The effect of WWTP products amendments on Phaseolus vulgaris rhizosphere and its ability to inactivate clarithromycin

Abstract

With increasing efforts to reuse wastewater treatment plant (WWTP) products in agriculture, assessing their impact on soil-plant systems is crucial, while the effects of accompanying antibiotic residues on soil microbial communities have not yet been adequately studied. This study focuses on clarithromycin (CLR), highly present in wastewater, and investigates the CLR-degradation potential of plant-associated microorganisms. Phaseolus vulgaris plants were grown in raised beds filled with Haplic Cambisol and amended with or without WWTP products (treated wastewater, biosolid, or composted biosolid), as a source of CLR residues. The rhizosphere microbiomes after biosolid amendments was significantly enriched by Pseudomonadaceae as assessed by 16S rRNA metagenomics and cultures enriched by CLR revealed dominance of Proteobacteria. However, no degradation of CLR by microbial consortia or enrichment cultures was observed, suggesting the multiplication of CLR-resistant bacteria with other resistance mechanisms. Cultivation-based approach combined with antibiotic modulation assays and subsequent LC-MS analysis confirmed the complete CLR removal by seven phylogenetic groups of actinomycetes in vitro. The proportion of isolates indicated that the rhizosphere is a natural reservoir for CLR-inactivating microorganisms; however, the amendment of soils with WWTP products can significantly increase their abundance and diversity.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-14953-6.

Introduction

The rhizosphere plays a crucial role in plant health and the functioning of the soil ecosystem. Complex root-microbe interactions improve plant fitness by facilitating nutrient uptake (nitrogen fixation, phosphorus solubilization, siderophore production), producing various growth-promoting hormones that help reduce the negative effects of environmental stressors such as drought and salinity^1,2 and by providing a broad spectrum of antimicrobial compounds that act as a barrier against pathogen invasion³. In addition to these fundamental functions, various plant-associated microorganisms degrade a wide range of anthropogenic pollutants in the environment. This capability is extensively utilized in phytoremediation practices, particularly within constructed wetland⁴. The micropollutants are ubiquitous and enter the soil and thus the rhizosphere through human activities. One of the main sources of pharmaceutical contaminants are municipal and industrial wastewater^5,6. Recently, the focus has shifted to the circular economy and the efficient use of by-products from wastewater treatment plants (WWTPs) in agriculture. Examples include the use of biosolids as plant fertilizer and the use of treated wastewater (TWW) for irrigation. This practice raises important questions about the impact of pharmaceutical residues in WWTP products on the soil and plant microbiome, and their subsequent fate in the environment.

The range and extent of pharmaceuticals present in WWTP products is related to the proportion of their consumption in the population⁷. One of the most worrying are antibiotics, as multiresistant pathogens are rapidly developing. Macrolide antibiotics are critically important antibiotics in human therapy⁸and as such they are among the most commonly prescribed antibiotics in Central and Western Europe⁹. They comprise a broad spectrum of molecules consisting of a macrocyclic lactone ring with attached deoxy-sugar or amino-sugar residues and have a bacteriostatic effect by inhibiting proteosynthesis. Their spectrum of activity depends on the size of the macrolactone ring (macrolides with 12- to 16-membered rings) and the presence or absence of cladinose. In human medicine, macrolides mainly comprises the well-known representative erythromycin (natural product, ERY) and its derivatives clarithromycin (CLR) and azithromycin (AZI).

In WWTP, the reduction of macrolide antibiotic concentrations involves both biotic and abiotic processes. While mechanisms such as sorption to sewage sludge and photodegradation via UV disinfection contribute to macrolide attenuation, microbial degradation in activated sludge systems is the main route for their removal¹⁰. Nevertheless, these processes are often insufficient to achieve complete elimination, resulting in the persistence of macrolide residues in the final effluents and sludge-derived products. The concentrations of the various macrolides in TWW then range from nanograms to micrograms per liter¹¹. Occasional deviations are observed in TWW from pharmaceutical companies¹². In biosolids, macrolide concentrations are found in tens of nanograms per gram^13–16. Thus, their concentrations often exceed the PNEC (Predicted No Effect Concentration) values, which were developed to assess the risks of pollutants to biological systems and the risk associated with the spread of antibiotic resistance mechanisms (PNEC_res)¹⁷ in surface waters. The most commonly prescribed in the EU is CLR¹⁸, 40% of which is excreted in active form¹⁹. CLR exceeded the PNEC values in 32.1% of 123,761 environmental samples analysed worldwide from various sources (hospital wastewater 66% > municipal wastewater 63% > surface water 5%²⁰) and was identified as the macrolide with the highest risk quotient (RQ) associated with resistance development during wastewater treatment²¹.

With the WWTP products, macrolides enter the freshwater and soils, resulting often in alarming concentrations in both matrices^22–26. Their fate in the receiving soil is influenced by various soil properties such as pH, cation exchange capacity, clay and organic matter resulting in high sorption of macrolides to soil particles and in their high persistence in the environment^27–29. However, experimental designs comparing sterile and non-sterile soils have also shown significant role of microorganisms in degradation of ERY^30–32 indicated often by sharp decrease of half-lives, DT₅₀. In addition, previous exposure of soil to macrolides indicated increased degradation and selection of degrading soil microbiota in a field study³³. What is more, direct in vitro degradation of ERY by soil and cave bacteria has occasionally been demonstrated^34–37. In addition, WWTP products have previously been associated with CLR-degrading microorganisms^38,39 and may increase the degradative capacity of plant-associated indigenous microbiome. However, none of such in vitro experiments were conducted with CLR. Furthermore, strain should be considered as non-pathogenic, contaminant-tolerant microbe with high environmental adaptability for use in remediation strategy. In this context, spore-forming filamentous bacteria of the genus Streptomyces are of great interest. Streptomycetes are typical soil saprophytic bacteria with a high metabolic versatility known for production of variety of extracellular enzymes that ensure broad spectrum of degradation capacities⁴⁰.

Despite the recognized potential of biodegrading microorganisms in the remediation of contaminated sites, there is little comprehensive information on the soil-plant microbiome responsible for CLR degradation in both uncontaminated and contaminated soils. Therefore, in our study, we aimed to characterize rhizosphere bacterial composition and its ability to degrade macrolide antibiotic CLR as affected by soil management type. We hypothesize, that the application of WWTP products containing residues of CLR will enhance degradation capacity of rhizosphere microbiome. We expected that higher and continuous CLR preasure will select CLR-degrading microorganisms. Although many Gram-negatives were previously associated with degradation of macrolides, we hypothesise that Gram-positive filamentous-like bacteria typical for the soil environment are abundant CLR-degrader as well. To achieve our goals, we characterized rhizosphere of plants grown in two year mesocosm trial representing four different management practices: CMW was irrigated with tap water (control soil), CME was irrigated with TWW, CMB was amended with biosolid, and CMC was amended with composted biosolid. In the crop-rotation system, rhizosphere of beans (Phaseolus vulgaris) grown in the second year of the trial (year 2022) was collected. Different approaches can be used to obtain and characterize degradative microorganisms: isolation of microbial consortia enhancing synergistic mechanisms^10,41or isolation of pure cultures^36,42. In addition, prior enrichment of the microbial consortium with macrolides at high concentrations under co-metabolic conditions can increase the yield of degrading microorganisms³⁹. Here, we followed all the approaches. First, we characterized the bacterial composition of the rhizosphere, and evaluated the degradation of CLR by whole microbial consortia (DG1) and by microbial consorcia enhanced by increasing enrichment of CLR (DG2). Secondly, in view of previous studies showing the importance of actinomycetes in the degradation of non-macrolide antibiotics and other various pollutants and their potential in bioremediation^40,43,44we focused on the CLR inactivating actinomycete cultures.

Results

Occurrence of selected macrolides in WWTP products and treated soils

The treated wastewater contained all three macrolides tested (ERY, CLR, AZI), with concentrations being lowest for ERY (median values 27 and 21 ng L^− 1 in years 2021 and 2022, respectively) and the highest for CLR (median values 280 and 230 ng L^− 1 in years 2021 and 2022, respectively). Concentrations were comparable in both years, with maximum values in April and October (start and end of vegetation season). In contrast, only AZI and CLR were detected in biosolid, and, surprisingly, concentrations varied from year to year with higher values in 2022. Only AZI was detected in composted biosolids. The tap water contained trace concentrations of CLR (5.9 ng L^− 1 and 1.5 ng L^− 1 in years 2021 and 2022, respectively). The concentrations of macrolides in the treated soils and in the water discharged from soil beds were below the limits of quantification. The results are shown in Table 1.

Table 1.

Concentrations of tested macrolides in matrices (TWW, cambisol, tap water, biosolid and composted biosolid).

	TWW 4–11/2021 (n = 30), min-max (median) values [ng L− 1]	Cambisol (n = 2), min-max (median) values [ng g− 1]	Composted biosolid (n = 2),min-max (median) values [ng g− 1]	Biosolid (n = 2), min-max (median) values [ng g− 1]	Tap water 9/2021 (n = 2), min-max (median) values [ng L− 1]
Year 2021
Azithromycin	10–220 (64)	< LOQ	< LOQ	9.4–20 (15)	< LOQ
Clarithromycin	140–670 (280)	< LOQ	< LOQ	7.9–19 (13)	5.6–6.2 (5.9)
Erythromycin	10–41 (27)	-	-	-	< LOQ

	TWW 4–9/2022 (n = 28), min-max (median) values [ng L− 1]	Cambisol, min-max (median) values [ng g− 1]	Composted biosolid (n = 3), min-max (median) values [ng L− 1]	Biosolid (n = 3), min-max (median) [ng g− 1]	Tap water 4/2022–11/2022 (n = 3), min-max (median) values [ng L− 1]
Year 2022
Azithromycin	21–170 (76)	-	24–30 (25)	270–290 (280)	< LOQ
Clarithromycin	98–660 (230)	-	< LOQ	41–47 (44)	1.2–1.7 (1.5)
Erythromycin	5.6–78 (21)	-	< LOQ	< LOQ	< LOQ

LOQ values are provided in Supplementary Table S8. Abbreviations: LOQ – limits of quantification.

Rhizosphere Microbiome as affected by soil management

As expected, soil amendments with biosolid (CMB) and compost (CMC) altered nutrient content when compared to the control (CMW, Supplementary Figure S1). Significant increase were observed in dissolved phosphorus (DP) content (p < 0.001 in both cases) and carbon content in microbial biomass (C_mic; p < 0.05 in CMC). Irrigation with TWW (CME) had no effect on the observed parameters. More details are shown in Supplementary Table S1.

These shifts were not reflected in the changes in bacterial α-diversity of respective rhizosphere (Supplementary Figure S2a), as neither the Shannon index nor Chao 1 differed between managements (p > 0.05 in all cases, Supplementary Table S2). However, the effect of management was evident in prokaryotic β-diversity as assessed by NMDS analyses. Both biosolids amendments (CMC and CMB) formed separate clusters, while CMW and CME showed a greater degree of similarity (Supplementary Figure S2b).

The prokaryotic communities of the rhizosphere were dominated by Proteobacteria (50.9% on average), followed by Actinobacteriota (15.4%). The most abundant families were Commamonadaceae, Nitrosomonadaceae, Pseudomonadaceae, SC-I-84, Gemmatimonadaceae and Xanthobacteraceae (Fig. 1a, b). Differential abundance analysis (DESeq2) showed the differences between the prokaryotic communities at family level as affected by soil management. Thermoactinomycetaceae increased their relative abundance in all soils amended by WWTP products (CME, CMC and CMB) compared to the control, while Oxalobacteraceae were increased only in CMB. The largest taxonomic shift was observed for the Pseudomonadaceae, whose relative abundance increased almost ninefold in CMC and CMB compared to the control (Fig. 1c).

Fig. 1.

Relative abundance of prokaryotic taxa in rhizosphere and enrichment. (a) Heatmap at the level of phylum. (b) Heatmap at the level of family. Only taxa comprising relative abundance higher than 1.5% in more than 20% of samples are shown. (c) Differential abundance analyses (DESeq2) performed on the level of family. Only statistically significant results are shown (p < 0.05). Abbreviations: CMW – soil irrigated with tap water, CME –soil irrigated with TWW, CMC – soil amended with composted biosolid and irrigate with TWW, CMB – soil amended with biosolid and irrigated with tap water; the treatment type supplemented with -E indicates the rhizosphere samples enriched for four months with clarithromycin in liquid minimal medium.

We did not observed any significant differences between managements in the culturable bacteria in the rhizosphere on non-selective medium, however, we observed changes on medium supplemented with CLR (p = 0.712). The CLR-resistant subpopulation was comparable in CMW, CME and CMC, but was significantly larger in CMB compared to CMW (p = 0.011, Fig. 2). The prevalence of CLR-resistant bacteria was three times higher in CMB than in the others (4.2% in CMB and 1.1–1.5% in the other managements). The results of the statistical analysis can be found in Supplementary Table S3.

Fig. 2.

Cultivable bacteria in rhizosphere. logCFU per gram of wet weight of rhizosphere (n = 3) after 6 days of cultivation at 28 °C. Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1. ). Abbreviations: CMW – soil irrigated with tap water, CME –soil irrigated with TWW, CMC – soil amended with composted biosolid and irrigate with TWW, CMB – soil amended with biosolid and irrigated with tap water.

Rhizosphere Microbiome as such showed no measurable degradation of CLR

The results of the liquid chromatography mass spectrometry (LC-MS) analysis (first degradation experiment, DG1) showed that the microbial population of the rhizosphere did not degrade CLR under the given experimental conditions (Supplementary Figure S3). Only the CMW and CME management types were included in DG1, as we assumed the previous selective pressure of CLR present in TWW used for continuous long-term irrigation of plants (e.g. growing seasons in years 2021 and 2022). Both management regimes showed a similar response to CLR as the sole carbon source. The total number of bacterial cells slightly decreased on day 3 and increased on day 11 compared to the control without CLR. The number of cultivable bacteria on day 6 were 9.6 × 10⁵ and 5.9 × 10⁵ CFU per mL in CMW and CME, respectively.

The enrichment cultures revealed amplification of resistant bacteria but no degradation of CLR

To increase the probability of obtaining a CLR-degrading culture, rhizosphere microbial communities were enriched with increasing concentrations of CLR under co-metabolic conditions for up to four months. All management types were included in the enrichment experiment (enriched cultures are referred here as CMW-E, CME-E, CMC-E and CMB-E). Once the final concentration of CLR reached the 1 mg L^− 1, the viability of the enriched cultures was checked by standard cultivation on agar plates. The CFU values were similar regardless of the management (p = 0.780; Supplementary Table S4). Although we expected an activation of the degradation mechanisms under the given experimental conditions, no degradation activity was revealed from LC-MS data. The number of cultivable bacteria on day 6 of degradation experiment using enriched cultures (DG2) showed no effect of soil management (only minor differences, p = 0.169). The total cell counts were also similar on day 3, but surprisingly, the total cell counts increased on day 11 of the experiment in the CMW-E, while it remained stable in the other enriched cultures). The results of DG2 are shown in Supplementary Figure S4.

To characterize the enriched microbial population, we performed sequencing of 16S rDNA gene amplicons from cultures obtained at the final enrichment concentration (1 mg L^− 1). Soil management had no effect on the α-diversity of prokaryotic communities (both Shannon and Chao 1 indices p > 0.05, Supplementary Figure S2s). As expected, the microbial communities of the enriched cultures differed significantly from the original rhizosphere. Enrichment with CLR reduced α-diversity (p < 0.001 in both indices) and altered the structure of the microbial communities significantly (Supplementary Figure S2a, c). The most abundant phylum was Proteobacteria (91.4%) with minor participation of Bdellovibrionota (3.2%), Verrucomicrobiota (1.4%) and Armatimonadota (1%). Proteobacteria increased their relative abundance almost one-fold in enrichments when compared to original rhizosphere (Fig. 1a) and were represented mainly by families Comamonadaceae (28%), Pseudomonadaceae (26.9%), Xanthomonadaceae (15.2%), Chromobacteriaceae (9%), and Caulobacteraceae (4%), followed by Bdellovibrionaceae (3.2%) (Fig. 1b; Supplementary Table S5). Community structure was the most different in CMC-E in comparison to other enrichments, which clustered together(Supplementary Figure S2c). DESeq2 analysis (p < 0.05) revealed that the increase of relative abundance of Sphingomonadaceae was associated with soils irrigated with TWW (both CME-E and CMC-E), and Oxalobacteraceae, whose relative abundance was lower in CME-E and CMB-E in comparison to control (CMW-E, Fig. 1c).

Pure isolates of actinomycetes from rhizosphere inactivate CLR

A total of 111 isolates were obtained from the rhizosphere with typical actinomycetes morphology, which could grow on CLR-supplemented medium. The isolates were tested using solid antibiotic modulation assay (AMA) and yielded 32 CLR-inactivating actinomycetes, of which 28 were unique (39.6% of the total). The lowest proportion was observed in CMW (7.7% of 13 isolates in total) and a comparable proportion in other soil managements: 30% (50 in total), 27.7% (22 in total) and 27% (26 in total) for CME, CMC and CMB, respectively. In addition, another 23 isolates, of which 14 were unique, showed an ambiguous inactivating response. Taxonomic assignment of 16S rRNA sequences of the total 42 isolates with positive and ambiguous response in solid-AMA revealed three bacterial genera of CLR-inactivating actinomycetes: Streptomyces (39 isolates), Nocardiopsis (2 isolates) and Nocardioides (1 isolate) (Fig. 3). Phylogenetic analysis revealed six clades of Streptomyces sp., dominated by isolates of S. albidoflavus group, S. xiamenensis and S. anulatus group. The strongest inactivation activity showed isolates identified as S. xiamenensis (CME and CMC), S. intermedius group (CMB), S. albidoflavus group (all soil management types) and Nocardiopsis alba (CMB), while the isolates with ambiguous response were identified as S. anulatus group (CME, CMC and CMB), S. zaomyceticus (CME), S. nigrescens group (CME) and Nocardioides albus (CME).

Fig. 3.

Maximum-likelihood (PhyML) 16S rRNA gene phylogenetic tree of rhizosphere isolates with positive or ambiguous results in antibiotic modulation assay. Soil management type of isolates are distinguished by colour of the isolate ID (BCCO 10_XXXX). The inner ring shows the clade assignment, the outer two rings show the results of solid- and liquid-antibiotic modulation assay. Abbreviations: CMW – soil irrigated with tap water, CME –soil irrigated with TWW, CMC – soil amended with composted biosolid and irrigate with TWW, CMB – soil amended with biosolid and irrigated with tap water.

The isolates were further tested using liquid-AMA, the selection process included the affiliation to a phylogenetic clade, type of soil management and intensity of CLR-inactivation in solid-AMA. Finally, a total of 30 isolates were tested, of which 70% showed a significant decrease in CLR concentration in the spent medium. The results of both antibiotic modulation assays are shown in Fig. 3. 67% of the isolates gave consistent results in both types of assays.

While some of CLR-inactivating Streptomyces clades gave identical results regardless of soil management type (S. albidoflavus group), some of the highly represented clades showed the management-specific response. Isolates of the S. anulatus group inactivated CLR more efficiently when isolated from rhizosphere irrigated with TWW. In addition, S. xiamenensis isolates were only recovered from the rhizosphere irrigated with TWW (CME, CMC). The accuracy of liquid-AMA was verified by measuring the residual concentration of CLR in the spent medium by LC-MS for selected isolates (Table 2). Total removal (100%) of CLR was confirmed for isolates with strong liquid-AMA activity (formation of no inhibition zone). False-positive results were ruled out by biomass sorption test, which showed no sorption of CLR to microbial biomass (biomass of isolates BCCO 10_2486, BCCO 10_2487, BCCO 10_2488 and BCCO 10_2496 tested).

Table 2.

Results for the isolates, whose biodegradation were verified by LC-HRMS.

Isolate ID	Soil management	The closest relatives (% sequence identity, accession no.)	Solid-AMA (decrease of zone size in mm)	Liquid-AMA (Visible zone of inhibition in mm)	Biomass sorption assay	Average initial concentration of CLR (µg L− 1) ± SD (n = 3), LOQ (min-max) = 1.8–4.8	Removal efficiency LC-HRMS (%)
BCCO_10_2490	CMB	Nocardiopsis alba (99.79%, NR_026340.1)	++	0	NT	1000 ± 60	100
BCCO_10_2489	CMB	Streptomyces intermedius/ koyangensis (99.72%, NR_041103.1/ CP031742.1)	++	0	NT	1000 ± 0	100
BCCO_10_2485	CME	Streptomyces koyangensis (99.72%, NR_025662.1)	+	0	NT	1000 ± 0	100
BCCO_10_2486	CME	Streptomyces anulatus/ pretensis/ praecox/ durocortorensis (100.00%, NR_112527.1/ NR_125619.1/ NR_112358.1/  MW582863.1)	+/-	0	not sorbed	1100 ± 0	98.9
BCCO_10_2496	CME	Streptomyces resistomycificus/ griseochromogenes (100.00%, NR_042100.1/ NR_042102.1)	++	0	not sorbed	1000 ± 60	100
BCCO_10_2488	CME	Streptomyces xiamenensis (99.51%, CP009922.3)	+++	0	not sorbed	1000 ± 0	100
BCCO_10_2487	CME	Streptomyces griseus (99.93%, EU048540.1)	++	0	not sorbed	1100 ± 60	100
BCCO_10_2492	CMC	Streptomyces xiamenensis (99.20%, CP009922.3)	+++	0	NT	1000 ± 0	100

Experiments in solid-AMA were performed on MH agar in 28 °C for 3 days, while in liquid-AMA in MH broth in 28 °C for 5 days. Abbreviation: +/- decrease in ZD by < 3 mm, + decrease in ZD by 4–7 mm, ++ decrease in ZD by 8–14 mm, +++ decrease in ZD by > 15 mm; NT - not tested; CMW – soil irrigated with tap water, CME –soil irrigated with TWW, CMC – soil amended with composted biosolid and irrigate with TWW, CMB – soil amended with biosolid and irrigated with tap water.

Discussion

Given the severe effort for reuse of WWTP products, there is a lack of research addressing the impact of the WWTP products on the rhizosphere microbiome of edible plants and the fate of certain micropollutants. We focused on the biodegradation of the second most common antibiotic, clarithromycin. As expected, CLR was detected in TWW in significantly higher concentrations than other macrolides, following the data about the prescriptions (provided in Supplementary Table S6) with seasonal peak in autumn. The concentrations of macrolides in biosolids varied between years, being higher in 2022, especially for AZI. The different occurrence patterns of tested macrolides in the matrices underlines the fact that the individual macrolide compounds have different physicochemical properties, e.g. higher sorption of AZI⁴⁵. In TWW and biosolid, CLR exceeded PNEC_res (liquid and soil^17,26, suggesting the selection and spread of CLR-resistant bacteria. CLR has previously been identified as one of the main contaminants among antibiotic residues in the aquatic environment²⁰ and is one of the main micropollutant requiring special attention in wastewater treatment⁴⁶. Despite the presence of macrolides in WWTP products, they did not reach measurable levels in the receiving soils even after two years of application. This may be caused by degradation of macrolides by soil microorganisms, however also by their leakage due to heavy rainfalls in the summer season⁴⁷. Trace concentrations of CLR found in tap water are alarming (1.2–6.2 ng L^− 1, Table 1). Previously⁴⁸, found various pharmaceuticals, including macrolides, in the drinking water reservoir (Švihov, the Czech Republic) downstream of outlet of TWW. This highlight the limitations of wastewater treatment methods and emphasize the importance of use of advanced treatment technologies. In contrast, no macrolides were detected in Cambisol prior to the construction of the soil mesocosms.

The rhizosphere is not only a hotspot of microbial activity in the soil, but also a pool of microorganisms that can establish closer relationships with plants or directly colonize their edible parts. Giving the increasing importance of the circular economy, many studies have focused on assessing the safety of WWTP products for agricultural use, but little is known about the effects of WWTP products on the functional attributes of plant-associated microbiome. In our study, we used a wide range of methods to obtain comprehensive information on the effects of WWTP products on the bacterial community in the rhizosphere of P. vulgaris and their potential to degrade CLR. First, using 16S rRNA metagenomics, we found a high resilience of bacteria to the use of TWW for irrigation over two years, similar to that described for bacteria in bulk soil after one year of the same management⁴⁷. On the other hand, amendment of both types of biosolids to the rhizosphere significantly increased the relative abundance of Pseudomonadaceae compared to the control. This is probably related to the improved nutrient status of the soil (i.e. increased DP content), and is also consistent with the high abundance of Pseudomonadaceae previously found in the rhizosphere of P. vulgaris grown in Phaeozem by Medina-Pérez et al.⁴⁹. These changes in the structure of the bacterial community were not reflected in the structure of the enriched cultures. Secondly, the cultivation approaches showed the effect of biosolid on the prevalence of CLR-resistant rhizosphere bacteria, which was not observed with composted biosolid or TWW (Fig. 2). Previously, an increase in resistant bacteria in vitro after biosolids addition had not been observed, however, studies tended to focus on bacteria considered indicative of fecal contamination rather than native soil microbiome⁵⁰. Third, by characterizing the CLR-inactivating streptomycetes isolated from the rhizosphere, we found differences in the observed bacterial taxa and in the CLR-inactivation efficiency related to the type of WWTP product introduced into the soil. The diversity of CLR-inactivating streptomycetes was highest in the rhizosphere irrigated with TWW (seven clades), followed by CMB (four), CMC (three) and the lowest in the untreated rhizosphere (one). The degradation of CLR were also verified for eight strains with positive CLR inactivation. This suggests that agricultural soils harbour indigenous streptomycetes that can degrade CLR and that this capacity is enhanced by the addition of WWTP products. This is probably related to the macrolide residues contained in WWTP products and persistent selection pressure on the rhizosphere microbiome. The knowledge about the occurrence of streptomycetes in WWTP products is limited. However, Streptomycetes are thought to frequently colonize humans⁵¹ and the phylotypes found in this study have previously been associated with human tissue⁵².

We aimed to investigate CLR-inactivation activity and taxonomic assignment of microorganisms in the rhizosphere of P. vulgaris grown in Cambisol under different management regimes: with or without amendments of TWW or biosolids. To characterise and obtain the CLR-degrading microorganisms in the experimental rhizosphere, we pursued two complementary approaches. The first approach involved the isolation of bacterial consortia from the rhizosphere that utilise a variety of microbial enzymes and cell-to-cell signaling pathways to degrade a broad range of pollutants. In addition, to select for CLR-degrading microbial community, we also performed enrichment of the rhizosphere culture with CLR in increasing concentration up to four months. In the second approach, pure Streptomyces-like cultures were isolated from the rhizosphere using standard cultivation techniques. Both approaches are widely used, but here the isolation of CLR-degrading microbial consortia failed, although the viability of the consortia were verified (cultivation on agar plates and microscopy, Supplementary Figures S3-S4). The rhizosphere as an environment with abundant labile organic substrates selects for copiotrophs with faster growth rate⁵³. Some of them were previously identified as ERY degraders in the aquatic environment⁵⁴. Shen et al.³¹ identified Proteobacteria and Bacteroidetes as one of the ERY degraders in soil. Copiotrophs dominated in the enriched cultures too (Comamonadaceae, Pseudomonadaceae and Xanthomonadaceae), and likely employed other CLR-resistance mechanisms for self-protection than CLR-degradation. These could be methylation of the macrolide target within the cell (mediated by erm genes), protection of the target by the production ABC-F protein (mediated by msr genes) or the use of broad-spectrum efflux pumps to expel macrolides from the cell^55,56. However, these were not the subject of this study.

We performed the isolation on agar with CLR-concentration of 1 mg L^− 1, which is below the epidemiological cut-off value (ECCOF) presented by The European Committee on Antimicrobial Susceptibility Testing for mycobacteria (as representatives of actinomycetes)⁵⁷ and below the clinical breakpoint values set for Nocardia and other actinomycetes⁵⁸. This ensured to retrieve actinomycetes posing various CLR-resistance mechanisms including the biodegradation. By the combination of solid and liquid AMA, we obtained representatives of three actinomycetes genera capable of inactivating CLR, which clustered into eight phylogenetic clades. The most of strains were assigned to the genus Streptomyces. In some cases (33%), the differences in the CLR removal efficiency between solid-AMA and liquid-AMA occurred, which could be explained by specific growth requirements, which are difficult to meet in the screening stage. Streptomycetes are typical soil saprophytes, often present active in the rhizosphere of various plants. They prefer to grow in a more structured environment for the complex cellular development and spore formation and have a long generation time⁵⁹. Therefore, in the enriched cultures, they were probably overgrown by fast-growers. Although they were previously isolated from human tissue, they are not considered obligate pathogens except those causing endemic actinomycetoma⁵². On the other hand, they are the main producers of natural antibiotics, including macrolides. Therefore, they may produce enzymes that modify and inactivate the antibiotic for self-defence (glycosylases encoded by ole genes⁶⁰. Moreover, the predictions of the annotated genome sequences of streptomycetes suggest that they may also utilize homologues of hydrolytic enzymes such as erythromycin esterases (Ere) or α/β-hydrolases (Est) able to irreversibly break down the macrolactone ring, the central structure of the macrolides. The action of previously unknown or overlooked genes coding for new enzymes involved in resistance mechanisms cannot be ruled out⁶¹.

The biodegradation of antibiotics is an ancient tool that provides self-protection for antibiotic-producing bacteria and serves as a natural mechanism for detoxifying the environment⁶². At the same time, some of the genes encoding antibiotic-degrading enzymes also confer resistance to antibiotics in pathogenic bacteria⁶³. This dual role of antibiotic-degrading genes needs to be considered and risk assessment related to their spread in the environment requires further investigation. Some of the resistance genes are being monitored globally, especially in clinically important pathogenic species both in clinical setting and in the environment⁶⁴. Previous studies have shown that the pool of resistance genes differs between pathogenic and antibiotic-producing taxa and that there is a phylogenetic barrier to their transmission⁶⁵. This argues for the use of antibiotic-degrading Streptomyces species (natural soil dwellers) in bioremediation to reduce environmental concentrations and thus slow the development and spread of antibiotic resistance in contaminated environments.

Conclusion

The increasing environmental pollution by antibiotics requires a comprehensive understanding of their fate in the environment. The role of the rhizosphere in the degradation of antibiotics is undisputed, but the underlying mechanisms are still largely unknown. Our results show that the rhizosphere of beans (Phaseolus vulgaris) grown in Haplic Cambisol serves as a natural reservoir for actinomycetes that can inactivate the semi-synthetic macrolide antibiotic clarithromycin. However, the application of WWTP products as fertilizer or as alternative water source for irrigation led to a significant increase in the abundance and taxonomic diversity of CLR-degrading actinomycetes. This could be due to the selection pressure exerted by the macrolide concentrations exceeding the PNEC_res in the matrices used, treated wastewater and biosolid. Through a comprehensive assessment of the rhizosphere microbiome and soil chemistry, we found that fertilisation with both biosolids altered the bacterial community structure and dissolved phosphorus content, while composted biosolid also increased soil microbial biomass. Although we did not find significant differences in either the diversity or the structure of the rhizosphere bacterial community after the sole application of treated wastewater compared to the control, the isolate-level approach provided insight into the fundamental changes at the functional level - the abundance and phylogenetic representation of CLR-degrading actinomycetes increased substantially. Our study emphasises the importance of employing both culture-dependent and culture-independent approaches in studies focused on antibiotic degradation. Furthermore, while degradation of antibiotics is a natural detoxification of soil bacteria, it is also a mechanism of antibiotic resistance. This dual role needs to be further investigated to fully understand its impact and assess its potential for bioremediation.

Materials and methods

Chemicals used

Clarithromycin (CLR) with a purity of more than 98.0% was purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). The stock solution (2 mg mL^− 1) was prepared in LC-MS grade ethanol (Merck, Darmstadt, Germany) and filter sterilized (nylon membrane, Thermo Scientific, NY, USA). For the purpose of degradation experiments with bacterial consortia in a growth medium containing CLR as the sole carbon source, CLR was dissolved directly in the culture medium at a final concentration of 1 mg L^− 1 and the medium was then filter sterilized. Nalidixic acid (Merck, Darmstadt, Germany), dissolved in deionized water, and cycloheximide (Merck, Darmstadt, Germany) dissolved in dimethyl sulfoxide (Merck, Darmstadt, Germany) were used to prepare agar media for the isolation of actinomyceter to avoid the growth of fast-growing bacteria nd fungi, respectively. Ultra-pure water (Milli-Q^® Direct 8, Merck, Germany) and acetonitrile (Merck, LC-MS grade) acidified with formic acid (Merck, LC-MS grade) were used for LC-MS analysis.

Experimental set-up

The two-year experiment was conducted at the WWTP in Hrdějovice (Czech Republic) with a current load of 112,000 population equivalents. In spring 2021, raised beds (1.5 × 1.0 × 0.8 m, i.e. width x depth x height) were constructed as previously described⁶⁶ and filled with Haplic Cambisol developed on paragneiss (sandy loam soil). We managed the soil beds in four different ways: (a) CMW was irrigated with tap water (control), (b) CME was irrigated with treated wastewater (WWTP effluent), (c) CMC contained the wet equivalent of 133 t of composted biosolids per hectare and was irrigated with tap water, and (d) CMB contained the wet equivalent of 44 t of stabilized biosolids per hectare and was irrigated with tap water. The experiment started in 2021 and continued in 2022 with one modification: Irrigation with tap water was replaced by irrigation with wastewater for CMC. The amendments of composted biosolids and biosolids were made in the first half of April in 2021 and 2022 (in equal amounts each year). The compost, which contained 15–30% (by volume) of biosolids, was obtained from the composting company DIWENDYS s.r.o. (České Budějovice, Czech Republic). Sanitized and stabilized biosolids (thermophilic anaerobic digestion) were obtained directly from WWTP. The chemical and microbial parameters of the soil were previously described by^47,66,67. In 2021, the beds were cropped with maize (Zea mays L. convar. Saccharata Koern), in 2022 with potatoes (Solanum tuberosum L.; April – July), soybeans (Glycine max L.; April – August) and beans (Phaseolus vulgaris L.; July – October). The experimental setup is shown schematically in Supplementary Figure S5.

Soil and rhizosphere sampling

Bulk soil and rhizosphere samples of P. vulgaris were collected at the time of harvest in October 2022. Bulk soil samples were taken from the topsoil (0–10 cm depth) in four replicates in an approximate amount of 0.5 kg and immediately transported to the laboratory on ice. For the collection of the rhizosphere, the roots of P. vulgaris including the soil and rhizosphere were excavated and transported on ice to the laboratory, where the bulk soil was removed and only the rhizosphere was collected (in three replicates). Both bulk soil and rhizosphere were then separately homogenized by sieving through a 2-mm sieve and stored at -76 °C for further molecular analyses. The bulk soil samples were immediately analysed for nutrients in the soil water extract and dry matter content. Samples of the rhizosphere for microbial cultivation analyses and degradation experiments were used immediately. The bulk soil used for the preparation of the enrichment medium (see below) was stored at 4 °C.

Soil characteristics

Dissolved organic carbon, nitrogen and phosphorous (DOC, DN and DP, respectively) in soil samples were determined as previously described by⁴⁷. Briefly, 10 g of homogenised bulk soil sample were shaken in 50 mL of deionized water (45 min) and centrifuged (1370xg for 30 min). The supernatant was filtered through 0.45 μm membrane (Whatman, ME 25/21 ST). DOC and DN in the filtrate were measured using a TOC analyser (model TOC-LCPH/CPN, Shimadzu), DP was determined spectrophotometrically⁶⁸. Microbial biomass carbon (C_mic) and nitrogen (N_mic) were analyzed by fumigation-extraction method as described in⁴⁷.

Bacterial colony forming units in rhizosphere and isolation of actinobacteria

Five grams of homogenized rhizosphere samples were suspended in 45 mL of 0.2% Calgon (sodium hexametaphosphate, Penta, Praha, Czech Republic) and sonicated for 4 min to detach the bacterial cells from the soil particles. The suspensions were serially diluted in sterile saline and plated on R2A agar (ready-to-use, Himedia, Mumbai, India) with or without selection of CLR (1 mg L^− 1). The bacteria were cultivated at 28 °C. All media were additionally supplemented with nalidixic acid (25 mg L^− 1) and cycloheximide (50 mg L^− 1) to prevent the growth of fast-growing bacteria and fungi, respectively. All soil samples were analysed in triplicate. Filamentous-like bacteria were picked from the R2A agar plates supplemented with CLR after 6 or more days of cultivation and stored as spore suspensions in 15% (v/v) glycerol stocks at -76 °C for long-term preservation. The isolates have been deposited in the Culture Collection of Actinomycetes of the Biology Centre Collection of Organisms (BCCO, www.actinomycetes.bcco.cz).

Degradation experiment with rhizosphere-bacterial consortia (DG1)

As we expected the highest concentration of CLR in TWW, the first degradation experiment with freshly collected and homogenized rhizosphere was performed in triplicate with CMW and CME in October 2022 (here referred to as DG1). Glassware and mineral medium (MM) were prepared as described⁴¹. Experiments were performed in 100 mL Erlenmeyer flasks filled with 50 mL of MM and CLR as the sole carbon source at a final concentration of 0.6 mg C L^− 1. The medium was prepared 1 day before the experiment, filtered with a 0.2 μm membrane and stored at 4 °C. To standardize the size of the inoculum, 5 g of rhizosphere was suspended in 45 mL of MM, sonicated for 2 min, and the total number of bacterial cells in the suspension was estimated by DAPI (4′, 6-diamidino-2-phenylindole) staining and microscopic counting⁶⁹. The suspension was then diluted with sterile MM and used as an inoculum to reach a final concentration of 10⁵ bacterial cells mL^− 1. A sterile, non-inoculated MM with (MM + CLR) or without (MM0) supplementation of CLR served as abiotic controls. The experiment was carried out for 11 days (previously shown as sufficient length for degradation of CLR in liquid conditions^38,39 at 25 °C and in the dark in a rotary shaker (120 rpm). The number of bacterial cells was measured on day 3 and day 11 as described above. The number of CFU was estimated on day 6 by plating a serially diluted suspension on R2A agar supplemented with CLR (1 mg L^− 1) and nalidixic acid (25 mg L^− 1), colonies were read after 6 days of cultivation at 28 °C in the dark. Samples for the measurement of CLR concentrations were collected aseptically in a sterile laminar flow hood with a sterile syringe and needle on day 0 immediately after inoculation and at the end of the experiment (day 11).

Degradation experiment with enrichmed cultures (DG2)

The second degradation experiment required the preparation of enrichment cultures. Enrichment was performed by all management types (CMW, CME, CMC and CMB) in triplicate as follows: Five grams of freshly collected and homogenized rhizosphere was weighed into 100 mL Erlenmeyer flask and suspended in 45 mL MM⁴¹. The experiment was performed at 25 °C in a rotary shaker (120 rpm) in the dark with an initial CLR concentration of 0.01 mg L^− 1. To acclimate microbial consortia CLR concentration was gradually increased over a period of 14 days by 0.124 mg L^− 1 up to final 1 mg L^− 1 of CLR after 4 months and continued simultaneously with the refreshment of the culture. The enrichment medium for the refreshing cultures was prepared by sterilization (autoclaving) of 5 g of bulk soil of the corresponding soil management, suspended in 45 mL of fresh MM. The cooled medium was supplemented with sterile CLR solution and inoculated with 5% (2.5 mL of the culture from previous 14 days-period. The microbial community obtained 7 days after reaching a CLR concentration of 1 mg L-1 is referred to herein as the enriched culture (denoted CMW-E, CME-E, CMC-E and CMB-E). To estimate the CFU in the enriched culture, 1 mL of the suspension was serially diluted and plated on the R2A agar supplemented with CLR (1 mg L^− 1) and nalidixic acid (25 µg L^− 1). The bacterial community structure of the enriched culture was assessed by 16S rRNA gene amplicon sequencing (see below).

The enriched cultures served as inoculants for the second degradation experiment (here referred to as DG2), using the same experimental setup as described for DG1, with the only exception that the initial concentration of CLR as the sole carbon source in DG2 was increased to 1.64 mg L^− 1 which represent 1 mg C L^− 1.

CLR-inactivation experiments using pure bacterial cultures

The bacterial isolates obtained from the rhizosphere were tested for their ability to inactivate CLR. Prior to testing, isolates were revived on yeast malt extract agar (ISP medium no. 2) and refreshed on oatmeal agar (ISP medium no. 3; both media preparation as described by⁷⁰ for 7 days at 28 °C.

Initially, all isolates were tested on solid medium using the antibiotic modulation assay (here solid-AMA) described by⁷¹ with minor modifications. Briefly, CLR (2 mg L^− 1) was applied to the centre of Mueller-Hinton agar plates (Dulab, Czech Republic; 9 cm diameter) in a volume of 5 µl and soaked for 2 h. Then a loop of the fully grown strain of interest was inoculated from the edge to the centre of the agar plate and cultured for 72 h at 28 °C in the dark. To demonstrate the inactivation ability of the test strain, agar plates were overlaid with an overnight culture of Staphylococcus aureus DSM 346 (100 µl of 0.5 McFarland) resuspended in 10 mL of Mueller-Hinton soft agar (MH broth, Carl Roth, Germany; 1% agar, VWR, Pennsylvania, USA). The soft agar was cooled to 42 °C before inoculation. The mixture was poured evenly onto the plate and cultured at 37 °C for 21 h. Each isolate was tested in two replicates. The visible decrease in the zone of inhibition of S. aureus near the tested culture, indicating a decrease in CLR concentration, was assumed to be a CLR-positive inactivation reaction. The intensity of inactivation was categorised as strong (> 15 mm), medium (8–14 mm), low (3–7 mm) or ambiguous (0–3 mm) based on the decrease in the inhibition zone diameter near the tested culture.

Secondly, the positive results of the solid-AMA were verified in a nutrient-rich liquid medium (MH broth) supplemented with CLR (1 mg L^− 1) (referred to here as liquid-AMA). The experiment was performed in 100 mL Erlenmeyer flasks covered with cellulose stoppers and aluminium foil in a final volume of 50 mL. The 72-hour pre-culture grown in MH broth served as inoculum. 200 µL of the preculture (washed twice with sterile saline), diluted to 2 McFarland scale values, was inoculated into the medium and cultured for up to 5 days at 28 °C in a rotary shaker (140 rpm) in the dark. Each isolate and the abiotic control were tested in triplicate. On day five, the spent medium was filter-sterilized using a regenerated cellulose membrane with a pore size of 0.2 μm (Chromafil, Macherey-Nagel, Düren, Germany). The spent medium was applied to a sterile paper disc (50 µL), followed by the disc diffusion method performed on MH agar (Dulab, Czech Republic). The development of the inhibition zone of the indicator strain formed by the CLR residue in the spent medium was compared with the inhibition of CLR in the one of abiotic control. If a decrease in the diameter of the inhibition zone was observed compared to the control, the CLR was inactivated by the given culture. Bacillus subtilis CCM 1718 served as the indicator strain here, as it formed larger inhibition zones with easily readable results. The results of the selected strains that showed the highest CLR inactivation and belonged to different clade or soil management (see below) were verified by LC-HRMS (spent medium collected on day 0 and 5 of cultivation).

To evaluate and exclude the effects of sorption of CLR on bacterial cells (false-positive results), a biomass sorption control was performed in triplicate as follows: 5-day-old cultures were autoclaved and then subjected to liquid-AMA assay as described above. The inhibition zone of indicator strain formed by CLR residue in the spent medium of tested isolates, identical to the inhibition zone of abiotic control, was assumed as not-sorbed CLR. The representative strains selected for the biomass sorption test were chosen based on their growth morphology in the liquid medium (cloudy appearance or growth in clumps; 4 isolates in total).

Analytics of clarithromycin in degradation experiments

The concentrations of macrolide compounds (CLR, AZI and ERY) in complex samples (treated wastewater, tap water, Cambisol, composted biosolid, biosolid, soil in the beds and water drained from the beds) were analysed by liquid chromatography connected with high-resolution mass spectrometry (LC/HRMS) using the methods described in detail in our previous publication⁷².

Samples from the degradation experiments DG1 and DG2 as well as liquid-AMA were immediately filter sterilized through a regenerated cellulose membrane with a pore size of 0.2 μm (Chromafil, Macherey-Nagel, Duren, Germany) and stored at -20 °C. The CLR concentration was quantified by LC-MS as follows: 5 ng of an internal standard (Clarithromycin-D3, Toronto Research Chemicals, Canada) was added to 500 µL of the sample. The samples were analysed by HPLC with MS/MS detection (TSQ Quantiva, Thermo Fisher Scientific). The separation of the analyte was performed at Arion polar C18 column (Chromservis, Czech Republic) by the gradient of acetonitrile in water (LC-MS grade, Merck, Darmstadt, both acidified with 0.1% of formic acid). Detailed information on the analysis can be found in Supplementary Table S7. Limits of quantifications (LOQ, ng mL^− 1) of macrolides in all the matrices are provided in Supplementary Table S8.

Bacterial pure cultures identification and phylogeny tree construction

All bacterial isolates with positive or ambiguous response in solid-AMA were identified by 16S rRNA gene similarity. DNA was extracted from the bacterial biomass using the Nucleospin Microbial DNA Kit (Macherey-Nagel, Düren, Germany) according to the manufacturer’s instructions for Gram-positive bacteria. PCR amplification was performed as previously described⁷³. Sequences were edited and assembled in Geneious v 8.1.9 (Biomatters, Auckland, New Zealand) and compared with GenBank sequences of type strains using the Basic Local Alignment Search Tool⁷⁴ to identify the most closely related species. To exclude identical strains from the following analytical methods, we compared the genetic fingerprint of selected isolated from the same soil management and assigned to the same species (clade) using box-PCR⁷⁵.

The phylogenetic tree was constructed in Geneious using the ClustalW alignment and the GTR + I + G model using PhyML. The final visualization was done in iTOL v. 6.9.1⁷⁶. The reference sequences for the phylogenetic tree were downloaded from Genbank and are listed in Supplementary Table S9.

Prokaryotic community analyses by 16S rRNA gene amplicon sequencing

Genomic DNA from the rhizosphere and enriched cultures was extracted using the NucleoSpin^®Soil Kit (Macherey-Nagel, Düren, Germany) according to the manufacturer’s instructions with a final elution volume of 60 µL. The quality and concentration of the DNA was checked using the NanoDropTM One UV-Vis spectrophotometer (ThermoFisher Scientific, Waltham, USA). The extracted DNA was stored at –18 °C.

For bacterial community structure analysis, the hypervariable region V4 of the bacterial 16S rRNA gene was amplified using primer pair 515F (5′-GTGCCAGCMGCCGCGGTAA-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′) designed by⁷⁷. PCR reactions were performed in duplicate using FastStart PCR Master (Roche, Basel, Switzerland) and Bovine Serum Albumin (BSA, final concentration 0.6 mg mL^− 1; ThermoFisher Scientific, Waltham, USA) in the final volume of 25 µL. The reaction conditions were as follows: initial denaturation 94 °C for 5 min, 28 cycles of 94 °C for 45 s, 50 °C for 45 s, 72 °C for 45 s and final elongation at 72 °C for 10 min. The purity and amount of DNA in the PCR product was checked on an agarose gel (2%). Pooled replicates were then submitted for paired-end sequencing (2 × 153 bp) using an Illumina MiniSeq platform (UIL, Chicago, USA).

Raw paired-end sequences were demultiplexed by the sequencing facility. The sequences were filtered (low quality sequences with a maximum expected error of 2), denoised, merged and analysed for the presence of chimeric sequences with final Amplicon Sequencing Variant (ASV) clustering using the DADA2 pipeline^78,79. ASV taxonomy was assigned using the SILVA 16S rRNA gene database (v.138.1)⁸⁰ with a confidence threshold of 99% using the RDP classifier⁸¹.

Statistical analysis

The effects of management on soil chemical properties and the number of CFU from the rhizosphere and degradation experiments were evaluated by one-way ANOVA using the aov function of R⁸². The effects of soil management and cultivation day in DG1 and DG2 on the number of bacterial cells and the effects of management and soil origin (either rhizosphere or enrichment) on the α-diversity of the respective microbiome were evaluated by two-way ANOVA. If ANOVA results were significant, a pairwise post-hoc comparison was performed with Tukey’s adjustment of p-values using the TukeyHSD function. Data were checked for normal distribution using the Shapiro-Wilk test and log-transformed where necessary.

Soil microbial community data were analysed in R using phyloseq v.1.44.0⁸³. Alpha diversity was estimated using the Chao1 richness index and the Shannon index⁸⁴ from datasets rarefied to minimal sequencing depth (2742 sequences). The effects of soil management and microbiome origin (either rhizosphere or enrichment) on both indices were assessed by a two-way ANOVA of logarithmized values, followed by a pairwise post-hoc comparison with Tukey adjustment of p-values. The effect of soil management on bacterial communities were visualized using non-metric multidimensional scaling (NMDS) of Bray-Curtis dissimilarities of square-root transformed ASV relative abundances. To determine the bacterial families that were significantly affected by soil management, differential gene expression analysis (DESeq2)⁸⁵ was used on non-rarefied and untransformed datasets. All graphical outputs were generated using the ggplot2 package⁸⁶ in the R environment (v. 4.3.1)⁸². All statistical tests were evaluated at significance level of α = 0.05.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

L.K. – conceptualisation, investigation, methodology, formal analysis, visualisation, writing – original draft; K.G. – data curation, formal analysis, writing – review and editing; H.Š. – data curation, formal analysis; A.V.S. – data curation, formal analysis; M.P. – formal analysis, methodology; R.G. – investigation, formal analysis, writing – review and editing; R.K. – funding acquisition, writing – review and editing; A.C. – data curation, investigation, writing – original draft.

Funding

The project was funded by The Ministry of Agriculture Czech Republic (QK21020080-The fate of selected micropollutants, which occur in treated water and sludge from wastewater treatment plants, in soil) and by The Grant Agency of the University of South Bohemia in České Budějovice (063/2023/P-Microbial degradation of macrolide antibiotics in soil-plant ecosystem treated with wastewater treatment plant products).

Data availability

All data generated or analysed during this study are included in this published article (and its Supplementary Information files). The 16S rRNA metagenomic sequence data obtained from microbial consortia have been deposited in the European Nucleotide Archive (ENA) at EMBL-EBI under accession number PRJEB75333 and are available at the following URL: https://www.ebi.ac.uk/ena/browser/view/PRJEB75333. 16S rRNA sequences of actinomycete isolates obtained in this study have been deposited in the GenBank under accession numbers PQ423591-PQ423632 (Supplementary Table S9).

Declarations

Competing interests

The authors declare no competing interests.