Available via license: CC BY-NC-SA 4.0
Content may be subject to copyright.
Available via license: CC BY-NC-SA 4.0
Content may be subject to copyright.
Biochemical
Analysis
of
Connexin43
Intracellular
Transport,
Phosphorylation,
and
Assembly
into
Gap
Junctional
Plaques
Linda
S
.
Musil
and
Daniel
A
.
Goodenough
Department
of
Anatomy
and
Cellular
Biology,
Harvard
Medical
School,
Boston,
Massachusetts
02115
Abstract
.
We
previously
demonstrated
that the
gap
junction
protein
connexin43
is
translated
as a
42-kD
protein
(connexin43-NP)
that
is
efficiently
phosphor-
ylated
to
a
46,000-M
*
species
(connexin43-P
2 )
in
gap
junctional
communication-competent,
but not
in
com-
munication-deficient,
cells
.
In
this
study,
we
used
a
combination
of
metabolic
radiolabeling
and
immuno-
precipitation
to
investigate the
assembly
of
connexin43
into
gap
junctions
and
the
relationship
of
this
event
to
phosphorylation
of
connexin43
.
Examination
of
the de-
tergent
solubility
of
connexin43
in
communication-
competent
NRK
cells
revealed
that
processing
of
con-
nexin43
to the
P
2
form was accompanied
by
acquisition
of
resistance
to solubilization in
1%
Triton
X-100
.
Immunohistochemical
localization
of
connexin43
in
Triton-extracted
NRK
cells
demonstrated
that
con-
nexin43-P
2
(Triton-insoluble)
was
concentrated
in
gap
AP
junctions
are
plasma
membrane
specializations
that
mediate
the
regulatable
transfer
of
small
mole-
cules
and
ions
between
adjoining
cells
(Gilula
et al
.,
1972
;
Loewenstein,
1981
;
Beyer
et
al
.,
1990)
.
Present
in vir-
tually
all
metazoan
tissues,
gap
junctions
are
involved
both
in
.
relaying
signals
and
in
maintaining
metabolic
continuity
between
connected
cells
.
In
electrically
excitable
tissues
in-
cluding
myocardium,
smooth
muscle,
and
nerve,
gap
junc-
tions
provide
low-resistance
electrical
pathways
between
cells
that
are
essential
for
their
specialized
functions
(De
Mello,
1987)
.
Gap
junction-mediated
intercellular
com-
munication
has
also
been
implicated
in
fundamental
cellular
processes
such
as
embryonic
development,
differentiation,
and growth
control
(Loewenstein,
1979
;
Mehta
et
a]
.,
1986
;
Guthrie
and
Gilula,
1989)
.
It
has
been
well
established
that
gap
junctions
in
ver-
tebrates
are
comprised
of
connexins,
members
of
a
family
of
closely
related
integral
membrane
proteins
(Stevenson
and
Paul,
1989)
.
The
events
involved
in the
assembly
of
newly
synthesized
connexin
monomers
into
functional
gap
junctional
plaques
are,
however,
poorly
understood
.
A
necessary
step
injunction
formation
is
oligomerization
of
six
connexin
monomers
into half
an
intercellular
channel
(a
con-
nexon)
in
an
as yet
undefined
cellular
compartment
.
A
con-
nexon
in
the
plasma
membrane
of
one
cell
must
then
join
©
The
Rockefeller
University
Press,
0021-9525/91/12/1357/18
$2
.00
The
Journal
ofCell
Biology,
Volume
115,
Number
5,
December
1991
1357-1374
1357
junctional
plaques,
whereas
connexin43-NP
(Triton-
soluble)
was
predominantly
intracellular
.
Using
either
a
20°C
intracellular
transport
block
or
cell-surface
pro-
tein
biotinylation,
we
determined
that
connexin43
was
transported
to
the
plasma
membrane
in the
Triton-sol-
uble
connexin43-NP
form
.
Cell-surface
biotinylated
connexin43-NP
was
processed
to
Triton-insoluble
con-
nexin43-P
2
at
37°C
.
Connexin43-NP
was
also
trans-
ported
to
the
plasma
membrane
in
communication
de-
fective,
gap
junction-deficient
S180 and
L929
cells
but
was
not
processed
to
Triton-insoluble
connexin43-P2
.
Taken
together,
these
results
demonstrate
that
gap
junc-
tion
assembly
is
regulated
after
arrival
of
connexin43
at
the
plasma
membrane
and
is
temporally
associated
with
acquisition
of
insolubility in
Triton
X-100 and
phosphorylation
to the
connexin43-P
2
form
.
with
a
connexon
in
an
opposing
cell
membrane
to
form
an
intercellular
channel
.
These
channels
become
concentrated
at
cell-cell
interfaces
into
very
high-density
clusters
(-10
4
channels/gm
2
of
membrane)
referred
to as
gap
junctional
plaques
or
maculae
(Loewenstein,
1981
;
Yamasaki,
1990)
.
It
is
not
known
how
these
various
assembly
steps
are
regu-
lated,
or
whether
they
serve
as
control
points
in
the
estab-
lishment
of
gap
junction-mediated
cell-cell
communication
.
Many
effectors
of
protein
kinases
modulate
gap
junctional
communication,
and
it
has
recently
been
demonstrated
that
certain
connexins themselves
are
phosphorylated
(reviewed
by
Musil
and
Goodenough,1990
;
Stagg
and
Fletcher,
1990)
.
Saez
et al
.
(1986)
reported
that
addition
of
8-bromo-cAMP
to
primary
rodent
hepatocytes
results
in
a
1
.6-fold
increase
in
incorporation of
32
P
into
immunoprecipitable
connexin32
within
30-60
min,
apparently
without
an
appreciable
in-
crease
in
the
total
amount
of
connexin32
protein
(Traub
et
al.,
1987)
.
Since
gap
junctional
conductance
was
also
in-
creased
50-75
%
during
this
period,
it
was
proposed
that
phosphorylation
of
connexin32
may
be
involved
in
the regu-
lation
of
gap
junctional
communication
.
Within
the
last
year
several
groups
have demonstrated
serine
phosphorylation
of
connexin43,
a
connexin
first
identified
in
heart
yet
present
also
in
a
wide
variety
of
other
tissues
(Crow
et al
.,
1990
;
Musil
et
al
.,
1990a,
b
;
Filson
et al
.,
1990
;
Swenson
et al
.,
on July 10, 2011jcb.rupress.orgDownloaded from
Published December 1, 1991
1990
;
Laird
et al
.,
1991)
.
Agonists
of
protein
kinase
A
and
protein
kinase
C
have
been
shown
to
affect
gap
junctional
conductance
in several
çonnexin43-containing
cell
types
(Loewenstein,
1985
;
Stagg
and
Fletcher,
1990
;
Spray
and
Burt,
1990)
;
as
with
connexin32,
the
mechanism
whereby
these agents influence
cell-cell
communication
is
not
known
.
A
potential
functional
relationship
between
serine
phos-
phorylation
of Connexin43
and gap
junctional
communica-
tion
was
suggested
by
studies
examining
the
posttranslational
processing
of
Connexin43
in cell
lines
that
differ
greatly
in
their
ability
to
form
morphologically
and
physiologically
recognizable
gap
junctions
(Musil
et al
.,
19906)
.
In
all
gap
junctional
communication-competent
cell
types
examined,
Connexin43
is
synthesized
as
a
single,
42-kD
species
that
is
converted
to
a species of
ti44-kD
(Connexin43-P,)
and
then
to
one
of
-46 kD
(Connexin43-P
2 )
by
the
addition
of
phos-
phate
onto
serine
residues
.
In
contrast,
certain
cell
lines
that
are severely
deficient
in
junctional
communication
(mouse
S180
and
L929
cells)
constitutively
synthesize
Connexin43
but
neither
process
it
to
the
P
2
form
nor
accumulate
Con-
nexin43
in
visible
gap
junctional
plaques
.
Conversion
of
S180
cells
to
a
communication-competent
phenotype
by
transfec-
tion
with
a
cDNA
encoding
the
cell-cell
adhesion molecule
L-CAM
induces
both
phosphorylation
of Connexin43
to the
P
2
form
and
assembly
of
gap
junctional
plaques
.
In
com-
plementary
experiments,
ordinarily
communication-com-
petent
cells
treated
with
known
inhibitors
of
gap
junction
permeability
no
longer
detectably
process
Connexin43
to
Connexin43-P
2 .
These
results
establish
a
strong
correlation
between
the
ability
of
cells
to
phosphorylate
Connexin43
to
the
P
2
form
and
to
form
morphologically
and
physiologi-
cally
recognizable
gap
junctions,
but
shed
little
light
on
the
functional
role
of
Connexin43
phosphorylation
.
In the
current
study
we
investigated
further
the
relation-
ship
between
Connexin43
phosphorylation,
gap
junction
as-
sembly,
and
cell-cell
communication
.
Biochemical
assays
for
transport
of
Connexin43
to the
plasma
membrane
and
for
accumulation
of
Connexin43
in
junctional
plaques
were
de-
veloped
and
used
to
analyze
Connexin43
processing
in
vari-
ous
cell
types
.
In
NRK
and
other
communication-competent
cells,
Connexin43
undergoes
a
dramatic
posttranslational
change
in
Triton
X-100
solubility
that
is
temporally
as-
sociated
both
with
phosphorylation
of
connexin43
to
tile
P2
form
and
with
assembly of Connexin43
into
gap
junctional
plaques
.
Both
acquisition
of
Triton
insolubility
and
phos-
phorylation
of
Connexin43
occur
(at
least
in
part) after
trans-
port
of
Connexin43
to the
plasma
membrane
.
Communica-
tion-defective,
gap
junction-deficient
S180
and
L929
cells
also
transport
Connexin43
to the
cell
surface
but
do
not pro-
cess
it
to
Triton-insoluble,
terminally
phosphorylated
con-
nexin43-P2
.
Processing
of
Connexin43
to
the
P
2
form
is
thus
not
required
for transport of
Connexin43
to
the
cell
surface
but
is
tightly
correlated
with
incorporation
of
Connexin43
into
junctional
plaques,
a
fact
suggesting
that
terminal
phos-
phorylation
of
Connexin43
may
be
involved
in
a
later
step in
gap
junctional
plaque
assembly
or
in functional
processes
.
Materials
and
Methods
Reagents
Tissue
culture
reagents
were
purchased
from
Gibco
(Grand
Island,
NY),
The
Journal
of
Cell
Biology,
Volume
115,
1991
except
for
FCS,
which
was
purchased
from
HyClone
(Logan,
Utah)
.
[
35
S]Methionine
(cell-labeling
grade)
was
from
New
England
Nuclear
(Boston,
MA)
;
Tran
35
S-label
was
obtained
from
ICN
Radiochemicals
(Div
.
ICN
Biomedicals
Inc
.,
Irvine,
CA)
.
NHS-LC-biotin
and
avidin-
agarose
beads
were
purchased
from
Pierce
Chemical
Co
.
(Rockford,
IL)
.
Lauryldimethylamine
oxide
(LDAO)
was
from
Calbiochem
Corp
.
(San
Diego,
CA)
.
Unless
otherwise
specified,
all
other
chemicals
were
obtained
from
Sigma
Chemical
Co
.
(St
.
Louis,
MO)
.
Cell
Culture
The
NRK,
5180,
and
L929
cell
lines
were
maintained
as
previously
de-
scribed
(Musil
et al
.,
19906)
.
Three-day-old,
newly
confluent
60-mm
cul-
tures
were used
for
all
experiments
unless
otherwise
specified
.
The
gap-
junctional
communication-competence
of
the
NRK
cells
was
confirmed
by
dye-coupling
experiments
(not
shown)
;
comparable
results
were
obtained
when
Lucifer
yellow
was
introduced
into cells
by
microinjection
(Schuetze
and
Goodenough,
1982)
or by scrape
loading
(El-Fouly
et
al
.,
1987)
.
Metabolic
Labeling
of
Cells,
Preparation
of
Cell Lysates,
and
Immunoprecipitation
Details
of
the
metabolic
labeling
of
cell
cultures
with
["S]methionine
are
given
elsewhere
(Musil
etal
.,
1990x)
.
Each
60-mm
cell
culture
was
labeled
with
100
WCi/ml
of
[
35
S]methioníne,
except
in
the
case
of
cell-surface
bi-
otinylation
experiments
for
which
350
gCi/ml
of
[
35
S]methionine or
Tran
35
S-label
were
used
.
At
the
end
of
the
labeling
or
chase
period,
the
cultures
were
solubilized
in the
presence of
0
.6°ío
SDS,
and
the resulting
cell
lysates
were
immunoprecipitated
according
to
the
procedure
of
Musil
et
al
.
(1990x)
.
The
anti-Connexin43
antibodies
used throughout
this
study
were
affinity
purified
from
a
rabbit
antiserum
generated
against
a
connexin43-
specific
synthetic
peptide
encoding
amino
acids
252-271
of
rat
heart
Con-
nexin43
(Beyer
et al
.,
1989)
as
previously
described (Musil
et
al
.,
19906)
.
SDS
Gel
Electrophoresis
and
Fluorography
Immunoprecipitated
samples
were
analyzed on
10%
SDS-polyacrylamide
gels
(Laemmli,
1970)
.
Gels
were
processed
for
fluorography
with
ENIHANCE
(New
England
Nuclear,
Boston,
MA)
using
the
supplier's
suggested
protocol
and
then
exposed
to
prefogged
XAR-5
film
(Eastman
Kodak
Co
.,
Rochester,
NY)
.
Quantitative
densitometry
was conducted
with
an
U1troScan
XL
laser
densitometer
(LKB
Instruments
Inc
.,
Broma,
Sweden)
.
Detergent
Solubilization
of
Connexin43
from
Cultures
and
Lenses
Metabolically
labeled
cell
cultures
were
scraped
from
the
tissue
culture
dish
with
a
rubber
spatula
into
4 ml
of
Leibovitz's
L-15
medium
supplemented
with
2
mM
PMSF
and
10
mM
N-ethylmaleimide
(4°C)
and
then
pelleted
by
centrifugation
at
150
g
for 7
min
.
Lenses
were
dissected
from
10-d
white
leghorn chicken
embryos
(taking
care
to
remove
associated
ciliary
epithelial
cells)
and
metabolically labeled
with
[
35
S]methionine as
described
else-
where
(Musil
et
al
.,
1990x)
.
In
the
standard
Triton
X-100
solubilization
assay,
cells
(or lenses)
were
then
resuspended
in
1
ml
of
lysis
buffer (5
mM
Tris base,
2
mM
EDTA,
2
mM
EGTA,
0
.5
mM
diisopropylfluorophosphate,
10
mM
N-ethylmaleimide,
2
mM
PMSF,
and
200
pM
leupeptin)
.
After
a
10-min
incubation
at
4°C,
the
swollen
cells
were
disrupted
by
repeated
passage
(25-30
times)
through
a
25-gauge
needle,
and
the
resulting
cell
lysates
were
brought
to
isotonicity
by
addition
of
100
td
of
a
lOx
PBS
stock
solution
.
Alternatively,
cells
were
disrupted
under
isotonic
conditions
in either
PBS
or
L-15
medium,
with
identical
results
.
20%
Triton
X-100
was
added
to a
final
concentration
of
1%
(wt/wt),
and
the lysates
were
incubated
for
30
min
at
4°C
with
occa-
sional
resuspension
by
vortexing
.
Half
of
the
sample
(575
pl)
was
then
reserved
at
4°C
(total
cell
lysate)
whereas
the
remainder
was
subjected
to
centrifugation
at
100,000
g for
50
min
at
4°C
.
The
supernatant
fraction
(Triton-soluble
lysate)
was
carefully
removed,
brought
to
1
.2% SDS,
and
boiled
for 3
min
.
When
cool,
the
sample
was
diluted
with
900
Al
of
immu-
noprecipitation
buffer
(Musil
et al
.,
1990x) supplemented
with
0
.5
M
su-
crose,
10
mM
N-ethylmaleimide,
2
mM
PMSF,
and
enough
Triton
X-100
to
bring
the
final
detergent
concentration
to
2%
Triton
X-100
and
0
.4%
SDS
.
The
total
cell
lysate
sample was
processed
identically
.
The
material
pelleted
at
100,000
g
(Triton-insoluble
lysate)
was
resuspended
to
575
kl
with
immunoprecipitation
buffer
supplemented
with
2
mM
PMSF
and
1
.2
%
135
8
on July 10, 2011jcb.rupress.orgDownloaded from
Published December 1, 1991
SDS
and
then
boiled
for
3
min,
after
which
it
was
diluted
with
Triton-
containing
immunoprecipitation
buffer
as
described
for
the
Triton-soluble
supernatant
.
All
three
fractions
were
then
immunoprecipitated
with
affinity-
purified
antibodies
to
connexin43
(252-271)
using
our
standard
protocol
(see
above)
.
In
certain
experiments
the
buffer conditions,
amount
of
Triton,
temperature,
and/ortime of
solubilization
were
altered as specified in
Table
I
.
For
solubilization
in
sarcosine,
NRK
cells
or
embryonic
chick
lenses
were
disrupted
in 5
mM
Tris base,
0
.5
mM
diisopropylfluorophosphate,
10
mM
N-ethylmaleimide,
2
mM
PMSF,
and 200
uM
leupeptin,
pH
10
.0,
and
incubated
for
10
min
at
4°C
.
N-laurylsarcosine
was
then
added
to a
final
concentration
of
0
.3
%,
after
which
the
lysates
were
incubated
for
10-15
min
at
25°C
with
occasional
vortexing
.
The
samples
were
then
centrifuged
at
100,000
g
and
processed
as
described
for
Triton-solubilized
cultures
.
In
Situ
Extraction
of
NRK
Cells with
Tlriton
NRK
cell
cultures
grown on
uncoated
35-mm
tissue
culture
dishes
were
rinsed
three
times
at
4°C
with
incubation
buffer (8 .0 g
NaCl,
0
.4
g
KCI,
0
.09
g
Na2HP04-7H20,
0
.047
g
KH2PO
4
,
0
.097
g
MgSO4,
0
.4
g
CaC12
.2H20,
and 4
.76
g Hepes
per
liter
of
H2O
;
pH
7
.5)
.
The
cells
were
then
equilibrated
in
the
same
buffer
for
20 min
at
WC,
after
which
the
buffer
was
removed
and
replaced
with
1
ml
of
incubation
buffer
supplemented
with
0
.5
mM
diisopropylfluorophosphate,
10
mM
N-ethylmaleimide,
2
mM
PMSF,
and
200
pM
leupeptin
in
either
the
absence
(mock-extracted
cells)
or
presence
(Triton-extracted
cells)
of
1%
Triton
X-100
.
The
cultures
were
gently
rotated
on an
orbital
shaker
at
14°C
for
30
min,
after
which
the ex-
traction buffer
was
carefully
removed
and
the
cultures
were
rinsed
five
times
with
incubation
buffer
containing
0
.5
mM
diisopropylfluorophosphate
and
2
mM
PMSF
In
the
case of
Triton-extracted
cultures,
this
buffer
contained
1%
Triton
for
the
first
three
rinses
.
Care
was
taken
not
to
disrupt
the
cell
monolayer
during
the
wash
steps
.
For
biochemical
analysis
of
connexin43
remaining
with
the
monolayer,
300
pl
of
lysis
buffer
(described
earlier)
containing
0
.6%
SDS
was
added
to the tissue
culture
dish,
and
the
lysates
boiled
for
3 min
before
itnmuno-
precipitation
of
connexin43
.
For
immunofluorescent
localization
of
connex-
in43,
the
extracted
or
mock-extracted
cells
were
fixed
for
1
h
at
room
tem-
peratures
in
1
%
formaldehyde
(prepared
freshly
from
paraformaldehyde)
in
PBS
(final
pH, 7
.4)
.
The
fixed
cultures
were
treated
with
PBS
containing
0.2%
Triton
X-100
and
5
%
normal
goat
serum,
incubated
overnight
at
4°C
with
a
1
:100
dilution
of
affinity-purified
antibodies
to
connexin43
(252-
271),
and
incubated
with
1
:500
rhodamine-conjugated
goat
anti-rabbit
IgG
(Boehringer
Mannheim
Biochemicals,
Indianapolis,
IN)
as
previously
de-
scribed
(Musil
et al
.,
1990b)
.
The
cultures
were
photographed
on
an
Axio-
scope
microscope
(Carl
Zeiss,
Inc
.,
Oberkochen,
Germwny)
fitted
with
the
appropriate
filters
.
Electron
Microscopy
NRK
cell
cultures
either
mock
extracted
in
the
absence
of
Triton
or
ex-
tracted
with
Triton
for
30
min
(at
14°C
or
4°C) were
fixed
by
addition
of
2.5%
glutaraldehyde
and
1%
tannic
acid
in
0
.1
M
cacodylate
(pH
7
.4)
directly to the tissue
culture
dish
.
The
specimens
were
then
postfixed
in 1
%
OsO4
and
stained
with
1%
uranyl
acetate
in
situ,
after
which
cell
sheets
were
scraped
from
the
tissue
culture
dish
prior
to
dehydration
and embed-
ding
in
epoxy
resins
.
Thin
sections
were
cut
and
observed
using
a
JEOL
100CX
electron
microscope
operating
at
60
kV
.
Cell
Surface
Biotinylation
Biotinylation
of
cell
monolayers
was
conducted
using
a
modification
of
the
procedure
described
by
Le
Bivic
et
al
.
(1989
and
1990b)
.
Metabolically
la-
beled
60
mm
cell
cultures
were
rinsed
three
times
at
4°C
with
PBS
contain-
ing
0
.1
mM
CaCh
and
1 .0
mM
MgC12
(PBS+)
and
incubated
on
ice for
10
min
in
the
same
buffer
.
After
rinsing
the
cultures
twice
with
PBS+,
cell-
surface
biotinylation
was
initiated
by
addition
of
2
nil
of
0
.5
mg/nil
NHS-LC-biotin
in
PBS
+
to each
dish
and
the
cultures
were
then
incubated
for
30 min
with
gentle
agitation
.
The
reaction
was
quenched
by
rinsing
the
cultures
five
times
with L-15
medium
supplemented
with 15
mM
glycine
;
there
was
a
10-min
incubation
in
L-15/glycine
between
the
third
and
fourth
washes
.
All
manipulations
were
conducted
on
ice in a
room
at
4°C
.
In
some
cases, the
biotinylated
monolayers
were
immediately
lysed
with
500
ul
of
lysis
buffer
(described
earlier)
supplemented
with
0.6%
SDS
and
10
mM
glycine
and
boiledfor
3
min, or
subjected
to
our
standard
Triton
solubiliza-
tion
assay
in the
presence of 10
mM
glycine
.
Alternatively,
the
cells
were
chased
for
1-3
h
at
37
°C
in
L-15
medium
supplemented
with 10
mM
Hepes,
10%
FCS,
and
0
.5
mM
methionine
before
lysis
.
Connexin43
was
then
im-
Musil
and
Goodenough
Gap
Junction
Assembly
munoprecipitated
from
all
samples
using
affinity-purified
antibodies
to
connexin43
(252-271)
and
protein
A-Sepharose
beads
as
previously
de-
scribed
.
Immunoprecipitated
connexin43
was
eluted
from
the
beads
by
boil-
ing
them
for
4 min
in
40
Ed
of
immunoprecipitation
buffer
containing
10
%
SDS
and
2
mM
PMSF
One
fifteenth
of
the
eluted
connexin43
was
added
to
SDS-PAGE
sample
buffer
and
reserved
as
a
sample
of
total
cellular
connexin43
.
The
remainder was
diluted
with
immunoprecipitation
buffer
supplemented
with
0.5% BSA,
10
mM
N-ethylmaleimide,
2
mM
PMSF,
and
enough
Triton
X-100
to
bring
the
final
detergent
concentration
to
1%
Triton
and
0.2%
SDS
.
Biotinylated
connexin43
was
recovered
from
this
sample
by
a
second
round
of
precipitation
with
avidin-agarose (35
pl,
50%
slurry), after
which
the
beads
were
washed
as
described
above
following
immunoprecipitation
of
connexin43
with
protein
A-Sepharose
.
The
bio-
tinylated
connexin43
was
eluted
from
the
avidin-agarose
by
boiling
in
SDS-
PAGE
sample
buffer
for 5
min
and
analyzed
(along
with
the
total
cellular
connexin43
sample)
by
SDS-PAGE
.
In
control
experiments,
connexin43
was
immunoprecipitated
from
[
35
S]methionine-labeled,
unbiotinylated
NRK
cells
and
then
subjected
to
a
second
round
of
precipitation
with
pro-
tein
A-Sepharose
instead
of
avidin-agarose
.
When
analyzed by
SDS-PAGE,
the
amount
and
phosphorylation
state
of
this
material
appeared
identical
to
[35
S]connexin43
that
had
been
subjected
to only
a single
round
of
precipi-
tation,
indicating
that
artifactual
dephosphorylation
or
degradation
of
con-
nexín43
did
not
occur
during
the
double
precipitation
procedure
(not
shown)
.
Inhibition
of
Junctional
Permeability
with
Heptanol
or
100%
CO
2
Heptanol
(Fisher
Scientific,
Fair
Lawn,
NJ)
was
diluted
fresh daily
1:4 in
ethanol
and
used
at a
final
concentration
of
3
.5
mM
as
previously
described
(Musil
et
al
.,
1990b)
.
For
cytoplasmic
acidification
of
NRK
cells,
hu-
midified
100%
C02
was
bubbled
through
MEM
formulated with Earles
salts
(plus
L-glutamine
and 1% FCS)
for
10
min
to
reduce
the
pH
of
the
medium
to
6
.0
.
NRK
cultures
were
then
immediately
incubated
in
this
medium
for
10-15
min
at
37°C
(Schuetze
and
Goodenough,
1982)
.
Results
Differential
TYiton
X-100
Solubility
of
Phosphorylated
and
Nonphosphorylated
Forms
of
Connexin43
Our
previous
studies
characterized
the
biosynthesis
of
con-
nexin43
in
NRK
cells,
a
gap
junctional
communication-
competent
cell
line
that
assembles
connexin43
into
large
gap
junctional
plaques
(Musil
et
al
.,
1990b)
.
When
NRK
cul-
tures
are
metabolically
labeled
with
[
35
S]methionine
for
5
h,
lysed,
and
immunoprecipitated
with
antibodies
affinity
purified
from
rabbit
anti-connexin43
(252-271)
serum
(Beyer
et al
.,
1989), three
[
35
S]methionine-labeled
connexin43-
related
species are
obtained
(Fig
.
1
A,
lane
1)
.
The
fastest
migrating
of
these
bands
(M
r ,
-42,000)
has
been
shown
to
represent
newly
synthesized
connexin43
and
comigrates
with
connexin43
translated
in
a
cell-free
reticulocyte
lysate
system
(Musil
et al
.,
1990a)
.
Since
this
species
is
not
meta-
bolically
labeled
with
32P
(Fig
.
1
A,
lane 2)
and
is
insensi-
tive
to
treatment
with
alkaline
phosphatase
(Musil
et
al
.,
1990a
and
b),
we
refer
to
it
as
connexin43-NP
(not
phos-
phorylated)
.
Pulse-chase
studies
have
determined
that
connexin43-NP
is
posttranslationally
modified
in
NRK
cells
to a
44-kD
and
then to
a 46-kD
species
.
Both of
these
bands
incorporate
32P
(Fig
.
1
A,
lane 2)
and
are
converted
to the
connexin43-
NP
form
by
alkaline
phosphatase
treatment
(Musil
et
al
.,
1990a
and
b)
.
This
finding
demonstrates
that
the
posttransla-
tional
shift
in
the
apparent
molecular
weight
of connexin43
is
due
exclusively
to the addition of
phosphate
.
The
lower
and
upper bands
of the
phosphorylated
connexin43
doublet
have
been
designated
connexin43-P,
and
connexin-43P
2 ,
respectively
(Musil
et
al
.,
1990b)
.
A
similar pattern
of
135
9
on July 10, 2011jcb.rupress.orgDownloaded from
Published December 1, 1991
Figure
1
.
Solubility
of
phosphorylated
and
nonphosphorylated
forms
of
connexin43
in
Triton
X-100
.
Confluent
cultures of
communication-
competent
NRK
(A)
or
communication-deficient
S180
(B)
or
L929
(C)
cells
were
metabolically
labeled
with
either
[
35
S]methionine
(lanes
1,
and
3-5)
or
P
2
P]O
a
(lane
2)
for
5
h
.
In
lanes
1
and
2,
the
cells
were
immediately
lysed
and immunoprecipitated
with
affinity-purified
antibodies
to
connexin43
(252-271)
.
Lanes
3-5
are
from
a
separate
experiment
in
which
cells
were homogenized
and
incubated
with
Triton
X-100 under
our
standard
solubilization
conditions
(1%
Triton
in
PBS,
30 min,
4°C)
.
Half
of
the
Triton-treated
lysate
was
reserved
on
ice
(total
cellular lysate),
whereas
the
remainder
was
subjected
to
centrifugation
at
100,000
g
for
50
min
.
Connexin43
was
then
immu-
noprecipitated
from
equal
amounts
of
the
total
cellular
lysate
(T,
lane
3),
Triton-soluble
supernatant
(S,
lane
4),
or
Triton-insoluble
pellet
(P,
lane
5)
fractions
.
Connexin43-P,
and
-P
Z
immunoprecipitated
from
the insoluble
fraction
routinely
migrated
slightly
slower
than
that
recovered
from
the
total
cellular
lysate
because
of
a
difference in
the Triton
content
of
the
two
fractions
during
the
SDS
denaturation
step
(A,
lanes
3
vs
.
5
;
see Materials
and
Methods)
.
connexin43
biosynthesis
and
phosphorylation has
been
ob-
served
in
other
communication-competent
cell
types
(Musil
et
al
.,
1990a
and
b)
.
In addition to these three
major forms
of
connexin43,
a
barely
detectable species
that
migrates
slightly
slower
than
connexin43-NP
and
that
can
be metabol-
ically
labeled
with
31
P was
occasionally
observed
;
this
band
may
represent
partially
dephosphorylated
connexin43 or
a
minor
additional
form
of
connexin43
and
will
not
be
consid-
ered
further
.
The
phosphorylated
and
nonphosphorylated
forms
of
connexin43
displayed
different
solubilities
in
the
nonionic
detergent
Triton
X-100
(Fig
.
1
A)
.
3-d-old,
newly
confluent
monolayers of
NRK
cells
were
metabolically labeled
for 5
h
with
[
35
SJmethioníne,
after
which
the
cells
were
scraped
from
the
tissue
culture dish
and
disrupted
by
repeated
pas-
The
Journal
of
Cell
Biology,
Volume
115,
1991
sage
through
a
25-gauge
needle
.
The
cell
lysates
were
then
incubated
under
isotonic
conditions
with
1%
(wt/wt,
final
concentration)
Triton
X-100
for
30
min
at
4°C
.
Half
of the
lysate
was
stored
on
ice
to
serve
as
the
total
connexin43 sam-
ple
;
the
remainder
was
subjected
to
centrifugation
at
100,000
g
for
50 min
to
separate
the Triton-soluble
from
the
Triton-insoluble
material
.
Immunoprecipitatíon
of
con-
nexin43
from
the
total
(Fig
.
1
A,
lane
3),
the
supernatant
(Triton-soluble,
lane
4),
and
pellet
(Triton-insoluble,
lane
5)
fractions
revealed
that
-80-90%
of
connexin43-NP
was
solubilized
by
1
%
Triton
X-100
.
In
contrast,
connexin43-P
2
was
quantitatively
recovered
in the
pellet
fraction
and
re-
mained
insoluble
upon
reextraction
with
1%
Triton
(not
shown)
.
Connexin43-P,
was
not
well
resolved
from
con-
nexin43-P
Z
after
Triton
treatment
and
displayed
an
inter-
136
0
on July 10, 2011jcb.rupress.orgDownloaded from
Published December 1, 1991
mediate
and somewhat
variable
sensitivity
to
Triton
;
most
of
the
connexin-43-P,
fractionated
with
connexin43-P
2
(Fig
.
1
A,
lane
S),
but
some
was
detectable
in
the
Triton-
soluble
supernatant
(lane
4)
.
Similar
results
were
obtained
with
2-5-day-old
cultures,
indicating
that cell
density
and/or
age
were
not
critical
factors
in
determining
the
resistance
of
connexin43
to
Triton
.
Connexin43
from
S180L
cells,
an-
other
communication-competent
cell
line that
forms
large
gap
junctional
plaques
(Mege
et al
.,
1988
;
Musil
et al
.,
19906),
displayed
a
pattern
of
Triton
X-100
solubility identi-
cal
to
that
observed
in
NRK
cells
(data
not
shown)
.
Connexin43
is
also
synthesized
by
certain
communication-
defective
cell
lines,
including
mouse
sarcoma
180
(5180)
and
fibroblastic
L929
cells
(Musil
et
al
.,
19906)
.
These
cells
have
been
shown
to
be
severely
deficient
(S180
cells)
or
com-
pletely
lacking
(L929
cells)
in
morphologically
or
physio-
logically
recognizable
gap
junctions
(Furshpan
and
Potter,
1968
;
Mege
et
al
.,
1988
;
Larson
et
al
.,
1990)
.
S1ß0
cells
process
connexin43
to the
connexin43-P,
form
to
a
lesser
extent
than
do
NRK
cells
(compare
Fig
.
1
B, lane
1
with
Fig
.
1
A,
lane
1)
and
lack
all
but
a
trace
of
connexin43-P
2 ,
which was
detectable
only
as a
fraction
of
32p
labeled
con-
nexin43
(Fig
.
1
B, lane 2)
.
L929
cells
contain
neither
con-
nexin43-P,
nor-P2
(Fig
.
1
C, lanes 1
and
2
;
see also
Musil
et al
.,
19906)
.
When
S180
(Fig
.
1
B,
lanes
3-S)
or
L929
(Fig
.
1
C,
lanes
3-5)
cultures
were
subjected
to the
Triton
solubility
assay
exactly
as
described
for
NRK
cells,
connex-
in43-NP
was
immunoprecipitated
only
from
the soluble
fraction
;
recovery
ranged
from
66-90
%
.
Connexin43
is
.
a
s
metabolically
stable
in
S180
and L929
cells
as
in
NRK
cells
(Musil
et al
.,
19906),
ruling
out the
possibility
that
com-
munication-deficient
cells
degraded
connexin43
at
a
rate
faster
than
the
rate
of
conversion
to
the Triton-insoluble
form
.
Together,
these
results
suggest
that
solubility
in
Triton
X-100
is
correlated
with
the
phosphorylation
state
of
connex-
in43
rather
than with
its
cell
of origin
.
Table
I
summarizes
the
solubility
properties
of
connexin43
in
Triton
X-100
.
The
differential
solubility
of
phosphorylated
and
nonphosphorylated
forms
of
connexin43
in
NRK
cells
depicted
in
Fig
.
1
A
were
obtained
with
Triton
X-100
con-
centrations
ranging
from 0
.04-4
.0%
and was
unaffected
by
the
presence
of
2-mercaptoethanol,
divalent
cations,
or
prolongation
of
the
solubilization
period
from
30
min
to
24 h
.
Thus,
the resistance
of
connexin43-P
2
to
Triton
was
not
due
to
disulfide
bonds
or
to
Ca
2
+-
or
Mgt+-dependent
polymerization,
possibilities
raised
by
earlier
studies
(Zam-
pighi
and
Robertson,
1973
;
Manjunath
and
Page,
1986)
.
The
addition
of
high
salt
concentration
(0
.5
M
NaCl)
to
the
sol-
ubilization buffer
did,
however,
render
connexin43-P
2
par-
tially
soluble
in
1
%
Triton
X-100,
a
fact
suggesting
that
ionic
interactions
may
be
involved
in
maintaining
connexin43-P
2
in
a Triton-insoluble
state
.
When
the
temperature
of
the
iso-
tonic
Triton
treatment
was
raised
from
4°C
to
25°C,
connex-
in43-P
2
remained
insoluble
in
0
.04%
Triton
but
was
par-
tially
solubilized
by
1%
Triton
X-100,
probably
reflecting
increased
detergent
activity
at
higher temperatures
.
Neither
phosphorylated
nor
nonphosphorylated
forms
ofconnexin43
were
solubilized
by
levels
of
Triton
X-100below
the
critical
micellar
concentration
under
any
of
the
conditions
tested
.
Solubility
of
Connexin43
in
Other
Detergents
Relative
insolubility
in the ionic
detergent
N-lauryl
sarcosine
is
the basis
for
several
procedures
for
isolation
of
gap
junc-
Musil
and
Goodenough
Gap
Junction
Assembly
Table
L
Detergent
Solubility
of
Connexin43
in
NRK
Cells
All
forms
of
connexin43
insoluble
0
.005%
Triton
X-100
;
4°C
or
25°C
(<cmc)
0
.005%
Triton
X-100
in
PBS
+
500
mM
NaCl
0
.005%
LDAO
(<cmc)
5
mM
Tris,
pH
10
.0
;
30
min,
25°C
Connexin43-NP
soluble,
connexin43-P2
insoluble
0
.04%-4
.0%
Triton
X-100
(>cmc)
1
.0%
Triton
X-100
;
0
.5-24
h,
4°C
1
.0%
Triton
X-100
:
f
5
%
2-mercaptoethanol
t
1
mM
Ca
2
*,
Mgt'
t
5
mM
EDTA,
EGTA
0
.04%
Triton
X-100
;
30
min,
25°C
Connexin43-NP
soluble,
connexin43-P
2
partially
soluble
1
.0%
Triton
X-100
;
30 min,
25°C
1
.0%
Triton
X-100
in
PBS
+
500
mM
NaCl
0
.3%
deoxycholate
t
10
mM
N-ethylmaleimide'
All
forms
of
connexin43
soluble
0
.3%
N-lauryl
sarcosine
in
5
mM
Tris,
pH
10
.0
;
10
min,
25 °C
0
.07%
LDAO
(>cmc)
Three-d-0Id,
newly
confluent
cultures
of
NRK
cells
were
metabolically
labeled
for 5
h
with
["S]methionine
and
then
homogenized
as
described
in
Materials
and
Methods
.
The
cell
lysates
were
subsequently
incubated
under
the condi-
tions specified
above
;
unless
noted
otherwise,
all
reactions
were
conducted
in
PBS,
2
mM
EDTA,
2
mM
EGTA,
pH
7
.4,
for
30
min
at
4°C
.
The
lysates
were
then
subjected
to
centrifugation
at
100,000
g
for
50
min
at
4°C,
after
which
connexin43
was
immunoprecipitated
from
the soluble
supernatant
and
insolu-
ble
pellet
fractions
and
analyzed
by
SDS-PAGE
.
'
Connexin43-P
2
was more
soluble in
0.3%
deoxycholate
in
PBS,
pH
7
.4,
than
in 5
mM
Tris,
pH
10
.0
.
LDAO,
lauryldimethylamine
oxide
;
cmc,
critical
micellar
concentration
.
tions
from
whole
organs
such
as heart
(Kensler
and Good-
enough,
1980
;
Manjunath
et al
.,
1984)
and
liver
(Good-
enough
and
Stoeckenius,
1972
;
Hertzberg
and
Gilula,
1979)
.
It
was
therefore
of
interest
to
examine
the
behavior
of
con-
nexin43
from
NRK
cells
in
this
detergent (Fig
.
2
A)
.
Lysates
were
prepared
from
[
35
S]methionine-labeled
NRK
cultures
in
5
mM
Tris,
pH
10
.0,
and
incubated
with
0.3%
N-lauryl
sarcosine
for
10
min
at
room
temperature,
solubilization
conditions
identical
to
those
used
during
the
preparation
of
connexin43-containing
gap
junctions
from
rat
heart
(Kensler
and
Goodenough,
1980
;
Manjunath
et al
.,
1984)
.
After
cen-
trifugation
at
100,000
g
for
50
min,
>97%
of
the
total
(both
phosphorylated
and
nonphosphorylated)
connexin43
recov-
ered
by
immunoprecipitation
was
in the soluble
fraction
(Fig
.
2
A,
lane 2)
.
Overexposure
of
the
fluorograph
revealed
a
small
amount
of
connexin43-P
2
in
the insoluble
fraction
;
whether
this
material represents
bona
fide
sarcosine-insolu-
ble
connexin43 or
is
due
to
nonspecific
trapping
of soluble
connexin43
in
the
pellet
is
unknown
.
The
ability
to
solubilize
connexin43-P2
was
not
unique
to
sarcosine
since
identical
results
were
obtained
with
the
nonionic
detergent
lauryl-
dimethylamine
oxide
.
Phosphorylated
forms
of connexin43
were
also
partially
soluble
in
deoxycholate,
the extent
of
solubilization
being
dependent
on
the
composition
of
the
ly-
sis
buffer (Table
I)
.
The
lack
of
resistance
ofconnexin43
in
NRK
cells
to
N-lau-
ryl
sarcosine
appeared
to
be
at
variance
with
the
reported
sarcosine-insolubility
of
connexin43
in
rat
heart
(Kensler
and
Goodenough,
1980
;
Manjunath
et al
.,
1984)
.
A
poten-
tial
explanation
for
this
discrepancy
was
that
connexin43
is
somehow
organized
or
assembled
differently
in
tissue
culture
136
1
on July 10, 2011jcb.rupress.orgDownloaded from
Published December 1, 1991
Figure
2
.
Comparison
of
the
detergent
solubility
of
connexin43
from
NRK
cells
and
from
chick
lenses
.
NRK
cultures
(A)
or
intact
em-
bryonic
chick
lenses (B)
were
metabolically
labeled
for
5
h
with
[
35
S]methionine,
homogenized,
and
incubated
either
with
0
.3%
N-lauryl
sarcosine
in
5
mM
Tris
(pH
10
.0)
for
10
min
at
25°C
(A
and
B,
lanes
1-3),
or
with
1%
Triton
X-100
under our
standard
solubilization
conditions
(PBS
pH
7
.4,
30
min,
4°C)
.
(B,
lanes
4-6)
.
Half
of
the
cell
lysate
was
then
centrifuged
at
100,000
g
for
50
min,
after
which
connexin43
was immunoprecipitated
from
equal
amounts
of
the
total
cellular
lysate (lanes
marked
T),
detergent-soluble
supernatant
(lanes
marked
S),
or
detergent-insoluble
pellet (lanes
marked
P)
fractions
.
cells
than
in
whole
organs
.
To
test this
possibility,
we
ex-
amined
the
detergent
solubility
of
connexin43
in
embryonic
(d
10)
chick
lens
.
Our
previous
studies
demonstrated
that
ep-
ithelial
cells
from
this
organ
synthesize
connexin43
and
in-
corporate
it
into
large,
communication-competent gap
func-
tional
plaques
(Musil
et
al
.,
1990x)
.
Intact
lenses
(stripped
of
associated
ciliary
epithelium)
were
incubated
with
[
35
S]-
methionine
for 5
h
.
The
labeled lenses
were
then
disrupted
at
4
°
C
in
hypotonic
Tris
buffer
and
incubated
with
1
%
Triton
X-100
(in
PBS,
at
4°C)
or
0
.3
%
N-lauryl sarcosine
(in
5
mM
Tris,
pH
10,
at
25°C),
exactly
as
described
for
NRK
cell
ly-
sates
.
As
shown
in
Figure
2
B,
the pattern
of
solubility
of
[
35
S]methionine-labeled
connexin43
in
chick
lens
epithelium
was
similar to
that
obtained
in
NRK
cells
in
both
detergents
(compare
Fig
.
1
A,
lanes
3-S
with
Fig
.
2
B,
lanes
4-6,
Fig
.
2
A,
lanes
1-3
with
Fig
.
2
B,
lanes
1-3)
.
By
these
criteria,
then,
the
solubility
properties
of connexin43
from
an
organ
(lens)
are
similar
to
those
of connexin43
from
tissue
culture
cells
(NRK
cells)
.
Possible
explanations
for
the
recovery
of
sarcosine-insoluble
gap
junctions
from
heart are
considered
in
the
Discussion
.
The
Journal
of
Cell
Biology,
Volume
115,
1991
time
Course
of
Acquisition
of
Mton
X-100
Resistance
by
Connexin43
In
communication-competent
cell
types
thus
far
examined,
connexin43-NP
is
the
kinetic
precursor
of
connexin43-P2
.
It
therefore
seemed
likely
that
connexin43
is
synthesized
in
a
Triton-soluble
form
that
acquires
Triton
resistance
as
it
matures
.
This
possibility
was
confirmed
in
a
pulse-chase
ex-
periment
(Fig
.
3)
.
Confluent
monolayers of
NRK
cells
were
labeled
with
[
35
S]methionine
for
40
min
and
then
chased
in
the
presence
of
an
excess
of
unlabeled
methionine
for
<6
h
before
cell
lysis
and
assessment
of connexin43
Triton
solu-
bility
under
our
standard
assay
conditions
.
As
expected,
I
35
S]methionine-connexin43
examined
immediately
after
the
pulse
period
was
entirely
in the
connexin43-NP form
and
was
quantitatively
solubilized
by
1
%
Triton
(Fig
.
3,
lanes
1-3)
.
After
1
h
of chase,
about
one
third
of
[
35
S]methio-
nine-connexin43
had
been
phosphorylated
to
connexin43-
P,,
with
no
detectable
conversion
to
connexin43-P
2
(Fig
.
3,
lane
4)
.
This
connexin43-P,
was
recovered
in
both
the
Triton-soluble (Fig
.
3,
lane
S)
and
-insoluble
(Fig
.
3,
lane
136
2
on July 10, 2011jcb.rupress.orgDownloaded from
Published December 1, 1991
Figure
3
.
Pulse-chase
analysis
of
the
Triton
solubility
of
Connexin43
in
NRK
cells
.
Confluent
monolayers
of
NRK
cells
were
metabolically
labeled
with
[
35
S]methionine
for
40
min
and
chased
for
0,
1,
4,
or 6
h
.
The
cultures
were
then
homogenized
and
subjected
to
the
standard
Triton
solubilization
assay
(1%
Triton
X-100
in
PBS,
30
min,
4°C),
after
which
equal
amounts
of
the
various
fractions
were
immunoprecipi-
tated
with
affinity
purified
anti-Connexin43
(252-271)
antibodies
.
Lanes
marked
TQ,
4,
7,
and
10),
total
cellular
Connexin43
;
lanes
marked
S
(2,
S, 8,
and
11),
Triton-soluble
Connexin43
;
lanes
marked
P
(3,
6,
9,
and
12),
Triton-insoluble
Connexin43
.
Lanes
7-12
were
exposed
longer
than
the
other
lanes
to
double
the
intensity
of
the
Connexin43
signal
.
Connexin43-P,
(lanes
6
and
9)
and Connexin43-P2
(lanes
9
and
12),
immunoprecipitated
from
the insoluble
fraction,
routinely
migrated
slightly
slower
than
that
recovered
from
the
corresponding
total
cellular
lysate
due
to
a
difference
in
the
Triton
content
of
the
two
fractions
during
the
SDS
denaturation
step
(see
Materials
and
Methods)
.
6)
fractions,
similar
to
the
distribution
of
Connexin43-P,
in
NRK
cells
labeled
to
steady
state
(Fig
.
1
A,
lanes
3-S)
.
Connexin43-NP
remained
predominantly
Triton-soluble,
al-
though
a
minor
fraction
(-20
%)
was
consistently
recovered
in
the
Triton-insoluble
pellet
beginning about
this
time
.
At
4 h
of
chase,
66%
of
the
[
35
S]methionine-Connexin43
syn-
thesized
during
the
pulse
period
had
been
degraded,
a
find-
ing consistent
with
the
2-2
.5
h
half-life
of
Connexin43
in
NRK
and
other
cell
types
(Musil
et al
.,
1990b)
.
The
remain-
ing
[
35
S]methionine-Connexin43
was
largely
processed
to
the
P
2
form
and
was
quantitatively
resistant
to
Triton
solu-
bilization,
whereas
the
residual
[
35
S]methionine-connex-
in43-NP
detectable
at this
time
was
mainly
Triton
soluble
(Fig
.
3,
lanes 7-9)
.
Low
levels
of
Triton-insoluble
connex-
in43-NP
as well as Triton-soluble
Connexin43-P, were,
how-
ever,
still
present
.
The
rapid
turnover
rate
ofConnexin43
thwarted
attempts
to
determine whether
Triton-insoluble
Connexin43-NP
served
as
a
precursor
to
Connexin43-P2
or
was
degraded
without
being
phosphorylated
;
similarly,
the
fate
of
Triton-soluble
Connexin43-P,
was
unknown
.
It
was
thus
not
possible
to
conclude
from
these
data
whether
most
of the
connexin43-
NP
was
phosphorylated
to the
Connexin43-P,
form
before
or
after
becoming
Triton insoluble
.
In
either case,
connex-
in43-P2
was
completely
resistant
to
Triton
and
appeared
to
remain
so
throughout
the
lifetime
of the protein (Fig
.
3,
lanes
10-12)
.
Taken
together,
these
results
indicate
a strong
tem-
poral
correlation
between
processing
of
Connexin43
to
the
Musil
and
Goodenough
Cap
Junction
Assembly
P
2
form
and
the
acquisition
of
insolubility
in
Triton
X-100
.
However,
whether
these
two
events
normally
occur
in
rapid
succession
or
simultaneously
is
not
known
.
Morphological
Localization
of
the
7Wton-insoluble
Pbol
of
Connexin43
We
exploited
the
Triton
resistance
of
Connexin43-P2
to
de-
termine
the
intracellular
distribution
of
the various
forms
of
Connexin43
using
an
in
situ
extraction
procedure
(Fig
.
4)
.
Monolayer
cultures
of
NRK
cells
were
washed
three
times
and
then
extracted
by
addition
of
1%
Triton
X-100-contain-
ing
isotonic buffer
directly
to
the tissue
culture
dish
.
After
a
30-min
incubation, the cultures
were
carefully
rinsed
five
times
to
remove
solubilized
cellular
components
.
The
Con-
nexin43 remaining
with
the
extracted
monolayer
was
ana-
lyzed
either
biochemically
by
immunoprecipitation
or
mor-
phologically
by
immunofluorescence
using
affinity-purified
antibodies
to
Connexin43
(252-271)
.
Preliminary
studies
demonstrated
inefficient
solubilization
of
Connexin43-NP
af-
ter
in
situ
extraction
of
[
35
S]methionine-labeled
NRK
cells
at
4°C
for
<2
h
(data not
shown)
.
When
the extraction
tem-
perature
was
raised
from 4°C
to
14°C,
however,
75%
of
[
35
S]methionine-Connexin43-NP
was
released
from
the
cells
within
30
min
(Fig
.
4
A,
inset,
lane
1
vs
.
2)
;
the
proportion
released
was
comparable
to
the
fraction
of
Connexin43-NP
solubilized
from
disrupted
NRK
cell
lysates
in the
standard
Triton
solubilization
assay (Fig
.
1
A,
lanes
3-S)
.
In
contrast,
136
3
on July 10, 2011jcb.rupress.orgDownloaded from
Published December 1, 1991
Figure
4
.
In
situ
extraction
of
Triton-soluble
connexin43-NP
in
NRK
cells
.
NRK
monolayers
were
incubated
at
14°C
for
30
min
under
isotonic
conditions
in
either
the
absence
(A)
or
presence
(B)
of
1%
Triton
X-100
.
After
extensive
washing
to
remove
solubilized
material,
the
cultures
were
fixed
and
processed
for
immunofluorescence
using
anti-connexin43
(252-271)
antibodies
followed
by rhodamine-labeled
goat
anti-rabbit
IgG
.
(Inset)
NRK
cultures
metabolically
labeled
with
[
35
S]methionine
for
5 h
were
extracted
either
with
(lane 2) or
with-
out
(lane
1)
1%
Triton
exactly
as
described
for
unlabeled
cells
and
then
immunoprecipitated
with
anti-connexin43
(252-271)
antibodies
.
Note
that
["S]methionine-connexin43-P2
is
resistant
to
Triton
solubilization
and
appears
to
be
localized
to brightly
staining
maculae
at
cell-cell interfaces
.
Figure
5
.
Thin-section
electron
microscopy of
gap
junctions
in
Triton-extracted
and
mock-extracted
NRK
cultures
.
(A)
Low-magnification
photomicrograph
of
a
gap
junction
between
two
mock-extracted
NRK
cells
.
Bar,
0
.2
Am
.
(B)
Gap
junction
in
an
NRK
culture
subjected
to
in
situ
extraction
with
1%
Triton
at
4°C
.
Bar,
0
.5
Am
.
(Inset)
Higher-magnification
view
of
the
same
junction
.
Bar,
0
.1
j,m
.
(C and
D)
Gap
junctions
in
NRK
cells
extracted
in
situ
with
1%
Triton
at
14°C
.
Some
junctions
appear morphologically
intact
(C),
whereas
others
have
partially
disassembled
into
short,
double-membrane
structures
(D)
.
Bar,
0
.2
am
.
(D,
inset)
Higher
magnification
of
D,
showing
pen-
talaminar
structure
of
the
gap
junction
fragments
.
Bar,
0
.1
jm
.
The
Journal
of
Cell
Biology,
Volume
115,
1991
136
4
on July 10, 2011jcb.rupress.orgDownloaded from
Published December 1, 1991
[
35
S]methionine-connexin43-P
2
was
resistant
to
this treat-
ment
and
was
quantitatively
recovered
with
the
extracted
monolayer
(Fig
.
4 A,
inset)
.
Immunohistochemical
studies
revealed
that
control
monolayers
incubated
with only
buffer
before
fixation
(Fig
.
4
A)
showed
a
pattern
of
macular
immu-
noreactivity
at
cell-cell
interfaces
typical
of
gap
junctions
(Beyer
et
al
.,
1989)
.
In
addition,
intracellular
structures
morphologically
similar
to the
Golgi
apparatus
were
la-
beled,
consistent
with
our
previous
results
with
these
cells
(Musil
et al
.,
1990b)
.
In
situ
extraction
of
NRK
cells
with
1%
Triton
X-100
at
14°C
before
fixation
abolished
this
Golgi-like
immunofluorescent
staining,
leaving the
macular
signal
at
cell-cell
interfaces
(Fig
.
4B)
.
A
fine,
speckled
pat-
tern
of
staining
distributed
throughout
the extracted
cells
was
also
visible
.
This
diffuse
immunoreactivity
was
obtained
with
preimmune
serum
as
well,
indicating
that
it
does
not
represent
connexin43
.
In
contrast,
the
plaque-like
signal
at
cell-cell
interfaces
was
not
detectable
when
preimmune
se-
rum
was
substituted
for
the antibodies to
connexin43
(not
shown)
.
To
confirm
that
the
Triton-resistant
maculae
were
gap
junctional
plaques,
Triton-extracted
and
unextracted
NRK
cells
were examined
by
thin-section
EM
(Fig
.
5)
.
Mock-
extracted
cells
not
exposed
to
Triton
contained pentalaminar
structures at
cell-cell
interfaces
that
are diagnostic
of
gap
junctions
(Fig
.
5
A)
.
These
profiles
were
maintained
when
NRK
cells
were
extracted
with
1%
Triton
at
4°C
before
in
situ
fixation,
conditions
under
which
connexin43-P
2
is
quantitatively
recovered
with
the
monolayer
but
recogniz-
able
intracellular
structures
have been
largely extracted (Fig
.
5
B)
.
When
the
temperature
of
the Triton
treatment
was
raised to 14
°
C
to
maximize
the extraction of
connexin43-
NP,
morphologically
intact
gap
junctional
plaques
were
still
detectable
(Fig
.
5
C),
but
some
appeared
to
be
in
the
process
of
dissociating
into
short,
double-membrane
structures
that
became
cross-linked to the extracted
monolayer
upon
fixa-
The
Journal
of
Cell
Biology,
Volume
115,
1991
Figure
6
.
Inhibition
of con-
nexin43
posttranslational
pro-
cessing
at
20°C
.
Duplicate
cul-
tures
of
confluent
NRK
cells
were
metabolically
labeled
with
1
35
5]methionine
at
20°C
for
4
h
.
The
monolayers
were
then
subjected
to
our
standard
Tri-
ton
X-100
solubility
assay
(I%
Triton
in
PBS,
30 min, 4°C),
either
immediately
after
label-
ing
at
20°C
(lanes
1-3) or
af-
ter
a
2-h chase
at
37°C
in
L-15
medium
supplemented
with
10
mM
HEPES,
10%
FCS,
and
0
.5
mM
methionine
(lanes
4-6)
.
Connexin43 was
immu-
noprecipitated
from
equal
vol-
umes
of
the
total
cell
lysate
(lanes
1
and
4),
Triton-soluble
(lanes
2
and
5),
and
Triton-
insoluble
(lanes
3
and
6)
frac-
tions
.
tion
(Fig
.
5
D)
.
Morphologically
similar
disassembly
of
gap
junctions
in the
presence
of
detergents
has
previously
been
observed
in
Mauthner
cell
synaptic
disks
(Zampighi
and
Robertson,
1973)
and
in
lens
fiber
junctions
(Kistler
and
Bullivant,
1988)
.
Thus,
although
Triton
treatment
at
14
°
C
affected
the
ultrastructure
of
gap
junctions,
they
were
still
recovered
with
the
extracted
monolayer
.
Taken
together,
the
results
depicted
in
Fig
.
4
and
5
indicate
that (Triton-
insoluble)
connexin43-P
2 is
localized
primarily
in
gap
func-
tional
plaques
.
Nonphosphorylated,
Triton-soluble
connex-
in43-NP
appears
to
be
predominantly
intracellular,
although
diffusely
distributed
connexin43
would
not
be
detectable by
morphological
techniques
.
Ransport
of
Connexin43-NP
to
the
Plasma
Membrane
in
NRK
Cells
Although
the
results
depicted
in
Fig
.
4
demonstrated
that
connexin43-P
2
accumulated
on
the
cell
surface in junctional
plaques,
the
cellular
location
of
conversion
of
newly
synthe-
sized
connexin43-NP
to the
P
z
form
was
not
established
.
Matlin
and
Simons
(1983)
found
that
incubation
of
mam-
malian
cells
at
20
°
C
reversibly
blocks
transport of
nascent
secretory
and
integral
membrane
proteins
within
the
trans-
Golgi
region
.
To
examine
whether
phosphorylation
of
con-
nexin43
and/or
its
acquisition
of
Triton
resistance
occurs
be-
fore
arrival
in
this
intracellular
compartment,
confluent
monolayers
of
NRK
cells
were
metabolically
labeled
with
[ 3
IS]methionine
for
4
h
at
20°C
.
Comparison
of
[ 3
'S]methi-
onine-connexin43
synthesized
at
20°C
(Fig
.
6,
lanes
1-3)
with
that
labeled
at
37°C
(Fig
.
1
A,
lanes
3-S)
revealed
that
connexin43
remained
in
the
Triton-soluble
connexin43-NP
state
at
the
lower
temperature
.
Reversal
of
the
20
°
C
block
by
incubation
of the
cells
for
an
additional
2
h
in
chase
medium
at
37
°
C
resulted
in
efficient
conversion
of
["S]me-
thionine-connexin43-NP
to
Triton-insoluble
[
3
5S]methio-
nine-connexin43-P,
and
-P
Z
(Fig
.
6,
lanes
4-6)
.
Thus,
both
136
6
on July 10, 2011jcb.rupress.orgDownloaded from
Published December 1, 1991
Figure
7
.
Cell-surface
biotinylation
of
connexin43
in
intact
NRK
cells
.
Confluent
NRK
cell
monolayers
labeled
with
[
35
S]methionine
at
37°C
for
5
h
were
subjected
to cell-surface biotinylation
at
4°C
.
The
cells
were
then
either
lysed
directly
into
SDS
(A)
or
processed
for
the
standard
Triton
solubility
assay
(1%
Triton
in
PBS,
30
min,
4°C)
(B)
.
(A)
Lane
1,
total
cellular
connexin43
immunoprecipitated
from
1/15
of a
biotinylated
60-mm
cell
culture
.
Lane
2
is
a
shorter
exposure
of
lane
1,
illustrating
the
lack
of
resolution
between
the
P,
and
Ps
forms
of
connexin43
after
the
biotinylation
procedure
.
Lane
3,
biotinylated
connexin43
recovered
from
the
remaining
14/15
of
the
60-
mm
cell
culture
by
sequential
precipitation
with
affinity-purified
anti-connexin43
(252-271)
antibodies
followed
by
avidin-agarose
.
Same
exposure
as
lane
1
.
Lanes
4-6,
same
as
lane
3,
except
connexin43
was
immunoprecipitated
in
the
presence
of
100
ug/ml
of
competing
connexin43
(252-271)
peptide
(lane
4)
;
NHS-LC-biotin
was
quenched
with
excess
glycine
prior
to
the
biotinylation
reaction
pane
6)
;
or
NHS-LC-biotin
was
omitted
duringthe
biotinylation
procedure
pane
5)
.
(B)
Cell-surface biotinylated
connexin43
recovered
from
equal
amounts
of
the
total
cellular
lysate
(lane
1),
Triton-soluble
supernatant
pane
2),
or
Triton-insoluble
pellet
pane
3)
fractions
.
acquisition
of
Triton
insolubility
and
phosphorylation
of
connexin43
to the P,
and
P
Z
forms
occurred
in a
cellular
compartment
distal
to the
medial
Golgi
.
A
similar
conclu-
sion
wasdrawn
from
experiments
in
which
intra-Golgi
trans-
port
was
inhibited
with
nontoxic
concentrations
of the
car-
boxylic
ionophore
monensin
(data not
shown)
.
In
this
case,
however,
the
block
in
connexin43
processing
was
less
com-
plete
than
at
20°C,
presumably
because
of
the
known
leak-
iness
of
monensirfs
effect
on
intracellular
transport
at
0
.5-1
.0
,uM
(Peters
et al
.,
1983)
.
The 20°C
block
experiments
raised the
possibility
that
phosphorylation
of
connexin43
and/or
acquisition of Triton
resistance
occurred
after
transport
of
connexin43
to the
cell
surface
.
This
possibility
was
investigated
using
the
technique
of
cell-surface biotinylation
(Le
Bivic
et
al.,
1989,
1990b)
to
selectively
label
and
monitor
connexin43
on
the
plasma
membrane
.
[
35
S]methionine-labeled
cell
monolayers
were
incubated
at
4°C
with
the
membrane-impermeant
protein
bi-
otinylating
reagent
NHS-LC-biotin,
which
in
intact
cells
reacts
covalently
with
primary
amine
groups
(mostly
lysine
residues)
located
in
extracellular
domains
of
plasma
mem-
brane
proteins
(Sargiacomo
et al
.,
1989)
.
After
30 min
the
reaction
was
quenched
with
glycine,
the
cells
were
lysed
in
SDS,
and
connexin43
was
immunoprecipitated by
our
stan-
dard
protocol
.
A
fraction
of
the
immunoprecipitated
connex-
in43
was
reserved
as
a
sample
of
total
cellular
connexin43,
and
the
remainder
was
subjected
to a
second
round
of
pre-
cipitation
with
avidin-agarose
to
selectively
recover
bio-
tinylated
connexin43
molecules
.
Musil and
Goodenough
Gap
Junction
Assembly
When
NRK
cells
were
metabolically
labeled
with
[
35
S]me-
thionine for 5
h
and
then
biotinylated
as
described
above,
two
distinct
[
35
S]methionine-connexin43
species
were
obtained
after
the
double
precipitation
procedure
and
analysis
on
SDS-PAGE
(Fig
.
7
A,
lane 3)
.
The
faster-migrating
form
had
the
same
electrophoretic
mobility
as
total
cellular
[3
sS]me-
thionine-connexin43-NP
(Fig
.
7
A,
lane
1)
and
was
soluble
in
1%
Triton
X-100
(Fig
.
7
B, lane 2)
.
The
other
biotin-la-
beled
band
comigrated with
phosphorylated
total
connex-
in43
and was
resistant
to
Triton
X-100
(Fig
.
7
B,
lane
3)
;
this
species
was
converted
to
connexin43-NP
by
alkaline
phos-
phatase
(data not
shown)
.
The
biotinylation
and
precipita-
tion
procedures
slightly
altered
the
migration
of
connexin43
on
SDS-PAGE
such
that
connexin43-P,
was
not
fully
re-
solved
from
connexin43-P
2
on
either
5
.5-
or 12
.5-cm
gels
.
We
will
therefore
refer
to
phosphorylated
forms
of
biotiny-
lated
connexin43
as
simply
connexin43-P
to
reflect
this
ambiguity
.
Several
control
experiments
confirmed
the
specificity
of
cell-surface biotinylation
of
connexin43
;
neither
[
35
S]methi-
onine-connexin43-NP
nor
[
35
S]methionine-connexin43-P
was
recovered
if
the
anti-connexin43
(252-271)
inununopre-
cipitation
step
was
conducted
in
the
presence
of
an
excess
of
the
peptide
against
which
the
antibody
was
raised (Fig
.
7
A,
lane
4),
or
if
the
biotinylation
reagent
was
either
omitted
(Fig
.
7,
lane
S) or
quenched
with
excess
glycine before incu-
bation
with
the
cell
monolayer
(Fig
.
7,
lane
6)
.
Similar
results
were
obtained
with
S180L
cells
(data
not
shown),
an-
other
communication-competent
cell
type
that
forms
large
136
7
on July 10, 2011jcb.rupress.orgDownloaded from
Published December 1, 1991
Figure
8
.
Pulse-chase
analysis of
transport
of
Connexin43
to
the
plasma
membrane
.
NRK
cell
monolayers
were
metabolically
la-
beled
with
[
35
S]methionine
for
30
min
and
chased
for
0
(lanes
1
and
2),
1
(lanes
3
and
4),
or
3
(lanes
S
and
6)
h
at
37°C
before
cell-surface
biotinylation
at
4°C
.
Lanes
marked
T,
total
cellular
Connexin43 immunoprecipitated
from
1/15
of
a
biotinylated
60-mm
cell
culture
.
Lanes
marked
B, biotinylated
Connexin43
purified
by
the
double
precipitation
procedure
from
the
remaining
14/15
of
the
corresponding
60-mm
cell
culture
.
gap
junctional
plaques
(Mege
et
al
.,
1988
;
Musil
et
AL,
19906)
.
The
recovery
of
cell-surface
biotinylated
Connexin43-NP
as
well
as
Connexin43-P
was
consistent
with
transport
of
Connexin43
to the
plasma
membrane
in
the
Connexin43-NP
form
.
This
was
confirmed
in a
pulse-chase,
experiment
in
which
NRK
cells
were
metabolically
labeled
with
["Slme-
thionine
for
30
min
and
then
chased
for
0-3
h
at
37°C
before
biotinylation
at
4°C
(Fig
.
8)
.
Virtually
no
[
35
S]methio-
nine-Connexin43
became
biotinylated
immediately
after
the
pulse
period
(Fig
.
8,
lane
2),
presumably
because
most
newly
synthesized
[
35
S]methionine-Connexin43
had
not
yet
reached
the
plasma
membrane
and was
inaccessible to
the
biotinylated
reagent
.
This
finding
indicates
that
biotinylation
was
confined
to the
cell
surface,
as
expected
.
In
contrast,
cultures
that
had
been
pulsed
for
30
min
and
then
chased
for
1
h
before
biotinylation
had
readily detectable
amounts
of
avidin-precipitable
[
35
S]methionine-Connexin43
(Fig
.
8,
lane 4)
.
It
is
significant
that
nearly
all
of
this
biotin-labeled
Connexin43
was
in
the
Connexin43-NP
form
and
was
Triton
soluble
(see
Fig
.
9
B,
lanes 1-3)
;
the
minor
amount
of
bio-
tinylated
[3
5
S]methionine-Connexin43-P
recovered
at
this
time
was
most
likely
Connexin43-P,
since
newly
synthe-
sized
Connexin43-NP
required
chase
periods
of
>1 h
in
order
for
detectable
amounts
to
be
converted
to the
P
2
form
(Fig
.
3
;
Musil
et al
.,
19906)
.
When
NRK
cells
were
biotinylated
after
a
3-h
chase,
[
35
S]methionine-Connexin43-NP
and
con-
nexin43-P
were
precipitated
in
approximately
equal
amounts
(Fig
.
8,
lane
6)
.
The
decrease
in
the
total
amount
of bio-
tinylated
[
35
S]methionine-Connexin43
recovered
at
this
time
relative
to
that
obtained
after
a
1-h
chase
was
most
likely
due
to a
combination
of
the
rapid
turnover
rate
of
Connexin43
(t,a,
N2-2
.5
h)
(Musil
et al
.,
19906)
and
the
decreased
efficiency
with
which
Connexin43
became
biotinylated
after
incorporation
into
gap
junctional
plaques
(see
Discussion)
.
Taken
together,
these
results
demonstrated
that
at
least
a
fraction
of
Connexin43
in
NRK
cells
arrived
on
the
plasma
The
Journal
of
Cell
Biology,
Volume
115,
1991
membrane
within
1-1
.5
h
of synthesis
and
in the
Triton-
soluble
Connexin43-NP form
.
Phosphorylation
of
Cell
Surface
Connexin43-NP
in
NRK
Cells
Cell-surface
biotinylation
has
been
shown
not
to
interfere
with
subsequent
intracellular
transport
and
processing
of
plasma
membrane
proteins
(Matter
et al
.,
1990
;
Le
Bivic
et
al
.,-
1990a,
b)
.
It
was-therefore
reasonable
to
study
the fate
of
cell
surface
Connexin43-NP
by
use
of
this
technique
(Fig
.
9)
.
NRK
cell
monolayers
were
metabolically
labeled
with
[
35
S]methionine
for
3
.0
min
and
chased
for
an
additional
hour
at
.
37°C
to
allow
transport
of
ne)Ni
y
synthesized
[
35
S]-
methionine-connexin43-io
.
.the
plasn
.
membrane
.
The
cul-
tures
were
then
biotinylated
at
4°C
After
which
the reaction
was
quenched
with
excess
glycine
aAd
the
cells
incubated
at
37
°
C
for
0-3 h
before
cell
lysiseand
purification
of
biotin-
labeled
Connexin43
.
Control
,e
riments
indicated
that
NRK
cultures
remained
well
coupled
for
at
least
3
h
after
bio-
tinylation
(not
shown)
.
In
the
absence
of
a
:ehasa
period
after
biotinylation,
most
(80-90%)
of
the,surfacs-labeled
[
35
S]methionine-connex-
in43
was
inthe
Connexin43-NP form
(Fig
.
9
A,
lane
1
;
Fig
.
8,
lane
4h
When
cells~were
chased
for
1
h
after
biotinylation,
the
profile-of'biotin-conjugated
["S]methionine-Connexin43
changed
botl
íqualitatively
and
quantitatively
(Fig
.
9
A,
lane
2)
.
.
The amount
of
biotinylated
["S]methionine-Connexin43
decreased
by
-40
%,
reflecting
the
rapid
turnover
rate
of
to-
tal
cellular
Connexin43
.
Of
the surface-labeled
[35
S]methio-
nine-Connexin43
remaining,
-50%
was
in
the
Connexin43-P
form
due
to
a
two-
to
threefold
increase
in
the
amount
of
biotinylated
[
35
S]methionine-Connexin43-P
between
0-1
h
of
chase
.
This
phosphorylated
Connexin43
migrated
slightly
slower
than
the
minor
amount
of
Connexin43-P
recovered
immediately
after
biotinylation,
a
finding
consistent
with
conversion
to the
P
2
form
(compare
Fig
.
9
A
lanes
1
and
2)
.
With
increasing
chase
time
after
biotinylation
(Fig
.
9
A,
lanes
3
and
4),
cell-surface-labeled
[
35
S]methionine-con-
nexin43-NP
became
undetectable
by
3
h
(Fig
.
9,
lane 4)
whereas
the
levels
of
biotinylated
[
35
S]methionine-connex-
in43-P
declined
more
slowly
at
a
rate
comparable
to
that
of
total
cellular
[
35
S]methionine-Connexin43-P2
(Musil
et al
.,
19906)
.
These
turnover
kinetics
suggested
that
the
loss
of
biotinylated
Connexin43-NP
observed
between
1-3
h
of
chase
(lanes
2-4)
was due
to
phosphorylation
to
Connexin43-P
However,
direct
degradation
of
at
least
a
fraction
of
this
bio-
tinylated
Connexin43-NP
without
conversion
to
Connexin43-P
cannot
be
definitively
ruled
out
.
Examination of
the
Triton
X-100
resistance
of
Connexin43
revealed
that
most
of
the
biotinylated
[
35
S]methionine-con-
nexin43
was
soluble
immediately
after
biotinylation
(Fig
.
9
B,
lanes
1-3)
.
This
observation
is
consistent
with
results
ob-
tained
with
total
cellular
[
35
S]methionine-Connexin43
after
a
similar
pulse-chase
protocol
(Fig
.
3,
lanes
4-6)
.
By
3 h
of
chase
after
biotinylation,
however,
surface-labeled
[
35
S]-
methionine-Connexin43
was
completely
insoluble in
Triton
(Fig
.
9 B,
lanes
4-6)
.
Taken
together,
these
observations
in-
dicate
that
cell-surface
biotinylated
Connexin43
is
converted
from
Triton-soluble
Connexin43-NP
to
a
phosphorylated
Connexin43-P
species
that
was
Triton-resistant
and
therefore
likely
to
be
part
of
a
gap
junctional
plaque
.
Like
total
cellular
[
35
S]methionine-Connexin43
(Musil
et al
.,
19906),
virtu-
136
8
on July 10, 2011jcb.rupress.orgDownloaded from
Published December 1, 1991
Figure 9
.
Processing
of
cell
surface
connexin43-NP
to
Triton
insoluble
connexin43-1?
NRK
cell
cultures
were
metabolically
labeled
for
30
min
with
[s
5
S]methionine
and
then incubated
in
the
presence
of an
excess
of
unlabeled
methionine
for
1
h
at
37°C
to
permit
transport
of
[s
5
S]methionine-connexin43
to
the
plasma
membrane
.
The
monolayers
were
then
cell-surface biotinylated
at
4°C,
after
which
the
cells
were
chased
at
37°C
for
0-3
h
to
follow the
fate
of
biotinylated
connexin43
.
(A)
At
the
end
of
the
specified
chase
period,
the
cells
were
lysed
directly
in
SDS,
and
biotinylated
connexin43
was
recovered
by
the
double
precipitation
procedure
.
(B)
Cells
were
subjected
to
the
standard
Triton
solubility
assay
(1%
Triton
in
PBS,
30
min, 4°C)
either
immediately
after
biotinylation
(lanes
1-3) or
after
a
3-h chase
(lanes
4-6)
.
Cell-surface
biotinylated
connexin43
was
recovered
from
equal
amounts
of
the
total
cellular
lysate
(lanes
marked
T),
Triton-
soluble
supernatant
(lanes
marked
S),
or
Triton-insoluble
pellet (lanes
marked
P)
fractions
.
ally
all
of
the biotinylated
connexin43
remaining
4 h
after
synthesis
had undergone
this
process
.
We
conclude
that
cell
surface
connexin43-NP
serves
as
a
precursor
to
connex-
in43-P,
suggesting
that
transport
of
connexin43
to the
plasma
membrane
normally
precedes
phosphorylation
to
either
the
P,
or
P
z
form
.
Tlransport
of
Connexin43
to
Plasma
Membrane
in
Communication-Deficient
Cell
Lines
To
determine
whether connexin43
was
also transported to the
surface
of
communication-deficient
cells,
confluent
mono-
layers
of
S180
or
L929
cells
were
metabolically
labeled
for
5 h
with
[35S]methionine
and
then
biotinylated
at
4°C
ex-
actly
as
described
for
NRK
cells
(Fig
.
10)
.
A
biotin-con-
jugated
species
corresponding
to
["S]methionine-connex-
in43-NP
was
recovered
from
both
cell
types,
the
amount
of
which
(relative
to
total cellular
PS]methionine-connexin43)
ranged
from0
.75
to
1
.5
times
as
much
as
in
NRK
cells
(Fig
.
10 A,
lanes
2 and
6)
.
Control
experiments
revealed
that
ra-
diolabeled
connexin43-NP
was
not
appreciably
biotinylated
immediately
after
a
30-min
pulse
with
['SS]methionine
but
became
accessible to the
biotinylation
reagent
after
a
1-h
chase
(Fig
.
10
B
;
compare
lane
2
with
lane
4
;
lane
6
with
lane
8)
.
Biotinylation
of
connexin43
thus
appeared
to
be
confined
to
the
plasma
membrane,
as
was
determined
for
NRK
cells
(Fig
.
8)
.
We
conclude
that
S180
and
L929
cells
transport
connexin43-NP
to the
cell
surface
but
are
deficient
in
the
ability
to
assemble
it
into
gap
junctions
.
Musil
and
Goodwough
Gap
Junction
Assembly
Effect
of
Junctional
Communication
Disruption
on
Thiton
Solubility
of
Connexin43
Our
studies
have
demonstrated
that
communication-compe-
tent
cell
types
convert
connexin43-NP
to the Triton-insolu-
ble,
plaque-associated
Pz
form
whereas
communication-
deficient
cells
do
not
.
To
investigate
the
dependence
of these
posttranslational
modifications
of
connexin43
on
ongoing
cell-cell
communication,
NRK
cells
were
rendered
com-
munication-deficient
by
treatment
with
the
uncoupling
reagent
heptanol
(Fig
.
11)
.
Heptanol
has
been
shown
to reversibly
block
transfer
of
dye
between
NRK
cells
(as
well
as
other
connexin43-containing
cell
types)
within
10
min
(Chanson
et al
.,
1989)
.
If
NRK
cultures
were
labeled
for
4 h
with
[s
5
S]methionine
in the
presence
of
3
.5
mM
heptanol,
none
of the
[
35
S]methionine-connexin43
synthesized
during
this
period
was
phosphorylated
to the
P
Z
form
(Fig
.
11
A,
lane
1)
.
This
finding
is
consistent
with
our
previous
results
(Musil
et al
.,
1990b)
.
This
[s
5
S]methionine-connexin43
was
solu-
ble in
1%
Triton
X-100
when
assayed
under
our
standard
conditions
(30
min, 4°C,
in
PBS)
(Fig
.
11
A,
lanes
1-3)
.
If,
however,
the
cells
were
labeled
with
[
35
S]methionine
for
4 h
in the
absence
of
heptanol
and
only
then
uncoupled
by
a
10
min
incubation
in
heptanol-containing
chase
medium,
the
[
35
S]methionine-connexin43-P
2
formed
during
the
pulse
period
remained
Triton
insoluble
and
fully
phosphorylated
(Fig
.
11
B,
lanes 4-6),
indistinguishable
from
untreated
con-
trols
(Fig
.
11
B,
lanes
1-3)
.
Immunohistochemical
studies
revealed
that
the
macular
anti-connexin43
(252-271)
stain-
136
9
on July 10, 2011jcb.rupress.orgDownloaded from
Published December 1, 1991
Figure
10
.
Expression
of connexin43
on
the
plasma
membrane
of
communication-deficient
5180
andL929
cells
.
(A)
Cultures
of S180
(lanes
1-4) or
L929
(lanes
S-7)
cells
were
metabolically
labeled
at
37°C
with
[
3
'S]methionine
for
5
h
before
cell-surface biotinylation at
4°C
.
The
cultures
were
lysed
in
SDS
and
total
cellular
connexin43
was
immunoprecipitated
;
a
fraction
of
this
was
then
precipitated
with
avidin-agarose
to
selectively
recover
biotinylated
connexin43
molecules
.
Lanes
1
and
S,
total
cellular
connexin43
immunoprecipitated
from
1/15
of
a
biotinylated
60-mm
cell
culture
.
Lanes
2 and
6,
biotinylated
connexin43
recovered
from
the
remaining
14/15
of
the
corre-
sponding
60-mm
cell
culture
.
Lanes
3
and
7,
same
as
lanes
2
and
6,
respectively,
except
that
NHS-LC-biotin
was
quenched
with
excess
glycine
prior
to
the
biotinylation
reaction
.
Lane
4
:
same
as
lane
2,
except
that
connexin43
immunoprecipitation
was
conducted
in
the
presence of
100,ug/ml
of
competing
connexin43
(252-271)
peptide
.
(B)
S180
(lanes
1-4)
or
L929
(lanes
S-8)
cells
were
pulsed
for
30
min
with
[
3
'S]methionine
and
then
chased
at
37°C
for
either
0h
(lanes
1,
2,
S,
and
6)
or
1
h
(lanes 3,
4,
7,
and
8)
.
Cell-surface
biotinyla-
tion
was
then
conducted
at
4°C
.
Lanes
marked
T,
total
cellular
connexin43 immunoprecipitated
from
1/15
of
a
biotinylated
60-mm
cell
culture
.
Lanes
marked
B,
biotinylated
connexin43
purified
by
the
double
precipitation
procedure
from
the
remaining
14/15
of
the
corre-
sponding
60-mm
cell
culture
.
The
Journal
of
Cell
Biology,
Volume
115,
1991
137
0
on July 10, 2011jcb.rupress.orgDownloaded from
Published December 1, 1991
Figure
11
.
Effect
of
heptanol
on
posttranslational
processing
of
Connexin43
in
NRK
cells
.
Confluent
NRK
cultures
were
metabolically
labeled
for
4 h with
[
33
S]methionine
in
either
the
presence
(A)
or
absence (B) of
uncoupling
levels
(3.5
mM)
of
heptanol
.
(B)
The
cultures
were
then chased
for
10
min
either
without
(lanes
1-
.1)
or with
(lanes
4-6)
3
.5
mM
heptanol
.
All
cultures
were
homogenized
and
subjected
to
the
standard
Triton
solubility
assay
(1%
Triton
in
PBS,
30
min,
4°C)
.
Connexin43 was
immunoprecipitated
from
equal
amounts
of
the
total
cellular
lysate
(lanes
marked
T),
Triton-soluble
supernatant
(lanes
marked
S),
or
Triton-insoluble
pellet
(lanes
marked
P)
frac-
tions
.
ing pattern
observed
at
cell-cell
interfaces
in
untreated
NRK
cells
was
retained
after
a
10-min
incubation
in
heptanol
(data
not
shown),
a finding consistent
with
the
lack
of
effect
of
brief
exposure
to
heptanol
on
the
ultrastructure
of
pancreatic
acinar
cell-gap junctions
(Meda
et
al
.,
1986
;
Bruzzone
et
al
.,
1987)
.
Similar
results
were
obtained
when
NRK
cells
la-
beled
for
4 h
with
['SS]methionine
were
rendered
communi-
cation-incompetent
by
a
different
mechanism,
i.e
.,
cytoplas-
mic
acidification
induced
by
brief (10-min)
exposure
to
100
C0
2
(Schuetze
and
Goodenough,
1982
;
data
not
shown)
.
Thus,
under
uncoupling
conditions, as
in
all
other
situations
we
have
tested,
Connexin43-NP
is
soluble
in
Triton
in
our
standard
assay
whereas
Connexin43-P
2
is
invariably
Triton-
insoluble
.
Discussion
A
combination
of electron
microscopy,
x-ray
diffraction,
and
other
physical
techniques
have demonstrated
that
gap
func-
tional
plaques
consist
of
arrays of
intercellular
channels,
each
of
which
is
composed
of
two
hexameric connexons
joined
head-to-head
to
form
a
transmembrane
pore
(Caspar
et
al
.,
1988)
.
We
have
studied
the
assembly
of
this
complex
structure
by
investigating
the
intracellular
transport
and
post-
translational
processing
of
Connexin43,
a
member
of
the
closely
related
family
of
integral
membrane
proteins
that
comprise
gap
junctions
(Beyer
et al
.,
1987
;
Stevenson
and
Paul,
1989)
.
We
have
shown
that
phosphorylation
of
connex-
Musil
and
Goodenough
Gap
Junction
Assembly
ín43 to the
mature
P
2
form
occurs
after
arrival
of
newly
synthesized
Connexin43-NP
on
the
cell
surface
and
is
accom-
panied by accumulation of
Connexin43
in Triton
X-100-
insoluble
gap
junctional
plaques
.
Connexin43
is
also
trans-
ported
to the
plasma
membrane
of
communication-deficient
cells
but
remains
Triton soluble
and
is
not
appreciably
pro-
cessed
to the
Connexin43-P
2
form
.
Taken
together,
these
re-
sults
are
consistent
with
a
model
in
which
Connexin43
is
con-
stitutively
transported
to the
plasma
membrane
regardless
of
the
ability
of the
cell
to
form
gap
junctions
.
Subsequent
phosphorylation
of
Connexin43
to
the
P
2
form
occurs
only
in
communication-competent
cells
and
is
temporally
associ-
ated
with,
and
may
be
functionally
involved
in,
assembly of
Connexin43
into
morphologically
and
physiologically
recog-
nizable
gap
junctional
plaques
.
Detergent
Solubility
of
Phosphorylated
and
Nonphosphorylated
Forms
of
Connexin43
A
major
finding
of
this
study
is
that
newly
synthesized
Connexin43
is
completely
soluble in
the
nonionic
detergent
Triton
X-100
but
acquires
Triton
resistance
concomitant
with
phosphorylation
to the
Connexin43-P
2
form
.
In
situ
extraction of
communication-competent
NRK
cultures
with
1%
Triton,
combined
with
immunofluorescent
localization
of
Connexin43,
revealed
that
Triton-insoluble
connexin43-
P
2
was
concentrated
in
gap
junctional
plaques
whereas
non-
phosphorylated
(Triton-soluble)
Connexin43-NP
was
pre-
dominantly
intracellular
.
Acquisition
of
Triton
resistance
137
1
on July 10, 2011jcb.rupress.orgDownloaded from
Published December 1, 1991
thus
appears
to
be
a
useful
biochemical
marker
for
accumu-
lation
of
connexin43
in
gap
junctional plaques,
although
we
cannot
rule
out
the
possibility
that
a small
amount
of
Triton-
insoluble
connexin43
(undetectable
by
immunofluorescence)
is
present
in
nonjunctional
membranes
as well
.
The
physical basis
for
the
insolubility
of
connexin43-P
2
in
Triton
is
unknown
.
Resistance
to
Triton
X-100
is
not,
however,
an
inevitable
consequence
of
tight
packing
of mul-
tisubunit integral
membrane
proteins since the
nicotinic
ace-
tylcholine
receptor
remains
Triton
soluble
even
after
assem-
bly
into
extremely
high-density
plasma
membrane
clusters
(Miledi
et al
.,
1971)
.
For
some
proteins,
posttranslational
acquisition
of
Triton
insolubility
has
been
shown
to
reflect
association
with
the
cytoskeleton
(Nelson,
1989)
;
however,
there
is
no
evidence
(at
least
in
liver
and
lens
fibers)
that
gap
junctional
plaques
interact
with
any
filamentous
submem-
branous
systems
(Hirokawa
and
Heuser,
1982
;
Peracchia
and
Peracchia,
1980)
.
Insolubility
of
connexin43-P
2
in Tri-
ton
is
also
unlikely
to
be
due
to
formation
of
intermolecular
disulfide
bonds
since
reducing
agents
fail
to
convert
connex-
in43-P
2
to a Triton-soluble
state
(Table
I)
.
One
potential
explanation
for
the
differential
solubility
of
phosphorylated
and
nonphosphorylated
forms
of
connexin43
is
that
assem-
bly
of connexin43
into
gap
junctions
results
in a
conforma-
tional
change
in
the
connexin43
molecule
that
renders
the
plaque
insoluble in Triton
.
Alternatively,
connexin43
could
become
posttranslationally
associated
with
Triton-resistant
plasma
membrane
lipids
(Yu
et al
.,
1973)
.
Evaluation
of
this
possibility
awaits
definitive
determination
of the
lipid
com-
position
of
gap
junctions
(Malewicz
et
al
.,
1990)
.
The
classic
method
used
to
isolate
gap
junctions
from
ro-
dent
liver
(composed
of
connexin32
and
connexin26)
as
well
as
from
heart
(containing
connexin43)
exploits
the
relative
resistance
of
these
structures
to
solubilization
in
the ionic de-
tergent
N-lauryl sarcosine
(Goodenough
and
Stoeckenius,
1972
;
Hertzberg
and
Gilula,
1979
;
Kensler
and
Goodenough,
1980)
.
The
observation
here
that
connexin43-P2
from
NRK
cells
is
virtually
quantitatively
solubilized
under
conditions
(0
.3
%
sarcosine,
5
mM
Tris,
pH
10
;
10
min, 25
°C)
that
yield
sarcosine-insoluble
gap
junctions
from
rat
heart
(Kensler
and
Goodenough,
1980)
thus
requires
explanation
.
Differen-
tial
organization
of connexin43
in
tissue-culture
cells
com-
pared
to
whole
organs
cannot
be
the
answer,
since
connex-
in43
in
embryonic
chick
lens
is
also soluble
in
sarcosine
in
our
assay
(Fig
.
2)
.
Three
explanations
seem
plausible
.
First,
the
gap
junctions
recovered
from
rat
heart
after
sar-
cosine
treatment
may
represent a
very
small
(and
perhaps
specialized)
fraction
ofthe
total
gap
junction
population
.
An
early
estimate
of the yield
of
gap
junctions
obtained
from
mouse
liver
using
a sarcosine insolubility-based
isolation
procedure
was
-10%
(Goodenough
and
Stoeckenius,
1972)
.
The
amount
of
protein
recovered
in the
final
gap
junction
fraction
has since
been
shown
to
be
-100-fold
less
(D
.
Good-
enough,
unpublished
data),
decreasing
this
figure
to
-0
.1%
.
Although
the
reason
for
this
low
yield
of
gap
junctions
has
not
been
systematically
studied,
the
observation
of
Hertz-
berg
and
Gilula (1979)
that
rat
liver
gap
junctional
plaques
are
in
fact
partially
soluble in
sarcosine
suggests
that
sig-
nificant loss
of
junctional
material
may
occur
during
sarco-
sine
treatment
.
Consistent
with
this
interpretation,
the
re-
covery
of
gap
junctions
from
rat
liver
increases
10-fold
when
a
detergent-free, rather
than
a
sarcosine-based,
isolation
pro-
The
Journal
of Cell
Biology,
Volume
115,
1991
cedure
is
used
(Hertzberg,
1984)
.
Given
that
cardiac
gap
junctions
are
less
resistant
to
sarcosine
than
those
from
liver
(Kensler
and
Goodenough,
1980),
it
is
therefore
likely
that
the
yield
of
sarcosine-insoluble
gap
junctions
from
heart
is
very
low
.
Second,
published
procedures
for
the
isolation
of
gap
junctions
from
rodent
heart
and
liver
(see
above)
involve
sev-
eral
fractionation
steps
before
sarcosine
treatment
.
It is
pos-
sible
that
during
these
manipulations
some
gap
junctions
are
artifactually
converted
to
a
sarcosine-resistant
form
in a
manner
that is
not
reproduced
in
our
solubilization
assay
utilizing
freshly
prepared
whole-cell
lysates
.
A
third
alternative
that
we
cannot
rule out
is
that
connex-
in43
in heart
intercalated
disks
is
genuinely
less
soluble in
sarcosine
that
connexin43
in
either
NRK
cells
or
embryonic
chick
lens
epithelium,
perhaps
reflecting
the
presence
of
as
yet unidentified
connexins
or other
structural
proteins
in
heart
gap
junctions
that
confer
stability
to
harsh
detergents
.
Direct
quantitation
of
the
sarcosine
solubility
of
gap
junc-
tions
in
cardiac
and
hepatic
tissue
will
be
required
to
resolve
which
of these three
possibilities
(if
any)
is
correct
.
Transport
of
Connexin43
to
the
Plasma
Membrane
We
developed
a
biochemical
assay for
transport
of
connex-
in43
to
the
plasma
membrane
based
on
the
technique
of
cell-
surface
biotinylation
(Le
Bivic
et
al
.,
1989
and
1990b)
.
Using
this
assay,
we
determined
that
connexin43
is
inserted
into
the
plasma
membrane
of
communication-competent
NRK
cells
prior to
conversion
to
either
the
connexin43-P,
or
the
connexin43-P2
form
.
This
cell-surface
connexin43-
NP
is
soluble
in
Triton,
suggesting
that
transport to
the
plasma
membrane
is
not
coincident
with
assembly of
con-
nexin43
into
junctional
plaques
.
This
conclusion
is
further
supported
by
our
finding
that
cell
lines
with very
few
(S1ß0
cells)
or
no
(L929
cells)
morphologically
or physiologically
detectable
gap
junctions
nevertheless transport
connexin43
to the
plasma
membrane
.
Taken
together,
these
studies
pro-
vide
the
first
biochemical
proof
for
"free"
(extrajunctional)
channel
precursors
in the
plasma
membrane,
the existence
of
which
has
long
been
hypothesized
(Loewenstein,
1981
;
Rook
et
al
.,
1990)
.
In
communication-competent
cells,
ex-
trajunctional
cell-surface
connexin43-NP
is
efficiently
con-
verted
into
Triton-insoluble
and
therefore
plaque-associated
connexin43-P
2
.
Our
data
thus
support
a
model
of
gap
junc-
tion
formation
in
which
channel
precursors
are inserted into
nonspecialized
regions
of the
plasma
membrane
and
subse-
quently
accumulate
at
sites
of
cell-cell
contact
by
lateral
migration
in
the
plane
of the
membrane
bilayer
(Loewen-
stein,
1981)
.
The
absence
of
cell-cell
adhesion
molecules
in
communication-deficient
S180
and
L929
cells
(Mege
et al
.,
1988
;
Nagafuchi
et al
.,
1987)
rules
out
an
obligatory
role
for
CAM-mediated
intercellular
association
in
the transport
of connexin43
to
the
plasma
membrane
.
Although
highly
selective
for
connexin43
on
the
plasma
membrane,
cell-surface
biotinylation
is
not a
quantitative
technique
.
In the four
cell
lines
we
examined
(NRK,
S180L,
S1ß0,
and
L929),
only
-1%
of
the
total
[ 3
IS]methionine-
connexin43
labeled
during
a
5-h
pulse
with
[
35
S]methionine
was
recovered
with
avidin-agarose
after
cell-surface
bio-
tinylation
at
4°C
(Figs
.
7
and
10)
.
This
low
percentage
is
highly reproducible
(range,
0
.7-1
.3
%
;
n
=
8
in
NRK
cells)
and
is
not
due
to
inefficiency
of
either
of the
precipitation
137
2
on July 10, 2011jcb.rupress.orgDownloaded from
Published December 1, 1991
steps,
incomplete
elution
of
connexin43
from
the
avidin-
agarose
beads,
or
to
obvious
degradation
of
connexin43
(data
not
shown)
.
Two
interrelated
factors
probably
account
for
the limited
yield of
biotinylated
connexin43
.
First
is
the
low
efficiency
of
the
biotinylation
reaction
itself,
which
has
been
estimated
to
vary
from
N9-50
%,
depending
on
the
par-
ticular
protein
examined
(Le
Bivic
et al
.,
1989
;
Matter
et al
.,
1990)
.
In
support
of
this
possibility,
we
have
found
that
rais-
ing
the
temperature
at
which
NRK
cells
were
biotinylated
from
4°C
to
37°C
increased
the
amounts
(relative
to
total
cellular
[
3
IS]methionine-connexin43)
of
avidin-precipitable
["S]methionine-connexin43-NP
threefold
and
of
[
11
S]me-
thionine-connexin43-P
tenfold,
despite
the
fact
that
intracel-
lular
connexin43 remained
inaccessible
to the
biotinylation
reagent
at
the
higher
temperature
(not
shown)
.
Thus,
an
ab-
solute
maximum
of
one
third
of
the
[
3
IS]methionine-con-
nexin43-NP
and
one
tenth
of
the
[
3
IS]methionine-connex-
in43-P
actually
present
on
the
plasma
membrane
becomes
biotinylated
under
standard
labeling
conditions
(4°C)
.
Should
the
efficiency
of
biotinylation
of
connexin43
at
37"C
be
considerably
less
than
100%,
this
fraction
would
drop
ac-
cordingly
.
The
second
probable
cause
for
the
low
yield of
biotiny-
lated
connexin43
in
communication-competent
cells
is
inac-
cessibility
of
plaque-associated
connexin43
to the
surface-
labeling
reagent
.
Comparison
of
the
phosphorylation
state
of
total
[
31
S]methionine-labeled
connexin43
in
NRK
cells
with
that
of
cell-surface
biotinylated
connexin43
reveals
that
pro-
portionately
less
connexin43-P
than
connexin43-NP
be-
comes
biotinylated
at
any
given
time
.
Although
readily
de-
tectable
under
steady-state
labeling
conditions
(Fig
.
7,
lane 1
vs
.
lane
3),
this
phenomenon
is
most
strikingly
illustrated
in
a
pulse-chase
experiment
(Fig
.
8)
in
which
phosphorylated
connexin43
represents
80%
of
the
total
cellular
[
3
IS]methi-
onine-connexin43
recovered
after
a
3-h
chase
(Fig
.
8,
lane
S)
but
only
45
%
of the
connexin43
biotinylated
at this
time
(Fig
.
8,
lane 6)
.
Control
experiments
indicated
that
this
under-representation
of
connexin43-P
after
biotinylation
was
not
due
to
artifactual
dephosphorylation
during
the
dou-
ble
precipitation
procedure
(see
Materials
and
Methods)
.
Since
phosphorylated
forms
of
connexin43
are Triton
insolu-
ble (Fig
.
1
A),
we
believe
that
the
most
likely
reason
why
connexin43-P
is
inefficiently
labeled
by
cell-surface
biotiny-
lation
is
that
it is
largely
incorporated
into
gap
junctional
plaques
.
Plaque-associated
connexin43
may
be
a
poor
sub-
strate
for
the
biotinylation
reagent
because
of
restricted
dif-
fusion
of
NHS-LC-biotin
(mol
mass
=
556
D
;
2
.24
nm)
in
the
2-nm
intrajunctional
"gap" or to
changes
in the
tertiary
or
quaternary
structure
of
connexin43
that
render
the pro-
tein's
reactive
lysine
residues
inaccessible
after
plaque
forma-
tion
.
Reduced
biotinylation
of
junctional
connexin43
would
also explain
why
approximately
the
same
fraction
(-I%)
of
total
cellular
[
33
S]methionine-connexin43
becomes
biotiny-
lated
in
NRK
cells
(with
abundant
large
plaques)
as
in
S180
and
L929
cells
that
lack
detectable
gap
junctions
.
In
light
of
these
quantitative limitations
of
the
cell-surface
biotinylation
assay,
we
cannot
calculate
the
t
i/2
of
transport
of
connexin43
to
the
plasma
membrane
or
directly
deter-
mine
the
fraction
of
total
connexin43
present
on
the
cell
sur-
face
at
a
given
time
.
Our
demonstration
that
phosphorylation
of
connexin43
to the
P
2
form
occurs
after
transport to the
plasma
membrane
and
that
this
species
accumulates
in cell
Musil
and
Goodenough
Gap Junction
Assembly
surface
plaques
.
does,
however,
allow
us to
conclude
that
most
(if
not
all)
connexin43-P2
is
localized to the
plasma
membrane
.
Conversely,
most
intracellular
connexin43
ap-
pears
to
be
nonphosphorylated, although
the
presence
of
in-
ternalized
connexin43-P2
destined for
degradation
cannot
be
ruled
out
.
Role
of
Connexin43
Phosphorylation
in
Gap
Junction
Formation
and
Function
On
the
basis
of
our
current
and
previous data
(Musil
et al
.,
1990a
and
b),
we
can
eliminate
certain
potential
functions
for
phosphorylation
of connexin43
to the P, or
P
2
form
.
First,
the
presence
of
connexin43-NP
on
the
plasma
mem-
brane
of
both
communication-competent and
-deficient
cells
rules
out
an
obligatory
role
for
processing
to
either
the
P,
or
P
2
form
in
the transport of
connexin43
to the
plasma
membrane
.
Our
demonstration
that
cell-surface
connexin43-
NP
is
converted
to
Triton-insoluble
connexin43-P
in
NRK
cells
suggests
that
transport
of
connexin43-NP
to the
plasma
membrane
is
part
of
the
normal
pathway
leading
to
gap
junc-
tion
formation
in
communication-competent
cells
.
Since
connexin43-P,
and
-P
2
are the
only
forms
of connexin43
to
become
appreciably
labeled
With
32
P
in
the
cell
types
we
have
examined,
these
results
are consistent
with
arrival
of
connexin43
at
the
cell
surface
in
a
nonphosphorylated
state
.
The
reported
presence
of an
additional,
minor
phosphory-
lated
form
of
connexin43
that
appears
to
be
a
kinetic
inter-
mediate
between
connexin43-NP
and
-P,
in vole
fibroblasts
(Crow
et al
.,
1990)
does,
however,
raise
the
possibility
that
in
at
least
some
cell
types
initial
phosphorylation
of
connex-
in43
may
begin
before
transport
to
the
plasma
membrane
and
conversion
to
the
P,
and
P
2
forms
.
Second,
phosphory-
lation
does
not
appreciably
affect
the
metabolic
stability
of
connexin43
:
the
tv2
of
degradation
of
connexin43
is
equiva-
lent in
cells
that
process
connexin43
to
both
the P,
and
P
2
forms
(NRK
cells)
and
those
that
do
not
(L929
cells)
(Musil
et
al
.,
1990b)
.
Lastly,
phosphorylation
of connexin43
is
not
exclusively
associated
with
actively
communicating
cells
(Fig
.
11)
.
The
presence
of
Triton-resistant
connexin43-P
2
in
uncoupled
as
well
as
actively
communicating
NRK
cells
sug-
gests
that
reversible
conversion
of
connexin43
to
and
from
the
Triton-insoluble
P
2
form
is
not the
mechanism
of
inter-
cellular
channel
opening
and
closure
.
Under
all
conditions
examined,
terminal
phosphorylation
of
connexin43
was
tightly
linked
to
acquisition
of
Triton
in-
solubility
and
therefore
to
accumulation
of
connexin43
into
gap
junctional
plaques
.
Phosphorylation
to the
connexin43-
P
2
form
could
thus
potentially
be
involved
in
establishment
and/or
maintenance
of
gap
junctional
plaques
.
In
this
case,
determination
of
the
role
ofconnexin43
phosphorylation
will
most
likely
require
detailed analysis
of
the
assembly
state
of
phosphorylation-deficient
connexin43
mutants
.
We
thank
Dr
.
Karl Matlin
for
advice
and
use
of
his
20°C
bath
and
Anelise
Horah
for
photographic
assistance
.
This
work
was
supported
by
National
Institutes
of Health
grants
GM-
18974
and
EY-02430
.
Received
for
publication
22
July
1991
and
in
revised
form
22
August
1991
.
References
Beyer,
E
.
C
.,
D
.
L
.
Paul,
and
D
.
A
.
Goodenough
.
1987
.
Connexin43
:
a protein
137
3
on July 10, 2011jcb.rupress.orgDownloaded from
Published December 1, 1991
from
rat
heart
homologous
to
a
gap
junction
protein
from
liver
.
J
.
Cell Biol
.
105
:2621-2629
.
Beyer,
E
.
C
.,
J
.
Kistler,
D
.
L
.
Paul,
and
D
.
A
.
Goodenough
.
1989
.
Antiser
a
directed against
connexin43
peptides react
with
a
43-kD
protein
localized
to
gap
junctions
in
myocardium
and
other
tissues
.
J
.
Cell
Biol
.
108
:595-605
.
Beyer,
E
.
C
.,
D
.
L
.
Paul,
and
D
.
A
.
Goodenough
.
1990
.
Connexi
n
family
of
gap
junction
proteins
.
J
.
Membrane
Biol
.
116
:187-194
.
Bruzzone,
R
.,
E
.
R
.
Trimble,
A
.
Gjinovci,
O
.
Traub,
K
.
Willecke,
and
P
.
Meda
.
1987
.
Regulation
of
pancreatic
exocrine
function
:
a role for
cell-to-
cell
communication? Pancreas
.
2
:262-271
.
Caspar,
D
.
L
.
D
.,
G
.
E
.
Sosinsky,
T
.
T
.
Tibbins,
W
.
C
.
Phillips,
and
D
.
A
.
Goodenough
.
1988
.
Ga
p
junction
structure
.
In
Gap
Junctions
.
E
.
L
.
Hertz-
berg
and
R
.
G
.
Johnson,
editors
.
Alan
R
.
Liss,
Inc
.,
New
York
.
117-133
.
Chanson,
M
.,
R
.
Bmzzone,
D
.
Bosco,
and
P
.
Meda
.
1989
.
Effects
of
n-alcohols
on
junctional
coupling
and
amylase
secretion of
pancreatic
acinar
cells
.
J
.
Cell
.
Physiol
.
139
:147-156
.
Crow,
D
.
S
.,
E
.
C
.
Beyer,
D
.
L
.
Paul, S
.
S
.
Kobe,
and
A
.
F
.
Lau
.
1990
.
Phos-
phorylatio
n
of
connexin43
gap
junction
protein
in
uninfected
and
Rous
sar-
coma
virus-transformed
mammalian
fibroblasts
.
Mol
.
Cell
.
Biol
.
10
:1754-
1763
.
De
Mello,
W
.
1987
.
Modulation
of
junctional
permeability
.
In
Intercellular
Communication
.
W
.
C
.
De
Mello,
editor
.
Plenum
Publishing
Corp
.,
New
York
.
29-64
.
El-Fouly,
M
.
H
.,
J
.
E
.
Trosko,
and
C
.
C
.
Chang
.
1987
.
Scrape-loadin
g
and
dye
transfer
.
A
rapid
and
simple
technique
to
study
gap
junctional
intercellu-
lar
communication
.
Exp
.
Cell
Res
.
168
:422-430
.
Filson,
A
.
J
.,
R
.
Azarnia,
E
.
C
.
Beyer,
W
.
R
.
Loewenstein,
and
J
.
S
.
Brugge
.
1990
.
Tyrosin
e
phosphorylation
of
a
gap
junction
protein correlates
with
in-
hibition
of
cell-to-cell
communication
.
Cell
Growth
and
Differentiation
.
1
:661-668
.
Furshpan,
E
.
J
.,
and
D
.
D
.
Potter
.
1968
.
Low-resistanc
e
junction
between
cells
in
embryos
and
tissue
culture
.
In
Current
Topics
in
Developmental
Biology
.
Vol
.
3
.
A
.
A
.
Moscona
and
A
.
Monroy,
editors
.
Academic
Press,
New
York
.
95-127
.
Gilula,
N
.
B
.,
O
.
R
.
Reeves,
and
A
.
Steinbach
.
1972
.
Metabolic
coupling,
ionic
coupling,
and
cell
contacts
.
Nature
(Lond
.)
.
235
:262-265
.
Goodenough,
D
.
A
.,
and
W
.
Stoeckenius
.
1972
.
The
isolation
of
mouse
hepato-
cyte
gap
junctions
.
Preliminary
chemical
characterization
and
x-ray
diffrac-
tion
.
J
.
Cell
Biol
.
54:646-656
.
Guthrie
.
S
.
C
.,
and
N
.
B
.
Gilula
.
1989
.
Ga
p
junction
communication
and
devel-
opment
.
Trends
Neurosci
.
12
:12-16
.
Hertzberg,
E
.
1984
.
A
detergent-independent
procedure
for
the
isolation
of
gap
junctions
from
rat
liver
.
J
.
Biol
.
Chem
.
259
:9936-9943
.
Hertzberg,
E
.
L
.,
and
N
.
B
.
Gilula
.
1979
.
Isolatio
n
and
characterization
of
gap
junctions
from
rat
liver
.
J
.
Biol
.
Chem
.
254
:2138-2147
.
Hirokawa,
N
.,
and
J
.
Heuser
.
1982
.
The
inside
and
outside
of
gap
junction
membranes
visualized
by
deep-etching
.
Cell
.
30:395-406
.
Kensler,
R
.
W
.,
and
D
.
A
.
Goodenough
.
1980
.
Isolatio
n of
mouse
myocardial
gap
junctions
.
J
.
Cell Biol
.
86:755-764
.
Kistler,
J
.,
and
S
.
Bullivant
.
1988
.
Dissociation
of
lens
fiber
gap
junctions
releases
MP70
.
J
.
Cell
Sci
.
91:415-421
.
Laemmli,
U
.
K
.
1970
.
Cleavage of
structural
proteins
during
the
assembly
of
the
head
of
bacteriophage
T4
.
Nature
(Lond
.)
.
227
:680-688
.
Laird,
D
.
W
.,
K
.
L
.
Puranam,
and]
.
P
.
Revel
.
1991
.
Turnover
and
phosphory-
lation
dynamics
ofconnexin43
gap
junction
protein in
cultured cardiac
myo-
cytes
.
Biochem
.
J
.
273
:67-72
.
Larson,
D
.
M
.,
C
.
C
.
Haudenschild,
and
E
.
C
.
Beyer
.
1990
.
Gap
junction
mes-
senger
RNA
expression
by
vascular wall
cells
.
Circ
.
Res
.
66
:1074-1080
.
Le
Bivic,
A
.,
F
.
X
.
Real,
and
E
.
Rodriguez-Boulan
.
1989
.
Vectorial
targeting
of
apical
and
basolateral
plasma
membrane
proteins
in
a
human
adenocarci-
noma
epithelial
cell line
.
Proc
.
Nati
.
Acad
.
Sci
.
(USA)
.
86:9313-9317
.
Le
Bivic,
A
.,
A
.
Quaroni,
B
.
Nichols,
and
E
.
Rodriguez-Boulan
.
1990a
.
Bio-
genetic
pathways of
plasma
membrane
proteins in
Caco-2,
a
human
intestinal
epithelial
cell
line
.
J
.
Cell
Biol
.
111
:1351-1362
.
Le
Bivic,
A
.,
Y
.
Sambuy,
K
.
Mostov,
and
E
.
Rodriguez-Boulan
.
19906
.
Vec-
torial
targeting
of
an
endogenous
apical
membrane
sialoglycoprotein
and
uvomorulin
in
MDCK
cells
.
J
.
Cell
Biol
.
110
:1533-1539
.
Loewenstein,
W
.
R
.
1979
.
Junctional
intercellular
communication
and
the
con-
trol
of growth
.
Biochim
.
Biophys
.
Acia
.
560
:1-65
.
Loewenstein,
W
.
R
.
1981
.
Junctional
intercellular
communication
:
the
cell-to-
cell
membrane
channel
.
Physiol
.
Rev
.
61
:829-913
.
Loewenstein,
W
.
R
.
1985
.
Regulation
of
cell-cell
communication
by
phos-
phorylation
.
Biochem
.
Soc
.
Symp
.
50
:43-58
.
Malewicz,
B
.,
V
.
V
.
Kumar,
R
.
G
.
Johnson,
and
W
.
J
.
Baumann
.
1990
.
Lipid
s
in
gap
junction
assembly
and
function
.
Lipids
.
25:419-427
.
Manjunath,
C
.
K
.,
and
E
.
Page
.
1986
.
Rat
heart
junctions
as
disulfide-bonded
connexon
multimers
:
their
depolymerization
and
solubilization
in
deoxycho-
late
.
J
.
Membrane
Biol
.
90
:43-58
.
Manjunath,
C
.
K
.,
G
.
E
.
Goings,
and
E
.
Page
.
1984
.
Cytoplasmic
surface
and
The
Journal
of
Cell
Biology,
Volume
115, 1991
intramembrane
components
of
rat
heart
gap
junctional
proteins
.
Am
.
J
.
Phys-
iol
.
246
:H865-H875
.
Madin,
K
.
S
.,
and
K
.
Simons
.
1983
.
Reduced
temperature
prevents
transfer
of
a
membrane
glycoprotein
to
the
cell
surface but
does
not
prevent
terminal
glycosylation
.
Cell
.
34
:233-243
.
Matter,
K
.,
M
.
Brauchbar,
K
.
Bucher,
and
H
.-P
.
Hauri
.
1990
.
Sortin
g
of
en-
dogenous plasma
membrane
proteins
occurs
from
two
sites
in cultured
hu-
man
intestinal epithelial
cells
(Caco-2)
.
Cell
.
60
:429-437
.
Meda,
P
.,
R
.
Bruzzone,
S
.
Knodel,
and
L
.
Orci
.
1986
.
Blockage of
cell-to-cell
communication
within
pancreatic
acini
is
associated with
increased
basal re-
lease of
amylase
.
J
.
Cell
Biol
.
103
:475-483
.
Mege,
R
.
M
.,
F
.
Matsuzaki,
W
.
J
.
Gallin, J
.
I
.
Goldberg,
B
.
A
.
Cunningham,
and
G
.
M
.
Edelman
.
1988
.
Constructio
n
of
epithelioid
sheets
by
transfection
of
mouse
sarcoma
cells
with
cDNAs
for
chicken
cell
adhesion
molecules
.
Proc
.
Nad
.
Acad
.
Sci
.
(USA)
.
85:7274-7278
.
Mehta,
P
.
P
.,
J
.
S
.
Bertram,
and
W
.
R
.
Loewenstein
.
1986
.
Growt
h
inhibition
of
transformed
cells
correlates
with
their
junctional
communication
with nor-
mal
cells
.
Cell
.
44:187-196
.
Miledi,
R
.,
P
.
Molinoff,
and
L
.
Potter
.
1971
.
Isolation
ofthe
cholinergic
recep-
tor
protein
of
Torpedo
electric
tissue
.
Nature
(Load
.)
.
229
:554-556
.
Musil,
L
.
S
.,
and
D
.
A
.
Goodenough
.
1990
.
Ga
p junctional
intercellular
com-
munication
and
the
regulation
of
Connexin
expression
and
function
.
Curr
.
Op
.
Cell
Biol
.
2
:875-880
.
Musil,
L
.
S
.,
E
.
C
.
Beyer,
and
D
.
A
.
Goodenough
.
1990a
.
Expressio
n
of
the
gap
junction
protein
connexin43
in
embryonic
chick
lens
:
molecular
cloning,
ultrastructural
localization,
and
post-translational
phosphorylation
.
J
.
Mem-
brane
Biol
.
116
:163-175
.
Musil,
L
.
S
.,
B
.
A
.
Cunningham,
G
.
M
.
Edelman,
and
D
.
A
.
Goodenough
.
19906
.
Differentia
l
phosphorylation
of the
gap
junction
protein
connexin43
injunctional
communication-competent
and
-deficient
cell
lines
.
J
.
Cell Biol
.
111
:2077-2088
.
Nagafuchi,
A
.,
Y
.
Shirayoshi,
K
.
Okazaki,
K
.
Yasuda,
and
M
.
Takeichi
.
1987
.
Transformation
of
cell
adhesion
properties
by
exogenously
introduced
E-cadherin
cDNA
.
Nature
(Zond
.)
.
329
:341-343
.
Nelson,
W
.
J
.
1989
.
Topogenesis
of plasma
membrane
domains
in
polarized
epithelial
cells
.
Curr
.
Op
.
Cell Biol
.
1
:660-668
.
Peracchia,
C
.,
and
L
.
L
.
Peracchia
.
1980
.
Gap
junction
dynamics
:
reversible
effects
of
divalent cations
.
J
.
Cell
Biol
.
87:708-718
.
Peters,
B
.
P
.,
M
.
Brooks,
R
.
J
.
Hartle,
R
.
F
.
Krzesicki,
F
.
Perini,
and
R
.
W
.
Ruddon
.
1983
.
The
use
of
drugs
to dissect the
pathway
for secretion
of
the
glycoprotein
hormone
chorionic
gonadotropin
by
cultured
human
tropho-
blastic
cells
.
J
.
Biol
.
Chem
.
258
:14505-14515
.
Rook,
M
.
B
.,
B
.
de
Jonge,
H
.
J
.
Jongsma,
and
M
.
A
.
Masson-Pevet
.
1990
.
Ga
p
junction
formation
and
functional
interaction
between
neonatal
rat
car-
diocytes
in culture
:
a
correlative
physiological
and
ultrastructural
study
.
J
.
Membrane
Biol
.
118
:179-192
.
Saez,
J
.
C
.,
D
.
C
.
Spray,
A
.
C
.
Nairn,
E
.
Hertzberg,
P
.
Greengard,
and
M
.
V
.
L
.
Bennett
.
1986
.
cAM
P
increases
junctional
conductance
and
stimu-
lates
phosphorylation
of
the
27-kDa
principal
gap
junction
polypeptide
.
Proc
.
Natl
.
Acad
.
Sci
.
(USA)
.
83
:2473-2477
.
Sargiacomo,
M
.,
M
.
Lisanti,
L
.
Graeve,
A
.
Le
Bivic,
and
E
.
Rodriguez-
Boulan
.
1989
.
Integral
and
peripheral
protein
composition
of the apical
and
basolateral
membrane
domains
in
MDCK
cells
.
J
.
Membrane
Biol
.
107
:277-286
.
Schuetze,
S
.
M
.,
and
D
.
A
.
Goodenough
.
1982
.
Dy
e
transfer
between
cells
of
the
embryonic
chick
lens
becomes
less
sensitive
to
C02
-
treatment
with
de-
velopment
.
J
.
Cell
Biol
.
92
:694-705
.
Spray,
D
.
C
.,
and
J
.
M
.
Burt
.
1990
.
Structure-activit
y
relations
of the
cardiac
gap
junction channel
.
Am
.
J
.
Physiol
.
258
:C195-C205
.
Stagg,
R
.
B
.,
and
W
.
H
.
Fletcher
.
1990
.
The
hormone-induced
regulation
of
contact-dependent
cell-cell
communication
by
phosphorylation
.
Endocr
.
Rev
.
11
:302-325
.
Stevenson,
B
.
R
.,
and
D
.
L
.
Paul
.
1989
.
Th
e
molecular
constituents
of
intercel-
lularjunctions
.
Curr
.
Op
.
Cell
Biol
.
1
:884-891
.
Swenson,
K
.
I
.,
H
.
Piwnica-Worms,
H
.
McNamee,
and
D
.
L
.
Paul
.
1990
.
Tyrosin
e
phosphorylation
of
the
gap
junction
protein
connexin43
is
required
for the
pp60v-src-induced
inhibition
of
communication
.
Cell
Regulation
.
1
:989-1002
.
Traub,
O
.,
J
.
Look,
D
.
Paul,
and
K
.
Willecke
.
1987
.
Cyclic
adenosine
mono-
phosphate
stimulates biosynthesis
and
phosphorylation
of the
26
kDa
gap
junction
protein
in
cultured
mouse
hepatocytes
.
Eur
.
J
.
Cell
Biol
.
43:48-54
.
Yamasaki,
H
.
1990
.
Gap
junctional
intercellular
communication
and
carcino-
genesis
.
Carcinogenesis
.
11
:1051-1058
.
Yu,
J
.,
D
.
A
.
Fischman,
and
T
.
L
.
Steck
.
1973
.
Selectiv e
solubilization
of
pro-
teins
and
phospholipids
from
red
blood
cell
membranes
by
nonionic
deter-
gents
.
J
.
Supramol
.
Struct
.
1
:233-248
.
Zampighi,
G
.,
and
J
.
D
.
Robertson
.
1973
.
Fin
e
structure
of the
synaptic
discs
separated
from
the goldfish
medulla
oblongata
.
J
.
Cell
Biol
.
56
:92-105
.
137
4
on July 10, 2011jcb.rupress.orgDownloaded from
Published December 1, 1991
