Computationally Designed Sesamol Derivatives Proposed as Potent Antioxidants.

Oxidative stress has been recognized to play an important role in several diseases, such as Parkinson's and Alzheimer's disease, which justifies the beneficial effects of antioxidants in ameliorating the deleterious effects of these health disorders. Sesamol, in particular, has been investigated for the treatment of several conditions because of its antioxidant properties. This article reports a rational computational design of new sesamol derivatives. They were constructed by adding four functional groups (-OH, -NH2, -COOH, and -SH) in three different positions of the sesamol molecular framework. A total of 50 derivatives between mono-, di-, and trisubstituted compounds were obtained. All the derivatives were evaluated and compared with a reference set of commercial neuroprotective drugs. The estimated properties are absorption, distribution, metabolism, excretion, toxicity, and synthetic accessibility. Selection and elimination scores were used to choose a first set of promising candidates. Acid-based properties and reactivity indexes were then estimated using the density functional theory. Four sesamol derivatives were finally selected, which are hypothesized to be potent antioxidants, even better than sesamol and Trolox for that purpose.

Introduction

Because of its biradical nature, molecular oxygen readily accepts electrons yielding a series of partially reduced species collectively known as reactive oxygen species (ROS). Some of them are superoxide radical anions (O2•–), hydrogen peroxide (H2O2), hydroxyl radicals (HO•), peroxyl radicals (ROO•), and alkoxyl radicals (RO•). They are frequently involved in the initiation and propagation of chain reactions, which are highly damaging to cells.[1−3] Oxidative stress (OS) is the result of the unregulated production of ROS and the imbalance in pro-oxidant/antioxidant homeostasis that leads to the generation of toxic ROS.[4,5] OS plays an important role in neuronal degeneration disorders such as Parkinson’s disease, Alzheimer’s disease, amyotrophic lateral sclerosis,[6−8] diabetic complications,[9,10] vascular diseases,[11] chronic inflammation, stroke and septic shock, atherosclerosis, rheumatoid arthritis, and cancer.[12]

In addition to their use as food additives, mostly to prevent rancidity, antioxidants are effective in protecting tissues or cells from oxidative damage. They are considered protectors against aging, poisoning by toxic agents,[13−15] and the already mentioned consequences of OS. Antioxidants react with short-lived free radicals, yielding the corresponding long-living radical intermediates. Because these intermediates are less reactive than the original radicals, the damage induced by ROS is lowered. Recently, there has been a growing interest in natural antioxidants. They are known to possess a variety of biological activities such as antitumor, antimitotic, antiviral, and other activities.[16]

Sesame, an important oil seed derived from Sesamum indicum, is one of the oldest oil seeds known to man, and it is considered to have not only nutritional value but also some medicinal effects.[17,18] Sesame oil contains several minor constituents including sesamin, sesamolin, and sesamol.[19] The latter is one partially responsible of the high resistance of sesame oil to oxidative rancidity as compared to other vegetable oils such as olive, peanut, and soybean oils.[20,21] Sesamol has shown to be an antiaging agent by preventing photodamage caused by chronic UV exposure,[22,23] to exhibit antimutagenic effects,[20] and to be a plasminogen activator in the prevention of atherosclerosis.[24] It has also been studied for an extensive variety of disease treatments.[25−31]

One of the most widely studied properties of sesamol is its antioxidant activity.[32−36] It has been proved to have effective capacity for inhibiting hydroxyl radical-induced deoxyribose degradation[37] and lipid peroxidation[23,37−39] which helps protecting plasma, low-density lipoprotein, and erythrocyte membrane from oxidation.[40−42] Sesamol also induces nitric oxide (the most important vascular relaxing factor) release from endothelial cells.[28,43] It has scavenging effects against hydroxyl (OH•), superoxide anion (O2•–), nitric oxide, hydrogen peroxide (H2O2), and other oxidants,[44−47] in some cases with an efficiency higher, or equal to those of ascorbic acid[48] and vitamin E.[49] Based on its free radical scavenging activity, it has been suggested to use sesamol in the management of Huntington’s, Alzheimer’s, and Parkinson’s diseases, auto-immunity disorders, and brain damage in autism.[50−54]

The interesting properties of sesamol have motivated investigations on the possible applications of its derivatives for different purposes. Some of them, already experimentally obtained, are shown in Scheme ,[55−58] while Scheme shows the computationally designed sesamol derivatives that have been previously studied.[59] Computational Chemistry has been proven to be a useful tool for the study of antioxidants in general[60] and of sesamol in particular, including the importance of its acid base equilibrium on its antioxidant activity.[61]

Scheme 1

Sesamol and Experimentally Obtained Sesamol Derivatives[55−58]

Scheme 2

Computationally Designed and Evaluated Derivatives[59]

In this work, the search for sesamol derivatives with an improved antioxidant behavior has been extended. To that purpose, a large set of this kind of molecules has been rationally designed. They were constructed by adding different functional groups in different positions of the parent molecule to generate mono-, di-, and trifunctionalized species. Absorption, distribution, metabolism, and excretion (ADME) properties, toxicity, and synthetic accessibility were estimated in silico. Selection and elimination score criteria, acid-based properties, and reactivity indexes allowed to reduce the set of derivatives and choose the most promissory candidates. These new molecules are hypothesized to be potent antioxidants, even better than sesamol.

Results and Discussion

Sesamol Derivatives and Selection Scores

A total amount of 68 derivatives were designed. For 18 of them, it was impossible to obtain the values of Ames mutagenicity and/or LD50, for that reason, only 50 derivatives were studied. Among them, 10 have only one added functional group, 34 have two, and six have three placed in R1, R2, and/or R3 positions. Table S7 describes them all, showing the corresponding labels and substituents. Their ADME properties and those for the reference set of molecules were calculated as previously described, and their values are presented in Tables S8 and S9, respectively. The estimated toxicity, expressed as LD50 and Ames mutagenicity (M); synthetic accessibility (SA), and selection scores (SS) are reported in Tables S10 and S11. Arithmetic averages were calculated for the reference set and used as reference for sesamol derivatives analyses. The selection score of the parent molecule was also used as a reference (Figure ).

Figure 1

Selection score (SS) for the sesamol derivatives designed in this work. Vertical lines mark the arithmetic mean of the reference set (red) and sesamol value (green).

It was found that most of the derivatives have higher selection scores than sesamol and also than the average value of the reference set. This indicates that we can expect them to have the desired properties of oral drugs, at least according to the analyzed properties. Taking into account that the highest values of SS are expected to be related to lower toxicity, better SA, and better ADME properties, the 12 derivatives with the highest selection scores were selected for the next stage of investigation. Their structures are shown in Scheme , and their elimination scores are reported in Table S12. For the reference set of molecules used here, the average and individual values of SE,ADME2 are in line with the previously reported ones for other medical drugs.[62,63] Individual values of SE for each derivative are shown in Figure , in ascending order.

Scheme 3

Structures and SS Values of the Subset of Sesamol Derivatives Selected as the Most Promising Ones Based on ADME Properties, Toxicity, and SA

Figure 2

Elimination score (SE) for the most promising sesamol derivatives.

Small increases in SE are associated with properties that do not have much weight in the deviations from the reference set, while large increases in SE are related to properties of greater relevance in the deviation. The SE,ADME2 criterion only contains two properties (log P and molecular weight) and most derivatives show similar values, which corroborates the need of using the remaining three SE criteria to further analyze the results. On the other hand, the additional six properties included in SE,ADME8 (PSA, HBD, HBA, XAt, RB, and MR) and toxicity (M and LD50) included in SE,ADMET have the highest contributions to the elimination score; while the SA, accounted for in SE,ADMETSA, has a small contribution to the deviations from the reference set. The individual contribution of each property to SE was also investigated; the results are shown in Table S13 and are also plotted in Figure to facilitate the analysis.

Figure 3

Individual contributions to the elimination score (SE) for the most promising sesamol derivatives.

The LD50 of the investigated derivatives is responsible for the largest deviations from the reference set of medical drugs. However, it seems worthwhile to analyze in detail the reason of such deviations. Please note that higher values of LD50 correspond to less toxic compounds, that is, a higher concentration of the compound would be required to induce death in rats. Most of the deviations in LD50 arise from values of the sesamol derivatives that are significantly higher than the average of the reference set. The exceptions are dS-40 and dS-43, which individual values have the opposite trend (LD50 = 361.93 and 746.16, respectively). However, even those derivatives have LD50 values that are significantly higher than the lower value in the reference set (LD50 = 44.98, bromocriptine). Moreover, according to the estimated LD50, dS-40, and dS-43 are less toxic to rats than ladostigil, lisuride, riluzole, and rivastigmine (Table S11).

When analyzing the Ames mutagenicity, lower values are associated with less mutagenic compounds. The average value of this property for the reference set is 0.41. This toxicity index was found to be lower for all the investigated derivatives, except for dS-48 (M = 0.44). However, several molecules in the reference set have M values higher than 0.44 including apomorphine, benserazide, cabergoline, dantrolene, and entacapone.

In addition, the values of PSA, HBD, HBA, XAt, RB, and MR properties reasonable deviations from the average of the reference set. In all cases, Lipinski, Veber, and Ghose criteria are still met.

Thus, based on these considerations, none of the 12 selected derivatives were eliminated as antioxidant candidates, that is, their toxicity, SA, bioavailability, and permeation behavior are not expected to be problematic. Selecting the most promising candidates, among them, would then be based on their reactivity indexes.

Electronic Calculations

The estimated pKa values of the 12 selected derivatives are reported in Table . The molar fractions of their acid-base species, at physiological pH (pH = 7.4) are listed in Table S14. The associated deprotonation routes and distribution diagrams are shown Figures S1 and S2, respectively. Those species with a nonnegligible molar fraction (Mf ≥ 10–3), at the pH of interest, were selected for the study of the remaining electronic properties. Figure shows the natural logarithm of sesamol derivatives molar fractions. The horizontal line represents the acceptation limit. The species located above the line are the ones selected for the next calculations.

Table 1

Estimated pKa Values of Sesamol Derivatives, at pH = 7.4

	pKa1	pKa2	pKa3	pKa4
dS-6	1.03	4.23	9.88
dS-9	2.57	8.79	12.79
dS-12	–0.29	4.40	12.61
dS-36	1.89	3.84	11.96
dS-40	1.69	11.99
dS-42	1.58	10.16	13.70
dS-43	1.25	7.56	13.83
dS-45	0.29	2.78	10.29	14.24
dS-46	3.36	8.01	11.70	13.82
dS-47	–0.12	5.13	10.13	14.00
dS-48	3.89	8.19	10.52	13.68
dS-49	7.80	10.11	12.88	15.42
Trolox	3.89	11.92

Figure 4

Analysis of the molar fractions (Mf) of sesamol derivatives acid-base species.

The values of the polar strength (PS) for the estimation of EI and EA, using the electron propagation theory (EPT), are reported in Table S15. All of them are in the acceptance range, which validates the calculations. The reactivity indexes of each derivative are provided in Table S16. A graphical tool, previously proposed,[64] and designed to simultaneously anticipate H and electron donor abilities was used (Figure ) to explore such capacities in the designed sesamol derivatives. It is known as the electron and hydrogen donating ability map for antioxidants (eH-DAMA). The species located at the left and bottom of this map are expected to be the best donors for both H and electron. Thus, they should be the more efficient for acting as antioxidants via hydrogen atom-transfer (HAT) and single electron-transfer (SET) mechanisms.

Figure 5

eH-DAMA for acid-base species of sesamol derivatives, sesamol, Trolox, and the oxidant the H2O2/O2•– pair.

SET reactions have the peculiarity that thermochemical considerations alone might lead to erroneous conclusions. This is because highly exergonic reactions (those involving donors with very low ionization energies) may be located in the inverted region of the Marcus parabola, that is, they may be very slow. Therefore, using IE directly to assess antioxidant ability, via SET, might be misleading. That is why in the eH-DAMA plot a different reactivity index is used to account for the electron donor ability of the investigated antioxidants, that is, the electrodonating power (ω–). It depends on the IE, but in a nonlinear way (Figure S3). Moreover, the shape of this dependence resembles that of the Marcus parabola. Species with very low IE, have high values of ω– and, therefore, they are not predicted to be efficient as free radical scavengers, via SET.[65,66] This could be the case of the dianionic species of sesamol derivatives.

Species with small ω– (at the bottom of the eH-DAMA) are expected to be particularly efficient as neutralizers of free radicals via SET. Species with low bond dissociation energies (BDE) (at the left of the eH-DAMA) are expected to act as free radical scavengers via hydrogen donation (HAT). The parent molecule, Trolox (an antioxidant frequently used as a reference regarding ROS neutralizing activity), and the oxidizing pair H2O2/O2•–, are also included in the map to facilitate comparisons and interpretation. According to the results shown in Figure , the monoanionic species of derivatives dS-48, dS-49, dS-9, dS-46, dS-47, dS-43, dS-45, and dS-42 were identified as the most efficient radical scavengers, through the above-mentioned chemical routes. They are expected to be better in neutralizing the hydroperoxyl pair than Trolox and sesamol. These compounds appear to be efficient through both mechanisms: SET and HAT. However, their relative abundance, under physiological conditions, is another important aspect to consider.

For the compounds best located in the eH-DAMA (dS-48, dS-49, dS-9, and dS-46), the neutral form is the most abundant at physiological pH, although the monoanions are expected to be present in nonnegligible amounts (Table S14). On the contrary, for the other four derivatives (dS-47, dS-43, dS-45, and dS-42) that are predicted to be better than Trolox and sesamol, as H and electron donors, the most abundant acid-base species at this pH are just the ones identified to be the key for the antioxidant behavior of the corresponding derivatives, that is, the monoanions. Thus, considering that neutral species cross biological barriers easier than charged ones that derivatives dS-48, dS-49, dS-9, and dS-46 are the best placed in eH-DAMA and the population of their anionic species are high enough (13.9, 28.5, 3.9, and 19.7%, respectively), they were selected as the most promising candidates to act as antioxidants by scavenging free radicals via HAT and SET.

Conclusions

Sesamol derivatives were computationally designed in a rational way, using a computer assisted protocol. Adding functional groups to the sesamol structure allowed to generate 50 compounds: 10 with one additional functional group and 34 with two and 6 derivatives with three. They were evaluated and compared to a reference set of currently used medical drugs using selection and elimination scores. This led to a fist selection of 12 sesamol derivatives.

The search was refined by electronic calculations, including estimations of pKas, and reactivity indexes that account for electron and H donation capabilities. The identification of the best donors allowed to propose four derivatives as the most promising candidates to act as chemical antioxidants, via HAT and SET. They are dS-48, dS-49, dS-9, and dS-46. They are predicted to be better antioxidants than sesamol and Trolox. Further investigations on these derivatives are still necessary to confirm or refute the proposal from this work.

Computational Methods

Design and First Evaluation of Sesamol Derivatives

Sesamol derivatives were built incorporating four functional groups (−OH, −NH2, −SH, and −COOH) in all the available positions of the phenolic ring: R1, R2 and R3 (Scheme ). All combinations of substitution in positions R1, R2, and R3, give the possibility to obtain mono-, di-, and trisubstituted sesamol derivatives. The added functional groups only represent a modest structural modification, which could help to preserve the desirable features of sesamol. They are explored to investigate if they surpass sesamol as free radical scavengers via hydrogen or electron donation, considering at the same time their bioavailability and permeation behavior necessary for oral drugs.

Scheme 4

Substitution Sites R1, R2 y R3 and the Original Sites of Sesamol a y b

To evaluate the derivatives in terms of their potential use as oral drugs, different properties were considered, namely permeability, toxicity, and SA. A reference set of drugs was used to compare and put the behavior of the investigated derivatives into context. The reference set consists of 35 neuroprotectors already in use (Table S1, Supporting Information). This set has been successfully used in a previous work.[64] The parameters described below are calculated in the same way for derivatives and the reference set.

ADME properties are closely related to three important empirical rules that allow predicting bioavailability, solubility, and permeability of oral drugs, based on Lipinski,[67] Ghose,[68] and Veber[69] criteria (Table S2). The calculated properties were the number of H bond donors (HBD), number of H bond acceptors (HBA), molecular weight (MW), octanol/water partition coefficient (Log P), molar refractivity (MR), number of non-hydrogen atoms (XAt), routable bonds (RB), and polar surface area (PSA).

The physicochemical properties were calculated using online software Molinspiration Property Calculation Service,a while the molar refractivity (MR) was estimated using DruLiTob software. Toxicity indexes (M: Ames mutagenicity and LD50 (mg/kg): lethal dose 50) were calculated with the Toxicity Estimation Software Tool (T.E.S.T.) version 4.1. SA is another important aspect to consider when designing new compounds in silico. It was calculated using the SYLVIA-XTc 1.4 program (Molecular Networks, Erlangen, Germany).[70] It classifies chemical compounds on a scale from 1 to 10 with the highest values representing the most difficult synthesis.[71]

A selection score criteria (SS) was used to assign a unique number to each derivative that serves to assess their likeliness as oral drugs. This criterion includes all the evaluated properties (Table S3) and was also used for the reference set. The higher the SS score the better drug-like properties. This score is general, so it could mask a case with high value of SS but with an undesired value for a single property. For double-checking if any molecule in the selected subset significantly deviates from the target values, an elimination score (SE) was also used (Table S4).

Electronic calculations were necessary to explore the antioxidant behavior of derivatives and their most abundant species (cationic, neutral, mono, di, and trianionic). Geometry optimizations and frequency calculations were performed with Gaussian 09.[72] To that purpose the density functional theory, in particular the M05-2X approximation,[73] was used combined with the 6-311+G(d,p) basis set and the solvation model density.[74] All data were obtained at 298 K. Local minima were identified by the absence of imaginary frequencies.

The molecules designed in this work have not been synthesized, so there is no knowledge of their acidity constants. This is an important property not only in the process of diffusion through lipid membranes but also in the antioxidant behavior that largely depends on the deprotonation degree. Therefore, they were theoretically estimated. To that purpose, the fitted parameters approach method was used.[75,76] This method consists in the calculation of pKas using the linear fitting expression pKa = mΔGBA + C0. The ΔGBA value is the Gibbs free energy difference between the conjugated base and the corresponding acid (GA– – GHA). Parameters m and C0 change considering the substituents and the level of calculation used. Details on this strategy for our calculation can be found in Table S5.

Several reactivity indexes were estimated for sesamol derivatives (Table S6). The ionization energies (IE) and the electron affinities (EA) were calculated using the EPT,[77] which is known to produce similar values to those experimentally determined. In particular, the partial third-order quasiparticle theory (P3) was used because of its low mean error, for open-shell systems, when comparing to other methods.[78] To validate the EPT calculation, the value of the pole strength (PS) was considered and checked to be larger than 0.80.[79,80] The indexes derived from IE and EA were used to estimate electron donor abilities, which are related to electron-transfer processes.

On the other hand, BDE were calculated for anticipating HAT capabilities. All hydrogen atoms in the molecule that could act as donors were studied. They are those in positions a and b belonging to the framework of the parent molecule as well as those in the groups placed at the substitution sites R1, R2 and R3 (Scheme ).