![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Vol. 292, Issue 3, 1104-1110, March 2000
Department of Neuroscience, University of Pittsburgh (Y.L.-S., J.W.J.) and Department of Neurobiology, University of Pittsburgh School of Medicine (E.A.), Pittsburgh, Pennsylvania
 |
Abstract |
|---|
 Top
 Abstract
 Introduction
Materials and Methods
Results
Discussion
References
|
|---|
Intracellular Mg2+ (Mgi2+) inhibits the N-methyl-D-aspartate (NMDA) subtype of glutamate receptors in cultured cortical neurons. To examine the effects of Mgi2+ on recombinant NMDA receptors composed of subunit combinations found in cortical neurons, we expressed heteromeric receptors composed of NR1/NR2A and of NR1/NR2B subunits in Chinese hamster ovary (CHO) cells. We recorded whole-cell currents from the recombinant receptors in the absence and presence of Mgi2+. The voltage dependence of control (0 Mgi2+) NMDA-activated currents obtained from CHO cells transfected with NR1/NR2A and with NR1/NR2B receptors showed outward rectification, a property that has been observed previously in native cortical NMDA receptors. The magnitude and voltage dependence of inhibition by Mgi2+ of NMDA-activated currents were similar in CHO cells transfected with NR1/NR2A receptors, CHO cells transfected with NR1/NR2B receptors, and in cultured neurons expressing native NMDA receptors. These observations suggest that Mgi2+ has uniform effects on the native NMDA receptors expressed in cortical neurons. Furthermore, inhibition by Mgi2+ must not depend on intracellular factors or post-translational receptor modifications that are specific to neurons. Finally, the results indicate that the previously observed differences between whole-cell and outside-out patch measurements of Mgi2+ inhibition could not result from poor control of voltage or Mgi2+ concentration in the dendrites of neurons. The most likely alternative explanation is that patch excision causes an alteration in NMDA receptors that results in more effective inhibition by Mgi2+.
| |
Introduction |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
The
N-methyl-D-aspartate (NMDA) subtype of
glutamate receptors is thought to be of fundamental importance in brain
physiology and pathology. Probably because of its intimate involvement
in the function and dysfunction of the brain, NMDA receptors are under
tight regulation by multiple endogenous factors (for review, see McBain
and Mayer, 1994
). One type of regulation, the voltage-dependent block
by extracellular Mg2+ (Mge2+;
Mayer et al., 1984; Nowak et al., 1984; Jahr and Stevens, 1990), is
well established to be critical for many of the roles that NMDA
receptors play (McBain and Mayer, 1994).
The channel of NMDA receptors also can be blocked by intracellular
Mg2+ (Mgi2+). Unlike block by
Mge2+, block by Mgi2+ increases with
membrane depolarization (Johnson and Ascher, 1990). Mge2+ and Mgi2+ block the channel of
NMDA receptors at two distinct sites on opposite sides of the
selectivity filter (Johnson and Ascher, 1990; Li-Smerin and Johnson,
1996a). Because its binding site is located at a position of critical
importance within the channel, Mgi2+ has become an
important tool for study of the structure and function of recombinant
NMDA receptors (Kupper et al., 1996, 1998; Wollmuth et al., 1998).
In this study, we had two main goals. First, we compared inhibition by
Mgi2+ of recombinant NMDA receptors composed of
NR1/NR2A and of NR1/NR2B subunits. Although there is extensive evidence
that the external channel-blocking site is similar in these two subunit
combinations, there have been no studies of the NR2 subunit dependence
of block at the Mgi2+ blocking site. Much of the
previous work on block by Mgi2+ (Johnson and Ascher,
1990; Li-Smerin and Johnson, 1996a,b) has been performed on native NMDA
receptors in cultured cortical neurons. NMDA receptors in this
preparation are composed predominantly of NR1 subunits coassembled with
NR2A and/or NR2B subunits (Stern et al., 1992, 1994; Zhong et al.,
1994; Béhé et al., 1995; Blanpied et al., 1997).
Comparisons of the Mgi2+ effects on NR1/NR2A and
NR1/NR2B receptors permitted us to test the validity of the assumption
that Mgi2+ has homogenous effects on NMDA receptors
expressed in cultured cortical neurons.
A second goal of this study was to examine possible explanations for
the observation that Mgi2+ inhibits steady-state NMDA
responses less effectively in whole-cell experiments than in
excised-patch experiments (Li-Smerin and Johnson, 1996b). Understanding
of this discrepancy is necessary to determine the effects of
Mgi2+ on NMDA responses under physiological conditions.
Steady-state current in excised patches was quantified as "mean patch
current," a measurement of the integral of NMDA-activated current
flow across the patch. The discrepancy between Mgi2+
inhibition of whole-cell and mean patch current suggests a difference between the preparations either in an experimental condition important to Mgi2+ block, or in NMDA receptor properties. A
difference in experimental conditions that might explain this
discrepancy results from the presence of dendrites on neurons. If
membrane potential and/or the concentration of Mgi2+
([Mg2+]i) of dendrites during
whole-cell recording were not adequately controlled, an artifactual
determination that whole-cell currents are inhibited less effectively
than mean patch current could result. To test this possibility, we
expressed recombinant NMDA receptors in Chinese hamster ovary (CHO)
cells, which are spatially compact. If voltage and/or
[Mg2+]i in the dendrites of cultured
neurons were inadequately controlled, then the results obtained from
CHO cells should resemble the results obtained from excised patches
rather than whole neurons. By comparing the Mgi2+
inhibition of transfected CHO cells, whole neurons, and excised patches, we determined whether the dendrites of neurons influence the
measured effects of Mgi2+. Our results support the
alternative explanation that patch excision affects
Mgi2+ inhibition by inducing a change in NMDA receptor properties.
| |
Materials and Methods |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
CHO Cell Culture and Transfection of NMDA Receptors.
CHO-K1
cells (ATTC CCL61) were chosen for expression studies because they are
compact, do not form syncytia, and are easily transfected (Boeckman and
Aizenman, 1994, 1996). CHO-K1 cells were grown in Ham's F-12 nutrient
medium with 10% fetal bovine serum and 1 mM glutamine (CHO medium) and
passaged approximately every 2 days. The cDNAs encoding NR1-1a, NR2A,
and NR2B subunits were subcloned into mammalian expression vectors as
described previously (Boeckman and Aizenman, 1994, 1996). CHO cells
were transiently transfected with the indicated combination of NMDA receptor subunit constructs with LipofectAMINE (Gibco-BRL, Paisley, Scotland). To facilitate recognition of positively transfected cells
used for whole-cell recordings, a marker protein, green fluorescent
protein (GFP), was cotransfected with NMDA receptors in CHO cells. The
expression vector for GFP (Chalfie et al., 1994) was generated as
described previously (Blanpied et al., 1997). Approximately 24 h
before transfection, cells were seeded in CHO medium at a density of
3 × 105 cells/35-mm culture dish.
Transfections were accomplished by addition of 1.3 µg of total DNA
and 6 µl of LipofectAMINE in 1 ml of serum-free CHO medium per dish,
followed by a 4- to 5-h incubation at 37°C. The marker plasmid
(pCI/GFP) and total DNA were transfected at a ratio of 1:4.3, and NR1
and NR2 subunits were transfected at a ratio of 1:3 (Cik et al., 1993).
Twenty-four hours after transfection, cells were trypsinized and
replated at a 1:2 dilution onto 12-mm-diameter glass coverslips in
35-mm plastic Petri dishes. 5,7-Dichlorokynurenic acid (1 mM) was added to the medium to prevent the cell death due to the toxic effects of
NMDA receptor expression (Cik et al., 1994; Anegawa et al., 1995;
Boeckman and Aizenman, 1996). Cells were used ~40 to 50 h after
the start of transfection.
Whole-Cell Patch-Clamp Recordings.
The conventional
whole-cell configuration was used to record membrane currents with
pipettes pulled from borosilicate thin-walled glass with filaments
(Clark Electromedical, Reading, England) and an Axopatch 1D amplifier
(Axon Instruments, Foster City, CA). Partial compensation for series
resistance (50-80%) was performed in some experiments. The pipettes
had a resistance of 2 to 4 M
when filled with either the control
solution, which contained 0 MgCl2, 125 mM CsCl,
10 mM HEPES, and 10 mM EGTA, or the Mg2+
solution, which contained 15 mM MgCl2, 105 mM
CsCl, 10 mM HEPES, and 10 mM EGTA. The pH was adjusted to 7.2 with
CsOH. The concentration of free Mg2+
([Mg2+]i) in the Mg2+
solution was calculated as 10 mM with a program that subtracted Mg2+ bound to EGTA from total
Mg2+ using apparent affinity constants of EGTA
for Mg2+ (Li-Smerin and Johnson, 1996a). However,
the [Mg2+] in the Mg2+
solution was found actually to be 6.2 mM when it was measured with the
Mg2+-sensitive fluorescent dye mag-indo-1
(Li-Smerin et al., 1996). The value of 6.2 mM was therefore used as the
[Mg2+] in the Mg2+
solution in this study. The control extracellular solution contained 140 mM NaCl, 2.8 mM KCl, 1 mM CaCl2, and 10 mM
HEPES. pH was adjusted to 7.2 with NaOH. To activated NMDA receptors,
stocks of NMDA (10 mM), and glycine (10 mM) were diluted into the
control extracellular solution to achieve 30 µM NMDA and 10 µM
glycine (NMDA solution). All the chemicals were purchased from Sigma
Chemical Co. (St. Louis, MO).
). Briefly, barrels made of
square capillary tubing were arrayed in parallel and connected to
solution reservoirs. After formation of the whole-cell configuration, the barrel perfusing control solution was positioned ~0.2 mm away from the cell under study. Subsequent changes of the perfusion solution
were achieved by a lateral movement of the array of barrels. For each
measurement of whole-cell NMDA-activated current, the cell was
superfused with NMDA solution for typically 10 to 15 s. This
application was preceded and followed by superfusion of the control
solution. Whole-cell NMDA-activated currents were recorded at membrane
potentials ranging from 
60 to +60 mV in the following order: 60,
+60, 60, 40, +40, 60, 20, +20, and 60 mV. Currents recorded
at 60 mV were used to monitor the consistency in the magnitude of
NMDA responses during an experiment. Any cell in which the current
recorded at 60 mV varied during an experiment by >25% from the
first measurement was excluded. Data were discarded if voltage drift at
the end of an experiment was >4 mV.
Whole-cell currents were low-pass filtered at 500 Hz and recorded on
chart paper (Thermal Arraycorder WP 7700; Western Graphtec, Irvine,
CA). Currents also were low-pass filtered at 10 kHz, sampled at 44 kHz
with a Neuro-Corder (DR-890; Neuro Data Instruments Corp., New York,
NY) and stored on magnetic tapes. All experiments were performed at
room temperature (20-25°C).
Calculations and Statistics.
Steady-state whole-cell
currents under control conditions and during applications of NMDA plus
glycine were measured from chart paper records. The difference between
the responses in the absence and the presence of the agonists was
calculated as the amplitude of NMDA-activated current. To pool data
from different cells, current measurements at each membrane potential
were normalized for each cell to current measured at 60 mV, a voltage
at which inhibition by Mgi2+ is minimal.
Mgi2+ does weakly inhibit NMDA responses at 60 mV; at
the [Mg2+]i used herein (6.2 mM),
the strongest action of Mgi2+ quantified here in be a
15% inhibition of mean patch current (Li-Smerin and Johnson, 1996b).
The maximal inaccuracy of 15% introduced by normalization would not
affect interpretation of any of the data presented herein.
. The
variance of the factional current
[s2(FC)] was first obtained as
follows: s2 (FC) = (IMg/Icon)2
·
[s2Mg/(nMg
· I2Mg) + s2con/(ncon
· I2con)],
where s2 is variance and n
is the number of cells in the presence (Mg) and
absence (con) of Mgi2+. A
two-tailed Student's t test, with Bonferroni corrections
for multiple pairwise comparisons, was then used to examine the
significance of the differences between fractional currents. The
t value for the difference between any two fractional
currents being compared (FCx versus
FCy) was calculated as t = (FCx FCy)/[s2(FCx) + s2(FCy)]0.5.
|
|
|
| |
Results |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
Inhibition by Mgi2+ of Whole-Cell NR1/NR2A
Receptor-Mediated Currents.
To investigate the effects of
Mgi2+ on the NMDA receptor subunit combinations that
predominate in native cortical neurons, we used CHO cells transfected
with NR1 and NR2A or with NR1 and NR2B subunits. Figure 1A shows
examples of current records obtained in 0 Mgi2+ and in
6.2 mM Mgi2+ from CHO cells transfected with NR1 and
NR2A subunits. The NMDA-activated currents at 60 and +60 mV were of
similar amplitude with a
[Mg2+]i of zero, but
current amplitude is smaller at +60 mV than at 60 mV with a
[Mg2+]i of 6.2 mM. This
suggests that at positive potentials Mgi2+ inhibits
whole-CHO cell currents mediated by NR1/NR2A receptors.
60 mV was
1.11 ± 0.05 (n = 6). A similar nonlinear I-V
relation of NMDA-activated currents of native receptors has previously
been described and attributed to a voltage-dependent change in the
channel open probability (Nowak and Wright, 1992; Li-Smerin and
Johnson, 1996b). In contrast, inward rectification was observed in the
presence of 6.2 mM Mgi2+ (n = 5).
Compared with the mean normalized current in the absence of
Mgi2+, the mean normalized current in the presence of
6.2 mM Mgi2+ was reduced by 14% at +20 mV, 29% at +40
mV, and 42% at +60 mV. The differences between 0 and 6.2 mM
Mgi2+ at +40 and +60 mV are statistically significant
(P < .05, two-tailed t test). The amplitude
of current recorded at +20 mV and negative potentials was not
significantly affected by Mgi2+.
A central goal of this study is to determine whether
Mgi2+ inhibits NMDA responses similarly in transfected
CHO cells and in neurons. We addressed this goal by predicting
Mgi2+ inhibition of CHO cell currents based on previous
whole-cell measurements in cultured cortical neurons (Li-Smerin and
Johnson, 1996b; see Fig. 1 legend). As shown in Fig. 1B, the predicted I-V relation based on Mgi2+ action in cultured neurons
(dotted line) closely resembles the currents measured in CHO cells in
the presence of Mgi2+ (squares and solid lines). These
data are consistent with two conclusions: first, Mgi2+
inhibits native NMDA receptors in cultured cortical neurons and recombinant NR1/NR2A receptors via quantitatively similar mechanisms; and second, the complex geometry of cultured cortical neurons did not
interfere with accurate measurement of the action of
Mgi2+ in whole-cell experiments. However, because
cultured cortical neurons are likely to express NR1/NR2B as well as
NR1/NR2A receptors, these conclusions require further experiments on
NR1/NR2B receptors.
Inhibition by Mgi2+ of Whole-Cell NR1/NR2B
Receptor-Mediated Currents.
Fig. 1C shows examples of current
measurements from CHO cells transfected with NR1 and NR2B subunits
obtained at 60 and +60 mV with 0 Mgi2+ and with 6.2 mM Mgi2+. Figure 1D presents I-V relations of
whole-cell NMDA-activated current in the absence and the presence of
6.2 mM Mgi2+ recorded from CHO cells transfected with
NR1 and NR2B subunits. Consistent with the results obtained with the
NR1/NR2A receptors (Fig. 1, A and B), the I-V relation in the absence
of Mgi2+ shows outward rectification: the ratio of
current at +60 mV to that at 60 mV was 1.14 ± 0.08 (n = 5). This nonlinear relation of current and voltage
once again resembles that observed in native NMDA receptors. The
current amplitude in the presence of 6.2 mM Mgi2+
(n = 6) also was reduced in NR1/NR2B receptors at
positive potentials. Compared with the mean of currents measured in the
absence of Mgi2+, the mean of normalized currents
measured in the presence of 6.2 mM Mgi2+ was reduced by
18% at +20 mV, 36% at +40 mV, and 44% at +60 mV. The differences
between normalized currents measured in 0 and in 6.2 mM
Mgi2+ at +40 and +60 mV are statistically significant
(P < .05, two-tailed t test). Thus,
NR1/NR2B receptors also appear to be subject to voltage-dependent
inhibition by Mgi2+.
Comparison among Preparations of NMDA Response Rectification and
Mgi2+ Inhibition.
The above-mentioned data permit
comparison in transfected CHO cells and cultured neurons both of
control current rectification and of inhibition by
Mgi2+. In Fig. 2, four types of normalized currents
recorded at three positive potentials in the absence and presence of
Mgi2+ are compared: whole-cell currents of NR1/NR2A
receptors, whole-cell currents of NR1/NR2B, whole-cell currents of
native NMDA receptors, and mean patch currents of native NMDA
receptors. Data from native NMDA receptors are taken from Li-Smerin and
Johnson (1996b).
; Li-Smerin and Johnson, 1996b) and in
patches from Xenopus oocytes expressing NMDA receptors (Kupper et al., 1998). However, outward rectification of control whole-cell currents was significantly greater in neurons than in CHO
cells transfected with either NR1/NR2A (at +60 and +40 mV) or NR1/NR2B
(at +60 mV) receptors (Fig. 2). In addition, all control whole-cell
currents exhibited significantly less rectification than control mean
patch current (Fig. 2). Whole-cell and mean patch currents depend on
single-channel current amplitude and on channel open probability. Under
the recording conditions used herein, single-channel current amplitudes
of both native and heterologously expressed NMDA receptors depend
nearly linearly on voltage. It therefore is likely that the differences
between rectification of transfected CHO whole-cell, neuronal
whole-cell, and neuronal mean patch control currents reflect
differences in the voltage dependence of channel open probability
(Nowak and Wright, 1992).
Mgi2+ inhibited whole-cell currents mediated by
NR1/NR2A and by NR1/NR2B receptors in a voltage-dependent manner,
similar to its effect on native receptors of cultured neurons. The only
significant difference among normalized currents in the presence of
Mgi2+ was between neuronal whole-cell and mean patch
current at +60 mV (Fig. 2). However, normalized currents in the
presence of Mgi2+ reflects both rectification of
control current and the inhibitory effects of Mgi2+. We
therefore quantified fractional inhibition by Mgi2+ of
each type of NMDA-activated current (Fig. 3).
To compare the effects of Mgi2+ on NMDA responses from
each of the preparations shown in Fig. 2, we plotted the fractional
current (normalized current in the presence of Mgi2+
divided by the normalized current in the absence of
Mgi2+) as a function of voltage (Fig. 3). Consistent
with the data of Fig. 1, B and D, the effect of Mgi2+
on whole-cell currents measured from CHO cells transfected with either
subunit combination and on cultured cortical neurons are similar. There
was no significant difference at any potential in fractional current
among these three preparations. In contrast, the fractional mean patch
current was significantly different from fractional current in each of
the three whole-cell preparations at each voltage (P < .05).
| |
Discussion |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
In cultured cortical neurons, the preparation in which we
previously examined inhibition by Mgi2+ (Li-Smerin and
Johnson, 1996a,b), the predominant NMDA receptor subunits are NR1,
NR2A, and NR2B. This conclusion is supported by the similarity of the
single-channel conductance and kinetics of NMDA receptors in cultured
cortical neurons (Antonov and Johnson, 1996; Li-Smerin and Johnson,
1996a) and in recombinant NR1/NR2A or NR1/NR2B receptors (Stern et al.,
1992, 1994; Béhé et al., 1995; Brimecombe et al., 1997). In
contrast, the single-channel properties of NR1/NR2C (Stern et al.,
1992) and NR1/NR2D (Wyllie et al., 1996) receptors differ considerably
from those of cultured cortical neurons. Furthermore, mRNA for NR1,
NR2A, and NR2B, but not NR2C, is found in cultured cortical neurons
(Zhong et al., 1994). Therefore, to further our understanding
Mgi2+ inhibition of the NMDA receptors in cortical
neurons, we studied recombinant NR1/NR2A and NR2/NR2B receptors.
Control I-V Relation of Recombinant NMDA Receptors Expressed in CHO Cells. In the absence of Mgi2+, the outward rectification previously observed in cultured cortical neurons also is exhibited by NR1/NR2A and NR1/NR2B receptors expressed in CHO cells (Fig. 1). However, the magnitude of the rectification is smaller in recombinant NMDA receptors of either subunit composition than in native NMDA receptors. This difference could result from neuron-specific molecules that influence the degree of NMDA receptor current rectification. The lack of this factor in CHO cells would result in less I-V relation nonlinearity of recombinant NMDA receptors. Alternatively, post-translational modifications of NMDA receptors that affect current rectification may differ in neurons and CHO cells.
In neurons, control (0 pipette Mg2+) mean patch currents rectify more strongly than do whole-cell currents. The conclusion that this difference is not due to residual intracellular Mg2+ during whole-cell recording with Mg2+-free pipette solutions (Li-Smerin and Johnson, 1996b) is supported by the even weaker rectification observed
herein of control CHO whole-cell currents. Patch excision may induce a
change in NMDA receptor properties that results in an increase in the
voltage dependence of channel open probability (Nowak and Wright,
1992).
Inhibition by Mgi2+ of Recombinant NMDA Receptors Expressed in CHO Cells. Whole-cell NMDA-activated currents of transfected CHO cells were reduced at positive potentials by Mgi2+. The magnitude of the voltage-dependent inhibition by Mgi2+ of recombinant NR1/NR2A or NR1/NR2B receptors were similar to each other and to inhibition of native NMDA receptors expressed in cultured neurons (Fig. 3). The characteristics of Mgi2+ inhibition of whole-cell currents thus appear to be independent of the type of cell in which NMDA receptors are expressed.
Subunits NR2A and NR2B of NMDA receptors have identical amino acid sequences in the region (M2) that has been proposed to line the channel (McBain and Mayer, 1994). The effects of Mgi2+ on the
single-channel current of NMDA receptors are influenced by residues
located in the M2 region of NR1/NR2A receptors (Kupper et al., 1996,
1998; Wollmuth et al., 1998). The similar effects of
Mgi2+ on NR1/NR2A and NR1/NR2B receptors presented here
further support the notion that these effects result from binding of
Mgi2+ to the pore-forming region.
The nonselective effects of Mgi2+ on NR1/NR2A and
NR1/NR2B receptors suggests that Mgi2+ has a
homogeneous effect on native NMDA receptors in preparations in which
only these subunits are expressed. We did not examine the effects of
Mgi2+ on CHO cells transfected with NR1, NR2A, and NR2B
receptors; we therefore cannot exclude the possibility that native
neurons express NR1/NR2A/NR2B receptors, and that Mgi2+
has a distinct effect on such receptors. There is considerable evidence
that NR1/NR2A/NR2B receptors can assemble and that their properties are
intermediate between NR1/NR2A and NR1/NR2B receptors (Sheng et al.,
1994; Brimecombe et al., 1997; Luo et al., 1997; Vicini et al., 1998;
Hawkins et al., 1999; Tovar and Westbrook, 1999). For
Mgi2+, we observed indistinguishable effects on
NR1/NR2A, NR1/NR2B, and native receptors. Thus, it is likely that the
effects of Mgi2+ on NR1/NR2A/NR2B receptors would be
similar to its effects on NR1/NR2A and NR1/NR2B receptors. Models in
which Mgi2+ is assumed to interact with a homogeneous
population of receptors, as used in Li-Smerin and Johnson (1996b),
should be applicable to cortical neurons.
Differential Effects of Mgi2+ on Steady-State
Responses of Whole Cells and Patches.
In cultured cortical
neurons, Mgi2+ inhibits whole-cell NMDA-activated
currents less effectively than currents measured in patches. Mgi2+ inhibits single-channel currents recorded from
cultured cortical neurons (Johnson and Ascher, 1990; Li-Smerin and
Johnson, 1996a) with relatively high affinity. Mgi2+
inhibition of patch currents is somewhat less effective (~1.4-fold lower affinity) because Mgi2+ binding affects channel
gating of NMDA receptors (Li-Smerin and Johnson, 1996b). The large and
surprising preparation dependence of Mgi2+ action was
observed when inhibition of NMDA-activated whole-cell and mean patch
currents from cultured neurons were compared (Li-Smerin and Johnson,
1996b). This difference implied that Mgi2+ action
differed in the two recording configurations. Interpretation of these
data was hindered, however, by a possible lack of control of
[Mg 2+]i or membrane potential in
neuronal whole-cell measurements. No comparison of
Mgi2+ inhibition of whole-cell and patch responses have
previously been made in any other system.
). However, previous work in heterologous
expression systems has demonstrated that the effect of
Mgi2+ on single-channel currents of NMDA receptors is
not preparation-dependent. Inhibition by Mgi2+ of
single-channel currents of NR1/NR2A receptors expressed in Xenopus oocytes (Kupper et al., 1998; Wollmuth et al., 1998)
and in human embryonic kidney-293 cells (Wollmuth et al., 1998) is quantitatively very close to inhibition of single-channel currents in
cultured cortical neurons. Mgi2+binding affects the
gating of some mutant NMDA receptors expressed in Xenopus
oocytes (Kupper et al., 1998), but no evidence for an effect on channel
gating was observed in wild-type receptors expressed in oocytes (Kupper
et al., 1998; Wollmuth et al., 1998). Thus, our observation herein of
nearly identical inhibition by Mgi2+ of whole-cell
currents from transfected CHO cells and from neurons (Fig. 3) indicates
that the effects of patch excision are not exclusive to neuronal preparations.
One possible explanation for the patch versus whole-cell difference
observed in neurons was inadequate control of voltage and/or
[Mg2+]i in dendrites
during whole-cell recordings, which could cause underestimation of the
magnitude of inhibition. If this were the case, then the magnitude of
Mgi2+ inhibition in CHO whole-cell recordings should
resemble inhibition in excised patches but should differ from
inhibition in whole neurons. Our data fully contradict this prediction.
Therefore, control of dendritic voltage and
[Mg2+]i in whole-cell
experiments was adequate to permit accurate measurement of
Mgi2+ inhibition.
The differential effects of Mgi2+ on patch- and
whole-cell currents also could have resulted from a difference in the
Mgi2+ sensitivity of NR1/NR2A and NR1/NR2B receptors;
if one receptor subtype were enriched at the neuron soma, then patch-
and whole-cell experiments would sample NMDA receptor subtypes with
different Mgi2+ sensitivity. This possibility can be
rejected because there was no difference between the effect of
Mgi2+ on NR1/NR2A and on NR1/NR2B receptors.
Thus, in heterologous expression systems as well as in neurons,
Mgi2+ inhibits whole-cell NMDA responses less
effectively than patch responses. The differential effects of
Mgi2+ cannot be explained by differences between
configurations in control of experimental conditions, nor by
heterogeneous effects of Mgi2+ on NMDA receptor
subtypes. Our results suggest that patch excision modifies NMDA
receptors in a way that weakens their sensitivity to inhibition by
Mgi2+. Several other properties of NMDA receptors are
modified by patch excision, including glycine-insensitive
desensitization (Sather et al., 1990), mechanosensitivity (Paoletti and
Ascher, 1994), channel open probability (Rosenmund et al., 1995), and
single-channel conductance (Clark et al., 1997). It is possible that
some or all of these changes result from disruption of an interaction between NMDA receptors and a cytoplasmic protein. NMDA receptors are
known to be capable of binding to many cytoplasmic proteins, including
PSD-95 (postsynaptic density-95) and related proteins that contain the
PDZ domain (a protein binding domain first found in PSD-95,
Discs large, and ZO-1), 
-actinin, yotiao,
neuronal intermediate filaments, calmodulin, and phospholipase C-
(Gurd and Bissoon 1997; Sheng and Wyszynski, 1997; O'Brien et al.,
1998). In some cases, the interactions can affect NMDA receptor
activity (Krupp et al., 1999; Yamada et al., 1999). If patch excision
does affect Mgi2+ inhibition through disruption of
interactions between NMDA receptors and intracellular proteins, the
data presented herein suggest that the cytoplasmic protein is
ubiquitous and can interact with both NR1/NR2A and NR1/NR2B receptors.
| |
Acknowledgments |
|---|
We thank Jessica C. Brimecombe and Faye A. Boeckman for preparation of and advice on the use of transfected CHO cells. We also thank William Potthoff and Karen Hartnett for excellent technical support, Dr. S. Nakanishi for NR1-1a cDNA, Dr. P. Seeburg for NR2A and NR2B cDNAs, and Dr. M. Chalfie for GFP cDNA.
| |
Footnotes |
|---|
Accepted for publication November 28, 1999.
Received for publication September 9, 1999.
1 This work was supported by National Institutes of Health Grants MH45817 and MH00944 (to J.W.J.) and NS29365 (to E.A.), and by National Institute of Mental Health Training Grant T32 MH18273 (to Y.L.-S.).
2 Current address: Molecular Physiology and Biophysics Unit, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, MD 20892.
Send reprint requests to: Jon W. Johnson, Department of Neuroscience, 446 Crawford Hall, University of Pittsburgh, Pittsburgh, PA 15260. E-mail: johnson{at}bns.pitt.edu
| |
Abbreviations |
|---|
NMDA, N-methyl-D-aspartate; CHO, Chinese hamster ovary; GFP, green fluorescent protein; FC, fractional current; I-V, current-voltage.
| |
References |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
.
J Neurochem
69:
623-630[Medline].-actinin with the NR1 subunit modulate Ca2+-dependent inactivation of NMDA receptors.
J Neurosci
19:
1165-1178
2/
1-subtype N-methyl-D-aspartate receptor by PSD-95.
J Biol Chem
274:
6647-6652This article has been cited by other articles:
|
|
 |
|
  T. A. Blanpied, R. J. Clarke, and J. W. Johnson Amantadine Inhibits NMDA Receptors by Accelerating Channel Closure during Channel Block J. Neurosci., March 30, 2005; 25(13): 3312 - 3322. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Li-Smerin, E. S Levitan, and J. W Johnson Free intracellular Mg2+ concentration and inhibition of NMDA responses in cultured rat neurons J. Physiol., June 15, 2001; 533(3): 729 - 743. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
|
 |
 |
 |
) and 6.2 mM
(
) Mgi2+ from NR1/NR2A (B) or NR1/NR2B (D)
receptors. The amplitude of each current measurement was normalized to
current amplitude measured at 
, different from corresponding
NR1/NR2B current; 
, different from corresponding nWCC current.
Symbols above the solid wide column refer to statistical differences
between currents recorded in 0 Mgi2+; symbol within the
open narrow column refers to statistical difference between currents
recorded in 6.2 mM Mgi2+. There was no significant
difference between NR1/NR2A and NR1/NR2B in the current recorded in the
absence or presence of Mgi2+ at any voltage.
) or NR1/NR2B (
) and mean patch current (
) were calculated from data
in Li-Smerin and Johnson (1996b)