Cell-specific STORM super-resolution imaging reveals nanoscale organization of cannabinoid signaling

Journal name:
Nature Neuroscience
Volume:
18,
Pages:
75–86
Year published:
DOI:
doi:10.1038/nn.3892
Received
Accepted
Published online

Abstract

A major challenge in neuroscience is to determine the nanoscale position and quantity of signaling molecules in a cell type– and subcellular compartment–specific manner. We developed a new approach to this problem by combining cell-specific physiological and anatomical characterization with super-resolution imaging and studied the molecular and structural parameters shaping the physiological properties of synaptic endocannabinoid signaling in the mouse hippocampus. We found that axon terminals of perisomatically projecting GABAergic interneurons possessed increased CB1 receptor number, active-zone complexity and receptor/effector ratio compared with dendritically projecting interneurons, consistent with higher efficiency of cannabinoid signaling at somatic versus dendritic synapses. Furthermore, chronic Δ9-tetrahydrocannabinol administration, which reduces cannabinoid efficacy on GABA release, evoked marked CB1 downregulation in a dose-dependent manner. Full receptor recovery required several weeks after the cessation of Δ9-tetrahydrocannabinol treatment. These findings indicate that cell type–specific nanoscale analysis of endogenous protein distribution is possible in brain circuits and identify previously unknown molecular properties controlling endocannabinoid signaling and cannabis-induced cognitive dysfunction.

At a glance

Figures

  1. 3D-STORM imaging of CB1 receptors on hippocampal GABAergic axon terminals.
    Figure 1: 3D-STORM imaging of CB1 receptors on hippocampal GABAergic axon terminals.

    (a) Immunofluorescence labeling of CB1 receptors in the CA1 stratum pyramidale of CB1+/+ mice showed a basket-like arrangement of CB1-positive axon terminals encircling a CB1-immunonegative pyramidal cell somata. (b) These axon terminals could also be visualized by immunofluorescence staining for the neuropeptide CCK, a neurochemical marker of CB1-expressing interneurons. (c) STORM imaging of CB1 receptors revealed identical basket-like orientation of the same perisomatic GABAergic axon terminals. (d,e) Magnifying the epifluorescence images failed to delineate the morphological structure of the respective profiles located in the boxed region or the precise nanoscale position of CB1 receptors with the axon terminals. (f) In contrast, the improved spatial resolution of 3D-STORM microscopy made it possible to reliably discern individual CB1-positive boutons located adjacently to one another, as well as to reveal localization points representing the nanoscale position of CB1 receptors. (g,h) To validate the specificity of our approach, we also processed and imaged hippocampal sections from littermate CB1−/− mice under the same conditions. In these sections, we found an almost complete absence of STORM localization points representing CB1 in CCK-immunolabeled boutons. (i) Histogram of CB1 NLPs in individual CCK-positive varicosities in the CA1 stratum pyramidale of sections derived from littermate CB1+/+ (n = 3 animals, n = 208 axon terminals) or CB1−/− (n = 2 animals, n = 200 axon terminals) mice. Note that the markedly reduced CB1 NLP value in CB1−/− samples validates the specificity of the antibody, the staining process and the STORM imaging protocol. Scale bar in panel a also applies to panels b and c; scale bar in panel d also applies to panels e and f.

  2. Cell type-specific integrated analysis of physiological properties, anatomical parameters and nanoscale molecular imaging.
    Figure 2: Cell type–specific integrated analysis of physiological properties, anatomical parameters and nanoscale molecular imaging.

    (a) Representative voltage traces in response to depolarizing and hyperpolarizing current steps recorded in whole-cell patch-clamp configuration from a multipolar neuron in CA1 stratum radiatum revealed the characteristic accommodating, regular-spiking pattern. (b) The neuron was filled with biocytin during recording to allow post hoc reconstruction. Although the dendrites (blue) arborized in all hippocampal layers, the axon cloud (red) was restricted to the stratum pyramidale. (c) Measurement of the relative laminar location of the recovered axonal varicosities confirmed a bouton distribution biased toward the pyramidal layer. The BDI based on the descriptive statistics of the distance distribution classified this cell as a perisomatic interneuron (BDI > 1). (d) Maximum intensity projection of a deconvolved confocal stack of the biocytin-filled axon terminals. (e,f) Subsequent STORM super-resolution imaging revealed CB1 immunolabeling on the identified boutons. (f) This STORM image depicts CB1 receptor localization points in the axon terminal shown in the boxed region in e. (g) Traces recorded from another interneuron also showed accommodating firing pattern. (h) This multipolar interneuron exhibited a distinct axonal morphology from the cell presented in b. (i) BDI classified this cell as a dendritic interneuron (BDI < 0.5). (jl) Combined confocal and STORM imaging visualized the nanoscale distribution of CB1 receptors in identified boutons of the same dendritic interneuron. Boutons were imaged at a similar tissue depth (~5 μm) to equalize antibody penetration probability and the effect of light scattering. We processed and evaluated 80 interneurons in the same manner.

  3. Perisomatic boutons are larger and contain proportionally more CB1 receptors compared with dendritic boutons.
    Figure 3: Perisomatic boutons are larger and contain proportionally more CB1 receptors compared with dendritic boutons.

    (a,b) Perisomatic interneuron terminals were larger than boutons of dendritic interneurons (n = 6 cells per cell type, Mann-Whitney U test, *P = 0.004, U = 1 for area, **P = 0.002, U = 0 for perimeter). (c) Perisomatic axon terminals also carried more presynaptic CB1 receptors (*P = 0.015, U = 3). (d) CB1 NLPs showed correlation with bouton size in both interneuron types (n = 279 and 334 boutons, Spearman's correlation). (e,f) In contrast, the density of CB1 was similar between the two cell types (Mann-Whitney U test, P = 0.7 and 0.3, U = 15 and 11 for NLP over area and perimeter, respectively; ns, P > 0.05). (g,j) Biocytin-filled axon terminals of CB1-positive perisomatic (g) and dendritic (j) interneurons labeled by a mixture of Alexa488- and Alexa647-conjugated streptavidin were imaged in confocal (g,j) and STORM (h,k) modes. (i,l) A three-dimensional convex hull was fitted on the localization points to visualize the surface of axon terminals. (m) Consistent with the size difference of CB1 two-dimensional convex hulls (b), the perimeter of biocytin two-dimensional convex hulls of perisomatic boutons was significantly larger (n = 5 cells per cell type, Mann-Whitney U test, *P = 0.012, U = 0). (n) Dual-channel directSTORM imaging of an interneuron bouton labeled by Alexa488-conjugated streptavidin and immunostained for CB1 using Alexa647. (o) Fitting three-dimensional convex hulls on the biocytin or on the CB1 localization points resulted in markedly similar contours of the structure. Representative images are shown from experiments repeated successfully on two biocytin-filled perisomatic interneurons. Graphs show raw data and median ± interquartile range (IQR).

  4. Cell type-specific multicolor STORM imaging illuminates homogeneous nanoscale distribution of CB1 receptors on GABAergic axon terminals.
    Figure 4: Cell type–specific multicolor STORM imaging illuminates homogeneous nanoscale distribution of CB1 receptors on GABAergic axon terminals.

    (a,d) Deconvolved confocal images show a biocytin-filled CA1 pyramidal neuron and adjacent CB1-positive boutons. (b,e) Putative afferents of the labeled neuron were identified. (c,f) Dual-channel 3D-STORM demonstrated accumulation of the active zone protein bassoon at the contact site, indicative of a perisomatic and dendritic afferent synapses. Representative images are shown from experiments repeated on ten biocytin-filled pyramidal neurons. (g,i) Confocal images of axon terminals of biocytin-filled perisomatic and dendritic interneurons, respectively. (h,j) Dual-channel 3D-STORM revealed the nanoscale bassoon and CB1 distribution. (k,m) Localization points were visualized in three dimensions. The Euclidean distances between CB1 and bassoon localization points could be readily calculated. (l,n) To allow intermolecular distance measurements along the structure surface, we determined the shortest paths between CB1 and bassoon localization points on the bouton surface after fitting a three-dimensional convex hull on the CB1 localization points. (o) Neither the Euclidean nor the surface distance was different between axon terminals of perisomatic and dendritic interneurons (n = 311 and 141 boutons, Mann-Whitney U test, P = 0.69, U = 21420 for Euclidean distance; n = 33 and 22 boutons, P = 0.31, U = 305 for surface distance; ns, P > 0.05). Graphs show raw data and median ± IQR. (p) The density of receptors (CB1 NLP in 200 nm from random points of the surface, median ± IQR) was plotted against surface distance from bassoon. Perisynaptic (within 250 nm) and extrasynaptic (1,000–1,250 nm) CB1 densities were similar in boutons, and between cell types. Scale bar in panel g also applies to panel h; scale bar in panel i also applies to panel j.

  5. Cell type-specific and size-independent active zone architecture of GABAergic interneurons.
    Figure 5: Cell type–specific and size-independent active zone architecture of GABAergic interneurons.

    (a,d) Deconvolved confocal images of boutons of perisomatic and dendritic interneurons. (b,e) 3D-STORM images of bassoon immunostaining in identified axon terminals. (c,f) Density-based clustering (DBSCAN) revealed individual bassoon clusters (highlighted in different colors). (g) Boutons of perisomatic and dendritic interneurons contained similar NLPs (n = 349 and 374 boutons from 10 perisomatic and 12 dendritic cells, respectively, Mann-Whitney U test). (h) In contrast with the comparable bassoon NLPs (Kolmogorov-Smirnov test), perisomatic boutons contained more CB1 localization points (n = 279 and 334 boutons from 6 and 6 cells, Kolmogorov-Smirnov test), as determined by separate single-staining experiments. (i) Axon terminals of perisomatic interneurons contained more bassoon clusters (Mann-Whitney U test), whereas NLPs in individual clusters was lower (j). Graphs show raw data and median ± IQR. (k) Dual-channel STORM imaging demonstrated that CB1 NLPs in any given distance from bassoon labeling was higher in axon terminals of perisomatic interneurons (n = 80–80 randomly selected boutons, Kolmogorov-Smirnov test, plot shows median ± IQR). (l) To analyze molecular properties independently of bouton size, subsamples of boutons comprising an identical size distribution were selected at random (n = 200–200 boutons, Kolmogorov-Smirnov test). (m) In these subsamples, perisomatic axon terminals contained less bassoon, but a similar amount of CB1. (n) Although the number of bassoon clusters was similar, clusters in boutons of perisomatic interneurons had less NLPs (o). (p) CB1 NLPs in any given distance from bassoon clusters were similar in identically sized subsamples of boutons of the two interneuron types (n = 80–80 boutons, Kolmogorov-Smirnov test).

  6. Chronic THC treatment induces long-lasting loss of CB1 from the plasma membrane of GABAergic axon terminals in a dose-dependent manner.
    Figure 6: Chronic THC treatment induces long-lasting loss of CB1 from the plasma membrane of GABAergic axon terminals in a dose-dependent manner.

    (a,d) Deconvolved confocal images of axon terminals of perisomatic interneurons, recorded in slices prepared from mice treated with vehicle or THC (intraperitoneal injections twice a day for 6.5 d). (b,e) STORM imaging revealed a marked THC-evoked reduction of CB1 localization points. A CB1 antibody with improved sensitivity derived from transgenic rabbits (Online Methods) was used to enable reliable visualization of CB1 receptors at low and high levels alike. (c,f) Visualization of CB1 localization points in three dimensions demonstrated an increased intraterminal accumulation of CB1 in THC-treated animals. (g) Following chronic administration of 10 mg per kg THC, CB1 NLPs were reduced by 74% (n = 185 and 117 axon terminals from 3 vehicle- and 2 THC-treated animals, Mann-Whitney U test). After a drug-free period of 11.5 d, 53% of CB1 receptors recovered, but the reduction in NLPs (35%) was still significant (n = 283 and 113 boutons from 3 and 3 animals). A longer recovery period of 6 weeks, however, allowed complete rescue (n = 113 and 92 boutons from 2 and 3 animals). Chronic treatment with a lower (1 mg per kg) dose resulted in a slight (16%) decrease in NLPs (n = 129 and 153 boutons from 2 and 3 animals). (h) The remaining CB1 localization points were shifted toward the center of the axon terminal after a high dose of THC (unpaired two-sided t test). However, internalization was not detectable after recovery for 11.5 d or 6 weeks, or after low dose of THC. Graphs show raw data and median ± IQR.

  7. Nanoscale localization precision and relative quantification of CB1 receptors by 3D-STORM imaging
    Supplementary Fig. 1: Nanoscale localization precision and relative quantification of CB1 receptors by 3D-STORM imaging

    (a) Schematic representation of the experimental paradigm for combined confocal/STORM imaging in HEK cells. Intrinsic fluorescence intensity of EGFP linked to the extracellular N-terminus of CB1 receptors was measured by confocal microscopy. The same CB1 proteins were immunolabeled by an antibody raised against the intracellular C-terminus, and subsequently by a secondary antibody tagged with an activator-reporter dye pair Cy3-A647, which was detected by STORM microscopy. (b) Confocal image of a representative HEK cell reveals CB1 receptor accumulation within the plasma membrane. (c) At lower magnification, STORM superresolution image of the same HEK cell depicts a similar subcellular CB1 localization pattern as confocal imaging. (d) To study whether the Number of Localization Points (NLP) obtained by STORM imaging and measured in a given anatomical profile reliably predicts relative protein quantity, randomly selected HEK cells (n = 46) expressing an EGFP-CB1 fusion construct at various expression levels were imaged by both confocal and STORM microscopy. Intrinsic EGFP fluorescence intensity (FI) values were measured by confocal microscopy and determined by subtracting mean background level (assessed from 3-3 Regions of Interest (ROIs), located in the extracellular blank regions (n = 138) from every image) from the mean pixel intensity. These values are plotted against the number of localization points (NLP) obtained by STORM microscopy in each imaged cell. The strong positive linear correlation between the two imaging readouts confirms that NLP values reliably predict relative protein quantity (n = 46 cells, Spearman's rank order correlation, p < 0.001). (e) The high magnification confocal image of the boxed plasma membrane segment in b illustrates the diffraction-limited blurred nature of the EGFP fluorescence signal. (f) In contrast, at higher magnification, the nanoscale localization precision of fluorophores representing CB1 receptors by STORM imaging results in crisp visualization of the same boxed plasma membrane segment compared to the confocal image shown in e. As determined from the same images, the high resolution of the STORM image was the consequence of a fluorophore localization precision of 5 nm in the lateral (x-y), and 32 nm in the axial (z) dimensions (measured as the standard deviation of localization distribution).

  8. Quantitative evaluation of the quality of STORM imaging in brain sections
    Supplementary Fig. 2: Quantitative evaluation of the quality of STORM imaging in brain sections

    (a-c) Localization precision displayed only a modest variation (1 nm or 5 nm at x-y or z directions, respectively) as a function of axial position ie. at different z distances from the focal plane. Clusters resulting from the multiple blinking of isolated fluorophores in the STORM images of CB1-immunostaining were identified, and the distribution of the distances of individual localization points from the center of mass of the clusters was plotted. (d) Histogram shows the distribution of the number of detected photons per localization point (n = 1341 LP) above the peak detection threshold in STORM images of 40000 cycles taken from CB1-immunostaining in hippocampal sections at 5 μm depth.

  9. Comparative analysis of the passive and active electrophysiological properties of regular-spiking CB1-positive perisomatic and dendritic interneurons in the CA1 subfield of the mouse hippocampus
    Supplementary Fig. 3: Comparative analysis of the passive and active electrophysiological properties of regular-spiking CB1-positive perisomatic and dendritic interneurons in the CA1 subfield of the mouse hippocampus

    (a-e) Whole-cell patch-clamp recordings from multipolar interneurons which were post-hoc classified as either perisomatic or dendritic interneurons based on their axonal distribution and their corresponding BDI value. Electrophysiological characterization reveals that these GABAergic interneuron types do not differ in their resting membrane potential (p = 0.19, t(30) = −1.34 summarized in a), input resistance (p = 0.17, t(30) = −1.4) (b), relative sag amplitude (p = 0.24, U = 94.5) (c), rebound potential (p = 0.25, U = 89) (d), and action potential amplitude (p = 0.15, t(30) = −1.47) (e). (f-h) However, dendritic interneurons fire faster action potentials (p = 0.0024, U = 48.5) (f), show a more marked afterhyperpolarization (p = 0.037, t(30) = −2.18) (g), and exhibit more adaptation of firing upon current injection (p = 0.013, t(30) = −2.63)) (h). Statistical comparisons of 14 perisomatic and 18 dendritic cells were performed by unpaired two-sided t-test (a,b,e,g,h) or by Mann-Whitney U test (c,d,f). Graphs show raw data and median±IQR. The measured data for each cell and for each parameter are presented in Supplementary Table.

  10. The quantitative power of CB1-immunostaining and subsequent STORM imaging is robust to various experimental conditions
    Supplementary Fig. 4: The quantitative power of CB1-immunostaining and subsequent STORM imaging is robust to various experimental conditions

    (a) Electrophysiological recordings from acute hippocampal slices, tissue processing and immunostaining were carried out in parallel in two different laboratories in Irvine, CA, USA and in Budapest, Hungary. Undertaking all STORM imaging on the same setup in the latter laboratory resulted in similar normalized distribution of the number of CB1 localization points per boutons, as demonstrated in the cumulative probability plot showing the NLP values per bouton, normalized for the cell mean. Statistical analysis was done by Kolmogorov-Smirnov test, p > 0.05, n = 150 and n = 279 boutons from n = 7 and n = 6 basket cells filled and processed in Irvine and Budapest, respectively, were used in the experiment. (b) Immunostaining with two independent primary antibodies raised either in a guinea-pig52, or a transgenic (Tg) rabbit that overexpress the neonatal Fc receptor (FcRn) and have improved humoral immune responses58, and with the two corresponding secondary antibodies custom-labeled with fluorophore pairs for STORM imaging, resulted in similar normalized distribution of the number of CB1 localization points per bouton (n = 279 and n = 302 boutons from n = 6-6 basket cells stained with anti-CB1 antibody produced in guinea-pig and rabbit, respectively, Kolmogorov-Smirnov test, p > 0.1). (c-d) Whole-cell patch-clamp recordings and biocytin-filling did not affect bouton size or CB1 NLP. Biocytin-filled boutons, and an equal number of neighboring biocytin-negative, randomly chosen boutons were selected from the same images for analysis. The size of the boutons was determined by using the CB1 STORM signal outlining the terminals. No significant differences were found in the number of CB1 localization points, or in the size of boutons with or without biocytin-labeling (n = 202-202 axon terminals, Mann-Whitney U test, p = 0.20, U = 4563.5 and p = 0.31, U = 29232, respectively). Graphs show raw data and median±IQR.

  11. Comparative morphological analysis of the axon terminals of CB1-positive perisomatic and dendritic interneurons
    Supplementary Fig. 5: Comparative morphological analysis of the axon terminals of CB1-positive perisomatic and dendritic interneurons

    (a-c) Deconvolved confocal images of n = 279 and n = 334 biocytin-filled boutons from n = 6 perisomatic and n = 6 dendritic cells, respectively, were analyzed for their shape descriptors. (a) Circularity describes how close the relation between the area and perimeter of the bouton shape is to that of a perfect circle. (b) Aspect ratio reflects the degree of elongation. (c) Solidity measures the area of the shape divided by its convex area, thereby describing convexity and smoothness of edges of the bouton shape. No significant differences are detected between cell types in either the circularity (p = 0.82, U = 16), aspect ratio (p = 0.70, U = 15), or solidity (p = 0.94, U = 17) of their boutons (Mann-Whitney U test). Graphs show raw data and median±IQR.

  12. CB1 density distribution on axon terminals of identified interneurons can be reliably determined under the specific imaging conditions used throughout the study
    Supplementary Fig. 6: CB1 density distribution on axon terminals of identified interneurons can be reliably determined under the specific imaging conditions used throughout the study

    (a) To assess whether the local density of STORM labeling on axon terminals is within the dynamic range of the measurement, 10-10 ROIs (30 by 30 confocal pixels each; pixel size = 0.0064 μm2) were placed on the plasma membrane of HEK cells and biocytin-filled axon terminals to measure labeling density. Comparison of the labeling density did not reveal significant difference between boutons of perisomatic and dendritic interneurons (n = 6 cells/cell type, Mann-Whitney U test, p = 0.59, U = 14), and the vast majority of boutons fell within the range defined by the two HEK cells with the highest and lowest expression levels. Graphs show raw data and median±IQR. (b) The NLP values of boutons were averaged for each interneuron, and the mean±SD range of NLP density for interneurons was determined. This range well represents the STORM CB1 labeling density found on GABAergic axon terminals, as it includes 10 of the 12 interneurons. HEK cells (n = 20) within this range show highly linear positive correlation between intrinsic fluorescence intensity and NLP (Spearman's rank order correlation, p < 0.0001). (c) The distribution of local densities (neighboring localizations within 50 nm 3D distance from each localization point) on the boutons from (a) shows a heavy right tail indicating that the signal is not likely to be saturated. (d) The different size of perisomatic and dendritic boutons is evident regardless of the duration of the imaging. Using the same data set as in Figure 3, the perimeter of a 2D-convex hull fitted on the localization points detected before the given period of imaging was measured, and was plotted as the ratio of the perimeter of dendritic axon terminals over perisomatic boutons (median±IQR of the mean values of 6 cells per cell type is shown). The difference was significant at every time point after 250 cycles (n = 6 cells/cell type, Mann-Whitney U test, p = 0.015, U = 3). (e) Similarly, the finding that axon terminals of dendritic cells had less CB1 receptors than perisomatic boutons was also independent of the imaging duration after 350 cycles, as the difference was significant at every time point afterwards (n = 6 cells/cell type, Mann-Whitney U test, p = 0.041, U = 5). The same approach was used to compare the number of CB1 localization points per bouton between interneuron types. (f) The variance of nearest-neighbor distances between CB1 localization points within boutons was constant in both cell types after 400 cycles into imaging, indicating that running the imaging for up to 1000 cycles for each image provides reliable measurement of CB1 distribution within GABAergic axon terminals.

  13. Layer-specific electron microscopic analysis of CB1-containing GABAergic axon terminals
    Supplementary Fig. 7: Layer-specific electron microscopic analysis of CB1-containing GABAergic axon terminals

    (a,b) To test whether quantitative information collected by 3D-STORM microscopy is comparable with data acquired by other well-established quantitative anatomical approaches, hippocampal sections were processed for immunogold labeling of CB1 receptors and analyzed by transmission electron microscopy (TEM). Random cross-sections of CB1-positive GABAergic boutons were selected from the stratum pyramidale and stratum radiatum of the CA1 subfield. The presumed GABAergic nature of an axon terminal was established by the presence of symmetrical synapses (arrowheads). Note that the vast majority of silver-intensified gold particles representing CB1 receptors are attached to the intracellular surface of the plasma membrane of axon terminals. (c,d) CB1-positive boutons found in the pyramidal layer are significantly larger (n = 3 animals, 100 boutons per layer per animal, paired two-sided t-test, p = 0.025, t(2) = 6.26 and p = 0.033, t(2) = 5.33 for area and perimeter, respectively). (g) In addition, the number of gold particles (NGP) on the plasma membrane of these perisomatic axon terminals is considerably higher than on boutons observed in the stratum radiatum (paired two-sided t-test, p = 0.036, t(2) = 5.13). (h) In contrast, the density of gold particles was not significantly different on boutons located in the two hippocampal layers (paired two-sided t-test, p = 0.064, t(2) = 3.76). These observations obtained at the synaptic population level are in general agreement with the cell-type-specific findings obtained by STORM imaging (see Figure 4). Graphs show raw data and median±IQR.

  14. Visualization of the localization coordinates of the STORM signal by 3D rendering and calculation of membrane surface distances
    Supplementary Fig. 8: Visualization of the localization coordinates of the STORM signal by 3D rendering and calculation of membrane surface distances

    (a) Analysis of nanoscale distribution of STORM data was carried out using the x, y and z coordinates of each individual localization point. Their 3D positions are illustrated as atoms in Visual Molecular Dynamics. (b) First, a convex hull was fitted on the localization points, giving an approximation of the structure surface. Note that single-molecule localization-microscopy gives a more precise estimate of structural dimensions than electron microscopy22, which underestimates these parameters due to anisotropic shrinkage during tissue dehydration, as well as compared to confocal microscopy, which overestimates size due to light diffraction25. Indeed, the average bouton size based on the cross-sectional area of the polyhedrons of perisomatic and dendritic terminals are 1.33±0.04 μm2 and 0.97±0.03 μm2, respectively (mean±SEM), when measured by 3D-STORM microscopy. These values fall in between the range of electron microscopic (0.29±0.01 μm2 and 0.22±0.01 μm2); and confocal measurements (1.65±0.04 μm2 and 1.20±0.03 μm2). These data independently indicate that the convex hull approach resulted in a realistic representation of the imaged part of the bouton. This feature has previously been an exclusive advantage of the laborious serial reconstruction of electron micrographs. (c) Secondly, the distance of each localization point from the convex hull was measured, and only localizations within 80 nm from the surface were included in further analysis. This approach could also be utilized to measure the proportion of membrane-attached localization points and those which are likely intracellular receptors. These subcellularly segregated receptor populations presumably represent functionally different receptor pools, hence it is important to distinguish between them, when quantifying receptor distribution or pathological redistribution in a cell-type-specific manner. (d) Finally, the distance of each localization point from the active zone was measured as the shortest route on the polyhedron surface.

  15. Validation of the hull surface distance calculations and density measurements on model boutons
    Supplementary Fig. 9: Validation of the hull surface distance calculations and density measurements on model boutons

    (a) To measure the sensitivity of the local density analysis on the polyhedron surface, a simulation of homogeneous and artificially enriched receptor distribution was conducted. “CB1 localization points” were randomly placed on model “bouton” comprising a sphere of 1000 nm radius, excluding a 230 nm radius area of the surface as the “active zone”. A convex hull was fitted onto these points. (b-c) To model different levels of potential perisynaptic enrichment, an increasing number of localization points were placed in random positions of the surface within a 250 nm radius of the active zone. (d) To measure localization point density on the surface, reference points were also placed at 40 nm spacing over the entire sphere, excluding the active zone. CB1 density, defined as the number of localization points within 200 nm distance of each reference point, was determined along with the distance from the edge of the active zone measured on the hull surface. The entire process was repeated 50 times for each enrichment condition. Finally, reference points per individual boutons were binned based on the distance from the synapse and averaged by bin. Altogether 50 mean values per bin per condition were used for generating graphs (median±IQR) and calculating p values using unpaired two-sided t-tests to compare perisynaptic and extrasynaptic regions. The difference between the 0–250 nm and the 1000-1250 nm bins is already significant at 20% (p < 0.0001), but not at 10% enrichment (p = 0.84). The graph for 1000% is not shown for sake of clarity, but was also highly significant (p < 0.0001).

  16. Combination of in vivo single neuron recording in a behaving mouse with post-hoc anatomical characterization and molecular imaging at the nanoscale
    Supplementary Fig. 10: Combination of in vivo single neuron recording in a behaving mouse with post-hoc anatomical characterization and molecular imaging at the nanoscale

    (a) Example traces from juxtacellular recording of a CA1 interneuron in an anesthesia-free mouse running on a spherical treadmill. A 1 sec-long recording, filtered for local field potential, and, spikes, and the raw trace are shown below each other, respectively. (b) After the recording, the cell was filled with neurobiotin. Morphological analysis of the labeled neuron was performed by confocal microscopy. Part of the axonal arbor (red), as well as its soma and some dendrites (blue), were reconstructed from four consecutive 20 μm thick sections by Neurolucida. (c) Laminar distribution analysis of more than 3000 axonal varicosities recovered from 14 serial sections classifies this cell as a dendritic interneuron (BDI < 0.5). (d) Maximum intensity projection of a deconvolved confocal image stack of individual axon terminals from the same in vivo recorded interneuron. (e) CB1-immunolabeling on these identified axon terminals is visualized by overlaying the corresponding confocal and STORM images. (f) Magnification of the boxed region of the STORM image in e illustrates the intraterminal distribution of CB1 localization points at the nanoscale level and reveals a qualitatively similar pattern as that observed in the case of in vitro-labeled interneurons in Figure 3. (g) Cumulative probability plots of normalized CB1 NLP found on individual terminals of in vivo- or in vitro-filled dendritic interneurons. The distribution of data obtained from the two experimental preparations were identical (Kolmogorov-Smirnov test, p > 0.1). Data were collected from single CB1 STORM imaging of 129 boutons from one in vivo, and of 141 boutons from 5 in vitro-filled cells and were normalized to the mean NLP value of the respective cell. (h) The variance of CB1 nearest neighbor distances is also comparable under the different experimental conditions indicating similar CB1 distribution at the nanoscale level in boutons of in vivo- or in vitro-filled dendritic interneurons (Kolmogorov-Smirnov test, p > 0.1). (i) In line with the in vitro findings, at the single bouton level, the number of CB1 localization points correlates linearly with bouton size (n = 129 axon terminals, Spearman's rank order correlation, p < 0.001). The experiment was performed on a single in vivo neurobiotin-filled neuron.

References

  1. Klausberger, T. & Somogyi, P. Neuronal diversity and temporal dynamics: the unity of hippocampal circuit operations. Science 321, 5357 (2008).
  2. Lichtman, J.W. & Denk, W. The big and the small: challenges of imaging the brain's circuits. Science 334, 618623 (2011).
  3. O'Rourke, N.A., Weiler, N.C., Micheva, K.D. & Smith, S.J. Deep molecular diversity of mammalian synapses: why it matters and how to measure it. Nat. Rev. Neurosci. 13, 365379 (2012).
  4. Hangya, B., Pi, H.J., Kvitsiani, D., Ranade, S.P. & Kepecs, A. From circuit motifs to computations: mapping the behavioral repertoire of cortical interneurons. Curr. Opin. Neurobiol. 26, 117124 (2014).
  5. Choquet, D. & Triller, A. The dynamic synapse. Neuron 80, 691703 (2013).
  6. Kano, M., Ohno-Shosaku, T., Hashimotodani, Y., Uchigashima, M. & Watanabe, M. Endocannabinoid-mediated control of synaptic transmission. Physiol. Rev. 89, 309380 (2009).
  7. Katona, I. & Freund, T.F. Multiple functions of endocannabinoid signaling in the brain. Annu. Rev. Neurosci. 35, 529558 (2012).
  8. Lovinger, D.M. Presynaptic modulation by endocannabinoids. Handb. Exp. Pharmacol. 184, 435477 (2008).
  9. Katona, I. et al. Presynaptically located CB1 cannabinoid receptors regulate GABA release from axon terminals of specific hippocampal interneurons. J. Neurosci. 19, 45444558 (1999).
  10. Wilson, R.I. & Nicoll, R.A. Endogenous cannabinoids mediate retrograde signaling at hippocampal synapses. Nature 410, 588592 (2001).
  11. Ohno-Shosaku, T., Maejima, T. & Kano, M. Endogenous cannabinoids mediate retrograde signals from depolarized postsynaptic neurons to presynaptic terminals. Neuron 29, 729738 (2001).
  12. Klausberger, T. et al. Complementary roles of cholecystokinin- and parvalbumin-expressing GABAergic neurons in hippocampal network oscillations. J. Neurosci. 25, 97829793 (2005).
  13. Tricoire, L. et al. A blueprint for the spatiotemporal origins of mouse hippocampal interneuron diversity. J. Neurosci. 31, 1094810970 (2011).
  14. Miles, R., Toth, K., Gulyas, A.I., Hajos, N. & Freund, T.F. Differences between somatic and dendritic inhibition in the hippocampus. Neuron 16, 815823 (1996).
  15. Lee, S.H., Földy, C. & Soltesz, I. Distinct endocannabinoid control of GABA release at perisomatic and dendritic synapses in the hippocampus. J. Neurosci. 30, 79938000 (2010).
  16. Hoffman, A.F., Oz, M., Yang, R., Lichtman, A.H. & Lupica, C.R. Opposing actions of chronic Δ9-tetrahydrocannabinol and cannabinoid antagonists on hippocampal long-term potentiation. Learn. Mem. 14, 6374 (2007).
  17. Puighermanal, E. et al. Cannabinoid modulation of hippocampal long-term memory is mediated by mTOR signaling. Nat. Neurosci. 12, 11521158 (2009).
  18. Puighermanal, E. et al. Dissociation of the pharmacological effects of THC by mTOR blockade. Neuropsychopharmacology 38, 13341343 (2013).
  19. Maglione, M. & Sigrist, S.J. Seeing the forest tree by tree: super-resolution light microscopy meets the neurosciences. Nat. Neurosci. 16, 790797 (2013).
  20. Rust, M.J., Bates, M. & Zhuang, X. Sub-diffraction-limit imaging by stochastic optical reconstruction microscopy (STORM). Nat. Methods 3, 793795 (2006).
  21. Vida, I., Halasy, K., Szinyei, C., Somogyi, P. & Buhl, E.H. Unitary IPSPs evoked by interneurons at the stratum radiatum-stratum lacunosum-moleculare border in the CA1 area of the rat hippocampus in vitro. J. Physiol. (Lond.) 506, 755773 (1998).
  22. MacGillavry, H.D., Song, Y., Raghavachari, S. & Blanpied, T.A. Nanoscale scaffolding domains within the postsynaptic density concentrate synaptic AMPA receptors. Neuron 78, 615622 (2013).
  23. Nair, D. et al. Super-resolution imaging reveals that AMPA receptors inside synapses are dynamically organized in nanodomains regulated by PSD95. J. Neurosci. 33, 1320413224 (2013).
  24. Specht, C.G. et al. Quantitative nanoscopy of inhibitory synapses: counting gephyrin molecules and receptor binding sites. Neuron 79, 308321 (2013).
  25. Hefft, S. & Jonas, P. Asynchronous GABA release generates long-lasting inhibition at a hippocampal interneuron-principal neuron synapse. Nat. Neurosci. 8, 13191328 (2005).
  26. Wilson, R.I., Kunos, G. & Nicoll, R.A. Presynaptic specificity of endocannabinoid signaling in the hippocampus. Neuron 31, 453462 (2001).
  27. Dani, A., Huang, B., Bergan, J., Dulac, C. & Zhuang, X. Superresolution imaging of chemical synapses in the brain. Neuron 68, 843856 (2010).
  28. Frank, T. et al. Bassoon and the synaptic ribbon organize Ca2+ channels and vesicles to add release sites and promote refilling. Neuron 68, 724738 (2010).
  29. Herlitze, S. et al. Modulation of Ca2+ channels by G-protein beta gamma subunits. Nature 380, 258262 (1996).
  30. Holderith, N. et al. Release probability of hippocampal glutamatergic terminals scales with the size of the active zone. Nat. Neurosci. 15, 988997 (2012).
  31. Wilhelm, B.G. et al. Composition of isolated synaptic boutons reveals the amounts of vesicle trafficking proteins. Science 344, 10231028 (2014).
  32. Bass, C.E. & Martin, B.R. Time course for the induction and maintenance of tolerance to Δ(9)-tetrahydrocannabinol in mice. Drug Alcohol Depend. 60, 113119 (2000).
  33. Heishman, S.J., Huestis, M.A., Henningfield, J.E. & Cone, E.J. Acute and residual effects of marijuana: profiles of plasma THC levels, physiological, subjective, and performance measures. Pharmacol. Biochem. Behav. 37, 561565 (1990).
  34. Varvel, S.A. et al. Delta9-tetrahydrocannbinol accounts for the antinociceptive, hypothermic and cataleptic effects of marijuana in mice. J. Pharmacol. Exp. Ther. 314, 329337 (2005).
  35. Carlson, G., Wang, Y. & Alger, B.E. Endocannabinoids facilitate the induction of LTP in the hippocampus. Nat. Neurosci. 5, 723724 (2002).
  36. Chevaleyre, V. & Castillo, P.E. Endocannabinoid-mediated metaplasticity in the hippocampus. Neuron 43, 871881 (2004).
  37. Bolla, K.I., Brown, K., Eldreth, D., Tate, K. & Cadet, J.L. Dose-related neurocognitive effects of marijuana use. Neurology 59, 13371343 (2002).
  38. Hirvonen, J. et al. Reversible and regionally selective downregulation of brain cannabinoid CB1 receptors in chronic daily cannabis smokers. Mol. Psychiatry 17, 642649 (2012).
  39. Stott, C.G., White, L., Wright, S., Wilbraham, D. & Guy, G.W. A phase I study to assess the single and multiple dose pharmacokinetics of THC/CBD oromucosal spray. Eur. J. Clin. Pharmacol. 69, 11351147 (2013).
  40. DeLong, G.T., Wolf, C.E., Poklis, A. & Lichtman, A.H. Pharmacological evaluation of the natural constituent of Cannabis sativa, cannabichromene and its modulation by Δ(9)-tetrahydrocannabinol. Drug Alcohol Depend. 112, 126133 (2010).
  41. Chen, K. et al. Long-term plasticity of endocannabinoid signaling induced by developmental febrile seizures. Neuron 39, 599611 (2003).
  42. Ludányi, A. et al. Downregulation of the CB1 cannabinoid receptor and related molecular elements of the endocannabinoid system in epileptic human hippocampus. J. Neurosci. 28, 29762990 (2008).
  43. Maccarrone, M. et al. Abnormal mGlu 5 receptor/endocannabinoid coupling in mice lacking FMRP and BC1 RNA. Neuropsychopharmacology 35, 15001509 (2010).
  44. Zhang, L. & Alger, B.E. Enhanced endocannabinoid signaling elevates neuronal excitability in fragile X syndrome. J. Neurosci. 30, 57245729 (2010).
  45. Jung, K.M. et al. Uncoupling of the endocannabinoid signalling complex in a mouse model of fragile X syndrome. Nat. Commun. 3, 1080 (2012).
  46. Shigemoto, R. et al. Target cell–specific concentration of a metabotropic glutamate receptor in the presynaptic active zone. Nature 381, 523525 (1996).
  47. Mikasova, L., Groc, L., Choquet, D. & Manzoni, O.J. Altered surface trafficking of presynaptic cannabinoid type 1 receptor in and out synaptic terminals parallels receptor desensitization. Proc. Natl. Acad. Sci. USA 105, 1859618601 (2008).
  48. Huestis, M.A. et al. Blockade of effects of smoked marijuana by the CB1-selective cannabinoid receptor antagonist SR141716. Arch. Gen. Psychiatry 58, 322328 (2001).
  49. Volkow, N.D., Baler, R.D., Compton, W.M. & Weiss, S.R. Adverse health effects of marijuana use. N. Engl. J. Med. 370, 22192227 (2014).
  50. Pertwee, R.G. Targeting the endocannabinoid system with cannabinoid receptor agonists: pharmacological strategies and therapeutic possibilities. Phil. Trans. R. Soc. Lond. B 367, 33533363 (2012).
  51. Tappe-Theodor, A. et al. A molecular basis of analgesic tolerance to cannabinoids. J. Neurosci. 27, 41654177 (2007).
  52. Fukudome, Y. et al. Two distinct classes of muscarinic action on hippocampal inhibitory synapses: M2-mediated direct suppression and M1/M3-mediated indirect suppression through endocannabinoid signalling. Eur. J. Neurosci. 19, 26822692 (2004).
  53. Zimmer, A., Zimmer, A.M., Hohmann, A.G., Herkenham, M. & Bonner, T.I. Increased mortality, hypoactivity, and hypoalgesia in cannabinoid CB1 receptor knockout mice. Proc. Natl. Acad. Sci. USA 96, 57805785 (1999).
  54. Nyilas, R. et al. Enzymatic machinery for endocannabinoid biosynthesis associated with calcium stores in glutamatergic axon terminals. J. Neurosci. 28, 10581063 (2008).
  55. Varga, C., Golshani, P. & Soltesz, I. Frequency-invariant temporal ordering of interneuronal discharges during hippocampal oscillations in awake mice. Proc. Natl. Acad. Sci. USA 109, E2726E2734 (2012).
  56. Dick, O. et al. The presynaptic active zone protein bassoon is essential for photoreceptor ribbon synapse formation in the retina. Neuron 37, 775786 (2003).
  57. Bates, M., Huang, B., Dempsey, G.T. & Zhuang, X. Multicolor super-resolution imaging with photo-switchable fluorescent probes. Science 317, 17491753 (2007).
  58. Baranyi, M., Cervenak, J., Bender, B. & Kacskovics, I. Transgenic rabbits that overexpress the neonatal Fc receptor (FcRn) generate higher quantities and improved qualities of anti-thymocyte globulin (ATG). PLoS ONE 8, e76839 (2013).
  59. Huang, B., Wang, W., Bates, M. & Zhuang, X. Three-dimensional super-resolution imaging by stochastic optical reconstruction microscopy. Science 319, 810813 (2008).
  60. Deschout, H. et al. Precisely and accurately localizing single emitters in fluorescence microscopy. Nat. Methods 11, 253266 (2014).
  61. Thévenaz, P., Ruttimann, U.E. & Unser, M. A pyramid approach to subpixel registration based on intensity. IEEE Trans. Image Process. 7, 2741 (1998).
  62. Ester, M., Kriegel, H.-P., Sander, J. & Xu, X. A density-based algorithm for discovering clusters in large spatial databases with noise. in Proc. 2nd Int. Conf. Knowl. Discov. Data Min. (eds. Simoudis, E., Han, J. & Fayyad, U.) 226231 (1996).
  63. Veatch, S.L. et al. Correlation functions quantify super-resolution images and estimate apparent clustering due to over-counting. PLoS ONE 7, e31457 (2012).
  64. McKinney, D.L. et al. Dose-related differences in the regional pattern of cannabinoid receptor adaptation and in vivo tolerance development to delta9-tetrahydrocannabinol. J. Pharmacol. Exp. Ther. 324, 664673 (2008).

Download references

Author information

  1. These authors equally contributed to this work.

    • Barna Dudok,
    • László Barna &
    • Marco Ledri

Affiliations

  1. Momentum Laboratory of Molecular Neurobiology, Institute of Experimental Medicine, Hungarian Academy of Sciences, Budapest, Hungary.

    • Barna Dudok,
    • László Barna,
    • Marco Ledri,
    • Szilárd I Szabó,
    • Eszter Szabadits,
    • Balázs Pintér,
    • Stephen G Woodhams,
    • Christopher M Henstridge,
    • Gyula Y Balla,
    • Rita Nyilas &
    • István Katona
  2. School of Ph.D. Studies, Semmelweis University, Budapest, Hungary.

    • Barna Dudok &
    • Gyula Y Balla
  3. Department of Anatomy and Neurobiology, University of California, Irvine, California, USA.

    • Csaba Varga,
    • Sang-Hun Lee &
    • Ivan Soltesz
  4. Alfred Renyi Institute of Mathematics, Hungarian Academy of Sciences, Budapest, Hungary.

    • Máté Matolcsi
  5. ImmunoGenes Kft, Budakeszi, Hungary.

    • Judit Cervenak &
    • Imre Kacskovics
  6. Department of Immunology, Eötvös Loránd University, Budapest, Hungary.

    • Imre Kacskovics
  7. Department of Anatomy, Hokkaido University School of Medicine, Sapporo, Japan.

    • Masahiko Watanabe
  8. Department of Biomedical Sciences, University of Cagliari, Cittadella Universitaria di Monserrato, Monserrato Cagliari, Italy.

    • Claudia Sagheddu,
    • Miriam Melis &
    • Marco Pistis
  9. CNR Neuroscience Institute, Cagliari, Italy.

    • Marco Pistis

Contributions

B.D. and L.B. designed and performed STORM experiments and corresponding analytical tools, analyzed data, prepared the figures, and wrote the manuscript. M.L. carried out patch-clamp electrophysiological recordings, cell reconstructions, STORM imaging and data analysis. S.I.S. conducted and analyzed in vitro electrophysiological measurements as well as participated in STORM imaging and data analysis. E.S., B.P., S.G.W. and R.N. developed immunostaining, tissue processing and antibody-labeling procedures for STORM imaging and contributed to data acquisition and analysis. C.M.H. performed cell culture and electron microscopy experiments and data analysis. G.Y.B., J.C., I. Kacskovics and M.W. developed and validated antibodies. C.V. performed in vivo electrophysiology. S.-H.L. carried out in vitro electrophysiological recordings. M. Matolcsi developed mathematical analysis tools. C.S., M. Melis and M.P. designed and performed the chronic drug administration experiment. I.S. designed and supervised in vitro and in vivo electrophysiological experiments and contributed to manuscript editing. I. Katona conceived and supervised the project, performed in vitro electrophysiological recordings, and wrote the manuscript.

Competing financial interests

J. Cervenak is a scientific researcher at ImmunoGenes Ltd.; I. Kacskovics is scientific co-founder of ImmunoGenes Ltd., a company specialized in the generation of FcRn transgenic animals for the production of polyclonal and monoclonal antibodies.

Corresponding author

Correspondence to:

Author details

Supplementary information

Supplementary Figures

  1. Supplementary Figure 1: Nanoscale localization precision and relative quantification of CB1 receptors by 3D-STORM imaging (231 KB)

    (a) Schematic representation of the experimental paradigm for combined confocal/STORM imaging in HEK cells. Intrinsic fluorescence intensity of EGFP linked to the extracellular N-terminus of CB1 receptors was measured by confocal microscopy. The same CB1 proteins were immunolabeled by an antibody raised against the intracellular C-terminus, and subsequently by a secondary antibody tagged with an activator-reporter dye pair Cy3-A647, which was detected by STORM microscopy. (b) Confocal image of a representative HEK cell reveals CB1 receptor accumulation within the plasma membrane. (c) At lower magnification, STORM superresolution image of the same HEK cell depicts a similar subcellular CB1 localization pattern as confocal imaging. (d) To study whether the Number of Localization Points (NLP) obtained by STORM imaging and measured in a given anatomical profile reliably predicts relative protein quantity, randomly selected HEK cells (n = 46) expressing an EGFP-CB1 fusion construct at various expression levels were imaged by both confocal and STORM microscopy. Intrinsic EGFP fluorescence intensity (FI) values were measured by confocal microscopy and determined by subtracting mean background level (assessed from 3-3 Regions of Interest (ROIs), located in the extracellular blank regions (n = 138) from every image) from the mean pixel intensity. These values are plotted against the number of localization points (NLP) obtained by STORM microscopy in each imaged cell. The strong positive linear correlation between the two imaging readouts confirms that NLP values reliably predict relative protein quantity (n = 46 cells, Spearman's rank order correlation, p < 0.001). (e) The high magnification confocal image of the boxed plasma membrane segment in b illustrates the diffraction-limited blurred nature of the EGFP fluorescence signal. (f) In contrast, at higher magnification, the nanoscale localization precision of fluorophores representing CB1 receptors by STORM imaging results in crisp visualization of the same boxed plasma membrane segment compared to the confocal image shown in e. As determined from the same images, the high resolution of the STORM image was the consequence of a fluorophore localization precision of 5 nm in the lateral (x-y), and 32 nm in the axial (z) dimensions (measured as the standard deviation of localization distribution).

  2. Supplementary Figure 2: Quantitative evaluation of the quality of STORM imaging in brain sections (48 KB)

    (a-c) Localization precision displayed only a modest variation (1 nm or 5 nm at x-y or z directions, respectively) as a function of axial position ie. at different z distances from the focal plane. Clusters resulting from the multiple blinking of isolated fluorophores in the STORM images of CB1-immunostaining were identified, and the distribution of the distances of individual localization points from the center of mass of the clusters was plotted. (d) Histogram shows the distribution of the number of detected photons per localization point (n = 1341 LP) above the peak detection threshold in STORM images of 40000 cycles taken from CB1-immunostaining in hippocampal sections at 5 μm depth.

  3. Supplementary Figure 3: Comparative analysis of the passive and active electrophysiological properties of regular-spiking CB1-positive perisomatic and dendritic interneurons in the CA1 subfield of the mouse hippocampus (107 KB)

    (a-e) Whole-cell patch-clamp recordings from multipolar interneurons which were post-hoc classified as either perisomatic or dendritic interneurons based on their axonal distribution and their corresponding BDI value. Electrophysiological characterization reveals that these GABAergic interneuron types do not differ in their resting membrane potential (p = 0.19, t(30) = −1.34 summarized in a), input resistance (p = 0.17, t(30) = −1.4) (b), relative sag amplitude (p = 0.24, U = 94.5) (c), rebound potential (p = 0.25, U = 89) (d), and action potential amplitude (p = 0.15, t(30) = −1.47) (e). (f-h) However, dendritic interneurons fire faster action potentials (p = 0.0024, U = 48.5) (f), show a more marked afterhyperpolarization (p = 0.037, t(30) = −2.18) (g), and exhibit more adaptation of firing upon current injection (p = 0.013, t(30) = −2.63)) (h). Statistical comparisons of 14 perisomatic and 18 dendritic cells were performed by unpaired two-sided t-test (a,b,e,g,h) or by Mann-Whitney U test (c,d,f). Graphs show raw data and median±IQR. The measured data for each cell and for each parameter are presented in Supplementary Table.

  4. Supplementary Figure 4: The quantitative power of CB1-immunostaining and subsequent STORM imaging is robust to various experimental conditions (160 KB)

    (a) Electrophysiological recordings from acute hippocampal slices, tissue processing and immunostaining were carried out in parallel in two different laboratories in Irvine, CA, USA and in Budapest, Hungary. Undertaking all STORM imaging on the same setup in the latter laboratory resulted in similar normalized distribution of the number of CB1 localization points per boutons, as demonstrated in the cumulative probability plot showing the NLP values per bouton, normalized for the cell mean. Statistical analysis was done by Kolmogorov-Smirnov test, p > 0.05, n = 150 and n = 279 boutons from n = 7 and n = 6 basket cells filled and processed in Irvine and Budapest, respectively, were used in the experiment. (b) Immunostaining with two independent primary antibodies raised either in a guinea-pig52, or a transgenic (Tg) rabbit that overexpress the neonatal Fc receptor (FcRn) and have improved humoral immune responses58, and with the two corresponding secondary antibodies custom-labeled with fluorophore pairs for STORM imaging, resulted in similar normalized distribution of the number of CB1 localization points per bouton (n = 279 and n = 302 boutons from n = 6-6 basket cells stained with anti-CB1 antibody produced in guinea-pig and rabbit, respectively, Kolmogorov-Smirnov test, p > 0.1). (c-d) Whole-cell patch-clamp recordings and biocytin-filling did not affect bouton size or CB1 NLP. Biocytin-filled boutons, and an equal number of neighboring biocytin-negative, randomly chosen boutons were selected from the same images for analysis. The size of the boutons was determined by using the CB1 STORM signal outlining the terminals. No significant differences were found in the number of CB1 localization points, or in the size of boutons with or without biocytin-labeling (n = 202-202 axon terminals, Mann-Whitney U test, p = 0.20, U = 4563.5 and p = 0.31, U = 29232, respectively). Graphs show raw data and median±IQR.

  5. Supplementary Figure 5: Comparative morphological analysis of the axon terminals of CB1-positive perisomatic and dendritic interneurons (48 KB)

    (a-c) Deconvolved confocal images of n = 279 and n = 334 biocytin-filled boutons from n = 6 perisomatic and n = 6 dendritic cells, respectively, were analyzed for their shape descriptors. (a) Circularity describes how close the relation between the area and perimeter of the bouton shape is to that of a perfect circle. (b) Aspect ratio reflects the degree of elongation. (c) Solidity measures the area of the shape divided by its convex area, thereby describing convexity and smoothness of edges of the bouton shape. No significant differences are detected between cell types in either the circularity (p = 0.82, U = 16), aspect ratio (p = 0.70, U = 15), or solidity (p = 0.94, U = 17) of their boutons (Mann-Whitney U test). Graphs show raw data and median±IQR.

  6. Supplementary Figure 6: CB1 density distribution on axon terminals of identified interneurons can be reliably determined under the specific imaging conditions used throughout the study (112 KB)

    (a) To assess whether the local density of STORM labeling on axon terminals is within the dynamic range of the measurement, 10-10 ROIs (30 by 30 confocal pixels each; pixel size = 0.0064 μm2) were placed on the plasma membrane of HEK cells and biocytin-filled axon terminals to measure labeling density. Comparison of the labeling density did not reveal significant difference between boutons of perisomatic and dendritic interneurons (n = 6 cells/cell type, Mann-Whitney U test, p = 0.59, U = 14), and the vast majority of boutons fell within the range defined by the two HEK cells with the highest and lowest expression levels. Graphs show raw data and median±IQR. (b) The NLP values of boutons were averaged for each interneuron, and the mean±SD range of NLP density for interneurons was determined. This range well represents the STORM CB1 labeling density found on GABAergic axon terminals, as it includes 10 of the 12 interneurons. HEK cells (n = 20) within this range show highly linear positive correlation between intrinsic fluorescence intensity and NLP (Spearman's rank order correlation, p < 0.0001). (c) The distribution of local densities (neighboring localizations within 50 nm 3D distance from each localization point) on the boutons from (a) shows a heavy right tail indicating that the signal is not likely to be saturated. (d) The different size of perisomatic and dendritic boutons is evident regardless of the duration of the imaging. Using the same data set as in Figure 3, the perimeter of a 2D-convex hull fitted on the localization points detected before the given period of imaging was measured, and was plotted as the ratio of the perimeter of dendritic axon terminals over perisomatic boutons (median±IQR of the mean values of 6 cells per cell type is shown). The difference was significant at every time point after 250 cycles (n = 6 cells/cell type, Mann-Whitney U test, p = 0.015, U = 3). (e) Similarly, the finding that axon terminals of dendritic cells had less CB1 receptors than perisomatic boutons was also independent of the imaging duration after 350 cycles, as the difference was significant at every time point afterwards (n = 6 cells/cell type, Mann-Whitney U test, p = 0.041, U = 5). The same approach was used to compare the number of CB1 localization points per bouton between interneuron types. (f) The variance of nearest-neighbor distances between CB1 localization points within boutons was constant in both cell types after 400 cycles into imaging, indicating that running the imaging for up to 1000 cycles for each image provides reliable measurement of CB1 distribution within GABAergic axon terminals.

  7. Supplementary Figure 7: Layer-specific electron microscopic analysis of CB1-containing GABAergic axon terminals (127 KB)

    (a,b) To test whether quantitative information collected by 3D-STORM microscopy is comparable with data acquired by other well-established quantitative anatomical approaches, hippocampal sections were processed for immunogold labeling of CB1 receptors and analyzed by transmission electron microscopy (TEM). Random cross-sections of CB1-positive GABAergic boutons were selected from the stratum pyramidale and stratum radiatum of the CA1 subfield. The presumed GABAergic nature of an axon terminal was established by the presence of symmetrical synapses (arrowheads). Note that the vast majority of silver-intensified gold particles representing CB1 receptors are attached to the intracellular surface of the plasma membrane of axon terminals. (c,d) CB1-positive boutons found in the pyramidal layer are significantly larger (n = 3 animals, 100 boutons per layer per animal, paired two-sided t-test, p = 0.025, t(2) = 6.26 and p = 0.033, t(2) = 5.33 for area and perimeter, respectively). (g) In addition, the number of gold particles (NGP) on the plasma membrane of these perisomatic axon terminals is considerably higher than on boutons observed in the stratum radiatum (paired two-sided t-test, p = 0.036, t(2) = 5.13). (h) In contrast, the density of gold particles was not significantly different on boutons located in the two hippocampal layers (paired two-sided t-test, p = 0.064, t(2) = 3.76). These observations obtained at the synaptic population level are in general agreement with the cell-type-specific findings obtained by STORM imaging (see Figure 4). Graphs show raw data and median±IQR.

  8. Supplementary Figure 8: Visualization of the localization coordinates of the STORM signal by 3D rendering and calculation of membrane surface distances (263 KB)

    (a) Analysis of nanoscale distribution of STORM data was carried out using the x, y and z coordinates of each individual localization point. Their 3D positions are illustrated as atoms in Visual Molecular Dynamics. (b) First, a convex hull was fitted on the localization points, giving an approximation of the structure surface. Note that single-molecule localization-microscopy gives a more precise estimate of structural dimensions than electron microscopy22, which underestimates these parameters due to anisotropic shrinkage during tissue dehydration, as well as compared to confocal microscopy, which overestimates size due to light diffraction25. Indeed, the average bouton size based on the cross-sectional area of the polyhedrons of perisomatic and dendritic terminals are 1.33±0.04 μm2 and 0.97±0.03 μm2, respectively (mean±SEM), when measured by 3D-STORM microscopy. These values fall in between the range of electron microscopic (0.29±0.01 μm2 and 0.22±0.01 μm2); and confocal measurements (1.65±0.04 μm2 and 1.20±0.03 μm2). These data independently indicate that the convex hull approach resulted in a realistic representation of the imaged part of the bouton. This feature has previously been an exclusive advantage of the laborious serial reconstruction of electron micrographs. (c) Secondly, the distance of each localization point from the convex hull was measured, and only localizations within 80 nm from the surface were included in further analysis. This approach could also be utilized to measure the proportion of membrane-attached localization points and those which are likely intracellular receptors. These subcellularly segregated receptor populations presumably represent functionally different receptor pools, hence it is important to distinguish between them, when quantifying receptor distribution or pathological redistribution in a cell-type-specific manner. (d) Finally, the distance of each localization point from the active zone was measured as the shortest route on the polyhedron surface.

  9. Supplementary Figure 9: Validation of the hull surface distance calculations and density measurements on model boutons (340 KB)

    (a) To measure the sensitivity of the local density analysis on the polyhedron surface, a simulation of homogeneous and artificially enriched receptor distribution was conducted. “CB1 localization points” were randomly placed on model “bouton” comprising a sphere of 1000 nm radius, excluding a 230 nm radius area of the surface as the “active zone”. A convex hull was fitted onto these points. (b-c) To model different levels of potential perisynaptic enrichment, an increasing number of localization points were placed in random positions of the surface within a 250 nm radius of the active zone. (d) To measure localization point density on the surface, reference points were also placed at 40 nm spacing over the entire sphere, excluding the active zone. CB1 density, defined as the number of localization points within 200 nm distance of each reference point, was determined along with the distance from the edge of the active zone measured on the hull surface. The entire process was repeated 50 times for each enrichment condition. Finally, reference points per individual boutons were binned based on the distance from the synapse and averaged by bin. Altogether 50 mean values per bin per condition were used for generating graphs (median±IQR) and calculating p values using unpaired two-sided t-tests to compare perisynaptic and extrasynaptic regions. The difference between the 0–250 nm and the 1000-1250 nm bins is already significant at 20% (p < 0.0001), but not at 10% enrichment (p = 0.84). The graph for 1000% is not shown for sake of clarity, but was also highly significant (p < 0.0001).

  10. Supplementary Figure 10: Combination of in vivo single neuron recording in a behaving mouse with post-hoc anatomical characterization and molecular imaging at the nanoscale (263 KB)

    (a) Example traces from juxtacellular recording of a CA1 interneuron in an anesthesia-free mouse running on a spherical treadmill. A 1 sec-long recording, filtered for local field potential, and, spikes, and the raw trace are shown below each other, respectively. (b) After the recording, the cell was filled with neurobiotin. Morphological analysis of the labeled neuron was performed by confocal microscopy. Part of the axonal arbor (red), as well as its soma and some dendrites (blue), were reconstructed from four consecutive 20 μm thick sections by Neurolucida. (c) Laminar distribution analysis of more than 3000 axonal varicosities recovered from 14 serial sections classifies this cell as a dendritic interneuron (BDI < 0.5). (d) Maximum intensity projection of a deconvolved confocal image stack of individual axon terminals from the same in vivo recorded interneuron. (e) CB1-immunolabeling on these identified axon terminals is visualized by overlaying the corresponding confocal and STORM images. (f) Magnification of the boxed region of the STORM image in e illustrates the intraterminal distribution of CB1 localization points at the nanoscale level and reveals a qualitatively similar pattern as that observed in the case of in vitro-labeled interneurons in Figure 3. (g) Cumulative probability plots of normalized CB1 NLP found on individual terminals of in vivo- or in vitro-filled dendritic interneurons. The distribution of data obtained from the two experimental preparations were identical (Kolmogorov-Smirnov test, p > 0.1). Data were collected from single CB1 STORM imaging of 129 boutons from one in vivo, and of 141 boutons from 5 in vitro-filled cells and were normalized to the mean NLP value of the respective cell. (h) The variance of CB1 nearest neighbor distances is also comparable under the different experimental conditions indicating similar CB1 distribution at the nanoscale level in boutons of in vivo- or in vitro-filled dendritic interneurons (Kolmogorov-Smirnov test, p > 0.1). (i) In line with the in vitro findings, at the single bouton level, the number of CB1 localization points correlates linearly with bouton size (n = 129 axon terminals, Spearman's rank order correlation, p < 0.001). The experiment was performed on a single in vivo neurobiotin-filled neuron.

PDF files

  1. Supplementary Text and Figures (4,034 KB)

    Supplementary Figures 1–10

  2. Supplementary Methods Checklist (504 KB)

Excel files

  1. Supplementary Table 1: Physiological and anatomical parameters of CB1-positive regular-spiking, multipolar interneurons in the CA1 subfield of the mouse hippocampus. (19 KB)

    AP, action potential; FWHM, full width at half maximum; AHP, after hyper polarization; BDI, bouton distribution index.

Additional data