Optical recording of neuronal activity with a genetically-encoded calcium indicator in anesthetized and freely moving mice

Abstract

Fluorescent calcium (Ca(2+)) indicator proteins (FCIPs) are promising tools for functional imaging of cellular activity in living animals. However, they have still not reached their full potential for in vivo imaging of neuronal activity due to limitations in expression levels, dynamic range, and sensitivity for reporting action potentials. Here, we report that viral expression of the ratiometric Ca(2+) sensor yellow cameleon 3.60 (YC3.60) in pyramidal neurons of mouse barrel cortex enables in vivo measurement of neuronal activity with high dynamic range and sensitivity across multiple spatial scales. By combining juxtacellular recordings and two-photon imaging in vitro and in vivo, we demonstrate that YC3.60 can resolve single action potential (AP)-evoked Ca(2+) transients and reliably reports bursts of APs with negligible saturation. Spontaneous and whisker-evoked Ca(2+) transients were detected in individual apical dendrites and somata as well as in local neuronal populations. Moreover, bulk measurements using wide-field imaging or fiber-optics revealed sensory-evoked YC3.60 signals in large areas of the barrel field. Fiber-optic recordings in particular enabled measurements in awake, freely moving mice and revealed complex Ca(2+) dynamics, possibly reflecting different behavior-related brain states. Viral expression of YC3.60 - in combination with various optical techniques - thus opens a multitude of opportunities for functional studies of the neural basis of animal behavior, from dendrites to the levels of local and large-scale neuronal populations.

Keywords: adeno-associated virus; barrel cortex; calcium; neocortex; two-photon microscopy; yellow cameleon.

Figure 1

AAV-delivered expression of YC3.60 in mouse neocortex. (A) A recombinant adeno-associated virus with the human synapsin promoter (P_hSYN) drives expression of YC3.60. (B) Dorsal view of a fixed mouse brain 4 weeks following stereotactic AAV-injection showing YC3.60 expression as a bright YFP fluorescence spot in barrel cortex in the left hemisphere. (C) Confocal image of a coronal section showing the expression pattern surrounding the injection site (asterisk) in a different brain together with NeuN counterstain. Dense YC3.60 labeling of cortical neurons is observed preferentially in L2/3 and L5. (D) A magnified view on the neocortical layers reveals that a large fraction of neurons (red) in L2/3 is also labeled with YC3.60. (E,F) Two-photon images of YFP fluorescence in fixed brain slices showing YC3.60 labeling of somata excluding the nucleus as well as of dendrites in L2/3 and in L5 neurons, respectively.

Figure 2

In vivo imaging of YC3.60-expressing neocortical neurons. (A) In vivo two-photon image stack of YC3.60-expressing neurons in mouse barrel cortex 4 weeks after AAV-injection. Three example YFP images are shown from superficial L1 (86 μm depth from pia, top), upper L2/3 (234 μm depth, middle) and deeper L2/3 (336 μm depth, bottom). Note labeled horizontal dendrites in L1 and dense labeling of somata in L2/3. (B) Vertically resliced side projection of the image stack in (A) showing apical dendrites of pyramidal neurons. Dashed line in the lower image in (A) indicates the section for vertical z-reslicing. Projection extension was about 20 μm. (C) In vivo two-photon image of neuronal YC3.60 expression (green, YFP channel) together with counterstaining of astrocytes with SR101 (red).

Figure 3

Ca²⁺ signals reported by YC3.60. (A) Two-photon line scan measurements of YC3.60 signals in the neuropil of hippocampal slice cultures. Fluorescence changes were evoked by extracellular stimulation at 100 Hz with different number of stimuli, averaged along the line length and over five stimulus presentations. Top: Fluorescence signals in the YFP and CFP channel (smoothed with a 30-ms window), Bottom: YFP/CFP ratio changes expressed as ΔR/R (smoothed with a 100-ms window). Note that intensity increases in the YFP-channel and decreases in the CFP-channel. (B) Example fluorescence transient from an individual YC3.60-expressing hippocampal neuron in vitro following synaptic stimulation, which evoked three action potentials. The Ca²⁺ transient is expressed either as relative percentage change in the YFP channel (ΔF/F_YFP) or as percentage change in the YFP/CFP ratio (ΔR/R). (C) Analogous in vivo example of a spontaneously occurring fluorescence transient in a YC3.60-expressing L2/3 neocortical neuron. In (A–C) raw fluorescence intensities of CFP and YFP are reported in arbitrary units.

Figure 4

YC3.60 sensitivity to action potentials in vitro and in vivo. (A) Characterization of YC3.60 sensitivity in vitro. Top: Cell-attached recording at room temperature from a hippocampal CA3 pyramidal neuron (left image with recording pipette) and simultaneous fluorescence measurement from a somatic region of interest (white box). Extracellular current and single-trial Ca²⁺ transients are shown in response to 1, 2 and 3 APs elicited by synaptic stimulation (arrows; stimulation pipette not shown; stimulus artifacts blanked). Bottom: Mean fluorescence traces (±SEM) in response to 1, 2 and 3 APs (n = 32 traces from 4 cells for 1 AP; n = 40 (5) for 2 APs; n = 35 (5) for 3 APs). (B) Characterization of YC3.60 sensitivity in vivo. Top: Juxtacellular voltage recording from a L2/3 neuron in barrel cortex (left image with recording pipette) and simultaneous two-photon Ca²⁺ measurement from the soma region. Example Ca²⁺ transients are shown for spontaneously occurring 1, 2 and 3 APs. Bottom: Mean fluorescence traces (±SEM) in response to 1, 2 and 3 APs (n = 78 traces from 9 cells for 1 AP; n = 33 (7) for 2 APs; n = 23 (7) for 3 APs). (C) Peak amplitudes of Ca²⁺ transients as a function of number of APs (in vitro, open circles, n = 32, 40, 35 and 4 transients for 1, 2, 3 and 5 APs, respectively; in vivo, filled circles, n = 138 transients from 11 cells, 61 (9), 30 (9), 18 (7), 12 (6), 4 (3), 6 (5) for 1–7 APs, respectively). (D) Decay time constants of exponential fits as a function of number of APs (in vitro, open circles, same n as in (C); in vivo, filled circles n = 78 (9), 33 (7), 23 (7), 16 (6), 11 (5), 3 (2), 5 (5) for 1–7 APs, respectively). Red lines are linear regressions. Error bars are shown as SEM. (E) Efficiency of AP detection in vivo was determined by estimating the distribution of the signal-to-noise ratio (SNR) under noise conditions and fitting with a Gaussian. From the fit, we determined the SNR cutoff at which less than 5% of baseline traces would be classified as false positives (SNR = 1.3). Using this threshold, 71% of single APs (97/137), 80% of doublets (48/60) and 93% of triplets (27/29) were correctly detected.

Figure 5

Single-cell and population YC3.60 Ca²⁺ signals in L2/3 of barrel cortex. (A) Simultaneous two-photon Ca²⁺ imaging in soma and dendrites of L2/3 neurons using vertical (xz-)imaging. Examples of spontaneous somatic (S, red) and apical dendritic (D, blue) YC3.60 Ca²⁺ transients for the cells depicted in the left image. Right: Mean decay times in dendrites compared to somata for 23 measurements (gray lines; mean ± SEM). (B) Simultaneous juxtacellular voltage recording and two-photon Ca²⁺ imaging from a neuron showing rare events of sustained and high-frequency AP firing that are accompanied by large YC3.60 Ca²⁺ transients with peak amplitudes of up to 30% ΔR/R. Top: Sustained AP firing leads to prolonged elevation of the fluorescence ratio. Bottom: A short burst of 11 APs is accompanied by a fast Ca²⁺ transient, which returns to baseline following a stereotypical exponential decay. (C) Two-photon Ca²⁺ imaging of a small population of neurons during sensory stimulation (seven times five air-puffs to contra-lateral whiskers at 5 Hz). Large Ca²⁺ transients in cell 1 (red trace) correlated with the spiking activity observed in the simultaneous juxtacellular voltage recording. Concomitant Ca²⁺ transients were also evoked in neighboring neuronal somata and in the nearby neuropil (NP). The response to the first stimulation episode (dashed box) is shown on expanded scale in the lower left, indicating that YC3.60 resolves the individual steps in the accumulated Ca²⁺ response. (D) Event-triggered average Ca²⁺ traces from somata and adjacent neuropil for spontaneous (n = 37 events of 1–3 APs) and evoked (n = 32 events of 1–5 APs) action potentials. Multi-whisker air puff-evoked Ca²⁺ transients in somata were significantly larger than those in the neuropil while spontaneous spikes were accompanied by somatic but no neuropil transients. Errors are shown as SEM.

Figure 6

Bulk recording of spontaneous and sensory-evoked YC3.60 Ca²⁺ signals in barrel cortex. (A) Large-area Ca²⁺ imaging using two-photon excitation and whisker stimulation by air puffs. Neuropil and two neurons visible about 100 μm deep in a 12-week-old mouse ∼6 weeks after virus infection. The entire frame was taken as region of interest. Bottom: Examples of single-trial spontaneous and sensory-evoked responses are shown along with average traces of 10 trials. (B) Large-area Ca²⁺ imaging using single-photon excitation and a camera and simultaneous local field potential (LFP) recording in barrel cortex of an anesthetized mouse (left schematic). Right: The mean YC3.60 fluorescence signal (ΔF/F in YFP-channel; red traces) correlated well with the LFP for both spontaneous activity (top) and upon air-puff whisker stimulation (bottom; dashed vertical lines). (C) Fiber-optic bulk recording of YC3.60 signals in barrel cortex in an anesthetized mouse (left schematic). Fluorescence excitation and detection were both accomplished through the optical fiber, the tip of which was placed on the cortical surface. Right: Examples of single-trial YC3.60 fluorescence traces (ΔF/F in YFP-channel) and mean of 10 traces for spontaneous activity (top) and upon air-puff whisker stimulation (bottom; dashed vertical lines).

Figure 7

Fiber-optic recording of brain area activity in freely moving mice using YC3.60. (A,B) Two examples of fiber-optic recording of YC3.60 signals in awake, freely moving mice. Bulk Ca²⁺ signals indicating neuronal activity were recorded in somatosensory cortex through a single-core optical fiber as shown in Figure 6C. Fluorescence changes in the YFP-channel are shown during 25–30 s periods together with the position of the mouse in an open field box. Animal behavior (sitting still, moving, touches, or having contact to the wall) is indicated by background colors. The trajectory of the animals’ movement is indicated with selected time stamps. (C) Six more examples of Ca²⁺ imaging from three mice, together with corresponding behavioral observations. Changes of the animal's behavioral state (e.g., start of movement) were frequently associated with marked discontinuities in the fluorescence trace, indicating complex underlying Ca²⁺ dynamics. (D) Control experiment showing that Ca²⁺ signals are blocked by local perfusion of the cortical region with Cd²⁺. (E) Control experiment demonstrating that a flat fluorescence trace is observed in the absence of YC3.60 expression.