Activity-dependent redistribution of Kv4.2 with AMPA stimulation. (A) AMPA treatment (100 μM) caused a redistribution of Kv4.2g away from synaptic sites and a punctate accumulation of Kv4.2g in the dendritic shaft as well as soma within 15 min. Redistribution of Kv4.2g was reversibly recovered within 6 h after withdraw of AMPA (right). Scale bar: 20 μm in top, 8 μm in bottom. (B) Surface proteins of Kv4.2g expressing neurons were labeled with NHS-SS-Biotin and probed with anti-GFP. In stimulated neurons the surface level of Kv4.2g was markedly decreased without a significant change in total protein level (100 μM AMPA, 15 min). Time course of surface Kv4.2g recovery showed that internalized Kv4.2g was recruited back into surface membrane within 6 h after washout of AMPA, consistent with the imaging results. Endogenous Rab4 are shown as cytoplasmic controls. (C) An example of endogenous transient and sustained K⁺ currents recorded from the same neuron before (pre) and 5 min after AMPA (50 μM) stimulation. Scale bar: 500 pA, 100 ms. We note here that we likely underestimate the degree of Kv4.2 internalization in our somatic electrophysiological recordings as we only measure K⁺ currents located in the soma and proximal dendrites whereas internalization requires synaptic activity and the majority of endogenous A-currents is in distal dendrites. (D) Normalized K⁺ current amplitudes showing that endogenous transient K⁺ current amplitudes were significantly decreased upon AMPA stimulation (76±5% of pre, n=12), while no significant change in sustained K⁺ current amplitude was observed. Error bars represent S.E.M. (E) Surface level of the endogenous Kv4.2 was determined by biotinylation assay using anti-Kv4.2 in unstimulated or AMPA-stimulated hippocampal slices (100 μM AMPA, 15 min). Consistent with Kv4.2g imaging and biotinylation data, endogenous Kv4.2 is internalized upon AMPA stimulation. Endogenous Rab4 are shown as cytoplasmic controls.

Mechanisms of AMPA-induced redistribution of Kv4.2g. (A) Upon 50 μM AMPA (15 min), most F-actin labeled spines lost Kv4.2g fluorescence. Spines were detected by TRITC-phalloidin labeling of F-actin. AMPA-induced redistribution of Kv4.2g was reduced when co-treated with the voltage-gated Na⁺ channel blocker (TTX, 1 μM), Tetanus neurotoxin (TeTN, 20 nM), the NMDAR antagonist D,L-APV (100 μM) or the membrane-permeable Ca²⁺ chelator BAPTA-AM (10 μM), 15 min prior to AMPA stimulation. Stars: Examples of Kv4.2g-positive spines; arrows: Kv4.2g-negative spines. Scale bar: 8 μm. (B) Summarized ratio of the number of Kv4.2g-positive spines to total spines for each treatment. Error bars represent S.E.M.

Time-lapse images of Kv4.2g redistribution. (A) Time-lapse images showing Kv4.2g fluorescent intensity decrease upon AMPA (50 μM) stimulation in spines of hippocampal neurons co-expressing Kv4.2g and tdTomato. (B) Kv4.2g fluorescence intensity was not significantly changed with APV co-application. Scale bar: 2 μm. (C) Time course of averaged fractional fluorescent changes (Δ;F/ F₀) of Kv4.2g in spines. AMPA stimulation resulted in a progressive decrease of Kv4.2g- specific fluorescent intensity in spines, with no significant change in tdTomato fluorescent intensity (inset). AMPA-mediated decreases in spine Kv4.2g fluorescence intensity were blocked by pre-application of 100 μM APV. Error bars represent S.E.M.

Clathrin-mediated endocytosis of Kv4.2 upon AMPA stimulation. (A and B) Dynamin inhibitory peptide (myrs-DYN, 50 μM), bath-applied 10 min prior to AMPA stimulation (50 μM, 15 min), blocked Kv4.2g internalization. Scrambled dynamin peptide (myrs-scramDYN, 50 μM) was used as a negative control. Stars: Kv4.2g-positive spines; arrows: Kv4.2g-negative spines. Scale bar: 8 μm. (C) An example of endogenous transient and sustained K⁺ currents recorded from the same neuron before (pre) and 5 min after AMPA (50 μM) stimulation with DYN (100 μg/ml) or scramDYN (100 μg/ml) included in the patch pipette. Scale bars: 200 pA, 50 ms. (D) Normalized K⁺ current amplitudes showing that AMPA (50 μM, 5 min)-induced decrease of endogenous transient K⁺ currents were blocked by DYN but not with scramDYN. No significant change in sustained K⁺ current amplitude was observed in either condition. Error bars represent S.E.M. (E) DYN but not scramDYN reduced AMPA-mediated membrane depolarization and reduced recovery times compared to control recordings with no intracellular peptide (−36.42±1.69 mV peak depolarization and 305.83±16.01 sec for recovery, n=12). Error bars represent S.E.M. Myristoylated DYN was used in imaging but not electrophysiology experiments. K⁺ channels were recorded 3–8 min after recovery from AMPA stimulation.

Kv4.2g and GluR1 differently redistributed upon AMPA stimulation. (A) Kv4.2g and GluR1, overlapped in spines in basal conditions, were not co-localized after AMPA stimulation (50 μM, 15 min). APV application resulted in Kv4.2g-only positive spines, indicating distinct internalization pathway of Kv4.2 and GluR1 upon AMPA stimulation. Stars: Kv4.2g-positive spines: arrows: Kv4.2g-negative spines. Scale bar: 8 μm. (B) Illustration of lines used to compare fluorescent intensity in spines and dendritic shaft. (C) Representative linear plot analysis of Kv4.2g (green) and GluR1 (red) distribution. X-axis indicates a length along dendrite and Y-axis fluorescent intensity (range 0–255, arbitrary units). In line scans, peaks correspond to spines containing Kv4.2g and/or GluR1 (left), or clusters in dendritic shaft (right). Upon AMPA stimulation, fluorescent peaks of both Kv4.2g and GluR1 were abolished in spines. In APV-treated spines, peaks of Kv4.2g were detected but not those of GluR1. In APV-treated dendritic shafts, peaks of Kv4.2g and GluR1 were not co-localized, once internalized by AMPA. Scale bar: 10 μm. (D) Summarized number of Kv4.2g and/or GluR1 containing spines (Kv4.2g-containing spines; 93±2.1%, GluR1-containing; 29.5±4.4%, overlapping 27±4.3%, n=417 from 11 neurons for AMPA+APV vs. 93±0.8%, 91±2.5%, 86±2%, n=439 from 8 neurons for basal). Error bars represent S.E.M.

Kv4.2 shapes mEPSCs. (A) Sample mEPSC recordings from control, Kv4.2g and Kv4.2g^W362F expressing neurons (left). Scale bar: 10 pA, 1 s. Aligned and averaged mEPSCs from the same cells as the traces to the left (right). Scale bar: 2.5 pA, 10 ms. (B) Pooled amplitude and charge data of mEPSCs from each experimental group. Both amplitude and charge are reduced in Kv4.2g expressing neurons (black bar) and enhanced in Kv4.2g^W362F neurons (white bar) compared with control (grey bar). (C) Normalized count of mEPSC charge for each experimental group. Charges for the first 200 mEPSCs recorded for each cell were used to create the distributions. With increasing functional level of Kv4.2, the charge profile broadened and the peak shifted toward larger charges. The inset shows the cumulative probability of charge (binned into 20 pA*ms bins) for Kv4.2g expressing neurons (black trace) and enhanced in Kv4.2g^W362F neurons (dashed trace) compared with control (grey trace). (D) Bath application of 4-AP (7 mM) significantly increased mEPSC charge in three of five cells tested. All recordings in A-D were made from a −60 mV holding potential. (E) AMPA stimulation reduces the A-type K⁺ current effect on mEPSCs. Example mEPSCs measured from holding potentials around rest (−60 mV) and outside of the Kv4.2 activation range (−80 mV) before and after AMPA (25μM, 2–3 min) stimulation. Each trace is the average of mEPSCs recorded for 5 min at each potential and condition. Scale bar: 5 pA, 5 ms. (F) In all six cells tested, AMPA stimulation increased the −60/−80 mV mEPSC amplitude ratio. This effect is consistent with Kv4.2 internalization as it is blocked by co-application of APV (50 μM) and mimicked by 4-AP (7 mM). Amplitude ratios in all conditions may be overestimates if we are measuring mEPSCs from a population of distal synapses in the −80 mV condition that, at −60 mV, do not rise above noise to a detectable level. These distal synapses have smaller amplitudes due to cable properties of the dendrite, thus reducing the average amplitude in the −80 recordings and increasing the ratio.

ChemLTP induces synaptic insertion of GluR1 but also Kv4.2g internalization. (A–D) Glycine (200 μM, 3–5 min) -induced LTP enhanced surface expression of GluR1 while reducing the number of Kv4.2g puncta co-labeled with synaptic markers. Stars: Kv4.2g-positive clusters; arrows: Kv4.2g-negative clusters. Scale bar: 8 μm. (E) No significant differences in intensity of immunostained GluR1 were observed between Kv4.2g expressing and control neurons. Levels of endogenous total and surface GluR1 were compared in Kv4.2g expressing (star) and control neurons (arrow). (F) Total and surface levels of Kv4.2g and GluR1 were examined before and after LTP induction in Kv4.2g neurons using biotinylation and immunoblots. (G) With chemLTP (200 μM glycine, 3–5 min), surface Kv4.2g was reduced (~82%) and surface GluR1 was enhanced (~155%). Total protein levels of Kv4.2g and GluR1 were normalized to the control protein Rab4. Surface levels of two proteins were normalized to the corresponding total protein level. Normalized total and surface protein levels were then compared before and after chemLTP induction. The enhanced level of surface GluR1 was similar in Kv4.2g and control neurons. Rab4 is shown as a cytoplasmic control. Error bars represent S.E.M. (H) Examples of the decrease in transient K⁺ currents during chemLTP. A-type K⁺ currents are decreased by Kv4.2 internalization during chemLTP. Transient currents of each group were recorded from the same cell at the times indicated. Each trace is an average of 5 sweeps. Scale bar: 1 nA, 100 ms. (I) Decrease in average transient K⁺ current amplitude during chemLTP. Transient K⁺ currents were gradually and significantly decreased after chemLTP induction in both groups. However, the remaining amplitude in Kv4.2g neurons was still greater than the basal level in control neurons, indicated by a dashed line. (J) Cumulative decrease of transient K⁺ currents during chemLTP shows identical time-courses and magnitudes for both groups. Error bars represent S.E.M.

Evidence for Kv4.2 internalization during synaptically evoked LTP. (A) An example of LTP induced in a CA1 pyramidal neuron from an organotypic slice. EPSCs were measured alternatively at −60 (red circles) and −80 mV (black circles) every 20 sec. LTP was induced using a depolarization pairing protocol. LTP was greater in the −60 mV traces than in those recorded from a holding potential of −80 mV. Inset shows EPSCs for each holding potential before (1) and 45 min after (2) LTP induction. As after AMPA stimulation in mEPSC recordings (Figure 6), LTP reduced the −60/−80 mV EPSC ratio, consistent with Kv4.2 internalization. Each trace is an average of 3 EPSCs. Scale bar: 25 pA, 25 ms. (B) Pooled LTP data from 7 cells. After confirming LTP in each neuron, we compared the degree of potentiation at the −60 mV holding potential to that recorded at −80 mV over the final 5 minutes of recordings. Potentiation was greater for the −60 mV holding potential in every cell tested. No potentiation was observed in at either holding potential in 6 cells where EPSCs from a second, control pathway (not receiving synaptic stimulation during the depolarization used to induce LTP) were evoked with a second stimulating electrode. Error bars represent S.E.M.