|Home | About | Journals | Submit | Contact Us | Français|
The rapidly activating and inactivating voltage-gated K+ (Kv) current, IA, is broadly expressed in neurons and is a key regulator of action potential repolarization, repetitive firing, back propagation (into dendrites) of action potentials, and responses to synaptic inputs. Interestingly, results from previous studies on a number of neuronal cell types, including hippocampal, cortical and spinal neurons, suggest that macroscopic IA is composed of multiple components and that each component is likely encoded by distinct Kv channel α-subunits. The goals of the experiments presented here were to test this hypothesis and to determine the molecular identities of the Kv channel α-subunits that generate IA in cortical pyramidal neurons. Combining genetic disruption of individual Kv α-subunit genes with pharmacological approaches to block Kv currents selectively, the experiments here revealed that Kv1.4, Kv4.2 and Kv4.3 α-subunits encode distinct components of IA that together underlie the macroscopic IA in mouse (male and female) cortical pyramidal neurons. Recordings from neurons lacking both Kv4.2 and Kv4.3 (Kv4.2-/-/Kv4.3-/-) revealed that, although Kv1.4 encodes a minor component of IA, the Kv1.4-encoded current was found in all the Kv4.2-/-/Kv4.3-/- cortical pyramidal neurons examined. Of the cortical pyramidal neurons lacking both Kv4.2 and Kv1.4, 90% expressed a Kv4.3-encoded IA larger in amplitude than the Kv1.4-encoded component. The experimental findings also demonstrate that the targeted deletion of the individual Kv α-subunits encoding components of IA results in electrical remodeling that is Kv α-subunit specific.
Voltage-gated K+ (Kv) currents play distinct roles in controlling neuronal action potential waveforms, repetitive firing patterns (Crill and Schwindt, 1983), responses to synaptic inputs (Kole et al., 2007), neurotransmitter release (Ishikawa et al., 2003) and synaptic plasticity (Schrader et al., 2002). Consistent with these diverse roles, multiple types of Kv currents with distinct time- and voltage-dependent properties are co-expressed in most neurons. The functional diversity of neuronal Kv currents is generated, in part, through the expression of multiple Kv channel pore forming (α) subunits. In cortical pyramidal neurons, for example, multiple Kv channel α-subunits from different subfamilies are co-expressed (Guan et al., 2006). The macroscopic Kv currents in these cells can be separated into four components based on differing time constants (τ) of inactivation: IA, which inactivates rapidly (τ≈25 ms); ID, characterized by an intermediate rate of inactivation (τ≈250 ms); IK, which inactivates slowly (τ≈2 seconds); and ISS, that is non-inactivating (Locke and Nerbonne, 1997b). Previous studies suggest that these kinetically distinct current components are encoded by molecularly distinct populations of channels (Yuan et al., 2005).
The rapidly activating and rapidly inactivating Kv current, IA, which is widely expressed in central and peripheral neurons (Rogawski, 1985), regulates multiple neuronal processes, including action potential repolarization, repetitive firing (Yuan et al., 2005), synaptic integration and the back propagation (into dendrites) of action potentials (Cai et al., 2004). Considerable, evidence also suggests that alterations in IA expression and/or function are associated with neuropathology. For example, IA availability is decreased in a mouse model of temporal lobe epilepsy (Bernard et al., 2004). Additionally, a mutation in Kv4.2 has been identified in a patient with temporal lobe epilepsy (Singh et al., 2006). Other studies utilizing experimental models of epilepsy have described alterations in the expression and the subcellular localization of Kv4.2 and Kv4.3 in the hippocampus (Lugo et al., 2008; Monaghan et al., 2008).
To understand the molecular mechanisms that regulate the expression, properties and functioning of IA in normal and pathological states, the pore forming and accessory subunits underlying the generation of IA channels have to be identified. Previous studies on neurons from mice (Kv4.2-/-) in which the Kcnd2 (Kv4.2) locus was disrupted revealed that Kv4.2 contributes importantly to the generation of IA in cortical pyramidal neurons (Nerbonne et al., 2008), hippocampal pyramidal neurons (Chen et al., 2006) and in neurons from the dorsal horn of the spinal cord (Hu et al., 2006). In each of these studies however, rapidly activating and inactivating currents similar to IA in wild type cells were observed suggesting that additional components of IA (i.e. not encoded by Kv4.2) are expressed in these cells. The studies presented here exploit pharmacology, in combination with genetic tools to disrupt the expression of individual Kv channel α-subunits, to identify the Kv channel α-subunits responsible for the generation of IA in cortical pyramidal neurons.
Wild type (WT) C57BL/6 mice and mice harboring targeted genetic disruption of the Kcna4 (Kv1.4-/-) (London et al., 1998), Kcnd2 (Kv4.2-/-) (Nerbonne et al., 2008), or Kcnd3 (Kv4.3-/-) (Niwa et al., 2008) locus in the C57BL/6 background were used in the experiments presented here. Neurons were isolated from the primary visual cortices of postnatal day 6-8 animals of both genders using previously described methods (Huettner and Baughman, 1986; Locke and Nerbonne, 1997a, b; Nerbonne et al., 2008). Briefly, mice were first anesthetized with halothane and then decapitated. After dissection of the visual cortex, the tissue, which contained all cortical layers, was chopped into small pieces and incubated in Neurobasal medium containing papain (Worthington, Lakewood, NJ) under 95% oxygen: 5% CO2 at 37°C for 30 minutes. Following the incubation, the tissue pieces were triturated using fire-polished pasteur pipettes. Isolated neurons were recovered by centrifugation through a bovine serum albumin gradient. Cells were resuspended in Neurobasal media (Invitrogen, Carlsbad, CA) and plated on previously prepared monolayers of cortical astrocytes (Huettner and Baughman, 1986). One hour after plating, the media was replaced with Minimum Essential Medium (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum and 0.14mM L-glutamine. Cultures were maintained in a 5% CO2 37°C incubator.
Whole-cell recordings were obtained from isolated cortical pyramidal neurons at room temperature (22-23°C). Recordings were obtained from pyramidal shaped neurons on the first and second day in culture prior to the elaboration of extensive processes to ensure, adequate voltage clamp control (see analysis below). Data were collected using an Axon 1D amplifier (Axon Instruments) controlled through a Digidata 1322 analog/digital interface (Molecular Devices, Sunnyvale, CA). Pipettes were fabricated from borosilicate glass (WPI, Sarasota, FL) with a Sutter model P-87 horizontal puller (Sutter Instruments, Novarto, CA). Using a standard pipette solution (see below), pipette resistances were between 2 and 4 MΩ. For recordings, bath solution contained (in mM): 140 NaCl, 4 KCl, 2 CaCl2, 2 MgCl2, 10 HEPES, 5 glucose, 0.001 TTX and 0.1 CdCl2, and (pH7.4 and 300mOsM). The recording pipette solution contained (in mM): 135 KCl, 10 HEPES, 5 glucose, 1.1 CaCl2, 2.5 BAPTA, and 3 MgATP and 0.5 NaGTP were added the day of recording (pH7.4; 300mOsM). The calculated free Ca2+ in this BAPTA buffered pipette solution was 100 nM (MAXCHELATOR)(Patton et al., 2004). The K+ channel blockers used, tetraethylammonium (TEA), 4-aminopyridine (4-AP), Heteropodatoxin-2, α-Dendrotoxin (Alomone Labs, Jerusalem, Israel) and Ba2+ were added to the bath solution immediately prior to recordings. All reagents were from Sigma (St. Louis, MO) unless otherwise noted.
In all experiments, junction potentials were zeroed prior to forming pipette-membrane seals. Signals were low pass filtered at 10 kHz, and sampled at 100 kHz. Whole-cell Kv currents were routinely evoked in response to 4 second depolarizing voltage steps to potentials between -40 mV and +40 mV (in 10 mV increments) from a holding potential of -70 mV. In parallel experiments, a prepulse paradigm was used to facilitate the isolation of the rapidly inactivating currents in each cell. In this case, currents evoked at test potentials from -40 to +40 mV (in 10 mV increments) following a brief (60 ms) step to -10 mV were recorded. Offline subtraction of the currents evoked following the prepulse from the currents evoked without the prepulse allowed the isolation of the rapidly inactivating outward K+ currents (see Figure 1).
Data were compiled and analyzed using ClampFit (Axon), Microsoft Excel, and Prism (Graphpad). Only data from cells with input resistances greater that 300MΩ and access resistances less than 15 MΩ were included in the analyses. Membrane capacitances were determined by analyzing the decays of capacitive currents elicited by short (25ms) voltage steps (± 10 mV) from the holding potential (-70 mV). Whole-cell membrane capacitances (Cm) were calculated for each cell by dividing the integrated capacitive transients by the voltage. Consistent with the short time in culture and lack of extensive processes, the capacitive transients of recorded cells had single exponential decay phases. Input resistances were calculated from the steady-state currents elicited by the same ± 10 mV steps (from the holding potential). For each cell, the series resistance was calculated by dividing the time constant of the decay of the capacitive transient (fit by a single exponential) by the Cm; the mean (± SEM) series resistance was 5.4 MΩ (± 0.1) (n=222). Series resistances were compensated electronically by greater than 80% in all cells. Voltage errors resulting from uncompensated series resistances, therefore, were small (<2 mV) and were not corrected. The inactivation phases of the Kv currents were analyzed using the equation y=A1e-t/τ1 +A2e-t/τ2 +A3e-t/τ3 +C where A1 (IA), A2 (ID), A3(IK) are the amplitudes of individual current components (see text), each with a characteristic time constant of decay (τ1, τ2, and τ3), and C is the non-inactivating component (ISS) of the total Kv current. For statistical analyses, current-voltage plots (IV plots) were compared using repeated measurement ANOVA. The statistical significance of the differences between individual IV plots were, subsequently, calculated using Tukey's multiple comparisons post hoc test and p values are reported in the text on in the figure legends.
For biochemical experiments mice were deeply anaesthetized with halothane, decapitated and the brains were rapidly removed. Posterior (~1 mm) cortices from 4 animals of each genotype (WT, Kv1.4-/-, Kv4.2-/- and Kv4.3-/-) were dissected and flash frozen in liquid nitrogen. Protein lysates were prepared from the posterior cortices using previously described methods (Brunet et al., 2004). The protein concentration in each sample was determined using a Bio-Rad protein assay kit (Bio-Rad, Hercules, CA), following the manufacturer's directions. Equal amounts of proteins were fractionated on 7.5% SDS-PAGE gels, and transferred to PVDF membranes, incubated in blocking buffer at room temperature for one hour, followed by incubation with primary antibodies against the individual Kv α-subunits overnight at 4°C. The monoclonal anti-Kv4.2 and anti-Kv4.3 antibodies were obtained from the UC Davis/NIH NeuroMab Facility, supported by NIH grant U24NS050606 and maintained by the Department of Neurobiology, Physiology and Behavior, College of Biological Sciences, University of California, Davis, CA. The polyclonal anti-Kv1.4 antibody used was obtained from Chemicon. Bound antibodies were detected using horseradish peroxidase conjugated rabbit anti-mouse IgG (Bethyl Labs, Montgomery, TX) or goat anti-rabbit IgG (GE Healthcare) and the Dura West chemiluminescence reagent (Pierce, Rockford, IL). Signals were detected and quantified using the Bio-Rad ChemiDoc system and Quantity One software (Bio-Rad). Blots were re-probed with primary antibodies against glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (Abcam, Cambridge, MA) to verify equal loading of lanes. For quantification the anti–GAPDH signals were used to normalize the Kv α-subunit signals measured from the same blot.
Previously we reported that the rapidly activating and rapidly inactivating Kv current, IA, that is prominent in wild type (WT) cortical pyramidal neurons (Figure 1A), was not evident in most (80%) cortical pyramidal neurons (Figure 1B) isolated from mice (Kv4.2-/-) harboring a targeted disruption of the Kcnd2(Kv4.2) locus (Nerbonne et al., 2008), findings consistent with previous suggestions that Kv4.2 is the critical Kv α-subunit encoding IA channels in cortical (and hippocampal) pyramidal neurons (Korngreen and Sakmann, 2000; Kim et al., 2005; Chen et al., 2006). In a small subset (~20%) of Kv4.2-/- cells, however, a rapidly inactivating current was observed (Nerbonne et al., 2008), suggesting the expression of a molecularly distinct (non-Kv4.2 encoded) IA in this subset of cortical pyramidal neurons.
The marked differences in the waveforms of the Kv currents in the vast majority (80%) of Kv4.2-/- cortical pyramidal neurons (Figure 1B) and those recorded from WT cells (Figure 1A) reflects the upregulation of delayed rectifier currents (IK and ISS), as evidenced by the sensitivity of the upregulated currents to the K+ channel blocker tetraethylammonium (TEA). As illustrated in Figure 1C, addition of 3 mM TEA to the bath reduced IK and ISS revealing a rapidly inactivating current component similar to IA in WT neurons (Figure 1C). Similar results were obtained in all (n=23) Kv4.2-/- neurons examined in the presence of TEA. To quantify the amplitude of IA in these (Kv4.2-/-) cells, a 60ms prepulse to -10mV was used to inactivate IA (Figure 1Db). Offline subtraction of the current records obtained with the prepulse from the control records in the same cell (Figure 1Da) allowed the isolation IA (Figure 1Da-b). The peak amplitudes of the subtracted current records were measured and normalized to cell capacitance to yield IA densities. The mean ± SEM density of IA in Kv4.2-/- neurons was significantly (p<0.01) lower than in WT cells (Figure 1F), an observation that might be interpreted as suggesting the presence of residual Kv4.2 encoded channels in Kv4.2-/- cells. Previous studies, however, have shown that no Kv4.2 protein or transcript is detectable in the cortices of Kv4.2-/- mice (Burkhalter et al., 2006; Nerbonne et al., 2008). Thus, the observation of a reduction in, but not the elimination of, IA in Kv4.2-/- neurons indicates that other Kv α-subunits (in addition to Kv4.2) contribute to the generation of functional IA channels in cortical pyramidal neurons
Subsequent experiments were focused on determining the molecular identity of the Kv α-subunit(s) underlying the non-Kv4.2-encoded component(s) of IA in cortical pyramidal neurons. In heterologous expression systems, subunits of the Kv1, Kv3, Kv4 and Kv12 subfamilies can generate rapidly activating and inactivating (A-type) Kv currents (Blair et al., 1991; Ruppersberg et al., 1991; Rettig et al., 1994; Trudeau et al., 1999). Previous studies (Camerino et al., 2007) also suggest that the channels encoded by the various Kv α-subunits can be distinguished by using selective pharmacologic blockers. In control experiments, 3 mM TEA, which effectively blocks Kv3 channels (Rudy and McBain, 2001; Gutman et al., 2003), had no measurable effects on IA in WT neurons, indicating that Kv3 channels do not contribute to the generation of IA in mouse visual cortical pyramidal neurons (data not shown). The remaining Kv subfamilies, Kv1, Kv4 and Kv12, can be, at least partially, separated based on differential sensitivities to 4-aminopyridine (4-AP). Previous studies have shown, for example, that Kv1 channels are blocked effectively by submillimolar concentrations of 4-AP (Grissmer et al., 1994; Rasmusson et al., 1995; Gutman et al., 2003), whereas Kv4 channels are blocked by 4-AP in the several millimolar range (Serodio et al., 1996) and Kv12 channels are insensitive to high millimolar concentrations of 4-AP (Trudeau et al., 1999; Gutman et al., 2003).
To explore the possible role of Kv1 channels, recordings were obtained from Kv4.2-/- and WT neurons exposed to 1mM 4-AP (in the presence of 3 mM TEA) (Figure 1E). Addition of 1 mM 4-AP to the bath solution markedly reduced IA in both Kv4.2-/- and WT neurons (Figure 1F), demonstrating the presence of a 1mM 4-AP-sensitive (Kv1 encoded?) component of IA in neurons of both (WT and Kv4.2-/-) genotypes. Although this concentration of 4-AP was selected to facilitate selective block of Kv1-encoded channels, Kv4-encoded channels might also be affected, albeit to a lesser extent (Serodio et al., 1996; Gutman et al., 2003). Consistent with the suggestion that Kv4.2 channels are not effectively blocked by 1 mM 4-AP, the magnitude of the reduction in IA density caused by 1 mM 4-AP was similar in Kv4.2-/- and WT neurons (Figure 1F). The mean ± SEM density of the 4-AP-resistant IA, however, was significantly (p<0.01) lower in Kv4.2-/-, than in WT, cells (Figure 1C-D), consistent with the elimination of the Kv4.2-encoded component of IA in Kv4.2-/- neurons. In recordings obtained from 10 (of 12) Kv4.2-/- neurons in 1 mM 4-AP a rapidly inactivating current remained, suggesting the presence of an additional, non-Kv1-encoded component of IA. There was no detectable 1 mM 4-AP-insensitive component of IA in the other 2 (of 12) Kv4.2-/- cells examined (see Discussion).
The 4-AP sensitivity of IA (Figure 1) suggests that one component of IA, in WT and Kv4.2-/- neurons is likely encoded by Kv1 channels. Previous studies in heterologous cells have shown that Kv1.4 generates rapidly activating and inactivating currents when expressed alone (Po et al., 1993) and that other Kv1 subfamily members can generate rapidly inactivating currents through heteromeric assembly with Kv1.4 or by combining with accessory Kvβ subunits, specifically Kvβ1 or Kvβ3 (Po et al., 1993; Heinemann et al., 1995; Leicher et al., 1998). To explore directly the hypothesis that Kv1.4 encodes a component of IA, whole-cell Kv currents were examined in cortical neurons isolated from mice (Kv1.4-/-) harboring a targeted disruption of the Kcna4 (Kv1.4) locus (London et al., 1998). The waveforms of the Kv currents in Kv1.4-/- neurons (Figure 2A) were similar to those in WT neurons (Figure 1A) and the inactivation phases of the currents were also well described by the sum of three exponentials. Analysis of the Kv current waveforms revealed that the mean ± SEM density of the peak Kv current as well as IA, ID, and IK densities were similar in Kv1.4-/- neurons (n=15) to those determined in WT cells (n=28). There was, however, a small, but statistically significant (P<0.01) difference, in ISS densities measured in Kv1.4-/- and WT neurons (Table 1). The observation that mean ± SEM IA densities are similar in Kv1.4-/- and WT neurons may indicate that Kv1.4 does not contribute to IA in WT cells or, alternatively, that other IA components are upregulated in Kv1.4-/- neurons masking the loss of the Kv1.4-encoded current. To explore these possibilities, pharmacological experiments were conducted.
In hippocampal pyramidal neurons barium (Ba2+) in the hundred micromolar range has been shown to reduce IA selectively (Gasparini et al., 2007; Andrasfalvy et al., 2008). In addition, the effect of Ba2+ on IA was reportedly reduced in Kv4.2-/- hippocampal neurons (Andrasfalvy et al., 2008) suggesting that, of the channels that may encode IA, Ba2+ is selective for Kv4-encoded channels, although block of inward rectifier channels has also been reported (Schoots et al., 1996; Hughes et al., 2000). Control experiments revealed that the effect of 400 μM Ba2+ on IA amplitudes in Kv4.2-/- neurons (n=18) was significantly (p<0.01) smaller than in WT cells (n=14). At +40 mV, for example, the calculated mean ± SEM density of the Ba2+ sensitive component of IA was 75 ± 4 pA/pF in WT neurons and 42 ± 3 pA/pF in Kv4.2-/- neurons. In subsequent experiments, therefore, Kv currents in Kv1.4-/- and WT neurons were examined in control bath solution and in the presence of 400 μM Ba2+ (Figure 2). As illustrated in Figure 2, addition of 400 μM Ba2+ to the bath significantly (p<0.01) reduced the density of IA in Kv1.4-/-neurons (Figure 2C). In addition, the mean ± SEM density of the Ba2+-resistant component of IA was significantly (p<0.05) larger in WT than in Kv1.4-/- neurons (Figure 2C), indicating a role for Kv1.4 in the generation of IA in WT neurons. This observation, together with the finding that mean ± SEM peak IA densities are similar in Kv1.4-/- and WT neurons, suggests the Kv4-encoded component of IA is upregulated in Kv1.4-/- neurons.
The results of the experiments described above utilizing pharmacology in combination with the targeted disruption of Kv4.2 or Kv1.4 revealed roles for both of these subunits (Kv4.2 and Kv1.4) in the generation IA current in cortical pyramidal neurons. To determine if there are additional Kv channels that contribute to the generation of macroscopic IA in cortical pyramidal neurons, we generated mice lacking both Kv1.4 and Kv4.2 (Kv1.4-/-/Kv4.2 -/-) and examined Kv currents in cells isolated from these animals. Similar to the Kv currents in Kv4.2-/- neurons (Figure 1B), marked increases in the delayed rectifier currents (IK and ISS) were evident in records obtained from Kv1.4-/-/Kv4.2 -/- neurons (Figure 3A). Also similar to the Kv4.2-/- neurons, addition of TEA to the bath blocked the delayed rectifier currents and unmasked a residual component of IA in 18 of 20 Kv1.4-/-/Kv4.2 -/- cells examined (Figure 3B-C). In 2 (of the 20) Kv1.4-/-/Kv4.2 -/- neurons studied, no rapidly inactivating currents were detected in the presence of TEA.
Subsequent experiments were focused on exploring the possible roles of Kv1 and Kv4 α-subunits in the generation of the component of IA remaining in Kv1.4-/-/Kv4.2-/- neurons (Figure 3). In addition to Kv1.4, several Kv1 α-subunits including Kv1.1, Kv1.2 and Kv1.6 are reportedly expressed in cortical pyramidal neurons (Guan et al., 2006). Initial experiments were performed using the peptide toxin α-Dendrotoxin (α-Dtx), which has been reported to block heterologously expressed Kv1.1-, 1.2-, and 1.6-encoded currents and has previously been used to examine the roles of Kv1 channels in cortical neurons (Harvey and Robertson, 2004; Guan et al., 2006; Kole et al., 2007). As was done in the experiments on Kv4.2-/- neurons (Figure 1), TEA was added to the bath to block the large delayed rectifier currents (IK and ISS) in Kv1.4-/-/Kv4.2-/- neurons (Figure 3A). The further addition of α-Dtx to the bath solution at a concentration of 100 nM had no significant effect on IA currents in Kv1.4-/-/Kv4.2-/- neurons (not illustrated). Similar experiments were carried out using Heteropdatoxin-2 (Hptx-2), which is specific for Kv4 channels (Zarayskiy et al., 2005). In contrast to α-Dtx, addition of Hptx-2 (at 1 μM) significantly (p<0.01) reduced IA amplitudes of in Kv1.4-/-/Kv4.2-/- neurons (Figure 3D), revealing a further role for Kv4 α-subunits (i.e. in addition to Kv4.2) in the generation of IA in cortical pyramidal neurons. Interestingly, previous reports demonstrated that Kv4.3 is expressed in cortical pyramidal neurons (Serodio and Rudy, 1998; Burkhalter et al., 2006), suggesting a role for Kv4.3.
To explore directly the role of Kv4.3 in the generation of IA in cortical pyramidal neurons, whole-cell Kv current recordings were obtained from neurons isolated from animals harboring a targeted disruption of the Kcnd3 (Kv4.3-/-) locus (Niwa et al., 2008). In most (20 of 24) of the Kv4.3-/- neurons studied, the waveforms of the Kv currents (Figure 4B) were similar to those recorded from WT neurons (Figure 4A) with a prominent rapidly inactivating IA component. Similar to the Kv currents in WT neurons, the inactivation phases of the Kv currents in these (20 of 24) Kv4.3-/- neurons were well described by the sum of three exponentials, consistent with the expression of IA, ID, IK and ISS (Table 1). The waveforms of the Kv currents in the remaining (4 of 24) Kv4.3-/- neurons (Figure 4C), however were distinct, and resembled the current waveforms seen in most Kv4.2-/- neurons (Figure 1B) with large delayed rectifier currents and without a prominent rapidly inactivating current component. The prepulse paradigm described previously was used to isolate (and allow the quantification of) IA in Kv4.3-/- neurons. These experiments revealed that the mean ± SEM density of IA was significantly (p<0.01) lower in Kv4.3-/- neurons (n=24) than in WT cells (n=22) (Figure 4D).
The results of the experiments presented above indicate that in addition to Kv4.2, Kv1.4 and Kv4.3 also encode IA channels and contribute to the generation of the macroscopic IA recorded in cortical pyramidal neurons. Accordingly, disruption of both Kv4.2 and Kv4.3 should leave only the Kv1.4-encoded component of IA, a current that would be expected to be blocked completely by 1mM 4-AP. This hypothesis was tested directly by obtaining recordings from neurons isolated from mice lacking both Kv4.2 and Kv4.3 (Kv4.2-/-/Kv4.3-/-), generated by crossing the Kv4.2-/- and Kv4.3-/- animals. The waveforms of the Kv currents in Kv4.2-/-/Kv4.3-/-cortical pyramidal neurons (Figure 5A) resembled those in Kv4.2-/- neurons (Figure1B) with large delayed rectifier currents and without a prominent rapidly inactivating current component. Inclusion of 3mM TEA in the bath unmasked a rapidly inactivating current in all (n=18) Kv4.2-/-/Kv4.3-/- neurons (Figure 5B). Interestingly, the mean ± SEM peak density of the IA remaining in Kv4.2-/-/Kv4.3-/- neurons is similar in magnitude to the IA component eliminated in Kv1.4-/-neurons (Figure 2B). In addition, as illustrated in Figure 5C, no rapidly inactivating currents remained in Kv4.2-/-/Kv4.3-/- neurons (n=13) in the presence of 1 mM 4-AP (Figure 5C).
To examine the relative expression levels of the Kv1.4, Kv4.2, and Kv4.3 proteins in cortex and the impact of the targeted deletion of individual α-subunits, Western blots were conducted on lysates prepared from posterior cortices (containing visual cortex) dissected from WT, Kv1.4-/-, Kv4.2-/- and Kv4.3-/- animals (n=4 for each genotype). As illustrated in Figure 6A-C, the Kv1.4, Kv4.2 and Kv4.3 proteins were readily detected in the cortical lysates from WT mice, demonstrating that all three of these Kv α-subunits are expressed. In addition, in the samples from Kv1.4-/-, Kv4.2-/- or Kv4.3-/- animals, no Kv1.4, Kv4.2 or Kv4.3 protein, respectively, was detected, confirming the specificity of each of the antibodies used. Each blot was also probed with an anti-GAPDH antibody to confirm equal protein loading in each lane. The anti-GAPDH signals were used to normalize the Kv α-subunit-specific antibody signals in each lane. Quantification of the normalized Kv α-subunit-specific antibody signals revealed that the expression levels of Kv1.4 and Kv4.3 in samples from mice with targeted disruption of the other Kv α-subunits were similar to those in WT cortices (Figure 6D). The expression of Kv4.2 appeared to be slightly increased in samples from Kv1.4-/- animals relative to those from WT animals, but the difference did not reach statistical significance (p=0.15 by t-test)(Figure 6D). These results indicate the total protein expression levels of the Kv α-subunits encoding individual components of IA do not undergo appreciable remodeling in response to the genetic disruption of the other IA encoding Kv α-subunits. Nevertheless, it is certainly possible that remodeling of the subcellular distribution of Kv subunits in different neuronal compartments occurs in one or more of the targeted deletion animals. Alternative experimental approaches need to be employed to explore this possibility.
The contributions of Kv1.4, Kv4.2 and Kv4.3 to the generation of the total macroscopic IA in cortical pyramidal neurons is revealed in direct comparisons of IA densities in cortical pyramidal neurons of the various genotypes and examined under different pharmacologic conditions (Figure 7). As discussed previously, in all experiments conducted on Kv4.2-/- (as well as Kv1.4-/-/Kv4.2-/- and Kv4.2-/-/Kv4.3-/-) neurons 3 mM TEA was used to facilitate the isolation of IA by reducing the large delayed rectifier currents (IK) that are present in Kv4.2-/- neurons (Figure 1). In control experiments on WT cells 3 mM TEA had no effect on IA (data not shown). Examination of the distributions of IA densities determined in individual cells reveals that there is considerable heterogeneity in peak IA densities among WT cortical pyramidal neurons (Figure 7A). In both Kv4.2-/- and Kv4.3-/- neurons, the distributions of IA densities in individual cells are similar to WT neurons, although the mean IA densities are lower. In the Kv1.4-/-/Kv4.2-/- cortical pyramidal neurons the mean IA density was similar to that measured in Kv4.2-/- neurons, although the distribution of IA densities in individual neurons is shifted considerably, revealing that more Kv1.4-/-/Kv4.2-/- neurons displayed low IA densities. In addition, IA densities in Kv4.2-/-/Kv4.3-/- neurons were low and tightly clustered (Figure 7A).
Analyses of results of the experiments conducted using the various pharmacological manipulations yielded similar conclusions. Consistent with roles for Kv4 and Kv1 channels in the generation of IA, for example, both Ba2+ and 4-AP reduced the density of IA in WT cortical pyramidal neurons. In addition, the 1 mM 4-AP-resistant component of IA was reduced in Kv4.2-/-, relative to WT, neurons, consistent with the loss of the Kv4.2-encoded component of IA. Further, the component of IA remaining in Kv4.2-/-/Kv4.3-/- neurons was eliminated completely by 1 mM 4-AP, indicating that the 4-AP resistant component of IA in Kv4.2-/- neurons is encoded by Kv4.3. Finally, complete block of the component of IA remaining in Kv4.2-/-/Kv4.3-/- cells by 1 mM 4-AP is consistent with Kv1.4 containing channels encoding this component of IA.
A systematic experimental approach, employing genetic disruption of Kv α-subunit expression paired with pharmacology, was utilized to identify the Kv channel α-subunits responsible for the generation of macroscopic IA in cortical pyramidal neurons. The results presented here demonstrate that Kv1.4, Kv4.2 and Kv4.3 all contribute to macroscopic IA in (mouse visual) cortical pyramidal neurons. In Kv1.4-/-, Kv4.2-/- and Kv4.3-/- neurons, a component of IA is lost. The components of IA encoded by Kv1.4 and Kv4.3 were individually isolated in neurons (Kv4.2-/-/Kv4.3-/- and Kv1.4-/-/Kv4.2-/- neurons, respectively) with combined genetic disruption of the other two Kv α-subunits. In each case, the remaining component of IA was blocked selectively by 4-AP or Hptx-2 (respectively). In addition, the component of IA remaining in Kv1.4-/-/Kv4.2-/- neurons was sensitive to Hptx-2 but not to α-Dtx, indicating that Kv1 α-subunits do not contribute to IA in cortical pyramidal neurons in the absence of Kv1.4. To the best of our knowledge, this study represents the first complete molecular dissection of IA in mammalian neurons and, in addition, provides the first direct demonstration of a native neuronal Kv1.4-encoded current.
Precise determination of the contributions of Kv1.4-, Kv4.2- and Kv4.3-encoded channels to the generation of the macroscopic IA in individual WT neurons has been limited by the lack of potent blockers specific for channels encoded by these α-subunits. The limitations of pharmacology can be seen in Figure 7 where the reduction of IA density in WT neurons due to 1mM 4-AP is larger than expected if 1 mM 4-AP specifically and selectively blocks only Kv1-encoded channels. The magnitude of the reduction in IA density by 1 mM 4-AP suggests, probably not surprisingly, some block of Kv4 channels at this (1 mM) concentration. The use of targeted gene disruption is specific, but is hindered by the now well documented electrical remodeling evident in neurons when the expression of the normal channel repertoire is altered (Marder and Goaillard, 2006; Van Wart and Matthews, 2006; Nerbonne et al., 2008). The summary data (Figure 7) indicate that Kv4.2 and Kv4.3 α-subunits are the major contributors to macroscopic IA in cortical pyramidal neurons. The cumulative results also indicate that Kv1.4-encoded channels contribute a minor component of IA, expressed at a lower density in cortical pyramidal neurons than the Kv4-encoded components (Figure 7A). The recordings here were obtained from young postnatal cortical pyramidal neurons and it is certainly possible that the relative contributions of Kv1.4, Kv4.2 and Kv4.3 to the total IA changes during development. Interestingly, the biochemical data revealed the robust expression of Kv1.4, Kv4.2 and Kv4.3 in adult cortex (Figure 6). Previous studies suggest that the physiological properties of cortical neurons do change during postnatal development although the major effects appear to be quantitative changes in current densities, rather than qualitative changes in current properties/types (McCormick and Prince, 1987; Kasper et al., 1994). It seems reasonable to suggest, therefore, that Kv1.4, Kv4.2 and Kv4.3 all contribute to IA is cortical pyramidal neurons throughout postnatal development. As neurons mature, however, the subcellular distribution patterns and/or functional roles of individual channel types may change. Further experiments will be necessary to explore these questions directly.
The experiments here also revealed that Kv current remodeling is evident in Kv1.4-/-, Kv4.2-/-, and Kv4.3-/- neurons. The characteristics of the Kv current remodeling, however, were different in each case. In Kv1.4-/- neurons, a small increase in ISS was seen in conjunction with an increase in a Ba2+ sensitive (Kv4-encoded) rapidly inactivating current. In the majority (~80%) of Kv4.2-/- neurons, marked differences in Kv current waveforms were evident, reflecting increased TEA-sensitive delayed rectifier currents (Figure 1) (Nerbonne et al., 2008). Surprisingly, only 20% of Kv4.3-/- neurons displayed a similar remodeling (Figure 4C) despite similar decreases in mean IA density in Kv4.2-/- and Kv4.3-/- neurons (Figure 7A). These observations illustrate the complexity of ascertaining the functional roles of individual channel α-subunits by using genetic disruption alone. Interestingly, other studies have reported changes occurring at the circuit level in Kv4.2-/- mice. Specifically, experiments on acute brain slices from Kv4.2-/- mice revealed increased inhibition of hippocampal pyramidal neurons mediated by an increase in tonic GABA currents (Andrasfalvy et al., 2008), suggesting that widespread, compensatory changes in neuronal properties and excitability can occur in response to alterations in Kv α-subunit expression.
The results presented here demonstrate that both Kv4.2 and Kv4.3 can form functional channels in cortical pyramidal neurons independent of each other. Finding functional Kv4.3 channels independent of Kv4.2 expression in cells that normally express both is somewhat surprising given that cardiac myocytes from Kv4.2-/- mice have no remaining Kv4-encoded current despite the expression of Kv4.3 (Guo et al., 2005). Although the experiments here demonstrate that Kv4.2 and Kv4.3 can function independently, previous studies have shown that Kv4.2 and Kv4.3 can be co-immunoprecipitated from brain, consistent with the presence of heteromultimeric Kv4.2/Kv4.3 channels (Marionneau et al., 2009). The formation of functional homomultimeric or heteromultimeric Kv4 channels is also consistent with the partially overlapping subcellular localization of Kv4.2 and Kv4.3 in cortical neurons (Burkhalter et al., 2006). The molecular diversity of neuronal Kv4 channels could enable precise and independent modulation of multiple neuronal processes, as well as differential sensitivities to multiple regulatory pathways. The functional diversity of neuronal Kv4 channels is likely further expanded by the co-expression of numerous accessory subunits, such as the K+ Channel Interacting Proteins, KChIPs, and Dipeptidyl Peptidases, DPP 6 and 10 (Schrader et al., 2002; Jerng et al., 2005; Maffie and Rudy, 2008).
Numerous recent studies suggest that neuronal Kv channels, like other types of ion channels, function as components of macromolecular protein complexes (Lai and Jan, 2006; Dai et al., 2009). Identification of the Kv α-subunits responsible for the generation of specific Kv currents is a critical first step in determining the composition of functional Kv channel complexes and the roles individual Kv α- and accessory subunits play in controlling channel properties and in regulating neuronal excitability. Knowing that Kv1.4, Kv4.2 and Kv4.3 encode distinct components of IA in cortical pyramidal neurons, therefore, provides a foundation for studies aimed at defining the physiological roles of accessory subunits and other regulatory proteins in the generation of functional IA channel complexes.
In spite of the many studies in heterologous expression systems, very little is known about the in situ functioning of Kv channel accessory subunits and translating findings from heterologous systems to native cells has proven difficult. For example, studies in heterologous cells suggest that DPP6 plays a dominant role in determining the kinetic properties of Kv4-encoded currents (Jerng et al., 2005). A recent study examining the effects of disrupting DPP6 expression in hippocampal neurons, however, described very small changes in the properties of IA, although there were marked and unexpected alterations in neuronal excitability (Kim et al., 2008). Although the authors interpreted the functional effects in terms of changes in IA, the experiential observations may, in part, reflect electrical remodeling with knockdown of DPP6, as is evident in response to disruption of Kv channel α-subunit expression (Guo et al., 2005; Nerbonne et al., 2008). Interestingly, several channel accessory subunits have been suggested to interact with and differentially regulate multiple types of ion channels (Li et al., 2005; Nerbonne and Kass, 2005; Sole et al., 2009; Thomsen et al., 2009), highlighting the possibility that accessory subunits may play complex roles in regulating different types of channels, as well as in orchestrating electrical remodeling. The demonstration here that Kv1.4, Kv4.2 and Kv4.3 each encode a component of IA in cortical pyramidal neurons and that varied electrical remodeling occurs in response to the disruption of Kv α-subunit expression will facilitate the design and, perhaps most importantly, the interpretation of experiments focused on defining the roles of IA channel accessory and regulatory proteins.
The authors thank Ms. Amy Huntley for assistance with tissue culture and Mr. Rick Wilson for the maintenance and screening of the mice used in this study. We would also like to thank Nick Foeger, Dr. Scott Marrus and Dr. Yarimar Carrasquillo for many valuable discussions and for comments on the manuscript. In addition, the authors acknowledge the financial support provided by National Institutes of Health (NS030676 to JMN); AJN was supported by an institutional training grant (T32-EY13360) from the National Eye Institute.