Synemin
and
Vimentin
are
Components
of
Intermediate
Filaments
in
Avian
Erythrocytes
BRUCE
L
.
GRANGER,
ELIZABETH
A
.
REPASKY,
and
ELIAS
LAZARIDES
Division
of
Biology,
California
Institute
of
Technology, Pasadena,
California
91125
ABSTRACT
Synemin,
a
high-molecular-weight
protein
associated
with
intermediate
filaments
in
muscle,
and
vimentin,
an
intermediate-filament
subunit
found
in
many
different
cell
types,
have
been
identified
by
immunologic
and
electrophoretic
criteria
as
components
of
interme-
diate filaments
in
mature
avian
erythrocytes
.
Desmin,
the
predominant
subunit
of
intermediate
filaments
in
muscle,
has
not
been
detected
in
these
cells
.
Two-dimensional
immunoautora-
diography
of
proteolytic
fragments
of
synemin
and
vimentin
demonstrates
that
the
erythrocyte
proteins
are highly
homologous,
if
not
identical, to
their
muscle
counterparts
.
Double
immu-
nofluorescence
reveals that
erythrocyte
synemin
and
vimentin
co-localize
in
a
cytoplasmic
network
of
sinuous
filaments
that
extends
from
the
nucleus
to
the
plasma
membrane
and
resists
aggregation
by
Colcemid
.
Erythrocytes
that
are
attached
to glass
cover
slips
can
be
sonicated
to
remove
nuclei
and
nonadherent
regions
of
the
plasma
membrane
;
this
leaves
elliptical
patches
of
adherent
membrane
that
retain
mats
of
vimentin-
and
synemin-containing
intermediate
filaments,
as
seen
by
immunofluorescence
and
rotary
shadowing
.
Similarly,
mechanical
enucleation
of
erythrocyte
ghosts
in
suspension
allows
isolation
of
plasma
mem-
branes
that
retain
a
significant
fraction
of
the
synemin
and
vimentin,
as
assayed
by
electropho-
resis,
and
intermediate
filaments,
as
seen
in
thin
sections
.
Both
synemin
and
vimentin
remain
insoluble,
along
with
spectrin
and
actin,
in
solutions
containing
nonionic
detergent
and
high
salt
.
However,
brief
exposure
of
isolated
membranes
to
distilled
water
releases
the
synemin
and
vimentin
together
in
nearly
pure
form,
before
the
release
of
significant
amounts
of
spectrin
and
actin
.
These
data
suggest
that
avian
erythrocyte
intermediate
filaments
are
somehow
anchored
to
the
plasma
membrane
;
erythrocytes
may
thus
provide
a
simple
system
for
the
study
of
intermediate
filaments
and
their
mode
of
interaction
with
membranes
.
In
addition,
these
data,
in
conjunction
with
previous
data
from
muscle,
indicate that
synemin
is
capable
of
associating
with
either
desmin
or
vimentin
and
may
thus
perform
a
special
role in
the
structure
or
function
of
intermediate
filaments
in
erythrocytes
as
well
as
in
muscle
.
Mature
avian
erythrocytes
are
nucleated,
biconvex,
elliptical
discs
that
contain
relatively
few
cytoplasmic
organelles
.
Re-
moval
of hemoglobin from
these
cells
by
hypotonic
lysis
(15)
reveals
an
equatorial
bundle
of
microtubules
known
as the
marginal
band
(5),
a
submembranous
spectrin-actin
shell
(9),
as
well
as
a
residual
network
of
cytoplasmic
filaments
that
surrounds
the
mitochondria
and
extends
from
the
nucleus
to
the
plasma
membrane
(27)
.
This
latter
network
of
filaments
is
probably
a
component
of
the
"trans
marginal
band
material"
noted
in
many
nonmammalian
vertebrate
erythrocytes
(13)
.
These
filaments
appear
to
be of
the type
known
as
intermediate
filaments
(33, 39),
due
to
their characteristic
ultrastructural
morphology
and
insolubility
in
nonionic
detergents
(59,
63)
.
THE
JOURNAL
OF
CELL
BIOLOGY
"
VOLUME
92
FEBRUARY
1982
299-312
©
The
Rockefeller
University Press
"
0021-9525/82/02/0299/14$1
.00
The
close
association
of
these
filaments
with
the
plasma
mem-
brane
and
nucleus,
as
shown
by
electron
microscopy,
suggests
that
they
might
function
to
maintain
the
shape
of
the
cell
or
position
the
nucleus
within
the
cell
(13,
27, 59,
63)
.
We
have examined
these
filaments
biochemically,
immuno-
logically,
and
ultrastructurally
and
have
determined
that
they
are
composed
predominantly
of
vimentin,
an
intermediate
filament
subunit
common
to
many
different
cell
types
(18,
39)
.
The
other
major
component
of
these
filaments
is
synemin,
a
high
molecular
weight
protein
originally isolated
from
avian
smooth
muscle
in
association
with
desmin
and
subsequently
shown
to
co-localize
with
desmin
and
vimentin
in
skeletal
muscle
(25)
.
Double
immunofluorescence
shows
that erythro-
299
cyte
synemin
and
vimentin
also coexist
in
a network
of
cyto-
plasmic
filaments
.
The
electrophoretic
and
immunologic
cri-
teria
used
to
identify
vimentin
and
synemin
in
these
cells
fail
to
detect
desmin,
the
major
intermediate
filament
subunit
of
muscle
(40,
53)
that
is
also
found
in
some
nonmuscle
cells (19,
55,58)
.
Various
cell
fractionation
procedures
based
on
differential
centrifugation
have
indicated
that
synemin
and
vimentin
sed-
iment
with
both
nuclear
and
membrane
fractions
.
In
this
study
we
have
concentrated
on
those
intermediate
filaments
that
remain
associated
with
the
erythrocyte
plasma
membrane
after
mechanical
enucleation
of
the
cells
.
These
filaments
resist
dissociation
from
the
membranes
by
sonication
and
treatment
with
high
salt
and
nonionic
detergent,
suggesting
that
they
are
in
some
way
anchored
to
the
membrane
cytoskeleton,
perhaps
to
the
spectrin-actin
network
.
However,
we
have
found
that
the
filaments
can
be
selectively
removed
from
the
membranes
by
treatment
with
low
ionic
strength
solutions
and
that
the
pre-
dominant
proteins thus
released
are
vimentin
and
synemin
.
The
evidence
presented
here
that
synemin
associates
with
vimentin
in
erythrocytes,
in
conjunction
with
the
evidence
presented
previously
that
synemin
associates
with
desmin
in
smooth
muscle
and
desmin
and
vimentin
in skeletal
muscle
(25),
suggests
that
synemin
may
be
capable
of
associating
with
different
intermediate
filament
subunits
in
different
cell
types
.
Because
synemin
is
not
detectable
in
all
cell
types,
it
may
play
a
special
role in
the
structure
or function of
certain
types
of
intermediate filaments
or
filaments
in
certain
cell
types
.
MATERIALS
AND
METHODS
Preparation
of
Erythrocyte
Membranes
White
leghorn
chickens
were
given
intravenous
injections
of3-4
mg
(500-700
USP
units)
of
heparin
(17)
and
then
-50
mg
of
sodium
pentobarbital
.
Blood was
collected
from
the
neck vein
into
a
solution
of0
.01%
heparin,
5
mM
HEPES
(N-
2-hydroxyethylpiperazine-N'-2-ethanesulfonic
acid),
155
mM
choline
chloride
(pH
7
.1
at
room
temperature)
.
Phosphate-buffered
saline
was
simultaneously
injected
into
a wing
vein,
and
the
perfusate
was
also
collected
until
it
became
relatively
clear
.
Alternatively,
blood
for
some
experiments
was
drawn
from wing
veins
or
neck
veins
of
uninjected
chickens
and
collected
in
1-2
vol
of
the
above
heparin-containing
solution
.
Blood
cells
were
pelleted
by
centrifugation
for
5
min
at
1,000
g,
and
the top
white layer
of
cells
(huffy
coat)
and
supernatant were
removed
by
aspiration
.
The
erythrocyte
pellet,
exclusive
of
a
dark-red
layer
adhering
to
the
bottom of
the
centrifuge
tube,
was
resuspended
in
155
mM
choline
chloride,
5
mM
HEPES
(pH
7
.1
at
room
temperature)
and
recentrifuged
.
Again, the
supernatant,
huffy
coat,
and
dark-red
layer
were
discarded
.
This
cycle
was
repeated
for
a
total
of
4-8
washes,
and was
performed
either
at
room
temperature
or
at
4°C
.
The
final pellet
of erythrocytes
was
rapidly
resuspended
in at
least
10
vol
of
ice-cold
hypotonic
lysis
buffer
(Buffer
H)
[5
mM
Tris-CI
(pH
7
.5),
5
mM
NaN,,,
5
MM
M9C1
2
, 1
mM
EGTA
(ethyleneglycol-bis(P-amino-ethyl
ether)
N,N'-tetra-
acetic acid),
1
mM
dithioerythritol
(DTE)
or
dithiothreitol
(DTT), 0
.5
mM
phenylmethyl
sulfonyl
fluoride
(PMSF))
.
MgCl
2
was
included
to
keep
the nuclei
intact,
EGTA
and
PMSF
were
included
as protease
inhibitors,
and
DTE
(or
DTT)
was
found
to
increase
both
the
yield
and
size
of
the
plasma
membrane
fragments
(see
below)
.
The
resulting
nucleated
erythrocyte
ghosts
were
pelleted
by
centrifugation
for
5
min
at
approximately
10,000
g,
then
resuspended
in
at
least
10
vol
of
the
same
solution
.
This
cycle
was
repeated
for
a
total
of
three
or
four
washes
in
Buffer
H
.
For
the
preparation
of
plasma membranes,
the
final
pellet
was
resuspended
in
2-4
vol
of
Buffer
H,
loaded
into
a
syringe,
and
forced
rapidly
through a
23
gauge
hypodermic
needle
bent
into
the
shape
of
a
Z
(with
two 30°
angles)
.
Centrifu-
gation
for
10-20
min
at
1,000
g
in
a
swinging
bucket
rotor
resulted
in
three
layers
:
a
firm,
white
pellet
of
free
nuclei
on
the
bottom,
a
loose,
pink
layer
of
undisrupted
cells
in
the
middle,
and a
supernatant
containing
soluble
proteins
and membrane
fragments
.
The
middle
layer
was
resuspended
in Buffer
H
and
again
forced
through
the needle
and
centrifuged
to
give
the
three
layers
.
This
was
usually
repeated
three
times
;
the
yield
of
membrane
fragments
increased
with each
cycle
.
The
supernatants
from
each
centrifugation
were
combined
and
recentrifuged
for
30
min
at
100-200
g
to
remove
any
remaining
nuclei
.
The
supernatant
of
this
low
30
0
THE
JOURNAL
OF
CELL
BIOLOGY
"
VOLUME
92,
1982
speed
centrifugation
was
respun
for
10
min
at
20,000
g
to
pellet
the
membranes
which were
subsequently
resuspended
in
10-20
vol
of
Buffer
H
and
stored
on
ice
.
Contaminating
nuclei
visible
by
phase
contrast
microscopy could
be
removed
by
another
low
speed
centrifugation
.
After
storage
for
a
month
on
ice,
little
change
was
seen
in
the
electrophoretic
protein
profile
as
judged by
one
dimensional
SDS-polyacrylamide
gel
electrophoresis
(SDS-PAGE)
(see
below)
.
Selective
Extraction
of
Synemin
and
Vimentin
From
Erythrocyte
Membranes
The
suspension
of
erythrocyte
membranes
was
mixed
with
10-30
vol
of2
mM
EDTA
10
mM
Tris
(pH
adjusted
with
HCl
to
7
.4 at
0°C)
;
after
10
min
on
ice
the
membranes
were
pelleted
by
centrifugation
at
20,000
g
for
10
min
.
The
pellet
was
resuspended
(with
a
Pasteur
pipette)
in
distilled
water
at
0°C and
centrifuged
as
above
.
The
supernatant
was
then
lyophilized
directly
for
electrophoretic
analysis,
or
first
recentrifuged
for
1-5
h
in
a
Beckman
SW
50
.1
swinging
bucket
rotor
at
50,000
RPM
(Beckman
Instruments,
Inc
.,
Fullerton,
Calif)
.
Polyacrylamide
Gel
Electrophoresis
One-dimensional
SDS-PAGE
was
based on
the
discontinuous,
Tris-glycine
system
of
Laemmli
(37),
as
modified
and
described
previously
(29)
.
Separating
gels
were
1l
x
14
x
0
.16
cm
and
contained
12
.5%
acrylamide
and
0
.11%
N,N'-
methylene-bisacrylamide
.
Samples were
solubilized
with
1%
SDS,
125
mM
Tris-
Cl
(pH
6
.8),
10%
glycerol,
1%
2-mercaptoethanol,
1
mM
EDTA,
0
.004% bromo-
phenol
blue
("l%
SDS
sample
buffer"),
and
immediately
placed
in
a
boiling
water
bath
for
-1
min
.
Two-dimensional
isoelectric
focusing
(IEF)/SDS-PAGE
was performed
ac-
cording
to
the
method
of
O'Farrell
(48),
as
modified
and
described
previously
(29),
except
that
nonidet
P-40
(NP-40)
was
omitted
from
all
gels
and
samples
(this
enhanced
the
resolution
of
the
isoelectric
variants
of
several
proteins of
interest
in
this
system)
.
Second-dimension
SDS
slab
gels
were
as
described
above
.
Phosphorylation
Incorporation
of
["P]-phosphate
into
erythrocyte
proteins
was
performed
essentially
as
described
by
Beam
et al
.
(4)
and
Alper
et al
.
(2)
.
Blood
was
collected
from
the
wing
vein of
an
adult
hen
turkey
in
cold
heparinized
choline
chloride
buffer
and
washed
as
described
above
.
Erythrocytes
were
then
washed once
at
room
temperature with
40
vol
of
157
.5
mM
NaCl,
2
.5
mM
KCI,
11
.1
MM
D-
glucose,
10
mM
HEPES
(brought
to
pH
7
.65
at
room
temperature
with
NaOH)
.
1
ml
of packed
cells
was
resuspended
in
9
ml
of
this
solution
;
2
mCi
of
'P-
phosphoric
acid
(New
England
Nuclear,
Boston,
MA
:
100
pl
of
carrier-free
in
20
mM
HCI)
were then
added
and
the
suspension
was
incubated
in
an
orbital
shaker
bath
at
39°C
.
After
3
.5
h
the
suspension
was
divided
in
half
;
to
one-half
was
added
50
pl
of
0
.1
mM
DL-isoproterenol-HC1
(Sigma Chemical
Co
.,
St
.
Louis,
MO)
in
the
above
solution
(final
concentration
1
pM)
.
Both
aliquots
were
incubated
for
another
20
min
at
39°C
and
then
processed
in
parallel
for
gel
electrophoresis
.
The
erythrocytes
were
spun
down
and
washed
once
with
40
vol
of
the
above
solution
at
39°C,
then
lysed
with 80
vol
of
10
mM
Tris-Cl
(pH
8
.0),
5
MM
MgC1
2
,
1
mM
EGTA,
1
mM
o-phenanthroline,
0
.5
mM
PMSF,
I
mM
NaF
at
0°C
.
The
nucleated ghosts
were
spun
down
and
washed
once with
this
solution,
then
disrupted
by one
passage
through a
bent
hypodermic
needle
as
described
above
.
Intact
cells
and
nuclei
were
removed
by
low-speed
centrifuga-
tion,
and
free
membrane
fragments
were
collected
by
high-speed
centrifugation
.
The
membrane
pellet
was
boiled
in
1%
SDS
sample
buffer
and
analyzed
by
SDS-
PAGE
.
After
staining
and
destaining,
the
gel
was
dried
on
filter
paper
and
exposed
to
Kodak
X-Omat
R
XR5
film
at
room
temperature
for
10
d
without
an
intensifying
screen
.
In
a
separate
experiment,
erythrocytes
were
collected,
washed
and
labeled
as
described above, except
that
the
labeling
period
was
20 h
and
no
isoproterenol
was
added
.
Cells
were
lysed
with
5
mM
sodium
phosphate,
5
mM
MgC12,
l
mM
EGTA,
1
mM
PMSF
(pH
7
.4
at
4°C)
and
disrupted
as
described
above
.
The
membrane
fraction
was
dissolved
in 10
M
urea
containing
1
%
2-mercaptoethanol,
analyzed
by
IEF/SDS-PAGE,
and
autoradiographed
.
Immunoautoradiography
Immunoautoradiography
was
performed
as
described
previously
(25)
.
Gels
were
incubated
with
antisera
diluted
1,000-fold,
followed
by
radioiodinated
protein
A
(4-7
pCi
'
2S
I/pg
protein
A
;
each
gel
was
incubated
with
20-30 pCi
in
100
ml
of
solution)
.
The
dried
gels
were
exposed
to
x-ray
film
for the following
times
with
(+)
or
without
(-)
an
intensifying
screen
:
Fig
.
3
b,
27
h
(-)
;
Fig
.
3
c,
15
h (+)
;
Fig
.
4
b,
36
h
(-)
;
Fig
.
4
c
;
55
h
(+)
;
Fig
.
5
b,
405
h
(-)
;
Fig
.
5
d,
42
h
Samples
for
electrophoresis
were
prepared
as
follows
.
Fig
.
3
:
The
white
cell
layer
(huffy
coat)
was
collected
from
the
first
centrifugation of
heparinized
chicken
blood
and
recentrifuged
twice
in
HEPES-buffered
saline
containing
1
mM
EDTA
to
give
an
erythrocyte-free
preparation
.
The
cells
were
mixed
with
40
vol
of
ethanol
at
0°C,
pelleted,
and
boiled
in
1°k
SDS
sample
buffer
.
Whole
adult
chicken-gizzard
smooth-muscle
tissue
was
frozen
and
pulverized
in
liquid
nitrogen,
thawed
in
ethanol,
pelleted,
and
boiled
in
1%
SDS
sample
buffer
.
Turkey
erythrocyte
membranes
were
extracted
at
0°C
for
20
min
with
10
mM
Tris,
1
mM
EDTA,
l
M
NaCl,
1%
2-mercaptoethanol,
I%
Triton X-100,
and
the
residue
was
pelleted
and
boiled in
1%
SDS
sample
buffer
.
Chicken
erythrocyte
membranes
were
extracted
at
0°C
for
70
min
with
10
mM
Tris,
2
mM
EDTA,
0
.5
M
KCI,
1°k
Triton
X-100
(pH
7
.4),
and
the
residue
was
pelleted
and
boiled in
1%
SDS
sample
buffer
.
Another
aliquot
of
chicken
erythrocyte
membranes
was
washed
at
0°C
with
10
mM
Tris-2
mM
EDTA
(pH
7
.4)
and
then
treated
for
4 h
at
0°C
with
10
mM
Tris-1
mM
EGTA
(pH
7
.4)
;
the
extract
was
dialyzed
against
water,
lyophilized,
and
boiled
in
1%
SDS
sample
buffer
.
Fig
.
4
:
Chicken
erythrocyte
membranes
were
treated
with
60
vol
of
ethanol
to
permeabilize the
vesicles,
pelleted,
and
dissolved
at
room
temperature
in
a
saturated
urea
solution
containing
1%
2-mercaptoethanol
.
Fig
.
5
:
Chicken
erythrocytes
were
extracted
twice
at
0°C
with
1%
Triton
X-
100,
150
mM
NaCl,
2
mM
EGTA,
5
mM
MgC1
2
,
1
mM
sodium
tetrathionate,
1
mM
e-amino-n-caproic
acid,
1
mM
o-phenanthroline,
1
mM
PMSF,
10
mM
Tris-
Cl
(pH
7
.2)
and
rinsed
twice
with
this
solution
without
the Triton
X-100
.
The
resulting
cytoskeletons
were then
extracted
with
this
latter
solution
containing
6
M
urea
for
1
.5
h
at
0°C
.
This
extract
was
dialyzed
against
water,
and
the
resulting
precipitate
was
collected
and
dissolved
in
a
saturated
urea
solution
containing
0.5%
2-mercaptoethanol
.
Whole
adult
chicken-gizzard
smooth-muscle
tissue
was
frozen
and
pulverized
in
liquid
nitrogen,
thawed
in ethanol,
pelleted,
boiled for
30
s
in
101
a1
of I% SDS,
and
then
dissolved
in
90
Wl
of
saturated
urea
containing
1%
2-mercaptoethanol
.
Immunofluorescence
Glass cover
slips
were
pretreated
with
Alcian
Blue
to
promote
erythrocyte
adhesion
(54)
.
Cover
slips
were
cleaned,
simmered
for
5
min
in
0.1%
Alcian
Blue
8GX
(Sigma Chemical
Co
.,
St
.
Louis,
Mo
.),
rinsed
with
distilled
water,
and
air
dried
.
Washed
erythrocytes
in
the
choline
chloride/HEPES
buffer
were
allowed
to
settle
on
the
cover
slips
for
5-10
min
at
room
temperature
;
nonadherent
cells
were
removed
by
rinsing
with
the
same
solution
.
For
the
sonication
experiments,
cover
slips
with
attached
erythrocytes
were
hypotonically
lysed
at
room
temperature
in
Buffer
H
and
then
placed
in
Buffer
F
[130
mM
KCI, 5
mM
NaCl,
l
mM
NaN3,
5
MM
MgC12,
1
mM
EGTA,
20
mM
potassium
phosphate
(pH
7
.5)]
.
Cover
slips
were
laid
face-up
in
a
beaker
of
buffer
F
and
sonicated
for
20
s
at
20
watts
using
a
Braunsonic
1510
sonicater
(B
.
Braun
Instruments,
San
Francisco, Calif) with
a
4-mm
titanium
probe
tip
positioned
3-4
cm
above
the
cells
.
A
glass
rod
was
used
to
hold
the
cover
slips
in
position
on
the
bottom
of
the
beaker during
sonication
.
Incubations
with
antisera
and
subsequent
washes
were
all
performed
at
room
temperature
in
buffer
TM
(0
.5%
Triton X-100, 130
mM
NaCl, 5
mM
KCI, 5
mM
NaN
3
,
5
MM
MgC1
2
,
1
mM
EGTA,
10
mM
Tris-Cl
(pH
7
.5)]
.
Alternatively,
sonication
was
performed
in
Buffer
H
and
antibody
incubations
and
washes
in
Buffer
F
without
the
MgC1
2
and
EGTA
(Fig
.
8
a,b,c,h)
.
To
investigate
the
Colcemid
sensitivity
of
the
filaments,
turkey
erythrocytes
attached
to
cover
slips
were
incubated
for
16
h
at
37°C
in
growth
medium
(Eagle's
minimal
essential
medium,
nonessential
amino
acids,
15%
horse
serum,
5%
chick
embryo
extract,
0
.01%
streptomycin
and
100
U
penicillin/ml)
containing
5
pct
Colcemid
(demecolcine
;
Calbiochem-Behring
Corp
.,
La
Jolla,
Calif)
.
Control
cells
were
treated
identically,
except
that
they
were
not
exposed
to
Colcemid
.
Cover
slips
were
placed
in Buffer
TM
containing
0
.5
mM
PMSF
for
1
min
at
room
temperature
to
make
cytoskeletons,
then
fixed
for
10
min
at
room
temperature
in
Buffer
F
containing
2%
formaldehyde
.
Subsequent
incubations
and
washes were done
in
Buffer
TM
containing
a
0
.5
mM
PMSF
.
Similar
results
were
obtained with
unfixed
cells,
with
cells
incubated
in
suspension
rather
than
attached
to
cover
slips,
and
with
cells
incubated
with
100
pct
Colcemid
.
Rabbit
anti-vimentin
was
prepared
using
antigen
from embryonic
chicken
skeletal
muscle
cytoskeletons
(24)
.
Rabbit
anti-desmin
was
prepared
using
desmin
present
in
a
low-salt
extract
of
chicken-gizzard
smooth muscle
(25)
.
Rabbit
anti-
synemin
was
prepared
using
antigen
obtained
from
chicken-gizzard
intermediate-
filament
proteins
that
had
been
solubilized
and
precipitated
three
times
in
acetic
acid (25)
.
All
antigens
were
ultimately
purified
by
preparative
SDS-PAGE
before
injection
.
Conjugation
of
anti-vimentin
with
rhodamine
B
was performed
as
described
(25)
.
Double
Immunofluorescence
was
performed by
the
indirect/direct
method
(32)
.
Fluorescein-conjugated
goat
anti-rabbit
IgG
was
purchased
from
Miles-
Yeda, Ltd
.
(Rehovot,
Israel)
and
diluted
150-fold
for use
.
Primary
antisera
were
partially
purified
by
precipitation
with
ammonium
sulfate
at
50%
saturation
and
used
at
-1/30
serum
concentration
.
Cover
slips
were
mounted
in
90%
glycerol
in
Tris-buffered
saline
and
photo-
graphed
with
Kodak
Tri-X
film
using
a
Leitz
phase/epifluorescence
microscope
and
filter
modules
K
and
N2
.
The
patches
of
plasma
membrane
in
Fig
.
8s
are
visible
by
phase-contrast
microscopy
because
an
air
bubble
was
present
under
that
portion
of
the
cover
slip
.
Electron
Microscopy
For
thin
sectioning,
chicken
erythrocyte
membranes
were
pelleted
and
fixed
with
1%
glutaraldehyde,
5
mM
MgCl
2
,
10
mM
sodium
phosphate
(pH
7
.5),
postfixed
in
1%
060
4
,
0
.1
M
sodium
cacodylate
(pH
7
.4),
stained
at
60°C
with
1%
uranyl
acetate,
0
.1
M
sodium
maleate
(pH
5
.15),
dehydrated
in
ethanol
and
propylene
oxide,
and embedded
in
a
9/16/1 mixture
of
1,2,7,8-diepoxyoctane
(Aldrich
;
Milwaukee,
Wis
.)/nonenyl
succinic
anhydride
(ICN
K/K
Laboratories,
Inc
.,
Plainview,
N
.
Y
.)/DMP-30
.
The
resin
was
cured
for
2
d
at
60°C
and
sectioned
with
glass
knives
on
a
Reichert
OmIJ2
ultramicrotome
.
Sections
were
stained
for
3
min
with
0
.2%
lead
citrate,
examined
with
a
Philips
EM201
at
80
kv,
and
photographed
on
35-mm
film
.
Shadowed
replicas
were
made
as
follows
:
Chicken
erythrocytes
adhering
to
Alcian
Blue-coated
cover
slips
were
lysed
and
sonicated
as
described
under
Immunofluorescence
.
The
cover
slips
were
immersed
in
buffered
1%
glutaralde-
hyde
followed
by
1%
080
4
,
dehydrated
with ethanol,
dried
in
a
carbon
dioxide
critical-point
drier,
rotary
shadowed
with
platinum/palladium
(80/20)
at an
angle
of
6°,
and
then
carbon
coated
.
Replicas
were
separated
from
cover
slips
with
5%
HF,
washed
with
water,
mounted
on
300-mesh
copper
grids,
and
viewed
by
transmission
electron
microscopy
.
RESULTS
Membrane
Fractionation
and
Ultrastructure
The
study of
minor
protein
components
of
avian
erythrocytes
is
hampered
by
the
relative
abundance
of
hemoglobin
and
chromatin
in
these
cells
.
Removal
of
these
two
components
makes
biochemical,
immunological
and
ultrastructural
char-
acterization
of
the
remaining
structures
easier
.
Hypotonic
lysis
(to
remove
hemoglobin)
and
subsequent
mechanical
enuclea-
tion
(to
remove
chromatin)
gives
a
preparation
of
membranes
that
can
be
studied in a
manner
analogous
to
the
study
of the
simpler
mammalian
erythrocyte
ghosts
.
To
this
end,
chicken
and
turkey
erythrocytes
were
isolated
from
fresh
blood
by
differential
centrifugation
and
lysed
in
a
low
osmolarity
buffer
containing
magnesium
ions
to
keep
the
nuclei
intact
(27)
.
These
nucleated
ghosts
were
then
disrupted
by
passage
through
a
bent
hypodermic
needle,
and
a
membrane
fraction
was
sepa-
rated
from
the
nuclei
and
unbroken
cells
by
differential
cen-
trifugation
.
This
membrane
fraction
is
the
main
object
of
this
study
and
will
hereafter
be
referred
to
as "erythrocyte
mem-
branes
."
Representative
thin
sections
of
the
erythrocyte
membranes
are
shown
in
Fig
.
1
.
The
preparation
is
composed
primarily
of
plasma membranes,
both complete
and
in
pieces
(compare
with
sections
of
whole
cells
in
references
4, 5,
63,
66)
.
Close
exami-
nation
reveals
that in
a given
thin
section
many
of
the
mem-
branes
have
filaments
associated
with
them
.
These
filaments
are
-9
nm
in
diameter
and
are
therefore
classified
as
inter-
mediate
filaments
.
They
are present
on
the
cytoplasmic
side
of
the
plasma
membrane
fragments
and
often
appear
to
be
in
close
apposition
to
the
protein
network
(analogous
to
the
spectrin
network
in
mammalian
erythrocytes)
just
inside the
lipid
bilayer
.
This
is
especially
apparent
in
grazing
sections
of
the
membrane
in
these
relatively
thick
sections
.
The
filaments
are
usually
curved
and
randomly
distributed
and
do
not
exhibit
any
obvious
association
with
specific
cell
structures
.
Thin
sections
of the
free
nuclei
(not
shown)
also
reveal
intermediate
filaments
associated
with
these
structures,
as
pre-
viously
shown
by
Woodcock
(63)
.
It
has not
been
determined
what
proportion
of the
filaments
remains
with
the
membranes
and
what
proportion
with
the
nuclei,
due
to
the
difficulty
of
GRANGER
ET
AE
.
Synemin
and
Vimentin
in
Avian
Erythrocytes
30
1
FIGURE
1
Thin
sections
of
chicken
erythrocyte
membranes
.
(a)
Medium-power
view
of pelleted
membranes
showing
that
filaments
can
be
seen
associated
with
many,
but not
all,
membranes
in
a
given
section
(x
11,000)
.
(b,
c)
High-power
views
.
Note
cytoplasmic
filaments
and
frequent close
associations
between
filaments
and
membranes
(b
:
x
38,000
;
c
:
x
47,000)
.
Bars,
500
nm
.
enucleating
the
ghosts
quantitatively
and
completely
separating
the
resulting
fractions
.
Such
a
quantitation
is
also
complicated
by
the
fact
that
the
proportion
of
the
filaments
associated
with
the
nucleus
before
cell
disruption
that
remains
associated
with
the
nuclei
after
fractionation
is
probably
a
function
of
the
severity
of
the
disruption
procedure
and
the
extent
of
loss
of
the
outer nuclear
membrane
.
However,
based
on
various
bio-
chemical
and
ultrastructural
data
(see
below),
it
appears
that
on
the
average
less
than
half
ofthe
cell's
intermediate
filaments
end
up
in
the
membrane
fraction
.
It
is
apparent
from
these
thin
sections that
the
erythrocyte
membrane
fraction
contains
low
levels
of
contamination
by
fragments
of
structures
other
than
the
plasma
membrane
.
Even
though
this
enucleation
procedure
results
in free
nuclei
that
appear
to
be
intact
by
phase-contrast
microscopy,
the
relatively
fragile
outer nuclear
membrane
may
become
partially
frag-
mented and
fractionate
with
the
plasma
membranes
(27,
65)
.
Fragments
of
mitochondrial
membranes
may
also
be
present
in this
fraction
.
However,
because
our
studies
were
concerned
primarily
with
intermediate
filaments
rather
than
specific
membrane
proteins,
further
purification
of
the
membrane
frac-
tion
was
not
deemed
necessary
for
subsequent
biochemical
studies
.
The
purpose
of
the fractionation
was
to
remove
chro-
matin
that
would
have
physically
interfered
with the
membrane
extraction
experiments,
and
this
was
accomplished
.
Negligible
amounts
of
histone
could
be
seen
when
the
membrane
fraction
was
analyzed
by
SDS-PAGE,
and
nuclear
membrane
lamins
(21,
52)
could
not be detected
by
IEF/SDS-PAGE,
showing
that
the
level
of
contaminating
material
was
low
.
Electrophoretic
Analysis
of
Membrane
Fraction
Analysis
of
the
protein
composition
of avian
erythrocyte
membranes
was
performed
with
regard
to
the
voluminous
work
on
mammalian
erythrocyte
ghosts
.
Similarities
between
the
two
systems
include
two
major
high-molecular-weight
proteins
in
avian
membranes
that
correspond
to
the
mammalian
eryth-
rocyte spectrins
(see
Figs
.
2
and
6)
.
Avian
a-spectrin
comigrates
by
SDS-PAGE
with
mammalian
a-spectrin,
but
the
,Q
variant
has
a
higher
mobility
and can
be
resolved
into
a
closely
spaced
doublet
on
underloaded
gels
(not
shown)
.
Both
systems contain
30
2
THE
JOURNAL
Of
CELL
BIOLOGY
"
VOLUME
92,
1982
actin
at
42,000
daltons
as
well
as
a
broad
band
of
membrane
proteins
around
100,000
daltons
(Band
3
;
reference
16)
.
Among
the
characteristic
differences
are the
presences
in
avian
mem-
branes
of
goblin,
a
hormonally-regulated
phosphoprotein
(4),
and
of the
intermediate
filament
proteins,
vimentin
and
syne-
min
.
The
presence
of
these
two
intermediate
filament
compo-
nents in
association
with
avian
erythrocyte
membranes
was
demonstrated
by
two-dimensional
gel
electrophoresis
(IEF/
SDS-PAGE
;
see
Fig
.
4a)
.
Erythrocyte
vimentin
coelectro-
phoreses
in
this
gel
system
with
vimentin
identified
in
other
avian
cell
types
(19,
24)
;
the
identification
of
synemin
was
tentative
at
this
stage
and
required
immunological
and
bio-
chemical
confirmation,
as
described
below
.
Desmin
was
not
detected
on
these
electrophoretograms
.
Initial
biochemical
studies
of
erythrocyte
membranes
began
with
attempts
to
remove
peripherally
bound
proteins
from
the
membrane
lipid
bilayer
.
We
found
that
the
solubilization
or
release
of
any
protein
components
from
the
erythrocyte
mem-
branes,
without
the use of
detergents
or
strongly
chaotropic
agents,
required
the
removal
of
divalent
cations
.
Therefore,
before
most
biochemical
experiments,
the
magnesium
ions
present
in
the
membrane
suspension
(in
the
hypotonic
lysis
buffer)
were
removed
from
the
membranes
by
washing
with
a
low-salt
buffer
containing
EDTA
.
Treatment
of
membranes
with
solutions
of
very
low
ionic
strength
was
expected
to
release spectrin,
by
analogy
to
the
mammalian
erythrocyte
system
(42,
43)
.
However,
it
was
ob-
served
that if
such
an
extraction
was
performed
briefly
at
0°C,
then
the
primary
protein
released
was
vimentin
.
Fig
.
2 a
shows
a
two-dimensional
gel
of
the
extract
obtained
by
treatment
of
chicken
erythrocyte
membranes
with
distilled
water
for
30
min
on
ice
.
In addition
to
the four or
five
isoelectric
variants
of
vimentin
at
52,000
daltons,
are
synemin
at
230,000
daltons
and
actin
at
42,000
daltons
.
Identification
of the
230,000
dalton
polypeptide
as
synemin
is
based
on
its
immunological
cross-
reactivity
with
smooth
muscle
synemin
and
its
immuncauto-
radiographic
peptide
map,
both
as
detailed
below,
as well as
its
copurification
with
vimentin
.
No
desmin
can
be
detected
on
this
gel
.
Distilled
water
was
found
to
be
the
optimal
solvent
for
extraction
of
relatively
pure vimentin
and
synemin,
but
other
low
ionic
strength
solutions (eg
.,
1-2
mM
EDTA
or
EGTA
FIGURE
2
Water
extracts
of
chicken
erythrocyte
membranes
.
(a)
Two-dimensional
gel
of
a
distilled
water
extract
of
erythrocyte
membranes
showing
predominance
of
vimentin
(V)
and
synemin
(S)
.
The
E
and
y
variants
of
actin
(A) are
present
.
(b)
One-
dimensional
gel
;
extract
as
in
(a)
was
centrifuged
for
1
h
at
240,000
g
(av
.)
;
lane
1
contains
the
supernatant
and
lane
2
the
pellet
.
For
comparison,
lane
3
shows
a
preparation
of
gizzard
desmin
(D)
and
synemin
that
was
cycled
twice
by
soiubiiization
and
precipitation
using
acetic
acid
.
(c)
Erythrocyte
membrane
residue
after
two
sequential
extractions
with
distilled
water
over
2-h
period
.
G
goblin
(see
Fig
.
6
for
documentation
of
this
identification)
; 1
and
2,
a
and
/3
spectrin,
A,
actin
.
(d,
e, f)
Sequential
water
extracts
of
a
single
aliquot
of
erythrocyte
membranes
;
extraction
periods
were
for
1
min,
0
.5 h
and 9
h,
respectively,
and each
gel
represents
all
of
the
protein
extracted
at
each
step
.
All
membrane
extractions
in this
figure
were
performed
at
0°C
.
Panel
a
represents
the
protein
extracted
from 35-40
pi
of
packed
erythrocyte
membranes,
and
lane
1
from
20-25
fil
of
membranes
.
brought
to
pH
7
.4
with
Tris)
will
produce
extracts
of
compa-
rable purity
but
lower
yield
(approximately
one-half)
.
If
the
pH
of
the
extracting
solution
is
raised,
release
of
several
components
in
addition
to
vimentin
is
favored,
resulting
in
a
less
pure
preparation
of
vimentin
and
synemin
.
At
pH
11,
the
membranes
are
nearly
quantitatively stripped
of
non-integral
proteins
(not
shown
;
see
reference
56)
.
Similarly,
many
proteins
can
be
released
from
the
membranes
by
treatment
with
acetic
acid
(cf
.
reference
41),
but
this
results
in
a
low-yield,
very
impure
preparation
of
vimentin
and
synemin
that
exhibits
extensive
proteolytic
degradation
.
Acetic acid thus
appears
to
be undesirable
for
use
in
the
extraction
and
purification
of
intermediate
filaments
from
erythrocytes
as
it
has
been
used
previously
for
smooth
muscle
(29,
53)
.
This
technique
of obtaining
highly
enriched
preparations of
vimentin
and
synemin
takes
advantage
of
the
fact
that
spectrin
and
actin
are
released
very
slowly
from
the
membranes
distilled
water
or
low
salt
buffers
at
0°C
.
Fig
.
2d
-f
show
sequential
extracts
of
a
membrane
aliquot
made
with
distilled
water
at
0°C
for
1
min,
30
min
and
9
h
.
Most
of
the
vimentin
and
synemin
are
released
from
the
membranes
within
I
min
.
The
ratio
of
actin
to
vimentin
increases
with
each
extraction,
and
spectrin
becomes
a
major
component
of
the
extract
after
a
few
hours
.
After
prolonged
extraction,
though,
even
in the
presence
of
EDTA
and
reducing
agents,
most of
the
spectrin
(and
much
of
the
actin)
is
still
associated
with
the
membrane
(not
shown)
.
If
the
extractions
are
performed
at
37°C
instead
of
0°C,
all
of
these
proteins
are
released
more
rapidly
;
spectrin
is
solubilized
rapidly
enough
to
become
the
major
component
of
short-interval
extracts,
making
vimentin
a
minor
component
(not
shown)
.
After
repeated
and
prolonged
extractions
with
distilled
water
at
0°C,
the
membranes
tend
to
break
up
into
smaller
fragments
and
vesicles
.
Partial
loss
of
the
spectrin
network
may
account
for
this
.
Lane
1
of Fig
.
2b
is
the
one-dimensional
gel
profile
of
a
distilled
water
extract identical
to
that
in
Fig
.
2a,
except
that
as
centrifuged
for
1
h
at
170,00-310,000
g
(240,000
g
av
.)
;
the
supernatant
was
lyophilized
and
run
in
Lane
1,
and
the
pellet
was
run
in
Lane
2
.
It is
apparent
that
under
these
conditions
little
of
the
protein
is
sedimented
.
A
similar
result
was
obtained
with
a 5-h
centrifugation
.
Minor
polypeptides
that
do
not
focus
discretely
in the
two-dimensional
gel
system
can
be
visualized
here
.
Electrophoresis
of
this
material
on
less
porous
polyacrylamide
gels
shows
that
most of
the material
migrating
with the
dye
front
in
Lane
1
is
residual
hemoglobin
.
Lane
3
is
a preparation
of
chicken-gizzard
smooth-muscle
intermediate
filaments
that
shows
synemin,
vimentin,
desmin,
and
actin
for
molecular
weight
comparison
.
Synemin
and
a
spectrin
are
difficult
to
resolve
from
one
another
on
normally
loaded
one-dimensional
gels
but are
clearly
resolved
on
under-
loaded
gels
and
two-dimensional
gels
.
Not
all
of the
vimentin
is
extracted
from
erythrocyte
mem-
branes
with
a
single
distilled
water
treatment
.
Fig
.
2e
shows
a
sample of
membranes
that
was
extracted
twice
with
water
over
a
2-h period
;
the
amount
of
vimentin
is less,
but
some
still
remains
.
Residual vimentin
is
evident
even
after
four
extrac-
tions
over a 5-d period (not
shown)
.
However,
in
thin
sections
of
membranes
treated
with
distilled
water
for
5
min,
cytoplas-
mic
intermediate
filaments
cannot
be
found
.
This
suggests
that
the
residual
vimentin
may
not
be
in
the
form
of
free
cytoplas-
mic
filaments
(see
Discussion)
.
GRANGER
ET
AL
.
Synemin
and
Vimentin
in
Avian
Erythrocytes
303
Immunological
Characterization
of Erythrocyte
Intermediate Filaments
The
technique
of
immunoautoradiography
(10),
which
uses
antibodies
to
detect
protein
antigens
in
polyacrylamide
gels,
was
used
in this
study
for
three
purposes
:
(a)
to
determine
whether
the
erythrocyte
intermediate-filament
subunits
were
antigenically crossreactive
with
their
muscle
counterparts
;
(b)
as
a
form
of peptide
map
analysis
to
determine
whether
the
subunits
in
erythrocytes
were
homologous
or
identical
to
their
muscle
counterparts
;
and
(c)
to
detect these
antigens
in
gels
with
a
sensitivity
much
greater
than
that
afforded
by
Coomassie
Blue
staining
.
Antisera
used
in this
study
were
all
elicited
against
chicken
muscle
proteins,
purified
by
SDS-PAGE,
and
each
appears
to
be
specific
for
its
respective
antigen
as
assayed
by
two-dimensional
immunoautoradiography
(24,
25)
.
Fig
.
3
shows
the presence
of
immunoreactive
forms
of
both
vimentin
and
synemin
in
various
fractions
of
chicken
and
turkey
erythrocyte
membranes
.
Fig
.
3
a
shows
a
Coomassie
Blue-stained
SDS-polyacrylamide
gel
of
a
variety
of
samples
;
this
gel
was
labeled
with
anti-synemin
followed
by
radioiodi-
nated
protein
A,
and
the
corresponding
autoradiogram
is
shown
in
Fig
.
3 c
.
A
duplicate
gel
was
processed
with
anti-
vimentin,
and
its
autoradiogram
is
shown
in
Fig
.
3
b
.
Lane
1
represents
whole
white
cells
from
chicken
blood,
examined
to
ensure
that
the
vimentin
and
synemin
being
studied
in the
erythrocyte
preparation
were
not
originating
from
the
extremely
low
level
of
contamination
by
white
cells
.
Little
if
any
synemin
is
detectable,
and
the
quantity
of
vimentin
is
low
relative to
the
amount
of
actin
present
in
these
cells
.
Lane
2
is
whole
chicken-gizzard
smooth-muscle
tissue
.
It is
the
tissue
from
which
synemin
was
originally
purified
and was
the
source of
the
synemin
used
for
immunization
(25)
.
Vimen-
tin
is
also
present
in
this
tissue
(25
;
Fig
.
3
b)
.
The
two
most
30
4
prominent
bands
near
the
top
of
the
lane
are filamin
and
myosin
;
the
autoradiogram
of
Fig
.
3
c
shows
that
synemin
migrates
between
these
two
proteins
(see also
reference
25)
.
a-Actinin
is
visible
at
100,000
daltons,
and
the
two
major bands
in
the
middle
are
desmin
and
actin
(50,000
and
42,000
daltons)
.
The
remaining
lanes
demonstrate
the
presence
of
vimentin
and
synemin
in
erythrocyte
membranes,
membrane
cytoskele-
tons,
and low
salt
extracts
.
Lanes
3
and
8
display
chicken
erythrocyte
membranes,
and
lane
4
displays
turkey
erythrocyte
membranes
.
Both
samples
contain
polypeptides
that
have
mo-
lecular
weights
and
antigenicities
similar
to
those
of
muscle
vimentin
and
synemin
.
Lanes
5
and
6
represent
turkey
and
chicken
erythrocyte
membranes
that
have
been
extracted
with
high
salt
and
Triton
X-100
.
Both
vimentin
and
synemin
remain
insoluble
in
the
cytoskeletal
residue, as
do
the
spectrin
and
actin
.
Goblin,
the
100,000
dalton
cluster,
a
44,000
dalton
polypeptide
and
many
minor components
are
partially
or
completely
solubilized
(compare
lanes
S
and
6 with
lanes
3
and
4)
.
Lane
7
contains
a
low-salt
extract
of chicken
erythrocyte
membranes
that
is
highly
enriched
in
vimentin
and
synemin
;
the
greater
amount
of
handling
and
processing
of
these
samples
relative to
the other
samples
on
the gel
probably
accounts
for
the
increased
quantity
of
proteolytic
fragments
of
vimentin
and
synemin
evident
in
these lanes
of
the
autoradiogram
.
Anti-desmin
does
not
label
any
of the
proteins
in
these
samples,
except
for
the
desmin
present
in
the
gizzard
tissue
(not
shown)
.
Fig
.
4
a
depicts
chicken
erythrocyte
membrane
proteins
re-
solved
by
IEF/SDS-PAGE
.
This
gel
was
processed in
immu-
noautoradiography
with
anti-vimentin,
and
the
resulting
au-
toradiogram
is
shown
in
Fig
.
4
b
.
An
identical gel
was
processed
with
anti-synemin,
and
its
autoradiogram
is
in
Fig
.
4
c
.
Proteins
readily
identified
by
Coomassie
Blue
staining
include
goblin,
a
and
,8
spectrin,
synemin,
the
multiple
isoelectric
variants
of
FIGURE
3
Immunoautoradiography
using
antivimentin
and
antisynemin
.
(a)
Polyacrylamide
gel
stained
with
Coomassie
Blue
after
labeling
with
antisynemin
and
radioiodinated
protein
A
;
(b)
Autoradiogram
of
duplicate
gel
labeled
ith
anti-vimentin
;
(c)
Autoradiogram
of
gel
in
(a)
.
Samples
are
from chickens
unless
otherwise
noted
.
Lane
1
:
Buffy
coat
.
The
identity
of
the
prominant
labeled
polypeptide
just
above
the
dye
front
is
not
known
.
It
labels
comparably
with
both
antisera
and
may
therefore
be an
IgG
receptor
or
protein
A
receptor
that
survives
SDS
denaturation
and
acetic
acid/ethanol
fixation
.
Lane 2
:
Whole
gizzard
muscle
.
Lane
3
:
Erythrocyte
membranes
.
Lane
4
:
Turkey
erythrocyte
membranes
.
Lane 5
:
High
salt
plus
detergent
residue
of
turkey
erythrocyte
membranes
.
Lane
6
:
High
salt
plus
detergent
residue
of
erythrocyte
membranes
.
Lane
7
:
Low
salt
extract
of
erythrocyte
membranes
.
Lane
6
:
Erythrocyte
membranes
.
A,
actin
;
V,
vimentin
.
Synemin
is
the
major
high
molecular
weight
protein
in
lane
7
;
it
migrates
just
beneath
a
spectrin
.
THE
JOURNAL
OF
CELL
BIOLOGY
"
VOLUME
92,
1982
FIGURE
4
Two-dimensional
immunoautordiography
of
chicken
erythrocyte
membranes
using
antivimentin
and
antisynemin
.
(a)
Two-dimensional
gel
of
chicken
erythrocyte
membranes
stained
with
Coomassie
Blue
after
labeling
with antivimentin
and
radioiodinated protein
A
.
(b)
Autoradiogram
of
same
gel
.
(c)
Autoradiogram
of
duplicate
gel
labeled
with
antisynemin
.
A,
actin
;
V,
vimentin
;
S,
synemin
;
1,
a
spectrin
2,,Q
spectrin
;
G, goblin
.
vimentin,
and
nearly
equal
amounts
of
a
and
y
actin
.
Anti-
vimentin
and
anti-synemin
label
only
their
respective
proteins
on
the
gel
and
do
not
crossreact
with
other
proteins
in
this
system
.
Desmin
is
not
detectable
by
Coomassie
Blue
staining,
nor
by
immunoautoradiography
with
anti-desmin
(not
shown)
.
The
diagonal
string
of polypeptides
smaller
and
more
acidic
than
vimentin,
visible
in
Fig
.
4b,
represents
breakdown
prod-
ucts
of
vimentin
(19,
24)
;
the
same
probably
holds
true
for the
numerous
polypeptides
under synemin
that
label
with
anti-
synemin
.
Because
the
vimentin
and
synemin
used
for
immu-
nization
were
excised
from an
SDS-polyacrylamide
gel,
it
seems
unlikely
that
all
of
these
minor
polypeptides
could
be
unrelated
contaminants
.
The
relative
quantity
of
these
other
polypeptides
increases
with
increased
processing of the
sam-
ples
.
They
can
be
reduced
in
amount
or
completely
eliminated
if
special
precautions
are
taken
to
inhibit
proteolytic
enzymes
(19,
25)
.
Also,
the
same
patterns
are
seen
in
samples
from
different
tissues
(references
19,
24, 25,
and
below)
.
To
determine
the
antigenic
homology
of
erythrocyte
synemin
and
gizzard
synemin,
a
fortuitous
form
of peptide
mapping
was
used
.
Fragments
of
synemin
generated
by
endogenous
proteases
during
processing
of the
tissues
were
detected
with
antibodies
to
synemin
as
visualized
by
two-dimensional
im-
munoautoradiography
.
An
antiserum
specific
for
a given
pro-
tein thus
allows
visualization
of
the protein's
peptide
map
without
the
necessity
of
prior
purification
of
that
protein
.
The
degradation
pattern
of
erythrocyte
vimentin
as seen
in
Fig
.
4
b
is
similar
to
that
already
published
for
muscle
vimentin
(24)
.
Fig
.
5
compares
the
synemin
present
in
chicken
erythrocyte
cytoskeletons
and
that
in
chicken
gizzard
smooth
muscle
tissue
.
Fig
.
5 a
shows
the
proteins
that
remain
after
extraction
of
erythrocytes
with
1%
Triton
X-100
in
a
physiological
salt
buffer
;
a
major
difference
in
the
protein
pattern,
compared
with
the
membranes
in
Fig
.
4a,
is
the
presence
of
the
nuclear
lamins
.
There
is
also
more
vimentin
relative to
the
amount
of
actin
present,
as
this
preparation
also
includes
the
nucleus-
associated intermediate filaments
.
The
autoradiogram
of
this
gel
after
processing
with
anti-synemin
is
shown
in
Fig
.
5 b
.
Similarly,
a
two-dimensional
gel
of
whole
gizzard
tissue
proc-
essed with
anti-synemin,
and
its
autoradiogram,
are
shown
in
Fig
.
5 c
and
d
.
The
similarities
in
the
two
synemin
patterns
are
striking
(see
also
Fig
.
4c)
;
in
each
case,
the parent
molecule
is
most
heavily
labeled,
and
the
arcs
of
daughter
products
ter-
minate
in
what
appears
to
be a
particularly
stable
fragment
at
-34,000
daltons
(pI
~4
.9)
.
We
have
noted
no
consistent
differences
between
erythrocyte
synemin
and
smooth
muscle
synemin
.
Minor
differences
in
the
fragment
patterns
may
be
attributable
to
different
endogenous
proteases,
different
processing
schemes,
or
slight
differences
in
electrophoresis
;
the
latter
two
would
explain
the
very
minor
differences
between
the
erythrocyte
synemins
in Figs
.
4
c
and
5
d
.
There
is
a
slight
variation
in
the
observed
isoelectric
points
of
erythrocyte
and
smooth-muscle
synemin,
but
there
is
varia-
tion
even
among
different
samples
of
erythrocyte
synemin
(compare
Figs
.
2 a
and
4
a)
.
The
focusing
of
synemin seems
to
be
influenced
by
the
amount
of
protein
loaded
on
the
isoelectric
focusing
gel
;
the
apparent
isoelectric
point
of
synemin
is
often
the
same
as
that
of desmin
or
vimentin
if
either
of the
latter
is
present
in
large
quantities
on
the
gel
(Fig
.
2 a
and
reference
25)
.
We
conclude
from
these
immunoautoradiographic
data
that
erythrocyte
synemin
and
muscle
synemin
are
similar
if
not
identical
;
similarities
in solubility
properties
and
cellular
dis-
tribution
(below)
strengthen
the
conclusion
that
the
erythrocyte
polypeptide
may
be
regarded
as
synemin
as defined
previously
in
smooth
muscle
.
Phosphorylation
Goblin
is
a
high
molecular
weight
protein
of
the turkey
erythrocyte
plasma
membrane
characterized
by
hormone
de-
pendent
phosphorylation
(4)
.
Both
goblin
and
synemin
have
reported
molecular
weights
of
230,000
daltons
;
although
their
solubility
properties
and
distributions
appeared
to
differ,
we
thought
it
was
necessary to
conclusively
determine
whether
goblin
and
synemin
were
indeed
different
proteins
.
We
iden-
tified
goblin
by
its
characteristic
properties
of
being
a
large
membrane-associated
protein
and
the
most
hyperphospho-
rylated
polypeptide
in
turkey
erythrocytes
treated briefly
with
the
,(3-adrenergic
agonist,
isoproterenol
(4)
.
Fig
.
6
a
shows
a
Coomassie
Blue-stained
gel
of
membranes
of
turkey
erythro-
cytes
labeled
with
["P]
inorganic
phosphate
;
those
on
the
left
were
also
treated
with
isoproterenol,
whereas
those
on
the
right
were
not
.
Fig
.
6 b
is
the
corresponding
autoradiogram
.
By
the
above
criteria,
we
conclude
that
the
band
designated
in
the
figure
is
goblin
.
In
this
gel
system,
goblin migrates
more
slowly
than
the
two
spectrin
variants,
rather
than
migrating
between
them
as
in
the
system
of
Beam
et
al
.
(4)
.
Using
their
gel
system,
we
found
that
the
electrophoretic
pattern
of
our
samples
was
indeed
different
:
the
relative
positions
of
some
bands
was
different,
and
goblin
and
the spectrins
were
not
resolved
as
well
.
A
44,000 dalton polypeptide
is
also
noticeably
hyperphos-
GRANGER
ET
AL
.
Synemin
and
Vimentin
in
Avian
Erythrocytes
30
5
FIGURE
5
Comparison
of
erythrocyte
and
gizzard
synemins by
two-dimensional
immunoautoradiography
.
(a, c)
Two-dimensional
gels
stained
with
Coomassie
Blue
after
labeling
with
antisynemin
and
radioiodinated
protein
A
.
(b,
d)
Corresponding
autoradi-
ograms
.
Samples
are
(a)
a
Triton
X-100
insoluble
cytoskeleton
of
chicken
erythrocytes,
and
(c)
whole
gizzard
smooth
muscle
tissue
.
A,
actin
;
D,
desmin
;
V,
vimentin
;
L,
lamins
A,
8,
and
C
(references
21 and 52)
;
M,
myosin
;
5,
synemin
;
F,
filamin
.
phorylated
in
the
presence
of
isoproterenol
(4),
and
this
hor-
mone-dependent
phosphorylation
seems
to
apply
to
a
lesser
degree
to
several
other
proteins
as
well
.
Vimentin
appears
to
be
one
of
these,
because
a
slight
increase
in
labeling
of
this
protein
with
isoproterenol
can
be
detected
on
the
autoradi-
ogram
(Fig
.
6
b)
.
This
is
consistent
with the
hormone-depend-
ent
phosphorylation
of
vimentin
that
has
been
observed
in
other
cell
types
(D
.
L
.
Gard
and
E
.
Lazarides,
manuscript
submitted
for
publication)
.
Synemin
migrates
too
close
to
a
spectrin
on
one-dimensional
gels
to
be
able
to
determine
whether
it
is
phosphorylated,
but
two-dimensional
gels
of
erythrocytes
phosphorylated
to
a steady
state
(1)
suggest
that
synemin
is
indeed
phosphorylated
(not
shown)
.
These
gels also
show
that
all
but
the
most
basic
isoelectric
variant
of
vimentin
are
phosphorylated,
as
is
the
case
with
vimentin
in
other
cell
types
(46,
47)
.
Localization
of
Synemin
and
Vimentin
by
Immunofluorescence
With
antibodies
specific
for
synernin
and
vimentin,
it
was
possible
to
determine
the
spatial
distributions
of
these
antigens
within the
avian
erythrocyte
by
immunofluorescence
.
The
abundance
of hemoglobin
in
these
cells
was
not
a
problem
because
synemin
and
vimentin
remained
insoluble
after
re-
moval
of
the
hemoglobin
by
hypotonic
lysis
or
detergent
lysis
.
Typically,
erythrocytes
were allowed
to
attach
to
Alcian
Blue-
coated
glass
cover
slips
(54),
then
lysed with a
physiological
salt
solution
containing Triton
X-100
and
magnesium
ions,
and
processed
for
examination
by
immunofluorescence
microscopy
without
any
fixation
.
The
result
with both
anti-vimentin
and
30
6
THE
JOURNAL
OF
CELL
BIOLOGY
"
VOLUME
92,
1982
anti-synemin
was
a cytoplasmic
network
of sinuous
filaments
extending
from
the
nucleus
to
the
plasma
membrane
.
This
network
gave
the
impression
of
being
composed
of
a
small
number
of
long
filaments,
since
relatively
few
free
ends
could
be
seen
.
A
high
concentration
of
filaments
was
often
noted
near
the
poles of the
nuclei
(Fig
.
7
d)
.
However,
the small
size
and
relatively
great
depth
of
field
of
the
cells,
coupled
with
the
dense
packing
of
the
filaments,
prevented
adequate
resolution
of
the
individual
filaments
after
photographic
reproduction
of
the
network
.
Nevertheless,
this
technique
of
immunofluores-
cence
allowed
us
to
determine
the
effects
of
Colcemid
on
the
filaments
in
these
cells
.
Colcemid
has
been
found
to
cause
an
aggregation
and
bundling
of intermediate
filaments
in
a
variety
of
cultured
cell
types
(8,
18,
23,
31, 33,
38)
.
When
turkey
erythrocytes
were
incubated
with
Colcemid
under
conditions
that
are
normally
used
for
cultured
cells,
their
intermediate
filaments
did
not
aggregate
.
Chicken
embryo
fibroblasts
incubated
in
the
same
petri plate
showed
normal
filament
aggregation
.
Fig
.
7
shows
the
distribution
of
the
erythrocyte
intermediate
filaments
after
this
incubation,
as revealed
by
immunofluorescence
using
anti-
vimentin
.
The
top
row
depicts
cells
that
received
Colcemid,
and
the
bottom
row
shows
control
cells
that
were
treated
identically,
except
that
they
received
no
Colcemid
.
After
in-
cubation
with
or
without
Colcemid,
the
cells
were
briefly
lysed
with
Triton
X-100
to
remove
hemoglobin,
then
fixed
with
formaldehyde
to
ensure
preservation
of filament
distribution
.
There
was
no
obvious
Colcemid-induced
aggregation
of
the
filaments,
and
what
appeared
to
be
individual
filaments
could
clearly
be
seen
extending
to
the
plasma
membrane
.
No
con-
sistent
difference
between
the
Colcemid-treated
cells
and
con-
FIGURE
6
Phosphorylation
of turkey
erythrocyte
plasma
membrane
proteins
:
identification
of
goblin
.
Erythrocytes
were
incubated
for
4
h
with
31p_
phosphate
;
half
were
treated
with
isoproterenol
for
the
final
20
min
.
Plasma
membranes
were
isolated
and
the
proteins
were
resolved by
SDS-PAGE
.
The
Coomassie
Blue-stained
gel
was
dried
(a)
and
auto
radiographed
(b)
.
Goblin
(G),
the
high-molecular-
weight
protein
that
is
most
noticeably
hyperphosphorylated
in
the
presence
(+) of isoproterenol,
is
distinct
from
synemin,
which
runs
just
beneath
a
spectrin
(1)
.
Abbreviations
:
G, goblin
;
1
and
2,
a
and
i3
spectrin
;
V,
vimentin
;
A,
actin
;
+
and
-
refer
to
the presence
and
absence
of
isoproterenol
in
the
incubation
mixture
.
trols
could
be detected
.
(This
filament
distribution
is
the
same
as
in
erythrocytes
that
have
not
been
subjected
to
an
in
vitro
incubation
but
examined
soon
after
removal
from
the
animal
.)
Similar
results
were
obtained
with
anti-synemin
.
Unfixed
cells
exhibit
a
similar
filament
distribution,
but
the
nuclei
shrink
somewhat
and become
more
dense
during
processing
.
High
concentrations of
Colcemid
(100
t,M
rather
than
5
/
.M)
also
had
no
effect
on
the intermediate
filaments,
and
cells
treated
in
suspension
rather
than
after
attachment
to
cover
slips
were
similarly
unaffected
.
To
examine
the
association
of
the
filaments
with
the
mem-
brane
and
allow
visualization
of
the
filament
network
more
clearly,
a
technique
was
developed
that
removed
the
erythro-
cyte
nuclei
and
most of
the
filaments
to
leave
a
residual
mat
of
membrane-associated
filaments
in
which
the
individual
strands
could
be
resolved
and
photographed
.
This
technique
is
based
on
ideas
that
stemmed
from
a
number
of
sources
(3,
26, 35, 44,
54)
.
It
involves attaching
erythrocytes
to
Alcian
Blue-coated
glass
cover
slips,
then
disrupting
the
cells
by
cavitation
(with a
sonicator)
forcefully
enough
to
remove
nuclei
and
other
cellular
structures
not
firmly
anchored
to
the cover
slip
.
This
results
in
a
cover
slip
covered
with
residual
elliptical
patches of
mem-
brane,
each with
its
most
firmly
associated
structures
.
Immu-
nofluorescence
shows
that
these
patches
often
have
vimentin-
and
synemin-containing
filaments
attached
to
them
.
This
was
the
first
good
indication
that
intermediate
filaments
might
in
some
way
be
physically
anchored
to
the
erythrocyte
plasma
membrane
.
Fig
.
8 is
a
montage
of
fluorescence
micrographs
showing
the
presence
and
distribution
of
vimentin
and
synemin
on
these
residual
patches of
membrane
.
Tangled
and
wavy
networks
of
filaments
and
fragments
of
filaments
can
be
visu-
alized
.
The
uniform-diameter
filaments
probably
represent
individual
intermediate
filaments
rather
than
bundles
(see
electron
microscopic
correlates
in
Fig
.
9)
.
Their
measured
diameter
of
200-300
nm
is
close
to
the
resolution
limit
of the
light
microscope
and
consistent
with
the
immunofluorescence
image
of
individual
microtubules
(49)
.
Anti-desmin
gives
neg-
ligible
fluorescence,
which
is
consistent
with
our
inability
to
detect
it
electrophoretically
.
Synemin
and
vimentin
preimmune
sera also give
negligible
fluorescence,
and
preadsorption
of
the
antisera
with
the
corresponding
purified
proteins
has
been
shown
previously
to
block
fluorescence
(25)
.
The
quantity of
filaments
remaining
on
the
membrane
patches
is
probably
a
function of
the
degree
of
sonication
.
Sonication
was
monitored
by
phase-contrast
microscopy
and
performed
at
a
level
that
removed
most
of the
nuclei
.
This
produced
a
wide
range
of anucleate
membrane
patches
;
most
patches
retained
no
intermediate
filaments
(see
Fig
.
8
s),
whereas
the
rest
displayed
patterns
ranging
from
short
frag-
ments
(Fig
.
81
p)
to
complex
networks
of
filaments
(Fig
.
8 a-
d)
.
Fig
.
8
s is
a
combination
phase-contrast/fluorescence
mi-
crograph
showing
bare
membrane
patches
as
well
as
filament-
containing patches
.
It is
noteworthy
that
the
remaining
fila-
ments
are not
always
distributed
uniformly
over
the
membrane
patch
.
Double
immunofluorescence
was
performed
to directly
com-
pare
the distributions
of
vimentin
and
synemin
on
these
patches
.
Fig
.
8 q
and
r
show
that
the distributions
are the
same
.
Both
antigens
appear
along
the
same
filaments,
and
nowhere
at
this
level
of
resolution
is
one
antigen present
and
the other
absent
.
However,
the
synemin
fluorescence
sometimes
gives
the
impression
of
being
slightly
punctate
along
filaments
that
show
uniform
vimentin
fluorescence
.
We
have
obtained
comparable
immunofluorescence
results
with
cells
sonicated
in
low
ionic
strength
and
physiologic
salt
buffers,
by
following
sonication
with
a
detergent/high-salt
extraction,
and
with
subsequent
incubations
and
rinses
in
the
presence
or
absence
of Triton
X-100
.
In
none
of the prepara-
tions
shown
in
Fig
.
8
were
the
samples
fixed
with
protein
crosslinkers
or
denaturants
.
Ultrastructure
of
Sonicated
Membranes
Sonicated
erythrocyte
membrane
patches
on
glass
cover
slips,
similar
to
those
used
for
immunofluorescence,
were
fixed
and
rotary-shadowed
for
examination
by
transmission
electron
microscopy
.
Fig
.
9
shows
portions
of
three
such
membrane
patches
with
their
adherent
filaments
.
The
pattern
of
the
filaments
is
similar
to
the
pattern
seen
in
immunofluorescence
;
the
filaments
tend
to
be
relatively
long,
and
can
be
straight,
wavy
or
curved
into
loops
.
Occasionally,
the
filaments
extend
beyond
the
edge
of
the
membrane
patch (Fig
.
9 b
and
c),
presumably
as a
result
of the
sonication
.
Similar
patterns
are
obtained
if
the
erythrocytes
are
sonicated
in
a
physiologic
salt
buffer
without
previous
hypotonic
lysis,
showing
that
the
fila-
ments
are
not
precipitated
on
the
membranes
as
a
result
of
the
low-salt
treatment
.
Treatment
with
Triton
X-100
also
does
not
GRANGER
ET
AE
.
Synemin
and
Vimentin
in
Avian
Erythrocytes
30
7
FIGURE
7
Distribution
of
vimentin
in
Colcemid-treated
erythro-
cytes
.
Phase
(a,
c)
and
corresponding
fluorescence
micrographs
(b,
d)
of turkey
erythrocytes
incubated
in
culture
medium
with
(a,
b)
or
without
(c,
d)
Colcemid
.
Intermediate
filaments
were
visualized
in
fixed
cytoskeletons
by
indirect
immunofluorescence
using
anti-
vimentin
.
Similar
patterns
were
obtained with
antisynemin
.
Bar,
10
ttm
.
x
1350
.
affect
this
pattern
.
Strands
of
chromatin
originating
from
the
erythrocyte nuclei
have
a
distinctive
morphology
and
are
evi-
dent
in
the
replicas
only
when
magnesium
ions
are
omitted
from
the
lysis
or
sonication
buffers
.
The
relatively
flat,
coarse,
upper
face
of
the
membrane
patch
probably
represents
the
spectrin
network
that
lines
the
cytoplasmic
surface
of the
plasma
membrane
.
The
intermediate
filaments
are
of
fairly
uniform diameter
;
it
is
not
known
whether
the
frequent,
slight
bulges,
constrictions,
and
discontinuities
have
a
molecular
basis
or
are
just
an
artifact
ofthe
shadowing
procedure
(cf
.
reference
28)
.
Detail
cannot
be resolved
sufficiently to
determine
whether
the
filaments
branch
or
merely
associated
laterally
in
various
places
.
We
have
not
been
able
to
positively
identify
structures
that
might
anchor
the
filaments
to
the
membrane
patch,
and
it
is
not evident
from
these
micrographs
how
abundant
such
linkers
might
be
.
From
the
distribution
of
the
filaments
on
the
sonicated
membranes,
it
appears
likely
that
they
are
not
ran-
domly
distributed
over
the
plasma
membrane
.
DISCUSSION
Intermediate
Filaments
in
Avian
Erythrocytes
Avian
erythrocytes
provide a
simple system
for the
study
of
30
8
THE
IOURNAL
OF
CELL
BIOLOGY
"
VOLUME
92,
1952
intermediate
filaments
.
These
cells
are
terminally
differen-
tiated,
nondividing,
nonadherent,
and
nonmotile
.
They
are
easily
obtained
and
purified
.
Their cytoplasms
are
relatively
simple
and
nondynamic
.
Interaction
of
intermediate
filaments
with both
the nucleus
and
plasma
membrane
can
thus
be
examined
without
the
variability
and
complexity
inherent
in
most
other
cell
types
.
Comparison
of
the
avian
erythrocyte
with the well
characterized
mammalian
erythrocyte
provides
insight
into
the
functions
and
relationships
between
compo-
nents
that
are not
common
to
both,
such
as
nuclei,
microtubules
and
intermediate
filaments
.
In
this
study
we
demonstrate
that
vimentin
and
synemin
are
the
major
components
of
avian
erythrocyte intermediate
fila-
ments
.
These
filaments
had
previously
been
postulated
to
be
composed
of
vimentin
(63),
and
vimentin
indeed
appears
to
be
their
major
subunit
.
Vimentin
and
synemin
appear
to
coexist
uniformly
in
the
filaments,
which
form
a
looping,
intertwined
network
in the
cytoplasm
.
A
portion
of
this
filament
network
is
associated
with
the
plasma
membrane
firmly
enough
to
resist
detachment
by
physical
disruption
forces
that
are
sufficient
to
remove
the
nucleus
and
fragment
the
plasma
membrane
.
How-
ever,
synemin
and
vimentin
can
be
selectively
released
from
the
plasma
membrane
by
treatment
with
divalent
cation-free,
low
ionic
strength
solutions
.
We
have used
chicken
and
turkey
erythrocytes
as
represent-
atives
of
avian
erythrocytes
in
general
for
this
work
.
Immu-
nofluorescence
was
performed
on
turkey
erythrocytes,
because
they
are
slightly
larger
than
chicken
erythrocytes
.
The
protein
goblin
had
been
defined
in
turkey
erythrocytes
(4),
so
we
used
these
cells
for the
phosphorylation
experiments
.
Biochemical
studies
were
based
primarily
on
chicken
erythrocytes,
which
were
more
readily
available
than
turkey
erythrocytes
;
electron
microscopy
of
thin sections
was
performed
on
chicken
eryth-
rocytes
to
correlate
with the
biochemistry
.
However,
it
is
evi-
dent
from
electrophoretograms
such
as in
Fig
.
3
a
that
chicken
and
turkey
erythrocytes
are not
identical
.
For
example,
turkey
erythrocyte
membranes
have
a
smaller
R-spectrin
and
less
protein
in
the
100,000 dalton
cluster
and
44,000
dalton
band
than
chicken
erythrocyte
membranes
.
Vimentin
and
synemin,
though,
appear
to
be
very
similar,
if
not
identical,
as
judged by
IEF/SDS-PAGE
and
immunoautoradiography
(not
shown)
.
We
believe, therefore, that
the
generalizations
made
about
vimentin
and
synemin
in this
paper
are
likely to
be
applicable
to
all
avian
erythrocytes
.
Studies
of
Intermediate
Filaments
Associated
with the
Plasma
Membrane
Our
studies
of
the
properties
of
erythrocyte intermediate
filaments
have
focused
on
those
filaments
that
fractionate
with
the
plasma
membrane
.
They
have
the
same
antigenicity
and
electrophoretic
mobility as
the
filaments
that
fractionate
with
the
nuclei,
so
we
regard
them
as equivalent
in
terms of
basic
composition
and
properties
.
Whether
a
given
segment
of
fila-
ment
will
fractionate
with the
nucleus
or
plasma
membrane
may
be
variable
and depend
as
much
on
the
homogenization
conditions
as
on
the
spatial
arrangement
of
the
filaments
.
We
have
developed
an
enucleation
technique
that
gives
fair
yields
of
whole
plasma
membranes
and
large pieces
of
mem-
brane
(referred
to
here
simply
as
"erythrocyte
membranes")
.
Many
of
these
membranes
retain
large
networks
of
intermedi-
ate
filaments
;
in
a given
thin
section,
though,
many
do
not
appear
to
possess
any
associated
filaments
.
This
may
be
a
result
of
loss
during
enucleation,
initial
lack
of
attachment
of
the
filaments
to
certain
regions
of the
plasma
membrane,
absence
FIGURE
8
Immunofluorescence
of turkey
erythrocyte
intermediate
filaments
.
Erythrocytes
adhering
to
cover
slips
were
hypoton-
ically
lysed
and
sonicated
to
remove
overlying
membranes
and
nuclei
.
Intermediate
filaments
remaining
attached
to
the
resulting
patches
of
plasma
membrane
were
visualized
by
immunofluorescence
using antibodies
to
synemin
(a,
r)
or
vimentin
(b-q,
s)
.
Specimens
were
not
fixed,
and
all
but
a, b, c,
and
h
were
treated
with Triton
X-100
.
Micrographs
a-p
are
indirect
immunofluores-
cence
images
;
q and rdemonstrate
colocalization of
vimentin
(q)
and
synemin
(r)
by double
immunofluorescence
.
Micrograph
s
is
a
combination
phase/fluorescence
image
showing
the
distribution
of
vimentin
on
the
elliptical
patches
of
plasma
membrane
;
note
that
many
patches
are
devoid
of
filaments
.
Bars, 5
Am
.
a-
r,
x
3040
;
s,
x
1330
.
FIGURE
9
Platinum
replicas
of
sonicated
chicken
erythrocyte
ghosts
.
Samples
were
prepared
as
in
Fig
.
8,
then
fixed,
dried,
and
rotary
shadowed
with platinum
for
examination
by
transmission
electron
microscopy
.
Intermediate
filaments
can be
seen
on
patches
of
plasma
membrane
that
remained
attached
to
the
cover
slip
during
sonication
.
In
b and
c,
a
portion
of
the
filaments
have
fallen
beyond
the
edge
of
the
membrane
patch
.
Magnification
:
Bar,
1
Am
.
X
16,000
.
from
certain
regions
of the
erythrocyte
cytoplasm,
or a
close
apposition
to
the
plasma
membrane
that
renders the
filaments
unresolvable
.
Previous
studies
involving
isolation
of the
avian
erythrocyte
plasma
membrane
by
differential
centrifugation
after
mechan-
ical
disruption
of the
cells
have
relied
on
pressure-release
homogenization
(7,
14,
61),
sonication
(4,
27),
rotating
blades
(11,
66),
or
a
tight-fitting
Dounce
(12,
22) or
Potter-Elvehjem
homogenizer
(6)
.
However,
the
presence
of
filaments
associated
with
the
isolated
membrane
fragments
was
noted
only
rarely
(27),
and,
in
comparisons
to
mammalian
erythrocyte
mem-
branes,
the
presence
of an
extra
polypeptide,
similar
in
molec-
ular
weight
to
vimentin,
was
rarely
mentioned
(11)
.
Some
of
these
disruption
techniques
produce
very
small
membrane
fragments
that
may
be
largely
stripped
of
filaments
;
alterna-
tively,
the
filaments
may
assume
a
distribution
or
configuration
in
which
they
are
not
readily
identifiable
by
electron
micros-
copy
.
The
gentler
disruption
techniques
appear
to
produce
membrane
fragments
similar
to
those
in
this
study,
but
associ-
ated
filaments
have
tended
to
escape
detection
.
Intermediate
GRANGER
ET
At
.
Synemin
and
Vimentin
in
Avian
Erythrocytes
30
9
filaments
have been most
apparent
in
detergent-insoluble
cy-
toskeletons
of
whole
erythrocytes
examined
by
thin
sectioning
or
negative
staining
(59,
63)
.
Treatment
of avian
erythrocyte
membranes
with
certain
low
ionic
strength
solutions
removes
the
associated
intermediate
filaments
.
Filaments
can
no
longer
be
found
with the
mem-
branes
in
thin
sections,
and
the
low-salt
extract
contains
nearly
pure vimentin
and
synemin
.
This
release
seems
to
depend
on
low
ionic
strength
and
absence
of
divalent
cations
and
be
independent
of
reducing
agents or
nonionic
detergents
.
Our
highest
yields
have been
obtained
using
distilled
water
.
Roughly
60-90%
of
the
vimentin
is
released
after
1
min
of
extraction
with
distilled
water
.
Selective
release
of
vimentin
and
synemin,
as
compared
to
spectrin
and
actin,
is
enhanced
by
low
temperature
and
brevity
of
treatment
.
Because
the
released
vimentin
and
synemin
cannot
be
sedimented
by
cen-
trifugation
for
5
h
at
240,000
g,
yet
appear
to
comigrate
in
a
gel filtration
column
with
an
exclusion
limit
of
15
million
daltons
(unpublished
observations),
they
must
exist
in
solution
as
some
sort
of
multimeric
complex
or
oligomer
.
This
implies
that
the
filaments
break
down
or
partially
depolymerize
during
or
after
release
from
the
membranes
.
Solubility
in
low
salt
has
similarly
been
described
for
other preparations of
native
inter-
mediate
filaments
(29,
30, 50, 51,
55, 57)
.
These
extraction
conditions
may
thus
be
resulting
in
a
dissolution
of
the
fila-
ments
rather
than
a
dissociation
of
the
filaments
from
the
membranes
.
It
is
conceivable
that
these extraction
conditions
have
no
disruptive
effect
on
the
anchorage
points
of
the
filaments
to
the
membranes,
which
would
explain
why
some
of the
vimentin remains with
the
membranes
after
extensive
extraction
with
water
.
This vimentin
may
be
a
distinct
popu-
lation
associated
with
anchorage
points
in
the
form
of
tightly
bound
monomers
or
oligomers
or
short
segments
of
filament
not
resolvable
in
thin
sections
.
These
extraction
data
thus
do
not
permit
a
conclusion
about
the nature
of
attachment
of
the
filaments
to
the
membranes
.
It
can
only
be
stated,
based
on
the
physical
data
of enucleation
and
sonication,
that
at
least
some
of the intermediate
filaments
in
avian
erythrocytes
are
somehow
anchored
to
the
plasma
membrane,
and
that
this
attachment
is
stable
in
the
presence
of
physiologic
salt,
high
salt,
and
nonionic
detergent
.
Comparison
of
Erythrocyte Proteins
An
aspect
of
comparative
biochemistry exemplified
by
this
study
is
the
difficulty
of
comparing
protein
profiles
of
a
given
preparation
by
different
SDS-PAGE
systems
.
Although
useful
for
general
comparisons,
different gel
systems
may
not
be
directly
comparable
with
regard
to specific
polypeptides
.
There
has
classically
been
disagreement
between
different
investiga-
tors
about
calculated
molecular
weights
;
even
the
relative
positions
of
different
polypeptides
may
not
be
consistent
in
different
gel
systems
(for
example,
the
high
molecular
weight
proteins
shown
in this
paper-see
Results)
.
This
stresses
cau-
tion
in
identifying
a
polypeptide
solely
by
its
mobility
on an
SDS-polyacrylamide
gel
.
Our
electrophoretic
profiles
of
avian
erythrocyte
membrane
proteins
differ
from
those
of
other
laboratories,
which
also
differ
among
themselves
(2,
4,
7, 11,
12,
34,
60,
61)
;
some
of
these
differences
have
been
noted
and
attributed
to
endogenous
proteases
or
proteases present
in
contaminating
leukocytes
(11,
34,
61)
.
Extrapolation
from
one
class
to
another
(for
example,
mammalian
(16)
to
avian
eryth-
rocyte
membranes)
may
not
be
justified
either
and
may
lead to
erroneous
identification
of
polypeptides
.
Two-dimensional
gel
31 0
THE
JOURNAL
OF
CELL
BIOLOGY
"
VOLUME
92,
1982
electrophoresis
makes
polypeptide
identification
less
ambigu-
ous,
because another
parameter
(isoelectric
point)
is
taken
into
account
and
has proved
useful
for
several
proteins
in
this
study
.
Nevertheless,
other
(nonelectrophoretic)
evidence
for
the
iden-
tity
of
a protein
band
on
a gel
is
essential
.
We
have
used
immunologic
and
solubility properties, in
addition
to
electro-
phoretic
characteristics, to
identify
synemin
and
vimentin
in
avian
erythrocytes,
and
phosphorylation
characteristics to
iden-
tify
goblin
(4)
.
Determination
of
why
similarly
prepared
sam-
ples
show
not
only
different
relative
mobilities
but
also
differ-
ent
relative
amounts
using
different gel
systems
awaits
further
study
.
The
Effects
of
Colcemid
One
indication
of
a
functional
or
interactive
difference
be-
tween
the intermediate
filaments
of
avian
erythrocytes
and
most
other
cell
types
grown
in
vitro
is
the
insensitivity
of
the
former
to
Colcemid
.
Treatments
with
Colcemid
that
will
cause
aggregation
and
perinuclear
bundling
ofintermediate
filaments
in
most
cultured
cells
(8, 18,
23, 31, 33,
38)
appear
to
have
no
effect
on
the
filaments
of
erythrocytes
.
Colcemid
sensitivity
might
thus
be
a function of
how
dynamic
a
cell
is,
or
perhaps
its
state
of
differentiation, as
appears
to
be
likely for
skeletal
muscle
cells
(20,
25)
but
not
be an
intrinsic
property
of
the
filaments
themselves
.
Related
to this
may
be
the observation
that
chick
erythrocyte
marginal
band
microtubules
are
resistant
to
depolymerization
by
Colcemid
(5)
.
Intermediate
Filament
Proteins
Synemin
was
originally
found
to
be
associated
with
inter-
mediate
filaments
in
smooth and
skeletal
muscle
(25)
.
Here
we
show
that
synemin
is
not a
muscle-specific
protein
but
is
present as well in
at
least
one
nonmuscle
cell,
the
mature
avian
erythrocyte
.
The
original
study
also
raised
the
possibility
that
synemin
was
a desmin-associated
polypeptide
;
here
we
show
that
synemin
can
also
exist
and
associate
with vimentin
.
In
both
muscle
cells
and
erythrocytes,
synemin
appears
to be a
component
of the
same
filaments
that
contain
desmin
and
vimentin,
as
determined
by
double
immunofluorescence
.
Den-
sitometric
scans of
Coomassie
Blue-stained
polyacrylamide
gels
of preparations
of
vimentin
and
synemin
from
chicken
eryth-
rocytes
give
a
vimentin-to-synemin
ratio
of
50
:1
.
This
is
similar
to
the
ratio
obtained
for
desmin
and
synemin
in
smooth
muscle
and
suggests
a constant stoichiometry
between synemin
and
intermediate
filaments
of
different
subunit
composition
.
This
ratio
is
a
very
rough
estimate,
not
taking
into
account
differential
proteolysis
of the
proteins
during
processing
and
possible
nonlinearity
in
dye
binding
and
densitometry,
and
should
therefore
not
be
regarded
as
the
true
ratio
.
It
is
useful,
however,
for
rough
comparisons
of
different
systems
.
We
have
taken
advantage
of
a
novel
form of
two-dimen-
sional
peptide
mapping
to
compare
synemins
from
different
tissues
.
This
combination
of
partial
hydrolysis
of
tissue
proteins
by endogenous
proteases
and
two-dimensional
immunoauto-
radiography
has demonstrated
a
high
degree
of
homology
between
synemins
from
avian
smooth
muscle
and
erythrocytes
.
Both
molecules
exhibit
an
S-shaped
string
of
fragments
that
terminates
in
a
proteolytically
stable
peptide
of 34,000 daltons
.
This
technique
is
extremely
sensitive,
detecting
peptides
much
too
scarce
to
be
seen
by
Coomassie
Blue
staining,
but
does
not
resolve
the
high
molecular
weight
peptides
sufficiently
to
allow
detailed
comparisons
.
Also,
slight
variations
from
gel
to
gel
do
not
allow
us
to
conclude
that
the
synemins
we
are
examining
are
identical
.
Minor
differences
in
the
maps
may
be
artifactual
or
may
reflect
functional
differences
in
the molecule,
perhaps
related
to
whether
synemin
is
found
in
association
with
desmin
or
with
vtmentm
.
These
data
do
not address the question
of
whether
synemin
is
an
integral
or
an
associated
filament
protein,
or
what
its
properties
are
independent
ofdesmin
and
vimentin
.
The
large
size
and
paucity of
synemin
relative to
desmin
and
vimentin
tend to favor a
role
for
synemin
as
an
associated
polypeptide
.
Perhaps
it
is
analogous
to
the
high
molecular
weight
polypep-
tide
of
neurofilaments,
which
appears
to
be
wrapped
helically
around
the core
filament
(62),
where
it
may
function
to
stabilize
the
filament,
promote
assembly
(45),
or
mediate
interactions
with other
molecules
or
organelles
.
The
presence of
nonmicrotubular
filaments
in
nucleated
erythrocytes
has
been
known
for
some
time
(27,
36),
but
only
recently
were
they
identified
as
intermediate
filaments
(59,
63)
.
These
filaments
were
usually
noted
and
studied
in
relation
to
the
nucleus
or
nuclear
membrane
.
In
this
paper
we
show
that
they
also
exhibit
a
close association
with
and
apparent
anchor-
age
to
the
plasma
membrane,and
that
they contain
the
inter-
mediate-filament
subunits
vimentin
and
synemin
.
Nucleated
erythrocytes
may
thus
be an
ideal
model
system
for the
study
of
filament-membrane
interactions
and
for
examining
inter-
mediate
filament
nucleation,
assembly,
and
deployment
during
differentiation
.
We
thank David
L
.
Gard
for his
helpful
comments
on
the
manuscript,
and
Dr
.
Jean-Paul
Revel
and
Mr
.
Patrick
F,
Koen
for
their
help
with
the
electron
microscopy
.
This
work
was
supported
by
grants
from
the
National
Institutes
of
Health
(NIH)
(GM
06965),
National
Science
Foundation,
Muscular
Dystrophy
Association
of
America,
and
a
Biomedical
Research
Sup-
port
Grant
to
the
Division of
Biology,
California
Institute
of
Technol-
ogy
.
B
.
L
.
Granger was
also
supported
by a
Predoctoral
Training
Grant
from
the
NIH
(GM
07616),
and
E
.
A
.
Repasky
by a
Postdoctoral
Fellowship
from
the
NIH(GM
07401)
.
E
.
Lazarides
is
the recipient
of
a
Research
Career
Development
Award
from
the
NIH
.
Received
for
publication
1
.?
July
1981,
and
in
revised
form
S
October
1981
.
REFERENCES
1 .
Alper,
S
.
L
.,
K
.
G, Beam,
and
P
.
Greengard
.
1980
.
Hormonal
control
of
Na*-K*
co-
transport
in
turkey
erythrocytes
.
Multiple
site
phosphorylation
of
goblin,
a
high
molecular
weight
protein
of
the
plasma
membrane
.
J
.
Biol
.
Chem
.
255
:4864-4871
.
2
.
Alper,
S
.
L
.,
H
.
C
.
Palfrey,
S
.
A
.
DeRiemer, and P
.
Greengard
.
1980
.
Hormonal
control
of
protein
phosphorylation
in
turkey
erythrocytes
.
Phosphorylation
by
cAMP-dependent
and
Ca
s
*dependent
protein kinases
of
distinct
sites
in
goblin,
a
high
molecular
weight
protein
of
the
plasma
membrane
.
J
.
Biol
.
Chem
.
255
:11029-11039
.
3
.
Batten,
B
.
E
.,
J
.1
.
Aalborg,
and
E
.
Anderson
.
1980
.
The
cytoplasmic
filamentous
network
in
cultured
ovarian
graaulosa
cells
.
Cell
.
21
:885-895
.
4
.
Beam,
K
.
G
.,
S
.
L
.
Alper,
G
.
E
.
Palade,
and
P
.
Greengard
.
1979
.
Hormonally
regulated
phosphoproteia
of
turkey
erythrocytes
.
Localization
to
plasma
membrane
.
J.
Cell
Biol
.
83
:
1-15
.
5
.
Behnke,
O
.
1970
.
A
comparative
study
of
microtubules
of
disk-shaped
blood
cells
.
J
.
Ultrastruct
.
Res
.
31
:61-75
.
6
.
Bilezikian,
J
.
P
.,
and
G
.
D
.
Aurbach
.
1973
.
A
0-adrenergic
receptor
of
the
turkey
erythrocyte
. 1 .
Binding
of
catecholamine
and
relationship to
adenylate
cyclase
activity
.
J
.
Biol
.
Chern
.
248
:5577-5583
.
7
.
Blanchet,
J
.
P
.
1974
.
Chicken
erythrocyte
membranes
:
Comparison
of
nuclear
and plasma
membranes
from
adults
and
embryos
.
Exp
.
Cell
Res
.
84
:159-166
.
8
.
Blose,
S
.
H
.,
and
S
.
Chacko
.
1976
.
Rings
of
intermediate
(100
A)
filament
bundles
in
the
perinuclear
region
of
vascular endothelial
cells
.
Their
mobilization
by
colcemid
and
mitosis
.
J
.
Cell
Biol
70
:459-466
.
9
.
Branton,
D
.,
C
.
M
.
Cohm
and J
.
Tyler
.
1981
.
Interaction
of
cytoskeletal proteins
on
the
human
erythrocyte
membrane
.
Cell
. 24
:24-32
.
10
.
Burridge,
K
.
1978
.
Direct
identification
of
spec
glycoproteins
and
antigens
in
sodium
dodecyl
sulfate
gels
.
Methods Enzymol
.
50
:54-64
.
11
.
Caldwel
,
A
.
B
.
1976
.
Proteins
of
the turkey
erythrocyte
membrane
.
Biochemistry
.
15
:2711-
2718
.
12
.
Chan,
L
.
N
.
L
.
1977
.
Change
s in
the
composition
of
plasma
membrane
proteins
during
differentiation
ofembryonic
chick
erythroid
cells
.
Proc
.
Nail
.
Acad
.
Sci
.
U
.
S
.
A
.
74
:1062-
1066
.
13
.
Cohen,
W
.
D
.
1978
.
Observations
on
the
marginal
band
system
of
nucleated
erythrocytes
.
J
.
Cell Biol
.
78
:260-273
.
14
.
Davoren,
P
.
R
.,
and
E
.
W
.
Sutherland
.
1963
.
Th
e
cellular
location of
adenyl
cyclase
in
the
pigeon
erythrocyte
.
J
.
Biol
.
Chem
.
238
:3016-3023
.
15
.
Dodge,
J
.
T
.,
C
.
Mitchell,
and
D
.
J
.
Hanahan
.
1963
.
Th
e
preparation
and
chemical
characteristics
of
hemoglobin-free
ghosts
of
human
erythrocytes
.
Arch
.
Biochem
.
Biophys
.
100
:119-130
.
16
.
Fairbanks,
G
.,
T
.
L
.
Stock,
and
D
.
F
.
H
.
Wallach
.
1971
.
Electrophoreti
c
analysis
of
the
major
polypeptides
of
the
human
erythrocyte
membrane
.
Biochemistry
.
10
:2606-2617
.
17
.
Fourie,
F
.
leR
.
1977
.
Effects
of
anticoagulants
on
the
hematocrit,
osmolarity
and
pH
of
avian blood
.
Point
.
ScL 56
:1842-1846
.
18
.
Franke,
W
.
W
.,
E
.
Schmid,
M
.
Osborn,
and
K
.
Weber
.
1978
.
Different intermediate-sized
filaments
distinguished
by
immunofluorescence microscopy
.
Proc
.
Nall
.
Aced
.
Sci
.
U
.
S
.
A
.
75
:5034-5038
.
19
.
Gard,
D
.
L
.,
P
.
B
.
Bell,
and
E
.
Lazarides
.
1979
.
Coexistence
of
desmin
and
the
fibroblastic
intermediate
filament
subunit
in
muscle
and
nonmuscle
cells
:
identification
and
compar-
ative
peptide
analysis
.
Proc
.
Natl
.
Acad
.
Sci
.
U
.
S
.
A
.
76
:3894-3898
.
20
.
Gard,
D
.
L
.,
and
E
.
Lazarides
.
1980
.
The
synthesis
and
distribution
of desmin and
vimentin
during
myogenesis
in
vitro
.
Cell
19
:263-275
.
21
.
Gerace,
L
.,
and
G
.
Blobel
.
1980
.
The
nuclear
envelope
lamina
is
reversibly
depolymerized
during
mitosis
.
Cell
.
19
:277-287
.
22
.
Ginsberg,
B
.
H
.,
C
.
R
.
Kahn,
and J
.
Roth
.
1976
.
The
insulin
receptor
of
the
turkey
erythrocyte
.
Characterization
of
the
membrane-bound
receptor
.
Biochim
.
Biophys
.
Acta
443
:227-242
.
23
.
Goldman,
R
.
D
.
1971
.
The
role
of
three
cytoplasmic
fibers
in
BHK-21
cell
motility
.
1
.
Microtubules
and
the
effects
of
Colchicine
.
J
.
Cell Biol
.
51
:752-762
.
24
.
Granger,
B
.
L
.,
and
E
.
Lazarides
.
1979
.
Desmin
and
vimentin
coexist
at
the periphery
of
the myofibril
Z
disc
.
Cell
.
18
:1053-1063
.
25
.
Granger,
B
.
L
.,
and E
.
Lazarides
.
1980
.
Synemin
:
a
new
high
molecular
weight
protein
associated
with
desmin and
vimentin
filaments
in
muscle
.
Cell
22
:727-738
.
26
.
Hall,
C
.
E
.
1956
.
Visualization
of
individual
macromolecules
with
the
electron
microscope
.
Proc
.
Nall
.
Acad
Sci
.
U
.
S
.
A
.
42
:801-806
.
27
.
Harris,
J
.
R
.,
and
l
.
N
.
Brown
.
1971
.
Fractionatio
n of
the
avian
erythrocyte
:
an
ultrastruc-
tural
study
.
L
Uttrastruct
.
Res
.
36
:8-23
.
28
.
Heuser,
J
.
E
.,
and
M
.
W
.
Kirschner
.
1980
.
Filamen
t
organization
revealed
in
platinum
replicas
of
freeze-dried
cytoskeletons
.
J
.
Cell
Biol
.
86
:212-234
.
29
.
Hubbard,
B
.
D
.,
and
E
.
Lazarides
.
1979
.
Copurification
of
actin
and desmin from
chicken
smooth muscle
and
their
copolymerization
in
vitro
to
intermediate
filaments
.
J
.
Cell
Biol
.
80
:166-182
.
30
.
Huiatt,
T
.
W
.,
R
.
M
.
Robson,
N
.
Arakawa, and
M
.
H
.
Stromer
.
1980
.
Desmi
n from
avian
smooth
muscle
.
Purification
and
partial
characterization
.
J
.
Biol
.
Chem
.
255
:6981-6989
.
31
.
Hynes~
R
.
0
.,
and
A
.
T
.
Destree
.
1978x
. 1
0
mm
filaments
in
normal
and
transformed
cells
.
Cell
.
13
:151-163
.
32
.
Hynes,
R
.
O
.,
and
A
.
Destree
.
1978b
.
Relationships
between
fibronectin
(LETS
protein)
and
actin
.
Cell
.
15
:875-886
.
33
.
Ishikawa,
H
R
.
Bischoff,
and
H
.
Holtzer
.
1968
.
Mitosis
and
intermediate-sized
filaments
in
developing
skeletal
muscle
.
J
.
Cell Biol
.
38
:538-555
.
34
.
Jackson,
R
.
C
.
1975
.
The
exterior
surface
of
the
chicken
erythrocyte
.
J
.
Biol
.
Chem
250
:
617-622
.
35
.
Jacobson,
B
.
S
.,
and
D
.
Branton
.
1977
.
Plasma
membrane
:
rapid
isolation
and
exposure
of
the
cytoplasmic
surface
by
use
of
postively
charged
beads
.
Science
(Wash
D
.
C
.)
.
195
:
302-304
.
36
.
Koehler,
J
.
K
.
1968
.
Freeze-etching
observations
on
nucleated
erythrocytes
with
special
reference
to
the
nuclear
and plasma
membranes
.
Z
.
Zellforsch
.
Mikrosk
.
Anat
.
85
:1-17
.
37
.
Laenmtli,
U
.
K
.
1970
.
Cleavage
of
structural
proteins
during
the
assembly
of
the
head of
bactenophage
T4
.
Nature
(Lord
.)
.
227
680--685
.
38
.
Lazarides~
E
.
1978
.
The
distribution
of
desmin
(100
A)
filaments
in
primary
cultures
of
embryonic
chick
cardiac
cells
.
Exp
.
Cell
Res
.
112
:265-273
.
39
.
Lazarides,
E
.
1980
.
Intermediate filaments
as
mechanical
integrators
of
cellular
space
.
Nature
(Land
.)
.
283
:249-256
.
40
.
Lazarides,
E
.,
and
B
.
D
.
Hubbard
.
1976
.
Immunologica
l
characterization
of
the
subunit
of
the
loo
A
filaments
from
muscle
cells
.
Proc
.
Nat
.
Acad
.
Sci
.
U
.
S
.
A
.
73
:4344-4348
.
41
.
Maddy,
A
.
H
.,
and P
.
G
.
Kelly
.
1971
.
Aceti
c
acid
as
a
solvent
for
erythrocyte
membrane
proteins
.
Biochim
.
Biophys
.
Acta
.
241
:290-301
.
42
.
Marchesi,
S
.
L
.,
E
.
Steers,
V
.
T
.
Marchesi,
and
T
.
W
.
Tillack
.
1970
.
Physica
l
and
chemical
properties
ofa
protein
isolated
from
red
cell
membranes
.
Biochemistry
.
9
:50-57
.
43
.
Marchesi,
V
.
T
.,
and
E
.
Steers
.
1968
.
Selective
solubilization
ofa
protein
component
of
the
red
cell
membrane
.
Science
(Wash
.
D
.
C
.)
.
159
:203-204
.
44
.
Marek,
L
.
F
.,
R
.
O
.
Kelley,
B
.
D
.
Perdue,
and P
.
O
.
Zamora
.
1979
.
Rol
e
of
the cytoskeleton
in
fibroblast
attachment
to
poly-L-lysine
treated
substrates
.
J
.
Cell
BioL
83
(2,
Pt
.
2)
:
324a
(Abstr
.)
.
45
.
Moon,
H
.
M
.,
T
.
Wisniewski,
P
.
Men,
l
.
DeMartini,
and
H
.
M
.
Wisniewski
.
Partial
purification
of
neurofilament
subunits
from
bovine
brains
and
studies
on
neurofilament
assembly
.
J
.
Cell
Mot
.
89
:560-567
.
46
.
O'Connor,
C
.
M
.,
D
.
R
.
Balzer,
and
E
.
Lazarides
.
1979
.
Phosphorylation
of
subunit
proteins
of
intermdiate
filaments
from
chicken
muscle and nonmuscle
cells
.
Proc
.
Nail
.
Acad
.
Sc
i
U
.
S
.
A
.
76
:819-823
.
47
.
O'Connor,
C
.
M
.,
D
.
L
.
Gard
and
E
.
Lazarides
.
1981
.
Phosphorylation
of
intermediate
filament proteins
by
cAMP-dependent
protein kinases
.
Cell
23
:135-143
.
48
.
O'Farrell,
P
.
H
.
1975
.
High
resolution
two-dimensional
electrophoresis
of
proteins
.
J
.
Biol
.
Chem
.
250
:4007-021
.
49
.
Osborn,
M
.,
R
.
E
.
Webster,
and
K
.
Weber
.
1978
.
Individual
microtubules
viewed by
immunofluorescence
and
electron
microscopy
in
the
samePtK2
Cell
.
J
Cell Biol
.
77:R27-
R34
.
50
.
Rueger,
D
.
C
.,
J
.
S
.
Huston,
D
.
Dahl,
and
A
.
Bigrami
.
1979
.
Formation
of
100
A
filaments
from
purified
glial
fibrillary
acidic
protein
in
vitro
.
J
.
Mal
.
Biol
.
135 :53-68
.
51
.
Schlaepfer,
W
.
W
.
1978
.
Observations
on
the
disassembly
of
isolated
mammalian
neuro-
filaments
.
J
.
Cell Biol
.
76
:50-56
.
52
.
Shelton,
K
.
R
.,
L
.
L
.
Higgins,
D
.
L
.
Cochran
. J
.
J
.
Ruffolo,
and
P
.
M
.
Egle
.
1980
.
Nuclea
r
lamins
of
erythrocyte
and
liver
.
J
Biol
.
Chem
255
:10978-10983
.
53
.
Small,
J
.
V
.,
and
A
.
Sobieszek,
1977
.
Studies
on
the
function
and
composition
of
the
to-
nm
(100-A)
filaments
of
vertebrate
smooth
muscle
.
J
Cell
Sci
.
23
:243-268
.
54
.
Sommer,
J
.
R
.
1977
.
To
cationize
glass
.
J
.
Cell
Biol
.
75
(2,
Pt
.
2)
:
245a
(Abstr
.)
.
55
.
Starger,
J
.
M
.,
W
.
E
.
Brown,
A
.
E
.
Goldman,
and
R
.
D
.
Goldman
.
1978
.
Biochemica
l
and
immunological
analysis
of
rapidly
purified
10-nm
filaments
from
baby
hamster
kidney
(BHK-21)
cells
.
J
.
Cell Biol
.
78
:93-109
.
56
.
Stock,
T
.
L
.,
and
J
.
Yu
.
1973
.
Selective
solubilization
of
proteins
from
red
blood
cell
membranes
by
protein
perturbants
.
J
.
Supramol
.
Struct
. 1
:220-232
.
57
.
Steinert,
P
.
M
.,
W
.
W
.
Idler,
F
.
Cabral,
M
.
M
.
Gottesman,
and
R
.
D
.
Goldman
.
1981
.
I
n
GRANGER
ET
AL
.
Synemin
and
Vimentin
in
Avian
Erythrocytes
31
1
vitro
assembly
of
homopolymer
and
copolymer
filaments
from
intermediate
filament
subunits
of
muscle
and
fibroblastic
cells
.
Proc
.
Nail
.
Acad
.
Sci
.
U
.
S
.
A
.
78
:3692-3696
.
58
.
Tuszynski,
G
.
P
.,
E
.
D
.
Frank,
C
.
H
.
Damsky,
C
.
A
.
Buck,
and L
.
Warren
.
1979
.
The
detection
of
smooth
muscle
desmin-like
protein
in
BHKs,/C,a
fibroblasts
.
J
.
Biol
Chem
.
254
:6138-6143
.
59
.
Virtanen,
L,
M
.
Kurkinen,
and
V
.
P
.
Lehto
.
1979
.
Nucleus-anchorin
g
cytoskeleton
in
chicken
red
blood
cells
.
Cell
Blot
.
Int
.
Rep
.
3
:157-162
.
60
.
Weise,
M
.
1
.,
and
L.
N
.
L
.
Chan
.
1978
.
Membran
e
protein
synthesis
in
embryonic
chick
erythroid
cells
.
J
.
Mot
.
Chem
.
253
:1892-1897
.
61
.
Weise,
M
.
1
.,
and
V
.
M
.
Ingram
.
1976
.
Protein
s
and
glycoproteins
of
membranes
from
developing
chick
red
cells
.
J
Biol
.
Chem
.
251
:6667-6673
.
62
.
Willard,
M
.,
and
C
.
Simon
.
1981
.
Antibody
decoration
of
neurofilaments
.
J
.
Cell
Biol
.
89
:
31
2
THE
JOURNAL
OF
CELL
BIOLOGY
-
VOLUME
92,
1982
198-205
.
63
.
Woodcock,
C
.
L
.
F
.
1980
.
Nucleus-associate
d
intermediate
filaments
from
chicken
eryth-
rocytes
.
J
.
Cell
Wol
.
85
:881-889
.
64
.
Yu, J
.,
D
.
A
.
Fischman,
and
T
.
L
.
Steck
.
1973
.
Selectiv
e
solubilization
of
proteins
and
phospholipids
from
red
blood
cell
membranes
by
nonionic
detergents
.
J
.
Supramol
Struct
.
1
:233-248
.
65
.
Zentgraf,
H
.,
B
.
Deumling,
and
W
.
W
.
Franke
.
1969
.
Isolatio
n and
characterization of
nuclei
from
bird
erythrocytes
.
Exp
.
Cell
Res
.
56
:333-337
.
66
.
Zentgraf,
H
.,
B
.
Deumling,
E
.-D
.
Jarasch,
and
W
.
W
.
Franke
.
1971
.
Nuclea
r
membranes
and
plasma
membranes
from hen
erythrocytes
. 1 .
Isolation,
characterization,
and compar-
ison
.
J
.
Biol
.
Chem
.
246
:2986-2995
.
