Co-expression of VAL- and TMT-opsins uncovers ancient photosensory interneurons and motorneurons in the vertebrate brain

Abstract

The functional principle of the vertebrate brain is often paralleled to a computer: information collected by dedicated devices is processed and integrated by interneuron circuits and leads to output. However, inter- and motorneurons present in today's vertebrate brains are thought to derive from neurons that combined sensory, integration, and motor function. Consistently, sensory inter-motorneurons have been found in the simple nerve nets of cnidarians, animals at the base of the evolutionary lineage. We show that light-sensory motorneurons and light-sensory interneurons are also present in the brains of vertebrates, challenging the paradigm that information processing and output circuitry in the central brain is shielded from direct environmental influences. We investigated two groups of nonvisual photopigments, VAL- and TMT-Opsins, in zebrafish and medaka fish; two teleost species from distinct habitats separated by over 300 million years of evolution. TMT-Opsin subclasses are specifically expressed not only in hypothalamic and thalamic deep brain photoreceptors, but also in interneurons and motorneurons with no known photoreceptive function, such as the typeXIV interneurons of the fish optic tectum. We further show that TMT-Opsins and Encephalopsin render neuronal cells light-sensitive. TMT-Opsins preferentially respond to blue light relative to rhodopsin, with subclass-specific response kinetics. We discovered that tmt-opsins co-express with val-opsins, known green light receptors, in distinct inter- and motorneurons. Finally, we show by electrophysiological recordings on isolated adult tectal slices that interneurons in the position of typeXIV neurons respond to light. Our work supports "sensory-inter-motorneurons" as ancient units for brain evolution. It also reveals that vertebrate inter- and motorneurons are endowed with an evolutionarily ancient, complex light-sensory ability that could be used to detect changes in ambient light spectra, possibly providing the endogenous equivalent to an optogenetic machinery.

Figure 1. Phylogenetic and sequence analyses of TMT-Opsins and Encephalopsins.

(A) Maximum likelihood (ML) and neighbor joining (NJ) trees group TMT-Opsins into three distinct subclasses conserved across vertebrates with high branch support. Encephalopsins form a fourth, closely related group. The topology of the NJ tree is shown, and support values are given as NJ/ML next to critical branches. Grey box, ciliary-type opsins; yellow, TMT-Opsins; blue, Encephalopsins. (B, C, D, F) Conserved sequence stretches in the c-terminus of the indicated opsin subfamilies. Numbers on x-axis refer to the amino acid positions in bovine rhodopsin (B, C, D, F) or human Encephalopsin (G). (E) Comparative analysis of characteristic opsin sequence features critical for photopigment function. Bovine rhodopsin is used as reference for counterion position. (G) Conserved sequence stretch in the c-terminus of Encephalopsins.

Figure 2. TMT-Opsins and Encephalopsin are functional light receptors.

(A) Example traces of Neuro2A (N2A) cell stimulation by two consecutive light pulses of 12 min and 10 min. Data are presented as means (N = 4). (B–D) Traces of N2A cells transfected with medaka tmt-opsins (red) versus mutated (L294A) tmt-opsins (black), 10 min light stimulation. Data are presented as mean ± SEM (grey lines) (N = 4). (E) Different kinetics of TMT-2 (blue) compared to TMT-1B and TMT-3A in N2A cells (see also Figure S3). Baseline normalized CI values were normalized to the maximum and data are presented as means (N = 4). (F) Quantification of opsin-dependent N2A cell responses to light. Relative light responses are displayed as mean ± SEM (N = 72–136). (G) Quantification of opsin-dependent N2A cell responses to different spectra compared to human rhodopsin. Data represent mean ± SEM (N = 12–36). (H) HEK cells transfected with medaka encephalopsin (red) versus Schiff base mutant version (black). Data represent mean ± SEM (grey lines) (N = 32). (I) Quantification of Encephalopsin-dependent cell responses in HEK and N2A cells. Data represent mean ± SEM (N = 44–120); **** p<0.0001; *** p<0.0005; ns, not significant; yellow background box, light stimulation. See Figures S4 and S13 for analyses details.

Figure 3. TMT-Opsin expression in inter- and motorneuron nuclei is maintained from larvae to adult stages.

(Left topmost panel) Dorsal view of a schematized medaka larva; red boxes indicate positions of sections displayed below. (Right topmost panel) lateral view of an adult medaka brain, transversal planes corresponding to sections below. ISH on 7 dpf larvae (A, C, E, G) and coronal sections of the adult brain (B, D, F, H). Magnifications of boxed areas on the right; corresponding expression domains in larvae are indicated with arrowheads. Scale bars, 50 µm. Expression domains: tmtops1b, granular layer of the olfactory bulb (A, B); tmtops3a, semicircular torus (C, D); tmtops2, dorsal tegmental nucleus (E, F), facial nerve nucleus of the hindbrain (G, H).

Figure 4. TMT-Opsin 1B is expressed in inter- and motorneuron nuclei on mRNA and protein level.

ISH (A, C) and immunohistochemistry (B, D, E, F) of TMT-Opsin 1b on coronal adult medaka brain sections. Magnification of black boxes in insets. Scale bars, 50 µm. (A) mRNA expression of tmtops1b in the dorsal tegmental nucleus. (B) Protein expression of TMTopsin1b in cells of the dorsal tegmental nucleus, the same area as in (A). Arrowheads indicate TMTopsin1b+ cells. (C) Multiple domains of mRNA expression of tmtops1b in the hindbrain. (D) TMTopsin1b+ cells localize to sites of mRNA expression indicated by a yellow box in (C). (E) Overview of TMTopsin1b protein expression in the hindbrain. Arrows and asterisks indicate protein expression domains that correspond to mRNA expression in (C). (F) Magnification of box in (E), z-stack: 13.87 µm. Note the projections (arrowheads) extending from a TMTopsin1b+ cell verifying its neuronal nature.

Figure 5. Co-expression of tmt-opsins with choline acetyltransferase in distinct inter- and motorneurons.

Two-color ISH of tmtops1b (blue) and chat1/2 (red/red fluorescence) on coronal adult medaka brain sections. Transversal planes of (A), (D), and (G) correspond to planes in Figure 3H, D, and F, respectively. (B, E, H) Magnification of boxed areas. (C, F, I) Fluorescent images of chat1/2 staining. Arrowheads, co-expressing cells. Scale bars, 50 µm. Co-staining in facial nerve motorneurons (A–C), in interneurons of the periventricular grey zone of the tectum (D–F), and in interneurons of the rostral tegmental nucleus (G–I). Note that fluorescent signal of chat1/2 expression can be quenched in areas of strong tmt-opsin staining (for higher magnification see Figure S10).

Figure 6. A subset of tectal interneurons are intrinsically light sensitive.

(A) Confocal ISH images of anti-ChAT staining (green) in the tectum of coronal whole-brain slices. Z-stack, 25.55 µm (B) Magnification of boxed area in (A). Z-stack, 25.55 µm; arrowheads, ChAT-positive interneurons; arrows, neurites projecting from ChAT-expressing interneurons. (C) Representative membrane potential traces from two tectal interneurons recorded under total darkness. (D) Membrane potential changes of interneurons recorded under darkness (N = 14). Values were calculated in relation to the reference point (0 min). No significant difference can be observed between t = −1 and t = +1. Example traces in (C) of one spiking (triangle) and one nonspiking (circle) neuron highlighted in red. Note that delta membrane changes of cells recorded under darkness never get close to 10 mV (dashed line). (E) Representative membrane potential traces from single non-light-responsive interneurons exposed to light (yellow box). (F) Membrane potential changes of interneurons exposed to 1 min light (N = 28). Example traces in (E) of one intrinsically spiking (triangle) and a nonspiking (circle) neuron that do not respond to light are highlighted in blue. Light-responsive (circle) and intrinsically spiking light-responsive (triangle) interneurons in (G) are marked in green. A significant difference can be observed between t = −1 and t = +1. (G) Representative membrane potential traces from single light-responsive interneurons exposed to light. The p values were assessed by paired Student's t test.

Figure 7. Tmt- and val-opsin co-expression domains in inter- and motorneuron cluster are conserved in vertebrates.

Two-color ISH of tmt-opsin1b (blue) and val-opsins (red/red fluorescence) on coronal medaka (A–F) and zebrafish (G–L) sections. Magnifications are indicated as boxes. Scale bars, 50 µm. Tmtops1b and valopb co-expression in the central posterior thalamic nucleus of medaka (A–B) and zebrafish (G–H), in the dorsal tegmental nucleus in medaka (C) and zebrafish (I) and in the facial nerve nucleus in medaka (D–F) and zebrafish (J–L).