1Interdepartmental Neuroscience Program, Yale University, New Haven, Connecticut 06520; and 2Department of Neurobiology and Behavior, University of California, Irvine, California 92697-4550
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Purcell, Angela L. and Thomas J. Carew. Modulation of Excitability in Aplysia Tail Sensory Neurons by Tyrosine Kinases. J. Neurophysiol. 85: 2398-2411, 2001. Tyrosine kinases have recently been shown to modulate synaptic plasticity and ion channel function. We show here that tyrosine kinases can also modulate both the baseline excitability state of Aplysia tail sensory neurons (SNs) as well as the excitability induced by the neuromodulator serotonin (5HT). First, we examined the effects of increasing and decreasing tyrosine kinase activity in the SNs. We found that tyrosine kinase inhibitors decrease baseline SN excitability in addition to attenuating the increase in excitability induced by 5HT. Conversely, functionally increasing cellular tyrosine kinase activity in the SNs by either inhibiting opposing tyrosine phosphatase activity or by direct injection of an active tyrosine kinase (Src) induces increases in SN excitability in the absence of 5HT. Second, we examined the interaction between protein kinase A (PKA), which is known to mediate 5HT-induced excitability changes in the SNs, and tyrosine kinases, in the enhancement of SN excitability. We found that the tyrosine kinases function downstream of PKA activation since tyrosine kinase inhibitors reduce excitability induced by activators of PKA. Finally, we examined the role of tyrosine kinases in other forms of 5HT-induced plasticity in the SNs. We found that while tyrosine kinase inhibitors attenuate excitability produced by 5HT, they have no effect on short-term facilitation (STF) of the SN-motor neuron (MN) synapse induced by 5HT. Thus tyrosine kinases modulate different forms of SN plasticity independently. Such differential modulation would have important consequences for activity-dependent plasticity in a variety of neural circuits.
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Tyrosine kinases are known to
play important roles in developmental processes such as cellular
proliferation and differentiation. Recently tyrosine kinases have also
been implicated in processes thought to underlie learning and memory.
For example, several tyrosine kinases, including Src and Fyn, are
involved in the induction of hippocampal long-term potentiation (LTP)
(Grant et al. 1992; Lu et al. 1998
;
O'Dell et al. 1991
; Yu et al. 1997
).
Furthermore, disruption of tyrosine kinase activation produces
behavioral deficits in the acquisition of some memory tasks
(Grant et al. 1992
; Whitechurch et al.
1997
). In addition, homozygous mutant mice lacking the TrkB
receptor tyrosine kinase exhibit an impairment in LTP (in CA1 of
hippocampus) as well as learning deficits (Minichiello et al.
1999
). Finally, mitogen-activated protein (MAP) kinase, activated downstream of receptor tyrosine kinases, is also required both for hippocampal LTP (English and Sweatt 1996
, 1997
)
and for associative and spatial forms of learning (Atkins et al.
1998
; Blum et al. 1999
; Schafe et al.
1999
).
Indirect evidence for tyrosine kinase function in synaptic plasticity
also exists in Aplysia. Long-term facilitation (LTF) of the
synapse between sensory neurons (SNs) and siphon motor neurons (MNs)
accompanies long-term sensitization of the defensive gill and
siphon-withdrawal reflex (Frost et al. 1985). Serotonin (5HT), a biogenic amine released during tail shock (Marinesco and Carew 2000
), also produces LTF of these monosynaptic
connections when repeatedly applied to either the intact CNS or
cocultures of SNs and MNs (Clark and Kandel 1993
;
Emptage and Carew 1993
; Montarolo et al.
1986
; Zhang et al. 1997
). Application of
brain-derived neurotrophic factor (BDNF), which activates the TrkB
receptor tyrosine kinase in vertebrates, to the intact CNS or SN-MN
cocultures produces facilitation of the SN-MN synapse 24 h later
(Giustetto et al. 1999
; McKay and Carew
1996
). Furthermore, treatment with TrkB-IgG protein blocks LTF
induced by 5HT, suggesting that an endogenous ligand of the TrkB
receptor tyrosine kinase is required for the induction of LTF
(Giustetto et al. 1999
). In addition, activation of MAP
kinase, a downstream substrate of receptor tyrosine kinases, has been
shown to be required for LTF induced by 5HT at the SN-MN synapse of
Aplysia (Martin et al. 1997
). Finally, the
Drosophila mutant linotte, which displays
impairments in odor-discrimination learning, has a disruption of a
receptor tyrosine kinase (Dura et al. 1993
, 1995
).
In addition to their function in synaptic plasticity, tyrosine kinases
have been shown to modulate ion channel activity (for review, see
Siegelbaum 1994), thereby regulating neuronal firing patterns and excitability. For example, a number of different potassium
channels have been shown to be regulated by tyrosine phosphorylation,
including Kv1.2 (Huang et al. 1993
; Lev et al. 1995
), Kv1.3 (Bowlby et al. 1997
; Fadool
et al. 1997
; Holmes et al. 1996a
), Kv1.5
(Holmes et al. 1996b
; Nitabach et al.
2001
), and Kv2.1 (Sobko et al. 1998
). Moreover,
both sodium current (Hilborn et al. 1998
;
Ratcliffe et al. 2000
) and calcium current (Hu et al. 1998
) can be controlled by direct tyrosine phosphorylation of their respective channels. Tyrosine phosphorylation inhibits a
nonselective cation current in pressure-sensitive cells in the leech
(Catarsi et al. 1995
), and in Aplysia,
dephosphorylation of tyrosine residues by a protein kinase A
(PKA)-activated tyrosine phosphatase switches the gating mode of a
nonselective cation current in bag cell neurons (Wilson and
Kaczmarek 1993
). Finally, tyrosine phosphorylation has also
been shown to regulate channel current through ionotropic receptors,
such as the N-methyl-D-aspartate (NMDA) receptor
(Yu et al. 1997
) and the nicotinic ACH receptor (Huganir 1991
).
In light of the evidence that tyrosine kinases and components of their signaling cascades can influence synaptic plasticity and electrical properties of cells, we were interested in exploring the role of tyrosine kinases in specific forms of plasticity in the Aplysia SNs and MNs. Here we show that manipulating the level of tyrosine kinase activity in the SNs modulates their excitability state in a bi-directional manner: blocking tyrosine kinase activity decreases SN excitability, while increasing kinase activity increases excitability. We also show that the tyrosine kinase pathway in the SNs can interact with the signaling cascade that mediates 5HT-induced increases in excitability. Finally, we demonstrate that tyrosine kinases can modulate different forms of short-term plasticity independently as tyrosine kinase inhibitors have no effect on 5HT-induced STF.
Some of the data presented in this paper have been reported previously
in abstract form (Purcell and Carew 1999, 2000
).
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Preparation
Wild-caught adult Aplysia californica (obtained from
Marinus, Long Beach, CA, or Marine Specimens Unlimited, Pacific
Palisades, CA) were anesthetized by injection of isotonic
MgCl2 (approximately 100 ml/100 g body wt).
Pleural-pedal ganglia were removed from the animal and incubated for
30-45 s in 0.4% glutaraldehyde to reduce contraction of the
connective tissue during 5HT application. Ganglia were pinned out in a
silicone elastomer (Sylgard)-coated recording dish containing a 1:1
mixture of isotonic MgCl2 and artificial sea
water (ASW) to prevent synaptic transmission during dissection. ASW
consisted of (in mM) 460 NaCl, 55 MgCl2, 11 CaCl2, 10 KCl, and 10 Tris, pH 7.6. The ganglia
were desheathed to expose the somatic clusters of the tail SNs in the
pleural ganglion and the tail MNs in the pedal ganglion (Walters
et al. 1983). During the experiment, the preparation was
perfused at room temperature (20-22°C) at a rate of approximately 4 ml/min with ASW or ASW containing specific drugs and/or 5HT. In some
cases where perfusion of drugs was cost prohibitive, experiments were
performed in a static bath (see Experimental procedures).
Intracellular recordings
The preparation was illuminated from below with a dark-field
condenser. Tail SNs and tail MNs were identified by their location in
the somatic clusters of the pleural and pedal ganglia, respectively. Glass microelectrodes filled with a 3 M KCl solution for intracellular recording had resistances between 8 and 15 M. Intracellular signals were amplified by Getting (model 5A, Iowa City, IA) and Axoclamp (model
2B, Axon Instruments, Foster City, CA) intracellular amplifiers. Data
were recorded and analyzed by a MacLab data-acquisition system (AD
Instruments, Mountain View, CA).
In all experiments, the resting membrane potential and the input
resistance of each neuron were monitored to assess the health of the
preparation. In synaptic experiments, the MN was hyperpolarized to 70
mV to prevent the cell from firing action potentials. The membrane
potential of the SNs was not altered during any of the experiments.
Excitability was determined by measuring the number of spikes elicited
in the SN during a 300-ms depolarizing current pulse. Intracellular
current injection was adjusted (typically from 0.5 to 2 nA) so that the
SN fired two to three spikes in the baseline condition. Monosynaptic
excitatory postsynaptic potentials (EPSPs) were evoked in the MNs by a
3-ms depolarizing current pulse in the SNs.
Experimental procedures
NEURONAL EXCITABILITY.
Tests were administered with an interstimulus interval (ISI) of 2 min.
After two stable baseline tests in ASW, the tyrosine kinase inhibitor
genistein (Sigma, St. Louis, MO), its inactive isomer, genistin
(Sigma), or vehicle (0.33% DMSO/0.43% ethanol) were perfused for 15 min prior to 5HT application, during which another two tests were
administered (at 13 and 15 min). The preparation was then exposed to a
solution containing 50 µM 5HT, and one of the drugs mentioned above
for at least 16 min, with tests taken every 2 min. In other
experiments, after stable baselines were established, two other
tyrosine kinase inhibitors, herbimycin A (Gibco Life Technologies,
Grand Island, NY) and lavendustin A (Calbiochem, La Jolla, CA),
dissolved in DMSO were added directly to a 2-ml static bath from
concentrated stock solutions achieving final bath concentrations of 5 and 50 µM, respectively. Fifteen minutes after drug addition, 220 µL of 5HT was added to the bath to a final concentration of 50 µM.
A similar testing protocol to that described for the genistein
experiments was used. In experiments using the tyrosine phosphatase
inhibitor potassium bisperoxo(1, 10-phenanthroline)oxovanadate(V) (bpV)
(Calbiochem), the preparation was exposed to the drug for 30 min in a
static bath prior to 5HT application, with tests taken every 5 min
beginning 15 min after exposure. 5HT was subsequently added to the
static bath and tests resumed every 2 min for at least 10 min. In
another series of experiments, the cAMP phosphodiesterase inhibitor
Ro-20-1724 and adenylyl cyclase activator
7-deacetyl-7-(O-N-methyl
piperazino)--butyryl-forskolin dihydrochloride (Calbiochem) were
perfused alone or with genistein for 15 min prior to testing. Perfusion
was stopped and testing performed in a static bath for 16 min before
5HT application. 5HT was added to the static bath (final bath
concentration of 50 µM) and testing continued for an additional 10 min.
SHORT-TERM SYNAPTIC FACILITATION. A nondepressed baseline EPSP amplitude was established in ASW by eliciting single spikes in the SN with an ISI of 15 min. The drug or vehicle solution was applied directly to a static bath 25 min before 5HT application during which a third baseline was obtained. The short-term test was taken immediately after a 5-min 50 µM 5HT pulse in a static bath. In a subset of experiments, SN excitability was monitored simultaneously with changes in synaptic efficacy. In these experiments, a second SN was impaled and used to monitor excitability changes.
Data analysis and statistics
Generally, two SNs per preparation were monitored for excitability changes. The responses of the two SNs were averaged to give a single value at each time point for the preparation. Due to the difference in drug application (perfusion vs. static bath), two sets of 5HT controls were performed (1 for each condition). No significant difference between the 5HT perfusion control and the 5HT static bath control was found at any of the time points [2-way ANOVA (treatment × time), F(1,4) = 0.707, NS]. Therefore the two groups were pooled for subsequent statistical analyses. EPSP amplitudes were measured as the peak voltage of the EPSP. The average of at least two pretests in ASW were taken as a baseline measure. Only EPSPs with pretests within 20% of the mean were used for further study (approximately 10% of all synapses were excluded by this criterion). All EPSP amplitudes were expressed as a percent of the baseline measure.
Spike amplitude and duration measurements were taken from SN recordings in which spikes were elicited by a 3-ms depolarizing stimulus. Spike amplitude was measured as the voltage change between the peak of the action potential and the average of the resting potential 15 ms before the current pulse. Spike duration was measured as the change in time between the peak of the action potential and the point at which the voltage was one-third of the spike amplitude. Input resistance measurements were obtained by injecting a 1-nA hyperpolarizing current pulse and recording the voltage at which the charge curve leveled out. For data analysis, both resting potential and input resistance were expressed as percentages of each cell's baseline value in normal ASW.
No significant effect of vehicle on 5HT-induced SN excitability was observed over any of the time points [2-way ANOVA (treatment × time), F(1,6) = 0.215, NS]. Therefore the vehicle control group was pooled with the 5HT control group for subsequent statistical analyses. In most cases, comparisons containing three or more between group variables were analyzed by an ANOVA followed by a Fisher's protected least significant difference (PLSD) post hoc analysis. In one case (Fig. 2B), a Dunnett's t-test was used as a post hoc analysis when multiple treatment groups were being compared with a control group. Simple comparisons between two groups were analyzed by an unpaired t-test. When comparing means within a group (for example, post-5HT treatment to pre-5HT treatment), a paired t-test was used. All probability values are two-tailed except for those reported for the three-electrode experiments. In this case, one-tailed t-tests were used because previous data predicted the expected direction of the effect. All data are displayed as the means ± SE.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Tyrosine kinase inhibitors attenuate 5HT-induced SN excitability
5HT induces several immediate changes in the tail SNs, including
increases in both excitability and spike duration. These changes have
been attributed mainly to activation of PKA and protein kinase C (PKC)
(for review, see Byrne and Kandel 1996). In light of the
recent evidence that growth factors can modulate plasticity induced by
5HT at Aplysia SN-MN synapses (Giustetto et al.
1999
; McKay and Carew 1996
; Zhang et al.
1997
), we examined whether tyrosine kinases, the primary
transducers of many growth factor signals, contributed to any of the
short-term modulatory effects produced by 5HT.
In the intact pleural ganglion, excitability was measured in the tail
SNs by determining the number of spikes elicited in the neuron in
response to a 300-ms intracellular depolarizing pulse. The current was
adjusted so that the neuron fired two to three spikes in the baseline
condition. Genistein, a broad-spectrum tyrosine kinase inhibitor
(Akiyama and Ogawara 1991; Akiyama et al.
1987
), or its vehicle (0.33% DMSO, 0.43% ethanol) were
applied to the preparation before and during the application of 50 µM 5HT. Representative results are shown in Fig.
1A. Summary data showing the
time course of excitability changes are depicted in Fig. 1B.
The increase in excitability induced by 5HT was stable over the time
period tested (Fig. 1B). However, when 5HT was accompanied by genistein, 5HT-induced excitability was clearly attenuated. Application of 50 µM genistein reduced the number of spikes fired during 5HT to approximately half of that exhibited under control conditions (mean number of spikes ± SE at 8-min time point:
genistein: 5.3 ± 1.0; vehicle: 9.1 ± 0.5). A twofold higher
concentration of genistein (100 µM) further reduced the number of
spikes elicited during 5HT application to a value less than baseline
(mean number of spikes ± SE at 8-min time point: genistein:
1.6 ± 0.6).
|
In addition to its effect on 5HT-induced excitability, we observed that genistein alone produced a significant decrease in baseline excitability of the SNs [ANOVA F(3,16) = 7.53, P < 0.05]. This reduction was evident comparing the genistein group both to vehicle controls (Fig. 1B; mean number of spikes ± SE at 0-min time point: vehicle: 1.5 ± 0.2; 100 µM genistein: 0.6 ± 0.2; Fisher's PLSD, P < 0.05) and to inactive isomer controls (data not shown; mean number of spikes ± SE at 0-min time point: 100 µM inactive isomer genistin: 1.8 ± 0.2; Fisher's PLSD, P < 0.05). These results suggest that there is a baseline level of tyrosine kinase activity in the SNs in the absence of 5HT stimulation and that reducing this basal kinase activity produces a corresponding reduction in the excitability state of the SNs.
We also observed that genistein altered the waveform of the SN action
potential. Application of 100 µM genistein in the absence of 5HT
decreased the amplitude of the action potential by approximately 22%
and increased the duration of the action potential by 93% (data not
shown). These effects were not observed with the inactive isomer or ASW
alone. Consistent with these observations, Jonas et al.
(1996) have shown that when tyrosine kinases are activated (as
opposed to being blocked, as in these experiments), the exact opposite
effects on cell excitability and spike waveform have been observed. In
bag cell neurons of Aplysia application of insulin, which
activates a tyrosine kinase receptor, increases excitability, increases
the amplitude of the action potential and decreases the duration of the
spike (Jonas et al. 1996
). Further experiments will be
directed at elucidating the ionic basis of these changes.
To examine the specificity of genistein's actions, we tested the
effect of genistein's inactive isomer, genistin, as well as two
additional tyrosine kinase inhibitors, herbimycin A and lavendustin A
(Onoda et al. 1989; Uehara and Fukazawa
1991
; Uehara et al. 1989
) on 5HT-induced SN
excitability (Fig. 2A). An
overall ANOVA revealed a significant difference of treatment in the
number of spikes elicited in 5HT [Fig. 2B;
F(5,34) = 9.20, P < 0.05]. While
genistein produced a significant concentration-dependent decrease in
5HT-induced excitability (50 µM, tD = 3.77 and 100 µM, tD = 7.47, P < 0.05), its inactive isomer, genistin (100 µM), had no significant effect (Fig. 2, A and B;
tD = 1.37, NS), demonstrating that the
effect of genistein was not due to nonspecific actions of the drug. In
addition, both 5 µM herbimycin A and 50 µM lavendustin A
significantly reduced the increase in excitability produced by 5HT
(Fig. 2, A and B;
tD = 4.67 and
tD = 3.87, respectively, P < 0.05). Similar results were obtained if the change
in number of spikes was analyzed to account for the shift in baseline
(data not shown). These data show that tyrosine kinase activity
modulates the increase in SN excitability produced by 5HT.
|
Inhibiting tyrosine phosphatase activity increases SN excitability
Since reducing tyrosine kinase activity attenuates SN excitability, the question arises whether the reverse is true: does increasing tyrosine kinase activity increase SN excitability? To explore this question, we unmasked the effects of endogenous tyrosine kinase activity by bath applying bpV, a membrane permeable tyrosine phosphatase inhibitor. Reducing the actions of opposing tyrosine phosphatase activity should lead to a net increase in the amount and/or duration of phosphorylation induced by underlying tyrosine kinase activity. We found that SN excitability was significantly increased in 100 µM bpV alone [Fig. 3, A and C1; ANOVA, F(2,11) = 5.13, Fisher's PLSD, P < 0.05]. The increase in excitability was stable over time (Fig. 3B). Subsequent addition of 5 µM 5HT produced a slight increase in excitability over that observed with bpV alone. However, the presence of the phosphatase inhibitor did not enhance the excitability produced by 5HT [Fig. 3C2; ANOVA, F(2,11) = 0.31, NS]. Moreover, bpV increased SN excitability in a concentration-dependent manner (Fig. 3C1) but had no effect on 5HT-induced excitability (Fig. 3C2).
|
Importantly, these results suggest that there is a persistent level of tyrosine kinase activity in the baseline state (i.e., in the absence of 5HT-induced excitability) that can be modulated to produce changes in the cell's excitability. The fact that the phosphatase inhibitor had no effect on the excitability produced by 5HT suggests that the signaling pathways of tyrosine kinases and 5HT may converge on a common downstream substrate, for example, an ion channel (see DISCUSSION).
None of the tyrosine kinase inhibitors produced significant changes in the SN resting membrane potential [ANOVA, F(5,22) = 1.92, NS]. However, an overall ANOVA revealed a significant difference in input resistance among the inhibitors [F(5,23) = 5.76, P < 0.05]. A significant decrease in input resistance occurred in the presence of herbimycin A, lavendustin A, and the highest concentration of genistein (tD = 3.18, tD = 4.04, and tD = 2.86, respectively, P < 0.05), while the inactive isomer, genistin, had no significant effect (tD = 1.59). Consistent with this finding, application of bpV significantly increased the input resistance of the SNs compared with ASW controls (t8 = 4.60, P < 0.05) while having no significant effect on SN resting potential (t8 = 0.32, NS). These changes in resting membrane conductance could reflect changes in baseline ion channel activity due to modulation of tyrosine kinase activity. Further experiments will investigate the nature of the changes in input resistance.
Intracellular injection of active Src kinase increases SN excitability
The cytoplasmic Src tyrosine kinase has been observed to
modulate different ion channels, including Kv1.3, Kv1.5, and the NMDA
receptor (Fadool et al. 1997; Holmes et al.
1996a
,b
; Yu et al. 1997
). To test whether this
kinase may also affect the excitability of Aplysia SNs, we
injected an active form of the human recombinant Src tyrosine kinase
into the SNs. SNs injected with the Src kinase showed a steady increase
in excitability over time (Fig.
4, A and
B1). Src induced a significant increase in excitability at both 30 and 60 min after injection compared with the control solution [Fig. 4B2; 30 min: ANOVA, F(2,26) = 9.04, Fisher's PLSD, P < 0.05; 60 min: ANOVA,
F(2,24) = 9.30, Fisher's PLSD, P < 0.05]. Subsequent addition of 5HT produced a significant increase in
excitability over that observed at 60 min in both Src-injected and
control-injected cells (Fig. 4B2; paired t-tests,
control: t8 = 7.87, P < 0.05; Src: t13 = 5.44, P < 0.05). However, there was no significant difference between Src-injected cells and control-injected cells in the
final levels of excitability induced by 50 µM 5HT [Fig. 4B2, ANOVA F(2,23) = 5.22, Fisher's PLSD,
NS]. This finding is consistent with our earlier results showing that
bpV application did not enhance 5HT-induced excitability.
|
It is possible that 5HT did not further increase excitability induced by Src injections because Src alone brought the SNs to a "ceiling" beyond which no further increase in excitability was possible. To examine the question, an additional 1-2 nA over the testing current was injected into Src-treated cells still in the presence of 50 µM 5HT. The additional current significantly increased the number of spikes elicited during the depolarizing pulse (5HT: 15.5 ± 1.1, n = 14; 5HT + current: 18.1 ± 1.5, n = 12; t11 = 3.51, P < 0.05; data not shown). These results show that a ceiling effect cannot account for the results shown in Fig. 4B2 and support the conclusion that the signaling cascades engaged by Src interact with those engaged by 5HT.
To confirm that the increase in excitability produced by Src was due to its tyrosine kinase activity, herbimycin A, a drug that irreversibly inactivates Src kinase, was applied to the preparation before and during Src injection. Herbimycin A completely blocked the enhancement of excitability induced by Src injection (Fig. 4, A and B1). In fact, the excitability of SNs injected with Src kinase in the presence of herbimycin A was not significantly different from that of SNs injected with control solution (Fig. 4B2, 30 and 60 min: Fisher's PLSD, NS). Consistent with our earlier results, herbimycin A significantly attenuated the increase in excitability induced by application of 50 µM 5HT [Fig. 4B2, ANOVA, F(2,23) = 5.22, PLSD, P < 0.01]. This series of experiments shows that injected constitutively active Src kinase causes an increase in Aplysia SN excitability through tyrosine phosphorylation. Furthermore, modulation by Src kinase can interact with 5HT-induced excitability (see DISCUSSION).
SN excitability induced by cAMP elevation is diminished by a tyrosine kinase inhibitor
We have shown that changes in the level of tyrosine kinase
activity can alter both baseline SN excitability and
excitability induced by 5HT. To further examine the possible site of
action of tyrosine kinases in the signaling cascade, we examined the effects of genistein on excitability induced by elevation of cAMP which
presumably leads to the activation of PKA, a kinase that has been
implicated in 5HT-induced changes in the SNs (Baxter and Byrne
1990; Ghirardi et al. 1992
; Goldsmith and
Abrams 1992
; Klein et al. 1986
). Intracellular
levels of cAMP were increased by applying 100 µM forskolin, an
activator of adenylyl cyclase, and 500 µM Ro-20-1724, a
phosphodiesterase inhibitor. Confirming a previous report
(Goldsmith and Abrams 1992
), we found that perfusion of
these cAMP elevators produced an increase in excitability that was
stable over time (Fig. 5A).
Moreover, addition of 50 µM 5HT to the bath, which already contained
the cAMP elevators, produced a further increase in excitability. Both
the cAMP-induced increase in excitability and the further increase
induced by 5HT were significantly reduced by blocking tyrosine kinase
activity with genistein. Specifically, we observed a significant
reduction in the amount of excitability induced by the cAMP elevators
when 50 µM genistein was included in the bath (Fig. 5, A
and B, 23-min time point,
t8 = 3.30, P < 0.05). Subsequent addition of 5HT to the bath containing the cAMP
elevators and genistein produced an increase in excitability over that
with drugs alone. However, this increase was significantly less than
that observed in preparations that were not exposed to genistein (Fig.
5B, 37-min time point, t8 = 3.26, P < 0.05).
|
These results show that genistein can attenuate the enhancement of excitability produced by elevation of cAMP as well as that produced by 5HT. Since elevation of cAMP levels leads to the activation of PKA and since 5HT is thought to mediate its effects through elevation of PKA activity, these results suggest that tyrosine kinase activity influences SN excitability at a site that is downstream of PKA's actions. Subsequent addition of 5HT produces a comparable further increase in excitability to that induced by cAMP elevation in both the presence or absence of genistein (Fig. 5A). This result suggests that inhibiting tyrosine kinase activity reduces baseline SN excitability, thereby attenuating the effects of 5HT modulation. Therefore the 5HT cascade may be functionally independent of tyrosine kinase activity.
Tyrosine kinase activity is not required for short-term synaptic facilitation
We have shown that manipulating the degree of tyrosine
kinase activity can modulate (both increase and decrease) the level of
excitability in SNs. We next examined whether tyrosine kinases play a
role in other rapidly induced modulatory effects of 5HT. It is well
established that a brief (5 min) pulse of 5HT produces an enhancement
of synaptic efficacy between tail SNs and tail MNs that lasts for
approximately 15 min (Mauelshagen et al. 1996; Walters et al. 1983
). To examine the effect of blocking
tyrosine kinase activity on short-term facilitation (STF), 100 µM
genistein or its vehicle were applied to the preparation 25 min prior
to a 5-min 5HT pulse. As shown in Fig. 6,
the drug alone had no significant effect on synaptic transmission (Fig.
6, A and B; t18 = 1.15, NS). In
addition, genistein had no effect on the input resistance of the MN
(t16 = 0.06, NS). Subsequent 5HT
exposure produced significant STF in both the vehicle control group
(t18 = 5.50, P < 0.05) and the group exposed to genistein
(t18 = 2.60, P < 0.05). Moreover, there was no significant difference in the amount of
STF observed between the genistein and vehicle groups
(t18 = 0.46, NS). In a subset of these
experiments, excitability was simultaneously monitored in a second SN
(Fig. 7, A1 and
B1). Although genistein did not block STF produced by 5HT,
consistent with our previous results, it completely blocked the
5HT-induced increase in excitability (Fig. 7, A2 and
B2). Thus there was a significant difference in the number
of spikes fired during 5HT between the vehicle and genistein groups
(t4 = 4.67, P < 0.05). Moreover, 5HT produced a significant increase in excitability
compared with vehicle alone (t4 = 5.74, P < 0.05), while no difference was observed
between genistein + 5HT and genistein alone
(t4 = 0.50, NS). Finally, significant STF was observed in both the vehicle and genistein groups
(vehicle; t4 = 7.15 and genistein;
t4 = 3.93, P < 0.05), and no difference was found in the amount of facilitation between groups (t4 = 0.08, NS). Therefore
excitability and STF induced by 5HT in the SNs exhibit independent
modulation: while tyrosine kinase activity clearly alters the level of
SN excitability, it does not appear to be involved in short-term
modulation of synaptic efficacy.
|
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Several studies have shown that the increase in membrane
excitability induced by 5HT in Aplysia SNs is mediated by
PKA (for review, see Byrne and Kandel 1996).
Pharmacological manipulations which increase cAMP levels or direct
injection of cAMP into SNs produce an increase in excitability similar
to that produced by 5HT application (Baxter and Byrne
1990
; Goldsmith and Abrams 1992
; Hochner
and Kandel 1992
; Klein et al. 1986
). Moreover,
inhibitors of PKA block or attenuate 5HT-induced increases in SN
excitability (Goldsmith and Abrams 1992
; Hochner
and Kandel 1992
). On a mechanistic level, 5HT and agents that
activate PKA both modulate a specific type of potassium current, the
"S" (serotonin) current (IKS)
(Baxter and Byrne 1989
; Klein et al.
1982
; Pollock et al. 1985
). Single-channel recordings revealed an S-type potassium channel is modulated by 5HT,
cAMP (Siegelbaum et al. 1982
), and PKA (Shuster
et al. 1985
). In addition, both 5HT and cAMP reduce a
calcium-activated potassium current
(IK,Ca) in SNs (Walsh and Byrne
1989
). In light of this evidence, the current model for
5HT-induced changes in SN excitability suggests that 5HT increases the
level of cAMP, thereby activating PKA (Byrne and Kandel
1996
), leading to the suppression of both IKS and
IK,Ca and the enhancement of SN
excitability (Baxter et al. 1999
).
In the present study, we provide evidence for a novel mechanism by which the excitability state of the SNs can be modulated. Table 1 summarizes the results of manipulations of tyrosine kinase activity on SN excitability. We find that inhibition of tyrosine kinase activity reduces baseline excitability as well as the increase in excitability produced by both 5HT and PKA. Conversely, we find that, in the absence of 5HT, functionally increasing the effects of cellular tyrosine kinase activity by application of a tyrosine phosphatase inhibitor or by direct injection of the active Src tyrosine kinase increases SN excitability. Moreover, the tyrosine kinase signaling cascade appears to interact with downstream components of 5HT modulation as their combined effects are not simply additive. These results show that the excitability state of the SNs is modulated by tyrosine kinases and this modulation can influence the subsequent effects of 5HT in the SNs.
|
Changes in tyrosine kinase activity modulate the excitability state of the SNs
Our data show that increasing or decreasing the level of tyrosine kinase activity produces corresponding changes in the excitability state of the SNs. An interesting implication of these results is that there is persistent tyrosine kinase activity within the SNs since application of a tyrosine phosphatase inhibitor produces an increase in excitability in the absence of 5HT. This finding suggests that tyrosine phosphatases tightly regulate the level of substrate phosphorylation on tyrosine residues. When the opposing phosphatase activity is removed by an inhibitor, the persistently active tyrosine kinases now become the dominant enzyme and are able to more effectively phosphorylate their substrates. Furthermore treatment with a tyrosine kinase inhibitor in the absence of 5HT decreases baseline SN excitability, suggesting that the baseline phosphorylation levels can be downregulated as well.
Several examples of modulation by persistent tyrosine phosphorylation
have been reported previously. The voltage-dependent potassium channel
Kv1.3 found in mammalian brain exhibits a basal level of tyrosine
phosphorylation (Holmes et al. 1996a). Moreover, when
cells expressing this channel are treated with a tyrosine phosphatase
inhibitor, increases in tyrosine phosphorylation of the channel occur
in a time- and dose-dependent manner, with a corresponding suppression
of current through the channel. In leech pressure-sensitive neurons, a
cation channel is inhibited by basal tyrosine phosphorylation as
phosphatase treatment of inside-out patches containing the channel
increases channel activity (Aniksztejn et al. 1997
). In
Aplysia bag cell neurons, the level of tyrosine phosphorylation of a cation channel or a closely associated channel protein determines the gating mode of the channel as well as its behavior to subsequent second-messenger signals (Wilson and
Kaczmarek 1993
). Collectively, these studies suggest several
candidate loci for tyrosine regulation that can regulate basal ion
channel activity and thus the firing properties and excitability state
of the neurons.
Two possible models of tyrosine kinase modulation
We have shown that blocking tyrosine kinase activity can attenuate
excitability induced by 5HT and increasing tyrosine kinase activity
increases SN excitability, which occludes further increases induced by
5HT. These data suggest that tyrosine kinases function downstream of
the 5HT receptor and either operate in series with the 5HT cascade or
converge on a common substrate. We have also shown that a tyrosine
kinase inhibitor can attenuate excitability induced by elevation of
cAMP levels, suggesting that the tyrosine kinase functions downstream
of PKA activation as well. These findings suggest two possible models
that describe how tyrosine kinases may modulate SN excitability in
conjunction with the 5HT cascade (Fig.
8). The first model proposes that
tyrosine kinases function in series with the 5HT cascade and operate
downstream of PKA (Fig. 8A). This scenario requires that PKA
would eventually lead to the activation of a tyrosine kinase. Evidence
for this type of signaling cascade has recently been reported in mouse
adipocytes (Fredriksson et al. 2000). Norepinephrine
induces the expression of vascular endothelial growth factor through a
cascade involving
-adrenoreceptors, PKA activation, and the Src
tyrosine kinase. Activation of a tyrosine kinase downstream of PKA has
also been observed in primary cultures of striatal neurons
(Vincent et al. 1998
). The tyrosine kinase inhibitor
genistein attenuates forskolin-induced MAP kinase activation in these
neurons. However, this type of signaling cascade is not yet well
understood and it is not yet clear how PKA leads to the activation of
tyrosine kinases.
|
In addition, we found that 5HT application was able to further increase
excitability over that induced by cAMP elevators in the presence or
absence of genistein, suggesting that tyrosine kinases may function
independently of the 5HT cascade. This result contrasts with two
previous studies (Baxter and Byrne 1990; Hochner and Kandel 1992
), which report that either bath application of cAMP analogues or direct injection of Sp-cAMPS into the SNs occluded further changes in excitability produced by 5HT. Several experimental factors could account for the differences in these observations, such
as different pharmacological agents and type of preparation. For
example, the experiments in both the study by Baxter and Byrne (1990)
and by Hochner and Kandel (1992)
were
performed in excised SNs and used cAMP analogues to elevate cAMP
levels, whereas we used the intact nervous system and
forskolin/Ro-20-1724 to increase cAMP levels in this study. Further
experiments will resolve this apparent difference.
A more likely model that is consistent with our results is that
tyrosine kinases are not activated downstream of 5HT but rather they
phosphorylate substrates that are modulated downstream of 5HT and PKA
(Fig. 8B). Changes in persistent tyrosine phosphorylation through changes in the balance between tyrosine kinases and tyrosine phosphatases would adjust the baseline ion channel activity and modify
the cell's response to subsequent exposure to neuromodulators, such as
5HT. As mentioned previously, there are several ion channels that have
basal tyrosine phosphorylation. It is conceivable that a channel, such
as the S channel, is modulated by both tyrosine kinases and PKA.
Alternatively, tyrosine kinases may modulate PKA directly. In
Aplysia, the catalytic subunit of PKA can have one of two
different amino termini (Beushausen et al. 1992). Only catalytic subunits containing the second amino terminus can be phosphorylated in vitro by Src kinase. Furthermore it has been shown
that the substrate selection of PKA depends on which amino terminus the
catalytic subunit possesses (Panchal et al. 1994
). An
intriguing possibility is that tyrosine phosphorylation could regulate
the dynamics of PKA substrate phosphorylation. Future experiments will
help distinguish between these two possible models.
STF is not modulated by tyrosine kinases
Although inhibiting tyrosine kinase activity clearly reduced
SN excitability, it had no effect on STF. This result suggests that
substrates for tyrosine kinases are not required for the enhancement of
transmitter release in the short term. It also demonstrates that two
different types of short-term plasticity that are induced by the same
neuromodulator (5HT) can be differentially regulated. Therefore stimuli
that induce changes in tyrosine kinase activity (i.e., growth factors,
cytokines) may affect the firing properties of the SNs and their
subsequent responses to neuromodulators without changing the synaptic
efficacy of the SNs onto their follower cells. In addition to
modulatory plasticity induced by 5HT, the SNs also exhibit intrinsic
(homosynaptic) plasticity that is activity dependent. Low rates of
firing give rise to homosynaptic depression (Armitage and
Siegelbaum 1998; Castellucci et al. 1970
), while high rates produce synaptic facilitation (e.g., post tetanic
potentiation) (Bao et al. 1997
; Schaffhausen et
al. 2001
; Walters and Byrne 1984
). Thus tyrosine
kinase modulation of excitability (but not STF) in the SNs could
provide a means of differential regulation of homosynaptic plasticity,
leaving heterosynaptic plasticity (e.g., induced by 5HT) unaltered.
This confers a high degree of flexibility in the induction of synaptic
plasticity at the SN-MN synapse.
Tyrosine kinase activity provides a mechanism for gain control of cellular plasticity
There are many different external signals that can trigger changes in cellular tyrosine kinase activity. Several cytokines and growth factors utilize tyrosine kinases to initiate gene expression after injury or during periods of development. Our results suggest that changes in the persistent level of tyrosine kinase activity can have a major impact on Aplysia SN excitability. Therefore alterations in the amount of tyrosine phosphorylation within a neuron can provide a cellular context which can affect the cell's response to subsequent signaling events.
This type of modulation of plasticity may be operative during
development as well. Growth factors more prevalent during particular periods of development could activate tyrosine kinase cascades unique
to that growth period. The increased SN excitability due to high
tyrosine phosphorylation levels may, in turn, facilitate the occurrence
of specific types of activity-dependent plasticity (Hawkins et
al. 1983; Sutton and Carew 2000
; Walters
and Byrne 1983
, 1984
).
If particular forms of ion channel function or plasticity are regulated
by tyrosine kinases, this could provide a mechanism for modulation
within a particular developmental window. An example of this type of
modulation of plasticity in a specific developmental window is observed
in vertebrates with the change of subunit composition of the NMDA
receptor. During development, the majority of NMDA receptors contain
NR2B subunits that cause the receptor to have a longer open time and
allow a greater calcium influx. As the animal matures, the relative
composition of the NMDA receptor changes, with proportionally greater
amounts of NR2A subunits; thus the channel open time is decreased
considerably (Carmignoto and Vicini 1992; Sheng
et al. 1994
). Consistent with this observation, the threshold
for LTP is lower in developing animals expressing NMDA receptors
containing NR2B subunits (Kirkwood et al. 1995
). Furthermore, in adult transgenic animals that only have NR2B-containing NMDA receptors, the threshold for LTP is lower and they perform better
than normal adults in learning tasks (Tang et al. 1999
).
In conclusion, tyrosine phosphorylation can, in principle, adjust the excitatory state of a neuron, which in turn could contribute to a variety of forms of plasticity both intrinsic to the cell and with its synaptic partners. Thus elucidation of the role of tyrosine kinase activity in different types of cellular plasticity will help to gain a more complete understanding of how differential regulation of specific second-messenger cascades may facilitate information processing in the nervous system.
![]() |
ACKNOWLEDGMENTS |
---|
The authors thank J. H. Byrne, C. M. Sherff, M. A. Sutton, and U. Mueller for helpful comments on an earlier version of the manuscript.
This work was supported by a predoctoral National Science Foundation fellowship to A. L. Purcell and National Institute of Mental Health Grant RO1 MH-14-1083 to T. J. Carew.
![]() |
FOOTNOTES |
---|
Address for reprint requests: T. J. Carew (E-mail: tcarew{at}uci.edu).
Received 7 December 2000; accepted in final form 14 March 2001.
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|