HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) All Versions of this Article: 91/3/1337    most recent 00414.2003v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (16) Citing Articles via Google Scholar Google Scholar Articles by Shen, W. Articles by Surmeier, D. J. Search for Related Content PubMed PubMed Citation Articles by Shen, W. Articles by Surmeier, D. J.

Kv1.2-Containing K⁺ Channels Regulate Subthreshold Excitability of Striatal Medium Spiny Neurons

Department of Physiology and Institute for Neuroscience, Feinberg School of Medicine, Northwestern University, Chicago, Illinois 60611

Submitted 28 April 2003; accepted in final form 11 September 2003

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
A slowly inactivating, low-threshold K⁺ current has been implicated in the regulation of state transitions and repetitive activity in striatal medium spiny neurons. However, the molecular identity of the channels underlying this current and their biophysical properties remain to be clearly determined. Because previous work had suggested this current arose from Kv1 family channels, high-affinity toxins for this family were tested for their ability to block whole cell K⁺ currents activated by depolarization of acutely isolated neurons. -Dendrotoxin, which blocks channels containing Kv1.1, Kv1.2, or Kv1.6 subunits, decreased currents evoked by depolarization. Three other Kv1 family toxins that lack a high affinity for Kv1.2 subunits, r-agitoxin-2, dendrotoxin-K, and r-margatoxin, failed to significantly reduce currents, implicating channels with Kv1.2 subunits. RT-PCR results confirmed the expression of Kv1.2 mRNA in identified medium spiny neurons. Currents attributable to Kv1.2 channels activated rapidly, inactivated slowly, and recovered from inactivation slowly. In the subthreshold range (ca. -60 mV), these currents accounted for as much as 50% of the depolarization-activated K⁺ current. Moreover, their rapid activation and relatively slow deactivation suggested that they contribute to spike afterpotentials regulating repetitive discharge. This inference was confirmed in current-clamp recordings from medium spiny neurons in the slice preparation where Kv1.2 blockade reduced first-spike latency and increased discharge frequency evoked from hyperpolarized membrane potentials resembling the "down-state" found in vivo. These studies establish a clear functional role for somato-dendritic Kv1.2 channels in the regulation of state transitions and repetitive discharge in striatal medium spiny neurons.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
In response to temporally coherent glutamatergic synaptic input, striatal medium spiny neurons move from a hyperpolarized "down-state" to a depolarized "up-state" that is near spike threshold (Stern et al. 1997; Wilson and Groves 1981; Wilson and Kawaguchi 1996). Previous studies of these state transitions have highlighted the importance of intrinsic K⁺ channel currents in determining membrane potential range visited during these states (Wilson and Kawaguchi 1996). In the down-state, Kir2 family channels dominate the conductance profile of the medium spiny neuron (Mermelstein et al. 1998; Nisenbaum and Wilson 1995). When the membrane potential moves away from the K⁺ equilibrium potential in response to synaptic input, these channels close as Mg²⁺ and polyamines are swept into the pore (Ficker et al. 1994; Ishihara et al. 1989). Near -60 mV, another family of K⁺ channels activate and begin to brake the synaptically driven depolarization. One class of channel that activates quickly in this voltage range is composed of Kv4 family subunits (Tkatch et al. 2000). Early in development, these channels are the predominant depolarization-activated K⁺ channels in medium spiny neurons (Akins et al. 1990; Surmeier et al. 1988). However, these channels inactivate relatively rapidly, making them poor regulators of sustained depolarizing inputs of the sort seen in adult neurons. This role is played by another K⁺ channel that emerges later in development. This channel opens in a subthreshold voltage range but unlike Kv4 channels inactivates slowly. Studies in vitro have suggested that a channel with these properties helps set the membrane potential and discharge rate in the up-state (Nisenbaum et al. 1994). In many respects, this current resembles what has been referred to as the D-type current (Storm 1988) or the slowly inactivating A-type current (Albert and Nerbonne 1995; Locke and Nerbonne 1997; Surmeier et al. 1991). In several types of central neuron, this current slows the subthreshold ramp potential induced by depolarizing current injection and slows repetitive spike rate (Bekkers and Delaney 2001; Foehring and Surmeier 1993; McCormick 1991; Storm 1988, 1990).

The molecular identity (or identities) of this slowly inactivating K⁺ channel in medium spiny neurons has yet to be determined. However, the biophysical and pharmacological data gathered to date point to a Kv1 family channel. In addition to its activation at subthreshold membrane potentials and slow inactivation, two properties are telling. First, the slowly inactivating current is sensitive to 4-aminopyridine (4-AP) at micromolar concentrations and to -dendrotoxin (-DTX) (Nisenbaum et al. 1994, 1996). Second, recovery from inactivation is slow, having a recovery time constant greater than a second at -90 mV (Surmeier et al. 1991). Both features are similar to those of Kv1.1, Kv1.2, and Kv1.6 channels studied in heterologous expression systems (Coetzee et al. 1999). Several other lines of evidence that are consistent with this inference as members of the Kv1 family are increasingly being seen as key determinants of somatodendritic excitability in neurons. In peripheral sensory neurons, these channels help determine somatic spike threshold and spike frequency in response to depolarizing stimuli (Glazebrook et al. 2002). Similarly, in neurons of the medial trapezoid body, heteromers of Kv1.1, Kv1.2, and Kv1.6 subunits are important regulators of repetitive spiking (Dodson et al. 2002). Cortical pyramidal neurons also express an -DTX-sensitive current with profound effects on repetitive discharge near rheobase (Bekkers and Delaney 2001). The experiments described here were undertaken to pinpoint the molecular identity of the slowly inactivating K⁺ channel in striatal medium spiny neurons, to quantitatively characterize its gating properties, and to determine its impact on spiking. Our results show that although Kv1.1, Kv1.2, and Kv1.6 mRNA are expressed in identified medium spiny neurons, the pharmacological profile of the slowly inactivating current in the somatic and proximal dendritic membrane is that of a channel containing Kv1.2 subunits. The biophysical features of the channel currents are consistent with this identifi-cation. Moreover, in the subthreshold voltage range, this Kv1.2 channel current constitutes nearly half of the total somatic K⁺ current, leading to a major role in regulating spike threshold and repetitive discharge.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Acute dissociation procedure and slice preparation

Striata were dissected from young adult (3-4 wk) Sprague-Dawley rats, and medium spiny neurons were acutely dissociated using previously described procedure (Surmeier et al. 1992). The rats were anesthetized with isoflurane and decapitated. The brains were quickly removed and immersed in an ice-cold, oxygenated sucrose solution comprising (in mM) 250 sucrose, 4 glucose, 2.5 KCl, 1 Na₂HPO₄, 4 MgSO₄, 15 HEPES, 1 kynurenic acid, 0.1 N-nitro-L-arginine, and 0.005 glutathione (pH 7.4, 300 mOsm/l). The brains were blocked, and 300- to 350-µm-thick coronal sections were cut with a microslicer (VT1000, Leica). The slices were then incubated for 1-5 h at room temperature (22-24°C) in NaHCO₃-buffered Earl's balanced salt solution (EBSS). The holding solution was bubbled with carbogen (95/5) and supplemented with (in mM) 1 kynurenic acid, 0.1 N-nitro-L-arginine, and 0.005 glutathione (pH 7.4, 300 mOsm/l). After incubation, the dorsal striata from each hemisphere were dissected and transferred to oxygenated HEPES-buffered Hank's buffered salt solution (HBSS) containing 0.8 mg/ml protease (type XIV), where they were incubated for 25 min at 35°C. The HBSS solution was supplemented with (in mM) 1 pyruvic acid, 1 kynurenic acid, 0.1 N-nitro-L-arginine, and 0.005 glutathione (pH 7.4, 300 mOsm/l). After enzyme digestion, tissue was rinsed three times in a dissociation solution, and the neurons were dissociated by triturating using three progressively smaller fire-polished Pasteau pipettes. The dissociation solution was oxygenated and comprised (in mM) 140 sodium isethionate, 2 KCl, 4 MgCl₂, 23 glucose, 15 HEPES, 1 kynurenic acid, 0.1 N-nitro-L-arginine, and 0.005 glutathione (pH 7.4, 300 mOsm/l). The cell suspension was then placed into a 35-mm plastic petri dish that was mounted onto the stage of an inverted microscope (Nikon, Tokyo). After the neurons settled to the bottom of the dish, they were continuously perfused with a physiological saline. The background saline solution contained (in mM) 140 NaCl, 2 KCl, 2 MgCl₂, 1 CaCl₂, 23 glucose, and 15 HEPES (pH 7.4, 300 mOsm/l). All recordings were conducted at 22-24°C.

For slice preparation, 250 µM coronal striatal slices were made in an artificial cerebrospinal fluid (ACSF) containing (in mM) 125 NaCl, 2.5 KCl, 2 CaCl₂, 1 MgCl₂, 1.25 NaH₂PO₄, 24 NaHCO₃, and 10 glucose (pH 7.4 with 5% CO₂, 300 mOsm/l). The slices were then allowed to recover in ACSF for 1 h before recording.

Electrophysiology

Whole cell voltage and current recordings were performed using standard techniques (Baranauskas et al. 1999; Hamill et al. 1981). Electrodes were pulled from 7052 glass on a Sutter P-97 puller and fire-polished before use. For recordings from acutely dissociated neurons, the internal pipette solution contained (in mM) 120 KMeSO₄, 2 MgCl₂, 5 EGTA, 10 HEPES, 2 Na₂ATP, 0.2 Na₂GTP, 12 phosphocreatine, and 0.1 leupeptin; pH was adjusted to 7.2 with KOH and osmolarity to 270-280 mOsm/l. The external recording solution had the following (in mM): 140 NaCl or sodium isethionate, 3 or 0.5 KCl, 2 CaCl₂, 2 MgCl₂, 10 glucose, and 10 HEPES; pH was adjusted to 7.2 with NaOH and osmolarity to 300 mOsm/l. TTX (1 µM) and 0.1 mM CdCl₂ were added to block voltage-dependent Na⁺ and Ca²⁺ channels, respectively. Solutions were applied with a gravity-driven sewer pipe system. An array of application capillaries was positioned a few hundred micrometers from the cell under study. Solution changes were made by altering the position of the array with a PMC200-P2 motion controller and an 850B high-speed actuator (Newport, Irvine, CA). Solution change was complete within 1 s. For the current-clamp study in tissue slices, the internal pipette solution contained (in mM) 135 KMeSO₄, 5 KCl, 0.5 CaCl₂, 5 EGTA, 5 HEPES, 2 MgATP, and 0.5 Na₂GTP, pH 7.2 and 280-290 mOsm/l. Agents dissolved in bath solution were perfused with a rate of 2-3 ml/min.

Whole cell recordings were obtained with an Axopatch 200 or a Multiclamp 700A amplifier and monitored with a PC running Clampex 8 (Axon Instruments, Union City, CA). Electrode resistance in the bath was typically 1.5-2.5 M for recording of acutely isolated neurons and 3-4 M for slice recording. After achieving the whole cell configuration, series resistance (7-12 M in dissociated cells and 10-20 M in slices) was compensated (70-90%) and periodically monitored. A sample of 38 dissociated medium spiny neurons showed a mean whole cell capacitance of 7.6 ± 0.2 (SE) pF.

Data were analyzed with Igor Pro (Version 4, Wavemetrics, Lake Oswego, OR), Clampfit 8 (Axon Instruments, Union City, CA) or Mini Analysis (Synaptosoft, Decatur, GA). Spike shape parameters were measured using the Mini Analysis routines. Specifically, spike threshold was determined by locating the membrane voltage at which the second derivative of the spike waveform exceeded three times its SD in a period before the spike onset. Spike amplitude was defined as the difference between the peak and the baseline of action potential. Spike width was measured at half the spike amplitude. The afterhyperpolalization was defined as the difference between the spike baseline and the minimum membrane voltage after the action potential peak. For samples >10, central tendency was estimated by computing a mean (± SE); for smaller samples, medians were computed. Differences between samples were analyzed with a Student's t-test (large samples) or a paired Wilcoxon signed-rank test (small samples).

Computer simulation

Kv1.2 currents were modeled using NEURON (Hines and Carnevale 1997). Channel gating was assumed to conform to a Hodgkin-Huxley-like formalism: I = G_maxm²(t,V)[(a)h(t,V) + (1 - a)](V - E_k), where I is current, G_max is the maximum conductance, V is the transmembrane voltage, E_k is the K⁺ equilibrium potential, m is the gating particle to control activation, and h is the gating particle to control inactivation and a is the fraction of the current that inactivates. Both m and h were functions of time and transmembrane voltage that satisfied the partial differential equations

where

This formalism was coded in a NEURON mod file and then virtual channels were distributed in the membrane of a spherical neuron (diameter = 20 µm). To generate estimates of the parameters governing activation and inactivation processes, three types of experiment were performed. First, the rising phase of -DTX-sensitive currents evoked by voltage steps in acutely isolated neurons were fit with the model. This was done using Igor Pro. These gave reasonable estimates of _m at depolarized membrane voltages (>-30 mV) where the current was readily resolvable. The deactivation time constant at -60 mV was estimated from mono-exponential fits to -DTX-sensitive tail currents. Next, the recovery kinetics of the -DTX-sensitive currents at -90 mV were estimated using voltage ramps. Third, voltage ramps of varying speed were used to generate an estimate of inactivation voltage dependence and kinetics. -DTX-sensitive ramp currents were generated in dissociated neurons with biphasic ("up and down") ramps of three different speeds. The parameters of the ramp protocol initially were chosen to span a range of kinetics thought to be relevant to Kv1.2 channels at room temperature. These ramps were then tested in a NEURON simulation for their ability to generate data that allowed NEURON's parameter estimation utility (Multirun Fitter) to reliably converge to accurate estimates of the model's starting values even when it was given initial estimates that deviated significantly from these values. Based on these simulations, voltage-clamp protocols were designed for use with real neurons as described in the following text.

The -DTX-sensitive currents generated in these experiments were then used by NEURON's Multirun Fitter to generate parameter estimates for the inactivation process; the activation process was constrained by the parameter estimates derived from the step data. The model was able to reasonably simulate the response to the ramp currents. For activation, the parameter estimates derived from this curve fitting procedure were: V_mh = -27 mV; V_mc = -16 mV; _m0 = 3.4 ms; C_m = 89.2; V_hm = -34.3 mV; V_cm = 30.1 mV; for the inactivation process, the parameters were: V_hh = -33.5 mV; V_hc = 21.5 mV; C_h = 548.7; V_h1 = -0.96 mV; V_c1 = 29.01 mV; V_h2 = -0.96 mV; V_c2 = 100 mV. To estimate Kv1.2 currents flowing during repetitive spike activity, whole cell recordings from medium spiny neurons in a slice preparation were ported to NEURON and used to drive a simulated voltage-clamp electrode attached to a spherical neuron containing modeled Kv1.2 channels. To correct for the difference in temperature, _h(V_m) and _m(V_m) functions were corrected using a Q₁₀ of 3 (Hille 2001).

Single-cell RT-PCR analysis

Isolated individual neurons were patched in the cell-attached mode and lifted into a stream of the physiological saline. Neurons were then aspirated into the electrode. Electrodes contained 1-2 µl of diethylpyrocarbonate (DEPC)-treated water. The glass used for making electrodes was heated to 180°C for 2-3 h. Sterile gloves were worn at all time during the procedure to minimize RNase contamination. After aspiration of the neuron, the electrode was removed from the holder, the tip was broken in a 0.5-ml Eppendorf tube containing 3.6 µl of DEPC-treated water, 0.7 µl of RNasin (28,000 U/ml), and 0.7 µl of oligo-dT or random hexamer (0.5 mg/ml), and the contents were ejected. The mixture was heated to 70°C to denature the nucleic acids and then placed on ice for >1 min. Single-strand cDNA was synthesized from the cellular mRNA by adding SuperScript II RT (1.2 µl; 200 U/ml), 10x PCR buffer (1 µl), MgCl₂ (2 µl; 25 mM), DTT (2 µl; 0.1 M), and mixed dNTPs (1 µl; 10 mM) and then incubating the mixture at 42°C for 120 min. The reaction was terminated by heating the mixture to 70°C for 15 min. The RNA strand in the RNA-DNA hybrid was then removed by adding 0.5 µl RNase H (2 U/ml) and incubating at 37°C for 20 min. All reagents except RNasin (Promega, Madison, WI) were obtained from Invitrogen (Carlsbad, CA).

The single-cell cDNA generated from the reverse transcription step was subjected to conventional PCR using a programmable thermal cycler (P-200, MJ Research, Watertown, MA). PCR primers were developed from GenBank sequences with commercially available OLIGO software (Version 6.6, National Biosciences, Plymouth, MN). Primers for substance P (SP), enkephalin (ENK), Kv1.1, and Kv1.2 have been described previously from this laboratory (Song et al. 1998; Surmeier et al. 1996). Kv1.3 mRNA (GenBank Accession No. X16001 [GenBank] ) was detected with a pair of primers 5'-GCT GCC AGC ACC TCT CCT C (position 1368) and 5'-AAC CAG GCA CCT GTT TCA GTA GA (position 1740), which give a PCR product of 396 base pairs (bp). Kv1.6 mRNA (GenBank Accession No. X17621 [GenBank] ) was detected with a pair of primers 5'-ACC TGA AGG CAA CGG ACA ATG (position 1896) and 5'-ACC CAG GCA TGA GAA CTC AGT CTT (position 2322), which give a PCR product of 450 bp. Because of the apparent low abundance of mRNA for Kv1.2, two-round PCR using nested primers was designed. For the first-round outer primers, the upper primer was 5'-GTT GTC CAA AGA TCC CGT CCT C (position 1863), and the lower primer was 5'-GCT GCA TTG TGC CTT TCT GAG A (position 2304). The inner primers were 5'-CAA CTG TAC CTT GGC TAA CAC AAA (position 2002) and 5'-CTC CAT GGG TGA CTC TCA TCT T (position 2276), yielding a predicted PCR product of 296 bp. "Touch-down" protocols were implemented for more efficient amplification of single-cell cDNA. This was done by modifying a standard 45-cycle PCR protocol in the following way: first, 35 cycles were run at the optimal annealing temperature for each primer set, then the annealing temperature was decreased by 1 for two cycles—five times—for a total of 45 cycles.

For an assessment of developmental change in Kv1.2 mRNA, serial dilutions of striatal cDNA from different ages were made to determine the detection threshold for Kv1.2 and -actin. The detection threshold for Kv1.2 cDNA in each sample was then normalized to that for -actin. The Kv1.2 expression level is presented as a percentage of -actin.

PCR procedures were performed using procedures designed to minimize the chance of cross-contamination (Cimino et al. 1990). Negative controls for contamination from extraneous and genomic DNA were run for every batch of neurons. To ensure that genomic DNA did not contribute to the PCR products, neurons were aspirated and processed in the normal manner except that the reverse transcriptase was omitted. Contamination from extraneous sources was checked by replacing the cellular template with water. Both controls were consistently negative in these experiments.

Chemicals and drugs

-Dendrotoxin (-DTX), r-agitoxin-2 (AgTX), dendrotoxin-K (DTX-K), r-margatoxin (MgTX), and TTX were obtained from Alomone Labs (Jerusalem, Israel). Bovine serum albumin (BSA, 0.1%) was added to -DTX-, AgTX-, DTX-K-, and MgTX-containing solutions (Martina et al. 1998). To control for any effects of BSA, it was included at equal concentrations in all of external solutions in experiments using these toxins. GTP, leupeptin, and ATP were purchased from Calbiochem (La Jolla, CA). All other chemicals were obtained from Sigma Chemicals (St. Louis, MO).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Pharmacological profile of the slowly inactivating K⁺ channels implicates Kv1.2 channels

-DTX has a high affinity for three Kv1 family subunits— Kv1.1, Kv1.2, and Kv1.6 (Coetzee et al. 1999). These subunits can form homomeric or heteromeric ion channels, which retain their sensitivity to -DTX. To determine if similar channels were expressed in medium spiny neurons, -DTX and other specific Kv1 subunit blockers were examined for their ability to block depolarization-activated K⁺ currents. To this end, voltage-clamp recordings were made from acutely isolated medium spiny neurons (Baranauskas et al. 1999). To isolate K⁺ currents, voltage-dependent Na⁺, Ca²⁺, and Ca²⁺-dependent K⁺ channels were blocked by addition of 1 µM TTX and 0.1 mM CdCl₂ to the perfusion solution. After a prepulse of 2 s to -90 mV to diminish resting voltage-dependent inactivation, currents were evoked by a depolarizing voltage step to +30 mV for 1 s. These currents decayed bi-exponentially (Fig. 1A). Application of -DTX at concentrations from 0.4 to 1 µM reduced current amplitude both at the peak of the transient and at latter time points in the current trace, suggesting that a rapidly activating, slowly inactivating current was being blocked (Fig. 1A). This inference was confirmed by examination of difference currents (Fig. 1A, bottom). -DTX (1 µM) reduced the current measured at the end of the step to +30 mV by 15.1 ± 1.9% (n = 18, P < 0.01, Fig. 1B).

View larger version (14K): [in this window] [in a new window]   FIG. 1. -Dendrotoxin (-DTX) blocks a component of depolarization activated K+ currents. A: K+ current traces in control conditions and in the presence of external -DTX. Currents were evoked by depolarizing test pulses to +30 mV after a prepulse of 2 s to -90 mV. Subtraction of current trace obtained in the presence of -DTX from in control gives rise to the difference current in A, bottom. B: percentage blockade by -DTX, r-agitoxin-2 (AgTX), dendrotoxin-K (DTX-K), r-margatoxin (MgTX). *P < 0.01.

Medium spiny neurons express Kv1.2 mRNA

If the pharmacological analysis is correct, medium spiny neurons should express Kv1.2 mRNA. Previous studies have shown Kv1.2 mRNA is present in the striatum (Scott et al. 1994), an observation confirmed by tissue level RT-PCR analysis (data not shown). To test the hypothesis that medium spiny neurons express Kv1.2 mRNA, single-cell RT-PCR profiling was performed. Initially, 15 medium spiny neurons—identified by their expression of enkephalin (ENK) or substance P (SP) mRNA—were examined using a single round of amplification. Kv1.1, Kv1.2, Kv1.3, and Kv1.6 mRNAs were examined. All four transcripts were readily detected at the tissue level. But, in individual medium spiny neurons the expression pattern was more restricted. Kv1.3 subunit mRNA was not detected in identified medium spiny neurons. On the other hand, the other transcripts were frequently evident (Fig. 2A). Kv1.1 was the most readily detected, being seen in >90% of the neurons profiled. Kv1.2 was less readily detected, being seen in 40% of the neurons with a single round of amplification.

View larger version (53K): [in this window] [in a new window]   FIG. 2. Striatal medium spiny neurons express Kv1.2 subunit mRNA. A: single-cell RT-PCR (scRT-PCR) results indicate that from this sample of medium spiny neurons (n = 15), 93% had detectable levels of Kv1.1 mRNA, 40% had detectable levels of Kv1.2 mRNA, and 27% had Kv1.6 mRNA, with none of the cells having detectable levels of Kv1.3 mRNA. B: with a 2-round PCR analysis, most of the cells (25/28) had either detectable levels of Kv1.2 and substance P (SP, left panel of gel) or Kv1.2 and enkephalin (ENK, right panel of gel). C: expression of Kv1.2 is developmentally regulated. A threefold change (arbitrary ratio to -actin, see METHODS section) was detected after the 1st 4 postnatal weeks.

To provide an additional test of the involvement of Kv1.2 channels, a developmental profile of Kv1.2 expression was generated. The depolarization activated K⁺ currents in early postnatal medium spiny neurons are dominated by Kv4, rapidly inactivating channels (Surmeier et al. 1988). The slowly inactivating K⁺ currents emerge in the third and fourth postnatal weeks. If this current is attributable to Kv1.2 channels in whole or part, Kv1.2 mRNA levels should increase in parallel. Semi-quantitative RT-PCR analysis of Kv1.2 expression revealed a roughly threefold increase in the relative levels of Kv1.2 mRNA abundance between the first and fourth postnatal weeks (Fig. 2C).

Kv1.2 channel currents activate at subthreshold membrane potentials

If slowly inactivating, Kv1.2 channels regulate synaptic integration and repetitive discharge, then these channels should be responsible for a significant portion of the total K⁺ current at subthreshold membrane potentials. This hypothesis was examined first using voltage steps (1 s) from hyperpolarized membrane potentials (-90 mV). The outward currents evoked by steps above -60 mV were reduced by -DTX (1 µM; Fig. 3A). Subtraction of the -DTX-sensitive currents revealed them to be activated at subthreshold potentials (ca. -55 mV). The voltage dependence of activation was assessed by converting the current amplitudes to chord conductances and then normalizing (Fig. 3B). Data points were fitted with a second-order Boltzmann function. The -DTX-sensitive, Kv1.2 channel currents had an average half-activation voltage of -27.4 ± 2.5 mV (n = 6) and a relatively shallow slope factor (16.3 ± 1.8 mV). By contrast, the residual currents had a more depolarized half-activation voltage (-15.3 ± 2.2 mV); slope factors were similar to the Kv1.2 currents (16.8 ± 1.7 mV; Fig. 3B). The kinetics of activation were determined by fitting a second-order Hodgkin-Huxley model to the rising phase of the Kv1.2 currents (see METHODS; Fig. 3C). Activation time constants were voltage dependent, decreasing over the potential range where they could be accurately measured from the step currents (Fig. 3D). For example, at 0 mV, the average activation time constant was 27.8 ± 2.4 ms (n = 6), whereas at +45 mV, it was 2.8 ± 0.4 ms. These data were pooled with the deactivation time constant estimate at -60 mV to generate a closed form approximation of the relationship between the activation kinetics and voltage. The data were reasonably well fit with commonly used kinetic formula (see METHODS) plotted as a solid line in Fig. 3D.

View larger version (26K): [in this window] [in a new window]   FIG. 3. Kv1.2 channel currents activate rapidly and inactivate slowly. A: families of K+ currents obtained in the absence and presence of 1 µM -DTX and the subtraction currents in the lower panel. Currents were evoked during 1-s voltage steps between -75 and +45 mV (15-mV increments) after prepulses of 2 s to -90 mV. The voltage-clamp protocol is illustrated at the top. B: normalized conductance curves were plotted as a function of membrane potentials. The data were fit to a Boltzmann function (see METHODS). The fits were V1/2 = -16.7 mV and k = 17.2 mV for the curve after application of -DTX; V1/2 = -28.4 mV and k = 17.4 mV for the curve of -DTX-sensitive current. C: the rising phases of the Kv1.2 currents obtained between 0 and +45 mV were extended for clarity and fit with a modified Hodgkin-Huxley formalism. —, the fitted curves. D: average time constants (tau) of activation derived from curve fits (as in C) are plotted. The deactivation time constant at -60 mV is from Fig. 4B. The function fit to the data are described in METHODS.

View larger version (25K): [in this window] [in a new window]   FIG. 4. Kv1.2 channel currents accounts for half of subthreshold K+ current. A: activation curves obtained by recording membrane current responses to a slowly depolarizing voltage-ramp (130 mV/s from -90 to +40 mV, inset) in the absence and presence of 1 µM -DTX. B: -DTX-sensitive currents obtained by subtraction of the traces shown in A. Inset: tail currents obtained by stepping to -60 mV; the deactivation time constant was estimated by fitting the tail currents with a single exponential function with a time constant of 56 ms (—). C: normalized conductance estimates generated from the currents in A and B were plotted as a function of membrane potentials. The data were then fit with a Boltzmann function as described in METHODS. The fits were V1/2 = -7.5 mV and k = 12.2 mV for the -DTX-insensitive conductance; V1/2 = -21.9 mV and k = 15.5 mV for the -DTX-sensitive conductance. D: percentage of current that was attributable to -DTX-sensitive, Kv1.2 channels over a voltage range of -60 and +40 mV.

In addition to -DTX, Kv1.2 channels are readily blocked by 4-aminopyridine (4-AP) at micromolar concentrations (Russell et al. 1994). Application of 100 µM 4-AP reduced currents evoked by the slow voltage ramp by approximately the same amount as did -DTX. At -20 mV, 4-AP reduced the current amplitude by nearly half (48.3 ± 2.1%, n = 4). Holding the membrane at -60 mV, to inactivate Kv1.2 channels, reduced the block by 4-AP to 14% (±1.8%, n = 4). The block by -DTX was similarly reduced by holding at -60 mV (8.7 ± 1.2%, n = 4).

Kv1.2 channels recover slowly from inactivation

The pharmacological and biophysical data presented thus far are consistent with the attribution of the -DTX-sensitive current to Kv1.2 channels. Another distinctive feature of these channels (Coetzee et al. 1999; Sprunger et al. 1996) is their slow recovery from inactivation. To determine if the channels in medium spiny neurons had similar properties, channels were inactivated by holding at -40 mV for 4 s and then stepped to -90 mV for varying periods of time before delivering a voltage ramp (Fig. 5A). To isolate the Kv1.2 current, the first conditioning period (25 ms) was used as a control, assuming that there would be little or no recovery in a few milliseconds. Ramp currents were aligned along a voltage axis and the amplitude measured at -20 mV (Fig. 5B). Plots of the relative current amplitude as a function of the duration of the prepulse to -90 mV were readily fit with a mono-exponential function. The average recovery time constant was 1,028.8 ± 108.2 ms (n = 5, Fig. 5C). This value is very similar to that expected of a Kv1.2 channel (Coetzee et al. 1999; Sprunger et al. 1996).

View larger version (25K): [in this window] [in a new window]   FIG. 5. Kv1.2 channels recover slowly from inactivation. A: recovery kinetics were measured with ramp currents that were proceeded by a 4-s prepulse to -40 mV and a conditioning step to -90 mV of variable duration. B: ramp currents at each recovery interval were aligned along the ramp voltage axis. The current amplitudes at -20 mV (arrow) were normalized to the amplitude after a 25-ms conditioning step, and the average normalized currents were plotted as a function of conditioning step duration in C. Data were well fit with a single exponential having a time constant of 987 ms.

The biophysical data presented thus far are not sufficient to make an adequately constrained model of Kv1.2 channel gating. The principal obstacle to the construction of such a model was the apparent variation in inactivation kinetics estimated from long voltage steps. The problem with these step protocols is that they are very slow, taking tens of minutes to complete over an appropriate voltage range. As a consequence, rundown (or other alterations in channel gating) can contaminate difference currents, leading to random variation in the kinetic estimates. To overcome this obstacle, voltage ramps of varying speed were used to evoke currents. These ramps could be completed in a few seconds (rather than minutes), minimizing the impact of channel instability on parameter estimation. Three ramp speeds were used that preliminary simulation work had shown adequately constrained the inactivation parameters of a modified Hodgkin-Huxley formalism. An example of the -DTX-sensitive currents evoked by these ramp protocols is shown in Fig. 6A. Note that the peak current amplitude decreased with increasing ramp durations, providing a measure of channel inactivation. Plots of the percent contribution of Kv1.2 channels to the total K⁺ current for the three ramps are shown as an inset. For the intermediate ramp speed, the contribution of Kv1.2 is very similar to what was shown earlier (Fig. 4). However, for faster ramps it is clear that Kv1.2 makes a more substantial contribution to the total current at more depolarized membrane potentials (blue trace). Conversely, for very slow ramps, the contribution of Kv1.2 channels diminishes across a broad range of potentials, reflecting voltage-dependent inactivation of these channels.

View larger version (39K): [in this window] [in a new window]   FIG. 6. Kv1.2 channel kinetics were simulated using a modified Hodgkin-Huxley model. A: -DTX-sensitive Kv1.2 channel ramp currents were estimated by subtraction of control currents evoked by voltage ramps of 3 different speeds from those evoked in the presence of -DTX (1 µM). The percentage contribution of Kv1.2 channels to total K+ currents for each ramp is illustrated inset. B, left: plots of the activation (m) time constant (taum), and steady-state (m) values as a function of membrane potential. Right: corresponding plots inactivation (h) time constant (tauh) and steady-state (h) values as a function of membrane potential. C: simulated Kv1.2 current (red line) captures inactivation properties of the Kv1.2 channel population as represented by 3 -DTX-sensitive currents evoked by the same membrane potential step. D: -DTX-sensitive Kv1.2 channel ramp currents were well fit with the Kv1.2 channel model (green line). E: simulated currents evoked by progressively more depolarized 1-s voltage steps from a holding potential of -90 mV (protocol is shown at the top). F: simulated currents evoked by a standard inactivation protocol having 5-s prepulses between -110 mV and -10 mV followed by a test step to +20 mV. Note the similarity between the simulated currents and those derived from whole cell recording.

Kv1.2 channels regulate first spike latency and repetitive discharge in medium spiny neurons

From hyperpolarized membrane potentials, rheobase current injection in medium spiny neurons produces a slow voltage ramp to the first spike (Nisenbaum et al. 1994). An example of this phenomenon is shown in Fig. 7A (black line) where the whole cell recording was taken from a medium spiny neuron in a tissue slice preparation. To estimate the contribution of Kv1.2 channels to this behavior, the recording was used as a voltage command in a NEURON simulation of Kv1.2 currents at 33°C (Q₁₀ = 3). The Kv1.2 channel current evoked by this voltage trajectory is shown as the blue line in Fig. 7A. This simulation shows several interesting features of the current. First, it is significantly activated during the approach to the first spike, suggesting it plays an important role in shaping the voltage trajectory. As shown in the preceding text, in this voltage range, Kv1.2 channels are a major component of the total K⁺ current. Second, the Kv1.2 channels are efficiently activated by the spike and deactivate relatively slowly, leading to substantial activation during the interspike period. This suggests that the fast afterhyperpolarization (fAHP) of the medium spiny neuron is likely to be strongly influenced by Kv1.2 channels. This can be seen more clearly in Fig. 7B, where the current and voltage traces have been scaled up.

View larger version (42K): [in this window] [in a new window]   FIG. 7. Kv1.2 channels prolong 1st spike latency and slow repetitive discharge. A: black line is the voltage response of a medium spiny neuron recorded in a slice to a near rheobase current injection. Blue line is the simulated Kv1.2 current evoked by the same voltage trajectory. Note that the Kv1.2 current deactivated slowly and remained activated during interspike period. B: the same data as in A at higher gain to illustrate the Kv1.2 channel currents during the interspike period. C: after bath application of 1 µM -DTX, the 1st spike latency of the medium spiny neuron shown in A was shortened, the fast afterhyperpolarization (fAHP) reduced and the repetitive discharge rate increased, as suggested by the Kv1.2 model. D: action potentials (APs) evoked by a symmetric 4-s-long current ramp shown at the top of A in the same neuron before and after application of -DTX (1 µM). Note that -DTX shortened the latency to the first AP and increased the repetitive firing. E: instantaneous frequency vs. the applied currents, measured during current ramps. Opened circles are measurements from the control data; solid circles from -DTX data. Data were fit to a simple linear equation. Note that -DTX increases the intercept of fitting line without a change in the slope. F: mean time to the 1st spike vs. the ramp slopes applied. -DTX shortened the time required to fire the 1st AP. G: mean number of spikes elicited by different current ramps to the values in the form of ramp slope. -DTX reduced the amount of current required to fire APs.

To characterize the role of Kv1.2 channels on the response to a time varying, suprathreshold current injection, ramp currents were applied to medium spiny neurons recorded in the slice. In the presence of -DTX, spike threshold— defined by the abrupt increase in dV/dt— decreased by almost 2 mV (1.9 ± 0.11 mV; from -45.07 ± 1.01 mV in control to -47.01 ± 0.99 mV in -DTX, n = 7, P < 0.02, paired Wilcoxon signed-rank test). As predicted by the slow deactivation kinetics of Kv1.2 current (Fig. 7B), the fAHP after the first action potential was reduced by 3.45 ± 0.36 mV (-6.78 ± 0.46 mV in control; -3.32 ± 0.20 mV in -DTX, n = 7, P < 0.02, paired Wilcoxon signed-rank test). There were no alterations in spike amplitude (90.05 ± 1.95 mV in control vs. 92.14 ± 2.15 mV in -DTX, n = 7, P > 0.05, paired Wilcoxon signed-rank test) or spike half-width (1.95 ± 0.04 ms in control vs. 1.94 ± 0.04 ms in -DTX, n = 7, P > 0.05, paired Wilcoxon signed-rank test). In addition, the application of -DTX (1 µM) consistently shortened first spike latency and increased instantaneous spike frequency (Fig. 7, D-F). The increased discharge frequency was seen throughout the ramp as shown in the plot of frequency as a function of current amplitude (Fig. 7E); -DTX shifted the relation to the left without significantly altering the slope of the relationship (P > 0.05, Wilcoxon, n = 5). First spike latency was shortened by -DTX the same amount for a broad range of ramp speeds as expected for the rapid activation kinetics of the Kv1.2 current (Fig. 7F). The number of spikes evoked during each ramp was increased, shifting the relationship between spike number and ramp slope upward without changing the linearity of the plot or its slope (P > 0.05, Wilcoxon, n = 5, Fig. 7G).

Another important feature of Kv1.2 channels is that prolonged residence at depolarized membrane potentials leads to inactivation. This property has been shown to underlie a form of short-term response plasticity (Turrigiano et al. 1996). In vivo, medium spiny neurons move between hyperpolarized membrane potentials (ca. -85 mV), where perhaps 30% of Kv1.2 channels should be inactivated, and more depolarized membrane potentials (ca. -55 mV), where 70% of the channels should be inactivated at steady state. A plot modeled Kv1.2 channel availability (h state variable) as a function of time after voltage steps to -60, -50, and -40 mV from a holding potential of -80 mV is shown in Fig. 8A (the simulation was run assuming a temperature of 33°C). The kinetics of change were slow, taking >20 s to stabilize. As predicted by this result, changing the somatic membrane potential to -60 mV in medium spiny neurons in the slice by current injection had an impact on first spike latency and discharge rate that was similar to that of -DTX (Fig. 8B, cf. Fig. 7C). Although smaller, Kv1.2 current is likely to still to make a contribution to the membrane potential trajectory as judged by currents generated by the model when the current-clamp waveform was used a voltage command (blue line, Fig. 8B). To test this inference, -DTX (1 µM) was applied to medium spiny neurons held at -60 mV and then the same current step applied. Now, -DTX did not significantly reduce first spike latency (51 ± 12 ms in control; 34 ± 4 ms in -DTX, n = 4, P > 0.05, Wilcoxon) but did slightly increase discharge frequency (136 ± 14%, n = 4, P < 0.05, Wilcoxon) and attenuate the fAHP (6.9 ± 0.39 mV in control vs. 3.7 ± 0.47 mV in -DTX, n = 4, P < 0.05, Wilcoxon) in agreement with the modeling results (Fig. 8C). These results suggest that prolonged residence in the up-state could lead to substantial inactivation of Kv1.2 channels and enhanced excitability. But recordings from medium spiny neurons in anesthetized animals suggest that up-state durations are unlikely to last 10-20 s (Wilson and Kawaguchi 1996). Spiking could make a difference in the extent of inactivation. This hypothesis was tested with the model using a brief (2 ms) depolarization to +20 mV to simulate a spike. The model's membrane potential was stepped from -80 to -55 mV and spike-like waveforms delivered at varying frequencies. Channel availability was tracked (h state variable) during the course of the simulation. At repetitive frequencies <20 Hz, spiking had virtually no impact on channel availability (Fig. 8D). Above this frequency, spiking had a progressively larger impact, increasing the inactivation of Kv1.2 channels. Simulation runs using 10-, 20-, 50-, and 100-Hz repetitive rates are shown in Fig. 8D. Although the inactivation of Kv1.2 channels increased with high-frequency spiking, the effect was very modest with the total inactivation increasing by just over 10% with 20 s of 100-Hz spiking.

View larger version (32K): [in this window] [in a new window]   FIG. 8. Prolonged depolarization leads to inactivation of Kv1.2 currents. A: Kv1.2 channel availability (h state variable) progressively decreases as membrane voltage steps to more depolarized potentials. B: black line is the voltage response of a medium spiny neuron in the slice to near rheobase current injection when the soma was depolarized to near -60 mV. Blue line is the simulated Kv1.2 current evoked by this voltage trajectory from a holding potential of -60 mV. C: the response of the neuron depicted in B after application of -DTX (1 µM). D: Kv1.2 channel availability is plotted as a function of different stimulating frequencies (10-100 Hz). Note that channel availability is not changed significantly at spiking frequencies <20 Hz and availability drops by only 10% over the range of 10-100 Hz.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

Thus all four lines of evidence point to channels containing Kv1.2 subunits as playing a major role in subthreshold integration and repetitive activity in medium spiny neurons. A similar finding has recently been reported in cortical pyramidal neurons (Bekkers and Delaney 2001). As in pyramidal neurons, Kv1.2 channels in medium spiny neurons contribute a relatively small proportion of the total K⁺ current (ca. 10-20%) at very depolarized membrane potentials. In spite of this fact, blockade of these channels has a major impact. In part, this is attributable to their strategic location in the spike generating region (Bekkers and Delaney 2001; Dodson et al. 2002). But another factor that governs their functional role is the voltage range in which they open. As shown here, in the voltage range just below spike threshold, Kv1.2 channel currents constitute a much larger fraction of the total K⁺ current— approaching half of the total current. This is particularly true for intermediate speed voltage trajectories in medium spiny neurons where channel inactivation is not prominent. Together these features—strategic location and subthreshold activation—make Kv1.2 channels particularly important contributors to the regulation of excitability.

Do other Kv1 family subunits contribute to -DTX-sensitive channels in medium spiny neurons? Our scRT-PCR experiments revealed that medium spiny neurons express (in addition to Kv1.2) readily detectable levels of Kv1.1 and Kv1.6 mRNA but not Kv1.3 mRNA. The ease with which these subunits were detected cannot be taken as an indication of mRNA abundance because of uncertainties in the efficiency of the reverse transcription and amplification reactions. Quantitative inferences will require further study. What can be said at this point is that medium spiny neurons express mRNA for all three Kv1 subunits. These subunits are known to form heteromers in vivo (Dodson et al. 2002; Koch et al. 1997; Wang et al. 1999a). However, as heteromers retain their toxin sensitivity (Dodson et al. 2002; Koch et al. 1997) the inability of DTX-K, MgTX or AgTX to block appreciable K⁺ current indicates that Kv1.1 and Kv1.6 subunits do not participate in the somatic/proximal dendritic Kv1 channels in medium spiny neurons. This conclusion is also supported by the ability of -DTX to block current in the presence of concentrations of TEA that block channels containing Kv1.1 and/or Kv1.6 subunits. Where then do channels containing these subunits reside if not in the somatic membrane? Kv1.1/Kv1.6 (and Kv1.2) subunits are frequently found in axons or axon terminals (Gu et al. 2003; Monaghan et al. 2001; Rasband et al. 1998; Scherer 1999; Shamotienko et al. 1997; Southan and Robertson 2000; Wang et al. 1994), and the terminal fields of medium spiny neurons display immunoreactivity for Kv1.1 protein (Scott et al. 1994). It is also possible that channels containing Kv1.1 or Kv1.6 subunits are found in more distal dendritic regions. However, there is little immunocytochemical support for this proposition and electrophysiological studies suggest that Kv1 family channels are not prominent in these regions in other types of neuron (Golding et al. 1999).

Is D-type current attributable to Kv1 channels?

The biophysical and pharmacological properties of the Kv1 channels in medium spiny neurons strongly resemble those attributed to D-type current in these and other brain neurons (Bekkers and Delaney 2001; Golding et al. 1999; Nisenbaum et al. 1994, 1996; Storm 1988; Surmeier et al. 1991). Given the striking similarities, it is tempting to postulate that all D-type currents are attributable to Kv1 channels. Variations in the subunit composition of Kv1 channels and inclusion of modifying subunits (Nakahira et al. 1996; Rhodes et al. 1997) may account for the described variation in biophysical properties of D-type current (Bekkers and Delaney 2001; Foehring and Surmeier 1993; Southan and Robertson 2000; Storm 1988). Variation in subunit composition may not only explain relatively minor differences in the properties of D-type currents but could explain how Kv1 family channels also give rise to delayed rectifiers (Hille 2001).

Kv1.2 channels control spike generation and repetitive activity

Kv1.2 channels in medium spiny neurons effectively play roles ascribed to both D-type channels and delayed rectifiers. That is, the Kv1.2 channel currents effectively delayed first spike latency and contributed to spike afterhyperpolarization and slowing of repetitive discharge. The delay in first spike latency in medium spiny neurons with near rheobase current injection has been associated with a 4-AP-sensitive current previously (Kita et al. 1985; Nisenbaum et al. 1994). Although medium spiny neurons also express Kv4 channels that display a modest affinity for 4-AP (Tkatch et al. 2000), the ability of -DTX to mimic the effects of 4-AP argues that Kv1.2 channels are largely responsible for these earlier observations. The delay in first spike latency was seen for a broad range of current intensities. These data and those showing a shift in spike threshold with -DTX application suggest that Kv1.2 channels are key regulators of spike generation in naturally occurring situations.

After spike generation, Kv1.2 channels are strongly activated and then deactivate slowly, allowing them to participate in the afterhyperpolarization, slowing the trajectory to the next spike during sustained depolarization. Kv1.2 blockade accelerated discharge frequency over the entire frequency-intensity plot of medium spiny neurons. The slow deactivation kinetics of Kv1.2 channels are similar to those of Kv2 family channels that are also expressed by medium spiny neurons (Baranauskas et al. 1999), suggesting that both contribute to the voltage trajectory after spikes and the regular spiking pattern. The maintained activation of the Kv1.2 channels during repetitive firing (e.g., Fig. 7) is also dependent on their relatively hyperpolarized activation voltage dependence and slow inactivation kinetics as shown for D-type currents in hippocampal neurons (Storm 1988). The sensitivity of the afterhyperpolarization in medium spiny neurons to holding potential between -90 and -60 mV shown here and previously (Pineda et al. 1992) is consistent with the proposition that Kv1.2 channels contribute to this process and regular firing.

Kv1.2 channels also are likely to be important regulators of up-states in medium spiny neurons (Wilson and Groves 1981; Wilson and Kawaguchi 1996). Previous studies by Wilson and Kawaguchi (1996) have shown that K⁺ currents are key determinants of the membrane potential achieved during synaptically driven state transitions. Kv1.2 channels undoubtedly figure prominently in this regulation as they constitute nearly half of the total K⁺ current at up-state potentials (ca. -55 mV). The inactivation kinetics of Kv1.2 channels at this potential ensures that their contribution will wane only slowly during sustained up-states, largely unaffected by spiking. Nevertheless, sustained residence in the up-state, as may occur in the awake, behaving state, undoubtedly increases excitability by inactivating Kv1.2 channels. Over long periods (tens of seconds), this process endows medium spiny neurons with a form of activity-dependent, short-term memory (Turrigiano et al. 1996). The available evidence (Dodson et al. 2002; Golding et al. 1999) suggests that this function of Kv1.2 channels is most important in the soma and initial segment where spikes are generated and is not likely to be important in the dendrites receiving synaptic input. This conjecture must be tested, however. Low micromolar concentrations of 4-AP have prominent excitatory effects on glutamate-mediated EPSPs in medium spiny neurons (Bargas et al. 1998). It is possible that in addition to promoting presynaptic depolarization, blockade of Kv1.2 channels in postsynaptic dendritic regions provide a brake on synaptically mediated depolarization—much as Kv4 channels do in hippocampal pyramidal neurons (Yuan et al. 2002).

	ACKNOWLEDGMENTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
We thank S. Ulrich and Y. Chen for technical support.

GRANTS

This work was supported by National Institutes of Health Grants NS-34696 and DA-12958.

	FOOTNOTES

 
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Address for reprint requests and other correspondence: D. J. Surmeier, Dept. of Physiology, Northwestern University, Feinberg School of Medicine, 303 E. Chicago Ave., Chicago, IL 60611 (E-mail: j-surmeier{at}northwestern.edu).

	REFERENCES

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Akins PT, Surmeier DJ, and Kitai ST. Muscarinic modulation of a transient K⁺ conductance in rat neostriatal neurons. Nature 344: 240-242, 1990.[CrossRef][Medline]

Albert JL and Nerbonne JM. Calcium-independent depolarization-activated potassium currents in superior co