Quan V. Vo*a,
Duong Thi Thuy Hoab,
Nguyen Thi Hoaa,
Manh Duc Tranb and
Adam Mechlerc
aThe University of Danang – University of Technology and Education, Danang 550000, Vietnam. E-mail: vvquan@ute.udn.vn
bThe University of Danang – University of Sciences and Education, Danang 550000, Vietnam
cDepartment of Biochemistry and Chemistry, La Trobe University, Victoria 3086, Australia
First published on 30th January 2024
Caffeoylquinic acids (CQAs) are well-known antioxidants. However, a key aspect of their radical scavenging activity – the mechanism of action – has not been addressed in detail thus far. Here we report on a computational study of the mechanism of activity of CQAs in scavenging hydroperoxyl radicals. In water at physiological pH, the CQAs demonstrated ≈ 104 times higher HOO˙ antiradical activity than in lipid medium (k(lipid) ≈ 104 M−1 s−1). The activity in the aqueous solution was determined by the hydrogen transfer mechanism of the adjacent hydroxyl group (O6′–H) of the dianion states (Γ = 93.2–95.2%), while the single electron transfer reaction of these species contributed 4.8–6.8% to the total rate constants. The kinetics estimated by the calculations are consistent with experimental findings in water (pH = 7.5), yielding a kcalculated/kexperimental = 2.4, reinforcing the reliability and precision of the computational method and demonstrating its utility for evaluating radical reactions in silico. The results also revealed the pH dependence of the HOO˙ scavenging activity of the CQAs; activity was comparable for all compounds below pH 3, however at higher pH values 5CQA reacted with the HOO˙ with lower activity than 3CQA or 4CQA. It was also found that CQAs are less active than Trolox below pH 4.7, however over pH 5.0 they showed higher activity than the reference. The CQAs had the best HOO˙ antiradical activity at pH values between 5.0 and 8.6. Therefore, in the physiological environment, the hydroperoxyl antiradical capacity of CQAs exhibits similarity to renowned natural antioxidants including resveratrol, ascorbic acid, and Trolox.
Chlorogenic acids have the ability to scavenge free radicals from a variety of sources, including 2,21-azino-bis(3-ethylbenzothiazoline-6-sulphonic acid) (ABTS) radicals, 1,1-diphenyl-2-picrylhydrazyl (DPPH) radicals, superoxide anions (O2), hydroxyl radicals (OH) and peroxynitrite (ONOO).13,16,17 Kinetics of 5CQA were determined experimentally: it reacts with the superoxide, peroxynitrite, and peroxyl radical with second-order kinetics and rate constants of 3.34 × 109, 9.60 × 105, 1.6 × 105 and 1.28 × 105 M−1 s−1, respectively,17 whereas the 5CQA reacted with HO˙ radicals with rate constant of 109–1010 M−1 s−1.18,19 Computational approaches were also used to evaluate the antioxidant activity of CQAs,18–23 however the mechanism and kinetics of the HOO˙ radical scavenging activity, especially in physiological environments and at different pH values have not been thoroughly investigated. In particular, the effect of different pH levels on the kinetics and mechanism of the radical scavenging activity of phenolic acids is well known,24–26 and thus a more comprehensive examination is desirable. Furthermore, the examination of radical reactions involving HO˙ and HOO˙ is significant not only in the advanced oxidation processes,27–29 but also in the radical scavenging activity of antioxidants.30–33 Accordingly, in this study, thermodynamic and kinetic calculations were used to examine the hydroperoxyl radical scavenging activity of the CQAs in physiological conditions and at varied pH levels.
The rate constant (k) was determined through the application of the usual transition state theory (TST) under the conditions of a 1 M standard state,44–48 and the details are shown in Table S1, ESI.†
(1) |
The computations were performed utilizing the Gaussian 16 suite of programs52 and the Eyringpy code, depending on the particular circumstance.53,54 Atom-in-molecule (AIM) analysis55 was performed by using the AIM2000 software.56
In order to evaluate the likelihood of reacting with free radicals of all possible X–H (X = C, O) bonds including C–H (C2, C3, C4, C5 and C6), and O–H (COOH, O1, O3, O5, O6′ and O7′, Fig. 1) key thermochemical characteristics: bond dissociation energies (BDEs), proton affinities (PAs), and ionization energies (IEs) that provide a first approximation for the probability of reactions following either of the three respective mechanisms, i.e. FHT (formal hydrogen transfer), SET (single electron transfer) and PL (proton loss),41,60,61 were first computed in physiological environments (and water (W) and pentyl ethanoate (P)). The results are shown in Fig. 2.
Fig. 2 The computed BDE, PA, IE (kcal mol−1) of CQAs and ΔG° (kcal mol−1) of the HOO˙ + CQAs reactions following the SET, PL, and FHT mechanisms (a: 3CQA; b: 4CQA; c: 5CQA). |
The lowest BDE values were observed at the O6′(7′)–H bonds in all of the studied acids with BDE(O6′(7′)–H) = 79.1–81.9 kcal mol−1 for the lipid medium and BDE(O6′(7′)–H) = 82.7–85.4 kcal mol−1 for the aqueous solution (Fig. 2a1–c1). The active site can be attributed in all cases to the formation of intramolecular hydrogen bonds between the hydrogen atoms of the adjacent hydroxyl groups and the O6′(O7′) radicals.60,62 While the values for other O–H bonds ranged from 104.9 to 131.6 kcal mol−1, the BDE(C–H) values were between 88.3 and 102.6 kcal mol−1. Surprisingly the H-abstraction of the COO–H bond was less likely with the BDE(COO–H) = 112.3–114.4 kcal mol−1 (to emphasize, this refers to hydrogen abstraction; proton dissociation is much more likely, see below). The BDE values in the water were slightly higher than those in the pentyl ethanoate solvent. As expected the PA and IE values in the polar medium were lower than those of the nonpolar environment. The deprotonation was in the order of COO–H > O7′–H > O6′–H in all of the studied compounds, whereas the IE values varied from 93.5 to 139.6 kcal mol−1.
The evaluation in the Gibbs free energies (ΔG°, Fig. 2a2–c2) of the HOO˙ + CQAs reactions following either of the three pathways revealed that the HOO˙ radical trapping activity of CQAs is only spontaneous via the hydrogen transfer of the O6′(7′)-H bonds (ΔG° = −3.6 to −6.2 kcal mol−1), whereas the other FHT reactions cannot happen in the studied media due to the positive ΔG° values. The SET mechanism is not spontaneous either in any of the studied environments, thus this reaction of the neutral states of CQAs can be safely ignored in the kinetic study. It is important to notice that the PL reactions are not spontaneous either in any of the studied environments; however, the PA values were substantially lower than the corresponding BDE values, thus the deprotonation of CQAs should be considered in the aqueous solution. Previous studies indicated that the addition reaction into the α,β-unsaturated bond had no contributions to the ROO˙ radical (i.e., HOO˙ and CH3OO˙) scavenging activity, particularly in the physiological environments,19,24,63 and RAF reaction is not supported for the π system of aromatic rings.64,65 Thus, this reaction was omitted in our study. Hence, in the lipid medium, the H-abstraction of the O6′(7′)–H bonds should be used to compute rate constants, whereas, in the aqueous solution, proton dissociation should be assessed before the kinetic investigation.
Comp. | Group | pKa | f (pH = 7.40)c | ||
---|---|---|---|---|---|
a Ref. 66.b Calculated in this work.c f Molar fraction. | |||||
3CQA | COOH | 1 | 3.95a | H3A | 0.000 |
O6′–H | 2 | 8.00b | H2A− | 0.799 | |
O7′–H | 3 | 12.59b | HA2− | 0.201 | |
A3− | 0.000 | ||||
4CQA | COOH | 1 | 4.14a | H3A | 0.000 |
O6′–H | 2 | 7.97b | H2A− | 0.788 | |
O7′–H | 3 | 12.52b | HA2− | 0.212 | |
A3− | 0.000 | ||||
5CQA | COOH | 1 | 3.95a | H3A | 0.000 |
O6′–H | 2 | 8.22b | H2A− | 0.868 | |
O7′–H | 3 | 12.27b | HA2− | 0.131 | |
A3− | 0.000 |
The pKa1 values for 3CQA and 5CQA are 3.95, while those for 4CQA are 4.14 (Table 1). The range of pKa2 values is 7.97 to 8.22, while the range of pKa3 values is 12.27 to 12.59. The calculated pKa values of 5CQA (pKa2 = 8.22 and pKa3 = 12.27) closely align with the experimental results (pKa2 = 8.21 and pKa3 = 12.5),67 providing evidence for the accuracy and validity of the computational approach. The mole fractions f(H2A−) and f(HA2−) range between 0.788 and 0.868 and between 0.131 and 0.212, respectively, while the H3A and A3− phases are not present in water at pH = 7.40. Therefore, the CQAs exist in both anionic and dianionic states in water with a pH of 7.4. These two states were examined in the subsequent investigation.
Comp. | Mechanisms | Pentyl ethanoate | Water | ||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
ΔG≠ | kapp | Γ | States | ΔG≠ | kapp | F | kf | Γ | |||
3CQA | SET | H2A− | 34.3 | 15.9 | 0.799 | 3.84 × 10−13 | 0.0 | ||||
HA2− | 6.4 | 1.20 × 108 | 0.201 | 2.41 × 107 | 4.8 | ||||||
FHT | O6′–H | 15.3 | 6.30 × 103 | 57.8 | H2A− | 15.9 | 4.80 × 103 | 0.799 | 3.84 × 103 | 0.0 | |
O7′–H | 15.2 | 4.60 × 103 | 42.2 | 16.6 | 1.00 × 103 | 0.799 | 7.99 × 102 | 0.0 | |||
O6′–H (HA2−) | HA2− | 1.0 | 2.40 × 109 | 0.201 | 4.82 × 108 | 95.2 | |||||
ktotal | 1.09 × 104 | 5.07 × 108 | |||||||||
4CQA | SET | H2A− | 32.3 | 17.2 | 0.788 | 1.10 × 10−8 | 0.0 | ||||
HA2− | 6.4 | 1.30 × 108 | 0.212 | 2.76 × 107 | 5.2 | ||||||
FHT | O6′–H | 14.9 | 8.00 × 102 | 4.8 | H2A− | 16.2 | 2.10 × 103 | 0.788 | 1.65 × 103 | 0.0 | |
O7′–H | 14.6 | 1.60 × 104 | 95.2 | 16.1 | 4.70 × 103 | 0.788 | 3.70 × 103 | 0.0 | |||
O6′–H (HA2−) | HA2− | 1.8 | 2.38 × 109 | 0.212 | 5.05 × 108 | 94.8 | |||||
ktotal | 1.68 × 104 | 5.32 × 108 | |||||||||
5CQA | SET | H2A− | 38.2 | 15.1 | 0.869 | 5.56 × 10−16 | 0.0 | ||||
HA2− | 6.8 | 6.60 × 107 | 0.131 | 8.65 × 106 | 6.8 | ||||||
FHT | O6′–H | 15.3 | 9.50 × 103 | 49.2 | H2A− | 18.0 | 2.40 × 102 | 0.869 | 2.09 × 102 | 0.0 | |
O7′–H | 15.0 | 9.80 × 103 | 50.8 | 17.5 | 3.10 × 102 | 0.869 | 2.69 × 102 | 0.0 | |||
O6′–H (HA2−) | HA2− | 5.4 | 9.00 × 108 | 0.131 | 1.18 × 108 | 93.2 | |||||
ktotal | 1.93 × 104 | 1.27 × 108 |
Lipid environment:
ktotal = kapp(FHT(O6′–H)-neutral) + kapp(FHT(O7′–H)-neutral) | (2) |
Water at physiological pH:
ktotal = kf(SET-HA−) + kf(FHT(O6′–H)-HA−) + kf(FHT(O7′–H)-HA−) + kf(SET-A2−) + kf(FHT(O6′–H)-A2−) | (3) |
kf = kapp·f | (4) |
As shown in Table 2, the calculations suggest that CQAs can be potent HOO˙ scavengers in the nonpolar environment, with ktotal = 1.09 × 104–1.93 × 104 M−1 s−1. The rate constant for the reaction between 5CQA and HOO˙ was found to be the highest, whereas the reaction between 3CQA and HOO˙ exhibited the lowest rate constant. Based on the calculated data, the HOO˙ radical trapping ability of CQAs in the lipid medium can be ranked as follows: 5CQA > 4CQA > 3CQA. Thus, the activity of CQAs in the nonpolar medium is comparable to reference antioxidants including ascorbic acid (k = 5.71 × 103 M−1 s−1),40 resveratrol (k = 1.31 × 104 M−1 s−1),64 and Trolox (k = 3.40 × 103 M−1 s−1).68
In water at pH = 7.40, the FHT mechanism of the neighboring hydroxyl group (O6′–H) of the dianion states determined the HOO˙ radical scavenging activity of the CQAs (Γ = 93.2–95.2%), while the SET reaction of these species contributed approximately 4.8–6.8% to the ktotal. It should be noted that the tunneling corrections (κ) had a negligible impact on the H-abstraction rate constant of the dianion state, the substantial imaginary frequencies (ν > 3000 cm−1) of these transition states notwithstanding. This suggests that the remarkably swift reaction rates are caused solely by the excessively low Gibb activation energy (ΔG≠ = 1.0–5.4 kcal mol−1) (kTST ≈ kD, where kD denotes the diffusion rate). The HOO˙ radical scavenging activity of the CQAs was not influenced by the monoanion states, despite the fact these states exist about 13.1% to 21.2% (Table 1) of the CQAs in water at pH = 7.4. All compounds exhibit outstanding HOO˙ antiradical activity with ktotal ≈ 108 M−1 s−1. 4CQA had the maximum activity with ktotal = 5.32 × 108 M−1 s−1 that is approximately 4.19 and 1.05 times faster than 5CQA and 3CQA, respectively. In water at pH = 7.4, the radical trapping activity of CQAs against HOO˙ is ranked as follows: 4CQA > 5CQA > 3CQA. In water at physiological pH, the CQAs demonstrated ≈ 104 times greater HOO˙ radical scavenging ability than in the nonpolar environment. In water, CQAs have greater HOO˙ antiradical capacity than Trolox (k = 8.96 × 104 M−1 s−1),68 resveratrol (k = 5.62 × 107 M−1 s−1),64 and ascorbic acid (k = 9.97 × 107 M−1 s−1),40 but the fairly similar activity to caffeic acid (k = 2.69 × 108 M−1 s−1), ferulic acid (k = 3.36 × 108 M−1 s−1) and dihydrocaffeic acid (k = 1.04 × 108 M−1 s−1).63 Consequently, CQAs are promising natural antioxidants.
According to the information provided above, the removal of a hydrogen atom from the neighboring hydroxyl group (O6′–H bond) of the dianion states is responsible for the scavenging activity of CQAs in water at a pH level that is characteristic of physiological conditions. In this section, AIM analysis was employed to examine the structural characteristics of the transition states (TSs) pertaining to the O6′–H bond. The findings are displayed in Table S3, ESI,† and Fig. 5. The analysis of the AIM data reveals that the stability of the TSs-DIANION-O6′-H-OOH species can be attributed to intermolecular interactions occurring at certain sites, namely O6′⋯H41, O43⋯H41, O7′⋯H42, and PCR(O6′–H41–O43–O44–H42–O7′–C7′–C6′) (Fig. 5), whereas the stability of TSs-ANION-O6′–H–OOH is given by the intramolecular hydrogen bonding at H41⋯O6′, H41⋯O44 and O45⋯C6′ and at PCR(C6′–O6′–H41–O44–O45). The formation of an intramolecular hydrogen bond between O7′ and H42 in the TSs-DIANION-O6′-H-OOH results in the creation of an 8-atom ring, which includes O6′–H41–O43–O44–H42–O7′–C7′–C6′ (Fig. 5). The electron density between O7′ and H42 exhibits characteristics of partial covalent bonding, as evidenced by ∇2ρ(r) > 0, G(r)/|V(r)| ≤ 1 and H(r) < 0.69,70 This electron density significantly contributes to the stability of the transition state, as revealed by the substantial negative values of EHD (−26.9 to −27.9 kcal mol−1). Thus, the total EHB of TSs-DIANION-O6′–H–OOH (EHB = −211.0 to −212.9 kcal mol−1) is about 1.13 times lower than those of TSs-ANION-O6′–H–OOH (EHB = −184.6 to −186.7 kcal mol−1, respectively). This underpins the increased stability of TSs-DIANION-O6′–H–OOH and consequently the decreased ΔG≠ values (ΔG≠ = 1.0–5.4 kcal mol−1) in comparison with the remaining TSs.
k = kapp(SET) + ∑kapp(FHT) | (5) |
kstate = k·f(CQA) | (6) |
ktotal = ∑kstate | (7) |
koverall = f(HOO˙)·ktotal | (8) |
Fig. 6 Calculated log(ktotal) (a) and log(koverall) (b) at 298.15 K, in the CQAs + HOO˙ in water as a function of pH values. |
Comp. | States | Mechanisms | ΔG≠ | kapp | k | |
---|---|---|---|---|---|---|
3CQA | H3A | FHT | O6′–H | 17.4 | 5.80 × 102 | 1.86 × 103 |
O7′–H | 17.1 | 1.28 × 103 | ||||
H2A− | SET | 34.3 | 4.80 × 10−13 | 5.80 × 103 | ||
FHT | O6′–H | 15.9 | 4.80 × 103 | |||
O7′–H | 16.6 | 1.00 × 103 | ||||
HA2− | SET | 6.4 | 1.20 × 108 | 2.52 × 109 | ||
FHT | O6′–H | 1.0 | 2.40 × 109 | |||
A3− | SET | 0.0 | 8.10 × 109 | 8.10 × 109 | ||
4CQA | H3A | FHT | O6′–H | 16.3 | 5.00 × 101 | 2.15 × 103 |
O7′–H | 16.6 | 2.01 × 103 | ||||
H2A− | SET | 32.3 | 1.40 × 10−11 | 6.80 × 103 | ||
FHT | O6′–H | 16.2 | 2.10 × 103 | |||
O7′–H | 16.1 | 4.70 × 103 | ||||
HA2− | SET | 6.4 | 1.30 × 108 | 2.51 × 109 | ||
FHT | O6′–H | 1.8 | 2.38 × 109 | |||
A3− | SET | 0.0 | 8.10 × 109 | 8.10 × 109 | ||
5CQA | H3A | FHT | O6′–H | 17.0 | 1.80 × 103 | 2.23 × 103 |
O7′–H | 18.0 | 4.30 × 102 | ||||
H2A− | SET | 38.2 | 6.40 × 10−16 | 5.50 × 102 | ||
FHT | O6′–H | 18.0 | 2.40 × 102 | |||
O7′–H | 17.5 | 3.10 × 102 | ||||
HA2− | SET | 6.8 | 6.60 × 107 | 9.66 × 108 | ||
FHT | O6′–H | 5.4 | 9.00 × 108 | |||
A3− | SET | 0.0 | 8.30 × 109 | 8.30 × 109 |
The log(ktotal) for the total rate constant (Fig. 6a) did not change below pH 4.7, however, it increased significantly between pH = 4.8 and 8.3 by 4–6 units and afterward grew progressively until pH 14. The sudden increase in the log(ktotal) figures at pH = 4.8 and 8.3 is due to the appearance of HA2− states and thus the onset of rapid SET processes. In acidic media (pH < 4.7), the reactions between CQAs and HOO˙ are sluggish because the majority of the CQAs exist in the H3A states (neutral states), acting via a slow FHT reaction.
It was demonstrated that pKa(HOO˙) is 4.80, and thus the f(HOO˙) value is zero at pH > 9.1. Since only pH levels below 9.1 had any impact on the koverall values of reactions between CQAs and HOO˙, only these were looked at (Fig. 6b). It was found that as the pH levels rose, the koverall changed. Most of the studied acids showed a rise in log(koverall) at pH 4; after a brief fall, the log(koverall) significantly rose at pH = 4.7–6.5, before decreasing once again. For this range, koverall was 0 because f(HOO˙) = 0 at pH > 9.2 (Fig. 6b).
In terms of the CQAs, the exhibition was fairly similar to the HOO˙ antiradical activity at pH < 3, however at the rest pH values, 5CQA reacted with the HOO˙ lower than 3CQA or 4CQA. It is important to notice that the 3CQA acid had a fairly similar HOO˙ radical scavenging activity to 4CQA in all of the studied pH levels. Compared with typical antioxidant-Trolox, at pH 4.7, CQAs had less HOO˙ radical scavenging activity than Trolox; nevertheless, at pH > 5.0, these acids reacted with the HOO˙ more quickly than the standard. According to the calculated data, in the pH range of 5.0–8.6, the CQAs had the highest HOO˙ antiradical activity (log(koverall) = 5.1–6.2). It was found that the calculated rate constant for the 5CQA + HOO˙ reaction (koverall(calculation) = 3.10 × 105 M−1 s−1) exhibit a high level of consistence with the empirical observations (kexp = 1.28 × 105 M−1 s−1, pH = 7.5).17 Therefore, the computed kinetic values are fairly accurate.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3ra08460d |
This journal is © The Royal Society of Chemistry 2024 |