Content uploaded by Jan Egebjerg
Author content
All content in this area was uploaded by Jan Egebjerg on Sep 26, 2017
Content may be subject to copyright.
Proc.
Nati.
Acad.
Sci.
USA
Vol.
90,
pp.
755-759,
January
1993
Neurobiology
Ca2+
permeability
of
unedited
and
edited
versions
of
the
kainate
selective
glutamate
receptor
GluR6
JAN
EGEBJERG*
AND
STEPHEN
F.
HEINEMANN
Molecular Neurobiology
Laboratory,
Salk
Institute,
La
Jolla,
CA
92037
Contributed
by
Stephen
F.
Heinemann,
October
12,
1992
ABSTRACT
The
Ca2+
permeability
of
the
kainate
selec-
tive
glutamate
receptor
GluR6
depends
on
the
editing
of
the
RNA
(or
DNA).
The
unedited
version
of
GIuR6,
GluR6Q,
encodes
a
glutamine
at
position
621
(Q/R
site)
and
exhibits
a
Ca2+/monovalent
ion
permeability
ratio
of
1.2,
while
the
edited
version
of
GIuR6,
GluR6R,
encodes
an
arginine
at
position
621
and
exhibits
a
permeability
ratio
of
0.47.
Kainate
activation
of
the
GluR6
receptor
results
in
currents
that
are
modulated
by
extracellular
calcium
ions.
Permeability
ratios
of
other
divalent
ions
indicate
that
the
Q/R
site
is
not
the
only
determinant
for
divalent
ion
permeability.
The
level
of
editing
of
the
receptor
will
determine
the
Ca2+
influx
through
the
GIuR6
receptor
channels
and,
consequently,
may
modulate
the
synaptic
activity.
Ca2+
flux
through
glutamate
receptors
is
thought
to
have
a
key
role
in
long-term
potentiation,
excitotoxic
cell
death,
and
epilepsy.
In
most
neurons
that
have
been
studied
the
non-N-
methyl-D-aspartate
(NMDA)
ionotropic
glutamate
receptors
exhibit
low
permeability
to
Ca2+
(1),
although
in
some
neurons
and
glia
cells
the
Ca2+/monovalent
ion
permeability
ratio
(Pca/Pmono)
is
considerably
higher
(2-6).
Two
types
of
non-
NMDA
glutamate
receptors
have been
observed
in
cultured
rat
hippocampal
neurons.
Type
I
receptors
have
a
low
PCa/
Pmono
and
exhibit
a
linear
current-voltage
(I-V)
relationship,
while
the
type
II
receptors
with
a
high
Pca/Pmoijo
exhibit
a
strongly
inward
rectifying
I-V
relationship
(4).
However,
non-NMDA
glutamate
receptors
present
in
salamander
bipo-
lar
cells
exhibit
a
high
PCa/PNa
but
a
linearI-V
relationship
(5).
Recent
cloning
experiments
have
revealed
the
existence
of
at
least
three
structurally
distinct
classes
of
non-NMDA
receptors
in
mammals:
GluR1-GluR4,
which
are
activated
by
both
kainate
and
DL-a-amino-3-hydroxy-5-methyl-4-
isoxazolepropionic
acid
(AMPA)
(7-11);
GluR5-GluR7,
of
which
GluR5
and
GluR6
generate
homomeric
channels
acti-
vated
by
kainate
but
not
by
AMPA
(12-15);
and
KA-1
and
KA-2,
which
do
not
generate
functional
homomeric
channels
but
bind
kainate
but
not
AMPA
(16-18).
Coexpression
of
KA-2
with
GluR5
or
GluR6
potentiates
the
response
com-
pared
to
the
homomeric
GluR5
or
GluR6,
respectively,
and
the
heteromeric
channels
are
activated
by
both
AMPA
and
kainate
(17,
18).
The
Ca2+
permeability
and
rectification
properties
of
the
GluR1-GluR4
class
of
subunits
depend
strongly
on
the
sub-
unit
composition
of
the
receptor.
GluR1,
GluR3,
and
GluR4,
in
combination
with
each
other
or
as
homomeric
channels,
generate
strongly
inward
rectifying
receptors
permeable
to
Ca2+-i.e.,
high
PCa/Pmono.
However,
when
the
GluR2
sub-
unit
is
a
constituent
of
the
receptor
complex,
the
I-V
rela-
tionship
becomes
linear
and
Ca2+
permeability
is
greatly
reduced
(19).
Mutagenesis
studies
have
revealed
that
the
presence
of
either
a
glutamine
or
an
arginine
in
the
putative
transmembrane
region
II
can
account
for
the
permeability
and
rectification
properties
(20,
21).
GluR1,
GluR3, and
GluR4
subunits
all
have
a
glutamine,
while
GluR2
has
an
arginine.
Thus,
the
presence
of
an
arginine
correlates
with
a
linear
I-V
relationship
and
low
PCa/Pmono.
Recently,
it
has
been
shown
that
this
critical
arginine
is
not
encoded
in
the
GluR2
gene
but
probably
results
from
almost
1001o
editing
of
the
mRNA.
Similar
editing
takes
place
in
the
expression
of
GluR5
and
GluR6
subunits
but
not
to
the
same
extent;
30%o
of
the
GluR5
and
"70%6
of
the
GluR6
mRNAs
are
edited
(22).
In
this
report,
we
study
the
effect
of
this
editing
process
on
the
divalent
ion
permeability
of
GluR6
receptors.
MATERIALS
AND
METHODS
The
GluR6Q
version
was
generated
by
mutagenesis
of
a
cDNA
coding
for
the
GluR6R
subunit
(23).
Oocytes
were
isolated
from
Xenopus
frogs
and
injected
with
5-15
ng
of
RNA
in
vitro
transcribed
as
described
(7).
Recordings
were
performed
4-14
days
after
injection
under
a
two-electrode
voltage
clamp
with
an
Axoclamp
2A
ampli-
fier.
Both
recording
and
current
electrodes
were
filled
with
3
M
KCl.
I-V
relationships
were
obtained
by
applying
2-sec
voltage
ramps
in
the
presence
of
agonist
and
subtracting
the
average
resting
I-V
curve
obtained
before
and
after
agonist
application.
When
changing
solutions,
the
oocyte
would
in
general
be
equilibrated
with
the
new
solution
for
2
min
before
applying
the
agonist.
I-V
curves
obtained
from
successive
applications
of
agonist
in
the
new
solution
established
that
the
2-min
perfusion
was
sufficient
to
exchange
the
solutions
and
the
internal
ion
concentrations
were
constant
for
the
duration
of
the
experiments.
The
following
solutions
were
used:
low-Ca2+/Ringer's,
15
mM
Hepes-NaOH,
pH
7.4/90
mM
NaCl/1
mM
KCl/0.1
mM
CaCl2/1
mM
MgCl2.
The
solution
used
for
the
divalent
permeability
studies
was
15
mM
Hepes/80
mM
N-methylglucamine
(pH
adjusted
to
7.4
by
HCl)/10
mM
MgCl2,
CaCl2,
SrC12,
or
BaCl2.
The
Ca2+
solutions
contain
15
mM
Hepes (pH
7.4)
and
either
2,
5,
10,
or
20
mM
CaCl2
and
N-methylglucamine
for
osmotic
balance.
A
correction
for
solution-dependent
junction
potentials
(3-6
mV)
is
included
in
the
reversal
potentials.
Oocytes
were
injected
with
50-100
nl
of
20
mM
1,2-bis(2-
aminophenoxy)ethane-N,N,N',N'-tetraacetic
acid
(pH
8.0)
(BAPTA)
5-180
min
prior
to
recording.
Each
oocyte
was
exposed
to
10
1LM
concanavalin
A
(Con
A;
Sigma
type
IV)
for
5
min
in
the
recording
chamber
if
not
otherwise
indicated.
Activities
were
calculated
by
Guggenheim's
modification
of
the
Debye-Huckel
expression
for
activity
coefficients
(24).
RESULTS
The
permeability
properties
for
divalent
cations
of
homo-
meric
channels
generated
from
GluR6
subunits
were
studied
Abbreviations:
NMDA,
N-methyl-D-aspartate;
Pca/Pmono,
Ca2+/
monovalent
ion
permeability
ratio;
AMPA,
DL-a-amino-3-hydroxy-
5-methyl-4-isoxazolepropionic
acid;
BAPTA,
1,2-bis(2-aminophe-
noxy)ethane-N,N,N',N'-tetraacetic
acid.
*To
whom
reprint
requests
should
be
addressed.
755
The
publication
costs
of
this
article
were
defrayed
in
part
by
page
charge
payment.
This
article
must
therefore
be
hereby
marked
"advertisement"
in
accordance
with
18
U.S.C.
§1734
solely
to
indicate
this
fact.
756
Neurobiology:
Egebjerg
and
Heinemann
in
Xenopus
oocytes
injected
with
in
vitro
transcribed
RNA.
The
receptors
were
activated
by
either
kainate
or
domoate
and
current
responses
were
recorded
with
a
two-electrode
voltage
clamp.
Most
experiments
were
performed
after
Con
A
treatment
of
the oocytes,
which
generally
potentiated
the
responses
at least
100-fold
(13).
This
treatment
was
particu-
larly
important
for analysis
of
the
GluR6R
variant,
which
only
gave
5-
to
50-nA
responses
to
kainate
before
Con
A
treatment.
Control
experiments
using
the
more
efficacious
agonist
domoate
established
that
Con
A
treatment
did
not
influence
the
permeability
properties
(data
not
shown).
The
Q/R
site
has
been
shown
to
control
the
rectifying
properties
of
the
GluRl-GluR4
class
of
subunits
when
cells
are
studied
in
high
sodium
solutions
(20,
25).
We
and
others
have
reported
a
similar
role
for
the
Q/R
site
in
controlling
rectification
in
GluR6
(17,
23).
GluR6Q
receptors
exhibited
a
strongly
inward
rectifying
I-V
relationship
in
contrast
to
GluR6R
receptors,
which
exhibited
a
slight
outward
rectifi-
cation
(Fig.
1B).
The
experiments
were
performed
in
low
Ca2+
(0.1
mM)/Ringer's
solution
to
avoid
interference
from
endogenous
Ca2+-activated
Cl-
channels.
Substitution
of
the
external
chloride
by
the
impermeable
methanesulfonate
did
not
change
the
reversal
potential
of
GluR6R
(13)
or
GluR6Q
(data
not
shown),
indicating
that
there
is
no
significant
Cl-
contribution
to
the
current.
To
investigate
the
Ca2+
permeability
of
the
two
variants
of
the
receptor,
we
performed
experiments
in
solutions
con-
taining
different
concentrations
of
Ca2+
as
the
only
perme-
able external
ion.
The
external
Na+
and
K+
ions
were
substituted
by
isoosmolar
concentrations
of
the
impermeable
cation
N-methylglucamine.
To
avoid
activation
of
endoge-
nous
Ca2+-activated
Cl-
channels
(26),
which
are
activated
in
the
rising
phase
of
the
GluR6
response,
the
fast
Ca2+
ion
A
GLUR6
FTLLNSFW
L
(R/Q)
Q
EL
GLUR1-4
FGIFNSLWSLG
(LR/Q)Q
DI
B
1.5
1
.0
0.5
20
1.5
-
FIG.
1.
(A)
Comparison
between
regions
of
GluR6
and
GluR1-
GluR4.
The
putative
transmembrane
region
is
underlined.
(B)
I-V
relationships
for
either
of
the
homomeric
GluR6
variants
obtained
from
oocytes
injected
with
RNA
encoding
either
GluR6Q
or
GluR6R.
I-V
was
recorded
during
a
2-sec
voltage
ramp
from
-80
to
30
mV
and
then
subtracting
the
average
resting
I-V
curve
obtained
before
and
after
agonist
application.
I-V
curves
were
normalized
to
the
current
observed
at
-70
mV.
chelator
BAPTA
was
injected
prior
to
the
recordings.
A
number
of
tests
were
performed
to
ensure
an
efficient
block
of
the
Ca2+-induced
Cl-
currents
after
BAPTA
treatment.
After
recordings
from
each
oocyte
(normally
one
to
four
agonist
applications
of
5
sec),
the
oocyte
was
exposed
to
agonist
in
high
Ca2+
(20
mM).
Only
oocytes
that
did
not
show
any
change
in
the
reversal
potential
after
30
sec
were
used
for
the
analysis
(70%o
and
90%o
of
the
GluR6Q
and
GluR6R
injected
oocytes,
respectively).
The
strong
rectification
of
the
GluR6Q
receptor
current
makes
determination
of
the
reversal
potential
of
the
GluR6Q
receptor
very
sensitive
to
an
additional
Cl-
current.
The
rectification
permitted
an
addi-
tional
test
of
whether
BAPTA
could
control
calcium
and
prevent
secondary
activation
of
the
Cl-
current.
If
this
current
had been
activated,
then
a
large
outward
current
would
be
expected
at
positive
potentials.
Indeed,
this
was
the
case
in
oocytes
not
injected
with
BAPTA,
but
only
a
very
small
outward
current
was
seen
in
BAPTA-injected
oocytes
(Fig.
2B).
Since
the
I-V
curve
of
the
Cl-
current
is
known
(ref.
27;
unpublished
data),
a
worst-case
estimate
of
the
contribution
of
the
Cl-
current
at
the
reversal
potential
can
be
made
by
assuming
that
all
the
current
at
50
mV
is
Cl-
current
(e.g.,
10
nA
in
Fig.
2B),
the
Cl-
contribution
to
the
total
current
at
-40
mV
is,
because
of
the
outward
rectifi-
cation
of
the
Cl-
current,
estimated
to
be
0.2
nA,
and
this
would
change
the
reversal
potential
by
<1
mV
in
the
negative
direction.
Only
the
GluR6Q-injected
oocytes
that
did
not
develop
any
additional
outward
current
after
30
sec
of
agonist
application
in
20
mM
Ca2+
solution
were
used
for
the
analysis.
In
summary,
injection
of
BAPTA
to
a
final
internal
concentration
of
=2
mM
in
the
oocyte
is
sufficient
to
inhibit
Ca2+-induced
activation
of
the
Cl-
channels,
and
under
these
conditions
Ca2+
permeation
can
be
assessed.
The
reversal
potentials
were
used
to
assess
the
permeation
of
Ca2+
for
the
homomeric
GluR6
channels.
Fig.
2
shows
that
the
reversal
potentials
for
both
the
GluR6Q
and
GluR6R
channels
become
more
depolarized
as
a
function
of
Ca2+
activity
in
the
extracellular
solution,
indicating
that
both
channels
are
permeable
to
Ca2+
ions.
The
reversal
potentials
for
GluR6R
channels
are
obviously
more
negative
than
for
GluR6Q
at
a
given
Ca2+
concentration,
suggesting
a
higher
Pca/Pmono
for
GluR6Q
than
for
GluR6R
(Fig.
2).
To
examine
the
Ca2+
permeability
more
quantitatively,
the
reversal
potentials
were
recorded
in
low
Ca2+/Ringer's
so-
lution
(containing
Na+
and
K+
as
charge
carriers)
and
after-
wards
in
Na+,K+-free
10
mM
Ca2+
solution.
The
reversal
potential
in
the
low
Ca2+/Ringer's
solution
was
used
to
estimate
the
intracellular
monovalent
ion
activity,
which
showed
some
variability
between
oocytes
from
different
batches
and
between
oocytes
from
the
same
batch
at
different
times
after
injection.
Assuming
that
the
intracellular
mono-
valent
ion
concentration
varies
only
marginally
during
the
shift
of
bathing
solution,
the
permeability
ratios
were
calcu-
lated
from
the
Goldman-Hodgkin-Katz
equation
modified
to
include
divalent
cations
(1),
assuming
no
anion
permeability
and
equal
permeability
for
sodium
and
potassium.
The
latter
was
confirmed
by
substituting
extracellular
sodium
with
potassium,
which
did
not
change
the
reversal
potential
sig-
nificantly
(0.4
+
0.5
mV;
n
=
6).
Fig.
2
B
and
C
shows two
typical
I-V
curves
for
the
GluR6Q
and
-R
variants.
The
permeability
ratio
was
calculated
for
each
oocyte
and
aver-
aged.
GluR6Q
showed
AV,
=
29
+
2
mV
(n
=
10)
negative
shift
of
the
reversal
potential
in
the
Ca2+
solution
compared
to
normal
Ringer's
solution,
implying
PCa/Pmono
=
1.2
+
0.1.
The
numbers
for
the
GluR6R
variant
were
iVv
=46
±
2
mV
(n
=
10)
and
PCa/Pmono
=
0.47
±
0.03.
A
similar
experiment
was
performed
with
60
mM
Ba2+
as
the
external
charge
carrier,
showing
a
PBa/Pmono
of
0.8
(n
=
6)
and
0.6
(n
=
6)
for
GluR6Q
and
GluR6R,
respectively.
The
reversal
potentials
were
less
than
or
equal
to
-110
mV
in
the
absence
of
external
Proc.
Natl.
Acad
Sci.
USA
90
(1993)
Neurobiology: Egebjerg
and
Heinemann
-20
r
-30
-
A
*GIuR6O
o
GluR6R
-60
Ba
Sr
-40
-
-50
F
-60
F
70
2
3
4
5
6
7
8
910
Iog
(aj
)(m
M)
200
r
Sr
Mg
20
R
100
F
Ca++
-40
-20
Na+
-100
-200
Membrane
potential
(mV)
20
c
c
Mg
g
7
/I*W
Sr
f j
Co
-40
-20
C
500
GluR6R
250
Ca++
60
-40
-20
Na+
c
1-
cJ
4)
0I
Membrane
potential
(mV)
20
-250
_
-50s
_
2-
FIG.
2.
(A)
Reversal
potentials
obtained
in
Ca2+/Ringer's
solu-
tion
containing
2-20
mM
Ca2+
for
GluR6Q
or
GluR6R.
(B)
I-V
curves
for
homomeric
GluR6Q
receptors
in
either
low
Ca2+/Ringer's
(Na+)
or
10
mM
Ca2+/Ringer's
(Ca2+)
solution.
(C)
As
in
B
for
GluR6R.
Oocytes
were
injected
with
100
nl
of
a
20
mM
BAPTA
solution
prior
to
recording
in
order
to avoid
activation
of
the
endogenous
Ca2+-
activated
Cl-
channel
(see
text).
Na+,
K+,
and
divalent
ions,
indicating
a
negligible
contribu-
tion
of
N-methylglucamine
to
the
inward
current.
We
also
studied
the
permeability
properties
of
other
alka-
line
earth
metal
ions.
As
shown
in
Fig.
3,
all
ions
tested
were
able
to
permeate
the
channels.
Determination
of
the
reversal
potentials
(Table
1)
indicated
a
different
order
of
permeability
for
the
two
GIuR6Q
and
-R
variants
with
the
order
Ca2+
2
-500
-1000
FIG.
3.
I-V
relationship
for
GluR6Q
(A)
and
GluR6R
(B)
in
solutions
containing
the
indicated
ion
at
10
mM
as
the
only
extra-
cellular
charge
carrier.
I-V
curves
were
obtained
as
2-sec
voltage
ramps
from
-80
or
-100
to
40
mV.
*,
Recording
in
Ca2+
solution
for
GluR6Q
was
performed
after
30-sec
Con
A
treatment,
since
a
longer
treatment
would
potentiate
the
response
to
a
level
where
the
Ca2+
influx
exceeds
the
Ca2+
buffering
capacity
of
the
injected
BAPTA,
consequently
making
inhibition
of
the
Ca2+-activated
Cl-
current
impossible.
An
additional
5-min
exposure
to
Con
A
was
performed
before
agonist
applications
in
the
other
solutions.
The
GluR6R-
injected
oocyte
was
treated
for
3
min
with
Con
A
before
kainate
was
applied
in
the
Ca2+
solution
and
for
an
additional
3
min
before
kainate
application
in
the
other
solutions.
(Inset)
To
assess
the
reversal
potential
more
precisely,
the
I-V
curves
for
GluR6R
obtained
in
the
Mg2+,
Sr2+,
and
Ba2+
solutions
were
fitted
to
a
third-order
polyno-
mial
function.
Mg2+
>
Sr2+
Ba2+
for
GluR6Q
and
Ba2+
Ca2+
>
Sr2+
Mg2+
for
GluR6R.
An
interesting
finding
is
that
the
magnitude
of
the
current
was
strongly
dependent
on
the
type
of
divalent
ion
(Fig.
3;
Table
1).
At
a
holding
potential
of
-80
mV,
the
inward
current
in
the
Ca2+
solution
was
much
larger
than
currents
measured
in
either
Mg2+,
Sr2+,
or
Ba2+
solutions
(Table
1).
The
currents
in
the
Ca2+
solution
were,
for
both
variants,
=5%
of
the
currents
measured
in
the
low
Ca2+/Ringer's
solution.
This
low
current
indicated
that
divalent
ions
might
inhibit
the
current
carried
by
monovalent
ions.
To
analyze
A
Proc.
Natl.
Acad.
Sci.
USA
90
(1993)
0
II)
0~
-a
(I)
L.
a)
(L)
20
757
B
GluR6Q
-100
-200
20
Potential
(mV)
-300
c
-
400
_
u
1500
Sr
Ba
20
40
758
Neurobiology:
Egebjerg
and
Heinemann
Table
1.
Permeability
ratios
determined
in
solutions
containing
the
indicated
ions
at
10
mM
as
the
only
external
charge
carrier
(n
=
5-9)
and
relative
current
carried
by
the
indicated
ions
compared
to
the
current
recorded
in
10
mM
Ca2+
solution
at
a
holding
potential
of
-80
mV
(n
=
4-6)
Relative
current
Permeability
ratio
(Px/Pmono)
(Ix/ICa,
%)
GluR6Q
GluR6R
GluR6Q
GluR6R
Mg2+
1.0
±
0.1
0.41
±
0.05
5.0
±
0.7
2.1
±
0.1
Ca2+
1.2
±
0.1
0.47
±
0.03
100
100
Sr2+
0.77
±
0.05
0.41
±
0.04
1.5
±
0.3
3.8
±
0.4
Ba2+
0.71
±
0.04
0.60
±
0.06
0.6
±
0.1 3.2
±
0.0
Current
ratio
between
Ca2+
and
Mg2+
was
measured
after
a
short
Con
A
application,
while
current
ratios
between
Mg2+,
Sr2+,
and
Ba2+
were
measured
after
extended
Con
A
treatment
(see
Fig.
3).
this
effect,
the
currents
were
measured
in
Ringer's
solution
containing
different
concentrations
of
Ca2+
or
Ba2+.
The
EC50
for
kainate
did
not
change
over
the
Ca2+
concentration
range
used
when
corrected
for
Ca2+-kainate
complexes
(28).
To
ensure
maximal
responses,
30
,uM
kainate
was
used
(i.e.,
30
times
the
EC50).
The
shapes
and
minima
of
the
curves
were
independent
of
the
magnitude
of
current
responses,
indicat-
ing
that
the
BAPTA
injection
sufficiently
inhibited
activation
of
the
Cl-
channels.
Fig.
4A
shows
that
Ca2+
at
-2
mM
gives
the
maximal
inhibition
with
an
increased
current
observed
for
higher
Ca2+
concentrations,
presumably
reflecting
the
in-
crease
of
the
extracellular
concentration
of
charge
carrier
and
perhaps
a
Ca2+-dependent
change
in
the
unitary
conductance
or
the
open
probability
of
the
receptor.
The
inhibitory
effect
of
Ba2+
is
not
overcome
at
higher
Ba2+
concentrations,
suggesting
a
difference
in
the
mechanism
of
Ba2W
permeation
(see
Discussion).
DISCUSSION
The
present
study
shows
that
homomeric
GluR6Q
receptors
exhibit
a
high
Pca/Pmono,
while
the
GluR6R
receptor
is
less
Ca2+
permeable.
Permeability
studies
performed
on
the
GluR1-GluR4
receptor
subunits
revealed
a
similar
pattern.
When
the
two
variants
of
GluR2
were
studied
in
a
mammalian
expression
system
PCa/PcS
was
found
to
be
1.2
for
the
GluR2Q
variant
and
0.05
for
the
GluR2R
variant
(21),
com-
pared
to
1.2
and
0.47
for
the
GluR6Q
and
GluR6R
variants,
respectively.
PBa/Pmono
was
estimated
to
be
2-3
for
homo-
meric
GluRl
and
GluR3
but
it
was
<0.02
for
heteromeric
receptors
containing
GluR2
(20).
However,
the
GluR6
vari-
ants
exhibit
a
similar
PBa/Pmono
of
0.8
and
0.6
for
GluR6Q
and
GluR6R,
respectively.
These
differences
in
divalent
ion
se-
lectivity
between
different
GluR
subunits
suggest
that
diva-
lent
ion
permeability
is
not
solely
dependent
on
the
amino
acid
at
the
Q/R
site.
The
Ca2+-dependent
reduction
of
the
current
mediated
by
the
GluR6
variants
makes
the
activity
of
the
receptors
sensitive
to
extracellular
variations
in
Ca2+
concentrations.
That
dependency
is
not
observed
for
the
GluR1-GluR4
class
of
glutamate
subunits
(19).
If
receptors
generated
from
this
class
of
subunits
contain
GluR2,
they
exhibit
a
weak
reduc-
tion
in
total
current
at
high
extracellular
Ca2+
concentrations,
while
receptors
without
GluR2
carry
more
current
at
higher
Ca2+
concentrations
in
accordance
with
the
increased
con-
centration
of
charge
carrier
(19).
The
mechanism
underlying
the
Ca2+-induced
modulation
of
the
response
could
involve
a
high-affinity
divalent
ion
binding
site
in
the
channel
or
a
Ca2+
(or
divalent
ion)-dependent
change
in
open
probabilities
or
unitary
conductance
as
observed
for
the
neuronal
acetyl-
choline
receptors
(29,
30).
The
dependence
of
the
current
on
Ca2+
concentration
suggests
a
blocking
mechanism
analo-
gous
with
the
block
of
the
voltage-gated
Ca2+
channel
(31,
32).
The
increased
current
at
high
Ca2+
concentrations
might
then
reflect
an
increased
exit
rate
due
to
electrostatic
repul-
sion
between
Ca2+
ions
in
the
channel.
A
blocking
model
could
explain
the
apparent
paradox
for
GluR6R,
where
Ba2+,
despite
a
permeability
ratio
to
monovalent
ions
higher
than
that
of
Ca2+,
shows
a
decrease
in
total
current
at
higher
Ba2+
concentrations
and
not
an
increase
as
observed
for
Ca2+
(Fig.
4).
If
indeed
Ba2+
binds
in
the
channel
with
a
higher
affinity
than
Ca2+
it
may
inhibit
the
monovalent
current
more
effi-
ciently
than
Ca2+.
The
consequently
slower
passage
of
Ba2+
ions
will
then
reduce
the
total
current
mostly
carried
by
Na+
more
efficiently.
The
observation
that
subtypes
of
the
non-NMDA
receptors
are
permeable
to
Ca2+
(4,
6,
19,
20)
suggests
an
additional
mechanism
for
glutamate-induced
Ca2+
influx.
This
mecha-
nism
is
most
efficient
at
polarized
or
hyperpolarized
poten-
tials
in
contrast
to
the
Ca2+
flux
through
the
NMDA
subtype
of
the
glutamate
receptors,
which
requires
a
depolarization
of
the
membrane
to
relieve
the
Mg2+
block
of
the
NMDA
0.1
1
10
External
Ca++
concentration
(mM)
3.5
0)
(0
+
0
+
0.
0.-
C-)C
3.0
V
2.5
I
0
GluR6R
2.0
F
1.5
F
1.0
F
0.5
F
0.0
100
0.1
1
10
External
Ba++
concentration
(mM)
FIG.
4.
Inward
current
measured
at
a
holding
potential
of
-70
mV
in
solutions
containing
different
Ca2+
(A)
or
Ba2+
(B)
concentrations.
The
Ca2+
or
Ba2+
concentration
was
varied
from
0.1
to
50
mM
in
Ringer's
solution
containing
90
mM
NaCl,
1
mM
KCI,
and
15
mM
Hepes
NaOH
(pH
7.5).
Oocytes
were
injected
with
BAPTA
and
treated
with
Con
A
prior
to
the
recordings.
3.5
3.0
0
+
0.
C+
C.)
(.)
0
U
N
E
,
_
C
%-
c
E3
0
0~
L.
z
3
B
*
GluR6(
2.5
2.0
1.5
1.0
0.5
0.0
100
Proc.
Nad.
Acad
Sci.
USA
90
(1993)
Proc.
Natl.
Acad.
Sci.
USA
90
(1993)
759
receptor.
The
glutamate-induced
Ca2+
flux
through
the
non-
NMDA
receptors
may
therefore
be
most
significant
for
stimuli
that
do
not
depolarize
the
membrane
sufficiently
to
activate
the
NMDA
receptor
or
voltage-activated
Na+
or
Ca2+
channels.
The
membrane
potential
may
remain
polar-
ized
due
to
simultaneous
inhibitory
stimuli,
or
the
Ca2+-
permeable
receptors
may
generate
a
positive
feedback
sys-
tem
through
Ca2+-activated
Cl-
or
K+
channels.
The
in-
creased
internal
Ca2+
concentration
may
induce
additional
long-term
effects
by
affecting
kinases
and,
consequently,
the
phosphorylation
level
of
other
receptor
systems.
Note
Added
in
Proof.
Control
experiments
using
a
whole-cell
patch-
clamp
technique
on
transfected
mammalian
293
kidney
cells
con-
firmed
that
both
versions
of
the
GluR6
(R/Q)
have
significant
permeability
ratios.
PCa/Pmono
ratios
were
the
same
in
oocytes
and
293
kidney
cells
for
GluR6Q.
However,
the
GluR6R
showed
a
higher
PCa/Pmono
(3.3
±
0.5)
in
293
cells.
This
unexpected
difference
remains
to
be
investigated.
We
thank
R.
Dingledine,
R.
Hume,
R.
Papke,
and
J.-P.
Pin
for
help
and
advice
throughout
these
studies.
We
thank
S.
Traynelis
for
sharing
unpublished
data
concerning
the
permeability
ratios
in
293
cells.
This
work
was
supported
by
the
Danish
Medical
Research
Council
(J.E.)
and
by
grants
to
S.F.H.
from
the
McKnight
Founda-
tion,
the
Human
Frontier
Science
Program,
the
Muscular
Dystrophy
Association
of
America,
and
the
National
Institutes
of
Health
(NS
28709
and
NS
11549).
1.
Mayer,
M.
L.
&
Westbrook,
G.
L.
(1987)
J.
Physiol.
(London)
394,
501-527.
2.
Murphy,
S.
N.,
Thayer,
S.
A.
&
Miller,
R.
J.
(1987)
J.
Neu-
rosci.
7,
4145-4158.
3.
Glaum,
S.
R.,
Holzwarth,
J.
A.
&
Miller,
R.
J.
(1990)
Proc.
Nati.
Acad.
Sci.
USA
87,
3454-3458.
4.
Iino,
M.,
Ozawa,
S.
&
Tsuzuki,
K.
(1990)
J.
Physiol.
(London)
424,
151-165.
5.
Gilbertson,
T.
A.,
Scobey,
R.
&
Wilson,
M.
(1991)
Science
251,
1613-1615.
6.
Burnashev,
N.,
Khodorova,
A.,
Jonas,
P.,
Helm,
P.
J.,
Wis-
den,
W.,
Monyer,
H.,
Seeburg,
P.
H.
&
Sakmann,
B.
(1992)
Science
256,
1566-1570.
7.
Hollmann,
M.,
O'Shea-Greenfield,
A.,
Rogers,
S.
W.
&
Heine-
mann,
S.
(1989)
Nature
(London)
342,
643-648.
8.
Boulter,
J.,
Hollmann,
M.,
O'Shea-Greenfield,
A.,
Hartley,
M.,
Deneris,
E.
S.
&
Heinemann,
S.
(1990)
Science
249,
1033-1037.
9.
Keinanen,
K.,
Wisden,
W.,
Sommer,
B.,
Werner,
P.,
Herb,
A.,
Verdoom,
T.
A.,
Sakmann,
B.
&
Seeburg,
P.
H.
(1990)
Science
249,
556-560.
10.
Nakanishi,
N.,
Shneider,
N.
A.
&
Axel,
R.
(1990)
Neuron
5,
569-581.
11.
Sakimura,
K.,
Bujo,
H.,
Kushiya,
E.,
Araki,
K.,
Yamazaki,
M.,
Yamazaki,
M.,
Meguro,
H.,
Warashina,
A.,
Numa,
S.
&
Mishina,
M.
(1990)
FEBS
Lett.
272,
73-80.
12.
Bettler,
B.,
Boulter,
J.,
Hermans-Borgmeyer,
I.,
O'Shea-
Greenfield,
A.,
Deneris,
E.
S.,
Moll,
C.,
Borgmeyer,
U.,
Hollmann,
M.
&
Heinemann,
S.
(1990)
Neuron
5,
583-595.
13.
Egebjerg,
J.,
Bettler,
B.,
Hermans-Borgmeyer,
I.
&
Heine-
mann,
S.
(1991)
Nature
(London)
351,
745-748.
14.
Bettler,
B.,
Egebjerg,
J.,
Sharma,
G., Pecht, G.,
Hermans-
Borgmeyer,
I.,
Moll,
C.,
Stevens,
C.
F.
&
Heinemann,
S.
(1992)
Neuron
8,
257-265.
15.
Sommer,
B.,
Burnashev,
N.,
Verdoorn,
T.
A.,
Kainanen,
K.,
Sakmann,
B.
&
Seeburg,
P.
H.
(1992)
EMBO
J.
11,
1651-1556.
16.
Werner,
P.,
Voight,
M.,
Keinanen,
K.,
Wisden,
W.
&
Seeburg,
P.
H.
(1991)
Nature
(London)
351,
742-744.
17.
Herb,
A.,
Burnashev,
N.,
Werner,
P.,
Sakmann,
B.,
Wisden,
W.
&
Seeburg,
P.
H.
(1992)
Neuron
8,
775-785.
18.
Sakimura,
K.,
Morita,
T.,
Kushiya,
E.
&
Mishina,
M.
(1992)
Neuron
8,
267-274.
19.
Hollmann,
M.,
Hartley,
M.
&
Heinemann,
S.
(1991)
Science
252,
851-853.
20.
Hume,
R.
I.,
Dingledine,
R.
&
Heinemann,
S.
(1991)
Science
253,
1028-1031.
21.
Burnashev,
N.,
Monyer,
H., Seeburg,
P.
H.
&
Sakmann,
B.
(1992)
Neuron
8,
189-198.
22.
Sommer,
B.,
Kohler,
M.,
Sprengel,
R.
&
Seeburg,
P.
H.
(1991)
Cell
67,
11-20.
23.
Dingledine,
R.,
Hume,
R.
I.
&
Heinemann,
S.
F.
(1992)
J.
Neurosci.
12,
4080-4087.
24.
Robinson,
R.
A.
&
Stokes,
R.
H.
(1960)
Electrolyte
Solutions
(Butterworths,
London),
2nd
Ed.
25.
Verdoorn,
T.
A.,
Brunashev,
N.,
Monyer,
H.,
Seeburg,
P.
H.
&
Sakmann,
B.
(1991)
Science
252,
1715-1718.
26.
Miledi,
R.
(1982)
Proc.
R.
Soc.
London
Ser.
B
215,
491-497.
27.
Miledi,
R.
&
Parker,
I.
(1984)
J.
Physiol.
(London)
357,
173-183.
28.
Gu,
Y.
&
Huang,
L.-Y.
M.
(1991)
Neuron
6,
777-784.
29.
Mulle,
C.,
Choquet,
D.,
Korn,
H.
&
Changeux,
J.-P.
(1992)
Neuron
8,
135-143.
30.
Vernino,
S.,
Amador,
M.,
Luetje,
C.
W.,
Patrick,
J.
&
Dani,
J.
A.
(1992)
Neuron
8,
127-134.
31.
Hess,
P.,
Lansman,
J.
B.
&
Tsien,
R.
W.
(1986)
J.
Gen.
Physiol.
88,
293-319.
32.
Hille,
B.
(1992)
Ionic
Channels
of
Excitable
Membranes
(Sin-
auer,
Sunderland,
MA),
2nd
Ed.
Neurobiology:
Egebjerg
and
Heinernann