Water concentration method and nucleic acid extraction for viruses and ARGs
Grab influent (n = 72) and effluent (n = 72) wastewater samples were collected along with dehydrated biosolid samples (n = 72) from 6 different WWTPs over a one-year period (January 2022 – December 2022). Samples were grabbed early in the morning (8 am) by collecting ~ 500 mL of wastewater in sterile HDPE plastic containers (Labbox Labware, Spain). Collected samples were transferred on ice to the laboratory, kept refrigerated at 4°C, and concentrated within 24 h. Samples were artificially contaminated with 106 PCR units (PCRU) of porcine epidemic diarrhea virus (PEDV) strain CV777, serving as a coronavirus model. Additionally, 106 PCRU of mengovirus (MgV) vMC0 (CECT 100000) were used as a non-enveloped counterpart for recovery efficiency assessment. Effluent wastewater samples were concentrated through a previously validated aluminium-based adsorption-precipitation method11,91. Alternatively, 40 mL of influent wastewater samples were processed with the Enviro Wastewater TNA Kit (Promega Corp., Spain) vacuum concentration system following the manufacturer's instructions92. For biosolid samples, 0.1g of biosolid were resuspended in 900 µL PBS for nucleic acid extraction prior to PCR analyses.
Nucleic acid extraction from influent and effluent wastewater concentrates and biosolid suspensions was performed by using the Maxwell® RSC Instrument (Promega, Spain) with the Maxwell RSC Pure Food GMO for viral and ARG extraction. Specific programs, namely 'Maxwell RSC Viral Total Nucleic Acid' and 'PureFood GMO and Authentication,' were employed for viral and ARG extractions, respectively.
Virus detection and quantification
The detection of process control viruses, PEDV and MgV, was carried out through RT-qPCR using the One Step PrimeScript™ RT-PCR Kit (Perfect Real Time) (Takara Bio Inc., USA) as detailed elsewhere93. Levels of HuNoV GI and GII, HAstV, RV, HAV and HEV were determined using the RNA UltraSense One-Step kit (Invitrogen, USA), following previously described procedures9,11. The occurrence of crAssphage was established using the qPCR Premix Ex Taq™ kit (Takara Bio Inc)94. PMMoV detection was determined using the PMMoV Fecal Indicator RT-qPCR Kit (Promega, Spain) following the manufacturer’s instructions. SARS-CoV-2 detection was performed by targeting the N1 region of the nucleocapsid gene. The One Step PrimeScript™ RT-PCR Kit (Perfect Real Time) was used with N1 primers and conditions described by CDC95. IAV detection followed the protocol described by CDC (2009) using primers from CDC (2020) and the One Step PrimeScript™ RT-PCR Kit (Perfect Real Time)96.
Different controls were used in all assays: negative process control consisting of PBS; whole process control to monitor the process efficiency of each sample (spiked with PEDV and MgV); and positive (targeted gene reference material) and negative (RNase-free water) RT-qPCR controls. The recoveries of PEDV and MgV, spiked as enveloped and non-enveloped viral process controls, respectively, ranged between 6.31 and 59.65% (data not included). The validation of results for targeted viruses adhered the criteria specified in ISO 15216-1:2017, where a recovery of the process control of ≥ 1% is required97.
Commercially available gBlock synthetic gene fragments (Integrated DNA Technologies, Inc., USA) of HuNoVs GI and GII, HAstV, RV, HAV, HEV, and crAssphage were used to prepare standard curves for quantification. For IAV and RSV quantification, Twist Synthetic InfluenzaV H1N1 RNA control (Twist BioScience, South San Francisco, CA, USA), and purified RNA of RSV (Vircell, S.L., Spain) were used. The PMMoV Fecal Indicator RT-qPCR Kit (Promega) provided PMMoV RNA for generating a standard curve. A table, featuring primers, probes, PCR conditions, limit of quantification (LOQ/L), and limit of detection (LOD/L) for all targeted viruses in this work is available in the Supplementary materials (Table S1).
Quantification of viable somatic coliphages, E. coli, and Extended Spectrum Beta-Lactamases producing E. coli.
Somatic coliphages were determined from wastewater samples filtered through sterile filters (0.45 µm pore) by using a commercial Bluephage Easy Kit for Enumeration of Somatic Coliphages (Bluephage S.L., Spain), following manufacturer’s instructions. For biosolid samples, 1g of biosolid was resuspended in 100 mL PBS for both somatic coliphages and E. coli enumeration.
For all water and biosolid samples, E. coli and Extended Spectrum Beta-Lactamases producing E. coli (ESBL-E. coli) enumeration was assessed by using selective culture media Chromocult coliform agar (Merck, Darmstadt, Germany) and CHROMagar ESBL (CHROMagar, Paris, France), respectively. Spread plating (0.1 mL) or membrane filtration (200 mL) was used depending on the anticipated bacterial concentration. Influent wastewater samples were diluted serially, and 0.1 mL aliquots were spread-plated. Effluent samples were filtered through a 0.45 µm cellulose nitrate membrane filter (Sartorius, Madrid, Spain). Following incubation at 37 ºC for 24 hours, results were interpreted, with. dark blue-violet colonies considered positive for E. coli and dark pink-reddish colonies considered positive for ESBL-E. coli. The analysis was performed in duplicate, and the results were expressed as CFU/100 mL. The detection limit (LOD) for E. coli and ESBL-E. coli counts in the influent and biosolid samples was 2.0 Log CFU/100 mL (100 CFU/100 mL), while in the effluents, the LOD was 0 Log CFU/100 mL (1 CFU/100 mL).
Detection and quantification of antimicrobial resistance genes in effluent waters and biosolids
In this study, 11 ARGs that confer resistance to Sulfonamides (sul1, sul2_1), beta-lactamase (pbp2b, blaCTX−M), phenicols (cmlA_2), nitroimidazoles (nimE), MLSB (ermB_1, ermA), tetracyclines (tetPB_3, tetA_1) and fluoroquinolones (qacA_1), were only detected in effluent waters and biosolids. The 16S rRNA gene was used as positive control for qPCR measurement. Quantification of the 12 selected genes was performed by high-throughput quantitative PCR (HT-qPCR) using the SmartChip™ Real-Time PCR system (TakaraBio, CA, USA) by Resistomap Oy (Helsinki, Finland). qPCR cycling conditions and processing of raw data were described elsewhere98–100. Each DNA sample was analysed in duplicate. Data processing and analysis were performed by using a python-based script by Resistomap Oy (Helsinki, Finland)101,102.
Digestion of organic material and isolation of MPs
Initial steps consisted on optimizing the protocol for the removal of organic material and the isolation of the maximum number of MPs from wastewater and biosolid samples. Different volumes of water, amounts of biosolids and digestion strategies for organic biomass removal were tested to remove the greatest amount of organic material without compromising the integrity of the MPs. Avoiding filter clogging was a requirement during the methodology development, to facilitate further identification of MPs. To reduce the risk of external contamination by MPs, laboratory consumables made of glass were used, the reagents were purified by filtering through a 0.2 µm pore size nitrocellulose filter (Whatman, Maidstone, UK), 100% cotton lab aprons were used, samples were processed in a laminar flow cabinet, the beakers were covered with a watch glass, disposable nitrile gloves were used and, before and after using the material, all used materials were rinsed thoroughly with deionized water. In order to assure that the isolation of MPs was effective and external contamination did not occur, a negative control (NC) was included every month and a positive control (PC) was carried out every 3 months. The positive control was made with fluorescent polystyrene microspheres (Invitrogen, Waltham, USA) of 1 µm in diameter. Specifically, a solution of 1000 beads/20 µL was prepared and 20 µL of this solution was incorporated before the pre-treatment and, the number of remaining microbeads after the digestion protocol was determined to calculate the percentage of recovery. The average value of particle recovery was 93.9%.
Two different pre-treatment protocols were finally defined:
1) Sieved > 300 µm or (S): With this pre-treatment, all solid particles (including MPs) larger than 300 µm were isolated from 2 L of wastewater or 5 g of biosolid samples after sieving, oxidative digestion, and filtration steps.
2) Total Particles or (T): With this pre-treatment all solid particles (including MPs) with a size between 1 µm and 5 mm were isolated from a 10 mL aliquot of wastewater after oxidative digestion, density separation, and filtration steps.
Through protocol (S), a larger and more representative amount of wastewater was treated, but particles smaller than 300 µm were lost. In the other hand, protocol (T) allowed the analysis of particles down to 1 µm in size, but the amount of analysed wastewater was much smaller to avoid filter clogging.
In both protocols (S) and (T), oxidative digestion was performed to remove organic material, adapting the method described by the National Oceanic and Atmospheric Administration (NOAA)103.
In the case of the Sieved 300 µm or (S) protocol (Fig. 11), 2L of wastewater or 5 g of biosolids were treated. The 5 g of biosolids were previously dispersed in 100 mL of ultrapure MilliQ water by applying stirring and heat during 30 minutes at 30 ºC. The wastewater or biosolid dispersion were subsequently poured through a 300 µm mesh stainless steel sieve. The retained particles were collected by washing with MilliQ water into a beaker and digested by adding an equivalent volume of NaClO (14%, VWR chemical, USA). After heating at 75 ºC for 3 h under stirring, the sample was sieved again to remove the disaggregated smallest particles. The particles retained on the sieve were collected by washing with MilliQ water on a 0.8 µm pore size nitrocellulose filter (Whatman, USA). The filter was protected from external contamination between a microscope glass slide and a glass cover, and finally dried at 40ºC for 24 h in a convection oven.
In the case of the Total Particles or (T) protocol, an oxidative digestion (Fenton reaction) was performed on a 10 mL wastewater sample by adding 20 mL of a H2O2 (30%, Sigma- Aldrich, USA) solution and 20 mL of a 0.05 M Fe (II) solution prepared by mixing FeSO4 (Sigma- Aldrich, USA), H2SO4 (96%, PanReac AppliChem, ITW Reagents, USA) and deionized water. The sample was then heated at 75°C for 30 min under stirring. The digestion step was repeated if any remaining organic material was visually. Thereafter, a density separation was performed after adding NaCl (99.5%, Sigma- Aldrich, USA) until saturation. Subsequently, the sample was left to sediment for 30 min in a separatory funnel and the supernatant was filtered through a 0.8 µm pore size nitrocellulose filter (Whatman, USA) under vacuum. The filter was also protected between glass slide and coverslip and dried at 40ºC for 24 hours.
Characterization of particles present in biosolid and wastewater samples.
Filters obtained after pre-treatment protocols (S) and (T) were photographed using an EVOCAM II macrophotography equipment (Vision engineering, Woking, UK) and the ViPlus software (2018, Vision Engineering). Two partially overlapping 2MPx color photos were taken for each filter, always at 20x magnification, with half of the filter appearing in each photo. These images were fused by digital stitching techniques using the mosaic J command of the FIJI software (ImageJ 1.49q Software, National Institutes of Health, USA). Each image showed a 25*15mm field of view. The pixel size was 13.3 microns, obtaining an image to calibrate in each photo session to have a precise external calibration data. A rough quantification was performed, and all particles, including MPs, were characterized using the Nis Elements BR 3.2 software (Nikon corporation, Japan). To achieve this, a macro of programmed actions was designed in which, firstly, the pixel size was calibrated in the complete image of the filter, then a matrix-iterative detection tool for particles less bright than the filter was applied, which facilitated a binary segmentation by brightness levels and achieve the selection of the particles of each filter in an automated way, only in the filtration zone. Finally, the data of all the particles were exported to obtain the count and the different morphological values of numerous parameters and perform the statistical calculations.
For the characterization, the particles were classified into 3 size ranges of 1-100 µm, 100–300 µm and 300–5000 µm. The particles were also classified according to their circularity, calculated from the measured perimeter and area of each particle according to Eq. 1, in 3 ranges: 0-0.4, 0.4–0.8 and 0.8-1. A circularity value of 1.0 indicates a perfect circle. As the value approaches 0.0, it indicates an increasingly elongated polygon. Particles with a circularity less than 0.4 were considered as fibers.
\(Circularity=4\pi (\frac{area}{{perimeter}^{2}}\) ) (1)
In addition, the efficiency of WWTPs in removing particles was calculated according to the following equation:
$$Efficiency=\frac{\text{i}\text{n}\text{f}\text{l}\text{u}\text{e}\text{n}\text{t}-\text{e}\text{f}\text{f}\text{l}\text{u}\text{e}\text{n}\text{t}}{\text{i}\text{n}\text{f}\text{l}\text{u}\text{e}\text{n}\text{t}}\times 100$$
2
Where: Efficiency = particle removal efficiency (%); influent = number of particles detected at the WWTP influent; effluent = number of particles detected at the WWTP effluent.
Quantification of microplastics present in biosolid and wastewater samples.
Quantification, identification and characterization of MPs was carried out only on samples from the odd months. The analysis was performed using an automated Raman microscope Alpha300 apyron (Witec, Ulm, Germany). First, each filter was mapped by acquiring a total of 1089 images, which after reconstruction represented a 27% of the filter area or 1 cm2. The present particles were detected and selected by performing image analysis using the ParticleScout 6.0 software in automatic mode.
After particle selection, analysis on each particle by Raman spectroscopy and subsequent identification were carried out. The optimal conditions for Raman spectra acquisition were as follows: 785 nm laser which facilitates to identify fluorescent particles, 300 lines/mm diffraction grating opening, spectral range between 0 and 3000 cm− 1, 10 accumulations, 0.2 second acquisition time, and 40 mW laser power. The spectrum of each particle was registered and compared with an in-house build spectral library of polymers. The reference polymer materials included in the spectral library were polyethylene (PE), polyethylene terephthalate (PET), polyamide (PA), polypropylene (PP), polystyrene (PS), polyvinyl chloride (PVC), polytetrafluoroethylene (PTFE), polyacrylamide (PAM), Polyarylsulfones (PSU), Polymethylmethacrylate (PMMA), nitrile rubber (NBR), Cellophane and Melamine. Particles that had a 75% or better match (HQI) between the sample and reference spectra were identified as composed of the same material or of a similar chemical nature. In addition, a visual inspection was carried out and the spectrum acquisition was repeated on the particles where a clear identification was not initially possible. Three rules were considered to discriminate between plastics and non-plastics and to prioritize the particles to be analysed: i) the object must not show cellular or natural organic structures; ii) the fibre thickness must be uniform along the entire length; iii) the colour of the particles must be clear and homogeneous104. The MPs already identified were classified based on material type, size, morphology, and area.
Statistical analysis
Results were statistically analysed and significance of differences was determined on the ranks with a one-way analysis of variance (ANOVA) and Tukey’s multiple comparison tests. In all cases, a value of p < 0.05 (confidence interval 95%) was deemed significant.