Luminopsins integrate opto- and chemogenetics by using physical and biological light sources for opsin activation

Abstract

Luminopsins are fusion proteins of luciferase and opsin that allow interrogation of neuronal circuits at different temporal and spatial resolutions by choosing either extrinsic physical or intrinsic biological light for its activation. Building on previous development of fusions of wild-type Gaussia luciferase with channelrhodopsin, here we expanded the utility of luminopsins by fusing bright Gaussia luciferase variants with either channelrhodopsin to excite neurons (luminescent opsin, LMO) or a proton pump to inhibit neurons (inhibitory LMO, iLMO). These improved LMOs could reliably activate or silence neurons in vitro and in vivo. Expression of the improved LMO in hippocampal circuits not only enabled mapping of synaptic activation of CA1 neurons with fine spatiotemporal resolution but also could drive rhythmic circuit excitation over a large spatiotemporal scale. Furthermore, virus-mediated expression of either LMO or iLMO in the substantia nigra in vivo produced not only the expected bidirectional control of single unit activity but also opposing effects on circling behavior in response to systemic injection of a luciferase substrate. Thus, although preserving the ability to be activated by external light sources, LMOs expand the use of optogenetics by making the same opsins accessible to noninvasive, chemogenetic control, thereby allowing the same probe to manipulate neuronal activity over a range of spatial and temporal scales.

Keywords: bioluminescence; hippocampus; luciferase; neural circuitry; substantia nigra.

Fig. 1.

Determining minimum light intensity for induction of action potential for Volvox channelrhodopsin 1 and comparison of light emission from variants of GLuc. (A) A cortical neuron from a rat embryo in culture expressing Volvox channelrhodopsin 1 (VChR1) was whole-cell current-clamped and light pulses (1 s, 480 nm) at different intensities were delivered, resulting in sub- and suprathreshold depolarizations. (B) Frequency of action potential firing was quantified as a function of light intensity. Fitting with the Hill equation resulted in maximum firing frequency of 11.8 ± 0.3 Hz, half-maximum light intensity of 56.7 ± 5.1 μW/mm², and the Hill coefficient of 2.6 ± 0.6. n = 8. Error bars indicate SEM in this and subsequent figures. (C) HEK cells were transfected with wild-type GLuc, super luminescent (slGLuc), and slow burn (sbGLuc) variants, and with NanoLuc luciferase all fused with VChR1, and were challenged with different concentrations of CTZ to obtain dose–response curves for bioluminescence using a plate reader. The data were fitted with the Hill equations. Hill coefficients were set to 2. n = 3 wells. (D) sbGLuc showed a ∼10-fold increase in bioluminescence compared with wild-type GLuc at a 50 µM CTZ concentration. Luminescence from Renilla luciferase and from NanoLuc with its own substrate, furimazine, was also plotted. The same dataset obtained from HEK cells shown in C was replotted. n = 3 wells.

Fig. 2.

Luciferases with higher light emission improve coupling efficiency. (A) Representative CTZ-induced currents with LMO1 (GLuc-ChR2; Top), LMO2 (GLuc-VChR1; Middle), and LMO3 (sbGLuc-VChR1; Bottom) in transfected HEK cells (sampling rate, 10 KHz). Bioluminescence was obtained simultaneously but at a much lower sampling rate (0.2 Hz) and is shown in the same scale for the three examples (cyan). (B) For all LMOs, the response to CTZ correlated with the maximum photocurrent elicited by direct illumination. However, for a given value of maximum photocurrent, responses to CTZ were higher for cells expressing LMO2 than for cells expressing LMO1 and substantially higher for cells expressing LMO3. n = 3–4 cells for each data point from 12 (LMO1; HEK cells), 10 (LMO2; HEK cells), and 14 (LMO3; neurons in culture) cells in total. The correlation between photocurrent and CTZ-induced current was also examined in the brain slice preparation (open symbols in red; n = 3 and 4 cells). (C) The efficiency of coupling between the GLuc variants and the channelrhodopsins was calculated from the same dataset shown in B by dividing the amplitude of the CTZ-induced current by that of the maximum photocurrent induced by direct illumination of the LMOs. The coupling efficiency of LMO3 was significantly higher than that of the first two LMOs. n = 12 (LMO1), 10 (LMO2), and 21 (LMO3). The coupling efficiency of LMO3 determined in neurons in culture and in brain slice preparation was not significantly different (P = 0.87; two-tailed unpaired Student’s t test; n = 14 and 7 for culture and slice preparation, respectively) and thus pooled together.

Fig. 3.

Bioluminescence controls action potential firing. (A) An LMO3-expressing neuron showed bioluminescence upon CTZ application (100 μM). (B) CTZ-induced bioluminescence (blue trace) in the LMO3-expressing neuron shown in A caused depolarization and action potential firing (black trace). (C) An iLMO-expressing neuron showed bioluminescence upon CTZ application (100 μM). (D) CTZ-induced bioluminescence (blue trace) in an iLMO-expressing neuron caused slight hyperpolarization and inhibition of action potential firing (black trace). Action potentials were induced by peri-threshold depolarizing currents (60 pA, 15 ms) at 1 Hz. (E) Mean membrane potential change caused by CTZ application in neurons expressing LMO2, LMO3, and iLMO. n = 8, 4, and 4, respectively.

Fig. S1.

Characterization of Mac photocurrent. (A) Neurons were transfected with Mac-EGFP and photocurrent was elicited with various wavelengths at different intensities under voltage clamp. Mean currents were fitted with the Hill equations with the following parameters: 25.5 pA, 0.75, and 12.9 mW/mm² with 480 nm; 24.3 pA, 0.85, and 5.0 mW/mm² with 540 nm; 20.6 pA, 1.14, and 3.7 mW/mm² with 560 nm; 17.8 pA, 0.97, and 3.2 mW/mm² with 575 nm for the maximal photocurrent, the Hill coefficient, and the half maximal irradiance, respectively. n = 3. Error bars denote SEM in this and the subsequent panels. (B and C) The data from each cell were individually fitted with the Hill equation, and means for the maximal photocurrent (B) and for the half maximal irradiance (C) were calculated. The same dataset was used as in A.

Fig. 4.

Bioluminescence elicits postsynaptic currents in vitro. (A) Cortical neurons from rat embryos in culture were transduced by lentivirus carrying the LMO3 gene, resulting in EYFP tag expression (yellow). A negative cell in the center was whole-cell patch clamped and filled with Texas Red dye (shown in red), whereas neighboring cells were activated by focal photostimulation with an arc lamp depicted as a cyan circle (not scaled; the actual spot size in wide-field illumination was 670 μm in diameter). (B) A light pulse (1 s; 480 nm; 1 mW/mm²; 670 µm in diameter) elicited fast inward currents in the cell under voltage clamp at –60 mV. The internal solution contained 12.5 mM chloride to keep the equilibrium potential of chloride close to the holding potential. (C) In another cell, fast inward currents elicited by photostimulation were blocked by a mixture of ionotropic glutamate antagonist, CNQX (100 µM) and kinurenic acid (KA; 3 mM), indicative of glutamatergic EPSCs. (D) CTZ (100 µM; 470 µL) was bath-applied to the same coverslip shown in A, causing widespread bioluminescence. Bioluminescence was expected in the entire coverslip (9 mm in diameter) depicted as a cyan rectangle (not scaled). (E) CTZ-induced bioluminescence (blue trace, Top) caused fast inward currents (black trace, Middle) while it lasted. EPSC frequency was quantified in 10-s bins (Bottom). All traces are shown in the same time scale. (F) Mean frequency of spontaneous EPSCs before stimulation (control), during stimulation by lamp, or by CTZ (*P < 0.05; paired Student’s t test; n = 8 for both lamp and CTZ). (G) Averaged waveforms of EPSCs recorded from the cell shown in A, B, and E during photostimulation by lamp and by bioluminescence with CTZ. n = 24 (lamp) and 101 (CTZ).

Fig. S2.

Light responses in LMO3-expressing neurons in the substantia nigra ex vivo. (A) Hyperpolarization in the membrane potential (V_m) by current injection (I_m), showing minimal contribution of I_h. Rebound action potential firing was truncated for clarity. (B) Photocurrent in response to a 2-s light pulse. (C) Optical mapping of light-induced voltage responses in a SN cell activated by a 559-nm laser; pixel color scale (at right) indicates size of light responses, with locations where laser stimulation evoked action potentials indicated in orange. (D) Representative light responses evoked in locations 1–3 in C.

Fig. S3.

Bioluminescence-induced responses in a SN cell. (A) Responses to a 2-s light pulse. Fast outward currents resembling IPSCs were superimposed on an inward photocurrent due to LMO3 expressed by the cell. (Inset) IPSCs are shown on an expanded time scale. (B) Bioluminescence (Top) and membrane current (Middle) simultaneously recorded from an SN cell. CTZ induced a slow inward current (red dashed line) and superimposed fast outward currents resembling IPSCs. Holding potential was –60 mV. (Inset) CTZ-induced IPSCs on an expanded time scale.

Fig. 5.

Bioluminescence changes in vivo single unit activity bidirectionally in awake behaving mice. (A) Firing rate of a representative neuron in the SNr that increased firing after CTZ in a mouse expressing LMO3, with its spike waveform on the right. (B) LMO3 population summary: average firing rates before and after injection [**P < 0.01; two-way ANOVA followed by post hoc test; n = 30 (vehicle) and 31 (CTZ) cells from 5 mice]. Error bars represent SEM. (C) Firing rate of a representative SNr neuron that decreased firing after CTZ in a mouse expressing iLMO, with its spike waveform on the right. (D) iLMO population summary: average firing rates before and after injection [***P < 0.001; two-way ANOVA followed by post hoc test; n = 36 (vehicle) and 46 (CTZ) cells from five mice]. (E) Firing rates of two representative neurons over 2 h following injection of CTZ (1 for LMO3 and 1 for iLMO). Shaded area indicates the estimated period during which CTZ remains effective in altering firing rates.

Fig. 6.

Bioluminescence modulates behavior in awake freely moving mice. (A) Representative trajectories of mouse bodies showing effects of intra-SNr injection of CTZ on turning behavior in mice expressing LMO unilaterally in the SNr. Unilateral excitation of nigral neurons produced more ipsiversive turning (LMO3), whereas unilateral inhibition produced more contraversive turning (iLMO), consistent with well-established observations of nigral functions. (B) In mice expressing LMO3, injection of CTZ into the SNr (n = 3) or the tail vein (n = 5) increased ipsiversive turning (*P = 0.0021 and *P = 0.04, respectively; two-way ANOVA followed by post hoc test). (C) In mice expressing iLMO, injection of CTZ into the SNr (n = 3) or the tail vein (n = 8) increased contraversive turning (*P = 0.016 and *P = 0.0003, respectively; two-way ANOVA followed by post hoc test).

Fig. 7.

Bimodal opto- and chemogenetic interrogation of neurons expressing luminopsin. (A–G) Hippocampal dentate gyrus and CA3 neurons were transduced by AAV-LMO3. (A–C) CTZ-induced bioluminescence and EPSCs in a CA1 pyramidal cell. CTZ treatment elicited bioluminescence (A), as well as fast inward currents, presumably EPSCs. B, Inset shows inward currents on a faster time scale. (C) Time course of CTZ-induced increase in EPSC frequency. (D–G) Mapping local excitatory inputs of a CA1 pyramidal cell. (D) Image shows fluorescence from the EYFP tag (green) overlaid on a bright-field image of a hippocampal slice. (E) CA1 pyramidal cell was whole-cell patch clamped with a pipette filled with Texas Red dye, with dye fluorescence revealing morphology of the neuron. (F, Left) Map of excitatory inputs activated by brief laser light spots scanned throughout the part of the CA1 region shown in E. Pseudo color scale (lower left) indicates amplitudes of EPSCs evoked when laser spot was positioned at each location, whereas the red image illustrates cell morphology (from E). (F, Right) Representative EPSCs evoked at locations indicated by corresponding numbers in map at left. (G) Light flash (blue) evoked EPSCs (black) that were blocked by treatment with the glutamate receptor antagonists CNQX and d-AP5 (red).