Introduction

Many biochemical reactions in our body involve the generation of reactive oxygen species (ROS) or reactive nitrogen species (RNS). Under normal conditions, the balance between the generation and reduction in ROS is controlled by the body’s antioxidant defence mechanism. However, under certain pathological conditions, when ROS are not effectively eliminated by the body’s antioxidant defence mechanism, the dynamic balance between the generation and elimination of ROS is broken. As a result, ROS may attack the biomolecules such as lipids, carbohydrates, proteins and DNA, causing cellular damage and resulting in oxidative stress. The consequence of cellular damage is the development of various degenerative diseases such as cardiovascular, neurodegenerative and Alzheimer’s diseases, and cancers [1–4]. In order to protect against cellular damage by ROS, both synthetic and natural antioxidant molecules are increasingly utilised to enhance the antioxidant capability of the body system. However, synthetic antioxidants are less preferred to natural products because of their adverse side effects on some of the body organs, such as liver [5–7]. Thus, considerable attention has been given to the search for natural antioxidants (often from plant sources). The current study focuses on the antioxidant potential of phenolic acid amide derivatives. These compounds have been reported both as natural products and as synthetic products. They possess various biological activities including antibacterial, antiviral, anti-inflammatory, antioxidative, and antiproliferative [8–14]. Phenolic acid amides are conjugates of aromatic amines to phenolic acids. Both aromatic amines and phenolic acids have been reported to possess antioxidant activities [15–20]. Therefore, a study of the antioxidant activity of phenolic acid amides would yield useful information on the effect of conjugation of the constituent molecules on the antioxidant properties of the phenolic acid amides. The aromatic amines selected for the study are phenylethylamine, tyramine and dopamine, and the phenolic acids selected for the study are cinnamic acid, coumaric acid and caffeic acid. The conjugated phenolic acid amides selected for the study include caffeoyltyramine (CaTy), caffeoyldopamine (CaD), caffeoylphenylethylamine (CaPEA), cinnamoyldopamine (CiD) and p-coumaroyldopamine (p-CoD). The structures of the selected compounds are shown in Fig. 1 together with the atom numbering interesting for the discussion of the study.

Fig. 1
figure 1

Schematic representation of the studied phenolic acid amides, and the atom numbering for the main skeletal is shown on caffeonyldopamine structure. The O atoms and the H atoms are given the number of the C atom to which they are bonded. The phenolic acid ring is denoted X and the aromatic amine ring is denoted Y

Full size image

Although extensive experimental and theoretical studies on the antioxidant activity of the isolated aromatic amines and phenolic acids have been reported [20–26], only experimental findings have been reported extensively on the antioxidant activity of phenolic acid amide derivatives [10, 12, 14, 27]. Therefore, the objective of this work is to study the electronic and molecular properties of these compounds in order to understand the mechanism of their antioxidant activity and to compare the information thereof with that of the constituent phenolic acids and aromatic amines. The study is performed utilising the density functional theory (DFT) with the B3LYP functional, for both the study of both the antiradical activity and the metal chelation mechanism. The calculations are performed in vacuo, in benzene, in acetonitrile and in water solvents. The investigation in different media is meant to mimic the different aqueous and lipophilic environments found in biological systems where antioxidants interact with both radical species and toxic transition metal ions [4, 28, 29].

The schematic representation of the selected phenolic acid amide derivatives and the atom numbering utilised in this work are shown in Fig. 1. The structures are arranged in order of increasing number of hydroxyl groups on the aromatic amine ring (ring Y) and decreasing number of hydroxyl groups on the phenolic acid ring (ring X). The two main torsional degrees of freedom about the aliphatic chain (the C1′–C12–C11–N10 and the C12–C11–N10–C9′ torsion angles) and the single C−O bonds, connecting the OH groups to the aromatic rings, are flexible, and their rotation may result in several possible conformations. Because of the presence of catecholic moiety (moieties) in all the compounds, it is expected that all the compounds can be stabilised by intramolecular hydrogen bonds (IHB) involving the phenolic OH groups at C3 and C4 or those at C3′ and C4′. IHB are known to play significant roles in stabilising molecular conformations and in influencing biological activities [30–42]. For instance, the antioxidant activities of several biological molecules have been reported to be enhanced by the presence of the IHB because of the role of this interaction is stabilising the radical species formed due to the transfer of the hydrogen atom or the electron [3, 4, 29, 43, 44].

The assessment of antioxidant properties by a single mechanism underestimates the antioxidant potential for that particular molecule because it reflects only the ability to inhibit a precise class of oxidants present. To this effect, the current study investigates the antioxidant mechanism of phenolic acid amides through both the antiradical and the metal chelation mechanisms. The antiradical-scavenging activity is often discussed through four mechanisms, the hydrogen atom transfer (HAT, [45–47]), the single-electron transfer (SET, [48]), the single-electron transfer followed by proton transfer (SET-PT, [49]) and the sequential proton loss electron transfer (SPLET, [50]) mechanisms. The HAT, SET-PT and SPLET mechanisms depend on the presence of free phenolic OH group(s). The HAT mechanism is related to the ability of the molecule to donate its phenolic H atom to the radical species:

$${\text{R}}^{ \cdot } + {\text{ArOH}} \rightarrow {\text{RH}} + {\text{ArO}}^{ \cdot }$$
(1)

The termination of further chain reactions depends on the stability of the neutral radical intermediates (ArO∙ radical species) formed. This means that factors enhancing the stability of the ArO∙ radical species, such as the resonance delocalisation of the electron within the aromatic ring and the formation of the quinine structure, increase the antiradical activity. The ability of phenolic antioxidants to donate a hydrogen atom is mainly governed by the homolytic O–H bond dissociation enthalpy (BDE).

The SET mechanism is governed by the capacity of an antioxidant compound to transfer an electron and reduce any other compound including metal ions and radical species. In the case of antioxidant reactions involving radical species, the antioxidant compound (ArOH) transfers a single electron to the radical species (R∙):

$${\text{R}}^{ \cdot } + {\text{ArOH}} \rightarrow {\text{ArOH}}^{ \cdot + } + {\text{R}}^{ - }$$
(2)

Relative reactivity in SET mechanism is primarily determined by adiabatic ionisation potential (AIP) value [51, 52] of the reactive functional group. The lower the AIP value, the more favourable the electron transfer reaction [48, 53].

The single-electron transfer followed by proton transfer (SET-PT, [49]) is a two-step reaction mechanism. In the first step, a phenolic antioxidant molecule reacts with the free radical giving rise to the cationic radical form of the phenolic antioxidant and an anionic form of the radical.

$${\text{R}}^{ \cdot } + {\text{ArOH}} \rightarrow {\text{ArOH}}^{ \cdot + } + {\text{R}}^{ - }$$
(3.1)

This reaction is a thermodynamically significant step of this two-step mechanism. In the second step, the cationic radical form of the phenolic antioxidant decomposes into a phenolic radical and a proton.

$${\text{ArOH}}^{ \cdot + } \rightarrow {\text{ArO}}^{ \cdot } + {\text{H}}^{+}$$
(3.2)

The numerical parameters related with the SET-PT mechanism are the adiabatic ionisation potential (AIP) for the first step and proton dissociation enthalpy (PDE) for the second step.

The sequential proton loss electron transfer reaction is also a two-step reaction mechanism. In the first step, the phenolic antioxidant dissociates into an anionic form and proton,

$${\text{ArOH}} \rightarrow {\text{ArO}}^{-} + {\text{H}}^{+}$$
(4.1)

In the second step, the ions created in the first reaction react with the free radical giving rise to the radical form of the phenolic antioxidant as well as a neutral molecule.

$${\text{ArO}}^{-} + {\text{X}}^{ \cdot } + {\text{H}}^{+} \rightarrow {\text{ArO}}^{ \cdot } + {\text{XH}}$$

The numerical parameters related with this mechanism are the proton affinity (PA), for the first reaction step, and electron transfer enthalpy (ETE) for the second step.

The other mechanism used to describe antioxidant properties is transition metals chelation (TMC, [45]). Transition metals have the ability to catalyse oxidative processes leading to the formation of hydroxyl radical species and

$${\text{ArOH}} \rightarrow {\text{ArO}}^{-} + {\text{H}}^{+}$$
(5.1)

Transition metals also have the ability to decompose hydroperoxides through the Fenton reaction:

$${\text{H}}_{2} {\text{O}}_{2} + {\text{M}}^{n + } \rightarrow {\text{HO}}^{-} + {\text{HO}}^{ \cdot } + {\text{M}}^{(n + 1) + }$$
(5.2)

In the TMC mechanism associated with the formation of the hydroxyl radical species (Eq. 5.1), each molecule which may dissociate has the ability to chelate transition metal ions. In particular, anions of polyphenols have the significant abilities to chelate transition metal ions. Since chelation of metals often occurs due to deprotonated hydroxyls in the polyphenols, the ability of a molecule to produce the proton is taken into consideration. The numerical parameter related with this mechanism is the gas-phase acidity, which is identified with the vacuum enthalpy of the compound (∆H acidity). For the calculations in solvent media, the free Gibbs energy (∆G acidity) is the parameter of interest.

Although the metal chelation mechanism is considered a minor reaction in comparison with radical-scavenging mechanisms [54], its occurrence in biological systems plays significant role in the removal of toxic transition metal ions through altering their redox potentials thereby rendering them inactive. In cases where the metal chelation mechanism is possible, it is governed by the ability of the antioxidant molecule to reduce the transition metal ion (by donating electrons to the metal ion) and to form a stable metal–ligand complex. The formation of a stable complex structure, as a result of the interaction of the antioxidant (ligand) with the toxic transition metal ion, shields biological targets from the metal ion (e.g. Fe2+, Fe3+, Cu2+, …), thereby preventing metal-catalysed free radical reactions [54, 55].

Phenolic acid amide derivatives, with their multiple hydroxyl groups, aromatic rings and the carbonyl group on the aliphatic chain, have several available sites for metal complexation. However, there has not been any investigation concerned with the formation of stable complexes between phenolic acid amides and transition metals ions. This study considers the possible 1:1 stoichiometry that metal/phenolic acid amide complexes can exhibit. The copper(II) cation (Cu2+) has been selected for the investigation mainly because Cu and Fe are the most abundant transition metal ions in biological systems.

Computational details

The geometry optimisations for the neutral species and the radical species of phenolic acid amides were performed with the DFT/B3LYP method using the 6-311+G(d,p) basis sets [28, 56–58]. The restricted wave function was employed for the closed-shell neutral species, and the unrestricted wavefunction was utilised for the open-shell radical species. Geometry optimisation of each ArO∙ radical was performed starting from the optimised structure of the parent molecule, after the H atom was removed from the possible position. Frequency calculations were performed, at the same level of calculations as the geometry optimisation, on fully optimised conformers, to determine the nature of the stationary points. Zero-point and thermal enthalpies corrections, computed at T = 298 K and p = 1 atm in the rigid rotor harmonic oscillator approximation, were performed to obtain free energies and enthalpies. Solvent effects on geometries and relative conformational stabilities have been taken into consideration using the solvent model density (SMD) model [59]. Calculations in different media were performed using the B3LYP/6-311+G(d,p) method and starting from the optimised geometries in vacuo. The utilisation of the geometry obtained in vacuo for calculations in the presence of the solvent allows for the possibility of tracking geometry changes as a result of the geometry relaxation in solution.

The adiabatic ionisation potential (AIP) was estimated using the equation [48, 53]

$${\text{AIP}} = {\text{H}}\left( {{\text{M}}^{ \cdot + } } \right){-}{\text{H}}\left( {\text{M}} \right)$$
(6)

where M is the lowest-energy conformer for each of the studied compounds, H(M˙+) is the enthalpy of the M cation radical species, and H(M) is the enthalpy of M. The bond dissociation enthalpy (BDE) was estimated by using the equation

$${\text{BDE}} = {\text{H}}_{\text{r}} \left( {{\text{ArO}}^{ \cdot } } \right) + {\text{H}}_{\text{h}} \left( {{\text{H}}^{ \cdot } } \right) - {\text{H}}_{\text{p}} \left( {\text{ArOH}} \right)$$
(7)

where Hr is the enthalpy of the radical generated by H abstraction, Hh is the enthalpy of the H atom, and Hp is the enthalpy of parent compound. A low O–H BDE value, usually associated with greater ability to donate the H atom, corresponds to high radical-scavenging ability by the phenolic compound [45, 46]. Proton dissociation enthalpy (PDE) has been calculated using the equation

$${\text{PDE}} = {\text{H}}_{\text{ArO}}^{ \cdot } + {\text{H}}_{\text{H}}^{+} {-}{\text{H}}_{\text{ArOH}}^{ \cdot + }$$
(8)

in which H ∙ArO is the enthalpy of the radical, H +H is the enthalpy of the proton, H ∙+ArOH is the enthalpy of cationic radical. Proton affinity has been estimated by using the equation

$${\text{PA}} = {\text{H}}_{\text{ArO}}^{-} + {\text{H}}_{\text{H}}^{+} {-}{\text{H}}_{\text{ArOH}}$$
(9)

where, H −ArO is the enthalpy of the anion, H +H is the enthalpy of the proton, HArOH is the enthalpy of the compound. Electron transfer enthalpy (ETE) has been estimated using the equation

$${\text{ETE}} = {\text{H}}_{\text{ArO}}^{ \cdot } {-}{\text{H}}_{\text{ArO}}^{-}$$
(10)

where H ∙ArO is the enthalpy of the radical and H −ArO is the enthalpy of the anion.

The CaD···Cu2+ complexes were optimised using UB3LYP/6-311+G(d,p) method. Harmonic vibrational frequencies were obtained at the same level of theory to classify the stationary points as local minima or transition states. For all the local minima, final energies were obtained, on the aforementioned UB3LYP/6-311+G(d,p) optimised geometries, by using a 6-311+G(3df,2p) basis set that includes diffuse functions on all heavy atoms, as well as high angular momentum functions, which may be important when describing π-complexes. For Cu the 6-311+G(3df,2p) basis set corresponds actually to the (14s9p5d/9s5p3d) Wachters–Hay [60–63] basis supplemented with a set of (1s2p1d) diffuse functions and with two sets of f functions (rather than d) and one set of g functions (rather than f).

The CaD···Cu2+-binding energy was estimated using the following equation:

$$\Delta E_{\text{binding}} = -\Delta E_{\text{inter}} = \left( {E_{{{\text{CaD}} \cdots {\text{Cu}}}}^{2 + } - E_{\text{Cu}}^{2 + } - E_{\text{CaD}} } \right)$$
(11)

where ΔE inter is the interaction energy, E 2+CaD···Cu is the total energy of the optimised complex, E 2+Cu is the total energy of the isolated Cu2+ cation, and E CaD is the total energy of the isolated CaD conformer. The greater the computed value of the binding energy, the stronger the affinity of M2+ ion to bind to the ligand [64, 65].

All calculations were performed with Gaussian09 programme [66]. The schematic representations were drawn using the ChemOffice package in the UltraChem 2010 version, and conformers were drawn using GaussView5 programme.

The natural bond order (NBO) and the atoms-in-molecule (AIM) population analysis schemes were utilised for gaining insight into the bonding situation of the metal···molecule complexes. NPA charges on the atoms were determined by means of the natural bond order (NBO) analysis scheme [67–69]. The atoms-in-molecule (AIM) procedure was performed using the AIMAll programme [70]. The number of critical points (CP) found for the metal–ligand systems are in agreement with the Poincare–Hopf rule. The following parameters of the bond critical point (BCP) were analysed: the electron density (ρ) and its Laplacian (∇2 ρ), the density of the total energy of electrons (H), and its two components, the Lagrangian kinetic electron density (G) and the potential electron density (V). The total energy density, H, was estimated as the sum of the kinetic electron density (G) and potential energy density (V):

$$H_{\text{BCP}} = G_{\text{BCP}} + V_{\text{BCP}}$$
(12)

Results and discussion

Geometrical and conformational stabilities for the isolated molecules

The optimised conformers of the neutral studied compounds are shown in Fig. 2, and their relative energies are reported in Table 1. The conformers shown are those differing in terms of the rotation about the OH groups. In all the structure, the lowest-energy conformers are stabilised by the IHB formed in the catecholic unit. The –CH2CH2NH– group prefers a near-perpendicular orientation with respect to ring Y, while the –CH=CHC(O)– prefers a planar arrangement with respect to ring X. In vacuo, the relative energy difference among conformers of each structure is so small (often <1 kcal/mol) that all conformers can be considered to exits. A study in different media shows that the energy difference between conformers is always below 1 kcal/mol and that as the solvent polarity increases, the energy order may be reversed. This result suggests that all the conformers (for each structure) can be considered populated and should be taken into consideration on further studies of the activities of the compound. However, because conformational differences have minimal effect on the antioxidant activity properties (such as BDE values [58, 71]), only the lowest-energy conformers have been taken into consideration for the study of the antiradical activity of phenolic acid amide derivatives.

Fig. 2
figure 2

Optimised lowest-energy conformers of the studied phenolic acid amides, B3LYP/6-311+G(d,p) results in vacuo. For a given structure, relative energy values are shown below each conformer. The results in solution are similar to the results in vacuo

Full size image
Table 1 Relative energy (kcal/mol) for the conformers of neutral species, neutral radical species and anionic species of the studied phenolic acid amides, B3LYP/6-311+G(d,p) results in different media
Full size table

Antioxidant properties and the hydrogen atom transfer (HAT) mechanism

The optimised neutral radical species for each of the studied phenolic amide derivative are shown in Fig. 3, the corresponding molecular spin density distributions are shown in Fig. S1, and their relative energies are reported in Table 1. The radical species are named by including the notation “rad” in the name, and the number of the O atom on which the H atom is abstracted. Only the H atom that is not engaged in IHB has been abstracted because of the knowledge that IHB is important in influencing antioxidant activity [29]. The optimised radical species are stabilised by the presence of IHB, either involving the radicalised O atom as the H atom acceptor or present in the ring on which there has not been H atom abstraction. In all the structures with catecholic unit, the lowest-energy radical species correspond to the abstraction of the H atom from the para position. When the H atom is abstracted from the phenolic acid ring, the energy difference between the para and the meta radical species is nearly 3 kcal/mol, while when the H atom is abstracted from the aromatic amine ring, the energy difference between the para and the meta radical species is lower, being <1 kcal/mol.

Fig. 3
figure 3

Optimised lowest-energy conformers of the studied phenolic acid amides neutral radical species, B3LYP/6-311+G(d,p) results in vacuo. For a given structure, relative energy values are shown below each conformer. The results in solution are similar to the results in vacuo

Full size image

The ability of the molecule to donate its hydroxyl H atom is better assessed by estimating the bond dissociation enthalpy for that O–H group. A lower BDE value means that there is high tendency for the molecule to donate its H atom, which in turn implies high reactivity and therefore high antiradical activity. Usually one compares the BDE value for the removal of any of the OH group in a given molecule with the BDE value for a removal of the H atom from some standard antioxidant molecules or molecules that are already known as good antioxidant. The BDE values for the studied phenolic acid amides are reported in Table 2. An abstraction of the H atom in para position results in lower BDE value than an abstraction of the H atom in meta position. The BDE value for the phenolic units (CaTy-aii-A, CaTy-ai-A, p-CoD-aii-C and p-CoD-ai-A) is higher (79–81 kcal/mol) than the O–H BDE values for the catecholic units (72–75 kcal/mol). The estimated BDE values, for the different radical species, are within the range of most substituted catechol and phenol derivatives. For instance, the estimated values of the BDE values of phenolic acid amides are close to those of rooperol and quercetin [29, 72].

Table 2 Antioxidant descriptors for all the studied compounds, B3LYP/6-311+G(d,p) results in different media
Full size table

A comparison of the BDE value obtained from catecholic units suggests that it is slightly lower for the case in which the catechol moiety is on the phenolic acid ring (CaPEA-ai-C and CaTy-ai-C) than for the aromatic amine ring (p-CoD-ai-A and CiD-ai-A). A similar comparison of the BDE values obtained from the phenolic unit of the aromatic amine ring and that of the phenolic acid ring shows a similar trend.

For each of the studied compound, the BDE values increase with the increase in solvent polarity, which implies that the phenolic acid amides are better H atom abstractor in non-polar media than in polar media. In polar media, an intermolecular H-bond can form between the phenolic hydroxyl, especially a catechol group, and the solvent molecule. The formation of such molecule···solvent intermolecular H-bonds may hamper the H atom-donating ability [73, 74], resulting in high BDE values in polar solvents.

Analysis of the spin density distribution (Fig. S1) suggests that the distribution of the single electron is determined by the site on the ring at which the H atom is abstracted. When the H atom is abstracted from a para position, the distribution of the single electron is delocalised furthest from the radicalised O atom than when the H atom is abstracted from the meta position. In the former case, the single electron may be distributed to a large extent up to the C=O and N–H groups of the aliphatic chain, while in the latter cases, the spin density distribution is largely on the ortho and para positions with respect to the radicalised O3 atom. An analysis of the atomic spin density of the radicalised O atom (Table 1) suggests that radical species with the least atomic spin density on the radicalised O atom have the lowest BDE value. Consistently, the more delocalised the spin density in the radical is, the easier is the radical formed and thus the lower is the BDE. The stability of the radical species of phenolic acid amides is therefore influenced to a large extent by the delocalisation of the spin density of the single electron, from the atom on which the H atom is abstracted [29]. Therefore, both IHB and spin delocalisation play significant role in the stabilisation of the neutral radical species formed from H atom transfer.

An investigation on the N–H bond dissociation enthalpy of the central N–H group was performed in order to be compared with the O–H BDE results. The outcome is reported in Table S2 and indicates that the H atom abstraction from the N–H group is very difficult because the N–H BDE value (~102 kcal/mol) is much higher than that of O–H BDE. A comparison across structures suggests that the N–H BDE value is constant across structures. The calculated spin density of O atom in phenoxyl radical derived from a given molecule is lower than that of N in N-centred radical derived from the same molecule. The two factors suggest that the hydrogen atom transfer antioxidant mechanism of phenolic acid amides is a result of H atom abstraction from the phenolic OH group rather than from the N–H group.

A comparison of the BDE values of the isolated aromatic amines (tyramine and dopamine) and phenolic acids (caffeonic acid and coumaric acid, Table S1) suggests that aromatic amines are better H atom donor than phenolic acids. A comparison of the BDE values of the isolated aromatic amines and phenolic acids with the BDE for the phenolic acid amides suggests that the conjugation of the two units results in lower BDE value.

The single-electron transfer (SET) mechanism

The calculated ionisation potential (IP) values, which characterise the electron transfer mechanism, are reported in Table 1 for all the compounds. In vacuo, the order of the IP values is such that CaD < p-CoD < CaTy < CiD < CaPEA. The IP values therefore depend on the number and the position of the OH groups present. The IP values of p-CoD and that of CiD are greater (≈1 kcal/mol) than those of CaTy and CaPEA, respectively, which suggests that when the aromatic amine moiety is of catecholic nature, the resulting phenolic acid amide is a better electron donor than the phenolic acid amide in which the phenolic acid moiety has a catecholic nature. This result is consistent with the results obtained on dopamine and caffeonic acid (Table S1), which shows that dopamine is a better electron donor than caffeonic acid. The preference for dopamine as a better electron donor may be related to the fact that the lone pair of electrons on N (present in dopamine) is readily available for donation than the lone pair of electrons on the O atom (present in the carbonyl group of caffeonic acid). The IP values for CaD, p-CoD and CaTy are lower than those of quercetin and taxifolin (170.3 and 183.4 kca/mol, respectively, obtained with the same calculation method, [75]), suggesting that phenolic acid amides have the ability to scavenge radical species through the ET mechanism.

In solution, the IP value of each compound decreases with the increase in solvent polarity, which implies that the ability of the compounds to act through SET mechanism is enhanced with the increase in solvent polarity. A comparison across media also suggests that as the polarity increases, the IP value changes minimally across structures, so that in the most polar solvent, the IP value of all the compounds has the same value of ~128 kcal/mol.

The resonance structures of the cation free radicals obtained from the electron abstraction can be observed from the distribution of the atomic spin densities for CaD, p-CoD, CaTy, CiD and CaPEA. Figure 4 shows the distribution of atomic spin densities for the cation free radicals for these compounds. The calculated spin density for the electron abstraction from CaD shows a global contribution of 0.208 for the oxygen of the OH groups, 0.351 for the global contributions of ring X, 0.178 for the global contributions of the aliphatic chain and 0.263 for the global contributions of ring Y. p-CoD shows a global spin density contributions of 0.189 for oxygen of the OH groups, a 0.367 for the global spin density contributions in ring X, a 0.225 for the global spin density contributions in aliphatic chain and a 0.238 global spin contribution for ring Y. The data for CaTy, which is a structural isomer of p-CoD, shows a global spin density contribution of 0.016 for oxygen of the OH groups, a 0.035 decrease in global spin density contributions in ring X, a 0.007 reduction in global spin density contributions in aliphatic chain and an increase of 0.031 for the global spin contribution of ring Y. An assessment of the spin density of the N atom in the cation radical species may provide an indication on their relative stability. The order of the spin density distribution on the N atom is such that CaD < p-CoD < CiD ~CaTy < CaPEA, which is close to the trend in the IP values. Therefore, the lowest localisation of the unpaired electron on the N atom may explain the highest stability of CaD cation radical when compared to the other phenolic acid amides studied in this work.

Fig. 4
figure 4

Atomic spin densities in the cation free radical species of the studied phenolic acid amide derivatives, B3LYP/6-311+G(d,p) results in vacuo for the lowest-energy conformer of each structure

Full size image

A comparison of the IP values of the separate aromatic amine moiety with the corresponding phenolic acid moiety (e.g. caffeonic acid with dopamine and coumaric acid with tyramine, Table S1) suggests that aromatic amines are better electron donor than the phenolic acids. The preference for the aromatic amines to donate electrons may be related to the presence of the N atom in the amine group. The N atom has greater tendency to donate electrons than the O atom in the C=O of phenolic acids.

The single electron transfer followed by proton transfer (SET-PT) mechanism

The formation of radical cations, in the first step of the SET-PT mechanism, requires an energy expense (kcal/mol) of 164–171 in vacuo, 145–149 in benzene, 129–130 in acetonitrile and 127–129 in water (IP values, Table 2). The PDE parameter characterises the second step of the SET-PT mechanism. The data in Table 2 indicate that in vacuo, the loss of the proton from the ArOH∙+ requires even greater energy than the electron transfer reaction because the PDE values are higher than the IP values. The high values of PDE in vacuum suggest that this mechanism is not preferred. However, in different solvents the PDE value is significantly lower than the IP value. Therefore, it is reasonable to infer that in solution, the PT transfer reaction is thermodynamically preferred to the ET reaction. The preference for the PT reaction increases with the increase in solvent polarity.

Among the isolated aromatic amine and phenolic acid compounds (Table S1), electron-donating mechanism is also preferred in vacuo, while the loss of the proton is preferred in solvent media. A comparison of the PDE value, among the isolated aromatic amines and phenolic acids, suggests that it is lower for the aromatic amines than for the phenolic acid amides for the situation in vacuo, and it is lower for the phenolic acids than for the aromatic amines in other media.

Sequential proton loss electron transfer (SPLET) mechanism

The anions resulting from the loss of the proton are shown in Fig. S2, and their relative energies are reported in Table 1. In CaPEA and CaTy, where the catecholic unit is on ring X, the deprotonation of the 4-OH group yields to the most stable anion that is characterised by an internal H-bond. In CaTy, the deprotonation on the 4′-OH position results in significantly unpopulated anionic conformer. The energetic gaps among the anionic species of CaPEA and CaTy decrease with the increase in solvent polarity. In p-CoD and CiD, where the catecholic unit is on ring Y, the deprotonation of the OH group yields conformers whose energetics are not significantly different, both in vacuo and in different solvents. These latter results indicate that all the anions of p-CoD and CiD may coexist in vivo. In CaD, the only significant anionic conformer is the one in which the deprotonation occurs at the 4-OH group.

The SPLET mechanism is better described through the use of the PA and ETE parameters. The data for these parameters are reported in Table 2. The first step in the SPLET mechanism is proton loss, and it is determined by the acidity of the OH group. The PA values for all investigated compounds decrease with an increase in solvent polarity, which suggests that the proton loss reaction is least preferred in vacuo and best preferred in polar medium. Electron transfer from the phenoxide ion represents the second step in SPLET mechanism. This step is better discussed in terms of the electron transfer enthalpy (ETE). The ETE values increase with the increase in solvent polarity. In comparison with the PA values, the ETE values are significantly lower in non-polar media (in vacuo and in benzene) and higher in acetonitrile and water than the PA values. These results suggest that the electron transfer reaction from phenoxide anions to free radical species is thermodynamically preferred in non-polar media, while the proton loss reaction is thermodynamically preferred in acetonitrile and in water. These results also imply that the first step (the dissociation of the phenolic antioxidant into phenoxide anion form and a proton) is the rate-determining step in vacuo and in benzene (non-polar media), while the electron transfer mechanism is the rate-determining step in acetonitrile and in water. The second step (that involves a single-electron transfer) is a significant step for electron-deficient radical scavenging. The phenoxide anions derived from polyphenols are better electron donors and radical scavengers than the parent molecules. Low ETE value means deep extent of the radical-scavenging reaction. Therefore, although the first step of the SPLET mechanism is the rate-determining step in vacuo and in benzene, the extent of radical scavenging is determined by the second step.

A comparison of the PA and ETE parameters for the isolated aromatic amine and the corresponding phenolic acid (e.g. caffeonic acid and dopamine, coumaric acid and tyramine, Table S1) suggests that, in all the media, the PA value is lower for the phenolic acid than for the corresponding aromatic amine, while the ETE value is lower for the aromatic amine than for the corresponding phenolic acid. It is therefore reasonable to consider that aromatic amines are thermodynamically better radical scavenger than phenolic acids through the SPLET mechanism.

Thermodynamically preferred radical-scavenging mechanism

The preference for a particular reaction mechanism depends on both the solvent and properties of the radical species involved in the reaction [76–78]. To investigate the preferred mechanism of action for the radical scavenging, it is necessary to compare the enthalpy values for the different processes. Because the calculated gas-phase ionisation potentials (IPs) and proton affinities (PAs) are significantly higher than phenolic O–H BDEs, it is reasonable to infer that that O–H bond dissociation represents the most probable process in the gas phase. The preference for the HAT mechanism in non-polar solvents may be related to the fact that it does not involve charge separation processes. The SET-PT mechanism is preferred in polar media due to the charge separation process [79]. The trend observed for the phenolic acid amides is also reproduced by the isolated phenolic acids and aromatic amine derivatives (Table S1). A comparison of the enthalpy values of the separate phenolic acids and aromatic amides with those of the phenolic acid amides clearly shows the preference for phenolic acid amides to act as radical scavengers.

Metal chelation mechanism: results of the study in vacuo

Acidities in the gas phase and in solution

Metal chelation mechanism often is a result of metal interacting with deprotonated hydroxyl in the polyphenolic compounds such as flavones [54]. The acidity values describe the ability of a molecule to lose a proton, ∆H acidity is associated with the ability of molecule to lose proton in vacuo, and ∆G acidity is associated with the ability of the molecule to lose a proton in solution. Both ∆H acidity and ∆G acidity values are reported in Table S3; they are generally higher than both BDE and the IP values (Table 2), indicating that the metal chelation mechanism is less preferred. Although this mechanism is minor with respect to radical-scavenging mechanisms, it is interesting to investigate how it occurs by determining the preferred complexes of Cu(II) with the phenolic acid amides. In this study, CaD molecule has been selected for the investigation because it combines the catecholic units on both aromatic amine and phenolic acid moieties.

Stability of the CaD···Cu2+ and the preferred binding sites for the Cu2+ ion

The calculated CaD···Cu2+ complexes are shown in Fig. 5. The relative energies (ΔE) and binding energies (ΔE binding) for the complexes, CaD···Cu bond distances, natural atomic charge Cu ion and the spin density of Cu ion in the complexes are reported in Table 3. The complexes are named in such a way that the name includes the original name of the starting structure (CaD), the symbol Cu for the cation and the Roman numbers indicating the possible site for Cu chelation on the CaD molecule. In principle, there are six possible sites on CaD for the Cu(II) chelation: two catechol moieties, two aromatic rings, the C=O group and the C=C double bond in the aliphatic chain. The coordination to these sites can give rise to six different complexes shown in Fig. 5.

Fig. 5
figure 5

Optimised structures and spin density distribution on the CaD···Cu complexes, B3LYP/6-311+G(d,p) results in vacuo. Representative CaD···Cu bond distances (Å) are indicated in blue colour

Full size image
Table 3 CaD···Cu bond distance, relative energies (ΔE, kcal/mol), interaction energy (ΔE inter, kcal/mol), spin density and charge on the Cu ion (obtained from NPA analysis) for the complexes of CaD and Cu ion, B3LYP/6-311+G(3df,2p)//B3LYP/6-311+G(d,p) results in vacuo
Full size table

In the lowest-energy complex, Cu(II) ion tri-dentated to CaD through both the oxygen atom of the C=O carbonyl group (1.869 Å) and the aromatic ring X through the atoms C1′ and C6′. The second lowest-energy complex (CaD···Cu-IV) is 14.247 kcal/mol above the lowest-energy complex structure. In this structure, the Cu ion is bi-dentated to CaD through the O atom of the C=O carbonyl group and the H7 atom. The rest of the complexes have relative energies >19 kcal/mol, which is an indication that these configurations are not preferred. The highest-energy complexes are those in which the Cu ion is coordinated through the benzene ring of each of the catechol unit, with the Cu ion preferring to bind through the X ring than the Y ring. The relative energy values of the last four complexes indicates that Cu(II) has poor chelating power towards both the catechol group and the benzene ring.

Because of the acidic nature of the OH groups in these compounds, it is possible to consider that these compounds may exist in some anionic form at physiological pH values. It was therefore considered interesting to study the interaction between the Cu(II) ion and some of the anionic forms of CaD. The Cu(II) cation may attack not only the deprotonation sites but also the other positions that are electron-rich centres. The investigated structures are shown in Fig. 6, and their relative energies and binding energies are reported in Table 3. The results of the study show that the Cu ion prefers to bind simultaneously to both the carbonyl O atom and the aromatic ring Y (through the π electrons of the aromatic ring). All other configurations of (CaD···Cu)+ have relatively high energy to be considered populated.

Fig. 6
figure 6

Optimised structures and spin density distribution on the CaD···Cu complexes for the deprotonated ligand, B3LYP/6-311+G(d,p) results in vacuo. Representative CaD···Cu bond distances (Å) are indicated in blue colour

Full size image

The binding energy (ΔE′binding), at different chelation sites, informs about the metal ion affinity (MIA) and also the preferred site on CaD for interaction with the metal cation. The trends in the ΔE′binding follow the same trend as the stabilisation energies. In this way, the MIA is highest when the Cu2+ cation is simultaneously chelated on the π system of aromatic ring Y and the lone pair of electrons of the O atom of the carbonyl group. A comparison of the ΔE′binding between the neutral complexes and the deprotonated complexes suggests that ΔE′binding is higher among the complexes with the deprotonated CaD-ai than those with the neutral CaD.

AIM analysis, NPA charges and spin density in the Cu2+···ligand complexes

The topology of the electron density distribution ρ(r), its first derivative ∇ρ(r) and second derivative (Laplacian) ∇ρ(r) provides information about the electronic structure of a molecule. Both electron density and electron energy density (H), obtained from the AIM analysis, at the different bond critical points (BCP), are often utilised to characterise and determine the strength of a given bond [80–82]. The relation between the electron density at BCP and the bond strength has been found for strong bonds (e.g. covalent bonds), and for weak interactions, such as metal–ligand interactions and hydrogen bonding [83–86]. Also, the values H and its components, the kinetic (G) and potential (V) electron energy densities provide valuable information on the nature of the chemical bond. For the strongest bonds, such as covalent bonds, V is usually negative and relatively large, G is positive, and H is negative. In such systems, both interacting atoms are sharing electrons, which are considerably localised in the internuclear space between the atoms of interest. The potential energy density estimated in BCP has large values because the electrons are relatively stable energetically in the internuclear region. Strong bonds are characterised by negative values of ∇2 ρ and are referred to as shared interactions. Weak bonds, such as H-bonding interactions, are characterised as closed-shell interactions. In this type of bonding, H has positive value and is close to zero, G has positive value, and V has negative value. However, the positive value of G predominates over the negative value of V, since the electrons are energetically less stable in the region between two closed-shell systems. Closed-shell interactions are also accompanied by positive values of ∇2 ρ estimated at the corresponding BCP, around which the electron density is depleted. The metal–ligand interaction has characteristics that represent the mix of shared and closed-shell interactions. For instance, the value of H is usually negative and close to zero, as found for shared interactions, but with a positive value of ∇2 ρ, as found for closed-shell interactions [87].

The bond critical point data for the studied CaD···Cu complexes are reported in Table 4. The largest value of ρ corresponds to Cu2+···O9 bond, indicating that this is the strongest site at which the metal ion interacts with the ligand. Therefore, the strongest inter-atomic interaction corresponds to the situation in which the M2+ ion is bonded to the sp 2 O atom. This result is in agreement with the fact that the sp 2 O has the highest available lone pair of electrons than other electron donor centres present in the molecule.

Table 4 Bond critical point data for the CaD···Cu complexes, result of the study in vacuo
Full size table

The value of H is always negative and close to zero with a range of 0.009–0.056, indicating closed-shell interaction. The ratio |V|/G is utilised to characterise a chemical bond; interactions with |V|/G < 1 are characteristic of closed-shell interactions; interactions with |V|/G > 2 are typically shared interactions; and interactions with 1 < |V|/G < 2 represent interactions of intermediate character [88, 89]. The results reported in Table 4 show that the CaD···Cu interactions have a range of 1.2–1.61, what is indicative of intermediate between covalent and ionic character. The values of ∇2 ρ are positive, indicative of a predominantly closed-shell (ionic) interaction. This means that although the interactions are intermediate between covalent and ionic character, they are more inclined towards ionic character.

The free Cu(II) cation has a charge of +2. The charge on the Cu(II) ion in the complexes with CaD (reported in Table 3) has the range of 0.852–1.057e. These results suggest that on interaction with CaD, the Cu(II) ion is reduced to Cu(I) ion. In the process, CaD is oxidised so that the complexes that are obtained with the copper ion are of the form [CaD···Cu]+. The spin density of Cu(II) ions in the isolated state is 1 and for the neutral CaD ligand is zero. The atomic spin density of Cu ion in the complex has the range of 0.0003–0.1337 for the situation in which the Cu ion binds to the neutral CaD molecule and a range of 0.0085–0.2750 for a situation in which the Cu ion binds to the anionic species. In both situation, therefore, the Cu(II) tends to have significantly small spin density with respect to the isolated Cu(II) ion.

Metal chelation mechanism in the presence of water solvent

The polarisation effects of the solvent molecules, on both the geometries and energies of the complexes of CaD with Cu2+ ions, were taken into account through using the PCM, SMD model. The relative energy (∆E soln), binding energy (∆E inter,soln), charges and spin densities on the Cu2+ ions are reported in Table S4; the bond critical point data are reported in Table S5. The geometries of the complexes are similar between the results in vacuo and the results in water solution. However, the CaD···Cu2+ bond distance is longer in water solution than in vacuo, indicating that the interaction between the ligand and the M2+ ion is stronger in vacuo than in solution. The relative energy order of the complexes in water solution is not significantly different from the order of the relative energies of the complexes in vacuo.

All the complexes in solution have ∆E′binding,soln > 0, indicating that Cu2+ has a tendency to bind to CaD. Among the lowest-energy complexes, the binding energy decreases significantly with respect to the binding energy obtained in vacuo, what suggests that CaD has less affinity for the Cu2+ ion in water than in vacuo. The decreased affinity of the ligand towards the metal ion may be related to the presence of polarisation effects in solution, which are absent in vacuo [29].

The spin density values have a range of 0.2424–0.7472. In comparison with the results in vacuo, the spin density on the Cu2+ ion is less depleted in solution than in vacuo. The charge on the Cu2+ ion decreases less (1.241−1.610e) in comparison with the results in vacuo, indicating that the positive charge is less dissipated in solution. An analysis of the bond critical point data suggests that in solution, the O9···Cu bond is weak in solution than in vacuo because of having smaller value of ρ in solution than in vacuo. However, the trend in all other properties is the same as that in vacuo, suggesting intermediate type of bonding with greater inclination towards ionic character.

Conclusion

Theoretical studies on the antioxidant properties phenolic acid amides were performed to investigate factors that contribute to their conformational preferences and to elucidate their antioxidant properties and mechanisms. The antioxidant activity has been considered through the ability of phenol acid amides to scavenge free radical species and through their ability to chelate metal ions.

The radical-scavenging activity has been investigated through the hydrogen atom transfer (HAT), the electron transfer mechanisms (ET), the single-electron transfer followed by proton transfer (SET-PT) and the sequential proton loss electron transfer mechanisms (SPLET). In general, the antioxidant activity through the HAT mechanism depends on the arrangement of hydroxyl substituents on the aromatic ring. The abstraction of the H atom from the phenolic OH group is preferred to the abstraction of the H atom from the NH group. Moreover, the abstraction of the H atom from the para OH group in a catecholic unit contributes the strongest strong free-radical-scavenging properties for the phenolic acid amides. Based on the calculated thermodynamic parameters, phenolic acid amides can be considered to be better radical scavenger than the isolated aromatic amines and phenolic acids.

The metal chelation properties of phenolic acid amides depend on the affinity of the Cu ion for the different binding sites.

The study suggests that the preferred complexation site for phenolic acid amides is the simultaneous binding to the carbonyl O atom and the benzene ring of the aromatic amide. An estimated metal ion affinity (MIA) is influenced by the nature of the ligand and the media; it is higher when the ligand associated with the metal ion is an anion than when it is a neutral molecule, and it is significantly lower in aqueous solution than in vacuo. The interaction of CaD with the Cu ion shows the ability of CaD to reduce the metal ions from a +2 state to a +1 state. An AIM analysis suggests that the CaD···Cu interactions correspond to the mixed type interactions with greater inclination towards ionic character.