Journal of Virological Methods 117 (2004) 19–25

Standardised evaluation of the performance of a simple membrane

filtration-elution method to concentrate bacteriophages
from drinking water
Javier Méndez a , Ana Audicana b , Ana Isern c , Julián Llaneza d , Belén Moreno e ,
Marı́a Luisa Tarancón f , Juan Jofre a , Francisco Lucena a,∗
a Departamento de Microbiologı́a, Universidad de Barcelona, Avda. Diagonal 645, 08028 Barcelona, Spain
b Dirección de Salud Pública del Gobierno Vasco, C/ Marı́a Dı́az de Haro, 60, 48010 Bilbao, Spain
c Laboratori Municipal de Barcelona, Institut Municipal de Salut Pública, Avda. de les Drassanes, 13-15, 08001 Barcelona, Spain
d Laboratorio de Salud Pública de Asturias, Finca de la Cedellada, Camino de Rubı́n s/n, 33011 Oviedo, Spain
e Laboratorio de Salud Pública de Gipuzkoa, Avda. Navarra, 4, 20013 Donostia-San Sebastián, Spain
f Conselleria de Salut i Consum, Plaça Hospital 3, 07012 Palma de Mallorca, Spain

Received 7 July 2003; received in revised form 17 November 2003; accepted 18 November 2003

Abstract

The bacteriophage elution procedure described further after adsorption to acetate–nitrate cellulose membrane filters allows better recovery
of phages concentrated from 1 l of water than elution proceduresm used previously. The improvement is due to the combined effect of the
eluent (3% (w/v) beef extract, 3% (v/v) Tween 80, 0.5 M NaCl, pH 9.0) and the application of ultrasound instead of agitation or swirling.
Average recovery of somatic coliphages, 82 ± 7%, was the greatest, and that of phages infecting Bacteroides fragilis, 56 ± 8%, the lowest,
with intermediate values for F-specific and F-specific RNA bacteriophages. Thus, the method allowed recovery of over 56% for all the phages
suggested as surrogate indicators. The method was then validated according to an International Standardisation Organisation validation
standard procedure and implemented in routine laboratories, which obtained reproducible results.
© 2003 Elsevier B.V. All rights reserved.

Keywords: Bacteriophages; Concentration; Membrane filter; Elution; Validation

1. Introduction approaches: the first is a presence/absence method, which

is only qualitative and the second is through concentration
Bacteriophages have been investigated as potential model and plaque assay, which is quantitative. Many methods
organisms for water quality assessment (IAWPRC, 1991; for concentrating phages from water samples had been
Grabow, 2001; Jofre, 2002). Phages proposed for water described (Logan et al., 1980; Schulze and Lenk, 1983;
quality assessment are somatic coliphages (Kott and Ben Shields and Farrah, 1986; Sobsey et al., 1990; Borrego
Ari, 1968), F-specific RNA bacteriophages (Havelaar et al., et al., 1991; Lucena et al., 1995; Sinton et al., 1996). Most
1986) and bacteriophages infecting Bacteroides fragilis of the methods have only been described for one type of
(Tartera and Jofre, 1987). Although phages are present in bacteriophages and generally validated with a laboratory
surface waters at concentrations high enough to allow di- strain. Among these, we find very feasible a method based
rect enumeration in small volumes of sample, this is not on adsorption to acetate–nitrate cellulose ester membrane
the case with groundwater or drinking water. In this case it filters described by Sobsey et al. (1990) that was modified
is necessary to test phages in larger, at least 1 l, volumes. further by Sinton et al. (1996). The method is very efficient
Detecting phages in 1 l volumes can be achieved by two for concentrating coliphages from 100 ml, but results for
1 l volumes are below 50% recovery (Hsu et al., 1998). We
∗ Corresponding author. Tel.: +34-934-021-484; tested the method as modified by Sinton et al. (1996) for
fax: +34-934-110-592. concentrating somatic coliphages from 1 l of drinking water
E-mail address: fluena@ub.edu (F. Lucena). and the recovery was also below 50% (data not shown).

0166-0934/$ – see front matter © 2003 Elsevier B.V. All rights reserved.
doi:10.1016/j.jviromet.2003.11.013
20 J. Méndez et al. / Journal of Virological Methods 117 (2004) 19–25

It is well known that concentration efficiencies obtained for somatic coliphages, F-specific RNA bacteriophages
when testing natural samples for viruses are frequently sig- and B. fragilis phages, respectively were used as con-
nificantly lower than those reported in the publications de- trols in enumeration tests and for spiking of water in the
scribing the method as developed with reference laboratory method development experiments indicated further. As well,
bacteriophages. Also, Round Robin tests among different sewage derived phage stocks for SOMCPH, F-TOTPH,
laboratories show significant differences in concentration ef- F-RNAPH, BFRPHHSP40 and BFRPHRYC2056 were
ficiencies between the different laboratories (Melnick et al., prepared as indicated further for method validation
1984; Vilaginés et al., 1997). Many factors other than the studies.
concentration method itself seem to influence the recovery:
the characteristics of the water sample, the amount of sam- 2.3. Spiking phage suspensions
ple, the numbers of bacteriophages spiked and the bacterio-
phages groups themselves. Because of the many sources of Two kinds of phage suspensions were used.
uncertainty it is important to establish “standardised” pro-
cedures for evaluating concentration procedures, based on 2.3.1. For developing the modified method
both reliability and efficiency. Recently, a standard proce- Volumes of suspensions of phages X174, MS2 and
dure for validation of methods for concentration of bacte- B40-8 containing between 1.5 × 104 and 5.0 × 105 pfu per
riophages from water has been adopted by the International ml were added to the water sample to be submitted to con-
Standardisation Organisation (Anonymous, 2002). This pro- centration to reach a final concentration of 60–200 plaque
cedure does not give specific details of concentration meth- forming units per litre. After vigorous agitation, phages
ods, but provides a general framework for the evaluation in the sample were enumerated, either directly or after
of the suitability of a particular method for a given type dilution.
and volume of water. Secondly, it also discusses how to use
this information in establishing the actual values of bacte- 2.3.2. For validating the modified method
riophages in a given sample taken into account the numbers Sewage derived phage stocks were prepared as indicated
recovered after concentration. in the ISO 10705-3 Standard (Anonymous, 2003). Briefly,
We describe modifications introduced to the method first primary or secondary sewage effluent was either centrifuged
reported by Sobsey et al. (1990) and then modified by Sinton for 10 min at 2000 × g or filtered through a 12 ␮m pore
et al. (1996). The modifications introduced were examined size membrane filter as described in Mendez et al. (2002).
for their impact upon the concentration of phages belonging Target bacteriophages were enumerated for initial titre. The
to the three groups reported above as potential indicators phage suspensions were stored at 4 ◦ C until the enumeration
from 1 l of drinking water. The final modified method was results were available. If necessary, the suspensions were
then validated according to the international validation stan- diluted to obtain concentrations of 60–200 plaque forming
dard mentioned above. Finally, the implementation of the units per ml of each of the phage groups. Since their numbers
method in routine microbiology laboratories and the repro- in sewage are different, these means that three sets of phage
ducibility of the results with sewage derived bacteriophages spike-stocks were prepared. Glycerol was added to a final
in these laboratories were assessed. concentration of 10% (v/v). Then, these suspensions were
distributed in 10 ml volumes in plastic containers, and frozen
2. Materials and methods at −20 or −70 ◦ C.
To ensure that sewage derived phages were distributed at
2.1. Phage assay random within and between the containers, the within and
between containers homogeneity was tested. Two contain-
Enumeration of bacteriophages was done by determin- ers for each phage group suspension (somatic coliphages,
ing the number of plaque forming units (pfu) by the dou- F-specific/F-specific RNA phages and phages infecting B.
ble agar layer technique according to standards ISO10705-1 fragilis infecting strains HSP40 and RYC2056) were thawed
(Anonymous, 1995), ISO10705-2 (Anonymous, 2000) and at room temperature, and from each container two 0.5 ml
ISO10705-4 (Anonymous, 2002), which describe methods aliquots for the target bacteriophages were assayed. The
for the detection and enumeration of F-specific (F-TOTPH) counts were then analysed for within (T1 ) and between con-
and F-specific RNA bacteriophages (F-RNAPH), somatic tainers (T2 ) homogeneity.
coliphages (SOMCPH) and phages infecting strains HSP40 T1 is a measure for the variation within one container of
(BFBHSP40) and RYC2056 (BFBRYC2056) of B. fragilis, spiking material (replicate variation). T2 is a measure for
respectively. the variation between different vials of one batch of spiking
material.
2.2. Bacteriophages
J
I

 

(zij − zi+ /J) 2
Phages X174, MS2 and B40-8, recommended by the T1 =
Standard ISO methods indicated above as reference phages (zi+ /J)
i=1 j=1
 J. Méndez et al. / Journal of Virological Methods 117 (2004) 19–25 21

where zij is the number of pfu per analytical portion (j) of 2.4.3. Eluting conditions
vial i, and The effect on elution efficiency of swirling agitation and
ultrasound was compared as follows. The eluting conditions
J

were assayed with the eluting solution 1 that was confirmed
zi+ = zij be the most efficient as indicated in the results section. The
j=1 swirling agitation of the membrane filter submerged in elut-
ing solution was applied as indicated by Sinton et al. (1996).
being the sum of numbers of pfu in all containers counted Ultrasound was applied by submerging the flask or tube con-
(I, that in this case is 2) and J (in this case 2) the number taining the membrane filter into an ultrasound cleaning bath
of analytical portions per container. Degrees of freedom for for 4 min.
T1 are I(J − 1).

J   2.5. Concentration method proposed and validated

(zi+ − z++ /J)
T2 =
z++ /I As indicated above, the method came up as a modification
j=1
of a membrane filter phage concentration method previously
where described (Sobsey et al., 1990; Sinton et al., 1996) accord-
ing to different changes tested herein. The method is recom-
I 


 mended for volumes ranging from 100 ml to 1 l of waters
z++ = zij with turbidity <2.0 NTU. For concentrating bacteriophages
i=1 from 1 l of water proceed as follows. First, add MgCl2 to the
water sample to reach a final concentration 0.05 M. Then,
is the total count of plaques over all vials and duplicates. filter the sample through an acetate–nitrate cellulose ester
Degrees of freedom for T2 are I − 1. membrane filter, 0.22 ␮m pore size, 47 mm diameter. Filter
As indicated in the ISO Standard the phage suspensions at a flow rate of approximately 1 l in 30 min. Cut the mem-
were considered as suitable for the validation of the method brane filter in eight fragments and place into a glass flask
if 0.01 < T1 < 5.99 and T2 < 3.84. containing 5 ml of eluting solution (1% beef extract, 0.5 M
NaCl and 3% Tween 80). Place the flask into an ultrasound
2.4. Modifications assayed for the phage concentration cleaning bath for 4 min. Count the eluted bacteriophages
method by the double layer agar method. Count the bacteriophages
retained in the membrane fragments placing the fragments
The experiments to test the effect of changes of protocol face down onto a host monolayer.
in concentration efficiency were done spiking bottled wa-
ter samples with suspensions of laboratory reference phages 2.6. Evaluation and validation of the method with sewage
followed by enumeration of phages before and after con- derived phages
centration. The protocol was applied to the assays described
further. This was carried out as indicated in the standard
ISO10705-3 (Anonymous, 2003).
2.4.1. Sample amendment Briefly, five samples of bottled water or tap water
The effect on virus adsorption of different MgCl2 concen- dechlorinated by adding sodium thiosulphate (Anonymous,
trations (0.005, 0.05 and 0.5 M, the last one being the con- 1998) were tested. Turbidity was measured after spiking
centration proposed by Sinton et al. (1996)) in the sample and ranged from 0.1 to 2.0 nephelometric turbidity units
was assayed. For these tests a 4.14 M MgCl2 ·6H2 O solution (NTUs). Since 1 l is the maximum volume proposed for
was prepared and the volumes necessary to reach the above this method, containers containing 125, 250, 500, 1000 ml
indicated final concentrations were added to the reference were prepared. One millilitre of sewage derived spiking
phage spiked water sample. material prewarmed at room temperature was added to each
container, and the remainder of the spiking material was
2.4.2. Eluting solutions kept on melting ice for enumeration.
The following eluting solutions were tested: (1) 3% (w/v)
beef extract, 3% (v/v) Tween 80, 0.5 M NaCl, pH 9.0; (2) 2.6.1. Calculation of recovery
1% (w/v) beef extract, 0.5 M NaCl, pH 9.0; (3) 3% (w/v) Let Nc be the total number of plaques recovered from a
beef extract, 0.5 M NaCl, pH 9.0; (4) 3% (w/v) beef extract, concentrate, and Ns be the number of plaques from 1 ml of
0.5 M NaCl, 0.4 M urea, pH 9.0; (5) 1% (w/v) beef extract, the spiking material. The percentage recovery was calculated
0.5 M NaCl, 0.4 M urea, pH 9.0; (6) 1% (w/v) beef extract, as R(%) = Nc /Ns × 100.
0.5 M NaCl, 0.25 M glycine, pH 9.0; (7) 0.5 M NaCl, 0.25 M The arithmetic average recovery, the standard deviation
glycine, pH 9.0; (8) 1% (w/v) beef extract, 0.25 M glycine, and the confidence interval were calculated for all assayed
pH 9.5. volumes. The R(%) plus the confidence intervals were
22 J. Méndez et al. / Journal of Virological Methods 117 (2004) 19–25

plotted against the volume of sample and examined for Table 1

linearity. If confidence intervals of R(%) for the different Comparison of elution efficiencies using different eluents
volumes overlapped, it was considered that there were no Eluent
significant volume effects. 1a 2 3 4 5 6 7 8
The results of all volumes tested were combined and the
arithmetic mean of R(%), the standard deviation and the rela- X174 100b 56.9 60.0 63.1 73.8 79.7c 23.4 73.5
MS2 100 52.0 68.1 63.4 50.7 41.4 2.9 64.4
tive standard deviation (standard deviation/arithmetic mean) B40-8 100 25.2 42.5 42.5 38.1 44.1 4.0 77.3c
were calculated. According to the ISO standard the method
Results are the average of nine independent experiments.
can be considered reliable if the relative standard deviation a Indicated in the Section 2.
is <0.5. b Recovery by eluent 1 has been assigned as 100%. All other recoveries

are expressed as relative (%) to recovery by eluent 1.

c The values were not significantly different from data in the first
2.7. Concentration in different laboratories
column.
Five routine laboratories, which in a previous collabora-
tive study aimed to ensure the correct implementation in the 3.1.2. Eluting solution
laboratory of the methods for enumerating bacteriophages The eluting solution number 1 (3% (w/v) beef ex-
succeeded in the implementation of the methods, partici- tract, 3% (v/v) Tween 80, 0.5 M NaCl, pH 9.0), which
pated in the trial. was established as 100% recovery (Table 1) gave re-
Each laboratory prepared its own spiking material from coveries significantly higher (ANOVA, P < 0.05) than
sewage as indicated above. any other of the eluting solutions tested for the three
They assayed (4/5 assays for each group of bacterio- bacteriophages, with the exception of solution number
phages per laboratory) the concentration of somatic col- 6 for phage X174 and solution number 8 for phage
iphages, F-specific bacteriophages and phages infecting B. B40-8. Consequently it was decided to elute the ad-
fragilis RYC2056 from 1 l of spiked water. sorbed phages with the following solution: 3% beef ex-
tract 3% Tween 80, 0.5 M NaCl, pH 9.0 in the modified
2.8. Statistical tests method.

Confidence intervals based

√ on Student’s
√ t-tests were 3.1.3. Eluting conditions
calculated as {X̄ − t × s/ n; X̄ + t × s/ n} where X̄ is Ultrasound gave significantly higher elutions (ANOVA,
the arithmetic average recovery; s the standard deviation, P < 0.05), but for phage X174, than swirling elution
n the numbers of R(%) values and t values of a Student’s and vigorous shaking (Fig. 1). Consequently it was de-
t-distribution (Anonymous, 1999). cided to include elution by ultrasound in the final modified
Statistical tests (ANOVA, Student’s t-test, Multiple method.
range test analysis using least significant differences be-
tween means using Student’s t-test) were performed using
the Statistical Package for Social Sciences (Anonymous, 100

1999). Differences were considered significant when 90

P < 0.05.
80

70
3. Results 60
% Recovery

50
3.1. Modifications assayed for the phage concentration
method 40

30
3.1.1. Amendment of the sample
The three reference bacteriophages studied (X174, MS2 20

and B40-8) adsorbed over 93% in the presence of 0.5 and 10

0.05 M MgCl2 . No significant differences (ANOVA, P > 0
0.05) were observed between the numbers of phages ad- N= 6 6 17 18 6 8

ØX174 (SE) ØX174 MS2 (SE) MS2 B40-8 (SE) B40-8

sorbed in the presence of the two concentrations of MgCl2 .
In contrast, the lower concentration, 0.005 M, of MgCl2 Phage
tested showed significantly (ANOVA, P < 0.05) lower ad-
Fig. 1. Comparison of swirling elution method (SE) and ultrasound for
sorption of bacteriophages ÔX174 and MS2. Consequently the elution of adsorbed laboratory reference phages. (䊐) Indicates the
it was decided to amend the samples with 0.05 M MgCl2 arithmetic mean and bars indicate the 95% confidence intervals of six
before the concentration step. independent experiments.
 J. Méndez et al. / Journal of Virological Methods 117 (2004) 19–25 23

3.1.4. Bacteriophages remaining in the filters The percentages of phages counted by platting the mem-
Even applying the best eluting solution and the best elut- brane filters varied from 13.5 to 21.6% for somatic col-
ing conditions, non-negligible numbers of phages, ranging iphages, 1.8–9.8% for F-specific bacteriophages, 0.8–8.8%
from 1 to 28% depending on the experiments and phages, for F-specific RNA phages, 11.6–19.6 for phages infecting
remained in the membrane filter. Consequently, it was rec- B. fragilis HSP40 and from 19 to 26.9% for phages infecting
ommended in the proposed method to include the count of strain RYC2056 of B. fragilis.
phages retained in the filter as part of the total number of The ISO validation procedure was applied to all phages
phages recovered as was recommended in the original meth- reported in Fig. 1 and results indicated neither volume ef-
ods (Sobsey et al., 1990; Sinton et al., 1996). fect, nor unacceptable relative standard deviation >0.5 for the
As a consequence of the results reported in this section, five groups of phages studied. The relative standard devia-
the final modified method was established as indicated in tions for the concentration efficiencies of the different phages
the Section 2. were: 0.19 for somatic coliphages, 0.23 for F-specific bacte-
riophages, 0.24 for F-specific RNA bacteriophages, 0.34 for
3.2. Evaluation and validation of the method bacteriophages infecting B. fragilis HSP40 and 0.28 for bac-
teriophages infecting B. fragilis RYC2056. Therefore, it was
The recovery efficiency of the method was evaluated for concluded that the method is reliable for volumes ranging
natural populations of phage, using phages partially purified from 100 ml to 1 l for somatic coliphages, F-specific RNA
from sewage as indicated in the material and methods sec- bacteriophages and bacteriophages infecting B. fragilis.
tion. Results expressed as the arithmetic mean of the recover- For brevity, only the results corresponding to the
ies for all volumes tested, since as indicated below there was F-specific RNA bacteriophage are reported in detail below.
no volume effect, are shown in Fig. 2. Recovery of somatic The sewage derived spiking material showed an excellent
coliphages was significantly higher (multiple range test anal- homogeneity both within and between vial since values of
ysis using least significant differences between means using T1 and T2 were 0.034 and 0.41 well below 5.99 and 3.84
Student’s t-test, P < 0.05) than the recoveries of F-specific, indicated in the ISO validation standard.
F-RNA bacteriophages and bacteriophages infecting B. Fig. 3 shows the plotted arithmetic average of the recover-
fragilis. In contrast, the recoveries of F-specific bacterio- ies (R(%)) plus the confidence intervals obtained by concen-
phages, F-specific RNA phages and phages infecting B. trating phages from 125, 250, 500 and 1000 ml. No signifi-
fragilis were similar (multiple range test analysis using least cant differences (ANOVA, P > 0.05) were detected. There
significant differences between means using Student’s t-test, was not volume effect in the percentage of phages detected
P > 0.05). by platting the membrane filter.
The calculation of the combined (all volumes) recoveries
(R(%)) gave an arithmetic mean average of value of 63.7 with
100 a standard deviation of 14.8. The relative standard deviation
90

80
100
70
90
60
80
% Recovery

50
70
40
60
% Recovery

30
50
20
40
10
30
0
N= 20 20 20 20 20
20
SOMCPH FTOTPH FRNAPH BFRPHHSP40 BFRPHRYC2056

Phage Group 10

0
Fig. 2. Average recovery of different groups of sewage derived phage from N= 5 5 5 5
1 l of water. (䊐) Indicates the arithmetic mean and bars indicate the 95% 125 250 ml 500 ml 1000 ml
confidence intervals. The arithmetic means correspond to the recoveries Volume
obtained in five tests performed at each of the four volumes. (SOMCPH),
somatic coliphages; (F-TOTPH) F-specific bacteriophages; (F-RNAPH), Fig. 3. Average recovery of sewage derived F-specific RNA phages from
F-specific RNA bacteriophages; (BFRPHHSP40) bacteriophages infecting four different volumes. (䊐) Indicates the arithmetic mean and bars indicate
B. fragilis strain HSP40 and (BFRPHRYC2056) bacteriophages infecting the 95% confidence intervals. Each value corresponds to five independent
of B. fragilis strain RYC2056. experiments.
24 J. Méndez et al. / Journal of Virological Methods 117 (2004) 19–25

100 phage groups were over 55%. Average recovery of somatic

90 coliphages was slightly over 80%, which is clearly supe-
rior to the original method method/s for 1 l of water which
80
is slightly below 50% (Hsu et al., 1998). The improvement
70 is due to the combined effect of the new eluent and the
60 ultrasound treatment. Results referring to elution of phage
X174 and the overall results for somatic coliphages may
% Recovery

50
seem contradictory, but it should be considered that the so-
40 matic coliphages are a very heterogeneous group (Ackerman
30 and Nguyen, 1983; Pedroso and Martins, 1995; Muniesa
et al., 1999) and that in municipal sewage the phages of the
20
family of X174 are not the most abundant (Muniesa, et al.,
10 1999).
0 The results of the standardised validation procedure in-
N= 20 21 20 22 20 32
SOMCPHc SOMCPH FRNAPHc FRNAPH BFRPHRYC2056c BFRPHRYC2056 dicate that the method is reproducible and that it can be
Phage Group easily implemented in routine laboratories with highly re-
Fig. 4. Average recoveries obtained by the laboratories participating in
producible results. The validation method used is simple
the interlaboratory exercise utilizing sewage derived phage. (䊐) Indicates and allows one to work with naturally occurring phages at
the arithmetic mean and bars indicate the 95% confidence intervals. concentrations approaching those found in nature. Some
For each phage group the first value (c) is the one obtained by the experiments performed with un-homogeneous phage sus-
central laboratory and the second, the average value obtained by the all pensions gave very variable recoveries (data not shown)
participating laboratories. (SOMCPH) somatic coliphages; (F-RNAPH)
F-specific RNA bacteriophages and (BFRPHRYC2056) bacteriophages
and consequently indicated that when working with low
infecting of B. fragilis strain RYC2056. numbers of phage, the homogeneity of spiking material is
essential for calculating recovery efficiencies of concen-
tration methods. This may explain some extreme results
gave a value of 0.23 well below 0.5 as recommended in the reported in Round Robin exercises when concentrations of
ISO standard. low numbers of phages were tested (Melnick et al., 1984).
In our opinion it will be worthwhile to assay the suit-
3.3. Results of the collaborative study ability of similar approaches for validating concentration
methods for animal viruses and oocysts of protozoa as
Results expressed as the arithmetic mean of the Cryptosporidium.
recoveries ± the confidence intervals obtained by all the
participating laboratories are shown in Fig. 4, with the
central laboratory recoveries isolated for comparison In the Acknowledgements
figure are also reported the values obtained. Though not
significantly different (Student’s t-test for of each group of This study was partially supported by REN2002-04035-
bacteriophages, P > 0.05), the average arithmetic mean re- C03-03/HID from CICYT, Ministerio de Ciencia y Tec-
coveries of somatic coliphages and F-specific phages were nologı́a, Spain; and by grant 1999SGR00023 and CeRBa
lower than those of the validation study, whereas recoveries (Biotechnology Center Reference) from the Generalitat de
of phages infecting Bacteroides were identical. Catalunya.
As well, the relative standard deviations were 0.27 for
somatic coliphages, 0.40 for F-specific RNA bacteriophages
and 0.36 for bacteriophages infecting B. fragilis RYC2056, References
all well below 0.5.
Ackerman, H.W., Nguyen, T.M., 1983. Sewage coliphages studied by
electronic microscopy. Appl. Environ. Microbiol. 45, 1049–1059.
4. Discussion Anonymous, 1995. ISO 10705-1: Water Quality. Detection and Enu-
meration of Bacteriophages. Part 1. Enumeration of F-specific RNA
Bacteriophages. International Standardisation Organisation. Geneva,
The method reported above introduces some modification Switzerland.
to the methods based in adsorption-elution to acetate–nitrate Anonymous, 1998. Standard methods for the examination of water and
cellulose ester membrane filters described first by Sobsey wastewater, 20th. ed. American Public Health Association, Washing-
et al. (1990) and modified further by Sinton et al. (1996). The ton, D.C.
Anonymous, 1999. Statistics program. SPSS for Windows, Version 10.0.6,
modified method demonstrated good recoveries of the three
1989–1999. Statistical Package for Social Sciences Inc. Chicago, IL.
groups of phages suggested as surrogate indicators from 1 l Anonymous, 2000. ISO 10705-2: Water Quality. Detection and Enumer-
of drinking water. Somatic coliphages were recovered more ation of Bacteriophages. Part 2. Enumeration of Somatic Coliphages.
efficiently that the others; however, average recoveries of all International Standardisation Organisation, Geneva, Switzerland.
 J. Méndez et al. / Journal of Virological Methods 117 (2004) 19–25 25

Anonymous, 2002. ISO 10705-3/FD: Water Quality. Detection and Enu- Melnick, J.L., Safferman, R., Rao, V.C., Goyal, S., Berg, G., Dahling,
meration of Bacteriophages. Part 3. Validation of Methods for Con- D.R., Wright, B.A., Akin, E., Stetler, R., Sorber, C., Moore, B.,
centration of Bacteriophages from Water. International Standardisation Sobsey, M.D., Moore, R., Lewis, A.L., Wellings, F.M., 1984. Round
Organisation. Geneva, Switzerland. Robin investigation of methods for the recovery of poliovirus from
Anonymous, 2003. ISO 10705-4: Water Quality. Detection and Enumera- drinking water. Appl. Environ. Microbiol. 47, 144–150.
tion of Bacteriophages. Part 4. Enumeration of Bacteriophages Infect- Mendez, J., Jofre, J., Lucena, F., Contreras, N., Mooijman, K., Araujo, R.,
ing Bacteroides fragilis. International Standardisation Organisation. 2002. Conservation of phage reference materials and water simples
Geneva, Switzerland. containing bacteriophages of enteric bacteria. J. Virol. Methods 106,
Borrego, J.J., Cornax, R., Preston, D.R., Farrah, S.R., McElhaney, B., Bit- 215–224.
ton, G., 1991. Development and application on new positively charged Muniesa, M., Lucena, F., Jofre, J., 1999. Study of the potential relationship
filters for recovery of bacteriophages from water. Appl. Environ. Mi- between the morphology of infectious somatic coliphages and their
crobiol. 57, 1218–1222. presence in the environment. Lett. Appl. Microbiol. 87, 402–409.
Grabow, W.O.K., 2001. Bacteriophages: update on application as model Pedroso, D.M.M., Martins, M.T., 1995. Ultra-morphology of coliphages
for viruses in water. Water SA 27, 251–268. isolated from water. Water Res. 29, 1190–1202.
Havelaar, A.H., Furuse, K., Hogeboom, E.H., 1986. Bacteriophages and Schulze, E., Lenk, J., 1983. Concentration of coliphages from drinking
indicators bacteria in human and animal faeces. J. Appl. Bacteriol. water by Mg(OH)2 flocculation. Naturwissenschaften 7, S: 612.
60, 255–262. Shields, P.A., Farrah, S., 1986. Concentration of viruses in beef extract
Hsu, F.C., Handzel, T.R., Lovelance, G., Stewart, J.R., Sobsey, M.D., by flocculation with ammonium sulphate. Appl. Environ. Microbiol.
1998. Improved methods to detect low levels of coliphages in wa- 51, 211–213.
ter by enrichment presence-absence and membrane filter methods. Sinton, L.W., Finlay, R.K., Reid, J.A., 1996. A simple membrane
In: Proceedings of the 1998 Water Quality Technology Conference, filtration-elution method for the enumeration of F-RNA, F-DNA and
American Water Works Association. Denver, CO. somatic coliphages in 100 ml water samples. J. Microbiol. Methods
IAWPRC Study Group on Health Related Water Microbiology, 1991. 25, 257–269.
Bacteriophages as model viruses in water quality control. Water Res. Sobsey, M.D., Schwab, K.J., Handzel, T.R., 1990. A simple membrane
25, 529–545. filter method to concentrate and enumerate male-specific RNA col-
Jofre, J., 2002. Bacteriophages as indicators. In: Bitton, G. (Eds.), Encyclo- iphages. J. Am. Water Works Assoc. 82, 52–59.
pedia of Environmental Microbiology. Willey, New York, pp. 354–363. Tartera, C., Jofre, J., 1987. Bacteriophages active against Bacteroides
Kott, Y., Ben Ari, H., 1968. Bacteriophages as marine pollution indicators. fragilis in sewage polluted waters. Appl. Environ. Microbiol. 53,
Rev. Int. Océanogr. Méd. ix, 207–217. 1632–1637.
Logan, K.B., Rees, G.E., Primrose, S.B., 1980. Rapid filtration of bacte- Vilaginés, P., Sarrette, B., Champsaur, H., Hughes, B., Dubrou, S., Joret,
riophages from large volumes of freshwater: evaluation of positively J.C., Laveran, H., Lesne, J., Paquin, J.L., Delattre, J.M., Oger, C.,
charged % microporous filters. J. Virol. Methods 1, 87–97. Alame, J., Grateloup, I., Perrollet, H., Serceau, R., Sinegre, F., Vi-
Lucena, F., Muniesa, M., Araujo, R., Jofre, J., 1995. Simple concentration laginés, R., 1997. Round Robin investigation of glass wool method
methods for bacteriophages of Bacteroides fragilis in drinking water. for poliovirus recovery from drinking water and sea water. Water Sci.
J. Virol. Methods 45, 121–130. Technol. 35, 445–449.