The Essential Oil of Petroselinum crispum (Mill) Fuss Seeds from Peru: Phytotoxic Activity and In Silico Evaluation on the Target Enzyme of the Glyphosate Herbicide

Petroselinum crispum (Mill) Fuss is an aromatic plant belonging to the Apiaceae family and used in gastronomy as a spice. Several studies have been developed in leaves but studies are limited in seeds, especially the essential oils obtained from seeds. The aim of this study was to determine the phytochemical profile of the volatile compounds of this essential oil by gas-chromatography–mass spectrometry (GC-MS) in order to evaluate its phytotoxic activity on Lactuca sativa seeds and to carry out an in silico analysis on the target enzyme of the herbicide glyphosate 5-enolpyruvylshikimate 3-phosphate synthase (EPSP). The essential oil was obtained by steam distillation for two hours and then was injected into a GC-MS, the phytotoxic assay was carried out on Lactuca seeds and the in silico evaluation on the EPSP synthase focused on the volatile compounds similar to glyphosate, docking analysis, and molecular dynamics to establish the protein–ligand stability of the most active molecule. The chromatographic analysis revealed 47 compounds, predominated by three compounds with the most abundant percentage in the total content (1,3,8-ρ-menthatriene (22.59%); apiole (22.41%); and β-phellandrene (15.02%)). The phytotoxic activity demonstrated that the essential oil had a high activity at 5% against L. sativa seed germination, inhibition of root length, and hypocotyl length, which is comparable to 2% glyphosate. The molecular docking on EPSP synthase revealed that trans-p-menth-6-en-2,8-diol had a high affinity with the enzyme EPSP synthase and a better stability during the molecular dynamic. According to the results, the essential oil of P. crispum seeds presented a phytotoxic activity and might be useful as a bioherbicide agent against weeds.


Introduction
Around the world, herbicide-resistant weeds represent a serious problem in food security with a current presence of 500 cases; the USA and Australia are the leading countries with a high number of herbicide-resistance cases, having reported 160 and 90 cases, respectively, followed by Canada, China, and Brazil [1]. In particular, the weeds that grow on crops such as maize, wheat, soybean, rice, canola, spring barley, and cotton have shown a high herbicide resistance [2]. Weeds are a type of plant that can rapidly reproduce and limit crop production by competing for several elements such as water, light, soil nutrients, and space for deep root systems, etc., [3]. An increase in weeds during agricultural production can trigger yield losses in crops and can depend on the type of weeds, crops, climate, weed density, weed emergence time, and others [4]. Additionally, these weeds might release a type of chemical substance called an allelopathic substance, which could negatively affect crops [5]. Despite the reduction management of weeds being a focus during the employment of synthetic herbicide, serious damage to the environment is produced due to its negative effects which include the acceleration of soil erosion, herbicide-resistant weeds, and contamination in vegetables and/or foods [2].
Bioherbicides are considered a powerful tool to combat weeds without causing any damage to the environment because they originate from nature sources and are less toxic than synthetic herbicides such as glyphosate [6]. Regarding glyphosate ((N-(phosphonomethyl) glycine)), this synthetic herbicide is sold at a low cost, but lately there have been concerns about its carcinogenic effect in humans [7]. Glyphosate biochemically works as a competitive inhibitor of phosphoenolpyruvate in the shikimic acid pathway, leading to the inhibition of the enzyme enolpyruvylshikimate-3-phosphate synthase (EPSP), a key enzyme in the generation of aromatic amino acids for plant growth [8]. Within the active site of EPSP synthase, several key residues play essential roles in catalysis and substrate binding. One of these critical residues is a conserved aspartate residue (Asp) and lysine (Lys), which aids in substrate binding and stabilizes the transition state during the reaction [9]. These active site residues work cooperatively to ensure efficient catalysis and precise substrate specificity, making EPSP synthase an attractive target for the development of herbicides and antimicrobial agents [10].
Certain bioherbicides have as their main chemical agent essential oils (EOs), which are also considered allelopathic substances [11]. EOs are volatile chemicals obtained by a distillation process or non-conventional methods (supercritical fluid CO 2 extraction, microwaves, ultrasound, etc.) [12] and their composition is mainly based on monoterpenes and sesquiterpenes compounds, although other compounds including aliphatic and aromatic esters have also been reported in EOs. Currently, the use of EOs in sustainable agriculture is being widely investigated to produce potential bioherbicides, insecticides, antibacterial, antifungal, antiviral, nematicides, and other effects to avoid crop losses due to weeds [13]. In addition, an issue linked to organic farming is the use of chemical herbicides which open up the possibility of searching for phytocompounds such as EOs with a bioherbicidal potential.
Petroselinum crispum (Mill) Fuss is a perennial herb (Figure 1), native to the Mediterranean region and used in the food, cosmetic, and pharmaceutical industries for its aromatic properties; it is commonly known as parsley and belongs to the Apiacea family [14]. It has been introduced and cultivated in many parts of the world due to its medicinal effects such as: antiseptic, stimulant, emmenagogue, antispasmodic, carminative, and stomachic effects, and its action on the uterine fiber [15][16][17]. On the other hand, the EO of P. crispum presents a high content of apiol [18], myrcene, 1,3,8-p-menthatriene, myristicin, β-phellandrene, and other terpenoids [19]. However, these chemical profiles can vary depending on several factors, mainly the origin, the extraction method, and the part of the plant used to obtain the EO [20]. The aim of this investigation was to determine the volatile compounds of P. crispum EO by gas-chromatography-mass spectrometry (GC-MS), in order to evaluate their phytotoxic activity on Lactuca sativa seeds and carry out an in silico analysis on EPSP synthase-the target enzyme of the herbicide glyphosate.

Chemical Characterization of EO of P. crispum Seeds
The obtained EO of P. crispum presented a light-yellow color with an extraction yield of 0.106%, a relative density of 0.9100 ± 0.02 g/mL at 20 °C, and a refractive index of 1.4910 at 20 °C. The chemical profile identified by GC-MS is reported in Table 1. The EO revealed 47 compounds (Table 1), 4 of which had unknown structures. The GC-MS identified the presence of 1,3,8-ρ-menthatriene as the major component with 22.59%, followed by apiole (22.41%) and β-phellandrene (15.02%). The studies reported about the EO of P. crispum seeds vary according to the precedence, type of extraction, pre-treatments, and environment factors. Piras et al. compared the EOs from seeds obtained by supercritical extraction and hydrodistillation, during which apiol was the most abundant compound with 65.4% and 82.1%, respectively, followed by myristicin with 20.1% and 11.4%. Furthermore, the yield percentage was better in supercritical extraction than in hydrodistillation [21]. The EO from Morocco presented apiole (23.45%), α-pinene (18.98%), and β-pinene (15.6%) as the main volatile compounds [22]. The EO from Italy had as its main volatile component myristicin with 19.4%, followed by 1,3,8-p-menthatriene (17.7%), β-phellandrene (17.2%), and apiole (15.5%). In addition, this EO was composed of monoterpene hydrocarbons (52.9%) and its content of phenylpropanoids was 34.9%, similar to our findings [23]. In a study carried out in China regarding the EO obtained by microwave and hydrodistillation, the content of myristicin was 79.58% and 85.59%, respectively [24]. Other reports of an EO from Turkey revealed that 3-methoxy-γ-asarone was the major compound with 34.19%, followed by myristicin with 23.83%. The variability in composition observed in comparison with the other P. crispum EOs from seeds discussed above could be due to the collection season, edaphic and climate factors, the previous pretreatment during the obtention of the EO in this study, and whether the seeds were milled and placed immediately into the distillation apparatus. However, the difference in percentages between 1,3,8ρ-menthatriene and apiole in the EO from Peru is not significant; this investigation is the first report of this type, contradicting those findings in which myristicin is the main component.
Essential oil Figure 1. The EO of P. crispum seeds obtained by steam distillation.

Chemical Characterization of EO of P. crispum Seeds
The obtained EO of P. crispum presented a light-yellow color with an extraction yield of 0.106%, a relative density of 0.9100 ± 0.02 g/mL at 20 • C, and a refractive index of 1.4910 at 20 • C. The chemical profile identified by GC-MS is reported in Table 1. The EO revealed 47 compounds (Table 1), 4 of which had unknown structures. The GC-MS identified the presence of 1,3,8-ρ-menthatriene as the major component with 22.59%, followed by apiole (22.41%) and β-phellandrene (15.02%). The studies reported about the EO of P. crispum seeds vary according to the precedence, type of extraction, pre-treatments, and environment factors. Piras et al. [21] compared the EOs from seeds obtained by supercritical extraction and hydrodistillation, during which apiol was the most abundant compound with 65.4% and 82.1%, respectively, followed by myristicin with 20.1% and 11.4%. Furthermore, the yield percentage was better in supercritical extraction than in hydrodistillation [21]. The EO from Morocco presented apiole (23.45%), α-pinene (18.98%), and β-pinene (15.6%) as the main volatile compounds [22]. The EO from Italy had as its main volatile component myristicin with 19.4%, followed by 1,3,8-p-menthatriene (17.7%), β-phellandrene (17.2%), and apiole (15.5%). In addition, this EO was composed of monoterpene hydrocarbons (52.9%) and its content of phenylpropanoids was 34.9%, similar to our findings [23]. In a study carried out in China regarding the EO obtained by microwave and hydrodistillation, the content of myristicin was 79.58% and 85.59%, respectively [24]. Other reports of an EO from Turkey revealed that 3-methoxy-γ-asarone was the major compound with 34.19%, followed by myristicin with 23.83%. The variability in composition observed in comparison with the other P. crispum EOs from seeds discussed above could be due to the collection season, edaphic and climate factors, the previous pretreatment during the obtention of the EO in this study, and whether the seeds were milled and placed immediately into the distillation apparatus. However, the difference in percentages between 1,3,8-ρ-menthatriene and apiole in the EO from Peru is not significant; this investigation is the first report of this type, contradicting those findings in which myristicin is the main component. In Table 2, it can be noted that there was no significant difference in some of the parameters evaluated to determine the phytotoxic effect between P. crispum EO at 0.5% and 1% with glyphosate, in seed germination (p = 0.1587 and p = 0.9690, respectively) or root length (p = 0.0784 and p = 0.1796, respectively), but in hypocotyl length, there was a significant difference at 0.1%, 0.5%, and 1% (p = 0.0004, p = 0.0067, p = 0.0424, respectively). Figure 2A shows the normal germination of L. sativa seed treated with 0.1% DMSO and does not evidence any alteration in growth. In Figure 2B-D, treated with the P. crispum EO, a reduction in root length was observed depending on the concentration of the used EO, as well as the hypocotyl length reduction. Figure 2E shows the effect of glyphosate in the germination process with a reduction in root and hypocotyl length.

Oxidative Stress Markers in the Phytotoxic Activity of the EO of P. crispum on Lactuca sativa Seeds
In Figure 3, the markers of oxidative stress such as malondialdehyde (MDA) and superoxide dismutase (SOD) are observably increased in those groups treated with the EO and the positive control glyphosate. As shown in Figure 3, there was no significant difference between P. crispum at 1% (1.72 ± 0.07 μM g −1 DW) and glyphosate (1.72 ± 0.07 μM g −1 DW) in MDA. For SOD, P. crispum at 1% (0.98 ± 0.12 U mg −1 DW) and glyphosate (1.05 ± 0.1 U mg −1 DW) provide similar values. According to Tijana et al. [40], lipid peroxidation represents a parameter of oxidative damage to molecules and cellular structures, especially those that contains lipids; in this research, MDA was evaluated and those seeds exposed to the EO at different concentrations showed an increased MDA at 1% overall. In addition, SOD is considered to be one of the main enzymatic systems to inhibit the stress generated by free radicals in plants [41]. Figure 3 shows that L. sativa seeds exposed to the EO increased their levels of SOD in response to the damage generated by the EO, similar to glyphosate. Several terpenoids such as α-pinene and β-pinene have shown an increase in MDA, catalase, SOD, peroxidase, and ascorbate as an action mechanism in phytotoxicity [42]. Certain species of plants have demonstrated phytotoxic activity via increased lipid peroxidation (MDA) such as the EOs of Pogostemon benghalensis, Monarda didyma, and Artemisia scoparia [43]. The use of essential oils as promissory bioherbicides is generating much interest due to the toxicity in the environment produced by synthetic herbicides, which can cause serious damage to the ecosystem. In this study, glyphosate was used as a positive control, which is used as a synthetic herbicide that blocks the enzyme EPSP synthase and catalyzes the sixth step in the shikimic acid pathway, reducing aromatic amino acids such as phenylalanine, tyrosine, and tryptophan [25] and its phytotoxic activity is also presented with the inhibition of seed germination, root length, hypocotyl length, and the rate of root length/stem length depending on the concentration, while the negative control (0.1% DMSO) did not show any phytotoxic activity. Germination was inhibited at the highest concentration, being similar to the positive control (1% glyphosate). Currently, there are no reports about the phytotoxicity of P. crispum EO from seeds. However, during a study in which the plant organ used to obtain the EO was not revealed, P. crispum at 0.64 µg/mL reduced the germination at 37.1%, reduced the root length at 29.1%, and resulted in a total fresh weight reduction at 79.7% [26].
In addition, the phytotoxic activity of the EO might be linked to its content of oxygenated terpenoids. This conclusion is supported by several studies of EO with phytotoxic activity in which an oxygenated terpenoid represents the major component, such as eucalyptol in species of Eucalyptus [35], carvacrol in Thymus proximus [36], and menthol in Mentha × piperita [37]. However, in this study, oxygenated terpenoids are only represented by oxygenated monoterpenes (2.07%) and oxygenated sesquiterpenes (0.792%) at low percentages of the total content. Additionally, apiol at 80 mM has shown a reduction in growth of Lemna paucicostata seeds at 75% [38]. Subsequently, monoterpenes and sesquiterpenes extracted from EOs have shown phytotoxic effects, causing anatomical and physiological changes in plants; proposed mechanisms involve the accumulation of lipid globules in the cytoplasm, oxidative stress, a reduction in mitochondria, and an inhibition of DNA [39].

Oxidative Stress Markers in the Phytotoxic Activity of the EO of P. crispum on Lactuca sativa Seeds
In Figure 3, the markers of oxidative stress such as malondialdehyde (MDA) and superoxide dismutase (SOD) are observably increased in those groups treated with the EO and the positive control glyphosate. As shown in Figure 3, there was no significant difference between P. crispum at 1% (1.72 ± 0.07 µM g −1 DW) and glyphosate (1.72 ± 0.07 µM g −1 DW) in MDA. For SOD, P. crispum at 1% (0.98 ± 0.12 U mg −1 DW) and glyphosate (1.05 ± 0.1 U mg −1 DW) provide similar values. According to Ðordević et al. [40], lipid peroxidation represents a parameter of oxidative damage to molecules and cellular structures, especially those that contains lipids; in this research, MDA was evaluated and those seeds exposed to the EO at different concentrations showed an increased MDA at 1% overall. In addition, SOD is considered to be one of the main enzymatic systems to inhibit the stress generated by free radicals in plants [41]. Figure 3 shows that L. sativa seeds exposed to the EO increased their levels of SOD in response to the damage generated by the EO, similar to glyphosate. Several terpenoids such as α-pinene and β-pinene have shown an increase in MDA, catalase, SOD, peroxidase, and ascorbate as an action mechanism in phytotoxicity [42]. Certain species of plants have demonstrated phytotoxic activity via increased lipid peroxidation (MDA) such as the EOs of Pogostemon benghalensis, Monarda didyma, and Artemisia scoparia [43].  Results are expressed as the average of three determinations. p ≤ 0.05 is considered as significant (Tukey's test). Table 3 represents the molecular similarity data for a set of compounds. The table includes various molecular descriptors such as the A-LogP (predicted octanol-water partition coefficient), molecular weight, number of hydrogen bond acceptors (HBA), number of hydrogen bond donors (HBD), number of rotatable bonds, number of rings, number of aromatic rings, and minimum distance. Based on the given data, the top 11 compounds are marked as similar, while the remaining compounds are marked as dissimilar. The top 11 compounds have molecular weights ranging from 138.20 to 222.36, and A-LogP values ranging from 1.313 to 3.202. These values suggest that the compounds are relatively small, with a moderate lipophilicity. They also have a similar number of hydrogen bond acceptors and donors, with values ranging from 1 to 4 which indicate the ability to form good hydrogen bonds with the target site. The number of rotatable bonds is relatively low, A B Results are expressed as the average of three determinations. p ≤ 0.05 is considered as significant (Tukey's test). Table 3 represents the molecular similarity data for a set of compounds. The table includes various molecular descriptors such as the A-LogP (predicted octanol-water partition coefficient), molecular weight, number of hydrogen bond acceptors (HBA), number of hydrogen bond donors (HBD), number of rotatable bonds, number of rings, number of aromatic rings, and minimum distance. Based on the given data, the top 11 compounds are marked as similar, while the remaining compounds are marked as dissimilar. The top  3.202. These values suggest that the compounds are relatively small, with a moderate lipophilicity. They also have a similar number of hydrogen bond acceptors and donors, with values ranging from 1 to 4 which indicate the ability to form good hydrogen bonds with the target site. The number of rotatable bonds is relatively low, ranging from 1 to 3, indicating a semi-flexible structure. These compounds also have a relatively low number of rings, ranging from 1 to 3, with only a few having aromatic rings. In contrast, the dissimilar compounds have molecular weights ranging from 132.202 to 238.28 and A-LogP values ranging from 2.772 to 4.939. These values suggest that these compounds are larger and more hydrophobic than similar compounds. Furthermore, most of these compounds have zero or very low values for HBA and HBD, indicating a lack of polar functional groups. These compounds also have a higher number of rotatable bonds and rings, ranging from 0 to 7, indicating a more flexible and complex structure. The minimum distance values shown in the table represent the similarity between each compound and a reference compound, the co-crystalized ligand 2AAY. The top 11 compounds have minimum distance values ranging from 2.07 to 3.11, indicating a high degree of similarity with the reference compound. In contrast, the dissimilar compounds have minimum distance values ranging from 3.5249 to 3.8803, indicating a low degree of similarity. In summary, the data in the table suggest that the top 11 compounds are similar in terms of their molecular properties, while the remaining compounds are dissimilar. The top 11 compounds have a moderate molecular weight and lipophilicity, with a similar number of polar functional groups, and a relatively rigid and simple structure. In contrast, the dissimilar compounds are larger and more hydrophobic, with fewer polar functional groups, and a more flexible and complex structure (Figure 4).

Docking Studies
The most similar compounds were docked against the EPSP synthase target site to study their binding mode and the degree of affinity towards this target site. The cocrystalized ligand showed an energy binding of −6.55 kcal/mol to the EPSP synthase target site, indicating its binding mode. the co-crystalized ligand formed six hydrogen bonds with Arg27, Ser23, Lys22, Asp313, and Lys340, with distances of 1.86, 1.78, 5.82, 2.46, 1.72, and 1.84 Å ( Figure 5).

Docking Studies
The most similar compounds were docked against the EPSP synthase target site to study their binding mode and the degree of affinity towards this target site. The co-crystalized ligand showed an energy binding of −6.55 kcal/mol to the EPSP synthase target site, indicating its binding mode. the co-crystalized ligand formed six hydrogen bonds with Arg27, Ser23, Lys22, Asp313, and Lys340, with distances of 1.86, 1.78, 5.82, 2.46, 1.72, and 1.84 Å ( Figure 5). In addition, cis-ρ-mentha-2,8-dien-1-ol had an energy binding of −5.47 kcal/mol to the EPSP synthase target site, and it formed two pi-alkyl interactions with Tyr200 and Lys340. It also had interactions with Asp313 and Lys22 through two hydrogen bonds with distances of 1.77 and 2.12 Å, as shown in Figure 6. On the other hand, terpineol had an energy binding of −5.69 kcal/mol to the EPSP synthase target site. It interacted with Lys22 In addition, cis-ρ-mentha-2,8-dien-1-ol had an energy binding of −5.47 kcal/mol to the EPSP synthase target site, and it formed two pi-alkyl interactions with Tyr200 and Lys340. It also had interactions with Asp313 and Lys22 through two hydrogen bonds with distances of 1.77 and 2.12 Å, as shown in Figure 6. On the other hand, terpineol had an energy binding of −5.69 kcal/mol to the EPSP synthase target site. It interacted with Lys22 and Asp313 through two hydrogen bonds with distances of 1.80 and 1.67 Å, respectively. Moreover, it formed a pi-alkyl interaction with Tyr200 ( Figure 7). and Asp313 through two hydrogen bonds with distances of 1.80 and 1.67 Å, respectively. Moreover, it formed a pi-alkyl interaction with Tyr200 ( Figure 7).  The binding orientation of the cis-ρ-mentha-1(7),8-dien-2-ol had an affinity score of −5.66 kcal/mol against the EPSP synthase target site, and it formed three pi-alkyl interactions with Tyr200, Pro312, and Lys340. It also had interactions with Asp313 and Lys22 through two hydrogen bonds with distances of 1.83 and 1.93 Å, respectively, as shown in Figure 8. Meanwhile, the trans-p-mentha-8-thiol-3-one had an energy binding of −5.56 kcal/mol against the EPSP synthase target site. It formed two pi-alkyl interactions with Lys340 and Tyr200 and two hydrogen bonds with Lys22 and Lys340 with distances of 1.74 and 2.35 Å, respectively, as shown in Figure 9. and Asp313 through two hydrogen bonds with distances of 1.80 and 1.67 Å, respectively. Moreover, it formed a pi-alkyl interaction with Tyr200 ( Figure 7).  The binding orientation of the cis-ρ-mentha-1(7),8-dien-2-ol had an affinity score of −5.66 kcal/mol against the EPSP synthase target site, and it formed three pi-alkyl interactions with Tyr200, Pro312, and Lys340. It also had interactions with Asp313 and Lys22 through two hydrogen bonds with distances of 1.83 and 1.93 Å, respectively, as shown in Figure 8. Meanwhile, the trans-p-mentha-8-thiol-3-one had an energy binding of −5.56 kcal/mol against the EPSP synthase target site. It formed two pi-alkyl interactions with Lys340 and Tyr200 and two hydrogen bonds with Lys22 and Lys340 with distances of 1.74 and 2.35 Å, respectively, as shown in Figure 9. The binding orientation of the cis-ρ-mentha-1(7),8-dien-2-ol had an affinity score of −5.66 kcal/mol against the EPSP synthase target site, and it formed three pi-alkyl interactions with Tyr200, Pro312, and Lys340. It also had interactions with Asp313 and Lys22 through two hydrogen bonds with distances of 1.83 and 1.93 Å, respectively, as shown in Figure 8. Meanwhile, the trans-p-mentha-8-thiol-3-one had an energy binding of −5.56 kcal/mol against the EPSP synthase target site. It formed two pi-alkyl interactions with Lys340 and Tyr200 and two hydrogen bonds with Lys22 and Lys340 with distances of 1.74 and 2.35 Å, respectively, as shown in Figure 9. The binding orientation of the cis-ρ-mentha-1(7),8-dien-2-ol had an affinity score of −5.66 kcal/mol against the EPSP synthase target site, and it formed three pi-alkyl interactions with Tyr200, Pro312, and Lys340. It also had interactions with Asp313 and Lys22 through two hydrogen bonds with distances of 1.83 and 1.93 Å, respectively, as shown in Figure 8. Meanwhile, the trans-p-mentha-8-thiol-3-one had an energy binding of −5.56 kcal/mol against the EPSP synthase target site. It formed two pi-alkyl interactions with Lys340 and Tyr200 and two hydrogen bonds with Lys22 and Lys340 with distances of 1.74 and 2.35 Å, respectively, as shown in Figure 9.  Trans-p-menth-6-en-2,8-diol's binding mechanism demonstrated an energy binding of −5.54 kcal/mol against the target site of EPSP synthase. Asp313, Lys340, and Lys22 were also interacted with by trans-p-menth-6-en-2,8-diol through three hydrogen bonds that were 1.75, 4.86, and 2.02 Å apart. (Figure 10). Additionally, apiole's binding mode demonstrated an energy binding of −6.85 kcal/mol against the target site of EPSP synthase; it interacted with Asp313 and Tyr200 via three pi-alkyl, pi-pi, and pi-anion interactions; and it also interacted with Lys340 and Arg124 via two hydrogen bonds with distances of 2.43 and 2.12 Å (Figure 11).  Trans-p-menth-6-en-2,8-diol's binding mechanism demonstrated an energy binding of −5.54 kcal/mol against the target site of EPSP synthase. Asp313, Lys340, and Lys22 were also interacted with by trans-p-menth-6-en-2,8-diol through three hydrogen bonds that were 1.75, 4.86, and 2.02 Å apart. (Figure 10). Additionally, apiole's binding mode demonstrated an energy binding of −6.85 kcal/mol against the target site of EPSP synthase; it interacted with Asp313 and Tyr200 via three pi-alkyl, pi-pi, and pi-anion interactions; and it also interacted with Lys340 and Arg124 via two hydrogen bonds with distances of 2.43 and 2.12 Å (Figure 11). Table 4 shows the DS, RMSD, and interactions in kcal/mol of each tested metabolite against the EPSP synthase target site, from which it can be concluded that apiole was the best molecule with a high affinity to the protein target −6.85 kcal/mol, similar to glyphosate at −6.55 kcal/mol. However, the compound trans-p-menth-6-en-2,8-diol had a greater number of hydrogen bonds than other compounds. Therefore, the molecular dynamic was carried out with this compound. The other values of the volatile compounds are presented in Table S1 (Supplementary Materials).
of −5.54 kcal/mol against the target site of EPSP synthase. Asp313, Lys340, and Lys22 were also interacted with by trans-p-menth-6-en-2,8-diol through three hydrogen bonds that were 1.75, 4.86, and 2.02 Å apart. (Figure 10). Additionally, apiole's binding mode demonstrated an energy binding of −6.85 kcal/mol against the target site of EPSP synthase; it interacted with Asp313 and Tyr200 via three pi-alkyl, pi-pi, and pi-anion interactions; and it also interacted with Lys340 and Arg124 via two hydrogen bonds with distances of 2.43 and 2.12 Å (Figure 11).   Table 4 shows the DS, RMSD, and interactions in kcal/mol of each tested metabolite against the EPSP synthase target site, from which it can be concluded that apiole was the Trans-p-menth-6-en-2,8-diol's binding mechanism demonstrated an energy binding of −5.54 kcal/mol against the target site of EPSP synthase. Asp313, Lys340, and Lys22 were also interacted with by trans-p-menth-6-en-2,8-diol through three hydrogen bonds that were 1.75, 4.86, and 2.02 Å apart. (Figure 10). Additionally, apiole's binding mode demonstrated an energy binding of −6.85 kcal/mol against the target site of EPSP synthase; it interacted with Asp313 and Tyr200 via three pi-alkyl, pi-pi, and pi-anion interactions; and it also interacted with Lys340 and Arg124 via two hydrogen bonds with distances of 2.43 and 2.12 Å (Figure 11).   Table 4 shows the DS, RMSD, and interactions in kcal/mol of each tested metabolite against the EPSP synthase target site, from which it can be concluded that apiole was the  According to the molecular docking results, trans-p-menth-6-en-2,8-diol, an oxygenated monoterpene of P. crispum EO, presented a high affinity for the EPSP synthase. In one study, several ligands were evaluated against EPSP synthase, and those with a high affinity for the enzyme had hydrogen-bond interactions with residues Lys23, Arg404, Arg105, Asp50, Gly101, Arg131, and Thr102. Furthermore, some of those residues were important in the interactions between phosphoenolpyruvate and glyphosate in the phosphoenolpyruvate-binding site [44]. These reports confirm that trans-p-menth-6-en-2,8diol from P. crispum EO might be used as an herbicide in the future via its interactions with similar residues to those reported by Oliveira et al. [44] such as Asp313, Lys340, and Lys22.

Molecular Dynamic Simulation
The study utilized molecular dynamics simulations to assess the stability and conformational changes of a protein-ligand complex. The root mean square deviation (RMSD) was used to evaluate the dynamic movements and conformational variations in both the protein and ligand, revealing a consistently low RMSD by the simulation. This indicates a high level of stability within both the apo and ligand-bound states. In order to identify the regions of the protein that exhibit flexibility during the simulation, the RMSF was calculated for each residue. Interestingly, no significant differences were noticed in the flexibility of any residues upon binding of the ligand. In order to assess the compactness of the complex, the radius of gyration (Rg) was calculated. The Rg values remained relatively constant throughout the simulation, indicating that the system remained compact. However, a slight increase in Rg was observed for the complex compared to the starting period. In order to investigate the interaction between the protein-ligand complex and the solvent, the solvent accessible surface area (SASA) was calculated. It was found that the protein had a lower SASA value at the end of the simulation, indicating a reduction in surface area and suggesting a more stable structure. Lastly, the study investigated the hydrogen bonding between the protein and ligand. It was observed that the protein and ligand could create up to four hydrogen bonds, which likely helped to keep the complex stable ( Figure 12).

The Molecular Mechanics Poisson-Boltzmann Surface Area (MMPBSA)
The study used the MM/PBSA method to calculate the binding free energy between the protein-ligand complex for the last 20 ns of the MD production run at intervals of 100 ps. The MmPbSaStat.py script was employed to obtain the average binding free energy and its standard deviation/error from the output files obtained from g_mmpbsa. The results showed that the binding free energy between the ligand and protein was −28 KJ/mol, indicating a favorable interaction. The study also identified the contribution of individual protein residues towards binding free energy. The study discovered that the LYS22, PRO312, TRP337, and ARG386 residues of the protein had a larger contribution of more than −2 KJ/mol toward the interaction with the ligand. This was done by breaking down the overall binding free energy into the per residue contribution energy. These residues were identified as hotspot residues in the binding of the ligand to the protein. A figure

The Molecular Mechanics Poisson-Boltzmann Surface Area (MMPBSA)
The study used the MM/PBSA method to calculate the binding free energy between the protein-ligand complex for the last 20 ns of the MD production run at intervals of 100 ps. The MmPbSaStat.py script was employed to obtain the average binding free energy and its standard deviation/error from the output files obtained from g_mmpbsa. The results showed that the binding free energy between the ligand and protein was −28 KJ/mol, indicating a favorable interaction. The study also identified the contribution of individual protein residues towards binding free energy. The study discovered that the LYS22, PRO312, TRP337, and ARG386 residues of the protein had a larger contribution of more than −2 KJ/mol toward the interaction with the ligand. This was done by breaking down the overall binding free energy into the per residue contribution energy. These residues were identified as hotspot residues in the binding of the ligand to the protein. A figure illustrating these results is provided (Figure 13). In another investigation, certain residues of the enzyme had a positive contribution with the ligand, enhancing its interaction, which were Arg124, Asp313, Glu341, His385, Lys22, Gln171, Arg344, Arg386, and Gly196 [45].

Plant Material
A quantity of 10 kg of P. crispum aerial parts was collected from Ica, Peru, located at 458 masl (14°04′40.5″ S and 75°43′15.5″ W)) in October 2022 (Figure 1). The material plant was authenticated in the Herbarium of the Universidad Nacional Mayor de San Marcos (070-2022-USM-MHN). Seeds (1 kg) were selected, cleaned, and milled. Then, seeds were placed into a Clevenger apparatus to obtain the EO by steam distillation for 2 h. Finally, the EO was separated from the hydrolate by decantation, and anhydrous Na2S O4 was added to purify the EO. Finally, the EO was kept in a sealed amber vial at 4 °C.

Identification of the Volatile Compounds by Gas Chromatography-Mass Spectrometry (GC-MS)
The volatile components of EO of P. crispum seeds were identified using an Agilent 6890N gas chromatograph coupled to an Agilent 5973 Mass Selective Detector operating in electron-ionization mode at 70 eV with a 5% diphenyl and 95% dimethylpolysiloxane capillary column (DB-5 MS, 30 m × 0.25 mm × 0.25 μm). The EO was diluted in CH2Cl2 of HPLC grade (1: 100 v/v). The chromatographic conditions were set according to a previous analysis protocol [46]. The identification of volatile components was based on a comparison of relative retention indices (RIs), mass spectra data (NIST-5 library), and published

Plant Material
A quantity of 10 kg of P. crispum aerial parts was collected from Ica, Peru, located at 458 masl (14 • 04 40.5 S and 75 • 43 15.5 W)) in October 2022 (Figure 1). The material plant was authenticated in the Herbarium of the Universidad Nacional Mayor de San Marcos (070-2022-USM-MHN). Seeds (1 kg) were selected, cleaned, and milled. Then, seeds were placed into a Clevenger apparatus to obtain the EO by steam distillation for 2 h. Finally, the EO was separated from the hydrolate by decantation, and anhydrous Na 2 SO 4 was added to purify the EO. Finally, the EO was kept in a sealed amber vial at 4 • C.

Identification of the Volatile Compounds by Gas Chromatography-Mass Spectrometry (GC-MS)
The volatile components of EO of P. crispum seeds were identified using an Agilent 6890N gas chromatograph coupled to an Agilent 5973 Mass Selective Detector operating in electron-ionization mode at 70 eV with a 5% diphenyl and 95% dimethylpolysiloxane capillary column (DB-5 MS, 30 m × 0.25 mm × 0.25 µm). The EO was diluted in CH 2 Cl 2 of HPLC grade (1: 100 v/v). The chromatographic conditions were set according to a previous analysis protocol [46]. The identification of volatile components was based on a comparison of relative retention indices (RIs), mass spectra data (NIST-5 library), and published literature. Each retention index was calculated in comparison to a homologous series of n-alkanes C9-C25 (C9, 99% BHD purity and C10-C25, 99% Fluka purity).

Phytotoxicity Assay on Lactuca sativa Seeds
Germination and growth bioassays were performed in Petri dishes (90 mm in diameter), with Whatman filter paper number 42. Commercial seeds of Lactuca sativa (Anasac ® , Batch N • 393368-61, distributed by Hortus S.A., Ate, Lima, Peru) were treated with different solutions of the EO solubilized in 0.1% DMSO at 0.1; 0.5; and 1.0%, respectively. The Petri dishes were then incubated at 20 • C for 5 days under a photoperiod of 12 h light. Several parameters were evaluated, including the number of germinated seeds, the root length, and the hypocotyl length. A solution of 0.1% DMSO was used as a negative control, and a 1% glyphosate solution (Fuego ® , NeoAgro, Lima, Peru) was used as a positive control. Each Petri dish had twenty seeds, and five replicates were utilized [47].

Determination of Oxidative Stress Parameters in L. sativa
The values of malondialdehyde (MDA) were measured using the reagent thiobarbituric acid. A total of 50 µL of L. sativa shoots homogenized in phosphate-buffered saline reacted with 200 µL 0.25% thiobarbituric acid solution (solution containing 10% trichloroacetic acid); then, these samples were heated at 95 • C for 30 min using a water bath. Next, samples were cooled to be centrifuged at 10,000× g for 10 min. The MDA content was measured using absorbance at 532 nm, and 155 mM −1 cm −1 was taken as the extinction coefficient. The formula used is: where A is the absorbance at 532 nm.
Superoxide dismutase (SOD) was determined using a commercial kit (Sigma Chemical Co, Saint Louis, MA, USA) according to the manufacturer's instructions. The SOD reaction was read on a microplate reader at 450 nm. Each assay was repeated in triplicate [48].
The formula used were: Increase in absorbance (uninhibited) per min (∆ODs min −1 ) at 450 nm; inhibition of absorbance per min (∆ODo min −1 ) by the sample at 450 nm.

Method of Molecular Similarity
The molecular similarity calculation is a computational method that evaluates the similarity between two molecules based on their ligands' structural and physicochemical properties, such as LogP, molecular weight, and distances between descriptors. In this study, six ligands with anticancer activity against an angiotensin-receptor blocker were analyzed for molecular similarity with glyphosate as a reference using Discovery studio software. The molecular properties analyzed included the rotatable bonds, cyclic rings, aromatic rings, hydrogen bond donors (HBD), hydrogen bond acceptors (HBA), partition coefficient (A-LogP), molecular weight (M.wt), and molecular fractional polar surface area (MFPSA) (Figure 4).

Method of Docking Study
The AUTODOCK VINA v.1.2.0 software was used to test ten natural metabolites against the EPSPS target site. To generate the binding sites, the co-crystallized ligand within crystal protein (PDB code: 2AAY) obtained from the RCSB was used [49]. The targeted proteins were prepared by removing water molecules, performing preparation options, adding missing amino acids, correcting unfilled valence atoms, and minimizing the protein peptides energy by applying CHARMM force fields. The essential amino acids of the protein were selected and prepared for screening. The natural metabolites were prepared by drawing their 2D structures using Chem-Bio Draw Ultra17.0 and saving them in SDF file format. The saved files were then opened in AUTODOCK VINA v.1.2.0 software, and the ligands were protonated, and energy was minimized using an MMFF94 force field with 0.1 RMSD kcal/mol. The minimized structures were stored for molecular docking. Molecular docking was performed using docking algorithms, where the targeted pocket was held rigid, and the ligands were allowed to be flexible. Each molecule was allowed to produce twenty different interaction poses with the protein during the refinement. The docking scores (affinity interaction energy) of the best-fitted poses with the active site at the EPSP synthase target site were recorded, and the 3D orientation was generated using the Biovia Discovery Studio 2019 Visualizer [50].

Method of Molecular Dynamic Simulation and MMPBSA
Molecular dynamics (MD) simulation is a computational method that investigates the behavior of molecules and atoms by numerically solving the equations of motion of a system. The widely used Gromacs software was employed to perform the MD simulations. The protein-ligand complex was prepared using Chimera software, and topology and parameter files were generated using Gromacs. The system was solvated with water molecules and equilibrated with NVT ensembles (constant number of particles, volume, and temperature) and NPT ensembles (constant number of particles, pressure, and temperature), both of which were carried out at 300 K and 1 atm for 100 picoseconds (ps). The thermostat and barostat for the simulation that was run, respectively, were chosen to be the V-rescale and Parrinello-Rahman. Finally, the production run was carried out for 100 nanoseconds (ns) at 300 K with a 2-fs time step. The molecular mechanics Poisson-Boltzmann surface area (MMPBSA) method was utilized to calculate the binding free energy between the protein and ligand. The trajectory files were analyzed using Gromacs tools to determine the RMSD, RMSF, SASA, hydrogen bonds, total binding energy, and radius of gyration. These were calculated using Gromacs in 100 ns to understand the dynamic behavior of the protein-ligand complex and the effect of ligand binding on protein structure and stability. The obtained results provided insights into the stability and interactions of the protein-ligand complex, which can aid in drug discovery and design [51].

Conclusions
The EO from the seeds of P. crispum collected in Ica, Peru, was analyzed for the first time by GC-MS and found to contain a high content of terpene hydrocarbons and phenylpropenes. The most abundant volatile compound was 1,3,8-ρ-menthatriene (22.59%), followed by apiol (22.41%). This finding indicates the existence of chemotypes of P. crispum which may be useful to identify and assess the quality control of the EO of Peruvian P. crispum seeds. Moreover, the EO produced a phytotoxic effect on L. sativa seeds at 1%, similar to glyphosate. The inhibition of germination, root length, and hypocotyl length were achieved at 0.5% and 1% concentrations, and its involved mechanism could be inducing oxidative stress due to a high concentration of MDA and SOD found in L. sativa shoots. Regarding the in silico analysis, 11 volatile compounds were similar to glyphosate during the molecular similarity analysis. In addition, trans-p-menth-6-en-2,8-diol had a high affinity to the enzyme EPSP synthase with −5.54 kcal/mol; furthermore, it presented a stability with the enzyme residues, Lys22, Pro312, Trp337, and Arg386 being important residues in the energy contribution during the molecular dynamic. This investigation will