol. The MCM 008 molecule has very close results for the various sites with the 8-OH position having the lowest value of 581.61 kJ/mol. The results of AIP for all the sites in this molecule are higher than that of phenol ( . These results show that MCM 015 at the various possible sites possess stronger electron donating ability than phenol with the 9-OH sites having the best potential. In aqueous solution, the AIP results are by far lower than the corresponding values in vacuum. Consequently, these molecules possess higher tendency to lose electrons to the free radical in aqueous solution than in vacuum. The lower AIP values in aqueous solution could be attributed to the high solvation enthalpies of the electron. The PDE characterizes the second step of the SET-PT mechanism. The general trend of PDE is analogous to that of BDE (Table 2). This is due to the fact that proton dissociation results in radical formation as in bond dissociation. The molecules of MCM 002, MCM 008 and MCM 015 have lowest PDE values of 1024.05 KJ/mol, 1017.95 kJ/mol and 1036.52 kJ/mol at the 8-OH, 6-OH and 9-OH positions respectively in vacuum. These values are lower than that of the reference molecule (phenol) which has a value of 1076.59 kJ/mol at the same computation level. This suggests that these molecules at the stated sites possess greater proton dissociation ability than phenol. Also, the observed results of PDE in water are lower than the corresponding values in vacuum. For reactions that involve multiple step mechanism such as the SET-PT, the first step is the most important from a thermodynamic view point. Therefore the results of AIP dominate those of PDE.

Analysis of SPLET mechanism
The proton affinity (PA) and electron transfer enthalpy (ETE) results for MCM 002, MCM 008 and MCM 015 in vacuum and water are presented in Table 2. The antioxidant parameter employed in the quantification of the first step of a SPLET mechanism is PA, whereas, the second step is quantified by ETE. For each of the title molecules, MCM 002, MCM 008 and MCM 015, the highest recorded PA results in vacuum occurred at the 5-OH, 8-OH and 15-OH positions with values of 1359.51 kJ/mol, 1391.63 kJ/mol and 1331.58 kJ/mol respectively. This demonstrates that the formation of phenoxide anion at these positions is more difficult than at other positions for each of the considered molecules. This could be attributed to the formation of an intramolecular hydrogen bond between the OH groups at these positions and the respective electronegative groups such as chlorine or oxygen. The lowest results of PA in vacuum for each of the molecules MCM 002, MCM 008 and MCM 015 occurred at the 7-OH site with values of 1309.47 kJ/mol, 1342.10 kJ/mol and 1261.16 kJ/mol respectively. These are the preferred sites for deprotonation in each of the molecules. The PA values in water are by far lower than the corresponding values in vacuum. This is a consequence of the high solvation enthalpy of the proton. Also, the computed results of PA at all the considered sites for each of the molecules MCM 002, MCM 008 and MCM 015 in both vacuum and water are lower than the corresponding values for phenol (1462.89 kJ/mol in vacuum; 141.47 in water).
For the molecule MCM 002, the lowest ETE value of 238.42 kJ/mol was observed at the 5-OH position in vacuum. In MCM 008, this was observed at the 8-OH position with ETE values of 220.16 kJ/mol and -988.37 kJ/mol in vacuum and water respectively. For MCM 015 molecule, the lowest ETE value was also observed in ring A in both media. The 5-OH and 8-OH positions are located on ring A of the coumarin moiety. This result indicates that the OH group in ring A contributes greatly to the second step of the SPLET mechanism compared to other rings of coumarin antioxidants. Observe that the ETE value in each case is lower in water than in air. This trend cuts across all the considered sites for the three molecules. From Table  2, the observed results of ETE for each of the molecules in vacuum are lower than those of AIP. Consequently, the single electron transfer from the anionic form of the coumarin antioxidant to the free radical is preferred to that from the neutral form. This is in line with literature results (Rimarcik et al., 2010;Xue et al., 2014;Wang et al., 2015).

Thermodynamically preferred mechanism
The Gibbs free energy changes of reactions of inactivation of HOO⋅ and CH 3 OO⋅ by coumarin antioxidant molecules via the HAT, SET-PT and SPLET mechanisms in vacuum and water are reported in Tables 3 and 4 respectively. Results that emanate from the ∆ r G computation are useful parameters employed in the assignment of the free radical scavenging potency of molecules, because more negative values of the reaction Gibbs free energy change (∆ r G) represent the thermodynamically more preferred reactions. The molecules of MCM 002, MCM 008 and MCM 015 are able to effectively scavenge HOO⋅ and CH 3 OO⋅ by HAT and SPLET mechanisms in vacuum and water at all their considered sites of free radical scavenge, because these processes are exergonic. In both the gas phase and aqueous solution, the exergonicity of ∆ r G PA in the first step overwhelmed the endergonicity of ∆ r G ETE in the second step of the SPLET mechanism, since a thermodynamically unfavourable reaction could be driven by thermodynamically favourable reaction that is coupled to it. These molecules are not able to scavenge HOO⋅ and CH 3 OO⋅ by SET-PT mechanism in vacuum, because the first step (∆ r G AIP ) which is the most important is endergonic, while the second step (∆ r G PDE ) is exergonic. Consequently, the entire SET-PT pathway is not feasible in vacuum. In aqueous solution, the first step (∆ r G AIP ) is exergonic and overwhelmed the endergonicity of the second step (∆ r G PDE ). Subsequently, the entire process is exergonic and thermodynamically feasible in aqueous solution.
A comparison of the results presented in Tables 3 and 4, reveals that the investigated coumarin antioxidants are better free radical scavengers for HOO⋅ than CH 3 OO⋅ radicals by HAT, SPLET and SET-PT mechanisms. Among the tested molecules, MCM 015 at the 9-OH position exhibited the highest level of exergonicity in its reactions based on the computed results of ∆ r G. Consequently, the reactions of inactivation of free radicals by MCM 015 via HAT, SPLET and SET-PT mechanisms are presented in Scheme 1 (a), (b) and (c) respectively.

Analysis of frontier molecular orbitals
The frontier orbital distributions and energy for MCM 002, MCM 008 and MCM 015 computed at the B3LYP/6-311G* level in the gas phase are presented in Figure 4. The LUMO distribution occurs almost over the entire molecules of MCM 002, MCM 008 and MCM 015. The HOMO appears to be delocalised over the entire molecule of MCM 015. For MCM 008, it is localised on the conjugated carbons of ring A and its associated oxygen atoms. For MCM 002, it Scheme 1: (a) HAT mechanism of MCM 015 in vacuum (b) SPLET mechanism of MCM 015 in water (c) SET-PT mechanism of 015 in water. is localised on the oxygen atoms of the hydroxyl group and the conjugated carbons of rings A and B.
From Figure 4, MCM 015 has the highest HOMO energy (-5.62 eV), followed by MCM 008 (-5.91 eV) and MCM 002 (-6.03 eV) with almost equal results. Since molecules with higher HOMO values have stronger electron donating potentials (Xue et al., 2014;Szeląg etal., 2011;Işın, 2016), MCM 015 will have a higher tendency to donate electrons. Also, the electron donating ability of these molecules predicted by   their HOMO energies follow the same sequence as that of their AIP and BDE at their preferred positions of radical scavenge (Table 2). Consequently, MCM 015 is predicted to have a higher antioxidant activity in comparison to the other two molecules.

CONCLUSION
In the present study, rational design of coumarin antioxidants with improved free radical scavenging properties has been effected by virtual screen using the proposed coumarin antioxidant model. With the aid of the developed model, the molecular structure and antioxidant activities of these molecules were successfully predicted. Among these molecules, MCMM 015 (1,6-dichloro-4,5,9,10-tetrahydroxy-2H-phenalenol[1,2,3-de]chromen-2-one) possessed the best antioxidant activity with a value of 12.898. Our results indicate that the employed screening method is time and cost effective in the design and activity determination of new set of antioxidants.
Three of the best deigned molecules namely, MCMM 02, MCMM 08 and MCMM 15 were further subjected to thermodynamic studies by the DFT method in the gas phase and aqueous solution. Reaction enthalpies such BDE, AIP, PDE, PA and ETE which characterise various reaction mechanisms such HAT, SET-PT and SPLET were calculated. In order to rationalise the differences in the BDE values for each molecule, the spin density distributions of the radicals were calculated. All considered coumarin molecules have better hydrogen atom donation ability compared to the reference molecule with MCMM 15 having the best ability as judged from the BDE values. In the gas phase, MCMM 002 at the 2-OH site possessed the strongest electron donation ability from the computed results of AIP. Also from the computed results of PA, the preferred site for deprotonation is the 7-OH position in each of the molecules MCMM 002, MCMM 008 and MCMM 015.
In order to investigate the preferred mechanism of free radical scavenge by each of the antioxidant molecules, the reaction Gibbs free energy change for scavenging HOO⋅ and CH 3 OO⋅ at the various OH sites were computed. From the exergonicity of the observed results, the HAT and SPLET mechanisms were thermodynamically feasible in the gas phase and aqueous solution. The SET-PT mechanism was thermodynamically prohibited in the gas phase, but favourable in aqueous solution. Also, the investigated coumarin antioxidants were better scavengers for HOO⋅ than CH 3 OO⋅ by these mechanisms. The present study is a gateway to new perspectives towards a rational exploitation of the coumarins in the field of pharmacy and food chemistry.