Detection of potentially pathogenic enteric viruses in environmental samples from Kenya using the bag-mediated filtration system

Enteric virus environmental surveillance via a highly sensitive method is critical, as many enteric viruses have low infectious doses and can persist in the environment for extended periods. This study determined the potential of the novel bag-mediated filtration system (BMFS) to recover human enteric viruses and pepper mild mottle virus (PMMoV) from wastewater and wastewater-impacted surface waters, examined PMMoV use as a fecal contamination indicator in Kenya, and identified potential BMFS process controls. From April 2015 to April 2016, BMFS samples were collected from seven sites in Kenya (n = 59). Enteroviruses and PMMoV were detected in 100% of samples, and human adenovirus, human astrovirus, hepatitis A virus, norovirus GI, norovirus GII, sapovirus, and human rotavirus were detected in the majority of samples. The consistent detection of enteroviruses and PMMoV suggests that these viruses could be used as indicators in similarly fecally contaminated sites and BMFS process controls. As contamination of surface water sources remains a global issue, enteric virus environmental surveillance is necessary. This study demonstrates an effective way to sample large volumes of wastewater and wastewater-impacted surface waters for the detection of multiple enteric viruses simultaneously.


INTRODUCTION
Waterborne diseases are responsible for approximately 3.4 million deaths per year globally, and enteric viruses are frequently implicated in waterborne disease outbreaks (Gibson ). Enteric virus shedding in stool from infected individuals can contaminate environmental waters, leading to waterborne outbreaks (Gibson ). Enteric viruses can be detected in wastewater and wastewater-impacted surface waters in areas with poor sanitation (Gibson et (Symonds et al. ). PMMoV has been used globally as a fecal contamination indicator in source waters; however, its presence in African waters has not yet been examined (Symonds et al. ).
A novel environmental surveillance method called the bag-mediated filtration system (BMFS) was developed for detection of poliovirus, an enteric virus, from environmental waters and first tested in Kenya (Fagnant et al. ). The BMFS enables in-field collection and filtration of large sample volumes (3-6 L wastewater or wastewater-impacted waters) (Fagnant et al. ; Zhou et al. ). Filters are further processed in the laboratory, including filter elution and secondary concentration (Fagnant et al. ). BMFS advantages include the ability to ship compact cartridge filters rather than large-volume, potentially hazardous water samples, and capacity to process large sample volumes for a high effective volume assayed.
The BMFS results in improved sensitivity for poliovirus surveillance compared to the gold standard method that processes 500-mL by aqueous two-phase separation (Zhou et al. ). This study examines the ability of the BMFS to sample and concentrate a diverse array of enteric viruses simultaneously.
The objective of this study was to (1) determine the potential for the BMFS to recover diverse domestic wastewater-related viruses from wastewater and wastewater-

Nucleic acid extraction
At UP, secondary concentrate aliquots were chloroform extracted, followed by addition of 5 × 10 4 copies of mengovirus (extraction control) and nucleic acid extraction. The semi-automated NucliSENS ® easyMAG ® instrument and accessory products (bioMérieux SA, Marcy-I'Étoile, France) were used according to manufacturer's instructions, with a 1,000-μL input volume and nucleic acid elution in 100-μL.
BMFS secondary concentrate aliquots were shipped to United States Centers for Disease Control and Prevention where they were processed for PMMoV analysis. Samples were chloroform extracted and RNA extraction was performed using QIAamp ® Viral RNA Kits (Qiagen, Hilden, Germany) according to manufacturer's instructions, with an increase in input volume to 280-μL, and RNA elution in 60-μL. RNA extracts were shipped to the University of Washington (UW) for PMMoV analysis.

RT-PCR
At UP, samples were analyzed for EV, HAstV, HAV, NoV GI, NoV GII, and HRV by direct real-time RT-PCR using CeeramTools ® kits (bioMérieux) according to manufacturer's instructions. In-house real-time PCR assays for HAdV were conducted using TaqMan      on some sampling days, including 28 July 2015 and 11

RESULTS AND DISCUSSION
August 2015 (p < 0.05) ( Figure S8, available online). The eight enteric viruses were detected at similar rates between all sites (p > 0.05).
Of the enteric viruses detected, HRV was detected the least frequently (Table 1) The frequent detection of eight enteric viruses throughout this study in samples collected in Mombasa, Garissa, and Kisumu suggests the BMFS is a robust method that could be used in a variety of situations (Table 1). Samples were filtered in the field, shipped to Nairobi on cold chain

PMMoV detection and quantification
PMMoV was detected in 100% of BMFS samples and cooccurred with the presence of human enteric viruses (n ¼ 59; Table 1). PMMoV was detected with similar Cq values as EV (Tables 2 and 3)   sampled (p > 0.05). The average estimated PMMoV concentration was 6.1 × 10 5 gene copies L À1 sample (Table 3). This is within the range seen for contaminated surface waters (10 1 -10 8 targets L À1 ), though slightly lower than the range for untreated domestic wastewater (10 6 -10 10 targets L À1 )