Estrogen Degraders and Estrogen Degradation Pathway Identified in an Activated Sludge

ABSTRACT

The environmental release and fate of estrogens are becoming an increasing public concern. Bacterial degradation has been considered the main process for eliminating estrogens from wastewater treatment plants. Various bacterial isolates are reportedly capable of aerobic estrogen degradation, and several estrogen degradation pathways have been proposed in proteobacteria and actinobacteria. However, the ecophysiological relevance of estrogen-degrading bacteria in the environment is unclear. In this study, we investigated the estrogen degradation pathway and corresponding degraders in activated sludge collected from the Dihua Sewage Treatment Plant, Taipei, Taiwan. Cultivation-dependent and cultivation-independent methods were used to assess estrogen biodegradation in the collected activated sludge. Estrogen metabolite profile analysis revealed the production of pyridinestrone acid and two A/B-ring cleavage products in activated sludge incubated with estrone (1 mM), which are characteristic of the 4,5-seco pathway. PCR-based functional assays detected sequences closely related to alphaproteobacterial oecC, a key gene of the 4,5-seco pathway. Metagenomic analysis suggested that Novosphingobium spp. are major estrogen degraders in estrone-amended activated sludge. Novosphingobium sp. strain SLCC, an estrone-degrading alphaproteobacterium, was isolated from the examined activated sludge. The general physiology and metabolism of this strain were characterized. Pyridinestrone acid and the A/B-ring cleavage products were detected in estrone-grown strain SLCC cultures. The production of pyridinestrone acid was also observed during the aerobic incubation of strain SLCC with 3.7 nM (1 μg/liter) estrone. This concentration is close to that detected in many natural and engineered aquatic ecosystems. The presented data suggest the ecophysiological relevance of Novosphingobium spp. in activated sludge.

IMPORTANCE Estrogens, which persistently contaminate surface water worldwide, have been classified as endocrine disruptors and human carcinogens. We contribute new knowledge on the major estrogen biodegradation pathway and estrogen degraders in wastewater treatment plants. This study considerably advances the understanding of environmental estrogen biodegradation, which is instrumental for the efficient elimination of these hazardous pollutants. Moreover, this study substantially improves the understanding of microbial estrogen degradation in the environment.

INTRODUCTION

Natural estrogens (estrone [E1], 17β-estradiol [E2], and estriol [E3]) are exclusively produced and excreted by animals. Primary sources of estrogen include human urine and livestock manure (1). Because of the increase in human population and increased demands for livestock products, estrogen pollution has become a global concern and challenge (2, 3). Lange et al. (4) reported that the annual estrogen excretion of farm animals in the European Union and the United States can be up to 33 and 49 metric tons, respectively. Estrogens are frequently detected in the environment (1, 5–7). As an example, surface water in Chinese, American, and Dutch rivers has been reported to contain estrogens in a nanogram per liter range (8–11). Concentrations of estrogens in Wulo Creek, which is near a concentrated livestock feedlot in southern Taiwan, were as high as 1,267 ng/liter (for E1), 313.6 ng/liter (for E2), and 210 ng/liter (for E3) (12).

The occurrence and persistence of estrogens in surface water ecosystems have become a major concern in environmental research and policy, because long-term exposure to these compounds, even at extremely low concentrations, may adversely affect animal physiology and behavior (7, 13, 14). Young et al. (15) proposed predicted-no-effect-concentration values for E2 (1 ng/liter) and E1 (3 ng/liter). In Europe, intersex fish were observed near sewage treatment plants (16). Widespread occurrence of intersex black basses (Micropterus spp.) in the United States was also documented (17). Furthermore, Lambert et al. (18) reported that offspring sex ratios in the wild frog populations in the United States are affected by estrogens released into suburban ponds. In addition to being endocrine disruptors, estrogens have been classified by the World Health Organization as group 1 carcinogens (http://monographs.iarc.fr/ENG/Classification/latest_classif.php).

In developed countries, wastewater treatment plants are crucial for removing steroid hormones produced by humans and livestock (6, 19). Numerous studies have reported the essential role of bacterial degradation in removing estrogens from engineered ecosystems (20–22). Estrogen-degrading bacteria belonging to the phyla Proteobacteria, Actinobacteria, Bacteroidetes, and Firmicutes (see reference 23 and references therein) have been isolated from various engineered ecosystems (e.g., activated sludge and compost) and natural environments (e.g., soils, sandy aquifers, and the Baltic Sea). Nonetheless, the literature on estrogen degradation pathways is limited. Thus far, at least four different degradation pathways for natural estrogens have been proposed (Fig. 1). In total, nine E2-derived degradation metabolites have been identified through mass spectrometry (MS) (24–29). Six of the nine compounds are metabolites of estrogen degradation pathway I (i.e., the 4,5-seco pathway). The other MS-identified metabolites, estratetraenol, compound IIIc, and compound IVd, are proposed to be involved in degradation pathways II, III, and IV, respectively. The calculated molecular masses for the pseudomolecular ions ([M+H]⁺ and [M+Na]⁺) of these degradation metabolites are shown in Fig. 1. In our recent study (24), we used Sphingomonas sp. strain KC8, a wastewater isolate (25), as the model organism for aerobic estrogen degradation, and we identified 4-hydroxyestrone, a meta-cleavage product, and pyridinestrone acid (also known as pyridinestrone-3-carboxylic acid; see Fig. 1 for the structures of these molecules) as likely degradation intermediates. Pyridinestrone acid was produced from the meta-cleavage product through an abiotic reaction with ammonium; therefore, pyridinestrone acid could be a dead-end product in the 4,5-seco pathway. We also sequenced the genome of strain KC8 and identified gene clusters likely to be involved in the degradation of A/B rings (gene clusters I and II) and C/D rings (gene cluster III) of E2 (24). Based on the results of the genomic and transcriptomic analyses as well as heterologous overexpression, we identified catabolic genes oecA and oecB, coding for 17β-estradiol dehydrogenase and estrone 4-hydroxylase, respectively. Furthermore, the meta-cleavage enzyme 4-hydroxyestrone 4,5-dioxygenase (OecC), encoded by gene cluster II, was purified and characterized from E1-grown cells of strain KC8. These studies have enabled us to propose the initial steps of the aerobic estrogen degradation pathway (i.e., the 4,5-seco pathway). Nevertheless, the downstream metabolites and the corresponding enzymes of this aerobic pathway remain to be uncovered. Alternative estrogen degradation pathways and metabolites have been proposed, but biochemical evidence for these pathways has not been reported (26–29).

FIG 1.

Proposed bacterial degradation pathways of natural estrogens. NMR, nuclear magnetic resonance.

The biochemical mechanisms underlying aerobic estrogen biodegradation have been largely studied in pure cultures (24–28). The question as to which estrogen biodegradation pathway(s) play a major role in the environment has yet to be resolved. Moreover, the distribution and abundance of estrogen-degrading bacteria in the environment have yet to be investigated. In the present study, we examined microbial estrogen degradation in the activated sludge of the Dihua Sewage Treatment Plant (DHSTP), which treats domestic wastewater produced by the 3 million residents of Taipei City, Taiwan. To enrich E1-degrading microbial populations, the activated sludge samples were incubated with 1 mM E1. To investigate the biochemical mechanisms and microorganisms involved in aerobic E1 degradation in the activated sludge, we used the following approaches: (i) identification of estrogen metabolites through ultraperformance liquid chromatography–electrospray ionization–high-resolution mass spectrometry (UPLC-ESI-HRMS), (ii) phylogenetic identification of E1-degrading bacteria through Illumina MiSeq sequencing, (iii) detection of the characteristic catabolic gene oecC using PCR, and (iv) isolation and characterization of a major estrogen degrader from the DHSTP sludge sample.

RESULTS

The DHSTP activated sludge samples collected in August 2014 and January 2015 were incubated under the following conditions: activated sludge with dimethyl sulfoxide (DMSO) alone, activated sludge with E1, and autoclaved sludge with E1. The incubation experiments for sludge collected in January 2015 were performed in duplicate. Estrogen metabolites, the bacterial 16S rRNA gene, and the estrogen catabolic gene oecC in the subsamples were analyzed using UPLC-ESI-HRMS, the Illumina MiSeq platform, and PCR-based functional assays, respectively.

UPLC-ESI-HRMS identification of estrogen metabolites in activated sludge.

Estrogen transformation was not observed in the autoclaved sludge incubated with 1 mM E1. Similarly, estrogen metabolites were not detected in activated DHSTP sludge incubated with DMSO alone. In contrast, the consumption of the substrate, as well as the production of three estrogen degradation metabolites, namely, pyridinestrone acid [extracted ion chromatogram (EIC) at m/z 300.16 ([M+H]⁺)], compound Id [EIC at m/z 331.15 ([M+Na]⁺)], and compound Ie [EIC at m/z 303.16 ([M+Na]⁺)], was observed during aerobic incubation of activated sludge with E1 (Fig. 2; see also Fig. S1 in the supplemental material). The ESI-HRMS spectra, as well as the UPLC retention time of E1 (8.10 min) and pyridinestrone acid (3.95 min), were comparable to those of the authentic standards. Authentic standards for compounds Id and Ie are currently unavailable. However, the fragmentation ions ([M-3H₂O+H]⁺, [M-2H₂O+H]⁺, and [M-H₂O+H]⁺) and pseudomolecular ions ([M+H]⁺ and [M+Na]⁺) of compounds Id and Ie could be identified in their ESI-HRMS spectra (Fig. S2). 4-Hydroxyestrone, estratetraenol, the ring cleavage product, compound IIIc, and compound IVd were not detected in the sludge treatments through the corresponding EICs.

FIG 2.

UPLC-ESI-HRMS analysis of ethyl acetate extracts of DHSTP sludge treatment samples. Activated sludge was collected in August 2014. See Fig. S1 for the production of the estrogen metabolites pyridinestrone acid, compound Id, and compound Ie in E1-amended activated sludge collected in January 2015.

Phylogenetic identification of estrogen-degrading bacteria in activated sludge.

DNA was extracted from the subsamples, which were withdrawn from the activated sludge treatments. The V3–V4 hypervariable region of the bacterial 16S rRNA gene was amplified through PCR, and the resulting amplicons were sequenced using an Illumina MiSeq sequencer (Illumina, San Diego, CA, USA). The sequences were analyzed using two pipelines: USEARCH (version 10.0.240) and mothur (version 1.35.1). Except for unassigned and “other” (individual classes with a relative abundance of <1%) bacteria, 29 classes were identified overall in the DHSTP activated sludge. The initial (0-day incubation) bacterial community structures in all of the treatments (amended with DMSO alone [solvent control] and with 1 mM E1 [dissolved in DMSO]) were highly similar, and the abundant bacterial classes were Alphaproteobacteria (4.9% to 5.9%), Betaproteobacteria (11.4% to 25.2%), Caldilineae (1.9% to 3.9%), Deltaproteobacteria (8.9% to 12.4%), and Sphingobacteriia (7.6% to 10.3%; Fig. 3A). Members of Betaproteobacteria were the most abundant bacteria in the initial sludge samples; however, betaproteobacteria were not enriched in all treatments. In the activated sludge incubated with DMSO alone (no exogenous E1), we did not observe remarkable temporal changes in the bacterial community structures. In contrast, in the presence of 1 mM E1, alphaproteobacteria were apparently enriched (increased from 10.4% [Fig. 3A, upper-middle panel] to 23.7% [Fig. 3A, upper-left panel] abundance after 6 days). Furthermore, we observed a considerable increase in the relative abundance of Novosphingobium spp. (class Alphaproteobacteria) in activated sludge incubated with E1 (increased to 5.7% to 17.1% abundance after 6 days; Fig. 3B). The enrichment of Novosphingobium spp. did not occur in the activated sludge incubated with DMSO alone. The results derived from further analyses (e.g., the phylogenetic analysis and relative abundance of individual operational taxonomic units [OTUs]) of the bacterial community structures are presented in Data Set S1.

FIG 3.

(A) Class-level phylogenetic analysis (Illumina MiSeq) revealed temporal changes in the bacterial community structures in various sludge treatment samples. (B) Novosphingobium spp. were apparently enriched in DHSTP sludge treatments amended with E1. d, days.

PCR amplification of oecC-like genes in activated sludge.

Degenerate gene-specific primers (Fig. S3C) were used to detect oecC in activated sludge samples collected from the DHSTP. The catabolic gene oecC encoding 4-hydroxyestrone 4,5-dioxygenase was identified in a variety of estrogen-degrading alphaproteobacteria, including strain KC8, Altererythrobacter estronivorus strain MH-B5, and Novosphingobium tardaugens NBRC 16725 (24). The designed primers were validated using chromosomal DNA of the three estrogen-degrading aerobes. To analyze primer specificity, DNA from androgen-degrading bacteria (Comamonas testosteroni ATCC 11996 and Cupriavidus taiwanensis DSM 17343) or cholesterol-degrading bacteria (Gordonia cholesterolivorans DSM 45229 and Mycobacterium smegmatis DSM 43277), which cannot degrade estrogens, was used as the negative control in these assays. PCR products with an expected size of approximately 830 bp were amplified from estrogen-degrading aerobes but not from androgen- or cholesterol-degrading bacteria (Fig. S3D). This result indicates that the degenerate primers are highly specific to oecC and cannot be used to amplify the extradiol dioxygenase genes tesB and hsaC, which are involved in aerobic degradation of androgens and cholesterol, respectively (30). Subsequently, DNA extracted from the E1-amended activated sludge treatment (collected in January 2015, replicate 1) was used as the template. PCR products amplified from the initial sludge sample (0-day incubation) and E1-enriched sludge sample (4-day incubation) were cloned in Escherichia coli. In total, 20 clones (initial clones 1 to 20) and 10 clones (E1-amended clones 1 to 10) were randomly selected from the initial and the E1-enriched sludge samples, respectively, for sequencing. The individual sequences of the amplified oecC fragments are shown in Table S1. It is worth mentioning that all of the oecC sequences obtained from the E1-enriched sludge sample were highly similar (60.7% to 72.0% amino acid sequence identity) to that of N. tardaugens NBRC 16725 (Fig. 4). Phylogenetic analysis of deduced amino acid sequences of the oecC fragments demonstrated high similarity (>50.6% amino acid sequence identity) between the environmental sequences and oecC sequences from the estrogen-degrading alphaproteobacteria. Moreover, these oecC sequences can be divided into four clades, and the DHSTP sequences are affiliated with clades I and III (Fig. 4). Our data thus indicated that oecC-like sequences amplified from the DHSTP activated sludge samples were clustered with oecC of the estrogen-degrading alphaproteobacteria and were separated from other extradiol dioxygenase genes, such as bphC, hsaC, and tesB.

FIG 4.

Phylogenetic tree of deduced amino acid sequences of oecC fragments obtained from DHSTP activated sludge. Initial and E1-amended sequences were amplified from the sludge treatment (January 2015, replicate 1) incubated with E1 for 0 and 4 days, respectively. Degenerate primers (see Fig. S3C for sequences) derived from proteobacterial oecC were used to amplify oecC fragments (see Table S1 for individual sequences) from sludge samples. Information regarding the extradiol dioxygenases is provided in Table S4. Alt., Altererythrobacter; Nov., Novosphingobium; Spm., Sphingomonas; Com., Comamonas; Myc., Mycobacterium; Rho., Rhodococcus; Pse., Pseudomonas; Bur., Burkholderia; Spb., Sphingobium; Geo., Geobacillus.

Isolation and characterization of estrogen-degrading Novosphingobium sp. strain SLCC.

Novosphingobium sp. strain SLCC was able to grow on a chemically defined medium solidified with agar. Strain SLCC formed circular convex milk-white colonies with smooth edges that were approximately 1.5 mm in diameter. The oecC (SLCC_oecC) and 16S rRNA (SLCC_16S rRNA) sequences of strain SLCC are presented in Table S2. The oecC sequence of strain SLCC showed 77.3% amino acid sequence identity to that of N. tardaugens NBRC 16725 (Fig. 4). We performed a quantitative PCR study to examine the temporal changes in the 16S rRNA gene of strain SLCC in different sludge treatments. The abundance of the strain SLCC gene in each sample was normalized by the total eubacterial 16S rRNA gene. An apparent increase in the 16S rRNA gene of strain SLCC was observed only in the E1-amended sludge treatments (Fig. 5). The relative abundance of the strain SLCC 16S rRNA gene increased (up to 6%) after 4 days of aerobic incubation. The real-time quantitative PCR results were consistent with those of the bacterial community analysis conducted using the Illumina MiSeq platform (Fig. 3), suggesting that strain SLCC is a major estrogen degrader in activated sludge.

FIG 5.

Real-time quantitative PCR indicated temporal changes in the 16S rRNA gene copies of the Novosphingobium sp. strain SLCC in E1-amended sludge treatments. Relative abundance of the strain SLCC 16S rRNA gene was calculated as a proportion of the total number of bacterial 16S rRNA gene copies. SLCC and Eub primer pairs were used to amplify the 16S rRNA gene of strain SLCC and total eubacterial population, respectively. Data are shown as the mean ± standard deviation (SD) of three experimental measurements. Standard curves of real-time quantitative PCR are shown in Fig. S4.

The results obtained from the physiological and metabolic characterization of strain SLCC are shown in Table S3. Strain SLCC showed optimal growth at pH 7.0 and 28°C and could aerobically grow with androst-4-en-3,17-dione, cholic acid, E1, E2, progesterone, or testosterone in minimal medium. However, it could not use androsta-1,4-diene-3,17-dione, cholesterol, or 17α-ethinylestradiol (EE2) as a sole source of carbon and energy. Strain SLCC was catalase and oxidase positive and exhibited β-glucosidase activity. This bacterial strain was not able to assimilate the tested carbohydrates. The major fatty acids of E2-grown cells are C_14:0 2-OH (6.5%), C_16:0 (8.6%), C_18:1 ω7c (48.1%), and 11-methyl C_18:1 ω7c (20.2%).

Estrogen metabolites involved in the aerobic E1 degradation by strain SLCC.

To identify the estrogen metabolites produced by strain SLCC, we first incubated strain SLCC cells with 1 mM E1 in a chemically defined medium. During the aerobic growth of these cells on E1, we detected the production of at least three intermediates, namely, pyridinestrone acid, compound Id, and compound Ie (Fig. S5A). We then incubated strain SLCC cells with a low concentration (1 μg/liter) of [3,4C-¹³C]E1 in a minimal medium. After 24 h of aerobic incubation, we detected the production of ¹³C-labeled pyridinestrone acid in the bacterial cultures (Fig. 6 and S5B). Pyridinestrone acid was not detected in the autoclaved culture incubated with [3,4C-¹³C]E1 or in a live culture incubated without [3,4C-¹³C]E1. Moreover, we detected pyridinestrone acid in the culture supernatant but not in the cell pellet. Our data thus indicate that this compound is excreted into extracellular environments after the production.

FIG 6.

UPLC-ESI-HRMS detection of ¹³C-labeled pyridinestrone acid in strain SLCC cultures incubated with [3,4C-¹³C]E1 (1 μg/liter). The [3,4C-¹³C]E1-amended treatment was performed in triplicate, and the UPLC-HRMS analysis results of the other two replicates are shown in Fig. S5.

DISCUSSION

The 4,5-seco pathway was the only detectable estrogen degradation pathway in E1-amended activated sludge.

In a recent study (24), we detected pyridinestrone acid, a characteristic metabolite of the 4,5-seco pathway, in E1-amended activated sludge as well as river water. However, estrogen metabolites for other degradation pathways in the environmental samples were not identified. In the present study, we applied UPLC-ESI-HRMS to detect estrogen metabolites that have been previously identified by MS (24–29). A highlight of this study is the detection of downstream metabolites, specifically the A/B-ring cleavage products (compounds Id and Ie), of the 4,5-seco pathway in the sludge samples. Thus far, biochemical mechanisms underlying the biotransformation of the A-ring cleavage product to compounds Id and Ie are largely unknown. One cannot exclude the possibility that these carboxylic compounds are dead-end products in this degradation pathway. The chemical compositions of compounds Id and Ie are C₁₇H₂₄O₅ and C₁₆H₂₄O₄, respectively, suggesting that the production of these two compounds from the A-ring cleavage product, a C₁₈ compound, might proceed through serial decarboxylation reactions. Nonetheless, the catabolic genes and enzymes involved in the bacterial production of compounds Id and Ie remain to be identified. The A/B-ring cleavage products are characteristic of the 4,5-seco pathway. Therefore, these metabolites and corresponding genes may be used as molecular markers for environmental studies of estrogen biodegradation.

In an earlier study, Lee and Liu (29) detected a lactone product (compound IIIc, Fig. 1) in the supernatant of activated sludge amended with E2 (up to 20 mg/liter). In the present study, we did not detect peaks corresponding to the pseudomolecular ions ([M+H]⁺ and [M+Na]⁺) of the proposed lactone product. We also did not detect other metabolites reported for estrogen degradation pathways II, III, and IV, suggesting that bacteria in the DHSTP sludge adopt the 4,5-seco pathway to degrade natural estrogens.

Novosphingobium spp. are the major degraders in E1-amended activated sludge.

Thayanukul et al. (31) applied microautoradiography-fluorescence in situ hybridization (MAR-FISH) to investigate potential estrogen degraders in activated sludge, and their results indicated that at submicrogram per liter concentrations, most [³H]E1-incorporating cells belong to Alphaproteobacteria. However, MAR-FISH provided little information on the metabolic fates of the tested substrates. Here, our current data, including estrogen metabolite profile analyses of the bacterial cultures and sludge samples, PCR-based functional assays, and the analysis of bacterial community structures suggest that Novosphingobium spp. are major estrogen degraders in activated sludge, which typically contains natural estrogens on a nanomolar scale (6, 32, 33). However, the degenerate primers were exclusively derived from alphaproteobacterial sequences; PCR-based functional assays cannot exclude the involvement of sludge bacteria belonging to other bacterial classes in estrogen degradation.

Thus far, all estrogen degradation pathways were described only in bacteria (24–29). Thus, we studied the changes in the bacterial community in the sludge treatments. The community structure analysis suggested that gammaproteobacteria and actinobacteria are not crucial estrogen degraders in the examined activated sludge. Furthermore, the Illumina MiSeq analysis revealed an apparent dominance of Novosphingobium spp. in the E1-amended DHSTP sludge. Thus far, only a few Novosphingobium strains are reportedly capable of aerobic estrogen degradation. Estrogen-degrading Novosphingobium spp. isolated from activated sludge included the Novosphingobium sp. strain JEM-1 (34) and the N. tardaugens strain ARI-1 (NBRC 16725) (35). Strain ARI-1 can degrade natural estrogens, including E1, E2, and E3, but not the synthetic estrogen EE2. In contrast, the JEM-1 strain can degrade E1 and E2, as well as EE2, on a millimolar scale. In the present study, we isolated estrogen-degrading strain SLCC, which can grow with various steroids, including natural estrogens, androgens, progesterone, and cholic acid. The estrogen metabolite profile analysis results of strain SLCC cultures and sludge extracts were comparable, suggesting that strain SLCC and related Novosphingobium spp. are responsible for the production of pyridinestrone and the A/B-ring cleavage products in the DHSTP activated sludge. Moreover, this strain produced pyridinestrone acid in a minimal medium containing 1 μg/liter (3.7 nM) of E1. This concentration was very close to the environmental levels. We provide strong evidence demonstrating that a bacterial isolate can degrade estrogens at environmentally relevant concentrations.

Our sequence data may enable the design of oligonucleotide probes for microarray analysis. The GeoChip, a comprehensive DNA microarray targeting microbial metabolism genes (36, 37), has been extensively applied to monitor microbial functionality in environmental samples, including biodegradation pathways in wastewater treatment plants (38, 39). Systems biology approaches, such as metatranscriptomics or metaproteomics coupled with metagenomics, can be applied in in situ estrogen biodegradation. This may facilitate the in situ identification of estrogen degraders and degradation genes as well as the investigation of microbial interactions during estrogen degradation in environmental samples.

MATERIALS AND METHODS

Chemicals and bacterial strains.

[3,4C-¹³C]E1 (99%) was purchased from Cambridge Isotope Laboratories. The other chemicals were of analytical grade and were purchased from Fluka, Mallinckrodt Baker, Merck, and Sigma-Aldrich. Comamonas testosteroni ATCC 11996 was obtained from the American Type Culture Collection (Manassas, VA, USA). Cupriavidus taiwanensis DSM 17343, Gordonia cholesterolivorans DSM 45229, Mycobacterium smegmatis DSM 43277, and Novosphingobium tardaugens NBRC 16725 (= DSM 16702) were obtained from the Deutsche Sammlung für Mikroorganismen und Zellkulturen (Braunschweig, Germany). These aerobic bacteria were aerobically cultured in Luria-Bertani medium.

Collection of activated sludge samples from the DHSTP.

The DHSTP is the largest municipal wastewater treatment plant (500,000 m³/day) in Taipei, Taiwan. In addition to domestic water, the DHSTP receives industrial, medical, and livestock wastewater, as well as groundwater (33). The hydraulic retention time is approximately 10 h, and the effluent is discharged into the Tamsui River, the largest river in Taipei and its suburbs (40). The plant is designed as an anoxic (denitrifying)/aerobic process for nitrogen and carbon removal (see references 33 and 40) for detailed information. Estrogens in influents are efficiently removed through a biodegradation process in the DHSTP (33). Activated sludge samples (10 liters) were collected from the aerobic tank of the DHSTP in August 2014 and January 2015. The activated sludge samples were placed in sterilized 20-liter glass bottles and delivered to the laboratory within 30 min.

Aerobic incubation of activated sludge with E1.

The DHSTP activated sludge samples (0.5 liters of sludge in 2-liter glass bottles) were incubated under the following conditions: activated sludge with DMSO (0.8% [vol/vol]; the solvent) alone, activated sludge with E1 (dissolved in DMSO), and autoclaved sludge with E1 (1 mM). The bottles were incubated at 25°C with stirring at 160 rpm for 2 weeks. Samples (10 ml) were withdrawn from the bottles every 2 days, and these subsamples were stored at −80°C before use. Estrogen metabolites in the subsamples were detected using UPLC-ESI-HRMS. The bacterial 16S rRNA gene and estrogen catabolic gene (oecC) in the subsamples were analyzed using the Illumina MiSeq platform (San Diego, CA, USA) and PCR-based functional assay, respectively.

UPLC-ESI-HRMS.

Ethyl acetate extractable samples were analyzed through UPLC-ESI-HRMS on a UPLC system coupled to an ESI-mass spectrometer. Separation was achieved on a reversed-phase C₁₈ column (Acquity UPLC BEH C₁₈, 1.7 μm, 100 by 2.1 mm; Waters) with a flow rate of 0.4 ml/min at 50°C (column oven temperature). The mobile phase comprised a mixture of two solvents: solvent A (2% [vol/vol] acetonitrile containing 0.1% [vol/vol] formic acid) and solvent B (methanol containing 0.1% [vol/vol] formic acid). Separation was achieved with a linear gradient of solvent B from 5% to 99% in 12 min. Mass spectral data were collected in +ESI mode in separate runs on a Waters quadrupole time of flight (Q-TOF) Synapt G2 high-definition mass spectrometer (Milford, MA, USA) operated in a scan mode of m/z 50 to 500. The capillary voltage was set at 3,000 V; the source and desolvation temperatures were 80°C and 250°C, respectively. The cone gas flow rate was 50 liters/h. The predicted elemental composition of individual intermediates was calculated using the MassLynx mass spectrometry software (Waters).

DNA extraction, amplification of 16S rRNA genes, and Illumina sequencing.

DNA was extracted from frozen sludge samples using the PowerSoil DNA isolation kit (Mo Bio Laboratories, Carlsbad, CA, USA). A 16S amplicon library was prepared according to the Illumina 16S metagenomic sequencing library preparation guide (https://support.illumina.com/content/dam/illumina-support/documents/documentation/chemistry_documentation/16s/16s-metagenomic-library-prep-guide-15044223-b.pdf [accessed 20 March 2018]), with slight modifications. Genomic sections flanking the V3–V4 region of the bacterial 16S rRNA gene were amplified from 24 sludge treatment samples by using HiFi HotStart ReadyMix (Kapa Biosystems, Boston, MA, USA) through PCR (95°C for 3 min; 25 cycles of 95°C for 30 s, 55°C for 30 s, 72°C for 30 s; and 72°C for 5 min). A primer pair flanked by the Illumina Nextera linker sequence (forward primer, 5′-TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCTACGGGNGGCWGCAG-3′; reverse primer, 5′-GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTACHVGGGTATCTAATCC-3′) was used. The PCR products were first separated on an agarose gel, and those with the expected size (approximately 445 bp) were excised from the gel and purified using the GenepHlow gel/PCR kit (Geneaid, New Taipei City, Taiwan). Next, Illumina Nextera XT index (Illumina) sequencing adapters were integrated to the ends of the amplicons through PCR (95°C for 3 min; 8 cycles of 95°C for 30 s, 55°C for 30 s, 72°C for 30 s; and 72°C for 5 min). The final libraries were purified using AMPure XP beads (Beckman Coulter, Pasadena, CA, USA) and quantified using the Qubit double-stranded DNA (dsDNA) HS assay kit (Life Technologies, Carlsbad, CA, USA). The library profiles were randomly analyzed using the Agilent high-sensitivity DNA kit for BioAnalyzer (Santa Clara, CA, USA). To ensure consistency in pooling, all 24 libraries were subjected to quantitative PCR (qPCR) normalization using Kapa library quantification kits to derive the molar concentration, and the final library mixture was verified using qPCR. The library pool was sequenced on the Illumina MiSeq V2 sequencer by using the MiSeq reagent kit V3 for paired-end reads (2 × 300 bp).

Analysis of sequencing data.

In our pipeline, USEARCH (version 10.0.240) (41) and mothur (version 1.35.1) (42) were applied. The assemblage of paired-end reads, primer removal, quality filtering, chimera and singleton detections, and read number normalization were performed using the sequence analysis tool USEARCH, and the taxonomic assignment of OTUs was performed against the Silva database (version 128; https://www.arb-silva.de/documentation/release-128/) by using mothur. The taxonomic assignment and abundance of individual OTUs are presented in Data Set S1. The commands used in the command line for analyzing the bacterial community structure are outlined as follows: (i) merge the reads from paired-end sequencing (use USEARCH): usearch –fastq_mergepairs _R1.fastq –relabel @ -fastq_maxdiffs 10 –fastq_pctid 80 –fastq_minmergelen 400 –fastq_maxmergelen 600 –fastqout ../outmerged. fq; (ii) strip primers (use USEARCH): usearch –fastx_truncate ../out/merged.fq –stripleft 17 –stripright 21 –fastqout ../out/stripped.fq; (iii) execute quality filtering (use USEARCH): usearch –fastq_filter ../out/stripped.fq –fastq_maxee 1.0 –fastaout ../out/filtered.fa; (iv) discard singletons (use USEARCH): usearch –fastx_uniques ../out/filtered.fa –fastaout ../out/uniques.fa –sizeout; (v) execute OTUs picking (use USEARCH): usearch –cluster_otus ../out/uniques.fa –minsize 2 –otus ../out/otus.fa –relable Otu; (vi) construct an OTU table and remove chimeras (use USEARCH): usearch –otutab ../out/merged.fq –otus ../out/otus.fa –otutabout ../out/otutab.txt –mapout ../out/map.txt; (vii) assign the OTUs to the taxonomy (use mothur): classify.seqs (fasta = otus.fa, template = silva.nr_v128.align, taxonomy = silva.nr_v128.tax, cutoff = 80, probs = F, iters = 100, processors = 4); (viii) normalize the size of sequence numbers of each subsamples (use USEARCH): usearch -otutab_norm ../out/otutable_Bacteria.txt -sample_size 24900 -output ../out/otutab_norm_Bacteria.txt classify.seqs.

Amplification of oecC genes from the activated sludge samples using degenerate primers.

Multiple alignment of oecC genes from three estrogen-degrading proteobacteria, including strain KC8, A. estronivorus strain MH-B5, and N. tardaugens NBRC 16725, was performed using the Align/Assemble function in Geneious 8.1.4. A degenerate primer pair (forward primer oecC-f1, 5′-YKCGGYYTGGGCTATGTSGG-3′; reverse primer oecC-r1, 5′-ATCGCGCCSCASCCRATYTC-3′) were derived from the conserved regions (Fig. S3). oecC fragments were amplified through PCR (94°C for 5 min; 30 cycles at 94°C for 30 s, 55°C for 30 s, and 72°C for 40 s; and finally, 72°C for 5 min). Amplified oecC sequences were cloned into Fast-Trans competent E. coli DH5α (Protech Technology Enterprise, Taipei, Taiwan) using the yT&A cloning kit (Yeastern Biotech, Taipei, Taiwan). oecC fragments (approximately 830 bp) were sequenced on an ABI 3730 xl DNA analyzer (Applied Biosystems, Waltham, MA, USA) with the BigDye Terminator kit, according to the manufacturer's instructions.

Isolation of the Novosphingobium sp. strain SLCC from activated sludge amended with E1.

Activated sludge (1 liter) was collected from the DHSTP in January 2015. The DHSTP sludge (100 ml) was aerobically incubated with 1 mM E1 for 7 days. The basic medium used for the isolation and routine cultivation of strain SLCC contained 2.0 g of NH₄Cl, 0.1 g of CaCl₂·2H₂O, 0.5 g of MgSO₄·7H₂O per liter, and 1 mM E1. After autoclaving, this basic medium was supplemented with the following sterile chemicals: 4.95 ml of 1 M KH₂PO₄, 20.05 ml of 1 M K₂HPO₄, 1 ml of an EDTA-chelated mixture of trace elements (43), 1 ml of a selenite and tungstate solution (44), and 1 ml of vitamin solution VL-7 (45). The E1-amended sludge (100 ml) was subcultured in the same medium. After E1 had been exhausted, the subculture was serially diluted (10⁻¹ to 10⁻⁷) in the growth medium. The 10⁻⁷ dilution was spread on R2A agar (BD Difco, Franklin Lakes, NJ, USA) containing 1 mM E1 and aerobically incubated at 28°C for 1 week. Two colonies were picked up from the agar plates and transferred to the chemically defined medium containing 1 mM E1. The aerobic degradation of E1 in the bacterial cultures was confirmed through high-performance liquid chromatography (HPLC). Purity was assessed microscopically and using growth tests in liquid R2A medium or R2A agar. The 16S rRNA genes of the two bacterial colonies capable of aerobic E1 degradation were PCR amplified using the bacterial 16S rRNA universal primers 27F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1492R (5′-GGTTACCTTGTTACGACTT-3′) (46); the resulting PCR products were sequenced to verify the identity of the isolates.

To test the steroid utilization of strain SLCC, this bacterial strain was aerobically grown in chemically defined medium (100 ml) containing different steroid substrates (1 mM), including androst-4-en-3,17-dione, androsta-1,4-diene-3,17-dione, cholesterol, cholic acid, E1, E2, 17α-ethinylestradiol, progesterone, and testosterone. The cultures were sampled daily, and the samples (1 ml) were extracted using ethyl acetate. The consumption of steroid substrates in the culture extracts was detected using thin-layer chromatography (TLC).

Physiological analyses of strain SLCC.

The catalase activity of SLCC was tested by incubating bacterial cells in a 3% H₂O₂ solution. Oxidase activity was determined using the Dropper oxidase reagent (BD Biosciences, San Jose, CA, USA). The assimilation pattern and biochemical characteristics of strain SLCC were determined using the API 20NE system (bioMérieux, Marcy l'Etoile, France), according to the manufacturer's instructions. The assay mixtures were incubated at 30°C for 48 h. Strain SLCC was incubated on R2A agar containing 1 mM E2 at 30°C for 96 h. Whole-cell fatty acid methyl esters of strain SLCC were prepared as previously described (47) and were analyzed using the Sherlock microbial identification system (Midi, Inc., Newark, DE, USA).

Real-time quantitative PCR of bacterial 16S rRNA genes.

A specific SLCC primer pair (forward primer SLCC-for, 5′-GGCGCAGCTAACGCATTAAG-3′; and reverse primer SLCC-rev, 5′-GCGCGTTGCTTCGAATTAAA-3′) and Eub general primer pair (48) (forward primer 341F, 5′-CCTACGGGAGGCAGCAG-3′ and reverse primer 534R, 5′-ATTACCGCGGCTGCTGGC-3′) were used to amplify the 16S rRNA gene of strain SLCC and total eubacterial population, respectively. Three replicates of real-time quantitative PCR experiments were performed using an ABI 7300 sequence detection system (Applied Biosystems, Waltham, MA, USA). The PCR mixture (20 μl) contained 10 μl of Power SYBR green PCR master mix (Applied Biosystems), 0.3 μM each primer, and 30 ng of environmental DNA. The thermal cycling conditions involved an initial denaturation step of 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 60 s.

Aerobic incubation of strain SLCC with 1 mM unlabeled E1.

Strain SLCC was aerobically grown in R2A medium (500 ml in a 2-liter Erlenmeyer flask) containing 0.1 mM E1. Cells were harvested through centrifugation (7,000 × g, 5 min, 25°C) in the exponential-growth phase at an optical density at 600 nm (OD₆₀₀) of 0.5 (optical path, 1 cm). The cell pellet was resuspended in chemically defined medium as described above. The cell suspension (OD₆₀₀, 5; 100 ml) was fed with 1 mM E1 and was aerobically incubated at 28°C with shaking (180 rpm) for 24 h. Subsequently, the estrogen metabolites in the cell suspension were extracted using ethyl acetate and were identified through UPLC-ESI-HRMS.

Aerobic incubation of strain SLCC with 1 μg/liter [3,4C-¹³C]E1.

Strain SLCC was aerobically grown in R2A medium (250 ml in a 1-liter Erlenmeyer flask). Bacterial cells (50 ml) in the exponential-growth phase (OD₆₀₀, 1) were transferred to chemically defined medium (1 liter) containing [3,4C-¹³C]E1 (1 μg/liter; 3.7 nM). The [3,4C-¹³C]E1-amended treatments were performed in triplicate. The negative controls included bacterial cells incubated in chemically defined medium (1 liter) without [3,4C-¹³C]E1 and autoclaved cells incubated with [3,4C-¹³C]E1. The treatments were aerobically incubated at 28°C with shaking (180 rpm) for 24 h. Cells were removed through centrifugation (10,000 × g, 10 min, 4°C). The supernatants were then vacuum filtered through a cellulose acetate membrane filter (pore size, 0.45 μm) before solid-phase extraction. Oasis HLB LP extraction cartridges (6 ml; Waters) with 0.5 g of sorbent were used to extract pyridinestrone acid from the samples. Cartridges were preconditioned sequentially with methanol (20 ml) and deionized water (10 ml). The filtered samples were loaded into the cartridges at a flow rate of approximately 5 ml/min. The cartridges were washed with deionized water (10 ml) and dried under a vacuum. Subsequently, pyridinestrone acid was eluted with 12 ml of methanol. The extracts were collected, dried, and redissolved in 500 μl of methanol. The samples were centrifuged at 13,500 × g for 5 min, and the estrogen metabolite, [3,4C-¹³C]pyridinestrone acid, was detected through UPLC-ESI-HRMS.

TLC.

The steroids were separated on silica gel aluminum TLC plates (silica gel 60 F₂₅₄; thickness, 0.2 mm; 20 by 20 cm; Merck, Darmstadt, Germany). Ethyl acetate-H₂O-acetate (85:10:10 [vol/vol/vol]) was used as the developing solvent system for the separation of pyridinestrone acid. Dichloromethane-ethyl acetate-methanol (14:4:1 [vol/vol/vol]) was used as the developing solvent system for the separation of neutral steroid substrates. The steroid compounds were visualized under UV light at 254 nm or by spraying the TLC plates with 30% (vol/vol) H₂SO₄.

HPLC.

A reversed-phase (RP) HPLC system (Hitachi, Tokyo, Japan) was used for quantifying estrogen metabolites. Separation was achieved on an analytical RP-C₁₈ column [Luna 18(2), 5 μm, 150 by 4.6 mm; Phenomenex, Torrance, CA, USA], with a flow rate of 0.5 ml/min. The separation was performed isocratically at 35°C, with 40% (vol/vol) methanol containing 0.1% (vol/vol) trifluoroacetic acid serving as an eluent. The steroid products were detected in the range of 200 to 450 nm using a photodiode array detector.

Accession number(s).

Strain SLCC was deposited at the Bioresource Collection and Research Center (BCRC), Hsinchu, Taiwan, under the accession number BCRC 81104. The nucleotide sequence data set was deposited in the National Center for Biotechnology Information (NCBI) Sequence Read Archive under accession number SRP123564.