Ankyrin-R regulates fast-spiking interneuron excitability through perineuronal nets and Kv3.1b K+ channels

Abstract

Neuronal ankyrins cluster and link membrane proteins to the actin and spectrin-based cytoskeleton. Among the three vertebrate ankyrins, little is known about neuronal Ankyrin-R (AnkR). We report AnkR is highly enriched in Pv⁺ fast-spiking interneurons in mouse and human. We identify AnkR-associated protein complexes including cytoskeletal proteins, cell adhesion molecules (CAMs), and perineuronal nets (PNNs). We show that loss of AnkR from forebrain interneurons reduces and disrupts PNNs, decreases anxiety-like behaviors, and changes the intrinsic excitability and firing properties of Pv⁺ fast-spiking interneurons. These changes are accompanied by a dramatic reduction in Kv3.1b K⁺ channels. We identify a novel AnkR-binding motif in Kv3.1b, and show that AnkR is both necessary and sufficient for Kv3.1b membrane localization in interneurons and at nodes of Ranvier. Thus, AnkR regulates Pv⁺ fast-spiking interneuron function by organizing ion channels, CAMs, and PNNs, and linking these to the underlying β1 spectrin-based cytoskeleton.

Keywords: ankyrin; cytoskeleton; extracellular matrix; ion channel; mouse; neuroscience; perineuronal net; spectrin.

Figure 1.. AnkR is expressed in select neurons in the cortex and hippocampus.

(A) Immunostaining of coronal mouse brain for AnkR. Arrowhead indicates a red blood cell. Scalebars, left, 250 μm, right, 20 µm. (B) Immunoblot of P1, P15, P30, and 6-month C57BL/6 brain homogenates for anti-AnkR and neurofilament-M (NFM). Quantification of immunoblot in ImageJ by comparison of AnkR to NFM loading control from three independent duplicate experiments; P1, P15 (N=four mice), P30, 6 months (N=three mice). Samples were normalized to NFM, then compared to 6-month animals. Statistical significance was determined by multiple unpaired t-tests with Welch correction and Holm-Šídák method for multiple comparisons. (C, D) Immunostaining for AnkR (red) and Parvalbumin (Pv, green) in 3-month cortex (C, N = three mice; n=923 cells) and hippocampus (D, N = three mice; n=342 cells). Scalebars, 50 µm; inset, 10 µm. (E) Immunostaining of human cortex for AnkR (red) and Parvalbumin (green). Nuclei are labeled using DAPI (blue). Scalebars, 20 µm. (F) Schematic of the Ank1 conditional allele. loxP sites (red triangles) flank exons 26 and 27 in the spectrin-binding domain (SBD), after the membrane-binding domain (MBD) and before the regulatory domain (RD). Cre-mediated recombination and removal of exons 26 and 27 generates a premature stop codon in exon 28. (G) Immunoblot of brain homogenates for anti-AnkR, and Neurofilament-M. Molecular weights are indicated at right in kDa. All error bars indicate mean ± SEM.

Figure 1—figure supplement 1.. AnkR immunostaining in control and Ank1 conditional knockout mice.

(A) Immunostaining of cortex (top) and hippocampus (bottom) for AnkR (red), Parvalbumin (Pv, green), and Somatostatin (SST, blue). Outlined arrowheads illustrate AnkR⁺Pv⁺ cells, solid arrowheads illustrate AnkR⁺SST⁺ cells, and white arrows show AnkR⁺Pv^-SST^- cells. Scalebars, 50 µm. (B) Immunostaining of cortex (top) and hippocampus (bottom) in mice with the indicated genotypes. Scalebars, 50 µm. (C) Immunostaining of cortex for AnkR (red) and AnkG (green) in the indicated genotypes. Scalebars, 20 µm.

Figure 2.. AnkR interacting proteins.

(A) Immunostaining of mouse cortex using antibodies against AnkR (red) and β1 spectrin (green). Scalebar, 20 µm. (B) Immunoblot of β1 spectrin, AnkR, and IgG immunoprecipitation reactions using antibodies against AnkR and β1 spectrin. IP, immunoprecipitation; IB, immunoblot. (C) Immunoblot of control and AnkR-deficient mouse brains using antibodies against β1 spectrin, AnkR, and neurofilament-M (NFM). Quantification of the β1 spectrin immunoblots normalized to NFM. Error bars indicate mean ± SEM. N=6. (D) Top AnkR-interacting candidates. Circle size corresponds to the mean PSM from IP mass spectrometry. Concentric circles radiating from Ank1 correspond to % reduction in PSMs from AnkR knockout mouse compared to control. Identified proteins are organized according to their putative functions. (E–H) Immunostaining of mouse cortex using antibodies against AnkR (red) and TnR (green, E), Bcan (green, F), Acan (green, G), and Vcan (green, H). Scalebars, 20 μm. (I) Immunostaining of mouse cortex and hippocampus using antibodies against AnkR (red) and fluorescent-labeled WFA (green). Scalebars, 250 µm (top) and 20 µm (bottom). (J) Immunostaining of control (left) and AnkR-deficient mouse brain (right) using antibodies against AnkR (red) and NrCAM (green). Scalebars, 50 µm. Normalized fluorescence intensity of perisomatic and axon initial segment (AIS) NrCAM in Ank1^F/F (blue) and Ank1^F/F;Nes-Cre (gray). Error bars indicate mean ± SEM. N=3. (K) Immunoblot of NrCAM, AnkR, and IgG immunoprecipitation reactions using antibodies against AnkR (top) and NrCAM (bottom). (L) Immunoblot of PlexinA4, AnkR, and IgG immunoprecipitation reactions using antibodies against AnkR (top). Immunoblot of Nrp1, AnkR, and IgG immunoprecipitation reactions using antibodies against AnkR (bottom).

Figure 2—figure supplement 1.. AnkR interacting proteins.

(A, B) Immunostaining of cortex (A) and hippocampus (B) in control and AnkR-deficient mouse brain using antibodies against AnkR (red), β1 spectrin (green), and the nuclear marker Hoechst. Scale bars, 50 µm. (C) Proteomics strategy to identify AnkR-interacting proteins. PSM, peptide spectral match. (D, E) Immunostaining of cortex (D) and hippocampus (E) using antibodies against AnkR (red) and NrCAM (green). Arrowheads in (E) indicate AnkR⁺/NrCAM⁺ neurons. Scalebars, 20 µm (D), 50 μm (E). (F) Immunofluorescence of hippocampus using fluorescent WFA (red) and antibodies against NrCAM (green). Arrowheads indicate WFA⁺ neurons. Scalebar, 25 µm. (G, H) Immunostaining of caudoputamen in control and AnkR-deficient mouse brain using fluorescent WFA (red) and antibodies against NrCAM (green). Arrowheads indicate WFA⁺ neurons. Scalebars, 50 µm. (I) Staining of deep cerebellar nuclei using antibodies against Nrp1 (green) and AnkR (red). Scalebar, 100 µm.

Figure 3.. AnkR maintains perineuronal net density and structure.

(A) Fluorescent WFA labeling of PNNs in coronal sections of cortex and hippocampus in one-month-old mice. Genotypes of the respective mice are indicated. Boxed regions are shown to the right. Scalebars, 250 µm and 100 µm. (B) Quantification of colocalization between WFA and Pv labeling per unit area (UA) in hippocampus and cortex. N=three mice/group. (C) Corrected total cell fluorescence (CTCF) in hippocampus (HPC) and cortex (CTX). N=three mice/group, n=230 cells/animal. (D) Semi-quantitative analysis of high magnification WFA showing increased hole size and disruption of PNNs in retrosplenial (RSP) cortex of 1-month mice. N=three mice/group, n=20 cells/animal. (E) Analysis of high-magnification WFA showing increased thickness of WFA and disruption of PNNs in somatosensory cortex of 1-month mice; Ank1^F/F (N=two mice; n=45 cells), Ank1^F/F; Dlx5/6-Cre (N=three mice; n=51 cells). Small solid dots indicate individual cells, large open dots indicate animal means. Error bars indicate mean ± SEM. (F) Fluorescent WFA label of PNNs in RSP cortex of 1-month-old mice. Genotypes of the respective mice are indicated. Scalebars, 5 µm. (G) Fluorescent WFA labeling of PNNs in coronal sections of cortex and hippocampus in 12-month-old mice. Genotypes of the respective mice are indicated. Boxed regions are shown to the right. Scalebars, 250 µm and 100 µm. (H) Quantification of colocalization between WFA and Pv labeling per unit area (UA) in hippocampus and cortex. N=three mice/group. (I) CTCF in HPC and CTX. N=three mice/group, n=230 cells/animal. (J) Semi-quantitative analysis of high magnification WFA showing increased hole size and disruption of PNNs in RSP cortex of 12-month mice. N=three mice/group, n=20 cells/animal. (K) Analysis of high-magnification WFA showing increased thickness of WFA and disruption of PNNs in somatosensory cortex of 12 month mice; Ank1^F/F (N=three mice; n=46 cells), Ank1^F/F; Dlx5/6-Cre (N=three mice; n=45 cells). Small solid dots indicate individual cells, large open dots indicate animal means. Error bars indicate mean ± SEM. (L) Fluorescent WFA label of PNNs in RSP cortex of 12-month-old mice. Genotypes of the respective mice are indicated. Scalebars, 5 µm. Error bars indicate mean ± SEM. N=3/group.

Figure 3—figure supplement 1.. Labeling of control and AnkR cKO neurons in cortex using fluorescent WFA and antibodies against AnkR.

Scalebar, 50 μm.

Figure 3—figure supplement 2.. WFA labeling is reduced across cortical regions.

(A, B) Corrected Total Cell Fluorescence (CTCF) across cortical regions at one- and 12 months of age. RSP, retrosplenial cortex; MOp, primary motor cortex; MOs, secondary motor cortex; SSp, primary somatosensory cortex; SSs, secondary somatosensory cortex. Error bars indicate mean ± SEM. N=3/group. (C) Schematic of cortical regions where CTCF was measured. (D) WFA labeling of PNNs showing grading scale used for analysis in Figure 3D and J. Scalebar, 5 µm. (E) WFA labeling of PNNs in somatosensory cortex of 1-month mice. Scalebar, 10 um. (F) WFA labeling of PNNs in somatosensory cortex of 12-month mice. Scalebar, 10 um.

Figure 4.. Mice lacking AnkR in GABAergic forebrain interneurons have reduced anxiety.

(A) Thirty min-long recordings of mouse trajectories in the open-field assay. Genotypes are indicated. (B) The percent of time spent in the center or perimeter of the open field. (C) The percent of the total path spent in the center or perimeter of the open field. (D) 10 minute-long recordings of mouse trajectories in the elevated plus maze. (E) The distance traveled in the open arm, closed arm, or the center of the elevated plus maze. (F) The time spent in the open arm, closed arm, or the center of the elevated plus maze. (G) The number of entries into the open arm, closed arm, or center of the elevated plus maze. In all panels, error bars indicate mean ± SEM. Black circles indicate the animals corresponding to the representative traces.

Figure 4—figure supplement 1.. Quantification of open field and elevated plus maze assays.

(A–C) Measurements of velocity (A), distance traveled (B), and rotations (C) in the open field test. (D, E) Measurements of velocity (D) and distance (E) in the elevated plus maze assay. In all panels, error bars indicate mean ± SEM. Black circles indicate the animals corresponding to the representative traces in Figure 4.

Figure 5.. WFA⁺ neurons in Ank1^F/F;Dlx5/6-Cre mice have abnormal action potentials and spike train characteristics.

(A) A representative action potential illustrating the measurement of action potential parameters. AHP, afterhyperpolarization. (B–D) Summary data showing the action potential latency (B), action potential threshold (C), and action potential amplitude (D) from the single action potential evoked by rheobase current. (E) Representative single action potentials evoked by rheobase currents from Ank1^F/F, Ank1^+/+;Dlx5/6-Cre, and Ank1^F/F;Dlx5/6-Cre cells. Action potentials are aligned at 50% of the rising phase on X axis and peak on Y axis. Note the wider action potential with shallower and delayed afterhyperpolarization in Ank1^F/F;Dlx5/6-Cre cell. (F–H) Summary data showing the action potential half-width (F), afterhyperpolarization time (G), and afterhyperpolarization amplitude (H) from the single action potentials evoked by rheobase currents. (I) Representative spike trains from Ank1^f/f, Ank1^+/+;Dlx5/6-Cre, and Ank1^F/F;Dlx5/6-Cre cells in response to 500 ms current injection. Left and right traces show the spike trains evoked by currents that are two and three times of the action potential threshold currents, respectively. Note the strong amplitude adaptation and premature depolarization block indicated by the arrow in the right trace of Ank1^F/F;Dlx5/6-Cre cell. (J) The average firing frequency during 500 ms current injection as a function of injected currents. Note, recording was stopped at maximal 1450 pA current or when cells reached depolarization block. Since 13 out of 15 Ank1^f/f;Dlx5/6-Cre+ cells reached depolarization block prior to 1450 pA current while only 3 out of 19 control cells reached depolarization block prior to 1450 pA, the firing frequency is overestimated in the high current range for Ank1^F/F;Dlx5/6-Cre neurons. (K–M) Summary data showing the average firing frequency during the first 100 ms (K), spike frequency adaptation (L), and amplitude adaptation (M) from the spike trains evoked by currents that are two times of the action potential threshold currents. (N) Summary data showing the minimal currents that caused the cells to enter depolarization block. If the maximal injected current (1450 pA) did not cause depolarization block then 1450 pA was recorded as the result. For all panels, each circle represents one neuron and the black circles indicate the representative cells in (E and I). Bar graphs represent mean ± SEM. Statistical significance was determined by one-way ANOVA or Kruskal-Wallis test with multiple comparisons.

Figure 5—figure supplement 1.. Flourescein-WFA labeling of perineuronal nets in live slices for electrophysiology.

(A) Top: low-magnification image of a coronal slice from the left hemisphere of an Ank1^f/f mouse after recording. Lines on the brain slice are the indentations resulted from the mesh used to hold the slice in place during recording. Bottom: high-magnification image of the somatosensory cortex layer five within the boxed region in the low-magnification image where recording was performed. (B, C) Same as in (A) except for Ank1^+/+;Dlx5/6-Cre (B) and Ank1^F/F;Dlx5/6-Cre (C). (D) Phase plane plots of action potentials from Ank1^F/F, Ank1^+/+;Dlx5/6-Cre, and Ank1^F/F;Dlx5/6-Cre cells. (E) Summary data of maximum dV/dt Error bars indicate mean ± SEM. Statistical significance was determined by one-way ANOVA.

Figure 5—figure supplement 2.. Analysis of miniature excitatory postsynaptic currents.

(A) Representative traces of miniature excitatory postsynaptic currents (mEPSCs) recorded at –70 mV (top) and miniature inhibitory postsynaptic currents (mIPSCs) recorded at +10 mV (bottom) from Ank1^F/F, Ank1^+/+;Dlx5/6-Cre, and Ank1^F/F;Dlx5/6-Cre cells. (B, C) Summary data of mEPSC frequency (B) and amplitude (C). (D, E) Summary data of mIPSC frequency (D) and amplitude (E). (F) Summary data of excitation/inhibition ratio that is the ratio between the integrated mEPSC charge per unit time and integrated mIPSC charge per unit time. Each circle represents one neuron and the black circles indicate the representative cells in (D). Error bars indicate mean ± SEM. Statistical significance was determined by one-way ANOVA or Kruskal-Wallis test with multiple comparisons.

Figure 6.. AnkR binds to Kv3.1b K⁺ channels and is both necessary and sufficient for its membrane localization and clustering.

(A, B) Immunostaining of 1-month-old somatosensory cortex for AnkR (red) and Kv3.1b (green). Low-magnification images are shown in (A) and high-magnification images in (B). The genotypes analyzed are shown. Scalebars, 50 µm in (A) and 10 µm in (B). (C) Immunostaining of human cortical biopsies from two separate patients using antibodies against AnkR (red) and Kv3.1b (green), and DAPI (blue) to label nuclei (Nu). Scalebars, 10 µm. (D) Quantification of Kv3.1b immunofluorescence intensity in control and Ank1^F/F;Dlx5/6-Cre mice. Error bars indicate mean ± SEM. N=3/group. (E) Immunoblots of brain homogenates from three one-month-old control and three 1-month-old AnkR-deficient brains using antibodies against Kv3.1b, AnkR, and NFM. (F) Quantification of Kv3.1b protein normalized to NFM. (G, H). Immunoblots of AnkR-GFP immunoprecipitations in cells co-expressing AnkR-GFP with Myc-tagged β1 spectrin, full length Flag-tagged Kv3.1b, or truncated versions of Flag-tagged Kv3.1b. The amino acids included in the Flag-tagged Kv3.1b truncation mutants are indicated. (I) The consensus AnkR-binding motif present in Kv3.1b and Kv3.3, but not Kv3.2. (J) Immunoblots of Kv3.1b, AnkR, and IgG immunoprecipitation reactions using antibodies against AnkR and Kv3.1b. (K) Immunostaining of ventral root nodes of Ranvier in Ank3^F/+ and Ank3^F/F;Chat-Cre mice using antibodies against AnkG (green), Kv7.2 (red), and neurofascin (NFasc, blue) on the left, and AnkR (green), Kv3.1b (red), and NFasc (blue) on the right. Scalebars, 1 μm. (L) Quantification of the percentage of nodes of Ranvier labeled for Kv7.2, Kv3.1b, AnkG, and AnkR in Ank3^F/+ and Ank3^F/F;Chat-Cre mice. Ank3^F/+ - Kv7 (N=two mice; n=217 nodes), Kv3.1b (N=three mice; n=301 nodes), AnkG (N=two mice; n=222 nodes), AnkR (N=three mice; n=286 nodes). Ank3^F/F; Chat-Cre - Kv7 (N=three mice; n=252 nodes), Kv3.1b (N=three mice; n=244 nodes), AnkG (N=three mice; n=251 nodes), AnkR (N=three mice; n=249 nodes). Error bars indicate mean ± SEM.

Figure 6—figure supplement 1.. Kv3.1b membrane localization requires AnkR.

(A) Immunolabeling of 12-month-old somatosensory cortex from Ank1^F/F and Ank1^F/F;Dlx5/6-Cre mice using antibodies against AnkR (red) and Kv3.1b (green). Scalebar, 50 μm. (B) Immunoblots of brain homogenates from three 12-month-old control and three 12-month-old AnkR-deficient brains using antibodies against Kv3.1b, AnkR, and NFM. (C) Quantification of Kv3.1b protein normalized to NFM. Error bars indicate mean ± SEM. N=3/group. (D) Immunolabeling of 1-month-old somatosensory cortex from Ank1^F/F and Ank1^F/F;Dlx5/6-Cre mice using antibodies against AnkR (red) and Kv3.1 (green). Scalebar, 50 μm. (E) Immunolabeling of 1-month-old somatosensory cortex from Ank1^F/F and Ank1^F/F;Dlx5/6-Cre mice using antibodies against AnkR (red) and Kv3.2 (green). Scalebar, 50 μm. (F) Quantification of Kv3.1 immunofluorescence intensity in somatosensory cortex of control and Ank1^F/F;Dlx5/6-Cre mice. Error bars indicate mean ± SEM. N=3/group. (G) Quantification of Kv3.2 immunofluorescence intensity in somatosensory cortex of control and Ank1^F/F;Dlx5/6-Cre mice. Error bars indicate mean ± SEM. N=3/group. (H) Immunolabeling of 1-month-old somatosensory cortex from Ank1^F/F mice using antibodies against AnkR (red) and Kv3.3 (green). Scalebar, 25 μm. (I) Immunostaining of Ank1^F/F spinal cord using antibodies against AnkR (red) and Kv3.1b (green). Scalebars, 10 μm and 1 μm. (J) Immunostaining of ventral root nodes of Ranvier in Ank3^+/+ and Ank3^F/F;Chat-Cre mice using antibodies against AnkR (red), Kv3.3 (green), and NFasc (blue). Scalebar, 1 μm. (K) The specific K⁺ channels and spectrins found at nodes of Ranvier is dictated by the Ankyrin scaffold to which they bind.

Figure 7.. AnkR is highly expressed in the perisomatic region of Pv⁺ interneurons which are surrounded by PNNs, a specialized ECM structure.

AnkR is a scaffolding protein that binds to and stabilizes PNN-associated CAMs (including NrCAM and PlexinA4) and ion channels (including Kv3.1b) by linking them to the β1-α2 spectrin-based cytoskeleton. Loss of AnkR results in (1) altered PNN morphology including reduced WFA intensity and decreased compactness of the nets; (2) molecular changes including reduced β1 spectrin, PNN-associated NrCAM, and Kv3.1b; (3) behavioral changes including decreased anxiety-like behaviors in the open field and elevated plus maze; and (4) electrophysiological changes including decreased AP latency and threshold, broader APs with shallower and delayed AHP, and decreased firing rate during current injection.

Author response image 1.