Paralemmin-1 controls the nanoarchitecture of the neuronal submembrane cytoskeleton

Abstract

The submembrane cytoskeleton of neurons displays a highly ordered 190-nanometer periodic actin-spectrin lattice, the membrane-associated periodic skeleton (MPS). It is involved in mechanical resilience, signaling, and action potential transmission. Here, we identify paralemmin-1 (Palm1) as a component and regulator of the MPS. Palm1 binds to the amino-terminal region of βII-spectrin, and MINFLUX microscopy localizes it in close proximity (<20 nanometers) to the actin-capping protein and MPS component adducin. Combining overexpression, knockout, and rescue experiments, we observe that the expression level of Palm1 controls the degree of periodicity of the MPS and also affects the electrophysiological properties of neurons. A single amino acid mutation (W54A) in Palm1 abolishes the MPS binding and remodeling activities of Palm1. Our findings identify Palm1 as a protein specifically dedicated to organizing the MPS and will advance the understanding of the assembly and plasticity of the actin-spectrin submembrane skeleton in general.

Fig. 1.. Palm1 is an MPS component.

(A) Palm1-interacting prey clones aligned with the βII-spectrin sequence. Each horizontal line represents one of 258 prey clones. Vertical dashed lines indicate the SMO. CH, calponin homology; PH, pleckstrin homology. (B and C) Representative two-color STED images of rat HPN (DIV 19, methanol fixation) immunolabeled for Palm1 and the βII-spectrin C terminus (B) or adducin (C). Scale bars, 1 μm. (D) Average autocorrelation (AC) and cross-correlation (CC) analyses of Palm1 and βII-spectrin of n = 27 axons from N = 3 independent neuronal cultures. (E) As (D) for Palm1 and adducin (n = 20, N = 3). (F) Confocal image of a rat HPN (DIV 19, PFA fixation) expressing endogenous Palm1 tagged with mEGFP and detected by using a nanobody against mEGFP. Arrows point at regions displayed in (G) and (H). Scale bar, 25 μm. (G) Close-up STED image of the axon coimmunostained against βII-spectrin. Scale bar, 1 μm. Dashed line on the merged image indicates the region on which the AC and CC analyses shown at the bottom have been performed. (H) Close-up STED images of a dendrite with dendritic spines. Scale bars, 1 μm.

Fig. 2.. Palm1 populates distal axonal regions before βII-spectrin, but is incorporated into the MPS later.

(A) Representative confocal images of rat HPN (DIV 1, 3, 5, 12, and 19, methanol fixation) immunolabeled against Palm1, βII-spectrin, and ankG. Scale bars, 25 μm. Shown are the maximum intensities projection of five z-stacks. Black arrowheads point at the neurite end/growth cone. (B and C) Normalized (A.U., arbitrary units) and smoothed (50 values) fluorescence intensities of Palm1, βII-spectrin, and ankG along the axons indicated by the dashed lines on the representative image at DIV 3 (B) and at DIV 19 (C). Gray area in (B) highlights the typical enrichment of Palm1 at the neurite end/growth cone. (D to G) Normalized fluorescence intensities (A.U.) and AC analyses of Palm1 [(D) and (E)] and βII-spectrin [(F) and (G)] along the proximal, middle, and distal axons at different DIVs. Axons analyzed in the same region for (D) and (E) in the proximal/middle/distal region: DIV 3: 35/31/21; DIV 5: 23/29/15; DIV 12: 25/33/26; DIV 19: 17/22/26. Axons analyzed for (F) and (G): DIV3: 24/18/20; DIV 5: 29/26/32; DIV 12: 22/13/23; DIV 19: 12/11/27. All from N = 3. Statistical analyses: One-way ANOVA with post hoc Tukey correction. All P values in file data S1. Histograms show mean ± SEM. (H) Representative image of the growth cone of an HPN (DIV 2, PFA fixation), immunolabeled against Palm1 and βII-spectrin, and phalloidin labeled for F-actin. Scale bar, 5 μm. (I) Ratio of mean fluorescence intensities between the central (CD) and the peripheral (PD) domains of growth cones, segmented according to the phalloidin signal. Growth cones analyzed: n = 56, from N = 3.

Fig. 3.. Overexpression of both Palm1 splice variants increases the complexity of neuronal morphology and enhances βII-spectrin periodicity.

(A) mRNA expression levels of Palm1 and Palm1ΔEx8 during development of HPN. Fold change relative to the housekeeping genes. N = 3, statistical analyses: One-way ANOVA with post hoc Tukey correction. All P values in file data S1. (B) Representative confocal images of rat HPN (DIV 3, PFA fixation) electroporated with plasmids encoding either YFP, YFP-CaaX, YFP-Palm1, or YFP-Palm1ΔEx8. Scale bars, 25 μm. (C) Sholl analysis of neurons overexpressing the indicated constructs. Intersections were counted every 1 μm. Cells analyzed: YFP, n = 34; YFP-CaaX, n = 24; YFP-Palm1, n = 31; YFP-Palm1ΔEx8, n = 27. All from N = 3. (D) STED images of rat HPN at DIV 3 overexpressing YFP, YFP-Palm1, and YFP-Palm1ΔEx8, and endogenous βII-spectrin. Proximal (left) and middle (right) regions of the same axon are shown. YFP was detected using nanobodies. (E) Same as (D) but for DIV 19. Scale bars, 2 μm. Corresponding axons shown in fig. S3. (F and G) AC amplitude analysis of endogenous βII-spectrin along different axonal regions in untransfected (UT) neurons, or after overexpression of YFP, YFP-Palm1, and YFP-Palm1ΔEx8 at DIV 3 (F) and DIV 19 (G). (H and I) Normalized fluorescence intensities (A.U.) of βII-spectrin along the same axonal regions measured in (F) and (G), respectively. (J and K) Correlation scatter plots of the periodicity of βII-spectrin versus Palm1 (J) or Palm1ΔEx8 (K). r, Pearson’s r coefficient; P, P value. Axons analyzed in the proximal/middle region in (F) and (H): WT, 63/62; YFP, 27/27; YFP-Palm1, 31/30; YFP-Palm1ΔEx8, 31/31; and in (G) and (I) to (K): WT, 32/35/30; YFP, 18/17/13; YFP-Palm1, 20/21/20; YFP-Palm1ΔEx8, 19/23/21. All from N = 3. Statistical analyses: one-way ANOVA with post hoc Tukey correction; all P values in file data S1. Histograms show mean ± SEM.

Fig. 4.. Palm1-KO mature neurons have a disorganized MPS.

(A) Representative STED images of βII-spectrin nanoscale organization along the proximal and middle axons in mouse WT and Palm1-KO neurons (DIV 19). Scale bars, 5 μm. (B and C) AC amplitudes calculated from the regions indicated by the dashed lines in (A) (P: proximal; M: middle). (D) AC amplitude analysis and (E) normalized fluorescence intensities (A.U.) of endogenous βII-spectrin in the proximal and middle axons of WT and Palm1-KO neurons. Axons analyzed in the proximal/middle/distal region in (D): WT, 38/43/21; Palm1-KO, 70/58/42; and in (E): WT, 22/23/11; Palm1-KO, 36/27/22. All from N = 3. Statistical analyses: one-way ANOVA with post hoc Tukey correction; all P values in file data S1. Histograms show mean ± SEM.

Fig. 5.. Palm1 reintroduction into Palm1-KO neurons rescues and enhances the periodic organization of the MPS.

(A) Representative STED images of recombinant YFP-Palm1 and endogenous βII-spectrin fluorescence periodicity, displaying an untransfected axon lacking Palm1 (top neurite) and an axon rescued by overexpression of YFP-Palm1ΔEx8 (bottom neurite; DIV 13, PFA fixation). AC analysis along the dashed lines shows the enhanced periodic pattern of βII-spectrin after Palm1ΔEx8 overexpression. Scale bar, 4 μm. (B) AC amplitude analysis of βII-spectrin along different axonal regions in untransfected Palm1-KO neurons, or after overexpression of YFP, YFP-Palm1, or YFP-Palm1ΔEx8 (DIV 13, transfection at DIV 5). (C) Normalized intensities (A.U.) of endogenous βII-spectrin along the same axonal regions analyzed in (B). (D) Correlation scatter plot of βII-spectrin periodicity versus Palm1 or (E) Palm1ΔEx8. r, Pearson’s r coefficient; P, P value. Axons analyzed for (B) to (E) in the proximal/middle/distal region: Palm1-KO, 23/23/12; YFP, 6/6/6; YFP-Palm1, 6/7/7; YFP-Palm1ΔEx8, 7/7/7. All from N = 1. Statistical analyses: one-way ANOVA with post hoc Tukey correction; all P values in file data S1. Histograms show mean ± SEM.

Fig. 6.. Molecular features involved in interactions between the MPS, Palm1, and β-spectrin.

(A) The Palm1 W54A mutation abolishes MPS integration and remodeling. STED images of rat HPN (DIV 14, PFA fixation) overexpressing YFP-Palm1(W54A), and endogenous βII-spectrin along the proximal (left) and middle axon (right). Scale bar, 1 μm. (B) AC amplitude analyses of YFP, YFP-Palm1, and YFP-Palm1(W54A) and (C) of βII-spectrin along axonal regions in untransfected (UT) neurons, or after electroporation with the indicated constructs (DIV 13 to 15). Axons analyzed for (B) and (C) in the proximal/middle region: YFP, 10/11; YFP-Palm1, 19/20; YFP-Palm1(W54A), 16/18; Untransfected, 20/21. All from N = 2 to 3. Statistical analyses: one-way ANOVA with post hoc Tukey correction; P values in file data S1. Histograms show mean ± SEM. (D) Scheme of βII-spectrin N-terminal region with Palm1-binding sequence (amino acids 261 to 307) and the interval blocking Palm1/βII-spectrin interaction (amino acids 169 to 184), as well as the collinear PIP2-binding sites in actinin-2. Below, the βII-spectrin prey constructs A to G are aligned, together with the yeast spot colonies detecting these interactions (right). (E) Selectivity of Palm1 and phosphomimetic mutant Palm1-E5 for five β-spectrin and three actinin isoforms as preys. (F) Palm1 and Palm1-E5 interact with βII-spectrin even at high 3AT concentrations. Smad/Smurf: positive control. (G) Yeast spot colonies of a C-terminal Palm1 deletion series and Palm1ΔEx8 as baits. (H) Yeast spot colonies of truncated Palm1(1-263), its ΔEx8 splice variant, and several missense mutants as baits. Yeast colony images are shown in the negative for better visualization. In experiments (D), (G), and (H), complete 3AT concentration series (0, 1, 2, 5, 10, 20, 50, 100, and 200 mM) were tested, but only selected concentrations with the most informative growth patterns are shown. In experiments (G) and (H), βII-spectrin prey G was used. (I) 3D structure of the βII-spectrin N-terminal domains relevant for Palm1 binding (movie S1), as predicted by AlphaFold2 (Q62261, amino acids 1 to 529).

Fig. 7.. Palm1 proximity to adducin at the MPS.

(A) Single-color 3D MINFLUX of periodic YFP-Palm1 along the axon of a mature neuron (DIV 18, PFA fixation). Scale bar, 500 nm. YFP-Palm1 was detected by using a nanobody against YFP in combination with DNA-PAINT. (B) YZ projection of the region highlighted in (A) shows a hollow structure, indicating that Palm1 localizes to the plasma membrane. (C) 2D projection of region indicated in (A). (D) First to fourth NN analysis of Palm1 molecules. Data from 6075 trace ID (TID, i.e., individual localization bursts) from n = 4. (E) Two-color 3D MINFLUX (exchange DNA-PAINT) reveals clear periodic patterns (highlighted by black arrowheads) for both YFP-Palm1 (blue) and adducin (magenta). Black boxes indicate Palm1 doublets flanking adducin molecules displayed in the close-ups below. (F) NN distances of Palm1 and adducin obtained from 918 TIDs corresponding to the image shown in (E).

Fig. 8.. Visualization of Palm1-induced cellular and molecular manifestations in relation to its expression levels.