Concentration of simian rotavirus SA-11 from tap water by membrane filtration and organic flocculation.

Vol. 45, No. 3APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Mar. 1983, p. 850-855
0099-2240/83/030850-06$02.00/0
Concentration of Simian Rotavirus SA-11 from Tap Water by
Membrane Filtration and Organic Flocculation
NAOMI GUTTMAN-BASS* AND ROBERT ARMON
Environmental Health Laboratory, Hadassah Medical School, Hebrew University, Jerusalem, Israel
Received 17 September 1982/Accepted 23 December 1982
Simian rotavirus SA-11 was concentrated from tap water by adsorption to and
elution from microporous filters, followed by organic flocculation. Two types of
filters were compared for their ability to concentrate the virus. Both Zeta Plus 60S
and Cox AA type M-780 filters were efficient for virus adsorption, but the
efficiency of virus elution was higher with Zeta Plus than with Cox filters.
Optimum conditions for virus recovery from Zeta Plus filters included an input
water pH of 6.5 to 7.5 and the use of 3% beef extract (pH 9.0) for elution. Under
these conditions, an average of 62 to 100% of the virus was recovered in the
concentrate. Organic flocculation was used as a second-step concentration
method, with average recoveries of 47 to 69%. When the two methods were used
to concentrate small numbers (7 to 75 PFU/liter) of input rotavirus, an average of
75 ± 40% recovery was achieved. With large volumes of input water, however,
recovery was reduced to 16 ± 7%.
Rotaviruses are viral agents of gastroenteritis
in animals and humans and are distributed
worldwide (5, 19). They are fecally excreted (5)
and thus enter the water environment as poten-
tial agents of waterborne epidemics of diarrhea.
They have been detected in sewage (13, 17), but
at present there is little firm evidence for their
role as waterborne disease agents. For determi-
nation of the extent of contamination of the
environment with these viruses it is essential to
have a reliable and sensitive technique for their
detection. Furthermore, these viruses may be
present in water in low concentrations, and an
efficient concentration technique is required for
their detection by standard virological proce-
dures.
Most methods of virus concentration have
been developed with poliovirus as a repre-
sentative virus. This virus differs from rotavirus-
es in many characteristics, including adsorptive
behavior and stability (2, 3). For determination
of which is the most efficient method for rota-
virus concentration from water it is necessary to
use one of the rotaviruses as an experimental
model. The simian rotavirus SA-11 has been
commonly used for laboratory investigations of
rotaviruses since it grows to high titer in tissue
cultures and can be quantitated by plaque assay
(12). Because this simian rotavirus resembles
the human rotaviruses in morphology and com-
position and shares cross-reacting antigens (11,
18), it is useful as a model for the human disease
agents.
In this study, the efficiency of simian rotavirus
SA-11 concentration from tap water was investi-
gated for two membrane filtration methods. The
usefulness of organic flocculation as a second-
step concentration method for rotaviruses was
also determined.
MATERIALS AND METHODS
Cells. The MA-104 line of fetal rhesus monkey
kidney cells was obtained from Y. Marzouk (Central
Virus Laboratory, Tel Aviv, Israel). The cells were
grown in RPMI-1640 medium containing 10o newborn
bovine serum (Flow Laboratories, Irvine, United
Kingdom); 200 jig of penicillin, 200 pLg of streptomy-
cin, 200 ,ug of kanamycin, 25 jig of neomycin, and 100jig of nystatin per ml; and 20 ,uM glutamine.
Virus and virus assay. Simian rotavirus SA-11, kind-
ly supplied by C. P. Gerba (Department of Microbiolo-
gy, University of Arizona, Tucson), was grown and
assayed in MA-104 cells under conditions similar to
those described previously (12). Specifically, virus
stocks were prepared in the presence of 12 ,ug of
trypsin per ml (1:250; tissue culture grade; Difco
Laboratories, Detroit, Mich.) in growth medium minus
serum. The titer of virus stocks was typically 101
PFU/ml. Plaque assays were carried out in 50-mm-
diameter tissue culture plates in the presence of 5%
carbon dioxide. Facilitators added to the agar overlay
to enhance plaque formation were 20 ,ug of trypsin and
100 ,ug of diethylaminoethyl dextran per ml (Sigma
Chemical Co., St. Louis, Mo.), conditions experimen-
tally determined to maximize virus titers. Plaques
were formed within 3 to 5 days postinfection, and an
agar overlay containing 170 ,ug of neutral red per ml
was added to aid in plaque visualization.
Virus concentration methods. (i) Membrane filtra-
tion. Virus concentration by adsorption to and elution
850

ROTAVIRUS CONCENTRATION FROM TAP WATER 851
TABLE 1. Concentration of rotavirus by Zeta Plus
filtersa
Virus % Virus
Expt Virus input eluted recovery
(PFU) efficiencyb
1 4.0 x 106 2.7 x 106 68
2 4.0 x 106 2.1 x 106 53
3 4.0 x 106 3.9 x 106 98
4 4.0 x 106 2.0 x 106 50
5 4.3 x 105 2.6 x 105 60
6 4.3 x 105 2.6 x 105 60
7 4.3 x 105 2.1 x 105 49
8 4.3 x 105 2.4 x 105 56
a Filtration was done with 2 liters of tap water at pH
6.5.
b The mean + standard deviation percent efficiency
was 62 ± 16.
from microporous filters was as follows, with excep-
tions as indicated. Virus was added to Jerusalem tap
water dechlorinated by the addition of 10 mg of sodium
thiosulfate per liter. The pH of the input water was
adjusted to various levels, and the sample was filtered
through a 142-mm-diameter filter at a pressure of 0.5
kg/cm. Eluent (75 ml of 3% beef extract [pH 9.0]) was
added, incubated on the filter for 15 min at room
temperature, and eluted. Filters used were a Zeta Plus
60S cellulose-diatomaceous earth-"charge-modified"
resin filter, with a 0.45-,um nominal pore size (AMF,
Cuno Div., Meriden, Conn.), and a Cox AA type M-
780 epoxy-fiberglass-asbestos filter, with a 0.45-p.m
nominal pore size (Cox Instrument Div., Lynch Corp.,
Detroit, Mich.). Virus recovery was determined by
titration of the virus stock, input water, filtrate, and
eluate by plaque assay.
(ii) Organic flocculation. Virus eluted from filters
with 3% beef extract was reconcentrated by organic
flocculation as described previously (6). Briefly, the
pH of the eluate was lowered to 3.5, and the sample
was stirred for 30 min at room temperature. The floc
which formed was concentrated by centrifugation and
suspended in a 1/20 volume of 0.15 M Na2HPO4 (pH
7.5). The amounts of virus in the floc supernatant were
determined by plaque assay.
Eluents. Eluents tested for their elution efficiency
included the following: beef extract (Lab-Lemco, Ox-
oid, London, England), tryptose phosphate broth
(Difco Laboratories), peptomeat (Biolife Italiana, Mi-
lan, Italy), and Triton X-100 (BDH Chemicals, Poole,
United Kingdom).
RESULTS
Concentration of rotavirus by membrane filtra-
tion. The ability of microporous filters to con-
centrate simian rotavirus SA-11 from tap water
was tested for two filter types, the electroposi-
tive Zeta Plus 60S filters and the Cox AA type
M-780 filters, which are electronegative at neu-
tral pH levels. The filtration procedures for both
filters were similar, except that to enhance ad-
sorption, the pHs of the input water were 6.5 for
the Zeta Plus filters and 3.5 for the Cox filters.
Virus was seeded into dechlorinated tap water,
the pH was adjusted, and the sample was fil-
tered. After filtration, the adsorbed virus was
eluted with 3% beef extract (pH 9.0), and the
eluates were assayed for virus.
The Zeta Plus filters efficiently concentrated
rotavirus from tap water. Virus recovered in the
eluate averaged 62 + 16% of the input virus,
with a range of recovery in individual experi-
ments of 49 to 98% (Table 1). Adsorption was
highly efficient, since no virus was recovered in
the filtrate. For comparison, a series of experi-
ments with poliovirus type 1 (Brunhilde strain)
was performed under the same conditions, re-
sulting in an average recovery of 73 + 5% (data
not shown). Thus, the Zeta Plus filters recov-
ered rotavirus from tap water with an efficiency
which did not differ markedly from their efficien-
cy for poliovirus.
In contrast, the Cox filters proved to be
inefficient for rotavirus concentration from tap
water. In a series of eight experiments, recovery
in the eluate averaged 3.1% (Table 2), a recovery
much lower than that found for Zeta Plus filters.
The Cox filters adsorbed the virus efficiently,
since no virus was detected in the filtrate, but
the virus was not eluted.
During filtration through Cox filters, the virus
may have been inactivated because of the low
pH of the input water. Experiments were per-
formed to measure the rate of loss of the virus
titer in the input tap water at pH 3.5. It was
found that for up to 1 h at room temperature, the
virus titer did not drop by more than 50% (data
not shown). This was not sufficient to account
for the low virus recovery, since during concen-
tration, the time of exposure of the virus to pH
3.5 was less than 10 min. Thus, it is probable
that the virus was not recovered from the filter
TABLE 2. Concentration of rotavirus by Cox
filtersa
Virus eluted % VirusExpt (PFU) recovery
efficiency'
1 0.06 X 106 4.0
2 0.02 x 106 1.3
3 0.05 x 106 3.3
4 0.01 x 106 0.7
5 0.01 x 106 0.7
6 0.04 x 106 2.7
7 0.11 x 106 7.3
8 0.07 x 106 4.7
a Filtration was done with 2 liters of tap water at pH
3.5. Virus input was 1.5 X 106 PFU for all experi-
ments.
b The mean + standard deviation percent efficiency
was 3.1 + 2.3.
VOL. 45, 1983

852 GUITMAN-BASS AND ARMON
P
GE
deW)
TOP WATER PH
I1
FIG. 1. Effect of input water pH on rotavirus re-
covery from Zeta Plus filters. Tap water (2 liters) was
seeded with rotavirus (2.4 x 106 to 4.0 x 106 PFU) and
adjusted to the indicated pHs. After filtration, the
virus was eluted, and the amounts of virus in the eluate
and filtrate were determined. No virus was detected in
the filtrate; virus recovery is indicated as the percent-
age of input virus recovered in eluate.
either because of inefficient elution or because
of inactivation of the virus on the filter.
Effect of input pH on rotavirus recovery. To
determine the range of usefulness of Zeta Plus
filters for rotavirus recovery from natural wa-
ters, we determined the recovery efficiency as a
function of input water pH. Rotavirus was seed-
ed into tap water, the pH was adjusted as
required, and the sample was filtered. The virus
was eluted, and recovery in the eluate was
determined. Virus recovery efficiency was at a
maximum between pH 6.5 and 7.5 (Fig. 1). Virus
was not detected in the filtrate at any input pH.
Thus, virus adsorption was efficient under all
conditions tested, but elution was successful
within a more limited pH range.
Efficiency of various eluents for rotavirus elu-
tion. To maximize the efficiency of rotavirus
recovery, we compared a number of eluents for
their ability to elute rotavirus from Zeta Plus
filters (Table 3). Several proteinaceous media
were tested, including 2.5% and 5% tryptose
phosphate broth and 5% peptomeat. Peptomeat
was unable to elute the virus. Tryptose phos-
phate broth functioned as an eluent, but 3% beef
extract was more effective.
An attempt was made to improve rotavirus
elution by use of the nonionic detergent Triton
X-100. Triton X-100 was added to the filter and
incubated for 10 min, followed by the addition of
3% beef extract. After an additional contact time
of 15 min, the sample was eluted, and virus
recovery was determined (Table 3). No im-
provement in virus recovery was found. Similar-
ly, a mixture of Triton X-100 and beef extract
was not a better eluent than beef extract alone.
It is possible that higher concentrations of Triton
X-100 would aid in elution, but they were not
used since they damaged the cell monolayer.
To determine whether the addition of beef
extract in two steps would release more virus
from the filter, we performed a second elution
after the first. Virus recovery did not improve
(Table 3). Thus, of the eluents and procedures
tested, 3% beef extract in a single step was the
most simple and efficient method for rotavirus
elution from Zeta Plus filters.
Effect of incubation time on elution efficiency.
To test whether increased incubation time of the
eluent on the filter would improve elution, we
performed the following experiment. Rotavirus
was added to 2-liter volumes of tap water and
filtered through four Zeta Plus filters in parallel.
Eluent was added and eluted at times ranging
from 15 to 60 min after addition to the filter.
Virus recovery did not improve with longer
incubation times; after 15 min, 41% of the virus
was recovered in the eluate, and 42% was recov-
ered after 60 min. Thus, there was no advantage
to extended incubation; 15 min was sufficient to
elute the virus. Note that the extended exposure
TABLE 3. Efficiency of rotavirus elutiona
Eluentb No. of % Avg virusexperiments recoveryc
3% BE 8 62 ± 16
2.5% TPB 4 28 ± 7
5% TPB 4 26 ± 10
5% PM 8 0.12 ± 0.25
2.7% BE + 0.001% 4 45 ± 8
TX-100
0.01% TX-100 + 2 31 ± 5
3% BEd
3% BE + elution + 4 65 ± 7
3% BEe
a Virus (7.7 x 105 to 5.5 x 106 PFU) was seeded into
2 liters of tap water and filtered through Zeta Plus
filters at pH 6.5. Eluent (75 ml, pH 9.0) was added and
eluted after 15 min. Exceptions are as noted.
b BE, Beef extract; TPB, tryptose phosphate broth;
PM, peptomeat; and TX-100, Triton X-100.
I The results are calculated as the percentage of
input virus recovered in the eluate and expressed as
the mean + standard deviation of each set of experi-
ments.
d The first eluent (10 ml, pH 7.0) was added to the
filter and incubated for 10 min, followed by the addi-
tion of the second eluent (65 ml) and incubation for 15
min, followed by elution.
eElution was done as described in a, followed by
the addition of 25 ml of 3% BE, incubation for 15 min,
and elution. The eluates were combined and titrated.
APPL. ENVIRON. MICROBIOL.

ROTAVIRUS CONCENTRATION FROM TAP WATER 853
TABLE 4. Concentration of rotavirus by membrane
filtration and organic flocculationa
% Virus recovered in: % Organic
Expt flocculation
Eluate Floc Supernatant efficiency
1 113 44 0 39
2 89 50 0 56
a Virus (1.5 x 10' PFU) was added to 2 liters of tap
water and filtered at pH 6.5. After elution, the virus
was reconcentrated by organic flocculation. The mean
percentages of virus recovered in the eluate, the floc,
and the supernatant were 101, 47, and 0, respectively.
The mean percent efficiency was 47.
of the virus to pH 9.0 did not reduce the virus
titer, indicating that extensive inactivation was
not occurring under these conditions.
Efficiency of organic flocculation for rotavirus
concentration. Virus isolation from water re-
quires concentration to a volume small enough
to be handled conveniently by the system used
for virus detection. For large volumes of water,
this generally means the use of a second concen-
tration step to attain a final volume of a few
milliliters. Organic flocculation is a method
which has been shown to be highly efficient for
poliovirus concentration (6, 7). It is particularly
applicable when beef extract is used as the filter
eluent, since it simply requires the reduction of
the pH of the sample to 3.5 for floc formation.
The floc is then concentrated by centrifugation
and resuspended in a small volume of phosphate
buffer. To determine the efficiency of this tech-
nique for rotavirus concentration, we performed
organic flocculation on the eluates of Zeta Plus
filters. The organic flocculation step had an
average efficiency of 47% (Table 4). In these
experiments, because of the high recovery from
the filters, the efficiency of the two steps togeth-
er was 47%. The supernatant which remained
after the organic floc was removed was also
assayed for virus, and none was detected, indi-
cating that the rotavirus adsorbed efficiently to
the floc.
TABLE 5. Concentration of rotavirus by organic
flocculation'
Virus input % Virus recovered in:Expt (PFU) Floc Supernatant
1 2.1 x 107 57 0.01
2 2.3 x 107 65 0.52
a Virus was seeded into 75 ml of 3% beef extract.
Organic flocculation was performed, and the indicated
fractions were assayed for virus. The mean percent-
ages of virus recovered in the floc and the supernatant
were 61 and 0.26, respectively.
TABLE 6. Recovery of small quantities of rotavirus
from tap water'
Virus Virus
Expt input recovered Recoveryb(PFU) (PFU)
1 70 37 53
2 70 69 99
3 70 61 87
4 70 82 117
5 150 29 19
6 150 25 17
7 15 17 113
8 15 14 93
a Virus was seeded into 2 liters of tap water, filtered
through Zeta Plus filters at pH 6.5, eluted, and recon-
centrated by organic flocculation.
b The mean ± standard deviation percent recovery
was 75 ± 40.
To determine whether there was virus loss as
a result of using beef extract which had served
as an eluant, we performed organic flocculation
with virus added to fresh beef extract. An aver-
age of 61% of the virus was recovered from the
organic floc, and <1% remained in the superna-
tant (Table 5), a slightly improved recovery.
Although organic flocculation successfully
concentrated rotavirus, there was some virus
loss. This may have been due to inactivation or
incomplete resuspension after flocculation. To
improve virus dispersion, we sonicated the re-
suspended concentrate for up to 10 min in a
water bath sonicator. This treatment did not
increase the virus titer (data not shown). Since
the virus was exposed to pH 3.5 for 45 min
during organic flocculation, the stability of rota-
virus in beef extract (pH 3.5) was investigated.
Virus was seeded into 3% beef extract, the pH
was adjusted to 3.5, and samples were removed
and titrated at various times. By 30 min, the
virus titer had dropped to 70% of the initial value
(data not shown), possibly accounting for most
of the virus loss.
Recovery of small quantities of rotavirus from
tap water. Viruses are found in small amounts in
natural waters, and a useful concentration meth-
od must recover small numbers of infectious
particles with high efficiency. The ability of a
combination of the membrane filtration and or-
ganic flocculation methods to concentrate small
quantities of rotavirus was tested in the follow-
ing experiments. Rotavirus (15 to 150 PFU) was
seeded into 2-liter volumes of tap water, the
water was filtered through Zeta Plus filters, and
the virus was eluted and reconcentrated by
organic flocculation. Virus recovery in the elu-
ate was not determined because of the small
amounts of seeded virus. Virus recovery in the
VOL. 45, 1983

854 GUTFMAN-BASS AND ARMON
TABLE 7. Concentration of rotavirus from large
volumes of tap watera
Input % Virus recovered % Organic
Exp water in:_ flocculationVol Organic efficiency(liters) Eluate floc
1 64.5 24 14 57
2 33 30 22 73
3 33 15 12 78
a Virus (4 x 1iO PFU) was seeded into the indicated
amount of tap water, filtered through Zeta Plus filters
at pH 6.5, eluted, and reconcentrated by organic
flocculation. The mean ± standard deviation percent-
ages of virus recovered in the eluate and the organic
floc were 23 ± 8 and 16 ± 7, respectively. The mean ±
standard deviation percent efficiency was 69 ± 11.
organic floc averaged 75% (Table 6), with a
range of 17 to 117%. This wide range may have
been due to the large statistical error associated
with assaying for small numbers of viruses. The
results indicate a high recovery efficiency for the
two methods when used with low virus concen-
trations.
Concentration of rotavirus from large volumes
of water. The ability of Zeta Plus filters to
concentrate rotavirus from large volumes of tap
water was tested in the following series of ex-
periments. Virus was added to 33 or 64.5 liters of
water, filtered, eluted, and reconcentrated by
organic flocculation. The largest volume of wa-
ter which could be filtered through a single filter
was 64.5 liters, with a filtration time of 2 h. The
amounts of virus in the eluate and in the organic
floc were measured (Table 7). The average virus
recovery of 16% was lower than that found with
smaller volumes of input water. Since recovery
of the virus after organic flocculation was rela-
tively high, virus loss occurred during the mem-
brane filtration step. Thus, the low virus recov-
ery was probably a result of reduced adsorption
efficiency or of inefficient elution from the filter.
DISCUSSION
Several studies have investigated methods for
rotavirus concentration from tissue culture har-
vests (4, 10) and water (8, 9, 13, 17). The
techniques used for tissue culture harvests are
not directly applicable to the handling of large
volumes of water. Successful concentration of
rotavirus from up to 100 liters of water by
ultrafiltration has been reported (8), but this
technique is both unwieldy for field virus con-
centration and time-consuming. The use of Talc-
Celite layers also has been reported to be effi-
cient for rotavirus concentration from water (9),
but this method is more difficult to use than
membrane filtration.
Field isolation of rotavirus from sewage has
been recently reported by two laboratories.
Steinmann (17) concentrated rotavirus from
sewage by filtration through Seitz electroposi-
tive filters, followed by reconcentration by ultra-
centrifugation. Rotavirus was detected by im-
munological techniques which did not
distinguish between viable and noninfectious
viruses. Unfortunately, the efficiency of the
methods used was not reported, so that compari-
sons with other techniques are difficult. In an-
other study (13), membrane filtration with Filter-
ite filters was used to concentrate rotavirus from
sewage, resulting in successful isolation of via-
ble virus. Simian rotavirus SA-11 was used to
measure the efficiency of the technique, and
recovery efficiency from 20 liters of tap water
averaged 54.6%. The major disadvantages of
these filters are the need for reduction of the
input water pH to 3.5 and the relative sensitivity
of the filters to interference by contaminants in
the water (14).
Zeta Plus filters have several advantages over
Filterite filters, including relative insensitivity to
changes in water quality (14) and efficient virus
recovery at a more neutral pH (16). As reported
here, the Zeta Plus filters were more successful
in recovering rotavirus from tap water than were
the electronegative Cox filters with which they
were compared. The Cox filters used have been
included as one of the filter types recommended
for use as a tentative standard method for virus
concentration (1) and have been found to effi-
ciently recover poliovirus from water when beef
extract is used as an eluent (6, 7). Although it
was found here that these filters adsorbed rota-
virus, elution was highly inefficient, and these
filters would not be appropriate for rotavirus
concentration.
In contrast, Zeta Plus filters both adsorbed
and eluted rotavirus, resulting in recovery of the
majority of the input virus. The filters were
useful at a neutral pH range, obviating the need
for pH adjustment in the filtration of many
natural waters. It was of interest that the input
pH of the water influenced elution rather than
adsorption efficiency. That the chemical nature
of the input water may interfere with virus
elution has also been observed with polyvalent
cations (15). Thus, the chemical condition of the
water may influence the virus-filter bond or the
state of virus aggregation and thus, virus elution.
Proteinaceous media used as virus eluents
have been generally found to be more efficient
than other materials (4, 7, 14, 15). Of the media
and procedures tested here, 3% beef extract (pH
9.0) was the most efficient for rotavirus elution.
It should be noted that it was not sufficient for
the eluent to be proteinaceous, since the two
other media tested were less effective than was
beef extract.
APPL. ENVIRON. MICROBIOL.