United States
Environmental Protection
Agency
Municipal Environmental Research
Laboratory ,
Cincinnati OH 45268 '
Research and Development
EPA-600/S2-82-022 August 1982
Project Summary
Virion Aggregation and
Disinfection of Water by
Chlorine and Bromine
D. Gordon Sharp
The state of aggregation among
virions suspended in water was mea-
sured and quantitatively characterized
by several methods described in the
report. Some of them provided physi-
cal assay (virion count) along with
aggregate counts by quantitative elec-
tron microscopy; others revealed, by
centrifugation, the fraction of single
virions and also the rate of change of
this fraction (physical stability of the
suspension).
All unpurified virus suspensions
contained some aggregates and all of
them contained 50 percent or more
singles. Stable purified suspensions
of single virions without detectable
aggregation were prepared and used
for comparison of inactivation kinet-
ics of several viruses under a variety of
conditions.
All of the viruses tested tended to
aggregate at acid pH, but the pH
below which aggregation began was
quite different for different viruses.
All the viruses tended to aggregate at
low ionic strength, but the kind of salt
or buffer present strongly influenced
the rate of aggregation at a given pH
and temperature. The rates of
aggregation at low pH were found to
be in good agreement with the von
Smoluchowski rapid coagulation
theory.
Virion aggregation caused depar-
tures from first order inactivation
kinetics. However, monodisperse sus-
pensions of certain viruses also
showed similar departures apparently
related to inactivation mechanisms.
Reversal of inactivation of mono-
disperse inactivated echovirus by
complementation also was demon-
strated.
Inactivation of poliovirus by hypo-
chlorite (OCH at pH 10 was very
weak compared with hypochlorous
acid (HOC!) at pH 6. Increasing sensi-
tivity of the virus to HOCI as pH was
increased to 7, 8, and 9 appears to be
duetoa change in the virus rather than
to OCr. However, addition of O.1M
sodium chloride (NaCI) at pH 10 in-
creased the inactivation rate more
than 100-fold. Chlorides of cesium
and potassium also enhanced the rate,
making it comparable with that of
HOCI at pH 6.
This Project Summary was devel-
oped by EPA's Municipal Environ-
mental Research Laboratory.
Cincinnati, OH, tr. announce key find-
ings of the research project that is
fully documented in a separate report
of the same title (see Project Report
ordering information at back).
Introduction
Treatment of water with free chlorine
is one of the most widely used and most
effective methods known for inactiva-
tion of both bacteria and viruses. Even
in micromolar concentrations, the de-
struction of viruses is usually very rapid,
but increased resistance has been en-
countered when the virus particles
(virions) have been aggregated together
in clumps. This can happen in the
-------
laboratory where high concentrations
of purified virus tend to coalesce, or it
may be encountered in the field where
virus particles have entered the water
supply in fecal material containing in-
fected cells in which there may be
thousands of closely packed virions.
One of the reasons for the increased
halogen resistance of clumped viruses
is readily understandable in terms of
protection of one or a few virions inside
clumps of various sizes that decrease to
different degrees the accessibility of the
halogen molecules to the protected
virions. This leads to a disinfection rate
that does not remain constant but de-
creases with time as increasingly re-
sistant clumps are the survivors at each
step of the process. Evidence for this is
presented herein as well as some evi-
dence for a second means of survival of
infectivity through clumping of the vir-
ions. A suspension of single echovirus
particles, which has been reduced by
chlorine in infectivity for tissue culture
cells by a factor of about 100,000, has
been reactivated subsequently 1000-
fold or more simply by inducing the
virions to aggregate into small clumps. It
is not yet clear to what extent either or
both of these survival mechanisms is
operative in the halogen treatment of
large volumes of drinking water, but it is
abundantly clear that further laboratory
work of the kind reported will be neces-
sary before adequate understanding is
achieved and before experimental work
can be effectively carried to the more
complex conditions that will doubtless
be encountered in actual practice.
Aggregation among virions and its
effect on the disinfection process has
been the subject of speculation in the
discussion section of many published
papers, but little physical evidence has
ever been presented to accompany it.
The research reported herein is
involved with both the physical and the
biological aspects of the problem.
Quantitative measurement and charac-
terization of the state of aggregation of a
suspension of virus particles has been
accomplished and described along with
methods for measuring the degree of
colloidal stability of suspension. Both
the ultracentrifuge and the electron
microscope (EM) have been employed
in this.
When the first EM analyses of aggre-
gation were obtained, they revealed a
problem that has remained unsolved.
They showed that we were unable to
produce two preparations of virus with
the same degree of aggregation. That is,
no two preparations had the same frac-
tion of single particles, pairs, triplets,
and other groupings, and furthermore,
the observed degree of aggregation
appeared to be subject to slow or some-
times rapid change, depending on some
readily recognizable and doubtless
other unknown parameters. This meant
that it would be exceedingly difficult if
not impossible to repeat experiments
with different disinfectants on viruses
with a given degree of aggregation or to
compare the effectiveness of a given
disinfectant on several viruses all with
the same degree of aggregation. Condi-
tions were revealed, however, that
were favorable to the preparation and
maintenance of several viruses in the
monodisperse (all singles) state. This
made it possible, for the first time, to
compare the kinetics of inactivation of
several viruses under a wide variety of
conditions under which there could be
no influence of virion aggregation.
Materials and Methods
Preparation and Maintenance
of Monodisperse Virus
Suspensions
Poliovirus, Mahoney and Brunhilde,
echovirus, Farouk, and Coxsackie B3
were prepared in multiply infected HEp-
2 cells and harvested before substantial
lysis had occurred. This allowed the re-
covery of most of the virus by simply sedi-
menting the infected cells, extracting
them at higher shear with freon, and
separating single virions from soluble
nitrogenous foreign material, ribo-
somes, and aggregated virus by velocity
banding. The velocity banding was
accomplished using a sucrose density
gradient made with phosphate or tris
buffered saline at pH 7.0 to 7.2. This
rendered the preparations sufficiently
free of chlorine or bromine demand so
that inactivation experiments could be
made without significant loss of either
free halogen at concentrations as low
as 2 fjM. These preparations, contain-
ing roughly 20 percent sucrose, could
be refrigerated for several months
unfrozen without either troublesome
growth of bacteria or serious aggrega-
tion and little loss of infectivity. Atypical
preparation of poliovirus (Mahoney)
containing 1.3 x 1012 virions per ml is
shown in Figure 1. The picture was
taken 2 months after the virus was puri-
fied. There is no evidence of aggrega-
tion here and there is no other test that
is more sensitive. In particular, no cen-
trifuge technique is capable of detecting
less that 1 percent of aggregation in a
virion suspension.
Reovirus was prepared from L cells in
essentially the same manner, but it was
difficult to keep it for long periods with-
out some aggregation.
Measurement and Characteri-
zation of Virion Aggregation
Quantitative electron microscopy
was used to assess the number and size
of virions and at the same time to count
the number of aggregates of each size
present. Figure-2 is an EM of slightly
aggregated reovirus. Counts from sev-
eral pictures of this kind can be used to
produce a frequency charge completely
characterizing the state of aggregation.
Two methods have been used in the
presentation of virus suspensions of the
types shown in Figures 1 and 2, and
they must be strictly controlled lest
aggregation be inadvertently produced
in the process. Both methods involve
attachment of the virions to collodion
films before any drying to remove sus-
pending liquid. Large viruses such as
reovirus were sedimented by centrifu-
gation from dilute suspensions onto a
suitable collecting surface, but this
proved unsatisfactory for the smaller
picorna viruses. These were gathered
on aluminized collodion films to which
they adhered during Brownian bom-
bardment from dilute suspensions
containing sufficient salt at neutral pH.
In this case, thorough washing was
required to ensure that no unattached
virions were present when drying was
done. These methods produce the most
complete characterization of the collod-
ion state of a virus suspension, but they
can be difficult and time consuming.
They require a high virus concentration,
and they give little useful information in
cases of heavy aggregation. This situa-
tion can often be handled best by cen-
trifugation.
If a quick comparison is needed of the
state of aggregation of several virus
preparations, the virion aggregation
test (VAT) has proved useful. This
involves velocity banding of the plaque
forming units (PFU) in a two-step
sucrose gradient. Plaque titration of the
virus found at the end of a centrifuge
run in the light, medium, and dense
regions of the tubes gives a set of three
values for comparison.
A more precise variation of the VAT
has been called the single particle anal-
ysis (SPA), which does not use a density
gradient. The virus suspension fills the
-------
Figure 1. Poliavirus (Mahoney) purified
for halogen inactivation
experiments. A stock suspen-
sion containing 1.3 x 7012
virions per ml in 20% sucrose
was diluted 10X with phos-
phate buffered saline for
this KA picture.
whole centrifuge tube at the start, and
after a predetermined centrifuge treat-
ment (speed, time, and temperature)
and very careful stop, the top half of the
tube's contents is removed, mixed, and
titrated. If an unaggregated sample of
the virus is available, it can be used as a
control for several others carried in the
same rotor at the same time. By stop-
ping the rotor when only single virions
remain above the arbitrary sampling
level, the fraction of single virions in the
suspension can be calculated from the
ratio of PFU remaining in supernatant of
the test sample or samples to the PFU in
the unaggregated control. In the
absence of an unaggregated control,
the expected fraction of singles in the
supernatant region of a completely
dispersed sample can be calculated if
virion size, shape, and density are
known or, better yet, if the sedimenta-
tion constant is known.
The SPA test, when done in the six
tube swinging bucket rotor (SW 50.1,
Beckman*) has been particularly useful
for obtaining five states of aggregation
that have occurred at five different time
periods after subjecting a virus to a
*Mention of trade names or commercial products
does not constitute endorsement or
recommendation for use
Figure 2. Reovirus, slightly aggregated,
prepared for physical assay by
the same KA method used for
the poliovirus of Figure 1.
given salt concentration, pH, or other
factor that influences aggregation
suspension stability data have been
obtained in terms of observed rate of
change in the remaining fraction of
single virions. One more use of the
fraction of singles is for direct
comparison with rapid coagulation
theory developed for colloids in general
by von Smoluchowski 60 years ago.
The Kinetic Experiment for
Virus Inactivation
Virion aggregation measurements
have shown that most suspensions of
virus contain clumps of many sizes.
Inactivation of such a complex mixture
of PFU must involve a reaction rate that
changes with time; therefore, it is
important that all phases of the kinetic
process be examined, not the least of
which is the initial velocity, which may
be great if most of the PFU are single
virions.
An apparatus has been constructed
for this work (Figure 3) with which virus
exposure to free chlorine or bromine for
intervals as short as V* second can be
readily controlled. Several previously
unknown transient changes in
inactivation kinetics were found in this
way.
Results
Inactivation of
Aggregated Virus
Initial demonstration experiments
showed that reovirus inactivation by
bromine began rapidly then became
much slower, even though only small
aggregates were present. Removal of
the aggregates produced a continued
rapid rate of inactivation. Essentially the
same behavior was demonstrated for
poliovirus inactivation by chlorine.
These data included an accurate
characterization of the aggregation that
existed among the virions in each
experiment. There was no doubt that
the observed effects were due to aggre-
gation, but it soon became apparent that
a series of controlled experiments on
aggregated virus could not be done be-
cause repeated preparation of virus
with the same degree of aggregation
was not possible, and no preparation of
virus remained for very long in the same
state of aggregation. One state of col-
loidal suspension did, unexpectedly,
prove to be quite stable and repro-
ducible. If pH, ionic strength, and salts
were all correct, the poliovirus
(Mahoney) would remain monodisperse
for long periods of time; therefore, many
experiments were performed under
these optimum conditions where free-
dom from aggregation could be demon-
strated.
Inactivation of Monodisperse
Polio (Mahoney) and
Reoviruses
At an early stage in the work, it was
recognized that whereas purified vir-
uses tended to aggregate at a pH less
than 7, the addition of a small quantity
of salt would often prevent it. At pH 6,
where free chlorine is essentially all
HOC), poliovirus aggregated slowly
unless 0.1 M NaCI was present in the
phosphate buffer, and 0.3 M NaCI would
prevent aggregation of this virus, even
at pH 5 during the experiments with
bromine (Br2). It was this that led us to
add some NaCI to the buffer in many
experiments to preserve uniformity in
the procedure, even though in many
cases it was not needed to prevent
aggregation.
Bromine inactivated poliovirus only a
little faster than the same molar concen-
trations of chlorine when comparing
hypobromous acid (HOBr) and HOCI but
hypobromite ion (OBr~) was much more
effective than either HOBr or OCI". Tri-
bromamine (NBr3)and HOBr were about
-------
equally effective, but dibromamine
(NHBrz) acted very slowly. Reovirus,
with which only a few initial experi-
ments were done, was about 16x more
sensitive to HOBrthan polio virus (Mah-
oney). One conspicuous fact that will
bear repetition here is the rapid inacti-
vation of poliovirus by OBr~ compared
with the much slower inactivation by
HOBr. With chlorine, this was reversed,
HOCI was much more effective than the
same molarity of free chlorine in the
form of OCf.
This report gives the results of many
inactivation experiments with monodis-
perse poliovirus using both bromine and
chlorine in several forms at different
temperatures and at different concen-
trations. Plotting log plaque survival
ratio versus time sometimes produced
a linear relationship, but more often it
did not. So, while we have shown above
that virion aggregation can produce a
sagging type of inactivation curve with
the rate decreasing with time, it appears
that just the same kind of curve often
occurs when no aggregation is present,
and the lagging curve (slow at first, then
faster) frequently occurs as well. These
results are highly reproducible, and it is
now clear that they are characteristic of
the action of these reagents with indi-
vidual virus particles and not a clumping
phenomenon. In the case of echovirus
(Farouk), a three-phase inactivation
curve was observed with HOCI at pH 6.
Plaque titer was reduced about 95 per-
cent in less than 1 second, and then it
remained constant for 20 seconds
before beginning a steady decline at a
much slower rate. Inasmuch as no
aggregation was present, it appears that
transient changes in the capsid proteins
occur.
Some Characteristics of Vir-
ion Aggregation
To examine the effects of chlorine on
viruses in water at low and high pH, one
must consider the different forms of free
chlorine that exist at different pH values
and possible changes in sensitivity of
the virus. Special attention must be
given to any pH-related changes in vir-
ion aggregation. Isoelectric points for
several viruses were determined here
that range from reovirus at pH 3.7 to
poliovirus (Mahoney) at pH 8.3. The poli-
ovirus shows no tendency to aggregate
at pH 8.3. The reovirus does aggregate
at pH 3.7, along with all the other viruses
tested (polioviruses, Mahoney and
Brunhilde, echovirus, and Coxsackie B3
and B5). Some of them aggregate at low
pH at the maximum rate calculated by
von Smolukowski. Data are presented in
Reynold's number
in flow tube
5300O
Down stream sample points
Tension springs
that pull the
sliding blocks
that hold I
Bromine
bottle
(20 liters)
Push release
Mix motor
cross section
Bar magnet
mixing bugs
Mix motor
lull view
Scale 10cm
Figure 3. Apparatus designed and built for this series of kinetic experiments on the
inactivation of viruses in water by bromine and chlorine.
Discharge
-------
the Project Report showing that single
virions disappear, through clump for-
mation, according to his theoretical pre-
diction:
N= No
1
+ t/t1A)
No = number of single particles at the
start when t = 0
N = number of single particles re-
maining at time t
tV4 = a constant, the time required
to reduce N to No/4
where IV* = 3>?
4N0KT
^rj - viscosity
K = Boltzmann's Constant
T = temperature
Under these conditions every collision
between virions results in adhesion
and, as the equation shows, high con-
centration promotes rapid coagulation.
Virus particle size has no effect, and
virions may remain dispersed for a very
long time if they are sufficiently dilute.
Apparently this kind of aggregation is
vanishingly small at virion concentrations
as low as those likely to exist in polluted
water, but it puts a maximum limit on
concentrations of purified virus that
can be maintained for experiment in the
laboratory. For example, if all collisions
in a monodisperse suspension of 2 x
1010 poliovirus per ml (10s PFU/ml)
resulted in adhesion only % of them
would remain single after 9 seconds
had elapsed. Conditions described ear-
lier for maintenance of viruses in
dispersed form must provide very little
opportunity for colliding virions to
adhere, but at some critical pH below 7,
all the viruses examined here begin to
aggregate at a rate that increases as pH
is reduced. Some buffers keep viruses
dispersed better than others, and sev-
eral salts have been tested and found
beneficial. The degree of effectiveness
varies greatly from one virus to another,
even among the few tested, and some
salts are shown here to exert a strong
effect upon the chlorine inactivation
rate of some viruses that is independent
of aggregation.
The Remarkable Effect of
Some Salts on the Chlorine
Inactivation of Some Viruses,
Particularly at High pH.
The use of the correct buffer and
some added salt at low pH, where slow
aggregation of poliovirus can occur, has
kept all the experiments free of
aggregation. Of course, the salt was
added, for uniformity, to all the buffers,
even those with a pH above 7, where it
was not needed. So' the belated
discovery was made that 0.1M NaCI,
when added to the phosphate buffer at
pH 6, doubled the HOC! inactivation rate
of poliovirus (Mahoney and Brunhilde)
but not of either Coxsackie B3 or B5. The
reaction rate of all viruses tested with
OCI~ at pH 10 was increased at least 30x
and that of poliovirus (Mahoney) over
100x when 0.1 M NaCI was added to
0.01 M buffers. Similar experiments
with cesium chloride (CsCI) showed it to
be less and potassium chloride (KCI)
more effective in this respect than the
sodium salt. Diva lent chlorides were not
tested except as mentioned above in
connection with virion aggregation.
Aggregation and Revival of
Chlorine-damaged Virus
Infectivity
When well dispersed echovirus was
reduced, several factors of ten in plaque
liter by exposure to HOCI and then
induced to aggregate by adjustment to
pH 4.5, the liter of the aggregated virus
was revived by factors varying from 10
to 3000. This virus was probably not
repaired and made whole again, but
when placed upon susceptible cell
monolayers in aggregated form, it was
10 to 3000 times more infectious than it
was in dispersed form. This effect was
not tested on an animal host.
Revival of Chlorine-damaged
Virus without Aggregation
When echovirus was treated with 20
A/M HOCI at pH 7, there was an
immediate loss of about 95 percent of
the PFU followed by a gradual increase
until most of the original infectivity was
regained. This recovery, however, was
only temporary since continued
chlorine treatment beyond 20 seconds
produced a steady decline in infectivity.
Exposure of this echovirus to ultraviolet
rays (2537 Au) produced a steady
decline in plaque titer with no transient
changes in rate, and the presence or
absence of 0.1 M NaCI was of no
consequence.
Stability of Virus Clumps in
Water
Purified poliovirus that has been
induced to form clumps by reducing pH
or salt content becomes dispersed again
when put into physiological saline or the
phosphate-buffered saline (PBS) of
Dulbecco. But, poliovirus released from
infected cells by several cycles of
freezing and thawing directly into PBS is
only partly dispersed. Purified
monodisperse virus tends to aggregate
when put into distilled water, but
monodisperse dilutions were made in
distilled water when the first step was
long enough, such as 1:7000. In this
way the virions became separated far
enough, quickly enough so that a state
of colloidal quasistability could persist
for a very long time. When this dilution
was made in smaller steps such as 1:10,
the purified virus aggregated in the first
step because of reduced ionic strength,
and all subsequent water dilution steps
just served to move the stable
aggregates further apart and produce
dilute but stable suspensions of
permanent aggregates.
Other Reports Based on This
Research
Essentially all of the significant results
of this research have been published
and are available as follows.
Floyd, R. 1979. Viral aggregation: mixed
suspensions of poliovirus and reovi-
rus. Appl. Env. Microbiol. 38:980-986.
Floyd, R., J.D. Johnson, and D.G. Sharp.
1978. Inactivation of single poliovirus
particles in water by hypobromite
ion, molecular bromine, dibromam-
ine and tribromamine. Env. Sci.
Technol. 12:1031-1035.
Floyd, R., and D.G. Sharp. 1978. Viral
aggregation: quantitation and kinet-
ics of the aggregation of poliovirus
and reovirus. Appl. Env. Microbiol.
35:1079-1083.
Floyd, R., and D.G. Sharp. 1978. Viral
aggregation: effects of salts on the
aggregation of poliovirus and reovi-
rus at low pH. Appl. Env. Microbiol.
35:1084-1094.
Floyd, R., and D.G. Sharp. 1979. Viral
aggregation: buffer effects in the
aggregation of poliovirus and reovi-
rus at low and high pH. Appl. Env.
Microbiol. 38:395-401.
Floyd, R., D.G. Sharp, and J.D. Johnson.
1979. Inactivation by chlorine of
single poliovirus particles in water.
Env. Sci. Technol. 13:438-442.
Jensen, H., K. Thomas, and D.G. Sharp.
1980. Inactivation of Coxsackievir-
uses B3 and B5 in water by chlorine.
Appl. Env. Microbiol. 40:633-640.
Sharp, D.G., and J. Leong. 1980. Inacti-
vation of poliovirus I. (Brunhilde) sin-
-------
gle particles by chlorine in water.
Appl. Env. Microbiol. 40:381 -385.
Sharp. D.G., D.C. Young, R. Floyd, and
J.D. Johnson. 1980. Effect of ionic
environment on the inactivation of
poliovirus in water by chlorine. Appl.
Env. Microbiol. 39:530-534.
Young, D.C., J.D. Johnson, and D.G.
Sharp. 1977. The complex reaction
kinetics of ECHO-1 virus with chlo-
rine in water. Proc. Soc. Exp. Biol.
Med. 156:496-499.
Young, D.C. and D.G. Sharp. 1979. Par-
tial reactivation of chlorine-treated
echovirus. Appl. Env. Microbiol.
37:766-773.
The full report was submitted in fulfill-
ment of Grant No. R-804587 by the
University of North Carolina under the
sponsorship of the U.S. Environmental
Protection Agency.
D. Gordon Sharp is with the University of North Carolina, Chapel Hill. NC 27514.
John C. Hoffis the EPA Project Officer (see below).
The complete report, entitled "Virion Aggregation and Disinfection of Water by
Chlorine and Bromine," (Order No. PB 82-230 889; Cost: $12.00, subject to
change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Municipal Environmental Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
*USGPO: 1982 — 559-092/0448
-------
United States Center for Environmental Research Fees Paid
Snyironmental Protection Information Environmental
Agency Cincinnati OH 45268 Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
RETURN POSTAGE GUARANTEED
Third-Class
Bulk Rate
MERL0063240
LOU M TILLEY
REGION V EPA
LIBRARIAN
230 S DEARBORN ST
CHICAGO IL 60604
------- |