Quantitative RT-PCR Detection of Human Noroviruses and Hepatitis A Virus in Surface Water and Fresh Produce

Vegetables can be considered as a vehicle for the transmission of human enteric viruses such as noroviruses (genogroups I and II) and hepatitis A virus (HAV) when irrigated with contaminated irrigation or when prepared by infected food handlers. In the current study, we investigated the presence of hepatitis A virus (HAV) and human noroviruses (genogroups I and II) in fresh produce and surface water used in cultivation of this produce using real-time PCR. Samples were collected from six different points in the Mansoura and Giza regions, Egypt. Our analysis showed that at least one virus was found in 41.6% (30/72) of surface water samples and 27% (13/48) of fresh produce samples. HAV (23/72) with a mean viral concentration = 4 × 10 6 genome copies/litter (GC/L) was the most frequently identied virus in surface irrigation water samples, followed by human norovirus genogroup II (HNoV GII) (15/72, with a mean concentration = 1.2 × 10 6 GC/L, and human noroviruses genogroup I (HNoV GI) (12/72, with a mean concentration = 1.4 × 10 4 GC/L). Additionally, HAV (10/48) with a mean concentration = 5.2 × 10 5 genome copies/gram (GC/g) was also the most frequently detected virus in the fresh produce samples, followed by HNoV GII (8/48, with a mean concentration = 1.7 × 10 4 GC/g), meanwhile HNoV GI (6/48) was less detected virus with a mean concentration = 3 × 10 3 GC/g. This work suggests a wide prevalence of human enteric viruses in surface irrigation waters and fresh produce, which is of concern when the fresh produce is eaten raw. Thus, Additional monitoring for viral pathogens in irrigation water and food is needed to increase produce safely.


Introduction
Water and food-borne outbreaks reported in 2018 in Europe were related to viruses (13.5% of all outbreaks), bacterial toxins (24.2% of all outbreaks), and bacterial agents (57.0% of all outbreaks). Viruses are known as important emerging infectious agents of food-borne illnesses (Bányai et al., 2018). In Egypt, the presence of enteric viruses, such as hepatitis A and E viruses, norovirus, rotavirus, adenovirus, human noroviruses, and human astrovirus has recently been reported in river water used in irrigation purpose (Elmahdy et al., 2020; Shaheen et al., 2020Shaheen et al., , 2019Shaheen et al., , 2018Elmahdy, 2019a, 2019b). This underlines that irrigation water can play a role in transferring of these viruses to fresh produce.
Hepatitis A virus (HAV) and human noroviruses (HNoV), are recognized as major public health concerns where they are responsible for the majority of acute hepatitis or non-bacterial gastroenteritis which are occasionally fatal (Vaughan et al., 2014;Koo et al., 2010). HAV belong to the Picornaviridae family while HNoV is classi ed in the Caliciviridae. Both viruses possess a single-stranded and positive-sense RNA genome as they are small non-enveloped viruses. HNoV which was classi ed into ten genogroups (HNoV GI- Viral contamination of irrigation, ground, surface water, and soil has been investigated in previous studies Santamaria and Toranzos, 2003). To ensure the safety of water and fresh produce, the development of accurate and reliable quantitative methods for detecting the presence of enteric viruses in water and food is needed. The aim of this study was to investigate the presence of HAV and HNoV in fresh produce and to establish whether the river water used in irrigation is a potential source of fresh produce contamination.

Material And Methods
Sampling A total of 72 surface water samples and 48 fresh produce (24 green onions and 24 lettuces) samples were collected from three sites at two regions (Giza and Mansoura) in Egypt. Three water samples were collected monthly from each region during the period March-2019 to February-2020. Also, 3 fresh produce samples from each kind were collected monthly from the same sites at each region according to the harvest season (November-2019 to February-2020).

Determination of virus recovery from fresh produce and water samples
In this study, we used murine norovirus (MNV-1) as a process control virus to investigate the virus recovery in all virus detection assays. In brief, autoclaved surface water samples (5 L) were inoculated with 200 μl of MNV-1 suspension (4.7 × 10 8 GC/mL) prior to concentration. Also, leaves of green and lettuce were chopped into small pieces then fty grams of each sample was weighed and sanitized in chlorinated water for 20 min. After sanitation, the leaves of each fresh produce were divided into two equal parts and each part was placed into a sterile bag under aseptic conditions. One portion of the fresh produce sample was inoculated with the same amount of virus prior to concentration, while the other portion was left as blank. The e ciency of MNV-1 recovery from inoculated samples was assessed using the protocols described below.
Furthermore, a representative sample was taken from green onion and lettuce collected during this study and inoculated with human MNV-1 as sample process control virus (SPCV) (previously tested negative for MNV-1 by RT-qPCR).

Virus concentration
Virus concentration from surface water samples was performed by using a protocol described by Katayama et al. (2002) while virus elution from fresh produce samples was done according to ISO/TS 15216-1 method (ISO 15216-1:2017 2017). Brie y, 25 g of green onion or lettuce in small pieces was mixed with 40 ml of Tris-glycine buffer (100 mM Tris-HCl, 50 mM glycine, and 1% beef extract, pH 9.5) in a sterile plastic bag. After 20 min at room temperature with constant rocking (approximately 70 oscillations/min) to elute the viruses from the surface of the fresh produce sample, the sample was distributed into clean centrifuge tubes. After centrifugation at 10,000×g for 30 min at 4 °C, the vegetable matter was discarded and the eluate was transferred into clean centrifuge tubes then the pH was adjusted 7.2 ± 0.3 using 1.0 N HCl. For the PEG precipitation, 0.25 volume of 50% (w/v) polyethylene glycol 8000/1.5 M NaCl were added to the eluates and incubated on rocking at 120 rpm at 4 °C for 1 h. After additional centrifugation for 30 min at 10,000×g at 4°C , the resulting pellets were dissolved in 500 μl of 10 mM PBS then stored at − 20 °C until use.

RNA extraction and virus quanti cation
The nal concentrates were used to extract viral RNA using a  2015), respectively, using onestep Rotor-Gene Probe RT-PCR Kit. All ampli cations were conducted in duplicate using Rotor-Gene system (QIAGEN, Germany). For each assay, PCR amplicon for each positive control was generated by cloning the amplicon into a plasmid (pGEM-T Easy Vector (Promega) for HAV strain HM175; and pCR2.1-TOPO vector (Promega) for HNoV GI, GII, and MNV-1), and concentrations of a puri ed plasmid DNA were calculated using a Nano Drop spectrophotometer (Thermo Fisher Scienti c, USA). Standard curves were prepared by tenfold serial dilutions of the positive control plasmids. Ultra-pure water was used as negative controls in each assay to ensure that there was no cross-contamination in the assay. To increase PCR e ciency, serial ten-fold dilutions were prepared to viral nucleic acids to dilute the inhibitors (if found). Viral nucleic acid was ampli ed in a 25 μL Real-time PCR mixture containing 5 μL of the RNA extract, 12.5 μL RT-PCR Master Mix (Qiagen, Germany), 400 nM each primer, 250 nM of probe, and nuclease-free water to complete the volume reaction up to 25 μL. This mixture was transferred into a 48-well microplate then loaded into the Rotor-Gene system. Fluorescence data were measured at the end of annealing step. All primers and probes used in this study, as well as the thermal cycling conditions of each virus, are shown in Table 1. The percentage of MNV-1 recovery rate from spiked water or fresh produce samples was calculated by using the following formula: amount of virus detected after spiking experiments/amount of viral inoculum X 100 (Hennechart-Collette et al., 2021).

Statistical analysis
Statistical analyses were performed using GraphPad Prism version 5.0 (USA) software. The critical P value for the test was set at < 0.05. The Pearson correlation was applied to evaluate the correlations between viral distributions and water/fresh produce samples. The mean viral load was compared between samples with a one-way analysis of variance. The detection limits of the qPCR was established according to the highest dilution where it was possible to virus quanti cation (The detection limits for HAV were 2.5 × 10 1 genome copies in 25 mg of fresh produce and in 10 L of water whereas the detection limit for both HNoV GI and HNoV GII were 2 × 10 1 genome copies in 25 mg of fresh produce and in 10 L of water).
Detection of HAV and HNoV in irrigation water samples Mansoura region than those collected from Cairo region ( Table 2). This variation was statistically signi cant (P ≤ 0.01).

Distribution of single and multiple viral agents in the positive water samples
As shown in Table 4 In the green onion samples, single viral agent was detected in 25% (6/24) of positive samples. HAV, HNoV GI, and HNoV GII were detected in 50% (3/6), 16.6% (1/6) and 33.3% (2/6) as a single viral agent in the positive samples, respectively. Only one positive sample contained two viral agents (HAV + HNoV GII) while two samples contained the three viral agents (HAV+ HNoV GI+ HNoV GII). In the lettuce samples, contamination with single viral agent was found in 20.8% (5/24) of the positive samples. HAV (3/5) was the most frequently found alone in these samples, followed by HNoV GI (1/5) and by HNoV GII (1/5). Contamination with two viruses was detected only in one positive samples that contained both HNoV GI and HNoV GII. Also, combination of the three viral was detected in 14.3% (1/7) of the positive samples (Table 4).

Discussion
Enteric viruses can be transferred to fresh produce by various routes during growth, harvest, packaging, . Indeed, monitoring these viruses in both the fresh produce and irrigation water can provide a useful tool to reduce the risk of foodborne disease linked to the occurrence of these viruses. In this context, the aim of this study was to investigate the presence of HAV and HNoV in irrigation waters and fresh produce irrigated by this water and collected from two regions (Giza and Mansoura) in Egypt.
HNoV was detected in 25% (30/120) of both matrices where HNoV GII (19.2%, 23/120 samples) was higher than HNoV GI (15%, 18/120 samples). In contrast, the detection rate of HNoV GI was higher than HNoV GII In this study, HAV detection in both matrices was higher than HNoV prevalence. This nding is similar to our previous study conducted on the same two regions (Shaheen et al., 2019), suggesting that surface irrigation water could be the source of fresh produce contamination. Since irrigation water can transmit the viral contamination for fresh produce, thus primary products must be produced only in regions where an appropriate water quality is used for irrigation purposes. If the quality control of irrigation water is not controlled, prior consumption, it is suggested to immerse the fresh produce with drinking water containing sodium hypochlorite (15-20 ppm free chlorine levels) for two minutes to reduce viral loads on the vegetable surfaces (Bosch et al., 2018).
This the rst report of HNoV GII detection in fresh produce in Egypt. However, this study has some identi ed HNoV (GI and GII) in 30% of diarrheal specimens collected from Mansura City. A recent study from Mansura region, HNoV was also detected in 70.5% of diarrheal samples with HNoV GII as the prevalent genotype (Zaki et al., 2019). This ndings agrees with our study that HNoV GII being the most prevalent genotype. In Cairo region, HNoV (GI and GII) and HAV were also identi ed in samples collected from patients with severe diarrhea or hepatitis-related symptoms (Kamel et al., 2011;2009). The HNoV genotypes as well as HAV detected in the clinical samples are also detected in fresh produce, suggesting that viral contamination of the fresh produce could be originated from contact with contaminated irrigation water.
In the current study, the viral load of HAV was higher than HNoV in the surface water and fresh produce samples. Comparatively, HAV had high variability in the viral loads, where some fresh produce samples had as low as 1.1 x 10 2 GC/g up to 8.3 x 10 7 GC/mL detected in one of the surface irrigation water samples tested. In surface water samples, most of the positive samples had viral concentrations between 10 4 and 10 5 GC/L while between 10 3 and 10 4 GC/g in fresh produce samples. High variability in the viral loads was also observed for HNoV where some fresh produce samples had as low as 2.2 x 10 2 GC/g which increased to 7.8 x 10 6 GC/mL detected in one of the surface irrigation water samples tested. Most of the positive surface water samples had HNoV concentrations between 10 3 and 10 4 GC/L. However, the HNoV in fresh produce samples had low variability with viral loads ranged from 2.2 x 10 2 to 8.2 x 10 4 GC/g. The low prevalence of HAV and HNoV with low concentrations in fresh produce than surface water may be due to their direct exposure on fresh produce surfaces to the ultraviolet radiation emitted from the sun.
The low viral concentrations found in the positive fresh produce and surface samples are by far higher than the infective dose required to induce disease by most of the enteric viruses, ranging between 10-100 viral particles for human rotavirus and HAV and even less for human HNoV (Yezli and Otter 2011). Thus, this low infective dose represents a potential risk for a viral outbreak if these fresh produce reach the consumers. In comparison, another study from Mexico found that the mean viral loads of HAV ranged from 2.8 x 10 2 to 2.4 x 10 3 GC/g while HNoV loads ranged from 2.1 x 10 2 to 1.3 x 10 3 GC/g (Felix-Valenzuela et al., 2012), which is lower than viral concentrations detected in this study. All precautions were applied in this study to prevent cross-contamination in the RT-qPCR reactions, and no ampli cation was found in negative controls.
In conclusion, The detection of HAV and HNoV both in surface irrigation waters and associated fresh produce, demonstrate that viral contamination of the fresh produce could be arisen, at least partially, during the production phase in the farm. Furthermore, the results obtained in this work highlight the importance of viral surveillance programme for fresh produce, due to the low infectious dose of most the gastroenteric viruses. We expect that this work will contribute to the management and control of microbiological risks in fresh food, reducing the public health risks from the eating of the contaminated foods.   M, refers to sample collected from Mansoura; G, refers to sample collected from Giza; -, refers to negative samples