• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Neuron. Author manuscript; available in PMC Aug 22, 2009.
Published in final edited form as:
PMCID: PMC2730466

K+ Channels at the Axon Initial Segment Dampen Near-Threshold Excitability of Neocortical Fast-Spiking GABAergic Interneurons


Fast-spiking cells (FS cells) are a prominent subtype of neocortical GABAergic interneurons with important functional roles. Multiple FS cell properties are coordinated for rapid response. Here, we describe an FS cell feature that serves to gate the powerful inhibition produced by FS cell activity. We show that FS cells in layer 2/3 barrel cortex possess a dampening mechanism mediated by Kv1.1-containing potassium channels localized to the axon initial segment. These channels powerfully regulate action potential threshold and allow FS cells to respond preferentially to large inputs that are fast enough to “outrun” Kv1 activation. In addition, Kv1.1 channel blockade converts the delay-type discharge pattern of FS cells to one of continuous fast spiking without influencing the high-frequency firing that defines FS cells. Thus, Kv1 channels provide a key counterbalance to the established rapid-response characteristics of FS cells, regulating excitability through a unique combination of electrophysiological properties and discrete subcellular localization.


Classically, neocortical neurons have been divided into two broad categories: spiny cells that release the excitatory neurotransmitter glutamate and nonspiny (or sparsely spiny) local-circuit interneurons that release the inhibitory neurotransmitter γ-aminobutyric acid (GABA) (Peters and Jones, 1984). A prominent subset of GABAergic interneurons in mammalian neocortex exhibit action potentials (APs) of brief duration and discharge sustained trains of APs at high frequency in response to prolonged stimulation: such cells have been alternately referred to as fast-spike units (FSUs; Simons, 1978) or fast-spiking (FS) cells (McCormick et al., 1985; Connors and Gutnick, 1990).

FS cells have been suggested to contribute to a host of neocortical functions at multiple levels of analysis, from feed-forward inhibition within cortical microcircuits (Miller et al., 2001; Swadlow, 2003; Cruikshank et al., 2007) to the generation of synchronous and rhythmic network activity (Traub et al., 2004) and the regulation of critical periods for sensory map formation in the intact animal (Hensch, 2005). Dysfunction of neocortical FS cells has been shown to lead to epilepsy (Lau et al., 2000; Ogiwara et al., 2007) and is hypothesized to be involved in the pathogenesis of schizophrenia (Lewis et al., 2005).

Consistent with this functional significance, FS cells possess many unique anatomical, molecular, and physiological specializations. FS cells are the most prevalent type of neocortical interneuron (~50% of the total), possess powerful output synapses localized to the perisomatic region of target cells, and form dense gap-junctional connectivity with one another (Kawaguchi and Kubota, 1997; Gibson et al., 1999: Galarreta and Hestrin, 1999). It has been proposed that FS cells may constitute the dominant inhibitory subsystem in neocortex (Beierlein et al., 2003).

Multiple FS cell features are coordinated for rapid response, including high-frequency repetitive firing, brief single spikes, a fast membrane time constant, and tight coupling of GABA release to presynaptic calcium (Ca2+) influx. These features are produced by complex molecular machinery that includes the specific expression of voltage-gated potassium (K+) channels of the Kv3 subfamily, AMPA receptors with rapid gating kinetics, sodium (Na+) channels with fast activation kinetics, and the exclusive use of P/Q-type Ca2+ channels in neurotransmitter release (Rudy and McBain, 2001; Jonas et al., 2004; Hefft and Jonas, 2006).

Here, we describe a feature of FS cells that, in marked contrast to the multiple previously described rapid-response properties of these cells, serves to suppress excitability. The molecular basis of this dampening influence on FS cell excitability is the spatially delimited expression of specific K+ channels at the FS cell axon initial segment (AIS), such that these channels have a profound but restricted influence on near-threshold excitability and AP threshold. High-frequency firing by FS cells is widely observed under in vivo conditions (Swadlow, 2003; Contreras, 2004). Nevertheless, a counter-regulatory mechanism to balance the rapid-response properties of FS cells is likely important given the high divergence of FS cell synaptic connectivity and the powerful perisomatic inhibition provided by these cells.

AP threshold is a basic biophysical property of all spiking neurons. However, little is known concerning the intrinsic conductances that regulate this parameter. Classical work indicated a role for membrane potential (Vm) and the slope of changes in Vm in regulating the state of Na+ channel inactivation and hence spike threshold (Noble, 1966). The existence of within-cell variance in the voltage threshold for AP generation (rather than a fixed AP threshold) as well as a role for Na+ channel inactivation in this variability has been demonstrated in recordings of cortical neurons (likely pyramidal cells) in vivo (Azouz and Gray, 2000; Henze and Buzsaki, 2001). However, K+ channels in general are also known to regulate the subthreshold properties of neurons (Connor and Stevens, 1971). Here, we show that, in FS cells, AP threshold is additionally regulated by a specific K+ channel (a Kv1 channel) via a dynamic interaction with Na+ channels.

Numerous attempts have been made to classify neocortical interneurons (e.g., Gupta et al., 2000; Petilla Interneuron Convention, online at www.columbia.edu/cu/biology/faculty/yuste/petilla), although many issues remain to be resolved. We show additionally that the very same K+ channel that controls AP threshold in FS cells also regulates the discharge pattern of FS cells in response to near-threshold depolarizations and thereby accounts for at least some of the firing pattern diversity observed within the FS cell class.

Finally, the Kv1 K+ channel described in this study is shown to contain Kv1.1 subunits, which is of further physiological relevance given the known role of this K+ channel subunit in various forms of pathological hyperexcitability in rodents and humans.

These results have been presented previously in abstract form (E.M.G. et al., Soc. Neurosci., abstract 234.15, 2006).


Near-Threshold Properties of FS Cells

We performed whole-cell recordings from FS cells in the barrel subfield of layer 2/3 mouse primary somatosensory cortex (barrel cortex) using the acute slice preparation. In many cases, we used transgenic mice expressing enhanced green fluorescent protein (EGFP) driven by the promoter of the calcium-binding protein parvalbumin (PV; a known marker of FS cells; Kawaguchi and Kubota, 1997) for the efficient targeting of FS cells (see the Supplemental Experimental Procedures available online). Confirmation of FS cell identity was based upon known features of single FS cell action potentials (APs) and of suprathreshold FS cell discharge patterns (Rudy and McBain, 2001). For example, at 30°C–32°C, AP half-width was 0.34 ± 0.07 ms (mean ± SD; n = 41), time from AP voltage threshold to the trough of the after-hyperpolarization (AHP) was 2.6 ± 1.5 ms, and the amplitude of the fast AHP (relative to threshold) was 18.7 ± 8.7 mV. FS cells also lacked spike-frequency adaptation (see below). In response to near-threshold current pulses, all FS cells (76/76; 100%) exhibited an abrupt onset of repetitive discharge, in that the first suprathreshold current injection (the threshold current, ITH; see Supplemental Experimental Procedures) elicited multiple spikes (Figure 1A). This near-threshold behavior produced a discontinuous relationship between injected current and firing frequency (f–I curve) as compared with regular-spiking pyramidal cells (PCs) recorded in parallel (Figures 1B and 1C). PCs were found to discharge one spike (or occasionally two spikes) with ITH. An abrupt onset of firing was widespread among FS cells in various neocortical areas (Figure S1 and Figure S2) and across layers. Initial firing frequency (fi) was 16.0 ± 8.5 Hz for FS cells (n = 14) and 2.3 ± 0.9 Hz for PCs (calculated for a 600 ms sweep; n = 5; p < 0.01 versus FS cells via Student’s t test) (Figure 1D). In addition, the gain of the f–I relation was 593.4 ± 151.1 Hz/nA for FS cells (n = 12) and 54.3 ± 28.9 for PCs (n = 5; p < 0.01) (Figure 1D).

Figure 1
Salient Electrophysiological Properties of FS Cells in Barrel Cortex

These observations were consistent with previous studies indicating that neocortical FS cells exhibit an abrupt onset of firing (e.g., McCormick et al.,1985; Kawaguchi, 1995). In addition, such data support the recent suggestion that FS cells exhibit Hodgkin type 2 membrane excitability, while regular-spiking PCs are of type 1 (Tateno et al., 2004).

Previous studies have also noted that FS cells exhibit little or no spike-frequency adaptation (McCormick et al., 1985; Gibson et al., 1999). We found that, on average, FS cells in layer 2/3 barrel cortex actually show spike-frequency acceleration during sustained discharge. The ratio of the first interspike interval (ISI1) to the tenth (ISI10) and last (ISIn) interspike intervals (ISI1/ISI10 and ISI1/ISIn) were 1.17 ± 0.13 and 1.03 ± 0.20, respectively (n = 31), with 26 of 31 FS cells (84%) recorded having ISI1/ISI10 ratios >1 and 18 of 31 FS cells (58.1%) having ISI1/ISIn ratios >1. This is in contrast to PCs in layer 2/3 barrel cortex, which had ISI1/ISI10 values of 0.35 ± 0.08 (n = 4; p < 0.01 versus FS cells) and ISI1/ ISIn of 0.31 ± 0.08 (n = 4; p < 0.01 versus FS cells) (Figure 1F). Closer examination of the instantaneous firing frequency (IFF) of FS cells during sustained trains revealed two processes: firing frequency increased early in trains, followed by slight spike-frequency adaptation later in trains, with the ratio of the smallest ISI in the train (ISImin, corresponding to the fastest IFF) to ISIn (ISImin/ISIn) being 0.87 ± 0.05 (n = 27) (Figure 1E).

Delay-Type Firing by FS Cells

The most distinctive feature of neocortical FS cells is the discharge of sustained trains of brief APs at high frequency. However, we observed that FS cells in barrel cortex often exhibited a prominent delay to first spike near ITH (Figure 1A and Figure 2A and Figure S1), a feature noted to be particularly frequent among (but not unique to) FS cells in supragranular layers.

Figure 2
Delay-Type Firing Is Associated with a Slow Ramp Depolarization and a Positive Shift in AP Voltage Threshold

When firing behavior was probed with small (1–5 pA) current steps near ITH, 70 of 76 FS cells (92% of the total) in layer 2/3 mouse barrel cortex exhibited a prominent delay to first spike. Delayed firing by FS cells was not simply due to the passive charging of the FS cell membrane, as these cells were found to have a fast membrane time constant (τm) of 5.7 ± 1.1 ms (range, 3.9–9.1 ms; n = 23). With rectangular current pulses of 600 ms duration, the average delay to first spike (measured from current onset) was 243.3 ± 115.7 ms (range, 76.3–535.7 ms; n = 39). While delay-type FS cells were observed across all layers of barrel cortex as well as in other neocortical areas (Figure S1), we focused our subsequent investigations on layer 2/3 FS cells of barrel cortex, as almost all of these cells exhibited a robust delay to first spike near threshold.

FS cell firing patterns that include such a delay to first AP with threshold current injection have been recorded previously (e.g., Kawaguchi and Kubota, 1997; Gibson et al., 1999). This firing pattern likely corresponds to the delayed nonaccommodating (dNAC) interneuron subtype of Gupta et al. (2000) and the FS cells with a delayed-onset response (dFS cells) defined by Petilla interneuron nomenclature. However, the underlying basis and physiological relevance of this FS cell feature has never been directly investigated.

The prevalence of dFS cells in neocortex may have been underestimated previously due to the fact that delayed firing is present specifically near ITH. To quantify the range of current injections across which a delay to first spike was observed, we divided the values for current injection at which dFS cells shifted to a continuous fast-spiking cell (cFS) pattern by the value of ITH, which normalized for variation in ITH across cells. We found that, among dFS cells in layer 2/3 barrel cortex, delayed firing existed within a range of 1.00 to 1.25 ± 0.13 times ITH (n = 23; data not shown); that is, the FS cell firing pattern converted from that of a dFS cell to that of a cFS cell at, on average, 1.25 times ITH. This range reflects a large amount of absolute current (95 ± 29 pA), as FS cells have a low input resistance and a depolarized voltage threshold for AP generation and hence exhibit large values for ITH (381 ± 116 pA; n = 23) as compared to PCs (175 ± 51; n = 16; p < 0.01 versus FS cells).

The 6/76 FS cells in layer 2/3 without a delay exhibited an onset spike (or a few spikes) at ITH, followed by a long pause (Figure S2). Nevertheless, 76 of 76 FS cells recorded in layer 2/3 barrel cortex (100% of total), irrespective of firing pattern near threshold, converted to a cFS pattern with sufficient current injection (Figure S2).

Delayed Firing and AP Threshold in FS Cells

We also noted that the voltage threshold for the generation of APs differed between the two modes of FS cell discharge (dFS and cFS) (Figures 2B and 2D). The first AP elicited at ITH occurred (following a delay) at a voltage threshold of −39.4 ± 3.0 mV (n = 12). However, during sustained high-frequency discharge in response to suprathreshold current (cFS pattern), the voltage threshold for AP generation shifted to more hyperpolarized levels, to −43.9 ± 3.3 mV for the first spike in cFS trains (n = 12; p < 0.01 versus dFS firing mode). Further examination revealed that the FS cell delay could be attenuated and ultimately eliminated via the delivery of graded subthreshold conditioning prepulses of increasing amplitude (Figures S4A and S4B). A conditioning pulse of a magnitude sufficient to eliminate the delay also produced a negative shift in the voltage threshold of the first AP elicited (from −38.7 ± 3.4 to −40.5 ± 2.5 mV; n = 4).

In dFS cells, the duration of the delay decreased progressively with increasing current injection (Figure 2E). Correspondingly, the voltage threshold for AP generation slowly shifted to more hyperpolarized values in a graded fashion as the cell transitioned from dFS to cFS in response to increasing current injection (Figure 2F). At intermediate current injections, some dFS cells displayed an onset spike separated from high-frequency tonic firing by a pause (e.g., Figure 2A). In cells that displayed this pause, a graded regulation of AP threshold was still observed (Figure 2F).

Two conclusions can be drawn from these results. First, in the suprathreshold range, all FS cells discharge sustained trains of APs at high frequency (i.e., cFS pattern). Second, in the near-threshold voltage range, most FS cells in supragranular layers of barrel cortex exhibit a prominent delay to first spike. The transition to cFS firing was accompanied by a large negative shift in the voltage threshold for the generation of APs. cFS discharge included early spike-frequency acceleration followed by slight spike-frequency adaptation during sustained trains. Although the delay itself was observed to be a near-threshold phenomenon, we hypothesized that delayed firing was related to the apparent spike-frequency acceleration and might reflect a more basic feature of the near-threshold behavior of FS cells. Subsequent experiments were directed toward elucidating the basis of delay-type firing, spike-frequency acceleration, and changes in AP threshold exhibited by dFS cells.

Mechanism of Delayed Firing by FS Cells: Powerful Regulation of AP Threshold by Kv1 Potassium Channels

Closer inspection of the near-threshold behavior of FS cells revealed a slow ramp-like depolarization of, on average, 2.8 ± 1.3 mV (range, 1.3–5.3 mV; n = 16; see Experimental Procedures) (Figure 2G). The kinetics of this slow depolarization was well fit by a single exponential with a τrise of 122.7 ± 46.8 ms (range, 57.2–212.4 ms; n = 18) (Figure 2G). We considered the possibilities that this slow depolarization might be due to either the slow accumulation of an inward current during subthreshold depolarizations (such as a persistent sodium [INa,p] or a calcium-activated nonspecific cation current [ICAN]) or slow inactivation of an outward current.

Both the slow depolarization and delayed firing persisted in low (0.1 mM) external Ca2+ as well as in the presence of 50 µM CdCl2 (to block voltage-gated Ca2+ channels) or 10 µM flufenamic acid (known to block ICAN) (n = 3 for each condition; data not shown). The slow depolarization and delayed firing also persisted following replacement of half the extracellular Na+ with choline, and the slow depolarization remained after TTX application (although spiking and hence delayed firing were blocked).

These negative findings suggest that the basis of the slow depolarization seen in delay-type FS cells may be due to the slow inactivation of a K+ current. We probed the apparent FS cell input resistance (Rm) with brief hyperpolarizing current pulses superimposed upon large, just-subthreshold rectangular current steps (Figures S4C and S4D) and found that Rm increased during the course of the slow depolarization. This indicates a change in membrane conductance, further consistent with the inference that the slow depolarization is due to the inactivation of K+ channels.

Among K+ channels, those formed by subunits of the Kv1 subfamily of voltage-gated K+ channels are known to operate at near-threshold potentials and exhibit varying degrees of inactivation, depending upon precise subunit composition (Coetzee et al., 1999). Hence, we considered Kv1 channels to be strong candidates for the molecular basis of the phenomena in question. To explore this possibility, we tested the Kv1-specific peptide toxin dendrotoxin-I (DTx-I). DTx-I blocks Kv1 channels that contain at least one Kv1.1, -1.2, and/or -1.6 subunit, which are the most widely expressed Kv1 subfamily members in neocortex (Wang et al., 1994; Coetzee et al., 1999). Bath application of 50–00 nM DTx-I had no effect on FS cell Rm near resting membrane potential (81.8 ± 13.8 MΩ before and 85.9 ± 22.0 MΩ after DTx-I; n = 7; p = 0.90) (Figure 3C). This suggests that little or no DTx-sensitive current is active in FS cells at rest (which was −71.3 ± 3.1 mV). However, DTx-I produced a decrease in ITH of 43%, from 280.8 ± 66.5 to 159.2 ± 37.7 pA (n = 7; p < 0.01), and decreased the voltage threshold for AP generation from −37.5 ± 0.9 to −48.7 ± 5.6 (n = 7; p < 0.05; calculated for the first AP elicited at ITH). While DTx-I decreased ITH, the I-f curve was similar before and after DTx-I application with far-suprathreshold current injections (Figure S5). In addition, DTx-I abolished the near-threshold slow ramp depolarization seen in dFS cells and eliminated delayed firing, converting the threshold discharge pattern of FS cells in layer 2/3 barrel cortex from dFS to cFS (Figures 3A and 3B). DTx-I also eliminated the firing pattern seen in some dFS cells at intermediate current injections that included an onset spike separated from tonic firing by a pause. Interestingly, even after DTx-I application, there remained a discontinuity in the I-f relation (Figure S5), suggesting that the type-2 behavior of FS cells does not require the Kv1 current.

Figure 3
DTx Blocks Delayed Firing by FS Cells

Kv1 channels appear to have a particularly prominent role in determining the near-threshold behavior of FS cells, as bath application of 50–100 nM DTx-I had only subtle effects on the near-threshold excitability of PCs in layers 2/3 and 5 (Figure S6B). Furthermore, layer 5 PCs do not exhibit a delay to first spike but instead fire on the rising phase of the passive charging of the membrane. Moreover, DTx-I had no effect on the time from onset of current injection to the peak of the first AP elicited at ITH in PCs (control, 188.3 ± 32.3 ms; DTx-I, 185 ± 35.9 ms; n = 5; p = 0.72) (Figure S6A). To further confirm this result, we performed a simultaneous whole-cell recording of a layer 2/3 FS cell and a layer 5 PC: 100 nM DTx-I decreased ITH in the FS cell by 88 pA and eliminated delayed firing while decreasing ITH in the PC by only 15 pA, with no effect on the measured time from current onset to the peak of the first spike (data not shown).

In order to gain additional information as to the precise subunit composition of the Kv1 channel molecular complexes present in dFS cells, we repeated the above experiments using the Kv1.1-specific blocker dendrotoxin-K (DTx-K; Robertson et al., 1996), which was found to reproduce the results obtained with DTx-I (Figure S3).

Application of DTx-I or DTx-K also abolished the acceleration of firing frequency during the initial phase of sustained FS cell discharge (Figures 3D and 3E). This result suggests that, at suprathreshold voltages, the DTx-sensitive current initially suppresses firing frequency early in the train but that slow inactivation of this current during the train allows for an elevation of firing rate, producing an ISI1/ISI10 ratio >1 (i.e., spike-frequency acceleration). After blockade of this current, the instantaneous firing frequency peaked with ISI1, and the ISI1/ ISI10 ratio was less than 1, unmasking a slight spike-frequency adaptation. The ISI1/ISI10 ratio was 1.14 ± 0.08 (n = 6) under control conditions and 0.98 ± 0.09 after application of 50 nM DTx-I (p < 0.01), while the ISI1/ISIn ratio was 1.07 ± 0.13 (n = 6) under control conditions and 0.90 ± 0.12 after application of 50 nM DTx-I (p < 0.05). As the Kv1-mediated current is likely small compared to Kv3-mediated current (which accounts for >70% of total outward current recorded from neocortical FS cells at +40 mV; Erisir et al., 1999), it has only a minor influence on cellular behavior at depolarized voltages, in contrast to the dramatic influence of the Kv1-mediated current near threshold.

We next performed voltage-clamp recordings of outside-out patches pulled from FS cell somata in an attempt to analyze the DTx-sensitive current implied by the current-clamp data (Figure S8). However, current sensitive to 50–100 nM DTx-I or DTx-K was absent from these patches (Figure S8). DTx-I had no effect on the outward K+ currents in the patch, even at concentrations up to 300 nM (n = 3) (300 nM being more than ten times the IC50 of DTx-I for heterologously expressed Kv1.1, -1.2, and -1.6 homomers) (Figure S8B). From these data, we concluded that the DTx-sensitive Kv1.1-containing K+ channels responsible for regulating AP generation by FS cells are likely not present in the somatic membrane.

Expression of Kv1 Channels by FS Cells: Kv1.1 K+ Channels at the Axon Initial Segment Regulate FS Cell Excitability

As our voltage-clamp experiments suggested nonsomatic expression of Kv1 protein by FS cells, we sought to determine the site of Kv1 localization. As delayed firing was blocked by DTx-K, we investigated the precise subcellular localization of Kv1.1 protein in FS cells using immunohistochemical methods. As shown in the low-magnification image in Figure 4A, faint Kv1.1 immunoreactivity was observed throughout neocortex, consistent with previous data indicating diffuse labeling of the cortical neuropil (Wang et al., 1994). However, a small number of neuronal somata were more strongly stained. Many of these strongly Kv1.1-positive cells were also GFP positive in PV-GFP transgenic mice (Figure 4A; note that not all PV-positive cells are GFP positive in these mice, but all GFP-positive cells express PV). Double staining with antibodies directed against Kv3.1 (which is known to be expressed in all neocortical FS cells; Chow et al., 1999; Rudy and McBain, 2001) showed that 97% of these strongly Kv1.1-immunoreactive neurons also expressed Kv3.1 (Figure S9). However, further analysis suggested that the pattern of Kv1.1 immunoreactivity was not actually produced by labeling of the somatic membrane of FS cells. In higher-magnification images of single optical sections (rather than in projections of image stacks), Kv1.1 immunoreactivity did not label the perimeter of the cell body (as would be expected for a membrane protein) but instead produced diffuse intracellular labeling (Figures 4B and 4C and Figure S9). This pattern presumably represented intracellular Kv1.1 protein destined for trafficking. In addition, strongly Kv1.1-positive neurons exhibited a single, prominent, Kv1.1-immunoreactive process emanating from the soma or a proximal dendrite (Figures 4B and 4C and Movie S1). This Kv1.1 immunoreactivity was observed to begin at 7.0 ± 1.8 µM from the soma and extended along the labeled process for 19.2 ± 3.7 µM (n = 7; see Experimental Procedures). These Kv1.1-immunoreactive processes likely represented FS cell axons, based on a number of considerations. FS cells tend to have a multipolar somatic morphology (Kawaguchi and Kubota, 1997), yet only a single process was ever observed to be labeled with Kv1.1 antibody. In addition, double labeling for Kv1.1 and ankyrin-G (a marker of the AIS; Kordeli et al., 1995) in PV-GFP-transgenic mice demonstrated near-perfect overlap between Kv1.1 and ankyrin-G in GFP-positive cells (Figures 4B and 4C). Ankyrin-G labeling began 4.3 ± 2.4 µM from FS cell somata and extended 22.9 ± 2.4 µM along the axon. Such data confirmed that FS cells express Kv1.1 and that, in FS cells, prominent expression of Kv1.1 protein occurs at the AIS.

Figure 4
Kv1.1 Protein Is Enriched at the FS Cell AIS

To verify that somatic Kv1.1 protein was indeed largely cytoplasmic (and not somatic membrane associated), we performed electron microscopic analysis of Kv1.1-immunolabeled tissue (Figure 5). Similar to results obtained using confocal microscopy, Kv1.1 protein was found to be present in the cytoplasm of Kv1.1-immunoreactive interneurons (Figures 5B and 5C) but was not expressed at the somatic plasma membrane (Figure 5C) or the membrane of the proximal axon (Figure 5D). However, there was dense labeling of the AIS (Figure 5E). Together with the immunofluorescence and voltage-clamp data, the EM results illustrated the striking enrichment of Kv1.1 protein at the AIS and provide strong evidence that Kv1.1 is not expressed at the FS cell somatic membrane.

Figure 5
Kv1.1 Protein Is Present at the AIS but Absent from Somatic Membrane of FS Cells in Layer 2/3 Barrel Cortex

Local application of DTx-K to the axon confirmed that axonally localized Kv1.1-containing K+ channels were responsible for the effects of bath-applied DTx-K (Figure S10; see Supplemental Experimental Procedures). Briefly, we filled FS cells via the patch pipette with a fluorescent dye (Figure S10A1) and followed the axon back to its approximate origin (Figures S10B and S10C1). Local application of DTx-K to the FS cell somata had no apparent effect on FS cell physiology, including no effect on ITH or on delayed firing (Figure S10A), while subsequent application of DTx-K near the origin of the axon abolished the delay to first spike seen at threshold and lowered ITH (Figure S10C).

Recent work has shown a role for Kv1 channels in the regulation of axonal spike width and neurotransmitter release from layer 5 PCs (Shu et al., 2007; Kole et al., 2007). To address a potential role of Kv1 channels in axonal spike broadening and regulation of synaptic transmission by FS cells, we recorded synaptically connected FS cell→PC pairs (Figure S7 and Supplemental Experimental Procedures). Somatic depolarization sufficient to inactivate Kv1 channels (as evidenced by absence of delayed firing; Figure S7A2) had no effect on the amplitude of unitary FS cell→PC GABA-mediated postsynaptic potentials (uGPSPs) (Figures S7A and S7C): normalized GPSP amplitude was 99.3% ± 1.5% of control (n = 5). In addition, bath application of 100 nM DTx-I had no effect on the amplitude of uGPSPs (Figures S7B and S7C): normalized GPSP amplitude was 99.0% ± 2.0% of control (n = 3). This indicates that Kv1 channels at the AIS do not regulate AP-evoked GABA release from FS cells.

Kv1 Channels Effectively Control Near-Threshold Excitability of dFS Cells If Present at the AIS: Evidence from Computer Simulations

To examine whether the functions of Kv1 channels in the regulation of spike generation and near-threshold properties in FS cells were dependent upon localization to the AIS, we constructed a two-compartment FS cell model containing a soma and an axon, with spike initiation occurring in the axon. The model supported the discharge of sustained trains of brief APs at high frequency in response to applied current, with little spike-frequency adaptation (Figure 6A). A Kv1-like conductance (IKv1; see Supplemental Experimental Procedures) was added to the axon (Figure 6B) or soma (Figure 6C), and the effects on AP generation and near-threshold excitability were analyzed (see Table S1).

Figure 6
In a Two-Compartment FS Cell Model, Kv1 Conductance Effectively and Specifically Dampens Near-Threshold Excitability If Present at the Axon

Addition of IKv1 (10 nS) to the axon increased ITH (to 133 pA, from 88 pA without Kv1) and produced a positive shift in the voltage threshold for AP generation (from −38mV without to −33 mV with the Kv1, for the first spike elicited at ITH). Furthermore, axonal IKv1 produced a slow ramp depolarization in response to near-threshold applied currents and clear delay-type firing over a broad range of current amplitude (from 133 to 165 pA, or 1.24 times ITH; Figure 6B); however, it had little effect on the high-frequency discharge pattern of FS cells produced in response to far-suprathreshold current injection (Figure 6).

When IKv1 was added to the soma, a delay to first spike was observed; however, the delay in this condition was not robust, only being present across an extremely narrow range of current injections (from 124 to 126 pA, or 2.4% of ITH). As seen in Figure 6C, with a somatic IKv1, near-threshold current injections tended to produce a firing pattern characterized by an onset spike followed by a depolarized plateau or pause. With a somatic Kv1, ITH was 124 pA, while the voltage threshold for the first spike elicited at threshold was −36 mV. These intermediate values indicate that the somatic Kv1 does suppress excitability, although not nearly to the extent achieved by axonal Kv1 localization.

To demonstrate that the effects of Kv1 location on FS cell near-threshold excitability were indeed robust (and not specific to a particular set of parameter values), we systematically varied the parameters of maximal conductance and time constants for activation and inactivation of IKv1 (Table S1). Markedly increasing the Kv1 conductance at the soma (from 10 nS to, for example, 50 nS) increased ITH (from 124 to 244 pA), but even this 5-fold increase in IKv1 did not reproduce the robust delayed firing seen with an axonal IKv1 (6 pA range of applied current, or 2.6% of ITH). Moreover, increased IKv1 progressively impaired the ability to generate high-frequency firing (50 nS somatic Kv1 conductance produced an 11.8% decrease in firing frequency in response to 1 nA applied current). Furthermore, increasing the activation kinetics of somatic IKv1 could not approximate the dampening effects on near-threshold excitability achieved by IKv1 located at the axon. Decreasing the activation time constant of the somatic IKv1 (from 3.7 ms at −50 mV) to 0.1 ms (i.e., to near-instantaneous activation) had no further effect on ITH and only modestly increased the range over which delayed firing occurred (from 124 to 138 pA, an additional 12.1% of ITH). Alternatively, increasing the activation time constant (and thus slowing down onset of Kv1 activity) at the axon resulted in a firing pattern that included an onset spike followed by a pause (instead of a robust delay).

Hence, Kv1 channels exert a dampening influence on FS cell excitability, but it is the specific axonal localization that allows Kv1 channels to exert powerful control over AP generation without affecting high-frequency discharge (see Discussion). Furthermore, a robust delay to first spike as observed under experimental conditions is only seen with axonally localized Kv1 channels, and this delay depended on fast activation and slow inactivation of near-threshold operating Kv1 channels.

Kv1 Channels Regulate the Responsiveness of FS Cells to Physiological Stimuli

To address the potential role of Kv1 channels in the processing of incoming synaptic stimuli by FS cells, we performed whole-cell current-clamp recordings of layer 2/3 FS cells and recorded EPSPs evoked via extracellular stimulation (Figure S11; see Supplemental Experimental Procedures). FS cells are known to receive strong recurrent excitatory drive via powerful intracortical connectivity (Dantzker and Callaway, 2000), with large-amplitude unitary connections (Beierlein et al., 2003). DTx-K had no effect on the amplitude of extracellularly evoked EPSPs (Figure S11A). However, as DTx-K lowered the voltage threshold for AP generation in response to synaptic stimulation (Figures S11B and S11C), spikes were recruited at much lower stimulation intensities. EPSPs that failed to elicit spikes under control conditions efficiently drove spikes in the majority of stimulus presentations after application of DTx-K. Thus, Kv1 channels likely regulate the number of synchronous incoming excitatory inputs required to drive FS cells and prevent aberrant recruitment of powerful FS-mediated perisomatic inhibition.

In order to gain additional insight into the potential physiological role of this Kv1-mediated current, we probed the responsiveness of delay-type FS cells to inputs with different stimulus features and temporal dynamics. We first stimulated dFS cells with current ramps. For a given cell, we established a current amplitude sufficient to reliably drive a single AP (in 10/10 trials) in response to a 10 ms ramp. For the example shown in Figure 7A, this current amplitude was 500 pA (ramp slope of 50 nA/s). Current ramps of identical peak amplitude but intermediate slope (from, in this case, <25.0 to >14.3 nA/s) failed to elicit APs. However, ramps with a still slower rise time (≤ 14.3 nA/s) did drive dFS cells to fire APs (in 10/10 trials). Thus, while current ramps of rapid rise time efficiently drove APs in dFS cells, ramps of intermediate slope appeared to engage the dampening influence of the Kv1 current and were suppressed. Current ramps of still longer rise time but the same amplitude elicited long-latency spikes preceded by a delay, a sequence that resembled the response to threshold rectangular pulses (i.e., Kv1-mediated suppression of early spiking, followed by delayed spiking), which we inferred was due to inactivation of Kv1. We found that the suppression of firing with current ramps of intermediate slope was eliminated following addition of DTx-K (Figure 7B), confirming the role of Kv1.1-containing Kv1 channels in this phenomenon. This feature was also reproduced in our FS cell model (data not shown). Furthermore, this phenomenon was not observed in PCs (either in layer 2/3 or layer 5) (Figure 7C) and in fact is likely to be quite unique to dFS cells, as previous studies have shown that ITH decreases with progressively slower current ramps in PCs (Fricker et al., 1999). This experiment illustrates that rapid inputs could escape (or “outrun”) the Kv1 current due to faster activation kinetics of Na+ channels relative to Kv1 channels. In contrast, slower inputs that engage the initial dampening influence of the Kv1-mediated current were suppressed.

Figure 7
Kv1 Current Functions to Suppress the Response of FS Cells to Slowly Rising Inputs

We then investigated the responsiveness of dFS cells to sinusoidal current injections, using a series of four cycles to simulate repetitive EPSPs (Figure 8). In the near- but just-subthreshold range, the voltage response to successive same-amplitude cycles increased slightly, a phenomenon analogous to the slow ramp depolarization seen in response to rectangular current injection (Figure 2G, Figure 8A and 8B). This effect was not due to summation, as voltage returned to resting membrane potential after each cycle. When applied current was increased such that the stimulus reached ITH, the first AP seen was always elicited in response to the third or fourth cycle in the series (and never to the first), even though the current amplitude was the same for each cycle. This suggested that earlier cycles in the series precondition the response to later stimuli. Further increases in the amplitude of the peak current eventually elicited spiking in response to the first cycle; however, subsequent cycles in the series then produced multiple spikes. Interestingly, in response to a given suprathreshold sine wave current injection, we also observed that the timing of each AP relative to the phase of the cycle advanced with successive cycles in the series (Figure 8D). These phenomena were eliminated by DTx-K: after application of 50 nM DTx-K, FS cells fired spikes in response to the first cycle in the series, did so at lower values of ITH application, and did not exhibit phase advancement during the series. Furthermore, these phenomena were not exhibited by PCs (Figures 8C and 8D3). This experiment reinforces the idea that the Kv1 conductance serves to dampen the response to near-threshold stimuli, in effect filtering out weak, slowly rising inputs that were not sufficiently strong to overcome the Kv1 current or not sufficiently fast to avoid it. This will bias the FS cell toward responsiveness to large, rapidly rising, synchronous input. In addition, this inhibitory force is time dependent, exerting a greater dampening influence early in a series of stimuli. This phenomenon could allow subthreshold events in FS cells to precondition the response to successive inputs.

Figure 8
Kv1 Current Functions to Dynamically Regulate the Near-Threshold Excitability of FS Cells in Response to Time-Varying Inputs


In the present study, we report a feature of fast-spiking neocortical GABAergic interneurons that is likely to be critical to the function of these cells within neocortical circuits. FS cells in layer 2/3 mouse barrel cortex express Kv1.1-containing potassium channels that are specifically localized to the AIS and exert a major influence on the excitability of FS cells via regulation of AP threshold and near-threshold responsiveness. Application of the Kv1-specific blocker DTx-I or the Kv1.1 subunit-specific blocker DTx-K eliminated the characteristic delay to first spike seen with near-threshold current injections, converting the discharge pattern from that of a dFS to a pattern of cFS. In addition, the Kv1-mediated current functioned as a dampening influence in the near-threshold range, powerfully regulating the threshold for AP generation. Blockade of this current produced a large reduction in both the threshold current injection (ITH) and the voltage threshold for AP generation. Thus, by regulating AP threshold—an elemental biophysical property of all neurons—Kv1 channels have a powerful influence on FS cell excitability. However, DTx-I and DTx-K had no effect on spontaneous resting membrane potential or apparent input resistance calculated near rest, indicating that the DTx-sensitive Kv1 current in FS cells was not open appreciably at resting membrane potential. On the other hand, sufficiently large, fast inputs could “outrun” this dampening influence and efficiently drive FS cells. Furthermore, via accumulation of Kv1 channel inactivation, subthreshold depolarizations were shown to modulate FS cell excitability.

Immunofluorescent confocal microscopy showed enrichment of Kv1.1 protein at the FS cell AIS, as demonstrated by colocalization with ankyrin-G. That Kv1.1 was indeed present at the membrane of the AIS and restricted from somatic membrane was subsequently confirmed using EM. Cytoplasmic immunoreactivity evident using both immunofluorescence and EM presumably reflected labeling of Kv1.1 protein trafficking to the axon. Moreover, local application of DTX-K to the FS cell axon, but not to the soma, reproduced the effects of bath-applied toxins, firmly demonstrating the axonal localization of the Kv1 dampening mechanism. A two-compartment FS cell model containing a soma and axon showed that robust delayed firing could be produced with the addition of a Kv1-like conductance to the axonal (but not to the somatic) compartment. Moreover, modeling showed that specific localization to the axon was required for the ability of Kv1 channels to efficiently dampen FS cell excitability in the near-threshold range while at the same time leaving unaffected the suprathreshold high-frequency discharge that defines FS cells.

While many features of FS cells are coordinated for rapid responsiveness, we conclude that Kv1 channels serve as a control mechanism, allowing FS cells to select among inputs and respond preferentially to large-amplitude or synchronous inputs with rapid kinetics while filtering out slow or out-of-phase inputs. These features may be important for the rhythmogenic properties ascribed to neocortical FS cells and may have implications for the pathophysiology and potential treatment of human diseases caused by Kv1.1 mutation.

Kv1 Channels: A Newly Identified Role as Regulators of FS Cell Function

Kv1 channels are formed by heteromultimerization among eight known pore-forming subunits (Kv1.1–1.8, of which Kv1.1–1.4 and -1.6 are widely and prominently expressed in CNS neurons) and association with modulatory Kvβ subunits (Coetzee et al., 1999; Pongs et al., 1999). Most, if not all, CNS neurons express several Kv1 pore-forming proteins along with one or more Kvβ subunits; hence, there is likely to be a vast diversity of Kv1-mediated currents, including a diversity of DTx-sensitive currents (also known as D currents, or ID; Storm, 1988; Wu and Barish, 1992).

We show here that neocortical FS cells express a D-type current mediated by axonally localized Kv1.1-containing Kv1 K+ channels and that these channels form the molecular basis of delayed firing in dFS cells. The fact that the ID at the FS cell AIS contains Kv1.1 subunits has important implications, as the rapid activation kinetics of Kv1.1 is likely to be critical for the functions subserved by Kv1 channels in FS cells. In addition, involvement of Kv1.1 has further relevance, as mutations in the Kv1.1 gene are associated with multiple forms of neuropathology (see below).

While DTx-I and DTx-K had potent effects on the near-threshold excitability of FS cells, these drugs did not impair the ability of the cells to fire at high frequency in response to suprathreshold depolarizations and did not alter AP width, although DTx did abolish the initial spike-frequency acceleration observed in FS cells early in sustained high-frequency AP trains (Figure 3). Due to the slow deactivation of Kv1 channels (especially relative to Kv3 channels), an appreciable density of Kv1-mediated current would impair high-frequency firing, as was shown in our computer simulations. Thus, the magnitude of the Kv1-mediated current is likely to be small; however, operation in the near-threshold range (where few other conductances are active) combined with strategic positioning near the site of AP initiation allows these channels to powerfully yet selectively influence spike initiation.

In the model, Kv1-mediated current increased the threshold for AP generation irrespective of location. However, localization to the axon was required for Kv1 channels to rapidly curtail AP generation and produce robust delayed firing at the onset of suprathreshold depolarizations. This phenomenon depended on the ability of Kv1-mediated current to counteract regenerative inward Na+ current. Kv1 channels placed at the axon establish a shunt at the precise region of membrane that must be discharged by inward Na+ current to successfully generate APs. In this situation, Na+ and Kv1 channels both respond to the same transmembrane voltage, and the appearance of a delay depends on the relative kinetics of the two currents and the effective time constant of the axonal membrane. However, with Kv1 restricted to the soma, the axial resistance of the axonal compartment confers a partial electrical isolation of the Na+ channels at the AIS, with the somatic transmembrane voltage—and hence the activation of somatic Kv1 channels—lagging behind the voltage at the axon. This effect is compounded by the larger (slower) time constant of the soma relative to the axon. In model simulations, faster Kv1 kinetics can (but only partially) compensate for the lack of this privileged subcellular localization but can never approach the ability of axonally localized Kv1 channels to dampen near-threshold excitability. Even with a clearly unrealistic activation τ of 0.1 ms, somatically localized Kv1 channels cannot reproduce the same dampening effects on spike generation and the robust delay to first spike observed with axonal channels. Thus, the membrane of the AIS is a privileged locale, and expression of Kv1.1-containing K+ channels at this site imparts a potent control of spike generation. It is clear that a unique combination of electrophysiological properties and discrete subcellular localization converge to position Kv1.1-containing channels as specific and powerful regulators of near-threshold FS cell excitability.

Kv1 Channels Contribute to the Diversity of FS Cell Discharge Patterns and Differences in Subthreshold Behavior of FS Neurons and PCs

FS cells have been subdivided into delay-type (dFS), cells with a pause preceded by one spike at onset; continuous fast-spiking (cFS); and stuttering (sFS) types, based mainly on firing patterns in slice recordings (see Petilla Interneuron Nomenclature). These categories overlap with the dNAC, cNAC, and stuttering (cSTUT) of Gupta et al. (2000). Here, we showed that DTx converted the near-threshold firing pattern of layer 2/3 FS cells from dFS to cFS (Figure 3 and Figure S3). The same result was achieved via inactivation of the Kv1 current with conditioning prepulses (Figures S4A and S4B). We also noted that all FS cells, across all neocortical layers and irrespective of firing pattern near threshold, convert to a cFS pattern with sufficient current injection (B.D.C., E.M.G., and B.R., Soc. Neurosci., abstract 476.4, 2007). Thus, FS cells exhibit a diversity of near-threshold discharge patterns that correlates with neocortical layer and may depend upon the additional expression of other Kv1 subunits and/or different Kvβs. Such considerations could also explain the differences in the influence of Kv1 channels in FS cells versus PCs.

In layer 5 PCs, an axonal ID influences axonal spike width (Shu et al., 2007; Kole et al., 2007) and contributes to regulation of neurotransmitter release (Kole et al., 2007). However, layer 5 PCs do not exhibit a delay to first spike (see Figure 1B and Figure S6; Shu et al., 2007), and Kv1 channel blockade has minimal effects on the subthreshold excitability of these cells (Figure S6). These differences between FS cells and layer 5 PCs could be related to, among other factors, cell-type-specific differences in the composition and precise subcellular localization of Kv1 channels. Neocortical layer 5 PCs are known to prominently express Kv1.2 protein at the AIS (Inda et al., 2006) and do not seem to express Kv1.1 (Figure 4 and Figure S9). While Kv1.1 subunits promote the fastest Kv1-mediated currents, Kv1.2 has the slowest kinetics (Grissmer et al., 1994). These differences in channel composition could result in a slower ID in layer 5 PCs, which may render the ID in PCs unable to produce a robust delay to first spike. Moreover, as we show here, the Kv1.1-containing Kv1 channels in FS cell axons are precisely localized to the ankyrin-containing domain and hence presumably overlap exactly with the region of Na+ channel localization (Figure 4). However, in PCs, Kv1.2 protein is localized to the more distal AIS, with extension beyond the region of Na+ channel expression (Inda et al., 2006). Taken together, these examples illustrate how specific Kv1 subunit composition and localization to precise subcellular microdomains renders it such that a diversity of ID has differential effects on cellular function.

Neurons of the medial nucleus of the trapezoid body (MNTB) in the auditory brainstem represent another example of this principle, where somatic Kv1 currents are important in determining the temporal fidelity and patterning of APs (Dodson et al., 2002). In these cells, long rectangular current injections elicit the discharge of a single-onset AP followed by abrupt cessation of firing in the form of a plateau potential (Dodson et al., 2002). It is likely that Kv1 current at the MNTB does not activate rapidly enough to curtail the onset spike and that incomplete inactivation of the Kv1 current completely suppresses subsequent discharge.

Physiological Significance of Kv1 Channels in FS Cells

FS cells function as feed-forward inhibitory elements in neocortical circuits (Swadlow, 2003; Gabernet et al., 2005; Cruikshank et al., 2007). These cells are efficiently driven by excitatory inputs (Gibson et al., 1999; Dantzker and Callaway, 2000; Bruno and Simons, 2002) and in turn provide powerful disynaptic inhibition of their synaptic targets via proximally targeted synapses containing multiple high-fidelity release sites (Tamas et al., 1997). In addition, FS cells discharge at high frequency, as has been shown in vitro and in vivo (Swadlow, 2003; Contreras, 2004), Thus, multiple features of FS cells appear coordinated to allow efficient recruitment of these cells by ongoing neocortical activity with resultant distribution of powerful inhibition to FS cell targets.

It is of particular interest that not only do the Kv1 channels at the FS cell AIS contain Kv1.1 but that the most prominent expression of Kv1.1 protein in neocortex is in fact at the FS cell AIS. Deletion of Kv1.1 produces profound epilepsy in mice (Smart et al., 1998), although the cellular basis of this pathology remains unclear. Our results raise the intriguing possibility that the observed phenotype in Kv1.1 knockout mice might involve FS cell dysfunction. Future work is required to confirm this hypothesis and, beyond that, to resolve how seemingly enhanced excitability of FS cells might lead to epilepsy.

Furthermore, multiple point mutations in the Kv1.1 gene have been described in humans, which result in hyperexcitability phenotypes including episodic ataxia type 1 and various forms of epilepsy (Zuberi et al., 1999; Liguori et al., 2001). Mutations in Kv1.1-associated proteins leucine-rich glioma inactivated gene 1 (Lgi1) and contactin-associated protein-like 2 (Caspr2) have also been shown to cause human epilepsy syndromes (Schulte et al., 2006; Strauss et al., 2006). Deletion of a host of genes known to be expressed by FS cell interneurons leads to epilepsy in mice, including Kv3.2 (Lau et al., 2000), SCN1A (NaV1.1; Ogiwara et al., 2007), and now, Kv1.1. FS cells are critical regulators of cortical inhibition and prominently express Kv1.1 protein; hence, FS cell dysfunction may also contribute to the cellular basis by which Kv1.1 mutations cause epilepsy in human.


Slice Preparation

Experiments were carried out in accordance with the NIH Guide for the Care and Use of Laboratory Animals. Acute mouse brain slices were prepared essentially as described previously (Goldberg et al., 2005). See Supplemental Experimental Procedures for details.


To facilitate FS cell identification, we used transgenic mice in which EGFP is expressed by PV-containing interneurons (B13 line; provided by Z.J. Huang), PV being a known marker of FS cells (Kawaguchi and Kubota, 1997). See Supplemental Experimental Procedures for additional information related to recording solutions and pharmacological agents used.

ITH was defined as the first 600 ms rectangular current injection that elicited a spike. Initial firing frequency (fi) was defined as mean firing frequency during the pulse at ITH.

τm was calculated in current clamp using a single-exponential fit to the voltage produced by small hyperpolarizing current injections at Vm.

Voltage threshold was defined as dV/dt = 10 mV/ms.

The slow ramp depolarization observed in dFS cells was quantified using 600 ms current injections that remained just subthreshold (Figure 2G).

Apparent Rm was calculated as the slope of the linear fit to the V-I curve generated using 600 ms current pulses around Vm (measured at steady state).

Population data are presented as mean ± SD.

Local DTx Application

FS cells were patched using a pipette solution containing Alexa Fluor 594 (Invitrogen, Carlsbad, CA). After ~10 min (to allow dialysis of the cell), we searched for long, thin fluorescent processes studded with presumptive synaptic boutons (Goldberg et al., 2005). The axon was observed to extend for hundreds of microns beyond the soma, consistent with the wide axonal arborization of FS cells (in comparison to a more local dendritic extent; Kawaguchi and Kubota, 1997). We then followed this process back to the soma, where it was seen to initiate from a proximal dendrite. DTx-K (2 µM; diluted in ACSF) was delivered to soma and AIS via a second pipette, using pressure ejection (2–5 ms; 2–5 psi) from a Picospritzer III (General Valve Corp.). Identical ejection parameters were used for somatic and axonal applications within a given experiment.

Extracellular Stimulation

Stimulation was performed using a bipolar stimulating electrode fashioned from a fine-tipped stainless-steel electrode (FHC) glued to a tungsten rod and threaded through a glass pipette. The electrode was positioned laterally within layer 2/3, and EPSPs were evoked by 0.2 ms pulses, typically of 5–50 µA so as to stimulate a small number of axons.


For localization of Kv1.1 at the AIS in FS cells, we used mouse monoclonal anti-Kv1.1 (K36/15; Neuromab; www.neuromab.org) and rabbit anti-ankyrin-G. See Supplemental Experimental Procedures for additional details.

Electron Microscopy

Tissue was prepared from layer 2/3 mouse barrel cortex essentially as previously described (Chow et al., 1999), with additional details provided in the Supplemental Experimental Procedures.

For sampling and measurement, ultrathin sections were examined on a JEOL 1010 electron microscope. Labeled somata and neuropil were imaged at 10,000X, 12,000X, and 20,000X magnification with a 16 megapixel CCD camera (SIA). Images were examined using Image Pro Plus v5.0 software (Silver Spring, MD), at final magnifications of 60,000X to 80,000X. Micrograph composites were assembled using Adobe Photoshop (San Jose, CA).

To identify FS cells, we searched for densely Kv1-immunolabeled neurons that lacked pyramidal morphology and that also possessed a cytoplasm rich in mitochondria (a known feature of PV-positive interneurons; Kita et al., 1990; Chow et al., 1999).

Computer Simulations

A two-compartment computational model of an FS cell soma and axon was implemented using XPPAUT software (Ermentrout, 2002; www.math.pitt.edu/~bard/xpp/xpp.html). See Supplemental Experimental Procedures for current balance equations and model parameters.

Supplementary Material



The Supplemental Data for this article can be found online at http://www.neuron.org/cgi/content/full/58/3/387/DC1/.


This research was supported by NIH grants NS30989 and NS045217 and NSF IBN-0314645 to B.R. and NRSA F30 NS47882 and a fellowship from the William Randolph Hearst Foundation to E.M.G. We thank Z. Josh Huang for the gift of transgenic mice and Max Schiff and Tim Vogels for scientific discussion.


  • Azouz R, Gray CM. Dynamic spike threshold reveals a mechanism for synaptic coincidence detection in cortical neurons in vivo. Proc. Natl. Acad. Sci. USA. 2000;97:8110–8115. [PMC free article] [PubMed]
  • Beierlein M, Gibson JR, Connors BW. Two dynamically distinct inhibitory networks in layer 4 of the neocortex. J. Neurophysiol. 2003;90:2987–3000. [PubMed]
  • Bruno RM, Simons DJ. Feedforward mechanisms of excitatory and inhibitory cortical receptive fields. J. Neurosci. 2002;22:10966–10975. [PubMed]
  • Chow A, Erisir A, Farb C, Nadal MS, Ozaita A, Lau D, Welker E, Rudy B. K+ channel expression distinguishes subpopulations of parvalbumin-and somatostatin-containing neocortical interneurons. J. Neurosci. 1999;19:9332–9345. [PubMed]
  • Coetzee WA, Amarillo Y, Chiu J, Chow A, Lau D, McCormack T, Moreno H, Nadal MS, Ozaita A, Pountney D, et al. Molecular diversity of K+ channels. Ann. N Y Acad. Sci. 1999;868:233–285. [PubMed]
  • Connor JA, Stevens CF. Prediction of repetitive firing behaviour from voltage clamp data on an isolated neurone soma. J. Physiol. 1971;213:31–53. [PMC free article] [PubMed]
  • Connors BW, Gutnick MJ. Intrinsic firing patterns of diverse neocortical neurons. Trends Neurosci. 1990;13:99–104. [PubMed]
  • Contreras D. Electrophysiological classes of neocortical neurons. Neural Netw. 2004;17:633–646. [PubMed]
  • Cruikshank SJ, Lewis TJ, Connors BW. Synaptic basis for intense thalamocortical activation of feedforward inhibitory cells in neocortex. Nat. Neurosci. 2007;10:462–468. [PubMed]
  • Dantzker JL, Callaway EM. Laminar sources of synaptic input to cortical inhibitory interneurons and pyramidal neurons. Nat. Neurosci. 2000;3:701–707. [PubMed]
  • Dodson PD, Barker MC, Forsythe ID. Two heteromeric Kv1 potassium channels differentially regulate action potential firing. J. Neurosci. 2002;22:6953–6961. [PubMed]
  • Erisir A, Lau D, Rudy B, Leonard CS. Function of specific K(+) channels in sustained high-frequency firing of fast-spiking neocortical interneurons. J. Neurophysiol. 1999;82:2476–2489. [PubMed]
  • Ermentrout B. Simulating, Analyzing, and Animating Dynamical Systems: A Guide to XPPAUT for Researchers and Students. Philadelphia, PA: Society for Industrial Mathematics; 2002.
  • Fricker D, Verheugen JA, Miles R. Cell-attached measurements of the firing threshold of rat hippocampal neurones. J. Physiol. 1999;517:791–804. [PMC free article] [PubMed]
  • Gabernet L, Jadhav SP, Feldman DE, Carandini M, Scanziani M. Somatosensory integration controlled by dynamic thalamocortical feed-forward inhibition. Neuron. 2005;48:315–327. [PubMed]
  • Galarreta M, Hestrin S. A network of fast-spiking cells in the neocortex connected by electrical synapses. Nature. 1999;402:72–75. [PubMed]
  • Gibson JR, Beierlein M, Connors BW. Two networks of electrically coupled inhibitory neurons in neocortex. Nature. 1999;402:75–79. [PubMed]
  • Goldberg EM, Watanabe S, Chang SY, Joho RH, Huang ZJ, Leonard CS, Rudy B. Specific functions of synaptically localized potassium channels in synaptic transmission at the neocortical GABAergic fast-spiking cell synapse. J. Neurosci. 2005;25:5230–5235. [PubMed]
  • Grissmer S, Nguyen AN, Aiyar J, Hanson DC, Mather RJ, Gutman GA, Karmilowicz MJ, Auperin DD, Chandy KG. Pharmacological characterization of five cloned voltage-gated K+ channels, types Kv1.1, 1.2, 1.3, 1.5, and 3.1, stably expressed in mammalian cell lines. Mol. Pharmacol. 1994;45:1227–1234. [PubMed]
  • Gupta A, Wang Y, Markram H. Organizing principles for a diversity of GABAergic interneurons and synapses in the neocortex. Science. 2000;287:273–278. [PubMed]
  • Hefft S, Jonas P. Asynchronous GABA release generates long-lasting inhibition at a hippocampal interneuron-principal neuron synapse. Nat. Neurosci. 2006;8:1319–1328. [PubMed]
  • Hensch TK. Critical period plasticity in local cortical circuits. Nat. Rev. Neurosci. 2005;6:877–888. [PubMed]
  • Henze DA, Buzsaki G. Action potential threshold of hippocampal pyramidal cells in vivo is increased by recent spiking activity. Neuroscience. 2001;105:121–130. [PubMed]
  • Inda MC, DeFelipe J, Munoz A. Voltage-gated ion channels in the axon initial segment of human cortical pyramidal cells and their relationship with chandelier cells. Proc. Natl. Acad. Sci. USA. 2006;103:2920–2925. [PMC free article] [PubMed]
  • Jonas P, Bischofberger J, Fricker D, Miles R. Interneuron Diversity series: Fast in, fast out-temporal and spatial signal processing in hippocampal interneurons. Trends Neurosci. 2004;27:30–40. [PubMed]
  • Kawaguchi Y. Physiological subgroups of nonpyramidal cells with specific morphological characteristics in layer II/III of rat frontal cortex. J. Neurosci. 1995;15:2638–2655. [PubMed]
  • Kawaguchi Y, Kubota Y. GABAergic cell subtypes and their synaptic connections in rat frontal cortex. Cereb. Cortex. 1997;7:476–486. [PubMed]
  • Kita H, Kosaka T, Heizmann CW. Parvalbumin-immunoreactive neurons in the rat neostriatum: a light and electron microscopic study. Brain Res. 1990;536:1–15. [PubMed]
  • Kole MH, Letzkus JJ, Stuart GJ. Axon initial segment kv1 channels control axonal action potential waveform and synaptic efficacy. Neuron. 2007;16:633–647. [PubMed]
  • Kordeli E, Lambert S, Bennett V. AnkyrinG. A new ankyrin gene with neural-specific isoforms localized at the axonal initial segment and node of Ranvier. J. Biol. Chem. 1995;270:2352–2359. [PubMed]
  • Lau D, Vega-Saenz de Miera EC, Contreras D, Ozaita A, Harvey M, Chow A, Noebels JL, Paylor R, Morgan JI, Leonard CS, Rudy B. Impaired fast-spiking, suppressed cortical inhibition, and increased susceptibility to seizures in mice lacking Kv3.2 K+ channel proteins. J. Neurosci. 2000;20:9071–9085. [PubMed]
  • Lewis DA, Hashimoto T, Volk DW. Cortical inhibitory neurons and schizophrenia. Nat. Rev. Neurosci. 2005;6:312–324. [PubMed]
  • Liguori R, Avoni P, Baruzzi A, Di Stasi V, Montagna P. Familial continuous motor unit activity and epilepsy. Muscle Nerve. 2001;24:630–633. [PubMed]
  • McCormick DA, Connors BW, Lighthall JW, Prince DA. Comparative electrophysiology of pyramidal and sparsely spiny stellate neurons of the neocortex. J. Neurophysiol. 1985;54:782–806. [PubMed]
  • Miller KD, Pinto DJ, Simons DJ. Processing in layer 4 of the neocortical circuit: new insights from visual and somatosensory cortex. Curr. Opin. Neurobiol. 2001;11:488–489. [PubMed]
  • Noble D. Applications of Hodgkin-Huxley equations to excitable tissues. Physiol. Rev. 1966;46:1–50. [PubMed]
  • Ogiwara I, Miyamoto H, Morita N, Atapour N, Mazaki E, Inoue I, Takeuchi T, Itohara S, Yanagawa Y, Obata K, et al. Na(v)1.1 localizes to axons of parvalbumin-positive inhibitory interneurons: a circuit basis for epileptic seizures in mice carrying an Scn1a gene mutation. J. Neurosci. 2007;27:5903–5914. [PubMed]
  • Peters A, Jones EG. Classification of cortical neurons. In: Jones EG, Peters A, editors. Cerebral Cortex. Volume I. New York, NY: Plenum Press; 1984. pp. 107–120.
  • Pongs O, Leicher T, Berger M, Roeper J, Bähring R, Wray D, Giese KP, Silva AJ, Storm JF. Functional and molecular aspects of voltage-gated K+ channel beta subunits. Ann. N Y Acad. Sci. 1999;868:344–355. [PubMed]
  • Robertson B, Owen D, Stow J, Butler C, Newland C. Novel effects of dendrotoxin homologues on subtypes of mammalian Kv1 potassium channels expressed in Xenopus oocytes. FEBS Lett. 1996;383:26–30. [PubMed]
  • Rudy B, McBain CJ. Kv3 channels: voltage-gated K+ channels designed for high-frequency repetitive firing. Trends Neurosci. 2001;24:517–526. [PubMed]
  • Schulte U, Thumfart JO, Klocker N, Sailer CA, Bildl W, Biniossek M, Dehn D, Deller T, Eble S, Abbass K, et al. The epilepsy-linked Lgi1 protein assembles into presynaptic Kv1 channels and inhibits inactivation by Kvbeta1. Neuron. 2006;49:697–706. [PubMed]
  • Shu Y, Yu Y, Yang J, McCormick DA. Selective control of cortical axonal spikes by a slowly inactivating K+ current. Proc. Natl. Acad. Sci. USA. 2007;104:11453–11458. [PMC free article] [PubMed]
  • Simons DJ. Response properties of vibrissa units in rat SI somatosensory neocortex. J. Neurophysiol. 1978;41:798–820. [PubMed]
  • Smart SL, Lopantsev V, Zhang CL, Robbins CA, Wang H, Chiu SY, Schwartzkroin PA, Messing A, Tempel BL. Deletion of the K(V)1.1 potassium channel causes epilepsy in mice. Neuron. 1998;20:809–819. [PubMed]
  • Storm JF. Temporal integration by a slowly inactivating K+ current in hippocampal neurons. Nature. 1988;336:379–381. [PubMed]
  • Strauss KA, Puffenberger EG, Huentelman MJ, Gottlieb S, Dobrin SE, Parod JM, Stephan DA, Morton DH. Recessive symptomatic focal epilepsy and mutant contactin-associated protein-like 2. N. Engl. J. Med. 2006;354:1370–1377. [PubMed]
  • Swadlow HA. Fast-spike interneurons and feedforward inhibition in awake sensory neocortex. Cereb. Cortex. 2003;13:25–32. [PubMed]
  • Tamas G, Buhl EH, Somogyi P. Fast IPSPs elicited via multiple synaptic release sites by different types of GABAergic neurone in the cat visual cortex. J. Physiol. 1997;500:715–738. [PMC free article] [PubMed]
  • Tateno T, Harsch A, Robinson HP. Threshold firing frequency-current relationships of neurons in rat somatosensory cortex: type 1 and type 2 dynamics. J. Neurophysiol. 2004;92:2283–2294. [PubMed]
  • Traub RD, Bibbig A, LeBeau FE, Buhl EH, Whittington MA. Cellular mechanisms of neuronal population oscillations in the hippocampus in vitro. Annu. Rev. Neurosci. 2004;27:247–278. [PubMed]
  • Wang H, Kunkel DD, Schwartzkroin PA, Tempel BL. Localization of Kv1.1 and Kv1.2, two K channel proteins, to synaptic terminals, somata, and dendrites in the mouse brain. J. Neurosci. 1994;14:4588–4599. [PubMed]
  • Wu RL, Barish ME. Two pharmacologically and kinetically distinct transient potassium currents in cultured embryonic mouse hippocampal neurons. J. Neurosci. 1992;12:2235–2246. [PubMed]
  • Zuberi SM, Eunson LH, Spauschus A, De Silva R, Tolmie J, Wood NW, McWilliam RC, Stephenson JP, Kullmann DM, Hanna MG. A novel mutation in the human voltage-gated potassium channel gene (Kv1.1) associates with episodic ataxia type 1 and sometimes with partial epilepsy. Brain. 1999;122:817–825. [PubMed]
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...