Experience results in long-lasting changes in dendritic spine size, yet how the molecular architecture of the synapse responds to plasticity remains poorly understood. Here a combined approach of multicolor stimulated emission depletion microscopy (STED) and confocal imaging in rat and mouse demonstrates that structural plasticity is linked to the addition of unitary synaptic nanomodules to spines. Spine synapses in vivo and in vitro contain discrete and aligned subdiffraction modules of pre- and postsynaptic proteins whose number scales linearly with spine size. Live-cell time-lapse super-resolution imaging reveals that NMDA receptor–dependent increases in spine size are accompanied both by enhanced mobility of pre- and postsynaptic modules that remain aligned with each other and by a coordinated increase in the number of nanomodules. These findings suggest a simplified model for experience-dependent structural plasticity relying on an unexpectedly modular nanomolecular architecture of synaptic proteins.
This is a preview of subscription content, access via your institution
Access Nature and 54 other Nature Portfolio journals
Get Nature+, our best-value online-access subscription
$29.99 / 30 days
cancel any time
Subscribe to this journal
Receive 12 print issues and online access
$209.00 per year
only $17.42 per issue
Rent or buy this article
Prices vary by article type
Prices may be subject to local taxes which are calculated during checkout
Similar content being viewed by others
Feldman, D. E. Synaptic mechanisms for plasticity in neocortex. Annu. Rev. Neurosci. 32, 33–55 (2009).
Trachtenberg, J. T. et al. Long-term in vivo imaging of experience-dependent synaptic plasticity in adult cortex. Nature 420, 788–794 (2002).
Matsuzaki, M., Honkura, N., Ellis-Davies, G. C. R. & Kasai, H. Structural basis of long-term potentiation in single dendritic spines. Nature 429, 761–766 (2004).
Sheng, M. & Hoogenraad, C. C. The postsynaptic architecture of excitatory synapses: a more quantitative view. Annu. Rev. Biochem. 76, 823–847 (2007).
Matsuzaki, M. et al. Dendritic spine geometry is critical for AMPA receptor expression in hippocampal CA1 pyramidal neurons. Nat. Neurosci. 4, 1086–1092 (2001).
Bosch, M. et al. Structural and molecular remodeling of dendritic spine substructures during long-term potentiation. Neuron 82, 444–459 (2014).
Nicoll, R. A. A brief history of long-term potentiation. Neuron 93, 281–290 (2017).
Ehrlich, I. & Malinow, R. Postsynaptic density 95 controls AMPA receptor incorporation during long-term potentiation and experience-driven synaptic plasticity. J. Neurosci. 24, 916–927 (2004).
Migaud, M. et al. Enhanced long-term potentiation and impaired learning in mice with mutant postsynaptic density-95 protein. Nature 396, 433–439 (1998).
Stein, V., House, D. R., Bredt, D. S. & Nicoll, R. A. Postsynaptic density-95 mimics and occludes hippocampal long-term potentiation and enhances long-term depression. J. Neurosci. 23, 5503–5506 (2003).
Bats, C., Groc, L. & Choquet, D. The interaction between Stargazin and PSD-95 regulates AMPA receptor surface trafficking. Neuron 53, 719–734 (2007).
Hafner, A.-S. S. et al. Lengthening of the stargazin cytoplasmic tail increases synaptic transmission by promoting interaction to deeper domains of PSD-95. Neuron 86, 475–489 (2015).
Schnell, E. et al. Direct interactions between PSD-95 and stargazin control synaptic AMPA receptor number. Proc. Natl. Acad. Sci. USA 99, 13902–13907 (2002).
Zeng, M. et al. Phase transition in postsynaptic densities underlies formation of synaptic complexes and synaptic plasticity. Cell 166, 1163–1175.e12 (2016).
Blanpied, T. A., Kerr, J. M. & Ehlers, M. D. Structural plasticity with preserved topology in the postsynaptic protein network. Proc. Natl. Acad. Sci. USA 105, 12587–12592 (2008).
Tang, A.-H. H. et al. A trans-synaptic nanocolumn aligns neurotransmitter release to receptors. Nature 536, 210–214 (2016).
Nair, D. et al. Super-resolution imaging reveals that AMPA receptors inside synapses are dynamically organized in nanodomains regulated by PSD95. J. Neurosci. 33, 13204–13224 (2013).
Harris, K. M. & Stevens, J. K. Dendritic spines of CA 1 pyramidal cells in the rat hippocampus: serial electron microscopy with reference to their biophysical characteristics. J. Neurosci. 9, 2982–2997 (1989).
Schikorski, T. & Stevens, C. F. Quantitative ultrastructural analysis of hippocampal excitatory synapses. J. Neurosci. 17, 5858–5867 (1997).
Meyer, D., Bonhoeffer, T. & Scheuss, V. Balance and stability of synaptic structures during synaptic plasticity. Neuron 82, 430–443 (2014).
Schoonderwoert, V., Dijkstra, R. & Luckinavicius, G. Huygens STED deconvolution increases signal-to-noise and image resolution towards 22 nm. Microscopy 21, 38–44 (2013).
Toni, N., Buchs, P. A., Nikonenko, I., Bron, C. R. & Muller, D. LTP promotes formation of multiple spine synapses between a single axon terminal and a dendrite. Nature 402, 421–425 (1999).
Susaki, E. A. et al. Whole-brain imaging with single-cell resolution using chemical cocktails and computational analysis. Cell 157, 726–739 (2014).
Tainaka, K. et al. Whole-body imaging with single-cell resolution by tissue decolorization. Cell 159, 911–924 (2014).
Holtmaat, A. & Svoboda, K. Experience-dependent structural synaptic plasticity in the mammalian brain. Nat. Rev. Neurosci. 10, 647–658 (2009).
Lu, W. et al. Activation of synaptic NMDA receptors induces membrane insertion of new AMPA receptors and LTP in cultured hippocampal neurons. Neuron 29, 243–254 (2001).
Nägerl, U. V., Willig, K. I., Hein, B., Hell, S. W. & Bonhoeffer, T. Live-cell imaging of dendritic spines by STED microscopy. Proc. Natl. Acad. Sci. USA 105, 18982–18987 (2008).
Park, M., Penick, E. C., Edwards, J. G., Kauer, J. A. & Ehlers, M. D. Recycling endosomes supply AMPA receptors for LTP. Science 305, 1972–1975 (2004).
Sun, Q. & Turrigiano, G. G. PSD-95 and PSD-93 play critical but distinct roles in synaptic scaling up and down. J. Neurosci. 31, 6800–6808 (2011).
Turrigiano, G. G. & Nelson, S. B. Homeostatic plasticity in the developing nervous system. Nat. Rev. Neurosci. 5, 97–107 (2004).
Li, Z. & Murthy, V. N. Visualizing postendocytic traffic of synaptic vesicles at hippocampal synapses. Neuron 31, 593–605 (2001).
McClelland, A. C., Sheffler-Collins, S. I., Kayser, M. S. & Dalva, M. B. Ephrin-B1 and ephrin-B2 mediate EphB-dependent presynaptic development via syntenin-1. Proc. Natl. Acad. Sci. USA 106, 20487–20492 (2009).
Nakata, T., Terada, S. & Hirokawa, N. Visualization of the dynamics of synaptic vesicle and plasma membrane proteins in living axons. J. Cell Biol. 140, 659–674 (1998).
Gross, G. G. et al. Recombinant probes for visualizing endogenous synaptic proteins in living neurons. Neuron 78, 971–985 (2013).
Südhof, T. C. The presynaptic active zone. Neuron 75, 11–25 (2012).
Steiner, P. et al. Destabilization of the postsynaptic density by PSD-95 serine 73 phosphorylation inhibits spine growth and synaptic plasticity. Neuron 60, 788–802 (2008).
Hruska, M., Henderson, N. T., Xia, N. L., Le Marchand, S. J. & Dalva, M. B. Anchoring and synaptic stability of PSD-95 is driven by ephrin-B3. Nat. Neurosci. 18, 1594–1605 (2015).
Petersen, C. C. H., Malenka, R. C., Nicoll, R. A. & Hopfield, J. J. All-or-none potentiation at CA3-CA1 synapses. Proc. Natl. Acad. Sci. USA 95, 4732–4737 (1998).
Liu, K. K., Hagan, M. F. & Lisman, J. E. Gradation (approx. 10 size states) of synaptic strength by quantal addition of structural modules. Phil. Trans. R. Soc. Lond. B 372, 20160328 (2017).
Harris, K. M., Jensen, F. E. & Tsao, B. Three-dimensional structure of dendritic spines and synapses in rat hippocampus (CA1) at postnatal day 15 and adult ages: implications for the maturation of synaptic physiology and long-term potentiation. J. Neurosci. 12, 2685–2705 (1992).
Sorra, K. E., Fiala, J. C. & Harris, K. M. Critical assessment of the involvement of perforations, spinules, and spine branching in hippocampal synapse formation. J. Comp. Neurol. 398, 225–240 (1998).
Broadhead, M. J. et al. PSD95 nanoclusters are postsynaptic building blocks in hippocampus circuits. Sci. Rep. 6, 24626 (2016).
Alvarez, V. A. & Sabatini, B. L. Anatomical and physiological plasticity of dendritic spines. Annu. Rev. Neurosci. 30, 79–97 (2007).
Yang, Y. et al. Selective synaptic remodeling of amygdalocortical connections associated with fear memory. Nat. Neurosci. 19, 1348–1355 (2016).
Sinnen, B. L. et al. Optogenetic control of synaptic composition and function. Neuron 93, 646–660.e5 (2017).
Zhang, Y., Cudmore, R. H., Lin, D.-T. T., Linden, D. J. & Huganir, R. L. Visualization of NMDA receptor-dependent AMPA receptor synaptic plasticity in vivo. Nat. Neurosci. 18, 402–407 (2015).
Huganir, R. L. & Nicoll, R. A. AMPARs and synaptic plasticity: the last 25 years. Neuron 80, 704–717 (2013).
Gray, N. W., Weimer, R. M., Bureau, I. & Svoboda, K. Rapid redistribution of synaptic PSD-95 in the neocortex in vivo. PLoS Biol. 4, e370 (2006).
Kim, I. H. et al. Spine pruning drives antipsychotic-sensitive locomotion via circuit control of striatal dopamine. Nat. Neurosci. 18, 883–891 (2015).
Spence, E. F., Kanak, D. J., Carlson, B. R. & Soderling, S. H. The Arp2/3 complex is essential for distinct stages of spine synapse maturation, including synapse unsilencing. J. Neurosci. 36, 9696–9709 (2016).
Kayser, M. S., McClelland, A. C., Hughes, E. G. & Dalva, M. B. Intracellular and trans-synaptic regulation of glutamatergic synaptogenesis by EphB receptors. J. Neurosci. 26, 12152–12164 (2006).
McClelland, A. C., Hruska, M., Coenen, A. J., Henkemeyer, M. & Dalva, M. B. Trans-synaptic EphB2-ephrin-B3 interaction regulates excitatory synapse density by inhibition of postsynaptic MAPK signaling. Proc. Natl. Acad. Sci. USA 107, 8830–8835 (2010).
Lois, C., Hong, E. J., Pease, S., Brown, E. J. & Baltimore, D. Germline transmission and tissue-specific expression of transgenes delivered by lentiviral vectors. Science 295, 868–872 (2002).
Taniguchi, H. et al. Silencing of neuroligin function by postsynaptic neurexins. J. Neurosci. 27, 2815–2824 (2007).
Nolt, M. J. et al. EphB controls NMDA receptor function and synaptic targeting in a subunit-specific manner. J. Neurosci. 31, 5353–5364 (2011).
Gehne, N. et al. Cross-over endocytosis of claudins is mediated by interactions via their extracellular loops. PLoS One 12, e0182106 (2017).
Rankin, B. R. et al. Nanoscopy in a living multicellular organism expressing GFP. Biophys. J. 100, L63–L65 (2011).
Richardson, D. S. & Lichtman, J. W. Clarifying tissue clearing. Cell 162, 246–257 (2015).
We thank R. Hodge and the other members of the Dalva laboratory and S. Datta and J. Kauer for suggestions and comments. Grants from NIDA (DA022727), NIMH (MH100093) and the Vickie and Jack Farber Foundation to M.B.D. supported this work.
The authors declare no competing interests.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Integrated supplementary information
Neurons transfected with tdTomato at DIV0 were fixed and stained for endogenous PSD-95 at DIV21-25. The indicated secondary antibodies were used to simultaneously visualize PSD-95 using conventional confocal (FWHM ~250–300 nm) and a) STED using a 592 nm CW depletion beam (FWHM ~80 nm) or b) gated STED using a 775 nm pulsed depletion beam (FWHM ~50 nm). a) Staining of PSD-95 simultaneously with Atto-425 and AlexaFluor-488. Images were acquired by exciting samples with the 442 nm and 488 nm lines respectively while using a 592 nm CW depletion beam on a Leica SP5 instrument. Arrow indicates PSD-95 labeling in spines (dotted outline). Compared to confocal, STED resulted in an improvement in resolution in both channels by enabling the visualization of two discrete clusters in the indicated spine. b) Staining of PSD-95 simultaneously with AlexaFluor-594 and Atto-647N. Images were acquired by exciting samples with the 594 nm and 647 nm lines while using a 775 nm pulsed depletion beam on a Leica SP8 instrument. Arrow indicates PSD-95 labeling in the indicated spine (dotted outline). Improvement in resolution compared to confocal was observed in both STED channels as visualized by the appearance of multiple distinct PSD-95 clusters in the spine after activation of the appropriate depletion lasers (arrows). These data indicate that the AlexaFluor and Atto dyes used in this study are compatible with STED nanoscopy. Similar results were obtained from three independent experiments. Scale bars in a and b, 1 µm.
Raw confocal and STED images from the control condition from which the spine in Fig. 5a was chosen. Dendritic arbors were labeled by transfecting neurons with tdTomato at DIV0. Images are of transfected neurons stained at DIV21-25 with primary antibodies to mRFP, PSD-95 and vGlut1 as indicated. a) Images for tdTomato were acquired in conventional confocal mode using a Cy3 conjugated secondary antibody and a 561 nm excitation laser line (AOTF = 7%). Imaging was performed using an SP5 Leica instrument. Anti-PSD-95 was visualized by using an Atto-425-conjugated secondary antibody that was excited by a 442 nm line (AOTF = 35%). Raw STED PSD-95 images were obtained by depleting with a 592 nm line (AOTF = 90%). Anti-vGlut1 was visualized by labeling with an AlexaFluor-488-conjugated secondary antibody excited with a 488 nm line (AOTF = 12%). Raw STED images were obtained by depleting with a 592 nm line (AOTF = 100%). Clear PSD-95 (green) and vGlut1 (red) puncta showing a high degree of apposition (merge) were obtained using this method. b) Raw STED images from panel a were subjected to background subtraction (mean intensity of all pixels in the image) followed by a Gaussian blur (2 pixel size). c) Deconvolution of raw STED images in panel a in SP5 Leica Application Suite Advance Fluorescence software using 80 nm microscope resolution. Scale bar for a, b and c, 2µm. d) Line profiles (dotted line) through the puncta in a control spine (also shown in Fig. 5a) demonstrated sub-diffraction size PSD-95 (FWHM = 210 nm) and vGlut1 puncta (FWHM = 190 nm) in unprocessed STED images. e) Gaussian blur/background subtraction performed on raw images resulted in visibly brighter puncta with FWHM = 220 nm (PSD-95) and FWHM = 200 nm (vGlut1). f) Deconvolution resulted in puncta fluorescence intensity that was comparable to the Gaussian blur/background subtraction method of analysis used in panels b and e. Deconvolution improved resolution of both PSD-95 puncta (FWHM = 190 nm) and vGlut1 puncta (FWHM = 150 nm), but it did not result in artifactual PSD-95 or vGlut1 clusters. Scale bar for d, e and f, 1µm. Similar results were obtained from three independent experiments (see also Fig. 5).
Neurons were transfected with EGFP at DIV 0 and stained for endogenous PSD-95 at DIV 21-25. EGFP labeling was enhanced by staining with a GFP antibody (Online Methods) and visualized by an AlexaFluor-488-conjugated secondary antibody. Endogenous PSD-95 was simultaneously visualized in two channels by co-staining with primary antibodies to PSD-95 with AlexaFluor-594 and Atto-647N secondary antibodies (as in Fig. S1). After acquiring a confocal image using a Leica SP8 3x gated STED system a series of ten STED images were obtained using a pulsed 775 nm depletion beam on a single field of view to induce signal weakening due to bleaching. Gated detectors were used for both AlexaFluor 594 (0.3 to 6 nsec) and Atto-647N (0.2 to 6 nsec). Images were then deconvolved offline using Huygens deconvolution software (Online Methods). This software conducts deconvolution based on the laser intensities used, microscope PSF, background and signal-to-noise ratio in each channel independently. A maximum of 40 iterations was used. The inset demonstrates two spines with high PSD-95 signal that appear to contain a single PSD-95 cluster in both channels in confocal mode. Appearance of two discrete PSD-95 clusters is observed in raw STED images in both Atto-647N and AlexaFluor-594 channels in one spine (double arrow), while the neighboring spine has only one cluster (single arrow). Deconvolution further improved the resolution, resulting in two clearly separated clusters in a double clustered spine, but did not induce cluster artifacts (single cluster spine is still single clustered after deconvolution). The same clusters could be visualized using deconvolution even after bleaching reduced the signal-to-noise ratio (after the tenth STED image), indicating that deconvolution is unlikely to generate puncta resembling PSD-95 even in poor signal-to-noise conditions. Similar results were obtained from three independent experiments. Scale bars, 1 µm.
Supplementary Figure 4 Modular organization of dendritic spine synapses in vitro—analysis of non-deconvolved images.
(a) The images from Fig. 1a are shown without deconvolution. PSD-95 (green) and vGlut1 (red) modules were imaged in STED (~80–100 nm FWHM) and tdTomato-labeled dendritic spines were imaged simultaneously in confocal mode (~300 nm FWHM, gray and dashed yellow outlines). Scale bar, 0.8 µm. Schematic (left panels) demonstrating the arrangement of multiple synaptic profiles found in individual spines for the images shown on the right. Similar results were obtained from three independent experiments (See also Fig. 1). (b) Positive correlation of the non-deconvolved PSD-95 clusters with spine size (green line, R2 = 0.4582, slope = 1.940 ± 0.1758, p < 0.0001, F-test, n = 146 spines). Dotted black line represents the deconvolved data from Fig. 1i (deconvolved vs. non-deconvolved, p = 0.9615, ANCOVA). (c) Positive correlation of the non-deconvolved vGlut1 clusters with spine size (red line, R2 = 0.2534, slope = 1.373 ± 0.1964, p < 0.0001, F-test, n = 146 spines). Dotted black line represents the deconvolved data from Fig. 1j (deconvolved vs. non-deconvolved, p = 0.8424, ANCOVA). (d and g) Quantification of the percentage of spines containing single and multiple (d) PSD-95 and (g) vGlut1 clusters (n = 146 spines). (e and h) Quantification of the average areas of individual (e) PSD-95 (233 clusters, n = 85 (single), n = 78 (double), n = 54 (triple), n = 16 (quadruple), p = 0.1533, F(3, 230) = 1.703, one-way ANOVA) and (h) vGlut1 (225 clusters, n = 84 (single), n = 94 (double), n = 39 (triple), n = 8 (quadruple), p = 0.2761, F (3, 221) = 1.031, one-way ANOVA). (f and i) Quantification of the total area of (f) PSD-95 (n = 85 (single), n = 39 (double), n = 18 (triple), n = 4 (quadruple) spines, ***p < 0.0001 all comparisons, F (3, 142) = 55.82), and (i) vGlut1 (n = 84 (single), n = 47 (double), n = 13 (triple), n = 2 (quadruple) spines, 1 vs. 2, 1 vs. 3, ***p < 0.0001; 2 vs. 3, *p = 0.0192, F (3, 142) = 25.22, one-way ANOVA with Fisher’s LSD post hoc). Graphs (d-i) represent mean +/− SEM.
Supplementary Figure 5 Multiple aligned pre-and postsynaptic modules represent discrete synaptic profiles within individual spines.
(a) A Leica TCS SP8 gated STED 3X instrument which allows for super-resolved imaging in three dimensions (~50 nm XY and ~200–250 nm in Z) was used to acquire dual-color 3D super-resolution images of endogenous PSD-95 (green) and vGlut1 (red) staining and confocal resolution images of tdTomato-filled dendritic spines (gray, white dashed outline) of DIV 21 cultured cortical neurons. Left and middle panels show orthogonal views (yellow dashed lines) of three distinct PSD-95 and vGlut1 nanomodules. Yellow, cyan and magenta arrows point to three discrete modules within the example spine. Scale bar, 1 µm. The right panel shows a 3D reconstruction of the same spine with super-resolved PSD-95 and vGlut1 nanomodules superimposed (Methods). Yellow, cyan and magenta arrows point to the 3D-rendered modules from the 2D image stacks in the left panel. Scale bar, 500 nm. Similar results were observed in at least three independent experiments. (b) Dual-color 3D super-resolved images of endogenous PSD-95 (green) and the pre-synaptic scaffolding protein Bassoon (red) staining in a confocal resolved tdTomato-filled spine (gray, white dashed outline). Orthogonal views (yellow dashed line) show two distinct, aligned PSD-95 and Bassoon modules (cyan and magenta arrows, middle panel) in the same spine. Scale bar, 1 µm. The right panel shows the 3D reconstructed image from the 2D image stacks in the left panel. Two discrete, 3D reconstructed PSD-95 and Bassoon modules are shown by cyan and magenta arrows, respectively. Scale bar, 500 nm. Similar results were observed in three independent experiments. (c-f) Positive correlation of the total synaptic area (measured as the sum of areas of all nanomodules in a spine) with the area of the spine head determined using Pearson’s correlation. All four synaptic markers (c) PSD-95 (R2 = 0.3208, p < 0.0001, F-test, n = 406 spines, CW non-gated STED at ~80 nm resolution), (d) vGlut1 (R2 = 0.2108, p < 0.0001, F-test, n = 406 spines, gated STED (AlexaFluor-594, 0.3 to 6 nsec) at ~50 nm resolution), (e) Bassoon (R2 = 0.1955, p < 0.0001, F-test, n = 217 spines, gated STED (Atto-647N, 0.2 to 6 nsec) at ~50 nm resolution) and (f) SYP-1 (R2 = 0.2860, p < 0.0001, F-test, n = 189 spines, gated STED (Atto-647N, 0.2 to 6 nsec) at ~50 nm resolution) are significantly correlated with the spine size.
Supplementary Figure 6 PSD-95 and vGlut1 clusters in vitro show poor colocalization with randomly distributed spines.
(a) Super-resolved 2D images of PSD-95 (green) and vGlut1 (red) nanomodules as in Fig. 1. Scale bar, 5 µm. (b) Binary PSD-95 (top panels) and vGlut1 (bottom panels) masks generated by thresholding the corresponding images shown in panel (a) as described in Online Methods. Magenta circles indicate 500 randomly positioned simulated spines of small (left panels) and large (right panels) variety that were created using the random number generator in ImageJ (NIH, Bethesda, MD). Simulated spine sizes reflect spine sizes from our data in Fig. 1 at 12.5, 25, 37.5, 50, 62.5, 75, 87.5 and 99 percentiles. The simulations were repeated using images from three independent transfection experiments. (c) Schematic of the inclusion and exclusion criteria in simulations were set to reflect criteria that were used in manual analysis of the data in Fig. 1: PSD-95 puncta were counted only if entirely inside the simulated spine (circle), vGlut1 puncta were counted if there was partial overlap with or if the cluster contacted the simulated spine (circle). Three independent simulations per spine size (500 randomizations in a single run) were performed on images acquired from three independent experiments. (d) Quantification of the average proportion of all simulated and real spines that did not colocalize with either PSD-95 or vGlut1 clusters (***p < 0.0001, t = 180, df = 26 (PSD-95), (vGlut1) t = 15.70 df = 26 (vGlut1), two-tailed Student’s t-test, circles represent averages of all empty spines from three independent simulations (total of 34770 (PSD-95) and 29239 (vGlut1) empty spines), and averages of empty spines per neuron as measured in our data, n = 25 neurons). (e-f) Quantification of the proportion of spines of indicated size that do not colocalize with (e) PSD-95 nanomodules (n = 34770 empty spines) or (f) vGlut1 nanomodules (29239 empty spines). The majority of simulated spines in each size category do not contain any PSD-95 or vGlut1 puncta. (g-h) Quantification of the proportion of simulated spines that contain (g) PSD-95 (n = 1561 spines), or contain/contact (h) vGlut1 puncta (n = 6761 spines). (g) The majority of simulated spines that associate with PSD-95 clusters contain only a single PSD-95 nanomodule (93.5%) with significantly fewer two (5.7%), three (0.6%), four (0.06%) and five (0.01%) nanomodule-containing spines (p < 0.0001, ANOVA, F(4, 334) = 4507). (h) The majority of vGlut1 containing spines have only a single vGlut1 nanomodule (81.4%) with significantly fewer two (15.9%), three (2.2%), four (0.2%) and five (0.04%) nanomodule-containing spines (p < 0.0001, F(4, 355) = 2631, one-way ANOVA). Bar graphs (d-h) show mean +/− SEM. The dots represent the averages of simulations using images from three independent experiments; each dot is an average of three independent randomization runs per spine size. (i-j) Graphs displaying the correlation between simulated spine sizes and the number of (i) PSD-95 (Sim. 1, n = 863, Sim. 2, n = 120, Sim. 3, n = 270 PSD-95-containing spines) and (j) vGlut1 nanomodules (Sim. 1, n = 2961, Sim. 2, n = 512, Sim. 3, n = 740 vGlut1-containing spines) using Pearson’s correlation. Only spines that contained clusters were included in this analysis. Averages for PSD-95 and vGlut1 from three independent simulation experiments are shown in Fig. 1i, j.
Supplementary Figure 7 Monte Carlo simulations of SYP-1, vGlut1, Bassoon and PSD-95 synaptic nanomodules.
Analysis was conducted as in Fig. S6 except as described (a-b) Super-resolved maximum intensity projection images of PSD-95 (green), vGlut1 (red), SYP-1 (blue) nanomodules as in Fig. 2 and Bassoon nanomodules as in Fig. 3. Images were acquired with a Leica SP8 3X instrument using gated detectors for vGlut1 (AlexaFluor-594, 0.3 to 6 nsec), SYP-1 (Atto-647N, 0.2 to 6 nsec) and Bassoon (Atto-647N, 0.2 to 6 nsec) in combination with a pulsed 775 nm depletion beam for a maximum resolution of ~50 nm. Scale bar for a and b, 5 µm. Magenta circles indicate simulated spine sizes reflecting spine sizes from our data in Fig. 2 and Fig. 3 at 12.5, 25, 37.5, 50, 62.5, 75, 87.5 and 99 percentiles. The simulations were repeated using images from three independent transfection experiments from Fig. 2 and 3. (c-f) Quantification of the proportion of simulated spines of the indicated size that did not colocalize with (c) PSD-95 nanomodules (n = 30375 spines), (d) vGlut1 nanomodules (n = 26131 spines), (e) SYP-1 nanomodules (n = 27511 spines) and (f) Bassoon nanomodules (27091 spines). (g-j) Quantification of the proportion of (g) PSD-95, (h) vGlut1, (i) SYP-1 and (j) Bassoon nanomodules in the remaining spines that contain clusters. The majority of simulated spines that associate with puncta contain only a single nanomodule with significantly fewer multi-nanomodule containing spines. (g) PSD-95 simulations (single, 96.8%; two, 3.1%; three, 0.09%; four and five, 0%, p < 0.0001, F(4, 10) = 11750, one-way ANOVA, n = 1124 spines). (h) vGlut1 simulations (single, 78.1%; two, 17%; three, 3.6%; four, 1% and five, 0.1%, p < 0.0001, F(4, 10) = 381, one-way ANOVA, n = 5415 spines). (i) SYP-1 simulations (single, 90.6%; two, 8.3%; three, 0.8%; four, 0.05% and five, 0%, p < 0.0001, F(4, 10) = 852.2, one-way ANOVA, n = 3958 spines). (j) Bassoon simulations (single, 88.5%; two, 9.1%; three, 0.7%; four, 0.09% and five, 0%, p < 0.0001, F(4, 10) = 8162, one-way ANOVA, n = 4332 spines). Graphs in (c-j) represent mean +/− SEM. (k-m) Pearson’s correlations between simulated spine sizes and the number of (k) PSD-95 (Sim. 1, n = 629, Sim. 2, n = 336, Sim. 3, n = 159 PSD-95-containing spines), (l) vGlut1 (Sim. 1, n = 1291, Sim. 2, n = 2148, Sim. 3, n = 1976 vGlut1-containing spines), (m) SYP-1 (Sim. 1, n = 1962, Sim. 2, n = 759, Sim. 3, n = 1237 SYP-1-containing spines) and (n) Bassoon (Sim. 1, n = 1800, Sim. 2, n = 1317, Sim. 3, n = 1215 Bassoon-containing spines). Only spines that contained clusters were included in this analysis. Averages from three independent simulations for PSD-95, vGlut1 and SYP-1 are shown in Fig. 2f–h and for Bassoon in Fig. 3h.
Supplementary Figure 8 Modular organization of spine synapses in brain sections not subjected to CUBIC.
Neurons in mouse cortex were transduced with EGFP at P7 to label spines and dendritic morphology as described in Fig. 4. At P28-35, brain sections were fixed and labeled with antibodies for PSD-95 and vGlut1 (Online Methods). Spines were analyzed for PSD-95 (green) and vGlut1 (red) modules by imaging in 3D STED mode on a Leica SP8 3X instrument. (a) High-contrast images and corresponding 3D reconstruction of one-module (top panels) and two-module (bottom panels) spines. Scale bar, 0.75 µm. Spines in 3D-rendered images are shown as transparent in order to enable the visualization of PSD-95 puncta (green). Scale bars, 0.3 µm (top panel), 0.2 µm (bottom panel). Similar results were obtained using brain sections from three different EGFP-injected animals. (b and e) Quantification of the percentage of spines (n = 3 independent injection experiments/animals) containing single and multiple (b) PSD-95 and (e) vGlut1 modules (n = 156 spines). (c and f) Quantification of the average areas of individual (c) PSD-95 (n = 255 clusters, n = 87 (single), n = 90 (double), n = 54 (triple), n = 24 (quadruple), p = 0.1075, F (3, 251) = 2.050, one-way ANOVA) and (f) vGlut1 (n = 258 clusters, n = 85 (single), n = 90 (double), n = 63 (triple), n = 20 (quadruple), p = 0.0562, F (3, 254) = 2.55, one-way ANOVA) nanomodules. (d and g) Quantification of the total area of (d) PSD-95 (n = 87 (single), n= 45 (double), n = 18 (triple), n = 6 (quadruple), ***p < 0.0001, *p = 0.0414, F (3, 152) = 83.58, one-way ANOVA), and (g) vGlut1 (n = 85 (single), n = 45 (double), n = 21 (triple), n = 5 (quadruple), ***p < 0.0001, **P = 0.0002, *p = 0.0034, F (3, 152) = 53.01, one-way ANOVA with Fisher’s LSD post hoc). (h and i) Plots of the relationship between cluster number and spine size. A positive correlation between (h) PSD-95 (green line, Pearson’s R2 = 0.5275, p < 0.0001, n = 156 spines) and (i) vGlut1 (red line, Pearson’s R2 = 0.4957, p < 0.0001, n = 156 spines) cluster number and spine size was observed. (j) A plot of the relationship between PSD-95 and vGlut1 nanomodules (Pearson’s R2 = 0.6224, p < 0.0001, n = 156 spines). Bar graphs show mean +/− SEM, with numbers of individual spines or clusters represented by dots.
Simulations were conducted as in Fig. S6 except as noted. (a) Super-resolved maximum intensity projections of PSD-95 (green) and vGlut1 (red) nanomodules. Maximum projections were made from a small substack that corresponded to the width of the majority of spines in the image (≤1.5 µm). Scale bar, 2 µm. Simulations using images from three independent injection experiments are shown. (b) Binary PSD-95 (top panels) and vGlut1 (bottom panels) masks generated by thresholding the corresponding images shown in panel (a) as described in Online Methods. Magenta circles indicate simulated spine sizes selected from spine sizes in our data set in Fig. 4 at the 14, 28, 42, 56, 70, 84 and 98 percentiles. (c) Quantification of the average proportion of all simulated and real spines that did not colocalize with either PSD-95 or vGlut1 clusters (***p < 0.0001, t = 92.38, df = 33 (PSD-95), t = 2703, df = 33 (vGlut1), two-tailed Student’s t-test, circles represent averages of all empty spines from three independent simulations (26603 and 23490 simulated spines not containing PSD-95 or vGlut1) and averages of empty spines per neuron as measured in our data, n = 32 neurons). (d-e) Quantification of the proportion of spines of the indicated sizes that did not colocalize with (d) PSD-95 nanomodules (n= 26603 spines) or (e) vGlut1 nanomodules (n = 23490 spines). The majority of simulated spines in each size category do not contain clusters. (f-g) Quantification of the proportion of (f) PSD-95 (n = 4897) and (g) vGlut1 (n = 8011) nanomodules in the remaining spines that contain clusters. (f) The majority of simulated spines that associate with PSD-95 clusters contain only a single PSD-95 nanomodule (93%) with significantly fewer two (6.4%), three (0.5%), four (0.05%) and five (0%) nanomodule-containing spines (1 vs. 2, 1 vs. 3 1 vs. 4, 1 vs. 5, p < 0.0001; 2 vs. 3, p = 0.011; 2 vs. 4 and 2 vs. 5, p = 0.007; 3 vs. 4 and 3 vs. 5, p = 0.8; 4 vs. 5, p = 0.9; F(4, 10) = 924.5, one-way ANOVA, Fisher’s LSD post hoc). (g) The majority of vGlut1-containing spines have only a single vGlut1 nanomodule (82.6%) with significantly fewer two (14.3%), three (2.5%), four (0.4%) and five (0.04%) nanomodule-containing spines (1 vs. 2, 1 vs. 3 1 vs. 4, 1 vs. 5, p < 0.0001; 2 vs. 3, p = 0.008; 2 vs.4 and 2 vs. 5, p = 0.002; 3 vs. 4, p = 0.42; 3 vs. 5, p = 0.34; 4 vs. 5, p= 0.88; F(4, 10) = 384.4, one-way ANOVA, Fisher’s LSD post hoc). Bar graphs (c-g) show mean +/− SEM. The dots represent the averages of simulations using images from three independent experiments; each dot is an average of three independent randomization runs of 500 spines per spine size. (h-i) Graphs displaying the Pearson’s correlation between simulated spine sizes and the number of (h) PSD-95 (Sim. 1, n = 2374, Sim. 2, n = 1195, Sim. 3, n = 1328 PSD-95-containing spines) and (i) vGlut1 nanomodules (Sim. 1, n = 2977, Sim. 2, n = 3072, Sim. 3, n = 1962 vGlut1-containing spines). Only spines that contained clusters were included in this analysis. Averages for PSD-95 and vGlut1 from three independent simulation experiments are shown in Fig. 4i, j.
Supplementary Figure 10 Stimulation of NMDARs with glycine induces long-lasting increases in spine size but does not lead to changes in PSD-95 nanomodule size during live-cell STED imaging.
(a) Quantification of the average change in dendritic spine size after a three-minute treatment with 200 µM glycine (vertical dashed line). Significant increases in spine size over baseline were observed only in the potentiated spines (green trace, n = 25 of 49 spines). Non-responsive spines did not increase in size upon glycine treatment (red trace, n = 24 of 49 spines). The increases in spine size were blocked by treatment with the NMDAR inhibitors (n = 21 spines, gray trace), APV (50 µM) and MK801 (10 µM). Control spines (n = 20 spines) that were not treated with glycine did not change size on average during the three-hour imaging period. Quantification of spine size increase was performed at every time point (30 min,***p < 0.0001, **p = 0.0002, *p = 0.018, one-way ANOVA with Tukey’s post hoc). Experiments were repeated ≥ 3 times. (b) Quantification of PSD-95 nanomodule size during three-hour live-cell STED in potentiated spines. The size of each PSD-95 nanomodule at every time point was normalized to the size before glycine treatment (-30 min timepoint) in every potentiated spine. Only nanomodules present during the entire duration of imaging were included (p = 0.4421, F(7, 224) = 0.9861, n = 23 PSD-95 clusters, one-way ANOVA at every timepoint). Error bars indicate mean ± SEM. Gray lines in panel b indicate individual PSD-95-EGFP clusters.
Supplementary Figure 11 Endogenous PSD-95 nanomodules visualized by PSD-95 intrabodies accumulate in potentiated spines following glycine treatment.
Neurons were transfected with tdTomato to visualize spines and neuronal morphology and the FingR-PSD-95-EGFP construct to label endogenous PSD-95 clusters at DIV 0. At DIV 21–25, neurons were subjected to live-cell STED for three hours. (a) Representative time-lapse images of tdTomato-labeled spines (red) containing EGFP-labeled endogenous PSD-95 clusters (green, white arrows) in control, potentiated and non-responsive spines. Scale bar, 650 nm. Control neurons were imaged for three hours in ACSF without glycine stimulation. Chemical LTP was induced with glycine (200 µM, 3 minutes, black arrow) and changes in spine morphology and PSD-95 nanomodule number were followed every 30 minutes for three hours. Similar results were obtained from three independent transfection experiments. (b) Quantification of the changes in endogenous PSD-95 nanomodule number in control (n = 49), potentiated (n = 34) and non-responsive (n = 38) spines over three hours. Measurements were performed at each time point (*p = 0.0248, (potentiated vs. control only); **p = 0.0012 (30 min), p = 0.0045 (60 min), ***p = 0.0002 (90 min), p = 0.0004 (120 min), p = 0.0002 (150 min), all comparisons, one-way ANOVA, Fisher’s LSD post hoc). (c) Quantification of the total distance moved over three hours (p ≤ 0.0001, F(2, 136) = 11.80, control, n = 55 modules; potentiated n = 39 modules; non-responsive, n = 45 modules; one-way ANOVA, Fisher’s LSD post hoc). (d) Quantification of the average change in dendritic spine size after a three-minute treatment with 200 µM glycine (vertical dashed line). Potentiated spines exhibited a sustained increase in size over three hours that was significantly different form non-responsive and control spines that were not treated with glycine (**p < 0.0001, one-way ANOVA, Fisher’s LSD post hoc). Graphs show mean +/− SEM at every timepoint.
Supplementary Figures 1–11
Representative field of view of super-resolved synaptic staining in a brain slice from mouse cortex. The movie shows a 3D reconstruction of a CUBIC1/2 cleared section demonstrating an EGFP-labeled dendrite that was imaged in confocal resolution (gray) and with dual-color STED to image endogenous PSD-95 (green) and vGlut1 (red) puncta at super resolution (xy ~80 nm and z ~200–300 nm, Online Methods). The movie was generated by reconstructing the resulting image stacks in Imaris software. Only the puncta that colocalized with dendritic spines are shown in the final 3D image reconstructions. This particular dendrite contains three spines, two of which contain two aligned PSD-95 and vGlut1 modules, while the third contains a single aligned PSD-95 and vGlut1 module. Experiments were repeated ≥3 times.
Representative time-lapse movie of a control-treated neuron. The white arrow indicates the spine that was selected for retrospective STED imaging of PSD-95 and vGlut1 and is shown in Fig. 5a. The spine shows dynamic movements, but does not increase in size during the 3-h imaging period. Scale bar, 2 µm. Experiments were repeated ≥3 times.
Supplementary Video 3 - Structural plasticity associated with cLTP is linked to synaptic module number.
Representative time-lapse movies of spine morphology following cLTP. Images were acquired using a spinning-disk confocal microscope every 6 min for 3 h using an adaptive focus control system. The top movie shows a spine (yellow arrow) that increased in size after a 3-min treatment with 200 µM glycine (green dot) and remained larger for the remainder of imaging. The middle movie shows a nonresponsive spine (white arrow) that did not become larger after 3-min glycine treatment. The bottom movie is a time-lapse sequence acquired from a neuron treated with NMDAR inhibitors APV (50 µM) + MK-801 (10 µM) prior to glycine treatment. Spines indicated by arrows are shown in Fig. 5b–d and are examples of spines subjected to retrospective STED imaging of PSD-95 and vGlut1. Scale bar, 2 µm. Experiments were repeated ≥3 times.
Supplementary Video 4 - Endogenous PSD-95 nanomodules visualized by PSD-95 intrabodies accumulate in potentiated spines following glycine treatment.
Representative time-lapse movies of spines containing EGFP-labeled endogenous PSD-95 subjected to cLTP or control treatments. Images were acquired every 30 min for three hours. Chemical LTP was induced with glycine (3 min, 200 µM, magenta dot) between the first and second time-lapse image. Endogenous PSD-95 (green, labeled with EGFP FingR intrabody) was imaged in STED mode (~100 nm FWHM) and spine morphology (red, labeled with tdTomato) was imaged in confocal mode (~300 nm FWHM). Movies show merged images. The conditions shown are as labeled and appear in Supplementary Fig. 11. Sustained enlargement is observed in a potentiated spine, which is associated with the appearance of a new PSD-95 nanomodule. The nonresponsive spine exhibits only transient enlargement and additional PSD-95 nanomodules do not accumulate in this condition. The spine in the control condition, not treated with glycine, does not increase in size and remains single-clustered for the remainder of imaging (3 h). Scale bar, 650 nm. Experiments were repeated ≥ 3times.
Supplementary Video 5 - Rapid remodeling of aligned pre- and postsynaptic modules underlies cLTP structural plasticity.
Representative time-lapse movies of spines imaged with live-cell STED after cLTP or control treatments. Live-cell STED images of PSD-95-EGFP (green) and SYP-mTurq2 (red) are shown with simultaneously collected confocal images of tdTomato (white). Movie shows overlays. The conditions shown are as labeled and appear in Fig. 6. PSD-95–EGFP (green) and SYP-mTurq2 (red) clusters are precisely aligned with one another in all four conditions and stay aligned during the 3 h of imaging. The top set of images shows a control-treated spine. Treatment with glycine (3 minutes, 200 µM) leads to a rapid increase in the mobility of both PSD-95-EGFP and SYP-mTurq2 at potentiated spines (top movie). Nonresponsive spines display no new modules and control levels of movement (lower movie). APV and MK-801 treatment reduced the mobility of PSD-95–EGFP and SYP-mTurq2 (bottom movie). Experiments were repeated ≥3 times.
Supplementary Video 6 - NMDAR-dependent plasticity is associated with fast modification of synaptic nanoarchitecture.
Representative time-lapse movies of spines imaged with live-cell STED after cLTP or control treatments. Images were collected every 12.5 min for 1 h. Live-cell STED images of PSD-95–EGFP (green) and SYP-mTurq2 (red) are shown with simultaneously collected confocal images of tdTomato (gray and white outlines). Movie shows each channel separately, as well as the merged image of PSD-95–EGFP and SYP-mTurq2. The conditions shown are as labeled and appear in Fig. 7a. Treatment with glycine (3 min, 200 µM, magenta dot) leads to a rapid increase in dynamics of pre- and postsynaptic nano-organization that is associated with the addition or disappearance of PSD-95-EGFP and SYP-mTurq2 modules in early stages following glycine treatment. New clusters of PSD-95-EGFP and SYP-mTurq2 can be seen by the end of imaging (1 h). The nonresponsive spine, which does not increase in size, does not exhibit increased mobility or addition of new nanomodules, similarly to the control spine that was not subjected to glycine treatment. Scale bar, 1 µm. Experiments were repeated ≥3 times.
Time-lapse movie of FRAP experiments after cLTP or control treatments. The FRAP experiment is 22 min long, with images acquired every 30 s. Chemical LTP was induced with glycine (200 µM, 5 min, magenta dot). Following cLTP, recovery of PSD-95 in the indicated conditions was monitored for additional 12.5 min. Conditions shown are as labeled and appear in Fig. 7b. Scale bar, 1 µm. Experiments were repeated ≥3 times.
About this article
Cite this article
Hruska, M., Henderson, N., Le Marchand, S.J. et al. Synaptic nanomodules underlie the organization and plasticity of spine synapses. Nat Neurosci 21, 671–682 (2018). https://doi.org/10.1038/s41593-018-0138-9
This article is cited by
CLARITY increases sensitivity and specificity of fluorescence immunostaining in long-term archived human brain tissue
BMC Biology (2023)
Fyn nanoclustering requires switching to an open conformation and is enhanced by FTLD-Tau biomolecular condensates
Molecular Psychiatry (2023)
Nature Reviews Neuroscience (2023)
Increasing dopamine synthesis in nigrostriatal circuits increases phasic dopamine release and alters dorsal striatal connectivity: implications for schizophrenia
MAxSIM: multi-angle-crossing structured illumination microscopy with height-controlled mirror for 3D topological mapping of live cells
Communications Biology (2023)