Identification of Antioxidant Methyl Derivatives of Ortho-Carbonyl Hydroquinones That Reduce Caco-2 Cell Energetic Metabolism and Alpha-Glucosidase Activity

α-glucosidase, a pharmacological target for type 2 diabetes mellitus (T2DM), is present in the intestinal brush border membrane and catalyzes the hydrolysis of sugar linkages during carbohydrate digestion. Since α-glucosidase inhibitors (AGIs) modulate intestinal metabolism, they may influence oxidative stress and glycolysis inhibition, potentially addressing intestinal dysfunction associated with T2DM. Herein, we report on a study of an ortho-carbonyl substituted hydroquinone series, whose members differ only in the number and position of methyl groups on a common scaffold, on radical-scavenging activities (ORAC assay) and correlate them with some parameters obtained by density functional theory (DFT) analysis. These compounds’ effect on enzymatic activity, their molecular modeling on α-glucosidase, and their impact on the mitochondrial respiration and glycolysis of the intestinal Caco-2 cell line were evaluated. Three groups of compounds, according their effects on the Caco-2 cells metabolism, were characterized: group A (compounds 2, 3, 5, 8, 9, and 10) reduces the glycolysis, group B (compounds 1 and 6) reduces the basal mitochondrial oxygen consumption rate (OCR) and increases the extracellular acidification rate (ECAR), suggesting that it induces a metabolic remodeling toward glycolysis, and group C (compounds 4 and 7) increases the glycolysis lacking effect on OCR. Compounds 5 and 10 were more potent as α-glucosidase inhibitors (AGIs) than acarbose, a well-known AGI with clinical use. Moreover, compound 5 was an OCR/ECAR inhibitor, and compound 10 was a dual agent, increasing the proton leak-driven OCR and inhibiting the maximal electron transport flux. Additionally, menadione-induced ROS production was prevented by compound 5 in Caco-2 cells. These results reveal that slight structural variations in a hydroquinone scaffold led to diverse antioxidant capability, α-glucosidase inhibition, and the regulation of mitochondrial bioenergetics in Caco-2 cells, which may be useful in the design of new drugs for T2DM and metabolic syndrome.


Introduction
Diabetes mellitus (DM) is a metabolic chronic disease, highly prevalent in both the developing and developed world [1].It is characterized by inadequate glycemic control, caused by either reduced insulin secretion (type 1 diabetes mellitus, T1DM) or by a reduced cellular response to insulin (type 2 diabetes mellitus, T2DM) [2][3][4], which is the main type of diabetes, with a prevalence of the 90-95% of the total incidence [5].In T2DM, enhanced ROS production via reduced intracellular antioxidant levels is observed [6,7], which, at the same time, accelerates the dysregulation of glucose metabolism and tissue damage through signal transductions.
Molecular crosstalk between microRNAs and oxidative stress in the pathogenesis of T2DM and its related health problems has been pointed out [6].Also, recently, it has been suggested that miRNAs and oxidative stress influence each other.Oxidative stress seems to regulate the function and biogenesis of several miRNAs and vice versa.An imbalance in miRNA expression can lead to oxidative stress by promoting free radical generation and (or) reducing the endogenous antioxidant capacity [8][9][10].
It is known that dietary polyphenols modulate post-prandial increases in glucose levels, intestinal integrity, and oxidative damage [22][23][24].Consequently, a growing interest in this type of compound has been observed in the last few years.It has been suggested that oxidative energy metabolism plays a crucial role in mitochondrial ATP production and in maintaining a gut barrier with high integrity [25].The disruption of this intestinal barrier function is related to inflammation [26] and dysmetabolic conditions, including T2DM [27].In this context, the human intestinal epithelial Caco-2 cell line has been used to assess the relationship between mitochondrial ATP production and intestinal permeability [28].
Previously, we have reported an ortho-carbonyl hydroquinone series that disrupts mitochondrial bioenergetics in cancer cells by Complex I inhibition and oxidative phosphorylation uncoupling without affecting non-cancer cells [29][30][31].On the other hand, it has been described that certain hydroquinones and quinones can induce redox cycling [32][33][34][35]; however, this is not the case with these hydroquinones [30].Since they exhibit a high antioxidant capability in cell-free analysis [36] and, as has been discussed above, oxidative stress is involved in the intestinal dysfunction associated with T2DM, this work aims to study the antioxidant capacity of the methyl derivatives of a bicyclic ortho-carbonyl hydroquinone (Table 1), which differs only in the number and position of methyl groups, and its effects as an AGI, and in modulating the energetic metabolism of Caco-2 cells.

Introduction
Diabetes mellitus (DM) is a metabolic chronic disease, highly prevalent in both the developing and developed world [1].It is characterized by inadequate glycemic control, caused by either reduced insulin secretion (type 1 diabetes mellitus, T1DM) or by a reduced cellular response to insulin (type 2 diabetes mellitus, T2DM) [2][3][4], which is the main type of diabetes, with a prevalence of the 90-95% of the total incidence [5].In T2DM, enhanced ROS production via reduced intracellular antioxidant levels is observed [6,7], which, at the same time, accelerates the dysregulation of glucose metabolism and tissue damage through signal transductions.
Molecular crosstalk between microRNAs and oxidative stress in the pathogenesis of T2DM and its related health problems has been pointed out [6].Also, recently, it has been suggested that miRNAs and oxidative stress influence each other.Oxidative stress seems to regulate the function and biogenesis of several miRNAs and vice versa.An imbalance in miRNA expression can lead to oxidative stress by promoting free radical generation and (or) reducing the endogenous antioxidant capacity [8][9][10].
It is known that dietary polyphenols modulate post-prandial increases in glucose levels, intestinal integrity, and oxidative damage [22][23][24].Consequently, a growing interest in this type of compound has been observed in the last few years.It has been suggested that oxidative energy metabolism plays a crucial role in mitochondrial ATP production and in maintaining a gut barrier with high integrity [25].The disruption of this intestinal barrier function is related to inflammation [26] and dysmetabolic conditions, including T2DM [27].In this context, the human intestinal epithelial Caco-2 cell line has been used to assess the relationship between mitochondrial ATP production and intestinal permeability [28].
Previously, we have reported an ortho-carbonyl hydroquinone series that disrupts mitochondrial bioenergetics in cancer cells by Complex I inhibition and oxidative phosphorylation uncoupling without affecting non-cancer cells [29][30][31].On the other hand, it has been described that certain hydroquinones and quinones can induce redox cycling [32][33][34][35]; however, this is not the case with these hydroquinones [30].Since they exhibit a high antioxidant capability in cell-free analysis [36] and, as has been discussed above, oxidative stress is involved in the intestinal dysfunction associated with T2DM, this work aims to study the antioxidant capacity of the methyl derivatives of a bicyclic ortho-carbonyl hydroquinone (Table 1), which differs only in the number and position of methyl groups, and its effects as an AGI, and in modulating the energetic metabolism of Caco-2 cells.

Synthesis of the Compounds
The new hydroquinones (Scheme 1) were obtained in two steps.The first consists of obtaining acylhydroquinones following the reported method [37,38] with a slight modification.To a Reaction Vial G10 (Anton-Paar, Graz, Austria) equipped with a stir bar, the following were added: 500 mg of corresponding hydroquinone, a 1.5 equivalent of the corresponding carboxylic acid, and 3 mL of Boron trifluoride dihydrate.The mixture was heated at 160 • C for 30 min in a Monowave 50 reactor (Anton-Paar).The mixture was allowed to cool to room temperature, neutralized with saturated sodium bicarbonate solution, and extracted with ethyl acetate.The extract was washed with distilled water, dried with anhydrous sodium sulfate, and then filtered and concentrated in a rotavapor.Afterward, the corresponding acylhydroquinone was purified by flash chromatography with hexane/ethyl acetate 8:1 v/v as an eluent.The second step was performed following reference [39], although this was also done with slight variations.A mixture of corresponding acylhydroquinone (1 equiv) and Ag 2 O (2.5 equiv) in 30 mL of dichloromethane was vigorously stirred for 1-2 h at r.t., yielding the corresponding quinone.This mixture was filtered through celite and added dropwise to a solution of the 4-(2-methyl-2-propenyl)morpholine or 4-(2-ethylbut-1-en-1-yl)morpholine at r.t. for 2 h, being monitored by thin-layer chromatography.Then, the solvent was evaporated under reduced pressure, and the residue was dissolved in a mixture of ethanol and hydrochloric acid and refluxed by 1 h.Next, it was poured on an ice/water mixture, and the product was extracted with 3 portions of 20 mL of ethyl acetate, which were dried with anhydrous sodium sulfate and then filtered; the solvent was evaporated under reduced pressure and purified by flash chromatography using hexane/ethyl acetate 4:1 v/v as an eluent.

Synthesis of the Compounds
The new hydroquinones (Scheme 1) were obtained in two steps.The first consists of obtaining acylhydroquinones following the reported method [37,38] with a slight modification.To a Reaction Vial G10 (Anton-Paar, Graz, Austria) equipped with a stir bar, the following were added: 500 mg of corresponding hydroquinone, a 1.5 equivalent of the corresponding carboxylic acid, and 3 mL of Boron trifluoride dihydrate.The mixture was heated at 160 °C for 30 min in a Monowave 50 reactor (Anton-Paar).The mixture was allowed to cool to room temperature, neutralized with saturated sodium bicarbonate solution, and extracted with ethyl acetate.The extract was washed with distilled water, dried with anhydrous sodium sulfate, and then filtered and concentrated in a rotavapor.Afterward, the corresponding acylhydroquinone was purified by flash chromatography with hexane/ethyl acetate 8:1 v/v as an eluent.The second step was performed following reference [39], although this was also done with slight variations.A mixture of corresponding acylhydroquinone (1 equiv) and Ag2O (2.5 equiv) in 30 mL of dichloromethane was vigorously stirred for 1-2 h at r.t., yielding the corresponding quinone.This mixture was filtered through celite and added dropwise to a solution of the 4-(2-methyl-2-propenyl)morpholine or 4-(2-ethylbut-1-en-1-yl)morpholine at r.t. for 2 h, being monitored by thin-layer chromatography.Then, the solvent was evaporated under reduced pressure, and the residue was dissolved in a mixture of ethanol and hydrochloric acid and refluxed by 1 h.Next, it was poured on an ice/water mixture, and the product was extracted with 3 portions of 20 mL of ethyl acetate, which were dried with anhydrous sodium sulfate and then filtered; the solvent was evaporated under reduced pressure and purified by flash chromatography using hexane/ethyl acetate 4:1 v/v as an eluent.
Scheme 1. Synthesis of methyl derivatives of ortho-carbonyl hydroquinones.R corresponds to those shown in Table 1.

In Vitro Antioxidant Capacity
The radical-scavenging activities of the hydroquinone series against peroxyl radicals were assessed (ORAC assays).The results are presented in Table 2.For the series with the gem-dimethyl on ring B (compounds 1-5), compound 1 showed the highest antioxidant capacity (8.8 + 0.2 mol TE/mol comp).Similarly, for the series with the gem-diethyl group on ring B (compounds 6-10), compound 6 presents the highest antioxidant capacity (7.0 + 0.3 mol TE/mol compound).These results indicate that introducing methyl substituents Scheme 1. Synthesis of methyl derivatives of ortho-carbonyl hydroquinones.R corresponds to those shown in Table 1.

In Vitro Antioxidant Capacity
The radical-scavenging activities of the hydroquinone series against peroxyl radicals were assessed (ORAC assays).The results are presented in Table 2.For the series with the gem-dimethyl on ring B (compounds 1-5), compound 1 showed the highest antioxidant capacity (8.8 + 0.2 mol TE/mol comp).Similarly, for the series with the gem-diethyl group on ring B (compounds 6-10), compound 6 presents the highest antioxidant capacity (7.0 + 0.3 mol TE/mol compound).These results indicate that introducing methyl substituents such as R1, R2, and R3 decreases the antioxidant capacity of the compounds.The replacement of the gem-dimethyl with the gem-diethyl group also decreases the antioxidant capacity.

Calculation of Main Thermodynamic Parameters
To obtain insight into the mechanism responsible for the antioxidant capacity of the compounds, we carried out DFT calculations.We considered the three mechanisms commonly associated with radical scavenging [40]: Hydrogen atom transfer (HAT), sequential electron transfer-proton transfer (SET-PT), and sequential proton loss-electron transfer (SPLET).These mechanisms can occur in parallel but at different rates.The thermodynamic parameters associated with each mechanism can help to assess their feasibility.The HAT mechanism is associated with the bond dissociation enthalpy (BDE) [40].The SET-PT mechanism is associated with ionization potential (IP) and proton dissociation enthalpy (PDE) [40].The SPLET mechanism is associated with proton affinity (PA) and electron transfer enthalpy (ETE) [40].We evaluated the BDE for both OH groups (1 and 2 in Table 2) present in all compounds (BDE 1 and BDE 2 in Table 2).The highest values correspond to BDE

Calculation of Main Thermodynamic Parameters
To obtain insight into the mechanism responsible for the antioxidant capacity of the compounds, we carried out DFT calculations.We considered the three mechanisms commonly associated with radical scavenging [40]: Hydrogen atom transfer (HAT), sequential electron transfer-proton transfer (SET-PT), and sequential proton loss-electron transfer (SPLET).These mechanisms can occur in parallel but at different rates.The thermodynamic parameters associated with each mechanism can help to assess their feasibility.The HAT mechanism is associated with the bond dissociation enthalpy (BDE) [40].The SET-PT mechanism is associated with ionization potential (IP) and proton dissociation enthalpy (PDE) [40].The SPLET mechanism is associated with proton affinity (PA) and electron transfer enthalpy (ETE) [40].We evaluated the BDE for both OH groups (1 and 2 in Table 2) present in all compounds (BDE 1 and BDE 2 in Table 2).The highest values correspond to BDE    To correlate to the experimental antioxidant capacity mechanism in relation to thermodynamic parameters, two approaches can be employed [41][42][43].Some reports consider the value of the parameter as the key factor associated with a mechanism [41,42]; namely, the lower value, the higher their contribution of the mechanism.However, an alternative To correlate to the experimental antioxidant capacity mechanism in relation to thermodynamic parameters, two approaches can be employed [41][42][43].Some reports consider the value of the parameter as the key factor associated with a mechanism [41,42]; namely, the lower value, the higher their contribution of the mechanism.However, an alternative method is to consider the correlation between the experimental values of assays and thermodynamic parameters [43].Following the first approach, the most favorable mechanism corresponds to the HAT, with the H-atom abstraction corresponding to the hydroxyl 2. Looking into the correlation between ORAC and the thermodynamic parameters (Figure 1), we found that the best correlation corresponds to the IP.However, the BDEs and IP slopes are positive (Figure 1A-C), indicating that high values of the thermodynamic parameter (theoretically a less potent antioxidant) correspond to the high values of ORAC (experimentally stronger antioxidants).
We also consider the hydroquinone anions, with the compounds deprotonated at hydroxyl 2, as contributing antioxidant species.The BDE for the O-H 1, obtained in this form, is presented in Table 2.These values were lower than those of neutral species.However, the correlation exhibited the same tendency regarding the experimental ORAC values (Figure 1D,E).These results suggest that the variation of the antioxidant capacity of the compounds is influenced by a factor not considered in the calculations.These variables (BDE, IP, and PA) do not consider the kinetic factors that can influence the experimental antioxidant capacity, and the steric hindrance in particular can lead to a reduction in the antioxidant capacity [44][45][46].In this case, the increase in the number of the methyl groups leads to a variation in the electronic density that should favor the antioxidant capacity; however, it also increases the steric hindrance to access the -OH group, which can explain the tendency observed in the ORAC results.
method is to consider the correlation between the experimental values of assays and thermodynamic parameters [43].Following the first approach, the most favorable mechanism corresponds to the HAT, with the H-atom abstraction corresponding to the hydroxyl 2. Looking into the correlation between ORAC and the thermodynamic parameters (Figure 1), we found that the best correlation corresponds to the IP.However, the BDEs and IP slopes are positive (Figure 1A-C), indicating that high values of the thermodynamic parameter (theoretically a less potent antioxidant) correspond to the high values of ORAC (experimentally stronger antioxidants).We also consider the hydroquinone anions, with the compounds deprotonated at hydroxyl 2, as contributing antioxidant species.The BDE for the O-H 1, obtained in this form, is presented in Table 2.These values were lower than those of neutral species.However, the correlation exhibited the same tendency regarding the experimental ORAC values (Figure 1D,E).These results suggest that the variation of the antioxidant capacity of the compounds is influenced by a factor not considered in the calculations.These variables (BDE, IP, and PA) do not consider the kinetic factors that can influence the experimental antioxidant capacity, and the steric hindrance in particular can lead to a reduction in the antioxidant capacity [44][45][46].In this case, the increase in the number of the methyl groups leads to a variation in the electronic density that should favor the antioxidant capacity; however, it also increases the steric hindrance to access the -OH group, which can explain the tendency observed in the ORAC results.

In Vitro and In Silico α-Glucosidase Inhibitory Activity
The effect of synthesized hydroquinones against α-glucosidase activity was evaluated (Table 2).Most of the evaluated compounds showed high inhibitory activity, with six of the ten compounds showing IC50 values lower than the well-known AGI acarbose (IC50 = 194.0+ 1.7 µM).Compounds 8 and 10 presented the highest inhibitory potency, with

In Vitro and In Silico α-Glucosidase Inhibitory Activity
The effect of synthesized hydroquinones against α-glucosidase activity was evaluated (Table 2).Most of the evaluated compounds showed high inhibitory activity, with six of the ten compounds showing IC 50 values lower than the well-known AGI acarbose (IC 50 = 194.0+ 1.7 µM).Compounds 8 and 10 presented the highest inhibitory potency, with IC 50 values of 60.0 + 1.5 µM and 78.0 + 1.1 µM, respectively, and compound 9 was the worst.These results suggest that the gem-diethyl group in ring B is a key structural factor in the inhibitory potency of the compounds, but the presence of methyl substituents in the rest of the structure needs detailed examination due to their variable effect.
To obtain a detailed explanation of the structure-activity relationship observed in the experimental inhibitory results, we carried out molecular docking.Several conformers and poses for each molecule were analyzed.The binding site was defined by the crystallographic structure obtained for α-glucosidase in complex with acarbose (PDB code: 3W37).The results were used as an input for an MMGBSA free energy calculation study.Figure 2A shows the values of MMGBSA ∆Gbind and the IC 50 presented as pIC 50 = -log[IC 50 ], and the correlation between them is shown in Figure 2B.The correlation between experimental pIC 50 values and calculated MMGBSA ∆Gbind gives a R2 = 0.71.
Figure 2 shows 3D maps of the binding interactions of compound 3 (Figure 2C), compound 10 (Figure 2D), and acarbose (Figure 2E) with the α-glucosidase enzyme.Acarbose, which corresponds to the model ligand to α-glucosidase, shows a three-center hydrogen bond interaction between one of their hydroxyl groups with aspartic acid 232 and arginine 552, a hydrogen bond interaction between another of their hydroxyls and aspartic acid 568, and a hydrogen bond interaction between two other hydroxyls and histidine 626.Additional hydrophobic interactions of acarbose with tryptophan 432 and phenylalanine 601 are also identified.Compound 3, which presented the highest MMGBSA ∆G bind values (lowest affinity), presents a three-center hydrogen bond interaction between one of their hydroxyls with aspartic acid 232 and arginine 552, like acarbose, hydrophobic interactions with phenylalanine 601 and phenylalanine 476, and π-π stacking interactions with tryptophan 432.Compound 10, instead, with the lowest MMGBSA ∆Gbind value (highest affinity), shows a hydrogen bond between one of their hydroxyls and aspartic acid 568.In addition, it presents hydrophobic interactions with aspartic acid 630, arginine 629, phenylalanine 476, and methionine 470, and additionally a π-π stacking interaction with tryptophan 432.
IC50 values of 60.0 + 1.5 µM and 78.0 + 1.1 µM, respectively, and compound 9 was the worst.These results suggest that the gem-diethyl group in ring B is a key structural factor in the inhibitory potency of the compounds, but the presence of methyl substituents in the rest of the structure needs detailed examination due to their variable effect.
To obtain a detailed explanation of the structure-activity relationship observed in the experimental inhibitory results, we carried out molecular docking.Several conformers and poses for each molecule were analyzed.The binding site was defined by the crystallographic structure obtained for α-glucosidase in complex with acarbose (PDB code: 3W37).The results were used as an input for an MMGBSA free energy calculation study.Figure 2A shows the values of MMGBSA ΔGbind and the IC50 presented as pIC50 =log[IC50], and the correlation between them is shown in Figure 2B.The correlation between experimental pIC50 values and calculated MMGBSA ΔGbind gives a R2 = 0.71. Figure 2 shows 3D maps of the binding interactions of compound 3 (Figure 2C), compound 10 (Figure 2D), and acarbose (Figure 2E) with the α-glucosidase enzyme.Acarbose, which corresponds to the model ligand to α-glucosidase, shows a three-center hydrogen bond interaction between one of their hydroxyl groups with aspartic acid 232 and arginine 552, a hydrogen bond interaction between another of their hydroxyls and aspartic acid MMGBSA binding free energies were estimated for the complex of the enzyme with compounds 3 and 10 (Table 3).Compound 10 shows a stronger net binding energy (∆G bind ) compared with compound 3 (−24.19vs. 6.00 kcal/mol).For both compounds, the Van der Waals energy dominated over all other energies.The favorable contribution of lipophilic and electrostatic binding energy (∆G bind LP and ∆G bind CB) is similar for both compounds.The most remarkable difference is the unfavorable contribution of the generalized Born electrostatic solvation energy (∆G bind GB) for compound 3 compared with compound 10.We evaluate the effect of ortho-carbonyl hydroquinones 1-10 on the intestinal Caco-2 cell line at 24 h of treatment.As Figure 3A shows, compounds 5 and 10 significantly increased the MTT reduction, suggesting that these compounds increase the cell number.Consistent with the antiproliferative effect previously described by us [30], compound 6, instead, decreased the MTT reduction.The other compounds of this series did not show significant changes in Caco-2 cell experiments.We previously described how ortho-carbonyl hydroquinone scaffolds inhibit oxidative phosphorylation in breast cancer cells [29][30][31]37].Since the reduction in glucose utilization and metabolism in intestinal cells is a promising target for T2DM treatment, the effect of these compounds on mitochondrial respiration and glycolysis in Caco-2 cells was determined.As Figure 3B shows, three groups of compounds were identified in a dot-plot from basal mitochondrial OCR and glycolysis (ECAR).Compared to the control (red dot), group A (compounds 2, 3, 5, 8, 9, and 10) reduces the glycolysis.Group B (compounds 1 and 6) reduces the basal mitoOCR and increases ECAR, suggesting that it induces a metabolic remodeling toward glycolysis, and group C (compounds 4 and 7) increases the glycolysis lacking effect on OCR.The effect of ortho-carbonyl hydroquinone series (100 µM) on five parameters of mitochondrial respiration (basal OCR, max OCR, spare capacity, We previously described how ortho-carbonyl hydroquinone scaffolds inhibit oxidative phosphorylation in breast cancer cells [29][30][31]37].Since the reduction in glucose utilization and metabolism in intestinal cells is a promising target for T2DM treatment, the effect of these compounds on mitochondrial respiration and glycolysis in Caco-2 cells was determined.As Figure 3B shows, three groups of compounds were identified in a dot-plot from basal mitochondrial OCR and glycolysis (ECAR).Compared to the control (red dot), group A (compounds 2, 3, 5, 8, 9, and 10) reduces the glycolysis.Group B (compounds 1 and 6) reduces the basal mitoOCR and increases ECAR, suggesting that it induces a metabolic remodeling toward glycolysis, and group C (compounds 4 and 7) increases the glycolysis lacking effect on OCR.The effect of ortho-carbonyl hydroquinone series (100 µM) on five parameters of mitochondrial respiration (basal OCR, max OCR, spare capacity, ATP-linked OCR, and proton leak-driven OCR), and two of glycolysis (glycolytic capacity and glycolysis), is represented in a dendrogram (Figure 3C).Interestingly, two clusters of compounds are shown with differential effects on maximal OCR and glycolysis.Cluster 1 is composed of glycolysis enhancers with a strong inhibitory effect on max.OCR: compounds 1, 6, 7, and 9. Cluster 2 is composed of glycolysis inhibitors with a reduced effect on max OCR: compounds 2, 3, 4, 5, 8, and 10.No correlations between OCR and ECAR parameters, α-glucosidase activity inhibition, or ORAC were observed (Supplementary Figure S1).

Compounds 5 and 10 Differentially Act on Mitochondrial Respiration, Glycolysis, and ROS Production in Caco-2 Cells
Based on the above results that identify compounds 5 and 10 as among the most active AGIs, we evaluated the effect of both compounds on mitochondrial respiration and glycolysis (Figure 4).Compound 5 reduces the basal and max OCR values without affecting the proton leak-driven OCR.In contrast, compound 10 increases the proton leak-OCR and decreases max OCR (Figure 4A-D).Both compounds significantly reduce the glycolysis (Figure 4E-H).No correlations between OCR and ECAR parameters, α-glucosidase activity inhibition, or ORAC were observed (Supplementary Figure S1).

Compounds 5 and 10 Differentially Act on Mitochondrial Respiration, Glycolysis, and ROS Production in Caco-2 Cells
Based on the above results that identify compounds 5 and 10 as among the most active AGIs, we evaluated the effect of both compounds on mitochondrial respiration and glycolysis (Figure 4).Compound 5 reduces the basal and max OCR values without affecting the proton leak-driven OCR.In contrast, compound 10 increases the proton leak-OCR and decreases max OCR (Figure 4A-D).Both compounds significantly reduce the glycolysis (Figure 4E-H).Using menadione as an ROS inducer [47,48], the preventive effect of compounds 5 and 10 on mitochondrial and cytosolic ROS production was evaluated.No changes in the mitochondrial ROS production were observed (Figure 4I).However, menadione-dependent cytosolic ROS production was significantly prevented by compound 5 (Figure 4J).Collectively, the results suggest that compounds 5 and 10 are AGIs that also reduce glycolysis and mitochondrial respiration, and compound 5 has a cytosolic antioxidant effect.

Discussion
The T2DM pathogenesis involves alterations in nutrient uptake and energetic metabolism, in which high oxidative stress and inflammation mediate altered cell signaling [49].Intestinal cells participate in the nutrient uptake and the catabolism of carbohydrates by the α-glucosidase enzyme, which is a current target for T2DM treatment [50].We evaluated the antioxidant capacity of methyl derivatives of a bicyclic ortho-carbonyl hydroquinone, which differ only in the number and position of methyl groups, and their effects as AGIs, and in modulating the energetic metabolism of Caco-2 cells.The antioxidant capacity results suggest that the increase in the methyl groups reduce antioxidant capacity due to the steric hindrance to access the -OH group, despite how the electronic effects should favor it.The docking results suggest that the gem-diethyl group in ring B is a key structural factor in the inhibitory potency of the compounds.Compounds 5 and 10 were among the most potent inhibitors of α-glucosidase and produced an increase in the cell number.These compounds correspond to those fully methylated at R 1 , R 2 , and R 3 , with gem dimethyl and gem diethyl at R 4 and R 5 , respectively (Table 1).Beyond clinical AGI drugs used in T2DM, it has been reported that new naturally occurring compounds act as AGIs [51,52], also producing a reduction in glucose-transporter-encoding gene expression and glucose uptake in vitro and in pre-clinical models [53].Our results at the cellular level allow us to identify three groups of compounds with differential actions on OXPHOS and the glycolysis of Caco-2 cells.Group A reduced the glycolysis, group B reduced the basal OCR and increased ECAR (metabolic remodeling-inducers), and group C increased the glycolysis lacking effect on OCR.Previously, we described that small structural modifications to the ortho-carbonyl hydroquinone scaffold studied produce Complex I inhibitors and OXPHOS uncouplers by a protonophoric mechanism [29][30][31] based on modification in the gem-dimethyl/ethyl substitution.Moreover, recent studies on the corresponding quinones showed that these methyl substitutions influence redox potential [54] and the ability to inhibit the mitochondrial electron transport chain [55].In particular, FRV-1, an ortho-carbonyl substituted quinone, was identified as a new inhibitor of the deactive conformational status of Complex I [55].Notably, our results in Caco-2 cells expand the information available on the structure-mitochondrial activity relationships of ortho-carbonyl substituted compounds.
Our results show that in Caco-2 cells, compound 5 was an OCR/ECAR inhibitor and compound 10 was a dual agent, increasing the proton leak-driven OCR and inhibiting the maximal electron transport flux.Despite this different action in OXPHOS, both compounds reduced the glycolysis, suggesting that compounds 5 and 10 could be useful for reducing intestinal glucose uptake and metabolization under a potential T2DM condition.

Conclusions
We found that the presence of methyl substituents leads to a reduction in the compounds' antioxidant capacity, and concomitantly an increase in their capacity for AGIs.Fully methylated ortho-carbonyl hydroquinones 5 and 10 present lower antioxidant capacity, and potent capacity as AGIs, and they affect the energetic metabolism and reduce glycolysis, promoting the proliferation of Caco-2 cells.Compound 6 decreased the proliferation of Caco-2 cells, while the other compounds in this series did not show significant changes.Compound 5 was an OCR/ECAR inhibitor in Caco-2 cells, while compound 10 was a dual agent, increasing the proton leak-driven OCR and inhibiting the maximal electron transport flux.Our results reveal that slight structural variations in a hydroquinone scaffold led to diverse antioxidant capability, α-glucosidase inhibition, and the regulation of mitochon-drial bioenergetics in Caco-2 cells.These findings may be useful for designing new drugs for T2DM and metabolic syndrome.

ORAC Assays
The antioxidant activity was determined by ORAC (oxygen radical absorbance capacity) assays [61].Briefly, the antioxidant capacity for ORAC-FL was estimated by measuring the changes in fluorescence after 120 min of reaction with the radical.Assays were performed using a 96-well microplate in a Synergy H1 hybrid multi-mode microplate reader (Biotek, Winooski, VT, USA), and the results were expressed as µmol Trolox 100 g −1 DW.Three replicates were accomplished for each analysis.

α-Glucosidase Inhibitory Activity
The inhibition of α-glucosidase was determined following a previous method [62] using p-nitrophenyl α-D-glucopyranoside as a substrate.The reaction mixture contained 160 µL of 0.1 M sodium phosphate buffer (pH = 6.9), 5 µL of enzyme (5.46 U/mL), and polyphenolic extract (0.17-1.36 µg/mL).After the pre-incubation of the reaction mixture on ice for 5 min, the enzyme reaction was started by adding 5 µL of 25 mM p-nitrophenyl α-D-glucopyranoside into this mixture.The reaction was incubated for 15 min at 37 • C.Then, 80 µL of 0.2 M sodium carbonate was added.The absorbance was measured at 405 nm in a microplate reader (Biotek ELx808, Agilent, Santa Clara, CA, USA).Enzyme inhibition was calculated using the following equation: % inhibition = (A 0 − A s )/A 0 × 100, where A 0 is the absorbance of the control (blank, without extract) and A s is the absorbance in presence of the extract.IC 50 values denote the µg GAE/mL required to inhibit the enzyme by 50%.

Quantum Chemical Calculations
Calculations were achieved using the Gaussian 09 software [63], at density functional theory (DFT) M06-2X/6-311+G(d,p) level.No imaginary vibrational frequencies were found at the optimized geometries, indicating that they are the true minimal of the potential energy surface.The calculated thermodynamic parameters were obtained as follows:

•
Bond dissociation enthalpy (BDE): H(RO  − ).where H(RO • ), H(ROH), and H(ROH •+ ) correspond to the enthalpies of the neutral radical and the neutral and radical cation H(RO − ) of the compounds, while H(e − ), H(H + ) and H(H • ) correspond to enthalpies of the electron, proton, and radical hydrogen.

Molecular Docking
Molecules 1 to 10 were sketched and prepared using Maestro and LigPrep (Schrödinger Release 2022-1: Maestro, LigPrep Schrödinger, LLC, New York, NY, USA, 2021), generating all molecules without adding charges.ConfGen [64] was used to generate at most 10 conformers per molecule.Protein-ligand docking was performed using Glide software (version 9.0, Schrödinger, LLC, New York, NY, USA, 2021) [65,66].The binding site was defined by the crystallographic structure obtained for alpha-glucosidase in a complex with acarbose (PDB code: 3W37).Before docking calculations, the protein was prepared using Maestro Protein Preparation Wizard [67] by removing ligands, metals, and water molecules; adding hydrogens; and ionizing residues at pH 7.4.the minimization of the protein structure was conducted with an OPLS4 force field.The molecules were then prepared using the OPLS force field to minimize energy.The grid box was defined using the ligand cocrystallized in α-glucosidase as the center of the box.The docking was performed with the Glide standard precision (SP) function [68] and the top-10 poses per docked ligands were selected and subjected to postprocessing and rescoring with the molecular mechanicsgeneralized Born surface area (MM-GBSA) with Prime [69].This computational method combines molecular mechanics energy and implicit solvation models, which enable the docking results to be rescored and correlated with the experimental activities (IC 50 ).The predicted binding energies (∆G bind ) between the ligands and the receptor were calculated as previously reported [70].

MTT Assay
The cell proliferation was determined using the MTT assay.The cells were incubated in 96-well plates at 7000 cells per well and incubated for 24 h.The cells were treated for 24 h with compounds (1, 5, 10, 30, 50, and 100 µM).After treatment, the culture medium was removed, and the cells were incubated with MTT for 1 h at 37 • C. Finally, 100 µL of DMSO was added and measured by spectrophotometry at 570 nm, as previously described [71].

Determination of Extracellular Acidification Rate and Oxygen Consumption Rate in Real-Time
To analyze the extracellular acidification rate (ECAR), Caco-2 cells (10,000 cells/well) were seeded on XFe96 V3-PS multi-well plates and kept for 24 h at 37 • C in 5% CO 2 with a DMEM culture medium supplemented with FBS.After 24 h, the cells were stimulated by compounds at 100 µM for 24 h.Then, the culture medium was replaced with assay media (unbuffered DMEM without red phenol, with 4 mM glutamine and 10 mM glucose, pH = 7.4) 1 h before the assay.Glycolysis was evaluated by the sequential injection of 10 mM glucose, 1 µM oligomycin, and 100 mM 2-deoxy-D-glucose (2-DG), and ECAR was 1, suggesting that the intramolecular hydrogen bond (IAHB) C-H•••O impedes O-H dissociation.The BDEs of the O-H not involved in an IAHB presented values around 70 kcal/mol.Based on the above, the SET-PT and SPLET mechanisms only were evaluated for the estimation of PDE and PA for OH2.

Table 2 .
In vitro radical-scavenging activities (ORAC), calculated BDEs, IP, PDE, PA, and ETE (values in kcal/mol), and inhibitory effect of ortho-carbonyl hydroquinones on α-glucosidase activity.Data are shown as means ± SD of three independent experiments.* Calculated BDE of the O-H 1 for anions deprotonated at O-H 2. n.d.: not determined.n : not significant, a : p < 0.001 versus acarbose.
1, suggesting that the intramolecular hydrogen bond (IAHB) C-H•••O impedes O-H dissociation.The BDEs of the O-H not involved in an IAHB presented values around 70 kcal/mol.Based on the above, the SET-PT and SPLET mechanisms only were evaluated for the estimation of PDE and PA for OH2.

Table 2 .
In vitro radical-scavenging activities (ORAC), calculated BDEs, IP, PDE, PA, and ETE (values in kcal/mol), and inhibitory effect of ortho-carbonyl hydroquinones on α-glucosidase activity.Data are shown as means ± SD of three independent experiments.* Calculated BDE of the O-H 1 for anions deprotonated at O-H 2. n.d.: not determined.n : not significant, a : p < 0.001 versus acarbose.

Figure 2 .
Figure 2. In silico molecular docking for ortho-carbonyl hydroquinone series against the α-glucosidase enzyme.(A) Binding free energy components for the α-glucosidase-hydroquinone (3 and 10) complexes calculated by MM-GBSA analysis; all energies are in kcal/mol and pIC50 for ortho-carbonyl hydroquinones.(B) Correlations of calculated MMGBSAΔG values with experimental IC50 inhibitory values, expressed as pIC50 = −log [IC50].The r square value from linear regression and p-value for Spearman's correlation (n = 10) are shown.(C-E) 3D maps binding interactions of compound 3 and 10 against the α-glucosidase enzyme.Ligand compound 3 (A,C) and compound 10 (B,D) exposure points are indicated with gray coloring, while the binding interactions of ligands with α-glucosidase are indicated with red lines.(C) Binding free energy components for the α-glucosidase-hydroquinone (3 and 10) complexes calculated by MM-GBSA analysis; all energies are in kcal/mol.

Figure 2 .
Figure 2. In silico molecular docking for ortho-carbonyl hydroquinone series against the α-glucosidase enzyme.(A) Binding free energy components for the α-glucosidase-hydroquinone (3 and 10) complexes calculated by MM-GBSA analysis; all energies are in kcal/mol and pIC50 for ortho-carbonyl hydroquinones.(B) Correlations of calculated MMGBSA∆G values with experimental IC 50 inhibitory values, expressed as pIC50 = −log [IC50].The r square value from linear regression and p-value for Spearman's correlation (n = 10) are shown.(C-E) 3D maps binding interactions of compound 3 and 10 against the α-glucosidase enzyme.Ligand compound 3 (A,C) and compound 10 (B,D) exposure points are indicated with gray coloring, while the binding interactions of ligands with α-glucosidase are indicated with red lines.(C) Binding free energy components for the α-glucosidase-hydroquinone (3 and 10) complexes calculated by MM-GBSA analysis; all energies are in kcal/mol.

18 Figure 3 .
Figure 3.Effect of ortho-carbonyl methyl-substituted hydroquinones on proliferation, mitochondrial respiration, and glycolysis in Caco-2 cells.(A) Effect of compounds on the MTT reduction in Caco-2.Cells were treated with compounds (1, 5, 10, 30, 50, and 100 µM) for 24 h.(B) Dot-plot for basal mitochondrial OCR (basal mitoOCR) vs. glycolysis.The red square represents to control condition.Three groups of compounds are highlighted in blue, yellow, and green.(C) Dendrogram with five and two parameters of mitochondrial respiration and glycolysis, respectively, affected by compounds (100 µM) for 24 h of treatment.Data are shown as means ± SD of three independent experiments.** p < 0.01, *** p < 0.001 control, and n.s.: not significant.

Figure 3 .
Figure 3.Effect of ortho-carbonyl methyl-substituted hydroquinones on proliferation, mitochondrial respiration, and glycolysis in Caco-2 cells.(A) Effect of compounds on the MTT reduction in Caco-2.Cells were treated with compounds (1, 5, 10, 30, 50, and 100 µM) for 24 h.(B) Dot-plot for basal mitochondrial OCR (basal mitoOCR) vs. glycolysis.The red square represents to control condition.Three groups of compounds are highlighted in blue, yellow, and green.(C) Dendrogram with five and two parameters of mitochondrial respiration and glycolysis, respectively, affected by compounds (100 µM) for 24 h of treatment.Data are shown as means ± SD of three independent experiments.** p < 0.01, *** p < 0.001 control, and n.s.: not significant.

1 , 6 , 7 , and 9 .
Int. J. Mol.Sci.2024,25,  x FOR PEER REVIEW 9 of 18 ATP-linked OCR, and proton leak-driven OCR), and two of glycolysis (glycolytic capacity and glycolysis), is represented in a dendrogram (Figure3C).Interestingly, two clusters of compounds are shown with differential effects on maximal OCR and glycolysis.Cluster 1 is composed of glycolysis enhancers with a strong inhibitory effect on max.OCR: compounds Cluster 2 is composed of glycolysis inhibitors with a reduced effect on max OCR: compounds 2, 3, 4, 5, 8, and 10.

Figure 4 .
Figure 4. (A-D) Effect of selected compounds (100 µM, compounds 5 and 10) on mitochondrial respiration and (E-H) glycolysis at 24 h of treatment.(I,J) Evaluation of antioxidant effects in the in vitro model of menadione-dependent ROS production.Effects of compounds 5 and 10 on (I) mitochondrial ROS and (J) ROS in Caco-2 cells treated with menadione.The mitoROS and cytosolic ROS

Figure 4 .
Figure 4. (A-D) Effect of selected compounds (100 µM, compounds 5 and 10) on mitochondrial respiration and (E-H) glycolysis at 24 h of treatment.(I,J) Evaluation of antioxidant effects in the in vitro model of menadione-dependent ROS production.Effects of compounds 5 and 10 on (I) mitochondrial ROS and (J) ROS in Caco-2 cells treated with menadione.The mitoROS and cytosolic ROS were evaluated with mitoSOX red and DHE, respectively.Data are shown as means ± SD of three independent experiments.* p < 0.05, ** p < 0.01, and *** p < 0.001, control.n.s.: not significant.#: ** p < 0.01 for compound vs. compound + Menadione.

Table 1 .
Compounds studied in this work.

Table 1 .
Compounds studied in this work. 3

Table 3 .
Calculated binding energies of molecules 3 and 10 in the complex.