Article Figures & Data
Figures
- Figure 1.
Detection of uncaging spEPSPs using a VSD. A, Z-projection image of basal dendrite region, with spine of interest 92 μm from soma. On-spine recording, and peri-spine uncaging targets shown. B, Simultaneous spine optical (top; spEPSP dark trace) and somatic whole-cell electrical (bottom; soEPSP dark trace, 0.53 mV) recordings of uncaging induced EPSPs. Averaged uncaging sweeps are superimposed on control sweeps (light traces). Red trace is dual-exponential fit to spEPSP. Average of N = 6 sweeps uncaging and control, σnoise = 0.63%. Vertical dashed line indicates time of uncaging pulse (0.5 ms pulses). C, Simultaneous optical (top; same spine) and somatic whole-cell electrical (bottom) recordings of APs (current step depicted below) used for calibration. Averages of N = 10 sweeps APs, 23.6% peak. Same scale bars apply to all optical recordings for ease of comparison.
- Figure 2.
Calibration of the voltage dependence of spine ΔF/F using optical recordings in spines of bAPs. A, Image of neuron loaded with VSD and a region of basal dendrites. Red dot shows juxtasomatic basal dendrite recording location. Scale bar, 5 μm. B, Superimposed electrical action potential and optical bAP recordings (average of N = 8 sweeps) from juxtasomatic site showing reduced optical sensitivity at subthreshold potentials (circled). Peak optical sensitivity = 13.8% ΔF/F; 200 pA, 50 ms current injection. C, Binning and plotting optical (peak normalized) versus electrical data shows nonlinear voltage-dependent VSD sensitivity (N = 5 trials, gray lines). Blue dashed line shows quadratic fit to sensitivity data, whereas red dashed line (shown for comparison) represents a linear relationship between rest and the peak of the AP. D, Adjusting the optical data with nonlinear sensitivity calibration from C demonstrates accuracy in both subthreshold and superthreshold potential ranges. E, Amplitude of the optical bAP in the spine (percentage ΔF/F) versus distance from soma for N = 48 spines across 32 different cells. F, Attenuation of bAPs as a function of distance from soma in basal dendrites of L5 pyramidal neurons as previously reported using a VSD (length constant λ = 400 μm; Acker and Antic, 2009).
- Figure 3.
Uncaging responses in spines that are below the optical measurement threshold: clear optical bAP measurements serve as positive controls and for signal calibration. A–C, Same as Figure 1 on a different dendrite with target spine 52 μm from soma. A clear soEPSP is seen in B (0.5 mV peak), but no clear optical spEPSP is seen in B (top; no fit is made), whereas at the same time bAPs are clearly measureable optically in C (top; serves as positive control). Averages of N = 10 sweeps APs, 18.3% peak. Average of N = 10 sweeps uncaging and control, σnoise = 0.3%.
- Figure 4.
Amplitude and duration of uncaging-evoked spEPSPs. A, Histogram of calibrated spEPSP amplitudes, produced from a dual-exponential fit, and taken only if fit amplitude exceeded noise levels (σnoise) by a factor of at least 2.5 (mean = 13.0 ± 6.7 mV, N = 20, calibrated using optical bAP measurements as shown in Fig. 2). B, Cumulative histograms of spEPSP amplitudes. Dark bars correspond to supra-noise threshold data from A. Additionally, noise thresholds from experiments with no measurable spEPSP (N = 28, but each still shows clear optical bAP from the spine and somatic EPSP as in Fig. 3) can be included because actual amplitudes must be less than this number (mean = 10.1 ± 5.6 mV, N = 48 total; light behind dark bars). C, Histogram of the uncaging-evoked spEPSP half-widths (taken from fit curves, N = 19 spines, 1 point at 45 ms omitted; mean = 11.7 ms).
- Figure 5.
Uncaging-evoked EPSP amplitudes in the spine versus the soma. A, Uncaging-evoked spEPSP amplitude versus the somatic soEPSP amplitude for spines with VSD signal above the detection threshold based on noise estimates (2.5 × σnoise, N = 20 spines). Error bars correspond to SD from time-series data, σnoise. B, Stem plot showing the 2.5 × σnoise threshold, which corresponds to the maximum voltage in the spine evoked by glutamate-uncaging, versus the somatic soEPSP for spines where the VSD signal was below the detection threshold (N = 28 spines). C, Histogram of the ratio of spine to somatic EPSP amplitudes. The mean amplitude ratio was 25.3 ± 12.2. D, Ratio of spine to somatic EPSP amplitudes versus distance from the soma.
- Figure 6.
FRAP of an intracellular cytosolic dye was used to estimate Rneck. A, Sample image with spine targeted for photobleaching. Cells from the same population of neurons were filled with the cytosolic dye AlexaFluor 488. B, After a 0.5 ms pulse of 770 nm excitation light, photobleaching of green fluorescence was apparent. Recovery approached control (black trace) and was fit to a single exponential (blue curve). C, Histogram of fit exponential time constant (τ) values from N = 34 spines (mean = 95 ± 11 SE, median 72 ms). D, Conversion to Rneck depends on the spine head volume (Eq. 2). We estimated spine head volume by 3D segmentation of Z-stack spine images (770 nm excitation) into binary images, eroding or dilating these binary 3D images to produce a series of sizes and then convolving with the 3D point spread function of the microscope obtained using 170 nm fluorescent beads. The resultant set of blurred images were then correlated with the experimental image to obtain the best fit. The source binary image which produced the best fit was then used to determine the spine head volume. E, Histogram of Rneck values calculated from FRAP experiments (mean 204 ± 21 SE MΩ), with inset examples of large and small resistance examples.
- Figure 7.
Estimating Rneck using biophysical models. A, A morphologically realistic multicompartmental model of a L5 pyramidal neuron was used (Mainen and Sejnowski, 1996). The passive electrical properties of the model were as follows: Rm = 30,000 Ω cm2, Cm = 0.75 μF cm−2, and Ra = 150 Ω cm. The resulting somatic input resistance was 43 MΩ. B, Basal dendrite region, which was sampled by moving a spine head and neck with passive properties sequentially between 117 different locations to test the role of the spine neck resistance Rneck. C, With the test spine at a single dendritic location, Rneck increases the spEPSP and values depend on the maximal synaptic conductance Gsyn (right side color bar for families of curves). D, At the same time, attenuation ratio of spine–somatic EPSP amplitude appears very linear and insensitive to Gsyn. E, For each Rneck value, EPSPs were generated at all 117 basal dendritic locations, and the spine–soma attenuation ratio was plotted versus the distance of each model site (open circles) along with a quadratic fit (blue line) and superimposed with the experimental attenuation values (N = 19, closed circles with error bars). F, For each experimental attenuation data point (distance, spine–soma attenuation), the model’s Rneck value that led to the quadratic fit (blue line) coinciding with that point was determined and plotted in a histogram (mean 179 ± 25 SE MΩ).
Tables
Analysis variable, Rneck Set N Obs Mean SD Median Minimum Maximum Lower 95% CL for mean Upper 95% CL for mean VSD measurements + model 19 179.4 109.1 210.2 22.9 419.6 132.1 227.7 FRAP experiments 34 204.3 125.3 161.1 51.8 520.6 164.8 247.6
In this issue
