Biodegradation Potential and Diversity of Diclofenac-degrading Microbiota in an Immobilized Cell Bioﬁlter

: Despite that diclofenac has been embodied to the European watch list of priority substances of concern, studies on diclofenac biodegradation are limited and the diversity of diclofenac-degrading microbiota remains unknown. In this work, an immobilized cell bioﬁlter was constructed and operated to evaluate its e ﬀ ectiveness to depurate high strength diclofenac wastewater and to identify the diclofenac-degrading community accommodated in activated sludge by employing high-throughput sequencing techniques. After a two-month adaptation period, bioﬁlter removal e ﬃ ciencies reached values as high as 97.63 ± 0.62%, whereas utilization of diclofenac in the immobilized cell bioﬁlter led to a drastic pH decrease. Based on Illumina sequencing, the major bacterial taxa identiﬁed in the immobilized cell bioﬁlter were members of the species Granulicella pectinivorans and Rhodanobacter terrae , followed by members of the species Castellaniella denitriﬁcans , Parvibaculum lavamentivorans , Bordetella petrii , Bryocella elongata and Rhodopseudomonas palustris . The ability of such taxa to utilize a wide range of carbon sources and to e ﬀ ectively adapt under acidic conditions seemed to be the main parameters, which favored their prevalence in the immobilized cell bioﬁlter. In addition, Wickerhamiella was the predominant fungal taxon in the immobilized cell bioﬁlter, which appears to be actively involved in diclofenac degradation in activated sludge systems.


Introduction
The detection of non-steroidal anti-inflammatory drugs (NSAIDs) in aquatic environments is a result of various sources of pollution, exhibiting a negative impact on health and the environment. In particular, pharmaceuticals are disposed of households, veterinary clinics, hospitals and pharmaceutical manufacturing facilities and enter wastewater treatment plants (WWTPs) for processing prior to their final environmental discharge in natural water resources. However, WWTPs are designed mainly to remove conventional pollutants contained in municipal and industrial wastewater and fail to effectively remove NSAIDs [1,2], a fact that has led to the detection of numerous pharmaceuticals in the effluents of WWTPs as well as in surface water and groundwater systems [3,4].
Diclofenac [2-(2,6-dichloranilino) phenylacetic acid] is one of the best-selling NSAIDs worldwide [5]. It is a common environmental contaminant regularly detected in drinking water, river through the operation of a 33 L/h peristaltic pump. The diclofenac-based wastewater was introduced to the biofilter by a liquid pump, so as a hydraulic retention time (HRT) of 3.5 days was achieved. During start-up, the immobilized cell biofilter was inoculated with activated sludge from a local full-scale WWTP.
Processes 2019, 7, x FOR PEER REVIEW 3 of 13 through the operation of a 33 L/h peristaltic pump. The diclofenac-based wastewater was introduced to the biofilter by a liquid pump, so as a hydraulic retention time (HRT) of 3.5 days was achieved. During start-up, the immobilized cell biofilter was inoculated with activated sludge from a local full-scale WWTP.

Physicochemical Analyses
Dissolved oxygen (DO) was recorded by using a WTW Oxi 320 m, whereas pH and electrical conductivity (EC) were measured by using the Metrohm 632 and Crison CM35 probe apparatus, respectively. Chemical oxygen demand (COD) and biochemical oxygen demand (BOD5) were determined according to the "Standard Methods for the Examination of Water and Wastewater" [29]. Physicochemical parameters were determined on regular basis, e.g., three times per week for pH, EC and COD. Total kjeldahl nitrogen (TKN) and ammonium nitrogen (NH4 + -N) measurements were performed through the implementation of the Kjeldahl and the ammonium distillation protocol, respectively. To measure NO3 − -Ν concentration, nitrates were reduced to nitrites through filtration in a Cd-copperized column and subsequently nitrite nitrogen (NO2 − -N) was determined spectrophotometrically at 543 nm as previously described by Clesceri et al. [29]. Determination of orthophosphates as PO4 3− -P was performed through the reduction of the molybdophosphoric acid, deriving from the reaction of orthophosphates with ammonium molybdate under acidic conditions, by stannous chloride solution.

HPLC Determination of Diclofenac
Prior to the analysis, samples from the influent and the effluent of the biofilter system obtained every 5 to 6 days were dissolved in methanol at concentrations of 1:100 and 1:5 v/v, respectively and passed through 0.45 μm membrane filters. A Shimadzu SCL-10AVP HPLC system (Kyoto, Japan), consisting of a Discovery HS C18 column (5 μm, 250 × 4.6 mm) (Supelco, Bellefonte, PA, USA), a LC-10 ADVP pump, a DGU-14A degassing system, a CTO-10ASVP column oven and a SPD-M10 AVP photodiode array detector was used for the chromatographic determination of diclofenac. The mobile phase consisted of a mixture of acetonitrile and 0.2% formic acid in water at mixing ratio of 60:40 v/v. The flow rate was 0.8 mL/min, the injection volume was 20 μL and the calibration standards or samples were analyzed by applying an isocratic elution program at 80 bar and 30 °C for

Physicochemical Analyses
Dissolved oxygen (DO) was recorded by using a WTW Oxi 320 m, whereas pH and electrical conductivity (EC) were measured by using the Metrohm 632 and Crison CM35 probe apparatus, respectively. Chemical oxygen demand (COD) and biochemical oxygen demand (BOD5) were determined according to the "Standard Methods for the Examination of Water and Wastewater" [29]. Physicochemical parameters were determined on regular basis, e.g., three times per week for pH, EC and COD. Total kjeldahl nitrogen (TKN) and ammonium nitrogen (NH 4 + -N) measurements were performed through the implementation of the Kjeldahl and the ammonium distillation protocol, respectively. To measure NO 3 − -N concentration, nitrates were reduced to nitrites through filtration in a Cd-copperized column and subsequently nitrite nitrogen (NO 2 − -N) was determined spectrophotometrically at 543 nm as previously described by Clesceri et al. [29]. Determination of orthophosphates as PO 4 3− -P was performed through the reduction of the molybdophosphoric acid, deriving from the reaction of orthophosphates with ammonium molybdate under acidic conditions, by stannous chloride solution.

HPLC Determination of Diclofenac
Prior to the analysis, samples from the influent and the effluent of the biofilter system obtained every 5 to 6 days were dissolved in methanol at concentrations of 1:100 and 1:5 v/v, respectively and passed through 0.45 µm membrane filters. A Shimadzu SCL-10AVP HPLC system (Kyoto, Japan), consisting of a Discovery HS C18 column (5 µm, 250 × 4.6 mm) (Supelco, Bellefonte, PA, USA), a LC-10 ADVP pump, a DGU-14A degassing system, a CTO-10ASVP column oven and a SPD-M10 AVP photodiode array detector was used for the chromatographic determination of diclofenac. The mobile phase consisted of a mixture of acetonitrile and 0.2% formic acid in water at mixing ratio of 60:40 v/v. The flow rate was 0.8 mL/min, the injection volume was 20 µL and the calibration standards or samples were analyzed by applying an isocratic elution program at 80 bar and 30 • C for a period of 20 min.
Photodiode array detection of diclofenac was performed at 200 nm and the limit of detection was 250 µg/L.

Amplicon Sequencing Analysis of the Diclofenac-degranding Microbiota in the Immobilized Cell Biofilter
Samples for amplicon sequencing analysis were obtained under steady state conditions. Genomic DNA was extracted in triplicates from the Siran beads containing immobilized biomass through the use of "Wizard Genomic DNA Purification Kit" (Promega, Madison, WI, USA), based on the instructions recommended by the manufacturer. In detail, the immobilized on Siran glass beads biomass were initially frozen with liquid nitrogen and pestled to dust. DNA extraction for bacterial amplicon sequencing was performed following the "Wizard Genomic DNA Purification Kit" protocol for bacteria, which includes additional steps for DNA extraction from Gram-positive bacteria, where 60 µL of 10 mg/mL lysozyme and 60 µL of 10 mg/mL lysostaphin were added to achieve cell lysis per each sample preparation. The "Wizard Genomic DNA Purification Kit" protocol for DNA extraction from yeasts, which included a lyticase treatment step by adding 7.5 µL of 75 U/µL lyticase per sample preparation, was employed. Amplification of the V1-V3 fragment of the 16S rRNA gene was performed by using the primer set 27F Purification of the amplified PCR products was performed by Ampure beads (USA). Illumina sequencing was carried out in a MiSeq apparatus and sequencing data was processed through demultiplexing and trimming of the amplicons [30]. Further processing excluded N(s)-containing, abnormal length size (<200 bp) or low-quality score reads (Q < 30) [31]. Using USEARCH v.11, the -fastq_filter was employed to improve the quality of the assembled sequences and the -fastx_uniques option was used to select the unique read sequences and their abundances [32]. Clustering of the selected read sequences into operational taxonomic units (OTUs) (number of read sequences ≥ 3) was performed through the use of the -cluster_otus command. Chimeric sequences were discarded by using the -unoise3 option of USEARCH v.11 [33]. Taxonomy was assigned to RDP (Ribosomal Database Project) via SPINGO (Species-level IdentificatioN of metaGenOmic amplicons) algorithm [34]. The 16S rRNA gene copies of the genomes of the bacterial taxa detected in the present study were estimated by using rrnDB version 5.5 [35]. AmpliTAXO within AmpliSAS was employed for clustering fungal reads [36]. The microbial community established in the immobilized cell bioreactor under steady state conditions was evaluated through three independent samplings. A total of 137,244 non-chimeric bacterial amplicons (three replicates of 45,764, 48,249 and 43,231 each) and a total of 184,850 non-chimeric fungal amplicons (three replicates of 62,247, 65,934 and 56,669 each) were sequenced and further deposited into the NCBI database under the BioProject number PRJNA530067.

Statistical Analysis
Relationships among variables were identified through calculation of Pearson's correlation coefficients using Past v.3.25 [37], whereas the Student's t-test was implemented to assess significant differences between influent and effluent physicochemical traits.

Operating Behavior and Effectiveness of the Immobilized Cell Biofilter to Degrade Diclofenac-Based Wastewater
The bioreactor system was fed with synthetic wastewater containing 400 mg/L of diclofenac in order to reveal the specialized members of the activated sludge, which were capable of breaking down this priority pollutant. Statistically significant differences were found between influent and effluent diclofenac concentrations (p < 0.01, in Student's t-test) throughout biofilter operation. A further decrease in effluent diclofenac concentration was observed from days 0-75 to 76-141 (p < 0.01, in Student's t-test), i.e., from 57.61 ± 6.21 to 7.87 ± 1.84 mg/L diclofenac. During start-up, an adaptation period of over 2 months was needed in order the microbial biomass to be effectively adapted to the new feeding conditions. Thereafter, diclofenac removal efficiency was highly improved, reaching up to 97.63 ± 0.62% (Figure 2), indicating the existence of an effective diclofenac-degrading population in the activated sludge.

Operating Behavior and Effectiveness of the Immobilized Cell Biofilter to Degrade Diclofenac-Based Wastewater
The bioreactor system was fed with synthetic wastewater containing 400 mg/L of diclofenac in order to reveal the specialized members of the activated sludge, which were capable of breaking down this priority pollutant. Statistically significant differences were found between influent and effluent diclofenac concentrations (p < 0.01, in Student's t-test) throughout biofilter operation. A further decrease in effluent diclofenac concentration was observed from days 0-75 to 76-141 (p < 0.01, in Student's t-test), i.e., from 57.61 ± 6.21 to 7.87 ± 1.84 mg/L diclofenac. During start-up, an adaptation period of over 2 months was needed in order the microbial biomass to be effectively adapted to the new feeding conditions. Thereafter, diclofenac removal efficiency was highly improved, reaching up to 97.63 ± 0.62% (Figure 2), indicating the existence of an effective diclofenac-degrading population in the activated sludge. Langenhoff et al. [19] used batch reactors containing biomass adapted to high diclofenac concentrations to treat wastewater containing 50 mg/L diclofenac and reported removal efficiencies of 65-70% in a 20 day-treatment period, which are much lower than that achieved in the immobilized cell biofilter operated at HRT of 3.5 days. In contrast to batch reactors, the continuous flow mode of operation employed in our study appears to restrict the accumulation of possible toxic metabolites, permitting the progressive development of an acclimatized biomass capable of effectively degrading high diclofenac concentrations at a shorter HRT. In addition, the immobilization of biomass also resulted in elevated biosystem efficiency, due to the fact that the sludge retention time (SRT) was continuously increasing since no biomass was wasted out. According to the literature, SRT values higher than 20 days are needed in order to reach removal efficiencies higher than 50-60% [10].
Regarding COD removal patterns, the immobilized cell biofilter was adapted in shorter period, resulting in a COD decrease from to 581 ± 6 mg/L in the influent to 125 ± 5 mg/L in the effluent (p < 0.01, in Student's t-test), whereas the respective effluent sCOD value was equal to 80.0 ± 4.6 mg/L ( Figure 3). This COD decline corresponded to COD removal efficiency of 78.5 ± 0.9 and 86.2 ± 0.8% in the case of total and soluble COD, respectively ( Figure 3). The significant correlation between diclofenac concentration decrease and COD removal (correlation coefficient r = 0.551, p < 0.01) points out that both diclofenac and COD removal profiles followed the same trend under steady state conditions.  Langenhoff et al. [19] used batch reactors containing biomass adapted to high diclofenac concentrations to treat wastewater containing 50 mg/L diclofenac and reported removal efficiencies of 65-70% in a 20 day-treatment period, which are much lower than that achieved in the immobilized cell biofilter operated at HRT of 3.5 days. In contrast to batch reactors, the continuous flow mode of operation employed in our study appears to restrict the accumulation of possible toxic metabolites, permitting the progressive development of an acclimatized biomass capable of effectively degrading high diclofenac concentrations at a shorter HRT. In addition, the immobilization of biomass also resulted in elevated biosystem efficiency, due to the fact that the sludge retention time (SRT) was continuously increasing since no biomass was wasted out. According to the literature, SRT values higher than 20 days are needed in order to reach removal efficiencies higher than 50-60% [10].
Regarding COD removal patterns, the immobilized cell biofilter was adapted in shorter period, resulting in a COD decrease from to 581 ± 6 mg/L in the influent to 125 ± 5 mg/L in the effluent (p < 0.01, in Student's t-test), whereas the respective effluent sCOD value was equal to 80.0 ± 4.6 mg/L ( Figure 3). This COD decline corresponded to COD removal efficiency of 78.5 ± 0.9 and 86.2 ± 0.8% in the case of total and soluble COD, respectively ( Figure 3). The significant correlation between diclofenac concentration decrease and COD removal (correlation coefficient r = 0.551, p < 0.01) points out that both diclofenac and COD removal profiles followed the same trend under steady state conditions. The pH in the immobilized cell biofilter was dropped drastically to acidic values (p < 0.01, in Student t-test) when diclofenac served as the sole carbon and energy source, resulting in a pH decrease from 7.16 ± 0.02 in the influent to 5.67 ± 0.04 in the effluent (Figure 4). Moreover, effluent pH significantly decreased from 5.82 ± 0.06 to 5.51 ± 0.05 from days 0-75 to 76-141 (p < 0.01, in Student's t-test). Diclofenac removal highly correlated with ΔpH (pHin-pHef) (r = 0.503, p < 0.01), indicating the influence of diclofenac biodegradation on pH drop. A similar observation was also addressed by Trapido et al. [38], who reported that diclofenac degradation during photolysis led to pH acidification, independent of the initial pH value of the diclofenac-containing wastewater. Furthermore, the electrical conductivity (EC) in the influent and the effluent of the biosystem were estimated to be 2.59 ± 0.04 and 2.47 ± 0.02 mS/cm, respectively (Figure 4), due to the fact that the growth medium was supplemented with trace elements. The marginal differences (p < 0.05, in Student's t-test), which were observed between influent and effluent EC values, could be attributed to the pooled values determined.
Based on the determination of nitrogenous compounds, it appears that the immobilized biomass could not oxidize ammonium to nitrate. A drastic pH drop to acidic values (below 6) has been reported to inhibit nitrifying activity [39]. Besides, differences in metabolic capabilities of  The pH in the immobilized cell biofilter was dropped drastically to acidic values (p < 0.01, in Student t-test) when diclofenac served as the sole carbon and energy source, resulting in a pH decrease from 7.16 ± 0.02 in the influent to 5.67 ± 0.04 in the effluent (Figure 4). Moreover, effluent pH significantly decreased from 5.82 ± 0.06 to 5.51 ± 0.05 from days 0-75 to 76-141 (p < 0.01, in Student's t-test). Diclofenac removal highly correlated with ∆pH (pHin-pHef) (r = 0.503, p < 0.01), indicating the influence of diclofenac biodegradation on pH drop.  The pH in the immobilized cell biofilter was dropped drastically to acidic values (p < 0.01, in Student t-test) when diclofenac served as the sole carbon and energy source, resulting in a pH decrease from 7.16 ± 0.02 in the influent to 5.67 ± 0.04 in the effluent ( Figure 4). Moreover, effluent pH significantly decreased from 5.82 ± 0.06 to 5.51 ± 0.05 from days 0-75 to 76-141 (p < 0.01, in Student's t-test). Diclofenac removal highly correlated with ΔpH (pHin-pHef) (r = 0.503, p < 0.01), indicating the influence of diclofenac biodegradation on pH drop. A similar observation was also addressed by Trapido et al. [38], who reported that diclofenac degradation during photolysis led to pH acidification, independent of the initial pH value of the diclofenac-containing wastewater. Furthermore, the electrical conductivity (EC) in the influent and the effluent of the biosystem were estimated to be 2.59 ± 0.04 and 2.47 ± 0.02 mS/cm, respectively (Figure 4), due to the fact that the growth medium was supplemented with trace elements. The marginal differences (p < 0.05, in Student's t-test), which were observed between influent and effluent EC values, could be attributed to the pooled values determined.
Based on the determination of nitrogenous compounds, it appears that the immobilized biomass could not oxidize ammonium to nitrate. A drastic pH drop to acidic values (below 6) has been reported to inhibit nitrifying activity [39]. Besides, differences in metabolic capabilities of  A similar observation was also addressed by Trapido et al. [38], who reported that diclofenac degradation during photolysis led to pH acidification, independent of the initial pH value of the diclofenac-containing wastewater. Furthermore, the electrical conductivity (EC) in the influent and the effluent of the biosystem were estimated to be 2.59 ± 0.04 and 2.47 ± 0.02 mS/cm, respectively (Figure 4), due to the fact that the growth medium was supplemented with trace elements. The marginal differences (p < 0.05, in Student's t-test), which were observed between influent and effluent EC values, could be attributed to the pooled values determined.
Based on the determination of nitrogenous compounds, it appears that the immobilized biomass could not oxidize ammonium to nitrate. A drastic pH drop to acidic values (below 6) has been reported to inhibit nitrifying activity [39]. Besides, differences in metabolic capabilities of specific microbial communities, due to the selection pressure applied by the high diclofenac concentration, cannot be excluded. However, organic nitrogen appeared to be mineralized as proven by the determined TKN and NH 4 + -N values in the influent and the effluent of the biosystem (Table 1), since statistically significant differences were identified between influent and effluent TKN (p < 0.01, in the Student's t-test). This provides evidence for almost complete mineralization of diclofenac, since the N content of this NSAID was ammonified and its COD was greatly reduced. Even though adequate aeration was provided (4.32 ± 0.22 mg/L) since Langenhoff et al. [19] showed that diclofenac degradation was sufficient only under aerobic rather than anoxic conditions, low levels of nitrate and nitrite nitrogen (1.81 ± 0.85 and 0.21 ± 0.12 mg/L, respectively) were detected as the result of the restricted nitrification process. Indeed, the N content of diclofenac-based wastewater used was mainly assimilated to the microbial cells rather than nitrified (Table 1). Moreover, sufficient phosphorus concentration was provided in the immobilized cells, determining residual phosphorus concentration in the effluent of 7.07 ± 1.75 mg/L ( Table 1). As the result of phosphorus assimilation in the biomass, statistically significant differences were found between the influent and effluent PO 4 3− -P concentration (p < 0.05, in Student's t-test). In addition, the BOD in the effluent of the immobilized cell bioreactor was as low as 8 mg/L (Table 1).

Bacterial Community Structure in the Immobilized Cell Biofilter Fed with Diclofenac-Based Wastewater
Illumina data revealed that Acidobacteria, Alphaproteobacteria, Gammaproteobacteria and Betaproteobacteria were the predominant higher bacterial taxa (95.23 ± 1.24% of bacterial reads in total), showing almost equal abundances, i.e., 26.49 ± 0.45%, 25.39 ± 0.40%, 22.10 ± 1.29% and 21.25 ± 1.19% of the total reads, respectively. Actinobacteria also accounted for the 3.05 ± 0.45% of the total reads. In addition, amplicon sequencing analysis at genus level showed that the microbial community in the immobilized cell bioreactor fed with diclofenac as the sole carbon and energy source was dominated by bacteria belonging to the genera Granulicella and Rhodanobacter, which accounted for 20.75 ± 0.56% and 20.56 ± 1.59% of the total reads, respectively (Figure 5a). In particular, the predominant taxa were associated with the species Granulicella pectinivorans and Rhodanobacter terrae (Figure 5b). Taken into consideration the 16S rRNA gene copies of the genomes of Granulicella and Rhodanobacter, the G. pectinivorans population was double that of R. terrae ( Figure 5). Little is known for Granulicella spp., although, as all members of the family Acidobacteriaceae, they are slow growers that are favored by acidic conditions [40] and are capable of utilizing a wide range of carbon sources, even if they are originated from oligotrophic habitants [41]. Rhodanobacter is a common habitant of activated sludge systems [42] and has been reported to be an effective degrader of recalcitrant organic compounds, including volatile organics (VOCs) [43] and other NSAIDs like ibuprofen [44]. Moreover, Rhodanobacter is an acid-tolerant denitrifier, which proliferates under acidic conditions [45]. The drop in the effluent pH can be attributed to the predominance of Acidobacteria and Rhodanobacter in the adapted to high diclofenac concentration biomass, which are common inhabitants of acidic environments, due to their ability to degrade recalcitrant compounds [46], as well as to the fact that bacteria can adjust environmental pH to their optimum growth [47]. Indeed, the ability of these taxa to utilize a wide range of carbon sources, including recalcitrant organic compounds, and to be effectively adapted under acidic conditions explain their prevalence in the immobilized cell biofilter.
Processes 2019, 7, x FOR PEER REVIEW 8 of 13 common inhabitants of acidic environments, due to their ability to degrade recalcitrant compounds [46], as well as to the fact that bacteria can adjust environmental pH to their optimum growth [47]. Indeed, the ability of these taxa to utilize a wide range of carbon sources, including recalcitrant organic compounds, and to be effectively adapted under acidic conditions explain their prevalence in the immobilized cell biofilter.
Beyond the two most dominant bacterial taxa mentioned above, members of the genera Castellaniella, Parvibaculum, Bordetella and Bryocella, showing relative abundance (as a % of total bacterial reads) of 12.32 ± 0.20%, 12.19 ± 0.36%, 7.71 ± 0.97% and 5.67 ± 0.38%, respectively, were detected in the bacterial community established in the immobilized cell biofilter (Figure 5a). Representatives of Rhodopseudomonas, Mycobacterium, Gluconacetobacter, Acidisphaera, Chthoniobacter and Filomicrobium also corresponded to 12.01 ± 1.25% of the total bacterial reads (Figure 5a). To the best of our knowledge, the endophyte Microbacterium sp. MG7 [48], a few soil bacterial strains belonging to Raoultella-Klebsiella group [49,50] and Brevibacterium sp. D4 [25] isolated from activated sludge are the only bacteria that have been found to be capable of degrading diclofenac. In addition, Labrys portucalensis, which was detected in the present study (Figure 5b), has been reported to be capable of completely degrading diclofenac co-metabolically in the presence of acetate [26]. However, no study regarding the diversity of the diclofenac-degrading microbiota, which utilize diclofenac as the sole carbon and energy source, has been performed until now. Moreover, the Alcaligenaceae representatives identified in the immobilized biomass, like Castellaniella denitrificans and Bordetella petrii, which covered 20.03 ± 1.04% of the total reads (Figure 5b), are considered as effective degraders of various recalcitrant compounds [51,52].
Apart from various white-rot fungi like Trametes versicolor, which are considered as effective degraders of endocrine disrupting substances [53,54] and a few other allochnonous fungi [55], no information on fungal microbiota from activated sludge that are capable of degrading diclofenac exists. At phylum/subphylum level, the fungal diversity in the immobilized cell biofilter fed with diclofenac as the sole carbon and energy source consisted almost exclusively of members of Saccharomycotina (Ascomycota), representing 98.54 ± 0.08% of the total fungal reads, followed by members of Pezizomycotina, and Ascomycota (1.45 ± 0.07%) (Figure 6a). Thus, Ascomycota accounted for 99.99% of the fungal total relative abundance, whereas the remaining 0.01% consisted of member of the phylum Basidiomycota (subdivisions Pucciniomycotina, Agaricomycotina and Ustilaginomycotina). At the genus level, Wickerhamiella was found to be the predominant fungal taxon in the immobilized cell biofilter with relative abundance of 78.56 ± 1.35%, followed by Yarrowia (19.96 ± 1.28%), Paecilomyces (0.94 ± 0.09%) and Exophiala (0.40 ± 0.05%) spp. (Figure 6b). Representatives of Rhodopseudomonas, Mycobacterium, Gluconacetobacter, Acidisphaera, Chthoniobacter and Filomicrobium also corresponded to 12.01 ± 1.25% of the total bacterial reads (Figure 5a). To the best of our knowledge, the endophyte Microbacterium sp. MG7 [48], a few soil bacterial strains belonging to Raoultella-Klebsiella group [49,50] and Brevibacterium sp. D4 [25] isolated from activated sludge are the only bacteria that have been found to be capable of degrading diclofenac. In addition, Labrys portucalensis, which was detected in the present study (Figure 5b), has been reported to be capable of completely degrading diclofenac co-metabolically in the presence of acetate [26]. However, no study regarding the diversity of the diclofenac-degrading microbiota, which utilize diclofenac as the sole carbon and energy source, has been performed until now. Moreover, the Alcaligenaceae representatives identified in the immobilized biomass, like Castellaniella denitrificans and Bordetella petrii, which covered 20.03 ± 1.04% of the total reads (Figure 5b), are considered as effective degraders of various recalcitrant compounds [51,52]. Apart from various white-rot fungi like Trametes versicolor, which are considered as effective degraders of endocrine disrupting substances [53,54] and a few other allochnonous fungi [55], no information on fungal microbiota from activated sludge that are capable of degrading diclofenac exists. At phylum/subphylum level, the fungal diversity in the immobilized cell biofilter fed with diclofenac as the sole carbon and energy source consisted almost exclusively of members of Saccharomycotina (Ascomycota), representing 98.54 ± 0.08% of the total fungal reads, followed by members of Pezizomycotina, and Ascomycota (1.45 ± 0.07%) (Figure 6a). Thus, Ascomycota accounted for 99.99% of the fungal total relative abundance, whereas the remaining 0.01% consisted of member of the phylum Basidiomycota (subdivisions Pucciniomycotina, Agaricomycotina and Ustilaginomycotina). At the genus level, Wickerhamiella was found to be the predominant fungal taxon in the immobilized cell biofilter with relative abundance of 78.56 ± 1.35%, followed by Yarrowia (19.96 ± 1.28%), Paecilomyces (0.94 ± 0.09%) and Exophiala (0.40 ± 0.05%) spp. (Figure 6b).  Wickerhamiella that was proliferated in the immobilized cell biofilter, where diclofenac served as the sole carbon source, was recently found as the dominant fungal taxon in 18 full-scale municipal WWTPs [56]. Moreover, Yarrowia has been recently reported to possess the capability of partially degrading diclofenac [55], while this is the first report on the potential involvement of Paecilomyces and Exophiala spp. on diclofenac degradation.
It is concluded that regarding the efficiency of the immobilized cell biofilter to degrade diclofenac as the sole carbon and energy source, an adaptation period of over two months under prolonged HRT (3.5 days) was required to achieve removal efficiencies as high as 97.63 ± 0.62%. Afterwards, an effective diclofenac-degrading population was enriched from the activated sludge that was used as the inoculum, where both diclofenac and COD removal patterns followed the same trend under steady state conditions. Utilization of diclofenac as the sole carbon and energy source in the immobilized cell biofilter also led to a drastic pH decrease. Evidence is also provided for almost complete mineralization of diclofenac, due to the ammonification of N content of this pollutant and the high COD removal efficiency achieved. Based on amplicon sequencing, the major bacterial taxa identified in the immobilized cell biofilter fed with diclofenac as the sole carbon and energy source were members of the genera Granulicella and Rhodanobacter. The ability of members of these genera to utilize a wide range of carbon sources and to be effectively adapted under acidic conditions were the main factors favoring their prevalence in the immobilized cell biofilter. Wickerhamiella was found to be the predominant fungal taxon in the immobilized cell biofilter, followed by Yarrowia, Paecilomyces and Exophiala. Wickerhamiella is a common fungal taxon in WWTPs, which was also favored in the immobilized cell biofilter when diclofenac served as the sole carbon source.