Journal of Virological Methods, 1( 1980) 87 -97 87

@ Elsevier/North-Holland Biomedical Press

RAPID CONCENTRATION OF BACTERIOPHAGES FROM LARGE VOLUMES

OF FRESHWATER: EVALUATION OF POSITIVELY CHARGED, MICROPOROUS
FILTERS

K.B. LOGAN, GE. REES, N.D. SEELEY and S.B. PRIMROSE

Department of Biological Sciences, University of Warwick, Coventry, CV4 7AL, England

(Accepted 28 January 1980)

Microporous, positively charged Zeta Plus 60 S filters were evaluated for bacteriophage recovery
from large volumes of natural water. A variety of phages adsorbed efficiently to the filter medium
at water pH levels below neutrality, but adsorption was reduced above pH 7. Adsorbed phages were
easily eluted with an arginine/l% beef extract solution, pH 9.0. A concentration procedure suitable
for the field isolation of bacteriophages from large volumes of river water was devised. The procedure
involves, (1) prefiltration through 10” cartridge depth filters, (2) adjustment of water pH to pH
5.5-6.0, (3) adsorption of the phages on to Zeta Plus 60 S filters, (4) elution of bound phage in a
small volume of eluent, (5) secondary concentration by ultrafiltration of the resulting eluates. Using
this procedure, bacteriophages in 65 1 of river water were concentrated to 35 ml with recoveries in
the range 50-60%.

1NTRODUCTlON

At present, little is known about the distribution and biological significance of viruses
in the aquatic environment. Bacteriophages and animal viruses frequently can be detected
by direct plating of faecally polluted water. However, studies on virus distribution in un-
polluted waters require a method for efficiently concentrating and recovering low numbers
of viruses from relatively large volumes of water. Recent concern about the potential
hazards of human enteric virus contamination of drinking waters has stimulated the devel-
opment of appropriate methods.
Although numerous methods have been developed for concentrating viruses from large
volumes of water, the most successful rely on virus adsorption to certain types of micro-
porous filters, followed by elution in a small volume of fluid (for review see Wallis et al.,
1979). However, efficient virus adsorption to the commonly used filter media occurs only
if the water is first acidified and/or multivalent cations are added. Thus Sobsey et al.
(1973) found that enterovirus adsorption was maximal at pH 3.5 and Seeley and Primrose
(1979) obtained maximal adsorption of a wide range of bacteriophages at pH 3.8. To
elute enteroviruses adsorbed to filters, Sobsey et al. (1973) used glycine buffer pH 11 S,
88

but Seeley and Primrose (1979) found that these conditions inactivated some bacterio-
phages, and that 3% beef extract, pH 8.5, was more suitable for the recovery of adsorbed
phages.
For studies on the distribution of viruses in the aquatic environment the above method
of adsorption and elution has a serious defect: it only concentrates acid resistant viruses.
At present there are few reports of efficient methods of virus concentration from large
volumes of water at neutral pH. Farrah et al. (1978) found that low concentrations of
AU, in tapwater formed floes that efficiently adsorbed poliovirus at pH 7. Adsorbed
viruses were recovered subsequently by elution in a small volume of glycine buffer, pH 11.
However, Seeley and Primrose (1979) obtained poor recoveries of phages with an alumi-
nium hydroxide floe technique. Recently, Sobsey and Jones (1979) reported that polio-
viruses could be adsorbed from tapwater by passage through certain types of positively
charged, microporous filters without any prior adjustment of pH or addition of salts.
In this report we show that these positively charged filters can be used to concentrate
bacteriophages from river water, but that pH affects the efficiency of phage recovery.

MATERIALS AND METHODS

Bacteria and bacteriophages

The bacteriophages used in this study, and their normal plating hosts, are described in
Table 1. Stocks of all bacteriophages were prepared by lysis of liquid cultures, clarified
by centrifugation at low speed and stored at 4°C.

TABLE i

Properties of bacteriophages

Phage Phage properties Bacterial host

MS2 Spherical, male-specific, single-stranded RNA Escherichia coli HfrH

T2 Long tail, contractile, double-stranded DNA Escherichia coli SPA0
8X174 Spherical, single-stranded DNA Escherichia coli C603
P22 Short tail, non-contractile, double-stranded DNA, Salmonella typhimurium
temperate his G46
PR4a Spherical, short tail, plasmid-specific, internal Escherichia coli W3 110
lipid layer, double-stranded DNA (RPD
86a Spherical, short tail, outer lipid layer, double- Pseudomonas phaseoiicola
stranded RNA HBIOYa

a Generously supplied by Dr. Dennis Bamford, Department of Genetics, University of Helsinki, Fin-
land. All other phage and bacterial stocks were from the University of Warwick culture collection.
 89

Media and solutions

LC medium (pH 7), used for the cultivation of Pseudomonas phaseozicola, contained
(g/l distilled water): Bacto-tryptone, 10.0; yeast extract, 5.0; NaCl, 5.0; glucose, 4.0;
MgSO‘, - 7Hz0, 0.05. All other media and solutions have been described by Seeley and
Primrose (1979).

Phage assay procedure

Bacteriophages were assayed by the soft agar overlay method (Primrose et al., 1968).

Prior to virus adsorption, all natural waters were clarified by filtration through a
5 pm porosity AMF-CUNO, Micro-Klean II, 10” cartridge depth filter (V.A. Howe
and Co., Ltd., 88, Peterborough Rd., London) followed in series by two 10” polypropy-
lene fibre-wound sediment removal filters of 1 pm porosity (IMI Mouldings Ltd., Fleming
Way, Crawley, Sussex).

Adsorbent filters

AMF-CUNO Zeta Plus (series S) positively charged microporous filters were obtained
from Flowtech Fluid Handling Ltd., Deacon Way, Reading, Berks.

Bacteriophage concentration

Samples of pre-filtered pond or river water were collected in 70 1 glass-fibre mixing

vessels. Where specified, the pH was adjusted by slow addition of 1 N HCl or 1 N NaOH
with constant stirring. For laboratory tests, sample volumes of pond water were then
seeded with a small volume of stock bacteriophages of known titre. Thorough mixing
of both acid and phage was accomplished with a portable 12 V stirrer (Seeley et al.,
1979) and pH levels were monitored using a portable pH meter (‘Expand-Mate’, Beck-
man-RIIC Ltd., Glenrothes, Fife). After mixing, two 5 ml aliquots were removed to
determine the input concentration of bacteriophages. The samples were then pumped
through adsorbent filters and the filtrates collected and assayed for bacteriophages.
For tests involving small volumes (< 1 litre), water was filtered under vacuum through
47 mm diameter adsorbent filters. For larger sample volumes, a 12 V Stuart No10 centri-
fugal pump (Stuart Turner Ltd., Henley on Thames, Oxon.) was used to pump water
through a 273 mm dia. adsorbent filter held in the perspex filter holder described by
Seeley et al. (1979).
90

Elution

The eluents used are specified in the text (see Results). Elution from the 273 mm dia.
filter was accomplished by recirculation of 500 ml of eluent for 15 min at ambient
temperature using a high-speed peristaltic pump (H.R. Flow Inducer, Watson-Marlow
Ltd., Falmouth, Cornwall, England). The pH of the eluate was immediately adjusted to
pH 7.4 by slow addition of 1 N HCl with constant stirring. After measuring the total
volume, samples of eluate were assayed for infectious phage.
Where specified, further concentration of phage in the eluate was achieved by ultra-
filtration in a Bio-Fiber 80 beaker (Bio-Rad Laboratories, Richmond, CA, U.S.A.).

RESULTS

Choice of adsorbent filter

Zeta Plus positively charged microporous filters are composed of cellulose fibres,
modified anion exchange resins and inorganic (diatomaceous earth) filter aids. The
material is supplied in a variety of porosity grades and resin types, but only the S series
is autoclavable. Of a number of the S grades tested, the 60 S grade (nominal porosity
0.5 pm) gave the best combination of practical flow rate and efficient virus adsorption
(data not shown) and was selected for use in this study.

Effect of pH on bacteriophage adsorption to Zeta Plus 60 S filters

Our initial aim was to determine the efficiency of phage adsorption to Zeta Plus filters
throughout the observed pH range of natural waters in our area (pH 5.558.0). 65 1 vol-
umes of pre-filtered pond water were adjusted to the pH values shown in Table 2 and
seeded with a mixture of bacteriophages (minimum 100,000 p.f.u. of each of T2, MS2,
P22 and PR4). After thorough mixing, aliquots were removed and assayed for the input
phage concentration. The 65 1 of seeded water were passed through separate 273 mm dia.
fnters at an initial flux of 2 mlsmin-’ -cm-*, and the filtrates assayed for bacteriophages.
The results are shown in Table 2.
All the bacteriophages tested efficiently absorbed to the filters in the pH range 5.5 -7.0
but at higher pH values the efficiency of bacteriophage adsorption was markedly reduced.
Thus the ambient pH of any natural water under test would clearly be an important
factor in the overall efficiency of virus concentration using Zeta Plus filters. Since the pH
of rivers is often higher than 7, particularly in the summer months, adjustment of the pH
of the water to a predetermined value would be essential to give consistent results and
maximum phage recovery.
 91

TABLE 2

Effect of pH on bacteriophage adsorption to Zeta Plus 60 S filters

Adsorption Total input phage adsorbed (%)

PH
Bacteriophage
T2 MS2 P22 PR4

5.5 99.8 99.9 99.7 99.9

6.0 99.3 99.8 99.3 99.7
6.5 96.1 99.8 98.2 97.6
7.0 92.2 97.4 95.5 96.8
7.5 73.2 66.2 68.0 82.0
8.0 28.6 21.9 23.5 29.7

Optimum pH value for recovery of bacteriophages from natural waters

The results
presented in Table 2 show that bacteriophages efficiently adsorb to Zeta
Plus filtersat pH 7. Initially this was considered to be the ideal pH value for it would
give minimal bacteriophage inactivation. However, tests in the field revealed that phage
recovery was considerably reduced when larger volumes of river water were filtered at
neutral pH. To investigate this ‘volume effect’ on bacteriophage adsorption we examined
the appearance of unadsorbed phage in the filtrate at different pH values.
65 1 volumes of prefitered pond water were adjusted to different pH values, seeded
with a mixture of bacteriophages and filtered as before, except that successive 5 1 volumes
of filtrate were assayed for phage. The results obtained with bacteriophage T2 are pre-
sented in Fig. 1 and again show that adsorption efficiency decrease with increasing pH
values. Furthermore, even at pH 7, increasing the sample volume led to a decrease in
phage adsorption. Similar results were obtained with bacteriophages MS2, P22 and PR4.
Clearly, when determining the optimum pH for phage recovery on Zeta Plus filters we
must balance the increased adsorptive capacity of the filters at pH values below 7 against
the sensitivity of viruses to increased acidity. In field trails we have found that adjustment
of the water to pH 5.5-6.0 gives greater than 80% bacteriophage adsorption when 100 1
volumes of natural freshwater are processed.

El&ion of bacteriophages adsorbed to Zeta Plus filters

Since phages do not adsorb efficiently to Zeta Plus filters under alkaline conditions
(Table 2) we tested a range of buffers of increasing pH as possible eluents. Results obtain-
ed in small-scale experiments with 200 ml volumes of T2-seeded pond water, and 100 ml
of glycine or arginine buffered eluents are shown in Table 3. Similar results were obtained
with bacteriophages MS2, P22 and PR4 (data not shown). The efficiency of elution not
 92

pH6.0

pH 7.0

50 pH 7.5

2b 4’0 6’0
Volume Filkote ( liwes )
Fig. 1. Effect of pH and sample volume on adsorption of bacteriophage T2 to Zeta Plus 60 S filters.
 93

TABLE 3

Effect of pH on elution of bacteriophage T2 adsorbed to Zeta Plus filters

Elution Adsorbed phage in eluate (%)

PH
5 0 mM glycine 50 mM arginine

7.0 0 0
7.5 0 0
8.0 1.1 12.8
8.5 4.7 22.8
9.0 6.6 34.2
9.5 30.9 69.8
10.0 41.7 86.8

only increased with increasing pH but was also influenced by the chemical nature of the
eluent. Even with arginine buffered eluents, relatively high pH values (pH >9..5) were
required to recover 80% of the input phages. However, some bacteriophages are inactivat-
ed above pH 9 (Seeley and Primrose, 1979), and so we examined a variety of other
eluents. The results presented in Table 4 show that addition of protein to the arginine
buffer significantly increased the efficiency of elution at pH 9, and that 1- 3% beef ex-
tract is optimal. Recoveries greater than 100% are presumably due to dispersion of aggre-
gates of phage present in the initial inoculum.

Phage recoveries from natural waters using optimum conditions

From the foregoing results we selected the optimum conditions for each step in the

TABLE 4

Elution of bacteriophages MS2 and T2 adsorbed to Zeta Plus filters

Additions to Adsorbed phage in eluate (%)

50 mM arginine
(PH 9) MS2 T2

None 13.9 36.0

0.1% beef extract (Oxoid) 72.3 55.0
1 .O%beef extract 111.9 98.8
3 .O%beef extract 112.2 108.1
3.0% Bacto-peptone (Difco) 66.3 77.1
3.0% Casamino acids (Difco) 12.4 40.1
0.35 M NaCl 2.7 21.6
94

procedure and combined them to produce an overall bacteriophage concentration system.

To test the efficiency of this procedure 65 1 volumes of pond-water were pre-filtered
and collected in large mixing vessels. The water was adjusted to pH 5.5-6.0, seeded with
a variety of bacteriophages and passed through a 273 mm diameter Zeta Plus 60 S filter.
Adsorbed phages were eluted with 500 ml 50 mM arginine, 1% (w/v) beef extract pH 9.0,
and the eluate immediately adjusted to pH 7.4 by slow addition of 1 N HCl. The results
(Table 5) show that with the exception of @X174, more than 70% of seeded phages
were concentrated and recovered by this procedure. Particularly significant is the high
recovery of @6, which is acid-labile and would be inactivated during concentration
by the method of Seeley et al. (1979). Recoveries of bacteriophage @X174 consistently
were low, apparently because of poor adsorption to the filter media. The reasons for this
are not clear but essentially similar results have been described by others (Goyal et al.,
cited in Wallis et al., 1979).

TABLE 5

Recoveries of test phages using optimum conditions of virus concentration

Phage type Phage adsorbed (%) Overall recovery of input

phage in eluate

MS2 98.9 93.2

T2 99.3 87.4
@X174 32.8 23.6
P22 98.3 81.8
PR4 94.2 79.5
@6 82.8 78.5

TABLE 6

Concentration of indigenous phages from 65 litres of River Avon water

Host Average titre/ml Titre/ml Overall phage

raw water eluate recovery (%)

Escherichia coli W3 110 90 8.4 X 10” 57.6

Klebsiella pneumoniae 889 110 1.3 x 10S 61.4
Salmonella typhirnurium his G46 1 1.1 x IO3 58.3
Aeromonas hydrophila 0 1.8 x lo*
Aeromonas salmonicida 0 2.1 x loz
Hyphomicrobium 0 8.3 X 10’
Rhodopseudomonas blastica 0 50
 95

Concentration of indigenous bacteriophages from river water

To test the effectiveness of our concentration method for field use, indigenous phages
in 65 1 of river water (River Avon, Warwick) were concentrated 1850-fold to 35 ml by
the procedure shown in Fig. 2. The eluates were titered on a variety of bacteria and the
results are shown in Table 6. Only with the three enteric bacterial hosts could phage be
detected by direct plating of the raw water. In all field trails to date, recoveries of phages
for these hosts have been in the range 50-60%. Therefore we consider this method to be
satisfactory for studies on phage distribution in the aquatic environment.

DISCUSSION

We have shown that Zeta Plus 60 S filters are effective adsorbents of bacteriophages,
and have subsequently developed a procedure for concentrating and recovering low
numbers of phages from large volumes of natural freshwater. We consider that this me-
thod offers a number of advantages over existing techniques of virus concentration. For
example, the use of these filters does not require the addition of multivalent cations, or
the drastic alterations in pH encountered in other virus concentration procedures (Sobsey
et al., 1973; Farrah et al., 1976). In addition, Zeta Plus filter sheets are considerably less

Raw water

#
pre-filters in
series (5 pm, 1 pm, 1 pm)
adjustment of
1
p mixing vessel
pH to 5.5 -6.0 I

+
27 3 mm diameter
Zeta Plus adsorbent

filtrate to / filter \Elut e with 500 ml

waste arginine/beef extract (pH 9)

c
Ultrafiltration

Final eluate volume 35 ml

Fig. 2. Procedure for concentrating bacteriophages from natural waters using microporous, positively
charged Zeta Plus filters.
96

expensive than the cartridge ‘in-depth’ or pleated membrane filters used by other workers
(Farrah et al., 1976; Payment et al., 1976; Sobsey et ai., 1977).
The absorption of viruses to falters is thought to be partly due to electrostatic inter-
actions between the virion and filter surfaces (Mix, 1974). Electrophoretic studies have
shown that most viruses have isoelectric points below pH 7, and hence bear a net negative
charge at neutral pH (Brinton and Lauffer, 19593. Sobsey and Jones (1979) have shown
that the Zeta Plus 60 S fnter medium has an isoelectric point between pH 5 and 6. In-
creasing the ambient pH above these values would therefore reduce the net positive
charge, and lead to a decrease in the number of virus adsorption sites within the filter
matrix. Thus we have observed that at pH levels above neutrality, the efficiency of
bacteriophage adsorption is considerably reduced. However, using our virus concentration
technique, only moderate adjustments of pH are necessary to maximize virus adsorption,
even when alkaline raw waters are to be processed.
A second factor that might affect the efficiency of phage adsorption to Zeta Plus
filters is the amount of organic material in the water sample under test. It seems likely
that not only viruses but all other pioteinaceous materials in the water will adsorb to the
filter medium. As there must be a finite number of adsorption sites within the filter
matrix, then clearly high levels of organic contamination in the water will reduce the
efficiency of bacteriophage adsorption. We have described a ‘volume effect’ on phage
adsorption to Zeta Plus filters whereby increasing the sample volume led to a progressive
decrease in the amount of phage adsorbing to the filter. Throughout this study we have
used natural freshwater for laboratory tests. It is likely therefore that ‘breakthrough’ of
phage occurs when all the adsorption sites within the filter matrix become saturated with
adsorbed organic material. We found that this ‘breakthrough’ of phage is reduced by
filtration at pH levels slightly below neutrality, presumably by increasing the number of
positively charged adsorption sites on the filter.
Adsorption of bacteriophage to Zeta Plus filters is reversible. Adsorbed phages are
easily recovered using arginine/beef extract eluents at pH 9. Because adsorbed bacterio-
phages are effectively recovered from Zeta Plus fileers using moderately alkaline eluents,
phage inactivation should be minimized.
Field tests showed that 50-60% of indigenous bacteriophages could be concentrated
from 65 1 of river water using the procedure outlined in Fig. 2. Since this procedure is so
effective we have constructed a portable concentrator for field use capable of concentrat-
ing up to 500 litres of river water. Such a concentrator should facilitate studies on the dis-
tribution of viruses in the aquatic environment, particularly since preliminary results in
our laboratory indicate that it can also be used to recover animal viruses from water.
 91

REFERENCES

Brinton, C. and M. Lauffer, 1959, in: Electrophoresis, ed. M. Bier (Academic Press Inc., New York).
Farrah, S.R., C.P. Gerba, C. Wallis and J.L. Melnick, 1976, Appl. Environ. Microbial. 31, 221.
Farrah, S.R., S.M. Goyal, C.P. Gerba, C. Wallis and J.L. Melnick, 1978, Appl. Environ. Microbial. 35,
624.
Mix, T.W., 1974, Develop. Ind. Microbial. 15, 136.
Payment, P., C.P. Gerba, C. Wallis and J.L. Melnick, 1976, Water Res. 10,893.
Primrose, S.B., L.R. Brown and C.E. Dowell, 1968, J. Virol. 2, 1308.
Seeley, N.D. and S.B. Primrose, 1979, J. Appl. Bacterial. 46, 103.
Seeley, N.D., G. Hallard and S.B. Primrose, 1979, J. Appl. Bacterial. 47, 145.
Sobsey, M.D. and B.L. Jones, 1979, Appl. Environ. Microbial. 37,588.
Sobsey, M.D., C. Wallis, M. Henderson and J.L. Melnick, 1973, Appl. Microbial. 26,529.
Sobsey, M.D., C.P. Gerba, C. Wallis and J.L. Melnick, 1977, Can. J. Microbial. 23, 770.
Wallis, C., J.L. Melnick and C.P. Gerba, 1979, Ann Rev. Microbial. 33, 413.