Multi-Target Herbicidal Effects of Agave lechuguilla Torr. Extract on Chenopodium album L.: Germination Inhibition, Metabolic Disruption, and Morpho-Physiological Alterations

TL;DR

A plant extract from Agave lechuguilla shows strong weed-killing effects by disrupting multiple biological processes in lambsquarters, a common weed.

Contribution

The study introduces a multi-target bioherbicide from Agave lechuguilla that disrupts germination, metabolism, and cell structure in weeds.

Findings

AGE extract inhibited seed germination in a concentration-dependent manner.

AGE caused metabolic disruption, including reduced carbohydrate content and altered enzyme activity.

AGE significantly reduced cell area by 51.1% and suppressed metabolic heat output.

Abstract

The pursuit of sustainable alternatives has spurred interest in plant-derived bioherbicides with multi-target actions to combat resistance. This study explored the herbicidal potential of Agave lechuguilla extract (AGE) against the widely problematic weed Chenopodium album L. (lambsquarters). Various methods, including germination assays, biochemical profiling, measurements of antioxidant enzyme activity, isothermal microcalorimetry, and both macroscopic and microscopic morphological analyses, were employed to evaluate the effects of AGE relative to glyphosate (1.5%). The results indicated that AGE inhibited seed germination in a concentration-dependent manner, with the 30 g/L dose exhibiting the most significant effect. Treatment with 30 g/L of AGE led to a notable decrease in total carbohydrate content and catalase activity, an increase in total lipids, and an enhancement of glutathione reductase activity. Additionally, it suppressed metabolic heat output and severely disrupted root and cellular architecture, resulting in a reduction of cell area by 51.1%. In contrast, glyphosate primarily increased lipid content and induced near-complete metabolic suppression while causing distinct morphological and enzymatic alterations. The findings demonstrate that AGE functions as a multi-target pre-emergence bioherbicide, disrupting processes related to germination, metabolism, oxidative balance, and morphology through mechanisms that differ from the single-target action of glyphosate. This underscores its potential for sustainable weed management.

Figures

Full text

1. Introduction

Sustainable weed management presents a significant challenge within contemporary agriculture [1]. The widespread application of synthetic herbicides, such as glyphosate, has generated substantial concerns regarding environmental contamination, adverse effects on non-target organisms, and the proliferation of herbicide-resistant weeds in valued crops [2]. This has necessitated a shift towards environmentally friendly weed management alternatives, including the search for and development of bioherbicides derived from plants, insects, or microbial extracts [3]. These bioherbicides can offer multiple modes of action, thereby reducing selection pressure and the development of resistance [4].

Allelopathy is a biological phenomenon in which plants produce bioactive compounds that can inhibit the growth of competing plants [5]. In this context, plant-derived bioherbicides are rich in allelopathic compounds [6]. The genus Agave is a promising source of bioactive compounds with phytotoxic potential [7]. Among its members, Agave lechuguilla Torr. is among the least studied for its herbicidal properties. This plant is harvested in northern and central Mexico for its vegetable fiber known as ‘ixtle’ (Tampico fiber), which is utilized in the production of brushes, scouring pads, bags, sacks, ropes, and various handicrafts [8,9]. Notably, this fiber constitutes only 15% of the plant; the remaining material, referred to as ‘guishe’ (waste), is often discarded. Consequently, there is growing interest in valorizing this byproduct, particularly for its potential as an extract containing bioactive compounds with herbicidal activity [10]. Several Agave species have been identified as having high steroidal saponin content, suggesting their potential as bioherbicides [11]. Chenopodium album L. (lambsquarters) ranks among the top 10 most problematic weeds globally, affecting a range of crops, including wheat, barley, chickpea, canola, maize, soybean, sorghum, potato, tomato, and sugar beet. This annual weed is widely distributed across North America and has shown resistance to multiple herbicide classes [12,13,14]. Given these characteristics, its global agricultural impact, and documented herbicide resistance, C. album was selected as the target species to evaluate the herbicidal effects of AGE [12]. Notably, the in vitro efficacy of Agave lechuguilla extract has been established as a promising pre-emergence bioherbicide in seed models [15,16]. Furthermore, in our previous transcriptional study, we emphasized the extract’s multi-target potential on the seeds of the weed C. album. While we identified the multi-target mode of action of this extract, the aim of this study is to determine its physiological, metabolic, and biochemical effects, as well as its macro- and micromorphological changes, using commercial glyphosate as a reference single-target comparator, because our study focused primarily on transcriptional analysis (mRNA-seq) [17].

While glyphosate is mainly utilized as a post-emergence herbicide, research has demonstrated its effectiveness against glyphosate-resistant weeds when applied pre-emergence [18,19]. The inclusion of this compound in the study is warranted due to its well-defined, single-target mechanism of action, specifically the inhibition of 5-enolpyruvylshikimate-3-phosphate (EPSP) synthase, which serves as a reliable reference point for comparison. We recognize the conceptual distinction between the typical post-emergence application of glyphosate and the pre-emergence evaluation of AGE. Nevertheless, glyphosate remains an important “single-target reference” because its extensive physiological and morphological effects are thoroughly documented, allowing us to highlight the multi-target, distinct mechanisms of AGE [20,21,22].

In this study, a range of key methodologies will be employed to achieve the stated objective. First, germination tests, biochemical profiling, assessments of antioxidant enzyme activity, isothermal microcalorimetry, and both macroscopic and microscopic evaluations will be conducted. These approaches aim to test the initial hypothesis that AGE exerts a multi-target inhibitory effect on the germination and seedling development of C. album, and to explore potential alterations beyond the glyphosate target.

2. Results

2.1. Inhibitory Effect on the Germination of Chenopodium album Seeds Exposed to the Agave lechuguilla Extract (AGE)

The tetrazolium chloride test indicated that of 97 seeds, 80 were viable and 17 were non-viable, yielding a viability rate of 82%.

The cumulative germination of C. album in the control group reached a maximum of 83% after 96 h. In contrast, the AGE treatments exhibited a concentration-dependent inhibitory effect. Notably, the 75 g/L and 30 g/L AGE treatments exhibited stronger inhibitory effects than 1.5% glyphosate and 10 g/L AGE, respectively (Figure 1).

The cumulative germination data for the control group align with the seed viability percentage obtained from the tetrazolium chloride test, which indicated a viability of 82.5%. This supports our observation of a cumulative germination rate of 83% in the control group at 96 h, confirming that low seed viability did not cause the observed inhibition of germination in the treatment groups.

Because the data did not meet the assumptions for a two-way ANOVA, we conducted a statistical analysis using a generalized linear model (GLM) with a binomial distribution. This analysis confirmed highly significant effects of both treatment (χ^2^ = 158.84, p < 0.0001) and time (χ^2^ = 43.82, p < 0.0001) on germination. Multiple comparisons conducted with Tukey’s test (Table 1) revealed that all treatments differed significantly from the control (p < 0.0001), with the 30 g/L AGE treatment exhibiting the most pronounced inhibitory effect (OR = 0.052, 95% CI: 0.030–0.08) (Table S1). No significant differences were observed between the 10 g/L treatments and the 75 g/L + 1.5% glyphosate combination (p = 0.91). Regarding the time factor, germination rates significantly increased from 24 to 48 h (OR = 2.65, 95% CI: 1.68–4.24) and reached a stabilization point after 72 h (Table S1).

2.2. Differential Biochemical Response of Seeds and Seedlings to Glyphosate and Agave lechuguilla Extract

Analysis of total carbohydrate content revealed statistically significant differences among the treatments (Figure 2A). The control group exhibited a mean concentration of 94.63 ± 10.2 mg/g. In contrast, treatment with the Agave lechuguilla extract (AGE 30 g/L) resulted in a notable reduction, with a mean value of 77.15 ± 16.8 mg/g. Conversely, the application of glyphosate (1.5%) did not demonstrate any inhibitory effect compared to the control, with a mean value of 97.36 ± 9.5 mg/g.

The total lipid profile was the parameter most significantly impacted by glyphosate application (Figure 2B). The control group of seedlings and seeds exhibited a lipid concentration of 54.18 ± 9.3 mg/g. In contrast, a substantial, statistically significant increase in lipid levels was observed in the treated groups. The AGE treatment increased lipid concentration to 76.43 ± 15.2 mg/g, whereas glyphosate had the most pronounced effect, nearly doubling the control level to 87.84 ± 21.8 mg/g.

In contrast to carbohydrates and lipids, the total protein content remained essentially unchanged across all treatments (Figure 2C). The average values were 11.24 ± 1.4 mg/g for the control group, 11.73 ± 1.0 mg/g for the bioherbicide treatment, and 11.33 ± 0.8 mg/g for the glyphosate treatment. This stability indicates that, under experimental conditions, the treatments did not significantly impact overall protein synthesis or degradation.

2.3. Enzymatic Activity Assays

An analysis of antioxidant enzyme activity revealed distinct patterns in Chenopodium album seedlings after 72 h of treatment with Agave lechuguilla extract (AGE) at 30 g/L and glyphosate at 1.5% (Figure 3). Catalase (CAT) activity was significantly reduced in the AGE treatment (25.43 ± 8.4 U/mg protein) compared to the control (203.13 ± 21.9 U/mg protein) and the glyphosate treatments (236.85 ± 87.51 U/mg protein) (Figure 3A). Conversely, ascorbate peroxidase (APX) decreased its activity in AGE (6.2 ± 1.4 U/mg protein) and glyphosate (5.93 ± 1.1 U/mg protein) relative to the control (9.82 ± 0.3 U/mg protein) (Figure 3B).

Superoxide dismutase (SOD) exhibited a moderate decline in activity under both herbicide treatments, with a more pronounced decrease observed in the AGE treatment (234.8 ± 33.0 U/mg protein) than in the glyphosate treatment (258.5 ± 5.1 U/mg protein), compared to the control (289.2 ± 88.3 U/mg protein) (Figure 3C). Glutathione reductase (GR) showed a significant increase in activity, with glyphosate (0.204 ± 0.05 U/mg protein), and AGE (0.131 ± 0.07 U/mg protein) compared to the control (0.120 ± 0.01 U/mg protein) (Figure 3D).

The antioxidant enzyme profile of Chenopodium album seedlings exhibited significant variations in the activity of ascorbate peroxidase (APX), superoxide dismutase (SOD), and catalase (CAT) enzymes across different treatments, as illustrated in the heat map based on Z-score values (Figure 3E). In the control group, APX showed a positive Z-score of 1.15, whereas SOD and CAT showed negative Z-scores of −0.50 and −0.65, respectively, indicating diminished activity relative to the mean. Under AGE treatment, APX levels remained elevated at 1.04; however, SOD decreased by 0.95, suggesting potential inhibition of this enzyme. Conversely, the glyphosate treatment produced an opposing pattern, with a reduction in APX activity to −0.76 and an increase in CAT activity to 1.13.

2.4. Effect of Treatments on the Root Morphology of Seedlings

Macroscopic evaluation of seedlings revealed significant morphological differences among the treatments applied (Figure 4). Seedlings from the control group exhibited a typical root system, characterized by well-pigmented radicles and an abundance of root hairs. In contrast, seedlings treated with AGE (30 g/L) displayed reduced root pigmentation, seemingly shorter root lengths, and a marked absence of root hairs when compared to the control. The glyphosate (1.5%) treatment produced the most pronounced phenotypic changes, with seedlings exhibiting the nearly complete absence of root pigmentation. Although the root length in this treatment was slightly greater than that observed with the AGE, it was characterized by a total absence of root hairs.

These qualitative observations were validated and quantified through image analysis using WinRhizo (Table 2). Statistical analysis demonstrated that the AGE treatment significantly reduced total root length by 68% compared to the control group (p < 0.001), leading to values even lower than those recorded for the glyphosate treatment. Similarly, the projected area and root volume experienced substantial reductions of 44% and 59%, respectively, in response to the AGE treatment compared to the control (p < 0.001). In contrast, glyphosate exhibited a moderate inhibitory effect, significantly decreasing root volume by 31% compared to the control, although no significant differences were observed in total length relative to the control. Additionally, parameters such as total surface area and average diameter did not reveal statistically significant differences across the treatments.

Principal component analysis (PCA) demonstrated a clear distinction between the root morphological profiles resulting from the various treatments (Figure 5). The first two components (PC1 (Dim1) and PC2 (Dim2)) accounted for 86% of the total variance in the data, with PC1 alone explaining 67.9% of the variance (Figure S1). This component primarily differentiated control seedlings from those treated with AGE or glyphosate. Control replicates clustered tightly in the negative region of PC1 values, while the AGE and glyphosate treatments occupied the positive region. Notably, the AGE treatment exhibited the most distinct morphological profile, clustering at the positive end of PC1, which clearly separated it from both the control and glyphosate treatments. This clustering pattern suggests that AGE not only inhibits growth but also induces a qualitatively unique and significantly altered root architecture.

2.5. Isothermal Microcalorimetry Detects Metabolic Inhibition in Treated Seeds

The metabolic activity of Chenopodium album seeds was measured using isothermal microcalorimetry (IMC), revealing that both Agave lechuguilla extract (AGE) and glyphosate treatments significantly suppressed this activity (Figure 6).

The total heat generation (J) was the integral of the heat-flow curve over the experimental period. Control seeds exhibited the highest metabolic activity, with a mean heat production of 9.46 ± 1.59 J. Treatment with AGE at 30 g/L resulted in a significant reduction in metabolic heat to 4.19 ± 0.42 J (p < 0.01). Glyphosate at a concentration of 1.5% showed the most pronounced inhibitory effect, lowering heat generation to 1.52 ± 0.30 J (p < 0.01) (Figure 6A).

The real-time heat flow profiles provided additional insights into the temporal dynamics of metabolic inhibition. In the AGE treatment system (Figure 6B), AGE-treated seeds exhibited lower heat flow than the water control throughout the measurement period. Interestingly, the AGE control (without seeds) exhibited detectable heat flow, indicating microbial activity within the extract. However, the heat flow from seeds treated with AGE was significantly above the background level, confirming that seed metabolic activity was real, although it was strongly inhibited.

In the glyphosate treatment system (Figure 6C), metabolic suppression was even more dramatic. Seeds treated with glyphosate exhibited minimal heat flow, which was barely distinguishable from the glyphosate control without seeds. This near-complete absence of metabolic activity stands in stark contrast to the robust heat flow observed in the seed + water control, emphasizing glyphosate’s potent herbicidal action through metabolic shutdown.

The differing patterns of inhibition between AGE and glyphosate suggest distinct mechanisms of action. While both treatments significantly reduced seed metabolism, AGE allowed for some residual metabolic activity, whereas glyphosate virtually eliminated it, reducing metabolic activity to levels comparable to those observed in seedless controls.

2.6. Effects of Agave lechuguilla Extract and Glyphosate on Cellular Morphology of Chenopodium album

After 72 h of exposure to various treatments, measurements were taken of the width, length, and cell area surrounding the hypocotyls of the seedlings. The results indicated a significant reduction in all measured parameters for the treatments involving Agave lechuguilla extract (AGE) at 30 g/L and glyphosate at 1.5% (Figure 7).

The effects of the different treatments on the cellular morphological parameters of Chenopodium album hypocotyl cells are summarized in Table 3 and illustrated in Figure 8. Linear mixed model analysis revealed highly significant treatment effects on all three morphological parameters (width, length, and area; p < 0.0001 for all global F-tests).

For cellular width, Type III ANOVA of the linear mixed models revealed a significant treatment effect on cellular length (F = 117.51, p < 0.0001). Control cells showed the greatest width (21.51 ± 0.39 µm), followed by glyphosate-treated cells (15.83 ± 0.37 µm) and AGE-treated cells (13.17 ± 0.49 µm). Post hoc Tukey comparisons confirmed that all three treatments differed significantly from each other (p < 0.0001 for all pairwise comparisons). Specifically, AGE treatment reduced cell width by 38.8% compared to controls, while glyphosate reduced it by 26.4%.

For cellular length, a significant treatment effect was also observed (F = 37.56, p < 0.0001), with control cells being the longest (28.41 ± 0.55 µm). Both glyphosate (22.61 ± 0.50 µm) and AGE (21.67 ± 0.86 µm) treatments significantly reduced cell length compared to controls (p < 0.0001 for both comparisons), resulting in reductions of 20.4% and 23.7%, respectively. However, no significant difference in cell length was detected between glyphosate and AGE treatments (p = 0.5697).

The most pronounced effect was observed for total cellular area (F = 163.21, p < 0.0001), which integrates changes across both dimensions. Control cells maintained the largest area (620.7 ± 13.3 µm^2^), while glyphosate treatment reduced cellular area by 38.5% to 381.9 ± 11.0 µm^2^, and AGE treatment caused a more substantial reduction of 51.1%, resulting in an area of 303.7 ± 17.2 µm^2^. All pairwise comparisons among treatments were statistically significant (p < 0.001), with AGE showing significantly greater inhibition than glyphosate (p = 0.0002).

The box plot visualization in Figure 8A–C provides a graphical representation of these treatment effects, clearly illustrating separation between treatment groups for all parameters. Notably, the area data (Figure 8C) demonstrates significant treatment effects and greater variability within the AGE treatment group, suggesting diverse cellular responses to this extract.

3. Discussion

In our previous research, we demonstrated that the Agave lechuguilla (AGE) extract at 30 g/L elicited a multi-target response, positioning it as a potential pre-emergence bioherbicide. This study provides physiological and functional validation of the observed transcriptional response. It found that transcriptional dysregulation induced by AGE—especially in genes associated with energy metabolism and oxidative stress—results in significant inhibition of germination, distinct metabolic changes, considerable alterations in oxidative enzymes, and a cessation of root and cell growth. This study confirms its multi-target herbicidal efficacy [17].

The cumulative germination results indicate a significant, dose-dependent inhibition of germination beginning at 30 g/L of AGE. The tetrazolium chloride assay supports this finding, confirming that the observed inhibition is not due to reduced seed viability, as germination outcomes are nearly identical to those from the cell viability test. The generalized linear model (GLM) revealed significant differences between treatments and over time, highlighting an inhibitory effect on germination, specifically in the 30 g/L AGE treatment. While various germination inhibitors are known, they generally function by inhibiting cotyledon elongation, with the molecular mechanisms differing depending on whether the inhibitor is hormonal or metabolic [23]. In this instance, we suspect multi-target response, as indicated in our previous research. Although glyphosate is not classified as a pre-emergence herbicide, it was used in this work as a control. This is due to its known molecular mechanism, which inhibits the enzyme 5-enolpyruvylshikimate-3-phosphate (EPSP) synthase [24]. Additionally, a reduction in germination was observed compared with the control group, reinforcing previous findings that glyphosate can interfere with the development of Chenopodium album seedlings [25].

The changes observed in total carbohydrate content indicate a significant reduction following AGE treatment at 30 g/L, whereas no such effect was observed with glyphosate. This suggests a potential metabolic disruption or energy mobilization during germination, making AGE a possible early target for investigation. In the Astragalus membranaceus model, it has been demonstrated that methyl jasmonate (MeJa) and abscisic acid (ABA) inhibit both germination and post-germination development, leading to alterations in lipid and protein mobilization and excessive carbohydrate energy consumption [26]. Our transcriptomic study identified genes associated with hormonal disruptions induced by AGE, which may account for the reduced carbohydrate content observed in AGE-treated seedlings. Regarding lipids, a statistically significant increase is observed in both the AGE and glyphosate treatments. Research indicates that during ABA-induced germination inhibition, triglyceride breakdown is suppressed, leading to their accumulation [27,28]. This phenomenon of mediated inhibition may account for the observed increase in lipid reserves in both treatments.

Enzymatic activity in seedlings after 72 h of exposure to AGE and glyphosate revealed several notable differences. Notably, catalase (CAT) activity was significantly suppressed in the AGE treatments, while ascorbate peroxidase (APX) decreased in response to both AGE and glyphosate treatments. Conversely, glutathione reductase (GR) exhibited a significant increase in both conditions. These patterns of oxidative stress can be understood in relation to the mechanisms of action for both treatments. Previous reports indicate that certain allelochemicals, such as p-hydroxybenzoic acid, inhibit cucumber root growth, an effect associated with elevated reactive oxygen species levels in root tips [29]. Enzyme kinetics studies have demonstrated that exceeding a threshold H_2_O_2_ concentration can inactivate the enzyme via a substrate suicide mechanism, which may explain the reduced CAT activity observed in the AGE treatment [30].

The enzymatic activity of ascorbate peroxidase (APX) tends to increase in response to defense mechanisms; however, elevated allelochemical concentrations and prolonged exposure can decrease this activity [31]. The results discussed here are from 72 h post-exposure, a timeframe that may have contributed to enzyme inactivation. Additionally, reactive oxygen species (ROS) can damage the structure of antioxidant enzymes. Conversely, instability of the substrate (ascorbate) may lead to its depletion and, consequently, reduce enzyme activity [32]. Moreover, certain allelochemicals have been reported to inhibit the expression of genes that encode the APX enzyme [33].

The results for GR activity compared to APX and CAT show a differential antioxidant response, since while CAT and APX decrease in the AGE treatment, GR increases. It has been demonstrated in situ that the concentration of specific allelochemicals increases progressively over time; this is likely to cause the peroxisomal (CAT) and chloroplastic (APX) membranes to release these enzymes, and the H_2_O_2_ concentration inactivates them. One reason for the increased activity of glutathione reductase (GR) may be linked to the backup phenomenon, which occurs when ascorbate peroxidase (APX) and catalase (CAT) are unable to effectively neutralize hydrogen peroxide (H_2_O_2_). This failure results in a significant accumulation of glutathione-related stress in the plant. In response, GR activity rises to restore the balance between reduced and oxidized glutathione (GSH/GSSG) [34]. In situ, it has been demonstrated that the concentration of specific allelochemicals increases progressively over time. This likely leads to the release of these enzymes from the peroxisome (for CAT) and chloroplast (for APX) membranes, while the increased concentration of hydrogen peroxide (H_2_O_2_) inactivates the enzymes [35]. One reason for GR activity may be the backup phenomenon, in which APX and CAT fail to neutralize H_2_O_2_. The plant undergoes massive glutathione-related stress, and in response, GR activity increases to restore the reduced/oxidized glutathione (GSH/GSSG) balance [36].

The consistent activity of superoxide dismutase (SOD) can be attributed to several physiological mechanisms. The most likely explanation is its role as a “first line of defense” against reactive oxygen species (ROS). Research has shown that many plants maintain a naturally high-to-moderate baseline level of SOD to regulate the oxidative burst, particularly when neither treatment increases SOD activity [37,38].

Cellular changes, along with macroscopic and microscopic morphological effects in the roots, indicate a reduction in cell numbers. The application of AGE resulted in a significant decrease in root length and volume, as well as in the formation of root hairs. Principal Component Analysis (PCA) results indicated a distinct root architecture relative to the control and glyphosate treatments. Although both treatments resulted in notable inhibition of cell expansion, this may be linked to biochemical disruptions that physically hinder macroscopic and cellular growth. The resulting oxidative stress, along with disturbances in energy and metabolism, compromises cell integrity and turgor pressure, leading to the observed morphological changes [39]. These findings are intriguing, as they suggest that AGE compounds can disrupt membranes, as many reported allelochemicals do [40]. In our previous transcriptomics study, we identified cytoskeletal genes that were downregulated during cell expansion. This macroscopic and microscopic evidence supports our previous transcriptional observations regarding protein-protein networks.

The total heat released during germination was calculated by integrating the heat curves. As expected, metabolic suppression was observed in both the AGE and glyphosate treatments. However, in the case of glyphosate, the suppression was nearly complete. This finding suggests that the mechanisms of action for the two treatments differ significantly. The results are consistent with glyphosate’s known mechanism of disrupting the shikimate pathway. In contrast, the AGE treatment exhibits multiple biochemical and morphological signatures, suggesting a multi-target mechanism that affects multiple metabolic pathways, thereby reducing metabolism. This observation is consistent with our previous transcriptional study, in which we noted alterations in oxidative phosphorylation and ATP synthesis, as well as disruptions in carbohydrate metabolism. Similar effects have been reported with specific allelochemicals, which also interfere with energy metabolism [41].

The combined results of this study validate AGE’s multi-target mode of action as a potential bioherbicide, as we previously suggested in our transcriptional research. The physiological responses observed—such as inhibition of germination, oxidative stress, carbohydrate depletion, and stunted growth—occur in the presence of AGE and are coordinated reactions. A crucial next step in the research is to isolate and identify the bioactive compounds responsible for these physiological effects. Additionally, proteomic validation is necessary to confirm enzymatic and physiological predictions. One major concern with bioherbicides is the potential for developing resistance. However, the advantage of this extract lies in its multi-target mode of action, which could mitigate the rapid emergence of resistance. Nonetheless, long-term selective studies are needed to support this hypothesis.

4. Materials and Methods

4.1. Biological Material

Agave lechuguilla extract (AGE) was obtained from collectors in the community of Cosme, located in the Ramos Arizpe municipality of Coahuila de Zaragoza, Mexico, who extract ixtle (tampico fiber). The fiber-free bagasse of A. lechuguilla was transported in polypropylene containers to the Environmental Biotechnology Laboratory at the Universidad Autónoma de Coahuila. The juice was extracted using a hydraulic press with a force of 1.014 kN. This juice was then mixed in a home blender (Proctor Silex, 62515RY) at maximum speed until a consistent foam was produced. The foam was subsequently dried in a tray dryer (Koleff, KL-10) at 50 °C for 3 h. The resulting powder was stored in airtight containers until needed. The bagasse waste collection adhered to the official Mexican standard NOM-008-SEMARNAT-1996, which permits the use of Agave lechuguilla waste without authorization.

Seeds of the weed Chenopodium album L. were collected from an agricultural field in the community of San Rafael, Veracruz, Mexico. Although it is neither an endangered nor a threatened plant, a collection permit was requested from the Instituto Nacional de Investigaciones Forestales, Agrícolas y Pecuarias (INIFAP). After transporting the seeds to the laboratory, molecular identification was performed using the ITS and tnrL genes (PP668985 and PP658427). Once identified, the seeds were deposited in the HCIB herbarium (Centro de Investigaciones Biológicas del Noroeste, S.C.) under the following accession number: HCIB32366.

4.2. Germination and Viability of Chenopodium album Seeds

To check the viability of C. album seeds, 97 seeds were soaked overnight at room temperature in distilled water. After rinsing with distilled water, the seeds were soaked overnight in a 0.5% tetrazolium chloride solution. The seeds were then washed with distilled water and allowed to dry at room temperature. Once dry, the seed coats were carefully broken using a scalpel. The seeds were visually inspected under a stereoscope, and the intensity of the staining was used to classify them as either viable or nonviable [42].

A 96 h germination assay was conducted to evaluate seed germination. Murashige-Skoog (MS) medium, supplemented with 6% agar, was prepared with three concentrations of A. lechuguilla extract powder (AGE): 10 g/L, 30 g/L, and 75 g/L. Additionally, a treatment with 1.5% glyphosate (in the form of isopropyl amine salt, from the Arraza brand, Agrochemical Tridente, Coyoacán, Mexico City, Mexico) was included, along with a control group containing only the MS medium.

The MS agar was poured into 100 × 15 mm Petri dishes under sterile conditions. To prepare, the seeds were washed with a 0.1% solution of Tween-20, followed by a 5 min immersion in a 3% solution of commercial sodium hypochlorite (HClO). The seeds were then rinsed five times with distilled water to remove any residual hypochlorite.

The washed seeds were soaked overnight in their respective treatments: control seeds were placed in distilled water only, AGE seeds were soaked in the specified concentrations (10 g/L, 30 g/L, and 75 g/L), and glyphosate seeds were treated with 1.5% glyphosate. A total of 12 seeds were placed in each Petri dish, with four replicates per treatment. The plates were sealed with plastic film and incubated at 25 °C under a 12 h photoperiod. Germination was recorded every 24 h.

Cumulative germination data, expressed as the proportion of germinated seeds, were analyzed using a generalized linear model (GLM) with a binomial distribution and a logit link function. Since complete separation occurred in the 75 g/L treatment at 24 h (0% germination in all replicates), the treatments were grouped into three categories: low concentration (10 g/L), medium concentration (30 g/L), and high concentration (75 g/L + 1.5% glyphosate), with the control serving as the reference. The final model included the main effects of treatment and time, with the interaction term excluded because it was not significant. Overdispersion was assessed using the ratio of residual deviance to degrees of freedom. The value obtained was 1.64, then a quasi-binomial model was used for the final inferences. Multiple comparisons between treatments and time points were performed using Tukey’s post hoc tests, with the estimated marginal means (emmeans) method and a significance level of α = 0.05. All analyses were performed in R v4.3.1, using the stats (https://cran.r-project.org/web/packages/STAT/index.html, accessed on 23 January 2026), emmeans (https://cran.r-project.org/web/packages/emmeans/index.html, accessed on 23 January 2026), and multcomp (https://cran.r-project.org/web/packages/multcomp/index.html, accessed on 23 January 2026) packages (Script S1).

4.3. Biochemical Composition of Germinated Seeds and Seedlings

An experiment was conducted under the conditions described in Section 4.2, with five replicates for each treatment: 30 g/L AGE, 1.5% glyphosate, and controls. Samples (ungerminated seeds and seedlings) were taken 72 h after germination was assessed and placed in 1.5 mL microcentrifuge tubes, where their fresh weight was recorded. The samples were then frozen at −80 °C and lyophilized for 16 h (VirTis, model 2KBTES-55). The dry weight of the lyophilized material was recorded, and the material was then pulverized in a FastPrep-40 bead homogenizer (MP Biomedicals, Santa Ana, CA, USA) for 40 s at 6 m/s. The powder was mixed with 100 µL of 20% trichloroacetic acid (TCA) and centrifuged at 5000× g for 10 min at 5 °C.

The total carbohydrate test was conducted using 25 µL of supernatant recovered, and 250 µL of 0.01% anthrone dissolved in 96% H_2_SO_4_ was added. The mixture was heated at 85 °C for 10 min, then placed in an ice bath. A 200 µL sample was read in a microplate spectrophotometer (MultiSkan FC ThermoFischer Scientific, Waltham, MA, USA) at 630 nm. A glucose curve of known concentrations (100, 50, 25, 12.5, 6.25, and 3.125 mg/dL) was used. From this curve, the total carbohydrate concentration was calculated in mg/g dry weight [43].

For total lipid test was performed on a 25 µL aliquot of the above supernatant, prepared in 20% TCA, to which 250 µL of concentrated H_2_SO_4_ was added. The tubes were shaken and heated at 90 °C for 10 min. The tubes were chilled on ice. A 20 µL sample was taken, and 200 µL of sulfo-phospho-vanillin was added. The mixture was incubated at room temperature for 40 min. The sample was read at 540 nm, and a standard curve was constructed using a triglyceride standard (200, 100, 50, 25, 12.5 mg/dL). Total lipid concentrations are reported in mg/g dry weight [44].

Total protein test was conducted using 25 µL of supernatant, and 500 µL of 0.1 N NaOH, and the mixture was left to digest for 2 h. 25 µL of the digest was taken, and 200 µL of BCA (bicinchoninic acid) reagent was added, incubated at 60 °C for 15 min, and read at 562 nm. A standard curve was performed using bovine serum albumin (BSA) (100, 50, 25, 12.5, and 6.25 mg/mL). The total protein concentration was calculated and reported in milligrams per gram of dry weight [45].

As the data did not conform to the assumptions of normality (evaluated using the Shapiro–Wilk test) and homoscedasticity (assessed via the Levene test), a nonparametric analysis of variance on ranks was employed, specifically the Kruskal–Wallis test. To identify differences among treatment groups (control, AGEs, and glyphosate), a post hoc Tukey’s multiple comparisons test was conducted (p < 0.05). All analyses were performed using SigmaPlot v12.0 (Systat Software, Inc., San Jose, CA, USA).

4.4. Determination of the Activity of Antioxidant Enzymes

In the experiment detailed in Section 4.2, five samples of seedlings and seeds, each weighing approximately 100 mg, were collected after 72 h. These samples were placed in 1.5 mL microcentrifuge tubes, and their fresh weights were recorded. To each tube, 250 µL of homogenization buffer (0.2 M potassium phosphate buffer (PPB), 1 mM EDTA, 1% PVP, 10% glycerol, pH 7.0) was added per 100 mg of fresh weight. The samples were then homogenized using a glass pestle in an ice bath. Following homogenization, the samples were centrifuged at 12,000× g for 10 min at 4 °C. The supernatant was carefully collected and stored in a clean tube, while the pellet was resuspended in an additional 250 µL of homogenization buffer per gram of fresh weight. This mixture was centrifuged again under the same conditions, and the new supernatant was combined with the previously collected one. This crude extract was subsequently used to measure enzymatic activity [5].

To quantify the specific enzymatic activity for each assay, we measured protein content by the fresh weight using the Bradford method. This was performed in a microplate format, with a calibration curve established using bovine serum albumin (BSA). The plates were read at 595 nm (3550-UV, Bio-Rad, Hercules, CA, USA).

To evaluate catalase (CAT) activity, a 1:200 dilution of the crude extract was introduced into a quartz cuvette containing 10 mM H_2_O_2_ in 50 mM potassium phosphate buffer (pH 7.0). The kinetics of H_2_O_2_ consumption were monitored by measuring absorbance at 230 nm at 10 s intervals for 2 min. For the spectrophotometer blank (Helios Omega UV-VIS, Thermo Scientific, Waltham, MA, USA), only the 10 mM H_2_O_2_ in phosphate buffer at pH 7.0 was analyzed. CAT activity was calculated using linear regression to determine the slope (ΔA/min), considering the extinction coefficient (ε) of H_2_O_2_, which is 0.0436 µM^−1^·cm^−1^. Catalase activity was expressed as units per milligram of protein, with 1 unit defined as the decomposition of 1 µM of H_2_O_2_ per minute [46].

The measurement of ascorbate peroxidase (APX) activity uses ascorbic acid as a substrate to reduce hydrogen peroxide (H_2_O_2_) via oxidation to dehydroascorbate (DHA). In this process, a mixture consisting of 50 mM BBP (pH 7.0), 0.5 mM H_2_O_2_, 0.5 mM ascorbic acid (AA), and a 1:100 dilution of crude extract was prepared in a quartz cuvette. The absorbance was measured at 290 nm every 30 s for 3 min. The spectrophotometer blank included the entire mixture, excluding the extract. APX activity was calculated using linear regression to determine the slope (ΔA/min), considering the extinction coefficient (ε) of ascorbate, which is 0.028 µMol^−1^·cm^−1^. APX activity was expressed as units per milligram of protein, with 1 unit defined as the decomposition of 1 µMol of ascorbate per minute [47].

To measure Superoxide Dismutase (SOD) activity, the method relies on the enzyme’s ability to inhibit the reduction in Nitro Blue Tetrazolium (NBT) to formazan by superoxide radicals generated via chemical oxidation. A reaction mixture was prepared at a final volume of 2 mL, consisting of 50 mM BPP at pH 7.8, 2 mM EDTA, 9.9 mM L-methionine, 55 µM NBT, and 0.025% Triton X-100, and was adjusted with deionized water. To this mixture, 40 µL of a 1:2 dilution of the extract and 20 µL of 1 mM riboflavin were added, and the mixture was gently mixed. The reaction was conducted in duplicate: one sample was exposed to a 15 W white-light tube positioned 12 cm away, with constant oscillation for 10 min, while the other was kept in total darkness (the reaction blank). Absorbance was measured at 560 nm using polypropylene test cells for both the light-exposed samples and the dark samples. SOD activity was expressed as units (U) of SOD per mL sample, with 1 U of SOD defined as the amount of enzyme needed to inhibit 50% of NBT reduction [48].

To assess the activity of the enzyme glutathione reductase (GR), the assay reduced GSSG (glutathione disulfide) with DNTB (5,5-dithiobis (2-nitrobenzoic acid)) in the presence of NADPH, producing TNB^−^ (2-nitro-5-thiobenzoate), which absorbs light at 412 nm. In a 1 cm polypropylene cuvette, a mixture is prepared to a final volume of 1 mL, containing 50 mM BPP and 2 mM EDTA, pH 7.8, with the addition of 0.75 mM DNTB, 0.1 mM NADPH, and a 1:100 dilution of the plant extract. The addition of 1 mM GSSG initiates the reaction, and absorbance is measured at 412 nm every 15 s for 5 min. For the blank control, the same mixture is used, except that the plant extract is omitted. GR activity was calculated using linear regression to determine the slope (ΔA/min), considering the extinction coefficient (ε) of TNB^−^, which is 14.15 M^−1^·cm^−1^. GR activity was expressed as units per milligram of protein, with 1 unit defined as the formation of 1 M of TNB^−^ per minute [49].

Antioxidant enzyme activity data (CAT, APX, SOD, GR) were analyzed using SigmaPlot v12.0 (Systat Software). Before analyzing variance, the data were assessed for compliance with parametric assumptions through normality testing (Kolmogorov–Smirnov) and homoscedasticity testing (Levene). Once verified, a one-way ANOVA (p < 0.01) was performed to assess whether significant differences existed among treatments. To identify specific group differences, Duncan’s post hoc test was applied (p < 0.01).

For the multivariate analysis of the antioxidant enzyme profile, hierarchical clustering was performed in R version 4.3.1. Previously, the enzyme activity data (CAT, APX, SOD, GR) were normalized using a Z-score transformation for each enzyme, thereby eliminating biases arising from differing measurement scales and enabling multivariate comparisons. Normalization was calculated as follows: Z = (X − μ)/σ, where X represents the enzyme activity value, μ denotes the mean of the corresponding enzyme, and σ indicates its standard deviation. Hierarchical clustering used Euclidean distance as the dissimilarity measure, and dendrograms were constructed using complete linkage. The heatmap visualization was generated using the pheatmap (https://cran.r-project.org/web/packages/pheatmap/index.html, accessed on 4 January 2026) package (Script S2).

4.5. Measurement of Seedlings’ Root Parameters

In the experiment described in Section 4.2, at least fifteen seedlings were selected after 72 h of exposure. The radicles were carefully washed and arranged in a transparent scanning container with a thin layer of water to prevent overlap. The samples were digitized using a flatbed scanner (Epson Expression 11000XL V3.49, Nagano, Japan) at 600 dpi. The resulting images were analyzed using WinRhizo Pro (version 2013; Regent Instruments Inc., Québec, QC, Canada) to assess total length, surface area, volume, average diameter, and root volume.

The parameters were analyzed using SigmaPlot v12.0 (Systat Software) to assess whether they met the assumptions of normality and homoscedasticity. The only parameter that did not satisfy these assumptions was total length. This parameter was assessed using ANOVA on ranks (Kruskal–Wallis) (p < 0.05), while the remaining parameters were analyzed using one-way ANOVA (p < 0.05). Post hoc analysis was conducted using the Tukey test (p < 0.05).

To assess the overall impact of treatments on complete root architecture, a Principal Components Analysis (PCA) was conducted in RStudio v4.3.1, using the factoextra (https://cran.r-project.org/web/packages/factoextra/index.html, accessed on 15 October 2025) and ggplot2 (https://cran.r-project.org/web/packages/ggplot2/index.html, accessed on 15 October 2025) packages. The input data matrix (Script S3) comprised 45 replicate observations (n = 15) along with six quantitative morphological variables: total length, total projected area, total surface area, average diameter, length-to-volume ratio, and root volume. Before proceeding with the analysis, the data were standardized (scale = TRUE) to mitigate bias arising from differing units of measurement across variables, ensuring that each parameter contributed equally to the model. Principal components were extracted from the correlation matrix, and the variance explained by each component was computed from its eigenvalue. The graphical representation of the replicates (individuals) was plotted on the plane defined by the first two principal components (PC1 and PC2), with points color-coded according to the treatment factor (“Control,” “AGE,” “Glyphosate”) and 95% confidence ellipses added to illustrate the grouping of the treatments (Script S3).

4.6. Isothermal Microcalorimetry (IMC) Monitoring of Seed Germination

The metabolic activity of Chenopodium album seeds was assessed using an isothermal microcalorimeter (TAM III; TA Instruments Inc., New Castle, DE, USA). Three independent experiments were conducted at 25 °C under controlled conditions. The first experiment served as a control, utilizing only the seeds. Within each glass ampoule, a segment of filter paper was placed. The first ampoule contained filter paper and distilled water (seedless), while the remaining five ampoules each contained one Chenopodium album seed along with distilled water. Following a 30 min equilibration period, heat-flow measurements were recorded continuously for 1 week. The second experiment investigated the effects of Agave lechuguilla extract at a concentration of 30 g/L (AGE). Six ampoules were prepared: the first served as a seedless control with filter paper and distilled water; the second contained filter paper and AGE without a seed; the third held filter paper, distilled water, and one seed; and the remaining three ampoules included AGE (30 g/L) and one seed each. All measurements were conducted at 25 °C after a 30 min equilibration period. The third experiment mirrored the design of the second, with the sole modification of replacing AGE with glyphosate at 1.5%.

The raw data were processed using TAM Assistant software version 2.0.65.0 (TA Instruments, Inc., New Castle, DE, USA). The integrated heat flow values (total heat generated, J) were derived from the run summaries. For the AGE treatment, a correction was applied to account for endogenous microbial activity observed in the seedless AGE control; specifically, the total heat value from the AGE control without seeds was subtracted from the integrated values obtained from the AGE treatments with seeds. This correction was not necessary for glyphosate treatments, as no significant heat flow was detected. The total heat release data, obtained from triplicate measurements, were analyzed using SigmaPlot v12 (Systat Software). A one-way ANOVA (p < 0.01) followed by Duncan’s post hoc test (p < 0.01) was conducted. Before analysis, data normality was verified using the Kolmogorov–Smirnov test, and homoscedasticity was confirmed with Levene’s test. Heat flow curves were visualized using R software v4.4.2, employing the ggplot2 (https://ggplot2.tidyverse.org/, accessed on 14 January 2026), data.table (https://cran.r-project.org/web/packages/data.table/index.html, accessed on 14 January 2026), and dplyr (https://cran.r-project.org/web/packages/dplyr/index.html, accessed on 14 January 2026) packages.

4.7. Histomorphological Analysis

From the experiment detailed in Section 4.3, three seedlings per treatment group (Control, AGE 30 g/L, and Glyphosate 1.5%) were randomly selected. The plant samples were fixed for 48 h in FAA solution (formalin-acetic acid-ethyl alcohol, 10:5:85 v/v). Following this, the tissues were dehydrated through a graded series of ethanol (70%, 80%, 90%, 95%, and 100%), cleared with xylene, and infiltrated with paraffin using a LEICA ASP200S automatic tissue processor. The samples were then embedded in liquid paraffin at 60 °C and cooled to −5 °C in a LEICA EG1150H and EG1150C embedding center. Cross-sections 4 µm thick were obtained with an automatic rotary microtome (LEICA RM2155). These sections were mounted on gelatin-treated slides and deparaffinized with xylene prior to staining. To enhance cellular structures, the sections were stained with 0.1% toluidine blue (pH 4.4) for 5 min. The stained samples were examined under a light microscope (Olympus BX41, Tokyo, Japan) equipped with a digital camera (Nikon Digital Sight DS-Ri1, Shinagawa-ku, Tokyo, Japan), and digital images were captured at 20× magnification, focusing specifically on the hypocotyl region. Morphometric analysis was conducted using Image-Pro Plus v9.0 software (Media Cybernetics; Bethesda, MD, USA). For each biological replicate (n = 3 per treatment), two independent micrographs were analyzed, yielding a total of 6 micrographs per treatment. In each micrograph, the dimensional parameters of 30 epidermal and cortical cells adjacent in the hypocotyl were measured to obtain cell width (µm), defined as the maximum distance along the transverse axis, cell length (µm), the maximum distance along the longitudinal axis, and cell area (µm^2^), representing the cross-sectional area.

Cellular morphological data were analyzed using linear mixed models (LMMs) with the lme4 (v1.1.35, https://cran.r-project.org/web/packages/lme4/index.html, accessed on 7 January 2026) and lmerTest (v3.1.3, https://rdrr.io/cran/lmerTest/man/lmerTest-package.html, accessed on 7 January 2026) packages in R version 4.4.2. For each response variable—width, length, and area—models were fitted using restricted maximum likelihood (REML) with Treatment as a fixed effect, and a nested random-effects structure (1|Replicate/Micrography) was used to account for the hierarchical nature of the experimental design. Model singularity was evaluated, revealing a boundary (singular) fit warning for the Replicate random effect, which indicated negligible variance among biological replicates. The significance of the fixed effects was assessed using Type III ANOVA tests with Satterthwaite’s approximation, and multiple comparisons among treatment groups were conducted using Tukey’s HSD correction (α = 0.05) via the emmeans (https://cran.r-project.org/web/packages/emmeans/index.html, accessed on 7 January 2026) package (v1.10.0), with degrees of freedom estimated using the Kenward-Roger method. Data visualization and manipulation were executed using ggplot2 (v3.5.1, https://ggplot2.tidyverse.org/, accessed on 7 January 2026), dplyr (v1.1.4, https://dplyr.tidyverse.org/, accessed on 7 January 2026), and tidyr (v1.3.1, https://dplyr.tidyverse.org/, accessed on 7 January 2026) [50]. The complete, reproducible analysis code, along with the raw dataset, is provided in Script S4. Each treatment comprised three independent biological replicates, with two micrographs per replicate and measurements from 30 cells per micrograph, yielding a total of 476 observations (Control: n = 180, AGE: n = 119, Glyphosate: n = 177).

5. Conclusions

This study demonstrates that Agave lechuguilla extract (AGE) serves as a potent, multi-target pre-emergence bioherbicide against Chenopodium album. The extract significantly inhibited germination, disrupted energy metabolism (by reducing carbohydrates and elevating lipids), altered the antioxidant enzyme profile (suppressing catalase and ascorbate peroxidase while increasing glutathione reductase), and suppressed metabolic activity. Additionally, it induced severe morphological alterations in root and hypocotyl cells. Collectively, these effects suggest a mode of action distinct from that of the single-target inhibitor glyphosate, which primarily leads to metabolic shutdown and lipid accumulation. The coordinated physiological, biochemical, and morphological disruptions confirm the multi-target herbicidal potential of AGE, as previously indicated by transcriptional analysis, offering a promising strategy to mitigate selection pressure for herbicide resistance.

Future research should concentrate on isolating, identifying, and characterizing the specific bioactive compounds within AGE responsible for the observed phytotoxic effects. This chemical characterization is crucial for standardizing the extract and understanding its structure-activity relationship. Additionally, proteomic and metabolomic studies are necessary to validate enzymatic predictions and to map the complete metabolic network disruptions. To evaluate practical agronomic value, greenhouse and field trials are needed to assess the efficacy, optimal application methods, and environmental fate of AGE under real-world conditions. Furthermore, a thorough investigation into its selectivity and impacts on non-target organisms, including crops and soil microbiota, is essential to ensure environmental safety. Finally, long-term studies are vital for monitoring the potential development of weed resistance to this multi-target extract, which, although theoretically lower-risk, requires validation.