VIRUS CONCENTRATION BY CROSSFLOW MEMBRANE
FILTRATION: EFFECT OF HYDRODYNAMIC CONDITIONS
AND MEMBRANE PROPERTIES
By
Elodie Pasco

A DISSERTATION
Submitted to
Michigan State University
In partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Environmental Engineering
2012

ABSTRACT

Crossflow membrane filtration is emerging as a concentration method for the recovery and
detection of viruses in large volumes of water. To increase virus recovery, the membrane is
usually “blocked” with a proteinaceous solution or a chemical dispersant. This study explores a
novel approach using controlled and rapid (<1 h) layer-by-layer adsorption of polyelectrolytes to
form an anti-adhesive, sacrificial layer on the membrane surface. Membranes coated with
polyelectrolyte multilayers (PEMs) can ensure up to 2-fold higher pre-elution recoveries of P22
bacteriophage than membranes blocked with calf serum (CS). Calculations of virus-membrane
interaction energies confirm that the higher recoveries with PEM-coated membranes are due to
higher charge and hydrophilicity relative to the CS-blocked counterpart. The pre-elution
recovery with PEM-coated membranes at high cross flow rates is >70%, which might eliminate
the need for elution and backflushing steps, thus reducing both the time required for pathogen
concentration and the volume of the final sample. Operating the filter at high transmembrane
pressures and crossflow rates leads to a significant increase in the concentration rate; for
2

example, for CS-blocked membranes the initial rate of 5.8 ± 0.3 L/(min·m ) was achieved
without affecting P22 recovery; this flux value is almost 4 times higher than values reported in
the literature. In addition, the effects of crossflow and high permeate flow rate on the recovery
and removal of infective virus from DI and surface water are evaluated using the CS-blocked
membrane. An increase in the average permeate flux (up to 5-fold for DI water and 2-fold for

surface water over values reported in literature) does not affect the recovery or removal of
infective P22. Finally, the performance of the anti-adhesive PEM membrane coatings with
complex water matrices is evaluated. The data are only preliminary as only a small number of
filtrations have been performed but results seem to indicate that PEM-coated membranes can
maintain their advantage over CS-blocked membrane in terms of P22 pre-elution recovery when
concentrating viruses from MBR effluent. When challenged by MBR effluent, PEM-coated
membranes also appear to be more resistant to fouling than their CS-blocked counterparts. With
surface water however, no difference was observed.

ACKNOWLEDGEMENTS

I would like to thank my advisor Dr. Volodymyr Tarabara for his time and guidance on this
research project and his patience and encouragement in difficult times.

I would like to thank my committee members Dr. Merlin L. Bruening, Dr. Syed A. Hashsham
and Dr. Irene Xagoraraki for their suggestions and help with this project.

I would also like to thank Lori Larner for being so helpful and efficient with all administrative
matters and more.

iv

TABLE OF CONTENTS

List of Tables ............................................................................................................................... viii
List of Figures ................................................................................................................................ ix
Abbreviations ................................................................................................................................. xi
Chapter 1. Introduction ................................................................................................................... 1
1.1 Waterborne pathogen collection and quantification: importance and challenges ............ 1
1.1.1
Waterborne disease outbreaks................................................................................... 1
1.1.2
Virus detection: Needs, challenges and current methods ......................................... 2
1.2 Polyelectrolyte multilayer films as anti-adhesive recoverable membrane coatings ...... 11
1.2.1
Polyelectrolyte multilayer (PEM) films .................................................................. 11
1.2.2
Anti-adhesive performance of PEMs ...................................................................... 11
1.2.3
PEM films as sacrificial layers ............................................................................... 15
1.3 Virus interaction with inert surfaces: Methods, challenges and current knowledge ...... 15
1.3.1
Virus surface characterization................................................................................. 15
1.3.2
Quantifying interactions between viruses and a surface ......................................... 19
1.3.3
Current knowledge .................................................................................................. 24
1.4 Dissertation overview..................................................................................................... 26
References ................................................................................................................................. 28
Chapter 2. Anti-adhesive membrane films for virus concentration and recovery from water prior
to detection .................................................................................................................................... 42
2.1 Introduction .................................................................................................................... 42
2.2 Materials and Methods ................................................................................................... 46
2.2.1
Reagents .................................................................................................................. 46
2.2.2
Bacteriophage ......................................................................................................... 47
2.2.3
Bacteriophage characterization ............................................................................... 47
2.2.4
Crossflow concentration apparatus ......................................................................... 48
2.2.5
Membrane preparation ............................................................................................ 49
2.2.6
Membrane characterization ..................................................................................... 50
2.2.7
Concentration procedure ......................................................................................... 50
2.2.8
Sample analysis....................................................................................................... 52
2.2.9
Data analysis and statistics...................................................................................... 53
2.2.10 Determination of the fouling resistance Rf ............................................................. 54
2.2.11 Blocking law analysis ............................................................................................. 55
2.2.12 Calculation of the energy of interaction between P22 phage and membrane surfaces
using the XDLVO approach ................................................................................................. 55
2.2.13 Effect of polyelectrolytes and CS on qPCR analysis .............................................. 57
2.3 Results and Discussion ................................................................................................... 58
2.3.1
PEM selection ......................................................................................................... 58
v

2.3.2
Properties of CS-blocked and PEM-coated membranes and P22 phage ................ 59
2.3.2.1 Membrane permeability and fouling mechanisms .............................................. 59
2.3.2.2 Surface properties of membranes and P22 phage ............................................... 60
2.3.2.3 Virus-membrane interaction energies ................................................................. 64
2.3.3
Virus recovery with PEM-coated and CS-blocked membranes ............................. 65
2.3.3.1 Pre-elution recovery ............................................................................................ 65
2.3.3.2 Post-elution recovery and potential of PEMs as sacrificial coatings .................. 67
2.3.3.3 Reproducibility .................................................................................................... 69
2.3.4
Virus log removal ................................................................................................... 71
2.3.5
Effect of operational conditions on virus concentration ......................................... 72
2.3.5.1 Operational conditions and flux .......................................................................... 72
2.3.5.2 Effect of Jcf / J p on virus recovery ................................................................... 75
2.3.6
Potential for practical applications.......................................................................... 77
Acknowledgements ................................................................................................................... 78
References ................................................................................................................................. 80
Chapter 3. Effect of flow conditions and water composition on the removal and recovery of
infective bacteriophage by crossflow filtration with CS-blocked membranes ............................. 86
3.1 Introduction .................................................................................................................... 86
3.2 Material and Methods..................................................................................................... 88
3.2.1
Reagents .................................................................................................................. 88
3.2.2
Water samples ......................................................................................................... 88
3.2.3
Bacteriophage ......................................................................................................... 89
3.2.4
Membrane preparation, crossflow concentration apparatus and concentration
procedure............................................................................................................................... 89
3.2.5
Sample analysis....................................................................................................... 90
3.2.6
Calf serum layer stability test ................................................................................. 91
3.3 Results and Discussion ................................................................................................... 92
3.3.1
Effect of filtration conditions on the recovery and removal of infective P22 ........ 92
3.3.2
qPCR vs. plaque assay ............................................................................................ 97
3.3.3
DI water vs. surface water .................................................................................... 100
3.3.4
Effect of filtration condition on CS layer integrity ............................................... 101
3.3.5
Conclusion ............................................................................................................ 103
References ............................................................................................................................... 105
Chapter 4. Virus recovery and concentration from complex water matrices using PEM-coated
membranes: Preliminary assessment .......................................................................................... 107
4.1 Introduction .................................................................................................................. 107
4.2 Materials and Methods ................................................................................................. 109
4.2.1
Reagents ................................................................................................................ 109
4.2.2
Water samples ....................................................................................................... 109
4.2.3
Water quality testing. ............................................................................................ 109
4.2.4
Bacteriophage ....................................................................................................... 110
4.2.5
Membrane preparation, crossflow concentration apparatus and concentration
procedure............................................................................................................................. 110
4.2.6
Sample analysis..................................................................................................... 111
vi

4.3 Results and Discussion ................................................................................................. 112
4.3.1
Water quality......................................................................................................... 112
4.3.2
P22 recovery and removal from MBR effluent and surface water samples ......... 113
4.3.3
Effect of backflushing on P22 recovery from surface water. ............................... 117
4.3.4
Fouling resistance of CS-blocked and PEM-coated membrane ........................... 117
4.3.5
Conclusion ............................................................................................................ 118
References ............................................................................................................................... 121
Chapter 5. Suggested future research.......................................................................................... 122
APPENDICES ............................................................................................................................ 126
Appendix A. Determination of crossflow and permeate fluxes used by Morales-Morales et
al. [1] and Hill et al. [2-4] .................................................................................................... 126
References ....................................................................................................................... 128
Appendix B. Addition of carrier RNA to sample: effect of DNA extraction/detection ...... 130
References ....................................................................................................................... 132
Appendix C. Compaction behavior of PEM-coated membranes and CS-blocked membranes.
............................................................................................................................................. 134
Appendix D. PEM layer stability during filtration and elution. .......................................... 135
Appendix E. Preliminary work on visualization of virus adhesion on surfaces.................. 137
References ....................................................................................................................... 142

vii

LIST OF TABLES

Table 1.1: Virus recovery using the VIRADEL method ................................................................ 4
Table 1.2: Virus recovery by methods other than crossflow filtration or VIRADEL .................... 9
Table 1.3: Virus recovery by crossflow filtration ......................................................................... 10
Table 2.1: Primer and probe sequence (from Masago et al. [45]) ................................................ 53
Table 2.2 Measured contact angles, calculated surface energy parameters and free energy of
interfacial interaction when immersed in water for P22 phage, CS-blocked membranes and PEMcoated membranes. ........................................................................................................................ 63
Table 4.1: MBR effluent and surface water TOC levels ............................................................ 113
Table 5.1: Virus recovery performance of PEM-coated membrane and calf serum-blocked
membrane at the pre-elution (rpre), post-elution (rpost) and coating removal - through elution or
backflush - (rsacrificial) stages. ................................................................................................... 124
Table B.1: Effect of carrier RNA addition on CP value obtained by qPCR ............................... 131

viii

LIST OF FIGURES

Figure 2.1: Comparison of the proposed virus pre-concentration approach based on
polyelectrolyte multilayer coatings (right) with other current methods. ...................................... 44
Figure 2.2: Crossflow filtration apparatus .................................................................................... 49
Figure 2.3: Filtration conditions in experiments on P22 preconcentration and recovery using CSblocked and PEM-coated membranes.. ......................................................................................... 51
Figure 2.4: Effect of various polyelectrolytes and calf serum on the P22 concentration
determined by qPCR. Error bars indicate the standard deviation interval.. ................................. 59
Figure 2.5: Blocking law analysis for fouling of CS-blocked and PEM-coated membranes during
filtration of DI water spiked with P22 stock. A) Distributions and contributions of the blocking
mechanisms in terms of permeate volume. B) Inverse cumulative flow rate (t/V) as a function of
permeate volume V for CS-blocked and PEM-coated membranes. ............................................. 61
Figure 2.6: XDLVO energy profiles for the interaction of P22 with a) CS-blocked membranes
and b) PEM-coated membranes at pH = 6 and in 10 mM KCl. .................................................... 65
Figure 2.7: Pre-elution and post-elution (see inset) recovery of P22 from DI water.................... 67
Figure 2.8: Relative frequency of occurrence of P22 (a) pre-elution and (b) post-elution
recoveries from DI water under the operational conditions optimized to maximize P22 recovery
....................................................................................................................................................... 70
Figure 2.9: P22 removal from DI water by CS-blocked and PEM-coated membrane at various
J cf / J p . ....................................................................................................................................... 71
Figure 2.10: Observed fouling resistance (averaged over the duration of experiment) for CSblocked and PEM-coated membranes during the concentration of 1L of DI water seeded with
P22 to 250 mL at various crossflow to permeate flow ratios. ...................................................... 73
Figure 2.11: Filtration conditions: average permeate flux and crossflow flux, used for P22
recovery with CS-blocked membrane and PEM-coated membranes ........................................... 74
Figure 2.12: Pre-elution recovery of P22 from DI water as a function of the fouling resistance to
membrane resistance ratio............................................................................................................. 77
Figure 3.1: Filtration conditions in experiments on infective P22 concentration and recovery
from DI water and surface water using CS-blocked membranes… ............................................. 90

ix

Figure 3.2: Effect of crossflow filtration on a) pre-elution recovery, b) post-elution recovery, c)
removal of infective P22 by crossflow filtration with CS-blocked membrane ............................ 94
Figure 3.3: Effect of crossflow filtration conditions on fouling, measured by the ratio of final
over initial permeate flux for DI water and surface water. ........................................................... 96
Figure 3.4: Effect of analysis method on P22 a) pre-elution recovery, b) post-elution recovery,
and c) removal by crossflow filtration with CS-blocked membrane. One filtration was run for
each set of conditions.. .................................................................................................................. 99
Figure 3.5: Dextran rejection by CS-blocked membrane under increasing Jcf / J p during a)
filtration and b) elution ............................................................................................................... 102
Figure 4.1: Filtration conditions in experiments on P22 concentration and recovery from MBR
..................................................................................................................................................... 111
Figure 4.2: a) P22 recovery and b) removal from DI water (data from Chapter 2), MBR effluent
and surface water with CS-blocked and PEM-coated membranes under “highest recovery”
conditions. ................................................................................................................................... 115
Figure 4.3: a) P22 recovery and b) removal from MBR effluent with PEM-coated membrane at
high and low permeate fluxes… ................................................................................................. 116
Figure 4.4: Fouling observed for CS-blocked membrane and PEM-coated membrane during
filtration tests performed to recover P22 from DI water, MBR effluent and surface water under a)
“highest recovery” conditions and b) “fastest filtration” conditions. ......................................... 119
Figure C.1: Compaction and decompaction behavior of bare PES membranes, CS-blocked
membranes and PEM-coated membranes. .................................................................................. 134
Figure D.1: Dextran rejection by PEM-coated membrane under increasing Jcf / J p during a)
filtration and b) elution.. ............................................................................................................. 136
Figure E.1: AFM topographic images of a) (PDADMA/PSS)4.5 on gold substrate b) and c)
(PDADMA/PSS)4.5 on mica dipped in P22 suspension ............................................................. 138
Figure E.2: Fluorescence images of serial dilution samples. a) corresponds to the 1:1000 dilution,
4
5
6
b) to the 1:10 dilution, c) to the 1:10 dilution and d) to the 1:10 dilution ............................. 140
Figure E.3: Fluorescence images of a) (PSS/PAH)4-coated ; b) (PSS/PAH)4.5-coated and c) bare
anodisc filters dipped in P22 suspension. ................................................................................... 141

x

ABBREVIATIONS

AFM

atomic force microscopy

BE

beef extract

BSA

bovine serum albumin

CHI

chitosan

CN

cellulose nitrate

CS

calf serum

DE

diatomaceous earth

DI

deionized

DLVO

Derjaguin, Landau, Verwey and Overbeek

E1

echovirus 1

FLOC

flocculation

HA

hyaluronic acid

HAdV

human adenovirus

HAV

hepatitis A virus

HE

heparin

HSA

human serum albumin

IEP

isoelectric point

J cf

crossflow rate

Jp

permeate flux

LRV

log removal

xi

MBR

membrane bioreactor

MNoV

murine norovirus

MWCO

molecular weight cut off

NaPP

sodium polyphosphate

NOM

natural organic matter

NoV

norovirus

PAA

poly(acrylic acid)

PAAm

polyacrylamide

PAH

poly(allylamine hydrochloride)

PAN

polyacrylonitrile

PDADMAC

poly(dimethyldiallylammonium)

PEG

poly(ethylene glycol)

PEI

polyethyleneimine

PEM

polyelectrolyte multilayer

PGA

poly(L-glutamic acid)

PLL

poly(L-lysine)

PSU

polysulfone

PV

poliovirus

QCM

quartz crystal microbalance

RV-A

group A rotavirus

SW

sewage

TW

tap water

XDLVO

extended Derjaguin, Landau, Verwey and Overbeek

xii

Chapter 1. Introduction

1.1

1.1.1

Waterborne pathogen collection and quantification: importance and challenges

Waterborne disease outbreaks

Waterborne diseases are infections transmitted through contact with or consumption of
contaminated waters. The main etiological agents are bacteria, protozoa and viruses, and the
main reservoirs of concern are recreational waters and surface and groundwater used as sources
of drinking water. Progress in drinking water treatment and regulation since the beginning of the
th

20 century [1] led to a decrease in the occurrence of waterborne disease outbreaks from
exposure to contaminated drinking water. As a result, such outbreaks have become relatively
uncommon in the U.S. and other developed countries [2]. For example, the Center for Disease
Control listed on average 18 disease outbreaks associated with drinking water per year for the
1999-2008 period in the United States (actual outbreak occurrences might have been higher as it
has been suggested that only 10% to 50% of outbreaks are actually reported [1]). These
outbreaks resulted in an average of about 1,000 sick individuals and 2 deaths per year [3-7].
Since 2000, approximately 15% of the outbreaks for which the etiological agent was identified
have been caused by virus contamination [4-7]. In contrast, occurrences of waterborne disease
outbreaks caused by exposure to contaminated recreational water in the U.S. are on the rise from
an average of 31 outbreaks resulting in about 1,200 illness per year for the 1999-2004 period to
an average of 53 outbreaks resulting in about 4,600 illness per year for the 2005-2008 period [3,
8-11]. Virus-associated outbreaks accounted for 3% to 10% of the events for which the
1

etiological agents has been identified, with norovirus being the most prevalent viral etiological
agent identified. A recent review of all published literature on virus-associated waterborne
disease outbreaks also found norovirus to be the largest cause of outbreaks with adenoviruses
and echoviruses coming in second and third position, respectively [12]. In swimming pools,
where almost half of outbreaks occurred, inadequate disinfection was often the culprit. In lakes
and ponds, the second largest affected recreational type of water, water quality surveillance is the
most effective means of preventing outbreaks [12].

1.1.2

Virus detection: Needs, challenges and current methods

Detection of waterborne pathogens is essential because: i) although deficiencies of the
water treatment system can be minimized, zero risk cannot be achieved [13] and the
consequences of one incident can be very large such as in the cases of the 2000 Walkerton, ON
outbreak that affected over 2,000 people and killed 6 [14]; or the 1993 Milwaukee outbreak
where more than 400,000 were affected [1, 15, 16]. And as the number of people more
susceptible to infection (individuals with weak immune systems) increases [2], the extent of
these consequences can be expected to increase . ii) The detection of pathogen in recreational
water is a key factor in preventing outbreaks through this route of exposure [12, 17]. iii) As
direct and indirect water reuse increases [18, 19], the potential for drinking water contamination
increases and so does the need for monitoring the microbiological quality of water. This is
especially true for viruses as several studies have shown that viruses were more resistant than
bacteria with respect to environmental conditions and other stressors including disinfection [2].
Infective viruses have been found in 5% of groundwater samples taken from 35 different states

2

[20] and were also found in drinking water that met bacteriological standards and contained an
adequate level of chlorine [1, 2, 21].

The methods currently employed to concentrate and quantify pathogens in water are
complex and time consuming. The first difficulty is that there is currently no approved method
to simultaneously collect and concentrate all microorganisms [22, 23]. The VIRADEL (VIRus
ADsorption-ELution) method, which is the EPA-approved method of virus collection, involves
the filtration of a large sample (up to 1,000 L for drinking water) with a microfilter either
positively charged [24-33] or negatively charged [26, 30, 32, 34-36]. Positively charged filters
are usually favored as the use of negatively charged ones requires the acidification of the sample
at a pH of 3.5 and sometimes the addition of multivalent cation salts [32, 36], which might affect
virus integrity and infectivity [37]. During filtration, viruses are adsorbed by the filter (mostly
through electrostatic interactions) and are later eluted using a basic 1.5% to 6% beef extract (BE)
solution and/or 0.05 M glycine solution. If further volume reduction is required, this step can be
followed by flocculation to precipitate viruses [32]. The method greatly differs from methods of
bacteria collection and protozoa collection. The former usually involves collection of bacteria at
the surface of a microfiltration membrane [38, 39] while the latter includes depth filtration
followed by elution with a detergent solution and centrifugation [40].

3

Table 1.1: Virus recovery using the VIRADEL method
Virus

Water

1

Flocculation
step

2
Filter (+/-) /
3
Treatment

Sample amendment

coliphage TW

no

NA

double 1-MDS (+)

coliphage TW

no

NA

CN (-)/PEI

coliphage TW

no

NA

coliphage SW

no

coliphage SW

Elution
6

Recovery (%) Ref.

3% BE , pH 9

12 ± 6

[25]

4

3% BE, pH 9

44 ± 8

[25]

DE (-)/Nalco

5

3% BE, pH 9 – drop by drop

103 ± 7

[25]

NA

DE (-)/Nalco

3% BE, pH 9 – drop by drop

71 ± 15

[25]

no

NA

double 1-MDS (+)

3% BE, pH 9

39 ± 5

[25]

coliphage SW

no

NA

DE (-)/Nalco

3% BE, pH 9

47 ± 15

[25]

coliphage SW

no

NA

Double 1-MDS (+)

3% BE, pH 9.5

14

[30]

coliphage SW

no

NA

50 S Zeta Plus (+)

3% BE, pH 9.5

14

[30]

coliphage SW

no

pH 3.5/ AlCl3

Filterite (-)

3% BE, pH 9.5

17

[30]

coliphage sea

no

Prefiltered

cellulose ester

1.5% BE, pH 9

27

[37]

4

Table 1.1 (cont’d)
MS2

TW

no

NA

DE (-)/FeCl3 + AlCl3 3% BE pH 9

53 ± 12

[41]

PV1

TW

no

NA

Nanoceram (+)

1.5 % BE/0.05M Gly pH 9

51 ± 26

[27]

PV1

TW

no

NA

1-MDS (+)

1.5 % BE/0.05M Gly pH 9

67 ± 6

[27]

PV1

TW

no

NA

MK (+)

3% BE/0.05M Gly pH 9.5

101 ± 57

[28]

PV1

TW

no

NA

50 S Zeta Plus (+)

0.3% BE/).05 M Gly, pH 9.5

56 ± 8

[29]

PV1

sea

yes

pH 3.5/ AlCl3

epoxy fiberglass (-)

6% BE pH 10.5

36 ± 5

[34]

PV3

TW

yes

NA

glass wool

3% BE, 0.5M Gly, PH 9.5

98 ± 24

[42]

HNoV

TW

no

25 mM MgCl2

cellulose esters (-)

0.5 mM H2SO4/1mM NaOH

55 to 104

[26]

HNoV

river

no

25 mM MgCl2

cellulose esters (-)

0.5 mM H2SO4/1mM NaOH

11 to 18

[26]

HNoV

TW

no

NA

1-MDS (+)

3% BE pH 9

14 to 46

[26]

5

7

Table 1.1 (cont’d)
HAdV

TW

yes

NA

glass wool

3% BE, 0.5M Gly, PH 9.5

28 ± 14

[42]

HAdV

GW

yes

NA

glass wool

3% BE, 0.5M Gly, PH 9.5

8±4

[42]

HAdV

TW

no

NA

MK (+)

1% BE, 0.25 N Gly, pH 9.5

0.007 to 0.02 [24]

1
2
3
4
5
6
7

type of water: TW: tap water, SW: sewage, GW: groundwater
filter surface charge
filter pretreatment
CN: cellulose nitrate
DE diatomaceous earth
BE: beef extract
Gly: glycine

6

The VIRADEL method also suffers, in some cases, from poor recovery efficiencies and
large standard deviations (Table 1.1) [23]. For example, Borrego et al. observed that less than
35% of the indigenous Escherichia coli phages present in tap water spiked with sewage were
adsorbed on the Virosorb 1-MDS filter (the only electropositive filter suggested by EPA for the
VIRADEL method), and less than 12% were recovered by elution [25]. Low elution recovery
number were also observed from Abdelzaher et al. who only eluted up to 35% of the coliphage
spiked in seawater and adsorbed on a negatively charged cellulose membrane [37]. The method
is also very sensitive to water quality. Haramoto et al. showed that recovery of norovirus (NoV)
spiked in surface water was 80% less than the recovery observed when NoV was spiked in
deionized water [26]. This leads to unacceptably low recoveries for certain kinds of viruses such
as adenovirus for which recovery as low as 0.02% has been reported [24]. Several groups have
investigated other positively charged filters as alternatives to Virosorb microfilters, because of
Virosorb’s high cost and vulnerability with respect to clogging by colloids [25, 27, 28, 30, 31,
43]. The MK filter was less efficient (lower recovery and greater variability) [28], while Zeta
Plus 50S,which seems as efficient as Virosorb to recover virus from various type of water, was
not available as a cartridge and thus limited to smaller water volumes [30, 31]. Nanoceram, a
cheaper positively charged filter was found to be as efficient as Virosorb [27]. Negatively
charged filters made of diatomaceous earth (DE) treated with cationic polymer solution (such as
Nalco 7111) showed improved coliphage recovery from sewage and sewage spiked tap water
samples [25] when compared with Virosorb filters, but unless a drop by drop elution was
performed, coliphage recovery was still less than 50% for sewage effluent samples (data not
available for tap water samples).

7

Other methods such as ultracentrifugation [44, 45], centrifugal ultrafiltration [46],
flocculation-re-dissolution-ultrafiltration [47], and tangential filtration have also been
investigated (Table 1.2). Tangential filtration (also known as crossflow filtration) has emerged
as the most promising alternative to VIRADEL [24, 48-54] due to its ability to concentrate large
volumes of water relatively fast. Centrifugation-based methods are still of interest for secondary
concentration purposes. The crossflow filtration method has been shown to give significantly
better results than the VIRADEL method [48] and can also be used to collect and concentrate
bacteria and protozoa [55-61]. Crossflow ultrafiltration also presents the advantage of
maintaining virus integrity as, ideally, virions remain suspended in the retentate throughout the
concentration process. At the same time, low molecular weight solutes, which can potentially
inhibit qPCR detection, can pass through the membrane and are not concentrated [22]. In
practice, hollow fiber modules have been the most commonly used membrane configuration [48,
50, 51, 60, 61, 65, 66]. These filters have the advantage of a high surface area to volume ratio
allowing practical transport of the concentration device to the filtration site if needed and
filtration of large volumes of water in a relatively short amount of time. They can also be
reusable, thus making the procedure more cost-effective.

8

Table 1.2: Virus recovery by methods other than crossflow filtration or VIRADEL
Virus

1
Water

Method

Recovery (%)

Ref.

RV-A

SW

Ultracentrifugation

Qbeta

TW

Electronegative membrane vortex

HAdV 2

eluates

Centrifugal filtration

33 ± 14

[63]

PV1

eluates

Centrifugal filtration

95 ± 5

[63]

3

sea

ultracentrifugation

76%

[64]

3

sea

flocculation

22%

[64]

Viruses
Viruses
1
2

34-60

[45]

2 26-67

[62]

type of water: TW: tap water, SW: sewage

i) Metallic cations are added to water samples. ii) Samples are then filtered on
electronegative microporous membranes. iii) Membranes are then vortexed in presence of
an elution buffer. iv) the elution buffer is centrifuged to separate membrane debris from
viruses in suspension
3
enumerated by epifluorescence and consequently species were not identified (other virus
were enumerated by plaque assay or qPCR)

Nevertheless, crossflow filtration also requires that the membrane surface be covered or
“blocked” by a layer of proteins (beef extract, calf serum, bovine serum albumin) or a chemical
dispersant (NaPP) [52, 57, 58] to improve recovery. Protein-blocked membranes might be prone
to contamination during storage and transport [57]. In addition, values and reproducibilities of
recoveries of certain viruses such as HAdV remain low (Table 1.3) [24].

9

Table 1.3: Virus recovery by crossflow filtration
Virus

Water

1

Membrane

Blocking

Recovery (%)

Ref.

no blocking

4.5 ± 2.9

[58]

DIW + E-coli 15-20 kDa PSU

no blocking

29.5 ± 7.1

[58]

DIW + E-coli 15-20 kDa PSU

3 % BE

54.3 ± 23.8

[58]

MS2

TW

15-20 kDa PSU

0.1 % NaPP

53 ± 13

[61]

MS2

TW

15-20 kDa PSU

5% CS

108 ± 16

[57]

MS2

TW

15-20 kDa PSU

5% CS

120 ± 22

[56]

T1

surface water 50 kDa PAN

5% CS

73 ± 17

[60]

PP7

surface water 50 kDa PAN

5% CS

62 ± 5

[60]

MNoV

DIW

15-20 kDa PSU

no blocking

2.0 ± 1.5

[58]

MNoV

DIW + E-coli 15-20 kDa PSU

no blocking

5.3 ± 6.2

[58]

MNoV

DIW + E-coli 15-20 kDa PSU

3 % BE

63.3 ± 29.7

[58]

NoV

surface water ND

0.1% BSA

<0.95

[24]

HAdV

surface water ND

0.1 % BSA

3.18 – 6.02

[24]

E1

TW

15-20 kDa PSU

0.1% NaPP

49 ± 15

[61]

PV1

DIW

10 kDa

no blocking

16.9 - 17.8

[52]

PV1

Animal-viruses

15-20 kDa PSU

MS2

1

DIW

MS2

Phages

MS2

DIW

10 kDa

3% BE

49.6 – 76.8

[52]

type of water: DIW: deionized water, TW: tap water, SW: sewage

10

1.2

1.2.1

Polyelectrolyte multilayer films as anti-adhesive recoverable membrane coatings

Polyelectrolyte multilayer (PEM) films

PEM films consist of alternating layers of polycations and polyanions. They can form by
the simple alternating adsorption of oppositely charged polyelectrolytes on a substrate (layer by
layer assembly) through electrostatic interactions [67]. The formation of PEM films through Hbonding has also been shown [68]. By choosing different polyelectrolytes to build the film and
by varying the conditions of polyelectrolyte deposition (e.g., pH and ionic strength of the
polyelectrolyte solution), films with a wide range of surface charge, hydrophilicity, swellability,
and stiffness [69-74] have been prepared. PEM films have been deposited onto porous supports
to prepare nanofiltration membranes with unusually large water permeabilities [75-78] but also
as anti-adhesive coatings on filtration membranes [79, 80] or other surfaces [68, 72, 81-90], to
decrease the adsorption of one or several proteins [68, 72, 80-83, 91, 92], selected mammalian
cells [68, 72, 84-88], or selected bacteria [81, 84, 89, 90].

1.2.2

Anti-adhesive performance of PEMs

Meier-Haack and Muller [83, 91] achieved a 10-fold decrease in the amount of human
serum albumin (HSA) irreversibly adsorbed on a polypropylene microfiltration membrane during
filtration by coating the membrane with a PEM composed of poly(dimethyldiallylammonium)
(PDADMAC) and poly(acrylic acid) PAA; coating with the PEM increased HSA rejection but
did not impair the filtration flux. Mendelsohn et al. showed that murine fibroblast cell adhesion

11

on a substrate was completely eliminated by coating the surface with ten poly(allylamine
hydrochloride) (PAH)/PAA bilayers. Yang et al. achieved the same result with only one bilayer
of PAA/polyacrylamide (PAAm), and the film anti-adhesive properties were preserved for more
than a month [68, 72]. Lichter et al. have demonstrated that a PAH/PAA coating on a titanium
alloy also decreased bacterial adhesion (Staphylococcus epidermis and Escherichia coli) by
almost two orders of magnitude [89].

PEMs built using natural polypeptides and polysaccharides have been of particular interest
in the field of biomedicine. Richert et al. used hyaluronic acid (HA) and chitosan (CHI) to form
PEM films that decreased rat chondrocyte adhesion by more than 90% and bacterial adhesion (E.
coli) by approximately 80% [84]. Poly(L-lysine) (PLL)/alginate films and heparin(HE)/CHI
films also demonstrated anti-adhesive properties for fibroblast cells and E. coli, respectively [86,
90].

The mechanisms involved in the adhesion of proteins, mammalian cells and bacteria are
not fully understood; however, we do know that although proteins play a large role as promoters
in mammalian cell and bacteria adhesion, there is not always a correlation between protein
adhesion to a surface and mammalian cell or bacteria adhesion to the same surface [72]. Protein
adhesion onto PEM films is mostly regulated by electrostatic forces. In conditions where
proteins are negatively charged they tend to adsorb a lot more onto a PEM that contains a
polycation as the outermost layer and vice versa. In addition pH and ionic strength strongly
affect this adsorption behavior by altering the protein and polyelectrolyte charge densities and
screening those charges, respectively [72, 80, 81, 83, 91]. When both the protein and the top

12

layer have the same charge, protein adsorption is low unless the pH is close to the protein and
top polyelectrolyte layer’s pI, where electrostatic repulsion is low enough that hydrophobic
interaction can play a role [91]; as expected in that case, increasing the ionic strength had no
effect on the adsorption of proteins [82]. The effect of the charge density of the PEM top layer
on adhesion can be of two types. As mentioned earlier, increasing the charge density of the PEM
top layer can increase electrostatic repulsion and attraction. In addition, it can also affect the
conformation of the layer [74]. At higher charge density, polyelectrolyte chains adopt a flatter
conformation and form thinner layer while at lower charge density layers tend to be thicker and
loopier [74]. If a condensed high charge density layer is deposited on top a loopier one, it is
possible for some loops and segment from the layer before last to reach the PEM surface and
affect adsorption behavior [90]. This phenomenon might explain observations by Muller et al.
who used a strong polyanion with a high negative charge density, instead of weak polyanions, to
build the PEM’s outermost layer and saw an increase in the adsorption of the negatively charged
human serum albumin (HSA) [91]. Muller et al. speculated that in the case of PEMs containing
strong polyanions, strong attraction between the protein’s positively charged segment (although
HSA’s net charge is negative, some segments are positively charged) and the polyanion lead to a
reorientation of the protein to favor binding between the two. In the case of PEMs containing
weak polyanions, attractive interactions between the protein’s positive segments and the
polyanions are not strong enough to orient the protein to allow binding and thus the repulsion
between the protein’s negatively charged domains and the polyanions dominate the interaction
[91]. Salloum et al. also found that the use of more swellable PEM films decreased significantly
the amount of protein adsorbed by the surface compared to other PEMs, including those charged
similarly to the sorbing protein. In two other studies, however, no correlation between PEM

13

swelling and protein adsorption was found [72, 82].

The PEM film’s ability to swell was also correlated with its ability to resist adhesion of
mammalian cells and bacteria [68, 72, 84]. In fact, poor swellability and stiffness [87, 89],
which both depend on the amount of cross-linking between the polycations and the polyanions
[72, 87, 88, 92], are the only two properties of the PEM film that could be correlated with the
adhesion of bacteria or mammalian cells. Neither the nature (polyanion or polycation) of the
outermost layer, the surface charge density, the roughness, the wettability, the conformation of
the layer (flat and smooth or thick and loopy) nor the interaction energy (estimated using the
Lifshitz-van der Waals acid-base approach described by van Oss [93]) of PEM film was shown
to impact cell attachment [72, 89].

To explain the correlation between cell adhesion and swellability, Mendelsohn et al.
hypothesized that in solution the water rich structure of the film does not allow cell’s adhesion
proteins to adapt a configuration supporting cell adhesion. The environment “looks” more like
water than a surface to the cell, which is why cell adhesion does not occur [72]. Lichter et al.
showed that a 100-fold increase in elastic modulus resulted in approximately a 100-fold decrease
in bacterial adhesion and speculated that as the surface become stiffer, bacteria fimbriae
interactions with the surface are altered. More precisely, the fimbriae retraction after collision
with the surface is slower with a stiffer surface because there is more mechanical resistance to
the retraction; this additional time of contact between the bacteria and the surface would enhance
bacterial adhesion [89].

14

1.2.3

PEM films as sacrificial layers

When appropriately constructed, a PEM can be disassembled by way of exposure to high
[69, 94], neutral [95, 96] or low [97] pH, higher ionic strength [98-100], a specific solvent [101]
or lower [102] or higher [103] temperature. These properties of PEM have been exploited to
create micropatterned [95], porous [100] or even self-standing [98] PEM films. For example
Dubas was able to create a free standing PSS/PDADMAC multilayer by depositing 20 bilayer of
PSS/PDADMAC on top on 20 bilayer layers of PAA/PDADMAC and exposing the film to a salt
concentration higher than 0.6 M. In these conditions PAA/PDADMAC dissociates while
PSS/PDADMAC remains stable [98]. Sacrificial PEM layers have been considered for the
controlled delivery of drugs [101], the development of easy to clean surfaces [97] and the control
of membrane fouling during membrane filtration [69]. In that latter study, a PEM was deposited
on a UF membrane to create a nanofiltration membrane; once fouled the PEM layer was
sacrificed allowing the almost complete removal of both the fouling layer and PEM from the UF
membrane surface. The PEM layer could be regenerated to create a new nanofiltration
membrane [69].

1.3

1.3.1

Virus interaction with inert surfaces: Methods, challenges and current knowledge

Virus surface characterization

To understand virus interaction with surfaces, one must first gather information on the
virus’ own surface properties. The surface of a virion (virus particle) consists of a protein shell,
called capsid and, in the case of so called enveloped viruses, of a lipid bilayer enveloping the
15

protein shell. Sequencing techniques permit the determination of the composition of the protein
shell and theoretical surface properties (e.g. isoelectric point). Such theoretical values, however,
do not always accurately predict values measured in experiments [104]. Experimental
determination of virus surface properties is thus necessary. Such characterization requires that
virions be separated from protein and host cell debris present in the virus growth medium.
Purification methods include differential centrifugation and equilibrium gradient centrifugation.
In the differential centrifugation method, high and low speed centrifugation are used
alternatively to separate soluble components, large macromolecules and cell debris from virions
[105]. In the equilibrium gradient centrifugation method, the suspensions of viruses is
introduced into a tube where solutions of increasing density have been layered on top of each
over. During the centrifugation process virions migrate to the layer corresponding to their
density [106, 107]. Others methods have also been used; these include precipitation followed by
centrifugation to recover the pellet [108], centrifugal filtration with multiple washings [109],
multiple dialysis [110], or successive microfiltration and ultrafiltration using membranes of
various molecular weight cut-off (MWCO) values [111].

Once purified, virions can be characterized. The isoelectric point (IEP), a crucial
property that indicates if the particle will be positively or negatively charged at a given pH, is
determined by analyzing the behavior of the particles in an electric field under different pH
conditions. Two methods have been frequently employed: isoelectric focusing (IEF) and
determination of the electrophoretic mobility (EM) by light scattering (ELS) or a microscopybased method [112]. In IEF, particles are subjected to an electric field and migrate through a
gradient of increasing pH. When the particles reach the pH zone corresponding to its IEP the

16

particle stops. In EM analysis, the electrophoretic mobility, which is the ratio of the migration
velocity to the electric field, is measured at various pH values by light scattering or microscopic
methods. The EM is used to calculate the zeta-potential for each pH and the pH at which the
zeta-potential changes sign is the IEP.

Hydrophobicity is another surface characteristic essential for understanding virion
interactions with surfaces. When particles are immersed in an aqueous solution, the free energy
of cohesion of water drives particles to agglomerate to minimize their surface area exposed to
water, this phenomenon is called hydrophobic attraction. Sometimes, this attractive force is
attenuated or reversed when interaction between particles and water are preferred to interaction
between particles; this component of the total free energy is called hydrophilic repulsion [113].
If hydrophobic repulsion prevails, the material seems to repel water and this property is called
hydrophobicity. Van Oss and Giese [114] formally defined the hydrophobicity of a material i
based on the interfacial free energy of interaction between two objects of the same material i
when immersed in water: Giwi , which is expressed in terms of free energy per unit surface
area and can be determined using the following equation:
2


LW 
 


Giwi  2  iLW   w   4  i i   w w   i w   i w 





(1.1)



LW
 w ,  w and  w are the surface tension parameters of water (the “+”, “-“ and “LW”
superscripts correspond to the electron acceptor, the electron donor and the Lifhsitz-van der
Waals components of the surface tension, respectively), and  iLW ,  i and  i are the surface
tension parameters of the material i .
17

When Giwi is negative, there is a cohesive attraction between the two objects and the
material is considered hydrophobic. The magnitude of Giwi can be interpreted as the degree
of hydrophobicity [114]. If Giwi is positive the material is hydrophilic. This definition is not
yet adopted by the entire scientific community. Often a surface is considered hydrophilic if it is
“wetted” by water that is if the water contact angle is smaller than 90º, which does not
necessarily corresponds to a positive Giwi . For example dry biological surfaces typically have
2
a negligible  i and a  iLW of approximately 40 mJ/m ; for such surfaces the transition from

hydrophilic to hydrophobic corresponds to a water contact angle of approximately 60º (equations
1.1 and 1.2).

The surface tension parameters of a given surface can be determined using the extended
Young equation (1.2) after measuring the contact angle of three probe liquids with known
surface tension parameters  .

TOT
1  cos  

 LW LW
 
 
 2  s  
 s   s 



(1.2)






In these equations,  is the contact angle for a given probe liquid and the subscripts

and 

correspond to the solid surface and the liquid probe, respectively. In the case of viruses [115117] and 
nanoparticles in general [118, 119], contact angle measurements can be done by

depositing layers of particles on a clean surface and determining the contact angles using the
sessile drop method. The roughness of the deposited layer, which can be independently

18

determined by atomic force microscopy (AFM), and remnant contamination will affect the
measured contact angle following equations 1.3 and 1.4, respectively:
cos apparent  r costheory

(1.3)

cos apparent  F1 cos1  F2 cos 2

(1.4)

where  is the contact angle, r the roughness and Fi the fractional amount of each material
present in the layer. It should be noted that in two out of the three studies where the contact
angle of a virus was measured [115, 116], the virus suspension was not purified before formation
of the virus layer. The measured contact angles were probably significantly affected by the
presence of bacterial debris and broth components within the virus deposit. Alternatively, virus
hydrophobicity has also been estimated qualitatively by hydrophobic interaction chromatography
[120].

1.3.2

Quantifying interactions between viruses and a surface

It is possible to directly measure the interfacial energy of interaction between a surface
and a particle using AFM [121-130]. In AFM, a surface is scanned by a tip, which is a part of a
flexible cantilever. The position of the cantilever depends on both surface morphology and
forces between the cantilever tip and the surface. In force spectroscopy, the vertical movements
of the cantilever, proportional to interaction forces, are recorded while the surface is approached
and moved away from the tip [131]. A single microparticle composed of silica [121, 123, 125],
alumina [121], polystyrene [121, 122], zirconia [124], lactose [127] or a bacterial spore [126],
can be glued to a tipless cantilever to study the particle’s interaction with various surfaces.
19

Cantilevers with tips have also been modified with nanoparticles: for example, Vakarelski et al.
[130] fabricated a cantilever tip terminated by a single gold nanoparticle; Das et al [129] glued a
nanoparticle agglomerate on a cantilever tip and Attinti et al. [117] opted to coat the cantilever
tip with nanoparticles. Others chose to deposit microparticles [128] or nanoparticles [132, 133]
on the surface scanned by the cantilever, and to functionalize the cantilever tip with molecules of
interest such as humic acid [128].

In the field of virology, AFM has mostly been used to image viruses and investigate virus
interactions with a host cell or other surfaces through imaging. For example, AFM was used to
visualize the pattern formed by phages assembled on a surface and to quantify the number of
viruses deposited and/or the presence of aggregates [134-138]). AFM was also used to quantify
the binding force between macromolecules present at the surface of a virus capsid and other
surfaces by functionalizing the AFM tip with the macromolecules of interest [139]. A few
studies used AFM to quantify interaction forces between viruses and surfaces. Attinti et al. [117]
studied virus-sand interaction using a virus-coated tip and Liu et al. [140] measured the force of
interaction between a tip functionalized with anti-hemagglutinin monoclonal antibody and a
single influenza virus deposited on a glass slide.

Virus interactions with surfaces such as clay or NOM-coated substrates have also been
investigated using Quartz Crystal Microbalance (QCM) gravimetry [141-146]. By measuring
very small changes in the mass of material deposited on the crystal surface, QCM enables
rigorous quantification of the mass of viruses deposited over time and can be used to study the
deposition kinetics. Other methods for studying virus-surface interactions include total internal

20

reflection fluorescence microscopy (TIRF), which has been used to study virus adsorption
kinetics [147], other fluorescence microscopies [115], transmission electron microscopy (TEM)
[148], and methods that employ radioactively labeled viruses [149]. Finally, in many studies the
interaction of a virus with surfaces was evaluated by measuring the difference in the
concentration of the virus in suspension (measured by traditional methods such as plaque assay)
before and after the suspension is exposed to a surface [116, 120, 150-161]. It should be noted,
however, that loss of infectivity or particle aggregation during virus interaction with the surface
can affect the results of this test when an infectivity assay is selected as the detection method.

Several authors employed the DLVO (Derjaguin, Landau, Verwey and Overbeek) [115,
116, 117, 141, 142, 144, 156, 157, 162-165] and XDLVO (extended DLVO) [115, 116, 117, 156,
157] theories to describe virus attachment to surfaces. The DLVO theory predicts surfacesurface interaction by taking into consideration Liftshitz-van der Waals (U
double layer interactions (U
considered (U

AB

EL

LW

) and electrostatic

). In the XDLVO model, Lewis acid-base forces are also

). Only one study [115] determined the virus surface tension parameters  LW ,

  and   , which allow for the rigorous determination of the Lifshitz-van der Waals and
Lewis acid-base interactions; however these viruses were not purified. All but the study by
Chattopadhyay et al. [116], computed van der Waals interaction based on the Hamaker method
[166]:

A rp
132
s
U vanderwaal  
6h

 14h 
1   



1

(1.5)

21

where r p is the virion radius, h the distance between the virion and the surface,  the
characteristic wavelength of the interaction and A132the Hamaker constant for the particlewater-surface system. Most of these studies (7 out of 10) [117, 141, 142, 144, 157, 162, 165]
used Hamaker constants based on Murray and Park’s [167] estimation of the poliovirus–waterclay system which itself was based on the Hamaker constant of the polystyrene–water-clay
system and the assumption that polystyrene and poliovirus shared similar dielectric properties.
By following this approach, these studies did not take into account the potential specificity of
each virus surface for the determination of the van der Waals interactions; however, it is true that
2

most biological surfaces have a  LW of about 40 mJ/m ; as a consequence van der Waals
interactions between viruses and surface might not differ greatly from one virus species to
another [111]. Three studies [156, 163, 164] used the Lifshitz theory [168], which links the
substance relative permittivity and refractive index to the Hamaker constant, to determine the
virus Hamaker constant.

Three studies computed the Lewis acid-base interactions [117, 156, 157] to use the
XDLVO model. They did so using Yoon’s empirical approach [169]. which is described in the
following equation:

K T
 cos1  cos 2 
U AB   123 with log K123  a
b
6h
2



(1.6)

where T is the temperature, h the separation distance, 1 and  2 the water contact angles of the
virus and the surface and a and b are system specific constants. Yoon’s approach only requires
the measurement of the water contact angle rather than the 3 liquid probe contact angles
necessary in van Oss’ theoretical approach [113]. Bergendahl and Grasso found that Yoon’s
22

empirical approach and van Oss’s theoretical approach (described in chapter 2) gave similar
trends in terms of which interactions were the most attractive; however, calculated values for
Lewis acid base interaction energies were quite different; and for one system Lewis acid base
interactions were found to be attractive using Yoon’s approach but repulsive using the van Oss
rigorous theoretical approach. In Yoon’s approach, variations in water contact angles are used to
calculate variations in hydrophobic interactions whereas variations in contact angles are also due
to changes in Lifshitz van der Waals interactions. In van Oss’ approach, measuring the contact
angle of a completely apolar liquid probe (diiodomethane for example) allows the determination
of the Lifshitz van der Waals interaction term independently from the Lewis acid-base term.
Additionally, measuring contact angles of two polar liquid probes with different electron donor
and electron acceptor characteristics allows the determination of these two parameters for the
surface studied [113].

Other models have been used to estimate hydrophobic interaction. Chattopadhyay et al.
[116], who defined hydrophobic interaction the same way as van Oss and Giuse [112] ( i.e.
hydrophobic interaction is the sum of van der Waals and Lewis acid-base interactions), used the
empirical method developed by Neumann et al. [170] to estimate the hydrophobic interaction
between bacteriophages and clay. Neumann’s method predicts interfacial tensions between
surfaces using only water contact angles [170]; however, several authors have demonstrated that
Neumann’s empirical approach is flawed [113, 171]. The hydrophobic interaction term used by
Lee et al. did not include the van der Waals interaction and was estimated using an empirical
model usually applied to determine protein hydrophobic energy [164].

23

1.3.3

Current knowledge

Sorption of viruses to various surfaces is the subject of numerous studies. Already in
1975 Bitton reviewed more than 50 studies looking at virus adsorption to materials such as
activated carbon, nitrocellulose, glass, clays, sands, polyelectrolytes, iron oxides, ion exchange
resins and membrane filters [172]. Viruses sorb to a wide range of surfaces including positively
charged surfaces such as alumina and cationic polyelectrolytes and negatively charged surfaces
such as nitrocellulose, glass, or anionic polymers. Adsorption, especially on negatively charged
surfaces, depended on the ionic strength [161, 172]. Indeed high salt concentrations decreased
the thickness of the double-layer, allowing negatively charged virus to get close enough to the
surface for attractive forces such as van der Waals forces to act. Localized interactions between
the positively charged fraction of the virus protein coat and a negatively charged surface could
also explain the attachment of viruses to such surfaces [172]. The presence of proteinaceous
material such as bovine albumin or bovine serum or natural organic matter hindered virus
adsorption on activated carbon, clay, sands, anion exchange resin and membrane filters [160,
172-174]. Some authors suggested these proteinaceous materials competed with virus for the
surface adsorption sites [172]; other authors have mentioned the possibility of the thick layer of
organic matter on the surfaces increasing steric repulsion between viruses and the surface [174].

It was observed in those earlier studies that not all viruses exhibited similar adsorption
behavior; however, only few reports considered viruses’ surface properties and morphologies
and how they correlated with the adsorption behavior [175, 176]. It was later found that viruses
present a wide range of surface properties and thus a range of sorption behaviors [112, 154].

24

Michen et al. [112] reviewed 48 studies where the isoelectric point of 104 viruses was measured.
The reported IEPs ranged from 1.9 to 8.4 but were mostly below 7, and depended not only on the
species but also on the strain. Several viruses (such as poliovirus, mengovirus and
coxsackievirus) had 2 IEPs, probably due to a change of conformation of the viral capsid at a
certain pH. Viruses also vary widely in terms of their hydrophobicity [116, 117, 120]. An
example of a very hydrophilic virus is phiX174 with a water contact angle between 26 ± 2 º
[117] and 42 ± 6 º [116]. An example of a hydrophobic virus is T2 with a water contact angle of
96 ± 9 º [116].

The IEP of viruses was a determining factor in virus adsorption indicating the importance
of electrostatic interactions for virus attachment to surface [142, 144, 150, 151, 159, 173].
Hydrophobic interactions have also been shown to be important in the attachment of MS2 [116,
151, 153, 156, 157, 173], a hydrophilic virus. In some cases, interaction with hydrophobic
surfaces disrupted the virus protein coat and lead to virus disassembly [147]. As ionic strenght
increases and repulsive electrostatic interactions decrease, the hydrophobic interaction becomes
increasingly important [151].

Other factors that seem to influence virus adsorption include the virus size (higher
surface areas present more sorption sites) [114,157], virus morphology (the presence of spikes,
fibers, tails, extruding protein loops) [116, 172, 177], and the specific interaction with divalent
cations especially Ca

2+

which can form bridges between functional groups on the virus and on

the surface [141, 143, 144].

25

In several studies DLVO theory failed to predict virus attachment behavior [144, 165],
probably because of the importance of hydrophobic interaction in virus adsorption to surfaces.
Also representing viruses as hard particles in models predicting double layer interactions is not
always accurate [178]. Although the XDLVO model predicted the virus attachment
experimentally observed in some studies well [116, 157, 164], others found discrepancies
between the predicted and experimentally observed virus behavior [117, 156].

1.4

Dissertation overview

This work aims at improving virus recovery and concentration by crossflow filtration. In
chapter 2 we designed highly controllable anti-adhesive membrane coatings to replace the
traditional protein based blocking layer and assessed their performance in terms of virus
recovery and concentration rate, under different filtration conditions that include filtration rates
more than double the highest rate reported in the literature. In chapter 3, the effect of high
crossflow and high permeate flow rate, which are necessary to hasten the filtration but might
affect virus infectivity, on recovery and removal of infective virus from DI and surface water are
evaluated using the CS-blocked membrane. The effect of those filtration conditions on the
stability of the blocking layer is also investigated. In chapter 4, we looked at the performance of
the anti-adhesive PEM membrane coatings with complex water matrices.

26

REFERENCES

27

REFERENCES

1.

Leclerc, H., L. Schwartzbrod, and E. Dei-Cas, Microbial agents associated with
waterborne diseases. Critical Reviews In Microbiology, 2002. 28(4): p. 371-409.

2.

Theron, J. and T.E. Cloete, Emerging waterborne infections: Contributing factors,
agents, and detection tools. Critical Reviews In Microbiology, 2002. 28(1): p. 1-26.

3.

Lee, S.H., et al., Surveillance for waterborne disease outbreaks-United States, 19992000, in CDC Surveillance Summaries2002.

4.

Blackburn, B.G., et al., Surveillance for waterborne-disease outbreaks associated with
drinking water - United States, 2001-2002. Morbidity and Mortality Weekly Report,
2004. 53(SS-8): p. 23-44.

5.

Liang, J.L., et al., Surveillance for waterborne disease and outbreaks associated with
drinking water and water not intended for drinking-United States, 2003-2004, in
Morbidity and Mortality Weekly Report Surveillance Summaries 2006.

6.

Yoder , J., et al., Surveillance for waterborne disease and outbreaks associated with
drinking water and water not intended for drinking — United States, 2005–2006.
Morbidity and Mortality Weekly Report Surveillance Summaries, 2008. 57(SS-09): p.
39-62.

7.

Brunkard, J.M., et al., Surveillance for waterborne disease outbreaks associated with
drinking water - United States, 2007-2008. Morbidity and Mortality Weekly Report
Surveillance Summaries, 2011. 60(SS-12): p. 1-78.

8.

Yoder, J.S., et al., Surveillance for waterborne disease outbreaks and other health events
associated with recreational water - United States, 2001-2002. Morbidity and Mortality
Weekly Report Surveillance Summaries, 2004. 53(SS-08): p. 1-22.

9.

Dziuban, E.J., et al., Surveillance for waterborne-disease outbreaks associated with
recreational water - United States, 2003-2004. Morbidity and Mortality Weekly Report,
2006. 55(SS-12): p. 1-24.

10.

Yoder, J.S., et al., Surveillance for waterborne disease and outbreaks associated with
recreational water use and other aquatic facility-associated health events — United
States, 2005–2006 and in Morbidity and Mortality Weekly Report Surveillance
Summaries2008.

11.

Hlavsa, M.C., et al., Surveillance for waterborne disease outbreaks and other health
events associated with recreational water - United States, 2007-2008. Morbidity and
Mortality Weekly Report Surveillance Summaries, 2011. 60(SS-12): p. 1-78.
28

12.

Sinclair, R.G., E.L. Jones, and C.P. Gerba, Viruses in recreational water-borne disease
outbreaks: a review. Journal of Applied Microbiology, 2009. 107(6): p. 1769-1780.

13.

Hrudey, S.E. and E.J. Hrudey, Safe drinking water: Lessons from recent outbreaks in
affluent nations2004, London: IWA Publishing.

14.

Waterborne outbreak of gastroenteritis associated with a contaminated municipal water
supply, Walkerton, Ontario, May-June 2000. Can Commun Dis Rep, 2000. 26(20): p.
170-3.

15.

Kramer, M.H., et al., Waterborne disease: 1993 and 1994. Journal American Water
Works Association, 1996. 88(3): p. 66-80.

16.

Fox, K.R. and D.A. Lytle, Milwaukee's crypto outbreak: Investigation and
recommendations. Journal American Water Works Association, 1996. 88(9): p. 87-94.

17.

Wade, T.J., et al., Rapidly measured indicators of recreational water quality are
predictive of swimming-associated gastrointestional illness. Environmental Health
Perspectives, 2006. 114(1): p. 24-28.

18.

Nellor, M.H., R.B. Baird, and J.R. Smyth, Health effects of indirect potable water reuse
American Water Works Association Journal, 1985. 77(7): p. 88-96.

19.

NRC, Issues in potable reuse: The viability of augmenting drinking water supplies with
reclaimed water 1998: Washington, DC.

20.

Abbaszadegan, M., M. LeChevallier, and C. Gerba, Occurrence of viruses in US
groundwaters. J. Am. Water Works Assn., 2003. 95(9): p. 107-120.

21.

Gerba, C.P., et al., Waterborne rotavirus: A risk assessment. Water Res., 1996. 30(12): p.
2929-2940.

22.

Straub, T.M. and D.P. Chandler, Towards a unified system for detecting waterborne
pathogens. Journal of Microbiological Methods, 2003. 53(2): p. 185-197.

23.

Schaudies, P. and D.A. Robinson, Literature review of molecular methods for
simulateneous detection of pathogens in water, USEPA, Editor 2007: Washington D.C.

24.

Albinana-Gimenez, N., et al., Comparison of methods for concentrating human
adenoviruses, polyomavirus JC and noroviruses in source waters and drinking water
using quantitative PCR. Journal of Virological Methods, 2009. 158(1-2): p. 104-109.

25.

Borrego, J.J., et al., Development and application of new positively charged filters for
recovery of bacteriophages from water. Applied And Environmental Microbiology, 1991.
57(4): p. 1218-1222.
29

26.

Haramoto, E., et al., Recovery of human norovirus from water by virus concentration
methods. Journal of Virological Methods, 2009. 160(1-2): p. 206-209.

27.

Karim, M.R., et al., New electropositive filter for concentrating enteroviruses and
noroviruses from large volumes of water. Applied and Environmental Microbiology,
2009. 75(8): p. 2393-2399.

28.

Ma, J.F., J. Naranjo, and C.P. Gerba, Evaluation of Mk filters for recovery of
enteroviruses from tap water. Applied and Environmental Microbiology, 1994. 60(6): p.
1974-1977.

29.

Roepke, D.C., et al., An adsorption-elution technique for the recovery of influenza-virus
from water. Avian Diseases, 1989. 33(4): p. 649-653.

30.

Rose, J.B., et al., Comparison of microporous filters for concentration of viruses from
wastewater. Applied and Environmental Microbiology, 1984. 47(5): p. 989-992.

31.

Sobsey, M.D. and B.L. Jones, Concentration of poliovirus from tap water using positively
charged microporous filters. Applied and Environmental Microbiology, 1980. 37(3): p.
588-595.

32.

USEPA, Manual of methods for virology, USEPA, Editor 2001: Washington D.C.

33.

Brassard, J., et al., Concentration and detection of hepatitis A virus and rotavirus in
spring water samples by reverse transcription-PCR. Journal of Virological Methods,
2005. 123(2): p. 163-169.

34.

Goyal, S.M. and C.P. Gerba, Viradel method for detection of rota virus from seawater.
Journal of Virological Methods, 1983. 7(5-6): p. 279-285.

35.

Katayama, H., A. Shimasaki, and S. Ohgaki, Development of a virus concentration
method and its application to detection of enterovirus and Norwalk virus from coastal
seawater. Applied And Environmental Microbiology, 2002. 68(3): p. 1033-1039.

36.

Scott, T.M., J. Lukasik, and S.R. Farrah, Improved method for recovery of bacteriophage
from large volumes of water using negatively charged microporous filters. Canadian
Journal Of Microbiology, 2002. 48(4): p. 305-310.

37.

Abdelzaher, A.M., et al., Sequential concentration of bacteria and viruses from marine
waters using a dual membrane system. J. Environ. Qual., 2008. 37(4): p. 1648-1655.

38.

APHA, 9222 B. Standard Total Coliform Membrane Filter Procedure, in Standard
Methods for the Examination of Water and Wastewater1999, American Public Health
Association, American Water Works Association, Water Environment Federation.

30

39.

USEPA, Method 1604: Total Coliforms and Escherichia coli in Water by Membrane
Filtration Using a Simultaneous Detection Technique (MI Medium), 2002.

40.

USEPA, Method 1623: Cryptosporidium and Giardia in Water by Filtration/IMS/FA,
2005.

41.

Farrah, S.R., et al., Use of modified diatomaceous-earth for removal and recovery of
viruses in water. Applied And Environmental Microbiology, 1991. 57(9): p. 2502-2506.

42.

Lambertini, E., et al., Concentration of enteroviruses, adenoviruses, and noroviruses
from drinking water by use of glass wool filters. Appl. Environ. Microbiol., 2008. 74(10):
p. 2990-2996.

43.

Preston, D.R., et al., Novel-Approach For Modifying Microporous Filters For Virus
Concentration From Water. Applied And Environmental Microbiology, 1988. 54(6): p.
1325-1329.

44.

Di Pasquale, S., et al., Comparison of different concentration methods for the detection of
hepatitis A virus and calicivirus from bottled natural mineral waters. Journal Of
Virological Methods, 2010. 165(1): p. 57-63.

45.

Fumian, T.M., et al., Detection of rotavirus A in sewage samples using multiplex qPCR
and an evaluation of the ultracentrifugation and adsorption-elution methods for virus
concentration. Journal Of Virological Methods, 2010. 170(1-2): p. 42-46.

46.

Tsai, U.L., et al., Simple method of concentrating enteroviruses and hepatitis A virus
from sewage and ocean water for rapid detection by reverse transcriptase-polymerase
chain reaction. Applied And Environmental Microbiology, 1993. 59(10): p. 3488-3491.

47.

Liu, J.Y., Q.P. Wu, and X.X. Kou, Development of a virus concentration method and its
application for the detection of noroviruses in drinking water in China. Journal Of
Microbiology, 2007. 45(1): p. 48-52.

48.

Hill, V.R., et al., Comparison of hollow-fiber ultrafiltration to the USEPA VIRADEL
technique and USEPA Method 1623. J. Environ. Qual., 2009. 38(2): p. 822-825.

49.

Adikane, H.V., et al., Concentration of foot-and-mouth disease virus by ultrafiltration.
Journal of Membrane Science, 1997. 132(1): p. 91-96.

50.

Belfort, G., Y. Rotem, and E. Katzenelson, Virus concentration using hollow fiber
membranes. Water Res., 1975. 9(1): p. 79-85.

51.

Belfort, G., Y. Rotem, and E. Katzenelson, Virus concentration using hollow fiber
membranes .2. Water Res., 1976. 10(4): p. 279-284.

31

52.

Berman, D., M.E. Rohr, and R.S. Safferman, Concentration of poliovirus in water by
molecular filtration. Appl. Environ. Microbiol., 1980. 40(2): p. 426-428.

53.

Chamorro, C., et al., Concentration of infectious pancreatic necrosis virus determined by
combining tangential flow filtration with exclusion filtration. Archivos De Medicina
Veterinaria, 2006. 38(1): p. 77-82.

54.

Dziewulski, D.M. and G. Belfort, Virus concentration from water using high-rate
tangential-flow hollow fiber ultrafiltration. Water Science And Technology, 1983. 15(5):
p. 75-89.

55.

Gibson, K.E. and K.J. Schwab, Detection of bacterial indicators and human and bovine
enteric viruses in surface water and groundwater sources potentially impacted by animal
and human wastes in Lower Yakima Valley, Washington. Applied And Environmental
Microbiology, 2011. 77(1): p. 355-362.

56.

Hill, V.R., et al., Multistate evaluation of an ultrafiltration-based procedure for
simultaneous recovery of enteric microbes in 100-liter tap water samples. Applied and
Environmental Microbiology, 2007. 73(19): p. 6327-6327.

57.

Hill, V.R., et al., Development of a rapid method for simultaneous recovery of diverse
microbes in drinking water by ultrafiltration with sodium polyphosphate and surfactants.
Applied and Environmental Microbiology, 2005. 71(11): p. 6878-6884.

58.

Kim, H., H. Park, and G. Ko, Hollow-fiber ultrafiltration for the concentration and
simultaneous recovery of multiple pathogens in contaminated foods. Journal of Food
Protection, 2009. 72(12): p. 2547-2552.

59.

Knappett, P.S.K., et al., Efficacy of Hollow-Fiber Ultrafiltration for Microbial Sampling
in Groundwater. Ground Water, 2011. 49(1): p. 53-65.

60.

Morales-Morales, H.A., et al., Optimization of a reusable hollow-fiber ultrafilter for
simultaneous concentration of enteric bacteria, protozoa, and viruses from water.
Applied and Environmental Microbiology, 2003. 69(7): p. 4098-4102.

61.

Polaczyk, A.L., et al., Ultrafiltration-based techniques for rapid and simultaneous
concentration of multiple microbe classes from 100-L tap water samples. Journal Of
Microbiological Methods, 2008. 73(2): p. 92-99.

62.

Haramoto, E., et al., Development of a novel method for simultaneous concentration of
viruses and protozoa from a single water sample. Journal of Virological Methods, 2012.
182(1-2): p. 62-69.

63.

Ikner, L.A., M. Soto-Beltran, and K.R. Bright, New method using a positively charged
microporous filter and ultrafiltration for concentration of viruses from tap water.
Applied and Environmental Microbiology, 2011. 77(10): p. 3500-3506.
32

64.

Prata, C., et al., Ultracentrifugation as a direct method to concentrate viruses in
environmental waters: virus-like particle enumeration as a new approach to determine
the efficiency of recovery. Journal of Environmental Monitoring, 2012. 14(1): p. 64-70.

65.

Olszewski, J., L. Winona, and K.H. Oshima, Comparison of 2 ultrafiltration systems for
the concentration of seeded viruses from environmnetal waters. Can. J. Microbiol., 2005.
51(4): p. 295-303.

66.

Rajal, V.B., et al., Validation of hollow fiber ultrafiltration and real-time PCR using
bacteriophage PP7 as surrogate for the quantification of viruses from water samples.
Water Res., 2007. 41(7): p. 1411-1422.

67.

Decher, G., Fuzzy nanoassemblies: Toward layered polymeric multicomposites. Science,
1997. 277(5330): p. 1232-1237.

68.

Yang, S.Y., J.D. Mendelsohn, and M.F. Rubner, New class of ultrathin, highly celladhesion-resistant polyelectrolyte multilayers with micropatterning capabilities.
Biomacromolecules, 2003. 4(4): p. 987-994.

69.

Shan, W.Q., et al., Polyelectrolyte multilayer films as backflushable nanofiltration
membranes with tunable hydrophilicity and surface charge. Journal of Membrane
Science, 2010. 349(1-2): p. 268-278.

70.

Ge, L.Q., et al., Fabrication of multilayered hollow nanofibers and estimation of its
Young's modulus. Japanese Journal of Applied Physics Part 1-Regular Papers Brief
Communications & Review Papers, 2007. 46(10A): p. 6790-6795.

71.

Miller, M.D. and M.L. Bruening, Correlation of the swelling and permeability of
polyelectrolyte multilayer films. Chemistry of Materials, 2005. 17(21): p. 5375-5381.

72.

Mendelsohn, J.D., et al., Rational design of cytophilic and cytophobic polyelectrolyte
multilayer thin films. Biomacromolecules, 2003. 4(1): p. 96-106.

73.

Burke, S.E. and C.J. Barrett, Swelling behavior of hyaluronic acid/polyallylamine
hydrochloride multilayer films. Biomacromolecules, 2005. 6(3): p. 1419-1428.

74.

Yoo, D., S.S. Shiratori, and M.F. Rubner, Controlling bilayer composition and surface
wettability of sequentially adsorbed multilayers of weak polyelectrolytes.
Macromolecules, 1998. 31(13): p. 4309-4318.

75.

Jin, W., A. Toutianoush, and B. Tieke, Use of polyelectrolyte layer-by-layer assemblies
as nanofiltration and reverse osmosis membranes. Langmuir, 2003. 19: p. 2550-2553.

76.

Miller, M.D. and M.L. Bruening, Controlling the nanofiltration properties of multilayer
polyelectrolyte membranes through variation of film composition. Langmuir, 2004.
20(26): p. 11545-51.
33

77.

Ouyang, L.Y., R. Malaisamy, and M.L. Bruening, Multilayer polyelectrolyte films as
nanofiltration membranes for separating monovalent and divalent cations. Journal of
Membrane Science, 2008. 310(1-2): p. 76-84.

78.

Stanton, B.W., et al., Ultrathin, multilayered polyelectrolyte films as nanofiltration
membranes. Langmuir, 2003. 19(17): p. 7038-7042.

79.

Lajimi, R.H., et al., Effect of LbL surface modification on characteristics and
performances of cellulose acetate nanofiltration membranes. Desalination, 2011. 266(13): p. 78-86.

80.

Muller, M., et al., Polyelectrolyte complex layers: a promising concept for antifouling
coatings verified by in-situ ATR-FTIR spectroscopy. Macromolecular Rapid
Communications, 1999. 20(12): p. 607-611.

81.

Boulmedais, F., et al., Polyelectrolyte multilayer films with pegylated polypeptides as a
new type of anti-microbial protection for biomaterials. Biomaterials, 2004. 25(11): p.
2003-2011.

82.

Salloum, D.S. and J.B. Schlenoff, Protein adsorption modalities on polyelectrolyte
multilayers. Biomacromolecules, 2004. 5(3): p. 1089-1096.

83.

Meier-Haack, J. and M. Muller, Use of polyelectrolyte multilayer systems for membrane
modification. Macromolecular Symposia, 2002. 188: p. 91-103.

84.

Richert, L., et al., Layer by layer buildup of polysaccharide films: Physical chemistry and
cellular adhesion aspects. Langmuir, 2004. 20(2): p. 448-458.

85.

Lu, Y.X., J.Q. Sun, and J.C. Shen, Cell adhesion properties of patterned poly(acrylic
acid)/poly(allylamine hydrochloride) multilayer films created by room-temperature
imprinting technique. Langmuir, 2008. 24(15): p. 8050-8055.

86.

Elbert, D.L., C.B. Herbert, and J.A. Hubbell, Thin polymer layers formed by
polyelectrolyte multilayer techniques on biological surfaces. Langmuir, 1999. 15(16): p.
5355-5362.

87.

Schneider, A., et al., Polyelectrolyte multilayers with a tunable Young's modulus:
Influence of film stiffness on cell adhesion. Langmuir, 2006. 22(3): p. 1193-1200.

88.

Thompson, M.T., et al., Tuning compliance of nanoscale polyelectrolyte multilayers to
modulate cell adhesion. Biomaterials, 2005. 26(34): p. 6836-6845.

89.

Lichter, J.A., et al., Substrata mechanical stiffness can regulate adhesion of viable
bacteria. Biomacromolecules, 2008. 9(6): p. 1571-1578.

34

90.

Fu, J.H., et al., Construction of anti-adhesive and antibacterial multilayer films via layerby-layer assembly of heparin and chitosan. Biomaterials, 2005. 26(33): p. 6684-6692.

91.

Muller, M., et al., Selective interaction between proteins and the outermost surface of
polyelectrolyte multilayers: Influence of the polyanion type, pH and salt. Macromolecular
Rapid Communications, 2001. 22(6): p. 390-395.

92.

Muller, M., et al., Polyelectrolyte multilayers and their interactions. Journal of Adhesion,
2004. 80(6): p. 521- 547.

93.

van Oss, C.J., Long-range and short-range mechanisms of hydrophobic attraction and
hydrophilic repulsion in specific and aspecific interactions. Journal of Molecular
Recognition, 2003. 16(4): p. 177-190.

94.

Sukhishvili, S.A. and S. Granick, Layered, erasable, ultrathin polymer films. JACS,
2000. 122(39): p. 9550-9551.

95.

Yang, S.Y. and M.F. Rubner, Micropatterning of Polymer Thin Films with pH-Sensitive
and Cross-linkable Hydrogen-Bonded Polyelectrolyte Multilayers. Journal of the
American Chemical Society, 2002. 124(10): p. 2100-2101.

96.

Ono, S.S. and G. Decher, Preparation of Ultrathin Self-Standing Polyelectrolyte
Multilayer Membranes at Physiological Conditions Using pH-Responsive Film Segments
as Sacrificial Layers. Nano Letters, 2006. 6(4): p. 592-598.

97.

de Vos, W.M., et al., Thin polymer films as sacrificial layers for easier cleaning. Colloids
and Surfaces a-Physicochemical and Engineering Aspects, 2010. 358(1-3): p. 6-12.

98.

Dubas, S.T., T.R. Farhat, and J.B. Schlenoff, Multiple membranes from "true"
polyelectrolyte multilayers. Journal of the American Chemical Society, 2001. 123(22): p.
5368-5369.

99.

Dubas, S.T. and J.B. Schlenoff, Polyelectrolyte multilayers containing a weak polyacid:
Construction and deconstruction. Macromolecules, 2001. 34(11): p. 3736-3740.

100.

Gui, Z., et al., A novel fast response fiber-optic pH sensor based on nanoporous selfassembled multilayer films. Journal of Materials Chemistry, 2010. 20(36): p. 7754-7760.

101.

Smith, R.C., et al., Layer-by-layer platform technology for small-molecule delivery.
Angewandte Chemie International Edition, 2009. 48(47): p. 8974-8977.

102.

Zhuk, A., S. Pavlukhina, and S.A. Sukhishvili, Hydrogen-Bonded Layer-by-Layer
Temperature-Triggered Release Films†. Langmuir, 2009. 25(24): p. 14025-14029.

35

103.

Gui, Z.L., et al., Controllable disintegration of temperature-responsive self-assembled
multilayer film based on polybetaine. Colloids and Surfaces a-Physicochemical and
Engineering Aspects, 2011. 380(1-3): p. 270-279.

104.

Goodridge, L., et al., Isoelectric point determination of norovirus virus-like particles by
capillary isoelectric focusing with whole column imaging detection. Analytical
Chemistry, 2004. 76(1): p. 48-52.

105.

Handa, H., et al., Recognition of Salmonella typhimurium by immobilized phage P22
monolayers. Surface Science, 2008. 602(7): p. 1392-1400.

106.

Ricciuti, C.P. and A.M. Haywood, Production of defective MS2 virions in resting
Escherichia-Coli. Virology, 1974. 58(1): p. 164-175.

107.

McLaughlin, M.R. and R.A. King, Characterization of Salmonella bacteriophages
isolated from swine lagoon effluent. Current Microbiology, 2008. 56(3): p. 208-213.

108.

Rusckowski, M., et al., Investigation of four 99mTc-labeled bacteriophages for infectionspecific imaging. Nuclear Medicine and Biology, 2008. 35(4): p. 433-440.

109.

Bae, J. and K.J. Schwab, Evaluation of murine norovirus, feline calicivirus, poliovirus,
and MS2 as surrogates for human norovirus in a model of viral persistence in surface
water and groundwater. Applied and Environmental Microbiology, 2008. 74(2): p. 477484.

110.

Langlet, J., et al., Aggregation and surface properties of F-specific RNA phages:
Implication for membrane filtration processes. Water Research, 2008. 42(10-11): p.
2769-2777.

111.

Mylon, S.E., et al., Influence of salts and natural organic matter on the stability of
bacteriophage MS2. Langmuir, 2010. 26(2): p. 1035-1042.

112.

Michen, B. and T. Graule, Isoelectric points of viruses. Journal of Applied Microbiology,
2010. 109(2): p. 388-397.

113.

van Oss, C.J., Interfacial forces in aqeuous media - second edition2006: Taylor &
Francis Group.

114.

van Oss, C.J. and R.F. Giese, The hydrophilicity and hydrophobicity of clay-minerals.
Clays and Clay Minerals, 1995. 43(4): p. 474-477.

115.

Kukavica-Ibrulj, I., et al., Hepatitis A virus attachment to agri-food surfaces using
immunological, virological and thermodynamic assays. Journal Of Applied
Microbiology, 2004. 97(5): p. 923-934.

36

116.

Chattopadhyay, S. and R.W. Puls, Adsorption of bacteriophages on clay minerals.
Environ. Sci. Technol. , 1999. 33: p. 3609.

117.

Attinti, R., et al., Virus (MS2, phiX174, and aichi) attachment on sand measured by
atomic force microscopy and their transport through sand columns. Environmental
Science & Technology, 2010. 44(7): p. 2426-2432.

118.

Kim, J., et al., Interactions of Aqueous NOM with Nanoscale TiO2: Implications for
Ceramic Membrane Filtration-Ozonation Hybrid Process. Environmental Science &
Technology, 2009. 43(14): p. 5488-5494.

119.

Xiao, Y. and M.R. Wiesner, Characterization of surface hydrophobicity of engineered
nanoparticles. Journal of Hazardous Materials, 2012. 215: p. 146-151.

120.

Shields, P.A. and S.R. Farrah, Characterization of virus adsorption by using DEAEsepharose and octyl-sepharose. Applied And Environmental Microbiology, 2002. 68(8):
p. 3965-3968.

121.

Brant, J.A. and A.E. Childress, Assessing short-range membrane-colloid interactions
using surface energetics. Journal of Membrane Science, 2002. 203(1-2): p. 257-273.

122.

Bowen, W.R., et al., A new technique for membrane characterisation: direct
measurement of the force of adhesion of a single particle using an atomic force
microscope. Journal of Membrane Science, 1998. 139(2): p. 269-274.

123.

Craig, V.S.J., B.W. Ninham, and R.M. Pashley, Study of the long-range hydrophobic
attraction in concentrated salt solutions and its implications for electrostatic models.
Langmuir, 1998. 14(12): p. 3326-3332.

124.

Hook, M.S., P.G. Hartley, and P.J. Thistlethwaite, Fabrication and characterization of
spherical zirconia particles for direct force measurement using the atomic force
microscope. Langmuir, 1999. 15(19): p. 6220-6225.

125.

Chin, C.-J., S. Yiacoumi, and C. Tsouris, Influence of Metal Ion Sorption on Colloidal
Surface Forces Measured by Atomic Force Microscopy. Environmental Science &
Technology, 2002. 36(3): p. 343-348.

126.

Chung, E., et al., Adhesion of spores of Bacillus thuringiensis on a planar surface.
Environmental Science & Technology, 2010. 44(1): p. 290-296.

127.

Willing, G.A., et al., New approach to the study of particle-surface adhesion using
atomic force microscopy. Journal of Colloid and Interface Science, 2000. 226(1): p. 185188.

37

128.

Abu-Lail, L.I., et al., Quantifying the adhesion and interaction forces between
Pseudomonas aeruginosa and natural organic matter. Environmental Science &
Technology, 2007. 41(23): p. 8031-8037.

129.

Das, T., T. Becker, and B.N. Nair, Measurements on hydrophobic and hydrophilic
surfaces using a porous gamma alumina nanoparticle aggregate mounted on Atomic
Force Microscopy cantilevers. Thin Solid Films, 2010. 518(10): p. 2769-2774.

130.

Vakarelski, I.U., et al., Nanoparticle-terminated scanning probe microscopy tips and
surface samples. Advanced Powder Technology, 2007. 18(6): p. 605-614.

131.

Capella, B. and G. Dietler, Force-distance curves by atomic force microscopy. Surface
Science Reports, 1999. 34: p. 1-104.

132.

Zhang, W., A.G. Stack, and Y. Chen, Interaction force measurement between E. coli cells
and nanoparticles immobilized surfaces by using AFM. Colloids and Surfaces B:
Biointerfaces, 2011. 82(2): p. 316-324.

133.

Guerrero, A.R., et al., Exploring the surface sharge on peptide-gold nanoparticle
conjugates by force spectroscopy. Langmuir, 2010. 26(14): p. 12026-12032.

134.

Ikai, A., et al., Atomic force microscope of bacteriophage T4 and its tube-baseplate
complex. FEBS Letters, 1993. 326(1–3): p. 39-41.

135.

Kolbe, W.F., D.F. Ogletree, and M.B. Salmeron, Atomic Force Microscopy Imaging of
T4 Bacteriophages on Silicon Substrates. Ultramicroscopy, 1992. 42: p. 1113-1117.

136.

Dubrovin, E.V., et al., Atomic force microscopy as a tool of inspection of viral infection.
Nanomed. Nanotechnol. Biol. Med., 2007. 3(2): p. 128-131.

137.

Zago, M., et al., Epifluorescence and atomic force microscopy: Two innovative
applications for studying phage-host interactions in Lactobacillus helveticus. Journal of
Microbiological Methods, 2012. 88(1): p. 41-46.

138.

Pereira, E.M.A., et al., Binding of dengue virus particles and dengue proteins onto solid
surfaces. ACS Applied Materials & Interfaces, 2010. 2(9): p. 2602-2610.

139.

Lee, J.W.-M., J.J.-H. Chu, and M.-L. Ng, Quantifying the specific binding between West
Nile virus envelope domain III protein and the cellular receptor alphaVbeta3 integrin.
Journal of Biological Chemistry, 2006. 281(3): p. 1352-1360.

140.

Liu, C.-H., et al., Localization and force analysis at the single virus particle level using
atomic force microscopy. Biochemical and Biophysical Research Communications, 2012.
417(1): p. 109-115.

38

141.

Tong, M., et al., Deposition kinetics of MS2 bacteriophages on clay mineral surfaces.
Colloids and Surfaces B: Biointerfaces, 2012. 92(0): p. 340-347.

142.

Yuan, B., M. Pham, and T.H. Nguyen, Deposition Kinetics of Bacteriophage MS2 on a
Silica Surface Coated with Natural Organic Matter in a Radial Stagnation Point Flow
Cell. Environmental Science & Technology, 2008. 42(20): p. 7628-7633.

143.

Pham, M., E.A. Mintz, and T.H. Nguyen, Deposition kinetics of bacteriophage MS2 to
natural organic matter: Role of divalent cations. Journal of Colloid and Interface
Science, 2009. 338(1): p. 1-9.

144.

Gutierrez, L., et al., Deposition and aggregation kinetics of rotavirus in divalent cation
solutions. Environmental Science & Technology, 2010. 44(12): p. 4552-4557.

145.

Rydell, G.E., et al., QCM-D studies of human norovirus VLPs binding to
glycosphingolipids in supported lipid bilayers reveal strain-specific characteristics.
Glycobiology, 2009. 19(11): p. 1176-1184.

146.

da Silva, A.K., et al., Adsorption and Aggregation Properties of Norovirus GI and GII
Virus-like Particles Demonstrate Differing Responses to Solution Chemistry.
Environmental Science & Technology, 2010. 45(2): p. 520-526.

147.

Britt, D.W., J. Buijs, and V. Hlady, Tobacco mosaic virus adsorption on self-assembled
and Langmuir-Blodgett monolayers studied by TIRF and SFM. Thin Solid Films, 1998.
327-329(0): p. 824-828.

148.

Lvov, Y., et al., Successive deposition of alternate layers of polyelectrolytes and a
charged virus. Langmuir, 1994. 10(11): p. 4232-4236.

149.

Valentine, R.C. and A.C. Allison, Virus particle adsorption I. Theory of adsorption and
experiments on the attachment of particles to non-biological surfaces. Biochimica et
Biophysica Acta, 1959. 34(0): p. 10-23.

150.

Penrod, S.L., T.M. Olson, and S.B. Grant, Deposition kinetics of two viruses in packed
beds of quartz granular media. Langmuir 1996. 12(23): p. 5576.

151.

van Voorthuizen, E.M., N.J. Ashbolt, and A.I. Schafer, Role of hydrophobic and
electrostatic interactions for initial enteric virus retention by MF membranes. Journal of
Membrane Science, 2001. 194(1): p. 69-79.

152.

Bales, R.C., et al., Bacteriophage adsorption during transport through porous media:
chemical perturbations and reversibility. Environ. Sci. Technol., 1991. 25: p. 2088-2095.

153.

Bales, R.C., et al., MS-2 and poliovirus transport in porous media: Hydrophobic effects
and chemical perturbations. Water Resour. Res., 1993. 29(4): p. 957-963.

39

154.

Gerba, C.P., et al., Quantitative assessment of the adsorptive behavior of viruses to soils.
Environ. Sci. Technol. , 1981. 15: p. 940-944.

155.

LaBelle, R.L. and C.P. Gerba, Influence of pH, salinity, and organic matter on the
adsorption of enteric viruses to estuarine sediments. Applied and Environmental
Microbiology, 1979. 38: p. 93-101.

156.

Michen, B., et al., Virus Removal in Ceramic Depth Filters Based on Diatomaceous
Earth. Environmental Science & Technology, 2012. 46(2): p. 1170-1177.

157.

Chrysikopoulos, C.V. and V.I. Syngouna, Attachment of bacteriophages MS2 and Phi
X174 onto kaolinite and montmorillonite: Extended-DLVO interactions. Colloids and
Surfaces B-Biointerfaces, 2012. 92: p. 74-83.

158.

Sadeghi, G., et al., Systematic study of effects of pH and ionic strength on attachment of
phage PRD1. Ground Water, 2011. 49(1): p. 12-19.

159.

Dowd, S.E., et al., Delineating the specific influence of virus isoelectric point and size on
virus adsorption and transport through sandy soils. Applied and Environmental
Microbiology, 1998. 64(2): p. 405-410.

160.

Gutierrez, L., et al., Adsorption of rotavirus and bacteriophage MS2 using glass fiber
coated with hematite nanoparticles. Water Research, 2009. 43(20): p. 5198-5208.

161.

Bitton, G., O. Pancorbo, and G.E. Gifford, Factors affecting the adsorption of poliovirus
to magnetite in water and wastewater. Water Res., 1976. 10: p. 978-980.

162.

Syngouna, V.I. and C.V. Chrysikopoulos, Interaction between viruses and clays in static
and dynamic batch systems. Environmental Science & Technology, 2010. 44(12): p.
4539-4544.

163.

Chattopadhyay, S. and R.W. Puls, Forces dictating colloidal interactions between viruses
and soil. Chemosphere, 2000. 41(8): p. 1279-1286.

164.

Lee, S.-Y., et al., Coagulation of tobacco mosaic virus in alcohol-water-LiCl solutions.
Journal of Colloid and Interface Science, 2008. 324: p. 92-98.

165.

Loveland, J.P., et al., The reversibility of virus attachment to mineral surfaces. Colloids
Surf. A, 1996. 107: p. 205-221.

166.

Gregory, J., Approximate expressions for retarded van der waals interaction. Journal of
Colloid and Interface Science, 1981. 83(1): p. 138-145.

167.

Murray, J.P. and G.A. Parks, Poliovirus adsorption on oxide surfaces, in Particulates in
Water1980, American Chemical Society. p. 97-133.

40

168.

Israelachvili, J.N., Intermolecular and surface forces - Second edition1991, New York:
Academic Press.

169.

Yoon, R.H., D.H. Flinn, and Y.I. Rabinovich, Hydrophobic interactions between
dissimilar surfaces. Journal of Colloid and Interface Science, 1997. 185(2): p. 363-370.

170.

Neumann, A.W., et al., Equation-of-state approach to determine surface tensions of lowenergy solids from contact angles. Journal of Colloid and Interface Science, 1974. 49(2):
p. 291-304.

171.

van De Ven, T.G.M., et al., The surface tension of solids. Journal of Colloid and Interface
Science, 1983. 91(1): p. 298-299.

172.

Bitton, G., Adsorption of viruses onto surfaces in soil and water. Water Res., 1975. 9: p.
473-484.

173.

Lytle, D.C. and L.B. Routson, Minimized virus binding for tests of barrier materials.
Applied and Environmental Microbiology, 1995. 61: p. 643-649.

174.

Brady-Estevez, A.S., et al., Impact of solution chemistry on viral removal by a singlewalled carbon nanotube filter. Water Research, 2010. 44(13): p. 3773-3780.

175.

Puck, T. and B. Sagik, Virus and cell interaction with ion exchangers. The Journal of
Experimental Medicine, 1953. 97(6): p. 807-820.

176.

Cookson, J.T. and W.J. North, Adsorption of viruses on activated carbon. Equilibriums
and kinetics of the attachment of Escherichia coli bacteriophage T4 on activated carbon.
Environmental Science & Technology, 1967. 1(1): p. 46-52.

177.

Kellenberger, E., et al., Mechanism of the long tail-fiber deployment of bacteriophages Teven and its role in adsorption, infection and sedimentation. Biophysical Chemistry,
1996. 59(1–2): p. 41-59.

178.

Nguyen, T.H., et al., The RNA core weakly influences the interactions of the
bacteriophage MS2 at key environmental interfaces. Soft Matter, 2011. 7(21): p. 1044910456.

41

Chapter 2. Anti-adhesive membrane films for virus concentration
and recovery from water prior to detection

By Elodie V. Pasco, Irene Xagoraraki, Syed A. Hashsham, Merlin L. Bruening, Volodymyr V.
Tarabara.
Chapter 1.
2.1

Introduction

Efficient pathogen detection in drinking and recreational water is vital for preventing
disease outbreaks. Because waterborne pathogens are present in low concentrations, a fast and
reliable concentration step is crucial for their detection. Crossflow filtration with hollow fiber
membranes, an emerging technique for concentrating various waterborne pathogens [1-13]
including viruses [7-13], offers several advantages over traditional pathogen concentration
methods (see below). First, membranes with an appropriate pore size simultaneously concentrate
microbial pathogens ranging from protozoa to bacteria to viruses [2, 3, 5, 6] without
concentrating molecular toxins and low molecular weight qPCR inhibitors [14]. Second,
crossflow decreases membrane fouling and results in increased concentration rates [15]. Third,
in contrast to filtration in a dead-end geometry, crossflow minimizes virus-membrane contact so
more viruses remain suspended in the retentate throughout the concentration process, facilitating
virus recovery [14].

The EPA-approved VIRADEL (VIRus ADsorption-ELution) method [16] for virus
concentration includes filtration of a water sample through a positively [7, 16-18] or negatively
42

[16, 18, 19] charged microfilter. During the filtration, viruses adsorb onto the filter, and elution
occurs in a high pH solution containing 1.5 to 6% beef extract and/or 0.05 M glycine (Fig. 2.1).
If necessary, addition of a flocculant to the eluate can precipitate viruses to further reduce
volume [16]. The VIRADEL method often gives low or poorly reproducible recoveries [20].
For example, Virosorb 1-MDS, the only electropositive VIRADEL filter recommended by EPA,
adsorbed less than 35% of the indigenous Escherichia coli phages in sewage-spiked tap water,
and the eluate contained only 12% of the phages [17]. The VIRADEL method is also sensitive
to various water constituents; indeed, norovirus recovery was 92% from deionized (DI) water but
only 19% from surface water [18]. In addition, highly adhesive viruses such as human
adenovirus (HAdV) [21], give recoveries as low as 0.02% from surface water [7], and norovirus
recoveries can also be low [7, 22]. To increase recovery, several studies explored
ultracentrifugation [23, 24], centrifugal ultrafiltration (UF) [25], and flocculation-re-dissolutionUF [26]; however, these methods are impractical for large volume samples.

Crossflow filtration with a “blocked” membrane (see below) gives recoveries equal to or
better than those with VIRADEL for a range of viruses in various types of water [7, 11].
Additionally, the large membrane surface area in hollow fiber cartridges and the use of crossflow
to mitigate fouling [15] allow rapid concentration of high-volume samples [2, 4, 6, 11, 12, 27].
Covering or “blocking” the membrane surface with a layer of proteins (beef extract [10, 13], calf
serum [13], bovine serum albumin [13]) or glycine [10] is crucial to improve virus recovery
during crossflow filtration (Fig. 2.1).

43

Figure 2.1: Comparison of the proposed virus pre-concentration approach based on
polyelectrolyte multilayer coatings (right) with other current methods.
For interpretation of the references to color in this and all other figures, the reader is referred
to the electronic version of this dissertation)

These blocking procedures usually employ complex mixtures of macromolecules of
various molecular weights to cover virus adsorption sites and possibly minimize virus passage
through the membrane filter [2, 3, 5, 7, 11, 28]. Although the term “blocking” has also been
used to describe treatment with small molecular weight compounds such as sodium
polyphosphate (e.g., 600 Da [3, 6]) and glycine (75 Da), their effect is most likely limited to
preventing adsorption of viruses to the membrane surface and does not include true pore
44

blockage. The blocking step can be time consuming (e.g., overnight deposition of a protein
layer), and a membrane blocked with proteins might be prone to contamination during storage
and transport [3]. In addition for certain viruses such as HAdV recovery values remain low and
variable [7]

To overcome these challenges we propose coating UF filters with rationally designed
polyelectrolyte multilayer (PEM) films [29, 30]. PEM deposition is in effect controlled
membrane blocking that employs alternating adsorption of two oppositely charged polymers.
PEMs can form quickly (in less than 1 h) and reproducibly [31]. Several studies described antiadhesive PEM coatings on various surfaces [32-35], including membrane filters [36, 37], to
decrease the adsorption of selected proteins [32, 33, 36, 37], mammalian cells [32, 34], or
bacteria [34, 35]. Although PEMs were employed to create ordered virus monolayers [38, 39],
no studies examined virus adhesion to PEM-modified surfaces or applied PEMs to prevent virus
adsorption or to enhance recovery for virus concentration and detection.

In addition to low and poorly reproducible recoveries, the long duration of the sample
concentration step remains a bottleneck in virus detection. For example, concentrating 100 L of
tap water to 400 mL takes 2 h with a hollow fiber cartridge, and further volume reduction to a
few mL [6] for qPCR analysis requires another 1 to 3 h [6]. Because virus concentrations of less
than 1 genomic copy per 100 L of water represent a significant risk of illness [40], virus
monitoring often requires testing up to 1,000 L for drinking water [16]. Increasing
transmembrane pressure can, in principle, increase filtration rate; however, this may exacerbate
membrane fouling, alter virus recovery and offset the gains in the permeate flux. To overcome

45

this limitation we operated the membrane at a high transmembrane pressure while
simultaneously increasing the retentate crossflow rate to maintain a high permeate flux while
limiting fouling to a manageable level.

In summary, this work aims to improve virus concentration by membranes. We designed
highly controllable anti-adhesive membrane coatings and assessed their performance in terms of
virus recovery and concentration rate. We also evaluated how crossflow rate and permeate flux
affect virus rejection, recovery, and rate of concentration with CS-blocked and PEM-coated
membranes. The anti-adhesive membrane coating required less than an hour to deposit and
increased pre-elution virus recovery up to 2-fold.

2.2

2.2.1

Materials and Methods

Reagents

Lysozyme (from chicken egg white), ethylenediaminetetraacetic acid (EDTA), calf
serum, heparin (sodium salt), chitosan (medium molecular weight), diiodomethane and sodium
polyphosphate (NaPP) were purchased from Sigma-Aldrich. Tween 80 (Fisher), tryptic soy
broth (Becton, Dickinson) and glycerol (J.T. Baker) were used as received.

46

2.2.2

Bacteriophage

Bacteriophage P22 (provided by Prof. Joan B. Rose, MSU) is a non-enveloped, double
stranded DNA phage that has been used to understand dsDNA virus assembly [41-44] and as a
surrogate for human viruses to understand their fate in the environment [45, 46]. P22 has an
icosahedral head with a diameter of 60 nm and a short tail [47]. P22 stock was grown by
inoculating Salmonella enterica serovar Typhimurium LT2 (also provided by Prof. Rose) in
tryptic soy broth. After incubation bacterial host cells were lysed by adding 0.1 mL of lysozyme
(50 mg/mL) and 0.75 mL of EDTA (0.5 M). The culture was then centrifuged for 10 min at
4000 rpm and the supernatant was filtered through a 0.45 μm cartridge filter. P22 stock
suspension (~ 3·10

10

plaque forming units per mL) was stored at 4°C. One liter of water was

spiked with 1 mL of P22 stock on the day of the filtration.

2.2.3

Bacteriophage characterization

To characterize the surface properties of P22 particles, they had to be separated from
other components present in the stock, i.e. tryptic soy broth components and host bacteria cell
debris. The purification protocol was modified from Mylon et al. [48]. First, 150 mL of P22
stock solution was filtered through a 0.1 μm membrane (Nucleopore, Whatman) using a stirred
Amicon 8050 dead-end UF cell (Millipore) to remove molecules and debris larger than 100 nm.
The permeate volume was then reduced 5- fold by filtration through a 50 kDa polyethersulfone
membrane (Omega, Pall) in an Amicon 8200 filtration cell. At this point the cell was connected
to a feed tank containing 1 L of 10 mM KCl. This solution was passed through the 50 kDa
47

membrane to wash away molecules smaller than 50 kDa such as broth components, until only 30
mL remained in the cell. The particle size distribution in this purified stock was measured by
dynamic light scattering at pH 6.. The purified P22 stock was then filtered through a 30 kDa
polyethersulfone membrane (Omega, Pall) to deposit the equivalent of ~ 9 layers of virus.
Contact angles of 3 probe liquids (deionized water, glycerol, diiodomethane) with these layers
were measured using the sessile drop method (FTÅ 200 contact angle analyzer, First Ten
Angstroms) to determine the surface tension parameters of P22 phage particles (see section
2.2.12). The  -potential of P22 was determined based on electrophoretic mobility using phase
1

analysis light scattering (ZetaPALS, Brookhaven Instrument Corporation) at pH 6 , using
purified P22 washed with 1 mM KCl (instead of 10 mM KCl used for particle sizing and contact
angle).

2.2.4

Crossflow concentration apparatus

Fig. 2.2 shows the schematic of the crossflow concentration unit. A high-pressure
peristaltic pump (model 621 CC, Watson-Marlow) drew the solution from the pressurized feed
tank to the membrane filtration cell (CF042, Sterlitech), where the membrane surface area
exposed to the feed was 41 cm2. The permeate was collected in an Erlenmeyer flask positioned
on an electronic mass balance (Adventurer Pro AV8101C, Ohaus) interfaced with a computer.
The transmembrane pressure varied from ~ 0.5 bars (7 psi) to ~ 2.8 bars (40 psi). An inline
flowmeter (101-7, McMilan) recorded the crossflow velocity.

he  -potential of P22 determined with this method is a good descriptor
of P22 surface potential only if P22 is a soft particle, which has not been verified here.
48

pressure sensor
permeate
peristaltic pump
membrane module
P

concentrate

flowmeter

data acquisition
system

pressurized 1 L
feed tank

mass balance

Figure 2.2: Crossflow filtration apparatus

2.2.5

Membrane preparation

Prior to modification, a polyethersulfone (PES) UF membrane (OMEGA 30 kDa, Pall)
was submerged for 24 h in 1 N NaOH and thoroughly rinsed with deionized water. The
membrane was either blocked with CS or coated with a PEM film. For CS blocking, the
membrane was loaded in the crossflow filtration cell, and 500 mL of 5 % CS solution was
circulated over the membrane surface for 16 to 18 h with no pressure differential across the
membrane. The filtration unit was then rinsed twice with DI water for 10 min. To coat the
membrane with a PEM film, the membrane was placed in a custom-made holder and the
membrane surface was alternately exposed to HE solution (1 mg/L, pH 5) and CHI solution (1
mg/L, pH 5) with a 1-min rinse with DI water in between. The duration of each exposure was 5
min. The total number of deposited bilayers was 4.5, where the last half bilayer was HE. Both
49

HE and CHI solutions also contained 0.15 M NaCl. Membranes were prepared immediately
prior to filtration.

2.2.6

Membrane characterization

The  -potential of the membranes was measured using an electrokinetic analyzer (BIEKA, Brookhaven Instrument Corp.) at pH 6. Membrane hydrophilicity was characterized by
measuring water contact angles using the sessile drop method. At least 3 different membranes
2

were used for each analysis. Contact angles of diiodomethane and glycerol were also measured
to determine the surface tension parameters of the CS-blocked and the PEM-coated membranes.
Water permeabilities of coated membranes were determined before and after membrane
compaction by measuring the permeate flux at a transmembrane pressure differential of 40 psi.

2.2.7

Concentration procedure

The membranes were first compacted by filtering DI water at 40 psi for 90 min.
Filtration of 1 L of the DI water spiked with 1 mL of P22 stock solution was then performed to
reduce the volume to 250 mL. The ratio of the crossflow flux to the initial permeate flux,
J cf / J p , ranged from 1,500 to 11,100 (Fig. 2.3). Crossflow flux was regulated from 6,800 to
2

26,800 L/(min·m ), which corresponded to Reynolds numbers from 400 to 1,700.

50

this study: CS-blocked membrane

initial permeate fllux Jp, L/(min·m2)

10

this study: PEM-coated membrane
earlier studies

9

high Jp

high Jp

8

high Jcf
fastest concentration conditions

low Jcf

7
6
5
4
3

(5)

2
(3) (2, 11)

1

highest recovery conditions

0
0.0

0.5

1.0

1.5

2.0

2.5

3.0

crossflow flux Jcf ·104, L/(min·m2)
Figure 2.3: Filtration conditions in experiments on P22 preconcentration and recovery using CSblocked and PEM-coated membranes. Number of tests: n=27 for CS, n=22 for PEM.

Following the filtration, viruses were eluted from the membrane using 0.01% Tween 80,
0.01% NaPP [2, 3]. The elution occurred with the membrane still in the filtration cell and at the
same crossflow rate as in the filtration, but with a retentate pressure of only 2.5 psi, which was
due to the crossflow. Aliquots of the feed, retentate, permeate and eluate were stored at 4°C.
Samples were also taken during compaction with DI water to verify that no contamination was

51

present before P22 was introduced. A new membrane was used for each filtration and the unit
was disinfected with CIDEX (Advanced Sterilization Products) between filtration tests.

2.2.8

Sample analysis

The P22 concentration in each sample was determined by qPCR. Within 24 h of the
filtration experiment, MagNA Pure Compact System (Roche) was used to extract P22 DNA from
390 μL grab samples into 100 μL eluates. RNA carrier (10 µL, 1 μg/μL) was added to the 390
μL samples before extraction to prevent bacteriophage adsorption on the MagNA Pure system
tips and magnetic beads. Each extract was then analyzed by qPCR in triplicate following the
protocols described by Masago et al.[45] PCR mixtures (20 μL) contained 5 μL of extracted
DNA sample, 2 μL each of forward and reverse primers (5 μmol/μL), 0.3 μL of Taqman Probe
(10 μmol/L), 10 μl of PCR master mix (LightCycler 480 Probes, Roche) and 0.7 μL of water.
qPCR was performed using the LightCycler 1.5 system (Roche). Primer and probe sequences
(Table 2.1), and qPCR conditions were similar to the ones described by Masago et al. [45].
QPCR mixtures were heated at 95°C for 10 min prior to 45 cycles of 95°C for 10 s, 60°C for 20 s
and 72°C for 10 s. Finally mixtures were cooled for 30 s at 40°C. To prepare a standard curve
linking qPCR crossing-point (CP) values with numbers of P22 DNA copies, P22 DNA standards
were prepared following the method described by Xagoraraki et al. [49]. The phage target DNA
sequence was amplified by PCR using the same primers that were used for qPCR, and the
amplicon was cloned into plasmid vector using the TOPO TA Cloning kit (Invitrogen). Plasmid
vector DNA was extracted and purified using QIAprep Spin Miniprep kit (Qiagen) and the DNA
concentration was quantified by spectrophotometry using Nanodrop ND-1000 (Thermo Fisher

52

Scientific). The number of DNA copies per unit volume could then be calculated based on the
plasmid vector DNA length of 4,000 base pairs (bp) and assuming an average mass per bp of 650
Da. The purified DNA was serially diluted and the dilutions run by qPCR to develop the
standard curve.

Table 2.1: Primer and probe sequence (from Masago et al. [45])
Primer/probe

Sequence

Reverse

CTT AAC AAG CTC TGA CTG CTC ATC A

Forward

CCA TCG CCT GTG ACT CGA T

Taqman Probe

FAM-TCG CAA CGA TGC AGA ACG ACT CGTAMRA

2.2.9

Data analysis and statistics

P22 log removal (LRV), percentage of pre-elution recovery ( rpre ), and percentage of
post-elution recovery ( rpost ) were calculated using the following equations:

C f
LRV  log
Cp

r pre 

(2.1)


,



(2.2)

C rVr
 100% ,
CfVf

r post 

C rVr  CeVe
 100% ,
CfVf

(2.3)

53

where C f , C r , and Ce are P22 concentrations in feed, retentate, and eluate samples,
respectively, and V f , V r , and Ve are the volumes of these samples. Virus recovery and virus
removal were plotted as a function of the Jcf / J p ratio. The relationships between Jcf / J p , on
the one hand, and rpre , rpost and LRV, on the other hand, for the PEM-coated membranes and
the CS-blocked membranes were statistically analyzed using the regression procedure of the SAS
software. An interaction term (between Jcf / J p and the type of membrane) was introduced in
the regression model. The software Minitab was used to statistically characterize the
distributions of pre-elution and post-elution recoveries.

2.2.10 Determination of the fouling resistance Rf

As we concentrate P22, permeate flux decreases due to fouling of the membrane by
virions and other components present in the feed (proteins and bacterial debris). The flow
resistance due to fouling at the end of filtration was calculated using the following equation:
(2.4)

J

R f  Rm  0  1
Jf




Where R f is the resistance due to fouling, Rm the resistance of the blocked or coated
membrane, J 0 the average permeate flux during the first minute of filtration and J f is the
permeate flux during the last minute of filtration.

54

2.2.11 Blocking law analysis

Dead-end filtrations of P22-containing feed solution (1 L of DI water spiked with 1 mL
of P22 stock) through CS-blocked and PEM-coated membranes were conducted using an
unstirred Amicon 8010 ultrafiltration cell connected to a pressurized feed tank. The permeate
flux was recorded continuously for 1 hour using an electronic mass balance connected to a
computer. For each experiment the membrane was first compacted for 90 min with DI water.
Permeate flux data were then obtained and fitted to the linearized equations describing the four
fouling mechanisms: complete blocking, standard blocking, intermediate blocking and cake
filtration [50]. To assign a fouling mechanism to each portion of the permeate flux dataset, we
used the procedure described by Wang et al. [50]. A core portion of the filtration dataset that
2

fitted a given mechanism with a high coefficient of determination (r > 0.999 unless otherwise
specified) was identified and expanded point by point by evaluating if the point belonged to the
95% prediction interval given by the regression. If it did, the point was added to the core and a
new linear regression analysis was performed. For more detail on the procedure, please refer to
the paper by Wang et al. [50].

2.2.12 Calculation of the energy of interaction between P22 phage and membrane surfaces using
the XDLVO approach

The XDLVO theory predicts the energy of interaction between particles and a surface by
taking into consideration Lifshitz-van der Waals (LW), electrostatic double layer (EL) and Lewis

55

acid-base (AB) forces [51, 52]. The total energy of interaction between a particle and a plane
surface can be written as [52]:
(2.5)

U XDLVO  U LW  U EL  U AB
XDLVO

where U
EL

U

is the total energy of interaction, U

is the electrostatic interaction energy and U

AB

LW

is the van der Waals interaction energy,

is the Lewis acid-base interaction energy.

Each of these terms can be expressed as a function of the separation distance ( ) using the
following equations [52]:
2
LW d   4   LW   LW   LW   LW  y 0 r p
U
 
 p

s



 d


d

EL d     r 2  ln 1  e
U
r 0 p
p s 
d

1 e





    2   2  ln1  e  2d

s  
  p
 


(2.6)







  


 

 
     s   p          s   p    



 exp y0  d 
U AB d   4r p  


  


 
 
  s  p   s  p




(2.7)

(2.8)

_
LW
where   ,   and   are the surface tension parameters of immersion liquid (the + and 

superscripts correspond to the electron acceptor and electron donor components of the surface

_
LW

tension, respectively),  p ,  p and  p are the surface tension parameters of the particle,
_
LW

 s ,  s and  s are the surface tension parameters of the planar surface, r p is the particle

radius, and  p and  s are the zeta potentials of the particle and the surface. The minimum
equilibrium cut-off distance, y 0 , was assigned a value of 0.16 nm [53]. For other symbols,

56

 r 0 is the dielectric permittivity of water,  is the inverse of the Debye length and  is the
characteristic decay length of Lewis acid-base interactions in water (0.6 nm) [53].

The surface tension parameters of the P22 phage and membrane surfaces were
determined using the extended Young equation (2.9) after measuring the contact angle of three
probe liquids with known surface tension parameters.
(2.9)

TOT  2  LW  LW          
(1  cos ) 
 s
s 
s  




 AB  2   

(2.10)

 TOT   AB   LW

(2.11)

In these equations,  is the contact angle for a given probe liquid, and the subscripts

and 

correspond to the solid surface and the liquid probe, respectively.





2.2.13 Effect of polyelectrolytes and CS on qPCR analysis

8

A P22 solution (~ 3·10 PFU/mL) was prepared by diluting 1 mL of P22 stock with
deionized (DI) water. One mL of 0.12 mg/mL solutions of HE, PAA, poly(styrene sulfonate) ,
CHI and PDADMAC, and 1 mL of a 0.24 mg/mL calf serum solution were seeded with 0.1 mL
of the P22 solution. The pH of each resulting solution was adjusted to 6.3 (similar to the feed pH
during filtration). The final concentration of polyelectrolyte and calf serum in each solution was
0.11 and 0.22 mg/L respectively, which is 100-fold higher than the concentration we would

57

expect if we would backflush our membrane with 100 ml of solution. The P22 concentration in
these samples was determined using the same methodology as for the filtration samples.

2.3

2.3.1

Results and Discussion

PEM selection

As described in Chapter 1 various polyelectrolytes have been used to prepare anti-adhesive
coatings. Based on these studies HA/CHI, HA/PAH, HE/CHI and HE/PAH films were deposited
on UF membranes. (PAA/PAH films are not highly permeable.) Contact angles of water with
the resulting surfaces and water permeability were measured. Based on a limited set of data it
appeared that HE based films were more hydrophilic than their HA counterparts.

Additionally we looked at the effect of polyelectrolytes on qPCR analysis. As Fig. 2.4
shows, polycations out of 3 significantly inhibited DNA extraction or qPCR while most
polyanions had no effect on qPCR results. Although CHI partially inhibited qPCR, PAH
completely annihilated any fluorescence signal and as a consequence P22 could not be detected
in samples containing PAH. As a result, CHI was considered a better candidate than PAH and
based partly on the results of these inhibition tests and partly on the measured values of water
contact angles and water permeabilities of candidate PEMs, we selected (HE/CHI)4.5 as one
suitable PEM film. Several polyelectrolytes such as PDADMAC or PAA, not considered in
these studies, were more compatible than CHI or HE with qPCR and should be considered in
future work
58

Log P22 concentration as determined by
qPCR

7
6
5
4
3
2
1
0

potential inhibitors
Figure 2.4: Effect of various polyelectrolytes and calf serum on the P22 concentration
determined by qPCR. Error bars indicate the standard deviation interval. n=3.

2.3.2

Properties of CS-blocked and PEM-coated membranes and P22 phage

2.3.2.1 Membrane permeability and fouling mechanisms

The DI water permeabilities of PES, CS-blocked PES, and PEM-coated PES membranes
2

after compaction were 280  20, 90  20, and 120  30 L/(m ·h·bar), respectively. Thus, the two
modification methods yielded membranes of similar permeabilities. We applied blocking laws
to P22 dead-end filtration data to interpret how feed components interacted with CS-blocked and
59

PEM-coated membranes. As Fig. 2.5 shows, for both membranes each of the four mechanisms
could fit at least a portion of the dataset. Concurrent fits to different mechanisms indicate either
a combination of mechanisms or that the data are insufficient to differentiate between
mechanisms. Clearly, however, a cake filtration law provides a high quality fit to plots of flux
versus time after filtering 5 mL of virus suspension through PEM-coated membranes, but only
after filtering 32 mL through CS-blocked membranes. (Filtration of one mL corresponds to one
2

P22 phage, quantified as DNA copies, deposited per 1 μm of membrane). The larger volume
required to achieve cake filtration with CS-blocked membranes suggests that P22 (or other
components not removed from the growth medium by centrifugation) can partially or fully enter
pores in the CS layer. For a description of different blocking mechanisms please refer to the
study by Wang et al. [50].

2.3.2.2 Surface properties of membranes and P22 phage

Several studies have shown that blocking the membrane with proteins improves virus
recovery [3, 10, 28]. While two studies reported that blocking a membrane with proteins had no
[4] or a negative [12] effect on virus recovery, most studies used CS or other proteinaceous
solutions to block membranes for microorganism concentration by crossflow filtration [2, 3, 5, 7,
10, 11, 13]. .

60

a)

b)
1.4
60
*

1.2
y = 0.0172x + 0.3635
r2 = 0.99965

40
30
*

20

*

0.8

y = 0.003x + 0.7033
R² = 0.5762

0.6
0.4

PEM-coated membrane
CS-blocked membrane
t/V=f(V) linearisation for V>5ml (PEM)
t/V = f(V) linearisation for V>30mL (CS)

cake filtration

10
0

complete
blocking
intermediate
blocking
standard
blocking

1.0
t/V (min/mL)

permeate volume (mL)

50

0.2
CS

PEM

0.0
0

20

40
V (mL)

60

Figure 2.5: Blocking law analysis for fouling of CS-blocked and PEM-coated membranes during filtration of DI water spiked with
2

P22 stock. a) Distributions and contributions of the blocking mechanisms in terms of permeate volume. * indicates that the r value
2

corresponding to fitting that blocking mechanism with the experimental data was > 0.99. When no * is present, r was > 0.999. b)
Inverse cumulative flow rate (t/V) as a function of permeate volume V for CS-blocked and PEM-coated membranes.
61

Some of these studies suggest that treating the membrane with beef extract or CS
solutions minimizes viral adsorption to the membrane [10, 13]; however, to our knowledge all
evidence for such an effect is indirect and based on higher values for virus recovery. Previous
studies did not consider the specific interactions between the virus and the membrane.

To quantify the anti-adhesiveness of the CS layer, we measured the surface energy (Table
2.2) and charge of the CS-blocked membrane. Based on literature values of the  -potential and
isoelectric point, at least 2 out of the 4 viruses were negatively charged [55, 56] in studies
showing that CS or beef extract increased recovery [10, 13, 54]. (Isoelectric points for the other
2 viruses are not known.) Therefore, we expected that CS renders the membrane surface more
hydrophilic and negatively charged to make it more anti-adhesive; however, CS-blocked
membranes were slightly positively charged at pH 6 (  = 3  2 mV) and not more hydrophilic (

 w = 66  13°) than the polyethersulfone support (  w = 67  3°). In fact, the CS-blocked
membranes were hydrophobic (see Table 2.2) according to van Oss and Giese’s definition of
hydrophobicity (a negative free energy of interfacial interaction when immersed in water; see
Chapter 1) [57]. Although the IEP of BSA, the main component of CS, is 4.7 so the protein
would have a slight negative charge at pH 6, the CS also includes components other than BSA
that might preferentially deposit and be responsible for the slightly positive charge of the CSblocked surface. Thus, although CS can prevent virus attachment to the UF support membrane
by occupying adsorption sites, negatively charged viruses might still adsorb onto CS surface
during filtration.

62

Table 2.2 Measured contact angles, calculated surface energy parameters and free energy of interfacial interaction when immersed in
water for P22 phage, CS-blocked membranes and PEM-coated membranes.
2

Giwi

Surface energy parameters (mJ/m )
Probe liquid

 LW





 AB

 TOT

Ultrapure water

21.8

25.5

25.5

51.0

34.0

3.9

57.4

30.0

64.0

Diiodomethane

50.8

0.0

0.0

0.0

2

(mJ/m )

72.8

Glycerol

Contact angle (º)

50.8

Material

Diiodo-

H2O

Glycerol
methane

P22

44

3

57

27

71

25

49 ± 8

75 ± 6

30 ± 2

CS-blocked membrane

41

0.004

17

0.5

42

-24

66 ± 13

64 ± 2

37 ± 3

PEM-coated membrane

42

0.07

53

4

46

36

36 ± 3

55 ± 4

34 ± 5

63

This could suggest that CS blocking improves recovery by limiting virus penetration into
the membrane but not by minimizing virus adhesion. Facile elution of viruses attached to the CS
might also enhance recovery. In contrast, the PEM-coated membranes were negatively charged
at pH 6 (  = -7  3 mV) and considerably more hydrophilic (  w = 36  3°) than the CS-blocked
membranes. Given that P22 phages were negatively charged at pH 6 (  = -17  1 mV) and
hydrophilic (  w = 49  8°), we expect the PEM to resist P22 adsorption. As a precedent in this
area, Hill et al. [3] and Polaczyk et al. [6] evaluated the efficiency of a 15-min blocking of 15-20
kDa UF membrane with a 600 Da NaPP solution. Although surface charges were not measured,
they hypothesized that deposition of the polyanion NaPP increased electrostatic repulsion
between the membrane and microorganisms. NaPP “blocking” was faster than overnight CSblocking and resulted in similar recovery for MS2 bacteriophage [3, 6]. To further quantify the
propensity of P22 to adhere to the CS-blocked and PEM-blocked membranes, we calculated the
energy of phage-membrane interactions.

2.3.2.3 Virus-membrane interaction energies

The XDLVO theory predicted secondary minima in the energy of the virus-membrane
interaction at separations of 4.3 nm and 10.9 nm for the CS-blocked and PEM-coated
membranes, respectively (Fig. 2.6). Due to attractive van der Waals and electrostatic
interactions between CS-blocked membranes and P22, the secondary minimum (-2.6 kT) could
be enough for some P22 to reversibly attach to the membrane surface despite not being able to
overcome the repulsive barrier (500 kT) [58]. The depth (-0.6 kT) of the secondary minimum in
the PEM-P22 interaction energy was less than the average thermal energy of particles, indicating
64

that attachment of P22 to this membrane was not efficient. This XDLVO analysis only takes
into account macroscopic phenomena, and adhesion of P22 to either surface could also occur
through microscopic attraction to discreet electron acceptor sites on the membrane surfaces [53]
as well as through non-XDLVO interactions.

3
interaction energy, kT units

a)

b)
3

2

2

1

1

0

0
0

10

van der Waals
Lewis acid-base
electrostatic
total interaction energy

0

20

-1

20

-1

-2

10

-2

-3

-3

separation distance, nm

separation distance, nm

Figure 2.6: XDLVO energy profiles for the interaction of P22 with a) CS-blocked membranes
and b) PEM-coated membranes at pH = 6 and in 10 mM KCl.

2.3.3

Virus recovery with PEM-coated and CS-blocked membranes

2.3.3.1 Pre-elution recovery

Figure 2.7 shows P22 recoveries in the retentate as a function of the J cf / J p ratio. The
Jcf / J p ratio is important as it influences the transport of viruses to and, possibly, across the

65

membrane [59]. Generally, if permeate flow increases, increased concentration polarization
leads to both more particle adhesion to the membrane and more particle passage through the
membrane. In contrast, increasing the crossflow velocity decreases concentration polarization
and facilitates particle removal from the membrane surface through shear-induced backtransport[15]; however, it has also been shown that, in conditions typical of crossflow
ultrafiltration, the minimum

Pre-elution recovery with
CS-blocked membranes ranged from 22% to 51%. These values are comparable to results of a
previous study with CS-blocked UF membranes where pre-elution recoveries were 38  22% for
T1 phage and 45  55% for PP7 phage [5]; we should note, however, that this prior study
employed surface water that could foul the membrane significantly [9]. When concentrating
bacteriophage MS2 from tap water, Hill et al. [3] reported pre-elution recoveries of 84  13%.
This high recovery probably stemmed from the relatively simple composition of the water
samples combined with the addition of 0.1% NaPP, a negatively charged chemical dispersant
thought to enhance charge repulsion between the membrane and viruses by altering the viruses’

 -potential [6].

P22 pre-elution recovery with PEM-coated membranes ranged from 32 to 85%. Over the
entire range of filtration conditions (Fig. 2.7), PEM-coated membranes statistically outperformed
CS-blocked membranes in terms of P22 pre-elution recovery, but the improvement was most
pronounced at higher Jcf / J p ratios. We attribute the observed higher pre-elution to the PEM’s
anti-adhesiveness stemming from its negative charge [30] and hydrophilicity [29].
66

recovery postelution, %

100

100

90

recovery pre-elution , %

80
70

90
0
80
CS

60

PEM

50
40
30
20
P22 pre-elution with CS-blocked membrane
P22 pre-elution with PEM-coated membrane

10
0
0.0

0.2

0.4

0.6

0.8

1.0

crossflow to permeate flux ratio Jcf/Jp

1.2

104

Figure 2.7: Pre-elution and post-elution (see inset) recovery of P22 from DI water. Because
there was no significant dependence of post-elution recovery on the value of Jcf / J p , only the
average recovery and the 90% confidence interval are reported.
Number of experiments: n = 26 for CS-blocked membranes and n = 22 for PEM-coated
membranes.

2.3.3.2 Post-elution recovery and potential of PEMs as sacrificial coatings

Post-elution recovery of P22 did not differ significantly between CS-blocked (99  11%)
and PEM-coated membranes (96  9%). These values for post-elution recovery were within the
67

range typically reported for crossflow filtration of bacteriophages [2, 5, 6, 11-13, 28]. For
example, Hill et al. [2] recovered 86  13% of the ΦX174 phage and 120  22% of MS2 phage
from a 100 L tap sample containing 0.01% NaPP. The high post-elution recovery with both CSblocked and PEM-coated membranes may reflect the simplicity of the water matrix or weak
adhesion of P22 to surfaces. We expect that with more adhesive viruses, recovery should
decrease, and the competitive advantage of the PEM as a more anti-adhesive surface will
increase. When 10-L surface water samples were concentrated by crossflow filtration followed
by elution, Albinana-Gimenez et al. [7] observed post-elution recoveries of only 18.7% for
polyomavirus, 5.06% for adenovirus and less than 0.95% for norovirus. Anti-adhesive properties
of PEM-coated membranes, evidenced by high pre-elution recoveries of P22, might help to
improve the recovery of such “sticky” and difficult to concentrate human viruses.

Treating the PEM film as a sacrificial layer and removing it from the support using a
backflush could also increase post-elution recovery. Such removal will not affect the supporting
UF membrane and should help to recover surface-associated viruses. After the backflush, a new
PEM layer can be deposited at the surface of the same support [61]. To implement such a
procedure, however, electrolytes of the dissolved PEM should minimally interfere with qPCR.
Tests with select polyelectrolytes indicate that although some polyanions and most polycations
interfere with qPCR analysis; there are polyanions and polycations that, at the concentration
expected in backflush samples, do not inhibit qPCR or DNA extraction (see Fig. 2.4). Indeed
adding PSS, HA, PAA or PDADMAC to P22 samples did not affect the count of P22 by qPCR
when compared to the control sample where P22 was suspended in DI water.

68

2.3.3.3 Reproducibility

Blocking with CS is inherently poorly reproducible, so we expected the PEM-coated
membranes to show more reproducible recoveries than CS-blocked membranes. We performed
15 filtrations under conditions that should lead to the highest recoveries (“low J p ; high J cf ”
domain) with CS-blocked membranes, and 14 filtrations under the same conditions with PEMcoated membranes. Figure 2.8 illustrates the distribution of pre-elution and post-elution
3

recoveries under these conditions for both membrane types. The interquartile range (IQR, also
called midspread) for pre- and post-elution recoveries were comparable for both types of
membranes. IQRs for pre-elution recovery distributions are 15% and 16% for CS-blocked
membranes and PEM-blocked membranes respectively, whereas IQRs for post-elution recovery
distributions were lower, 10% and 9%, respectively for CS-blocked and PEM-coated
membranes. We conclude that there was no significant difference in recovery reproducibility
between the two types of membranes.

The interquartile range is the difference between the upper and lower quartiles of the
cumulative distribution curve.
69

relative frequency of occurrence

a)

calf serum

0.5

PEM

0.4
0.3
0.2
0.1
0.0

pre-elution recovery (%)

relative frequency of occurrence

b)
0.5

calf serum
PEM

0.4
0.3
0.2
0.1
0.0

post-elution recovery (%)
Figure 2.8: Relative frequency of occurrence of P22 (a) pre-elution and (b) post-elution recoveries
from DI water under the operational conditions optimized to maximize P22 recovery
( J cf / J p > 6,400). Number of experiments: n= 15 for CS-blocked membranes and n=14 for
PEM-coated membranes.
70

2.3.4

Virus log removal

LRV values varied in the 3.9  1.2 range for CS-blocked membranes and in the 4.1  1.1
range for PEM-coated membranes (Fig. 2.9). There was no significant difference between the
two types of membranes and no significant influence of Jcf / J p on virus removal.

6

CS-blocked membrane
PEM-coated membrane

log removal

5
4
3
2
1
0
0.0

0.2

0.4

0.6

crossflow to permeate flux ratio Jcf/Jp

0.8

1.0

104

Figure 2.9: P22 removal from DI water by CS-blocked and PEM-coated membrane at various
J cf / J p .

71

2.3.5

Effect of operational conditions on virus concentration

2.3.5.1 Operational conditions and flux

Figure 2.3 illustrates the range of operational conditions employed in this work. The J p
values are the permeate fluxes averaged over the first minute of filtration. For experiments
conducted at lower transmembrane pressures (Fig. 2.3; “low J p ” domain), plotted J p values
also represent the average J p as minimal or no fouling occurred in these tests. In experiments
run at higher transmembrane pressures (Fig. 2.3; two “high J p ” domains) fouling did occur,
especially for PEM-coated membranes (Fig. 2.10), so average permeate fluxes were lower than
the plotted initial values of J p (Fig. 2.11 shows average rather than initial fluxes). For
comparison, Fig. 2.3 and Fig. 2.11 also show J p values from earlier studies where crossflow
membrane filtration was used to recover viruses and where values of both J p and J cf were
available (see Appendix A).

Compared to earlier studies [2, 3, 5, 6, 11, 13], we evaluated a relatively large range of
operational conditions by varying both J p (2.0 to 8.6 (L/min·m2)) and J cf (6,800 to 26,800
2

(L/min·m )). This allowed us to assess the virus concentration performance of CS-blocked and
PEM-coated membranes under conditions optimal for either highest virus recovery or fastest
sample concentration (Fig. 2.7). The highest recovery conditions belong to the “low J p ; high

72

J cf ”domain, while the fastest concentration conditions correspond to the “high J p ; high J cf

” domain (Fig. 2.3 and 2.11).

9.E+12
CS-blocked membrane

8.E+12
fouling resistance Rf × 1012, m-1

PEM-coated membrane
7.E+12
6.E+12
5.E+12
4.E+12
3.E+12
2.E+12
1.E+12
0.E+00
0

0.5

1

1.5

2

crossflow to permeate flux ratio Jcf/Jp × 104
Figure 2.10: Observed fouling resistance (averaged over the duration of experiment) for CSblocked and PEM-coated membranes during the concentration of 1L of DI water seeded with
P22 to 250 mL at various crossflow to permeate flow ratios.

73

average permeate fllux Jp, L/(min·m2)

this study: CS-blocked membrane
this study: PEM-coated membrane
earlier studies

6
low Jp

high Jp

low Jcf

5

high Jcf

4
3
2

(5)
(2, 11)

1

low Jp
high Jcf

(3)

0
0.0

0.5

1.0

1.5

2.0

2.5

3.0

crossflow flux Jcf ·104 , L/(min·m2)
Filtration conditions: average permeate flux and crossflow flux, used for P22
recovery with CS-blocked membrane and PEM-coated membranes. n=27 for CS, n=22 for PEM.
In contrast to Fig. 2.3, this graph presents values of permeate flux averaged over the duration of
filtration experiment rather than over the first minute.

Under “high J p ; high J cf ” conditions where J cf was in the 20,400  3,600
2

L/(min·m ) range, fouling decreased the concentration rate; for the CS-blocked membrane, the
2

initial J p (averaged over the first minute of filtration) was 5.8  0.3 L/(min·m ), whereas the
2

J p averaged over the length of the filtration was 4.9  0.3 L/(min·m ). Similarly, for PEM-

coated membranes, under the fastest concentration conditions the initial J p was 7.2  1.4
74

2

2

L/(min·m ) and the average J p was 3.9  0.6 L/(min·m ). Thus the concentration rates were
similar for CS-blocked and PEM-blocked membranes, but the PEM may show somewhat more
fouling due to a higher initial flux.

With polysulfone membranes of similar MWCO (30 kDa) blocked with bovine serum,
2

Hill et al. [2, 3, 11] obtained permeate flow rates between 0.4 and 0.6 L/(min·m ) with tap water
samples at a transmembrane pressure of ~ 13 psi. Morales-Morales et al.[5] reported a J p of
2

1.5 L/(min·m ) at a transmembrane pressure of 12 psi for surface and tap water samples with a
50 kDa polyacrylonitrile membrane also blocked with CS. By increasing both the
transmembrane pressure and cross-flow rate we achieve significantly higher (~ 3-fold) rates of
concentration. The comparison is not direct, however, because of different feed water matrices
2

that were spiked by viruses and different concentration ratios: 4.9 mL of feed per cm of
2

membrane surface for surface/tap water in the study by Morales-Morales et al. [5], 5.5 mL/cm
2

for tap water in the study by Hill et al. [2] versus 18.3 mL/cm for DI water in this study.

2.3.5.2 Effect of J cf / J pon virus recovery

P22 pre-elution recovery by CS-blocked membranes did not significantly increase with
an increase in Jcf / J p (Fig. 2.7). Others looked briefly at the dependence of virus recovery on
hydraulic conditions [8]. Belfort et al. noticed no significant difference in poliovirus recovery by
filtration through a non-blocked cellulose acetate membrane when tripling the crossflow rate; the
75

observation, however, was based on only one filtration performed at a lower crossflow rate [8].
More recently, Polaczyk et al. [6] observed no statistical difference in recoveries of MS2 phage
and echovirus 1 when operating a membrane blocked with NaPP at two different permeate flow
rates; however, in their study, the permeate flow rate and crossflow rate were linked, and as a
consequence the Jcf / J p ratio changed only from 5,800 to 6,500.

In contrast, the pre-elution recovery of P22 with PEM-coated membranes significantly
increased with an increase in Jcf / J p . The recovery is 55 ± 15% for the fastest concentration
conditions (“high J p ; high J cf ” domain) and 71  11% for the highest recovery conditions
(“low J p ; high J cf ” domain). This effect of Jcf / J p on virus recovery is expected. At
higher J p , the force driving the virus toward the membrane is higher. In addition at higher
Jcf / J p there is more fouling as seen in Fig. 2.10. Fouling can conceal, at least partially, the

membrane surface and its anti-adhesive properties which decreased pre-elution recovery as seen
in Fig 2.12. Such high recoveries may render the elution step unnecessary. Indeed this level of
recovery is comparable to what might be obtained by other virus concentration methods that
involve elution. Eliminating the need for elution has several advantages. Time is gained and the
volume of the concentrated sample is smaller. In addition, all viruses are recovered pre-elution.
These viruses interacted less with the membrane surface than viruses recovered by elution and as
a consequence their integrity is better maintained [62]. Post-elution recovery for both types of
membrane is not affected by Jcf / J p .

76

P22 pre-elution with CS-blocked membrane

100

P22 pre-elution with PEM-coated membrane

90
recovery pre-elution, %

80
70
60

y = -17.432x + 78.831
R² = 0.4028

50
40
30
20

y = -20.37x + 37.902
R² = 0.1821

10
0
0.0

0.5

1.0

1.5

2.0

2.5

fouling resistance over membrane resistance ratio
Figure 2.12: Pre-elution recovery of P22 from DI water as a function of the fouling layer
resistance to membrane resistance ratio.

2.3.6

Potential for practical applications

Preparation of membranes with a large surface area is critical for successfully scaling up
the proposed sample concentration approach [63, 64]. PEMs have already been deposited on flat
sheet membranes [65], tubular membranes [66], and hollow fibers [67]. The proposed approach
could be scaled-up by depositing a PEM using alternating circulation of cationic and anionic
polyelectrolyte solutions through a standard hollow fiber cartridge [2, 3, 5, 6, 11]. Elution, backflushing, and coating removal and regeneration could occur in situ by flushing the cartridge with
appropriate solutions. Other components of a scaled-up portable system would include a back77

pressure valve for generating and controlling the transmembrane pressure and a bypass valve for
regulating the crossflow rate. When appropriate, the proposed PEM-based approach can be
combined with the addition of chemical dispersant, such as NaPP, as described in several earlier
studies [2, 3, 5, 6, 11] to further reduce virus attachment to the membrane surface and improve
recovery.

Acknowledgements

This work was supported by the NSF Partnerships for International Education and Research
grant OISE-0530174, a U.S. EPA Science to Achieve Results grant R833010, and a MSU
Foundation Strategic Partnership grant. The authors would also like to acknowledge Mr. Hang
Shi for his help with  -potential measurements and Mr. Tyler Rafferty for his help with
filtration tests.

78

REFERENCES

79

REFERENCES

1.

Gibson, K.E. and K.J. Schwab, Detection of bacterial indicators and human and bovine
enteric viruses in surface water and groundwater sources potentially impacted by animal
and human wastes in Lower Yakima Valley, Washington. Applied And Environmental
Microbiology, 2011. 77(1): p. 355-362.

2.

Hill, V.R., et al., Multistate evaluation of an ultrafiltration-based procedure for
simultaneous recovery of enteric microbes in 100-liter tap water samples. Applied and
Environmental Microbiology, 2007. 73(19): p. 6327-6327.

3.

Hill, V.R., et al., Development of a rapid method for simultaneous recovery of diverse
microbes in drinking water by ultrafiltration with sodium polyphosphate and surfactants.
Applied and Environmental Microbiology, 2005. 71(11): p. 6878-6884.

4.

Liu, P.B., et al., Hollow-fiber ultrafiltration for simultaneous recovery of viruses,
bacteria and parasites from reclaimed water. Journal Of Microbiological Methods, 2012.
88(1): p. 155-161.

5.

Morales-Morales, H.A., et al., Optimization of a reusable hollow-fiber ultrafilter for
simultaneous concentration of enteric bacteria, protozoa, and viruses from water.
Applied and Environmental Microbiology, 2003. 69(7): p. 4098-4102.

6.

Polaczyk, A.L., et al., Ultrafiltration-based techniques for rapid and simultaneous
concentration of multiple microbe classes from 100-L tap water samples. Journal Of
Microbiological Methods, 2008. 73(2): p. 92-99.

7.

Albinana-Gimenez, N., et al., Comparison of methods for concentrating human
adenoviruses, polyomavirus JC and noroviruses in source waters and drinking water
using quantitative PCR. Journal of Virological Methods, 2009. 158(1-2): p. 104-109.

8.

Belfort, G., Y. Rotem, and E. Katzenelson, Virus concentration using hollow fiber
membranes. Water Res., 1975. 9(1): p. 79-85.

9.

Belfort, G., Y. Rotem, and E. Katzenelson, Virus concentration using hollow fiber
membranes .2. Water Res., 1976. 10(4): p. 279-284.

10.

Berman, D., M.E. Rohr, and R.S. Safferman, Concentration of poliovirus in water by
molecular filtration. Appl. Environ. Microbiol., 1980. 40(2): p. 426-428.

11.

Hill, V.R., et al., Comparison of hollow-fiber ultrafiltration to the USEPA VIRADEL
technique and USEPA Method 1623. J. Environ. Qual., 2009. 38(2): p. 822-825.

80

12.

Rajal, V.B., et al., Validation of hollow fiber ultrafiltration and real-time PCR using
bacteriophage PP7 as surrogate for the quantification of viruses from water samples.
Water Res., 2007. 41(7): p. 1411-1422.

13.

Winona, L.J., et al., Efficient and predictable recovery of viruses from water by small
scale ultrafiltration systems. Canadian Journal of Microbiology, 2001. 47(11): p. 10331041.

14.

Straub, T.M. and D.P. Chandler, Towards a unified system for detecting waterborne
pathogens. Journal of Microbiological Methods, 2003. 53(2): p. 185-197.

15.

Sablani, S.S., et al., Concentration polarization in ultrafiltration and reverse osmosis: a
critical review. Desalination, 2001. 141(3): p. 269-289.

16.

USEPA, Manual of methods for virology, USEPA, Editor 2001: Washington D.C.

17.

Borrego, J.J., et al., Development and application of new positively charged filters for
recovery of bacteriophages from water. Applied And Environmental Microbiology, 1991.
57(4): p. 1218-1222.

18.

Haramoto, E., et al., Recovery of human norovirus from water by virus concentration
methods. Journal of Virological Methods, 2009. 160(1-2): p. 206-209.

19.

Katayama, H., A. Shimasaki, and S. Ohgaki, Development of a virus concentration
method and its application to detection of enterovirus and Norwalk virus from coastal
seawater. Applied And Environmental Microbiology, 2002. 68(3): p. 1033-1039.

20.

Schaudies, P. and D.A. Robinson, Literature review of molecular methods for
simulateneous detection of pathogens in water, USEPA, Editor 2007: Washington D.C.

21.

Gibbons, C.D., et al., Evaluation of positively charged alumina nanofibre cartridge filters
for the primary concentration of noroviruses, adenoviruses and male-specific coliphages
from seawater. Journal of Applied Microbiology, 2010. 109(2): p. 635-641.

22.

Karim, M.R., et al., New electropositive filter for concentrating enteroviruses and
noroviruses from large volumes of water. Applied and Environmental Microbiology,
2009. 75(8): p. 2393-2399.

23.

Di Pasquale, S., et al., Comparison of different concentration methods for the detection of
hepatitis A virus and calicivirus from bottled natural mineral waters. Journal Of
Virological Methods, 2010. 165(1): p. 57-63.

24.

Fumian, T.M., et al., Detection of rotavirus A in sewage samples using multiplex qPCR
and an evaluation of the ultracentrifugation and adsorption-elution methods for virus
concentration. Journal Of Virological Methods, 2010. 170(1-2): p. 42-46.

81

25.

Tsai, U.L., et al., Simple method of concentrating enteroviruses and hepatitis A virus
from sewage and ocean water for rapid detection by reverse transcriptase-polymerase
chain reaction. Applied And Environmental Microbiology, 1993. 59(10): p. 3488-3491.

26.

Liu, J.Y., Q.P. Wu, and X.X. Kou, Development of a virus concentration method and its
application for the detection of noroviruses in drinking water in China. Journal Of
Microbiology, 2007. 45(1): p. 48-52.

27.

Gibson, K.E. and K.J. Schwab, Tangential-flow ultrafiltration with integrated inhibition
detection for recovery of surrogates and human pathogens from large-volume source
water and finished drinking water. Applied and Environmental Microbiology, 2011.
77(1): p. 385-391.

28.

Kim, H., H. Park, and G. Ko, Hollow-fiber ultrafiltration for the concentration and
simultaneous recovery of multiple pathogens in contaminated foods. Journal of Food
Protection, 2009. 72(12): p. 2547-2552.

29.

Bales, R.C., et al., MS-2 and poliovirus transport in porous media: Hydrophobic effects
and chemical perturbations. Water Resour. Res., 1993. 29(4): p. 957-963.

30.

Farrah, S.R., et al., Use of modified diatomaceous-earth for removal and recovery of
viruses in water. Applied And Environmental Microbiology, 1991. 57(9): p. 2502-2506.

31.

Decher, G., Fuzzy nanoassemblies: Toward layered polymeric multicomposites. Science,
1997. 277(5330): p. 1232-1237.

32.

Mendelsohn, J.D., et al., Rational design of cytophilic and cytophobic polyelectrolyte
multilayer thin films. Biomacromolecules, 2003. 4(1): p. 96-106.

33.

Salloum, D.S. and J.B. Schlenoff, Protein adsorption modalities on polyelectrolyte
multilayers. Biomacromolecules, 2004. 5(3): p. 1089-1096.

34.

Richert, L., et al., Layer by layer buildup of polysaccharide films: Physical chemistry and
cellular adhesion aspects. Langmuir, 2004. 20(2): p. 448-458.

35.

Fu, J.H., et al., Construction of anti-adhesive and antibacterial multilayer films via layerby-layer assembly of heparin and chitosan. Biomaterials, 2005. 26(33): p. 6684-6692.

36.

Lajimi, R.H., et al., Effect of LbL surface modification on characteristics and
performances of cellulose acetate nanofiltration membranes. Desalination, 2011. 266(13): p. 78-86.

37.

Muller, M., et al., Polyelectrolyte multilayers and their interactions. Journal of Adhesion,
2004. 80(6): p. 521- 547.

82

38.

Yoo, P.J., et al., Spontaneous assembly of viruses on multilayered polymer surfaces.
Nature Materials, 2006. 5(3): p. 234-240.

39.

Yoo, P.J., et al., Solvent-assisted patterning of polyelectrolyte multilayers and selective
deposition of virus assemblies. Nano Letters, 2008. 8(4): p. 1081-1089.

40.

Gerba, C.P., et al., Waterborne rotavirus: A risk assessment. Water Res., 1996. 30(12): p.
2929-2940.

41.

Teschke, C.M. and K.N. Parent, "Let the phage do the work": Using the phage P22 coat
protein structures as a framework to understand its folding and assembly mutants.
Virology, 2010. 401(2): p. 119-130.

42.

Parent, K.N., et al., P22 coat protein structures reveal a novel mechanism for capsid
maturation: Stability without auxiliary proteins or chemical crosslinks. Structure, 2010.
18(3): p. 390-401.

43.

Parent, K.N., et al., Cryo-reconstructions of P22 polyheads suggest that phage assembly
is nucleated by trimeric interactions among coat proteins. Physical biology, 2010. 7(4):
p. 045004.

44.

Cortines, J.R., et al., Decoding bacteriophage P22 assembly: Identification of two
charged residues in scaffolding protein responsible for coat protein interaction.
Virology, 2011. 421(1): p. 1-11.

45.

Masago, Y., T. Shibata, and J.B. Rose, Bacteriophage P22 and Staphylococcus aureus
attenuation on nonporous fomites as determined by plate assay and quantitative PCR.
Applied and Environmental Microbiology, 2008. 74(18): p. 5838-5840.

46.

Shen, C., et al., Evaluating bacteriophage P22 as a tracer in a complex surface water
system: The Grand River, Michigan. Environmental Science & Technology, 2008. 42(7):
p. 2426-2431.

47.

Ackermann, H.W., Tailed bacteriophages: The order Caudovirales, in Advances in Virus
Research, Vol 511999, Academic Press Inc: San Diego. p. 135-201.

48.

Mylon, S.E., et al., Influence of salts and natural organic matter on the stability of
bacteriophage MS2. Langmuir, 2010. 26(2): p. 1035-1042.

49.

Xagoraraki, I., et al., Occurrence of human adenoviruses at two recreational beaches of
the great lakes. Applied and Environmental Microbiology, 2007. 73: p. 7874-7881.

50.

Wang, F.L. and V.V. Tarabara, Pore blocking mechanisms during early stages of
membrane fouling by colloids. Journal of Colloid and Interface Science, 2008. 328(2): p.
464-469.

83

51.

van Oss, C.J., Acid-base interfacial interactions in aqueous media. Colloids Surf. A,
1993. 78: p. 1-49.

52.

Brant, J.A. and A.E. Childress, Assessing short-range membrane-colloid interactions
using surface energetics. Journal of Membrane Science, 2002. 203(1-2): p. 257-273.

53.

van Oss, C.J., Interfacial forces in aqeuous media - second edition2006: Taylor &
Francis Group.

54.

Oshima, K.H., Efficient and predictable recovery of viruses and Cryptosporidium parvum
oocysts from water by ultrafiltration system, in New Mexico: Water Resources Research
Institute. Technical Completion Report, W.R.R. Institute, Editor 2001, Water Resources
Research Institute.

55.

Brorson, K., et al., Characterization and purification of bacteriophages using
chromatofocusing. Journal of Chromatography A, 2008. 1207(1–2): p. 110-121.

56.

Murray, J.P. and G.A. Parks, Poliovirus adsorption on oxide surfaces, in Particulates in
Water1980, American Chemical Society. p. 97-133.

57.

van Oss, C.J. and R.F. Giese, The hydrophilicity and hydrophobicity of clay-minerals.
Clays and Clay Minerals, 1995. 43(4): p. 474-477.

58.

Redman, J.A., S.L. Walker, and M. Elimelech, Bacterial adhesion and transport in
porous media: Role of the secondary energy minimum. Environmental Science &
Technology, 2004. 38(6): p. 1777-1785.

59.

Herath, G., K. Yamamoto, and T. Urase, The effect of suction velocity on concentration
polarization in microfiltration membranes under turbulent flow conditions. Journal of
Membrane Science, 2000. 169(2): p. 175-183.

60.

Wiesner, M.R. and S. Chellam, Mass-transport considerations for pressure-driven
membrane processes. Journal American Water Works Association, 1992. 84(1): p. 88-95.

61.

Shan, W.Q., et al., Polyelectrolyte multilayer films as backflushable nanofiltration
membranes with tunable hydrophilicity and surface charge. Journal of Membrane
Science, 2010. 349(1-2): p. 268-278.

62.

Furiga, A., et al., Effects of ionic strength on bacteriophage MS2 behavior and their
implications for the assessment of virus retention by ultrafiltration membranes. Applied
and Environmental Microbiology, 2011. 77(1): p. 229-236.

63.

Ahmadiannamini, P., et al., Influence of polyanion type and cationic counter ion on the
SRNF performance of polyelectrolyte membranes. Journal of Membrane Science, 2012.
403-404: p. 216-226.

84

64.

Bruening, M. and D. Dotzauer, Polymer films: Just spray it. Nat Mater, 2009. 8(6): p.
449-450.

65.

Jin, W., A. Toutianoush, and B. Tieke, Use of polyelectrolyte layer-by-layer assemblies
as nanofiltration and reverse osmosis membranes. Langmuir, 2003. 19: p. 2550-2553.

66.

Dotzauer, D.M., et al., Catalytic membranes prepared using layer-by-layer adsorption of
polyelectrolyte/metal nanoparticle films in porous supports. Nano Lett., 2006. 6(10): p.
2268-2272.

67.

Lu, O.Y., et al., Catalytic hollow fiber membranes prepared using layer-by-layer
adsorption of polyelectrolytes and metal nanoparticles. Catalysis Today, 2010. 156(3-4):
p. 100-106.

85

Chapter 3. Effect of flow conditions and water composition on the
removal and recovery of infective bacteriophage by crossflow
filtration with CS-blocked membranes
3.1

Introduction

As mentioned in Chapter 1, efficient virus detection in drinking and recreational water is
vital for preventing disease outbreaks and this requires an efficient concentration method. The
ideal concentration/quantification method should be able to differentiate infectious viruses from
non-infectious ones [1] as only the former are a concern in terms of public health, and
quantification of viral nucleic acid is not necessarily a good indicators of the presence of
infectious virus [2]. It is thus important that the concentration method used does not affect virus
infectivity.

Crossflow filtration is emerging as a concentration method for the collection, concentration
and quantification of viruses in large volumes of water [3-9], but has the potential to disrupt
virus infectivity. Indeed, a decrease in infective virus titer in the retentate of crossflow filtration
units has been observed in several studies [9-12] but others did not observe such trends [13]. In
two studies, virus aggregation, which affects virus infectivity assays, was initially considered a
culprit but eventually ruled out as the reason for the observed loss of infectivity [11, 12]. In
some studies increasing ionic strength was found to correlate with virus loss of infectivity in tap
water [11, 12], while others, using more complex water samples did not observe any correlation
between sample conductivity (an approximate measure of ionic strength) and loss of infectivity

86

[9]. Virus morphology, the presence of hydrophobic surfaces in the concentration device, and
the pumping-induced shear stress, have also been mentioned as possible factors contributing to
the loss of viral infectivity during crossflow filtration [10, 11]. Another factor that might
influence infectivity assays, and thus recovery of infective viruses by crossflow filtration, is the
presence of natural organic matter or cations which have been shown to affect virus aggregation
[14]. Finally, changes in filtration conditions such as crossflow rate and permeate flow rate
translate into changes in the forces applied to the viruses and, therefore, might affect virus
infectivity, virus aggregation, or both.

In Chapter 2, we showed that it was possible to increase the initial permeate flux through
CS-blocked membranes by up to 4 times (over the values reported in the literature) without
affecting P22 recovery as determined by qPCR. In this Chapter we will look at the effect of
filtration conditions and water matrix on the removal and the pre- and post-elution recoveries of
infective bacteriophage P22 by CS-blocked membranes. In Chapter 2 we also showed that
increasing Jcf / J p increased P22 pre-elution recovery by PEM-coated membranes, whereas for
CS-blocked membrane, increasing Jcf / J p had no effect on P22 recovery or removal. This was
contrary to our expectation. Indeed decreasing J p should decrease concentration polarization
and lead to less particle adhesion to the membrane [15]. In addition, increasing the crossflow
rate can decrease concentration polarization [15] and has been shown to decrease passage of
virus through membrane [15, 16]; however, this effect of crossflow might be minimal for viruses
as their size corresponds to a minimum in Brownian and shear-induced particle back-transport
from the membrane surface [17]. It is possible that the stable P22 removal at increasing
Jcf / J p is a result of two mutually counterbalancing effects: 1) a partial loss of integrity of the

87

CS layer at high crossflow and 2) a decrease in the net flux of viruses towards the membrane
surface at higher Jcf / J p . To determine which scenario is operative we assessed the integrity
of the CS layer under various filtration conditions.

3.2

3.2.1

Material and Methods

Reagents

Lysozyme (from chicken egg white), ethylenediaminetetraacetic acid (EDTA), calf
serum, sodium polyphosphate (NaPP), and dextran standard (25 kDa) were purchased from
Sigma-Aldrich. Tween 80 was purchased from Fisher and tryptic soy broth from Becton,
Dickinson.

3.2.2

Water samples

Deionized (DI) water was supplied by a commercial ultrapure water system equipped
with a terminal 0.2 µm capsule microfilter (PolyCap, Whatman Plc., Sanford, ME). Ten liters
surface water samples were collected from Lake Lansing (Lansing, MI) as grab samples in May
and September 2009 and in February 2010, and filtered through a 0.45 μm membrane on the day
of collection. Samples were kept at 4°C until used. Surface water samples were used as early as
the day following their collection and within 4 months of their collection. Before use, the
surface water container was agitated to resuspend any settled particles.

88

The total organic carbon (TOC) content of filtered Lake Lansing samples was determined
using a Total Organic Carbon analyzer (1010, O I Analytical, USA). Each sample was analyzed
in triplicate. Calibration standards and blanks were run for each set of measurements.

3.2.3

Bacteriophage

Bacteriophages were cultured as described in Chapter 2, section 2.2.2.

3.2.4

Membrane preparation, crossflow concentration apparatus and concentration procedure

The CS-blocked membranes were prepared as described in Chapter 2 section 2.2.5.
Membranes were first compacted for 90 min at 40 psi with DI water. The filtration apparatus
and concentration procedure are described in Chapter 2 sections 2.2.4 and 2.2.7. Elution was
performed with non-spiked surface water amended with 0.01% Tween 80 and 0.01% NaPP. The
filtration conditions used in this study are shown in Fig. 3.1.

89

initial permeate fllux Jp, L/(min·m2)

10

DI water

9

surface water

earlier studies
high Jp

high Jp

8

high Jcf

low Jcf

fastest
concentration
conditions

7
6
5
4
3
(d)

2
(b, c)
(a)

1
0
0.0

0.5

low Jp

high Jcf

low Jcf
1.0

low Jp

highest recovery conditions
1.5

2.0

2.5

3.0

crossflow flux Jcf ·104, L/(min·m2)
Filtration conditions in experiments on infective P22 concentration and recovery
from DI water and surface water using CS-blocked membranes. Number of tests: n=23 for DI
water, n=18 for surface water. Earlier studies: a: Hill et al. [18], b: Hill et al. [19], c: Hill et al.
[5], d: Morales-Morales et al. [20].

3.2.5

Sample analysis

The P22 concentration in each sample was determined by plaque assay using the double
agar layer procedure within 24 h of the filtration experiment. Ten-fold serial dilutions of each
sample were prepared. The bacterial host, S. typhimurium, was grown to log phase in tryptic soy
broth and added to a 1% soft agar solution, which was inoculated with 1 mL dilution aliquots and
poured on top of an agar plate. Plates were inoculated overnight at 37ºC. Plaques (clear zone of
90

infected bacterial cells) were then counted and this count was used in combination with the
dilution factor to determine P22 concentration in plaque forming unit (PFU)/mL in the sample.
Plaque assays were run in triplicate for each dilution of each sample. Samples from 3 filtrations
corresponding to 3 different filtration conditions (low J p - low Jcf ; low J p - high Jcf and
high J p - low Jcf ) were also stored at -80ºC in presence of glycerol (10%) to be later analyzed
by qPCR. The procedure for qPCR analysis is described in Chapter 2 section 2.2.8.

P22 log removal (LRV) percentage of pre-elution recovery ( r pre ), and percentage of
post-elution recovery ( r post ) were calculated using the equations (1), (2) and (3) from Chapter
2. Data were compared with the Student t-test. P < 0.05 was considered to show statistical
significance. A Bonferroni correction was performed when multiple comparison were done.

3.2.6

Calf serum layer stability test

The stability of the calf serum layer during filtration was tested in the filtration unit
described in section 2.2.4. CS-blocked membranes were prepared and compacted as described in
section 2.2.5. The membranes were then used to filter a 20 mg/L aqueous solution of 25 kDa
dextran using a transmembrane pressure of approximately 40 psi. The crossflow rates was
2

2

increased from 1,500 L/(min m ) to 22,000 L/(min m ), which corresponds to Jcf / J p ratios
from 900 to 8,900. Each crossflow rate was maintained for 5 min. and permeate samples were
collected before and after the crossflow rate was increased. Retentate samples were also
collected several times during the test to monitor dextran concentration. Concentrations of
91

dextran in the retentate and in the permeate were determined using the total organic carbon
analyzer. Each sample was analyzed in triplicate. Calibration standards and blanks were run for
each set of measurements. Rejection ( R ) of dextran at a given time was determined using
equation (3.1):
(3.1)

 C p (t ) 
  100%
R(t )  1 

C r (t ) 



where C p and C r are, respectively, dextran concentrations in the permeate and retentate at that
given time. C r at any given time was determined using linear interpolation of retentate
concentration values measured at several times during each experiment

The stability of the calf serum layer during elution was also tested. The test was similar to
the one described above except that pressure was only applied to collect samples. CS layer
stability during filtration was tested in triplicate, only one test was performed to determine the
CS layer stability during elution.

3.3

3.3.1

Results and Discussion

Effect of filtration conditions on the recovery and removal of infective P22

A preliminary test was performed to determine if infective viruses were lost due to the
adhesion to the components of the filtration unit, other than the membrane or due to loss of
6

infectivity. We showed that circulating 1 L of DI water spiked with P22 (1.3·10 PFU/mL) for 3
h in the filtration unit, with no membrane in place and no pressure applied, did not affect P22
92

counts in the infectivity assay indicating that there was no loss of infectivity due to contact with
components of the concentration unit as some have observed [11] or that this loss of infectivity
was coincidentally offset by virus disaggregation.

Figure 3.2 shows the recovery and removal of infective P22 from DI water and surface
water by CS-coated membrane under various filtration conditions. At the pre-elution level, there
was no effect of the filtration conditions on the recovery of infective P22 from either DI water or
surface water. At the post-elution level, although the recovery of infective P22 from DI water in
the least favorable conditions (high J p , low Jcf ) seemed lower on average than under other
conditions, a t-test analysis showed that the difference was not significant. In the case of surface
water, again no effect of filtration conditions on infective P22 recovery is observed.

Recovery of infective P22 from DI water or surface water was not affected by filtration
conditions. As mentioned in Chapter 2, a similar trend has been observed by others [21, 22],
however the scope of their studies was limited in terms of the number of filtrations performed
[21] or range of Jcf / J p ratios explored [21]. The trend is counterintuitive as it was expected
that increasing Jcf / J p would decrease concentration polarization and fouling (Fig. 3.3), which
should increase pre-elution recovery as virions have a lesser chance to be trapped near the
membrane.

93

pre-elution recovery, %

a)

100

DI water

surface water

80
60
40
20
0

filtration conditions
b)
post-elution recovery, %

160

DI water

surface water
*

120
80
40
0

filtration conditions

94

c)

DI water

10

*

8
log removal

surface water

6
4
2
0

filtration conditions

95

DI water

surface water

Jp final/ Jp initial

1.0
0.8
0.6
0.4
0.2
0.0

filtration conditions

For microfiltration (MF) membranes it was shown that increasing the Jcf / J p ratio
leads to an increase in the removal of MS2 and Q phages [16] and that higher J p and stirring
increase removal of poliovirus [23]. In the present study, however, changing filtration conditions
had no effect on P22 removal from DI water or surface water. In the case of surface water, the
formation of a thicker fouling layer preventing virus passage at lower Jcf / J p might have
opposed the effect of increased concentration polarization and helped maintaining removal levels
as high as at higher J cf / J p . Others have observed

96

3.3.2

qPCR vs. plaque assay

As reported in Chapter 2, no significant effect of Jcf / J p on P22 pre and post-elution
recoveries or removal (as determined by qPCR) by CS-blocked membrane was observed. In this
Chapter we report the absence of such trends observed for infective P22 as well. In addition, the
average removal and pre-elution recoveries of infective P22 observed in this study were not
significantly different than the average removal and recoveries based on qPCR analysis as
reported in Chapter 2 for CS-blocked-membranes; however, post-elution recovery of infective
virus was significantly lower than recovery measured by qPCR (74 ± 31% and 99 ± 11%
respectively).

Data from the three filtrations for which samples were analyzed by both qPCR and
plaque assay, confirm this observation (Fig. 3.4). Similar elution and removal recovery were
observed, except at the post-elution level where a significant difference was detected at high J p ,
low Jcf . If this observation is confirmed to be statistically significant, it would indicate that,
under high J p , low Jcf , viruses recovered during elution have either lost infectivity (for
example due to interaction with the membrane during the filtration step or because of the
additive -Tween 80 and NaPP- used for elution) or have aggregated. Since NaPP is often used as
a dispersant when concentrating various pathogens including viruses [20], the decrease in viral
count observed by plaque assay was most probably due to the loss of P22 infectivity rather than
97

aggregation of the phage. This could be confirmed by measuring the particle size in the retentate
and eluate; however, particle size analysis requires higher virus titer than the one that we had in
our filtrations and concentrating those samples was not an option as any concentration could
affect the aggregation state of the virus suspension.

Another difference between results obtained by qPCR and plaque assay is the
reproducibility of virus recovery and removal data. Standard deviations of pre-elution
recoveries, post-elution recoveries and LRV values obtained by qPCR (8%, 11% and 1.2
respectively) were lower than standard deviations obtained in plaque assays (17%, 31% and 1.4
LRV respectively). This might be explained by a greater standard deviation associated with the
plaque assay method compared to the qPCR method (as seen in Fig. 3.4), which is due, in part, to
the fact that plaque assay was performed on 3 different grab samples while qPCR was performed
in triplicate on one grab sample.

98

100
pre-elution recovery, %

a)

plaque assay

80
60
40
20
0
high Jp, low Jcf

low Jp, low Jcf

low Jp, high Jcf

filtration conditions

post-elution recovery, %

b)
100

plaque assay

80
60
40
20
0
high Jp, low Jcf

low Jp, low Jcf

low Jp, high Jcf

filtration conditions

Effect of analysis method on P22 a) pre-elution recovery, b) post-elution
recovery, c) removal, by crossflow filtration with CS-blocked membrane. One filtration
was run for each set of conditions. The error bar indicates the standard deviation
associated with the experimental method.

99

Figure 3.4 (cont’d)
c)

5
plaque assay
log removal

4
3
2
1
0
high Jp, low Jcf

3.3.3

low Jp, low Jcf
filtration conditions

low Jp, high Jcf

DI water vs. surface water

As Fig. 3.2 demonstrates, at the pre-elution level there was no significant difference
between in the recovery of infective P22 from surface water and from DI water. At the postelution level, average recovery from surface water was consistently higher than recovery from DI
water, however, the difference is not statistically significant. We did not expect post-elution
recovery from surface water to be higher than post-elution recovery from DI water, especially
under high J p - low Jcf . In these conditions, virus can get trapped in the fouling layer, which
is more important for surface water than for DI water (Fig 3.3).

100

3.3.4

Effect of filtration condition on CS layer integrity

To explain the absence of influence of Jcf / J p on the recovery and removal of infective
P22 and P22 DNA, we hypothesized that the calf serum layer was disintegrating at higher
Jcf / J p values thus offsetting the benefit of lower concentration polarization and fouling.

However, as seen in Fig. 3.5, 25 kDa dextran rejection by CS-blocked membrane was not
affected by filtration conditions, indicating that the layer integrity is maintained. The test was
performed at 40 psi and, as a consequence, Jcf / J p values higher than 8,000 were not tested
(although those conditions were used in P22 filtration tests).

101

a)
100
90
dextran rejection, %

80
70
60
50
40
30
20

immediately after Jcf increase
5 min after Jcf increase

10
0
0.0

b)

0.2
0.4
0.6
0.8
crossflow to permeate flux ratio Jcf/Jp 104

1.0

100
90

dextran rejection, %

80
70
60
50
40
30
20
immediately after Jcf increase
5 min after Jcf increase

10
0
0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

crossflow flux Jcf 104, L/min.m2

Dextran rejection by CS-blocked membrane under increasing Jcf / J p during a)
filtration and b) elution. n=3 for filtration, error bar indicates standard deviation; n=1 for elution.

102

During elution, dextran rejection decreased as Jcf increased, indicating that the layer
integrity was compromised. (Additional testing should be performed to confirm this trend as only
one elution test was performed). We do not expect this phenomenon, if confirmed, to affect P22
recovery. On the contrary, stripping the layer off the membrane may be an effective approach for
recovering those viruses that penetrated that layer during filtration. It is possible that when
submitted to the pressure of 20 psi, the calf serum layer also loses integrity, which would balance
the effect of decreasing the permeate flux in terms of virus adhesion to the membrane. In
addition, increasing crossflow only has a minimal effect on decreasing concentration polarization
of P22 [17]. Together, these phenomena could explain the lack of improvement in P22 preelution recovery or removal at increasing Jcf / J p However the calf serum layer stability at 20
psi should be assessed to confirm this hypothesis.

3.3.5

Conclusions

Permeate flux and crossflow rate were found to have no statistically significant effect on
the recovery and removal of infective P22. The average permeate flux could be increased up to
5-fold for DI water and 2-fold for surface water over values reported in the literature without
affecting the recovery or removal of infective P22. However, virus infectivity and/or
aggregation did seem to be affected by crossflow filtration. Specifically, under high J p - low
Jcf conditions, viruses recovered at the post-elution stage seemed to have lower infectivity or

be aggregated, which might be due to interactions with the membrane surface or the presence of
additive at the elution stage.

103

REFERENCES

104

REFERENCES

1.

Straub, T.M. and D.P. Chandler, Towards a unified system for detecting waterborne
pathogens. Journal of Microbiological Methods, 2003. 53(2): p. 185-197.

2.

Richards, G.P., Limitations of molecular biological techniques for assessing the
virological safety of foods. Journal of Food Protection, 1999. 62(6): p. 691-697.

3.

Winona, L.J., et al., Efficient and predictable recovery of viruses from water by small
scale ultrafiltration systems. Canadian Journal of Microbiology, 2001. 47(11): p. 10331041.

4.

Grassi, T., et al., Rotavirus detection in environmental water samples by tangential flow
ultrafiltration and RT-nested PCR. Environmental Monitoring and Assessment, 2010.
164(1-4): p. 199-205.

5.

Hill, V.R., et al., Comparison of hollow-fiber ultrafiltration to the USEPA VIRADEL
technique and USEPA Method 1623. J. Environ. Qual., 2009. 38(2): p. 822-825.

6.

Olszewski, J., L. Winona, and K.H. Oshima, Comparison of 2 ultrafiltration systems for
the concentration of seeded viruses from environmnetal waters. Can. J. Microbiol., 2005.
51(4): p. 295-303.

7.

Oshima, K.H., et al., The removal of phages T1 and Pp7, and poliovirus from fluids with
hollow-fiber ultrafilters with molecular-weight cutoffs of 50000, 13000, and 6000. Can. J.
Microbiol., 1995. 41(4-5): p. 316-322.

8.

Soule, H., et al., Ultrafiltration and reverse transcription-polymerase chain reaction: An
efficient process for poliovirus, rotavirus and hepatitis A virus detection in water. Water
Res., 2000. 34(3): p. 1063-1067.

9.

Rajal, V.B., et al., Validation of hollow fiber ultrafiltration and real-time PCR using
bacteriophage PP7 as surrogate for the quantification of viruses from water samples.
Water Res., 2007. 41(7): p. 1411-1422.

10.

Sorber, C.A., B.P. Sagik, and J.F. Malina, Virus Rejection by Reverse OsmosisUltrafiltration Processes. Water Research, 1972. 6(11): p. 1377.

11.

Furiga, A., et al., Effects of ionic strength on bacteriophage MS2 behavior and their
implications for the assessment of virus retention by ultrafiltration membranes. Applied
and Environmental Microbiology, 2011. 77(1): p. 229-236.

12.

Pierre, G.l., et al., Protocol for the assessment of viral retention capability of membranes.
Journal of Membrane Science, 2011. 381(1-2): p. 41-49.
105

13.

Michalsky, R., et al., Effects of temperature and shear force on infectivity of the
baculovirus Autographa californica M nucleopolyhedrovirus. Journal of Virological
Methods, 2008. 153(2): p. 90-96.

14.

Mylon, S.E., et al., Influence of salts and natural organic matter on the stability of
bacteriophage MS2. Langmuir, 2010. 26(2): p. 1035-1042.

15.

Sablani, S.S., et al., Concentration polarization in ultrafiltration and reverse osmosis: a
critical review. Desalination, 2001. 141(3): p. 269-289.

16.

Herath, G., K. Yamamoto, and T. Urase, The effect of suction velocity on concentration
polarization in microfiltration membranes under turbulent flow conditions. Journal of
Membrane Science, 2000. 169(2): p. 175-183.

17.

Wiesner, M.R. and S. Chellam, Mass-transport considerations for pressure-driven
membrane processes. Journal American Water Works Association, 1992. 84(1): p. 88-95.

18.

Hill, V.R., et al., Development of a rapid method for simultaneous recovery of diverse
microbes in drinking water by ultrafiltration with sodium polyphosphate and surfactants.
Applied and Environmental Microbiology, 2005. 71(11): p. 6878-6884.

19.

Hill, V.R., et al., Multistate evaluation of an ultrafiltration-based procedure for
simultaneous recovery of enteric microbes in 100-liter tap water samples. Applied and
Environmental Microbiology, 2007. 73(19): p. 6327-6327.

20.

Morales-Morales, H.A., et al., Optimization of a reusable hollow-fiber ultrafilter for
simultaneous concentration of enteric bacteria, protozoa, and viruses from water.
Applied and Environmental Microbiology, 2003. 69(7): p. 4098-4102.

21.

Polaczyk, A.L., et al., Ultrafiltration-based techniques for rapid and simultaneous
concentration of multiple microbe classes from 100-L tap water samples. Journal Of
Microbiological Methods, 2008. 73(2): p. 92-99.

22.

Belfort, G., Y. Rotem, and E. Katzenelson, Virus concentration using hollow fiber
membranes. Water Res., 1975. 9(1): p. 79-85.

23.

Madaeni, S.S., A.G. Fane, and G.S. Grohmann, Virus removal from water and
wastewater using membranes. Journal of Membrane Science, 1995. 102: p. 65-75.

24.

Wickramasinghe, S.R., et al., Understanding virus filtration membrane performance.
Journal of Membrane Science, 2010. 365(1-2): p. 160-169.

25.

Jacangelo, J.G., S.S. Adham, and J.-M. Laine, Mechanism of cryptosporidium, giardia,
and MS2 virus removal by MF and UF. American Water Works Association Journal,
1995. 87(9): p. 107-121.

106

Chapter 4. Virus recovery and concentration from complex water
matrices using PEM-coated membranes: Preliminary assessment
4.1

Introduction

In Chapter 2 we showed that PEM-coated membranes show promising results for the
recovery of viruses present in DI water; however, viruses often need to be concentrated from
more complex matrices: tap water, groundwater, surface water as well as water treated for reuse
applications. It has been shown that increasing water complexity can affect virus recovery by
crossflow filtration [1]. For example, Belfort et al. showed that poliovirus recovery decreased
from 77% with DI water to 52% with tap water. In at least one study, however, water
complexity did not affect virus recovery as there was no significant difference in infective MS2
and PRD1 recovery between drinking and surface water [2].

Even for the same type of water, variation in water composition can greatly affect virus
recovery [3, 4]. T1 phage recovery from surface water from 6 different locations varied from
34% to 75% [3], and PP7 recoveries from storm water from 21 different locations varied from
10% to 98 % when assessed by qPCR, and from 25% to 90% when assessed by plaque assay.
However, since filtrations were not replicated it is not possible to conclude whether the
variations were due to differences in the water composition or to an inherent irreproducibility of
the combination of the recovery and detection methods. In a study on virus recovery from
drinking water from 8 different locations, Hill et al. [5] did perform replicate measurements and
found no statistically significant correlations between the 7 parameters tested (pH, turbidity,

107

alkalinity, conductivity, iron concentration, TOC, and DOC) and recoveries of MS2 or ΦX174
phages.

Water composition can affect both concentration and quantification steps. The formation
of a cake layer may prevent virus passage through the membrane [6, 7] or conceal a membrane’s
anti-adhesive properties with respect to a virus. Virus surface properties also depend on the
solution makeup (such as ionic strength, presence of divalent cations and dissolved NOM). In
addition, materials present in the water might affect the detection method [8], especially in the
case of highly concentrated samples. This is why it is important to evaluate the concentration
and detection methods with a wide range of water compositions.

In this Chapter the performance of PEM-coated membrane in recovering P22 from MBR
effluent and surface water is evaluated and compared to the performance of CS-blocked
membrane. The effect of adjusting TOC levels in samples before DNA extraction is assessed so
that we can determine if the observed effects of water complexity on recovery or removal
occurred at the concentration or detection steps. In addition, since the possibility of virion
entrapment at the membrane surface increases with water complexity, the effect of adding a
backflush step after elution on final P22 recovery is studied for surface water samples.

108

4.2

4.2.1

Materials and Methods

Reagents

The reagents used in this chapter were the same as those described in corresponding
sections of previous Chapters.

4.2.2

Water samples

A 25 L membrane bioreactor (MBR) effluent sample was collected from the Traverse
City wastewater treatment plant (Traverse City, MI) as a grab sample in August 2011. A 10 L
surface water sample was collected from Lake Lansing (Lansing, MI) as a grab sample in
September 2011, and filtered through a 0.45 μm membrane on the day of collection. Samples
were kept at 4°C until used. Water samples were used as early as the day following their
collection and within 1 months of their collection. Before use, the surface water container was
agitated to resuspend any settled particles.

4.2.3

Water quality testing.

Total organic carbon (TOC) contents of MBR effluent sample and filtered Lake Lansing
water were determined using a Total Organic Carbon analyzer (1010, O I Analytical, USA). To
determine how filtration affected the TOC content of retentate and permeate samples, filtrations
of MBR effluent and surface water were performed with a PEM-coated membrane and a CS109

blocked membrane in conditions similar to the ones used in virus concentration tests (except for
the filtration of MBR effluent with a CS-blocked membrane which was performed at 40 psi
instead of 20 psi), and TOC analysis was performed on feed, retentate and permeate samples.
Each sample was analyzed at least in triplicate. Calibration standards and blanks were run for
each set of measurements.

4.2.4

Bacteriophage

Bacteriophages were cultured as described in Chapter 2, section 2.2.2.

4.2.5

Membrane preparation, crossflow concentration apparatus and concentration procedure

The PEM-coated membranes and CS-blocked membranes were prepared as described in
Chapter 2 section 2.2.5. The filtration apparatus and concentration procedure are described in
Chapter 2 sections 2.2.4 and 2.2.7. For both types of water, filtrations at high J cf and low J p
(i.e., “highest recovery conditions”) were performed at least in triplicate. For the MBR effluent,
PEM performance at high J cf and high J p was also evaluated. Filtration conditions used in
this study are shown in Fig. 4.1. Elution was performed with 500 mL of permeate amended with
0.01% Tween 80 and 0.01% NaPP. Filtrations with surface water included an additional
backflush step after elution. To backflush the membrane, it was flipped in the filtration cell so
that the membrane skin, coated by PEM or blocked with CS, was on the permeate side. The
backflush was performed by filtering 100 mL of permeate through the membrane at the applied
transmembrane pressure of approximately 40 psi.
110

MBR effluent with CS-blocked
membrane
surface water with CS-blocked
membrane
MBR effluent with PEM-coated
membrane
surface water with PEM-coated
membrane
earlier studies

initial permeate fllux Jp, L/(min·m2)

10
9
8
7
6

high Jp
high Jcf
fastest concentration conditions

5
4
3
(5)

2

(2, 11)
(3)

1

low Jp
high Jcf

highest recovery conditions

0
0.0

0.5

1.0

1.5

2.0

2.5

3.0

crossflow flux Jcf ·104, L/(min·m2)
Figure 4.1: Filtration conditions in experiments on P22 concentration and recovery from MBR
effluent and surface water using CS-blocked membranes and PEM-coated membrane. Number of
tests: n=9 for MBR effluent (3 with CS-blocked membranes, 6 with PEM-coated membranes),
n=7 for surface water (3 with CS-blocked membranes, 4 with PEM-coated membranes).

4.2.6

Sample analysis

P22 concentration in feed, retentate, permeate, eluate and backflush samples were
measured by qPCR as described in Chapter 2 section 2.2.8. For comparison, all samples except
the one from filtration tests performed in “fastest concentration” conditions were also run as
described in Chapter 2 section 2.2.8 with the exception that prior to DNA extraction, retentate
and feed samples were diluted with PCR grade water (Roche) so that TOC levels in those
111

samples matched TOC levels in permeate samples. Values of P22 recovery and removal from
MBR effluent and surface water samples were compared to corresponding values recorded in an
experiment that used DI water and was run in similar filtration conditions (data from Chapter 2).
Data were compared using the Student t-test. P < 0.1 was considered to show statistical
significance. A Bonferroni correction was performed when multiple comparisons were done.

4.3

4.3.1

Results and Discussion

Water quality

Results from TOC analysis are presented in Table 4.1. TOC levels in surface water, 9.0 ±
0.2 ppm, are slightly higher than TOC levels in MBR effluent, 7.3 ± 1.0 ppm. TOC from surface
water seemed to be better recovered than TOC from MBR effluent by both CS-blocked and
PEM-coated membranes. (The experiment should be replicated to determine if the difference is
statistically significant.) When surface water was filtered through CS-blocked membranes, the
combined TOC values of the permeate and the retentate exceeded the TOC values measured in
the feed. This could either be due to calf serum shed by the blocking layer to the retentate
stream, or the result of an experimental error. Additional experiments are necessary to identify
which explanation is valid.

112

Table 4.1: MBR effluent and surface water TOC levels
Type of water

TOC level

TOC

TOC

(ppm)

rejection

recovery

Feed (1L)

8.0

13 %

27 %

Retentate (250 mL)

8.5

Permeate (750 mL)

CS-blocked

Sample (Sample
volume)

MBR effluent

Membrane

7.0
18 %

28 %

5%

41%

19%

36%

PEM-coated Feed (1L)

6.6

Retentate (250 mL)
Permeate (750 mL)
Feed (1L)

8.8
15.5

Permeate (750 mL)

CS-blocked

5.4

Retentate (250 mL)

Surface water

7.2

8.4

PEM-coated Feed (1L)

9.1

Retentate (250 mL)
Permeate (750 mL)

4.3.2

13.7
7.4

P22 recovery and removal from MBR effluent and surface water samples

We expected that the higher TOC levels in the retentate samples would enhance P22
DNA extraction in retentate sample compared to P22 extraction from samples with lower TOC
(feed and permeate). Indeed the presence of organic matter in samples has been shown to
increase P22 extraction by MagNA Pure Compact system hypothetically by blocking adsorption
sites on the surfaces of various system components such as tips or magnetic particles to prevent
113

the loss of P22 DNA due to adsorption (see Appendix B). As a consequence, any observed
increase in recovery could be due to either a more efficient collection/concentration step or only
a more efficient DNA extraction/detection step. For this reason, TOC levels in feed and retentate
were adjusted to the TOC level in the permeate by dilution with PCR grade water. As can be
seen in Fig. 4.2, except for the recovery of P22 from MBR effluent with PEM-coated
membranes, recoveries obtained from TOC-adjusted qPCR analyses were on average lower than
when TOC levels were not adjusted, however, the difference was not significant.

Under the “highest recovery” conditions the pre-elution recovery of P22 by PEM-coated
membranes was lower on average for MBR effluent and surface water than for DI water but the
difference was not statistically significant. Lower pre-elution recovery recoveries may result
from the “masking” of the anti-adhesive properties of the membrane by the organic fouling layer,
from the entrapment of viruses within this fouling layer, or from both of these phenomena (Fig.
4.2). The post-elution recovery by PEM membranes was not significantly affected by the type of
water. Increasing the permeate flux did not significantly affect the pre-elution or post-elution
recovery of P22 from MBR effluent using PEM membranes (Fig. 4.3).

For CS-blocked membrane, pre-elution and post-elution recoveries were not significantly
different for DI water, MBR effluent and surface water.

114

a)

post-elution

pre-elution

TOC adjusted

TOC not adjusted

120

recovery, %

100
80
60
40
20
0
CS

PEM

CS

DI water

PEM
MBR
effluent

CS

PEM

Surface water

Type of membrane and type of water
b)

TOC adjusted

6

TOC not adjusted

5

log removal

4
3
2
1
0
CS

PEM
DI water

CS

PEM
MBR
effluent

CS
PEM
Surface water

Type of membrane and type of water
Figure 4.2: a) P22 recovery and b) removal from DI water (data from Chapter 2), MBR effluent
and surface water with CS-blocked and PEM-coated membranes under “highest recovery”
conditions. Error bars indicate standard deviation.
115

a)

b)
120

post-elution

5

pre-elution
4

80

log removal

recovery, %

100

60
40

2
1

20
0

3

0

high Jp
low low high Jcf high Jp, high Jcf
Jp, Jp
high Jcf
high Jcf
filtration conditions

low Jp
low Jp, high Jcf high highhigh Jcf
Jp, Jp
high Jcf
high Jcf
filtration conditions

Figure 4.3: a) P22 recovery and b) removal from MBR effluent with PEM-coated membrane at
high and low permeate fluxes. Error bars indicate standard deviation. Data reported here
corresponds to samples where TOC was not adjusted.

Although pre-elution recovery from MBR effluent by PEM-coated membranes was
noticeably higher than by CS-blocked membrane, the difference was not significant. No
difference was observed for surface water. At the post-elution level there was no difference
between CS-blocked membranes and PEM-coated membranes for both types of water, which is
also what was observed for DI water and was attributed to the weak adhesion of P22 to surfaces
compared to other viruses. P22 removal was not significantly affected by the type of membrane
or the type of water.
116

4.3.3

Effect of backflushing on P22 recovery from surface water.

The fouling layer can entrap virions and render them less recoverable by elution. In that
case adding a backflushing step can help recover viruses entrapped in the fouling layer. We
performed backflushing after tests on the filtration of surface water spiked with P22. For these
tests, we already showed (Table 4.2) that post-elution recoveries were not significantly lower
than post-elution recoveries from DI water which means that the loss of viruses due to
entrapment was probably minimal. It is thus not surprising that backflushing did not
significantly increase P22 recovery. For all 6 filtrations where a backflush was performed, the
post-backflush recovery was only marginally (less than 1%) higher the post- elution recovery.

4.3.4

Fouling resistance of CS-blocked and PEM-coated membrane

Figure 4.4 shows the ratio of final to initial permeate flux. This ratio is a measure of the
amount of fouling occurring during filtration. Initial and final permeate flux were averaged over
the first and last minute of filtration, respectively. For DI water experiments the permeate flux
reduction was higher for PEM-coated membranes than for CS-blocked membranes. For MBR
effluent, the trend seemed to be reversed and PEM-coated membranes seemed to experience less
fouling under both “highest recovery” and “fastest filtration” conditions (the difference was only
significant under “fastest filtration” conditions) compared to CS-blocked membranes. Fouling
resistance of the PEM-coated membranes stemmed from its higher hydrophilicity compared to
CS--blocked membranes (see Chapter 2). For surface water, the amount of fouling observed for

117

CS- blocked membranes was higher on average than for PEM-coated membrane, but the
difference was not statistically significant.

4.3.5

Conclusion

Due to the small number of filtrations performed the data presented here are only preliminary. It
was not possible to observe significant difference between PEM-coated membrane and CSblocked membrane in complex water matrices. However, PEM-coated membranes seemed to
maintain their advantage in terms of P22 pre-elution recovery when concentrating viruses from
MBR effluent. PEM-coated membranes also seemed to be more resistant to fouling. In
experiments with surface water, however, no difference was observed. Additional filtration tests
are needed to validate these trends.

118

a)

Jp final/ Jp initial

1.0

CS-blocked membrane

highest recovery
conditions

0.8
0.6
0.4
0.2
0.0
DI water

MBR effluent

surface water

type of water
b)
1.0

CS-blocked membrane

fastest concentration
conditions

Jp final/ Jp initial

0.8
0.6
0.4
0.2
0.0
DI water

MBR effluent
type of water

Figure 4.4: Fouling observed for CS-blocked membrane and PEM-coated membrane during
filtration tests performed to recover P22 from DI water, MBR effluent and surface water under a)
“highest recovery” conditions and b) “fastest filtration” conditions.

119

REFERENCES

120

REFERENCES

1.

Belfort, G., Y. Rotem, and E. Katzenelson, Virus concentration using hollow fiber
membranes .2. Water Res., 1976. 10(4): p. 279-284.

2.

Gibson, K.E. and K.J. Schwab, Tangential-flow ultrafiltration with integrated inhibition
detection for recovery of surrogates and human pathogens from large-volume source
water and finished drinking water. Applied and Environmental Microbiology, 2011.
77(1): p. 385-391.

3.

Olszewski, J., L. Winona, and K.H. Oshima, Comparison of 2 ultrafiltration systems for
the concentration of seeded viruses from environmnetal waters. Can. J. Microbiol., 2005.
51(4): p. 295-303.

5.

Rajal, V.B., et al., Validation of hollow fiber ultrafiltration and real-time PCR using
bacteriophage PP7 as surrogate for the quantification of viruses from water samples.
Water Res., 2007. 41(7): p. 1411-1422.

6.

Hill, V.R., et al., Multistate evaluation of an ultrafiltration-based procedure for
simultaneous recovery of enteric microbes in 100-liter tap water samples. Applied and
Environmental Microbiology, 2007. 73(19): p. 6327-6327.

6.

Jacangelo, J.G., S.S. Adham, and J.-M. Laine, Mechanism of cryptosporidium, giardia,
and MS2 virus removal by MF and UF. American Water Works Association Journal,
1995. 87(9): p. 107-121.

7.

Wickramasinghe, S.R., et al., Understanding virus filtration membrane performance.
Journal of Membrane Science, 2010. 365(1-2): p. 160-169.

8.

Guy, R.A., et al., Real-time PCR for quantification of Giardia and Cryptosporidium in
environmental water samples and sewage. Applied and Environmental Microbiology,
2003. 69(9): p. 5178-5185.

121

Chapter 5. Suggested future research

Chapters 2 and 4 describe results of what is to our knowledge the first study of PEM
membranes’ ability to remove and concentrate viruses. The membranes showed promising
performance in terms of virus removal and pre-elution recovery; the latter was attributed to the
negative charge and hydrophilicity of the PEM surface that resulted in unfavorable virus-PEM
interactions. Our findings regarding the performance of PEM-coated membrane and calf serumblocked membrane are summarized in Table 5.1.

A large number of experiments was necessary to evaluate the effect of crossflow and
permeate flow rates on virus recovery. To make such screening feasible, we chose to work with
the bacteriophage P22, which is a non-pathogenic virus. This choice was later proved be not
optimal as relatively high post-elution recovery of P22 could be achieved by an existing method
(using CS-blocked membranes). We expect that in the case of more adhesive viruses, such as
adenovirus, the competitive advantage of the anti-adhesive PEM surface will be more evident.
Specifically we expect the post-elution recovery obtained with PEM-coated membranes to be
significantly higher than the one obtained with CS-blocked membranes. Thus, we suggest that
PEM membranes are evaluated with adhesive, difficult-to-recover viruses. We also propose that
these tests include backflushing the membrane to recover viruses that do attach to the surface.

The performance of PEM-coated membranes in recovering viruses from complex water
was evaluated with only three filtration tests for each water matrix (MBR effluent and surface
water). Additional tests should be performed to confirm or disprove the advantage of PEM122

coated membrane over CS-blocked membranes in terms of recovery and/or reproducibility at the
post-elution stage. Using a backflushing step instead of the elution step should also be evaluated
as a possibility to improve reproducibility. The polyelectrolytes used in this study inhibited P22
infectivity and as a consequence the performance of PEM-coated membranes for the recovery of
infective virus was not investigated. Future work should consider alternative PEM coatings that
would not inhibit infectivity assay; in this context, multilayer formed through H-bonding rather
polyelectrolyte interaction are an interesting option. Finally the PEM-based approach should be
evaluated in a scaled-up system with a reusable, high surface area hollow fiber cartridge to
concentrate large-volume samples. As both membrane surface area and sample volume will be
increased proportionally, we expect that PEM-coated membrane performance will be maintained.

The role of surface properties such as hydrophobicity, charge, roughness and stiffness in
determining virus adhesion needs to be better understood. We suggest experimenting with
carefully selected bacteriophages that present a wide range of surface properties and using a
QCM-D instrument to rigorously quantify bacteriophage adhesion to various PEMs. Both PEMcoated surfaces, prepared using various polyelectrolytes combinations as well as various pH and
ionic strength, and phage surfaces would need to be thoroughly characterized to determine their
surface tension parameters and surface charge. In addition, atomic force microscopy would need
to be used to characterize PEMs in terms of their roughness, swelling and stiffness. Based on the
comparison of results of QCM-D tests, results of AFM measurements, and computed XDLVO
energies of virus-PEM interactions, conclusions may be drawn on the factors determining virus
adhesion to surfaces.

123

Table 5.1: Virus recovery performance of PEM-coated membrane and calf serum-blocked
membrane at the pre-elution (rpre), post-elution (rpost) and coating removal - through elution or
backflush with an adequate solution - (rsacrificial) stages. Note: +and ++ indicate that we expect
the membrane well and very well, respectively. – indicates that we expect the membrane to
perform poorly and ? indicates that additional experiments need to be performed to confirm our
assumption

rpre
PEM-coated membrane

rpost

rsacrificial

2

-?

++

+

infective virus

-

-

?

total virus
1

+?

total virus
calf serum-blocked membrane

infective virus

-

+

1

?

P22-specific inhibition
If PEM used does not inhibit qPCR

124

1

+

2

APPENDICES

125

Appendix A. Determination of crossflow and permeate fluxes used
by Morales-Morales et al. [1] and Hill et al. [2-4]

Morales-Morales et al. [1] reported concentrating pathogen from a 10 L sample using a permeate
flow of 300 mL/min and a crossflow rate of 3600 mL/min in a Microza hollow fiber cartridge
2

with 0.2 m membrane surface (model (AHP-1010). The permeate flux obtained by dividing the
2

permeate flow by the membrane surface area was 1.5 L/(min·m ). Pall Corporation, Microza’s
manufacturer, reports on its website [5] that a crossflow rate of 724 L/h corresponds to a velocity
of 1 m/s. For a 3,600 mL/min crossflow rate the velocity was thus 0.3 m/s, so the fluid
3

2

2

volumetric flux is 0.3 m /(m ·s) (or 18,000 L/(min·m )) [6].

In 2005, Hill et al. [2] reported concentrating pathogen from 10 L tap water samples with a
permeate flow of 800 mL/min and a crossflow rate of 900 mL/min using a Fresenius Hemaflow
2

F80A hollow fiber filter with a membrane surface of 1.8 m . Again by dividing the permeate
2

flow by the membrane surface area gives a permeate flux of 0.44 L/(min·m ). Ronco et al. [7]
reported that the Hemaflow F80A cartridge contains 12,288 fibers, each with an inner diameter
-4

2

of 200 μm corresponding to a total crossflow cross-sectional area of 3.86·10 m . Thus, a
2

crossflow rate of 900 mL/min corresponds to a crossflow volumetric flux of 2,300 L/(min·m ).

126

In 2007 and 2009, Hill et al. [3, 4] used a Fresenius F200NR hollow fiber cartridge with a
crossflow rate of 1,700 mL/min and a permeate flow rate of 1,200 L/min. The membrane surface
2

2

area was 2 m , which corresponds to a permeate flux of 0.6 L/(min.m ). The Fresenius F200NR
was also described by Ronco et al. [7] The cartridge contained 13,824 fibers, each with an inner
-4

2

diameter of 200 μm, which corresponded to a total cross-sectional area of 4.24·10 m . We thus
determined that a crossflow rate of 1,200 L/min corresponds to a crossflow volumetric flux of
2

3,900 L/(min.m ).

127

REFERENCES

128

REFERENCES

1.

Morales-Morales, H.A., et al., Optimization of a reusable hollow-fiber ultrafilter for
simultaneous concentration of enteric bacteria, protozoa, and viruses from water.
Applied and Environmental Microbiology, 2003. 69(7): p. 4098-4102.

2.

Hill, V.R., et al., Development of a rapid method for simultaneous recovery of diverse
microbes in drinking water by ultrafiltration with sodium polyphosphate and surfactants.
Applied and Environmental Microbiology, 2005. 71(11): p. 6878-6884.

3.

Hill, V.R., et al., Multistate evaluation of an ultrafiltration-based procedure for
simultaneous recovery of enteric microbes in 100-liter tap water samples. Applied and
Environmental Microbiology, 2007. 73(19): p. 6327-6327.

4.

Hill, V.R., et al., Comparison of hollow-fiber ultrafiltration to the USEPA VIRADEL
technique and USEPA Method 1623. J. Environ. Qual., 2009. 38(2): p. 822-825.

5.

Microza membranes. 2012 [cited 2012; Available from:
http://www.pall.com/main/Food-and-Beverage/Product.page?id=41053.

6.

Reible, D.D., Fundamentals of engineering1999: CRC Press LLC.

7.

Ronco, C., et al., Flow distribution analysis by helical scanning in polysulfone
hemodialyzers: Effects of fiber structure and design on flow patterns and solute
clearances. Hemodialysis International, 2006. 10(4): p. 380-388.

129

Appendix B. Addition of carrier RNA to sample: effect of DNA
extraction/detection

QPCR analysis of sample of P22 spiked in DI water and extracted using Roche’s MagNA Pure
resulted in higher CP value than expected. This phenomenon (lower CP value for DNA material
extracted from “clean samples” using MagNA Pure compared to environmental samples) had
also been observed by Dr. Kelvin Wong when he was working in Dr. Xagoraraki’s lab and was
attributed to the adsorption of virions and DNA to the magnetics beads and other components of
the MagNA Pure instrument. The solution developed by Dr. Wong was to add carrier RNA to
the sample before the extraction step. Carrier RNA had been used previously to increase DNA
extraction efficiency [1]. In Table A.1 we show the results of tests we performed to determine
the optimal amount of carrier RNA to add to filtration samples.

Carrier RNA (obtained from Qiagen) solutions were prepared at a concentration of 1 mg/mL in
PCR grade water or at a concentration of 0.01 mg/mL in viral lysis buffer (buffer AVL from
Qiagen) and amount ranging from 5 µL to 15 µL of carrier RNA solution were added to 385 µL
P22 in DI water samples. Samples were extracted using MagNA Pure instruments and qPCR
was run in triplicate as described in Chapter 2 section 2.2.8.

130

Table B.1: Effect of carrier RNA addition on CP value obtained by qPCR
Tube number
Volume introduced (µL)

1

2

3

4

6

sample

385

385

385

385

385

Carrier RNA in water

NA

NA

10

NA

NA

5

10

NA

15

NA

PCR grade water

10

5

5

NA

NA

total

400

400

400

400

400

qPCR results (mean CP

19.81 ± 0.22

19.66 ±

19.18 ±

19.63 ±

23.62 ±

0.01

0.03

0.12

0.02

(1 mg/mL)
Carrier RNA in buffer
(0.01 mg/mL)

value ± stdev)

The lowest CP value, corresponding to the highest P22 concentration detected, was obtained for
tube number 3 (which corresponds to the addition of 10 µL of carrier RNA at 1 mg/mL in qPCR
grade water). Ten microliters of carrier RNA (1 mg/mL) were added to all samples used in this
study before DNA extraction.

131

REFERENCES

132

REFERENCES

1.

Kishore, R., et al., Optimization of DNA Extraction from Low-Yield and Degraded
Samples Using the BioRobot® EZ1 and BioRobot® M48. Journal of Forensic Sciences,
2006. 51(5): p. 1055-1061.

133

Appendix C. Compaction behavior of PEM-coated membranes and
CS-blocked membranes.

Permeate flux data were averaged over the first and last minute of compaction and over the first
minute of filtration in order to look at possible difference of behavior during compaction and
decompression (between the compaction stage and the filtration stage) of the membrane. Results
are shown in Fig. C.1. There is no significant difference of behavior between CS-blocked
membrane and PEM-coated membrane.
600

compaction initial
compaction final
Filtration initial

permeability (L/hr.m2.bar)

500
400
300
200
100
0
bare PES 30kDa

CS
Type of membrane

PEM

Compaction and decompaction behavior of bare PES membranes, CS-blocked
membranes and PEM-coated membranes. Error bar indicates the 95% confidence intervals.
Number of test: n = 3 for bare PES membrane, n = 72 for CS-blocked membrane and n = 37 for
PEM coated membrane

134

Appendix D. PEM layer stability during filtration and elution.

The stability of the PEM layer during filtration was tested as described in section 3.2.6. The test
was performed in triplicate for filtration and elution conditions. On Fig D.1 we can see that 25
kDa dextran rejection by (HE/CHI)4.5 was highly variable and was only approximately 50% on
average (whereas CS-blocked membrane rejected approximately 60 to 70 % of 25 kDa dextran).
A higher molecular weight dextran might have been a better probe to assess the stability of
(HE/CHI)4.5. The PEM stability seems to be maintained at low Jcf / J p however, for Jcf / J p
value higher than 7,000 rejection of 25 kDa dextran seems to decrease. During elution, dextran
rejection dramatically decreased even at the lowest crossflow rate used in these experiments
which indicates that (HE/CHI)4.5 probably lose integrity during elution.

135

a)
100

immediately after Jcf increase
5 min after Jcf increase

90
dextran rejection, %

80
70
60
50
40
30
20
10
0
0.0

0.2

0.4

0.6

0.8

1.0

crossflow to permeate flux ratio Jcf/Jp

1.2

104

b)
100
immediately after Jcf increase
5 min after Jcf increase

90
dextran rejection, %

80
70
60
50
40
30
20
10
0
0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

crossflow flux Jcf 104, L/min.m2
Dextran rejection by PEM-coated membrane under increasing Jcf / J p during a)
filtration and b) elution. n=3 for filtration and elution, error bar indicates standard deviation.
136

Appendix E. Preliminary work on visualization of virus adhesion on
surfaces

As mentioned in Chapter 1 it is possible to visualized virions and aggregates deposited on
surfaces using AFM [1-5]. Aside from Pereira et al. study [5], viruses were deposited on clean
and smooth surface such as silicon or mica. In Pereira study, dengue virus adsorption on silicon
wafers and protein covered silicon wafers. On those they were observed aggregated and isolated
particles with height of approximately 10 to 15 nm which the authors considered to be
hemagglutinins particles secreted by dengue virus, and 25 to 30 nm which the authors considered
to be dengue particles with a collapsed lipid envelope [5].

In this preliminary study (PDADMA/PSS)4.5 were deposited on mica support (polyelectrolyte
solution were prepared at a polyelectrolyte concentration of 0.02M and a NaCl concentration of
0.5M). The PEM-coated supports were then dipped for 30 min. in P22 suspension. To prepare
this suspension, P22 stock was purified by centrifugal filtration at 4000 rpm using Amicon Ultra
4 centrifugal filter device with a 100,000 kDa cellulose membrane (Millipore). The virions
retained by the membrane were resuspended in KCl 0. M. P22 concentration was approximately
9

10 PFU/mL. The supports were then dried and analyzed by AFM using the Nanoscope IV
Multimode SPM from Veeco in tapping mode at room temperature using silicon cantilever with
NSC15 tip from Mikromasch.

The obtained AFM topographic imaged are shown in Fig. E.1
137

a)

b)

90 nm

80 nm

2

4

6

8

µm
6

4

2

10

c)

8

µm

40 nm
20 nm
0 nm

0
0

10

µm

AFM topographic images of a) (PDADMA/PSS)4.5 on gold substrate b) and c)
(PDADMA/PSS)4.5 on mica dipped in P22 suspension

In Fig. E.1, image b which is the (PDADMA/PSS)4.5 dipped in P22 suspension scan, shows
more high structure (up to 80 nm) than image a which is the scan of a (PDADMA/PSS)4.5 which
could be attributed to P22 particles (~60 nm). However, it is not possible to make a direct
comparison as the layers were not deposited on similar surface.

138

Another potential method to visualize virus is epifluorescence [4]. Epifluorescence has been
used to enumerate virions from water samples [4, 6-11]. Viruses are recovered on a filter which
is strained with SYBR Green, DAPI or Yo-Pro (fluorescent dyes that specifically bind to DNA).
The virus can then be visualized and enumerated by epifluorescence microscopy.

In this preliminary study a serial dilution of P22 suspensions was prepared. Ten milliliters from
each dilution was filtered with 0.02µm anodisc filter from Whatman backed with a 0.45 µm
cellulose ester filter. The filters were then dried and following Noble 2001 protocols a drop of
staining solution (0.25% SYBR Green I gel strain from Invitrogen) was placed on the filter for
15 min, the drop was then removed and the filter dried. The filter was mounted on a glass slide
with a drop of anti-fading solution (0.1% p-phenylenediamine in a mixture of 50% glycerol and
50% phosphate buffer saline) between the sample and the cover slip. Samples were frozen until
analyzed using a Zeiss Pascal laser scanning confocal microscope with a 488 nm illumination.
The emission was detected through a long pass 505 nm filter.

In addition to the serial dilution (PSS/PAH)4.5 and (PSS/PAH)4 were deposited on anodisc filters
8

which were then dipped for 30 min in P22 suspension (~10 PFU/mL). The filters were then
dried and stained as described above.

Results are shown in Fig. E.2 and E.3.

139

a)

50 µm

50 µm

c)

d)

50 µm

50 µm

Fluorescence images of serial dilution samples. a) corresponds to the 1:1000
4

5

6

dilution, b) to the 1:10 dilution, c) to the 1:10 dilution and d) to the 1:10 dilution
Pictures were taken at ×40 magnification

There does not seem to be any correlation between the number of fluorescent spots and the
dilution factor. One possibility is that focusing was done on viral aggregates and that the rest of
the virions would be visible at a different focal plan.

140

a)

b)

100 µm

100 µm

c)

100 µm

Fluorescence images of a) (PSS/PAH)4-coated ; b) (PSS/PAH)4.5-coated and c) bare
anodisc filters dipped in P22 suspension.
Pictures were taken at ×20 magnification

Fluorescent spots are visible on both PEM-coated anodisc but barely any are visible on the bare
anodisc filter. Unfortunately, because no controls were performed it is not possible to say if these
fluorescent spot are attributable to P22 particles or caused by agglomerates of polyelectrolyte
itself.
141

REFERENCES

142

REFERENCES

1.

Ikai, A., et al., Atomic force microscope of bacteriophage T4 and its tube-baseplate
complex. FEBS Letters, 1993. 326(1-3): p. 39-41.

2.

Kolbe, W.F., D.F. Ogletree, and M.B. Salmeron, Atomic Force Microscopy Imaging of
T4 Bacteriophages on Silicon Substrates. Ultramicroscopy, 1992. 42: p. 1113-1117.

3.

Dubrovin, E.V., et al., Atomic force microscopy as a tool of inspection of viral infection.
Nanomed. Nanotechnol. Biol. Med., 2007. 3(2): p. 128-131.

4.

Zago, M., et al., Epifluorescence and atomic force microscopy: Two innovative
applications for studying phage-host interactions in Lactobacillus helveticus. Journal of
Microbiological Methods, 2012. 88(1): p. 41-46.

5.

Pereira, E.M.A., et al., Binding of dengue virus particles and dengue proteins onto solid
surfaces. ACS Applied Materials & Interfaces, 2010. 2(9): p. 2602-2610.

6.

Noble, R.T. and J.A. Fuhrman, Use of SYBR green I for rapid epifluorescence counts of
marine viruses and bacteria. Aquat. Microb. Ecol., 1998. 14: p. 113-118.

7.

Hara, S., K. Terauchi, and I. Koike, Abundance of viruses in marine waters - Assessment
by epifluorescence and transmission electron-microscopy. Applied and Environmental
Microbiology, 1991. 57(9): p. 2731-2734.

8.

Hennes, K.P. and C.A. Suttle, Direct counts of viruses in natural waters and laboratory
cultures by epifluorescence microscopy. Limnol. Oceanog., 1995. 40: p. 1050-1055.

9.

Weinbauer, M.G. and C.A. Suttle, Comparison of epifluorescence and transmission
electron microscopy for counting viruses in natural marine waters. Aquat. Microb. Ecol.,
1997. 13: p. 225-232.

10.

Wen, K., A.C. Ortmann, and C.A. Suttle, Accurate estimation of viral abundance by
epifluorescence microscopy. Applied and Environmental Microbiology, 2004. 70(7): p.
3862-3867.

11.

Rinta-Kanto, J.M., et al., Rapid enumeration of virus-like particles in drinking water
samples using SYBR green I-staining. Water Research, 2004. 38(10): p. 2614-2618.

143