The effects of arbuscular mycorrhizal fungi on pesticides behavior in constructed wetlands

Yingrun Chen, MSc

Disertační práce

Abstract:

1. Research significance This thesis addresses an important environmental problem, namely pesticide contamination in wetland related water systems and its sustainable treatment. Tebuconazole and S-metolachlor are two widely used pesticides in agricultural production. Both compounds can leave farmland through runoff, precipitation, and leaching. They may then enter surface water, groundwater, and the food chain (Mei et al., 2011; Stara et al., 2019; Yu et al., 2021). Their environmental risks have therefore received increasing attention. Tebuconazole has been reported as a moderately toxic compound. Previous studies showed that excessive exposure may affect fetal development, liver and kidney function, enzyme activity, and reproductive health. Its potential carcinogenicity has also been reported (Kamrin, 1997; Komárek et al., 2010; Andrioli et al., 2023; Dong, 2024; Othmène et al., 2021; Šiviková et al., 2013; Jia et al., 2022; Li et al., 2022; Zhang et al., 2020; Fuerst, 1987). S-metolachlor also shows toxic effects. High concentrations may induce liver toxicity, and this herbicide can interfere with chlorophyll synthesis in plants (Komárek et al., 2010; Nikoloff et al., 2013). Although several treatment methods have been developed for pesticide removal, such as adsorption, membrane separation, advanced oxidation, and hybrid processes, these methods often require high cost, high energy input, and complex operation. They may also produce byproducts and cause secondary pollution (Pandit et al., 2025; Wang et al., 2025; Yogarathinam et al., 2025). In this context, CWs are a promising and sustainable option for pesticide contaminated water treatment. The significance of this thesis also lies in its attempt to address important research gaps related to pesticide behavior in CWs and the role of AMF. CWs remove organic pollutants through the combined action of plants, substrates, and microorganisms, and their ability to remove pesticides has been reported in previous studies (Vymazal and Březinová, 2015; Gikas et al., 2018; Chen et al., 2017). Among the biological factors involved, AMF are of particular interest because they can establish symbiosis with host plants and influence rhizosphere conditions, root exudation, litter properties, enzyme activity, and key microbial groups (Hu et al., 2021a; Hu et al., 2021b; Jiang et al., 2024; Jia et al., 2022; Jiang et al., 2021; Marschner et al., 1997). However, existing studies have mainly focused on the overall removal of pesticides from water. Much less attention has been paid to their fate in different wetland compartments, including water, substrate, roots, and leaves. As a result, the migration, accumulation, and transformation of pesticides in CWs are still not fully understood. This gap is important for tebuconazole and S-metolachlor because these two pesticides differ in chemical properties and biological effects, but both show environmental persistence and mobility. In addition, although AMF may affect pollutant uptake and distribution, their effects on pesticide partitioning, metabolite formation, and compartment-specific behavior in CWs remain unclear. The influence of AMF on microbial community composition, functional genes, and metabolic pathways under fungicide and herbicide stress has also not been sufficiently studied. Therefore, this thesis is significant because it links pesticide behavior, AMF symbiosis, and microbial regulation in CWs. It helps fill the gap between removal performance and process level understanding. It also provides theoretical support for the development of more effective and sustainable strategies for pesticide-polluted water treatment. 2. Achievement of the thesis aims Based on the points stated above, three objectives were established for this thesis. (1) Quantify the distribution of both pesticides and their metabolites across key CW compartments (water, substrates, roots, and leaves) under AMF colonization; (2) Identify and charact

Abstract:

erize metabolite profiles to resolve major transformation and degradation pathways; (3) Characterize microbial community responses associated with AMF colonization and link them to pesticide removal by evaluating microbial diversity and key taxa (including bacteria, archaea, fungi, and viruses), together with functional genes and metabolic pathways related to carbon, nitrogen, and phosphorus cycling under pesticide stress. 3. Thesis by publication map The thesis was arranged in chronological order according to their publication dates. The overall structure and key results of the publications were as shown in Fig. 6.1. 4. Methodological overview of this thesis 4.1 Experimental setup, operation, and sampling All experiments included in this thesis were conducted at the Czech University of Life Sciences Prague under ambient conditions, with temperatures ranging from 5 to 25 °C. The systems were protected from rainfall throughout the experimental period. PVC-U columns were used as the experimental units. Each column had a diameter of 150 mm and a height of 550 mm. The substrate was arranged in three layers. The bottom layer consisted of gravel and extended from 0 to 150 mm. The middle layer was used as the treatment zone and extended from 150 to 350 mm. The top layer consisted of sand and extended from 350 to 500 mm. Before use, both sand and gravel were heat sterilized at 120 °C for 3 h to reduce background variation. Iris pseudacorus was used as the host plant in all studies. In March 2022, six seeds were planted in each column at a depth of 10 cm. In the AMF treatments, 350 g of Rhizophagus irregularis (BEG 140) inoculum was mixed with sand and added to the middle layer. In the non-AMF treatments, the middle layer received sand only. This basic design was applied throughout the thesis to keep the experimental conditions comparable among chapters. The CWs were operated in batch mode and fed with simulated agricultural runoff. In Chapters III and IV, four treatment groups were included, namely control, AMF only, pesticide only, and AMF plus pesticide. Each treatment included five replicates. In chapter V, six treatment groups were established to compare AMF effects under tebuconazole and S-metolachlor stress. Five replicates were also used for each group. A consistent sampling scheme was used across the thesis. Inflow samples were collected on day 0. Outflow samples were collected on days 1, 2, 4, 8, 16, and 32. Rhizosphere substrate samples for metagenomic sequencing were collected before and after the experiment. Root samples were also taken before and after the experiment to evaluate AMF colonization. In the pesticide studies, plant tissues and rhizosphere substrate were collected at the end of the experiment for the analysis of pesticide distribution and metabolite profiles. 4.2 Analytical methods and data analysis AMF colonization was assessed by microscopy after root staining. Three standard indices were used, including the frequency of mycorrhizal presence (F%), the intensity of colonization (M%), and arbuscular abundance (A%). Pesticides and their metabolites were analyzed by liquid chromatography-mass spectrometry (LC-MS) using a high-resolution Q-TOF system with electrospray ionization. Conventional water quality parameters were measured using the same analytical framework across the studies. TN, TOC, TC, and IC were determined with a TOC analyzer. Nitrate, nitrite, phosphate, and sulfate were measured by ion chromatography. pH and ORP were measured with a portable multi-parameter meter. NH4+-N was determined according to a standard method. Metagenomic libraries were prepared by Hangzhou Lianchuan Biotechnology Co., Ltd. The sequencing data were deposited in the Paired End database. The same metagenomic workflow was used across the studies for taxonomic annotation, diversity analysis, and functional profiling. KEGG was used to analyze metabolic pathways and functional genes. LEfSe was used to identify taxa that differed among treatments. In the microbiome study, alpha diversity, beta diversity, and Random Forest analysis were also performed to identify taxa associated with treatment separation. Because water volume changed during the experiment as a result of evaporation and plant transpiration, pesticide and nutrient removal were evaluated using mass-based calculations rather than concentration alone. Student’s t-tests were used to compare water quality variables and removal-related parameters. Differences were considered significant at p < 0.05. All statistical analyses and figures were prepared in R version 4.3.1. 5. Summary of key results 5.1 AMF colonization rate under tebuconazole and S-metolachlor stress in CWs AMF colonization was consistently detected in Iris pseudacorus roots in AMF-CWs under both tebuconazole and S-metolachlor stress. This confirms that AMF symbiosis remained stable during CW operation. Colonization was evaluated by the frequency of mycorrhizal presence (F%), the intensity of colonization (M%), and arbuscular abundance (A%). In chapter III (with or without tebuconazole stress), F% ranged from 56.31% to 81.47%, and M% ranged from 43.65% to 54.49%, which indicates robust colonization. In chapter IV (with or without S-metolachlor stress), F% increased slightly after operation, from 53.6% to 56.5% in AMF-CWs without S-metolachlor and from 64.7% to 67.9% in AMF-CWs with S-metolachlor. M% also increased slightly, from 42.5% to 43.3% in AMF-CWs without S-metolachlor and from 41.7% to 43.5% in AMF-CWs with S-metolachlor. These increases were small and not statistically significant (p > 0.05). Across both chapters, A% remained low, at 0.86-1.47% under tebuconazole and 1.6-2.1% under S-metolachlor. Overall, AMF colonization was not suppressed by either fungicide or herbicide exposure during the 32-day operation, although arbuscule formation appeared more sensitive than overall root colonization. 5.2 Pesticide removal, distribution, and metabolite transformation in CWs Both tebuconazole and S-metolachlor were removed rapidly in CWs, and most of the mass decrease occurred at the early stage of operation. Under tebuconazole stress, a pronounced reduction was observed within the first day in all systems. Removal efficiencies were high in both AMF- and AMF+ systems, ranging from 94.10-97.78% and 95.87-97.83%, respectively, with no significant difference between treatments (p = 0.885). This rapid decrease was mainly due to fast substrate adsorption at the beginning, while later removal was more closely related to degradation and transformation. Under S-metolachlor stress, mass also declined sharply by day 1 and reached the lowest level by day 32. Removal efficiencies were 90.2% in AMF- systems and 92.7% in AMF+ systems, again without a significant difference (p = 0.47). A short increase in S-metolachlor mass between days 2 and 4 was observed, which likely reflected the partial release of previously adsorbed pesticide from the substrate. These results indicate that AMF colonization did not markedly change total pesticide removal efficiency for either pesticide. Although total removal was similar, the internal distribution of pesticides differed among phases and between AMF treatments. Under tebuconazole stress, residues were mainly detected in substrates and plant tissues, with higher concentrations in roots than in leaves. Average concentrations in AMF- systems were 0.69 ± 0.31 mg kg-1 in roots, 0.217 ± 0.10 mg kg-1 in leaves, and 0.04 ± 0.01 mg kg-1 in substrates. Lower values were observed in AMF+ systems, at 0.51 ± 0.14 mg kg-1 in roots, 0.15 ± 0.07 mg kg-1 in leaves, and 0.03 ± 0.006 mg kg-1 in substrates. This suggests that AMF modified tebuconazole distribution within CW compartments. Under S-metolachlor stress, the substrate was the dominant sink at the end of the experiment, with 27.1-30.1 μg retained, followed by the liquid phase with 16.3-21.9 μg, roots with 0.4-0.5 μg, and leaves with 0.3-0.5 μg. Root-associated S-metolachlor mass was higher in AMF+ systems (0.37 μg) than in AMF- systems (0.22 μg), which suggests enhanced plant uptake under AMF colonization. Across both pesticides, substrate adsorption was the main removal pathway, especially during the early phase, while plant uptake and microbial degradation played secondary roles. Metabolite profiling showed clear phase-specific patterns for both pesticides and AMF-related shifts. For tebuconazole, four metabolites were reported in CWs for the first time: tebuconazole carboxy acid, tebuconazole hydroxy, tebuconazole lactone, and tebuconazole dechloro. All four were detected in the liquid phase in both AMF- and AMF+ systems. Carboxy acid appeared on day 1 and showed an increase, then a decrease pattern, with lower relative abundance in AMF+ than in AMF- in liquid phases. Hydroxy appeared on day 2, peaked on day 4. Lactone was mainly detected in the liquid phase and was absent in plant tissues and substrate, with higher relative abundance in AMF+. Dechloro declined over time and showed lower relative abundance in AMF+ in the liquid phase, which suggests that AMF may have inhibited dechlorination. In plants, dechloro was mainly detected in leaves and was absent in roots and substrate, while carboxy acid and hydroxy were much higher in roots than in leaves and substrate. Under AMF+ conditions, plant-associated dechloro and carboxy acid were relatively higher, whereas hydroxy was relatively lower. For S-metolachlor, metabolite diversity was higher than for tebuconazole. Seven known metabolites were detected in influent or effluent, including metolachlor carbinol, deschloro, desmethoxy, ESA, OA, hydroxy, and morpholinone, together with ten unknown metabolites (Cmp.1-Cmp.10). In the liquid phase, ESA was detectable from day 1 and increased steadily, but its level was lower in AMF+ systems, with significant differences on days 1, 2, and 4 (p < 0.05). OA first appeared on day 4 and increased over time, with no significant AMF effect (p > 0.05). Other known metabolites generally increased and were lower in AMF+ systems, with significant differences for hydroxy on day 1, deschloro on day 8, and desmethoxy on day 32 (p < 0.05). Among the unknown metabolites, Cmp. 1, Cmp. 2, Cmp. 4, and Cmp.6 were present in influent, decreased on day 1, stabilized, and then rose again on day 2. Cmp.10 increased throughout the experiment. Cmp. 3, Cmp. 5, Cmp. 7, Cmp. 8, and Cmp. 9 were absent in influent but emerged and accumulated during operation. Most unknown metabolites were lower in AMF+ systems, with significant differences for Cmp. 1, Cmp. 2, Cmp.4, Cmp. 9, and Cmp. 10 (p < 0.05). By day 32, 19 compounds, including the parent compound and metabolites, were detected in outflow. In the liquid phase, 17 metabolites, excluding Cmp. 7 and Cmp. 10, were lower in AMF+, and the most abundant metabolites, excluding the parent compound, were Cmp. 9, Cmp. 8, and metolachlor deschloro. In the substrate, five metabolites were detected, and all were higher in AMF+. A total of 10 metabolites were detected in roots. Among them, 8 metabolites were lower in AMF+, while Cmp. 7 and Cmp. 10 were not. OA, Cmp. 9, and ESA were the dominant metabolites. In leaves, 9 metabolites were detected. Of these, 8 were lower in AMF+, with the exception of Cmp. 10. The most abundant metabolites were Cmp. 7, Cmp. 9, and OA. Overall, AMF was associated with lower metabolite levels in water and plant tissues for S-metolachlor but higher levels in substrate. 5.3 Bacterial community responses to AMF colonization under pesticide stress Under tebuconazole stress, metagenomic analysis also showed that AMF+ systems were enriched in Xanthomonadales, Xanthomonadaceae, and Lysobacter. KEGG analysis showed no AMF-related differences at Level 1 after tebuconazole addition, but significant shifts were detected at Level 3, including Type I polyketide structures, suberine and wax biosynthesis, biosynthesis of vancomycin group antibiotics, biosynthesis of enediyne antibiotics, bacterial chemotaxis, nonribosomal peptide structures and biosynthesis of type II polyketide backbone. Several Level 4 genes, including codA, NAD dehydrogenase, deaD, SurE, and tesA, generally showed higher relative abundance in AMF+ systems. Under S-metolachlor stress, the number of bacterial taxa with significant differences increased from 24 to 35. AMF+ systems were characterized by enrichment of Acidobacteriota and Catenulisporales/Catenulispora, while Pseudanabaenales, Leptolyngbyaceae, and Leptolyngbya were more abundant in AMF- systems. Several bacterial taxa, including Myxococcota, Nannocystaceae, and Terriglobia, persisted across the experimental period only in AMF+ systems, which suggests greater community stability. KEGG pathway analysis showed that AMF shifted bacterial functions toward enhanced biosynthesis and metabolic capacity, including steroid biosynthesis, siderophore group nonribosomal peptides, and ansamycin biosynthesis. Key genes such as CcmB, nrdA, and glpX also showed higher proportions in AMF+ treatments. 5.4 Carbon related responses to AMF colonization under pesticide stress Carbon dynamics showed a clear response to AMF colonization. Under tebuconazole stress, TOC mass was consistently higher in AMF+ treatments than in AMF- treatments. This difference was significant in systems without fungicide addition (p = 0.003), but not significant when tebuconazole was present (p = 0.06), which suggests that tebuconazole weakened the AMF effect on carbon dynamics. TC also differed significantly between AMF- and AMF+ treatments under tebuconazole stress (p = 0.008). Under S-metolachlor stress, TOC mass was again higher in AMF+ systems, with significant differences observed both in the absence (p = 0.004) and presence (p = 0.02) of herbicide. These results indicate that AMF consistently increased organic carbon availability in CWs. At the whole microbial level, carbon cycling microorganisms included 16 of 33 phyla and 15 of 93 genera above the 1% abundance threshold. Major decomposers, Pseudomonadota and Bacteroidota, declined across treatments and showed the clearest decrease under tebuconazole stress. In contrast, Cyanobacteriota increased in most treatments, which suggests higher tolerance of photosynthetic carbon fixing taxa. Stress tolerant groups such as Chloroflexota, Actinomycetota, and the fungus Stachybotrys also increased. Rhizophagus increased in all systems after operation, with a larger increase under AMF + tebuconazole (312%) than under tebuconazole without AMF (127%). At the KEGG level, seven carbon-related KEGG Level 3 pathways declined overall, but AMF-colonized systems showed lower decline rates than non-colonized systems (p = 0.01). This AMF buffering effect was especially clear under tebuconazole stress (p = 0.001). At KEGG Level 4, seven carbon cycle genes also declined, but the mean decline was much lower with AMF (1.44%) than without AMF (7.96%). Together, these results suggest that AMF increased carbon availability and reduced the loss of carbon related microbial functions under pesticide stress. 5.5 Nitrogen related responses to AMF colonization under pesticide stress Nitrogen removal performance was consistently improved in AMF+ treatments under both pesticide stresses. Under tebuconazole stress, the masses of TN, NO3--N, and NH4+-N generally decreased during operation, with better performance in AMF+ systems. In chapter III, significant differences between AMF+ and AMF- treatments were observed for TN (p = 0.012 and 0.011) and NO3--N (p = 0.002 and 0.001). In chapter IV, TN and NO3--N masses were also lower in AMF+ treatments. Significant differences were detected between systems without AMF and without pesticide and systems with AMF and without pesticide for TN (p = 0.01) and NO3--N (p = 0.002). These results indicate that AMF colonization enhanced nitrogen removal across both fungicide and herbicide scenarios. Nitrogen cycling microorganisms included 15 of 33 phyla and 16 of 93 genera above the 1% abundance threshold. Overall trends were similar between AMF and non-AMF conditions (p > 0.05), but clear contrasts emerged under tebuconazole stress. Several nitrogen-related taxa decreased under tebuconazole without AMF but increased when AMF were present under tebuconazole stress, including Candidatus Bathyarchaeota, Nitrososphaerota, Acidobacteriota, Mucoromycota, Candidatus Bathyarchaeota, and Klosneuvirus. Rhizophagus increased significantly under AMF + tebuconazole (p = 0.02). Nitrogen-fixing genera such as Bradyrhizobium and Mesorhizobium increased across treatments, which suggests relative tolerance to pesticide stress. At the community level, under S-metolachlor stress, the number of key microbial species identified by random forest increased from 10 in non-AMF systems to 15 in AMF-colonized systems, and the number of nitrogen cycling taxa increased from 3 to 8. Bacterial community analysis also showed clear AMF effects. At the functional level, five nitrogen related KEGG Level 3 pathways declined after operation in all treatments, but the mean decline rate was much lower with AMF than without AMF (2.66% vs 48.66%, p = 0.0003). This contrast was strongest under tebuconazole stress, where the decline rate was far higher without AMF than with AMF (125.07% vs 4.02%, p = 0.001). At KEGG Level 4, four nitrogen cycle genes declined less under AMF, with a mean decline of 6.33%, than without AMF, with a mean decline of 9.42%. Under tebuconazole stress, two regulators showed smaller declines with AMF than without AMF: histidine kinase (2.08% vs 2.84%) and a two-component transcriptional regulator (7.07% vs 17.06%). These results indicate that AMF improved nitrogen removal and buffered nitrogen-related microbial and functional responses under pesticide stress. 5.6 Phosphorus related responses to AMF colonization under pesticide stress Phosphorus removal showed a different pattern from carbon and nitrogen. Under tebuconazole stress, no significant differences in PO43- removal were detected between AMF+ and AMF- treatments. Under S-metolachlor stress, PO43- concentrations declined rapidly within the first two days in all CWs, regardless of AMF presence. These results suggest that phosphorus removal was less sensitive to AMF colonization than carbon and nitrogen under the tested conditions. The rapid early decrease also suggests that physical or chemical processes, especially substrate adsorption, dominated phosphorus removal. Phosphorus cycling microorganisms included 21 of 33 phyla and 14 of 93 genera above the 1% abundance threshold. Several taxa, including Candidatus Thermoplasmatota, Acidobacteriota, and Microsporidia, declined in most treatments, but were less reduced under AMF + tebuconazole CWs. Two key phosphorus-linked fungal groups increased strongly under AMF + tebuconazole: Mucoromycota increased 1.91 times (p = 0.002), and Rhizophagus increased 6 times (p = 0.02). Bradyrhizobium also increased significantly under tebuconazole with AMF (p = 0.0005), whereas Mesorhizobium showed no change (p = 0.89). These results suggest that AMF had a buffering effect on some phosphorus-related taxa under fungicide stress. However, phosphorus related functions were less responsive than carbon and nitrogen related functions. At KEGG Level 3, three phosphorus related pathways showed no significant AMF effect in the overall comparison (p = 0.95) or under tebuconazole stress (p = 0.32). At KEGG Level 4, five phosphorus cycle genes increased after operation, but differences between AMF and non-AMF systems were not significant (p > 0.05). Overall, AMF had limited effects on phosphorus related functional pathways under the tested conditions. 5.7 Whole microbial community responses and overall ecological implications Across all treatments, the whole microbial community remained taxonomically complex but broadly stable at the higher abundance level. Using a >1% abundance threshold, archaea included 7 phyla and 10 genera, bacteria included 12 phyla and 35 genera, fungi included 10 phyla and 24 genera, and viruses included 4 phyla and 24 genera. The composition above this threshold was highly similar before and after the experiment, which indicates a stable background environment during operation. At the genus level, more than 50% of archaeal, bacterial, and fungal reads were assigned to Others. This finding highlights the diversity and complexity of the microbial community. Alpha diversity increased in all systems, but the increase was consistently larger in AMF-colonized CWs than in non-AMF CWs. Chao1 increased by 2.22% in AMF-colonized systems and by 1.56% in non-AMF systems, while observed species increased by 2.83% and 2.48%, respectively. Significant increases in Chao1 were observed in AMF- colonized systems under herbicide stress, as well as in both AMF-colonized and non-AMF systems under fungicide stress. Observed species increased significantly in control systems, AMF-colonized herbicide systems, and both fungicide treatments. Under tebuconazole stress, AMF-colonized CWs showed stronger increases in alpha diversity, with Chao1 and observed species increasing by 4.47% and 5.03%, respectively, compared with 2.31% and 3.30% in non-AMF CWs. Beta diversity analyses by PCoA and NMDS showed tight clustering before operation but clear separation after operation, especially under pesticide stress. The first two PCoA axes explained 29.55% of total variation, with PCo1 accounting for 19.62% and PCo2 for 9.93%. Bray-Curtis analysis confirmed significant differences between control systems and non-AMF herbicide systems (p = 0.001, R = 0.23) and between AMF-colonized and non-AMF systems under fungicide stress (p = 0.01, R = 0.38). Random forest analysis further showed pesticide specific responses. Under tebuconazole stress, the number of key microbial species was strongly reduced, with only two detected in non-AMF systems and none in AMF-colonized systems, which suggests a stronger suppressive effect of the fungicide on community structure. Under S-metolachlor stress, the number of key microbial species increased under AMF colonization. A pesticide specific signal was also observed in metabolic and transport functions. After S-metolachlor stress, several ABC transporter genes increased significantly under AMF symbiosis (p = 0.017-0.047), but this pattern was not detected under tebuconazole. Gene enrichment analysis supported these trends. Under S-metolachlor, AMF-colonized systems showed stronger enrichment of oxidative phosphorylation and porphyrin metabolism. Under tebuconazole, non-AMF systems showed strong enrichment of ABC transporters and two-component systems, with about 12,000-16,000 enriched genes, whereas AMF-colonized systems lacked ABC transporter enrichment and showed reduced two-component enrichment. Overall, CWs achieved high removal efficiencies for both tebuconazole and S-metolachlor, mainly through early substrate adsorption. AMF did not markedly increase total pesticide removal, but it altered pesticide distribution and metabolite patterns, increased carbon availability, improved nitrogen removal, buffered the loss of carbon and nitrogen related functions, and enhanced microbial diversity and resilience under pesticide stress. 6. Main conclusions (1) AMF formed a stable symbiosis with Iris pseudacorus in CWs under both tebuconazole and S-metolachlor stress. Neither pesticide suppressed overall AMF colonization, although arbuscular abundance remained low. (2) AMF improved carbon and nitrogen dynamics in CWs. AMF colonized systems generally showed higher TOC and better nitrogen removal, while phosphorus removal was less affected and was mainly controlled by substrate-related processes. (3) Both tebuconazole and S-metolachlor were removed efficiently, with substrate adsorption as the dominant pathway. AMF did not significantly increase total pesticide removal, but it changed pesticide partitioning within wetland compartments. (4) AMF influenced pesticide transformation and metabolite distribution. Four tebuconazole metabolites were identified in CWs for the first time, and AMF was generally associated with lower S-metolachlor metabolite abundance in water and plant tissues. (5) AMF altered microbial community composition and increased microbial diversity under pesticide stress. Its effects were linked to the enrichment of taxa associated with organic matter degradation, nutrient cycling, and stress tolerance. (6) AMF helped maintain carbon and nitrogen related microbial pathways and functional genes, especially under tebuconazole stress. This indicates that AMF buffered pesticide induced functional loss and supported microbial stability in CWs. (7) Overall, the main role of AMF in pesticide-stressed CWs was not to increase total pesticide removal, but to improve system resilience, regulate pesticide behavior, and support microbial functioning. 7. Scientific contribution and novelty This thesis is novel in that it systematically investigated the role of AMF in CWs under pesticide stress within a unified experimental framework. Previous studies have mainly focused on overall treatment performance or on the general ecological functions of AMF in soil-plant systems. In contrast, this thesis examined how AMF influences pesticide fate beyond simple removal efficiency. The results showed that AMF affected pesticide transformation, metabolite formation, and compartment specific distribution in water, substrate, and plant tissues. In particular, four tebuconazole metabolites were reported in CWs for the first time, while AMF also reduced the abundance of many S-metolachlor metabolites in the liquid phase. These findings extend current understanding from parent compound removal to transformation behavior and internal distribution in wetland systems. Another important novelty of this thesis is that it linked pesticide behavior with rhizosphere microbial responses at the metagenomic level. The results showed that AMF altered microbial diversity, community composition, metabolic pathways, and functional genes related to carbon, nitrogen, and phosphorus cycling under pesticide stress. These changes suggest that AMF contributes to microbial stability, stress resilience, and functional maintenance in CWs. By combining pollutant fate analysis, metabolite profiling, and metagenomic investigation under consistent experimental conditions, this thesis provides new insight into AMF mediated bioremediation processes and offers a scientific basis for the future use of AMF in wetland treatment systems exposed to pesticides. This thesis demonstrates that integrating AMF into CWs can improve system robustness under fungicide and herbicide stress. AMF formed a stable symbiosis during operation and was not suppressed by either tebuconazole or S-metolachlor, confirming its feasibility for long term wetland application. Although overall pesticide removal remained dominated by substrate adsorption, AMF altered internal fate and stress responses within the system. AMF increased organic carbon availability and consistently enhanced nitrogen removal, while phosphorus removal was less sensitive to AMF colonization. At the mechanistic level, AMF reshaped pesticide distribution and transformation, reducing residues in water and plant tissues for some compounds and modifying metabolite patterns across compartments. Microbial analyses further showed that AMF increased community stability and buffered functional losses under pesticide stress, particularly for carbon and nitrogen cycling pathways, while inducing pesticide specific metabolic and transport responses. Overall, the results highlight AMF as an effective ecological regulator rather than a direct driver of pesticide removal. The main contribution of this work is linking wetland treatment performance with microbial and functional responses, which clarifies how AMF help maintain system function under pesticide stress. These findings support the use of AMF assisted CWs as a nature-based strategy to stabilize nutrient removal and microbial function in agricultural runoff treatment systems exposed to pesticides. 8. Limitations This thesis is based on laboratory-scale CWs operated under controlled conditions, which simplifies hydrology, vegetation, and microbial complexity compared with field systems. The experiments focused on single plant and AMF species, defined pesticide concentrations, and short- to medium-term operation, which may not capture long term dynamics, mixed contaminant effects, or seasonal variability. Microbial functions were inferred from metagenomic potential rather than direct activity measurements. Therefore, the results mainly reflect mechanistic responses under controlled conditions and should be cautiously extrapolated to full-scale wetlands. Furthermore, the specific mechanisms by which pesticides are transformed within plants colonized with AMF have not yet been explored. This point should be studied through a separate experiment. 9. Future work Future work should focus on disentangling AMF driven mechanisms beyond removal efficiency, particularly the regulation of pesticide partitioning, metabolite profiles, and microbial functional stability. Comparative studies across different pesticide classes, concentration regimes, and AMF plant combinations would clarify whether the observed buffering effects are compound specific or represent a generalizable ecological function. Integrating isotope tracing, enzyme assays, or flux-based measurements could strengthen links between microbial potential and process rates. Such efforts would improve mechanistic understanding and support the rational design of AMF-enhanced CWs for agricultural pollution control. As mentioned in the Limitations section of this thesis, the mechanisms by which plants colonized with AMF absorb pesticides through mycelium and how these pesticides are transformed within plant tissues still require further investigation. The experimental setup conceived by the authors is shown in Fig. 6.2 below, which will further elucidate the mechanisms of AMF.

Keywords

Arbuscular mycorrhizal fungi, Constructed wetlands, Pesticides, Metagenomics

Jazyk práce: angličtina

Datum vytvoření / odevzdání či podání práce: 26. 3. 2026

Obhajoba závěrečné práce

• Vedoucí: doc. Zhongbing Chen, Ph.D.
• Oponent: Miroslav Černík, externi, Pedro Carvalho, externi

Česká zemědělská univerzita v Praze

Fakulta životního prostředí

Doktorský studijní program:
Applied and Landscape Ecology