key: cord-0682883-2yw9uamt
authors: Monteiro, Silvia; Rente, Daniela; Cunha, Mónica V.; Marques, Tiago A.; Cardoso, Eugénia; Álvaro, Pedro; Vilaça, João; Ribeiro, Jorge; Silva, Marco; Coelho, Norberta; Brôco, Nuno; Carvalho, Marta; Santos, Ricardo
title: Recovery of SARS-CoV-2 from large volumes of raw wastewater is enhanced with the inuvai R180 system
date: 2021-12-15
journal: J Environ Manage
DOI: 10.1016/j.jenvman.2021.114296
sha: e83ff56befb691f80a98422ecce237198c0033a2
doc_id: 682883
cord_uid: 2yw9uamt

Wastewater-based epidemiology (WBE) for severe acute respiratory syndrome Coronavirus 2 (SARS-CoV-2) is a powerful tool to complement syndromic surveillance. Although detection of SARS-CoV-2 in raw wastewater may be prompted with good recoveries during periods of high community prevalence, in the early stages of population outbreaks concentration procedures are required to overcome low viral concentrations. Several methods have become available for the recovery of SARS-CoV-2 from raw wastewater, generally involving filtration. However, these methods are limited to small sample volumes, possibly missing the early stages of virus circulation, and restrained applicability across different water matrices. The aim of this study was thus to evaluate the performance of three methods enabling the concentration of SARS-CoV-2 from large volumes of wastewater: i) hollow fiber filtration using the inuvai R180, with an enhanced elution protocol and polyethylene glycol (PEG) precipitation; ii) PEG precipitation; and iii) skimmed milk flocculation. The performance of the three approaches was evaluated in wastewater from multiple wastewater treatment plants (WWTP) with distinct singularities, according to: i) effective volume; ii) percentage of recovery; iii) extraction efficiency; iv) inhibitory effect; and v) the limits of detection and quantification. The inuvai R180 system had the best performance, with detection of spiked control across all samples, with average recovery percentages of 68% for porcine epidemic diarrhea virus (PEDV), with low variability. Mean recoveries for PEG precipitation and skimmed milk flocculation were 9% and 14%, respectively. The inuvai R180 enables the scalability of volumes without negative impact on the costs, time for analysis, and recovery/inhibition. Moreover, hollow fiber ultrafilters favor the concentration of different microbial taxonomic groups. Such combined features make this technology attractive for usage in environmental waters monitoring.

Coronavirus 2 (SARS-CoV-2) is a powerful tool to complement syndromic 24 surveillance. Although detection of SARS-CoV-2 in raw wastewater may be prompted 25 with good recoveries during periods of high community prevalence, in the early stages 26 of population outbreaks concentration procedures are required to overcome low viral 27

concentrations. Several methods have become available for the recovery of SARS-28

CoV-2 from raw wastewater, generally involving filtration. However, these methods while other studies show all data performance, including LoD, LoQ and recovery 103 percentages (Philo et al., 2021) . Additionally, different studies calculate the LoD and 104

LoQ differently. The information collected from different studies should inform 105 laboratories on method performance. 106

In the present study, we evaluated the efficiency of SARS-CoV-2 recovery from raw 107 wastewater using three concentration methods: i) a newly developed hollow-fiber 108 ultrafilter, inuvai R180 (inuvai, a division of Fresenius Medical Care), with an improved 109 elution protocol; ii) PEG precipitation; and iii) skimmed milk flocculation. The 110 performance of the three methods was compared in raw wastewater according to 111 several characteristics, including: i) effective volume tested; ii) frequency and 112 consistency of detection; iii) percentage of recovery; iv) extraction efficiency; v) 113 inhibitory effect on reverse transcription-qPCR (RT-qPCR); and vi) concentration 114 information (including, Limit of Detection (LoD) and Limit of Quantification (LoQ)). 115

These three methodologies were chosen primarily because they allow for the 116 concentration of large volumes of water which is of extreme relevance at the early 117 stages of the epidemiological curve and because they allow for the concentration of 118 different types of water, including treated wastewater and environmental waters, 119 important when analyzing the entire SARS-CoV-2 urban water cycle. This study 120 benchmarks new and old methodologies for the detection of SARS-CoV-2 from raw 121 wastewater for WBE applications. incubated at 37 (± 1) ºC in 5% CO2 for 4 days. PEDV were recovered following three 137 cycles of freeze/thawing and centrifugation at 1,100 xg for 10 min. Quantification was 138 performed by reverse transcription digital PCR (RT-dPCR) as described on section 139 2.5 using the primers and probes from Supplementary Table S1 (Zhou et al., 2017) , 140 following nucleic acid extraction as described on section 2.4. After absolute 141 quantification by RT-dPCR (as described below), a stock solution was prepared in 142

DNase/RNase free water to obtain a PEDV final concentration of 1.21 x 10 4 GC/L in 143 wastewater. The same stock was used in all experiments described below. 144 145

Twenty-four-hour composite samples were collected between July 6-10, 2020, from 147 wastewater, without problems such as clogging or compromising of the membrane 159

structure. All methods were tested using the same initial volume of wastewater (1-L) 160

for a more accurate comparison. Method 2 used PEG precipitation: 20% PEG 8000 was added directly to 1-L of raw 171 wastewater, with overnight precipitation followed by centrifugation as described above fold dilutions in RNase/DNase-free water. Positive and NTC controls were also added 227 to each PCR assay. Positive and NTC controls were also added to each PCR assay. 228

Limits of detection (LoD) and quantification (LoQ) were determined in RNase/DNase-229 free water. The LoD was considered the lowest concentration of target that could be 230 consistently detected (in more than 95% replicates tested) (Burd et al., 2010) and LoQ, 231 the lowest concentration at which the performance of the method is acceptable, with 232 a coefficient of variation below 35% (Klymus et al., 2020) . percentage of recovery was obtained with the inuvai R180 system at 68% (± 11%) for 300 PEDV, with recoveries varying between 52 and 82% (Fig. 1A) . 301 PEG precipitation had the lowest percentage of recovery for PEDV (9% (± 5%)). 302

Recovery with skimmed milk performed only slightly better (14% (± 8%)) (Fig. 1A) . 303

There were statistically significant differences in the recovery percentage of PEG and 304 skimmed milk compared to inuvai R180 (KW = 15.989, df = 2, p < 0.05), although no 305 difference was observed between skimmed milk flocculation and PEG precipitation (p 306 < 0.05). 307

The inuvai R180 was the single method that consistently led to nucleic acid detection 308 in all samples. Concentration using PEG and skimmed milk led to the detection of 309 PEDV in 50% of the samples. The method using the inuvai R180 system led to detection by RT-qPCR of the highest 318 mean genome copies, 4.25 GC/reaction for PEDV, respectively. Concentration with 319 PEG (0.21 GC/reaction) and skimmed milk (0.28 GC/reaction) showed similar results 320 (Fig. 1B) . 321

Our recovery values using the inuvai R180 system were similar to those reported for coupled with an enhanced elution strategy with three steps that we adopted, and/or to 336 the secondary concentration protocol. The inuvai R180 ultrafilter has a reduced 337 nominal pore size (≤ 5.5 nm with a correspondent cut-off ≤ 18.8 Kda) compared to the 338 Rexeed 15S, which has a more open pore structure. Additionally, the ultrafilter used 339 in our study has a larger membrane area (1.8 m 2 for inuvai R180 vs 1.5 m 2 for Rexeed 340 S15) and larger fiber inner diameters (220 μm for inuvai R180 vs 185 μm for Rexeed 341 S15 precipitate properly  pH adjustment is required  Cannot be applied in the field  Requires the use of an orbital shaker, a relatively expensive equipment

The extraction efficiency using MNV as proxy averaged 70% (±19%) for inuvai R180 352 protocol. Extraction efficiencies for PEG precipitation and skimmed milk flocculation 353 averaged 50% (±15%) and 36 (±13%), respectively. RNA extraction was conducted 354

using QIAamp Fast DNA Stool mini kit (QIAGEN, Germany) that although being used 355 for DNA extraction, can efficiently co-extract RNA. The reagents provided by the 356 manufacturer as well as the columns provided are not specific for DNA. The handbook 357

for the kit indicates that RNase should be used to remove RNA from the final eluate. 358

Detection at a high rate demonstrated for MNV showed high efficiency, demonstrating 359 therefore the application of this DNA extraction kit for the co-extraction of RNA. 360

Detection of PEDV using the inuvai R180 system was consistently achieved with the 361 1/4-fold dilution, while for undiluted spiked samples, only 38% could be detected 362 without inhibition. PEG precipitation was the single method that detected both targets 363 from undiluted samples, although inhibition still occurred (as evidenced subsequently 364 by testing the 4-and 10-fold dilution). As for the skimmed milk concentration method, 365 detection in undiluted concentrates was found for 75% of the samples, although 366 inhibition still occurred (as measured by the dilutions). These results indicate that 367 inhibitory effects exerted upon RT-qPCR could be confirmed for the three methods 368 under comparison. Difference between concentration using the combined R180 369 system (including elution step and PEG concentration) versus simply using PEG 370 concentration may arise from the elution buffer used and high performance of the 371 ultrafilter itself. Also, co-concentration of PCR inhibitors occurred for all three methods, 372 although with a better extraction recovery using MNV obtained for the combined R180 373 system, which indicates that concentration using only PEG and skimmed milk 374 potentiated higher co-concentration of inhibitors. Although the ultrafilters concentrated 375 both liquid and solid phases and possible inhibitors were co-concentrated, many might 376 have been directly eliminated during the filtration process and other may not have 377 been eluted during the elution procedure. Therefore, one of the major driving forces 378 for the different results, particularly when comparing both PEG-based methods, might 379 have been the co-concentration of inhibitors. 380

Overall, our results showed that the inuvai R180 system coupled with an improved 381 elution protocol is highly suitable for the detection of PEDV, used as a surrogate for 382 SARS-CoV-2 in this study, exhibiting the highest percentage of detection and mean 383 recovery value. Additionally, this method also showed greater extraction efficiency and 384 larger volume processing without increased cost or time for downstream analyses. 385

Furthermore, the performance of the inuvai system showed consistency across raw 386 wastewater samples from different catchments / WWTP, including the Serzedelo 387 WWTP, which is highly impacted by industrial effluents (tannery industry) and 388 therefore an extremely complicated matrix to work with altogether, a result 389 corroborated by the Pan-European Umbrella study (Gawik et al., 2021) . In the 390 Umbrella study, raw wastewater samples from different European countries were 391 collected and sent for analysis in a centralized laboratory. In parallel, the same 392 samples were also analyzed in each country for comparison of results. The centralized 393

European laboratory was unable to recover SARS-CoV-2 RNA from Serzedelo raw 394 wastewater presenting low recovery percentages (0.1%) and lower concentrations of 395 crAssphage compared to the other samples analyzed. The same sample, analyzed by 396 our group and using the inuvai R180 system, was positive for SARS-CoV-2 and the 397 concentration of crAssphage was 3-log above that detected by the centralized 398 laboratory. These results demonstrate the difficulty of working with this raw 399 wastewater, highlighting the need to test method performance in raw wastewater from 400 different origins. 

After establishing the inuvai R180 system as gold-standard for primary concentration, 436 the efficiency of the relative quantification method (RT-qPCR) was assessed by 437 calculating the LoD and LoQ for the E_Sarbeco and RdRp assays. 438

The LoD was 3.99 GC and 5.52 GC per reaction for the E_Sarbeco and RdRp assays, 439

respectively. This corresponded to a method LoD of 2.73 x 10 3 GC/L for E_Sarbeco 440 and 3.79 x 10 3 GC/L for RdRp using the inuvai R180 system. As for the LoQ, the 441 results were 66 GC and 178 GC per reaction for the E_Sarbeco and RdRp assays, 442

respectively. This corresponded to a method LoQ of 4.56 x 10 4 GC/L for E_Sarbeco 443 and 1.22 x 10 5 GC/L for RdRp assay. 444

The LoD obtained in our study were inferior to those obtained by Philo et al. (2021) . 

Comparison 493 of virus concentration methods for the RT-qPCR-based recovery of murine hepatitis 494 virus, a surrogate for SARS-CoV-2 from untreated wastewater

Detection of murine norovirus 1 by plaque assay, transfection assay, and real-498 time reverse transcription-PCR before and after heat exposure

Scientific 501 Interest Group Obépine

Epidemiological surveillance of SARS-CoV-2 by genome quantification in wastewater 503 applied to a city in the northeast of France: comparison of ultrafiltration-and protein 504 precipitation-based methods

Glass wool concentration optimization for the detection of 508 enveloped and non-enveloped waterborne viruses

Validation of laboratory-developed molecular assays for infectious 510 diseases

Development and 513 sentinel system

COVID-19 surveillance in Southeastern 541

Virginia using wastewater-based epidemiology

Development of 543 virus concentration methods for the recovery of koi herpervirus in water

First environmental 546 surveillance for the presence of SARS-CoV-2 RNA in wastewater and river

Washington state 2019-nCoV case investigation team

First case of 2019 novel coronavirus in the United States

Benchmarking virus concentration methods for quantification of SARS-CoV-2 in 556 raw wastewater

Reporting the limits of detection and quantification 561 for environmental DNA assays

First detection of SARS-CoV-2 in 565 untreated wastewater in Italy

Evaluating recovery, cost, and throughput of different 569 concentration methods for SARS-CoV-2 wastewater-based epidemiology

Development of a large volume concentration 573 method for recovery of coronavirus from wastewater

SARS-Coronavirus-2 RNA in sewage and correlation with reported COVID-19

prevalence in the early stage of the epidemic in The Netherlands

Nanofluidic digital PCR for the quantification of 580 Norovirus for water quality assessment

A wastewater-based epidemiology tool for COVID-19 584 surveillance in Portugal

Comparing analytical methods to detect SARS-CoV-2 in 588 wastewater

The University of Arizona says it caught a dorm's COVID-19 outbreak 590 before it started. Its secret weapon: poop. The Washington Post

A 593 comparison of SARS-CoV-2 wastewater concentration methods for environmental 594 surveillance

SARS-CoV-2 RNA in wastewater anticipated COVID-19 occurrence in a low 597 prevalence area

Clinical characteristics of 138 600 hospitalized patients with 2019 novel coronavirus-infected pneumonia in Wuhan

CoV-2 in different types of clinical specimens

Timeline of WHO's response to COVID-19 -30

Evaluation of lockdown effect on SARS-CoV-2 609 dynamics through viral genome quantification in waste water

Evidence for 613 gastrointestinal infection of SARS-CoV-2

Characteristics of 616 pediatric SARS-CoV-2 infection and potential evidence for persistent fecal viral 617 shedding

Comparison and evaluation of conventional RT-PCR SYBR 620

green I and TaqMan real-time RT-PCR assays for the detection of porcine epidemic 621 diarrhea virus

We thank all the workers from Águas de Portugal Group who contributed with wastewater samples and 476 those who contributed with critical discussion.