Two-Photon Bidirectional Control and Imaging of Neuronal Excitability with High Spatial Resolution In Vivo

Abstract

Sensory information is encoded within the brain in distributed spatiotemporal patterns of neuronal activity. Understanding how these patterns influence behavior requires a method to measure and to bidirectionally perturb with high spatial resolution the activity of the multiple neuronal cell types engaged in sensory processing. Here, we combined two-photon holography to stimulate neurons expressing blue light-sensitive opsins (ChR2 and GtACR2) with two-photon imaging of the red-shifted indicator jRCaMP1a in the mouse neocortex in vivo. We demonstrate efficient control of neural excitability across cell types and layers with holographic stimulation and improved spatial resolution by opsin somatic targeting. Moreover, we performed simultaneous two-photon imaging of jRCaMP1a and bidirectional two-photon manipulation of cellular activity with negligible effect of the imaging beam on opsin excitation. This all-optical approach represents a powerful tool to causally dissect how activity patterns in specified ensembles of neurons determine brain function and animal behavior.

Keywords: digital holography; optogenetics; patterned illumination; two-photon excitation; two-photon imaging.

Graphical abstract

Figure 1

Two-Photon Holographic Stimulation of ChR2-Expressing Cells with Extended Shapes In Vivo (A) Optical setup for holographic illumination. S, laser source; P, Pockels cell; λ/2, half-wave plate; L_1–4, lenses; SLM, spatial light modulator; G, galvanometric mirrors; SL, scan lens; TL, tube lens; D_1–2, dichroic mirrors; PMT, photomultiplier tube; OBJ, objective. (B) Left: intensity profiles in the focal plane (xy, top) and along the axial direction (xz, bottom) of an extended shape illuminating a thin (thickness: ∼150 nm) fluorescent layer and generated with the setup displayed in (A). λ_exc = 920 nm. Right: schematic of the single-cell holographic stimulation paradigm. An elliptical shape was drawn on the soma of opsin-expressing neurons, resulting in an extended illumination volume covering the cell body of the target cell. (C) Two-photon image of a layer 2/3 cortical neuron co-expressing ChR2-mCherry and tdTomato. The cell was targeted for simultaneous holographic stimulation and juxtasomal electrophysiological recording, monitoring the fluorescence of tdTomato in vivo. The dotted lines indicate the recording pipette. (D) Top: electrophysiological trace recorded before, during, and after holographic stimulation (red bar, laser power: 80 mW). Middle: raster plot showing cell response over consecutive trials for the same neuron displayed in the top panel. Bottom: AP distribution for the trials shown in the middle panel (time bin: 100 ms). (E) Firing frequency before (Pre), during (Stim), and after (Post) holographic stimulation of ChR2-expressing layer 2/3 neurons. p = 2E−9, Friedmann test with Dunn’s correction, N = 28 cells from 10 mice. Average laser power: 54 mW, range: 30–92 mW. (F) Firing frequency increase versus displacement of the excitation volume in the radial (top) and axial (bottom) directions during holographic illumination of layer 2/3 cells expressing ChR2-eYFP. Top: N = 9 cells from 6 mice. Bottom: N = 8 cells from 6 mice. Black line represents the average and SEM, and individual experiments are depicted in gray. ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001. See also Figures S1–S3 and Tables S1–S3.

Figure 2

Opsin Somatic Targeting Increases the Spatial Resolution of Holographic Stimulation In Vivo (A) Confocal image of layer 2/3 cells expressing soma-targeted ChR2-eYFP (green). (B) Top: juxtasomal electrophysiological trace recorded Pre, during (Stim), and Post holographic stimulation (red bar, laser power: 30 mW; λ_exc = 920 nm) of a cortical layer 2/3 neuron expressing the soma-targeted ChR2 in vivo. Middle: raster plot showing cell response over consecutive trials for the same cell displayed in the top panel. Bottom: AP distribution for the trials shown in the middle panel (time bin: 100 ms). (C) Average firing frequency Pre, during (Stim), and Post holographic stimulation of layer 2/3 cells expressing the soma-targeted ChR2. p = 1.3E−7, Friedmann test with Dunn’s correction, N = 21 cells from 6 mice. Laser power: 30 mW. (D) Firing frequency increase versus displacement in the radial (top) and axial (bottom) directions during holographic illumination. Top: N = 9 cells from 3 mice. Bottom: N = 11 cells from 3 mice. In this figure, the black line represents the average and SEM, individual experiments are depicted in gray. See also Figure S4 and Tables S1–S3.

Figure 3

Two-Photon Holographic Stimulation across Cell Types and Layers In Vivo (A) Two-photon image of one layer 2/3 SST⁺ interneuron (left) and one layer 2/3 PV⁺ interneuron (middle) expressing ChR2-mCherry (red). One layer 4 Scnn⁺ neuron expressing ChR2-eYFP (green), together with tdTomato (red), is shown on the right. Neurons were recorded in the juxtasomal configuration with a glass pipette (dotted white line) filled with Alexa Fluor 488 (green) in vivo. (B) Top: electrophysiological traces recorded Pre, during (Stim), and Post holographic stimulation (red bar) for one SST⁺ cell (left), one PV⁺ cell (middle), and one Scnn⁺ cell (right). λ_exc = 920 nm. Laser power: 30 mW for SST⁺ and PV⁺ cells and 50 mW for Scnn⁺ neurons. Middle: raster plot showing cell response over consecutive trials for the same neurons displayed in the top panel. Bottom: AP distribution for the trials shown in the middle panel (time bin: 100 ms) for all cell types (SST⁺, left; PV⁺, middle; Scnn⁺, right). (C) Average firing frequency Pre, during (Stim), and Post holographic stimulation of ChR2-expressing layer 2/3 SST⁺ neurons (left), layer 2/3 PV⁺ neurons (middle), and layer 4 Scnn⁺ neurons (right). SST⁺ cells: p = 1.3E−9, Friedmann test with Dunn’s correction, N = 31 cells from 7 mice. PV⁺ cells: p = 1.2E−7, ANOVA test with Bonferroni’s correction, N = 22 cells from 7 mice. Scnn⁺ cells p = 5.2E−7, Friedmann test with Dunn’s correction, N = 19 cells from 8 mice. Laser power: 30 mW for SST⁺ and PV⁺ cells and 50 mW for Scnn⁺ cells. (D) Firing frequency increase versus displacement in the radial (top) and axial (bottom) directions during holographic illumination for layer 2/3 SST⁺ neurons (left), layer 2/3 PV⁺ neurons (middle), and layer 4 Scnn⁺ neurons (right). SST⁺ cells: top, N = 13 cells from 3 mice; bottom, N = 12 cells from 3 mice. PV⁺ cells: top, N = 10 cells from 4 mice; bottom, N = 11 cells from 4 mice. Scnn⁺ cells: top and bottom, N = 11 cells from 5 mice. In this figure, the black line represents the average and SEM, individual experiments are depicted in gray. See also Figure S5 and Tables S1–S3.

Figure 4

Two-Photon Holographic Activation of GtACR2 Allows Optogenetic Inhibition with High Spatial Resolution In Vitro and In Vivo (A) Left: schematic of the experimental configuration for slice recording. Neurons expressing GtACR2 were recorded in voltage-clamp configuration (V_c) and held at −50 mV while holographic illumination with an elliptical shape was performed. Chloride equilibrium potential: −68 mV. Right: two-photon image of a layer 2/3 neuron expressing GtACR2-eGFP that was recorded in voltage-clamp configuration. The glass pipette (dotted lines) was filled with Alexa Fluor 488 (green cytosolic signal). (B) Left: GtACR2-mediated photocurrents evoked by holographic illumination at different illumination powers. Traces are averages of 3 stimulation trials. Right: average peak photocurrent evoked by holographic illumination at various laser powers. λ_exc = 920 nm. In each cell, photocurrent values were normalized to the maximal photocurrent recorded at 30 mW light power. The inset displays the nonlinear dependence of photocurrents on the laser power at low power values. N = 7–11 cells from 2 to 3 mice. (C) Left: traces showing GtACR2 photocurrents evoked by holographic illumination at different wavelengths. Traces are averages of 4 stimulation trials. Average laser power: 25 mW. Right: average GtACR2 peak photocurrent evoked by holographic illumination as a function of the stimulation wavelength. Photocurrent values were normalized to the peak photocurrent at 920 nm. Values were fitted with a three-parameter Weibull function. N = 9 cells from 4 mice. Laser power: 20–30 mW. (D) Top: membrane potential of one layer 2/3 GtACR2-expressing neuron recorded in whole-cell configuration in vivo during simultaneous current injection (gray line below the trace; current amplitude: 70 pA) and two-photon holographic illumination (red bar). Laser power: 80 mW; stimulus duration: 500 ms. Inset: schematic of the experimental configuration in vivo. Middle: raster plot showing cell response over consecutive trials for the same cell displayed in the top panel. Bottom: AP distribution for the trials shown in the middle panel (time bin: 100 ms). (E) Average firing frequency Pre, during (Stim), and Post holographic stimulation of GtACR2-expressing layer 2/3 neurons. Holographic illumination was performed while injecting a small depolarizing current. Average current amplitude: 74 pA, range: 50–100 pA. p = 1.8E−5, Friedmann test with Dunn’s correction, N = 14 cells from 7 mice. Average laser power: 50 mW, range: 10–80 mW. (F) Firing frequency decrease versus displacement of the excitation volume in the radial (top) and axial (bottom) directions during holographic illumination. Top: N = 7 cells from 4 mice. Bottom: N = 8 cells from 5 mice. (G) Same as (D), but for a prolonged period of illumination (10 s). Injected current amplitude: 70 pA; laser power: 80 mW. (H) Same as (E) but for prolonged illumination. Average injected current amplitude: 50 pA, range: 30–70 pA. p = 9.4E−3, ANOVA with Bonferroni’s correction, N = 5, from 2 mice. Laser power: 80 mW. (I) Membrane potential Pre, during (Stim), and Post prolonged inhibition of layer 2/3 neurons with holographic stimulation of GtACR2 (time bins: 2.5 s). p = 7E−6, ANOVA test with Bonferroni’s correction, N = 5, from 2 mice. Laser power: 80 mW. In this figure, the black line represents the average and SEM, individual experiments are depicted in gray. See also Tables S1 and S2.

Figure 5

Scanning with Infrared-Shifted Wavelengths Does Not Modify the Activity of Cells Expressing Blue Light-Sensitive Opsins (A) Two-photon image of layer 2/3 cells expressing the soma-targeted ChR2-eYFP (green) in vivo. One ChR2⁺ neuron was recorded with a glass pipette (dotted white lines) while two-photon raster scanning inside the indicated area (dashed white line) was performed at 11 Hz. (B) Traces recorded in the juxtasomal configuration from one soma-targeted ChR2-expressing neuron during epochs (gray bars) of two-photon raster scanning at wavelength 1,100 nm (top) and wavelength 920 nm (bottom). Laser power: 30 mW in both conditions. (C) Average AP frequency during epochs of spontaneous activity (Spont) and during raster scanning (Scan). Left: results when scanning was performed at λ = 1,100 nm (laser power: 30 and 50 mW). Right: results when scanning was done at λ = 920 nm (laser power: 30 mW). Frame rate: 11 Hz; scanned area: ∼60 × 60 μm² for all experimental conditions. p = 0.47 for λ = 1,100 nm and 30 mW, p = 0.85 for λ = 1,100 nm and 50 mW, p = 5E−4 for λ = 920 nm and 30 mW, Wilcoxon signed rank test, N = 12 FOVs from 3 mice. In this figure, the black line represents the average and SEM, individual experiments are depicted in gray.

Figure 6

Simultaneous Two-Photon Imaging of Red-Shifted Indicator and Two-Photon Holographic Stimulation of Blue-Shifted Excitatory Opsin In Vivo (A) Schematic of the optical setup for simultaneous two-photon imaging (λ_exc = 1,100 nm) and two-photon holographic illumination (λ_exc = 920 nm). S₁, stimulation laser source; S₂, imaging laser source; P_1–2, Pockels cells; G_1–2, galvanometric mirrors; SL, scan lens; TL, tube lens; D_1–3, dichroic mirrors; PMT, photomultiplier tube; OBJ, objective; Hol. Module, holographic module (comprising the SLM, the λ_1/2, and L_1–4 displayed in Figure 1). (B) Calcium transients in a SST⁺ interneuron during simultaneous two-photon imaging (λ = 1,100 nm; laser power: 25 mW; frame rate: 11 Hz; scanned area: ∼90 × 90 μm²) and holographic stimulation (λ = 920 nm; laser power: 50 mW). ΔF/F₀ (gray trace) was smoothed with a moving average filter (black trace). The inset shows one layer 2/3 SST⁺ interneuron co-expressing ChR2-eYFP (green) and jRCaMP1a (red). (C) Two-photon image showing layer 2/3 neurons expressing soma-targeted ChR2-eYFP (green) and jRCaMP1a (red) in vivo. (D) Calcium transients recorded from jRCaMP1a-positive cells (imaging power: 30 mW; frame rate: 11 Hz). The numbers on the left refer to the neurons indicated in (B). Neurons 1–4 (red arrows) were simultaneously stimulated with four elliptical shapes covering the cell somata. Each stimulation episode is indicated by a red bar (stimulation power per cell: ∼50 mW). Periods of stimulation are blanked (see Experimental Procedures). See also Figures S6 and S7.

Figure 7

Simultaneous Two-Photon Imaging and Two-Photon Holographic Inhibition In Vivo (A) Left: image of a layer 2/3 PV⁺ interneuron co-expressing jRCaMP1a and the soma-targeted GtACR2 in vivo. The neuron was imaged and simultaneously recorded in the juxtasomal configuration. The patch pipette containing Alexa 594 is indicated by the dashed lines. Right: trace showing jRCaMP1a fluorescence (top, excitation wavelength: λ = 1,100 nm) and the spiking activity (bottom) for the neuron shown in (A). The cell was illuminated with a two-photon elliptical shape (red bar, λ = 920 nm; stimulation power: 50 mW; stimulus duration: 5 s). (B) Average decrease (ΔF/F₀) (black trace) in jRCaMP1a fluorescence induced by holographic illumination at λ = 920 nm (pink bar) in cells co-expressing jRCaMP1a and the soma-targeted GtACR2. The gray traces represent single experiments. N = 12 cells from 3 mice. Imaging power range: 20–34 mW; stimulation power: 50 mW. (C) Area below the fluorescence trace (left) and fluorescence baseline level (middle) Pre, during (Stim), and Post holographic illumination (excitation wavelength: λ = 920 nm; stimulation power: 50 mW) in cells co-expressing jRCaMP1a and the soma-targeted GtACR2. Average area: p = 4E−7, ANOVA test with Bonferroni correction, N = 12 cells from 3 mice; average fluorescence: p = 1E−8, ANOVA test with Bonferroni correction, N = 12 from 3 mice. (D) Firing frequency of PV⁺ interneurons Pre, during (Stim), and Post holographic illumination. Excitation wavelength: λ = 920 nm; stimulation power: 50 mW; stimulus duration: 5 s. p = 8E−6, ANOVA test with Bonferroni correction, N = 12 stimulation trials from 3 neurons. In this figure, the black line represents the average and SEM, individual experiments are depicted in gray.