Fig 1.
Distribution of SP in dendritic spines of hippocampal cultured neurons.
(A) Localisation of endogenous SP in dissociated cultured hippocampal neurons transfected with TMD-pHluorin (green) for the visualisation of the spine morphology and immunolabelled for SP (red). Higher magnifications of the selected area (white box) are shown in A1 (TMD-pHluorin), A2 (SP) and A3 (merged image). The arrowhead indicates a SP cluster located in a spine neck, whereas the arrow points to a spine devoid of SP. Scale bars: 5 μm in A, 2 μm in A1-3. (B) Data quantification shows that 87.0 ± 1.1% (mean ± SD) of spines overall are positive for SP. A more refined analysis indicates a preferential distribution in the spine neck compared to head or spine base compartments (neck, 68.9 ± 1.5%; head, 14.8 ± 1.1%; base, 3.2 ± 0.6%; n = 998 spines from 3 independent experiments; Table A in S1 File). (C) Recombinant mRFP-SP distribution in hippocampal neurons. Neurons were co-transfected with TMD-pHluorin (green) and mRFP-SP (red). Zoomed images of the selected area (white box) are shown in C1 (TMD-pHluorin), C2 (SP) and C3 (merged image). (D) Quantitative analysis indicates that 86.7% ± 1.1% of all spines contain mRFP-SP clusters. The distribution within the neck, head and base of the spine is similar to that observed for endogenous SP (neck, 67.0 ± 1.5%; head, 16.3 ± 1.2%; base, 3.4 ± 0.6%; n = 965 spines from 4 experiments; Table A in S1 File).
Fig 2.
SP distribution within the spine neck.
(A) Conventional fluorescence microscopy of a lentivirus-infected hippocampal neuron expressing dendra-SP (red). F-actin filaments were labelled with A647-phalloidin (green). (B) Super-resolution STORM/PALM imaging of the same dendritic segment. Single dendra and A647 fluorophore detections were rendered with a 2D Gaussian distribution with σ = 10 nm and represented in false colours (red and green, respectively). B1 and B2 are high magnifications of individual spines (white boxes in B), where SP is clearly visible along the spine neck, while phalloidin stains both neck and spine head. (C) Analysis of the full width at half maximum (FWHM) of the SP (red) and phalloidin domains (green) along a line through the spine neck in rendered STORM/PALM images. Measurements from an individual spine are shown in C1 and C2, the shape of the spine head is indicated in the upper panel in C1 (white outline based on the phalloidin staining). Note that the profile peaks in C2 were manually aligned. (D) Quantification of the FWHM of phalloidin and SP domains in spine necks. The box indicates the median, 5, 25, 75 and 95% of the spine population, the mean width is shown as a dot (n = 33 spines, 5 cultures, see also Table C in S1 File). (E,F) Analysis of SP and F-actin profiles in a 200 nm wide segment across the spine neck (red square in E1), based on the single molecule detections in pointillist images. An individual spine is shown in E1 (phalloidin: top, SP: bottom) with the corresponding detection profiles in E2. The population measurements are given in panel F. Scale bars: 2 μm in A, 200 nm in B2 and C1, 500 nm in E1.
Fig 3.
Role of SP on membrane protein diffusion in the spine neck.
(A) Schematic representation of three membrane and membrane-associated constructs used in this study: GPI-anchored GFP (GFP-GPI), TMD-pHluorin with a single transmembrane domain (TMD) and a short intracellular sequence, and pHluorin-mGluR5, containing seven TMDs and a cytoplasmic domain of 352 amino acid residues (drawn to scale). All constructs have an extracellular fluorophore used for antibody coupling with quantum dots (QD). (B,C) QD trajectories were recorded in SP-negative (B) and SP-containing spines (C). Expression of pHluorin-mGluR5 is shown in green, and SP in yellow (arrowhead); scale bar: 1 μm. (D-F) Quantification of QD diffusion in spine necks for GFP-GPI (D), TMD-pHluorin (E) and pHluorin-mGluR5 (F). Trajectories were analysed either in spines negative for SP (SP-) or positive for SP (SP+). For the latter, traces on top of SP clusters (SP area) or in areas devoid of SP (no SP area) were considered separately. The diffusion coefficient was calculated on the longitudinal component of displacements along the spine neck axis (D1Dlong; boxes indicate 5, 25, 50, 75 and 95% of all trajectories; dots: mean value; ns: not significantly different, * p < 0.05, ** p < 0.01, *** p < 0.001, KS test; n ≥ 54 trajectories, 3–5 cultures; see also Table D in S1 File).
Fig 4.
Effect of neuronal activity on SP distribution and mGluR5 diffusion.
(A) Immunocytochemistry of endogenous SP (red) and A647-phalloidin staining (green) in hippocampal neurons under control conditions (top panel) or after 30 minutes of 50 μM 4AP incubation (bottom). Scale bar: 2 μm. (B,C) Normalised total fluorescence intensity of phalloidin (B) and SP levels (C) in control and 4AP treatment, quantified within SP-positive clusters (Table B in S1 File). (D) Measurements of rendered super-resolution images show no change in the widths of the SP and phalloidin domains in the spine neck after 4AP application (4AP: n = 30 spines, 7 cultures) compared to control (see Fig 2D, Table C in S1 File). (E) Cumulative distribution of pHluorin-mGluR5 diffusion coefficients tracked in SP- (blue) and in SP+ spines, either on top (green) or outside of SP clusters (red), in control (solid lines) and after 4AP treatment (dashed lines). (F) Diffusion coefficients of pHluorin-mGluR5 in SP- and SP+ spine necks after 4AP incubation (n ≥ 135 trajectories from 3–5 cultures; Table E in S1 File).
Fig 5.
Effect of actin depolymerisation on SP distribution.
(A,B) Endogenous SP (red) and A647-phalloidin (green) labelling in control neurons (A) and after 5 min incubation with 5 μM latrunculin A (B). (C) Quantification of the total phalloidin and SP fluorescence intensity in SP clusters at different time points of latA application (n ≥ 100 cells, 3 cultures, normalised mean fluorescence ± SD). All time points were significantly different from the baseline value (phalloidin: p < 0.001 for all time points versus time zero; SP: p < 0.05 at 5 min and p < 0.001 for 10–60 min versus time zero, ANOVA). (D,E) Super-resolution imaging of lentivirus-expressed dendra-SP (red) and A647-phalloidin (green) in spines in control conditions (D) and after 5 minutes latA treatment (E). (F) Quantification of the FWHM of SP domains in spine necks in control and after latA treatment (5 μM, 5 min). Actin depolymerisation induced a significant expansion of the SP domain (p < 0.001, MW, nctr = 34, nlat = 33; Table C in S1 File). (G) Time-lapse imaging of mRFP-SP clusters in transfected neurons confirms the increase of cluster sizes during latA treatment (red circles) compared to control (black circles; mean ± SEM; p < 0.001 for the 15–30 min time points between the two conditions, ANOVA). FRAP recovery rates in absolute terms are not different between control (black squares) and latA (red squares; p = 0.7 for the pooled data from 10–30 min, MW). (H) Normalisation of the FRAP data to their respective baseline disclosed a reduced exchange rate during latA application (p = 0.001 for the pooled data from 10–30 min between the two conditions, MW). Scale bars: 5 μm in A, 2 μm in A1, 1 μm in D,E.
Fig 6.
Effect of actin depolymerisation on mGluR5 diffusion.
(A-C) Diffusion coefficients D1Dlong of pHluorin-mGluR5 in spine necks under control conditions (A, same data as in Fig 3F), during 5–10 min (B) and 15–20 min (C) of latrunculin A exposure (n ≥ 89 trajectories from 3–5 cultures; Table E in S1 File). (D) Direct comparison of the data shown in A-C shows that the diffusion coefficients within SP domains of the spine neck increase during 5 μM latA application. (E,F) No drastic changes in diffusion occur in SP-negative regions of the spine neck (E) and in SP- spines (F) during latA treatment.