Fig 1.
Presynaptic strengths segregate depending on both the dendritic location and the identity of presynaptic inputs.
(A) Experimental scheme to compare synaptic inputs from two presynaptic cells on the same postsynaptic neuron. Patch-clamp recordings were used to monitor evoked synaptic responses; FM4-64 dye loading of the presynaptic terminals of each of the two presynaptic cells was used to map the distribution of presynaptic strengths across dendrites of the postsynaptic neuron filled with an Alexa Fluor dye. (B) Example of average EPSC traces to paired stimuli (50-ms interstimulus interval) for each input. (C–D) Scatter plot comparing the EPSC amplitude (C) and PPR (D) of the two inputs. Plots include previously reported data set (n = 29 triplets from 12 cultures), which was obtained under exactly the same experimental conditions as the present study (Supplementary Information S7B Fig in [48]). (E) Experimental scheme for estimating RRP size at individual boutons from the two inputs. FM4-64 was loaded by 40 APs at 20 Hz by stimulating presynaptic cells via the patch pipette. Images were acquired before and after unloading the FM4-64 dye with 600 APs at 10 Hz, and the remaining signal was taken as background. (F) Example images of dye-loaded synapses (left), after first (middle) and second (right) unloading stimulations. Scale bar, 5 μm. Red and green arrowheads indicate boutons before (filled) and after (open) the unloading. (G) Image showing the overlay of the fluorescence signals corresponding to each of the two inputs (green and red). The signals for input 1 (red) and input 2 (green) were obtained by subtracting image 2 from image 1 and image 3 from image 2, respectively, as shown in F. Scale bar = 5 μm. (H) Summary graph showing the fluorescence intensity difference for pairs of boutons belonging to the same or different axons and contacting the same or different dendrites (n = 150 boutons from 3 triplets, Kruskall–Wallis test followed by Dunn multiple comparison test, *p < 0.05, ***p < 0.001). Underlying data can be found in S1 Data. AP, action potential; EPSC, excitatory postsynaptic current; norm., normalized; ns, not significant; PPR, paired-pulse ratio; Rec, recording pipette; RRP, readily releasable pool.
Fig 2.
Postsynaptic strengths segregate depending on dendritic branches and locally, on presynaptic inputs.
(A) Triple recording scheme with two independent presynaptic neurons forming convergent inputs onto a postsynaptic neuron. (B) Example traces of aEPSCs recorded in Sr2+-containing aCSF in response to stimulating each of the two convergent inputs. (C) Comparison of average aEPSC amplitude of two inputs (R2 = 0.95, p < 0.0001, n = 8 triplets). Linear regression and 95% confidence interval (gray) are shown. (D) Histograms from a representative triple recording experiment showing quantal event amplitudes for each of the two convergent inputs (gray and white bars). (E) Comparison of mean aEPSC frequency of two convergent inputs shows no correlation (R2 = 0.083, p = 0.4885, n = 8 triplets). (F) Scheme to probe the distribution of surface AMPA receptors. A postsynaptic neuron expresses SEP-GluA1 or SEP-GluA2 (yellow), and a nearby presynaptic partner is filled with an Alexa Fluor dye (blue). (G) Fluorescence images from a representative experiment. Left, postsynaptic SEP-GluA1–positive dendrites are contacted by an Alexa Fluor dye–labeled axon; some SEP-GluA1 puncta are contacted by the labeled axon (filled arrowheads), and others are presumed to receive synaptic inputs from unlabeled axons (open arrowheads). Right, same view shows punctate SEP-GluA1 signal representing putative postsynapses whose fluorescence intensity is color-coded. Scale bar, 10 μm. (H) Summary graph of the fluorescence intensity difference for pairs of SEP-GluA1 (left) or SEP-GluA2 (right) puncta apposed to different or same axon and belonging to different or same dendrites (SEP-GluA1: n = 461 puncta from 8 cells, SEP-GluA2: n = 1,066 puncta from 9 cells, one-way ANOVA followed by Tukey multiple comparison test, **p < 0.01, ***p < 0.001). Underlying data can be found in S1 Data. aCSF, artificial cerebrospinal fluid; aEPSC, asynchronous EPSC; EPSC, excitatory postsynaptic current; AMPA, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; GluA, AMPA receptor subunit; Rec, recording pipette; SEP, superecliptic pHluorin.
Fig 3.
Activity normalizes presynaptic strengths in the stimulated axon.
(A) Scheme for imaging synaptic vesicle dynamics at individual boutons. A presynaptic neuron expressing VGLUT1-pHluorin (green) and a postsynaptic partner filled with an Alexa Fluor 594 dye (red) were patch clamped. (B) Experimental scheme for monitoring activity-dependent changes in the RRP size of individual boutons contacting the postsynaptic neuron as shown in (A). (C) Fluorescence image showing Alexa Fluor dye–filled dendrites (red) contacted by VGLUT1-pHluorin–positive puncta (green). Putative synaptic contacts are indicated by arrowheads. Scale bar, 30 μm. (D) An example of eliciting presynaptic plasticity in a single connection. Images of VGLUT1-pHluorin fluorescence change (top) and fluorescence intensity traces (bottom) at indicated boutons (arrowheads) triggered by a 40-AP, 20-Hz stimulation are shown before and 20 min after the conditioning stimulation to elicit plasticity. (E) Plot comparing the RRP size before and after the conditioning stimulation for individual boutons from the example shown in (C,D) (n = 25). (F) Comparison of the distribution of RRP size before and after the conditioning stimulation for all boutons analyzed (n = 134 boutons from 7 cells). (G) Summary plot comparing the change in RRP induced by the conditioning stimulation versus the initial RRP size (R2 = 0.13, p < 0.0001, n = 134 boutons from 7 cells). The red line represents the linear regression. (H) Summary of RRP size difference analysis for pairs of boutons apposed to different dendritic branches or the same dendritic branch, before and after the conditioning stimulation, expressed as mean ± SEM (one-way ANOVA followed by Tukey multiple comparison test, *p < 0.05, ***p < 0.001). Underlying data can be found in S1 Data. AF, Alexa Fluor; AP, action potential; RRP, readily releasable pool; stim, stimulation; VGLUT1, vesicular glutamate transporter 1
Fig 4.
Activity-dependent downscaling of postsynaptic strengths compensates for presynaptic potentiation.
(A) Experimental scheme to compare the effect of the conditioning stimulation on quantal properties of convergent inputs in Sr2+-containing aCSF. The conditioning stimulation (1 Hz, 3 min) is given to one of the two presynaptic cells in Ca2+-containing aCSF. (B) Example traces of aEPSCs recorded in Sr2+-containing aCSF for stimulated (blue) and nonstimulated (black) input, before and >15 min after the conditioning stimulation. The trace on the top includes the first synchronous peak (low scale); the two bottom traces show asynchronous quantal events (high scale). (C,D) Cumulative distributions of stimulated (C) and nonstimulated (D) aEPSC amplitude before and >15 min after the conditioning stimulation. Insets show the scaled cumulative distributions. (E,F) Summary plots showing the comparison of rank-ordered aEPSC amplitudes after versus before the stimulation at stimulated (E: R2 = 0.99, p < 0.0001) and nonstimulated (F: R2 = 0.99, p < 0.0001) inputs (n = 8 experiments). Linear regression is shown in red where m indicates the slope. (G–J) Summary of mean aEPSC amplitude (G,H) and frequency (I,J) before and after the stimulation at stimulated (G,I) and nonstimulated (H,J) inputs (n = 8 triplets). Data were evaluated using the paired, two-tailed Student t test. *p < 0.05. (K) Comparison of the percent change between stimulated and nonstimulated connections in mean aEPSC amplitude (R2 = 0.69, p = 0.0104), with linear regression (black line) and 95% confidence interval (gray). (L) Comparison of the percent change between aEPSC amplitude and aEPSC frequency for the stimulated connection (R2 = 0.77, p = 0.004), with linear regression (black line) and 95% confidence interval (gray). Underlying data can be found in S1 Data. aCSF, artificial cerebrospinal fluid; aEPSC, asynchronous EPSC; EPSC, excitatory postsynaptic current; Rec, recording pipette; stim, stimulation.
Fig 5.
Uniform downscaling of postsynaptic AMPA receptors across the dendritic tree in primary hippocampal cultures.
(A) Scheme to visualize individual postsynaptic strengths of stimulated and nonstimulated synapses. (B) Representative EPSC trace evoked by a pair of brief depolarizations (+100 mV, 50-ms interval) of the presynaptic neuron. (C–D) Time course of EPSC amplitude (C) and PPR (D) from an example recording. (E) Fluorescence image of a SEP-GluA2–expressing postsynaptic neuron (yellow) receiving putative synaptic contacts from an axon of a neuron filled with an Alexa Fluor dye (blue). Scale bar, 40 μm. (F) Details of the boxed area in (E). Arrowheads indicate putative synaptic contacts between the stimulated presynaptic cell and the postsynaptic cell. Scale bar, 10 μm. (G) Illustration of the basis for discriminating between stimulated and nonstimulated synapses. SEP-GluA2 puncta apposed to the labeled axon are considered as stimulated synapses, while the others are considered as nonstimulated. (H) Time-lapse images of color-coded fluorescence intensity showing the decrease of SEP-GluA2 signal after the conditioning stimulation for the area corresponding to (F). Individual SEP-GluA2 puncta associated to the labeled axon shown in (F) are indicated (arrowheads). Scale bar, 10 μm. (I,J) Time course of SEP-GluA2 fluorescence intensity at putative stimulated synapses (I: paired, two-tailed Student t test, ****p < 0.0001, n = 12 synapses) and nonstimulated synapses (J: paired, two-tailed Student t test, ****p < 0.0001, n = 34 synapses) for the same postsynaptic neuron. (K,L) Plots of the SEP-GluA2 fluorescence intensity before versus after the stimulation at stimulated synapses and nonstimulated synapses for the same example in (E–J) (K: R2 = 1.00; L: R2 = 0.92, p < 0.0001). (M) Comparison of the extent change in fluorescence intensity of SEP-GluA2 at stimulated versus nonstimulated synapses (R2 = 0.77, p = 0.0209, n = 6 cells), with linear regression shown. Data are expressed as mean ± SEM. m indicates the slope. (N) Comparison of the extent change in fluorescence intensity of SEP-GluA2 versus PPR change at stimulated synapses (R2 = 0.77, p = 0.0209, n = 6 cells), with linear regression shown. m indicates the slope. Underlying data can be found in S1 Data. AMPA, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; AP, action potential; a.u., arbitrary unit; EPSC, excitatory postsynaptic current; GluA, AMPA receptor subunit; PPR, paired-pulse ratio; SEP, superecliptic pHluorin; stim, stimulation.
Fig 6.
Spine head volume of CA3 cells segregate depending on dendritic branches and, locally, on presynaptic CA3 inputs.
(A) Scheme illustrating a single pair of CA3 pyramidal cells from an organotypic slice that have been transfected to express GFP and tdTomato. (B) Experimental strategy to express different fluorescent proteins in a pair of CA3 cells. Whole-cell recordings are used to ensure the functional connectivity between the two cells and to infuse plasmids encoding GFP and tdTomato, allowing for visualization and discrimination of their axons and dendrites 24–36 h later. (C) Traces showing current-voltage relationship from a transfected CA3 pyramidal cell (100 pA current injection steps for 600 ms, from -200 pA to +700 pA). (D) Summary of RMP and Ri of transfected cells and control cells in same recording conditions (transfected: n = 10 cells, control n = 20 cells, Mann–Whitney test). (E) Example traces recorded from a connected CA3 cell pair showing synaptic currents (gray traces) evoked in response to an AP (red trace) triggered in the presynaptic cell 24–36 h after the transfection (second patch). The thick black trace at the bottom represents the average EPSC trace. (F) Confocal images acquired with a scanning laser microscope showing morphological interactions between CA3 neurons from a slice fixed 24–36 h after transfection. Left panel shows low magnification image of the 2D projection of a confocal stack. Right panels show high magnification views of the morphological contacts (arrowheads) between axonal boutons and dendritic spines from the CA3 neurons expressing GFP or tdTomato. Scale bar: Left panel, 50 μm; right panels, 5 μm. (G) Summary graph showing the spine head volume difference for pairs of spines apposed to different or same axon and belonging to different or same dendritic segments (n = 318 spines from 5 pairs, Kruskall–Wallis test followed by Dunn multiple comparison test, *p < 0.05, ****p < 0.0001). Underlying data can be found in S1 Data. AP, action potential; CA3, Cornu Ammonis 3; DG, dentate gyrus; EPSC, excitatory postsynaptic current; GFP, green fluorescent protein; ns, not significant; Ri, input resistance; RMP, resting membrane potential.
Fig 7.
LTD induction at unitary CA3 recurrent connections triggers functional and structural plasticity at both stimulated and nonstimulated synapses.
(A,C) Left: representative traces showing synaptic currents in the postsynaptic CA3 neuron evoked by a pair of APs triggered in the presynaptic CA3 neuron (2–3 nA, 50-ms interval), before and 20 min after LTD induction in absence (A) or presence (C) of 50 μM D-AP5. Middle: summary of the time course of EPSC amplitude (n = 10 cell pairs) for untreated (A) and D-AP5 (C). The gray shaded box represents the LTD induction. Right: graph plots showing PPR values before and after LTD induction in absence (A) or presence (C) of D-AP5 (untreated: n = 10 cell pairs; D-AP5: n = 7 cell pairs; paired two-tailed Student t test, *p < 0.05). (B,D) Left: example traces of sEPSCs recorded before (Pre-stim) and after (Post-stim) LTD induction in absence (B) or presence (D) of 50 μM D-AP5. Middle: graph plots showing sEPSC amplitude before and after LTD induction in absence (B) or presence (D) of D-AP5 (untreated: n = 9 cell pairs, D-AP5: n = 7 cell pairs, paired two-tailed Student t test, *p < 0.05). Right: cumulative distributions of sEPSC amplitudes before and after LTD induction in absence (B) or presence (D) of D-AP5. (E) 2D projection of optical sections acquired with an LSM from a slice fixed after the live-imaging session. The image shows a dendritic segment (green) with a spine apposed to a presynaptic terminal (arrowhead) of the labeled presynaptic CA3 cell axon (red). Scale bar, 5 μm. (F) 2D projections of time-lapse spinning disk confocal images acquired before (Pre-stim) and 5 min, 15 min, or 25 min after LTD induction and whose fluorescence intensity is color-coded. Images on the top show the same dendritic segment as in (E). The stimulated spine, the neighboring spines, and the distal spines are indicated with red, orange, and purple arrowheads, respectively. The image on the bottom shows a segment of dendrite with no spines contacted by the axon (control dendrite). Scale bar, 5 μm. (G) Time course of spine fluorescence integrated intensity normalized to the baseline for corresponding spines (stimulated, n = 9 spines from 4 slices; neighboring, 26 spines from 4 slices; distant, n = 28 spines from 4 slices; nonstimulated dendrite, n = 27 spines from 4 slices; two-way ANOVA test followed by Tukey and Dunett multiple comparison tests, *p < 0.05, **p < 0.01, ***p < 0.001). Underlying data can be found in S1 Data. AP, action potential; CA3, Cornu Ammonis 3; D-AP5, D-2-amino-5-phosphonovalerate; EPSC, excitatory postsynaptic current; LSM, laser scanning microscope; LTD, long-term depression; ns, not significant; PPR, paired-pulse ratio; sEPSC, spontaneous EPSC; stim, stimulation.
Fig 8.
Model for the activity-dependent distribution of pre- and postsynaptic strengths across incoming axons and dendrites.
Two axons (input 1 and input 2 in red and gray, respectively) contact several spines from two dendritic branches of the same target neuron. Input specificity of synaptic strengths is best represented at the level of the dendritic branch, where both pre- and postsynaptic strengths segregate depending on the specific presynaptic cells, suggesting that synapses interact locally to compete for or share local resources. In the schematic, different postsynaptic strengths are represented using different shades of blue or orange, and different presynaptic strengths are represented with a different number of presynaptic vesicles. The input-specific stimulation of a single presynaptic cell (red axon) results in the normalization of presynaptic strengths along the axon. The magnitude of presynaptic potentiation (or increased pr, indicated by upward red arrows) depends on the initial presynaptic strength. On the postsynaptic side, the long-lasting presynaptic potentiation accompanies a postsynaptic depression (inward red arrows) that spreads to nonstimulated neighboring spines and that may revert into potentiation at more distant spines (outward black arrows). pr, neurotransmitter release probability.