Ca2+-activated KCa3.1 potassium channels contribute to the slow afterhyperpolarization in L5 neocortical pyramidal neurons

Layer 5 neocortical pyramidal neurons are known to display slow Ca2+-dependent afterhyperpolarization (sAHP) after bursts of spikes, which is similar to the sAHP in CA1 hippocampal cells. However, the mechanisms of sAHP in the neocortex remain poorly understood. Here, we identified the Ca2+-gated potassium KCa3.1 channels as contributors to sAHP in ER81-positive neocortical pyramidal neurons. Moreover, our experiments strongly suggest that the relationship between sAHP and KCa3.1 channels in a feedback mechanism underlies the adaptation of the spiking frequency of layer 5 pyramidal neurons. We demonstrated the relationship between KCa3.1 channels and sAHP using several parallel methods: electrophysiology, pharmacology, immunohistochemistry, and photoactivatable probes. Our experiments demonstrated that ER81 immunofluorescence in layer 5 co-localized with KCa3.1 immunofluorescence in the soma. Targeted Ca2+ uncaging confirmed two major features of KCa3.1 channels: preferential somatodendritic localization and Ca2+-driven gating. In addition, both the sAHP and the slow Ca2+-induced hyperpolarizing current were sensitive to TRAM-34, a selective blocker of KCa3.1 channels.

Scientific RepoRtS | (2020) 10:14484 | https://doi.org/10.1038/s41598-020-71415-x www.nature.com/scientificreports/ Although the sAHP of ER81+ neocortical pyramidal neurons and the sAHP of CA1 pyramidal neurons are similar and share some common features such as norepinephrine sensitivity 7,16 , which channels underlie the sAHP in L5 neocortical neurons remains unknown. Thus, the pharmacological properties of the unknown www.nature.com/scientificreports/ channels underlying the sAHP in L5 neurons, and how they are linked to the adaptation of spiking, an important characteristic of genetically and morphologically distinct neuronal subtypes in L5 remain to be investigated.
In the present work, we analyze the adaptation patterns of L5 neurons and how they are linked to the sAHP in a rodent model. We explore the sensitivity of the sAHP in strongly adapting L5 neurons to potassium channel blockers. Most importantly, we demonstrate the sensitivity of Ca 2+ -evoked currents L5 neurons with strong sAHP to a selective blocker of KCa3.1, as well as a co-localization of the ER81 and KCa3.1 immunoreactivity in L5 neurons.

Results
Er81-positive neurons are known to display stronger adaptation than glt-positive neurons in the visual cortex, thus they can be referred to as adapting and nonadapting, respectively 5 . To measure the strength of adaptation of L5 cells, we recorded their spiking (~ 14-15 action potentials) induced by a continuous current step (duration: 1 s) and calculated the ratio of the last interspike interval to the third interspike interval (adaptation ratio; modified from 5 ). To characterize the sAHP and correlate it with the adaption ratio, we induced 10 action potentials by a stimulation train at 50 Hz 7 and measured the amplitude of afterhyperpolarization 300 ms after the end of the train (sAHP 300 ). The amplitude of the train current was selected to evoke a single action potential with each pulse.
Previous studies have demonstrated that sAHP depends on calcium influx induced by APs 7,17 . In the next experiment, we elevated the cytoplasmic capacity of sAHP cells to buffer intracellular Ca 2+ by loading the cells with BAPTA, a fast Ca 2+ chelating agent (Fig. 2a). BAPTA loading resulted in a significant decrease of sAHP amplitudes at 300 ms compared to the pretreatment control measurements (control: 3.41 ± 0.76 mV, BAPTA: 0.90 ± 0.23 mV; paired t-test, t = − 5.53, n = 5, p < 0.01; Fig. 2b).
To explore the possibility of KCa3.1 involvement in sAHP in more detail, we recorded currents induced by Ca 2+ uncaging inside voltage-clamped neurons filled with caged Ca 2+ through the patch pipette (Fig. 4a). Uncaging in the region of interest (ROI) encompassing the proximal axon induced no detectable currents in neither sAHP+ cells (n = 5; Fig. 4b) nor in non-sAHP cells (n = 5, Fig. 4c). Uncaging in the ROI of soma + proximal basal dendrites induced hyperpolarizing currents in both sAHP+ and non-sAHP cells (Fig. 4b,c). Overall, sAHP+ cells displayed significantly larger amplitudes of Ca 2+ -induced currents compared to non-sAHP cells . Also, the sAHP+ cells displayed a slowly-deactivating component that was not pronounced in non-sAHP cells. These results allowed us to hypothesize that the late slowly deactivating current was related to the KCa3.1 channels and the sAHP. Next, we applied TRAM-34, a selective blocker of KCa3.1 channels, and compared the amplitudes of late slowly deactivating currents evoked by Ca 2+ uncaging in sAHP+ cells to the pretreatment control currents (Fig. 5a). To characterize the late current, we measured it at 3 s after flashing as the dynamics of uncaging-evoked Ca 2+ elevation is much slower compared to that induced by action potentials 18 . In our experiments, TRAM-34 application to the bath significantly reduced the amplitudes of late currents (TRAM-34: 17 ± 6 pA, control: 63 ± 19 pA; paired t-test, t = 3.36, n = 4, p < 0.05). In the next series of reference experiments, the effect of apamin, an SK channel blocker, on the late current of sAHP+ cells was tested in sAHP+ cells. SK channels are common to both sAHP+ and non-sAHP cells, and blockade of SK channels diminishes medium afterhyperpolarization 7,19 . Like KCa3.1, the SK channel can be activated intracellularly by the CaM/Ca 2+ complex. Thus, we applied apamin to  www.nature.com/scientificreports/ dissect the current underlying the sAHP. In our experiments, apamin application did not reduce late currents in sAHP+ cells compared to pretreatment control (apamin: 32 ± 6 pA, control: 33 ± 6 pA; paired t-test, n = 4, t = 1.5, P = 0.25; Fig. 5b). TRAM-34 application followed by apamin application shows that KCa3.1 channels rather than SK channels contributed to the late slowly deactivating currents evoked by Ca 2+ elevation (Fig. 5c). Adding to this, our experiments with apamin application to non-sAHP cells demonstrated that apamin greatly reduced the amplitudes of currents induced by Ca 2+ uncaging in these cells (apamin: 20 ± 3 pA; control: 170 ± 50 nA, paired t-test, t = 4.88, n = 4, p < 0.05; Fig. 5d). Previous work has identified much larger sAHPs in ETV1 (ER81)-positive neurons with slender-tufted morphology located in sublayer 5A 7 . Thus, if KCa3.1 channels mediate the sAHP, they should co-localize with the To test this hypothesis, we double-labeled neocortical slices with antibodies to ER81 and KCa3.1 (Fig. 6). Immunostaining experiments revealed a layer-like distribution of ER81 immunofluorescence in large pyramidal-like neurons (Fig. 6a). Within the layer, ER81 immunofluorescence (green) in the somas co-localized with KCa3.1 immunofluorescence in the somatic membranes (red, Fig. 6b). We analyzed 12 areas (200 × 200 μm) within the putative layer of co-localization, where we observed 221 ER81positive cells and 228 KCa3.1-positive cells in total (co-localization in the same cells: 94.4%). Some pharmacological and genetic data suggest that Kv7 channels may mediate the sAHP 20,21 . However, in our uncaging experiments, we recorded no detectable Ca 2+ mediated currents when we targeted the axon, a site of Kv7 channel expression in L5 pyramidal neurons 22 . To explore this possibility in more detail, we tested the effect of XE-991, a selective blocker of Kv7 channels on amplitudes of sAHP 300 and adaptation ratios of sAHP+ cells (Fig. 7). Paradoxically, application of XE-991 increased adaptation in sAHP+ cells (XE-991: 2.01 ± 0.17; Control: 1.79 ± 0.13, paired t-test, t = 2.68, n = 6, p < 0.05; Fig. 7a,c). Unlike sAHP + cells, non-sAHP cells displayed bursting after XE-991 application (n = 4; Fig. 7b), which confirmed the previous observations 22 . No effect on the sAHP amplitude at 300 ms was detected using our sAHP-induction protocol (XE-991: 2.69 ± 0.44 mV; control: 2.6 ± 0.4 mV; paired t-test, t = − 1.3, n = 6, P = 0.25; Fig. 7d,e). XE-991 application induced similar normalized membrane depolarization in both sAHP + and non-sAHP cell groups (sAHP + cells: 5.6 ± 1.4% non-sAHP cells: 6 ± 1.1%, ANOVA, P = 0.55, n = 6, 5).

Discussion
Here, we establish a relationship between the Ca 2+ -dependent sAHP and KCa3.1 channels in an intrinsic feedback mechanism that links activity in L5 pyramidal neurons to the adaptation of their spiking discharge. We show that the contribution of KCa3.1 channels is significant when the neuron fires a train of action potentials. To the best of our knowledge, our study is the first to link the KCa3.1 channels and the sAHP in neocortical neurons. Even more importantly, we demonstrate this link using several parallel methods: electrophysiology, pharmacology, immunohistochemistry, and photoactivatable probes. With immunohistochemistry, we identified the KCa3.1 cells as ER81-positive pyramidal neurons located in the neocortical layer 5. Targeted Ca 2+ uncaging confirmed the two main features of KCa3.1 channels: preferential somatodendritic localization 15 and Ca 2+ -driven gating. In addition, both the sAHP and the slow Ca 2+ -induced hyperpolarizing current were sensitive to TRAM-34, a blocker of KCa3.1 channels.
Our experiments reveal a strong similarity of the key features of KCa3.1 function in neocortical L5 neurons to those described in CA1 hippocampal neurons. Previous studies demonstrated that intermediate-conductance Ca 2+ -dependent KCa3.1 channels underlie the sAHP generated by trains of synaptic input or postsynaptic stimuli in CA1 hippocampal pyramidal cells 10 . The role of KCa3.1 channels in the sAHP has been independently confirmed 13 .
Compared to conventional electrophysiological studies, the uncaging of Ca 2+ activates Ca 2+ -gated channels without the need for membrane depolarization 23 . Sah and Clements 24 studied sAHPs in CA1 cells with Ca 2+ uncaging using the same photosensitive component (DMNP-EDTA) we employed in this work. The shape of the Ca 2+ -evoked currents was generally similar to that recorded in our experiments from layer 5 neocortical Kv7 potassium channels that underlie the "M current" were regarded as a candidate contributing to the sAHP 17 . However, the CaM/Ca 2+ complex downregulates the neuronal isoforms (Kv7.2/7.3) of M-channel 27 . Our experiments with Ca 2+ uncaging in the axons, the sites of Kv7 localization in L5 neurons 22 , demonstrated no Ca 2+ -evoked currents, whereas the same uncaging procedure evoked hyperpolarizing currents in the somatodendritic compartment. KCa3.1 immunolabeling was localized primarily in the somatic region of excitatory cells www.nature.com/scientificreports/ in cortical structures 15 . Moreover, using channel blockers, Tivari et al. 13 addressed the role of Kv7 channels in hippocampal CA1 pyramidal neurons and found no evidence for their contribution to the sAHP. Apart from Ca 2+ -activated K + channels, other contributors to the sAHP have been reported. Two groups emphasized an important role for a Na-K pump in producing the sAHP in hippocampal CA1 pyramidal cells 13,28 and L5 neocortical pyramidal cells 28 . Moreover, a Slack/Slo co-assembly was strongly suggested as contributing to intermediate K + conductance of the cortical neurons 29 .
In our work, the L5 cells fell into two distinct clusters: sAHP+ cells and non-sAHP cells. However, due to selection of the largest cells in the upper part of the layer, the L5 cells described in this work may not represent the whole population of layer 5.
ER81-expressing regular spiking L5 neurons with strong adaptation have been found in the barrel cortex 5 . These cells are distinguishable by the polarity of plasticity: they display deprived whisker depression but no spared whisker potentiation 30 .

Methods
Slice preparation. All experimental protocols were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and approved by the Bioethics Committee of the Institute of Higher Nervous Activity and Neurophysiology of Russian Academy of Sciences. Briefly, Wistar rats (P15-22) of both sexes were deeply anesthetized with isoflurane and decapitated. Brains were rapidly removed and placed in ice-cold artificial cerebrospinal fluid (ACSF). Frontal brain slices (350 μm) were cut using a vibratome (VT1200 S, Leica) from the primary visual cortex of the right hemisphere. The slicing angle was finely tuned to prevent damage to the apical dendrites of the surface-located L5 cells. ACSF contained 125 mM NaCl, 25 mM NaHCO 3 , 27.5 mM glucose, 2.5 mM KCl, 1.25 mM NaH 2 PO 4 , 2 mM CaCl 2 , and 1.5 mM MgCl 2 (all Sigma BioXtra-graded; pH 7.4) and was aerated with 95% O 2 and 5% CO 2 . The slices were incubated at room temperature for 90 min. Experiments commenced 90-120 min after slicing (for more detail see our previous work 18,31 ).
Electrophysiology. Patch pipettes (5 to 6 megaohms) were filled with intracellular patch solution: 132 mM K-gluconate, 20 mM KCl, 4 mM Mg-adenosine triphosphate, 0.3 mM Na2GTP, 10 mM Na-phosphocreatine, and 10 mM HEPES (pH 7.3) (all from Sigma). Brain slices were placed into a chamber continuously perfused with ACSF. Experiments were performed at or near physiological temperature (33 °C to 34 °C) and visualized under differential interference contrast (DIC) infrared optics (Axioskop 2 FS mot., Zeiss) with a Retiga Electro IR camera (QImaging). Membrane potential was recorded in whole-cell current clamp mode with an Axoclamp-2B amplifier and sampled at 20 to 50 kHz with a Digidata 1440A ADC (analog-to-digital converter) board using the Clampex 10 software (all from Molecular Devices (for more detail see our previous work 18,31 ).
The adaptation ratio was calculated as average interspike time of the 14-15-th interspike interval (ISI) normalized to the average of the third ISI (i.e. excluding the 2 first ISIs). Thus, an adaptation ratio of 2 means that on average, the ISI interval increases ~ twofold after 10 spikes evoked by a continuous current step (Modified from Groh et al. 5 ). Current amplitudes were selected to evoke ~ 14-15 spikes with a 1-s step. For our experiments, we selected the largest cells, which represent a significant portion of L5 pyramidal neurons, but not the whole population. Repetitively bursting cells were omitted and constituted ~ 1% of the cells recorded. To avoid a rundown of afterhyperpolarization due to compromised Ca 2+ buffering capacity in weak cells, we omitted the cells that died or displayed a dramatic decrease in the amplitudes of their action potentials (> 10%) in the 10 min after the end of the experimental protocols.
We recorded currents in continuous voltage-clamp mode at a holding potential of − 55 mV, which is close to the calculated Nernst reversal potential of Cl − . The calculated reversal potential of K − is ~ − 110 mV, which guaranteed a substantial driving force through K + -permeable channels.
Confocal imaging and uncaging. Live-cell imaging was performed with an LSM 5 LIVE DuoScan confocal microscope (Zeiss) equipped with a chromatically corrected water immersion lens (Plan Apochromat IR DIC 63×, 1.0 NA, Zeiss). Neurons were filled with the morphological tracer Alexa Fluor 594 hydrazide (100 μM, Invitrogen) and DMNP-EDTA (~ 5 mM; caged Ca 2+ , Biotium) by passive diffusion from the patch pipette for ~ 40 min. This caged compound is reliably decomposed inside neurons using a 405-nm laser without neuronal damage 32 . We imaged the silhouette of the cell using a 532-nm laser (50 mV) and 550 long-pass (LP) emission filter at ~ 1% of nominal laser power. We used the Alexa 594 signal to target relevant regions of the axon without inducing unwanted uncaging. A rectangular-shaped uncaging ROI was selected to encompass the target structure while minimizing the background region. The effectiveness of uncaging in increasing the local Ca 2+ concentration was confirmed in our previous work in experiments that combined uncaging and Ca 2+ imaging with a Ca 2+ sensor 18 . In experiments with repetitive uncaging and blockers, we allowed 20-30 min after pretreatment control uncaging to reload the cell with DMNP-EDTA to restore its concentration before starting a second uncaging trial with a blocker in the bath. To ensure the matching of the uncaging region with the target structures established with Alexa 594 imaging, we projected the 405-nm uncaging laser beam using a chromatically corrected lens and the same scanner used for imaging. Confocal uncaging in voltage-clamped neurons allowed us to dissect Ca 2+ -activated transmembrane ionic currents from those activated by depolarization or hyperpolarization (for more detail see our previous work 18 ).