Synaptic Gpr85 Influences Cerebellar-Granule-Cell Electrical Properties and Light-Induced Behavior in Zebrafish

Characterization of newly generated knock-in zebrafish gpr85 reporter lines

Firstly, to define the tissues expressing gpr85 in zebrafish embryos, we performed WISH at different developmental stages from 1 to 6 dpf. The signal emerged at 1 dpf in the telencephalic and ventral midbrain regions, thereafter expanding to the entire brain by 6 dpf (Fig. S1A). Transcripts were also observed in the retina of 3 dpf zebrafish larvae (Fig. S1B).

To better define gpr85 expression, we generated a zebrafish reporter line using CRISPR/Cas9 technology with targeted integration repair (Wierson et al., 2020; Jordan et al., 2021). The GAL4-VP16 open reading frame was inserted downstream of the endogenous gpr85 start codon, creating a loss-of-function allele (Fig. S1C). The correct insertion of the GAL4 cassette at the gpr85 locus was validated using PCR genotyping (Fig. S1D) and Sanger sequencing. We then introduced the gpr85^GAL4 allele into either the Tg(UAS:GFP-CAAX)^m1230 or Tg(UAS:LifeAct-eGFP)^mu271 backgrounds. Phenotypic observations of these two lines did not reveal any developmental or adult behavioral defects in the heterozygous gpr85^GAL4/+ state. Adult zebrafish were fertile, and their offspring followed the expected Mendelian inheritance ratios. All observations described hereafter were consistent across both lines.

We first characterized the general expression profile of the GFP-CAAX reporter during the embryonic development of gpr85^GAL4/+, Tg(UAS:GFP-CAAX) individuals. Whole-body imaging of live embryos revealed that the GFP signal emerged at 1 dpf in the telencephalic and ventral midbrain regions (Fig. 1A,B). At 2 dpf, the expression expanded to include forebrain and midbrain (Fig. 1C), persisting through the end of embryogenesis (3 dpf; Fig. 1D) and into postembryonic stages (6 dpf; Fig. 1E). Consistent with WISH experiments, GFP was detected in the retina (Fig. S1E–H″). Immunodetection on coronal sections of 3 dpf embryos (Fig. S1E, Fʺ) and 6 dpf (Fig. S1G,Hʺ) larvae, revealed GFP-positive (GFP⁺) cells in the ganglion cell layer (GCL) and the inner nuclear layer (INL), but not in the photoreceptor layer (Fig. S1F,H). In contrast to the less sensitive WISH method (Fig. S1A,B), we also observed GFP reporter signal in both the ventral and dorsal regions of the developing spinal cord of 3 dpf embryos (Fig. S1I-Iʺ) and 6 dpf larvae (Fig. S1J-Jʺ).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Gpr85 is broadly expressed by maturing neurons in the zebrafish brain parenchyma during late embryogenesis. A–E, Lateral views of live gpr85^GAL4/+, Tg(UAS:GFP-CAAX) zebrafish embryos at 1 dpf with bright-field (A) and fluorescence imaging (B–E) showing the GFP-CAAX signal in the brain and spinal cord at 1 dpf (B), 2 dpf (C), 3 dpf (D), and 6 dpf (E). (N = 3) Scale bars, 50 μm. F–I, Maximum projection confocal images of coronal sections of the brain from a 3 dpf gpr85^GAL4/+, Tg(UAS:Lifeact-eGFP) zebrafish larva stained with anti-GFP (green; F′–I′), anti-HuC (red; (F″–I″) and DAPI (blue; F″′–I″′). Lifeact-eGFP⁺ cells are observed in the OB, subpallium (F′), Ha, retina, eminentia thalami, preoptic region, pallium (G′), tectum opticum, hypothalamus (H′), cerebellar plate, and MO (I′). Boxed regions are enlarged in panels J to M. Scale bars, 50 µm. J–M, Enlarged optical sections of the areas boxed in panels (F–I), showing Lifeact-eGFP⁺/HuC⁺ neurons in the brain parenchyma. (N = 3) Scale bars, 25 µm. N, Schematic representation of the coronal sections of the brain illustrated in F–I, with the forebrain in pink, the midbrain in turquoise, and the hindbrain in peach. CeP, cerebellar plate; DT, dorsal thalamus; EmT, eminentia thalami; H, rostral hypothalamus; Ha, habenula; Hc, caudal hypothalamus; MO, medulla oblongata; N, region of the nucleus of medial longitudinal fascicle; M2, migrated posterior tubercular area; M3, migrated area of EmT; OB, olfactory bulb; Oe, olfactory epithelium; P, pallium, Po, preoptic region; PTv, ventral part of posterior tuberculum; Re, retina; S, subpallium; TeO, tectum opticum; T, midbrain tegmentum; dpf, days postfertilization.

Furthermore, we observed GFP signal in the intestine of 6 dpf larvae, where specific cells were labeled all along the tract (Fig. S2A,B). Using gpr85^GAL4/+, Tg(UAS:GFP-CAAX; NBT:DsRed) fish, in which neural-specific beta tubulin (NBT) marks neurons, we identified the GFP⁺ cells as DsRed⁺ cells in the developing intestine of 6 dpf zebrafish (Fig. S2C,D).

We then focused on the developing brain, where gpr85 was already robustly expressed by 2 dpf, to further analyze its spatiotemporal pattern.

Gpr85 is broadly expressed in zebrafish brain neurons at the end of embryogenesis

To delineate more precisely the anatomical structures of the developing zebrafish brain that express gpr85, we performed anti-GFP immunodetection on coronal sections of 3 and 6 dpf gpr85^GAL4/+, Tg(UAS:Lifeact-eGFP) embryos. At 3 dpf, Lifeact-eGFP expression was detected throughout the brain (Fig. 1F–I). Robust expression was observed in the olfactory bulb (OB) and subpallium while notably absent from the olfactory epithelium at this stage (Fig. 1Fʹ). Gpr85 reporter expression was observed in cell clusters from the left and right habenula (Ha; Fig. 1Gʹ). In more caudal brain regions, Lifeact-eGFP⁺ cells were scattered across the brain parenchyma (Fig. 1Hʹ,Iʹ).

To confirm the neuronal identity of Lifeact-eGFP⁺ cells, coimmunostaining was performed with HuC, a pan-neuronal marker during zebrafish embryogenesis (Kim et al., 1996; Park et al., 2000). All Lifeact-eGFP⁺ cells were also positive for HuC (Fig. 1J–M) except for a few Lifeact-eGFP⁺/HuC⁻ cells which were observed in the midbrain ventral neuronal progenitor/precursor zone (Fig. S3A-A′′, arrowheads). These cells are likely postmitotic neuronal precursors migrating toward the developing brain parenchyma, as no PCNA⁺/Lifeact-eGFP⁺ cells were detected in this region (Fig. S3B,C″, arrowheads). Thus, at the end of embryogenesis, Lifeact-eGFP⁺ cells can largely be classified as maturing neurons.

A similar expression pattern of the gpr85 reporter was observed in our gpr85^GAL4/+, Tg(UAS:Lifeact-eGFP) line at 6 dpf in the hypothalamus (H; Fig. S3D), OB, pallium, subpallium (Fig. S3E), Ha (Fig. S3F,H), TL, TeO, cerebellum, and medulla oblongata (MO; Fig. S3G; Shainer et al., 2023).

We then compared our gpr85 expression data in embryos with a previous zebrafish brain single-cell transcriptomic dataset named “Daniocell,” which covers zebrafish development from 3 to 120 h postfertilization (hpf; Farrell et al., 2018; Sur et al., 2023). Consistent with our reporter line, gpr85 transcripts were undetectable before the midsegmentation stage (14–21 hpf; Fig. S4A). Afterward, transcripts were observed in the neural system, eye, and spinal cord throughout embryonic and postembryonic development (Fig. S3B). Initial expression appeared in developing forebrain neurons, motor neurons, and spinal cord interneurons (Fig. S4C). At midsegmentation, gpr85-expressing neurons were identified in the pallium and ventral forebrain/diencephalon (Fig. S4C).

Later during embryonic development, gpr85 expression expanded to additional brain regions, where it was observed in various neuronal cell types, including GABAergic, glutamatergic, glycinergic, and dopaminergic neurons. The expression of gpr85 was detected in neuronal subpopulations of the retina INL and GCL, consistent with our observations (Fig. S4F, H), emerging from 36 hpf, in amacrine cells (ACs) and in ON/OFF bipolar cells (BPCs). Finally, the transcriptomic data revealed no gpr85 expression in neural progenitors, endothelial cells, or immune cells of the central nervous system (CNS).

Taken together, these results confirm that our newly generated gpr85^GAL4 reporter lines are robust and ideal tools for identifying and studying gpr85-expressing cell populations in the zebrafish CNS and retina.

Gpr85 expression is maintained in the brain, retina, and intestine of adult zebrafish

The expression profile of gpr85 in the adult zebrafish brain has been previously documented by others through qPCR analysis (Breton et al., 2023), though with limited spatial resolution. Taking advantage of our gpr85 reporter lines, we further characterized the anatomical distribution of gpr85-expressing cells in the adult zebrafish brain. We first performed tissue clearing and imaging of an entire hemisphere of gpr85^GAL4/+, Tg(UAS:Lifeact-eGFP) adult brains. As during development, GFP⁺ cells were observed in various regions of the adult brain (Fig. 2A).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Cartography of gpr85 expression in the brain of adult zebrafish. A, Maximum projection confocal images of native Lifeact-eGFP fluorescence in a cleared adult brain hemisphere (sagittal view) from a gpr85^GAL4/+, Tg(UAS:Lifeact-eGFP) fish. Scale bar, 500 µm. B–H, Maximum projection confocal images of brain sections from a gpr85^GAL4/+, Tg(UAS:GFP-CAAX) adult zebrafish stained with anti-GFP (green) and DAPI (blue). B, Sagittal sections of corpus cerebelli (Cce). The boxes indicate the boxes in A indicate the regions enlarged in panels D–G. C, Sagittal sections of valvula cerebelli (Va). The box in C indicates the region enlarged in panel H. D, GCL. E, PCL. F, Molecular layer. G, Lca. H, valvula cerebelli. (N = 3) Scale bars, (C, D, E, F, G, H) 50 µm, (B) 500 µm.

We then analyzed sagittal and coronal sections from gpr85^GAL4/+, Tg(UAS:GFP-CAAX) brains. In the hindbrain, we observed that the highest expression was observed in the cerebellum with signal observed in different regions of the corpus cerebelli (Fig. 2B, D–G) and valvula (Fig. 2C,H) cerebelli. We observed GFP⁺ cells in granule cell layer (GCL; Fig. 2D), molecular layer (Fig. 2F), and lobus caudalis (Lca; Fig. 2G). No signal was detected in the Purkinje cell layer (PCL; Fig. 2E). Only scattered GFP⁺ cells were observed in the MO (Fig. 2A).

In the midbrain, a strong signal was detected in the torus longitudinalis (TL; Fig. 2A; Fig. S5A,B′) and the TeO (Fig. 2A; Fig. S5C-C′). In the forebrain, GFP⁺ cells were scattered in the OB (Fig. 2A), pallium/subpallium of the telencephalon (Tel; Fig. 2A), and the hypothalamus (H) where a strong GFP signal was observed in the posterior recess (Fig. 2A; Fig. S5D–E′).

Consistent with our embryonic observations, clusters of GFP⁺ cells were detected in the right and left Ha (Fig. 2A; Fig. S5F-F′). Notably, while GFP signal was prominent in the pituitary (Fig. S5G, H′), GFP⁺ cells were sparse, suggesting that this signal likely arises from neuronal projections rather than from intrinsic pituitary cells.

Altogether, strong gpr85 expression was detected in the cerebellum, Ha, hypothalamus, and TL of the adult zebrafish brain—mirroring patterns observed during development.

In addition to the brain, we investigated gpr85 reporter expression in the intestine. Using the gpr85^GAL4/+, Tg(UAS:GFP-CAAX; NBT:DsRed) line, we confirmed the presence of GFP⁺/DsRed⁺ neurons in the adult intestine (Fig. S2E). In addition, sparse GFP^low/DsRed⁻ cells were scattered within the intestinal villi (Fig. S2F,G).

Since GPR85 expression has been previously reported in the human testis (Matsumoto et al., 2000), we investigated its presence in zebrafish gonads. In the testis, GFP signal was detected in all cells (Fig. S2H-H′), whereas no GFP⁺ cells were observed in the ovaries.

Consistent with embryonic findings, GFP⁺ cells were also detected in the INL and the GCL of the adult retina (Fig. 3A-A′).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Gpr85 is enriched in the pre- and postsynaptic compartments of developing and adult IPL retinal ribbon synapses. A, Maximum projection confocal images of the retina from a gpr85^GAL4/+, Tg(UAS:GFP-CAAX) adult zebrafish stained with anti-GFP (green) and DAPI (blue). (N = 3) Scale bar, 50 µm. B–I, Confocal images of coronal retinal sections from 6 dpf (B–E″) or 6 mpf (F–I″) gpr85^GAL4/+, Tg(UAS:gpr85-eGFP) zebrafish stained with anti-GFP (green), anti-Ribeye-A (red), and DAPI (blue). B, F, Larval or adult Gpr85-eGFP⁺ AC somas and signal within the IPL are shown, with boxed regions enlarged in C-C″ and G-G″, respectively. White arrowheads highlight Gpr85-eGFP signals adjacent to Ribeye⁺ ribbon terminals. D, H, Larval or adult Gpr85-eGFP⁺/Ribeye-A⁺ BPC ribbon presynaptic terminals are shown, with boxed regions enlarged in E-E″ and I-I″, respectively. White arrowheads show examples of Gpr85-eGFP signal present in Ribeye⁺ ribbon terminals. (N = 3) Scale bars, (B, D) 25 µm, (F, H) 50 µm, (C, E, G, I) 10 µm. J, Electron microscopy coronal view of the structure of 6 dpf retina from gpr85^GAL4/+, Tg(UAS:gpr85-mCherry) zebrafish larvae. (N = 3) Scale bar, 10 µm. K, Enlarged image of the ribbon presynaptic terminals (yellow) from the boxed area in panel J. Scale bar, 1 µm. L, M, Enlarged images of synaptic boutons from the boxed area in panel K. White arrowheads highlight postsynaptic densities while black arrowheads point to presynaptic vesicles. Scale bars, 200 nm. N–S, Immunogold labeling of Ribeye (N–P) and mCherry-tagged Gpr85 (Q–S) focusing on the retinal inner plexiform layer from 6 dpf gpr85^GAL4/+, Tg(UAS:gpr85-mCherry) larvae. Dashed yellow lines highlight plasma membranes. Boxes indicate regions enlarged in panels O, P, R, and S. Scale bars, 200 nm. dpf, days postfertilization; mpf, months postfertilization; ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; GCL ganglion cell layer.

Taken together, these results show that gpr85 expression is maintained from late embryogenesis through adulthood in the brain, retina, and intestine, supporting its persistent and region-specific role in the zebrafish nervous system and gut.

Gpr85 receptor localizes to synapses in the retina and brain

To gain insight into the biological function of Gpr85, we characterized its subcellular localization in vivo, with a focus on neurons of the zebrafish brain and retina. A previous in vitro study suggested an interaction between GPR85 and neuroligin-bound PSD-95, a postsynaptic marker of excitatory synapses (Fujita-Jimbo et al., 2015). Based on this, we hypothesized that Gpr85 localizes to chemical synapses in zebrafish neurons.

To test this, we used our gpr85^GAL4/+ line to overexpress the Gpr85 receptor fused to either eGFP or mCherry at its C-terminal end. Both gpr85^GAL4/+, Tg(UAS:gpr85-eGFP) and gpr85^GAL4/+, Tg(UAS:gpr85-mCherry) lines were fertile and did not exhibit developmental defects. While no ectopic expression was observed, both lines exhibited incomplete labeling compared with the Lifeact-eGFP and GFP-CAAX reporter lines.

We first assessed subcellular localization in the retina. Coronal sections of the retina from gpr85^GAL4/+, Tg(UAS:gpr85-eGFP) fish revealed Gpr85-eGFP⁺ ACs (Fig. 3B,F) and BPCs (Fig. 3D,H) during development and adulthood. In ACs, the Gpr85-eGFP signal was localized in the soma (Fig. 3B,F, black arrowhead) and dendritic arborization of the cells within the IPL at both 6 dpf (Fig. 3B) and 6 months postfertilization (mpf; Fig. 3F). In BPCs, the Gpr85-eGFP signal was also observed in the soma (Fig. 5D, black arrowhead) and neurites/terminals in larvae IPL (Fig. 3D) while only found in the IPL of adult zebrafish (Fig. 3H). The dendritic arborization of BPCs was essentially devoid of Gpr85-eGFP.

To assess synaptic localization, we performed coimmunostaining for αRibeye-A, a presynaptic ribbon synapse marker. Gpr85-eGFP signal was enriched at the postsynaptic sites of ribbon synapses in the dendritic arborization of developing (Fig. 3C-C″, white arrowheads) and adult (Fig. 3G-G″, white arrowheads) ACs. In developing (Fig. 3E-E″, white arrowheads) and adult (Fig. 3I-I″, white arrowheads) BPCs, Gpr85-eGFP displayed a strong signal at the level of the Ribeye-A⁺ presynaptic terminals of ribbon synapses.

These findings provide the first in vivo evidence that Gpr85 is enriched at both pre- and postsynaptic sites of ribbon synapses in the developing and adult retina.

Given that Gpr85 is a seven-transmembrane receptor, it was anticipated to localize at the plasma membrane. However, no direct evidence of such localization in vivo had been reported previously. Using gpr85^GAL4/+, Tg(UAS:gpr85-mCherry) larvae, we investigated the distribution of the receptor by electron microscopy using immunogold staining. The initial characterization of the retinal structure revealed the ribbon terminals within the IPL (Fig. 3J,K, yellow). Presynaptic compartments were identifiable due to the presence of presynaptic vesicles (Fig. 3L,M, black arrows), while postsynaptic densities were identified as thick and electron-dense structures (Fig. 3L,M, white arrowheads). Ribeye-A immunogold labeling confirmed presynaptic identity (Fig. 3N–P). Gpr85-mCherry exhibited localization at the plasma membrane in the IPL (Fig. 3Q–S) confirming its membrane localization in vivo.

Next, we assessed the subcellular localization of the chimeric receptor in the developing and adult brain, using the Tg(UAS:FinGR(PSD)-GFP) and Tg(UAS:FinGR(GPHN)-mCherry) reporter lines to discriminate between excitatory and inhibitory synapses, respectively (Son et al., 2016). These lines were crossed to gpr85^GAL4/+, Tg(UAS:gpr85-mCherry) or Tg(UAS:gpr85-eGFP) backgrounds. Coimmunostaining was performed on 6 dpf coronal sections and adult sagittal sections of the brain.

As in the retina, Gpr85-eGFP⁺ and Gpr85-mCherry⁺ puncta were identified in neuronal projections at both larval and adult stages (Fig. 4). We observed both Gpr85-mCherry⁺/PSD⁺ (Fig. 4A,B, white arrowheads) and Gpr85-eGFP⁺/GPHN⁺ (Fig. 4C,D, white arrowheads) synapses in various regions of the developing and adult brain, including the cerebellum, Tel, hypothalamus, and TeO. As observed in the retina, the expression patterns of Gpr85 fusion proteins are incomplete, resulting in the detection of a fraction of PSD⁺ or GPHN⁺ cells which are also expressing Gpr85 fusion proteins.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Gpr85 is enriched at the level of excitatory and inhibitory synaptic inputs of gpr85-expressing neurons throughout the developing and adult brain. A, Confocal images of coronal brain sections from 6 dpf gpr85^GAL4/+, Tg(UAS:gpr85-mCherry; UAS:FinGR(PSD)-GFP) larvae stained with anti-GFP (green), anti-mCherry (red), and DAPI (blue). White arrowheads show examples of Gpr85-mCherry signal localized at the level of PSD⁺ excitatory synapses in the Tel and TeO. Scale bar, 10 µm. B, Confocal images of coronal brain sections from 6 mpf gpr85^GAL4/+, Tg(UAS:gpr85-mCherry; UAS:FinGR(PSD)-GFP) adult zebrafish stained with anti-GFP (green), anti-mCherry (red), and DAPI (blue). White arrowheads show examples of Gpr85-mCherry signal localized at the level of PSD⁺ excitatory synapses in the indicated brain regions. Scale bar, 10 µm. C, Confocal images of coronal brain sections from 6 dpf gpr85^GAL4/+, Tg(UAS:gpr85-eGFP; UAS:FinGR(GPHN)-mCherry) larvae stained with anti-GFP (green), anti-mCherry (red), and DAPI (blue). White arrowheads highlight examples of Gpr85-eGFP signal present at the level of the GPHN⁺ inhibitory synapses in the Tel and TeO. Scale bar, 10 µm. D, Confocal images of coronal brain sections from 6 mpf gpr85^GAL4/+, Tg(UAS:gpr85-eGFP; UAS:FinGR(GPHN)-mCherry) adult zebrafish stained with anti-GFP (green), anti-mCherry (red), and DAPI (blue). White arrowheads highlight examples of Gpr85-eGFP signal present at the level of the GPHN⁺ inhibitory synapses of the indicated brain regions. Scale bar, 10 µm. (for all N = 3) dpf, days postfertilization; mpf, months postfertilization. GPHN, Gephyrin; PSD, postsynaptic density protein 95.

In the adult cerebellum, we did not observe any somatic signal in the granule cells (GCs) of the corpus cerebelli. Our data suggest that the endogenous receptor is not evenly distributed along the membrane of GCs but is instead enriched at the postsynaptic level of the glomeruli (Fig. 4B,D, white arrowheads).

Overall, our findings provide the first in vivo evidence that Gpr85 is a synaptic membrane receptor, localized at both excitatory and inhibitory chemical synapses in the zebrafish retina and CNS across development and adulthood.

Gpr85 is dispensable for the development, viability, and formation of excitatory inputs of gpr85-expressing optic tectal neurons

We previously showed that gpr85 is expressed in multiple differentiating neuronal subpopulations during early zebrafish brain development. In addition, we provided evidence that Gpr85 is a synaptic membrane receptor in vivo. These findings raised the question of whether Gpr85 is required for the development or maturation of gpr85-expressing neurons.

To address this, we generated a complementary constitutive Gpr85-KO model (gpr85^Δ4/Δ4; Fig. 5A). Similar to the gpr85^GAL4/GAL4 line, homozygous mutants were viable and fertile and displayed no noticeable developmental defects (Fig. 5B,C). The offspring of these mutants followed a Mendelian inheritance ratio (Fig. 5D).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Zebrafish gpr85 loss-of-function models are phenotypically normal, and cell and excitatory synaptic input densities of Gpr85-deficient neurons are unaffected through development. A, Genomic sequence of WT and gpr85 deletion mutant (Δ4/Δ4) zebrafish. The mutant gpr85 sequence, generated via CRISPR/Cas9, results in a 4 base pair deletion (boxed) in exon 2, introducing a premature stop codon 17 amino acids later (p.Ala13Glyfs*30). B, No phenotypic differences were observed between 3 dpf WT or Δ4/Δ4 larvae. Scale bar, 500 µm. C, Adult gpr85Δ^4/Δ⁴ zebrafish show no macroscopic defects. D, Genotypic analysis of adult fish from heterozygous intercrosses. A normal Mendelian inheritance ratio is observed. Each dot color represents an independent cross (n = 5; at least 25 fish genotyped per cross; means: +/+, 24.7%; +/Δ4, 50%; Δ4/Δ4, 25.3%). E, Body length measurements of adult zebrafish with gpr85^Δ4/Δ4 (n = 5) and gpr85^GAL4/GAL4 (n = 28) loss-of-function mutations, compared with their respective control siblings (n = 18 for Δ4 siblings; n = 37 for GAL4 siblings). Statistical analysis, +/+ vs Δ4/Δ4, ns, p = 0,68; GAL4/+ vs GAL4/GAL4, ns, p = 0.59; M–W test. F, Wet brain weight normalized to body length for adult zebrafish with gpr85^Δ4/Δ4 and gpr85^GAL4/GAL4 loss-of-function mutations, compared with their respective control siblings (+/+ vs Δ4/Δ4, ns, p = 0.41; GAL4/+ vs GAL4/GAL4, ns, p = 0.91; M–W test). G, gpr85 transcript levels measured by RT-QPCR in gpr85^+/+ (n = 9), gpr85^Δ4/Δ4 (n = 5) and gpr85^GAL4/GAL4 (n = 5) 6 dpf larvae (+/+ vs Δ4/Δ4, **p = 0.002; +/+ vs GAL4/GAL4, **p = 0.007; M–W test). H, Schematic coronal view of a 6 dpf zebrafish larva brain. The box indicates the TeO region imaged in I. I, Maximum projection confocal images of TeO coronal sections from 6 dpf gpr85^GAL4/+, Tg(UAS:FinGR(PSD)-GFP) larvae immunostained with anti-GFP (green) and DAPI (blue). (N = 3) Scale bar, 50 µm. J, Quantification of PSD⁺ gpr85-expressing neuronal soma in the TeO SPV at 6 dpf. gpr85^GAL4/+ controls (n = 45) are compared with gpr85^GAL4/Δ4 fish (n = 26; gpr85^GAL4/+ vs gpr85^GAL4/Δ4; ns, not significant; p = 0.8; unpaired t test). K, Schematic dorsal view of a 6 dpf zebrafish larva brain. The box indicates the NP regions imaged in L. L, Live confocal imaging of the NP area from a 6 dpf gpr85^GAL4/+ vs gpr85^GAL4/Δ4, Tg(UAS:FinGR(PSD)-GFP) larva (dorsal view) showing PSD⁺ excitatory synapses from gpr85-expressing neurons. Scale bars, 10 µm. M, Quantification of PSD⁺ excitatory synapses belonging to gpr85-expressing neurons in the TeO NP at 6 dpf. gpr85^GAL4/+ controls (n = 25) are compared with gpr85^GAL4/GAL4 fish, as gpr85-deficient model (n = 28; gpr85^GAL4/+ vs gpr85^GAL4/Δ4; ns, not significant, p = 0.169; unpaired t test). SPV, stratum paraventriculare; NP, neuropil. Data are presented as mean ± SEM.

Although a prior study reported increased brain weight in GPR85-KO mice (Matsumoto et al., 2008), we found no significant differences in body length or in the wet brain weight-to-body length ratio in either of the Gpr85-KO zebrafish models (Fig. 5E,F). Nevertheless, we observed that gpr85 transcript levels were approximately twice as high in both Gpr85-KO models compared with wild-type 6 dpf larvae (Fig. 5G), suggesting that the absence of functional Gpr85 triggers upregulation of the gene and/or increases the stability of its transcripts.

To further investigate the impact of Gpr85-KO on the development of gpr85-expressing neurons, we used gpr85^GAL4/Δ4, Tg(UAS:FinGR(PSD)-GFP) larvae to quantify the developing optic tectal gpr85-expressing neurons in the SPV and their excitatory inputs within the NP. The NP, almost devoid of neuronal soma, exhibits a uniform distribution of excitatory inputs at 6 dpf (Fig. 5H–M).

We observed no significant changes in the density of FinGR(PSD)-GFP⁺ cells in the SPV of 6 dpf larvae in the Gpr85-KO larvae compared with controls (Fig. 5I,J). Similarly, the density of FinGR(PSD)-GFP⁺ synaptic inputs within the tectal NP remained unaffected in whole-mount immunostaining of Gpr85-KO larvae (Fig. 5L,M).

These results show that Gpr85 is dispensable for the development, viability, and formation of excitatory inputs in gpr85-expressing neurons during zebrafish development.

Transcriptomic analysis of Gpr85-KO cerebellar GCs reveals changes in genes related to neuronal activity

Although Gpr85 appears dispensable for the development of gpr85-expressing neurons and developmental synaptogenesis, it may influence neuronal homeostasis and/or synaptic activity. To explore the potential impact of constitutive Gpr85 deficiency on adult gpr85-expressing neurons, we conducted single-cell RNA sequencing (scRNAseq) on eGFP⁺ sorted cells from the adult brains of gpr85^GAL4/GAL4, Tg(UAS:Lifeact-eGFP) and gpr85^GAL4/+, Tg(UAS:Lifeact-eGFP) zebrafish.

After quality control and filtering, data from both genotypes were merged for clustering, which identified 11 distinct clusters under both control and KO conditions (gpr85^GAL4/+, Ctl, n = 4,219 cells; gpr85^GAL4/GAL4, KO, n = 3,341 cells; Fig. 6A; Data S1). As expected, 96% of cells expressed the eGFP transgene (Fig. 6B). Gpr85 transcripts were detected, though not in all cells (Fig. 6B), likely due to expression levels below the detection threshold of the sequencing method and/or dynamic expression patterns not captured by the stable eGFP reporter. Unlike our RT-qPCR findings at 6 dpf, this adult dataset showed no global or cluster-specific upregulation of gpr85 transcripts in mutants.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

scRNAseq analysis of eGFP⁺ sorted cells from adult gpr85^GAL4/+ and gpr85^GAL4/GAL4, Tg(UAS:lifeact-eGFP) dissociated brains reveals changes in gene expression related to neuronal activity. A, Experimental strategy for assessing transcriptomic changes in gpr85-expressing cells from the Gpr85-deficient adult brain. GFP⁺ cells were sorted from gpr85^GAL4/+ (Ctl) and gpr85^GAL4/GAL4 (KO), Tg(UAS:lifeact-eGFP) dissociated brains (cell sorting strategy shown with 10,000 events per condition). The right panel shows the UMAPs of cells from the Ctl and KO conditions (n = 4,219 cells and n = 3,341 cells, respectively) after filtering and clustering. B, UMAPs of Ctl and KO cells merged, displaying expression of the UAS:lifeact-eGFP transgene, gpr85, the two pan-neuronal markers elavl3 and rbfox3a, as well as the expression of the GC markers, sls17s7a, neurod1, cbln12, and fat2a. C, Table of the most up- and downregulated genes within the isolated cerebellar clusters (0 to 6) in the KO condition, with a minimal FC of two and an adjusted p < 10⁻⁴, expressed by at least 6% of the cells. D, Violin plots of most significantly DEGs with a minimal FC of 1.4 and an adjusted p < 10⁻⁴. All genes referenced are upregulated (mean represented by the dots). Top panel, Selection of genes related to exocytosis. Intermediate panel, Selection of genes encoding channels related to neuronal excitability/activity (voltage-dependent channels, clcn2a, cacna1ba, kcnh3, cacng8b; Na⁺/K⁺ ATPase subunits, atp1a3b, atp1b1b, atp1a1a.1, atp1b2a, syn2b). Lower panel, genes documented as IEGs.

All clusters, except Cluster 9, displayed neuronal identities, expressing the pan-neuronal markers elavl3 and rbfox3a (Fig. 6B). Based on gad1a, gab1b, and gad2 expression (Fig. S6C), we observed that 98% of the cells were non-GABAergic neurons. Cluster 9 was identified as non-neuronal, expressing leukocyte and microglial markers such as cd74b, mhc2a, and apoeb (Fig. S6A,B; Rovira et al., 2024).

Clusters 0–6 were identified as cerebellar GCs based on markers like slc17a7a, neurod1, cbln12, and fat2 (Fig. 6B, Fig. S7; Bae et al., 2009; Kaslin et al., 2009; Takeuchi et al., 2017, Xu et al., 2018). However, we did not identify Bergmann glia (slc1a3b, fabp7a), Purkinje, Golgi, and stellate cells (gad1a, aldoca, ca8; Bae et al., 2009; Fig. S6C). Moreover, Olig2 in the cerebellum marks a subpopulation of eurydendroid cells (Bae et al., 2009; Harmon et al., 2020), but olig2 expression was not observed in our data, showing the absence of this subpopulation in our dataset (Fig. S6C). Cluster 8, based on the expression of the neuroligin (nlgn) genes (Davey et al., 2010), was proposed to represent ventral Tel neurons (Fig. S6B,C). Cluster 10 was identified as dorsal pallium neurons characterized by expression of eomesa (Fig. S6B,C; Ganz et al., 2015), and Cluster 11 was identified as kiss1⁺ neurons from the ventral Ha (Fig. S6B; Kitahashi et al., 2009). Overall, ∼90% cells were classified as cerebellar GCs.

To identify transcriptional changes resulting from the loss of Gpr85, we focused on cerebellar GCs from Clusters 0 to 6 to assess differential expressed genes (DEGs; Data S1). Firstly, we selected genes with a fold change (FC) ≥2, adjusted p < 10⁻⁴ and expressed by ≥6% of cells. We identified 27 upregulated and 23 downregulated genes (Fig. 6C).

The most upregulated gene was the ornithine decarboxylase antizyme oaz2b (FC = 5.13), and oaz1a was also among the upregulated genes. The histone-like hist1h4l.18 was the most downregulated gene. Remarkably, upregulated genes included three genes encoding proteins involved in neuronal activity and excitability, the gamma-aminobutyric acid type A receptor subunit delta (gabrd), an ATPase Na⁺/K⁺ transporting subunit (atp1b1b), and a potassium voltage-gated channel (kcnh3; Fig. 6C). We also noticed that several of the most upregulated DEGs are classified as immediate early genes (IEGs; npas4a, egr1, egr4; Fig. 6C).

Based on these findings, we examined the DEG list for further enrichment in genes related to neuronal activity and excitability. In total, we observed upregulation of 11 genes related to neuronal activity-dependent exocytosis, 8 genes encoding ATPase Na⁺/K⁺ exchangers or voltage-dependent channels, and 10 IEGs (Fig. 6D; Table S1; FC > 1.4; adjusted, p < 10⁻⁴).

To validate the differential expression patterns observed in our scRNAseq dataset, we performed qPCR analysis on dissected adult cerebella from gpr85^GAL4/+ (Ctl) and gpr85^GAL4/+ (KO) siblings (Fig. S7). We selected several upregulated genes, including oaz2b (the most upregulated gene), chga (exocytosis), atp1b1b (channel), egr1, npas4a, and dusp5 (IEGs; Fig. 6C,D; Table S1). This analysis confirmed the upregulation of oaz2b, chga, and dusp5, while atp1b1b showed a trend toward increased expression, although it did not reach statistical significance (Fig. S7). In contrast, no significant changes were detected for the IEGs npas4a and egr1. Given that several IEGs have been reported to be induced during tissue dissociation and scRNAseq workflows (Ishikawa and Sakurai, 2015; Lacar et al., 2016; Marsh et al., 2020), the upregulation of IEGs observed in gpr85^GAL4/GAL4 (KO) may reflect an increased sensitivity to the stress associated with dissociation. However, this hypothesis requires further formal confirmation. Supporting this, we observed that while npas4a and egr1 were essentially upregulated in the GC Cluster 2 (IEGs⁺ cluster), dusp5, oaz2b, chga, and atp1b1b were more broadly distributed and upregulated across multiple GC clusters (Fig. S7). Thus, we validated the upregulation of dusp5, oaz2b, and chga in the Gpr85-KO cerebellum.

Altogether, these findings confirm that gpr85 is predominantly expressed in cerebellar GCs within the adult zebrafish brain. The absence of Gpr85 alters the transcriptional profile of these neurons, particularly genes associated with neuronal activity and excitability suggesting a role of Gpr85 in modulating cerebellar GCs electrophysiological activity.

Spontaneous activity of adult cerebellar Gpr85-KO GCs is increased ex vivo

Based on our transcriptomic analysis, we hypothesized that the electrophysiological properties of cerebellar GCs might be altered in Gpr85-KO models. To investigate this, we conducted electrophysiological recordings on ex vivo horizontal cerebellar slices from adult gpr85^GAL4/+ (or gpr85^GAL4/GAL4), Tg(UAS:Lifeact-eGFP) zebrafish (Fig. 7A). Recordings focused on gpr85-expressing GCs (eGFP⁺) to assess their spontaneous neuronal activity and passive membrane properties (Fig. 7A–C).

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

Gpr85-KO impacts adult cerebellar GCs electrophysiological properties. A, Schematic representation of the experimental setup. Cerebellar horizontal sections (220 µm) were prepared, and GCs were identified based on GFP fluorescence. Recordings were made in cell-attached and whole-cell voltage–clamp configurations. The chart illustrates the percentage of GCs in cell-attached configuration that exhibited or not spontaneous activity. B, Quantification of GCs spontaneous activity in a cell-attached configuration. Firing patterns are shown, as well as the firing rates measured over a 60 s recording (gpr85^GAL4/+, n = 13, vs gpr85^GAL4/GAL4, n = 15; *p = 0.0347; unpaired t test with Welch's correction). C, Quantification of GCs passive properties (membrane capacitance, Cm; membrane resistance, Rm; membrane time constant, τ) in a whole-cell voltage–clamp configuration (gpr85^GAL4/+, n = 26, vs gpr85^GAL4/GAL4, n = 25; Cm, ns, p = 0.9132; τ, ns, p = 0.1724; M–W U tests; Rm, *p = 0.0113; unpaired t test with Welch's correction). Ce, cerebellum; GCL, granule cell layer. Data are presented as mean ± SEM (B, C).

Cell-attached recordings revealed that in both genotypes, GCs displayed either no spontaneous activity (gpr85^GAL4/+: 17/30, 57% vs gpr85^GAL4/GAL4: 10/25, 40%) or some spontaneous activity (gpr85^GAL4/+: 13/30, 43% vs gpr85^GAL4/GAL4: 15/25, 60%; Fig. 7A).

Among spontaneously active neurons, the firing rate was significantly higher in Gpr85-KO GCs compared with controls (Fig. 7B; Hz, gpr85^GAL4/+, 4.69 ± 1.24 vs gpr85^GAL4/GAL4, 12.62 ± 3.24; p = 0.035; t test). This result provides the first direct evidence that Gpr85 regulates the excitability of the adult cerebellar GCs.

We next investigated the passive membrane properties of neurons using a whole-cell patch-voltage-clamp configuration. While the capacitance (Cm) was not significantly different between our groups (Fig. 7C), we observed a significant increase in membrane resistance (Rm) in Gpr85-KO GCs (Cm, gpr85^GAL4/+, 2.46 ± 0.25 pF vs gpr85^GAL4/GAL4, 2.58 ± 0.34 pF; p = 0.913, M–W test; Rm, gpr85^GAL4/+, 3,079 ± 188 MΩ; n = 26 vs gpr85^GAL4/GAL4, 4,663 ± 554 MΩ; p = 0,011; t test). This was not accompanied by significant changes in the membrane time constant (mS, gpr85^GAL4/+, 7.47 ± 0.93 ms vs gpr85^GAL4/GAL4, 10.81 ± 1.65 ms; p = 0.172; M–W test; Fig. 7C).

Altogether, these results indicate that Gpr85, in a direct or indirect manner, influences the electrophysiological properties of adult cerebellar GCs. Specifically, its loss is associated with increased firing rates and elevated membrane resistance, suggesting a role of Gpr85 in the regulation of neuronal excitability.

Gpr85-KO larvae exhibit a stronger light-triggered motor response

We previously highlighted that Gpr85 is enriched at excitatory and inhibitory synapses in both the retina and the brain. We also reported that the receptor is expressed in key anatomical regions of the locomotor network and visual system, including the retina, TeO, cerebellum, and spinal cord. Moreover, transcriptomic analyses of adult Gpr85-KO cerebellar GCs revealed gene expression changes related to neuronal activity and excitability. Finally, electrophysiological studies showed that Gpr85 deficiency alters the intrinsic neuronal properties of adult cerebellar GCs. Together, these findings led us to question if Gpr85 deficiency could have an impact on stress-response of Gpr85-KO larvae.

Adapting previously reported protocols (Scott et al., 2016; Campanari et al., 2020), we used the ZebraBox system to evaluate motor responses to dark/light stimuli. Gpr85 heterozygous fish were incrossed, and their 6 dpf offspring were randomly and individually distributed into 48-well plates for analysis. We first recorded 20 min of activity in the dark to assess spontaneous locomotor activity under basal conditions (Fig. 8A). No significant differences in basal activity were observed between Gpr85-KO larvae and their siblings (gpr85^Δ4/Δ4 vs siblings^Δ4; p = 0.36; gpr85^GAL4/GAL4 vs siblings^GAL4; p = 0.94; M–W test; Fig. 8B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 8.

Gpr85-KO impacts larval light-induced motor response. A, Schematic representation of the experimental setup for assessing spontaneous locomotor activity. B, Quantification of the distance swum by 6 dpf Gpr85-KO larvae compared with siblings over 20 min in the dark (gpr85^Δ4/Δ4, n = 69 vs siblings, n = 210; p = 0.36; and gpr85^GAL4/GAL4, n = 77, vs siblings, n = 190; p = 0.94; ns, not significant; M–W U tests). C, Schematic representation of the experimental setup for measuring light-induced motor responses. The motor response was recorded within the first second following light onset. D, The motor response was observed in controls as a significant increase in the distance swum following light onset (n = 210 siblings of gpr85^Δ4/Δ4 mutants; dark vs motor response; *p = 0.02; n = 247 siblings of gpr85^GAL4/GAL4 mutants, dark vs motor response; ****p < 0.0001; Wilcoxon matched-paired signed–rank tests). The motor response was significantly more pronounced in gpr85-deficient animals compared with controls (n = 80 gpr85^Δ4/Δ4 mutants vs siblings; *p = 0.015; n = 83 gpr85^GAL4/GAL4 mutants vs siblings, *p = 0.011; M–W U tests). SLA, spontaneous locomotor activity; SR, motor response. Data are presented as mean ± SD (B, D).

We next assessed the light-triggered motor response of Gpr85-KO larvae by measuring the distance swam during the first second following light onset (Fig. 8C). In control larvae, this stimulus reliably evoked a motor response, as evidenced by a significantly longer distance swam after light stimulation compared with the dark condition (siblings^Δ4, dark vs motor response; p = 0.02; siblings^GAL4, dark vs motor response; p < 0.0001; Wilcoxon test; Fig. 8D; Movie 1).

Both Gpr85-KO models displayed a significantly enhanced light-triggered motor response, reflected by a significantly increased distance swam following light induction as compared with their respective controls (gpr85^Δ4/Δ4 mutants vs siblings^Δ4; p = 0.015; gpr85^GAL4/GAL4 mutants vs siblings^GAL4; p = 0.011; M–W test; Fig. 8D). These results suggest that Gpr85 could be involved in the modulation of neuronal responses to light stimuli, potentially contributing to the regulation of visual and locomotor system activity in larvae.