Fifteen collection points were preselected in the state of Rio de Janeiro (estuarine beach, lagoon, river, human water supply, domestic sewage, and hospital sewage). Three factors determined these locations: (i) the influence of the discharge of treated or untreated sewage; (ii) proximity to urban agglomerations; and (iii) proximity to pre-existing or field hospitals.

To facilitate sample analysis, we categorized all collection points into aquatic matrices: coastal seawater, brackish water, freshwater river, drinking water, treated and untreated domestic raw sewage and raw hospital sewage (Table 1). hich were collected in polypropylene barrels (40 L volume), all inland and coastal seawater samples were collected in polypropylene bottles (1.5 L volume) and processed, concentrated, and stored at −20˚C.

The concentration method of coastal, brackish, river, consumption, domestic and hospital sewage samples w by ultracentrifugation as described by Pina and coworkers [7].

Table 1. Aquatic matrices, locations and methods of collection with their respective abbreviations.

The physicochemical parameters of the aquatic matrices were determined based on parameters established by the INEA. For the evaluation of these parameters, a conventional thermometer, Secchi disk, and pH strip were used, with data confirmation performed using a benchtop pH meter, conductivity, temperature, and depth probe, and oximeter, as stipulated in CONAMA Resolution 357/2005 [8].

Enterococci sp. are the first-choice marine microbiological indicators according to the United States Environmental Protection Agency and the World Health Organization [9][10]. To determine the presence of Enterococci sp., the commercial kit Colilert and Enterolert, based on defined substrate technology (DST), developed by IDEXX Laboratories, was used. Briefly, a specific water-soluble substrate was added and reacted with the target bacteria in the samples. After bacterial multiplication in sealed Quant-Tray cards and storage in an incubator, the samples were evaluated by staining and quantified from a table provided by the manufacturer, where concentrations range from 1 to 2 × 10³ microorganisms per 100 mL of sample.

For Escherichia coli, quantification was performed using the commercial kit Colilert-Quanti-Tray/2000 (IDEXX Laboratories, Inc., Westbrook, ME, USA), and the results were expressed as concentration/100mL. All samples used a dilution of 1:10. Thermotolerant coliforms were quantified using the multiple-tube method described in the Food and Drugs Administration Bacteriologic Analytical Manual [11]. Serial dilutions of 1:10, 1:100, and 1:1000 were performed from 10 mL of each water sample in 90 mL of Butterfield Phosphate Diluent. The substrate was then added to the tube containing the 1:1000 dilution. After homogenization, the suspension was divided into five tubes (20 mL each) and incubated at 35˚C for 24 h. Tubes showing turbidity and color change were transferred to Escherichia coli soup and incubated at 44.5˚C for 24 h to confirm the presence of thermotolerant coliforms [12], which was indicated by the presence of gas in the tube [13].

SARS-CoV-2 was evaluated using three analytical approaches: RT-qPCR (real-time polymerase chain reaction), nucleotide molecular characterization, and transmission electron microscopy. Initially, the samples were processed using different techniques (described in the “Study area and sampling” section), and the final concentrated volume was used for transmission electronic microscopy, nucleic acid extraction, reverse transcription, and RT-qPCR.

Viral RNA was extracted using a commercial extraction kit (Allprep PowerViral DNA/RNA kit—Qiagen) on a one-step instrument, as described by the manufacturer.

Briefly, 200 µL of the initial volume (viral concentrated) was used for extraction, followed by the addition of 600 µL of PM1 solution previously incubated for 10 minutes at 55˚C and 6 µL of β-mercaptoethanol. The mixture was homogenized for 30 s and incubated for 5 min at room temperature. Next, 150 µL of Interspersed Repetitive Sequence Polymerase Chain Reaction was added to the sample, homogenized, and incubated between 2 and 8˚C for 5 min. At the end of the incubation period, the mixture was centrifuged for 1 min at 13,000 x g and the supernatant transferred to a collector tube. A total of 600 µL of PM3 and PM4 solutions were added to the collector tube, homogenized, and centrifuged for 1 min at 13,000 x g. The clarified content was discarded, and the centrifugation process was repeated three more times. Then, 600 µL of PM5 solution was added to the column and centrifuged for 1 min at 13,000 x g, followed by removal of the supernatant, addition of 600 µL of PM4 solution, centrifugation for 1 min at 13,000 x g, removal of the supernatant, and another centrifugation for 2 min at 13,000 x g. After sample preparation, the column was placed in a new collector tube with the addition of 100 µL of RNase-free water, incubated for 1 min, and centrifuged again at 13,000 x g for 1 min. Finally, the column was removed and the liquid containing RNA and DNA was stored at −80˚C.

Samples were analysed at the Laboratory of Molecular Virology and Marine Biotechnology (VM/UFF) using the PikoReal^TM 96 real-time PCR system (Thermo Scientific^TM, USA). The SARS-CoV-2 N1 + N2 primer sequences used followed the CDC protocol. Sample cycle Threshold (Ct) values between 16 and 38 were considered positive for SARS-CoV-2 RNA.

Amplification of the S protein gene was performed with six sets of primers targeting the S segment and two sets flanking it, designed based on sequences deposited in GISAID up to September 2020. The SARS-CoV-2 RT-qPCR positives with a Ct value ≤ 16 were selected for confirmation. Selected samples were further analysed by conventional PCR targeting regions of the S gene, followed by 1.2% agarose gel electrophoresis as a complementary molecular analysis. A 100-nucleotide overlap was programmed (Figure 1).

Figure 1. Schematic representation of the fragments of each pair of primers used, where: PR1—primer pair 1, PR2—primer pair 2, PR3—primer pair 3, PR4—primer pair 4, PR5—primer pair 5, PR6—primer pair 6, PR7—primer pair 7 and PR8—primer pair 8.

Table 2 presents the primer sequences used. Standard RT-qPCR was performed using the Superscript III one-step RT-qPCR kit according to the manufacturer’s instructions. Amplified fragments were analysed by electrophoresis on a 1.2% agarose gel.

Table 2. Sequence of primer pairs used for identification of protein S, by conventional PCR technique.

*Spike protein segment (S protein)—Primer identification; **Size of the amplified fragment in base pairs (bp).

Transmission electron microscopy was performed using 50 µL aliquots of the samples fixed in glutaraldehyde (2.5%) and supplemented with 2% uranyl acetate (final concentration), followed by negative staining.

Copper grids, 400 mesh were coated with formalin and sprayed with carbon for stabilization. Subsequently, the samples were adhered to the grid for 30 sec had their humidity removed with film paper, and the negative contrast was evaluated by adding uranyl acetate (5%), incubating for 30 s, followed by drying with filter paper.

Transmission electronic microscopy images (Hitachi 7800 - 4K) with an acceleration voltage of 100 kV were observed at magnifications of 20,000 to 120,000 x, captured and evaluated using a digital system and the manufacturer’s software. The characterization of viral particles follows parameters as morphology, size, and structural uniformity [14].

Statistical analysis was performed using GraphPad PRISM 8.0.1 software. The Shapiro-Wilk test was used for normality testing, without subsequent analysis of the normal (Gaussian) distribution. Non-parametric analyses were performed and considered significant at p ≤ 0.05 (*). Qualitative results were presented as detection percentage values, and multinomial linear regression was used to compare physicochemical, bacteriological, and SARS-CoV-2 data. The non-parametric Spearman correlation test was used to evaluate the confidence interval of the correlation coefficient between bacterial and viral detection and physicochemical parameters, being considered significant when p ≤ 0.05.

As shown in Table 4, the brackish water matrix presented the highest observed concentrations of total coliforms (5.5 × 10³ MPN/100mL), thermotolerant coliforms (2,25 × 10³ MPN/100mL) and E. coli (3.15 × 10³ MPN/100mL) among the analysed matrices. Enterococci (36,2 × 10³ MPN/100mL) were found in the highest concentration in untreated hospital raw sewage matrices and treated domestic sewage matrices. The lowest concentration of thermotolerant coliforms (1.52 × 10³ MPN/100mL) was observed in untreated hospital raw sewage matrices. The freshwater river, treated, and untreated domestic wastewater matrices recorded the lowest average concentrations of thermotolerant coliform (1.10 × 10³ MPN/100mL). E. coli and enterococci showed the lowest concentrations in the matrices of untreated domestic wastewater (1.10 × 10³ MPN/100mL) and coastal sea water matrices (13 MPN/100mL), respectively.

Of the 45 samples analysed, contamination indicators were detected in 42 samples, with total coliforms being the most prevalent, followed by E. coli, thermotolerant coliforms, and enterococci when considering the overall averages. When we classify the aquatic matrices as suitable or unsuitable according with bacteriological parameters established by CONAMA Resolution 274/00 [12], which considers the presence above the permitted level of one bacteriological agent among the four most important (total coliforms, thermotolerant coliforms and E. coli and enterococci), six of the seven aquatic matrices (coastal seawater, brackish water, freshwater river, untreated domestic raw sewage, treated domestic raw sewage and untreated hospital raw sewage) were classified such as improper.

Table 4. Bacterial indicators of fecal contamination by collection point.

^Ct indicates the detection of SARS-CoV-2.

The presence of SARS-CoV-2 RNA was also analysed by RT-qPCR. SARS-CoV-2 RNA was detected in samples from six of the seven aquatic matrices analysed, whereas the drinking water sample was negative (Table 5).

Table 5. Detection of SARS-CoV-2 RNA by RT-qPCR in each aquatic matrix.

Samples with Ct values within the established detection range were considered positive for SARS-CoV-2 RNA. Selected samples were further analysed by conventional PCR targeting regions of the S gene, followed by 1.2% agarose gel electrophoresis as a complementary molecular analysis (Figure 2).

Figure 2.Electrophoresis on 1.2% agarose gel showing amplification of the S gene fragments. The first marker is the molecular weight pattern, with the highest weight bands near 1Kb are observed. The individual bands are the PR1 - PR8 primer pairs, from left to right in interleaved wells (Dendê beach).

When SARS-CoV-2 RNA detection and Ct values were correlated with bacterial concentrations, total coliforms, thermotolerant coliforms, and E. coli showed a positive correlation with the SARS-CoV-2 RNA detection. In contrast, enterococci showed an inverse correlation withSARS-CoV-2 RNA detection. The correlations between SARS-CoV-2 RNA detection and bacterial and physicochemical parameters were further evaluated. Thus, our results demonstrated a positive association between SARS-CoV-2 RNA detection and the concentrations of total coliforms, thermotolerant coliforms, and E.coli (Figure 3).

Figure 3.Correlation between bacterial concentration and SARS-CoV-2 RNA detection. (T. Coli.—Total coliforms; Thermo. Coli.—Thermotolerant Coliforms; E. coli—Escherichia coli and Enterococci.

Spearman’s correlation (Figure 4) was used to evaluate associations between SARS-CoV-2 RNA detection, bacterial concentration, and physicochemical parameters of aquatic matrices. An association was observed between SARS-CoV-2 RNA detection, pH, and E. coli. No similar association was observed with enterococci, total coliforms, or thermotolerant coliforms. In addition, it was observed that the bacteria showed weak positive correlations with each other. When we established the correlation between physicochemical parameters and bacterial concentration, we observed that bacterial concentration is more influenced by temperature.

Figure 4. Spearman’s correlation to evaluate the relationship between physicochemical and bacteriological parameters and the detection of SARS-CoV-2 RNA. T. Coli.—Total coliforms; Thermo. Coli.—Thermotolerant Coliforms; E. coli—Escherichia coli; pH—Hydrogen ion potential; EC—electrical conductivity; DO—dissolved oxygen and Temp—Temperature.

Electron microscopy results showed that the two particles, approximately 50 nm, are similar in size to enteric viruses (Figure 5(A)). Virus-Like particles (VLPs) of approximately 100 nm, with morphology compatible with coronavirus-like particles, were observed in Figure 5(B). Both Figure 5(A) and Figure 5(B) correspond to the brackish water matrices. VLPs measuring 40 to 60 nm, with different morphologies and dimensions, were also observed in Figure 5(C), while particles of 100nm with morphology similar to SARS-CoV-2 VLP were identified in Figure 5(D), with Figure 5(C) and Figure 5(D) corresponding to the freshwater river matrices. In the untreated sanitary effluent matrix (Figure 5(E)), a structure containing slightly more than 10 particles sizing between 60 and 95 nm, similar to enteropathogens, was found. Figure 5(F) shows a tailed bacteriophage, approximately 50 to 80 nm in size, similar to Siphovirus, identified in the coastal seawater matrix. In the last two images (Figure 5(G)—coastal seawater matrix and Figure 5(H)—untreated hospital raw sewage matrix), particles of 100 to 148 nm and 100 to 130 nm, respectively, compatible with the morphology and size of SARS-CoV-2 VLPs, were identified.

Figure 5. Electron micrograph of environmental matrices. Images A and B—brackish water matrix; C and D—freshwater river matrix; E—untreated domestic sewage matrix; F—coastal seawater matrix; G—coastal seawater matrix; and H—untreated hospital raw sewage matrix.

In addition, Table 6 shows the overall percentage of samples positive for SARS-CoV-2 VLP by RT-qPCR and transmission electron microscopy (TEM), considering the collection site and grouping by aquatic matrices. For the coastal seawater matrix, two collection sites were positive for SARS-CoV-2 VLP in at least one of the samples, both by RT-qPCR and TEM, suggesting the presence of the virus in 22.2% of the matrix. For the brackish water matrices, one sample from each collection point showed viral presence in both RT-qPCR and TEM, representing an aggregate of 33.3% of SARS-CoV-2 VLP.

The freshwater river matrix was positive in only one sample from the same collection point for both RT-qPCR and TEM techniques, with SARS-CoV-2 VLP being present in only 11.1% of the matrix composition. The drinking water matrix was negative for SARS-CoV-2 VLP in both techniques. In contrast, the untreated domestic sewage matrix was positive in at least two samples from each collection point for both techniques, resulting in viral presence in 66.6% of the matrix. When we analysed the treated domestic sewage matrice, only RT-qPCR showed a positive result for SARS-CoV-2 VLP in at least one sample, whereas transmission electron microscopy showed all negative results. For untreated hospital raw sewage matrices, in both techniques, at least one sample from each collection point was positive, totaling a presence in 55.5% of the matrix.

Table 6. Detection of SARS-CoV-2 by RT-qPCR compared with TEM by collection point and aquatic matrix.