Bioactive Compounds with Pesticide Activities Derived from Aged Cultures of Green Microalgae

Simple Summary Overusing synthetic pesticides has resulted in environmental problems and human health risks while contributing to pesticide-resistant pests. Microalgae, being natural producers of highly diverse bioactive compounds, represent a promising alternative to replace synthetic pesticides as a more environmentally friendly and safe alternative. In this study, we screened biomass extracts from 10 green microalgae from the Chlorophyta phylum to identify molecules with potential pesticide activities. All tested strains exhibited herbicidal, nematocidal, or algicidal activities. Metabolomics analysis of six strains showed the presence of fatty acids, isoquinoline alkaloids, and aldehydes, which may be responsible for the observed pesticide activities. Our study highlights the algicidal, herbicidal, and nematocidal activities of biomass extracts derived from green microalgae and the metabolites and metabolic pathways potentially involved in producing the bioactive molecules. Our study supports the potential of microalgae for developing more effective and eco-friendlier biopesticides for sustainable agriculture. Abstract The excessive use of synthetic pesticides has caused environmental problems and human health risks and increased the development of resistance in several organisms. Allelochemicals, secondary metabolites produced as part of the defense mechanisms in plants and microorganisms, are an attractive alternative to replace synthetic pesticides to remediate these problems. Microalgae are natural producers of a wide range of allelochemicals. Thus, they provide new opportunities to identify secondary metabolites with pesticide activities and an alternative approach to discover new modes of action and circumvent resistance. We screened 10 green microalgae strains belonging to the Chlorophyta phylum for their potential to inhibit the growth of photosynthetic and nonphotosynthetic organisms. Bioassays were established to assess microalgae extracts’ effectiveness in controlling the growth of Chlorella sorokiniana, Arabidopsis thaliana, Amaranthus palmeri, and the model nematode Caenorhabditis elegans. All tested strains exhibited herbicidal, nematocidal, or algicidal activities. Importantly, methanol extracts of a Chlamydomonas strain effectively controlled the germination and growth of a glyphosate-resistant A. palmeri biotype. Likewise, some microalgae extracts effectively killed C. elegans L1 larvae. Comprehensive metabolic profiling using LC-MS of extracts with pesticide activities showed that the metabolite composition of Chlamydomonas, Chlorella, and Chloroidium extracts is diverse. Molecules such as fatty acids, isoquinoline alkaloids, aldehydes, and cinnamic acids were more abundant, suggesting their participation in the pesticide activities.


Introduction
Synthetic pesticides in agriculture have reached record levels globally, with approximately 5.2 billion pounds being utilized annually worldwide [1]. Synthetic pesticides have proven effective in controlling pests and enhancing food production. However, their extensive use has resulted in several detrimental consequences. Pests have developed resistance to various pesticide classes, thus, imposing significant challenges to building in vitro. Prior to in vitro germination, seeds were surface disinfected with 70% ethanol for 7 min, followed by 20% chloride for 7 min, and rinsed twice with sterile distilled water for 10 min. Seeds were then stratified at 4 • C overnight. Seeds were germinated in 1% strength Murashige and Skoog (MS) [30] growth medium agar plates (1.5% w/v, Bacto agar) under a 16:8 light/dark photoperiod with a light intensity of 50 µmol m −2 s −1 and a constant temperature of 21 ± 2 • C in a Conviron growth chamber model A1000 (Controlled Environments Inc., Pembina, ND, USA). Arabidopsis seed was increased under greenhouse conditions following standard practices. A. palmeri seeds were not increased in our laboratory.

Solvent Extraction of Lyophilized Microalgae Biomass
Prior to biomass harvesting, 30-day-old microalgae batch cultures were spotted onto Lysogeny Broth (LB) agar plates (1.5% w/v, Bacto agar) and incubated at 37 • C in a New Brunswick™ Innova incubator (Eppendorf, Hauppauge, NY, USA). After 24 h, agar plates were visually inspected for microbial contamination. Only cultures that exhibited no signs of contamination were considered for further studies. Microalgae cultures were harvested by centrifugation (4000 rpm, 15 min). The biomass was flash-frozen in liquid nitrogen and freeze-dried using a FreeZone 2.5 L, −84 • C Benchtop Freeze Dryer (Labconco, Kansas City, MO, USA) under 0.15 mbar vacuum pressure and a collector temperature of −80 • C for 48 h. From the resulting dried biomass, aliquots of 250 mg were transferred to glass vials and extracted separately with 8 mL of four different solvents (hexane, methanol, chloroform, and ethanol). The biomass with each solvent was sonicated for 30 min to favor tissue disruption and then filtered with Whatman No. 2 filter paper. The filtered extracts were dried out in the fume hood for 30 h. When completely dried, the samples were resuspended in 250 µL of the corresponding extraction solvent and used in the different assays.

Assessment of the Algicidal Effect of Microalgae Extracts
The agar diffusion test was established using strain 3016 as the target organism and TAP agar plates to test the algicidal effect of microalgae extracts. Briefly, different volumes (5, 10, and 15 µL) of the microalgae extracts were added to filter paper disks (Whatman No. 2) of 8 mm diameter, which were allowed to dry out in a fume hood. The paper discs containing the microalgae samples were placed immediately onto the agar plates (50 × 15 mm D, VWR International LLC, Radnor, PA, USA) freshly inoculated with the target microalgae. The strain 3016 (200 µL of OD 750 0.5 culture) was spread onto agar plates with sterile borosilicate beads. For the controls, 10 µL of each solvent and 5 µL of kanamycin (100 µg/mL) were applied instead of the algal extracts following the abovementioned procedures. Culture viability controls were set up. Agar plates were incubated, as mentioned above, for seven days. Independent experiments were repeated three times with three technical replicates each.

Assessment of the Nematocidal Effect of Microalgae Extracts
To test the capacity of algal extracts to inhibit C. elegans growth, 10-15 L1 larvae nematodes were inoculated in an S-basal medium [29] per replicate. C. elegans worms were treated with three volumes of each algal extract (2, 5, and 10 µL) for each solvent extraction. For the negative control, worms were treated with 2, 5, and 10 µL of the corresponding solvent. To analyze the survival rate of L1 larvae, dead and alive worms were visually inspected 24 and 48 h after treatment and counted after 72 h of treatment using a stereomicroscope (Olympus SZX10, Center Valley, PA, USA). Independent experiments were repeated two times with five technical replicates.

Assessment of the Herbicidal Effect of Microalgae Extracts
The bioassay using the different algal extracts was performed in 24-well agar plates, following the above standard procedures. Three different volumes (2,5, and 10 µL) of the microalgae extract were used for the assay. The corresponding volume was placed onto the agar medium, allowed to dry out, and then A. thaliana seeds were sowed. For the negative control, 2, 5, or 10 µL of the corresponding solvent was used instead, and then A. thaliana seeds were sowed. Positive controls were implemented using phosphinothricin (15 µg/mL) and kanamycin (50 µg/mL). Seed germination was verified in all experiments by sowing seed onto TAP agar medium. All plates were incubated, as mentioned above. Independent experiments were repeated two times with four technical replicates.

Metabolite Profiling of Microalgae Extracts
Mass spectrometry for untargeted metabolomics of biomass extracts of microalgae strains was performed in a Vanquish Flex Ultra High-Performance Liquid Chromatography (UHPLC) instrument with an Orbitrap Exploris 240 mass spectrometer (MS) (Thermo Fisher Scientific, Waltham, MA, USA). An XBride PREMIER Peptide BEH C18 Waters column (2.1 × 150 mm, 2.5 mm particle size) was used, and the system operated in positive ionization mode with an ESI voltage of 3500V. Biomass extracts were prepared and filtered using OnPu syringe filters PTFE 0.22 µm (Gorzizen, China). The analysis was performed using the AcquireX workflow for the automated generation of the background exclusion list and management of data-dependent acquisition (DDA) and dynamic exclusion (DE). The exclusion list was automatically generated from the matrix blank (methanol) and exported to the data acquisition methods for subsequent analyses. The automatic gain control (AGC) target value was set at 2 × 10 5 for the entire MS and 5 × 10 4 for the MS/MS spectral acquisition. The mass resolution was set to 120,000 for full scan MS and 60,000 for MS/MS events and a scan range (m/z) of 70-800. High-energy collision dissociation (HCD) was performed with a stepped collision energy of 30, 50, and 80%. The inclusion list was generated using a pool of each sample (20 mL of each sample corresponding to each strain) which also served as quality control. The solvents used were HPLC-grade water (solvent A) (W5-4, Thermo Fisher Scientific, Waltham, MA, USA) and HPLC-grade methanol (solvent B) (A452-4, Thermo Fisher Scientific, Waltham, MA, USA), both containing 0.1% formic acid (Sigma-Aldrich, St. Louis, MO, USA). The total LC run time was 25 min and consisted of a gradient with a 0.2 mL/min flow rate. The gradient started in the 5% mobile phase B for 4 min. It was gradually increased to 100% after 13 min. Later, it was kept constant at 100% phase B for 4 min. Phase B was reduced to 5% at min 21 and was maintained in the same condition until min 25 to re-equilibrate the column.

Data Analysis
The acquired data were analyzed using Compound Discoverer v3.3 (Thermo Fisher Scientific, Waltham, MA, USA), a commercial software package developed by Thermo Fisher Scientific to analyze data from their mass spectrometers. LC-MS RAW files were uploaded, and study factors in the form of categorical factors were defined. The untargeted workflow was used, adding the following databases integrated into the CD software: BioCyc, Carotenoids, LipidMAPS, Chemical Biology Department, Max Planck Institute of Molecular Physiology, Marine Drugs, Nature Chemical Biology, NIST, and Pesticide Common Names. This workflow performed retention time alignment, identified compounds using mzCloud (HighChem LLC, Bratislava, Slovakia) and ChemSpider (The Royal Society of Chemistry, Piccadilly, London, UK) databases, and performed ANOVA and Tuckey comparisons statistical analyses. The compound table was filtered by peak quality (peak rating > 5).
To perform a functional analysis of each metabolite and an over-representation analysis of the metabolites present in each sample, we upload their numeric mass (m/z), retention times (RT) and adj. p-values to MetaboAnalyst server (http://www.metaboanalyst.ca/, accessed on 27 June 2023) [31,32] using a mass tolerance of 5.0 ppm. Masses were searched in positive mode, and the enrichment analysis was performed using the Mummichog algorithm version 2.0 (http://mummichog.org/, accessed on 27 June 2023) [33] with a p-value cutoff of 1.0 × 10 −5 and selecting the A. thaliana library integrated into MetaboAnalyst. For the enrichment analysis, we selected the chemical structure main-class library [34]. Subsequently, for metabolic pathways analysis we used the C. vulgaris (green algae) pathway library from the Kyoto Encyclopedia of Genes and Genomes (KEGG) integrated into MetaboAnalyst. An UpSet plot was used to visualize the intersections between metabolites in the different strains using the R package UpSetR [35,36]. The identified metabolites among the different strains and the abundance of intermediates of the isoquinoline alkaloid biosynthesis pathway were compared via heatmap analysis using the R package ComplexHeatmap [37]. In the latter case, the analysis employed the group areas derived from the output table of CD, specifically focusing on the group areas associated with the identified compounds involved in the isoquinoline alkaloid biosynthesis pathway. Principal Component Analysis (PCA) of metabolite profiles was conducted using as input the normalized area values for each technical replicate and each compound with an assigned name from the Compound Discoverer output tables. PCA was conducted in R with the prcomp function from the stats library [36], and the corresponding coordinates for each PCA dimension were plotted using ggplot2 [35].

Aged Cultures of Green Microalgae Produce Metabolites with Algicidal Activity
To investigate the presence of allelochemicals with a potential pesticide effect in microalgae cultures, ten green microalgal strains, four of the genus Chlorella, one of the genus Chloroidium, and five of the genus Chlamydomonas, were evaluated. Upon 30 days of culture under controlled conditions (see Section 2), all strains were tested for contaminants and cell viability, and then the biomass was harvested by centrifugation ( Figure 1). Stationary cultures exhibited a color change to dark green pigmentation and no noticeable pH changes ( Figure 1a). Although freshly prepared medium was not resupplied to the culture throughout the experiment, cells of all strains were viable after 30 days of culture ( Figure S1). Only those cultures free of contaminants were harvested; the biomass was freeze-dried, and metabolites were extracted from the biomass using methanol, ethanol, hexane, and chloroform (Figures 1b and S1).
We first tested the activity of the biomass extracts in inhibiting the growth of a fastgrowing photosynthetic organism, i.e., other microalgae. To this end, we established a paper disk-diffusion assay using different volumes of biomass extracts to determine their capacity to inhibit the growth of strain 3016 as the target organism. This microalga strain was selected based on its remarkable growth rate, with a duplication time of 2.88 −1 under similar growth conditions to those employed in our study (temperature and nutrient concentration) [38]. The four extracts prepared from each strain were tested on the same plate along with kanamycin as a positive control, whereas solvents alone were tested on a separate plate. As expected, a sharp halo of inhibition was observed for the kanamycin (13 mm diameter), whereas no negative effects were observed for the treatment with solvents only (Figures 1c and S2). We found that the several microalgae biomass extracts generated using the four solvents compromised the growth of strain 3016 with different efficacies (Figures 1 and S3, Table S1). In some cases, we observed that only methanol and ethanol extracts from the same strain effectively inhibited microalgal growth, as observed in the halo of inhibition. In contrast, in other cases, chloroform extracts also presented an algicidal effect (Figures 2 and S3, Table S1). For instance, methanol and ethanol extracts of strains 124, 4764, and 247 presented algicidal activity. In the case of strain 1798, chloroform and methanol extracts presented algicidal activity, whereas only methanol extracts presented the same activity in strains 162 and 3054 (Figures 1 and S3, and Table S1). These data suggest that metabolites with different properties may underlie the growth-inhibiting activity over strain 3016. and methanol extracts presented algicidal activity, whereas only methanol extracts presented the same activity in strains 162 and 3054 (Figures 1 and S3, and Table S1). These data suggest that metabolites with different properties may underlie the growth-inhibiting activity over strain 3016.

Crude Extracts of Green Microalgae Control Seed Germination of An Aggressive Glyphosate-Resistant Weed
Arabidopsis is a small weedy plant that has been adopted as a model for many studies in plant biology. Thus, it is a good model to test the herbicidal activity of microalgal extracts. We then established a simple and effective method of germinating Arabidopsis seedlings in 24-well plates as an initial step to assess the potential herbicidal activity of the microalgae extracts. Different volumes (2, 5, and 10 µL) of each extract were applied to the agar medium, and Arabidopsis seeds were sowed. Wells with solvent and MS growth medium only were used as controls and evaluated after seven days (see section 2).
We identified extracts from the different microalgal strains which affected seedling growth or completely aborted seed germination (Figures 2 and S4, and Table S1). The inhibitory activity of the microalgae biomass extracts was variable based on the solvent used for extraction and the volume of crude extract utilized for the assay (Table S1). Remarkably, the biomass extract of some strains showed a noticeable effect on seedling growth even with the lowest volume used, 2 µL ( Figure 2). For instance, 5 µL of methanol extract from strain 124 affected seedling growth, whereas 5 µL of ethanol extract from the same

Crude Extracts of Green Microalgae Control Seed Germination of an Aggressive Glyphosate-Resistant Weed
Arabidopsis is a small weedy plant that has been adopted as a model for many studies in plant biology. Thus, it is a good model to test the herbicidal activity of microalgal extracts. We then established a simple and effective method of germinating Arabidopsis seedlings in 24-well plates as an initial step to assess the potential herbicidal activity of the microalgae extracts. Different volumes (2, 5, and 10 µL) of each extract were applied to the agar medium, and Arabidopsis seeds were sowed. Wells with solvent and MS growth medium only were used as controls and evaluated after seven days (see Section 2).
We identified extracts from the different microalgal strains which affected seedling growth or completely aborted seed germination (Figures 2 and S4, and Table S1). The inhibitory activity of the microalgae biomass extracts was variable based on the solvent used for extraction and the volume of crude extract utilized for the assay (Table S1). Remarkably, the biomass extract of some strains showed a noticeable effect on seedling growth even with the lowest volume used, 2 µL ( Figure 2). For instance, 5 µL of methanol extract from strain 124 affected seedling growth, whereas 5 µL of ethanol extract from the same strain aborted or inhibited seed germination (Figures 2 and S5). Interestingly, in the case of strains 1602 and 247, extracts prepared with the four different solvents showed an inhibitory effect of Arabidopsis seed germination ( Figure S4 and Table S1). Seeds sowed on MS and MS + solvent controls germinated and grew normally. Similar experiments with methanol extracts of strain 124 growing Arabidopsis seedlings vertically onto the agar showed that 5 µL were effective at impeding shoot and root growth (Figures 2a and S5). Suggesting that the type of metabolites and their concentration in this biomass extract are effective enough to control weeds.
Glyphosate-resistant A. palmeri biotypes are among the most aggressive weeds that infest cotton, peanut, and soybean fields in the U.S. and other regions of the world [39]. They show an aggressive growth habit and prolific seed production. Glyphosate application rates exceeding the average use rate of 0.84 kg ae ha −1 by three and twelve times allow control of glyphosate-resistant A. palmeri biotypes in the field by only 17 and 82%, respectively [40]. To evaluate the efficacy of microalgae extracts in controlling the growth of glyphosate-resistant A. palmeri, methanol biomass extract from strain 124 that was effective at controlling Arabidopsis shoot and root growth at low volumes was selected for in vitro testing. Phosphinothricin and kanamycin were used as positive controls. These two compounds are known to affect plant growth and are widely used as selectable markers. Phosphinothricin was more effective than kanamycin in controlling the growth of glyphosate-resistant A. palmeri (Figure 2b). Importantly, we found that 2 µL of the methanol extract compromised seed germination and seedling growth and was as effective as the phosphinothricin concentration used in this study. Still, about 10% of the tested seeds were able to germinate under both treatments. Complete seed germination abortion was observed with 5 µL of the methanol extract, which was more effective than phosphinothricin and kanamycin in controlling the growth of glyphosate-resistant A. palmeri (Figure 2b). The carrier solvent methanol (solvent control) did not negatively affect seed germination and seedling development. Well plates were maintained for over 20 days under the same conditions to investigate whether the seed could still germinate, possibly due to the degradation of the bioactive molecules in the microalgae extracts or the lack of effectiveness of the volume used. We observed that about 20% of the seeds treated with 2 µL of methanol extract of strain 124 were able to germinate and grow ( Figure S6). However, in the case of the treatment with 5 µL of biomass extract, only a few seeds showed radicle emergence, but none of them thrived even after 20 days of treatment ( Figure S6). These findings highlight the potential of the biomass extract of strain 124 as an effective herbicide against A. palmeri and further suggest its potential as a natural alternative to conventional herbicidal agents.

Crude Extracts of Green Microalgae Kill C. elegans L1 Larvae Nematode
Weeds represent a reservoir of pests and diseases in the field, further complicating pest management. Pathogenic nematodes belonging to diverse genera and species represent a serious threat to a wide range of hosts. Some of the most damaging nematodes are rootknot (Meloidogyne spp.), cyst (Heterodera and Globodera spp.), and reniform (Rotylenchulus reniformis) [41]. In this context, plant-host resistance has resulted in an effective and eco-friendly control strategy for pathogenic nematodes. However, nematodes often form disease complexes with fungi like Fusarium, thus further complicating pest management. Therefore, the use of several pesticides is often an unavoidable practice. We assessed the potential of the microalgal extracts to disrupt the growth of the free-living nematode C. elegans because it provides a suitable model to investigate their potential to control or kill pathogenic nematodes that infest crop fields. We used 96-well plates for these assays, placing 12-16 L1 larvae per well for testing the extracts (see Section 2). The potential toxic effect of the solvents used as metabolite carriers (methanol, ethanol, hexane, and chloroform) was tested. Among these solvents, only chloroform exhibited toxic effects on the nematode, while the other solvents (methanol, ethanol, and hexane) showed no negative effects (Figure 3). We found that biomass extracts killed L1 nematodes 72 h after treatment. Dead nematodes were straight, not motile, and did not produce offspring after the treatment (Figures 3 and S7, Tables S1 and S2). Methanol and ethanol biomass extracts of Chlorella strains 162, 1230, 1798, 1602, and Chloroidium species 247 showed lethal effects on the nematode, as 100% died after 72 h, while hexane extracts of all microalgae strains had no toxic effects against C. elegans (Figures 3 and S7). Interestingly, when exposed to ethanol biomass extract of strain 4764, 100% of the larvae died, whereas 57% survived under methanol treatment ( Figure S7, Table S2). An inverse scenario was observed for methanol and ethanol extracts of strain 125 that caused the death of 100% and 38% of the larvae, respectively ( Figure S7, Table S2). In contrast, methanol and ethanol extracts of strains 124 and 3054 effectively killed 100% of C. elegans larvae ( Figure S7, Table S2). These results suggest the potential presence of molecules that act as nematicides that can be employed as a natural alternative to synthetic pesticides.
iology 2023, 12, x FOR PEER REVIEW 10 of 20 nematodes were straight, not motile, and did not produce offspring after the treatment (Figures 3 and S7, Tables S1 and S2). Methanol and ethanol biomass extracts of Chlorella strains 162, 1230, 1798, 1602, and Chloroidium species 247 showed lethal effects on the nematode, as 100% died after 72 h, while hexane extracts of all microalgae strains had no toxic effects against C. elegans (Figures 3 and S7). Interestingly, when exposed to ethanol biomass extract of strain 4764, 100% of the larvae died, whereas 57% survived under methanol treatment ( Figure S7, Table S2). An inverse scenario was observed for methanol and ethanol extracts of strain 125 that caused the death of 100% and 38% of the larvae, respectively ( Figure S7, Table S2). In contrast, methanol and ethanol extracts of strains 124 and 3054 effectively killed 100% of C. elegans larvae ( Figure S7, Table S2). These results suggest the potential presence of molecules that act as nematicides that can be employed as a natural alternative to synthetic pesticides.

Non-Targeted Metabolomics Analysis of Microalgae Crude Extracts Suggests Highly Diverse Chemical Classes Potentially Involved in Pesticide Activities
A nontargeted metabolomics analysis of methanolic extracts was performed to gain insights into the diversity of chemical classes of metabolites in the microalgae extracts and molecules potentially associated with the algicidal, herbicidal, and nematocidal effects observed in this study. Methanolic extracts of six strains, namely 3054, 124, 162, 1798

Non-Targeted Metabolomics Analysis of Microalgae Crude Extracts Suggests Highly Diverse Chemical Classes Potentially Involved in Pesticide Activities
A nontargeted metabolomics analysis of methanolic extracts was performed to gain insights into the diversity of chemical classes of metabolites in the microalgae extracts and molecules potentially associated with the algicidal, herbicidal, and nematocidal effects observed in this study. Methanolic extracts of six strains, namely 3054, 124, 162, 1798, 1602, and 1230, were analyzed by UHPLC-MS (see Section 2). The metabolome of each strain consisted of chromatographic peaks with signal-to-noise ratio > 3.0 and peak quality > 5.0 in the normalized chromatographic area. The overall composition of the metabolome was determined through spectral searches against mzCloud MS/MS and additional databases (see Section 2). The metabolomic dataset for strains 3054, 124, 162, 1798, 1602, and 1230 consisted of 3372, 4206, 3405, 3429, 3498, and 3986 compounds, respectively (Dataset S1). These findings highlight the comprehensive nature of our metabolomic profiling, providing insights into the rich chemical composition of the tested extracts and strains.
To obtain an overview of the similarities or differences between the pool of metabolites for the different microalgae extracts, we did a Principal Component Analysis (PCA) for each sample and each injection replicate. A plot of dimensions 1 and 2 (PC1 and PC2, Figure S8a) showed that samples from strains 1230 and 124 have very distinct metabolite profiles, while 1798, 1602, 162, and 3054 cluster together, suggesting they have similar profiles, with 1798 and 1602, and 162 and 3054 being almost indistinguishable from each other. However, when dimensions 3 and 4 were plotted ( Figure S8b), none of the samples from different strains clustered together, suggesting different metabolite profiles. Differences among metabolite profiles became evident when we plotted dimensions 1 and 2 of a PCA analysis of the 100 most abundant metabolites in each microalga extract ( Figure S8c). This indicates that the metabolites present in strains 1230 and 124 are distinct and that those of 1798, 1602, 162, and 3054 have similarities in general terms but that important differences exist between the identity of the most abundant metabolites in each extract.
To further investigate the chemical classes to which the metabolites profiled belong, all databases were analyzed using the MetaboAnalyst server (http://www.metaboanalyst.ca/, accessed on 27 June 2023) [31,32] and the Mummichog algorithm (http://mummichog.org/, v2, accessed on 27 June 2023) [33]. Because this analysis utilized a library of known molecules from A. thaliana, the pool of annotated metabolites for strains 3054, 124, 162, 1798, 1602, and 1230 was reduced to 26.8%, 27.4%, 33.1%, 26.1%, 26.4%, and 27.6%, respectively, of the total metabolites detected. This suggests that all those molecules not assigned a chemical class have not been detected in higher plants yet and may be unique to microalgae. We found several enriched metabolite categories in all the methanolic extracts analyzed, including benzamides, amino acids and peptides, fatty acids and conjugates, sterols, and isoprenoids (Figure 4a, Tables S3-S8). Interestingly, some metabolite categories were exclusive to certain strains. For instance, the indolyl carboxylic acids category was exclusive to 3054; the alcohols and polyols category was exclusive to strain 124, and the oligosaccharides category was exclusive to strain 1602; the indoles category was enriched in 124 and 1798, and pyridine carboxyaldehydes were enriched in 162, 124, and 1230 ( Figure 4a). An UpSet analysis was performed to determine which metabolites are common and unique to the six strains. A total of 577 metabolites were shared among all the strains, and 208, 137, 80, 54, and 44 metabolites were unique to strains 124, 1230, 1602, 1798, 162, and 3054, respectively (Figure 4b, Dataset S1). An UpSet analysis of the 100 more abundant metabolites in each methanol extract revealed that only 23 metabolites are shared between all strains, whereas 36, 23, 22, 21, 18, and 16 are exclusive to strains 124, 1230, 1602, 162, 1798, and 3054, respectively (Figure 4c, Dataset S1). The more abundant shared metabolites included (9Z)-9-Octadecenamide, Traumatin, Umbelliferone, Xestoaminol, 4-Linoleamide, and palmitamide.  The Kyoto Encyclopedia of Genes and Genomes (KEGG) and MetaboAnalyst were then used to integrate metabolic pathway enrichment and topology analyses. Metabolic pathways with impact values greater than 0.1 were analyzed, obtaining 39 pathways ( Figure 5, Dataset S1). Pathways enriched in all six analyzed strains were thirty-one out of thirtynine, whereas two metabolic pathways were enriched exclusively in strain 1602 and one in strain 1230 (Figure 5a). Betalain biosynthesis followed by linoleic acid were the pathways with the higher impact value in all strains. Additionally, the pathways with higher impacts (>0.5) in at least one strain are beta-Alanine, galactose, tyrosine, isoquinoline alkaloid biosynthesis, and arachidonic acid metabolism. Importantly, molecules related to some of these pathways protect plants against several stresses or possess bioactivity. For instance, betalains are secondary metabolites reported to protect plants against abiotic stresses such as low and high temperature, metal ions, light, and exposure to polyamines, spermidine, and putrescine [42,43] and play a crucial role in defense against pathogenic fungi [44]. Another interesting pathway represented in the metabolite profiles of the six strains is isoquinoline alkaloid biosynthesis. Isoquinoline alkaloids are a class of N-based heterocyclic compounds synthesized from L-tyrosine that have been described to possess a broad range of bioactivities such as antibacterial, antifungal, antiviral, antiparasitic, and insecticidal [45]. The isoquinoline alkaloids synthesis pathway is complex, and importantly, several of the intermediates were found in our metabolite profiles, including 3,4-Dihydroxy-L-phenylanaline, Tyramine, and 3-(4-hydroxyphenyl) pyruvate (Figure 5b). The LC-MS and CD analysis yielded valuable insights into the relative abundance of intermediates in the synthesis pathway of isoquinoline alkaloids. The normalized peak areas revealed that 3-(4-Hydroxyphenyl) pyruvate was the most abundant intermediate, followed by Tyramine, and then 3,4-Dihydroxy-L-phenylalanine. However, L-tyrosine was detected only in samples of strain 124 (Figure 5c). These data provide valuable insights into the chemical classes of metabolites in the biomass extracts and the related metabolic pathways that may contribute to the pesticide activity of Chlorella, Chlamydomonas, and Chloroidium strains. Furthermore, they represent valuable resources to explore the use of these strains for biocontrol strategies and molecule discovery.

Discussion
Microalgae are recognized as prolific producers of secondary metabolites encompassing a diverse range of compounds such as aldehydes, alkaloids, fatty acids, peptides, and carotenoids [46]. These metabolites are often accumulated within the microalgae cells or secreted into the surrounding environment during growth [11,15,16,47,48]. Some of these secondary metabolites have been described to possess growth-inhibiting properties. Previous screening studies primarily focused on the study of Cyanobacteria, particularly species belonging to Anabaena, Calothrix, Fischerella, Nostoc, Oscillatoria, and Syctonema, in the search for algicidal metabolites [49]. However, investigations on green microalgae species and their potential as sources of biopesticides remain largely untapped. In this study, we tested the biomass of green microalgae of the Chlamydomonas, Chlorella, and Chloroidium genera for bioactive compounds with algicidal, herbicidal, and nematocidal activities. Our results revealed that the tested microalgae extracts exhibited inhibitory effects against the three target organisms employed in this study, i.e., C. sorokiniana, A. thaliana, and C. elegans (Figures 1-3, Tables S1 and S2). These results suggest that the microalgae cells produce compounds toxic to these organisms, possibly allelochemicals, in response to the prolonged period of culture. Similar studies have identified algicidal activity in aged cultures of different C. reinhardtii and C. vulgaris strains, and the Volvocaceae family beyond 20 days after inoculation [24,26]. The authors attributed this growth-inhibiting activity to a combination of fatty acids that might act as allelochemicals produced in response to the state known as conditional senescence resulting from nutrient limitation [50,51]. Humby et al. (2013) [51] demonstrated that when C. reinhardtii is grown in TAP medium under mixotrophic conditions, similar conditions employed in our study, nutrient depletion occurs as early as seven days after inoculation. Phosphorus and nitrogen, both essential nutrients, were the first nutrients to be exhausted [51]. Although we did not determine nutrient levels in our harvested cultures, conditional senescence might cause stress-inducing secondary metabolite production in the 10 strains we employed.
Early studies with C. vulgaris reported that this microalga's aged cultures (2-3 weeks) possess algicidal activity. Although the responsible compound was not isolated and further characterized, it was named Chlorellin, and described as a mixture of fatty acids and hydrocarbons [24,52,53]. Similarly, 20-day-old cultures of C. reinhardtii excrete a fatty acid mixture that inhibits the growth of Haematococcus lacustris (formerly Haematococcus pluvialis) [25]. Untargeted metabolomics analysis using UHPLC-MS enabled metabolite profiling of the methanolic extracts of six of ten microalgae strains screened in our study. We identified a wide diversity of metabolites produced by each species. The metabolites detected belong to 45 different chemical classes, suggesting a highly diverse chemical profile. Among the more abundant metabolites across all strains, notable compounds include (9Z)-9-Octadecenamide, Traumatin, Umbelliferone, Xestoaminol, 4-Linoleamide, and palmitamide. Umbelliferone, also known as 7-hydroxycoumarin, belongs to the coumarin family and has been shown to possess antimicrobial properties, functioning particularly as an antibacterial [54] and antifungal agent [55], as well as exhibiting antioxidant properties [56]. However, it has not been tested as an inhibitor of growth or seed germination. The fact that another member of the coumarin family, imperatorin, isolated from the plant Esenbeckia yaxhoob, inhibits radicle growth in L. esculentum and seed germination in L. sativa and L. esculentum [57] suggests that umbelliferone, present in the methanol extract of strain 124 effective to control A. palmeri seed germination, exerts an herbicidal effect as well. Interestingly, Razavi (2009) [58] reported the cytotoxic effects of the methanolic extract derived from Z. absinthifolia fruits containing coumarins on L. sativa. This extract was found to significantly reduce seed germination, as well as impede shoot and root growth in lettuce [58]. These suggest that green microalgae are significant sources of coumarin molecules with important biotechnological applications.
The extract of the strain 162 contained three known bioactive compounds Z)-9octadecenamide, palmitamide, and cinnamic acid. (9Z)-9-octadecenamide, the amide derivative of oleic acid obtained from the marine sponge Dendrilla nigra, and palmita-mide, another fatty acid amide, have been reported to have antimicrobial properties [59]. Cinnamic acid, a natural aromatic carboxylic acid widely distributed in plants and microalgae [60,61], was present in higher amounts in extracts of strain 162 than in other strains. Numerous studies have reported various beneficial effects of this carboxylic acid, including antioxidant, antimicrobial, and anti-inflammatory properties [60]. Moreover, cinnamic acid has been shown to induce oxidative stress and cause root cell death in cucumber seedlings, as it enhances the production of reactive oxygen species (ROS) and promotes lipid peroxidation, destroying cell membranes [62]. This suggests that the pesticide activities we observed in a single extract might be due to the concerted action of multiple metabolites, acting as a cocktail of molecules against diverse organisms.
Metabolite profiling was combined with enriched metabolic pathways identification to gain insights into their functional significance. We identified 31 pathways with an enrichment ratio greater than 0.1 common to the six analyzed strains (Figure 5a), including betalain biosynthesis, linoleic acid metabolism, galactose metabolism, tyrosine metabolism, isoquinoline alkaloid biosynthesis, porphyrin and chlorophyll metabolism, and sphingolipid metabolism. However, strain-specific enrichment was observed, such as the citrate cycle (TCA cycle) and glyoxylate and dicarboxylate metabolism pathways in strain 1602, and the arachidonic acid metabolism pathway enriched in strain 1230. Additionally, specific pathways, including vitamin B6 metabolism, riboflavin metabolism, glycerophospholipid metabolism, and tryptophan metabolism, were found to be present only in two or three strains, indicating that the associated metabolites may confer distinct properties to these specific strains. The linoleic acid metabolism and isoquinoline alkaloid biosynthesis pathways are of particular interest, which were common to all analyzed strains. Prior studies have reported that a mixture of fatty acids, including linoleic and alpha-linolenic acids, can inhibit the growth of other microalgae or induce autoinhibition [11,24,25]. Additionally, herbicidal activity against Lemna minor (phylum Magnoliophyta) has been attributed to 2,5dimethyldodecanoic acid, a fatty acid produced by the cyanobacterium Lyngbya aesturii (Cyanobacteria) [63]. As part of the linoleic acid metabolism pathway, we detected fatty acids such as linoleic acid, alpha-linolenic acid, and gamma-linolenic acid in the metabolic profiles of all strains (Dataset S1), which suggests these fatty acids may contribute to the pesticidal activity detected. It has been reported that many microalgae species important to biofuel production due to their high fatty acid and neutral lipid content also release some fatty acids that may reach high enough levels to cause self-inhibition or toxicity to the strains of interest or other strains. Therefore, ultimately hurting biomass productivity. Such negative effects of fatty acids are attributed to the damage to the plasma membrane, which may cause leakage of cellular content and cellular lysis [64].
Another category of interesting metabolites is the isoquinoline alkaloids biosynthetic pathway, which was detected in all six analyzed strains ( Figure 5). Isoquinoline alkaloids have been reported to possess diverse activities such as antiviral, antifungal, and enzyme inhibitory activities [45,65]. In a study with Chelidonium majus, methanol extracts exhibited antimicrobial activity; the extract contained seven different isoquinoline alkaloids that were toxic to the Gram-positive pathogen S. aureus [66]. Additionally, the antifungal activity of twelve isoquinoline alkaloids was evaluated against eight plant pathogenic fungi, demonstrating different strengths of fungicide activities. However, to our knowledge, the effect of isoquinoline alkaloids against microalgae, nematodes, or plants is unknown. In our study, we detected intermediates of the isoquinoline alkaloids biosynthesis ( Figure 5), such as L-tyrosine (C00082), 3,4-Dihydroxyl-L-phenylalanine (C00355), tyramine (C00483), and 3-(4-Hydroxyphenyl) pyruvate (C01179), the latter being more abundant in all strains ( Figure 5). Three of these metabolites (C00355, C00483, and C01170, Figure 5) were identified in the remaining strains. Importantly, methanol extracts of 124, 3054, 125, 4764, 1798, 1230, and 1602 strains effectively killed C. elegans in our study. The fact that the enrichment pathway analysis revealed that the analyzed strains (124, 3054, 1230, 162, 1602, and 1798) produce metabolites related to the isoquinoline alkaloid biosynthesis pathway suggests that molecules related to this pathway could be responsible for the nematocidal effect. This is further supported by a filed patent that protects the use of isoquinoline alkaloids as part of a formulation with insecticidal, acaricidal, and nematocidal action [67]. Interestingly, a previous study demonstrated the nematocidal effect of 15-day-old cultures of the cyanobacterium Nostoc oryzae (formerly Anabaena oryzae) (Cyanobacteria) and the two green algae C. vulgaris and Tetradesmus obliquus (formerly Scenedesmus obliquus) against the nematode M. incognita in banana plants. However, the specific compound responsible for this effect remain uncharacterized [68].
To the best of our knowledge, there are no reports of using green microalgae as sources of growth-inhibiting substances for weed control. Only a few studies have been conducted to assess specifically cyanobacterial toxins as herbicides [48,69,70]. We demonstrated that methanolic extracts of strain 124 effectively control the germination and growth of a glyphosate-resistant A. palmeri, suggesting the presence of molecules with herbicidal activity. Glyphosate, an extensively utilized nonselective and broad-spectrum herbicide, is a phosphonic acid derived from the reaction between the methyl group of methylphosphonic acid and the amino group of glycine (N-[phosphonomethyl]glycine) [71]. This herbicide targets the enzyme 5-enolpyruvylshikimate-3-phosphate (EPSP) synthase [71], thus functioning as an inhibitor of amino acid synthesis. As a result, the levels of aromatic amino acids are depleted, leading to the disruption of biosynthetic metabolic pathways and the eventual demise of the plant [72]. In general, glyphosate resistance has been reported to be acquired by several weeds due to different mechanisms, classified as target-site resistance (single base pair alteration, multiple base pair alteration, and gene amplification) and nontarget site resistance (enhanced metabolism and translocation). When target-site resistance occurs, an alteration of the target enzyme EPSPS which confers glyphosate resistance occurs, or an increased enzyme level is produced by gene amplification that increases tolerance to this herbicide. On the other hand, nontarget site resistance refers to a condition where the target enzyme remains susceptible to glyphosate, but a specific mutation has conferred a mechanism that decreases the amount of the herbicide which reaches the EPSPS enzyme [73]. The fact that the methanolic extract of strain 124 controlled the germination and growth of this weed suggests the presence of molecules showing a different mode of action. Interestingly, we detected widely diverse secondary metabolites in the methanolic extracts analyzed, and only~26-33% were annotated following the conventional LS-MS libraries or enzymatic pathways, suggesting that microalgae produce novel yet-to-be-characterized metabolites. Therefore, these novel metabolites or pathways can potentially be the source of the inhibitors and should be studied in more detail in the search for biopesticide agents. An interesting alternative to biomass microalgae in search of metabolites with pesticidal activities is to test the supernatants as they contain all compounds released by the microalgae cells into the growing environment. After prolonged periods of culture, these allelopathic compounds tend to accumulate in the growth media, thus providing an additional source of bioactive compounds.

Conclusions
In conclusion, our study highlights the algicidal, herbicidal, and nematocidal activities of biomass extracts derived from microalgae of the Chlamydomonas, Chlorella, and Chloroidium genera. The observed toxic effects on various biological targets coupled with the characterization of the metabolite profiles, potential metabolites responsible for the pesticide activities, and their associated metabolic pathways provide valuable insights into the potential bioactive compounds identified and their underlying mechanisms. Further investigations into these extracts, their composition, and the metabolic pathways involved hold promise for developing more effective and eco-friendly biopesticides with novel bioactive molecules for agriculture. Moreover, future studies employing targeted metabolomic approaches, isolation, and characterization of the specific metabolites would further enhance our understanding of their bioactivity.  Figure S2. Solvents do not produce cytotoxic effects on Chlorella sorokiniana UTEX 3016 growth. Three different volumes, 5, 10, and 15 µL of each solvent [methanol (1), ethanol (4), hexane (3), and chloroform (2)], were tested. The lack of effect of solvents was also assessed. Kanamycin (km, 50 µg/mL) was used as a positive control. Pictures represent technical triplicates of the experiments. The control illustrates the growth of strain 3016 onto TAP agar Petri dishes without solvents. Figure S3. Biomass extracts of Chlamydomonas, Chlorella, and Chloroidium strains inhibit the growth of C. sorokinana UTEX 3016. Disk-diffusion assay was conducted to evaluate the impact of biomass solvent extracts [methanol (1), ethanol (2), hexane (3), and chloroform (4)] produced from stationary-phase cultures of Chlamydomonas (125 and 4533), Chlorella (1602, 1798, and 1230), and Chloroidium (247) strains (see Section 2). The strain C. sorokiniana UTEX 3016 was used as the target organism. Kanamycin (km, 50 µg/mL) was used as a positive control; TAP agar Petri dishes show strain growth with no solvents or microalgae extracts. Independent experiments were repeated three times with three technical replicates. Figure S4. Biomass extracts of Chlamydomonas, Chlorella, and Chloroidium strains present herbicidal activity. Assessment of biomass extracts of Chlorella (162, 1230, 178, 1602), Chloroidium (247), and Chlamydomonas (3054, 4764, 125, 4533) strains for Arabidopsis growth and seed germination inhibition. Control conditions using MS growth medium with no microalgae extracts and solvents only were used to test seed germination. Figure S5. Assessment of biomass extracts of strain C. reindhardtii CC124 for Arabidopsis growth and seed germination inhibition with two volumes of the extract (2 µL and 5 µL). Control conditions using MS growth medium with no microalgae extracts and solvents only were used to test seed germination. Figure S6. Methanol extract of C. reinhardtii CC124 inhibits the growth of A. palmeri. Top view of A. palmeri 20-day-old seedlings treated with methanol biomass extracts of strain 124. Rows represent technical replicates of representative plants. Control indicates growth medium without solvent or biomass extract. Phosphinotricin (ppt, 15 µg/L) and kanamycin (km, 50 µg/mL) were used as positive controls. Volume indicates the volume of biomass extract utilized. Figure S7. Biomass extracts of Chlamydomonas, Chlorella, and Chloroidium strains present nematocidal activity. Assessment of the toxic effect of biomass extracts from Chlorella (162, 1230, 1798, 1602), Chloroidium (247), and Chlamydomonas (3054, 4764, 125, 4533, 124) strains on C. elegans. Living nematodes are visible as curved white strings, and dead nematodes appear as straight strings after 72 h of treatment. Control conditions with solvents only and no microalgae extracts were used to evaluate their effect on nematode viability. Figure S8. Principal Component Analysis of the named compounds (a,b) and 100 more abundant metabolites (c) from our analysis of the metabolite profiles of strains 1230, 1602, 1798, 124, 3054, and 162 with Compound Discoverer. The normalized areas of named compounds were used for the analysis. Each dot represents a technical replicate. Table S1. Pesticide activities of microalgae biomass extracts tested against C. sorokiniana UTEX 3016 (Cs), A. thaliana (At), and C. elegans (Ce). Table S2. C. elegans L1 larvae survival (%) after 72 h of treatment with microalgae biomass extracts. Table S3. Major chemical classes of metabolites identified in methanol extract of strain C. sorokiniana UTEX 1230. Table S4. Major chemical classes of metabolites identified in methanol extract of strain C. sorokiniana UTEX 1602. Table S5. Major chemical classes of metabolites identified in methanol extract of strain Chlorella sp. 'anitrata' 1798. Table S6. Major chemical classes of metabolites identified in methanol extract of strain C. reinhardtii CC124. Table S7. Major chemical classes of metabolites identified in methanol extract of strain C. reinhardtii CC3054. Table S8. Major chemical classes of metabolites identified in methanol extract of strain Chlorella sp. UTEX EE162. Dataset S1. Metabolite profiles of methanol extracts of Chlorella, Chloroidium, and Chlamydomonas strains, UpSet analysis, and metabolic pathways.