GABAergic Synaptic Transmission Regulates Calcium Influx During Spike-Timing Dependent Plasticity

Coincident pre- and postsynaptic activity of hippocampal neurons alters the strength of gamma-aminobutyric acid (GABAA)-mediated inhibition through a Ca2+-dependent regulation of cation-chloride cotransporters. This long-term synaptic modulation is termed GABAergic spike-timing dependent plasticity (STDP). In the present study, we examined whether the properties of the GABAergic synapses themselves modulate the required postsynaptic Ca2+ influx during GABAergic STDP induction. To do this we first identified GABAergic synapses between cultured hippocampal neurons based on their relatively long decay time constants and their reversal potentials which lay close to the resting membrane potential. GABAergic STDP was then induced by coincidentally (±1 ms) firing the pre- and postsynaptic neurons at 5 Hz for 30 s, while postsynaptic Ca2+ was imaged with the Ca2+-sensitive fluorescent dye Fluo4-AM. In all cases, the induction of GABAergic STDP increased postsynaptic Ca2+ above resting levels. We further found that the magnitude of this increase correlated with the amplitude and polarity of the GABAergic postsynaptic current (GPSC); hyperpolarizing GPSCs reduced the Ca2+ influx in comparison to both depolarizing GPSCs, and postsynaptic neurons spiked alone. This relationship was influenced by both the driving force for Cl− and GABAA conductance (which had positive correlations with the Ca2+ influx). The spike-timing order during STDP induction did not influence the correlation between GPSC amplitude and Ca2+ influx, which is likely accounted for by the symmetrical GABAergic STDP window.


IntroductIon
GABAergic synaptic transmission can be either excitatory or inhibitory at different stages of nervous system development (Kaila, 1994;Blaesse et al., 2009), or during various pathological states (Kahle et al., 2008). The polarity of GABAergic transmission (hyperpolarizing versus depolarizing) depends on the intracellular concentration of Cl − ([Cl − ] i ); this is because the GABA A receptor is a Cl − -permeable ion channel (Kaila, 1994). The [Cl − ] i is largely determined by the cation-chloride cotransporters expressed in the neuronal membrane: NKCC1 accumulates Cl − into the neuron (Yamada et al., 2004;Dzhala et al., 2005), while KCC2 transports it out (Rivera et al., 1999). When neuronal Cl − is relatively high due to the dominant expression of NKCC1 during early development, the reversal potential for GABA (E GABA ≈ E Cl ) is depolarized with respect to the resting membrane potential and so GABAergic transmission is depolarizing and sometimes excitatory. In contrast, when neuronal Cl − is low due to the expression of KCC2 in the mature nervous system, E Cl is hyperpolarized with respect to the resting membrane potential making GABAergic transmission inhibitory.
We examined Ca 2+ dynamics during STDP induction in hippocampal neurons which have formed either depolarizing or hyperpolarizing GABAergic synapses. Using perforated patch-clamp recordings (with gramicidin) and imaging postsynaptic Ca 2+ (using Fluo4), we investigated how the polarity and strength of GABAergic transmission regulates Ca 2+ influx. We further analyzed this relationship by examining several aspects of GABAergic transmission, including GABA A conductance, E Cl , and Cl − driving force. We found that GABA A -mediated transmission regulates Ca 2+ influx during the induction of STDP, with the strength of the synapse significantly altering the magnitude of the postsynaptic Ca 2+ influx in a linear fashion.

MaterIals and Methods hIppocaMpal cultures
Low-density cultures of dissociated embryonic rat hippocampal neurons were prepared as previously described . In brief, embryonic day 18 (E18) pregnant Sprague-Dawley rats were briefly exposed to carbon dioxide and cervically dislocated in accordance with guidelines from the University of Toronto Animal Care Committee and the Canadian Council on Animal Care. Hippocampi were then removed and treated with trypsin for 15 min at 37°C, followed by gentle trituration. The dissociated cells were plated at a density of 50,000 cells/mL on poly-l-lysine coated 25 mm glass coverslips (in 35 mm Petri dishes). Cells were plated in Neurobasal medium (Invitrogen, Carlsbad, California, USA), supplemented with 2% B-27 (Invitrogen, Carlsbad, California, USA). Twenty-four hours after plating, half of the medium was replaced with the original plating medium containing 20 mM KCl. Forty-eight hours after plating, and every 3 days following, one third of the medium was replaced with DMEM (Invitrogen, Carlsbad, California, USA) supplemented with 10% heat-inactivated fetal bovine serum (Invitrogen, Carlsbad, California, USA), 10% Ham's F12 with l-glutamine (Invitrogen, Carlsbad, California, USA), 1% penicillin-streptomycin (Sigma-Adrich, Oakville, Ontario, Canada), 10 mM KCl and 15 mM HEPES. Both glia and neurons were present under these culture conditions. Cells were recorded from after 8-13 days in culture.

electrophysIology
Whole-cell perforated patch recordings using gramicidin (50 μg/ mL; Sigma-Adrich, Oakville, Ontario, Canada) were performed on pairs of synaptically connected cultured hippocampal neurons. The recording pipettes were made from glass capillaries (World Precision Instruments Inc., Sarasota, Florida, USA), with a resistance of 4-10 MΩ. The pipettes were filled with an internal solution containing 150 mM KCl, 10 mM HEPES, and gramicidin, pH 7.4, osmolarity = 300 mOsmol. The cultures were continuously perfused (approximately 1 mL/min) with extracellular recording solution containing (in mM): 150 NaCl, 3 KCl, 3 CaCl 2 ·2H 2 O, 2 MgCl 2 ·6H 2 O, 10 HEPES, 5 Glucose, pH 7.4, osmolarity = 307-315 mOsmol. Recordings were performed with a MultiClamp 700B (Molecular Devices Inc., Sunnyvale, California, USA) patch-clamp amplifier. Signals were filtered at 5 kHz using amplifier circuitry. Data was acquired and analyzed using Clampfit 9 (Molecular Devices Inc., Sunnyvale, California, USA). Recordings started after the series resistance had dropped below 30 MΩ. For assaying synaptic connectivity, each neuron was stimulated at a low frequency (0.05 Hz) by a 1 ms step depolarization from −70 to +20 mV in voltage-clamp mode. GPSCs were distinguishable from excitatory postsynaptic currents (EPSCs) by longer decay times. Upon occasion we did detect autaptic GABAergic synapses in our cultures, however we did not examine these synapses in the present study. During the STDP induction protocol both neurons were switched to current-clamp mode and injected with current (minimal stimulation, 2 ms) both pre-and postsynaptically to generate an action potential in each cell at a frequency of 5 Hz for 30 s. The interval between spike induction was ±5 ms, which resulted in a spike-timing interval of ±1 ms between onset of the GABAergic postsynaptic potential (GPSP) and the postsynaptic action potential. This protocol resulted in 150 pairs of pre-and postsynaptic action potentials. All recordings were performed at room temperature (25°C).
The resting membrane potential was determined in currentclamp mode in the absence of current injection or synaptic activity. E Cl was determined by varying the holding potential of the postsynaptic cell in 10 mV increments and measuring the resulting GPSC amplitude; each set of current-voltage (I-V) measurements was repeated after a 5-min interval. A linear regression of both sets of GPSC amplitude measurements was then used to calculate the voltage dependence of GPSCs. The intercept of this line with the abscissa was taken as E Cl . The slope of the same line was taken as GPSC conductance. The difference between the resting membrane potential and E Cl was taken as the driving force.
GABA A receptors are permeable to both HCO 3 − and Cl − (∼0.2-0.4 ratio; Kaila, 1994). Due to the relatively positive HCO 3 − equilibrium potential (∼−10 mV), which is set by mechanisms that control intracellular pH regulation (Kaila and Voipio, 1987), HCO 3 − mediates an inward, depolarizing current (Kaila and Voipio, 1987;Kaila et al., 1993;Gulledge and Stuart, 2003). However, our experiments were performed in bicarbonate-free solution buffered with HEPES, and thus GABA A receptor activation was solely mediating a Cl − current. For this reason we report E Cl and not E GABA .

Fluorescence IMagIng
To assess the effect of STDP induction on postsynaptic Ca 2+ influx the hippocampal neurons were loaded with the membrane-permeable fluorescent Ca 2+ indicator Fluo4-AM (Invitrogen; Carlsbad, CA, USA) for 30 min at 37°C, 5% CO 2 . The Fluo4 was dissolved in dimethyl sulfoxide (DMSO) and 20% pluronic acid to a stock concentration of 1 mM and then diluted to 1 μM in our extracellular recording solution. Following dye-loading the cells were thoroughly washed with extracellular recording solution. Cells were then transferred to the recording chamber of an inverted microscope (Olympus IX71) equipped with an Olympus 0.6 NA × 40 objective. Fluo4 was excited at 488 nm through a monochromator (Photon Technology International (Canada) Inc., London, ON), controlled by the ImageMaster software (Photon Technology International (Canada) Inc., London, ON). Fluorescence emission of labeled cells at 510 nm was detected with a 16-bit CCD camera (Cascade 650, Photometrics, Roper Scientific, Tuscon, AZ, USA). Images of 653 × 492 pixels were accumulated at 500-1000 ms intervals.

Fluorescence analysIs
Analysis of the fluorescence signals was performed off-line on the image sequences as they were originally acquired. Analysis was performed on regions of interest (ROIs) that encompassed approximately 80% of the soma (which ranged from 10-50 μM in diameter). Fluorescence was plotted against time to yield a graph of the fluorescence changes over the STDP induction period. F 0 was taken to be the fluorescence from the last image before induction began. F peak was taken to be the maximum fluorescence level reached over the course of the induction. F 30 was taken to be the fluorescence from the final image during the induction period. ∆F 30 was calculated as the percentage difference between F 0 and F 30 ; ∆F peak was calculated as the percentage difference between F 0 and F peak . The area under the graph was normalized to F 0 to provide F area . We determined that photobleaching did not impact our Ca 2+ analysis. This determination was made by comparing the fluorescence of quiescent cells during the first and last 5 s of a 30-s image acquisition and finding no significant difference in fluorescence (p = 0.06). Thus, we did not alter the image analysis further to account for photobleaching.

statIstIcal analysIs
All data are presented as mean ± SEM. Linear regression analysis was used to obtain correlation coefficients and corresponding p-values. One-way ANOVA was used to compare the F area values of depolarizing and hyperpolarizing synapses to those of neurons fired alone (post only). Paired t-tests were used to compare E Cl and resting membrane potential values when those values were obtained from the same neurons. All other statistical analysis used unpaired t-tests. All statistical analysis was performed using SigmaStat 2.03.

results characterIzatIon and localIzatIon oF gaBaergic synapses
In order to examine Ca 2+ influx during STDP induction, we first had to locate synaptically connected neurons and characterize the synapse between them. We did this using dual perforated patchclamp recordings from pairs of hippocampal neurons cultured at a low density. Hippocampal cultures were prepared from E18 rats and recorded from after 8-13 days in culture; they contained glia, pyramidal neurons and GABAergic interneurons. Synaptic connections were identified by stimulating one neuron and monitoring for the presence of postsynaptic currents in the other. After we located a synapse, our first indication that it was GABAergic came from the relatively long time course of the GPSC in voltage-clamp mode (35.50 ± 3.37 ms for GPSCs, as opposed to 5.76 ± 0.64 ms for glutamatergic currents; Balena and Woodin, 2008). If the synapse had a time course consistent with GABAergic currents we then determined the E Cl by constructing an I-V curve; the intersection of the curve with the x-axis was taken to be E Cl (Figures 1A,B). Because the extracellular recording solution was free of HCO 3 − , E Cl ≈ E GABA . Based on the relation of E Cl to resting membrane potential, we characterized GABAergic synapses as either: (1) depolarizing, when E Cl was more positive than the resting membrane potential ( Figure 1A); or (2) hyperpolarizing, when E Cl was more negative than the resting membrane potential ( Figure 1B). Depolarizing GABAergic synapses had an average E Cl of −58.33 ± 2.63 mV, which was significantly different from the resting membrane potential of those neurons (−70.33 ± 1.17 mV; n = 12; paired t-test p < 0.001; Figure 1C). In all cases E Cl was hyperpolarizing with respect to action potential threshold, and thus depolarizing synapses were not excitatory. Hyperpolarizing synapses had an average E Cl of −75.87 ± 1.74 mV, which was significantly different from  (Figures 2A,B). A sequence of fluorescence images was then acquired at 1-2 Hz during STDP induction. GABAergic STDP was induced in current-clamp mode; both the pre-and postsynaptic neurons were induced to fire action potentials (using minimal stimulation) at a frequency of 5 Hz for 30 s. The interval between spike induction was ±5 ms, which resulted in a spiketiming interval of ±1 ms between onset of the GPSP and the postsynaptic action potential. This yielded a graph representing the changes in Fluo4 fluorescence (as a measure of Ca 2+ ) over time ( Figure 2C). This data was normalized to the baseline fluorescence level and expressed as a percentage increase. We analyzed three measures of the change in fluorescence during STDP induction: (1) ∆F 30 (%), the change in fluorescence from F 0 to F 30 ; (2) ∆F peak (%), the change in fluorescence from F 0 to F peak ; and (3) F area , the area their resting membrane potential values (−63.40 ± 2.44 mV; n = 10; paired t-test p < 0.001; Figure 1C). There was a significant difference between the E Cl values (p < 0.001) and resting membrane potential values (p = 0.014) of depolarizing and hyperpolarizing synapses. The slope of each I-V curve provided the GABA A conductance of that synapse. Depolarizing GABAergic synapses had a conductance of 1.85 ± 0.34 pS (n = 12), and hyperpolarizing synapses had a conductance of 1.74 ± 0.35 pS (n = 10); there was no significant difference between the conductances of the two populations (p = 0.824).

sIMultaneous stdp InductIon and IMagIng oF ca 2+ dynaMIcs
Following characterization of the GABAergic synapse we examined the Ca 2+ dynamics during the induction of STDP. To do this, we loaded neurons with the Ca 2+ -sensitive cell permeant fluorescent At glutamatergic synapses the order of spiking (pre/post versus post/pre) during STDP induction determines the polarity of the plasticity (LTP versus LTD, respectively; Bi and Poo, 1998). However, at GABAergic synapses the order of spiking during induction does not determine the nature of plasticity (as evidenced by the symmetrical spike-timing window Woodin et al., 2003). We thus hypothesized that Ca 2+ influx during STDP induction should be independent of the spike-timing order. We found no significant difference in Ca 2+ influx (F area ) between synapses induced with positive or negative spiketiming intervals; this was true for both depolarizing (p = 0.96) and hyperpolarizing synapses (p = 0.77; Figure 3C). We also found no significant difference between the time taken for the Ca 2+ influx to reach its maximum between pre/post and post/pre synapses (p = 0.968).

gaBaergic synapse propertIes deterMIne ca 2+ dynaMIcs durIng stdp InductIon
The Ca 2+ influx during STDP induction was affected by GPSC amplitude. GPSC amplitude is determined by both E Cl and GABA A conductance (Kaila, 1994); thus we asked which of these properties was the most influential in regulating postsynaptic Ca 2+ . We found that the linear regression of E Cl versus Ca 2+ influx (F area ) yielded low r-squared values [0.31 (p = 0.078) for pre/post synapses, 0.25 (p = 0.114) for post/pre synapses, and 0.26 (p = 0.015) for all synapses] ( Figure 4A). Low r-squared values indicate that the trend line does not accurately predict the relationship between E Cl and Ca 2+ influx. Thus, the difference between the slopes of the trend lines for pre/post and post/pre synapses does not necessarily indicate that the spike-timing order influences the Ca 2+ influx. E Cl also did not correlate strongly with ∆F 30 [pre/post r-squared = 0.26 (p = 0.114), post/pre r-squared = 0.27 (p = 0.104), all synapses r-squared = 0.26 (p = 0.017)] or ∆F peak [pre/post r-squared = 0.29 (p = 0.086), post/pre r-squared = 0.25 (p = 0.119), all synapses r-squared = 0.26 (p = 0.016)].

dIscussIon
At GABAergic synapses, the induction of STDP requires an increase in postsynaptic Ca 2+ (Woodin et al., 2003;Xu et al., 2008;Ormond and Woodin, 2009). Depending on the stage of nervous system under the curve normalized to F 0 . Increasing the image acquisition frequency to 100-200 ms intervals did not significantly change the ∆F peak for depolarizing GABAergic synapses (p = 0.898), indicating that the standard acquisition rate was sufficient to resolve the peak of the Ca 2+ fluorescence. We chose to exclusively examine the Ca 2+ influx at the soma of the postsynaptic neurons for two reasons. First, the majority of GABAergic neurons innervate the proximal dendrites and soma of postsynaptic neurons (Cobb et al., 1995;Freund and Buzsaki, 1996;Di Cristo et al., 2004;Huang, 2006). Second, we recently demonstrated that GABAergic STDP is induced by feed-forward interneurons which target the soma of pyramidal neurons in the CA1 region of the hippocampus (Ormond and Woodin, 2009).

gpsc polarIty deterMInes the MagnItude oF ca 2+ InFlux durIng stdp InductIon
Ca 2+ influx is required for GABAergic STDP (Woodin et al., 2003), but whether or not the dynamics of the influx depend on the properties of the synapse had not yet been determined. After recording from a population of neurons that included both depolarizing and hyperpolarizing GPSCs and imaging their Ca 2+ dynamics, we examined the relationship between GPSC amplitude and Ca 2+ influx during STDP induction. GPSC amplitude correlated strongly with Ca 2+ influx (F area ); the linear regression analysis yielded similarly high r-squared value regardless of the spike-timing order [pre/post r-squared = 0.66 (p = 0.002), post/ pre r-squared = 0.65 (p = 0.003), all synapses r-squared = 0.65 (p < 0.001); Figure 3A]. GPSC amplitude also correlated strongly with other measure of Ca 2+ influx, ∆F 30 [pre/post r-squared = 0.65 (p = 0.002), post/pre r-squared = 0.60 (p = 0.005), all synapses r-squared = 0.62 (p < 0.001)] and ∆F peak [pre/post r-squared = 0.62 (p = 0.004), post/pre r-squared = 0.63 (p = 0.004), all synapses r-squared = 0.62 (p < 0.001)]. Thus, regardless of the fluorescence measure examined, there was a strong relationship between Ca 2+ influx and the nature of the GABAergic synapse; depolarizing synapses correlated with large increases in Ca 2+ , while hyperpolarizing synapses correlated with smaller increases.
The correlation between Ca 2+ influx and GPSC amplitude was further quantified by comparing the fluorescence increase during STDP induction between all depolarizing synapses (E Cl = −58.33 ± 2.63 mV; n = 12), all hyperpolarizing synapses (E Cl = −75.87 ± 1.74 mV; n = 10), and neurons with no synapses (which we call "post only"; n = 11). The post only neurons were also stimulated at 5 Hz for 30 s. There were significant differences between the Ca 2+ influx at depolarizing and hyperpolarizing synapses regardless of the measure analyzed (F area p = 0.004; F peak p = 0.005; F 30 p = 0.004; Figure 3B). Ca 2+ influx at hyperpolarizing synapses were also significantly different from the influx at post only neurons that fired in the absence of a synapse when F area and F 30 were analyzed (p = 0.04 and p = 0.048, respectively); however when F peak was analyzed there was not a significant difference between hyperpolarizing synapses and post only neurons (p = 0.06). Thus, we can conclude that depolarizing neurons let in the same amount of Ca 2+ during STDP induction as the postsynaptic neurons spiking alone (independent of a synapse). However, when the GPSC becomes hyperpolarizing it has a strong ability to decrease the Ca 2+ influx.  medium, E GABA ≈ E Cl ; thus the driving force through the GABA A receptors was largely determined by E Cl . We found that neither the channel conductance nor the Cl − driving force alone could predict the relationship between GPSC amplitude and Ca 2+ influx. This indicates that it is the combination of these properties of GABAergic transmission that is important in regulating Ca 2+ influx during STDP induction. At glutamatergic synapses, positive and negative spike-timing intervals lead to long-term potentiation and depression, respectively, resulting in an asymmetric spike-timing window (Markram et al., 1997;Poo, 1998, 2001;Debanne et al., 1998;Zhang et al., 1998). In contrast, the spike-timing window for GABAergic synapses is symmetric, with coincident activity (within ±15 ms) resulting in decreased inhibition, independent of the spike-timing order (Woodin et al., 2003). This likely accounts for the nonsignificant differences in Ca 2+ -influx between positive and negative spike-timing intervals. However, the present study only examined spike-timing intervals of <5 ms; whether the results are similar for intervals >5 ms remains to be determined.
We identified a significant difference in the resting membrane potentials between depolarizing and hyperpolarizing synapses. This difference may result from our ability to identify GABAergic synapses where E Cl sits close to the resting membrane potential; if E Cl ≈ resting membrane potential there would be no driving force for Cl − and thus we did not characterize a synapse electrophysiologically. This may have biased our selection of synapses for those with larger driving forces; depolarizing synapses would be more likely to be found onto neurons with relatively hyperpolarized resting membrane potentials, and hyperpolarizing synapses would be common onto neurons with relatively depolarized resting membrane potentials.
Following acute neuronal trauma (van den Pol et al., 1996;Toyoda et al., 2003), oxygen-glucose deprivation (Galeffi et al., 2004), and seizure activity (Galanopoulou, 2007), there is a depolarization of E Cl which renders GABAergic transmission depolarizing (Fiumelli and Woodin, 2007;Kahle et al., 2008). Based on our present results, this switch in the polarity of GABAergic transmission increases the amount of Ca 2+ influx during subsequent neuronal activity. In fact, the magnitude of the E Cl depolarization following neuronal insults is so large it often renders GABAergic transmission excitatory (Kahle et al., 2008), which should produce even larger Ca 2+ influxes than those observed in the present study. This may be particularly relevant given that the large Ca 2+ influxes resulting from neuronal injury contribute to cell death (Bano and Nicotera, 2007).
Taken together, a model emerges where postsynaptic Ca 2+ influx is required for STDP induction at GABAergic synapses, and where the magnitude of this influx is regulated by the GABAergic transmission itself. Further work will be need to elucidate both how the Ca 2+ influx in turn regulates E Cl , and how the Ca 2+ influx is regulated when GABAergic and glutamatergic STDP are induced simultaneously.
acknowledgMents This work was supported by a Natural Sciences and Engineering Research Council of Canada (NSERC) Discover Grant to Melanie A Woodin, and a NSERC Fellowship to Trevor Balena. development GABAergic synapses can be either depolarizing or hyperpolarizing, which led us to hypothesize that the nature of GABAergic transmission regulates the magnitude of the Ca 2+ influx. Our present results found this hypothesis to be true; depolarizing GABAergic synapses characteristic of immature neuronal circuits produced larger Ca 2+ influxes during STDP induction than hyperpolarizing GABAergic synapses, which are more commonly found in the mature central nervous system. Our analysis further revealed that this relationship between GABAergic synapses and Ca 2+ influx can be accounted for by the two main properties of GABAergic synapses (the driving force for Cl − and the GABA A receptor conductance) but did not depend on the order of spike timing.
It is already well known that synaptic transmission contributes to the postsynaptic Ca 2+ influx during glutamatergic STDP induction (Markram et al., 1997;Poo, 1998, 2001;Debanne et al., 1998;Zhang et al., 1998). At these excitatory synapses, Ca 2+ influx occurs primarily via VGCCs opened by the back-propagating action potential, and via NMDARs opened by the coincident occurrence of the back-propagating action potential and postsynaptic glutamate binding. How can our understanding of the Ca 2+ influx during glutamatergic STDP induction be related to GABAergic STDP? It depends on the polarity of the GABAergic synapse. When E Cl sits below the resting membrane potential the GPSC will hyperpolarize the postsynaptic membrane during STDP induction, resulting in a smaller Ca 2+ influx than if that same postsynaptic neuron was spiking alone (at 5 Hz for 30 s, in the absence of synaptic transmission). This smaller Ca 2+ influx presumably results from hyperpolarizing GPSCs decreases the opening of VGCCs (which are activated by action potential firing during STDP induction). However, at depolarizing GABAergic synapses, the Ca 2+ influx during STDP induction for both positive and negative spike-timing intervals does not differ significantly from when the postsynaptic neuron spikes alone. This indicates that the additional depolarization is insufficient to open more VGCCs, either because the majority of available VGCCs have already been opened by the action potential, or because the magnitude of depolarization is not sufficient to open VGCCs.
We already know that the required Ca 2+ influx during GABAergic STDP occurs partly through L-type VGCCs (Woodin et al., 2003;Ormond and Woodin, 2009). However, this cannot be the only source of Ca 2+ influx because when GABAergic synapses are blocked or absent (post only) the same spiking pattern which also opens VGCCs fails to induce plasticity. This indicates that there are either additional sources of Ca 2+ influx required for STDP, or that a component of the GABAergic signaling combines with the L-type Ca 2+ influx to induce plasticity. We have preliminary evidence for the involvement of T-type VGCCs during hyperpolarizing GABAergic STDP (Balena and Woodin, 2009); these channels require membrane hyperpolarization to be removed from their inactive state but also require subsequent membrane depolarization to become activated (Magee et al., 1995;Perez-Reyes and Lory, 2006). Thus during hyperpolarizing GABAergic transmission we believe that the Ca 2+ influx occurs both through L-type VGCCs (which have a reduced opening compared to post only) and through T-type VGCCs.
The strength of GABAergic synapses depends upon both the conductance of the channel and on the driving force for ions flowing through the channel. Because we recorded in a HCO 3