ww.sciencedirect.com

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 6 5 1e2 6 6 4

Available online at w

journal homepage: www.elsevier .com/locate/watres

Removal of MS2, Qb and GA bacteriophages during drinkingwater treatment at pilot scale

Nicolas Boudaud a,*, Claire Machinal a, Fabienne David a, Armelle Freval-Le Bourdonnec a,Jerome Jossent a, Fanny Bakanga a, Charlotte Arnal a, Marie Pierre Jaffrezic a,Sandrine Oberti a, Christophe Gantzer b

aVeolia Environment Research and Innovation, Chemin de la Digue, BP76, 78608 Maisons-Laffitte Cedex, Franceb Laboratoire de Chimie Physique et Microbiologie pour l’Environnement (LCPME), Nancy Universite/CNRS, Faculte de Pharmacie,

5 rue Albert Lebrun, 54000 Nancy, France

a r t i c l e i n f o

Article history:

Received 1 March 2011

Received in revised form

27 January 2012

Accepted 11 February 2012

Available online 3 March 2012

Keywords:

Bacteriophages

Virus removal

Drinking water

Treatment

Clarification

Filtration

Chlorine disinfection

* Corresponding author. Tel.: þ33 1 34 93 31E-mail address: nicolas.boudaud@veolia.

0043-1354/$ e see front matter ª 2012 Elsevdoi:10.1016/j.watres.2012.02.020

a b s t r a c t

The removal of MS2, Qb and GA, F-specific RNA bacteriophages, potential surrogates for

pathogenic waterborne viruses, was investigated during a conventional drinking water

treatment at pilot scale by using river water, artificially and independently spiked with

these bacteriophages. The objective of this work is to develop a standard system for

assessing the effectiveness of drinking water plants with respect to the removal of MS2, Qb

and GA bacteriophages by a conventional pre-treatment process (coagulation–flocculation–

settling-sand filtration) followed or not by an ultrafiltration (UF) membrane (complete

treatment process). The specific performances of three UFmembranes alone were assessed

by using (i) pre-treated water and (ii) 0.1 mM sterile phosphate buffer solution (PBS), spiked

with bacteriophages. These UF membranes tested in this work were designed for drinking

water treatment market and were also selected for research purpose.

The hypothesis serving as base for this study was that the interfacial properties for

these three bacteriophages, in terms of electrostatic charge and the degree of hydropho-

bicity, could induce variations in the removal performances achieved by drinking water

treatments.

The comparison of the results showed a similar behaviour for both MS2 and Qb

surrogates whereas it was particularly atypical for the GA surrogate. The infectious char-

acter of MS2 and Qb bacteriophages was mostly removed after clarification followed by

sand filtration processes (more than a 4.8-log reduction) while genomic copies were

removed at more than a 4.0-log after the complete treatment process. On the contrary, GA

bacteriophage was only slightly removed by clarification followed by sand filtration, with

less than1.7-log and 1.2-log reduction, respectively. After the complete treatment process

achieved, GA bacteriophage was removed with less than 2.2-log and 1.6-log reduction,

respectively.

The effectiveness of the three UF membranes tested in terms of bacteriophages removal

showed significant differences, especially for GA bacteriophage. These results could

provide recommendations for drinking water suppliers in terms of selection criteria for

membranes.

MS2 bacteriophage is widely used as a surrogate for pathogenic waterborne viruses in

Europe and the United States. In this study, the choice of MS2 bacteriophage as the best

87; fax: þ33 1 34 93 31 10.com (N. Boudaud).ier Ltd. All rights reserved.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 6 5 1e2 6 6 42652

surrogate to be used for assessment of the effectiveness of drinking water treatment in

removal of pathogenic waterborne viruses in worst conditions is clearly challenged. It was

shown that GA bacteriophage is potentially a better surrogate as a worst case than MS2.

Considering GA bacteriophage as the best surrogate in this study, a chlorine disinfection

step could guaranteed a complete removal of this model and ensure the safety character of

drinking water plants.

ª 2012 Elsevier Ltd. All rights reserved.

1. Introduction case scenario” in terms of virus removal (Schijven and

The importance of water as a vehicle for the transmission of

human pathogenic viruses is now well known, as are the

various potential sources of contamination (Bosch, 1998;

Szewzyk et al., 2000). Viral pollution of drinking water may

depend on water resource quality and drinking water treat-

ment efficiency (Springthorpe and Sattar, 2007). The occur-

rence of viruses, their persistence and aggregation in hydric

environments influence engineering or natural processes of

water treatment (Lechevallier and Au, 2004).

Studies designed to assess the vulnerability of water

resources with respect to viral risk were also undertaken to

specifically search for norovirus I and II, rotavirus, adenovirus,

astrovirus, enterovirus and reoviruses (Cavereau et al., 2009;

Lodder et al., 2010). Adenovirus was shown to be the most

abundant virus in the resources studied, while very few

enterovirus and hepatitis A virus (HAV) could be detectedwith

the methods currently available (Cavereau et al., 2009). Given

the necessity to minimize public health risks, the effective

removal of pathogenic viruses from water intended for

human consumption becomes a major issue because of

urbanization, demographic growth and the reuse of waste

water (Sattar et al., 1999; Fane et al., 2002).

Apart from some data reported on the reduction of

waterborne pathogenic viruses by the drinking water treat-

ments (Springthorpe and Sattar, 2007), virus detection means

remain limited because of difficulties related to extraction,

culture and implementation of standardized methods (Butot

et al., 2007). To overcome these analytical constraints, many

studies use F-specific RNA bacteriophages as potential surro-

gates of enteric viruses in a wide range of hydric environment

and water treatment processes (IAWPRC, 1991; Van

Voorthuizen et al., 2001; Huertas et al., 2003; Templeton

et al., 2007; Langlet et al., 2009; Shirasaki et al., 2009a).

Because of its similarities to enteric viruses, MS2 bacterio-

phage is the surrogate most commonly proposed and used in

Europe and the United States to assess the performance of

drinking water treatment processes (Jacangelo et al., 1995,

1997; Meng and Gerba, 1996; Redman et al., 1997; Sobsey et al.,

1998; Herath et al., 1999; Huertas et al., 2003; Meschke and

Sobsey, 2003; Shin and Sobsey, 2003; Thurston-Enriquez

et al., 2003; Zhu et al., 2005; Fiksdal and Leiknes, 2006;

Aronino et al., 2009; Langlet et al., 2009; Shirasaki et al., 2009a,

2009b), while Qb bacteriophage is the surrogate typically used

in Japan (Urase et al., 1996; Otaki et al., 1998). These bacte-

riophages have been chosen as indicator of the water treat-

ment process efficiencies because of their specific

physicochemical features which give them a status of “worst

Hassanizadeh, 2000; Van Voorthuizen et al., 2001; Langlet

et al., 2009; Shirasaki et al., 2009a). Qb and MS2 bacterio-

phages have similar size (20e30 nm) and isoelectric point (IEP)

butmay express differences on their hydrophilic/hydrophobic

balance. IEP of both bacteriophages are estimated to be lower

than 3.9 indicating a particle charge significantly negative at

neutral pH (Langlet et al., 2008).

To date, no practical bacteriophage resistance study has

been developed based on a surface water treatment process

sufficiently representative of drinking water production

plants. Most of thework has been confined to laboratory batch

tests (“Jar test”) to assess the performance of the clarification

process (coagulation, flocculation, settling) (Huertas et al.,

2003; Shirasaki et al., 2009a), or the bacteriophage removal

by membrane filtration (Jacangelo et al., 1997; Herath et al.,

1999; Hu et al., 2003; Langlet et al., 2009; Shirasaki et al., 2009b).

Unlike the MS2 and Qb models, GA bacteriophage sub-

jected to the physical barriers of drinking water treatments

has never been studied. Only fundamental data on structural,

interfacial and genomic properties were reported (Inokuchi

et al., 1986; Gott et al., 1991; Tars et al., 1997; Langlet et al.,

2008). Langlet et al. (2008) have previously shown that the

IEP is similar to MS2 and Qb bacteriophages but may express

higher hydrophobicity. Compared to MS2 and Qb models, the

study of GA bacteriophage behaviour under drinking water

treatment processes is of double interest. First, it seems rele-

vant to compare the performance of its removal compared to

the usual MS2 and Qb “reference” models. Secondly, the

surface features of these bacteriophages could serve as

selection criteria for the definition of the model the most

difficult to remove by the drinking water treatment processes.

The objective of this work is to develop a standard system

for assessing the effectiveness of drinking water plants with

respect to the removal of MS2, Qb and GA bacteriophages by

a conventional pre-treatment process (clarification coupled to

a sand filtration) followed or not by an ultrafiltration (UF)

membrane (complete treatment process). The initial hypoth-

esis is based on the fact that the charge and the interfacial

properties of these three bacteriophages could induce varia-

tions in their removal by these processes. A test pilot unit at

a 1:100 scale of the size of conventional surface water treat-

ment plants was implemented while ensuring operational

conditions as close as possible to industrial processes. The

impact of the treatments used were assessed by using the

detection of infectious phages (expressed in plaque forming

units) and reverse transcription-polymerase chain reaction

(RT-PCR) method to detect phage nucleic acids removal (free

and packaged RNA). Using both techniques potentially help to

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 6 5 1e2 6 6 4 2653

highlight aggregation and/or virucidal phenomena specific to

the treatment implemented (Langlet et al., 2009; Shirasaki

et al., 2009a).

2. Material and methods

2.1. Preparation of MS2, Qb, and GA bacteriophagesuspensions

Three representative F-specific bacteriophage strains, MS2,

GA and Qb, belonging to the genogroup I (GI), GII and GIII

respectively, were selected. Some of their properties have

been described by Langlet et al. (2008). These three bacterio-

phages were replicated according to standard procedure (ISO

10705-1, 1995) without the CHCl3 lysis step and using Escher-

ichia coli Hfr K12 (ATCC 23631) as bacterial host. After repli-

cation, the phage suspensions were centrifuged and the

supernatant filtered through a 0.22 mm membrane as

described by Langlet et al. (2008). The phage suspension was

stored as stock suspension at 4 �C prior to experiments.

The initial quantification of the number of infectious units

of these three bacteriophages was determined. The final

phage concentration for the different bacteriophage stock

suspensions was about 1011 PFU/mL.

2.2. Bacteriophage assays

2.2.1. Detection of infectious bacteriophagesMS2, GA and Qb bacteriophages were quantified by double

agar layer plaque assay method (ISO 10705-1, 1995). The

infectious phage concentration was expressed in plaque

forming units per millilitre (PFU/mL).

2.2.2. Real-time RT-PCR methodBacteriophage RNA was quantified by the real-time RT-PCR

method,which detects total bacteriophages regardless of their

infectivity. We defined concentration measured by the real-

time RT-PCR method as total bacteriophage concentration.

The quantification of viral RNA genome was described by

Langlet et al. (2009), using the method developed by Ogorzaly

and Gantzer (2006) with several modifications of the reverse

transcriptase procedure.

The resulting cDNA was then quantified by TaqMan real-

time PCR using primers and probes designed for the GI

(MS2), GII (GA) and GIII (Qb) as previously described by

Ogorzaly and Gantzer (2006).

Because substances such as natural organic matter (NOM)

present in riverwater are known to inhibit the amplification of

the viral genome by PCR (Abbaszadegan et al., 1993), each

sample was diluted 10 and 100-fold with demineralized water

before the real-time RT-PCR quantification.

The standard curve for the real-time RT-PCR method was

based on the relationship between the infectious bacterio-

phage concentration of a freshly prepared stock suspension

(diluted in 0.1 M PBS solution) measured by the ISO 10705-1

method and the number of cycles for amplification in the PCR

method, which is based on the assumption that the stock

suspension did not contain any inactivated or aggregated viral

particles.

2.3. Reagents and analytical procedure

Thecoagulant PAX-XL7 (aluminiumpoly-hydroxychlorosulfate,

8.5% Al2O3eKemira) was used for the clarification process of

surface water treatment.

For chlorine disinfection, a commercial solution of bleach

(9.6%-115 g Cl2/L) diluted in demineralizedwater (100mgCl2/L)

was used. The residual chlorine was determined using N,N-

diethyl-p-phenylenediamine (DPD) colourimetric method

(Hach Lang kit 8021, USEPA accepted method. The procedure

is adapted from the Standard Methods for the Examination of

Water and Waste water, 2005; 4500-Cl G for drinking water).

Turbidity was measured using a Hach 800DR turbidimeter

(Hach Company, Loveland, CO).

2.4. Water sources

River water was sampled from the Seine River (Maisons Laf-

fitte, France). Pre-treated water was obtained after clarifica-

tion followed by sand filtration of river water during drinking

water treatment at pilot scale.

2.5. Experimental design of drinking water treatment atpilot scale

The drinking water treatment process at a pilot scale of 1:100,

combining a pre-treatment (clarification followed by sand

filtration) and UF membrane processes is described Fig. 1. The

pilot was implemented with the aim of being as representa-

tive as possible of the actual size of the surface water treat-

ment plants typically operated by drinking water suppliers, in

order to assess the abilities of these drinking water plants in

removing enteric viruses.

Raw water pre-treatment process (100 L) involved

sequential coagulation (1 L), flocculation (5 L), lamellar settling

(5 L) and sand filtration (filtration velocity 2.3m/h, sand height

80 cm, column inner diameter 10 cm, sand volume 6.3 L) steps.

Pre-treated water at the sand filtration output was collected in

the pre-treated water tank (100 L) by means of a 10 mm inner

diameter polyvinyl chloride (PVC) pipe. Applied pre-treatment

flow rate was controlled at 18 L/h by a flowmeter. Raw water

transfer to the coagulation tank was carried out with a 14 mm

inner diameter PVC pipe. The coagulant PAX-XL 7was injected

on-line in the coagulation tank at an initial concentration of

25.2 mg/L Al2O3, using a 0.25 mm inner diameter iso-versinic

tube. Feeding flow rate was controlled at 72 mL/h in order to

get a final concentration of 4.25 mg/L Al2O3 in the coagulation

tank. Raw and pre-treated water tank were made of poly-

ethylene (PE), whereas pre-treatment process-specific tanks

were all made of PVC.Water stored in the raw and pre-treated

water tanks was stirred at a 50 rpm rate in order to maintain

some homogeneity. In the coagulation tank, water was stirred

quickly at a 350 rpm rate, whereas in the flocculation tank,

water was stirred at slower rate of 200 rpm so as to promote

the formation of aluminium floc particles. After settling, the

floc particles were extracted from the lamellar clarifier and

discarded by pumping (after complete pre-treatment of 100 L

of raw water). Conductivity and pH probes were introduced

into the raw and pre-treated water tanks to assess water

quality changes on-line.

Fig. 1 e Schematic of the drinking water treatment at pilot scale (1:100) experimental setup.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 6 5 1e2 6 6 42654

Pre-treated water UF process was carried out by means of

an assembledmembrane Amodule contained in PVC pressure

vessels (n ¼ 3). The characteristics of this membrane were

described below, according to the manufacturer data.

Flow rate was set at 0.84 L/h, with an initial imposed

pressure of 0.3 bars at the beginning of the filtration. Pre-

treated water transport to the membrane filtration was

carried out with a 4 mm inner diameter hexa-canal tube. All

UF membranes studied in this work were chemically treated

beforehand by the membrane manufacturer in order to

provide a hydrophilic membrane material (the contact angles

go from 0� to 40�). No further information has been provided

by the manufacturer regarding the chemical treatment used.

Each membrane module was tested prior to experiment in

order to check its integrity and permeability (Machinal et al.,

2006). A module was considered intact when no pressure

loss was detected, and when permeability measurements on

membrane modules from a batch of 12 modules did not

deviate from more than 15% (data not shown). The UF

membrane A was designed for the drinking water treatment

market and was selected for research purpose since its cutoff

is about the size of the bacteriophages studied.

2.6. Experimental methodology

In order to assess the MS2, GA and Qb bacteriophages removal

performances by the pre-treatment process, the UF

membrane and the combination of both processes (complete

treatment process), repeatability and reproductibility

measurements were performed to ensure data robustness

(three analysis campaigns with triplicates per phage). The

behaviour of each bacteriophage subjected to these physical

barriers for drinking water treatment was specifically moni-

tored to determine their inactivation (PFU) and removal

(genome).

Pre-treatment process efficiency for each bacteriophage

was assessed by treating 100 L of raw water which had been

spiked beforehand with a bacteriophage stock suspension to

contain about 106 PFU/mL (described in Section 2.1). Samples

were collected before (in the raw water tank) and after clari-

fication followed by sand filtration (in the pre-treated water

tank). For each raw and pre-treated water sample, three

analyses to detect infectious (PFU) and total bacteriophages

(genome) were performed. In order to measure virus removal

performances which are representative of typical surface

water treatment plants operated by drinking water suppliers,

pre-treated water samples were also collected based on the

experimental determination of the hydraulic Residence Time

Distribution (RTD) by means of a tracer [Naþ, Cl�] (Villermaux,

1993). This tracer was directly injected into the raw water in

order to increase the conductivity of the medium by about

three times. The tracer behaviour was then monitored by

conductivity at the start (rawwater tank) and at the end of the

pre-treatment process (pre-treated water tank). Based on the

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 6 5 1e2 6 6 4 2655

RTD determination, raw and pre-treated water samples were

collected after 240min of operating time of the drinking water

treatment at pilot scale to measure the optimal effectiveness

of clarification followed by sand filtration.

The bacteriophage removal performances of the complete

treatment process were measured in the same manner as for

the pre-treatment process evaluation, except that samples

were taken from the raw water tank (T0 and T240 min) and after

UF membrane A of pre-treated water. For each of the three

membrane A modules fed with pre-treated water, permeate

sampleswere collected after 30min and 120min ofmembrane

filtration, and after a backwash (30 s at 0.8 bars) followed by

a 30 min filtration. A single analysis to detect infectious (PFU)

and total (genome) bacteriophages was performed for each

permeate collected.

Assessment of the UF membrane A alone was carried out

on 20 L of water spiked with either MS2, GA or Qb bacterio-

phage stock suspension to a final concentration of about

106 PFU/mL. Samples were collected before (in the pre-treated

water tank) and after membrane filtration (as described

below). A single analysis to detect infectious and total bacte-

riophages was performed for each permeate, as described

above.

With the aim of assessing the UF membrane behaviour in

terms of bacteriophages removal according to water quality,

comparative tests were conducted on 5 L of 0.1 mM PBS

solution containing around 106 PFU/mL with membranes A

and two others membranes (B and C). Thesemembranes were

designed for drinking water treatment market and were

selected for research applications. According to the manu-

facturer data, the characteristics of these UFmembranes were

described in Table 1.

The experimental designs of the filtration equipment and

the analytical methodology deployed to characterize UF

membranes are the same as those described by Langlet et al.

(2009).

Raw water, pre-treated water and permeate samples

(100 mL) were stored for 24 h at 4 �C before analysis.

2.7. Chlorine disinfection assays

Chlorine-based disinfectants are applied in most plants for

disinfection purposes and also to maintain a residual disin-

fectant in the distribution system. The initial disinfectant

doses were chosen to achieve oxidant residuals

Table 1 e Characteristics of the UF membranes A, B and C sele

Membrane UF memb

Material PVDFa

Number of fibres 6

Active surface area (cm2) 100

Filtration mode External–in

Cutoff – pore size (mm)/MWCOc (KDa) 0.03 mm

Theoretical permeability at 20 �C (L h-1 m-2 bar�1) 300

a PVDF: polyvinylidene fluoride.

b PES: polyethersulfone.

c MWCO: molecular-weight-cutoff.

representative of typical drinking water treatments. The

target residual levels were 0.3 mg/L for chlorine, after

a contact time of 30 min.

Preliminary tests were conducted on demineralized water

to which the bacteriophage culture medium was added.

Residual chlorine levels were thenmonitored over 1 h. Results

have shown no effect of the culture medium into the residual

oxidant reduction because the injected volumes of bacterio-

phage stock suspensions were below 15 mL with an initial

concentration of 1.5 mg Cl2/L (data not shown).

2 L of spiked permeate and demineralized water (with an

initial concentration of about 106 PFU/mL for MS2 and GA

bacteriophages) were disposed in hermetic glassware covered

with aluminium. Permeate was produced by drinking water

treatment at pilot scale. Chlorine was added while stirring

gently with a magnetic stirrer. Samples were taken after

predetermined contact times (0, 2, 5, 10, 30 and 60 min) for

measurement of the residual chlorine, infectious and total

bacteriophages. Samples were collected in commercial

sampling bottles treated with thiosulfate (Gosselin�) in order

to neutralize the disinfectant and ensure the integrity of

bacteriophages. Analyses were done in triplicate and samples

were stored at 4 �C during 24 h before PFU and RT-PCR anal-

ysis. Ct4log values (disinfectant concentration, C [mg/L],

multiplied by the contact time, t [min]) were used to express

the effectiveness of the various disinfectants required to

remove a 4-log reduction of bacteriophages.

3. Results and discussion

3.1. Control of water quality and pre-treatment processefficiency

Physicochemical characteristics of raw and pre-treated water

used for all experiments were measured in order to monitor

raw and pre-treated water quality and pre-treatment process

efficiency (clarification followed by sand filtration). The

minimum and maximum values obtained during the various

test campaigns are summarized in Table 2. These data helped

to highlight variations in the quality of raw water.

The physicochemical characteristics of pre-treated water

were particularly homogenous from one parameter to the

other (Table 2). Based on turbidity data (<0.25 FNU) and total

organic carbon (TOC) (<2.5 mg/L) of pre-treated water, the

cted for research purpose.

rane A UF membrane B UF membrane C

PVDFa PESb

6 7

100 100

ternal External–internal Internal–external

200 KDa 100 KDa

300 600

Table 2 e Characteristics of the experimental natural water (Seine river) and pre-treated water.

Parameters Raw water characteristics(min–max values)a

Pre-treated water characteristics(min–max values)b

Conductivity (mS/cm) 398–524 385–496

pH 7.6–8.2 7.6–8.2

Temperature (�C) 12.6–21.9 NA

Total Alkalinity (F�) 16.5–19.8 15.2–17.9

Turbidity (FNU) 5.4–18.3 <0.2–0.25

Total Organic Carbon (mg/L) 2.5–3.1 2.2–2.5

Chloride (mg/L) 24.4–28.2 28.5–36.2

Sulphate (mg/L) 32.9–45.2 34.1–42.1

Calcium (mg/L) 73–86 68–82

Magnesium (mg/L) 6.2–8.6 6.1–9.2

Potassium (mg/L) 3.7–4.3 3.9–5.8

Sodium (mg/L) 14–17 15–18

Manganese (mg/L) <0.005–0.014 <0.005–0.400

Iron (mg/L) 0.005–0.012 <0.005

Ammonia (mg/L) 0.08–0.37 0.05–0.32

Aluminium (mg/L) 0.005–0.010 0.016–0.072

NA: not available.

a Raw water was used for the pre-treatment experiments.

b Pre-treated water was used for the membrane ultrafiltration experiments.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 6 5 1e2 6 6 42656

process developed at pilot scale demonstrated a good effi-

ciency after clarification followed by sand filtration (regardless

of raw water quality). Note that, compared to surface water

treatment plants typically operated by drinking water

suppliers; the pre-treatment applied at pilot scale is much less

drastic. Indeed, tolerance thresholds for the quality of water

clarified after sand filtration in plants must be <0.2 FNU for

turbidity and <2.0 mg/L for TOC.

For 4.25 mg Al2O3/L of coagulant PAX-XL 7 in the pilot

coagulation tank, concentrations <0.072 mg Al/L were found

at the sand filtration output. This result is in agreement with

WHO (2006) which states that, depending on the production

capacity of drinking water facilities, the threshold at the plant

output should be <0.1e0.2 mg Al/L.

Fig. 2 e Hydraulic control of the pre-

Raw and pre-treated water pH was ranging from 7.6 to 8.2

(Table 2). According to their respective IEP bacteriophages

(Langlet et al., 2008), this result indicates that these viruses

were all negatively charged in the natural water studied and,

therefore, tended to be isolated from each other.

Based on RTD determination, the results presented in Fig. 2

show that the pre-treatment process used was perfectly

homogenous in terms of hydrodynamic after about 200min of

operating at an 18 L/h flow rate, since the measured conduc-

tivity in pre-treated water reached a stationary state equiva-

lent to the one of the raw water (w950 mS/cm). It can be

assumed that for bacteriophages this plateau could take more

time to be reached because of interactions with NOM and di-

and trivalent cations throughout the pre-treatment process

treatment process at pilot scale.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 6 5 1e2 6 6 4 2657

(Pham et al., 2009; Mylon et al., 2010). Consequently, the

optimal effectiveness of the pre-treatment process at pilot

scale in terms of removal of bacteriophageswas assessed after

240 min of operating time of the drinking water treatment at

pilot scale.

3.2. Impact of pre-treatment process followed byultrafiltration on MS2, Qb and GA bacteriophages

In order to investigate the behaviour of MS2, GA and Qb

bacteriophages submitted to drinking water treatment at pilot

scale involving sequential clarification-sand filtration-UF

membrane A (complete treatment process), infectious (log

PFU/mL) and total phage concentrations (log particles/mL) in

raw and pre-treated waters and in permeates were measured

(Fig. 3). First, the constitutive content of infectious bacterio-

phages of raw water was quantified (<100 PFU/mL). This

Infectious bacte

0,0

1,0

2,0

3,0

4,0

5,0

6,0

7,0

T0 Raw water T240min Raw water T240min Pre-treatment process

(clarification + sandfiltration)

After comtreatmen

Drinking water trea

Lo

g in

fe

ctio

us

b

ac

te

rio

ph

ag

e c

on

ce

ntratio

n (lo

g(P

FU

/m

L))

Total bacteri

0,0

1,0

2,0

3,0

4,0

5,0

6,0

7,0

8,0

T0 Raw water T240min Raw water T240min Pre-treatment process

(clarification + sandfiltration)

After cotreatme

Drinking water tre

Lo

g to

ta

l b

ac

te

rio

ph

ag

e c

on

ce

ntra

tio

n (lo

g(p

artic

les

/m

L))

a

b

Fig. 3 e Changes in the concentration of logarithmic infectious

during drinking water treatment at pilot scale (sequence treatm

natural content can be regarded as negligible compared to

the concentration imposed in raw water after spiking

(w106 PFU/mL). The impact of phage culture media on raw

water quality was also not significant since about 1 mL of

phage stock suspension at 1011 PFU/mLwas injected in 100 L of

raw water to obtain around 106 PFU/mL (dilution effect).

Following the RTD determination, the logarithmic reduc-

tion of infectious (<0.39-log) and total bacteriophages (<0.07-

log) after 240 min incubation in raw water indicates that

spontaneous inactivation and aggregation of the phages were

low in the raw water studied (Fig. 3).

The effectiveness of the pre-treatment process showed

significant differences in terms of logarithmic removal of GA

bacteriophage compared to MS2 and Qb viruses. The infec-

tious character of MS2 and Qb bacteriophages was similar

after raw water pre-treatment (Fig. 3a). The removal of Qb

bacteriophage was complete (5.44-log reduction) (i) when the

riophages

plete pre-t process

Permeate 30 min Permeate 120 min Permeate 150 min

tment at pilot scale

Log Qβ (PFU) Log GA (PFU) Log MS2 (PFU)

UF membrane A

Backwash: 30 sec, 0.8 bars

ophages

mplete pre-nt process

Permeate 30 min Permeate 120 min Permeate 150 min

atment at pilot scale

Log Qβ (RT-PCR) Log GA (RT-PCR) Log MS2 (RT-PCR)

Backwash: 30 sec, 0.8 bars

UF membrane A

bacteriophages (a) and logarithmic total bacteriophages (b)

ent: clarification-sand filtrationeultrafiltration).

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 6 5 1e2 6 6 42658

pre-treatment process was homogenous with respect of

hydrodynamic RTD (after 240min of operating time), and (ii) at

the end of the pre-treatment of 100 L of spiked rawwater (after

pre-treatment completion). The same trend was observed for

MS2 bacteriophage, with an infectivity decrease of 5.14-log

reduction and 4.75-log reduction after 240 min and after pre-

treatment completion, respectively. On the other hand, the

behaviour of GA bacteriophage was atypical and clearly

differed from those of MS2 and Qb surrogates when subjected

to the pre-treatment process, with an infectivity decrease of

only 1.54-log reduction after 240 min and 1.65-log reduction

after pre-treatment completion.

For each bacteriophage, the pattern of the removal

behaviour was similar between infectious and total bacterio-

phages throughout the raw water pre-treatment process, but

not the concentrations (Fig. 3b). GA bacteriophage removal

was consistently low after pre-treatment (w1.0-log reduction),

while MS2 and Qb bacteriophages followed a similar evolution

in their logarithmic reduction imposed by the pre-treatment.

Even though total MS2 and Qb bacteriophage concentrations

were close to 3.5-log particles/mL at the end of the pre-

treatment; their logarithmic reduction differed (around 4.4-

log and 2.7-log reduction, respectively).

After raw water spiking, total and infectious initial bacte-

riophage concentrations showeddifferences ranging from0.9-

log for Qb to 1.7-log for MS2. This phenomenon may be

explained by the fact that the phage stock suspensions were

not purified before to be used and might have contained

phages having lost their infectivity and/or free genomic copies

in the culture medium.

The results obtained are in agreementwith those described

for “Jar tests” experiments on MS2 and Qb bacteriophages

(Shirasaki et al., 2009a). To explain the differences in terms of

inactivation and removal of bacteriophages during the pre-

treatment process, these same authors pointed out a higher

sensitivity of Qb bacteriophage to virucidal activity of

aluminium coagulant (in addition to aggregation and inacti-

vation effects induced), while other studies suggested a major

inactivating effect of aluminium coagulant on Qb (Matsui

et al., 2003; Matsushita et al., 2005).

Table 3 e Changes in the logarithmic removal of infectious anpilot scale.

Qb

PFU eq-PFU

Initial log infectious (PFU) and total (e

5.44 6.30

log PFU and eq-PFU r

After complete pre-treatmentb 5.44 2.74

+Permeate UF A 30 min 5.44 3.96

+Permeate UF A 120 min 5.44 3.95

+Permeate UF A 30 min (after backwash) 5.44 3.92

Global inactivation/removal 5.44 3.92-3.96

a Bacteriophages were spiked independently in raw water, initial concen

b Sampling were carried out in the pre-treated tank after complete pre-t

Considering these bacteriophages, the hypothesis put

forward in this study to explain the variations observed after

the pre-treatment is based on the interfacial properties of the

bacteriophages and their interactions with NOM. Negative

surface charges of the three bacteriophages in the resource

studied, influenced by RNA genome (Schaldach et al., 2006),

could generate electrostatic interactions with the aluminium

coagulant and NOM during the coagulation and flocculation

steps (trapping into aluminium floc particles).

Phage aggregation catalyzed by the coagulant could be

concurrently promoted by divalent cations, such as Mg2þ and

Ca2þ (Pham et al., 2009; Mylon et al., 2010), naturally occurring

in raw water. A higher hydrophobicity of GA bacteriophage

compared to MS2 and Qb models (Langlet et al., 2008), could

potentially reduce its removal during the clarification process

and, consequently, promote its diffusion because retention of

phages on the sand column mainly depends on their particle

size (Aronino et al., 2009). This hypothesis remains to be

confirmed by electrophoretic mobility measurements and by

characterizing in more details the interfacial properties of

bacteriophages. This result was quite surprising, given the

structural and genetic similarities between GA and MS2

bacteriophages (Inokuchi et al., 1986; Tars et al., 1997;

Golmohammadi et al., 1996) but suggest that the more

hydrophobic bacteriophage have the best chance to pass such

water treatment.

After clarification followed by sand filtration of raw water

spiked with bacteriophages, pre-treated water was then sub-

jected to the UF membrane A in order to assess the perfor-

mance of the complete treatment process at pilot scale.

Infectious and total bacteriophage concentrations were

measured for MS2, GA and Qb surrogates after a 30 min and

a 120 min filtration, and after a 30 min filtration following

a backwash (Fig. 3). The results showed that residual infec-

tious MS2 bacteriophage were removed early in the UF

membrane A (1.25-log reduction), while very few infectious

GA bacteriophages were retained by the same UF membrane

(<0.56-log reduction) (Fig. 3a). As for the infectious Qb bacte-

riophages, they were totally inactivated after pre-treatment

completion. The same trend was observed for total

d total bacteriophages during drinking water treatment at

Bacteriophages

GA MS2

PFU eq-PFU PFU eq-PFU

q-PFU) concentrations (average)a

5.13 6.50 6.00 7.71

emoval (average)

1.65 1.03 4.75 4.24

1.89 1.31 6.00 6.19

2.21 1.61 6.00 6.37

1.84 1.34 6.00 6.03

1.84-2.21 1.31-1.61 6.00 6.03-6.37

tration targeted 106 PFU/mL.

reatment of raw water spiked (100 L).

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 6 5 1e2 6 6 4 2659

bacteriophages (Fig. 3b). Total MS2 and Qb phages showed

significant logarithmic reduction of 2.0-log and 1.2-log,

respectively, whereas the decrease in GA concentrationwas at

best of 0.58-log reduction.

MS2, GA and Qb bacteriophages logarithmic reductions

obtained after raw water pre-treatment followed by UF

membrane A are summarized in Table 3 to draw overall the

removal (PFU, genome) achieved by the drinking water treat-

ment at pilot scale.

Based on our previous hypothesis, the UF membrane A,

which is electronegative at neutral pH, hydrophilized and

with a 0.03 mm cutoff according to the manufacturer, would

not retain non-aggregated GA bacteriophages as much as

aggregated MS2 and Qb bacteriophages.

Infectious b

0,0

0,5

1,0

1,5

2,0

2,5

3,0

3,5

4,0

4,5

5,0

emrePnim 03 etaemreP

Lo

g in

ac

tiv

atio

n o

f b

ac

te

rio

ph

ag

es

(IS

O 1

07

05

-1

m

eth

od

)

UF membrane

Total ba

0,0

0,5

1,0

1,5

2,0

2,5

3,0

3,5

4,0

4,5

5,0

emrePnim 03 etaemreP

Lo

g rem

oval o

f b

acterio

ph

ag

es (R

T-P

CR

m

eth

od

)

UF membran

a

b

Fig. 4 e Logarithmic removal of infectious bacteriophages (a) an

bacteriophage-spiked pre-treated water.

These results are consistent with published data on MS2

and Qb models subjected to membrane-based treatments

(Jacangelo et al., 1995; Langlet et al., 2009). The behaviour of

the bacteriophages in contact with the UF membrane A and

two others membranes tested (B, C) will be discussed below

with the evaluation of the performance of UF treatment alone.

3.3. Impact of ultrafiltration process on MS2, Qb and GAbacteriophages in pre-treated river water and 0.1 mM PBSsolution

The results of the removal of MS2, GA and Qb bacteriophages

by UF A of pre-treated water spiked to about 106 PFU/mL are

shown in Fig. 4. Mineral and organic qualities of the pre-

acteriophages

nim 051 etaemrePnim 021 eta

Log Qβ (PFU) Log GA (PFU) Log MS2 (PFU)

A, pre-treated water

Backwash:

30sec, 0.8 bars

cteriophages

nim 051 etaemrePnim 021 eta

Log Qβ (RT-PCR) Log GA (RT-PCR) Log MS2 (RT-PCR)

e A, pre-treated water

Backwash:

30sec, 0.8 bars

d total bacteriophages (b) after UF membrane A of

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 6 5 1e2 6 6 42660

treatedwater usedwere relatively stable fromone assay to the

other (data not shown). As previously observed, logarithmic

reduction of infectious GA bacteriophages was low compared

to bothMS2 andQbmodels during the UF A treatment (Fig. 4a).

GA bacteriophage removalwas less than 0.23-log reduction for

the various filtration times, while removal of infectious MS2

and Qb bacteriophages were consistently above 2.62-log and

2.59-log, respectively. These experiments also showed that

backwashing after 120 min of filtration did not significantly

impact bacteriophage retention performances of the UF A

tested. Logarithmic reductions of total and infectious bacte-

riophages by UF A were comparable to each other (Fig. 4b).

These data suggested a good correlation between both

detection methods used.

These results confirm the specific resistance of GA bacte-

riophage when subjected to the various physical barriers of

the drinking water treatment used. As described by Van

Voorthuizen et al. (2001), hydrophobic interactions play an

important role as well as electrostatic interactions during

membrane filtration. These same authors showed that MS2

bacteriophage retention with a hydrophobic membrane was

up to 2.0-log reduction higher than the retention with

a hydrophilic membrane for a pore diameter of 0.22 mm.

Therefore, the properties of the UF membrane A used (elec-

tronegative, hydrophilized, pore size 0.03 mm) could facilitate

the passing-through of GA bacteriophage because of its higher

hydrophobicity compared to MS2 and Qb (Langlet et al., 2008).

Interactions between bacteriophage surface and NOM asso-

ciated to divalent cations (Pham et al., 2009; Mylon et al., 2010)

could involve the aggregation state of the phages according to

their interfacial properties. In order to overcome NOM

involvement in assessing UF A performance, measurements

were performed on 0.1 mM PBS solution spiked to about

106 PFU/mL. The corresponding results for the removal ofMS2,

GA and Qb bacteriophages are shown in Fig. 5. Overall, the

Fig. 5 e Inactivation (PFU) and removal (genome) of bacteriophag

0.1 mM PBS solution.

results confirm those obtainedwith the same UFmembrane A

of spiked pre-treated water. Infectious and total bacterio-

phage concentrations decreased by 3.8e4.2-log reduction and

3.1e4.1-log reduction for MS2 and Qb, respectively, whereas

GA bacteriophage did not seems to be affected by the UF A

treatment since the logarithmic reductions of infectious and

total bacteriophages were below of 0.5-log reduction.

No significant differences in UF A performance for removal

of the studied bacteriophages were observed between pre-

treated water and PBS solution, except for infectious MS2

and Qb bacteriophages. However, it is difficult to conclude on

the involvement of NOM in phage retention by UF membrane

A used, because the distribution index of membrane pores is

not identical at 100% from one membrane module to the

other.

In order to implement the UF performances overlooked

removal bacteriophages, two others UF membranes B and C

were tested with 0.1 mM PBS solution. UFmembranes A and B

showed an only difference at cutoff level while UF membrane

C presented differences regarding material, filtration mode,

cutoff and permeability (Table 1). As described in Fig. 5 for UF

membrane B, infectious and total bacteriophage concentra-

tions decreased by 5.6e6.2-log reduction and 4.6e6.8-log

reduction for MS2 and GA, respectively. The same approach

was applied with UF membrane C and showed 4.2e5.2-log

inactivation and 4.9e6.0-log removal for GA and MS2,

respectively. For the conditions tested, the inactivation and

removal of MS2 phage was lower than GA phage with UF

membrane B.

The characterization of UF membranes tested with 0.1 mM

PBS solution, designed for drinking water treatment market

and actually used for research purpose, showed different

behaviours in terms of inactivation and removal particles,

especially for GA model. These results confirmed specific

interactions between membrane cutoff, electrostatic and

es after UF membranes A, B and C of bacteriophage-spiked

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 6 5 1e2 6 6 4 2661

hydrophobic properties of virus particles and UF membranes

(even if no data were available regarding hydrophobic prop-

erties of membranes tested). It is important to notice that

there is no consensus regarding a connection between pore

size and molecular-weight-cutoff (MWCO) for the UF

membrane cutoff determination (Causserand and Aimar,

2010). Moreover, no information regarding the pore size and

MWCO methods used has been provided by the membrane

manufacturers. The behaviour of bacteriophages during UF

processes is strongly dependent on interfacial properties of

bacteriophages but also membranes, associated to NOM and

biofouling. Further investigations should be undertaken to

improve fundamental aspects regarding these interfacial

properties, in order to explore mechanisms of phage removal

during UF processes.

From the suppliers point of view, this assessment of

membrane performances could then provide recommenda-

tions in terms of selection criteria for membranes for the

drinking water plant construction.

Infectious bact

0,0

1,0

2,0

3,0

4,0

5,0

6,0

7,0

0 10 20 30

Contact tim

Lo

g in

fectio

us b

acterio

ph

ag

e c

on

ce

ntra

tio

n

(lo

g(P

FU

/m

L))

Lo

Re

Total bacterio

0,0

1,0

2,0

3,0

4,0

5,0

6,0

7,0

8,0

0 10 20 30

Contact tim

Lo

g to

ta

l b

ac

te

rio

ph

ag

e co

nc

en

tra

tio

n

(lo

g(p

article

s/m

L))

Log

Resi

a

b

Fig. 6 e Logarithmic inactivation of bacteriophages (a) and loga

disinfection (after a bacteriophage concentration step in drinkin

The evaluation of the performances of the raw water pre-

treatment process, UF membranes alone or the combination

of both processes (complete treatment process) at pilot scale

could globally suggest that GA bacteriophagewould be a better

candidate than MS2 or Qb for assessing the capacities of

drinking water treatments with respect to viral risk in worst

conditions. However, further studies should be undertaken to

confirm its atypical status using other drinking water treat-

ments (e.g. ozonation, ultra-violets).

3.4. Inactivation of MS2 and GA bacteriophages byresidual chlorine

Drinking water treatment plants typically operated by

drinking water suppliers ensure at least 4.0-log removal for

viral risk based on the MS2 model. The results obtained at

pilot scale are in agreement with the recommendations for

water plants, since at least 4.2-log removal were already

achieved for MS2 bacteriophage by the end of the pre-

eriophages

40 50 60 70

e (min)

0

0,2

0,4

0,6

0,8

1

1,2

1,4

1,6

1,8

2

2,2

Ch

lo

rin

e resid

ual (m

g/L

)

g GA (PFU) Log MS2 (PFU)

sidual chlorine GA (mg/L) Residual chlorine MS2 (mg/L)

phages

40 50 60 70

e (min)

0

0,2

0,4

0,6

0,8

1

1,2

1,4

1,6

1,8

2

2,2

Ch

lo

rin

e resid

ual (m

g/L

)

GA (RT-PCR) Log MS2 (RT-PCR)

dual chlorine GA (mg/L) Residual chlorine MS2 (mg/L)

rithmic removal of bacteriophages (b) by residual chlorine

g water permeates).

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 6 5 1e2 6 6 42662

treatment process (Table 3). On the other hand, overall GA

bacteriophage removal was around 1.3e1.6-log reduction

after the pre-treatment process followed by the UF A

treatment.

Chlorine disinfection is widely used in drinking water

plants prior to distribution. Thus, in order to ensure secure

processes for drinking water based on the GA model (>4.0-log

reduction), kinetics of inactivation and removal of GA andMS2

bacteriophages by residual chlorine were performed (Fig. 6).

An initial chlorine concentration of 1.5 mg/L was injected into

the permeate after UF of the pre-treated water to obtain

a residual chlorine concentration of about 0.3 mg/L after

a contact time of 5 min. The sharp drop in residual chlorine

between 0 and 2minwas due to oxidation of NOM found in the

permeate, generating chlorination by-products with less

activity than residual chlorine.

Inactivation kinetics showed a rapid decrease of infectious

GA and MS2 bacteriophage concentrations during the first

2 min of contact time (w3.0-log and 4.8-log reduction,

respectively) (Fig. 6a). Logarithmic reduction of phage

concentrations was then slower because residual chlorine

concentration ranged around 0.3e0.4 mg/L. Within this

resource, MS2 bacteriophage appeared more sensitive to

residual chlorine than GA bacteriophage at equivalent contact

times. This result was confirmed by the kinetics of removal of

total GA and MS2 bacteriophages, although the decrease in

genomic RNA concentrationwas slower for both phages under

the same conditions (Fig. 6b). After a contact time of 60 min,

total GA and MS2 bacteriophage concentrations were 1.86-log

and 0.55-log particles/mL, respectively. The lowest loga-

rithmic reduction of total bacteriophages compared to infec-

tious bacteriophages is due to the sequential action of residual

chlorine which first affects the phage capsid integrity before

to degrade genomic RNA (Nuanualsuwan and Cliver, 2002;

Cliver, 2009).

MS2 bacteriophage higher sensitivity to residual chlorine

was unexpected since this disinfectant induces non-specific

oxidative damages against bacteriophages. Therefore, the

same disinfection tests were repeated on demineralizedwater

spiked to about 106 PFU/mL with either GA or MS2 bacterio-

phages. In order to determine residual chlorine effectiveness,

Ct4log values were calculated for both bacteriophages in both

resource types. In demineralized water, the results showed

that residual chlorine was indeed inactivating GA and MS2

bacteriophages in a non-specific way since Ct4log were similar,

with 1.1 and 1.5 mg min L-1, respectively. On the other hand,

Ct4log values for the permeate produced by the drinking water

treatment pilot were much higher, with a difference of

0.8 mg min L-1 between GA (7.0 mg min L-1) and MS2 bacte-

riophages (6.2 mg min L-1). This result suggests that permeate

NOM could impact on the effectiveness of GA and MS2 inac-

tivation during the disinfection tests.

Residual chlorine doses typically used in drinking water

plants classically operated by Veolia Water are usually tar-

geted at 0.3 mg/L for 30 min, i.e. an actual Ct of 9.0 mgmin L-1.

Insofar as the calculated Ct4log for GA bacteriophage are lower

than the actual Ct implemented in drinking water plants,

these plants can be considered as safe with respect to the viral

risk based on the GA model (>4-log viral reduction

guaranteed).

4. Conclusions

This study qualitatively and quantitatively assessed resis-

tance of MS2, GA and Qb bacteriophages towards surface

water treatments at pilot scale (clarification, sand filtration

andUF processes) representative of drinkingwater production

plants. The results indicated that MS2 and Qb bacteriophages

had a same behaviour during the application of the treatment

physical barriers, while GA displayed distinctive features.

Infectious and total bacteriophage logarithmic reductions for

MS2 and Qbwere above or equal to 4.0-log after the raw water

pre-treatment process and the UF membrane A (complete

treatment process), but under 2.2 and 1.6-log, respectively, in

the case of GA. Regarding GA bacteriophage, a chlorine

disinfection step after the complete treatment process ensure

the safety of drinking water plants in the conditions tested

(>4.0-log reduction).

Comparison of removal bacteriophages by the three UF

membranes tested, designed for drinking water treatment

market and actually used for research purpose, showed

specific relationship between membrane cutoff, electrostatic

interactions and hydrophobic properties, especially for GA

bacteriophage. This assessment of membrane performances

could provide recommendations for drinking water suppliers

in terms of selection criteria for membranes.

In the context of the assays undertaken at pilot scale, all

the results suggest that GA bacteriophage might be a more

suitable model for assessing the performance of drinking

water treatment processes in removing enteric viruses in

worst conditions than MS2, which is currently considered as

the reference model in the literature.

Further research is underway (i) to evaluate the effective-

ness of GA bacteriophage removal by other drinking water

treatments and (ii) to improve the knowledge on bacterio-

phage interfacial properties in relation to UF membrane-

based treatments.

Acknowledgements

This work was supported by Program grant ANR-07-PNRA-008

(ADHERESIST) from the ANR (Agence Nationale de la

Recherche).

r e f e r e n c e s

Abbaszadegan, M., Huber, M.S., Gerba, C.P., Pepper, I.L., 1993.Detection of enteroviruses in groundwater with thepolymerase chain reaction. Applied and EnvironmentalMicrobiology 59 (5), 1318e1324.

Aronino, R., Dlugy, C., Arkhangelsky, E., Shandalov, S., Oron, G.,Brenner, A., Gitis, V., 2009. Removal of viruses from surfacewater and secondary effluents by sand filtration. WaterResearch 43, 87e96.

Bosch, A., 1998. Human enteric viruses in the water environment:a minireview. International Microbiology 1, 191e196.

Butot, S., Putallaz, T., Sanchez, G., 2007. Procedure for rapidconcentration and detection of enteric viruses from berries

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 6 5 1e2 6 6 4 2663

and vegetables. Applied and Environmental Microbiology 73(1), 186e192.

Causserand, C., Aimar, P., 2010. Comprehensive membranescience and engineering. In: Drioli, Enrico, Giorno, Lidietta(Eds.), 2010. Basic Aspects of Membranes Science andEngineering, vol. 1. Elsevier Publishers, p. 368.

Cavereau, D., Lenes, D., Arnal, C., Arfi, C., Joret, J.C., 2009.Assessment of virus contamination risk in water resources.European Journal of Water Quality 40, 1e15.

Cliver, D.O., 2009. Capsid and infectivity in virus detection. Foodand Environmental Virology 1, 123e128.

Fane, S.A., Ashbolt, N.J., White, S.B., 2002. The implications ofsystem scale for cost and pathogen risk. Water Science andTechnology 46, 281e288.

Fiksdal, L., Leiknes, T., 2006. The effect of coagulation with MF/UFmembrane filtration for the removal of virus in drinkingwater. Journal of Membrane Science 279 (1e2), 364e371.

Golmohammadi, R., Fridborg, K., Bundule, M., Valegard, K.,Liljas, L., 1996. The crystal structure of bacteriophages Qb at3.5 A resolution. Structure 4 (5), 543e554.

Gott, J.M., Wilhelm, L.J., Uhlenbeck, O.C., 1991. RNA bindingproperties of the coat protein from bacteriophage GA. NucleicAcids Research 19 (23), 6499e6503.

Herath, G., Yamamoto, K., Urase, T., 1999. Removal of viruses bymicrofiltration membranes at different solutionenvironments. Water Science and Technology 40 (4e5),331e338.

Hu, J.Y., Ong, S.L., Song, L.F., Feng, Y.Y., Liu, W.T., Tan, T.W.,Lee, L.Y., Ng, W.J., 2003. Removal of MS2 bacteriophage usingmembrane technologies. Water Science and Technology 47(12), 163e168.

Huertas, A., Barbeau, B., Desjardins, C., Galarza, A.,Figueroa, M.A., Toranzos, G.A., 2003. Evaluation of Bacillussubtilis and coliphage MS2 as indicators of advanced watertreatment efficiency. Water Science and Technology 47 (3),255e259.

IAWPRC Study Group on Health Related Water Microbiology,1991. Bacteriophages as model viruses in water qualitycontrol. Water Research 25, 529e545.

Inokuchi, Y., Takahashi, R., Hirose, T., Inayama, S., Jacobson, A.B.,Hirashima, A., 1986. The complete nucleotide sequence of thegroup II RNA coliphage GA. Journal of Biochemistry 99,1169e1180.

ISO 10705-1, 1995. Water Quality-Detection and Enumeration ofBacteriophages, Part 1. Enumeration of F-specific RNABacteriophages. International Organization forStandardization, Geneva, Switzerland.

Jacangelo, J.G., Adham, S.S., Laine, J.M., 1995. Mechanism ofCryptosporidium, Giardia, and MS2 virus removal by MF and UF.Journal of the American Water Works Association (AWWA) 87(9), 107e121.

Jacangelo, J.G., Rhodes-Trussel, R., Watson, M., 1997. Role ofmembrane technology in drinking water treatment in theUnited States. Desalination 113, 119e127.

Langlet, J., Gaboriaud, F., Duval, J.F.L., Gantzer, C., 2008.Aggregation and surface properties of F-specific RNA phages:implication for membrane filtration processes. WaterResearch 42, 2769e2777.

Langlet, J., Ogorzaly, L., Schrotter, J.C., Machinal, C., Gaboriaud, F.,Duval, J.F.L., Gantzer, C., 2009. Efficiency of MS2 phage and Qb

phage removal by membrane filtration in water treatment:applicability of real-time RT-PCR method. Journal ofMembrane Science 326, 111e116.

Lechevallier, M.W., Au, K.K., 2004. Water Treatment and PathogenControl: Process Efficiency in Achieving Safe Drinking Water.World Health Organization.

Lodder, W.J., Van den Berg, H.H., Rutjes, S.A., De RodaHusman, A.M., 2010. Presence of enteric viruses in source

waters for drinking water production in The Netherlands.Applied and Environmental Microbiology 76 (17), 5965e5971.

Machinal, C., Torres, S., Schrotter, J.C., 2006. Expertise centre ofexcellence ARAMIS dedicated to membrane plant design andoperation. Desalination 200 (1e3), 757e759 (11/12).

Matsui, Y., Matsushita, T., Sakuma, S., Gojo, T., Mamiya, T.,Suzuoki, H., Inoue, T., 2003. Virus inactivation in aluminiumand polyaluminium coagulation. Environmental Science andTechnology 37 (22), 5175e5180.

Matsushita, T., Matsui, Y., Shirasaki, N., Kato, Y., 2005. Effect ofmembrane pore size, coagulation time, and coagulant dose onvirus removal by a coagulation-ceramic microfiltration hybridsystem. Desalination 178, 21e26.

Meng, Q.S., Gerba, C.P., 1996. Comparative inactivation of entericadenovirus, poliovirus and coliphages by ultravioletirradiation. Water Research 30 (11), 2665e2668.

Meschke, J.S., Sobsey, M.D., 2003. Comparative reduction ofNorwalk virus, poliovirus type 1, Fþ RNA coliphage MS2 andEscherichia coli in miniature soil columns. Water Science andTechnology 47 (3), 85e90.

Mylon, S.E., Rinciog, C.I., Schmidt, N., Gutierrez, L., Wong, G.C.,Nguyen, T.H., 2010. Influence of salts and natural organicmatter on the stability of bacteriophage MS2. Langmuir 26 (2),1035e1042.

Nuanualsuwan, S., Cliver, D.O., 2002. Pretreatment to avoidpositive RT-PCR results with inactivated viruses. Journal ofVirological Methods 104, 217e225.

Ogorzaly, L., Gantzer, C., 2006. Development of real-time RT-PCRmethods for specific detection of F-specific RNA bacteriophagegenogroups: application to urban raw wastewater. Journal ofVirological Methods 138, 131e139.

Otaki, M., Yano, K., Ohgaki, S., 1998. Virus removal ina membrane separation process. Water Science andTechnology 37 (10), 107e116.

Pham, M., Mintz, E.A., Nguyen, T.H., 2009. Deposition kinetics ofbacteriophage MS2 to natural organic matter: role of divalentcations. Journal of Colloid and Interface Science 338, 1e9.

Redman, J.A., Grant, S.B., Olson, T.M., 1997. Filtration ofrecombinant Norwalk virus particles and bacteriophage MS2in quartz sand: importance of electrostatic interactions.Environmental Science and Technology 37 (12), 3378e3383.

Sattar, S.A., Tetro, J., Springthorpe, V.S., 1999. Impact of changingsocietal trends on the spread of infections in American andCanadian homes. American Journal of Infection Control 27,S4eS21.

Schaldach, C.M., Bourcier, W.L., Shaw, H.F., Viani, B.E.,Wilson, W.D., 2006. The influence of ionic strength on theinteraction of viruses with charged surfaces underenvironmental conditions. Journal of Colloid and InterfaceScience 294 (1), 1e10.

Schijven, J.F., Hassanizadeh, M., 2000. Removal of viruses by soilpassage: overview of modeling, processes, and parameters.Critical Reviews in Environmental Science and Technology 30,49e127.

Shin, G.A., Sobsey, M.D., 2003. Reduction of Norwalk virus,poliovirus 1, and bacteriophage MS2 by ozone disinfection ofwater. Applied and Environmental Microbiology 69 (7),3975e3978.

Shirasaki, N., Matsushita, T., Matsui, Y., Urasaki, T., Ohno, K.,2009a. Comparison of behaviors of two surrogates forpathogenic waterborne viruses, bacteriophages Qb and MS2,during the aluminium coagulation process. Water Research43, 605e612.

Shirasaki, N., Matsushita, T., Matsui, Y., Kobuke, M., Ohno, K.,2009b. Comparison of removal performance of two surrogatesfor pathogenic waterborne viruses, bacteriophage Qb andMS2, in a coagulation-ceramic microfiltration system. Journalof Membrane Science 326, 564e571.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 2 6 5 1e2 6 6 42664

Sobsey, M.D., Battigelli, D.A., Shin, G.A., Newland, S.S., 1998. RT-PCR amplification detects inactivated viruses in water andwastewater. Water Science and Technology 38 (12), 91e94.

Springthorpe, S., Sattar, S.A., 2007. Virus removal during drinkingwater treatment. Human Viruses in Water 6, 109e126.

Szewzyk, U., Szewzyk, R., Manz, W., Schleifer, K.H., 2000.Microbiological safety of drinking water. Annual Review ofMicrobiology 54, 81e127.

Tars, K., Bundule, M., Fridborg, K., Liljas, L., 1997. The crystalstructure of bacteriophage GA and a comparison ofbacteriophages belonging to the major groups of Escherichiacoli Leviviruses. Journal of Molecular Biology 271 (5), 759e773.

Templeton, M.R., Andrews, R.C., Hofmann, R., 2007. Removal ofparticle-associated bacteriophages by dual-media filtration atdifferent filter cycle stages and impacts on subsequent UVdisinfection. Water Research 41, 2393e2406.

Thurston-Enriquez, J.A., Haas, C.N., Jacangelo, J., Riley, K.,Gerba, C.P., 2003. Inactivation of feline calicivirus and

adenovirus type 40 by UV radiation. Applied andEnvironmental Microbiology 69 (1), 577e582.

Urase, T., Yamamoto, K., Ohgaki, S., 1996. Effect of pore structureof membranes and module configuration on virus retention.Journal of Membrane Science 115 (1), 21e29.

Van Voorthuizen, E.M., Ashbolt, N.J., Schafer, A.I., 2001. Role ofhydrophobic and electrostatic interactions for initial entericvirus retention by MF membranes. Journal of MembraneScience 194, 69e79.

Villermaux, J., 1993. Genie de la reaction chimique: conception etfonctionnement des reacteurs. TEC & DOC-Lavoisier,167e174.

WHO, 2006. Guidelines for Drinking-Water Quality. IncorporatingFirst and Second Addenda, third ed. World HealthOrganization.

Zhu, B., Clifford, D.A., Chellam, S., 2005. Virus removal by ironcoagulation-microfiltration. Water Research 39 (20),5153e5161.