This article was originally published by IWA Publishing. IWA Publishing recognizes the retention of the right by the author(s) to photocopy or make single electronic copies of the paper for their own personal use, including for their own classroom use, or the personal use of colleagues, provided the copies are not offered for sale and are not distributed in a systematic way outside of their employing institution. Please note that you are not permitted to post the IWA Publishing PDF version of your paper on your own website or your institution’s website or repository. Please direct any queries regarding use or permissions to wst@iwap.co.uk

Provided for non-commercial research and educational use only. Not for reproduction or distribution or commercial use.

© IWA Publishing 2014 Water Practice & Technology Vol 9 No 3370 doi: 10.2166/wpt.2014.039

Detection and removal of pathogenic norovirus employing tertiary

treatment in wastewater and water treatment facilities

K. Fitzhenryb,d,*, M. Barretta,d, V. O’Flahertya,d, W. Doreb, S. Keaveneyb, M. Cormicanc

and E. Cliffordd,e

a Microbial Ecology Lab, School of Natural sciences, NUI, Galway

bMarine Institute, NUI, Galway

c Medical Microbiology, UHG, Galway

d College of Engineering and Informatics, NUI, Galway

e Ryan Institute for Environment, Marine and Energy, NUI, Galway

* Corresponding author. E-mail: Kelly.fitzhenry@nuigalway.ie

Abstract

Human pathogenic viruses (HPVs) with particular reference to norovirus (NoV) commonly known as ‘the wintervomiting bug’ is the leading causes of viral gastroenteritis worldwide. NoV is present in wastewater in high num-bers and coupled with a low infectious dose can endanger public health. Secondary wastewater treatmentprocesses are not designed to eradicate HPVs thus a tertiary disinfection step is required. Typical pathogenremoval processes include slow sand filtration and membrane barrier methods. NoV cannot be cultured in a lab-oratory and so its sole detection is limited to expensive and time-consuming molecular methods. The use of analternative virus, the FRNA bacteriophages has been suggested as a surrogate for NoV due to its physicochemicalsimilarities and cultivability. This research focuses on the investigation of a barrier method of tertiary disinfectionusing membrane filters and tangential flow filtration. In addition to this, the suitability of FRNA bacteriophages asa cheaper alternative indicator of NoV was investigated.

Key words: barrier method, disinfection, membrane, pathogen removal, surrogate, virus, wastewater

INTRODUCTION

The release of treated and untreated effluent into water-bodies is a global occurrence in both the devel-oped and developing world. Currently, European regulations do not requiremonitoring of viral loads intreated wastewater. The human pathogenic viruses (HPVs) of main concern with regards to wastewaterinclude the norovirus (NoV), as outbreaks of ‘winter vomiting bug’ can have significant health and econ-omic impacts. In the USA, Hoffmann et al. (2012) estimated food borne NoV to have an economicimpact of about $2 billion in terms of cost of illness and productivity loss each year. In the UK, aNoV outbreak in the winter of 2002/2003 was estimated to have cost the Health Service Executive(HSE) in excess of £100 million (Lopman et al. 2004). Outbreaks of reported cases in Ireland rangefrom 1,500–2,000 annually however it is believed that for every one case reported approximatelythree hundred cases go unreported due to the self-limiting nature of the virus (Tam et al. 2011).Those most at risk in the community often include the elderly and the immunodeficient. The mostcommon transport pathway forHPVs is via the faecal-oral route. They can be present in human effluentandmunicipalwastewaters;with viral numbers as high as 1011 viruses/gramof faecalmatter (Simmons&

Water Practice & Technology Vol 9 No 3371 doi: 10.2166/wpt.2014.039

Xagoraraki 2011). It is widely accepted that secondary wastewater treatment systems do not fullyremove viruses from a water body thus a tertiary treatment is desirable. Discharges from municipalwastewaters into aquatic environments can lead to NoV outbreaks by contaminating drinking watersupplies (Riera-Montes et al. 2011) and infiltrating food chains (Maunula et al. 2009).Tertiary treatment processes may include sand filtration, ultraviolet irradiation or membrane fil-

tration. Low pressure membrane filtration (namely microfiltration and ultrafiltration) in particularhas become increasingly popular over the past decade due to its associated low costs and remarkableefficacy in producing high quality effluent (Huang et al. 2009). However, a primary disadvantage com-monly linked to this method is that of membrane fouling which is highly dependent on the quality ofthe influent water. Means of alleviating this problem have included pre-treatment of the source watervia methods such as coagulation. Membrane filtration is separated into four classifications of pore sizeor molecular weight cut-off (MWCO); microfiltration (MF; 0.1–1 μm), ultrafiltration (UF; 0.01–0.1 μm), nanofiltration (NF; 0.01–0.001 μm) and reverse osmosis (RO; ,-0.001 μm). MWCO refersto the molecular mass of a macrosolute and so is not measured in length (μm) but rather in kiloDal-tons (kD). This type of measurement is generally limited to the smaller pore size classifications; UFand below. It should be noted that the classification of pore sizes (μm) for filtration tend to varyand overlap slightly as no one standard appears to be used internationally. With regards to pathogeneradication both microfiltration and ultrafiltration are generally used in accordance with pathogensize and morphology. The removal of microbial contaminants is typically expressed in terms oflog10 reduction or log10 removal and is defined by equation 1.

Log removal ¼ log10Cin

Cout

� �(1)

where Cin¼ Feed concentration and Cout¼ Filtrate concentration.Larger pathogens such as protozoan cysts e.g. Giardia and Cryptosporidium would be readily

removed via MF as they are two orders of magnitude larger than a typical MF filter (Hsu & Yeh2003). Most species of bacteria would also be removed via microfiltration as they too fall above theMF cut-off pore size. Viruses, on the other hand, are significantly smaller in size and so are more com-parable to the pore size of ultrafiltration membranes. Figure 1 displays a summary review of virusremoval (bacteriophage MS2) via microfiltration and ultrafiltration. In the context of comparisonMS2 is approximately 0.025 μm or 25 nm whereas NoV is approximately 28–30 nm.

Figure 1 | Summary of MF and UF studies (1991–2000) documenting virus (MS2) removal efficiency (USEPA 2001).

Membrane filtration may occur in one of two ways; dead end filtration or cross-flow/tangential flow fil-tration (TFF). Dead end filtration requires the entire sample to pass through the membrane until cloggingeventuallyoccurs. In the caseof cross-flowfiltration, feedwater is directed tangentially across themembranewith a portion of the sample permeating through the membrane (permeate) while the remainder flows

Water Practice & Technology Vol 9 No 3372 doi: 10.2166/wpt.2014.039

across themembrane (retentate). The samplemay then be re-circulated as desired; thismethod tends to limitthe amount of membrane biomass build up or ‘cake layer’ formation. Virus removal can also occur viaadsorption to the cake layer.As this layer thickens, the ability of themembrane to remove the virus improvesestablishing an alternative mechanism for virus removal. Huang et al. (2012) observed an increased log10reduction (2.1 to 3.0) of MS2 bacteriophage with filtration time in high fouling water. Removal levels forthe same virus in low fouling water ranged from 0.8 to 1.7 log10 using the same test conditions.Due to a lack of a suitable cell culture system, NoV may not be cultured in a laboratory. This poses

significant drawbacks when investigating the behaviour of the virus. To date, the most widely useddetection method is real-time reverse-transcription quantitative PCR (RT-qPCR). This type of molecu-lar analysis is advantageous as it can target specific genogroups and allows for quantification howeverit also has the potential to overestimate the amount of infectious virus present in a sample. Other dis-advantages of molecular assays are the high costs involved and the extensive training required to carrythem out. For these reasons, surrogate viruses namely F-specific RNA (FRNA) bacteriophages havebeen suggested due to their similar physical and physiochemical characteristics (Doré et al. 2000;Flannery et al. 2012). Bacteriophages are viruses that infect bacteria; they are present in the environ-ment the year round and most importantly they are culturable. Quick, easy and cost-effectivemicrobiological assays may be carried out on FRNA bacteriophages which also allows for infectivityquantification not currently possible via RT-qPCR.This study investigates whether membrane filtration; microfiltration and ultrafiltration may be used as

an efficient tertiary disinfection method for norovirus removal from wastewaters, with emphasis on theinfluence of pore size on viral removal rates and membrane clogging. The study also investigates the useof FRNAbacteriophages as a potential surrogate indicator of NoV. The infectivity of norovirus is assessedthrough the use of FRNA bacteriophages via a microbiological assay only available to cultivable viruses.

MATERIALS AND METHODS

TFF system

The TFF acted as a barrier method employing cross-flow filtration action across a membrane cassette(Pall’s Omega™ polyethersulfone [PES]). Approximately 5 L batches of secondary treated wastewaterwere seeded with FRNA bacteriophage (GA strain; from this point on it will be referred to as GA bac-teriophage) and pumped into a bench scale platform TFF system at a permeate flow rate ofapproximately 2 L/hr. It was decided at the beginning of the study that only the membranes exhibitingsignificant reductions would then be spiked with NoV stocks (genogroup II [GII]) in addition to thebacteriophage due to the costly nature of the molecular assays. Microfiltration cassettes analysed poresizes 0.45 and 0.2 μm while ultrafiltration cassettes analysed 100 kD pore size. Each cassette wastrialled at least four times before moving on to the next pore size.

Virus enumeration

Samples were processed via two methods molecular analysis and microbiological analysis. NoV GIIanalysis was carried out via molecular methods while GA bacteriophage processed by microbiologicalanalysis.

Molecular analysis

Approximately 40 mL of wastewater was concentrated to 500 μL using a virus adsorption-elutionmethod. The 500 μL sample was then brought forward for RNA extraction using NucliSENS

Water Practice & Technology Vol 9 No 3373 doi: 10.2166/wpt.2014.039

miniMAG extraction platform and NucliSENS magnetic extraction reagents. Briefly, the sample wasplaced in lysis buffer to release viral RNA. Magnetic silica was then added to the lysed samples towhich the RNA attached. This sample was then washed in a series of buffers to remove inhibitorsand contaminants and is finally eluted off the silica using elution buffer and placed in �20 °C storagefor PCR analysis. RT-qPCR was carried out on samples to test for NoV GII. Five microlitre aliquots ofsample RNA were added in duplicate to the wells of a 96 – well optical reaction plate. Twenty micro-litres of appropriate one-step mastermix were also added to the well including specific primers andprobes for the target virus genome to give a final reaction volume of 25 μL. In addition, a double-stan-dard DNA (dsDNA) curve was included for quantitation and positive control and molecular biologygrade water was used as a negative control. The use of a non-related virus (Mengo virus; Costafredaet al. 2006) was also employed as an ‘internal process control’ (IPC) to determine the extraction effi-ciency. PCR inhibition was also controlled for in the assay as described by (Flannery et al. 2012). TheRT-qPCR assay was ran in an AB7500 real-time PCR instrument under the appropriate run conditions(Flannery et al. 2012). Results were determined by examining Ct values and comparing them againstthe dsDNA curve generated by the PCR instrument to give genome copies per 100 mL of sample(copies/100 mL). Assay conditions and further details are outlined in Flannery et. al. (2012). Thelimit of quantification (LOQ) for this test was 125 detectable copies/100 mL and the limit of detection(LOD) was 25 detectable copies/100 mL.

Microbiological analysis

A double – layer overlay plaque assay was employed for this analysis (ISO standard method 10705-1).Briefly, 1 mL of appropriately diluted sample was added to 1 mL of host culture (Salmonlla Typhimur-ium WG49) and 2.5 mL of molten tryptone – yeast glucose agar and held at 45 °C. This mix was thenpoured onto hardened tryptone – yeast glucose agar plates and left to solidify before being transferredto a 37 °C incubator for 18+ 2 h. Once the incubation period had passed, the plates were removedand characteristic plaques were counted where each plaque was assumed to originate from oneGA bacteriophage. The results were expressed as plaque forming unit (pfu)/mL. The LOQ for thistest was 1 pfu/mL.

Suspended Solids (SS) test

0.45 μm cylindrical filter papers were labelled and weighed before duplicate 100 mL samples of sec-ondary effluent were passed through using a vacuum pump. The filter papers were then dried in a hotoven (105 °C) for 24 hours. Once dried, the filter papers were removed and re-weighed to determineany SS present. An average of two weights was determined for each sample tested.

RESULTS

Analysis of the microfiltration cassettes (0.45, 0.2 μm) exhibited at best, a 1 log10 reduction of GA bac-teriophage but no significant virus removal occurred which is consistent with current literature;microfiltration pore sizes appear to be too large to create a barrier for successful virus eradication(see Figure 2). This trend was observed regardless of the extent of membrane clogging observed.Following microfiltration, analysis then focused on the ultrafiltration cassettes. Initial tests on the

100 kD pore size displayed a 5 log10 reduction of GA bacteriophage (Figure 3).As significant GA bacteriophage reductions were observed using the 100 kD cassette, it was

decided to commence spiking with human pathogenic NoV GII alongside the GA bacteriophagevirus. HYPERLINK \l “bookmark3” Figure 4 illustrates virus reduction using the 100 kD cassette

Figure 2 | (a) GA bacteriophage log10 reduction via barrier method; feed vs. permeate using a 0.45 μm pore size cassette and(b) GA bacteriophage log10 reduction via barrier method using 0.2 μm cassette.

Figure 3 | GA bacteriophage log10 reduction via ultrafiltration using 100 kD poresize cassette.

Figure 4 | Comparative reduction of GA bacteriophage via infectivity assay (pfu/100 mL), NoV GII via PCR assays and infectivityassays (copies/100mL), employing ultrafiltration: 100 kD poresize cassette.

Water Practice & Technology Vol 9 No 3374 doi: 10.2166/wpt.2014.039

over four runs. GA bacteriophage was analysed via overlay plaque assay (pfu/100 mL) while NoV GIIwas analysed via RT-qPCR (copies/100 mL). A 3–5 log10 reduction of GA bacteriophage was achievedthroughout the trial period whereas NoV GII was reduced to below the limit of detection (approxi-mately 2–3 log10 reduction) for each run.Figure 5 compares average GA bacteriophage log10 reductions for membrane pore sizes investi-

gated to date. The microfiltration cassettes (0.45 and 0.2 μm) exhibited at best a 1 log10 reduction.For the 100 kD cassette an average removal of 4 log10 was achieved.Figure 6 exhibits the cumulative surface area loading of total suspended solids (TSS) vs. the effluent

volume ratio. The effluent volume ratio is displayed on the x-axis as the permeate volume divided bythe total influent volume, this is also divided by the pressure (bar) at which each cassette was trialled

Figure 5 | Average log10 reduction of GA bacteriophage vs. membrane cassette pore size (100 kD (0.01 μm), 0.2 and 0.45 μm)).

Figure 6 | Cumulative surface area loading of TSS vs. effluent volume ratio (Vp/VT) per unit bar on three cassettes trialled.

Water Practice & Technology Vol 9 No 3375 doi: 10.2166/wpt.2014.039

at. For each cassette the effluent volume flux is decreasing highlighting the fact that suspended solidloading on the cassette is having a negative impact on permeate flux.

DISCUSSION AND CONCLUSIONS

The use of microfiltration membranes as an inadequate barrier system for virus removal has beendemonstrated previously in the literature and by the results in this study. However, informationregarding the potential of MF membranes combined with solids loading for the removal of virusesis limited. The microfiltration membranes exhibited at best, a 1 log10 reduction of GA bacteriophagebut no significant removal was achieved. Furthermore, suspended solid loading did not appear toincrease virus removal. Analysis of the ultrafiltration cassette (100 kD) exhibited an average removalof 4 log10 of GA bacteriophage and a removal to below the limit of detection for NoV GII; approxi-mately 2–3 and 4 log10. The potential of FRNA bacteriophages as a surrogate for NoV has been widelyinvestigated due to similarities shared between the viruses. To date, this study has not shown completeremoval of GA bacteriophage but has reduced NoV GII to undetectable levels. Factors worth notingare (i) the difference in influent concentrations between viruses; GA bacteriophage concentrationswere approximately 108/100 mL while NoV GII concentrations were 104/100 mL (ii) the LOD forthe PCR assay. As the influent concentrations for both viruses were not equal it is difficult toassume a direct comparison. The LOQ for PCR is 125 copies/100 mL therefore there may be NoVvirus present in the permeate but it may not have been detected. The question as to whether FRNAbacteriophages are a good surrogate for NoV performance in this case is still under investigation.Studies will continue with the analysis of a final ultrafiltration cassette (500 kD) where influentvirus concentrations will be further optimised with a view for an improved comparison betweenboth viruses. The potential of using FRNA bacteriophages as a surrogate for NoV removal in a

Water Practice & Technology Vol 9 No 3376 doi: 10.2166/wpt.2014.039

WWTP would be cost-effective and much easier to implement rather than relying on costly molecularmethods.The impact of SS on TFF performance was also monitored throughout the study. For all pore sizes

investigated a rapid decrease in both retentate flux and permeate flux rate was observed over time,particularly in the case of the ultrafiltration cassette. This would be of particular concern at waste-water facilities where permanent on-site operators are not present. While scouring andbackwashing can alleviate such issues, the pore sizes necessary for viral removal may pose oper-ational challenges for designers. Overall, the practicality and cost of the implementation and use ofmembranes as a barrier method for tertiary disinfection in wastewater may need further study, par-ticularly for smaller decentralised facilities. The impact of SS on the maintenance of large-scalemembranes would most likely be too labour intensive and unrealistic.

REFERENCES

Costafreda, M. I., Bosch, A. & Pintó, R. M. 2006 Development, evaluation and standardisation of real – time TaqMan reversetranscription – PCR assay for quantification of Hepatitis A virus in clinical and shellfish samples. Applied andEnvironmental Microbiology 72 (6), 3846–3855.

Doré, W. J., Henshilwood, K. & Lees, D. N. 2000 Evaluation of F-specific RNA bacteriophage as a candidate human entericvirus indicator for bivalve molluscan shellfish. Applied and Environmental Microbiology 66, 1280–1285.

Flannery, J., Keaveney, S., Rajko-Nenow, P., O’Flaherty, V. & Doré, W.J. 2012 Concentration of norovirus during wastewatertreatment and its impact on oyster contamination. Applied and Environmetal Microbiology 78 (9), 3400–3406.

Hoffmann, S., Batz, M. B. & Morris, Glen, J. 2012 Annual cost of illness and quality-adjusted life year losses in the United Statesdue to 14 foodborne pathogens. Food Protection 75 (7), 1292–1302.

Hsu, B. M. & Yeh, H. H. 2003 Removal of Giardia and Cryptosporidium in drinking water treatment: a pilot-scale study. WaterResearch 37 (5), 1111–1117.

Huang, H., Schwab, K. & Jacangelo, J. G. 2009 Critical review pretreatment for low pressure membranes in water treatment : areview. Environmental Science and Technology 43 (9), 3011–3019.

Huang, H., Young, T. A., Schwab, K. J. & Jacangelo, J. G. 2012 Mechanisms of virus removal from secondary wastewatereffluent by low pressure membrane filtration. Membrane Science 409–410, 1–8.

Lopman, B. A., Reacher, M. H., Vipond, I. B., Hill, D., Perry, C., Halladay, T., Brown, D. W., Edmunds, W. J. & Sarangi, J. 2004Epidemiology and cost of nosocomial gastroenteritis, Avon, England, 2002–2003. Emerging Infectious Diseases 10,1827–1834.

Maunula, L. C., Roivainen, Keranen, M., Mäkelä, S., Söderberg, K., Summa, M. & Von Bonsdorff, C. H. 2009 Detection ofhuman norovirus from frozen raspberries in a cluster of gastroenteritis outbreaks. European Surveillance 14 (49), 19435.

Riera-Montes, M., Sjolander, K. B., Allestam, G., Hallin, E., Hedlund, K. O. & Lofdahl, M. 2011Waterborne norovirus outbreakin a municipal drinking-water supply in Sweden. Epidemiological Infections 139, 1928–1935.

Simmons, F. J. & Xagoraraki, I. 2011 Release of Infectious Human Enteric Viruses by Full-Scale Wastewater. Water Research45, 3590–3598.

U.S. Environmental Protection Agency (USEPA) 2001 Low Pressure Membrane Filtration For Pathogen Removal: Application,Implementation and Regulatory Issues. EPA 815-C- 01- 001.

Tam, C. C., Rodrigues, L. C., Viviani, L., Dodds, J. P., Evans, M. R., Hunter, P. R., Gray, J. J., Letley, L. H., Rait, G., Tompkins,D. S. & O’Brien, S. J. 2011 Longitudinal study of infectious intestinal disease in the UK (IID2 study): incidence in thecommunity and presenting to general practice. Gut 61 (1), 69–77.