High frequency neural spiking and auditory signaling by ultrafast red-shifted optogenetics

Abstract

Optogenetics revolutionizes basic research in neuroscience and cell biology and bears potential for medical applications. We develop mutants leading to a unifying concept for the construction of various channelrhodopsins with fast closing kinetics. Due to different absorption maxima these channelrhodopsins allow fast neural photoactivation over the whole range of the visible spectrum. We focus our functional analysis on the fast-switching, red light-activated Chrimson variants, because red light has lower light scattering and marginal phototoxicity in tissues. We show paradigmatically for neurons of the cerebral cortex and the auditory nerve that the fast Chrimson mutants enable neural stimulation with firing frequencies of several hundred Hz. They drive spiking at high rates and temporal fidelity with low thresholds for stimulus intensity and duration. Optical cochlear implants restore auditory nerve activity in deaf mice. This demonstrates that the mutants facilitate neuroscience research and future medical applications such as hearing restoration.

Fig. 1

Channelrhodopsin mutants with accelerated closing kinetics. a Helix F and helix C of channelrhodopsin. Residues changing the off-kinetics are highlighted (ChR2 numbering). b ClustalW alignment ot the helix F of ChR2, Chrimson, ReaChR and VChR1. Colored boxes show the channelrhodopsin mutants. c–h NG cells heterologously expressing channelrhodopsin variants were investigated by whole-cell patch-clamp experiments at a membrane potential of −60 mV. Typical photocurrents of ChR2-EYFP (black trace), ChR2-EYFP F219Y (red trace) (c), VChR1-EYFP (black trace), VChR1-EYFP F214Y (red trace) (d), ReaChR-Citrine (black trace), ReaChR-Citrine F259Y (red trace) (e), Chrimson-EYFP (black trace) and Chrimson-EYFP Y261F (red trace) (f) immediately after cessation of 0.5 s illumination at a saturating light intensity of 23 mW/mm^-2 and a wavelength of c λ = 473 nm, d λ = 532 nm, e λ = 532 nm and f λ = 594 nm. g Typical photocurrents of Chrimson-EYFP mutants, which were measured in response to 3 ms light-pulses (23 mW/mm², λ = 594 nm). h For clear illustration solely the photocurrents of Chrimson-EYFP, Chrimson-EYFP Y261F/S267M (f-Chrimson-EYFP) and Chrimson-EYFP K176R/Y261F/S267M (vf-Chrimson-EYFP) are shown. Photocurrents were normalized for comparison. Scale bars: c 10 ms, d, f 30 ms, e 100 ms, g, h 20 ms

Fig. 2

Light-induced spiking in rat hippocampal neurons. a–d Spiking traces at different light-pulse frequencies. Rat hippocampal neurons heterologously expressing Chrimson-EYFP (a), Chrimson-EYFP K176R/Y261F/S267M (vf-Chrimson-EYFP) (b) and Chrimson-EYFP Y261F/S267M (f-Chrimson-EYFP) (c, d) were investigated by whole cell patch-clamp experiments under current-clamp conditions (λ = 594 nm, pulse width = 3 ms, saturating intensity of 11–30 mW/mm²). c Traces from two different cells at a stimulation frequency of 60 Hz. d Traces from one cell at stimulation frequencies of 80 Hz and 100 Hz. e–g The dependence of spike probability on light pulse intensity for Chrimson-EYFP (e) (15 different cells), Chrimson-EYFP Y261F/S267M (f-Chrimson-EYFP) (f) (15 different cells) and Chrimson-EYFP K176R/Y261F/S267M (vf-Chrimson-EYFP) (g) (15 different cells). The action potentials were triggered by 40 pulses (λ = 594 nm, pulse width = 3 ms, ν = 10 Hz) of indicated light intensities. In order to determine the spike probability, the number of light-triggered spikes was divided by the total number of light pulses. Scale bars: y-axis: 10 mV, time-axis: (a, b, 10 Hz) 500 ms (a, b, 20 Hz) 300 ms (a, b, 40 Hz) 200 ms (c, 60 Hz) 100 ms (d, 80 Hz) 70 ms (d, 100 Hz) 50 ms

Fig. 3

vf-Chrimson drives fast interneurons to the frequency limit. a Example recording of a neocortical parvalbumin-positive interneuron in an acute brain slice. Current injection (500 ms, 550 pA) elicits high frequency firing (322 Hz), consistent with the fast spiking phenotype of these interneurons. b When tested with constant current injection, the input–output curve of PV-interneurons plateaus at a maximum firing rate of 270 ± 33 Hz (n = 8). c Example traces of the vf Chrimson-expressing PV-interneuron from a activated by light pulses (565 nm, 0.5 ms) at frequencies ranging from 50–500 Hz. Note that this interneuron reliably followed frequencies of up to 400 Hz. d Spiking probabilities of PV-interneurons at different optical stimulation frequencies. On average (black), PV-interneurons followed stimulation up to 300 Hz reliably (94 ± 5% spiking probability), and could still encode input frequencies of up to 400 Hz with a reliability of 68 ± 16% (n = 7; three whole-cell, four cell-attached recordings). e Action potential latency (assessed at peak) and action potential jitter (s.d. of latencies) after light pulse onset for all stimulation frequencies with reliable spiking (>85%). Error bars are s.e.m. Scale bars: a 50 ms, 10 mV c 50 ms, 10 mV

Fig. 4

f-Chrimson expression after postnatal AAV-transduction of SGNs. a Scheme of the future oCI as implanted into the human ear: the oCI passes through the middle ear (limited left by ear drum and right by inner ear) near the ossicles, enters the cochlea and spirals up in scala tympani. It will likely contain tens of microscale emitters (orange spots on oCI) that stimulate (orange beams) SGNs housed in the modiolus (central compartment of the cochlea), that encode information as APs. SGNs form the auditory nerve (right) which carries the information to the brain (not displayed). b pAAV vector used in the study to express f-Chrimson-EYFP under the control of the hSynapsin promoter (top) upon early postnatal injection of AAV2/6 into scala tympani via a posterior tympanotomy (lower left) to expose the round window (white circle in right lower panel). c Photocurrents of a representative culture f-Chrimson-EYFP-positive SGN isolated from an injected ear at postnatal day 14. Light pulses of 2 ms duration were applied at the indicated intensities in the focal plane and photocurrents recorded at −73 mV at room temperature. Scale bar: 2 ms, 50 pA. d Fraction of EYFP-positive (EYFP⁺) SGNs (identified by parvalbumin immunofluorescence, parvalbumin⁺) and e density of parvalbumin⁺ SGNs (#cells per 10⁴ µm²) obtained from data as in f. Symbols mark results from individual animals (n = 5), box–whisker plots show 10th, 25th, 50th, 75th and 90th percentiles of the injected (orange) and non-injected control (magenta) cochleae (Kruskal–Wallis ANOVA, P = 0.6538, H = 0.98; post-hoc Dunn’s test for comparison of expression, P > 0.05 for all pairwise comparisons; Mann–Whitney U test for comparison of density, L_apex vs. R_apex, L_mid vs. R_mid, L_base vs. R_base, P > 0.05 for all comparisons). f Projections of confocal cryosections with YFP (green) and parvalbumin (magenta) immunofluorescence of SGNs in three cochlear regions (scale bar: 50 µm). Insets (scale bar: 10 µm) show close-up images of single z-sections of the same images

Fig. 5

Single-channel oCIs drive oABRs in hearing and deaf mice. a Experimental set-up for oABR-recordings in mice: a 50 µm optical fiber coupled to a 594 nm Obis laser was implanted into scala tympani via a posterior tympanotomy and the round window. Recordings of far-field optically evoked potentials were performed by intradermal needle electrodes. For aABR recordings a free-field speaker was employed (lower panel). b Comparing oABRs (upper panel) and aABRs (lower panel) at strong stimulation levels for four mice (average of 1000 trials). oABRs were recorded in response to 1 ms long, 11 mW, 594 nm laser pulse at 10 Hz, aABRs of the same mice in response to 80 dB (SPL peak equivalent) clicks. Bars indicate the stimulus timing. c oABRs (upper panel, 594 nm, 1 ms at 10 s^-1) and aABRs (lower panel, clicks at 10 s^-1, values in SPL [peak equivalent]) recorded from an exemplary AAV-injected mouse at increasing stimulus intensities. d–f Normalized P1-N1-amplitude as a function of laser intensity (d 1 ms at 20 Hz), pulse duration (e 11 mW at 20 Hz), and stimulus rate (f 11 mW, 1 ms). Group average (lines) and s.d. (error bars) are shown in orange (same for g–i). g–i P1-latency as a function of laser intensity (g as in d), duration (h as in e), and rate (i as in f). j Exemplary aABR recordings done as in a–c using a 9 months-old mouse (following postnatal AAV-Chrimson-EYFP injection: elevated acoustic thresholds (around 60 dB [SPL], compare to c). k oABR recordings done as in a–c in the same mouse as in j, using 1 ms long laser pulses: thresholds similar to injected mice at 2–3 months of age (around 1 mW, compare to c). l P1-N1-amplitude of oABR (orange) and P1-N1-amplitude of aABR (gray) as function of stimulus intensity in young (2–3 months-old) and old (9 months-old) mice (n = 5 for each group, means (lines) ± s.e.m. (error bars) are shown. Symbols in d–i mark results from individual animals. Scale bars (b, c, j, k): 1 ms, 5 µV

Fig. 6

f-Chrimson enables SGNs spiking at near physiological rates. a Experimental set-up for recording optogenetic responses of SGNs in mice: a 50 µm optical fiber coupled to a 594 nm laser was implanted into scala tympani via the round window (lower panel, see cylindrical structure in the upper half) and microelectrodes were advanced into the cochlear nucleus via a craniotomy (upper panel). b Exemplary spikes of a neuron (1 ms, 5.5 mW for 100, 300 Hz; 11 mW for 500 Hz). Raster plot (right panel): spike times in response to laser pulses (orange bars: 2 ms @5.5 mW for 20-400 Hz, 1 ms @11 mW for 500-700 Hz and above: 0.5 ms @11 mW): spikes cluster in time for stimulus rates up to hundreds of Hz, temporal jitter increases with stimulation rates. Scale bar: 50 ms, 2 mV. c Activity of an exemplary neuron in response to 900 ms trains of laser pulses (1 ms) at three different rates leaving an inter-train recovery time of 100 ms (first 400 ms are shown and analyzed). Panels to the right side of raster plots show polar plots: synchronicity and probability of firing decay with increasing stimulus frequency. Spike probability 200 Hz: 0.8, 300 Hz: 0.33, 400 Hz: 0.04. Vector strength 200 Hz: 0.92, 300 Hz: 0.83, 400 Hz: 0.57 (Rayleigh-test: P < 0.001 in all cases). d Box-whisker plots showing 10th, 25th, 50th, 75th and 90th percentiles of the vector strength (orange) and spike probability (purple) of 40 units from five mice, stimulated at different rates as described for c. Symbols represent values from every unit. Gray circles are means of vector strength of SGNs in wild-type mice found with transposed tones at the characteristic frequency at 30 dB relative to spike threshold, for comparison. Numbers at the bottom of the graph indicate number of units clustered below them. e Temporal jitter of spikes across stimulation rates 50–400 Hz. Gray area represents the hazard function obtained in response to simulated Poisson spike trains. Data points show mean (lines) ± s.e.m. (error bars). Number of units included for each stimulation frequency (color coded) is shown