Valorization of Agave potatorum byproducts as a source of volatile α‐glucosidase inhibitors

Abstract

BACKGROUND

Agave (Agave potatorum) leaves are an overlooked byproduct of mezcal production. They possess a wealth of secondary metabolites with potential antidiabetic activity. The present study investigated this by identifying volatile constituents associated with in vitro inhibition of α‐glucosidase derived from bio‐guided hexane fractions of A. potatorum Zucc. leaves. The process involved extraction followed by silica‐gel column chromatography, along with bioautography and ultraviolet‐visible (UV‐visible) assays, using acarbose as the positive control. The two most promising bio‐guided fractions were characterized using Attenuated Total Reflection Fourier‐Transform Infrared (ATR‐FTIR) spectroscopy, phytochemical screening, and gas chromatography–mass spectrometry (GC–MS). Their in silico properties were evaluated through absorption, distribution, metabolism, and excretion (ADME) analysis and molecular docking.

RESULTS

Two fractions exhibited higher α‐glucosidase inhibitory activity with inhibitory concentration (IC₅₀) values significantly lower than that of acarbose; however, none of the fractions achieved 50% inhibition of α‐amylase. The inhibitory activity may be attributed to unsaturated metabolites with terpenoid‐like structures. Twenty‐three volatile organic compounds (VOCs) were identified in these two fractions by GC–MS. Among the compounds potentially associated with inhibitory activity of α‐glucosidase are dodecan‐1‐ol, 2,4‐di‐tert‐butylphenol, tetradecane, and benzophenone. The in silico analysis indicated that the VOCs that were identified were not expected to accumulate in the human body, and molecular docking suggested that the inhibitory effects of the identified ligands may be mediated by hydrophobic interactions and hydrogen bonding.

CONCLUSION

This research supports the revalorization of agave byproducts as a viable source of secondary metabolites capable of influencing glucose absorption by α‐glucosidase, which is a therapeutic target for addressing type 2 diabetes mellitus, the second leading cause of death worldwide. © 2026 Society of Chemical Industry.

INTRODUCTION

The distilled spirits market represents a significant segment of the alcoholic beverage industry. According to the World Health Organization (WHO) Global Status Report on Alcohol and Health and Treatment of Substance Abuse Disorders (2024), distilled spirits account for the largest proportion of recorded alcohol consumption worldwide. ¹ This sector has been registering constant growth since 2000. ² Agave‐distilled spirits reported a 2.9% increase in sales, totaling $6.7 billion in 2024. ³ These beverages were first developed by ancient pre‐Hispanic cultures and held significant cultural importance before their introduction to Spanish colonizers.

Agave uses range from beverages and food to fibers, construction materials, and pharmaceuticals; however, agave spirits (mezcal, tequila, bacanora, among others) are currently their most economically important products. Although mezcal and tequila possess denominations of origin, the global demand for agave spirits has led countries such as the USA and Australia to produce agave‐based distillates. ⁴ Depending on the species, a liter of mezcal requires between 10 and 20 kg of the heart (piña), reaching up to 35 kg in some cases. ⁵ This profitable activity generates a significant amount of waste, including vinasses, bagasse, and agave leaves. The leftover leaves generally remain unused due to the effort required for their collection and processing, and the lack of economic incentives.

Among the numerous Mexican mezcals, tobalá mezcal ranks in third position in national production according to Consejo Mexicano Regulador de la Calidad del Mezcal (COMERCAM). ² Its production requires large quantities of Agave potatorum, which is native to the states of Oaxaca and Puebla.^{6, 7, 8} Agave potatorum leaves represent up to 50% of the biomass generated during the jima (harvest),^{9, 10} and despite their considerable environmental impact, they are regarded as having limited agro‐industrial relevance. ¹¹ Like other agave species, the leaves contain diverse bioactive compounds, including alkaloids, coumarins, tannins, terpenes, phenolic compounds, flavonoids, steroids, and saponins with reported antibacterial, antifungal, antiparasitic, antioxidant, anticancer, immunomodulatory, anti‐inflammatory, and antihypertensive properties.^{10, 12} Some of these metabolites may also modulate carbohydrate‐digesting enzymes such as α‐glucosidase.^{13, 14} Previous studies have highlighted the potential to valorize A. potatorum leaf residues as a source of α‐glucosidase modulators, supported by their antioxidant activity, broad availability, and the need for sustainable management of mezcal industry byproducts. ¹⁴

Beyond its pharmacological relevance, the modulation of carbohydrate‐digesting enzymes has become an important strategy in food science. Natural plant‐derived compounds demonstrating α‐glucosidase inhibitory activity can be incorporated into functional foods with the aim of reducing carbohydrate absorption and lowering glycemic load. ¹⁵ This approach is increasingly attractive due to the consumer demand for natural ingredients and the gastrointestinal discomfort associated with synthetic inhibitors such as acarbose and voglibose.^{16, 17}

Type 2 diabetes mellitus (T2DM) is characterized by insulin resistance and a progressive decline in pancreatic function.^{18, 19} It remains one of the ten leading causes of death worldwide, ²⁰ accounting for 3 million deaths in 2024. ¹⁹ Dietary strategies targeting carbohydrate digestion, particularly through α‐amylase and α‐glucosidase modulation, are widely recognized as effective approaches to manage postprandial glycemia.^{21, 22, 23}

In the context of agro‐industrial valorization, exploring the functional potential of metabolites present in agave leaf residues aligns with current efforts to promote sustainability. The main objective of this study was therefore to identify volatile secondary metabolites from A. potatorum leaf extracts through bio‐guided fractionation and to evaluate their potential as natural inhibitors of α‐glucosidase activity.

EXPERIMENTAL

Reagents and solvents

All reagents and solvents were of analytical grade and used without further purification. 2‐Naphthyl‐α‐d‐glucopyranoside was obtained from Goldbio (St Louis, MO, USA), and α‐glucosidase (from Saccharomyces cerevisiae), α‐amylase (from porcine pancreas), p‐nitrophenyl‐α‐d‐glucopyranoside (PNPG), fast blue B salt (FBS), potato starch, imidazole, and silica gel (70–230 mesh) for column chromatography were acquired from Sigma‐Aldrich (Toluca, Mexico). Potassium iodide (KI), iodine (I₂), lugol solution, sodium acetate (NaAcO), acetic acid, methanol (MeOH), anhydrous silica gel (4–8 mesh) with indicator, distilled water (H₂O), phosphate buffer pH 7.0 (PBS), hexane (Hex), ethyl acetate (EtOAc), ferric chloride III (FeCl₃), cobalt chloride (CoCl₂), potassium hydroxide (KOH), phosphoric acid, diphenylamine, vanillin, and bismuth nitrate were all obtained from Meyer (Mexico City, Mexico).

Ammonium acetate (NH₄AcO) and dimethyl sulfoxide (DMSO) were purchased from J.T. Baker (Paris, KY, USA); polyethylene glycol 4000 (PEG) from Merck (Toluca, Mexico); and hydrochloric acid (HCl), sulfuric acid (H₂SO₄), acetic anhydride, and glacial acetic acid from Fermont (Monterrey, Mexico).

Centro de Investigación y de Estudios Avanzados (CINVESTAV) del Instituto Politécnico Nacional, Unidad Zacatenco, kindly provided diosgenin and ethinylestradiol standards.

Acarbose (positive control) was extracted from commercial tablets (Pisa Pharmaceuticals, Tlajomulco de Zuñiga, Jalisco, Mexico). Twenty tablets (50 mg each) were dissolved in 20 mL of MeOH, stirred, filtered, and concentrated under reduced pressure using a rotary evaporator, and characterized by Fourier transform infrared (FTIR) spectroscopy to ensure their purity (see Supporting Information, Fig. S1).

Instrumentation and equipment

Spectrophotometry

The absorbance was measured at multiple wavelengths using a microplate reader (Multiskan FC 357; Thermo Fisher Scientific, Shanghai, China, Serial Number 357–912 913).

Mid‐range infrared spectroscopy

Fourier Transform Infrared spectra were acquired using a Nicolet 6700 spectrometer (Thermo Scientific, Madison, WI, USA) with 32 scans at 4 cm⁻¹ resolution. An Attenuated Total Reflection (ATR) accessory with a ZnSe/diamond crystal was used to record spectra in the range of 4000–650 cm⁻¹. Absorption maxima are reported in wavenumbers (cm⁻¹).

Gas chromatography–mass spectrometry

Chromatography analyses were conducted using SCION 456 gas chromatography (GC) equipment (Bruker, Billerica, MA, USA) coupled to a triple quadrupole (TQ) mass detector and a CombiPAL autosampler. Separation was achieved using an Rxi‐5Sil MS fused silica column (30 m × 0.32 mm × 1.0 μm film thickness) (Restek, Bellefonte, PA, USA).

Helium was used as carrier gas (flow rate 1 mL min⁻¹). The injector was maintained at 220 °C with a split ratio of 10:1. The oven temperature program was as follows: the initial temperature was set at 55 °C and held for 1 min; then the temperature was increased to 155 °C at a rate of 20 °C min⁻¹, and held for 2 min. Afterwards, the temperature was then increased from 155 °C to 255 °C at 10 °C min⁻¹, and held for 5 min. Finally, the temperature was increased to 280 °C at 10 °C min⁻¹ and held for 5 min.

Electron ionization (EI) was performed at 70 eV, and mass spectra were processed using MS Data Review software with the National Institute of Standards and Technology (NIST) mass spectral library.

Biological material

Agave potatorum Zucc. leaves were selected as described in previous reports. ¹⁴ The leaves, collected from 5‐year‐old plants grown in San Pablo Etla (Valles Centrales, Oaxaca, Mexico), were identified according to morphological descriptors,^{6, 8} which were corroborated by comparison with the voucher OAX 21200 deposited in the herbarium of the Centro Interdisciplinario de Investigación para el Desarrollo Integral Regional, Unidad Oaxaca (CIIDIR Oaxaca) by Dr Abisaí Josué García Mendoza, ⁶ curator of the National Collection of Agavaceae and Nolinaceae at the Jardín Botánico del Instituto de Biología of the Universidad Nacional Autónoma de México (UNAM). All plants were harvested before floral scape development.

Leaves from three independent plants were washed with purified water, suitable for human consumption, cut into 0.5–0.7 cm longitudinal strips, and dried at room temperature (20 ± 2 °C) in darkness under constant airflow until constant weight (~3 days). Dried leaves were stored in sealed polyethylene bags at room temperature in a dry environment.

Extraction by ascending polarity and fractionation of secondary metabolites from leaves of A. potatorum Zucc.

Dried leaves were ground to ~1 mm particle size using a Pulverisette 19 universal cutting mill (Fritsch, Idar‐Oberstein, Germany, SN 19.1010/00821). Ultrasound‐assisted extraction was performed at 40 kHz in a $5.7$ L ultrasonic bath (Branson Ultrasonics, Danbury, CT, USA; SN BGJ101625780B) for 240 min at 20 ± 1 °C using solvents of increasing polarity – Hex, EtOAc, MeOH, and H₂O – at a solid to solvent ratio of 500 g L⁻¹, with four solvent renewals. Extracts were decanted, filtered, and concentrated at 40 °C under vacuum. The concentrate was then dried under an air stream for 24 h. Samples were stored in anhydrous silica at 4 °C until analysis.

Extractions were performed in triplicate. Yields were calculated using Eqn (1):

$$\text{yield}\left(\mathrm{g}{\mathrm{kg}}^{-1}\right)=\frac{w_i}{w_{\mathit{\text{Agave}}}}\times {10}^3$$ (1)

where $w_i$ is the mass obtained for solvent $i$ and $w_{\mathit{\text{Agave}}}$ corresponds to the initial dried leaf mass.

Bioautographic assays

Inhibition of α‐amylase

Qualitative α‐amylase inhibition was assessed using the method reported by Juárez‐Niño and Soto‐Castro ¹⁴ without modification.

Inhibition of α‐glucosidase

Qualitative α‐glucosidase inhibition was evaluated using the methodology described by Galindo‐Vargas et al. ¹³

Bio‐guided fractionation of the hexane extract

A total of 9.69 g of the Hex extract was subjected to silica gel column chromatography (5 × 20 cm, 70–230 mesh) using a gradient of Hex to Hex–EtOAc (7:3). Six primary fractions (identified as F1, F2, …, F6) were collected, monitored by Thin‐Layer Chromatography (TLC), concentrated at 40 °C, and dried. The F6 fraction showed inhibition of α‐glucosidase (IAG) activity and was further fractionated using a second column (1.5 × 40 cm) with Hex–EtOAc mixtures of increasing polarity; the ten fractions were obtained and labeled as F6‐1 to F6‐10. Bioactivity was monitored until two active fractions (F6‐5 and F6‐6) were obtained, and they were subjected to phytochemical screening, FTIR, gas chromatography–mass spectrometry (GC–MS), and quantitative inhibition assays. Details of the mobile phase, ratio, yield, and retention factors (R _f ) are available in the Supporting Information (Tables S1 and S2).

Phytochemical screening

Qualitative phytochemical analysis was conducted using TLC eluted of F6‐5 and F6‐6 according to the established methodology ²⁴ . Alkaloids, phenolics, flavonoids, anthraquinones, coumarins, triterpenes, steroids, saponins, terpenoids, and phenylpropanoids were classified as high (+++), medium (++), or low (+). Detailed protocols are provided in the Supporting Information, Table S3.

Quantitative inhibitory assays

Inhibition of α‐amylase

Assays were performed (in quadruplicate) with two independent replicates in accordance with the method described by Juárez‐Niño and Soto‐Castro. ¹⁴ Percentage inhibition of α‐amylase (%IAA) was calculated using equation 2, where Abs_sample and Abs_control refer to the absorbance at 590 nm of the sample and control respectively:

$$\%\mathit{IAA}=\left(\frac{{\mathit{Abs}}_{\mathit{\text{sample}}}}{{\mathit{Abs}}_{\mathit{\text{control}}}}\right)\times 100$$ (2)

Inhibition of α‐glucosidase

The percentage of inhibition of α‐glucosidase (%IAG) activity was determined using the methodology established by Juárez‐Niño and Soto‐Castro ¹⁴ by reading the absorbances at 405 nm and applying Eqn (3), where Abs_sample refers to the absorbance of the sample with enzymatic solution, Abs_{blank sample} to absorbance of the sample without enzyme solution (with buffer instead), Abs_control to the absorbance of the enzymatic solution without sample (only its solvent), and Abs_blank to the absorbance of solvents in both sample and the enzymatic solution:

$$\%\mathit{IAG}=\left[1-\frac{\left({\mathit{Abs}}_{\mathit{\text{sample}}}-{\mathit{Abs}}_{\mathit{\text{blank sample}}}\right)}{{\mathit{Abs}}_{\mathit{\text{control}}}-{\mathit{Abs}}_{\mathit{\text{blank}}}}\right]\times 100$$ (3)

In silico absorption, distribution, metabolism, and excretion properties

The structures of identified metabolites were input as Simplified Molecular Input Line Entry System (SMILES) strings into SwissADME (http://www.swissadme.ch/). Physicochemical properties, pharmacokinetic predictions, and drug‐likeness parameters were retrieved (Supporting Information, Table S4). Upon completion of the processing, a variety of physicochemical properties, pharmacokinetics, pharmacological similarity, and other pharmaceutical chemical data related to the molecules were obtained.

Molecular docking

A molecular docking study was conducted to estimate the binding interactions and affinities of the identified volatile organic compounds (VOCs) with the active site of α‐glucosidase. The crystal structure of α‐glucosidase from Saccharomyces cerevisiae (PDB ID: 7DCH) was used. To maintain consistency with the in vitro experimental conditions, ²⁵ the docking process was performed using the SwissDock web server (http://www.swissdock.ch/). The protein targets were subjected to the AutoDock Tools program for removing crystal water molecules and their complexes’ inhibitors. A docking approach was employed with a search space of 30 × 30 × 30 Å centered on the catalytic site. The AutoDock Vina program was employed for docking calculations at an exhaustiveness level of 6 to ensure a thorough sampling of the ligand's conformational space.^{26, 27} For each ligand, the optimized complex in .PDB format represented the most stable conformation (lowest binding energy), which was generated using Visual Molecular Dynamics (VMD). ²⁸ These optimized complexes were subsequently analyzed using the protein‐ligand interaction profiler (PLIP) ²⁹ to characterize specific molecular interactions, including hydrogen bonds and hydrophobic contacts within a 4.0 Å cutoff. Acarbose was docked under identical parameters as a positive control.

RESULTS AND DISCUSSION

Extraction of metabolites from A. potatorum leaves

Sequential ultrasound‐assisted extractions were performed with ascending polarity in Hex, EtOAc, MeOH, and H₂O with yields of 9 ± 0.35, 7.7 ± 0.1, 66.2 ± 8.1, and 377.5 ± 13.1 g kg⁻¹, respectively. The average yields obtained from ultrasound‐assisted extraction with non‐polar solvents, such as hexane and EtOAc, were generally less than 1%. In contrast, higher yields were obtained with polar solvents, indicating that most of the secondary metabolites contained in the leaves of A. potatorum were highly polar, as mentioned by Galindo‐Vargas et al. ¹³ However, the extraction yields obtained for all extracts were different from those reported by Galindo‐Vargas et al. ¹³ for the 8‐year‐old plant, hexane (12.3 ± 0.3 g kg⁻¹), EtOAc (8.9 ± 0.3 g kg⁻¹), MeOH (38.4 ± 0.4 g kg⁻¹), and H₂O (220.2 ± 3.5 g kg⁻¹). This could be due to differences in the locations and methods of cultivation of the species because previous studies indicated that the synthesis of secondary metabolites and accumulation of biomass were influenced strongly by environmental stressors such as water availability, soil composition, and solar radiation, which vary significantly between different geographic regions in Oaxaca. ³⁰

Inhibitory activity of α‐amylase and α‐glucosidase enzymes by bioautography of extracts from A. potatorum Zucc.

Figure 1(a), (b) show the qualitative assays of α‐amylase inhibition (IAA) and α‐glucosidase enzyme inhibition from extracts obtained by ascending polarity without elution. The lipophilic extracts, obtained in Hex and EtOAc, exhibited the strongest capacity to inhibit both enzymes. The extracts obtained in MeOH and H₂O did not inhibit the α‐amylase enzyme. From Fig. 1(a), it is evident that none of the A. potatorum crude extracts inhibit the α‐amylase enzyme to the same extent as acarbose, even at the highest loading amount (400 μg per spot); this contrasts with the results observed for α‐glucosidase inhibition.

Figure 1.

Bioautographic evaluation under natural light of (a) α‐amylase and (b) α‐glucosidase enzyme inhibition of extracts of Agave potatorum leaves.

Figure 1(b) shows that the leaves of A. potatorum have secondary metabolites capable of inhibiting IAG in a concentration‐dependent dose–response relationship, which agrees with that reported by Galindo‐Vargas et al. ¹³

Bio‐guided fractionation of the A. potatorum hexane extract leaves in an open chromatographic column

Coarse fractionation of 9.69 g of dried hexane extract from A. potatorum leaves was performed using an open chromatographic column (Fig. 2(a). Six fractions were obtained (identified as F1 to F6) and monitored by TLC (Fig. 2(b)) and bioautographic assay at a concentration of 200 μg per spot to determine the IAG for each fraction.

Figure 2.

Thin layer chromatography (TLC) of different fractions obtained by fractionation in an open chromatographic column by ascending polarity of the crude hexane extract. (a) Chromatographic column fractionation. (b) Thin layer chromatography monitoring of the obtained fractions eluted with Hex:EtOAc 85:15, (c) Bioautographic assay of α‐glucosidase enzyme inhibition. White squares in the TLC plate indicate the α‐glucosidase inhibition zones of the compounds in the fractions.

As Fig. 2(c) shows, fractions F1 and F2 did not present α‐glucosidase inhibition zones, whereas fraction F3 showed two inhibition zones with R _f of 0.23 and 0.53. Fractions F4, F5, and F6 displayed inhibition zones ranging from R _f of 0.11 to 0.53, indicating that fractions of higher polarity have a more complex chromatographic profile, taking into consideration the detectable spots, possibly because of the number of secondary metabolites present with the capacity to inhibit the α‐glucosidase enzyme. These fractions were the most abundant, with yields ranging from 6% to 30% (see Supporting Information, Table S1). Fractionation by ascending polarity of F6 was also performed because this fraction appeared to contain a lower number of compounds than F5, and these were easier to monitor due to their distinguishable ultraviolet‐visible (UV‐visible) response. To do this, 2.6967 g of F6 was fractionated in an open chromatographic column, and ten fractions (F6‐1 to F6‐10, see Supporting Information, Table S2) were obtained.

The bioautography assay of the ten fractions was performed with 200 μg aliquots of each fraction, using acarbose as a positive control, on eluted TLC plates (see Fig. 3). Spots of each of the non‐eluted fractions were also placed on top of the TLC plate.

Figure 3.

Bioautographic assay of α‐glucosidase enzyme inhibition from ten fractions obtained from fraction F6. Fractions F6‐1 to F6‐5 were eluted in Hex:EtOAc 7:3, while fractions F6‐6 to F6‐10 were eluted in Hex:EtOAc 6:4. All fractions and acarbose were deposited at 200 μg per spot.

It was observed (see Fig. 3 and Supporting Information, Table S2) that all fractions exhibited at least one zone of inhibition with a higher apparent intensity than that of acarbose. Figure 3 also shows that fractions F6‐5 and F6‐6 had the highest inhibitory intensity and were among the most abundant, with yields of 0.6418 and 0.3871 g, respectively. Regarding the non‐eluted spots, fractions F6‐3, F6‐5, F6‐6, F6‐7, F6‐9, and F6‐10 seem to exhibit stronger apparent inhibition than acarbose. Notably, F6‐5, F6‐6, and F6‐7 showed the highest inhibition, with F6‐5 being the most notable.

For fractions F6‐5 and F6‐6, α‐glucosidase enzyme inhibition quantification was performed. As Fig. 4 shows, both fractions had a significantly lower IC₅₀ (90 ± 2 and 97 ± 13 μg mL⁻¹, respectively) than acarbose (2260 ± 193 μg mL⁻¹) and showed linear dose‐dependent behavior in the concentration range of 125–3 and 250–3 μg mL⁻¹; however, with increasing concentration, the behavior became asymptotic. The same occurred with the positive control, which showed linear behavior in the range of 4000–125 μg mL⁻¹.

Figure 4.

Dose–response inhibition curve of α‐glucosidase enzyme with fractions obtained from the second fractionation of the hexane extract of Agave potatorum Zucc. leaves. Different letters expressed as superscripts of IC₅₀ indicate significant differences (P < 0.05, Tukey test).

The qualitative inhibitory activity against α‐amylase of the hexane crude extract and the fractions F6‐5 and F6‐6 were assayed by TLC bioautography using 400 μg per spot (see Fig. 5(a)). Here, three inhibition zones corresponding with different intensities were observed for both the hexane extract and the fractions; however, only the inhibition zones with R _f from 0.36 to 0.74 became as intense as acarbose applied at 50 μg per spot.

Figure 5.

Assay of α‐amylase enzyme inhibition: (a) bioautographic and (b) dose–response curve, from hexane crude extract and fractions F6‐5 and F6‐6.

Figure 5(b) indicates the quantitative inhibitory activity against α‐amylase at a concentration up to 2000 μg mL⁻¹ and shows that none of the fractions reached 50% inhibition, in comparison to acarbose (IC₅₀ 330 ± 10 μg mL⁻¹), which correlates with the low intensity of the inhibition halos observed by bioautography. Fractions F6‐5 and F6‐6 exhibited a maximum inhibition of 33% at 2000 μg mL⁻¹. It was therefore expected that secondary metabolites with high IAG and moderate IAA would decrease the side effects caused by commercial inhibitors, which inhibit both enzymes,^{31, 32} as carbohydrates would be partially hydrolyzed, but would fail to achieve absorption as they would not be hydrolyzed to monosaccharides by α‐glucosidase.

Phytochemical analysis

Phytochemical screening was performed to determine the families of compounds present in the hexane extract and its fractions. Table 1 reports the presence of these metabolite families as strong (+++), moderate (++), weak (+), and null (−). The developed TLC plate is shown in the Supporting Information, Table S5. As expected for a lipophilic extract, such as the one obtained in hexane and its respective fractions, the phytochemical screening revealed the presence of terpenic structures; this is consistent with previous reports for non‐polar extracts of A. potatorum Zucc. leaves obtained with hexane, dichloromethane, and EtOAc ¹³ and ethanolic extracts. ⁹ The Liebermann–Burchard test yielded positive results, indicating a strong presence of triterpenes, steroids, and/or saponins in both the hexane extract and fraction F6‐6, whereas fraction F6‐5 showed only a moderate presence. The vanillin‐phosphoric acid test also indicated a strong presence of terpenoids and lignans in the hexane extract, but only moderate levels in the fractions. Regarding saponins, both the hexane extract and its fractions tested strongly positive, in agreement with the reported presence of these compounds in the methanolic extract of A. potatorum leaves. ³³ Finally, the vanillin–sulfuric acid test confirmed a strong presence of terpenoids and/or phenylpropanoids in the hexane extract and in fraction F6‐6, whereas fraction F6‐5 showed a moderate presence.

Table 1.

Phytochemical screening by TLC of hexane extracts from A. potatorum Zucc. and their fractions

Test	Secondary metabolite family	Hex	F6‐5	F6‐6
Liebermann–Burchard	Triterpenes, steroids, saponins	+++	++	+++
Vanillin‐phosphoric acid	Terpenoids and lignans	+++	++	++
	Saponins	+++	+++	+++
Vanillin‐sulfuric acid	Terpenoids, phenylpropanoids	+++	++	+++
Dragendorff reagent	Alkaloids	−	−	−
FeCl3	Tannins	−	−	−
Fast blue salt	Phenolic compounds	++	+	++
KOH	Anthraquinones	+	++	+
	Anthrones	+++	++	+
	Coumarins	−	−	++
NP–PEG	Coumarins, phenylcarboxylic acids	+	+	++
	Anthrones, anthranols	+++	−	−
	Flavanols, flavones	−	−	−
CoCl2	Coumarins	+	−	++

Alkaloid and tannin detection with Dragendorff's reagent and FeCl₃ confirmed their absence in the hexane extract and fractions, which is consistent with their lipophilic nature and with previous reports for A. potatorum ethanolic extracts. ⁹ The fast blue salt test indicated a moderate presence of phenolics; nevertheless, more specific assays showed that the KOH test revealed variable anthraquinones, strong anthrones, and no coumarins in the hexane extract, whereas fraction F6‐5 presented moderate anthraquinones/anthrones and no coumarins, and fraction F6‐6 showed weak anthraquinones/anthrones but moderate coumarins, consistent with Galindo‐Vargas et al. ¹³ The natural products–polyethylene glycol (NP–PEG) test confirmed the presence of anthrones/anthranols in the hexane extract but flavanols or flavones were not detected in all samples. It also indicated the presence of weak coumarins/phenylcarboxylic acids in the hexane extract, with none detected in F6‐5 and moderate levels in F6‐6. Finally, the CoCl₂ test confirmed the presence of weak coumarins in the hexane extract and F6‐5, and mild levels in F6‐6.

From Figs 3 and 5(a), the inhibition zones of α‐amylase and α‐glucosidase enzymes from the bioautographic assays coincide with the zone of highest intensity in the TLC and correspond to compounds with R _f values from 0.30 to 0.74. According to the phytochemical profile (Supporting Information, Table S5), the main inhibitors of the enzymes are terpenic and/or steroidal in nature. Due to the overlap of spots in that zone, it is possible that, within the structure of the secondary metabolites there are phenolic groups, as indicated by a weakly positive response to fast blue salt. Likewise, from this comparison and the report on the α‐glucosidase enzyme inhibitory activity of terpenic structures, ³⁴ it is possible to attribute the inhibitory activity to terpenoids, triterpenes, steroids, or saponins present in fractions F6‐5 and F6‐6. The results published by Galindo‐Vargas et al. ¹³ were confirmed, suggesting that the coumarin‐ and terpene‐type compounds present in the hexane extract of A. potatorum Zucc. leaves may act as α‐glucosidase inhibitors.

Chemical characterization of fractions with α‐glucosidase enzyme inhibition

Fourier transform infrared spectroscopy

Fourier Transform Infrared (FTIR) spectra were acquired to identify the presence of the predominant functional groups of the compounds in fractions F6‐5 and F6‐6 (Fig. 6 and Table 2). In the case of fraction F6‐5 (Fig. 6(a)), bands associated with the stretching of olefinic C=C‐H bonds (3010 cm⁻¹) and CH, CH₂, and CH₃ bonds of aliphatic compounds (2923–2854 cm⁻¹) were observed. The band at 1709 cm⁻¹ also corresponded to the stretching of the carbonyl groups, which was corroborated by the asymmetric stretching band located at 1284 cm⁻¹; however, this last signal could also be associated with the C—C stretching of aliphatic chains. The band associated with the C=C vibration cannot be detected due to a possible overlap with the signal associated with the carbonyl group.

Figure 6.

Fourier transform infrared spectroscopy–attenuated total reflectance (FTIR–ATR) spectra of Agave potatorum Zucc. fractions: (a) F6‐5 and (b) F6‐6.

Table 2.

Assignment of the most intense experimental frequencies of the fractions F6‐5 and F6‐6

IR ν (cm−1)		Assignment
F6‐5	F6‐6
3010	3010	C=C‐H st (olefinic)
2923–2854	2923–2854	CHn st (aliphatic)
1709	1750, 1709, 1644	C=O and C=C st
1463–1413	1462	C‐H bas
1284	1375	C‐O st; C—C st
1121–1072	1178	C‐H bout; C—H bin; C—O st

b_as, asymmetrical bending; b_in, in‐plane bending; b_out, out‐of‐plane bending; b_s, symmetrical bending; st, stretching.

In contrast, despite a strong similarity to the spectrum of fraction F6‐5, Fig. 6(b) shows the spectrum of fraction F6‐6 with some distinctive signals. For instance, the stretching of a C=C double bond is visible at 1644 cm⁻¹, possibly pointing to compounds with unsaturation or conjugated systems. Furthermore, the region associated with the stretching of hydroxyl groups is more pronounced. This may indicate the presence of phenolic hydroxyl groups capable of forming hydrogen bonds, which is consistent with the phenolic‐type compounds observed in the fraction by the fast blue salt phytochemical screen.

F6‐6 also presents a visible band at 1178 cm⁻¹, which may be associated with oxygenated functional groups containing the C—O stretching band; this signal, in conjunction with the shown phytochemistry tests (KOH and CoCl₂ tests), could be associated with the presence of lactones and esters (see also the GC results and the identification of volatile compounds in Table 3).

Table 3.

Volatile organic compounds (VOCs) identified by gas chromatography (GC) from hexane fractions of the leaves of A. potatorum Zucc.

No.	Compound	Relative abundance (%)		Bioactivity reported
		F6‐5	F6‐6
1	(2E,4E)‐hepta‐2,4‐dienal	2.142,	16.368	Antifungal 35 Antioxidant, antimicrobial, and fumigant 36
2	Octanoic acid	—	2.142	Antioxidant, antimicrobial, and insect repellent 36 Antifungal, antibacterial 37 Stimulant of insulin production in humans and rats 38 Anticarcinogenic, antibacterial, antimicrobial 39
3	2‐(Pentadec‐12‐yn‐1‐yloxy)tetrahydro‐2H‐pyran	2.640	—	Possible anticancer activity 40
4	Nonanoic acid	—	6.249	Antidiabetic 41 Antimicrobial 42 Pesticide 43
5	8‐Methylene‐3‐oxatricyclo[5.2.0.03, 5]nonane	—	8.797	—
6	(2E,4Z,7Z)‐trideca‐2,4,7‐trienal	0.751	—	Anticancer, antidote, cancer preventive, cytochrome‐P450‐2E1 inhibitor, decreases C‐telopeptide excretion, decreases deoxypyridinoline excretion, decreases endothelial leukocyte adhesion, decreases endothelial platelet adhesion, decreases epinephrine production, decreases oxalate excretion 44
7	(2E,4E)‐deca‐2,4‐dienal	2.093, 2.731	—	—
8	(2E)‐dodec‐2‐en‐4‐yne	—	10.870	—
9	6,10‐Dimethyl‐undeca‐5,9‐dien‐2‐one	0.806	—	Antibacterial 45
10	Dodecan‐1‐ol	—	7.209	Antibacterial39, 46 Blocks yeast‐to‐hyphal transition in C. albicans 47 Insecticide 48 Antidiabetic 49
11	Tridecan‐2‐one	2.138	—	Antifungal50, 51 Antibacterial 52 Larvicide 53
12	2,4‐Di‐tert‐butylphenol	75.918	—	Antioxidant, anti‐inflammatory, antimicrobial, cytotoxic, insecticidal, nematocidal, allelopathic, auto‐toxic 54 Antibacterial55, 56 Antibacterial and antiproliferative 57 Antibiofilm 58 Antioxidant 59 Herbicide 60 Antidiabetic in silico 61 Antioxidant y antidiabetic 62 Antidiabetic in vitro 63
13	N, N‐diethyl‐4‐methylbenzamide (diethyltoluamide)	0.723	5.462	Insect repellent 64 Antimicrobial 65
14	Hexadecane‐1‐ol	0.096	7.191	Antimicrobial66, 67, 68 Antibacterial 69 Antioxidant 70 Antibiofilm 71
15	Tetradecane	0.402	—	Antidiabetic in silico 72 Antimicrobial 73 Antioxidant and antifungal 74 Antibacterial 75
16	(E)‐10‐methyltridec‐11‐en‐1‐yl propionate	0.291	—	Antimicrobial 76 Anti‐atherosclerotic in silico 77 Anti‐inflammatory, anticarcinogenic, diuretic 78
17	Benzophenone	0.628	—	Anticarcinogenic, antimicrobial, anti‐inflammatory, antiviral79, 80 Antifungal, anti‐HIV, antiviral, antimicrobial, antioxidant 81 Antifungal 82 Anti‐tumor, anti‐inflammatory, anti‐diabetic 83 Hypoglycemic, anti‐inflammatory 84 Antinematicidal, antimicrobial, antimycobacterial, antialgae, cytotoxic, antioxidant, immunosuppressant, anticoccidial, antimalarial, anti‐inflammatory, anti‐osteoclastogenic, α‐glucosidase inhibitor, proteasome, tyrosine phosphatase, antihyperlipidemic, O‐acyltransferase, protein kinase, phytotoxic, insecticide. 85
18	Methyl 2‐(3‐oxo‐2‐pentylcyclopentyl)acetate	—	2.050	—
19	Cyclotridecane	—	5.091	—
20	2‐(tert‐butyl)‐5‐(4,4‐dimethoxy‐pentyl)‐5‐methyl‐1,3‐dioxolan‐4‐one	0.270	—	—
21	Nonadec‐1‐ene	1.277	8.441	Antituberculosis, anticarcinogenic, antioxidant, antimicrobial 86 Antibacterial, antifungal 87
22	2‐Ethylhexyl 2‐hydroxybenzoate (2‐ethylhexyl salicylate)	0.977	3.090	Ultraviolet filter88, 89
23	Isopropyl tetradecanoate (isopropyl myristate)	1.776	2.988	Antibacterial 90

Only two compounds of the 23 identified have been reported in Agave: 12 in A. salmiana ⁹¹ and 15 in A. attenuata. ⁹²

Gas chromatography–mass spectrometry

Fractions F6‐5 and F6‐6 were analyzed using gas chromatography GC–MS. Table 3 summarizes the results, and the chromatograms and additional data are presented in the Supporting Information (Fig. S2 and Table S6). For fraction F6‐5, it was possible to identify 16 compounds, and for F6‐6, it was possible to identify 13 compounds. Of the 29 compounds identified, six were in both fractions; therefore, 23 different compounds were identified. Only two of them have been reported in plants of the genus Agave: 2,4‐di‐tert‐butylphenol (12) in A. salmiana ⁹¹ and tetradecane (15) in A. attenuata. ⁹² However, 12 and 15 have also been reported as part of the volatile organic compounds (VOCs) in tequila and its vinasses.^{93, 94} In this sense, octanoic (2) and nonanoic (4) acids were reported as forming part of the VOCs of mezcal from A. angustifolia and A. potatorum, ⁹⁵ and 4 has also been reported in pulque. ⁹⁶ Finally, dodecan‐1‐ol (10) has been reported as a VOC of tequila ⁹⁷ and hexadecane‐1‐ol (14) as a VOC of mezcal and other beverages.^{93, 95, 97, 98}

Regarding the reported bioactivity of the metabolites identified in the fractions F6‐5 and F6‐6 (Table 3), only 4, ⁹⁹ 10, ⁴⁹ 12,^{61, 63} 15, ⁷² and benzophenone (17) have been reported as antidiabetic compounds.^{84, 85} Sipahutar et al. ⁴¹ reported that 4 stimulates glucagon‐like peptide (GLP‐1) and peptide YY (PYY) through OR51E1 olfactory receptor signaling in L cells. The antidiabetic activity of compounds 10, 12, 15, and 17 was attributed to their inhibition of α‐amylase and α‐glucosidase enzymes. Specifically for 10, an in vitro study that exhibited significant inhibition of these enzymes was performed; ⁴⁹ and for 12 ^{61, 63}, and 15, ⁷² in vitro and in silico molecular docking studies were performed.

In the case of 17, it was reported that some of its glycosidic derivatives and polyhydroxybenzophenones, such as polyhydroxybenzophenone, promote glucose oxidation and favor its utilization as fuel through carbohydrate metabolism ⁸⁴ and exhibit inhibition of α‐glucosidase.^{84, 100, 101} Thus, considering the literature to date, 10 is one of the compounds potentially contributing to the inhibitory activity of α‐amylase and α‐glucosidase enzymes in F6‐6, and in F6‐5, some of the possible contributors are 12, 15, and 17.

In silico analysis of the VOCs absorption, distribution, metabolism, and excretion properties

The absorption, distribution, metabolism, and excretion (ADME) properties of the VOCs identified by GC were determined from in silico analysis. Based on the physicochemical properties related to crossing biological barriers, absorption, and distribution, ¹⁰² VOCs were characterized by small size and lipophilic nature (Supporting Information, Tables S4 and S7). In addition, all identified VOCs could undergo gastrointestinal absorption by passive diffusion (human intestinal absorption, see Fig. S3) and could enter cells through cellular transporters.^{103, 104} This suggests that they function as hypoglycemic agents not only by inhibiting gastrointestinal enzymes but also through other mechanisms.

Regarding pharmacokinetic parameters related to metabolism and excretion (Supporting Information, Table S8), the inhibition of certain cytochrome P450 (CYP) enzymes involved in drug biotransformation and elimination was evaluated. Understanding the inhibition of these CYPs can help predict adverse effects resulting from the accumulation of the compound or its metabolites in the body; these effects are more pronounced when molecules rely on a single enzyme for metabolism. Although metabolites 3, 10, 11, 12, 15, 16, 17, and 21–23 may inhibit at least one of these CYPs, accumulation issues are unlikely because other CYPs could metabolize them. As a result, they may be promising oral drug candidates with high intestinal absorption and the ability to modulate carbohydrate digestion through pathways beyond α‐glucosidase inhibition, without risking systemic accumulation.

In summary, four of the VOCs present in A. potatorum leaves can potentially contribute to enzyme inhibitory activity. However, other major compounds must be responsible for this inhibition. This is because the identified VOCs do not coincide in both fractions, which would be expected given the overlapping enzyme inhibition zones in the bioautographic assays (Figs 3 and 5(a)). Consequently, high‐performance liquid chromatography (HPLC) and nuclear magnetic resonance (NMR) analyses are necessary to identify other contributing metabolites and to develop strategies for their extraction, conservation, and administration to humans. Nevertheless, these results indicate that A. potatorum leaf residues have the potential to serve as a valuable source of functional compounds.

Molecular docking analysis

The 23 VOCs analyzed by docking interacted within the catalytic site of the 7DCH α‐glucosidase by hydrophobic interactions, displaying typical distances of less than 4 Å^{95, 96, 97} (Table 4). Specifically, the highest binding affinities were determined for compounds 20 (2‐(tert‐butyl)‐5‐(4,4‐dimethoxy‐pentyl)‐5‐methyl‐1,3‐dioxolan‐4‐one), 17 (benzophenone), and 12 (2,4‐di‐tert‐butylphenol), which exhibited the lowest binding energies of −5.645, −5.642, and −5.493 kcal mol⁻¹, respectively. These three compounds possess an aromatic structure that contributes to hydrophobic interactions and H‐bonding, enabling them to interact with the amino acids by pi‐stacking. Compound 20 established a high‐stability H‐bond with the residue ARG 539 at 2.83 Å; this interaction is critical, as ARG 539 plays a key role in the stabilization of the transition state within the catalytic pocket of Saccharomyces cerevisiae α‐glucosidase. The distances and binding energies of 2,4‐di‐tert‐butylphenol correspond to the data reported by Sanseya et al. ²⁵ Compounds 3, 13, and 22 rank next in terms of binding energy; they interact in the catalytic pocket by H‐bond of slightly higher energy than compounds 20, 17, and 12, since the distance of H‐bond are greater. Nonetheless, a consistent pattern of hydrophobic interactions was observed across the most active ligands; these molecules interacted with a cluster of nonpolar residues, specifically LEU 421, PRO 466, and ILE 525. This suggests that these lipophilic metabolites may contribute to the overall enzymatic inhibition by sterically hindering substrate access to the catalytic triad.

Table 4.

Molecular docking analysis of VOCs from the hexane extract of Agave potatorum leaves in the binding site of Saccharomyces cerevisiae α‐glucosidase (7DCH)

No.	Name	Binding energy (kcal mol−1)	Hydrophobic interaction – (amino acid‐residue number (distance in Å)	Hydrogen bond – amino acid residue number and (distance H–A in Å, donor angle)
1	(2E,4E)‐hepta‐2,4‐dienal	−3.745	TYR‐17A (3.51) VAL‐19A (3.56) TYR‐134A (3.66) ASP‐426A (3.7) ALA‐427A (3.57)
2	Octanoic acid	−4.184	LEU‐421A (3.92) LEU‐421A (3.54) LEU‐421A (3.92) ASN‐429A (3.92) PRO‐466A (3.78) PRO‐466A (3.78)	ARG‐539A (3.27, 130.96)
3	2‐(pentadec‐12‐yn‐1‐yloxy)tetrahydro‐2H‐pyran	−5.020	THR‐141A (3.56) LEU‐421A (3.7) LEU‐421A (3.81) PRO‐466A (3.99) PRO‐466A (3.66) ILE‐525A (3.85)	ASP‐522A (3.05, 150.12)
4	Nonanoic acid	−4.074	LEU‐421A (3.73) LEU‐421A (3.79) LEU‐421A (3.67) ASP‐426A (3.76) PRO‐466A (3.62) ILE‐525A (3.84)	ASP‐522A (3, 158.66) ASP‐523A (2.67, 103.13) ASP‐523A (2.05, 157.49)
5	8‐Methylene‐3‐oxatricyclo[5.2.0.02,4]nonane	−4.835	LEU‐421A (3.9) LEU‐421A (3.48) ASP‐426A (3.62) ASN‐429A (3.94) PRO‐466A (3.59) PRO‐466A (3.67) ILE‐525A (3.97)
6	(2E,4Z,7Z)‐trideca‐2,4,7‐trienal	−4.228	THR‐141A (3.69) PRO‐466A (3.83)	ARG‐539A (3.25, 115.63)
7	(2E,4E)‐deca‐2,4‐dienal	−3.945	LEU‐421A (3.8) ASP‐426A (3.73) PRO‐466A (3.65) PRO‐466A (3.83)
8	(2E)‐dodec‐2‐en‐4‐yne	−4.383	THR‐141A (3.89) LEU‐421A (3.67) LEU‐421A (3.62) ASN‐429A (3.71) PRO‐466A (3.71) PRO‐466A (3.67) ILE‐525A (3.78)
9	6,10‐Dimethyl‐undeca‐5,9‐dien‐2‐one	−4.794	THR‐141A (3.8) LEU‐421A (3.58) ASN‐429A (4) PRO‐466A (3.9)
10	Dodecan‐1‐ol	−4.278	LEU‐421A (3.7) PRO‐466A (3.7)	ALA‐427A (2.51, 129.13)
11	Tridecan‐2‐one	−4.644	LEU‐421A (3.69) LEU‐421A (3.8) LEU‐421A (3.66) PRO‐466A (3.6) VAL‐468A (3.81) ILE‐525A (3.85)	ARG‐513A (3.09, 104.27)
12	2,4‐di‐tert‐butylphenol	−5.493	LEU‐421A (3.57) ASP‐426A (3.85) ASN‐429A (3.95) PRO‐466A (3.78) PRO‐466A (3.83) ILE‐525A (3.68)	ARG‐539A (2.8, 121.3)
13*	N, N‐Diethyl‐4‐methylbenzamide (diethyltoluamide)	−5.153	LEU‐421A (3.76) PRO‐466A (3.67)	ASP‐523A (3.56, 104.37) ASP‐523A (2.56, 127.18)
14	Hexadecane‐1‐ol	−4.005	LEU‐421A (3.7) LEU‐421A (3.53) ASP‐426A (3.94) PRO‐466A (3.7)
15	Tetradecane	−4.311	ILE‐67A (3.93) LEU‐421A (3.81) LEU‐421A (3.69) ASN‐429A (3.87)
16	(E)‐10‐methyltridec‐11‐en‐1‐yl propionate	−4.502	LEU‐421A (3.67) ASN‐429A (3.72) PRO‐466A (3.89)
17	Benzophenone	−5.645	LEU‐421A (3.55) ASP‐426A (3.67) PRO‐466A (3.7) ILE‐525A (3.39)
18	Methyl 2‐(3‐oxo‐2‐pentylcyclopentyl)acetate	−4.947	LEU‐421A (3.88) LEU‐421A (3.58) ASN‐429A (3.89) PRO‐466A (3.82)	ARG‐539A (3.34, 120.12)
19	Cyclotridecane	−2.905	TYR‐17A (3.61) VAL‐19A (3.55) ALA‐136A (3.55) ASP‐426A (3.97) ALA‐427A (3.62) ALA‐427A (3.68)
20	2‐(tert‐butyl)‐5‐(4,4‐dimethoxy‐pentyl)‐5‐methyl‐1,3‐dioxolan‐4‐one	−5.642	THR‐141A (3.77) LEU‐421A (3.63) ASN‐429A (3.96) ILE‐525A (3.55)	ARG‐539A (2.83, 118.79)
21	Nonadec‐1‐ene	−4.610	THR‐141A (3.75) LEU‐421A (3.7) LEU‐421A (3.69) ASP‐426A (3.92) ASN‐429A (3.89) PRO‐466A (3.6)
22	2‐Ethylhexyl 2‐hydroxybenzoate (2‐ethylhexyl salicylate)	−5.229	LEU‐421A (3.88) LEU‐421A (3.72) LEU‐421A (3.43) ASP‐426A (3.76) ASN‐429A (3.63) PRO‐466A (3.86) ILE‐525A (3.93)	ASP‐522A (2.99, 155.15) ASP‐523A (2.69, 107.85) ASP‐523A (2.52, 111.14) ARG‐539A (2.93, 113.34)
23	Isopropyl tetradecanoate	−4.281	THR‐141A (3.45) ASN‐429A (3.66) VAL‐468A (3.49)
Positive control	Acarbose	−7.712	ALA‐427A (3.86)	GLN‐133A (2.77, 112.5) TYR‐134A (2.66, 102.86) ALA‐136A (1.9, 152.46) THR‐141A (2.3149.28) PRO‐466A (3.2130.26) ARG‐513A (2.37133.85) ALA‐518A (2.5112.66) ASP‐522A (2.31120.82) ASP‐522A (2.91151.65)

^*In addition to hydrogen bonds and hydrophobic interactions, Compund 13 presents an additional salt bridge with ASP‐523A (4.48 Å).

Additionally, molecular docking simulations of compounds 2, 4, and 1 – which had the highest similarity index according to the NIST MS library used, with 92%, 73%, and 67%, respectively (Supporting information, Table S6) – showed modest binding energies of −4.184, −4.074, and −3.745 kcal mol⁻¹, respectively, compared to acarbose. However, these compounds showed competitive positioning within the catalytic pocket. Compound 4 interacted through an H‐bond with ASP‐522, a residue also involved in acarbose stabilization, whereas compound 1 shared the ALA‐427 binding site by hydrophobic interaction at a shorter distance. These results suggest that these molecules may exert inhibitory effects by mimicking key interactions of the reference drug within the yeast enzyme model. Finally, although acarbose displayed a higher binding affinity (−7.712 kcal mol⁻¹) due to its extensive network of nine H‐bonds, the identified VOCs from A. potatorum leaves offer a complementary inhibitory mode based on hydrophobic pocket occupation. These findings may suggest a combined effect of the different VOCs present in the extract, providing a molecular basis for the inhibitory activity observed in the in vitro experimental data. Figure 7 shows the 3D representation of the molecular docking interaction analysis of compounds 4, 13, 20, and the reference inhibitor acarbose as examples.

Figure 7.

Three‐dimensional representation of the molecular docking interaction analysis of compounds 4, 13, 20, and the reference inhibitor acarbose within the active site of Saccharomyces cerevisiae α‐glucosidase. Blue solid lines indicate H‐bonds, while grey dashed lines represent hydrophobic interactions. Amino acid residues involved in stabilizing the complexes (e.g., ASP‐522, ALA‐427) are shown as blue sticks, and the inhibitor is shown in orange. Visualizations were generated using the protein‐ligand interaction profiler (PLIP).

CONCLUSIONS

The chemical composition of the metabolites contained in agave leaves is very complex. However, regardless of the polarity used in the extraction, all extracts contain metabolites capable of inhibiting α‐glucosidase and, to a lesser extent, α‐amylase. Bio‐guided fractionation was therefore conducted using column chromatography of the hexane extract, and two fractions were selected. The two promising fractions contained hydrophobic compounds from A. potatorum Zucc. leaves and exhibited significantly higher α‐glucosidase inhibition (lower IC₅₀ values) than acarbose, although none of the fractions reached 50% α‐amylase inhibition. According to the FTIR characterization and phytochemical profile, the two fractions exhibit an unsaturated, steroidal, and/or terpenic nature characterized by aliphatic chains and oxygenated functional groups such as lactones and carbonyls. Gas chromatography–mass spectrometry identified dodecan‐1‐ol, 2,4‐di‐tert‐butylphenol, tetradecane, and benzophenone as compounds potentially contributing to inhibitory activity against α‐glucosidase and α‐amylase. From the in silico analysis, none of the VOCs identified are expected to accumulate in the body. Meanwhile, the molecular docking showed that the inhibitory effects of A. potatorum metabolites identified in silico may depend on the hydrophobic and H‐bond interactions with the catalytic pocket of Saccharomyces cerevisiae α‐glucosidase. Nonetheless, total purification, NMR characterization, and/or mass spectrometry analysis are needed to confirm the structure of these compounds. This study provides a basis for further investigations of A. potatorum Zucc. leaf metabolites.

FUNDING INFORMATION

This work was supported by the Secretaria de Ciencia, Humanidades, Tecnología e Innovación (SECIHTI), formerly known as the Consejo Nacional de Humanidades, Ciencias y Tecnologías (CONAHCYT), under project ID CBF2023‐2024_3008.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

AUTHOR CONTRIBUTIONS

Elia‐Donají Juárez‐Niño: writing – original draft, methodology, investigation, and formal analysis. Angélica Cruz‐Gómez: formal analysis, writing – review, and editing. Rosa Santillan: visualization, formal analysis, writing – review. María Eugenia Ochoa‐Becerra: conceptualization and writing – review. Delia Soto‐Castro: project administration, visualization, methodology, conceptualization, formal analysis, writing – review, and editing.

Supporting information

Figure S1. FTIR spectra of purified acarbose.

Table S1. Open chromatographic column fractionation of the hexane extract of A. potatorum Zucc. leaves.

Table S2. Open chromatographic column fractionation of fraction F6 from A. potatorum Zucc. leaves.

Table S3. Methodologies of phytochemical screening.

Table S4. Identifiers of secondary metabolites identified by CG in hexane fractions of A. potatorum Zucc. leaves.

Figure S2. Chromatograms of GC of hexane fractions from A. potatorum Zucc. leaves.

Table S6. VOCs identified by GC from hexane fractions of A. potatorum Zucc. leaves.

Table S7. Physicochemical properties, lipophilicity, and solubility of VOCs identified by GC from hexane fractions of A. potatorum Zucc. leaves.

Fig. S3. Representation of boiled egg for VOCs identified by GC.

Table S8. Pharmacokinetic parameters of VOCs identified by GC from hexane fractions of A. potatorum Zucc. leaves.

Table S5. Phytochemical screening of hexanoic extract from A. potatorum Zucc. leaves and their fractions.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.