|Home | About | Journals | Submit | Contact Us | Français|
Thiabendazole (TBZ), imazalil (IMZ), ortho-phenylphenol (OPP), diphenylamine (DPA), and ethoxyquin (EQ) are used in fruit-packaging plants (FPP) with the stipulation that wastewaters produced by their application would be depurated on site. However, no such treatment systems are currently in place, leading FPP to dispose of their effluents in agricultural land. We investigated the dissipation of those pesticides and their impact on soil microbes known to have a key role on ecosystem functioning. OPP and DPA showed limited persistence (50% dissipation time [DT50], 0.6 and 1.3 days) compared to TBZ and IMZ (DT50, 47.0 and 150.8 days). EQ was rapidly transformed to the short-lived quinone imine (QI) (major metabolite) and the more persistent 2,4-dimethyl-6-ethoxyquinoline (EQNL) (minor metabolite). EQ and OPP exerted significant inhibition of potential nitrification, with the effect of the former being more persistent. This was not reflected in the abundance (determined by quantitative PCR [qPCR]) of the amoA gene of ammonia-oxidizing bacteria (AOB) and archaea (AOA). Considering the above discrepancy and the metabolic pattern of EQ, we further investigated the hypothesis that its metabolites and not only EQ were toxic to ammonia oxidizers. Potential nitrification, amoA gene abundance, and amoA gene transcripts of AOB and AOA showed that QI was probably responsible for the inhibition of nitrification. Our findings have serious ecological and practical implications for soil productivity and N conservation in agriculturally impacted ecosystems and stress the need to include metabolites and RNA-based methods when the soil microbial toxicity of pesticides is assessed.
During postharvest handling, fruits are subjected to treatments with fungicides like thiabendazole (TBZ), imazalil (IMZ), and ortho-phenyl phenol (OPP) (1, 2) and antioxidants like diphenylamine (DPA) and ethoxyquin (EQ) (3), resulting in the production of large volumes of pesticide-contaminated wastewaters. The direct environmental discharge of those effluents without prior depuration entails a serious risk for the integrity of receiving ecosystems, considering the high aquatic toxicity of the pesticides (4, 5). This concern has been addressed by the European Commission (EC), and authorization for those pesticides was given with the stipulation “that an efficient treatment of the produced wastewaters should be operative at a local scale” (6, 7).
Various processes and systems have been tested for the treatment of wastewaters produced by fruit-packaging plants (FPP) including adsorption onto activated carbon (8), photocatalytic degradation (9, 10), and biological degradation (11, 12). However, their high cost, their elevated technological requirements, and the frequent production of oxidized metabolites which are of unknown toxicity compared to the parent compound have hampered their full-scale implementation. Thus, to date effluents from FPP are either discharged in municipal wastewater treatment systems or they are disposed via land-spreading on the surface of nearby field sites. The latter practice could lead to the accumulation of high pesticide concentrations in agricultural soil. However, little if anything is known regarding the effects of pesticides used in FPP on soil microorganisms. This is mainly due to the conviction that their indoor postharvest application entails minimum exposure risk for soil microorganisms. Nevertheless, mishandling of wastewaters from FPP indicates a possible risk of the soil microbial community being exposed to high levels of those pesticides. Data assessing the effects of those pesticides on the overall microbial community and especially on important nontarget soil microorganisms controlling major steps in biochemical cycles are required.
Based on the current regulatory framework, the assessment of the toxicity of pesticides on soil microbes relies on simple C and N mineralization tests (European Commission regulation 1107/2009 ) which do not provide a comprehensive view of potential effects on the microbial diversity and key microbe-mediated functions in soil. In light of the methodological revolution in soil microbial ecology and the recent standardization of several of those methods (13), a revision of the current framework for the assessment of the toxicity of pesticides on soil microorganisms was proposed (14). This will involve the assessment of the toxicity of pesticides for key functional microbial groups like ammonia-oxidizing microorganisms (AOMs) (15). The latter regulate the first and rate-limiting step of nitrification, the oxidation of ammonia to nitrite (16). This key function is performed by ammonia-oxidizing bacteria (AOB) of the Betaproteobacteria and Gammaproteobacteria subgroups of Proteobacteria (17) and ammonia-oxidizing archaea (AOA) of the phylum Thaumarchaeota (18). Their key role in N cycling combined with their general sensitivity to external disturbances make AOMs excellent bioindicators for potential use in pesticide soil microbial toxicity assessment (19, 20).
The main objective of the current study was to provide a thorough insight into the unknown impact of pesticides used in FPP on the structure and function of the soil microbial community. To achieve this, a range of advanced methods focusing on broad microbial functional and structural endpoints (phospholipid fatty acids [PLFAs], microbial respiration, enzymatic activities, and potential nitrification) and on key functional microbial guilds like AOMs were employed. Parallel measurement of the dissipation of the studied pesticides and the metabolism of EQ enabled evaluation of the level and duration of soil microflora exposure to the chemicals applied or formed. Based on the consistent inhibitory effect of EQ on potential nitrification and the metabolic pattern of the antioxidant observed in our first microcosm study, we further focused on the interactions of EQ and its metabolites with AOMs using DNA- and RNA-based approaches, which allowed us to identify the causal agent of the observed inhibition of nitrification.
Analytical standards of TBZ (99.8% purity), IMZ (99.7%), OPP (99.9%), DPA (99.9%), and EQ (97.8%) were purchased from Fluka (Fluka Analytical, Switzerland) and were used for analytical purposes. The metabolites of EQ, quinone imine (QI) and dimethyl-ethoxyquinoline (EQNL), were synthesized as described by Thorisson et al. (21). Their chemical structures, along with those of the pesticides studied, are shown in Fig. S1 in the supplemental material. Commercial formulations of IMZ (Fungazil 50EC) and DPA (No Scald 31.8EC) were provided by ALFA Agricultural Supplies and OPP (Foamer 12.8 SL) by Fomesa Hellas, while TBZ (Hykeep 30 EC) and EQ (Xedaquine 72 EC) were kindly donated by a local FPP.
The study was carried out in a soil collected from a noncultivated plot in the farm of the Aristotle University of Thessaloniki, Greece, characterized as clay (50% sand, 9% silt, 41% clay) with 1.04% organic C content, an N-NO3− content of 37 mg kg (dry weight) soil−1, and pH 7.85. Soil samples were taken from 5 selected points of the top soil (up to 10 cm depth) by following the W nonsystematic pattern of sampling, according to ISO 10381-1 and -2 guidelines (22, 23). The individual samples were mixed thoroughly to provide a single bulk soil sample. The soil was partially air dried, sieved to pass through a 2-mm mesh sieve, and divided into six subsamples (2.5 kg each). The first five received 50 ml of aqueous solutions of the pesticides TBZ, IMZ, OPP, DPA, and EQ, to achieve a dose in soil of 50 mg kg (dry weight) soil−1. The final subsample received the same amount of water without pesticide to serve as an untreated control. This pesticide dose rate was selected based on the following hypotheses: (i) the maximum recommended dose rates for the indoor use of those pesticides are used, (ii) the concentration of the pesticides in the final wastewater needing treatment or disposal was 10% of the initial dose (upon reuse and recirculation, which results in degradation and dilution), (iii) the volume of wastewaters produced by a medium-size fruit-packaging plant is 25 m3, and (iv) wastewaters are disposed of at a field site of 0.1 ha and pesticide residues are distributed at the top 10 cm of the soil profile. After all applications, the soils were left to equilibrate for an hour, and further water was added to adjust moisture to 40% of the water holding capacity. Soil samples were then separated into 150-g portions and placed in aerated plastic bags, which were incubated in the dark at 25°C for 100 days. Immediately after pesticide application and at regular time intervals thereafter, triplicate samples from each treatment were removed from the incubator and stored either at −20°C for pesticide dissipation measurements and soil DNA extraction or at 4°C for phospholipid fatty acid (PLFA) analysis and enzymatic activity measurements. Fresh soil samples were always used for potential nitrification measurements. The moisture content of the samples was kept constant with regular addition of deionized water. In cases where rapid pesticide dissipation was observed (OPP, DPA, and EQ), the experiment was repeated, but its duration was adjusted to 24 h to accurately estimate the 50% dissipation times (DT50s) of the pesticides studied.
To further explore the interactions of EQ and its metabolites with AOMs, a second microcosm experiment was conducted. Thus, 2 kg of the soil used in the first microcosm study was divided into five subsamples of 0.4 kg each. The first soil subsample was treated with an aqueous suspension of EQ at a rate of 50 mg kg−1 soil. The next three subsamples were treated with 5 ml of acetonitrile solutions of QI, EQNL, or their mixture, resulting in soil concentrations of 16.1 and 1.4 mg kg−1 respectively. These intended doses were selected to reflect the maximum concentrations of QI and EQNL formed in soil during the first microcosm study. The final subsample was treated with an equivalent amount of acetonitrile and water to serve as an untreated control. Furthermore, EQ-treated samples were also treated with 5 ml acetonitrile to ensure that all treatments received the same volume of water and acetonitrile. Subsequent soil handling, dissipation, and potential nitrification measurements were performed on soil samples collected at 3, 10, 30, 60, and 90 days postapplication as described above, while soil samples for DNA and RNA extraction were collected at 3, 30, and 90 days and stored at −80°C until further processed.
Residues of TBZ, IMZ, OPP, EQ, and DPA were extracted from soil as described by Karas et al. (24). Special care was given to minimize the transformation of EQ to its oxidation derivatives during extraction (in a dark cold room, at 4°C).
High-performance liquid chromatography (HPLC) analysis was performed in an 1100 HP Agilent HPLC system equipped with a UV detector and a reverse-phase (RP) C18 GraceSmart (4.6 mm by 150 mm, 5 μm) column. The injection volume was 20 μl, and the flow rate of the mobile phase was 1 ml min−1. Mixtures of methanol and ammonium acetate solution (0.1% [wt/vol]) with different elution strengths were used as the mobile phase to analyze TBZ (70:30 by volume) and IMZ (80:20 by volume), and detection was achieved at 241 nm and 204 nm, respectively. Similarly, chromatographic separation of OPP, DPA, EQ, and the EQ oxidation derivatives QI and EQNL was achieved using mixtures of acetonitrile and ammonium acetate solution (0.1% [wt/vol]) either 70:30 by volume (OPP, EQ, and metabolites) or 80:20 by volume (DPA). OPP, EQ, and metabolite residues were detected at 254 nm, while DPA residues were detected at 210 nm. In all cases, method validation tests showed mean recoveries of >81% and relative standard deviations lower than 12%.
Extraction of PLFAs from soil was carried out as described by Papadopoulou et al. (25). Microbial respiration and total hydrolytic activity were determined by the alkali titration method (26) and the fluorescein diacetate activity (FDA) method (27), respectively. The activities of β-glucosidase and acid and alkaline phosphatases were determined as described by Eivazi and Tabatabai (28, 29). Potential nitrification was determined following the method of Kandeler (30).
RNA was extracted from 2 g of soil using the PowerSoil total RNA isolation kit (MoBio Laboratories, Inc., USA), and DNA was extracted during the same procedure by using the RNA PowerSoil DNA elution accessory kit (Mo Bio Laboratories, Inc., USA). DNA/RNA purity was checked spectrophotometrically and their integrity (shearing and degradation) was checked by electrophoresis. Nucleic acids were quantified using Qubit fluorometer (Invitrogen).
The effect of pesticides on the abundance of AOB and AOA was assessed via qPCR of the amoA gene, which was amplified from soil DNA using the primers amoA-1F/amoA-2R (31) and Arch-amoAF/Arch-amoAR (32), respectively. Amplifications for both AOB and AOA were performed in a 20-μl reaction mixture containing 1× KAPA SYBR Fast qPCR universal master mix (2×), a 0.2 μM concentration of each primer, 1× ROX Low (50×), 400 ng ml−1 bovine serum albumin (BSA), and ca. 10 ng DNA. The thermal cycling conditions used were as described elsewhere (33). The copy numbers of the bacterial and archaeal amoA gene were determined via external standard curves as described by Rousidou et al. (33). qPCR amplification efficiencies were 89.0 and 94.0%, with r2 values of 0.994 and 0.997 for AOB and AOA, respectively.
The effect of EQ and its metabolites on the function of AOB and AOA was further tested via reverse transcription-qPCR (RT-qPCR) of the amoA gene. Soil RNA (~250 ng) was treated with DNase I (1 U/μl) (amplification grade; Invitrogen), reverse transcribed, using 200 U Superscript II (Invitrogen) and a 0.2 μM concentration of amoA2R and Arch-amoAR primers (HPLC grade) for AOB and AOA, respectively, according to the manufacturers' instructions. Amplification of cDNAs for both AOB and AOA was performed following the procedure described above for DNA.
PLFA yields, potential nitrification, enzymatic activities, and qPCR data were subjected to two-way analysis of variance (ANOVA), followed by Tukey's post hoc test (P ≤ 0.05) to identify significant effects of time, pesticide and their interactions. The data of total microbial respiration during 35 days were analyzed via one-way ANOVA and the least significant difference (LSD) test (P ≤ 0.05). The relative abundance data from the PLFA analysis were also subjected to principal component analysis (PCA). Ordination of the different samples according to their PLFA content was performed for the first two principal components (PCs), which included most of the variance (>50%) of the original data set.
Pesticide dissipation data were fitted to a simple first-order kinetics (SFO) model. In cases where degradation data deviated from SFO, the biphasic hockey stick model was used. In all cases, statistical analysis was performed with SPSS 20.0v.
Dissipation of pesticides (Fig. 1) was mostly biphasic and was best described by the hockey stick model (x2 < 15, r2 > 0.75) (Table 1). IMZ was the most persistent pesticide, with an extrapolated DT50 of 150.8 days, followed by TBZ (DT50 = 47.0 days), while OPP and DPA were rapidly dissipated, with estimated DT50s of 0.6 and 1.3 days, respectively (Table 1). EQ was almost simultaneously transformed to quinone imine (QI) and 2,4-dimethyl-6-ethoxyquinoline (EQNL) (Fig. 1c and andd),d), which constituted 28 and 2.8%, respectively, of the total amount of the pesticide recovered in the soil at time zero. The DT50 of total residues of EQ (ΕQ+QI+EQNL) was 2.2 days (Table 1). QI showed limited persistence in soil and reached undetectable levels at 7 days postapplication, in contrast to EQNL, which persisted in soil at trace amounts throughout the experiment (Fig. 1c). The same pattern of dissipation and metabolism of EQ was observed in the 24-h experiment, with QI constituting over 90% of the total residues of EQ 24 h after application of the parent compound (Fig. 1d).
In total, 20 PLFAs were detected in all soil samples at relative abundances above 0.5%: 15:0, a15:0, i15:0, i16:0, 17:0, and i17:0 (Gram-positive bacteria); 16:1ω7, cy17:0, and cy19:0 (Gram-negative bacteria); 18:2ω6,9 (fungi), 10Me16:0, 10Me17:0, and 10Me18:0 (actinobacteria); 16:0, 18:1ω9cis, and 18:1ω9trans (Gram-negative bacteria and fungi); 23:0 and 24:0 (eukaryotic microorganisms); and 14:0, 18:0, and 20:0 (not associated with any specific microbial group) (34).
Statistical analysis showed that regardless of the treatment applied (pesticide or control), a significant reduction (P < 0.001) in PLFA concentration with time (main effect) was observed (Fig. 2a) in the order 3 days > 10 days > 30 days > 90 days = 60 days (see Table S1 in the supplemental material). When pesticide effects were explored, significant differences (P < 0.05) in the total PLFAs between pesticide-treated (mostly in EQ and OPP treatments) and control samples were evident at 30, 60, and 90 days (Fig. 2a), although no clear temporal patterns of pesticide-driven effects were evident in any of the treatments studied (Fig. 2a). PCA of the PLFA data verified the absence of a clear pesticide-driven effect on the structure of the soil microbial community (Fig. 2b), in contrast to time which was the main factor driving the structure of the soil microbial community (Fig. 2c). Soil samples collected at 60 and 90 days regardless of the pesticide treatment were clearly separated from the samples collected at 3, 10, and 30 days along PC1 and/or PC2, in line with the significant reduction of total PLFAs in all treatments at these sampling days.
Soil microbial respiration was not significantly affected (P = 0.278) by pesticides (data not shown). In contrast, all pesticides except EQ significantly increased (main effect P < 0.001) the hydrolytic activity in soil compared to the control. Further analysis of the interactions between treatments and time detected some significant effects of pesticides at certain days which did not show a clear temporal pesticide-driven pattern (see Fig. S2a in the supplemental material).
β-Glucosidase activity was significantly reduced by EQ (P < 0.001) for the first 10 days of incubation, but a recovery was observed from 30 days onwards. The effects of the other pesticides were transient and/or did not show a clear temporal pattern (see Fig. S2b in the supplemental material). No temporal pesticide effect on acid phosphatase activity was observed (data not shown), whereas alkaline phosphatase activity was significantly inhibited (P < 0.05) by EQ and IMZ at the later stages of the incubation (30 days onwards) (see Fig. S2c in the supplemental material).
Regarding potential nitrification, EQ and OPP induced a significant reduction (P < 0.001) compared to the control (Fig. 3a). The inhibitory effect of EQ was persistent, and no recovery was observed by the end of the experiment. In contrast, the inhibitory effect of OPP was restricted during the period between 30 to 60 days. Application of TBZ and DPA significantly stimulated potential nitrification for up to 60 and 10 days postapplication, respectively, whereas no significant effect on potential nitrification was induced by IMZ.
Regardless of the treatments employed, bacterial amoA gene copies were significantly more abundant than archaeal (Fig. 3b and andc).c). Despite the clear inhibitory effects of EQ and of OPP on potential nitrification, no significant effect of pesticides on amoA gene copy number in either AOB (P = 0.107) or AOA (P = 0.058) was observed (Fig. 3b and andcc).
Taking together the data on dissipation, potential nitrification, and abundance of AOMs on EQ-treated soils, a follow-up microcosm study was conducted to gain insight into the interactions of EQ and its metabolites with AOMs. The dissipation of EQ, QI, and EQNL in the different treatments (EQ, QI, EQNL, and the combination of QI and EQNL) showed patterns similar to the ones observed in the first microcosm study (see Fig. S3 in the supplemental material). The application of EQ, QI, and QI+EQNL induced a significant reduction in potential nitrification compared to the control for up to 30 days postapplication (Fig. 4a). On the other hand, EQNL induced a significant increase (P < 0.05) in potential nitrification compared to the control at 3 days, but nitrification reverted to levels similar to that of the control from 10 days onwards.
The abundance of the amoA gene copies of AOB was not significantly affected by pesticide treatments (P = 0.369) (Fig. 4b). On the other hand, significantly lower copy numbers of the amoA gene of AOA (P < 0.05) were measured at 3 days after treatment in the samples treated with EQ, QI, EQNL, and QI+EQNL compared to the control samples (Fig. 4c). However, AOA abundance in those treatments reverted to levels similar to the control from 30 days onwards. When transcript levels were determined, more clear and compound-dependent effects were observed. Thus, significantly lower transcripts of the amoA gene of both AOB (Fig. 4d) and AOA (Fig. 4e) were measured in the EQ-, QI-, and QI+EQNL-treated samples within the first 30 days after application than in the control, while in EQNL-treated samples a significant inhibition was evident only at 3 days. The transcript numbers of the amoA gene in both microbial guilds in those treatments were similar to those of the control at the end of the experiment.
Land-spreading of pesticide-contaminated wastewaters from FPP on agricultural land constitutes a major threat of contamination of soil, which acts as a sink for the further movement of pesticides to surface water and groundwater. Considering the pivotal role of microbes on soil ecosystem functioning (35), we investigated the impact of the pesticides contained in those agroindustrial effluents on the soil microbial community structure, function, and abundance using an array of molecular and biochemical tools. Parallel measurements of pesticide dissipation and metabolism enabled us to determine the duration of exposure of the soil microbes. Particular attention was given to the impact of EQ, whose behavior in soil was largely unknown, on the activity of AOMs, which control the rate-limiting step of N cycling.
The pesticides studied showed marked differences in their soil dissipation. TBZ and IMZ showed moderate to high persistence (Table 1) in accordance with previous reports which showed DT50 of 41 to 135 days for IMZ (36, 37), and 37 to 1100 days for TBZ (38, 39). OPP, DPA, and EQ dissipated rapidly (DT50s < 2.2 days) in line with previous studies (4, 24). The transformation of EQ was studied further, since it was found to be almost simultaneously oxidized in soil to QI and EQNL (Fig. 1c and andd).d). The former was the major metabolite, but it was rapidly degraded, in contrast to the latter, which was formed in small amounts but persisted for the duration of the experiment (Fig. 1d). This is in agreement with the only other study which investigated the metabolism and dissipation of EQ in the soil environment (24). All other studies have looked into the metabolism of EQ in food and animal tissues. Brannegan (40) reported a rapid EQ transformation to three major oxidation products (QI, EQNL, and a dimer of EQ) in lean beef and beef fat, while Gallagher and Stahr (41) reported the presence of EQNL in animal feed. Similarly, He and Ackman (42) detected QI and DM in fish meals and fish feeds. However, the dimer of EQ was not detected in our study.
The application of pesticides at the levels tested in the current study did not induce any clear temporal pattern of pesticide-driven changes on the size and on the structure of the soil microbial community (Fig. 2). In contrast, time was the main factor driving the size and the structure of the soil microbial community in both control and pesticide-treated soils. Similar time-dependent changes either in the size or in the structure of the soil microbial community have been reported in previous microcosm studies investigating the effect of botanical (43) and synthetic (44) pesticides, and they have been attributed to the gradual limitation of resources for the microbial biomass at the latter stages of soil incubations (45).
The impact of pesticides on the total metabolic microbial activity was estimated via microbial respiration and FDA hydrolytic activity measurements. Neither of the two microbial endpoints was significantly impaired by the pesticides tested. Further tests assessed the impact of pesticides on key soil enzymes, β-glucosidases and phosphatases, which participate in C and P cycling, respectively (46), and which were previously identified as the most sensitive enzymatic endpoints for assessing effects of pesticides on soil microbial function (47). The activity of β-glucosidase and phosphatase was affected by pesticide application; however, no clear temporal pattern of pesticide-driven effect was observed (see Fig. S2b and c in the supplemental material).
Potential nitrification was the most responsive microbial endpoint, suggesting strong interactions between pesticides and AOMs, either beneficial (TBZ and DPA) or detrimental (EQ and OPP). Despite their rapid soil dissipation EQ (mostly) and OPP induced inhibitory effects on potential nitrification which lasted beyond the persistence of those pesticides (Fig. 3a), in contrast to the recalcitrant fungicides TBZ and IMZ, which did not show any inhibitory effect on potential nitrification. Our findings suggested a lack of correlation between pesticide persistence and soil microbial toxicity. This is in line with several previous studies which also did not observe any correspondence between pesticides persistence and the activity or abundance of AOMs (48,–51). The high response of potential nitrification to pesticide treatments, especially to EQ, together with the key role of this process in global N cycling forced us to further investigate the impact of pesticides on AOMs via molecular approaches.
In contrast to previous reports (52), AOA gene copy numbers were significantly lower than those of AOB in all soil treatments. This was not surprising considering the relevantly high soil pH (7.85) and the low depth of the soil sampling (up to 10 cm), which are two of the main drivers thought to provide a competitive numerical advantage for AOB over AOA (53, 54). Regarding pesticide effects, DNA-based qPCR showed that TBZ, IMZ, and DPA had no effect on the abundance of AOB and AOA (Fig. 3b and andc).c). This is in line with previous studies which reported effects of pesticides on AOMs only upon exposure to levels well above the recommended rates (55, 56). It should be noted that for the pesticides used in FPP, there are no recommended dose rates for soil application, and their effects were assessed based on application rates calculated according to a realistic wastewater disposal scenario employed by FPP in Europe.
The qPCR data on EQ-treated samples were not in agreement with the persistent inhibitory effect of EQ on potential nitrification. This discrepancy was not surprising, considering that those two methods provide different information which could be viewed as complementary rather than comparable. Indeed, the sole presence of a functional gene like amoA as detected by DNA-based qPCR does not necessarily reflect operative function (16). Based on the above results, we decided to look further into the interactions of EQ with AOMs via RNA-based qPCR determination of the transcription of the amoA gene. Despite the fact that RNA-based analysis has been proposed as the most accurate way to assess the impact of stressors on soil microbial activity (15, 57) and especially on functional microbial guilds like the autotrophic, slow-growing AOMs (58, 59), their use in such studies is very limited (15), and mostly DNA-based data have been employed (49, 56). All the above combined with the almost instantaneous transformation of EQ in soil and its effects on potential nitrification led us to further explore the hypothesis that not EQ (or not only EQ) but its metabolic products, QI and EQNL, might have a role in the inhibition of the nitrification activity.
Thus, in a second microcosm soil study, EQ and its metabolites were applied in the same soil either alone or in combination (Fig. 4). DNA-based qPCR measurements indicated a transient or no effect of all chemical treatments on the abundance of AOA and AOB, respectively, which was visible only for the first 3 days postapplication. However, functional endpoints like potential nitrification and transcription of the amoA gene clearly showed a significant inhibition induced by all treatments where QI was either applied or formed (EQ, QI, and QI+EQNL) and lasted for at least 30 days. It should be noted that potential nitrification and RT-qPCR describe different processes: the abundance of the amoA gene transcripts serves as a measure of the rate of ammonia transformation to the intermediate product hydroxylamine, while potential nitrification also considers the rate of hydroxylamine conversion to nitrite, finally expressing the combined rate of both steps (16). The general agreement between potential nitrification and amoA transcription and the comparative inhibition observed in the transcripts of both AOA and AOB suggest that QI acts mainly on the first step of ammonia oxidation.
The rapid formation and dissipation of QI and its parent compound contrast the slower recovery of ammonia oxidation activity by AOB and AOA (at 30 days). The latter recovery of the ammonia oxidation activity could be attributed to the slow-growing lifestyle of AOB and AOA (μmax = 0.005 to 0.044 h−1 and 0.015 to 0.027 h−1, respectively) (60,–62), which might necessitate a long recovery time upon exposure to drastic toxicants like QI. Previous studies with acetylene, a suicide substrate of ammonia monooxygenase, showed that AOB required at least 10 days to recover upon release from exposure (63). The sacrificial antioxidant character of EQ and its oxidation products (64) implies that QI, which is produced in substantially larger amounts than EQNL, might have a direct impact on ammonia oxidation. On the other hand, EQNL, despite its longer persistence, induced only a transient effect on the transcription of the amoA gene of AOMs, and this might be related either to its low levels formed in soil or to its inherent lower toxicity. In vitro tests of the toxicity of EQ, QI, and EQNL on representative strains of AOB and AOA are pending and are expected to verify our observations in soil and test the biochemical basis of the inhibition observed.
In terms of ecosystem services, nitrification is a key component of N cycling, ensuring the conversion of ammonia released from organic N to nitrate, which is the preferred substrate of plants and aerobic soil microorganisms (18). Thus, inhibitory effects imposed by EQ-containing wastewaters on this process are expected to have undesirable effects on soil fertility and crop growth. On the other hand, nitrification greatly contributes to N losses and environmental pollution through nitrate leaching and denitrification, and its suppression has been used for the management of anthropogenic nitrogen pollution in disturbed agricultural ecosystems (65). Verification of the inhibitory mechanism of QI on both AOA and AOB might be exploited for the development of a new universal nitrification inhibitor. This will be highly beneficial considering that in comparative soil and in vitro tests, all the currently available nitrification inhibitors, including allylthiourea (66, 67) dicyandiamide, nitrapyrin (68), and 3,4-dimethylpyrazole phosphate (69), have shown more prominent inhibition on AOB rather than on AOA. This may have significant implications for agricultural practice, since AOA would be able to contribute to nitrogen fertilizer loss under conditions where AOB are inhibited, further reinforcing the need for novel nitrification inhibitors showing equivalent inhibitory thresholds for both AOA and AOB.
Soil disposal of wastewaters from FPP could result in an undesirable exposure of terrestrial ecosystems to pesticides contained in those effluents. Because of the pivotal role of soil microbes in ecosystem functioning, we evaluated the impact of those pesticides on soil microorganisms. No consistent adverse effects of TBZ, IMZ, OPP, and DPA on soil microbes were observed. In contrast, QI, the major transformation product of EQ in soil, induced a persistent inhibitory effect on the activity of AOMs. Our findings have both practical and ecological implications. On the one hand the disposal of EQ-containing effluents in agricultural land (common practice by FPP in Europe) could result in deleterious effects on N cycling, loss of soil fertility, and reduced crop productivity in regions adjacent to FPP. On the other hand, the uniform inhibitory effect of QI on the activity of AOB and AOA might lead to the development of a novel nitrification inhibitor for more efficient N conservation in agricultural and pasture soils. Apart from the above, our findings reinforce issues relevant in a future revision of the regulatory framework regarding pesticides soil microbial toxicity assessment, as follows. (i) Toxicity assessment should be always extended to relevant metabolites which, as shown in case of QI, could disrupt essential microbially mediated functions. (ii) Experimental approaches assessing the toxicity of pesticides on soil microbes should always involve temporal measurements to provide ample time for recovery. (iii) Broad microbial measurements, structural and functional, might fail to detect significant pesticide effects, and more specialized key soil microbial functional guilds, like AOMs, might identify effects not previously apparent. Thus, nitrification and nitrifiers could be a key endpoint in pesticide soil microbial toxicity assessment. (iv) Use of RNA-based approaches has a great potential to provide the necessary sensitivity to identify effect of pesticides on significant microbial functions which are not detected by DNA-based and biochemical endpoints.
We thank Theodora Matsi for soil physicochemical analysis and Maria Tourna for critical reviewing of the manuscript prior to its submission. Thanks also go to Katerina Souna and Manolis Karazafeiris for the preliminary LC/MS-MS analysis of EQ total residues.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/AEM.03437-15.