EPA-600/2-76-287
December 1976
Environmental Protection Technology Series
VIRUS PARTICLE AGGREGATION AND
HALOGEN DISINFECTION OF WATER SUPPLIES
Municipal Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socio-economic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental degradation from point and non-point sources of pollution. This
work provides the new or improved technology required for the control and
treatment of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-76-287
December 1976
VIRUS PARTICLE AGGREGATION AND
HALOGEN DISINFECTION OF WATER SUPPLIES
D. Gordon Sharp
University of North Carolina
Chapel Hill, North Carolina 2751^
Grant No. R803771
Project Officer
John C. Hoff
Water Supply Research Division
Municipal Environmental Research Laboratory
Cincinnati, Ohio 45268
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO ^5268
-------
DISCLAIMER
This report has been reviewed by the Municipal Environmental
Research Laboratory, U. S. Environmental Protection Agency, and approved
for publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U. S. Environmental Protection
Agency, nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
FOREWORD
The Environmental Protection Agency was created because of increasing
public and government concern about the dangers of pollution to the health
and welfare of the American people. Noxious air, foul water, and spoiled
land are tragic testimony to the deterioration of our natural environment.
The complexity of that environment and the interplay beween its components
require a concentrated and integrated attack on the problem.
Research and development is that necessary first step in problem
solution and it involves defining the problem, measuring its impact, and
searching for solutions. The Municipal Environmental Research Laboratory
develops new and improved technology and systems for the prevention, treat-
ment, and management of wastewater and solid and hazardous waste pollutant
discharges from municipal and community sources, for the preservation and
treatment of public drinking water supplies, and to minimize the adverse
economic, social, health, and aesthetic effects of pollution. This pub-
lication is one of the products of that research; a most vital communications
link between the researcher and the user community.
Using methods developed under a previous EPA Grant (R 8029^-6) additional
information on virus inactivation by halogens and the effects of virus
aggregation on disinfection is provided. This information contributes to
our understanding of disinfection and its use in water treatment processes.
Francis T. Mayo
Director
Municipal Environmental Research
Laboratory
111
-------
ABSTRACT
Wgter suspensions have been produced that contain essentially all single
virus particles. Both bromine and chlorine react very rapidly with these
preparations.] Reovirus is more sensitive to bromine than poliovirus by a
factor of tfiSfe than 10. Bromine destroys poliovirus a little faster than
chlorine but both are very fast acting on single virions and special kinetic
equipment had to be built to determine the kinetics. Tribromamine is about
as active as HOBr on single polio virions in equal molar concentrations.
But both reo and polio viruses in aggregated form, as seen directly by the
electron microscope, are much more resistant to bromine action than single
particlesj Aggregated seems to be the normal state of virus in water unless
it is put there in dispersed form. Dilution of laboratory stock virus in
water tends to aggregate it. Special methods were devised to detect this at
high dilutions. There are natural, unbroken aggregates too that are diffi-
cult to disperse. I Practical sterilization of water supplies will doubtless
have to deal with tfiese resistant aggregates, both natural and induced,
rather than with single virus particles.
IV
-------
CONTENTS
Page
Foreword
Abstract iv
List of Figures and Tables
Acknowledgments ^
Introduction -^
Conclusions 3
Recommendations 5
Materials and Methods ->
Experiments and Results ^2
Reovirus ^2
Poliovirus ^9
Physical State of Viruses in Water 24
Reovirus 24
Poliovirus 28
Electron Microscope Observations 29
Dilutions of Poliovirus in Water 31
Natural Aggregates of Poliovirus 33
Discussion 39
Inactivation of Reovirus 39
-------
CONTENTS (Continued)
Page
Survival of Reovirus Through Clumping 40
Physical Stability of Virus in Water 44
References 47
VI
-------
FIGURES AND TABLES
Fig.
1. Apparatus for halogen treatment of virus in flowing water.
2. Inactivation of reovirus in water by bromine (HOBr) at 2C.
3. Frequency distribution of reovirus aggregates.
4. Bromine action on a known mixture of reovirus aggregates.
5. Reovirus aggregates in starting preparation of Fig. 6.
6. Sedimentation velocity spectrum of the reovirus of Fig. 5.
7. Partition of reovirus between singles and aggregates.
8. Poliovirus purified by freon extraction and velocity banding.
9. Count data from pictures of the fractions of Fig. 8.
10. Inactivation of poliovirus by bromine (HOBr) at 2C.
11. Inactivation rate of poliovirus vs. bromine concentration.
12. Inactivation of poliovirus by bromine (HOBr) at IOC.
13. Inactivation of poliovirus by bromine (HOBr) at 20C.
14. Inactivation of poliovirus by chlorine (HOC1) Agg. and Disp.
15. Inactivation rate of poliovirus vs. HOC1 concentration.
16. Inactivation of poliovirus by Tribromamine.
17. Stock reovirus diluted 10 x then 20 x more with PBS.
18. Group size distribution in dilute stocks of Figs. 17 and 19.
19. Stock reovirus diluted 10 x with water then 20 x with PBS.
20. Typical purified stock suspensions of poliovirus.
vii
-------
FIGURES AND TABLES (Continued)
21. Dilution of poliovirus (Fig. 20) 10 x with water.
22. Further dilution of poliovirus (Fig. 21) with PBS.
23. Poliovirus aggregates at pH 5 in 0.05 M acetate buffer.
24. Reovirus emerging from infected cells in large aggregates.
25. Aggregates of poliovirions found in dialyzed crude extracts.
26. Inactivation of crude poliovirus (Fig. 25) by bromine.
27. Sedimentation velocity spectra of poliovirus PFU.
28. Simple virion aggregation test (Red, White, and Blue).
29. Poisson approximation to the reovirus curve of Fig. 4.
30. Straight line approximation to the reovirus curve of Fig. 4.
Table
1. Purified poliovirus is randomly spread and mostly singles.
2. The physical state of poliovirus in water.
viii
-------
ACKNOWLEDGMENTS
This work was supported in part by Grant #R803771 from the U.S. Environmental
Protection Agency and in part by the U.S. Army Medical Research and Develop-
ment Command under contract DAMD17-17-A013. It has also been materially
aided by Mr. Richard Boynton of the Sperry Rand Company of Durham, North
Carolina, who provided the special motor windings for the magnetic mixing
devices used in the Flash Flood Experiment.
The oil-driven OTD2 ultracentrifuge and the ARC-1 slow start device were
gifts of the Ivan Sorvall company.
Most of the experiments were done by Roger Floyd and D. C. Young with the
excellent technical assistance of Florence Stubbs, electron microscopist,
who made the virion counts and aggregation analyses.
All of the devices for preparation of virus for kinetic attachment electron
microscopic count and for the Flash Flood experiment were made by the author
in his home machine shop.
IX
-------
INTRODUCTION
There have been many papers written about the survival of virus during expo-
sure to disinfecting agents and most of them include, in the discussion of
their results, some speculation regarding the possible influence of virus
particle aggregation. This account differs from the others. It provides
direct evidence of the kind and degree of aggregation that was present among
the virions that were treated with bromine as well as the plaque forming unit
(PFU) titer of the starters and the survivors. This physical evidence comes
from electron microscopy (EM) of virus samples prepared by methods especially
contrived to avoid the aggregation artifacts which are inherent in ordinary
preparative techniques that have been designed primarily to show ultrastruc-
ture. Two methods for doing this have been devised here. Both of them are
based upon the premise that the virus particles and clumps must become at-
tached to the surface for observation while the surface is in contact with
the whole body of the water suspension; they must not be deposited there
during the time the water suspension of virus particles is drying because the
attendant forces of surface tension exert strong aggregating effects on sus-
pended particles. The first of these techniques, which works well with the
larger viruses, including reovirus, involves sedimentation of the particles
by centrifugation onto an agar receiving surface from which they can be
transferred by pseudoreplica to the EM. The other involves kinetic attach-
ment of the particles by Brownian bombardment of an aluminum coated collodion
film which, after thorough washing, can be examined directly in the electron
microscope. This method gives count and aggregation data on small viruses,
such as polio, which do not yield to the sedimentation-pseudoreplica method.
Any series of experiments which inquire into the behavior of virus, aggre-
gated or not, in the presence of disinfectants requires the preparation of
a concentrated stock from which starting virus is taken for each experiment.
Such a stock is usually kept frozen to preserve a high infectivity titer but
in the present work it was necessary to avoid aggregation as well and we
were unable to meet both requirements by freezing. Collection and storage
of virus stock suspensions in concentrated sucrose solution has served both
ends quite well. The preparation and handling of these virus stocks is a
significant part of this project.
Observations on the physical state of aggregation of a virus in water require
particle concentrations great enough so that EM pictures will show enough
particles and groups of various sizes so that their numbers can be establish-
ed with the necessary statistical significance. Also, they must be suffic-
iently purified so that there is no doubt about recognition of the particles,
as virus, and no doubt about distinguishing the separate individuals in the
-------
Q
groups. This means about 10 per ml for reovirus (by sedimentation count)
and about 10 per ml for polio by the kinetic attachment (KA) method. '
Such concentrations may,seem very high, but actually the corresponding PFU
titers are about 2 x 10 per ml for reo and 10 per ml for polio, both of
which are well within the range normally experienced in unconcentrated
lysates of infected cell cultures.
While EM provides quantitative data with a satisfying directness in experi-
ments with virion aggregation, it can be applied only within the limited
concentration range just described. In the stock concentrates as well as
in the very dilute conditions likely to prevail in virus-polluted natural
resource water, the physical state of the virus cannot be observed so direct-
ly. However, the relative rates of sedimentation of clumps and single virus
particles are essentially the same no matter how many or few are present.
Appropriate techniques have, therefore, been designed for zonal and swinging
bucket rotors to detect and, to some extent, to measure virion aggregation
under these extreme conditions. Centrifugation is better than filtration
for this purpose because the results are more readily interpreted quantita-
tively.
Survival of both reo and polio viruses is shown in this work with bromine to
be much longer for aggregates than for single virus particles. So it becomes
important to know in experiments, as well as in the field, what state the
virus is in. This work reveals some difficulties encountered when stock
viruses were diluted in water for disinfection experiments. Sometimes they
aggregated. Some of these aggregates were dispersed when diluted further
with suitable solvent solutions and others were not. An effort has been
made to define the conditions that lead to aggregation for the benefit of
those who wish avoid them and the confusion caused by unexpected plaque
titration results.
The virus particle aggregates that occur when monodisperse stock suspensions
are diluted in water are not to be confused with natural aggregates that are
present in virus preparations that have not been thoroughly dispersed during
extraction from the infected cells. Evidence is provided that such natural
aggregates may not disperse when suspended at high dilution in either labora-
tory distilled water or natural fresh water with or without normal particu-
late content.
Most of the experiments here dealing with virus inactivation have been done
with bromine under conditions where the principal active agent was HOBr. A
few were made with tribromamine and a few others with HOC1.
-------
CONCLUSIONS
The general state of virus particles in water suspension is aggregated.
Many viruses are produced as aggregates in infected cells and when we have
examined suspensions of reovirus by zonal centrifugation or by quantitative
EM, we have found a substantial fraction (often over 50%) sedimenting much
faster than single particles, and in this fraction many aggregates can be
seen by EM.
When aggregates are present in a reovirus suspension containing bromine,
the disinfection rate of the mixture proceeds at a constantly decreasing
rate which appears to be determined by the frequency of the different
aggregate sizes in the mixture.
When aggregates are removed from the above mixture by taking advantage of
their differences in sedimentation velocity, the rate of disinfection of
the remainder by bromine becomes essentially a straight line. This is the
basic reaction rate for single virions, and it is the only reliable basis
for comparing the resistance of one virus with another with different
disinfectants under different conditions.
When working with bromine and suspensions of single virions it is necessary
to make very short time exposures. Even with poliovirus and 3 yM bromine,
they must be as short as 5 sec and with reovirus, 1/2 sec or less. The
continuous flow or "Flash Flood" apparatus that we have built has met this
need and it promises to be of broad general value in such work.
Poliovirus suspensions of single particles are about 25 X more resistant
to bromine than similar suspensions of reovirus, and there are some anomalies
too. In a limited range of temperature and bromine concentration there is a
slight delay in the onset of the disinfection process. Such delays have been
seen by others but always case of what was presumed to be an aggre-
gated virus. This delay must now be attributed to something else, possibly
penetration of the virus capsid by the bromine. Poliovirus disinfection
rates are proportional to bromine concentration at IOC and at 20C but at 2C
they are not. Reasons for this are not apparent. This leads to a linear
Arrhenius plot only at about 10 yM bromine concentration.
The disinfection of suspensions of single poliovirions by HOC1 is generally
first order with rate constants at 2 and IOC less than those for HOBr but
under no conditions less than half. Experimental difficulties encountered
at pH 6, where HOC1 predominates, are due to colloidal instability. The
virions tend to aggregate.
-------
The rate of disinfection of poliovirus at IOC is roughly proportional to
concentration of HOC1, but at 2C it behaves like HOBr in that the disin-
fection rate does not increase linearly with concentration.
We find in our preparation of both polio and reo viruses, two kinds of
aggregates, natural and induced. The former are presumably unbroken parts
of the cytoplasmic concentrations in which they were generated, and the
latter occur in previously dispersed suspensions as a result of one or more
of several destabilizing conditions.
Induced aggregates of reovirus are quite stable.
Induced aggregates of poliovirus break up quite easily.
Natural aggregates of both viruses are generally more stable than induced
aggregates.
Dilution of either virus from monodispersed stock into distilled or natural
resource water can induce aggregation; it depends upon how it is done. Often
a single large dilution step will avoid aggregation.
Both reo and polio viruses aggregate in pH 5 buffer and usually in pH 6
buffer, especially at low ionic strength.
Natural aggregates of poliovirus are not dispersed when diluted several
thousand fold in distilled or natural resource water.
Stock suspensions of single poliovirus particles do not become aggregated
when diluted at one step (5000 fold) with distilled or natural lake water,
but they do aggregate when water is used after treatment in the filter
plant with lime and alum.
Aggregated reovirus has survived for 300 seconds the same bromine treat-
ment that destroyed single virions in 1 second. Aggregated poliovirus shows
the same relative resistance of clumps over single particles. These facts
show that the kinetics of halogen disinfection of virus in water is very much
dependent on the physical state of virus in the water that is, in turn, a
highly variable thing.
-------
RECOMMENDATIONS
1. Continue experiments on chlorine disinfection of single polio and
reo viruses, at least until a good comparison can be made with the
bromine data now on hand.
2. Examine the survival of polio and reovirus aggregates, both natural
and induced, when treated with chlorine and bromine.
3. Examine the effect of small-number aggregation on the halogen sur-
vival of poliovirus for comparison with similar data now in hand
on reovirus. See if there is complementation in plaque formation.
See if the data fit-one of the schemes presented in this report or
discussed by Clark or by Wei and Chang
4. Continue the effort to titrate poliovirus in such a way that induced
aggregates can actually be presented to the cell monolayer.
5. Examine, in detail, the conditions that determine the physical state
of the virus particles in water suspension with particular attention
to natural resource waters, their pH, ion content and concentration
etc. Look at divalent and travalent ions in particular.
6. Look for better ways to prepare and keep stocks of virus in both a
dispersed and aggregated condition.
7. See if particles suspended in the water have any effect on virion
aggregation, survival in the field or resistance to halogen action.
Check bentonite and other substances that are frequently found in
water to see how viruses interact with them.
8. Develop further the coordination between zonal centrifugation and
EM, because they contribute independent data on virion aggregation,
natural, induced and complexed with other particles. The centrifuge
is blind but very sensitive to aggregation. The EM sees aggregates
clearly but only in a very limited concentration range. The two
together are a powerful analytical combination.
9. Examine carefully the possibilities for virion aggregation that
occurrs when virus stock suspensions are employed in the dynamic
flow apparatus for disinfection studies. The changes in concen-
tration of virus, ions, sucrose, and pH that can occur can pro-
bably be influenced to some extent by optimum management, including
quick mixing, choice of buffer etc.
-------
It is of vital importance that the physical state of several of the
enteric viruses be examined. The data gained so far about just two.
show major differences, and they differ from those gathered by others
using bacteriophages. We need to see if polio, echo and coxsackie
for example, are alike or different before any overall concepts can
be formed about the behavior of enteric viruses in water.
10. Methods must be found for observing the natural aggregates of virus
particles that are released by infected cells directly into water.
It is likely to be these, rather than those induced from purified
preparations, that will be the persistent survivors of disinfecting
processes.
11. Virus isolated by change from low to high pH and back may be held or
not by the filters because of electric charge alone but we believe
induced and dispersed aggregates are involved as well. Inasmuch as
this process is now the method of choice for concentrating virus
from large volumes of water, it may be well to investigate this point
before an incomplete concept of the mechanism congeals into textbook
dogma.
12. As soon as we are able to produce by induced aggregation or select
from a mixture of natural aggregates, a preparation of all pairs
or containing only a narrow range of aggregate sizes, we want to
get on with the measurement of plaquing efficiency vs. group size,
particularly for survivors of halogen treatment.
13. Exploit as far as possible the improved resolving power now available
in swinging bucket centrifuges because of the newly-developed slow-
start and slow-stop techniques with density gradients.
14. Examine other sucrose gradient configurations and substances other
than sucrose to improve the resolving power of aggregate spectra
from the zonal centrifuge and also short-cut aggregation tests,
such as the Half and Half as well as the Red, White, and Blue
described later.
15. Employ each bit of new information on adhesion of virions to each
other and to other things to best advantage in improving the
technique of quantitative EM.
It is not expected that all of these areas can be pursued in depth,
but certainly several of them should be.
-------
MATERIALS AND METHODS
PREPARATION OF STOCK VIRUS FOR DISINFECTION EXPERIMENTS WITH DISPERSED AND
AGGREGATED VIRIONS.
Reovirus type 3 (Dearing strain) was grown in and all plaque titrations were
made on monolayer cultures of L cells, as previously described , except that
cells were maintained in 200 ml milk dilution bottles and passaged into 32 oz
prescription bottles for growth of virus. Virus was harvested at 16 to 18 h
after infection at 10-20 PFU/cell rather than 20 to 24 h as previously noted.
In addition, the virus was extracted from the cell in 6 ml of phosphate buf-
fered saline (PBS) without calcium and magnesium and 4 ml of Freon 113 for
2 min at one half speed in a Sorvall Omni-Mixer. The virus was placed on
20 •> 40% w/w sucrose gradients in 0.05 M phosphate buffer at pH 7.2, and
centrifuged at 25,000 RPM in the Beckman SW 27 rotor at 4C for 1 h. The virus,
which was collected from the lower of two bands seen wl^th a collimated beam
of light, was allowed to remain in the sucrose at 4-6C and was not pelletized.
The virus stored in this manner retains its infectivity for several weeks,
and the state of physical aggregation is quite stable. There is no bacterial
or fungal growth, and the sucrose does not produce any detectable bromine
demand under the conditions of our experiments.
Poliovirus type I (Mahoney strain)was obtained from Dr. Gerald Berg, Envi-
ronmental Protection Agency, Cincinnati, Ohio, and was serially passaged in
Human Epidennoid Carcinoma cells, HEp-2. The cells were grown in medium 199
containing 0.105% NaHCO and 5% fetal calf serum (FCS). After the cells
reached confluency, they were maintained on the same medium with 2% FCS.
Stocks of virus were produced in HEp-2 cells at 37C by infection at a multi-
plicity of 10-20 PFU/cell under a maintainence medium of 199 + 0.105% NaHCO^
and 2% FCS. After 18-24 h when the cytopathic effect was 100%, the cells
were frozen and thawed 3 times and the cell debris removed by low-speed
centrifugation (^800G) for 10 min. The supernatant fluids had titers of 4 x
10 PFU/ml and were kept frozen at -70C.
Purified virus was produced in the same cells grown for 3 days in 32 oz
prescription bottles inoculated with the above stock virus in Dulbecco's PBS
containing 12.5 mM MgCl , at a multiplicity of VLOOO PFU/cell. The virus
was allowed to adsorb for 1 h at 37C, then 40 ml of maintainence medium con-
taining 12.5 mM MgCl2 were added to each bottle and the cells were further
incubated at 37C for 11 h. The cells were chilled to 4C, and the supernatant
fluid was decanted. The monolayers were washed 2 times with PBS and the cells
of each bottle were scraped into 10 ml of PBS and pelletized at 250 G for 10
min. Cells remaining in the supernatant fluid were pelletized similarly and
pooled with the scraped cells and pelletized again. The pellet of cells was
-------
resuspended in 6 ml of PBS not containing Mg or Ca . Four ml of Freon 113
were added, and the cells were homogenized at one-half speed in a Sorvall
Omni-Mixer for 2 min. The phases were separated at 800G for 10 min, and the
upper aqueous phase was removed and held in an ice bath.
The Freon phase was re-extracted with a further 6 ml of PBS, and the phases
were again separated. The aqueous phase was pooled with the previous one,
and the Freon phase was again re-extracted with 6 ml of PBS. All aqueous
phases were combined and made to a volume of 20 ml. Ten ml of this extract
were placed on each of two 10 -* 30% w/w sucrose gradients made in 0.05 M
phosphate buffer, pH 7.2. The gradients were centrifuged at 25,000 RPM in
the Beckman SW27 rotor at 4C for 135 min. Fractions of 2 ml each were col-
lected from the tubes and examined by the kinetic attachment for the,.
presence of the virus. Twenty fractions were obtained from each tube , and
the highest count of virus was most frequently found in fraction 14. All
relevant fractions were pooled, and the virus was stored at refrigerator
temperature without any attempt being made to remove the sucrose.
PLAQUE TITRATION
Plaque titrations of reovirus were performed on 3-day-old monolayer cultures
of L cells ' in tightly stoppered 1 oz prescription bottles under an overlay
of 1% Difco agar containing Medium 199, 5% FCS, 0.245% NaHCO , and 0.003%
neutral red. Plaques were counted after 6 days at 37C.
Plaque assays of poliovirus were performed on 4-day-old monolayers of HEP-2
cells in 1 oz prescription bottles under an overlay of 1% Difco agar contain-
ing medium 199 plus 5% FCS, 0.210% NaHCO , 0.003% neutral red, and 5 mM
MgCl_. The plaques were read after 3 days incubation at 37C.
PHYSICAL ASSAY (Particle counting and aggregation analysis)
For reovirus the physical assay by electron microscopy on most preparations
was done by the sedimentation and agar pseudoreplica method. This will be
called the spin down or SD method. A few counts on reovirus were made by
the kinetic attachment or KA method. Poliovirus counts and aggregation
analyses were done by the KA method.
BROMINE, GLASSWARE, AND WATER
Bromine measurement and preparation of glass apparatus and demand-free water
were the same as previously reported except that amperometric titrations
were employed together with the method of Taylor , and all phosphate buf-
fers were made with Fisher primary standard potassium salts.
-------
ZONAL CENTRIFUGATION
Resolution of both crude and purified virus particle populations into sedi-
mentation velocity spectra was done with a BXIV titanium zonal rotor (pro-
vided by Dr. Norman Anderson of the Oak Ridge National Laboratories)
operating in a Sorvall OTD2, fluid-drive ultracentrifuge.
SEDIMENTATION VELOCITY PARTITION OF MIXTURES OF SINGLE VIRIONS AND AGGREGATES
The different sedimentation velocities of single virions and aggregates of
various sizes have been utilized to provide fractions for plaque titration.
Two types of experimental procedures have been employed for this, using the
small plastic tubes that fit the Beckman SW50 or SW50.1 centrifuge rotors.
The simpler involves partially sedimenting a homogeneous mixture of the virus
sample. Aggregated preparations show a lower than normal fraction of infec-
tivity remaining above the center line after spinning. The second variation,
called the Red, White, and Blue, has been very useful in detection of virion
aggregation. It involves placing a 2 ml layer of virus sample over a 2 1/2
ml layer of 20% sucrose solution containing trypan blue, over 1 ml of 50%
sucrose solution containing neutral red and spinning enough to sediment about
half of the PFUs from a well-dispersed preparation out of the top (white)
into the center (blue) fraction. Test samples with high aggregation will
leave most PFUs in the blue and red fractions.
Both of these techniques will be described in more detail, together with the
particular experiments in which they were used.
APPARATUS AND PROCEDURE FOR EXPOSING VIRUS TO DISINFECTION BY BROMINE
Long-time experiments (exposure times of one or more minutes) were made by
putting virus into beakers containing water with suitable buffer and bromine
concentration and withdrawing samples for quenching and titrating at appro-
priate times- These experiments have been described in detail in our earlier
publication.
Short-time experiments involving exposure of virus to bromine for 1/2 to 20
seconds could not be made in the beakers. For them, a kinetic experiment was
devised that would permit injection of virus into a fast-flowing stream of
turbulent bromine water and sampling with instant quenching of residual
bromine at several points down stream.
The apparatus must provide enough volume of water to ensure stability of
temperature, bromine concentration, and flow rate through the reaction
siphon tube (Fig. 1) for several minutes. The reaction tube is 3/8 inch
inside diameter and the bottle (20 1) height is adjustable so that turbulent
flow at the desired rate can be achieved. Because turbulence is essential,
the Reynolds Number has been kept equal to or greater than 3000 in all
experiments. The reaction tube has 5 access ports covered by disposable
serum bottle stoppers. Five 5 ml syringes with needles are inserted. The
-------
3CALE lOen
CONTINUOUS FLOW APPARATUS
for
VIRUS DISINFECTION EXPERIMENT
REYNOLDS NUMBER
IN FLOW TUBE
s 3000
DOWN STREAM SAMPLE POINTS
DISCHARGE
Fig. 1 Apparatus for injection into and recovery of virus from a turbulent stream of
water containing a disinfecting agent, such as bromine, in the Flash Flood Experiment.
-------
first of these (at the left) carries the virus inoculum (5 ml) which, when
triggered, starts the weight-driven plunger which delivers the inoculum
steadily for 5 sec, into the turbulent stream of buffered bromine water.
There is a 6 mm diameter disk located on the axis of the tube, just down-
stream of the inoculum needle tip, to aid in mixing.
The time of transit from injection point to sampling point will be determined
only by the flow rate and the distances D.,, D_, etc. It is not in any way
dependent on when the sample is taken as long as the sample is taken within
the 5 sec interval during which the "polluted" water passes the sampling
port in question.
Each of the 4 sampling syringes contains 1 ml of 2 mM sodium thiosulphate
solution, 1 ml of air, and a small Teflon-coated magnetic stirring bar. The
syringe plungers are held in metal blocks attached to helical springs, each
under tension and restrained by a release pin that can be pushed at the
desired time. When the spring is released, the syringe draws exactly 1 ml
from the flowing stream. This is mixed with the thiosulphate solution to
stop the reaction of the bromine on the virus. For virus-bromine exposure
times of 1 sec or less, additional precautions were taken to ensure instant
mixing. This is done by means of the magnetic stirring bar, which is made
to rotate at 1800 RPM by 2-pole AC motor windings surrounding each sampling
syringe. These are stator windings used in small electric servo motors
kindly provided for us by the Sperry Rand Company of Durham, N. C.
Operation
Twenty liters of buffered chlorine-demand-free water are prepared and adjusted
to the temperature required for the experiment. Enough bromine is added 17 h
before the experiment to bring the concentration in the bottle to approxir-
mately the right value. Just before the experiment, more bromine is added to
adjust the concentration to exactly the required level. Previous experiment
has established the exact level of liquid in the bottle to produce the
desired flow rate (20 to 40 ml/sec in different experiments). This level
is marked for easy reference and all experiments start there. Beginning
with the liquid at a higher level, the siphon is started and the flow con-
tinued for a few minutes to allow the tube to adjust to the temperature in
the bottle and when the liquid level reaches the marked starting point, the
virus injection is started. The stirring bars in all four sampling syringes
are already running in their 1 ml volumes of thiosulphate solution, and it
only remains to trigger each of the syringes as the flood passes and also to
catch a sample of the virus-bromine mixture at the discharge end to deter-
mine its bromine concentration for comparison with that in the bottle. Neu-
tralized bromine-virus mixtures from the sampling syringes and the virus
from the inoculation syringe are then assayed for virus by the plaque method.
The dilution of virus as it enters the flowing stream is just the ratio of
the rates of injection and stream flow. The injection syringe is timed by
direct observation and stopwatch. The stream flow rate is easily obtained
by timing the outflow of some conveniently measured volume, such as 500 ml.
11
-------
Variation in flow rate with liquid level in the bottle is negligible during
the experiment time, which has not been over 20 sec, because of the large
(approx. 700 cm ) surface area.
Tygon tubing is convenient for the flowing stream, but we have found that it
absorbs substantial quantities of bromine. Glass is difficult to manage
without catastrophic accidents. However, "High Density" polyethylene tubing
has performed well. It is readily cut to desired lengths, and it exerts
negligible bromine demand.
EXPERIMENTS AND RESULTS
REOVIRUS
Several experiments were made at IOC with samples taken at 4-sec intervals.
These were made with-virus prepared in the same way as described in our
earlier publication , which involved pelletizing the purified virus to
remove sucrose. These were done at 3, 5.8 and 5.9 yM bromine concentrations
(Fig. 2), and they all show approximately the same level of persistent resis-
tance previously reported in exposures of longer duration. They show also
that the initial fast reaction is taking place during the first 4-sec inter-
val. At this point we added the magnetic mixing devices to the sampling
syringes and reduced the length of the flow tube to give samples at one- and
later at 1/2-sec intervals. The temperature was reduced also, from 10 to 2C,
to slow the Reaction rate and still permit the use of bromine concentrations
high enough to hold constant when confronted with expected bromine demand
of the virus.
Three experiments were made at 2.8-3.0 yM bromine concentration with unpel-
letized virus'. Two were done with 1-sec time intervals and one with the
intervals reduced to 1/2 sec. The frequency of aggregates observed by EM
§ -31
Q.
o
-5
Fig. 2 Survival of reovirus
infectivity in water containing
3 yM (triangles), 5.8 yM (cir-
cles), and 5.9 yM (squares)
bromine at 2C and pH 7.0. The
rapid phase of the reaction is
done in less than 4 sec, and the
slower phase has been shown to
be caused by a few larger
aggregates.
4 8 12
EXPOSURE TIME (seconds)
16
12
-------
in this virus preparation is shown in the graph, Fig. 3. Of the whole par-
ticle population, 73% are singles, and the groups appear to be distributed
in such a way that the log frequency of each group size is^ linear function
of the log of the number in that group. We and others have observed
this same distribution with other viruses and also with polystyrene and acryl
particles of comparable size. The inactivation of the virus by bromine in
these three experiments is shown in the semilog plot of survival ratio vs
time in Fig. 4. It is curved all the way. A single line has been drawn by
inspection through all three sets of points. Correlation of this result
with the aggregation distribution (Fig. 3) will be made later.
Zonal Centrifuge Experiments, Group 1
An effort was made to obtain a preparation of virus containing only single
particles. This is needed to provide a frame of reference for comparing
.2 .4 .6 .8 1.0 1.2
LOG 10 GROUP SIZE
-4
I Z 3
BROMINE TREATMENT (seconds)
Fig 3 (left) Frequency distirbution of aggregates observed by EM (circles).
The'dotted line indicates the same distribution for a more concentrated sus-
pension containing a total of 10,000 particles.
Fig 4 (right) Inactivation of reovirus during the first 4 sec of contact
with 2.9 yM bromine in water at PH 7 and 2C. Three separate experiments are
shown with different symbols. The frequency of aggregates of different sizes
is that shown in Fig. 3. Filled circles show the nearly linear response of a
preparation containing nearly all single virions. The broken line is a pre-
diction for the survival of the same kind of particles as those of the
straight line but aggregated as shown in Fig. 3 and subjected to degradation
according to the multiple target concept as calculated by the Poisson limit.
13
-------
the above results with aggregated virus. Fifteen ml of a purified virus pre-
paration, aggregated as shown in Fig. 5 (circles), were layered over a 15-30%
w/w sucrose gradient with the BXIV zonal centrifuge rotor running at 2000 KPM.
The total number of virions put into the rotor was 10 _ by EM count. After
operating at 25,000 RPM for 30 min (Z o> t = 1.44 x 10 ) at 20C, the 640-ml
volume of the rotor was completely displaced by 35% sucrose piston fluid and
25-ml fractions were collected in 1-oz bottles. The virus particle count on
each was made by EM after dilution with 0.85% filtered sodium chloride solu-
tion. A dilution of at least 1/10 is required, otherwise the agar block
that receives the sedimented virions in the pseudoreplica counting process
will float to the surface because of residual sucrose. Particle counts are
plotted against fraction numbers and a scale of radial distances is included
on Fig. 6. Densities in the sucrose gradient were measured by direct pykno-
meter weighings.
Fractions 4 through 10 contained 69% of the recovered virus. In the peak
fraction (#9) there were 93% single particles. Six percent consisted of
pairs and the remaining one percent comprised triplets and groups of four
(see bottom line of the graph Fig. 5). Thirty-one percent of the total
recovered virus made a broad secondary peak with maximum count in bottle
#13; 63% of the particles contained in it were., in pairs. The total count
of all the fractions indicated 110% of the 10 particles that were put in.
Apparently there was no detectable loss in the partition process, and cal-
culations from the EM frequency distribution chart (upper line of Fig. 4)
predict that 71% of the in-going population was single particles, a figure
in excellent agreement with that observed in this actual sedimentation
velocity spectrum from the zonal centrifuge.
Fig. 5 Frequency distribution of
aggregates observed in starting virus
preparation for zonal centrifuge ex-
periment #1 (circles). The squares
show how the distribution was altered
by removal of most of the aggregates
from the region of the principal peak
(Fig. 6).
0.2 0.4 0.6 0.8
LOGIOGROUP SIZE
.0
14
-------
I! 13 15
FRACTION NUMBER
40
45 5.0
RADIAL DISTANCE (cm)
6.0
Fig. 6. Sedimentation velocity spectrum obtained in the zonal centrifuge of
spontaneously aggregated reovirus (solid line). Ordinates are particle counts
made by electron microscopy. Sixty-nine percent of the virus is in the major
peak, which has 93% single particles and a few remaining small aggregates as
shown in Fig. 5. The secondary peak contained mostly pairs and triplets in
Fraction 13. The dotted line shows a similar spectrum of crude (unpurified)
reovirus showing that the original state of freeze-thaw-release virus contains
very few small aggregates. Large aggregates have sedimented to the rotor rim.
Zonal Centrifuge Experiments, Group 2
Inasmuch as zonal selection provides a method of obtaining very high percen-
tage single reovirus particle suspensions we prepared a fresh quantity of
purified virus containing about 50 x more virus than that of the above zonal
fractionation. This preparation yielded the velocity spectrum with a singles
peak like that of Fig. 6, but the secondary peak constituting 31% of the
first population examined now contained aggregates in the same size range but
relatively few (less than 10%) of the total. Apparently freshly prepared
freon-extracted virus has very few aggregated particles. In the peak frac-
tion (#10) there yere 87% singles. It is possible that at this high concen-
tration (2.5 x 10 virions/ml) some spontaneous aggregation took place
before preparations could be made for EM. A part of this fraction was frozen
immediately to preserve both infectivity and physical dispersion for subse-
quent bromine inactivation experiments.
Zonal Centrifuge Experiments, Group 3
One more zonal velocity spectrum experiment was made with crude freeze-thaw
lysate of infected L cells without even low-speed clarification. This crude
preparation contained the virus from the same number of infected cells as
15
-------
that providing virus for the experiment just described. Again, a single
particle peak was observed as in Fig. 6, and again there was very little
virus in the region where aggregates of 2 to 8 were seen in substantial
numbers in the first experiment. This unexpected result drew attention to
the fact that the total quantity of virus recovered from the crude starting
material was only about 1/5 of that usually recovered by freon extraction of
an equal number of infected cells. It would appear that an excellent fraction
of single particles can be obtained in this way with no previous purification,
and that there are very few small aggregates in such a preparation but that
the major part of the virus must have been in large aggregates or combined
with larger cell debris which sedimented beyond the sampling range and
reached the rim of the zonal rotor.
Bromine Inactivation of "Singles" fraction from velocity spectrum:
The singles peak fraction from zonal velocity spectrum (Group 2 above) was
frozen at -40C in 1-ml vials to preserve infectivity and state of dispersion.
Subsequently thawed samples had not lost infectivity and the relative number
of single particles and groups in the size range of 2 to 8 had not changed,
but bromine inactivation showed that changes had, nevertheless, taken place.
After a rapid start, the reaction slowed and appeared to reach a resistant
level at about 10 survival after 2 sec at 3 yM and 2C. If these surviving
PFUs are aggregates, they must have "grown" from single particles during the
short time interval before freezing or in the thawing process. This has lead
us to doubt the value of freezing stock virus preparations to prevent the
state of dispersion.
Five ml of the same virus were treated with bromine in the same way, except
that it was centrifuged 17 min at 20,000 RPM at 20C in a Beckman SW50 rotor.
Only the top 4 ml were used in the bromine experiment. All large aggregates
should have been removed from this preparation. The resulting rapid linear
decline in PFUs (Fig. 4) provides excellent confirmation.
Partition of crude reovirus between singles and large aggregates by means of
a sucrose gradient in the swinging bucket centrifuge rotor.
Zonal sedimentation velocity spectra of crude reovirus (Group 3 above) yields
a prominent sharp band of single virions and relatively few small aggregates,
but most of the virus seems to have sedimented much faster than either. An
experiment was done in the large (37-ml) buckets of the Beckman SW-27 cen-
trifuge rotor to determine, if possible, the quantity and physical state of
virions and the plaque titer of the two major components of this particle
population, the singles band and the pelletized material. To do this, a
15-30% sucrose gradient was established and 10 ml of the freeze-thaw lysate
(the same as that used in experiment 3) was layered over it. After spinning
at 25,000 RPM (mean centrifugal field 81,000 XG) for 1 h, the supernatant
fluid above the visible singles band was discarded. Ten ml, including the
singles band, were collected and then the pelletized material was resuspended
in the remainder of the supernatant fluid by pumping with a pipette. These
two fractions will be called Band 1 and Pellet 1.
16
-------
Fig. 7 The total yeild of reovirus extracted
from a given number of infected cells is about
the same whether it is done by 3 freeze-thaw
cycles (top panel) or by freon extraction
(bottom panel), but only 27% of it appears as a
band of single particles when the crude virus is
sedimented in a 15-30% w/w sucrose gradient. The
rest is in the pellet. For freon-extracted virus
the partition is 53% to 47%. Fig. 6 shows that
there are very few small aggregates present.
PARTITION OF CRUDE VIRUS
8 O.B
PARTITION OF FREON EXTRACTED
VIRUS
An equal quantity of the crude virus preparation
was extracted with freon and layered over a
similar sucrose gradient, centrifuged, and
harvested in the same manner, Band 2 and Pellet 2,
Half of the resuspended Pellet 1 material was
extracted with freon, made up to the same volume
(10 ml), and banded and harvested like the pre-
vious ones, Band 3 and Pellet 3.
A part of fractions Bl and PI were treated 30 sec with 20-KHz acoustic waves
from the micro tip of a Branson Sonifier Model LS-75. The sample volumes
treated were 5 ml each, and they were immersed in ice water during treatment.
The maximum temperature attained during treatment was 15C. There were 8
samples in all prepared for virion count and observation of aggregation in
the electron microscope and for plaque titration. The results of physical
assay are shown in Fig. 7. Particle count of pelletized fractions from freon-
extracted preparations contained a large amount of cellular debris, which
made the virions difficult to count. Nevertheless the total particle yield
(sum of counts from pellet and band) was approximately the same. Freon
extraction yielded 57% of the particles in the band, but with the crude
virus only 28% was in the band. Freon extraction of virus from the pelletized
fraction of the crude virus experiment (PI) yielded 67% of the particles in
the band. Apparently the freon extraction process is quite efficient for
extraction of reovirus in monodisperse form from the infected cells.
Plaque titration of banded and pelletized fractions from several identical
partition experiments have shown 1/7 of the freon-extracted virus to be in
the pellet and 6/7 in the band. The partition of PFUs in crude preparations
has been erratic. The pelletized fraction always has much more than the
band, and once there were 50 X as many PFUs in the pellet as there were in
the band. Treatment of banded and pelletized fractions with 20-KHz acoustic
waves has not made any substantial change in particle count but in this case
it reduced the plaque titer of the pelletized fraction by a factor of 10.
The same treatment has made no change in the titer of the well-dispersed
virus in the banded fraction.
17
-------
POLIOVIRIONS
VELOCITY BAND
IN
BECKMAN
SW-27 Tube
10-30%
SUCROSE
Fig. 8 Poliovirus velocity banded in a sucrose density gradient after freon
extraction from infected HEp-2 cells.
18
-------
STARTING VOLUME
occupied by
Freon Extroct
of
nfected Hep2 Cells
8 10 12
FRACTION NUMBER
10 II 12 13
ROTOR RADIAL POSITION (incm)
Fig. 9 Graphical display of poliovirus particle count by EM, taken from the
fractions and pictures shown in Fig. 8.
POLIOVIRUS
The virus used in these experiments was extracted from the infected HEp2 cells
with freon and velocity-banded in a sucrose gradient, and the particle content
of several successive fractions from such a gradient is shown in Fig. 8. The
concentration of particles calculated from counts from such pictures is plot-
ted against radial position in Fig. 9. Many particles smaller than virus,
probably ribosomes, can be seen in fractions 11 and 12 just above the virus
peak. These were not counted. At no point below the main virus peak was
there any secondary peak in the region where pairs, triplets, etc., would
have been expected had they been present in substantial numbers. It appears
from this that the freon extraction process has yielded a virus preparation
with a very high proportion of single particles.
The strip pictures of Fig. 8 were taken at high particle concentration,
presenting enough particles so the reader may judge their relative number
and degree of purity. Other pictures made with more dilute preparations
were used to determine the true degree of aggregation, free of the coinci-
dence effect of one particle falling upon another by accident. Calculation
of the number of such accidental pairs is made in the discussion. They
indicate that the starting virus for these experiments contains at least
95% single particles.
Inactivation by bromine at 2C and pH 7
Six experiments were made at 2C, three of which are shown graphically in Fig.
10. All were characterized by an initial linear phase, constant decline in
19
-------
1°
cc
•'
to -2
UJ
ID
o
g
4 8 12 16
REACTION TIME (seconds)
Fig. 10 Inactivation of poliovirus at 3 different concentrations of bromine
at pH 7 and 2C; top line 0.6 pM, middle line 2.2 yM, and bottom line 22 yM.
4 8 12 16 20 24
BROMINE CONCENTRATION (>uM)
Fig. 11 Inactivation rates (Slopes in Logs/second taken from Fig. 10, 12,
and 13) for poliovirus at pH 7 as function of temperature and bromine con-
centration.
20
-------
log PFU survivor ratio per unit time of exposure. None of these reactions
showed any tendency toward delay during the first time interval which was 4
sec. The slopes of all the reactions ranging from 0.6 to 22 yM bromine con-
centration have been plotted (Fig. 11) where it can be seen that increasing
bromine concentration does not produce a linear increase in reaction rate.
The increase becomes progressively less with increasing bromine concentration,
indicating a progressive decrease in efficiency. Similar results have been
calculated from the data of Weidenkopf
as previously published.
In several inactivation experiments at bromine concentrations of 3.5 yM or
greater, there has been a slight increase in the reaction rate after a sub-
stantial period of linearity (see the 22 yM line of Fig. 10). We have not
been able to produce this effect regularly. Slopes for Fig. 11 were taken
from the linear part of each survival graph.
Inactivation by bromine at IOC and pH 7
At IOC the inactivation rate was constant at 1.9 yM bromine concentration
(Fig. 12). There is no indication, in this experiment, that there is any
time delay before inactivation begins. However, at higher bromine concen-
trations the linear part of the graph begins only after the first 4-sec
time interval (Fig. 12). It seems unlikely that the initial irate at the
higher concentrations could be less than that of the straight line for 1.9
yM bromine so the lower curves have been drawn tangent to it at zero time.
0
QL -I
-2
<
cn
UJ
Q.
o
ef -4
o
-5
4 8 12 16
REACTION TIME (seconds)
Fig. 12 Inactivation of poliovirus at 3 concentrations of bromine at pH 7
and IOC, top line 1.9 yM, middle line 5.9 yM and bottom line 10 yM.
21
-------
<
o:
g
cc
Z)
O
o
o
4 6
REACTION TIME (Seconds)
Fig. 13 Inactivation of poliovirus at 3 concentrations of bromine at pH 7
and 2QC; top line 1.9 pM, middle line 5.9 yM, and bottom line 10 yM.
The reaction rates, slopes of the linear parts of the three kinetic experi-
ments, are shown (Fig. 11) to be a linear function of bromine concentration
with an intercept indicating zero reaction for zero bromine.
Inactivation by bromine at 20C and pH 7
At 20C the inactivation of the virus proceeds faster at all bromine concen-
trations that it did at IOC (Fig. 13). There is some indication in the
graph for the 9.5 yM experiment that aggregation is showing its effect at
survival levels below 10 . Reaction rates were determined from the straight
part of each line.
Inactivation by chlorine
Published rates of inactivation of poliovirus by chlorine have been approxi-
mately one fifth of the rates we have found for suspensions of single polio-
virus particles in bromine. We suspected that a part of this difference
was due to the presence of aggregates in virus preparations used in earlier
work so we made some chlorine experiments using the same apparatus and
technique that were used for bromine, except that the pH was reduced to 6
in order that the major active agent would be HOC1 for direct comparison
with the HOBr of the earlier experiments. The buffer was the same (Phos-
phate), and its concentration in the flowing stream was 0.01 M.
Seventeen inactivation experiments were performed with chlorine concentra-
tions ranging from 1 to 40 .yM and temperature of 2C and IOC. Then eight
more experiments of the same type were made without any chlorine. The
22
-------
o o
< -I
cr-
_:
0- -3
O
o
o
-4
1
1
!
4 8 12 16
EXPOSURE TIME (Seconds)
10 20 30 40
CHLORINE CONCENTRATION (juM)
Fig. 14 (left) Inactivation of poliovirus by HOC1 at pH 6 sometimes proceeds
at a constant rate (circles) and sometimes it appears to start quickly then
stop (squares).
Fig. 15 (right) Inactivation rate of poliovirus at 2C and at IOC as a func-
tion of HOC1 concentration.
results were, at first, very confusing. Some of the experiments showed strict-
ly linear semi-log disinfection behavior like that shown on Fig. 14 and others
were characterized by a quick initial drop in titer followed by an interval
of very little change (also shown on Fib. 14). When the slopes of all the
experimental lines that were straight are plotted together as a function of
HOC1 concentration (Fig. 15), the relationship does not appear to be a simple
one. At 2C the rate increases from about 0.08 logs/sec at 1 yM to double
that value at 20 yM but there was no further increase in disinfection rate at
40 yM HOC1. The line looks as though it would intersect the vertical axis
before reaching zero chlorine concentration. Inasmuch as this is not possible,
the decline in rate below 1 yM HOC1 must be very fast indeed. This in in-
dicated by the dotted line on the graph.
At IOC the increase in disinfection rate with concentration of HOC1 is linear
from 5 to 20 yM but the line lies below that for 2C; the reaction is defini-
tely slower in this concentration range, although it appears that it might
rise above the 2C line if extended to higher concentrations. Something like
this was encountered in the bromine experiments with poliovirus (also
reproduced on Fig. 15 for comparison), but the reason for it is still obscure.
In an effort to explain the curved line of Fig. 14, a series of eight experi-
ments was made without chlorine in the flowing stream. In the previous
chlorine experiments the stock virus was diluted with 0.001 M phosphate buf-
fer (pH 6) just before injection into the flowing stream of chlorine-water
containing 0.01 M phosphate buffer at pH 6. Suspecting that dilution of the
virus concentrate with reduction in both sucrose and salt concentration and
23
-------
change in pH may have caused the virus to aggregate, we made dilutions from
stock in chlorine-demand-free (CDF) water and reduced the buffer concentra-
tion in the stream from 0.01 to 0.002 M. Experiments run at 2, 10 and 25C
all give titration curves like that of Fig. 14 with different degrees of
initial drop, ranging from 0.3 to 1.4 log units. When the stock virus was
diluted with 0.14 M Nad solution and injected into 0.020 M buffer at pH 6
in the stream, there was no drop in titer other than that expected from dilu-
tion at the point of injection. Also, when the pH of the streams was raised
from 6 to 7, there was no drop in titer with the passage of the virus through
the apparatus. These experiments removed the possibility that virus was
being inactivated by residual chlorine or chlorine compounds in the water
after dechlorination, and they strongly suggested that curvature in some of
the semi-log plots of survival titer may have been caused by aggregation of
the virus during the experiments. As a result of this, a study was made of
all experimental data taken with both reo and poliovirus during the past 3
years to learn what conditions, likely to be encountered in water experiments,
would promote virion aggregation. Findings of this study and results of new
experiments that it suggested are presented below in Physical State of Viruses
in Water.
Inactivation by Tribromamine
13
Nitrogen tribromide was generated at pH 4.5 by pouring together and stirring
equal volumes of buffered ammonium chloride and standard bromine solution.
The final molar ratio (NH^iEr-) was 1:3 in the test reactor. Identical ratio
solutions were scanned in the UV on a Gary 14 Spectrophotometer against a
buffered ammonium chloride reference to check the purity of the NBr_ formed
at 258 and 323 nm.
Reaction rates with polio virus at 4C were quite similar to those for HOBr
at the same molar concentration, temperature, and pH. No evidence of cur-
vature appears in the semi-log plot of survival ratio against treatment time,
and the dependence of reaction rate
on tribromamine concentration (Fig.
0.30h ^ 16) is quite similar to that of
HOBr (Fig. 11).
0.25
o>
Q.
O
O
O
d 0.20
O
i
0.15
< 0.10
0 0.05
UJ
0 10 20 30 40 50
TRIBROMAMINE CONCENTRATION (juM)
THE PHYSICAL STATE OF VIRUSES IN
WATER.
Reovirus
An example of many concentrates
prepared by freon extraction and
Fig. 16 Inactivation of polio-
virus by tribromamine.
24
-------
Fig. 17 A stock reovirus preparation diluted 10 x then 20 x with PBS,
The virions are well dispersed.
25
-------
.2
.4
.6 .8
GROUP SIZE
Fig. 18 The distribution of group
sizes counted from 5 pictures like
Fig. 17 and from 5 more pictures
aggregated like Fig. 19.
velocity banding in sucrose con-
taining 0.05 m phosphate buffer pH
7.0 is shown in EM picture Fig. 17.
The stock was diluted in two steps,
10 x then after 3 h, 20 x more
with PBS and sedimented for agar
pseudoreplica counts by EM. The
number of single particles, pairs,
etc., per thousand counted is plot-
ted as squares on the frequency
chart Fig. 18. The degree of
aggregation is very nearly the same
in every preparation. Single particles are 79% of the total, which was 1.5
x 10 virions per ml.
Quite a different picture is obtained when this stock is diluted the first
10 x step with water and then 3 hours later the second 20 x step with PBS
as above. The water dilution reduces the sucrose to about 3%, the buffer
concentration to 5 mM, and the virion concentration to 1.5 x 10 /ml. These
conditions regularly produce the aggregation seen on Fig. 19, and it is
apparent that the aggregates were 1) not present in the stock concentrate
and 2) they were not dissolved when PBS was added in the final 20 x dilu-
tions. Their frequency is plotted as circles on Fig. 18, where the decreased
slope shows the change in state of dispersion.
The physical state of the virus in the undiluted concentrate cannot be seen
by EM, but pictures of virus deposited by Brownian bombardment of an alumin-
ized collodion film in contact with a 5 x PBS dilution of the concentrate
look just like Fig. 17 and the frequency distribution of group sizes from
such a set of pictures is like the squares of Fig. 18. This preparation was
diluted 100 X in one step with 5 mM phosphate buffer at pH 7 and sedimented
for agar pseudoreplica count and group frequency counts. The plot of fre-
quency distribution was indistinguishable from that of the 5 x dilution
above. Presumably 100 x dilution did not change the relative frequency of
the aggregates. They neither increased nor broke up.
Sedimentation velocity analyses of freon-extracted concentrates have shown
faster moving peaks corresponding to pairs and other small groups, but
the presence and size of such peaks seems to be related to storage time
(4-6C). Freshly prepared virus has not shown any sedimenting peak where
pairs should be. Equally sharp single peaks have been seen when the sucrose
gradient contained 0.85% NaCl without buffer, 0.01 M phosphate buffer at pH
7 or no salts at all. But dilution of concentrates with no salt at all have
given irregular results. Dilution of such a concentrate containing 1.5 X
10 virions per ml with 9 volumes of PBS produced aggregates on 15 min of
26
-------
Fig.
PBS.
19 A stock reovirus preparation diluted 10 x with water then 20 x with
The virions are aggregated. Group frequencies are shown on Fig.
27
-------
standing that were not dispersed by 15-fold further dilution with PBS, but
when the two dilution steps were made without delay or if the 150 x dilution
were made in one step, there was no aggregation.
The water-banded virus (sucrose gradient without salt) had only a single peak
that we proceeded to use as a source of single particles for a bromine inacti-
vation experiment, but we found after the experiment that the virus had aggre-
gated when it was diluted into the 0.01 M phosphate buffer at pH 7 Fig. 19.
It is not possible to say just how much of the aggregation we saw was present
during the 4 sec of the bromine action, but the semilog plot of the surviving
virus titer was curved like Fig. 4, very different from that obtained sub-
sequently when we were able to produce suspensions of singles only. So it is
almost certain that this curve was caused by aggregation of the virions either
just before or during the disinfection experiment, even though the starting
virus showed no aggregation.
Other independent evidence of aggregation of reovirus particles, particularly
when they are diluted from sucrose-banded concentrates into water, comes from
partition experiments in a swinging bucket centrifuge tube. If a centrifuge
tube is filled with a diluted suspension of virus and spun at such a speed
and time as to just sediment all the single virions out of the top quarter
of the tube, and the rotor is carefully stopped and the top half of the
supernatant fluid is collected, mixed and titrated, there should be 1/2 of
the original titer remaining if all the particles in the tube were singles.
If there was aggregation present in the suspension, the remaining PFU must
be less than 1/2. While this test will not give a stiictly quantitative
measure of the aggregation, it has the advantage of providing evidence for
or against aggregation at virion concentrations much less that those required
for electron microscopy. This Ifolf and Half experiment was tried on a stock
of poliovirus containing 3 x 10 virions per ml in 0.05 M phosphate buffer
at pH 7 in 33% sucrose. This stock virus was diluted 10 x with water and
with PBS. After standing for 3 h at room temperature, both dilutions were
diluted another factor of 10 with the same diluent as before. Both were
titrated and the PBS dilution revealed 5.7 x 10 PFU/ml, while the water
dilution made 3.2 x 10 plaques per ml. This small reduction in titer
might be evidence of aggregation of the virions in the water. Both dilutions
of virus were then subjected to the sedimentation velocity partition experi-
ment described above. The saline dilution left 56% of the PFUs behind in
the top half of the supernatant fluid, while the water dilution left only 21%.
Apparently many of the PFUs in the water dilution sedimented faster than the
single virions, which have been shown to be the chief component of the saline
dilution.
PolioviEus
The physical state of poliovirus after freon extraction and velocity-
banding in a density gradient of sucrose has not been determined as well
as that for reovirus. Poliovirus cannot be prepared for EM counting and
aggregation analysis by the agar pseudoreplica method. Only the KA method
has been successful in producing EM pictures for aggregation analysis, there-
fore only virion concentrations 10 and above can be examined directly.
28
-------
When poliovirus is extracted from infected cells with freon and velocity-
banded on a sucrose gradient at 5 x 10 virions per ml, the suspension does
not scatter enough light to be visible to the unaided eye. This water-clear
virus suspension contains .05 M phosphate buffer at pH 7 and 22% sucrose.
Dilution with PBS by a factor of 5 or more permits kinetic attachment of
single virions and aggregates to aluminized films for EM. Four pictures
from a 10 x PBS dilution of virus were divided into 5 equal areas each and
counts of single particles, pairs, triplets, etc., were made on all 20 areas.
A part of one such picture is shown (Fig. 20), and the group frequency data
are shown in Table 1, where it can be seen that the mean singles count per
area was 124 with a standard error a=15 and a mean frequency of pairs was
observed to be 3.4T of the total count. Inasmuch as 124 particles would
"exclude" 6% of the picture area, we estimate an average of 7 pairs per
picture would.^have been produced by one single particle accidentally falling
upon another. The expected accidental frequency of triplets and groups of
four has not been calculated, but it appears from the number of pairs ob-
served that the actual amount of aggregation in this polio stock suspension
is very near zero, even though the total particle -concentration is 6 x 10
per ml. These pictures were made 19 days after this stock solution was
prepared. Another observation made 81 days after preparation made from
plates averaging 1104 total virions showed an average of 86 pairs per picture.
Since 64 were expected by coincidence, there seems to have been an irrever-
sible pair production of 22 per 1104 or about 2% in this time.
EM Observations
As with reovirus, dilution of poliovirus stock with water (without buffer
or neutral salt) tends to aggregate the virions, but the nature of the aggre-
gation is very different from that seen with reovirus. In Fig. 21 is shown
a 10 x water dilution of the same stock virus from which the salt dilution
gave Fig. 20. The polio aggregates are fewer, contain many more virions
than reo aggregates, and do not spread out flat when deposited on the alumin-
ized film for EM. They appear to be more tightly bound than the reo aggre-
gates that spread out and rarely appear to pile up. Still, the polio aggre-
gates formed by water dilution are readily broken up by addition of salt to
the suspension. A 2 x further dilution of the virus of Fig. 21 with PBS
dispersed the aggregates quite completely (Fig. 22).
At pH 7 in .05 M phosphate or tris buffer, the virus is physically stable
(as it is in .05 M borate buffer at pH 8 and at pH 9), but at pH 6 and .05
M phosphate buffer, aggregates form as they do also at pH 5 in .05 M acetate
buffer (Fig. 23). There is some evidence already that these isoelectric
aggregates are dispersed when the pH is increased to or above the neutral
range.
We have not observed any conditions under which any virus which aggregates
when diluted with a given diluent solution will deaggregate on further
dilution with the same diluent. In the present work, we were limited in
search of such data by the lack of sensitivity of the KA method. Neverthe-
less, it is a matter of considerable importance to know what happens to
29
-------
TABLE 1.
Group Size
EXAMPLE OF NUMERICAL DATA FROM 20 PICTURE AREAS TAKEN WITH
POLIOVIRUS PREPARATIONS LIKE THOSE ON FIG. 20.
Group
Frequency
Standard
Error a
Total Virions
Counted
:
_
3
4
124
4.5
0.35
0.05
15
2.1
2480
89
1
i
Fig. 20 Typical poliovirus preparation for EM count by the kinetic attach-
ment method. The virions are well dispersed.
30
-------
Fig. 21 (left) Water dilution (10 x) of poliovirus stock suspension aggre-
gates most of the virions.
Fig. 22 (right) The aggregated poliovirus of Fig. 21 is dispersed by dilution
with PBS.
such aggregates as are formed when stock viruses are diluted 5 to 10 x in
water, when they may be diluted much further with water.
Dilutions of poliovirus in water - beyond the level visible in the EM.
It was determined by experiment that 20 min at 30,000 RPM would sediment
about 50% of well-dispersed virus out of the top half of an SW 50.1 centri-
fuge tube at 20C. When such a tube was filled with a 10 x water dilution of
-------
Fig. 23 Poliovirus aggregates in .05 M Acetate buffer at pH 5.
32
-------
stock virus, allowed to stand for 1 h at 25C, and then centrifuged (using
slow start and coast stop) 4% of the original PFU remained in the top half
of the test tube. When this same water-dilution was diluted 2 x further
with PBS and centrifuged as before, 48% of the PFUs were found in the top
half of the tube. This experiment, performed at about 2 x 10 PFUs/ml
(2 x 10 virions/ml) , shows that the aggregates were present in the initial
water dilution but that they were essentially dispersed by further dilution
in PBS.
3
When 10 x PBS dilutions of virus were diluted with more PBS to ca 10 PFUs/
ml. 58% of the original PFU remained in the top half of the centrifuge tube
after partition, showing good dispersion of the virions. Water dilutions of
the original 10 x water dilution left much less virus in the supernatant
region, ranging from 14% down to 1.3%. Apparently the aggregates that were
formed by the first 10 x water dilution and seen at that level of concentra-
tion by EM do persist when diluted 100,000 times further.
Because dilutions for plaque titration are normally made with PBS, it was
not surprising to find that 10 x water dilutions of stock poliovirus usually
gave the same plaque titers as 10 x salt dilutions, even though the former
must have been severely aggregated at the start of the dilution series for
titration. When dilutions of the two were made with 1/4 molar sucrose
solution and monolayer cells were washed with the same before receiving
the virus inoculum, the titer of 10 x water dilutions (aggregated virus)
was usually lower than that of the salt dilutions but never less than half,
and sometimes the two titers were the same. This is the only attempt so far
to dilute the virus and put it on the cell monolayer in aggregated form.
Stepwise dilution of stock virus by factors of 10 in water have produced
stable aggregates, expecially if there is some time delay after the first
step. However, dilution of the same stock virus by a one-step factor of
100 with water has sometimes produced suspensions in which no aggregation
could be detected by sedimentation partition.
Natural Aggregates of Poliovirus Particles
The reovirus can be photographed directly in the EM as it emerges from infec-
ted L cells (Fig. 24). With poliovirus, evidence for original aggregation
is not quite so direct, but cytoplasmic crystals of poliovirus are shown in
the textbooks, and we have dialyzed crude freeze-thaw lysates of infected
HEp2 cells over Amicon XM300 filter membranes and managed to get clear
pictures of large polio particle aggregates (Fig. 25).
These aggregates are generally different from the induced aggregates just
described. They are usually not dispersed by any of the conditions that
easily disperse the induced aggregates. They are probably present along
with induced aggregates in such preparations as the dialyzed one mentioned
above. In fact, Fig. 25 might be showing both. There are groups showing
evidence of two dimensional symmetry which may be reassembled from singles,
and there are other groups enmeshed with what appears to be the cellular
material in which they were formed.
33
-------
Fig. 24 Reovirus within and emerging from the periphery of an infected L
cell. Such large clumps are very common.
34
-------
Fig. 25 Poliovirus, dialyzed crude preparation showing aggregates of
virions enmeshed with cellular debris and other aggregates that appear to
have reassembled after dispersion.
35
-------
"0 3 6 9 12
REACTION TIME IN MINUTES
Fig. 26 Inactivation of poliovirus,
dialyzed crude preparation, by 10 yM
bromine at 2C.
The resistance of plaque titer to the
action of bromine on suspensions of aggre-
gated poliovirus has been compared with
that of single particles.
In Fig. 26 are the results with 10 pM
HOBr at pH 7 acting at IOC upon crude
virus rendered essentially demand-free by
dialysis across Amicon XM 300 membranes.
The semi-log plot of survival in HOBr
was very similar to that for aggregated
reovirus. After the curvature begins, at
the 2 log level, the survival of aggre-
gated PFUs quickly becomes several factors
of 10 greater than the projected survival
of singles (Fig. 12). This discrepancy gives every indication of increasing
with continued treatment.
Most of our virus preparations above started with the release of the virus
from infected cells by freeze-thaw cycles in buffered saline solutions.
Virus in field water will probably not come about that way. So we have
collected some virus by freeze-thaw release directly into distilled water.
Such preparations have some unexpected characteristics, as revealed by the
zonal centrifuge. After velocity spectrum runs like that of Fig. 6, titra-
tions were made on the water virus and there was no peak where the single
particles ought to be (Fig. 27). A repeat of this same experiment with
another virus preparation months later gave the same results. A broad peak
of PFU appeared where groups of 4, 5, or 6 would be expected, but no peak
appeared where it did when dispersed virus was centrifuged the same way.
We cannot count poliovirus at these concentrations in the EM, but these
titration experiments show that it is aggregates with which we must deal
when this work moves into the field where the water is not nicely buffered
with appropriate salt solutions at physiological concentrations.
These experiments and some suggestions by Dr. Stephen Schaub led us to
examine the.state of poliovirus put into water at about the concentration
level of 10 PFUs/ml. This involved preparing virus that was monodisperse
and virus that consisted mostly of natural aggregates. Each of these was
into water at about 10 PFUs/ml, and an appropriate test was made immediately
and again after 1 h to tell whether either preparation had become more or
less aggregated.
The "appropriate test" we chose was the one shown Fig. 28 involving a small
plastic horizontal centrifuge tube with 1 ml of 50% sucrose in the bottom,
2% ml of 20% sucrose above it, and 2 ml of the dilute virus sample to be
tested at the top. A control tube is prepared the same way except that
36
-------
10 12 14 16
FRACTION NUMBER
Fig. 27 Zonal Sedimentation Velocity spectrum of the poliovirus PFU in a
15-30% w/w sucrose gradient. Peak of single particles shows clearly in the
freon-extract preparation (open circles), but when virus was released from
infected cells by 3 freeze-thaw cycles, there was no single peak, only a
broad band in which there were small aggregates (open squares).
VIRION AGGREGATION TEST
(Red, White ond Blue)
BEFORE SPIN
r
AFTER SPIN
E
o
10
1
H 3 E
-------
oo
TABLE 2. THE PHYSICAL STATE OF POLIOVIRUS IN WATER.
DATA FROM THE RED, WHITE AND BLUE EXPERIMENT OF FIG. 28.
University Lake
Alum-Treated
(Finished)
Phosphate
Water Type
Input-Single Virions
1 min W
B
R
60 min W
B
R
Input Aggregated Virus
1 min W
B
R
60 min W
B
R
Distilled
36
62
2
32
62
6
7
21
72
5
26
69
Turbid
0
57
31
12
58
33
9
22+
70
54+
Clarified
0
57
39
7
51
45
4
15
34
51
14
43
43
Dechlorinated
^
43
30;
44
10
19
71
7
19
74
Buffered Saline
60
37
3
55
44
1
?£
4+
37+
Average of 4 experiments.
Average of 2 experiments.
-------
the top section contains the best preparation of single virions we can pro-
duce. A little neutral red in the bottom section and some trypan blue in
the middle section helps locate all three (red, white, blue) for plaque
titration after centrifugation. The pairs of tubes are spun long enough to
sediment single particles about half way through the top (white) section.
This has been determined by experiment, and the value (30,000 RPM at 20C for
20 min) agrees quite_well with that calculated for polio whose sedimentation
constant is 160 x 10 S. When this is done, the other half of the PFUs
in the control tube appears in the middle (blue zone) and only traces have
been found in the red. Samples containing aggregated virus together with
singles will leave less than 50% of the PFUs in the white zone with the
rest distributed between blue and red, according to the frequency of small
and large aggregates. The red, white, and blue (RWB) diagnostic technique
has been used also to separate a single crude (freeze-thaw) virus preparation
into two parts. The white part will contain practically all singles, while
the red zone contains only aggregates with very few singles. These two
fractions have been diluted about 5000 x with distillled water, one drop in
200 ml in 250 ml Erlenmeyer flasks with magnetic stirring bar. These were
stirred gently and held at 25C. One sample of each was removed after 1 min
and another after 1 h. Each was put in an RWB tube, and spun at 30,000 RPM
for 20 min at 20C. All three sections of each tube were then titrated on
monolayer cultures of HEp-2 cells, and the results were compared with pre-
vious runs with specially prepared singles. The results showing the aggre-
gation spectrum of virus in distilled water are in Table 2, where it can be
seen that single particles put in distilled water remain singles and that
natural aggregates remain aggregated at least for the 1 h of the experiment.
Similar data were obtained with natural resource water (University Lake -
water supply for the city of Chapel Hill) with and without particles. But
the results were very different when the finished water from the filter
plant was used. The table shows that this water aggregated the virus. This
experiment has been repeated 4 times with the same result. It is possible
that trivalent aluminum and divalent calcium ions left over from the lime
and alum treatment of the water are causing the virus to aggregate when it
is put into this treated water. We expect to investigate this soon.
DISCUSSION
INACTIVATION OF REOVIRUS
Reovirus was chosen for a part of this work from among the many water viruses
because of its size which makes it readily amenable to physical assay by two
essentially independent methods. Earlier work has shown that the reo
particle counts and aggregation analyses made by the^KA method of preparation
for EM yield the same results with suspensions of 10 virions/ml as those
obtained by the sedimentation or agar pseudoreplica method on the same sus-
pension after approximately 100-fold dilution. Demonstration of this fact
is essential, of course, for if it were not so, the dilution of suspensions
of aggregated reovirus for plaque assay would yield equivocal information
about the bromine resistance or plaquing efficency of the aggregates seen
in the EM. For this reason, reovirus may provide a useful model of refer-
39
-------
ence when experiments of this kind are made with the many smaller picorna
viruses that are of greater significance as water pollutants but which are
more difficult to assay by physical means.
The reaction of bromine (3 uM) with preparations of reovirus in water at pH
7 is shown here to destroy 99.9% of the PFUs in 1 sec at 2C if the prepar-
ation contains predominately single virions. After the most strenuous
efforts to remove all aggregated virus, the logarithm of the survival ratio
is an almost linear function of the time of exposure to the bromine at least
to the level of 1:10,000 surviving PFUs, the titer falling to the rate of
3.0 logs^Q/sec. This is 25 x faster than thefirate observed with poliovirus
under similar conditions of bromine exposure. _It is more than 100 times
faster than rates indicated by Scarpino et al and Weidenkopf , respec-
tively, for chlorine at pH 6 and the same molar concentration and approxi-
mately the same temperature. Comparison of bromine rates at pH 7 with
those of chlorine at pH 6 seems most appropriate for comparison of the
effects of HOC1 and HOBr.
While reovirus, velocity banded in a sucrose gradient, is usually 90 to 95%
single particles, it gradually aggregates in such a way that the log of the
frequency of any group size becomes and remains a linear function of the log
of the number of particles in the group (Figs. 3 and 18). Inactivation by
bromine of such aggregated virus is characterized by a continuously decreas-
ing decay constant. This indicates that plaque-forming aggregates are more
resistant to bromine than single virions.
SURVIVAL OF REOVIRUS PFU THROUGH VIRION CLUMPING - A CLOSER LOOK
If about one PFU in 10,000 of the starting titer were an aggregate large
enough to protect one potential plaque-forming virion from destruction by
the bromine, this would be enough to account for the persistent infec-
tivity seen in the experiments of Fig. 2. If such an aggregate consisted
only of virions (no extraneous material) there would seem to be no way
that such small groups as 2, 3, or 4 could conspire to protect even one
of their number from attack by the large number of small bromine molecules
or ions. Effective protection would not seem possible with much less than
a complete monolayer of particles surrounding the protected one, and this
can happen with spheres of equal size only if there are at least 17 in
the clump. Clumps,of this size would be easy to see with the EM, but at
a frequency of 10 they might go unnoticed. The pictures did show that
no groups larger than 7 virions were present at a frequency greater than
1:1000 and such a cluster could not quite cover one half of the surface
of one of them. It could not afford that one much protection. How then,
can such a survival curve as that shown for the slightly aggregated reovirus
of Fig. 4 come about? There have been suggestions ' .
A part of the graph (Fig. 4) has been repeated twice for demonstration of
two different qualitative analyses. If all the virions are alike in this
mixture of singles and aggregates and the probability that any single
particle will make a plaque is E, then the starting titer when t = 0 (be-
fore bromine treatment begins) must be given by:
40
-------
where
N
2> N3> etc., are the numbers of singles, pairs, triplets, etc.,
M is the maximum clump size present in significant numbers, and C?, C0, C
etc. are coefficients, the values of which we will now seek.
3' 4'
If we were dealing with perfect virus (E=1.00), then each single, each pair,
each triplet, etc., would make just one plaque and no more; C=C~=C, etc.
=1. But the reovirus used here had about 40 X more particles like those
of Fig. 17 than the plaques it produced. This is saying that on the average
only one virion in about 40 does make a plaque. If random pairs, triplets,
etc., are made from such singles there is only 1:40 chance of finding 2 of
the plaque- forming singles in one such pair so there will be little loss
through redundancy, and the probability of plaque formation by a pair, trip-
let, etc., would seem to be very nearly 2 X, 3 X etc., greater than that for
single particles. Briefly, if the plaquing efficiency of the virus is low,
the probability of plaque formation by a small clump of i particles would be
approximately i times that for any one particle and equation 1 above can
be approximated by:
If this is the starting titer, its logarithm will be subtracted from that
of each survival titer so the E, whatever its value, does not appear in the
difference that is the logarithm of the survival ratio which is plotted on
the disinfection graph. The numbers N , N , N~, etc. are supplied by the EM
(like Fig. 17), and the logarithms of the corresponding set N , 2N7, 3N ,
etc., up to MN are plotted as
shown (Fig. 29*f along the verti-
cal zero time axis. (The set used
to plot the dotted line on Fig. 4
as it appeared in the original
<
tr
ir
en
publication
was simply N ,
N~
~,
etc.). The EM data do not go
much beyond 1 group per 1000
particles so this analysis pro-
bably should not include groups
of greater than 7 particles (M=7)
and the approximation to the ex-
perimental curve should be
accurate until single particles
have been reduced 99.9% by the
bromine .
1234
TREATMENT TIME IN SECONDS
Fig. 29 Poisson approximation
to the reovirus disinfection
curve (Fig. 4) using aggregate
frequencies observed by EM.
41
-------
Having gotten the starting points for each group size in the mixture, one can
draw immediately the best straight line through the data points for inactiva-
tion of single particles. This would indicate a titer reduction from To to
Tm after a mean dose m of bromine; T = T e . If one assumes that the
survival of a group of i particles is predicted by the multi-hit target con-
cept then the Poisson limit gives
and the survival titer of the' whole mixture will be
We can now use the straight single's line to determine m in terms of t, then
calculate from equation 3 and draw on Fig. 29 the predicted survival curve
beginning at the proper point indicated above for pairs !„ = E (1 + m)e
for triplets T_ = E(l + m + yf)e etc. making 7 lines in all. Now, for
any chosen time, the predicted"survival titer for the mixture will be given
by the sum of the antilogs of the seven intersections with this vertical
time line. The locus of the logarithms of these sums forms the curve drawn
(Fig. 29) for comparison with the experimental data from Fig. 4. The fit
of the predicted curve to the data is quite good.
Another, perhaps equally good, fit to the experimental data can be obtained
by drawing straight lines (Fig. 30) from the same set of starting points that
were used in Fig. 29. The slopes of these lines (except the single's line
that has been determined by experiment) can be found by trial until the sums
of the contributions from all the significant group sizes make a total the
log n of which falls on or near the experimental curve for all times within
the range tested. By means of such graphical experimenting it is not dif-
ficult to achieve a good fit, as may be.seen in Fig. 30, and each of the
lines can be expressed as T. = T. e , where K. is the reaction constant
for groups of i particles. "TJnlike the multi-hit Poisson concept of virus
disinfection which gives the exact curve for each group size, this method
shows only that if all the
reactions are first order
(straight lines) a set of
slopes can be found, consis-
tent with the known numbers
of each aggregate size pre-
< \\\ \ sent, to account quite well
for the observed results.
-2
0-
O -I
tr
a -3
o
o
-5
1
TREATMENT TIME IN SECONDS
Both of these graphical
analyses provide a tolerable
fit to the experimental
Fig. 30 Straight-line
approximation to the reo-
virus disinfection curve
(Fig. 4) using aggregated
frequencies observed by EM.
42
-------
disinfection curv.e down to the point where the survival ratio of single
particles is 10 and that of the aggregated preparation is 10 . Beyond
this point the EM does not supply the needed accurate data on the frequency
of clumps containing 8 or more virions. One is tempted to extrapolate the
observed linear relationship between Log Ni and Log i to provide a rational
basis for extending the predicted disinfection curve, but this must be done
with care because of increasing closeness of the lines (Fig. 30) or curves
(Fig. 29) for each group size. This can be seen on either graph, and it
signals a limit to useful extrapolation. While this might discourage those
who seek a simple method of predicting the course of disinfection beyond any
given experimental limit, the above exercise has provided a means of answer-
ing a significant question about the mechanism of survival through aggregation.
The question is whether small clumps of virus, even pairs, are more resistant
to bromine action than single particles or whether, after exposure to bromine,
the virions of a pair or small clump can help each other in the act of plaque
formation that might not occur if they were acting separately. If we con-
sider the point at which the PFU titer of single particles has dropped by a
factor of 10 , some reason must be found for the observed survival titer of
the aggregated virus which is 10 times greater. There are not enough large
aggregates present to account for this, and only large aggregates can sur-
round and protect one or more of their number from contact with the bromine.
Even small aggregates are not frequent enough (Fig.3)—unless they survive
much better than single particles. By either of the graphs it is clear that
when the mean treatment (m=7) has reduced singles to e , which is just about
10 of their original titer, pairs, for example, must be 8 to 10 times more
resistant than singles to contribute enough to the total observed titer.
This author thinks it unlikely that two virions can contrive to protect each
other to this extent, therefore their demonstrated survival must be a
synergistic effect resulting from some form of complementation in plaque
formation. This is probably the dominant mechanism of reovirus PFU survival
in small clumps down to 10 , and perhaps farther, but earlier work
showed that the large aggregates were dominant in survival at about the 10
level. It seems, therefore, that both modes of survival are operating with
aggregated reovirus exposed to bromine.
Preparations of reovirus containing 90 to 95% single particles can be obtained
easily from either freon-extracted lysates or infected L cells or directly
from crude lysates in which virus has been released from the cells by 3
cycles of freezing and thawing. The freon extracts yield about 2 x as much
virus, but both appear equally well purified when harvested from the sucrose
gradient in the singles band in either the swinging bucket or the zonal
centrifuge rotor. Slow spontaneous aggregation occurs in preparations of
reovirus stored at 4-6C, and in these mixtures of aggregates there is a
continuum of sizes, the logarithm of whose frequency is a linear function
of the logarithm of the number of virions in the group. As aggregation
continues, the slope of this line changes but the line remains straight.
These aggregates appear to contain nothing but virions (like those in Fig.
19). They are "pure" aggregates. But the major part of the virus released
from infected cells by freezing and thawing alone, without homogenization
with freon, must consist of aggregates of a different kind. One difference
is immediately seen in the sedimentation velocity spectrum of such crude
43
-------
virus which is produced in the zonal centrifuge rotor. While about 1/5 of
the total virions and 1/50 of the PFUs appear in a prominent "singles" band,
there are very few groups of 2 to 10, such as those seen in spontaneously
aggregated preparations (Fig. 6). The supporting experiment in the swinging
bucket rotor supplied the quantitative data on the partition of the virus in
the crude preparations. It indicates clearly that the fast-reaction kinetics
established in this report, and certainly required to understand the bromine-
virus reaction, have been obtained with about 1/5 of the virions normally
present in crude extracts of infected cells, while the remaining 4/5 is in
quite a different physical state, one that appears to be much more resistant
to bromine action. One can reasonably expect that the virus found in polluted
water will be in a state more like the crude. The disinfection rates for
HOC1 and HOBr on suspensions of single polio virions (Fig. 15) are not very
different. Due to the small amount of chlorine experimentation we have done
and the aggregation problems encountered at pH 6, we prefer to defer discus-
sion of this point until more chlorine data are available.
PHYSICAL STABILITY OF VIRUS IN WATER
The earlier work with the EM has lead us to expect some of the results that
we have since detected by sedimentation velocity techniques in work with
very dilute virus in water. As the work progresses it becomes more and more
apparent that by "water" we do not mean a buffered physiological saline
solution. Nor do we mean water that has been distilled and/or demineralized.
Ultimately, we must deal with water that has an ionic strength much lower
than that commonly used to suspend virus in the laboratory (or much higher
for sea water), and it may contain a substantial concentration of divalent
and trivalent ions and it may have a very low buffering power. The experi-
ments we have done by diluting stock virus in distilled water and in weak
buffers to change the pH have, therefore, opened some doors and prepared us
somewhat for the real problem of virus in natural resource water. One of
the first things they have done is emphasize the importance of a "stock virus"
preparation for such laboratory work (see page 1).
One must have a stock virus suspension as starting material for experiments
involving halogen disinfection and the influence of virion aggregation upon
survival and titration efficiency. The stock must have a high infectivity
titer so that its decline may be observed for several powers of 10. Its
infectivity must be stable so that it can be kept and used as starting
material for several, perhaps many, experiments. It must not contain extra-
neous material which can exert halogen demand, and the virions must remain
monodispersed. This author is confident that no one has or ever has had
such a stock of virus. But the degree to which an investigator can approxi-
mate such a virus preparation will doubtless be a measure of the experimental
success of his work. With such a stock virus he can make comparisons of the
resistance of different viruses to a given disinfecting agent, etc., and he
may induce aggregation in various ways and experiment with disinfection of
such aggregates. Eventually, however, he will want one more stock virus of
a different kind—that is aggregated exactly like it will be when it breaks
out of an infected cell and is released into a water supply. Our attempts
44
-------
to produce these stocks of virus and the resulting physical states attained
when they have been diluted in water constitute a major part of this report.
While a maximum yield of single particles of either reo or poliovirus comes
by way of freon extraction followed by velocity banding on a sucrose gradient,
the virus preparations banded without the freon extraction step seem to be
just as free of aggregation and physically indistinguishible, except for
concentration. Therefore, by using virus released by freezing and thawing
we have been able to get both kinds of stock virus from a single crude cell
lysate under conditions somewhat like those that might obtain in nature. The
fact that neither reo nor poliovirus aggregates rapidly in concentrated
sucrose solutions has been most helpful for the single particle band, and
the large aggregate band can be collected separately and kept right in the
sugar in which they are banded. Since no bacteria grow in these prepara-
tions at 5C, there is no need to freeze them and jeopardize the monodisperse
condition by the temporary forming of phase boundaries.
Dilution of stock virus with water is probably not like anything that occurs
in nature, but it is a laboratory necessity. Sucrose, virus, and dissolved
salts are all reduced in the process, and present data make it appear that
there is a critical dilution range in which virion aggregation occurs. Others
have reported the clumping of plant viruses when infected sap is purified
and the clumping of bacteriophage when it is put in seawater . The
latter authors did not find their bacterial virus stock clumping in water
of low or zero salt content; only when the salts approached the concentration
of seawater did clumping occur. Our two enteric viruses both show physical
instability (colloidal instability) with decreasing ionic strength, even at
pH 7, and if the virus particle concentration is still great enough when the
sucrose and salts have been diluted to the critical range, rapid clumping
occurs. In some cases, dilution with water can be done without clumping the
virus if the first step of the dilution is big enough that the transport
time for their getting together is long. One conspicuous exception to this
is the 5000-fold, one—step dilution of stock virus in the finished water
from the Chapel Hill filter plant. In this case, aggregation seems to be
instantaneous on addition of one drop to a gently stirred 200-ml volume of
water.
The two viruses we have examined show one conspicuous difference when aggre-
gated by dilution from stock into water. The reo aggregates do not break
up when diluted further with the balanced salt solutions that are used for
plaque titration in tissue cultues. Poliovirus aggregates, formed in the
same way, do break up when diluted with balanced salt solutions for titration.
This behavior could, and probably has, led to confusion in the past when a
virus stock has been diluted, become aggregated, and exposed to halogen
disinfection in this condition, and then dispersed again during dilution to
detect survivors by the plaque method.
The aggregation of both reo and polio viruses in weak buffers at pH 7 and
at pH 5 is another observation that is doubtless significant because it means
that survival of virus under these slightly acidic, but by no means unusual,
conditions will certainly be complicated by the clumping. It probably will
be prolonged and titration of the results will be complicated by dispersal
45
-------
or non-dispersal of the clumps in the titration process. Present titration
techniques do not permit the introduction of virus samples of low ionic
strength and pH 5 directly onto monolayer cultures of cells.
In summary, some of these experiments on aggregation of virus in water—not
balanced and buffered salt solutions but natural resource water—have reveal-
ed a means for extended survival of virus and a need for study of more than
just the two viruses that have been investigated here.
46
-------
REFERENCES
1. Sharp, D. G. Sedimentation Counting of Particles by Electron
Microscopy. Proc. 4th International Conference on Electron Microscopy.
Springer-Verlag, Berlin, 1960, pp. 542-548.
2. Sharp, D. G. Physical Assay of Purified Virus, Particularly the
Small Ones. 32nd Ann. Proc. Electron Microscopy Soc. Amer. pp. 264-265.
3. Clark, R. M. A Mathematical Model of the Kinetics of Viral Devitaliza-
tion. Mathematical Biosciences 2;413-423, 1968.
4. Wei, J. H. and S. L. Chang. A Multi-Poisson Distribution Model for
Treating Disinfection Data, Chapter 2 in Disinfection-Water and
Wastewater, J. D. Johnson, ed. Ann. Arbor Publishers, 1975.
5. Sharp, D. G., R. Floyd and J. D. Johnson. Nature of the Surviving
Plaque-Forming Unit of Reovirus in Water Containing Bromine. Applied
Microbiology 29:94-101, 1975.
6. Floyd, Roger, J. D. Johnson and D. G. Sharp. The Inactivation by
Bromine of Single Poliovirus Particles in Water. Applied Microbiology
31:298-303, 1976.
7. Sharp, D. G. , R. Floyd and J. D. Johnson. The Initial Fast Reaction of
Bromine (HOBr) on Reovirus in Turbulent Flowing Water. Applied
Microbiology 31:173-181, 1976.
8. Taylor, D. G., and J. D. Johnson. In A. J. Rubin (ed) Chemistry of
Water Supply Treatment and Distribution. Ann Arbor Science Publishers
1974, pg. 369.
9. Sharp, D. G. and P. M. McGuire. Spectrum of Physical Properties Among
the Virions of a Whole Population of Vaccinia Virus Particles. J_.
Virology 5:275-281. 1970.
10. Anderson, N. G. The Development of Zonal Centrifuges. National Cancer
Institute Monograph 21. U. S. Government Printing Office, Washington,
1966.
11. Kim, K. S. and D. G. Sharp. Electron Microscopic Observations on the
Nature of Vaccinia Virus Particle Aggregation. J_. Immunol. 97 ; 197-202,
1966.
47
-------
12. Geister, R. G. and D. H. A. Peters. Quantitative Use of the Electron
Microscope in Virus Research. Methods of Analyzing and Predicting
Biologic liters of Aggregated Virus Suspensions with a New Law of Aggre-
gation. Laboratory Investigation 14^864-874, 1975.
13. LaPointe, T. F., Guy Inman and J. D. Johnson In: Disinfection-Water
and Wastewater. J. D. Johnson, ed., Ann Arbor Science Publishers,
1975, Ch. 15.
14. Sharp, D. G. and M. J. Buckingham. Electron Microscopic Measure of
Virus Particle Dispersion in Suspension. Biochim. et Biophys. Acta
19_:13-21, 1956.
15. Scarpino, P. V., G. Berg, S. L. Chang, D. Dahling and M. Lucas.
A Comparative Study of the Inactivation of Viruses in Water by
Chlorine. Water Research 6:959-965, 1972.
16. Weidenkopf, S. J. Inactivation of Type I Poliomyelitis Virus with
Chlorine. Virology 5:56-67, 1958.
17. Brakke, M. K. Dispersion of Aggregated Barley Stripe Mosaic Virus
by Detergents. Virology 9:506-521, 1959.
18. Gerba, C. P. and G. E. Schaiberger. Effect of Particulates on Virus
Survival in Seawater. J. Water Pollution Control Federation 47:93-
103, 1975.
48
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-76-287
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Virus Particle Aggregation and Halogen Disinfection
of Water Supplies
5. REPORT DATE
December 1976
Issuing Date'
6, PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO
D. Gordon Sharp
9. PERFORMING ORGANIZATION NAME AND ADDRESS
School of Medicine
Department of Bacteriology
University of North Carolina
Chapel Hill, North Carolina
10. PROGRAM ELEMENT NO.
ICC 6lk
11. CONTRACT/GRANT NO.
R 803771
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory - Gin., OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio
13
T4. SPONSORING AGENCY CODE
TYPE OF REPORT AND PERIOD COVERED
Final 06/75 - 08/76
EPA/600/14
15. SUPPLEMENTARY NOTES
For supplementary information see "Virion Aggregation and Disinfection of Water
Viruses by Bromine," EPA-600/2-76-l63, PB 253 087.
16. ABSTRACT
Using a dynamic system the inactivation of polio and reo virus preparations
containing essentially all single and preparations containing aggregated viruses
was examined. Differences in resistance to bromine and chlorine were shown to
be caused both by inherent differences between different virus grouns and
by state of aggregation. Electron microscopic observations of polio and reo
viruses as they emerge from infected cells and in crude extracts indicated that
substantial numbers of viruses may be released from infected cells in an
aggregated state. Differences in stability of aggregates of polio and reo-
viruses and effects of dilution, ionic strength, and pH on aggregation are also
shown.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Viruses
Agglomeration
Water treatment
Disinfection
Halogens
Virus inactivation
13 B
6 M
13. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
59
20. SECURITY CLASS (Thispage)
Tin pi BHRI f i pri
22. PRICE
EPA Form 2220-1 (9-73)
49
if U. S. GOVERNMENT PRINTING OFFICE 1977-757-056/556') Region No. 5-11
------- |