Antioxidant properties comparative study of natural hydroxycinnamic acids and structurally modified derivatives: Computational insights

Computational and Theoretical Chemistry xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Computational and Theoretical Chemistry journal homepage: www.elsevier.com/locate/comptc

Antioxidant properties comparative study of natural hydroxycinnamic acids and structurally modified derivatives: Computational insights Gloria Mazzone a,⇑, Nino Russo b, Marirosa Toscano b a b

Dipartimento di Ingegneria Informatica, Modellistica, Elettronica e Sistemistica, Università della Calabria, I-87036 Arcavacata di Rende, Italy Dipartimento di Chimica e Tecnologie Chimiche, Università della Calabria, I-87036 Arcavacata di Rende, Italy

a r t i c l e

i n f o

Article history: Received 11 September 2015 Received in revised form 7 October 2015 Accepted 8 October 2015 Available online xxxx Keywords: Hydroxycinnamic acid derivatives DFT Antioxidant mechanisms UV–Vis characterization

a b s t r a c t Density functional theory (DFT) and time-dependent formulation of DFT (TDDFT) have been used to explore the antioxidant and absorption properties, respectively of naturally occurring cinnamic acids, caffeic and ferulic acids, and some derivatives recently synthesized from a structural modification of the ethylenic spacer between the aromatic ring and the carboxylic functionality. The main mechanisms proposed in the literature for the antioxidant action of polyphenols as radical scavengers, that are hydrogen atom transfer (HAT), electron transfer followed by proton transfer (SET-PT), and sequential proton loss electron transfer (SPLET), were discussed in details. From the outcomes the HAT mechanism results to be the most probable one for the antioxidant action of this class of compounds. The simulated UV–Vis spectra are characterized by a broad band centered around 340 and 380 nm for naturally occurring and synthesized compounds, respectively, generated by a H ? L electronic transition. The absorption spectra of natural antioxidants are in good agreement with the experimental counterpart, supporting the reliability of the spectra computed at TDDFT level of theory also for the derivatives. One of the derivative has been identified as the most promising candidate as antioxidant. Accordingly, our calculations encourage the synthesis of derivatives arising from ad-hoc structural modifications which could improve the antioxidant properties. Ó 2015 Published by Elsevier B.V.

1. Introduction In the last twenty years there has been a growing attention on the role of free radicals in biology, as they are able to damage the most important macromolecules of the human organism. They are mainly produced from normal essential metabolic processes in the aerobic living organisms, usually as reactive oxygen species (ROS), reactive nitrogen species (RNS) and reactive sulfur species (RSS). Hence they are naturally present in the human organism, and exert a beneficial effect at low/moderate concentrations in defense against infectious agents and in the function of a number of cellular signaling systems. ROS can be produced also by extracellular stress, such as irradiation, air pollutants, and exposure to toxic agents [1,2]. In a normal, healthy organism or human body, the generation of prooxidants in the form of ROS and RNS are effectively kept in check by the various levels of antioxidant defense, which guarantee a balance between produced and neutralized highly reactive species. However, when this ratio is threatened such reactive species can attack proteins, lipids, DNA, RNA and sug⇑ Corresponding author. E-mail address: [email protected] (G. Mazzone).

ars causing the so called oxidative stress of body cells [2,3], which is implicated in a several human diseases [2,4,5]. In recent years, the possible role of nutrition in prevention of human diseases has taken a leading role. An external supplement of antioxidants has been indicated as a suitable way to maintain the concentration of free radicals as low as possible and then avoiding the oxidative stress [6]. Indeed, in the last years several studies have reported the beneficial effects of antioxidants on human health, including anticancer, anti-inflammatory, antibacterial, antiviral, cardioprotective and neuroprotective properties [7– 12], just associated with their ability of preventing oxidative stress (OS). There is also evidence supporting their potential therapeutic use on diseases, such as diabetes, osteoporosis, arthritis, and cataract [13–16]. Their action is exerted by reacting with highly reactive species forming more stable and innocuous radicals for cells with respect to the inhibited ones, or turning off the radical chain reactions, helping to prevent the attack of such a species to biological macromolecules, limiting their damages. It is widely established that the most essential structural characteristics which provide effective antioxidant activity are the presence of phenolic OH groups, which enhance the ability of such a molecules to quench the free radicals. Plants produce thousands

http://dx.doi.org/10.1016/j.comptc.2015.10.011 2210-271X/Ó 2015 Published by Elsevier B.V.

Please cite this article in press as: G. Mazzone et al., Antioxidant properties comparative study of natural hydroxycinnamic acids and structurally modified derivatives: Computational insights, Comput. Theoret. Chem. (2015), http://dx.doi.org/10.1016/j.comptc.2015.10.011

2

G. Mazzone et al. / Computational and Theoretical Chemistry xxx (2015) xxx–xxx

of phenolic compounds as secondary metabolites [17] and consequently they are naturally present in fruits, vegetables and cereals. Hence, there are several naturally occurring compounds that present the structural characteristics needed to exert the role of free radical scavenger. Among them phenolic acids, benzoic and cinnamic acid derivatives, have been found to have very good antioxidant properties [18,19]. Phenolic compounds (ArOH) can exhert the free radical scavengin activity by following essentially three important mechanisms (see Scheme 1): (i) the hydrogen atom transfer (HAT) in which the radical abstract the hydrogen from the antioxidant molecule in one single step; (ii) the single-electron transfer followed by proton transfer (SET-PT), which takes place in two steps: first the radical cation ArOHþ is formed via electron transfer from the antioxidant to the free radical, and then it deprotonates yielding the ArO radical, followed by ROH formation; (iii) the sequential proton loss electron transfer (SPLET) takes place through two consecutive steps that are deprotonation of the phenolic compound and the consecutive electron transfer from the phenoxide anion to RO to form the phenoxyl radical. On the basis of theoretical calculations, the HAT mechanism has been proposed as a key reaction mechanism in the antioxidant activity of several compounds, not only polyphenols [20] and derivatives [21–27], but also nonphenolic compounds [28–30]. Anyhow, it is noteworthy that the other reaction mechanisms yield to the same products as HAT, and properly discriminating among them could be a complex task. Other important information on the working mechanism of antioxidants can come from the knowledge of the frontier orbital energies, EHOMO and ELUMO, and distributions. The lower HOMO energy is responsible for the minor ability of a molecule to donate a proton. On the contrary, an higher HOMO energy implies that the molecule is a good electron donor. Since the H-abstraction reaction

H+ RO· (ii) ArOH

ArOH+ · + RO-

-e(i)

RO·

ArO· + ROH

involves the electron transfer, the HOMO composition of a phenolic compound can give a qualitative idea of which sites are easily attacked by free radicals and other reactive agents. From the difference between HOMO and LUMO energies, indication about chemical activity of the molecule can be extracted. The lower DEH–L is connected with lower antioxidant activity of the molecule [31]. In spite of the enormous number of compounds known to exhert antioxidant properties, the synthesis of molecules that combine structural characteristics, and then beneficial effects, introducing modifications to naturally occurring antioxidants is becoming an increasingly robust strategy to achieve compounds with enhanced biological activities, as well as therapeutic potential [32–36]. Chavarria et al. [36] have exploited the significantly high antioxidative efficiency of the hydroxycinnamic acids [37], characterized by the presence of the CH@CHACOOH group, to synthesize derivatives compounds with improved antioxidant activity. Therefore, aiming at enhancing the absorption, distribution, metabolism and excretion (ADME), they have modeled the cinnamic acid scaffold in such a way to extend the ethylenic spacer between the aromatic ring and the carboxylic acid functionality (see Scheme 2). In this work, they have evaluated several properties, such as radical scavenging activity, redox properties, iron(II) chelating activity, octanol/water partition coefficient, cytotoxicity and neuroprotection in a cellular model, of caffeic (CA) and ferulic acids (FA) and their synthesized derivatives (1–3). The obtained results highlight the potential of the derivatives as drugs for the treatment of oxidative stress associated diseases, especially as neuroprotective agents. As the theoretical evaluation of intrinsic properties of such a compounds can help to gain more insight on their preferred action mechanism, we report here a systematic DFT-based study on such a compounds. In our investigation, the relationship between the experimental radical scavenging activity of cinnamic and ferulic acids, CA and FA, and derivatives 1–3 has been interpreted in terms of some key thermochemical parameters (BDE, IP, PDE, PA and ETE). The comparison between the scanvenging activity of the natural antioxidants and their recently synthesized derivatives [36] have been also reported. In addition, the simulation of UV–Vis absorption spectra has been provided in order to help their identification. 2. Computational details

(iii) -H+

H+

ArO- + RO· e-

Scheme 1. Schematic representation of (i) HAT, (ii) SET-PT and (iii) SPLET mechanisms.

All the electronic calculations have been performed using the Gaussian 09 computational package [38]. The geometries of all the investigated compounds, including radicals, radical cations, and anions, have been fully optimized employing the hybrid functional M05-2X [39] combined with the standard 6-31+G⁄⁄ basis set. Such a functional has been also suggested, together with other

Natural compounds

Synthesized derivatives O

O R1

O

OH H

R1

OH

R2 1: R1=OCH3, R2=OH 2: R1=OH, R2=OH 3: R1=OCH3, R2=OCH3

CA: R1=OH FA: R1=OCH3

Scheme 2.

Please cite this article in press as: G. Mazzone et al., Antioxidant properties comparative study of natural hydroxycinnamic acids and structurally modified derivatives: Computational insights, Comput. Theoret. Chem. (2015), http://dx.doi.org/10.1016/j.comptc.2015.10.011

3

G. Mazzone et al. / Computational and Theoretical Chemistry xxx (2015) xxx–xxx

PDE ¼ HðArO Þ þ HðHþ Þ  HðArOHþ Þ

The Cartesian coordinates of the optimized structures of the investigated compounds (see Scheme2) are reported in the supporting material (S1). Due to the presence of the double bonds which link the aromatic ring and the carboxylic acid functionality, the investigated hydroxycinnamic acids present a completely planar structure. In the most stable structure of all these compounds, the carboxylic functionality is arranged so that the hydrogen of the hydroxyl group is oriented in a cis fashion with respect to the oxygen atom of the carbonyl one. Derivatives 1 and 2 differ from ferulic and caffeic acids, respectively, for the presence of a second double bond as spacer between the aromatic ring and the carboxylic acid functionality which increases the conjugation effects [36]. In the optimized structures the hydroxyl groups are oriented in such a way to maximize Hbond-like interactions. Therefore, in FA and 1 the hydrogen atom of the 4OH is oriented to establish a hydrogen bond with the oxygen atom of the methoxy group in position 3. While, the most stable conformation of CA and compound 2, shows an intramolecular hydrogen bond at the catechol level. Compound 3 is characterized by two methoxy groups in positions 3 and 4 and it is experimentally synthesized [36] starting from the 3,4-dimethoxycinnamic acid, which is not investigated here.

PA ¼ HðArO Þ þ HðHþ Þ  HðArOHÞ

3.2. Descriptors of antioxidant properties

hybrid functionals, for kinetic data calculations in radical-molecule reactions, including scavenging ones [40]. The solvation effects have been computed by using the SMD solvation model [41]. The dielectric constant of 24.8 has been used in order to simulate the solvent (ethanol) used in the experimental measurements. All of the optimized structures have been confirmed to be real minima by harmonic vibrational frequencies calculation. The unrestricted open-shell approach has been used for radical species. In these cases, no spin contamination has been found, being the ‹S2› value about 0.750. Refined energies have been obtained by performing singlepoint calculations on the optimized geometries at the same level of theory and employing a larger standard basis set, 6-311++G⁄⁄, for all the atoms, taking into account also the impact of the solvent. BDE, IP, PDE, PA and ETE have been calculated in ethanol at 298.15 K following the procedure already applied on similar systems [42–46], according to the following expressions:

BDE ¼ HðArO Þ þ HðH Þ  HðArOHÞ IP ¼ HðArOHþ Þ þ Hðe Þ  HðArOHÞ







ETE ¼ HðArO Þ þ Hðe Þ  HðArO Þ The calculated H(H+) and H(e) enthalpy values are 1.48 and 0.75 kcal/mol, respectively [47]. The enthalpies of the reactions sketched in Scheme 1 have been calculated in ethanol at 298.15 K using the following expressions:

DHBDE ¼ HðArO Þ þ HðROHÞ  HðArOHÞ  HðRO Þ DHIP ¼ HðArOHþ Þ þ HðRO Þ  HðArOHÞ  HðRO Þ DHPDE ¼ HðArO Þ þ HðROHÞ  HðArOHþ Þ  HðRO Þ DHPA ¼ HðArO Þ þ HðROHÞ  HðArOHÞ  HðRO Þ DHETE ¼ HðArO Þ þ HðRO Þ  HðArO Þ  HðRO Þ HðRO Þ, HðROHÞ and HðRO Þ represent the enthalpies of the species arising from hydroxyl OH, peroxyl OOH, methoxyl OCH3 and methyl peroxyl OOCH3 radicals. Absorption spectra were computed as vertical electronic excitations from the minima of the ground-state structures by using time-dependent density functional response theory (TDDFT) [48] as implemented in the Gaussian 09 code. These calculations were carried out in ethanol medium, using the standard 6-31+G⁄⁄ basis set and M06, M05-2X [49], wB97XD [50,51], PBE0 [52,53] and B3LYP [54,55] exchange–correlation (XC) functionals. Computations of single-point spin densities have been performed using the above mentioned protocol for the most stable open shell species. This computational protocol has already been successfully applied to investigate the antioxidant properties of a large series of polyphenols [41–43,56–59].

The antioxidant activity parameters computed for all the species, including radicals and anions, of the investigated compounds are reported in Table 1. Results show that the energies put into play in the three considered mechanisms are of a very different magnitude, which can provide a primary indication of which mechanism is favored over another. The BDE represents the best reliable thermochemical parameter to describe the HAT mechanism, which involves the transferring of an H atom from a hydroxyl group of the antioxidant molecule to the free radical. The weakest OAH bond (lowest BDE) is expected to lead to the most likely reaction and then to the greatest antioxidant activity. Looking at the BDE values of CA and its derivative 2, it is evident that the radicalization of the hydroxyl group in position 4 leads to the most stable radical formation. Analyzing the BDE values of natural compounds, the CA results to be slightly more active than the FA in transferring one hydrogen atom directly to the free radical, albeit the BDE difference is of only 2 kcal/mol. Both model compounds, 1 and 2, exhibit BDEs lower than those computed for correspondent naturally occurring antioxidants, ferulic and caffeic acids, respectively.

Table 1 M05-2X/6-311++G⁄⁄ bond dissociation enthalpies (BDE), difference in radicals stability (DE), OAH proton dissociation enthalpies (PDE), proton affinities (PA), electron transfer enthalpies (ETE) and adiabatic ionization potentials (IP) are reported in kcal/mol. Compound

Site

BDE

DE

IP

PDE

PA

ETE

FA

10 OH 4OH

97.7 82.0

15.7 0.0

136.4

276.6 261.0

283.2 289.4

129.9 108.0

CA

10 OH 3OH 4OH

99.6 81.8 80.0

17.8 1.8 0.0

138.3

276.7 259.0 257.1

283.8 288.1 285.3

131.2 109.2 110.1

3.1. Molecular geometries

1

10 OH 4OH

93.6 79.6

14.0 0.0

131.4

277.6 263.6

283.4 290.3

125.5 104.7

Detailed knowledge of the structural and electronic characteristics of antioxidant compounds is of crucial importance in elucidating their scavenging behavior. The phenolic acids are characterized by one or more phenolic function and a carboxylic acid group.

2

10 OH 3OH 4OH

94.9 81.0 77.5

17.4 3.5 0.0

132.6

277.8 263.9 260.3

284.0 288.4 286.0

126.4 108.1 106.9

3

10 OH

95.4

–

131.9

278.9

283.7

127.1

3. Results and discussion

Please cite this article in press as: G. Mazzone et al., Antioxidant properties comparative study of natural hydroxycinnamic acids and structurally modified derivatives: Computational insights, Comput. Theoret. Chem. (2015), http://dx.doi.org/10.1016/j.comptc.2015.10.011

4

G. Mazzone et al. / Computational and Theoretical Chemistry xxx (2015) xxx–xxx

Among the investigated compounds, the derivative of caffeic acid, 2, with the lowest BDE value (77.5 kcal/mol), seems to be the best candidate to act as scavenger following the HAT mechanism, confirming an enhancement of antioxidant activity in the model compound with respect to the natural one (CA, BDE: 80.0 kcal/mol). All the compounds, except 3, are characterized by the presence of one or two hydroxyl groups directly bound to the aromatic ring. As mentioned before this is an essential condition to promotes the delocalization of the unpaired electron after the radicalization. The only possibility for system 3 to quench the free radicals according to the HAT mechanism is represented by the radicalization of the hydroxyl of the carboxylic group. Thus, although highly unlikely, the abstraction of the H atom from such group (10 OH) has been analyzed for 3 and all other compounds. As expected, the BDEs computed for this radicalization are significantly higher than those calculated for the phenolic hydroxyl ones. This means that the acidity of hydrogen atom obstructs the transferring of the hydrogen atom as a single entity. The possibility of significant conjugation effects and the presence of the ortho-diphenolic moiety on the antioxidant molecules [60], significantly contribute to stabilize their radical species. This can be confirmed by the spin density analysis. According to Parkinson, the more delocalized the spin density in the radical, the more easily is formed the radical [61]. In Fig. 1 are reported the spin density values and distribution of cinnamic acids and derivatives obtained at M05-2X/6-31+G⁄⁄ level of theory. Comparing the natural antioxidants and their derivatives, it appears clear that the extension of the ethylenic spacer between the aromatic ring and the carboxylic acid functionality contributes to enhance the conjugation effects, which in turn confirms the important role of the di-ethylenic moiety to evoke higher antioxidant activity of derivatives with respect to natural compounds [36]. Indeed, from the computations concerning CA and compound 2, or FA and compound 1, the unpaired electron is visibly delocalized also on a C atom of the second double bond. This is reflected on the BDEs values, which result smaller for derivatives than for natural antioxidants, as reported above. The stepwise mechanisms, SET-PT and SPLET, are both related to the predisposition of molecules to give electrons. In the former mechanism, an electron is transferred from the antioxidant to the free radical leading to the radical cation ArOHþ formation, which subsequently deprotonates. Therefore, adiabatic ionization potential (IP) and proton dissociation enthalpy (PDE) are the most relevant properties in analyzing the feasibility of this mechanism. In the latter mechanism, the electron transfer is preceded by a proton

one, thus proton affinity (PA) and electron transfer enthalpy (ETE) are useful to establish trends regarding the viability of such a mechanism. Analyzing the IP and PDE values reported in Table 1, the energy amount required to form the radical cation is significantly lower than that needed to accomplish the SET-PT second step. Thus, the deprotonation of radical cation ArOHþ is the most probable step that limits the reaction rate of such a mechanism in polar solvents. Molecules with low adiabatic ionization potential (IP) values are more susceptible to ionization and have a stronger antioxidant property. The trend of ionization potentials appears to be different with respect to that of bond dissociation energies, indeed, among the investigated compound, the derivative 1 show the lowest IP, confirming again the major activity of derivatives with respect to natural compounds. Anyhow, the energies required to accomplish the whole SET-PT reaction suggest that this mechanism is not preferred by neither of the examined cinnamic acids. From the other hand, the formation of the phenoxide anion in the first step of SPLET mechanism (PA values in Table 1) requires much more energy to occur than the electron transfer from the phenoxide anion to the free radical (ETEs in Table 1), indicating that the first step should be the slowest one for this mechanism. Also in this case, the energies put into play seem to exclude that this mechanism can occur. From the analysis of the antioxidant descriptors, as both SET-PT and SPLET reaction mechanisms involve a greater energy expense with respect to the HAT one, the latter seems to be the reaction path preferred by this class of compounds in polar solvents. Anyhow, it is noteworthy that each reaction step involved in the analyzed mechanisms should be fully explored to unequivocally determine the reasons for which such mechanism is the favoured one. This escapes the main goal of the present work which aims to provide a comparison on the antioxidant ability of some natural compounds and their derivatives by evaluating the known descriptors of antioxidant properties. However, the reaction enthalpies associated to the each step involved in the three antioxidative mechanisms (see Scheme 1) can contribute to identify the most probable reaction path, taking into account also the influence of a specific radical [62,63]. If the reaction between them results to be exothermic, the formed radical is clearly more stable than the starting one, hence the reaction is favored. If, instead, the reaction is endothermic, the path is obviously unlikely to occur. Lower reaction enthalpies can be associated to a greater ability of phenolic compounds to act as antioxidant following a particular

0.233 0.304

0.167

0.273

0.267 0.257

0.251

0.381

0.183

0.242 0.312

0.378

0.271

0.196

0.368 0.209

0.258 0.253

0.254

0.186

0.365 0.152

Fig. 1. Spin density values and distribution for the 4O-radical computed for CA, FA, 1, and 2 compounds at M05-2X/6-31+G⁄⁄ level of theory. The plots were accomplished with isodensity value of 4  103 a.u and only spin density values greater than 0.1 are reported.

Please cite this article in press as: G. Mazzone et al., Antioxidant properties comparative study of natural hydroxycinnamic acids and structurally modified derivatives: Computational insights, Comput. Theoret. Chem. (2015), http://dx.doi.org/10.1016/j.comptc.2015.10.011

5

G. Mazzone et al. / Computational and Theoretical Chemistry xxx (2015) xxx–xxx

mechanism. For this class of compounds the reaction enthalpies that have to be taken into account are DHBDE (HAT), DHPDE (SETPT) and DHETE (SPLET). For this purpose, hydroxyl OH, peroxyl OOH, methoxyl OCH3 and methyl peroxyl OOCH3 radicals have been selected. In Table 2 are collected the enthalpies of reaction of the investigated compounds with these radicals. From data in Table 2 it appears clear that DHIP and DHPDE are too high to consider the SET-PT a plausible mechanism, as neither of the investigated compounds reacts exothermically with the con-

Table 2 M05-2X/6-311++G⁄⁄ reaction enthalpies (kcal/mol) for the reaction of CA, FA and 1–3 compounds with OH, OOH, OCH3 and OOCH3 radicals. Compound

Radical

DHBDE

DHIP

DHPDE

DHPA

DHETE

FA



OH OOH OCH3  OOCH3

36.7 3.3 21.9 2.4

11.9 34.8 28.1 36.1

87.1 97.6 85.7 97.2

20.2 9.7 21.5 10.1

16.6 6.4 0.3 7.6



OH OOH  OCH3  OOCH3

38.7 5.3 23.9 4.4

13.7 36.6 29.9 37.9

85.1 95.6 83.7 95.2

24.2 13.7 25.6 14.1

14.5 8.4 1.7 9.7



OH OOH OCH3  OOCH3

39.2 5.7 24.3 4.8

6.8 29.7 23.0 31.0

84.7 95.2 83.3 94.8

19.3 8.8 20.6 9.2

19.9 3.1 3.7 4.3



OH OOH  OCH3  OOCH3

41.2 7.8 26.4 6.9

8.0 30.9 24.2 32.2

82.6 93.1 81.2 92.7

23.6 13.0 24.9 13.4

17.7 5.3 1.5 6.5



23.3 10.2 8.4 11.0

7.4 30.3 23.6 31.6

100.5 11.1 99.2 110.7

25.8 15.3 27.2 15.7

2.5 25.5 18.7 26.7

 

CA



1

 

2



3

OH OOH OCH3  OOCH3  

sidered radical species. For the other paths, both positive and negative reaction enthalpies have been found. In particular, the reaction of all the radicals with either natural or synthesized compounds is more exothermic when they follow the HAT mechanism (see DHBDE values) rather than the SPLET one (see DHETE values). This can be verified observing the behavior of CA and its derivative 2. From the data concerning the HAT mechanism (DHBDE values), regardless the nature of the free radical, the derivative 2 is confirmed as the most active compound in scavenging reactions. In addition, from data in Table 2 it emerges that the reaction with hydroxyl radical entails the major gain of energy whatever the path followed (see DHBDE, DHPDE and DHETE values), supporting its propensity to react with almost any molecule in its vicinity and then the need to prevent its action toward the most important biological macromolecules. Other information about the antioxidant activity can be extracted from the energies of the frontier orbitals. In Fig. 2 are reported the energy diagram of the four Gouterman’s orbitals for all the examined molecules, together with a graphical representation of the aforementioned orbitals. The HOMO energy is an important electronic parameter for describing the propensity of a molecule to participate in electron transfer reactions. A molecule with low values of HOMO energy has a weak electron donating ability. Otherwise, a higher HOMO energy implies that the molecule is a good electron-donor [64,65]. Looking at Fig. 2, the molecule that exhibit higher HOMO energy is the derivative 1, which could be indicated as the best electron donor among the studied compounds. This evidence confirms the trend of adiabatic IPs analyzed above, from which the caffeic acid appears as the worst candidate to participate in electron transfer reactions. Indeed, also analyzing the H–L energy gaps, the electron promotion from HOMO to LUMO in 1 is relatively easier to occur than in the other investigated molecules. Therefore, hypothetically,

6 4 2 0

L+1 L

0.66

0.73

-1.01

-1.01

0.74

0.69

0.72

-1.27

-1.27

-1.34

E (eV)

-2 -4

6.33

6.26

5.68

5.74

5.86

-6 -8

H -7.34 H-1 -8.35

-7.27

-6.95

-7.01

-7.20

-8.33

-8.14

-8.18

-8.09

FA

1

2

3

-10 -12 -14

CA

Fig. 2. Energy diagram of the four ‘‘Gouterman’s orbitals” for CA, FA and derivatives 1–3. Energy gaps (eV) between H and L are shown and indicated by solid arrows. Graphical representation of MOs (isodensity value of 4 104 a.u.) are sketched.

Please cite this article in press as: G. Mazzone et al., Antioxidant properties comparative study of natural hydroxycinnamic acids and structurally modified derivatives: Computational insights, Comput. Theoret. Chem. (2015), http://dx.doi.org/10.1016/j.comptc.2015.10.011

6

G. Mazzone et al. / Computational and Theoretical Chemistry xxx (2015) xxx–xxx

it could work well following the SET-PT mechanism in which the removal of the hydrogen atom occurs only after one electron is transferred. The HOMOs and LUMOs plots reported in Fig. 2 show the typical p-like molecular orbital characteristics. In all cases, while the HOMO are outspreaded on the whole system, with an implication of the carbonyl group, in the LUMO plots the OH of carboxylic group is also included, contrary to the group in position 3, which, differently from the HOMOs plot, results excluded.

1,0

1 2 3 CA FA

oscillator strength

0,8

0,6

0,4

3.3. Absorption spectra

0,2

In Table 3 are reported the outcomes of TDDFT calculations made to obtain the main transitions that define the absorption spectra of the investigated compounds. Although the experimental UV–Vis characterization for derivatives 1–3 is not available, there are in literature several studies on the absorption properties of caffeic and ferulic acids. Therefore, comparing the theoretical results with the available experimental ones, we are confident that also the computed electronic spectra for 1–3 derivatives can help the characterization of such compounds. In Fig. 3 are included the electronic spectra computed for all the studied acids at B3LYP/6-31+G⁄⁄ level of theory. All the hydroxycinnamic acids show the most intense absorption band that falls in the range of 340–390 nm, passing from natural compounds CA and FA (340 and 346 nm, respectively) to the synthesized ones 1–3 (375, 379 and 381 nm). The only contribution to this band is provided by the HOMO–LUMO (H–L) transition, which, looking at the MO’s plot reported in Fig. 2, occurs in all cases between states with p and p⁄ character, respectively. Analyzing the data reported Table 3 for the natural compounds, it appears that computed absorption spectra, obtained by using different exchange and correlation functionals, are all in satisfactory agreement with the available experimental ones. Anyhow, the dif-

0,0 150

200

250

300

350

400

450

500

550

wavelenght (nm) Fig. 3. Absorption spectra of CA, FA and derivatives 1–3 in ethanol, computed at TD-B3LYP/6-31+G⁄⁄ level of theory.

ferences among the employed functionals will be highlighted and commented. At B3LYP level of theory the computed maximum absorption wavelengths result to be blue shifted of 27 and 34 nm for CA and FA, respectively with respect to the experimental counterpart. B3LYP, PBE0 and M06 functionals overestimate the excitation energy (in terms of wavelength) computed for the H ? L transition of caffeic and ferulic acids. Conversely, M05-2X and wB97XD underestimate such energies of 14 and 12 nm, respectively, in the case of CA and 10 and 7 nm in the case of FA. However, the latter XC functionals are those that better reproduce the main absorption band of such a compounds. Focusing the attention on the CA data, the second absorption band (in the range of 250–300 nm) computed employing B3LYP

Table 3 Main excitation energies (DE), oscillator strength (f) and MO contribution (%) computed for all the investigated compounds in ethanol by using different exchange and correlation functionals and 6-31+G⁄⁄ basis set. All electronic states belong to 1A. The available experimental values are also included. MO contribution

a b

B3LYP

M05-2X

PBE0

M06

wB97XD

DE eV, nm

f

DE eV, nm

f

DE eV, nm

f

DE eV, nm

f

DE eV, nm

f

exp nm

CA H ? L (97%) H-1 ? L (85%) H ? L + 1 (64%)

3.64, 340 4.23, 293 5.04, 246

0.514 0.207 0.317

4.15, 299 4.83, 257 5.64, 220

0.725 0.048 0.467

3.76, 329 4.38, 283 5.18, 239

0.559 0.178 0.344

3.71, 334 4.29, 289

0.549 0.135

4.12, 301 4.71, 263 5.57, 223

0.717 0.048 0.468

313a 300 214

FA H ? L (97%) H-1 ? L (86%) H ? L + 1 (78%)

3.58, 346 4.22, 293 5.03, 246

0.490 0.233 0.367

4.10, 302 4.82, 257 5.61, 221

0.709 0.055 0.640

3.71, 334 4.37, 284 5.16, 240

0.536 0.202 0.397

3.65, 339 4.29, 289

0.527 0.155

4.07, 305 4.70, 264 5.53, 224

0.701 0.054 0.543

312b 287 215

1 H ? L (98%) H-1 ? L (91%) H-2 ? L (46%) H ? L + 1 (76%)

3.22, 3.96, 4.66, 4.81,

385 313 266 258

0.962 0.284 0.269 0.036

3.69, 336 4.73, 262

1.256 0.052

1.034 0.247 0.284 0.028

1.034 0.194

3.70, 335 4.66, 266

1.271 0.034

0.351

372 301 258 248

3.29, 376 4.08, 304

5.27, 235

3.34, 4.12, 4.81, 4.99,

5.22, 238

0.348

2 H ? L (98%) H-1 ? L (90%) H-2 ? L (39%) H ? L + 1 (46%)

3.27, 3.98, 4.67, 4.84,

379 311 265 256

0.996 0.234 0.253 0.022

3.73, 321 4.74, 262

1.271 0.038

3.38, 367 4.15, 299

1.064 0.202

3.34, 372 4.10, 302

3.74, 332 4.66, 266

1.285 0.024

5.31, 233

0.313

4.82, 257

0.267

5.27, 235

0.306

3 H ? L (96%) H-1 ? L (92%) H-2 ? L (57%) H ? L + 1 (42%)

3.34, 3.89, 4.77, 4.98,

371 319 260 249

0.847 0.482 0.160 0.050

3.82, 325 4.66, 266

1.269 0.133

1.280 0.113

0.273

0.952 0.412 0.176 0.043

3.82, 324 4.60, 269

5.41, 229

3.47, 4.04, 4.94, 5.15,

358 307 251 241

3.41, 363 4.00, 310 4.84, 256

1.061 0.154

0.984 0.326 0.196

Ref. [66]. Ref. [67].

Please cite this article in press as: G. Mazzone et al., Antioxidant properties comparative study of natural hydroxycinnamic acids and structurally modified derivatives: Computational insights, Comput. Theoret. Chem. (2015), http://dx.doi.org/10.1016/j.comptc.2015.10.011

G. Mazzone et al. / Computational and Theoretical Chemistry xxx (2015) xxx–xxx

functional deviates from the experimental value of only 7 nm, contrariwise to those obtained with the functionals just mentioned, by which a red-shift of 43 and 37 nm, respectively has been found. As can be seen in Fig. 3, in all cases a not intense band (with oscillator strength < 0.2) obscured by the main one, has been found. This is more evident in the data obtained with M05-2X and wB97XD functionals that detect an electronic transition with oscillator strength of only 0.048. A similar behavior can be observed for ferulic acid. In this case, M06 and PBE0 functionals overestimate such an absorption of only 3 and 2 nm, respectively. This band is mainly due to H1 ? L transition, which is again generated by a p ? p⁄ type transition. Experimental spectra for CA and FA present another absorption band at 214 and 215 nm, respectively. M05-2X and wB97XD blue shift this band of only 6 and 9 nm, respectively, showing the best agreement with experimental data. The absorption is due to a H ? L + 1 transition, that in both cases involves a virtual orbital localized on the phenol ring (see Fig. 2). The good agreement between computational results, obtained at TDDFT level of theory, and experimental absorption data obtained in this work, confirms what it is already reported in literature on simulated spectra of caffeic acid and other naturally occurring antioxidants [67–69]. Examining the data computed for derivatives 1–3, the H ? L transition has been systematically found at higher wavelengths compared to the natural compounds from which they derive. Indeed, from Fig. 3 it can be easily observed that derivatives 1–3 show a blue shift of the maximum absorption wavelength (385 vs 346 nm for CA and derivative 2 and 379 vs 340 nm for FA and derivative 1). This is clearly due to the inclusion of a double bond in the ethylenic spacer between aromatic ring and carboxylic acid functionality. Differently from naturally occurring antioxidants, the band located at lower wavelengths seems to be due to the H-2 ? L and H ? L + 1 electronic transitions, albeit not all XC functionals agree on the contribution of the former transition to this band. The derivative that absorbs at higher wavelength is the derivative 1, followed by the other derivatives and natural compounds. This result is consistent with the H–L energy gaps discussed in the previous section, which result to be the smallest one in the case of derivative 1. Considering that the basic structure of the model compounds is the same, the substituents on the aromatic ring play a pivotal role in shifting the absorption band.

4. Conclusions A comparative DFT-based study on the antioxidant properties of two naturally occurring compounds, caffeic and ferulic acids, and some synthesized derivatives have been carried out. The main mechanisms proposed in the literature for the antioxidant action of polyphenols as radical scavengers were discussed in details, and structural and electronic features of species arising from these mechanisms were provided. – The investigated compounds are completely planar molecules so that the conjugation and delocalization effects results to be maximized and crucial for stabilization of their radical, cationic and anionic forms. Spin density analysis indicates that such effects are enhanced in derivatives with respect to natural compounds. – On the basis of the computed values of the antioxidant activity descriptors, BDE, IP, PDE, PA and ETE, the most likely mechanism by which hydroxycinnamic acids and derivatives exert their antioxidant activity, seems to be the HAT one. BDE values computed for the natural compounds fall in the range of

–

–

–

–

7

80–82 kcal/mol, while, among the studied derivatives, the compound 2 show a lower BDE (77.5 kcal/mol) which makes it the better candidate to be proposed as antioxidant. The assessment of reaction enthalpies of the investigated compounds with selected radical species, confirms the HAT mechanism as the most probable one as well as the derivative compound 2 as the most active antioxidant among those investigated. From our outcomes, both SET-PT and SPLET reaction mechanisms involve a significant energetic disadvantage with respect to HAT. The analysis of the frontier orbital energies confirms the IP adiabatic trend, according to which compound 1 is expected to be the best electron donor. The simulated UV–Vis spectra of natural antioxidants, CA and FA, are in good agreement with the experimental counterpart, supporting the reliability of the TDDFT results obtained for the derivatives investigated in this study. Such a spectra are characterized by a broad band centered around 340 and 380 nm for naturally occurring and synthesized compounds, respectively, generated by a H ? L electronic transition. Other two absorption bands, with lower oscillator strengths, have been found arising from H-1 ? L and H ? L + 1 transitions. In the case of derivatives 1–3 the band at lower wavelengths involves H-2 ? L and H ? L + 1 electronic transitions.

Our investigation confirms the antioxidant properties of hydroxycinnamic acids and derivatives. The structural modification of the ethylenic spacer between the aromatic ring and the carboxylic functionality enhances the ability of derivatives to act as antioxidants. Accordingly, our calculations strongly encourage the synthesis of derivatives arising from ad-hoc structural modifications which could improve the antioxidant properties. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.comptc.2015.10. 011. References [1] B. Halliwell, Free radicals and other reactive species in disease, eLS (2001) 1–7. [2] L.A. Pham-Huy, H. He, C. Pham-Huy, Free radicals, antioxidants in disease and health, Int. J. Biomed. Sci. 4 (2008) 89–96. [3] V. Lobo, A. Patil, A. Phatak, N. Chandra, Free radicals, antioxidants and functional foods: impact on human health, Pharm. Rev. 4 (2010) 118–126. [4] T.M. Florence, The role of free radicals in disease, Aust. N. Z. J. Ophthalmol. 23 (1995) 3–7. [5] B. Halliwell, Free radicals and other reactive species in disease, eLS (2015) 1–9. [6] P.G. Pietta, Flavonoids as antioxidants, J. Nat. Prod. 63 (2000) 1035–1042. [7] F.M.F. Roleira, E.J. Tavares-Da-Silva, C.L. Varela, S.C. Costa, T. Silva, J. Garrido, F. Borges, Plant derived and dietary phenolic antioxidants: anticancer properties, Food Chem. 183 (2015) 235–258. [8] H. Parhiz, A. Roohbakhsh, F. Soltani, R. Rezaee, M. Iranshahi, Antioxidant and anti-inflammatory properties of the citrus flavonoids hesperidin and hesperetin: an updated review of their molecular mechanisms and experimental models, Phytother. Res. 29 (2015) 323–331. [9] P. Widsten, C.D. Cruz, G.C. Fletcher, M.A. Pajak, T.K. McGhie, Tannins and extracts of fruit byproducts: antibacterial activity against foodborne bacteria and antioxidant capacity, J. Agric. Food Chem. 62 (2014) 11146–11156. [10] R.G. Panchal, S.P. Reid, J.P. Tran, A.A. Bergeron, J. Wells, K.P. Kota, J. Aman, S. Bavari, Identification of an antioxidant small-molecule with broad-spectrum antiviral activity, Antiviral Res. 93 (2012) 23–29. [11] J. Tinkel, H. Hassanain, S.J. Khouri, Cardiovascular antioxidant therapy: a review of supplements, pharmacotherapies, and mechanisms, Cardiol. Rev. 20 (2012) 77–83. [12] C.C. Danta, P. Piplani, The discovery and development of new potential antioxidant agents for the treatment of neurodegenerative diseases, Expert Opin. Drug Discov. 9 (2014) 1205–1222. [13] G. Marrazzo, I. Barbagallo, F. Galvano, M. Malaguarnera, D. Gazzolo, A. Frigiola, N. D’Orazio, G. Li Volti, Role of dietary and endogenous antioxidants in diabetes, Crit. Rev. Food Sci. Nutr. 54 (2014) 1599–1616.

Please cite this article in press as: G. Mazzone et al., Antioxidant properties comparative study of natural hydroxycinnamic acids and structurally modified derivatives: Computational insights, Comput. Theoret. Chem. (2015), http://dx.doi.org/10.1016/j.comptc.2015.10.011

8

G. Mazzone et al. / Computational and Theoretical Chemistry xxx (2015) xxx–xxx

[14] X.T. Yan, S.H. Lee, W. Li, Y.N. Sun, S.Y. Yang, H.D. Jang, Y.H. Kim, Evaluation of the antioxidant and anti-osteoporosis activities of chemical constituents of the fruits of Prunus mume, Food Chem. 156 (2014) 408–415. [15] A. Vysakh, M. Ratheesh, T.P. Rajmohanan, C. Pramod, S. Premlal, B. Girish Kumar, P.I. Sibi, Polyphenolics isolated from virgin coconut oil inhibits adjuvant induced arthritis in rats through antioxidant and antiinflammatory action, Int. Immunopharmacol. 20 (2014) 124–130. [16] P. Sunkireddy, S.N. Jha, J.R. Kanwar, S.C. Yadav, Natural antioxidant biomolecules promises future nanomedicine based therapy for cataract, Colloids Surf. B Biointerf. 112 (2013) 554–562. [17] G.G. Duthie, P.T. Gardner, J.A. Kyle, Plant polyphenols: are they the new magic bullet?, Proc Nutr. Soc. 62 (2003) 599–603. [18] C. Sanchez-Moreno, J.A. Larrauri, F. Saura-Calixto, A procedure to measure the antiradical efficiency of polyphenols, J. Sci. Food Agric. 76 (1998) 270–276. [19] Z. Sroka, W. Cisowski, Hydrogen peroxide scavenging, antioxidant and antiradical activity of some phenolic acids, Food Chem. Toxicol. 41 (2003) 753– 758. [20] L. Cao, H. Yu, S. Shao, S. Wang, Y. Guo, Evaluating the antioxidant capacity of polyphenols with an off-on fluorescence probe and the mechanism study, Anal. Methods 6 (2014) 7149–7153. [21] A. Martínez, A. Galano, R. Vargas, Free radical scavenger properties of amangostin: thermodynamics and kinetics of HAT and RAF mechanisms, J. Phys. Chem. B 115 (2011) 12591–12598. [22] Y. Xue, Y. Zheng, L. An, L. Zhang, Y. Qian, D. Yu, X. Gong, Y. Liu, A theoretical study of the structure-radical scavenging activity of hydroxychalcones, Comput. Theor. Chem. 982 (2012) 74–83. [23] A.M. Mendoza-Wilson, S.I. Castro-Arredondo, R.R. Balandrán-Quintana, Computational study of the structure-free radical scavenging relationship of procyanidins, Food Chem. 161 (2014) 155–161. [24] A. Galano, A. Martínez, Capsaicin, a tasty free radical scavenger: mechanism of action and kinetics, J. Phys. Chem. B 116 (2012) 1200–1208. [25] J.M. Dimitric´, D. Markovic´, D. Milenkovic´, M. Amic´, I. Mojovic´, Z.S. Pašti Markovic´, The preferred radical scavenging mechanisms of fisetin and baicalein towards oxygen-centred radicals in polar protic and polar aprotic solvents, RSC Adv. 4 (2014) 32228–32236. [26] R. Praveena, K. Sadasivam, V. Deepha, R. Sivakumar, Antioxidant potential of orientin: a combined experimental and DFT approach, J. Mol. Struct. 1061 (2014) 114–123. [27] D. Mikulski, K. Eder, M. Molski, Quantum-chemical study on relationship between structure and antioxidant properties of hepatoprotective compounds occurring in Cynara scolymus and Silybum marianum, J. Theor. Comput. Chem. 13 (2014) 1450004–1450027. [28] A. Galano, J.R. Alvarez-Idaboy, Glutathione: mechanism and kinetics of its non-enzymatic defense action against free radicals, RSC Adv. 1 (2011) 1763– 1771. [29] R. Castañeda-Arriaga, J.R. Alvarez-Idaboy, Lipoic acid and dihydrolipoic acid. A comprehensive theoretical study of their antioxidant activity supported by available experimental kinetic data, J. Chem. Inf. Model. 54 (2014) 1642–1652. [30] D. Farmanzadeh, M. Najafi, On the antioxidant activity of the tryptophan derivatives, Bull. Chem. Soc. Jpn. 86 (2013) 1041–1050. [31] S. Sebastian, N. Sundaraganesan, S. Manoharan, Molecular structure, spectroscopic studies and first-order molecular hyperpolarizabilities of ferulic acid by density functional study, Spectrochim. Acta A 74 (2009) 312– 323. [32] S. Vazquez-Rodriguez, R. Figueroa-Guíñez, M.J. Matos, L. Santana, E. Uriarte, M. Lapier, J.D. Maya, C. Olea-Azar, Synthesis of coumarin–chalcone hybrids and evaluation of their antioxidant and trypanocidal properties, Med. Chem. Commun. 4 (2013) 993–1000. [33] E. Pontiki, D. Hadjipavlou-Litina, K. Litinas, G. Geromichalos, Novel cinnamic acid derivatives as antioxidant and anticancer agents: design, synthesis and modeling studies, Molecules 19 (2014) 9655–9674. [34] M. Miliovsky, I. Svinyarov, E. Prokopova, D. Batovska, S. Stoyanov, M.G. Bogdanov, Synthesis and antioxidant activity of polyhydroxylated transrestricted 2-arylcinnamic acids, Molecules 20 (2015) 2555–2575. [35] M.J. Matos, F. Mura, S. Vazquez-Rodriguez, F. Borges, L. Santana, E. Uriarte, C. Olea-Azar, Study of coumarin-resveratrol hybrids as potent antioxidant compounds, Molecules 20 (2015) 3290–3308. [36] D. Chavarria, T. Silva, D. Martins, J. Bravo, T. Summavielle, J. Garrido, F. Borges, Exploring cinnamic acid scaffold: development of promising neuroprotective lipophilic antioxidants, Med. Chem. Commun. 6 (2015) 1043–1053. [37] P.J. White, Y. Xing, Antioxidants from cereals and legumes, in: F. Shahidi (Ed.), Natural Antioxidants, Chemistry, Health Effects, and Application, AOCC Press, Champaign, IL, 1997, pp. 25–63. [38] Gaussian 09, Revision D.01, M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J.A. Montgomery, Jr., J.E. Peralta, F. Ogliaro, M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S.S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth, P. Salvador, J.J. Dannenberg, S. Dapprich, A.D. Daniels, Ö. Farkas, J.B. Foresman, J.V. Ortiz, J. Cioslowski, D.J. Fox, Gaussian Inc., Wallingford CT, 2009.

[39] Y. Zhao, N.E. Schultz, D.G. Truhlar, Design of density functionals by combining the method of constraint satisfaction with parametrization for thermochemistry, thermochemical kinetics, and noncovalent interactions, J. Chem. Theory Comput. 2 (2006) 364–382. [40] A. Galano, J.R. Alvarez-Idaboy, Kinetics of radical-molecule reactions in aqueous solution: a benchmark study of the performance of density functional methods, J. Comput. Chem. 35 (2014) 2019–2026. [41] A.V. Marenich, C.J. Cramer, D.G. Truhlar, Universal solvation model based on solute electron density and on a continuum model of the solvent defined by the bulk dielectric constant and atomic surface tensions, J. Phys. Chem. B 113 (2009) 6378–6396. [42] G. Mazzone, N. Malaj, N. Russo, M. Toscano, Density functional study of the antioxidant activity of some recently synthesized resveratrol analogues, Food Chem. 141 (2013) 2017–2024. [43] G. Mazzone, M. Toscano, N. Russo, Density functional predictions of antioxidant activity and UV spectral features of Nasutin A, Isonasutin, Ellagic Acid, and one of its possible derivatives, J. Agric. Food Chem. 61 (2013) 9650– 9657. [44] G. Mazzone, N. Malaj, A. Galano, N. Russo, M. Toscano, Antioxidant properties of several coumarin–chalcone hybrids from theoretical insights, RSC Adv. 5 (2015) 565–575. [45] Z. Markovic´, D. Amic´, D. Milenkovic´, J.M. Dimitric´-Markovic´, S. Markovic´, Examination of the chemical behavior of the quercetin radical cation towards some bases, Phys. Chem. Chem. Phys. 15 (2013) 7370–7378. [46] E. Klein, V. Lukes, DFT/B3LYP study of the substituent effect on the reaction enthalpies of the individual steps of single electron transfer–proton transfer and sequential proton loss electron transfer mechanisms of phenols antioxidant action, J. Phys. Chem. A 110 (2006) 12312–12320. [47] J.E. Bartmess, Thermodynamics of the electron and the proton, J. Phys. Chem. 98 (1994) 6420–6424. [48] M.E. Cassida, in: D.P. Chong (Ed.), Recent Advances in Density Functional Methods, Part I, World Scientific, Singapore, 1995, pp. 155–192. [49] Y. Zhao, D.G. Truhlar, The M06 suite of density functionals for main group thermochemistry, thermochemical kinetics, noncovalent interactions, excited states, and transition elements: two new functionals and systematic testing of four M06-class functionals and 12 other functionals, Theor. Chem. Acc. 120 (2008) 215–241. [50] J.-D. Chai, M. Head-Gordon, Long-range corrected hybrid density functionals with damped atom–atom dispersion corrections, Phys. Chem. Chem. Phys. 10 (2008) 6615–6620. [51] J.-D. Chai, M. Head-Gordon, Systematic optimization of long-range corrected hybrid density functionals, J. Chem. Phys. 128 (2008). 84106-1-15. [52] C. Adamo, V. Barone, Toward reliable density functional methods without adjustable parameters: the PBE0 model, J. Chem. Phys. 110 (1999) 6158–6170. [53] M. Ernzerhof, G.E. Scuseria, Assessment of the Perdew–Burke–Ernzerhof exchange-correlation functional, J. Chem. Phys. 110 (1999) 5029–5036. [54] A.D. Becke, Density-functional thermochemistry. III. The role of exact exchange, J. Chem. Phys. 98 (1993) 5648–5652. [55] C. Lee, W. Yang, R.G. Parr, Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density, Phys. Rev. B 37 (1988) 785– 789. [56] M. Leopoldini, S.G. Chiodo, N. Russo, M. Toscano, Detailed investigation of the OH radical quenching by natural antioxidant caffeic acid studied by quantum mechanical models, J. Chem. Theory Comput. 7 (2011) 4218–4233. [57] M. Leopoldini, T. Marino, N. Russo, M. Toscano, Antioxidant properties of phenolic compounds. H-atom versus electron-transfer mechanism, J. Phys. Chem. A 108 (2004) 4916–4922. [58] M. Leopoldini, T. Marino, N. Russo, M. Toscano, Density functional computations of the energetic and spectroscopic parameters of quercetin and its radicals in the gas phase and in solvent, Theor. Chem. Acc. 111 (2004) 210–216. [59] M. Leopoldini, N. Russo, M. Toscano, Gas and liquid phase acidity of natural antioxidants, J. Agric. Food Chem. 54 (2006) 3078–3085. [60] M. Leopoldini, N. Russo, M. Toscano, The molecular basis of working mechanism of natural polyphenolic antioxidants, Food Chem. 125 (2011) 288–306. [61] C.J. Parkinson, P.M. Mayer, L. Radom, An assessment of theoretical procedures for the calculation of reliable radical stabilization energies, J. Chem. Soc. Perkin Trans. 2 (1999) 2305–2313. [62] J. Ðorovic´, J.M. Dimitric´ Markovic´, V. Stepanic´, N. Begovic´, D. Amic´, Z. Markovic´, Influence of different free radicals on scavenging potency of gallic acid, J. Mol. Model. 20 (2014) 2345. [63] Z. Markovic´, J. Ðorovic´, J.M. Dimitric´ Markovic´, M. Zˇivic´, D. Amic´, Investigation of the radical scavenging potency of hydroxybenzoic acids and their carboxylate anions, Monatsh. Chem. 145 (2014) 953–962. [64] A.N. Queiroz, B.A.Q. Gomes, W.M. Moraes Jr., R.S. Borges, A theoretical antioxidant pharmacophore for resveratrol, Eur. J. Med. Chem. 44 (2009) 1644–1649. [65] K.M. Honório, A.B.F. da Silva, An AM1 study on the electron-donating and electron-accepting character of biomolecules, Int. J. Quant. Chem. 95 (2003) 126–132. [66] R. S´wisłocka, Spectroscopic (FT-IR, FT-Raman, UV absorption, 1H and 13C NMR) and theoretical (in B3LYP/6-311++G⁄⁄ level) studies on alkali metal salts of caffeic acid, Spectrochim. Acta A 100 (2013) 21–30. [67] R.A. Holser, International Scholarly Research Network ISRN Spectroscopy, vol. 2012, Article ID 493203, 5 pages, doi: 10.5402/2012/493203.

Please cite this article in press as: G. Mazzone et al., Antioxidant properties comparative study of natural hydroxycinnamic acids and structurally modified derivatives: Computational insights, Comput. Theoret. Chem. (2015), http://dx.doi.org/10.1016/j.comptc.2015.10.011

G. Mazzone et al. / Computational and Theoretical Chemistry xxx (2015) xxx–xxx [68] J.-P. Cornard, C. Lapouge, Absorption spectra of caffeic acid, caffeate and their 1:1 complex with Al(III): density functional theory and time-dependent density functional theory investigations, J. Phys. Chem. A 110 (2006) 7159– 7166.

9

[69] A. Belay, E. Libnedengel, H.K. Kima, Y.-H. Hwanga, Effects of solvent polarity on the absorption and fluorescence spectra of chlorogenic acid and caffeic acid compounds: determination of the dipole moments, Luminescence (2015), http://dx.doi.org/10.1002/bio.2932.

Please cite this article in press as: G. Mazzone et al., Antioxidant properties comparative study of natural hydroxycinnamic acids and structurally modified derivatives: Computational insights, Comput. Theoret. Chem. (2015), http://dx.doi.org/10.1016/j.comptc.2015.10.011