Ghost cells as sorption matrix for virus concentration from water.

Vol. 47. No. 6APPLIED AND ENVIRONMENTAL MICROBIOLOGY. June 1984. p. 1337-1340
0099-2"40/84/061337-04$02.00/0
Copyright © 1984. American Society for Microbiology
Ghost Cells as Sorption Matrix for Virus Concentration from Water
ROBERT ARMON.' YEHUDA KOTT,- AND ISHAK NEEMANI*
DeparOtments of Food Engineer-ing and BiotecIhno/ogv' anId En vironmenta/l anld Water Resou (res Enginzeer-ing,2 Te(cnion-
Israe/l In.stitiute o)f Te(hno/ogy, Haifi 32000, Isr-ael
Received 27 December 1983/Accepted 9 March 1984
Pig erythrocyte membranes were used as adsorbent material for the concentration of bacteriophage 4x-
174, MS-2., and f2 from 5 ml of saline solution. The adsorption was carried out at pH 3.5. and the elution was
carried out at pH 7.0. Compared with adsorption on 3% beef extract, bacteriophage adsorption on
erythrocyte ghost cells yielded 93 to 100%, and elution was 92 to 100%, whereas beef extract organic
flocculation yielded adsorption of 0 to 98.8% and elution of 61 to 86.6%. The same procedure but with
poliovirus LSC-1 gave 100% adsorption and 91 to 129% elution.
The adsorption-elution method (6) with membrane filters
is an essential tool in virus detection methodology. Recovery
of different viruses from water by this method depends on
four major factors: water sources, adsorption characteris-
tics, eluent composition, and pH. Glycine, beef extract,
tryptose phosphate broth, tryptic soy broth, and fetal calf
serum at pH 9.0 to 11.0 have all been used (11). The recovery
percentage varied with each eluent, virus, and water sample
volume.
Water solution of beef extract at 3% is the most commonly
used eluent due to its property of forming an organic floc at
pH 3.5 (15). However, recent experiments have shown that
3% beef extract (used as eluent; pH 9.0) failed to produce an
organic floc for enterovirus concentration when the pH was
lowered to 3.5. This phenomenon was confirmed by others
(R. Vasel and Y. Kott, personal communication) and is
probably due to a new process in beef extract production.
Concentration of viruses using erythrocytes has been
reported and is based on utilization of hemagglutination
interaction (13, 19) or passive hemagglutination (23). How-
ever, whole erythrocytes cannot be used in water due to
hemolysis at low osmotic pressure and loss of hemagglutina-
tion properties (12).
A search for a new organic adsorbent with the same or
better properties as beef extract led us to examine pig
erythrocyte ghost membranes.
The purpose of the study was to investigate the sorption
capability of ghosts cells for concentration of phage and
poliovirus from water.
MATERIALS AND METHODS
Pig erythrocyte membranes (ghosts). Blood samples were
collected at the Nazareth (Israel) slaughterhouse (mixed
blood groups) in bottles containing Alsever solution (19), 1:1
(wt/vol), and centrifuged at 1,750 x g for 10 min. The
supernatant was removed by aspiration.
The packed erythrocytes were washed three times with
0.85% saline and centrifuged at 367 x g for 10 min. Washed
packed erythrocytes were lysed by pouring into 10 to 20
volumes of distilled water, CO, saturated and well mixed,
until hemolysis was complete (14). Ghost cells were collect-
ed from the large volume of distilled water by centrifugation
using a Sorvall KSB-R continuous flow system, washed two
times with cold 0.2% NaCI, and fixed.
Corresponding author.
Fixation was achieved by suspending the ghost cells in 3%
formaldehyde-0.01 M phosphate-buffered saline (for 1 liter
of whole blood, 5 liters of 3% formaldehyde and 5 liters of
0.01 M phosphate-buffered saline were used) overnight at
37°C with continuous shaking. The ghosts cells were concen-
trated by centrifugation and washed with 0.85% saline three
times to remove excess formaldehyde. Before use, the
erythrocyte membranes (ghosts cells) were washed twice
more with 0.85% saline. To prevent microbial contamination
of the erythrocyte membranes, 0.01% sodium azide was
added. Ghost cells prepared by this method retained their
adsorptive properties for 4 years (at 4°C with 0.2% sodium
azide).
Phage. Bacteriophages Xx-174 (supplied by A. Razin,
Hebrew University, Jerusalem, Israel), MS-2, and f2 (from
the laboratory stock) stored at 4°C were used as experimen-
tal phages.
Escherichia coli CR,U was used as the host bacteria for (x-
174, and E. (.0li K Hfr was used for MS-2 and f2. Phage
titration was performed by the agar layer technique (1) using
media described by Davis and Sinsheimer (7) for MS-2, by
Loeb and Zinder (17) for f2, and by Sinsheimer (22) for 4bx-
174. Phage stock (108 PFU/ml) was kept at 4°C. Phage
dilutions were performed in 0.01 M phosphate-buffered
saline, pH 7.2.
Virus and virus assay. Poliovirus LSC-1 was grown and
assayed in Buffalo green monkey (BGM) cells (21).
Beef extract. Beef extract, 3% (wt/vol), from the followirlg
sources was used: (i) Lab-Lemco powder, Oxoid Ltd.,
London, England (code L-29); (ii) beef extract, Difco Labo-
ratories, Detroit, Mich. (code 0126-01); and (iii) beef extract
powder, Inolex Corp. (code 49-137-03).
Protein assay. Ghost cells were kept overnight in 0.1 M
KOH (to establish solution hemogeneity), and the various
beef extracts were examined for protein concentration by
the method of Lowry et al. (18).
Ghost cell quantitative estimation. Ghost cells were mea-
sured quantitatively by hematocrit results expressed as
percent (vol/vol) (5).
Adsorption-elution method. A three-step adsorption-elu-
tion method was used in this study.
Step 1. Phage (0.5 ml) at pH 7.0 was introduced into test
tubes containing 4.5 ml of erythrocytes (ghost cells).
Step 2. About 0.1 ml of 0.1 N HCI was added to lower the
pH to 3.5, and the tubes were shaken for 10 min. The tubes
were then centrifuged for 10 min to sediment the erythrocyte
membranes.
1337

1338 ARMON, KOTT, AND NEEMAN
TABLE 1. Yield comparison of adsorption at various pHs of
phages with ghost cells and beef extract
Adsorption efficiency (%)
pXx-174 MS-2 f2H_
Erythrocyte Beef Erythrocyte Beef Erythrocyte Beef
ghost cells" extract' ghost cells extract ghost cell extract
3.5 100 0 86 0 97 4
4.5 93.4 0 31 0 98 0
5.5 8.7 0 21.4 0 80 0
6.5 0 0 0 0 77 0
7.0 0 0 0 0 6 0
' The pig erythrocyte ghost cell concentration was 18% (vol/vol).
b Beef extract (Difco; catalog no. 0126-01) was used at a 3%
concentration.
Step 3. The pH was increased to 7.0, again using ca. 0.1 ml
of 0.1 N NaOH, and the pellet of ghosts was resuspended
homogeneously with a magnetic stirrer.
Phage assay. The phages were assayed at three stages
during adsorption-elution: (i) immediately after seeding; (ii)
after lowering the pH to 3.5 and centrifuging, from the
supernatant; and (iii) after the pH was increased to 7.0 and
the solution was mixed well.
Adsorption-elution calculations. The results were calculat-
ed as follows. The phage in step 1 is the 100% value (0%
adsorption). Adsorption is given as 100 - (b x 100)/a, and
elution is given as (c x 100)/l where b is the phage count in
the supernatant of step 2, and c is the phage count in step 3.
Results are an average of two experiments if not stated
otherwise.
RESULTS
Adsorption of 4x-174, MS-2, and f2 on erythrocyte mem-
branes compared with 3% beef extract at different pHs. The
adsorption pH was examined for Xx-174, MS-2, and f2
(Table 1) using 18% erythrocyte membranes and 3% beef
extract (Difco) at a pH range of 3.5 to 7.0. At pH 3.5, 86 to
100% of all phages were adsorbed using erythrocyte ghosts
as adsorbent material, whereas only 4% of phage f2 was
adsorbed by 3% beef extract.
Adsorption of 4)x-174, MS-2, and f2 on erythrocyte mem-
branes at different concentrations. Xx-174, MS-2, and f2 were
adsorbed on several erythrocyte membrane concentrations
at pH 3.5, and ghost cell concentrations varied from <0.5 to
32%. As can be seen from Table 2, 4% was shown to yield
optimal adsorption.
Adsorption and elution of Xx-174, MS-2, and f2 using
erythrocyte membranes was performed by change in pH, as
is done with the organic flocculation technique. Adsorption
of Wx-174 was 93%, that of MS-2 was 96%, and that off2 was
100% (Table 3). The elution was 98 to 102%. A comparison
of adsorption-elution of Xx-174, MS-2, and f2 using 3% beef
extract was done. The best results were obtained for the
three phages with the Oxoid beef extract, whereas the Inolex
and Difco products were inferior. At the same time, erythro-
cyte membranes in distilled water or saline yielded adsorp-
tion of 98 to 100% and elution of 92 to 126%, which were
higher than those achieved with 3% beef extract alone. Table
4 gives the results of adsorption-elution experiments per-
formed with three 3% beef extract solutions examined with
and without the addition of 13% erythrocyte membranes.
The results show that the addition of membrane caused a
significant increase in the adsorption capacity of the beef
TABLE 2. Phage adsorption with various concentrations of
erythrocyte ghost cells at pH 3.5
Erythrocyte ghost Adsorption efficiency (%)
cell concn (%) Ox-174 MS-2 f2
32 96 ND ND
26 94 95 100
23 94 ND ND
18 ND" 94 ND
12 ND 93 95
11 90 ND ND
5 86 ND ND
4 ND 78.8 87
3 84.6 ND ND
2 79 ND 62
1 ND 14 ND
0.9 ND ND 53
<0.5 75 ND ND
"ND, Not done.
extract in all three phages examined. However, there was no
significant change in elution capacity.
Adsorption-elution of the phages was examined at 4, 24,
and 37°C. Xx-174 showed the best adsorption, whereas MS-2
and f2 exhibited less adsorption capability, and at each
temperature the elution percentage differed.
Adsorbent material protein content. It has been suspected
that commercial beef extracts used in water virology do not
contain the same protein concentration or sorption proper-
ties. It was therefore interesting to compare the various
products with ghost cell membranes. The Inolex and Oxoid
(3% [wt/vol]) samples had 20.4 and 18.8 mg of protein per ml,
respectively, and the Difco (3%) sample had 7.8 mg of
protein per ml. (The Difco sample was in paste form,
whereas the others were powders.) For ghost cells at a 50%
(vol/vol) concentration the protein value was 17.7 mg/ml,
which is close to the Inolex and Oxoid (3%) protein values. It
is worthwhile to note that the erythrocyte ghost cell concen-
tration of 13%, which was used most in this study, had 7.2
mg of protein per ml (similar to Difco 3% beef extract).
Adsorption-elution of poliovirus LSC-1 in saline with pig
erythrocyte membranes. Poliovirus LSC-1 was used as a
representative model for enteroviruses in adsorption-elution
experiments with pig erythrocyte membranes. A high ad-
sorption of 100% and elution of 91 to 129% were obtained
(Table 5).
DISCUSSION
T-bacteriophages are known to be sensitive and may suffer
osmotic shock in deionized water (1). Bacteriophages such
as Xx-174, MS-2, and f2 are often used as models for
enterovirus studies in water (2, 3, 11, 16, 20). These phages
show higher resistance to adverse environments and are
often mentioned as viral pollution indicators (16), as they
grow quickly and are inexpensive to assay.
As the various bacteriophages often show sensitivity to
ionic strength changes, the results in Table 3 indicate that
despite the isotonic solutions used, a slight inactivation or
aggregation of phages occurred, as noted in control experi-
ments (saline only). It appeared that erythrocyte membranes
protected the phages from rapid changes in pH and ionic
strength.
In an attempt to substitute for 3% beef extract by using the
organic flocculation method, pig erythrocyte membranes
APPL. ENVIRON. MICROBIOL.

VIRUS CONCENTRATION FROM WATER 1339
TABLE 3. Phage adsorption-elution with erythrocyte ghost cells
PFU/ml
pH" XX-174 MS-2 f2
Controlb Expt' Control Expt Control Expt
7.0 3.5 x 103 5.8 x 2.1 x 102 7.8 x 102 3.6 x 102 3.6 x 102
3.5 3.3 x 103 4.3 x 102 2.2 x 102 0.3 x 102 3.8 x 102 0
7.0 3.7 x 103 5.7 x 103 1.4 x 102 8.0 x 102 3.5 x 102 3.7 x 102
% Adsorption 6 93 0 96 0 100
% Elution 105 98 66 102 97 102
a The pH changes were made by the addition of 0.1 N HCl and 0.1 N NaOH, respectively.
b Phage in saline.
' 13% erythrocyte ghost cells in saline.
were investigated and found to adsorb and elute phages (and
poliovirus LSC-1) under the same conditions.
As erythrocyte membranes are composed of 52% protein
of two major groups, glycoproteins (13.8%) and other pro-
teins (38.2%) (4, 10), it is believed that the relatively high
protein content distributed over the entire surface area
promoted the erythrocyte ghost cell adsorption-elution of
the phages. It could be thought that pig erythrocyte mem-
branes, as used for adsorption-elution of viruses, may be-
have as in hemagglutination interaction, but the data in
Tables 1 and 3 show that the behavior of the ghost cells is not
due to hemagglutination (12), which is a specific interaction
requiring strict control of such conditions as pH, ionic
strength, and temperature. In preliminary experiments with
X~x-174, MS-2, and f2, no agglutination of pig erythrocytes
was found. Results obtained by us showed that protein
quality rather than protein quantity has a much more impor-
tant effect on phage adsorption. It should be mentioned,
however, that different dilutions of erythrocyte membranes
did not yield a definite linear ratio when plotted against
protein concentration.
Pig erythrocytes have a lower electrophoretic mobility
than other erythrocytes (9) due to their higher positive net
charge, which is caused by lower sialic acid content. This
property possibly accounts for their good adsorbent proper-
ties for bacteriophages and viruses at pH 3.5. As indicated in
Table 1, at pH 6.5 to 7.0 erythrocyte ghost cells showed no
adsorption properties for phages, due to net decrease in the
positive charge. An additional explanation may be that the
adsorbent properties partially depend on van der Walls or
hydrophobic forces. Preliminary neuraminidase treatment
may lower the sialic acid content of the erythrocyte mem-
branes, resulting in a more positive charge that can adsorb
viruses better (9).
Differing water quality (2) does not favor the use of whole
erythrocytes due to hemolysis. Smith and Courtney (23)
tried passive hemagglutination to concentrate phage MS-2
but had only partial success. This method requires prelimi-
nary preparations, including antibodies, which makes it an
inconvenient method for routine use. Consequently, eryth-
rocyte membranes were used as adsorbent material for the
three bacteriophages, (x-174, MS-2, and f2, as well as for
one enterovirus.
The process of adsorption-elution described in the experi-
mental procedures was based on the method described by
Katzenelson et al. (15) using 3% beef extract organic floccu-
lation. The results obtained in this study with erythrocyte
ghost cells were higher and more stable compared with 3%
beef extract organic flocculation when bacteriophages were
used. In addition, the membranes also gave high adsorption
even at low concentrations (Table 2). The percent elution of
bacteriophages from erythrocyte membranes varied when
the experimental conditions changed. Low elution results
depend on the physicochemical conditions existing in these
experiments. A rise in temperature generally increases ad-
sorption due to the unfolding of proteins on the adsorbent
surface, which is an endothermic reaction (24). However, in
the current study only slight differences in the results were
obtained due to changes in temperature.
A combination of erythrocyte membranes with 3% beef
extract improved adsorption significantly (Table 4) but re-
sulted in lower elution, which can probably be attributed to
the interaction between beef extract and ghost cell proteins.
The results obtained up to now show that the use of pig
erythrocyte membranes can lead to a very efficient recovery
of bacteriophages and poliovirus (Tables 3 and 5). It is felt,
however, that further studies would show how to apply it to
virus recovery from large volumes of water of different
qualities. For example, studies on the use of erythrocyte
ghosts covalently linked to agarose beads in column (8) for
TABLE 4. Adsorption-elution of phages with 13% erythrocyte ghost cells in 3% beef extract suspension
Xx-174 MS-2 f2
Beef % Adsorption % Elution % Adsorption % Elution % Adsorption % Elution
extract
Witha WithoUtb With Without With Without With Without With Without With Without
Oxoid 100 91.8 32 31 100 58.5 79.7 61 96 54 65 52
Difco 98.5 55 20 24 47.4 4.6 65.7 78.8 71 29 93 100
Inolex 100 79 26 30 94.4 76.6 70 58 97.7 83.7 70 42
a 3% beef extract containing 13% erythrocyte ghosts (vol/vol).
b 3% beef extract only.
VOL. 47, 1984

1340 ARMON, KOTT, AND NEEMAN
TABLE 5. Adsorption-elution of poliovirus LSC-1 with
erythrocyte ghost cells
Expt Virus input' % Virus recovered in:Expt
~(PFU x 104) Step 2b Step 3"
1 2.0 0 100
2 2.2 1 91
3 1.9 0 94
4 1.8 0 129"
5 1.6 0 104
6 1.8 0 126
a Virus (1.6 x 104 to 2.2 x 104 PFU) was added to a 5-ml final
volume of saline containing 13% erythrocyte ghost cells.
b 100% adsorption of virus on erythrocyte ghost cells, except
experiment 2, in which adsorption was 99% (see text).
' Results represent percent elution of virus from erythrocyte
ghost cells (see text).
d Elution above 100% was due to virus aggregation in input.
phage and virus concentration from larger volumes of water
and wastewater effluents are now in progress in our labora-
tory.
LITERATURE CITED
1. Adams, M. H. 1959. Bacteriophages. Wiley-Interscience, New
York.
2. Ayres, P. A. 1978. Coliphages in sewage and the marine environ-
ment, p. 275-298. In F. A. Skinner and J. M. Shewan (ed.),
Aquatic microbiology. Academic Press, Inc., New York.
3. Berry, S. A., and B. G. Noton. 1976. Survival of bacteriophages
in seawater. Water Res. 10:323-327.
4. Bishop, C., and D. M. Surgenor. 1974. The red blood cell, 2nd
ed., vol. 1, p. 95. Academic Press, Inc., New York.
5. Borner, J. D., G. Toro, and P. G. Ackerman. 1962. Bray's
clinical laboratory methods, p. 149. The C. V. Mosby Company,
St. Louis, Mo.
6. Cliver, D. C. 1965. Factors in the membrane filtration of
enteroviruses. Appl. Microbiol. 39:85-91.
7. Davis, J. E., and R. L. Sinsheimer. 1963. The replication of
bacteriophage MS-2. J. Mol. Biol. 6:203-207.
8. Eshdat, Y., and J. A. Prujansky. 1979. Isolation of human
erythrocyte membranes and their outer surface proteins by
covalent fractionation on solid support. FEBS Lett. 101:43-46.
9. Eylar, E. H., M. A. Madoff, 0. V. Brody, and J. L. Oncley. 1962.
The contribution of sialic acid to the surface charge of the
erythrocyte. J. Biol. Chem. 237:1992-2000.
10. Finean, J. B., and R. H. Michell. 1981. Membrane structure, p.
10. In New comprehensive biochemistry, vol. 1. Elsevier/North-
Holland Biomedical Press, New York.
11. Goyal, S. M., K. S. Zerda, and C. P. Gerba. 1980. Concentration
of coliphages from large volumes of water and wastewater.
Appi. Environ. Microbiol. 39:85-91.
12. Holm, K. L., and L. Philipson. 1974. Early interaction between
animal viruses and cells. Monogr. Virol. 9:50-55.
13. Howe, G., and L. T. Lee. 1972. Virus-erythrocyte interactions.
Adv. Virus Res. 17:1-45.
14. Kabat, E. A. 1961. Experimental immunochemistry, p. 110.
Charles C. Thomas, Publisher, Springfield, Ill.
15. Katzenelson, E., B. Fattal, and T. Hoskovesky. 1976. Organic
flocculation: an efficient second step concentration method for
the detection of viruses in tapwater. Appl. Environ. Microbiol.
32:638-639.
16. Kott, Y., N. Roze, S. Sperber, and N. Betzer. 1974. Bacterio-
phages as viral pollution indicators. Water Res. 8:165-171.
17. Loeb, T., and N. D. Zinder. 1961. A bacteriophage containing
RNA. Proc. Natl. Acad. Sci. U.S.A. 47:282.
18. Lowry, C. H., N. S. Rosenbrough, and A. L. Farr. 1951. Protein
measurement with the Folin phenol reagent. J. Biol. Chem.
193:265-275.
19. Rosen, L. 1969. Hemagglutination with animal viruses, p. 285.
In K. Habel and N. P. Salzman (ed.), Fundamental techniques
in virology. Academic Press, Inc., New York.
20. Seeley, N. D., and S. B. Primrose. 1979. Concentration of
bacteriophages from natural waters. J. Appi. Bacteriol. 46:103-
116.
21. Shuval, H. I., A. Thompson, B. Fattal, S. Cymbalista, and Y.
Wiener. 1971. Natural virus inactivation processes in sea water.
J. Sanito Eng. Div. Proc. Am. Soc. Civ. Eng. 97:587-600.
22. Sinsheimer, R. L. 1959. Purification and properties of bacterio-
phage 4x-174. J. Mol. Biol. 1:37-42.
23. Smith, J. E., and R. J. Courtney. 1965. Concentration and
detection of viral particles by passive hemagglutation, p. 89-
107. In G. Berg (ed.), Transmission of viruses by the water
route. Wiley-Interscience, New York.
24. Zittle, C. A. 1953. Adsorption studies of enzymes and other
proteins. Adv. Enzymol. 44:319.
APPL. ENVIRON. MICROBIOL.