Factors influencing virus adsorption to solids


Material Information

Factors influencing virus adsorption to solids
Physical Description:
vi, 191 leaves : ill. ; 28 cm.
Shields, Patricia Ann, 1958-
Publication Date:


Subjects / Keywords:
Adsorption (Biology)   ( lcsh )
Viruses   ( lcsh )
Microbiology and Cell Science thesis Ph. D
Dissertations, Academic -- Microbiology and Cell Science -- UF
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1986.
Bibliography: leaves 176-190.
Statement of Responsibility:
by Patricia Ann Shields.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000566781
oclc - 14584647
notis - ACZ3222
sobekcm - AA00004871_00001
System ID:

Full Text







To Mom and Dad, and Tom, Sharon, Carrie and Jennifer


I gratefully acknowledge the help and guidance given me by the

members of my committee, Dr. Samuel R. Farrah, Dr. Lonnie O. Ingram, Dr.

Stephen G. Zam, Dr. Gabriel Bitton and Dr. Dinesh O. Shah. I am forever

indebted to my chairman, Dr. Farrah, who has been my friend as well as

my boss. The example of scientific excellence he has set will never be


In addition, I extend thanks to those individuals who have helped,

in many ways, to make this journey enjoyable, especially Dr. Phillip

Scheuerman, Orlando Lanni, Sharon Berenfeld, Kathi Moody, Vivienne

Thompson, Dave Preston, Lena Dingler, Jane Strandberg and Gail Waldman.

In particular, I am beholden to Lena Dingler.and Jane Strandberg, both

of whom have shown me the true meaning of friendship.

Finally, I extend thanks to my family for the love and support they

have given me -in this and all other endeavors. This dissertation is

dedicated to them.




ACKNOWLEDGMENTS................ .. ................................ iii

ABSTRACT..... ................. .................................... v


I INTRODUCTION....................... ........................ 1


Review of the Literature..................................... 7
Materials and Methods....................................... 25
Results ...................................................... 33
Discussion................................. .................... 65

III VIRUS ASSOCIATION WITH OTHER SOLIDS.......................... 76

Review of the Literature ..................................... 76
Materials and Methods........................................ 95
Results.............................................. 99
Discussion........................................... ...... 123

IV CHARACTERIZATION OF VIRUSES................................. 133

Review of the Literature.................................... 133
Materials and Methods....................................... 141
Results............................................. ...... 143
Discussion....................................................... 156

V SUMMARY...................................................... 166


PROPAGATION AND ASSAY.................................. 168

CULTURE WORK.......................................... 172

BIBLIOGRAPHY ..... ................................................ 175

BIOGRAPHICAL SKETCH.............................................. 190

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Patricia Ann Shields

May 1986

Chairman: Samuel R. Farrah
Major Department: Microbiology and Cell Science

Several factors influence the association of viruses with solids.

The natures of the adsorbent and virus studied are among the most

important factors that influence this association. Differences in virus

adsorptive behavior have been related to the relative strength of

hydrophobic and electrostatic interactions between the virus and solid.

The main objective of this study was to determine the relative

hydrophobic and electrostatic nature of viruses and various virus-

adsorbing solids.

Membrane filters used in virus concentration procedures were

characterized by contact angle and capillary rise measurements. Among

electropositive filters tested, Seitz S filters were found to be the

most hydrophobic, while Virosorb 1MDS filters were found to be the most

hydrophilic. Among electronegative filters tested, Millipore HA filters

were found to be the most hydrophobic, while Filterite filters were

found to be the most hydrophilic. These physical measurements were

found to correlate well with the ability of certain solutions to elute

viruses adsorbed to the filters. Based on these differences, several

two-step concentration procedures for viruses in water were developed

that used dissimilar filters for each stage.

Beef extract solutions are often used for the elution of viruses

adsorbed to membrane filters. The use of ammonium sulfate flocculation

to concentrate viruses in these solutions was compared to organic

flocculation. The recovery of seeded enteroviruses from sludge was 62%

using ammonium sulfate flocculation as compared to 34% when organic

flocculation was used as a concentration step. Ammonium sulfate

flocculation was also found to improve recoveries of viruses from

sewage effluents and soils.

The relative hydrophobic and electrostatic nature of viruses was

determined. Hydrophobic interaction chromatography and hydrocarbon

adherence tests were used to assess the relative hydrophobicity of the

viruses. Adsorption to DEAE-sepharose was used to assess the relative

electrostatic nature of the viruses. Based on these tests, echovirus-5

and bacteriophage MS2 were found to have the highest relative

hydrophobicity. Bacteriophages T2, T4 and MS2 were found to adsorb very

strongly to DEAE-sepharose, indicating these viruses had the strongest

relative electrostatic interactions of the viruses tested.


Viruses that are shed in fecal matter are collectively referred to

as enteric viruses. More than 100 different enteric viruses are known

to be excreted in the human feces (60, 68, 158). Examples of these

viruses and the diseases they cause are shown in Table 1. A human

showing no symptoms of disease can excrete up to a million infectious

virus particles per gram of feces and this number may be higher in

individuals manifesting disease (52, 68, 158).

Theoretically, any virus excreted in feces and capable of producing

infection when ingested could be transmissible by inefficiently treated

water (68, 88). However, epidemiological evidence for water-borne

transmission of human viruses is limited to Hepatitis A virus, Norwalk

virus and adenovirus infections associated with swimming pools (88).

For most enteric viruses, demonstration of water-borne transmission is

hampered by several factors. Many of these enteric viruses cause

asymptomatic infections (52). As such, the role of short-term carriers

of viruses who do not develop disease is not easily assessable (88,

158). In addition, the possibility that other secondary transmission

sources may exist makes accurate epidemiological studies virtually

impossible (68, 88).

Studies have shown that some human enteric viruses can survive

sewage treatment and persist in wastewater sludges and effluents (89,

115, 122, 172, 186). At the end of the treatment cycle, wastewater

Table 1. Human enteric viruses and associated diseases.

Virus Group of Types Disease

Poliovirus 3 Paralytic poliomyelitis, aseptic

Group A

Group B


Hepatitis A




Norwalk virus


3 (?)

Herpangina, aseptic meningitis,
paralysis, fever

Pleurodynia (Bornholm disease),
aseptic meningitis, acute
infantile myocarditis, rash,
fever, respiratory disease

Aseptic meningitis, fever,
rash, diarrheal disease,
respiratory disease

Infectious hepatitis

Fever, respiratory disease,

Severe diarrhea, vomiting,
low grade fever, dehydration

Respiratory and eye infections

Diarrhea, vomiting

a Adapted from 59, 60 and 68.

sludges and effluents are often disposed of on land (174). Once

disposed of in this manner, viruses associated with sludge flocs or in

effluents interact with soils can lead to viral contamination of crops,

soils, and surface and groundwaters (65, 68, 172, 174).

While the number of viruses in raw sewage may be quite high, the

concentration of viruses after treatment, and in subsequently

contaminated solids and water supplies is low. Even at these low

numbers, enteric viruses are capable of causing disease and thus are

cause for concern. Because the numbers of viruses are low, it is not

always practical to screen samples for the presence of these pathogens.

Therefore, it is logical to screen for the presence of an indicator

organism whose presence may indicate fecal pollution of the sample. The

presence of coliform organisms, such as E. coli, has been used

extensively as a measurement of fecal contamination (2). However,

studies have shown that enteric viruses are more resistent to

environmental conditions and sewage treatment processes than are the

coliform organisms (7, 8, 88). Enteric viruses have been isolated from

water supplies which have met coliform limits (7, 31, 57, 76, 187).

Therefore, there appears to be a need to find another, more suitable

organism to serve as an indicator for enteric viruses.

The ideal microbial indicator should have the following

characteristics: i) It should be present when pathogens are present and

absent when pathogens are absent, ii) The growth characteristics and

persistence of the indicator should be similar to the pathogen. iii)

Both the pathogen and the indicator should occur in a constant ratio so

that counts of the indicator will estimate the amount of pathogen

present. iv) The indicator should be found in greater numbers than the

pathogen. v) The indicator should be at least as resistant to

environmental factors as is the pathogen. vi) The indicator should be

nonpathogenic, easily quantifiable, and be able to be tested for in a

variety of environments (63, 167). Several investigators have advanced

the idea of using bacteriophage as indicators for the presence of

enteric viruses in fecally contaminated water (35, 78, 98, 167).

Recently, Stetler monitored coliphages in conjunction with indicator

bacteria and enteroviruses in a drinking water plant (167). Statistical

analysis of the data indicated that coliphages were much better

indicators of enteroviruses than were coliforms or fecal streptococci.

On the other hand, other investigators have noted that bacteriophage

levels do not correlate well with the presence of enteric viruses in

field samples (91, 175). If bacteriophages are to serve as indicator

organisms, then simple and reliable methods must be developed to detect

and concentrate them from natural samples. Procedures used to

concentrate animal viruses often required extremes in pH, which are

unsuitable methods for phage concentration (149, 150). Investigators

have developed methods to concentrate phage using materials which do not

require these fluctuations in pH (67, 113, 150, 157). The development

of a method which could be used to detect both bacterial and animal

viruses in the same sample would simplify and speed up assay procedures


Virus interactions with solid surfaces play a major role in their

removal by sewage treatment processes, as well as their response to,

transportation in and persistence in their environment. The phenomenon

of virus adsorption is also important in many methods used for the

detection and concentration of viruses in the environment (53). Thus,


it is important to study the mechanism of and factors influencing virus

adsorption to solids.

Several factors may influence the association of viruses with

solids. These factors include pH (21, 121, 171, 183), the concentration

and nature of salt in solution (102, 123, 177, 178, 180), the presence

of organic compounds (17, 55, 183), the flow rate of virus through the

adsorbent (61, 103, 147, 174), the nature of the virus (53, 55, 56, 117,

189) and the nature of the adsorbent surface (11, 123, 124, 164, 174).

Accordingly, these factors have been studies to better understand

virus-solid interactions and determine the mechanism of viral


The adsorption of viruses to solids was though to be mainly

electrostatic in nature (94, 173). Since viruses are essentially

biocolloids, theories which describe colloidal behavior have been used

to describe viral interaction with solids.

If the charges on both the virus particle and adsorbing surface are

opposite, electrostatic attraction results in viral adsorption (94, 117,

190). Conversely, if both are of the same charge, electrostatic

repulsion occurs. In most natural water systems, virus particles are

above their isoelectric point and assume a negative charge. By altering

such factors as salt concentration and pH, adsorption to electronegative

surfaces can occur. This phenomenon can be explained by the double

layer theory (52, 53, 173). Essentially, this theory says that ions

associated with both surfaces will attract counterions from the medium

into a compact layer (called the stern layer) around the surfaces. This

layer of counterions will partially neutralize the charge on the

surface. In order for the system to maintain electrical neutrality,

there also exists a diffuse layer containing a lower number of

counterions some distance away. If the bulk solution of counterions

increases by altering the pH or the addition of cationic salts, the

thickness of this diffuse layer decreases. This is because less volume

is needed to contain enough counterions to neutralize the surface

charge. As the size of this layer is decreased, the two surfaces can

come closer to each other and short range attractive forces may take

over. Thus interactions between two disperse particles are thought to

result from a balance between repulsive double layer interactions and

attractive short-range forces, such as van der Waals forces.

Accordingly, altering the pH or addition of cations allows for increased

adsorption to occur (52, 53).

Many studies have confirmed that electrostatic interactions are a

primary mechanism for adsorption of viruses to solids (53, 183, 190).

However, such complex theories such as the double layer theory or simple

theories such as salt bridging fail to explain some of the experimental

results. For example, at high pH, viral adsorption to negatively charge

filters is maintained by the presence of multivalent metal ions.

However, the addition of a chelating agent, which should effectively

remove such ions and prevent salt bridge formation, does not promote

viral elution (40, 42). Furthermore, elution of adsorbed viruses at

high pH has been accomplished using neutral solutions such as urea or

nonionic detergents (42, 48, 180, 183). These variable effects of

different solutions on the elution of viruses adsorbed to solids cannot

be explained in purely electrostatic terms, indicating that additional

forces are involved in viral adsorption to solids. These forces

include hydrogen bonding and more importantly, hydrophobic interactions.

The role of hydrophobic interactions in virus adsorption phenomena

has been examined in several studies and results have shown that these

interactions are influenced by salts, and the nature of the virus and

adsorbent studies (32, 40, 49, 50, 154, 156, 177).

The overall objective of this study was to examine the

contributions of electrostatic and hydrophobic interactions in the

association of viruses with solids, in the hope of better understanding

the factors involved in viral adsorption.

Chapter II of this dissertation examines the association of viruses

to membrane filters. These filters are characterized by several

methods, and the differences in the nature of these filters are used to

explain the differences in viral elution patterns seen. In addition,

several concentration procedures are developed for the detection of

animal and bacterial viruses in water samples. Finally, a new

concentration procedure is shown that allows for simultaneous

concentration of animal and bacterial viruses from the same sample.

Chapter III examines the association of viruses to solids such as

clay, soil and wastewater sludges. The extent of hydrophobic

interactions involved in virus association to clays is examined.

Additionally, methods are developed to recover viruses adsorbed to soils

and sludges.

Chapter IV describes the characterization of viruses themselves.

Viruses are examined by adsorption to or elution from ion exchange gels

and hydrophobic gels. Viruses are further characterized by adherence to

hydrocarbons. These results are then used to explain differences in the

adsorption/elution patterns of different viruses.


Review of the Literature

Concentration of Viruses on Membrane Filters

A wide variety of methods are available for the detection and

concentration of viruses in water samples, however, methods which employ

adsorption to and subsequent elution from membrane filters are still

considered to be among the best (53, 64, 136). In an early study,

Ellford (37) noted a decrease in bacteriophage titer when salt solutions

containing viruses were passed through nitrate cellulose filters. He

found that dilution of the phage into protein solution prior to passage

through the filter avoided loss of viruses onto the filter by

adsorption. In 1967, Cliver (25) found that some enteroviruses were

adsorbed to cellulose nitrate and cellulose acetate filters. In these

experiments, Cliver was attempting to rid solutions of bacteria so as to

assay these solutions for the presence of virus, and therefore this

adsorption was not desirable. He found, much the same as Ellford, that

proteins (in this case serum or gelatin) blocked this viral adsorption.

Later, work in the same laboratory indicated that 99% of enteroviruses

seeded into deionized water were adsorbed when passed through Millipore

(cellulose nitrate) filters (26). Over 80% of these viruses were

subsequently recovered by soaking the membrane in phosphate-buffered

saline containing 30-50% agamma chicken serum.

In 1967, Wallis and Melnick reported the first applied use of

membrane filters to concentrate viruses from water samples (180). This

report detailed several important discoveries. First, they showed in

this correspondence that the addition of salts (0.05 M MgCl2) greatly

enhanced virus adsorption to Millipore filters. The adsorbed virus was

recovered by grinding the filter with a mortar and pestle in the

presence of Melnick B media supplemented with 10% fetal bovine serum.

Over the course of several months, using this procedure, 2795 virus

isolates from sewage water were detected in the concentrate, while only

4 isolates were found in unconcentrated samples from the same period

(180). In addition, during the course of this study, Wallis and Melnick

noted interference with adsorption after the passage of about 100 ml of

sewage through the membrane filter. They suggested that this was due to

the high amount of organic material in the sewage. When the sewage was

passed through anion resins, the interfering material was removed and

virus adsorption was enhanced.

In an subsequent paper (181), Wallis and Melnick reported that

protein and other materials present in crude virus harvests interfered

with virus adsorption to Millipore filters. They called these materials

"membrane coating components" or MCC. These MCC could be removed by

adsorption to anion resins. Once the crude cell harvest was treated to

remove MCC, viruses in these harvests were efficiently adsorbed to

Millipore filters in the presence of salts at pH 5.0. Adsorbed virus

was eluted with a small volume of fetal bovine serum, and in virtually

all cases, 80 to 100-fold concentrations were achieved.

A study by Rao and Labzoffsky (138) confirmed and expanded on the

work of Wallis and Melnick. They found that low levels of calcium (200

ppm) facilitated virus adsorption to Millipore filters. In addition,

this was the first report of using a prefilter (Millipore AP25) in

series with the adsorbing filter. They suggested that in some water

samples, viruses attached to small solids were lost if the prefiltering

stage was separate from the adsorption stage. By using a filter series,

virus retained on the prefilter was eluted along with virus adsorbed to

the cellulose nitrate filter by treatment with 3% beef extract.

Since different eluting solutions had been used to recover viruses

adsorbed to membrane filters, a comparative study of these eluates was

conducted by Konawalchuk and Speirs (97). Coxsackievirus B5 and

poliovirus-1 adsorbed to Millipore filters were eluted using different

concentrations of beef extract, yeast extract and fetal bovine serum.

It was found that elution improved with increases in concentration of

eluting agent and the volume of the eluent. These researchers found

that the best eluate was undiluted fetal bovine serum, which recovered

over 80% of the adsorbed enteroviruses tested, and did not result in any

toxicity problems during viral assay procedures.

The ability of salts other than MgCl2 and CaCI2 to promote viral

adsorption was examined by Wallis et al. (178). They found that 0.5 mM

AlCl was as effective in promoting poliovirus adsorption to Millipore

HA filters as 50 mM MgC12, and was therefore more cost-effective when

large volumes of water were processed.

The processing of large volumes of water generally requires larger

filters, and subsequently larger volumes of the eluting solution are

necessary for efficient viral elution. In order to reduce the amount of

sample to be assayed, there is a need for a second or reconcentration

step. Since the protein solutions commonly used for viral elution

interfered with viral readsorption (25, 37, 180, 181), Wallis et al.

(178) sought another eluate. They found that a high pH (pH= 11.5)

glycine buffer solution eluted viruses adsorbed to Millipore filters

effectively and these eluted viruses were easily readsorbed to a smaller

diameter filter of the same type by lowering the pH and adding AlC13.

This study laid the groundwork for the development of an apparatus for

concentrating viruses from large volumes of water.

Wallis and coworkers (179) developed an apparatus for separating

virus from water contaminants so that the virus contained in the

clarified water could be readily concentrated on adsorbents at high flow

rates. In its final form, the apparatus consisted of five non-viral

adsorbing texile fibers in series, pretreated with Tween 80 to aid in

removal of contaminants without removal of virus. After passage of the

water through this filter series, the water was treated by passage

through an anion-exchange resin to remove organic which may interfere

with virus adsorption. Prior to virus adsorption onto cellulose nitrate

filters, the water was conditioned with MgC12 to a final concentration

of 0.05 M to facilitate adsorption. Elution was accomplished using a

glycine buffer at pH 11.5, and virus was reconcentrated by adsorption to

and elution from a smaller diameter filter (178). Poliovirus-1 was

seeded at high levels into 150 gallons of tapwater, and using this

apparatus, 84% of the virus was recovered. WHen very low levels of

virus were added to the tapwater, virus recovery was still over 60%.

This original model was modified by Homma and coworkers (83) and

this prototype was developed for commercial use by the Carborundum

Company (Niagara Falls,NY) as the "Aquella Virus Concentrator" (64,

136). In this system, the water was passed through a series of

prefilters consisting of 5 pm and 1 pm polyester depth cartridge filters

followed by a Tween 80 treated, 1 pm cotton cartridge. After passage

through these filters, the pH of the water was adjusted to 3.5 and A1C13

was added instead of MgC12 since previous studies had shown the use of

A1C13 to be more cost effective than MgCl2 (178). Viruses were adsorbed

to a 1 pm fiberglass or cellulose acetate cartridge filter and eluted

with one liter of 0.5 M glycine buffer, pH 11.5, and reconcentrated onto

smaller diameter cellulose nitrate membrane filters.

Sobsey and colleagues (165) found that further modifications of the

Wallis-Melnick concentrator were in order. After adjustment of water to

pH 3.5, it was passed through a fiberglass depth prefilter (K-27)

followed by a 142 mm epoxy-fiberglass-asbestos Cox filter in series.

Elution of adsorbed virus was accomplished using one liter of high pH

glycine buffer solution and virus was reconcentrated to a 10 ml volume

by adsorption onto a smaller diameter (47 mm) Cox filter series. Using

this modified system, they were able to recover over 75% of poliovirus-1

seeded into 100 gallons of tapwater. Problems developed with this

system when it was scaled up to process larger volumes of tapwater (44,

46). Farrah et al. (44) found that there was clogging of the initial

virus adsorbent when large volumes of water were processed. This lead

to a reduced flow rate and organic, such as humic acids (46) were

concentrated along with the viruses and interfered with the

reconcentration of the viruses by adsorption to a smaller diameter

filter. The initial clogging problem appeared to be due to the use of a

flat filter at the primary adsorption step. This problem was eliminated

by using a pleated membrane filter cartridge (epoxy-fiberglass Filterite

filter). This cartridge filter provided a larger surface area for virus

adsorption and was capable of adsorbing virus in tapwater at

considerably higher flow rates. The adsorbed virus was eluted with 2 1

of 0.05M glycine, pH 10.5. To overcome the problem of organic

concentrated along with viruses interfering with readsorption to smaller

filters, this step was eliminated and replaced by an aluminum

flocculation procedure (44). THe initial eluate was neutralized and

A1Cl3 was added to a 0.003 M concentration. This lowered the pH of the

water to 4. The solution was neutralized and a floc formed, which was

collected by centrifugation. Virus in the floc was eluted by mixing

the floc with an equal volume of 1.0 M glycine, pH 11.5. The complete

process could be done in 3 hours and virus recovery from 1900 1 of tap

water averaged between 40 and 50%.

Up to this time, most work done on the development of concentration

procedures to detect viruses had used tapwater seeded with virus. With

the development and subsequent modification of the Wallis-Melnick virus

concentrator, researchers began to examine its use with more polluted

water supplies. Using this concentration system, enteroviruses were

efficiently recovered from estuarine water (45, 120, 133), raw sewage

(54, 83, 151), and seawater (54).

Prior to 1978, work done on the concentration of viruses from water

supplies using membrane chromatography centered around the use of

negatively-charged filters such as Filterite, Millipore and Cox filters.

Efficient viral adsorption was accomplished by conditioning of the water

sample by addition of salts and acidification. In 1979, Sobsey and

Jones (164) reported the first use of electropositive filters in the

concentration of viruses from water samples. They tested two positively

charged filters: Seitz S, composed of cellulose-asbestos, and Zeta plus

grade S, composed of cellulose-diatomaceous earth-"charged-modified"

resin. These filters were compared with two negatively charged filters

used in virus concentration, Filterite and Cox filters. At pH values

near neutrality (pH 7.5), less than 20% of poliovirus-1 in unmodified

tapwater was adsorbed by either the Filterite or Cox filters, while

greater than 99% was adsorbed by the Zeta plus and Seitz S filters.

Zeta plus filters were used to concentrate poliovirus-1 from 100 gallons

of tapwater by adsorption to a 267 mm diameter 60S filter, and followed

by elution with 1 1 glycine buffer, pH 10. Virus in the eluate was

further concentrated by dropping the pH to approximately 7 and adsorbing

to a similar, smaller diameter filter followed by elution in a smaller

volume. Using this procedure, recovery of the poliovirus-1 equaled 75%.

Because these electropositive filters did not require acidification

of water samples, researchers began exploring the use of these filters

for the concentration of viruses sensitive to low pH, such as

bacteriophages (67, 113, 136, 149). While bacteriophages are of no

health concern to humans, this work has been done for several reasons.

Several investigators have advanced the idea of using bacteriophages as

indicators of fecally contaminated water (35, 78, 98, 167). Others have

indicated that concentration of naturally occurring phages is necessary

in order to learn more about bacteriophage ecology (136, 150).

Additionally, bacteriophages have been used as models to study the fate

of animal viruses in sewage treatment plants (140).

Logan et al. (113) evaluated the ability of positively charged

filters to adsorb coliphages at different pH values. They found that

greater than 90% of phages MS2 and T2 were adsorbed at pH 7 by the Zeta

plus filters. Efficient elution of virus was accomplished using a

solution of 1% beef extract in 50 mM arginine at pH 9.0. Next, they

seeded 65 1 of prefiltered pond water adjusted to pH 6.0 with various

coliphages. The pond water was passed through a 273 mm Zeta plus 60S

filter and adsorbed phage was eluted with 500 ml of the beef extract

solution. Recoveries of 87% for T2 and 93% of MS2 were obtained using

this method, while only 25% of XX174 was recovered. Use of this method

with river water resulted in 50-60% recovery of naturally occurring


Goyal and coworkers (67) also evaluated the use of Zeta plus

filters for the concentration of phage. They found that MS2, T2 and T4

were all efficiently adsorbed when suspended in tapwater, sewage and

lake water. Again, #X174 was found to adsorb less efficiently than the

other phage tested. Different eluting solutions for viruses adsorbed to

positively charged filters were tested and it was found that 4% beef

extract plus 0.5 M NaCl gave the best recoveries. To determine the

efficiency of this method, 500 ml of sewage water was passed through a

90 mm diameter Zeta plus 50S filter. Adsorbed indigenous bacteriophages

were eluted with 40-50 ml of the beef extract-NaCl solution with an

average efficiency of recovery of 56.5%.

The search group headed by Seeley and Primrose developed (148) and

modified (136) a portable concentrator using positively charged filters.

In this system, water is passed through a series of three prefilters,

precoated with 0.1% Tween 80 to prevent virus adsorption. The clarified

water was adjusted to pH 6.0 and pumped through a 500 mm diameter Zeta

plus filter. Adsorbed virus eluted with a solution of 1.0% beef extract

in 50 mM arginine. Recovery of naturally occurring bacteriophage in

pond water was greater than 60% (136).

The use of positively charged membrane filters for concentrating

viruses from very large volumes of water was hindered due to the lack of

availability of these filters in cartridge form. Previous work with

negatively charged filters had shown that these pleated cartridge

filters were capable of processing several thousand liters of tap water

without clogging (44, 46), while flat disc filters tended to clog

rapidly (44, 165). Sobsey and Glass (163) reported the use of a new

positively charged filter, Virosorb 1MDS, composed of fiberglass-

cellulose-"surface-modified resin" that was available as a double layer,

pleated-sheet cartridge. They compared these filters to the

electronegative Filterite pleated cartridge filters for the recovery of

poliovirus-l seeded into 1000 1 of tapwater. Using beef extract as the

eluate and an organic flocculation procedure (92) for reconcentration,

they found virtually no difference in virus recovery between the

Filterite and Virosorb 1 MDS filter (33% and 30% respectively). However,

the Virosorb 1MDS filters were much easier for field use as no

conditioning of the water prior to adsorption was necessary.

Currently, both positively charged and negatively charged filters

are being used for the processing of large volumes of natural water

samples to detect low levels of viruses. Zeta plus 30S filters were

used to detect indigenous enteroviruses in activated sludge effluents

(23). Virus types isolated included poliovirus-l,-2,-3 and

coxsackievirus B3. During an outbreak of gastroenteritis and hepatitis

in Texas (76), concentration of water supplies by both Filterite and

Virosorb 1MDS filters resulted in the detection of several enteroviruses

such as coxsackieviruses B2 and B3, and Hepatitis A virus.

Factors Influencing Viral Adsorption to Filters

Several factors influence the adsorption of viruses to membrane

filters. These factors include the nature of the virus, the composition

of the filter, and the nature of the solution, including pH, the

concentration and type of salt in solution, the presence of organic

compounds such as proteins and humic acids, and the flow rate through

the filter.

Adsorption of viruses to membrane filters is greatly dependent on

the virus being studied (53). Viruses have protein coats composed of

many amino acids. These amino acids may be acidic, basic or

hydrophobic. The ionization of these amino acids is determined by the

pH values of the viral suspension. At the virus isoelectric point,

there is a net charge of zero. While determination of an isoelectric

point provides some insight to the charged nature of a virus, there are

many problems in drawing any conclusions based on this data alone.

First, studies have shown that the isoelectric point is not only type

dependent, but strain dependent (117, 189). In addition, isoelectric

point data tell us nothing about the charge density of the virus (53).

Little work has been done to elucidate data concerning the charge

density of viruses.

In addition to electrostatic forces effecting the adsorption of

viruses to filters, the hydrophobic nature of the virus may also play a

role. An examination of the amino acid sequence of the coat protein of

MS2 phage has indicated that most spans of amino acids along the

sequence are hydrophobic in nature (188). The relevance of this data

has yet to be determined. It is possible that an understanding of the

relative degree of hydrophobicity of viruses may be useful in predicting

the adsorption behavior of different viruses to membrane filters.

The type and composition of a filter can influence the adsorption

of viruses to the surface. Early work by several investigators (25, 26,

180, 181) found that viruses adsorb efficiently to filters composed of

cellulose acetate, cellulose nitrate and cellulose triacetate. Cliver

(26) found that filters composed of nylon did not adsorb virus as

efficiently as others. This was explained as being due to the greater

hydrophilic nature of nylon compared to the other filters examined

(121). Mix examined various filters used in virus concentration (121)

and concluded that the relative contributions of ionic and hydrophobic

interactions by the filter determined the extent of virus adsorption.

Kessick and Wagner (94) evaluated the electrophoretic mobilities of

virus adsorbing filters as a function of pH, ionic strength, and salt

type. The filters studies were Millipore (cellulose nitrate), Filflo

W10A-7 (cellulose acetate) and Filterite (epoxy-fiberglass).

Electrophoretic mobility measurements through a pH range of 2.0 to 7.0

indicated that all three filters were negatively charged in this pH

range. They also found that increased concentration and valency of

salts worked to decrease the net negative charge of the filters tested.

This provided an explanation of why salt addition helped to promote

virus adsorption to membrane filters (178, 181).

Sobsey and Jones evaluated some positively charged filters using

electrophoretic mobility (164). They found that the electrophoretic

mobility of Zeta plus filters (composed of cellulose-diatomaceous earth-

"charge-modified" resin) and Seitz S (composed of cellulose-asbestos)

filters become more electropositive with a decrease in pH. The Zeta

plus filters had an isoelectric point between 5 and 6, while the Seitz S

filter was found to have an isoelectric point near 7.

Methods such as electrophoretic mobility measurements have limited

application. These measurements are done on very small particles of the

filter material obtained by shredding of the filter material (94). As

such, these measurements may not be indicative of conditions existing on

the intact filter surface. Indeed, Sobsey and Jones (164) data on the

isoelectric point of Seitz S filters was found to be considerably lower

than expected. They indicated that in the shearing process, fibers from

the inner layers of the filter become exposed, possibly lowering the

isoelectric point of the filter.

Despite the known effect of detergents (83, 179) and proteins (25,

37, 94) on viral adsorption to membrane filters, little or no work has

been done to measure the hydrophobic nature of these virus adsorbing


As discussed earlier, the pH of a solution can have a tremendous

effect on the adsorption of viruses to filters. At low pH values,

viruses are often below their isoelectric point and are thus positively

charged, and electronegative filters possess a net negative charge. The

resultant electrostatic attraction is likely a factor in promoting virus

adsorption (94, 117). The converse is true at high pH values, where

viruses are above their isoelectric point and are negatively charged.

Adsorption to electropositive filters is promoted by this electrostatic

attractive force (94).

Studies of virus adsorption to membrane filters have shown that

adsorption is enhanced in the presence of cations (181, 183). Moreover,

trivalent cations, such as aluminum, have been found to be as effective

in this promotion as divalent ions, such as magnesium, at a lower

concentration (178). Three possible mechanisms for the observed

enhancement of virus adsorption in the presence of cations have been

suggested. It has been proposed that these cations act as bridges

between the negative charges on the viruses and the filters (94). A

second suggestion is that the cations are adsorbed by the filters and

reverse the net charge of the filter (117). Electrophoretic mobility

measurements in the presence of cations (94, 163) tend to support this

theory. Another possibility is that the addition of cations (and

therefore an increase in ionic strength) reduces the diffuse layer of

ions surrounding the virus and the filter. This permits short-range

attractive forces to overcome the electrostatic barriers and permits

adsorption of the viruses (173). All of these proposals indicate that

salts influence electrostatic interactions between viruses and membrane


Recent work by Farrah and his research group (40, 49, 156) has

indicated that salts may also effect adsorption of viruses by

influencing hydrophobic interactions between filters and viruses.

Chaotropic salts are able to increase the solubility of certain proteins

(73). These chaotropic ions are relatively large, singly charged ions

such as trichloroacetate (TCA), thiocyanate, and iodide (74). It has

been suggested that hydrophobic interactions are a result of the

unfavorable interactions of apolar groups with water, and accordingly,

chaotropic ions have been viewed as decreasing the structure of water

and therefore making aqueous solutions more lipophilic. In contrast,

antichaotropic ions are generally small, singly charged ions such as

floride, or multivalent ions such as citrate, calcium or magnesium ions.

These antichaotropic ions have been found to promote hydrophobic

interactions, presumably by increasing water structure (73, 74). Farrah

et al. (49) found that at high pH, solutions of chaotropic salts, such

as TCA, eluted over 95% of poliovirus-1 adsorbed to Millipore filters,

while an antichaotropic salt, sodium phosphate, eluted little adsorbed

virus. The authors hypothesized that the ability of chaotropic salts to

elute adsorbed virus indicated that hydrophobic interactions were a

major factor in maintaining virus associations to filters at high pH.

In a further study (40), antichaotropic salts, such as magnesium

sulfate, efficiently promoted adsorption of MS2 to Zeta plus filters at

pH 6.0, while TCA effectively blocked adsorption, again indicating that

hydrophobic interactions were involved. The use of agents involved in

promoting or weakening hydrophobic interactions has yet to be

incorporated in a concentration method for the recovery of viruses from

water samples.

The presence of organic materials, such as humic acids, serum, and

other proteins can effectively reduce virus adsorption to filters. In

early work, Ellford (37) and Cliver (25) both prevented virus adsorption

to filters by pretreating the filters with protein solutions. Wallis

and Melnick-found that materials in crude virus harvest fluids (MCC)

interfered with viral adsorption to Millipore filters (181).

Later, the research group headed by Melnick (83, 178, 179) observed

that organic in water samples were adsorbed to membrane filters and

influenced viral adsorption. Elution of viruses from these filters also

resulted in elution of these organic impurities, which interfered with

readsorption of viruses to a smaller filter (44, 45). These impurities,

humic acids and other organic compounds, were characterized in a

subsequent publication (46). They found that these organic adsorbed at

low pH, were eluted at high pH, formed flocs at low pH and could be

removed by anion-exchange resins.

Recently, humic and fulvic acids were tested for their ability to

interfere with adsorption to some positively charged filters (69). It

was discovered that fulvic acid had little or no effect on the

adsorption of poliovirus-1 to Zeta plus or Seitz S filters. Humic acid,

while interfering with viral adsorption to these filters to some extent,

still allowed up to 40% of the virus to be recovered.

Organic compounds are believed to interfere with viral adsorption

to membrane filters by competing for membrane adsorption sites (53, 64).

Since organic can compete with viruses for adsorption sites, they have

been used to elute viruses from filter surfaces. Some such solutions

used for viral elution include beef extract (64, 107, 136, 138, 183),

yeast extract (97), tryptose phosphate broth (70, 159), casein (16, 42),

Tween 80 (154, 156), and fetal bovine serum (138, 180, 181).

The final factor that can effect virus adsorption to membrane

filters is the flow rate of water through the adsorbing filters, Scutt

(147) examined the adsorption of poliovirus-1 and reovirus-1 to

Millipore filters at different flow rates. While reovirus-1 was

adsorbed at all flow rates tested, breakthrough was observed with

poliovirus-1 as the flow rate was increased. These results indicate

that a minimum contact time is required for an electrostatic bond to

form between the virus and the adsorption site (64, 147).

The flow rate at which virus adsorption becomes diminished depends

on factors influencing the strength of the electrostatic forces between

viruses and filters. During the development of the Wallis-Melnick virus

concentrator, it was discovered that the flow rate of sewage through

glass filters effected virus adsorption (83). At low flow rates (1.9

1/min), 96% of virus as adsorbed compared to only 68% adsorption when

the flow rate was increased to 9.5 1/min.

Reconcentration Procedures

In concentration procedures, as the volume of water processed is

increased, larger filters, and subsequently larger primary eluates, are

used. As such, a second, or reconcentration step is often necessary.

This second step may lower virus recovery, but it reduces the amount of

eluate to a manageable volume for subsequent assays (64). Some

procedures used to reconcentrate viruses after initial adsorption to and

elution from membrane filters include: hydroextraction,

ultrafiltration, ultracentrifugation, readsorption to and subsequent

elution from smaller filters of the same type, and inorganic and organic


One means of reducing the volume of the initial eluate and thus

reconcentrate viruses is hydroextraction. The principle of

hydroextraction is simple. An aqueous solution containing virus is

placed in a semipermeable membrane, often dialysis tubing, and the

tubing is placed into a hygroscopic material overnight. Water passes

through the membranes while large macromolecules, such as viruses are

retained and concentrated (43). Farrah et al. (45) used hydroextraction

with polyethylene glycol as a reconcentration step in the recovery of

enteroviruses from estuarine water. Using this procedure, 80% of seeded

poliovirus-1 was recovered. Others (149) found that hydroextraction of

been extract solutions containing coliphages gave very low recoveries

(less than 5%) and were very difficult to handle due to the high

viscosity of the concentrated material.

Ultrafiltration works on a principle similar to hydroextraction.

Solutions are passed through membranes designed to permit the passage of

water and low molecular weight substances, while larger macromolecules,

such as viruses are retained and concentrated in liquid left behind

(64). Logan et al. (113) used ultrafiltration as a reconcentration step

in the recovery of bacteriophages from large volumes of river water,

with final recoveries in the range of 50-60%. Reconcentration of

enteroviruses using ultrafiltration has been successful in their

recovery from samples of estuarine water (45).

Ultracentrifugation (180) has been used with limited success in the

concentration of viruses. One major problem with this method is the use

of expensive and non-portable equipment, thus its use in field studies

is impractical.

The early models of the Wallis-Melnick virus concentrator used

readsorption to and subsequent elution from a smaller diameter filter

similar to the one used in the first adsorption step of the procedure

(179). This readsorption was facilitated by dropping the pH of the

glycine eluate from 11.5 to 3.5 (178). Further modification of the

model included the addition of A1C13 to 0.5 mM final concentration to the

eluate before readsorption to the second smaller filter (165). This

reconcentration procedure was efficient for smaller volumes of water.

However, difficulties were encountered in applying these procedures to

the detection of viruses in estuarine water (162) or in larger volumes

or tap water (44). Both organic compounds and viruses were adsorbed to

and eluted from the filters in the first stage. These organic compounds

and other impurities formed flocs at the low pH values that were

required for the adsorption of viruses to negatively charged filters

(44, 46, 162). These flocs clogged the filters and made a second stage

reconcentration with negatively charged filters impractical. Some of

these difficulties were overcome by the development of an inorganic

flocculation procedure (44) for a reconcentration step. Previous work

(182) had indicated that viruses readily adsorbed to flocs of aluminum

or calcium salts. Farrah and coworkers used this as the basis for

development of an inorganic flocculation procedure using AlC13 (44).

Initial eluates were conditioned by the addition of A1Cl3 to 3 mM. This

led to a drop in pH to about 4 or 5. Eluates were then adjusted to pH

7.5, resulting in the formation of flocs, which were collected by

centrifugation and resuspended in glycine. Virus recoveries averaged


Payment et al. (133) used FeC12 at 3 mM concentration to form an

inorganic floc during the recovery of virus from estuarine water

samples. They recovered 53% overall of poliovirus-1 added to the water

samples. The authors noted that the efficiency of the reconcentration

step was nearly 100%, but it was found to be impractical for eluates

with a high concentration of organic because of the formation of a very

large floc (133).

Protein solutions, such as beef extract, had been shown to be very

effective in the elution of viruses adsorbed to membrane filters (64,

107, 136, 138, 183), but this did not allow for reconcentration of the

eluate onto smaller filters. In 1976, Katzenelson and coworkers (92)

developed an organic concentration procedure for beef extract that

involved flocculating the beef extract by lowering the pH of the

solution to 3.5 (the isoelectric point of beef extract). Viruses in

solution adsorbed to these flocs, which were recovered by

centrifugation. Viruses were recovered by solubilizing the flocs in

small volumes of phosphate solution. Results in their laboratory

indicated that this method, used as a second step in conjunction with a

primary adsorption/elution step yielded a mean enterovirus recovery of

about 75%. Organic flocculation using non-fat dry milk and casein (16)

has also yielded efficient (over 70%) recoveries.

The organic flocculation step developed by Katzenelson (92) has

seen much use in the concentration and detection of enteroviruses from

natural samples (31, 69, 71, 76, 95). However, recent studies in

different laboratories (5, 84, 125, 132) have indicated that the source

and lot number of beef extract can effect its ability to flocculate,

thus lowering considerably the efficiency of this step. Payment et al.

(132) found that the addition of 2.5 mM ferric chloride to beef extract

solution which failed to flocculate at pH 3.5 cause the formation of

sufficient floc for efficient virus concentration (over 75% recovery).

In addition to flocculation variabilities, workers (87, 138) have

noted that beef extract concentrates have been shown to be toxic to cell

cultures. Despite these problems, organic flocculation continues to be

the main means of reconcentration of viruses in proteinaceous solutions.

Materials and Methods

A list of chemicals and their sources, and routine methods used in

animal and bacterial virus preparation and assay is presented in

appendix A. A complete list of media and solutions used in cell culture

work is presented in appendix B.


The filters used in this study, their characterization and sources,

are shown in Table 2.

Water Samples

The natural water samples used for indigenous phage concentration

and their concentration are described in Table 3.

Enterovirus Concentration from Tapwater

Tapwater (114 1) was dechlorinated by the addition of sodium

thiosulfate; absence of residual chlorine was ascertained by the

orthotolidine method (2). Following dechlorination, tapwater was

adjusted to a pH between 3.3 and 3.5 by the addition glacial of acetic

acid. Ten milliliters of the tapwater was removed and seeded with a

known amount of one of the enteroviruses tested. From 10 to 10

plaque-forming units (PFU) of virus were added in each trial. When low

numbers of viruses were added, the eluate from the Seitz filter was

assayed directly after adjustment of pH 7 to 7.5. When higher numbers

were added, viruses were assayed after dilution in phosphate-buffered

saline (PBS) with 2% fetal calf serum (FCS) (See appendix B). A 5-ml

portion of the viral suspension was added to the 114 liters and the

remaining 5 ml was assayed. The seeded tapwater was passed through a

10-in (ca. 25.4-cm), 0.2 Um Filterite filter. The adsorbed viruses were

eluted with 1 liter of 0.2 M trichloroacetic acid (TCA)-0.2 M lysine

that had been adjusted to pH 9 with sodium hydroxide. This eluate was

passed through a 47-mm Seitz S filter under positive pressure without pH

adjustment. Viruses adsorbed to the Seitz filter were recovered by one















- 1

0 0


0 Z

4 U

r -4





4J -4

4-1 -H



























Z 1 1 I 1

0 m 0 0 0

0 %o n 0 0n


ul M

0 0 a0 0 0




0 H N *

0 0
i* 0 --1

N t
H Hn

(U U On

Nd 44 *.44 M-
fl >O a4 0

41 0-1 a C

0 e B O
E- I B t
(U I I c CU

r0 LA I

0 e O 4.
4 4 C H 0

0l K0
ei u M i- i x

of two methods. In the first procedure, the filter was removed from the

holder and placed in a holder used for vacuum filtration (Gelman

Sciences, Inc., Ann Arbor, MI). Two 8 ml portions of 3% casitone were

pulled through the filter with a vacuum. After the first 2 to 3 ml of

each portion had passed through the filter, the vacuum was removed and

the solution was allowed to soak the filter for 5 minutes. After the

soaking period, the remaining part was pulled through the filter. The

two portions were pooled, neutralized and assayed. In the second

procedure, two 8 ml portions of FCS (pH 9) were pulled through the

filter with a 50 ml syringe. The two portions were pooled, neutralized

and assayed.

Adsorption/Elution of T7 using Electropositive Filters

Tapwater was dechlorinated as described above. After approximately

107 PFU of T7 phage were added, 10 ml of the tapwater was passed through

the filters, which were held in 25 mm holders, at a rate of 1 ml/s.

Viruses in the initial tapwater solution and in the filter effluent were

assayed to confirm viral retention by the filters. Next, 10 ml of a

test eluent (see Tables) was passed through the filters the eluted

viruses were assayed. Finally, 10 ml of 3% beef extract was passed

through the filter to elute residual virus. The data obtained from beef

extract controls indicated that phage was not being inactivated or

aggregated by test solutions (data not shown). The virus eluted was

expressed as a percentage of the virus present in the initial tapwater

solution. Values were obtained in 2-3 trials and represent the means

and standard deviations of 4 to 6 determinations.

Contact Angle Measurements

A cubic cell (3 cm x 3 cm x 3 cm) was filled with distilled water

and a filter was placed in the cell and allowed to stand for 1 hr in

order to be saturated with water. A small drop of carbon tetrachloride

was injected in the water just above the filter surface using a 50 il

syringe (Hamilton Co., Reno, NV). The contact angle between the filter

and droplet was measured using a NRL Gonimeter model A-100, Rame-Hart,

Inc. (129).

Capillary Rise Measurements

Filters were cut in strips measuring 20 cm long and one cm wide.

These strips were then dipped in distilled water or other liquid one cm

deep. The time at which the strip was dipped in the liquid was taken as

the zero reference time. The rise of liquid on the filter was measured

at one minute intervals for 10 to 12 minutes. The height of the liquid

on the filter (or difference of liquid and water) versus time was

plotted using the mean values of at least three trials (77).

Concentration of Indigenous Bacteriophage using Filters

Samples of water (see Table 3) were collected in pressure vessels

and brought back to the laboratory for immediate processing. A five ml

aliquot was removed to determine the input concentration of

bacteriophage. For the one-step procedure, samples were passed through

an AP20 prefilter (Millipore Corp., Bedford, MS) and 3 layers of

Virosorb 1MDS filters without pH adjustment. The filtrates were

collected and assayed to confirm bacteriophage adsorption. Next, 1.0 M

NaCl in 0.05 M imidazole buffered at pH 7 was used to elute the

adsorbed phage. The eluate volume was 5 ml when filters were held in 25

mm holders, 10 ml for 47 mm filters and 40 ml for 90 mm filters. The

two-step concentration procedure started as the one-step procedure

described above, with 90 mm filters used. The 40 ml sodium chloride

eluate was pulled through a 47 mm Seitz S filter using vacuum filtration

without pH adjustment and collected to check for phage adsorption.

Next, 10 ml of a 4% beef extract and 0.1% Tween 80 solution at pH 7 was

placed on the filter and 2 to 3 ml was pulled through using a vacuum to

displace residual fluid and ensure that the beef extract/detergent

eluate completely soaked the filter. The remaining 7 to 8 ml of the

eluate remained on the filter, soaking for 5 minutes, then was pulled

through and assayed for phage elution.

Flocculation of Seeded Beef Extract

Forty ml of 10% beef extract was seeded with 10 PFU of either

bacteriophage or animal virus. One half of this beef extract solution

was flocculated by lowering the pH to 3.5. The floc was collected by

centrifugation at 3000 x g for 10 minutes, and resuspended in two ml of

0.15 M sodium phosphate solution, pH 9, according to the method

described by Katzenelson et al. (92). The other half was mixed with 2

volumes of a saturated ammonium sulfate solution. Floc immediately

formed when these solutions were mixed, and the pH remained buffered

between 7 and 7.5. The beef extract/ammonium sulfate was centrifuged at

14,500 x g for 20 minutes at 5 100C and the floc was solubilized in 2

ml sterile distilled water.

Concentration of Phage and Enterovirus from Sewage Effluents

Unchlorinated secondary effluent from the University of Florida

treatment plant was collected in a sterile container and brought back to

the laboratory for immediate processing. One liter was seeded with 10

PFU enterovirus (see Table 9) and 5 ml was removed to assay for initial

input of enterovirus and indigenous bacteriophage. E. coli C (ATCC

13706) was used as host bacterium for indigenous bacteriophage assay

(75). The one liter sample was passed through three layers of Virosorb

1MDS filters with an AP20 prefilter encased in a 47 mm filter housing

(Millipore) without pH adjustment. Twenty ml of 10% beef extract, pH 9,

was used to elute adsorbed virus. The beef extract was adjusted to pH

by the addition of 1 N HC1, and divided into two equal portions. One

half of the eluate was reconcentrated using organic flocculation at pH

3.5, while the other half was reconcentrated by ammonium sulfate

flocculation, as described above. The final volume with both procedures

was 2 ml. The procedure was the same when other protein solutions (10%

tryptose phosphate broth and 10% brain heart infusion) were used in the

elution step, except that only one volume of saturated ammonium sulfate

was needed for flocculation.

Recovery of Low Levels of Indigenous Bacteriophage

Samples of water were passed through Seitz S filter with an AP20

prefilter, held in a 90 mm stainless steel holder. Direct assay of

these water samples indicated that no bacteriophage was present. Next,

10% beef extract, pH 9, was used to elute any phage that may have been

adsorbed. Beef extract was dropped to pH 7 and flocculated with

saturated ammonium sulfate as described above.


Concentration of Enteroviruses from Large Volumes of Tapwater

We developed a modified method to concentrate enteroviruses from

large volumes of tapwater (Table 4). We used acetic acid to adjust the

pH of the tapwater to 3.5, in lieu of HC1 (44) to aid in virus

adsorption to Filterite filters. A chaotropic salt, trichloroacetic

acid was used to elute viruses from the filter. Virus was efficiently

eluted, and readsorption of the virus to a second dissimilar filter

(Seitz S) was accomplished without any modification of this eluting

solution. Viruses adsorbed to the Seitz filters were eluted with pH 9

solutions of either 3% casitone or fetal calf serum (FCS). Casitone is

less expensive than FCS, but a soaking period was required when casitone

was used as the eluate. The mean recovery of enteroviruses with

casitone was 54%, and that with FCS was 43%.

Differential Elution of Phage from Positively Charged Filters

Initial studies were conducted to determine the adsorption of

bacteriophage T7 to four positively charged membrane filters: Virosorb

1MDS, Zeta plus C-30, Posidyne N66, and Seitz S. In these experiments,

tapwater was dechlorinated and adjusted to three different pH values

(4, 7 and 9), and in all cases, adsorption was nearly complete (>99%

adsorbed, data not shown). Based on these results, dechlorinated

tapwater was used as the adsorbing solution in subsequent studies with

T7 and electropositive filters.

The results of preliminary studies on the ability of salts and

detergents to elute T7 phage adsorbed to these filters are shown in

Table 5. A solution of the buffer alone (50 mM imidazole) eluted less

than 10% of the adsorbed phage from all of the filters tested. A

TABLE 4. Concentration of viruses from tapwater.

Number of Percent of added
Virus trials virus recovered Eluent
mean SD

Poliovirus-1 3 67 8

Coxsackievirus B3 3 46 18

Coxsackievirus B4 2 43 20 3% casitone, pH 9

Echovirus-1 2 63 12

Total 10 54 18


Coxsackievirus B3

Coxsackievirus B5



35 7

40 7

44 4 Fetal calf serum, pH 9

58 9

43 11

a Tap water was dechlorinated, adjusted to pH 3.3 to 3.5 by the
addition of acetic acid, and seeded with the indicated virus.
The sample was passed through a 10", 0.25-um porosity Filterite
filter at 3 to 5 gal/min. Next, 1,000 ml of 0.2 M NaTCA + 0.2 M
lysine, pH 9, was passed through the filter. This solution was
then passed through a 47-mm Seitz S filter. Virus adsorbed to the
Seitz filter was recovered by treating the filters with two 8-ml
volumes of 3% casitone or FCS, pH 9, as described in the text.

b SD: Standard deviation.

co 4




0 N r 0o

+1 +1 +1 +1

U (CN 00 (N
0 r- Nr (N

+t 0 +1 0

+1 +1 +1 0

co 0c r

+1 0 +1 0

U) m
Q4 r-I U2

0 04 >4
0 N

4.4 4J O
4> N 4 0


















solution of detergent (Tween 80) at pH 7 did not elute any T7 adsorbed

to the Seitz filters and less than 10% of the phage adsorbed to the

Virosorb 1MDS filters. This detergent solution eluted greater then 35%

of the T7 adsorbed to the Zeta plus C-30 or Posidyne N66 filters. A

solution of EDTA eluted 74% of the virus adsorbed to the Virosorb IMDS

filters but less than 5% of the virus adsorbed to the other filters. It

had previously been shown that solutions containing salts and detergents

were very efficient at eluting viruses adsorbed to electronegative

filters (156). When such a solution was used with the positively

charged filters, greater than 70% of the T7 adsorbed to the Zeta plus

C-30, Posidyne N66 and Virosorb 1MDS filters was eluted. However, only

22% of the T7 adsorbed to the Seitz S filters was eluted.

Preliminary results indicated that the ionic strength of the

eluting solution affected the amount of virus eluted from these filters.

Figure 1 shows the influence of ionic strength in solutions of detergent

at pH 7 on the elution of T7 adsorbed to membrane filters. In the

presence of detergent, solutions with ionic strengths of 0.4 or above

efficiently eluted the bacteriophage adsorbed to the Zeta plus C-30 and

the Posidyne N66 filters, while solutions with ionic strengths below 0.4

eluted the virus adsorbed to the Virosorb 1MDS filters. Again little or

no virus adsorbed to the Seitz S filters was eluted.

The effect of detergent concentration in the presence of 1 M NaC1

on viral elution was determined (Figure 2). Again, a solution of buffer

alone at pH 7 eluted less than 10% of the T7 adsorbed to all filters

tested. A solution of 1 M NaC1 at pH 7 did not elute any virus adsorbed

to either the Zeta plus C-30 or Seitz S filters, while less than 20% of

the adsorbed T7 was eluted from the Posidyne N66 filter. However,

FIGURE 1. Elution of bacteriophage T7 adsorbed to various positively
charged filters by solutions of 0.1 % Tween 80 and NaCl or
Na2SO4 at pH7.

Virosorb I MDS

Zeta plus C-30

Posidyne N66

o.II 0
0.1 0.2






Seitz S



= mm Z Z


II i-C .
C r+


o z- 0

D< M


29 o

rt "

II .


M En





N (

coc0oooooocoo Z



a 000 0 0 0 0 0 Z
0 0 0

fa A

o co w Cv

031n173 sn7IA %

greater than 75% of the T7 adsorbed to the Virosorb 1MDS was eluted

using the salt alone. Tween 80 concentrations of 0.001% in the presence

of NaCI greatly increased the amount of phage eluted from the Posidyne

N66 filters, yet had no effect on the elution of phage adsorbed to Zeta

plus C-30 or Seitz S filters. A solution of 0.1% Tween 80 in presence

of salt eluted created than 75% of the T7 adsorbed to the Virosorb 1MDS,

Posidyne N66 and Zeta plus C-30 filters, yet had no effect on elution of

T7 adsorbed to the Seitz S filters. Tween 80 was not soluble in the

salt solution in higher concentrations.

Indigenous bacteriophage from several water sources (Table 3) were

concentrated using a one step adsorption/elution procedure using

Virosorb 1MDS filters as the adsorbent and 1 M NaC1 as the eluting

solution. Table 6 shows the percent recovery of indigenous

bacteriophages using filters of various sizes and input volumes of

water. In general, recovery of virus was better when the smaller filter

holders were used. In addition, there was no significant difference in

viral recovery using E. coli B or E. coli C-3000 as the host bacteria.

The average recover of the 41 samples from 4 different sources was 55%

using E. coli B as the host and 61% using E. coli C-3000 as the host.

In samples with relatively high numbers of phages, a one-stage

concentration procedure may be sufficient to detect viruses. However,

where low levels are found, a two-stage procedure may be required to

provide a sufficient concentration factor to detect indigenous phages.

The results of this study have shown that some solutions that permit

virus adsorption to one filter are capable of eluting virus adsorbed to

another filter. This raises the possibility that a simple two-step

virus concentration procedure can be developed in which virus adsorbed

to one filter can be eluted using a solution permits virus adsorption to

0 o

o 0 C o r-4 v
t.0o co t nm r










a )



o +

u 1

-4 L L

Io oI

In n 000
r r 000

n ro un 0 Ln r o
(i vr m7 (N Is T' 0) N Iq cN IV ON

MmO N HrI-n m r r IT
M m (N r-

mrm r-T O(N (NOM
m co w N co m







~~aw ma m mmN

IV 110 O m ve O M CN

0) C14 ko HI 0 rY %J00 mm 1 l
1-4 Hl H- H- l-I N lil m l-l

a second smaller filter. The elution-adsorption step can be

accomplished with little or no modification of the solution. Based on

the results of this study, several two step procedures using two

different filters in series are possible. One such procedure is

outlined in Figure 3. We have used this procedure to concentrate phage

from several secondary effluent samples (Table 7). Using this

procedure, the average recovery of bacteriophages was 26% using E. coli

B as the host, and 63% E. coli C-3000 as the host. In two samples,

bacteriophage was detected in the final concentrates when none detected

in the initial sample (data no shown).

Characterization of Virus-Adsorbing Filters

In order to determine the relative hydrophobic-hydrophilic nature

of filters used for virus adsorption, the contact angle of carbon

tetrachloride on membrane filters submerged under water was measured

(Table 2). The contact angle of carbon tetrachloride on the Seitz S

filter was the smallest (1130), the contact angle on the Virosorb 1MDS

was the largest (1510), while the contact angles on the Zeta plus C-30

(1280) and Posidyne N66 (128*) were intermediate. Under the conditions

used to measure the contact angle, carbon tetrachloride was spread

immediately on the two negatively charged filters tested (Millipore HA

and Filterite), therefore, measurement was not possible.

The filters were further characterized using the capillary rise

method. Figure 4 shows the rise of deionized water on positively

charged membrane filters. The most rapid raise of water was observed in

the Virosorb 1MDS filters, while the smallest rate of water rise was




















0 nO)












o **


m nO In 0 +1

m Ln H %T
H (N (N +1


x x

\0 CO
















FIGURE 4. The rise of water on various positively charged filters.



-4-- Viroorb I MDS

2 --0---0- Zeta plus C-30
-A---A Seitz S
----*- Posidyne N66


2 4 6 8 10 12

TIME (min)

observed for the Posidyne N66 filters. Zeta plus C-30 and Seitz S

filters were intermediate in nature. Figure 5 shows the results of the

same test on negatively charged filters. The rise of deionized water on

Filterite filters was more rapid than that of Millipore HA filters.

Figures 6 through 8 show the change in rise (solvent water) of

the positively charged filters. In all cases, the change in rise was

the most positive for the Seitz S filters, while the difference was most

negative for the Virosorb 1MDS filters. As with contact angle

measurements, Zeta plus C-30 and Posidyne N66 filters were intermediate

in nature. The change in rise (chloroform water) was plotted for

negatively charged filters (Figure 9). In this instance, there was

little or no change in rise for Millipore HA filters, while this change

was slightly negative for the Filterite filters.

Ammonium Sulfate Flocculation as a Reconcentration Step

The ability of certain salt solutions, such as sodium phosphate,

sodium citrate, ammonium sulfate or sodium sulfate, to precipitate

proteins has been well documented (73, 93, 146). Preliminary

experiments in our laboratory indicated that optimum precipitation

occurred when beef extract was mixed with saturated ammonium sulfate.

The efficiency of this flocculation step was compared to the organic

flocculation procedure as described by Katzenelson et al. (92) and the

results are shown in Table 8. Bacteriophage recovery using organic

flocculation was less than 3% for all three phage tested, while

virtually all was recovered using ammonium sulfate flocculation. In

addition, animal virus recovery was high (greater than 70% for all

three tested) using the ammonium sulfate method as compared to organic

flocculation, which gave the highest recovery with poliovirus-1 at 53%.

FIGURE 5. The rise of water on various negatively charged filters.

A A Millipore HA
A A Filterite

TIME (min)

FIGURE 6. The change in capillary rise on various positively charged
filters: carbon tetrachloride water.


S,/, 4 6 8 10
S' TIME (MIN)' '

I -5D



FIGURE 7. The change in capillary rise on various positively charged
filters: butanol water.


v TIME (MI4)

- -2

5I -4

J -5

Z -87

. -9

I -10





FIGURE 8. The change in capillary rise on various positively charged
filters: chloroform water.


o 4.

oe -2m

2 4 6 8





FIGURE 9. The change in capillary rise on various negatively charged
filters: chloroform water.






0.5 1



TABLE 8. Virus recovery from seeded 10% beef extract, pH 7, by ammonium
sulfate flocculation versus organic flocculation.

% Recovery
Ammonium Sulfate Organic
Virus Flocculation Flocculation

MS2 101 8a 1 0

XX174 100 17 1 11

T3 107 8 2 1

Poliovirus-1 97 11 53 8

Echovirus-5 84 12 22 5

Coxsackievirus B5 72 8 10 1

a standard deviation

It should be noted that during the course of these experiments, three

lots of beef extract were tested before one was found that flocculated

at low pH. Both in our laboratory and in others (71, 92, 132), higher

recoveries using the organic flocculation procedure have been obtained

with some lots of beef extract.

This ammonium sulfate flocculation procedure was compared with the

organic flocculation procedure for recovering viruses in natural water

samples (Table 9). The overall mean recovery of indigenous

bacteriophage using the ammonium sulfate flocculation was 85% as

compared to 12% for the organic flocculation method. It was found that

the mean recovery of seeded enteroviruses using the ammonium sulfate

flocculation method was 73% as compared to 42% for organic flocculation.

Other proteinaceous solutions were tested for enterovirus and

indigenous bacteriophage recovery using this ammonium sulfate

precipitation method, and these results are shown in Table 10. For the

recovery of indigenous bacteriophage, there was no difference for the

solutions used. Ten percent tryptose phosphate broth (TPB), 10% brain

heart infusion (BHI) and 10% beef extract all recovered 71% of the phage

when used as a primary eluting solution followed by ammonium sulfate

flocculation. For three enteric viruses tested, the mean recovery was

76% when beef extract was used as the primary eluting solution, 62% for

BHI and 73% for TPB.

Finally, this procedure was also used to recover indigenous

bacteriophage in less heavily contaminated waters (Table 11). Direct

assay of these waters indicated that no phage was present. After

reconcentration of the beef extract eluate using ammonium sulfate

TABLE 9. Recovery of indigenous bacteriphage and seeded enterovirus
from sewage effluent by ammonium sulfate flocculation versus
organic flocculation.

% recovery % recovery

Trial Virus added ASFb OFc ASFb OFc

1 Poliovirus-1 47 33 115 15

2 Poliovirus-1 80 71 86 15

3 Echovirus-5 91 39 61 11

4 Echovirus-5 73 23 76 5

Mean SDd 7316 4218 8520 124

a. Unchlorinated effluent was seeded with 107 PFU of enterovirus
listed and assayed for initial enterovirus and indigenous
bacteriophage. One liter of sample was passed through a 47mm, 3
layers of Virosorb 1MDS filters. Adsorbed virus was eluted with
20ml 10% beef extract, pH 9. Beef extract was adjusted to pH 7 and
split into equal halves (see text).

b. ASF ammonium sulfate flocculation.
eluate was mixed with two volumes of
floc was centrifuged and resuspended

One half of beef extract
saturated ammonium sulfate,
in 2 ml distilled water (see

c. OF organic flocculation. One half of beef extract eluate was
adjusted to pH 3.5, floc was centrifuged and resuspended in 2 ml
0.15 M Na2HPO4 (see text).
2 4

d. SD standard deviation.

iN r-4 ( o0 Ln cN
N- r. co 0 %D 10
~rlmbD LO

0" 0 r- (0
in r- P. 0

to c


U > 0i
Z: 0
0 Io

z (D
W 4

E -

0 0


m u
in in

S3 34

> > H4 -4 0 0
0 0 l C > >

0 H 0 0 ) U
-4 r-4 0 0 x x
0 0 U U 0 0
4 4 C a d U

r-4 (N e n Lo LD

10 *H
.4- 4-4

0 .C 0
a) C4 -4a

Ul O U
4-0 M U1

0 r-a '
rU W d< 0 4-)
UC 44 1
1)'0 0 0
cu 0 m -
U) -4

- ,-o 0 0
Q4 ) 4-

.J1 w H
.1 40U)

> 40 (0 a
0 0 U4-4 10

41> Q0 U -4
rC 41 01
Q) -4 0 4
4-4 q -4 .J
4 U) m

40 ) >U C <
C 0 5

.a Oa 4 u
o n a r.(

r- 10001

*4 4 0 04

.-- > u n 4
. 4 f 1

(0a :j U) -1-4
0 04-J ,0 v

0 C-4 40
4) 10 .- 4- U)

(O 0 >(a0
4. 0 C u
4.) M4-4 0- V

4.> 6 4

= 44) r a 1
-4 >i C
U 4-' -4 U) r
^1 >)^ 4
Q 0 0)'0 -r-
r- 7 >, a)
j3 ) l0 J 4

0) IT Ln 0 L. n u J

N 1' 0 r, 0
r- N

TABLE 11. Recovery of low levels of indigenous bacteriophage by
ammonium sulfate flocculation.

SAMPLE Volume (ml) PFU/ml Volume (ml) PFU/ml Volume (ml) PFU/ml

effluent 3200 <1 42 <1 5 12.5

Lake water 2000 <1 34 1400 5 8000

Lake water 2500 <1 50 234 3 4350

a. Sample was passed through Seitz S filter with AP 20 Prefilter held
in 90 mm stainless filter holder. Next, 10% beef extract, pH 9,
was used to elute adsorbed virus. Beef extract was dropped to pH
7, and mixed with two volumes saturated ammonium sulfate to form
floc. Floc obtained by centrifugation was resuspended in sterile
distilled water (see text).

precipitation, phage was detected in all three samples, and there

appeared to be no probelms, such as toxicity, when the final concentrate

was plated directly.


Concentration of Enteroviruses from Large Volumes of Tapwater

Several concentration procedures to detect viruses in tapwater have

been developed and subsequently modified. Wallis et al. (178) and

Sobsey et al. (165) described two-stage concentration procedures in

which the same type (but different sizes) of filters were used in both

stages. The use of the same filter in a reconcentration procedure

involved reconditioning of the primary eluting solution to allow for

readsorption. Problems with clogging and co-concentration of organic

compounds (44, 162) and other impurities made a second stage

concentration procedure with negatively charged filters impractical.

This problem was overcome by the use of an inorganic flocculation step

for reconcentration (44), yet viral recovery was low. Katzenelson et

al. (92) increased the amount of viral recovery by using a second-stage

organic flocculation procedure with beef extract. While this procedure

has proven to be useful, problems such as large final volumes and

variability in the flocculation of beef extract supplies led us to

investigate possible modifications of currently used methods.

Our investigations led to three modifications of existing

procedures for concentrating enteroviruses from large volumes of

tapwater. The first modification is the use of acetic acid in place of

hydrochloric acid to adjust the pH of the tap water to pH 3.5.

Adsorption of viruses to negatively charged filters such as Filterite

filters is facilitated at a low pH (165). Hydrochloric acid has been

used to adjust water samples to the low pH values required for viral

adsorption in many studies. We have found that acidification of

tapwater with either hydrochloric or acetic acid permits the adsorption

of greater than 99% of the enteroviruses studied to Filterite filters.

However, the acidification of water under filed conditions with Dema

injectors (134) is easier when acetic acid is used in place of HC1.

Changes in the flow rate or pressure may cause fluctuations in the

amount of acid injected when Dema or pressure injectors are used to

adjust water pH. With HC1, these changes may cause wide fluctuations in

the pH unless the injection rate is carefully controlled. We have found

that these fluctuations are minimized when 1 or 2 M acetic acid is used

for acidification. Acetic acid (pH = 4.75) will buffer more closely to

the desired pH range than will HC1, which dissociates completely.

The second modification was the use of sodium trichloroacetate

(NaTCA) as the primary eluate. The association of viruses with solids

such as membrane filters has been shown to be influenced by both

hydrophobic and electrostatic interactions (40, 49, 50, 53, 156).

Previous studies have shown that detergents, which disrupt hydrophobic

interactions, can be used to elute viruses adsorbed to membrane filters

(49, 156). In addition, recent studies have shown that certain salts

may disrupt hydrophobic interactions (49). These salts, called

chaotropic salts, are relatively large, singly charged ions such as TCA

and thiocyanate. It has been suggested that these compounds disrupt

hydrophobic interactions by decreasing the structure of water and making

aqueous solutions more lipophilic. Chaotropic agents have been used to

elute viruses adsorbed to various solids by disrupting hydrophobic

interactions between the viruses and the solids (48, 49, 177). A

solution of the chaotropic salt, NaTCA, eluted 77% of the enteroviruses

used in this study that were adsorbed to Filterite filters.

Reconcentration of the viruses in the primary eluate with a second,

smaller electronegative filter requires acidification of the high-pH

eluent used to recover viruses adsorbed to the primary virus-adsorbing

filter (165, 178). It was previously shown (44) that such a drop in pH

may be accompanied by the precipitation of organic compounds that may

have been concentrated along with the viruses. These organic compounds

may interfere with the reconcentration process by reducing the

adsorption of viruses to the filters or by clogging the filters (44).

Therefore, readsorption of viruses to a second filter at the same pH

that was used for the elution of viruses from the first filter was


Sobsey and Jones (164) studied the adsorption of poliovirus-1 to

positively charged filters at several pH values. They found that both

Zeta plus filters and Seitz filters efficiently adsorbed polioviruses at

a relatively high pH. The last modification of our procedure introduced

was the use of Seitz filters for the second-stage concentration. When

the NaTCA-lysine eluate from the Filterite filters was passed through

the Seitz filters, greater than 95% of the viruses were adsorbed.

Viruses adsorbed to the Seitz filters were eluted efficiently with pH 9

solutions of either 3% casitone or fetal calf serum.

These three modifications make the pH adjustment step easier, and

permit detection of viruses in water by a two-stage concentration

procedure in which different types of membrane filters are used in each

stage. This use of different types of filters allows readsorption

without any modification of the primary eluate, simplifying the overall


Differential Elution of Phage from Positively Charged Filters

The use of negatively charged filters, such as the Filterite

filters used in the primary adsorption step discussed above, requires

the lowering of the pH of the water sample and/or the addition of

cations (178, 181) to facilitate viral adsorption. Since it had been

previously shown that these concentration procedures are unsuitable for

bacteriophage recovery (149, 150), investigators have explored the

possibility of the use of electropositive filters for the concentration

of bacteriophages (67, 150, 157). These filters are more positively

charged at the pH range of most natural water samples (pH range 6.5 8)

than are the electronegative filters, and so adsorption to the

electropositive filters can be accomplished with little or no

manipulation of the water sample.

We have examined four filters, Virosorb 1MDS, Zeta plus C-30,

Posidyne N66 and Seitz S filters. All four are more electropositive at

pH 5 8 than are negatively charged filters and adsorb virus in water

at these pH values. However, their composition is quite varied (Table

2). We have attempted to determine what effects these differences in

composition would have on the association of bacteriophage with the


Although all of the filters adsorbed T7 in tapwater at pH values

between 4 and 9, differences in the ability of certain solutions to

elute the adsorbed phage were noted. Previous studies have shown that

certain solutions of a neutral detergent, such as Tween 80, can disrupt

hydrophobic interactions between viruses and membrane filters (40, 50,

181). Once the hydrophobic interactions have been disrupted, the

addition of certain salts to this solution will disrupt electrostatic

interactions and result in the elution of adsorbed viruses (50, 156).

When hydrophobic interactions were disrupted at pH 7 by the addition of

detergent, only T7 adsorbed to Posidyne N66 and Zeta plus C-30 filters

was eluted to any appreciable extent. If salt was added to the

detergent solution, thus disrupting both hydrophobic and electrostatic

interactions, most of the adsorbed phage was eluted from all filters

tested except the Seitz S filter. These findings indicated that there

were distinct differences in the relative strength of hydrophobic and

electrostatic interactions with the phage among these four filters.

It was found that the ionic strength of the eluting solution, in

the presence of a detergent, affected the amount of adsorbed phage that

was eluted. Solutions of detergent with sodium chloride or sodium

sulfate with ionic strengths greater than 0.4 eluted most of the

adsorbed virus from all filters tested except the Seitz S. It is

assumed that increasing the ionic strength decreases the strength of the

electrostatic interactions between the virus and the filter (156).

Therefore, the filters that form relatively strong electrostatic

associations with viruses require solutions with relatively high ionic

strength to elute adsorbed viruses.

A titration of the amount of detergent necessary to elute adsorbed

T7 in the presence of 1 M sodium chloride provided the most striking

results. A solution of sodium chloride alone eluted most of the T7

adsorbed to the Virosorb 1MDS filter. Since no detergent was added, it

can be assumed that no hydrophobic interactions were disrupted. Since

salts can disrupt electrostatic interactions but promote hydrophobic

interactions (40, 49), this result can be explained by assuming that

hydrophobic interactions are not important in the association of phage

to the Virosorb 1MDS filter. Similarly, since a solution of a salt

alone did not elute viruses adsorbed to the other three positively

charged filters, some hydrophobic interactions are probably involved in

maintaining viral association to these filters. Since less detergent in

the presence of salt was necessary to elute phage adsorbed to the

Posidyne N66 filter than was required to elute phage adsorbed to the

Zeta plus C-30 filter, it can be assumed that hydrophobic interactions

are relatively stronger for the Zeta plus C-30 filters than for the

Posidyne N66 filters. Again, these specific solutions eluted little or

no phage adsorbed to the Seitz S filters.

Characterization of Virus-Adsorbing Filters

These results led us to characterize further these four

electropositive filters, as well as two electronegative filters often

used in virus adsorption schemes (Filterite and Millipore filters). We

used two means of characterizing the surface of these filters, which can

be done without destruction of the filter surface. One such test was

contact angle measurement. The contact angle determines the affinity of

a liquid (in this case carbon tetrachloride) for a surface submerged

under water. If carbon tetrachloride preferentially wets the submerged

filter, the resulting contact angle between the filter surface and

carbon tetrachloride droplet will be small (129). This would, in this

case, be an indication of the preferential hydrophobicity of the filter.

It was found that Virosorb 1MDS filters had the largest contact angle,

indicating that it was the least hydrophobic of the filters tested.

This was consistent with elution studies which had previously shown that

hydrophobic interactions were not important in phage association with

the Virosorb 1MDS filters. The Seitz S filters had the smallest contact

angle, and therefore were found to be the most hydrophobic. This may be

one reason why elution from these filters using defined solutions was so

poor. Indeed, in our laboratory and others (70, 155, 163), the eluting

solutions found to be most effective for virus adsorbed to Seitz S

filters are undefined protein solutions, such as beef extract.

Consistent with elution studies, Zeta plus C-30 and Posidyne N66 filters

were found to be intermediate in the hydrophobic nature of their

filters. Due to the nature of the experimental setup, contact angles on

the two negatively charged filters were not able to be determined.

The second method used for characterization of the filters was the

capillary rise method. In this case, the movement of a liquid through

the filters is measured (77). If the liquid is water, the rate of rise

is related to the hydrophilicity of a filter. Those filters which are

more hydrophilic will have a correspondingly higher rate of rise of

water. The results of this test were also in agreement with elution

studies. The Virosorb 1MDS filters, previously shown to be least

hydrophobic were found to be most hydrophilic, while Zeta plus C-30,

Seitz S and Posidyne N66, respectively, were found to be less


The rate of rise of some hydrophobic solvents, such as carbon

tetrachloride, chloroform, or butanol, when measured and plotted as the

difference between the solvents and water, gives another indication of

the relative hydrophobicity of these filters. These capillary rise

measurements were in complete agreement with both elution studies and

contact angle measurements. In all cases, the Virosorb 1MDS filters

were once again shown to be the least hydrophobic, while the Seitz S

filters were found to be the most hydrophobic of the positively charged

filters tested.

These capillary rise measurements were also done with two

negatively charged filters. Although both Millipore and Filterite

filters are electronegative, their composition is different (Table 2).

It was found that Filterite filters were more hydrophilic than the

Millipore filters, while the hydrophobic nature of both these filters

appeared to be slight, with Millipore being slightly more hydrophobic as

evidence by little of no change in capillary rise for chloroform -

water. Differences in the elution patterns of these two filters have

been documented by researchers (42). It was found that elution of

poliovirus-1 adsorbed to Filterite filters could be accomplished using

casamino acids or individual amino acids such as glycine or lysine,

while these solutions had little effect on the elution of viruses

adsorbed to Millipore filters. Since hydrolyzed proteins, such as

casamino acids have been increased number of free carboxyl and amine

groups, the ionic strength of the solution is also increased (42). It

is likely that this effect of increased ionic strength promoting elution

of adsorbed viruses from Filterite filters is related to the relative

hydrophilicity of these filters or the relatively more hydrophobic

nature of the Millipore filters.

The availability of Virosorb 1MDS filters in a pleated-sheet

cartridge allows for large volumes of water to be processed. Sobsey and

Glass (163) found that the concentration of poliovirus-1 from seeded

tapwater was equally efficient using Virosorb 1MDS filter cartridges or

the highly electronegative Filterite filters, yet the Virosorb 1MDS

filters were much simpler to use, due to little preconditioning of water

necessary for adsorption. The Virosorb lMDS filters have been used to

concentrate indigenous virus from sewage, well water and chlorinated

tapwater collected during an outbreak of gastroenteritis (76). The use

of Virosorb 1MDS filters in the concentration of animal viruses led us

to develop a method for concentrating bacteriophages with them. These

filters were found to adsorb indigenous phage from the four natural

water samples tested without any modification or treatment of the

sample, and concentration of the indigenous phage was accomplished by

using sodium chloride by pH 7 for elution. The results of this study

indicated that sodium chloride, which could be used to elute virus

adsorbed to the Virosorb 1MDS filters, was ineffective for eluting

viruses adsorbed to any of the other three electropositive filters

examined. This raises the possibility of using two different filters in

series for a simple two-step concentration procedure for bacteriophages

based on differences in the elution characteristics of the filters used.

This method has the advantage of maintaining a constant pH value

throughout the concentration procedure, thus minimizing inactivation of

the bacteriophage due to pH extremes. In addition, the method is

relatively simple and requires no costly instrumentation, and,

therefore, it is suitable for field use. One possible combination of

filters was used successfully in this study to concentration indigenous

bacteriophages. It is likely that other filters and eluting solutions

may be used in improved procedures.

Ammonium Sulfate Flocculation as a Reconcentration Step

The development of the organic flocculation method by Katzenelson

et al. (92) greatly simplified the reconcentration of viruses in beef

extract solutions, and has been incorporated into many concentration

schemes (44, 64, 70). A major difficulty with this procedure has

developed recently, that is, the variability of different lots and

sources of beef extract with regard to flocculation capacity (5, 84,

125, 132). Another problem with organic flocculation is that the use of

low pH prohibits its use for the reconcentration of bacteriophage, which

have been shown to be sensitive to extremes in pH (149, 150). Our use

of a saturated ammonium sulfate solutions to cause flocculation

eliminates both of these problems. Ammonium sulfate flocculation of

beef extract containing bacteriophage was found to be very superior to

organic flocculation. Moreover, animal virus recovery was also improved

by ammonium sulfate floccuation.

We incorporated this ammonium sulfate flocculation step into a

procedure to recover seeded animal viruses and indigenous bacteriophage

from wastewater effluent samples. The results using this method were

consistently higher than the recoveries using organic flocculation.

The recoveries of bacteriophage were not unexpected, as the pH drop to

3.5 necessary for organic flocculation probably inactivated many

indigenous phages in the beef extract eluate.

The availability of organic flocculation as a reconcentration step

has made beef extract the eluting solution of choice in many

concentration procedures (44, 64, 70). However, other protein solutions

may work as well as, if not better than beef extract, yet

reconcentration becomes a problem (70). Ammonium sulfate flocculation

has been found to work efficiently with other protein solutions, such as

typtose phosphate broth and brain heart infusion.

This procedure was also used to recover very low levels of

indigenous bacteriophage. When bacteriophages are serving as models for

enteric viruses, or as indicator organisms, the use of a concentration

method suitable for both animal and bacterial viruses would simplify

assay procedures considerably. It appears that ammonium sulfate

flocculation works equally as well with animal viruses as with


Ammonium sulfate flocculation appears to be an efficient method for

reconcentration of beef extract eluates when a fluctuation in pH is

undesirable, or when a beef extract lot fails to flocculate.


Review of the Literature

Enteric viruses can be excreted in concentrations as high as one

million viruses per gram of feces (52, 68, 158), and enter domestic

wastewater. Although large numbers of viruses may be present in

wastewater, conventional wastewater treatment plants have essentially

been designed to reduce the amount of organic material and suspended

material discharged to natural waters, and not to reduce pathogens. In

fact, the only step intentionally included for the reduction of

microbial pathogens is disinfection, usually via chlorination (166).

However, several of the steps in wastewater treatment facilities do

reduce the population of microorganisms in raw sewage.

Raw sewage may go through several treatment processes before

discharge from a treatment plant (Figure 10). This process usually

involves primary treatment followed by secondary treatment. In some

cases, wastewater is further conditioned by tertiary treatment.

Primary treatment is the first major process in conventional

wastewater treatment facilities. This is essentially a physical process

designed to remove heavy solids and other materials. The removal of

viruses during this stage of treatment is minimal at best (68, 166).

Effluent from primary treatment often passes on for secondary treatment.

Secondary treatment is a biological process that will metabolize and

flocculate soluble organic in the sewage. One type of secondary

treatment is the activated sludge process. In this system, air is








S 0

0 U

H 1 (0 t
>4. t0.4 U
z O U




-4 C

? u w u


O a
1-4 0

Z -.




1 .1

U- C

44 o



0 0


.4 1

,4 0
'O 0

o 0
H 0

>. C
-4 =3







added, creating an aerobic environment to ensure adequate mixing of the

sludge. This "activated sludge" contains organisms that are thought to

be the main agents responsible for production of various polymers from

the oxidation of soluble organic matter that aid in flocculation (12).

After settling, the liquid portion is removed (possibly for further

treatment) and sludge flocs are separated into two portions. One is

called "return sludge", and is used as an inoculum for primary effluent

entering the aeration tanks. The other part is called "wasted sludge",

which will undergo further treatment, then disposal. The activated

sludge treatment is very effect at removal of viruses from the liquid

portion of wastewater (68, 166).

After final treatment, both sludges (115, 170, 186) and wastewater

effluents (68, 135, 185) have been demonstrated to still contain viable

virus particles. Disposal of these "treated" liquids and solids is,

therefore, of major public health importance. Some methods used for

their disposal are shown in Table 12. As can be seen, in one form or

another, treated sewage is likely to be disposed of on land. Land

spreading of wastewater sludges and effluents has many advantages,

(Table 12) such as crop and soil renovation due to the addition of

nutrients, and water conservation and aquifer recharge. A major problem

with land disposal methods is groundwater and surface water

contamination as a result of virus transport through the soil (174).

With this in mind, it is important to look at the factors involved

in the association of viruses with solids, such as sludges, soils and



0 1 0 01
0 > 0 -4

o ,- ..C-,-4 C C :
S44 .,- 4J o 0 o o
n 00 3d -3H U -I

4 4 W C 0, 4-4 0 w "1 m "41 (i "1
o U 0 .14 -4 0 0 0 0 4
(12 4 t1 E -4 0 Z -HO
.l 2 c0o 0 o- '> 0 '
1 -,o 4 ,4i .,. o u 1 4 o e o
n 0 )0c c c 0 1-i o 0 0 5
S 4- 0 0 0 4J 0 4-4 4- 44 ,*HO 4 4
.(d O U u 0 .r0 4 r. Ja 0 r
O H -H O 0o 0 0
l 0 0 c p 0 0 0 o .
> -H (0 4 .-4 0 0 u z u 0
cc B ) 0 0 *)- H 0 4-i .1 .0 4-4
4j 41 4. 41 ( 3 4 .1 4-J H .H 4-4 e 4 4-H 4
.i < )4- (d (d -4 (d ( 4-1 Q 4-1 0 MU0 4-JO Q
4-4 H 9: r :r3 M(U 4J (3 N 44- Z 4) ( Z 4-
10 0 ,-.-- 44 0 j r' rU 0 0 z 0 .z
U 04r.W44 r -4-H -4 0 3-,144 3:
a E o 00 40 c to c

0 :3 S 24 4 'd 043 5. --
C-4 P4 w w Ow al' $ e 0 c > 0 0 > 0
co *-4. 3 -4A 0)0 r 0 44 W 0 14-

4 H C w 4-4

o S O o
0 a) <,) 0

ft 44 4 0 41 0
m HO 'UV C44O CO '

C O 4-- 401 4 90D: m z
4 4uO 44 0 4 (4- 4 M U a 4- o4
- > r-l W 4J -4 > 3 41 -4C
o o to Mt 4o 00 "dt 1 0 m 0
) a, a >i30 0o 0 u W-
Sa 4 04 0 04-4 4- 4-1 ( 0 4 404- 0 -4 0
S> -H (0O W04 4 w .4 -' -d >,44-' W -404 -H 41 404 04 4t Ql)

41 cO C k
0 0 ) 0 ) (d '0 Q) ra (A' a) 0ra U C ,-4-
V4-'-HO 04V V 030'U0V-J *.-4l 04 U *i-fl

a a -4 W M 1-4 & C a w 0d 4& 4 -4 4

if 4J tp 0 04 V 4) C (0 04 V 4l UO
U) V 0 04 a 0ta 0 k 0 0 'U A

4J I- i 4J 1 (n \ O i- C O
4 O O 4 c< o
4-4e C O

41l 44 0--
t 0 f(S 4-4 r. O V
4- E-1 -H 'U
O -q -44 00 Q ( :
S> 0 o 0 c l
40 O 4-) 'U Ct( rO V- r 0
mf 0 'Ud u o4 ?a -H 4Ui >n (a F 0 0 0

Oi (n O .C r-i ( mu1
a4 0 '0 .404 3-4 r. 0 04 90 6 00
3 r-I U r U 0404 0 4 t-4 U (d- 03 41J
0 W4 44 UZ -r4 1$ m
44 4 -1 4.) 10 4) a) (a r- 0f 4
(12 l r)0 r-4 OL -4 *z *H U > 4
UL) w 41 w 0I 4J A 4J 'U 0 r-I aa
V In ( d 4 tn r.i rr -4 -44-4

0 0 1 *)44 04--4 4'1 -0
4 3o 1 :3 4 a : 0 3) a aM
4J .-I 0 -4 )-404040 (0U241 0 U4-)
D 4-4' t4J 44 U r 4- 4-4 U) d4J -4 a -4
0 44 U 44 (a 44 'U ) 2

o 0 0 0
V E-4 4J 0 44 0
( i<
U -H 0 44 <
0 -' an
4-4P 44 t- l V
4 4 -H -H -A1 r01
H '4-4 ( L-a uH to

4 -4 (ara 'Vr
A.4 Q3 04 :sW 0 -4
E 0 (2

Association of Viruses with Sludge Flocs

Researchers in various laboratories have shown that virus

adsorption to sludge flocs occurs rapidly (13, 24, 118). Using

coxsackievirus A9 and poliovirus-1 (Mahoney), Clarke et al. (24) found

that within 4 hours, 99% and 90%, respectively, of these viruses were

associated with the sludge flocs, and that the removal of the viruses

from the liquid to solid fraction conformed to the Freundlich isotherm.

Using a laboratory-scale activated sludge unit, Malina et al. (118)

found that over 95% of poliovirus-1 was adsorbed to the sludge in 10

minutes. Both of these studies indicated that viruses were inactivated

over time once adsorbed.

Recent work (131) has indicated that the type of sludge can

influence the amount of adsorption seen. After one hour, only 57% of

seeded poliovirus-1 was found to be associated to mixed liquor, 70% to

anaerobically digested sludge, and 95% to aerobically digested sludge.

The authors suggested that the lower pH values found in aerobically

digested sludge may explain the difference in adsorption.

Due to the complex and variable nature of sludge, little or no work

has been done on the determination of viruses to sludge. Recent work in

this laboratory (32) explored the factors involved in bacteriophage

association to sludge flocs. Using various salt and detergent

solutions, it was found that hydrophobic interactions do not appear to

be a major factor in promoting viral association to sludge flocs.

Eluting solutions of increasing ionic strength were found to increase

the percent elution of indigenous bacteriophage adsorbed to sludge. The

author hypothesized that electrostatic interactions appeared to be the

major force involved in the association of bacteriophages with sludge


Methods for the Recovery of Viruses from Sludge

Because viruses have been shown to retain their infectivity when

adsorbed to solids (122, 143), it is important to monitor their fate

during and after treatment. To follow the fate of viruses during

treatment and disposal of sludge, methods have been developed to detect,

and in some cases, concentrate viruses adsorbed to sludge flocs.

An early study by Clarke et al. indicated that virus appeared to

be inactivated once adsorbed to the sludge flocs (24). However, the

solution used to elute the adsorbed viruses (versene and buffer

solutions) has been shown to be a poor eluting solution (49), and the

conclusion drawn by the investigators may be inaccurate. In addition to

the choice of eluting solution used, studies have shown that viruses may

be embedded in, as well as adsorbed to, sludge and therefore some

mechanical means is necessary to agitate the sludge-eluting solutions

mixture (47, 85, 186).

Hurst et al. (85) found that the method of agitation and the

eluting solution used affected the amount of poliovirus-1 recovered from

sludge flocs. When mixing with a magnetic stirrer was used, recovery

was greater if 0.05 M glycine was the eluting solution. If sonication

was used, then 3% beef extract as the primary eluate gave higher

recoveries. In the final procedure, elution was accomplished using

glycine at pH 11 and stirring the glycine-sludge mixture, followed by

concentration of the viruses using inorganic flocculation or adsorption

to filters. Poliovirus recovery from anaerobically digested sludge was

55%, while recovery from aerobically digested sludge was 25%.

In another study, three techniques were used to free embedded

virus: sonication, mechanical stirring and freon extraction (186). It

was found that no one treatment gave consistent results, but a larger

number of isolates were recovered from samples treated by sonication and

mechanical stirring, although stirring required more time.

The method developed by Hurst (85) was used by Farrah et al. with

limited success (48). This method resulted in overall volumes that were

quite large. Additionally this method did not effectively recover virus

adsorbed to aerobically digested sludge. In order to overcome these

difficulties, they developed a urea-lysine method for the recovery of

viruses adsorbed to sludge (48). Briefly, 4 M urea 1% lysine, pH 9

was mixed for five minutes with the sludge, resulting in 70% elution of

poliovirus-1 adsorbed to sludge flocs. Once eluted, the virus was

separated from the sludge flocs by centrifugation and concentrated by

aluminum hydroxide flocculation (44). The virus were eluted from these

flocs by with 0.1 M EDTA 3% beef extract, and further reconcentrated

by organic flocculation. This resulted in a very low final volume with

approximately 40% recovery of poliovirus-1 and coxsackievirus B3 (48).

However, problems, such as inactivation of virus by urea over time, and

length of time necessary to process samples preclude field use of this


Recently, two methods were compared for their ability to recover

viruses from different sludge types under the auspices of the American

Society for Testing Materials (ASTM) in the hopes of establishing a

standard method (66). One method tested was the Environmental

Protection Agency (EPA) method, developed by Donald Berman of EPA (6,

10). Briefly, this method involves mixing sludge on a magnetic stirrer,

followed by the addition of A1C13 to a final concentration of 0.5 mM

and pH adjustment to 3.5. Elution was accomplished by mixing the sludge

for 30 minutes with 10% buffered beef extract. After centrifugation to

separate viruses in beef extract from sludge flocs, the beef extract

solution was filtered through a Filterite series to remove contaminants.

The filtrate was diluted with sterile distilled water at a ratio of 7 ml

of water for every 3 ml of beef extract, and reconcentrated using

organic flocculation. The second method was the Glass method, developed

by Steven Glass of New Mexico State University (62). In this method,

sludge was placed in a blender and dry beef extract (final concentration

ca. 4%) and antifoam were added. This was blended for 3 minutes,

adjusted to pH 9 and stirred for 25 minutes longer. The viruses were

separated from the sludge flocs by centrifugation and concentrated by

organic flocculation. The final concentrate was detoxified by a

dithizone-chloroform mixture. The results of this comparison indicated

that the EPA procedure was slightly more sensitive than the Glass method

for recovering viruses from all sludge types but on (66).

All methods detailed here require extremes in pH, and are therefore

unsuitable for the recovery of bacteriophages from these samples.

Association of Viruses with Soils

The application of sewage treatment solids and liquids onto land

can lead to the adsorption of any surviving viruses to the soil. The

desorption of virus, and its subsequent transport through the soil

matrix can lead to groundwater contamination (142, 187). The

dissociation of viruses from soils is due to changes in the

physicochemical properties within the soil. Because the potential of

waterborne outbreaks of viral diseases is great, it is important to

study the mechanisms of and the factors associated with the adsorption of

viruses to soils.

The adsorption of viruses to soils is a reversible process that can

be described in terms of an equilibrium isotherm. Two such isotherms

have been used to examine the kinetics of virus adsorption to soils, the

Langmuir and Freundlich isotherms (53, 174). When the relationship

between the amount of adsorbed and unadsorbed virus is found to be

linear, a Freundlich isotherm is used. It has been proposed that

conformance to the Freundlich iostherm indicates that the surface

contains adsorption sites heterogeneous with regard to the strength of

bond formation (21). The Langmuir isotherm indicates that there are a

large but finite number of adsorption sites on a solid, thus equilibrium

Langmuir isotherms would indicate saturation-limited adsorption (174).

There is considerable debate whether viral adsorption to soils is

best described by either of these two isotherms. The linear plot of the

Freundlich isotherm has been used to describe a variety of virus

adsorption phenomenon, such as the adsorption of poliovirus-1 to loamy

sand (58), the adsorption of T2 and F2 to silty loam (34), #X174

adsorption to clay and silt loans (21) and poliovirus-2 adsorption to

over 30 soils and minerals (124). However, it has been hypothesized

(176) that most adsorption processes were most likely saturation-

limited, and therefore conformed to the Langmuir isotherm. The author

suggested that in most adsorption studies, insufficient concentrations

of viruses were used, and as such, the saturation region of a Langmuir

plot was never seen, and the linear section of such an isotherm give the

appearance of conformance to the linear Freundlich isotherm. Several

studies have dictated that this indeed may be the case. Moore et al.

(124) studied the adsorption of radioactive poliovirus-2 to 34 minerals

and soils and found that at low virus concentrations, data conformed to

the Freundlich isotherm. However, when saturated conditions were

examined, their data conformed well to the Langmuir equation. The

authors, however, also indicated that in systems where saturation

appears to be occurring, it may likely be due to a coating of soil

adsorption sites with competing material. In addition, the Langmuir

isotherm has been used to describe MS2 adsorption to Indian soils (99),

T4 adsorption to activated carbon (27) and the adsorption of OX174 to a

silt loam (21).

As these data above indicate, adsorption of viruses to soils is a

complicated process. The adsorption, and subsequent desorption of

viruses from various soil types can be affected by many factors. These

factors are listed in Table 13 and are examined below.

When viruses dispersed in liquid are applied to soils, they must

come into close contact.with one another in order for adsorption to

occur. This is directly related to the infiltration or flow rate of the

applied liquid (61, 174). Lance and coworkers (101, 103, 104, 105) have

studied the effect of various infiltration rates of sewage effluent

waters on the adsorption and elution of several enteroviruses. Using a

250 cm coarse sand column, and poliovirus-1 (101) or echovirus-1 (105),

they found that increased flow rates decreased the amount of virus

retained at the top portions of the column. At flow rates greater than

1.2 m/d, both poliovirus (101) and echovirus-1 (105) were leached

through the column. While the percentage of virus found in the column

effluent was low (<1%), the fact that virus will percolate through soil

at higher flow rates indicated the importance of contact time in viral

retention by soils.








r-4 O -4
o -.

* ,-






Q0 Q)
i rl

O ,-

IQ 0

0 r
U 0
On V

c 0
m 0

C 0




"- 44
U *
Ia (

- C



U 4J
o z
0 a
0 wC

> 4
4C -4

a4 4-)

(0) 1) r4 ()
0 *C

0 r> Q

4-0 0 0
0- 0 0

(a 0 w4
4 a4 -Hw

0 un
000 >
' .rl ( 1

0 En
U 0 o 4-
-I In 0
) 00

Qe rna

4-0 4 I) *r

4J 0 4-
0)U (C 04 U



C -


O -4
M 0

4 n




I n

a -4
0 -H



0 mI
-H 0
4-) to
04 0
.4 10

4 4

H 4 0
S0 r
-4 0 U

'H -H
SC 0

4 C 4)

oU -4

O-4In 0
0l C

S-4 1
(3 0 *rl

10 1-

V1 0

w (

S0 4-
A p -


0 4

., 41
C -4

0 4



C 0
0 -H

0 0

r l



-H 0


0 4


> 0
4) o.
-H <

r 0
0 U

* rl
4 .

0) 1-


> -4


aI -
cJ A



3 C


1 --



-c -4 I
4.) C i

W 4-
04 *-l

0 a

0 U) 0

r-l 0
C -r
S0 i
-04 -

'0 0
0 0

C -rl
w Q)

W In


-4 o

-4 0

> 0

0 M Ul

-H C

S*r 0



04 U

More recent work in the same laboratory (108) compared the movement

of poliovirus-1 in a 250 cm long soil column during saturated and

unsaturated flow of sewage water. In was found that during unsaturated

flow, poliovirus-1 was not detected below the 40 cm mark. However,

during maximum flow rates, poliovirus-1 penetrated to 160 cm, and in

some cases, past the 200 cm level. These data indicated the importance

of low infiltration rates to maximize virus retention by soils.

In natural systems, both viruses and most soils are negatively

charged. As such, pH and salt content can have a tremendous impact on

virus adsorption to and subsequent elution from the soil matrix. Taylor

et al. (171) studied the interaction of radiolabelled poliovirus-2 with

various soils, sands and clay. They found that at the pH values above

9, virus adsorption to these substrates was limited. At pH values near

neutrality, adsorption of poliovirus-2 to soils was greatly facilitated

by the addition of 10 mM CaC12 or Na2SO4, while adsorption was limited

in solutions which did not contain these ions. These authors found that

for each substrate, there was a characteristic pH region of transaction

from strong to weak attraction for virus uptake, and .in order for

adsorption to occur, these electrostatic repulsive forces must be

overcome. In a study comparing the adsorption of X174 to five

different soils (21), it was found that the low adsorption capacity of a

loamy sand was related to its high pH.

The importance of cation addition in the promotion of virus

adsorption to soils has been well documented. Retention of f2 and

poliovirus-1 by sand was greatly increased by the presence of 10 mM

CaC12 or MgCl2 (108). Scheuerman et al. (144) found the adsorption of

poliovirus-1 to soils with high organic content was less than 20%. The

addition of 10 mM CaCl2 increased its adsorption to greater than 98%.

Moore et al. (123) found that adsorption of reovirus-3 to various sands

and silts was enhanced by the addition of 1 mM CaC12 or MgC12 to the

suspending medium. Using soil columns, Lance and Gerba (102)

demonstrated that the addition of 0.2 mM AC13 prevented the leaching of

poliovirus-1 in sewage effluents from the soil column. They suggested

that the increase in virus adsorption associated with increasing ionic

strength of the suspending solution was probably due to a decrease in

the thickness of the layers of charged ions around soil particles and

viruses. This allowed for short-range attractive forces, such as van

der Waals interactions, to take over.

The desorption of adsorbed viruses by low ionic strength solutions

such as rainwater has been well documented (36, 104, 106, 185, 187).

Duboise et al. (36) found that poliovirus-1 adsorbed to soil columns in

sewage effluents were easily desorbed by the addition of deionized water

to simulate rainfall. Studies by Wellings et al. (185, 187)

demonstrated the presence of viruses in wells 10 to 20 feet below soil

that had been treated with sewage effluents. These viruses appeared

after a period of heavy rainfall, and the authors hypothesized that the

increase in water/soil ratio resulting from the rainfall led to

desorption of the attached virus. Lance and coworkers (104) found that

the viral desorption and movement caused by simulated rainwater was

greatly reduced or eliminated by the addition of salts, again

implicating the importance of ionic strength in the association of

viruses to soils.

The presence of soluble organic matter in suspending solutions can

interfere with viral adsorption to soils by competing with viruses for

adsorption sites (61). Gerba and Goyal (55) found that virus suspended

in secondarily treated sewage, which has a high content of soluble

organic, adsorbed less to soils than virus suspended in deionized water

adsorbed to the same soil. Work by Bitton et al. (17) indicated that

humic substances interfered with the adsorption of coliphage T2 and

poliovirus-1 to a sandy soil.

The organic content of the soil itself may also play a role in the

association of viruses to soils. In studies examining adsorption of

virus to over 30 different soils and minerals, Moore and coworkers found

that both poliovirus-2 (124) and reovirus-3 (123) were adsorbed poorly

in soils with higher organic content. Burge and Enkiri found that XX174

adsorbed well to four of five soils tested, the exception being a clay

loam high in organic content (21). Soils with high amount of humic and

fulvic acids were shown to be poor adsorbers of poliovirus-1 (144).

In addition to organic components, other constituents of the soils

will influence the association of viruses with them. The adsorption of

XX174 to soils was found to be correlated with the cation exchange

capacity and specific surface area, as well as organic content (21).

The natural pH of a soil has been found to effect virus adsorption (21,

34, 55). Generally, soils with lower pH's tended to adsorb virus more

effectively. The presence of iron oxides, such as magnetite and

hematite, in soil increases the adsorption capacity for viruses (11, 61,

124). Clay content has been shown to be very important component in

soils that promotes virus adsorption (61, 174) and will be discussed


The nature of the virus itself will play a paramount role in

adsorption of viruses to soil surfaces. Burge and Enkiri (20) found

that *X174 purified by density gradient centrifugation was still

heterogenous. They found 2 fractions with different adsorption rates.

Landry et al. (106) compared lab strains and wild type strains of

poliovirus-1 and discovered that field strains were eluted from soils by

rainwater to a greater extent (>33%) than were lab strains (<3% eluted).

Gerba and Goyal (55) examined the adsorption of 27 different

reference strains of enteroviruses to a loamy sand. They found that

most viruses adsorbed very well (>90%) but there were some exceptions

including echoviruses-1 (55%), -12 (78%) and -29 (14%). They further

examined different strains of the same virus and found that adsorption

to loamy sand was very strain dependent. For six different strains of

echovirus-1, adsorption varied from 0 to 99%. This same group

statistically analyzed their data to determine variables associated with

virus-soil interactions (56). They found that different types and

strains of viruses could be grouped by their ability to similarly

effected by certain soil characteristics. Group I, which included

echovirus-l, coxsackievirus B4, X174 and MS2, was found to be

influenced by pH, organic matter and exchangable iron content of the

soil. Group II (poliovirus-l, echovirus-7, coxsackievirus B3, T2 and

T4) did not appear to be influenced significantly by any soil factor.

The authors concluded that viruses in group I were more likely to be

influenced by changes in soil characteristics than group II as far as

adsorption/elution of these viruses was concerned.

Association of Viruses with Clays

Clay is perhaps the most important constituent of soils that

governs the adsorption of viruses (61, 174). Studies have shown that

the higher the clay content of the soil, the higher the viral adsorption

to that soil (34, 169, 174). The adsorption of MS2 to Indian soils was

shown to be depending on clay content (99). Lateritic soil, with a clay

content of 32%, was found to adsorb more MS2 than Black Cotton soil or

Kampur soil, with clay contents of 28% and 10% respectively. Sobsey et

al. (161) found that reovirus-3 and poliovirus-1 adsorption to different

soils could be related to clay content. Gerba and Goyal compared

adsorption of selected enteroviruses to nine different soils, and found

increased clay content in the soil allowed for more virus adsorption

(55). Work by Drewry and Eliassen (34) indicated that there was a

direct relationship between the clay content of nine soils from Arkansas

and California and their retention of bacteriophage.

High virus retention by clays has been related to their ion

exchange capacity (21, 112, 169) and their large surface area available

(21, 36, 61, 143) for adsorption. The research group headed by Stotzky

has examined the adsorption of various viruses to pure clay minerals to

determine the factors involved in this association (110, 111, 112, 145,


The adsorption of coliphages T1 and T7 to kaolin, a two-layer non-

expanding clay, and bentonite, a three layer expanding clay, was

examined (145). The adsorption of T7 to these clays was found to be

related to the cation exchange capacity (CEC) of these clays, indicating

that T7 adsorbed to the negative sites on the clay surface. T1

adsorption, however, was not correlated with CEC. Pretreatment of clays

with 1% sodium metaphosphate, which blocked positive sites on the clay,

reduced T1 adsorption to kaolin by 54% and to bentonite by 28%. The

authors surmized that T1 adsorption appeared to be related more to anion

exchange capacity (AEC) of these clays. Similar experiments with

reovirus-3 (110) indicated that CEC of these clays was important in

reovirus adsorption. The investigators also studied the adsorption of

reovirus-3 to bentonite which was made homoionic to various cations. It

was shown that adsorption was increased in the presence of bentonite made

homoionic to higher valency ions; that is, adsorption of reovirus-3 was

greater to bentonite homoionic to aluminum than bentonite homoionic to

calcium or magnesium. The authors suggested that this increase in

adsorption was due to reduction of the net negative charge on clays,

which allowed virus to come close enough to the clays to become

protonated and subsequently adsorb by cation exchange (110).

The effect of various proteins on virus adsorption to clays has

been evaluated. Carlson et al. (22) found that egg albumin or bovine

albumin decrease the adsorption of poliovirus-1 or T2 to kaolin. Lipson

and Stotsky (111) Showed that chymotrypsin decreased reovirus-3

adsorption to kaolin by 26% and to bentonite by 66%. This reduction of

adsorption was also seen with the use of ovalbumin. It was suggested

that the blockage of negatively charged sites on the clay by these

proteins was the reason for this decrease. In subsequent study (112),

the adsorption of T1 to bentonite and kaolin was blocked completely by

pretreatment of the clay with minimal essential media (MEM) supplemented

with 10% fetal calf serum. In this case, the authors hypothesized that

the organic in the MEM blocked positively charged sites on the clay

that T1 appears to adsorb to. These authors also examined the

adsorption of a lipid-enveloped virus, herpesvirus hominis type 1