1Laboratory for Memory and Learning, Brain Science Institute, Institute of Physical and Chemical Research (RIKEN), Saitama 351-0198; and 2Laboratory for Speciation Mechanisms I, National Institute for Basic Biology, Okazaki 444-8585, Japan
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Karachot, Laddawan,
Yoshinori Shirai,
Réjan Vigot,
Tetsuo Yamamori, and
Masao Ito.
Induction of Long-Term Depression in Cerebellar Purkinje Cells
Requires a Rapidly Turned Over Protein.
J. Neurophysiol. 86: 280-289, 2001.
Evidence is presented
indicating that the induction of long-term depression (LTD) in Purkinje
cells (PCs) requires a rapidly turned over protein(s) during a critical
time period within 15 min after the onset of LTD-inducing stimulation
and that synthesis of this protein is maintained by mRNAs supplied via
transcription. LTD was induced in granule cell axon (GA)-to-PC synapses
by stimulation of these synapses at 1 Hz for 5 min in conjunction with
the climbing fibers (CFs) forming synapses on the same PCs and
represented by a persistent reduction in the GA-induced excitatory
postsynaptic potentials (EPSPs). Not only a prolonged but also a brief
(5 min) pulse application of translational inhibitors (anisomycin,
puromycin, or cycloheximide) effectively blocked the LTD induction.
Pulses applied during the period from 30 min before to 10 min after the onset of conjunctive stimulation blocked the LTD induction, but those
applied 15 min after were ineffective. The three translational inhibitors blocked the LTD induction similarly, suggesting that the
effect is due to their common action of inhibiting protein synthesis.
Infusion of a mRNA cap analogue (7-methyl GTP) into PCs also blocked
LTD induction, ensuring that the postsynaptic protein synthesis within
PCs is required for LTD induction. Transcriptional inhibitors,
actinomycin D and
5,6-dichloro-l--D-ribofuranosyl-benzimidazole, also
blocked the LTD induction, but this effect was apparent when 5-min
pulses of the transcriptional inhibitors preceded the conjunctive stimulation by 30 min or more. This time lag of 30 min is presumed to
be required for depletion of the protein(s) required for LTD induction.
The presently observed effects of translational and transcriptional
inhibitors on the LTD induction are of temporal characteristics
corresponding to their depressant effects on the type-1 metabotropic
glutamate-receptor (mGluR1)-mediated slow EPSPs in PCs as we have
reported recently. An antagonist of mGluR1s [(RS)-1-aminoindan-1,5-dicarboxylic acid], however, did not block LTD
induction when it was applied during the 10-min period following conjunctive stimulation, where translational inhibitors effectively blocked LTD induction. This discrepancy in time course suggests that
the rapidly turned over protein(s) required for LTD induction is
involved in a process occurring downstream of the activation of mGluR1s.
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Roles of protein synthesis in
synaptic plasticity have been proposed based on the findings that
translational and transcriptional inhibitors inhibit long-term
potentiation (LTP) in hippocampal neurons (Frey and Morris
1997; Frey et al. 1988
; Nguyen et al. 1994
) and sensitization in Aplysia sensory neurons
(Montarolo et al. 1986
). Translational inhibitors also
inhibit the neurotrophin-induced LTP in hippocampal neurons
(Kang and Schuman 1996
) and metabotropic glutamate
receptor (mGluR)-mediated long-term depression (LTD) in hippocampal
neurons (Huber et al. 2000
). Protein synthesis and
processing occur in ribosomes, the endoplasmic reticulum, and the Golgi
complex, which are largely contained in neuronal somata. Recently,
however, neuronal dendrites have drawn attention as local sites of
protein synthesis and processing specific to synapses since dendrites
contain various mRNAs (Gao 1998
; Martone et al.
1998
; Steward 1995
; Tian et al.
1999
) as well as ribosomes, polyribosomes, the endoplasmic
reticulum, and part of the Golgi complex (Gardiol et al.
1999
; Steward and Reeves 1988
; Steward et
al. 1996
). Dendrites exhibit an increased expression level of
certain proteins or their mRNAs following the induction of LTP in
hippocampal and neocortical neurons (Klintsova and Greenough 1999
; Osten et al. 1996
; Roberts et al.
1998
; Schuman 1999
).
In cerebellar Purkinje cells (PCs), LTD is induced following
conjunctive activation of two excitatory inputs to PCs, namely, granule
cell axons (GAs) and climbing fibers (CFs) (Ekerot and Kano
1985; Ito and Kano 1982
; Karachot et al.
1994
; Sakurai 1987
). The so-induced LTD is
represented by persistent depression in the GA-evoked excitatory
postsynaptic potentials (EPSPs) or currents (EPSCs). Inducing a
persistent reduction in glutamate sensitivity of cultured PCs by
conjunctive application of glutamate and membrane depolarization,
Linden (1996)
reported a depressant action of translational and transcriptional inhibitors on this reduced form of
LTD. In the present study, we confirmed that translational inhibitors
effectively blocked LTD in cerebellar slices, and the results were
interesting in two respects. First, to block protein synthesis, neural
tissues are usually incubated with translational inhibitors for 1-3 h
(Frey et al. 1988
; Linden 1996
;
Montarolo et al. 1986
); however, a brief 5-min pulse
application of translational inhibitors was found to be sufficient to
block the induction of LTD. Second, translational inhibitors abolished
the entire LTD including its early phase, contrary to the general
presumption that protein synthesis is required for the late phase of
synaptic plasticity but not for the early phase (Frey and Morris
1997
; Frey et al. 1988
; Linden
1996
; Montarolo et al. 1986
; Nguyen et al. 1994
).
We have recently reported that 5-min pulse applications of
translational or transcriptional inhibitors effectively suppress slow
EPSPs mediated by type 1 metabotropic glutamate receptors (mGluR1-EPSPs) in PCs (Karachot et al. 2000). Since
activation of mGluR1s is known to be required for the induction of LTD
(Aiba et al. 1994
; Conquet et al. 1994
;
Hartell 1994
; Shigemoto et al. 1994
), it
is possible that the protein required for the generation of
mGluR1-EPSPs is a key factor responsible for the unique protein synthesis dependence of LTD. To test this possibility, we measured the
time course of the depressant effect of translational and transcriptional inhibitors on LTD by applying 5-min inhibitor pulses
and found that it is consistent with the previously reported time
course of their effects on mGluR1-EPSPs (Karachot et al. 2000
). However, we also found that the effect of translational inhibitors of depressing LTD induction outlasted the inactivation of
mGluR1s by an antagonist. This suggests that the rapidly turned over
protein synthesis is not directly involved in activation of mGluR1s but
probably in a process occurring downstream of the activation of mGluR1s.
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Sagittal slices (0.4-mm thick) were obtained from Wistar rat
cerebella dissected under ether anesthesia. Adult (postnatal week 6)
rats were used for intracellular recording from PCs and young rats
(postnatal week 3-4) were preferred for whole cell patch-clamp
recording. The standard perfusate (pH = 7.3) containing (in mM)
118 NaCl, 4.7 KCl, 2.5 CaCl2, 25.0 NaHCO3, 1.18 KH2PO4, 1.19 MgSO4, and 11.0 glucose plus 20 µM picrotoxin
was equilibrated with 95% O2-5%
CO2 gas and was warmed to maintain the
temperature of the perfusion chamber at 31-32°C. GAs were bipolarly
stimulated near the pial surface at an intensity sufficient to induce
EPSPs of 8-10 mV (Fig. 1A).
CFs were bipolarly stimulated in the white matter at an intensity
slightly higher than the threshold for evoking CF responses from each
impaled PC (Fig. 1C). In most experiments, PCs were impaled
at their dendritic shafts using glass microelectrodes (60-80 M)
filled with 3 M KCl solution. In the experiments illustrated in Fig. 6,
EPSPs were recorded with a whole cell configuration current-clamp using
a whole cell clamp pipette (5-8 M
) containing (in mM) 134 K-gluconate, 6 KCl, 4 NaCl, 10 HEPES, 0.2 EGTA, 4 MgATP, 0.3 Tris-GTP,
and 14 phosphocreatine; the pH was adjusted to 7.25 with KOH. In part
of these experiments, the pipette also contained 250 µ M 7-methyl GTP
(m7GpppG), a mRNA cap analogue (Chu and
Rhoades 1980
; Huber et al. 2000
).
|
To observe LTD, fast EPSPs mediated by
-amino-3-hydroxy-5-methyl-4-isoxazolone propionate (AMPA) receptors
were evoked by stimulating GAs every 5 s, and five successive
measurements were averaged. In Figs. 2, 3, 6-8, and 10, the initial
slope of the GA-evoked EPSPs is plotted at 1-min intervals. Their 100%
values were determined by averaging the five measurements during the
5-min period prior to conjunctive stimulation. No plot was made during
conjunctive stimulation and the succeeding 1-min period to avoid the
effect of the transitional change of the GA-stimulating conditions from the 0.2-Hz-test stimulation to 1-Hz conjunctive stimulation and vice
versa. LTD was induced using the protocol of Karachot et al.
(1994)
of repetitive simultaneous stimulation of GAs and CFs at
1 Hz for 5 min (300 pulses). ANOVA was adopted for statistical analysis, and Duncan's new multiple range test was used for comparison between data obtained with different timings of conjunction and drug applications.
To induce mGluR1-EPSPs, brief tetanic stimulations (50 Hz, 8 pulses at
30-s intervals) were applied to GAs bipolarly near the pial surface in
the presence of 10 µM 2,3-dioxo-6-nitro-1,2,3,4 tetrahydrobenzo[f]quinoxaline-7-sulfonamide (NBQX), an antagonist for
AMPA-selective glutamate receptors. An antagonist of
N-methyl-D-aspartate (NMDA)-selective receptors,
namely, 30 µM D()-2-amino-5-phosphonopentanoic acid
(D-AP-5), was also added routinely. Even though adult PCs themselves do not form NMDA receptors, D-AP-5 eliminated a
possible NMDA receptor-mediated activation of granule cells due to
spread of stimuli to mossy fibers. The mGluR1-EPSPs thus obtained were confirmed to be blocked by 500 µM (RS)-a-methyl-4-carboxyphenyl glycine (MCPG), a nonspecific antagonist of mGluRs. A mGluR1-specific antagonist, (RS)-1-aminoindan-1,5-dicarboxylic acid (AIDA)
(Bachelor et al. 1997
; Pellicciari et al.
1995
) was used for the purpose of testing the involvement of
mGluR1-EPSPs in LTD induction (Fig. 10).
7-Hydroxyiminocyclopropan[b]chromen-1a-carboxylate ethyl ester (CPCCOEt) (Pagano et al. 2000
) was also tested, but DMSO
used to dissolve it showed by itself a depressant effect, gradually developing in 10-20 min, on mGluR1-EPSPs even at a concentration as
low as 0.01-0.05%; CPCCOEt was not used further for testing LTD.
NBQX, D-AP-5, and AIDA were purchased from Tocris, and
other chemicals were from Sigma. The concentrations of translational and transcriptional inhibitors in the perfusates were selected according to the results of our previous study (Karachot et al. 2000): anisomycin, 25 µM; puromycin, 50 µM; cycloheximide,
30 µM, actinomycin D, 25 µM;
5,6-dichloro-1-
-D-ribofuranosylbenzimidazole (DRB), 30 µM. These were dissolved in the perfusates, except for DRB whose
dissolution required sonication. AIDA was dissolved in concentrations
of 200-650 µM, and pH was adjusted to 7.3 with NaOH.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
In recording from PCs, LTD was characterized by a reduction in the
rising slope of the GA-evoked EPSP, which followed conjunctive stimulation, as shown in Fig. 1, A and B. That no
significant change occurred in CF responses was taken as evidence
indicating that the reduction in GA-evoked EPSPs was not due to
deterioration of the examined PCs (Fig. 1, C and
D). The plot of the magnitude of the rising slope of
GA-evoked EPSPs in Fig. 2 (open circles) shows a gradual development of LTD over 60 min. To determine the magnitude of LTD, we measured the average reduction in the EPSP slopes
at two phases: 7-16 min (early phase) and 35-44 min (late phase)
after the onset of conjunctive stimulation (Fig. 2, i and ii). In 15 control PCs examined in this study, the magnitude of LTD was 20.1 ± 1.0% at 7-16 min and 37.8 ± 1.2% at 35-44 min (means and SE). The
latter values are comparable with those reported previously
(Karachot et al. 1994; Miyata et al.
1999
).
|
Effects of translational inhibitors
We used three translational inhibitors, namely, anisomycin
(Grollman 1968), puromycin (Stanton and Sarvey
1984
), and cycloheximide (Deadwyler et al. 1987
;
Stanton and Sarvey 1984
). Emetine (Bennett et al.
1964
) was also tested, but since it induced sustained
depression of GA-evoked EPSPs by itself, its effect on LTD was
difficult to evaluate; the data obtained with emetine were therefore
excluded. The translational inhibitors presently used did not affect
the membrane resting potential, input resistance, GA-evoked EPSPs, or
CF responses of PCs, but they exerted a clear inhibitory effect on LTD.
Figure 1, E-H, similarly shows GA-evoked EPSPs and CF responses under continuous perfusion with 25 µM anisomycin that started no less than 30 min before conjunctive stimulation, and GA-evoked EPSPs no longer exhibited LTD. As illustrated in Fig. 2
(closed circle), LTD was abolished at both the early (3.6 ± 1.3%) and late phases (1.4 ± 1.4%, 5 PCs) after conjunctive stimulation.
One of the findings in this study is that even 5-min pulse applications
of anisomycin effectively block LTD when applied simultaneous to
conjunctive stimulation (Fig.
3A, GA-evoked EPSPs at the
late phase: 0.2 ± 3.1%, 6 PCs). Blockade of LTD occurred also when 5-min applications of anisomycin preceded conjunction by 10 min (Fig.
3B, 10.1 ± 4.1%, 9 PCs) or even 30 min (Fig.
3C, 15.1 ± 3.1%, 7 PCs). Blockade of LTD also occurred
even when 5-min anisomycin pulses were delayed by 5 min (Fig.
3D, 9.0 ± 4.7%, 8 PCs) or 10 min (Fig. 3E, 19.3 ± 5.4%, 8 PCs) after the onset of conjunction. However, when 5-min
pulse applications of anisomycin were delayed by 15 min, LTD
occurred normally (Fig. 3F, 41.0 ± 5.9%, 5 PCs; P > 0.05 as compared with that of the control).
|
Figure 4A summarizes similar measurements obtained using the three translational inhibitors on the late phase of LTD at 35-44 min after the onset of conjunction. It is noted that these inhibitors suppressed the magnitude of LTD statistically significantly (P < 0.01) when 5-min pulses of the translational inhibitors were applied 20 min before to 10 min after the onset of conjunctive stimulation. However, cycloheximide applied 30 min before conjunction was ineffective (P > 0.05), whereas anisomycin similarly applied was still effective. Figure 4B summarizes the measurements on the early phase of LTD. Differing from the inhibition of the late phase, the inhibition of the early phase of LTD was insignificant (P > 0.05) when translational inhibitors were applied 20-30 min before (Fig. 3C) or 15 min after (Fig. 3F) the onset of conjunction. The reason why translational inhibitors applied 10 min after the onset of conjunction reduced the magnitude of the early phase of LTD to a lesser degree than that of the late phase (compare Fig. 4, A and B) is presumably because the early phase of LTD has already developed before it was aborted by translational inhibitors. The difference obtained between Fig. 4, A and B, when translational inhibitors were applied 20 or 30 min before conjunction, however, would suggest that the early phase of LTD is less sensitive to translational inhibitors as compared with the late phases (DISCUSSION).
|
When applied 10 min after the onset of conjunction, the inhibitors acted after the very initial development of LTD and aborted it (Fig. 3E). Events at this critical timing are illustrated in Fig. 5 with cycloheximide. Cycloheximide pulses applied 10 min, but not 15 min, after the onset of conjunctive stimulation effectively aborted the LTD induction. These measurements in Figs. 3-5 using 5-min pulse applications of translational inhibitors indicate that protein synthesis is required critically during the period within 15 min after the onset of 5-min conjunctive stimulation.
|
While anisomycin is widely used as a protein synthesis inhibitor, it is
known also to activate 45- and 55-kDa protein kinases (p45/p55) at a
concentration as low as 0.037 µM (10 ng/ml) (Cano et al.
1994). However, 5-min pulses of 0.1 µM anisomycin applied simultaneously with conjunctive stimulation did not inhibit the induction of LTD (LTD at 35-44 min: 29.1 + 2.6%, 4 PCs,
P > 0.05 compared with the 15 control samples shown in
Fig. 2). Anisomycin is also known to activate the 70- and 85-kDa S6
protein kinases (p70/85S6k), which are homologous
of the mitogen-activated kinase-activated protein kinase-2 (MAPKAPK-2)
(Cano et al. 1994
, 1996
). However, 25 nM rapamycin that
blocks the anisomycin activation of p70/85S6k
(Cano et al. 1994
) did not alter the blocking effect on
the LTD induction of the coapplied 25 µM anisomycin simultaneously
with conjunctive stimulation (LTD at 35-44 min: 5.2 + 1.6%, 3 PCs, P > 0.05 compared with 6 samples
subjected to 5-min 25 µM anisomycin pulses alone).
Effects of a mRNA cap analogue
The mRNA cap analogue (m7GpppG) inhibits
translation by competing with the endogenous capped mRNA for the
cap-binding brain elF-4E (Gingras et al. 1999). Figure
6 illustrates that, without loading of
m7GpppG in the whole cell clamp pipette, LTD
developed as observed in intracellular recording shown in Fig. 2. The
magnitudes of LTD at 7-16 and 35-44 min after the onset of
conjunctive stimulation are 16.1 ± 3.8% (7 PCs) and 37.6 ± 3.2% (5 PCs), respectively, which are not significantly (P > 0.05) different from those obtained in the intracellularly recorded 15 control PCs (20.1 ± 1.0% at 7-16 min and 37.8 ± 1.2% at
35-44 min). When the pipette was loaded with
m7GpppG, LTD was significantly depressed
(P < 0.01) in both early and late phases, the
magnitudes of LTD at 7-16 min being
1.0 ± 2.2% (4 PCs) and that at
35-44 min being
11.2 ± 3.9% (3 PCs). Hence,
m7GpppG infused through the whole cell clamp
pipette blocked LTD similarly to translational inhibitors.
|
Effects of transcriptional inhibitors
We used two types of transcriptional inhibitors, actinomycin D and
DRB. Perfusion of 25 µM actinomycin D for 5 min immediately decreased
the amplitudes of GA-evoked AMPA-EPSPs (see the downward shift of shown in Fig. 7). This effect is too
sudden to be ascribed to transcriptional inhibition; it presumably
arose from a side effect of actinomycin D, which may have a direct
effect on GA-to-PC transmission. However, this transient effect of
actinomycin D did not affect our observation of LTD (Fig. 7,
). The
magnitude of LTD measured 35-44 min after the onset of conjunctive
stimulation, immediately succeeding 5-min pulse applications of
actinomycin D (Fig. 7), was 31.2 ± 2.4% (9 PCs), which was not
significantly different from that of the control LTD (37.8%,
P > 0.05). LTD was suppressed only when 5-min pulse
applications of actinomycin D preceded the conjunctive stimulation by
30 min or more (Fig. 8,
). DRB (30 µM) also blocked the LTD induction after a delay of 30 min (Fig. 8,
). It is notable that DRB did not induce an immediate depression of
the GA-evoked EPSPs during its application such as that caused by
actinomycin D. It can be observed from Fig. 8 that actinomycin D and
DRB blocked the entire LTD including its early phase. Figure
9 summarizes these observations and shows that conjunctive stimulation at 30-60 min after 5-min pulse
applications of either actinomycin D or DRB induced significantly
(P < 0.01) smaller LTD magnitude as compared with
control LTD.
|
|
|
Effect of mGluR1 antagonist
We reported previously that a rapidly turned over protein plays a
key role in the metabotropic glutamatergic transmission that produces
characteristic slow mGluR1-EPSPs in PCs (Karachot et al.
2000). Since an antagonist (Hartell 1994
), an
antibody (Shigemoto et al. 1994
), and gene deficiency
(Aiba et al. 1994
; Conquet et al. 1994
)
against mGluR1s block LTD induction, it is possible that translational and transcriptional inhibitors block LTD
induction by interfering with mGluR1s. We, therefore compared the
effect of a mGluR1 antagonist and protein synthesis inhibitors on LTD induction.
Five-minute pulse applications of AIDA at 500 µM, but not at 200-300
µM, rapidly suppressed the GA-evoked mGluR1-EPSPs, down to 30%
within 2 min, as shown in Fig.
10A. Applications of AIDA simultaneous to conjunctive stimulation blocked LTD and instead induced
potentiation of GA-induced AMPA-EPSPs as described previously (Hartell 1994) (Fig. 10B). However, 5-min
pulse applications of AIDA shortly following conjunctive stimulation
did not interfere with LTD induction as shown in Fig. 10, C
and D. This contrasts to the effect of translational
inhibitors, which blocked LTD induction even when applied 10 min after
the onset of conjunctive stimulation (Figs. 3E and 5). The
discrepancy suggests that the effect of translational inhibitors is not
due to direct inactivation of mGluR1 receptors
(DISCUSSION). A transient downward shift of the plotted
points is noted in Fig. 10, C and D, during 5-min
pulse application of AIDA. This may be due to a reversible effect of AIDA on GA-mediated AMPA-EPSPs, which, however, did not influence the
present observation of the much longer lasting LTD.
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Specificity of protein synthesis inhibitors
All the three types of translational inhibitors blocked the
induction of LTD rapidly and persistently (Fig. 4). We previously reported that these translational inhibitors rapidly and persistently suppressed amino acid incorporation into cerebellar slices
(Karachot et al. 2000). This general agreement in these
two sets of data supports the view that protein synthesis is required
for the induction of LTD. However, a discrepancy was noted in that
translational inhibitors applied 20 or 30 min before the onset of
conjunction did no reduce the magnitude of the early phase of LTD at
7-16 min (Fig. 4B), even while methionine incorporation
into cerebellar slices continues to be inhibited for 30 min after 5-min
pulse applications of translational inhibitors (Karachot et al.
2000
). The discrepancy is smaller for the effect of
translational inhibitors on the late phase of LTD at 35-44 min; yet,
cycloheximide pulses applied 30 min after the onset of conjunction did
not reduce significantly the late of phase of LTD (Fig. 4A).
These observations suggest that the concentration of translational
inhibitors gradually declines within 30 min after their pulse
applications even though this recovery is not reflected in the
methionine incorporation into cerebellar slices measured by
Karachot et al. (2000)
. The discrepancy may not be
surprising since the present LTD measurements reflect events only in
PCs, while the methionine incorporation into cerebellar slices involves
protein synthesis in not only PCs but also in other types of cerebellar neurons.
Translational inhibitors activate certain protein kinases independent
of protein synthesis inhibition (Cano et al. 1994). However, all the three types of kinases known to be activated by
anisomycin [p45/p55, p70/85S6k, and c-Jun
NH2-terminal kinases/stress-activated protein
kinases (JNK/SAPKs)] can be excluded as a cause of the blockade of LTD for the following reasons. Anisomycin did not block the LTD induction at a concentration of 0.1 µM (see RESULTS), which is
subthreshold for inhibiting protein synthesis yet effective in
activating p45/p55 (Cano et al. 1994
). The report that
puromycin is ineffective in activating P45/p55 (Cano et al.
1994
) justifies the exclusion of p45/p55 because the LTD
induction is blocked by puromycin similarly to anisomycin (Fig. 4).
p70/85S6k kinases are unlikely to be involved in
the LTD induction because rapamycin, an inhibitor of anisomycin-induced
activation of p75/80S7k, did not prevent the
anisomycin-induced blockade of LTD induction (see RESULTS).
And finally, responses of JNK/SAPKs to application of
anisomycin become distinct after 15 min (Cano et al.
1995
), being too slow to account for the rapid
anisomycin-induced inhibition of LTD induction (Fig. 4). While
phosphorylation of JNK/SAPKs proceeds within 5 min of
10-µM-anisomycin application, it takes 60 min with
10-µM-cycloheximide application, and only slight phosphorylation occurs within 60 min of 10-µM-puromycin application (Sidhu and Omiecinski 1998
), unlike the uniformly rapid inhibitory action of the three translational inhibitors on LTD induction (Fig. 4).
Actinomycin D blocks transcription by intercalating DNAs (Sobell
1973), while DRB inhibits mRNA synthesis by blocking the initiation of RNA chain elongation (Tamm and Sehgal
1978
). The close similarity of the effects on the LTD induction
in spite of the distinct mechanisms of action of the two
transcriptional inhibitors supports the view that their common action
of inhibiting mRNA synthesis underlies the blockade of LTD induction.
The 30-min delay required for the action of both actinomycin D and DRB
can be accounted as the time required for the gradual degradation of
preexisting dendritic mRNAs to a level lower than that required for LTD
induction (see following text).
Comparison with other forms of synaptic plasticity
Protein synthesis inhibitors affect the late phase, but not the
early phase, of LTP (Frey et al. 1988; Nguyen et
al. 1994
) or LTD (Kauderer and Kandel 2000
) in
hippocampal neurons or sensitization in Aplysia
(Montarolo et al. 1986
). When a translational inhibitor was continuously applied to cultured PCs immediately after conjunctive stimulation, LTD started to decline 45 min after the onset of the
glutamate/depolarization conjunctive stimulation (Linden
1996
). The phase of LTD within 45 min after the onset of
conjunctive stimulation in cultured PCs is thus independent of protein
synthesis in contrast to the present results in PCs of cerebellar
slices that both of the early and late phases of LTD are dependent on protein synthesis. However, the present results obtained in PCs of
cerebellar slices are in accordance with the reports on the neurotrophin-induced LTP (Kang and Schuman 1996
) and
mGluR5-mediated LTD (Huber et al. 2000
) in hippocampal
neurons, which are inhibited by translational inhibitors in both its
early and late phases.
To explain the different effects of translational inhibitors on PCs in
cerebellar slices and cultures, a question may arise as to whether
translational inhibitors act on GAs, presynaptic to PCs in cerebellar
slices, which are absent in cultured PCs used by Linden
(1996). However, as far as we assume that the rapid protein
synthesis required for LTD induction is the same as that required for
the generation of mGluR1-EPSPs (Karachot et al. 2000
), it must occur within PCs because translational inhibitors do not affect
GA-mediated AMPA-EPSPs and because the characteristic delay of action
of transcriptional inhibitors on LTD induction is in accordance with
the slow decline of mGluR1-EPSPs (see following text). The present
result that a mRNA cap analogue infused into PCs also blocked LTD
induction (Fig. 6) strongly supports the view that postsynaptic protein
synthesis is required for LTD induction.
There are three factors to be considered in relation to the different
protein synthesis dependencies of PCs in cerebellar slices and
cultures. 1) The LTD-inducing stimulating conditions are
considerably different between the two types of PC preparations. In
cerebellar slices, LTD induction requires application of 300 pulses to
GAs and CFs, while two to six conjunctive stimulations by a glutamate
pulse and a 3-s membrane depolarization to 0 mV are sufficient in
cultured PCs. 2) The time courses of LTD are also
considerably different. LTD in cerebellar slices develops slowly over
1 h (Fig. 2), while the reduced form of LTD in cultured PCs
develops rapidly to reach a plateau within a few minutes (Linden 1996). And 3) signal transduction processes
underlying LTD also show certain differences. It is known that LTD
induction in tissue cultures does not require either nitric oxide
(Linden et al. 1995
) or inositol trisphosphate
(Narasimhan et al. 1998
), which are required for LTD
induction in cerebellar slices (Boxall and Garthwaite 1996
; Inoue et al. 1998
; Shibuki and
Okada 1991
). The corticotropin-releasing factor (CRF) contained
in CFs, which is indispensable for LTD induction in cerebellar slices
(Miyata et al. 1999
), could be absent in tissue cultures
lacking CFs. Thus the two forms of LTD, which are induced under
different stimulation conditions and which developed with different
time courses, could be underlain by signal transduction processes to
which protein synthesis contributes differently.
The protein-synthesis-dependency, however, may not uniform between the early and late phases of LTD in PCs of cerebellar slices because when translational inhibitors were applied 20-30 min before the onset of conjunction, the late phase of LTD was in most cases significantly depressed (Fig. 4A), whereas the early phase of LTD was insignificantly affected (P > 0.05, Fig. 4B). It appears that the early phase of LTD is less sensitive to translational inhibition.
Roles of a rapidly turned over protein in LTD induction
Since the translational inhibitors suppressed the entire LTD
including its early phase (Figs. 2 and 3), one may assume that the
protein required for LTD induction is rapidly synthesized during
conjunctive stimulation. Such synaptically driven rapid protein
synthesis is possible. For example, Weiler et al. (1997) have recently reported that stimulation of type-1 metabotropic glutamate receptors in brain tissues triggered a rapid association of
mRNAs of the Fragile-X mental retardation protein (FMRP) with the
translational machinery within 1-2 min, resulting in an increased expression level of FMRP near synapses. Huber et al.
(2000)
suggest that the synthesis of proteins required for the
induction of mGluR5-mediated LTD in hippocampal neurons occurs within
minutes after cessation of LTD-inducing stimulation.
An alternative possibility is that the protein(s) required for the initial development of LTD is constitutively present and maintained critically at a level required for the induction of LTD by an unusually high turnover rate against fast degradation. When translational inhibitors block protein synthesis, the level of the protein(s) would decrease, resulting in blockade of LTD induction. In either case, the protein(s) is rapidly turned over, however, the difference between the two possibilities is whether production of the protein(s) depends on the LTD-inducing stimulation or not.
The translational inhibitors still blocked LTD induction even when
applied 10 min, but not 15 min or more, after the onset of conjunctive
stimulation (Figs. 3E, 4, and 5). This indicates that
protein synthesis is required for the initial development of LTD within
15 min after the onset of conjunctive stimulation but not for its
further development or maintenance. Therefore contrary to the current
presumption that protein synthesis is required for the development and
maintenance of the late phase of LTP or LTD, the present results
indicate that a rapidly turned over protein(s) is required for the
initial development of LTD only within 15 min following the onset of
conjunctive stimulation. During this critical period, there could be a
regenerative process such as the autophosphorylation of CaM kinase type
II that has been suggested to be involved in the initial phase of LTP
(Giese et al. 1998). The self-regeneration must be
transformed to a more stable process based on a change in receptors or
related molecular structures in synapses (Lisman and Falton
1999
). Thereafter, LTD would continually be enhanced even if
the protein synthesis is inhibited due to the takeover by a more
structurally based stable process, for example, the recently
demonstrated clathrin-mediated endocytosis (Wang and Linden
2000
).
Sites of translational and transcriptional inhibition
We have reported previously that translational inhibitors rapidly
suppress mGluR1-EPSPs in PCs (Karachot et al. 2000); the peak amplitude of mGluR1-EPSPs decreased to 50-60% during 5-min applications of anisomycin (55%), puromycin (51%), and cycloheximide (62%). This rapid onset of the effect of translational inhibitors suggests that the translation required for the generation of
mGluR1-EPSPs and presumably also for LTD induction occurs locally near
the synapses under observation. Nevertheless the relative contribution of somatic and dendritic translations to mGluR1-EPSPs and LTD should be
further investigated.
Our previous study (Karachot et al. 2000) also revealed
that it took 30 min to obtain a similar extent of decrease to
translational inhibitors after 5-min pulse applications of actinomycin
D (57%) and DRB (52%). It appears that LTD induction is blocked when
the peak amplitude of mGluR1-EPSPs decreases to 50-60% regardless of
the time course of the decrease. This close correspondence between the
observations on LTD and mGluR1-EPSPs is explicable if we assume that
both mGluR1-EPSPs and LTD depend on the local concentrations of the
same protein(s).
The time course of the depletion of the local protein(s) under
transcriptional inhibition will depend on the rate of transportation of
mRNAs from the transcription site in the nucleus to the translation site in somata and/or dendrites and that of proteins so produced from
the latter site to the dendritic synapses under observation as well as
on the rate of degradation of the mRNAs and protein(s). It is difficult
to precisely relate the presently obtained time lag of 30 min to these
factors because of lack of relevant data. It may, however, be pointed
out that the rate 242 µm/h measured for dendritic transport of mRNAs
(Muslimov et al. 1997) is too slow because it allows
mRNAs to travel a distance of only 120 µm even if the entire 30-min
lag time is spent for the transport. Since motor proteins, such as
kinesin and dynein superfamily proteins, are transported along
microtubules at rates of up to 4,320 µm/h (720 µm/10 min)
(Hirokawa 1998
), it is possible that the mRNAs and/or
protein(s) required for LTD induction travel faster and reach remote
sites of PC dendrites within a fraction of the 30-min time lag.
Relationship with metabotropic glutamate-mediated transmission
Because the translational inhibitors did not affect appreciably,
the GA-evoked EPSPs or the CF-evoked responses consisting of EPSPs and
Ca2+ spikes (Fig. 1, E-H), the
maintenance of ionotropic glutamate receptors or voltage-dependent
Ca2+ channels involved in the generation of these
electric signals does not require rapidly turned over proteins. On the
other hand, as reported previously (Karachot et al.
2000), generation of slow mGluR1-EPSPs in PCs requires a
rapidly turned over protein(s). Because an antagonist (Hartell
1994
), an antibody (Shigemoto et al. 1994
) and
gene deficiency (Aiba et al. 1994
; Conquet et al. 1994
) against mGluR1s block LTD induction, it may appear that translational and transcriptional inhibitors block LTD induction by
inactivating mGluR1s. However, inactivation of mGluR1s with AIDA did
not parallel the translational-inhibitor-caused blockade of LTD
induction in terms of the time course of their sensitive phases
following conjunctive stimulation (compare Figs. 3D and 5
and Fig. 10, C and D).
As suggested in the preceding text, rapid protein synthesis appears to
occur commonly in the signal transduction pathways linking mGluR1s to
both the generation of mGluR1-EPSPs and induction of LTD. The
well-known pathway linking mGluR1s to phospholipase C (PLC) via Gq
proteins in turn activates signal transduction cascades resulting in
production of diacylglycerol and protein kinase C, as well as inositol
trisphosphate, that eventually causes the release of
Ca2+ ions from intracellular stores. These
PLC-mediated pathways are involved in the induction of LTD
(Inoue et al. 1998; Linden and Connor
1991
), but not in the generation of mGluR1-EPSPs (Hirono et al. 1998
; Tempia et al. 1998
). Therefore it
is unlikely that the PLC pathways involve the presently postulated
rapid protein synthesis. The protein synthesis required for LTD
induction should occur in another, yet unidentified pathway, at its
stage of activation of G proteins or further downstream processes such
as activation of effector molecules, second messengers, and/or their
target molecules.
![]() |
FOOTNOTES |
---|
Address for reprint requests: M. Ito, Laboratory for Memory and Learning, Brain Science Institute, RIKEN, Wako, Saitama 351-0198, Japan (E-mail: Ito-BSI{at}brain.riken.go.jp).
Received 24 July 2000; accepted in final form 10 April 2001.
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|