Content uploaded by Dennis Drayna
Author content
All content in this area was uploaded by Dennis Drayna on Aug 25, 2015
Content may be subject to copyright.
Proc.
Natl.
Acad.
Sci.
USA
Vol.
74,
No.
12,
pp.
5744-5748,
December
1977
Microbiology
Molecular
basis
of
reovirus
virulence:
Role
of
the
SI
gene
(cell
tropism/neurovirulence/ependyma
and
neurons/reovirus
recombinants)
HOWARD
L.
WEINER*t,
DENNIS
DRAYNA*,
DAMON
R.
AVERILL,
JR.O,
AND
BERNARD
N.
FIELDS*t
*
Department
of
Microbiology
and
Molecular
Genetics,
Harvard
Medical
School,
Boston,
Massachusetts
02115;
t
Department
of
Medicine,
Sections
of
Neurology
and
Infectious
Diseases,
Peter
Bent
Brigham
Hospital,
Boston,
Massachusetts
02115;
and
f
Department
of
Pathology,
Harvard
Medical
School,
an(l
Department
of
Neuropathology,
Children's
Hospital
Medical
Center,
Boston,
Massachusetts
02115
Communicated
by
Robert
M.
Chanock,
September
20,
1977
ABSTRACT
A
genetic
approach
has
been
used
to
define
the
molecular
basis
for
the
different
patterns
of
virulence
and
central
nervous
system
cell
tropism
exhibited
by
reovirus
types
1
and
3.
Intracerebral
inoculation
of
reovirus
type 3
into
new-
born
mice
causes
a
necrotizing
encephalitis
(without
ependymal
damage)
that
is
uniformly
fatal.
Animals
inoculated
with
reo-
virus
type
1
generally
survive
and
may
develop
ependymal
cell
damage
(without
neuronal
necrosis)
and
hydrocephalus.
Using
recombinant
clones
derived
from
crosses
between
reovirus
types
1
and
3,
we
have
been
able
to
determine
that
the
SI
genome
segment
is
responsible
for
the
differing
cell
tropism
of
reovirus
serotypes
and
is
the
major
determinant
of
neurovirulence.
The
type
1
S1
genome
segment
is
responsible
for
ependymal
damage
with
subsequent
hydrocephalus;
the
type
3
SI
genome
segment
is
responsible
for
neuronal
necrosis
and
neurovirulence.
We
postulate
that
these
differences
are
due
to
the
specific
interac-
tion
of
the
al
outer
capsid
polypeptide
(the
protein
coded
for
by
the
SI
genome
segment)
with
receptors
on
the
surface
of
ei-
ther
ependymal
cells
or
neuronal
cells.
The
unique
specificities
of
individual
viruses
for
certain
host
species,
as
well
as
different
tissues
within
a
host,
are
hallmarks
of
viral
infections.
The
molecular
basis
of
these
properties
is
poorly
understood.
The
different
serotypes
of
reovirus
[defined
on
the
basis
of
neutralizing
and
hemagglutination
inhibition
antibodies
(1,
2)]
exhibit
differing
tropisms
for
cells
in
the
central
nervous
system
and,
presumably
as
a
consequence
of
their
differing
tropisms,
produce
differing
patterns
of
virulence.
Specifically,
when
reovirus
type
3
is
inoculated
intracerebrally
into
newborn
mice,
an
acute
encephalitis
develops
that
is
fatal
in
virtually
100%
of
animals
and
is
accompanied
by
destruction
of
neuronal
cells
without
damage
to
ependymal
cells
(3,
4).
Intracerebral
inoc-
ulation
of
reovirus
type
1
produces
a
nonfatal
infection
in-
volving
the
ependymal
cells
that
line
the
ventricular
cavities
of
the
brain,
with
little
or
no
effect
on
neurons
(5).
Hydro-
cephalus
often
develops
as
a
sequel
to
the
ependymal
damage
(6).
The
basis
for
this
difference
in
virulence
and
cell
tropism
between
reovirus
types
1
("avirulent")
and
3
("virulent")
is
unknown.
Reoviruses
are
animal
viruses
whose
genomes
consist
of
10
segments
of
double-stranded
RNA
(dsRNA)
named
according
to
size
classes:
3
large
segments
(LI,
L2,
L3),
3
medium
seg-
ments
(MI,
M2,
M3),
and
4
small
segments
(S1,
S2, S3,
S4).
The
dsRNA
is
surrounded
by
a
double
capsid
shell,
the
outer
capsid
consisting
of
three
polypeptides,
ar1,
or3,
and
g2,
derived
from
genome
segments
S1,
S4,
and
M2,
respectively
(7,
8).
Using
recombinant
clones
derived
from
crosses
between
reovirus
types
1
and
3,
and
consisting
of
genome
segments
de-
rived
from
both
parents,
we
have
previously
shown
that
the
S1
gene,
the
dsRNA
segment
that
codes
for
the
ol
outer
capsid
polypeptide,
is
responsible
for
type
specificity
[as
determined
by
neutralization
(9)]
and
for
hemagglutination
(H.
L.
Weiner,
unpublished
observations).
Using
the
same
genetic
approach,
we
have
been
able
to
determine
that
the
S1
dsRNA
segment
is
the
gene
responsible
for
the
differing
cell
tropisms
of
reovirus
serotypes
and
is
a
major
determinant
of
neurovirulence.
MATERIALS
AND
METHODS
Cells.
Mouse
L
cells
were
maintained
in
suspension
culture
in
Joklik's
modified
Eagle's
minimal
essential
medium
(Grand
Island
Biological
Co.,
Grand
Island,
NY)
supplemented
with
5%
fetal
calf
serum
[International
Biological
Laboratories
(IBL),
Rockville,
MDI.
Monkey
CV-1
cells
were
maintained
in
mo-
nolayers
in
IMEMZO
("improved
minimal
essential
media
supplemented
with
zinc,
insulin
and
Hepes
buffer,"
IBL)
supplemented
with
10%
fetal
calf
serum
(IBL).
Virus.
Reovirus
type
1
(Lang
strain)
and
type
3
(Dearing
strain)
were
the
same
as
previously
described
(10).
Recombinant
clones
were
prepared
by
mixedly
infecting
L
cells
with
equal
multiplicities
of
temperature-sensitive
(ts)
mutants
of
type
.3
reovirus
(11)
and
clones
were
collected
at
390
(a
nonpermissive
temperature
for
the
temperature-sensitive
mutants)
and
shown
to
be
recombinants.
Eight
such
clones
were
utilized
in
the
present
analysis
(Fig.
1
and
Table
1).
The
origin
of
these
clones
is
described
in
detail
elsewhere.11
Analysis
of
Viral
RNA.
Viral
RNA
was
analyzed
as
pre-
viously
described
(10).
Briefly,
CV-1
monolayers
were
inocu-
lated
with
the
two
serotypes
of
reovirus
or
hybrid
clones
at
a
multiplicity
of
infection of
10
and
were
grown
at
310
on
60-
X
15-mm
plastic
Falcon
tissue
culture
dishes
(Falcon,
Oxnard,
CA)
in
5
ml
of
IMEMZO.
At
2.5
hr
after
infection,
medium
containing
0.25
mCi
Of
32p
was
added.
At
48
hr,
infected
mo-
nolayers
were
washed
with
an
isotonic
buffer
(140
mM
NaCl/10
mM
Tris-hydrochloride
at
pH
7.4/1.5
mM
MgCl2;
buffer
A)
and
cells
were
collected.
The
cells
were
treated
with
0.5%
Nonidet
P-40
in
buffer
A,
nuclei
were
removed
by
cen-
trifugation,
and
RNA
was
precipitated
at
-20°
in
ethanol.
The
precipitates
were
pelleted
by
high-speed
centrifugation,
dried,
dissolved
in
gel
sample
buffer,
and
applied
to
slab
gels
(10%
acrylamide/0.267%
bisacrylamide)
as
described
by
Laemmli
(12).
After
electrophoresis,
gels
were
fixed
and
autoradiography
was
performed.
Abbreviations:
dsRNA,
double-stranded
RNA;
PFU,
plaque-forming
units.
§
Sharpe,
A.
H.,
Rlamig,
R.
F.,
Mustoe,
1'.
A.
&
Fields,
B.
N.,
"A
genetic
map
of
reovirus.
I.
Correlation
of
genlome
RNAs
between
serotvpes
1,
2,
and
3,"
Virology,
in
press.
Ramig,
R.
F.,
Mustoe,
T.
A.,
Sharpe,
A.
H.
&
Fields,
B.
N.,
"A
genetic
map
of
reovirus.
II.
Assignment
of
the
double-stranded
RNA
negative
mutant
groups
C.
D,
and
E
to
genome
segments,
Virology,
in
press.
5744
The
costs
of
publication
of
this
article
were
defrayed
in
part
by
the
payment
of
page
charges.
This
article
must
therefore
be
hereby
marked
"advertisement"
in
accordance
with
18
U.
S.
C.
§1734
solely
to
indicate
this
fact.
Proc.
Natl.
Acad.
Sci.
USA
74
(1977)
5745
TYPE
TYPE
204
3
802
L
I
urn~
~ ~
~~~0
-M
L1
_1
_0
L
2
_
m
-a
_L3
L3
0
431M2
mM2
A_
~M
i
4_M
I
41
MO-N
ris
M3
4E1
_3
b'
^'
'_SI
Vol
S2
-
-*S2
3g
o
d
34S3Iig
FIG.
1.
Autoradiograms
of
reovirus
dsRNA
segments
from
ser-
otypes
1
and
3
and
hybrid
clones
204
and
802,
electrophoresed
in
a
Tris/glycine
gel
system
(10).
The
three
large
(Li,
L2,
L3),
three
me-
dium
(M1,
M2,
M3),
and
four
small
(Si,
S2,
S3,
S4)
genome
segments
of
serotypes
1
and
3
are
labeled.
Each
of
the
genome
segments
from
the
two
serotypes
can
be
distinguished
on
the
basis
of
migrational
differences.
The
numbering
of
the
medium
and
small
segments
is
in
accordance
with
numbering
of
these
segments
in
Tris/acetate
gel
systems.
Clone
204
contains
genome
segments
LI
and
Si
(arrows)
from
type
3
and
the
remainder
of
segments
from
type
1.
Clone
802
contains
genome
segments
LI
and
Si
(arrows)
from
type
1
and
the
remainder
of
segments
from
type
3.
Inoculation
of
Animals
and
Viral
Titrations.
Pregnant
BALB/c
mice
were
obtained
from
Charles
River
Laboratories,
North
Wilmington,
MA,
and
newborns
were
inoculated
in-
tracerebrally
within
24
hr
of
birth
with
various
amounts
of
virus
13
X
105
to
3
X
102
plaque-forming
units
(PFU)/mousej
in
a
volume
of
0.03
ml
using
a
tuberculin
syringe
with
a
26-guage
needle.
Triplicate
litters
(a
total
of
15-20
animals)
were
inoc-
ulated
with
each
clone
or
parental
serotype
at
each
dilution
of
virus.
Survival
curves
were
determined
by
recording
the
number
of
animals
alive
up
to
15
days
after
inoculation.
Ani-
mals
dying
on
day
1
or
2
after
inoculation
(approximately
4%
of
animals
inoculated)
were
not
included
in
survival
curves.
Animals
were
observed
twice
daily
and
the
brains
from
dead
animals
were
either
fixed
in
formalin
for
routine
hematoxy-
lin/eosin
pathologic
analysis
or
frozen
at
-70°
for
determina-
tion
of
viral
titers
in
brain
tissue.
Surviving
animals
were
sac-
rificed
at
day
15;
occasionally,
healthy
animals
were
sacrificed
at
day
7.
All
viral
titrations
of
brain
tissues
were
done
on
animals
injected
with
3
X
105
PFU.
For
viral
titrations,
brains
were
thawed,
added
to
2
ml
of
gelatin
saline
(to
make
a
10%
sus-
pension),
and
sonicated.
Cell
debris
were
removed
by
low
speed
centrifugation
(600g
for
15
min),
and
the
virus-containing
su-
pernatant
was
then
assayed
on
L
cell
monolayers
according
to
the
plaque
assay
of
Fields
and
Joklik
(11).
RESULTS
Disease
Pattern
in
Animals
Injected
with
Reovirus
Type
3.
We
have
confirmed
the
finding
that
type
3
reovirus,
when
injected
intracerebrally
into
newborn
mice,
is
uniformaly
fatal
at
doses
ranging
from
3
X
10"
PFU
injected/mouse
to
3
X
103
PFU
injected/mouse;
at
3
X
102
PFU/mouse
20%
of
the
ani-
mals
survive
(3,
4).
At
higher
doses
(3
X
105)
animals
become
sick
at
day
5-6
and
are
dead
by
day
8.
At
lower
doses
(3
X
102)
animals
become
sick
at
day
8-9
and
most
are
dead
by
day
10-12
(Fig.
2).
Histological
sections
of
the
brains
of
animals
dying
following
intracerebral
injection
of
type
3
reovirus
show
a
necrotizing
encephalitis
with
destruction
of
neuronal
cells
manifest
by
fragmentation
of
the
nucleus
and
eosinophilic
cytoplasm.
The
ependymal
cells
that
line
the
ventricular
cavities
of
the
brain
are
unaffected
(Fig.
3).
Viral
titers
of
brains
from
dead
animals
that
were
injected
with
3
X
105
PFU
of
type
3
reovirus
were
consistently
greater
than
5
X
109
PFU/ml.
Disease
Pattern
in
Animals
Injected
with
Reovirus
Type
1.
In
contrast
to
the
lethal
effects
of
reovirus
type
3,
the
majority
of
newborn
mice
injected
with
type
1
reovirus
at
doses
ranging
from
3
X
105
PFU
injected/mouse
to
3
X
103
PFU
injected/
mouse
survive
(Fig.
2).
Some
of
the
animals
injected
with
type
1
reovirus
show
necrosis
of
the
ependymal
cells
that
line
the
ventricular
cavities
of
the
brain
and
hydrocephalus
(Fig.
3);
many
animals
show
normal
histologic
patterns.
In
contrast
to
type
3
reovirus,
there
is
no
involvement
of
neuronal
cells.
Thus,
the
sporadic
deaths
seen
in
animals
injected
with
type
1
are
apparently
not
due
to
the
same
mechanism
as
that
seen
with
type
3,
because
there
is
no
neuronal
involvement.
In
fact,
the
brains
of
such
animals
are
histologically
normal.
The
viral
titers
in
brains
of
animals
injected
with
type
3
reovirus
that
were
dying
of
encephalitis
on
day
6-8
were
>5
X
109
PFU/ml,
while
viral
titers
in
brains
of
animals
injected
with
3
X
105
PFU
of
type
1
reovirus
on
a
comparable
day
were
--3
loglo
lower
(-5
X
106
PFU/ml).
By
day
15,
viral
titers
in
animals
injected
with
type
1
had
fallen
to
-5
X
12
PFU/ml.
Disease
Patterns
in
Animals
Injected
with
Hybrid
Clones.
Recombinant
clones
that
segregated
each
of
the
10
dsRNA
segments
were
selected
for
animal
inoculation
(Table
1).
Clone
65
was
studied
first
because
all
the
dsRNA
segments
coding
for
the
outer
capsid
polypeptides
(M2,
S1,
S4)
were
derived
from
type
1,
while
most
of
the
genes
coding
for
internal
and
non-
structural
polypeptides
were
derived
from
type
3.
The.
pattern
of
survival
of
animals
inoculated
with
this
clone
was
similar
to
the
pattern
in
those
inoculated
with
type
1
(Fig.
2).
In
addition,
histological
study
demonstrated
that
some
animals
had
epen-
dymal
cell
necrosis
with
hydrocephalus;
no
animals
showed
evidence
of
neuronal
necrosis.
Furthermore,
titers
of
virus
from
brain
tissue
of
sacrificed
animals
were
comparable
to
those
seen
with
type
1.
Because
the
general
pattern
of
virulence
and
cell
damage
induced
following
inoculation
with
clone
65
most
resembled
that
of
type
1,
these
early
experiments
suggested
that
one
or
more
components
of
the
outer
capsid
were
playing
a
primary
role
in
the
neurotropism
of
reovirus.
Accordingly,
subsequent
experiments
focused
on
recombi-
nants
that
segregated
genes
coding
for
the
polypeptides
of
the
outer
capsid
(M2,
SI,
S4).
Because
the
Si
genome
segment
codes
for
the
protein
(al)
that
confers
type
specificity
for
reovirus
(both
in
terms
of
neutralization
and
hemagglutination
prop-
erties),
a
hybrid
clone
was
next
chosen
to
determine
if
the
Si
gene
was
also
responsible
for
the
type-specific
virulence
of
reovirus
type
3.
Hvbrid
clone
204
(Fig.
1
and
Table
1)
contains
an
S1
dsRNA
segment
from
type
3
and
M2
and
S4
dsRNA
segments
from
type
1;
the
remainder
of
genome
segments
are
Microbiology:
Weiner
et
al.
5746
Microbiology:
Weiner
et
al.
Table
1.
Patterns
of
virulence,
viral
titers,
and
histology
following
intracerebral
inoculation
of
hybrid
reovirus
clones
into
newborn
mice
Origin
of
genome
segment*
Non-
Pattern
Outer
capsid
Core
structural
Uncertain
of
Viral
Clone
M2
Si
S4
Li
L2
S2
M3
S3
L3
MI
Diseaset
titerst
Histology§
65
1
1
1
3
3
1
3
1
3
.3
1
1
Normal
&
1
54
1
1
33
1
3
3
3
1
1
1
1
Normal
&
1
802
.3
1
3
1
3
3
3
3 3 3
Intermediate
Intermediate
Normal
&
1
80
3
1
3
1
3
1
3
3
1
3
Intermediate
Intermediate
Normal
&
1
103
.3
1
11
3
3 3
3 3 3
Intermediate
1
Normal
&
1
204
1
3
1
3
1
1 1 1
1
1
3
3
:3
63
1
3
3
3
1
3
1
1
3
1
3
.3
3
94
3
3
1
1
3
1
3
1
3
1
3
3
3
*
Numbers
in
the
table
indicate
from
which
type
of
reovirus
the
genome
segment
of
the
hybid
clone
originated.
The
segments
are
grouped
according
to
the
location
of
their
protein
products
in
the
virus.
t
Pattern
of
disease:
Type
1,
infected
animals
generally
survive
without
overt
acute
illness.
Type
3,
infected
animals
die
after
an
incubation
period
of
5-10
days
with
occasional
survivors
at
3
X
102
PFU/mouse.
Intermediate
pattern,
death
usually
occurs
later
than
with
type
.3
and
most
animals
survive
at
lower
doses
of
viral
inoculation
(3
X
103
to
3
X
102
PFU/mouse).
Viral
titers:
Infected
brains
have
revealed
three
levels
of
peak
virus
titers.
Type
1
=
105-106
PFU/ml.
Type
3
>5
X
109
PFU/ml.
Intermediate
level
=
107-108
PFU/ml.
§
Histology:
Type
1,
ependymal
damage
with
associated
hydrocephalus;
neuronal
cells
are
normal.
Type
.3,
necrotizing
encephalitis
affecting
neurons;
ependymal
cells
are
normal.
Normal,
no
abnormal
neuropathologic
findings.
from
type
1
except
for
LI,
which
is
from
type
3.
Thus
clone
204
contains
eight
genes
from
type
1
and
two
genes
(Si
and
Li)
from
type
3.
As
shown
in
Fig.
2,
clone
204
is
exceptionally
virulent,
analogous
to,
and
perhaps
even
more
virulent
than,
type
3.
Following
intracerebral
inoculation,
we
have
had
no
survivors
with
clone
204
at
3
X
102
PFU/per
mouse
(-20%
of
animals
injected
with
3
X
102
PFU
of
reovirus
type
3
survive).
Histologic
study
has
shown
a
necrotizing
encephalitis;
epen-
dymal
cells
have
been
intact
in
all
animals
studied
and
no
ani-
mals
have
had
hydrocephalus.
Viral
titers
of
these
brains
have
been
>5
X
109.
These
experiments
thus
suggested
that
the
virulence
and
cell
tropism
of
type
3
is
a
property
of
the
Si
gene
product.
Because
the
Si
gene
of
type
3
conferred
neuronal
tropism
and
virulence
into
a
recombinant
(clone
204)
containing
pri-
marily
type
1
genes,
it
was
important
to
determine
if
a
recip-
rocal
recombinant
(i.e.,
containing
a
type
1
SI
gene
on
a
pre-
dominantly
type
3
background)
would
be
avirulent
and
pro-
duce
ependymitis
and
hydrocephalus.
Hybrid
clone
802
(Fig.
1
and
Table
1)
contains
SI
and
Li
segments
from
type
1,
while
all
other
genome
segments
are
derived
from
type
3.
As
shown
in Fig.
2,
clone
802
is
virulent
at
high
doses
(3
X
105)
but
is
relatively
avirulent
at
low
doses
(3
X
103
and
3
X
102).
However,
it
is
not
as
virulent
as
clone
204
or
as
the
type
3
parent
even
at
higher
doses
(death
occurs
between
days
9
and
I1
following
inoculation
with
clone
802,
whereas
death
occurs
between
days
5
and
7
following
injection
with
clone
204
or
type
3).
Histologic
study
reveals
that,
while
most
animals
dying
after
inoculation
with
clone
802
have
normal
brains,
some
animals
show
epen-
dymal
necrosis
and
hydrocephalus.
No
animals
had
evidence
of
neuronal
necrosis.
Viral
titers
from
infected
brains
were
-5
X
108
PFU/ml,
higher
than
those
of
animals
injected
with
type
1
but
lower
than
those
from
animals
injected
with
type
3.
These
results
provide
strong
evidence
that
a
type
1
gene
in
the
SI
position
is
responsible
for
the
ependymal
tropism
of
reovirus
type
1.
However,
while
the
site
of
tissue
destruction
appears
linked
to
the
SI
gene,
the
fact
that
there
is
an
intermediate
pattern
of
virulence
in
clone
802
suggests
a
role
for
additional
genes
in
the
virulence
pattern
of
this
clone.
To
further
confirm
the
findings
obtained
with
clones
65,
802,
and
204,
additional
hybrid
clones
(54,
63,
80,
94,
103)
that
further
segregated
each
of
the
viral
genes
were
inoculated
in-
tracerebrally
into
newborn
mice.
The
results
with
these
clones
are
summarized
in
Table
1
and
Fig.
2.
In
each
instance,
the
presence
of
a
type
3
Si
gene
(clones
63
and
94)
conferred
a
high
degree
of
virulence
associated
with
neuronal
destruction,
while
a
type
1
Si
gene
led
to
ependymal
damage
and
hydrocephalus
(or
a
normal
histologic
appearance)
associated
with
a
type
1
(clone
54)
or
an
intermediate
(clones
80
and
103)
pattern
of
death
(Table
1).
These
results
confirmed
our
prior
conclusions
concerning
the
primary
role
of
the
S1
gene.
DISCUSSION
The
results
we
have
obtained
utilizing
recombinant
clones
derived
from
crosses
between
reovirus
type
1
and
type
3
indi-
cate
that
a
single
gene
is
responsible
for
the
differing
neuro-
tropism
of
reovirus.
The
presence
of
an
SI
gene
from
type
1
is
responsible
for
the
destruction
of
ependymal
cells
and
the
subsequent
development
of
hydrocephalus.
Ependymal
damage
only
occurs
when
the
Si
gene
is
from
type
1
(clones
54,
65, 80,
103,
and
802).
This
is
most
dramatically
illustrated
with
clone
802,
which
has
only
two
genome
segments
from
type
1
(SI
and
Li).
The
LI
segment
in
this
hybrid
is
not
responsible
for
the
ependymal
damage
because
ependymal
cells
are
normal
in
animals
injected
with
clone
94
(which
has
a
type
1
LI
seg-
ment).
The
presence
of
an
Si
gene
from
type
3
is
responsible
for
the
acute
necrotizing
encephalitis
with
neuronal
destruction
associated
with
type
3
virus
(clones
63,
94,
and
204).
This
is
dramatically
illustrated
with
clone
204,
which
has
only
two
genome
segments
from
type
3
(S1
and
LI).
The
L1
segment
is
not
responsible
for
the
encephalitis
and
neuronal
destruction
because
neuronal
cells
are
normal
following
inoculation
with
clones
65
and
54,
whose
LI
segments
derive
from
type
3.
Thus,
the
neuronal
destruction
caused
by
reovirus
type
3
is
related
to
the
SI
genome
segment.
While
the
relationship
of
the
SI
gene
to
distribution
of
virus
in
the
brain
("cell
tropism")
appears
to
be
absolute,
the
rela-
tionship
of
the
SI
gene
to
death
of
the
infected
animals
is
somewhat
more
complex.
Recombinants
consisting
of
an
SI
from
type
3
inserted
into
a
background
of
predominanty
type
1
genes
(such
as
clone
204)
are
extraordinarily
virulent.
How-
ever,
the
reciprocal
recombinants,
i.e.,
those
in
which
the
Proc.
Natl.
Acad.
Sci.
USA
74
(1977)
Proc.
Natl.
Acad.
Sci.
USA
74
(1977)
5747
0~
~~~~~~~~~#0
3°
650
#204
l
*,100
\-\\
:
\
6
3
100
0.
~~
~~~~~~~~~94
63
(105)
(1O5)
50
105
~~~~~~~~~~~~~~~~~50
L54
,,,,
1#94
&
63
100
10
2
2
5
100
102
50
50
10~
#802
14
10
5
10
15
5
10
15
Days
after
injection
Fi(;.
2.
Survival
patterns
of
mice
after
intracerebral
inoculation
with
parental
reovirus
types
1
and
3
and
various
recombinant
clones.
Newborn
mice
were
injected
intracerebrally
within
24
hr
of
birth
with
various
amounts
of
virus
and
studied
for
15
days.
Survival
curves
are
labeled
according
to
amount
of
virus
injected;
102,
10;3,
104,
and
10"
represent
3
X
102,
3
X
103,
3
X
104,
and
3
X
105
PFU
of
virus
injected
per
mouse.
background
is
predominantly
type
3
but
which
contain
a
type
1
SI
gene
(such
as
802,
103,
and
54)
show
a
reduced
and
delayed
pattern
of
death
that
is
intermediate
between
types
1
and
3
(802,
103)
or
essentially
like
type
1
(54).
Histologically,
the
brains
of
animals
that
die
following
injection
with
these
clones
are
devoid
of
neuronal
destruction,
showing
either
ependymal
damage
or
no
abnormalities.
Thus,
the
mechanism
of
death
is
almost
certainly
different
than
that
associated
with
the
type
3
SI
neuronal
destruction.
The
differences
in
neurotropism
and
the
different
patterns
of
virulence
caused
by
reovirus
types
1
and
3
are
most
likely
due
to
specific
interaction
of
the
ol
outer
capsid
protein
(the
protein
coded
for
by
the
SI
genome
segment)
with
receptors
on
the
surface
of
either
ependymal
cells
or
neuronal
cells.
According
to
this
scheme,
the
al
protein
from
type
1
interacts
with
a
cell
surface
receptor
on
ependymal
cells,
allowing
viral
penetration
into
the
cell
and
subsequent
ependymal
cell
damage.
In
an
analogous
fashion
the
c1
protein
from
type
3
interacts
with
a
cell
surface
receptor
on
neuronal
cells.
Furthermore,
the
specificity
of
cell
tropism
exhibited
by
reovirus
types
1
and
3
suggests
that
there
is
no
crossreactivity
between
the
c1
protein
of
type
1
and
neuronal
cells
or
the
c1
protein
of
type
3
and
ependymal
cells.
The
fact
that
c1
is
the
hemagglutinin
poly-
peptide
further
supports
its
role
in
receptor
interactions
(13).
Other
investigators
have
suggested
that
cell
receptors
play
an
important
role
in
determining
viral
tropism
and
virulence.
Sabin
showed
that
adsorption
of
an
avirulent
strain
of
polio-
myelitis
to
central
nervous
system
tissue
did
-not
occur
under
conditions
that
facilitated
the
adsorption
of
a
virulent
strain
(14).
In
a
similar
fashion,
Holland
(15)
has
shown
that
susceptibility
to
coxsackie
viruses
correlates
with
the
presence
of
receptors
for
viral
adsorption
on
the
cell
surface;
in
several
animal
hosts,
homogenates
of
susceptible
organs
adsorb
the
virion,
while
homogenates
of
insusceptible
organs
do
not.
Mayer
et
al.
(16)
used
a
genetic
approach
to
investigate
the
neurovirulence
of
a
particular
strain
of
influenza,
A/NWS/cc-p
(HON1).
Recombinants
between
this
virulent
strain
and
a
nonneurovirulent
strain
of
influenza,
A/Jap/305
(H2N2).,
demonstrated
no
link
between
neurovirulence
and
either
the
viral
hemagglutinin
or
neuraminidase.
Earlier
genetic
studies
of
neurovirulent
strains
of
influenza
(17-19)
demonstrated a
variable
expression
of
virulence
and
it
was
generally
felt
that
virulence
was
a
multigenetic
phenomenon.
Our
studies
thus
differ
from
those
with
influenza
because
the
Microbiology:
Weiner
et
al.
5748
Microbiology:
Weiner
et
al.
Proc.
Natl.
Acad.
Sci.
USA
74
(1977)
V..~~~~~~~~~~~w
A
*
4
*.
o
E
~
a
C
-i*;
At
*;
,*'s
'.'qfe,
'(''
C'''
-0
a
a
.-e
00,i~
%*
*
* t
*
*
"
SW~~~~
,
B
*
1~~fs.;;;
*
wI.'
+
'
.
D
*N
S'
n
S
.S-i#,w
,ws
'.9
L
.-
A
s
*
S
S~~~~~~~~L'I...~~~~~~~~~.b'4
~ ~ ~ ~
~~l
*
.
~~~~~~~~~~~~~~~I
$',',4
*
;
.f
4
~
,;4
K
W
1
e
',
,
e 1
-s
'
V
:;1r<6*
I;
..SP i+
I.*
.*r
,w*
iv<
.l1
ut
b
Fic.
3.
Hematoxylin/eosin
stains
of
brains
from
mice
injected
with
reovirus
types
1
and
3.
(A)
Normal
mouse
brain
from
7-day-old
animal:
histologic
appearance
of
the
caudate/putamen
(c),
lateral
ventricular
ependyma
(arrow),
and
septal
nucleus
(f).
(X85.)
(B)
Brain
of
mouse
injected
with
reovirus
type
3
and
dying
on
day
7:
extensive
neuronal
necrosis
within
th6
septal
nucleus
(arrow)
with
preservation
of
ependymal
cells.
(X85.)
(C)
Brain
of
mouse
injected
with
reovirus
type
1
and
sacrificed
on
day
15:
greatly
enlarged
lateral
ventricle
(L)
with
preservation
of
septal
nucleus
and
caudate/putamen.
Arrow
marks
site
of
detail
in
D
below.
(X21.)
(D)
Detail
of
ventricular
wall
from
C
(above)
showing
ependymal
cell
necrosis
represented
by
karyorrhexis
(arrow).
There
is
no
evidence
of
neuronal
necrosis.
(X85.)
r1
polypeptide,
the
reoviral
hemagglutinin,
is
linked
to
both
neurovirulence
and
cell
tropism.
However,
our
results
using
recombinant
clones
lacking
a
type
3
gene
in
the
S1
position
(802,
103,
54)
but
containing
type
3
genome
segments
in
the
M2
and/or
S4
positions
(genes
coding
for
the
other
two
outer
capsid
proteins)
suggest
that
multiple
genes
may
be
capable
of
pro-
ducing
virulence
in
an
altered
manner
when
the
primary
vir-
ulence
gene
is
absent.
It
maybe
that
these
"secondary
virulence
genes"
produce
their
effects
through
control
of
viral
replication
rather
than
by
determining
specific
virus/host-cell
interac-
tions.
We
would
like
to
thank
Sandra
Kleiderman
and
Martha
Hubbard
for
their
excellent
secretarial
support.
This
work
was
supported
by
National
Institutes
of
Health
Grant
1
RO1
Al
13178-01
and
research
support
from
the
Milton
Fund
of
Harvard
University.
H.L.W.
is
the
recipient
of
National
Institutes
of
Health
Research
Fellowship
Award
1
F32
A105461-01
and
Teacher
Investigator
Award
1
K07
NS
00237-01
from
the
National
Institute
of
Neurological
and
Communicative
Dis-
orders
and
Stroke.
1.
Rosen,
L.
(1960)
Am.
J.
Hyg.
71,
242-249.
2.
Rosen,
L.
(1962)
Ann.
N.Y.
Acad.
Sci.
101,
461-465.
3.
Margolis,
G.,
Kilham,
L.
&
Gonatos,
N.
K.
(1971)
Lab.
Invest.
24,
101-109.
4.
Raine,
C.
S.
&
Fields,
B.
N.
(1973)
J.
Neuropathol.
Exp.
Neurol.
32,
19-33.
5.
Kilham,
L.
&
Margolis,
G.
(1969)
Lab.
Invest.
21,
189-198.
6.
Masters,
C.,
Alpers,
M.
&
Kakulas,
B.
(1977)
Arch.
Neurol.
34,
18-27.
7.
Smith,
R.
E.,
Zweerink,
H.
J.
&
Joklik,
W.
K.
(1969)
Virology
39,
791-810.
8.
Joklik,
W.
K.
(1974)
Comp.
Virol.
2,
297-298.
9.
Weiner,
H.
L.
&
Fields,
B.
N.
(1977)
J.
Exp.
Med.,
146,
1305-
1310.
10.
Ramig,
R.
F.,
Cross,
R.
K.
&
Fields,
B.
N.
(1977)
J.
Virol.
22,
726-733.
11.
Fields,
B.
N.
&
Joklik,
W.
K.
(1969)
Virology
37,
335-342.
12.
Laemmli,
U.
K.
(1970)
Nature
227,
680-685.
13.
Winzler,
R.
J.
(1969)
in
Cellular
Recognition,
eds.
Smith,
R.
T.
&
Good,
R. A.
(Appleton,
New
York),
p.
11.
14.
Sabin,
A.
B.
(1957)
in
Cellular
Biology,
Nucleic
Acids
and
Vi-
ruses,
ed.
St.
Witlock,
0.
V.
(New
York
Academy
of
Science,
New
York),
p.
113.
15.
Holland,
J.
J.
(1961)
Virology
15,312-326.
16.
Mayer,
V.,
Schulman,
J.
L.
&
Kilbourne,
E.
D.
(1973)
J.
Virol.
11,
272-278.
17.
Burnet,
F.
M.
&
Lind,
P.
E.
(1954)
Nature
173,
627-630.
18.
Fraser,
K.
B.
(1959)
Virology
9,
191-201.
19.
Gotlieb,
T.
&
Hirst,
G.
K.
(1956)
Virology
2,
235-248.