2 Introduction

The ability to detect and respond to light is fundamental for many organisms. Vision relies on specialized photoreceptor cells that convert light into neural signals through a process called phototransduction. This allows the brain to interpret visual information and guide behavior. The rhodopsin chromoproteins, composed of the seven-helix membrane apoprotein opsin and the chromophore retinal, play a central role in this process. When a photon is absorbed, the chromophore undergoes a conformational change, transforming from 11-cis-3-hydroxyretinal to all-trans-3-hydroxyretinal. Then, structural changes occur in the protein, initiating signal transduction. As G protein-coupled receptors (GPCRs), rhodopsins first activate a trimeric G protein, which then initiates a cascade of downstream signaling events in phototransduction (Xiong and Bellen, 2013).

While the primary function of rhodopsins is image formation, some are also known to participate in non-image-forming functions. Based on their specific roles, rhodopsins are classified as visual or non-visual (Kingston and Cronin, 2016; Leung and Montell, 2017; Terakita, 2005). Several studies have highlighted the diverse functions of rhodopsins, including their involvement in thermosensation (Shen et al., 2011), hearing (Senthilan et al., 2012), proprioception (Zanini et al., 2018), and taste (Leung et al., 2020).

Opsins are classified into different groups, including c-opsins (ciliary) and r-opsins (rhabdomeric). While c-opsins trigger a G_αt (G_alpha transducin)-mediated signaling pathway leading to Na⁺ channel closure and consequent hyperpolarization, r-opsins trigger a G_αq-mediated depolarizing signaling cascade. In humans, c-opsins in rods and cones are essential for visual perception, while melanopsin (OPN4), an r-opsin expressed in intrinsically photosensitive retinal ganglion cells (ipRGCs), is involved in the synchronization of circadian rhythms and enhances contrast sensitivity (Sonoda et al., 2018). Drosophila melanogaster (Dmel) has seven rhodopsins (RH1-RH7), all belonging to the r-opsin group. RH1 (RHODOPSIN 1), the primary rhodopsin, is encoded by the ninaE (neither inactivation nor afterpotential E) gene and is expressed in the outer receptor cells R1-R6 of the retina in the compound eyes of flies. These cells project to the monopolar cells in the lamina, where the signals are processed and transmitted to the medulla. The inner receptor cells R7 (expressing RH3 and RH4) and R8 (expressing RH5 and RH6) form synapses directly within the medulla. The medulla receives all signals, processes them further, and transmits them to the lobula and lobula plate. The R8 rhodopsins, RH5 and RH6, are also expressed in the larval eye, Bolwig’s organ, and its adult remnant, the Hofbauer-Buchner (H-B) eyelet. RHODOPSIN 2 (RH2) is mainly expressed in the ocelli, while the expression of RHODOPSIN 7 (RH7) remains debated.

Phylogenetically, Drosophila rhodopsins cluster into three groups: The first group comprises RH3, RH4 and RH5; the second group consists of RH1, RH2 and RH6; and the last group is formed only by RH7. Homologs of RH7 occur broadly across arthropods, including hexapods, crustaceans, and chelicerates (Feuda et al., 2016; Senthilan and Helfrich-Förster, 2016). While the first six rhodopsins are well characterized, the function of RH7 remains unclear. RH7 shares typical rhodopsin features (seven transmembrane helices, chromophore-binding lysine, LRTPXN motif) but differs in having extended N- and C-terminal tails twice as long as in other rhodopsins (Figure 1A) (Senthilan and Helfrich-Förster, 2016). Moreover, RH7 and its homologs lack the highly conserved QAKKMNV motif in the third intracellular loop (ICL3), important for G protein interaction (Figure 1A, Supplementary Material S2A) (Senthilan and Helfrich-Förster, 2016). Unlike other rhodopsins, whose non-palindromic RCSI (Rhodopsin Core Sequence I) motif lies ∼100 bp upstream of their transcription start sites, the RCSI of Rh7 is fully palindromic (TAATCAGATTA) and located within its second intron, about 7 kb downstream of the transcription start site. Generally, the unique, non-palindromic RCSI sequences regulate cell-specific rhodopsin expression, while palindromic sequences correlate with broader gene expression (Datta and Rister, 2022; Rister et al., 2015; Senthilan and Helfrich-Förster, 2016). The unusual palindromic RCSI sequence, along with its positioning within the Rh7 gene, may account for the distinct and broader expression of Rh7, which is consistent with TAPIN (tandem affinity purification of INTACT nuclei)-seq data. These data show that Rh7 is absent in retinal photoreceptors but present in visual processing neurons in the lamina and medulla (Figure 1B) (Davis et al., 2020).

Figure 1

Previous studies on the expression and function of RH7 have been contradictory (Grebler et al., 2017; Kistenpfennig et al., 2017; Leung et al., 2020; Leung and Montell, 2017; Ni et al., 2017; Senthilan et al., 2019). Two factors may explain these discrepancies. First, the genetic background of the tested flies, particularly the presence of the white and yellow mutations as markers, appears to influence the Rh7 phenotype, possibly due to overlapping expression sites (Davis et al., 2020). Second, RH7 is relatively large but likely adopts a compact, self-contained structure (Figure 1A). Its highly phosphorylatable N and C termini may interact with other proteins and further obscure relevant epitopes, making antibody detection nearly impossible. However, all previous studies, despite their differences, agree that Rh7 is an atypical rhodopsin with non-visual functions, a hypothesis that is further supported by this study. We show that RH7 is distinct from other Drosophila rhodopsins. RH7 is expressed in non-retinal neurons and likely acts via a different signaling pathway, contributing to darkness detection and nocturnal activity. As an ancient rhodopsin, RH7 may support basic visual functions such as contrast processing and circadian synchronization, similar to melanopsins in mammals.

3 Materials and Methods

3.1 Bioinformatics

We searched for r-opsin homologs in selected animals of the Protostomia using BLASTP from NCBI (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE=Proteins, accessed December 2023). Amino acid sequences of Drosophila melanogaster RH1-RH7 served as reference queries. Searches were performed against the non-redundant protein database (nr) using the default BLASTP settings (BLOSUM62 matrix, gap costs 11/1, conditional compositional score matrix adjustment). From the initial output, we curated a representative dataset of 53 species, prioritizing those with RH7 homologs. Protein hits were compiled in Excel and imported into Geneious Prime (2025.1.3) for downstream analysis. In total, more than 400 sequences were retrieved. Multiple sequence alignments were performed using global alignment with BLOSUM62 and gap costs of 11/1. A phylogenetic tree was generated using the neighbor-joining method with the Jukes-Cantor distance model. To minimize redundancy due to isoforms, gene duplications, or duplicate annotations, the dataset was manually curated across multiple iterations, resulting in a final set of 280 unique sequences. Protein names, sequences, and accession numbers are provided in Supplementary Material S1. To confirm phylogenetic assignments, reciprocal BLAST searches were performed against the D. melanogaster proteome, ensuring clade consistency and homology validation.

3.2 GsX assay

G protein coupling of GPCR constructs was examined using HEK293T or HEK293-ΔG7 cells (lacking GNAS, GNAL, GNAQ, GNA11, GNA12, GNA13, and GNAZ) obtained from A. Inoue at Tohoku University (Bowin et al., 2019). HEK293T cells were cultured in DMEM supplemented with 10% FBS (Pan Biotech, Germany), 100 U/mL penicillin and 100 mg/mL streptomycin at 37 °C in 5% CO₂. The GsX assay was modified from Ballister et al. (Ballister et al., 2018). White 96-well plates (Greiner Bio-One) were coated with 0.1 mg/mL poly-L-lysine (Sigma Aldrich, USA) at 37 °C for 1 h prior to cell seeding. Receptor plasmids, G-protein chimeras and Glo22F (Promega) were co-transfected using Lipofectamine 2000 (Thermo Fisher, USA) and incubated at 37 °C in 5% CO₂ for 24 hours. The medium was replaced with L-15 medium (phenol red-free, 1% FBS) containing 2 mM beetle luciferin (in 10 mM HEPES, pH 6.9) and 10 mM 9-cis-retinal or all-trans retinal and incubated for 1 hour at room temperature in the dark. cAMP-dependent luminescence was measured using a Berthold Mithras multimode plate reader (Berthold Tech., Germany). Cells were activated with a 1-second light pulse (470 nm) using an LED light plate (Phlox Corp., France) or CoolLED pE-4000 (CoolLED, UK). Two technical duplicates were performed, and data were normalized to pre-activation baselines.

3.3 Locomotor Activity

The locomotor activity of individual male flies was monitored at 1-minute intervals using the Drosophila Activity Monitoring (DAM) system (Trikinetics, Inc., Waltham, MA), which detects infrared (IR) beam interruptions. Male flies (3 to 7 days old) were individually placed in recording glass tubes (∼5 mm diameter) containing agar/sugar medium (2% agar; 4% sucrose) filled to one-third capacity and sealed with an air-permeable plug. Activity was quantified by counting the number of infrared (IR) beam crossings in the center of each tube. All recordings were performed under constant conditions of 60% relative humidity and 25 °C, except for the "hot day” experiments. For "hot day" experiments, the DAM monitors were placed inside a programmable incubator (I-36NL; Percival Scientific, Inc., Perry, IA) to simulate temperature cycles as described by Yoshii et al. (2009) and Bywalez et al. (2012). A "normal day" was defined as 12 h light (30 µW/cm² corresponding to 100 lux) and 12 h darkness, while “short days” consisted of 8 h light (30 µW/cm²) and 16 h darkness. To assess light-off startle responses, which are brief clock-independent increases in activity after sudden light-off during the day, flies were repeatedly exposed to a 30-minute dark pulse following 1.5 h of light. The experiments were conducted under “normal day” conditions consisting of 12 h light (30 µW/cm²) and 12 h darkness. The activity data were normalized to account for differences in baseline activity between genotypes. Specifically, the mean activity during all light phases was calculated, excluding time points corresponding to the dark pulses. Each data point was then divided by this genotype-specific average to obtain normalized activity values. For all experimental conditions, data from four consecutive days were averaged, omitting the first day to reduce effects of acclimatization.

For the startle response assay (4-hour assay), flies were rendered arrhythmic by exposure to constant light (130 µW/cm² ∼500 lux) for two days. They were then exposed to repeated 4-hour light-dark cycles, each consisting of 3.5 hours of light (30 µW/cm²) followed by 30 minutes of darkness, repeated six times per day, for a total duration of three days. Startle activity was recorded during each cycle. Values from all cycles across the three days were averaged, with the first cycle excluded to avoid carryover effects from the prior light regime. Since each light phase lasted 210 minutes, data were plotted with the 30-minute dark pulse centered on the time axis, resulting in two light periods of 105 minutes each, before and after the dark pulse. For normalization, the mean activity during the 105 minutes preceding light-off was set to one, and activity after light-off was expressed relative to this baseline.

3.4 Fly lines, Primers, and Antibodies

The lines listed below were used for the experiments in this study. To minimize background effects, all lines were previously brought on the same X chromosomal background derived from the control fly CS having a y⁺ and w⁺ genotype.

3.5 Generating the Rh7-Gal4 line

The Rh7 promoter construct comprised 1,080 bp upstream of exon 1 extending into exon 2, including the translation start site located in exon 2. Due to the overall length of the target region (>8,000 bp), the construct was assembled in four fragments, which were sequentially cloned into the pPTGAL vector (Sharma et al., 2002). The four inserts were amplified from D. melanogaster genomic DNA using specific Gal4 primer pairs (see Table 2). Insert 1 and insert 4 included artificial restriction sites (Eco52 and KpnI, respectively), while all other restriction sites were naturally occurring. Cloning began with insert 4 (see Supplementary Material S2B) and proceeded stepwise with the remaining fragments. Final plasmids were amplified in E. coli NEB 10-beta and the construct integrity was confirmed by sequencing. The plasmid was injected into w¹¹¹⁸ flies by TheBestGene Inc., and the resulting transgenic line was backcrossed into the same y⁺w⁺ background as the controls.

Table 1

Table 2

3.6 Generating UAS-constructs

For the rescue experiments, a UAS-Rh7 construct and a UAS-empty (UAS-()) construct were generated as controls. The Rh7 construct was prepared from the D. melanogaster cDNA using the primers UAS-Rh7-5’ and UAS-Rh7-3’ (Table 2), digested with EcoRI and BglII enzymes and ligated into the pUASTattB vector (Bischof et al., 2007). For the UAS-empty construct, the pUASTattB vector was used unmodified. The vectors were then injected into flies with CytoSite 2R, 55C4 by the company TheBestGene Inc. All lines were brought into a +;+;Rh7⁰ background and finally crossed with flies with respective Gal4-constructs in +;+;Rh7⁰ background and assayed. The UAS-Rh7^Darkfly construct was prepared with the primers UAS-Rh7-5’ and UAS-Darkfly-3’ (Table 2) with the same steps described above. All transgenic lines were backcrossed into the same y⁺w⁺ background as the controls.

3.7 Optomotor response

Male flies were individually tested for their optomotor walking behavior. The setup comprised: (1) an upright cylinder (ø 8 cm; H 4.5 cm) with vertical black and white stripes (2.2 cm width and 4 cm height) inside for visual stimulation and (2) a transparent Plexiglas arena (ø 3 cm; H 1.5 cm) with a tube for inserting the flies. Although the arena was 1.5 cm high, the open space available for movement was only about 2 mm, limiting the flies to a single horizontal walking plane (Supplementary Material S2C). The LED-lit cylinder with 6 white and 6 black stripes rotated clockwise (CW) and counterclockwise (CCW) at 16 rpm, equivalent to a frequency of 1.6 Hz. The flies had a 2-minute acclimatization without light, followed by 2 minutes under CW rotation with light, a 1-minute rest, and then 2 minutes under CCW rotation with light. For the optomotor response (OR), the number of full circles walked and complete self-turns in the direction of the pattern movement within the two 2 minutes were added, while movements in the opposite direction were subtracted. The result was then divided by the total number of cylinder rotations (4×16).

3.8 qPCR

For quantitative PCR (qPCR), total RNA was extracted from either five heads, five brains, 10 retinas or 10 laminas of control flies. Total RNA was extracted using the Quick-RNA™ Microprep Kit (Zymo), and cDNA was synthesized with the QuantiTect Reverse Transcription Kit (Qiagen). For qPCR, three biological replicates were applied for each tissue, with two technical replicates. qPCRs were performed using the SensiFAST™ SYBR® No-ROX Kit (Bioline) in the Qiagen Rotor-Gene Q qPCR machine, with raw data fit by dynamic tube and slope correction calculation using the Rotor-Gene Q Series software. Relative mRNA levels were calculated using the Δ-CT equation, with α-tubulin used as a reference gene. qPCR was performed with Rh7 (Rh7 qPCR-5’, Rh7 qPCR-3’) and control (alphaTub-5’, alphaTub-3’) primers, each at 0.1 mM (Table 2).

3.9 Immunohistochemistry

Fluorescent immunohistochemistry was performed on 3- to 5-day-old male flies. Whole flies were fixed for 3 h in 4% paraformaldehyde in 0.5% PBST (1xPBS with 0.5%Triton X-100) followed by three 10-minute rinses with PBS. The brains were dissected in 0.5% PBST and incubated overnight at 4 °C with 5% normal goat serum (NGS) or normal donkey serum (NDS) and 0.02% NaN₃ in 0.5% PBST, depending on the host species of the secondary antibodies. Antibody solutions contained the antibody in the concentration specified in Table 3 and 5% NGS or NDS with 0.02% NaN₃ in 0.5% PBST. The brains were incubated for two days at 4 °C in primary antibody solution followed by five 10-minute washes with 0.5% PBST. Subsequently, the brains were incubated for at least 4 hours at room temperature or overnight at 4 °C with the secondary antibody solution containing the corresponding secondary antibodies (Table 3). The samples were washed five times for 10 minutes with 0.5% PBST before embedding in VectaShield 1000 mounting medium (Vector Laboratories, Burlingame, CA, USA) using object slides with spacers (∼170 µm). The high-precision coverslips (170 ± 5 µm, No.1.5H) were sealed with FixoGum. Brains with retinas attached were embedded with retinas facing the object slide to ensure easy imaging from the posterior. Object slides were stored in the dark at 4 °C until scanning. In addition, individual neurons were visualized using the MultiColor FlpOut (MCFO) technique (Nern et al., 2015), which stochastically labels single cells with distinct membrane-targeted fluorescent tags. Flippase-mediated excision of FRT-flanked stop cassettes enables sparse and color-resolved labeling, allowing detailed analysis of neuronal shape and projections. Labeled brains were processed and mounted as described above.

Table 3

3.10 In situ hybridization

For RNA in situ hybridization, we used the hybridization chain reaction v3 (HCR) method developed by Molecular Instruments (Choi et al., 2018). The flies were fixed and dissected as described for the fluorescent immunohistochemistry. The tissue was either used directly for in situ hybridization or dehydrated and stored in 99% methanol at −20 °C for up to 14 days. Dehydration was performed by sequential incubation in 30%, 50%, 70% and 90% methanol in PBS (2 minutes each). Rehydration was performed in the opposite order. In situ HCR was performed using a modified protocol from Bruce et al.(Bruce et al., 2021). In short, Rh7 probe mixes containing 20 probe pairs were designed by Molecular Instruments. Instead of using Proteinase K or detergent-based solutions, tissues were incubated in 0.5% PBST for 30 minutes. This was followed by a 30-minute incubation at 37 °C in preheated probe hybridization buffer (Molecular Instruments), and then an overnight incubation (≥20 h) at 37 °C in preheated probe hybridization buffer containing the Rh7 probe mix (12 nM). After hybridization the tissue was washed four times for 15 minutes each with preheated probe washing buffer (Molecular instruments) at 37 °C. This was followed by two washes at room temperature (5 minutes each) in 5xSSCT (5xSSC buffer with 1% Tween20). Next, the tissues were incubated for 30 minutes at room temperature in amplification buffer (Molecular Instruments) followed by incubation for at least 20 h at room temperature in amplification buffer containing the hairpins required for signal amplification. Hairpins were prepared by heating 3 µM h1 hairpins and 3 µM h2 hairpins separately in amplification buffer to 95 °C for 90 seconds. They were then cooled to room temperature for 30 minutes, protected from light, before being mixed and applied to the tissue. When hairpins were reused, the already mixed hairpins were reheated together. After five washes with 5x SSCT (2 x 5 minutes, 3 x 30 minutes) the tissues were either directly embedded in VectaShield as described above, or the normal fluorescent immunohistochemistry protocol was used for additional antibody labeling starting with blocking in 5% NGS or NDS.

3.11 Image acquisition and analysis

Images were acquired with a Leica TCS SP8 confocal microscope equipped with a photon multiplier tube and hybrid detector. A white light laser (Leica Microsystems, Wetzlar, Germany) was used for excitation. We used a 20-fold glycerol immersion objective (0.73 NA, HC PL APO, Leica Microsystems, Wetzlar Germany) for wholemount scans. We obtained 8-bit confocal stacks with 2048 × 1024 pixels with a maximal voxel size of 0.3 x 0.3 x 2 μm and an optical section thickness of 3.12 μm. Weak signals were amplified using line accumulation. Images were analyzed using Fiji. No changes in pixel intensity were applied other than linear changes in brightness and contrast.

4 Results

4.1 Phylogenetic and functional distinction of RH7 from other r-opsins

To determine the time of origin of RH7 and other r-opsins, a phylogenetic tree was carried out. This analysis clearly identifies the three clades of Drosophila rhodopsins including the RH7 clade (Figure 2A). While RH1, RH2, and RH6 form the first rhodopsin clade (Figure 2A, blue), which seems to be more distinct from the RH7 clade, the clade formed by RH3, RH4, and RH5 (Figure 2A, magenta) is more closely related to RH7. In contrast to the other two clades that include only rhodopsin homologues of other arthropods, the RH7 clade also includes homologues from Onychophora (Euperipatoides kanangrensis) and Tardigrada (e.g. Hypsibius dujardini). Rhodopsin sequences of other arthropods that are not related to Drosophila rhodopsins are shown in gray (Figure 2A). The detailed tree in higher resolution can be found in Supplementary Materials S3. The presence of RH7 in Tardigrada and Onychophora, unlike other rhodopsins, indicates an early origin within panarthropods (Figure 2B).

Figure 2

To test the functionality of RH7, we used a GsX assay designed to determine which G_α proteins can bind to a given GPCR, in this case RH7 (Ballister et al., 2018). This assay employs chimeric G_α proteins that share a core derived from G_αs, which triggers cAMP production upon activation. The C terminal 14 amino acids of each chimera, critical for GPCR binding, are replaced with sequences from other G_α subtypes, such as G_αq or G_αz. This allows each chimera to mimic the coupling specificity of a distinct G protein while ensuring that any productive GPCR interaction results in a measurable increase in cAMP levels. For this analysis, we co-expressed RH7 with various GsX chimeras in HEK cells and stimulated RH7 with brief light pulses. We observed robust activation with the G_αz chimera, regardless of whether 9-cis-retinal or all-trans-retinal was present (Figure 2C). Although there is no G_αz homolog in Drosophila, related G_α subunits such as G_αo and G_αi are expressed in the fly. Nevertheless, these results revealed two key findings: (1) RH7 can be activated by light, and (2) RH7 does not appear to signal via the G_αq-mediated pathway, unlike other Drosophila rhodopsins. These findings indicate that RH7 exhibits a signaling profile distinct from other insect rhodopsins.

4.2 Rh7⁰ mutants show decreased activity during darkness

To study the influence of RH7 on the daily activity of flies, wild-type (WT) control flies and Rh7⁰ mutants were examined under different light conditions. We found significant differences in activity profiles between the mutants and controls on standard 12-hour light/dark cycles (LD 12:12) (Figure 3A). During the day, mutants and wild-type controls showed the morning and evening activity typical of Drosophila melanogaster with a siesta in between. Even though the overall activity of the flies remained the same throughout the day, Rh7⁰ mutants were clearly less active during the night and compensated for this deficit with increased daytime activity (Figure 3A*). The startle response after light-off, a brief activity burst lasting 5-10 minutes, was greatly reduced in Rh7⁰ mutants. Daily temperature fluctuations (29 °C day, 24 °C night) further amplified the difference between Rh7⁰ mutants and wild-type controls. Under these conditions, control flies shifted their activity to the cooler dark phase, while Rh7⁰ mutants showed only a slight shift. After light-off, Rh7⁰ mutants quickly reduced activity and entered the resting phase, whereas wild-type controls remained active longer (Figure 3B, B*). Under short day conditions (LD 8:16), wild-type control flies strongly shifted their activity to the dark phase, while Rh7⁰ mutants showed a weaker shift (Figure 3C, C*).

Figure 3

4.3 Startle response to light-off persists without a functional clock

The startle response, triggered by the sudden light-off, differs from circadian-regulated evening activity and can occur during the day as well. To test the startle response, we performed 30-minute dark pulses every 1.5 hours during the light phase on a standard LD 12:12 day and recorded fly activity. We observed that the startle responses of both wild-type controls and mutants were generally lower in the morning than in the afternoon (Figure 4A). The circadian clock and thus the internal state of flies appear to influence the startle response, as can be seen in clockless per⁰¹ mutants. Per⁰¹ mutants exhibited a startle response that remains constant throughout the day. This startle response is significantly and consistently reduced in the per⁰¹;;Rh7⁰ double mutants (Figure 4A). These results show that the startle response is influenced by the circadian clock but functions independently, with RH7 playing a key role.

Figure 4

To focus solely on the startle response and exclude any circadian clock effects, flies were rendered arrhythmic by exposing them to continuous bright light (130 µW/cm²) for two days. This led to sustained activation of the blue light receptor Drosophila type CRYPTOCHROME (DCRY or CRY) and continuous degradation of TIMELESS, resulting in a standstill of the molecular clock (Deppisch et al., 2022; Dolezelova et al., 2007; Emery et al., 2000). The light was then switched off for 30 minutes every 3.5 hours for three days (4-hour assay). To quantify the behavioral response, startle reactions were normalized as described above and averaged across trials within each genotype. Under these conditions, Rh7⁰ mutants exhibited significantly reduced startle responses compared to wild-type controls (Figure 4B, top row). Notably, Rh7^MiMIC mutants, previously examined by Kistenpfennig et al.(Kistenpfennig et al., 2017) displayed a similar phenotype. qPCR analysis revealed that Rh7^MiMIC mutants retain reduced but detectable mRNA levels, indicating an additional Rh7 mutant line with altered gene expression, while Rh7⁰ mutants completely lack Rh7 transcripts (Supplementary Material S2D). This phenotype remained when the two alleles (Rh7⁰ and Rh7^MiMIC) were tested in trans to each other but disappeared in heterozygous x/+ flies, confirming the critical role of RH7 in mediating the startle response (Figure 4B, bottom row).

4.4 Rh7 is expressed in the optic lobe and clock neurons

To determine the tissue-specific expression of Rh7, we performed qPCR analyses using α-tubulin as a reference gene. The expression of Rh7 was highest in the lamina, followed by the brain, while only minimal levels were detected in the retina (Figure 5A). Additionally, we performed Rh7 RNA in situ hybridization chain reaction (HCR), which verified the strongest Rh7 expression in the lamina and medulla (Figure 5B). We further observed considerably weaker expressions in neurons of the central brain (Figure 5B). No in situ hybridization was detected in Rh7⁰ mutants (Figure 5C).

Figure 5

To further investigate Rh7 expression, we generated a Rh7 promoter-driven Gal4 line (Rh7-Gal4). Since the Rh7 gene expression regulatory elements are not only located within the promoter but also extend into the transcriptional region, we cloned a large part of this region, including exon 1, intron 1 and part of exon 2, together with the start codon (Supplementary Material S2B).

When comparing GFP expression driven by the newly generated Rh7-Gal4 with the Rh7 RNA in situ hybridization and qPCR data, we again observed the strongest expression in the lamina and medulla. Interestingly, also the Rh7-Gal4 line showed additional expression in neurons of the central brain (Figure 5D). Some of the stained dorsally located neurons are part of the clock network as can be seen from the colocalization with the antibody against the clock protein PERIOD. This subset of neurons includes some DN_1p, DN_1a, and DN₂ neurons (Figure 5H). To test whether these clock neurons also express cry, we performed simultaneous RNA in situ hybridization for cry and Rh7. As suggested by single cell RNA sequencing data (Ma et al., 2021; Reinhard et al., 2024), the clock neurons expressing Rh7 are CRY-negative with the DN_1a forming the only exception (Supplementary Material S2E). To validate the expression pattern of the Rh7-Gal4 line, we performed in situ hybridization with Rh7-Gal4>GFP flies. All GFP-labeled cells were also detected by RNA in situ hybridization, although the signal was less pronounced in cells of the central brain. Some cells, particularly in the optic system, were detected only by in situ hybridization and not by the Gal4 line. This suggests an even broader expression of Rh7 than shown by the Gal4 line (Figure 5F, G). To further characterize Rh7-expressing neurons, we used MultiColor FlpOut (MCFO) (Nern et al., 2015), which stochastically labels neurons when crossed to a Gal4-driver line. Using MCFO, we identified at least two cell types of the optic system based on their characteristic arborization pattern in the medulla: lamina neuron 5 (L5) and medulla intrinsic neuron 1 Mi1 (Figure 5E). Both neurons were also present in the TAPIN RNA-seq analysis, with L5 showing the strongest Rh7 expression (Davis et al., 2020).

4.5 C-terminal truncation of RH7 affects dark-phase behavior

The so-called Darkfly was kept in complete darkness for 57 years (1400 generations) accumulating mutations that may have enhanced its adaptation to perpetual darkness (Izutsu et al., 2012). Notably, the mutant carries a nonsense mutation in the Rh7 gene, resulting in the loss of the final 21 amino acids at the C-terminus. qPCR analyses indicate that Rh7 transcript levels are unchanged, suggesting that the truncated RH7 protein is indeed expressed (Supplementary Material S2D). When we exposed the Darkfly to dark pulses in a 12:12 light-dark cycle, they exhibited the opposite phenotype of Rh7⁰ mutants, showing a stronger startle response (Figure 6A), possibly reflecting increased RHODOPSIN 7 activity in the Darkfly strain (Figure 6A).

Figure 6

To validate the role of Rh7 in the startle response, we conducted a rescue experiment by re-expressing Rh7 in Rh7⁰ mutants using the UAS-Rh7 construct driven by Rh7-Gal4. As a control for cloning and insertion artifacts, we used a UAS-empty line with the same insertion site as UAS-Rh7, driven by Rh7-Gal4 in the Rh7⁰ background. In the dark pulse experiment, Rh7-Gal4>UAS-empty;Rh7⁰ flies behaved like mutants, while Rh7-Gal4>UAS-Rh7;Rh7⁰ flies resembled wild-type controls (Figure 6B). To express a truncated RH7 protein lacking the last 21 amino acids, as found in the Darkfly, we crossed the UAS-Rh7^Darkfly; Rh7⁰ line with the Rh7-Gal4; Rh7⁰ line. We observed an increased startle response in the offspring, most prominently during the second half of the light phase (Figure 6B). In the 4-hour assay, in which the flies were made arrhythmic we found a clear rescue of the startle response in Rh7-Gal4>UAS-Rh7;Rh7⁰ flies compared to the Rh7-Gal4>UAS-empty;Rh7⁰ controls. Furthermore, Rh7-Gal4>UAS-Rh7^Darkfly;Rh7⁰ flies showed a significantly increased startle response (Figure 6C).

4.6 RH7 modulates optomotor responses

The Drosophila optomotor response is a reflex that stabilizes locomotion by counteracting perceived environmental motion with its own motion. The optic system, particularly the lamina neurons, plays an important role in this response (Yang and Clandinin, 2018). We found that Rh7⁰ mutants show deficits in the optomotor response, while the Darkfly showed an increased optomotor response (Figure 6D). When the rescue flies were tested, the Rh7-Gal4>UAS-empty;Rh7⁰ flies behaved like the Rh7⁰ mutants, whereas the Rh7-Gal4>UAS-Rh7;Rh7⁰ flies showed a significant rescue, with no significant difference from wild-type controls. In contrast, Rh7-Gal4>UAS-Rh7^Darkfly;Rh7⁰ flies showed a significantly higher optomotor response than the wild-type controls and the Rh7-Gal4>UAS-Rh7;Rh7⁰ flies. This response was similar to the Darkfly (Figure 6E).

5 Discussion

5.1 RH7 is an ancestral and structurally distinct opsin

Although RH7 has been known for nearly 25 years, its function remains poorly understood. This is partly due to its atypical expression outside the retina and its divergence from canonical visual rhodopsins. Phylogenetic analyses indicate that RH7 is the most ancestral opsin within the Drosophila genus. Homologs identified in tardigrades and onychophorans (Figure 2A) point to a deep evolutionary origin within Panarthropoda, suggesting that RH7 belongs to an ancient class of opsins (Figure 2B). In onychophorans and tardigrades, two early branching panarthropod groups, RH7-like opsins have been linked to monochromatic vision (Beckmann et al., 2015; Fleming et al., 2021; Hering and Mayer, 2014).

The ancestral character of RH7 is further reflected in its unusual structural features. Unlike other Drosophila rhodopsins, RH7 has markedly longer N- and C-terminal regions, which are roughly twice as long and contain a high density of predicted phosphorylation sites. This suggests a high potential for post-translational regulation and possibly increased signaling flexibility. The extended termini distinguish RH7 from the more compact structure of typical rhodopsins and may represent an intermediate stage between canonical GPCRs and specialized visual rhodopsins (Figure 7A). In addition, RH7 lacks several conserved motifs found in other Drosophila rhodopsins, which likely emerged during the specialization of visual opsins. The absence of these features, together with the extended termini, supports the idea that RH7 occupies a structurally and evolutionarily distinct position. A more detailed comparison of these motifs and their functional implications is presented in section 5.5.

Figure 7

Besides its structural divergence, the developmental origin and central localization of RH7-expressing neurons support its ancestral photoreceptive role. RH7 localizes to central visual neuropils such as the lamina and medulla that process input from peripheral photoreceptors. Unlike canonical R-opsins, which are primarily expressed in specialized retinal photoreceptor cells, RH7 expression in central brain regions suggests an ancient form of non-image-forming photoreception integrated directly within neural circuits. This idea is further supported by the presence of opsins expressed in brain regions of other early-diverging bilaterians, such as annelids and mollusks (Arendt et al., 2004; Vöcking et al., 2015; Wollesen et al., 2019). RH7 may thus represent an evolutionarily conserved opsin for central light sensing that predates the functional separation of sensory input and processing centers.

5.2 Divergence from canonical signaling: first insights into RH7 phototransduction

The signaling properties of RH7 were explored using heterologous GsX assays, which clearly indicated that RH7 can be activated by light (Figure 2C) but does not interact with G_αq. This lack of interaction appears plausible, as G_αq expression is restricted to retinal cells, whereas RH7 is expressed in lamina and medulla neurons (Davis et al., 2020). The G_αq-binding domains are highly conserved between mammals and Drosophila (93% identity, Supplementary Material S2F), meaning that RH7 should interact with mammalian G_αq if it engages in a G_αq-mediated signaling pathway. Instead of interacting with G_αq, RH7 appears to interact with mammalian G_αz. Although Drosophila lacks a direct homolog of G_αz, the related inhibitory subunits G_αi and G_αo from the same family may serve as functional analogs.

If RH7 indeed couples to a G_αi family member, this would represent a notable exception among Drosophila rhodopsins by engaging an inhibitory G protein pathway. This possible reversal of signaling direction recalls the mammalian system, where melanopsin in ipRGCs activates an excitatory G_αq pathway, while classical retinal rhodopsin signals through an inhibitory transducin (G_αt) pathway belonging to the G_αi family (Contreras et al., 2021; Gao et al., 2019; Koyanagi and Terakita, 2008). The use of opposing G protein pathways in these systems, despite their evolutionary distance, may reflect a conserved principle that links photoreceptor polarity to downstream signaling.

5.3 RH7 in the optic lobe: A modulator of early visual processing

Rh7 is broadly expressed in the optic lobe, particularly in the lamina and medulla, which are key regions for early visual processing. These areas receive input from outer photoreceptors and are involved in contrast and motion detection. We found strong Rh7 expression in L5 lamina neurons and Mi1 neurons of the medulla. L5 neurons pass signals from outer photoreceptors to the medulla, where Mi1 neurons help process motion and contrast information.

Our behavioral data support a role for RH7 in modulating these visual functions. Flies lacking Rh7 show reduced optomotor responses, suggesting they are less sensitive to moving visual stimuli. In contrast, the Darkfly strain, which expresses a truncated variant of RH7, shows an increased response in the same assay (Figure 6D). Rescue experiments confirm that this effect depends on Rh7: expression of full-length RH7 in Rh7⁰ mutants restores normal behavior, while the truncated form of the Darkfly enhances the optomotor response. These results indicate that RH7 influences visual processing in the optic lobe. The enhanced behavior seen with the Darkfly variant suggests that changes in RH7 structure can affect its function. This possibility is discussed further in section 5.5.

5.4 RH7 in circadian regulation and behavioral flexibility

Fly activity is shaped by both the circadian clock and external light cues, which are detected by specialized photoreceptors. In Drosophila, RH7 contributes to this light-dependent regulation. Under standard 12:12 light-dark cycles, Drosophila melanogaster displays two peaks of locomotor activity in the morning and evening separated by a midday siesta (Helfrich-Förster, 2000). This rhythm is controlled by the circadian clock and its associated neuronal network. Flies experience two rest phases: a pronounced one at night and a shorter one during the siesta, both primarily regulated by dorsal clock neurons (DNs) (Guo et al., 2018). The DN₁ subclass plays a key sleep-promoting role and includes CRY-positive and CRY-negative neurons that may contribute differently to the two rest phases (Guo et al., 2016; Lamaze and Stanewsky, 2020).

Although Drosophila is generally active during the light phase, activity can shift into the dark phase under specific conditions such as heat stress or short photoperiods. This behavioral flexibility is reduced in Rh7⁰ mutants. They show decreased night activity and fail to adjust their behavior to darkness, pointing to a role for RH7 in conveying light input to the circadian system (Figure 3). Daily activity levels remain similar between mutants and wild-type controls, suggesting that these effects reflect changes in circadian regulation rather than general locomotor impairment.

RH7 is expressed in dorsal clock neurons, especially in CRY-negative DN₁, with the exception of the two DN_1a neurons (Figure 5H). Notably, nearly all sequenced species that possess RH7 also contain DCRY (Deppisch et al., 2023; Senthilan and Helfrich-Förster, 2016), suggesting a potential functional interplay between these two photoreceptors within the circadian light input network.

While we could confirm RH7 expression in the dorsal clock neurons, we did not detect it in lateral clock neurons (Figure 5B, D), as previously suggested by Ni et al. (Ni et al., 2017). This finding is consistent with findings reported in Senthilan et al. (Senthilan et al., 2019). In addition to its role in dorsal neurons, RH7 may also influence circadian regulation via the optic lobe. Its expression in the lamina and medulla, along with possible synaptic connections between Mi1 neurons and PDF-expressing neurons (Reinhard et al., 2024), and the presence of PDF receptors in lamina neurons (Bastian et al., 2025; Davis et al., 2020), suggest a functional link between early visual processing centers and the circadian network.

5.5 RH7 is essential for locomotor activity during darkness

Although activity during darkness is regulated by the circadian clock, the startle response to light-off is independent of it. Dark-pulse experiments show that both wild-type controls and per⁰ mutants, which lack a functional circadian clock, exhibit a strong startle response. While the circadian clock seems to influence the magnitude of the response, it is not essential for its occurrence (Figure 4A). Moreover, this phenotype persists in flies rendered arrhythmic during the 4-hour assay. The critical role of RH7 is supported by trans-heterozygous combinations with Rh7^MiMIC, which show the same phenotype as Rh7⁰ mutants, and by rescue experiments restoring normal behavior (Figure 4B).

The Darkfly strain, bred in complete darkness for many generations, provides an intriguing case for RH7 function in sustained darkness. It carries an RH7 variant with a 21-amino acid deletion in the C-terminal tail, a region rich in predicted phosphorylation sites. Structural modeling based on AlphaFold predicts an interaction between this tail and the receptor core, specifically the unique ICL3 loop of RH7 (Figure 1A, Supplementary Material S2A), which lacks the QAKKMNV motif found in other Drosophila rhodopsins. Such intramolecular interactions are known to mediate autoinhibition in GPCRs, where the C-terminal tail stabilizes an inactive conformation that shifts to an active state upon receptor stimulation (Maggio et al., 2023).

We hypothesize that the deletion in the Darkfly variant weakens this autoinhibitory interaction and thus facilitates constitutive activation (Figure 7B). This is supported by the observation that deleting the final 21 amino acids increases dark-phase activity and optomotor response, approaching levels seen in Darkfly flies (Figure 6). However, while the RH7 mutation largely accounts for the increased dark activity, the enhanced phenotype in the Darkfly strain likely reflects the combined effects of additional mutations accumulated during long-term dark rearing (Izutsu et al., 2012), which may also influence neural development or synaptic connectivity. Overall, these findings highlight the pivotal role of RH7 in modulating behavior in darkness and demonstrate that structural changes can fine-tune RH7 signaling to adapt to environmental conditions.

5.6 Parallels and divergence between Drosophila RH7 and other noncanonical Opsins

This study highlights significant similarities between RH7 and other noncanonical rhodopsins, particularly mammalian melanopsin (OPN4). Both RH7 and melanopsin are expressed in neurons downstream of classical photoreceptors, with RH7 in lamina and medulla neurons of the optic lobe and melanopsin in the ipRGCs, suggesting analogous roles in processing visual and circadian information. Functionally, both opsins contribute to circadian regulation as well as contrast and motion detection. Structurally, they share elongated N- and C-terminal regions that are intermediate in length relative to canonical rhodopsins and typical GPCRs. These extended regions harbor multiple phosphorylation sites likely crucial for their unique signaling mechanisms.

Furthermore, RH7 and melanopsin activate phototransduction cascades distinct from those of classical visual opsins, potentially producing opposing effects. This reflects an evolutionary adaptation that supports diverse functions in both visual and nonvisual photoreception. However, while melanopsin expression in mammals is largely confined to the visual system, Drosophila RH7 appears to have acquired additional functions in the brain. This divergence may reflect evolutionary adaptations to differing anatomical constraints: the small, light-permeable brain of Drosophila may permit direct light sensing beyond the eyes, whereas in mammals, the evolution of a larger brain enclosed by a light-impermeable skull may have limited such extraocular photoreception to specialized retinal pathways.

Although melanopsin is predominantly expressed in retinal ganglion cells, it has also been detected in various brain regions (Aviles-Trigueros et al., 2012; Nissilä et al., 2017). Additionally, other non-canonical opsins such as encephalopsin and neuropsin are present in the human brain, with their functional roles still under investigation (Buhr et al., 2015; Nissilä et al., 2012; Upton et al., 2022, 2021). In this context, the expression of RH7 in the Drosophila brain, outside of clock neurons, may reflect similar non-visual roles. While its brain expression is relatively weak compared to its prominent presence in the optic lobes, it is consistently observed using both Rh7-Gal4 reporters and in situ hybridization. The identity of the non-clock neuronal populations expressing Rh7 remains to be determined, and further work, such as double labeling with cell-type-specific markers, will be necessary to clarify this. In these neurons, RH7 might contribute to higher-order visual processing, sensorimotor control, general arousal, or aspects of neuronal development.

These findings open new possibilities to study how noncanonical opsins like RH7 contribute to complex brain functions beyond just sensing light. Understanding the roles of RH7 can give important clues about how light-sensing systems evolved and how they work together with brain circuits to control behavior. Future research using genetics, anatomy, and physiology will be important to fully understand all the functions of RH7 in Drosophila and other animals.

6 Conclusion

Our study provides a key insight into the unresolved functional diversity of opsins. Unlike canonical rhodopsins involved in image formation, RH7 acts as a fundamental photoreceptor mediating simple light-on and light-off signals to diverse neuronal populations. As an evolutionarily ancient opsin, RH7 likely represents a primordial photoreceptive system predating the specialization of complex visual pathways. Structurally and functionally, RH7 exhibits characteristic features of G protein-coupled receptors (GPCRs), positioning it as an evolutionary intermediate between ancestral GPCRs and specialized visual opsins. This is reflected in its atypical sequence motifs, extended terminal domains, and distinct signaling properties, which collectively blur the distinction between classical visual and non-visual photoreception.

Functionally, RH7 modulates early visual processing in the lamina and might contribute to circadian regulation in dorsal clock neurons. Its phylogenetic and mechanistic parallels to mammalian melanopsin further support its role as a GPCR-like photoreceptor integrating environmental light cues with intracellular signaling cascades to regulate physiology and behavior. In summary, RH7 exemplifies how ancient opsins have diversified to couple photoreception with GPCR-mediated signaling beyond image formation. Its expression in brain regions highlights a conserved evolutionary strategy: the integration of light sensitivity into central neural circuits, underscoring the deep-rooted link between sensory and neurodevelopmental architectures.

Acknowledgements

We thank Kathrin Sauter, Christina Brunner, and Marina Müller for technical assistance. We also acknowledge Maria Gallant and Barbara Mühlbauer for their support, and Giulia Manoli for help with data analysis. We are grateful to Naoyuki Fuse for providing Darkfly, to E. Kostenis and A. Inoue for the HEK293ΔG7 cell line, and to Martin Göpfert and Marion Silies for valuable discussions. This work was supported by the German Research Foundation (DFG), SE 2320/1-1, SO 1337/2-2, and the DFG Heisenberg program SO 1337/6-1.

Additional information

Data availability

All data supporting this study are available within the main text or the Supplemental Information. Additional details can be obtained from the corresponding author upon request.

Author contributions

V.K. performed fly activity assays, optomotor experiments, generated UAS-construct lines, and carried out MCFO stainings and qPCR analyses. N.R. performed immunohistochemistry and in situ hybridizations. H.H. generated the Rh7-Gal4 line. A.M. contributed additional fly activity assays. D.R. established experimental setups and contributed to experimental design. P.S. carried out the GsX assay. C.H.-F. contributed to conceptual discussions, provided critical input throughout the project, and performed dissections for qPCR. P.R.S. designed the experiments, coordinated the study, analyzed experimental results, performed phylogenetic analyses, and wrote the manuscript. All authors reviewed and approved the final manuscript.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work the authors used DeepL and ChatGPT in order to improve the readability and language of the manuscript. After using these tools, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article.

Funding

Deutsche Forschungsgemeinschaft (DFG) (SE 2320/1-1)

• Pingkalai R Senthilan

Deutsche Forschungsgemeinschaft (DFG) (SO 1337/2-2)

• Peter Soba

Deutsche Forschungsgemeinschaft (DFG) (SO 1337/6-1)

• Peter Soba

Additional files

Supplementary Material S1

Supplementary Material S2

Supplementary Material S3