Differential distribution of endoplasmic reticulum controls metabotropic signaling and plasticity at hippocampal synapses Niklaus Holbro, Åsa Grunditz, and Thomas G. Oertner1 Friedrich Miescher Institute for Biomedical Research, Maulbeerstrasse 66, CH-4058 Basel, Switzerland Edited by Roger A. Nicoll, University of California, San Francisco, CA, and approved July 22, 2009 (received for review May 8, 2009) Synaptic plasticity is considered essential for learning and storage signaling in synapses made on spines containing ER and spines of new memories. Whether all synapses on a given neuron have without ER. To stimulate identified synapses in intact tissue, we the same ability to express long-term plasticity is not well under- combined two-photon imaging with two-photon glutamate un- stood. Synaptic microanatomy could affect the function of local caging. We show that ER is specifically associated with potent signaling cascades and thus differentially regulate the potential for synapses and governs the potential for mGluR-mediated synap- plasticity at individual synapses. Here, we investigate how the tic depression. In spines containing ER, calcium influx through presence of endoplasmic reticulum (ER) in dendritic spines of CA1 NMDARs was dwarfed by large calcium release events that are pyramidal neurons affects postsynaptic signaling. We show that likely to play a key role in plasticity induction. The preferential the ER is targeted selectively to large spines containing strong association of ER with the most potent synapses suggests that synapses. In ER-containing spines, we frequently observed synap- mGluR-mediated depression plays an important role in the tically triggered calcium release events of very large amplitudes. balance of excitation, counteracting the tendency of potent Low-frequency stimulation of these spines induced a permanent synapses to become even stronger over time. depression of synaptic potency that was independent of NMDA receptor activation and specific to the stimulated synapses. In Results contrast, no functional changes were induced in the majority of Genetic Approach To Identify ER-Containing Spines in CA1 Pyramidal NEUROSCIENCE spines lacking ER. Both calcium release events and long-term Cells. To visualize the ER in intact hippocampal tissue, we depression depended on the activation of metabotropic glutamate constructed a green ER label by fusing EGFP with ER-targeting receptors and inositol trisphosphate receptors. In summary, spine and ER-retention sequences (9). Organotypic hippocampal slice microanatomy is a reliable indicator for the presence of specific cultures were cotransfected biolistically with the ER label and a signaling cascades that govern plasticity on a micrometer scale. cytoplasmic red fluorescent protein (RFP) to visualize cell morphology (Fig. 1A). Two-photon microscopy was used to long-term depression ͉ metabotropic glutamate receptor ͉ image transfected CA1 pyramidal cells at high resolution and metaplasticity ͉ spine apparatus ͉ dendritic spines identify ER-containing (ERϩ) and other (ERϪ) spines on oblique dendrites close to the soma (Fig. 1B). Spines were ctivity-dependent changes in synaptic strength are thought considered ERϩ if they had a clear GFP signal inside the head Ato be essential for learning and the formation of new (Fig. 1C); spines with traces of ER in the neck region were not memories (1). The intracellular signaling cascades underlying analyzed. Analysis of several transfected CA1 pyramidal cells different forms of synaptic plasticity have been studied exten- showed that 18.7 Ϯ 2.3% of spines were positive for the ER label sively at the CA3 to CA1 projection in the hippocampus. (n ϭ 318 spines, 8 cells). A similar fraction of ER-containing Although long-term potentiation at these synapses is strictly spines has been reported in acute slices by using electron NMDA receptor-dependent, at least two mechanistically distinct microscopic reconstruction of CA1 cell dendrites (6). To test forms of long-term depression (LTD) have been described, whether spine ER was continuous with dendritic ER, we per- triggered by the activation of NMDA receptors (NMDARs) and formed fluorescence recovery after photobleaching (FRAP) metabotropic glutamate receptors (mGluRs), respectively (2). experiments on the ER label. A brief laser pulse (930 nm) was Although the potential for NMDAR-dependent plasticity can be used to bleach the green ER label inside the spines (Fig. 1D). On regulated by the subunit composition of the receptor itself, much average, ER fluorescence recovered with a time constant of less is known about the regulation of mGluR-dependent plas- 210 Ϯ 39 ms (n ϭ 33 spines). These data indicate that dendritic ticity (3). Aberrant mGluR signaling and dysregulated synaptic spines contain ER structures that are connected to dendritic ER plasticity have been implicated in severe mental disorders, such and that ER constituents (such as ions and proteins) can rapidly as fragile X mental retardation (4). The induction of mGluR- diffuse inside the ER lumen from the spine to the dendrite and dependent LTD is known to involve activation of postsynaptic vice versa. group I mGluRs and inositol trisphosphate (IP3)-mediated calcium release from the endoplasmic reticulum (ER) (reviewed ER-Containing Spines Have Large Heads and Often Contain a Spine in ref. 5). Interestingly, only a small subset of dendritic spines on Apparatus. Quantification of spine morphology showed that CA1 pyramidal cells contains ER (6). The heterogeneous dis- ERϩ spines had significantly larger cytoplasmic volumes tribution of this organelle very well could affect the plasticity of (0.058 Ϯ 0.005 ␮m3, n ϭ 49) than the rest of the population individual synapses (7, 8). In all previous studies of synaptic depression, plasticity was induced at large numbers of synapses simultaneously. However, Author contributions: N.H. and T.G.O. designed research; N.H. and Å.G. performed re- this strategy does not allow the investigation of functional search; N.H. analyzed data; and N.H. and T.G.O. wrote the paper. differences between individual synaptic connections. Differ- The authors declare no conflict of interest. ences in synaptic microanatomy, such as the presence of spe- This article is a PNAS Direct Submission. cialized organelles, could strongly regulate plasticity on a local 1To whom correspondence should be addressed. E-mail: [email protected] . scale, resulting in functional heterogeneities between individual This article contains supporting information online at www.pnas.org/cgi/content/full/ contacts. To test this hypothesis, we compared postsynaptic 0905110106/DCSupplemental. www.pnas.org͞cgi͞doi͞10.1073͞pnas.0905110106 PNAS Early Edition ͉ 1of6 Downloaded by guest on September 26, 2021 A B C D 300 200 20 % 100 ∆G/G 100 ms FRAP time constant (ms) 0 E 0.09 F 300 ) 3 * 0.06 200 0.03 100 spine head volume (µm FRAP time constant (ms) FRAP 0.00 0 ER+ ER- ER+ ER- Fig. 1. ER labeling and spine properties. (A) Organotypic hippocampal Fig. 2. ER-containing spines bear strong synapses. (A) Stimulation of ER- cultures were transfected biolistically with a cytoplasmic RFP (red) and an containing spine by two-photon glutamate uncaging (arrowhead). (Scale bar: ER-targeted EGFP (GFP-ER, green). (B) Two-photon image (maximum intensity 1 ␮m.) (B) Average uEPSCs for ER-containing (n ϭ 30) and other spines (n ϭ 44). projection) of transfected CA1 pyramidal cell. White box indicates region of Colored region represents SEM. (C) Peak amplitude of uEPSC was significantly analyzed oblique dendrites. (Scale bar: 50 ␮m.) (C) Dendrite of a transfected larger in ERϩ spines (n ϭ 30) compared with that in ERϪ spines (n ϭ 44). Values CA1 pyramidal cell with one large ER-containing spine (arrow). Overlay of red in C represent mean Ϯ SEM. (D) Relationship between spine volume and uEPSC (RFP) and green (GFP-ER) fluorescence results in yellow color. (Scale bar: 1 ␮m.) amplitude (n ϭ 62). (D) FRAP of the GFP-ER label was performed on ERϩ spines, and GFP diffusion inside the ER lumen was monitored (n ϭ 33 spines). (E) Spine head volume of ERϩ spines was significantly larger compared with that of ERϪ spines (ERϩ, In spines of pyramidal cells, ER often forms a specialized n ϭ 49; ERϪ, n ϭ 91) (F) Cytoplasmic FRAP time constants were not different organelle consisting of stacked membrane discs, the spine ap- ϩ Ϫ ϩ ϭ Ϫ ϭ in ER and ER spines (ER , n 26; ER , n 35). Values in D–F represent paratus (6). To assess which fraction of ERϩ spines in our mean Ϯ SEM. sample contained this organelle, we combined live ER imaging with posthoc immunohistochemistry against synaptopodin, a (0.028 Ϯ 0.002 ␮m3, n ϭ 91; Fig. 1E). Spine heads are separated protein associated with the spine apparatus (13) (Fig. S1A). We from their parent dendrite by a thin spine neck, allowing found that the majority (78%) of ERϩ spines were also positive biochemical compartmentalization of second messengers (10). for synaptopodin (Fig. S1B) and thus very likely contained a The ER could affect the diffusional coupling between spine head spine apparatus. and parent dendrite by physically obstructing the neck. To address this issue, we measured FRAP time constants (␶)of Synapses on ER-Containing Spines Have a High Potency. To assess the cytoplasmic RFP in ER-containing and other spines. We functional properties of spine synapses, we stimulated individual bleached Ϸ30% of the cytoplasmic red fluorescence inside the spines by two-photon glutamate uncaging. First, we identified spine and monitored fluorescence recovery. The recovery time ERϩ spines on oblique dendrites (Fig. 2A). After spine prese- constants in the two groups of spines were not significantly lection, cells were patch-clamped, and glutamate was uncaged on different (␶ERϩ ϭ 235 Ϯ 40 ms, n ϭ 26; ␶ERϪ ϭ 218 Ϯ 20 ms, n ϭ identified spines. On average, uncaging-evoked excitatory 35; Fig. 1F), demonstrating that the ER did not block cytoplas- postsynaptic currents (uEPSCs) had amplitudes of 11.4 Ϯ 0.7 pA mic diffusion between spine head and dendrite.