Bioherbicidal activity and putative mode of action of Dittrichia graveolens against Amaranthus palmeri and A. retroflexus Authors: Mesut Sırrı, Ahlam Khalofah, Yunus Başar, Mustafa Abdullah Yılmaz, Shahid Farooq Published: 2026 ABSTRACT AbstractBackgroundChallenges in controlling Amaranthus weeds and rising cases of herbicide resistance necessitate the development of alternative weed control strategies. Plant-derived extracts may serve as valuable sources of phytotoxic chemicals. Nevertheless, their species-specific effects on germination and early seedling development require investigation. This study evaluated the allelopathic activity of aqueous extracts of Dittrichia graveolens on Amaranthus palmeri and A. retroflexus.MethodsAqueous extracts of D. graveolens were evaluated against A. palmeri and A. retroflexus under controlled laboratory conditions using a concentration gradient of 0, 6.25, 12.5, 25, 50, and 100%. The final germination, root and shoot lengths, inhibition in germination and growth, and germination timing indices were analyzed. Species-specific dose–response models were used to estimate IC50 and IC90 values, and comparative statistical analyses were conducted to assess interspecific variations in sensitivity. LC–MS/MS profiling was used to determine the phytochemical makeup of the extract. Molecular docking was conducted to predict potential interactions between the identified compounds and selected germination-related enzymes.ResultsThe LC–MS/MS profiling indicated a phenolic-rich extract, with quinic acid as the major chemical component (23.472 mg g⁻¹ extract), followed by fumaric acid (0.577 mg g⁻¹), salicylic acid (0.254 mg g⁻¹), protocatechuic acid (0.238 mg g⁻¹), and piceid (0.176 mg g⁻¹). The aqueous extract induced a significant dose-dependent suppression of germination and seedling development in both weed species. Higher final germination values were recorded under control and 6.25% extract treatments, decreased to 72% at 12.5%, and further reduced to 44–46% at 25%. Nevertheless, extract doses of 50% and 100% completely inhibited germination (< 5%). Root development exhibited significant sensitivity. Root inhibition reached 87.6% in A. retroflexus, compared with 56.0% in A. palmeri at a 25% extract concentration. A comparative dose–response analysis indicated that A. retroflexus was more sensitive than A. palmeri, as shown by lower IC50 values for germination inhibition (16.4% compared to 20.8%), root inhibition (11.4% vs. 20.5%), and shoot inhibition (13.1% relative to 27.8%). Germination timing indices indicated that higher extract concentrations decreased germination speed, delayed mean germination time, and altered germination synchrony. Molecular docking predicted comparatively favorable binding energies between selected flavonoids and evaluated germination-related enzymes. Luteolin and apigenin showed the most favorable predicted binding energies for phytase, whereas amentoflavone showed favorable predicted binding energies for catechol oxidase and β-glucanase.ConclusionThe aqueous extract of D. graveolens exhibited concentration-dependent allelopathic effects on A. palmeri and A. retroflexus, with A. retroflexus showing greater sensitivity overall. The extract inhibited germination and seedling development and altered germination timing. Phytochemical profiling identified candidate phenolic and flavonoid compounds, while molecular docking generated testable hypotheses regarding their possible interactions with germination-related enzymes. Biochemical, greenhouse, and field validation is required before their mechanistic involvement or the practical bioherbicidal use of the extract can be established.Graphical abstract Keywords: Allelopathy, Dose–response analysis, Phenolic compounds, Phytotoxicity, Weed management, Seed germination inhibition Introduction Weeds are among the most important biological constraints on crop yield and quality in agricultural production systems. Weeds exert significant economic losses by competing with cultivated plants for water, light, nutrients, and space [1]. Amaranthus species are considered among the most problematic weeds worldwide [2] due to their rapid growth, broad ecological tolerance, and high seed production capacity [3–5]. Amaranthus palmeri S. Watson and A. retroflexus L. are invasive and aggressive weeds that cause significant yield losses in many agricultural systems [3, 5–8]. A. palmeri is reported to cause 11–94% yield losses in various crops, including maize [9], soybean [10, 11], sweet potato [12], sorghum [13] and cotton [14]. Similarly, A. retroflexus causes significant yield losses (47–63%) in various crops, including kidney beans [15] and snap beans [16]. Hence, effective management of these highly competitive species is necessary to improve crop productivity. Weed management in modern agriculture largely relies on herbicides [17]. However, their intensive and prolonged use has led to the evolution of herbicide resistance in many weed species [18–20]. A. palmeri populations have developed resistance to multiple herbicides with different modes of action, making management extremely challenging [21–27]. Similarly, A. retroflexus has evolved resistance to numerous herbicides with multiple modes of action [28–33]. The increasing reports of herbicide resistance in both weed species necessitate a shift from sole reliance on herbicides toward alternative weed management approaches. Plant-derived extracts with allelopathic and bioherbicidal properties have emerged as potential complementary tools for reducing reliance on conventional herbicides [34–36]. These extracts may potentially be incorporated into integrated weed management programs to diversify control practices and reduce selection pressure for herbicide resistance [37, 38]. Plant-derived natural compounds have attracted considerable attention as potential bioherbicides in recent years [37, 38]. However, it is important to emphasize that natural origin does not imply environmental or toxicological safety. Several bioactive natural compounds exhibit toxicity levels comparable to, or exceeding, those of synthetic pesticides (e.g., rotenone, nicotine, pyrethrins, azadirachtin). Thorough ecotoxicological evaluation is therefore essential prior to any field application [34]. Secondary metabolites (i.e., phenolic compounds, flavonoids, terpenoids, and alkaloids) synthesized by plants can exhibit allelopathic effects that inhibit the germination and growth of other plants [39, 40]. These natural compounds may provide additional options for weed management, although their efficacy, selectivity, environmental fate, and toxicological safety require careful evaluation [34–36]. These attributes support further evaluation of plant-derived compounds as potential bioherbicides that could complement conventional herbicides and diversify integrated weed management programs. Dittrichia graveolens (L.) Greuter is an aromatic plant in the Asteraceae family, widely distributed in the Mediterranean and Western Asian flora [41]. It is considered an aggressive invasive species due to its high seed production, rapid growth, and strong ecological plasticity [41, 42]. The invasion success is partially attributed to the production of bioactive secondary metabolites with strong allelopathic potential [43, 44]. Essential oils and extracts of this species are rich in phenolic compounds, flavonoids, and various terpenoids [45–47]. Furthermore, phytochemical studies have shown that D. graveolens contains biologically active metabolites, including monoterpenes, sesquiterpenes, and various phenolic compounds [48–50]. The reported presence of potentially allelopathic secondary metabolites makes D. graveolens a relevant candidate for evaluation as a source of phytotoxic compounds. Although these compounds have been associated with antioxidant, antimicrobial, and other biological activities, information on the allelopathic and potential bioherbicidal effects of D. graveolens against problematic weeds remains limited. In vitro germination and seedling-growth assays are commonly used to evaluate the phytotoxic effects of plant extracts on weeds. Molecular docking can complement these assays by predicting energetically favorable interactions between identified compounds and selected protein targets and by prioritizing candidate ligand–target combinations for subsequent investigation [51, 52]. However, docking results are hypothesis-generating and cannot independently demonstrate physical binding, enzyme inhibition, biological efficacy, or a specific mode of action. This study determined the bioherbicidal effects of D. graveolens extracts on seed germination and seedling growth of A. palmeri and A. retroflexus. It was hypothesized that aqueous extracts of D. graveolens would concentration-dependently inhibit seed germination and early seedling growth in A. palmeri and A. retroflexus and that the two species would differ in their sensitivity to the extract. The specific objectives were: (i) to quantify the effects of increasing extract concentrations on final germination, germination timing, and seedling growth; (ii) to characterize and quantify the principal phytochemical constituents of the extract using LC–MS/MS; (iii) to estimate species-specific IC₅₀ and IC₉₀ values and calculate sensitivity ratios for comparative assessment; and (iv) to use molecular docking to generate testable hypotheses regarding potential interactions between selected compounds and germination-related protein targets. Materials and methods Plant material and experimental conditions The D. graveolens plants were collected from the Kezer Campus of Siirt University (37.970536 °N, 41.845020 °E), Türkiye, during July 2025. The aboveground parts (leaves and stems) were harvested at the flowering stage, air-dried in the shade at room temperature (25 ± 2 °C), and subsequently ground into a fine powder for extraction. The plant material was taxonomically identified by Prof. Dr. Mehmet Fidan (Siirt University, Türkiye), and a voucher specimen was deposited in the Herbarium of Siirt University (Voucher No.: DG-2025-131). Seeds of A. palmeri and A. retroflexus were collected from naturally infested agricultural fields in Mardin province, Türkiye. The seeds were cleaned, air-dried, and stored in paper bags at room temperature (20–25 °C) until use. All experiments were conducted in the Plant Pathology Laboratory, Department of Plant Protection, Faculty of Agriculture, Siirt University, Türkiye, under controlled environmental conditions (25 ± 2 °C, 60–70% relative humidity, and a 12 h light/12 h dark photoperiod). A light intensity of ~ 150 µmol m⁻² s⁻¹ photosynthetically active radiation (PAR) was provided by cool-white, fluorescent lamps. Preparation of aqueous extracts and experimental design The aqueous extraction procedure was adapted from previously reported allelopathic bioassays using Dittrichia and Inula species [53, 54]. Water was selected as the extraction solvent to recover predominantly polar constituents and to approximate the aqueous release of plant metabolites. The aboveground parts of D. graveolens were air-dried at room temperature (25 ± 2 °C) in the shade until a constant weight was achieved. The dried plant material was then ground into a fine powder using an electric grinder and passed through a 0.5 mm sieve to ensure uniform particle size. For crude extract preparation, 100 g of powdered plant material was mixed with 1 L of distilled water (1:10 w/v) and homogenized for 5 min using a laboratory blender. The homogenate was subsequently placed on an orbital shaker at 120 rpm for 24 h at room temperature to facilitate extraction. The mixture was first filtered through double-layer muslin cloth, then through Whatman No. 1 filter paper to remove fine debris. The filtrate was further clarified by centrifugation at 4000 rpm (≈ 3000 × g) for 15 min, and the supernatant was collected as the stock (100%) extract. The obtained extract was transferred into sterilized amber glass bottles and stored at 4 °C until use (maximum storage duration: 7 days to avoid degradation). The resulting supernatant was designated the stock extract, corresponding to 100 g dry plant material L⁻¹ of extraction solvent. For treatment preparation, the undiluted stock was operationally defined as 100% and diluted with distilled water to obtain 50, 25, 12.5, and 6.25% of the stock concentration. Distilled water served as the 0% control. Thus, the nominal treatment concentrations corresponded to 0, 6.25, 12.5, 25, 50, and 100% of the original stock extract. The prepared extracts were evaluated for their effects on seed germination (%), root length (cm), and shoot length (cm) of both weed species. The experiments were arranged in a completely randomized design (CRD) with four replicates per treatment and repeated twice. Each replicate consisted of one Petri dish (experimental unit) containing 20 seeds of either A. palmeri or A. retroflexus. Treatments included six concentrations of aqueous extracts: 100% (crude extract), 50%, 25%, 12.5%, 6.25%, and 0% (distilled water as the control). A fixed volume of 5 mL extract solution was applied to each Petri dish lined with Whatman No. 1 filter paper. Germination assays Weed seeds were surface sterilized in 2% sodium hypochlorite (NaOCl) solution for 5 min, followed by three rinses with sterile distilled water (5 min each). For each treatment, 20 seeds were placed in sterile Petri dishes (90 × 15 mm) lined with double-layer autoclaved Whatman No. 1 filter paper. Each Petri dish (experimental unit) received 5 mL of the corresponding extract, while the control treatment received 5 mL of distilled water. The dishes were sealed with parafilm to minimize moisture loss and incubated in a growth chamber at 25 ± 2 °C, 60–70% relative humidity, and a 12 h light/12 h dark photoperiod [55]. Germination was recorded daily for 14 days, and a seed was considered germinated when the radicle reached at least 2 mm in length. This duration was selected in accordance with established germination testing protocols for Amaranthus species [56]. At the end of the incubation period, the root and shoot lengths of all germinated seedlings in each Petri dish were measured, and the corresponding mean value per Petri dish was used as the experimental-unit observation. Germination percentage was computed by using Eq. 1: where GT is the number of germinated seeds, and TT is the total number of seeds. Daily germination counts were used to evaluate temporal germination dynamics by calculating the germination speed index, mean germination time, and germination synchrony index. The germination speed index was determined using the formula Σ(Gi/ti), where Gi represents the number of seeds germinated on day i and ti denotes the respective day of observation. The mean germination time was computed as Σ(niti)/Σni, where ni represents the count of freshly germinated seeds on day i. Germination synchrony was determined using the formula ΣC(ni, 2)/C(Σni, 2), where C(ni, 2) = ni(ni − 1)/2; values approaching 1 indicate more synchrony in germination. The inhibition in germination, root, and shoot lengths was calculated by using Eq. 2: where C is the mean value of the control (germination, root length, or shoot length), and T is the corresponding value under extract treatment. Positive values indicate inhibition relative to control, whereas negative values indicate stimulation. The inhibitory effects of extracts were expressed separately for germination, root growth, and shoot growth. LC–MS/MS analysis LC–MS/MS analysis was performed to identify and quantify phenolic compounds present in the aqueous extracts. For sample preparation, 100 mg of each extract was weighed and dissolved in 5 mL of water–methanol (50:50, v/v) in a volumetric flask. The solution was vortexed and sonicated for 10 min to ensure complete dissolution. Subsequently, 1 mL of the solution was diluted to a final volume of 5 mL with the same solvent mixture. An aliquot of 1.5 mL was filtered through a 0.22 μm PTFE syringe filter and transferred into LC vials. Chromatographic separation was performed using a reverse-phase C18 column (e.g., 150 × 2.1 mm, 3 μm particle size) maintained at 30 °C. The mobile phase consisted of (A) water with 0.1% formic acid and (B) methanol with 0.1% formic acid, using a gradient elution program. The flow rate was set at 0.3 mL min⁻¹, and the injection volume was 10 µL [57, 58]. Mass spectrometric detection was performed using an electrospray ionization (ESI) source operating in negative- and/or positive-ion mode, depending on the target compounds. The capillary voltage, nebulizer gas, drying gas temperature, and collision energies were optimized for each analyte. Quantification was performed in multiple reaction monitoring (MRM) mode, using external calibration curves prepared with analytical standards. During analysis, samples were maintained at 15 °C in the autosampler to ensure stability. Data acquisition and processing were conducted using the instrument’s dedicated software. In Silico molecular docking analysis Molecular docking analysis was performed using ligands (acacetin, aconitic acid, amentoflavone, apigenin, chlorogenic acid, fumaric acid, gallic acid, luteolin, piceid, protocatechyl acid, protocatechyl aldehyde, quinic acid, salicylic acid) and reference herbicides (pendimethalin and propyzamide) downloaded in SDF format from https://pubchem.ncbi.nlm.nih.gov/. Ligands were selected directly from the phytochemical profile to ensure ecological relevance and direct linkage between chemical composition and in silico analysis. Ligand structures were calculated using Chem3D, with energy minimization prior to docking, and saved in the SYBYL mol2 format. Pendimethalin and propyzamide were used as reference compounds only to provide a comparative standard for binding affinity values. These herbicides predominantly inhibit microtubule assembly (HRAC Group 3), which is mechanistically separate from the germination-related enzyme targets assessed herein. This difference is recognized as a limitation of the study. The 3D structures of the target proteins associated with seed germination, catechol oxidase (PDB: 1BT1, resolution: 2.70 Å), β-glucanase (PDB: 1GBG, resolution: 1.80 Å), and phytase (PDB: 1QLG, resolution: 2.20 Å), were obtained from the Protein Data Bank. Protein preparation was performed using the Molegro Virtual Docker 6 program, which removed co-crystallized ligands, water molecules, and other heteroatoms. Molecular docking simulations were performed with PyRx using the AutoDock Vina algorithm for binding affinity calculations [59]. The docking search domain was defined by a grid box centered on the active site of each protein [60]. The grid box dimensions were adjusted to fully cover the binding pocket while allowing flexibility in ligand orientation. The grid spacing was set to 1.0 Å. For docking calculations, the best conformations in AutoDock Vina were selected based on the lowest binding energy values (kcal/mol). The mean square deviation (RMSD) between experimental and predicted ligand positions was calculated. An RMSD value below 2.0 Å was considered acceptable. Based on the obtained docking positions, Discovery Studio was used to determine the basic interactions of the most stable ligand-protein complexes [61]. Phytotoxicity index A composite phytotoxicity index was calculated for each experimental unit as the arithmetic mean of germination, root-growth, and shoot-growth inhibition. The three components were assigned equal weight. The index was used as a descriptive summary of the overall response and was not interpreted as a validated mechanistic or toxicity index. Statistical analysis All statistical analyses were performed using the R statistical environment [62]. The final germination percentage, root length, shoot length, and their respective inhibition values were analyzed to determine the impacts of extract concentration, weed species, and their interaction. The proportion of inhibition in germination and seedling growth characteristics was determined relative to the species-specific control. For continuous response variables, model residuals were evaluated for normality and homogeneity of variance using residual plots and appropriate diagnostic tests. A two-way analysis of variance was conducted to directly compare A. palmeri and A. retroflexus, using weed species, extract concentration, and their interaction as fixed variables. When a significant treatment effect was detected, estimated marginal means were compared using Tukey-adjusted pairwise comparisons at P < 0.05. Species were also compared within each extract concentration using multiplicity-adjusted contrasts. Species-specific concentration–response relationships were fitted using four-parameter log-logistic models implemented in the drc package [63]. Concentrations producing 50 and 90% inhibition relative to the species-specific control were estimated as IC₅₀ and IC₉₀, respectively. Approximate 95% confidence intervals were calculated using the delta method. Sensitivity ratios were calculated separately for each endpoint and inhibition level by using Eq. 3. Values greater than 1 indicate that A. retroflexus required a lower extract concentration than A. palmeri to achieve the specified inhibition level and was therefore more sensitive. Approximate 95% confidence intervals for sensitivity ratios were calculated on the logarithmic scale. Concentration–response curves and inhibition heatmaps were generated using the ggplot2 package. Retardation values (percentage reduction relative to the control) for germination, root length, and shoot length were presented as heat maps. Mean retardation values were used to illustrate the relative inhibitory effects of treatments across weed species and dose levels. All graphical visualizations, including heat maps and dose–response curves, were generated using the ‘ggplot2’ package [64]. Results Seed germination and seedling growth Seed germination was significantly influenced by extract dose in both weed species (Fig. 1). Rapid and uniform germination was observed under control conditions, and final germination reached 96–100% by day 4–6. A similar pattern was observed at the 6% extract dose (~ 96–100% seed germination) with a slight delay in germination. The delayed germination suggests slight phytotoxic effects at low concentrations. In contrast, increasing extract concentrations reduced both the germination rate and final germination percentage. Cumulative germination was reduced to 72% of the control at the 12% dose (~ 25–30% reduction), with a clear delay in germination. This inhibitory effect intensified at the 25% dose, where final germination declined to ~ 44–46% (~ 55–60% reduction relative to the control), and germination stabilized after day 6. Seed germination was entirely inhibited at higher extract concentrations (50% and 100%). Final germination remained < 5% (> 95% inhibition), with sporadic germination events occurring only at later time points (days 8–10). These findings demonstrate a near-complete suppression of germination at elevated extract doses. Seed germination (%) of Amaranthus palmeri and Amaranthus retroflexus under different extract concentrations (Control, 6%, 12%, 25%, 50%, and 100%). Values represent the mean ± standard error (SE) of four replicates The extract exhibited a clear dose-dependent inhibitory effect, reducing both the speed and capacity of germination. The consistent pattern across both weed species indicates strong phytotoxic potential, with higher extract concentrations effectively retarding seed germination. Temporal germination dynamics validated that the extract influenced not only the germination percentage but also the temporal pattern of germination. The germination speed index decreased significantly as extract concentration increased in both weed species. A. palmeri exhibited the greatest germination speed index under the control treatment, followed by a slight reduction at 6%, a considerable decrease at 12%, and a significant decline at 25%. The same trend was observed for A. retroflexus; however, the decrease was more pronounced at intermediate doses, particularly at 25% (Fig. 2). Germination timing metrics of Amaranthus palmeri and A. retroflexus under different concentrations of Dittrichia graveolens aqueous extract. Bars represent mean ± SE. Different letters indicate significant differences among treatments according to Tukey’s test at P < 0.05 The mean germination time increased with extract concentration, suggesting delayed germination due to phytotoxic stress. Mean germination time remained comparatively low from the control to 25% extract concentration for A. palmeri, but increased significantly at 50% and 100% concentrations. On the other hand, mean germination time significantly increased at 25% and 50% in A. retroflexus, whereas germination was almost retarded at 100%. The germination synchrony index varied with extract concentration. Control and low-dose treatments often exhibited slight synchronous germination, but intermediate and high extract concentrations lowered synchronization. These results indicated that D. graveolens extract reduced germination ability, impeded germination rate, delayed germination timing, and altered synchrony, with more pronounced temporal disruption in A. retroflexus at intermediate dosages (Fig. 2). The four-parameter log-logistic models indicated species-specific differences in inhibitory concentrations (Table 1; Fig. 3). For germination, the IC₅₀ was lower for A. retroflexus (16.4%; 95% CI: 14.5–18.4) than for A. palmeri (20.8%; 18.7–22.9), giving a sensitivity ratio of 1.27. By contrast, the IC₉₀ estimates were 56.5% (41.1–71.9) and 49.0% (40.1–57.9), respectively, and their sensitivity ratio was 0.87, indicating convergence and substantial uncertainty at the high-inhibition level. Table 1Species-specific inhibitory concentrations and sensitivity ratios estimated from four-parameter log-logistic modelsTraitLevelA. palmeri % (95% CI)A. retroflexus % (95% CI)Sensitivity ratio(95% CI)GerminationIC₅₀20.8 (18.7–22.9)16.4 (14.5–18.4)1.27 (1.10–1.46)IC₉₀49.0 (40.1–57.9)56.5 (41.1–71.9)0.87 (0.64–1.17)Root inhibitionIC₅₀20.5 (17.1–23.8)11.4 (10.0–12.8)1.80 (1.49–2.17)IC₉₀53.5 (39.1–68.0)30.7 (23.0–38.5)1.74 (1.24–2.44)Shoot inhibitionIC₅₀27.8 (22.1–33.4)13.1 (10.4–15.8)2.12 (1.63–2.76)IC₉₀51.4 (8.0–94.8)35.4 (20.4–50.5)1.45 (0.61–3.44)Values are model estimates with approximate 95% confidence intervals. The sensitivity ratio is IC for A. palmeri divided by IC for A. retroflexus; a ratio > 1 indicates greater sensitivity of A. retroflexus at that inhibition level. Ratio confidence intervals were calculated on the log scale. IC₅₀ and IC₉₀ denote concentrations associated with 50 and 90% inhibition, respectively Species-specific inhibitory concentrations and sensitivity ratios estimated from four-parameter log-logistic models Values are model estimates with approximate 95% confidence intervals. The sensitivity ratio is IC for A. palmeri divided by IC for A. retroflexus; a ratio > 1 indicates greater sensitivity of A. retroflexus at that inhibition level. Ratio confidence intervals were calculated on the log scale. IC₅₀ and IC₉₀ denote concentrations associated with 50 and 90% inhibition, respectively Dose-response relationship of extract concentrations and seed germination, root length, and shoot length of Amaranthus palmeri and A. retroflexus. Data points represent observed means ± SE fitted with a four-parameter log-logistic model. Calculated effective doses (IC50 and IC90) are provided in each panel to indicate the concentrations required for 50% and 90% inhibition of the respective growth parameter Species differences were more pronounced for seedling growth. The root IC₅₀ was 11.4% for A. retroflexus and 20.5% for A. palmeri (sensitivity ratio 1.80), while the shoot IC₅₀ values were 13.1 and 27.8%, respectively (ratio 2.12). IC₉₀ ratios likewise favored greater sensitivity of A. retroflexus for root and shoot inhibition, although the wide confidence interval for the A. palmeri shoot IC₉₀ indicates considerable uncertainty in that estimate (Table 1). The inhibition heatmap showed concentration-dependent increases in germination, root growth, and shoot growth inhibition in both species (Fig. 4). At 25%, root inhibition reached 87.6% in A. retroflexus and 56.0% in A. palmeri. At 6.25%, germination inhibition was 0% in A. palmeri and 17.0% in A. retroflexus. Root inhibition was generally greater than shoot inhibition at corresponding concentrations. At 100%, complete inhibition of all three endpoints was observed in A. retroflexus, whereas A. palmeri retained small amounts of germination and seedling growth. Heatmap illustrating the inhibitory effects of extract concentrations on the germination, root length, and shoot length of A. palmeri and A. retroflexus. Values represent the percentage of inhibition relative to the control, where 0 indicates no effect, and 100 indicates complete growth suppression The composite phytotoxicity index increased with extract concentration in both weed species, reflecting the combined concentration-dependent inhibition of germination, root growth, and shoot growth (Fig. 5). At 6.25, 12.5, and 25%, the index was higher in A. retroflexus than in A. palmeri, indicating greater combined inhibition in A. retroflexus at low and intermediate concentrations. The index approached its maximum at 50 and 100% in both species, showing near-maximal combined inhibition under the controlled laboratory conditions. Composite phytotoxicity index of Amaranthus palmeri and A. retroflexus under different concentrations of Dittrichia graveolens aqueous extract The composite phytotoxicity index approached its maximum in both species at 50% and 100% dose, indicating complete inhibition of germination and seedling development. The convergence of both species at elevated concentrations indicates that the extract was highly efficacious against both weeds when applied at high doses. The strong response of A. retroflexus at lower and intermediate doses indicates its greater susceptibility to the aqueous extract. Phytochemical profiling Quinic acid (23.472 mg g⁻¹ extract) was identified as the dominant compound in the aqueous extract, representing the major peak in the chromatogram (Fig. 6; Table 2). Fumaric acid (0.577 mg g⁻¹), salicylic acid (0.254 mg g⁻¹), protocatechuic acid (0.238 mg g⁻¹), and piceid (0.176 mg g⁻¹) were present at moderate concentrations among the remaining constituents. Contrastingly, gallic acid, chlorogenic acid, protocatechuic aldehyde, apigenin, amentoflavone, acacetin, and luteolin were detected at relatively low levels. Representative LC–MS/MS chromatogram of the aqueous extract of Dittrichia graveolens. Peaks correspond to detected phenolic and related secondary metabolites, identified by retention times and mass spectra Table 2LC–MS/MS identification and quantification of phenolic and related compounds in the aqueous extract of Dittrichia graveolensAnalyteRT (min)Content (mg g⁻¹ extract)SD (±)Calibration equationR²LOD (mg L⁻¹)LOQ (mg L⁻¹)Quinic acid1.8223.4720.412y = 14532x + 120.40.99920.0150.050Fumaric acid2.350.5770.021y = 13845x + 98.70.99900.0100.035Aconitic acid2.910.0130.002y = 12760x + 85.20.99870.0080.025Gallic acid3.480.1230.006y = 15230x + 110.60.99950.0120.040Protocatechuic acid4.760.2380.009y = 14980x + 105.30.99930.0110.038Chlorogenic acid6.920.0660.004y = 15820x + 130.50.99960.0130.045Protocatechuic aldehyde7.350.0290.002y = 14025x + 95.40.99890.0090.030Piceid9.180.1760.007y = 16240x + 142.10.99940.0140.048Salicylic acid10.640.2540.011y = 15060x + 118.70.99920.0120.042Luteolin13.720.0060.001y = 17120x + 150.20.99850.0070.022Apigenin15.380.0390.003y = 16850x + 145.80.99910.0100.033Amentoflavone19.840.0110.001y = 17540x + 160.40.99880.0080.026Acacetin21.270.0100.001y = 16980x + 148.60.99900.0090.030Compounds were identified based on retention time (RT) and mass spectral data and quantified using external calibration curves. Values are expressed as mg analyte per g extract (mean ± SD)RT Retention time, LOD Limit of detection, LOQ Limit of quantification LC–MS/MS identification and quantification of phenolic and related compounds in the aqueous extract of Dittrichia graveolens Compounds were identified based on retention time (RT) and mass spectral data and quantified using external calibration curves. Values are expressed as mg analyte per g extract (mean ± SD) RT Retention time, LOD Limit of detection, LOQ Limit of quantification The chromatographic profile was dominated by quinic acid and other polar organic and phenolic acids, whereas the detected flavonoids occurred at considerably lower concentrations (Fig. 6; Table 2). These results characterize the chemical composition of the aqueous extract but do not establish the individual contribution of any compound to the observed phytotoxic effects. Molecular docking Molecular docking generated a range of predicted AutoDock Vina scores for the identified compounds against catechol oxidase (1BT1), β-glucanase (1GBG), and phytase (1QLG) (Table 3). Flavonoids generally produced more negative predicted scores than phenolic and organic acids across the selected protein targets. Table 3Binding energies (kcal mol⁻¹) of phytochemicals identified in Dittrichia graveolens aqueous extract against selected germination-associated enzymesMolecules1BT11GBG1QLGAcacetin-7.1-7.7-9.4Aconitic acid-5.6-5.5-6.5Amentoflavone-9.3-9.1-8.2Apigenin-7.1-7.9-10.2Chlorogenic acid-7.0-7.9-9.6Fumaric acid-4.9-4.8-5.7Gallic acid-5.8-5.7-6.9Luteolin-7.3-7.6-10.3Piceid-7.1-8.2-10.0Protocatechuic acid-6.0-6.0-6.6Protocatechuic aldehyde-5.5-5.6-5.8Quinic-5.6-5.8-7.1Salicylic acid-5.4-5.9-6.4Pendimethalin -5.7 -6.4 -5.2 Propyzamide -6.1 -6.7 -7.0 1BT1 = catechol oxidase, 1GBG = β- glucanase, 1QLG = phytase. Values are predicted AutoDock Vina scores expressed in kcal mol⁻¹. More negative values indicate more favorable predicted docking scores under the applied computational conditions but do not confirm physical ligand–protein binding or enzyme inhibition. Pendimethalin and propyzamide were included solely as computational reference compounds and not as mechanistic equivalents Binding energies (kcal mol⁻¹) of phytochemicals identified in Dittrichia graveolens aqueous extract against selected germination-associated enzymes 1BT1 = catechol oxidase, 1GBG = β- glucanase, 1QLG = phytase. Values are predicted AutoDock Vina scores expressed in kcal mol⁻¹. More negative values indicate more favorable predicted docking scores under the applied computational conditions but do not confirm physical ligand–protein binding or enzyme inhibition. Pendimethalin and propyzamide were included solely as computational reference compounds and not as mechanistic equivalents Amentoflavone produced the most negative predicted scores for catechol oxidase and β-glucanase, with values of − 9.3 and − 9.1 kcal mol⁻¹, respectively. For phytase, luteolin, apigenin, and piceid produced the most negative predicted scores, with values of − 10.3, − 10.2, and − 10.0 kcal mol⁻¹, respectively. Organic acids such as fumaric acid, aconitic acid, and salicylic acid produced less negative scores, ranging from − 4.8 to − 6.5 kcal mol⁻¹, whereas the scores for quinic acid ranged from − 5.6 to − 7.1 kcal mol⁻¹. These values rank candidate ligand–target combinations under the applied computational conditions but do not establish physical binding, enzyme inhibition, or biological activity. Pendimethalin and propyzamide produced predicted scores ranging from − 5.2 to − 7.0 kcal mol⁻¹. These compounds were included solely as descriptive computational benchmarks. More negative scores for some extract-derived compounds do not demonstrate greater herbicidal efficacy or indicate a shared mode of action with reference herbicides. The predicted poses of the highest-ranked ligand for each protein target contained hydrogen bonds, hydrophobic contacts, and π-based interactions (Figs. 7, 8 and 9). These contact maps describe computationally predicted poses and should not be interpreted as evidence of stable ligand–enzyme complexes. Predicted two-dimensional (2D) interaction diagram of the compound exhibiting the highest binding affinity with catechol oxidase (1BT1). Hydrogen bonds, π–π interactions, and hydrophobic contacts contributing to ligand–protein stability are displayed Predicted two-dimensional (2D) interaction diagram of the compound exhibiting the highest binding affinity with β-glucanase (1GBG). Key intermolecular interactions (hydrogen bonding and hydrophobic interactions) are shown Predicted two-dimensional (2D) interaction diagram of the compound exhibiting the highest binding affinity with phytase (1QLG). Interaction types responsible for strong binding (including hydrogen bonds and π-based interactions) are highlighted Across the evaluated targets, several flavonoids produced the most negative predicted scores for phytase, whereas amentoflavone ranked highest for catechol oxidase and β-glucanase (Fig. 10). Mean predicted scores for the extract-derived compounds were more negative than those of the two reference compounds for some targets (Fig. 11). These comparisons are descriptive and hypothesis-generating and do not confirm enzyme inhibition or a mechanism of phytotoxic action. Predicted AutoDock Vina scores of selected compounds identified in the Dittrichia graveolens aqueous extract against enzyme targets. More negative values represent more favorable predicted docking scores under the applied computational conditions Mean predicted AutoDock Vina scores of extract-derived compounds and reference compounds against the evaluated protein targets. Bars represent descriptive means, and error bars indicate the SE across compounds within each group. The error bars do not represent uncertainty from replicated docking simulations Analysis of binding strength indicated that the compounds obtained from the extract exhibited varying projected affinities across the three enzymes associated with germination. The strongest binding affinities among the evaluated targets were observed for phytase, particularly for luteolin, apigenin, and piceid, with binding strengths of around 10 kcal mol⁻¹. Amentoflavone had the most robust anticipated interaction with catechol oxidase and β-glucanase, suggesting potential multi-target binding capability. Generally, flavonoid compounds exhibited superior predicted binding compared to simpler phenolic or organic acids (Fig. 10). The mean binding energies of the extract-derived compounds were equivalent to or greater than those of the reference herbicides across the evaluated enzyme targets. These differences were evident for phytase, as the extract compounds exhibited a more negative mean binding energy than the reference herbicides. Nonetheless, these findings should be regarded as preliminary computational evidence of potential ligand–enzyme interactions rather than definitive validation of a particular herbicidal mechanism of action (Fig. 11). Discussion Effects on germination and seedling development Seed germination bioassays demonstrated a strong dose-dependent inhibitory effect of D. graveolens extracts on both A. palmeri and A. retroflexus. Germination percentage declined exponentially with increasing extract concentration, with nearly complete inhibition (> 95%) at higher doses (50–100%). These findings indicate that allelochemicals interfere with early metabolic activation processes such as enzymatic reserve mobilization and hormonal regulation [65, 66]. Temporal germination indices strengthened bioassay results. The extract not only reduced germination but also altered the germination kinetics. The decreased germination speed index and prolonged mean germination time suggest that allelochemical exposure hindered the physiological shift from imbibition to radicle emergence. This delay is physiologically significant because rapid, synchronized germination is crucial for early seedling establishment and competitive advantage. The synchronization response indicates that the extract disturbed the consistency of germination. Under controlled, low-dose conditions, germination occurred quickly and uniformly, but higher extract concentrations diminished synchrony or led to sparse, delayed germination. This trend suggests that the extract exerted physiological stress on the initial germination processes; however, the current results do not elucidate the specific biochemical mechanism underlying this effect. Temporal patterns revealed that low concentrations (6%) caused only a slight delay in germination, whereas intermediate concentrations (12–25%) significantly reduced both the germination rate and final germination percentage. High concentrations (≥ 50%) almost completely inhibited germination. These results are consistent with previous studies demonstrating strong allelopathic effects of Inula species on various weed taxa [67–69]. Aqueous extracts of I. graveolens suppressed germination and growth in several weed species, including A. retroflexus, Portulaca oleracea, and Sorghum halepense [53]. Similarly, reduced seed germination was observed for A. retroflexus treated with I. viscosa [70]. Several other studies have confirmed that extracts from various Inula species significantly suppress germination and seedling development [44, 54, 67, 69]. The concentration–response analysis indicated that interspecific sensitivity depended on both the measured endpoint and the inhibition level. A. retroflexus was more sensitive at the germination IC₅₀ and at the root and shoot IC₅₀ levels, as the corresponding sensitivity-ratio confidence intervals did not include 1. At the IC₉₀ level, a clear interspecific difference was supported for root inhibition, whereas the confidence intervals for germination and shoot inhibition included 1. Therefore, the apparent differences in the latter endpoints should be interpreted cautiously. Although germination was identified as the most sensitive stage, increasing extract concentrations also markedly suppressed early seedling growth, particularly root and shoot elongation. Root growth was generally more strongly inhibited than shoot growth at corresponding concentrations, indicating that root elongation was a particularly sensitive endpoint in the present bioassay. This greater sensitivity may be associated with the direct exposure of emerging roots to the extract, the high metabolic activity of root meristems, or the susceptibility of cell division and elongation processes to phytotoxic compounds. These mechanisms were not directly evaluated in the present study [71, 72]. The observed reduction in root length reflects disruptions in fundamental physiological processes, including cell division and elongation, membrane integrity, and water and nutrient uptake [73]. In contrast, shoot growth was comparatively less affected, suggesting differential allocation of stress responses between plant organs. Such root-specific sensitivity is widely recognized as a key indicator of allelopathic activity and has been consistently reported in previous studies [74, 75]. Overall, the pronounced inhibition of root development highlights the critical role of allelochemicals in impairing early plant establishment and supports the use of root growth parameters as reliable indicators for evaluating phytotoxic and bioherbicidal effects. Phytochemical analysis The LC–MS/MS profile indicated that the extract was dominated by polar metabolites, particularly phenolic and organic acids, with quinic acid as the major component. The observed extract composition is consistent with the use of water as an extraction solvent, which preferentially recovers polar compounds [76]. Similar phytochemical profiles reported for other Inula and Dittrichia species are broadly consistent with the chemical classes detected in the present extract, although differences may arise from extraction methods, environmental conditions, plant developmental stage, and analytical procedures [46, 47, 77]. Phytochemical screening confirmed the presence of diverse secondary metabolite classes, including anthraquinones, flavonoids, glycosides, coumarins, tannins, sesquiterpenes, steroids, and terpenoids [45, 78]. Several of the detected compound classes have previously been associated with phytotoxic or allelopathic effects. Their occurrence in the extract therefore identifies plausible candidate constituents, although the present study did not determine the biological contribution of individual compounds or compound combinations [70]. The predominance of polar phenolics indicated that aqueous extraction effectively isolated compounds with well-documented allelopathic properties. Phenolic compounds are widely known to inhibit seed germination through multiple biochemical mechanisms [79], including disruption of membrane integrity, inhibition of key enzymatic activities, and induction of oxidative stress [80]. Phenolic acids such as protocatechuic, salicylic, and chlorogenic acids are well documented for their allelopathic and phytotoxic effects, including inhibition of germination and disruption of cellular processes [81, 82]. The extract contained a chemically diverse mixture dominated by quinic acid together with several organic acids, phenolic acids, and flavonoids. The observed phytotoxicity may therefore reflect the combined activity of multiple constituents rather than the action of a single compound. However, the present crude-extract bioassay cannot distinguish between additive, antagonistic, or synergistic interactions. Experiments using isolated compounds and defined mixtures are required to determine their individual and combined effects [83]. This supports the concept of synergistic allelopathy, in which combined effects exceed those of individual constituents [84]. The phytochemical profile provides a basis for selecting candidate constituents for further investigation. Nevertheless, the present results establish only the coexistence of the detected metabolites and phytotoxic activity in the crude extract; they do not demonstrate which compounds were biologically active or whether interactions among them were additive, antagonistic, or synergistic. Molecular insights from docking analysis Molecular docking provided hypothesis-generating information on the predicted interactions of selected extract-derived compounds with the evaluated germination-related protein targets. The calculated AutoDock Vina scores enabled relative ranking of ligand–target combinations under the applied computational conditions but did not demonstrate physical binding or enzyme inhibition. The selected compounds produced predicted AutoDock Vina scores ranging from − 7.0 to − 10.3 kcal mol⁻¹ for phytase, β-glucanase, and catechol oxidase. More negative values represented more favorable predicted docking scores within the applied computational framework but did not confirm the formation or stability of ligand–enzyme complexes. Among these, flavonoids such as luteolin and apigenin showed the highest binding affinity toward phytase, while Amentoflavone produced comparatively favorable predicted docking scores across multiple evaluated protein targets. These findings align with the structural properties of flavonoids, whose hydroxyl-rich and aromatic frameworks enable strong interactions with protein targets [85]. The predicted poses contained hydrogen bonds, π-based interactions, and hydrophobic contacts that may contribute to ligand accommodation within the selected binding regions. These computationally predicted contacts do not establish the stability of the complexes under biological conditions. These interactions are putatively relevant because phytase regulates phosphorus mobilization, β-glucanase facilitates cell wall degradation and radicle protrusion, and catechol oxidase is involved in phenolic oxidation and redox balance [86–88]. Experimental inhibition of these enzymes could potentially interfere with phosphorus mobilization, cell-wall modification, and redox-related processes during germination [89]. However, the present docking analysis did not establish that the extract-derived compounds inhibited these enzymes in treated seeds. Several extract-derived compounds produced more negative predicted docking scores than pendimethalin and propyzamide for the selected protein targets [81]. This comparison provides a descriptive computational benchmark only and does not indicate greater herbicidal efficacy. Moreover, the established action of pendimethalin and propyzamide on microtubule assembly differs from the protein targets evaluated here, and no common mechanism is implied. Nonetheless, they are only in silico predictions; validation requires specific biochemical enzyme inhibition experiments and in vivo studies. Molecular docking does not account for bioavailability, metabolic transformation, or in vivo chemical stability. Overall, the docking results identify candidate compounds and protein targets for subsequent biochemical investigation. They generate a testable hypothesis that multiple germination-related proteins may potentially be involved but do not confirm enzyme inhibition, biological efficacy, or a multi-target mode of action. Validation requires direct enzyme-inhibition assays, experiments with isolated compounds, and measurements of target-related responses in treated plants. Integration of in vitro and in silico findings The bioassay, phytochemical profiling, and molecular docking components provide complementary but distinct forms of information. The bioassays experimentally demonstrated concentration-dependent inhibition of germination and seedling growth by the crude aqueous extract. LC–MS/MS analysis identified candidate constituents present in the extract, whereas molecular docking ranked potential interactions between selected compounds and protein targets under computational conditions. The coexistence of phytotoxic activity, a chemically diverse metabolite profile, and favorable predicted docking scores provides a basis for developing mechanistic hypotheses. However, no formal relationship was established between the concentration of an individual compound, its predicted docking score, and the magnitude of phytotoxic inhibition. Therefore, the present evidence cannot assign the observed biological activity to a particular constituent or confirm a specific enzymatic mechanism. Future studies should combine bioassay-guided fractionation, testing of isolated compounds and defined mixtures, direct enzyme-inhibition assays, and greenhouse and field experiments. Such approaches would clarify compound-specific activity, possible interactions among constituents, target engagement, crop selectivity, environmental safety, and practical performance. Limitations Despite promising results, this study has some limitations that should be considered. Experiments were conducted under laboratory conditions, which may not fully reflect field complexity. Furthermore, natural origin does not guarantee safety; thorough ecotoxicological assessment will be required before any field application of D. graveolens-derived preparations. Molecular docking provides predictive insights only and does not account for bioavailability, metabolic transformation, or in vivo stability. The use of crude extracts limits attribution of effects to specific compounds. Future research should incorporate enzyme inhibition assays, transcriptomic analyses, and greenhouse and field trials. Conclusion The aqueous extract of Dittrichia graveolens exhibited significant concentration-dependent phytotoxic activity against Amaranthus palmeri and A. retroflexus under controlled laboratory conditions. Germination and initial seedling development decreased in a dose-dependent manner, with near-total suppression observed at elevated extract concentrations, especially ≥ 50%. LC–MS/MS profiling revealed 13 predominant compounds in the extract, with phenolic compounds and organic acids (notably quinic acid and fumaric acid) dominating. These metabolites represent candidate constituents for further investigation, although their individual contributions to the observed phytotoxicity were not determined. A. retroflexus generally exhibited lower IC₅₀ values than A. palmeri, particularly for root- and shoot-growth inhibition. However, interspecific differences at the IC₉₀ level were endpoint-dependent, and estimates with wide confidence intervals should be interpreted cautiously. The germination timing analysis revealed that the extract not only reduced ultimate germination but also slowed germination speed, delayed mean germination time, and altered germination synchrony. The composite phytotoxicity index validated a cohesive inhibitory response in germination and seedling development characteristics. Molecular docking identified candidate ligand–target combinations and generated testable hypotheses regarding possible interactions between selected compounds and germination-related protein targets. Nonetheless, docking data must be interpreted with care since they serve as hypothesis-generating evidence rather than definitive proof of enzyme inhibition or a particular mechanism of action. Overall, D. graveolens represents a potential source of phytotoxic natural compounds for further bioherbicide research. However, direct biochemical validation, testing of isolated compounds and defined mixtures, crop-selectivity and ecotoxicological assessments, and greenhouse and field experiments are required before practical use can be considered. Acknowledgements The authors are grateful to the anonymous reviewers and editor for their constructive feedback. Generative AI The authors used Grammarly to translate and improve the manuscript’s language. The authors carefully controlled the content. Authors’ contributions M.S.: investigation, methodology, validation, visualization, resourcer, software, supervision, writing – original draft, and writing-review and editing. A.K. conceptualization, data curation, formal analysis, funding acquisition, investigation, writing-original draft. M.A.Y.: Conceptualization, data curation, investigation, methodology, software, validation, visualization, writing-review and editing, Y.B: Conceptualization, investigation, methodology, software, validation, visualization, writing-original draft. S.F.: conceptualization, data curation, formal analysis, investigation, writing-original draft. Funding This study received no funding. Data availability Data will be available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Consent for publication All authors consent to publication. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Mesut Sırrı, Email: m.sirri@siirt.edu.tr. Shahid Farooq, Email: shahid@harran.edu.tr. References Associated Data Data Availability Statement Data will be available from the corresponding author on reasonable request.