Volume 5 Number 10 October 1978
Nucleic Acids Research
Semi-conservative transcription in particles of a double-stranded RNA mycovirus
G.Ratti and K.W.Buck
Department of Biochemistry, Imperial College of Science and Technology, London SW7 2AZ, UK
Received 24 July 1978
ABSTRACT
During transcription in vitro catalysed by the virion RNA polymerase
of Aspergillus foetidus virus AfV-S in the presence of tritiated UTP, the
virus double-stranded RNA becomes labelled in one strand, which has the
same sequence as the single-stranded RNA transcripts produced. Most of
the label incorporated into double-stranded RNA could be chased into singlestranded RNA by further reaction with excess unlabelled nucleoside triphosphates. In reactions containing tritiated UTP the single-stranded RNA
transcripts released after the first round of transcription were unlabelled.
It is concluded that transcription in virions of AfV-S occurs by displacement of one of the strands of double-stranded RNA by the RNA strand being
newly synthesised i.e. the reaction is semi-conservative with respect to
double-stranded RNA.
INTRODJCTION
Aspergillus foetidus virus AfV-S is a small isometric virus with a
genome of three dsRNA1 components2, dsRNA 1 (mol.wt. 4.oxIO6), dsNA 2
(mol.wt. 2.6x1o6) and dsRNA 3 (mol.wt. 0.26x106). ds RNAs i and 2 are each
encapsidated separately in l2a and Sla particles respectively, whereas
dsRNA 3 is found only together with either dsRNA 1 or dsRNA 2 in S2b or
Sib particles3.
polypeptides.
The capsids of all particles are constructed from the same
17
Recently it has been shown that the virion RNA polymerase of
AfV-S is a transcriptase 4 . Most of the RNA polymerase activity in vitro
is found in Sla particles, and the major products of transcription, which
are released from the particles, are full length ssRNA1 copies of one of
the strands of dsRNA 2. Re-initiation of transcription occurs and, during
the course of a 48 h reaction, 6 to 8 ssRNA transcripts are produced, on
average, per molecule of dsRNA 2.
In reactions containing 3H-UTP, label is incorporated also into dsRNA 2
within virions, reaching a maximum level after 4 h. Most of this activity
is also associated with Sla particles. In the present paper evidence is
presented that labelling of dsRNA 2 occurs as part of the transcription
reaction. It is shown that transcription in virions of AfV-S occurs by a
C) Information Retrieval Limited 1 Falconberg Court London WI V 5FG England
3843
Nucleic Acids Research
mechanism involving repeated displacement of
the RNA strand being newly
synthesised,
so
of the strands of dsRNA by
one
that at the end of each round of
transcription, the displaced strand is released from its template and from
the virion, and the dsRNA remaining within the virion contains
and
one
conserved
newly synthesised strand i.e. the reaction is semi-conservative with
In this respect therefore AfV-S RNA polymerase activity
respect to dsRNA.
proves
one
to be different from the other virion associated, dsRNA dependent,
RNA polymerases until
described.
now
METHODS
Growth of the fungus, preparation and purification of virus, RNA
polymerase reactions, preparation of virus RNA, separation of ssRNA and
dsRNA, ssRNA/dsRNA hybridisations and analysis of RNA by gel electrophoresis
were
carried out
as
described previously
Analytical ultracentrifugation.
Equilibrium density gradient centri-
fugation of dsRNA in caesium sulphate solutions
man,Model E ultracentrifuge equipped with
a
was
carried out in
monochromator and
ultraviolet absorption optical system, with photoelectric
plexer
Samples
accessory.
were
a
a
Beck-
double-beam
scanner
and multi-
placed in cells with double-sector charcoal
filled Epon centrepieces and -1° wedge top windows in the 4 place, AN-F
rotor, and centrifuged at 34,000 rev/min for 70 h at 250C.
Densities of
caesium sulphate solutions were calculated from refractive indices5 and
buoyant densities of RNA
were
calculated
as
described by Szybalski6.
RESULTS
Incorporation of 3H-UMP into one strand of dsRNA 2
It has been shown previously4 that during the AfV-S transcription
reaction 3H-UMP is incorporated into dsRNA 2 within Sla particles the amount
varying with different virus preparations and corresponding
the synthesis of 40 to 80 % of one RNA strand.
one
molecule of dsRNA 2 and
no
on average to
Since Sla particles contain
detectable ssRNA this amount of 3H-UMP
incorporation cannot be accounted for either by synthesis of dsRNA on a
ssRNA template,
with nascent subvirions of reovirus7, or by "filling in"
as
of short ssRNA tails
with the virion RNA
on
predominantly dsRNA molecules,
polymerase8
of bacteriophage
46.
as has been found
There remain two
other possibilities: (a) transcription could occur by semi-conservative
strand displacement of dsRNA, so that the newly synthesised strand becomes
part of the duplex,
phage
3844
@6 dsRNA;
as is
thought to
occur9
in vivo in the replication of
(b) complete replication of dsRNA could occur in some
Nucleic Acids Research
particles, as has been found with the virion RNA polymerase of a Penicillium
virus10.
In (a) only one strand of the dsRNA duplex should become labelled,
whereas in (b) both dsRNA strands should become labelled. Hence the two
possibilities can be distinguished by competition hybridisation of denatured
3H-labelled dsRNA 2 with an excess of unlabelled ssRNA 2 transcript.
Theoretically if both dsRNA strands are labelled, 50% of the label should
be displaced, whereas if only one strand is labelled either 100% or 0%
of the label should be displaced depending on whether the polarity of the
labelled strand is the same or opposite to that of ssRNA 2. Accordingly
total RNA was isolated from an i8 h RNA polyimerase reaction with Sla
particles in the presence of 3H-UTP, and labelled dsRNA 2 was separated
precipitation. Heat denatured 3H-labelled
dsRNA 2 was then annealed with increasing amounts of unlabelled ssRNA 2
and the amount of 3H-dsRNA resulting was determined as the amount of RNA
resistant to the action of RNAases1 A and TI at high salt concentrations
from ssRNA 2 by repeated LiC
(buffer A)1, as described previously4k1l. The results given in Table 1
show that 97% of the label can be displaced from dsRNA 2 by its ssRNA 2
transcripts. It is clear that during transcription dsRNA 2 is labelled
in only one strand which has the same sequence as that of ssRNA 2.
Formation of hybrid density labelled dsRNA 2
The tritium labelling experiments show that during AfV-S transcription
dsRNA 2 becomes labelled to the extent, on average, of 40 to 80 % in one
strand, but, do not distinguish whether some molecules are completely labelled and others are unlabelled, or whether all molecules become partially
TABIE
I
Competition Hybridisation
RNAase resistant,
Ratio
ssRNA/dsRNA
Sample
(w/w)
4000
3700
3H-dsRNA 2*
3H-dsRNA 2**
3H-dsRNA 2 +
unlabelled ssRNA 2
% RNAase
insolublereitn
resistant
TCA'
label (cts/min)
100
92.5
31
400
10.0
62
119
3.0
132
3.3
3H-dsRNA 2 +
unlabelled ssRNA 2
3H-dsRNA 2
+
97***
unlabelled ssRNA 2
* unheated blank value
** denatured and annealed
a different batch of ssRNA 2 was used
*
3845
Nucleic Acids Research
labelled.
a
Evidence for the complete displacement of
proportion of the molecules
ments.
An RNA polymerase reaction with AfV-S
dsRNA 2 strand in
was
carried out with 5-bromo-
Transcription occurred with about the
UTP in place of UTP.
as
was
one
obtained from density labelling experi-
with UTP and after 18 h ssRNA 2 labelled with 5-bromo-UMP
same
was
efficiency
formed in
molar amount equal to about 3 times that of its dsRNA template i.e.
re-
initiation of transcription also occurred readily with the brominated
substrate.
DsRNA
was
isolated from the reaction mixture and centrifuged
caesium sulphate density gradient in the analytical
to equilibrium in
a
ultracentrifuge.
The
u.v. scan,
shown in Fig. lc, shows that three main
components, all shown to be dsRNA by their resistance to RNAases A and TI
in buffer A,
were
resolved.
Scans of unreacted dsRNAs
I
and 2 (Fig. lb)
and dsRNA 2 alone (Fig. la) showed that the two lighter bands correspond
602
1615
DENSITY (g/ml)
Fig.l. Equilibrium density gradient centrifugation of dsRNA in caesium
sulphate. Scans at 265 rnm obtained after centrifuging for 70 h at 34,000
rev/min. Approximately 2 pg of RNA were used in each experiment.
(a) unreacted dsRNA 2 from Sla particles; (b) unreacted dsRNA 2 + dsRNA I
from purified, unfractionated AfV-S; (c) dsRNA obtained from a 24 h AfV-S
polymerase reaction mixture in which 5-bromo-UTP was used in place of UTP.
3846
Nucleic Acids Research
in buoyant density to dsRNA 2 (47% G+C) and dsRNA 1 (60% G+C) respectively;
dsRNA 3 which comprised less than 2% of the total RNA was not detected.
It is also clear that after reaction the amount and density of dsRNA J was
unchanged, whereas the amount of unlabelled dsRNA 2 had decreased by about
50% and a corresponding amount of a denser band of bromo-UMP labelled dsRNA
2 has been formed. The 2.5% increase in density of dsRNA 2 is consistent
with the expected value based on the increased molecular weight of a hybrid
dsRNA containing one brominated and one unbrominated strand, (U:BrU) RNA,
or in comparison with the observed 2.7% increase in density on formation
of hydrid (T:BrU) DNA, after one round of semi-conservative replication of
Escherichia coli DNA in the presence of 5-bromouracil12. It is concluded
that in this experiment complete strand displacement occurred in about 50%
of dsRNA 2 molecules.
Pulse-chase experiments
The above experiments establish that, during the AfV-S RNA polymerase
reaction, one RNA strand in 40 to 80 % of dsRNA 2 molecules is replaced by
a newly synthesised strand, but do not show that this reaction occurs more
than once.
In order to determine whether the label incorporated into dsRNA
2 was stable or was continually turning over in repeated rounds of transcr-
iption, two experiments were carried out. In the first, an AfV-S preparation was incubated in an RNA polymerase reaction mixture containing 0.1 mM
ATP, CTP,
GTP and
3H-UTP, and 3 mM
Mg2+
for I h and then the reaction
mixture was diluted into 3 volumes of a solution of unlabelled nucleoside
triphosphates such that the final concentrations were 3 mM UTP, 0.1 uM ATP,
CTP and GTP and 7 mM Mg2+, and incubation was continued (the chase). RNA
was prepared by phenol extraction of samples taken at different time intervals during the chase and 3H-RNA resistant and susceptible to RNAase A
digestion in buffer A was measured by using TCA precipitation. The results
given in Fig. 2, show that 80% of the 3H-UMP incorporated into ribonuclease
resistant RNA during the I h labelling period become ribonuclease sensitive
during the chase.
In the second experiment virus was incubated in a standard polymerase
reaction mixture containing all unlabelled nucleoside triphosphates for 4 h
and then 3H-UTP was added. After further incubation for ao mins. (the
pulse), the reaction mixture was diluted as above and incubation was continued (the chase). RNAase resistant and sensitive RNA were measured in
samples taken at the end of the pulse and after 3.5 h chase as before
(Table 2). It has been shown previously4 that synthesis of ssRNA 2 contin3847
Nucleic Acids Research
c
z
F-
s
50
LU
z
0
5
10
15
25
20
TIME(h)
Fig. 2. Turn-over of label in dsRNA. A reaction mixture containing 3H-UTP
(spec. act. 714 mCi/rmole) was incubated for I h at 300C and then the specific activity of 3H-UTP was decreased 12D fold by adding UTP (see text).
The high specific activity label incorporated into dsRNA during the first
phase of the reaction was then monitored during the second phase (chase) by
taking samples at various intervals and stopping the reaction with EDTA.
After phenol extraction RNA was dissolved in buffer A and the TCA insoluble
label present after RNAase digestion was determined. This is expressed as
the percentage of the RNAase resistant label (32,000 cts/min) present at
the beginning of the chase (4 C ir the graph).
ues
for
up
to 48 h whereas labelling of dsRNA 2 is maximal after
ca.
4 h.
The results in Table 2 show that (a) incorporation of UMP into dsRNA 2
was
still occurring after 4 h of reaction and (b) this incorporation could be
displaced by further reaction.
Further insight
was
obtained by examination of RNA prepared from
samples at the end of the pulse and after 3.5 h chase by electrophoresis
in polyacrylamide gels containing 8
urea.
M
The results showed that at the
end of the pulse 3H-labelled RNA was found in two bands, one at the top of
the gel and one with the same mobility
between the two bands (Fig. 3);
no
as
dsRNA 2, and in
peak of 3H-RNA
was
tion of ssRNA 2, although toluidine blue staining of
parallel showed that
an
a
a
small proportion
found in the posi-
similar gel
run
in
appreciable quantity of unlabelled ssRNA 2 had been
synthesised, probably mainly during the 4 h of the reaction prior to the
pulse.
If the RNA isolated at the end of the pulse was incubated with
RNAase A in buffer A prior to electrophoresis, the band at the top of the
gel
was no
longer observed, but the 3H-incorporation
had increased from 18,O0
3848
counts/min
to
into the dsRNA 2 band
34,000 counts/min (Fig. 3b).
This
Nucleic Acids Research
TABLE
2
Pulse-chase experiment
Sample
RNAase resistant,
TCA insoluble label
RNAase sensitive
label
(cts/min)
(cts/min)
(a)
RNA synthesised
between 4 h and
4 h 20 min (pulse)
o,o000
(b)
After 3.5 h chase
5,000
12,800
27,800
RNA was prepared by phenol extraction from equivalent amounts of reaction
mixture: (a) at the end of a 3H-UTP pulse (spec. act. 1760 cts min-1
pmole1), between 4 h and 4 h 20 min from the start of the polymerase
reaction; (b) after a further 3.5 h of incubation in a reaction mixture
in which the specific activity of 3H-UTP had been reduced 1D times. The
concentration of dsRNA in each sample was adjusted to be the same and 25 pl
(containing 350 ng of dsRNA 2) per assay was used for determination of
ribonuclease resistant and sensitive label. The composition of these RNA
samples was also analysed by gel electrophoresis (Fig. 3).
shows that all the incorporation of 3H-UTP into RNAase resistant RNA was
into structures containing dsRNA 2, presumably with ssRNA tails of varying
length.
After the chase a peak of RNAase sensitive 3H-RNA with tbe mobility
of ssRNA 2 appeared (Fig. 3c) and the RNAase resistant 3H-RNA with the
mobility of dsRNA 2 had decreased from 34,20 to 8,800 counts/min (Fig. 3d)
i.e. about 75% of the 3H-incorporation into dsRNA 2 during the pulse was
chased into high molecular weight ssRNA, in good agreement with the results
obtained by direct TCA precipitation.
Displacement of one strand from dsRNA 2
A critical test of the semi-conservative transcription mechanism is to
demonstrate the displacement and release of one strand from parental virion
dsRNA. If an RNA polymerase reaction, in which transcription occurs semiconservatively, is carried out in the presence of 3H-UTP it would be
expected that the ssRNA liberated at the end of the first round of transcription would be unlabelled since it is derived from the unlabelled dsRNA
template; subsequent rounds would then release 3H-labelled ssRNA by displacement of the newly synthesised, labelled strand from the dsRNA duplex.
In contrast the dsRNA template should become labelled in the first round of
transcription.
In order ;to test these
predictions,
a
preparation of Sla particles,
which contain only dsRNA 2 and no detectable ssRNA, was incubated in a
standard RNA polymerase reaction containing 3H-UTP and at different time
3849
Nucleic Acids Research
x
c
E
U14
z
Sl i ce
number
Fig. 3. Pulse-chase experiment: analysis of RNA by electrophoresis in
polyacrylamide gels containing 8 M urea. 0.05 ml of RNA (same samples as
in Table 2) were used for each experiment. After electrophoresis gels were
sliced and counted. The positions of ssRNA 2 and dsRNA 2 were determined
by scanning gels at 260 nm. (a) pulse sample; (b) pulse sample after
RNAase treatment; (c) 3.5 h chase sample; (d) 3.5 h chase sample, after
RNAase treatment.
intervals, samples were taken and ssRNA 2 released into the supernatant was
examined by polyacrylamide gel electrophoresis. Gels were scanned at 260nm
and then sliced and counted for 3H-UMP incorporated. Incorporation of 3H3850
Nucleic Acids Research
UMP into dsRNA, after isolation from virions,
examined in
was
The results (Fig. 4) showed that labelling of dsRNA 2
medium in
an
at 260 nm showed that ssRNA 2
scans
was
released into the
amount corresponding to the displacement of
20% of dsRNA 2 molecules.
After this time
label
no
way.
Moreover after 30 min
of ssRNA 2 during the early part of the reaction.
reaction, gel
similar
a
greater than that
was
was
band of ssRNA 2 whereas the band of dsRNA 2 contained
one
strand from
detected in the
more
than 20,000 cts/
min.
DISCUSSION
The results show that transcription in virions of AfV-S occurs'by
mechanism involving repeated displacement of
by the RNA strand being newly synthesised,
of the strands of dsRNA 2
one
so
a
that at the end of each round
of transcription the displaced strand is released from its template and
from the virion, and the dsRNA 2 remaining within the virion contains
conserved and
one
newly synthesised strand i.e. the reaction is semi-
conservative with respect to dsRNA.
A diagrammatic representation of the
reaction for two rounds of transcription is shown in Fig. 5.
may
one
This mechanism
be contrasted with that of reovirus,' in which transcription
incorporation of labelled nucleoside triphosphate
con,servatively, with
no
residues into
dsRNA13,14.
genome
occurs
The evidence for semi-conservative trans-
cription in AfV-S is summarised below.
(a)
During the polymerase reaction, in the
2 becomes labelled in
one
strand, which has the
presence
of 3H-UTP, dsRNA
same sequence as
the ssRNA
transcription product (ssRNA 2).
(b)
If 3H-UTP is present from the beginning of the reaction, labelling
However it the reaction
is allowed to proceed for 4 h with unlabelled UTP and then 3H-UTP'is intro-
of dsRNA 2 reaches
a
maximum after 4 h reaction4.
duced, dsRNA 2 becomes labelled in the subsequent reaction and the rate of
this labelling is consistent with the rate of ssRNA 2 synthesis after this
period of reaction (Table 2, Fig. 3).
(c)
About 80% of the label incorporated into dsRNA 2 can be chased
out by subsequent reaction with unlabelled substrates (Fig. 2 and Table 2),
resulting in the release of labelled ssRNA 2 (Fig. 3).
out 100% of the label from dsRNA 2 is probably
deceleration of the initial reaction rate.
Failure to chase
a result of the
rapid
It has been observed
previously4
that the rate of reaction after 6 h is only about 5% of the initial rate.
This may be due partly to failure of some intact particles to re-initiate
3851
Nucleic Acids Research
50
40
30
C
E
u 20
-
10
30
60
120
180
240
TIME(min)
Fig. 4. Incorporation of 3H-UTP into complete ssRNA 2 transcripts and dsRNA
2 template by Sla particles. Sla particles, isolated from purified AfV-S
by two cycles of sucrose gradient centrifugation, were incubated in standard
reaction conditions4 containing 3H-UTP (spec. act. 180 cts min-1 pmole-1)
at a final virus concentration of 0.08 mg/ml. At the times indicated 0.1
ml samples were taken and the reaction was stopped with EDTA. The amounts
of label in the ssRNA 2 released into the medium and in the dsRNA 2 remaining within the virus particles were analysed by polyacrylamide gel electrophoresis and plotted against time. The ordinate values correspond to 0.075
ml of original reaction mixture. Open circles, ssRNA 2; closed circles,
dsRNA 2.
the reaction and partly to the disruption of some particles which occurs
during the reaction 4
(d) The major products of a 20 min labelling period have the
properties of transcriptive intermediates, consisting of dsRNA 2 with long
ssRNA tails, of the type shown in Fig. 5, since in gel electrophoresis they
moved as a slow band near the top of the gel, but after trimming off the
ssRNA tails with RNAase, under conditions where ssRNA is degraded but high
molecular weight dsRNA is stable1, moved as a band with the same mobility
as dsRNA 2 (Fig. 3). In contrast intermediates of this type have not been
isolated from a reovirus transcription reaction, which occurs conservatively;
in that case the major products after pulse labelling were incomplete, ssRNA
transcripts which di&6,not remain bound to the dsRNA template after phenol
extractioni5.
(e) Comparison of the time course of labelling of dsRNA 2 and ssRNA 2
(Fig. 4) with the production of ssRNA 2, measured from u.v. scans of gels
3852
Nucleic Acids Research
_
I\
/
Fig. 5. Diagram of the proposed mechanism of transcription of AfV-S dsRNA s
catalysed , within the virus particle, by the virion associated RNA polymerase . The broken lines represent RNA synthesi sed during the in vitro
reaction and do not imply any fragmentation of the product. (F7o-ra reaction
carried out in the presence of 3H-UTP, the broken line would represent labelled RNA and the unbroken line, unlabelled RNA). Presumably transcription
starts at the 5' end of the displaced strand. Re-initiation could easily
occur if the ends of the dsRNA template are in close proximity within the
virion e.g. if the RNA is arranged in a circular conformation.
after electrophoresis, shows tha-t labelling of dsRNA 2 was greater than
that of ssRNA 2 during the early part of the reaction and that ssRNA 2
released after 30 min. reaction was unlabelled.
This is consistent with
semi-conservative mechanism in which unlabelled transcripts would be
released at the end of the first round of transcription and the dsRNA
a
template would be labelled (Fig. 5).
From the proportion of hybrid (U :BrU) dsRNA 2 molecules produced in
the density labelling experiments (Fig. 1) and the amount of 3H-UMP incorporation into dsRNA 2 after 4 h (F'ig. 4) it is clear that in these experiments initially 40 to 50 % of Sla particles were involved in semiconservative transcription. The amount of ssRNA 2 released after 30 min
reaction corresponds to displacement of one strand from dsRNA 2 in 50% of
the active Sla particles. The reaction was not completely synchronous and
maximum labelling of dsRNA 2 took 4 h. It is probable that eithe;r initia3853
Nucleic Acids Research
tion of transcription
was
asynchronous,
varied in different virions
as
or
that the rate of transcription
suggested for reovirus
cores
The amount of ssRNA 2 released after 30 min reaction
on
was
calculated,
the basis of its content of 25% U and the specific activity of the 3H-
UTP used in the experiment (see legend to Fig. 4), to be equivalent to
17,000 counts/min if it were fully labelled with UTP as in a conservative
Since no label was detected in the band of ssRNA 2 after this
mechanism.
time and allowing for
a
sensitivity of detection of
the background in gel slices), it
transcription, if
any,
can
800
counts/min (double
be deduced that less than 5% of the
could have occurred by
a
conservative mechanism.
ACKNOWLEIDGEMENT
The work
was
supported by
a
grant from the Science Research Council
(to K.W.B.).
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
3854
dsRNA, double-stranded RNA; ssRNA, single-stranded RNA;
RNAase, ribonuclease; TCA, trichloroacetic acid; buffer A, 0.3 M-NaCl
+ 0.03 M-Na citrate, pH 7.0.
Buck, K.W. & Ratti, G. (1977). J. Gen. Virol. 37, 215-219.
Buck, K.W. & Ratti, G. (1975). J. Gen. Virol. 27, 211-224.
Ratti, G. & Buck, K.W. (1978). J. Gen. Virol. in press.
Brakke, M. (1967). In Methods in Virology (Maramarosch, K. & Koprowski,
H., Eds), Vol. 2, pp 93-97, Academic Press, London and New York.
Szybalski, W. (1968). In Methods in Enzymology (Grossman, L. and
Moldave, K., eds), Vol. 12B, pp 330-360, Academic Press, London and
New York.
Zweerink, H.J., Ito, Y. & Matsuhisa, T. (1972). Virology 50, 349-358.
Van Etten, J.L., Vidaver, A.K., Koski, R.K. & Semancik, J.S. (1973).
J. Virol. 12, 464-471.
Coplin, D.L., Van Etten, J.L., Koski, R.K. & Vidaver, A.K. (1975).
Proc. Natl. Acad. Sci. (USA) 72, 849-853.
Buck, K.W. (1975). Nucleic Acids Res. 2, 1889-1902.
Bellamy, A.R., Shapiro, L., August, J.T. & Joklik, W.K. (i967). J. Mol.
Biol. 29, 1-17.
Pritchard, R.H. & Lark, K.G. (1964). J. Mol. Biol. 9, 288-307.
Skehel, J.J. & Joklik, W.K. (1969). Virology 39, 822-831.
Gomatos, P.J. (1970). J. Virol. 6, 61o-62o.
Banerjee, A.K. & Shatkin, A.J. (1970). J. Virol. 6, 1-11.
Bellamy, A.R. & Harvey, J.D. (1976). Virology 70, 28-36.
Ratti, G. & Buck, K.W. (1975). Biochem. Biophys. Res. Commun. 66,
706-711.
Abbreviations: