Detection of Protoparvovirus in wastewater and human adenovirus in a green leafy vegetable in an Environmental Education Center in southern Brazil

Hydric systems have frequently suffered negative impacts due to the lack of water resource management and sanitation. This scenario occurs mainly because of sewage disposal in springs and soil, and consequently can cause waterborne diseases that affect humans and many animal species (Bacci & Pataca, 2008). According to data from the National Information and Sanitation System for 2019, only 61.9% of the Brazilian population is served by sewage collection in urban areas. In addition, of all the sewage generated, 78.5% was collected, of which 49.1% received treatment (Sistema Nacional de Informações sobre Saneamento, Brazil). Tracking microbiological contamination can be used strategically for hydric basin monitoring and management. Coliforms from soil, Escherichia coli, which is solely of fecal origin (warm-blooded species), and viruses from human or animal enteric tracts can be used as markers. All of these have previously been defined as environmental biomarkers (Buzanello et al., 2008; Rames et al., 2016).

Viruses present in water can cause gastroenteritis in both humans and animals. They can easily spread in the environment due to infiltration into the soil, thereby reaching groundwater sources (Rames et al., 2016). In addition, water used for irrigation can cause food contamination, particularly in vegetables that are eaten raw (Kokkinos et al., 2012). In Sicily, a study investigated the contamination of vegetable samples (n = 70; including Swiss chard, chicory, parsley, celery, escarole, endive, and lettuce) and detected Norovirus (NoV) and Enterovirus (EV) in 2.9% of the samples and hepatitis E virus (HEV) in 1.4% of the samples (Purpari et al., 2019). In Argentina, NoV, Rotavirus (RV), EV, and human astrovirus were detected in green leafy crops (lettuce, spinach, chicory, and arugula; Prez et al., 2018). In Brazil, to the authors’ knowledge, there has been only one study identifying RV type A (RVA), human Mastadenovirus (HAdV), hepatitis A virus (HAV), and NoV in vegetables (cucumber, spring onions, lettuce, and arugula) (Carvalho et al., 2020). RVA was detected in arugula, cucumber, and spring onions, and HAdV in cucumber species (Carvalho et al., 2020).

Adenoviruses (AdVs), apart from EV, HEV, HAV, NoV, and RV, are commonly utilized in environmental monitoring studies; however, they are usually used as a reference for bioindicators in different environmental matrices (Smith et al., 2010). AdVs can be excreted in large quantities by infected hosts such as humans and animals, and they have relevant resistance in the environment. Nevertheless, virus markers are not yet a requirement for public water policy in Brazil (Prado & Miagostovich, 2014).

AdV is a non-enveloped virus with double-stranded DNA belonging to the Adenoviridae family, which is divided into five genera. HAdV belongs to the Mastadenovirus genus and is classified into seven species (A–G) (Han et al., 2013). HAdV can infect a range of tissues, with the intestinal and respiratory tracts being the most affected (Straus & Straus, 2008). HAdV-C, B, and E are responsible for respiratory diseases, and they can be excreted in feces from carrier hosts (Roy et al., 2009). Although HAdV-F (gastrointestinal) is used for tracking environmental sources, HAdV-C and -E are commonly found in environmental matrices (Bibby & Peccia, 2013). Mastadenovirus can also infect animal species such as dogs. Canine mastadenovirus (CAV) is subdivided into Type 1 and Type 2 (Benkő et al., 2002). CAV-1 induces a systemic disease that causes hepatitis; CAV-2 replicates in the respiratory epithelium when co-infected with Bordetella sp. and canine parainfluenza virus, and can evolve into kennel cough (Buonavoglia & Martella, 2007). Transmission may occur through direct or indirect contact, via fomites (water and food), or from other species (Hu et al., 2001).

The presence of parvovirus in environmental matrices is not commonly investigated, even though it has the same transmission mechanism as the viruses mentioned above. Only a few studies investigating parvovirus genomes from animal and/or human species in the environment have been reported (Blinkova et al., 2009). Human bocavirus (HBoV) has been detected in effluent samples in the USA (Blinkova et al., 2009). In Brazil, HBoV DNA has been revealed in surface water in the metropolitan region of Rio Grande do Sul (RS) State in southern Brazil (Kluge et al., 2013). In Spain, chicken parvovirus and turkey parvovirus were found after mapping fecal contamination of avian origin (Carratalá et al., 2012). These viruses have not yet been reported in Brazil.

Parvoviruses are small non-enveloped viruses with a single-stranded DNA genome that can infect a wide range of hosts (mammals, birds, and reptiles), including the order Carnivora (ICTV, 2019). Carnivore protoparvovirus 1 is the most important parvovirus that infects carnivores, and includes CPV, feline panleukopenia virus, and mink enteritis virus (Steinel et al., 2001). CPV belongs to the genus Protoparvovirus, subfamily Parvovirinae, and family Parvoviridae. CPV is classified into Type 1 and Type 2 (CAV-1 and CAV-2, respectively). CAV-1 can cause myocarditis in neonatal pups and CPV-2 is responsible for severe gastroenteritis. Genetic alterations have occurred in CPV-2a, CPV-2b subtypes, and more recently CPV-2c (Buonovaglia et al., 2001). Infection with different types of CPV usually occurs through oronasal exposure, feces, and contaminated environments. Insects, rodents, humans, and fomites can spread this virus, which can infect both vaccinated and unvaccinated dogs (Decaro et al., 2007; Strottmann et al., 2008).

The main objectives of this study were to (a) analyze the presence of E. coli, TC, AdVs (human and canine species), and CPV in different sources of EEC water; (b) investigate the presence of AdV in lettuce irrigated with water sources available in the EEC; and (c) investigate CAV and CPV in fecal samples from dogs kept in the kennel.

The study was conducted at the EEC (coordinates 29°44′27.1″ S and 51°03′06.2″ W) and in the city kennel (coordinates 29°44′20.06″ S and 51°03′03.4″ W; see Figure 1) in mid-October 2017 and late February 2018 in spring and summer, respectively. In both periods, surface water, wastewater, dog feces, and lettuce were collected.

Water samples were collected from one stream, three weirs, and one wastewater facility (animal shelter septic tank), which were designated Point 1–Point 5 (P1–P5; see Figure 1). All water samples were aseptically collected in 500 mL sterile flasks, packed in isothermal boxes, and sent to the Laboratory of Molecular Microbiology (LMM) of Feevale University RS State, Brazil, for further processing. Samples from P1 to P5 collected in 2017 were labeled LMM 3940–3944, and those collected in 2018 were named LMM 4160–4164.

The dog feces were collected in the kennel as soon as the dogs were excreted (see Figure 1 and the epidemiological aspects in Table 1). Dog handling procedures agreed with all ethical principles for animal experimentation: the 01.17.055/2016 protocol approved by the Ethics Committee of Animal Use of Feevale University (CEUA/Feevale). Dog fecal matter collected in 2017 was labeled LMM 3948–3965 and 3970–3971, and that collected in 2018 was named LMM 4165–4177.

Lettuce (Lactuca sativa) samples were collected from the EEC garden (Figure 1) and were watered through the irrigation system at P2. Lettuce samples were designated H1–H4, and three parts of each plant were analyzed (leaf, stem, and root) and renamed in the format H1L, H1S, and H1R. Each plant (H1–H4) generated three samples, totaling 12 samples per collection and 24 samples in total. Lettuce samples were named LMM 3966–3969 and 3972–3979 for the 2017 samples, and LMM 4312–4323 for 2018.

TC and E. coli were detected using the Colilert^® substrate enzyme method (IDEXX^® Laboratories, Westbrook, USA) according to the manufacturer's instructions. All the water samples were tested within 24 h of collection. The sample was positive for TC if the large and small wells of the tray showed yellow in natural light and positive for E. coli if the wells fluoresced blue following exposure to 300 nm UV light. Negative results were observed in the absence of color and/or fluorescence, respectively. The results were expressed as the most probable number in 100 mL of water (MPN/100 mL) according to the table provided by the manufacturer.

Samples were concentrated by applying the ultracentrifugation method, in which 36 mL of each was centrifuged at 10,000 × g at 8 °C for 3 h. Pellets were resuspended in 1 mL of Tris-EDTA buffer (pH 8.0) and homogenized under vigorous agitation for 1 min according to the standard protocol described in previous studies (Girardi et al., 2018). The final volume was aliquoted, and DNA was extracted.

For the processing of feces samples, 0.2 g of fecal matter was diluted in 1 mL of Eagle's Minimum Essential Medium, vortexed for 1 min, and centrifuged for 3 min at 10,000 × g. From the supernatant, 1 mL was used for viral DNA extraction according to Heldt et al. (2016).

The lettuce samples were chopped and macerated. Subsequently, 1 g of each sample was placed in a 15 mL reaction tube and 9 mL of phosphate-buffered saline buffer was added. The tube was placed into an incubator at 22 °C with constant stirring at 2,000 × g for 1 h. Subsequently, 1 mL of supernatant was aliquoted and sent for DNA extraction using the protocols of the LMM that have been used to analyze HEV in processed pork following the protocol of Heldt et al. (2016).

Viral DNA was extracted using the Promega^® extraction kit following the manufacturer's instructions. Aliquots of 200 μL of each sample were used, and the final elution was performed in microtubules free of DNAse and RNAse using the polymerase chain reaction (PCR). All steps required careful criteria to prevent the contamination of the samples.

Nested-PCR, a DNAPol target gene, was used to detect different AdVs in all samples. In the first round, 1 μL of each primer pol-F (5′-CAGCCKCKGTTRTGYAGGGT-3′) and pol-R (5′-GCHACCATYAGCTCCAACTC-3′) were used, both with 20 pmol concentrations, along with 18 μL of DNAse- and RNAse-free water, 25 μL of mix (Promega®), and 5 μL of extracted DNA, totaling 50 μL of reaction volume. In the second round, the same reagents and volume were used as that of the first round, and the 5 μL of DNA extracted from the product of the first PCR was replaced, as well as the set of oligonucleotides for the oligonucleotides sense pol-nF (5′GGGCTCRTTRGTCCAGCA-3′) and reverse pol-nR (5′-TAYGACATCTGYGGCATGTA-3′) (Li et al., 2010). The resultant amplicon was approximately 300 bp. In the first round, negative and positive controls, RNAse- and DNAse-free water, and HAdV-41 were used, respectively, and both samples were used again as templates in the second round.

The DNA extracted from dog fecal samples was concomitantly subjected to PCR specific for CAV and CPV. A positive control was used for CAV analysis detected by the previous LMM sample. The standard PCR for CAV was performed in a final volume of 50 μL containing 5 μL of DNA diluted in 25 μL of mix (Promega^®), 18 μL of water, and 20 pmol of each primer. The primers used for canine AdV were CAV-F1, 5′-CACGATGTGACCACTGAGAG-3′, and CAV-R1, 5′-GGTAGGTATTGTTTGTGACAGC-3′ (20 pmol dilutions). The resultant amplicon was 300 and 350 bp of the gene encoding the CAV-1 and CAV-2 hexon protein, respectively (Monteiro et al., 2015).

For CPV detection, the Vanguard^® HTLP 5/CV-L vaccine, which contained the target canine viruses used in this study, was used as a positive control. Reactions of conventional PCR for CPV were performed using the same final volume as that used for CAV. The primers for CPV were CPV-555-F, 5′-CAGGAAGATATCCAGAAGGA-3′ and CPV-555-R, 5′-GGTGCTAGTTGATATGTAAT3ACA-3′ (Buonavoglia et al., 2001). The generated PCR product was 555 bp of the capsid protein gene (VP2, viral protein).

All positive PCR samples were subjected to purified assays using the PureLink^® kit (Invitrogen) according to the manufacturer's instructions, followed by sequencing. The ABIPrism 3100 system/company ACTGene equipment was used to characterize the species and/or type.

The resulting nucleotide sequences were analyzed using the CAP3 program implemented in BioEdit 7.0.5. Alignments were performed using Clustal Omega (Sievers et al., 2011). Phylogenetic trees were obtained using the neighbor-joining method (Saitou & Nei, 1987) and Kimura 2 (1980) MEGA7 software (Kumar et al., 2016).

A total of 51 samples were collected, including 33 fecal, 8 surface water, 2 wastewater, and 8 lettuce samples. Although 51 samples were collected, 67 were analyzed, as each lettuce plant generated three samples, totaling 12 samples per collection and 24 in total.

E. coli and TC were detected in all surface water and wastewater samples during at least one of the collection periods.

For the viruses examined in this study, at least one was detected in 29.8% (20/67) of the samples analyzed. Co-detection was observed in one wastewater sample (CAV-1 with Raccoon CPV-like) and two dog feces samples (CAV-1 with CPV-2a).

DNA AdV was detected in all types of samples analyzed and was observed in 13.4% of the total (9/67). Human AdV type C (HAdV-C) was detected in one water sample and E (HAdV-E) in three water samples. One sample of water positive for HAdV-E was co-contaminated with HAdV-C. HAdV-E was also detected in one lettuce sample (stem). CAV-1 was found in two fecal samples and one water sample, and one AdV that was not characterized was detected in dog feces.

Different concentrations of TC and E. coli were observed in each sampling period (2017 and/or 2018) and at all points (P1–P5). In samples in which it was impossible to determine the MPN for E. coli and TC, the test was performed without dilution (Table 3).

Table 2 shows the PCR results for AdV and CPV detection in water samples. P1 and P2 were negative for both viruses during both sampling periods (2017 and 2018). In samples collected in 2017, HAdV-C was detected at P3, P4, and P5. In 2018, HAdV-E was found at P3, and co-contamination with CAV-1 and a virus from wild animal species (Raccoon CPV-like) was observed at P5.

Viruses were detected in 57.5% (16/33) of the fecal samples. Considering the number of animals, this detection ratio falls to 42.4% (14/33) because two animals (one female and one male) presented co-infection with CAV-1 and CPV-2a (see Table 1).

Three of the four lettuce samples were irrigated with water from P2. The control was irrigated with cistern water (EEC in Figure 1). Only one sample (stem) collected in 2018 was positive for HAdV-E: LMM 4319 (see Figure 2).

The nucleotide sequences resulting from the samples that were positive for Mastadenovirus and Protoparvovirus were compared to 38 and 24 sequences from the GenBank databases. The phylogenetic tree from the 8 AdV (human and canine species) and 14 Protoparvovirus characterized is shown in Figures 2 and 3, respectively. One AdV that could not be characterized was not included in the analysis.

This study reveals the presence of coliforms, AdV, and CPV in water; AdV and CPV in dog feces and AdV in lettuce in an EEC that contains a city kennel. To the best of our knowledge, the detection of an animal Protoparvovirus in wastewater samples and HAdV-E in lettuce have not been observed previously in southern Brazil.

When we analyzed each water collection point, TC and E. coli showed different values in October 2017 and February 2018. Rain occurred prior to the October collection, which could account for the high values observed in P1 (see Table 3). This is not only due to the infiltration of soil surface residues but also relates to the source of this stream (P1). P1 is in a seaside resort whose upper bed contains pig, sheep, and bird farms. Pigs are an environment-degrading species due to their large volume of excrement (Druzian et al., 2007). However, before the second collection in 2018, there was a drought period, which may be related to the lower concentration of TC and E. coli in P1 (Table 3). No viruses were detected at this location.

A reduction in TC and E. coli values in the second sampling was also observed at P2 (weir, irrigation pond) due to the same pluviometric influence observed at P1. P2 acts as a nursery for carp (Cyprinus carpio), which feed on organic matter from vegetation and sediments (Reidel et al., 2004). Although this fish has varied eating habits and is widely used to control aquatic macrophytes (Marques et al., 2004), its presence may have influenced the low concentrations of coliforms and the absence of viruses. However, P2 was the origin of the lettuce irrigation system, in which HAdV-E DNA was detected in a stem (LMM4319). We cannot confirm that the contamination by HAdV-E in lettuce was related to the water used in irrigation, because in P2, which was the source of irrigation, no virus was detected.

Lettuce is normally consumed raw and its contamination may represent a potential risk of enteric infection. A study performed in South Korea detected AdV in raw lettuce, chicory, and spinach irrigated with groundwater. The infectivity assay by cell culture PCR (ICC-PCR) was positive only for spinach species (Cheong et al., 2009). However, this tool was not used in the present study. These results could provide information on Brazilian water policy for irrigation systems, because green leafy vegetables with viruses may result in clinical or subclinical waterborne diseases that directly impact public health.

At P3 and P4, coliform concentrations were lower than those at other points in the first sample collection (2017). However, HAdV-C was detected in both sampling periods, and HAdV-E was found in samples from P3 in 2018 (see Table 2). This detection of coliforms with viruses has been commonly reported but is not clearly understood (Cheong et al., 2009; Comerlato et al., 2011). The presence of HAdVs in these water samples may be related to anthropogenic circulation such as through employees, visitors, and hikers. Fluctuations in water volume due to rainfall rates and the presence of Eichhornia crassipes (water hyacinth) could also have contributed to this outcome. The water hyacinth has a high nutrient retention capacity, which may cause eutrophication of aquatic environments, providing the protection of the viral particles against environmental adversities such as solar rays, temperature, and pH (Alves et al., 2003).

The critical location in both samplings was P5 (animal shelter septic tank) where high levels of TC and E. coli were observed (Table 2). In addition, P5 showed viruses not yet detected at other points, such as CAV-1 and Raccoon CPV-like (LMM4164) (Table 2 and Figure 3). A study of postmortem intestinal tract samples from raccoons (Procyon lotor) in Canada (2009–2017) resulted in the epidemiological and molecular characterization of protoparvoviruses (Carratalá et al., 2012). The phylogenetic tree constructed in the present study grouped LMM4164 in the same clade as the raccoon (see Figure 3), since MF069443 and MF069444 were characterized as CPV-like by Canuti et al. (2017). Therefore, it is probable that the isolated LMM4164 was derived from raccoons. However, another virus, Raccoon protoparvovirus (RPV), which is a member of the Protoparvovirus genus, has been isolated from postmortem or other biological samples from the same animal species (Kapil et al., 2015; ICTV, 2019). As this virus was found in water, likely from the washing of the kennel stalls, we can suggest that the origin comes from wild species of raccoons that exist in the region. These animals circulate in the area and may be infected and threatened due to the virulence of this virus, or perhaps the dogs have been previously infected by the raccoons circulating in the area; therefore, we have designated it as Raccoon CPV-like.

In southern Brazil, a raccoon is known as a guaxinim. It is a nocturnal omnivore that has adapted to urban habitats. Raccoons prefer to live near riverbanks and close to crops and farms that provide shelter and food. The proximity to domestic animals creates transmission between the two groups. They are susceptible to diseases that affect carnivores, especially canine distemper, rabies, canine adenovirus, leptospirosis, and parvovirus (Kapil et al., 2015). This description strengthens the hypothesis that the Raccoon CPV-like might have come from raccoon species and not from domestic dogs.

Raccoons are the main hosts for the spread and evolution of Carnivore protoparvovirus, highlighting their role in the emergence of CPV-2a in dogs (Canuti et al., 2017) due to two key mutations in VP2 (87Leu and 101Thr) (Allison et al., 2012). These CPV-like strains originated from multiple cross-species transmissions, followed by sustained transmission and evolution in raccoons (Allison et al., 2013).

Another important finding that received our attention was the detection of carnivorous viruses in P5 in 2018 (the second sampling). Two animals tested positive for CAV-1, which was also detected in the water (Tables 1 and 2). This resulted in the main hypothesis that the origin of CAV-1 in P5 was from dogs kept in shelters. In contrast, CPV-2a detected in dogs in the second sampling was not found in P5, and only the Raccoon CPV-like was observed. This finding reinforces our hypothesis that the virus originated from wild animals and not from domestic dogs. In addition, the detection of CAV-1 in surface water and in dog feces during the same period strengthens the idea that viruses were detected in water because there was a host infected with the virus (symptomatic or asymptomatic). Water resources must be protected to avoid the spread of pathogens and the contamination of rivers and streams. Environmental monitoring of these sources to track contamination is of paramount importance for epidemiological mapping for both public and animal health.

The results observed here corroborate those of other similar studies conducted in different aquatic matrices of the same region: river water (Dalla Vecchia et al., 2015), marine waters (Gularte et al., 2019), streams (Peteffi et al., 2018; Girardi et al., 2019), recreational waters (Girardi et al., 2019; Gularte et al., 2019), and surface and groundwaters (Demoliner et al., 2021).

CAV-1, uncharacterized AdV, and CPV-2a were detected in the dog fecal matter analyzed in this study. Although these viruses were detected, no animals became ill (as observed by the authors and reported by caregivers). Prior to the 2017 sample collection, animals were immunized against the target pathogens of the study, mainly CPV. This could be the main cause for the outcome, as only one animal tested positive for viruses (Table 1). Vaccination normally occurs once a year usually in springtime according to caregivers in the EEC city kennel.

The presence of CPV-2a is not surprising because this virus is highly prevalent in canine populations worldwide, including Brazil, regardless of the vaccination status. In addition, it is frequently detected in young dogs, and no gender predilection was found in this study (Table 1).

The results of the present study directly reflect the importance of strengthening and implementing sanitation systems in rural areas. Water is necessary for the existence of life whether human, animal, or plant, and river sources are often found in rural areas such as those of the Rio do Sinos. These sources form the Rio dos Sinos Basin, which is responsible for supplying the population of the metropolitan region of Rio Grande do Sul State in southern Brazil.

The viruses and coliforms found in the waters analyzed in this study reflect only a small percentage of the actual conditions of the water resources. Coliforms are recognized parameters for detecting water potability; however, it is evident that they are not sufficient to measure the microbiological quality of water bodies based on the frequent detection of viruses in water. Therefore, we highlight the importance of utilizing viral analysis for parameters relating to water potability including those in irrigation systems, as is already occurring in developed countries. The water used for crop irrigation requires quality parameters to guarantee a high standard of the yield, particularly for foods normally consumed raw, since pathogens can lodge in the veins and confluence of the leaves.

Environmental preservation affects all parameters of human and animal life and health, and the exploitation of forests and preservation areas forces wild fauna to approach rural residences and breeding sites of domestic animals in search of food resources. This induces disease transmission, where domestic and wild species become carriers of viruses and contribute to the epidemiology of infections. The transmission of viral pathogens by fomites in areas common to these animals may justify, for example, the presence of Raccoon CPV-like in the P5 effluent.

The search for sustainable environmental solutions, such as the biotreatment of effluents, would significantly contribute to reducing the contamination of water matrices. Reducing the spread of pathogens, especially enteric varieties, could contribute to a better quality of life for both humans and animals, reduce human gastroenteritis rates, and improve public health.

We are very grateful to all the LMM for the welcoming and assistance made available at this stage and for the knowledge acquired. A special thanks to the supervisor Andréia Henzel for the excellent orientation and affection.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.